Determination of Royalty Rates and Terms for Making and Distributing Phonorecords (Phonorecords III), 54406-54486 [2023-14925]

Download as PDF 54406 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations LIBRARY OF CONGRESS Copyright Royalty Board 37 CFR Part 385 [Docket No. 16–CRB–0003–PR (2018–2022) (Remand)] Determination of Royalty Rates and Terms for Making and Distributing Phonorecords (Phonorecords III) Copyright Royalty Board, Library of Congress. ACTION: Final rule and order. AGENCY: The Copyright Royalty Judges announce their final determination after remand of the rates and terms for making and distributing phonorecords for the period beginning January 1, 2018, and ending on December 31, 2022. DATES: Effective date: August 10, 2023. Applicability date: The regulations apply to the license period beginning January 1, 2018, and ending December 31, 2022. ADDRESSES: The final determination after remand is posted in eCRB at https://app.crb.gov/. For access to the docket to read the final determination after remand and submitted background documents, go to eCRB and search for docket number 16–CRB–0003–PR (2018–2022) (Remand). FOR FURTHER INFORMATION CONTACT: Anita Brown, CRB Program Assistant, (202) 707–7658, crb@loc.gov. SUPPLEMENTARY INFORMATION: SUMMARY: Final Determination After Remand ddrumheller on DSK120RN23PROD with RULES2 On October 26, 2020, the United States Court of Appeals for the D.C. Circuit (D.C. Circuit) issued its mandate vacating and remanding in part the original Determination 1 issued by the Copyright Royalty Judges (Judges) in the captioned proceeding. See Johnson v. Copyright Royalty Board, 969 F.3d 363 (D.C. Cir. 2020). In its ruling on appeal, the D.C. Circuit found that in the original Determination, the Judges (1) failed to give adequate notice to participants of their overhaul of the royalty rate structure combined with significantly increased and uncapped rates for section 115 licenses; (2) failed 1 Determination of Royalty Rates and Terms for Making and Distributing Phonorecords (Phonorecords III), 84 FR 1918 (Feb. 5, 2019) (final rule and order) (original Determination); see also Final Determination, 16–CRB–0003–PR (2018– 2022) (Nov. 5, 2018). The original Determination was issued by two of the Judges (Majority) and was accompanied by a dissenting opinion (Dissent) authored by the third Judge. The Dissent is appended to and part of the same document as the original Determination. VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 to explain why they rejected a benchmark based on a past settlement agreement 2 in lieu of overhauling of the rate structure and significantly increasing rates; and (3) failed to identify their legal authority to redefine a material term after they promulgated a definition of that term in the original Initial Determination circulated to the participants. See Johnson, 969 F.3d at 367, 381; Initial Determination, Determination of Royalty Rates and Terms for Making and Distributing Phonorecords (Phonorecords III), 16– CRB–0003–PR (2018–2022) (Jan. 27, 2018). After receipt of the D.C. Circuit’s ruling and mandate, the Judges consulted with the parties to the appeal and established procedures for the remand proceeding. See Order Adopting Schedule for . . . Remand (Dec. 23, 2020).3 Each side offered opening submissions, responsive submissions, additional evidentiary filings, and further supplemental briefing requested by the Judges. The parties’ submissions included legal briefing and incorporated evidence from the original proceeding as well as evidence newly developed for the remand proceeding. After preliminary deliberations, the Judges asked for supplemental briefing from the parties responsive to a proposed alternative rate structure. See Notice and Sua Sponte Order Directing the Parties to Provide Additional Materials (Dec. 9, 2021). With respect to redefinition of the material term Bundled Revenue, the Judges also sought legal analysis from the parties relating to the D.C. Circuit’s directive that the Judges either provide ‘‘a fuller explanation of the agency’s reasoning at the time . . .’’ or take ‘‘new agency action accompanied by the appropriate procedures.’’ See Johnson, 969 F.3d at 392 (citing Department of Homeland Security v. Regents of the Univ. of Cal., 140 S. Ct. 1891, 1908). On February 9, 2022, the Judges invited additional briefing on the Bundled Revenue definition issue, specifically permitting the parties to offer additional analysis of possible characterization of the 2 The referenced settlement agreement formed the basis for regulatory terms relating to section 115 musical works royalties and was adopted as a final rule in Adjustment [or] Determination of Compulsory License Rates for Mechanical and Digital Phonorecords, Docket No. 2011–3 CRB Phonorecords II, 78 FR 67938 (Nov. 13, 2013). See also Technical Amendment at 78 FR 76987 (Dec. 20, 2013). 3 Following the original remand scheduling order, the Judges amended the remand proceeding schedule by, e.g., permitting additional briefing, changing due dates, and seeking additional input with regard to specific issues. See, e.g., Order . . . Modifying Scheduling Orders (Dec. 13, 2021) (eCRB no. 25973). PO 00000 Frm 00002 Fmt 4701 Sfmt 4700 Copyright Owners’ motion for clarification following the Determination as a motion for rehearing under the Copyright Act, title 17, United States Code at sec. 803(c)(2). See Sua Sponte Order Regarding Additional Briefing (Feb. 9, 2022). At the request of the parties, the Judges agreed to forego live testimony. On March 8, 2022, all parties were afforded an opportunity to present oral argument on all remand issues.4 On July 1, 2022, the Judges issued an Initial Ruling and Order after Remand (Initial Ruling) 5—applying Johnson and considering the entire record developed pre-remand and post-remand. In the Initial Ruling, the Judges directed the parties to attempt to submit jointly agreed-upon regulatory provisions implementing the Initial Ruling for the Judges to consider. The Judges further ruled that, if the parties could not agree on all the regulatory language, they should make separate submissions regarding regulatory provisions in dispute. See Initial Ruling at 114. The parties agreed to many regulatory provisions but disagreed as to several such provisions. Accordingly, they filed separate submissions and respective replies regarding the regulatory provisions. Services’ Joint Submission of Regulatory Provisions (July 18, 2022); Copyright Owners’ Submission of Regulatory Provisions to Implement the Initial Ruling (July 18, 2022); Services’ Joint Response to Copyright Owners’ Submission of Regulatory Provisions (Aug. 5, 2022); Copyright Owners’ Response to Judges’ July 27, 2022 Order Soliciting Responses Regarding Regulatory Provisions (Aug. 5, 2022). The Judges considered those submissions and entered an order addressing the disputed regulatory provisions. See Corrected Order regarding Regulatory Provisions 4 Copyright Owners and Services divided the time for oral argument. George Johnson dba GEO Music Group waived oral argument. 5 The Initial Ruling (eCRB no. 26938) is included in Related Rulings and Orders as section A. The findings and conclusions in the Initial Ruling were adopted by a majority of the Judges, but two Judges filed separate opinions. See Initial Ruling at 2 n.5. One Judge, former Chief Judge Suzanne Barnett, dissented from the Majority’s conclusion in the Initial Ruling regarding the Phonorecords II rate structure (section II of the Initial Ruling), though not from the exception to that benchmark with regard to the headline rate of 15.1% and the imposition of a cap on the TCC rate prong. See Dissent in Part re Benchmark (July 1, 2022) (eCRB no. 26943). The other opinion was issued by Judge Strickler, who dissented from the reasoning relating to the adoption of the definition of Service Revenue (section V), but concurred in the adoption of that definition. See Dissent in Part as to Section IV of the Initial Ruling and Order after Remand . . . (July 1, 2022) (eCRB no. 26965). E:\FR\FM\10AUR2.SGM 10AUR2 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations ddrumheller on DSK120RN23PROD with RULES2 Following Initial Ruling and Order (after Remand) (Nov. 10, 2022) (November 10th Order).6 On November 30, 2022, the parties filed a Joint Submission in which they provided joint regulatory language no longer in dispute that applied the binding rulings of the Judges and the D.C. Circuit.7 However, the parties identified the single issue in dispute that relates to the ‘‘Total Content Cost’’ (‘‘TCC’’) rates for nine offerings made by interactive streaming services. Joint Submission . . . Regarding Regulatory Provisions Following Initial Ruling and Order (after Remand) (Nov. 30, 2022) (Joint Submission) (eCRB no. 27337). Having considered the parties’ submissions (including the Joint Submission), the Initial Ruling, and all other pertinent material, the Judges adopted the several TCC rates set forth in the Phonorecords II-based benchmark as proposed by the Services. See Order 43 on Phonorecords III Regulatory Provisions (eCRB no. 28210).8 Based on the entirety of the record, the Judges adopt in toto 9 the Initial Ruling and the Order 43 on Phonorecords III Regulatory Provisions which are set out in this document. Accordingly, those two documents are adopted by reference in this Final Determination After Remand. Additionally, the regulatory terms that will codify this Final Determination After Remand are set out in this document.10 6 The November 10th Order corrected an otherwise substantively identical order issued two days earlier, on November 8, 2022, which had inadvertently included a small amount of text. See November 10th Order at 1 (eCRB no. 27312). 7 The Judges largely adopt the regulations in the Joint Submission, which reflect the substance of the Judges’ post-remand rulings, the substance and formatting that the Judges had adopted in the preremand Final Determination that were not raised as issues on appeal, and updates to references to subparagraphs of Section 115 to conform to statutory amendments made pursuant to the Music Modernization Act in 2018. Any differences in language or style are made for ease of reference, consistent with the parties’ post-remand joint filings. 8 The Judges also found good cause to adopt a joint proposal for modified language regarding late fees, in 37 CFR 385.3. Order 43 on Phonorecords III Regulatory Provisions at 9. 9 But see Judge Strickler’s Dissent, cited at n.5 supra, in which—although he agrees with the Majority as to the definition of a Service Revenue Bundle—he disagrees as to the legal reasoning supporting that conclusion. 10 The documents are: Initial Ruling and Order After Remand, designated as Related Rulings and Orders, section A; Order 43 on Phonorecords III Regulatory Provisions, designated as Related Rulings and Orders, section B; Dissent in Part as to Section IV of the Initial Ruling and Order after Remand by Judge David R. Strickler, designated as Related Rulings and Orders, section C; and Dissent in Part re Benchmark, designated as Related Rulings and Orders, section D. VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 On the basis of the foregoing, the Judges propound the rates and terms described in this Final Determination After Remand for the period January 1, 2018, through December 31, 2022.11 No participant having filed a timely petition for rehearing, the Judges have made no substantive alterations to the body of the Initial Determination After Remand. The Register of Copyrights reviewed the Judges’ Final Determination After Remand for legal error in resolving a material issue of substantive law under title 17, United States Code, and has closed her review. Non-substantive typos have been corrected and non-substantive formatting changes have been made to the version reviewed by the Register in order to accommodate the Federal Register’s formatting standards. The Librarian shall cause the Judges’ Final Determination After Remand, and any correction thereto by the Register, to be published in the Federal Register no later than the conclusion of the Register’s 60-day review period. Related Rulings and Orders A. Initial Ruling and Order After Remand (Redacted Version With Federal Register Naming and Formatting Conventions) On October 26, 2020, the United States Court of Appeals for the D.C. Circuit (D.C. Circuit) issued its mandate vacating and remanding in part the Determination 12 issued by the Copyright Royalty Judges (Judges) in the captioned proceeding. See Johnson v. Copyright Royalty Board, 969 F.3d 363 (D.C. Cir. 2020). In its ruling on appeal, the D.C. Circuit found that in the Determination, the Judges (1) failed to 11 The regulations applicable to the period 2018 through 2022, as set forth following this SUPPLEMENTARY INFORMATION section, will appear in the CFR as appendix A to the current regulations. Although these Phonorecords III regulations adopt the substance of the Phonorecords II-based benchmark where the Judges so require, in §§ 385.21 and 385.22, these Phonorecords III regulations are structured, consistent with the parties’ Joint Submission, in the same consolidated manner as set forth in the pre-remand Phonorecords III regulations (a structure as to which no party appealed). See Exhibit A to the Joint Submission at 16, n. 47; see also Exhibit B to the Joint Submission at n.17 (red-lined version of Exhibit A, supra). 12 Determination of Royalty Rates and Terms for Making and Distributing Phonorecords (Phonorecords III), 84 FR 1918 (Copyright Royalty Board Feb. 5, 2019) (final rule and order) (‘‘Determination’’); See also Final Determination, 16–CRB–0003–PR (2018–2022) (Nov. 5, 2018) (citations to the Determination and to the Dissent in this Initial Ruling and Order after Remand (Initial Ruling) are found in this document). The Determination was issued by two of the Judges (Majority) and was accompanied by a dissenting opinion (Dissent) authored by the third Judge. The Dissent is appended to and part of the same document as the Determination. PO 00000 Frm 00003 Fmt 4701 Sfmt 4700 54407 give adequate notice to participants of their overhaul of the royalty rate structure combined with significantly increased and uncapped rates for section 115 licenses; (2) failed to explain why they rejected a benchmark based on a past settlement agreement 13 in lieu of overhauling of the rate structure and significantly increasing rates; and (3) failed to identify their legal authority to redefine a material term after they promulgated a definition of that term in the Initial Determination circulated to the participants. See Johnson, 969 F.3d at 367, 381; Initial Determination, Determination of Royalty Rates and Terms for Making and Distributing Phonorecords (Phonorecords III), 16–CRB–0003–PR (2018–2022) (Jan. 27, 2018). After receipt of the D.C. Circuit’s ruling and mandate, the Judges consulted with the parties to the appeal and established procedures for the remand proceeding. See Order Adopting Schedule for . . . Remand (Dec. 23, 2020).14 Each side offered opening submissions, responsive submissions, additional evidentiary filings and further supplemental briefing requested by the Judges. The parties’ submissions included legal briefing and incorporated evidence from the original proceeding as well as evidence newly developed for the remand proceeding. After preliminary deliberations, the Judges asked for supplemental briefing from the parties responsive to a proposed alternative rate structure. See Notice and Sua Sponte Order Directing the Parties to Provide Additional Materials (Dec. 9 Order). The Judges also sought legal analysis from the parties relating to the D.C. Circuit’s directive that the Judges either provide ‘‘a fuller explanation of the agency’s reasoning at the time . . .’’ or take ‘‘new agency action accompanied by the appropriate procedures.’’ See Johnson, 969 F.3d at 392 (citing Dep’t of Homeland Sec. v. Regents of the Univ. of Cal., 140 S. Ct. 1891, 1908 (Regents)). On February 9, the Judges invited additional briefing on 13 The referenced settlement agreement formed the basis for regulatory terms relating to section 115 musical works royalties and was adopted as a final rule in Adjustment of Determination of Compulsory License Rates for Mechanical and Digital Phonorecords, Docket No. 2011–3 CRB Phonorecords II, 78 FR 67938 (Nov. 13, 2013), Technical Amendment at 78 FR 76987 (Dec. 20, 2013). In this Initial Ruling, references to Phonorecords II, PR II, and PR II-based benchmark are references to this final rule. 14 Following the original remand scheduling order, at the request of parties or on their own motion, the Judges amended the remand proceeding schedule by, e.g., permitting additional briefing, changing due dates, and seeking additional input with regard to specific issues. See, e.g., Order . . . Modifying Scheduling Orders (Dec. 13, 2021). E:\FR\FM\10AUR2.SGM 10AUR2 54408 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations the service bundle definition issue, specifically permitting the parties to offer additional analysis of possible characterization of the Copyright Owners’ motion for clarification following the Determination as a motion for rehearing under the Copyright Act, title 17, United States Code (Act) at sec. 803(c)(2). At the request of the parties, the Judges agreed to forego live testimony. On March 8, 2022, all parties were afforded an opportunity to present oral argument on all remand issues.15 Following oral argument, the Judges deliberated and now issue this Initial Ruling after Remand. After due consideration of all of the evidence and oral argument of counsel, the Judges 16 determine: 17 (1) With regard to the applicable rates and rate structure, the percent-ofrevenue all-in headline royalty rate for the mechanical license shall be set at 15.1%, phased-in, as set forth below: 2018–2022 ALL-IN HEADLINE ROYALTY RATES Percent of Revenue ................................................................................. 2019 2020 2021 2022 11.4% 12.3% 13.3% 14.2% 15.1% the D.C. Circuit has already held that the fact that some of the Streaming Services’ proposals contemplated continued use of an uncapped total content cost prong for some categories ‘‘does not mean they anticipated that the [Judges] would uncap the total content cost prong across the board . . . [which] is quite different.’’ Johnson, 369 F.3d at 382. The difference, according to Johnson, is that ‘‘[u]ncapping the total content cost prong across all categories leaves the Streaming Services exposed to potentially large hikes in the mechanical license royalties they must pay.’’ Id. Accordingly, the Judges find that Copyright Owners indeed do bear the burden of proof with regard to the appropriateness of uncapped rate structure and rates they are proposing on remand and the Services bear the burden of proof with regard to the appropriateness of the Phonorecords IIbased benchmark they are continuing to advance on remand. As a preliminary matter, the Judges address the issue of burden of proof raised by both parties. Pursuant to the Administrative Procedure Act (APA), ‘‘the proponent of a rule or order has the burden of proof.’’ 5 U.S.C. 556(d). See also Initial Remand Submission of Copyright Owners at 48 (Apr. 1, 2021) (‘‘CO Initial Submission’’) (citing section 556(d) of the APA as setting forth ‘‘a basic rule of these rate-setting proceedings that a participant is required to provide evidence establishing the propriety of all aspects of its own proposed rates and terms, including all aspects of the participant’s proposed rate structure.’’). Accordingly, it is clear to the Judges that the Services should continue to bear the burden of proof regarding the sufficiency of their proffered Phonorecords II-based benchmark in this remand proceeding. And, in like fashion, because on remand Copyright Owners have assumed the mantle of pursuing the vacated rate structure and rates, they bear the burden of proof with regard to their proposal. However, Copyright Owners assert that it is the Services who bear the burden of proof as to Copyright Owners’ proposal regarding the appropriateness, vel non, of an uncapped TCC rate prong. According to Copyright Owners, this burden falls on the Services because ‘‘only the Services . . . proposed TCC prongs at the hearing,’’ in the form of the mix of capped and uncapped TCC prongs contained in the Services’ Phonorecords II benchmark. Id. at 47. The Judges find that the fact that the Phonorecords II-based benchmark advanced by the Services contains this mix of capped and uncapped TCC prongs does not bear on Copyright Owners’ duty, under 5 U.S.C. 556(d), to satisfy the burden of proof with regard to the rates and rate structure they are advancing on this remand. Moreover, 15 Copyright Owners and Services divided the time for oral argument. George Johnson dba GEO Music Group waived oral argument. 16 The findings and conclusions in this Initial Ruling are adopted by a majority of the Judges. One Judge dissents from the adoption of the entirety of the Phonorecords II rate structure (section II), though not from the exception to that benchmark with regard to the headline rate of 15.1% and the imposition of a cap on the TCC rate prong. One Judge dissents in part from the reasoning relating to adoption of the definition of Service Revenue (section V), but not from the adoption of that definition. 17 As addressed infra, the Judges also order that the participants in this remand proceeding prepare and submit regulatory provisions consistent with this ruling. See Footnote 174. 18 The Services include in their Joint Rate Proposal a chart summarizing the proposed rates for their offerings. That chart is attached as an Addendum to this Initial Ruling. In all other respects, the rates and rate structure of the Phonorecords II-based benchmark proposed by the Services (as that benchmark is defined herein) shall constitute the rates and rate structure for the Phonorecords III period.18 To be clear: the 15.1% headline percentage rate substitutes for the headline percentage rates in subparts B and C of the Services Phonorecords IIbased benchmark, and the definition of ‘‘Service Revenue’’ for bundles shall be the definition contained in 37 CFR 385.11 (paragraph (5) for the ‘‘Service Revenue’’ definition) as proposed in the Services’ Phonorecords II-based benchmark. (2) The Services’ Phonorecords IIbased benchmark is the better of the benchmarks proposed by the parties and satisfies the requirements of 17 U.S.C. 801(b)(1) in all respects. However, as noted supra, to be consistent with this statutory section and the decision in Johnson, the royalty rate of 10.5% in that benchmark shall be replaced with the 15.1% rate set forth in paragraph (1) above. (3) To reiterate for clarity, consistent with the adoption of the Phonorecords II-based benchmark, and for the reasons more fully developed herein, the Judges adopt the definition of ‘‘Service Revenue for Bundled Services’’ as it appeared in the Initial Determination in the underlying proceeding. Following are the Judges’ analysis and ruling after remand. I. Preliminary Issue: Burden of Proof ddrumheller on DSK120RN23PROD with RULES2 2018 VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 PO 00000 Frm 00004 Fmt 4701 Sfmt 4700 II. Rate Structure and Rates A. Relevant Rulings in Johnson In establishing a royalty rate structure and the rates within it in the context of this remand proceeding, the Judges are guided by the rulings in Johnson. 1. Percent of Revenue Prong The D.C. Circuit noted that the Judges found the royalties in the Phonorecords II period were too low and that record companies were receiving a disproportionate share of the sum of the mechanical and sound recording royalties. Johnson, 969 F.3d at 384–85. The D.C. Circuit acknowledged that ‘‘[t]he Judges . . . then carefully E:\FR\FM\10AUR2.SGM 10AUR2 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations analyzed the competing testimony and drew from it rates that were grounded in the record and supported by reasoned analysis.’’ Id. at 385. The D.C. Circuit found that the Judges acted well within their discretion and not arbitrarily, relying on substantial evidence in establishing the ‘‘zone of reasonableness’’ for the rates. Id. As the D.C. Circuit noted, the Judges’ process was ‘‘the type of line-drawing and reasoned weighing of the evidence [that] falls squarely within the [Judges’] wheelhouse as an expert administrative agency.’’ Id. at 385–86 (emphasis added). ddrumheller on DSK120RN23PROD with RULES2 2. Uncapped TCC Prong The D.C. Circuit found fault, however, in the Judges’ determination to establish an uncapped and increased percentagebased total content cost (TCC).19 Id. at 380. This approach ‘‘removed the only structural limitation on how high the [TCC] . . . can climb.’’ Id. The D.C. Circuit reasoned that uncapping the TCC alternative rate prong across all categories of service exposed the Services to potentially large hikes in the overall mechanical royalties they must pay. Id. at 382. The D.C. Circuit noted: ‘‘As the [Judges] acknowledge, sound recording rightsholders have considerable market power vis-a`-vis interactive streaming service providers . . . . The interactive streaming services are . . . exposed to the labels’ market power and record companies could, if they so chose, put those services out of business entirely . . . . [B]y virtue of their oligopoly power, the sound recording copyright holders have extracted ‘inflated’ royalties. . . .’’ Id. (cleaned up). While the Services had advocated uncapping the TCC alternative rate prong for some categories of service, that ‘‘does not mean they anticipated that the [Judges] would uncap the total content cost prong across the board. That is quite different.’’ Id. at 382. The D.C. Circuit found that the Judges ‘‘failed to provide adequate notice of the drastically modified rate structure [they] ultimately adopted.’’ Id. at 381. The D.C. Circuit emphasized that the failure to provide adequate notice of their intentions ‘‘is no mere formality [because] [i]nterested parties’ ability to provide evidence and argument . . . not 19 ‘‘TCC’’ refers to ‘‘Total Content Cost,’’ and is defined as ‘‘a percentage of the royalties paid by the service . . . to sound recording copyright holders.’’ Johnson, 969 F.3d at 370; see also Determination at 13 n.38 (‘‘TCC’’ is an industry acronym for ‘‘Total Content Cost’’, a shorthand reference to the extant regulatory language describing generally the amount paid by a service to a record company for the section 114 right to perform digitally a sound recording.’’). VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 only protects the parties’ interests, it also helps ensure that the [Judges’] ultimate decision is well-reasoned and grounded in substantial evidence.’’ Id. at 381–82. To support their adoption of an uncapped TCC rate prong, the Judges ‘‘predicted that the sound recording copyright owners’ royalty rates would naturally decline in the course of their negotiations with interactive streaming services.’’ Id. at 372. The Judges found persuasive the rebuttal testimony of one of Copyright Owners’ economic expert witnesses, Professor Watt, that an increase in mechanical royalties payable by the Services would lead to a corresponding decrease in the Services’ sound recording royalty obligations. See Determination at 73–74 (‘‘[S]ound recording royalty rates in the unregulated market will decline in response to an increase in the compulsory license rate for musical works [and] Professor Watt’s bargaining model predicts that the total of musical works and sound recordings royalties would stay ‘‘almost the same’’ in response to an increase in the statutory royalty.’’). The Services painstakingly criticized this ‘‘see-saw’’ theory. The D.C. Circuit concluded that, on remand, if and when the Judges consider the ‘‘uncapped’’ rate structure, they shall address all substantive challenges to that approach raised by the Services, including the issue of whether ‘‘an increase in mechanical license royalties would lead to a decrease in sound recording royalties.’’ Id. at 383. Thus, the D.C. Circuit held, the Judges erred procedurally in adopting an uncapped TCC alternative rate prong. The D. C. Circuit therefore instructed the Judges to provide the parties with the opportunity to fully address the issues regarding the uncapped TCC prong, and for the Judges to address the ‘‘substantive challenges’’ raised by the Services. 3. Four Itemized Statutory Objectives The statutory standard found in section 801(b)(1) instructs the Judges to set rates that are not only ‘‘reasonable,’’ but also reflective of four itemized objectives, or factors, which, as the D.C. Circuit stated, set forth ‘‘competing priorities.’’ 17 U.S.C. 801(b)(1)(A)–(D); Johnson, 969 F.3d at 387.20 With regard 20 These competing objectives are: (A) To maximize the availability of creative works to the public; (B) To afford the copyright owner a fair return for his or her creative work and the copyright user a fair income under existing economic conditions; (C) To reflect the relative roles of the copyright owner and the copyright user in the product made available to the public with respect PO 00000 Frm 00005 Fmt 4701 Sfmt 4700 54409 to these four priorities, the D.C. Circuit found that the Judges properly analyzed and applied the first objective (Factor A). Id. at 387–88. In particular, the D.C. Circuit did not disturb the Judges’ ruling that an increase in the royalty rates for mechanical licenses was necessary in order to satisfy Factor A. Johnson, 369 F.3d at 387–88. According to Johnson, in making this finding, the Judges had engaged in a ‘‘reasonable reading of the record’’ and had relied on ‘‘substantial evidence.’’ Id. at 388. Thus, Factor A (when considered without regard to the other three objectives) indicated that the statutory rate needed to be higher than it was during the Phonorecords II period.21 With regard to the other three objectives, Johnson stated that ‘‘[t]he question whether the [Judges] adequately addressed factors B through D . . . is intertwined with the nature of the rate structure ultimately imposed by the [Judges].’’ Id. at 389. Accordingly, the D.C. Circuit concluded that it ‘‘need not . . . address whether the [Judges] adequately considered these remaining factors.’’ Id.22 Within the parameters of the holdings in Johnson, the Judges consider the record facts and the arguments made in this remand proceeding, together with the pertinent facts and arguments made in the original proceeding. B. Rate Evidence for the 33-Months From January 2018 Through September 2020 After the Determination was issued, from its effective inception on January 1, 2018, through September 30, 2020— a 33-month period—the parties operated under the rates and rate structure set forth in that ruling. In light of the D.C. Circuit’s decision in Johnson, as of October 1, 2020, the parties reverted to the Phonorecords II rates. The Services have asserted in this remand proceeding that, during the 33-month period when the Majority’s new and higher to relative creative contribution, technological contribution, capital investment, cost, risk, and contribution to the opening of new markets for creative expression and media for their communication; and (D) To minimize any disruptive impact on the structure of the industries involved and on generally prevailing industry practices. Id. 21 However, as the D.C. Circuit also noted, because the four section 801(b)(1) objectives reflect ‘‘competing priorities, id’’ at 387, the holding that Factor A militates toward a higher rate is not ultumately dispositive. Rather, it must be weighed with the other statutory factors. 22 The phrase ‘‘intertwined with the nature of the rate structure’’ requires emphasis because the Majority independently considered how to weigh Factors B and C specifically as to the 15.1% revenue rate, without regard to the overall rate structure, as discussed infra. E:\FR\FM\10AUR2.SGM 10AUR2 54410 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations Phonorecords III rates were in effect, [REDACTED]. By contrast, Copyright Owners, on remand, looking at the same data over this 33-month period, aver that they prove the existence of the seesaw theory. 1. Services’ Position According to the Services, [REDACTED]. [REDACTED]. Moreover, according to the Services, [REDACTED]. The Services further maintain that, [REDACTED]. The Services make the [REDACTED]. And, [REDACTED]. Id. ¶¶ 5, 9–13, 16– 19, 22–23, 26–27. The Services claim that [REDACTED]. More particularly, [REDACTED]. [REDACTED]. [REDACTED]. The Services’ economic experts rushed to judgment upon learning of these facts, claiming that they disproved the seesaw theory. See Katz WDRT ¶¶ 25–27 (relying on testimonies cited supra and concluding that seesaw theory was disproved, based on [REDACTED]); Marx WDRT ¶¶ 48–51 (relying on same testimonies and likewise finding because [REDACTED]); Leonard WDRT ¶ 17. ([REDACTED]). ddrumheller on DSK120RN23PROD with RULES2 2. Copyright Owners’ Position Copyright Owners analyzed the royalty data over the same 33-month period (January 2018 through September 2020) and reach the opposite conclusion. One of their economic expert witnesses, Dr. Jeffrey Eisenach, testified that [REDACTED]. Moreover, he opined that [REDACTED]. See Eisenach RWRT sec. 2(A) & appx. C. Based on this analysis, Professor Watt declares empirical vindication of his seesaw theory. Watt RWRT ¶¶ 41–42, 46 (‘‘The [Judges’] bargaining theory insights about the relationship between royalty rates were correct. . . . [REDACTED]. . . .’’). 3. Analysis and Decision Regarding Evidence of Post-Determination Rates The Judges are perplexed by the willingness of the expert economic witnesses on both sides to opine that the rate changes from January 2018 through September 2020 can serve as confirmation of their clients’ respective positions. The issue to be considered empirically was whether the sound recording rate would decrease in response to the increase in the mechanical rate. That is, if the record labels had previously set royalties at a level that would allow the Services merely to survive, would the record labels agree to lower their sound recording rate if more of the Services’ surplus were acquired by Copyright Owners? To answer this question, the VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 economists on both sides applied sophisticated bargaining models and critiques to explain the nature of the negotiations that would ensue. In the process, the economists lost track of an obvious, elementary point: The Phonorecords III rates were being challenged by the Services’ appeal, and might not persist. Indeed, the rates were ultimately vacated and the parties returned in October 2020 to the Phonorecords II rates.23 Now, the rates will be changed again by this postremand Determination, and going forward may be subject to further potential change, consistent with the provisions of title 17. In light of such ongoing fundamental uncertainty, why would any economist or businessman assume that the sound recording companies would agree to adjust their rates in response to a change in the mechanical rate? The Judges are amazed that the economic experts neglected even to raise this uncertainty as a complicating issue, let alone a dispositive one.24 Moreover, no party called as a witness any representatives of the Majors, or subpoenaed their testimony or documents, to provide the Judges with evidence of how these record companies perceived the seesaw issue, whether as a permanent phenomenon or as an uncertain matter, given the pendency of the legal proceedings regarding the ultimate mechanical rate. Any of the parties could have requested that the Judges subpoena a sound recording industry witness to give testimony and produce documents as to this issue, pursuant to 17 U.S.C. 803(b)(6)(C)(ix), but none did so. Further, Copyright Owners, who are representing the music publishing interests of inter alios, Sony, Universal, Warner, and Merlin, likely could have produced such sound recording witnesses without the need for a subpoena. Witnesses from these entities who negotiated with the Services after the Phonorecords III rates and rate structure became effective certainly would have knowledge relevant to the testimony of the Services’ witnesses [REDACTED] who claimed that [REDACTED]. 23 There also was uncertainty as to the effective inception date of the Phonorecords III rate period, because the Services had appealed (ultimately unsuccessfully) the CRB Judges’ finding that the period commenced, retroactively, as of January 1, 2018. 24 To place this point in the economic context of this proceeding, the Judges characterize the ongoing ‘‘legal uncertainty’’ as another ‘‘independent variable’’ to add to the economic experts’ list of such variables, discussed infra, that affect the ‘‘dependent variable,’’ viz., the sound recording rate. PO 00000 Frm 00006 Fmt 4701 Sfmt 4700 Simply put, the period from period from January 2018 through September 2020 was a time the Judges construe as ‘‘33-months of uncertainty,’’ see 3/8/22 Tr. 87, 91 (Closing Argument) when no party could ascertain with any assuredness the ultimate Phonorecords III rates and rate structure. Thus, for the economists and the parties to claim vindication for their arguments by reliance on how the record labels did or did not respond to the challenged and ever-shifting rates during this ‘‘33 months of uncertainty’’ reflects the elevation of adversarial zeal over objective judgment. Accordingly, the Judges place no weight on the purported changes or stability of the sound recording rates during the Phonorecords III rate period. C. Percent-of-Revenue Rate Prong 1. Copyright Owners’ Position In their initial remand submission, Copyright Owners provided no new evidence to support any aspect of the 15.1% revenue-based rate (or for that matter, any new evidence to support the rates or rate structure in the Determination), and elected to rely on the pre-remand record. In fact, in their initial remand submission, Copyright Owners do not so much as mention the 15.1% revenue rate derived by the Judges. However, in their reply remand submission (which the Judges found also to constitute, in part, a substantive initial submission 25) Copyright Owners do address the 15.1% revenue rate. In the reply submission, Copyright Owners simply stated: ‘‘[T]he Circuit affirmed the Board’s derivation of rate percentages, including raising the revenue rate to 15.1%.’’ Copyright Owners’ Reply Brief on Remand (in Reply Remand Submission of Copyright Owners, Vol. 1) at 64, n.48 (July 2, 2021) (‘‘CO Reply’’). In a subsequent submission, Copyright Owners added that ‘‘[t]he narrow mandate on this Remand does not allow for reopening the rate percentage determination in the [ ]Determination.’’ Copyright Owners’ Motion for Reconsideration or Clarification at 15 & n.10 (Dec. 17, 2021) (emphasis added) (Dec. 17th Motion). Thereafter, Copyright Owners asserted that the D.C. Circuit’s affirmance of the 25 See Order Denying in Part and Granting in Part Services’ Motion to Strike Copyright Owners’ Expert Testimony and Granting Services’ Request to File Supplemental Testimony and Briefing at 11 (Oct. 1, 2021) (Oct. 1st Order) (The Judges found that ’’with one exception . . . the challenged testimonial evidence of Copyright Owners’ economic expert witnesses serve the dual purposes of direct and rebuttal statements’’ and, as a consequence, ‘‘provide[d] the Services an opportunity to file supplemental testimony and briefing in opposition. E:\FR\FM\10AUR2.SGM 10AUR2 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations [Judges’] revenue percentage rate calculation was ‘‘strong[ ]’’ and ‘‘detailed.’’ Copyright Owners’ Reply in Further Support of Motion for Reconsideration or Clarification at 4 (January 5, 2022). Moreover, Copyright Owners took note that the Services had relied on substantively identical language in Johnson to support their argument that other statements in that D.C. Circuit decision should be deemed affirmed. See id. at 4–5 (noting Services’ reliance on Johnson’s description of the Judges’ rulings regarding student and family discounts (‘‘grounded in substantial record evidence . . . based on the weight and credibility of the evidence [and] squarely within the Judges’ expertise)’’ as demonstrating that the D.C. Circuit had affirmed those rulings) (emphasis added); see also Copyright Owners’ Brief in Response to the Additional Materials Orders at 2, 6– 7 (Jan. 24, 2022) (‘‘CO Additional Submission’’) (again asserting that ‘‘the 15.1% revenue rate . . . was specifically affirmed in detail by Johnson.’’). ddrumheller on DSK120RN23PROD with RULES2 combined royalties being paid by the services than are currently paid in the marketplace. . . The discrepancy in total royalties between the models and the real world is explained, in part, by the absence of supranormal complementary oligopoly profits in the Shapley model, and the presence of those profits in the actual market.’’). Id. (quoting Phonorecords III, 84 FR 1952). By this approach, the Services maintain, ‘‘the Majority awarded the Copyright Owners the full 15.1% of revenue dictated by its model (phased in over time), and left it up to the Services to convince the complementary oligopolist major labels to dramatically lower sound recording rates.’’ Id. at 54– 55. The Services argue that, instead, the Majority should have applied to Professor Marx’s [REDACTED]% total royalty obligation what they characterize as ‘‘any of the[ ] real-world ratios in place of the [REDACTED] ratio taken from ‘‘Professor Gans’ ‘‘Shapleyinspired’’ model. Id. at 54. According to the Services, these lower ratios would have reduced the revenue percentage rate well below 15.1%. Id. 2. Services’ Position Alternatively, the Services propose, In their initial submission after the through Professor Marx’s post-remand remand, the Services objected to any written testimony, that the Judges now continued application by the Judges of adopt ‘‘a more balanced, burden-sharing the 15.1% revenue rate because, ‘‘as the approach’’ to address what she Majority acknowledged, this particular described as the Majority’s ‘‘imbalance’’ division of revenues will never happen problem. Id. at 57; see also Marx WDRT in the real world because of the ¶¶ 52–63.27 Essentially, her proposal complementary oligopoly power of the begins with an assumption, based on record labels.’’ Services’ Joint Opening record evidence, that labels typically Brief (in Services’ Joint Written Direct take specific shares of service revenue, Remand Submission at Tab D) at 52 including shares of [REDACTED]%, (‘‘Services’ Initial Submission’’) (Apr. 1, [REDACTED]% and [REDACTED]%.28 2021). More particularly in this regard, These shares are significantly higher the Services note that Professor Marx’s than the [REDACTED]% that Professor Shapley Value Model,26 which served as Marx generated from her Shapley an input for the generation of the 15.1% model. Next, Professor Marx’s postrevenue rate, also indicated that only remand burden-sharing approach uses [REDACTED]% of the interactive as inputs the 15.1% of service revenue streaming revenue should be paid out as and the [REDACTED]% of service royalties to the sound recording revenue that would be retained by the rightsholders, with the remaining musical works owners and the Services [REDACTED]% of these revenues respectively.29 Putting these two factors retained by the interactive streaming together, she sets forth the basic math: services. Id. (‘‘Both Professor Marx’s and Using her [REDACTED]% sound Professor Watt’s models show lower recording share as an example, she 26 Generally, a Shapley Value Model is a game theory analysis. It models a hypothetical bargain that assigns each ‘‘player’’ the average marginal value it contributes to the bargain and (after accounting for the costs that each ‘‘player’’ would need to recover) the remaining ‘‘surplus’’ is allocated among the players according to their relative contributions. See Johnson, 969 F.3d at 372. For the reasons discussed infra, in the present case, the Shapley surplus from the streaming revenue is split essentially equally by the owners of the sound recording and musical works owners inter se, but the royalty rates themselves that would result from their bargaining would be different as between these two inputs, because of their differing costs. See, e.g., Gans WDT ¶ 73. VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 27 Claiming consistency with the Majority’s analysis, Professor Marx appears to maintain that her ‘‘burden-sharing’’ approach generates the statutorily-required ‘‘reasonable’’ rate as well as a rate that satisfies the ‘‘fair return’’/‘‘fair income’’ objectives of statutory Factor B. See Marx WDRT ¶ 52 (introducing her correction of the alleged ‘‘imbalance’’ problem by noting that ‘‘the ‘‘right’’ mechanical royalty rate is one that is ‘‘reasonable’’ and achieves the four objectives laid out in Section 801(b)(1).’’ 28 See Marx WDRT, fig. 7 ([REDACTED]). 29 The [REDACTED]% of revenue that the services would retain is based on one of Professor Marx’s ‘‘Shapley Value Models.’’ Shapley Value modeling is discussed infra. PO 00000 Frm 00007 Fmt 4701 Sfmt 4700 54411 notes that there is not enough revenue for the labels to take this [REDACTED]% share, if the musical works owners also receive 15.1% and the Services also retain the [REDACTED]% derived from her model ([REDACTED]% + 15.1% + [REDACTED]% = [REDACTED]%, an irrational result). See Services’ Joint Opening Brief at 57. Professor Marx engages in an analysis based on the following math and logic (again, using the [REDACTED]% sound recording rate as an example of the fixed amount taken by the labels): (1) [REDACTED]% of the streaming revenues remain available to be split between the services and the musical works copyright owners; (2) adding the 15.1% revenue rate and her [REDACTED]% revenue retention percentage equals [REDACTED]%; and (3) the 15.1% revenue rate, as a percent of this [REDACTED]%, is [REDACTED]%; and (4) [REDACTED]% of the [REDACTED]% available for splitting between the services and the musical works copyright owners is [REDACTED]% (rounded). Id. at fig.8. Thus, she identifies her version of a ‘‘fair’’ result: The Services and Copyright Owners would split the residual revenue remaining after the labels have exercised their complementary oligopoly power to take an outsized fixed share—with the split proportional to the 15.1%-to[REDACTED]% revenue amounts calculated respectively by the Judges (the 15.1% musical works rate) and Professor Marx (the [REDACTED]% service revenue retention). Id. 59, table. 8.30 In their final post-remand submission, the Services also flatly state: ‘‘[T]he D.C. Circuit did not ‘‘affirm’’ the 15.1% rate—it vacated that rate.’’ Services’ Joint Rebuttal Brief Addressing the Judges’ Working Proposal at 2 (Feb. 24, 2022) (‘‘Services’ Additional Submission’’). However, the Services do not support that quoted statement with any citation to Johnson. See id. Further, the Services assert that the 15.1% revenue rate is not immune from postremand review and reduction because ‘‘the D.C. Circuit withheld judgment ‘‘on whether that final percentage satisfies factors B through D of Section 801(b)(1). . . .’’ Id. at 3. 30 Using the same logic and calculation method, Professor Marx finds that the services would retain |[REDACTED]% ÷ [REDACTED]%, which equals |[REDACTED]%. Assuming again that [REDACTED]% of the steaming revenue is available to split (because the labels have appropriated [REDACTED]%), the services would retain [REDACTED]% [REDACTED]% rounded) of the streaming revenue. Id. E:\FR\FM\10AUR2.SGM 10AUR2 ddrumheller on DSK120RN23PROD with RULES2 54412 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations 3. Analysis and Decision Regarding 15.1% Revenue Rate Prong The Judges determine that they are clearly bound by the D.C. Circuit’s decision in Johnson to maintain the 15.1% revenue rate, as phased-in by the Determination. Several reasons support this decision. First, the Judges conclude that the D.C. Circuit’s decision in Johnson is conclusive and unambiguous regarding the revenue percentage rate. The D.C. Circuit rejected the Services’ assertion that the Judges acted ‘‘arbitrarily’’ as to this particular issue, noting that the Services had misstated the relevant facts. Johnson, 969 F.3d at 385–86 (responding to Services’ misdescription of Judges’ analysis and explaining what Services described as ‘‘not what happened.’’). Moreover, the D.C. Circuit held that with regard to the construction of the 15.1% revenue rate, the Judges had ‘‘engaged in the type of linedrawing and reasoned weighing of the evidence [which] falls squarely within the [Judges’] wheelhouse as an expert administrative agency.’’ Id. at 386. The D.C. Circuit further noted that the Judges ‘‘proceed[ed] cautiously’’ to set the 15.1% revenue rate by establishing a ‘‘zone of reasonableness’’ for the revenue rate. Id. at 385. Indeed, with regard to each aspect of this revenue rate analysis, the D.C. Circuit found that the Judges’ decision making was ‘‘grounded in the record and supported by reasoned analysis’’ and that ‘‘[s]ubstantial evidence supports [their] judgment.’’ Id. at 385. Second, when the D.C. Circuit reviewed the Determination, it applied ‘‘the same standards set forth in the Administrative Procedure Act, 5 U.S.C. 706.’’ Id. at 375 (noting that 17 U.S.C. 803(d)(3) cross-references 5 U.S.C. 706); see also id. (‘‘[W]e will set aside the [ ] Determination ‘only if it is arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law, or if the facts relied upon by the agency have no basis in the record.’’). Here, the D.C. Circuit explicitly found that the Judges’ analysis and findings in connection with the 15.1% revenue rate are not arbitrary and capricious, and that the facts relied upon by the Judges have a sufficient basis in (are ‘‘grounded in’’) the record. It seems beyond dispute that the D.C. Circuit affirmed the Judges in their setting of the 15.1% revenue rate as a rate that is reasonable, and thus satisfies that aspect of the section 801(b)(1) standard.31 Indeed, it would 31 The CRB Judges intentionally distinguish between the ‘‘reasonable’’ rate standard in the initial body of section 801(b)(1) and the objectives set forth as Factors A–D of section 801(b)(1). A rate VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 border on the Orwellian to misconstrue the D.C. Circuit’s unequivocal and obvious affirmance of the reasonableness of the 15.1% revenue rate as a vacating of that finding. Third, the Judges note that Johnson conspicuously declines to identify the Judges’ setting of the 15.1% percent-ofrevenue rate as one of the findings to be revisited on remand. Rather, Johnson states that the three overarching issues for resolution on remanded are the Majority’s failure: (1) ‘‘to provide adequate notice of the rate structure it adopted,’’ (2) ‘‘to explain its rejection of a past settlement agreement as a benchmark for rates going forward; and (3) ‘‘[to] identif[y] the source of its asserted authority to substantively redefine a material term after publishing its Initial Determination.’’ Johnson, 369 F.3d at 367. The Majority’s finding that the 15.1% royalty rate is ‘‘reasonable’’ was not identified by the D.C. Circuit as a finding that was vacated and subject to further review and, indeed, as noted supra, the appellate panel credited what it characterized as the Majority’s careful analysis and line-drawing in arriving at that finding. The clarity of the D.C. Circuit’s affirmance of the royalty rate of 15.1% for the percent-of-revenue prong moots the issue of whether Professor Marx’s attempt, described supra, to correct the so-called ‘‘imbalance’’ problem has merit. However, the Judges note that, even if this issue had not been conclusively decided in Johnson, they would reject her approach as futile. That is, Professor Marx fails to acknowledge that any surplus that her approach would appear to provide to the Services would be siphoned off by the Majors, given their complementary oligopoly power. More particularly, the sound recording royalty rates she posits ([REDACTED]%, [REDACTED]% and [REDACTED]%) are all functions of the sound recording companies’ understanding of the Services’ noncontent costs (costs that the Services must recover out of retained revenues in order to remain in operation, i.e., to ‘‘survive’’) and the then-existing musical works content (royalty) costs (comprised of the mechanical rate and the performance rate). If, as Professor Marx contemplates, the mechanical rate is reduced so that Copyright Owners ‘‘share the burden’’ of the complementary oligopoly effect on can satisfy the statutory ‘‘reasonable rate’’ requirement yet require adjustment (higher or lower) to reflect the balancing of the four additional factors. Accordingly, the Judges defer to a subsequent section, infra, a discussion of how Factors A–D should be addressed on this remand. PO 00000 Frm 00008 Fmt 4701 Sfmt 4700 sound recording rates, that ‘‘burden sharing’’ would increase the revenues retained by the Services (that is the purpose of Professor Marx’s approach!). But such an increase would raise the Services’ revenue above their ‘‘survival’’ rate, as understood by the record labels. Thus, the record labels, given their complementary oligopoly power, would increase the Services’ royalty rate above what it otherwise would have been. Alternately stated, when Professor Marx hypothesizes a given sound recording royalty rate in column 1 of Figure 8 in her WDRT, that rate is assumed, by the logic of the complementary oligopoly theory, to have already allowed the services to cover only their non-content costs and musical works royalties, as understood by the record labels. So, her assumed rate in column 1 is not a fixed parameter, but rather an independent variable, which is a function of, inter alia, the costs incurred by the services, i.e., their non-content costs plus their musical works royalty costs.32 If those service costs decreased (for example, in an attempt to reduce the services’ burden of bearing the full brunt of the labels’ complementary oligopoly power as in Professor Marx’s attempt to correct the imbalance problem), the percentage in column 1 of Figure 8 would increase, as the labels siphoned off that surplus over the services’ survival revenue requirements. To find otherwise would be to refute the logic of the dynamics of the complementary oligopoly effect.33 Moreover, the defect in Professor Marx’s attempt to remedy the so-called ‘‘imbalance’’ problem is a consequence of the statutory licensing and royalty scheme. To recap, the licensing of content used by the interactive services is bifurcated. The sound recording royalties paid by the interactive services to the record labels are not regulated, and complementary oligopoly power exists in that market, inflating sound recording royalty rates above an effectively competitive level. See 32 The interactive services also pay a separate royalty for the performance license necessary to transmit a song. However, under the Judges’ ‘‘AllIn’’ royalty structure, that performance royalty is deducted from the ‘‘All-In’’ calculation to determine the mechanical royalty. Also, the performance royalty paid to the largest Performing Rights Organization (PROs) are subject to determination by federal judges in the Southern District of New York (the so-called ‘‘rate court’’). 33 To be clear, the Judges are not stating that the Services’ retention of only enough revenue to allow them to cover their noncontent costs and thus merely ‘‘survive’’ is indicia of an effectively competitive (or even healthy) market—but are merely acknowledging the state of affairs given the unregulated nature of the sound recording royalties and the complementary oligopoly power that exists in that market. E:\FR\FM\10AUR2.SGM 10AUR2 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations ddrumheller on DSK120RN23PROD with RULES2 Determination at 73 (‘‘[T]he existence of complementary oligopoly conditions in the market for sound recordings’’ is the basis for ‘‘the record companies’ ability to obtain most of the available surplus’’ generated by interactive streaming.) 34 However (and to state the obvious), the mechanical rate paid by the interactive services for musical works is regulated, pursuant to 17 U.S.C. 115 and, until the 2018 enactment of the Music Modernization Act,35 according to the rate standards in 17 U.S.C. 801(b)(1). Thus, there is no statutory or regulatory impediment to prevent record labels from responding to a decrease in the mechanical rate by increasing the unregulated sound recording rate if such an increase is in their economic interest.36 Accordingly, any attempt by the Judges to reduce the mechanical royalty rate in order to allow the Services to retain more of the surplus would fail; it would be like pouring water into a bucket with a siphon at its base. More water would not remain in the bucket, but rather would accumulate wherever the siphon leads—in this case, to the record labels. The Judges could keep mechanical royalty rates depressed and allow this to occur, but that would harm Copyright Owners while providing no relief to the Services. And despite the old adage that ‘‘misery loves company,’’ the Judges detect no directive under section 801(b)(1) that they harm Copyright Owners without providing a gain for the interactive streaming services—and that they provide a windfall for the record labels, to boot. Although Professor Marx’s attempt to reduce the Services’ ‘‘misery’’ by sharing it with Copyright Owners is unavailing, the statutory scheme and market forces do appear to combine to mitigate the burden created by the complementary oligopoly power of the sound recording companies. If 34 As the Judges have consistently noted, this complementary oligopoly power is generated by the concentration of ownership of sound recording licenses for ‘‘Must Have’’ repertoires among the three Majors (Sony Music Group, Warner Music Group and Universal Music Group), plus Merlin (a consortium of Indies sometimes referred to as ‘‘the fourth Major’’), as indicated by their reported collective 85% share of Spotify’s streams in 2018, the first year of the rate period at issue here. See https://www.midiaresearch.com/blog/smallerindependents-and-artists-direct-grew-fastest-in2020. 35 In subsequent rate periods, the rate remains regulated, but is subject to a different standard—the ‘‘willing buyer-willing seller marketplace standard,’’ for shorthand) under 17 U.S.C. 115. 36 The inverse relationship between changes in the mechanical royalty rate and changes in the sound recording royalty rate has been characterized as the ‘‘seesaw’’ effect, which is discussed in further detail infra, with regard to the uncapped TCC rate prong. VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 interactive streaming revenue were to grow over the rate period,37 then the phase-in to the 15.1% rate will reflect fixed annual percentages of a larger base, allowing services to retain a higher dollar level of the interactive streaming revenues.38 [REDACTED]. See, e.g., Diab WDRT ¶¶ 10–11 (Google agreements); Mirchandani WDRT ¶¶ 16–17 (Amazon agreements); Bonavia WDRT ¶¶ 8; 14– 19 (Spotify agreements); White WDRT ¶¶ 6; 8–14; 19; 24; 27–28 (Pandora agreements). Additionally, the Services’ headline sound recording rates [REDACTED]. Services’ Joint Remand Reply Brief at 40 (and record citations therein). Thus, assuming no increase in non-content costs (or increases smaller than the increases in streaming revenue), the Services will realize increased revenue above and beyond what they needed to survive. The Services and Copyright Owners recognize the mitigation of harm to the Services generated by these facts (although they may well disagree with the Judges’ application of these facts). During colloquy with counsel for Pandora and Spotify during closing arguments on remand, the Judges asked why they should in essence apply the ‘‘misery loves company’’ adage: [JUDGE STRICKLER] [T]he problem is . . . the sound recording [rates] are unregulated in the interactive market . . . . Congress did not want that to be controlled at all. So every time I see . . . the services’ argument about how we have [to] set a rate that’s fair even though there’s this ability of the sound recording [companies] to take more, my margin note is always this: ‘‘Are they arguing that ‘misery loves company?’ ’’ [W]hy shouldn’t that misery be shared with Copyright Owners? . . . Isn’t that really Professor Marx’s argument in her proposed split . . . using the 15.1 percent figure . . . ? [COUNSEL] [Regarding] Judge Strickler[’s] . . . ‘‘misery loves company’’ issue. . . . I 37 Because this proceeding was appealed and remanded, the Judges have the benefit of knowing the ‘‘future’’ (beyond 2017), during which U.S. interactive streaming revenues have continued to grow, a fact that is undisputed, and as to which the Judges take administrative notice. See, e.g., RIAA 2018 Year-End Music Industry Revenue Report (available at https://www.riaa.com/wp-content/ uploads/2019/02/RIAA-2018-Year-End-MusicIndustry-Revenue-Report.pdf; RIAA 2020 Year-End Music Industry Revenue Report (available at https://www.riaa.com/wp-content/uploads/2021/ 02/2020-Year-End-Music-Industry-RevenueReport.pdf (interactive streaming revenue increased within this rate period from (approximately) $1.6 billion in 2018 to $7.7 billion in 2019 and $8.8 billion in 2020). 38 For example, if a royalty is set at a flat rate of 15.1% when a revenue base is $1,000, then the royalty is $151, leaving $849 in revenues to cover other costs which, for this example, are held constant. If the revenue base doubles to $2,000, the same flat 15.1% royalty rate generates $302 in royalties, leaving $1,698 in revenues to cover other costs which, if constant, allow for the additional revenue ($1,698¥$849 = $849) to generate profits. PO 00000 Frm 00009 Fmt 4701 Sfmt 4700 54413 think . . . the way [Judge Strickler] put it during the trial was, even if I thought rates needed to come down, how would that help you; wouldn’t the labels just take all that surplus for themselves based on their complementary oligopoly power? . . . . I want[ ] to address it right off the bat . . . . in open session. Relat[ed] to . . . the seesaw . . . our point is that these label rates are sticky in both directions. If you see an increase in musical works rates, you do not see a quick decrease in label rates, and the opposite is true. These rates are sticky. . . . There’s a lot of friction with respect to the ability of label rates to change quickly in response to the dynamic marketplace or the dynamic for business reasons or because of regulatory changes in musical works rate. These are multi-year contracts. They take a long time to negotiate. They are complex, et cetera. So, I do think it’s right that at a minimum you can buy time where the ratio is more aligned with the 801(b) factors. In other words, you don’t have to worry that the labels will take it all right away, even if you believe they will ultimately take that. [JUDGE STRICKLER] So you are saying we have something that reduces misery for a period of time until the misery returns? [COUNSEL] That’s right. And I think that would have been true in 2018 when you were sitting drafting the decision. It’s even more true today in 2022 when the label rates, as I mentioned, are effectively set, bought and paid for. 3/8/22 Tr. 29–30, 43–46 (Closing Argument) (emphasis added). Similarly, on this topic, Copyright Owners’ counsel accurately characterized the Judges’ adoption of the static 15.1% Shapley-based rate as the inevitable consequence of ‘‘regulatory lag,’’ that requires a regulator to keep a rate constant over the statutory term because there is no sufficient data to project future rates. Id. at 273–75; see generally A. Kahn, 2 The Economics of Regulation at 48 (1971) ‘‘The regulatory lag [is] the inevitable delay that regulation imposes in the downward . . . [and] upward adjustments’’ to rate levels, and ‘‘thus is to be regarded not as a deplorable imperfection of regulation but as a positive advantage [because] companies can for a time keep the higher profits they reap from a superior performance. . . .’’).39 39 The Judges emphasize two points that mitigate any negative impact on Copyright Owners from the static nature of the 15.1% revenue rate. First, as a percent-of-revenue rate, it generates more royalty revenue in a growing market, so the quantum of revenue is not static. Second, Copyright Owners’ own economic expert witness, Professor Gans, testified that the data in the ‘‘market observations’’ from the Goldman Sachs Report on which he relied were the result of ‘‘negotiated rates in the free market and thus ‘‘presumed to . . . fully consider[ ] E:\FR\FM\10AUR2.SGM Continued 10AUR2 54414 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations ddrumheller on DSK120RN23PROD with RULES2 4. Consideration of Factors A–D in Section 801(b)(1) Finally, the Judges consider the impact of Factors A–D of section 801(b)(1) in connection with the setting of the revenue percentage rate of 15.1%.40 Regarding Factor A, it cannot be gainsaid that the D.C. Circuit has left this issue unresolved. Rather, Johnson unambiguously affirmed the Majority’s finding that an increase in the mechanical royalty rate was warranted. Specifically, Johnson states that the Majority’s decision in this regard met the ‘‘test’’ that it be ‘‘supported by substantial evidence [and] reflect a reasonable reading of the record.’’ Johnson, supra, at 388. Moreover, with regard to the level of the increase, the D.C. Circuit did not disturb the finding by the Majority that ‘‘[t]he rates determined by the Judges represent a 44% increase over the current headline rate, and thus satisfies the Factor A objective. . . .’’ Determination at 85.41 With regard to Factors B and C,42 even if Johnson were construed as permitting . . . expectations of future costs and revenues . . . . incorporate[ing] expectations of future values.’’ Gans WRT ¶¶ 37–38. On this issue, it is noteworthy that both the Majority and the D.C. Circuit credited Professor Gans’s reliance on these projections. See Determination at 70 (‘‘The Judges . . . find Professor Gans’ reliance on financial analysts’ projections for the respective industries to be reasonable.’’); Johnson, 969 F.3d at 386 (holding that ‘‘[t]he CRB Judges’ finding that Gans’s . . . reliance on Goldman Sachs’ profit projections’’ was ‘‘reasonable’’ and the] . . . type of line-drawing and reasoned weighing of the evidence [that] falls squarely within the [Copyright Royalty Board’s] wheelhouse as an expert administrative agency.’’) Thus, dynamic changes going forward in the rate term are embodied in the 15.1% revenue rate, and dynamic market expectations are incorporated in the modeling data used to establish that rate. 40 The D.C. Circuit ruled, with regard to the ‘‘nature of the rate structure,’’ that because it had ‘‘vacat[ed] and remand[ed] . . . for lack of notice’’ ‘‘[t]he question whether the [Judges] adequately addressed factors B through D is bound up with the [Judges’] analysis of sound recording rightsholders’ likely responses to the new rate structure.’’ Johnson, supra, at 389. However, the 15.1% revenue rate, viewed separately, is not bound up in the ‘‘rate structure’’ issue, which relates to the uncapped TCC prong and how the 15.1% revenue rate may be ‘‘intertwined’’ with that second rate prong. As explained infra, the Judges are not adopting an uncapped TCC rate prong, so the 15.1% rate is no longer ‘‘bound up’’ with the vacated and remanded ‘‘rate structure’’ issue, making moot the argument that a new post-remand analysis of Factors B through D is necessary or appropriate. However, on remand, Copyright Owners have placed in issue the ‘‘disruption’’ element of Factor D, claiming that the Services have not proven that the uncapped TCC rates and rate prong have or will cause disruption. 41 The 44% figure cited by the Majority reflects the percentage increase of the headline rate, from 10.5% to 15.1%. 42 Factors B and C are typically considered jointly, because of the overlap in the objectives of providing a ‘‘fair return’’ and a ‘‘fair income’’ to the licensors and licensees respectively (the Factor B objectives) and reflecting their relative roles in making the streamed music available to the public VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 the Judges to revisit this issue, they would not adjust the 15.1% revenue rate on the basis of these two factors. In this regard, the Judges note that the Majority found that the 15.1% revenue rate was not only ‘‘reasonable,’’ but also a ‘‘fair allocation of revenue between copyright owners and services.’’ Determination at 87 (emphasis added). The Majority thus found explicitly that ‘‘with regard to Factors B and C . . . there is no basis to depart from [its] determination of the reasonable . . . rate structure and rates as set forth supra.’’ Id. More particularly, the Majority calculated the 15.1% rate by utilizing the total royalty percentage revenue of only [REDACTED]% as calculated by Spotify’s economic expert witness, Professor Marx, whose economic modeling intentionally reflected a conception of fairness by reducing the effect of the labels’ complementary oligopoly market power. See Determination at 67–68 (noting that Professor Marx testified that this aspect of her model ‘‘represents a fair outcome in the absence of market power [and] . . . eliminates . . . market power’’ which . . . if left in the economic analysis would ‘‘render[ ] . . . the analysis incompatible with the objectives of Factors B and C of section 801(b)(1).)’’) (emphasis added).43 Accordingly, the Judges find it would be substantively unwarranted to engage in any new consideration on remand of the impact, if any, of Factors B and C on the otherwise reasonable 15.1% revenue rate.44 (the Factor C objectives). See Johnson, 969 at 388 (noting without criticism the joint consideration of Factors B and C; Determination at 85–86 (noting without criticism the several experts’ joint consideration of Factors B and C). 43 Additonal facts support the Majority’s finding that the 15.1% revenue rate is fair. The record evidence indicates that the headline percent-ofrevenue sound recording rate was between approximately [REDACTED]% to [REDACTED]% in 2017. See Marx WDRT ¶ 58, fig 7. When the 15.1% mechanical rate is added to that rate range, the range of the total royalty obligation (based on headline rates) is [REDACTED]% to [REDACTED]%. (Plus, given the phase-in of the rates expressly to avoid disruption, the total royalty obligation would be even lower before 2022, at current sound recording rates.) The evdence preremand indicated that the Services were ‘‘surviving’’ while incurring noncontent of costs of approximately [REDACTED]% of revenue, leaving about [REDACTED]% of revenue available to pay royalties while still remaining in business. See Eisenach WRT ¶ 79 (Copyright Owners’ expert economic witness); McCarthy WDT ¶¶ 28–29 (Spotify’s Chief Financial Officer.) Thus, even if the Judges were to engage in a de novo analysis of the potential applicability of Factors B and C to the 15.1% rate, they would not find any basis sufficient to warrant a downward rate adjustment, beyond the phase-in adopted in the Determination. 44 However, the Judges take note of their further observation, discussed supra, that the combined impact of ‘‘sticky’’ sound recording royalty rates PO 00000 Frm 00010 Fmt 4701 Sfmt 4700 The final itemized statutory factor— Factor (D)—instructs the Judges to consider the ‘‘competing priority’’ of ‘‘minimiz[ing] any disruptive impact on the structure of the industries involved and on generally prevailing industry practices.’’ 17 U.S.C. 801(b)(1)(D). As with Factors B and C, even if Johnson were construed to allow the Judges to revisit this issue on remand with respect to the 15.1% revenue rate, the Judges would not change the Majority analysis or findings. In the Determination, the Judges adopted the following interpretation of this standard set forth in previous determinations: [T]he Judges reiterated their understanding of Factor D, concluding that a rate would need adjustment under Factor D if that rate directly produces an adverse impact that is substantial, immediate and irreversible in the short-run because there is insufficient time for either [party] to adequately adapt to the changed circumstance produced by the rate change and, as a consequence, such adverse impacts threaten the viability of the music delivery service currently offered to consumers under this license. Determination at 86 (emphasis added). Also, in order to minimize any economic disturbance to the Services’ businesses, the Majority decided to phase-in the 15.1% rate over the fiveyear rate term, setting annual percent-ofrevenue rates as follows: 11.4% in 2018; 12.3% in 2019; 13.3% in 2020; and 14.2% in 2021, before the full 15.1% rate became effective in 2022 the final year of the rate term. Id. at 87–88. On remand, the Services have not made any argument that the rate structure or rates set by the Majority were ‘‘disruptive under this standard.’’ 45 In sum, there is insufficient basis for the Judges to change the Majority’s application of Factor (D) to the 15.1% revenue rate finding by the Majority.46 and the inevitable regulatory lag provide an additional modicum of fairness with regard to the mechanical royalty rate. 45 The Judges further discuss the Factor D ‘‘disruption issue infra in connection with their analysis of the uncapped TCC prong. 46 Additional facts further support the Majority’s finding that the 15.1% revenue rate is would not be disruptive under Factor D. The record evidence indicates that the headline percent-of-revenue sound recording rate was between approximately [REDACTED]% to [REDACTED]% in 2017. See Marx WDRT ¶¶ 14, 19. When the 15.1% mechanical rate is added to that rate range, the range of the total royalty obligation (based on headline rates) is [REDACTED]% to [REDACTED]%. (Plus, given the phase-in of the rates expressly to avoid disruption, the total royalty obligation would be even lower before 2022, at current sound recording rates.) The evidence pre-remand indicated that the Services were ‘‘surviving’’ while incurring noncontent costs of approximately [REDACTED]% of revenue, leaving about [REDACTED]% of revenue available to pay royalties while still remaining in business. See Eisenach WRT ¶ 79 (Copyright Owners’ expert E:\FR\FM\10AUR2.SGM 10AUR2 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations 5. Conclusion Regarding the 15.1% Revenue Rate For the forging reasons, the Judges do not disturb the Majority’s finding that the percent-of-revenue rate at 15.1%, phased-in annually over the rate period, constitutes a ‘‘reasonable’’ rate under section 801(b)(1) to be used as the statutory rate for the 2018 to 2022 period.47 D. Uncapped TCC Rate Prong 1. Two Post-Remand Rationales for Uncapped TCC Rate Prong The Determination set forth the following two primary reasons for adopting a ‘‘greater-of’ rate structure that also included an uncapped TCC rate prong: ddrumheller on DSK120RN23PROD with RULES2 First, the use of an uncapped TCC metric is the most direct means of implementing a key finding . . . by the experts for participants on both sides in this proceeding: the ratio of sound recording royalties to musical works royalties should be lower than it is under the current rate structure. Incorporating an uncapped TCC metric into economic witness); McCarthy WDT ¶¶ 28–29 (Spotify’s Chief Financial Officer). Thus, even if the Judges were to engage on remand in a de novo analysis of the potential applicability of Factor D to the 15.1% rate, they would not find any disruption sufficient to warrant a downward rate adjustment, beyond the phase-in adopted in the Determination. 47 The Services’ assert that the Judges previously found that the reasonableness of the 15.1% rate was subject to revision on remand. In support of this position, the Services cite to the Judges’ Order Granting in Part and Denying in Part Copyright Owners’ Motion for Reconsideration or, in the Alternative, Clarification at 3, 4 n.7 (January 6, 2022) (Jan. 6th Order). But the Judges said in that interlocutory proposal merely that Copyright Owners were incorrect in their extreme assertion that the Judges could not make an ‘‘alternative rate and rate structure finding . . . except for the readoption of the vacated rate and rate structure approach in the Phonorecords III Determination [because] . . . [t]hat . . . would . . . be inconsistent with Johnson [and] . . . would render the D.C. Circuit’s vacating and remanding of the proceeding without force or effect.’’ Id. at 4, n.7. That did not mean that certain elements of the D.C. Circuit’s ruling could be ignored. Further, when the Judges provided the parties with the Judges’ explicitly tentative ‘‘Working Proposal,’’ they did not declare that the 15.1% revenue rate calculation could be revisited. Rather, the Judges ‘‘express[ed] a concern, not that the foregoing calculations could be overridden, but rather that this analysis . . . is ‘incomplete’ . . .’’ Jan. 6th Order at 6 (emphasis added). The parties’ submissions in response to the Judges’ ‘‘Working Proposal’’ demonstrated that the 15.1% revenue rate calculation was not ‘‘incomplete’’ in the manner that had raised the Judges’ concern. Nothing the Judges said in this interlocutory and tentative ‘‘Working Proposal’’ constituted a definitive statement regarding the Judges’ view of what was and was not subject to review on remand. See generally merriamwebster.com (defining the adjective ‘‘working’’ in this context as ‘‘assumed or adopted to permit or facilitate further work or activity . . . a working draft.’’). Indeed, a primary purpose of the ‘‘Working Proposal’’ was to allow the Judges and the parties to address potential issues and resolutions, without prejudice going forward. VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 the rate structure permits the Judges to influence that ratio directly. Second, an uncapped TCC rate prong effectively imports into the rate structure the protections that record companies have negotiated with services to avoid the diminution of revenue. Determination at 35–36.48 2. Copyright Owners’ Position Copyright Owners claim that the uncapped TCC prong should be adopted. They contend that the D.C. Circuit remand was merely ‘‘procedural’’ rather than substantive, and the Judges thus are not precluded from readopting the uncapped TCC prong in this remand proceeding. CO Initial Submission at 35–38 (and record citations therein). They further contend that the uncapped TCC prong was adopted to provide protection against revenue deferment and displacement occasioned by the Services choosing to elevate the growth of subscribers and other listeners over revenue maximization. Id. at 38–43 (and record citations therein). The uncapped TCC prong was first proposed by Google to persuade the Judges to reject Copyright Owners’ proposed ‘‘greater-of’’ rate structure containing a per-play prong and a per subscriber prong. Id. at 43–46 (and record citations therein). Copyright Owners argue that the uncapped TCC prong should be adopted because: (1) the Services have not shown any actual or threatened ‘‘disruption’’ or other harm resulting from the uncapped TCC prong during the 33-month period; (2) the Services actually experienced ‘‘unprecedented growth and profit’’ during this period; and (3) the Services paid lower percentages of revenues in mechanical and total royalties when the uncapped TCC prong was in effect. Copyright 48 The Majority added two other reasons that are not germane to this remand. In particular, the Majority stated that, compared to the Phonorecords II benchmark proposed by the Services, the ‘‘greater-of’’ structure with the uncapped TCC rate prong was ‘‘simpler’’ to understand than the ‘‘Rube Goldberg-esque’’ nature of the Phonorecords II rate structure. Id. at 36. This issue apparently was not raised on appeal, as it was not mentioned in Johnson, and Copyright Owners have not raised the issue on remand. See CO Initial Submission, supra. (However, the Judges do consider this issue in their analysis of the PR II-based benchmark, infra.) The final reason provided by the Majority was that its adoption of an uncapped TCC rate prong was supported by evidence of Google’s agreements with labels that included an uncapped rate structure, on which Google had relied to propose, post-hearing, the same greater-of rate structure. Id. However, the D.C. Circuit found that Google’s proposal was distinguishable, as it was based on a far lower TCC rate (15%) as well as a far lower percent-of-revenue rate (10.5%). The D.C. Circuit thus declined to rely on the Google-based approach as support for the uncapped TCC rate prong. Johnson, 969 F.3d at 383. PO 00000 Frm 00011 Fmt 4701 Sfmt 4700 54415 Owners’ Reply Brief on Remand at 34– 48 (and record citations therein). Relatedly, according to Copyright Owners the Services’ argument that the ‘‘see-saw’’ effect is unsupported by empirical evidence has collapsed, given the evidence relating to market performance. Further Copyright Owners maintain that this argument is irrelevant to the rate structure issue. Id. at 48–50 (and record citations therein). 3. Services’ Position The Services argue on remand that the uncapped TCC rate prong must be rejected. The Services reject the ‘‘seesaw’’ theory claiming it is disproved by the experience of the parties during the 33-month period. Services’ Joint Opening Brief at 48–49; Services’ Joint Supplemental Brief at 7– 13 (Nov. 15, 2021) (and record citations therein). The Services further contend that Copyright Owners have disavowed the ‘‘seesaw’’ theory as understood by the Majority. The Services allege that Copyright Owners now claim that the theory was nothing more than ‘‘a nod’’ to certain ‘‘core principles’’ of bargaining theory, rather than a specific prediction of a commensurate inverse relationship between increases in the mechanical royalty rate and decreases in the sound recording royalty rate. Services’ Joint Supplemental Brief at 2, 5–7 (and record citations therein). With regard to the uncapped TCC rate prong, the Services assert that Copyright Owners have not even attempted to demonstrate—nor could they demonstrate—that the uncapped TCC rate prong is consistent with all four statutory objectives set forth in section 801(b)(1). Services’ Joint Reply Brief at 1, 3–4, 33–34 36 (July 2, 2021) (‘‘Services’ Reply’’); see also Services’ Joint Opening Brief at 44–64 (and record citations therein). The Services claim that ‘‘yoking’’ the mechanical rate to the ‘‘complementary oligopoly rates extracted by the labels is plainly unreasonable.’’ Services’ Joint Opening Brief at 44–46. The Services argue that the existence, vel non, of any ‘‘disruptive impact’’ arising from the uncapped TCC rate prong, is misguided and not dispositive, because it is only one of the four separately itemized factors and, as this factor relates to Copyright Owners’ proposed uncapped TCC prong, they bear the burden of proof. Services’ Reply at 35–37. Finally, the Services contend that Copyright Owners have failed to explain their self-contradictory pre-remand argument that ‘‘an uncapped TCC prong ‘does nothing to protect Copyright Owners from the Services’ revenue E:\FR\FM\10AUR2.SGM 10AUR2 54416 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations displacement and deferment.’ ’’ Services’ Reply at 43. 4. Application of Johnson Findings Regarding Uncapped TCC Rate Prong The Judges conclude that the D.C. Circuit affirmed the Majority’s derivation and calculation of the 26.1% TCC rate, but vacated and remanded the Judges’ application and inclusion of that rate prong in the rate structure. The D.C. Circuit noted that, on appeal, the Services contended that ‘‘it was arbitrary and capricious for the [Judges] to rely on information drawn from different expert analyses in calculating the mechanical royalty rates.’’ Johnson, 969 F.3d at 384. Thus, the Services were making the same ‘‘information’’-based argument in opposition to the calculation of both aspects of the mechanical royalty rates—the revenue percentage prong and the TCC prong. See also id. (‘‘the Streaming Services separately leveled objections to the particular percentages adopted by the Copyright Royalty Board to calculate the revenue and total content cost prongs.’’) (emphasis added) In fact, both rate prongs were indeed derived from the same analyses. See Determination at 75 (table) (showing that both 15.1% revenue rate and 26.2% TCC rate derived from same data— Professor Marx’s model showing total royalties as high as [REDACTED]% [Majority’s lower bound] and Professor Gans’s ‘‘Shapley-inspired’’ model showing TCC percent should be [REDACTED]%.) 49 It is also clear from Johnson that the D.C. Circuit found that the Majority had reasonably derived and calculated the 26.2% TCC rate: ddrumheller on DSK120RN23PROD with RULES2 When it came to . . . the ratio of sound recording to musical work royalties that Gans derived from his analysis the [CRB Judges] specifically found . . . reasonable Gans’ equal value assumption [for dividing the Shapley surplus . . . between sound recording and musical works owners] and his reliance on Goldman Sachs’ profit projections. That type of line-drawing and reasoned weighing of the evidence falls squarely within the Board’s wheelhouse as an expert administrative agency. See Johnson, 969 F.3d at 385–86 (cleaned up) (emphasis added). Accordingly, because the identical analysis was performed by the Judges to derive the 26.2% TCC rate as was done to derive the 15.1% revenue rate, the Majority’s finding with regard to the derivation and calculation of the TCC rate likewise is not subject to further consideration on remand by the Judges. 49 The reciprocal of Professor Gans’s [REDACTED]ratio of sound recording:musical works royalties is [REDACTED], or [REDACTED]%. VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 However, it is equally clear that the D.C. Circuit vacated and remanded the Majority’s application and inclusion of the 26.2% TCC rate in a separate ‘‘greater-of’’ TCC prong. The defect that generated the vacating on this issue was procedural— ‘‘the Streaming Services had no notice that they needed to defend against and create a record addressing such a significant, and significantly adverse, overhaul of the mechanical license royalty scheme . . .’’ Id. at 382. The consequence of the D.C. Circuit’s action, however, was substantive. The D.C. Circuit stated: This is no mere formality. Interested parties’ ability to provide evidence and argument bearing on the essential components and contours of the [Judges’] ultimate decision not only protects the parties’ interests, it also helps ensure that the [Judges’] ultimate decision is well-reasoned and grounded in substantial evidence. . . . The Streaming Services separately challenge the uncapped rate structure as arbitrary and capricious. In particular, they argue that the rate structure formulated by the [Judges] failed to account for the sound recordings rightsholders’ market power. They also object that the [Judges] failed to provide a ‘satisfactory explanation, or root in substantial evidence, [their] conclusion that an increase in mechanical license royalties would lead to a decrease in sound recording royalties [the ‘‘inverse relationship’’ a/k/a the ‘‘seesaw’’ effect]. Id. at 381–83 (cleaned up) (emphasis added). Thus, the D.C. Circuit explicitly declined to address these substantive issues, because of the deficient procedure. Instead, the D.C. Circuit remanded these substantive issues back to the Judges. Id. Simply put, Johnson found that the absence of notice here could be outcome-determinative. Thus, the Judges categorically reject Copyright Owners’ assertion that the remand as to the uncapped TCC rate structure was merely ‘‘procedural.’’ The Judges do not accept the notion that the Majority simply committed some ministerial faux pas that could be summarily corrected so that the uncapped TCC rate structure could be rubber-stamped on remand. Rather, the Judges’ error rendered it impossible for them to consider the pros and cons of such a rate structure without the necessary input from the Services (and, for that matter, Copyright Owners as well). Because the procedural infirmity precluded the D.C. Circuit from deciding whether the Majority’s decision was ‘‘well-reasoned and grounded in substantial evidence,’’ there also can be no substantive presumption of the appropriateness of the uncapped TCC rate prong, as suggested by Copyright Owners. To the contrary, the D.C. Circuit’s opinion PO 00000 Frm 00012 Fmt 4701 Sfmt 4700 makes it clear that on remand the Judges must engage in a fresh consideration of the statutory appropriateness, vel non, of the uncapped TCC rate prong, by weighing and contextualizing the competing evidence and testimony entered into the record both before and after the remand. Accordingly, although Copyright Owners correctly assert that Johnson did not find the uncapped TCC rate structure to be ‘‘unfair, unreasonable or inequitable,’’ Johnson just as clearly did not find that structure to be ‘‘fair, reasonable or equitable.’’ Rather, the purpose of the remand was for the Judges to make these determinations. Accordingly, the Judges next examine whether setting the statutory mechanical rate as an uncapped TCC rate is ‘‘reasonable,’’ as required by section 801(b)(1).50 5. Determining Whether Uncapped TCC Rate Prong is ‘‘Reasonable’’ a. Rejection of First Rationale for Including Uncapped TCC Rate Two substantive issues are implicated raised with regard to the issue of reasonableness: (1) whether the ‘‘seesaw’’ theory is valid; and (2) if it is valid, whether there exist sufficient data to support the phased-in 26.2% uncapped TCC rate.51 To demonstrate that this uncapped TCC rate prong and the (phased-in) 26.2% rate are reasonable, Copyright Owners rely on the combined application of two economic models—the Shapley Value model and a Nash Bargaining Model. Accordingly, it is necessary to consider how these two models relate to each other and how these models and their interrelationship impact the setting of the statutory rate. The D.C. Circuit described the Shapley Value Model methodology: The Shapley methodology is a game theory model that seeks to assign to each market player the average marginal value that the player contributes to the market. This methodology first determines the costs that 50 The Judges consider infra whether any of the four itemized statutory factors require an adjustment to this analysis. 51 As noted supra, in the Judges’ recitation of the parties’ remand arguments regarding the uncapped TCC rate prong, they make other arguments as well, specifically regarding: (1)) whether it would be necessary and/or appropriate to adopt this uncapped TCC rate prong to offset revenue deferral and/or displacement by the Services; (2) whether this rate prong has caused, or would cause, economic ‘‘disruption’’ to the Services (under Factor D of section 801(b)(1)); (3) whether the uncapped TCC rate prong would satisfy Factors B and C of section 801(b)(1); and (4) whether this rate prong improperly imports the complementary oligopoly power of sound recording licensors. The Judges consider these issues after addressing the issues relating to the ‘‘seesaw’’ theory. E:\FR\FM\10AUR2.SGM 10AUR2 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations each player should recover, then divides the ‘‘surplus’’ among the players in proportion to the value of their contributions to the worth of the hypothetical bargain that would be struck. Johnson, 969 F.3d at 372. The Judges provided a consistent but more detailed definition: The Shapley value gives each player his average marginal contribution to the players that precede him, where averages are taken with respect to all potential orders of the players. The Shapley value approach models bargaining processes in a free market by considering all the ways each party to a bargain would add value by agreeing to the bargain and then assigns to each party their average contribution to the cooperative bargain. The idea of the Shapley value is that each party should pay according to its average contribution to cost or be paid according to its average contribution to value. It embodies a notion of fairness. The Shapley model is a game theory model that is ultimately designed to model the outcome in a hypothetical ‘fair’ market environment. It is closely aligned to bargaining models, when all bargainers are on an equal footing in the process. Determination at 62–63 (cleaned up). To apply a Shapley Value Model in a rate proceeding, the economic modeler must obtain usable cost and revenue data to be inputted into the model. More particularly for this proceeding, the modeler must identify the parties’ input costs, including the Services’ noncontent costs, and the revenue derived from interactive streaming.52 The difference between these revenues and the Services’ noncontent costs represents the Shapley ‘‘surplus’’ that can be shared among the Services, the sound recording companies and Copyright Owners. ddrumheller on DSK120RN23PROD with RULES2 (i) The Shapley Approach of the Parties’ Economic Expert Witnesses (a) Professor Gans’s ‘‘Shapley-Inspired’’ Model Professor Gans, Copyright Owners’ expert, utilized royalty and profit interactive streaming data for record companies and music publishers that he obtained from ‘‘a [then] recent music industry equity analysis report,’’ namely, a Goldman, Sachs Equity Research report dated October 4, 2016 entitled ‘‘Music in the Air, Stairway to Heaven.’’ Gans WDT ¶ 76 & n.39. As the Majority summarized Professor Gans’s approach, ‘‘[h]e found that, for the 52 Identifying useful data is a vexing problem. As one of Copyright Owners’ expert economic witnesses, Professor Watt, has written: ‘‘[T]he main problem with the Shapley approach . . . a particularly pressing problem [is] that of data availability.’’ R. Watt, Fair Copyright Remuneration: The Case of Music Radio, 7 Rev. Econ. Rsch Copyright. Issues at 21, 27 (2010). VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 music publishers to recover their costs and achieve profits commensurate with those of the record companies under his approach, the ratio of sound recording royalties to musical works royalties derived from his Shapley-inspired analysis was [REDACTED] (which attributes equal profits to both classes of rights holders and acknowledges the higher costs incurred by record companies compared to music publishers).’’ Determination at 69 (citing Gans WDT ¶ 77 tbl.3) (emphasis added). Regarding Professor Gans’s Shapleyinspired analysis, the Majority stated: [T]he Judges find the ratio of sound recording to musical work royalties that Professor Gans derived from his analysis to be informative. Professor Gans computed this ratio based on an assumption of equal Shapley values between musical works and sound recording copyright owners. The Judges find this assumption to be reasonable . . . . [53] Determination at 70. This is part and parcel of the ‘‘line-drawing’’ undertaken by the Majority that the D.C. Circuit affirmed. Thus, on remand, the Judges do not find cause to reconsider the Majority’s limited adoption of Professor Gans’s Shapley-inspired analysis.54 (b) Professor Marx’s Shapley Value Model Professor Marx constructed two Shapley Value Models, one of which was relied upon by the Majority. In the model credited by the Majority, Professor Marx assumed one collective owner of sound recording copyrights and one collective owner of musical works. She also assumed the presence of a single interactive service. See Determination at 64–68. That approach yielded a total royalty obligation for sound recordings and musical works ranging between [REDACTED]% and [REDACTED]% of the hypothetical service’s revenue. Dissent at 133. Copyright Owners criticized Professor Marx’s decision to assume in her model only one interactive streaming service, rather than the multiple services that actually existed. They contend that assumption reduced the market power of the licensors in her model. According to Copyright Owners’ economic experts, 53 The assumption of equal Shapley values is based on the understanding that a sound recording license and a musical works license are both necessary (i.e., perfect complements) in order for a service to stream a song. Determination at 69 & n.122 therein. 54 Because the ratio of sound recording to musical works royalties that Professor Gans derived from the data and other evidence was the only portion of his testimony on which the Majority relied, and because that reliance was affirmed by the D.C. Circuit, the criticisms of other aspects of Professor Gans’s modeling are no longer relevant. PO 00000 Frm 00013 Fmt 4701 Sfmt 4700 54417 Professor Marx’s approach was a misuse of the Shapley Value Model. They aver that the Shapley Value approach is intended only to eliminate from the rate derivation the bargaining ability of a ‘‘Must Have’’ input supplier (like the sound recording companies and Copyright Owners) to ‘‘hold-out’’ and thus squeeze licensees for higher royalties. By modeling every possible ‘‘arrival ordering,’’ they contend, the ‘‘hold-out’’ problem is avoided. They further contend that Professor Marx misconstrued the purpose of the Shapley approach by wrongly modeling market participants in a manner that significantly reduced the actual market power of these ‘‘Must Have’’ input suppliers. Determination at 66–67. The Majority agreed with Professor Marx. The two Judges in the Majority found that her modeling reasonably ‘‘attempts to eliminate a separate factor—market power—that she asserts renders a market-based Shapley Analysis incompatible with the objectives of Factors B and C of section 801(b)(1).’’ Id. at 68. Although the Majority ultimately relied upon Professor Marx’s modeling in this regard, the Majority found that her data inputs were problematic. Determination at 65. Specifically, Professor Marx relied on 2015 data from Warner/Chappell and Warner Music Group for music publisher sound recording company noncontent costs, respectively. The Majority found that 2015 data was less probative than 2016 data and understated the percentage of revenue to be paid to the two classes of content providers. However, the Majority ultimately found only that this one-year older data served to ‘‘understate’’ the allocation of surplus to the upstream content providers, and thus rejected only her lower [REDACTED]% bound for total royalties, The Majority did decide to adopt her upper bound of [REDACTED]% value for total royalties, which could (and ultimately did) ‘‘constitute a lower bound for total royalties in computing a royalty rate,’’ applied by the Majority in order to make a downward adjustment to offset the complementary oligopoly effect of ‘‘Must Have’’ inputs. Id. at 73, 75. (c) Professor Watt’s Criticisms of and Adjustments to Professor Marx’s Shapley Modeling Professor Richard Watt was called by Copyright Owners as a rebuttal witness at the hearing, for the purpose of reviewing Professor Marx’s WDT. Watt WRT ¶ 3. He concluded that Professor Marx’s Shapley Value Model contains important methodological and data E:\FR\FM\10AUR2.SGM 10AUR2 54418 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations flaws which, in his opinion, caused her to significantly understate the mechanical and overall (musical works + sound recording) royalty rates to be paid by interactive services pursuant to a proper Shapley analysis. Id. at ¶ 5. Professor Watt also criticized her Shapley Value Model for failing to incorporate the fact that ‘‘the different interactive streaming companies— Spotify, Apple Music, Rhapsody/ Napster, Google Play Music, Amazon, etc.—do all compete (and rather fiercely) among themselves, offering (perhaps perfectly) substitutable services.’’ Id. at ¶ 25. Even more strongly in this vein, Professor Watt relied on the following description of the substitutability of the streaming services, inter se: ddrumheller on DSK120RN23PROD with RULES2 Each [interactive streaming] service in the increasingly crowded field is working frantically to overcome the perception that the main distinction among the uniformly priced $9.99 a month offering is little more than font style, quirky playlist title and color scheme. . . . [M]usic platforms have long fought against the perception that they’re . . . selling a nearly interchangeable product . . . You’re getting sold the same car [with] just got a different lick of paint on it.’’). Id. at ¶ 32 n.19. Professor Watt claimed that incorporating this downstream competition into the model would reduce the Shapley values of the Services and increase the Shapley values for the input suppliers, by recognizing which players provide ‘‘essential inputs’’ and which are in competition with other suppliers of substitutable inputs. Id. He further criticized Professor Marx for including in her model ‘‘other distributors’’ who are not interactive streaming services. Id. at ¶ 27. According to Professor Watt, these other distributors ‘‘do not belong in a properly constructed Shapley Value Model because their presence would ‘‘show up’’ in the model as lower revenues for interactive services as their subscribers or listeners left for these other distributors (such as noninteractive services). Id. Additionally, because he criticized Professor Marx’s use of 2015 data (as noted supra), Professor Watt re-worked Professor Marx’s model by examining how the use of 2016 data, as opposed to her 2015 data, would ‘‘better reflect[ ] . . . the reality of the market. Id. at ¶ 37; see also id. at ¶ 44. When using the (higher) 2016 revenues (and making some relatively more minor adjustments he found necessary), Professor Watt estimated that the share of streaming revenues that would be paid out in total royalties (for musical works + sound VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 recordings) in Professor Marx’s model would range from [REDACTED]% to [REDACTED]%. Id. at ¶¶ 50–52.55 After analyzing these Shapley analyses,56 the Majority found that the mechanical royalty rate needed to be increased in order to provide Copyright Owners with a reasonable rate as required by section 801(b)(1). As a matter of arithmetic though, if the mechanical rate increased and the sound recording rate did not decrease by a corresponding amount, then the total royalties paid by the Services would increase. That issue brings the Judges to consideration of Professor Watt’s bargaining model, on which the Majority relied to posit an inverse relationship (the seesaw effect), by which an increase in the mechanical rate would result in a commensurate reduction in the sound recording rate. (ii) Professor Watt’s Bargaining Model Professor Watt’s Nash Bargaining Model is the linchpin that connects: (a) the higher mechanical royalty rates generated by the Shapley Value results relied upon by the Majority with (b) the assumed lower sound recording rates— a connection that the Majority found to render ‘‘reasonable’’ and ‘‘fair’’ its uncapped TCC prong. See Determination at 73–74 (‘‘As to the issue of applying a TCC percentage to a sound recording royalty rate that is artificially high as a result of musical works rates being held artificially low through regulation, the Judges rely on Professor Watt’s insight (demonstrated by his bargaining model) that sound recording royalty rates in the unregulated market will decline in response to an increase in the compulsory license rate for musical works.’’). Alternately stated, Professor Watt’s bargaining model result, i.e., the seesaw effect, if sufficiently supported in the record, is the phenomenon that would allow the Judges on remand to apply the Shapley results by increasing the mechanical rate, without unduly exposing the Services to the risk of higher total royalties. 55 As noted supra, when the Majority weighed and credited Professor Watt’s entire Shapley analysis, in which his estimate of total royalties was [REDACTED]%, those Judges contextualized Professor Marx’s [REDACTED]% total royalty calculation as the lower bound of a zone of reasonable rates, and applied it as a measure that, in their analysis, would offset the complementary oligopoly effect of real-world royalties. Determination at 75 (text and tbl.). 56 Because his testimony was made in rebuttal, leaving the Services no procedural right to file written testimony in opposition, the Majority gave little weight to Professor Watt’s total royalty projections and no weight to his proffered ratios of sound recordings-to-musical works royalties. Determination at 75. PO 00000 Frm 00014 Fmt 4701 Sfmt 4700 More particularly, the Majority recognized a potential problem that those Judges would have to resolve before utilizing the Shapley Value approach to create an uncapped TCC prong: ‘‘This is problematic because the sound recording rate against which the TCC rate would be applied is inflated . . . both by . . . complementary oligopoly [market] conditions . . . and the record companies’ ability to obtain most of the available surplus due to the music publishers’ absence from the bargaining table.’’ Determination at 73.57 But the Majority found that Professor Watt had provided a rationale which permitted them to resolve the second problem: As to the issue of applying a TCC percentage to a sound recording royalty rate that is artificially high as a result of musical works rates being held artificially low through regulation, the Judges rely on Professor Watt’s insight . . . that sound recording royalty rates in the unregulated market will decline in response to an increase in the compulsory license rate for musical works. 3/27/17 Tr. 3090 (Watt) (‘‘[T]he reason why the sound recording rate is so very high is because the statutory rate is very low. And if you increase the statutory rate, the bargained sound recording rate will go down.’’). Determination at 73–74; see also Watt WRT ¶ 23 n.13 (‘‘[I]in my Appendix 3, I show that . . . if the musical works rate is increased to what would be a realistically fair and reasonable rate, then the negotiated fee for sound recordings would decrease almost dollar for dollar . . . .’’); see also id. at ¶ 36 (‘‘The statutory rate for mechanical royalties . . . is significantly below the predicted fair rate, and the statutory rate effectively removes the musical works rightsholders from the bargaining table with the services. Since this leaves the sound recording rightsholders as the only remaining essential input, bargaining theory tells us that they will successfully obtain most of the available surplus.’’).58 57 The other problem the Majority needed to resolve was how to deflate the market-based sound recording royalty rates to mitigate the complementary oligopoly effect in those rates. Id. As discussed supra, the Judges resolved this problem by applying the low total royalty payment sum, [REDACTED]%, from Professor Marx’s Shapley Value Model. 58 In full detail, Professor Watt concluded: ‘‘[F]or every dollar that the statutory rate for musical works undercuts a fair and reasonable rate, the freely negotiated rate for sound recordings will increase by an estimated [REDACTED] cents. That is, if the musical works rate is increased to what would be a realistically fair and reasonable rate, then the negotiated fee for sound recordings would decrease almost dollar for dollar, with only a minor change in the total royalty rate for all copyrights combined.’’ Id. at ¶ 23, n.13; see also id., appx. 3 at 12. E:\FR\FM\10AUR2.SGM 10AUR2 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations To repeat: This inverse relationship is what has been described as the ‘‘seesaw’’ effect. The question in this regard on remand is whether the record proves that the seesaw theory is valid and measurable going forward. Alternately stated, does the record prove that Professor Watt’s bargaining model serves as the linchpin that would allow the Judges to apply the Shapley results by increasing the mechanical rate, without unduly exposing the Services to the risk of higher total royalties? To resolve this issue, the Judges examine this bargaining model dispute in detail, as it bears on whether the uncapped TCC rate structure can be incorporated into the statutory rate. (a) Bargaining Model Dispute ddrumheller on DSK120RN23PROD with RULES2 Professor Watt utilized a general Nash Bargaining Model.59 In his particular application, Professor Watt modeled the streaming services and the labels each as a ‘‘single unit,’’ asserting (as is common in Shapley analyses) that this single-unit modeling was done ‘‘for simplicity.’’ Watt WRT, appx. 3 at 10. Applying this and other modeling assumptions, Professor Watt posited: ‘‘If there were to be no successful deal, then each of these two bargainers [the assumed ‘‘single’’ interactive service and ‘‘single’’ label] would earn 0, since in that case the interactive streaming service could not operate.’’ Id. In his oral testimony at the hearing, Professor Watt did not opine as to whether changes in variables other than musical works royalties would also have an impact on the level of sound recording royalty rates, even as higher musical works rates would otherwise place virtually 1:1 downward pressure on the sound recording rate. However, in his written rebuttal hearing testimony, i.e., his WRT, Professor Watt did make varying assumptions regarding the changes in the Services’ non-content costs, by which he did change the total revenue share for content providers. Watt WRT ¶¶ 50–52. He concluded from this varying replication of Professor Marx’s Shapley model ‘‘that the results that it delivers are very dependent upon the amount of total interactive 59 The Nash Bargaining Model is one type of game-theoretic approach used by economists to model the distribution of ‘‘gains from trade’’ between two parties ‘‘in a manner that reflects ‘fairly’ the bargaining strength of the different agents. Marx WDRT ¶ 28 n.33 (citing A. Mas-Colell, M. Whinston, and J. Green, Microeconomic Theory 838 (1995)). To understand the parties’ modeling dispute, it is necessary to appreciate the essential elements of the Nash Bargaining Model, as previously summarized by the Judges: ‘‘In the Nash Framework [for full quotation, see eCRB no. 27063 n.48].’’ SDARS III Final Determination, 83 FR 65210, 65215 & n.32 therein (Dec. 19, 2018). VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 streaming revenue and the fraction of that revenue that is taken up by downstream non-content costs.’’ Id. at ¶ 53 (emphasis added).60 The Services had no procedural right under part 351 of the Judges’ regulations to proffer surrebuttal written testimony from economic witnesses to challenge Professor Watt’s assertion, made for the first time in rebuttal, of the seesaw relationship between changes in the musical works royalty rate and the sound recording royalty rate paid by interactive services. Moreover, the Services and their economists also had no opportunity to weigh in on the Majority’s application of same (which was not revealed until the Judges rendered their decision). See Johnson, 969 F.3d at 381 (‘‘Streaming Services had no notice that they needed to defend against and create a record addressing such a significant, and significantly adverse, overhaul of the mechanical license royalty scheme.’’).61 Now though, on this remand, the Services have been afforded the opportunity to present these criticisms, through their expert witnesses. (b) Professor Katz’s Principal Criticism Pandora’s economic expert, Professor Michael Katz, levied several criticisms of the bargaining model proffered by Professor Watt and applied by the Majority. The most important problem with Professor Watt’s analysis, according to Professor Katz, is that the former’s model assumes an ‘‘extremely unrealistic’’ zero payoff to the label in the absence of an agreement with a streaming service—an assumption which is ‘‘far from . . . innocuous.’’ Written Direct Remand Testimony of Professor Michael Katz (Katz WDRT) ¶¶ 16, 20. Professor Katz opines that this zero payoff assumption is equivalent to assuming, contrary to undisputed market facts, that: (1) subscribers and listeners to an interactive service would not switch to other interactive services if that service failed to reach an agreement with the labels; and (2) the interactive service is a ‘‘Must-Have’’ 60 The Judges take note here of Professor Watt’s presentment of alternative scenarios, because, as discussed infra, the Services and their economists accuse Professor Watt of changing his testimony, post-remand, by limiting the scenarios in which his ‘‘seesaw’’ argument would apply in order to salvage the credibility of his bargaining model. 61 The Services could have sought leave to file surrebuttal testimony, and could have challenged the Majority’s understanding of Professor Watt’s testimony, after the Initial Determination, by filing a Motion for Rehearing pursuant to 37 CFR 353.1. However, a party is not required to engage in either of these procedural approaches, but rather may challenge the Determination on appeal, as has occurred here. PO 00000 Frm 00015 Fmt 4701 Sfmt 4700 54419 input supplier. Katz WDRT ¶¶ 17–18. In terms of Nash modeling, according to Professor Katz, Professor Watt’s assumption is thus equivalent to ‘‘assum[ing] that the sound recording copyright owners have no outside option.’’ Katz WDRT ¶ 127 (app. A) (emphasis added). Moreover, not only does Professor Katz assert the indisputability that such substitution would occur, he points out that Professor Watt himself acknowledged in his own testimony that such substitution would occur. Katz WDRT ¶ 19.62 Beyond this purported inconsistency, Professor Katz finds Professor Watt’s nosubstitution assumption to be a serious modeling error because, in order to quantify accurately each Nash bargainer’s contribution to the net surplus to be divided, the extent of substitutability on each side of the market must be captured by the modeling. Katz WDRT ¶ 20. That is, he opines that ‘‘Professor Watt’s assumption that there is no substitution dramatically biases his model toward finding a large seesaw effect and renders his analysis unreliable . . . lead[ing]to a prediction that the share of an increase in musical works royalties that will fall on the streaming services is approximately eight times larger than Professor Watt’s prediction. Id. at ¶ 21. As a matter of music business dynamics, Professor Katz interprets Professor Watt’s substitutability error as follows. The assumption that a label receives a zero payoff if it does not reach agreement with a streaming service is equivalent to assuming that, if a streaming service shut down, none of the consumers who would otherwise have used that streaming service will switch to alternative streaming services or other sources of licensed music. The two forms of the assumption are equivalent because, when the services are substitutes, failure to reach an agreement with one service will not drive a label’s payoffs from interactive streaming to zero. It will not result in the loss of all of the benefits that could be enjoyed by reaching an agreement. Instead, many consumers would engage in substitution and choose other streaming services, which will allow the label to earn profits from the additional royalties that would be paid to it by those other services. Id. at ¶ 18. Professor Katz attempts to adjust Professor Watt’s Nash Bargaining Model to account for this substitution effect. In his Appendix A, Professor Katz— acknowledging the reality of multiple interactive services—changes Professor Watt’s assumed single label’s payoff 62 The Judges have quoted Professor Watt’s testimony in this regard supra. E:\FR\FM\10AUR2.SGM 10AUR2 54420 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations (designated as parameter ‘‘A’’ in the Nash Bargaining Model) from a value of zero to a value equal to ‘‘the share of revenues that would be diverted to other streaming services’’ multiplied by ‘‘the royalty rate that the label receives from the other interactive streaming services.’’ Id. ¶¶ 119, 127. Professor Katz asserts that the diversion to other streaming services represents an ‘‘outside option’’ available to a label. Id. ¶ 127. Professor Katz incorporates this ‘‘outside option’’ in his revised version of Professor Watt’s Nash Bargaining Model. In addition, Professor Katz asserts that Professor Watt’s modeling is unreliable because ‘‘his prediction of the size of the see-saw effect is very sensitive to the assumed values of various other parameters.’’ Id. at ¶ 23. For example, Professor Katz asserts that a change in the royalty rate paid to the labels could materially affect the balance or even the existence of the seesaw effect. Id. at ¶ 127. As further support for his opinion, Professor Katz relies on the testimony of one of Copyright Owners’ own economic expert witnesses, who gave testimony clearly indicating that the ‘‘seesaw’’ effect was not at all likely to occur. Id. ¶ 24, n.16 (citing Gans WRT ¶ 32).63 In sum, Professor Katz finds Professor Watt’s Nash Bargaining Model to be unusable as a foundation to set royalty rates because, although ‘‘there are theoretical reasons to believe that a seesaw effect may occur, . . . there are complications and it is difficult to predict how big the effect will be.’’ Id. ¶ 24 (emphasis added). ddrumheller on DSK120RN23PROD with RULES2 (c) Professor Watt’s Rebuttal to Professor Katz In rebuttal to Professor Katz’s criticisms, Professor Watt states that ‘‘the record needs to be straight on Nash bargaining theory,’’ in order to explain ‘‘the foundational error’’ committed by Professor Katz. Watt RWRT ¶ 52. This basic mistake, according to Professor Watt, is Professor Katz’s erroneous assertion that the bargaining model must account for a label’s ‘‘outside option.’’ Id. ¶ 53. Relying on economic authority regarding bargaining theory, Professor Watt defines an ‘‘outside option’’ as ‘‘the best alternative that a player can command if he withdraws 63 In this regard, Professor Gans testified: ‘‘[When considering] the general distribution of profit when royalty rates for musical works rightsholders are increased[,] [i]n principle, those funds could come from a decrease in service profit, a decrease in sound recording royalties, or an increase in consumer pricing . . . . The general redistribution of profit in response to increased musical works royalties is fundamentally an empirical question. . . .’’ Gans WRT ¶ 32. VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 unilaterally from the bargaining process.’’ Id. ¶ 59 (emphasis added); see also id. ¶ 53 (‘‘An outside option is a payoff that the label would receive if negotiations with the service do not result in an agreement.’’) (emphasis added).64 Connecting this principle of bargaining theory to economic theory, Professor Watt explains his understanding of the relationship of the ‘‘outside option’’ to the more familiar economic concept of ‘‘opportunity cost’’: An outside option could also be referred to as an ‘‘opportunity cost,’’ since it is the value of what would be foregone should a deal with the service actually be struck. It is . . . useful to recognize the equivalence between an outside option and an opportunity cost, because economics in general has a very long history of understanding how opportunity costs weigh in on economic decision making. Id. Professor Watt then opines how Professor Katz confused the ‘‘outside option’’ with the disagreement (a/k/a threat) point in the Nash Bargaining Model: [Professor] Katz claim[s] that the outside option value that the labels would enjoy should they not reach an agreement with the services should be included as part of the ‘‘disagreement point’’ within the bargaining model and reimbursed like a cost prior to bargaining. Doing this can dramatically alter the results of the model. It is also definitively not how such an option should be modelled. [Professor] Katz [is] guilty of misunderstanding the Nash bargaining model, and concretely, the meaning of a ‘‘disagreement point,’’ and the way that an outside option should be brought into the model. Id. ¶ 55. More particularly, according to Professor Watt, these outside options/ opportunity costs do not belong in a Nash Bargaining Model, because they are ‘‘not the types of status quo actual financial payments that may be modelled as disagreement points.’’ Id. ¶ 57. Rather, he asserts that, as Professor Katz essentially acknowledged, they are ‘‘payoffs from substitution, [i.e.,] an option instead of the deal, and they are not actual financial payments, but opportunity costs. Id. 64 The phrase ‘‘outside option’’ suggests the existence of an ‘‘inside option.’’ Indeed, a treatise cited by Professor Watt identifies the ‘‘inside option,’’ defining it as ‘‘[t]he payoff the [bargainer] obtains while the parties temporarily disagree’’— contrasting it with the ‘‘outside option’’ as (consistent with Professor Watt’s testimony) ‘‘the payoff [the bargainer] obtains if she chooses to permanently stop bargaining, and chooses not to reach an agreement with [the counterparty].’’ A. Muthoo, Bargaining Theory with Applications at 137 (1999). PO 00000 Frm 00016 Fmt 4701 Sfmt 4700 Professor Watt then explains that an outside option/opportunity that by definition exists as an alternative to a bargain between two parties lies outside the two parties’ bargain, and is thus outof-place within a proper Nash Bargaining Model: In the case at hand, if the parties never stop negotiating and never take up substitute options, then no joint enterprise is offered and there is no surplus to share, so each necessarily gets a payoff equal to 0, just as I assumed in my model. . . . [A]gainst this backdrop, an outside option (a potential payoff that is not directly related to a share of the surplus that is being negotiated) . . . comes in [to the model] as a constraint upon the set of feasible deals that could be struck, exactly as an opportunity cost would be treated. Id. ¶¶ 57–58. (d) Dr. Leonard’s Criticisms of Professor Watt’s Bargaining Model According to Google’s economic expert witness, Dr. Gregory Leonard, the Majority wrongly relied on Professor Watt’s bargaining model because it is ‘‘highly stylized’’ and theoretically ‘‘simplified’’ in ways that make it unable to predict that ‘‘an increase in the musical works royalty would be offset nearly dollar-for-dollar by a decrease in the sound recording royalties (the ‘‘seesaw effect’’), thus leaving the services virtually unaffected by the proposed increase in musical works royalties.’’ Leonard WDRT ¶ 8. Pointedly, Dr. Leonard criticizes Professor Watt’s bargaining model as comprised of a ‘‘veneer of ‘complexity’ . . . mathematical formulas and [a] reference to John Nash,’’ adopted to provide a rationalization for adoption of his Shapley Value modeling that would significantly increase the mechanical royalty rate.’’ Id. ¶ 16. These modeling deficiencies, Dr. Leonard asserts, are not merely ‘‘simplifying assumptions [that] better focus on the specific question the model is meant to address,’’ but rather ‘‘simplify away economic characteristics . . . entirely abstract[ing] away economic characteristics . . . central to the question at hand.’’ Id. ¶ 18. In particular, Dr. Leonard avers that Professor Watt’s bargaining model materially abstracts away from, inter alia: (1) the nature of consumer demand for streaming services and competing forms of music; (2) how services decide to enter or exit the streaming market; (3) the nature of the oligopolistic interaction among the labels; (4) the nature and timing of the bargaining between each label and each service; (5) E:\FR\FM\10AUR2.SGM 10AUR2 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations the potential for ‘‘hold-up’’ 65 by labels that perceive the services to be in a vulnerable bargaining position due to their previous industry-specific investments made under their assumption that the pre-existing statutory structure would be maintained; and (6) the failure of Professor Watt’s bargaining model to grapple with the complementary oligopoly structure of the sound recording market. Id. ¶¶ 18, 20. These factors, he posited, are ‘‘important for determining how sound recording royalties would actually change in response to a change in the statutory musical works royalty.’’ Id. Professor Leonard concludes that, by not modeling these factors, Professor Watt’s ‘‘prediction of a virtual dollar for dollar decrease in sound recording royalties is unreliable as a basis for formulating policy.’’ Id. ¶ 20. Regarding the complementary oligopoly structure of the market and its impact on the bargaining process, Professor Leonard emphasizes that an important ‘‘real-world hurdle’’ assumed away by Professor Watt’s modeling of a single label entity is that ‘‘each label would prefer to have the other labels lower their sound recording royalties while maintaining its own royalties at pre-existing levels . . . .’’ Id. ¶ 21. More particularly, Dr. Leonard explains that ‘‘even if a label were to recognize that it is more efficient for overall sound recording royalties to be lower, the label may not be willing to lower its royalty rate without assurance that the other labels will do the same,’’ a result which he asserts ‘‘is unlikely to happen absent some form of collusive behavior.’’ Id. Thus, Dr. Leonard maintains that the existence and size of any ‘‘seesaw’’induced decrease in sound recording royalties remains indeterminate, and it remains ‘‘within the realm of theoretical possibility that the labels do not agree to any reduction in sound recording royalties even if a reduction in overall royalties would be economically efficient. Id. ddrumheller on DSK120RN23PROD with RULES2 (e) Professor Watt’s Rebuttal to Dr. Leonard’s Criticisms Professor Watt replies with a spirited defense of economic modeling in 65 A hold-up problem occurs when: (1) parties to a future transaction must make specific investments prior to the transaction in order to prepare for it; and (2) the exact form of the optimal transaction (e.g., how many units if any, what quality level, the time of delivery) cannot be specified with certainty ex ante. W. Rogerson, Contractual Solutions to the Hold-Up Problem, Rev. Econ. Stud. 777 (1992). Here, the interactive services may need to commit to paying for long-term investments, even though they cannot know the level of their largest costs (content royalties) beyond a single rate term. VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 general and his economic bargaining model in particular. He begins by pointing out that models are not supposed to be ‘‘perfect representations of reality [but rather] are intended to isolate what is important, in order to expose a useful insight on some issue of relevance.’’ Watt RWRT ¶ 105. He adds that economic models (not merely his bargaining model) ‘‘do not necessarily deliver predictions of situations that are immune to changes in variables outside the model, but rather the results inform conclusions about the relationships between the variables and parameters within the model, [which is] by nature a crude representation[ ] of reality, but the lessons and insights that they provide can be very relevant to realworld applications.’’ Id. ¶¶ 106–07 (emphasis added). With particular regard to his bargaining model, Professor Watt takes issue with Dr. Leonard’s assertion that in the former’s model the surplus is a ‘‘fixed constant.’’ See Watt RWRT ¶¶ 110–111. Rather, Professor Watt avers that his bargaining model assume[s] that when the surplus . . . whatever value it takes . . . is to be shared, the parties understand that the amount to be shared is, at that moment, given.’’ Id. ¶ 111 (emphasis added). Turning to Dr. Leonard’s critique regarding the purported distortionary effect of Professor Watt’s modeling assumption of a single label and a single interactive service, Professor Watt responds by acknowledging that, if he had modeled multiple labels and services in the bargaining process, that would be ‘‘not particularly enlightening vis-a`-vis the single bargain setting, as it will not lead to different insights than those distilled by the [Majority].’’ Id. ¶ 113.66 Further, Professor Watt characterizes this criticism as ‘‘empty,’’ because under either his two-player Nash model or Dr. Leonard’s posited multi-player (Nash-in-Nash) model, the labels will not respond to a musical works royalty increase ipso facto with a reduction in the sound recording royalty (i.e., the seesaw effect will not occur if there is ‘‘a change in some other variable.’’). Id. ¶ 114. (f) Professor Marx’s Criticisms of Professor Watt’s Bargaining Model Professor Marx criticizes Professor Watt’s application of the Nash Bargaining Model because, in her opinion, its ‘‘precise prediction’’ of the 66 Professor Watt describes Dr. Leonard’s multiple simultaneous negotiations in a bargaining model as a ‘‘Nash-in-Nash’’ model, but the former does not explain why he concludes that this approach ‘‘will not lead to different insights’’ than those the Majority distilled from his two-party Nash model. PO 00000 Frm 00017 Fmt 4701 Sfmt 4700 54421 nearly one-to-one seesaw relationship ‘‘depends critically on the assumptions that he makes and the numerical inputs that he uses.’’ Marx WDRT ¶ 33. First, criticizing his modeling assumptions, like Professor Katz, she criticizes his decision to abstract from reality by positing a single label and a single interactive streaming service. She opines that his one label/one service modeling assumption ineluctably leads to his conclusion that each of these two parties ‘‘has a ‘disagreement payoff’ of zero [meaning that] each party ends up with nothing in the absence of a deal.’’ Id. ¶ 34. But this zero ‘‘disagreement payoff’’ is merely a product of Professor Watt’s abstraction from reality, according to Professor Marx, because ‘‘[i]n reality, if interactive streaming went away, a share of the music listening that had occurred through interactive streaming services would migrate to other forms of music distribution, generating revenues for the label . . . meaning that the disagreement payoff would be positive for the label). Id. (emphasis added).67 Consistent with Professor Katz, she maintains that Professor Watt himself acknowledged the presence of this substitution effect when he testified that ‘‘[t]he existing interactive streaming companies do not hold an essential input, as first they compete with the non-interactive services . . . .’’ Id. ¶ 35, n.43 (citing Watt WRT, app. 3). More particularly, Professor Marx maintains, a record label’s disagreement payoff must be considered realistically ‘‘in any accounting of what would happen if record labels and interactive streaming services failed to reach an Agreement . . . .’’ Marx RWDT ¶ 35. And, she opines, when this real-world substitution effect is taken into account, the seesaw effect that Professor Watt estimates is reduced dramatically, because ‘‘[t]he greater . . . the substitution between streaming and other forms of distribution, the greater is the revenue that the record label can capture in the event of disagreement 67 Professor Marx’s reference to a substitution from a shutdown interactive service to ‘‘other forms of music distribution’’ is different from, but analytically analogous to, Professor Katz’s assertion that the shutdown of any one interactive service would result in migration of its subscribers and other users to the remaining interactive services. These analogous critiques are complementary. See Marx WDRT ¶ 37 (‘‘One would expect the same decrease in the estimated see-saw effect by including a second, competing interactive streaming service in the market instead of just the one that Professor Watt uses. In that case, if no deal is reached, users would migrate to an even closer substitute—a competing interactive streaming service—resulting in an even higher degree of profit migration and thus an even lower estimated see-saw effect’’). E:\FR\FM\10AUR2.SGM 10AUR2 54422 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations ddrumheller on DSK120RN23PROD with RULES2 and the lower is the estimated see-saw effect.’’ Id.68 Professor Marx opines that modeling the bargaining process without these real-world particulars diminishes the value of Professor Watt’s Nash model in several significant ways. First, because his model fails to incorporate the presence of three major record labels, ‘‘each with substantial complementary oligopoly power,’’ it fails to capture the fact that ‘‘each record label does not fully internalize the impact of its rates on the viability of the industry.’’ Id. ¶ 39. She points to the Judges’ Final Determination in Web IV, where the Judges note how this aspect of complementary oligopoly compromises the value of a rate as a useful benchmark. Id. ¶ 39 n.45 (quoting Web IV Final Determination). More particularly, she opines that when, as here, ‘‘there are multiple negotiations between multiple record labels and multiple services,’’ sound recording rates can be affected ‘‘by the order of negotiations’’ among the several label:service negotiating pairs—a factor that Professor Watt’s bargaining model fails to capture. Marx WRDRT ¶ 41. Next, Professor Marx avers that Professor Watt’s bargaining model ‘‘does not explain how or over what time frame the market would move to a new equilibrium.’’ Id. ¶ 40. More particularly, she testifies, because interactive services’ ‘‘agreements with record labels often contain multi-year terms and can take many years to negotiate . . . there may be little incentive or practical ability for both sides to move to a new rate before the contract expires’’. Id. ¶ 41. She takes note that this point was established at the hearing during questioning of Professor Watt from the bench: JUDGE STRICKLER: What of the situation . . . that the . . . time period for the existing agreements between the . . . labels and the interactive streamers is such that they’ve already locked in a particular rate and then we set a rate that’s higher for the mechanical to reflect the fact that the sound recording royalty should drop, but it’s locked in for a period of time? Are we running the risk, then, of disrupting the market by having a total royalty that’s greater than what is indicated by your Shapley testimony, simply because of the disparity of times in which the rates are . . . implemented? PROFESSOR WATT: That’s a very fair point. And I didn’t even think of that until 68 In the context of the bargaining model, Professor Marx identifies Professor Watt’s choice of ‘‘a market structure that is completely symmetric between record labels and services not reflective of the real world’’ as forcing his model ‘‘to attribute[ ] all the . . . surplus division to . . . bargaining power . . . and none of it to the market structure.’’ Id. ¶ 38. VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 you’ve mentioned it . . . [T]he model I have done is . . . assuming that . . . the bargained thing happens at the same time as the—or in the same general period of time as a change in the statutory rate. You’re absolutely correct. 3/27/17 Tr. 3091–92 (Watt); see Marx WRDRT ¶ 42, n.46 Third, Professor Marx points out that Professor Watt’s Nash model does not attempt to capture the effects of the heterogeneous and asymmetric distribution of information relevant to the bargain available to each party at the time of negotiation. Id. ¶ 41. Lastly, Professor Marx avers that Professor Watt’s Nash Bargaining Model fails to address, on a more general basis beyond informational issues, other ‘‘asymmetries among record labels and among services.’’ Marx WDRT ¶ 41. In sum, Professor Marx concludes that these foregoing real-world points all preclude the Judges from relying on Professor Watt’s testimony to identify a stable relationship between changes in the mechanical royalty rate and the sound recording royalty rate because they all share a common defect—they ‘‘lie outside Professor Watt’s model.’’ Marx WRDT ¶ 41. To be clear, Professor Marx does not criticize Professor Watt for neglecting to include these points in his bargaining model; rather, she acknowledges that ‘‘[t]hese are difficult features to capture in a tractable equilibrium model.’’ Id. Indeed, she urges the Judges to appreciate that relying on such a necessarily limited model, as the Majority did, can have ‘‘dramatic effects’’ on the royalty rates derived. Id. Professor Marx emphasizes that all of these inherent modeling deficiencies are especially pernicious, if the bargaining model is applied yet again on remand, to set specific rates over a five-year period, when other variables will have independent effect on royalty rates. Id. (g) Professor Watt’s Rebuttal to Professor Marx Because Professor Marx’s criticisms are of a similar nature to Professor Katz’s criticisms, Professor Watt responds to Professor Marx as he did to Professor Katz. To summarize, Professor Watt responds to Professor Marx’s points as follows: • Her criticism is centered on what he characterizes as her ‘‘bogus’’ argument that he supposedly had predicted almost a ‘‘dollar for dollar’’ sound recording rate reduction in response to an increase in the musical works rate (the seesaw effect). Watt RWRT ¶ 19. Professor Watt finds this argument ‘‘particularly disheartening,’’ because Nash bargaining theory explains why PO 00000 Frm 00018 Fmt 4701 Sfmt 4700 the seesaw would apply to the splitting of the surplus based on the available data, and that ‘‘there are quite apparent reasons why available surplus may not decrease even if the musical works rate increased, because of simultaneous changes to other variables in the model.’’ Id. ¶ 34 (emphasis added). • Professor Marx implicitly contradicts her own reliance on the complementary oligopoly power of the Major labels by modifying his bargaining model through the insertion of a lower value for their bargaining power. Id. ¶¶ 19, 22–24, 26. • Professor Marx misconstrues the purpose of his Nash model, which was to serve ‘‘as a reply’’ to Professor Marx’s direct testimony, and ‘‘to show bargaining insights that bore upon aspects of the case.’’ Id. ¶ 29. • Professor Marx, like Professor Katz, improperly includes in her bargaining model a potential payoff for the label arising from an ‘‘outside option,’’ i.e., from an alternative that the label can choose only if the Nash bargaining terminates. Id. ¶¶ 53—68. (h) Professor Marx’s Reply to Professor Watt’s Criticism 69 In her supplemental remand testimony, Professor Marx challenged several of Professor Watt’s criticisms contained in his remand testimony. First, she takes issue with what he identified as two ‘‘core’’ economic principles of bargaining: (1) that all of the available net surplus will be shared; and (2) that neither of the two bargainers will demand a share such that more than the total net surplus is shared. Marx WSRT ¶¶ 7–8. As an initial matter, she disputes the notion that these are ‘‘core’’ principles of bargaining. Id. ¶ 8. More particularly, she states that, in the present case, because ‘‘the label does not know with exactitude the precise maximum that a service would be willing to pay (i.e., its ‘‘survival’’ rate), and the service likewise does not know the exact minimum that the label would be willing to accept,’’ the simple bargaining model must be expanded to address ‘‘the potential for delay and/or bargaining breakdown.’’ Id. As a further criticism, Professor Marx avers that ‘‘[i]n the real world, the negotiated royalty outcomes do not involve just two parties, but rather a sequence of overlapping, interrelated, 69 The Judges found that Professor Watt’s remand testimony, denoted as ‘‘rebuttal,’’ also provided de facto ‘‘direct’’ testimony, to which the Services could respond with supplemental testimony and argument. Oct. 1st Order at 11–12. Professor Marx’s response in the following text was set forth in Spotify’s permitted supplemental testimony. E:\FR\FM\10AUR2.SGM 10AUR2 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations bilateral bargains involving multiple competing services and multiple record labels with complementary oligopoly power.’’ Id. ¶ 12.70 This complication, she opines, exacerbates the informational deficit noted in the immediately preceding paragraph, such that negotiations within the several pairings of labels and services ‘‘are affected by uncertainty and private information and . . . Professor Watt’s discussion of bargaining theory [thus] does not support any particular realworld see-saw outcome.’’ Id. (iii) Resolution of the Bargaining Dispute (a) Professor Watt’s Nash Bargaining Model Does Not Support Adoption of Uncapped TCC Rate ddrumheller on DSK120RN23PROD with RULES2 The purpose of Professor Watt’s Nash Bargaining Model was to allay the Judges’ concern that increasing the mechanical rate would lead to higher total royalties for the Services. His bargaining model was understood by the Majority to show that such higher total royalties would not result, because the model demonstrated the ‘‘seesaw’’ effect, whereby the sound recording rate would fall almost dollar-for-dollar with the increase in the mechanical rate. See Determination at 73–74 (‘‘[T]he Judges rely on Professor Watt’s insight . . . demonstrated by his bargaining model that sound recording royalty rates in the unregulated market will decline in response to an increase in the compulsory license rate for musical works. . . . Professor Watt’s bargaining model predicts that the total of musical works and sound recordings royalties would stay ‘almost the same’ in response to an increase in the statutory royalty.’’) (emphasis added).71 On the surface, the economic experts on both sides appear to be at loggerheads regarding the existence and applicability of the seesaw relationship. However, as discussed below, on further analysis of their respective positions, in light of Professor Watt’s remand testimony regarding a key assumption in his bargaining model, their disagreement narrows considerably 70 In like manner, Professor Marx opines that Professor Spulber’s discussion of bargaining theory is irrelevant to any assessment of ‘‘the complexities affecting real-world negotiations’’ and the presence, vel non, of a seesaw outcome. Id. ¶ 13. 71 Copyright Owners note the Majority’s recognition that, regardless of the rate structure, i.e., uncapped TCC or otherwise, Professor Watt’s ‘‘insight’’ from ‘‘bargaining theory’’ would still apply. See Determination at 74, n.138. That being the case, the Majority’s first rationale for adopting an uncapped TCC rate is undermined. VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 and—in an important respect—vanishes completely.72 To recap: In his WRT, Professor Watt stated [W]ith an appropriately modelled bargaining analysis . . . in my Appendix 3 . . . I show that for every dollar that the statutory rate for musical works undercuts a fair and reasonable rate, the freely negotiated rate for sound recordings will increase by an estimated [REDACTED] cents. That is, if the musical works rate is increased to what would be a realistically fair and reasonable rate, then the negotiated fee for sound recordings would decrease almost dollar for dollar, with only a minor change in the total royalty rate for all copyrights combined. Watt WRT ¶ 23 & n.13. But nowhere in his WRT did he qualify this statement by explicitly acknowledging that in his bargaining model there are certain assumptions lurking, i.e., that his ‘‘concrete’’ analysis is subject to the ‘‘ceteris paribus’’ constraint—that all other things are held constant (i.e., equal before and after the change in the musical works rate) other things being equal).73 It is only in his later remand testimony—after the D.C. Circuit’s remand had compelled him to confront criticism from adverse economists—that Professor Watt expresses this assumption overtly, making explicit the ‘‘understanding’’ that he had theretofore only tacitly assumed: In other words, a model in which only the two copyright rates are permitted to change . . . as was the understanding in my original model, allows the system to derive a clear 72 This is unsurprising. The difference of opinion among economists often lies in their assumptions, which may be left unstated or opaque (intentionally or not). Once those assumptions are laid upon the table, their differences often evaporate. As the esteemed economist Fritz Machlup noted more than sixty years ago: ‘‘The most prolific source of disagreement lies in differences of factual assumptions. It is not customary for experts to state all the assumptions that underlie their conclusions; it would be much too cumbersome. But when they have reached very different conclusions, then we are forced to go back and find out what implicit assumptions they have made.’’ F. Machlup, Why Economists Disagree, 109 Proceedings of the American Philosophical Society 1, 3 (1965). In the modern world of more formal economic modeling as well, the obfuscation of assumptions continues to be an important source of dispute, according to a book written by a leading game theorist upon which Professor Watt relies in his testimony. A. Rubinstein, Economic Fables at 20 (2012) (‘‘[T]he model’s formal mantle enables economists . . . to conceal from the layman the assumptions the model uses.’’); see J. Schlefer, The Assumptions Economists Make at 29 (2012) ([S]ome assumptions made by economists capture important insights, others are insane. All you have to do is decide which capture insights, which are insane, and in which situations.’’) 73 In his oral testimony, Professor Watt likewise did not qualify his opinion by taking note of his ceteris paribus assumption. See 3/27/17 Tr. 3026 et seq. (Watt). PO 00000 Frm 00019 Fmt 4701 Sfmt 4700 54423 relationship between those two rates, and that relationship is that an increase in one leads to a decrease in the other, that is, the ‘see-saw effect.’ But if . . . something else changes along with the musical works rate . . . then the net effect does not predict that the negotiated rate of the labels will decrease.’’ Watt RWRT ¶ 35 (emphasis added). Indeed, as noted supra, Professor Watt did give a nod to the relaxing of his implied ceteris paribus assumption in his WRT, by identifying varying ‘‘scenarios’’ in which he considered the impact of potential changes in service revenues and service non-content costs, leading to different percentages of royalties paid to content providers. Watt WRT ¶¶ 45–52. Professor Watt then used these several assumptions and scenarios to opine as follows: ‘‘The message that should be taken from this exercise . . . is that the results . . . are very dependent upon the amount of total interactive streaming revenue and the fraction of that revenue that is taken up by downstream non-content costs.’’ Id. ¶ 53.74 Professor Spulber, on behalf of Copyright Owners, likewise emphasizes on remand the importance of the ceteris paribus assumption in economic modeling: [A]long with an increase in the compulsory license rate, all other things being equal, we would expect to see a decrease in sound recording royalty rates. . . . ‘‘All other things being equal’’ (ceteris paribus in Latin), is a central principle for economic modelling. This economic analysis of bargaining highlights an important relationship between two content cost variables. However, that relationship does not exist in a vacuum. Many other variables affect the bargaining situation and, for any given period, the net effect of all of the different variables may be different than the effect of the modeled variable alone. Thus, this economic analysis of bargaining will not assure that a streaming service will not face disruption in the real world for any reason. . . . Economic modeling is supposed to simplify the situation in order to distill useful principles and teachings. Spulber RWRT ¶¶ 26–28 (emphasis added). The Judges agree that the ceteris paribus principle 75 is a fundamental 74 Further, in his remand testimony, Professor Watt points out that Professor Katz made clear in his testimony that he applied the ‘‘all else equal’’ assumption expressly in his own Nash bargaining analysis at the hearing. Watt RWRT ¶ 20 (quoting Katz WRT ¶ 67). 75 The phrase is often translated into English as ‘‘all other things equal.’’ However, that is somewhat ambiguous. Equal to what? Not to other things. Rather, every ‘‘thing’’ (i.e., every other independent variable) whose effects are not being measured E:\FR\FM\10AUR2.SGM Continued 10AUR2 54424 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations principle in economic analysis and modeling. Professor Watt succinctly makes this point, quoting the Nobel laureate economist James Buchanan, for the following proposition: At the heart of any analytical process lies simplification or abstraction, the whole purpose of which is that of making problems scientifically manageable. In the economic system we recognize, of course, that ‘everything depends on everything else,’ and also that ‘everything is always changing’. Watt RWRT ¶ 32 (quoting J. Buchanan, Ceteris paribus: Some Notes on Methodology, 24 So. Econ. J. 259, 259 (1958). However, Professor Watt does not quote another portion of Professor Buchanan’s article that makes a point that looms large in the present proceeding, to wit, the limitations inherent in applying the necessary ceteris paribus condition: Real problems require the construction of models, and the skill of the scientist is reflected in the predictive or explanatory value of the model chosen. We simplify reality to construct these models, but the fundamental truth of interdependence must never be forgotten. . . . [However,] [f]ew, if any, meaningful results may be achieved by using ceteris paribus to eliminate the study of large numbers of variables. If such variables are closely related, they must be studied simultaneously; there is no escape route open. ddrumheller on DSK120RN23PROD with RULES2 Id. at 259–60 (emphasis added); see also A. Rubinstein, Comments on Economic Models, Economics, and Economists: Remarks on Economics Rules by D. Rodrik, 55 J. Econ. Lit.162, 167 (2017) ‘‘[W]hat matters to the empirical relevance of a model is the realism of its critical assumptions’’) (emphasis added).76 This is not to say that Professor Watt was unaware of this caveat. As noted supra, he recognizes the difficulty of extrapolating from a ceteris paribus world to the real world. The present panel of Judges likewise recognizes this. However, the Majority missed this distinction in the Determination when it applied Professor Watt’s correct but ceteris paribus ‘‘insight’’ for a constant remain ‘‘constant,’’ or ‘‘controlled,’’ i.e., ‘‘equal’’ to their measure prior to the change of the independent variable being examined. See W. Nicholson, Microeconomic Theory: Basic Principles and Extensions at 649 (9th ed. 2005) (defining ‘‘ceteris paribus’’ as ‘‘[t]he assumption that all other relevant factors are held constant when examining the influence of one particular variable in an economic model’’). 76 The Judges note now that Professor Watt did not claim that his bargaining model generated any predictions, but rather that it explained the splitting of the Shapley surplus by the sound recording and musical works copyright owners, respectively, and the impact of that split on royalty rates, given the assumptions and the data in his model. VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 real-world relationship between sound recording and musical works royalty rates. Again, not a single economist made this improper analytical leap or proposed an uncapped TCC rate in order to set a TCC ratio across the entire rate term. Indeed, on careful inspection, no economist states in his or her remand testimony that Professor Watt’s bargaining model provides economic support for the uncapped TCC rate prong. With the foregoing testimony in mind, the Judges see particularly relevant several additional points in Professor Watt’s remand rebuttal testimony that pertain to the appropriateness, vel non, of a TCC rate prong. Referring to the application of his bargaining model to the present case, Professor Watt made these crucial statements regarding the lack of a seesaw effect that would generate decreases in sound recording rates when the mechanical rate is increased: [T]he actual effects one would expect to see several years later would be based on the actual data at that time. Moreover, I would expect many other variables to have a larger effect on the bargains than the relatively small changes in the musical works rate. . . . [U]nderstanding actual market outcomes requires understanding these variables. . . . [A]n attempt to capture all aspects of the real world is too complex for a simple statistical exercise involving an econometric regression. There is no obvious data to actually use for some of the independent variables, such as consumer demand equations, costs of entry and exit, a measure of oligopolistic interaction, different timings of different rate bargains, and the actual values of outside options. Watt WRWT ¶¶ 6(iv), 118.77 Although Professor Watt was hardly transparent in disclosing his ceteris paribus assumption in his original testimony, it seems clear that he always understood its presence, and that, when this assumption was relaxed, ‘‘the actual effects . . . several years later would be based on the actual data at that time [and] many other variables [with] a larger effect on the bargains than the relatively small changes in the musical works rate.’’ Id. ¶ 6(iv) (emphasis added). Professor Spulber likewise opined that the absence of an explicit statement of these assumptions in Professor Watt’s 77 In the language of econometrics, Professor Watt describes this problem as the ‘‘almost sure[ ] impossibil[ity] of ‘‘introduce[ing] a control variable for each and every possible aspect that could potentially impinge upon the relationship [that] could easily lead to such a low R2, and/or statistically insignificant key coefficients, as to make the regression meaningless.’’ Id. ¶ 118. PO 00000 Frm 00020 Fmt 4701 Sfmt 4700 testimony was unremarkable and appropriate: [A]ll other things being equal’. . . should be generally read into economic modeling conclusions or predictions, whether or not the words are repeated in each instance. Economists do not typically repeat these words in each place where they apply, since it would lead to constant repetition. Spulber RWRT ¶ 46, n.8. Regardless of whether economists invariably identify the existence of implicit assumptions lurking in each other’s models, Professor Watt overlooked a cardinal rule of communication: Know your audience. Here, his audience is comprised of three Judges, only one of whom is also an economist.78 Failing to appreciate Professor Watt’s implied ceteris paribus assumption, the Majority transformed his limited (albeit important) ‘‘insight’’ regarding the equal split of the Shapley surplus between the two classes of rights holders—and the seesaw effect that would have if the mechanical rate were increased when the split was imposed—into a justification for the imposition of an uncapped TCC rate prong over the five-year rate term. The Majority’s language reveals this point clearly: As to the issue of applying a TCC percentage to a sound recording royalty rate that is artificially high as a result of musical works rates being held artificially low through regulation, the Judges rely on Professor Watt’s insight . . . demonstrated by his bargaining model that sound recording royalty rates in the unregulated market will decline in response to an increase in the compulsory license rate for musical works. See 3/27/17 Tr. 3090 (Watt) (‘‘[T]he reason why the sound recording rate is so very high is because the statutory rate is very low. And if you increase the statutory rate, the bargained sound recording rate will go down.’’) Professor Watt’s bargaining model predicts that the total of musical works and sound recordings royalties would stay ‘‘almost the same’’ in response to an increase in the statutory royalty. Id. at 3091. Determination at 73–74 (emphasis added). Making the point ever so plainly, Professor Watt now expressly acknowledges that his ‘‘ ‘see-saw effect’ was never really a ‘prediction’ ’’ at all! Watt RWRT ¶ 117. Rather, he now cautions the present panel of Judges, that, ‘‘to make the jump from the model to the actual real-world effects, one cannot ignore the words that are omnipresent in all economic modeling, 78 The dissenting Judge (the only economist on the panel) warned that the seesaw effect was rife with assumptions that rendered it too speculative to be relied upon to support the uncapped TCC rate prong. See Dissent at 7–8. E:\FR\FM\10AUR2.SGM 10AUR2 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations ddrumheller on DSK120RN23PROD with RULES2 that predictions about causal relationships are understood to be ‘‘all else equal.’’ Id. ¶ 32. Without the benefit of these caveats regarding an extrapolation of the ‘‘seesaw’’ theory to the real-world, and with absence of an explicit statement of the ceteris paribus assumption, the Majority misapplied his testimony as a basis to adopt a fixed TCC rate, based upon data from a snapshot in time (2016) to cement that rate relationship for the entire five-year period.79 The Majority misapplied Professor Watt’s correct insight from bargaining theory regarding the use of a fixed ratio for the equal division by two ‘‘Must Have’’ input suppliers of the Shapley surplus to set royalty rates in a period, by using that insight incorrectly to establish a fixed ratio of royalty rates over the rate term.80 79 The importance of Professor Watt’s failure to make explicit the ceteris paribus assumption in his WRT is demonstrated by his need to make it explicit in his RWRT. But even now, rather than acknowledge that the Majority missed the point, he claims that the Services’ are wrongly blaming the Majority for failing to understand this assumption: ‘‘The Services’ testimony on this remand seems primarily focused on creating a ‘‘straw man’’ argument . . . accus[ing] the [Majority] of something that the [Majority] did not do—that is, rely on a guarantee of a particular decrease in sound recording royalty rates—and the Services then attack the Board’s determination by claiming that the decrease did not occur.’’ Watt RWRT ¶ 5. As shown supra, however, this is precisely how the Majority interpreted Professor Watt’s ‘‘insight.’’ The Judges understand that, as a matter of tact and tactics, Copyright Owners may be reluctant to acknowledge that the error lies in the combination of their witness’s opaque testimony and the Majority’s lack of understanding of the assumptions economists make. Copyright Owners might prefer to cast the Majority as the victims of the Services’ incorrect accusation. But the plain language of the Determination belies Copyright Owners’ characterization as to how the confusion arose. 80 The forgoing analysis as applied to the uncapped TCC rate needs to be contrasted with the application of Professor Watt’s bargaining model to increase the percent of-revenue rate to 15.1%. That higher rate was set by the Majority after its consideration of the same Shapley approaches, pursuant to the Judges’ combination of inputs from Professor Gans model (his [REDACTED] round recording-to-musical works ratio) and the Shapley Value Model of Professor Marx that adjusted for complementary oligopoly power by establishing a lower total royalty level ([REDACTED]%). But the difference is that the 15.1% revenue rate was set by applying the Shapley results based on actual and projected market data, see Gans WRT ¶ 38, whereas the uniform uncapped TCC rate (26.2%) was based on the ceteris paribus assumption that held constant the actual data regarding the aforementioned independent variables. As explained above though, Professors Watt and Spulber make it clear that the ‘‘insight’’ from bargaining theory did not have implications to allow for a ‘‘prediction’’ of rates in future periods. Thus, when the Majority engaged in its analysis and ‘‘line-drawing’’ to apply the data and market projections relied upon by Dr. Gans’s data, the Majority was operating—to use the D.C. Circuit’s phrase—in its ‘‘wheelhouse,’’ making a finding that withstood appeal. Johnson, supra, 969 F.3d at 385– 86; see also Determination at 69–70 (‘‘Professor VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 Additionally, an examination of the expert economists’ testimony reveals that their facial disagreements vanish once the necessary assumptions are laid bare. Professor Watt and the Services’ three economists all identify the following independent variables that will impact the relative levels of sound recording and musical works rates paid by interactive services: (1) the level of downstream consumer demand; (2) entry costs; (3) exit costs; (4) oligopolistic interaction; (5) the timing of sound recording agreements vis-a`-vis statutory rate setting; and Professor Watt and the three Service economists agree with regard to the relevancy of these six independent variables. Compare Watt RWRT ¶¶ 6(iv), 118 (identifying all five independent variables) with Leonard WDRT ¶ 18 (identifying independent variables 1–4 above); Marx WDRT ¶¶ 4–5, 42; (identifying independent variables 1–5 above); Katz WDRT ¶¶ 127, 134 n.115 (identifying independent variables 4 and 6 above). Accordingly, the remand record shows a consensus as to the lack of modeling of independent variables that would be important to estimate an uncapped TCC royalty ratio that could be utilized by the Judges to lock-in a ratio over the rate term. Indeed, as noted supra, a careful reading of the remand testimony by Copyright Owners’ economists, Professors Watt and Spulber, reveals that neither of them actually testifies that there is sufficient theoretical and empirical evidence to support the uncapped TCC rate prong and the 26.2% TCC rate phased in on that prong. Rather, those two witnesses testify to something far narrower: the alleged correctness of Professor Watt’s ‘‘seesaw’’ theory as demonstrating an equal splitting of the surplus between the two ‘‘Must Have’’ input suppliers, and the effect of that split when all other relevant independent variable are held constant. In this regard, it is noteworthy that none of Copyright Owners’ several economic experts in this proceeding (Dr. Eisenach, Professor Gans, Dr. Rysman, or Professor Watt) ever proposed an uncapped TCC rate prong in any form, let alone within a greater-of formulation. Such a proposal would have been improper, because, as the expert Gans utilized data from projections in a Goldman Sachs analysis to identify the aggregate profits of the record companies and the music publishers, respectively. . . . The Judges also find Professor Gans’s reliance on financial analysts’ projections for the respective industries to be reasonable.’’). PO 00000 Frm 00021 Fmt 4701 Sfmt 4700 54425 testimony described above makes clear, the ceteris paribus assumption, reasonable for modeling purposes to provide insight as to the surplus split, lacks the input of the omitted variables that the experts on both sides find relevant to the application of economic modeling in this proceeding. A further review of Copyright Owners’ economic expert witness testimony on remand— the first time any of them had occasion to weigh-in on the appropriateness of the uncapped TCC prong—reveals that they also have not endorsed the uncapped TCC rate prong as a proper form of rate setting. To be sure, they strongly endorse the insight first described by Professor Watt in his WRT that the Nash surplus would be split essentially evenly between the two suppliers of essential content, given his simplifying assumptions. But such endorsement is hardly the same as endorsement of the uncapped rate prong itself. For these reasons, the Judges find erroneous the Majority’s identification of a fixed relationship between the sound recording and mechanical royalty rates that could serve as a basis for the Majority’s first rationale for yoking the mechanical rate to an uncapped TCC rate prong. (b) The Services Have Not Rebutted Copyright Owners’ Prima Facie Showing That Professor Watt’s Model Demonstrates a More Limited ‘‘Seesaw’’ Effect The foregoing analysis and decision related to the absence of a fixed relationship between the sound recording and mechanical royalty rates. A separate fixed relationship—the one Professor Watt has clarified he was demonstrating all along—is that if the Judges increase the mechanical royalty rate, the Shapley surplus realized by the labels will decrease almost dollar-fordollar with the increase in the mechanical rate. The Services’ economists aver that even this version of the seesaw is defective. According to Professors Katz and Marx, the Nash Bargaining Model constructed by Professor Watt is deficient because it fails to properly characterize the ‘‘disagreement payoff’’ to the sound recording company when it and an interactive service fail to reach an agreement. More particularly, as explained supra, they assert that Professor Watt’s model omits the value of ‘‘outside options’’ available to the sound recording company. This criticism relates to the issue of whether the seesaw effect would occur as posited in Professor Watt’s model. That is, the increase in the sound recording E:\FR\FM\10AUR2.SGM 10AUR2 ddrumheller on DSK120RN23PROD with RULES2 54426 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations company’s ‘‘disagreement payoff’’ (a/k/ a ‘‘threat point’’) would lead to a higher royalty in the Nash bargain between the sound recording company and the interactive service than needed to generate the seesaw effect to offset the higher mechanical royalty rate. As the several experts’ positions in this regard, discussed supra, make clear, however, each side has a different understanding of whether an ‘‘outside option’’ is properly included in the definition and calculation of the ‘‘disagreement payoff.’’ On the one hand, Professors Katz and Marx claim that the existence and value of ‘‘outside options’’ should be included in the ‘‘disagreement payoff.’’ However, they provide no economic authority for that assertion. By contrast, Professor Watt cites to multiple economic game theory publications and authorities for the proposition that the presence and value of ‘‘outside options’’ are not to be included in the ‘‘disagreement payoff’’ contained in a Nash Bargaining Model. See A. Muthoo, Bargaining Theory with Applications at 105 (1999) (‘‘I thus emphasize that the outside option point does not affect the disagreement point.’’); M. Osborne & A. Rubinstein, Bargaining and Markets at 88 (1990) (‘‘it is definitely not appropriate to take as the disagreement point an outside option. . . .’’); K. Binmore, A. Rubinstein & A. Wolinsky, The Nash Bargaining Solution in Economic Modeling, 17 RAND J. Econ. 176, 185 (1986) (‘‘An outside option is defined to be the best alternative that a player can command if he withdraws unilaterally from the bargaining process.’’). According to Professor Watt and these authorities, the reason for excluding ‘‘outside options’’ from the Nash Bargaining Model is fundamental to the nature of the model itself. In the Nash approach, the negotiating parties are bargaining with each other only over the surplus their deal can generate, and they are attempting to agree upon an allocation of that surplus that exists within the bounds of their respective ‘‘disagreement payoffs.’’ Each may have ‘‘inside options,’’ which are alternatives available to them while bargaining is ongoing and they temporarily disagree. See Muthoo, supra, at 137. However, ‘‘outside options’’ are available to a Nash bargaining party only in lieu of continuing the Nash bargaining with the original counterparty if it ‘‘withdraws’’ from the Nash bargaining process. See Binmore et al., supra. Professor Watt characterizes the distinction as follows: [T]he Nash bargaining model [is] designed as [a] self-contained portrayal[ ] of negotiating VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 behavior. . . . Given a surplus to share, the Nash model . . . provide[s] allowance for financial payments that a party is actually receiving, only while negotiations are ongoing, without walking away for another option, and that would cease as a result of the deal, to be factored into modelling as a cost in some situations.’’) . . . [A]n outside option (a potential payoff that is not directly related to a share of the surplus that is being negotiated) . . . comes in as a constraint upon the set of feasible deals that could be struck. . . .’’ Watt RWRT ¶¶ 56, 58.81 The Services never sought to introduce further testimony regarding this important dispute. This is particularly striking because the Services filed a motion to strike certain portions of the CO Reply, or for leave to file supplemental testimony responsive to those itemized portions. The portions the Services identified in their motion did not include Professor Watt’s criticisms as to the inclusion of ‘‘outside options’’ in their experts’ Nash modeling. Further, after the Judges granted the Services’ motion by providing them leave to file supplemental testimony—consistent with the designations in their motion— the supplemental testimonies did not address this ‘‘outside options’’ issue. In the course of discussions among the parties and the Judges regarding remand procedures, the Judges invited the parties to produce witnesses for a hearing, at which one or more of the Services’ economic expert witnesses could have addressed this ‘‘outside options’’ issue. However, the Services (and Copyright Owners) waived the opportunity to produce witnesses at a hearing. Rather, they offered, and the Judges agreed, that they would stand on their written testimonies and proceed to closing arguments by counsel. In the closing arguments, each side argued numerous points of controversy and provided the Judges with dozens of demonstrative aids summarizing record evidence and the parties’ arguments, but none of those arguments or demonstrative aids so much as mentioned this ‘‘outside options’’ dispute. Moreover, when the Judges inquired during closing arguments as to whether Services’ counsel would be addressing any of the experts’ ‘‘modeling disputes,’’ counsel said that they were resting on their papers. 3/8/ 22 Tr. 86–87 (Closing Argument). Similarly, when the Judges inquired of Copyright Owners’ counsel whether he 81 Professor Marx in fact cites several of these authorities (for other points), without noting the distinction they make between the appropriate inclusion of ‘‘inside options’’ and exclusion of ‘‘outside options’’ in Nash modeling. See id. ¶ 59. PO 00000 Frm 00022 Fmt 4701 Sfmt 4700 would be addressing the modeling ‘‘dust-up’’ between Professors Watt and Katz, counsel demurred, stating that although he would ‘‘love to engage on it but . . . ‘‘there would be too many slides. . . .’’ Id. at 262–64. Simply put, the Services’ economic experts made an assertion regarding the need for Professor Watt to have included ‘‘outside options’’ in his Nash Bargaining Model, but Professor Watt presented authority clearly stating that such inclusions would be improper. Thus, Copyright Owners made a prima facie showing that in a Nash Bargaining Model, the surplus generated by the streaming surpluses acquired by the content providers would be split equally as between the sound recording licensors and musical works licensors, and that, ceteris paribus, an increase in the mechanical rate to provide Copyright Owners more of the surplus (per the Shapley-based results relied on by the Majority) would be essentially offset through a nearly 1:1 reduction in the sound recording rate. In response to Copyright Owners’ prima facie case, the Services stood mute in response to the rebuttal argument claiming that their experts misapprehended the Nash modeling distinctions between ‘‘inside options’’ and ‘‘outside options.’’ 82 Accordingly, the Judges find that the Services’ criticisms in this regard are insufficient to rebut Copyright Owners’ prima facie showing that Professor Watt’s Nash Bargaining Model properly 82 The third economic expert for the Services, Dr. Leonard, did not utilize the ‘‘outside option’’ phraseology to describe his critiques. Rather, he first criticized Professor Watt for assuming the existence of a ‘‘fixed surplus.’’ Leonard WDRT ¶ 16. However, as discussed supra, that assumption came from the Majority’s extrapolation from Professor Watt’s hearing testimony. His explicit statement regarding the ceteris paribus assumption makes clear that he was not assuming a ‘‘fixed surplus.’’ Watt RWRT ¶¶ 110–11. (Again, the only ‘‘fixed’’ surplus was not ‘‘assumed,’’ but rather quantified, in order to establish the Majority’s percent-ofrevenue prong royalty rate of 15.1%.) Dr. Leonard next claims that Professor Watt’s assumption that the labels would bear virtually the entirety of an increase in the statutory rate, because they previously ‘‘have captured almost all’’ [the] surplus,’’ has been contradicted by the evidence. Specifically, he refers to the 33-month period in which the Phonorecords III rates were effective (January 2018 through September 2020). Leonard WDRT ¶ 16. However, as the Judges find in this Determination, that 33-month period was marked by significant uncertainty with regard to the ultimate rates and rate structure (and the rates were being phased-in), so no findings could reliably be made based on sound recording rate changes during that period. The remainder of Dr. Leonard’s critique concerns issues that would make a fixed TCC ratio inappropriate over the rate term. The Judges agree with those criticisms as previously discussed, but they do not pertain to this narrower issue of whether the surplus generated by interactive streaming would be split in a manner consistent with Professor Watt’s Nash Bargaining Model. E:\FR\FM\10AUR2.SGM 10AUR2 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations identified and valued the ‘‘disagreement payoff.’’ 83 84 ddrumheller on DSK120RN23PROD with RULES2 b. Rejection of Second Rationale for Including Uncapped TCC Rate Prong In the Determination, as noted supra, the Majority also justified the adoption of the uncapped TCC rate prong because it had the effect of ‘‘import[ing] into the rate structure the protections that record companies have negotiated with services to avoid the undue diminution of revenue through the practice of revenue deferral.’’ Determination at 36; see also Johnson, 369 F.3d at 372 (‘‘By pegging the mechanical license royalties 83 To be clear, the Judges’ ruling is narrow; they make no finding beyond crediting this prima facie showing and the failure of the Services to rebut sufficiently that showing. It might be the case that the existence and definition of ‘‘outside options’’— and their relationship to ‘‘inside options’’—have other implications vis-a-vis a Nash Bargaining Model applied in the context of a rate setting proceeding. However, the Judges may not introduce and rely on analytical approaches not developed by the parties. See Johnson, 969 F.3d at 381 (the Judges must not ‘‘procedurally blindside[ ]’’ the parties with an ‘‘approach . . . first presented in the determination and not advanced by any participant.’’). See generally P. Wald, Limits on the Use of Economic Analysis in Judicial Decisionmaking, 50 J. L. & Contemporary Problems 225, 228 (1987) (’’ judicial analysis, economic or otherwise, takes place only in the context of lawsuits between two or more parties imposes a practical constraint on the judge’s ability to use economic analysis.’’). 84 Professor Katz also criticizes Professor Watt’s assumption that ‘‘a label’s non-content costs are proportional to licensing revenues.’’ Katz WDRT ¶ 22. More particularly, Professor Katz claims that this is not ‘‘plausible’’ because ‘‘the royalty rate does not directly affect the sound recording copyright owners’ non-content cost.’’ Id. ¶ 133. The effect of eliminating this assumption, according to Professor Katz, is to reduce the seesaw effect in Professor Watt’s model of [REDACTED] slightly further away from a 1:1 ratio, to .92. Id. In rebuttal, Professor Watt says this criticism is inconsistent with Professor Katz’s own analysis, because the latter also ‘‘sets the cost equal to a fraction of revenue. . . .’’ Watt ¶ 82 n.31 (referring apparently to a comparison of Katz WDRT ¶ 129 with id. ¶ 133). Professor Watt concludes that not only does ‘‘[Professor] Katz’s own model contain the same feature that he is critical of in my model,’’ it is also ‘‘not a flaw in the bargaining model.’’ Watt ¶ 82. As a substantive matter, Professor Watt defends the assumption that non-content costs would rise with royalty income, because ‘‘[g]reater revenue should be directly equated with a larger scale of business’’ and ‘‘the additional royalty income would have to be managed (i.e., distributed to those who need to be paid from it, such as artists), implying higher administration costs.’’ Id. ¶ 79. The Judges find that the common use by both experts of this assumed proportionality of a label’s non-content costs to licensing revenues alone blunts Professor Katz’s criticism of Professor Watt’s modeling. Further, Professor Watt reasonably posits that higher revenue would imply a larger scale of business with associated general cost increases. (But the Judges do no agree that it was reasonable for Professor Watt to assume that distribution and administrative costs in particular would increase merely because of an increase in royalty rates; simply paying more money, ceteris paribus, is not self-evidently associated with an increase in costs.) VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 to an uncapped total content cost prong, the Board sought to ensure that owners of musical works copyrights were neither undercompensated relative to sound recording rightsholders, nor harmed by the interactive streaming services’ revenue deferral strategies. . . .’’) (emphasis added). (i) Parties’ More Specific Arguments Copyright Owners likewise argue that the uncapped TCC rate structure should be ‘‘adopted to provide protection against revenue deferment and displacement in a revenue-based rate structure.’’ CO Initial Submission at 38; see also id. at 40 (describing uncapped TCC rate prong as ‘‘critical backstop in a revenue-based rate structure.’’). Whereas Copyright Owners echo the Majority, the Services adopt the reasoning of the Dissent. They argue as follows: [A] rate structure with a capped TCC prong, like the Phonorecords II settlement, achieves the same goal of protecting the Copyright Owners from any potential revenue deferral through a ‘‘structure that provides alternate rate prongs and floors, below which the royalty revenue cannot fall,’’ . . . and does so without allowing Copyright Owners to impermissibly share in the labels’ complementary oligopoly power. . . . [T]he streaming industry has twice concluded, after extensive negotiations, that the appropriate way to address any concerns regarding revenue deferral is to have a rate structure that includes a capped TCC prong. Phono I, 74 FR 4510; Phono II, 78 FR 67938. Services’ Joint Opening Brief at 62 (quoting Dissent, 84 FR 1990) (emphasis added). In their Reply, Copyright Owners argue that the Majority maintained the benefits of price discrimination contained in the prior Phonorecords II framework, but balanced that goal with added protection against Service revenue deferral and displacement. Copyright Owners’ Reply Brief on Remand at 49 (‘‘In adopting a rate structure with [an uncapped] TCC for all service offerings, the [Majority] balanced its concerns about fostering price discrimination while also protecting against proven revenue diminution by the Services.’’). The Services, in their Reply, take note that pre-remand, Copyright Owners had strenuously objected to any yoking of the mechanical royalty rate to the sound recording rate, maintaining that, although the Copyright Owners now advocate for an uncapped TCC rate to protect against revenue displacement and diminution: [I]n their [pre-remand] reply proposed findings, the Copyright Owners had expressed a very different view, arguing that PO 00000 Frm 00023 Fmt 4701 Sfmt 4700 54427 an uncapped TCC prong ‘‘does nothing to protect Copyright Owners from the Services’ revenue displacement and deferment’’ [and] Copyright Owners have not even tried to explain away their complete about-face on this issue. Services’ Reply at 43. (ii) Analysis and Decision Regarding Revenue Diminution or Deferral The Judges find that the second rationale put forth to support an uncapped TCC rate does not justify the adoption of that rate prong. Several reasons support this finding. First, there is insufficient evidence to show how the sound recording companies contractually structure their own royalty rates, which would constitute the rate base for an uncapped TCC rate for the mechanical royalty. The sound recording royalty rate, when proffered for use as a mechanical royalty rate base, is analogous to pegging the value of a foreign currency to the U.S. dollar. That is no mere benchmark. The Judges must have the benefit of sufficient record evidence to demonstrate that the pegging (or, to use the D.C. Circuit’s word in Johnson, ‘‘yoking’’) of a statutory rate to an unregulated rate serves the statutory purposes for the rate at issue, here, the mechanical rate. But Copyright Owners presented virtually no evidence regarding how the sound recording companies structure their interactive service royalties. Indeed, in the hearing, Dr. Eisenach acknowledged that the ‘‘relative value of sound recording [to] musical works licenses may depend on a variety of factors,’’ but he intentionally eschewed unnecessary ‘‘assumptions, complexities and uncertainties associated with theoretical debates’’ as to why the particular market ratios existed. See Determination at 44. Indeed, the Majority found fault with Dr. Eisenach’s willful ignoring of these issues, agreeing with the Services’ criticism that Dr. Eisenach’s ‘‘use of sound recording royalties paid by interactive services embeds within his analysis the inefficiently high rates that arise in that unregulated market through the complementary oligopoly structure of the sound recording industry and the Cournot Complements inefficiencies that arise in such a market. See Determination at 47. The uncapped TCC rate advocated now by Copyright Owners suffers from the same affliction. The only reference to such sound recording rate formulae in Copyright Owners’ voluminous PFF after the hearing was its statement that the effective revenue calculations in two of the Major labels’ agreements with the E:\FR\FM\10AUR2.SGM 10AUR2 ddrumheller on DSK120RN23PROD with RULES2 54428 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations services was based on [REDACTED]. See Copyright Owners’ PFF ¶¶ 72, 91 (cited post-remand at Copyright Owners’ Motion for Reconsideration or Clarification at 25, n.14). On remand, the Services have provided a further summary of the types of [REDACTED]. See White WDRT ¶¶ 6–7, 14–15, 20, 24– 26, 28–29 ([REDACTED]); Bonavia WDRT ¶¶ 15–17 ([REDACTED]); Mirchandani WDRT ¶¶ 16, 21–24 ([REDACTED]). Clearly, the levels of [REDACTED] would have to be weighed and the impact of complementary oligopoly power would need to be identified in order to adjust the rate prongs to account for that power. But the record is devoid of such details. Second, compounding this problem, because the uncapped TCC rate is embedded in a ‘‘greater-of’’ rate structure, the labels can exploit their complementary oligopoly power when creating the switching points that toggle royalty payments between and among rate prongs. As the Judges have explained previously, in declining to import a ‘‘greater of’’ structure from the unregulated interactive market, this structure[it] is based on ‘‘agreements [which] were all negotiated in a market characterized by the lack of effective competition, and that the lack of competition would affect the structure as well as the level of rates.’’ SDARS III, 83 FR 65210, 65228 (Dec. 19, 2018) (emphasis added). Further, the Judges held therein that the ‘‘advantageous’’ nature of a ‘‘greater-of’’ structure to sound recording licensors ‘‘may well represent an example of what licensors can and would obtain when they exploit their ‘‘must have’’ status for a special competitive advantage.’’ Id.; see also Dissent at 47 (in absence of testimony explaining how greater-of structure is consonant with effective competition, use by licensor suggests a game of ‘‘heads I win tails you lose.).’’ Thus, there is insufficient evidence or testimony that would permit the Judges to make any adjustment for the complementary oligopoly power that may be built into each prong of the sound recording royalty rate structures. Third, as the Services note, Copyright Owners pre-remand, opposed the identical rate structure—consisting of a percent-of-revenue prong and an uncapped TCC prong—before Copyright Owners were in favor of it, postremand.85 Although Copyright Owners 85 When Copyright Owners opposed the concept of an uncapped TCC rate prong in a greater-of structure, the proposed uncapped TCC rate was Google’s 15% (and its proposed percent-of-revenue rate was 10.5%). Determination at 13. But after the Majority set the uncapped TCC rate at 26.2%—a 75% increase over the 15% TCC rate—Copyright VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 took a 180-degree turn on this issue, they never stated they were wrong to oppose it previously. Indeed, the Dissent relied upon Copyright Owners’ strenuous objection to an uncapped TCC rate, quoting it verbatim: Copyright Owners rightly note that they obtain no legal protection under such a TCC prong. In making this argument regarding displacement and deferral of revenue, Copyright Owners lay out comprehensively all the problems inherent in an uncapped TCC prong set in a greater of rate structure, such as adopted in the majority opinion: The notion that [the] TCC prong will provide protection from revenue gaming, deferral and displacement, and other revenue prong problems is unsupported and speculative. Relying on just the TCC to solve those admitted problems leaves the Copyright Owners’ protection from such problems entirely outside the statute. . . . the per-user rates in the label deals are what protects the Copyright Owners from price-slashing by the services. What is left unanswered . . .is . . . how can it be reasonable to ask the Judges to set a rate that does not itself provide for a fair return . . . but simply puts the Copyright Owners’ fair return in the hands of the labels to negotiate terms that will adequately protect the publishers and songwriters as well? The labels do not have a mandate to ensure that the Services provide a fair return to the Copyright Owners, and cannot be directed to ensure such. Indeed, labels may not have the same incentives as songwriters and publishers to negotiate such protections in their deals. To wit, a label could make an agreement with a service that includes only a revenue prong in exchange for equity or some other consideration that it may never include in the applicable revenue subject to the TCC. . . . [W]hat if Google purchased one or more record labels and did not have to pay any label royalties? Or what if Spotify chose to avail itself of the compulsory license to create its own master recordings embodying musical works—which it is already doing . . . and chose to compensate itself for its use of the master recordings on a sweetheart basis (or not at all)? Or what if one or more labels decided to enter the interactive streaming market and did not have to pay themselves royalties? In each case, the Copyright Owners’ protection—the protection that the Services admit the Copyright Owners need and is provided by the TCC—would be gone. Dissent at 5–6 (quoting Copyright Owners’ RPFF-Google at 39–41) (emphasis added). To make the identical point post-remand, but from the Services’ perspective, Pandora’s economic expert witness, Professor Katz, simply utilizes Copyright Owners’ verbatim language (bolded above), but substitutes the word ‘‘Services’’ for ‘‘Copyright Owners’’ (and ‘‘income’’ for ‘‘return’’) to highlight how reliance on Owners became zealous converts to the concept of an uncapped TCC rate proper. PO 00000 Frm 00024 Fmt 4701 Sfmt 4700 the sound recording royalty rate is improper: What is left unanswered . . . is . . . how can it be reasonable to ask the Judges to set a rate that does not itself provide for a fair income . . . but simply puts the Services’ fair income in the hands of the labels to negotiate terms that will adequately protect the Services as well? The labels do not have a mandate to ensure that the Copyright Owners provide a fair income to the Services, and cannot be directed to ensure such. Katz WDRT ¶ 71. The Judges find this argument persuasive, both in its own right and in the fact that it has been advanced by Copyright Owners and the Services alike.86 Fourth, the Judges note that the Majority did not find that revenue diminution, via displacement, deferral, or otherwise was pervasive, as Copyright Owners aver. Compare CO Initial Submission at 40 (‘‘The record overwhelmingly established that the percent of revenue prong often results in musical works royalties that are too low . . . drive[n] [by] . . . . revenue deferral [and] revenue displacement’’) with Determination at 21 (‘‘The Judges agree that there is no support for any sweeping inference that cross-selling has diminished the revenue base.’’) (emphasis added) and 36 (‘‘The Judges find that the present record indicates that the Services do seek to engage to some extent in revenue deferral in order to promote their long-term growth strategy.’’) (emphasis added). Given that the Majority found revenue diminution through displacement and/ or deferral exists only ‘‘to some extent’’ and is not a ‘‘sweeping’’ issue, the Judges on remand find that the uncapped TCC rate structure creates the potential for unbalanced harm. As noted supra, the only protection against runaway mechanical rates, the seesaw hypothesis, cannot justify yoking the mechanical rate to a fixed ratio with the 86 At Closing Arguments on remand, Judge Strickler queried counsel for Copyright Owners regarding their prior rejection of an uncapped TCC prong within a ‘‘greater-of’’ rate structure. Counsel’s response was that an uncapped TCC doesn’t provide enough protection against revenue diminution: ‘‘It provides more than the Phonorecords II rates, but not as much as we want,’’ although ‘‘still better than’’ the negotiated Phonorecords II approach. 3/8/22 Tr. 240– 41(Closing Argument). But Copyright Owners have neither distinguished nor disavowed their persuasive legal point quoted in the text above, to wit that an uncapped TCC rate would be unreasonable if the ‘‘protection’’ it affords lies ‘‘entirely outside the statute.’’ Whether the ‘‘protection’’ relates to Copyright Owners’ concern over revenue diminution or to the Services’ concern over uncapped mechanical rates, the legal defect is the same—the unreasonableness of leaving the purported protection ‘‘entirely outside the statute.’’ E:\FR\FM\10AUR2.SGM 10AUR2 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations unregulated sound recording rate.87 By contrast, and as discussed infra, the Phonorecords II-based benchmark approach, despite its own imperfections, is superior in this regard, because its series of alternate rate prongs and floors represents a negotiated compromise (negotiated by trade associations with countervailing power) between the potential for revenue diminution that would harm Copyright Owners, on the one hand, and the potential for runaway mechanical rates (yoked to the sound recording companies’ complementary oligopoly power) that would injure the Services, on the other. ddrumheller on DSK120RN23PROD with RULES2 (iii) Distinction Between the ‘‘Reasonable’’ Rate Statutory Standard and the Factor (D) Objective To Minimize ‘‘Disruptive Impact’’ The Judges next consider an issue emphasized by Copyright Owners: whether the Services have demonstrated that the uncapped TCC rate prong would cause a ‘‘disruptive impact’’ as set forth in Factor (D) of section 801(b)(1).88 87 Even Google, the party that, post-hearing, broached in its PFF the idea of an uncapped TCC prong, candidly identified the risk arising from an uncapped TCC: ‘‘Having no cap on TCC . . . leaves the services exposed to the labels’ market power, and would warrant close watching if adopted. . . .’’ Google PFF ¶ 73 (emphasis added). But as the Dissent noted, there is no satisfactory way to monitor an uncapped TCC rate prong: ‘‘Who would do the ‘‘watching’’? When would such watching occur? Congress directed the Judges to be the ‘‘watchers,’’ and Congress instructed that the ‘‘watching’’ should occur only through rate proceedings. . . .’’ Dissent at 4 (emphasis in original). 88 Separate and apart from the ‘‘disruptive impact’’ argument made by Copyright Owners, there is no need to consider how this prong would relate to Factor D, because the Judges find the uncapped TCC rate prong with the (phased-in) 26.2% rate to be ‘‘unreasonable.’’ If it were necessary to separately consider the four itemized factors, the Judges would confirm that Factor A is satisfied, because, as the D.C. Circuit found, the Majority reasonably found that rates should increase from the Phonorecords II period, and the 15.1% revenue rate represents a 44% increase. The Judges would also find Factors B and C to be satisfied without a separate uncapped TCC rate prong. The reason is that, under the section 801(b)(1) standard, the ‘‘reasonableness’’ standard filters out more statutorily infirm rates than the fairness objectives. By contrast, when a rate does satisfy the ‘‘reasonableness’’ standards under section 801(b)(1), the Judges must also consider the rate through the finer ‘‘fairness’’ filter. Cf. Determination at 68 & n.120 (distinguishing between: (1) a Shapley Value analysis that filters out unreasonable rates by reducing licensors’ ability to abuse market power by threatening or exercising their refusal to license (‘‘hold-out or ‘‘hold-up’’ power); and (2) a Shapley Value analysis that further filters out unfair rates by going beyond eliminating abuse of market power to also make a ‘‘market power adjustment’’ explicitly to address Factors B and C). Finally, as the text infra, explains, the Judges also find no basis under Factor D to alter their analysis. VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 Section 801(b)(1) provides that one of the competing priorities of the Judges in setting the mechanical rate is ‘‘[t]o minimize any disruptive impact on the structure of the industries involved and on generally prevailing industry practices.’’ 17 U.S.C. 801(b)(1)(D). In Johnson, the D.C. Circuit did not identify any argument by the Services that was predicated on a claim that this statutory form of ‘‘disruption’’ had occurred, or was likely to occur, as a consequence of the Majority’s rates and rate structure. Additionally, the D.C. Circuit did not ground its decision to vacate and remand the Judges’ uncapped TCC rate and rate structure rulings based on the potential that these rulings would be disruptive to the Services, let alone would cause a statutory ‘‘disruptive impact.’’ After the D.C. Circuit’s ruling, an argument regarding ‘‘disruption’’ was first made by Copyright Owners, not the Services. Copyright Owners argued that the vacated rates should nonetheless be maintained as interim rates, during the pendency of the remand proceeding. Motion of Copyright Owners to Adopt Interim Rates and Terms Pending the Remand Determination, passim (Nov. 2, 2020). Copyright Owners argued that reverting to the rates that existed before the Determination would constitute a ‘‘disruption’’ and self-servingly predicted that the Services would attempt to argue that the uncapped TCC rate and rate structure were themselves ‘‘disruptive.’’ Copyright Owners opined that such an argument would be a ‘‘hollow exercise.’’ Id. at 12, n.5; see id. at 2–3, 9 (claiming absence of disruption from uncapped TCC rate and structure despite absence of such argument by Services). In response to that motion, the Services did not assert that the Majority’s uncapped TCC rates and rate structure would constitute disruption or have disruptive impact, whether under statutory Factor D or otherwise. See Services’ Opposition to the National Music Publishers’ Association (NMPA) and Nashville Songwriters Association International’s (NSAI) ‘‘Interim Rates Motion’’ (Nov. 18, 2020). In reply, Copyright Owners shifted from anticipating a ‘‘disruption’’ argument to misinterpreting Johnson, asserting, without citation: ‘‘With respect to the TCC prong, the remand directs only that services be given opportunity to offer evidence of disruption from rates that have now been in effect for three years without any disruption.’’ Copyright Owners’ Reply in Support of Motion to Adopt Interim Rates at 7–8 (Nov. 25, 2020) (emphasis added). PO 00000 Frm 00025 Fmt 4701 Sfmt 4700 54429 On December 10, 2020, the Services submitted to the Judges their Proposal for Remand Proceedings, in which they made no argument that the uncapped TCC rates and rate structure (or, for that matter, any aspect of the Determination) would cause disruption or have a disruptive impact, whether under statutory Factor D or otherwise. By contrast, in their remand proposal, Copyright Owners reference twelve times that, for the Judges to reject the uncapped TCC rates and structure, the Services must show the presence of ‘‘disruption’’ arising from the Majority’s uncapped TCC rates and structure. Copyright Owners made this argument notwithstanding that the ‘‘reasonable’’ rate standard is separate from the ‘‘disruptive impact’’ issue, which is an itemized objective (one of four) to be considered as an adjustment to what would otherwise constitute a ‘‘reasonable’’ rate. See Proposal of Copyright Owners for the Conduct and Schedule of the Resolution of the Remand at 2, 7–8, 22–24 (Dec. 10, 2020).89 In the CO Initial Submission, Copyright Owners assert, without citation to any of the Services’ filings: ‘‘The Services contend that, had they been given such an opportunity [at the hearing], they supposedly could have established that an ‘‘uncapped’’ TCC is disruptive because the market for sound recordings is not effectively competitive.’’ Id. at 5. Copyright Owners further aver that the Services must ‘‘provide evidence, consistent with the [CRB Judges’] well-established disruption standard, that because of the labels’ supposed market power, the TCC structure adopted by the Board has actually, substantially, immediately and irreversibly threatened the continued viability of the interactive streaming industry’’ in a manner that will ‘‘threaten the viability of the music delivery service currently offered to consumes under [the] license.’’ Id. at 7, 56 (citations omitted). Copyright Owners then assert that the Services bear the burden of proving disruption under Factor D from the 89 When Copyright Owners do address an argument that the Services actually made (on appeal) regarding the uncapped TCC rates and structure, they note not that the Services had made a ‘‘disruption’’ argument, but rather that ‘‘the Services appealed for the reversal of the TCC prong as substantively unreasonable.’’ Id. at 22 (emphasis added). But Copyright Owners then assert, coyly, that ‘‘this request was not granted by the Circuit’’ (citing Johnson, 969 F.3d at 383), when in actuality, the D.C. Circuit did not rule against the Services on this point, but rather stated only that it was not addressing substantive arguments made by the Services ‘‘[b]ecause we have vacated the rate structure devised by the [Judges] for lack of notice. . . .’’ Id. E:\FR\FM\10AUR2.SGM 10AUR2 54430 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations ddrumheller on DSK120RN23PROD with RULES2 uncapped rates and rate structure embodied within the rate proposal (even though only Copyright Owners are pursuing this approach on remand). Further, Copyright Owners assert that the Services’ objection to the uncapped rates and rate structure must fail unless they can show that such a disruptive impact occurred during the 33-month period (from January 2018 through September 2020) when the Phonorecords III rates were in effect. Id. at 56. In their initial substantive remand briefing, the Services once more did not assert that the Determination’s uncapped TCC rates and structure would cause disruption pursuant to Factor D of section 801(b)(1), or even assert a non-statutory disruption arising therefrom. Rather, the Services directly attacked this rate approach as inconsistent with the statutory ‘‘reasonable’’ rate requirement, maintaining that ‘‘[t]ying the mechanical rates directly to the complementary oligopoly sound recording rates in the manner of the Majority’s uncapped TCC rates and rate structure is plainly unreasonable.’’ Services’ Joint Opening Brief at 46 (Apr. 1, 2021) (emphasis added). The Services also asserted that the uncapped TCC rates and rate structure are ‘‘unreasonable’’ because they do not promote the statutory objectives of Factor B (‘‘fair income’’ to the copyright user) and Factor C (reflecting the copyright users’ itemized role in making the musical works ‘‘available to the public.’’). Id. at 45, 50–51, 55.90 In the Services’ Reply, the Services attack Copyright Owners’ ‘‘singular focus on the disruptive impact of the 90 The Services’ only references to the concept of ‘‘disruption’’ relate to their argument that their own benchmark premised on the prior Phonorecords II rate structure and rates would not be disruptive. Id. at 4, 24, 29–30. That argument is properly made by Services in this context, because a party seeking to persuade the Judges to adopt its proposal bears the burden of proof, pursuant to section 556(d) of the APA, regarding the consonance of its proposal with all the standards contained in section 801(b)(1). The Judges do note that one of the Services’ expert witnesses, Professor Katz, found the Majority’s attempt to avoid disruption by phasing-in the new rate provisions insufficient ‘‘to mitigate the risk of short-term market disruption’’. That testimony does not constitute a direct reliance by the Services on the statutory disruption objective in Factor D, but rather emphasizes the Majority’s own concern with such disruption and the witness’s concern that the phase-in did not prevent the disruptive effect that the Majority itself had contemplated. In any event, Professor Katz, as an economist, cannot make a legal argument regarding the applicability of the Factor D objective, the Services did not rely on his testimony in that regard and, as noted, the Services made no legal Factor D ‘‘disruption’’ argument on remand. Thus, the Judges do not give any weight to Professor Katz’s testimony in this regard. VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 uncapped TCC prong.’’ Services’ Reply at 35. In particular, the Services argue: 1. they have maintained and demonstrated that Copyright Owners’ uncapped rates and rate structure are ‘‘unreasonable,’’ separate and apart from demonstrating that this uncapped approach also fails to satisfy the four itemized statutory factors; 2. the burden of proof with regard to Factor D disruption lies with Copyright Owners, because they are the ones who are advocating for the uncapped TCC rates and rate structure; 3. the presence of Factor D disruption, vel non, is not dispositive, because section 801(b)(1) and Johnson require the Judges to apply the entirety of the statutory standard (which consists of the ‘‘reasonable rate’’ requirement and consideration of all four itemized Factors; and 4. the ‘‘full extent of the disruption to the Services from an uncapped TCC prong was never tested in the marketplace [because] [t]he Majority set escalating rates, and the [ ] Determination was vacated before the significant hike in rate levels was fully implemented.’’ Id. at 35–36. In their Remand Reply, with regard to the issue of ‘‘disruption,’’ Copyright Owners assert: 1. The Services have ‘‘completely abandoned’’ their appellate argument asserting disruption, and admit to having no evidence that the Board’s adopted rate structure has any materially disruptive impact. Copyright Owners’ Reply Brief on Remand at 5 (July 2, 2021). 2. The Services have not even attempted to show any Factor D related effect or other disruption from the adopted rates and structure. Id. at 15, n.9. 3. The failure of the Services to provide evidence of disruption or to pursue the argument that disruption had occurred was inconsistent with their prior assertions that the uncapped TCC rates and rate structure created ‘‘a real risk of economic harm’’ and the ‘‘impact’’ or ‘‘harm’ that the uncapped approach generated. Id. at 35. 4. Each of the Services, in response to Copyright Owners’ discovery requests, acknowledges that it was not offering new evidence regarding the ‘‘impact’’ of the Phonorecords III rates and rate structure. Id. at 36–38. 5. The Services did not merely suffer no disruption, they experienced unprecedented growth and profit under PO 00000 Frm 00026 Fmt 4701 Sfmt 4700 the uncapped TCC rate prong. Id. at 45.91 6. The Services on remand have attempted to replace their prior ‘‘disruption’’ assertion with a claim of ‘‘unreasonableness.’’ Id. at 50, n.36. (iv) Analysis and Decision Regarding ‘‘Disruption’’ Issue The full Factor D ‘‘disruption’’ standard, as set forth by the Judges, states that an adjustment is warranted by Factor D if the rate analysis made by the Judges would otherwise: directly produce[ ] an adverse impact that is substantial, immediate and in the short-run because there is insufficient time for either [party] to adequately adapt to the changed circumstance produced by the rate change and, as a consequence, such adverse impacts threaten the viability of the music delivery service currently offered to consumers under this license. Determination at 87. Factor D is not applicable, particularly as proposed by Copyright Owners. Thus, the Judges reject Copyright Owners’ assertion that the uncapped TCC prong should be adopted because of the absence of evidence of ‘‘disruptive impact’’ proffered by the Services. This rejection is based on several findings of fact and conclusions of law. First, the issue of ‘‘disruptive impact’’ pertains here to the proposal advanced by Copyright Owners, not the Services. Thus, the burden of proving that this uncapped TCC rate prong proposal satisfies the elements, including Factor D, of the section 801(b)(1) standard in a sufficient manner lies with Copyright Owners, not the Services. See 5 U.S.C. 556(d). Accordingly, the fact that the Services did not affirmatively assert an argument of ‘‘disruptive impact’’ is of no consequence. Moreover, as the review of the Services’ filing makes clear, the Services never abandoned that argument, because they never made it. 91 The Judges allowed the Services to make a supplemental filing in response to Copyright Owners’ remand reply, because those papers contained direct as well as reply materials. In their supplemental filing, the Services argued that they had not ‘‘thrived,’’ that the financial data on which Copyright Owners’ relied did not isolate revenue attributable to interactive services, was not limited to U.S. generated revenue, and used changes in the market capitalization of Amazon and Alphabet (Google’s parent corporation) as a proxy for the economic fortunes of their interactive services. Services’ Joint Supplemental Brief at 13–15. As explained supra, the Judges find the permanency of the Phonorecords III rate structure during the 33month period from January 2018 through September 2020 to have been in question, pending the appeal that resulted in the vacating and remanding of the Determination and the reversion back to the Phonorecords II rates and rate structure. Given that uncertainty, the Judges find it wholly inappropriate to draw any conclusions from the change or stasis in the sound recording rates or the total royalty payments by a Service over that period. E:\FR\FM\10AUR2.SGM 10AUR2 ddrumheller on DSK120RN23PROD with RULES2 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations Rather, they have consistently argued that the uncapped TCC rate prong was unreasonable, not that it was statutorily ‘‘disruptive’’ as that standard has been applied by the Judges. Second, Copyright Owners did not demonstrate with sufficient evidence or testimony that the uncapped TCC rate would be consistent with Factor D. To be clear, by this the Judges do not mean that Copyright Owners were obliged to prove a negative. Rather, they needed to prove, and indeed attempted to do so, that it was unlikely that their rates would cause a ‘‘disruptive impact.’’ In this regard, as an empirical matter, Copyright Owners proffered the testimony of an economic expert witness, Dr. Eisenach, who opined that the Services’ [REDACTED]. Eisenach WRT ¶¶ 12–41 ([REDACTED]) CO Reply at 40–41. However, as the Judges discuss supra, that period reflected ‘‘33 months of uncertainty,’’ during which no one could predict the final mechanical rate and structure that would be adopted by the Judges and/or the D.C. Circuit after appeals. Accordingly, that factual evidence is unpersuasive. Further, as a theoretical matter, Copyright Owners rely on Professor Watt’s testimony regarding the ‘‘seesaw’’ effect. In that regard, and as discussed supra, the Majority took comfort in what it understood to be Professor Watt’s ‘‘prediction’’ that increases in mechanical royalties would be offset almost dollar-for-dollar by reductions in the sound recording royalty. However, as also discussed supra, Professor Watt has now clarified on remand that he never made such a ‘‘prediction,’’ and that his testimony regarding the socalled ‘‘seesaw’’ was limited to shifts in the share of the surplus to Copyright Owners and from sound recording companies as a consequence of an increase in the mechanical rate, holding all other factors unchanged (the ceteris paribus assumption). Moreover, Professor Watt further explained that many other factors would likely impact the sound recording rate together with an increase in the mechanical rate, including ‘‘a measure of oligopolistic interaction, different timings of different rate bargains, and the actual values of outside options.’’ Watt RWRT ¶ 118. Professor Watt candidly acknowledged that he has not modeled these independent variables, and he further notes that the data may not exist to allow for such modeling. Id. But the inability to model the impact of independent variables does not mean that their potential to cause disruption can be ignored. VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 In particular, the purpose of the ‘‘seesaw’’ contention was that it prevented economic harm to the Services in connection with a rise in the mechanical rate. Although not of Professor Watt’s design, that connection is intentionally built into the Majority’s uncapped TCC rate. See Determination at 35 (‘‘Incorporating an uncapped TCC metric into the rate structure permits the Judges to influence that ratio directly.’’) But the ‘‘measure of oligopolistic interaction’’ referenced by Professor Watt was the very concern expressed by the Dissent, which cautioned that there was no evidence that the sound recording companies would be compelled to maintain the same industry structure and accept the loss of substantial royalty income. See Dissent at 4 (‘‘[T]he record companies may decide to keep their rates high despite the increase in mechanical rates, or decide it is in their interest to avoid a reduction in royalty revenue by creating a completely different paradigm for streaming, by which the record companies move the streaming service in-house and effectively destroy the existing services.’’).92 Also, the ‘‘different timings of different rate bargains,’’ another independent variable identified in Professor Watt’s remand testimony, was an issue raised to him at the hearing by Judge Strickler. Professor Watt candidly agreed that the Judge was ‘‘absolutely correct’’ that there is a ‘‘risk, then, of disrupting the market by having a total royalty that’s greater than what is indicated by your Shapley testimony, simply because of the disparity of times in which the rates are . . . implemented.’’ 3/27/17 Tr. 3091–92 (Watt) (emphasis added). However, this admitted risk of disruption was not addressed by sufficient record evidence.93 92 The Dissent noted that this risk was speculative in nature because there was no evidence proffered at the hearing regarding the reactions of the sound recording companies. But no such evidence was forthcoming in the remand proceeding either, and, as noted supra, the burden of proof in this regard falls on Copyright owners as the proponents of the uncapped TCC rate prong. In fact, because the major publishers who are members of the NMPA (a constituent of Copyright Owners) are part of the same corporate structure as the sound recording Majors, the burden of producing evidence would fall on Copyright Owners as well regarding the sound recording companies’ reaction to the ‘‘seesaw’’ effect. 93 As noted supra, Copyright Owners did not call any sound recording industry witnesses, or provide evidence from sound recording companies, indicating that labels would even be amenable to considering such renegotiated rate reductions. Instead, at the hearing, Professor Watt merely speculated that the sound recording companies might renegotiate their rates downward to reflect the seesaw effect when mechanical rates increased. PO 00000 Frm 00027 Fmt 4701 Sfmt 4700 54431 Third, disruption in the narrow sense of Factor D as applied by the Judges previously is not relevant to the present problem. An increase in total royalties is not a short-run immediate issue, but rather an ever-present possibility that the seesaw analysis does not sufficiently address. Rather, the uncapped nature of the TCC rate prong renders it unreasonable rather than narrowly disruptive. Balancing the foregoing considerations, the Judges find that Copyright Owners’ disruption-based argument lacks merit. 6. Conclusion Regarding Uncapped TCC Rate Prong For the foregoing reasons, the Judges decline to adopt the uncapped TCC rate tier proposed on remand by Copyright Owners. III. Rejection of Phonorecords II Settlement as a Benchmark A. D.C. Circuit Ruling Each of the Streaming Services advanced somewhat different rate plans, but all four proffered a benchmark that ‘‘broadly sought to maintain the Phonorecords II rate structure,’’ while lowering or eliminating the mechanical floor.94 Johnson, 969 F.3d at 371. With regard to the Services’ proposed benchmark based on the Phonorecords II rates, rate structure, and terms (hereinafter, PR II-based benchmark),95 the Judges are guided by several rulings in Johnson. In particular, the D.C. Circuit found the Judges’ treatment of the PR II-based benchmark to be ‘‘muddled.’’ Johnson, 969 F.3d at 387. The D.C. Circuit emphasized that the Judges ‘‘failed to Tr.3/27/17 3093–94 (Watt) (‘‘I’m not able to comment on how, you know, how possible it is to take an agreement that’s in force and then change it.’’). Not only was that mere speculation, it was provided by an economist who is neither a music industry executive nor an attorney, and the witness did not testify that he had spoken to anyone who would have industry knowledge regarding whether a label would even be amenable to considering such rate reductions. 94 The ‘‘mechanical floor’’ refers to an alternative rate calculation. ‘‘If the All-In Rate calculation results in a dollar royalty payment below the stated Mechanical Floor rate, then that floor rate would bind.’’ Determination at 26 n.59. 95 See Services’ Joint Rate Proposal (in Services’ Joint Written Direct Remand Submission at Tab C) (Apr. 1, 2021). According to the Services, their rate proposal in this proceeding is meant to ‘‘update the Phonorecords II terms to include terms of the Determination, as amended during the implementation of the Music Modernization Act, that were upheld in Johnson . . . including terms relating to student and family plan products, or that were not challenged by either the Copyright Owners or the Services.’’ Id. at 2. The Services include in their Joint Rate Proposal a chart summarizing the proposed rates for their offerings. That chart is attached as an Addendum to this Initial Ruling. E:\FR\FM\10AUR2.SGM 10AUR2 54432 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations explain’’ their rejection of the PR IIbased benchmark. Id. at 367. See also id. at 376 (Judges ‘‘failed to ‘‘reasonably explain’’ rejection). In the appeal, Copyright Owners attempted to defend the Judges’ reliance on the absence of evidence of the settling parties’ subjective intent in reaching the Phonorecords II terms. Id. at 387. The D.C. Circuit dismissed Copyright Owners’ post hoc attempt, noting that ‘‘nowhere does the [ ] Determination explain why evidence of the parties’ subjective intent in negotiating the Phonorecords II settlement is a prerequisite to its adoption as a benchmark.’’ Id. at 387 (emphasis added). The D.C. Circuit also criticized the attempt by the Judges’ appellate counsel to ‘‘change tack’’ and argue that their rejection of the PR II-based benchmark was reasonable because: (1) evidence showed that the prior rates had been set far ‘‘too low’’ and (2) it was ‘‘outdated’’. The D.C. Circuit found that those arguments also were ‘‘nowhere to be found in the [ ] Determination’s discussion’’ of the appropriateness of the Phonorecords II settlement as a potential benchmark. Id. at 387 (emphasis added).96 In the end, the D.C. Circuit agreed with the Streaming Services that, inter alia, the Judges failed to reasonably explain their rejection of the benchmark and, for all of the reasons cited, vacated and remanded the adopted rate structure and percentages for further proceedings. Id. at 381. B. Remand Procedure Regarding the PR II-Based Benchmark On December 15, 2020, subsequent to the D.C. Circuit’s decision, the Judges entered an Order Regarding Proceedings on Remand, in which the Judges stated: The Judges accept the parties’ proposals to resolve the issues concerning the use of the Phonorecords II settlement as a benchmark. . . . . . . The Services and Copyright Owners also agree that the Judges should resolve this issue based on the existing record, after receiving two rounds of additional briefing from the parties. ddrumheller on DSK120RN23PROD with RULES2 Remand Order at 1–2. 96 In the present remand ruling, the Judges do not rely on their appellate counsel’s ad hoc arguments that the D.C. Circuit found to be absent from the Determination. The Judges note though (as discussed in more detail infra) that in this Initial Ruling they are increasing the 10.5% royalty rate in the Phonorecords II rates by 44% to 15.1% (as phased-in by the Determination), thus addressing appellate counsel’s ad hoc assertion that the Phonorecords II rates were ‘‘too low.’’ Similarly, as discussed infra, the Judges address the notion that the PR II-based benchmark is outdated. VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 Based on the ruling in Johnson the Judges reject Copyright Owners’ position that they need not engage in a full analysis of the issue. The Judges conclude that they must engage in, and fully articulate, a reasoned analysis that adequately addresses ‘‘the issues concerning the use of the Phonorecords II settlement as a benchmark.’’ Id. (emphasis added). If the Judges determine that the Majority properly rejected the Services’ proposed use of the PR II-based benchmark, the rejected portions will play no part in the Judges’ remand ruling. On the other hand, if the Judges find, after engaging in that analysis, that the PR II-based benchmark was not properly rejected then, as a matter of law and logic, the Judges must weigh the Services’ PR II-based benchmark for application, in whole or in part. The Judges reject Copyright Owners’ reading of Johnson as holding that the Judges cannot fully consider the PR IIbased benchmark on remand. Copyright Owners argue that the D.C. Circuit ‘‘did not suggest the [Judges] substantively erred’’ in rejecting that benchmark, or that they ‘‘needed to reconsider [their] decision,’’ but had ‘‘merely remanded for a ‘reasoned analysis’ . . . as to why it did so.’’ CO Initial Submission at 10; see also Copyright Owners’ Reply Remand Brief at 7–8. Because Johnson ruled that the Majority’s reasoning was muddled, indiscernible, unexplained and lacking in reason, the D.C. Circuit obviously neither accepted nor rejected the Majority’s disregard for the PR IIbased benchmark—thus requiring the CRB Judges to take a comprehensive look at that benchmark. In this regard, the Judges agree with the Services that, pursuant to apposite case law, if the outcome of the remand as to this issue was preordained pending the further ‘‘reasoned analysis,’’ the D.C. Circuit would have expressed a desire simply to remand without vacating as to this issue. Services’ Joint Remand Reply Brief at 7–8 (citing Allied-Signal, Inc. v. NRC, 988 F.2d 146, 150–51 (D.C. Cir. 1993) (‘‘The decision whether to vacate depends on the seriousness of the order’s deficiencies (and thus the extent of doubt whether the agency chose correctly) and the disruptive consequences of an interim change that may itself be changed.’’)).97 Because Johnson held that the Majority’s reasoning was muddled, indiscernible, unexplained, and lacking 97 However, the Judges note that section 803(d)(3) may require the D.C. Circuit to remand rather than reverse when the issue concerns more than rates alone. Thus, the statute appears to require a remand in order for the Judges to apply their statutory authority and expertise in toto. PO 00000 Frm 00028 Fmt 4701 Sfmt 4700 in reason, the D.C. Circuit obviously neither accepted nor rejected the Majority’s disregard for the PR II-based benchmark. Thus, the Judges take a comprehensive look at that benchmark’s rates and rate structure to evaluate its usefulness in this proceeding. Relatedly, the Judges also reject Copyright Owners’ assertion that the Judges can only consider on remand the Phonorecords II rates, and cannot consider on remand the relative strengths and weaknesses of the structure in which those rates are embedded. See Copyright Owners’ Reply Brief on Remand at 14. This distinction is impractical and unworkable. If the (non-‘‘headline’’ rates 98) themselves can be reviewed and found acceptable (as they are infra) into what structure would they be placed? There are multiple provisions in the Phonorecords II rate structure providing for different rates, designed to balance (1) the ability of services to attract consumers with a low Willingness-toPay and/or a low Ability-to-Pay (the price discriminatory and differentiated features 99) with (2) the revenue diminution protections for which Copyright Owners had successfully negotiated. Moreover, the D.C. Circuit has vacated the Determination, and in doing so did not make any rulings critical of the rate structure in the Phonorecords II-based benchmark that would suggest the cramped review advocated by Copyright Owners. Indeed, the D.C. Circuit explicitly stated, without distinguishing between rates and structure, that it ‘‘agree[s] with the Streaming Services that the [Judges] . . . failed to reasonably explain [their] rejection of the Phonorecords II settlement as a benchmark . . .’’ See Johnson, 969 F.3d at 376; see also id. at 389 (issues relating to ‘‘rates’’ and ‘‘rate structure’’ are ‘‘intertwined’’). Further, the Judges emphasize that the rate structure of the PR II-based benchmark provides protection sought by Copyright Owners against revenue diminution by the Services—protection they would otherwise lose—because in this Initial Ruling the Judges are not adopting the vacated uncapped TCC prong for which Copyright Owners are now advocating, and which they claim 98 As explained elsewhere in this Initial Ruling, the Judges are increasing the ‘‘headline’’ rate from 10.5% to 15.1%. 99 Specifically, the PR II-based benchmark would incorporate the price discriminatory features for product differentiation as between: (1) subscription vs. ad-supported services; (2) portable and nonportable services; and (3) unbundled vs. bundled services. See Determination at 10; Dissent at 26. The third category—bundled vs. unbundled—is discussed infra in the context of the Bundled Revenue definition. E:\FR\FM\10AUR2.SGM 10AUR2 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations ddrumheller on DSK120RN23PROD with RULES2 would have protected them in that regard. Cf. CO Additional Submission at 4–6 (acknowledging PR II-based benchmark provided some TCC provisions, allowing for protection against revenue diminution). Thus, the Judges’ remand rulings on the PR IIbased benchmark rates and on the uncapped TCC rate prong are inextricably interlaced. See Johnson, 969 F.3d at 381 (absence of ‘‘reasoned explanation’’ for rejecting PR II-based benchmark was problematic because it occurred ‘‘when’’ Judges adopted an alternative proposal that called for ‘‘setting . . . total content cost and revenue rates.’’) (emphasis added). The Judges weigh each benchmark’s intrinsic strengths and weaknesses, as well as its comparative advantages and disadvantages vis-a`-vis other proffered benchmarks. On remand, the interrelationships of the competing benchmarks are of particular importance, given Copyright Owners’ need for the aforementioned protections against revenue diminution via price discrimination.100 100 The Judges categorically reject Copyright Owners’ assertion that the PR II-based benchmark cannot be considered because the parties agreed in the Phonorecords II settlement that any future statutory mechanical rate determination would made ‘‘de novo’’ vis-a`-vis that settlement determination. In fact, the industrywide representatives (NMPA and Digital Media Association (DiMA)) who entered into the settlement conspicuously did not agree that the existing rate structure or rates could not be considered as the bases for future rate determinations. By contrast, the Phonorecords I settlement agreement expressly stated ‘‘[s]uch royalty rates shall not be cited, relied upon, or proffered as evidence or otherwise used in the [Phonorecords II] Proceeding.’’ Trial Ex.6013, Phonorecords I Agreement at sec. 3. Compare Trial Ex. 6014, Phonorecords II Agreement at sec. 5.5 (omitting clause precluding reliance on evidentiary value of Phonorecords II royalty rates and including full-integration clause). This change objectively demonstrates that the parties to the 2012 settlement understood the evidentiary value of the Phonorecords II settlement in the next section 115 proceeding, i.e., this proceeding. See Dissent at 15– 16. On the other hand, the Judges reject the Services’ argument that the Phonorecords II rates and structure should be retained merely because the Services relied on their continuation to make investments in their business models. As Copyright Owners note, the applicable regulations provide that ‘‘[i]n any future proceedings the royalty rates payable for a compulsory license shall be established de novo.’’ 37 CFR 385.17; see also 37 CFR 385.26. A party may feel confident that past is prologue and that the parties will agree to rollover the extant rates for another period; a party could be sanguine as to its ability to make persuasive arguments as to why the rates should remain unchanged; a party might even conclude that the mechanical rate is such a small proportion of the total royalty obligation that its increase would be unlikely to alter long-term business plans. But for sophisticated commercial entities to claim that they simply assumed the rates would rollover—without the reasonable possibility of significant adjustment or outright abandonment— VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 Through this approach, the Judges ultimately may adopt only one of the parties’ benchmarks or other methodologies, or they may modify the proposals by combining them, provided such a modification is ‘‘within a reasonable range of contemplated outcomes . . . piecing together a rate structure, the economic and policy consequences of which had already been explored and developed by the parties in the record.’’ Johnson, 369 F.3d at 382. In their consideration of the PR IIbased benchmark, the Judges are not suggesting that this benchmark is the optimal tool to use in order to identify rates and terms among all approaches that might have been proffered (but were not). But the Judges are cabined by the evidence they receive. See 17 U.S.C. 803(a)(1) (‘‘the Judges shall act . . . on the basis of a written record . . . .’’); see also P. Wald, supra, (noting that parties’ economic proposals made in an action ‘‘impose[ ] a practical constraint’’ on judge who will, ‘‘for the most part, be limited by what the parties serve up to her.’’). Based upon the available record evidence, the Judges find that the Services’ PR II-based benchmark— although not necessarily perfect—is more than sufficient to satisfy the legal requisites for application, as well as a practical benchmark, when used in conjunction with the 15.1% headline revenue rate advocated by Copyright Owners. See generally Nat’l Cable Television Ass’n v. Copyright Royalty Tribunal, 724 F.2d 176, 182 (D.C. Cir. 1983) (rate-setting is an intensely practical affair). C. Parties’ Remand Arguments Regarding PR II-Based Benchmark 101 1. Services’ Arguments The Services maintain that their PR IIbased benchmark satisfies the strikes the Judges as so irrational and reckless as to raise serious doubts about the credibility of that position. (If the Services had made a persuasive argument that certain fixed cost investments were ‘‘sunk’’ and had useful lives that substantially exceeded the five-year rate term, then such costs could be considered under Factor C of section 801(b)(1), but they did not make a persuasive argument in this regard. Cf. SDARS II, 78 FR 23054, 23069 (Apr. 17, 2013) (adjusting rates downward under Factor C, and distinguishing internet music transmissions, to reflect that—because Sirius XM needed to make ‘‘unique and substantial’’ investments in the form of ‘‘sunk’’ costs paid for satellites with a useful life of l2–15 years—‘‘it is not unreasonable for Sirius XM to expect to recoup a certain amount of those costs over the expected useful life of the [s]atellites,’’ which exceeded the five-year rate term.) 101 The parties made arguments both in the original hearing and in this remand proceeding regarding the Services’ proffer of the PR II-based benchmark. Each party’s pre-remand and postremand arguments overlap to some extent. PO 00000 Frm 00029 Fmt 4701 Sfmt 4700 54433 ‘‘reasonable’’ rate requirement and is consistent with the four itemized factors set forth in section 801(b)(1). They make several arguments in favor of this position. First, they aver that their PR II-based benchmark possesses all the characteristics of an ‘‘ideal’’ benchmark. Services’ Joint Opening Brief at 19. In this regard, they argue that their proffered benchmark ‘‘involves the same sellers, the same or similar buyers, and the same rights as at issue in this proceeding,’’ and that there has been ‘‘no material change in the economic circumstances of the marketplace that would warrant adjusting the rate levels or rate structure in the benchmark.’’ Id. at 20. Applying the facts to these benchmark characteristics, the Services assert that the first three elements—same sellers (here, licensors), same buyers (here, licensees) and same rights (the mechanical license for interactive streaming) are satisfied. In particular, they note that the majority of the participants in the present proceeding either directly participated in the Phonorecords II settlement process or were active in the market contemporaneous with that settlement. Id. at 20–21. Turning to the next benchmark characteristic—the absence of a ‘‘material change in the economic circumstances of the marketplace that would warrant adjusting the rate levels or rate structure in the benchmark’’— they emphasize that the PR II-based benchmark contains different rate levels for different product offerings, to account for (a) consumers’ varying willingness-to-pay (WTP) and (b) the zero marginal physical cost of digital reproductions of sound recordings containing musical works. Id. at 21–22 (citing multiple experts). Next, the Services point to the fact that the ‘‘headline’’ 102 royalty rate is based on a percent-of-revenue, so that revenue growth (or decline) on this rate prong allows for royalty payments to directly adjust in tandem. Id. Further, the Services assert that the importance of streaming as ‘‘the future of the music industry’’ was known to the Phonorecords II negotiators, as evidence by the then-recent launch in the United Examination of the pre-remand arguments is also necessary because of the findings in Johnson and because the parties agreed that the evidentiary record on this remanded issue would not be enlarged. 102 The Judges and the parties characterize the percent-of-revenue of revenue rate as the ‘‘headline’’ rate. See Johnson, 969 F.3d at 383 n.10. E:\FR\FM\10AUR2.SGM 10AUR2 ddrumheller on DSK120RN23PROD with RULES2 54434 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations States of the popular Spotify service. Id. at 23. Beyond these benchmark requisites, the Services also emphasize that the PR II-based benchmark is the product of a settlement whose negotiated features burnish the value of this benchmark as reflective of effective competition. Specifically, they note: • The settlement was negotiated in the same statutory context, concerning the identical rate standard and factors as applicable to the present proceeding. • Neither side would have accepted a deal materially worse than what it expected from a section 115 proceeding applying the section 801(b)(1) considerations. • The statutory alternative diminishes any additional licensor-side negotiating power arising from ‘‘Must-Have’’ complementary oligopoly of the licensors of the musical works publishers. Id. at 22. Moving from the negotiating context to market performance under this standard, the Services aver that this approach has borne fruit for the industry as a whole. They point to the evidence of the licensors’ consistent profitability and the licensees’ ability to ‘‘benefit’’ from the Phonorecords II approach. Id. at 23. The Services also maintain that the Phonorecords II structure ‘‘addresses any concerns with bundling and the potential for revenue deferment.’’ Id. at 24.103 They assert that these issues were specifically addressed by Copyright Owners during the Phonorecords II negotiation, because ‘‘multiproduct firms such as Yahoo and Microsoft’’ that offered streaming services had the capacity to make bundled offerings to consumers. These concerns were addressed in the Phonorecords II rate structure, the Services note, through the use of ‘‘multiple rate prongs, minima and floors,’’ ensuring that ‘‘the total musical works royalty for certain types of offerings does not fall below a specified level,’’ thereby ‘‘mitigat[ing] the effect of any potential revenue deferrals and appropriately address[ing] any concerns with bundling.’’ Id.104 Finally, the Services maintain that ‘‘[d]irect agreements between Copyright Owners and Services also support adoption of the PR II-based benchmark.’’ Id. at 34. In particular, they note that many of the royalty rates (and terms) in these direct agreements apply the Phonorecords II rates. Moreover, the Services maintain, because these direct agreements are in the nature of blanket license of a publisher’s entire catalog, they provide an added ‘‘access’’ value in the form of full-repertoire licensing. These direct agreements do not include a rate above Phonorecords II levels; thus, the Services contend, they underscore the reasonableness of the Phonorecords II rates. Id. 105 Finally, the Services aver that the PR II-based benchmark satisfies the itemized four section 801(b)(1) factors. With regard to Factor A, they maintain that: (1) the Phonorecords II framework has corresponded with an increase in the supply of musical works; (2) the PR II-based benchmark will increase the likelihood that the Services will increase subscriber counts, generating profitability, which will make streaming available to more listeners; and (3) the price discriminatory aspects of this royalty rate structure allows the Services to afford to offer streamed music to listeners with a low willingness (or ability)-to-pay, at lower rates or through ad-supported services. Services’ Joint Opening Brief at 25–27. Regarding Factors B and C (the ‘‘fair return’’ and ‘‘relative contributions’’ objectives), the Services emphasize that the PR II-based benchmark satisfies these statutory elements because it: (1) was the result of negotiations between industrywide representatives who had every incentive to obtain a ‘‘fair’’ return and to receive recompense for their ‘‘contributions’’ to streaming; and (2) allowed interactive streaming to become ‘‘a significant means for consumers to listen to music’’ while simultaneously generating growth in annual royalties for Copyright Owners.’’ Id. at 27–29. Lastly on the subject of the statutory factors—regarding Factor D (minimizing disruptive impact)—the Services make a 103 The issue of bundling is addressed in this Initial Ruling infra, in connection with the Judges’ definition of Service Revenue generated through the offering of sound recordings as part of a bundle containing other goods or services. 104 The Services also reiterate their pre-remand argument that the Phonorecords III settlement of subpart A rates for sales of physical and digital download phonorecords (now reorganized in subpart B) confirms the appropriateness of the Phonorecords II-based benchmark. However, any further reliance by the Services on that argument is moot, because the D.C. Circuit affirmed the Majority’s analysis of the subpart A rates. Johnson, 969 F.3d at 386 (noting that the Majority adequately explained treatment of the subpart A rates as ‘‘ ‘at best’ a floor’’ below which the mechanical royalty rates paid by the Services for interactive streaming cold not fall). 105 Under section 115—prior to the effective date of the 2008 Music Modernization Act—an interactive service was required to serve a ‘‘Notice of Intent’’ to use the copyright license (NOI) with the owner of a copyright for each musical work before streaming the sound recording embodying that musical work. By contrast, a direct license with a publisher covers more than an individual musical work by providing ‘‘access’’ value to an entire catalog, without the transaction cost burden of filing multiple individual NOIs. VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 PO 00000 Frm 00030 Fmt 4701 Sfmt 4700 succinct argument: ‘‘By renewing the rate levels and structure of Phonorecords II, there is minimal risk of disruption.’’ Id. at 29–30. The Services also address several further criticisms of the PR II-based benchmark contained in the Determination. Focusing first on an issue specifically addressed in Johnson, they assert the irrelevancy of the ‘‘subjective intent’’ of the parties that negotiated the Phonorecords II settlement—a factor on which the Majority relied in deciding not to adopt the PR II-based benchmark. In this regard, the Services are also responding to the D.C. Circuit’s concern regarding this issue. See Johnson, 969 F.3d at 387 (‘‘In rejecting that settlement as a possible benchmark, the [Judges] faulted the Streaming Services for failing to explain why the parties to the Phonorecords II settlement agreed to the rates in that settlement . . . [b]ut nowhere does the [ ] Determination explain why evidence of the parties’ subjective intent in negotiating the Phonorecords II settlement is a prerequisite to its adoption as a benchmark.’’). The Services note that no benchmark evidence presented by any party is proffered with supporting evidence of the subjective intent of the bargainers who negotiated the benchmark. Moreover, they note that the Majority in fact acknowledged that ‘‘[r]elying on a benchmark as objectively useful without [the need for] further inspection’’ is ‘‘typical and appropriate for the benchmarking method.’’ Id. at 35 (quoting Determination at [55] & n.106 (emphasis added)). With regard to other criticisms of the Majority’s failure to use the PR II-based benchmark, the Services argue that the Majority misapplied their previous rulings that they ‘‘cannot and will not set rates to protect any particular streaming service business model.’’ Id. at 37 (quoting Phonorecords III, 84 FR 1945). The Services find this principle inapposite, because their point is that the multiple price-discriminatory aspects of the Phonorecords II approach made it ‘‘a valuable benchmark . . . because it had allowed for different service types to emerge and grow, which benefits the entire market.’’ Id. at 37. The Services also take issue with the Majority’s assertion that the Phonorecords II rate structure was too complex, deriding it as a ‘‘RubeGoldberg-esque’’ contraption. Id. at 38. Rather, the Services maintain that the structure was as complex as necessary to effectuate the parties’ needs, particularly the price discriminatory features and the protections against E:\FR\FM\10AUR2.SGM 10AUR2 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations revenue diminution. Id. at 38–39. Further, the Services note that the record is devoid of any testimony or evidence indicating any actual confusion caused by the Phonorecords II rate structure. Id. at 39. Finally in this regard, the Services maintain that the rate structure adopted by the Majority is essentially as complex as the structure in Phonorecords II, with the only major change being the replacement of the capped TCC rates with uncapped TCC rates.106 Id. The Services address another criticism—that the rates in the PR IIbased benchmark are too low. This issue is largely moot, as the D.C. Circuit’s affirmance of the Majority’s expert ‘‘line-drawing’’ and ‘‘reasoned weighing of the evidence’’ confirmed that a rate increase was necessary. In this Initial Ruling, the Judges have acknowledged specifically the appropriateness of the 15.1% revenue rate—a 44% increase over the 10.5% headline rate in the PR II-based benchmark.107 ddrumheller on DSK120RN23PROD with RULES2 2. Copyright Owners’ Arguments Copyright Owners assert that the record evidence overwhelmingly supports the Judges’ rejection of the PR II-based benchmark. At the outset, they maintain that the Judges found—and the D.C. Circuit affirmed—that a rate increase was required in the Phonorecords III terms. CO Initial Submission at 13. (As noted, an increase in the headline rate by 44%, to 15.1%, is adopted in this Initial Ruling.) Next, Copyright Owners maintain that the evidence established that ‘‘market conditions’’ were ‘‘radically different’’ at the time of the Phonorecords III proceeding compared with when the parties entered into their 2012 industrywide agreement in Phonorecords II. Id. at 17. In particular, Copyright Owners point to testimony describing the streaming industry as ‘‘nascent’’ in 2012, with fewer streams, subscribers, services, and choices of music; operating in a consumer 106 As discussed infra, the relative complexity or simplicity of the rate structure is not a statutory factor, nor is it a decisive element of a reasonable rate structure, when the details of that structure effectuate price discriminatory configurations that would increase the availability of music and streaming revenues and otherwise satisfy the statutory criteria. 107 The Judges characterize this issue as largely moot because the PR II-based benchmark includes on its ‘‘lesser of’’ prongs price discriminatory rates, discussed infra. But those ‘‘lesser of’’ rates are overridden by the ‘‘greater’’ 15.1% rate. As also discussed infra, Mechanical Floors continue to bind at lower mechanical royalty levels (without reducing the songwriters’ ‘‘All-In’’ musical works royalty that includes the performance royalties), because these floors were retained in the Determination and were not the subject of appeal. VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 environment when download purchases and Pandora’s noninteractive service were the predominant means for consumers to listen digitally to music. Id. at 18–21. In sum, Copyright Owners maintain, that streaming was ‘‘economically insignificant’’ to the music industry when the PR II provisions were adopted. Id. at 20. Copyright Owners particularly emphasize the substantial increase in streaming revenue during the Phonorecords II period. They point out that while ‘‘total streaming revenue had ranged from approximately $150 million in 2005 to $212 million in 2010, . . . after 2012[,] annual [streaming] revenue exploded to reach approximately $1.6 billion by 2015.’’ Id. at 23. Further, they note there is no evidence that the music publishers or anyone else had predicted this substantial rise in streaming and the revenues it generated, and that in no way could it be inferred that those rates had ‘‘baked-in’’ future growth. In fact, Copyright Owners assert at the hearing that the PR II rates were merely ‘‘experimental’’—consistent with the relatively nascent stage of the streaming industry. Id. at 25. Additionally, Copyright Owners maintain that the identities of the parties involved in the Phonorecords III proceeding are different from those who established the Phonorecords II framework. Although they acknowledge the presence of current interactive services Spotify and Rhapsody in this market prior to the Phonorecords II framework agreed to by the trade associations for the interactive services and the music publishers, they point out that ‘‘[n]one of the other participants in this proceeding even entered the streaming business until after the Phonorecords II settlement.’’ Id. at 21. Next, Copyright Owners assert that the Services’ evidence is inadequate to support a finding that the rates in their PR II-based benchmark are suitable for use in setting royalty rates in this proceeding. First, they echo the Determination, which stated that the Services (1) did not examine in detail the particular rates within the existing rate structure; (2) relied on the 2012 rates as objectively useful without further inspection; and (3) did not call witnesses to testify regarding the 2012 settlement negotiations. Id. at 27 (citing Determination, 84 FR 1944 & n.106). Because of the absence of the foregoing evidence, Copyright Owners assert that the Services were left with ‘‘no evidence explaining how the particular rates and percentages in those settlements were calculated or derived, how they were negotiated, or how they were reasonable in light of the explosive growth in the PO 00000 Frm 00031 Fmt 4701 Sfmt 4700 54435 streaming marketplace between the time of those settlements and the Phonorecords III proceeding.’’ Id. at 28. The absence of such evidence, according to Copyright Owners, meant that the Services had failed to carry their burden of proof under 5 U.S.C. 556(d) with respect to their proposal, a burden Copyright Owners assert the Services acknowledged they bore. Id. at 29–30. Additionally, Copyright Owners claim that the D.C. Circuit found ‘‘validity’’ in Copyright Owners’ assertion that the subjective intent of the parties to the Phonorecords II settlement is relevant because it would have revealed whether the agreed-upon rates were based on economic realities or instead were driven by other considerations. Id. at 30–31 (citing Johnson, 969 F.3d at 387). However, Copyright Owners acknowledge that, because this was not a reason given by the Majority, it carried no weight with the D.C. Circuit on appeal. Id. at 31. 3. Analysis and Decision Regarding PR II-Based Benchmark 108 a. PR II-Based Benchmark Meets Most of the Requisites for a Useful Benchmark The four classic characteristics of an appropriate benchmark are: (1) the degree of comparability of the negotiating parties to the parties contending in the rate proceeding, (2) the comparability of the rights in question, (3) the similarity of the economic circumstances affecting the earlier negotiators and the current litigants, and (4) the degree to which the assertedly analogous market under examination reflects an adequate degree of competition to justify reliance on agreements that it has spawned. In re Pandora Media, 6 F.Supp.3d 317, 354 (S.D.N.Y. 2014, aff’d sub nom Pandora Media Inc. v. ASCAP, 785 F.3d 73 (2d. Cir. 2015). As discussed below, the PR II-based benchmark meets criteria (1), (2) and (4), but requires adjustment to fully satisfy criterion (3). First, the PR II-based benchmark obviously pertains to the same rights at issue in this proceeding, as it reflects the licensing provisions from the immediately preceding mechanical license proceeding. Second, the licensors (songwriters and music publishers) and licensees (interactive streaming services) are 108 The setting of statutory royalty rates involves to a significant degree the application of economic analysis. Accordingly, the Judges find it appropriate to set forth certain key aspects of microeconomics that guide the application of the section 801(b)(1) standard in the present proceeding. That guidance is set forth more fully in the Dissent at 29–39. E:\FR\FM\10AUR2.SGM 10AUR2 54436 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations ddrumheller on DSK120RN23PROD with RULES2 comparable (albeit not identical). While Copyright Owners emphasize the different identities and market involvement of the licensees, particularly the greater market penetration of Amazon, Apple, and Google, the Services note that even prior to the more significant entry of these three entities, similar multiproduct firms, such as Yahoo and Microsoft, were active licensees. The Judges find that the changing identities of the large multiproduct technology firms does not demonstrate the absence of comparability between and among such firms in the Phonorecords II and Phonorecords III rate periods. The shifting market entries, exits, strategies, successes and setbacks of otherwise comparable firms are expected occurrences in a dynamic capitalist market system and are not factors that materially diminish the necessary comparability of the parties for benchmarking purposes. Third, important economic fundamentals of the marketplace are sufficiently similar in crucial respects. First, the heterogeneity of the willingness-to-pay among subscribers and listeners in the downstream market continues to support price discrimination and thus differentiated royalty rates upstream pursuant to the concept of ‘‘derived demand.’’ See Determination at 19 (and record citations therein) (‘‘Weighing all the evidence and based on the reasoning in this Determination, the Judges conclude that a flexible, revenue-based rate structure is the most efficient means of facilitating beneficial price discrimination in the downstream market.’’); Dissent at 32, 51, 86, 121, 126 (and record citations therein).109 Second, the items being licensed for transmission—‘‘second copies’’ of sound recordings (with embedded musical works)—have a marginal physical cost of zero, a critical economic point on which the experts for both parties concur, and as to which the Majority and the Dissent repeatedly and significantly rely. See Determination at 18, 21, 36, 59, 80 (and record citations therein); Dissent at 30–31, 33–34, 37, 47, 109 The Determination asserts that it includes a price discriminatory feature because a revenue percentage-based rate is itself price discriminatory, in that it does not set royalties on a per-play basis. Determination at 35 n.71. But that ‘‘blunt’’ form of price discrimination does not capture the granular discriminatory features that the parties had negotiated. There is no sufficient basis for the Judges to substitute their own blunt conception of the appropriate form and extent of price discrimination for the structure generated in negotiations by the market participants. See Dissent at 37. VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 49–50, 59, 122, 127–128 (and record citations therein).110 Copyright Owners are clearly correct, however, in noting a substantial change in economic circumstances that distinguished the Phonorecords II negotiations from the current proceeding; viz., the dramatic growth of interactive streaming revenues.111 The economic impact of this revenue growth is incorporated into the experts’ Shapley Value Models and the Judges’ analysis of same. This analysis has generated the 44% increase in the headline royalty rate, from 10.5% to 15.1% (as phasedin by the Majority and again in this Initial Ruling).112 Simply put, three economic principles co-exist. First, the downstream interactive streaming market remains differentiated among listeners with different willingnesses and abilities to pay, based on varied preferences (utility) and disparities in income. Second, streaming of the ‘‘second copy’’ of the sound recordings (with embedded musical works) remains physically costless (but generates potential ‘‘opportunity costs’’). But, third, streaming revenues have grown substantially. There is no incompatibility or inconsistency in the simultaneity of these economic principles. Each of them must be taken into account and they are in this Initial Ruling. This economic context refutes the arguments made during oral argument at the D.C. Circuit that the PR II-based benchmark should be rejected in toto 110 It bears emphasis that the fact ‘‘second copy’’ reproductions are physically costless does not even suggest that the market price should be zero. Rather, in this ‘‘second-best’’ economic context, pricing above marginal physical costs is imperative in order for Copyright Owners to recover their ‘‘first copy’’ costs, avoid ‘‘opportunity costs,’’ and earn profits. See Dissent at 36–38. 111 Copyright Owners also cite data demonstrating the increase in listeners and the number of streams. The Judges find those data to be causal for the key point in rate setting in this proceeding—the significant increase in revenues. 112 At first blush it may seem that the increase in interactive revenues is not an economic fundament that would support an increase in a percentage-ofrevenue based royalty formula. However, as more fully discussed herein, under the Shapley Value approach, the increase in revenues has generated an increased ‘‘Shapley Surplus’’ (roughly analogous to interactive streaming industry profits), which the two ‘‘Must Have’’ input suppliers (record companies and Copyright Owners) will essentially split equally. If this surplus increases faster than the interactive services’ non-content costs (or if those costs remain stable or fall), the increased revenues would flow disproportionately to theses input suppliers, thus causing the increase in revenues to support an increase in the royalty rate, all other things held constant. And, because the ‘‘Must Have’’ input suppliers have complementary oligopoly power, the Majority relied on a Shapley model constructed by Spotify’s expert, Professor Marx, that adjusted for this market power. PO 00000 Frm 00032 Fmt 4701 Sfmt 4700 because it was supposedly ‘‘outdated.’’ The heterogeneity of the downstream demand of listeners and the zero physical cost of ‘‘second copies’’ are enduring features that affect the upstream market via the principle of derived demand. The substantial growth of streaming revenues, however, necessitated an increase in the headline rate from 10.5% to 15.1% (as phasedin), for the reasons discussed in the Judges’ analysis in this Initial Ruling of the interrelationship among: (1) Shapley Value modeling; (2) Nash Bargaining; (3) complementary oligopoly power; and (4) effective competition. Further, the foregoing analysis also undermines the pre-remand argument made by Copyright Owners that the PR II-based benchmark reflects a market that was not yet ‘‘mature,’’ or was only ‘‘experimental.’’ Markets are not ‘‘mature’’ as opposed to, say, ‘‘adolescent.’’ Indeed, the metaphor is strained because all economic models are subject to revision if the salient facts have changed, without rendering the prior models mere ‘‘experiments.’’ Markets simultaneously exhibit enduring characteristics—here, heterogeneous customers and zero marginal physical costs and dynamic change—here, significant revenue increases.113 And yet, Copyright Owners seek to deny the idea that these principles could exist simultaneously. In an attempt to disqualify the application of the PR II-based benchmark, Copyright Owners complain: [W]hile streaming activity and revenues grew under the Phonorecords II royalty rates, the [REDACTED]. For example . . . [REDACTED]. CO Initial Submission at 15–16 (emphasis added). But as the Services explained, the economic defect in Copyright Owners’ analysis, is that it ignores the principle of price discrimination and its beneficial effects: [A]s [Professor] Hubbard explained, it is ‘‘meaningless’’ to compare growth in streams to growth in royalties in the context of Prime Music in particular because the record showed that Prime Music brings ‘‘new people into the market.’’ . . . If not for the flexibility (and beneficial price discrimination) the 113 If one were to indulge the ‘‘maturity’’ metaphor, the ongoing creative destruction in the streaming industry has only reinforced the fact that, according to one of Copyright Owners’ own economic expert witnesses, the interactive streaming market (as of the Phonorecords III hearing) was not yet mature, but rather remained ‘‘a relatively new enterprise.’’ Watt WRT ¶¶ 39–40. Thus, it is hardly clear from the record that interactive streaming has ‘‘matured’’ in a manner that would render anachronistic the enduring marketplace characteristics. E:\FR\FM\10AUR2.SGM 10AUR2 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations existing Service Provider Revenue definition and rate structure facilitated, the Copyright Owners ‘‘would have gotten zero’’ from those new listeners. . . . ‘‘So they’re better off by that amount’’ of royalty growth. . . . The undisputed fact that [REDACTED]—reflects that the existing rule enables beneficial price discrimination that expands the total royalty pool and benefits Copyright Owners. ddrumheller on DSK120RN23PROD with RULES2 Services’ Reply at 58–59. This rebuttal by Professor Hubbard is an example of the important distinction between ‘‘increases in demand’’ (when the demand curve shifts outward) and movements ‘‘down the demand curve’’ (when sellers use price discrimination to generate more revenue without additional cost to attract buyers with a lower willingness or ability to pay). The parties’ otherwise dueling economists agreed on this point. Compare 4/3/17 Tr. 4373–74 (Rysman) (Copyright Owners’ witness acknowledging that under the current rate regime overall revenues might be increasing because of movements ‘‘down the demand curve’’ (i.e., changes in quantity demanded in response to lower prices), rather than because of, or in addition to, an outward shift of the demand curve (i.e., increase in demand at every price)) with 3/13/17 Tr. 701 (Katz) (the Services’ witness who likewise noted that the present structure enhances variable pricing that allows streaming services ‘‘to work[ ][their]way down the demand curve.’’). Moreover, Copyright Owners baldly cherry-pick the data they present. [REDACTED] CO Initial Submission at 15–16. So, by their own data, presented in their own brief, they acknowledge that [REDACTED]. See Services’ Reply at 57–58 (Copyright Owners have proven the ‘‘opposite’’ of what they intended). This is precisely what beneficial price discrimination is designed to accomplish.114 The appropriateness of adopting the price discriminatory rate provisions of the PR II-based benchmark is further underscored by Copyright Owners’ candid acknowledgement at the hearing that they were essentially urging the Judges to adopt what is known as the ‘‘Bargaining Room’’ approach to rate setting. See Dissent at 24 (and record citations therein).115 114 Further, [REDACTED] because: (1) the marginal physical cost of ‘‘second-copy’’ streams is zero; (2) royalties were calculated [REDACTED]; and (3) Copyright Owners’ original proposed a perplay (i.e., per-stream) metric, which was rejected by all three of the Judges. 115 The Bargaining Room approach was first proposed for incorporation into the statutory license standard in 1967 by the NMPA, to be included in the predecessor section, later reorganized in section 801(b)(1) that governs this proceeding. See Dissent at 22–24 (and citations VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 In the present proceeding, the appropriateness, vel non, of the Bargaining Room approach boils down to the following: Copyright Owners emphasize the inability of the Judges (or anyone) to identify present market rates precisely, let alone over the fiveyear rate period because the compulsory license set by the Judges cannot possibly contemplate every single business model that may develop in the ensuing time. . . . If the statutory rate is set below market rates, then the parties will never negotiate upward toward the market rates, because the licensees will always prefer to invoke the right to use the licensed work at the belowmarket statutory rates. However, if the Judges set the statutory rate above what they find to be market rates, different licensees who each have a maximum willingness to pay (WTP) below such a statutory rate would seek to negotiate lower rates with the licensors. In response to such requests to negotiate, according to this argument, Copyright Owners would respond by negotiating various lower rates for those licensees, provided lower rates were also in the selfinterest of Copyright Owners. Dissent at 24–25 (and record citations therein). The Judges find no reason to depart from the policy decision in Phonorecords I that the rate setting policies made explicit in section 801(b)(1) are best discharged if the Judges eschew the Bargaining Room approach and continue to identify rate structures and rates that reflect the standards set forth in the statutory provision. To supplant the statutory factors with a Bargaining Room approach would essentially be to adopt a purely market-based rate-setting approach that is inconsistent with section 801(b)(1) and with the Judges’ application of that statute to set rates, rate structures, and terms consonant with effective competition. With this background in mind, the Judges turn specifically to the interrelationship between the price discrimination aspects of the rates in the PR–II benchmark and the Bargaining Room approach. Copyright Owners have demonstrated (albeit tacitly) their understanding that, therein). Ultimately, Congress punted on the Bargaining Room approach, and adopted into law the four-factor language set forth in section 801(b)(1). A subsequent attempt by NMPA to have the Copyright Royalty Tribunal (CRT) (a predecessor to the Judges) adopt the Bargaining Room theory was rejected by the CRT, a rejection that was affirmed on appeal. See Recording Industry Ass’n. of America v. Copyright Royalty Tribunal, 662 F.2d 1, 37 (D.C. Cir. 1981), aff’g Adjustment of Royalty Payable under Compulsory License for Making and Distributing Phonorecords, 46 FR 10466, 10478 (1981). See generally, F. Greenman & A. Deutsch, The Copyright Royalty Tribunal and the Statutory Mechanical Royalty: History and Prospect, 1 Cardozo Arts & Ent. L.J. 1, 53, 64 (1982). PO 00000 Frm 00033 Fmt 4701 Sfmt 4700 54437 if the statutory provisions did not contain a price discriminatory rate structure to reflect the varying WTP, they would have to invent it. This finding is apparent from their advocacy for the adoption of a Bargaining Room approach to rate-setting. See, e.g., 4/3/17 Tr. 4390, 4431 (Rysman) (lauding bargaining room approach as reflecting ‘‘economical element of price discrimination . . . the [licensor] is picking its prices carefully.’’) (emphasis added); id. at 4431 (explaining that under this approach, when negotiating with Spotify regarding a rate for adsupported service, ‘‘Must Have’’ music publishers would ‘‘have the right . . . to set that price.’’); 4/4/17 Tr. 483–45 (Eisenach) (acknowledging Copyright Owners’ approach was consistent with Bargaining Room theory because they were seeking rates so high as to force would-be licensees to negotiate for the ‘‘Must Have’’ mechanical license.). Thus, the Judges find there to be no real dispute as to whether there is a market-based need for an upstream discriminatory rate structure.116 Rather, the parties appear to be in disagreement as to who shall be in control of the setting of rates, the Judges, through their application of law, or Copyright Owners, through the exercise of their complementary oligopoly power. The resolution of this choice is clear; the Judges, not the licensors, are statutorilycharged with establishing provisions that are reasonable and otherwise properly reflect the itemized objectives of section 801(b)(1). Fourth, the PR II-based benchmark reflect a rate structure with an adequate degree of competition, because there 116 The Majority recognized this point as well when—regarding the ‘‘increase the total revenue that price discrimination enables—they ask (and answer) rhetorically: ‘‘How could Copyright Owners and their economic experts argue against a rate structure that inures to their benefit as well? The answer is: They do not. . . . [T]hey advocate for a rate set under the bargaining room theory, through which mutually beneficial rate structures can still be negotiated, but not subject to the ‘‘reasonable rate’’ and itemized factor analysis required by law.’’ Determination at 85 & n.153. The Judges also note that Copyright Owners’ acknowledgement that they too would set price discriminatory rates and structures is not simply a feature of this market. Rather, ‘‘discriminatory pricing . . . is the normal attribute of equilibrium . . . in a broad range of market types and conditions where consumers can be separated into distinct groups with different demand elasticities.’’ W. Baumol, Regulation Misled by Misread Theory: Perfect Competition and Competition-Imposed Price Discrimination at 2 (2002). See also Dissent at 38, n.74. Given the ubiquity of discriminatory pricing, the Judges also find that the adoption into the statutory license of such pricing is not—as Copyright Owners contend—simply the inappropriate favoring of a particular business model, but rather a necessary reflection of the fundamental nature of market demand, particularly, the varied WTP among listeners. E:\FR\FM\10AUR2.SGM 10AUR2 54438 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations was a balance of bargaining power between the two negotiating industrywide trade associations, offsetting the complementary oligopoly effects in place when a ‘‘Must Have’’ licensor bargains separately with each licensee. Recently, the Judges discussed in detail how the presence of countervailing bargaining power generates royalty rates at effectively competitive levels. See Web V, 86 FR 59452, 59457 (Oct. 27, 2021). Further with regard to this fourth point, the parties have been operating over the past ten years under this basic rate structure, with profits accruing to the licensors and admittedly tolerable losses befalling the licensees. Moreover, after experience with these rates and this rate structure in the Phonorecords I period, they renewed and expanded this structure for use in the Phonorecords II period, when the alternative of a statutory rate proceeding was available to licensors and licensee alike. Their mutual willingness to continue in this manner is important evidence of the workability and reasonableness of this approach. ddrumheller on DSK120RN23PROD with RULES2 b. Evidence of Subjective Intent Not Prerequisite to Partial Adoption of the PR II-Based Benchmark 117 The Judges rely on the PR II-based benchmark as an objective benchmark. Thus, the absence of testimony regarding what went through the minds of the negotiators of the Phonorecords II agreement (and the predecessor Phonorecords I agreement) does not diminish the objective value of this benchmark. The Judges view the provisions of the PR II-based benchmark as they would any benchmark, in the context of the requisite benchmarking elements identified and discussed supra. This approach allows the factfinder to analyze the benchmark through the lens of its service in the marketplace as an objective model for the market at issue, the Phonorecords III market. See, e.g., 3/13/17 Tr. 550–51, 117 At the outset, the Judges reject Copyright Owners’ contention that the D.C. Circuit found ‘‘validity’’ in their assertion that there was merit in Copyright Owners’ assertion of the ‘‘subjective intent issue.’’ Rather, on this issue, Johnson first held: ‘‘[N]owhere does the [ ] Determination explain why evidence of the parties’ subjective intent in negotiating the Phonorecords II settlement is a prerequisite to its adoption as a benchmark.’’ Johnson, 969 F.3d at 387. Then, when Copyright Owners’ appellate counsel attempted to cure that failure by making their own ‘‘subjective intent’’ argument, the D.C. Circuit responded to that ‘‘subjective intent’’ argument with a single word: ‘‘Perhaps.’’ Id. (emphasis added). This does not in any way suggest that Johnson found ‘‘validity’’ in the ‘‘subjective intent’’ argument, but rather was a non-committal response, consistent with the D.C. Circuit’s ruling finding that the Determination had not explained this point. VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 566 (Katz) (knowledge of why parties negotiated specific provisions is unnecessary, because objective results demonstrate satisfactory performances of market). Both Professors Katz and Hubbard noted that the current rate structure remains useful, not based on consideration of the parties’ subjective understandings at the time of its creation, but because the market has not since changed in a manner that would create a basis for departure. Katz WDT ¶ 80 (‘‘My analysis has identified no changes in industry conditions since then [2012] that would require changing the fundamental structure of the percentage-of-revenue prong.’’); 4/13/17 Tr. 5977–78 (Hubbard) (changes in market are ‘‘not uncorrelated with the structure that was in place’’ in 2012).118 In this regard, it bears emphasis that Copyright Owners’ own witness, Dr. Eisenach, relied on several potential approaches that the Majority characterized as benchmarks for his rate analysis, without attempting to examine the subjective intent of the parties who negotiated those agreements. Indeed, the Majority found that the PR II Rates were properly considered as an objective benchmark, in the same manner as Dr. Eisenach’s proffered benchmarks: The Services do not examine in detail the particular rates within the existing rate structure. Rather, they treat the rates within that structure as benchmarks, i.e., generally indicative of a sufficiently analogous market that has ‘‘baked-in’’ relevant economic considerations in arriving at an agreement. Dr. Eisenach did not analyze why he chose the levels for the rates and ratios on which he relied as benchmarks or consider the subjective understandings of the parties who negotiated his benchmarks. Similarly, the Services’ economists elected to rely on the 2012 rates as objectively useful without further inspection. This point is not made to be critical of Dr. Eisenach’s approach, but rather to show that the Services’ reliance on the 2012 settlement as a benchmark shares this similar analytical characteristic, typical and appropriate for the benchmarking method. (The factual wrinkle here is that, hypothetically, the Services could have called witnesses and presented testimony regarding the negotiations that led to the 2012 (and 2008) settlements, but did not, rendering the 2012 benchmark similar to other benchmarks taken from other markets.) Determination at 55 & n.106.119 118 As noted supra, the relevant material change since the Phonorecords II agreement was reached is the significant growth in streaming revenues. That change is reflected in the Judges’ application of the Shapley Value analyses, by which the Judges increased the headline royalty rate by 44%, from 10.5% to 15.1% (phased-in). 119 Copyright Owners do not deny that they did not offer evidence of subjective intent for Dr. Eisenach’s benchmarks. Rather, they assert Dr. PO 00000 Frm 00034 Fmt 4701 Sfmt 4700 Copyright Owners also aver that they entered into the Phonorecords II settlement simply to avoid litigation costs. Copyright Owners’ Reply Brief on Remand at 29. At the hearing, this assertion was presented by David Israelite, NMPA’s President. Israelite WRT ¶ 28; 3/29/17 Tr. 3649–52 (Israelite) (claiming NMPA lacked financial position to fund rate litigation). The Services countered by noting that there was no evidence to support Mr. Israelite’s testimony in this regard, or how it may have impacted the NMPA decision to participate. And, the Services pointed out, notwithstanding his testimony regarding financial constraints, NMPA had incurred the expense of a year-long negotiation with the Services to seek higher rates, create new service categories in subpart C, and change the TCC calculations. Id. at 159, 161–64; 3/29/17 Tr. 3856 (Israelite). Further, as a general principle, a party’s mere assertion that the Phonorecords II approach was the product of a settlement that was predicated on the avoidance of litigation costs savings does not invalidate its use as a benchmark in proceedings before the Judges, especially because, by statute, the Judges are authorized to consider such agreements. See Music Choice v. Copyright Royalty Board, 774 F.3d 1000, 1014–15 (D.C. Cir. 2014) (testimony alleging agreement was reached to avoid litigation costs does not invalidate evidentiary use of that agreement for rate setting purposes, absent other evidence demonstrating settlement was involuntary or otherwise unreasonable.). Thus, the Judges find that the evidentiary record does not support Copyright Owners’ position that this ‘‘litigation cost avoidance’’ assertion constituted a separate, idiosyncratic value that diminishes the Eisenach’s reliance on benchmarks without examining the subjective understandings of the negotiators of the benchmarks is irrelevant because: (1) Copyright Owners were not seeking the adoption in toto of the rates contained in any specific benchmark cited by Dr. Eisenach; (2) Dr. Eisenach analyzed multiple benchmarks to derive a reasonable range of rates; (3) his benchmarks were not adopted; and (4) his benchmarks and are not at issue on this remand. Copyright Owners Reply Brief on Remand at 28 n.19. But Copyright Owners confuse evidentiary standards with evidentiary application. Benchmarks are subject to the same evidentiary standards, regardless of the breadth of purpose for which they are proffered and regardless of whether they were adopted or rejected. Further, the fact that Dr. Eisenach’s chosen benchmarks are ‘‘not at issue on this remand’’ does not render Copyright Owners’ reliance on purely objective benchmarks uninformative as to their own understanding of the irrelevancy of the subjective thoughts of benchmark negotiators. See generally Web IV, 81 FR 26370 (proposed benchmark adjustment based on alleged ‘‘additional value’’ should be supported by ‘‘record evidence . . . to provide a basis for such for such an adjustment.’’). E:\FR\FM\10AUR2.SGM 10AUR2 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations Judges’ partial reliance on the PR II Rates in this Initial Ruling. Copyright Owners also mistakenly rely on the fact that the Services bore the burden of proof regarding the absence of any subjective idiosyncratic factors that hypothetically could have diminished the useful value of the PR IIbased benchmark. Id. at n.21. The Services indeed bore the burden of proof (i.e., persuasion) with regard to their proffered benchmark PR II Rates, and they presented adequate objective evidence and testimony that this approach has worked in the marketplace to serve as prima facie proof to support the Judges’(partial) use of this benchmark in this remand proceeding. And, as explained above, such subjective intent was not a necessary element of their benchmark proofs. But, with regard to Copyright Owners’ rebuttal to those proofs, Copyright Owners bore the burden of production, to present sufficient evidence and/or testimony that the Judges could rely on to reject the (partial) use of the PR IIbased benchmark. This Copyright Owners failed to do.120 In fact, given Copyright Owners’ reliance on the subjective intent of the parties to a benchmark, the Judges attempted to identify potential subjective evidence of how the capped TCC rates in the PR II-based benchmark 121 were derived, during the examination of Dr. Eisenach at the hearing: ddrumheller on DSK120RN23PROD with RULES2 [JUDGE STRICKLER] Do you discuss, Dr. Eisenach, . . . in your written direct or written rebuttal testimony how the parties arrived . . . at the ratios for sound recording to musical works in [witness interrupts] [DR. EISENACH] That process is opaque to me, Your Honor. [JUDGE STRICKLER] Did you [witness interrupts] [DR. EISENACH] I know—I know there was a 2008 negotiation. I know there was a 2012 negotiation. I wasn’t . . . present, and I’m not privy to any of the details. [JUDGE STRICKLER] You were not informed by your client or by any other source of information as to how they arrived at those particular ratios? [DR. EISENACH] When I’ve asked the question, I’ve found people chuckle and— and there doesn’t seem to have been too much system—systematic thought that went into it, but I don’t really know that. I just— 120 As described in this Initial Ruling, the Judges identified this same distinction between the burden of proof and the burden of production to find in favor of Copyright Owners’ proffered expert testimony in support of their Nash Bargaining analysis, testimony which constituted prima facie proof that was not adequately rebutted by the production of sufficient testimony from the Services’ expert economic witnesses. 121 The ‘‘capped’’ TCC rates are elements of the Phonorecords II rates. VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 when I ask the question, people say: Nobody really knows. . . . Someone may know, but that’s what I’ve been told. 4/4/17 Tr. 4611 (Eisenach) (emphasis added). The Judges find it perplexing, to say the least, that Copyright Owners would ‘‘chuckle’’ when asked by their expert witness for the very subjective evidence which they claim to be relevant. But of perhaps greater relevance is Dr. Eisenach’s further testimony, quoted above, that he was also told by Copyright Owners that ‘‘nobody really knows’’ how the parties arrived at those rate ratios. Copyright Owners’ ‘‘chuckle,’’ in response to its expert’s critical inquiry as to the derivation of rates—and that expert’s understanding that his client simply did not know how those rates were derived—undercut Copyright Owners’ claim that subjective understanding of those rates could undermine their usefulness in the benchmark.122 c. Substantial Evidence Demonstrates That PR II Rates, Other Than the Headline Rate, Are Not ‘‘Too Low’’ As noted supra, one reason the D.C. Circuit vacated and remanded the Determination was because it declined to entertain the argument made only by appellee’s counsel that ‘‘the prior rates had been set far too low, thus negating the usefulness of the prior settlement as a benchmark.’’ Johnson, 969 F.3d at 387. The Judges have noted throughout this Initial Ruling their adoption of the Shapley Value modeling analysis undertaken by the Majority, and raised the headline royalty rate by 44% from 10.5% to 15.1% (as phased-in), rendering moot appellate counsel’s suggestion regarding the rate level. Here, the Judges further consider whether other rates within the PR IIbased benchmark are reasonable, not only because they are part and parcel of the workable structure of that benchmark, but also to determine if they are supported by record evidence. To put this issue in context, those rates would apply on the second prong of the ‘‘greater-of’’ rate structure in the PR IIbased benchmark. The first prong in the PR II-based benchmark rates is the 10.5% revenue rate—increased to 15.1% (as phased-in) by this Initial Ruling. The second prong consists of the ‘‘lesser of’’ a TCC rate or a per subscriber rate.123 122 The Judges also find Copyright Owners’ assertion that they did not know how those rates were established is not credible, given that they and their representatives negotiated those rates. 123 This second prong contains only a TCC rate (i.e., an uncapped rate) for: (1) the ad-supported the service, because there are no subscribers to such a service; and for (2) bundled subscription service, for which there is a $0.25 per month floor but no per- PO 00000 Frm 00035 Fmt 4701 Sfmt 4700 54439 For certain delivery configurations, these rates also cannot fall below any applicable Mechanical Floor. See Johnson, 969 F.3d at 370.124 The Services describe the key feature of these non-headline rates as the fostering of beneficial price discrimination, i.e., the adoption of ‘‘different rate levels for different product offering,’’ in order [t]to account for consumers’ different willingness to pay [WTP] for music. Services’ Joint Opening Brief (on Remand) at 21. As an example of how these price discriminatory rates impacted the market, the Services compare and contrast two Amazon offerings, Amazon Music Unlimited (for Echo) and Amazon Prime Music. Amazon Music Unlimited, with more than 30 million available songs as of the Phonorecords II proceeding period, see Mirchandani WDT ¶ 41, [REDACTED].125 By contrast, Amazon Prime Music, calculated as a ‘‘bundled subscription’’ configuration, makes available only an abridged repertoire of 2 million songs, see Mirchandani, supra, and [REDACTED]. See id. at § 385.13(a)(4). Thus, Amazon pays [REDACTED] for listening by the more casual consumers who use the limited catalog Prime Music service at no additional charge beyond their Prime membership fee, compared to consumers who want the full repertoire provided by Amazon Music Unlimited on their Echo devices. See Services’ Joint Opening Brief at 71. These royalty obligations demonstrate the combination of price discrimination, product differentiation and ‘‘derived demand’’ in action; that is, the [REDACTED] are derived from the lower demand of consumers of the limited Amazon Prime Music service compared with subscribers to Amazon Music subscriber cap, and Service Revenue for such bundles is calculated pursuant to 37 CFR 385.11 (‘‘Service Revenue’’ definition, ¶ 5). 124 As Johnson explained, the CRB Judges ‘‘retained the mechanical floor’’ because, like so much of the PR II-based benchmark, it ‘‘‘appropriately balances the [streaming service providers’] need for the predictability of an All-In rate with publishers’ and songwriters’ need for a failsafe to ensure that mechanical royalties will not vanish[.]’’ Id. at 371–72. It is noteworthy that Copyright Owners urged the Judges (successfully) to maintain the Mechanical Floor provisions, which are the product of the Phonorecords II (and Phonorecords I) negotiations. Thus, it seems apparent that Copyright Owners as well as the Services consider provisions from the negotiated rates and rate structure to be in the nature of benchmarks, although differing as to which elements such be included or excluded. (The Services unsuccessfully argued for the elimination of the Mechanical Floors.) This perspective underscores the correctness of the Judges’ decision on remand to treat the PR II-based benchmark as useful. 125 [REDACTED]. E:\FR\FM\10AUR2.SGM 10AUR2 54440 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations ddrumheller on DSK120RN23PROD with RULES2 Unlimited on their Echo devices, which in turn drive higher revenues. It is also important to note that these differential rates on the second prong of the ‘‘greater-of’’ structure of the PR II Rates are overridden by the revenue percentage rate on the first prong if that first prong rate generates more revenue. For example, [REDACTED], see Dissent at 29 (Table) and 116; see also [REDACTED]. With the headline rate now increased on a phased-in basis, the price discriminatory royalty generated by this [REDACTED]. It is noteworthy that Johnson affirmed the Majority’s setting of other price discriminatory features, e.g., the family and student plan provisions, based on the Judges’ reliance on the Services’ expert testimony regarding the benefits of ‘‘having a way . . . where low willingness to pay consumers can still access music in a way that still allows more monetization of that provision of that service.’’ Johnson, 969 F.3d at 392– 93. In similar fashion, the multi-tiered rates in the PR II-based benchmark likewise were supported by the same type of testimony; indeed, from expert testimony proffered by both parties, as considered below. First, Professor Katz notes that the existing rate structure captures two important aspects of the economics of the interactive streaming market: (1) the variable WTP among listeners; and (2) the corollary variable demand for streaming services. See 3/13/17 Tr. 586– 87 (Katz); see also Marx WRT ¶ 239 et seq.; 4/7/17 Tr. 5568 (Marx) (noting that the present structure serves differentiated products offered to customer segments with a variety of preferences and WTP). In more formal economic terms, Professor Katz notes that the present structure enhances variable pricing that allows streaming services ‘‘to work [their] way down the demand curve,’’ i.e., to engage in price discrimination that expands the market, providing increased revenue to the Copyright Owners as well as the Services. 3/13/17Tr. 701 (Katz). Second, in similar testimony, Professor Hubbard captures the interrelationship between the economics of this market and the existing rate structure: [F]rom an economic perspective, you can think of this market and this industry as being composed of different customer segments by tastes and preferences and willingness to pay. And so no rate structure can really work without understanding that, and no business model can really work without understanding that. [I]n terms of rate structures, the Phonorecords II framework from the previous proceeding does offer a benchmark to start VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 because it provides for differences in distinct product categories in terms of music service offerings, pricing possibilities, and so on. And it has encouraged a very diverse digital music offering set from actual competitors. 3/21/17 Tr. 2175–76 (Hubbard). Moreover, Professor Hubbard [REDACTED] 4/13/17 Tr. 5978 (Hubbard); see also Hubbard WDT ¶ 4.7 (the 2012 rate structure provides the ‘‘necessary flexibility to accommodate the underlying economics of Amazon’s various digital music service offerings.’’). See also 3/15/17 Tr. 1176 (Leonard) (notwithstanding changes and growth in the streaming marketplace over current rate period, underlying economic structure of marketplace, which made percent-of-revenue based royalty appropriate, has not changed). Third, the Services’ experts further assert that the multiple pricing structures necessary to satisfy the WTP and the differentiated quality preferences of downstream listeners relate directly to the upstream rate structure to be established in this proceeding. For example, Professor Marx opines that the appropriate upstream rate structure is derived from the characteristics of downstream demand. 3/20/17 Tr. 1967 (Marx) (agreeing that rate structure upstream should be derived from need to exploit willingness to pay of various users downstream via percentage of revenue because downstream listeners have varying willingness to pay that should be exploited for mutual benefit of copyright licensees and licensors). Professor Marx further acknowledged that this upstream:downstream consonance in rate structures represents an application of the concept of ‘‘derived demand,’’ whereby the demand upstream for inputs is dependent upon the demand for the final product downstream. Id. Moreover, Dr. Leonard notes that reliance on the Services to identify segmented demand and develop price discriminatory approaches is appropriate because ‘‘the downstream company is going to have a lot more information about . . . the business, about what makes sense.’’ 4/ 6/17 Tr. 5238 (Leonard). Regarding a comparison of revenue growth to streaming growth, Professor Hubbard dismisses as economically ‘‘meaningless’’ Copyright Owners’ argument that they have suffered relative economic injury under the current rate structure simply because the increase in their revenues from interactive streaming has been proportionately less than the growth in the number of interactive streams, leading mathematically to a lower implicit or effective per stream royalty PO 00000 Frm 00036 Fmt 4701 Sfmt 4700 rate. That is, he notes there is no evidence to rebut this prima facie indication of beneficial price discrimination, i.e., no contrary evidence indicating that, if the Services had sought to increase the price of the services available to these low to zero WTP listeners because of higher royalties, they would have paid the higher price, rather than declined to utilize a royalty-bearing interactive streaming service. See 4/13/17 Tr. 5971– 73 (Hubbard); see also Dissent at 52. The Services also link their price discrimination argument to the fact that the marginal physical cost of streaming is zero to the need for a flexible rate structure such as now exists. In this regard, Professor Hubbard notes that, because ‘‘[t]he marginal production cost at issue here is—is zero. . . . it’s not clear why it’s not better to bring new customers into the market on which royalties would be paid and, of course, zero marginal cost incurred.’’ 4/13/17 Tr. 5917–18 (Hubbard). See also Marx WDT ¶ 97 (‘‘Setting the price of marginal downstream listening at its marginal cost of zero induces more music consumption and variety than per-song or per-album pricing.’’). Professor Marx makes the same argument as to the salutary nature of price discrimination in this context with regard to Spotify’s ad-supported approach. Focusing on the first purpose, Spotify is attracting ad-supported listeners who have a relatively low WTP, whether they have low incomes, (a budget constraint) or low interest in music (low ‘‘utility,’’ in the parlance of economists). These listeners, and the advertising revenue they generate are real and reflect the WTP of a large swath of all interactive listeners. See Marx WRT ¶ 115–16 & Fig. 9 (‘‘While I agree that one aspect of the ad-supported service is to provide an on-ramp to paid services, it also has another important aspect, namely to serve low WTP customers. . . . Copyright Owners’ economists err in not calculating the impact of the Copyright Owners’ proposal on ad-supported services. Adsupported services currently make up a majority of subscribers and [REDACTED]% of all streams in the industry.’’). Accordingly, a separate tier for an adsupported service accounts for the different nature of the downstream listenership, allowing the upstream royalty to be based on that characteristic. This differentiation was essentially acknowledged by Copyright Owners late (too late, actually) when they proposed in their post-hearing filing that ‘‘if the Judges intend to include the Spotify ad-supported E:\FR\FM\10AUR2.SGM 10AUR2 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations service in the rate structure and rate calculations, that they do so by establishing separate rates and terms for the ad-supported service. See COPCOL (Corrected) ¶ 228 & n.34. But the PR IIbenchmark already incorporates separate rates for free/ad-supported services!126 Another important evidentiary factor buttressing the need for price discriminatory rates and structures was the testimony of the Services’ survey expert, Mr. Robert Klein, Chair and cofounder of Applied Marketing Systems, Inc. Mr. Klein surveyed 2,101 people (the Klein Survey) who were listeners to streamed music and found, inter alia, that: (1) the majority of listeners would not pay for a monthly streaming subscription; and (2) for those who do subscribe, their demand was elastic, with increases in subscription prices causing overall greater percentage reductions in quantity demanded, moving customers to free, ad-supported and non-streaming alternatives. See Klein WRT ¶¶ 60–67. By contrast, Copyright Owners did not present any survey testimony. The Determination fully credited the Klein Survey, finding as follows: ddrumheller on DSK120RN23PROD with RULES2 It is important to note that Copyright Owners’ attacks on the Klein Survey are not levelled by any witnesses, nor contradicted by their own survey expert, because Copyright Owners elected not to proffer such an expert in their direct (or rebuttal) cases. Rather, Copyright Owners elected to make a descriptive argument regarding the elasticity of demand among different segments of the market, as opposed to a survey-based or econometric study of price elasticity. [Although] Copyright Owners attack the Klein Survey on several fronts[,] [t]he arguments made by Copyright Owners are insufficient . . . to seriously weaken the probative value of the Klein Survey. In the end, the Judges are not persuaded by the Copyright Owners’ revenue bundling arguments not to adopt a flexible, revenuebased royalty rate. Determination at 22–23 & n.53; see also Dissent at 64–67 (including point-bypoint rejection of Copyright Owners’ non-expert criticisms of Klein Survey). The Services also note that the existing rate structure has produced generally positive practical consequences in the marketplace. Their joint accounting expert, Professor Mark Zmijewski, testified that the [REDACTED] from the sale of product under (former) Subpart A since 2014 has been [REDACTED] over the same period. Expert Report of Mark E. 126 Copyright Owners also belatedly proposed that the Judges establish specific functionality limits on a separate ad-supported prong to avoid cannibalization of subscriber-based streaming with fuller functionality. Id. [REDACTED]. VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 Zmijewski February 15, 2017 ¶¶ 38, 40 (Zmijewski WRT); 4/12/17 Tr. 5783 (Zmijewski); see also 4/13/17 Tr. 5897 (Hubbard) (‘‘the evidence that I reviewed suggests that the copyright holders have actually benefitted from this structure. . . .’’). More particularly, Professor Zmijewski testified that: • Total revenues reported by the NMPA for NMPA members from all royalty sources [REDACTED]. Zmijewski WRT ¶ 41. • This [REDACTED]. Id. • The [REDACTED]. Id. • Mechanical royalty revenue for the sale of downloads and physical phonorecords [REDACTED]. Id. ¶ 38.127 In sum, the foregoing analysis demonstrates the economic reasonableness and appropriateness of the price discriminatory Phonorecords II rate structure and its negotiated safeguards to address the real possibility of revenue diminution. As discussed below, the record evidence also supports royalty rates within the PR IIbased benchmark.128 127 By contrast, Copyright Owners assert that the appropriate approach would only consider interactive service payment of mechanical royalties, and exclude performance royalties. On that basis, revenue, for the sale of digital downloads and physical phonorecords mechanical royalty revenue [REDACTED] from [REDACTED] in 2014 to [REDACTED] (as noted in (4) above, whereas mechanical royalty from streaming [REDACTED] from [REDACTED] in 2014 to [REDACTED] in 2015. Thus, the [REDACTED] in mechanical royalty revenue from streaming [REDACTED] in mechanical royalty revenue from the sale of digital and physical phonorecords. The Judges do not agree with Copyright Owners. Performance royalty and mechanical royalty payments made by the Services are for perfect complements—neither license has any value to the Services unless they acquire both. Indeed, that is a critical reason why the mechanical rate is calculated on an ‘‘All-In’’ basis. Thus, it makes sense to make the comparison in the manner undertaken by Professor Zmijewski. 128 Again, to be clear, the Judges are substituting the 15.1% revenue rate for the 10.5% revenue rate as the headline rate in the ‘‘greater-of’’ structure of the Phonorecords II benchmark. Thus, the price discriminatory royalty rates discussed below would apply only if they generated a ‘‘greater’’ level of revenue than the headline 15.1% revenue rate. And, although the Mechanical Floor rate is not tied directly as an alternative to the ‘‘greater-of’’ revenue rate (now 15.1% as phased-in), it is not a floor that ignores the effect of that ‘‘greater of’’ rate. For example, assume the popular standalone portable subscription streaming service that people access on their mobile phones would pay an ‘‘All-In’’ musical works royalty of 15.1% based on the application of the two ‘‘greater-of’’ prongs. However, assume also the ‘‘Performance Royalty’’ that must be subtracted is 12%. That would leave 3.1% of service revenue attributable to the mechanical right. However, if that revenue rate of 3.1% yielded mechanical royalty revenue that was less than the royalty revenue generated by the applicable monthly mechanical floor of $0.50 per subscriber, then the mechanical floor would control. This application, like any other application of the mechanical floor, does not diminish the value of the 15.1% right, but rather limits its reduction PO 00000 Frm 00037 Fmt 4701 Sfmt 4700 54441 The PR II-based benchmark contain several alternate rates explicitly calculated as a percentage of payments made by interactive streaming services to the record companies for sound recording rights. See Addendum to this Initial Ruling. In the Subpart relating to streaming, the (former) subpart B category, the TCC is 22% for adsupported services and 21% for portable subscriptions. Id.; see also 37 CFR 385.13(b)(2) and (c)(2). These percentage figures correspond to sound recording: musical works royalty ratios of 4.55:1 and 4.76:1, respectively. With regard to these ratios, Copyright Owners’ economic expert witness, Dr. Eisenach, stated: ‘‘In my opinion, the evidence . . . indicates that the relative valuation ratios implied by the current Section 115 compulsory license . . . represent an upper bound on the relative market valuations of the sound recording and musical works rights.’’ Id. ¶ 92 (emphasis added). (As an ‘‘upper bound,’’ these ratios would represent the lower bound on the relative market valuations of the reciprocal percentage of the value musical works rights relative to sound recording rights, again, 22% and 21%.129) Thus, there appears to be consensus between Copyright Owners’ witness and the Services (who advocate for applying these rates on the price discriminatory tier of their benchmark) that these rates constitute ‘‘relative market valuations’’ (even if they are not Dr. Eisenach’s preferred market valuations within the bounded zone of such values). Dr. Eisenach’s testimony regarding the ‘‘bounds’’ of useful market valuations is noteworthy because his acknowledgement is consonant with judicial precedent. The Judges’ setting of reasonable rates often requires them to identify a ‘‘zone of reasonableness,’’ within which they identify appropriate statutory rates. See, e.g., Intercollegiate Broadcasting System, Inc. v. Copyright Royalty Board, 684 F.3d 1332, 1340 (D.C. Cir. 2012) (The CRB Judges’ rate setting can necessitate the finding of a ‘‘zone of reasonableness [because] ‘‘[s]tatutory reasonableness is an under the ‘‘All-In’’ calculation. Recall also that the Determination, Dissent and Johnson do not disturb the All-In and Mechanical Floor features of the Phonorecords II benchmark.) And finally, with regard to the actual per subscriber monetary values in the mechanical floors, no party suggested changes from rate levels in the PR II-based benchmark, including in the mechanical floor rates. The Judges recognize, as did Dr. Katz, Pandora’s economic expert witness, that alternate values might have been preferable for rates contained in the PR II-based benchmark, but none were in the record. See 4/15/17 Tr. 5056–58 (Katz). 129 1 ÷ 4.55 = .219, or 22% (rounded); 1 ÷ 4.76 = .210 (21%). E:\FR\FM\10AUR2.SGM 10AUR2 54442 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations ddrumheller on DSK120RN23PROD with RULES2 abstract quality represented by an area rather than a pinpoint.’’). The 21% and 22% TCC rates within section 115 identified by Dr. Eisenach as generating the ‘‘lower bound on the relative market valuations’’ imply certain approximate percent-of-revenue rates, i.e., percent of total service revenue (not percent of sound recording revenue). See Dissent at 91, n.133 (sound recording rates clustered between [REDACTED]% and [REDACTED]% of revenue). For example, if the sound recording royalty rate for interactive streaming is [REDACTED]% of revenue, then the musical works rate would be calculated as 0.21 × [REDACTED], which equals [REDACTED]%, (or as .22 × [REDACTED] which equals [REDACTED]%). At the low end of the range, if the sound recording royalty rate is [REDACTED]%, then, applying these TCC figures, the implied musical work royalty rate would be calculated as [REDACTED]% (.21 × [REDACTED]) or [REDACTED]% (.22 × [REDACTED]).130 It is important to emphasize and detail the context of these price discriminatory rates. These capped TCC rates are on the ‘‘greater of prong’’ that is compared with the headline 15.1% revenue rate (phased-in) that the Judges are also adopting in this Initial Ruling. As phased in, the headline rate is greater than all the capped TCC-based rates identified in Dr. Eisenach’s testimony, supra, [REDACTED]. For 2019, the phased-in headline percentage rate, 12.3%, is [REDACTED] the [REDACTED]% and [REDACTED]% revenue rates derived if the sound recording rates was [REDACTED]%. For 2018, the phased-in headline percentage rate, 11.4%, is [REDACTED] all the rates derived from the capped TCC rates Dr. Eisenach identified as ‘‘market valuations’’ (albeit the lower bound in his opinion). But that is of no negative consequence for Copyright Owners, 130 Dr. Eisenach’s identification of the 21%–22% TCC as within the bounds of market valuations may appear surprising at first in light of the higher 26.2% uncapped TCC rate pursued (unsuccessfully) on remand by Copyright Owners. But in the context of his testimony, Dr. Eisenach’s opinion is understandable. The former headline rate of 10.5%, when sound recording rates ranged from approximately [REDACTED]% to [REDACTED]% of streaming revenues, yielded TCC rates between [REDACTED]% and [REDACTED]%. Thus, Dr. Eisenach was identifying a market valuation [REDACTED] (at his lower bound) between [REDACTED]% (the difference between 21% and [REDACTED]%) and [REDACTED]% (the difference between 22% and [REDACTED]%). Again, for context, this Initial Ruling raises the percentage rate by 44% when fully phased-in (based on the experts’ Shapley analyses, significantly above the TCC rates advocated by Dr. Eisenach, even assuming the [REDACTED]%–[REDACTED]% sound recording rates on which he relied. VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 because they would get paid on the ‘‘greater-of’’ metric (capped TCC or headline rate) under the Phonorecords II-based rate structure the Judges are adopting (For the portable subscriptions, even though the 80 cents/ subscriber ‘‘lesser-of’’ portion of the non-headline prong would apply on that prong if it was lower than the capped TCC rate, the actual rate could not be lower than the phased-in headline rate.) Dr. Eisenach also examined direct agreements between record companies and interactive streaming services that contain rates for sound recordings and mechanical royalties, respectively. See, e.g., id. ¶¶ at 84–91. In such cases, the ratio of sound-recording to musicalworks royalties ranged tightly between [REDACTED] and [REDACTED], closely tracking the regulatory ratios implicit in the section 115 TCC. Id. ¶ 92. (The [REDACTED] ratio equates to a TCC rate of [REDACTED]%, and the [REDACTED] ratio equates to a mechanical rate of [REDACTED]%.). He concluded, as he did with regard to the actual section 115 license rates: ‘‘In my opinion, the evidence presented . . . indicates that the relative valuation ratios implied by the . . . negotiations under [the statutory] shadow—ranging from [REDACTED] [[REDACTED]%] to [REDACTED] [[REDACTED]%]— represent an upper bound on the relative market valuations of the sound recording and musical works rights.’’ Eisenach WDT ¶ 92 (emphasis added). Dr. Eisenach also identified several additional useful benchmarks. First, he identified what was coined the ‘‘Pandora Opt-Out Agreement’’ benchmark,131 which reflected a ratio of 131 Pandora was only a noninteractive service at that time, and thus only paid the performance right royalty, not the mechanical right royalty, for the right to use musical works. Because the parties agree that the performance right and the mechanical right are perfect complements, Pandora’s payments for the performance right are thus relevant and probative, as they reflect the full value of the musical works royalty to a noninteractive service. These factors became relevant because major music publishers had negotiated direct licensing agreements with Pandora for its noninteractive service covering the period from 2012 through 2018. Eisenach WDT ¶ 103. They negotiated these direct agreements after certain publishers had decided to ‘‘opt-out,’’ i.e., to withdraw their digital music performance rights from PROs, and asserted the right to negotiate directly with a digital streaming service. Pandora thus negotiated several such ‘‘Opt-Out’’ Agreements with an understanding that the rates contained in those direct agreements might not be subject to rate court review and thus could reflect market-based rates. Given this unique circumstance, and given that the markets and parties involved in the Pandora Opt-Out agreements are somewhat comparable to the markets and parties at issue in this proceeding, Dr. Eisenach concluded that these agreements provided ‘‘significant insight into the relative value of the sound recording and musical works rights in this proceeding.’’ Id. (emphasis added). (The Judges did PO 00000 Frm 00038 Fmt 4701 Sfmt 4700 [REDACTED] of sound-recordings to musical-works in a comparable benchmark setting. This ratio translates to a TCC percent of [REDACTED]%. With sound recording royalty rates of approximately [REDACTED]% to [REDACTED]%, this TCC reflects an effective percentage of total revenue equal to [REDACTED]% to [REDACTED]%. Second, Dr. Eisenach identified YouTube agreements with music publishers that relate to the combination of a commercial sound recording and a ‘‘static image.’’ The YouTube agreements contain an explicit royalty of [REDACTED].132 That [REDACTED]% royalty is a denominator in the ratio concept utilized by Dr. Eisenach, and the numerator is the [REDACTED] sound recording royalty paid to the record companies. YouTube had agreed to pay [REDACTED]% of its revenues, and had agreed to pay [REDACTED] and other record companies [REDACTED]% of revenues. The [REDACTED] ratio reduces to [REDACTED], implying a TCC ([REDACTED]) of [REDACTED]%. The [REDACTED] ratio reduces to [REDACTED], implying a TCC ([REDACTED]) of [REDACTED]%. See Dissent at 101–102. These additional rates identified in Dr. Eisenach’s testimony further confirm the reasonableness of the non-headline rates within the PR II-based benchmark. Finally, the Judges look at the effective rates paid by Spotify, the largest interactive streaming service in terms of in terms of the number of subscriber-months and the number of plays. See Marx WRT ¶¶ 37–38 & Figs. 8 & 9. Under the PR II based benchmark, Spotify paid on its subscription service an effective ‘‘All-In’’ royalty rate of [REDACTED]% of its total revenues. See Dissent at 80, 115, 149 (and record citations therein). Spotify paid this effective percent-of-revenue rate [REDACTED]. See id. at 29 (Table). Turning to Spotify’s free/adsupported offering (and as noted supra), Spotify paid royalties under the PR II Rates at an effective ‘‘All-In’’ royalty rate of [REDACTED]%. Spotify paid this effective percent-of-revenue rate [REDACTED]. See id. When Spotify’s two tiers are blended and averaged, the effective percent-of-revenue rate is [REDACTED]% of revenue. See id. at 116. The average rate has salience in this proceeding because Spotify’s two not adopt Dr. Eisenach’s speculation that this performance royalty would continue to grow after 2018. See Determination at 51; Dissent at 102–103.) 132 Dr. Eisenach preferred to use YouTube agreements that included [REDACTED], but the Judges relied on [REDACTED] as more comparative. Determination at 50; Dissent at 102. E:\FR\FM\10AUR2.SGM 10AUR2 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations tiers are interrelated, in that free/adsupported listeners constitute a pool of potential converts to the subscription tier under this ‘‘freemium’’ model, even as this offering generates royalties under the PR II-based benchmark. ddrumheller on DSK120RN23PROD with RULES2 d. Copyright Owners’ Concern Regarding Revenue Diminution Is Insufficient To Reject the PR II-Based Benchmark Copyright Owners argue that what the Services tout as beneficial price discrimination generates an ‘‘incredible’’ level of revenue diminution, including displacement, resulting in a ‘‘major problem’’ that reduces reportable revenues and thus the royalty base. See, e.g., 3/7/22 Tr. 193 (Copyright Owners’ counsel). This argument is based upon documents and evidence that demonstrated the following: • [REDACTED]; • [REDACTED]. • [REDACTED]; • [REDACTED]; • [REDACTED]; • [REDACTED] • [REDACTED]; • [REDACTED]; • [REDACTED]; and • Copyright Owners’ expert, Dr. Rysman, testified that interactive services often elect to forgo current profit maximization, e.g., by charging lower prices, in order to build a customer base and greater long-run profitability or value, from selling music and non-music products or services to its customers. CO Initial Submission at 40–42 (and record citations herein). The Services’ economic experts do not ignore the fact that there can be revenue attribution problems when interactive streaming is combined with other products or services. They acknowledge that, even absent any wrongful intent with regard to the identification and measurement of revenue, attribution of revenue across product/service lines of various services can be difficult and imprecise. See, e.g., 4/5/17 Tr. 5000 (Katz) (problem of measuring revenue ‘‘certainly a factor that goes into thinking about reasonableness.’’). However, Professor Katz testified that the existing rate structure agreed to by the parties accommodates these bundling, deferral, and displacement issues via the use of an alternative rate prong that would be triggered if the royalty revenue resulting from the headline rate of 10.5% of streaming revenue fell below the royalty revenue generated by that second prong. Katz VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 WDT ¶¶ 82–83; 3/13/17 Tr. 670 (Katz). Moreover, Professor Katz concluded that, because the marketplace appears to be functioning (in the sense that publishers are earning profits and new and existing interactive streaming services continue to operate despite accounting losses), these revenuemeasurement issues are being adequately handled by the alternative rate prong, even if an altered second prong might work better. Id. at 738–39. More generally, Professor Katz further noted that, the existing rates within the PR II-based benchmark were performing well, and even if alternative minima might be preferable, no such alternative rates were in the record. See 4/15/17 Tr. 5056–58 (Katz) (under the PR II-based benchmark ‘‘the industry . . . was performing well,’’ but ‘‘if someone had a proposal [with] a specific reason why we should adjust this minimum that’s something I would have examined,’’). But Copyright Owners did not propose alternative rates or minima within the PR II-based benchmark, but instead urged the Judges to disregard the benchmark writ large. Accordingly, there were no alternative rates or minima in the record. Professor Katz further noted that the PR II-based benchmark rates were established when ‘‘ecosystem’’ entities such as Yahoo—akin to Amazon, Apple, and Google—were in the marketplace. 4/5/17 Tr. 5055–57 (Katz); see also Determination at 31 (and record citations therein) (noting the presence of Microsoft as well as Yahoo as licensees in the interactive market during the Phonorecords II negotiations). More broadly, the Services’ position regarding the use of the two prongs and their alternate rates to ameliorate the revenue-measurement problems is summed up by Professor Katz as follows: [T]he primary reason [for the two rate prongs] . . . is because of the measurement issues that can come up when having royalties based on a . . . percentage of revenues because there can be issues about how to appropriately assign revenues to a service. And so I think the minim[a] can play an important role when those—you know, when those measurement problems are severe, you can turn to the minimum instead. . . . [W]hat I have in mind, right, is that what would happen if you could imagine an entrepreneur coming along and saying we want to have a service and have some incredibly low price and not a very good monetization model, where a copyright owner would say—in an effectively competitive market, would say, wait a minute, I don’t want to license to you on those terms. It’s—I just think the possibility of getting a return is so low, I’m not going to do it, even though you, as an entrepreneur, PO 00000 Frm 00039 Fmt 4701 Sfmt 4700 54443 are willing to try this. I as the copyright owner want some sort of, you know, return on it. And that’s what the minimum also helps to do. 3/13/17 Tr. 599 (Katz.); see also 3/20/17 Tr. 1900–01 (Marx) (minima protect against revenue measurement problems); 4/7/17 Tr. 5584 (Marx) (statutory minima play ‘‘two roles’’— protecting the Copyright Owners from ‘‘revenue mismeasurement’’ by creating the ‘‘greater of’’ prong,’’ but incorporating per subscriber rate prong in ‘‘lesser of’’ component to protect services from the record companies’ use of their market power to engage in ‘‘manipulation of the sound recording royalties’’ on which the TCC prong is calculated). After considering the record, the Judges determine that the Majority had not found—as Copyright Owners claim—that the activities and strategies by the Services were ‘‘incredible’’ or a ‘‘major problem. Rather, the Majority’s characterization was measured, stating repeatedly that the Services engaged ‘‘to some extent’’ in revenue diminution because they ‘‘might focus on long-term profit maximization to promote their long-term growth strategy, which occurs ‘‘even absent wrongful intent.’’ Determination at 20–21, 36, 90; accord, Dissent at 59. In fact, the Majority specifically stated: ‘‘The Judges agree that there is no support for any sweeping inference that cross-selling has diminished the revenue base.’’ Id. at 21 (emphasis added). The Majority (and the Dissent) thus acknowledged the reasonableness of both sides of this issue, recognizing both the Services’ use of price discriminatory approaches that can lower per user or per-stream revenues but grow royalties, market share and revenue, as well as Copyright Owners’ concomitant desire to protect themselves from reductions in the royalty revenue base, however limited in extent, that would only serve to diminish royalties. One way the input supplier can avoid this impact is to refuse to accept a percent of revenue form of payment and move to a fixed per-unit input price. This is what Copyright Owners originally and unsuccessfully sought in this proceeding, subject to a bargaining room approach by which they could switch back to the old approach (or any other approach) through purely marketbased negotiations, unbounded by the statutory and regulatory standards of ‘‘fairness’’ and ‘‘effective competition.’’ See Dissent at 60. The Judges must reconcile the parties’ competing considerations. A way by which they are both accommodated is through a pricing structure with E:\FR\FM\10AUR2.SGM 10AUR2 54444 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations alternate rate prongs and floors, below which the royalty revenue cannot fall. This is precisely the bargain struck between Copyright Owners and services in 2008 and 2012, and that has been the rate structure through 2017. And, because the Majority and the Dissent found that revenue diminution occurred only ‘‘to an extent,’’ rather than in the pervasive (sweeping’’) manner averred by Copyright Owners, there is no sufficient reason in the record to depart from the bargained-for multi-tiered rate structure in Phonorecords II that allows for price discrimination but tempers its impact on royalties through the use of minima and floors. e. Copyright Owners’ Claim of ‘‘Inherent’’ Economic Value Is Belied by the Record, Including Their Own Arguments Pre-remand, Copyright Owners approached this rate setting process with an overarching premise: A musical work has an ‘‘inherent value’’ that must be reflected in the royalty rates. As the NMPA’s president, Mr. Israelite testified, when asked how ‘‘inherent’’ value is defined: [W]hoever owns an individual copyright is the one to define it. I think that would be the most appropriate definition of it. What someone is willing to license it for would be that inherent value to that owner . . . That would be market value. ddrumheller on DSK120RN23PROD with RULES2 3/29/17 Tr. 3707 (Israelite). If the market for musical works was as atomistic as the above quote assumes, the songwriter of an individual musical work could indeed set his or her own royalty rate, and refuse to license to any streaming service or other distributor who refused to pay that royalty. But that is not how the licensing market works.133 Songwriters typically assign their licensing rights to music publishers (to avoid ruinous transaction costs). These music publishers control huge ‘‘Must Have’’ repertoires that are offered under blanket licenses to streaming services. (The musical works market of course is subject to a compulsory license, but this is precisely how the unregulated market works for the licensing of sound recordings by labels to interactive streaming services.) 133 The record does not include evidence of selfmarketing by songwriters through social media or via negotiation of individual royalty contracts by the exercise of overwhelming star power, whether through traditional payment mechanisms or new methods, such as the murky mechanism of nonfungible tokens (NFTs). The absence of incidents of such self-marketing from the record evidence in this proceeding suggests that they likely constitute but a small segment of the songwriter/publisher market. Accordingly, such self-marketing and individual negotiations do not impact the Judges’ setting of statutory rates in this proceeding. VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 It is acknowledged even by Copyright Owners’ own expert witness, Professor Watt that the creation of these large collectives generates market power that necessitates rate regulation. See R. Watt, Copyright and Economic Theory: Friends or Foes at 163, 190 (2000) (quoted in Dissent at 35). Further, this ‘‘inherent’’ market value notion is antiquated as a matter of economics. Although an individual Copyright Owner can announce his or her ‘‘asking’’ royalty, that is not sufficient to generate a ‘‘market’’ royalty, unless and until a licensee agrees to pay it. In market-based economics. that is to say, the economic consensus that has governed economics since the ‘‘marginal revolution’’ in the mid to late 19th century, value is ascertained through the intersection of supply and demand, with the price established at the margin representing the market value of the good or service bought and sold.134 If there is no demand for a product, be it a musical work or anything else, it has no economic value. Even though costs have been incurred to produce the product, those costs cannot be recovered (or profit earned) absent a sufficient WTP in the market. And, as noted supra, the product being offered and at issue here is comprised of ‘‘second copies’’ of sound recordings (with embedded musical works), which are costless to reproduce for streaming purposes. Of course, these ‘‘second copies’’ do have actual value when they are in demand, and the royalties that their licensing 134 As one scholar has summarized the 19th century transition from classical to neoclassical economics: ‘‘By the early 1870s, economics reached a tipping point, and it ushered in a revolution in thought, signaling the beginning of the ‘‘modern,’’ or ‘‘neoclassical’’ era. Marginalists flipped classical economics on its head. Instead of focusing on the production side of economics, they turned to consumption. It is the satisfaction of the wants of consumers that matters for value, not the labor required for production. What established the overall value of a good is the value fetched by the final unit of that item on the market. As more units of a good are produced, the marginal value of the last unit tends to decrease. . . . According to marginal utility, the consumer, not the producer, therefore drives the valuation process.’’ J. Wasserman, The Marginal Revolutionaries at 28 (2019). This transformation reflected the abandonment of the ‘‘labor theory of value’’—the cornerstone of Marxian economics. See E.R. Canterbury, A Brief History of Economics at 111 (2001) (‘‘Marx’s devotion to a labor theory of value was complete.’’). It initially appears as irony that Copyright Owners espouse a Marxian approach to value while preaching the virtues of unregulated markets. The initial whiff of irony dissipates when one appreciates that a collective licensor with the market power of control over a ‘‘Must Have’’ input has every incentive to urge a pricing or valuation method that takes the focus away from the force of consumer demand in an effectively competitive market, which is a hallmark of neoclassical economics. PO 00000 Frm 00040 Fmt 4701 Sfmt 4700 generates must cover: (1) the first copy (creative) costs; (2) the ‘‘opportunity cost’’ (measured by the next best alternative for royalty earnings if the ‘‘second copies’’ could have been supplied through another distribution channel that paid higher royalties to attract the end-user/consumer at issue); and (3) profits to induce the creation of musical works. Second, the fact that Copyright Owners originally proposed a persubscriber alternative rate to their perplay rate itself belies their conviction that some ‘‘inherent’’ economic value exists. When the metric of value switches from ‘‘per-play’’ to ‘‘persubscriber,’’ the focus of value likewise shifts from an emphasis on producer value to consumer value. That is, if there is truly an ‘‘inherent’’ value for a product or service, that singular value cannot divide into two distinct values with the ‘‘greater-of’’ the two controlling. Such an argument gives away the game, so to speak, demonstrating, perhaps unsurprisingly, that economic arguments (not unlike legal advocacy) are often situational— designed to support maximalist positions and the exercise of market power, however acquired. See also Determination at 28 n.64 (rejecting the ‘‘inherent value’’ argument). f. PR II-Based Benchmark Not ‘‘Too Complex’’ Copyright Owners and the Majority complained that the PR II-based benchmark is too complex. See Copyright Owners’ PFF ¶ 12 (criticizing complexity of PR II Rates as lacking ‘‘transparency’’); Determination at 36 (characterizing parties’ negotiated, renewed, and expanded rate structure as Rube-Goldberg-esque in complexity and impenetrability.’’) After considering this issue on remand, the Judges disagree. If some songwriters or lyricists have been confused by their royalty statements, their confusion of course should be resolved. However, one of the benefits of a collective is that it possesses the expertise and resources to identify and explain how royalties are computed and distributed. Moreover, this claim of complexity cannot serve as a basis to override the multi-part negotiated benchmark that the parties, through their respective trade associations, negotiated and implemented. As the Dissent stated: ‘‘There is no good reason why the rate structure that is consonant with the parties’ ten-year history and with the relevant economic model should be sacrificed on the slender E:\FR\FM\10AUR2.SGM 10AUR2 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations ddrumheller on DSK120RN23PROD with RULES2 argument that ‘‘simpler is better than complicated.’’ Dissent at 88.135 Further, section 801(b)(1) does not identify ‘‘simplicity’’ as a statutory goal for the setting of rates, rate structure, and terms. Although there is certainly no need for gratuitous complexity, the price discriminatory structure and the associated levels of rates in the PR IIbased benchmark that were eliminated by the Majority (while maintaining all the remaining complexity) were most certainly not gratuitous, but rather designed, after negotiations, to establish a structure that would expand the revenues and royalties to the benefit of Copyright Owners and Services alike, while also protecting Copyright Owners from potential revenue diminution by the Services. Moreover, when the market itself is complex—in that the WTP across consumer groups is heterogeneous and the offerings reflect that fact—it is unsurprising that the regulatory provisions would resemble the complex terms in a commercial agreement negotiated in such a setting. For the Judges to demand simplicity in this context would be to sacrifice the specificity that an effectively competitive market requires. See Dissent at 88 (rejecting the simplicity argument by invoking the advice attributed to Albert Einstein that ‘‘[e]verything should be made as simple as possible, but no simpler.’’ g. So-Called Statutory ‘‘Shadow’’ Does Not Diminish Value of the PR II-Based Benchmark Rates Copyright Owners maintain that the rates in the PR II-based benchmark are infirm because, like any benchmark for which a statutory rate is the default, they are not actual market rates. That is, such a rate is said to exist in the socalled ‘‘shadow’’ of the statutory rate. See Dissent at 70 (and citations therein). The Judges reject this argument for several reasons. First, the argument is undercut by the explicit language of section 115 of the Copyright Act, which states: ‘‘In addition to the objectives set forth in section 801(b)(1), in establishing such rates and terms, the Copyright Royalty Judges may consider rates and terms under voluntary license agreements described in subparagraphs (B) and (C).’’ 17 U.S.C. 115(c)(3)(D). Subparagraphs (B) and (C), respectively, refer to agreements on ‘‘the terms and 135 Copyright Owners’ concern for transparency has apparently evaporated in connection with its eagerness to adopt the proffered uncapped TCC rates. Under that approach, the definition of revenue, the handling of bundled products and the exclusion of certain consideration from the royalty base will remain opaque to songwriters—and to the Judges. VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 rates of royalty payments under this section’’ by ‘‘persons entitled to obtain a compulsory license under [17 U.S.C. 115(a)(1)]; and ‘‘licenses’’ covering ‘‘digital phonorecord deliveries.’’ Id. Thus, it is beyond dispute that Congress has authorized the Judges, in their discretion, to consider such agreements as evidence, irrespective of—or perhaps because of—the shadow cast by the compulsory license. Thus, the appropriate question is how much weight the Judges, in their discretion, should afford such benchmarks in any particular proceeding. There is no basis to find, as Copyright Owners suggest, that statutorily-based or influenced benchmarks, including specifically the PR II-based benchmark in this proceeding, are per se inferior to other benchmarks or alternative economic evidence (e.g., from models, surveys or experiments) that may be unaffected by the shadow. Those other benchmarks or forms of evidence will also be subject to their own imperfections and incompatibilities with the target market and must be identified and weighed accordingly.136 Thus, the Judges must not only consider (i) the importance, vel non, of any potential so-called ‘‘shadow-based’’ distortionary effects from a benchmark derived from a regulated statutory benchmark market, but also (ii) how any such purported ‘‘shadow’’ effects compare to any distortions generated by other proffered benchmarks and competing alternative economic evidence, e.g., distortions based on complementary oligopoly power, bargaining constraints and product differentiation in other benchmarks, models, surveys or experiments.137 The Services’ experts discount the foregoing shadow-based criticism. Moreover, the Services laud a statutorily-influenced benchmark in general, and the specific PR II-based benchmark in particular, because the latter reflects more equal bargaining power between licensors and licensees. In this regard, one of the Services’ 136 It has been famously and wisely said that ‘‘all models are wrong, but some are useful.’’ G. Box & N. Draper, Empirical Model-Building at 424 (1987). Benchmarks, Shapley, and Nash models, surveys and experiments are all models, in that ‘‘[a] model is a representation of something beyond itself . . . being used as a representative of that something, and in prompting questions of resemblance between the model and that something . . . substitute systems . . . directly examined . . . to indirectly acquire information about their target systems.’’). U. Maki, Models are Experiments, Experiments are Models, 12 J. Econ. Meth. 303 (2005). 137 It is also important to note that the reasonable rate and rate structure identified under the section 801(b)(1) standard (before considering the four itemized statutory factors) need not be a marketbased rate, as discussed infra. PO 00000 Frm 00041 Fmt 4701 Sfmt 4700 54445 economic expert witnesses, Professor Katz, points out that rates set voluntarily by the parties in a settlement under the ‘‘shadow’’ provide two important benefits. First, with a statutory rate-setting proceeding as a backstop, large licensors cannot credibly threaten to ‘‘hold out’’ and ‘‘walk away’’ from the negotiations without an agreement, thereby negating their ability to use their ‘‘must have’’ status to obtain rates above effectively competitive levels. Second, when, as here, such negotiations are conducted with all the parties at the figurative table—including here, trade associations—no single party has disproportionate market power in the negotiations. See 3/13/17 Tr. 661 (Katz). The Judges agree that settlement agreements reached in the statutory shadow are useful. Although imperfect when considered in isolation, in that the statutory proceeding is the default backstop, in context they negate the power of any entity simply to refuse to strike a deal. The negation of that power blunts the complementary oligopoly power of licensors of ‘‘Must Have’’ repertoires (whether musical works or sound recordings), making a benchmark agreement reached in the so-called ‘‘shadow’’ advantageous in establishing an effectively competitive rate. See Web IV, supra, 26,316, 26,330–31 (May 2, 2016) (noting counterbalancing effect of statutory license in establishing effectively competitive rates). Further, when such settlement agreements are industrywide, they tend to eliminate disproportionate market power, See Dissent at 72; Web III, 79 FR 23102, 23111 (Apr. 25, 2014), aff’d Intercollegiate Broad. Sys., Inc. v. Copyright Royalty Bd., Case No. 14– 1098 (D.C. Cir. Aug. 11, 2015) (relying on two settlement agreements). Nonetheless, Copyright Owners are correct to note that, hypothetically, some licenses might have otherwise been negotiated at rates higher than the settlement rate that was affected by the so-called shadow. But that is simply the tradeoff that the statutory scheme makes in its identification of settlement rates as evidentiary benchmarks. Such a theoretical problem cannot serve to override the salutary aspects of benchmark settlement agreements. See Web IV, supra at 26,630 (rejecting same argument as speculative and ‘‘too untethered from the facts to be predictive or useful in adjusting for the supposed shadow of the existing statutory rate.’’). Lastly, with regard to a benchmark affected by the so-called ‘‘shadow,’’ the Judges find that, with regard to the application of the itemized factors in E:\FR\FM\10AUR2.SGM 10AUR2 54446 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations section 801(b)(1), they have the same duty to independently weigh those factors as they do for all otherwise reasonable rates. Thus, the Judges reject the idea that rates and terms reached through a settlement must be understood to supersede—or can be assumed to embody—the Judges’ current thinking as to the application of the statutory elements set forth in section 801(b)(1). The Judges are obliged to conduct the four-factor analysis anew when considering a previously adopted settlement in a subsequent proceeding— and they do so infra. Of course, if on such further analysis, the Judges find that the provisions in an otherwise useful benchmark agreement (including those in a benchmark influenced by the so-called ‘‘shadow’’) do appropriately reflect the four itemized statutory factors in section 801(b)(1), then the Judges may adopt the provisions of that settlement without a factor-based adjustment. ddrumheller on DSK120RN23PROD with RULES2 h. Conclusion Regarding PR II-Based Benchmark Accordingly, the Judges find the PR II Rates to be a useful benchmark. However, this benchmark is modified by the Judges’ substitution of the 15.1% headline percentage rate for the 10.5% headline percentage rate in the benchmark. D. Precedent Permits Judges To Apply Elements of PR II Rates, Rate Structure and Terms Even if Those Are Not Proffered as Benchmarks The D.C. Circuit has previously held that the Judges have the authority to adopt elements from the existing rate provisions, if they find that those prevailing provisions better satisfy the statutory requisites than any other proposed structures and rates discernible from the record evidence. Music Choice v. Copyright Royalty Bd., 774 F.3d 1000, 1009 (D.C. Cir. 2014). This authority exists even when no party has proffered those provisions in the form of a benchmark. In Music Choice (concerning the setting of satellite radio royalty rates under the same section 801(b)(1) standard), the CRB Judges rejected the parties’ proffered benchmarks and instead relied on a percent-of-revenue rate (13%) that was neither a benchmark nor even the prior statutory rate, but merely ‘‘a component of a prior determination.’’ Id. at 1009. The licensor-party, SoundExchange, argued, like Copyright Owners here, that this component of a prior rate was ‘‘stale,’’ ‘‘outdated,’’ or ‘‘obsolete.’’ Rejecting this argument as ‘‘erroneous,’’ the D.C. Circuit stated that ‘‘the Judges did not VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 consider the 13% rate as a current benchmark,’’ but rather used it to ‘‘bridge the gap’’ caused by the inadequacies of the parties’ rejected benchmarks. Id. In so doing, the D.C. Circuit held that the Judges properly resolved ‘‘serious problems’’ with the licensor’s proposal, even as it had ‘‘partially credited it’’ and also ‘‘used permissible indicia of reasonableness to help fix the rate.’’ Id. Music Choice is highly instructive. Here, on remand, the Judges adopt a modified version of the prior rate structure and rates in Phonorecords II. The fact that it was also proffered as a benchmark, in another modified form by the Services, does not render Music Choice inapposite. Rather, because the Phonorecords II provisions were proffered as benchmark evidence, these provisions were placed squarely into the record, allowing the parties and the Judges to address the relative merits. A fortiori, Music Choice underscores the propriety of the Judges approach in this proceeding. That is, even if the Services had not proffered this approach as a benchmark, Music Choice allows the Phonorecords II approach to serve as a guidepost for establishing the rates and rate structure in this proceeding. Further, here the Judges are adopting actual elements from the prior rate provisions, rather than, as in Music Choice, a mere ‘‘component’’ used to generate the prior rate. A fortiori yet again, Music Choice allows the Judges to prudently utilize the prior rate and rate structure regulations to synthesize a determination in this proceeding. The analogous nature of Music Choice is also seen in the Judges’ use in the present case of the ‘‘headline’’ 15.1% revenue rate proposed by Copyright Owners on remand combined with elements from the PR–II regulatory provisions, including its price discriminatory rates. In Music Choice, the Judges likewise ‘‘partially credited’’ the licensor’s proposal, which, as noted supra, the D.C. Circuit affirmed. Finally, the Judges take note that Music Choice also addressed the Judges’ findings regarding the setting of another statutory license, for Preexisting Subscription Services (PSS), by using a rate in a settlement from a prior period. This context is also analogous here, because Copyright Owners object to the use of the Phonorecords II rate structure and rates as the product of a settlement. It is instructive to consider how the arguments of the licensor (SoundExchange) in Music Choice mirror those of Copyright Owners in this proceeding: • SoundExchange notes that this rate ‘‘is the product of settlement PO 00000 Frm 00042 Fmt 4701 Sfmt 4700 negotiations that occurred in SDARS I between Music Choice and SoundExchange.’’ • SoundExchange argues that the Judges arbitrarily rejected . . . more recent data points in favor of the ‘‘outdated’’ settlement rate. • SoundExchange maintains that the Judges conceded that the prevailing rate had limited value, as the settlement rate ‘‘was negotiated in the shadow of the statutory licensing system and cannot properly be said to be a market benchmark rate.’’ • SoundExchange also argues that simply reciting that ‘‘nothing in the record persuades the Judges’’ that the prevailing rate is unreasonable . . . does not show that [it] is reasonable, or that it is supported by the written record. • [G]iven the lack of creditable benchmarks in the record, the Judges did not err when they used the prevailing rate as the starting point of their Section 801(b) analysis. • The Copyright Act contemplates that the Judges would . . . consider ‘‘prior determinations’’ and rates established ‘‘under voluntary license agreements.’’ • [T]he Judges did not err when relying on the settlement rate. The Judges conceded that the settlement rate does not represent a market rate. . . . But . . . the relevant portion of the Copyright Act does not use the term ‘‘market rates,’’ nor does it require that the term ‘‘‘reasonable rates’’ be defined as market rates. . . . The Act authorizes the Judges to consider rates set ‘‘under voluntary license agreements.’’ • Music Choice complains that it agreed to a higher rate to avoid litigation costs, but has not introduced evidence that the settlement was involuntary or otherwise unreasonable. It was not arbitrary, then, for the Judges to consider the voluntary settlement rate. Music Choice, 774 F.3d at 1012–15. These aspects of Music Choice are highly instructive, considering the Judges’ parallel findings regarding the same and similar arguments as discussed supra regarding prior settlement agreements and the so-called ‘‘shadow’’ of the statutory rates. In sum, Music Choice provides ample support for the conclusion that, even if the Services had not proffered their PR II-based benchmark, the Judges would have acted well within their authority to give the same weight to the PR II rates and structure as they have in this Initial Ruling.138 138 This ruling is in no way conflicts with the Judges’ duty to set rates, rate structures, and terms de novo in each rate proceeding, as discussed supra. E:\FR\FM\10AUR2.SGM 10AUR2 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations E. Four Itemized Factors in Section 801(b)(1) The Judges have considered the application of the four itemized statutory factors A through D, in connection with their application of the 15.1% revenue rate and their partial use of the PR II-based benchmark. ddrumheller on DSK120RN23PROD with RULES2 1. Factor A The Judges have explained supra that price discrimination is a ‘‘win-win’’ for Copyright Owners and the Services. By serving low WTP listeners, it brings in new listeners and subscribers who increase royalty payments as well as revenues. Any licensor would prefer to increase its royalties, rather than ‘‘leave money on the table,’’ and a rate structure that effects such an increase (through the concept of ‘‘derived demand’’) is appropriate. Moreover, for purposes of applying Factor A, a rate structure that increases royalties, ceteris paribus, would induce more production of musical works, a result that Copyright Owners should desire. This point appears to raise a question: How could Copyright Owners and their economic experts object to a rate structure that inures to their benefit as well? The answer is: They do not object. They are not economic naifs. As stated supra, they advocate for a rate set under the bargaining room theory, through which rate structures can still be negotiated, but not subject to the ‘‘reasonable rate’’ and itemized factor analysis required by law. In those negotiations, as Dr. Eisenach candidly acknowledged, Copyright Owners would have a different threat point to use in order to obtain better rates and terms. 4/4/17 Tr. 4845–46 (Eisenach). Second, given a heterogeneous downstream WTP, it would not be more profitable simply to equate ‘‘availability’’ with a higher rate. As noted supra, any product that is priced beyond the WTP of a significant portion of the public is unavailable to that segment.139 Royalties that are aligned The de novo process requires the Judges to weigh new evidence regarding potential new rates, rate structures, and terms, but that is not inconsistent with the Judges’ ability, as explicated by the D.C. Circuit in Music Choice, to adopt prior rates, rate structures, and terms in whole or in part if, in their discretion, the new evidence is deficient. See Music Choice, supra, at 1012 (‘‘The Judges were under no obligation to salvage benchmarks they found to have fundamental problems.’’). 139 The concept of willingness-to-pay (WTP) as used by economists is an antiseptic phrase, because it includes not merely people who do not value a music streaming subscription highly, but also individuals and families who are ‘‘income constrained’’ (yet another antiseptic phrase, read ‘‘low income’’ people and families) who lack the ‘‘ability-to-pay’’ for an interactive subscription. That segment of the population likely reflects a VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 with the varying WTP of different classes of listeners will make downstream price discrimination more affordable to the services, driving new revenue and royalties—precisely as the PR II-based benchmark allows.140 In this regard, Copyright Owners have taken a cramped and unrealistic view of such incentives. In particular, the Judges disagree with Copyright Owners’ expert economic witness, Professor Rysman, who startlingly asserted in response to a hypothetical from the bench that even a $10,000 per month subscription price would increase ‘‘availability.’’ 4/3/17 Tr. 4397 (Rysman). The Judges find Professor Rysman misapprehends the nature of a price signal. If the price is so high as to eliminate or reduce total revenue to creators, in no way will higher rates simply induce the supply of creative works over time. Indeed, even monopolists do not seek the highest price possible, but rather seek to maximize profits. See E. Mansfield & G. Yohe, Microeconomics at 362–63 (11th ed. 2004) (‘‘Monopolies maximize profits by producing where marginal cost equals marginal revenue.’’). Thus, even monopolists, who have the most market power, are constrained in their pricing by the demand curve and the marginal revenue it creates. Simply put, although a higher royalty rate might have an immediate superficial appeal, if the consequence will be lower revenues, the high per-play rate would reveal itself as a form of fool’s gold. In sum, the Judges find that the Factor A objective of ‘‘maximizing the availability of creative works’’ is furthered by an upstream rate structure that contains multiple royalty rates reflective of and derived from downstream variable WTP, because it will facilitate beneficial price discrimination. Such price discrimination allows for access to be afforded ‘‘down the demand curve,’’ significant portion of the nation, because ‘‘40% of Americans would struggle to come up with even $400 to pay for an unexpected bill,’’ let alone pay for a music streaming service. See https:// www.minneapolisfed.org/article/2021/what-a-400dollar-emergency-expense-tells-us-bout-theeconomy. When the royalty rates paid by interactive services enable streaming services to satisfy the demand of these low-income consumers (through the principle of ‘‘derived demand’’) that segment of American society can enjoy the benefits of listening to interactive streamed music, even if the offerings they can afford lack the large catalogs and ‘‘bells and whistles’’ of a pricier service. 140 To be sure, royalties will not increase in equal proportions with increases in the number of streams or listeners, but that is a feature of price discrimination, not a bug. The goal is to generate revenues from low WTP listeners who otherwise would be lost as sources of revenues and royalties to both the interactive services and Copyright Owners. PO 00000 Frm 00043 Fmt 4701 Sfmt 4700 54447 making musical works available to more members of the public. However, there is no evidence to suggest that the price discriminatory rates should be changed, in order to address the connection between price discrimination and the objective of Factor (A). Accordingly, the Judges find no basis to adjust either the rate structure or the rates based on Factor (A). 2. Factors B and C The concepts of ‘‘fair income,’’ ‘‘fair return’’ and recompense for costs and other contributions was considered in connection with the setting of the 15.1% revenue rate. In that context, the Judges analyzed the Shapley Value modeling that was designed to generate ‘‘fair’’ rates that allowed the parties to recover their costs and to share the surplus (over and above costs) in a manner that: (1) prevented the ‘‘Must Have’’ Input Suppliers (the record companies and Copyright Owners) from using the essential aspect of their inputs to engage in hold-up by threatening to withhold their respective repertoires; and (2) allocated surplus shares according to each party’s contribution to the surplus (as calculated though the ‘‘arrival orderings’’ in the Shapley model).141 The PR II-based benchmark was the product of an industrywide negotiation, with the music publishers represented by the NMPA and the interactive streaming services represented by DiMA, their respective trade associations. As explained in the Dissent, supra, at pp. 137–39, when an industrywide settlement is reached, particularly when the default procedure is a contested rate proceeding before the Judges, it contains the same benefits with regard to the avoidance of the ‘‘hold-out’’ effect and the equalizing of bargaining power as produced by Professor Marx’s Shapley value modeling. See 3/13/17 Tr. 577 (Katz) (‘‘I think of the shadow as balancing the bargaining power between the two 141 As noted elsewhere in this Initial Ruling, Professor Marx, Spotify’s economic expert witness, reduced the relative market power of the input suppliers in her model which she claimed would be consonant with the ‘‘fairness’’ objectives in Factor B. On behalf of Copyright Owners, Professor Watt disagreed, arguing that the Shapley approach takes the existing market power as reflective of the parties’ market contributions, and thus needs no adjustment. The Majority utilized Professor Marx’s Shapley-based calculation of a total royalty payment of [REDACTED]% of service revenue in setting a 15.1% revenue rate (phased-in), which the Judges are adopting in this Initial Ruling. The Majority also used Professor Marx’s calculation to find that Factors B and C were satisfied without further adjustment. See Determination at 68 & n.120, 75, 86–87. But this issue is not relevant to the present discussion of Factors B and C with regard to the application of the PR II-based benchmark. E:\FR\FM\10AUR2.SGM 10AUR2 54448 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations parties.’’); Katz CWRT 136, n.236 (‘‘there are market forces that promote the achievement of the statutory objectives in private agreements, such as the 2012 Settlement, when the parties are equally matched (it was an industrywide negotiation) and the negotiations are conducted in the shadow of a pending rate-setting proceeding that can be expected to set reasonable rates in the event that the private parties do not reach agreement.’’). Accordingly, this benchmark already incorporates the dynamics of a negotiation between parties with mutually countervailing power (although those dynamics required updating of the headline rate to 15.1% to account for the higher revenues, as undertaken by the Majority’s Shapley analysis). See Web V, 86 FR 59452, 59456 (Oct. 27, 2021) (‘‘the licensor-side complementary oligopoly power could be ameliorated by the ‘‘countervailing power’’ of a licensee’’). Therefore, the Judges do not make any adjustment in their application of the PR II-based benchmark pursuant to Factors B and C. 3. Factor D As noted supra, the Judges understand that a Factor D adjustment is warranted if the rate the Judges would otherwise establish ddrumheller on DSK120RN23PROD with RULES2 directly produces an adverse impact that is substantial, immediate and irreversible in the short-run because there is insufficient time for either [party] to adequately adapt to the changed circumstance produced by the rate change and, as a consequence, such adverse impacts threaten the viability of the music delivery service currently offered to consumers under this license. Determination at 87. There is no record evidence to suggest that the Services’ PR II-based benchmark, as utilized by the Judges in this Initial Ruling, would create the requisite ‘‘adverse impact’’ to trigger Factor D. The Services certainly do not assert that their own proffered benchmark would be disruptive. With regard to Copyright Owners, the Judges cannot identify any aspect of the PR IIbased benchmark that would cause the type of disruption that can serve as an adjustment under the statutory language of Factor D or the Judges’ application of same, as quoted above. The Judges understand Copyright Owners’ complaint to be principally that [REDACTED] during the Phonorecords II period, [REDACTED] the number of musical works streamed via sound recordings performed on interactive services. However, that is most certainly not any sort of disruption, let alone a disruption cognizable under section VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 801(b)(1) and under the Judges’ application of that provision. F. Subpart C Offerings Covered by Foregoing Analysis The Phonorecords II parties also negotiated several new service types— paid locker services, purchased content locker services, mixed service bundles, music bundles and limited offerings. These service configurations were described in subpart C of 37 CFR 385 under the Phonorecords II regulatory provisions.142 Parness WDT ¶ 13; Levine WDT ¶¶ 38–39; Israelite WDT ¶¶ 28–30. These negotiations spanned more than a year. See 3/29/17 Tr. 3652– 55 (Israelite) (involved protracted bargaining, in which NMPA rejected some categories, while others were accepted and became part of subpart C). Id. at 3654–56. The parties ultimately agreed on a structure for subpart C that resembled the subpart B structure, including a headline percentage of revenue royalty rate and per-subscriber and TCC minima. Parness WDT ¶ 14; see also 37 CFR 385.22. As with the bundling negotiations relating to subpart B, the parties negotiated and created a bundled service category under subpart C (with certain adjustments to the definition of ‘‘revenue.’’) 3/8/17 Tr. 161–64 (Levine); 37 CFR 385.21. Copyright Owners urge the elimination of the subpart C provisions as essentially obsolete because locker services for ‘‘purchased content’’ (new download purchases) and for ‘‘paid’’ downloads (already owned) have largely disappeared, as listeners transitioned away from ownership models to access models. See 3/8/17 Tr. 159–160 (Levine); 3/16/17 Tr. 1458–1461 (Mirchandani); Mirchandani WDT ¶ 33; 3/22/17 Tr. 2523 (Dorn). Copyright Owners also re-assert the same arguments with respect to subpart C as they have for interactive streaming in subpart B. See CORPFF–JS at p.2. The Services argue that Copyright Owners do not point to any evidence to show that locker services have completely disappeared, emphasizing that Apple and Amazon continue to offer locker service. Joyce WDT ¶ 5; Mirchandani WDT ¶¶ 16–17; 3/22/17 Tr. 2523–25 (Dorn); Ramaprasad WDT, Table 3. More generally, the Services urge the Judges to use the subpart C rate structure as the benchmark for rates in the forthcoming period for the same 142 The interactive steaming (and limited download) provisions that are the principal subject of this proceeding were contained in subpart B of the Phonorecords II (and Phonorecords I) regulations. (These subparts were reorganized pursuant to the now vacated Determination.) PO 00000 Frm 00044 Fmt 4701 Sfmt 4700 reasons as they urge the use of the subpart B rates as an appropriate benchmark. See Mirchandani WDT ¶¶ 58–62. The Judges find no reason on remand to treat the subpart C offerings differently than the manner in which they are treating the subpart B interactive streaming offerings, for the reasons set forth in the Dissent at 118– 119. That means, however, that the various ‘‘headline’’ rates for these subpart C offerings must also adjust to 15.1%,143 whereas the alternative rates (identified in subpart C as ‘‘minima’’ and ‘‘subminima)’’ rates shall remain unchanged. IV. Change in Definition of Service Revenue for Bundles 144 The Judges analyze the definition of ‘‘Service Revenue’’ for bundled offerings in the context of the partial adoption of the PR II-based benchmark. As discussed supra, the Judges have found that the PR II-based benchmark is a useful benchmark, particularly because of its features that incentivize beneficial downstream price discrimination that generates more listeners, revenues, and royalties. A. Background In their Initial Determination, the Judges adopted a definition of ‘‘Service Revenue’’ (i.e., a royalty base) for a ‘‘Bundle’’ 145 that provided, in pertinent part: Service Revenue shall be the revenue recognized from End Users for the Bundle less the standalone published price for End Users for each of the other component(s) of the Bundle . . . Initial Determination, Attachment A at 7 (§ 382.2 therein).146 143 Accordingly, in the PR II-based benchmark, the subpart C ‘‘headline’’ rates that shall adjust to 15.1% are: 11.35% for Mixed Service Bundles; 11.35% for Music Bundles; 10.5% for Limited Offerings; 12% for Paid Locker Services; and 12% for Purchased Content Locker Services. See 37 CFR 385.22(a)(1) (Step 1); 385.23(a)(1) through (5). 144 Judge Strickler disagrees with the procedural analysis of a different majority by which they readopt the Bundled Revenue definition from the Initial Determination, and he dissents on that specific issue. However, Judge Strickler concurs and joins with the Majority regarding the substantive re-adoption of that definition from the Initial Determination. Judge Strickler has drafted a separate opinion on this Bundled Revenue issue. 145 For interactive streaming, the Judges’ Initial Determination defined a ‘‘bundle’’ (in pertinent part) as an offering which combined the delivery of streamed music: ‘‘together with one or more nonmusic services . . . or non-music products . . . as part of one transaction without pricing for the music services or music products separate from the whole offering. . . .’’ Initial Determination, Attachment A at 2 (§ 385.2 therein). 146 The definition added: ‘‘[I]f there is no standalone published price for a component of the Bundle, then the Service shall use the average E:\FR\FM\10AUR2.SGM 10AUR2 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations ddrumheller on DSK120RN23PROD with RULES2 After the Judges issued their Initial Determination, Copyright Owners submitted a Motion for Clarification or Correction of Typographical Errors and Certain Regulatory Terms which disclaimed any intent to seek rehearing, but sought ‘‘clarification or correction’’ of certain regulatory terms to conform them to what Copyright Owners claimed to be the apparent intent of the Initial Determination. (Motion for Clarification).147 Copyright Owners purported to bring their motion under the Judges’ general regulations governing motions. See 37 CFR 303.3 and 303.4 (formerly codified at 37 CFR 350.3 and 305.4). The Motion for Clarification argued, among other things, that the definition of Service Revenue as applied to bundled offerings should be reworked. Copyright Owners argued that defining the revenue as the total price of the bundle, minus the standalone published prices for the non-streaming offerings in the bundle, undervalued the revenue created by the streaming offerings. They proposed that ‘‘Service Revenue’’ for bundled offerings be defined as the standalone price of the offering (or comparable offerings). The Services objected to Copyright Owners’ styling of their motion as something other than a motion for rehearing. The Services also objected that Copyright Owners had not previously proposed a definition of ‘‘Service Revenue’’ for bundled offerings, and that their ‘‘late-proposed’’ definition was unsupported by the record. On October 29, 2018, the Judges issued an Order concluding neither party had met the exceptional standard for granting rehearing motions,148 stating that the parties had failed to present ‘‘even a prima facie case for rehearing under the applicable standard’’. Amended Order Granting in Part and Denying in Part Motions for Rehearing (Order on Rehearing) (Jan. 4, 2019).149 The Judges explained that they nevertheless found it appropriate to standalone published price for End Users for the most closely comparable product or service in the U.S. or, if more than one comparable exists, the average of standalone prices for comparables.’’ Id. at 7–8. 147 Streaming Services submitted a motion for rehearing that was limited to fixing clerical errors and clarifying existing ambiguities in the proposed regulatory terms appended to the Initial Determination. 148 The standard is set forth in the Order on Rehearing at 2 n.3. The Judges discuss and apply this standard infra, pursuant to Johnson, and in the context of this remand proceeding. 149 Judge Strickler, who had dissented from the Initial Determination and the Determinations, did not join in this Order on Rehearing. VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 resolve the issues that the parties had raised. Order on Rehearing at 2. The Judges added that, to the extent such resolution could be considered a rehearing under 17 U.S.C. 803(c)(2), the Judges resolved the motions on the papers without oral argument. Id. Regarding the definition of ‘‘Service Revenue’’ for bundled offerings, the Judges summarized the parties’ competing arguments: Copyright Owners presented evidence that the existing approach led, in some cases, to an inappropriately low revenue base—but did so in service to their argument that the Judges should reject revenue-based royalty structures. They did not present evidence to support a different measure of bundled revenue because their rate proposal was not revenue-based. The Services rely on the fact that the approach to bundled revenue in the extant regulations is derived from the 2012 Settlement. The Judges have, however, declined to rely on the 2012 Settlement as a benchmark, as the basis for the rate structure, or, therefore, as regulatory guidance. The Services have observed correctly that the evidentiary records in Web IV and SDARS III differ from the record in this proceeding.150 Order on Rehearing at 17 (emphasis added). Despite these arguments, the Judges found that neither party presented evidence adequate to support the approach advocated in postdetermination filings, because ‘‘the ‘economic indeterminacy’ problem inherent in bundling’’ remained unresolved.’’ Id.151 The Judges stated that the Services were the party in possession of the relevant information, and concluded that the Services bore the burden of providing evidence that might mitigate the ‘‘indeterminacy problem’’ inherent in bundling. Because the Judges concluded that the Services had not met that burden, they ruled that they must adopt an approach to valuing bundled revenue that is in line with what the Copyright Owners proposed. As a result, the Judges discarded the formula in the Initial Determination and ruled, instead, that streaming service providers will use their own standalone price (or comparable) for the music component (not to exceed the value of 150 In Web IV and SDARS III, unlike under the Phonorecords II-based benchmark, there were no minima or floors to provide licensors with royalties in the event bundled offerings would otherwise fail to generate royalties. 151 The ‘‘economic indeterminacy’’ problem was described in SDARS III: ‘‘Such bundling [for full quotation, see eCRB no. 27063 n.140].’’ SDARS III, 83 FR 65264. As discussed in this Initial Ruling, this indeterminacy problem was addressed by the Phonorecords II-based benchmark through negotiated alternative royalty provisions for bundled offerings. PO 00000 Frm 00045 Fmt 4701 Sfmt 4700 54449 the entire bundle) when allocating bundled revenue. Id. at 16–18. Consistent with the Judges’ Order on Rehearing, the Judges’ replaced the definition of ‘‘Service Revenue’’ for a ‘‘Bundle’’ that they had included in the Initial Determination with a new definition in the Determination. The final definition provided, in pertinent part: Service Revenue shall be the lesser of the revenue recognized from End Users for the bundle and the aggregate standalone published prices for End Users for each of the component(s) of the bundle that are Licensed Activities . . . [or] if there is no [such] standalone price, then the average standalone . . . price . . . for the most closely comparable product or service . . . or . . . the average of standalone prices for comparables. Determination, Attachment A at 8. The Services, Copyright Owners and George Johnson appealed the Judges’ Determination to the D.C. Circuit. See Johnson, 969 F.3d 363. The Services challenged both the Judges’ legal authority and the substantive soundness of the decision to reformulate the definition of ‘‘Service Revenue’’ for bundled offerings, after the Judges had issued the Initial Determination. The D.C. Circuit examined several authorities under which the Judges may revisit and amend a determination. It addressed the three ways identified in the statute: ‘‘(i) order rehearing ‘in exceptional cases’ in response to a party’s motion, 17 U.S.C. 803(c)(2)(A); (ii) correct ‘technical or clerical errors,’ id. § 803(c)(4); and (iii) ‘modify the terms, but not the rates’ of a royalty payment, ‘in response to unforeseen circumstances that would frustrate the proper implementation of [the] determination.’ ’’ Johnson, 969 F.3d at 390. The D.C. Circuit found that the Judges’ reformulation of the definition of ‘‘Service Revenue’’ fit none of those categories. The D.C. Circuit noted that the Judges were explicit that they did not treat the Motion for Clarification as a motion for rehearing under 17 U.S.C. 803(c)(2). Id. Furthermore, the D.C. Circuit noted the Judges’ own findings that the Motion for Clarification did not meet the exceptional standard for granting rehearing motions under section 803(c)(2) and that the Copyright Owners failed to make even a prima facie case under the rehearing standard. In Johnson, the D.C. Circuit found that the change to the definition of Service Revenue for bundled offerings was not an exercise of the Judges’ authority under section 803(c)(4) to ‘‘correct any technical or clerical errors in the determination[.]’’ 17 U.S.C. 803(c)(4). E:\FR\FM\10AUR2.SGM 10AUR2 54450 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations The D.C. Circuit observed the substantive nature of the change to the definition and determined that there was nothing technical or clerical about the amendment. The D.C. Circuit found that the Judges did not even purport to modify the terms in response to unforeseen circumstances that would frustrate the proper implementation of the Initial Determination. The D.C. Circuit observed that the Judges never mentioned section 803(c)(4) or unforeseen circumstances as the basis for revamping the Service Revenue definition. Beyond the explicit statutory authorities for amendments to determinations, the D.C. Circuit addressed arguments for inherent authority to make sua sponte any appropriate substantive or fundamental changes after the Initial Determination. The D.C. Circuit foreclosed reliance on inherent authority, finding that Congress’s decision to limit rehearing to exceptional cases, and to confine other post hoc amendments to cases involving technical or clerical errors, would be a nullity if the Judges also had plenary authority to revise their determinations whenever they thought appropriate. The D.C. Circuit noted that the Judges’ decision to amend the definition said nothing of the sort, and prior decisions are silent on that topic. In sum, the D.C. Circuit found that the Judges failed to explain the legal authority for reformulating the definition of ‘‘Service Revenue.’’ In relevant part, the D.C. Circuit ruled ddrumheller on DSK120RN23PROD with RULES2 we must vacate the [ ] Determination’s bundled offering Service Revenue definition and remand for the [Judges] . . . either to provide ‘a fuller explanation of the agency’s reasoning at the time of the agency action[,]’ or to take ‘new agency action’ accompanied by the appropriate procedures. Id. at 392 (citing Regents, 140 S.Ct. at 1908). Because the D.C. Circuit determined that the Judges failed to identify any legal authority for adopting the new Service Revenue definition, it found no occasion to address the Streaming Services’ separate argument that the definition was arbitrary, capricious, or unsupported by substantial evidence. Id. The Services and Copyright Owners agreed that the Judges should resolve the definitional issue based on the existing record, after receiving two rounds of additional briefing from the parties.152 See Services’ Proposal for Remand Proceedings (Dec. 10, 2020) 152 As indicated below, during the remand proceedings, the Judges solicited two rounds of additional briefing addressing specific issues. VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 (Services’ Proposal) at 5–6, 9–10; Proposal of the Copyright Owners for Conduct and Resolution of the Remand (Public) (Dec. 10, 2020) (Copyright Owners’ Proposal) at 4–6. The Judges issued an Order Regarding Proceedings on Remand, which, in part, opened briefing on the issue of the adoption of a revised definition of ‘‘service revenue’’ for bundled offerings between issuing the Initial Determination and the Determination. Order Regarding Proceedings on Remand (Dec. 15, 2020). The Judges received the following relevant briefing. • CO Initial Submission • Services’ Initial Submission • CO Reply • Services’ Reply On December 9, 2021, the Judges requested additional briefing. Dec. 9 Order. The Dec. 9 Order sought additional briefing setting forth the parties’ views on whether this proceeding constitutes the type of new agency action addressed by the D.C. Circuit, which would allow adoption of a Service Revenue definition without limitation to the definition expressed in the Initial Determination. Additionally, the Judges requested additional evidence that the parties might offer to support adoption of the Service Revenue definitions expressed in either the Initial Determination or the Determination. In response to the Dec. 9 Order, the Judges received the following relevant briefing. • CO Additional Submission • Services’ Additional Submission On February 9, 2022, the Judges solicited further briefing on ‘‘Whether the D.C. Circuit’s Johnson decision permitting the Judges to engage in new agency action in this remand proceeding allows the Judges to engage in new agency action through a reconsideration of Copyright Owners’ February 12, 2018 Motion for Clarification as a Motion for ‘rehearing’ pursuant to 17 U.S.C. 803(c)(2)(A) and 37 CFR 353.1.’’ Sua Sponte Order Regarding Additional Briefing (Feb. 9 Order). In response to the Feb. 9 Order, the Judges received the following relevant briefing. • Copyright Owners’ Brief Responding to Judges’ February 9, 2022 Sua Sponte Order Regarding Additional Briefing on New Agency Action Question, and Replying to Services’ New Agency Action Arguments in their Joint Supplemental Brief Addressing the Judges’ Working Proposal (in Additional Materials Rebuttal Submission of Copyright Owners at Tab B) (Feb. 24, 2022) (‘‘CO Further Briefing’’) PO 00000 Frm 00046 Fmt 4701 Sfmt 4700 • Services’ Joint Response to the Judges’ February 9, 2022 Sua Sponte Order Regarding Additional Briefing and Rebuttal Regarding ‘‘New Agency Action’’ (Feb. 24, 2022) (‘‘Services’ Further Briefing’’) B. Authority for Modification to the Initial Determination 1. Copyright Owners’ Position Copyright Owners assert that this remand proceeding offers a straightforward path to take new agency action and that the law makes clear that new agency action can consist of issuing a new determination on remand. CO Initial Submission at 71. Copyright Owners maintain that: [T]the new agency action here is a determination after remand proceedings, the Board is largely free to chart its own procedural course, and the Board has done so in its December 15 Order. The Board is not required to undertake any of the procedural steps set forth in 17 U.S.C. 803(b) in order to take such ‘‘new agency action.’’ See 17 U.S.C. 803(d)(3) (requiring only that on remand further proceedings be taken ‘‘in accordance with subsection (a)’’); 37 CFR 351.15; Intercollegiate Broad. Sys., Inc., 796 F.3d at 125 (‘‘[N]either the Copyright Act nor the Board’s regulations prescribe any particular procedures on remand.’’) The Circuit’s instruction that the action be ‘‘accompanied by the appropriate procedures[,]’’ Johnson, 969 F.3d at 392, does not dictate what those ‘‘appropriate procedures’’ must be but instead plainly refers to these flexible rules. See also Oceana, Inc., 321 F. Supp. 3d at 136 (explaining that when remanding to an agency, a court generally ‘‘may not dictate to the agency the methods, procedures, or time dimension, for its reconsideration’’). CO Initial Submission at 71, FN 33. Copyright Owners acknowledge the Services’ position that the asserted procedural error is an ‘‘absence of authority’’ that can never be cured. Id. at 74 (citing Services’ Proposal for Remand Proceedings at 10). They note that the D.C. Circuit did not say the Judges lacked the authority to revisit the service revenue definition for bundles on remand. Nor, they observe, did it say the Judges have no authority to review the record evidence and the parties’ arguments and reach the same conclusion or a different conclusion on remand. Copyright Owners opine that if the only possible outcome were for the Judges to reinstate a definition that lacked any explanation or evidentiary support solely because it was present in the Initial Determination, then the D.C. Circuit would not have remanded the issue but would have simply reversed and reinstated the Initial Determination definition. But instead, they note, the D.C. Circuit remanded and said the E:\FR\FM\10AUR2.SGM 10AUR2 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations ddrumheller on DSK120RN23PROD with RULES2 Judges could take new agency action precisely to cure the asserted procedural defect. Copyright Owners assert that the remand allowed the parties to present the record evidence and their arguments so that the Judges can address the definition ‘‘afresh’’ in the remand determination. Id. at 74. Copyright Owners argue that 17 U.S.C. 803(d)(3) states only that proceedings on remand must be in accordance with 17 U.S.C. 803(a). They contend that remand proceedings need not be confined to procedures the Services claim are too late in the game for the Judges to follow. The Copyright Owners point to the D.C. Circuit’s ruling in Intercollegiate Broad. Sys., Inc. v. Copyright Royalty Bd., that ‘‘neither the Copyright Act nor the Board’s regulations prescribe any particular procedures on remand.’’ 796 F.3d 111, 125 (D.C. Cir. 2015) (citing 17 U.S.C. 803(a), (d)(3)). Accordingly, they argue, the Judges can reaffirm the adopted bundled service revenue definition following their review of the parties’ submissions without regard to section 803(c)(2) or 803(c)(4). CO Reply at 65– 66.153 Copyright Owners further argue that the Judges may properly justify the changed definition under section 803(c) as a fuller explanation of the agency’s reasoning at the time it was made. They urge that the Judges could explain that, especially in light of the evidence of how the Services misused the prior definition to make service revenue completely disappear, carrying over the prior bundle service revenue from Phonorecords II into the Initial Determination was unintended and inadvertent.154 CO Reply at 69. Copyright Owners also assert that the Judges could explain that Copyright Owners had, in their Motion for Clarification, identified an ‘‘exceptional case’’ under section 803(c)(2) because the prior definition failed to comport 153 Copyright Owners reiterate this argument in the CO Additional Submission. Copyright Owners added that the parties in this remand were afforded the opportunity for further briefing and, if they wished, to submit additional evidence on this issue, thus providing broader opportunity for submission than in Fisher v. Pension Benefit Guaranty Corp., 994 F.3d 664, 670 (D.C. Cir. 2021), in which the D.C. Circuit upheld new agency action after remand even though the agency did not provide appellant the opportunity to submit new briefing or exhibits. CO Additional Submission at 35–36; 38. 154 Copyright Owners assert that the definition in the Initial Determination conflicted with, the Board’s findings in the Initial Determination, including its findings that the adopted rates and terms would afford copyright owners a fair return for their creative works, thereby satisfying factor B of the 801(b) standard and thus needed to be revised so as to not ‘‘frustrate the proper implementation of’’ the Final Determination. CO Reply at 69 (citing 17 U.S.C. 801(b) and 803(c)(4)). VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 with Judges’ precedent and economic principles, and was unsupported by evidence.155 In addition, Copyright Owners note that the Judges reheard the evidence and legal arguments as presented in the parties’ briefs on the issue and, as a result, may choose to adopt the revised definition. Copyright Owners maintain that for the Judges to do so would not be impermissible post-hoc reasoning, because the D.C. Circuit remanded precisely because the Judges did not provide any reason in the Determination for revising the bundle revenue definition. CO Reply at 69–71. 2. Services’ Position The Services assert that the D.C. Circuit found only ‘‘three ways in which the Board can revise Initial Determinations’’ and that the Judges had failed to establish that the change to the service revenue definition fit any of those three categories. Services’ Initial Submission at 64–65 (citing Johnson at 390). According to the Services the first way the Judges may revise an Initial Determination is to ‘‘order rehearing ‘in exceptional cases’ in response to a party’s motion, 17 U.S.C. 803(c)(2)(A).’’ Services’ Initial Submission at 65 (citing Johnson at 390).156 The Services argue that the D.C. Circuit held in Johnson that the Judges’ ‘‘material revision of the ‘Service Revenue’ definition for bundled offerings does not fall within the Board’s rehearing authority under section 803(c)(2)(A)’’ because ‘‘the Board itself . . . was explicit that it ‘did not treat the [Copyright Owners’] motion[ ]’ . . . ‘as [a] motion[ ] for rehearing under 17 U.S.C. 803(c)(2).’ ’’ The D.C. Circuit also noted that ‘‘as the Board found, the Copyright Owners’ motion did ‘not meet [the] exceptional standard for granting rehearing motions’ under section 803(c)(2).’’ Id. (citing Johnson at 390). The Services assert that the Judges were not able to make ‘‘a volte-face’’ and justify on appeal their revision to the definition as an exercise of rehearing authority. As the D.C. Circuit held, agency action must be justified by ‘‘reasons invoked by the agency at the time it took the challenged action,’’ and post-hoc rationalizations 155 In response to an Order by the Judges, Copyright Owners provided additional briefing regarding reconsideration of the motion for clarification as a motion for ‘‘rehearing’’ which is addressed separately infra. 156 In response to an Order by the Judges, the Services provided additional briefing regarding reconsideration of the motion for clarification as a motion for ‘‘rehearing’’ which is addressed separately infra. PO 00000 Frm 00047 Fmt 4701 Sfmt 4700 54451 are insufficient. Id. (citing Johnson at 390). The Services add their view that the Judges cannot revisit the decision to deny rehearing without engaging in impermissible post-hoc reasoning. They note that the Supreme Court has explained that, while an agency may ‘‘elaborate later’’ on its ‘‘initial explanation’’ of the reason (or reasons) for its action, it ‘‘may not provide new ones.’’ Services’ Initial Submission at 66 (citing e.g., Regents, 140 S. Ct. at 1908). The Services offer that the Judges, having stated that they did not consider the Copyright Owners’ motion to revise the definition to be a motion for rehearing, cannot now conclude that the motion qualified as one for rehearing and that the Judges in fact engaged in rehearing. Id. The Services add that under section 803(c)(2)(A), the Judges can only use their rehearing authority ‘‘ ‘in exceptional cases’ in response to a party’s motion.’’ Id. (citing Johnson at 390). The Services argue that the Motion for Clarification cannot be found to have satisfied that standard. The Copyright Owners did not argue that their motion satisfied the ‘‘exceptional cases’’ standard before the Judges or the D.C. Circuit, and have therefore waived that argument. Id. According to the Services, the second way the Judges may revise an Initial Determination, viz. action to correct a technical or clerical error under section 803(c)(4), cannot be used now to justify any modification of the Service Revenue definition in the Initial Determination. The Services note that the D.C. Circuit held specifically that the Judges’ change to the Service Revenue definition could not be construed as correcting a technical or clerical error because it involved a substantive rewrite of the Service revenue definition. Id. at 67 (citing Johnson at 391). The Services aver that the third way the Judges may revise the terms in an Initial Determination is in response to unforeseen circumstances that would frustrate the proper implementation of the determination. Id. at 67. The Services note that the D.C. Circuit held in Johnson that this authority did not justify the Judges’ change to the Service Revenue definition because the Judges did not invoke this authority and ‘‘the need to ground the original definition in the record’’ could not credibly be described as ‘‘an unforeseen circumstance.’’ Id. (citing Johnson at 391). The Services also note that the D.C. Circuit rejected the argument that the Judges have ‘‘inherent authority’’ to make changes to the Initial E:\FR\FM\10AUR2.SGM 10AUR2 ddrumheller on DSK120RN23PROD with RULES2 54452 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations Determination. The D.C. Circuit explained that the specific restrictions Congress placed on the Judges’ authority in section 803 ‘‘would be a nullity if the Board also had plenary authority to revise its determinations whenever it thought appropriate.’’ Id. (citing Johnson at 391–92). The Services add that even if the Judges offered a new source of authority capable of justifying substantive changes to the Service Revenue definition now, the Judges would be unable to rely on this ‘‘uninvoked authority’’ without engaging in impermissible post-hoc reasoning. Id. The Services counter Copyright Owners’ position that the Judges need not respond to the error the D.C. Circuit identified with this aspect of the Determination and that the Judges’ ‘‘new agency action’’ may consist of issuing a new determination on remand. The Services argue that failure to address the legal and factual issues on which the D.C. Circuit remanded would violate the D.C. Circuit’s order and would result in a second remand. The Services surmise that the issue of authority to make the changes to the Initial Determination are particularly important in this context, where the D.C. Circuit recognized that the Copyright Act places limits on the Judges’ authority to alter an initial determination by defining conditions for rehearing and the types of changes that are permitted absent a rehearing. In this regard, the Services maintain that the Judges cannot do on remand what they lacked authority to do in the first instance. The Services assert that the Judges must resolve the legal question whether there is authority to alter the revenue definition in the Initial Determination. They urge that the remanded issue is not what the substance of the service revenue definition should be as a matter of first impression, but instead is whether the Judges have properly exercised authority to alter the Initial Determination’s definition. Services Reply at 52–54.157 The Services assert that the Judges have two paths available to them: (1) to provide a ‘‘fuller explanation’’ of the prior conclusion that the Judges had legal authority to revise the Service Revenue definition in the Initial Determination or (2) answer that threshold question through new agency action. The Services maintain that, if 157 The Services agree that this remand proceeding qualifies as ‘‘new agency action’’ but again urge that failure to address the legal and factual issues on which the court remanded would nonetheless violate the D.C. Circuit’s order. Services’ Additional Submission at 38–42. VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 they pursue the ‘‘fuller explanation’’ path, the Judges are limited to elaborating on what they said previously, and that they cannot add new reasons they did not initially provide. With regard to what may constitute new agency action, the Services assert that path gives the Judges freedom to consider new reasons that the Copyright Act provided the Judges with the authority to make this change to the Initial Determination. The Services argue, however, that undertaking a new agency action does not, as Copyright Owners claim, obviate the need for the Judges to identify proper legal authority before substantively changing the Initial Determination, such authorities being limited to the authority of section 803(c)(4) or the rehearing authority of section 803(c)(2). Id. at 54–55. The Services address Copyright Owners’ position that if the only possible outcome were for the Judges to reinstate a definition that lacked any explanation or evidentiary support solely because it was present in the Initial Determination, then the D.C. Circuit would not have remanded the issue but would have simply reversed and reinstated the Initial Determination definition. The Services urge that the D.C. Circuit could not reverse because the Department of Justice raised for the first time on appeal new justifications for the Judges’ decision to change the Initial Determination. Instead, the Services maintain, the D.C. Circuit had to remand and give the Judges the opportunity to address the Department of Justice’s new justifications in the first instance, as the D.C. Circuit could not rule them out given the posture of the appeal. Id. at 56. In the Services’ Additional Submission, they concede that this remand proceeding is new agency action and that the Judges have provided the parties with sufficient procedural opportunities to present any new evidence and raise any additional arguments regarding the question the D.C. Circuit remanded. Services’ Additional Submission at 38. But the Services still insist that the Judges may not alter the Service Revenue definition without first identifying legal authority in the Copyright Act for modifying the Initial Determination. In the Services’ view the new agency action avenue provided by the D.C. Circuit merely offers a singular path beyond the Judges’ ability to offer a ‘‘fuller explanation’’ of their previous reasoning for revisiting the definition in the Rehearing Order. According to the Services’ argument, the new agency action provided for in this remand only offers the additional PO 00000 Frm 00048 Fmt 4701 Sfmt 4700 opportunity to offer new reasons supporting any legal authority for altering the Initial Determination’s Service Revenue definition, beyond those that were raised in the appeal. Services’ Additional Submission at 38– 42 C. Reconsideration of Motion for Clarification as Motion for ‘‘Rehearing’’ 158 1. Copyright Owners’ Position Copyright Owners argue that the Judges have the authority to engage in new agency action in this remand proceeding through a reconsideration of the Motion for Clarification as a motion for rehearing, pursuant to 17 U.S.C. 803(c)(2)(A) and 37 CFR 353.1. Copyright Owners urge, however, that proceeding in that fashion would add an entirely unnecessary and complicating step. They again suggest that there is no need to reconsider or recharacterize the Motion for Clarification as a motion for rehearing because the remand itself affords the opportunity for the Judges to take new agency action, which, as in a rehearing, permits them to reconsider evidence and arguments, but, unlike a rehearing, is not limited by the constraints of section 803(c)(2). CO Further Briefing, Tab B at 7–8. Copyright Owners posit that if the Judges engage in new agency action to reconsider the Motion for Clarification as a motion for rehearing under 803(c), and to decide that motion based on all of the evidence in the record supporting the adopted bundle revenue definition and showing the prior bundle revenue definition to be unsupported and unreasonable, they may properly do so. They assert that the while they did not make a request for rehearing on the face of the Motion for Clarification, that is not the same as a finding that the standard could not have been met. The Judges may consider whether, based on the evidence in the record, the rehearing standard has been satisfied on this remand. In Copyright Owners’ view, the Judges could conclude, revisiting on remand the question of whether the rehearing standard has now been met, that Copyright Owners have satisfied the ‘‘exceptional case’’ standard for granting rehearing motions under 158 The Judges consider the briefs filed in response to the Feb. 9, 2022 Order only to the extent that they are responsive to the Feb. 9, 2022 Order, which requested briefing on the specific matter of whether the D.C. Circuit’s Johnson decision permitting the Judges to engage in new agency action in this remand proceeding allows the Judges to engage in new agency action through a reconsideration of Copyright Owners’ February 12, 2018 Motion for Clarification as a Motion for ‘‘rehearing,’’ pursuant to 17 U.S.C. 803(c)(2)(A) and 37 CFR 353.1. E:\FR\FM\10AUR2.SGM 10AUR2 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations section 803(c)(2). Copyright Owners note that if the Judges do engage in new agency action that reconsiders the Motion for Clarification as a motion for rehearing, the Judges should fully explain their reasoning. Id. Tab B at 8– 10.159 ddrumheller on DSK120RN23PROD with RULES2 2. Services’ Position The Services assert that the Judges cannot invoke their rehearing authority by construing the Motion for Clarification as a rehearing motion. They maintain that the D.C. Circuit expressly found that the revision of the Service Revenue definition for bundled offerings does not fall within the Judges’ rehearing authority under section 803(c)(2)(A). The Services assert that Copyright Owners did not satisfy either prong of section 803(c)(2)(A), which authorizes rehearing only ‘‘upon motion of a participant’’ and ‘‘in exceptional cases.’’ They note that the D.C. Circuit agreed with the Judges’ decision not to treat Copyright Owners’ motion as one for rehearing and that the D.C. Circuit also agreed with the Judges’ further finding that ‘‘Copyright Owners’ motion did not meet the exceptional standard for granting rehearing motions.’’ Services’ Further Briefing at 7 (citing Johnson at 390). The Services add their view that the Judges are bound by the D.C. Circuit’s conclusions on this issue. They maintain that because the Judges’ section 803(c)(2)(A) rehearing authority is among the grounds that Johnson addressed and determined, the Judges cannot rely on that authority on remand. Id. at 8–9. The Services urge that the Judges already correctly concluded that the Motion for Clarification was not a motion for rehearing, and note that Copyright Owners never presented their motion as one for rehearing. The Services add that because Copyright Owners did not challenge that decision on appeal, it is too late for them to do so now.160 Id. at 9–10. The Services argue that Copyright Owners’ Motion did not make any attempt to satisfy the exceptional cases 159 With regard to the obligation to fully explain their reasoning for any reconsideration, the Copyright Owners point to United Food & Com. Workers Union, Loc. No. 663 v. U.S. Department of Agriculture., 532 F. Supp. 3d 741, 769 (D. Minn. 2021) (‘‘When an agency takes a new course of action, it must ‘display awareness that it is changing position’ and ‘show that there are good reasons for the new policy.’ ’’), quoting FCC v. Fox Television Stations, Inc., 556 U.S. 502, 515 (2009) (emphasis in original). 160 In fact, the issue of whether to recharacterize the Motion for Clarification as a motion for rehearing is not one raised by Copyright Owners, but by the Judges sua sponte. The Services’ estoppel argument as to the Copyright Owners cannot apply to the Judges’ action. VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 standard set out in 17 U.S.C. 803(c)(2)(A). They argue that Copyright Owners did not purport to identify any new evidence, new legal authority, or even a substantive error in the Judges’ reasoning in the Initial Determination, but instead the motion asserted that the Judges’ inclusion of the definition of service revenue in the Initial Determination was supposedly inadvertent. The Services add that Copyright Owners did not identify any specific evidence in the Phonorecords III record or any aspect of the Initial Determination that suggested the inclusion of this definition was a mistake. Id. at 10. The Services point out that Copyright Owners’ motion did not comply with the procedural requirements for a motion for rehearing. They then urge that the Judges cannot invoke their section 803(c)(2)(A) authority by rewriting a participant’s motion to say it is seeking rehearing when that participant specifically and unambiguously disclaimed any intent to seek rehearing. Id. at 11. The Services note that the Judges previous conclusion that even if the Motion for Clarification had requested rehearing, that motion would not and does not meet that exceptional standard for granting rehearing and failed to make even a prima facie case for rehearing. The Services observe that the Judges apply a strict standard to rehearing motions to prevent parties from using the rehearing process to seek a second bite at the apple by advancing theories and arguments that could have been advanced earlier during the proceeding. Id. at 12. The Services reiterate their view that Copyright Owners’ motion did not point to any evidence in the Phonorecords III record at all, and, that the only evidence in the Phonorecords III record concerning bundles supports the longstanding definition of Service Revenue which has been effective in encouraging the Services to offer bundles that benefit Copyright Owners by growing the market for music streaming services. Id. at 14. The Services finally assert that this is not an extraordinary case where a party has identified an error that, if left uncorrected, would result in manifest injustice. Id. at 15–16. The Services conclude by urging that given this procedural history and the unchanged state of the record since the initial hearing, any claim that Copyright Owners have somehow now satisfied the exceptional case standard would be clear error. Id. at 17. PO 00000 Frm 00049 Fmt 4701 Sfmt 4700 54453 D. Record Evidence Regarding Definition of Service Revenue 1. Copyright Owners’ Position Copyright Owners assert that the prior bundle revenue definition (published in the Initial Determination) failed to address the ‘‘ ‘economic indeterminacy’ problem inherent in bundling’’ appropriately and in a way consistent with Judges’ precedent. CO Initial Submission at 75 (citing Order on Rehearing at 16–18). Copyright Owners proceeded to cite several portions of testimony from the Services’ economic experts who acknowledged this problem. Id. They then point to hearing testimony in which Copyright Owners repeatedly raised the ‘‘economic indeterminacy’’ problem and demonstrated what they characterized as the absurd results to which the prior definition had led. Id. at 76. They point out that under the prior definition, service revenue for bundled subscriptions started with revenues recognized from the bundle (i.e., the price paid by the subscriber) and subtracted ‘‘the standalone published price’’ for all non-music components of the bundle. [REDACTED]. Id. Copyright Owners point out that the Judges already found with respect to other licenses that such an approach is not only fundamentally unfair, but ‘‘absurd.’’ Id. (citing 81 FR 26316, 26382 (May 2, 2016) (webcaster licenses)); see also 83 FR 65210, 65264 (Dec. 19, 2018) (SDARS licenses) (rejecting proposed deductions by service for bundle revenues because of the ‘‘acknowledged ‘economic indeterminacy’ problem inherent in bundling’’). The Copyright Owners concur with the Judges’ correct conclusion that the same reasoning applies to Phonorecords III. Id. at 76–77 (citing Order on Rehearing at 18) (‘‘the ‘economic indeterminacy’ problem inherent in bundling is common to all three proceedings.’’). The Copyright Owners offer that Spotify conceded to this flaw in the definition in the Initial Determination, but offered an alternative that contained the same loophole. Id. at 77–78. Copyright Owners point out that the proponent of a term bears the burden of proof as to adoption. The Judges made clear that the licensee who wishes to offer bundles must bear the burden of providing evidence that might mitigate the acknowledged economic indeterminacy problem inherent in bundling, because any such evidence would be in its possession, not in the possession of the licensors. Id. at 79 (citing SDARS III Determination, 83 FR 65210, 65264) (‘‘bundling [is] undertaken to increase [the Services’] E:\FR\FM\10AUR2.SGM 10AUR2 54454 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations ddrumheller on DSK120RN23PROD with RULES2 revenues and it would be reasonable to assume that [the Services have] information relevant to the economic allocation of the bundled revenue.’’). The Copyright Owners contend they presented unrebutted evidence showing the unreasonableness of the Services’ proposed definition while the Services offered no evidence to support their definition. Id. at 78, 79 (citing Order on Rehearing at 18). Copyright Owners maintain that no Service offered evidence concerning the separate values of the constituent parts of the bundles, or any other evidence concerning the economic allocation of bundled revenue, let alone the reasonableness of the definition in the Initial Determination. Id. at 80. Copyright Owners assert that in the absence of evidence to support the proposed definition, the Judges may adopt or fashion a definition of service revenue for bundled offerings that comports with the record evidence, which is precisely what the Judges did and can, through new agency action, do again. Id. at 81. Copyright Owners dispute the Services’ assertion that there is support for the Phonorecords II approach to bundles in the record of this proceeding. Instead, Copyright Owners argue, the Services’ purported evidence at most supports the benefits of the practice or strategy of bundling. They maintain that the strategy of bundling covered music services with other products or services has nothing to do with whether the Services should be free to reduce the revenue allocable to music to zero. They offer that the definition in the Initial Determination has nothing to do with such benefits, and that those benefits may be equally served by a definition that ensures value is apportioned to the music component in the bundle. CO Reply at 73–76. 2. Services’ Position The Services argue that the evidence in the existing written record addressing bundles shows both that this definition is supported by the Phonorecords II benchmark and that it has proven, industry-wide benefits. Services’ Initial Submission at 68. They offer that the Copyright Owners did not propose an alternative definition of service revenue until after the Judges issued the Initial Determination and that any definition they propose now would fail the basic requirement that the Judges must adopt rules ‘‘on the basis of a written record.’’ Id. (citing 17 U.S.C. 803(a)(1) and 803(c)(3)). Addressing the merits of the definition contained in the Initial Determination, the Services argue that it best serves the goals of the Copyright VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 Act; that as a bright-line, easily administered rule, it continues the broad industry agreement from Phonorecords II. The Services contend the prior definition increases output and incentivizes beneficial price discrimination to reach listeners who would otherwise not pay for music. They argue that the record evidence confirms that the prior treatment of bundles enabled experimentation and variation in the distribution of music with long-term benefits for all parties. They state that Copyright Owners’ argument that Services [REDACTED] also demonstrates the broad benefits of the definition of Service Revenue in Phonorecords II because the record showed that arrangement enabled funneling of many of listeners into fullpriced, full-catalog services—such treatment of bundles enabled the flexibility and price discrimination that yielded beneficial growth of the royalty pool.161 The Services allege that Copyright Owners also ignore the extensive royalties that were generated. They add that with the per-subscriber minimum guarantees that the Copyright Owners will still be paid a fair royalty. The Services then cite several portions of testimony from various Services’ economic experts who point out the realization of an expanded royalty pool, which the Services offer as proving a functioning marketplace. Id. at 68– 74.162 The Services then assert that no other definition of service revenue for bundles that has been before the Judges combines both the administrative simplicity of the Initial Determination’s definition and the broad price discrimination benefits of promoting discounted bundles. They maintain that while neither the Services nor Copyright Owners submitted evidence specifically addressing the way that customers, Services, or Copyright Owners might value the component parts of bundles, such subjective valuations are unnecessary for the Judges to find ample support for the Phonorecords II approach to bundles in the record. Id. at 75–76. The Services also argue that while the Judges’ decision in SDARS III did involve valuation of the music and nonmusic components of a bundle, the resolution in SDARS III is inapposite because, here, the rate structure has a way of ensuring that Copyright Owners 161 Notably, the Services do not deny that the former definition did, in fact, [REDACTED]. 162 The Services’ Reply reiterates this point and offers that the testimony cited by the Copyright Owners also shows why the Initial Determination’s Service Revenue definition works for bundles and grows royalties. Services Reply at 57–58. PO 00000 Frm 00050 Fmt 4701 Sfmt 4700 are fairly compensated for bundles: the statutory minimum payment. Services Reply at 62. E. Analysis and Conclusions Regarding Definition 1. Remand Proceeding as New Agency Action Having considered the entirety of the record of this proceeding, a majority of the Judges (Definition Majority) conclude that this remand constitutes ‘‘new agency action’’ and meets all of the criteria to qualify as new agency action. The Judges thus have the opportunity to consider the issue afresh consistent with their procedural rules regarding remands. The Definition Majority finds that it is unnecessary to attempt to distinguish new ‘‘agency action’’ from ‘‘new agency action.’’ Neither approach is endorsed clearly by the varied judicial interpretations of a new agency action. See R.J. Krotoszynski, Jr., Administrative Law Discussion Forum: ‘‘History Belongs to the Winners’’: the Bazelon-Leventhal Debate and the Continuing Relevance of the Process/ Substance Dichotomy in Judicial Review of Agency Action, 58 Admin. L. Rev. 995 (Fall 2006). As noted by Judge Bazelon, the D.C. Circuit ‘‘believed in process-based review, [but] he argued that it was improper for judges to prescribe specific procedures.’’ Id. at 1001. Judge Bazelon’s remand orders focused on providing ‘‘genuine opportunities to participate in a meaningful way’’ and ‘‘genuine dialogue’’ with interested parties, while leaving the agency ‘‘free to decide which specific procedures to undertake.’’ Id. Several reported cases point to new action as an alternative to a fuller explanation. But few define ‘‘new agency action’’ other than to say, as did the Johnson court, that the agency must take it ‘‘accompanied by the [unspecified] appropriate procedures.’’ Johnson, 969 F.3d at 392. Parties to the original action, already familiar with the issue and the factual and legal background, recognized that the D.C. Circuit identified the adoption of a modified definition in the Determination as one of three issues on remand. In repeated rounds of remand submissions, both the Services and the Copyright Owners included the definition issue. The Judges were not satisfied with the parties’ lack of focus on the issue, however, and ordered expressly further briefing on the new agency action issue and sub-issues relating to the adoption of a definition of Service Revenue as it relates to E:\FR\FM\10AUR2.SGM 10AUR2 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations ddrumheller on DSK120RN23PROD with RULES2 bundled service offerings. See (Dec. 9 Order) at 4; Sua Sponte Order Regarding Additional Briefing (Feb. 9, 2022). New agency action is not synonymous with justification, or confirmation, of the prior action. New agency action is a procedural mechanism for reconsideration of the record, reopening the record for additional evidence and argument, and adoption of a conclusion based on the expanded record. In this instance, the presentations, written and oral, of participants on remand, together with a re-examination of the original record, support reversion to the definition originally announced in the Initial Determination. Ultimately, given repeated opportunities for legal analysis on the issue, both sides agreed that the remand proceeding itself, with ample notice and multiple opportunities for input was sufficient to constitute new agency action. See CO Further Briefing at 3, 7. The Services argued, however, that notwithstanding this appropriate new agency action, the Judges remained without authority to adopt the revised definition as a term governing the royalty rates determined in this proceeding. Their arguments regarding procedures undertaken in the Determination are superseded by the Judges’ conduct of extensive remand proceedings.163 The gravamen of the Administrative Procedure Act is transparency in agency 164 rulemaking. Agencies must publish notice of their intentions, provide opportunities for interested parties to comment and object, and finalize regulations only after reconciling objections with the policies and purposes of proposed regulations. The adjudication of this remand proceeding was conducted openly. Interested parties had ample opportunity to object, to comment, and to brief legal and factual issues relating to the Judges’ approach to promulgating an appropriate definition of bundled service revenue. The present analytic approach merely takes the position that the Judges engaged in new agency action by conducting a fully open and broadly explored remand proceeding. Unlike a 163 Furthermore, the issue of the Judges’ authority to take an action in issuing the Determination is moot. The Judges, after new agency action, have chosen not to defend the definition in the Determination but rather to conclude, following that new agency action, that the definition in the Initial Determination is more appropriate in these circumstances. Whether the Judges had the authority in the first instance is not at issue, as they are not repeating the former action. 164 The proceedings of the Copyright Royalty Board (CRB) are subject to the standards of the Administrative Procedure Act. See 17 U.S.C. 803(a)(1). VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 rehearing or exercise of continuing jurisdiction, this remand proceeding is not limited by the constraints of sections 803(c)(2) or 803(c)(4). Contrary to the Services’ assertion, the Judges address the issue on which the D.C. Circuit remanded, the need to exercise authority within the lines drawn by the authorizing statute. This remand proceeding does not, therefore, violate the D.C. Circuit’s order. The Johnson opinion clearly states the two paths by which the Judges may address the issues presented to them on remand; they may either (1) provide ‘‘a fuller explanation of the agency’s reasoning at the time of the agency action[,]’’ or (2) to take ‘‘new agency action’’ accompanied by the appropriate procedures. Johnson, 369 F.3d at 392. The Judges chose to pursue the second option: this new agency action. The Judges reiterate: the Services concede that, through this proceeding the Judges have provided the participants with adequate procedural opportunities to present any new evidence on the proper Service Revenue definition for bundles. The Judges also acknowledge, but disagree with, the Services’ position that that they must return to the issues as they were presented after issuance of the Initial Determination, regardless of the admittedly complete and valid remand procedure, which constitutes new agency action. The Judges (the majority on this issue) determine that any confining action on remand to the provisions of sections 803(c)(2)(A) or 803(c)(4) would misconstrue the clear expression of the ‘‘new agency action’’ alternative presented by the D.C. Circuit,165 as well as chapter 8 of title 17. As the Copyright Owners correctly observed, in a remand proceeding, the Judges are not required to undertake any of the procedural steps set forth in section 803(b) nor are the Judges compelled to consider or be limited by sections 803(c)(2)(A) or 803(c)(4). The statute only requires that the Judges’ remand proceedings are in accordance with section 803(a).166 The D.C. Circuit observed that the Judges have ‘‘considerable freedom to determine [their] own procedures.’’ SoundExchange v. CRB, 904 F.3d 41 at 61. The D.C. Circuit also cautions that 165 The case that the D.C. Circuit points to for the new agency action path clarifies that ‘‘An agency taking this [new agency action] route is not limited to its prior reasons but must comply with the procedural requirements for new agency action.’’ Regents, 140 S. Ct. at 1908). 166 ‘‘The court [United States Court of Appeals for the District of Columbia Circuit] may also vacate the determination of the Copyright Royalty Judges and remand the case to the Copyright Royalty Judges for further proceedings in accordance with subsection (a).’’ 17 U.S.C. 803(d)(3). PO 00000 Frm 00051 Fmt 4701 Sfmt 4700 54455 such flexibility must be exercised within the lines drawn by the authorizing statute. Here, the Judges operate within the lines drawn with respect to remand proceedings set forth in chapter 8 of title 17. 2. ‘‘Fuller Explanation’’ of Modification to Initial Determination Case law regarding development of a ‘‘fuller explanation’’ of an agency’s action emphasizes that the agency cannot adopt post hoc reasoning on the same record. See, e.g., SEC v. Chenery Corp., 332 U.S. 194, 201 (1947) (after remand, agency bound to ‘‘deal with the problem afresh . . . .’’). Certainly, adopting a post hoc argument of appellate counsel, just because it offers a rationale for the agency’s original action is impermissible.167 On the other hand, if the record in the initial proceeding is sufficiently robust to support a reinterpretation or additional reasoning, the agency may justify its initial action with that ‘‘fuller explanation’’ without considering any new evidence. See, Fisher v. Pension Benefit Guar. Corp., 468 F.Supp.3d 7, 20 (D.C.D.C. 2020), aff’d Fisher v. Pension Benefit Guar. Corp., 994 R.3d 664 (D.C. Cir. 2021), rehearing en banc denied, Fisher v. Pension Ben. Guar. Corp., 2021 U.S. App. LEXIS 18793 (D.C. Cir., June 23, 2021) (requirement of new evidence a ‘‘novel proposition of law’’ without precedent). On remand, an agency may elaborate on its prior reasoning, but it may not provide new reasons for the original decision. Fisher, 994 F.3d at 669. If the Judges had chosen in this remand to rest on their Determination regarding the service revenue definition, they might have done so only if they could elaborate on the existing record.168 In the alternative, the Judges issue a new decision after new agency action. Id. The Judges, having engaged in new agency action to settle on the definition of service revenue for bundled offerings, do not find a need to address the statutory avenues or the confines that are provided for rehearing or continuing jurisdiction, nor do the Judges pursue the propriety of reconsideration of the 167 A rationalization is not post hoc simply because it is iterated by counsel. Denomination of a rationalization as post hoc is a matter of timing, not of the offeror. 168 In this instance, had the Judges decided to keep the definition in the Determination, they probably could have given a fuller explanation based on the record in the underlying proceeding. Because the Judges have opted to rely on the freshlook approach in the ‘‘new agency action’’ alternative and because the prior definition is appropriate given adoption of the PR II rate structure, development of that fuller explanation based on the record is unnecessary. E:\FR\FM\10AUR2.SGM 10AUR2 54456 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations Motion for Clarification as a motion for rehearing.169 ddrumheller on DSK120RN23PROD with RULES2 3. Substantive Analysis of Dueling Definitions of Bundled Revenue The fundamental difference between the impact of the two alternative definitions is simply stated: Under the Initial Determination: downstream bundling and its price discriminatory effect would be incentivized by a royalty structure that reflects the lower WTP of consumers who subscribe by paying for a Bundle; Under the Determination: downstream bundling and its price discriminatory effect would not be incentivized by a royalty structure that reflects the lower WTP of consumers who subscribe by paying for a Bundle. To explain this difference, the Judges find it helpful to describe (as in the Determination and Dissent) how bundling facilitates price discrimination and how lower royalties for bundled streaming services incentivize such bundling. Price discrimination occurs when a seller offers different units of output at different prices. See, e.g., H. Varian, Intermediate Economics at 462 (8th ed. 2010). The benefit to the seller arises from attempting to ‘‘charge each customer the maximum price that the customer is willing to pay for each unit bought.’’ R. Pindyck & D. Rubinfeld, Microeconomics at 401 (8th ed. 2013). For all goods, and intellectual property goods such as copyrights in particular,170 the social benefit is that price discrimination more closely matches the quantity sold with the competitive quantity as the seller or licensor better aligns the price with the WTP of different categories of buyers or licensees. See W. Fisher, Reconstructing the Fair Use Doctrine, 101 Harv. L. Rev. 1659, 1701 (1988). A seller can engage in price discrimination in several ways. One form is known as ‘‘second-degree price discrimination,’’ by which buyers selfsort the packages and quantities they purchase.171 See W. Adams & J. Yellen, 169 The Judges also find no need to consider any inherent authority that may remain for consideration. 170 Streamed copies of intellectual property, such as musical works and sound recordings, have a marginal production cost of essentially zero, making price discrimination particularly beneficial, because charging any positive price, even to a buyer with the lowest WTP, still exceeds the zero marginal production costs. See Dissent at passim. 171 ‘‘First-degree’’ price discrimination is a hypothetical construct by which a seller can identify the WTP of every buyer. ‘‘Third-degree’’ price discrimination occurs when the seller offers different prices to buyers based on their different characteristics (e.g., a senior citizen discount). See Pindyck & Rubinfeld, supra, at 402, 404–05. VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 Commodity Bundling and the Burden of Monopoly, 90 Q. J. Econ. 470, 476 (1976) (the profitability of bundling ‘‘stem[s] from its ability to sort customers into groups with different reservation price [WTP] characteristics.’’). Bundling, i.e., the ‘‘practice of selling two or more products as a package,’’ Pindyck & Rubinfeld, supra at 419, is thus a type of second-degree price discrimination. See A. Boik & H. Takahashi, Fighting Bundles: The Effects of Competition on Second Degree Price Competition, 12 a.m. Econ. J. 156, 157 (2020). The applicability of these basic economic principles was understood and explained by the parties’ experts at the hearing. See, e.g., 3/15/17 Tr. 1224– 25 (Leonard) (Google’s economic expert testifying that price discrimination through bundling is ‘‘very, very common . . . even by pretty competitively positioned firms . . . to sort out customers into willingness-topay groups.’’); 3/30/17 Tr. 3983 (Gans) (Copyright Owners’ economic expert acknowledging that bundling is a form of price discrimination); see also Dissent at 69 (same). How does this downstream (retail level) benefit of price discrimination impact the setting of upstream royalty rates? As the Majority explained (in summarizing the Services’ expert testimony) the linkage is explained by the economic concept of ‘‘derived demand’’: [M]ultiple pricing structures necessary to satisfy the WTP and the differentiated quality preferences of downstream listeners relate directly to the upstream rate structure to be established in this proceeding. Professor Marx opines that the appropriate upstream rate structure is derived from the characteristics of downstream demand. 3/20/ 17 Tr. 1967 (Marx) (rate structure upstream should be derived from need to exploit WTP of users downstream via a percentage of revenue). This upstream to downstream consonance in rate structures represents an application of the concept of ‘‘derived demand,’’ whereby the demand upstream for inputs is dependent upon the demand for the final product downstream. Id.; see P. Krugman & R. Wells, Microeconomics at 511 (2d ed. 2009) (‘‘[D]emand in a factor market is . . . derived demand . . . [t]hat is, demand for the factor is derived from the [downstream] firm’s output choice’’). Determination at 19; accord Dissent at 32 (noting that ‘‘the upstream demand of the interactive streaming services for musical works (and the sound recordings in which they are embodied)—known as ‘‘factors’’ of production or ‘‘inputs’’—is derived from the downstream demand of listeners to and users of the interactive streaming services . . . This interdependency PO 00000 Frm 00052 Fmt 4701 Sfmt 4700 causes upstream demand to be characterized as ‘‘derived demand.’’). In the present proceeding, the PR IIbased benchmark embodies the parties’ negotiated definition of Bundled Revenue for purposes of calculating royalties on bundled interactive offerings. This is definition in the Initial Determination. Copyright Owners’ preferred definition for Bundled Revenue—the Determination’s definition—would not only ignore this agreed-upon definition, but would also de-link the royalty rate from the WTP of purchasers of bundles.172 The Judges recognize that Copyright Owners have expressed concern the Services could use such bundling in order to diminish revenue otherwise payable on a higher royalty tier. However, the Majority noted that the evidence indicated such diminishment only occurred ‘‘in some cases.’’ Clarification Order at 17. Thus, the Judges find that eliminating the incentive for price discrimination via bundling would be a disproportionate response and inconsistent with the broad price discriminatory PR II-based benchmark they find useful in this proceeding. Expert testimony in this regard is ‘‘substantial evidence’’ on which the Judges can rely. For example, the D.C. Circuit also relied in Johnson on the testimony of the same witness, Spotify’s economic expert witness, Professor Marx, who explained how a downstream ‘‘lower willingness (or ability) to pay’’ among some cohorts of consumers supports definitional terms, for student and family subscribers, that lower royalty rates in order to further ‘‘economic efficiency’’ in a manner that 172 To see the incentivizing effect of the link between the royalty level and variable WTP, consider the following example. Assume a hypothetical bundle consists of a subscription to the ‘‘Acme’’ interactive music streaming service and the sports service NFL Sunday Ticket. Assume also that Acme and NFL Sunday Ticket have standalone monthly subscription prices of $9.99/month and $149.99/month respectively, so that purchasing both separately would cost $159.98/month. But assume the bundle price is only $140/month. Acme’s purpose in bundling its interactive music streaming service subscription offering with NFL Sunday Ticket would be to attract customers who had a WTP for the standalone Acme service below $9.99/month, but a WTP at or above the $140/ month for the bundle. Under the definition in the Determination, royalties would be paid on the standalone $9.99/ month Acme price. But the purpose of the bundling was to attract subscribers who would not pay the standalone $9.99/month price, so no such wouldbe subscribers would sign-up, and no royalties would be generated by them. By contrast, under the Initial Determination, the standalone price of NFL Sunday Ticket, $159.98/ month, would be subtracted from the $140/month bundle price. Although that would preclude a payment of royalties on a revenue prong, royalties still would be paid, under a different tier or on the mechanical floor. E:\FR\FM\10AUR2.SGM 10AUR2 ddrumheller on DSK120RN23PROD with RULES2 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations ‘‘still allows more monetization of that provision of that service.’’ Johnson at 392–93. Broadening her lens, Professor Marx also explained that this price discriminatory approach is appropriate ‘‘across all types of services and subscribers,’’ as in ‘‘[t]he current law [and in the PR II-based benchmark]’’ which ‘‘accommodates . . . adsupported services . . . and ‘bundled services’ through different rate provisions.’’ Marx WRT ¶ 41 (emphasis added). See also 3/21/17 2182–83 (Hubbard) (Amazon’s expert witness testifying that ‘‘Prime Music, which is bundled with an Amazon Prime service . . . sort[s] out customers’ willingness to pay, with an idea of trying to maximize the number of customers,’’ and agreeing that this approach constitutes ‘‘sorting by way of bundling.’’) (emphasis added). Further, Professor Hubbard opined that, given the revenue attribution ‘‘measurement problem’’ associated with bundled products, the ‘‘Phonorecords II’’ approach ‘‘with the different categories and the minima . . . address this sort of problem [in] a very good way.’’ 3/15/17 Tr. 1221 (Hubbard). As in the case of family and student price discrimination, the beneficial effect of such differential pricing was supported by industry witnesses as well as expert witnesses. See, e.g., Mirchandani WDT ¶ 71 (Amazon executive citing the Phonorecords IIbased benchmark provisions regarding bundling that ‘‘allowed Amazon to bundle Prime Music with Amazon Prime, enabling Amazon to bring a limited catalog of music [REDACTED]’’). In sum, the same type of witness testimony that the D.C. Circuit found sufficient to support price discriminatory student and family plans also supports the use of the price discriminatory bundled definition contained in the Initial Determination. Given the overall benefits from price discrimination, at first blush it is curious that Copyright Owners would risk ‘‘leaving money on the table’’ by removing the royalty-based incentive for price discrimination via bundling. The Judges have identified this problem earlier in this Initial Ruling, in connection with the broader issue of the overall beneficial price discriminatory structure of the PR II-based benchmark. As the Judges noted in that general price discrimination context, Copyright Owners’ own expert economic witnesses acknowledged that they would not irrationally ‘‘leave money on the table.’’ In fact, Copyright owners’ aim, according to that testimony, is to create an unregulated space—per the Bargaining Room theory—and to use VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 their complementary oligopoly power to negotiate price discriminatory rates (in bundles or otherwise), which would free them from the section 801(b)(1) requirements of reasonableness and fairness. The Judges further find that their prior ruling on this issue in SDARS III is distinguishable. There, a proffered bundled revenue definition eliminated the payment of any royalty at all. Copyright Owners quite correctly describe that result as ‘‘absurd,’’ but that is not the result here. Rather, in the present case, the parties’ negotiated an approach that the Judges adopted in the Initial Determination requiring royalties to be paid on interactive services bundled with other products or services. Even more distinguishable is Copyright Owners’ assertion that Web IV provides support for their preferred definition of service revenue. The argument is immediately suspect, because Web IV involved per-play royalty rates—not percent-of-revenue rates, making the definition of revenue wholly inapposite. Further, the discussion of the price of an ‘‘ice cream cone’’ in Web IV—on which Copyright Owners rely—had nothing to do with bundling or isolating the WTP for different products or services. Rather, there the Judges criticized a bizarre argument made by a licensee (who had a quantity discount for plays steered in its direction), that was tantamount to arguing that if a vendor sells one ice cream cone for $1.06 but a buyer could buy two for $1.06, that the market price of an ice cream cone is thus only $.06. This argument was indeed fallacious, because the price of an ice cream cone would be the average of the total cost for the two cones, i.e., $.53/cone. Here, the issue is how to address the WTP of different classes of buyers with heterogeneous WTP, not the pricing of a discount for all purchasers buying the same quantity. The parties utilized the Bundled Revenue definition from the PR II-based benchmark (and in the Initial Determination) to address the indeterminacy inherent in the variable WTP among purchasers of the bundles, by setting floors and minima, rather than attempt to sort out the WTP of individual (or individual blocs) of subscribers.173 173 Accordingly, Copyright Owners’ assertion that the Services did not satisfy their burden of proof with regard to the Bundled Revenue definition misses the point. The Services’ burden was to show the reasonableness of utilizing the Bundled Revenue definition in the PR II-based benchmark, not to show that their proffered approach measured the WTP of individual subscribers (or blocs of subscribers). Such an alternative approach might have had merit but no alternative approach was presented to the Judges. PO 00000 Frm 00053 Fmt 4701 Sfmt 4700 54457 For the foregoing reasons, the Judges find that the definition in the Initial Determination (unlike the definition in the Determination) is consistent with the Judges’ other substantive rulings herein. That is, just as the Majority abandoned its Bundled Revenue definition in its Initial Determination because it refused to credit the PR IIbased benchmark (even as ‘‘guidance’’), the Judges here do partially rely on the PR II-based benchmark, and thus find that it supports the Bundled Revenue definition contained in the Initial Determination. 4. Application of Four Itemized Statutory Factors As the forgoing analysis explains, bundling is a form of price discrimination. Accordingly, the Judges’ explanation of how price discriminatory rates in the PR II-based benchmark interrelate with the Factor (A) through (D) objectives in section 801(b)(1) are equally applicable here. Accordingly, the Judges adopt by reference their discussion of those four factors set forth supra in connection with the PR IIbased benchmark, and find that there is no basis pursuant to those four factors to adjust the PR II-based benchmark definition of Bundled Revenue. V. Conclusion On the basis of the foregoing analyses, and in consideration of the entirety of the record, the Judges make the following determination relating to the issues on remand from the D.C. Circuit. To be clear, the Judges are not declaring that an alternative Bundled Revenue definition and/or alternative rates and structures for bundle, might not have been preferable. See 4/15/17 Tr. 5056–58 (Katz) (‘‘[I]f someone had a proposal [with] a specific reason why we should adjust this minimum that’s something I would have examined’’); see also 3/15/17 Tr. 1227–28 (Leonard) (Google’s economic expert testifying that ‘‘if somebody had . . . suggest[ed] . . . a different sort of bucket that should be created . . . that’s a good idea.’’). But Copyright Owners did not propose such alternatives at the hearing, and the alternative in their Motion for Clarification simply eviscerated the ‘‘derived demand’’-based link between royalties and bundled offerings. As the Judges have noted supra, in the words of Judge Patricia Wald, all judges are cabined by the record evidence introduced by the parties. Therefore (in the absence of a way in which to synthesize the parties’ proposals in a manner that does not ‘‘blindside’’ the parties) the Judges must choose between the proposals that are in the record, not potentially superior proposals that are not in the record. Here, the Judges favor the Bundled Revenue definition in the Initial Determination that was negotiated by the parties, incentivizes price discrimination and pays royalties on the bundled music, over the substituted definition in the Determination pursued by Copyright Owners that would eliminate price discrimination, except under the terms Copyright Owners could impose via their complementary oligopoly power, and without regard to the statutory requirements of a ‘‘reasonable rate’’ and a ‘‘fair income’’ for the Services. E:\FR\FM\10AUR2.SGM 10AUR2 54458 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations As noted at the outset, the headline rate for all offerings throughout the Phonorecords III period shall be as follows: 2018–2022 ALL-IN HEADLINE ROYALTY RATES Percent of Revenue ................................................................................. In all other respects, the rates and rate structure of the PR II-based benchmark shall be effective as the rates and structure throughout the Phonorecords III period. The definition of Service Revenue for bundled offerings throughout the Phonorecords III period shall be the definition contained in the Initial Determination. VI. Order In light of the foregoing analyses and conclusions, the Judges hereby order that the participants in this remand proceeding prepare and submit regulatory provisions consistent with this ruling.174 The participants shall file agreed regulatory language within ten days of the date of this ruling. The Judges further order that if the participants cannot agree on a joint submission, the Judges will accept separate submissions respectively from (1) Copyright Owners and (2) Services, jointly. In absence of an agreed submission, the participants shall file 2018 2019 2020 2021 2022 11.4% 12.3% 13.3% 14.2% 15.1% separate submissions not later than 15 days after the date of this ruling.175 The Judges further order that parties shall not file, and the Judges shall not consider, briefing or legal argument beyond necessary explanatory notes to the proposed language, section by section, not to exceed 250 words per proposed section.176 The Judges specifically admonish the parties that they shall not use these submissions as a basis to object to this Initial Ruling, either explicitly or implicitly by proposing regulatory provisions inconsistent with this Initial Ruling. The Judges further order that, within 30 days of the date of this Initial Ruling and the attendant dissenting documents, the parties shall file an agreed redacted version of this Initial Ruling, and the dissents, for public viewing. After the Judges have reviewed the parties’ regulatory submissions, the Judges shall adopt and format the necessary regulatory language format terms relevant to this ruling and issue a restricted Initial Determination after Remand, which shall embody their determination of rates and terms. The parties will have an opportunity to suggest redactions from the Initial Determination after Remand before it is issued as a public version. The parties shall not file any motions seeking rehearing or reconsideration of this Initial Ruling. Subsequent to the Judges’ issuance of their Initial Determination after Remand as identified in the immediately preceding paragraph, any party may file a Motion for Rehearing within 15 days of the issuance of said Initial Determination after Remand. After ruling on any and all Motions for Rehearing as identified in the immediately preceding paragraph, the Judges shall issue a Final Determination after Remand. So ordered. Issue Date: July 1, 2022. Stephen S. Ruwe, Copyright Royalty Judge. David R. Strickler, Copyright Royalty Judge. Suzanne M. Barnett, Chief Copyright Royalty Judge. ADDENDUM TO FINAL RULING AND ORDER % of Service provider revenue (percent) ddrumheller on DSK120RN23PROD with RULES2 Offering TCC % or TCC amount ‘‘Mechanical-only’’ royalty floor 15 cents per subscriber per month. 30 cents per subscriber per month. 50 cents per subscriber per month. 25 cents per month for each Active Subscriber during that month. n/a. n/a. n/a. n/a. n/a. Standalone Non-Portable Subscription Offering— Streaming Only. Standalone Non-Portable Subscription Offering— Mixed. Standalone Portable Subscription Offering ........... 10.5 Bundled Subscription Offering .............................. 10.5 The lesser of 22% of TCC for the Accounting Period or 50 cents per subscriber per month. The lesser of 21% of TCC for the Accounting Period or 50 cents per subscriber per month. The lesser of 21% of TCC for the Accounting Period or 80 cents per subscriber per month. 21% of TCC for the Accounting Period ............... Mixed Service Bundle ........................................... Limited Offering ..................................................... Paid Locker Service .............................................. Purchased Content Locker Service ...................... Free nonsubscription/ad-supported services free of any charge to the End User. 11.35 10.5 12 12 10.5 21% of TCC for the Accounting Period ............... 21% of TCC for the Accounting Period ............... 20.65% of TCC for the Accounting Period .......... 22% of TCC for the Accounting Period ............... 22% of TCC for the Accounting Period ............... 174 The Judges adopt this process in order to avoid a dispute regarding the regulatory provisions issued in connection with their ruling. Because this is a remanded proceeding, the Judges are not restricted to the procedures that would control in an original proceeding, and are exercising their authority to ‘‘make any necessary procedural . . . VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 10.5 10.5 rulings in any proceeding under this chapter.’’ 17 U.S.C. 801(c). 175 In their agreed upon or separate submissions, the parties shall address the issue identified in note 135 infra, regarding Copyright Owners’ assertion that the Services omitted from their proposed PO 00000 Frm 00054 Fmt 4701 Sfmt 4700 subpart C rates a portion of the Phonorecords II rates. 176 A section of the regulations is designated by a number following the decimal after the part number, for example, § 385.5. The regulations relevant to this proceeding are found in part 385. E:\FR\FM\10AUR2.SGM 10AUR2 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations B. Order 43 on Phonorecords III Regulatory Provisions (Public Version With Federal Register Naming and Formatting Conventions) Introduction The present Order concerns a single issue in dispute among the parties 177 regarding regulatory language implementing the Judges’ Initial Ruling and Order after Remand (‘‘Initial Ruling’’) entered in this proceeding.178 Subsequent to filing dueling submissions (see footnote 2 infra), the parties filed a Joint Submission, informing the Judges that they had ‘‘agree[d] on all of the regulatory language’’ except for certain rate percentages contained in Table 2 of the proposed § 385.21. Joint Submission . . . Regarding Regulatory Provisions Following Initial Ruling and Order (after Remand) at 1 (Nov. 30, 2022) (‘‘Joint Submission’’) (eCRB no. 27337). The Regulatory Language in Dispute The dispute between the parties is whether the Judges should adopt in the Phonorecords III regulations: (1) the several ‘‘Total Content Cost’’ (‘‘TCC’’) rates 179 set forth in the Phonorecords IIbased benchmark; or (2) the single 26.2% TCC rate discussed in the Initial Ruling. This dispute relates to nine offerings made by interactive streaming services, as detailed below: Copyright owners’ proposal (percent) Offering 54459 Services’ proposal Standalone Non-Portable Subscription Offering—Streaming Only. Standalone Non-Portable Subscription Offering—Mixed ........... 26.2 Standalone Portable Subscription Offering ................................ 26.2 Bundled Subscription Offering .................................................... Free nonsubscription/ad-supported services free of any charge to the End User. Mixed Service Bundle ................................................................. Purchased Content Locker Service ............................................ Limited Offering ........................................................................... Paid Locker Service .................................................................... 26.2 26.2 The lesser of 22% of TCC for the Accounting Period or 50 cents per subscriber per month. The lesser of 21% of TCC for the Accounting Period or 50 cents per subscriber per month. The lesser of 21% of TCC for the Accounting Period or 80 cents per subscriber per month. 21% of TCC for the Accounting Period. 22% of TCC for the Accounting Period. 26.2 26.2 26.2 26.2 21% of TCC for the Accounting Period. 22% of TCC for the Accounting Period. 21% of TCC for the Accounting Period. 20.65% of TCC for the Accounting Period. 26.2 Sources: Offering column text from Exhibit A to Joint Submission . . . Regarding Regulatory Provisions Following Initial Ruling and Order (after Remand) at 17 (Nov. 30, 2022) (eCRB no. 27338); Services’ Proposal column text from Services’ Joint Submission of Regulatory Provisions Ex. A at 11 (July 18, 2022) (eCRB no. 27005). Judges in connection with this proceeding. The Issue At a high level, the remaining regulatory issue is the following: Background Whether a 26.2% TCC rate identified in the hearing record, and discussed both on appeal and on remand by the D.C. Circuit, should substitute for TCC rates in the Phonorecords III period, or whether these uncapped TCC rates should be set at the specific levels ranging between 20.65% and 22% set forth in the Phonorecords II-based benchmark adopted by the Judges in the Initial Ruling. ddrumheller on DSK120RN23PROD with RULES2 To frame, address, and rule on this issue, in this Order the Judges place the parties’ dispute in the context of the prior rulings by the D.C. Circuit and the 177 The parties who have joined on this dispute (through filings after the issuance of the Initial Ruling) are the National Music Publishers’ Association and Nashville Songwriters Association International (collectively, ‘‘Copyright Owners’’) and Amazon.com Services LLC, Google LLC, Pandora Media, LLC, and Spotify USA Inc. (collectively, the ‘‘Services’’). (Copyright Owners have informed the Judges that another party, George Johnson, joins in Copyright Owners’ position with respect to the issue considered in this Order.) 178 The Judges instructed the parties to ‘‘prepare and submit regulatory provisions consistent with this ruling.’’ Initial Ruling and Order after Remand VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 On January 5, 2016, the Judges initiated proceedings to determine the appropriate mechanical license royalty rates and terms for the January 1, 2018 to December 31, 2022 period. See Notice Announcing Commencement of Proceedings in Phonorecords III, 81 FR 255 (Jan. 5, 2016). After the parties filed their written and rebuttal testimonies and engaged in discovery, they participated in a five-week evidentiary hearing presided over by the Judges. See at 114 (July 1, 2022) (eCRB nos. 26938, 27063). The Judges further instructed that, ‘‘if the participants cannot agree on a joint submission, the Judges will accept separate submissions respectively from (1) Copyright Owners and (2) Services, jointly.’’ Id. The parties did not initially file an agreed-upon joint submission as to regulatory provisions, but rather filed the permitted separate submissions. 179 TCC is defined in the Initial Ruling as ‘‘a shorthand reference to the extant regulatory language describing generally the amount paid by a service to a record company for the section 114 right to perform digitally a sound recording.’’ Initial Ruling at 4 n.8 (citations omitted). PO 00000 Frm 00055 Fmt 4701 Sfmt 4700 Determination of Royalty Rates and Terms for Making and Distributing Phonorecords, 84 FR 1918, 1920, 1923– 1925 (Feb. 5, 2019).180 In the Majority Opinion, the Judges adopted a ‘‘greater-of’’ royalty rate structure for the mechanical license, which contained a TCC rate applicable to all categories of offerings.181 See 84 FR 1963; see also Johnson v. Copyright Royalty Board, 969 F.3d 363, 372 (D.C. Cir. 2020) (summarizing the Majority Opinion). More particularly, the Majority adopted the following rates and rate structure: 180 The Determination was not unanimous. Judge David Strickler dissented from the Majority’s setting of the TCC rate, and he proposed that the appropriate rates should essentially be those proposed in the Phonorecords II-based benchmark proposed by several of the Services. Thus, for clarity, this Order refers to the ‘‘Majority Opinion’’ and the ‘‘Dissenting Opinion,’’ rather than the ‘‘Final Determination,’’ when discussing the respective opinions. 181 The other prong in the ‘‘greater-of’’ rate structure is the percent-of-revenue generated by the interactive streaming service, i.e., ‘‘service revenue.’’ E:\FR\FM\10AUR2.SGM 10AUR2 54460 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations 2018–2022 ALL-IN ROYALTY RATES: THE GREATER OF: 2018 Percent of Revenue ............................................................. Percent of TCC .................................................................... ddrumheller on DSK120RN23PROD with RULES2 Majority Opinion at 1918, 1960. The Services appealed.182 Among their arguments were the assertions— pertinent to this Order—that the Majority: (i) violated the Services’ procedural right to fair notice by choosing a structure that was not advanced by any party; (ii) acted arbitrarily and capriciously by simultaneously combining a TCC prong (phased-in to 26.2% of TCC) with an increase in the percentages on the revenue prong (phased-in to 15.1%); and (iii) failed to reasonably explain its rejection of the Phonorecords II settlement as a benchmark. Johnson, supra, at 376, 380–81.183 Copyright Owners argued in opposition that: (i) the Services’ procedural rights were not violated because ‘‘every component’’ of the Majority’s approach was contained in the hearing record; (ii) the Majority’s rate and rate structure rulings were well-reasoned, factually supported and, therefore, not arbitrary and capricious; and (iii) sufficient reasons existed in the record to support the Majority’s rejection of the Phonorecords II-based benchmark. Johnson, supra, at 382–383; 387. The D.C. Circuit vacated and remanded. More particularly, Johnson holds as follows: 1. The Majority Determination ‘‘failed to provide adequate notice of the drastically modified rate structure [they] ultimately adopted,’’ which was beyond ‘‘a reasonable range of contemplated outcomes’’ in ‘‘the parties’ pre-hearing proposals, the arguments made at the evidentiary hearing, and the preexisting rate structures.’’ Johnson at 381– 82. Accordingly, as to this issue, ‘‘[i]f the [Judges] wish[ ] to pursue [their] novel rate structure, [they] will need to reopen the evidentiary record.’’ Id. at 383. 2. The appellate issue of whether the Majority’s adoption of the (phased-in) 26.2% TCC royalty rate was ‘‘arbitrary and capricious’’ could not be addressed—given the absence of ‘‘adequate notice’’ cited in point (1) above. Id. 3. The Majority’s derivation, calculation and application of the royalty rate of 15.1% on the revenue prong was proper.184 The D.C. 182 The Copyright Owners and George Johnson also appealed; all three parties’ appeals were consolidated by the D.C. Circuit. Johnson at 375. 183 The annual phased-in rates are set forth in the Table supra. 184 The italicization of the word ‘‘application’’ serves to foreshadow a critical point discussed infra: The D.C. Circuit did not affirm any VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 2019 11.4% 22.0% 12.3% 23.1% Circuit explained that, as to this issue, the Majority had engaged in the ‘‘type of linedrawing and reasoned weighing of the evidence [that] falls squarely within the [Judges’] wheelhouse as an expert administrative agency.’’ Johnson at 386. More particularly, the D.C. Circuit approved of the Judges’ reliance on ‘‘substantial evidence’’ in the form of expert testimony to set the 15.1% service revenue rate. Johnson, at 384–85 (emphasis added). See also id. at 388 (finding ‘‘substantial evidence’’ for the Judges’ finding that an increase in the mechanical royalty rate was necessary to address a ‘‘marked decline in mechanical royalty income. . . .’’). 4. The Majority’s rejection of the Phonorecords II-based benchmark is remanded because the D.C. Circuit ‘‘cannot discern the basis on which the [Judges] rejected the Phonorecords II rates as a benchmark in [their] analysis, that issue is remanded to the [Judges] for a reasoned analysis.’’ Johnson at 387. On remand, the Judges adopted procedures that mainly followed the parties’ requests. More particularly, the Judges followed the D.C. Circuit’s directive and reopened the evidentiary record to receive evidence and testimony relating to the TCC issues. See Order Regarding Proceedings on Remand at 2 (Dec. 15, 2020). The postremand supplementary record added: (1) rate evidence for the 33-months from January 2018 through September 2020, when the parties operated under the Majority’s new (but subsequently vacated) regulations including the TCC rates; and (2) new testimony from economic expert witnesses on behalf of Copyright Owners and the Services. See Initial Ruling, passim. However, none of the post-remand evidence submitted and relied upon by the parties specifically addressed as a separate issue the rates for the nine offerings that are the subject of the present Order. On July 1, 2022, the Judges issued their Initial Ruling 185—applying application of the 26.2% TCC rate, except for the use of that 26.2% rate as an input derived from a specific dataset, to set the 15.1% service revenuebased royalty rate. Johnson, supra, at 385–86; see also at 386 n.11. 185 The findings and conclusions in the Initial Ruling were adopted by a majority of the Judges, but two Judges filed separate opinions. See Initial Ruling at 2 n.5. One Judge, former Chief Judge Suzanne Barnett, dissented from the Majority’s adoption in the Initial Ruling regarding the Phonorecords II rate structure (section II of the Initial Ruling), though not from the exception to that benchmark with regard to the headline rate of 15.1% and the imposition of a cap on the TCC rate PO 00000 Frm 00056 Fmt 4701 Sfmt 4700 2020 2021 13.3% 24.1% 14.2% 25.2% 2022 15.1% 26.2% Johnson and considering the entire record developed pre-remand and postremand. In their Initial Ruling, the Judges made several findings that bear upon the issue at hand, viz., whether to adopt in the Phonorecords III regulations the 26.2% TCC rate or the TCC rates (ranging from 20.65% to 22%) from the Phonorecords II-based benchmark. In particular, in the Initial Ruling, the Judges stated the following: 1. The Phonorecords II-based benchmark incorporates price discriminatory features for product differentiation as between: (a) subscription and ad-supported services; (b) portable and non-portable services; and (c) unbundled and bundled services. See Initial Ruling at 67–68 (noting the salutary price discriminatory nature of the Phonorecords IIbased benchmark). 2. The Phonorecords II-based benchmark ‘‘reflect[s] a rate structure with an adequate degree of competition, because there was a balance of bargaining power [‘‘countervailing power’’] between the two negotiating industrywide trade associations, offsetting the complementary oligopoly effects in place when a ‘‘Must Have’’ licensor bargains separately with each licensee.’’ Initial Ruling at 69. 3. Based upon the available record evidence, the Judges find . . . the Services’ Phonorecords II-based benchmark . . . ‘‘more than sufficient to satisfy the legal requisites for application, as well as a practical benchmark, when used in conjunction with the 15.1% headline revenue rate advocated by Copyright Owners.’’ Initial Ruling at 59. 4. ‘‘Substantial evidence demonstrates that the Phonorecords II-based benchmark rates, other than the headline rate, are not ‘too low.’ ’’ Initial Ruling at 73. 5. A Copyright Owner expert witness opined that ‘‘the evidence . . . indicates that the relative valuation ratios implied by the current Section 115 compulsory license [i.e., the Phonorecords II-based benchmark] implies a ‘‘lower bound on the relative market valuations of the reciprocal percentage of the value musical works rights relative to sound recording rights [i.e., TCC rates] [of] 22% and 21%.’’ Initial Ruling at 78 (emphasis therein). 6. The royalty rates and terms within subpart C of the Phonorecords II-based benchmark—which include the rates and term for the offerings at issue in this Order— prong. See Chief Judge Barnett’s ‘‘Dissent re Benchmark’’ (July 1, 2022) (eCRB no. 26943). The other opinion was issued by Judge Strickler, who dissented from the reasoning relating to the adoption of the definition of Service Revenue (section V), but concurred in the adoption of that definition. See Judge Strickler’s ‘‘Dissent in Part as to Section IV of the Initial Ruling and Order after Remand’’ (July 1, 2022) (eCRB no. 26965). E:\FR\FM\10AUR2.SGM 10AUR2 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations are expressly ‘‘covered by [the] foregoing analysis.’’ Initial Ruling at 93. In rejecting all of Copyright Owners’ arguments for different treatment of Phonorecords II-based benchmark rates in Subpart C therein, the Judges declined to adopt Copyright Owners’ ‘‘re-assert[ion] [of] the same arguments with respect to subpart C’’ that Copyright Owners advanced in opposing the Phonorecords IIbased benchmark ‘‘for interactive streaming in subpart B.’’ See Initial Ruling at 93–94 (‘‘The Judges find no reason on remand to treat the subpart C offerings differently than the manner in which they are treating the subpart B interactive streaming offerings . . . . That means, however, that the various ‘‘headline’’ rates for these subpart C offerings must also adjust to 15.1%, 131 whereas the alternative rates (identified in subpart C as ‘‘minima’’ and ‘‘subminima’’) rates shall remain unchanged.’’) (emphasis added). 7. The D.C. Circuit had affirmed that: (a) the ‘‘headline’’ percentage royalty rate (not a TCC rate) of 10.5% was too low; and (b) that the Majority had not improperly exercised its authority when it increased that revenue royalty rate to 15.1% (as phased-in over the five-year rate term). Accordingly, on remand, the Judges maintained the 15.1% (phased-in) percentage royalty rate. See, e.g., Initial Ruling at 4, 17. 8. The D.C. Circuit affirmed the Majority’s derivation and calculation of the 26.2% TCC rate for use as an input in calculating the 15.1% (phased-in) service revenue percentage royalty rate. However, Johnson vacated and remanded the Majority’s application and inclusion of the 26.2% TCC rate. Initial Ruling at 19–20. ddrumheller on DSK120RN23PROD with RULES2 For these reasons, the Judges decided in the Interim Ruling that: (1) the overall Phonorecords II rates comprise a ‘‘useful benchmark,’’ when the 15.1% headline percentage rate replaces the 10.5% headline percentage rate for the offerings in Subparts B and C of the Phonorecords II-based benchmark; and (2) ‘‘[t]he (phased-in) 26.2% rate [is] unreasonable.’’ Initial Ruling at 50 n.77; 88; and 93–94. Procedures Following the Post-Remand Initial Ruling In the Initial Ruling, the Judges directed the parties to attempt to submit jointly agreed-upon regulatory provisions implementing the Initial Ruling, for the Judges to consider. The Judges further ruled that, if the parties could not agree on all the regulatory language, they should make separate submissions regarding regulatory provisions in dispute. See Initial Ruling at 114. The parties agreed to many regulatory provisions but disagreed as to several such provisions. Accordingly, they filed separate submissions and respective replies, regarding the regulatory provisions. Services’ Joint Submission of Regulatory Provisions (July 18, 2022); Copyright Owners’ Submission of VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 Regulatory Provisions to Implement the Initial Ruling (July 18, 2022); Services’ Joint Response to Copyright Owners’ Submission of Regulatory Provisions (Aug. 5, 2022); Copyright Owners’ Response to Judges’ July 27, 2022 Order Soliciting Responses Regarding Regulatory Provisions (Aug. 5, 2022). The Judges considered those submissions and entered an order addressing the disputed regulatory provisions. See Corrected Order regarding Regulatory Provisions following Initial Ruling and Order (After Remand) (Nov. 10, 2022) (‘‘November 10th Order’’).186 In the November 10th Order, the Judges directed the parties once more to file a joint submission ‘‘of regulatory provisions that embody the rulings set forth in Johnson, the Initial Ruling and this [November 10th] Order, and any aspects of the [Majority] Determination (pre-remand) that the parties understand to remain effective after the foregoing rulings.’’ November 10th Order at 31. On November 30, 2022, the parties made the Joint Submission (as also identified at the outset of the present Order), in which they provided joint regulatory language no longer in dispute that applied the binding rulings of the Judges and the D.C. Circuit. However, as also noted above, the parties identified the single issue in dispute that relates to the nine service offerings described supra.187 The Parties’ Respective Arguments in Their November 30th Joint Submission Copyright Owners’ Arguments According to Copyright Owners, the Initial Ruling ‘‘appears to plainly acknowledge that, in light of Johnson, the derivation and calculation of the (phased-in) 26.2% TCC rate percentage cannot be changed.’’ Joint Submission at 6. More particularly, Copyright Owners aver that, according to the Judges’ Initial Ruling, ‘‘the D.C. Circuit affirmed the Majority’s derivation and calculation of 186 The November 10th Order corrected an otherwise substantively identical order issued two days earlier, on November 8, 2023, which had inadvertently included a small amount of text. See November 10th Order at 1. 187 On January 10, 2023, Spotify USA Inc., Amazon.com Services LLC, Google LLC, Pandora Media, LLC, National Music Publishers’ Association, Inc. and the Nashville Songwriters Association International filed a joint Motion (eCRB no. 27418) requesting modification of the previously proposed language for 37 CFR 385.3, which governs fees owed for late payment. There was no opposition to the January 10, 2023 joint Motion. The Judges find good cause to adopt the modified language, which provides that ‘‘where payment is due to the mechanical licensing collective under 17 U.S.C. 115(d)(4)(A)(i), late fees shall accrue from the due date until the mechanical licensing collective receives payment.’’ PO 00000 Frm 00057 Fmt 4701 Sfmt 4700 54461 the 26.[2]% . . . TCC rate’’ and further that ‘‘both rate prongs’’—the service revenue rate and the TCC rate—were ‘‘derived from the same analyses.’’ Initial Ruling at 19; Joint Submission at 6–7 (quoting Initial Ruling at 19 (emphasis removed)). Further to this point, Copyright Owners rely on the Judges’ additional language in the Initial Ruling that the pre-remand Final Determination’s ‘‘derivation and calculation of the TCC rate [i.e., the 26.2% rate] . . . is not subject to further consideration on remand by the Judges.’’ Joint Submission at 7 (quoting Initial Ruling at 20 (emphasis in Initial Ruling)).188 According to Copyright Owners, the foregoing points are consistent with the limited scope of the remand, which ‘‘was not opened for new evidence concerning TCC rate percentages.’’ Joint Submission at 7 (citations omitted). Accordingly, Copyright Owners emphasize that ‘‘there is no evidence in the record after remand to support changing the (phased-in) 26.2% TCC rate percentage.’’ Joint Submission at 7. Copyright Owners—characterizing the former Phonorecords II TCC rates now at issue as newly derived and calculated—maintain that these ‘‘new’’ TCC rate percentages therefore are ‘‘foreclosed’’ by the Initial Ruling and post-remand orders cited above. Joint Submission at 7–8. Copyright Owners also assert that the TCC rate at issue here—‘‘was not appealed by the Services or challenged during the remand, nor called into question by the Circuit in Johnson.’’ Joint Submission at 8 (emphasis removed). The absence of an appeal as to this issue, according to Copyright Owners, means that the only TCC rate supported by Johnson is the 26.2% TCC rate. Joint Submission at 8. The Services’ Arguments According to the Services, the Judges should adopt in the regulations the TCC percentage rates—ranging from 20.65% to 22%—because those rates are contained in the Phonorecords II-based benchmark adopted by the Judges and thus essentially have been ‘‘expressly set out by the Judges’’ in two prior decisions. Joint Submission at 2 (citing Initial Ruling at 2; November 10th Order at 6 n.13). In light of these prior Orders, the Services characterize Copyright Owners’ position as the new argument, improperly seeking regulatory provisions that ‘‘reflect the 26.2% rate 188 However, Copyright Owners disregard the Initial Ruling’s observation that Johnson vacated and remanded the Majority’s application and inclusion of the 26.2% TCC rate. Initial Ruling at 19. E:\FR\FM\10AUR2.SGM 10AUR2 54462 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations ddrumheller on DSK120RN23PROD with RULES2 previously imposed by the [M]ajority in the now-vacated pre-remand Final Determination.’’ Id. More pointedly, the Services argue that the Judges’ Initial Ruling already expressly considered and rejected application of the 26.2% TCC rate. Id. (citations omitted). Further, the Services maintain that it is because the Judges rejected the 26.2% TCC rate in the Initial Ruling that the Judges had no need to ‘‘substantively address the topic of TCC rates’’ in their November 10th Order. Id. at 4. The Services further maintain that ‘‘Johnson does not compel the Judges to simply reinstate their original preremand TCC rates.’’ Id. To this point, the Services rely on the Judges’ postremand finding that, although the error made by the Majority in adopting the 26.2% TCC rate in the pre-appeal Phonorecords III Determination was procedural, the ‘‘consequence . . . was substantive.’’ Id. (emphasis herein). For the above reasons, the Services maintain that the Judges could not possibly be required on remand to adopt an express 26.2% in any portion of the Phonorecords III regulations. Turning from their argument that the 26.2% TCC rate was rejected by the Judges, the Services focus on the Judges’ finding in the post-remand Initial Ruling that the ‘‘Phonorecords II benchmark . . . is the ‘better of the benchmarks proposed by the parties . . . one that satisfies the requirements of 17 U.S.C. 801(b)(1) in all respects,’ ’’ Joint Submission at 5 (quoting Initial Ruling at 2). Because the Phonorecords II benchmark includes the TCC rates now at issue—ranging from 20.65% to 22%—the Services maintain that those rates should properly be included in the Phonorecords III regulations. Id.189 189 The Services also argue that Copyright Owners’ assertion at this time that the 26.2% TCC rate should substitute for the Phonorecords II-based benchmark rates is procedurally untimely and improper. The Judges only partially agree with Services’ argument in this regard. If Copyright Owners had wanted to timely make this argument, they should have done so during the post-remand period before the Judges entered their Initial Ruling (or, of course, during the initial proceeding preappeal). In that sense, Copyright Owners failed to avail themselves procedurally of the right to make this substantive challenge. However, the Judges have afforded the parties the procedural right to propose regulatory language that they claim would implement the Initial Ruling; a procedural right exercised by both parties, as evidenced by, for example, their arguments in the Joint Submission. In that narrow sense, Copyright Owners’ present argument is not procedurally improper. As a matter of substance though, as explained in ‘‘The Judges Analysis and Ruling’’ infra, the Judges have considered herein Copyright Owners’ present arguments and found them inconsistent with the Initial Ruling. Finally, with regard to subsequent substantive challenges to the Initial Ruling, the parties correctly VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 The Judges’ Analysis and Ruling Having considered the parties’ submissions, the Initial Ruling and all other pertinent material, the Judges rule that the 26.2% TCC rate cannot and shall not be applied in the regulatory provisions now at issue. Rather, the Judges rule that the TCC rates set forth in the Phonorecords II-based benchmark shall be applied in the nine regulatory provisions now at issue, because they are consistent with and give effect to the Judges’ Initial Ruling. The more particular bases for this ruling are set forth below. Most fundamentally, the Judges note at the outset that in the Initial Ruling they expressly did not apply the 26.2% TCC rate in any manner other than as an input—using that TCC rate only as the D.C. Circuit directed—to calculate the 15.1% of service-revenue royalty rate. See, e.g., Initial Ruling at 41 (‘‘[A] careful reading of the remand testimony by Copyright Owners’ economists, Professors Watt and Spulber, reveals that neither of them actually testifies that there is sufficient theoretical and empirical evidence to support the . . . 26.2% TCC rate . . . .’’) (emphasis in original). See also id. at 40–41 n.69 (contrasting the improper application of the 26.2% TCC as a separate statutory rate from the use of the 26.2% TCC rate as input from a ‘‘bargaining model’’ solely to increase the service revenue rate to 15.1%.).190 In this regard, the Initial Ruling has relied upon the clear distinction made in Johnson between the 15.1% service revenue rate and the 26.2% TCC rate. Compare Johnson, supra, at 385 (affirming the Majority’s application of the ‘‘revenue rate of 15.1%’’ as ‘‘the type of line-drawing and reasoned weighing of the evidence falls squarely within understand that such challenges can be made after the Judges issue their post-remand ‘‘Initial Determination’’ (a statutorily-mandated ruling). See Joint Submission at 9 (Services agreeing with Copyright Owners’ understanding that they continue to properly ‘‘reserve all rights with respect to the Initial Ruling, any implementing regulations and any Initial and Final Determination, including the right to challenge any of the foregoing.’’). 190 The Services claim that this distinction constitutes a semantic twisting of words. See Joint Submission at 7. The Judges reject that characterization. Rather, their ruling is substantive, not semantic, because they have relied upon the testimony of several economic expert witnesses, including one of Copyright Owners’ own economic experts, who identified five reasons that the Judges found to preclude adoption of the 26.2% TCC rate as a separate statutory rate. See, e.g., Initial Ruling at 41. Moreover, not a single economist who testified at the hearing proposed that the Judges adopt the 26.2% TCC rate as a statutory rate, see Initial Ruling at 38, further supporting the Judges’ adoption in the Initial Ruling of the consensual negotiated TCC rates contained in the Phonorecords II-based benchmark for the nine offerings at issue. PO 00000 Frm 00058 Fmt 4701 Sfmt 4700 the[ir] wheelhouse as an expert administrative agency’’) with id. at 382– 83 (vacating the Majority’s decision for ‘‘significantly hiking the TCC rate to 26.2% from approximately 17% to 22%’’ without allowing the Services an opportunity to address the issue—an error that was even ‘‘worse’’ than the elimination of caps on certain other TCC offerings.). Further, the offerings now at issue were contained in the Phonorecords IIbased benchmark, and the Judges’ application of that benchmark in the Initial Ruling is unambiguous: Other than the new and increased headline rate of 15.1%, ‘‘the rates and rate structure of the Phonorecords II-based benchmark proposed by the Services . . .) shall constitute the rates and rate structure for the Phonorecords III period.’’ Initial Ruling at 2. Accordingly, with regard to the single remaining issue, pertaining to the nine offerings listed supra, the regulatory provisions proposed by the Services in the Joint Submission are fully consistent with the Initial Ruling. By contrast, Copyright Owners’ proposed language introduces a change in the Phonorecords II-based benchmark rates that was never the subject of an evidentiary proceeding pre-or postremand, whether through live or written testimony. But perhaps more importantly, as a matter of substance, Copyright Owners’ proposed regulatory provisions are inconsistent with the language and a key purpose of the Initial Ruling, which is to adopt the Phonorecords II-based benchmark rates, the basis of which were generated consensually by the parties, through negotiations between industrywide trade associations, which prevented unwarranted and disproportionate complementary oligopoly market power from affecting the royalty rates. See Initial Ruling at 69–70.191 191 The Judges also note that their adoption of these 20.65% through 22% TCC rates in the Phonorecords II-based benchmark—because they are lower than the 26.2% rate proposed by Copyright Owners—is consistent with their rationale for adopting that benchmark. As the Judges explained repeatedly and throughout the Initial Ruling, their adoption of the Phonorecords IIbased benchmark purposefully incorporates into the Phonorecords III regulations the beneficial price discriminatory features that are hallmarks of that benchmark. See, e.g., Initial Ruling at 65 n.98 (‘‘[T]the granular discriminatory features that the parties had negotiated . . . reflect an ‘‘appropriate form and extent of price discrimination . . . .’’ The Judges emphasized this point repeatedly. See generally Initial Ruling, passim. Further, as the Services note, Copyright Owners themselves—even when advocating for an otherwise across-the-board 26.2% TCC prong—had continued to propose the 20.65% to 22% TCC rates for the nine offerings at issue now. See Copyright Owners’ Submission of Regulatory Provisions to E:\FR\FM\10AUR2.SGM 10AUR2 ddrumheller on DSK120RN23PROD with RULES2 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations The Judges also reject Copyright Owners’ argument that by maintaining the 20.65% through 22% TCC rates in the Phonorecords II-based benchmark they would be violating their prior rulings regarding the scope of the remand. Citing to the Judges’ Order Regarding Proceedings on Remand at 1 (eCRB no. 23390) (‘‘Remand Order’’), Copyright Owners state in their Joint Submission that that the remand ‘‘was not opened for new evidence concerning TCC rate percentages.’’ Joint Submission at 7. But the decision to reopen the existing, and robust, evidentiary record only as to rate structure, did not limit the scope of the remand itself, nor consideration of evidence from the underlying proceeding. Moreover, the Judges find no language in either the Remand Order or the Remand Scheduling Order, and no other basis, that would support Copyright Owners’ characterization of the 20.65% through 22% TCC rates in the Phonorecords III-based benchmark as new evidence, given that they were expressly included in that benchmark which had been proffered at the hearing prior to the remand. Further, the present issue of whether the regulatory provisions implementing the Initial Ruling should apply the Phonorecords II-based benchmark TCC rates or the 26.2% TCC rate is not a dispute regarding the derivation or calculation of a new TCC rate. The Phonorecords II-based benchmark rates are self-evidently not new rates, because they existed in that prior benchmark. Moreover, the present dispute relates to whether the language and reasoning in the Initial Ruling are consistent with maintaining the rates contained in the Phonorecords II-based benchmark for the nine offerings at issue, or whether the Initial Ruling calls for abandoning those benchmark rates and replacing them with the 26.2% TCC rate proffered by Copyright Owners. As explained supra, the 26.2% TCC rate was properly utilized by the Majority as an input (combined with other evidence) in order to calculate the 15.1% service revenue royalty rate. The record reflects no other context in which the 26.2% TCC rate can be utilized, let alone must be utilized. Indeed, as explained supra, the record reflects the Judges’ rejection of the 26.2% TCC rate as a stand-alone statutory royalty rate. The Judges also reject Copyright Owners’ argument that the Services somehow waived their argument for maintaining the 20.65% through 22% Implement the Initial Ruling at 15–16) (July 18, 2022); see also Joint Submission at 6. VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 TCC Phonorecords II-based benchmark rates. More particularly, Copyright Owners incorrectly assert that these rates were ‘‘not appealed by the Services. . . .’’ Joint Submission at 8. Rather, the D.C. Circuit stated unambiguously: ‘‘[T]he Streaming Services object to the [Judges’] . . . rejection of the Phonorecords II . . . settlement[ ] as [a] rate benchmark[ ].’’ Johnson, 969 F.3d at 384; see also id. at 386 (‘‘The Streaming Services argue . . . that the [Judges] arbitrarily rejected . . . [a] potential rate benchmark[ ] . . . the Phonorecords II settlement—without adequate explanation.’’). Moreover, the D.C. Circuit repeatedly noted that it was vacating and remanding the Majority’s Determination with regard to, inter alia, the Majority’s improper decision to reject the Phonorecords II-based benchmark writ large, i.e., without qualification by the appellate panel that some parts of that proffered benchmark might have been correctly rejected. See Johnson, 969 F.3d at 367, 376, 381, 387. Obviously, virtually all the elements of the Phonorecords II-based benchmark— including the offerings now at issue— were appealed, and not waived, foregone or forfeited by the Services. Likewise, Copyright Owners are wrong in their claim that the Services had never ‘‘challenged’’ these rate issues ‘‘during the remand.’’ Joint Submission at 8. Rather, the Services argued on remand for the Phonorecords II-based benchmark to be applied comprehensively, without itemizing every element of that proffered benchmark. See Services’ Joint Opening Brief (post-remand) at 19–44 (Apr. 1, 2021) (detailing why ‘‘the Services’ proposal based on the Phonorecords II settlement is reasonable . . . .’’); see also Services’ . . . Submission of Regulatory Provisions at 2 (July 18, 2022) (‘‘Services’ July 18th Submission’’) (‘‘[T]he Services have faithfully implemented the task at hand—to use the rates and rate structure of the ‘‘Phonorecords II-based benchmark’’ proposed by the Services during the remand proceeding . . . .’’).192 Finally, the Judges find and conclude that their ruling in this Order sets forth 192 The decision in Johnson could be construed as rejecting one element of the Phonorecords II-based benchmark, viz., the 10.5% headline rate, because the appellate panel affirmed the higher Majority’s adoption of the (phased-in) 15.1% headline royalty revenue rate. The Initial Ruling is consistent with that ruling, and this rate is not now in dispute. See Services’ July 18th Submission at 2 (the Services acknowledge that in their proposed regulatory provisions they ‘‘replac[ed] the headline rate’’ of 10.5% with the headline royalty rate ‘‘set by the Judges [15.1%] in the Initial Ruling.’’). PO 00000 Frm 00059 Fmt 4701 Sfmt 4700 54463 reasonable rates satisfying the four objectives in the then-applicable (but now superseded) statutory rate standard contained in 17 U.S.C. 801(b)(1).193 First, with regard to Factor (A),194 the Judges recognize and follow the D.C. Circuit’s ruling that the Majority’s decision to increase in the ‘‘headline’’ service revenue royalty rate by 44% from 10.5% to 15.1% was supported by substantial evidence. Johnson at 387–88. Further with regard to Factor (A), the Judges understand their analysis and reasoning in the Initial Ruling— applying the Phonorecords II-based benchmark and thus rejecting the 26.2% TCC rate—to be applicable to the present dispute regarding the adoption of regulations to implement the Initial Ruling. Accordingly, the Judges adopt by reference herein their analysis and reasoning set forth at pages 90–91 of the Initial Ruling. For those reasons, the Judges decide, as they did in the Initial Ruling, that there is no basis for yet a further increase in the royalty rate based on Factor (A), finding ‘‘no evidence to suggest that the price discriminatory rates should be changed, in order to address the connection between price discrimination and the objective of Factor (A).’’ Id. at 91. Next, in considering Factors (B) and (C),195 the Judges’ Initial Ruling adopts the Majority’s reasoning that the 15.1% service revenue royalty rate provided a ‘‘fair allocation of revenue between copyright owners and services’’ and it would be ‘‘substantively unwarranted to engage in any new consideration on remand of the impact, if any, of Factors 193 The D.C. Circuit expressly declined to adopt most of the Majority’s application of the explicit statutory objectives. As to Factor (A), regarding the objective of ‘‘maximiz[ing] the availability of creative works to the public,’’ the D.C. Circuit held that the Majority’s finding that ‘‘an increase in the royalty rates for mechanical licenses was necessary to ensure the continued viability of songwriting as a profession’’ was ‘‘supported by substantial evidence.’’ Johnson at 387–388. However, with regard to the remaining statutory factors, Johnson instead vacated and remanded consideration of those matters to the Judges. See Johnson at 389. The Initial Ruling after remand considered these statutory objectives in detail. See Initial Ruling at 90–93. (The parties made no express argument regarding the application of these statutory objectives in their Joint Submission.). 194 Factor (A) provides that rates shall be calculated to achieve the objective of ‘‘maximize[ing] the availability of creative works to the public.’’ 17 U.S.C. 801(b)(1)(A). 195 The Factor (B) objectives (providing a ‘‘fair return’’ and a ‘‘fair income’’ to the licensors and licensees respectively) and Factor (C) objectives reflecting their relative roles in making the streamed music available to the public) are typically considered jointly, because of their overlapping concerns. See Initial Ruling at 15 n.31 (citing Johnson, 969 at 388). In this Order, the Judges likewise jointly address Factors (B) and (C). E:\FR\FM\10AUR2.SGM 10AUR2 54464 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations ddrumheller on DSK120RN23PROD with RULES2 (B) and (C) on the otherwise reasonable 15.1% revenue rate.’’ Id. at 15–16. In their Joint Submission, the parties have presented no arguments specifically addressing how Factors (B) or (C) might support their proposed TCC rates now at issue. Examining the record, the Judges find and conclude that maintaining the Phonorecords IIbased rates ranging from 20.65% to 22% embodies the fairness associated with rates negotiated between industrywide trade associations wielding relatively comparable bargaining power, as discussed supra and in the Initial Ruling.196 This notion of fairness is embodied in the determination of the reasonable rate and, as can be the case, when one of the four itemized statutory objectives of section 801(b)(1) is boundup and appropriately addressed within the broader context of setting a reasonable rate, no further adjustment is necessary through an invocation of an itemized statutory factor. See Determination of Royalty Rates and Terms for Making and Distributing Phonorecords (Phonorecords III) 84 FR 1918, 1955, 2015 (Feb. 5, 2019) (Majority and Dissenting Opinions agreeing that ‘‘to the extent market factors may implicitly address any (or all) of the four itemized factors, the reasonable, market-based rates may remain unadjusted.’’). Finally, the Judges see no reason to alter their adoption of the Phonorecords II-based benchmark rates for the nine offerings at issue in this Order based upon the final listed statutory objective, Factor (D).197 In the Joint Submission, Copyright Owners did not make an express argument relating to this factor (nor did the Services). Independently considering the potential application of Factor (D), the Judges find no evidence that the continuation of the Phonorecords II-based benchmark rates for the offerings at issue in this Order would cause any disruption that Factor (D) is intended to address. Further, as noted supra, the Judges have phased-in an increase in the headline service revenue royalty rate from 10.5% to 196 In this regard, the Judges agree with the Services’ argument. See Initial Ruling at 61 (summarizing the Services’ position as to Factors (B) and (C)). 197 ‘‘Factor (D) . . . instructs the Judges to consider the ‘competing priority’ of ‘minimiz[ing] any disruptive impact on the structure of the industries involved and on generally prevailing industry practices.’’’ Initial Ruling at 16. More particularly, ‘‘disruption’’ potentially remediable under Factor (D) requires that the contemplated rate ‘‘directly produce[ ] an adverse impact that is substantial, immediate and in the short-run because there is insufficient time for either [party] to adequately adapt to the changed circumstance produced by the rate change . . . .’’ Initial Ruling at 53–54. VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 15.1%—a 44% increase—rendering unreasonable any argument that the present decision to maintain the Phonorecords II-based TCC rates is ‘‘disruptive’’ to Copyright Owners under the statutory Factor (D) standard. Moreover, the Judges reassert their point in the Initial Ruling that there is no need to independently consider any potential disruption under the Factor (D) standard because the Judges have already found an application of that rate to be unreasonable. See Initial Ruling at 50 n.77. Further, the D.C. Circuit was aware of the existence of the 20.65% to 22% TCC rates in the Phonorecords IIbased benchmark for these nine offerings now at issue, and not only declined to affirm the Majority’s increase in those rates to 26.2%—a significant increase of 19% to 27% 198— but also condemned that increase. See Johnson at 383 (‘‘Worse still . . .the [Judges] also raised the total content cost [TCC] rate to 26.2%. . . .That rate previously fell between approximately 17% and 22%’’). Nothing in the record suggest that the Judges can or should utilize the narrow statutory ‘‘disruption’’ standard in Factor (D) of section 801(b)(1) as a basis to override the position of the D.C. Circuit or the Judges’ analysis in the Initial Ruling as to the inapplicability of the proffered 26.2% royalty rate. Order For the foregoing reasons, the Judges shall adopt in the regulatory provisions 199 the several ‘‘Total Content Cost’’ (‘‘TCC’’) rates set forth in the Phonorecords II-based benchmark as proposed by the Services.200 Within two days of the date of issuance of this Restricted Order, the parties shall file an agreed proposed redacted version for public viewing. Issue Date: April 26, 2023. David P. Shaw Chief Copyright Royalty Judge 198 An increase from 20.65% to 26.2% is a 5.55 percentage point increase, which is an increase of 27% (rounded). An increase from 22% to 26.2% is a 4.2 percentage point increase, which is an increase of 19% (rounded). 199 As addressed herein, the Judges find good cause to adopt the joint proposal for modified language regarding late fees, in 37 CFR 385.3. 200 The Initial Determination shall issue forthwith. PO 00000 Frm 00060 Fmt 4701 Sfmt 4700 C. Dissent in Part as to Section IV of the Initial Ruling and Order After Remand by Judge David R. Strickler 201 (Redacted Version With Federal Register Naming and Formatting Conventions) I. The Contours of This Partial Dissent I respectfully Dissent from Section IV of the Initial Ruling and Order after Remand (Initial Ruling). As explained herein, I conclude that the D.C. Circuit’s rulings in Johnson preclude the Judges from engaging in ‘‘new ‘agency action.’ ’’ 202 See Johnson v. Copyright Royalty Board, 969 F.3d 363, 386 (D.C. Cir. 2020). Accordingly, I cannot join with the present Majority in its determination that this remand proceeding constitutes ‘‘new ‘agency action’ ’’ consistent with Johnson. That argument is circular and renders useless the D.C. Circuit’s careful analysis of the procedures that are and are not available to the Judges after they have issued their Initial Determination. As further explained herein, the argument is circular because it begins with the D.C. Circuit’s ruling that the Determination 203 was improper because it invented a new procedure to change 201 I am concurring in the Majority’s substantive re-adoption of the Bundled Service Revenue definition from the Initial Determination. As explained herein, I disagree with the Majority regarding the procedural manner in which the Judges may reach this result. Thus, it would be more accurate to describe this ‘‘Dissent’’ as a ‘‘Concurring Opinion’’, or an ‘‘Opinion Concurring in Part and Dissenting in Part.’’ However, the Copyright Act does not expressly authorize Judges to issue a ‘‘concurring opinion,’’ but rather references the issuance of a ‘‘dissenting opinion.’’ See 17 U.S.C. 803(a)(3). Accordingly, I identify this opinion as a ‘‘Dissent in Part as to Section IV of the Initial Ruling and Order after Remand.’’ 202 I place the phrase agency action within quotation marks inside the broader phrase new agency action to avoid potential ambiguity and inconsistency with the directives in Johnson. There, the D.C. Circuit held that the Judges cannot assert ‘‘plenary authority to revise [their] determinations whenever [they] thought appropriate,’’ because such a power grab would render ‘‘a nullity . . . the lines drawn by the authorizing statute . . . to confine . . . post hoc amendments’’ to statutorily identified circumstances.’’ Johnson at 392. So, ‘‘new’’ means the new application of an existing statutorily available ‘‘agency action’’ that had not previously been invoked—not ‘‘new’’ in the sense of a form of action conjured up to meet the moment. (When this phrase is used in a quotation I do not use the double quotation marks.) This distinction is important because the Majority and Copyright Owners advance new forms of (extra-statutory) agency action, not merely new applications of statutorily-authorized agency actions. 203 Determination of Royalty Rates and Terms for Making and Distributing Phonorecords (Phonorecords III), 84 FR 1918 (Copyright Royalty Board Feb. 5, 2019) (final rule and order) (‘‘Determination’’); See also Final Determination, 16–CRB–0003–PR (2018–2022) (Nov. 5, 2018) (citations to the Determination and to the Dissent in this Dissent in Part are found in this document). The Dissent is appended to and part of the same document as the Determination. E:\FR\FM\10AUR2.SGM 10AUR2 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations the Bundled Revenue definition that was in the Initial Determination,204 only to circle back to where it started by creating—through the D.C. Circuit’s own remand no less—a further and extrastatutory ‘‘new ‘agency action’’’. The Majority also renders Johnson useless, by adopting a process by which—after the D.C. Circuit has remanded an issue because the Judges lacked procedural authority to rule—the procedural error is essentially honored in the breach, because the remand neuters the effect of the D.C. Circuit’s ruling.205 I join with the Majority though on its substantive decision to re-adopt the definition of Bundled Revenue set forth in the Initial Determination. As explained infra, I too find that it is clearly preferable to the definition that was swapped into the (Final) Determination. But as explained herein, I reconcile the procedural and substantive points differently. I apply what I believe to be the proper understanding of the D.C. Circuit’s ruling—finding, contrary to the Majority, no avenue for ‘‘new ‘agency action’’’ post-remand. Rather, the Judges must revert to the original—and substantively appropriate—definition of Bundled Revenue in the Initial Determination. To explicate the bases of this Dissent, my opinion as to this issue is set forth below. ddrumheller on DSK120RN23PROD with RULES2 II. Introduction The Majority and I analyze the definition of ‘‘Service Revenue’’ from ‘‘Bundled Offerings’’ (henceforth ‘‘Bundled Revenue’’ definition) in the context of our partial adoption of the PR II-based benchmark. As discussed supra, the Remand Majority found that the PR II-based benchmark is a useful benchmark, particularly because of its features that incentivize beneficial downstream price discrimination and generate more listeners, revenues, and royalties. As explained below, the Bundled Revenue definition—itself an element within the PR II-based benchmark—also embodies such price discriminatory incentives. Thus, the Judges’ analysis of the PR II-based benchmark and the Bundled Revenue definition are connected. 204 Initial Determination, 16–CRB–0003–PR (2018–2022) (Jan. 27, 2018). 205 The Initial Ruling suggests that the Judges could have utilized a ‘‘further explanation’’ for the switched Bundled Revenue definition, as opposed to using ‘‘new ‘agency action.’’’ I do not dissent from that general point. However, even though the Majority did not utilize this alternative approach on remand, I dissent to the extent that section could be read to allow a fuller explanation that would conflict with Johnson. VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 In the Determination, the earlier Majority likewise found the issues relating to the PR II-based benchmark to be bound-up with the question of the appropriate Bundled Revenue definition. But because that earlier Majority rejected the PR II-based benchmark, it likewise rejected the Bundled Revenue definition contained in the Initial Determination. The definition in the Determination thus eliminated the royalty-based incentive to engage in price discrimination via bundling. In the interregnum between the Initial Determination and the (Final) Determination, the Judges considered Copyright Owners’ post-hearing motion which sought, inter alia, to strike the Bundled Revenue definition in the Initial Determination. The Majority agreed with Copyright Owners that the definition in the Initial Determination should be replaced. An important rationale—highly relevant in the present context—was as follows: ‘‘The Judges have . . . declined to rely on the 2012 . . . benchmark . . . as the basis for the rate structure, or, therefore, as regulatory guidance.’’ Amended Order Granting in Part and Denying in Part Motions for Rehearing at 17 (Jan. 4, 2019) (Clarification Order).206 Unlike in the Determination, in this Initial Remand Ruling the Judges do rely on the PR II-based benchmark in part because of its price discriminatory aspects. More particularly, because the bundling of interactive services also constitutes a form of price discrimination, the Judges find the PR II-based benchmark definition of Bundled Revenue set forth in the Initial Determination to be substantively reasonable and otherwise consistent with the four itemized factors in section 801(b)(1). As a procedural matter though, I can neither: (1) offer any further or fuller explanation for why the Majority made this change in the Bundled Revenue definition nor (2) identify any ‘‘new 206 This January 4, 2019 Order was issued in response to two motions; the Services’ ‘‘Joint Motion for Rehearing to Clarify the Regulations’’ and Copyright Owners’ ‘‘Motion for Clarification or Correction of Typographical Errors and Certain Regulatory Terms.’’ As explained infra, Copyright Owners did not style their motion as a ‘‘rehearing’’ motion and expressly declined to argue that their motion met the statutory and regulatory requisites for rehearing. This remand issue pertains only to the post-hearing switch in the Bundled Revenue definition sought and obtained by Copyright Owners via their motion. Accordingly, it is clearer to refer herein to the Judges’ January 4, 2019 Order as the ‘‘Clarification Order,’’ rather than as a ‘‘Rehearing Order,’’ because the semantic distinction carries substantive overtones. (I had dissented from the Initial Determination and the Determination, and thus did not join in the Clarification Order.) PO 00000 Frm 00061 Fmt 4701 Sfmt 4700 54465 ‘agency action’’’ that would permit this definitional switch. And contrary to present Majority on remand, I also cannot identify a ‘‘new ‘agency action’’’ that the Judges can now take to return to the definition in the Initial Determination. But, as explained infra, the Judges need not identify such action, because the absence of a justification for the definitional switch requires the Judges to revert back to the definition in the Initial Determination. As a substantive matter though, the Judges unanimously agree to replace the post-hearing definition of Bundled Revenue in the Determination and reinstate the definition set forth in the Initial Determination. III. Background In this remand proceeding, the parties propose two starkly different definitions of Bundled Revenue. Each has a dramatically different impact on the use of the royalty structure and levels to incentivize price discrimination in the downstream market. The Services argue in favor of the language contained in the Initial Determination, i.e., in their PR II-based benchmark, which defines Bundled Revenue, in pertinent part, as the revenue recognized from End Users [i.e., consumers] for the Bundle less the standalone published price for End users for each of the other component(s) of the Bundle . . . . Initial Determination, Attachment A at 7 (§ 382.2 therein). By contrast, Copyright Owners support the Majority’s substituted language contained in the Determination, which defines Bundled Revenue, in pertinent part, as the lesser of the revenue recognized from End Users [i.e., consumers] for the bundle and the aggregate standalone published prices for End Users for each of the component(s) of the bundle that are License Activities . . . . Determination, Attachment A at 8 (§ 382.2 therein). In Johnson, the D.C. Circuit succinctly summarized these conflicting definitions as follows: In its Initial Determination, the [Judges] directed that the revenue from streaming services that are included in bundled offerings would generally be measured by the value remaining after subtracting the prices attributable to the other products in the bundle. When the Copyright Owners objected to the substance of that definition in their motion for ‘‘clarification,’’ the Board adopted an entirely new definition of Service Revenue for bundled offerings. . . . This new definition generally measured the value of the streaming component of a bundle as the standalone price of the streaming component. E:\FR\FM\10AUR2.SGM 10AUR2 54466 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations Johnson at 389.207 In the Clarification Order, the Judges succinctly summarized the parties’ respective positions. Id. at 17. They noted that Copyright Owners had presented evidence that the PR II-based benchmark definition contained in the Initial Determination ‘‘led in some cases to an inappropriately low revenue base,’’ although the Judges ‘‘agree that there is no support for any sweeping inference that cross-selling has diminished the revenue base.’’ Id. at 17, 21 (emphasis added). The Judges further noted the Services’ assertion that the Bundled Revenue definition in the Initial Determination is consistent with the Judges’ ‘‘endorsement of the classic price discrimination enabled by bundling strategies.’’ Id.208 The Majority resolved this issue in the Clarification Order in favor of Copyright Owners. Specifically, the Majority found that, because of the ‘‘indeterminacy problem’’ 209 inherent in bundling, ‘‘the Services—not the Copyright Owners—. . . are in a position to provide evidence of how they price bundles and value the component parts thereof.’’ Id. at 17–18. However, according to the Majority, although the Services ‘‘bore the burden of providing evidence concerning the proper economic allocation of bundled revenue,’’ they ‘‘failed to do so,’’ and ‘‘[b]y default . . . the Judges must adopt an approach to valuing bundled revenue that is in line with what the Copyright Owners have proposed.’’ Id. at 18. ddrumheller on DSK120RN23PROD with RULES2 IV. The Rulings in Johnson Regarding the Bundled Revenue Definition The Services appealed the Majority’s abandonment of the Bundled Revenue definition in their Initial Determination. Their appeal ‘‘challenge[s] both the legal authority and the substantive soundness’’ of this switch. 207 As explained infra (including by way of an example), the Bundled Revenue definition in the Initial Determination aligns with and incentivizes price discrimination in the downstream market, but the definition in the Determination does not. 208 The parties’ substantive arguments are discussed in more detail infra. 209 The ‘‘economic indeterminacy arises when ‘‘the input suppler . . . is paid as a percent of retail revenue, and the bundled revenue consists of some revenue attributable to the royalty base and other revenue excluded from the royalty base, the economic indeterminacy of the revenue attributable to each bucket creates a measurement problem, absent further information regarding the WTP [Willingness-to-Pay] of buyers/subscribers to the bundle.’’ SDARS III, 83 FR 65264. As explained infra, the PR II-based benchmark addresses this informational uncertainty with the parties’ negotiated alternative rate prongs and floors that guarantee royalties are paid, whereas the definition in the Determination eliminated the alignment of royalties to price discriminatory bundles designed to increase downstream access to musical works. VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 First, the Services argued that the Majority failed to identify and explain the procedural basis for making the switch after the hearing had concluded. Second, the Services argued that, substantively, the replacement definition in the Determination ‘‘was arbitrary, capricious, or unsupported by substantial evidence.’’ Johnson at 389, 392. The D.C. Circuit agreed with the Services regarding the procedural issue and therefore vacated and remanded that aspect of the Bundled Revenue definitional switch. In light of its procedural ruling, the D.C. Circuit explicitly declined to rule on the Services’ substantive argument relating to the definitional switch. Id. at 392. (‘‘Because the Board failed to explain the legal authority for its late-breaking rewrite, we vacate and remand that aspect of the decision [and] we have no occasion to address the Streaming Services’ separate argument that the definition was arbitrary, capricious, or unsupported by substantial evidence.’’). The D.C. Circuit’s rulings in Johnson pertaining to this Bundled Revenue Definition were clearly articulated. The D.C. Circuit found that the Majority ‘‘failed to explain under what authority’’ it made a material change to the definition ‘‘so late in the game.’’ Johnson at 389, 392. The D.C. Circuit noted that the Judges expressly declined to treat the Clarification Motion as a motion for rehearing; consequently, the motion did not request and the Judges did not reconsider either evidence or legal argument. Id. at 390. Although appellate counsel offered rationales, the D.C. Circuit rejected counsel’s post hoc reasoning. Id. and 391–92. Ultimately, the D.C. Circuit remanded the adopted regulation ‘‘either to provide ‘a fuller explanation of the [Judges’] reasoning at the time of the agency action[,]’ or to take ‘new agency action’ accompanied by the appropriate procedures.’’ Id. at 392, citing Dep’t of Homeland Sec. v. Regents of Univ. of Cal. 140 S.Ct. 1891, 1908 (2020). To be precise, I take note of the following specific rulings in Johnson: 1. ‘‘The problem is that the [Majority] has completely failed to explain under what authority it was able to materially rework that definition so late in the game.’’ Id. at 389. 2. ‘‘The [Majority] did not treat Copyright Owners’ motion to have the definition changed as a motion for rehearing . . . [because] Copyright Owners’ motion did not request a literal rehearing of evidence or legal argument.’’ Id. at 390 (cleaned up). 3. ‘‘The [Majority] nowhere in its order or the [ ] Determination explains PO 00000 Frm 00062 Fmt 4701 Sfmt 4700 the source of its power to make ‘fundamental’ changes under the authorizing statute . . . .’’ Id. at 392. [same as #1] 4. ‘‘[I]t should go without saying that we may not sustain the Board’s action based on its attorney’s theorizing at oral argument . . . vacillating gestures to uninvoked authority will not do.’’ Id. at 391–92 (the D.C. Circuit alluding to its rejection of arguments also made only by appellate counsel in support of the Majority’s rejection of the PR II-based benchmark earlier in the decision).210 ‘‘We must vacate the [ ] Determination’s bundled offering Service Revenue definition and remand for the [CRB Judges] either to provide ‘fuller explanation of the agency’s reasoning at the time of the agency action[,]’ or to take ‘new agency action’ accompanied by the appropriate procedures.’’ Id. at 392. V. Remand Procedure Regarding Bundled Revenue Definition Post-remand, the Judges stated their understanding, as well as the parties’ understanding, of the issue on remand with respect to the Bundled Revenue definition: The Services and Copyright Owners agree that the proceedings on remand should be limited to three issues: * * * [3] the adoption of a revised definition of ‘‘service revenue’’ for bundled offerings between issuing their Initial Determination and [their] Determination. Order Regarding Proceedings on Remand at 1 (Dec. 15, 2020) (Remand Order). The parties proposed, and the Judges agreed, that the record would not be reopened with regard to the Bundled Revenue definitional issue. Rather, the Remand Order permitted the parties only to provide further briefing on this matter. Id. Specifically, the Judges subsequently permitted each party to file simultaneous Initial Remand Submissions and simultaneous Reply Remand Submissions. See Order Adopting Schedule for Proceedings on Remand (Dec. 20, 2020). Thereafter, seeking further analysis on the question of ‘‘new agency action,’’ the Judges solicited, and received, further briefing on this issue. See Notice and Sua Sponte Order Directing the Parties to Provide Additional Materials (Dec. 9, 210 Going beyond the Majority’s actual rulings, the CRB Judges’ appellate counsel argued that the Majority’s authority for this definitional switch fell under either or both of the ‘‘inherent’’ statutory powers of the Judges or their ‘‘rehearing power.’’ Id. at 392. (The D.C. Circuit rejecting appellate counsel’s argument that it was unnecessary ‘‘for this Court to address which one it is because . . . it could properly be understood as both.’’). E:\FR\FM\10AUR2.SGM 10AUR2 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations 2021) (Feb. 9, 2021); Sua Sponte Order Regarding Additional Briefing (Feb. 9, 2021). VI. The Parties’ Submissions Regarding Bundled Revenue Definition In their respective briefing, Copyright Owners and the Services made arguments relating to: (1) the procedural issue, i.e., the Judges’ authority, vel non, to switch to a new Bundled Revenue definition in the Determination; and (2) the substantive issue, i.e., the relative merits of the two conflicting Bundled Revenue definitions. See Initial Remand Submission of Copyright Owners at 7– 10 (Apr. 1, 2021) (CO Initial Submission); Services’ Joint Opening Brief (in Services’ Joint Written Direct Remand Submission at Tab D) at 64–76 (Apr. 1, 2021) (Services’ Initial Submission); Copyright Owners’ Reply Brief on Remand (in Reply Remand Submission of Copyright Owners, Vol. 1) at 64–88 (CO Reply); Services’ Joint Reply Brief at 52–63 (Services’ Reply). A. The Procedural Issue ddrumheller on DSK120RN23PROD with RULES2 1. Copyright Owners’ Arguments Copyright Owners assert first that the Judges can preserve their post-hearing switch of the Bundled Revenue definition by sidestepping the D.C. Circuit’s holding and rationale in Johnson. That is, Copyright Owners maintain that this remand proceeding itself constitutes the necessary form of ‘‘new ‘agency action’’’ that Johnson invites, while also liberating the Judges from the consequences of the procedural infirmities identified by the D.C. Circuit. More particularly, Copyright Owners argue: [T]he new agency action here is a determination after remand proceedings[.] [T]he [Judges are] largely free to chart [their] own procedural course, and [they] ha[ve] done so in [their] [Remand] Order. The [Judges are] not required to undertake any of the procedural steps set forth in 17 U.S.C. 803(b) in order to take such ‘‘new agency action.’’ See 17 U.S.C. 803(d)(3) (requiring only that on remand further proceedings be taken ‘‘in accordance with subsection (a)’’); 37 CFR 351.15; Intercollegiate Broad. Sys., Inc., 796 F.3d at 125 (‘‘[N]either the Copyright Act nor the [Judge’s] regulations prescribe any particular procedures on remand.’’) The Circuit’s instruction that the action be ‘‘accompanied by the appropriate procedures[,]’’ Johnson, 969 F.3d at 392, does not dictate what those ‘‘appropriate procedures’’ must be but instead plainly refers to these flexible rules. See also Oceana, Inc., 321 F. Supp. 3d at 136 (explaining that when remanding to an agency, a court generally ‘‘may not dictate to the agency the methods, procedures, or time dimension, for its reconsideration’’). CO Initial Submission at 71 n.33. VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 Copyright Owners reject the Services’ position that the asserted procedural error is an ‘‘absence of authority’’ that can never be cured. Id. at 74 (citing Services’ Proposal for Remand Proceedings at 10). They note that the D.C. Circuit did not say the Judges lacked the authority to revisit the service revenue definition from bundles on remand. Nor, they observe, did it say the Judges have no authority to review the record evidence and the parties’ arguments and reach the same conclusion or a different conclusion on remand. Copyright Owners further opine that if the only possible outcome were for the Judges to reinstate a definition that lacked any explanation or evidentiary support solely because it was present in the Initial Determination, then the D.C. Circuit would not have remanded the issue but would have simply reversed and reinstated the Initial Determination definition. But instead, they note, the D.C. Circuit remanded and said the Judges could take ‘‘new agency action’’ precisely to cure the asserted procedural defect. Copyright Owners assert that the remand allowed the parties to present the record evidence and their arguments so that the Judges can address the definition ‘‘afresh’’ in the remand determination. Id. at 74. Further, Copyright Owners argue that 17 U.S.C. 803(d)(3) states only that proceedings on remand must be in accordance with 17 U.S.C. 803(a). They contend that remand proceedings need not be confined to procedures the Services claim are too late in the game for the Judges to follow, again relying on the holding in Intercollegiate Broad. Sys., supra, that ‘‘neither the Copyright Act nor the Board’s regulations prescribe any particular procedures on remand.’’ Id. at 125. Accordingly, they argue, the Judges can reaffirm the adopted bundled service revenue definition following their review of the parties’ submissions without invoking section 803(c)(2) or 803(c)(4) that were ruled inapplicable in Johnson. CO Reply at 65–66. Also, Copyright Owners argue that the Judges may properly justify the changed definition under section 803(c) as a fuller explanation of the agency’s reasoning at the time it was made. They urge that the Judges could explain that, especially in light of the evidence of how (in Copyright Owners’ characterization) the Services misused the prior definition to make service revenue completely disappear, the Judges carry-over of the prior Bundled Revenue definition from Phonorecords II into the Initial Determination was PO 00000 Frm 00063 Fmt 4701 Sfmt 4700 54467 unintended and inadvertent.211 CO Reply at 69. Copyright Owners also assert that, on remand, the Judges could explain that Copyright Owners had, in their Motion for Clarification, identified an ‘‘exceptional case’’ under section 803(c)(2) because the prior definition failed to comport with Judges’ precedent and economic principles, and was unsupported by evidence. In addition, the Judges reheard the evidence and legal arguments as presented in the parties’ briefs on the issue and, as a result, chose to adopt the revised definition. Copyright Owners maintain that for the Judges to do so would not be impermissible post-hoc reasoning. They note that the D.C. Circuit remanded precisely because the Judges did not provide any reason in the Determination for revising the Bundled Revenue definition. Copyright Owners note that it was the Services, not Copyright Owners, who appealed the Judges’ modification of the bundled service revenue definition; thus, Copyright Owners cannot be penalized for not making every possible argument for affirmance. CO Reply at 70. Further, and again notwithstanding the holding in Johnson, Copyright Owners argue that the Judges have the authority to engage in new agency action in this remand proceeding through a recasting of the Motion for Clarification as a motion for rehearing, pursuant to 17 U.S.C. 803(c)(2)(A) and 37 CFR 353.1. In this regard, Copyright Owners dismiss the point, raised by the D.C. Circuit, that their Motion for Clarification could not be recast as a motion for rehearing because Copyright Owners had explicitly disavowed that their motion sought rehearing under the statute, and that the Judges agreed. Rather, Copyright Owners maintain that the foregoing is not the same as a finding that the standard could not have been met. In Copyright Owners’ view, the Judges could revisit on remand the question of whether the rehearing standard has now been met, and find that Copyright Owners have satisfied the ‘‘exceptional case’’ standard for granting rehearing motions under section 803(c)(2).212 Copyright Owners 211 Copyright Owners assert that the definition in the Initial Determination conflicted with the CRB Judges’ finding in the Initial Determination that the adopted rates and terms would afford Copyright Owners a fair return for their creative works, thereby satisfying Factor B of the 801(b) standard. Thus, they maintain that the definitional switch was necessary so as to not ‘‘frustrate the proper implementation of’’ the Determination. CO Reply at 69 (citing 17 U.S.C. 801(b) and 803(c)(4)). 212 The Majority set forth the rehearing standard in the Clarification Order: ‘‘According to the E:\FR\FM\10AUR2.SGM Continued 10AUR2 54468 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations add that if the Judges do engage in new agency action that reconsiders the Motion for Clarification as a motion for rehearing, the Judges should fully explain their reasoning. Id. at 8–10. However, Copyright Owners urge that proceeding in that fashion would add an entirely unnecessary and complicating step. They again suggest that there is no need to reconsider or recharacterize the Motion for Clarification as a motion for rehearing because the remand itself affords the opportunity for the Judges to take new agency action, which, as in a rehearing, permits them to reconsider evidence and arguments, but, unlike a rehearing, is not limited by the constraints of section 803(c)(2). See Copyright Owners’ . . . Additional Briefing on New Agency Action . . . Question, etc., Tab B at 7–8 (Feb. 24, 2021). ddrumheller on DSK120RN23PROD with RULES2 2. The Services’ Arguments The Services’ arguments are based on the reasoning of the D.C. Circuit in Johnson. Specifically, they assert that the D.C. Circuit found only ‘‘three ways in which the [Judges] can revise Initial Determinations’’ via ‘‘new agency action,’’ and the Judges failed to establish that the change to the service revenue definition fit any of those three categories. Services’ Initial Submission at 64–65 (citing Johnson at 390).213 According to the Services, the first statutory way the Judges may revise an Initial Determination is to ‘‘order rehearing ‘in exceptional cases’ in response to a party’s motion, 17 U.S.C. 803(c)(2)(A).’’ Services’ Initial Submission at 65 (citing Johnson at 390). The Services argue that the D.C. Circuit held in Johnson that the Judges’ ‘‘material revision of the ‘[Bundled] Revenue’ definition . . . does not fall within the [Judges’] rehearing authority under section 803(c)(2)(A)’’ because ‘‘the [Judges] [themselves] . . . w[ere] explicit that [they] ‘did not treat the Copyright Act, the Judges may grant a motion for rehearing in exceptional circumstances, provided the moving party shows that an aspect of the determination is ‘‘erroneous.’’ See 17 U.S.C. 803(c)(2); 37 CFR 353.1. The moving participant must identify the aspects of the determination that it asserts are ‘‘without evidentiary support in the record or contrary to legal requirements.’’ 37 CFR 353.2. In general, the Judges grant rehearing only ‘‘when (1) there has been an intervening change in controlling law; (2) new evidence is available; or (3) there is a need to correct a clear error or prevent manifest injustice.’’ See, e.g., Order Denying Motion for Reh’g at 1, Docket No. 2006–1 CRB DSTRA (Jan. 8, 2008) (SDARS I Rehearing Order) (applying federal district court standard under Fed. R. Civ. P. 59(e)).’’ Clarification Order at 2, n.3. 213 The Services acknowledge that the Judges could alternatively have attempted to provide on remand a fuller explanation of their prior reasoning (in lieu of engaging in ‘‘new ‘agency action’’’). That issue is considered infra. VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 [Copyright Owners’] motion[ ]’ . . . ‘as [a] motion[ ] for rehearing under 17 U.S.C. 803(c)(2).’’’ Id. The D.C. Circuit also noted that ‘‘as the [Judges] found, . . . Copyright Owners’ motion did ‘not meet [the] exceptional standard for granting rehearing motions’ under section 803(c)(2).’’ Id. (citing Johnson at 390). The Services assert, quoting Johnson once more, that the Judges were not able to make ‘‘a volte-face’’ and justify on appeal their revision to the definition as an exercise of rehearing authority. As the D.C. Circuit held, agency action must be justified by ‘‘reasons invoked by the agency at the time it took the challenged action,’’ and post-hoc rationalizations are insufficient. Id. (citing Johnson at 390). The Services add their view that the Judges cannot revisit the decision to deny rehearing without engaging in impermissible post-hoc reasoning. They note the Supreme Court has explained that, while an agency may ‘‘elaborate later’’ on its ‘‘initial explanation’’ of the reason (or reasons) for its action, it ‘‘may not provide new ones.’’ Services’ Initial Submission at 66, citing e.g., Regents at 1908. The Services offer that the Judges, having stated that they did not consider the Copyright Owners’ motion to revise the definition to be a motion for rehearing, cannot now conclude that the motion qualified as one for rehearing and that the Judges in fact engaged in rehearing. Id.214 The Services next argue, relatedly, that the Judges cannot simply recast the Services Motion for Clarification as a rehearing motion in an attempt to satisfy the rehearing standard. In this regard, they maintain that Copyright Owners did not argue before the Judges or the D.C. Circuit that their Motion for Clarification satisfied the ‘‘exceptional cases’’ standard, and have therefore waived that argument. Id. The Services assert that the second statutory way the Judges may revise an Initial Determination, viz. taking ‘‘new agency action’’ to correct a technical or clerical error under section 803(c)(4), cannot be used to justify the modification of the Bundled Revenue definition in the Initial Determination. The Services note that the D.C. Circuit held specifically that the Judges’ change in the Bundled Revenue definition could not be construed as correcting a technical or clerical error because it involved a substantive rewrite of the Service revenue definition. Id. at 67 (citing Johnson at 391). 214 In fact, the issue of whether to characterize Copyright Owners’ Motion for Clarification as a motion for rehearing is not one raised by Copyright Owners, but rather by the Judges sua sponte. PO 00000 Frm 00064 Fmt 4701 Sfmt 4700 The Services argue that the third and final statutory justification for the Judges to engage in ‘‘new agency action’’ is to revise the terms in an Initial Determination is in response to ‘‘unforeseen circumstances’’ that would frustrate the proper implementation of the determination. Id. at 67. The Services note that the D.C. Circuit held in Johnson that this authority did not justify the Judges’ change to the Bundled Revenue definition because the Judges did not invoke this authority and ‘‘the need to ground the original definition in the record’’ could not credibly be described as ‘‘an unforeseen circumstance.’’ Id. (citing Johnson at 391). The Services also note that the D.C. Circuit rejected the argument that the Judges have an ‘‘inherent authority’’— unmentioned in the statute—to make changes to the Initial Determination. The D.C. Circuit explained that the specific restrictions Congress placed on the [Judges’] authority in section 803 ‘‘would be a nullity if [they] also had plenary authority to revise [their] determinations whenever [they] thought appropriate.’’ Id. (citing Johnson at 391– 92). The Services add that even if the Judges offered a new source of authority capable of justifying substantive changes to the [Bundled] Revenue definition now, the Judges would be unable to rely on this ‘‘uninvoked authority’’ without engaging in impermissible post-hoc reasoning. Id. The Services also reject Copyright Owners’ position that the Judges may sidestep the D.C. Circuit’s ruling by issuing a new determination on remand and simply arguing that any ruling after remand qualifies as new agency action pursuant to Johnson. The Services argue that failure to address the legal and factual issues on which the court remanded would violate the D.C. Circuit’s decision and would result in yet another remand. The Services emphasize that the issue of authority to make the changes to the Initial Determination are especially important in this context, because the D.C. Circuit recognized that the Copyright Act places limits on the Judges’ authority to alter an initial determination by defining conditions for rehearing and the types of changes that are permitted absent a rehearing. In this regard, the Services maintain that the Judges cannot do on remand what they lacked authority to do in the first instance. The Services assert that the Judges must resolve the legal question of whether authority exists to alter the revenue definition in E:\FR\FM\10AUR2.SGM 10AUR2 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations the Initial Determination. Services’ Reply at 52–54.215 The Services also take note of the alternative path available to the Judges: to provide a ‘‘fuller explanation’’ of the prior conclusion that the Judges had legal authority to revise the Service Revenue definition. The Services maintain that if the Judges pursue the ‘‘fuller explanation’’ path, the Judges are limited to elaborating on what they said previously, and that they cannot add new reasons they did not initially provide. Id. at 54–55; see also Services’ Joint Rebuttal Brief Addressing the Judges’ Working Proposal at 38–42 (Feb. 24, 2022) (‘‘Services’ Additional Submission’’). The Services address Copyright Owners’ position that if the only possible outcome were for the Judges to reinstate a definition that lacked any explanation or evidentiary support solely because it was present in the Initial Determination, then the D.C. Circuit would not have remanded the issue but would have simply reversed and reinstated the Initial Determination definition. The Services urge that the D.C. Circuit could not reverse because the CRB’s appellate counsel had raised—for the first time on appeal— new justifications for the Judges’ decision to change the Initial Determination. Instead, the Services maintain, the D.C. Circuit had to remand and give the Judges the opportunity to address appellate counsel’s new justifications in the first instance, as the D.C. Circuit could not rule them out given the posture of the appeal. Services’ Reply at 56. ddrumheller on DSK120RN23PROD with RULES2 VII. Analysis and Decision A. The Procedural Issue: Is There ‘‘New Agency Action’’ Available to the Judges? Having considered the parties’ arguments, I conclude that the rulings in Johnson, which clearly rejected all of the Majority’s procedural arguments seeking to justify their switch in the Bundled Revenue definition, foreclose any avenue for procedurally justifying this definitional switch. More particularly, I conclude that none of the procedural avenues proffered by Copyright Owners would constitute ‘‘new ‘agency action’’’ consonant with the holdings in Johnson. Further, I cannot identify any other procedural device (i.e., an extra-statutory form of agency action) that would permit the 215 In The Services agree that this remand proceeding qualifies as a ‘‘new agency action’’ but do not maintain that a ruling on remand that is inconsistent with Johnson would be the type of ‘‘new ‘agency action’’’ that Johnson permits. See Services Additional Submission at 38–42. VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 switched definition in a manner consistent with Johnson.216 In addition, I cannot identify any further or fuller explanation that might support the Majority’s procedural reasoning for swapping out the Bundled Revenue definition in the Initial Determination and substituting the definition in the Determination. In reaching this conclusion, I take note of the following specific language in Johnson: Section 803 identifies three ways in which the Board can revise Initial Determinations. It can (i) order rehearing ‘‘in exceptional cases’’ in response to a party’s motion; (ii) correct ‘‘technical or clerical errors,’’; and (iii) ‘‘modify the terms, but not the rates’’ of a royalty payment, ‘‘in response to unforeseen circumstances that would frustrate the proper implementation of [the] determination.’’ Johnson at 390 (citations omitted). After identifying these three alternatives, the D.C. Circuit concluded that the CRB Judges ‘‘rollout of an entirely new manner for calculating the streaming service revenue from bundled offerings fit none of those categories.’’ Id. First, I consider whether in the present case they can engage in ‘‘new ‘agency action’’’ pursuant to 17 U.S.C. 803(c)(2)(A) by recasting Copyright Owners’ Motion for Clarification as a Motion for Rehearing. I conclude that this avenue has been unambiguously cut-off by Johnson and, indeed (as noted in Johnson), by the Judges’ own prior ruling: The [CRB Judges’] material revision of the Bundled Revenue definition . . . does not fall within [their] rehearing authority under Section 803(c)(2)(A). We have that on no less an authority than the [CRB Judges themselves], [who were] explicit that [they] ‘‘did not treat the Copyright Owners’ motion’’ to have the definition changed ‘‘as a motion] for rehearing under 17 U.S.C. 803(c)(2).’’ That is because the Copyright Owners’ motion did not ‘‘request[ ] a literal rehearing of evidence or legal argument.’’ Nor could they have because, as the [CRB Judges] found, the Copyright Owners’ motion did ‘‘not meet [the] exceptional standard for 216 In this section, Copyright Owners’ arguments regarding recasting their Motion for Clarification as a request for rehearing, a correction for technical or clerical errors, or for unforeseen circumstances would constitute a new application of an existing ‘‘form of agency action’’ that the D.C. Circuit had rejected. But Copyright Owners’ argument in favor of the Judges’ supposed ‘‘inherent authority’’ to enlarge their post-hearing jurisdiction is an argument creating a new form of agency action, not an argument in favor of new application of an existing form of authority. Likewise, the next approach proffered by Copyright Owners, i.e. construing the remand itself as generating the requisite agency action, which is also the Majority’s approach, is an example of an agency action that is not statutorily specified and, as explained infra, is inconsistent with section 803(a). PO 00000 Frm 00065 Fmt 4701 Sfmt 4700 54469 granting rehearing motions’’ under Section 803(c)(2). . . . [The CRB Judges] explain[ed] that . . . Copyright Owners ‘‘failed to make even a prima facie case for rehearing under the [rehearing] standard’’. Johnson, 369 F.3d at 390. Further cutting off this ‘‘rehearing’’ approach, Johnson also expressly holds that it is a ‘‘forceful’’ principle that the D.C. Circuit ‘‘cannot sustain action on grounds that the agency itself specifically disavowed. Id. Moreover, in this Initial Remand Ruling I echo the Majority’s ruling in the Clarification Order that Copyright Owners had failed to present ‘‘even a prima facie case for rehearing under the applicable standard’’. Clarification Order at 2.217 Next, I consider whether the Judges can engage in ‘‘new ‘agency action’’’ by recharacterizing their switch of the Bundled Revenue definition as an attempted correction of ‘‘technical or clerical errors,’’ pursuant to their ‘‘continuing jurisdiction’’ under section 803(c)(4). Once again, they cannot, and the D.C. Circuit has effectively explained why this is so: The [Judges] do[ ] not even try to squeeze [their] substantive rewrite of the Service Revenue definition into that [§ 803(c)(4)] category. Quite the opposite, the [Judges] admit[ ] that the new definition ‘‘represent[s] a departure’’ from the definition in the Initial Determination, and was a substantive swap designed to ‘‘mitigate’’ the alleged ‘‘problem’’ of the original definition leaving the interactive streaming service providers free to ‘‘obscure royalty-based streaming revenue by offering product bundles that include music service offerings with other goods and services[.]’’ . . . To that same point, the order itself labels the initial and new definitions ‘‘diametrically-opposed approaches to valuing bundled revenues.’’ . . . . Nothing technical or clerical about that. Johnson at 391. On remand, I am unable to ascertain any basis for describing or justifying the changed Bundled Revenue definition as a technical or clerical correction. Thus, I conclude that the Judges cannot engage in ‘‘new ‘agency action’’’ pursuant to this section. Next, I consider whether the Judges can engage in ‘‘new ‘agency action’’’— 217 The first two bases for rehearing under the statute, viz., change in the controlling law and the availability of new evidence, clearly do not apply. The third basis, i.e., to correct a clear error or prevent manifest injustice, also does not apply. As explained herein, the substantive difference between the conflicting Bundled Revenue definitions should be resolved consistent with the Judges’ adoption of the PR II-based benchmark and the parties’ negotiated compromise of the ‘‘price discrimination vs. revenue diminution’’ dilemma. This resolution does not constitute an ‘‘error,’’ let alone a ‘‘clear error,’’ and maintaining the parties’ rate architecture from the Initial Determination does not generate any ‘‘injustice,’’ ‘‘manifest’’ or otherwise. E:\FR\FM\10AUR2.SGM 10AUR2 54470 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations by trying to squeeze the square peg of their definitional swap into the round hole that is the ‘‘unforeseen circumstances’’ clause in section 803(c)(4). That provision permits the Judges to exercise ‘‘continuing jurisdiction’’ if necessary to modify a regulatory term in a determination in response to ‘‘unforeseen circumstances,’’ if the absence of modification would frustrate the proper implementation of the determination. Once again, Johnson shuts the door: Come oral argument, the [Judges] attempted to explain that ‘‘the unforeseen circumstances would be that [they] initially adopted a definition that was not supported by the record, and that was in fact substantively unreasonable and would frustrate the proper implementation of their determination.’’ . . . It is hard to see how the need to ground the original definition in the record was an unforeseen circumstance. That is Administrative Law 101. See also 17 U.S.C. 803(c)(3) (‘‘A determination of the [Judges] shall be supported by the written record.’’). ddrumheller on DSK120RN23PROD with RULES2 Johnson at 391 (cleaned up). I agree. The present panel of Judges is bound by the D.C. Circuit’s ruling that the overlooking of the need to ground in the factual record the Bundled Revenue definition in the Initial Determination cannot constitute an ‘‘unforeseen circumstance.’’ Accordingly, I am unable to ascertain any basis for describing or justifying the changed Bundled Revenue definition as an ‘‘unforeseen circumstance’’ that would justify their invocation of ‘‘continuing jurisdiction.’’ I further consider the argument (made by the Judges’ appellate counsel and by Copyright Owners) that the Judges have the ‘‘inherent authority sua sponte to make any ‘appropriate’ substantive . . . or ‘fundamental’ changes after the Initial Determination . . . that [they] believe[ ] serve ‘the interests of enhancing the clarity and administrability of the regulatory terms accompanying the [ ] Determination.’’’ Johnson at 391. The D.C. Circuit made short work of this argument as well, stating that, although the CRB Judges have ‘‘considerable freedom’’ with regard to determining their own procedures that flexibility must be exercised within the lines drawn by the authorizing statute. Congress’s decision to limit rehearing to ‘‘exceptional cases,’’ and to confine other post hoc amendments to cases involving ‘‘technical or clerical errors,’’ would be a nullity if the [Judges] also had plenary authority to revise [their] determinations whenever [they] thought appropriate. The [Judges] nowhere in [their] order or the [ ] Determination explain[ ] the source of [their] power to make ‘‘fundamental’’ changes under the authorizing statute . . . any time [they] VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 deem such changes ‘‘appropriate’’ . . . even after the Initial Determination. Johnson at 392.218 As with regard to the proffered rationales discussed supra, I cannot identify any authority that would allow the Judges to declare for themselves in the present factual and legal context an ‘‘inherent’’ authority to override the Copyright Act and declare their right to engage in ‘‘new ‘agency action.’’’ Finally, I consider Copyright Owners’ suggestion that the remand itself by the D.C. Circuit permits the Judges, pursuant to the Copyright Act, to engage in any procedure necessary to support their switch in the Bundled Revenue definition. The present Majority essentially adopts this procedural approach. However, I reject that argument as meritless. The argument begins with a correct premise but seriously veers off course. Copyright Owners correctly note (and the Services do not disagree) that this remand proceeding constitutes ‘‘new ‘agency action.’’’ Copyright Owners then maintain that, because the Copyright Act does not provide for procedures that govern remand proceedings, the Judges are statutorily unconstrained with regard to the procedures they may adopt. This premise, although perhaps correct in other contexts, is most definitely incorrect in this specific context, given the clear holding in Johnson. Here, the D.C. Circuit has been unequivocal in identifying the statutory limitations that precluded the Judges from switching out the Bundled Revenue definition in their Initial Determination and replacing it with a different definition in the Determination that was, to use the Majority’s phrase, ‘‘diametrically opposed’’ to the prior definition, in that it would eliminate the royalty-based incentive to price discriminate via bundling.219 But Copyright Owners assert that the remand itself clothes the Judges with the procedural authority to make the very switch that Johnson forbids! I do not understand the D.C. Circuit to have admonished the Majority for its failure to respect the boundaries of its jurisdiction, only to provide them, via remand, with a back-door through which they may circle-back and exceed those very boundaries. A reading of section 803(a), upon which Copyright Owners rely, provides 218 By the same reasoning, Johnson also rejected the Judges’ explanation in the Determination that they were permitted to treat Copyright Owners’ request as a general motion under § 350.4) of their regulations. Id. 219 This substantive impact of the definitional switch is discussed infra. PO 00000 Frm 00066 Fmt 4701 Sfmt 4700 a further demonstration of the error in this argument. This subsection lists the authorities whose pronouncements the Judges must ‘‘act in accordance with,’’ including, quite unsurprisingly, ‘‘the decisions of the court of appeals under this chapter.’’ 17 U.S.C. 803(a). In the instant case, the D.C. Circuit has unambiguously held that the Judges lacked the statutory authority to make the definitional switch at issue. For the Judges to construe that clear ruling as an implicit invitation to create new extrastatutory remand procedures that contradict the D.C. Circuit’s rationale for the remand would be inexplicable and would render useless the procedural ruling in Johnson.220 In sum, I cannot and do not understand that the D.C. Circuit intended in Johnson simply to write a meaningless procedural opinion that the Judges could not merely ignore, but use to cleanse the very procedural error the D.C. Circuit had condemned.221 Accordingly, the Bundled Revenue definition in the Initial Determination should be reinstated. As explained in the portion of the Initial Remand Ruling in which I join, this reinstatement is harmonious with the entirety of the Judges’ findings and conclusions regarding the other remanded issues. 220 In fact, this argument is dangerous. The CRB Judges or any administrative agency, could willfully engage in extra-statutory procedures to obtain a particular substantive result. If there is no appeal, the extra-statutory procedure would be successful. But if the extra-statutory procedure was the subject of a successful appeal resulting in a remand, the CRB Judges (or any agency) could declare the remand as license to engage once more in extrastatutory procedures in order to obtain the same substantive result. This is a ‘‘heads-I-win, tails-youlose’’ strategy. 221 Copyright Owners also argue that if the D.C. Circuit had intended in Johnson to prohibit the Judges from engaging in ‘‘new ‘agency action’’’ on remand, they would have reversed and reinstated the Initial Determination, rather than vacated and remanded that aspect of the Determination. But that argument confuses prudence with uncertainty. The D.C. Circuit prudently allowed the Judges, who are presumed to have particular knowledge of their duties, to consider whether there exist further explanations of their reasoning or ‘‘new ‘agency actions’’’ they could invoke to support their definitional switch. That prudence hardly suggests that the D.C. Circuit was sanguine about the existence of further explanations or additional actions that might support the switch. Also, 17 U.S.C. 803(d)(3) explicitly allows the D.C. Circuit to ‘‘vacate [a] determination of the . . . Judges and remand the case to the . . . Judges for further proceedings,’’ but only expressly allows the court to ‘‘enter its own determination’’’’ in connection with ‘‘the amount or distribution of royalty fees and costs, and order the repayment of any excess fees, the payment of any underpaid fees and the payment of interest pertaining respectively thereto . . . .’’ Id. Thus, it is hardly clear that the D.C. Circuit understood it had any choice upon vacating, save to remand for further proceedings. E:\FR\FM\10AUR2.SGM 10AUR2 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations B. The Substantive Issue: The Dueling Definitions of Bundled Revenue 1. Introduction: The Issue as Framed in the Clarification Order Regarding the definition of ‘‘Service Revenue’’ from bundled offerings, the Judges summarized the parties’ competing arguments: Copyright Owners presented evidence that the existing approach led, in some cases, to an inappropriately low revenue base—but did so in service to their argument that the Judges should reject revenue-based royalty structures. They did not present evidence to support a different measure of bundled revenue because their rate proposal was not revenue-based. The Services rely on the fact that the approach to bundled revenue in the extant regulations is derived from the 2012 Settlement. The Judges have, however, declined to rely on the 2012 Settlement as a benchmark, as the basis for the rate structure, or, therefore, as regulatory guidance. The Services have observed correctly that the evidentiary records in Web IV and SDARS III differ from the record in this proceeding.222 ddrumheller on DSK120RN23PROD with RULES2 Clarification Order at 17 (emphasis added). Despite these arguments, the Judges found that neither party presented evidence adequate to support the approach advocated in postdetermination filings, because the ‘‘economic indeterminacy problem inherent in bundling’’ remained unresolved. Id. The Judges stated that the Services were the party in possession of the relevant information, and concluded that the Services bore the burden of providing evidence that might mitigate the ‘‘indeterminacy problem’’ inherent in bundling. Because the Judges concluded that the Services had not met that burden, they ruled that they must adopt an approach to valuing bundled revenue that is in line with what the Copyright Owners proposed. As a result, the Judges discarded the formula in the Initial Determination and ruled, instead, that streaming service providers will use their own standalone price (or comparable) for the music component (not to exceed the value of the entire bundle) when allocating bundled revenue. Id. at 16–18. On remand, the parties have made the following arguments regarding the substance of the Bundled Revenue definition: 2. Copyright Owners According to Copyright Owners, the prior Bundled Revenue definition in the 222 In Web IV and SDARS III, unlike under the Phonorecords II-based benchmark, there were no minima or floors to provide licensors with royalties in the event bundled offerings would otherwise fail to generate royalties. VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 Initial Determination failed to address the ‘‘ ‘economic indeterminacy’ problem inherent in bundling’’ appropriately and in a way consistent with Judges’ precedent. CO Initial Submission at 75 (citing Clarification Order at 16–18). Copyright Owners proceeded to cite several portions of testimony from the Services’ economic experts who acknowledged this problem. Id. They then point to hearing testimony in which Copyright Owners repeatedly raised the ‘‘economic indeterminacy’’ problem and demonstrated what they characterized as the absurd results to which the prior definition had led. Id. at 76. They pointed out that under the Initial Determination, the first step in computing Bundled Revenue was to identify revenues recognized from the entire bundle (i.e., the price paid by the subscriber). The second step was to subtract ‘‘the standalone published price’’ for all non-music components of the bundle. According to Copyright Owners, [REDACTED]. Id. at 76, 83. Copyright Owners point out that the Judges already found with respect to other licenses that such an approach is not only fundamentally unfair, but ‘‘absurd.’’ Id. (citing Web IV, 81 FR 26316, 26382 (May 2, 2016) (webcaster licenses); see also SDARS III, 83 FR 65210, 65264 (Dec. 19, 2018) (SDARS licenses) (rejecting proposed deductions by service from bundle revenues because of the ‘‘acknowledged ‘economic indeterminacy’ problem inherent in bundling’’). Copyright Owners concur with the Judges’ conclusion that the same reasoning applies to Phonorecords III. Id. at 76–77 (citing Clarification Order at 18 (‘‘the ‘economic indeterminacy’ problem inherent in bundling is common to all three proceedings.’’)). Copyright Owners offer that Spotify conceded to this flaw in the definition in the Initial Determination, but offered an alternative that contained the same loophole. Id. at 77–78. Copyright Owners also point out that the proponent of a term bears the burden of proof as to adoption. The Judges made clear that the licensee who wishes to offer bundles must bear the burden of providing evidence that might mitigate the acknowledged economic indeterminacy problem inherent in bundling, because any such evidence would be in its possession, not in the possession of the licensors. Id. at 79 (citing SDARS III, 83 FR 65264 (‘‘bundling [is] undertaken to increase [the Services’] revenues and it would be reasonable to assume that [the Services have] information relevant to the economic allocation of the bundled revenue.’’)). Copyright Owners contend PO 00000 Frm 00067 Fmt 4701 Sfmt 4700 54471 they presented unrebutted evidence showing the unreasonableness of the Services’ proposed definition while the Services offered no evidence to support their definition. Id. at 78, 79 (citing Clarification Order at 18). Copyright Owners maintain that no Service offered evidence concerning the separate values of the constituent parts of the bundles, or any other evidence concerning the economic allocation of bundled revenue, let alone the reasonableness of the definition in the Initial Determination. Id. at 80. Copyright Owners assert that in the absence of evidence to support the proposed definition, the Judges may adopt or fashion a definition of service revenue for bundled offerings that comports with the record evidence, which is precisely what the Judges did and, through new agency action, do again. Id. at 81. They further argue that the hearing record and the Judges’ precedent and reasoning further explain the unreasonableness of the prior definition and support the adopted bundle revenue definition. Id. at 82. Copyright Owners offer that in contrast to the Services’ evidentiary failure, they have provided sufficient evidence showing the unreasonableness of the Services’ proposed definition. They maintain that the definition adopted by the Judges in the Determination was consistent with the statutory factors and the evidence in the proceeding showing how the prior definition had been manipulated and ‘‘led, in some cases, to an inappropriately low revenue base.’’ Id. at 83 (citing Clarification Order at 17– 18). Copyright Owners dispute the Services’ assertion that there is support for the Phonorecords II approach to bundles in the record of this proceeding. Instead, Copyright Owners argue, the Services’ purported evidence at most supports the benefits of the practice or strategy of bundling. They maintain that the strategy of bundling covered music services with other products or services has nothing to do with [REDACTED]. They offer that the definition in the Initial Determination has nothing to do with such benefits, and that those benefits may be equally served by a definition that ensures value is apportioned to the music component in the bundle. CO Reply at 73–76. 3. The Services The Services argue that the evidence in the existing written record addressing bundles shows both that this definition is supported by the Phonorecords II benchmark and that it has proven industry-wide benefits. Services’ Initial Submission at 68. They emphasize that E:\FR\FM\10AUR2.SGM 10AUR2 ddrumheller on DSK120RN23PROD with RULES2 54472 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations Copyright Owners did not propose an alternative definition of service revenue until after the Judges issued the Initial Determination and that any definition they propose now would fail the basic requirement that the Judges must adopt rules ‘‘on the basis of a written record.’’ Id. (citing 17 U.S.C. 803(a)(1) and 803(c)(3)). Addressing the merits of the definition contained in the Initial Determination, the Services argue that it best serves the goals of the Copyright Act; that as a bright-line, easily administered rule, it continues the broad industry agreement from Phonorecords II, which ‘‘was negotiated voluntarily between the Services and . . . Copyright Owners—strong evidence that its terms are mutually beneficial.’’ Services’ Initial Submission at 69. The Services contend the prior negotiated definition increases output and incentivizes beneficial price discrimination to reach casual and passive listeners who would otherwise not pay for music and thus would not generate revenue from which royalties could be paid. With regard to [REDACTED]. Id. at 71 (and record citations therein). They further state that the definition of Bundled Revenue in Phonorecords II also enabled funneling of many of listeners into full-priced, full-catalog services. The Services allege that Copyright Owners also ignore the extensive royalties that were generated. They add that, for casual/passive listeners and those who may be funneled to subscription services, the per-subscriber minimum guarantees that the Copyright Owners will still be paid a fair royalty. The Services then cite several portions of testimony from various Services’ economic experts who point out the realization of an expanded royalty pool, which the Services offer as proving a functioning marketplace. Id. at 68–74.223 The Services maintain that while neither the Services nor Copyright Owners submitted evidence specifically addressing the way that customers, Services, or Copyright Owners might value the component parts of bundles, such subjective valuations are unnecessary—given that the parties’ negotiated handling of the bundling issues provides the Judges with ample support for the PR II-based benchmark definition in the Initial Determination. See id. at 75–76. 223 The Services’ Reply reiterates this point and offers that the testimony cited by the Copyright Owners also shows why the Initial Determination’s Service Revenue definition works for bundles and grows royalties. Services’ Reply at 57–58. VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 The Services also argue that while the Judges’ decision in SDARS III did involve valuation of the music and nonmusic components of a bundle, the resolution in SDARS III is inapposite because, here, the rate structure has a way of ensuring that Copyright Owners are fairly compensated from bundles: the statutory minimum payment. Services’ Reply at 62. C. Analysis and Decision The fundamental difference between the impact of the two alternative definitions is simply stated: Under the Initial Determination: downstream bundling and its price discriminatory effect would be incentivized by a royalty structure that reflects the lower WTP of consumers who subscribe by paying for a Bundle; Under the (Final)Determination: downstream bundling and its price discriminatory effect would not be incentivized by a royalty structure that reflects the lower WTP of consumers who subscribe by paying for a Bundle. To explain this difference, the Judges find it helpful to describe (as in the Determination and Dissent) how bundling facilitates price discrimination and how lower royalties for bundled streaming services incentivize such bundling. Price discrimination occurs when a seller offers different units of output at different prices. See, e.g., H. Varian, Intermediate Economics at 462 (8th ed. 2010). The benefit to the seller arises from attempting to ‘‘charge each customer the maximum price that the customer is willing to pay for each unit bought.’’ R. Pindyck & D. Rubinfeld, Microeconomics at 401 (8th ed. 2013). For all goods, and intellectual property goods such as copyrights in particular,224 the social benefit is that price discrimination more closely matches the quantity sold with the competitive quantity as the seller or licensor better aligns the price with the WTP of different categories of buyers or licensees. See W. Fisher, Reconstructing the Fair Use Doctrine, 101 Harv. L. Rev. 1659, 1701 (1988). A seller can engage in price discrimination in several ways. One form is known as ‘‘second-degree price discrimination,’’ by which buyers selfsort the packages and quantities they purchase.225 See W. Adams & J. Yellen, 224 Streamed copies of intellectual property, such as musical works and sound recordings, have a marginal production cost of essentially zero, making price discrimination particularly beneficial, because charging any positive price, even to a buyer with the lowest WTP, still exceeds the zero marginal production costs. See Dissent, passim. 225 ‘‘First-degree’’ price discrimination is a hypothetical construct by which a seller can PO 00000 Frm 00068 Fmt 4701 Sfmt 4700 Commodity Bundling and the Burden of Monopoly, 90 Q. J. Econ. 470, 476 (1976) (the profitability of bundling ‘‘stem[s] from its ability to sort customers into groups with different reservation price [WTP] characteristics.’’). Bundling, i.e., the ‘‘practice of selling two or more products as a package,’’ Pindyck & Rubinfeld, supra at 419, is thus a type of second-degree price discrimination. See A. Boik & H. Takahashi, Fighting Bundles: The Effects of Competition on Second Degree Price Competition, 12 a.m. Econ. J. 156, 157 (2020). The applicability of these basic economic principles was understood and explained by the parties’ experts at the hearing. See, e.g., 3/15/17 Tr. 1224– 25 (Leonard) (Google’s economic expert testifying that price discrimination through bundling is ‘‘very, very common . . . even by pretty competitively positioned firms . . . to sort out customers into willingness-topay groups.’’); 3/30/17 Tr. 3983 (Gans) (Copyright Owners’ economic expert acknowledging that bundling is a form of price discrimination); see also Dissent at 69 (same). How does this downstream (retail level) benefit of price discrimination impact the setting of upstream royalty rates? As the Majority explained (in summarizing the Services’ expert testimony) the linkage is explained by the economic concept of ‘‘derived demand’’: [M]ultiple pricing structures necessary to satisfy the WTP and the differentiated quality preferences of downstream listeners relate directly to the upstream rate structure to be established in this proceeding. Professor Marx opines that the appropriate upstream rate structure is derived from the characteristics of downstream demand. 3/20/ 17 Tr. 1967 (Marx) (rate structure upstream should be derived from need to exploit WTP of users downstream via a percentage of revenue). This upstream to downstream consonance in rate structures represents an application of the concept of ‘‘derived demand,’’ whereby the demand upstream for inputs is dependent upon the demand for the final product downstream. Id.; see P. Krugman & R. Wells, Microeconomics at 511 (2d ed. 2009) (‘‘[D]emand in a factor market is . . . derived demand . . . [t]hat is, demand for the factor is derived from the [downstream] firm’s output choice’’). Determination at 19; accord Dissent at 32 (noting that ‘‘the upstream demand of the interactive streaming services for musical works (and the sound recordings in which they are embodied)—known as ‘factors’ of identify the WTP of every buyer. ‘‘Third-degree’’ price discrimination occurs when the seller offers different prices to buyers based on their different characteristics (e.g., a senior citizen discount). See Pindyck & Rubinfeld, supra, at 402, 404–05. E:\FR\FM\10AUR2.SGM 10AUR2 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations ddrumheller on DSK120RN23PROD with RULES2 production or ‘inputs’—is derived from the downstream demand of listeners to and users of the interactive streaming services . . . This interdependency causes upstream demand to be characterized as ‘‘derived demand.’’). In the present proceeding, the PR IIbased benchmark embodies the parties’ negotiated definition of Bundled Revenue for purposes of calculating royalties on bundled interactive offerings. This is the definition in the Initial Determination. Copyright Owners’ preferred definition for Bundled Revenue—the Determination’s definition—would not only ignore this agreed-upon definition, but would also de-link the royalty rate from the WTP of purchasers of bundles.226 The Judges recognize that Copyright Owners have expressed concern the Services could use such bundling in order to diminish revenue otherwise payable on a higher royalty tier. However, the Majority noted that the evidence indicated such diminishment only occurred ‘‘in some cases’’ and that such practices were not ‘‘sweeping.’’ Clarification Order at 17, 21. Thus, the Judges find that eliminating the incentive for price discrimination via bundling would be a disproportionate response and inconsistent with the broad price discriminatory PR II-based benchmark they find useful in this proceeding. Expert testimony in this regard is ‘‘substantial evidence’’ on which the Judges can rely. For example, the D.C. Circuit also relied in Johnson on the testimony of the same witness, Spotify’s economic expert witness, Professor Marx, to affirm the inclusion of the price discriminatory structure for 226 To see the incentivizing effect of the link between the royalty level and variable WTP, consider the following example. Assume a hypothetical bundle consists of a subscription to the ‘‘Acme’’ interactive music streaming service and the sports service NFL Sunday Ticket. Assume also that Acme and NFL Sunday Ticket have standalone monthly subscription prices of $9.99/month and $149.99/month respectively, so that purchasing both separately would cost $159.98/month. But assume the bundle price is only $140/month. Acme’s purpose in bundling its interactive music streaming service subscription offering with NFL Sunday Ticket would be to attract customers who had a WTP for the standalone Acme service below $9.99/month, but a WTP at or above the $140/ month for the bundle. Under the definition in the Determination, royalties would be paid on the standalone $9.99/ month Acme price. But the purpose of the bundling was to attract subscribers who would not pay the standalone $9.99/month price, so no such wouldbe subscribers would sign-up, and no royalties would be generated by them. By contrast, under the Initial Determination, the standalone price of NFL Sunday Ticket, $159.98/ month, would be subtracted from the $140/month bundle price. Although that would preclude a payment of royalties on a revenue prong, royalties still would be paid, under a different tier or on the mechanical floor. VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 student and family plans. Johnson, 969 F.3d at 392–94. Professor Marx explained how a downstream ‘‘lower willingness (or ability) to pay’’ among some cohorts of consumers supports definitional terms, for student and family subscribers, that lower royalty rates in order to further ‘‘economic efficiency’’ in a manner that ‘‘still allows more monetization of that provision of that service.’’ Johnson at 392–93. Broadening her lens, Professor Marx also explained that this price discriminatory approach is appropriate ‘‘across all types of services and subscribers,’’ as in ‘‘[t]he current law [and in the PR II-based benchmark]’’ which ‘‘accommodates . . . adsupported services . . . and ‘bundled services’ through different rate provisions.’’ Marx WRT ¶ 41 (emphasis added). See also 3/21/17 2182–83 (Hubbard) (Amazon’s expert witness testifying that ‘‘Prime Music, which is bundled with an Amazon Prime service . . . sort[s] out customers’ willingness to pay, with an idea of trying to maximize the number of customers,’’ and agreeing that this approach constitutes ‘‘sorting by way of bundling.’’) (emphasis added). Further, Professor Hubbard opined that, given the revenue attribution ‘‘measurement problem’’ associated with bundled products, the ‘‘Phonorecords II’’ approach ‘‘with the different categories and the minima . . . address this sort of problem [in] a very good way.’’ 3/15/17 Tr. 1221 (Hubbard). As in the case of family and student price discrimination, the beneficial effect of such differential pricing was supported by industry witnesses as well as expert witnesses. See, e.g., Mirchandani WDT ¶ 71 (Amazon executive citing the Phonorecords IIbased benchmark provisions regarding bundling that ‘‘allowed Amazon to bundle Prime Music with Amazon Prime, enabling Amazon to bring a limited catalog of music [REDACTED]’’). In sum, the same type of witness testimony that the D.C. Circuit found sufficient to support price discriminatory student and family plans also supports the use of the price discriminatory bundled definition contained in the Initial Determination. Given the overall benefits from price discrimination, at first blush it is curious that Copyright Owners would risk ‘‘leaving money on the table’’ by seeking to remove the royalty-based incentive for price discrimination via bundling. The Judges have identified this problem earlier in this Initial Remand Ruling, in connection with the broader issue of the overall beneficial price discriminatory structure of the PR- PO 00000 Frm 00069 Fmt 4701 Sfmt 4700 54473 based benchmark. As the Judges noted in that general price discrimination context, Copyright Owners’ own expert economic witnesses acknowledged that they would not irrationally leave money on the table. In fact, Copyright owners’ aim, according to that testimony, is to create an unregulated space—per the Bargaining Room theory—and to use their complementary oligopoly power to negotiate price discriminatory rates (in bundles or otherwise), which would free them from the section 801(b)(1) requirements of reasonableness and fairness. The Judges further find that their prior ruling on this issue in SDARS III is distinguishable. There, a proffered bundled revenue definition eliminated the payment of any royalty at all. Copyright Owners quite correctly describe that result as ‘‘absurd,’’ but that is not the result here. Rather, in the present case, the parties’ negotiated an approach that the Judges adopted in the Initial Determination requiring royalties to be paid on interactive services bundled with other products or services. Even more distinguishable is Copyright Owners’ assertion that Web IV provides support for their preferred definition of service revenue. The argument is immediately suspect, because Web IV involved per-play royalty rates—not percent-of-revenue rates, making the definition of revenue wholly inapposite. Further, the discussion of the price of an ‘‘ice cream cone’’ in Web IV—on which Copyright Owners rely—had nothing to do with bundling or isolating the WTP for different products or services. Rather, there the Judges criticized a bizarre argument made by a licensee (who had a quantity discount for plays steered in its direction), that was tantamount to arguing that if a vendor sells one ice cream cone for $1.06 but a buyer could buy two for $1.06, that the market price of an ice cream cone is thus only $.06. This argument was indeed fallacious, because the price of an ice cream cone would be reasonably identified as the average of the total cost for the two cones, i.e., $.53/cone, and never as $.06 per cone. Here, the issue, is how to address the WTP of different classes of buyers with heterogeneous WTP, not the pricing of a quantity discount. The parties addressed this issue by utilizing the Bundled Revenue definition contained in the PR II-based benchmark (and in the Initial Determination) to address the indeterminacy inherent in the variable WTP among purchasers of the bundles, by setting floors and minima, rather than attempt to sort out the WTP of individual (or individual blocs) of E:\FR\FM\10AUR2.SGM 10AUR2 54474 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations subscribers. The ‘‘ice cream cone’’ issue in Web IV is wholly unrelated, and the SDARS III situation, as explained supra, is also distinguishable.227 For the foregoing reasons, I find that— even if the Judges had a procedural mechanism by which to support the switch in the Bundled Revenue definition—I would decline to utilize it in this Initial Remand Ruling, because the definition in the Initial Determination (unlike the definition in the Determination) is consistent with the Judges’ other substantive rulings herein. That is, just as the Majority abandoned its Bundled Revenue definition in its Initial Determination because it refused to credit the PR IIbased benchmark (even as ‘‘guidance’’), the Judges here do partially rely on the PR II-based benchmark, and thus find that it supports the Bundled Revenue definition contained in the Initial Determination. ddrumheller on DSK120RN23PROD with RULES2 VIII. Application of the Four Itemized Statutory Factors As the forgoing analysis explains, bundling is a form of price discrimination. Accordingly, the Judges’ explanation of how price discriminatory 227 The foregoing analysis also explains why Copyright Owners’ assertion that the Services did not satisfy their burden of proof with regard to the Bundled Revenue definition misses the point. The Services’ burden was to show the reasonableness of utilizing the Bundled Revenue definition in the PR II-based benchmark, not to show that their proffered approach measured the WTP of individual subscribers (or blocs of subscribers). Such an alternative approach might have had merit but no alternative approach was presented to the Judges. To be clear, the Judges are not declaring that an alternative Bundled Revenue definition and/or alternative rates and structures for bundle, might not have been preferable. See 4/15/17 Tr. 5056–58 (Katz) (‘‘[I]f someone had a proposal [with] a specific reason why we should adjust this minimum that’s something I would have examined,’’). See also 3/15/17 Tr. 1227–28 (Leonard) (Google’s economic expert testifying that ‘‘if somebody had . . . suggest[ed] . . . a different sort of bucket that should be created . . . that’s a good idea.’’). But Copyright Owners did not propose such alternatives at the hearing, and the alternative in their Motion for Clarification simply eviscerated the ‘‘derived demand’’-based link between royalties and bundled offerings. As the Judges have noted supra, in the words of Judge Patricia Wald, all judges are cabined by the record evidence introduced by the parties. Therefore (in the absence of a way in which to synthesize the parties’ proposals in a manner that does not ‘‘blindside’’ the parties) the Judges must choose between the proposals that are in the record, not potentially superior proposals that are not in the record. Here, the Judges favor the Bundled Revenue definition in the Initial Determination that was negotiated by the parties, incentivizes price discrimination and pays royalties on the bundled music, over the substituted definition in the Determination pursued by Copyright Owners that would eliminate price discrimination, except under the terms Copyright Owners could impose via their complementary oligopoly power, and without regard to the statutory requirements of a ‘‘reasonable rate’’ and a ‘‘fair income’’ for the Services. VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 rates in the PR II-based benchmark interrelate with the Factor A through D objectives in section 801(b)(1) are equally applicable here. Accordingly, the Judges incorporate by reference here their discussion of those four factors set forth supra in connection with the PR II-based benchmark, and find that there is no basis pursuant to those four factors to adjust the PR II-based benchmark definition of Bundled Revenue. IX. Conclusion This Dissent in part is issued as a RESTRICTED document. Within 30 days of the date of issuance, the participants shall file a version of this Dissent with agreed redactions to permit viewing by the public. 2022.228 Issue Date: July 2, DAVID R. STRICKLER, Copyright Royalty Judge D. Dissent in Part Re Benchmark (Redacted Version With Federal Register Naming and Formatting Conventions) The Copyright Royalty Judges (Judges) sit as a panel in all determination proceedings. See 17 U.S.C. 803(a)(2). A majority of two Judges is sufficient to issue a determination. See 17 U.S.C. 803(a)(3). If any Judge dissents from the majority determination, that dissenting Judge may issue a dissenting opinion and file it with the majority’s determination. Id. The Judges accept this same standard with regard to their issuance of the present Initial Ruling and Order after Remand (Initial Ruling). The undersigned Judge, author of this dissent in part (Benchmark Dissent) respectfully dissents 229 from the Initial Ruling of the majority (Remand 228 Technical difficulties on July 1 caused the delay in filing of this Dissent until July 2. 229 The dissenting Judge does not fault the economic analysis of the Remand Majority on this issue. The dissenting Judge is not the Judge selected for ‘‘a significant knowledge of economics.’’ See 17 U.S.C. 802(a)(1). This Benchmark Dissent is based on a broader reading of the requirements of section 801 of the Copyright Act, viz. ‘‘to make determinations of reasonable terms and rates. . .’’ consistent, of course, with the record evidence and sound legal and economic analysis. The role of the Judge is to weigh evidence; two Judges might rightfully and respectfully disagree on where that scale balances. The Remand Majority’s analysis led those Judges to conclude that they were bound to re-introduce the rate structure devised in the Phonorecords II proceeding. The Benchmark Dissent concludes that the economic analysis outlined in the Initial Ruling supports, but does not dictate, that result, but that the goal of reasonableness can be met with different structure(s). The Benchmark Dissent does not construct or propose a detailed, different structure. To do so would be an inefficient application of judicial resources at this late stage of this proceeding. The Benchmark Dissent finds, however, that both licensor and licensee participants agreed in this proceeding that a less complex rate structure is warranted. PO 00000 Frm 00070 Fmt 4701 Sfmt 4700 Majority) on the issue of adopting as a benchmark for current rates and terms the rates and terms adopted after a settlement by the parties to the preceding phonorecords proceeding.230 It should be noted that the Remand Majority adopts the rate structure from Phonorecords II, but retains the headline percent-of-revenue rate adopted in the Determination.231 I. Areas of Concurrence A. Background Statements The Benchmark Dissent adopts the statements regarding the background and procedural posture of this remand proceeding. See Initial Ruling at 1–2. B. Percent of Revenue Rate The Benchmark Dissent agrees with the Remand Majority’s retention of the headline percent-of-revenue rate and its phase-in over the period at issue. C. Definition of Service Revenue for Bundled Offerings For the reasons articulated in the Initial Ruling and the reasoning of the judge dissenting from that portion of the Initial Ruling, the definition of Service Revenue for bundled offerings contained in the Initial Determination must be adopted. See Initial Determination (Jan. 27, 2018). Adoption of the Phonorecords II (PR II) rate structure requires that the original definition pertain. II. Area of Dissent The first function of the Judges is ‘‘to make determinations . . . of reasonable terms and rates of royalty payments. . . .’’ 17 U.S.C. 801(b)(1). Under the statute in effect during the captioned proceeding, the rates shall be calculated to achieve four statutory objectives. Id. The terms of payment of the rates, however, are not subject to any particular statutory restrictions or guidelines. See, e.g., Live365 v. Copyright Royalty Bd., 698 F. Supp. 2d 25, 29–30 (D.D.C. 2010) (‘‘In performing their duties, the [Judges have] broad discretion to . . . impose regulations 230 The preceding proceeding, referred to as Phonorecords II, consisted of a final rule adopting the participants’ settlement agreement as regulatory terms and rates. See Final Rule, Adjustment of Determination of Compulsory License Rates for Mechanical and Digital Phonorecords, Docket No. 2011–3 CRB Phonorecords II, 78 FR 67938 (Nov. 13, 2013), Technical Amendment at 78 FR 76987 (Dec. 20, 2013). In this partial dissent, references to Phonorecords II, PR II, and PR II-based benchmark are references to this final rule. 231 Determination of Royalty Rates and Terms for Making and Distributing Phonorecords (Phonorecords III), 84 FR 1918 (Copyright Royalty Board Feb. 5, 2019) (final rule and order) (Determination); See also Final Determination, 16– CRB–0003–PR (2018–2022) (Nov. 5, 2018). E:\FR\FM\10AUR2.SGM 10AUR2 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations governing the rates and terms of copyright royalties. . . .’’).232 In general, in promulgating regulations the Judges aim to effect efficient and effective payment of royalty license fees. Regulations relating to license royalty rates describe the rates the Judges determine to be reasonable, whether presented by agreement of the affected parties or after adjudication. The regulations include, where necessary, methods of calculation of the payable royalties. The regulations also include such provisions as recordkeeping requirements, late fee assessments, and audit authority. As the Remand Majority points out, simplicity and clarity were not among the statutory factors applicable to determining royalty rates in the captioned underlying proceeding. Simplicity and clarity should, however, be paramount among the Judges’ considerations in governing rate payment procedures. In recent proceedings, the Judges have emphasized that the statute requires that they set both rates and terms. At the end of a different royalty rate proceeding, having been confronted with competing proposed regulations, or even with largely agreed regulatory terms, upon which the parties had proffered no evidence, the Judges cautioned counsel in this proceeding: Please be reminded that the Judges have an obligation to set both rates and terms. . . . In any proceeding, just because a regulation is in the current Code of Federal Regulations does not mean that the Judges are adopting that term. . . . The Judges cannot determine rates or terms without an evidentiary record. . . . The Judges cannot adopt any terms of royalty administration unless the parties present evidence to support their proposed terms. ddrumheller on DSK120RN23PROD with RULES2 Tr. 03/08/2017 (Barnett, J.) While chapter 8 of the Copyright Act encourages settlement, the Judges are not mandated to adopt parties’ settlements if they find they face opposition that discounts reasonableness or if the proposed regulations are contrary to law. See, e.g., Determination of Royalty Rates and Terms . . . (Phonorecords IV), 87 FR 18342, 18347, 18349 (Mar. 30, 2022). In the proceeding underlying the Determination, the parties proffered a variety of proposed regulations.233 Copyright Owners contended that the 232 The Judges’ regulations are, of course, subject to approval by the Librarian of Congress. 17 U.S.C. 802(f)(A)(i); see Live365 v. Copyright Royalty Bd., 698 F. Supp. 2d 25, 29–30 (D.D.C. 2010). 233 Spotify, as the only pure-play service, offered simplified regulations, but only because it did not propose any rates or terms for bundled or locker services. Spotify advocated elimination of the persubscriber stop-gap alternative in the greater-of percent-of-revenue/percent-of-TCC calculation. VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 extant rate structure ‘‘should be modified and simplified.’’ Copyright Owners’ Amended Proposed Rates and Terms (5/17/2017) at 2. Copyright Owners argued that the ten different rate categories should be ‘‘no longer applicable’’ as Copyright Owners proposed application of the same rates and rate structure to ‘‘all interactive streams and/or limited downloads [except bundles], regardless of the business model employed.’’ Id. at 3. Copyright Owners’ rate proposal hinged on a per-unit calculation across the board: the greater of a per-play amount or a per end user amount. Amazon proposed retaining the PR II rate structure. See Proposed Findings . . . of Amazon (May 13, 2017) ¶ AM– F–25. Amazon argued that the PR II rate structure ‘‘enabled Amazon to develop a varied assortment of services. . . .’’ Id. Amazon contended that the different royalty rates permit price discrimination by the Services. Id. ¶¶ AM–F–47, 49. Amazon conflates price discrimination with provision of heterogenous musical tastes and preferences. Id. ¶ AM–F–48. Amazon’s proposal mimicked the regulations adopted by agreement in the immediately prior proceeding. Apple proposed a per-play rate calculation, which would render the PR II rates and rate structure obsolete. Notwithstanding the different structure, however, Apple offered valid criticisms of the PR II rate structure. Apple termed the PR II rate structure ‘‘problematic.’’ See Apple Inc.’s Findings . . . and Conclusions . . . (May 11, 2017) at 30. Apple argued that the PR II rate structure was ‘‘overly complex, economically unsound, and unpredictable.’’ Id. ¶ APL–F65. Apple acknowledged that these shortcomings resulted in ‘‘a loss of trust and overall dissatisfaction with interactive streaming among songwriters. . . .’’ Id. Apple noted that across the ten rate categories in the PR II rates, ‘‘there are roughly 79 different calculations that can be made.’’ Id. ¶ APL–F67. Apple argued that the PR II rate structure was ‘‘not transparent or easy to understand’’ for copyright owners and created ‘‘uncertainty for services, who may find it difficult to predict which prong . . . will kick in in any given month.’’ Id. ¶¶ 68–69. Apple opined that, rather than encouraging new business models, the PR II rate structure ‘‘tends to stifle innovation around new pricing or distribution models, as services are incentivized to create businesses that fit into the ten pre-defined ‘boxes.’ ’’ Id. ¶ 70. Apple further argued that the PR II rates were economically unsound because they are based on revenue, PO 00000 Frm 00071 Fmt 4701 Sfmt 4700 54475 which is unrelated to demand for a given copyright owner’s song. Id. ¶ 71. Google’s proposal, from which the Majority derived the uncapped TCC rate prong of the Determination, contended that the ‘‘fragmented service categories are unnecessary under [its] proposal. . . .’’ Google, Inc.’s Proposed Findings . . . and Conclusions. . . (May 11, 2017) ¶ GPFF58. Google acknowledged questions regarding the complexity of the PR II rate structure. Google, therefore proposed a rate structure that would both streamline the regulations and protect Copyright Owners’ concerns regarding Services’ revenue deferment and displacement. Id. ¶ GPFF57. In the captioned underlying proceeding, the Judges heard little evidence offered in resounding support or vehement objection to the regulations the parties proffered. No party argued or supported the proposition that the PR II rate structure was the only way, or even the best way, to achieve license fee payment.234 In this remand proceeding, no party argued against the all-in approach to rate calculation. The parties disagreed regarding retention of ‘‘mechanical floors’’ for configurations for which the Services must pay mechanical royalties both to Copyright Owners in this proceeding under section 115 and to Performing Rights Organizations (PROs) according to the determinations of the ‘‘Rate Court.’’ 235 The parties disagreed over imposition of a cap on the TCC prong 236 in the greater-of percent-ofrevenue calculation. They also disagreed over retention or elimination of the per subscriber sub-minima that were featured in the PR II rates. The Remand Majority cites with approval the remand parties’ criticism of the simplified rate structure in the Determination, viz., that it is ‘‘virtually as complex as’’ the PR II rate structure. See Services’ Joint Opening Brief (Apr. 1, 2021) at 39. This characterization is 234 The Benchmark Dissent does not argue that the PR II rate structure did not achieve its purpose. Indeed, the all-in, greater-of, lesser-of scheme with payment minima and mechanical floors achieved the goals of (1) supporting increased absolute revenue through downstream price discrimination and (2) protecting creators from potential loss resulting from licensees’ revenue deferral or displacement. The Judges have never denied the value of price discrimination in these or other rate setting proceedings. 235 The District Court of the Southern District of New York determines performing rights royalties. Parties to those rate proceedings refer to that court, when engaged in the rate-setting cases, as the ‘‘Rate Court.’’ 236 ‘‘TCC’’ refers to a streaming services’ costs of content, referring in this proceeding to the cost of sound recording royalties the streaming services pay to record companies. E:\FR\FM\10AUR2.SGM 10AUR2 54476 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations a bit of hyperbole. The rate structure in the Determination is an all-in rate with ‘‘mechanical floors’’ where those are warranted. Except for the fundamentally different configurations included in subpart B, it does not set out separate calculations for different delivery configurations. On remand, the Remand Majority chooses to reinstate the PR II rate structure in its entirety, with all of its 79 permutations, changing only the headline percent-of-revenue rate and adding a cap on the TCC rate prong (which is an element of the structure itself). The Benchmark Dissent does not dispute the necessity and propriety of the increased headline percent-ofrevenue rate or the cap on the TCC rate prong. Indeed, as noted in the Remand Majority, the D.C. Circuit endorsed the rate increase as well-reasoned and determined well within the Judges’ discretion. The D.C. Circuit also found fault with ‘‘yoking’’ the TCC rate alternative to sound recording royalty rates, not subject to the Judges’ control, without reins. The basis of this Benchmark Dissent is simply that the regulatory scheme is not efficient, transparent, or mandated by credible evidence; nor is the structure necessary to achieve the purposes of reasonableness and equity.237 ddrumheller on DSK120RN23PROD with RULES2 A. Acceptance of Phonorecords II Settlement as a Proper Benchmark This is a Dissent in Part. The undersigned Judge does not disagree with the headline rate being retained at 15.1% or with the imposition of a TCC cap, for the reasons elucidated by the Remand Majority. Nonetheless, the Benchmark Dissent continues to disagree with adoption of the entirety of the rate structure adopted by Phonorecords II. As noted above, the Judges solicited evidence to support adoption of regulatory language to effect payment of the rates they established. Copyright Owners, Google, and Apple submitted rate proposals that greatly simplified the rate structure. Their rate structure regulation proposals were crafted to support their varying approaches to rate calculations not adopted by the Judges. Their criticisms of the PR II rate structure are valid, nonetheless, and support the Benchmark Dissent’s analysis. 237 As part of the Judges’ discretion to promulgate regulations to effect license rate collection, the Majority reorganized the regulations in part 385. This reorganization was completed to further the goal of clarity and conciseness. No party objected to or sought to overturn that reorganization of the regulations. Apparently, the perceived sanctity of the PR II rate structure is not unassailable. Reorganization can perhaps be seen as a first step to toward clarity, transparency, and simplicity for licensors and licensees. VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 In the underlying proceeding, the Majority declined to label the rate structure and resulting rates incorporated in the regulations promulgated after the Phonorecords II proceeding (rates and rate structure) as a benchmark, or starting point, for determination of new rates and terms in that proceeding. In the Determination in the extant proceeding, the Majority alluded to reasons they found the PR II rates to be inadequate to serve current circumstances.238 The D.C. Circuit noted that appellate counsel offered further explanation on appeal for the rejection of the PR II rates and rate structure as a benchmark. See Johnson v. Copyright Royalty Board, 969 F.3d 363, 387 (D.C. Cir. 2020). Nevertheless, the D.C. Circuit faulted the Majority for not providing adequate explanation of their rejection of a PR II-based benchmark in the first instance. See id. Indeed, the D.C. Circuit found the Majority’s reasoning on the issue in the Determination to be ‘‘muddled.’’ Id. at 386–87. Copyright Owners argue that the D.C. Circuit’s remand for further explanation did not equate to finding error in the Judges’ rejection of the PR II-based benchmark. See Initial Remand Submission of Copyright Owners (Apr. 1, 2021) 1, 10 (CO Initial Submission). Notably, the Services did not address the question of a finding of error, but proposed on remand a rate structure substantially similar to that in PR II and offered a benchmark analysis therefor. See Services’ Joint Opening Brief (in Services’ Joint Written Direct Remand Submission at Tab D) (Apr. 1, 2021) at 19 (Services’ Initial Submission). While the Copyright Owners’ parsing of Johnson might be technically correct, the Benchmark Dissent nonetheless accepts the wisdom of revisiting the analysis of the PR II rates and rate structure, focusing on the intricacies of the structure that ultimately come into play in determining the amount of royalty payable. The Benchmark Dissent disagrees that the record in this case demands adoption of the PR II rate structure as a suitable benchmark. The Benchmark Dissent hereby provides a full analysis of this issue, which includes a fuller explanation of the conclusions in the Determination and 238 The D.C. Circuit found that the Majority articulated a reasoned and reasonable rejection of the negotiated rates applicable to the categories of phonorecords included in ‘‘Subpart A’’ of the regulations as a benchmark in this proceeding. The issue on remand is articulation of a reason for not using the other subparts of 37 CFR 385 as a benchmark in this proceeding. See Johnson v. Copyright Royalty Board, 969 F.3d 363, 386 (D.C. Cir. 2020). PO 00000 Frm 00072 Fmt 4701 Sfmt 4700 supports and justifies rejection of the Phonorecords II rate structure. B. Attributes of a Useful Benchmark As repeated by the parties in the initial proceeding and in their remand submissions, for an exemplar to serve as a useful benchmark, it must be compared to the target market. The hallmarks of a useful benchmark are: (1) unity of products, (2) unity of sellers, and (3) unity of buyers. In addition, (4) economic circumstances and market conditions can influence the value of a benchmark. See Services’ Initial Submission at 20 (citing Determination of Royalt[ies] for Transmission of Sound Recordings. . ., 83 FR 65210, 65214 (Dec. 19, 2018) (SDARS III). In the Remand Majority opinion, the Judges argue that the PR II rate structure meets ‘‘most of the requisites for a useful benchmark.’’ See Initial Ruling, section III. C. 3. Assuredly, in the real world one is unlikely to find a perfect benchmark; consequently, the Judges in these proceedings look to the best available benchmark(s) and make adjustments to compensate for their shortcomings when compared to the attributes and circumstances of the target rates. The Benchmark Dissent is not so sanguine about one’s ability to reconcile the PR II rate structure with current market circumstances pertaining to music streaming (including participants and volumes of sales) almost a decade after the parties agreed to that structure. Because of the recognized gulf in market conditions between Phonorecords II and this Phonorecords III proceeding, the Benchmark Dissent rejects attempts to fit that square peg into the current round hole. 1. Unity of Products—the Same Rights The PR II rates regulated ‘‘sales’’ of the same licensing rights as those at issue in the current underlying proceeding, viz., the statutory license to utilize musical works embodied in the sound recordings that are the lifeblood of the music streaming services. This factor was not and is not in controversy. In this respect, the Judges could look to the PR II rates as a benchmark. 2. Unity of Sellers—Rightsholders The songwriter or songwriters own the copyright for musical works, that is, the musical notes and lyrics. In general, songwriters sell or license their works to publishers who fix the works to a physical medium, for example, piano rolls or sheet music. Music publishers also market the musical works licenses to record companies for their sound recordings. In today’s market, E:\FR\FM\10AUR2.SGM 10AUR2 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations ddrumheller on DSK120RN23PROD with RULES2 publishers and songwriters exist in a symbiotic relationship. Without new works, the publishers have no new product to market.239 To ensure a flow of new product, publishers often subsidize songwriters by providing working space or monetary advances on future sales of licensed work, or publishers might purchase outright the songwriters’ copyrights. Whether the rightsholder is a writer, composer, or publisher, the rights are the same, those derived from 17 U.S.C. 106 and limited by 17 U.S.C. 115. See 17 U.S.C. 106(1), (3) (exclusive rights); sec. 115 (compulsory licensing). The sellers’ interests are aligned. 3. Unity of Buyers—Streaming Services The Services argue unity of rights and sellers between the time of the PR II rates and the current proceeding. With respect to buyers, the Services allege that the current buyers are ‘‘the same or similar. . . .’’ Services’ Initial Submission at 20. The Services argue that the PR II rates involved ‘‘either the same type of buyers or the very same buyers as this proceeding.’’ Id. The license delimits the users it binds. It is axiomatic that current licensees are ‘‘of the same type’’ as licensees in 2012. Describing participants as ‘‘similar to those currently in the market’’ or ‘‘of the same type’’ as current participants is sufficiently imprecise to call into question the unity of buyers required to give great weight to a potential benchmark. The Services allege that ‘‘[m]ost of the participants in Phonorecords III were either directly involved in the Phonorecords II settlement or operated in the market at the time of the settlement.’’ Id. ‘‘Most of the participants’’ does not reveal which participants were active in Phonorecords II or the reasons for their participation. Amazon began an MP3 digital music service in 2004; it launched steaming in mid-2014. See Written Direct Testimony of Jeffrey Eisenach (Nov. 3, 2016) (Eisenach WDT) ¶ 51. Tab. 2. Apple launched its streaming service in 2019. During the Phonorecords II negotiations, Apple’s primary interest was digital downloads from the iTunes store. According to one of its witnesses, Google was, at the time of the Phonorecords II negotiations, ‘‘planning to launch a store, a locker, 239 Publishers may retain rights to songs no longer considered ‘‘new’’ or ‘‘popular’’ that might nonetheless still be subject to the section 115 license. The Services’ revenue is driven, however, by streaming new music. They understand that reselling older music, even in new packaging (covers) would lower their desirability and decrease the sources of revenue, their end users. VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 and a subscription service.’’ Google’s participation in the Phonorecords II negotiations was ‘‘primarily designed to make sure that our interests were met in—for our forthcoming music service.’’ 3/8/17 Tr. 157:2–158:2 (Zahavah Levine). Although the Services argue that the buyers in the current market are the same as, or similar to, buyers at the time of adoption of the PR II rates, the Services then and now advocate differing rate calculations for each music delivery configuration. Indeed, between 2008 and 2012, the delivery configurations multiplied and the parties negotiated different rate structures for those multiple configurations. Acknowledging participation by a service with one configuration—or a plan to launch one configuration—is insufficient to establish a unity of buyers for purposes of rate setting. Almost a decade after the effectuation of the 2012 rates, with new businesses tacking music streaming onto their digital ecosystems, the development of new and different delivery configurations continues to evolve.240 Nonetheless, the Services would have the Judges adopt a rate structure that specifies current delivery configurations but excludes some current innovations and cannot encompass the next innovations, whatever form they might take. The Benchmark Dissent acknowledges that buyers of the musical works for which licenses are at issue in this proceeding are of the ‘‘same type’’ as the Phonorecords II buyers. In some instances, they are the same participants. In the current landscape, however, the interests of those buyers are vastly different. The extent to which Apple, Amazon, and Google, were involved in Phonorecords II negotiations bears no resemblance to the interests of those services and their current service configurations. Without greater unity of buyers, the Benchmark Dissent must discount the viability of the PR II rates or rate structure as a useful benchmark in this proceeding. 4. Economic and Market Conditions The Services argue that the music streaming industry in 2018 was essentially unchanged from 2008 or 2012.241 See Services’ Initial Submission 240 Some services offer different levels of access to consumers using their proprietary devices, e.g., Amazon Echo. Some (non-satellite) music streaming services are now available directly via a button on a vehicle dashboard. 241 The PR II rates and rate structure were the product of a negotiated settlement that began and ended with reference to the negotiated rates adopted in 2008. Some additional categories of PO 00000 Frm 00073 Fmt 4701 Sfmt 4700 54477 at 20–21. The evidence in this proceeding compels a contrary conclusion. In 2008, musical works distribution consisted primarily of sound recordings reproduced in physical formats (vinyl and CDs) and digital downloads. See Eisenacht WRT ¶ 33 (Feb. 13, 2017). The record reflects that in 2008, of record labels’ revenues 96% were derived from sales of physical and digitally downloaded sound recordings; 2.5% from interactive streaming.242 By 2012, at the inception of the rates that were re-adopted as the PR II rates, musical works sales were beginning to shift from physical media to digital forms. In 2012, 8.1% of record label revenues were attributable to interactive streaming. Id. By 2015, evidence available in this proceeding showed that record labels’ revenues from digital downloads approximately equaled revenues from streaming and digital sales were more than double the sales of physical configurations, such as vinyl and CDs. Id. ¶¶ 44–45 and accompanying tables. Spotify, the dominant pure play streaming service in the U.S., did not enter the U.S. market until mid-2011. See CO Initial Submission at 20–21 (Apr. 1, 2021) and evidence cited therein. Spotify did not participate in the negotiations leading up to the adoption of the 2012 musical works royalty rates. See Eisenacht WRT ¶ 35, n.38. In fact, the record contains evidence that music streaming was not a major factor in setting mechanical license rates in 2008 or 2012.243 See CO Initial Submission at 19–21, and evidence cited therein. As more and larger streaming services entered the market, music consumption changed in character. Music consumption in the 2018 market had changed character completely from an ownership model to an access model. See Determination at 6. Further, three of the Services participating in the current proceeding are not pure play streaming services but are multidimensional marketing firms for whom music streaming is only one small facet of the business. From the perspective of those current licensees, the music streaming license is relatively insignificant to their overall financial service were added to the 2008 structure, e.g. locker services. Of those categories added in 2012, few remain a significant part of the current streaming industry. 242 The difference is attributable to sound recording revenues from non-interactive streaming. 243 The Services argue that only Mr. Israelite testified that the 2008 and 2012 rates were ‘‘experimental’’ and that the market is significantly changed since 2012. The Majority found, based upon the totality of the evidence, that Mr. Israelite’s testimony was credible and accorded it due weight. E:\FR\FM\10AUR2.SGM 10AUR2 54478 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations health. The Judges must, therefore, value the license objectively to assure the conglomerate licensees do not manipulate their revenues so as to reduce music streaming rights below what is fair and reasonable to the rightsholders. The Services further advocate use of the PR II rates and rate structure as a benchmark because they assert that the multifaceted rate structure is reflective of the Services’ own price discriminatory services. The Majority noted the Services’ price discrimination as a way to optimally monetize segments of the market with a lesser willingness to pay.244 Greater accommodation of users less willing to pay results in more streaming and more revenue for the Services at minimal to no marginal cost. A rate determined as a percentage of a service’s revenue allows that price discrimination to continue, resulting in additional royalties. The Benchmark Dissent contends, however, that the Judges need not adopt a rate structure with ten different service categories to allow the Services to continue their price discriminatory downstream sales. The payable royalties are a percent of revenue. If the Services receive relatively less revenue by marketing a family plan, for instance, that reduced revenue is the basis for the royalty calculation. Nothing in a simplified rate structure would inhibit price discriminatory service plans. The PR II rates’ multi-category structure might encompass the price discrimination the Services employ, but that does not make it a mandatory benchmark for current rates, especially if the target rate structure permits the same flexibility. C. Adoption of PR II Rates and Rate Structure in Direct Licenses ddrumheller on DSK120RN23PROD with RULES2 The Services assert that the PR II rates and rate structure have been adopted in negotiated direct licenses they have signed with rightsholders rendering those rates and that rate structure a valuable benchmark. The Services’ witnesses analyzed direct licenses and concluded that the rates closely matched the rates in the PR II regulations. [REDACTED].245 Analysis of direct licenses executed belie the Services’ assertion that the PR II rates 244 The adopted Phonorecords III rate regulations acknowledged price discrimination by, inter alia, permitting Services to account for discounted subscriptions in different ways. See Determination at 34. 245 The [REDACTED] direct licenses reportedly adopt the rates in part 385, which open-ended adoption could indicate acceptance of both rates and rate structure. VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 a. Significance of the Passage of Time structure is embraced by rightsholders.246 D. Additional Shortcomings of PR II Rates as a Benchmark The D.C. Circuit dismissed the Majority’s argument on appeal that (1) the PR II rates were too low and (2) the PR II rates were outdated. The D.C. Circuit noted that these two reasons might support the Majority’s conclusions, but they could not be asserted in the first instance on appeal. See Johnson at 386. 1. Rates Too Low The D.C. Circuit found that the Judges’ finding that the PR II rates were too low was not fully articulated until the matter was on appeal. As a result, the D.C. Circuit could not evaluate that reason as support for the final rates. Indirectly, however, the D.C. Circuit nonetheless accepted that underlying reason for the rate changes when it approved the higher rates themselves. See Johnson at 384–86. The adopted rates were soundly grounded in the record evidence. See id. By implication, acceptance of increased rates means the PR II rates were too low to be continued. With or without the ‘‘too low’’ rationale, the final adopted rates prove the point.247 2. Rate Structure Outdated In the Determination, the Majority cited several factors that implied the inadequacy of the PR II rates and rate structure as a compelling benchmark for Phonorecords III. As discussed above, the music streaming industry in 2018 was completely transformed from 2008 or 2012. Both the buyers and the economic market conditions were markedly changed. Referring to the PR II rates as ‘‘outdated’’ encompasses both a temporal element and a structural component. 246 [REDACTED] See AWDT Leonard ¶¶ 63–64. [REDACTED]. See Leonard AWDT ¶ 70–71. [REDACTED]. See AWDT Leonard ¶ 54. (calculation is ‘‘effectively simplified’’). [REDACTED]. [REDACTED]. 247 The Services argue that an agreed continuation of the Subpart A (now Subpart B) rates for, inter alia, physical phonorecords and permanent downloads, proves that the Phonorecords II rates are appropriate. See Services’ Initial Submission at 30. This argument asserts a false equivalency. Physical Phonorecords and permanent downloads are fundamentally different in character from streamed music. Further, the evidence indicates that the prominence of streaming access over ownership of recordings is waning. The parties’ agreement to maintain the Phonorecords II rates for this declining segment of the market does not equate to a mandate to adopt the entirety of the PR II rate structure. PO 00000 Frm 00074 Fmt 4701 Sfmt 4700 Music streaming in the earlier rate setting periods was in its infancy. Listeners had not yet fully embraced the subscribed access model for music consumption. By 2018, listeners could choose from ‘‘a diverse array of streaming offerings.’’ See WDT of Rishi Mirchandani ¶ 63. Such industry shifts alone could render the PR II rates ‘‘outdated.’’ b. Clarity and Simplicity Another salient factor the Majority addressed is the rate structure itself. To understand the PR II rate structure, one needed ten separate full-page flow chart diagrams, each featuring three formulae for calculating greater-of and lesser-of rate components. See Trial Ex. 846. The rates for some consumption configurations included a per-subscriber ‘‘mechanical floor’’ as a failsafe against overreaching by PROs, should the Rate Court increase their rates to an extent that all of the section 115 all-in percent of revenue royalty be consumed by the PROs. See, e.g., [FORMER] 37 CFR 385.13(a)(1) (Standalone non-portable subscription—streaming only [$.15 per subscriber]); [FORMER] 385.13(a)(2) (Standalone portable subscription— mixed use [$.50 per subscriber]) (2018).248 Other consumption configurations included ‘‘minima;’’ that is a lesser-of calculation comparing a percent of sound recording license costs (TCC) and a per subscriber amount. See, e.g., [FORMER] 37 CFR 385.13(b) (2018). Further, rate calculations differed depending upon, for example, whether the listener streamed on a portable device or a non-portable device; or whether the listener purchased access to the music alone from a pure-play streaming service or as part of a bundled offering, such as ‘‘free’’ streaming for a limited period included in the purchase price of the streaming device.249 The rationale for these convoluted rate calculation differences is 248 The Majority reintroduced these ‘‘mechanical floor’’ safeguards, notwithstanding a lack of evidence to explain, let alone justify, the difference between $0.15 and $0.50 per subscriber (the latter being 300% greater than the former) simply because one consumer listened to a song on a standalone non-portable device and another consumer listened to a song on a standalone portable device. 249 The Services have not offered convincing, substantive evidence or argument to support the fractured structure of the PR II rates. Tellingly, the user’s choice of consumption device is not a factor in license rates for other services. See, e.g., 17 CFR 380.10 (Webcasters rates differentiate between commercial and non-commercial licensees, not based on users’ reception devices); §§ 382.3, 382.12 (rates for satellite radio and pre-existing subscription services do not differentiate based on users’ reception devices). E:\FR\FM\10AUR2.SGM 10AUR2 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations unknown.250 They were the product of confidential negotiations among the parties involved in the music streaming business in the first decade of the 21st century. One side of the negotiating table sought reconsideration of those rates. The current licensees are not the same as those who negotiated the 2012 rollover of the 2008 rate scheme. Music streaming business models have witnessed significant growth and change. Meanwhile, the business models employed by songwriters and publishers remain largely unchanged— and not realizing a proportionate capture of the stream of dollars realized by the Services’ monetization of evermore consumption configurations. The marginal cost to the Services of additional streams, regardless of the business configuration or the user’s reception device, is zero. The Services, therefore, are in a position to capture increased revenue without an increase in cost of goods sold. In the end, a sound recording embodying a licensed musical work is being delivered to an end ‘‘user’’: one song; one listener. The calculation of what royalty the songwriter is entitled to should not rest on the medium of transmission or the location of the listening. See WDS Steve Bogard ¶ 34 (‘‘Streaming music anytime, any place, on any device is the way today’s music fans want to enjoy their music. Notwithstanding that the inherent value of a song is the same whether the consumer chooses to buy an album, permanently download an album or a single, or stream music on demand. . . .’’). The incremental difference in value to the listener of hearing a song in the car as opposed to through earbuds during a workout is not likely measurable. Certainly, no participant in this proceeding presented any evidence of the relative value of a song to a listener depending on the delivery configuration.251 ddrumheller on DSK120RN23PROD with RULES2 250 Prof. Katz asserted that ‘‘economic analysis’’ indicates that varying rates based on the characteristics of the service ‘‘facilitates continuing innovation, experimentation, and differentiation in means of making music accessible to consumers.’’ Katz WDT ¶ 85. Prof. Katz did not identify that economic analysis. He asserted that the fractured rates allow services to benefit despite different consumers’ willingness to pay. Nothing in the PR III rate structure at issue in any way inhibited services adapting to meet consumers’ willingness to pay. The rates are, in the main, revenue based— even if the services choose to market the service at a lower rate to a particular segment of the market. 251 The Remand Majority dubs analysis of value based on the cost of production rather than willingness to purchase as old-fashioned economic analysis. So it may be. In the modern economist’s widget market, if buyers are unwilling to pay enough to cover the cost of widget components, then widget production ceases. But in the oldfashioned creativity market, the goods are not VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 In the interest of making government more transparent and accessible to interested citizens, less is more. Opaque systems and formulae are or should be, in a word, outdated. The fact of settlement does not cure or even address the unnecessary complication of paying a royalty for the use of a statutory license under the PR II rates structure. More importantly, owners of the copyrights being licensed should be able to comprehend, calculate, and verify the sources and amounts of their royalty payments. 3. Not Business Model Neutral The Services contend that the PR II rate structure is preferable as it is business model neutral. Nothing in the record supports that assertion. In fact, Apple argued that the PR II rate structure stifled innovation as streaming services sought to fit any new business into a business model already defined as one of the ten identified models in the Phonorecords II regulations. The statute does not require that rate structures be business model neutral. The reasonableness requirement demands, however, that the Judges find and adopt reasons for differentiation in rates based on business models. 4. No Evidence of Settling Parties’ Subjective Intent Copyright Owners participating in the current proceeding argued that the Judges should consider the subjective intent of the parties in agreeing to ‘‘roll over’’ the 2008 rates and rate structure into the PR II regulations. The Services countered that subjective intent is irrelevant, as the product of those negotiations serves as objective evidence of the parties’ intents. On this question, the Services are correct. The negotiated rates show, objectively, that the negotiating parties agreed to a certain rate structure. The D.C. Circuit fungible. The inputs to a hit song are ephemeral; sometimes plentiful, sometimes elusive; they either coalesce or they do not. Songwriters will persevere because they cannot do otherwise. The demand for music continues to grow with each new innovation in delivery methods. The United States Constitution provides for protection of art and the creators of art. U.S. Const. art. I, sec. 8. Congress has specified how to protect, inter alia, the copyrights of songwriters. The Judges’ small part in that effort is to continue to assure that royalty rates are reasonable—for both creators and exploiters. In the music streaming industry, the evidence supports devoting a greater share of licensees’ increased wealth to the ‘‘widget makers.’’ The Dissent contends that the increase in the percent-of-revenue headline rate is a good step forward, but only the first step to assuring equity in the market. Streamlining, simplifying, and generally ‘‘cleaning up’’ payment calculations would go a long way in the right direction by removing twists and turns and confusing signals along the path of the royalty dollar from end user to creator. PO 00000 Frm 00075 Fmt 4701 Sfmt 4700 54479 criticized the Majority for not including in the Final Determination an explanation of why the subjective intent of the parties to the settlement was a ‘‘prerequisite’’ to adoption of that settlement as a benchmark. See Johnson at 387. The Judges need not, however, accept that objective evidence uncritically. Negotiating parties’ subjective state of mind can serve as convincing evidence of the economic circumstances and the state of the market at the time of the negotiations. While ascertaining the parties’ subjective intent in reaching the settlement is not a ‘‘prerequisite’’ to examination of the terms as a benchmark, the Benchmark Dissent finds subjective intent informative and useful as one factor in weighing the value of the settlement as a benchmark. E. Statutory Factors The Services argued to the D.C. Circuit that the Majority’s rejection of the PR II rates and rate structure was erroneous because the Majority failed to evaluate that structure and those rates under the statutory factors delineated in 17 U.S.C. 801(b)(1). Evaluation under section 801(b)(1) is required by the statute applicable to this proceeding.252 Nothing in section 801(b)(1) compels the Judges to evaluate compliance with the statutory factors of every proposed potential rate or rate structure. Neither are the Judges required to evaluate every potential benchmark or past rate structure under section 801(b)(1). The Judges are obliged to evaluate any rate structure they intend to adopt against the requirements of section 801(b)(1). If the Judges’ promulgated rate structure meets the section 801(b)(1) standard, then the promulgated rate structure can be adopted. Whether other possible proposals might also meet the section 801(b)(1) standard is not at issue in a proceeding. 1. Maximize the Availability of Creative Works to the Public The Services argue that the PR II rates and rate structure support and contribute to the maximization of musical works. As evidence, they cite the growth of music streaming overall, the profitability of all segments of the music industry.253 It is beyond question that music consumption has grown exponentially since the co-incident 252 With the passage of the Orrin G. Hatch—Bob Goodlatte Music Modernization Act, Congress eliminated the four statutory factors for evaluating license royalty rates. See Public Law 115–264, 132 Stat. 3676 (2018) (codified in scattered sections of title 17, U.S.C. 253 According to the Services, all segments of the music industry are thriving [REDACTED]. E:\FR\FM\10AUR2.SGM 10AUR2 54480 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations introduction of portable devices and streaming services. Growth continues as those devices and services become increasingly easy to actuate in vehicles. No participant alleged, however, that music industry success is caused by or even correlated to the PR II rate structure. Coincidence is not probative evidence. ddrumheller on DSK120RN23PROD with RULES2 2. Assure Fair Return to Copyright Owner and Fair Income to the Licensee The Services argued they were receiving a fair income and copyright owners were receiving a fair return under the PR II regulations. Although the Services argued that overall music royalties absorbed an inordinate portion of their revenues, none expressly laid that lack of available revenue at the door of mechanical royalties. Amazon’s witness, Dr. Glenn Hubbard described a growing increase in streaming industry revenues and forecasts of continuing growth. See WRT of Glenn Hubbard (Feb. 15, 2017) ¶ 2.23–24 (Hubbard WRT). Dr. Hubbard deconstructed Amazon’s increased revenues and concluded that the growth in streaming services’ revenue resulted in increased royalty payments to music publishers and other rights holders. Id. ¶ 3.10. When royalty rates are calculated on a percent-of-revenue, the royalty payments increase when revenues increase. The difficulty with this tautological argument is that revenue growth as between services and rightsholders has not been proportional. And, as Copyright Owners have argued, the rate at which the services share with mechanical rightsholders is the issue in this proceeding. The Judges are not called upon to set annual royalty payment dollar amounts; rather they are mandated to set the rates that drive those dollar amounts. And to adopt regulations that most closely effectuate actual payment to rightsholders, minimizing revenue deferral and other such loopholes. For all of the reasons provided in the Determination and in this Benchmark Dissent, the PR II-based rates and the controlling rate structure do not balance the section 115 fair income-fair return scale appropriately and reasonably. 3. Weigh Relative Roles of Licensors and Licensees in Making the Works Available to the Public No participant presented evidence to elucidate specifically the relative roles of the parties relating to musical works. Economic evidence assumed that the marginal cost of streaming more music is minimal. This does not discount the services’ sunk costs, such as the original VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 technological or capital investments. With respect to the contributions of the copyright owners, the contribution is clear. It all begins with a song. Without new music, the Services could continue by streaming unregulated works, new arrangements or covers of existing works, and non-music content. Whether they would continue to enjoy the growth they have enjoyed over the last decade is unknown. The PR II rates might be a contributing factor to both stability and growth of the industry, but based on the totality of the evidence, the Dissent concludes that with regard to musical works, the relative role of the creator of the musical works, and to a lesser extent, the music publisher, is undervalued. 4. Minimize Disruption The language for the fourth statutory factor requires the Judges to establish a rate structure in such a way as ‘‘[t]o minimize any disruptive impact on the structure of the industries involved and on generally prevailing industry practices.’’ [FORMER] 17 U.S.C. 801(b)(1)(D). The Services argue that the change in rate structure determined by the Majority in this proceeding is massively, and potentially fatally, disruptive to music streaming services. Ironically, the music industry has been in a constant state of disruption since the introduction of digital music. From peer-to-peer sharing, to purchased permanent downloads, to interactive and non-interactive streaming, the history of modern music consumption has been a model of disruption. Entry into the streaming market by multifaceted digital ecosystem providers is just the latest significant change in music delivery to consumers. Innovation in music delivery is constant. Allegedly to minimize disruption, the Services advocated retention of the PR II rates and rate structure.254 While every aspect of the music industry is experiencing explosive growth, maintenance of the inadequate rates for mechanical licenses is unfathomable. Some change, phased in over time, might be uncomfortable for the licensees, but failure to change rates to acknowledge the music delivery revolution is not an option. With such a dynamic history and uncertain future, a change in mechanical license rates is not just inevitable, but mandatory. Indeed, the Benchmark Dissent’s approach in this proceeding advances 254 Tellingly, on remand, the Services did not pursue any argument that the changes in the rates or rate structure in the Determination were disruptive. PO 00000 Frm 00076 Fmt 4701 Sfmt 4700 the notion that streamed music is streamed music. This is certainly true from the viewpoint of the songwriters and publishers, and of music consumers. Rather than introduce separate rate structures for each new delivery technology or streaming business model, the Judges need to establish a rate that will fairly compensate Copyright Owners for the use of their works and permit a fair return to licensees, regardless of what next technological disruption they might choose to introduce to the industry. In the captioned proceeding, the Majority declined to label the rate structure and resulting rates incorporated in the regulations promulgated after the Phonorecords II proceeding as a benchmark, or starting point, for determination of new rates and terms in this proceeding. In the Determination, the Majority alluded to reasons they found the PR II rates to be inadequate to serve current circumstances.255 Nevertheless, the D.C. Circuit faulted the Majority for not providing adequate explanation of their rejection of the PR II benchmark in the first instance. See Johnson at 386–87. Indeed, the D.C. Circuit found the Majority’s reasoning on the issue in the Determination to be ‘‘muddled.’’ Id. F. Rate Structure For all of the reasons outlined above, the Remand Majority’s acceptance and adoption of the Phonorecords II rate structure results in a rate structure in this proceeding that suffers from the same deficits the Benchmark Dissent believes to be inherent in that rate structure. Changing the headline rate and capping the TCC rate prong do not cure the ills of the rate structure itself. True, the PR II-based rates permit price discrimination, which increases revenue, and therefore royalties, in absolute terms. Reinstatement of minima in the TCC prong introduces a failsafe to runaway TCC-based rates. The mechanical floors adopted in the Determination continue, protecting mechanical license rightsholders from runaway performance royalties. The Benchmark Dissent maintains that all these goals could be met equally well with a streamlined, transparent, fair, and reasonable rate structure, as several of the participants in this proceeding advocated. 255 The D.C. Circuit found that the Majority articulated a reasoned and reasonable rejection of the negotiated rates applicable to the categories of phonorecords included in [FORMER] subpart A of the regulations as a benchmark in this proceeding. The issue on remand is articulation of a reason for not using the other subparts of 37 CFR part 385 as a benchmark in this proceeding. See Johnson at 386. E:\FR\FM\10AUR2.SGM 10AUR2 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations III. Conclusion This Dissent in part is issued as a RESTRICTED document. Within 30 days of the date of issuance, the participants shall file a version of this Dissent with agreed redactions to permit viewing by the public. Issue Date: July 1, 2022. Suzanne M. Barnett Chief Copyright Royalty Judge List of Subjects in 37 CFR Part 385 Copyright, Phonorecords, Recordings. For the reasons set forth in the preamble, the Copyright Royalty Judges amend 37 CFR part 385 as follows. PART 385—RATES AND TERMS FOR USE OF NONDRAMATIC MUSICAL WORKS IN THE MAKING AND DISTRIBUTING OF PHYSICAL AND DIGITAL PHONORECORDS 1. The authority citation for part 385 continues to read as follows: ■ Authority: 17 U.S.C. 115, 801(b)(1), 804(b)(4). ■ 2. Add appendix A to read as follows: Appendix A to Part 385—Part 385 Applicable to the Period January 1, 2018, through December 31, 2022, as clarified on August 10, 2023 Note: Cross-references to part 385 in this appendix are to those provisions as contained within this appendix. PART 385—RATES AND TERMS FOR USE OF MUSICAL WORKS UNDER COMPULSORY LICENSE FOR MAKING AND DISTRIBUTING PHYSICAL AND DIGITAL PHONORECORDS Subpart A—Regulations of General Application 385.1 General. 385.2 Definitions. 385.3 Late payments. 385.4 Recordkeeping for promotional or free trial non-royalty-bearing uses. ddrumheller on DSK120RN23PROD with RULES2 Subpart B—Physical Phonorecord Deliveries, Permanent Downloads, Ringtones, and Music Bundles 385.10 Scope. 385.11 Royalty rates. Subpart C—Eligible Interactive Streaming, Eligible Limited Downloads, Limited Offerings, Mixed Service Bundles, Bundled Subscription Offerings, Locker Services, and Other Delivery Configurations 385.20 Scope. 385.21 Royalty rates and calculations. 385.22 Royalty floors for specific types of Offerings. Subpart D—Promotional Offerings, Free Trial Offerings and Certain Purchased Content Locker Services 385.30 Scope. VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 385.31 Royalty rates. Subpart A—Regulations of General Application § 385.1 General. (a) Scope. This part establishes rates and terms of royalty payments for the use of nondramatic musical works in making and distributing of physical and digital phonorecords in accordance with the provisions of 17 U.S.C. 115. This subpart contains regulations of general application to the making and distributing of phonorecords subject to the license under 17 U.S.C. 115 (section 115 license). (b) Legal compliance. Licensees relying on the compulsory license detailed in 17 U.S.C. 115 shall comply with the requirements of that section, the rates and terms of this part, and any other applicable regulations. This part describes rates and terms for the compulsory license only. (c) Interpretation. This part is intended only to set rates and terms for situations in which the exclusive rights of a Copyright Owner are implicated and a compulsory license pursuant to 17 U.S.C. 115 is obtained. Neither this part nor the act of obtaining a license under 17 U.S.C. 115 is intended to express or imply any conclusion as to the circumstances in which a user must obtain a compulsory license pursuant to 17 U.S.C. 115. (d) Relationship to voluntary agreements. The rates and terms of any license agreements entered into by Copyright Owners and Licensees relating to use of musical works within the scope of those license agreements shall apply in lieu of the rates and terms of this part. § 385.2 Definitions. For the purposes of this part, the following definitions apply: Accounting Period means the monthly period specified in 17 U.S.C. 115(c)(2)(I) and (d)(4)(A)(i), and any related regulations in this chapter, as applicable. Active Subscriber means an End User of a Bundled Subscription Offering who has made at least one Play during the Accounting Period. Affiliate means an entity controlling, controlled by, or under common control with another entity, except that an affiliate of a Sound Recording Company shall not include a Copyright Owner to the extent it is engaging in business as to musical works. Bundled Subscription Offering means a Subscription Offering providing Licensed Activity consisting of Eligible Interactive Streams or Eligible Limited Downloads that is made available to End Users with one or more other products or services (including products or services subject to other subparts) as part of a single transaction without pricing for the subscription service providing Licensed Activity separate from the product(s) or service(s) with which it is made available (e.g., a case in which a user can buy a portable device and one-year access to a subscription service providing Licensed Activity for a single price). Copyright Owner(s) are nondramatic musical works copyright owners who are PO 00000 Frm 00077 Fmt 4701 Sfmt 4700 54481 entitled to royalty payments made under this part pursuant to the compulsory license under 17 U.S.C. 115. Digital Phonorecord Delivery has the same meaning as in 17 U.S.C. 115(e)(10). Eligible Interactive Stream means a Stream in which the performance of the sound recording is not exempt from the sound recording performance royalty under 17 U.S.C. 114(d)(1) and does not in itself, or as a result of a program in which it is included, qualify for statutory licensing under 17 U.S.C. 114(d)(2). Eligible Limited Download means a Limited Download as defined in 17 U.S.C. 115(e)(16) that is only accessible for listening for— (1) An amount of time not to exceed one month from the time of the transmission (unless the Licensee, in lieu of retransmitting the same sound recording as another Eligible Limited Download, separately, and upon specific request of the End User made through a live network connection, reauthorizes use for another time period not to exceed one month), or in the case of a subscription plan, a period of time following the end of the applicable subscription no longer than a subscription renewal period or three months, whichever is shorter; or (2) A number of times not to exceed 12 (unless the Licensee, in lieu of retransmitting the same sound recording as another Eligible Limited Download, separately, and upon specific request of the End User made through a live network connection, reauthorizes use of another series of 12 or fewer plays), or in the case of a subscription transmission, 12 times after the end of the applicable subscription. End User means each unique person that: (1) Pays a subscription fee for an Offering during the relevant Accounting Period; or (2) Makes at least one Play during the relevant Accounting Period. Family Plan means a discounted Subscription Offering to be shared by two or more family members for a single subscription price. Free Trial Offering means a subscription to a Service Provider’s transmissions of sound recordings embodying musical works when: (1) Neither the Service Provider, the Sound Recording Company, the Copyright Owner, nor any person or entity acting on behalf of or in lieu of any of them receives any monetary consideration for the Offering; (2) The free usage does not exceed 30 consecutive days per subscriber per two-year period; (3) In connection with the Offering, the Service Provider is operating with appropriate musical license authority and complies with the recordkeeping requirements in § 385.4; (4) Upon receipt by the Service Provider of written notice from the Copyright Owner or its agent stating in good faith that the Service Provider is in a material manner operating without appropriate license authority from the Copyright Owner under 17 U.S.C. 115, the Service Provider shall within 5 business days cease transmission of the sound recording embodying that musical work and withdraw it from the repertoire available as part of a Free Trial Offering; E:\FR\FM\10AUR2.SGM 10AUR2 ddrumheller on DSK120RN23PROD with RULES2 54482 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations (5) The Free Trial Offering is made available to the End User free of any charge; and (6) The Service Provider offers the End User periodically during the free usage an opportunity to subscribe to a non-Free Trial Offering of the Service Provider. GAAP means U.S. Generally Accepted Accounting Principles in effect at the relevant time, except that if the U.S. Securities and Exchange Commission permits or requires entities with securities that are publicly traded in the U.S. to employ International Financial Reporting Standards in lieu of Generally Accepted Accounting Principles, then that entity may employ International Financial Reporting Standards as ‘‘GAAP’’ for purposes of this subpart. Licensee means any entity availing itself of the compulsory license under 17 U.S.C. 115 to use copyrighted musical works in the making or distributing of physical or digital phonorecords. Licensed Activity, as the term is used in subpart B of this part, means delivery of musical works, under voluntary or statutory license, via physical phonorecords and Digital Phonorecord Deliveries in connection with Permanent Downloads, Ringtones, and Music Bundles; and, as the term is used in subparts C and D of this part, means delivery of musical works, under voluntary or statutory license, via Digital Phonorecord Deliveries in connection with Eligible Interactive Streams, Eligible Limited Downloads, Limited Offerings, mixed Bundles, and Locker Services. Limited Offering means a Subscription Offering providing Eligible Interactive Streams or Eligible Limited Downloads for which— (1) An End User cannot choose to listen to a particular sound recording (i.e., the Service Provider does not provide Eligible Interactive Streams of individual recordings that are ondemand, and Eligible Limited Downloads are rendered only as part of programs rather than as individual recordings that are on-demand); or (2) The particular sound recordings available to the End User over a period of time are substantially limited relative to Service Providers in the marketplace providing access to a comprehensive catalog of recordings (e.g., a product limited to a particular genre or permitting Eligible Interactive Streams only from a monthly playlist consisting of a limited set of recordings). Locker Service means an Offering providing digital access to sound recordings of musical works in the form of Eligible Interactive Streams, Permanent Downloads, Restricted Downloads or Ringtones where the Service Provider has reasonably determined that the End User has purchased or is otherwise in possession of the subject phonorecords of the applicable sound recording prior to the End User’s first request to use the sound recording via the Locker Service. The term Locker Service does not mean any part of a Service Provider’s products otherwise meeting this definition, but as to which the Service Provider has not obtained a section 115 license. Mixed Service Bundle means one or more of Permanent Downloads, Ringtones, Locker VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 Services, or Limited Offerings a Service Provider delivers to End Users together with one or more non-music services (e.g., internet access service, mobile phone service) or nonmusic products (e.g., a telephone device) of more than token value and provided to users as part of one transaction without pricing for the music services or music products separate from the whole Offering. Music Bundle means two or more of physical phonorecords, Permanent Downloads, or Ringtones delivered as part of one transaction (e.g., download plus ringtone, CD plus downloads). In the case of Music Bundles containing one or more physical phonorecords, the Service Provider must sell the physical phonorecord component of the Music Bundle under a single catalog number, and the musical works embodied in the Digital Phonorecord Delivery configurations in the Music Bundle must be the same as, or a subset of, the musical works embodied in the physical phonorecords; provided that when the Music Bundle contains a set of Digital Phonorecord Deliveries sold by the same Sound Recording Company under substantially the same title as the physical phonorecord (e.g., a corresponding digital album), the Service Provider may include in the same bundle up to 5 sound recordings of musical works that are included in the stand-alone version of the set of digital phonorecord deliveries but not included on the physical phonorecord. In addition, the Service Provider must permanently part with possession of the physical phonorecord or phonorecords it sells as part of the Music Bundle. In the case of Music Bundles composed solely of digital phonorecord deliveries, the number of digital phonorecord deliveries in either configuration cannot exceed 20, and the musical works embodied in each configuration in the Music Bundle must be the same as, or a subset of, the musical works embodied in the configuration containing the most musical works. Offering means a Service Provider’s engagement in Licensed Activity covered by subparts C and D of this part. Paid Locker Service means a Locker Service for which the End User pays a fee to the Service Provider. Performance Royalty means the license fee payable for the right to perform publicly musical works in any of the forms covered by subparts C and D this part. Permanent Download has the same meaning as in 17 U.S.C. 115(e)(24). Play means an Eligible Interactive Stream, or a play of an Eligible Limited Download, lasting 30 seconds or more and, if a track lasts in its entirety under 30 seconds, an Eligible Interactive Stream or a play of an Eligible Limited Download of the entire duration of the track. A Play excludes an Eligible Interactive Stream or a play of an Eligible Limited Download that has not been initiated or requested by a human user. If a single End User plays the same track more than 50 straight times, all plays after play 50 shall be deemed not to have been initiated or requested by a human user. Promotional Offering means a digital transmission of a sound recording, in the form of an Eligible Interactive Stream or an PO 00000 Frm 00078 Fmt 4701 Sfmt 4700 Eligible Limited Download, embodying a musical work, the primary purpose of which is to promote the sale or other paid use of that sound recording or to promote the artist performing on that sound recording and not to promote or suggest promotion or endorsement of any other good or service and: (1) A Sound Recording Company is lawfully distributing the sound recording through established retail channels or, if the sound recording is not yet released, the Sound Recording Company has a good faith intention to lawfully distribute the sound recording or a different version of the sound recording embodying the same musical work; (2) For Eligible Interactive Streams or Eligible Limited Downloads, the Sound Recording Company requires a writing signed by an authorized representative of the Service Provider representing that the Service Provider is operating with appropriate musical works license authority and that the Service Provider is in compliance with the recordkeeping requirements of § 385.4; (3) For Eligible Interactive Streams of segments of sound recordings not exceeding 90 seconds, the Sound Recording Company delivers or authorizes delivery of the segments for promotional purposes and neither the Service Provider nor the Sound Recording Company creates or uses a segment of a sound recording in violation of 17 U.S.C. 106(2) or 115(a)(2); (4) The Promotional Offering is made available to an End User free of any charge; and (5) The Service Provider provides to the End User at the same time as the Promotional Offering Stream an opportunity to purchase the sound recording or the Service Provider periodically offers End Users the opportunity to subscribe to a paid Offering of the Service Provider. Purchased Content Locker Service means a Locker Service made available to End User purchasers of Permanent Downloads, Ringtones, or physical phonorecords at no incremental charge above the otherwise applicable purchase price of the Permanent Downloads, Ringtones, or physical phonorecords acquired from a qualifying seller. With a Purchased Content Locker Service, an End User may receive one or more additional phonorecords of the purchased sound recordings of musical works in the form of Permanent Downloads or Ringtones at the time of purchase, or subsequently have digital access to the purchased sound recordings of musical works in the form of Eligible Interactive Streams, additional Permanent Downloads, Restricted Downloads, or Ringtones. (1) A qualifying seller for purposes of this definition is the entity operating the Service Provider, including Affiliates, predecessors, or successors in interest, or— (i) In the case of Permanent Downloads or Ringtones, a seller having a legitimate connection to the locker service provider pursuant to one or more written agreements (including that the Purchased Content Locker Service and Permanent Downloads or Ringtones are offered through the same third party); or E:\FR\FM\10AUR2.SGM 10AUR2 ddrumheller on DSK120RN23PROD with RULES2 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations (ii) In the case of physical phonorecords: (A) The seller of the physical phonorecord has an agreement with the Purchased Content Locker Service provider establishing an integrated offer that creates a consumer experience commensurate with having the same Service Provider both sell the physical phonorecord and offer the integrated locker service; or (B) The Service Provider has an agreement with the entity offering the Purchased Content Locker Service establishing an integrated offer that creates a consumer experience commensurate with having the same Service Provider both sell the physical phonorecord and offer the integrated locker service. (2) [Reserved] Relevant Page means an electronic display (for example, a web page or screen) from which a Service Provider’s Offering consisting of Eligible Interactive Streams or Eligible Limited Downloads is directly available to End Users, but only when the Offering and content directly relating to the Offering (e.g., an image of the artist, information about the artist or album, reviews, credits, and music player controls) comprises 75% or more of the space on that display, excluding any space occupied by advertising. An Offering is directly available to End Users from a page if End Users can receive sound recordings of musical works (in most cases this will be the page on which the Eligible Limited Download or Eligible Interactive Stream takes place). Restricted Download means a Digital Phonorecord Delivery in a form that cannot be retained and replayed on a permanent basis. The term Restricted Download includes an Eligible Limited Download. Ringtone means a phonorecord of a part of a musical work distributed as a Digital Phonorecord Delivery in a format to be made resident on a telecommunications device for use to announce the reception of an incoming telephone call or other communication or message or to alert the receiver to the fact that there is a communication or message. Service Provider means that entity governed by subparts C and D of this part, which might or might not be the Licensee, that with respect to the section 115 license: (1) Contracts with or has a direct relationship with End Users or otherwise controls the content made available to End Users; (2) Is able to report fully on Service Provider Revenue from the provision of musical works embodied in phonorecords to the public, and to the extent applicable, verify Service Provider Revenue through an audit; and (3) Is able to report fully on its usage of musical works, or procure such reporting and, to the extent applicable, verify usage through an audit. Service Provider Revenue, as used in this part: (1) Subject to paragraphs (2) through (5) of this definition and subject to GAAP, Service Provider Revenue shall mean: (i) All revenue from End Users recognized by a Service Provider for the provision of any Offering; (ii) All revenue recognized by a Service Provider by way of sponsorship and VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 commissions as a result of the inclusion of third-party ‘‘in-stream’’ or ‘‘in-download’’ advertising as part of any Offering, i.e., advertising placed immediately at the start or end of, or during the actual delivery of, a musical work, by way of Eligible Interactive Streaming or Eligible Limited Downloads; and (iii) All revenue recognized by the Service Provider, including by way of sponsorship and commissions, as a result of the placement of third-party advertising on a Relevant Page of the Service Provider or on any page that directly follows a Relevant Page leading up to and including the Eligible Limited Download or Eligible Interactive Stream of a musical work; provided that, in case more than one Offering is available to End Users from a Relevant Page, any advertising revenue shall be allocated between or among the Service Providers on the basis of the relative amounts of the page they occupy. (2) Service Provider Revenue shall: (i) Include revenue recognized by the Service Provider, or by any associate, Affiliate, agent, or representative of the Service Provider in lieu of its being recognized by the Service Provider; and (ii) Include the value of any barter or other nonmonetary consideration; and (iii) Except as expressly detailed in this part, not be subject to any other deduction or set-off other than refunds to End Users for Offerings that the End Users were unable to use because of technical faults in the Offering or other bona fide refunds or credits issued to End Users in the ordinary course of business. (3) Service Provider Revenue shall exclude revenue derived by the Service Provider solely in connection with activities other than Offering(s), whereas advertising or sponsorship revenue derived in connection with any Offering(s) shall be treated as provided in paragraphs (2) and (4) of this definition. (4) For purposes of paragraph (1) of this definition, advertising or sponsorship revenue shall be reduced by the actual cost of obtaining that revenue, not to exceed 15%. (5) In instances in which a Service Provider provides an Offering to End Users as part of the same transaction with one or more other products or services that are not Licensed Activities, then the revenue from End Users deemed to be recognized by the Service Provider for the Offering for the purpose of paragraph (1) of this definition shall be the revenue recognized from End Users for the bundle less the standalone published price for End Users for each of the other component(s) of the bundle; provided that, if there is no standalone published price for a component of the bundle, then the Service Provider shall use the average standalone published price for End Users for the most closely comparable product or service in the U.S. or, if more than one comparable exists, the average of standalone prices for comparables. (6) In the case of a Mixed Service Bundle, the revenue deemed to be recognized from End Users for the Offering for the purpose of paragraph (1) of this definition shall be the greater of— PO 00000 Frm 00079 Fmt 4701 Sfmt 4700 54483 (i) The revenue deemed to be recognized pursuant to paragraph (5) of this definition; and (ii) Either— (A) In the case of a Mixed Service Bundle that either has 750,000 subscribers or other registered users, or is reasonably expected to have 750,000 subscribers or other registered users within 1 year after commencement of the Mixed Service Bundle, 40% of the standalone published price of the licensed music component of the bundle (i.e., the Permanent Downloads, Ringtones, Locker Service, or Limited Offering); provided that, if there is no such standalone published price for the licensed music component of the bundle, then the average standalone published price for End Users for the most closely comparable licensed music component in the U.S. shall be used or, if more than one such comparable exists, the average of such standalone prices for such comparables shall be used; and further provided that in any case in which royalties were paid based on this paragraph (6)(ii)(A) due to a reasonable expectation of reaching 750,000 subscribers or other registered users within 1 year after commencement of the Mixed Service Bundle and that does not actually happen, applicable payments shall, in the accounting period next following the end of such 1-year period, retroactively be adjusted as if paragraph (6)(ii)(B) of this definition applied; or (B) Otherwise, 50% of the standalone published price of the licensed music component of the bundle (i.e., the Permanent Downloads, Ringtones, Locker Service, or Limited Offering); provided that, if there is no such standalone published price for the licensed music component of the bundle, then the average standalone published price for End Users for the most closely comparable licensed music component in the U.S. shall be used or, if more than one such comparable exists, the average of such standalone prices for such comparables shall be used. Sound Recording Company means a person or entity that: (1) Is a copyright owner of a sound recording embodying a musical work; (2) In the case of a sound recording of a musical work fixed before February 15, 1972, has rights to the sound recording, under 17 U.S.C. chapter 14, that are equivalent to the rights of a copyright owner of a sound recording of a musical work under title 17, United States Code; (3) Is an exclusive Licensee of the rights to reproduce and distribute a sound recording of a musical work; or (4) Performs the functions of marketing and authorizing the distribution of a sound recording of a musical work under its own label, under the authority of the Copyright Owner of the sound recording. Standalone Non-Portable Subscription Offering—Mixed means a Subscription Offering through which an End User can listen to sound recordings either in the form of Eligible Interactive Streams or Eligible Limited Downloads but only from a nonportable device to which those Eligible Interactive Streams or Eligible Limited Downloads are originally transmitted. E:\FR\FM\10AUR2.SGM 10AUR2 ddrumheller on DSK120RN23PROD with RULES2 54484 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations Standalone Non-Portable Subscription Offering—Streaming Only means a Subscription Offering through which an End User can listen to sound recordings only in the form of Eligible Interactive Streams and only from a non-portable device to which those Eligible Interactive Streams are originally transmitted while the device has a live network connection. Standalone Portable Subscription Offering means a Subscription Offering through which an End User can listen to sound recordings in the form of Eligible Interactive Streams or Eligible Limited Downloads from a portable device. Stream means the digital transmission of a sound recording of a musical work to an End User— (1) To allow the End User to listen to the sound recording, while maintaining a live network connection to the transmitting service, substantially at the time of transmission, except to the extent that the sound recording remains accessible for future listening from a Streaming Cache Reproduction; (2) Using technology that is designed such that the sound recording does not remain accessible for future listening, except to the extent that the sound recording remains accessible for future listening from a Streaming Cache Reproduction; and (3) That is subject to licensing as a public performance of the musical work. Streaming Cache Reproduction means a reproduction of a sound recording embodying a musical work made on a computer or other receiving device by a Service Provider solely for the purpose of permitting an End User who has previously received a Stream of that sound recording to play the sound recording again from local storage on the computer or other device rather than by means of a transmission; provided that the End User is only able to do so while maintaining a live network connection to the Service Provider, and the reproduction is encrypted or otherwise protected consistent with prevailing industry standards to prevent it from being played in any other manner or on any device other than the computer or other device on which it was originally made. Student Plan means a discounted Subscription Offering available on a limited basis to students. Subscription Offering means an Offering for which End Users are required to pay a fee to have access to the Offering for defined subscription periods of 3 years or less (in contrast to, for example, a service where the basic charge to users is a payment per download or per play), whether the End User makes payment for access to the Offering on a standalone basis or as part of a bundle with one or more other products or services. Total Cost of Content or TCC means the total amount expensed by a Service Provider or any of its Affiliates in accordance with GAAP for rights to make Eligible Interactive Streams or Eligible Limited Downloads of a musical work embodied in a sound recording through the Service Provider for the Accounting Period, which amount shall equal the Applicable Consideration for those rights at the time the Applicable VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 Consideration is properly recognized as an expense under GAAP. As used in this definition, Applicable Consideration means anything of value given for the identified rights to undertake the Licensed Activity, including, without limitation, ownership equity, monetary advances, barter or any other monetary and/or nonmonetary consideration, whether that consideration is conveyed via a single agreement, multiple agreements and/or agreements that do not themselves authorize the Licensed Activity but nevertheless provide consideration for the identified rights to undertake the Licensed Activity, and including any value given to an Affiliate of a Sound Recording Company for the rights to undertake the Licensed Activity. Value given to a Copyright Owner of musical works that is controlling, controlled by, or under common control with a Sound Recording Company for rights to undertake the Licensed Activity shall not be considered value given to the Sound Recording Company. Notwithstanding the foregoing, Applicable Consideration shall not include in-kind promotional consideration given to a Sound Recording Company (or Affiliate thereof) that is used to promote the sale or paid use of sound recordings embodying musical works or the paid use of music services through which sound recordings embodying musical works are available where the in-kind promotional consideration is given in connection with a use that qualifies for licensing under 17 U.S.C. 115. § 385.3 Late payments. A Licensee shall pay a late fee of 1.5% per month, or the highest lawful rate, whichever is lower, for any payment owed to a Copyright Owner and remaining unpaid after the due date established in 17 U.S.C. 115(c)(2)(I) or (d)(4)(A)(i), as applicable and detailed in part 210 of this title. Late fees shall accrue from the due date until the Copyright Owner receives payment, except that where payment is due to the mechanical licensing collective under 17 U.S.C. 115(d)(4)(A)(i), late fees shall accrue from the due date until the mechanical licensing collective receives payment. § 385.4 Recordkeeping for promotional or free trial non-royalty-bearing uses. (a) General. A Licensee transmitting a sound recording embodying a musical work subject to section 115 and subparts C and D of this part and claiming a Promotional Offering or Free Trial Offering zero royalty rate shall keep complete and accurate contemporaneous written records of making or authorizing Eligible Interactive Streams or Eligible Limited Downloads, including the sound recordings and musical works involved, the artists, the release dates of the sound recordings, a brief statement of the promotional activities authorized, the identity of the Offering or Offerings for which the zero-rate is authorized (including the internet address if applicable), and the beginning and end date of each zero rate Offering. (b) Retention of records. A Service Provider claiming zero rates shall maintain the records required by this section for no less time than PO 00000 Frm 00080 Fmt 4701 Sfmt 4700 the Service Provider maintains records of royalty-bearing uses involving the same types of Offerings in the ordinary course of business, but in no event for fewer than five years from the conclusion of the zero rate Offerings to which they pertain. (c) Availability of records. If a Copyright Owner or agent requests information concerning zero rate Offerings, the Licensee shall respond to the request within an agreed, reasonable time. Subpart B—Physical Phonorecord Deliveries, Permanent Downloads, Ringtones, and Music Bundles § 385.10 Scope. This subpart establishes rates and terms of royalty payments for making and distributing phonorecords, including by means of Digital Phonorecord Deliveries, in accordance with the provisions of 17 U.S.C. 115. § 385.11 Royalty rates. (a) Physical phonorecord deliveries and Permanent Downloads. For every physical phonorecord and Permanent Download the Licensee makes and distributes or authorizes to be made and distributed, the royalty rate payable for each work embodied in the phonorecord or Permanent Download shall be either 9.1 cents or 1.75 cents per minute of playing time or fraction thereof, whichever amount is larger. (b) Ringtones. For every Ringtone the Licensee makes and distributes or authorizes to be made and distributed, the royalty rate payable for each work embodied therein shall be 24 cents. (c) Music Bundles. For a Music Bundle, the royalty rate for each element of the Music Bundle shall be the rate required under paragraph (a) or (b) of this section, as appropriate. Subpart C—Eligible Interactive Streaming, Eligible Limited Downloads, Limited Offerings, Mixed Service Bundles, Bundled Subscription Offerings, Locker Services, and Other Delivery Configurations § 385.20 Scope. This subpart establishes rates and terms of royalty payments for Eligible Interactive Streams and Eligible Limited Downloads of musical works, and other reproductions or distributions of musical works through Limited Offerings, Mixed Service Bundles, Bundled Subscription Offerings, Paid Locker Services, and Purchased Content Locker Services provided through subscription and nonsubscription digital music Service Providers in accordance with the provisions of 17 U.S.C. 115, exclusive of Offerings subject to subpart D of this part. § 385.21 Royalty rates and calculations. (a) Applicable royalty. Licensees that engage in Licensed Activity covered by this subpart pursuant to 17 U.S.C. 115 shall pay royalties therefor that are calculated as provided in this section, subject to the royalty floors for specific types of services E:\FR\FM\10AUR2.SGM 10AUR2 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations described in § 385.22, provided, however, that Promotional Offerings, Free Trial Offerings, and certain Purchased Content Locker Services shall instead be subject to the royalty rates provided in subpart D of this part. (b) Rate calculation. Royalty payments for Licensed Activity in this subpart shall be calculated as provided in this paragraph (b). If a Service Provider includes different Offerings, royalties must be calculated separately with respect to each Offering taking into consideration Service Provider Revenue and expenses associated with each Offering. (1) Step 1: Calculate the all-in royalty for the Offering. For each Accounting Period, the all-in royalty for each Offering under this 54485 subpart shall be the greater of the applicable percent of Service Provider Revenue, as set forth in table 1 to this paragraph (b)(1), and the result of the TCC Prong Calculation for the respective type of Offering, as set forth in table 2 to this paragraph (b)(1): TABLE 1 TO PARAGRAPH (b)(1) Royalty year 2018 2019 2020 2021 2022 Percent of Service Provider Revenue ..................................................... 11.4 12.3 13.3 14.2 15.1 TABLE 2 TO PARAGRAPH (b)(1) Type of offering TCC prong calculation Standalone Non-Portable Subscription Offering—Streaming Only .......... The lesser of 22% of TCC for the Accounting Period and 50 cents per subscriber per month. The lesser of 21% of TCC for the Accounting Period and 50 cents per subscriber per month. The lesser of 21% of TCC for the Accounting Period and 80 cents per subscriber per month. 21% of TCC for the Accounting Period. 22% of TCC for the Accounting Period. Standalone Non-Portable Subscription Offering—Mixed ......................... Standalone Portable Subscription Offering .............................................. ddrumheller on DSK120RN23PROD with RULES2 Bundled Subscription Offering .................................................................. Free nonsubscription/ad-supported services free of any charge to the End User. Mixed Service Bundle ............................................................................... Purchased Content Locker Service .......................................................... Limited Offering ........................................................................................ Paid Locker Service ................................................................................. (2) Step 2: Subtract applicable Performance Royalties. From the amount determined in step 1 in paragraph (b)(1) of this section, for each Offering of the Service Provider, subtract the total amount of Performance Royalty that the Service Provider has expensed or will expense pursuant to public performance licenses in connection with uses of musical works through that Offering during the Accounting Period that constitute Licensed Activity. Although this amount may be the total of the Service Provider’s payments for that Offering for the Accounting Period, it will be less than the total of the Performance Royalties if the Service Provider is also engaging in public performance of musical works that does not constitute Licensed Activity. In the case in which the Service Provider is also engaging in the public performance of musical works that does not constitute Licensed Activity, the amount to be subtracted for Performance Royalties shall be the amount allocable to Licensed Activity uses through the relevant Offering as determined in relation to all uses of musical works for which the Service Provider pays Performance Royalties for the Accounting Period. The Service Provider shall make this allocation on the basis of Plays of musical works or, where per-play information is unavailable because of bona fide technical limitations as described in step 4 in paragraph (b)(4) of this section, using the same alternative methodology as provided in step 4. (3) Step 3: Determine the payable royalty pool. The payable royalty pool is the amount payable for the reproduction and distribution of all musical works used by the Service Provider by virtue of its Licensed Activity for VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 21% of TCC for the Accounting Period. 22% of TCC for the Accounting Period. 21% of TCC for the Accounting Period. 20.65% of TCC for the Accounting Period. a particular Offering during the Accounting Period. This amount is the greater of: (i) The result determined in step 2 in paragraph (b)(2) of this section; and (ii) The royalty floor (if any) resulting from the calculations described in § 385.22. (4) Step 4: Calculate the per-work royalty allocation. This is the amount payable for the reproduction and distribution of each musical work used by the Service Provider by virtue of its Licensed Activity through a particular Offering during the Accounting Period. To determine this amount, the result determined in step 3 in paragraph (b)(3) of this section must be allocated to each musical work used through the Offering. The allocation shall be accomplished by dividing the payable royalty pool determined in step 3 for the Offering by the total number of Plays of all musical works through the Offering during the Accounting Period (other than Plays subject to subpart D of this part) to yield a per-Play allocation, and multiplying that result by the number of Plays of each musical work (other than Plays subject to subpart D of this part) through the Offering during the Accounting Period. For purposes of determining the per-work royalty allocation in all calculations under this paragraph (b)(4) only (i.e., after the payable royalty pool has been determined), for sound recordings of musical works with a playing time of over 5 minutes, each Play shall be counted as provided in paragraph (c) of this section. Notwithstanding the foregoing, if the Service Provider is not capable of tracking Play information because of bona fide limitations of the available technology for Offerings of that nature or of devices useable with the Offering, the per-work royalty PO 00000 Frm 00081 Fmt 4701 Sfmt 4700 allocation may instead be accomplished in a manner consistent with the methodology used for making royalty payment allocations for the use of individual sound recordings. (c) Overtime adjustment. For purposes of the calculations in step 4 in paragraph (b)(4) of this section only, for sound recordings of musical works with a playing time of over 5 minutes, adjust the number of Plays as follows: (1) 5:01 to 6:00 minutes—Each Play = 1.2 Plays. (2) 6:01 to 7:00 minutes—Each Play = 1.4 Plays. (3) 7:01 to 8:00 minutes—Each Play = 1.6 Plays. (4) 8:01 to 9:00 minutes—Each Play = 1.8 Plays. (5) 9:01 to 10:00 minutes—Each Play = 2.0 Plays. (6) For playing times of greater than 10 minutes, continue to add 0.2 Plays for each additional minute or fraction thereof. (d) Accounting. The calculations required by paragraph (b) of this section shall be made in good faith and on the basis of the best knowledge, information, and belief at the time payment is due, and subject to the additional accounting and certification requirements of 17 U.S.C. 115(c)(2)(I) and (d)(4)(A)(i) and part 210 of this title. Without limitation, statements of account (where applicable) shall set forth each step of the calculations with sufficient information to allow the assessment of the accuracy and manner in which the payable royalty pool and per-play allocations (including information sufficient to demonstrate whether and how a royalty floor pursuant to § 385.22 does or does not apply) were E:\FR\FM\10AUR2.SGM 10AUR2 54486 Federal Register / Vol. 88, No. 153 / Thursday, August 10, 2023 / Rules and Regulations determined and, for each Offering reported, also indicate the type of Licensed Activity involved and the number of Plays of each musical work (including an indication of any overtime adjustment applied) that is the basis of the per-work royalty allocation being paid. (e) Computation of subscriber months in TCC Prong Calculation. In connection with the TCC Prong Calculation in step 1 in paragraph (b)(1) of this section for an Accounting Period, to the extent applicable, the total number of subscriber-months for the Accounting Period shall be calculated, taking all End Users who were subscribers for complete calendar months, prorating in the case of End Users who were subscribers for only part of a calendar month, and deducting on a prorated basis for End Users covered by an Offering subject to subpart D of this part. The product of the total number of subscriber-months for the Accounting Period and the specified number of cents per subscriber shall be used as the subscriberbased component (if any) in step 1 for the Accounting Period. ddrumheller on DSK120RN23PROD with RULES2 § 385.22 Royalty floors for specific types of Offerings. (a) In general. The following royalty floors for use in step 3 of § 385.21(b)(3)(ii) shall apply to the respective types of Offerings. (1) Standalone Non-Portable Subscription Offering—Streaming Only. Except as provided in paragraph (a)(4) of this section, in the case of a Subscription Offering through which an End User can listen to sound recordings only in the form of Eligible Interactive Streams and only from a nonportable device to which those Streams are originally transmitted while the device has a live network connection, the royalty floor is the aggregate amount of 15 cents per subscriber per month. (2) Standalone Non-Portable Subscription Offering—Mixed. Except as provided in paragraph (a)(4) of this section, in the case of a Subscription Offering through which an End User can listen to sound recordings either in the form of Eligible Interactive Streams or Eligible Limited Downloads but only from a non-portable device to which those Streams or Eligible Limited Downloads are originally transmitted, the royalty floor is VerDate Sep<11>2014 18:35 Aug 09, 2023 Jkt 259001 the aggregate amount of 30 cents per subscriber per month. (3) Standalone Portable Subscription Offering. Except as provided in paragraph (a)(4) of this section, in the case of a Subscription Offering through which an End User can listen to sound recordings in the form of Eligible Interactive Streams or Eligible Limited Downloads from a portable device, the royalty floor is the aggregate amount of 50 cents per subscriber per month. (4) Bundled Subscription Offering. In the case of a Bundled Subscription Offering, the royalty floor is the aggregate amount of 25 cents per month for each Active Subscriber. (b) Computation of royalty floors. For purposes of paragraph (a) of this section, to determine the royalty floor, as applicable to any particular Offering, the total number of subscriber-months for the Accounting Period shall be calculated by taking all End Users who were subscribers for complete calendar months, prorating in the case of End Users who were subscribers for only part of a calendar month, and deducting on a prorated basis for End Users covered by an Offering subject to subpart D of this part, except in the case of a Bundled Subscription Offering, subscriber-months shall be determined with respect to Active Subscribers. The product of the total number of subscriber-months for the Accounting Period and the specified number of cents per subscriber (or Active Subscriber, as the case may be) shall be used as the subscriber-based component of the royalty floor for the Accounting Period. A Family Plan shall be treated as 1.5 subscribers per month, prorated in the case of a Family Plan subscription in effect for only part of a calendar month. A Student Plan shall be treated as 0.50 subscribers per month, prorated in the case of a Student Plan End User who subscribed for only part of a calendar month. Subpart D—Promotional Offerings, Free Trial Offerings and Certain Purchased Content Locker Services § 385.30 Scope. This subpart establishes rates and terms of royalty payments for Promotional Offerings, Free Trial Offerings, and certain Purchased PO 00000 Frm 00082 Fmt 4701 Sfmt 9990 Content Locker Services provided by subscription and nonsubscription digital music Service Providers in accordance with the provisions of 17 U.S.C. 115. § 385.31 Royalty rates. (a) Promotional Offerings. For Promotional Offerings of audio-only Eligible Interactive Streams and Eligible Limited Downloads of sound recordings embodying musical works that the Sound Recording Company authorizes royalty-free to the Service Provider, the royalty rate is zero. (b) Free Trial Offerings. For Free Trial Offerings for which the Service Provider receives no monetary consideration, the royalty rate is zero. (c) Certain Purchased Content Locker Services. For every Purchased Content Locker Service for which the Service Provider receives no monetary consideration, the royalty rate is zero. (d) Unauthorized use. If a Copyright Owner or agent of the Copyright Owner sends written notice to a Licensee stating in good faith that a particular Offering subject to this subpart differs in a material manner from the terms governing that Offering, the Licensee must within 5 business days cease Streaming or otherwise making available that Copyright Owner’s musical works and shall withdraw from the identified Offering any End User’s access to the subject musical work. Dated: July 3, 2023. lllllllllllllllllllll David P. Shaw, Chief Copyright Royalty Judge lllllllllllllllllllll David R. Strickler, Copyright Royalty Judge lllllllllllllllllllll Steve Ruwe, Copyright Royalty Judge Approved by: lllllllllllllllllllll Carla D. Hayden, Librarian of Congress. [FR Doc. 2023–14925 Filed 8–9–23; 8:45 am] BILLING CODE 1410–72–P E:\FR\FM\10AUR2.SGM 10AUR2

Agencies

[Federal Register Volume 88, Number 153 (Thursday, August 10, 2023)]
[Rules and Regulations]
[Pages 54406-54486]
From the Federal Register Online via the Government Publishing Office [www.gpo.gov]
[FR Doc No: 2023-14925]



[[Page 54405]]

Vol. 88

Thursday,

No. 153

August 10, 2023

Part II





Library of Congress





-----------------------------------------------------------------------





Copyright Royalty Board





-----------------------------------------------------------------------





37 CFR Part 385





Determination of Royalty Rates and Terms for Making and Distributing 
Phonorecords (Phonorecords III); Final Rule

Federal Register / Vol. 88 , No. 153 / Thursday, August 10, 2023 / 
Rules and Regulations

[[Page 54406]]


-----------------------------------------------------------------------

LIBRARY OF CONGRESS

Copyright Royalty Board

37 CFR Part 385

[Docket No. 16-CRB-0003-PR (2018-2022) (Remand)]


Determination of Royalty Rates and Terms for Making and 
Distributing Phonorecords (Phonorecords III)

AGENCY: Copyright Royalty Board, Library of Congress.

ACTION: Final rule and order.

-----------------------------------------------------------------------

SUMMARY: The Copyright Royalty Judges announce their final 
determination after remand of the rates and terms for making and 
distributing phonorecords for the period beginning January 1, 2018, and 
ending on December 31, 2022.

DATES: 
    Effective date: August 10, 2023.
    Applicability date: The regulations apply to the license period 
beginning January 1, 2018, and ending December 31, 2022.

ADDRESSES: The final determination after remand is posted in eCRB at 
https://app.crb.gov/. For access to the docket to read the final 
determination after remand and submitted background documents, go to 
eCRB and search for docket number 16-CRB-0003-PR (2018-2022) (Remand).

FOR FURTHER INFORMATION CONTACT: Anita Brown, CRB Program Assistant, 
(202) 707-7658, [email protected].

SUPPLEMENTARY INFORMATION: 

Final Determination After Remand

    On October 26, 2020, the United States Court of Appeals for the 
D.C. Circuit (D.C. Circuit) issued its mandate vacating and remanding 
in part the original Determination \1\ issued by the Copyright Royalty 
Judges (Judges) in the captioned proceeding. See Johnson v. Copyright 
Royalty Board, 969 F.3d 363 (D.C. Cir. 2020). In its ruling on appeal, 
the D.C. Circuit found that in the original Determination, the Judges 
(1) failed to give adequate notice to participants of their overhaul of 
the royalty rate structure combined with significantly increased and 
uncapped rates for section 115 licenses; (2) failed to explain why they 
rejected a benchmark based on a past settlement agreement \2\ in lieu 
of overhauling of the rate structure and significantly increasing 
rates; and (3) failed to identify their legal authority to redefine a 
material term after they promulgated a definition of that term in the 
original Initial Determination circulated to the participants. See 
Johnson, 969 F.3d at 367, 381; Initial Determination, Determination of 
Royalty Rates and Terms for Making and Distributing Phonorecords 
(Phonorecords III), 16-CRB-0003-PR (2018-2022) (Jan. 27, 2018).
---------------------------------------------------------------------------

    \1\ Determination of Royalty Rates and Terms for Making and 
Distributing Phonorecords (Phonorecords III), 84 FR 1918 (Feb. 5, 
2019) (final rule and order) (original Determination); see also 
Final Determination, 16-CRB-0003-PR (2018-2022) (Nov. 5, 2018). The 
original Determination was issued by two of the Judges (Majority) 
and was accompanied by a dissenting opinion (Dissent) authored by 
the third Judge. The Dissent is appended to and part of the same 
document as the original Determination.
    \2\ The referenced settlement agreement formed the basis for 
regulatory terms relating to section 115 musical works royalties and 
was adopted as a final rule in Adjustment [or] Determination of 
Compulsory License Rates for Mechanical and Digital Phonorecords, 
Docket No. 2011-3 CRB Phonorecords II, 78 FR 67938 (Nov. 13, 2013). 
See also Technical Amendment at 78 FR 76987 (Dec. 20, 2013).
---------------------------------------------------------------------------

    After receipt of the D.C. Circuit's ruling and mandate, the Judges 
consulted with the parties to the appeal and established procedures for 
the remand proceeding. See Order Adopting Schedule for . . . Remand 
(Dec. 23, 2020).\3\ Each side offered opening submissions, responsive 
submissions, additional evidentiary filings, and further supplemental 
briefing requested by the Judges. The parties' submissions included 
legal briefing and incorporated evidence from the original proceeding 
as well as evidence newly developed for the remand proceeding. After 
preliminary deliberations, the Judges asked for supplemental briefing 
from the parties responsive to a proposed alternative rate structure. 
See Notice and Sua Sponte Order Directing the Parties to Provide 
Additional Materials (Dec. 9, 2021). With respect to redefinition of 
the material term Bundled Revenue, the Judges also sought legal 
analysis from the parties relating to the D.C. Circuit's directive that 
the Judges either provide ``a fuller explanation of the agency's 
reasoning at the time . . .'' or take ``new agency action accompanied 
by the appropriate procedures.'' See Johnson, 969 F.3d at 392 (citing 
Department of Homeland Security v. Regents of the Univ. of Cal., 140 S. 
Ct. 1891, 1908). On February 9, 2022, the Judges invited additional 
briefing on the Bundled Revenue definition issue, specifically 
permitting the parties to offer additional analysis of possible 
characterization of the Copyright Owners' motion for clarification 
following the Determination as a motion for rehearing under the 
Copyright Act, title 17, United States Code at sec. 803(c)(2). See Sua 
Sponte Order Regarding Additional Briefing (Feb. 9, 2022).
---------------------------------------------------------------------------

    \3\ Following the original remand scheduling order, the Judges 
amended the remand proceeding schedule by, e.g., permitting 
additional briefing, changing due dates, and seeking additional 
input with regard to specific issues. See, e.g., Order . . . 
Modifying Scheduling Orders (Dec. 13, 2021) (eCRB no. 25973).
---------------------------------------------------------------------------

    At the request of the parties, the Judges agreed to forego live 
testimony. On March 8, 2022, all parties were afforded an opportunity 
to present oral argument on all remand issues.\4\ On July 1, 2022, the 
Judges issued an Initial Ruling and Order after Remand (Initial Ruling) 
\5\--applying Johnson and considering the entire record developed pre-
remand and post-remand.
---------------------------------------------------------------------------

    \4\ Copyright Owners and Services divided the time for oral 
argument. George Johnson dba GEO Music Group waived oral argument.
    \5\ The Initial Ruling (eCRB no. 26938) is included in Related 
Rulings and Orders as section A. The findings and conclusions in the 
Initial Ruling were adopted by a majority of the Judges, but two 
Judges filed separate opinions. See Initial Ruling at 2 n.5. One 
Judge, former Chief Judge Suzanne Barnett, dissented from the 
Majority's conclusion in the Initial Ruling regarding the 
Phonorecords II rate structure (section II of the Initial Ruling), 
though not from the exception to that benchmark with regard to the 
headline rate of 15.1% and the imposition of a cap on the TCC rate 
prong. See Dissent in Part re Benchmark (July 1, 2022) (eCRB no. 
26943). The other opinion was issued by Judge Strickler, who 
dissented from the reasoning relating to the adoption of the 
definition of Service Revenue (section V), but concurred in the 
adoption of that definition. See Dissent in Part as to Section IV of 
the Initial Ruling and Order after Remand . . . (July 1, 2022) (eCRB 
no. 26965).
---------------------------------------------------------------------------

    In the Initial Ruling, the Judges directed the parties to attempt 
to submit jointly agreed-upon regulatory provisions implementing the 
Initial Ruling for the Judges to consider. The Judges further ruled 
that, if the parties could not agree on all the regulatory language, 
they should make separate submissions regarding regulatory provisions 
in dispute. See Initial Ruling at 114.
    The parties agreed to many regulatory provisions but disagreed as 
to several such provisions. Accordingly, they filed separate 
submissions and respective replies regarding the regulatory provisions. 
Services' Joint Submission of Regulatory Provisions (July 18, 2022); 
Copyright Owners' Submission of Regulatory Provisions to Implement the 
Initial Ruling (July 18, 2022); Services' Joint Response to Copyright 
Owners' Submission of Regulatory Provisions (Aug. 5, 2022); Copyright 
Owners' Response to Judges' July 27, 2022 Order Soliciting Responses 
Regarding Regulatory Provisions (Aug. 5, 2022).
    The Judges considered those submissions and entered an order 
addressing the disputed regulatory provisions. See Corrected Order 
regarding Regulatory Provisions

[[Page 54407]]

Following Initial Ruling and Order (after Remand) (Nov. 10, 2022) 
(November 10th Order).\6\
---------------------------------------------------------------------------

    \6\ The November 10th Order corrected an otherwise substantively 
identical order issued two days earlier, on November 8, 2022, which 
had inadvertently included a small amount of text. See November 10th 
Order at 1 (eCRB no. 27312).
---------------------------------------------------------------------------

    On November 30, 2022, the parties filed a Joint Submission in which 
they provided joint regulatory language no longer in dispute that 
applied the binding rulings of the Judges and the D.C. Circuit.\7\ 
However, the parties identified the single issue in dispute that 
relates to the ``Total Content Cost'' (``TCC'') rates for nine 
offerings made by interactive streaming services. Joint Submission . . 
. Regarding Regulatory Provisions Following Initial Ruling and Order 
(after Remand) (Nov. 30, 2022) (Joint Submission) (eCRB no. 27337).
---------------------------------------------------------------------------

    \7\ The Judges largely adopt the regulations in the Joint 
Submission, which reflect the substance of the Judges' post-remand 
rulings, the substance and formatting that the Judges had adopted in 
the pre-remand Final Determination that were not raised as issues on 
appeal, and updates to references to subparagraphs of Section 115 to 
conform to statutory amendments made pursuant to the Music 
Modernization Act in 2018. Any differences in language or style are 
made for ease of reference, consistent with the parties' post-remand 
joint filings.
---------------------------------------------------------------------------

    Having considered the parties' submissions (including the Joint 
Submission), the Initial Ruling, and all other pertinent material, the 
Judges adopted the several TCC rates set forth in the Phonorecords II-
based benchmark as proposed by the Services. See Order 43 on 
Phonorecords III Regulatory Provisions (eCRB no. 28210).\8\
---------------------------------------------------------------------------

    \8\ The Judges also found good cause to adopt a joint proposal 
for modified language regarding late fees, in 37 CFR 385.3. Order 43 
on Phonorecords III Regulatory Provisions at 9.
---------------------------------------------------------------------------

    Based on the entirety of the record, the Judges adopt in toto \9\ 
the Initial Ruling and the Order 43 on Phonorecords III Regulatory 
Provisions which are set out in this document. Accordingly, those two 
documents are adopted by reference in this Final Determination After 
Remand. Additionally, the regulatory terms that will codify this Final 
Determination After Remand are set out in this document.\10\
---------------------------------------------------------------------------

    \9\ But see Judge Strickler's Dissent, cited at n.5 supra, in 
which--although he agrees with the Majority as to the definition of 
a Service Revenue Bundle--he disagrees as to the legal reasoning 
supporting that conclusion.
    \10\ The documents are: Initial Ruling and Order After Remand, 
designated as Related Rulings and Orders, section A; Order 43 on 
Phonorecords III Regulatory Provisions, designated as Related 
Rulings and Orders, section B; Dissent in Part as to Section IV of 
the Initial Ruling and Order after Remand by Judge David R. 
Strickler, designated as Related Rulings and Orders, section C; and 
Dissent in Part re Benchmark, designated as Related Rulings and 
Orders, section D.
---------------------------------------------------------------------------

    On the basis of the foregoing, the Judges propound the rates and 
terms described in this Final Determination After Remand for the period 
January 1, 2018, through December 31, 2022.\11\ No participant having 
filed a timely petition for rehearing, the Judges have made no 
substantive alterations to the body of the Initial Determination After 
Remand. The Register of Copyrights reviewed the Judges' Final 
Determination After Remand for legal error in resolving a material 
issue of substantive law under title 17, United States Code, and has 
closed her review. Non-substantive typos have been corrected and non-
substantive formatting changes have been made to the version reviewed 
by the Register in order to accommodate the Federal Register's 
formatting standards. The Librarian shall cause the Judges' Final 
Determination After Remand, and any correction thereto by the Register, 
to be published in the Federal Register no later than the conclusion of 
the Register's 60-day review period.
---------------------------------------------------------------------------

    \11\ The regulations applicable to the period 2018 through 2022, 
as set forth following this SUPPLEMENTARY INFORMATION section, will 
appear in the CFR as appendix A to the current regulations. Although 
these Phonorecords III regulations adopt the substance of the 
Phonorecords II-based benchmark where the Judges so require, in 
Sec. Sec.  385.21 and 385.22, these Phonorecords III regulations are 
structured, consistent with the parties' Joint Submission, in the 
same consolidated manner as set forth in the pre-remand Phonorecords 
III regulations (a structure as to which no party appealed). See 
Exhibit A to the Joint Submission at 16, n. 47; see also Exhibit B 
to the Joint Submission at n.17 (red-lined version of Exhibit A, 
supra).
---------------------------------------------------------------------------

Related Rulings and Orders

A. Initial Ruling and Order After Remand (Redacted Version With Federal 
Register Naming and Formatting Conventions)

    On October 26, 2020, the United States Court of Appeals for the 
D.C. Circuit (D.C. Circuit) issued its mandate vacating and remanding 
in part the Determination \12\ issued by the Copyright Royalty Judges 
(Judges) in the captioned proceeding. See Johnson v. Copyright Royalty 
Board, 969 F.3d 363 (D.C. Cir. 2020). In its ruling on appeal, the D.C. 
Circuit found that in the Determination, the Judges (1) failed to give 
adequate notice to participants of their overhaul of the royalty rate 
structure combined with significantly increased and uncapped rates for 
section 115 licenses; (2) failed to explain why they rejected a 
benchmark based on a past settlement agreement \13\ in lieu of 
overhauling of the rate structure and significantly increasing rates; 
and (3) failed to identify their legal authority to redefine a material 
term after they promulgated a definition of that term in the Initial 
Determination circulated to the participants. See Johnson, 969 F.3d at 
367, 381; Initial Determination, Determination of Royalty Rates and 
Terms for Making and Distributing Phonorecords (Phonorecords III), 16-
CRB-0003-PR (2018-2022) (Jan. 27, 2018).
---------------------------------------------------------------------------

    \12\ Determination of Royalty Rates and Terms for Making and 
Distributing Phonorecords (Phonorecords III), 84 FR 1918 (Copyright 
Royalty Board Feb. 5, 2019) (final rule and order) 
(``Determination''); See also Final Determination, 16-CRB-0003-PR 
(2018-2022) (Nov. 5, 2018) (citations to the Determination and to 
the Dissent in this Initial Ruling and Order after Remand (Initial 
Ruling) are found in this document). The Determination was issued by 
two of the Judges (Majority) and was accompanied by a dissenting 
opinion (Dissent) authored by the third Judge. The Dissent is 
appended to and part of the same document as the Determination.
    \13\ The referenced settlement agreement formed the basis for 
regulatory terms relating to section 115 musical works royalties and 
was adopted as a final rule in Adjustment of Determination of 
Compulsory License Rates for Mechanical and Digital Phonorecords, 
Docket No. 2011-3 CRB Phonorecords II, 78 FR 67938 (Nov. 13, 2013), 
Technical Amendment at 78 FR 76987 (Dec. 20, 2013). In this Initial 
Ruling, references to Phonorecords II, PR II, and PR II-based 
benchmark are references to this final rule.
---------------------------------------------------------------------------

    After receipt of the D.C. Circuit's ruling and mandate, the Judges 
consulted with the parties to the appeal and established procedures for 
the remand proceeding. See Order Adopting Schedule for . . . Remand 
(Dec. 23, 2020).\14\ Each side offered opening submissions, responsive 
submissions, additional evidentiary filings and further supplemental 
briefing requested by the Judges. The parties' submissions included 
legal briefing and incorporated evidence from the original proceeding 
as well as evidence newly developed for the remand proceeding. After 
preliminary deliberations, the Judges asked for supplemental briefing 
from the parties responsive to a proposed alternative rate structure. 
See Notice and Sua Sponte Order Directing the Parties to Provide 
Additional Materials (Dec. 9 Order). The Judges also sought legal 
analysis from the parties relating to the D.C. Circuit's directive that 
the Judges either provide ``a fuller explanation of the agency's 
reasoning at the time . . .'' or take ``new agency action accompanied 
by the appropriate procedures.'' See Johnson, 969 F.3d at 392 (citing 
Dep't of Homeland Sec. v. Regents of the Univ. of Cal., 140 S. Ct. 
1891, 1908 (Regents)). On February 9, the Judges invited additional 
briefing on

[[Page 54408]]

the service bundle definition issue, specifically permitting the 
parties to offer additional analysis of possible characterization of 
the Copyright Owners' motion for clarification following the 
Determination as a motion for rehearing under the Copyright Act, title 
17, United States Code (Act) at sec. 803(c)(2).
---------------------------------------------------------------------------

    \14\ Following the original remand scheduling order, at the 
request of parties or on their own motion, the Judges amended the 
remand proceeding schedule by, e.g., permitting additional briefing, 
changing due dates, and seeking additional input with regard to 
specific issues. See, e.g., Order . . . Modifying Scheduling Orders 
(Dec. 13, 2021).
---------------------------------------------------------------------------

    At the request of the parties, the Judges agreed to forego live 
testimony. On March 8, 2022, all parties were afforded an opportunity 
to present oral argument on all remand issues.\15\ Following oral 
argument, the Judges deliberated and now issue this Initial Ruling 
after Remand.
---------------------------------------------------------------------------

    \15\ Copyright Owners and Services divided the time for oral 
argument. George Johnson dba GEO Music Group waived oral argument.
---------------------------------------------------------------------------

    After due consideration of all of the evidence and oral argument of 
counsel, the Judges \16\ determine: \17\
---------------------------------------------------------------------------

    \16\ The findings and conclusions in this Initial Ruling are 
adopted by a majority of the Judges. One Judge dissents from the 
adoption of the entirety of the Phonorecords II rate structure 
(section II), though not from the exception to that benchmark with 
regard to the headline rate of 15.1% and the imposition of a cap on 
the TCC rate prong. One Judge dissents in part from the reasoning 
relating to adoption of the definition of Service Revenue (section 
V), but not from the adoption of that definition.
    \17\ As addressed infra, the Judges also order that the 
participants in this remand proceeding prepare and submit regulatory 
provisions consistent with this ruling. See Footnote 174.
---------------------------------------------------------------------------

    (1) With regard to the applicable rates and rate structure, the 
percent-of-revenue all-in headline royalty rate for the mechanical 
license shall be set at 15.1%, phased-in, as set forth below:

                                     2018-2022 All-In Headline Royalty Rates
----------------------------------------------------------------------------------------------------------------
                                                2018          2019          2020          2021          2022
----------------------------------------------------------------------------------------------------------------
Percent of Revenue........................        11.4%         12.3%         13.3%         14.2%         15.1%
----------------------------------------------------------------------------------------------------------------

    In all other respects, the rates and rate structure of the 
Phonorecords II-based benchmark proposed by the Services (as that 
benchmark is defined herein) shall constitute the rates and rate 
structure for the Phonorecords III period.\18\
---------------------------------------------------------------------------

    \18\ The Services include in their Joint Rate Proposal a chart 
summarizing the proposed rates for their offerings. That chart is 
attached as an Addendum to this Initial Ruling.
---------------------------------------------------------------------------

    To be clear: the 15.1% headline percentage rate substitutes for the 
headline percentage rates in subparts B and C of the Services 
Phonorecords II-based benchmark, and the definition of ``Service 
Revenue'' for bundles shall be the definition contained in 37 CFR 
385.11 (paragraph (5) for the ``Service Revenue'' definition) as 
proposed in the Services' Phonorecords II-based benchmark.
    (2) The Services' Phonorecords II-based benchmark is the better of 
the benchmarks proposed by the parties and satisfies the requirements 
of 17 U.S.C. 801(b)(1) in all respects. However, as noted supra, to be 
consistent with this statutory section and the decision in Johnson, the 
royalty rate of 10.5% in that benchmark shall be replaced with the 
15.1% rate set forth in paragraph (1) above.
    (3) To reiterate for clarity, consistent with the adoption of the 
Phonorecords II-based benchmark, and for the reasons more fully 
developed herein, the Judges adopt the definition of ``Service Revenue 
for Bundled Services'' as it appeared in the Initial Determination in 
the underlying proceeding. Following are the Judges' analysis and 
ruling after remand.

I. Preliminary Issue: Burden of Proof

    As a preliminary matter, the Judges address the issue of burden of 
proof raised by both parties. Pursuant to the Administrative Procedure 
Act (APA), ``the proponent of a rule or order has the burden of 
proof.'' 5 U.S.C. 556(d). See also Initial Remand Submission of 
Copyright Owners at 48 (Apr. 1, 2021) (``CO Initial Submission'') 
(citing section 556(d) of the APA as setting forth ``a basic rule of 
these rate-setting proceedings that a participant is required to 
provide evidence establishing the propriety of all aspects of its own 
proposed rates and terms, including all aspects of the participant's 
proposed rate structure.''). Accordingly, it is clear to the Judges 
that the Services should continue to bear the burden of proof regarding 
the sufficiency of their proffered Phonorecords II-based benchmark in 
this remand proceeding. And, in like fashion, because on remand 
Copyright Owners have assumed the mantle of pursuing the vacated rate 
structure and rates, they bear the burden of proof with regard to their 
proposal.
    However, Copyright Owners assert that it is the Services who bear 
the burden of proof as to Copyright Owners' proposal regarding the 
appropriateness, vel non, of an uncapped TCC rate prong. According to 
Copyright Owners, this burden falls on the Services because ``only the 
Services . . . proposed TCC prongs at the hearing,'' in the form of the 
mix of capped and uncapped TCC prongs contained in the Services' 
Phonorecords II benchmark. Id. at 47. The Judges find that the fact 
that the Phonorecords II-based benchmark advanced by the Services 
contains this mix of capped and uncapped TCC prongs does not bear on 
Copyright Owners' duty, under 5 U.S.C. 556(d), to satisfy the burden of 
proof with regard to the rates and rate structure they are advancing on 
this remand. Moreover, the D.C. Circuit has already held that the fact 
that some of the Streaming Services' proposals contemplated continued 
use of an uncapped total content cost prong for some categories ``does 
not mean they anticipated that the [Judges] would uncap the total 
content cost prong across the board . . . [which] is quite different.'' 
Johnson, 369 F.3d at 382. The difference, according to Johnson, is that 
``[u]ncapping the total content cost prong across all categories leaves 
the Streaming Services exposed to potentially large hikes in the 
mechanical license royalties they must pay.'' Id.
    Accordingly, the Judges find that Copyright Owners indeed do bear 
the burden of proof with regard to the appropriateness of uncapped rate 
structure and rates they are proposing on remand and the Services bear 
the burden of proof with regard to the appropriateness of the 
Phonorecords II-based benchmark they are continuing to advance on 
remand.

II. Rate Structure and Rates

A. Relevant Rulings in Johnson

    In establishing a royalty rate structure and the rates within it in 
the context of this remand proceeding, the Judges are guided by the 
rulings in Johnson.
1. Percent of Revenue Prong
    The D.C. Circuit noted that the Judges found the royalties in the 
Phonorecords II period were too low and that record companies were 
receiving a disproportionate share of the sum of the mechanical and 
sound recording royalties. Johnson, 969 F.3d at 384-85. The D.C. 
Circuit acknowledged that ``[t]he Judges . . . then carefully

[[Page 54409]]

analyzed the competing testimony and drew from it rates that were 
grounded in the record and supported by reasoned analysis.'' Id. at 
385. The D.C. Circuit found that the Judges acted well within their 
discretion and not arbitrarily, relying on substantial evidence in 
establishing the ``zone of reasonableness'' for the rates. Id. As the 
D.C. Circuit noted, the Judges' process was ``the type of line-drawing 
and reasoned weighing of the evidence [that] falls squarely within the 
[Judges'] wheelhouse as an expert administrative agency.'' Id. at 385-
86 (emphasis added).
2. Uncapped TCC Prong
    The D.C. Circuit found fault, however, in the Judges' determination 
to establish an uncapped and increased percentage-based total content 
cost (TCC).\19\ Id. at 380. This approach ``removed the only structural 
limitation on how high the [TCC] . . . can climb.'' Id. The D.C. 
Circuit reasoned that uncapping the TCC alternative rate prong across 
all categories of service exposed the Services to potentially large 
hikes in the overall mechanical royalties they must pay. Id. at 382. 
The D.C. Circuit noted: ``As the [Judges] acknowledge, sound recording 
rightsholders have considerable market power vis-[agrave]-vis 
interactive streaming service providers . . . . The interactive 
streaming services are . . . exposed to the labels' market power and 
record companies could, if they so chose, put those services out of 
business entirely . . . . [B]y virtue of their oligopoly power, the 
sound recording copyright holders have extracted `inflated' royalties. 
. . .'' Id. (cleaned up).
---------------------------------------------------------------------------

    \19\ ``TCC'' refers to ``Total Content Cost,'' and is defined as 
``a percentage of the royalties paid by the service . . . to sound 
recording copyright holders.'' Johnson, 969 F.3d at 370; see also 
Determination at 13 n.38 (``TCC'' is an industry acronym for ``Total 
Content Cost'', a shorthand reference to the extant regulatory 
language describing generally the amount paid by a service to a 
record company for the section 114 right to perform digitally a 
sound recording.'').
---------------------------------------------------------------------------

    While the Services had advocated uncapping the TCC alternative rate 
prong for some categories of service, that ``does not mean they 
anticipated that the [Judges] would uncap the total content cost prong 
across the board. That is quite different.'' Id. at 382. The D.C. 
Circuit found that the Judges ``failed to provide adequate notice of 
the drastically modified rate structure [they] ultimately adopted.'' 
Id. at 381. The D.C. Circuit emphasized that the failure to provide 
adequate notice of their intentions ``is no mere formality [because] 
[i]nterested parties' ability to provide evidence and argument . . . 
not only protects the parties' interests, it also helps ensure that the 
[Judges'] ultimate decision is well-reasoned and grounded in 
substantial evidence.'' Id. at 381-82.
    To support their adoption of an uncapped TCC rate prong, the Judges 
``predicted that the sound recording copyright owners' royalty rates 
would naturally decline in the course of their negotiations with 
interactive streaming services.'' Id. at 372. The Judges found 
persuasive the rebuttal testimony of one of Copyright Owners' economic 
expert witnesses, Professor Watt, that an increase in mechanical 
royalties payable by the Services would lead to a corresponding 
decrease in the Services' sound recording royalty obligations. See 
Determination at 73-74 (``[S]ound recording royalty rates in the 
unregulated market will decline in response to an increase in the 
compulsory license rate for musical works [and] Professor Watt's 
bargaining model predicts that the total of musical works and sound 
recordings royalties would stay ``almost the same'' in response to an 
increase in the statutory royalty.''). The Services painstakingly 
criticized this ``see-saw'' theory.
    The D.C. Circuit concluded that, on remand, if and when the Judges 
consider the ``uncapped'' rate structure, they shall address all 
substantive challenges to that approach raised by the Services, 
including the issue of whether ``an increase in mechanical license 
royalties would lead to a decrease in sound recording royalties.'' Id. 
at 383.
    Thus, the D.C. Circuit held, the Judges erred procedurally in 
adopting an uncapped TCC alternative rate prong. The D. C. Circuit 
therefore instructed the Judges to provide the parties with the 
opportunity to fully address the issues regarding the uncapped TCC 
prong, and for the Judges to address the ``substantive challenges'' 
raised by the Services.
3. Four Itemized Statutory Objectives
    The statutory standard found in section 801(b)(1) instructs the 
Judges to set rates that are not only ``reasonable,'' but also 
reflective of four itemized objectives, or factors, which, as the D.C. 
Circuit stated, set forth ``competing priorities.'' 17 U.S.C. 
801(b)(1)(A)-(D); Johnson, 969 F.3d at 387.\20\ With regard to these 
four priorities, the D.C. Circuit found that the Judges properly 
analyzed and applied the first objective (Factor A). Id. at 387-88. In 
particular, the D.C. Circuit did not disturb the Judges' ruling that an 
increase in the royalty rates for mechanical licenses was necessary in 
order to satisfy Factor A. Johnson, 369 F.3d at 387-88. According to 
Johnson, in making this finding, the Judges had engaged in a 
``reasonable reading of the record'' and had relied on ``substantial 
evidence.'' Id. at 388. Thus, Factor A (when considered without regard 
to the other three objectives) indicated that the statutory rate needed 
to be higher than it was during the Phonorecords II period.\21\
---------------------------------------------------------------------------

    \20\ These competing objectives are: (A) To maximize the 
availability of creative works to the public; (B) To afford the 
copyright owner a fair return for his or her creative work and the 
copyright user a fair income under existing economic conditions; (C) 
To reflect the relative roles of the copyright owner and the 
copyright user in the product made available to the public with 
respect to relative creative contribution, technological 
contribution, capital investment, cost, risk, and contribution to 
the opening of new markets for creative expression and media for 
their communication; and (D) To minimize any disruptive impact on 
the structure of the industries involved and on generally prevailing 
industry practices. Id.
    \21\ However, as the D.C. Circuit also noted, because the four 
section 801(b)(1) objectives reflect ``competing priorities, id'' at 
387, the holding that Factor A militates toward a higher rate is not 
ultumately dispositive. Rather, it must be weighed with the other 
statutory factors.
---------------------------------------------------------------------------

    With regard to the other three objectives, Johnson stated that 
``[t]he question whether the [Judges] adequately addressed factors B 
through D . . . is intertwined with the nature of the rate structure 
ultimately imposed by the [Judges].'' Id. at 389. Accordingly, the D.C. 
Circuit concluded that it ``need not . . . address whether the [Judges] 
adequately considered these remaining factors.'' Id.\22\
---------------------------------------------------------------------------

    \22\ The phrase ``intertwined with the nature of the rate 
structure'' requires emphasis because the Majority independently 
considered how to weigh Factors B and C specifically as to the 15.1% 
revenue rate, without regard to the overall rate structure, as 
discussed infra.
---------------------------------------------------------------------------

    Within the parameters of the holdings in Johnson, the Judges 
consider the record facts and the arguments made in this remand 
proceeding, together with the pertinent facts and arguments made in the 
original proceeding.

B. Rate Evidence for the 33-Months From January 2018 Through September 
2020

    After the Determination was issued, from its effective inception on 
January 1, 2018, through September 30, 2020--a 33-month period--the 
parties operated under the rates and rate structure set forth in that 
ruling. In light of the D.C. Circuit's decision in Johnson, as of 
October 1, 2020, the parties reverted to the Phonorecords II rates. The 
Services have asserted in this remand proceeding that, during the 33-
month period when the Majority's new and higher

[[Page 54410]]

Phonorecords III rates were in effect, [REDACTED]. By contrast, 
Copyright Owners, on remand, looking at the same data over this 33-
month period, aver that they prove the existence of the seesaw theory.
1. Services' Position
    According to the Services, [REDACTED]. [REDACTED]. Moreover, 
according to the Services, [REDACTED]. The Services further maintain 
that, [REDACTED].
    The Services make the [REDACTED]. And, [REDACTED]. Id. ]] 5, 9-13, 
16-19, 22-23, 26-27.
    The Services claim that [REDACTED]. More particularly, [REDACTED].
    [REDACTED]. [REDACTED].
    The Services' economic experts rushed to judgment upon learning of 
these facts, claiming that they disproved the seesaw theory. See Katz 
WDRT ]] 25-27 (relying on testimonies cited supra and concluding that 
seesaw theory was disproved, based on [REDACTED]); Marx WDRT ]] 48-51 
(relying on same testimonies and likewise finding because [REDACTED]); 
Leonard WDRT ] 17. ([REDACTED]).
2. Copyright Owners' Position
    Copyright Owners analyzed the royalty data over the same 33-month 
period (January 2018 through September 2020) and reach the opposite 
conclusion. One of their economic expert witnesses, Dr. Jeffrey 
Eisenach, testified that [REDACTED]. Moreover, he opined that 
[REDACTED]. See Eisenach RWRT sec. 2(A) & appx. C.
    Based on this analysis, Professor Watt declares empirical 
vindication of his seesaw theory. Watt RWRT ]] 41-42, 46 (``The 
[Judges'] bargaining theory insights about the relationship between 
royalty rates were correct. . . . [REDACTED]. . . .'').
3. Analysis and Decision Regarding Evidence of Post-Determination Rates
    The Judges are perplexed by the willingness of the expert economic 
witnesses on both sides to opine that the rate changes from January 
2018 through September 2020 can serve as confirmation of their clients' 
respective positions. The issue to be considered empirically was 
whether the sound recording rate would decrease in response to the 
increase in the mechanical rate. That is, if the record labels had 
previously set royalties at a level that would allow the Services 
merely to survive, would the record labels agree to lower their sound 
recording rate if more of the Services' surplus were acquired by 
Copyright Owners? To answer this question, the economists on both sides 
applied sophisticated bargaining models and critiques to explain the 
nature of the negotiations that would ensue.
    In the process, the economists lost track of an obvious, elementary 
point: The Phonorecords III rates were being challenged by the 
Services' appeal, and might not persist. Indeed, the rates were 
ultimately vacated and the parties returned in October 2020 to the 
Phonorecords II rates.\23\ Now, the rates will be changed again by this 
post-remand Determination, and going forward may be subject to further 
potential change, consistent with the provisions of title 17. In light 
of such ongoing fundamental uncertainty, why would any economist or 
businessman assume that the sound recording companies would agree to 
adjust their rates in response to a change in the mechanical rate? The 
Judges are amazed that the economic experts neglected even to raise 
this uncertainty as a complicating issue, let alone a dispositive 
one.\24\
---------------------------------------------------------------------------

    \23\ There also was uncertainty as to the effective inception 
date of the Phonorecords III rate period, because the Services had 
appealed (ultimately unsuccessfully) the CRB Judges' finding that 
the period commenced, retroactively, as of January 1, 2018.
    \24\ To place this point in the economic context of this 
proceeding, the Judges characterize the ongoing ``legal 
uncertainty'' as another ``independent variable'' to add to the 
economic experts' list of such variables, discussed infra, that 
affect the ``dependent variable,'' viz., the sound recording rate.
---------------------------------------------------------------------------

    Moreover, no party called as a witness any representatives of the 
Majors, or subpoenaed their testimony or documents, to provide the 
Judges with evidence of how these record companies perceived the seesaw 
issue, whether as a permanent phenomenon or as an uncertain matter, 
given the pendency of the legal proceedings regarding the ultimate 
mechanical rate. Any of the parties could have requested that the 
Judges subpoena a sound recording industry witness to give testimony 
and produce documents as to this issue, pursuant to 17 U.S.C. 
803(b)(6)(C)(ix), but none did so. Further, Copyright Owners, who are 
representing the music publishing interests of inter alios, Sony, 
Universal, Warner, and Merlin, likely could have produced such sound 
recording witnesses without the need for a subpoena. Witnesses from 
these entities who negotiated with the Services after the Phonorecords 
III rates and rate structure became effective certainly would have 
knowledge relevant to the testimony of the Services' witnesses 
[REDACTED] who claimed that [REDACTED].
    Simply put, the period from period from January 2018 through 
September 2020 was a time the Judges construe as ``33-months of 
uncertainty,'' see 3/8/22 Tr. 87, 91 (Closing Argument) when no party 
could ascertain with any assuredness the ultimate Phonorecords III 
rates and rate structure. Thus, for the economists and the parties to 
claim vindication for their arguments by reliance on how the record 
labels did or did not respond to the challenged and ever-shifting rates 
during this ``33 months of uncertainty'' reflects the elevation of 
adversarial zeal over objective judgment.
    Accordingly, the Judges place no weight on the purported changes or 
stability of the sound recording rates during the Phonorecords III rate 
period.

C. Percent-of-Revenue Rate Prong

1. Copyright Owners' Position
    In their initial remand submission, Copyright Owners provided no 
new evidence to support any aspect of the 15.1% revenue-based rate (or 
for that matter, any new evidence to support the rates or rate 
structure in the Determination), and elected to rely on the pre-remand 
record. In fact, in their initial remand submission, Copyright Owners 
do not so much as mention the 15.1% revenue rate derived by the Judges. 
However, in their reply remand submission (which the Judges found also 
to constitute, in part, a substantive initial submission \25\) 
Copyright Owners do address the 15.1% revenue rate. In the reply 
submission, Copyright Owners simply stated: ``[T]he Circuit affirmed 
the Board's derivation of rate percentages, including raising the 
revenue rate to 15.1%.'' Copyright Owners' Reply Brief on Remand (in 
Reply Remand Submission of Copyright Owners, Vol. 1) at 64, n.48 (July 
2, 2021) (``CO Reply''). In a subsequent submission, Copyright Owners 
added that ``[t]he narrow mandate on this Remand does not allow for 
reopening the rate percentage determination in the [ ]Determination.'' 
Copyright Owners' Motion for Reconsideration or Clarification at 15 & 
n.10 (Dec. 17, 2021) (emphasis added) (Dec. 17th Motion).
---------------------------------------------------------------------------

    \25\ See Order Denying in Part and Granting in Part Services' 
Motion to Strike Copyright Owners' Expert Testimony and Granting 
Services' Request to File Supplemental Testimony and Briefing at 11 
(Oct. 1, 2021) (Oct. 1st Order) (The Judges found that ''with one 
exception . . . the challenged testimonial evidence of Copyright 
Owners' economic expert witnesses serve the dual purposes of direct 
and rebuttal statements'' and, as a consequence, ``provide[d] the 
Services an opportunity to file supplemental testimony and briefing 
in opposition.
---------------------------------------------------------------------------

    Thereafter, Copyright Owners asserted that the D.C. Circuit's 
affirmance of the

[[Page 54411]]

[Judges'] revenue percentage rate calculation was ``strong[ ]'' and 
``detailed.'' Copyright Owners' Reply in Further Support of Motion for 
Reconsideration or Clarification at 4 (January 5, 2022). Moreover, 
Copyright Owners took note that the Services had relied on 
substantively identical language in Johnson to support their argument 
that other statements in that D.C. Circuit decision should be deemed 
affirmed. See id. at 4-5 (noting Services' reliance on Johnson's 
description of the Judges' rulings regarding student and family 
discounts (``grounded in substantial record evidence . . . based on the 
weight and credibility of the evidence [and] squarely within the 
Judges' expertise)'' as demonstrating that the D.C. Circuit had 
affirmed those rulings) (emphasis added); see also Copyright Owners' 
Brief in Response to the Additional Materials Orders at 2, 6-7 (Jan. 
24, 2022) (``CO Additional Submission'') (again asserting that ``the 
15.1% revenue rate . . . was specifically affirmed in detail by 
Johnson.'').
2. Services' Position
    In their initial submission after the remand, the Services objected 
to any continued application by the Judges of the 15.1% revenue rate 
because, ``as the Majority acknowledged, this particular division of 
revenues will never happen in the real world because of the 
complementary oligopoly power of the record labels.'' Services' Joint 
Opening Brief (in Services' Joint Written Direct Remand Submission at 
Tab D) at 52 (``Services' Initial Submission'') (Apr. 1, 2021). More 
particularly in this regard, the Services note that Professor Marx's 
Shapley Value Model,\26\ which served as an input for the generation of 
the 15.1% revenue rate, also indicated that only [REDACTED]% of the 
interactive streaming revenue should be paid out as royalties to the 
sound recording rightsholders, with the remaining [REDACTED]% of these 
revenues retained by the interactive streaming services. Id. (``Both 
Professor Marx's and Professor Watt's models show lower combined 
royalties being paid by the services than are currently paid in the 
marketplace. . . The discrepancy in total royalties between the models 
and the real world is explained, in part, by the absence of supranormal 
complementary oligopoly profits in the Shapley model, and the presence 
of those profits in the actual market.''). Id. (quoting Phonorecords 
III, 84 FR 1952).
---------------------------------------------------------------------------

    \26\ Generally, a Shapley Value Model is a game theory analysis. 
It models a hypothetical bargain that assigns each ``player'' the 
average marginal value it contributes to the bargain and (after 
accounting for the costs that each ``player'' would need to recover) 
the remaining ``surplus'' is allocated among the players according 
to their relative contributions. See Johnson, 969 F.3d at 372. For 
the reasons discussed infra, in the present case, the Shapley 
surplus from the streaming revenue is split essentially equally by 
the owners of the sound recording and musical works owners inter se, 
but the royalty rates themselves that would result from their 
bargaining would be different as between these two inputs, because 
of their differing costs. See, e.g., Gans WDT ] 73.
---------------------------------------------------------------------------

    By this approach, the Services maintain, ``the Majority awarded the 
Copyright Owners the full 15.1% of revenue dictated by its model 
(phased in over time), and left it up to the Services to convince the 
complementary oligopolist major labels to dramatically lower sound 
recording rates.'' Id. at 54-55. The Services argue that, instead, the 
Majority should have applied to Professor Marx's [REDACTED]% total 
royalty obligation what they characterize as ``any of the[ ] real-world 
ratios in place of the [REDACTED] ratio taken from ``Professor Gans' 
``Shapley-inspired'' model. Id. at 54. According to the Services, these 
lower ratios would have reduced the revenue percentage rate well below 
15.1%. Id.
    Alternatively, the Services propose, through Professor Marx's post-
remand written testimony, that the Judges now adopt ``a more balanced, 
burden-sharing approach'' to address what she described as the 
Majority's ``imbalance'' problem. Id. at 57; see also Marx WDRT ]] 52-
63.\27\ Essentially, her proposal begins with an assumption, based on 
record evidence, that labels typically take specific shares of service 
revenue, including shares of [REDACTED]%, [REDACTED]% and 
[REDACTED]%.\28\ These shares are significantly higher than the 
[REDACTED]% that Professor Marx generated from her Shapley model. Next, 
Professor Marx's post-remand burden-sharing approach uses as inputs the 
15.1% of service revenue and the [REDACTED]% of service revenue that 
would be retained by the musical works owners and the Services 
respectively.\29\ Putting these two factors together, she sets forth 
the basic math: Using her [REDACTED]% sound recording share as an 
example, she notes that there is not enough revenue for the labels to 
take this [REDACTED]% share, if the musical works owners also receive 
15.1% and the Services also retain the [REDACTED]% derived from her 
model ([REDACTED]% + 15.1% + [REDACTED]% = [REDACTED]%, an irrational 
result). See Services' Joint Opening Brief at 57.
---------------------------------------------------------------------------

    \27\ Claiming consistency with the Majority's analysis, 
Professor Marx appears to maintain that her ``burden-sharing'' 
approach generates the statutorily-required ``reasonable'' rate as 
well as a rate that satisfies the ``fair return''/``fair income'' 
objectives of statutory Factor B. See Marx WDRT ] 52 (introducing 
her correction of the alleged ``imbalance'' problem by noting that 
``the ``right'' mechanical royalty rate is one that is 
``reasonable'' and achieves the four objectives laid out in Section 
801(b)(1).''
    \28\ See Marx WDRT, fig. 7 ([REDACTED]).
    \29\ The [REDACTED]% of revenue that the services would retain 
is based on one of Professor Marx's ``Shapley Value Models.'' 
Shapley Value modeling is discussed infra.
---------------------------------------------------------------------------

    Professor Marx engages in an analysis based on the following math 
and logic (again, using the [REDACTED]% sound recording rate as an 
example of the fixed amount taken by the labels): (1) [REDACTED]% of 
the streaming revenues remain available to be split between the 
services and the musical works copyright owners; (2) adding the 15.1% 
revenue rate and her [REDACTED]% revenue retention percentage equals 
[REDACTED]%; and (3) the 15.1% revenue rate, as a percent of this 
[REDACTED]%, is [REDACTED]%; and (4) [REDACTED]% of the [REDACTED]% 
available for splitting between the services and the musical works 
copyright owners is [REDACTED]% (rounded). Id. at fig.8.
    Thus, she identifies her version of a ``fair'' result: The Services 
and Copyright Owners would split the residual revenue remaining after 
the labels have exercised their complementary oligopoly power to take 
an outsized fixed share--with the split proportional to the 15.1%-to-
[REDACTED]% revenue amounts calculated respectively by the Judges (the 
15.1% musical works rate) and Professor Marx (the [REDACTED]% service 
revenue retention). Id. 59, table. 8.\30\
---------------------------------------------------------------------------

    \30\ Using the same logic and calculation method, Professor Marx 
finds that the services would retain [verbar][REDACTED]% / 
[REDACTED]%, which equals [verbar][REDACTED]%. Assuming again that 
[REDACTED]% of the steaming revenue is available to split (because 
the labels have appropriated [REDACTED]%), the services would retain 
[REDACTED]% [REDACTED]% rounded) of the streaming revenue. Id.
---------------------------------------------------------------------------

    In their final post-remand submission, the Services also flatly 
state: ``[T]he D.C. Circuit did not ``affirm'' the 15.1% rate--it 
vacated that rate.'' Services' Joint Rebuttal Brief Addressing the 
Judges' Working Proposal at 2 (Feb. 24, 2022) (``Services' Additional 
Submission''). However, the Services do not support that quoted 
statement with any citation to Johnson. See id. Further, the Services 
assert that the 15.1% revenue rate is not immune from post-remand 
review and reduction because ``the D.C. Circuit withheld judgment ``on 
whether that final percentage satisfies factors B through D of Section 
801(b)(1). . . .'' Id. at 3.

[[Page 54412]]

3. Analysis and Decision Regarding 15.1% Revenue Rate Prong
    The Judges determine that they are clearly bound by the D.C. 
Circuit's decision in Johnson to maintain the 15.1% revenue rate, as 
phased-in by the Determination. Several reasons support this decision.
    First, the Judges conclude that the D.C. Circuit's decision in 
Johnson is conclusive and unambiguous regarding the revenue percentage 
rate. The D.C. Circuit rejected the Services' assertion that the Judges 
acted ``arbitrarily'' as to this particular issue, noting that the 
Services had misstated the relevant facts. Johnson, 969 F.3d at 385-86 
(responding to Services' misdescription of Judges' analysis and 
explaining what Services described as ``not what happened.''). 
Moreover, the D.C. Circuit held that with regard to the construction of 
the 15.1% revenue rate, the Judges had ``engaged in the type of line-
drawing and reasoned weighing of the evidence [which] falls squarely 
within the [Judges'] wheelhouse as an expert administrative agency.'' 
Id. at 386. The D.C. Circuit further noted that the Judges 
``proceed[ed] cautiously'' to set the 15.1% revenue rate by 
establishing a ``zone of reasonableness'' for the revenue rate. Id. at 
385. Indeed, with regard to each aspect of this revenue rate analysis, 
the D.C. Circuit found that the Judges' decision making was ``grounded 
in the record and supported by reasoned analysis'' and that 
``[s]ubstantial evidence supports [their] judgment.'' Id. at 385.
    Second, when the D.C. Circuit reviewed the Determination, it 
applied ``the same standards set forth in the Administrative Procedure 
Act, 5 U.S.C. 706.'' Id. at 375 (noting that 17 U.S.C. 803(d)(3) cross-
references 5 U.S.C. 706); see also id. (``[W]e will set aside the [ ] 
Determination `only if it is arbitrary, capricious, an abuse of 
discretion, or otherwise not in accordance with law, or if the facts 
relied upon by the agency have no basis in the record.'').
    Here, the D.C. Circuit explicitly found that the Judges' analysis 
and findings in connection with the 15.1% revenue rate are not 
arbitrary and capricious, and that the facts relied upon by the Judges 
have a sufficient basis in (are ``grounded in'') the record. It seems 
beyond dispute that the D.C. Circuit affirmed the Judges in their 
setting of the 15.1% revenue rate as a rate that is reasonable, and 
thus satisfies that aspect of the section 801(b)(1) standard.\31\ 
Indeed, it would border on the Orwellian to misconstrue the D.C. 
Circuit's unequivocal and obvious affirmance of the reasonableness of 
the 15.1% revenue rate as a vacating of that finding.
---------------------------------------------------------------------------

    \31\ The CRB Judges intentionally distinguish between the 
``reasonable'' rate standard in the initial body of section 
801(b)(1) and the objectives set forth as Factors A-D of section 
801(b)(1). A rate can satisfy the statutory ``reasonable rate'' 
requirement yet require adjustment (higher or lower) to reflect the 
balancing of the four additional factors. Accordingly, the Judges 
defer to a subsequent section, infra, a discussion of how Factors A-
D should be addressed on this remand.
---------------------------------------------------------------------------

    Third, the Judges note that Johnson conspicuously declines to 
identify the Judges' setting of the 15.1% percent-of-revenue rate as 
one of the findings to be revisited on remand. Rather, Johnson states 
that the three overarching issues for resolution on remanded are the 
Majority's failure: (1) ``to provide adequate notice of the rate 
structure it adopted,'' (2) ``to explain its rejection of a past 
settlement agreement as a benchmark for rates going forward; and (3) 
``[to] identif[y] the source of its asserted authority to substantively 
redefine a material term after publishing its Initial Determination.'' 
Johnson, 369 F.3d at 367. The Majority's finding that the 15.1% royalty 
rate is ``reasonable'' was not identified by the D.C. Circuit as a 
finding that was vacated and subject to further review and, indeed, as 
noted supra, the appellate panel credited what it characterized as the 
Majority's careful analysis and line-drawing in arriving at that 
finding.
    The clarity of the D.C. Circuit's affirmance of the royalty rate of 
15.1% for the percent-of-revenue prong moots the issue of whether 
Professor Marx's attempt, described supra, to correct the so-called 
``imbalance'' problem has merit. However, the Judges note that, even if 
this issue had not been conclusively decided in Johnson, they would 
reject her approach as futile. That is, Professor Marx fails to 
acknowledge that any surplus that her approach would appear to provide 
to the Services would be siphoned off by the Majors, given their 
complementary oligopoly power.
    More particularly, the sound recording royalty rates she posits 
([REDACTED]%, [REDACTED]% and [REDACTED]%) are all functions of the 
sound recording companies' understanding of the Services' non-content 
costs (costs that the Services must recover out of retained revenues in 
order to remain in operation, i.e., to ``survive'') and the then-
existing musical works content (royalty) costs (comprised of the 
mechanical rate and the performance rate). If, as Professor Marx 
contemplates, the mechanical rate is reduced so that Copyright Owners 
``share the burden'' of the complementary oligopoly effect on sound 
recording rates, that ``burden sharing'' would increase the revenues 
retained by the Services (that is the purpose of Professor Marx's 
approach!). But such an increase would raise the Services' revenue 
above their ``survival'' rate, as understood by the record labels. 
Thus, the record labels, given their complementary oligopoly power, 
would increase the Services' royalty rate above what it otherwise would 
have been.
    Alternately stated, when Professor Marx hypothesizes a given sound 
recording royalty rate in column 1 of Figure 8 in her WDRT, that rate 
is assumed, by the logic of the complementary oligopoly theory, to have 
already allowed the services to cover only their non-content costs and 
musical works royalties, as understood by the record labels. So, her 
assumed rate in column 1 is not a fixed parameter, but rather an 
independent variable, which is a function of, inter alia, the costs 
incurred by the services, i.e., their non-content costs plus their 
musical works royalty costs.\32\ If those service costs decreased (for 
example, in an attempt to reduce the services' burden of bearing the 
full brunt of the labels' complementary oligopoly power as in Professor 
Marx's attempt to correct the imbalance problem), the percentage in 
column 1 of Figure 8 would increase, as the labels siphoned off that 
surplus over the services' survival revenue requirements. To find 
otherwise would be to refute the logic of the dynamics of the 
complementary oligopoly effect.\33\
---------------------------------------------------------------------------

    \32\ The interactive services also pay a separate royalty for 
the performance license necessary to transmit a song. However, under 
the Judges' ``All-In'' royalty structure, that performance royalty 
is deducted from the ``All-In'' calculation to determine the 
mechanical royalty. Also, the performance royalty paid to the 
largest Performing Rights Organization (PROs) are subject to 
determination by federal judges in the Southern District of New York 
(the so-called ``rate court'').
    \33\ To be clear, the Judges are not stating that the Services' 
retention of only enough revenue to allow them to cover their 
noncontent costs and thus merely ``survive'' is indicia of an 
effectively competitive (or even healthy) market--but are merely 
acknowledging the state of affairs given the unregulated nature of 
the sound recording royalties and the complementary oligopoly power 
that exists in that market.
---------------------------------------------------------------------------

    Moreover, the defect in Professor Marx's attempt to remedy the so-
called ``imbalance'' problem is a consequence of the statutory 
licensing and royalty scheme. To recap, the licensing of content used 
by the interactive services is bifurcated. The sound recording 
royalties paid by the interactive services to the record labels are not 
regulated, and complementary oligopoly power exists in that market, 
inflating sound recording royalty rates above an effectively 
competitive level. See

[[Page 54413]]

Determination at 73 (``[T]he existence of complementary oligopoly 
conditions in the market for sound recordings'' is the basis for ``the 
record companies' ability to obtain most of the available surplus'' 
generated by interactive streaming.) \34\ However (and to state the 
obvious), the mechanical rate paid by the interactive services for 
musical works is regulated, pursuant to 17 U.S.C. 115 and, until the 
2018 enactment of the Music Modernization Act,\35\ according to the 
rate standards in 17 U.S.C. 801(b)(1). Thus, there is no statutory or 
regulatory impediment to prevent record labels from responding to a 
decrease in the mechanical rate by increasing the unregulated sound 
recording rate if such an increase is in their economic interest.\36\
---------------------------------------------------------------------------

    \34\ As the Judges have consistently noted, this complementary 
oligopoly power is generated by the concentration of ownership of 
sound recording licenses for ``Must Have'' repertoires among the 
three Majors (Sony Music Group, Warner Music Group and Universal 
Music Group), plus Merlin (a consortium of Indies sometimes referred 
to as ``the fourth Major''), as indicated by their reported 
collective 85% share of Spotify's streams in 2018, the first year of 
the rate period at issue here. See https://www.midiaresearch.com/blog/smaller-independents-and-artists-direct-grew-fastest-in-2020.
    \35\ In subsequent rate periods, the rate remains regulated, but 
is subject to a different standard--the ``willing buyer-willing 
seller marketplace standard,'' for shorthand) under 17 U.S.C. 115.
    \36\ The inverse relationship between changes in the mechanical 
royalty rate and changes in the sound recording royalty rate has 
been characterized as the ``seesaw'' effect, which is discussed in 
further detail infra, with regard to the uncapped TCC rate prong.
---------------------------------------------------------------------------

    Accordingly, any attempt by the Judges to reduce the mechanical 
royalty rate in order to allow the Services to retain more of the 
surplus would fail; it would be like pouring water into a bucket with a 
siphon at its base. More water would not remain in the bucket, but 
rather would accumulate wherever the siphon leads--in this case, to the 
record labels. The Judges could keep mechanical royalty rates depressed 
and allow this to occur, but that would harm Copyright Owners while 
providing no relief to the Services. And despite the old adage that 
``misery loves company,'' the Judges detect no directive under section 
801(b)(1) that they harm Copyright Owners without providing a gain for 
the interactive streaming services--and that they provide a windfall 
for the record labels, to boot.
    Although Professor Marx's attempt to reduce the Services' 
``misery'' by sharing it with Copyright Owners is unavailing, the 
statutory scheme and market forces do appear to combine to mitigate the 
burden created by the complementary oligopoly power of the sound 
recording companies. If interactive streaming revenue were to grow over 
the rate period,\37\ then the phase-in to the 15.1% rate will reflect 
fixed annual percentages of a larger base, allowing services to retain 
a higher dollar level of the interactive streaming revenues.\38\ 
[REDACTED]. See, e.g., Diab WDRT ]] 10-11 (Google agreements); 
Mirchandani WDRT ]] 16-17 (Amazon agreements); Bonavia WDRT ]] 8; 14-19 
(Spotify agreements); White WDRT ]] 6; 8-14; 19; 24; 27-28 (Pandora 
agreements). Additionally, the Services' headline sound recording rates 
[REDACTED]. Services' Joint Remand Reply Brief at 40 (and record 
citations therein). Thus, assuming no increase in non-content costs (or 
increases smaller than the increases in streaming revenue), the 
Services will realize increased revenue above and beyond what they 
needed to survive.
---------------------------------------------------------------------------

    \37\ Because this proceeding was appealed and remanded, the 
Judges have the benefit of knowing the ``future'' (beyond 2017), 
during which U.S. interactive streaming revenues have continued to 
grow, a fact that is undisputed, and as to which the Judges take 
administrative notice. See, e.g., RIAA 2018 Year-End Music Industry 
Revenue Report (available at https://www.riaa.com/wp-content/uploads/2019/02/RIAA-2018-Year-End-Music-Industry-Revenue-Report.pdf; RIAA 2020 Year-End Music Industry Revenue Report 
(available at https://www.riaa.com/wp-content/uploads/2021/02/2020-Year-End-Music-Industry-Revenue-Report.pdf (interactive streaming 
revenue increased within this rate period from (approximately) $1.6 
billion in 2018 to $7.7 billion in 2019 and $8.8 billion in 2020).
    \38\ For example, if a royalty is set at a flat rate of 15.1% 
when a revenue base is $1,000, then the royalty is $151, leaving 
$849 in revenues to cover other costs which, for this example, are 
held constant. If the revenue base doubles to $2,000, the same flat 
15.1% royalty rate generates $302 in royalties, leaving $1,698 in 
revenues to cover other costs which, if constant, allow for the 
additional revenue ($1,698-$849 = $849) to generate profits.
---------------------------------------------------------------------------

    The Services and Copyright Owners recognize the mitigation of harm 
to the Services generated by these facts (although they may well 
disagree with the Judges' application of these facts). During colloquy 
with counsel for Pandora and Spotify during closing arguments on 
remand, the Judges asked why they should in essence apply the ``misery 
loves company'' adage:

    [JUDGE STRICKLER] [T]he problem is . . . the sound recording 
[rates] are unregulated in the interactive market . . . . Congress 
did not want that to be controlled at all. So every time I see . . . 
the services' argument about how we have [to] set a rate that's fair 
even though there's this ability of the sound recording [companies] 
to take more, my margin note is always this: ``Are they arguing that 
`misery loves company?' '' [W]hy shouldn't that misery be shared 
with Copyright Owners? . . . Isn't that really Professor Marx's 
argument in her proposed split . . . using the 15.1 percent figure . 
. . ?
    [COUNSEL] [Regarding] Judge Strickler['s] . . . ``misery loves 
company'' issue. . . . I think . . . the way [Judge Strickler] put 
it during the trial was, even if I thought rates needed to come 
down, how would that help you; wouldn't the labels just take all 
that surplus for themselves based on their complementary oligopoly 
power? . . . . I want[ ] to address it right off the bat . . . . in 
open session.
    Relat[ed] to . . . the seesaw . . . our point is that these 
label rates are sticky in both directions. If you see an increase in 
musical works rates, you do not see a quick decrease in label rates, 
and the opposite is true. These rates are sticky.
. . .
    There's a lot of friction with respect to the ability of label 
rates to change quickly in response to the dynamic marketplace or 
the dynamic for business reasons or because of regulatory changes in 
musical works rate. These are multi-year contracts. They take a long 
time to negotiate. They are complex, et cetera.
    So, I do think it's right that at a minimum you can buy time 
where the ratio is more aligned with the 801(b) factors. In other 
words, you don't have to worry that the labels will take it all 
right away, even if you believe they will ultimately take that.
    [JUDGE STRICKLER] So you are saying we have something that 
reduces misery for a period of time until the misery returns?
    [COUNSEL] That's right. And I think that would have been true in 
2018 when you were sitting drafting the decision. It's even more 
true today in 2022 when the label rates, as I mentioned, are 
effectively set, bought and paid for.

3/8/22 Tr. 29-30, 43-46 (Closing Argument) (emphasis added).
    Similarly, on this topic, Copyright Owners' counsel accurately 
characterized the Judges' adoption of the static 15.1% Shapley-based 
rate as the inevitable consequence of ``regulatory lag,'' that requires 
a regulator to keep a rate constant over the statutory term because 
there is no sufficient data to project future rates. Id. at 273-75; see 
generally A. Kahn, 2 The Economics of Regulation at 48 (1971) ``The 
regulatory lag [is] the inevitable delay that regulation imposes in the 
downward . . . [and] upward adjustments'' to rate levels, and ``thus is 
to be regarded not as a deplorable imperfection of regulation but as a 
positive advantage [because] companies can for a time keep the higher 
profits they reap from a superior performance. . . .'').\39\
---------------------------------------------------------------------------

    \39\ The Judges emphasize two points that mitigate any negative 
impact on Copyright Owners from the static nature of the 15.1% 
revenue rate. First, as a percent-of-revenue rate, it generates more 
royalty revenue in a growing market, so the quantum of revenue is 
not static. Second, Copyright Owners' own economic expert witness, 
Professor Gans, testified that the data in the ``market 
observations'' from the Goldman Sachs Report on which he relied were 
the result of ``negotiated rates in the free market and thus 
``presumed to . . . fully consider[ ] . . . expectations of future 
costs and revenues . . . . incorporate[ing] expectations of future 
values.'' Gans WRT ]] 37-38. On this issue, it is noteworthy that 
both the Majority and the D.C. Circuit credited Professor Gans's 
reliance on these projections. See Determination at 70 (``The Judges 
. . . find Professor Gans' reliance on financial analysts' 
projections for the respective industries to be reasonable.''); 
Johnson, 969 F.3d at 386 (holding that ``[t]he CRB Judges' finding 
that Gans's . . . reliance on Goldman Sachs' profit projections'' 
was ``reasonable'' and the] . . . type of line-drawing and reasoned 
weighing of the evidence [that] falls squarely within the [Copyright 
Royalty Board's] wheelhouse as an expert administrative agency.'')
    Thus, dynamic changes going forward in the rate term are 
embodied in the 15.1% revenue rate, and dynamic market expectations 
are incorporated in the modeling data used to establish that rate.

---------------------------------------------------------------------------

[[Page 54414]]

4. Consideration of Factors A-D in Section 801(b)(1)
    Finally, the Judges consider the impact of Factors A-D of section 
801(b)(1) in connection with the setting of the revenue percentage rate 
of 15.1%.\40\ Regarding Factor A, it cannot be gainsaid that the D.C. 
Circuit has left this issue unresolved. Rather, Johnson unambiguously 
affirmed the Majority's finding that an increase in the mechanical 
royalty rate was warranted. Specifically, Johnson states that the 
Majority's decision in this regard met the ``test'' that it be 
``supported by substantial evidence [and] reflect a reasonable reading 
of the record.'' Johnson, supra, at 388. Moreover, with regard to the 
level of the increase, the D.C. Circuit did not disturb the finding by 
the Majority that ``[t]he rates determined by the Judges represent a 
44% increase over the current headline rate, and thus satisfies the 
Factor A objective. . . .'' Determination at 85.\41\
---------------------------------------------------------------------------

    \40\ The D.C. Circuit ruled, with regard to the ``nature of the 
rate structure,'' that because it had ``vacat[ed] and remand[ed] . . 
. for lack of notice'' ``[t]he question whether the [Judges] 
adequately addressed factors B through D is bound up with the 
[Judges'] analysis of sound recording rightsholders' likely 
responses to the new rate structure.'' Johnson, supra, at 389. 
However, the 15.1% revenue rate, viewed separately, is not bound up 
in the ``rate structure'' issue, which relates to the uncapped TCC 
prong and how the 15.1% revenue rate may be ``intertwined'' with 
that second rate prong. As explained infra, the Judges are not 
adopting an uncapped TCC rate prong, so the 15.1% rate is no longer 
``bound up'' with the vacated and remanded ``rate structure'' issue, 
making moot the argument that a new post-remand analysis of Factors 
B through D is necessary or appropriate. However, on remand, 
Copyright Owners have placed in issue the ``disruption'' element of 
Factor D, claiming that the Services have not proven that the 
uncapped TCC rates and rate prong have or will cause disruption.
    \41\ The 44% figure cited by the Majority reflects the 
percentage increase of the headline rate, from 10.5% to 15.1%.
---------------------------------------------------------------------------

    With regard to Factors B and C,\42\ even if Johnson were construed 
as permitting the Judges to revisit this issue, they would not adjust 
the 15.1% revenue rate on the basis of these two factors. In this 
regard, the Judges note that the Majority found that the 15.1% revenue 
rate was not only ``reasonable,'' but also a ``fair allocation of 
revenue between copyright owners and services.'' Determination at 87 
(emphasis added). The Majority thus found explicitly that ``with regard 
to Factors B and C . . . there is no basis to depart from [its] 
determination of the reasonable . . . rate structure and rates as set 
forth supra.'' Id. More particularly, the Majority calculated the 15.1% 
rate by utilizing the total royalty percentage revenue of only 
[REDACTED]% as calculated by Spotify's economic expert witness, 
Professor Marx, whose economic modeling intentionally reflected a 
conception of fairness by reducing the effect of the labels' 
complementary oligopoly market power. See Determination at 67-68 
(noting that Professor Marx testified that this aspect of her model 
``represents a fair outcome in the absence of market power [and] . . . 
eliminates . . . market power'' which . . . if left in the economic 
analysis would ``render[ ] . . . the analysis incompatible with the 
objectives of Factors B and C of section 801(b)(1).)'') (emphasis 
added).\43\
---------------------------------------------------------------------------

    \42\ Factors B and C are typically considered jointly, because 
of the overlap in the objectives of providing a ``fair return'' and 
a ``fair income'' to the licensors and licensees respectively (the 
Factor B objectives) and reflecting their relative roles in making 
the streamed music available to the public (the Factor C 
objectives). See Johnson, 969 at 388 (noting without criticism the 
joint consideration of Factors B and C; Determination at 85-86 
(noting without criticism the several experts' joint consideration 
of Factors B and C).
    \43\ Additonal facts support the Majority's finding that the 
15.1% revenue rate is fair. The record evidence indicates that the 
headline percent-of-revenue sound recording rate was between 
approximately [REDACTED]% to [REDACTED]% in 2017. See Marx WDRT ] 
58, fig 7. When the 15.1% mechanical rate is added to that rate 
range, the range of the total royalty obligation (based on headline 
rates) is [REDACTED]% to [REDACTED]%. (Plus, given the phase-in of 
the rates expressly to avoid disruption, the total royalty 
obligation would be even lower before 2022, at current sound 
recording rates.) The evdence pre-remand indicated that the Services 
were ``surviving'' while incurring noncontent of costs of 
approximately [REDACTED]% of revenue, leaving about [REDACTED]% of 
revenue available to pay royalties while still remaining in 
business. See Eisenach WRT ] 79 (Copyright Owners' expert economic 
witness); McCarthy WDT ]] 28-29 (Spotify's Chief Financial Officer.) 
Thus, even if the Judges were to engage in a de novo analysis of the 
potential applicability of Factors B and C to the 15.1% rate, they 
would not find any basis sufficient to warrant a downward rate 
adjustment, beyond the phase-in adopted in the Determination.
---------------------------------------------------------------------------

    Accordingly, the Judges find it would be substantively unwarranted 
to engage in any new consideration on remand of the impact, if any, of 
Factors B and C on the otherwise reasonable 15.1% revenue rate.\44\
---------------------------------------------------------------------------

    \44\ However, the Judges take note of their further observation, 
discussed supra, that the combined impact of ``sticky'' sound 
recording royalty rates and the inevitable regulatory lag provide an 
additional modicum of fairness with regard to the mechanical royalty 
rate.
---------------------------------------------------------------------------

    The final itemized statutory factor--Factor (D)--instructs the 
Judges to consider the ``competing priority'' of ``minimiz[ing] any 
disruptive impact on the structure of the industries involved and on 
generally prevailing industry practices.'' 17 U.S.C. 801(b)(1)(D). As 
with Factors B and C, even if Johnson were construed to allow the 
Judges to revisit this issue on remand with respect to the 15.1% 
revenue rate, the Judges would not change the Majority analysis or 
findings. In the Determination, the Judges adopted the following 
interpretation of this standard set forth in previous determinations:

    [T]he Judges reiterated their understanding of Factor D, 
concluding that a rate would need adjustment under Factor D if that 
rate directly produces an adverse impact that is substantial, 
immediate and irreversible in the short-run because there is 
insufficient time for either [party] to adequately adapt to the 
changed circumstance produced by the rate change and, as a 
consequence, such adverse impacts threaten the viability of the 
music delivery service currently offered to consumers under this 
license.

Determination at 86 (emphasis added).
    Also, in order to minimize any economic disturbance to the 
Services' businesses, the Majority decided to phase-in the 15.1% rate 
over the five-year rate term, setting annual percent-of-revenue rates 
as follows: 11.4% in 2018; 12.3% in 2019; 13.3% in 2020; and 14.2% in 
2021, before the full 15.1% rate became effective in 2022 the final 
year of the rate term. Id. at 87-88.
    On remand, the Services have not made any argument that the rate 
structure or rates set by the Majority were ``disruptive under this 
standard.'' \45\ In sum, there is insufficient basis for the Judges to 
change the Majority's application of Factor (D) to the 15.1% revenue 
rate finding by the Majority.\46\
---------------------------------------------------------------------------

    \45\ The Judges further discuss the Factor D ``disruption issue 
infra in connection with their analysis of the uncapped TCC prong.
    \46\ Additional facts further support the Majority's finding 
that the 15.1% revenue rate is would not be disruptive under Factor 
D. The record evidence indicates that the headline percent-of-
revenue sound recording rate was between approximately [REDACTED]% 
to [REDACTED]% in 2017. See Marx WDRT ]] 14, 19. When the 15.1% 
mechanical rate is added to that rate range, the range of the total 
royalty obligation (based on headline rates) is [REDACTED]% to 
[REDACTED]%. (Plus, given the phase-in of the rates expressly to 
avoid disruption, the total royalty obligation would be even lower 
before 2022, at current sound recording rates.) The evidence pre-
remand indicated that the Services were ``surviving'' while 
incurring noncontent costs of approximately [REDACTED]% of revenue, 
leaving about [REDACTED]% of revenue available to pay royalties 
while still remaining in business. See Eisenach WRT ] 79 (Copyright 
Owners' expert economic witness); McCarthy WDT ]] 28-29 (Spotify's 
Chief Financial Officer). Thus, even if the Judges were to engage on 
remand in a de novo analysis of the potential applicability of 
Factor D to the 15.1% rate, they would not find any disruption 
sufficient to warrant a downward rate adjustment, beyond the phase-
in adopted in the Determination.

---------------------------------------------------------------------------

[[Page 54415]]

5. Conclusion Regarding the 15.1% Revenue Rate
    For the forging reasons, the Judges do not disturb the Majority's 
finding that the percent-of-revenue rate at 15.1%, phased-in annually 
over the rate period, constitutes a ``reasonable'' rate under section 
801(b)(1) to be used as the statutory rate for the 2018 to 2022 
period.\47\
---------------------------------------------------------------------------

    \47\ The Services' assert that the Judges previously found that 
the reasonableness of the 15.1% rate was subject to revision on 
remand. In support of this position, the Services cite to the 
Judges' Order Granting in Part and Denying in Part Copyright Owners' 
Motion for Reconsideration or, in the Alternative, Clarification at 
3, 4 n.7 (January 6, 2022) (Jan. 6th Order). But the Judges said in 
that interlocutory proposal merely that Copyright Owners were 
incorrect in their extreme assertion that the Judges could not make 
an ``alternative rate and rate structure finding . . . except for 
the re-adoption of the vacated rate and rate structure approach in 
the Phonorecords III Determination [because] . . . [t]hat . . . 
would . . . be inconsistent with Johnson [and] . . . would render 
the D.C. Circuit's vacating and remanding of the proceeding without 
force or effect.'' Id. at 4, n.7. That did not mean that certain 
elements of the D.C. Circuit's ruling could be ignored. Further, 
when the Judges provided the parties with the Judges' explicitly 
tentative ``Working Proposal,'' they did not declare that the 15.1% 
revenue rate calculation could be revisited. Rather, the Judges 
``express[ed] a concern, not that the foregoing calculations could 
be overridden, but rather that this analysis . . . is `incomplete' . 
. .'' Jan. 6th Order at 6 (emphasis added). The parties' submissions 
in response to the Judges' ``Working Proposal'' demonstrated that 
the 15.1% revenue rate calculation was not ``incomplete'' in the 
manner that had raised the Judges' concern. Nothing the Judges said 
in this interlocutory and tentative ``Working Proposal'' constituted 
a definitive statement regarding the Judges' view of what was and 
was not subject to review on remand. See generally merriam-webster.com (defining the adjective ``working'' in this context as 
``assumed or adopted to permit or facilitate further work or 
activity . . . a working draft.''). Indeed, a primary purpose of the 
``Working Proposal'' was to allow the Judges and the parties to 
address potential issues and resolutions, without prejudice going 
forward.
---------------------------------------------------------------------------

D. Uncapped TCC Rate Prong

1. Two Post-Remand Rationales for Uncapped TCC Rate Prong
    The Determination set forth the following two primary reasons for 
adopting a ``greater-of' rate structure that also included an uncapped 
TCC rate prong:

    First, the use of an uncapped TCC metric is the most direct 
means of implementing a key finding . . . by the experts for 
participants on both sides in this proceeding: the ratio of sound 
recording royalties to musical works royalties should be lower than 
it is under the current rate structure. Incorporating an uncapped 
TCC metric into the rate structure permits the Judges to influence 
that ratio directly.
    Second, an uncapped TCC rate prong effectively imports into the 
rate structure the protections that record companies have negotiated 
with services to avoid the diminution of revenue.

Determination at 35-36.\48\
---------------------------------------------------------------------------

    \48\ The Majority added two other reasons that are not germane 
to this remand. In particular, the Majority stated that, compared to 
the Phonorecords II benchmark proposed by the Services, the 
``greater-of'' structure with the uncapped TCC rate prong was 
``simpler'' to understand than the ``Rube Goldberg-esque'' nature of 
the Phonorecords II rate structure. Id. at 36. This issue apparently 
was not raised on appeal, as it was not mentioned in Johnson, and 
Copyright Owners have not raised the issue on remand. See CO Initial 
Submission, supra. (However, the Judges do consider this issue in 
their analysis of the PR II-based benchmark, infra.) The final 
reason provided by the Majority was that its adoption of an uncapped 
TCC rate prong was supported by evidence of Google's agreements with 
labels that included an uncapped rate structure, on which Google had 
relied to propose, post-hearing, the same greater-of rate structure. 
Id. However, the D.C. Circuit found that Google's proposal was 
distinguishable, as it was based on a far lower TCC rate (15%) as 
well as a far lower percent-of-revenue rate (10.5%). The D.C. 
Circuit thus declined to rely on the Google-based approach as 
support for the uncapped TCC rate prong. Johnson, 969 F.3d at 383.
---------------------------------------------------------------------------

2. Copyright Owners' Position
    Copyright Owners claim that the uncapped TCC prong should be 
adopted. They contend that the D.C. Circuit remand was merely 
``procedural'' rather than substantive, and the Judges thus are not 
precluded from readopting the uncapped TCC prong in this remand 
proceeding. CO Initial Submission at 35-38 (and record citations 
therein).
    They further contend that the uncapped TCC prong was adopted to 
provide protection against revenue deferment and displacement 
occasioned by the Services choosing to elevate the growth of 
subscribers and other listeners over revenue maximization. Id. at 38-43 
(and record citations therein). The uncapped TCC prong was first 
proposed by Google to persuade the Judges to reject Copyright Owners' 
proposed ``greater-of'' rate structure containing a per-play prong and 
a per subscriber prong. Id. at 43-46 (and record citations therein).
    Copyright Owners argue that the uncapped TCC prong should be 
adopted because: (1) the Services have not shown any actual or 
threatened ``disruption'' or other harm resulting from the uncapped TCC 
prong during the 33-month period; (2) the Services actually experienced 
``unprecedented growth and profit'' during this period; and (3) the 
Services paid lower percentages of revenues in mechanical and total 
royalties when the uncapped TCC prong was in effect. Copyright Owners' 
Reply Brief on Remand at 34-48 (and record citations therein).
    Relatedly, according to Copyright Owners the Services' argument 
that the ``see-saw'' effect is unsupported by empirical evidence has 
collapsed, given the evidence relating to market performance. Further 
Copyright Owners maintain that this argument is irrelevant to the rate 
structure issue. Id. at 48-50 (and record citations therein).
3. Services' Position
    The Services argue on remand that the uncapped TCC rate prong must 
be rejected. The Services reject the ``seesaw'' theory claiming it is 
disproved by the experience of the parties during the 33-month period. 
Services' Joint Opening Brief at 48-49; Services' Joint Supplemental 
Brief at 7-13 (Nov. 15, 2021) (and record citations therein). The 
Services further contend that Copyright Owners have disavowed the 
``seesaw'' theory as understood by the Majority. The Services allege 
that Copyright Owners now claim that the theory was nothing more than 
``a nod'' to certain ``core principles'' of bargaining theory, rather 
than a specific prediction of a commensurate inverse relationship 
between increases in the mechanical royalty rate and decreases in the 
sound recording royalty rate. Services' Joint Supplemental Brief at 2, 
5-7 (and record citations therein).
    With regard to the uncapped TCC rate prong, the Services assert 
that Copyright Owners have not even attempted to demonstrate--nor could 
they demonstrate--that the uncapped TCC rate prong is consistent with 
all four statutory objectives set forth in section 801(b)(1). Services' 
Joint Reply Brief at 1, 3-4, 33-34 36 (July 2, 2021) (``Services' 
Reply''); see also Services' Joint Opening Brief at 44-64 (and record 
citations therein). The Services claim that ``yoking'' the mechanical 
rate to the ``complementary oligopoly rates extracted by the labels is 
plainly unreasonable.'' Services' Joint Opening Brief at 44-46. The 
Services argue that the existence, vel non, of any ``disruptive 
impact'' arising from the uncapped TCC rate prong, is misguided and not 
dispositive, because it is only one of the four separately itemized 
factors and, as this factor relates to Copyright Owners' proposed 
uncapped TCC prong, they bear the burden of proof. Services' Reply at 
35-37.
    Finally, the Services contend that Copyright Owners have failed to 
explain their self-contradictory pre-remand argument that ``an uncapped 
TCC prong `does nothing to protect Copyright Owners from the Services' 
revenue

[[Page 54416]]

displacement and deferment.' '' Services' Reply at 43.
4. Application of Johnson Findings Regarding Uncapped TCC Rate Prong
    The Judges conclude that the D.C. Circuit affirmed the Majority's 
derivation and calculation of the 26.1% TCC rate, but vacated and 
remanded the Judges' application and inclusion of that rate prong in 
the rate structure. The D.C. Circuit noted that, on appeal, the 
Services contended that ``it was arbitrary and capricious for the 
[Judges] to rely on information drawn from different expert analyses in 
calculating the mechanical royalty rates.'' Johnson, 969 F.3d at 384. 
Thus, the Services were making the same ``information''-based argument 
in opposition to the calculation of both aspects of the mechanical 
royalty rates--the revenue percentage prong and the TCC prong. See also 
id. (``the Streaming Services separately leveled objections to the 
particular percentages adopted by the Copyright Royalty Board to 
calculate the revenue and total content cost prongs.'') (emphasis 
added)
    In fact, both rate prongs were indeed derived from the same 
analyses. See Determination at 75 (table) (showing that both 15.1% 
revenue rate and 26.2% TCC rate derived from same data--Professor 
Marx's model showing total royalties as high as [REDACTED]% [Majority's 
lower bound] and Professor Gans's ``Shapley-inspired'' model showing 
TCC percent should be [REDACTED]%.) \49\
---------------------------------------------------------------------------

    \49\ The reciprocal of Professor Gans's [REDACTED]ratio of sound 
recording:musical works royalties is [REDACTED], or [REDACTED]%.
---------------------------------------------------------------------------

    It is also clear from Johnson that the D.C. Circuit found that the 
Majority had reasonably derived and calculated the 26.2% TCC rate:

    When it came to . . . the ratio of sound recording to musical 
work royalties that Gans derived from his analysis the [CRB Judges] 
specifically found . . . reasonable Gans' equal value assumption 
[for dividing the Shapley surplus . . . between sound recording and 
musical works owners] and his reliance on Goldman Sachs' profit 
projections. That type of line-drawing and reasoned weighing of the 
evidence falls squarely within the Board's wheelhouse as an expert 
administrative agency.

See Johnson, 969 F.3d at 385-86 (cleaned up) (emphasis added). 
Accordingly, because the identical analysis was performed by the Judges 
to derive the 26.2% TCC rate as was done to derive the 15.1% revenue 
rate, the Majority's finding with regard to the derivation and 
calculation of the TCC rate likewise is not subject to further 
consideration on remand by the Judges.
    However, it is equally clear that the D.C. Circuit vacated and 
remanded the Majority's application and inclusion of the 26.2% TCC rate 
in a separate ``greater-of'' TCC prong. The defect that generated the 
vacating on this issue was procedural-- ``the Streaming Services had no 
notice that they needed to defend against and create a record 
addressing such a significant, and significantly adverse, overhaul of 
the mechanical license royalty scheme . . .'' Id. at 382. The 
consequence of the D.C. Circuit's action, however, was substantive. The 
D.C. Circuit stated:

    This is no mere formality. Interested parties' ability to 
provide evidence and argument bearing on the essential components 
and contours of the [Judges'] ultimate decision not only protects 
the parties' interests, it also helps ensure that the [Judges'] 
ultimate decision is well-reasoned and grounded in substantial 
evidence. . . .
    The Streaming Services separately challenge the uncapped rate 
structure as arbitrary and capricious. In particular, they argue 
that the rate structure formulated by the [Judges] failed to account 
for the sound recordings rightsholders' market power. They also 
object that the [Judges] failed to provide a `satisfactory 
explanation, or root in substantial evidence, [their] conclusion 
that an increase in mechanical license royalties would lead to a 
decrease in sound recording royalties [the ``inverse relationship'' 
a/k/a the ``seesaw'' effect].

Id. at 381-83 (cleaned up) (emphasis added). Thus, the D.C. Circuit 
explicitly declined to address these substantive issues, because of the 
deficient procedure. Instead, the D.C. Circuit remanded these 
substantive issues back to the Judges. Id. Simply put, Johnson found 
that the absence of notice here could be outcome-determinative. Thus, 
the Judges categorically reject Copyright Owners' assertion that the 
remand as to the uncapped TCC rate structure was merely ``procedural.'' 
The Judges do not accept the notion that the Majority simply committed 
some ministerial faux pas that could be summarily corrected so that the 
uncapped TCC rate structure could be rubber-stamped on remand. Rather, 
the Judges' error rendered it impossible for them to consider the pros 
and cons of such a rate structure without the necessary input from the 
Services (and, for that matter, Copyright Owners as well).
    Because the procedural infirmity precluded the D.C. Circuit from 
deciding whether the Majority's decision was ``well-reasoned and 
grounded in substantial evidence,'' there also can be no substantive 
presumption of the appropriateness of the uncapped TCC rate prong, as 
suggested by Copyright Owners. To the contrary, the D.C. Circuit's 
opinion makes it clear that on remand the Judges must engage in a fresh 
consideration of the statutory appropriateness, vel non, of the 
uncapped TCC rate prong, by weighing and contextualizing the competing 
evidence and testimony entered into the record both before and after 
the remand.
    Accordingly, although Copyright Owners correctly assert that 
Johnson did not find the uncapped TCC rate structure to be ``unfair, 
unreasonable or inequitable,'' Johnson just as clearly did not find 
that structure to be ``fair, reasonable or equitable.'' Rather, the 
purpose of the remand was for the Judges to make these determinations. 
Accordingly, the Judges next examine whether setting the statutory 
mechanical rate as an uncapped TCC rate is ``reasonable,'' as required 
by section 801(b)(1).\50\
---------------------------------------------------------------------------

    \50\ The Judges consider infra whether any of the four itemized 
statutory factors require an adjustment to this analysis.
---------------------------------------------------------------------------

5. Determining Whether Uncapped TCC Rate Prong is ``Reasonable''
a. Rejection of First Rationale for Including Uncapped TCC Rate
    Two substantive issues are implicated raised with regard to the 
issue of reasonableness: (1) whether the ``seesaw'' theory is valid; 
and (2) if it is valid, whether there exist sufficient data to support 
the phased-in 26.2% uncapped TCC rate.\51\ To demonstrate that this 
uncapped TCC rate prong and the (phased-in) 26.2% rate are reasonable, 
Copyright Owners rely on the combined application of two economic 
models--the Shapley Value model and a Nash Bargaining Model. 
Accordingly, it is necessary to consider how these two models relate to 
each other and how these models and their interrelationship impact the 
setting of the statutory rate.
---------------------------------------------------------------------------

    \51\ As noted supra, in the Judges' recitation of the parties' 
remand arguments regarding the uncapped TCC rate prong, they make 
other arguments as well, specifically regarding: (1)) whether it 
would be necessary and/or appropriate to adopt this uncapped TCC 
rate prong to offset revenue deferral and/or displacement by the 
Services; (2) whether this rate prong has caused, or would cause, 
economic ``disruption'' to the Services (under Factor D of section 
801(b)(1)); (3) whether the uncapped TCC rate prong would satisfy 
Factors B and C of section 801(b)(1); and (4) whether this rate 
prong improperly imports the complementary oligopoly power of sound 
recording licensors. The Judges consider these issues after 
addressing the issues relating to the ``seesaw'' theory.
---------------------------------------------------------------------------

    The D.C. Circuit described the Shapley Value Model methodology:

    The Shapley methodology is a game theory model that seeks to 
assign to each market player the average marginal value that the 
player contributes to the market. This methodology first determines 
the costs that

[[Page 54417]]

each player should recover, then divides the ``surplus'' among the 
players in proportion to the value of their contributions to the 
worth of the hypothetical bargain that would be struck.

Johnson, 969 F.3d at 372. The Judges provided a consistent but more 
detailed definition:

    The Shapley value gives each player his average marginal 
contribution to the players that precede him, where averages are 
taken with respect to all potential orders of the players. The 
Shapley value approach models bargaining processes in a free market 
by considering all the ways each party to a bargain would add value 
by agreeing to the bargain and then assigns to each party their 
average contribution to the cooperative bargain. The idea of the 
Shapley value is that each party should pay according to its average 
contribution to cost or be paid according to its average 
contribution to value. It embodies a notion of fairness. The Shapley 
model is a game theory model that is ultimately designed to model 
the outcome in a hypothetical `fair' market environment. It is 
closely aligned to bargaining models, when all bargainers are on an 
equal footing in the process.

Determination at 62-63 (cleaned up).
    To apply a Shapley Value Model in a rate proceeding, the economic 
modeler must obtain usable cost and revenue data to be inputted into 
the model. More particularly for this proceeding, the modeler must 
identify the parties' input costs, including the Services' non-content 
costs, and the revenue derived from interactive streaming.\52\ The 
difference between these revenues and the Services' noncontent costs 
represents the Shapley ``surplus'' that can be shared among the 
Services, the sound recording companies and Copyright Owners.
---------------------------------------------------------------------------

    \52\ Identifying useful data is a vexing problem. As one of 
Copyright Owners' expert economic witnesses, Professor Watt, has 
written: ``[T]he main problem with the Shapley approach . . . a 
particularly pressing problem [is] that of data availability.'' R. 
Watt, Fair Copyright Remuneration: The Case of Music Radio, 7 Rev. 
Econ. Rsch Copyright. Issues at 21, 27 (2010).
---------------------------------------------------------------------------

(i) The Shapley Approach of the Parties' Economic Expert Witnesses
(a) Professor Gans's ``Shapley-Inspired'' Model
    Professor Gans, Copyright Owners' expert, utilized royalty and 
profit interactive streaming data for record companies and music 
publishers that he obtained from ``a [then] recent music industry 
equity analysis report,'' namely, a Goldman, Sachs Equity Research 
report dated October 4, 2016 entitled ``Music in the Air, Stairway to 
Heaven.'' Gans WDT ] 76 & n.39. As the Majority summarized Professor 
Gans's approach, ``[h]e found that, for the music publishers to recover 
their costs and achieve profits commensurate with those of the record 
companies under his approach, the ratio of sound recording royalties to 
musical works royalties derived from his Shapley-inspired analysis was 
[REDACTED] (which attributes equal profits to both classes of rights 
holders and acknowledges the higher costs incurred by record companies 
compared to music publishers).'' Determination at 69 (citing Gans WDT ] 
77 tbl.3) (emphasis added).
    Regarding Professor Gans's Shapley-inspired analysis, the Majority 
stated:

    [T]he Judges find the ratio of sound recording to musical work 
royalties that Professor Gans derived from his analysis to be 
informative. Professor Gans computed this ratio based on an 
assumption of equal Shapley values between musical works and sound 
recording copyright owners. The Judges find this assumption to be 
reasonable . . . . [53]
---------------------------------------------------------------------------

    \53\ The assumption of equal Shapley values is based on the 
understanding that a sound recording license and a musical works 
license are both necessary (i.e., perfect complements) in order for 
a service to stream a song. Determination at 69 & n.122 therein.

Determination at 70. This is part and parcel of the ``line-drawing'' 
undertaken by the Majority that the D.C. Circuit affirmed. Thus, on 
remand, the Judges do not find cause to reconsider the Majority's 
limited adoption of Professor Gans's Shapley-inspired analysis.\54\
---------------------------------------------------------------------------

    \54\ Because the ratio of sound recording to musical works 
royalties that Professor Gans derived from the data and other 
evidence was the only portion of his testimony on which the Majority 
relied, and because that reliance was affirmed by the D.C. Circuit, 
the criticisms of other aspects of Professor Gans's modeling are no 
longer relevant.
---------------------------------------------------------------------------

(b) Professor Marx's Shapley Value Model
    Professor Marx constructed two Shapley Value Models, one of which 
was relied upon by the Majority. In the model credited by the Majority, 
Professor Marx assumed one collective owner of sound recording 
copyrights and one collective owner of musical works. She also assumed 
the presence of a single interactive service. See Determination at 64-
68. That approach yielded a total royalty obligation for sound 
recordings and musical works ranging between [REDACTED]% and 
[REDACTED]% of the hypothetical service's revenue. Dissent at 133.
    Copyright Owners criticized Professor Marx's decision to assume in 
her model only one interactive streaming service, rather than the 
multiple services that actually existed. They contend that assumption 
reduced the market power of the licensors in her model. According to 
Copyright Owners' economic experts, Professor Marx's approach was a 
misuse of the Shapley Value Model. They aver that the Shapley Value 
approach is intended only to eliminate from the rate derivation the 
bargaining ability of a ``Must Have'' input supplier (like the sound 
recording companies and Copyright Owners) to ``hold-out'' and thus 
squeeze licensees for higher royalties. By modeling every possible 
``arrival ordering,'' they contend, the ``hold-out'' problem is 
avoided. They further contend that Professor Marx misconstrued the 
purpose of the Shapley approach by wrongly modeling market participants 
in a manner that significantly reduced the actual market power of these 
``Must Have'' input suppliers. Determination at 66-67.
    The Majority agreed with Professor Marx. The two Judges in the 
Majority found that her modeling reasonably ``attempts to eliminate a 
separate factor--market power--that she asserts renders a market-based 
Shapley Analysis incompatible with the objectives of Factors B and C of 
section 801(b)(1).'' Id. at 68.
    Although the Majority ultimately relied upon Professor Marx's 
modeling in this regard, the Majority found that her data inputs were 
problematic. Determination at 65. Specifically, Professor Marx relied 
on 2015 data from Warner/Chappell and Warner Music Group for music 
publisher sound recording company noncontent costs, respectively. The 
Majority found that 2015 data was less probative than 2016 data and 
understated the percentage of revenue to be paid to the two classes of 
content providers. However, the Majority ultimately found only that 
this one-year older data served to ``understate'' the allocation of 
surplus to the upstream content providers, and thus rejected only her 
lower [REDACTED]% bound for total royalties, The Majority did decide to 
adopt her upper bound of [REDACTED]% value for total royalties, which 
could (and ultimately did) ``constitute a lower bound for total 
royalties in computing a royalty rate,'' applied by the Majority in 
order to make a downward adjustment to offset the complementary 
oligopoly effect of ``Must Have'' inputs. Id. at 73, 75.
(c) Professor Watt's Criticisms of and Adjustments to Professor Marx's 
Shapley Modeling
    Professor Richard Watt was called by Copyright Owners as a rebuttal 
witness at the hearing, for the purpose of reviewing Professor Marx's 
WDT. Watt WRT ] 3. He concluded that Professor Marx's Shapley Value 
Model contains important methodological and data

[[Page 54418]]

flaws which, in his opinion, caused her to significantly understate the 
mechanical and overall (musical works + sound recording) royalty rates 
to be paid by interactive services pursuant to a proper Shapley 
analysis. Id. at ] 5.
    Professor Watt also criticized her Shapley Value Model for failing 
to incorporate the fact that ``the different interactive streaming 
companies--Spotify, Apple Music, Rhapsody/Napster, Google Play Music, 
Amazon, etc.--do all compete (and rather fiercely) among themselves, 
offering (perhaps perfectly) substitutable services.'' Id. at ] 25. 
Even more strongly in this vein, Professor Watt relied on the following 
description of the substitutability of the streaming services, inter 
se:

    Each [interactive streaming] service in the increasingly crowded 
field is working frantically to overcome the perception that the 
main distinction among the uniformly priced $9.99 a month offering 
is little more than font style, quirky playlist title and color 
scheme. . . . [M]usic platforms have long fought against the 
perception that they're . . . selling a nearly interchangeable 
product . . . You're getting sold the same car [with] just got a 
different lick of paint on it.'').

Id. at ] 32 n.19.
    Professor Watt claimed that incorporating this downstream 
competition into the model would reduce the Shapley values of the 
Services and increase the Shapley values for the input suppliers, by 
recognizing which players provide ``essential inputs'' and which are in 
competition with other suppliers of substitutable inputs. Id.
    He further criticized Professor Marx for including in her model 
``other distributors'' who are not interactive streaming services. Id. 
at ] 27. According to Professor Watt, these other distributors ``do not 
belong in a properly constructed Shapley Value Model because their 
presence would ``show up'' in the model as lower revenues for 
interactive services as their subscribers or listeners left for these 
other distributors (such as noninteractive services). Id.
    Additionally, because he criticized Professor Marx's use of 2015 
data (as noted supra), Professor Watt re-worked Professor Marx's model 
by examining how the use of 2016 data, as opposed to her 2015 data, 
would ``better reflect[ ] . . . the reality of the market. Id. at ] 37; 
see also id. at ] 44. When using the (higher) 2016 revenues (and making 
some relatively more minor adjustments he found necessary), Professor 
Watt estimated that the share of streaming revenues that would be paid 
out in total royalties (for musical works + sound recordings) in 
Professor Marx's model would range from [REDACTED]% to [REDACTED]%. Id. 
at ]] 50-52.\55\
---------------------------------------------------------------------------

    \55\ As noted supra, when the Majority weighed and credited 
Professor Watt's entire Shapley analysis, in which his estimate of 
total royalties was [REDACTED]%, those Judges contextualized 
Professor Marx's [REDACTED]% total royalty calculation as the lower 
bound of a zone of reasonable rates, and applied it as a measure 
that, in their analysis, would offset the complementary oligopoly 
effect of real-world royalties. Determination at 75 (text and tbl.).
---------------------------------------------------------------------------

    After analyzing these Shapley analyses,\56\ the Majority found that 
the mechanical royalty rate needed to be increased in order to provide 
Copyright Owners with a reasonable rate as required by section 
801(b)(1). As a matter of arithmetic though, if the mechanical rate 
increased and the sound recording rate did not decrease by a 
corresponding amount, then the total royalties paid by the Services 
would increase. That issue brings the Judges to consideration of 
Professor Watt's bargaining model, on which the Majority relied to 
posit an inverse relationship (the seesaw effect), by which an increase 
in the mechanical rate would result in a commensurate reduction in the 
sound recording rate.
---------------------------------------------------------------------------

    \56\ Because his testimony was made in rebuttal, leaving the 
Services no procedural right to file written testimony in 
opposition, the Majority gave little weight to Professor Watt's 
total royalty projections and no weight to his proffered ratios of 
sound recordings-to-musical works royalties. Determination at 75.
---------------------------------------------------------------------------

(ii) Professor Watt's Bargaining Model
    Professor Watt's Nash Bargaining Model is the linchpin that 
connects: (a) the higher mechanical royalty rates generated by the 
Shapley Value results relied upon by the Majority with (b) the assumed 
lower sound recording rates--a connection that the Majority found to 
render ``reasonable'' and ``fair'' its uncapped TCC prong. See 
Determination at 73-74 (``As to the issue of applying a TCC percentage 
to a sound recording royalty rate that is artificially high as a result 
of musical works rates being held artificially low through regulation, 
the Judges rely on Professor Watt's insight (demonstrated by his 
bargaining model) that sound recording royalty rates in the unregulated 
market will decline in response to an increase in the compulsory 
license rate for musical works.''). Alternately stated, Professor 
Watt's bargaining model result, i.e., the seesaw effect, if 
sufficiently supported in the record, is the phenomenon that would 
allow the Judges on remand to apply the Shapley results by increasing 
the mechanical rate, without unduly exposing the Services to the risk 
of higher total royalties.
    More particularly, the Majority recognized a potential problem that 
those Judges would have to resolve before utilizing the Shapley Value 
approach to create an uncapped TCC prong: ``This is problematic because 
the sound recording rate against which the TCC rate would be applied is 
inflated . . . both by . . . complementary oligopoly [market] 
conditions . . . and the record companies' ability to obtain most of 
the available surplus due to the music publishers' absence from the 
bargaining table.'' Determination at 73.\57\ But the Majority found 
that Professor Watt had provided a rationale which permitted them to 
resolve the second problem:
---------------------------------------------------------------------------

    \57\ The other problem the Majority needed to resolve was how to 
deflate the market-based sound recording royalty rates to mitigate 
the complementary oligopoly effect in those rates. Id. As discussed 
supra, the Judges resolved this problem by applying the low total 
royalty payment sum, [REDACTED]%, from Professor Marx's Shapley 
Value Model.

    As to the issue of applying a TCC percentage to a sound 
recording royalty rate that is artificially high as a result of 
musical works rates being held artificially low through regulation, 
the Judges rely on Professor Watt's insight . . . that sound 
recording royalty rates in the unregulated market will decline in 
response to an increase in the compulsory license rate for musical 
works. 3/27/17 Tr. 3090 (Watt) (``[T]he reason why the sound 
recording rate is so very high is because the statutory rate is very 
low. And if you increase the statutory rate, the bargained sound 
---------------------------------------------------------------------------
recording rate will go down.'').

Determination at 73-74; see also Watt WRT ] 23 n.13 (``[I]in my 
Appendix 3, I show that . . . if the musical works rate is increased to 
what would be a realistically fair and reasonable rate, then the 
negotiated fee for sound recordings would decrease almost dollar for 
dollar . . . .''); see also id. at ] 36 (``The statutory rate for 
mechanical royalties . . . is significantly below the predicted fair 
rate, and the statutory rate effectively removes the musical works 
rightsholders from the bargaining table with the services. Since this 
leaves the sound recording rightsholders as the only remaining 
essential input, bargaining theory tells us that they will successfully 
obtain most of the available surplus.'').\58\
---------------------------------------------------------------------------

    \58\ In full detail, Professor Watt concluded: ``[F]or every 
dollar that the statutory rate for musical works undercuts a fair 
and reasonable rate, the freely negotiated rate for sound recordings 
will increase by an estimated [REDACTED] cents. That is, if the 
musical works rate is increased to what would be a realistically 
fair and reasonable rate, then the negotiated fee for sound 
recordings would decrease almost dollar for dollar, with only a 
minor change in the total royalty rate for all copyrights 
combined.'' Id. at ] 23, n.13; see also id., appx. 3 at 12.

---------------------------------------------------------------------------

[[Page 54419]]

    To repeat: This inverse relationship is what has been described as 
the ``seesaw'' effect. The question in this regard on remand is whether 
the record proves that the seesaw theory is valid and measurable going 
forward. Alternately stated, does the record prove that Professor 
Watt's bargaining model serves as the linchpin that would allow the 
Judges to apply the Shapley results by increasing the mechanical rate, 
without unduly exposing the Services to the risk of higher total 
royalties?
    To resolve this issue, the Judges examine this bargaining model 
dispute in detail, as it bears on whether the uncapped TCC rate 
structure can be incorporated into the statutory rate.
(a) Bargaining Model Dispute
    Professor Watt utilized a general Nash Bargaining Model.\59\ In his 
particular application, Professor Watt modeled the streaming services 
and the labels each as a ``single unit,'' asserting (as is common in 
Shapley analyses) that this single-unit modeling was done ``for 
simplicity.'' Watt WRT, appx. 3 at 10. Applying this and other modeling 
assumptions, Professor Watt posited: ``If there were to be no 
successful deal, then each of these two bargainers [the assumed 
``single'' interactive service and ``single'' label] would earn 0, 
since in that case the interactive streaming service could not 
operate.'' Id.
---------------------------------------------------------------------------

    \59\ The Nash Bargaining Model is one type of game-theoretic 
approach used by economists to model the distribution of ``gains 
from trade'' between two parties ``in a manner that reflects 
`fairly' the bargaining strength of the different agents. Marx WDRT 
] 28 n.33 (citing A. Mas-Colell, M. Whinston, and J. Green, 
Microeconomic Theory 838 (1995)). To understand the parties' 
modeling dispute, it is necessary to appreciate the essential 
elements of the Nash Bargaining Model, as previously summarized by 
the Judges: ``In the Nash Framework [for full quotation, see eCRB 
no. 27063 n.48].'' SDARS III Final Determination, 83 FR 65210, 65215 
& n.32 therein (Dec. 19, 2018).
---------------------------------------------------------------------------

    In his oral testimony at the hearing, Professor Watt did not opine 
as to whether changes in variables other than musical works royalties 
would also have an impact on the level of sound recording royalty 
rates, even as higher musical works rates would otherwise place 
virtually 1:1 downward pressure on the sound recording rate. However, 
in his written rebuttal hearing testimony, i.e., his WRT, Professor 
Watt did make varying assumptions regarding the changes in the 
Services' non-content costs, by which he did change the total revenue 
share for content providers. Watt WRT ]] 50-52. He concluded from this 
varying replication of Professor Marx's Shapley model ``that the 
results that it delivers are very dependent upon the amount of total 
interactive streaming revenue and the fraction of that revenue that is 
taken up by downstream non-content costs.'' Id. at ] 53 (emphasis 
added).\60\
---------------------------------------------------------------------------

    \60\ The Judges take note here of Professor Watt's presentment 
of alternative scenarios, because, as discussed infra, the Services 
and their economists accuse Professor Watt of changing his 
testimony, post-remand, by limiting the scenarios in which his 
``seesaw'' argument would apply in order to salvage the credibility 
of his bargaining model.
---------------------------------------------------------------------------

    The Services had no procedural right under part 351 of the Judges' 
regulations to proffer surrebuttal written testimony from economic 
witnesses to challenge Professor Watt's assertion, made for the first 
time in rebuttal, of the seesaw relationship between changes in the 
musical works royalty rate and the sound recording royalty rate paid by 
interactive services. Moreover, the Services and their economists also 
had no opportunity to weigh in on the Majority's application of same 
(which was not revealed until the Judges rendered their decision). See 
Johnson, 969 F.3d at 381 (``Streaming Services had no notice that they 
needed to defend against and create a record addressing such a 
significant, and significantly adverse, overhaul of the mechanical 
license royalty scheme.'').\61\ Now though, on this remand, the 
Services have been afforded the opportunity to present these 
criticisms, through their expert witnesses.
---------------------------------------------------------------------------

    \61\ The Services could have sought leave to file surrebuttal 
testimony, and could have challenged the Majority's understanding of 
Professor Watt's testimony, after the Initial Determination, by 
filing a Motion for Rehearing pursuant to 37 CFR 353.1. However, a 
party is not required to engage in either of these procedural 
approaches, but rather may challenge the Determination on appeal, as 
has occurred here.
---------------------------------------------------------------------------

(b) Professor Katz's Principal Criticism
    Pandora's economic expert, Professor Michael Katz, levied several 
criticisms of the bargaining model proffered by Professor Watt and 
applied by the Majority. The most important problem with Professor 
Watt's analysis, according to Professor Katz, is that the former's 
model assumes an ``extremely unrealistic'' zero payoff to the label in 
the absence of an agreement with a streaming service--an assumption 
which is ``far from . . . innocuous.'' Written Direct Remand Testimony 
of Professor Michael Katz (Katz WDRT) ]] 16, 20.
    Professor Katz opines that this zero payoff assumption is 
equivalent to assuming, contrary to undisputed market facts, that: (1) 
subscribers and listeners to an interactive service would not switch to 
other interactive services if that service failed to reach an agreement 
with the labels; and (2) the interactive service is a ``Must-Have'' 
input supplier. Katz WDRT ]] 17-18. In terms of Nash modeling, 
according to Professor Katz, Professor Watt's assumption is thus 
equivalent to ``assum[ing] that the sound recording copyright owners 
have no outside option.'' Katz WDRT ] 127 (app. A) (emphasis added).
    Moreover, not only does Professor Katz assert the indisputability 
that such substitution would occur, he points out that Professor Watt 
himself acknowledged in his own testimony that such substitution would 
occur. Katz WDRT ] 19.\62\
---------------------------------------------------------------------------

    \62\ The Judges have quoted Professor Watt's testimony in this 
regard supra.
---------------------------------------------------------------------------

    Beyond this purported inconsistency, Professor Katz finds Professor 
Watt's no-substitution assumption to be a serious modeling error 
because, in order to quantify accurately each Nash bargainer's 
contribution to the net surplus to be divided, the extent of 
substitutability on each side of the market must be captured by the 
modeling. Katz WDRT ] 20. That is, he opines that ``Professor Watt's 
assumption that there is no substitution dramatically biases his model 
toward finding a large seesaw effect and renders his analysis 
unreliable . . . lead[ing]to a prediction that the share of an increase 
in musical works royalties that will fall on the streaming services is 
approximately eight times larger than Professor Watt's prediction. Id. 
at ] 21.
    As a matter of music business dynamics, Professor Katz interprets 
Professor Watt's substitutability error as follows.

    The assumption that a label receives a zero payoff if it does 
not reach agreement with a streaming service is equivalent to 
assuming that, if a streaming service shut down, none of the 
consumers who would otherwise have used that streaming service will 
switch to alternative streaming services or other sources of 
licensed music. The two forms of the assumption are equivalent 
because, when the services are substitutes, failure to reach an 
agreement with one service will not drive a label's payoffs from 
interactive streaming to zero. It will not result in the loss of all 
of the benefits that could be enjoyed by reaching an agreement. 
Instead, many consumers would engage in substitution and choose 
other streaming services, which will allow the label to earn profits 
from the additional royalties that would be paid to it by those 
other services.

Id. at ] 18.
    Professor Katz attempts to adjust Professor Watt's Nash Bargaining 
Model to account for this substitution effect. In his Appendix A, 
Professor Katz--acknowledging the reality of multiple interactive 
services--changes Professor Watt's assumed single label's payoff

[[Page 54420]]

(designated as parameter ``A'' in the Nash Bargaining Model) from a 
value of zero to a value equal to ``the share of revenues that would be 
diverted to other streaming services'' multiplied by ``the royalty rate 
that the label receives from the other interactive streaming 
services.'' Id. ]] 119, 127. Professor Katz asserts that the diversion 
to other streaming services represents an ``outside option'' available 
to a label. Id. ] 127. Professor Katz incorporates this ``outside 
option'' in his revised version of Professor Watt's Nash Bargaining 
Model.
    In addition, Professor Katz asserts that Professor Watt's modeling 
is unreliable because ``his prediction of the size of the see-saw 
effect is very sensitive to the assumed values of various other 
parameters.'' Id. at ] 23. For example, Professor Katz asserts that a 
change in the royalty rate paid to the labels could materially affect 
the balance or even the existence of the seesaw effect. Id. at ] 127. 
As further support for his opinion, Professor Katz relies on the 
testimony of one of Copyright Owners' own economic expert witnesses, 
who gave testimony clearly indicating that the ``seesaw'' effect was 
not at all likely to occur. Id. ] 24, n.16 (citing Gans WRT ] 32).\63\
---------------------------------------------------------------------------

    \63\ In this regard, Professor Gans testified: ``[When 
considering] the general distribution of profit when royalty rates 
for musical works rightsholders are increased[,] [i]n principle, 
those funds could come from a decrease in service profit, a decrease 
in sound recording royalties, or an increase in consumer pricing . . 
. . The general redistribution of profit in response to increased 
musical works royalties is fundamentally an empirical question. . . 
.'' Gans WRT ] 32.
---------------------------------------------------------------------------

    In sum, Professor Katz finds Professor Watt's Nash Bargaining Model 
to be unusable as a foundation to set royalty rates because, although 
``there are theoretical reasons to believe that a see-saw effect may 
occur, . . . there are complications and it is difficult to predict how 
big the effect will be.'' Id. ] 24 (emphasis added).
(c) Professor Watt's Rebuttal to Professor Katz
    In rebuttal to Professor Katz's criticisms, Professor Watt states 
that ``the record needs to be straight on Nash bargaining theory,'' in 
order to explain ``the foundational error'' committed by Professor 
Katz. Watt RWRT ] 52. This basic mistake, according to Professor Watt, 
is Professor Katz's erroneous assertion that the bargaining model must 
account for a label's ``outside option.'' Id. ] 53. Relying on economic 
authority regarding bargaining theory, Professor Watt defines an 
``outside option'' as ``the best alternative that a player can command 
if he withdraws unilaterally from the bargaining process.'' Id. ] 59 
(emphasis added); see also id. ] 53 (``An outside option is a payoff 
that the label would receive if negotiations with the service do not 
result in an agreement.'') (emphasis added).\64\
---------------------------------------------------------------------------

    \64\ The phrase ``outside option'' suggests the existence of an 
``inside option.'' Indeed, a treatise cited by Professor Watt 
identifies the ``inside option,'' defining it as ``[t]he payoff the 
[bargainer] obtains while the parties temporarily disagree''--
contrasting it with the ``outside option'' as (consistent with 
Professor Watt's testimony) ``the payoff [the bargainer] obtains if 
she chooses to permanently stop bargaining, and chooses not to reach 
an agreement with [the counterparty].'' A. Muthoo, Bargaining Theory 
with Applications at 137 (1999).
---------------------------------------------------------------------------

    Connecting this principle of bargaining theory to economic theory, 
Professor Watt explains his understanding of the relationship of the 
``outside option'' to the more familiar economic concept of 
``opportunity cost'':

    An outside option could also be referred to as an ``opportunity 
cost,'' since it is the value of what would be foregone should a 
deal with the service actually be struck. It is . . . useful to 
recognize the equivalence between an outside option and an 
opportunity cost, because economics in general has a very long 
history of understanding how opportunity costs weigh in on economic 
decision making.

Id.
    Professor Watt then opines how Professor Katz confused the 
``outside option'' with the disagreement (a/k/a threat) point in the 
Nash Bargaining Model:

    [Professor] Katz claim[s] that the outside option value that the 
labels would enjoy should they not reach an agreement with the 
services should be included as part of the ``disagreement point'' 
within the bargaining model and reimbursed like a cost prior to 
bargaining. Doing this can dramatically alter the results of the 
model. It is also definitively not how such an option should be 
modelled. [Professor] Katz [is] guilty of misunderstanding the Nash 
bargaining model, and concretely, the meaning of a ``disagreement 
point,'' and the way that an outside option should be brought into 
the model.

Id. ] 55.
    More particularly, according to Professor Watt, these outside 
options/opportunity costs do not belong in a Nash Bargaining Model, 
because they are ``not the types of status quo actual financial 
payments that may be modelled as disagreement points.'' Id. ] 57. 
Rather, he asserts that, as Professor Katz essentially acknowledged, 
they are ``payoffs from substitution, [i.e.,] an option instead of the 
deal, and they are not actual financial payments, but opportunity 
costs. Id.
    Professor Watt then explains that an outside option/opportunity 
that by definition exists as an alternative to a bargain between two 
parties lies outside the two parties' bargain, and is thus out-of-place 
within a proper Nash Bargaining Model:

    In the case at hand, if the parties never stop negotiating and 
never take up substitute options, then no joint enterprise is 
offered and there is no surplus to share, so each necessarily gets a 
payoff equal to 0, just as I assumed in my model.
. . .
    [A]gainst this backdrop, an outside option (a potential payoff 
that is not directly related to a share of the surplus that is being 
negotiated) . . . comes in [to the model] as a constraint upon the 
set of feasible deals that could be struck, exactly as an 
opportunity cost would be treated.

Id. ]] 57-58.
(d) Dr. Leonard's Criticisms of Professor Watt's Bargaining Model
    According to Google's economic expert witness, Dr. Gregory Leonard, 
the Majority wrongly relied on Professor Watt's bargaining model 
because it is ``highly stylized'' and theoretically ``simplified'' in 
ways that make it unable to predict that ``an increase in the musical 
works royalty would be offset nearly dollar-for-dollar by a decrease in 
the sound recording royalties (the ``seesaw effect''), thus leaving the 
services virtually unaffected by the proposed increase in musical works 
royalties.'' Leonard WDRT  8.
    Pointedly, Dr. Leonard criticizes Professor Watt's bargaining model 
as comprised of a ``veneer of `complexity' . . . mathematical formulas 
and [a] reference to John Nash,'' adopted to provide a rationalization 
for adoption of his Shapley Value modeling that would significantly 
increase the mechanical royalty rate.'' Id. ] 16. These modeling 
deficiencies, Dr. Leonard asserts, are not merely ``simplifying 
assumptions [that] better focus on the specific question the model is 
meant to address,'' but rather ``simplify away economic characteristics 
. . . entirely abstract[ing] away economic characteristics . . . 
central to the question at hand.'' Id. ] 18.
    In particular, Dr. Leonard avers that Professor Watt's bargaining 
model materially abstracts away from, inter alia: (1) the nature of 
consumer demand for streaming services and competing forms of music; 
(2) how services decide to enter or exit the streaming market; (3) the 
nature of the oligopolistic interaction among the labels; (4) the 
nature and timing of the bargaining between each label and each 
service; (5)

[[Page 54421]]

the potential for ``hold-up'' \65\ by labels that perceive the services 
to be in a vulnerable bargaining position due to their previous 
industry-specific investments made under their assumption that the pre-
existing statutory structure would be maintained; and (6) the failure 
of Professor Watt's bargaining model to grapple with the complementary 
oligopoly structure of the sound recording market. Id. ]] 18, 20.
---------------------------------------------------------------------------

    \65\ A hold-up problem occurs when: (1) parties to a future 
transaction must make specific investments prior to the transaction 
in order to prepare for it; and (2) the exact form of the optimal 
transaction (e.g., how many units if any, what quality level, the 
time of delivery) cannot be specified with certainty ex ante. W. 
Rogerson, Contractual Solutions to the Hold-Up Problem, Rev. Econ. 
Stud. 777 (1992). Here, the interactive services may need to commit 
to paying for long-term investments, even though they cannot know 
the level of their largest costs (content royalties) beyond a single 
rate term.
---------------------------------------------------------------------------

    These factors, he posited, are ``important for determining how 
sound recording royalties would actually change in response to a change 
in the statutory musical works royalty.'' Id. Professor Leonard 
concludes that, by not modeling these factors, Professor Watt's 
``prediction of a virtual dollar for dollar decrease in sound recording 
royalties is unreliable as a basis for formulating policy.'' Id. ] 20.
    Regarding the complementary oligopoly structure of the market and 
its impact on the bargaining process, Professor Leonard emphasizes that 
an important ``real-world hurdle'' assumed away by Professor Watt's 
modeling of a single label entity is that ``each label would prefer to 
have the other labels lower their sound recording royalties while 
maintaining its own royalties at pre-existing levels . . . .'' Id. ] 
21. More particularly, Dr. Leonard explains that ``even if a label were 
to recognize that it is more efficient for overall sound recording 
royalties to be lower, the label may not be willing to lower its 
royalty rate without assurance that the other labels will do the 
same,'' a result which he asserts ``is unlikely to happen absent some 
form of collusive behavior.'' Id. Thus, Dr. Leonard maintains that the 
existence and size of any ``seesaw''-induced decrease in sound 
recording royalties remains indeterminate, and it remains ``within the 
realm of theoretical possibility that the labels do not agree to any 
reduction in sound recording royalties even if a reduction in overall 
royalties would be economically efficient. Id.
(e) Professor Watt's Rebuttal to Dr. Leonard's Criticisms
    Professor Watt replies with a spirited defense of economic modeling 
in general and his economic bargaining model in particular. He begins 
by pointing out that models are not supposed to be ``perfect 
representations of reality [but rather] are intended to isolate what is 
important, in order to expose a useful insight on some issue of 
relevance.'' Watt RWRT ] 105. He adds that economic models (not merely 
his bargaining model) ``do not necessarily deliver predictions of 
situations that are immune to changes in variables outside the model, 
but rather the results inform conclusions about the relationships 
between the variables and parameters within the model, [which is] by 
nature a crude representation[ ] of reality, but the lessons and 
insights that they provide can be very relevant to real-world 
applications.'' Id. ]] 106-07 (emphasis added).
    With particular regard to his bargaining model, Professor Watt 
takes issue with Dr. Leonard's assertion that in the former's model the 
surplus is a ``fixed constant.'' See Watt RWRT ]] 110-111. Rather, 
Professor Watt avers that his bargaining model assume[s] that when the 
surplus . . . whatever value it takes . . . is to be shared, the 
parties understand that the amount to be shared is, at that moment, 
given.'' Id. ] 111 (emphasis added).
    Turning to Dr. Leonard's critique regarding the purported 
distortionary effect of Professor Watt's modeling assumption of a 
single label and a single interactive service, Professor Watt responds 
by acknowledging that, if he had modeled multiple labels and services 
in the bargaining process, that would be ``not particularly 
enlightening vis-[agrave]-vis the single bargain setting, as it will 
not lead to different insights than those distilled by the 
[Majority].'' Id. ] 113.\66\ Further, Professor Watt characterizes this 
criticism as ``empty,'' because under either his two-player Nash model 
or Dr. Leonard's posited multi-player (Nash-in-Nash) model, the labels 
will not respond to a musical works royalty increase ipso facto with a 
reduction in the sound recording royalty (i.e., the seesaw effect will 
not occur if there is ``a change in some other variable.''). Id. ] 114.
---------------------------------------------------------------------------

    \66\ Professor Watt describes Dr. Leonard's multiple 
simultaneous negotiations in a bargaining model as a ``Nash-in-
Nash'' model, but the former does not explain why he concludes that 
this approach ``will not lead to different insights'' than those the 
Majority distilled from his two-party Nash model.
---------------------------------------------------------------------------

 (f) Professor Marx's Criticisms of Professor Watt's Bargaining Model
    Professor Marx criticizes Professor Watt's application of the Nash 
Bargaining Model because, in her opinion, its ``precise prediction'' of 
the nearly one-to-one seesaw relationship ``depends critically on the 
assumptions that he makes and the numerical inputs that he uses.'' Marx 
WDRT ] 33. First, criticizing his modeling assumptions, like Professor 
Katz, she criticizes his decision to abstract from reality by positing 
a single label and a single interactive streaming service. She opines 
that his one label/one service modeling assumption ineluctably leads to 
his conclusion that each of these two parties ``has a `disagreement 
payoff' of zero [meaning that] each party ends up with nothing in the 
absence of a deal.'' Id. ] 34. But this zero ``disagreement payoff'' is 
merely a product of Professor Watt's abstraction from reality, 
according to Professor Marx, because ``[i]n reality, if interactive 
streaming went away, a share of the music listening that had occurred 
through interactive streaming services would migrate to other forms of 
music distribution, generating revenues for the label . . . meaning 
that the disagreement payoff would be positive for the label). Id. 
(emphasis added).\67\ Consistent with Professor Katz, she maintains 
that Professor Watt himself acknowledged the presence of this 
substitution effect when he testified that ``[t]he existing interactive 
streaming companies do not hold an essential input, as first they 
compete with the non-interactive services . . . .'' Id. ] 35, n.43 
(citing Watt WRT, app. 3).
---------------------------------------------------------------------------

    \67\ Professor Marx's reference to a substitution from a 
shutdown interactive service to ``other forms of music 
distribution'' is different from, but analytically analogous to, 
Professor Katz's assertion that the shutdown of any one interactive 
service would result in migration of its subscribers and other users 
to the remaining interactive services. These analogous critiques are 
complementary. See Marx WDRT ] 37 (``One would expect the same 
decrease in the estimated see-saw effect by including a second, 
competing interactive streaming service in the market instead of 
just the one that Professor Watt uses. In that case, if no deal is 
reached, users would migrate to an even closer substitute--a 
competing interactive streaming service--resulting in an even higher 
degree of profit migration and thus an even lower estimated see-saw 
effect'').
---------------------------------------------------------------------------

    More particularly, Professor Marx maintains, a record label's 
disagreement payoff must be considered realistically ``in any 
accounting of what would happen if record labels and interactive 
streaming services failed to reach an Agreement . . . .'' Marx RWDT ] 
35. And, she opines, when this real-world substitution effect is taken 
into account, the seesaw effect that Professor Watt estimates is 
reduced dramatically, because ``[t]he greater . . . the substitution 
between streaming and other forms of distribution, the greater is the 
revenue that the record label can capture in the event of disagreement

[[Page 54422]]

and the lower is the estimated see-saw effect.'' Id.\68\
---------------------------------------------------------------------------

    \68\ In the context of the bargaining model, Professor Marx 
identifies Professor Watt's choice of ``a market structure that is 
completely symmetric between record labels and services not 
reflective of the real world'' as forcing his model ``to attribute[ 
] all the . . . surplus division to . . . bargaining power . . . and 
none of it to the market structure.'' Id. ] 38.
---------------------------------------------------------------------------

    Professor Marx opines that modeling the bargaining process without 
these real-world particulars diminishes the value of Professor Watt's 
Nash model in several significant ways. First, because his model fails 
to incorporate the presence of three major record labels, ``each with 
substantial complementary oligopoly power,'' it fails to capture the 
fact that ``each record label does not fully internalize the impact of 
its rates on the viability of the industry.'' Id. ] 39. She points to 
the Judges' Final Determination in Web IV, where the Judges note how 
this aspect of complementary oligopoly compromises the value of a rate 
as a useful benchmark. Id. ] 39 n.45 (quoting Web IV Final 
Determination). More particularly, she opines that when, as here, 
``there are multiple negotiations between multiple record labels and 
multiple services,'' sound recording rates can be affected ``by the 
order of negotiations'' among the several label:service negotiating 
pairs--a factor that Professor Watt's bargaining model fails to 
capture. Marx WRDRT ] 41.
    Next, Professor Marx avers that Professor Watt's bargaining model 
``does not explain how or over what time frame the market would move to 
a new equilibrium.'' Id. ] 40. More particularly, she testifies, 
because interactive services' ``agreements with record labels often 
contain multi-year terms and can take many years to negotiate . . . 
there may be little incentive or practical ability for both sides to 
move to a new rate before the contract expires''. Id. ] 41. She takes 
note that this point was established at the hearing during questioning 
of Professor Watt from the bench:

    JUDGE STRICKLER: What of the situation . . . that the . . . time 
period for the existing agreements between the . . . labels and the 
interactive streamers is such that they've already locked in a 
particular rate and then we set a rate that's higher for the 
mechanical to reflect the fact that the sound recording royalty 
should drop, but it's locked in for a period of time? Are we running 
the risk, then, of disrupting the market by having a total royalty 
that's greater than what is indicated by your Shapley testimony, 
simply because of the disparity of times in which the rates are . . 
. implemented?
    PROFESSOR WATT: That's a very fair point. And I didn't even 
think of that until you've mentioned it . . . [T]he model I have 
done is . . . assuming that . . . the bargained thing happens at the 
same time as the--or in the same general period of time as a change 
in the statutory rate. You're absolutely correct.

3/27/17 Tr. 3091-92 (Watt); see Marx WRDRT ] 42, n.46

    Third, Professor Marx points out that Professor Watt's Nash model 
does not attempt to capture the effects of the heterogeneous and 
asymmetric distribution of information relevant to the bargain 
available to each party at the time of negotiation. Id. ] 41.
    Lastly, Professor Marx avers that Professor Watt's Nash Bargaining 
Model fails to address, on a more general basis beyond informational 
issues, other ``asymmetries among record labels and among services.'' 
Marx WDRT ] 41.
    In sum, Professor Marx concludes that these foregoing real-world 
points all preclude the Judges from relying on Professor Watt's 
testimony to identify a stable relationship between changes in the 
mechanical royalty rate and the sound recording royalty rate because 
they all share a common defect--they ``lie outside Professor Watt's 
model.'' Marx WRDT ] 41.
    To be clear, Professor Marx does not criticize Professor Watt for 
neglecting to include these points in his bargaining model; rather, she 
acknowledges that ``[t]hese are difficult features to capture in a 
tractable equilibrium model.'' Id. Indeed, she urges the Judges to 
appreciate that relying on such a necessarily limited model, as the 
Majority did, can have ``dramatic effects'' on the royalty rates 
derived. Id. Professor Marx emphasizes that all of these inherent 
modeling deficiencies are especially pernicious, if the bargaining 
model is applied yet again on remand, to set specific rates over a 
five-year period, when other variables will have independent effect on 
royalty rates. Id.
(g) Professor Watt's Rebuttal to Professor Marx
    Because Professor Marx's criticisms are of a similar nature to 
Professor Katz's criticisms, Professor Watt responds to Professor Marx 
as he did to Professor Katz. To summarize, Professor Watt responds to 
Professor Marx's points as follows:
     Her criticism is centered on what he characterizes as her 
``bogus'' argument that he supposedly had predicted almost a ``dollar 
for dollar'' sound recording rate reduction in response to an increase 
in the musical works rate (the seesaw effect). Watt RWRT ] 19. 
Professor Watt finds this argument ``particularly disheartening,'' 
because Nash bargaining theory explains why the seesaw would apply to 
the splitting of the surplus based on the available data, and that 
``there are quite apparent reasons why available surplus may not 
decrease even if the musical works rate increased, because of 
simultaneous changes to other variables in the model.'' Id. ] 34 
(emphasis added).
     Professor Marx implicitly contradicts her own reliance on 
the complementary oligopoly power of the Major labels by modifying his 
bargaining model through the insertion of a lower value for their 
bargaining power. Id. ]] 19, 22-24, 26.
     Professor Marx misconstrues the purpose of his Nash model, 
which was to serve ``as a reply'' to Professor Marx's direct testimony, 
and ``to show bargaining insights that bore upon aspects of the case.'' 
Id. ] 29.
     Professor Marx, like Professor Katz, improperly includes 
in her bargaining model a potential payoff for the label arising from 
an ``outside option,'' i.e., from an alternative that the label can 
choose only if the Nash bargaining terminates. Id. ]] 53--68.
(h) Professor Marx's Reply to Professor Watt's Criticism \69\
---------------------------------------------------------------------------

    \69\ The Judges found that Professor Watt's remand testimony, 
denoted as ``rebuttal,'' also provided de facto ``direct'' 
testimony, to which the Services could respond with supplemental 
testimony and argument. Oct. 1st Order at 11-12. Professor Marx's 
response in the following text was set forth in Spotify's permitted 
supplemental testimony.
---------------------------------------------------------------------------

    In her supplemental remand testimony, Professor Marx challenged 
several of Professor Watt's criticisms contained in his remand 
testimony. First, she takes issue with what he identified as two 
``core'' economic principles of bargaining: (1) that all of the 
available net surplus will be shared; and (2) that neither of the two 
bargainers will demand a share such that more than the total net 
surplus is shared. Marx WSRT ]] 7-8.
    As an initial matter, she disputes the notion that these are 
``core'' principles of bargaining. Id. ] 8. More particularly, she 
states that, in the present case, because ``the label does not know 
with exactitude the precise maximum that a service would be willing to 
pay (i.e., its ``survival'' rate), and the service likewise does not 
know the exact minimum that the label would be willing to accept,'' the 
simple bargaining model must be expanded to address ``the potential for 
delay and/or bargaining breakdown.'' Id.
    As a further criticism, Professor Marx avers that ``[i]n the real 
world, the negotiated royalty outcomes do not involve just two parties, 
but rather a sequence of overlapping, interrelated,

[[Page 54423]]

bilateral bargains involving multiple competing services and multiple 
record labels with complementary oligopoly power.'' Id. ] 12.\70\ This 
complication, she opines, exacerbates the informational deficit noted 
in the immediately preceding paragraph, such that negotiations within 
the several pairings of labels and services ``are affected by 
uncertainty and private information and . . . Professor Watt's 
discussion of bargaining theory [thus] does not support any particular 
real-world see-saw outcome.'' Id.
---------------------------------------------------------------------------

    \70\ In like manner, Professor Marx opines that Professor 
Spulber's discussion of bargaining theory is irrelevant to any 
assessment of ``the complexities affecting real-world negotiations'' 
and the presence, vel non, of a seesaw outcome. Id. ] 13.
---------------------------------------------------------------------------

(iii) Resolution of the Bargaining Dispute
(a) Professor Watt's Nash Bargaining Model Does Not Support Adoption of 
Uncapped TCC Rate
    The purpose of Professor Watt's Nash Bargaining Model was to allay 
the Judges' concern that increasing the mechanical rate would lead to 
higher total royalties for the Services. His bargaining model was 
understood by the Majority to show that such higher total royalties 
would not result, because the model demonstrated the ``seesaw'' effect, 
whereby the sound recording rate would fall almost dollar-for-dollar 
with the increase in the mechanical rate. See Determination at 73-74 
(``[T]he Judges rely on Professor Watt's insight . . . demonstrated by 
his bargaining model that sound recording royalty rates in the 
unregulated market will decline in response to an increase in the 
compulsory license rate for musical works. . . . Professor Watt's 
bargaining model predicts that the total of musical works and sound 
recordings royalties would stay `almost the same' in response to an 
increase in the statutory royalty.'') (emphasis added).\71\
---------------------------------------------------------------------------

    \71\ Copyright Owners note the Majority's recognition that, 
regardless of the rate structure, i.e., uncapped TCC or otherwise, 
Professor Watt's ``insight'' from ``bargaining theory'' would still 
apply. See Determination at 74, n.138. That being the case, the 
Majority's first rationale for adopting an uncapped TCC rate is 
undermined.
---------------------------------------------------------------------------

    On the surface, the economic experts on both sides appear to be at 
loggerheads regarding the existence and applicability of the seesaw 
relationship. However, as discussed below, on further analysis of their 
respective positions, in light of Professor Watt's remand testimony 
regarding a key assumption in his bargaining model, their disagreement 
narrows considerably and--in an important respect--vanishes 
completely.\72\
---------------------------------------------------------------------------

    \72\ This is unsurprising. The difference of opinion among 
economists often lies in their assumptions, which may be left 
unstated or opaque (intentionally or not). Once those assumptions 
are laid upon the table, their differences often evaporate. As the 
esteemed economist Fritz Machlup noted more than sixty years ago: 
``The most prolific source of disagreement lies in differences of 
factual assumptions. It is not customary for experts to state all 
the assumptions that underlie their conclusions; it would be much 
too cumbersome. But when they have reached very different 
conclusions, then we are forced to go back and find out what 
implicit assumptions they have made.'' F. Machlup, Why Economists 
Disagree, 109 Proceedings of the American Philosophical Society 1, 3 
(1965). In the modern world of more formal economic modeling as 
well, the obfuscation of assumptions continues to be an important 
source of dispute, according to a book written by a leading game 
theorist upon which Professor Watt relies in his testimony. A. 
Rubinstein, Economic Fables at 20 (2012) (``[T]he model's formal 
mantle enables economists . . . to conceal from the layman the 
assumptions the model uses.''); see J. Schlefer, The Assumptions 
Economists Make at 29 (2012) ([S]ome assumptions made by economists 
capture important insights, others are insane. All you have to do is 
decide which capture insights, which are insane, and in which 
situations.'')
---------------------------------------------------------------------------

    To recap: In his WRT, Professor Watt stated

    [W]ith an appropriately modelled bargaining analysis . . . in my 
Appendix 3 . . . I show that for every dollar that the statutory 
rate for musical works undercuts a fair and reasonable rate, the 
freely negotiated rate for sound recordings will increase by an 
estimated [REDACTED] cents.
    That is, if the musical works rate is increased to what would be 
a realistically fair and reasonable rate, then the negotiated fee 
for sound recordings would decrease almost dollar for dollar, with 
only a minor change in the total royalty rate for all copyrights 
combined.

Watt WRT ] 23 & n.13. But nowhere in his WRT did he qualify this 
statement by explicitly acknowledging that in his bargaining model 
there are certain assumptions lurking, i.e., that his ``concrete'' 
analysis is subject to the ``ceteris paribus'' constraint--that all 
other things are held constant (i.e., equal before and after the change 
in the musical works rate) other things being equal).\73\
---------------------------------------------------------------------------

    \73\ In his oral testimony, Professor Watt likewise did not 
qualify his opinion by taking note of his ceteris paribus 
assumption. See 3/27/17 Tr. 3026 et seq. (Watt).

    It is only in his later remand testimony--after the D.C. Circuit's 
remand had compelled him to confront criticism from adverse 
economists--that Professor Watt expresses this assumption overtly, 
making explicit the ``understanding'' that he had theretofore only 
---------------------------------------------------------------------------
tacitly assumed:

    In other words, a model in which only the two copyright rates 
are permitted to change . . . as was the understanding in my 
original model, allows the system to derive a clear relationship 
between those two rates, and that relationship is that an increase 
in one leads to a decrease in the other, that is, the `see-saw 
effect.' But if . . . something else changes along with the musical 
works rate . . . then the net effect does not predict that the 
negotiated rate of the labels will decrease.''

Watt RWRT ] 35 (emphasis added).
    Indeed, as noted supra, Professor Watt did give a nod to the 
relaxing of his implied ceteris paribus assumption in his WRT, by 
identifying varying ``scenarios'' in which he considered the impact of 
potential changes in service revenues and service non-content costs, 
leading to different percentages of royalties paid to content 
providers. Watt WRT ]] 45-52. Professor Watt then used these several 
assumptions and scenarios to opine as follows: ``The message that 
should be taken from this exercise . . . is that the results . . . are 
very dependent upon the amount of total interactive streaming revenue 
and the fraction of that revenue that is taken up by downstream non-
content costs.'' Id. ] 53.\74\
---------------------------------------------------------------------------

    \74\ Further, in his remand testimony, Professor Watt points out 
that Professor Katz made clear in his testimony that he applied the 
``all else equal'' assumption expressly in his own Nash bargaining 
analysis at the hearing. Watt RWRT ] 20 (quoting Katz WRT ] 67).
---------------------------------------------------------------------------

    Professor Spulber, on behalf of Copyright Owners, likewise 
emphasizes on remand the importance of the ceteris paribus assumption 
in economic modeling:

    [A]long with an increase in the compulsory license rate, all 
other things being equal, we would expect to see a decrease in sound 
recording royalty rates.
. . .
    ``All other things being equal'' (ceteris paribus in Latin), is 
a central principle for economic modelling. This economic analysis 
of bargaining highlights an important relationship between two 
content cost variables. However, that relationship does not exist in 
a vacuum. Many other variables affect the bargaining situation and, 
for any given period, the net effect of all of the different 
variables may be different than the effect of the modeled variable 
alone. Thus, this economic analysis of bargaining will not assure 
that a streaming service will not face disruption in the real world 
for any reason.
. . .
    Economic modeling is supposed to simplify the situation in order 
to distill useful principles and teachings.

Spulber RWRT ]] 26-28 (emphasis added).
    The Judges agree that the ceteris paribus principle \75\ is a 
fundamental

[[Page 54424]]

principle in economic analysis and modeling. Professor Watt succinctly 
makes this point, quoting the Nobel laureate economist James Buchanan, 
for the following proposition:
---------------------------------------------------------------------------

    \75\ The phrase is often translated into English as ``all other 
things equal.'' However, that is somewhat ambiguous. Equal to what? 
Not to other things. Rather, every ``thing'' (i.e., every other 
independent variable) whose effects are not being measured remain 
``constant,'' or ``controlled,'' i.e., ``equal'' to their measure 
prior to the change of the independent variable being examined. See 
W. Nicholson, Microeconomic Theory: Basic Principles and Extensions 
at 649 (9th ed. 2005) (defining ``ceteris paribus'' as ``[t]he 
assumption that all other relevant factors are held constant when 
examining the influence of one particular variable in an economic 
model'').

    At the heart of any analytical process lies simplification or 
abstraction, the whole purpose of which is that of making problems 
scientifically manageable. In the economic system we recognize, of 
course, that `everything depends on everything else,' and also that 
---------------------------------------------------------------------------
`everything is always changing'.

Watt RWRT ] 32 (quoting J. Buchanan, Ceteris paribus: Some Notes on 
Methodology, 24 So. Econ. J. 259, 259 (1958).
    However, Professor Watt does not quote another portion of Professor 
Buchanan's article that makes a point that looms large in the present 
proceeding, to wit, the limitations inherent in applying the necessary 
ceteris paribus condition:

    Real problems require the construction of models, and the skill 
of the scientist is reflected in the predictive or explanatory value 
of the model chosen. We simplify reality to construct these models, 
but the fundamental truth of interdependence must never be 
forgotten. . . . [However,] [f]ew, if any, meaningful results may be 
achieved by using ceteris paribus to eliminate the study of large 
numbers of variables. If such variables are closely related, they 
must be studied simultaneously; there is no escape route open.

Id. at 259-60 (emphasis added); see also A. Rubinstein, Comments on 
Economic Models, Economics, and Economists: Remarks on Economics Rules 
by D. Rodrik, 55 J. Econ. Lit.162, 167 (2017) ``[W]hat matters to the 
empirical relevance of a model is the realism of its critical 
assumptions'') (emphasis added).\76\
---------------------------------------------------------------------------

    \76\ The Judges note now that Professor Watt did not claim that 
his bargaining model generated any predictions, but rather that it 
explained the splitting of the Shapley surplus by the sound 
recording and musical works copyright owners, respectively, and the 
impact of that split on royalty rates, given the assumptions and the 
data in his model.
---------------------------------------------------------------------------

    This is not to say that Professor Watt was unaware of this caveat. 
As noted supra, he recognizes the difficulty of extrapolating from a 
ceteris paribus world to the real world. The present panel of Judges 
likewise recognizes this. However, the Majority missed this distinction 
in the Determination when it applied Professor Watt's correct but 
ceteris paribus ``insight'' for a constant real-world relationship 
between sound recording and musical works royalty rates. Again, not a 
single economist made this improper analytical leap or proposed an 
uncapped TCC rate in order to set a TCC ratio across the entire rate 
term. Indeed, on careful inspection, no economist states in his or her 
remand testimony that Professor Watt's bargaining model provides 
economic support for the uncapped TCC rate prong.
    With the foregoing testimony in mind, the Judges see particularly 
relevant several additional points in Professor Watt's remand rebuttal 
testimony that pertain to the appropriateness, vel non, of a TCC rate 
prong. Referring to the application of his bargaining model to the 
present case, Professor Watt made these crucial statements regarding 
the lack of a seesaw effect that would generate decreases in sound 
recording rates when the mechanical rate is increased:

    [T]he actual effects one would expect to see several years later 
would be based on the actual data at that time. Moreover, I would 
expect many other variables to have a larger effect on the bargains 
than the relatively small changes in the musical works rate. . . . 
[U]nderstanding actual market outcomes requires understanding these 
variables.
. . .
    [A]n attempt to capture all aspects of the real world is too 
complex for a simple statistical exercise involving an econometric 
regression. There is no obvious data to actually use for some of the 
independent variables, such as consumer demand equations, costs of 
entry and exit, a measure of oligopolistic interaction, different 
timings of different rate bargains, and the actual values of outside 
options.

Watt WRWT ]] 6(iv), 118.\77\
---------------------------------------------------------------------------

    \77\ In the language of econometrics, Professor Watt describes 
this problem as the ``almost sure[ ] impossibil[ity] of 
``introduce[ing] a control variable for each and every possible 
aspect that could potentially impinge upon the relationship [that] 
could easily lead to such a low R\2\, and/or statistically 
insignificant key coefficients, as to make the regression 
meaningless.'' Id. ] 118.
---------------------------------------------------------------------------

    Although Professor Watt was hardly transparent in disclosing his 
ceteris paribus assumption in his original testimony, it seems clear 
that he always understood its presence, and that, when this assumption 
was relaxed, ``the actual effects . . . several years later would be 
based on the actual data at that time [and] many other variables [with] 
a larger effect on the bargains than the relatively small changes in 
the musical works rate.'' Id. ] 6(iv) (emphasis added).
    Professor Spulber likewise opined that the absence of an explicit 
statement of these assumptions in Professor Watt's testimony was 
unremarkable and appropriate:

    [A]ll other things being equal'. . . should be generally read 
into economic modeling conclusions or predictions, whether or not 
the words are repeated in each instance. Economists do not typically 
repeat these words in each place where they apply, since it would 
lead to constant repetition.

Spulber RWRT ] 46, n.8.
    Regardless of whether economists invariably identify the existence 
of implicit assumptions lurking in each other's models, Professor Watt 
overlooked a cardinal rule of communication: Know your audience. Here, 
his audience is comprised of three Judges, only one of whom is also an 
economist.\78\ Failing to appreciate Professor Watt's implied ceteris 
paribus assumption, the Majority transformed his limited (albeit 
important) ``insight'' regarding the equal split of the Shapley surplus 
between the two classes of rights holders--and the seesaw effect that 
would have if the mechanical rate were increased when the split was 
imposed--into a justification for the imposition of an uncapped TCC 
rate prong over the five-year rate term. The Majority's language 
reveals this point clearly:
---------------------------------------------------------------------------

    \78\ The dissenting Judge (the only economist on the panel) 
warned that the seesaw effect was rife with assumptions that 
rendered it too speculative to be relied upon to support the 
uncapped TCC rate prong. See Dissent at 7-8.

    As to the issue of applying a TCC percentage to a sound 
recording royalty rate that is artificially high as a result of 
musical works rates being held artificially low through regulation, 
the Judges rely on Professor Watt's insight . . . demonstrated by 
his bargaining model that sound recording royalty rates in the 
unregulated market will decline in response to an increase in the 
compulsory license rate for musical works. See 3/27/17 Tr. 3090 
(Watt) (``[T]he reason why the sound recording rate is so very high 
is because the statutory rate is very low. And if you increase the 
statutory rate, the bargained sound recording rate will go down.'')
    Professor Watt's bargaining model predicts that the total of 
musical works and sound recordings royalties would stay ``almost the 
same'' in response to an increase in the statutory royalty. Id. at 
3091.

Determination at 73-74 (emphasis added).
    Making the point ever so plainly, Professor Watt now expressly 
acknowledges that his `` `see-saw effect' was never really a 
`prediction' '' at all! Watt RWRT ] 117. Rather, he now cautions the 
present panel of Judges, that, ``to make the jump from the model to the 
actual real-world effects, one cannot ignore the words that are 
omnipresent in all economic modeling,

[[Page 54425]]

that predictions about causal relationships are understood to be ``all 
else equal.'' Id. ] 32.
    Without the benefit of these caveats regarding an extrapolation of 
the ``seesaw'' theory to the real-world, and with absence of an 
explicit statement of the ceteris paribus assumption, the Majority 
misapplied his testimony as a basis to adopt a fixed TCC rate, based 
upon data from a snapshot in time (2016) to cement that rate 
relationship for the entire five-year period.\79\ The Majority 
misapplied Professor Watt's correct insight from bargaining theory 
regarding the use of a fixed ratio for the equal division by two ``Must 
Have'' input suppliers of the Shapley surplus to set royalty rates in a 
period, by using that insight incorrectly to establish a fixed ratio of 
royalty rates over the rate term.\80\
---------------------------------------------------------------------------

    \79\ The importance of Professor Watt's failure to make explicit 
the ceteris paribus assumption in his WRT is demonstrated by his 
need to make it explicit in his RWRT. But even now, rather than 
acknowledge that the Majority missed the point, he claims that the 
Services' are wrongly blaming the Majority for failing to understand 
this assumption: ``The Services' testimony on this remand seems 
primarily focused on creating a ``straw man'' argument . . . 
accus[ing] the [Majority] of something that the [Majority] did not 
do--that is, rely on a guarantee of a particular decrease in sound 
recording royalty rates--and the Services then attack the Board's 
determination by claiming that the decrease did not occur.'' Watt 
RWRT ] 5. As shown supra, however, this is precisely how the 
Majority interpreted Professor Watt's ``insight.'' The Judges 
understand that, as a matter of tact and tactics, Copyright Owners 
may be reluctant to acknowledge that the error lies in the 
combination of their witness's opaque testimony and the Majority's 
lack of understanding of the assumptions economists make. Copyright 
Owners might prefer to cast the Majority as the victims of the 
Services' incorrect accusation. But the plain language of the 
Determination belies Copyright Owners' characterization as to how 
the confusion arose.
    \80\ The forgoing analysis as applied to the uncapped TCC rate 
needs to be contrasted with the application of Professor Watt's 
bargaining model to increase the percent of-revenue rate to 15.1%. 
That higher rate was set by the Majority after its consideration of 
the same Shapley approaches, pursuant to the Judges' combination of 
inputs from Professor Gans model (his [REDACTED] round recording-to-
musical works ratio) and the Shapley Value Model of Professor Marx 
that adjusted for complementary oligopoly power by establishing a 
lower total royalty level ([REDACTED]%). But the difference is that 
the 15.1% revenue rate was set by applying the Shapley results based 
on actual and projected market data, see Gans WRT ] 38, whereas the 
uniform uncapped TCC rate (26.2%) was based on the ceteris paribus 
assumption that held constant the actual data regarding the 
aforementioned independent variables. As explained above though, 
Professors Watt and Spulber make it clear that the ``insight'' from 
bargaining theory did not have implications to allow for a 
``prediction'' of rates in future periods.
    Thus, when the Majority engaged in its analysis and ``line-
drawing'' to apply the data and market projections relied upon by 
Dr. Gans's data, the Majority was operating--to use the D.C. 
Circuit's phrase--in its ``wheelhouse,'' making a finding that 
withstood appeal. Johnson, supra, 969 F.3d at 385-86; see also 
Determination at 69-70 (``Professor Gans utilized data from 
projections in a Goldman Sachs analysis to identify the aggregate 
profits of the record companies and the music publishers, 
respectively. . . . The Judges also find Professor Gans's reliance 
on financial analysts' projections for the respective industries to 
be reasonable.'').
---------------------------------------------------------------------------

    Additionally, an examination of the expert economists' testimony 
reveals that their facial disagreements vanish once the necessary 
assumptions are laid bare. Professor Watt and the Services' three 
economists all identify the following independent variables that will 
impact the relative levels of sound recording and musical works rates 
paid by interactive services:
    (1) the level of downstream consumer demand;
    (2) entry costs;
    (3) exit costs;
    (4) oligopolistic interaction;
    (5) the timing of sound recording agreements vis-[agrave]-vis 
statutory rate setting; and
    Professor Watt and the three Service economists agree with regard 
to the relevancy of these six independent variables. Compare Watt RWRT 
]] 6(iv), 118 (identifying all five independent variables) with Leonard 
WDRT ] 18 (identifying independent variables 1-4 above); Marx WDRT ]] 
4-5, 42; (identifying independent variables 1-5 above); Katz WDRT ]] 
127, 134 n.115 (identifying independent variables 4 and 6 above). 
Accordingly, the remand record shows a consensus as to the lack of 
modeling of independent variables that would be important to estimate 
an uncapped TCC royalty ratio that could be utilized by the Judges to 
lock-in a ratio over the rate term.
    Indeed, as noted supra, a careful reading of the remand testimony 
by Copyright Owners' economists, Professors Watt and Spulber, reveals 
that neither of them actually testifies that there is sufficient 
theoretical and empirical evidence to support the uncapped TCC rate 
prong and the 26.2% TCC rate phased in on that prong. Rather, those two 
witnesses testify to something far narrower: the alleged correctness of 
Professor Watt's ``seesaw'' theory as demonstrating an equal splitting 
of the surplus between the two ``Must Have'' input suppliers, and the 
effect of that split when all other relevant independent variable are 
held constant.
    In this regard, it is noteworthy that none of Copyright Owners' 
several economic experts in this proceeding (Dr. Eisenach, Professor 
Gans, Dr. Rysman, or Professor Watt) ever proposed an uncapped TCC rate 
prong in any form, let alone within a greater-of formulation. Such a 
proposal would have been improper, because, as the expert testimony 
described above makes clear, the ceteris paribus assumption, reasonable 
for modeling purposes to provide insight as to the surplus split, lacks 
the input of the omitted variables that the experts on both sides find 
relevant to the application of economic modeling in this proceeding. A 
further review of Copyright Owners' economic expert witness testimony 
on remand--the first time any of them had occasion to weigh-in on the 
appropriateness of the uncapped TCC prong--reveals that they also have 
not endorsed the uncapped TCC rate prong as a proper form of rate 
setting. To be sure, they strongly endorse the insight first described 
by Professor Watt in his WRT that the Nash surplus would be split 
essentially evenly between the two suppliers of essential content, 
given his simplifying assumptions. But such endorsement is hardly the 
same as endorsement of the uncapped rate prong itself.
    For these reasons, the Judges find erroneous the Majority's 
identification of a fixed relationship between the sound recording and 
mechanical royalty rates that could serve as a basis for the Majority's 
first rationale for yoking the mechanical rate to an uncapped TCC rate 
prong.
(b) The Services Have Not Rebutted Copyright Owners' Prima Facie 
Showing That Professor Watt's Model Demonstrates a More Limited 
``Seesaw'' Effect
    The foregoing analysis and decision related to the absence of a 
fixed relationship between the sound recording and mechanical royalty 
rates. A separate fixed relationship--the one Professor Watt has 
clarified he was demonstrating all along--is that if the Judges 
increase the mechanical royalty rate, the Shapley surplus realized by 
the labels will decrease almost dollar-for-dollar with the increase in 
the mechanical rate. The Services' economists aver that even this 
version of the seesaw is defective.
    According to Professors Katz and Marx, the Nash Bargaining Model 
constructed by Professor Watt is deficient because it fails to properly 
characterize the ``disagreement payoff'' to the sound recording company 
when it and an interactive service fail to reach an agreement. More 
particularly, as explained supra, they assert that Professor Watt's 
model omits the value of ``outside options'' available to the sound 
recording company. This criticism relates to the issue of whether the 
seesaw effect would occur as posited in Professor Watt's model. That 
is, the increase in the sound recording

[[Page 54426]]

company's ``disagreement payoff'' (a/k/a ``threat point'') would lead 
to a higher royalty in the Nash bargain between the sound recording 
company and the interactive service than needed to generate the seesaw 
effect to offset the higher mechanical royalty rate.
    As the several experts' positions in this regard, discussed supra, 
make clear, however, each side has a different understanding of whether 
an ``outside option'' is properly included in the definition and 
calculation of the ``disagreement payoff.'' On the one hand, Professors 
Katz and Marx claim that the existence and value of ``outside options'' 
should be included in the ``disagreement payoff.'' However, they 
provide no economic authority for that assertion.
    By contrast, Professor Watt cites to multiple economic game theory 
publications and authorities for the proposition that the presence and 
value of ``outside options'' are not to be included in the 
``disagreement payoff'' contained in a Nash Bargaining Model. See A. 
Muthoo, Bargaining Theory with Applications at 105 (1999) (``I thus 
emphasize that the outside option point does not affect the 
disagreement point.''); M. Osborne & A. Rubinstein, Bargaining and 
Markets at 88 (1990) (``it is definitely not appropriate to take as the 
disagreement point an outside option. . . .''); K. Binmore, A. 
Rubinstein & A. Wolinsky, The Nash Bargaining Solution in Economic 
Modeling, 17 RAND J. Econ. 176, 185 (1986) (``An outside option is 
defined to be the best alternative that a player can command if he 
withdraws unilaterally from the bargaining process.'').
    According to Professor Watt and these authorities, the reason for 
excluding ``outside options'' from the Nash Bargaining Model is 
fundamental to the nature of the model itself. In the Nash approach, 
the negotiating parties are bargaining with each other only over the 
surplus their deal can generate, and they are attempting to agree upon 
an allocation of that surplus that exists within the bounds of their 
respective ``disagreement payoffs.'' Each may have ``inside options,'' 
which are alternatives available to them while bargaining is ongoing 
and they temporarily disagree. See Muthoo, supra, at 137. However, 
``outside options'' are available to a Nash bargaining party only in 
lieu of continuing the Nash bargaining with the original counterparty 
if it ``withdraws'' from the Nash bargaining process. See Binmore et 
al., supra. Professor Watt characterizes the distinction as follows:

    [T]he Nash bargaining model [is] designed as [a] self-contained 
portrayal[ ] of negotiating behavior. . . . Given a surplus to 
share, the Nash model . . . provide[s] allowance for financial 
payments that a party is actually receiving, only while negotiations 
are ongoing, without walking away for another option, and that would 
cease as a result of the deal, to be factored into modelling as a 
cost in some situations.'')
. . .
    [A]n outside option (a potential payoff that is not directly 
related to a share of the surplus that is being negotiated) . . . 
comes in as a constraint upon the set of feasible deals that could 
be struck. . . .''

Watt RWRT ]] 56, 58.\81\
---------------------------------------------------------------------------

    \81\ Professor Marx in fact cites several of these authorities 
(for other points), without noting the distinction they make between 
the appropriate inclusion of ``inside options'' and exclusion of 
``outside options'' in Nash modeling. See id. ] 59.
---------------------------------------------------------------------------

    The Services never sought to introduce further testimony regarding 
this important dispute. This is particularly striking because the 
Services filed a motion to strike certain portions of the CO Reply, or 
for leave to file supplemental testimony responsive to those itemized 
portions. The portions the Services identified in their motion did not 
include Professor Watt's criticisms as to the inclusion of ``outside 
options'' in their experts' Nash modeling. Further, after the Judges 
granted the Services' motion by providing them leave to file 
supplemental testimony--consistent with the designations in their 
motion--the supplemental testimonies did not address this ``outside 
options'' issue.
    In the course of discussions among the parties and the Judges 
regarding remand procedures, the Judges invited the parties to produce 
witnesses for a hearing, at which one or more of the Services' economic 
expert witnesses could have addressed this ``outside options'' issue. 
However, the Services (and Copyright Owners) waived the opportunity to 
produce witnesses at a hearing. Rather, they offered, and the Judges 
agreed, that they would stand on their written testimonies and proceed 
to closing arguments by counsel.
    In the closing arguments, each side argued numerous points of 
controversy and provided the Judges with dozens of demonstrative aids 
summarizing record evidence and the parties' arguments, but none of 
those arguments or demonstrative aids so much as mentioned this 
``outside options'' dispute. Moreover, when the Judges inquired during 
closing arguments as to whether Services' counsel would be addressing 
any of the experts' ``modeling disputes,'' counsel said that they were 
resting on their papers. 3/8/22 Tr. 86-87 (Closing Argument). 
Similarly, when the Judges inquired of Copyright Owners' counsel 
whether he would be addressing the modeling ``dust-up'' between 
Professors Watt and Katz, counsel demurred, stating that although he 
would ``love to engage on it but . . . ``there would be too many 
slides. . . .'' Id. at 262-64.
    Simply put, the Services' economic experts made an assertion 
regarding the need for Professor Watt to have included ``outside 
options'' in his Nash Bargaining Model, but Professor Watt presented 
authority clearly stating that such inclusions would be improper. Thus, 
Copyright Owners made a prima facie showing that in a Nash Bargaining 
Model, the surplus generated by the streaming surpluses acquired by the 
content providers would be split equally as between the sound recording 
licensors and musical works licensors, and that, ceteris paribus, an 
increase in the mechanical rate to provide Copyright Owners more of the 
surplus (per the Shapley-based results relied on by the Majority) would 
be essentially offset through a nearly 1:1 reduction in the sound 
recording rate. In response to Copyright Owners' prima facie case, the 
Services stood mute in response to the rebuttal argument claiming that 
their experts misapprehended the Nash modeling distinctions between 
``inside options'' and ``outside options.'' \82\
---------------------------------------------------------------------------

    \82\ The third economic expert for the Services, Dr. Leonard, 
did not utilize the ``outside option'' phraseology to describe his 
critiques. Rather, he first criticized Professor Watt for assuming 
the existence of a ``fixed surplus.'' Leonard WDRT ] 16. However, as 
discussed supra, that assumption came from the Majority's 
extrapolation from Professor Watt's hearing testimony. His explicit 
statement regarding the ceteris paribus assumption makes clear that 
he was not assuming a ``fixed surplus.'' Watt RWRT ]] 110-11. 
(Again, the only ``fixed'' surplus was not ``assumed,'' but rather 
quantified, in order to establish the Majority's percent-of-revenue 
prong royalty rate of 15.1%.)
    Dr. Leonard next claims that Professor Watt's assumption that 
the labels would bear virtually the entirety of an increase in the 
statutory rate, because they previously ``have captured almost all'' 
[the] surplus,'' has been contradicted by the evidence. 
Specifically, he refers to the 33-month period in which the 
Phonorecords III rates were effective (January 2018 through 
September 2020). Leonard WDRT ] 16. However, as the Judges find in 
this Determination, that 33-month period was marked by significant 
uncertainty with regard to the ultimate rates and rate structure 
(and the rates were being phased-in), so no findings could reliably 
be made based on sound recording rate changes during that period.
    The remainder of Dr. Leonard's critique concerns issues that 
would make a fixed TCC ratio inappropriate over the rate term. The 
Judges agree with those criticisms as previously discussed, but they 
do not pertain to this narrower issue of whether the surplus 
generated by interactive streaming would be split in a manner 
consistent with Professor Watt's Nash Bargaining Model.
---------------------------------------------------------------------------

    Accordingly, the Judges find that the Services' criticisms in this 
regard are insufficient to rebut Copyright Owners' prima facie showing 
that Professor Watt's Nash Bargaining Model properly

[[Page 54427]]

identified and valued the ``disagreement payoff.'' 83 84
---------------------------------------------------------------------------

    \83\ To be clear, the Judges' ruling is narrow; they make no 
finding beyond crediting this prima facie showing and the failure of 
the Services to rebut sufficiently that showing. It might be the 
case that the existence and definition of ``outside options''--and 
their relationship to ``inside options''--have other implications 
vis-a-vis a Nash Bargaining Model applied in the context of a rate 
setting proceeding. However, the Judges may not introduce and rely 
on analytical approaches not developed by the parties. See Johnson, 
969 F.3d at 381 (the Judges must not ``procedurally blindside[ ]'' 
the parties with an ``approach . . . first presented in the 
determination and not advanced by any participant.''). See generally 
P. Wald, Limits on the Use of Economic Analysis in Judicial 
Decisionmaking, 50 J. L. & Contemporary Problems 225, 228 (1987) ('' 
judicial analysis, economic or otherwise, takes place only in the 
context of lawsuits between two or more parties imposes a practical 
constraint on the judge's ability to use economic analysis.'').
    \84\ Professor Katz also criticizes Professor Watt's assumption 
that ``a label's non-content costs are proportional to licensing 
revenues.'' Katz WDRT ] 22. More particularly, Professor Katz claims 
that this is not ``plausible'' because ``the royalty rate does not 
directly affect the sound recording copyright owners' non-content 
cost.'' Id. ] 133. The effect of eliminating this assumption, 
according to Professor Katz, is to reduce the seesaw effect in 
Professor Watt's model of [REDACTED] slightly further away from a 
1:1 ratio, to .92. Id.
    In rebuttal, Professor Watt says this criticism is inconsistent 
with Professor Katz's own analysis, because the latter also ``sets 
the cost equal to a fraction of revenue. . . .'' Watt ] 82 n.31 
(referring apparently to a comparison of Katz WDRT ] 129 with id. ] 
133). Professor Watt concludes that not only does ``[Professor] 
Katz's own model contain the same feature that he is critical of in 
my model,'' it is also ``not a flaw in the bargaining model.'' Watt 
] 82. As a substantive matter, Professor Watt defends the assumption 
that non-content costs would rise with royalty income, because 
``[g]reater revenue should be directly equated with a larger scale 
of business'' and ``the additional royalty income would have to be 
managed (i.e., distributed to those who need to be paid from it, 
such as artists), implying higher administration costs.'' Id. ] 79.
    The Judges find that the common use by both experts of this 
assumed proportionality of a label's non-content costs to licensing 
revenues alone blunts Professor Katz's criticism of Professor Watt's 
modeling. Further, Professor Watt reasonably posits that higher 
revenue would imply a larger scale of business with associated 
general cost increases. (But the Judges do no agree that it was 
reasonable for Professor Watt to assume that distribution and 
administrative costs in particular would increase merely because of 
an increase in royalty rates; simply paying more money, ceteris 
paribus, is not self-evidently associated with an increase in 
costs.)
---------------------------------------------------------------------------

b. Rejection of Second Rationale for Including Uncapped TCC Rate Prong
    In the Determination, as noted supra, the Majority also justified 
the adoption of the uncapped TCC rate prong because it had the effect 
of ``import[ing] into the rate structure the protections that record 
companies have negotiated with services to avoid the undue diminution 
of revenue through the practice of revenue deferral.'' Determination at 
36; see also Johnson, 369 F.3d at 372 (``By pegging the mechanical 
license royalties to an uncapped total content cost prong, the Board 
sought to ensure that owners of musical works copyrights were neither 
undercompensated relative to sound recording rightsholders, nor harmed 
by the interactive streaming services' revenue deferral strategies. . . 
.'') (emphasis added).
(i) Parties' More Specific Arguments
    Copyright Owners likewise argue that the uncapped TCC rate 
structure should be ``adopted to provide protection against revenue 
deferment and displacement in a revenue-based rate structure.'' CO 
Initial Submission at 38; see also id. at 40 (describing uncapped TCC 
rate prong as ``critical backstop in a revenue-based rate 
structure.'').
    Whereas Copyright Owners echo the Majority, the Services adopt the 
reasoning of the Dissent. They argue as follows:

    [A] rate structure with a capped TCC prong, like the 
Phonorecords II settlement, achieves the same goal of protecting the 
Copyright Owners from any potential revenue deferral through a 
``structure that provides alternate rate prongs and floors, below 
which the royalty revenue cannot fall,'' . . . and does so without 
allowing Copyright Owners to impermissibly share in the labels' 
complementary oligopoly power. . . . [T]he streaming industry has 
twice concluded, after extensive negotiations, that the appropriate 
way to address any concerns regarding revenue deferral is to have a 
rate structure that includes a capped TCC prong. Phono I, 74 FR 
4510; Phono II, 78 FR 67938.

Services' Joint Opening Brief at 62 (quoting Dissent, 84 FR 1990) 
(emphasis added).
    In their Reply, Copyright Owners argue that the Majority maintained 
the benefits of price discrimination contained in the prior 
Phonorecords II framework, but balanced that goal with added protection 
against Service revenue deferral and displacement. Copyright Owners' 
Reply Brief on Remand at 49 (``In adopting a rate structure with [an 
uncapped] TCC for all service offerings, the [Majority] balanced its 
concerns about fostering price discrimination while also protecting 
against proven revenue diminution by the Services.'').
    The Services, in their Reply, take note that pre-remand, Copyright 
Owners had strenuously objected to any yoking of the mechanical royalty 
rate to the sound recording rate, maintaining that, although the 
Copyright Owners now advocate for an uncapped TCC rate to protect 
against revenue displacement and diminution:

    [I]n their [pre-remand] reply proposed findings, the Copyright 
Owners had expressed a very different view, arguing that an uncapped 
TCC prong ``does nothing to protect Copyright Owners from the 
Services' revenue displacement and deferment'' [and] Copyright 
Owners have not even tried to explain away their complete about-face 
on this issue.

Services' Reply at 43.
(ii) Analysis and Decision Regarding Revenue Diminution or Deferral
    The Judges find that the second rationale put forth to support an 
uncapped TCC rate does not justify the adoption of that rate prong. 
Several reasons support this finding.
    First, there is insufficient evidence to show how the sound 
recording companies contractually structure their own royalty rates, 
which would constitute the rate base for an uncapped TCC rate for the 
mechanical royalty. The sound recording royalty rate, when proffered 
for use as a mechanical royalty rate base, is analogous to pegging the 
value of a foreign currency to the U.S. dollar. That is no mere 
benchmark. The Judges must have the benefit of sufficient record 
evidence to demonstrate that the pegging (or, to use the D.C. Circuit's 
word in Johnson, ``yoking'') of a statutory rate to an unregulated rate 
serves the statutory purposes for the rate at issue, here, the 
mechanical rate.
    But Copyright Owners presented virtually no evidence regarding how 
the sound recording companies structure their interactive service 
royalties. Indeed, in the hearing, Dr. Eisenach acknowledged that the 
``relative value of sound recording [to] musical works licenses may 
depend on a variety of factors,'' but he intentionally eschewed 
unnecessary ``assumptions, complexities and uncertainties associated 
with theoretical debates'' as to why the particular market ratios 
existed. See Determination at 44. Indeed, the Majority found fault with 
Dr. Eisenach's willful ignoring of these issues, agreeing with the 
Services' criticism that Dr. Eisenach's ``use of sound recording 
royalties paid by interactive services embeds within his analysis the 
inefficiently high rates that arise in that unregulated market through 
the complementary oligopoly structure of the sound recording industry 
and the Cournot Complements inefficiencies that arise in such a market. 
See Determination at 47. The uncapped TCC rate advocated now by 
Copyright Owners suffers from the same affliction.
    The only reference to such sound recording rate formulae in 
Copyright Owners' voluminous PFF after the hearing was its statement 
that the effective revenue calculations in two of the Major labels' 
agreements with the

[[Page 54428]]

services was based on [REDACTED]. See Copyright Owners' PFF ]] 72, 91 
(cited post-remand at Copyright Owners' Motion for Reconsideration or 
Clarification at 25, n.14). On remand, the Services have provided a 
further summary of the types of [REDACTED]. See White WDRT ]] 6-7, 14-
15, 20, 24-26, 28-29 ([REDACTED]); Bonavia WDRT ]] 15-17 ([REDACTED]); 
Mirchandani WDRT ]] 16, 21-24 ([REDACTED]). Clearly, the levels of 
[REDACTED] would have to be weighed and the impact of complementary 
oligopoly power would need to be identified in order to adjust the rate 
prongs to account for that power. But the record is devoid of such 
details.
    Second, compounding this problem, because the uncapped TCC rate is 
embedded in a ``greater-of'' rate structure, the labels can exploit 
their complementary oligopoly power when creating the switching points 
that toggle royalty payments between and among rate prongs. As the 
Judges have explained previously, in declining to import a ``greater 
of'' structure from the unregulated interactive market, this 
structure[it] is based on ``agreements [which] were all negotiated in a 
market characterized by the lack of effective competition, and that the 
lack of competition would affect the structure as well as the level of 
rates.'' SDARS III, 83 FR 65210, 65228 (Dec. 19, 2018) (emphasis 
added). Further, the Judges held therein that the ``advantageous'' 
nature of a ``greater-of'' structure to sound recording licensors ``may 
well represent an example of what licensors can and would obtain when 
they exploit their ``must have'' status for a special competitive 
advantage.'' Id.; see also Dissent at 47 (in absence of testimony 
explaining how greater-of structure is consonant with effective 
competition, use by licensor suggests a game of ``heads I win tails you 
lose.).''
    Thus, there is insufficient evidence or testimony that would permit 
the Judges to make any adjustment for the complementary oligopoly power 
that may be built into each prong of the sound recording royalty rate 
structures.
    Third, as the Services note, Copyright Owners pre-remand, opposed 
the identical rate structure--consisting of a percent-of-revenue prong 
and an uncapped TCC prong--before Copyright Owners were in favor of it, 
post-remand.\85\ Although Copyright Owners took a 180-degree turn on 
this issue, they never stated they were wrong to oppose it previously. 
Indeed, the Dissent relied upon Copyright Owners' strenuous objection 
to an uncapped TCC rate, quoting it verbatim:
---------------------------------------------------------------------------

    \85\ When Copyright Owners opposed the concept of an uncapped 
TCC rate prong in a greater-of structure, the proposed uncapped TCC 
rate was Google's 15% (and its proposed percent-of-revenue rate was 
10.5%). Determination at 13. But after the Majority set the uncapped 
TCC rate at 26.2%--a 75% increase over the 15% TCC rate--Copyright 
Owners became zealous converts to the concept of an uncapped TCC 
rate proper.

    Copyright Owners rightly note that they obtain no legal 
protection under such a TCC prong. In making this argument regarding 
displacement and deferral of revenue, Copyright Owners lay out 
comprehensively all the problems inherent in an uncapped TCC prong 
set in a greater of rate structure, such as adopted in the majority 
opinion:
    The notion that [the] TCC prong will provide protection from 
revenue gaming, deferral and displacement, and other revenue prong 
problems is unsupported and speculative. Relying on just the TCC to 
solve those admitted problems leaves the Copyright Owners' 
protection from such problems entirely outside the statute. . . . 
the per-user rates in the label deals are what protects the 
Copyright Owners from price-slashing by the services. What is left 
unanswered . . .is . . . how can it be reasonable to ask the Judges 
to set a rate that does not itself provide for a fair return . . . 
but simply puts the Copyright Owners' fair return in the hands of 
the labels to negotiate terms that will adequately protect the 
publishers and songwriters as well? The labels do not have a mandate 
to ensure that the Services provide a fair return to the Copyright 
Owners, and cannot be directed to ensure such. Indeed, labels may 
not have the same incentives as songwriters and publishers to 
negotiate such protections in their deals. To wit, a label could 
make an agreement with a service that includes only a revenue prong 
in exchange for equity or some other consideration that it may never 
include in the applicable revenue subject to the TCC. . . . [W]hat 
if Google purchased one or more record labels and did not have to 
pay any label royalties? Or what if Spotify chose to avail itself of 
the compulsory license to create its own master recordings embodying 
musical works--which it is already doing . . . and chose to 
compensate itself for its use of the master recordings on a 
sweetheart basis (or not at all)? Or what if one or more labels 
decided to enter the interactive streaming market and did not have 
to pay themselves royalties? In each case, the Copyright Owners' 
protection--the protection that the Services admit the Copyright 
Owners need and is provided by the TCC--would be gone.

Dissent at 5-6 (quoting Copyright Owners' RPFF-Google at 39-41) 
(emphasis added). To make the identical point post-remand, but from the 
Services' perspective, Pandora's economic expert witness, Professor 
Katz, simply utilizes Copyright Owners' verbatim language (bolded 
above), but substitutes the word ``Services'' for ``Copyright Owners'' 
(and ``income'' for ``return'') to highlight how reliance on the sound 
recording royalty rate is improper:

    What is left unanswered . . . is . . . how can it be reasonable 
to ask the Judges to set a rate that does not itself provide for a 
fair income . . . but simply puts the Services' fair income in the 
hands of the labels to negotiate terms that will adequately protect 
the Services as well? The labels do not have a mandate to ensure 
that the Copyright Owners provide a fair income to the Services, and 
cannot be directed to ensure such.

Katz WDRT ] 71.
    The Judges find this argument persuasive, both in its own right and 
in the fact that it has been advanced by Copyright Owners and the 
Services alike.\86\
---------------------------------------------------------------------------

    \86\ At Closing Arguments on remand, Judge Strickler queried 
counsel for Copyright Owners regarding their prior rejection of an 
uncapped TCC prong within a ``greater-of'' rate structure. Counsel's 
response was that an uncapped TCC doesn't provide enough protection 
against revenue diminution: ``It provides more than the Phonorecords 
II rates, but not as much as we want,'' although ``still better 
than'' the negotiated Phonorecords II approach. 3/8/22 Tr. 240-
41(Closing Argument). But Copyright Owners have neither 
distinguished nor disavowed their persuasive legal point quoted in 
the text above, to wit that an uncapped TCC rate would be 
unreasonable if the ``protection'' it affords lies ``entirely 
outside the statute.'' Whether the ``protection'' relates to 
Copyright Owners' concern over revenue diminution or to the 
Services' concern over uncapped mechanical rates, the legal defect 
is the same--the unreasonableness of leaving the purported 
protection ``entirely outside the statute.''
---------------------------------------------------------------------------

    Fourth, the Judges note that the Majority did not find that revenue 
diminution, via displacement, deferral, or otherwise was pervasive, as 
Copyright Owners aver. Compare CO Initial Submission at 40 (``The 
record overwhelmingly established that the percent of revenue prong 
often results in musical works royalties that are too low . . . 
drive[n] [by] . . . . revenue deferral [and] revenue displacement'') 
with Determination at 21 (``The Judges agree that there is no support 
for any sweeping inference that cross-selling has diminished the 
revenue base.'') (emphasis added) and 36 (``The Judges find that the 
present record indicates that the Services do seek to engage to some 
extent in revenue deferral in order to promote their long-term growth 
strategy.'') (emphasis added).
    Given that the Majority found revenue diminution through 
displacement and/or deferral exists only ``to some extent'' and is not 
a ``sweeping'' issue, the Judges on remand find that the uncapped TCC 
rate structure creates the potential for unbalanced harm. As noted 
supra, the only protection against runaway mechanical rates, the seesaw 
hypothesis, cannot justify yoking the mechanical rate to a fixed ratio 
with the

[[Page 54429]]

unregulated sound recording rate.\87\ By contrast, and as discussed 
infra, the Phonorecords II-based benchmark approach, despite its own 
imperfections, is superior in this regard, because its series of 
alternate rate prongs and floors represents a negotiated compromise 
(negotiated by trade associations with countervailing power) between 
the potential for revenue diminution that would harm Copyright Owners, 
on the one hand, and the potential for runaway mechanical rates (yoked 
to the sound recording companies' complementary oligopoly power) that 
would injure the Services, on the other.
---------------------------------------------------------------------------

    \87\ Even Google, the party that, post-hearing, broached in its 
PFF the idea of an uncapped TCC prong, candidly identified the risk 
arising from an uncapped TCC: ``Having no cap on TCC . . . leaves 
the services exposed to the labels' market power, and would warrant 
close watching if adopted. . . .'' Google PFF ] 73 (emphasis added). 
But as the Dissent noted, there is no satisfactory way to monitor an 
uncapped TCC rate prong: ``Who would do the ``watching''? When would 
such watching occur? Congress directed the Judges to be the 
``watchers,'' and Congress instructed that the ``watching'' should 
occur only through rate proceedings. . . .'' Dissent at 4 (emphasis 
in original).
---------------------------------------------------------------------------

(iii) Distinction Between the ``Reasonable'' Rate Statutory Standard 
and the Factor (D) Objective To Minimize ``Disruptive Impact''
    The Judges next consider an issue emphasized by Copyright Owners: 
whether the Services have demonstrated that the uncapped TCC rate prong 
would cause a ``disruptive impact'' as set forth in Factor (D) of 
section 801(b)(1).\88\
---------------------------------------------------------------------------

    \88\ Separate and apart from the ``disruptive impact'' argument 
made by Copyright Owners, there is no need to consider how this 
prong would relate to Factor D, because the Judges find the uncapped 
TCC rate prong with the (phased-in) 26.2% rate to be 
``unreasonable.'' If it were necessary to separately consider the 
four itemized factors, the Judges would confirm that Factor A is 
satisfied, because, as the D.C. Circuit found, the Majority 
reasonably found that rates should increase from the Phonorecords II 
period, and the 15.1% revenue rate represents a 44% increase. The 
Judges would also find Factors B and C to be satisfied without a 
separate uncapped TCC rate prong. The reason is that, under the 
section 801(b)(1) standard, the ``reasonableness'' standard filters 
out more statutorily infirm rates than the fairness objectives. By 
contrast, when a rate does satisfy the ``reasonableness'' standards 
under section 801(b)(1), the Judges must also consider the rate 
through the finer ``fairness'' filter. Cf. Determination at 68 & 
n.120 (distinguishing between: (1) a Shapley Value analysis that 
filters out unreasonable rates by reducing licensors' ability to 
abuse market power by threatening or exercising their refusal to 
license (``hold-out or ``hold-up'' power); and (2) a Shapley Value 
analysis that further filters out unfair rates by going beyond 
eliminating abuse of market power to also make a ``market power 
adjustment'' explicitly to address Factors B and C). Finally, as the 
text infra, explains, the Judges also find no basis under Factor D 
to alter their analysis.
---------------------------------------------------------------------------

    Section 801(b)(1) provides that one of the competing priorities of 
the Judges in setting the mechanical rate is ``[t]o minimize any 
disruptive impact on the structure of the industries involved and on 
generally prevailing industry practices.'' 17 U.S.C. 801(b)(1)(D). In 
Johnson, the D.C. Circuit did not identify any argument by the Services 
that was predicated on a claim that this statutory form of 
``disruption'' had occurred, or was likely to occur, as a consequence 
of the Majority's rates and rate structure. Additionally, the D.C. 
Circuit did not ground its decision to vacate and remand the Judges' 
uncapped TCC rate and rate structure rulings based on the potential 
that these rulings would be disruptive to the Services, let alone would 
cause a statutory ``disruptive impact.''
    After the D.C. Circuit's ruling, an argument regarding 
``disruption'' was first made by Copyright Owners, not the Services. 
Copyright Owners argued that the vacated rates should nonetheless be 
maintained as interim rates, during the pendency of the remand 
proceeding. Motion of Copyright Owners to Adopt Interim Rates and Terms 
Pending the Remand Determination, passim (Nov. 2, 2020). Copyright 
Owners argued that reverting to the rates that existed before the 
Determination would constitute a ``disruption'' and self-servingly 
predicted that the Services would attempt to argue that the uncapped 
TCC rate and rate structure were themselves ``disruptive.'' Copyright 
Owners opined that such an argument would be a ``hollow exercise.'' Id. 
at 12, n.5; see id. at 2-3, 9 (claiming absence of disruption from 
uncapped TCC rate and structure despite absence of such argument by 
Services).
    In response to that motion, the Services did not assert that the 
Majority's uncapped TCC rates and rate structure would constitute 
disruption or have disruptive impact, whether under statutory Factor D 
or otherwise. See Services' Opposition to the National Music 
Publishers' Association (NMPA) and Nashville Songwriters Association 
International's (NSAI) ``Interim Rates Motion'' (Nov. 18, 2020). In 
reply, Copyright Owners shifted from anticipating a ``disruption'' 
argument to misinterpreting Johnson, asserting, without citation: 
``With respect to the TCC prong, the remand directs only that services 
be given opportunity to offer evidence of disruption from rates that 
have now been in effect for three years without any disruption.'' 
Copyright Owners' Reply in Support of Motion to Adopt Interim Rates at 
7-8 (Nov. 25, 2020) (emphasis added).
    On December 10, 2020, the Services submitted to the Judges their 
Proposal for Remand Proceedings, in which they made no argument that 
the uncapped TCC rates and rate structure (or, for that matter, any 
aspect of the Determination) would cause disruption or have a 
disruptive impact, whether under statutory Factor D or otherwise. By 
contrast, in their remand proposal, Copyright Owners reference twelve 
times that, for the Judges to reject the uncapped TCC rates and 
structure, the Services must show the presence of ``disruption'' 
arising from the Majority's uncapped TCC rates and structure. Copyright 
Owners made this argument notwithstanding that the ``reasonable'' rate 
standard is separate from the ``disruptive impact'' issue, which is an 
itemized objective (one of four) to be considered as an adjustment to 
what would otherwise constitute a ``reasonable'' rate. See Proposal of 
Copyright Owners for the Conduct and Schedule of the Resolution of the 
Remand at 2, 7-8, 22-24 (Dec. 10, 2020).\89\
---------------------------------------------------------------------------

    \89\ When Copyright Owners do address an argument that the 
Services actually made (on appeal) regarding the uncapped TCC rates 
and structure, they note not that the Services had made a 
``disruption'' argument, but rather that ``the Services appealed for 
the reversal of the TCC prong as substantively unreasonable.'' Id. 
at 22 (emphasis added). But Copyright Owners then assert, coyly, 
that ``this request was not granted by the Circuit'' (citing 
Johnson, 969 F.3d at 383), when in actuality, the D.C. Circuit did 
not rule against the Services on this point, but rather stated only 
that it was not addressing substantive arguments made by the 
Services ``[b]ecause we have vacated the rate structure devised by 
the [Judges] for lack of notice. . . .'' Id.
---------------------------------------------------------------------------

    In the CO Initial Submission, Copyright Owners assert, without 
citation to any of the Services' filings: ``The Services contend that, 
had they been given such an opportunity [at the hearing], they 
supposedly could have established that an ``uncapped'' TCC is 
disruptive because the market for sound recordings is not effectively 
competitive.'' Id. at 5. Copyright Owners further aver that the 
Services must ``provide evidence, consistent with the [CRB Judges'] 
well-established disruption standard, that because of the labels' 
supposed market power, the TCC structure adopted by the Board has 
actually, substantially, immediately and irreversibly threatened the 
continued viability of the interactive streaming industry'' in a manner 
that will ``threaten the viability of the music delivery service 
currently offered to consumes under [the] license.'' Id. at 7, 56 
(citations omitted).
    Copyright Owners then assert that the Services bear the burden of 
proving disruption under Factor D from the

[[Page 54430]]

uncapped rates and rate structure embodied within the rate proposal 
(even though only Copyright Owners are pursuing this approach on 
remand). Further, Copyright Owners assert that the Services' objection 
to the uncapped rates and rate structure must fail unless they can show 
that such a disruptive impact occurred during the 33-month period (from 
January 2018 through September 2020) when the Phonorecords III rates 
were in effect. Id. at 56.
    In their initial substantive remand briefing, the Services once 
more did not assert that the Determination's uncapped TCC rates and 
structure would cause disruption pursuant to Factor D of section 
801(b)(1), or even assert a non-statutory disruption arising therefrom. 
Rather, the Services directly attacked this rate approach as 
inconsistent with the statutory ``reasonable'' rate requirement, 
maintaining that ``[t]ying the mechanical rates directly to the 
complementary oligopoly sound recording rates in the manner of the 
Majority's uncapped TCC rates and rate structure is plainly 
unreasonable.'' Services' Joint Opening Brief at 46 (Apr. 1, 2021) 
(emphasis added). The Services also asserted that the uncapped TCC 
rates and rate structure are ``unreasonable'' because they do not 
promote the statutory objectives of Factor B (``fair income'' to the 
copyright user) and Factor C (reflecting the copyright users' itemized 
role in making the musical works ``available to the public.''). Id. at 
45, 50-51, 55.\90\
---------------------------------------------------------------------------

    \90\ The Services' only references to the concept of 
``disruption'' relate to their argument that their own benchmark 
premised on the prior Phonorecords II rate structure and rates would 
not be disruptive. Id. at 4, 24, 29-30. That argument is properly 
made by Services in this context, because a party seeking to 
persuade the Judges to adopt its proposal bears the burden of proof, 
pursuant to section 556(d) of the APA, regarding the consonance of 
its proposal with all the standards contained in section 801(b)(1). 
The Judges do note that one of the Services' expert witnesses, 
Professor Katz, found the Majority's attempt to avoid disruption by 
phasing-in the new rate provisions insufficient ``to mitigate the 
risk of short-term market disruption''. That testimony does not 
constitute a direct reliance by the Services on the statutory 
disruption objective in Factor D, but rather emphasizes the 
Majority's own concern with such disruption and the witness's 
concern that the phase-in did not prevent the disruptive effect that 
the Majority itself had contemplated. In any event, Professor Katz, 
as an economist, cannot make a legal argument regarding the 
applicability of the Factor D objective, the Services did not rely 
on his testimony in that regard and, as noted, the Services made no 
legal Factor D ``disruption'' argument on remand. Thus, the Judges 
do not give any weight to Professor Katz's testimony in this regard.
---------------------------------------------------------------------------

    In the Services' Reply, the Services attack Copyright Owners' 
``singular focus on the disruptive impact of the uncapped TCC prong.'' 
Services' Reply at 35. In particular, the Services argue:
    1. they have maintained and demonstrated that Copyright Owners' 
uncapped rates and rate structure are ``unreasonable,'' separate and 
apart from demonstrating that this uncapped approach also fails to 
satisfy the four itemized statutory factors;
    2. the burden of proof with regard to Factor D disruption lies with 
Copyright Owners, because they are the ones who are advocating for the 
uncapped TCC rates and rate structure;
    3. the presence of Factor D disruption, vel non, is not 
dispositive, because section 801(b)(1) and Johnson require the Judges 
to apply the entirety of the statutory standard (which consists of the 
``reasonable rate'' requirement and consideration of all four itemized 
Factors; and
    4. the ``full extent of the disruption to the Services from an 
uncapped TCC prong was never tested in the marketplace [because] [t]he 
Majority set escalating rates, and the [ ] Determination was vacated 
before the significant hike in rate levels was fully implemented.''

Id. at 35-36.

    In their Remand Reply, with regard to the issue of ``disruption,'' 
Copyright Owners assert:
    1. The Services have ``completely abandoned'' their appellate 
argument asserting disruption, and admit to having no evidence that the 
Board's adopted rate structure has any materially disruptive impact. 
Copyright Owners' Reply Brief on Remand at 5 (July 2, 2021).
    2. The Services have not even attempted to show any Factor D 
related effect or other disruption from the adopted rates and 
structure. Id. at 15, n.9.
    3. The failure of the Services to provide evidence of disruption or 
to pursue the argument that disruption had occurred was inconsistent 
with their prior assertions that the uncapped TCC rates and rate 
structure created ``a real risk of economic harm'' and the ``impact'' 
or ``harm' that the uncapped approach generated. Id. at 35.
    4. Each of the Services, in response to Copyright Owners' discovery 
requests, acknowledges that it was not offering new evidence regarding 
the ``impact'' of the Phonorecords III rates and rate structure. Id. at 
36-38.
    5. The Services did not merely suffer no disruption, they 
experienced unprecedented growth and profit under the uncapped TCC rate 
prong. Id. at 45.\91\
---------------------------------------------------------------------------

    \91\ The Judges allowed the Services to make a supplemental 
filing in response to Copyright Owners' remand reply, because those 
papers contained direct as well as reply materials. In their 
supplemental filing, the Services argued that they had not 
``thrived,'' that the financial data on which Copyright Owners' 
relied did not isolate revenue attributable to interactive services, 
was not limited to U.S. generated revenue, and used changes in the 
market capitalization of Amazon and Alphabet (Google's parent 
corporation) as a proxy for the economic fortunes of their 
interactive services. Services' Joint Supplemental Brief at 13-15. 
As explained supra, the Judges find the permanency of the 
Phonorecords III rate structure during the 33-month period from 
January 2018 through September 2020 to have been in question, 
pending the appeal that resulted in the vacating and remanding of 
the Determination and the reversion back to the Phonorecords II 
rates and rate structure. Given that uncertainty, the Judges find it 
wholly inappropriate to draw any conclusions from the change or 
stasis in the sound recording rates or the total royalty payments by 
a Service over that period.
---------------------------------------------------------------------------

    6. The Services on remand have attempted to replace their prior 
``disruption'' assertion with a claim of ``unreasonableness.'' Id. at 
50, n.36.
(iv) Analysis and Decision Regarding ``Disruption'' Issue
    The full Factor D ``disruption'' standard, as set forth by the 
Judges, states that an adjustment is warranted by Factor D if the rate 
analysis made by the Judges would otherwise:

directly produce[ ] an adverse impact that is substantial, immediate 
and in the short-run because there is insufficient time for either 
[party] to adequately adapt to the changed circumstance produced by 
the rate change and, as a consequence, such adverse impacts threaten 
the viability of the music delivery service currently offered to 
consumers under this license.

Determination at 87. Factor D is not applicable, particularly as 
proposed by Copyright Owners. Thus, the Judges reject Copyright Owners' 
assertion that the uncapped TCC prong should be adopted because of the 
absence of evidence of ``disruptive impact'' proffered by the Services. 
This rejection is based on several findings of fact and conclusions of 
law.
    First, the issue of ``disruptive impact'' pertains here to the 
proposal advanced by Copyright Owners, not the Services. Thus, the 
burden of proving that this uncapped TCC rate prong proposal satisfies 
the elements, including Factor D, of the section 801(b)(1) standard in 
a sufficient manner lies with Copyright Owners, not the Services. See 5 
U.S.C. 556(d). Accordingly, the fact that the Services did not 
affirmatively assert an argument of ``disruptive impact'' is of no 
consequence. Moreover, as the review of the Services' filing makes 
clear, the Services never abandoned that argument, because they never 
made it.

[[Page 54431]]

Rather, they have consistently argued that the uncapped TCC rate prong 
was unreasonable, not that it was statutorily ``disruptive'' as that 
standard has been applied by the Judges.
    Second, Copyright Owners did not demonstrate with sufficient 
evidence or testimony that the uncapped TCC rate would be consistent 
with Factor D. To be clear, by this the Judges do not mean that 
Copyright Owners were obliged to prove a negative. Rather, they needed 
to prove, and indeed attempted to do so, that it was unlikely that 
their rates would cause a ``disruptive impact.''
    In this regard, as an empirical matter, Copyright Owners proffered 
the testimony of an economic expert witness, Dr. Eisenach, who opined 
that the Services' [REDACTED]. Eisenach WRT ]] 12-41 ([REDACTED]) CO 
Reply at 40-41. However, as the Judges discuss supra, that period 
reflected ``33 months of uncertainty,'' during which no one could 
predict the final mechanical rate and structure that would be adopted 
by the Judges and/or the D.C. Circuit after appeals. Accordingly, that 
factual evidence is unpersuasive.
    Further, as a theoretical matter, Copyright Owners rely on 
Professor Watt's testimony regarding the ``seesaw'' effect. In that 
regard, and as discussed supra, the Majority took comfort in what it 
understood to be Professor Watt's ``prediction'' that increases in 
mechanical royalties would be offset almost dollar-for-dollar by 
reductions in the sound recording royalty. However, as also discussed 
supra, Professor Watt has now clarified on remand that he never made 
such a ``prediction,'' and that his testimony regarding the so-called 
``seesaw'' was limited to shifts in the share of the surplus to 
Copyright Owners and from sound recording companies as a consequence of 
an increase in the mechanical rate, holding all other factors unchanged 
(the ceteris paribus assumption).
    Moreover, Professor Watt further explained that many other factors 
would likely impact the sound recording rate together with an increase 
in the mechanical rate, including ``a measure of oligopolistic 
interaction, different timings of different rate bargains, and the 
actual values of outside options.'' Watt RWRT ] 118. Professor Watt 
candidly acknowledged that he has not modeled these independent 
variables, and he further notes that the data may not exist to allow 
for such modeling. Id. But the inability to model the impact of 
independent variables does not mean that their potential to cause 
disruption can be ignored.
    In particular, the purpose of the ``seesaw'' contention was that it 
prevented economic harm to the Services in connection with a rise in 
the mechanical rate. Although not of Professor Watt's design, that 
connection is intentionally built into the Majority's uncapped TCC 
rate. See Determination at 35 (``Incorporating an uncapped TCC metric 
into the rate structure permits the Judges to influence that ratio 
directly.'') But the ``measure of oligopolistic interaction'' 
referenced by Professor Watt was the very concern expressed by the 
Dissent, which cautioned that there was no evidence that the sound 
recording companies would be compelled to maintain the same industry 
structure and accept the loss of substantial royalty income. See 
Dissent at 4 (``[T]he record companies may decide to keep their rates 
high despite the increase in mechanical rates, or decide it is in their 
interest to avoid a reduction in royalty revenue by creating a 
completely different paradigm for streaming, by which the record 
companies move the streaming service in-house and effectively destroy 
the existing services.'').\92\
---------------------------------------------------------------------------

    \92\ The Dissent noted that this risk was speculative in nature 
because there was no evidence proffered at the hearing regarding the 
reactions of the sound recording companies. But no such evidence was 
forthcoming in the remand proceeding either, and, as noted supra, 
the burden of proof in this regard falls on Copyright owners as the 
proponents of the uncapped TCC rate prong. In fact, because the 
major publishers who are members of the NMPA (a constituent of 
Copyright Owners) are part of the same corporate structure as the 
sound recording Majors, the burden of producing evidence would fall 
on Copyright Owners as well regarding the sound recording companies' 
reaction to the ``seesaw'' effect.
---------------------------------------------------------------------------

    Also, the ``different timings of different rate bargains,'' another 
independent variable identified in Professor Watt's remand testimony, 
was an issue raised to him at the hearing by Judge Strickler. Professor 
Watt candidly agreed that the Judge was ``absolutely correct'' that 
there is a ``risk, then, of disrupting the market by having a total 
royalty that's greater than what is indicated by your Shapley 
testimony, simply because of the disparity of times in which the rates 
are . . . implemented.'' 3/27/17 Tr. 3091-92 (Watt) (emphasis added). 
However, this admitted risk of disruption was not addressed by 
sufficient record evidence.\93\
---------------------------------------------------------------------------

    \93\ As noted supra, Copyright Owners did not call any sound 
recording industry witnesses, or provide evidence from sound 
recording companies, indicating that labels would even be amenable 
to considering such renegotiated rate reductions. Instead, at the 
hearing, Professor Watt merely speculated that the sound recording 
companies might renegotiate their rates downward to reflect the 
seesaw effect when mechanical rates increased. Tr.3/27/17 3093-94 
(Watt) (``I'm not able to comment on how, you know, how possible it 
is to take an agreement that's in force and then change it.''). Not 
only was that mere speculation, it was provided by an economist who 
is neither a music industry executive nor an attorney, and the 
witness did not testify that he had spoken to anyone who would have 
industry knowledge regarding whether a label would even be amenable 
to considering such rate reductions.
---------------------------------------------------------------------------

    Third, disruption in the narrow sense of Factor D as applied by the 
Judges previously is not relevant to the present problem. An increase 
in total royalties is not a short-run immediate issue, but rather an 
ever-present possibility that the seesaw analysis does not sufficiently 
address. Rather, the uncapped nature of the TCC rate prong renders it 
unreasonable rather than narrowly disruptive.
    Balancing the foregoing considerations, the Judges find that 
Copyright Owners' disruption-based argument lacks merit.
6. Conclusion Regarding Uncapped TCC Rate Prong
    For the foregoing reasons, the Judges decline to adopt the uncapped 
TCC rate tier proposed on remand by Copyright Owners.

III. Rejection of Phonorecords II Settlement as a Benchmark

A. D.C. Circuit Ruling

    Each of the Streaming Services advanced somewhat different rate 
plans, but all four proffered a benchmark that ``broadly sought to 
maintain the Phonorecords II rate structure,'' while lowering or 
eliminating the mechanical floor.\94\ Johnson, 969 F.3d at 371. With 
regard to the Services' proposed benchmark based on the Phonorecords II 
rates, rate structure, and terms (hereinafter, PR II-based 
benchmark),\95\ the Judges are guided by several rulings in Johnson.
---------------------------------------------------------------------------

    \94\ The ``mechanical floor'' refers to an alternative rate 
calculation. ``If the All-In Rate calculation results in a dollar 
royalty payment below the stated Mechanical Floor rate, then that 
floor rate would bind.'' Determination at 26 n.59.
    \95\ See Services' Joint Rate Proposal (in Services' Joint 
Written Direct Remand Submission at Tab C) (Apr. 1, 2021). According 
to the Services, their rate proposal in this proceeding is meant to 
``update the Phonorecords II terms to include terms of the 
Determination, as amended during the implementation of the Music 
Modernization Act, that were upheld in Johnson . . . including terms 
relating to student and family plan products, or that were not 
challenged by either the Copyright Owners or the Services.'' Id. at 
2. The Services include in their Joint Rate Proposal a chart 
summarizing the proposed rates for their offerings. That chart is 
attached as an Addendum to this Initial Ruling.
---------------------------------------------------------------------------

    In particular, the D.C. Circuit found the Judges' treatment of the 
PR II-based benchmark to be ``muddled.'' Johnson, 969 F.3d at 387. The 
D.C. Circuit emphasized that the Judges ``failed to

[[Page 54432]]

explain'' their rejection of the PR II-based benchmark. Id. at 367. See 
also id. at 376 (Judges ``failed to ``reasonably explain'' rejection).
    In the appeal, Copyright Owners attempted to defend the Judges' 
reliance on the absence of evidence of the settling parties' subjective 
intent in reaching the Phonorecords II terms. Id. at 387. The D.C. 
Circuit dismissed Copyright Owners' post hoc attempt, noting that 
``nowhere does the [ ] Determination explain why evidence of the 
parties' subjective intent in negotiating the Phonorecords II 
settlement is a prerequisite to its adoption as a benchmark.'' Id. at 
387 (emphasis added).
    The D.C. Circuit also criticized the attempt by the Judges' 
appellate counsel to ``change tack'' and argue that their rejection of 
the PR II-based benchmark was reasonable because: (1) evidence showed 
that the prior rates had been set far ``too low'' and (2) it was 
``outdated''. The D.C. Circuit found that those arguments also were 
``nowhere to be found in the [ ] Determination's discussion'' of the 
appropriateness of the Phonorecords II settlement as a potential 
benchmark. Id. at 387 (emphasis added).\96\ In the end, the D.C. 
Circuit agreed with the Streaming Services that, inter alia, the Judges 
failed to reasonably explain their rejection of the benchmark and, for 
all of the reasons cited, vacated and remanded the adopted rate 
structure and percentages for further proceedings. Id. at 381.
---------------------------------------------------------------------------

    \96\ In the present remand ruling, the Judges do not rely on 
their appellate counsel's ad hoc arguments that the D.C. Circuit 
found to be absent from the Determination. The Judges note though 
(as discussed in more detail infra) that in this Initial Ruling they 
are increasing the 10.5% royalty rate in the Phonorecords II rates 
by 44% to 15.1% (as phased-in by the Determination), thus addressing 
appellate counsel's ad hoc assertion that the Phonorecords II rates 
were ``too low.'' Similarly, as discussed infra, the Judges address 
the notion that the PR II-based benchmark is outdated.
---------------------------------------------------------------------------

B. Remand Procedure Regarding the PR II-Based Benchmark

    On December 15, 2020, subsequent to the D.C. Circuit's decision, 
the Judges entered an Order Regarding Proceedings on Remand, in which 
the Judges stated:

    The Judges accept the parties' proposals to resolve the issues 
concerning the use of the Phonorecords II settlement as a benchmark. 
. . .
. . .
    The Services and Copyright Owners also agree that the Judges 
should resolve this issue based on the existing record, after 
receiving two rounds of additional briefing from the parties.

Remand Order at 1-2.
    Based on the ruling in Johnson the Judges reject Copyright Owners' 
position that they need not engage in a full analysis of the issue. The 
Judges conclude that they must engage in, and fully articulate, a 
reasoned analysis that adequately addresses ``the issues concerning the 
use of the Phonorecords II settlement as a benchmark.'' Id. (emphasis 
added). If the Judges determine that the Majority properly rejected the 
Services' proposed use of the PR II-based benchmark, the rejected 
portions will play no part in the Judges' remand ruling. On the other 
hand, if the Judges find, after engaging in that analysis, that the PR 
II-based benchmark was not properly rejected then, as a matter of law 
and logic, the Judges must weigh the Services' PR II-based benchmark 
for application, in whole or in part.
    The Judges reject Copyright Owners' reading of Johnson as holding 
that the Judges cannot fully consider the PR II-based benchmark on 
remand. Copyright Owners argue that the D.C. Circuit ``did not suggest 
the [Judges] substantively erred'' in rejecting that benchmark, or that 
they ``needed to reconsider [their] decision,'' but had ``merely 
remanded for a `reasoned analysis' . . . as to why it did so.'' CO 
Initial Submission at 10; see also Copyright Owners' Reply Remand Brief 
at 7-8. Because Johnson ruled that the Majority's reasoning was 
muddled, indiscernible, unexplained and lacking in reason, the D.C. 
Circuit obviously neither accepted nor rejected the Majority's 
disregard for the PR II-based benchmark--thus requiring the CRB Judges 
to take a comprehensive look at that benchmark. In this regard, the 
Judges agree with the Services that, pursuant to apposite case law, if 
the outcome of the remand as to this issue was preordained pending the 
further ``reasoned analysis,'' the D.C. Circuit would have expressed a 
desire simply to remand without vacating as to this issue. Services' 
Joint Remand Reply Brief at 7-8 (citing Allied-Signal, Inc. v. NRC, 988 
F.2d 146, 150-51 (D.C. Cir. 1993) (``The decision whether to vacate 
depends on the seriousness of the order's deficiencies (and thus the 
extent of doubt whether the agency chose correctly) and the disruptive 
consequences of an interim change that may itself be changed.'')).\97\
---------------------------------------------------------------------------

    \97\ However, the Judges note that section 803(d)(3) may require 
the D.C. Circuit to remand rather than reverse when the issue 
concerns more than rates alone. Thus, the statute appears to require 
a remand in order for the Judges to apply their statutory authority 
and expertise in toto.
---------------------------------------------------------------------------

    Because Johnson held that the Majority's reasoning was muddled, 
indiscernible, unexplained, and lacking in reason, the D.C. Circuit 
obviously neither accepted nor rejected the Majority's disregard for 
the PR II-based benchmark. Thus, the Judges take a comprehensive look 
at that benchmark's rates and rate structure to evaluate its usefulness 
in this proceeding.
    Relatedly, the Judges also reject Copyright Owners' assertion that 
the Judges can only consider on remand the Phonorecords II rates, and 
cannot consider on remand the relative strengths and weaknesses of the 
structure in which those rates are embedded. See Copyright Owners' 
Reply Brief on Remand at 14. This distinction is impractical and 
unworkable. If the (non-``headline'' rates \98\) themselves can be 
reviewed and found acceptable (as they are infra) into what structure 
would they be placed? There are multiple provisions in the Phonorecords 
II rate structure providing for different rates, designed to balance 
(1) the ability of services to attract consumers with a low 
Willingness-to-Pay and/or a low Ability-to-Pay (the price 
discriminatory and differentiated features \99\) with (2) the revenue 
diminution protections for which Copyright Owners had successfully 
negotiated. Moreover, the D.C. Circuit has vacated the Determination, 
and in doing so did not make any rulings critical of the rate structure 
in the Phonorecords II-based benchmark that would suggest the cramped 
review advocated by Copyright Owners. Indeed, the D.C. Circuit 
explicitly stated, without distinguishing between rates and structure, 
that it ``agree[s] with the Streaming Services that the [Judges] . . . 
failed to reasonably explain [their] rejection of the Phonorecords II 
settlement as a benchmark . . .'' See Johnson, 969 F.3d at 376; see 
also id. at 389 (issues relating to ``rates'' and ``rate structure'' 
are ``intertwined'').
---------------------------------------------------------------------------

    \98\ As explained elsewhere in this Initial Ruling, the Judges 
are increasing the ``headline'' rate from 10.5% to 15.1%.
    \99\ Specifically, the PR II-based benchmark would incorporate 
the price discriminatory features for product differentiation as 
between: (1) subscription vs. ad-supported services; (2) portable 
and non-portable services; and (3) unbundled vs. bundled services. 
See Determination at 10; Dissent at 26. The third category--bundled 
vs. unbundled--is discussed infra in the context of the Bundled 
Revenue definition.
---------------------------------------------------------------------------

    Further, the Judges emphasize that the rate structure of the PR II-
based benchmark provides protection sought by Copyright Owners against 
revenue diminution by the Services--protection they would otherwise 
lose--because in this Initial Ruling the Judges are not adopting the 
vacated uncapped TCC prong for which Copyright Owners are now 
advocating, and which they claim

[[Page 54433]]

would have protected them in that regard. Cf. CO Additional Submission 
at 4-6 (acknowledging PR II-based benchmark provided some TCC 
provisions, allowing for protection against revenue diminution). Thus, 
the Judges' remand rulings on the PR II-based benchmark rates and on 
the uncapped TCC rate prong are inextricably interlaced. See Johnson, 
969 F.3d at 381 (absence of ``reasoned explanation'' for rejecting PR 
II-based benchmark was problematic because it occurred ``when'' Judges 
adopted an alternative proposal that called for ``setting . . . total 
content cost and revenue rates.'') (emphasis added).
    The Judges weigh each benchmark's intrinsic strengths and 
weaknesses, as well as its comparative advantages and disadvantages 
vis-[agrave]-vis other proffered benchmarks. On remand, the 
interrelationships of the competing benchmarks are of particular 
importance, given Copyright Owners' need for the aforementioned 
protections against revenue diminution via price discrimination.\100\
---------------------------------------------------------------------------

    \100\ The Judges categorically reject Copyright Owners' 
assertion that the PR II-based benchmark cannot be considered 
because the parties agreed in the Phonorecords II settlement that 
any future statutory mechanical rate determination would made ``de 
novo'' vis-[agrave]-vis that settlement determination. In fact, the 
industrywide representatives (NMPA and Digital Media Association 
(DiMA)) who entered into the settlement conspicuously did not agree 
that the existing rate structure or rates could not be considered as 
the bases for future rate determinations. By contrast, the 
Phonorecords I settlement agreement expressly stated ``[s]uch 
royalty rates shall not be cited, relied upon, or proffered as 
evidence or otherwise used in the [Phonorecords II] Proceeding.'' 
Trial Ex.6013, Phonorecords I Agreement at sec. 3. Compare Trial Ex. 
6014, Phonorecords II Agreement at sec. 5.5 (omitting clause 
precluding reliance on evidentiary value of Phonorecords II royalty 
rates and including full-integration clause). This change 
objectively demonstrates that the parties to the 2012 settlement 
understood the evidentiary value of the Phonorecords II settlement 
in the next section 115 proceeding, i.e., this proceeding. See 
Dissent at 15-16.
    On the other hand, the Judges reject the Services' argument that 
the Phonorecords II rates and structure should be retained merely 
because the Services relied on their continuation to make 
investments in their business models. As Copyright Owners note, the 
applicable regulations provide that ``[i]n any future proceedings 
the royalty rates payable for a compulsory license shall be 
established de novo.'' 37 CFR 385.17; see also 37 CFR 385.26. A 
party may feel confident that past is prologue and that the parties 
will agree to roll-over the extant rates for another period; a party 
could be sanguine as to its ability to make persuasive arguments as 
to why the rates should remain unchanged; a party might even 
conclude that the mechanical rate is such a small proportion of the 
total royalty obligation that its increase would be unlikely to 
alter long-term business plans. But for sophisticated commercial 
entities to claim that they simply assumed the rates would roll-
over--without the reasonable possibility of significant adjustment 
or outright abandonment--strikes the Judges as so irrational and 
reckless as to raise serious doubts about the credibility of that 
position. (If the Services had made a persuasive argument that 
certain fixed cost investments were ``sunk'' and had useful lives 
that substantially exceeded the five-year rate term, then such costs 
could be considered under Factor C of section 801(b)(1), but they 
did not make a persuasive argument in this regard. Cf. SDARS II, 78 
FR 23054, 23069 (Apr. 17, 2013) (adjusting rates downward under 
Factor C, and distinguishing internet music transmissions, to 
reflect that--because Sirius XM needed to make ``unique and 
substantial'' investments in the form of ``sunk'' costs paid for 
satellites with a useful life of l2-15 years--``it is not 
unreasonable for Sirius XM to expect to recoup a certain amount of 
those costs over the expected useful life of the [s]atellites,'' 
which exceeded the five-year rate term.)
---------------------------------------------------------------------------

    Through this approach, the Judges ultimately may adopt only one of 
the parties' benchmarks or other methodologies, or they may modify the 
proposals by combining them, provided such a modification is ``within a 
reasonable range of contemplated outcomes . . . piecing together a rate 
structure, the economic and policy consequences of which had already 
been explored and developed by the parties in the record.'' Johnson, 
369 F.3d at 382.
    In their consideration of the PR II-based benchmark, the Judges are 
not suggesting that this benchmark is the optimal tool to use in order 
to identify rates and terms among all approaches that might have been 
proffered (but were not). But the Judges are cabined by the evidence 
they receive. See 17 U.S.C. 803(a)(1) (``the Judges shall act . . . on 
the basis of a written record . . . .''); see also P. Wald, supra, 
(noting that parties' economic proposals made in an action ``impose[ ] 
a practical constraint'' on judge who will, ``for the most part, be 
limited by what the parties serve up to her.''). Based upon the 
available record evidence, the Judges find that the Services' PR II-
based benchmark--although not necessarily perfect--is more than 
sufficient to satisfy the legal requisites for application, as well as 
a practical benchmark, when used in conjunction with the 15.1% headline 
revenue rate advocated by Copyright Owners. See generally Nat'l Cable 
Television Ass'n v. Copyright Royalty Tribunal, 724 F.2d 176, 182 (D.C. 
Cir. 1983) (rate-setting is an intensely practical affair).

C. Parties' Remand Arguments Regarding PR II-Based Benchmark 
101
---------------------------------------------------------------------------

    \101\ The parties made arguments both in the original hearing 
and in this remand proceeding regarding the Services' proffer of the 
PR II-based benchmark. Each party's pre-remand and post-remand 
arguments overlap to some extent. Examination of the pre-remand 
arguments is also necessary because of the findings in Johnson and 
because the parties agreed that the evidentiary record on this 
remanded issue would not be enlarged.
---------------------------------------------------------------------------

1. Services' Arguments
    The Services maintain that their PR II-based benchmark satisfies 
the ``reasonable'' rate requirement and is consistent with the four 
itemized factors set forth in section 801(b)(1). They make several 
arguments in favor of this position.
    First, they aver that their PR II-based benchmark possesses all the 
characteristics of an ``ideal'' benchmark. Services' Joint Opening 
Brief at 19. In this regard, they argue that their proffered benchmark 
``involves the same sellers, the same or similar buyers, and the same 
rights as at issue in this proceeding,'' and that there has been ``no 
material change in the economic circumstances of the marketplace that 
would warrant adjusting the rate levels or rate structure in the 
benchmark.'' Id. at 20.
    Applying the facts to these benchmark characteristics, the Services 
assert that the first three elements--same sellers (here, licensors), 
same buyers (here, licensees) and same rights (the mechanical license 
for interactive streaming) are satisfied. In particular, they note that 
the majority of the participants in the present proceeding either 
directly participated in the Phonorecords II settlement process or were 
active in the market contemporaneous with that settlement. Id. at 20-
21.
    Turning to the next benchmark characteristic--the absence of a 
``material change in the economic circumstances of the marketplace that 
would warrant adjusting the rate levels or rate structure in the 
benchmark''--they emphasize that the PR II-based benchmark contains 
different rate levels for different product offerings, to account for 
(a) consumers' varying willingness-to-pay (WTP) and (b) the zero 
marginal physical cost of digital reproductions of sound recordings 
containing musical works. Id. at 21-22 (citing multiple experts).
    Next, the Services point to the fact that the ``headline'' \102\ 
royalty rate is based on a percent-of-revenue, so that revenue growth 
(or decline) on this rate prong allows for royalty payments to directly 
adjust in tandem. Id. Further, the Services assert that the importance 
of streaming as ``the future of the music industry'' was known to the 
Phonorecords II negotiators, as evidence by the then-recent launch in 
the United

[[Page 54434]]

States of the popular Spotify service. Id. at 23.
---------------------------------------------------------------------------

    \102\ The Judges and the parties characterize the percent-of-
revenue of revenue rate as the ``headline'' rate. See Johnson, 969 
F.3d at 383 n.10.
---------------------------------------------------------------------------

    Beyond these benchmark requisites, the Services also emphasize that 
the PR II-based benchmark is the product of a settlement whose 
negotiated features burnish the value of this benchmark as reflective 
of effective competition. Specifically, they note:
     The settlement was negotiated in the same statutory 
context, concerning the identical rate standard and factors as 
applicable to the present proceeding.
     Neither side would have accepted a deal materially worse 
than what it expected from a section 115 proceeding applying the 
section 801(b)(1) considerations.
     The statutory alternative diminishes any additional 
licensor-side negotiating power arising from ``Must-Have'' 
complementary oligopoly of the licensors of the musical works 
publishers.

Id. at 22. Moving from the negotiating context to market performance 
under this standard, the Services aver that this approach has borne 
fruit for the industry as a whole. They point to the evidence of the 
licensors' consistent profitability and the licensees' ability to 
``benefit'' from the Phonorecords II approach. Id. at 23.
    The Services also maintain that the Phonorecords II structure 
``addresses any concerns with bundling and the potential for revenue 
deferment.'' Id. at 24.\103\ They assert that these issues were 
specifically addressed by Copyright Owners during the Phonorecords II 
negotiation, because ``multiproduct firms such as Yahoo and Microsoft'' 
that offered streaming services had the capacity to make bundled 
offerings to consumers. These concerns were addressed in the 
Phonorecords II rate structure, the Services note, through the use of 
``multiple rate prongs, minima and floors,'' ensuring that ``the total 
musical works royalty for certain types of offerings does not fall 
below a specified level,'' thereby ``mitigat[ing] the effect of any 
potential revenue deferrals and appropriately address[ing] any concerns 
with bundling.'' Id.\104\
---------------------------------------------------------------------------

    \103\ The issue of bundling is addressed in this Initial Ruling 
infra, in connection with the Judges' definition of Service Revenue 
generated through the offering of sound recordings as part of a 
bundle containing other goods or services.
    \104\ The Services also reiterate their pre-remand argument that 
the Phonorecords III settlement of subpart A rates for sales of 
physical and digital download phonorecords (now reorganized in 
subpart B) confirms the appropriateness of the Phonorecords II-based 
benchmark. However, any further reliance by the Services on that 
argument is moot, because the D.C. Circuit affirmed the Majority's 
analysis of the subpart A rates. Johnson, 969 F.3d at 386 (noting 
that the Majority adequately explained treatment of the subpart A 
rates as `` `at best' a floor'' below which the mechanical royalty 
rates paid by the Services for interactive streaming cold not fall).
---------------------------------------------------------------------------

    Finally, the Services maintain that ``[d]irect agreements between 
Copyright Owners and Services also support adoption of the PR II-based 
benchmark.'' Id. at 34. In particular, they note that many of the 
royalty rates (and terms) in these direct agreements apply the 
Phonorecords II rates. Moreover, the Services maintain, because these 
direct agreements are in the nature of blanket license of a publisher's 
entire catalog, they provide an added ``access'' value in the form of 
full-repertoire licensing. These direct agreements do not include a 
rate above Phonorecords II levels; thus, the Services contend, they 
underscore the reasonableness of the Phonorecords II rates. Id. \105\
---------------------------------------------------------------------------

    \105\ Under section 115--prior to the effective date of the 2008 
Music Modernization Act--an interactive service was required to 
serve a ``Notice of Intent'' to use the copyright license (NOI) with 
the owner of a copyright for each musical work before streaming the 
sound recording embodying that musical work. By contrast, a direct 
license with a publisher covers more than an individual musical work 
by providing ``access'' value to an entire catalog, without the 
transaction cost burden of filing multiple individual NOIs.
---------------------------------------------------------------------------

    Finally, the Services aver that the PR II-based benchmark satisfies 
the itemized four section 801(b)(1) factors. With regard to Factor A, 
they maintain that: (1) the Phonorecords II framework has corresponded 
with an increase in the supply of musical works; (2) the PR II-based 
benchmark will increase the likelihood that the Services will increase 
subscriber counts, generating profitability, which will make streaming 
available to more listeners; and (3) the price discriminatory aspects 
of this royalty rate structure allows the Services to afford to offer 
streamed music to listeners with a low willingness (or ability)-to-pay, 
at lower rates or through ad-supported services. Services' Joint 
Opening Brief at 25-27.
    Regarding Factors B and C (the ``fair return'' and ``relative 
contributions'' objectives), the Services emphasize that the PR II-
based benchmark satisfies these statutory elements because it: (1) was 
the result of negotiations between industrywide representatives who had 
every incentive to obtain a ``fair'' return and to receive recompense 
for their ``contributions'' to streaming; and (2) allowed interactive 
streaming to become ``a significant means for consumers to listen to 
music'' while simultaneously generating growth in annual royalties for 
Copyright Owners.'' Id. at 27-29.
    Lastly on the subject of the statutory factors--regarding Factor D 
(minimizing disruptive impact)--the Services make a succinct argument: 
``By renewing the rate levels and structure of Phonorecords II, there 
is minimal risk of disruption.'' Id. at 29-30.
    The Services also address several further criticisms of the PR II-
based benchmark contained in the Determination. Focusing first on an 
issue specifically addressed in Johnson, they assert the irrelevancy of 
the ``subjective intent'' of the parties that negotiated the 
Phonorecords II settlement--a factor on which the Majority relied in 
deciding not to adopt the PR II-based benchmark. In this regard, the 
Services are also responding to the D.C. Circuit's concern regarding 
this issue. See Johnson, 969 F.3d at 387 (``In rejecting that 
settlement as a possible benchmark, the [Judges] faulted the Streaming 
Services for failing to explain why the parties to the Phonorecords II 
settlement agreed to the rates in that settlement . . . [b]ut nowhere 
does the [ ] Determination explain why evidence of the parties' 
subjective intent in negotiating the Phonorecords II settlement is a 
prerequisite to its adoption as a benchmark.'').
    The Services note that no benchmark evidence presented by any party 
is proffered with supporting evidence of the subjective intent of the 
bargainers who negotiated the benchmark. Moreover, they note that the 
Majority in fact acknowledged that ``[r]elying on a benchmark as 
objectively useful without [the need for] further inspection'' is 
``typical and appropriate for the benchmarking method.'' Id. at 35 
(quoting Determination at [55] & n.106 (emphasis added)).
    With regard to other criticisms of the Majority's failure to use 
the PR II-based benchmark, the Services argue that the Majority 
misapplied their previous rulings that they ``cannot and will not set 
rates to protect any particular streaming service business model.'' Id. 
at 37 (quoting Phonorecords III, 84 FR 1945). The Services find this 
principle inapposite, because their point is that the multiple price-
discriminatory aspects of the Phonorecords II approach made it ``a 
valuable benchmark . . . because it had allowed for different service 
types to emerge and grow, which benefits the entire market.'' Id. at 
37. The Services also take issue with the Majority's assertion that the 
Phonorecords II rate structure was too complex, deriding it as a 
``Rube-Goldberg-esque'' contraption. Id. at 38. Rather, the Services 
maintain that the structure was as complex as necessary to effectuate 
the parties' needs, particularly the price discriminatory features and 
the protections against

[[Page 54435]]

revenue diminution. Id. at 38-39. Further, the Services note that the 
record is devoid of any testimony or evidence indicating any actual 
confusion caused by the Phonorecords II rate structure. Id. at 39. 
Finally in this regard, the Services maintain that the rate structure 
adopted by the Majority is essentially as complex as the structure in 
Phonorecords II, with the only major change being the replacement of 
the capped TCC rates with uncapped TCC rates.\106\ Id.
---------------------------------------------------------------------------

    \106\ As discussed infra, the relative complexity or simplicity 
of the rate structure is not a statutory factor, nor is it a 
decisive element of a reasonable rate structure, when the details of 
that structure effectuate price discriminatory configurations that 
would increase the availability of music and streaming revenues and 
otherwise satisfy the statutory criteria.
---------------------------------------------------------------------------

    The Services address another criticism--that the rates in the PR 
II-based benchmark are too low. This issue is largely moot, as the D.C. 
Circuit's affirmance of the Majority's expert ``line-drawing'' and 
``reasoned weighing of the evidence'' confirmed that a rate increase 
was necessary. In this Initial Ruling, the Judges have acknowledged 
specifically the appropriateness of the 15.1% revenue rate--a 44% 
increase over the 10.5% headline rate in the PR II-based 
benchmark.\107\
---------------------------------------------------------------------------

    \107\ The Judges characterize this issue as largely moot because 
the PR II-based benchmark includes on its ``lesser of'' prongs price 
discriminatory rates, discussed infra. But those ``lesser of'' rates 
are overridden by the ``greater'' 15.1% rate. As also discussed 
infra, Mechanical Floors continue to bind at lower mechanical 
royalty levels (without reducing the songwriters' ``All-In'' musical 
works royalty that includes the performance royalties), because 
these floors were retained in the Determination and were not the 
subject of appeal.
---------------------------------------------------------------------------

2. Copyright Owners' Arguments
    Copyright Owners assert that the record evidence overwhelmingly 
supports the Judges' rejection of the PR II-based benchmark. At the 
outset, they maintain that the Judges found--and the D.C. Circuit 
affirmed--that a rate increase was required in the Phonorecords III 
terms. CO Initial Submission at 13. (As noted, an increase in the 
headline rate by 44%, to 15.1%, is adopted in this Initial Ruling.)
    Next, Copyright Owners maintain that the evidence established that 
``market conditions'' were ``radically different'' at the time of the 
Phonorecords III proceeding compared with when the parties entered into 
their 2012 industrywide agreement in Phonorecords II. Id. at 17. In 
particular, Copyright Owners point to testimony describing the 
streaming industry as ``nascent'' in 2012, with fewer streams, 
subscribers, services, and choices of music; operating in a consumer 
environment when download purchases and Pandora's noninteractive 
service were the predominant means for consumers to listen digitally to 
music. Id. at 18-21. In sum, Copyright Owners maintain, that streaming 
was ``economically insignificant'' to the music industry when the PR II 
provisions were adopted. Id. at 20.
    Copyright Owners particularly emphasize the substantial increase in 
streaming revenue during the Phonorecords II period. They point out 
that while ``total streaming revenue had ranged from approximately $150 
million in 2005 to $212 million in 2010, . . . after 2012[,] annual 
[streaming] revenue exploded to reach approximately $1.6 billion by 
2015.'' Id. at 23. Further, they note there is no evidence that the 
music publishers or anyone else had predicted this substantial rise in 
streaming and the revenues it generated, and that in no way could it be 
inferred that those rates had ``baked-in'' future growth. In fact, 
Copyright Owners assert at the hearing that the PR II rates were merely 
``experimental''--consistent with the relatively nascent stage of the 
streaming industry. Id. at 25.
    Additionally, Copyright Owners maintain that the identities of the 
parties involved in the Phonorecords III proceeding are different from 
those who established the Phonorecords II framework. Although they 
acknowledge the presence of current interactive services Spotify and 
Rhapsody in this market prior to the Phonorecords II framework agreed 
to by the trade associations for the interactive services and the music 
publishers, they point out that ``[n]one of the other participants in 
this proceeding even entered the streaming business until after the 
Phonorecords II settlement.'' Id. at 21.
    Next, Copyright Owners assert that the Services' evidence is 
inadequate to support a finding that the rates in their PR II-based 
benchmark are suitable for use in setting royalty rates in this 
proceeding. First, they echo the Determination, which stated that the 
Services (1) did not examine in detail the particular rates within the 
existing rate structure; (2) relied on the 2012 rates as objectively 
useful without further inspection; and (3) did not call witnesses to 
testify regarding the 2012 settlement negotiations. Id. at 27 (citing 
Determination, 84 FR 1944 & n.106). Because of the absence of the 
foregoing evidence, Copyright Owners assert that the Services were left 
with ``no evidence explaining how the particular rates and percentages 
in those settlements were calculated or derived, how they were 
negotiated, or how they were reasonable in light of the explosive 
growth in the streaming marketplace between the time of those 
settlements and the Phonorecords III proceeding.'' Id. at 28. The 
absence of such evidence, according to Copyright Owners, meant that the 
Services had failed to carry their burden of proof under 5 U.S.C. 
556(d) with respect to their proposal, a burden Copyright Owners assert 
the Services acknowledged they bore. Id. at 29-30.
    Additionally, Copyright Owners claim that the D.C. Circuit found 
``validity'' in Copyright Owners' assertion that the subjective intent 
of the parties to the Phonorecords II settlement is relevant because it 
would have revealed whether the agreed-upon rates were based on 
economic realities or instead were driven by other considerations. Id. 
at 30-31 (citing Johnson, 969 F.3d at 387). However, Copyright Owners 
acknowledge that, because this was not a reason given by the Majority, 
it carried no weight with the D.C. Circuit on appeal. Id. at 31.
3. Analysis and Decision Regarding PR II-Based Benchmark \108\
---------------------------------------------------------------------------

    \108\ The setting of statutory royalty rates involves to a 
significant degree the application of economic analysis. 
Accordingly, the Judges find it appropriate to set forth certain key 
aspects of microeconomics that guide the application of the section 
801(b)(1) standard in the present proceeding. That guidance is set 
forth more fully in the Dissent at 29-39.
---------------------------------------------------------------------------

a. PR II-Based Benchmark Meets Most of the Requisites for a Useful 
Benchmark
    The four classic characteristics of an appropriate benchmark are:
    (1) the degree of comparability of the negotiating parties to the 
parties contending in the rate proceeding,
    (2) the comparability of the rights in question,
    (3) the similarity of the economic circumstances affecting the 
earlier negotiators and the current litigants, and
    (4) the degree to which the assertedly analogous market under 
examination reflects an adequate degree of competition to justify 
reliance on agreements that it has spawned.
    In re Pandora Media, 6 F.Supp.3d 317, 354 (S.D.N.Y. 2014, aff'd sub 
nom Pandora Media Inc. v. ASCAP, 785 F.3d 73 (2d. Cir. 2015). As 
discussed below, the PR II-based benchmark meets criteria (1), (2) and 
(4), but requires adjustment to fully satisfy criterion (3).
    First, the PR II-based benchmark obviously pertains to the same 
rights at issue in this proceeding, as it reflects the licensing 
provisions from the immediately preceding mechanical license 
proceeding.
    Second, the licensors (songwriters and music publishers) and 
licensees (interactive streaming services) are

[[Page 54436]]

comparable (albeit not identical). While Copyright Owners emphasize the 
different identities and market involvement of the licensees, 
particularly the greater market penetration of Amazon, Apple, and 
Google, the Services note that even prior to the more significant entry 
of these three entities, similar multiproduct firms, such as Yahoo and 
Microsoft, were active licensees. The Judges find that the changing 
identities of the large multiproduct technology firms does not 
demonstrate the absence of comparability between and among such firms 
in the Phonorecords II and Phonorecords III rate periods. The shifting 
market entries, exits, strategies, successes and setbacks of otherwise 
comparable firms are expected occurrences in a dynamic capitalist 
market system and are not factors that materially diminish the 
necessary comparability of the parties for benchmarking purposes.
    Third, important economic fundamentals of the marketplace are 
sufficiently similar in crucial respects. First, the heterogeneity of 
the willingness-to-pay among subscribers and listeners in the 
downstream market continues to support price discrimination and thus 
differentiated royalty rates upstream pursuant to the concept of 
``derived demand.'' See Determination at 19 (and record citations 
therein) (``Weighing all the evidence and based on the reasoning in 
this Determination, the Judges conclude that a flexible, revenue-based 
rate structure is the most efficient means of facilitating beneficial 
price discrimination in the downstream market.''); Dissent at 32, 51, 
86, 121, 126 (and record citations therein).\109\ Second, the items 
being licensed for transmission--``second copies'' of sound recordings 
(with embedded musical works)--have a marginal physical cost of zero, a 
critical economic point on which the experts for both parties concur, 
and as to which the Majority and the Dissent repeatedly and 
significantly rely. See Determination at 18, 21, 36, 59, 80 (and record 
citations therein); Dissent at 30-31, 33-34, 37, 47, 49-50, 59, 122, 
127-128 (and record citations therein).\110\
---------------------------------------------------------------------------

    \109\ The Determination asserts that it includes a price 
discriminatory feature because a revenue percentage-based rate is 
itself price discriminatory, in that it does not set royalties on a 
per-play basis. Determination at 35 n.71. But that ``blunt'' form of 
price discrimination does not capture the granular discriminatory 
features that the parties had negotiated. There is no sufficient 
basis for the Judges to substitute their own blunt conception of the 
appropriate form and extent of price discrimination for the 
structure generated in negotiations by the market participants. See 
Dissent at 37.
    \110\ It bears emphasis that the fact ``second copy'' 
reproductions are physically costless does not even suggest that the 
market price should be zero. Rather, in this ``second-best'' 
economic context, pricing above marginal physical costs is 
imperative in order for Copyright Owners to recover their ``first 
copy'' costs, avoid ``opportunity costs,'' and earn profits. See 
Dissent at 36-38.
---------------------------------------------------------------------------

    Copyright Owners are clearly correct, however, in noting a 
substantial change in economic circumstances that distinguished the 
Phonorecords II negotiations from the current proceeding; viz., the 
dramatic growth of interactive streaming revenues.\111\ The economic 
impact of this revenue growth is incorporated into the experts' Shapley 
Value Models and the Judges' analysis of same. This analysis has 
generated the 44% increase in the headline royalty rate, from 10.5% to 
15.1% (as phased-in by the Majority and again in this Initial 
Ruling).\112\
---------------------------------------------------------------------------

    \111\ Copyright Owners also cite data demonstrating the increase 
in listeners and the number of streams. The Judges find those data 
to be causal for the key point in rate setting in this proceeding--
the significant increase in revenues.
    \112\ At first blush it may seem that the increase in 
interactive revenues is not an economic fundament that would support 
an increase in a percentage-of-revenue based royalty formula. 
However, as more fully discussed herein, under the Shapley Value 
approach, the increase in revenues has generated an increased 
``Shapley Surplus'' (roughly analogous to interactive streaming 
industry profits), which the two ``Must Have'' input suppliers 
(record companies and Copyright Owners) will essentially split 
equally. If this surplus increases faster than the interactive 
services' non-content costs (or if those costs remain stable or 
fall), the increased revenues would flow disproportionately to 
theses input suppliers, thus causing the increase in revenues to 
support an increase in the royalty rate, all other things held 
constant. And, because the ``Must Have'' input suppliers have 
complementary oligopoly power, the Majority relied on a Shapley 
model constructed by Spotify's expert, Professor Marx, that adjusted 
for this market power.
---------------------------------------------------------------------------

    Simply put, three economic principles co-exist. First, the 
downstream interactive streaming market remains differentiated among 
listeners with different willingnesses and abilities to pay, based on 
varied preferences (utility) and disparities in income. Second, 
streaming of the ``second copy'' of the sound recordings (with embedded 
musical works) remains physically costless (but generates potential 
``opportunity costs''). But, third, streaming revenues have grown 
substantially. There is no incompatibility or inconsistency in the 
simultaneity of these economic principles. Each of them must be taken 
into account and they are in this Initial Ruling.
    This economic context refutes the arguments made during oral 
argument at the D.C. Circuit that the PR II-based benchmark should be 
rejected in toto because it was supposedly ``outdated.'' The 
heterogeneity of the downstream demand of listeners and the zero 
physical cost of ``second copies'' are enduring features that affect 
the upstream market via the principle of derived demand. The 
substantial growth of streaming revenues, however, necessitated an 
increase in the headline rate from 10.5% to 15.1% (as phased-in), for 
the reasons discussed in the Judges' analysis in this Initial Ruling of 
the interrelationship among: (1) Shapley Value modeling; (2) Nash 
Bargaining; (3) complementary oligopoly power; and (4) effective 
competition.
    Further, the foregoing analysis also undermines the pre-remand 
argument made by Copyright Owners that the PR II-based benchmark 
reflects a market that was not yet ``mature,'' or was only 
``experimental.'' Markets are not ``mature'' as opposed to, say, 
``adolescent.'' Indeed, the metaphor is strained because all economic 
models are subject to revision if the salient facts have changed, 
without rendering the prior models mere ``experiments.'' Markets 
simultaneously exhibit enduring characteristics--here, heterogeneous 
customers and zero marginal physical costs and dynamic change--here, 
significant revenue increases.\113\
---------------------------------------------------------------------------

    \113\ If one were to indulge the ``maturity'' metaphor, the 
ongoing creative destruction in the streaming industry has only 
reinforced the fact that, according to one of Copyright Owners' own 
economic expert witnesses, the interactive streaming market (as of 
the Phonorecords III hearing) was not yet mature, but rather 
remained ``a relatively new enterprise.'' Watt WRT ]] 39-40. Thus, 
it is hardly clear from the record that interactive streaming has 
``matured'' in a manner that would render anachronistic the enduring 
marketplace characteristics.
---------------------------------------------------------------------------

    And yet, Copyright Owners seek to deny the idea that these 
principles could exist simultaneously. In an attempt to disqualify the 
application of the PR II-based benchmark, Copyright Owners complain:

    [W]hile streaming activity and revenues grew under the 
Phonorecords II royalty rates, the [REDACTED]. For example . . . 
[REDACTED].

CO Initial Submission at 15-16 (emphasis added).
    But as the Services explained, the economic defect in Copyright 
Owners' analysis, is that it ignores the principle of price 
discrimination and its beneficial effects:

    [A]s [Professor] Hubbard explained, it is ``meaningless'' to 
compare growth in streams to growth in royalties in the context of 
Prime Music in particular because the record showed that Prime Music 
brings ``new people into the market.'' . . . If not for the 
flexibility (and beneficial price discrimination) the

[[Page 54437]]

existing Service Provider Revenue definition and rate structure 
facilitated, the Copyright Owners ``would have gotten zero'' from 
those new listeners. . . . ``So they're better off by that amount'' 
of royalty growth. . . . The undisputed fact that [REDACTED]--
reflects that the existing rule enables beneficial price 
discrimination that expands the total royalty pool and benefits 
Copyright Owners.

Services' Reply at 58-59.
    This rebuttal by Professor Hubbard is an example of the important 
distinction between ``increases in demand'' (when the demand curve 
shifts outward) and movements ``down the demand curve'' (when sellers 
use price discrimination to generate more revenue without additional 
cost to attract buyers with a lower willingness or ability to pay). The 
parties' otherwise dueling economists agreed on this point. Compare 4/
3/17 Tr. 4373-74 (Rysman) (Copyright Owners' witness acknowledging that 
under the current rate regime overall revenues might be increasing 
because of movements ``down the demand curve'' (i.e., changes in 
quantity demanded in response to lower prices), rather than because of, 
or in addition to, an outward shift of the demand curve (i.e., increase 
in demand at every price)) with 3/13/17 Tr. 701 (Katz) (the Services' 
witness who likewise noted that the present structure enhances variable 
pricing that allows streaming services ``to work[ ][their]way down the 
demand curve.'').
    Moreover, Copyright Owners baldly cherry-pick the data they 
present. [REDACTED] CO Initial Submission at 15-16. So, by their own 
data, presented in their own brief, they acknowledge that [REDACTED]. 
See Services' Reply at 57-58 (Copyright Owners have proven the 
``opposite'' of what they intended). This is precisely what beneficial 
price discrimination is designed to accomplish.\114\
---------------------------------------------------------------------------

    \114\ Further, [REDACTED] because: (1) the marginal physical 
cost of ``second-copy'' streams is zero; (2) royalties were 
calculated [REDACTED]; and (3) Copyright Owners' original proposed a 
per-play (i.e., per-stream) metric, which was rejected by all three 
of the Judges.
---------------------------------------------------------------------------

    The appropriateness of adopting the price discriminatory rate 
provisions of the PR II-based benchmark is further underscored by 
Copyright Owners' candid acknowledgement at the hearing that they were 
essentially urging the Judges to adopt what is known as the 
``Bargaining Room'' approach to rate setting. See Dissent at 24 (and 
record citations therein).\115\
---------------------------------------------------------------------------

    \115\ The Bargaining Room approach was first proposed for 
incorporation into the statutory license standard in 1967 by the 
NMPA, to be included in the predecessor section, later reorganized 
in section 801(b)(1) that governs this proceeding. See Dissent at 
22-24 (and citations therein). Ultimately, Congress punted on the 
Bargaining Room approach, and adopted into law the four-factor 
language set forth in section 801(b)(1). A subsequent attempt by 
NMPA to have the Copyright Royalty Tribunal (CRT) (a predecessor to 
the Judges) adopt the Bargaining Room theory was rejected by the 
CRT, a rejection that was affirmed on appeal. See Recording Industry 
Ass'n. of America v. Copyright Royalty Tribunal, 662 F.2d 1, 37 
(D.C. Cir. 1981), aff'g Adjustment of Royalty Payable under 
Compulsory License for Making and Distributing Phonorecords, 46 FR 
10466, 10478 (1981). See generally, F. Greenman & A. Deutsch, The 
Copyright Royalty Tribunal and the Statutory Mechanical Royalty: 
History and Prospect, 1 Cardozo Arts & Ent. L.J. 1, 53, 64 (1982).
---------------------------------------------------------------------------

    In the present proceeding, the appropriateness, vel non, of the 
Bargaining Room approach boils down to the following:

    Copyright Owners emphasize the inability of the Judges (or 
anyone) to identify present market rates precisely, let alone over 
the five-year rate period because the compulsory license set by the 
Judges cannot possibly contemplate every single business model that 
may develop in the ensuing time. . . . If the statutory rate is set 
below market rates, then the parties will never negotiate upward 
toward the market rates, because the licensees will always prefer to 
invoke the right to use the licensed work at the below-market 
statutory rates. However, if the Judges set the statutory rate above 
what they find to be market rates, different licensees who each have 
a maximum willingness to pay (WTP) below such a statutory rate would 
seek to negotiate lower rates with the licensors. In response to 
such requests to negotiate, according to this argument, Copyright 
Owners would respond by negotiating various lower rates for those 
licensees, provided lower rates were also in the self-interest of 
Copyright Owners.

Dissent at 24-25 (and record citations therein).
    The Judges find no reason to depart from the policy decision in 
Phonorecords I that the rate setting policies made explicit in section 
801(b)(1) are best discharged if the Judges eschew the Bargaining Room 
approach and continue to identify rate structures and rates that 
reflect the standards set forth in the statutory provision. To supplant 
the statutory factors with a Bargaining Room approach would essentially 
be to adopt a purely market-based rate-setting approach that is 
inconsistent with section 801(b)(1) and with the Judges' application of 
that statute to set rates, rate structures, and terms consonant with 
effective competition.
    With this background in mind, the Judges turn specifically to the 
interrelationship between the price discrimination aspects of the rates 
in the PR-II benchmark and the Bargaining Room approach.
    Copyright Owners have demonstrated (albeit tacitly) their 
understanding that, if the statutory provisions did not contain a price 
discriminatory rate structure to reflect the varying WTP, they would 
have to invent it. This finding is apparent from their advocacy for the 
adoption of a Bargaining Room approach to rate-setting. See, e.g., 4/3/
17 Tr. 4390, 4431 (Rysman) (lauding bargaining room approach as 
reflecting ``economical element of price discrimination . . . the 
[licensor] is picking its prices carefully.'') (emphasis added); id. at 
4431 (explaining that under this approach, when negotiating with 
Spotify regarding a rate for ad-supported service, ``Must Have'' music 
publishers would ``have the right . . . to set that price.''); 4/4/17 
Tr. 483-45 (Eisenach) (acknowledging Copyright Owners' approach was 
consistent with Bargaining Room theory because they were seeking rates 
so high as to force would-be licensees to negotiate for the ``Must 
Have'' mechanical license.).
    Thus, the Judges find there to be no real dispute as to whether 
there is a market-based need for an upstream discriminatory rate 
structure.\116\ Rather, the parties appear to be in disagreement as to 
who shall be in control of the setting of rates, the Judges, through 
their application of law, or Copyright Owners, through the exercise of 
their complementary oligopoly power. The resolution of this choice is 
clear; the Judges, not the licensors, are statutorily-charged with 
establishing provisions that are reasonable and otherwise properly 
reflect the itemized objectives of section 801(b)(1).
---------------------------------------------------------------------------

    \116\ The Majority recognized this point as well when--regarding 
the ``increase the total revenue that price discrimination enables--
they ask (and answer) rhetorically: ``How could Copyright Owners and 
their economic experts argue against a rate structure that inures to 
their benefit as well? The answer is: They do not. . . . [T]hey 
advocate for a rate set under the bargaining room theory, through 
which mutually beneficial rate structures can still be negotiated, 
but not subject to the ``reasonable rate'' and itemized factor 
analysis required by law.'' Determination at 85 & n.153. The Judges 
also note that Copyright Owners' acknowledgement that they too would 
set price discriminatory rates and structures is not simply a 
feature of this market. Rather, ``discriminatory pricing . . . is 
the normal attribute of equilibrium . . . in a broad range of market 
types and conditions where consumers can be separated into distinct 
groups with different demand elasticities.'' W. Baumol, Regulation 
Misled by Misread Theory: Perfect Competition and Competition-
Imposed Price Discrimination at 2 (2002). See also Dissent at 38, 
n.74. Given the ubiquity of discriminatory pricing, the Judges also 
find that the adoption into the statutory license of such pricing is 
not--as Copyright Owners contend--simply the inappropriate favoring 
of a particular business model, but rather a necessary reflection of 
the fundamental nature of market demand, particularly, the varied 
WTP among listeners.
---------------------------------------------------------------------------

    Fourth, the PR II-based benchmark reflect a rate structure with an 
adequate degree of competition, because there

[[Page 54438]]

was a balance of bargaining power between the two negotiating 
industrywide trade associations, offsetting the complementary oligopoly 
effects in place when a ``Must Have'' licensor bargains separately with 
each licensee. Recently, the Judges discussed in detail how the 
presence of countervailing bargaining power generates royalty rates at 
effectively competitive levels. See Web V, 86 FR 59452, 59457 (Oct. 27, 
2021).
    Further with regard to this fourth point, the parties have been 
operating over the past ten years under this basic rate structure, with 
profits accruing to the licensors and admittedly tolerable losses 
befalling the licensees. Moreover, after experience with these rates 
and this rate structure in the Phonorecords I period, they renewed and 
expanded this structure for use in the Phonorecords II period, when the 
alternative of a statutory rate proceeding was available to licensors 
and licensee alike. Their mutual willingness to continue in this manner 
is important evidence of the workability and reasonableness of this 
approach.
b. Evidence of Subjective Intent Not Prerequisite to Partial Adoption 
of the PR II-Based Benchmark \117\
---------------------------------------------------------------------------

    \117\ At the outset, the Judges reject Copyright Owners' 
contention that the D.C. Circuit found ``validity'' in their 
assertion that there was merit in Copyright Owners' assertion of the 
``subjective intent issue.'' Rather, on this issue, Johnson first 
held: ``[N]owhere does the [ ] Determination explain why evidence of 
the parties' subjective intent in negotiating the Phonorecords II 
settlement is a prerequisite to its adoption as a benchmark.'' 
Johnson, 969 F.3d at 387. Then, when Copyright Owners' appellate 
counsel attempted to cure that failure by making their own 
``subjective intent'' argument, the D.C. Circuit responded to that 
``subjective intent'' argument with a single word: ``Perhaps.'' Id. 
(emphasis added). This does not in any way suggest that Johnson 
found ``validity'' in the ``subjective intent'' argument, but rather 
was a non-committal response, consistent with the D.C. Circuit's 
ruling finding that the Determination had not explained this point.
---------------------------------------------------------------------------

    The Judges rely on the PR II-based benchmark as an objective 
benchmark. Thus, the absence of testimony regarding what went through 
the minds of the negotiators of the Phonorecords II agreement (and the 
predecessor Phonorecords I agreement) does not diminish the objective 
value of this benchmark. The Judges view the provisions of the PR II-
based benchmark as they would any benchmark, in the context of the 
requisite benchmarking elements identified and discussed supra. This 
approach allows the factfinder to analyze the benchmark through the 
lens of its service in the marketplace as an objective model for the 
market at issue, the Phonorecords III market. See, e.g., 3/13/17 Tr. 
550-51, 566 (Katz) (knowledge of why parties negotiated specific 
provisions is unnecessary, because objective results demonstrate 
satisfactory performances of market).
    Both Professors Katz and Hubbard noted that the current rate 
structure remains useful, not based on consideration of the parties' 
subjective understandings at the time of its creation, but because the 
market has not since changed in a manner that would create a basis for 
departure. Katz WDT ] 80 (``My analysis has identified no changes in 
industry conditions since then [2012] that would require changing the 
fundamental structure of the percentage-of-revenue prong.''); 4/13/17 
Tr. 5977-78 (Hubbard) (changes in market are ``not uncorrelated with 
the structure that was in place'' in 2012).\118\
---------------------------------------------------------------------------

    \118\ As noted supra, the relevant material change since the 
Phonorecords II agreement was reached is the significant growth in 
streaming revenues. That change is reflected in the Judges' 
application of the Shapley Value analyses, by which the Judges 
increased the headline royalty rate by 44%, from 10.5% to 15.1% 
(phased-in).
---------------------------------------------------------------------------

    In this regard, it bears emphasis that Copyright Owners' own 
witness, Dr. Eisenach, relied on several potential approaches that the 
Majority characterized as benchmarks for his rate analysis, without 
attempting to examine the subjective intent of the parties who 
negotiated those agreements. Indeed, the Majority found that the PR II 
Rates were properly considered as an objective benchmark, in the same 
manner as Dr. Eisenach's proffered benchmarks:

    The Services do not examine in detail the particular rates 
within the existing rate structure. Rather, they treat the rates 
within that structure as benchmarks, i.e., generally indicative of a 
sufficiently analogous market that has ``baked-in'' relevant 
economic considerations in arriving at an agreement. Dr. Eisenach 
did not analyze why he chose the levels for the rates and ratios on 
which he relied as benchmarks or consider the subjective 
understandings of the parties who negotiated his benchmarks. 
Similarly, the Services' economists elected to rely on the 2012 
rates as objectively useful without further inspection.
    This point is not made to be critical of Dr. Eisenach's 
approach, but rather to show that the Services' reliance on the 2012 
settlement as a benchmark shares this similar analytical 
characteristic, typical and appropriate for the benchmarking method. 
(The factual wrinkle here is that, hypothetically, the Services 
could have called witnesses and presented testimony regarding the 
negotiations that led to the 2012 (and 2008) settlements, but did 
not, rendering the 2012 benchmark similar to other benchmarks taken 
from other markets.)

Determination at 55 & n.106.\119\
---------------------------------------------------------------------------

    \119\ Copyright Owners do not deny that they did not offer 
evidence of subjective intent for Dr. Eisenach's benchmarks. Rather, 
they assert Dr. Eisenach's reliance on benchmarks without examining 
the subjective understandings of the negotiators of the benchmarks 
is irrelevant because: (1) Copyright Owners were not seeking the 
adoption in toto of the rates contained in any specific benchmark 
cited by Dr. Eisenach; (2) Dr. Eisenach analyzed multiple benchmarks 
to derive a reasonable range of rates; (3) his benchmarks were not 
adopted; and (4) his benchmarks and are not at issue on this remand. 
Copyright Owners Reply Brief on Remand at 28 n.19. But Copyright 
Owners confuse evidentiary standards with evidentiary application. 
Benchmarks are subject to the same evidentiary standards, regardless 
of the breadth of purpose for which they are proffered and 
regardless of whether they were adopted or rejected. Further, the 
fact that Dr. Eisenach's chosen benchmarks are ``not at issue on 
this remand'' does not render Copyright Owners' reliance on purely 
objective benchmarks uninformative as to their own understanding of 
the irrelevancy of the subjective thoughts of benchmark negotiators. 
See generally Web IV, 81 FR 26370 (proposed benchmark adjustment 
based on alleged ``additional value'' should be supported by 
``record evidence . . . to provide a basis for such for such an 
adjustment.'').
---------------------------------------------------------------------------

    Copyright Owners also aver that they entered into the Phonorecords 
II settlement simply to avoid litigation costs. Copyright Owners' Reply 
Brief on Remand at 29. At the hearing, this assertion was presented by 
David Israelite, NMPA's President. Israelite WRT ] 28; 3/29/17 Tr. 
3649-52 (Israelite) (claiming NMPA lacked financial position to fund 
rate litigation). The Services countered by noting that there was no 
evidence to support Mr. Israelite's testimony in this regard, or how it 
may have impacted the NMPA decision to participate. And, the Services 
pointed out, notwithstanding his testimony regarding financial 
constraints, NMPA had incurred the expense of a year-long negotiation 
with the Services to seek higher rates, create new service categories 
in subpart C, and change the TCC calculations. Id. at 159, 161-64; 3/
29/17 Tr. 3856 (Israelite).
    Further, as a general principle, a party's mere assertion that the 
Phonorecords II approach was the product of a settlement that was 
predicated on the avoidance of litigation costs savings does not 
invalidate its use as a benchmark in proceedings before the Judges, 
especially because, by statute, the Judges are authorized to consider 
such agreements. See Music Choice v. Copyright Royalty Board, 774 F.3d 
1000, 1014-15 (D.C. Cir. 2014) (testimony alleging agreement was 
reached to avoid litigation costs does not invalidate evidentiary use 
of that agreement for rate setting purposes, absent other evidence 
demonstrating settlement was involuntary or otherwise unreasonable.). 
Thus, the Judges find that the evidentiary record does not support 
Copyright Owners' position that this ``litigation cost avoidance'' 
assertion constituted a separate, idiosyncratic value that diminishes 
the

[[Page 54439]]

Judges' partial reliance on the PR II Rates in this Initial Ruling.
    Copyright Owners also mistakenly rely on the fact that the Services 
bore the burden of proof regarding the absence of any subjective 
idiosyncratic factors that hypothetically could have diminished the 
useful value of the PR II-based benchmark. Id. at n.21. The Services 
indeed bore the burden of proof (i.e., persuasion) with regard to their 
proffered benchmark PR II Rates, and they presented adequate objective 
evidence and testimony that this approach has worked in the marketplace 
to serve as prima facie proof to support the Judges'(partial) use of 
this benchmark in this remand proceeding. And, as explained above, such 
subjective intent was not a necessary element of their benchmark 
proofs. But, with regard to Copyright Owners' rebuttal to those proofs, 
Copyright Owners bore the burden of production, to present sufficient 
evidence and/or testimony that the Judges could rely on to reject the 
(partial) use of the PR II-based benchmark. This Copyright Owners 
failed to do.\120\
---------------------------------------------------------------------------

    \120\ As described in this Initial Ruling, the Judges identified 
this same distinction between the burden of proof and the burden of 
production to find in favor of Copyright Owners' proffered expert 
testimony in support of their Nash Bargaining analysis, testimony 
which constituted prima facie proof that was not adequately rebutted 
by the production of sufficient testimony from the Services' expert 
economic witnesses.
---------------------------------------------------------------------------

    In fact, given Copyright Owners' reliance on the subjective intent 
of the parties to a benchmark, the Judges attempted to identify 
potential subjective evidence of how the capped TCC rates in the PR II-
based benchmark \121\ were derived, during the examination of Dr. 
Eisenach at the hearing:
---------------------------------------------------------------------------

    \121\ The ``capped'' TCC rates are elements of the Phonorecords 
II rates.

    [JUDGE STRICKLER] Do you discuss, Dr. Eisenach, . . . in your 
written direct or written rebuttal testimony how the parties arrived 
. . . at the ratios for sound recording to musical works in [witness 
interrupts]
    [DR. EISENACH] That process is opaque to me, Your Honor.
    [JUDGE STRICKLER] Did you [witness interrupts]
    [DR. EISENACH] I know--I know there was a 2008 negotiation. I 
know there was a 2012 negotiation. I wasn't . . . present, and I'm 
not privy to any of the details.
    [JUDGE STRICKLER] You were not informed by your client or by any 
other source of information as to how they arrived at those 
particular ratios?
    [DR. EISENACH] When I've asked the question, I've found people 
chuckle and--and there doesn't seem to have been too much system--
systematic thought that went into it, but I don't really know that. 
I just--when I ask the question, people say: Nobody really knows. . 
. . Someone may know, but that's what I've been told.

4/4/17 Tr. 4611 (Eisenach) (emphasis added). The Judges find it 
perplexing, to say the least, that Copyright Owners would ``chuckle'' 
when asked by their expert witness for the very subjective evidence 
which they claim to be relevant. But of perhaps greater relevance is 
Dr. Eisenach's further testimony, quoted above, that he was also told 
by Copyright Owners that ``nobody really knows'' how the parties 
arrived at those rate ratios. Copyright Owners' ``chuckle,'' in 
response to its expert's critical inquiry as to the derivation of 
rates--and that expert's understanding that his client simply did not 
know how those rates were derived--undercut Copyright Owners' claim 
that subjective understanding of those rates could undermine their 
usefulness in the benchmark.\122\
---------------------------------------------------------------------------

    \122\ The Judges also find Copyright Owners' assertion that they 
did not know how those rates were established is not credible, given 
that they and their representatives negotiated those rates.
---------------------------------------------------------------------------

c. Substantial Evidence Demonstrates That PR II Rates, Other Than the 
Headline Rate, Are Not ``Too Low''
    As noted supra, one reason the D.C. Circuit vacated and remanded 
the Determination was because it declined to entertain the argument 
made only by appellee's counsel that ``the prior rates had been set far 
too low, thus negating the usefulness of the prior settlement as a 
benchmark.'' Johnson, 969 F.3d at 387. The Judges have noted throughout 
this Initial Ruling their adoption of the Shapley Value modeling 
analysis undertaken by the Majority, and raised the headline royalty 
rate by 44% from 10.5% to 15.1% (as phased-in), rendering moot 
appellate counsel's suggestion regarding the rate level.
    Here, the Judges further consider whether other rates within the PR 
II-based benchmark are reasonable, not only because they are part and 
parcel of the workable structure of that benchmark, but also to 
determine if they are supported by record evidence. To put this issue 
in context, those rates would apply on the second prong of the 
``greater-of'' rate structure in the PR II-based benchmark. The first 
prong in the PR II-based benchmark rates is the 10.5% revenue rate--
increased to 15.1% (as phased-in) by this Initial Ruling. The second 
prong consists of the ``lesser of'' a TCC rate or a per subscriber 
rate.\123\ For certain delivery configurations, these rates also cannot 
fall below any applicable Mechanical Floor. See Johnson, 969 F.3d at 
370.\124\
---------------------------------------------------------------------------

    \123\ This second prong contains only a TCC rate (i.e., an 
uncapped rate) for: (1) the ad-supported the service, because there 
are no subscribers to such a service; and for (2) bundled 
subscription service, for which there is a $0.25 per month floor but 
no per-subscriber cap, and Service Revenue for such bundles is 
calculated pursuant to 37 CFR 385.11 (``Service Revenue'' 
definition, ] 5).
    \124\ As Johnson explained, the CRB Judges ``retained the 
mechanical floor'' because, like so much of the PR II-based 
benchmark, it ```appropriately balances the [streaming service 
providers'] need for the predictability of an All-In rate with 
publishers' and songwriters' need for a failsafe to ensure that 
mechanical royalties will not vanish[.]'' Id. at 371-72. It is 
noteworthy that Copyright Owners urged the Judges (successfully) to 
maintain the Mechanical Floor provisions, which are the product of 
the Phonorecords II (and Phonorecords I) negotiations. Thus, it 
seems apparent that Copyright Owners as well as the Services 
consider provisions from the negotiated rates and rate structure to 
be in the nature of benchmarks, although differing as to which 
elements such be included or excluded. (The Services unsuccessfully 
argued for the elimination of the Mechanical Floors.) This 
perspective underscores the correctness of the Judges' decision on 
remand to treat the PR II-based benchmark as useful.
---------------------------------------------------------------------------

    The Services describe the key feature of these non-headline rates 
as the fostering of beneficial price discrimination, i.e., the adoption 
of ``different rate levels for different product offering,'' in order 
[t]to account for consumers' different willingness to pay [WTP] for 
music. Services' Joint Opening Brief (on Remand) at 21. As an example 
of how these price discriminatory rates impacted the market, the 
Services compare and contrast two Amazon offerings, Amazon Music 
Unlimited (for Echo) and Amazon Prime Music.
    Amazon Music Unlimited, with more than 30 million available songs 
as of the Phonorecords II proceeding period, see Mirchandani WDT ] 41, 
[REDACTED].\125\ By contrast, Amazon Prime Music, calculated as a 
``bundled subscription'' configuration, makes available only an 
abridged repertoire of 2 million songs, see Mirchandani, supra, and 
[REDACTED]. See id. at Sec.  385.13(a)(4).
---------------------------------------------------------------------------

    \125\ [REDACTED].
---------------------------------------------------------------------------

    Thus, Amazon pays [REDACTED] for listening by the more casual 
consumers who use the limited catalog Prime Music service at no 
additional charge beyond their Prime membership fee, compared to 
consumers who want the full repertoire provided by Amazon Music 
Unlimited on their Echo devices. See Services' Joint Opening Brief at 
71. These royalty obligations demonstrate the combination of price 
discrimination, product differentiation and ``derived demand'' in 
action; that is, the [REDACTED] are derived from the lower demand of 
consumers of the limited Amazon Prime Music service compared with 
subscribers to Amazon Music

[[Page 54440]]

Unlimited on their Echo devices, which in turn drive higher revenues.
    It is also important to note that these differential rates on the 
second prong of the ``greater-of'' structure of the PR II Rates are 
overridden by the revenue percentage rate on the first prong if that 
first prong rate generates more revenue. For example, [REDACTED], see 
Dissent at 29 (Table) and 116; see also [REDACTED]. With the headline 
rate now increased on a phased-in basis, the price discriminatory 
royalty generated by this [REDACTED].
    It is noteworthy that Johnson affirmed the Majority's setting of 
other price discriminatory features, e.g., the family and student plan 
provisions, based on the Judges' reliance on the Services' expert 
testimony regarding the benefits of ``having a way . . . where low 
willingness to pay consumers can still access music in a way that still 
allows more monetization of that provision of that service.'' Johnson, 
969 F.3d at 392-93. In similar fashion, the multi-tiered rates in the 
PR II-based benchmark likewise were supported by the same type of 
testimony; indeed, from expert testimony proffered by both parties, as 
considered below.
    First, Professor Katz notes that the existing rate structure 
captures two important aspects of the economics of the interactive 
streaming market: (1) the variable WTP among listeners; and (2) the 
corollary variable demand for streaming services. See 3/13/17 Tr. 586-
87 (Katz); see also Marx WRT ] 239 et seq.; 4/7/17 Tr. 5568 (Marx) 
(noting that the present structure serves differentiated products 
offered to customer segments with a variety of preferences and WTP). In 
more formal economic terms, Professor Katz notes that the present 
structure enhances variable pricing that allows streaming services ``to 
work [their] way down the demand curve,'' i.e., to engage in price 
discrimination that expands the market, providing increased revenue to 
the Copyright Owners as well as the Services. 3/13/17Tr. 701 (Katz).
    Second, in similar testimony, Professor Hubbard captures the 
interrelationship between the economics of this market and the existing 
rate structure:

    [F]rom an economic perspective, you can think of this market and 
this industry as being composed of different customer segments by 
tastes and preferences and willingness to pay. And so no rate 
structure can really work without understanding that, and no 
business model can really work without understanding that.
    [I]n terms of rate structures, the Phonorecords II framework 
from the previous proceeding does offer a benchmark to start because 
it provides for differences in distinct product categories in terms 
of music service offerings, pricing possibilities, and so on. And it 
has encouraged a very diverse digital music offering set from actual 
competitors.

3/21/17 Tr. 2175-76 (Hubbard). Moreover, Professor Hubbard [REDACTED] 
4/13/17 Tr. 5978 (Hubbard); see also Hubbard WDT ] 4.7 (the 2012 rate 
structure provides the ``necessary flexibility to accommodate the 
underlying economics of Amazon's various digital music service 
offerings.''). See also 3/15/17 Tr. 1176 (Leonard) (notwithstanding 
changes and growth in the streaming marketplace over current rate 
period, underlying economic structure of marketplace, which made 
percent-of-revenue based royalty appropriate, has not changed).
    Third, the Services' experts further assert that the multiple 
pricing structures necessary to satisfy the WTP and the differentiated 
quality preferences of downstream listeners relate directly to the 
upstream rate structure to be established in this proceeding. For 
example, Professor Marx opines that the appropriate upstream rate 
structure is derived from the characteristics of downstream demand. 3/
20/17 Tr. 1967 (Marx) (agreeing that rate structure upstream should be 
derived from need to exploit willingness to pay of various users 
downstream via percentage of revenue because downstream listeners have 
varying willingness to pay that should be exploited for mutual benefit 
of copyright licensees and licensors). Professor Marx further 
acknowledged that this upstream:downstream consonance in rate 
structures represents an application of the concept of ``derived 
demand,'' whereby the demand upstream for inputs is dependent upon the 
demand for the final product downstream. Id. Moreover, Dr. Leonard 
notes that reliance on the Services to identify segmented demand and 
develop price discriminatory approaches is appropriate because ``the 
downstream company is going to have a lot more information about . . . 
the business, about what makes sense.'' 4/6/17 Tr. 5238 (Leonard).
    Regarding a comparison of revenue growth to streaming growth, 
Professor Hubbard dismisses as economically ``meaningless'' Copyright 
Owners' argument that they have suffered relative economic injury under 
the current rate structure simply because the increase in their 
revenues from interactive streaming has been proportionately less than 
the growth in the number of interactive streams, leading mathematically 
to a lower implicit or effective per stream royalty rate. That is, he 
notes there is no evidence to rebut this prima facie indication of 
beneficial price discrimination, i.e., no contrary evidence indicating 
that, if the Services had sought to increase the price of the services 
available to these low to zero WTP listeners because of higher 
royalties, they would have paid the higher price, rather than declined 
to utilize a royalty-bearing interactive streaming service. See 4/13/17 
Tr. 5971-73 (Hubbard); see also Dissent at 52.
    The Services also link their price discrimination argument to the 
fact that the marginal physical cost of streaming is zero to the need 
for a flexible rate structure such as now exists. In this regard, 
Professor Hubbard notes that, because ``[t]he marginal production cost 
at issue here is--is zero. . . . it's not clear why it's not better to 
bring new customers into the market on which royalties would be paid 
and, of course, zero marginal cost incurred.'' 4/13/17 Tr. 5917-18 
(Hubbard). See also Marx WDT ] 97 (``Setting the price of marginal 
downstream listening at its marginal cost of zero induces more music 
consumption and variety than per-song or per-album pricing.'').
    Professor Marx makes the same argument as to the salutary nature of 
price discrimination in this context with regard to Spotify's ad-
supported approach. Focusing on the first purpose, Spotify is 
attracting ad-supported listeners who have a relatively low WTP, 
whether they have low incomes, (a budget constraint) or low interest in 
music (low ``utility,'' in the parlance of economists). These 
listeners, and the advertising revenue they generate are real and 
reflect the WTP of a large swath of all interactive listeners. See Marx 
WRT ] 115-16 & Fig. 9 (``While I agree that one aspect of the ad-
supported service is to provide an on-ramp to paid services, it also 
has another important aspect, namely to serve low WTP customers. . . . 
Copyright Owners' economists err in not calculating the impact of the 
Copyright Owners' proposal on ad-supported services. Ad-supported 
services currently make up a majority of subscribers and [REDACTED]% of 
all streams in the industry.'').
    Accordingly, a separate tier for an ad-supported service accounts 
for the different nature of the downstream listenership, allowing the 
upstream royalty to be based on that characteristic. This 
differentiation was essentially acknowledged by Copyright Owners late 
(too late, actually) when they proposed in their post-hearing filing 
that ``if the Judges intend to include the Spotify ad-supported

[[Page 54441]]

service in the rate structure and rate calculations, that they do so by 
establishing separate rates and terms for the ad-supported service. See 
COPCOL (Corrected) ] 228 & n.34. But the PR II-benchmark already 
incorporates separate rates for free/ad-supported services!\126\
---------------------------------------------------------------------------

    \126\ Copyright Owners also belatedly proposed that the Judges 
establish specific functionality limits on a separate ad-supported 
prong to avoid cannibalization of subscriber-based streaming with 
fuller functionality. Id. [REDACTED].
---------------------------------------------------------------------------

    Another important evidentiary factor buttressing the need for price 
discriminatory rates and structures was the testimony of the Services' 
survey expert, Mr. Robert Klein, Chair and co-founder of Applied 
Marketing Systems, Inc. Mr. Klein surveyed 2,101 people (the Klein 
Survey) who were listeners to streamed music and found, inter alia, 
that: (1) the majority of listeners would not pay for a monthly 
streaming subscription; and (2) for those who do subscribe, their 
demand was elastic, with increases in subscription prices causing 
overall greater percentage reductions in quantity demanded, moving 
customers to free, ad-supported and non-streaming alternatives. See 
Klein WRT ]] 60-67. By contrast, Copyright Owners did not present any 
survey testimony. The Determination fully credited the Klein Survey, 
finding as follows:

    It is important to note that Copyright Owners' attacks on the 
Klein Survey are not levelled by any witnesses, nor contradicted by 
their own survey expert, because Copyright Owners elected not to 
proffer such an expert in their direct (or rebuttal) cases. Rather, 
Copyright Owners elected to make a descriptive argument regarding 
the elasticity of demand among different segments of the market, as 
opposed to a survey-based or econometric study of price elasticity.
    [Although] Copyright Owners attack the Klein Survey on several 
fronts[,] [t]he arguments made by Copyright Owners are insufficient 
. . . to seriously weaken the probative value of the Klein Survey. 
In the end, the Judges are not persuaded by the Copyright Owners' 
revenue bundling arguments not to adopt a flexible, revenue-based 
royalty rate.

Determination at 22-23 & n.53; see also Dissent at 64-67 (including 
point-by-point rejection of Copyright Owners' non-expert criticisms of 
Klein Survey).
    The Services also note that the existing rate structure has 
produced generally positive practical consequences in the marketplace. 
Their joint accounting expert, Professor Mark Zmijewski, testified that 
the [REDACTED] from the sale of product under (former) Subpart A since 
2014 has been [REDACTED] over the same period. Expert Report of Mark E. 
Zmijewski February 15, 2017 ]] 38, 40 (Zmijewski WRT); 4/12/17 Tr. 5783 
(Zmijewski); see also 4/13/17 Tr. 5897 (Hubbard) (``the evidence that I 
reviewed suggests that the copyright holders have actually benefitted 
from this structure. . . .'').
    More particularly, Professor Zmijewski testified that:
     Total revenues reported by the NMPA for NMPA members from 
all royalty sources [REDACTED]. Zmijewski WRT ] 41.
     This [REDACTED]. Id.
     The [REDACTED]. Id.
     Mechanical royalty revenue for the sale of downloads and 
physical phonorecords [REDACTED]. Id. ] 38.\127\
---------------------------------------------------------------------------

    \127\ By contrast, Copyright Owners assert that the appropriate 
approach would only consider interactive service payment of 
mechanical royalties, and exclude performance royalties. On that 
basis, revenue, for the sale of digital downloads and physical 
phonorecords mechanical royalty revenue [REDACTED] from [REDACTED] 
in 2014 to [REDACTED] (as noted in (4) above, whereas mechanical 
royalty from streaming [REDACTED] from [REDACTED] in 2014 to 
[REDACTED] in 2015. Thus, the [REDACTED] in mechanical royalty 
revenue from streaming [REDACTED] in mechanical royalty revenue from 
the sale of digital and physical phonorecords. The Judges do not 
agree with Copyright Owners. Performance royalty and mechanical 
royalty payments made by the Services are for perfect complements--
neither license has any value to the Services unless they acquire 
both. Indeed, that is a critical reason why the mechanical rate is 
calculated on an ``All-In'' basis. Thus, it makes sense to make the 
comparison in the manner undertaken by Professor Zmijewski.
---------------------------------------------------------------------------

    In sum, the foregoing analysis demonstrates the economic 
reasonableness and appropriateness of the price discriminatory 
Phonorecords II rate structure and its negotiated safeguards to address 
the real possibility of revenue diminution. As discussed below, the 
record evidence also supports royalty rates within the PR II-based 
benchmark.\128\
---------------------------------------------------------------------------

    \128\ Again, to be clear, the Judges are substituting the 15.1% 
revenue rate for the 10.5% revenue rate as the headline rate in the 
``greater-of'' structure of the Phonorecords II benchmark. Thus, the 
price discriminatory royalty rates discussed below would apply only 
if they generated a ``greater'' level of revenue than the headline 
15.1% revenue rate. And, although the Mechanical Floor rate is not 
tied directly as an alternative to the ``greater-of'' revenue rate 
(now 15.1% as phased-in), it is not a floor that ignores the effect 
of that ``greater of'' rate. For example, assume the popular 
standalone portable subscription streaming service that people 
access on their mobile phones would pay an ``All-In'' musical works 
royalty of 15.1% based on the application of the two ``greater-of'' 
prongs. However, assume also the ``Performance Royalty'' that must 
be subtracted is 12%. That would leave 3.1% of service revenue 
attributable to the mechanical right. However, if that revenue rate 
of 3.1% yielded mechanical royalty revenue that was less than the 
royalty revenue generated by the applicable monthly mechanical floor 
of $0.50 per subscriber, then the mechanical floor would control. 
This application, like any other application of the mechanical 
floor, does not diminish the value of the 15.1% right, but rather 
limits its reduction under the ``All-In'' calculation. Recall also 
that the Determination, Dissent and Johnson do not disturb the All-
In and Mechanical Floor features of the Phonorecords II benchmark.) 
And finally, with regard to the actual per subscriber monetary 
values in the mechanical floors, no party suggested changes from 
rate levels in the PR II-based benchmark, including in the 
mechanical floor rates. The Judges recognize, as did Dr. Katz, 
Pandora's economic expert witness, that alternate values might have 
been preferable for rates contained in the PR II-based benchmark, 
but none were in the record. See 4/15/17 Tr. 5056-58 (Katz).
---------------------------------------------------------------------------

    The PR II-based benchmark contain several alternate rates 
explicitly calculated as a percentage of payments made by interactive 
streaming services to the record companies for sound recording rights. 
See Addendum to this Initial Ruling. In the Subpart relating to 
streaming, the (former) subpart B category, the TCC is 22% for ad-
supported services and 21% for portable subscriptions. Id.; see also 37 
CFR 385.13(b)(2) and (c)(2). These percentage figures correspond to 
sound recording: musical works royalty ratios of 4.55:1 and 4.76:1, 
respectively.
    With regard to these ratios, Copyright Owners' economic expert 
witness, Dr. Eisenach, stated: ``In my opinion, the evidence . . . 
indicates that the relative valuation ratios implied by the current 
Section 115 compulsory license . . . represent an upper bound on the 
relative market valuations of the sound recording and musical works 
rights.'' Id. ] 92 (emphasis added). (As an ``upper bound,'' these 
ratios would represent the lower bound on the relative market 
valuations of the reciprocal percentage of the value musical works 
rights relative to sound recording rights, again, 22% and 21%.\129\) 
Thus, there appears to be consensus between Copyright Owners' witness 
and the Services (who advocate for applying these rates on the price 
discriminatory tier of their benchmark) that these rates constitute 
``relative market valuations'' (even if they are not Dr. Eisenach's 
preferred market valuations within the bounded zone of such values).
---------------------------------------------------------------------------

    \129\ 1 / 4.55 = .219, or 22% (rounded); 1 / 4.76 = .210 (21%).
---------------------------------------------------------------------------

    Dr. Eisenach's testimony regarding the ``bounds'' of useful market 
valuations is noteworthy because his acknowledgement is consonant with 
judicial precedent. The Judges' setting of reasonable rates often 
requires them to identify a ``zone of reasonableness,'' within which 
they identify appropriate statutory rates. See, e.g., Intercollegiate 
Broadcasting System, Inc. v. Copyright Royalty Board, 684 F.3d 1332, 
1340 (D.C. Cir. 2012) (The CRB Judges' rate setting can necessitate the 
finding of a ``zone of reasonableness [because] ``[s]tatutory 
reasonableness is an

[[Page 54442]]

abstract quality represented by an area rather than a pinpoint.'').
    The 21% and 22% TCC rates within section 115 identified by Dr. 
Eisenach as generating the ``lower bound on the relative market 
valuations'' imply certain approximate percent-of-revenue rates, i.e., 
percent of total service revenue (not percent of sound recording 
revenue). See Dissent at 91, n.133 (sound recording rates clustered 
between [REDACTED]% and [REDACTED]% of revenue). For example, if the 
sound recording royalty rate for interactive streaming is [REDACTED]% 
of revenue, then the musical works rate would be calculated as 0.21 x 
[REDACTED], which equals [REDACTED]%, (or as .22 x [REDACTED] which 
equals [REDACTED]%). At the low end of the range, if the sound 
recording royalty rate is [REDACTED]%, then, applying these TCC 
figures, the implied musical work royalty rate would be calculated as 
[REDACTED]% (.21 x [REDACTED]) or [REDACTED]% (.22 x [REDACTED]).\130\
---------------------------------------------------------------------------

    \130\ Dr. Eisenach's identification of the 21%-22% TCC as within 
the bounds of market valuations may appear surprising at first in 
light of the higher 26.2% uncapped TCC rate pursued (unsuccessfully) 
on remand by Copyright Owners. But in the context of his testimony, 
Dr. Eisenach's opinion is understandable. The former headline rate 
of 10.5%, when sound recording rates ranged from approximately 
[REDACTED]% to [REDACTED]% of streaming revenues, yielded TCC rates 
between [REDACTED]% and [REDACTED]%. Thus, Dr. Eisenach was 
identifying a market valuation [REDACTED] (at his lower bound) 
between [REDACTED]% (the difference between 21% and [REDACTED]%) and 
[REDACTED]% (the difference between 22% and [REDACTED]%). Again, for 
context, this Initial Ruling raises the percentage rate by 44% when 
fully phased-in (based on the experts' Shapley analyses, 
significantly above the TCC rates advocated by Dr. Eisenach, even 
assuming the [REDACTED]%-[REDACTED]% sound recording rates on which 
he relied.
---------------------------------------------------------------------------

    It is important to emphasize and detail the context of these price 
discriminatory rates. These capped TCC rates are on the ``greater of 
prong'' that is compared with the headline 15.1% revenue rate (phased-
in) that the Judges are also adopting in this Initial Ruling. As phased 
in, the headline rate is greater than all the capped TCC-based rates 
identified in Dr. Eisenach's testimony, supra, [REDACTED]. For 2019, 
the phased-in headline percentage rate, 12.3%, is [REDACTED] the 
[REDACTED]% and [REDACTED]% revenue rates derived if the sound 
recording rates was [REDACTED]%. For 2018, the phased-in headline 
percentage rate, 11.4%, is [REDACTED] all the rates derived from the 
capped TCC rates Dr. Eisenach identified as ``market valuations'' 
(albeit the lower bound in his opinion). But that is of no negative 
consequence for Copyright Owners, because they would get paid on the 
``greater-of'' metric (capped TCC or headline rate) under the 
Phonorecords II-based rate structure the Judges are adopting (For the 
portable subscriptions, even though the 80 cents/subscriber ``lesser-
of'' portion of the non-headline prong would apply on that prong if it 
was lower than the capped TCC rate, the actual rate could not be lower 
than the phased-in headline rate.)
    Dr. Eisenach also examined direct agreements between record 
companies and interactive streaming services that contain rates for 
sound recordings and mechanical royalties, respectively. See, e.g., id. 
]] at 84-91. In such cases, the ratio of sound-recording to musical-
works royalties ranged tightly between [REDACTED] and [REDACTED], 
closely tracking the regulatory ratios implicit in the section 115 TCC. 
Id. ] 92. (The [REDACTED] ratio equates to a TCC rate of [REDACTED]%, 
and the [REDACTED] ratio equates to a mechanical rate of [REDACTED]%.). 
He concluded, as he did with regard to the actual section 115 license 
rates: ``In my opinion, the evidence presented . . . indicates that the 
relative valuation ratios implied by the . . . negotiations under [the 
statutory] shadow--ranging from [REDACTED] [[REDACTED]%] to [REDACTED] 
[[REDACTED]%]--represent an upper bound on the relative market 
valuations of the sound recording and musical works rights.'' Eisenach 
WDT ] 92 (emphasis added).
    Dr. Eisenach also identified several additional useful benchmarks. 
First, he identified what was coined the ``Pandora Opt-Out Agreement'' 
benchmark,\131\ which reflected a ratio of [REDACTED] of sound-
recordings to musical-works in a comparable benchmark setting. This 
ratio translates to a TCC percent of [REDACTED]%. With sound recording 
royalty rates of approximately [REDACTED]% to [REDACTED]%, this TCC 
reflects an effective percentage of total revenue equal to [REDACTED]% 
to [REDACTED]%.
---------------------------------------------------------------------------

    \131\ Pandora was only a noninteractive service at that time, 
and thus only paid the performance right royalty, not the mechanical 
right royalty, for the right to use musical works. Because the 
parties agree that the performance right and the mechanical right 
are perfect complements, Pandora's payments for the performance 
right are thus relevant and probative, as they reflect the full 
value of the musical works royalty to a noninteractive service. 
These factors became relevant because major music publishers had 
negotiated direct licensing agreements with Pandora for its 
noninteractive service covering the period from 2012 through 2018. 
Eisenach WDT ] 103. They negotiated these direct agreements after 
certain publishers had decided to ``opt-out,'' i.e., to withdraw 
their digital music performance rights from PROs, and asserted the 
right to negotiate directly with a digital streaming service. 
Pandora thus negotiated several such ``Opt-Out'' Agreements with an 
understanding that the rates contained in those direct agreements 
might not be subject to rate court review and thus could reflect 
market-based rates. Given this unique circumstance, and given that 
the markets and parties involved in the Pandora Opt-Out agreements 
are somewhat comparable to the markets and parties at issue in this 
proceeding, Dr. Eisenach concluded that these agreements provided 
``significant insight into the relative value of the sound recording 
and musical works rights in this proceeding.'' Id. (emphasis added). 
(The Judges did not adopt Dr. Eisenach's speculation that this 
performance royalty would continue to grow after 2018. See 
Determination at 51; Dissent at 102-103.)
---------------------------------------------------------------------------

    Second, Dr. Eisenach identified YouTube agreements with music 
publishers that relate to the combination of a commercial sound 
recording and a ``static image.'' The YouTube agreements contain an 
explicit royalty of [REDACTED].\132\ That [REDACTED]% royalty is a 
denominator in the ratio concept utilized by Dr. Eisenach, and the 
numerator is the [REDACTED] sound recording royalty paid to the record 
companies. YouTube had agreed to pay [REDACTED]% of its revenues, and 
had agreed to pay [REDACTED] and other record companies [REDACTED]% of 
revenues. The [REDACTED] ratio reduces to [REDACTED], implying a TCC 
([REDACTED]) of [REDACTED]%. The [REDACTED] ratio reduces to 
[REDACTED], implying a TCC ([REDACTED]) of [REDACTED]%. See Dissent at 
101-102.
---------------------------------------------------------------------------

    \132\ Dr. Eisenach preferred to use YouTube agreements that 
included [REDACTED], but the Judges relied on [REDACTED] as more 
comparative. Determination at 50; Dissent at 102.
---------------------------------------------------------------------------

    These additional rates identified in Dr. Eisenach's testimony 
further confirm the reasonableness of the non-headline rates within the 
PR II-based benchmark.
    Finally, the Judges look at the effective rates paid by Spotify, 
the largest interactive streaming service in terms of in terms of the 
number of subscriber-months and the number of plays. See Marx WRT ]] 
37-38 & Figs. 8 & 9. Under the PR II based benchmark, Spotify paid on 
its subscription service an effective ``All-In'' royalty rate of 
[REDACTED]% of its total revenues. See Dissent at 80, 115, 149 (and 
record citations therein). Spotify paid this effective percent-of-
revenue rate [REDACTED]. See id. at 29 (Table).
    Turning to Spotify's free/ad-supported offering (and as noted 
supra), Spotify paid royalties under the PR II Rates at an effective 
``All-In'' royalty rate of [REDACTED]%. Spotify paid this effective 
percent-of-revenue rate [REDACTED]. See id. When Spotify's two tiers 
are blended and averaged, the effective percent-of-revenue rate is 
[REDACTED]% of revenue. See id. at 116. The average rate has salience 
in this proceeding because Spotify's two

[[Page 54443]]

tiers are interrelated, in that free/ad-supported listeners constitute 
a pool of potential converts to the subscription tier under this 
``freemium'' model, even as this offering generates royalties under the 
PR II-based benchmark.
d. Copyright Owners' Concern Regarding Revenue Diminution Is 
Insufficient To Reject the PR II-Based Benchmark
    Copyright Owners argue that what the Services tout as beneficial 
price discrimination generates an ``incredible'' level of revenue 
diminution, including displacement, resulting in a ``major problem'' 
that reduces reportable revenues and thus the royalty base. See, e.g., 
3/7/22 Tr. 193 (Copyright Owners' counsel). This argument is based upon 
documents and evidence that demonstrated the following:
     [REDACTED];
     [REDACTED].
     [REDACTED];
     [REDACTED];
     [REDACTED];
     [REDACTED]
     [REDACTED];
     [REDACTED];
     [REDACTED]; and
     Copyright Owners' expert, Dr. Rysman, testified that 
interactive services often elect to forgo current profit maximization, 
e.g., by charging lower prices, in order to build a customer base and 
greater long-run profitability or value, from selling music and non-
music products or services to its customers.

CO Initial Submission at 40-42 (and record citations herein).

    The Services' economic experts do not ignore the fact that there 
can be revenue attribution problems when interactive streaming is 
combined with other products or services. They acknowledge that, even 
absent any wrongful intent with regard to the identification and 
measurement of revenue, attribution of revenue across product/service 
lines of various services can be difficult and imprecise. See, e.g., 4/
5/17 Tr. 5000 (Katz) (problem of measuring revenue ``certainly a factor 
that goes into thinking about reasonableness.'').
    However, Professor Katz testified that the existing rate structure 
agreed to by the parties accommodates these bundling, deferral, and 
displacement issues via the use of an alternative rate prong that would 
be triggered if the royalty revenue resulting from the headline rate of 
10.5% of streaming revenue fell below the royalty revenue generated by 
that second prong. Katz WDT ]] 82-83; 3/13/17 Tr. 670 (Katz). Moreover, 
Professor Katz concluded that, because the marketplace appears to be 
functioning (in the sense that publishers are earning profits and new 
and existing interactive streaming services continue to operate despite 
accounting losses), these revenue-measurement issues are being 
adequately handled by the alternative rate prong, even if an altered 
second prong might work better. Id. at 738-39. More generally, 
Professor Katz further noted that, the existing rates within the PR II-
based benchmark were performing well, and even if alternative minima 
might be preferable, no such alternative rates were in the record. See 
4/15/17 Tr. 5056-58 (Katz) (under the PR II-based benchmark ``the 
industry . . . was performing well,'' but ``if someone had a proposal 
[with] a specific reason why we should adjust this minimum that's 
something I would have examined,''). But Copyright Owners did not 
propose alternative rates or minima within the PR II-based benchmark, 
but instead urged the Judges to disregard the benchmark writ large. 
Accordingly, there were no alternative rates or minima in the record.
    Professor Katz further noted that the PR II-based benchmark rates 
were established when ``ecosystem'' entities such as Yahoo--akin to 
Amazon, Apple, and Google--were in the marketplace. 4/5/17 Tr. 5055-57 
(Katz); see also Determination at 31 (and record citations therein) 
(noting the presence of Microsoft as well as Yahoo as licensees in the 
interactive market during the Phonorecords II negotiations).
    More broadly, the Services' position regarding the use of the two 
prongs and their alternate rates to ameliorate the revenue-measurement 
problems is summed up by Professor Katz as follows:

    [T]he primary reason [for the two rate prongs] . . . is because 
of the measurement issues that can come up when having royalties 
based on a . . . percentage of revenues because there can be issues 
about how to appropriately assign revenues to a service. And so I 
think the minim[a] can play an important role when those--you know, 
when those measurement problems are severe, you can turn to the 
minimum instead. . . . [W]hat I have in mind, right, is that what 
would happen if you could imagine an entrepreneur coming along and 
saying we want to have a service and have some incredibly low price 
and not a very good monetization model, where a copyright owner 
would say--in an effectively competitive market, would say, wait a 
minute, I don't want to license to you on those terms. It's--I just 
think the possibility of getting a return is so low, I'm not going 
to do it, even though you, as an entrepreneur, are willing to try 
this. I as the copyright owner want some sort of, you know, return 
on it. And that's what the minimum also helps to do.

3/13/17 Tr. 599 (Katz.); see also 3/20/17 Tr. 1900-01 (Marx) (minima 
protect against revenue measurement problems); 4/7/17 Tr. 5584 (Marx) 
(statutory minima play ``two roles''--protecting the Copyright Owners 
from ``revenue mismeasurement'' by creating the ``greater of'' prong,'' 
but incorporating per subscriber rate prong in ``lesser of'' component 
to protect services from the record companies' use of their market 
power to engage in ``manipulation of the sound recording royalties'' on 
which the TCC prong is calculated).
    After considering the record, the Judges determine that the 
Majority had not found--as Copyright Owners claim--that the activities 
and strategies by the Services were ``incredible'' or a ``major 
problem. Rather, the Majority's characterization was measured, stating 
repeatedly that the Services engaged ``to some extent'' in revenue 
diminution because they ``might focus on long-term profit maximization 
to promote their long-term growth strategy, which occurs ``even absent 
wrongful intent.'' Determination at 20-21, 36, 90; accord, Dissent at 
59. In fact, the Majority specifically stated: ``The Judges agree that 
there is no support for any sweeping inference that cross-selling has 
diminished the revenue base.'' Id. at 21 (emphasis added). The Majority 
(and the Dissent) thus acknowledged the reasonableness of both sides of 
this issue, recognizing both the Services' use of price discriminatory 
approaches that can lower per user or per-stream revenues but grow 
royalties, market share and revenue, as well as Copyright Owners' 
concomitant desire to protect themselves from reductions in the royalty 
revenue base, however limited in extent, that would only serve to 
diminish royalties.
    One way the input supplier can avoid this impact is to refuse to 
accept a percent of revenue form of payment and move to a fixed per-
unit input price. This is what Copyright Owners originally and 
unsuccessfully sought in this proceeding, subject to a bargaining room 
approach by which they could switch back to the old approach (or any 
other approach) through purely market-based negotiations, unbounded by 
the statutory and regulatory standards of ``fairness'' and ``effective 
competition.'' See Dissent at 60.
    The Judges must reconcile the parties' competing considerations. A 
way by which they are both accommodated is through a pricing structure 
with

[[Page 54444]]

alternate rate prongs and floors, below which the royalty revenue 
cannot fall. This is precisely the bargain struck between Copyright 
Owners and services in 2008 and 2012, and that has been the rate 
structure through 2017. And, because the Majority and the Dissent found 
that revenue diminution occurred only ``to an extent,'' rather than in 
the pervasive (sweeping'') manner averred by Copyright Owners, there is 
no sufficient reason in the record to depart from the bargained-for 
multi-tiered rate structure in Phonorecords II that allows for price 
discrimination but tempers its impact on royalties through the use of 
minima and floors.
e. Copyright Owners' Claim of ``Inherent'' Economic Value Is Belied by 
the Record, Including Their Own Arguments
    Pre-remand, Copyright Owners approached this rate setting process 
with an overarching premise: A musical work has an ``inherent value'' 
that must be reflected in the royalty rates. As the NMPA's president, 
Mr. Israelite testified, when asked how ``inherent'' value is defined:

    [W]hoever owns an individual copyright is the one to define it. 
I think that would be the most appropriate definition of it. What 
someone is willing to license it for would be that inherent value to 
that owner . . . That would be market value.

3/29/17 Tr. 3707 (Israelite).
    If the market for musical works was as atomistic as the above quote 
assumes, the songwriter of an individual musical work could indeed set 
his or her own royalty rate, and refuse to license to any streaming 
service or other distributor who refused to pay that royalty. But that 
is not how the licensing market works.\133\ Songwriters typically 
assign their licensing rights to music publishers (to avoid ruinous 
transaction costs). These music publishers control huge ``Must Have'' 
repertoires that are offered under blanket licenses to streaming 
services. (The musical works market of course is subject to a 
compulsory license, but this is precisely how the unregulated market 
works for the licensing of sound recordings by labels to interactive 
streaming services.) It is acknowledged even by Copyright Owners' own 
expert witness, Professor Watt that the creation of these large 
collectives generates market power that necessitates rate regulation. 
See R. Watt, Copyright and Economic Theory: Friends or Foes at 163, 190 
(2000) (quoted in Dissent at 35).
---------------------------------------------------------------------------

    \133\ The record does not include evidence of self-marketing by 
songwriters through social media or via negotiation of individual 
royalty contracts by the exercise of overwhelming star power, 
whether through traditional payment mechanisms or new methods, such 
as the murky mechanism of non-fungible tokens (NFTs). The absence of 
incidents of such self-marketing from the record evidence in this 
proceeding suggests that they likely constitute but a small segment 
of the songwriter/publisher market. Accordingly, such self-marketing 
and individual negotiations do not impact the Judges' setting of 
statutory rates in this proceeding.
---------------------------------------------------------------------------

    Further, this ``inherent'' market value notion is antiquated as a 
matter of economics. Although an individual Copyright Owner can 
announce his or her ``asking'' royalty, that is not sufficient to 
generate a ``market'' royalty, unless and until a licensee agrees to 
pay it. In market-based economics. that is to say, the economic 
consensus that has governed economics since the ``marginal revolution'' 
in the mid to late 19th century, value is ascertained through the 
intersection of supply and demand, with the price established at the 
margin representing the market value of the good or service bought and 
sold.\134\ If there is no demand for a product, be it a musical work or 
anything else, it has no economic value. Even though costs have been 
incurred to produce the product, those costs cannot be recovered (or 
profit earned) absent a sufficient WTP in the market. And, as noted 
supra, the product being offered and at issue here is comprised of 
``second copies'' of sound recordings (with embedded musical works), 
which are costless to reproduce for streaming purposes. Of course, 
these ``second copies'' do have actual value when they are in demand, 
and the royalties that their licensing generates must cover: (1) the 
first copy (creative) costs; (2) the ``opportunity cost'' (measured by 
the next best alternative for royalty earnings if the ``second copies'' 
could have been supplied through another distribution channel that paid 
higher royalties to attract the end-user/consumer at issue); and (3) 
profits to induce the creation of musical works.
---------------------------------------------------------------------------

    \134\ As one scholar has summarized the 19th century transition 
from classical to neoclassical economics: ``By the early 1870s, 
economics reached a tipping point, and it ushered in a revolution in 
thought, signaling the beginning of the ``modern,'' or 
``neoclassical'' era. Marginalists flipped classical economics on 
its head. Instead of focusing on the production side of economics, 
they turned to consumption. It is the satisfaction of the wants of 
consumers that matters for value, not the labor required for 
production. What established the overall value of a good is the 
value fetched by the final unit of that item on the market. As more 
units of a good are produced, the marginal value of the last unit 
tends to decrease. . . . According to marginal utility, the 
consumer, not the producer, therefore drives the valuation 
process.'' J. Wasserman, The Marginal Revolutionaries at 28 (2019). 
This transformation reflected the abandonment of the ``labor theory 
of value''--the cornerstone of Marxian economics. See E.R. 
Canterbury, A Brief History of Economics at 111 (2001) (``Marx's 
devotion to a labor theory of value was complete.''). It initially 
appears as irony that Copyright Owners espouse a Marxian approach to 
value while preaching the virtues of unregulated markets. The 
initial whiff of irony dissipates when one appreciates that a 
collective licensor with the market power of control over a ``Must 
Have'' input has every incentive