Federal Electricity: Retail Competition Could Create Government Savings
(Letter Report, 09/30/97, GAO/RCED-97-244).

The federal government--the largest consumer of electricity in the
United States--could benefit by purchasing electricity in a competitive
market.  GAO estimated potential savings by referring to several
published forecasts of the extent to which competition may reduce
electricity prices. GAO found that the government could save anywhere
from $1 billion to more than $8 billion over the 18-year period from
1998 to 2015 if it purchased electricity competitively.  However,
falling electricity prices would likely cause the government to buy more
electricity than it would have bought at higher prices.  After adjusting
for that likelihood, GAO estimated that federal spending on electricity
would decrease by $0.6 billion to $6.5 billion during the same period.

--------------------------- Indexing Terms -----------------------------

 REPORTNUM:  RCED-97-244
     TITLE:  Federal Electricity: Retail Competition Could Create 
             Government Savings
      DATE:  09/30/97
   SUBJECT:  Energy consumption
             Federal procurement
             Energy costs
             Electric energy
             Competition
             Energy industry
             Price regulation
             Projections
             Utility rates

             
******************************************************************
** This file contains an ASCII representation of the text of a  **
** GAO report.  Delineations within the text indicating chapter **
** titles, headings, and bullets are preserved.  Major          **
** divisions and subdivisions of the text, such as Chapters,    **
** Sections, and Appendixes, are identified by double and       **
** single lines.  The numbers on the right end of these lines   **
** indicate the position of each of the subsections in the      **
** document outline.  These numbers do NOT correspond with the  **
** page numbers of the printed product.                         **
**                                                              **
** No attempt has been made to display graphic images, although **
** figure captions are reproduced.  Tables are included, but    **
** may not resemble those in the printed version.               **
**                                                              **
** Please see the PDF (Portable Document Format) file, when     **
** available, for a complete electronic file of the printed     **
** document's contents.                                         **
**                                                              **
** A printed copy of this report may be obtained from the GAO   **
** Document Distribution Center.  For further details, please   **
** send an e-mail message to:                                   **
**                                                              **
**                                            **
**                                                              **
** with the message 'info' in the body.                         **
******************************************************************


Cover
================================================================ COVER


Report to Congressional Requesters

September 1997

FEDERAL ELECTRICITY - RETAIL
COMPETITION COULD CREATE
GOVERNMENT SAVINGS

GAO/RCED-97-244

Federal Electricity

(141036)


Abbreviations
=============================================================== ABBREV

  CSE - Citizens For A Sound Economy Foundation
  DOD - Department of Defense
  DOE - Department of Energy
  DRI -
  EIA - Energy Information Administration
  FEMP - Office of Federal Energy Management Programs
  GRI - Gas Research Institute
  kWh - kilowatt-hours
  WEFA -

Letter
=============================================================== LETTER


B-277876

September 30, 1997

Congressional Requesters

The federal government is the largest consumer of electricity in the
United States, spending about $2.8 billion in fiscal year 1995 for
the approximately 50 billion kilowatt-hours (kWh) of electricity it
used domestically, at an average price of 5.6 cents per kWh.\1
Electricity restructuring, which is now being implemented or
discussed in many states and in the Congress, would allow most retail
customers, including the federal government, to purchase electricity
in a competitive market, much as customers now choose among different
long-distance telephone providers.  Many of these various regulatory
reforms to restructure the electricity industry share the common goal
of bringing competition to the electricity market by guaranteeing
customers a choice of service providers. 

As requested, we evaluated whether the federal government could
realize savings\2 if, as expected, lower prices result from retail
competition in the electricity market.  To do this, we estimated the
quantities of electricity the federal government would have purchased
without retail competition and the prices it would have paid during
1998 through 2015.  This projection of prices and quantities is known
as our baseline projection. 

To estimate possible savings, we used the results of several
published forecasts of the extent to which retail competition is
expected to reduce electricity prices, and we applied these results
to our baseline estimates of the government's projected spending for
electricity.  Unless otherwise stated, all dollar estimates represent
1996 present values and are based on the period 1998 through 2015.\3
We note that secondary effects--that is, the ripple effects of lower
electricity prices on the economy--could be substantial and result in
even greater savings to the federal government.  For example,
declining electricity prices could result in greater economic growth,
leading to increased tax revenues; they also could result in lower
inflation rates, leading to smaller increases in federal spending for
those programs with benefit levels that are adjusted for inflation. 
(App.  I contains a more detailed discussion of our methodology.)


--------------------
\1 A watt is the basic unit used to measure electricity.  A kilowatt
is 1,000 watts.  A kilowatt-hour is equal to 1 kilowatt of power
applied for 1 hour.  The average household in the United States uses
about 10,000 kWh of electricity annually, according to the Department
of Energy's Energy Information Administration (EIA).  In our March
1997 report entitled Energy Consumption:  Federal Agencies'
Electricity Use and Cost (GAO/RCED-97-97R, Mar.  21, 1997), we
provided usage and cost data across federal agencies.  For this
report, we revised the fiscal year 1995 data to reflect only that
portion that would be affected by retail competition and converted
the data to constant 1996 dollars. 

\2 As used in this report, the term "savings" does not represent the
budgetary savings that might be realized annually in the federal
budget from lower electricity costs. 

\3 We limited our estimates of these discounted future savings
through 2015 because EIA's projections of electricity prices, which
we needed to develop our estimates, do not go beyond 2015. 


   RESULTS IN BRIEF
------------------------------------------------------------ Letter :1

Retail competition in the electricity industry would create savings
for the government; however, the actual amount of savings is highly
uncertain.  Using published forecasts of electricity prices under
competition, we modeled the federal government's electricity
purchases and estimate that the government could cumulatively save
from $1 billion to slightly over $8 billion during the 18-year period
from 1998 through 2015 if it purchased the baseline quantities of
electricity--that is, the quantities of electricity it would have
purchased without retail competition.  Our wide range of estimates
reflects, among other things, the substantial uncertainty surrounding
the future pace of the implementation of retail competition in the
United States and the prices paid and quantities purchased by the
government.  We note that our estimated baseline quantities reflect
the effects of uncertainty over non-price factors, such as federal
mandates for energy efficiency, on the government's electricity
purchases.  Nevertheless, holding these other factors constant, we
believe that falling electricity prices would likely cause the
government to buy more electricity than the baseline quantities
simply because the price would have declined more under
competition.\4 When this happens, the government could be thought of
as "spending" some of its savings (when purchasing the baseline
quantities) to buy more electricity.  As a result, the government's
spending on electricity might not fall by as much as its savings if
purchases were held at the baseline quantities.  Adjusting our
savings estimate to reflect this case, we estimate that federal
spending on electricity could cumulatively decrease by $0.6 billion
to $6.5 billion during the same 18-year period--that is, 1998 through
2015--because of the decline in prices resulting from retail
competition. 


--------------------
\4 In stating that falling electricity prices would likely cause the
government to buy more electricity, we are implicitly holding
constant all other factors that could affect the government's
electricity purchases, such as the size of the government.  In so
doing, we are able to analyze and discuss the sole effect of lower
prices resulting from retail competition on the government's spending
for electricity. 


   BACKGROUND
------------------------------------------------------------ Letter :2

Federal and state governments are actively considering regulatory
reforms to restructure the electricity industry--an industry with
total assets worth about $500 billion and net revenues of over $200
billion annually.\5 In 1992, the Congress enacted the Energy Policy
Act, which, among other things, promotes market competition in the
wholesale electricity industry.  (The wholesale electricity industry
is comprised of intermediaries such as electric utilities that resell
electricity to their retail customers.) In addition, under the act,
the states may pursue their own reforms in the retail electricity
market.  Currently, a number of bills to restructure the retail
electricity industry to promote a more efficient and market-driven
industry are before the Congress. 

Forty-nine states have considered reforming their retail electricity
markets, and 13 states have implemented plans to restructure the
industry, either by enacting legislation or by issuing comprehensive
regulatory orders.\6 For example, in California, where a plan is
furthest along, competitive electricity markets and consumer choice
are scheduled to begin in 1998.  California's plan calls for a rate
cut of no less than 10 percent for all small customers by January 1,
1998.  The plan also calls for the full recovery of utilities'
stranded costs over a 10-year period.\7 However, not all states'
plans call for the full recovery of stranded costs.  For example, the
New Hampshire plan calls for all customers to have retail choice by
January 1, 1998 (with a possible delay until July 1, 1998 or later);
utilities will be allowed to recover only some of their stranded
costs.  The full recovery of stranded costs is unlikely because, as
the plan points out, the primary goal of restructuring is to lower
electricity rates. 

During their deliberations, the Congress and the states are
discussing the numerous potential benefits of electricity
restructuring in order to bring about retail competition.  These
potential benefits include, among other things, (1) increased
competition, (2) lower prices, (3) lower operating costs for
business, (4) smaller regional differences in costs, (5) more jobs,
(6) more reliable service, and (7) a cleaner environment. 


--------------------
\5 The markets for the transmission and distribution of electricity
are likely to remain regulated for the foreseeable future. 

\6 The 13 states are Arizona, California, Maine, Massachusetts,
Montana, Nevada, New Hampshire, New Jersey, New York, Oklahoma,
Pennsylvania, Rhode Island, and Vermont. 

\7 Stranded costs are investments or assets owned by regulated
electric utilities that are not likely to be competitive in a
restructured marketplace.  For example, some utilities own nuclear
power plants that have very high operating costs, and these plants
are not likely to be competitive in the new retail market.  In
addition, such power plants often have high debt levels.  Estimates
of these stranded costs vary widely, from a low of roughly $10
billion to a high of about $500 billion.  The California plan calls
for the recovery of stranded costs through a non-bypassable
competitive transition charge.  (Non-bypassable means that the charge
will be imposed in such a way that consumers cannot avoid paying it,
whether they stay with their current utility or choose a new
supplier.)


   GOVERNMENT SAVINGS IN
   ELECTRICITY SPENDING COULD
   RESULT FROM RETAIL COMPETITION
------------------------------------------------------------ Letter :3

If the federal government could purchase its electricity on a
competitive basis, we estimate that it could cumulatively save from
$1.0 billion to $8.2 billion during the 18-year period from 1998
through 2015 if it purchased the baseline quantities of
electricity--that is, the quantities of electricity it would have
purchased without retail competition.  We note that the average
annual savings in 1998 through 2015 range from a low of about $60
million to a high of about $460 million.  We further note, however,
that our estimates could be below or above this range and are subject
to, among other things, substantial uncertainty over the (1) future
pace of implementing retail competition in the United States, (2)
extent to which competition would reduce electricity prices for the
government, (3) quantities of electricity that the government would
have purchased and of the prices that it would have paid without
retail competition, and (4) prices for electricity that the
government would pay with retail competition.  In addition, our
estimated baseline quantities reflect the effects of uncertainty over
non-price factors, such as federal mandates to increase energy
efficiency. 

Nevertheless, falling electricity prices would likely cause the
federal government to buy more than the baseline quantities simply
because the prices would have declined more under competition.\8 As a
result, the government's spending on electricity would not fall by as
much as its overall savings.  If the government buys more than the
estimated baseline quantities of electricity because of lower prices,
we estimate that the government's spending on electricity could
cumulatively decrease by $0.6 billion to $6.5 billion during the same
18-year period, that is, 1998 through 2015, because of the decline in
prices resulting from retail competition.  We note that the average
annual declines in spending in 1998 through 2015 range from a low of
about $30 million to a high of about $360 million. 

To estimate the level of savings, we used our baseline for the
government's future spending on electricity.  We developed this
baseline using the projections of (1) future prices published by the
Department of Energy's Energy Information Administration (EIA)\9 for
electricity under existing circumstances, that is, assuming no
prevailing retail competition and (2) the government's baseline usage
for three different growth cases.  We then subtracted from this
baseline the projections of lower government spending for electricity
using the baseline usage multiplied by the competitive prices.  We
used three competitive pricing projections:  those reported by
DRI/McGraw-Hill (DRI),\10 the Gas Research Institute (GRI),\11 and
Citizens For A Sound Economy Foundation (CSE).\12 Finally, we
discounted the resulting stream of future annual dollar savings to
determine the present value of these savings.  (A more detailed
discussion of our methodology is contained in app.  I.)

In developing our baseline of the federal government's projected
electricity usage, we started with the quantity of kWh the government
used domestically in fiscal year 1995, the latest year for which data
were available.  We adjusted this quantity by factors that can both
increase and decrease the quantity of electricity.  Increases in
usage are likely because even without the introduction of retail
competition, future electricity prices are expected to fall,
according to EIA's projections, and the government as a consumer will
tend to buy more of a product--even electricity--if the cost of the
product declines.  On the other hand, factors unrelated to the price
of electricity could result in a net decrease in electricity usage. 
For example, current federal mandates that require federal agencies
to implement energy efficiency measures and possible future efforts
to downsize these agencies could decrease the government's
electricity usage, offsetting, for example, the increased usage from
the greater use of computers.  Thus, the net effect of these
non-price factors could be to lower the federal government's usage of
electricity. 

Other types of uncertainty are also associated with any estimates of
the federal government's savings on spending for electricity.  The
estimates of electricity prices under retail competition vary widely,
depending on the key assumptions used in making the estimates and the
applicability of the estimates to the federal sector.  For example,
we used published estimates of the extent to which retail competition
would decrease electricity prices for specific customer classes,\13
but the government's electricity usage is spread among all classes. 
Therefore, simply taking the results of any one of these classes to
determine the government's savings may overstate or understate these
savings.\14 Moreover, because the government generally pays a
relatively low price for the electricity it currently buys,\15 the
actual savings from retail competition may be less than they would be
for other customers and therefore are more likely to be at the lower
end of our range of estimates. 

Furthermore, the recovery of stranded costs could affect prices under
retail competition.  For example, DRI's and GRI's price projections
under retail competition include the recovery of large amounts of
stranded costs that are passed on to consumers through higher
electricity rates, while CSE's price projections include none.  We
found that CSE's estimates resulted in the greatest savings in the
government's spending.  While we realize that other significant
factors also affect the amount of the price declines that would
result from retail competition, we believe that the assumption that
stranded costs will be recovered is an important factor that helps to
explain to some degree the differences in the estimates of the
savings.  (App.  II contains a more detailed discussion of the
results of our simulation analyses of electricity savings.)

In a competitive retail market for electricity, the reduction in
federal spending for electricity would be less than the government's
estimated savings if federal usage were to increase above the
baseline quantities.  As electricity prices declined below the
baseline prices, federal usage would likely increase in response to
this additional decline in prices.  Thus, spending would decrease
less than it would have if usage had remained at the baseline
quantities.  We estimate that the government's spending on
electricity could cumulatively decrease by $0.6 billion to $6.5
billion during 1998 through 2015 (or about 23 percent less, on
average, than the decline in spending if the quantities purchased
were held at our baseline) because of the decline in prices resulting
from retail competition.\16 We note that the average annual declines
in spending during 1998 through 2015 range from a low of about $30
million to a high of about $360 million. 

We note that this estimate is subject to the same uncertainties as in
our estimate of savings.  Furthermore, this estimate is subject to
the same uncertainty as the baseline about the extent to which the
government would likely increase its usage of electricity simply
because of lower prices--in this instance, lower competitive prices. 
(App.  II presents a more detailed discussion of the results of our
analysis of price effects.)


--------------------
\8 The price of electricity is one of many factors that influence the
amount of electricity that the government purchases.  Other factors
include, for example, the size of the government, the prices and
availability of alternative forms of energy, and the existence of any
federal mandates that require federal agencies to use less energy or
that in some way affect electricity usage.  These other factors can
change at the same time that electricity prices change, making it
difficult to identify the separate effect of any one factor.  By
stating that falling electricity prices would likely cause the
government to buy more electricity, we are implicitly holding
constant these other factors to allow us to analyze solely the effect
of the expected lower electricity prices resulting from retail
competition. 

\9 These projections are contained in EIA's report entitled Annual
Energy Outlook 1997 With Projections To 2015 (Washington, D.C.:  Dec. 
1996). 

\10 DRI/McGraw-Hill, World Energy Service:  U.S.  Outlook (Spring
1997). 

\11 Gas Research Institute, GRI Baseline Projection of U.S.  Energy
Supply and Demand To 2015, 1997 Edition (Mar.  1997). 

\12 Michael T.  Maloney and Robert E.  McCormick with Raymond D. 
Sauer, Customer Choice, Consumer Value:  An Analysis of Retail
Competition in America's Electric Industry.  Prepared for and
published by Citizens For A Sound Economy Foundation (Washington,
D.C.:  1996). 

\13 EIA assigns retail consumers to one of three customer
classes--residential, commercial, or industrial--depending on certain
characteristics of the consumers' facilities.  In addition, the
classes are aggregated to form a fourth class--average.  Because the
federal government's facilities vary widely, these facilities are
scattered among the first three different classes. 

\14 App.  II provides simulation analyses that estimate a range of
savings to the government using all four customer classes. 

\15 In fiscal year 1995, the federal government paid, on average, 5.6
cents per kWh, while the average retail price was 7.3 cents per kWh
(prices in constant 1996 dollars). 

\16 To estimate the increase in the quantities of electricity that
would be purchased because of the further decline in prices brought
about by retail competition, we used the estimates of price
elasticity--the degree to which the quantity purchased responds to
price changes--obtained from (1) a discussion with an EIA official
and (2) the CSE study. 


   CONCLUSIONS
------------------------------------------------------------ Letter :4

While future prices for electricity are expected to be lower for the
federal government, even under the current regulatory structure, the
restructuring of the electricity industry in order to foster retail
competition is likely to result in even lower prices.  Our analysis
shows that the federal government would be likely to receive
substantial financial benefits if it had the ability to purchase its
electricity on a competitive basis.  However, there are uncertainties
that would affect the magnitude of these benefits. 


   AGENCY COMMENTS AND OUR
   EVALUATION
------------------------------------------------------------ Letter :5

We provided a draft of this report to DOE for its review and comment. 
DOE stated that our report appears to fairly assess the major issues
that affect federal electricity use and identifies the great
uncertainty in the economic environment within which the federal
government buys electricity.  In addition, DOE said that our
analytical approach appears sound in terms of answering the narrow
question of how much the federal government could save on electricity
expenditures in a restructured environment, and that our range of
savings appears reasonable given the large uncertainties surrounding
the timing of the advent of retail competition and the treatment of
stranded costs.  DOE also stated that federal agencies are currently
working under Executive Order 12902, which requires a reduction in
energy usage by 2005.  We note that we included such possible
reductions in electricity usage in calculating our savings estimates. 
In commenting on the secondary effects of lower electricity prices on
economic growth and inflation, DOE stated that our discussion appears
to be stated with excessive caution.  In addition, DOE said that some
assessment of the relative magnitude of the effects on the federal
budget through reduced inflation compared with the effects resulting
from changes in federal electricity costs might be of great interest
to readers of this report.  We recognize the potential importance of
reduced inflation and other secondary effects on the federal budget,
and as we stated in the draft report, these effects could be
substantial and result in even greater savings to the federal
government.  However, for this review, we concentrated our efforts on
quantifying the direct effect of lower electricity prices on federal
expenditures, and while we agree that some assessment of the
magnitude of these secondary effects would be useful, we did not
explicitly estimate their magnitude.  DOE's complete response is
presented in appendix III. 


---------------------------------------------------------- Letter :5.1

We conducted our work from May 1997 through September 1997 in
accordance with generally accepted government auditing standards.  As
arranged with your offices, unless you publicly announce its contents
earlier, we plan no further distribution of this report until 30 days
after the date of this letter.  At that time, we will send copies to
the Secretary of Energy, the Administrator of the Energy Information
Administration, and the Secretary of Defense.  We will also make
copies available to others on request. 

If you or your staff have any questions concerning this report,
please call me at (202) 512-3841.  Major contributors to this report
are listed in appendix IV. 

Susan D.  Kladiva
Acting Associate Director, Energy,
 Resources, and Science Issues


Congressional Requesters

The Honorable Thomas J.  Bliley, Jr.
Chairman, Committee on Commerce
House of Representatives

The Honorable Dan Schaefer
Chairman, Subcommittee on Energy and Power
Committee on Commerce
House of Representatives

The Honorable James M.  Inhofe
Chairman
The Honorable Charles S.  Robb
Ranking Minority Member
Readiness Subcommittee
Committee on Armed Services
United States Senate


OBJECTIVES, SCOPE, AND METHODOLOGY
=========================================================== Appendix I

We evaluated whether the federal government could realize savings\1
if lower prices in the electricity market result from retail
competition, as is expected.  In making this savings estimate, we
developed a baseline representing the projections of government
spending for electricity purchased without retail competition.  We
subtracted from this baseline the projections of spending for
electricity that are based on these same quantities but with lower
electricity prices resulting from retail competition.  In addition,
we estimated the government's decline in spending, which is less than
its savings because the quantities purchased will increase further
than in the baseline as competitive prices fall below baseline
prices.  In estimating savings and the decline in spending, we did
not develop our own model to estimate the extent to which retail
competition is expected to reduce electricity prices for retail
customers.  Instead, we applied the results of several published
forecasts to our model of the government's projected spending for
electricity.  In addition, our analyses apply only to the effects of
retail competition on the generation of electricity--transmission and
distribution markets for electricity would likely still be regulated. 

We note, however, that our savings and spending decline estimates
could be below or above our estimated ranges and are subject to major
imprecision because of, among other things, the substantial
uncertainty over the (1) future structure and pace of implementing
retail competition in the United States, (2) quantities of
electricity that the government would purchase with and without
retail competition, and (3) prices that the government would pay with
and without retail competition.  In addition, in making these
estimates, we calculated ranges because they could more reliably
present the uncertainty over these estimates.  Moreover, we note that
the secondary effects--that is, the ripple effects of lower
electricity prices on the economy--could be substantial and result in
even greater savings and declines in spending for the federal
government. 

In estimating the range of possible savings to the federal
government, we (1) developed a baseline representing the projections
of the federal government's annual demand (both the prices paid and
the quantities consumed) for electricity under existing
circumstances, that is, without prevailing retail competition; (2)
subtracted from the baseline the projections of lower government
spending for electricity on baseline quantities, but with the lower
prices that could result from retail competition; and (3) discounted
the resulting stream of future annual dollar savings over the 18-year
period 1998-2015 to determine the 1996 present values.  To determine
the range of possible declines in spending for the federal
government, we used the above procedure for estimating savings, but
with the projections of lower government spending for electricity as
a result of retail competition based on larger quantities than in the
baseline.  That is, the government could be thought of as "spending"
some of its savings to buy more electricity. 

To construct the baseline, we (1) revised the 1995 data on actual
federal electricity expenditures obtained from the Department of
Energy's (DOE) Office of Federal Energy Management Programs (FEMP)\2
to reflect only that portion of the data affected by retail
competition and (2) projected future government prices on the basis
of prices for the AEO97 reference case contained in DOE's Energy
Information Administration (EIA) report entitled Annual Energy
Outlook 1997 With Projections To 2015.\3 We did not assess the
reliability of the computer data or verify the accuracy of the FEMP
data and the basis for the EIA projections used to construct the
baseline.  However, we did discuss FEMP's procedures for validating
the data submitted by the agencies. 

To identify possible savings and the declines in spending associated
with the competitive purchase of electricity, we used the price
declines projected by DRI/McGraw-Hill (DRI),\4

the Gas Research Institute (GRI),\5 and the Citizens For A Sound
Economy Foundation (CSE).\6 We used these competitive pricing
projections because they were the only ones we were able to identify
that considered the effects of retail competition for all consumers
and were not limited to a particular geographic region.  We reviewed
each study and contacted officials at these organizations to further
discuss the results of their analyses.  However, we did not verify
the accuracy of, or the basis for, these projections, which we used
to determine declines in electricity prices resulting from retail
competition. 

In addition to the above studies, we reviewed the following studies
to identify other possible sources to use for our cost savings
analysis:  a March 1996 Department of Defense report to the Congress,
Procurement of Electricity From Most Economical Source; a January
1997 report by the Heritage Foundation, Energizing America:  A
Blueprint For Deregulating The Electricity Market; a November 1996
report by the Congressional Research Service, Electricity:  The Road
Toward Restructuring; and a March 1997 DOE presentation on
"Electricity Prices in a Restructured Electric Power Industry."
Although time constraints did not allow us to include the results of
price declines from retail competition estimated by EIA in its August
1997 report, Electricity Prices in a Competitive Environment: 
Marginal Cost Pricing of Generation Services and Financial Status of
Electric Utilities--A Preliminary Analysis Through 2015, we did
review the report. 


--------------------
\1 As used in this report, the term "savings" does not represent the
budgetary savings that might be realized annually in the federal
budget process from lower electricity costs. 

\2 See Energy Consumption:  Federal Agencies' Electricity Use and
Cost (GAO/RCED-97-97R, Mar.  21, 1997).  We provided FEMP's data on
usage and cost across federal agencies.  For this report, we revised
the fiscal year 1995 data to reflect only that portion that would be
affected by retail competition and converted the data to constant
1996 dollars. 

\3 The EIA reference case estimates average, residential, commercial,
and industrial electricity prices under the assumption of "limited
competition." It assumes (1) competitive pressures from the wholesale
electricity markets and (2) the more limited "anticipation effects,"
but not the actual implementation, of full retail competition. 

\4 DRI/McGraw-Hill, World Energy Service:  U.S.  Outlook (Spring
1997). 

\5 Gas Research Institute, GRI Baseline Projection of U.S.  Energy
Supply and Demand To 2015, 1997 Edition (Mar.  1997).  We note that
GRI's price projection incorporates only some of the details of
retail restructuring--those aspects that could be dealt with through
their model assumptions.  GRI did not deal with those aspects that
required a more involved effort, such as methodology updates. 
Therefore, GRI's electricity price declines can be considered lower
than what would otherwise be expected if GRI's projections had
assumed full retail restructuring. 

\6 Michael T.  Maloney and Robert E.  McCormick with Raymond D. 
Sauer, Customer Choice, Consumer Value:  An Analysis of Retail
Competition in America's Electric Industry.  Prepared for and
published by Citizens For A Sound Economy Foundation (Washington,
D.C.:  1996). 


ANALYSES OF SAVINGS AND OF THE
DECLINE IN THE GOVERNMENT'S
SPENDING FOR ELECTRICITY
========================================================== Appendix II

This appendix provides the detailed results of our analyses to
provide a range of estimates of savings to the federal government
that might occur if the federal government had the ability to procure
its electricity competitively.  We note that our analyses apply only
to the generation of electricity.  Transmission and distribution
markets for electricity would likely still be regulated. 
Specifically, we estimated the (1) savings that would likely occur
with lower prices resulting from retail competition to purchase our
baseline quantities, that is, the quantities the government would
likely have purchased without retail competition and (2) decline in
the government's spending that would likely occur with lower prices
resulting from retail competition to purchase our baseline quantities
plus the additional quantities above our baseline because competitive
prices decline more than baseline (noncompetitive) prices.\1 We note
that under both the competitive and noncompetitive scenarios,
electricity prices would fall, but the decline would be substantially
more under retail competition. 

In estimating savings and the decline in the government's spending,
we did not develop our own model to estimate lower prices from retail
competition but used the results of three expert studies and applied
their results to project usage and prices for the government.  The
three studies used were (1) DRI/McGraw-Hill (DRI),\2 (2) GRI,\3 and
(3) the Citizens For A Sound Economy Foundation (CSE).\4 We note,
however, that any estimates of savings and the decline in the
government's spending resulting from retail competition are subject
to major imprecision because of, among other things, the high level
of uncertainty surrounding the future structure and pace of
implementing retail competition in the U.S.  electricity industry,
the future quantities of and prices for electricity purchased, and
the extent to which retail competition reduces electricity prices for
the government in particular. 

We estimated for the 18-year period 1998 through 2015, in terms of
1996 present values, (1) the savings that would likely occur from
lower electricity prices, because of the implementation of full
retail competition, for the purchase of the baseline quantities of
electricity under noncompetitive circumstances, that is, under
"existing circumstances" and (2) the decline in the government's
spending that would likely occur from lower prices because of retail
competition, taking into account the baseline quantities purchased
plus the purchase of additional quantities of electricity, which
would likely result from competitive prices falling below baseline
prices.  By definition, the decline in the government's spending
would be less than its savings because the quantities purchased for
the competitive scenario are larger.  We have defined both our
savings and the decline in spending as relative to a baseline for
prices, quantities, and dollar spending.  This allows us to estimate
how much less the federal government would spend than it would
otherwise have spent over future years--isolating the effect of
retail competition on spending.\5

Specifically, for any future year, our estimated savings is defined
as the noncompetitive price minus the competitive price, multiplied
by the quantity of electricity that would likely be purchased under
the noncompetitive price scenario.  The decline in the government's
spending is defined as the noncompetitive price multiplied by the
quantity of electricity that would likely be purchased under the
noncompetitive price scenario, minus the competitive price multiplied
by the quantity of electricity that would likely be purchased under
the competitive price scenario.  For either measure, we discounted
each of the resulting streams of future savings and declines in
spending over these years, and summed each of them over the period
1998-2015 to obtain their 1996 present values, that is, their values
in discounted 1996 constant dollars.\6


--------------------
\1 As the term noncompetitive is used here, we mean the electricity
generation charges under the "existing circumstances" scenario.  That
is, retail electricity prices are regulated, but cost reductions are
assumed to result from competitive pressures in the wholesale market
for electricity, as well as from suppliers' preparations for retail
competition. 

\2 DRI/McGraw-Hill, World Energy Service:  U.S.  Outlook (Spring
1997). 

\3 Gas Research Institute, GRI Baseline Projection of U.S.  Energy
Supply and Demand To 2015, 1997 Edition (Mar.  1997).  We note that
GRI's price projection incorporates only some of the details of
retail restructuring--those aspects that could be dealt with through
their model assumptions.  GRI did not deal with those aspects that
require a more involved effort, such as methodology updates. 
Therefore, GRI's electricity price declines can be considered lower
than what would otherwise be expected if GRI's projections had
assumed full retail restructuring. 

\4 Michael T.  Maloney and Robert E.  McCormick with Raymond D. 
Sauer, Customer Choice, Consumer Value:  An Analysis of Retail
Competition in America's Electric Industry.  Prepared for and
published by Citizens For A Sound Economy Foundation (Washington,
D.C.:  1996). 

\5 It is also possible to define both savings and the decline in the
government's spending as relative to a base year (e.g., 1997) price
rather than to baseline prices.  However, we did not choose these
definitions because they would include increased savings and declines
in the government's spending that are not attributable to retail
competition.  This is because these increases would have occurred
even without retail competition. 

\6 We use a real (i.e., inflation-adjusted) discount rate of 3.8
percent.  This rate is based on a 30-year federal government bond
nominal rate of 6.5 percent (the approximate rate when our analysis
was performed), minus a forecast average annual inflation rate of
about 2.7 percent over the period 1998 through 2015.  The inflation
rate used was the annual percentage change in the gross domestic
product implicit deflator from WEFA's first quarter 1997 forecast in
its U.S.  Long-Term Economic Outlook [Revised] (Vol.  1).  A real,
rather than a nominal, discount rate is used because our data are
already in 1996 constant dollars. 


   FEDERAL GOVERNMENT USAGE AND
   AVERAGE RATES IN 1995
-------------------------------------------------------- Appendix II:1

Using FEMP's fiscal year 1995 data on actual federal electricity
usage and cost,\7 we adjusted the data to reflect only that portion
that would be affected by retail competition.  For example, according
to the Department of Defense's (DOD) records, total spending for
electricity in fiscal year 1996 includes nondomestic spending of
about 25.5 percent of its total spending and electricity usage of
about 12.6 percent of its total kilowatt hours (kWh).\8 Because
spending on electricity used at facilities outside the United States
would not be affected by the restructuring of the electricity
industry in the United States, we subtracted this 25.5-percent share
from the total 1995 DOD spending; we made a similar adjustment of
12.6 percent for 1995 kWh usage in fiscal year 1995.  In addition, we
made other appropriate adjustments to FEMP's fiscal year 1995
spending and kWh usage data for other agencies.  We converted these
adjusted dollar figures to reflect constant 1996 dollars.  We then
calculated the 1995 average price--that is, the rate in cents per kWh
of electricity used by the federal government--by dividing the
government's total spending on electricity by its total kWh. 


--------------------
\7 FEMP coordinates federal energy efficiency efforts and reports
annually to the Congress on federal energy consumption and
conservation activities by executive branch agencies.  For fiscal
year 1995, FEMP's data were for 28 executive branch agencies and did
not include the judicial and legislative branch agencies.  According
to the data provided by legislative branch officials, the legislative
and judicial branches' usage is less than 1 percent of the
government's electricity usage. 

\8 Fiscal year 1996 was the first year DOD collected data on
nondomestic usage. 


   FEDERAL GOVERNMENT KWH USAGE
   OVER 1998-2015
-------------------------------------------------------- Appendix II:2

The extent to which the federal government's kWh usage of electricity
will increase or decrease in future years depends upon its future
price as well as upon non-price factors affecting demand.  As with
any commodity, as its price decreases (increases), the quantity
purchased will increase (decrease), all other factors affecting
demand remaining constant.  Non-price factors that might increase
future demand are, for example, the increased use of computers. 
Non-price factors that might decrease future demand are, for example,
the increased use of energy conservation measures and the downsizing
of the federal government.  In our analyses, we treat the price
effect on quantity purchased separately from the non-price effect on
quantity.  For example, electricity usage in kWh could be increasing
because of price reductions yet simultaneously decreasing because of
the combined effect of all non-price factors--leading to a net
positive or negative overall effect on quantity purchased. 

We assume that the quantity of kWh used will increase over the level
in the base year (1995) because of the price effects--price is
assumed to fall under both the competitive and noncompetitive
scenarios.  Quantity increases more under the competitive scenario
because the price falls more.  In either scenario, the amount of the
increase depends on the estimated demand elasticity for electricity
with respect to electricity price.\9 For the quantity effects,
because of the non-price factors, we assume a range of three
alternatives:  (1) an increase of 1 percent per year, (2) a zero
percent increase per year (no effect), and (3) a decrease of 1
percent per year.  The net effect on total kWh usage is determined by
the combined effects of both price and non-price factors. 


--------------------
\9 This demand elasticity is defined as the percentage change in the
quantity demanded divided by the percentage change in its own price,
all other factors affecting demand held constant.  Because quantity
rises as price falls, and vice versa, this elasticity is of a
negative sign.  Note, however, that higher (lower) elasticity values
represent elasticities that are higher (lower) in terms of absolute
value. 


   FEDERAL GOVERNMENT ELECTRICITY
   PRICES OVER 1998-2015
-------------------------------------------------------- Appendix II:3

We did not develop our own model to estimate electricity prices for
the federal government in 1998 through 2015 under either the
noncompetitive or competitive scenarios.  For the noncompetitive
scenario, we used EIA's AEO97 reference case, which estimates
average, residential, commercial, and industrial electricity prices
under the assumption of "limited competition." It assumes (1)
competitive pressures from the wholesale electricity markets and (2)
supplier preparation for, but not the actual implementation of,
retail competition.  We believe that this scenario best reflects the
extension of the current situation into the future, without any
further action by either the states or the federal government with
respect to electricity markets.  We assumed that the federal
government's average base price in 1995 would decline over the
forecast period at the same percentage rates as did the EIA forecast
prices for the average, residential, commercial, or industrial
classes, respectively. 

For the competitive scenario, we used the average, residential,
commercial, and industrial electricity price estimates from three
alternative studies:  (1) CSE,\10 (2) DRI, and (3) GRI.  We assumed
that the federal government's average base price in 1995 would
decline over the 1996-2015 period at the same percentage rates as did
the CSE, DRI, and GRI forecast prices.  The CSE study assumes
immediate full retail competition with no recouping of stranded costs
by the utilities.  The DRI study estimates electricity prices under
the assumption of reaching full retail competition in the electricity
markets by 2001 and assumes substantial amounts of stranded cost
recovery.  The GRI study assumes only some of the effects of retail
restructuring but also assumes substantial amounts of stranded cost
recovery.\11 Therefore, CSE's forecast prices are generally much
lower than those of either DRI or GRI. 


--------------------
\10 For CSE, we analyzed the electricity price declines for the
"average" customer class only. 

\11 Stranded costs are investments or assets owned by regulated
electric utilities that are not likely to be competitive in a
restructured marketplace.  For example, some utilities own nuclear
power plants that have very high operating costs, and these plants
are not likely to be competitive in the new retail market.  In
addition, such power plants often have high debt levels.  The
estimates of these stranded costs vary widely, from a low of roughly
$10 billion to a high of about $500 billion.  DRI estimates stranded
costs at about $247 billion, of which at least 90 percent are
recovered from ratepayers through a non-bypassable transmission
access charge and a rate surcharge.  GRI estimates stranded costs at
$100 billion, of which 80 percent is recovered from ratepayers
through some type of transmission charge. 


   SAVINGS AND THE DECLINE IN THE
   GOVERNMENT'S SPENDING: 
   SIMULATIONS--A RANGE OF
   ESTIMATES
-------------------------------------------------------- Appendix II:4

The estimated values for savings and the declines in spending for
federal electricity will vary, depending upon a number of
assumptions, any one of which may prove in the future to have been
inaccurate.  We performed simulation analyses to account for at least
some of the potential sources of variation and to estimate how much
such changes in our assumptions will affect our estimates of savings
and declines in spending.  Our simulations account for four major
causes of variation in our estimates of savings and declines in
spending.  First, to account for variations in competitive prices, we
used the estimates of competitive electricity prices from three
sources--DRI, GRI, and CSE.  Each source uses (explicitly or
implicitly) a different model of the U.S.  economy, with different
equations and assumed (or model-derived) forecast values for such
variables as real gross domestic product, inflation, labor force and
productivity growth, interest rates, and perhaps most importantly,
natural gas, coal, and petroleum prices. 

Second, to account for differences in prices for different types of
consumers, we used the percentage declines in the price forecasts for
average, residential, commercial, and industrial customers,
respectively.  Which category the federal government most resembles
is subject to debate.  Third, to account for the effects of non-price
factors on kWh used by the federal government, we assumed a growth
range represented by three cases:  (1) an annual growth rate of 1
percent, (2) zero annual growth, and (3) an annual decline rate of 1
percent.\12 With these different assumptions, over the 18-year period
from 1998 through 2015, this total kWh change because of non-price
factors would range from an increase of 19.6 percent to a decrease of
16.5 percent. 

Finally, we needed to account for price elasticity.  The estimates
for elasticity are also subject to much debate.  For the short
short-run (assumed to be reached in the year 2000) and the long
short-run (assumed to be reached in 2005), for the customer class
"average," we used elasticity estimates of -0.05 and -0.15,
respectively--elasticity numbers suggested as reasonable by an EIA
analyst.  For the long-run (assumed to be reached in 2013), for the
customer class "average," we used CSE's long-run elasticity estimate
of -0.976.  We also used CSE's long-run elasticity estimates for the
residential, commercial, and industrial customer classes, -0.795,
-0.450, and -1.702, respectively. 

The elasticity estimates for the three customer classes cited above
for the two short-run cases were adjusted up or down from the
short-run elasticity estimates for the average customer class by the
percentage that their long-run elasticity estimates were above or
below the corresponding long-run estimate for the average customer
class.  Additional simulations that we could have performed include
using a series of higher (or lower) sets of elasticity estimates. 
All other factors being held equal, such simulations would result in
higher (or lower) estimated savings.\13 For the decline in the
government's spending, the effect is more complex.  If the
elasticities were increased, the spending decline would decrease.  If
the elasticities were decreased, the spending decline would
increase.\14

We also could have performed many other mix-and-match simulations
using average, residential, commercial, or industrial price behavior
(percentage declines) with average, residential, commercial, or
industrial price elasticities--for a total of 16 different
simulations.  For example, we believe that a reasonable scenario
would be to use the percentage changes in industrial prices with the
commercial class elasticities.  This is because (1) the government's
average base price is already fairly low compared with the average
price for retail consumers--the industrial class has the lowest
prices--and (2) the government's price elasticities may be low--the
commercial class has the lowest elasticities. 


--------------------
\12 To select possible growth ranges to use in our analyses, we used
actual growth rates for the federal government over the last 10 years
and last 5 years, respectively.  According to our March 1997 report
entitled, Energy Consumption:  Federal Agencies' Electricity Use and
Cost, (GAO/RCED-97-97R, Mar.  21, 1997), the annual growth rate was a
positive 1 percent over the 10-year fiscal year period 1986-95, and
over the last 5 years, a negative 1 percent. 

\13 As elasticity increases (decreases), savings increase (decrease)
because of the rise (fall) in the quantity increase that results from
the decline in noncompetitive prices.  A larger (smaller) quantity
increase results from higher (lower) estimates of elasticity.  These
larger (smaller) increases are multiplied by the difference between
the noncompetitive and competitive prices. 

\14 The federal government's decline in spending can be divided into
three components.  The first two components are the difference
between noncompetitive and competitive prices multiplied by,
respectively, the base year quantity and the increase in quantity
because of non-price factors.  The third component is the
noncompetitive price multiplied by the increase in quantity because
of the decline in noncompetitive prices, minus the competitive price
multiplied by the increase in quantity because of the decline in
competitive prices.
The sum of the first two components is always positive over the
1998-2015 period since competitive prices are always less than
noncompetitive prices.  The estimate of elasticity is not included in
either of these two components; the estimate of elasticity is
included as a multiplicand in the third component.  However, the
third component is generally negative.  Although competitive prices
are generally much lower than noncompetitive prices, the increase in
quantity because of the decline in competitive prices (compared with
its comparable increase because of the decline in noncompetitive
prices) is, on a percentage basis, very much larger than is the
percentage decrease in competitive price below noncompetitive price.
For example, according to our analysis of the CSE case, in 1998
through 2015, the federal government would purchase about 123 billion
kWh more electricity than it would have purchased had its usage
remained at the 1995 base year level of about 50 billion kWh over the
period.  This increase is about 260 percent more than the about 34
billion kWh additional electricity purchased over the period in our
baseline compared with the government's usage had that usage remained
at the 1995 base year level.  This percentage greatly exceeds the
percentage decline in competitive prices below baseline prices,
resulting in a negative value for the third component.
We then multiply the increase (decrease) factor in the elasticity
estimate by this negative-value component--for example, by a factor
of 1.5, if elasticity increases by 50 percent and by 0.5, if
elasticity decreases by 50 percent.  Therefore, if elasticity
increases (decreases), the decline in spending decreases (increases). 


   SAVINGS AND THE DECLINE IN THE
   GOVERNMENT'S SPENDING: 
   ESTIMATES
-------------------------------------------------------- Appendix II:5

Tables II.1, II.2, and II.3\15 show the cumulative savings estimates,
reflecting the kWh usage assumed in our baseline, that is, the
noncompetitive price scenario.  The estimates for the present value
of savings in 1998 through 2015 range from (1) $1.0 billion for the
GRI industrial projection, assuming a one-percent yearly decline in
usage from non-price factors (table II.3) to (2) $8.2 billion for the
CSE average projection, assuming a one-percent yearly increase in
usage from non-price factors (table II.1).  The average annual values
over 1998 through 2015 range from a low of about $60 million to a
high of about $460 million. 



                               Table II.1
                
                   Savings on Electricity Spending If
                   Fiscal Year 1995 Usage Grows by 1
                 Percent Annually Because of Non-Price
                           Factors, 1998-2015

                         (Dollars in billions)

                                                         Present value
                                                      (discounted 1996
Studies' estimates for each type of class                     dollars)
--------------------------------------------------  ------------------
DRI (average)                                                     $5.5
DRI (residential)                                                  6.1
DRI (commercial)                                                   6.4
DRI (industrial)                                                   2.9
GRI (average)                                                      2.3
GRI (residential)                                                  3.1
GRI (commercial)                                                   2.4
GRI (industrial)                                                   1.3
CSE (average)                                                      8.2
----------------------------------------------------------------------
Note:  Includes quantity increases because of the elasticity effect
of lower prices. 



                               Table II.2
                
                   Savings on Electricity Spending If
                  Fiscal Year 1995 Usage Grows by Zero
                 Percent Annually Because of Non-Price
                           Factors, 1998-2015

                         (Dollars in billions)

                                                         Present value
                                                      (discounted 1996
Studies' estimates for each type of class                     dollars)
--------------------------------------------------  ------------------
DRI (average)                                                     $4.9
DRI (residential)                                                  5.5
DRI (commercial)                                                   5.7
DRI (industrial)                                                   2.6
GRI (average)                                                      2.1
GRI (residential)                                                  2.7
GRI (commercial)                                                   2.1
GRI (industrial)                                                   1.1
CSE (average)                                                      7.3
----------------------------------------------------------------------
Note:  Includes quantity increases because of the elasticity effect
of lower prices. 



                               Table II.3
                
                   Savings on Electricity Spending If
                  Fiscal Year 1995 Usage Declines by 1
                 Percent Annually Because of Non-Price
                           Factors, 1998-2015

                         (Dollars in billions)

                                                         Present value
                                                      (discounted 1996
Studies' estimates for each type of class                     dollars)
--------------------------------------------------  ------------------
DRI (average)                                                     $4.3
DRI (residential)                                                  4.8
DRI (commercial)                                                   5.0
DRI (industrial)                                                   2.3
GRI (average)                                                      1.8
GRI (residential)                                                  2.4
GRI (commercial)                                                   1.9
GRI (industrial)                                                 1.0\a
CSE (average)                                                      6.4
----------------------------------------------------------------------
\a $1.01 billion, rounded to two decimal places.\

Note:  Includes quantity increases because of the elasticity effect
of lower prices. 

Tables II.4, II.5, and II.6\16 show the estimates for the federal
government's cumulative decline in spending, reflecting the two kWh
usages assumed in our baseline and competitive price scenarios.  The
estimates for the present value of spending declines in 1998 through
2015 range from (1) $0.6 billion for the GRI industrial projection,
assuming a 1-percent yearly decline in usage from non-price factors
(table II.6), to (2) $6.5 billion for the CSE average projection,
assuming a 1-percent yearly increase in usage from non-price factors
(table II.4).  The average annual values over 1998 through 2015 range
from a low of about $30 million to a high of about $360 million. 



                               Table II.4
                
                  Declines in Electricity Spending If
                   Fiscal Year 1995 Usage Grows by 1
                 Percent Annually Because of Non-Price
                           Factors, 1998-2015

                         (Dollars in billions)

                                                         Present value
                                                      (discounted 1996
Studies' estimates for each type of class                     dollars)
--------------------------------------------------  ------------------
DRI (average)                                                     $4.3
DRI (residential)                                                  5.0
DRI (commercial)                                                   5.7
DRI (industrial)                                                   2.0
GRI (average)                                                      1.7
GRI (residential)                                                  2.4
GRI (commercial)                                                   2.1
GRI (industrial)                                                   0.8
CSE (average)                                                      6.5
----------------------------------------------------------------------
Note:  Includes quantity increases because of the elasticity effect
of lower prices. 



                               Table II.5
                
                  Declines in Electricity Spending If
                  Fiscal Year 1995 Usage Grows by Zero
                 Percent Annually Because of Non-Price
                           Factors, 1998-2015

                         (Dollars in billions)

                                                         Present value
                                                      (discounted 1996
Studies' estimates for each type of class                     dollars)
--------------------------------------------------  ------------------
DRI (average)                                                     $3.7
DRI (residential)                                                  4.3
DRI (commercial)                                                   5.0
DRI (industrial)                                                   1.7
GRI (average)                                                      1.4
GRI (residential)                                                  2.1
GRI (commercial)                                                   1.8
GRI (industrial)                                                   0.7
CSE (average)                                                      5.5
----------------------------------------------------------------------
Note:  Includes quantity increases because of the elasticity effect
of lower prices. 



                               Table II.6
                
                  Declines in Electricity Spending If
                  Fiscal Year 1995 Usage Declines by 1
                 Percent Annually Because of Non-Price
                           Factors, 1998-2015

                         (Dollars in billions)

                                                         Present value
                                                      (discounted 1996
Studies' estimates for each type of class                     dollars)
--------------------------------------------------  ------------------
DRI (average)                                                     $3.1
DRI (residential)                                                  3.7
DRI (commercial)                                                   4.4
DRI (industrial)                                                   1.5
GRI (average)                                                      1.2
GRI (residential)                                                  1.7
GRI (commercial)                                                   1.6
GRI (industrial)                                                   0.6
CSE (average)                                                      4.7
----------------------------------------------------------------------
Note:  Includes quantity increases because of the elasticity effect
of lower prices. 

For either savings or the government's decline in spending, tables
II.1 through II.6 show that the estimates of CSE are larger than
those of DRI, which in turn are larger than those of GRI.  For
example, in table II.1, the savings estimates for the CSE average
customer are $8.2 billion; for the DRI average customer, $5.5
billion; and for the GRI average customer, $2.3 billion.  The lower
savings estimate for GRI because of its higher price projections
reflects, in part, the fact that its modeling methodology did not
incorporate all of the aspects of full retail restructuring in the
electricity markets. 

In addition, tables II.1 through II.6 show that the industrial price
behavior assumption yields the smallest savings or decline in
spending estimates compared with the estimates for the average,
residential, and commercial customer classes.  Although the
industrial price elasticity assumptions were the highest, the
percentage decreases in industrial prices were much less than those
for the other customer classes.  For example, for DRI, in table II.1,
the savings estimates decrease from $6.4 billion for the commercial
class to a low of $2.9 billion for the industrial class.  The latter
result reflects the much smaller percentage declines in forecast
prices for the industrial customer. 

Tables II.1, II.2, and II.3, for savings, reflect the 1-percent,
zero-percent, and minus 1-percent, respectively, annual growth rates
in the quantity of electricity used because of non-price factors. 
Tables II.4, II.5, and II.6, for the decline in spending, reflect the
1-percent, zero-percent, and minus 1-percent, respectively, annual
growth rates in the quantity of electricity used because of non-price
factors.  As expected, the savings and the decline in spending
decrease as the quantity increase is lowered because of non-price
factors.  For example, for the DRI average customer, in tables II.1
through II.3, the estimates of savings in 1998 through 2015 fall from
$5.5 billion (with 1-percent annual growth because of non-price
factors) to $4.3 billion (with 1-percent annual decline because of
non-price factors).  Thus, the less that electricity usage expands
from non-price factors, the less will be the savings resulting from
the implementation of full retail competition.  For the estimates of
the decline in spending, shown in tables II.4 through II.6, the
effect is the same, with these spending declines falling from $4.3
billion to $3.1 billion. 

If the assumed elasticity estimates for the federal government were
to decrease by 50 percent, the savings estimates would be smaller,
all other factors held constant.  For example, the lowest end of our
range, $1.01 billion (table II.3) for savings decreases to $0.96
billion (a $0.05 billion decrease).  For the decline in spending, the
lowest end of our range, $0.6 billion (table II.6) increases to $0.7
billion (a $0.1 billion increase). 

If we assume, as a mix-and-match simulation, the industrial sector
price reductions, but with the elasticity estimates for the
commercial sector that are used in tables II.1 through II.6, our
savings estimates are lower because commercial sector elasticities
are lower than those of the industrial sector.  The estimates for the
declines in spending are higher because the elasticities for the
commercial sector are lower.  The small percentage declines in
industrial prices keep both savings and the decline in spending low. 
For example, for savings, using GRI's forecast for industrial prices
and a 1-percent annual decline in usage because of non-price factors,
savings are $1.0 billion in table II.3.  This number falls to $0.9
billion (a $0.1 billion decrease) using the commercial sector
elasticities.  For the decline in spending, using GRI's forecast for
industrial prices and a 1-percent annual decline in usage because of
non-price factors, the decline is $0.6 billion in table II.6.  This
number rises to $0.8 billion (approximately an $0.2 billion increase)
using the commercial sector elasticities. 



(See figure in printed edition.)Appendix III

--------------------
\15 All savings estimates are given in discounted 1996 constant
dollars (billions).  The estimates for present value reflect the
discounting of the savings in 1998 through 2015 using a real discount
rate of 3.8 percent. 

\16 All declines in spending estimates are given in discounted 1996
constant dollars (billions).  The estimates for present value reflect
the discounting of the declines in spending in 1998 through 2015
using a real discount rate of 3.8 percent. 


COMMENTS FROM THE DEPARTMENT OF
ENERGY
========================================================== Appendix II



(See figure in printed edition.)


MAJOR CONTRIBUTORS TO THIS REPORT
========================================================== Appendix IV

RESOURCES, COMMUNITY, AND ECONOMIC
DEVELOPMENT DIVISION, WASHINGTON
D.C. 

Peg Reese
Jay R.  Cherlow
James M.  Kennedy
Michael S.  Sagalow
Daren K.  Sweeney
Daniel G.  Williams

OFFICE OF THE CHIEF ECONOMIST

Joseph Kile

OFFICE OF GENERAL COUNSEL

Susan W.  Irwin


*** End of document. ***