(1)
Scope and Purpose. This Part sets forth basic
requirements for forecasting and applies to electric and gas forecasts and
supplements filed on and after December 1, 1978. 980 CMR 7.09(2) applies to all
forecasting methodologies while 980 CMR 7.09(3) applies only where econometric
approaches are employed.
(2)
General Requirements. With respect to all forecasting
methodologies:
(a) Each is to be designed to
accommodate sensitivity testing of major assumptions and parameters.
(b) Where computer programs are employed,
these are to be made available to the Council upon request.
(c) Where computer programs are employed,
source listings of computer programs, descriptions of the structure of each
program and subprogram, and descriptions of the relationships between logical
parts are to be made available to the Council upon request.
(d) Each forecasting methodology must
explicitly consider and quantify the following factors:
1. conservation programs and policies of the
Commonwealth;
2. conservation
programs and policies of the federal government;
3. the penetration of alternative energy
technologies, including but not limited to, wind and solar devices,
cogeneration, and biomass conversion;
4. improvements in the efficiencies of new
and existing appliances and machinery, including building insulation;
5. behavioral changes with respect to energy
use by customers, such as reduction in heating, cooling, and lighting
levels;
6. responses to higher
price levels, potential or actual changes in rate structure, and load
management.
(3)
Econometric Forecasting
Models. Documentationof econometric forecasting methodologies for
final form equations shall include:
(a) a
statement of the a priori theoretical or empirical basis for functional form
and variable selection;
(b) a
statement of the procedures by which variables have been added or deleted in
the development of final form specification;
(c) justification of all estimated
coefficients which have incorrect signs;
(d) identification of the method of
estimation;
(e) statistical
documentation of each equation; including standard error of estimation,
residual sum of squares, F-test, R-squared (adjusted or unadjusted), and a test
for serial correlation for time series data;
(f) statistical documentation of each
estimated parameter of each independent variable, including t-value or standard
error at the same specified level of significance, mean of the dependent
variables, and the standard deviation of each dependent variable.
Upon request of the Council, the petitioner or lead company
shall provide copies of the original computer outputs of all regressions, all
data files together with identification of the sources of the data, and an
index of variable names for cross-referencing between computer outputs and the
forecast documentation.
