FTC Comment Before the Public Utility Commission for the Commonwealth of Pennsylvania Concerning the Energy Efficiency and Conservation Program

1 Commonwealth of Pennsylvania, Pennsylvania Public Utility Commission, Energy
Efficiency and Conservation Program and EDC Plans, Docket No. M-2008-2069887 (Nov. 14,
2008), available at http://www.puc.state.pa.us/PCDOCS/1026558.doc.
2 Richard J. Hudson, Jr., testimony on behalf of the Retail Energy Supply Association,
before the Pennsylvania Public Utility Commission, En Banc Hearing on “Alternative Energy,
Before the
Public Utility Commission
Commonwealth of Pennsylvania
Energy Efficiency and Conservation 	)
Program and EDC Plans 	)
Reply to Comments on the Staff’s 	) 	Docket No. M-2008-2069887
Implementation Plan for the EE&C 	)
Program of November 26, 2008 	)
Reply Comment of the
Federal Trade Commission
December 17, 2008
I. Summary
The Federal Trade Commission (“Commission” or “FTC”) appreciates this opportunity
to comment on efforts by the Pennsylvania Public Utility Commission (“PA PUC”) to increase
energy efficiency, conservation, and demand response. In part, this comment is a reply to
comments on the November 26, 2008, PA PUC staff proposals for an implementation plan for
the Energy Efficiency and Conservation Program (and to the plan itself).1 It also replies to the
November 18, 2008, presentation by the Retail Energy Supply Association (“RESA”) at the PA
PUC en banc hearing. 2 The PA PUC’s efforts to increase energy efficiency, conservation, and

Energy Conservation and Efficiency, and Demand Side Response” (Nov. 18, 2008), available at
http://www.puc.state.pa.us/electric/pdf/EnBanc-DSR/Ttmy-RESA111908.pdf.
2
demand response have the potential to enhance consumer welfare and increase economic
efficiency at both the wholesale and retail levels. More efficient pricing, advanced metering, and
improvements in the technology used to determine when (and how much) energy is consumed
are all critical to the future performance of the power industry in Pennsylvania and in the United
States as a whole. We commend the Pennsylvania Legislature and the PA PUC for taking
initiatives on these important topics.
Although many of the specific questions posed by the PA PUC at the en banc hearing
pertained to Conservation Service Providers, we agree with RESA that these questions must be
framed in the context of empowering customers to manage their peak and overall loads. We
encourage the PA PUC to center the Energy Efficiency and Conservation Program around
providing customers with incentives and opportunities to better manage those loads. This
comment describes several aspects of encouraging energy efficiency and conservation as a
means to deliver consumer benefits. We believe that addressing the topics covered in this
comment, either in the short term or during subsequent policy reviews, will benefit
Pennsylvania’s electric power customers and U.S. power customers in general.
The FTC encourages the PA PUC to employ dynamic electric power pricing and demand
response to involve customers in addressing the power systems’ most pressing problems. Well-
designed dynamic pricing and demand response programs can enlist customers to help meet
important challenges facing the power system by, for example:
• 	managing peak load;

3
• 	ensuring that load never exceeds generation;
• 	keeping system costs down;
• 	reducing the need for ratepayers to pay to build, maintain, and operate peak-load
generating facilities;
• 	pricing pollution into the time-varying cost of power, thereby encouraging
customers to shift power demand to periods when the marginal plant is a low-
cost, low-emissions generator (and away from periods when a less attractive plant
is on the margin);
• 	making good use of the wind and solar generators being installed in response to
environmental concerns, market forces, and Pennsylvania’s renewable portfolio
standard;
• 	complementing unpredictable, intermittent wind and solar generators with
flexible demand rather than flexible supply, when flexible demand is more cost-
effective; and
• 	facilitating the appropriate investment in and use of technologies by end-users,
generators, and utilities, including plug-in vehicles and onsite generation when
they are cost-effective.
Automation, feedback, and incentives can enable consumers to become partners in
addressing all of these problems. The PA PUC has an important opportunity to give every
customer the option to sign up for real-time or critical peak pricing. The PA PUC also has an
opportunity to give utilities and Conservation Service Providers the incentives and mandate to
develop programs that are attractive to customers and that benefit customers who participate in

4
the programs, non-participating customers, utilities and the electric system. This requires firms
to develop communication strategies that help customers to make good choices about whether to
enroll and to respond appropriately to sound economic incentives. These programs can benefit
participating and non-participating customers if the programs have effective marketing, user-
friendly implementation, and structures that better align customers’ incentives to use power with
its cost to the grid. Early experience with dynamic pricing programs suggests that recruitment of
customers requires careful design and marketing; once enrolled, however, customers respond to
the new incentives and are pleased to save money and to gain a sense of better understanding and
controlling their energy use.
Developments in metering, feedback, and control technologies are creating new, cost-
effective opportunities for consumers to participate in managing load. At the same time, the
growing use of wind, solar, and nuclear generation is increasing the proportion of generators
over which dispatchers have limited control. These parallel developments in demand and supply
technology create significant additional opportunities to empower consumers to improve the
efficiency of the electric power system.
We encourage the PA PUC to continue to make a concerted, ongoing effort to support
programs that empower consumers to manage their loads and address the challenges facing
Pennsylvania’s electric system. The PA PUC’s role may involve encouraging firms to take
innovative approaches to empowering consumers, mitigating risks to firms during the
transformation, and updating regulations that stand in the way of progress. Dynamic pricing is a
compelling economic idea, but it requires an ongoing commitment to make it work for
regulators, utilities, and customers.

3 See, e.g., Deborah Platt Majoras, Chairman, Federal Trade Commission, Opening
Remarks at the FTC Conference on Energy Markets in the 21st Century: Competition Policy in
Perspective (Apr. 10, 2007), available at
http://www.ftc.gov/speeches/majoras/070410energyconferenceremarks.pdf. FTC merger cases
involving electric power markets have included DTE Energy/MCN Energy (2001) (consent
order), available at http://www.ftc.gov/os/2001/05/dtemcndo.pdf; and PacifiCorp/Peabody
Holding (1998) (consent agreement), available at
http://www.ftc.gov/os/1998/02/9710091.agr.htm. (The FTC subsequently withdrew the
PacifiCorp settlement when the seller accepted an alternative acquisition offer that did not pose a
threat to competition.)
4 FTC Staff Report, Competition and Consumer Protection Perspectives on Electric
Power Regulatory Reform: Focus on Retail Competition (Sept. 2001), available at
http://www.ftc.gov/reports/elec/electricityreport.pdf; FTC Staff Report, Competition and
Consumer Protection Perspectives on Electric Power Regulatory Reform (July 2000), available
at http://www.ftc.gov/be/v000009.htm (compiling previous comments that the FTC staff
provided to various state and federal agencies).
5
II. 	Interest of the Federal Trade Commission
The FTC is an independent agency of the federal government responsible for maintaining
competition and safeguarding the interests of consumers through enforcement of the antitrust and
consumer protection laws and through competition policy research and advocacy. The FTC
often analyzes regulatory or legislative proposals that may affect electric industry competition or
allocative efficiency. It reviews proposed mergers that involve electric and gas utility
companies. In the course of this work, as well as in antitrust and consumer protection research,
investigation, and litigation, the FTC applies established legal and economic principles and
recent developments in economic theory and empirical analysis.
The energy sector, including electric power, has been an important focus of the FTC’s
antitrust enforcement and competition advocacy.3 The FTC’s competition advocacy program has
produced two staff reports on electric power industry restructuring issues at the wholesale and
retail levels,4 and FTC staff also contributed to the work of the Electric Energy Market

5 Electric Energy Market Competition Task Force, Report to Congress on Competition in
Wholesale and Retail Markets for Electric Energy (2007), available at
http://www.ferc.gov/legal/fed-sta/ene-pol-act/epact-final-rpt.pdf.
6 The most recent FTC conference on energy issues was Energy Markets in the 21st
Century: Competition Policy in Perspective, held on April 10-12, 2007 (conference materials
available at http://www.ftc.gov/bcp/workshops/energymarkets/index.shtml). See also the FTC’s
public workshop on Market Power and Consumer Protection Policies Issues Involved with
Encouraging Competition in the U.S. Electric Industry, held on September 13-14, 1999
(workshop materials available at http://www.ftc.gov/bcp/elecworks/index.shtm); and the
Department of Justice and FTC Electricity Workshop, held on April 23, 1996.
7 See, e.g., Fed. Trade Comm’n, Comment Before the Federal Energy Regulatory
Commission on Wholesale Competition in Regions with Organized Electric Markets (Apr. 17,
2008), available at http://www.ftc.gov/be/v070014b.pdf.
FTC competition advocacy filings after mid-1994 are available in reverse chronological
order at http://www.ftc.gov/opp/advocacy_date.shtm. FTC competition advocacy efforts
regarding the electric power sector began in 1994 with a Comment of the Staff of the FTC
Bureau of Economics to the South Carolina Legislative Audit Council on the Statutes and
Regulations Covering the South Carolina Public Service Commission (Feb. 28, 1994).
8 The FTC and the Department of Justice participate as United States delegates in a
number of international organizations, such as the Organisation for Economic Co-operation and
Development. As part of this process, the FTC staff contributes to the United States’ “country
reports” on competition topics. See, e.g., United States Department of Justice and Federal Trade
Commission, “Note by the US Department of Justice and US Federal Trade Commission,”
OECD Roundtable on Energy Security and Competition Policy (Feb. 21-22, 2007), available at
http://www.ftc.gov/os/2007/02/WD200725OilGasUnited%20States.pdf. When requested by the
Department of State, the FTC staff also contributes to comments by the United States on
proposed regulatory reforms in other nations.
6
Competition Task Force, which issued a report to Congress in spring 2007. 5 The Commission
also has held public conferences on energy topics.6 The FTC and its staff have filed numerous
competition advocacy comments with FERC and the states concerning electricity restructuring
initiatives.7 The FTC staff also participates in preparing United States Government filings
before international competition organizations regarding energy policy matters. 8

9 Traditional retail rates generally are time-invariant, i.e., the rates do not depend on
when the customer uses power, regardless of scarcity conditions and wholesale prices. Some
traditional utility systems, however, include seasonal differences in rates that otherwise are time-
invariant. Over time, changes in average wholesale prices can (and do) lead to adjustments in
traditional retail rates. Such adjustments, however, involve substantial lags, and do not
contemporaneously track the daily cycles in generation and transmission costs that produce
fluctuations in wholesale prices.
7
III. Facilitating Demand Response and Dynamic Pricing, and Ending Regulatory Practices
That Can Inhibit Demand Response
By facilitating the implementation of demand response and dynamic pricing in
Pennsylvania, the PA PUC has an opportunity to reduce costs to ratepayers, increase reliability,
and ease the integration of intermittent wind and solar resources. We encourage the PA PUC to
seize this opportunity. Utilities may want to roll out dynamic pricing first as an attractive opt-in
program. Once such a program has been reviewed and refined, and after it has satisfied
thousands of customers for a few years, dynamic pricing could become the default option for all
new accounts (and perhaps could serve as an opt-out rate for existing customers of provider-of-
last-resort service). The widespread availability of advanced meters facilitates the low-cost
expansion of dynamic pricing programs.
The PA PUC may wish to coordinate with other regulators and market operators about
how the expansion of dynamic pricing and demand response should affect policies in the PJM
Interconnection and Midwest ISO wholesale power markets that supply Pennsylvania’s power.
This conversation might explore the future of market power mitigation and capacity market
policies that currently are used as substitutes for consumer participation.
A. Background
Traditional retail electric prices do not change contemporaneously when higher demand
leads to higher wholesale prices.9 This problem is particularly acute when generation or

10 During critical peak periods, wholesale prices sometimes reach $1000/MWh or more.
During off-peak periods, wholesale prices in some areas are as low as $20/MWh (or lower). For
example, the California Independent System Operator is seeking authority to institute a price
ceiling of $2500/MWh and a price floor of -$2500/MWh during the initial period of its revised
nodal pricing system. Simulations suggest that nodal prices may fall below the proposed floor or
exceed the proposed ceiling during extreme operating conditions. California Independent
System Operator, Market Redesign and Technical Upgrade Tariff Amendment to Adopt Price
Cap and Floor (filed Nov. 3, 2008), available at
http://www.caiso.com/2074/2074da39205e0.pdf.
8
transmission equipment problems and extreme weather conditions lead to excess demand for
electricity that cannot be supplied by the system, and also to far higher costs to produce,
transmit, and deliver electricity than normal. With time-invariant retail prices, customers pay the
same price to run their air conditioners and dryers during low-demand periods and high-demand
periods – two situations during which the real cost of having the marginal generator available
and operating, plus the cost of transmitting that power to the customer, may differ by a factor of
50 or more. 10 Fixed retail prices lead to a serious waste of resources. People who pay time-
invariant prices buy and use peak-period power that they might not be willing to purchase if they
had to pay the marginal cost of generating and transmitting that power. Moreover, the fact that
large increases in wholesale prices during high-demand periods do not lead to increases in retail
prices raises the likelihood of market power problems in wholesale electricity markets. If retail
prices – and therefore consumption – do not respond in real time to increases in wholesale
prices, generators are more likely to have market power, and to exercise it by raising wholesale
prices, than would be the case if retail prices increased (and trimmed consumption) when
wholesale prices increased.

11 Electric power’s social (opportunity) cost is the lesser of (1) the cost of building and
running an electricity system large enough to prevent the blackout or (2) the cost at which other
customers will reduce their consumption of electric power enough to prevent the blackout.
12 Dynamic pricing encourages owners of plug-in hybrid and all-electric vehicles to shift
the charging of their vehicles to periods in which low-cost power plants are on the margin.
9
Traditional, fixed retail pricing also increases the volatility of wholesale prices, increases
the risk of blackouts and brownouts, and raises the average costs of the electric power system.
For example, during an unusual heat wave that drove temperatures above 100 degrees in
Southern California in early September 2007, Southern California Edison reported that
approximately 20,000 customers were subject to extended blackouts. It was estimated that 90
percent of those blackouts were due to peak demand that exceeded the capacity of local
distribution equipment under such extreme temperatures. If retail prices had adjusted to reflect
wholesale prices in real time, people would not have used power that they valued at less than its
social cost. 11 As a result, there would have been a more efficient allocation of the limited
amount of electricity in the short term, and power system suppliers would have had stronger
incentives to build the efficient amount of generation, transmission, and distribution capacity.
Reliance on fixed retail prices poses major threats connected with two major new
technologies. One of these technologies is electric vehicles, of either the plug-in hybrid or the
all-electric variety. Time-invariant retail pricing does nothing to encourage plug-in hybrid
customers to recharge during off-peak hours. If large numbers of plug-in vehicles recharge
during peak demand periods, the power system will require the costly construction, maintenance,
and operation of peakers and additional transmission and distribution capacity in order to
maintain system reliability.12

13 Database of State Incentives for Renewables & Efficiency, Pennsylvania Incentives for
Renewable Energy, available at
http://www.dsireusa.org/library/includes/incentive2.cfm?Incentive_Code=PA06R&state=PA&C
urrentPageID=1.
10
The second major technology consists of wind and solar power and other forms of
intermittent renewable generation. Many states have prescribed a “renewable portfolio standard”
that requires a portion of generation to be from renewable energy sources. Pennsylvania’s
renewable portfolio standard requires that 18 percent of the state’s energy come from renewable
sources by 2020, including 0.5 percent from solar photovoltaic generation.13 Once a technology
such as wind or solar photovoltaic is in place, it is likely to be dispatched whenever it is
available because the marginal cost of wind and solar photovoltaic is close to zero. With fixed
retail prices, customers have no incentive to curtail their consumption when the wind dies down
or clouds roll in. A lack of demand response forces fossil-fuel generators with higher marginal
costs to produce more electricity. Without customer demand response, the costs of integrating
intermittent, albeit environmentally attractive, power sources into the power system may be
higher, as will the costs of reducing adverse environmental effects to mandated levels.
Unpredictable, intermittent wind and solar generators require flexible complements to balance
generation and load minute-by-minute. Dynamic pricing and demand response programs may
offer a flexible demand complement that is more cost-effective than flexible supply.
Dynamic pricing is a collection of approaches, including real-time pricing and critical
peak pricing, that allow retail prices to change on short notice in response to fluctuations in
wholesale prices. Real-time pricing sets one retail price for each hour (or a smaller unit of time,
such as quarter-hour or a five-minute segment) as a function of the spot market wholesale price.

14 Critical peak pricing schedules predetermined price periods and permits retail suppliers
to declare a limited number of critical periods that invoke the critical price. Utilities choose to
designate critical events when forecasted conditions are likely to cause electricity scarcity,
system unreliability, or high wholesale prices. Under the program, customers are notified about
critical events through automated phone calls, e-mails, or notification of a programmable
communicating thermostat. Although critical peak pricing notification policies vary, typically
customers are notified the day before the event that they need to adjust their thermostats by hand.
Shorter notice is possible if a customer has a programmable communicating thermostat or a
“gateway” system that automatically reduces his or her electricity consumption during critical
periods.
Utilities commit in advance to the number of critical hours or events. This limit typically
is about 1 percent of all hours, or about 15 events per year. Sometimes utilities commit to limits
on the timing of critical events. For example, California’s Statewide Pricing Pilot program
called critical events only between 2:00 pm and 7:00 pm on weekdays. Sometimes critical peak
pricing programs set forth conditions that will suffice to call a critical event. For instance,
temperatures below freezing or exceeding 95 degrees Fahrenheit suffice for Gulf Power to
trigger a critical event, while the California Statewide Pricing Pilot announced that any Stage 1
power emergency would trigger a critical event. Regarding the Gulf Power situation, see Gulf
Power Co., “GoodCents Select: Advanced Energy Management Program,” available at
http://www.ewh.ieee.org/r3/nwflorida/presentations/01_19_06.ppt; regarding California, see
Charles River Associates, “Impact Evaluation of the California Statewide Pricing Pilot” (Mar.
16, 2005), available at
http://www.energy.ca.gov/demandresponse/documents/group3_final_reports/2005-03-
24_SPP_FINAL_REP.PDF. In general, utilities also retain the flexibility to declare a critical
event on any day on which they forecast high power costs or low system reliability, so long as
such a declaration would not exceed the annual limit on the number of critical events they can
call.
11
Critical peak pricing is a simpler, dynamic (time-varying) pricing system. Typical critical peak
pricing programs define peak and off-peak periods and specify a peak price that is higher than
the off-peak price. These programs also allow utilities to designate about 1 percent of all hours
as critical scarcity periods, during which the price is significantly higher than during other (non-
critical) peak periods.14 This comment advocates dynamic pricing programs that increase
economic efficiency by making retail prices significantly better reflect the marginal cost of
power as it fluctuates over time.

15 The FERC Staff Report, Assessment of Demand Response and Advanced Metering,
FERC Docket AD-06-2-000 (Aug. 2006), available at http://www.ferc.gov/legal/staff-
reports/demand-response.pdf, also discussed a variety of regulatory barriers to demand response.
Experience and research have continued to develop in this area since the release of this FERC
Staff Report.
16 Owning an onsite generator can increase a customer’s responsiveness to prices by
enabling the customer to substitute self-generated power for power from the grid when the costs
of the latter exceed the costs of the former.
12
Five types of regulations are particularly likely to prevent or undermine demand
response, efficient pricing, and conservation.15 The first type includes regulations that forbid
time-varying retail pricing. Time-invariant pricing subsidizes consumption when power is most
costly and when increases in consumption are most likely to cause blackouts or other reliability
problems. A second type of unproductive regulatory action is approval of perfunctory dynamic
pricing schemes and implementation, which can lack measures to reduce the cost and risk of
participation and also can suffer from inadequate design, testing, and implementation of
communications, education, and marketing. These plans can be ineffective because they are
based on an oversimplified view of customers’ needs or suffer from underinvestment in
marketing. A third type comprises regulations and ratemaking systems that penalize utilities
financially if demand response increases between rate cases. The fourth type consists of
regulations that allow utilities to discourage efficient customer investment in onsite generation
by charging inefficiently high prices for standby service.16 The fifth type includes regulation that
deprives customers who want to offer demand response of the opportunity to customize their
offers.

17 State regulators, utilities, and researchers have gathered evidence about demand
response programs that are effective and consequently tend to reduce the need to buy power
when it is most expensive. Some demand response approaches, however, can entail significant
costs that must be compared to their benefits in any evaluation of their effectiveness. For
example, real-time prices encourage customers to reduce consumption during peak demand
periods and to invest in ways to respond efficiently to price fluctuations. Real-time pricing,
however, requires advanced meters that can be expensive. “Experiences in New York, Georgia,
California, and other states and pricing experiments have demonstrated that customers do take
actions to adjust their consumption, and are responsive to price (i.e., they have a nonzero price
elasticity of demand). Georgia Power Company’s successful real-time pricing tariff option has
demonstrated that industrial customers who receive real-time prices based on an hour-ahead
market are relatively price-responsive (price elasticities ranging from approximately -0.2 at
moderate price levels, to -0.28 at prices of $1/kWh or more) given the short-time period in which
to act. Among day-ahead real-time pricing customers, price elasticities range from
approximately -0.04 when prices are at moderate levels to -0.13 when customers are exposed to
higher prices. A critical peak-pricing experiment in California in 2004 determined that small
residential and commercial customers are price responsive and will produce significant
reductions. Participants reduced load 13 percent on average, and as much as 27 percent, when
price signals were coupled with automated controls such as controllable thermostats.” FERC
Staff Report, supra note 15, at 13-14 (footnotes omitted).
13
B. Removing Regulations that Prevent Retail Real-Time, Marginal-Cost Pricing of
Electric Power
The most direct way to increase the elasticity of wholesale demand with respect to
wholesale prices is to redesign retail rates to ensure that those rates better reflect the marginal
costs of supplying electricity to retail customers from the grid. These costs vary with changes in
aggregate and local demand and supply conditions, such as weather-driven demand for air
conditioning or generation and transmission equipment failures. A first step toward developing
retail prices that reflect marginal cost is to remove regulations that prohibit dynamic prices. One
way to achieve marginal-cost pricing of electric power is to base real-time retail prices on
wholesale market-clearing prices. 17 Retail pricing that tracks wholesale prices reduces or
eliminates the waste of resources that time-invariant pricing encourages, including generation of
electricity at a marginal cost that exceeds the value of the electricity to some of those who

18 Numerous power system simulations show that demand response lowers wholesale
prices during peak-demand periods. See, e.g., Severin Borenstein, “The Long-Run Efficiency of
Real-Time Electricity Pricing,” 26:3 Energy J. 93 (2005). Regarding reliability, ERCOT (the
system operator in most of Texas) reported that demand response avoided a blackout and
restored a frequency decline when a rapid, unanticipated decline in wind speeds caused an
associated decline in wind generation. This example illustrates that demand response can
substitute for generation-based reserves. Electric Reliability Council of Texas, “ERCOT
Demand Response Program Helps Restore Frequency Following Tuesday Evening Grid Event,”
ERCOT Press Release (Feb. 27, 2008), available at
http://www.ercot.com/news/press_releases/2008/pr_print_1_174210_174210.
19 Real-time metering already is extensively deployed for large commercial and industrial
customers in several states that allow customers to select their own electricity supplier. Several
pilot projects for residential customers have been completed or are underway, but mass
deployment of advanced meters for residential customers is rare in the United States. After an
extensive analysis of costs and benefits, the three major investor-owned utilities in California
have undertaken programs to deploy advanced meters for all classes of customers. Mass
deployments lower the average costs of the meters, reduce average installation costs, and yield
significant labor savings because the new meters report usage electronically and do not require
human meter readers to go door to door.
14
consume it during peak periods. Such retail pricing also would reduce the frequency and extent
of blackouts and brownouts, as well as required operating reserves. 18 In deciding which
approach to take to retail pricing, state regulators should consider the implementation and
transaction costs. Dynamic pricing’s transaction costs include the cost of having people monitor
and respond to price changes or the cost of computerized thermostats or appliance controllers to
do so. Implementation costs appear to be declining in the wake of technical advances in
metering and billing. 19
With dynamic pricing, some customers could see an increase in the variability of their
electric bills from month to month. Utilities, however, can offer billing plans under which
customers pay a constant amount each month in order to smooth out payments. Also with
dynamic pricing, customers may face a risk that their electric bills will increase unexpectedly

20 For a more extensive discussion of this subject, see Severin Borenstein, Center for the
Study of Energy Markets, Univ. of Cal. Energy Inst., Working Paper #155, “Customer Risk from
Real-Time Retail Electricity Pricing: Bill Volatility and Hedgability” (2006) (also published at
28:2 Energy J. 111 (2007)).
21 Gulf Power Company, “GoodCents Select: Advanced Energy Management Program,”
available at http://www.ewh.ieee.org/r3/nwflorida/presentations/01_19_06.ppt; Dan York and
Martin Kushler, “Exploring the Relationship Between Demand Response and Energy Efficiency:
A Review of Experience and Discussion of Key Issues,” American Council for an Energy-
Efficient Economy, Report No. U052 (Mar. 2005), available at
http://www.aceee.org/pubs/u052.pdf; Dan Merilatt (V.P. Program Development, GoodCents),
15
because of an unusual increase in market-wide demand or an unusual decrease in supply. In
theory, customers may be able to hedge (or buy insurance) against this in financial markets.
Regulators should consider whether customers are able to buy such insurance if they are willing
to pay its costs. For example, utilities could allow customers to have the option to buy blocks of
power in advance at prices that reflect the utilities’ costs of purchasing power in wholesale
markets; simultaneously, the utilities could purchase blocks of power in advance in wholesale
futures markets or enter into financial hedging contracts.20 A single program can both mitigate
bill volatility and offer good marginal incentives. Letting people prepay for a power
consumption schedule at its expected cost and any applicable risk premium reduces bill volatility
from transient, unexpected price changes. The pricing of deviations from the prepaid
consumption schedule at the dynamic price offers good marginal incentives.
So long as the social benefits of dynamic pricing exceed the costs of its implementation,
dynamic pricing should be the basic (or default) service option.
C. Successful Dynamic Pricing Implementation Requires Attention to Detail
Customers who opt for dynamic pricing programs – such as Gulf Power’s GoodCents
Select or Illinois’ EnergySmart Pricing – generally reduce their power consumption during peak
periods, save money, indicate satisfaction with the program, and continue participation. 21

“Demand Response Programs: New Considerations, Choices & Opportunities” (Jan. 2004),
available at http://www.enertouch.com/info/Demand%20Response%20Programs.pdf; Kathryn
Tholin, “Real-time Pricing for Illinois Consumers,” Center for Neighborhood
Technology/Community Energy Cooperative (Nov. 8, 2006), available at
http://peaklma.com/new%20folder/documents/tholin.ppt#256,1,Real-time Pricing for Illinois
Consumers.
22 Charles River Associates, “Impact Evaluation of the California Statewide Pricing
Pilot” (Mar. 16, 2005), available at
http://www.energy.ca.gov/demandresponse/documents/group3_final_reports/2005-03-
24_SPP_FINAL_REP.PDF; Ahmad Faruqui and Stephen George, “Quantifying Customer
Response to Dynamic Pricing,” 18:4 Electricity J. 53 (May 2005), available at
http://www.enertouch.com/info/Quantifying%20Customer%20Response.pdf.
23 Galen Barbose, Charles Goldman, and Bernie Neenan, “A Survey of Utility Experience
with Real Time Pricing,” Lawrence Berkeley Nat’l Lab., Paper LBNL-54238 (Dec. 1, 2004),
available at http://repositories.cdlib.org/lbnl/LBNL-54238.
24 We recognize, of course, that no practical new pricing system is likely to improve
every customer’s situation. Some users now pay less over the year than the marginal cost of
supplying their power. Heavy users of peak-time power that cannot easily change the time of
their electricity use will not switch voluntarily. But they should not be subsidized forever.
16
Customers reported similar reactions in pilot programs – such as California’s Statewide Pricing
Pilot – that offered a new rate, paid customers to participate, and appealed to their sense of civic
responsibility. Customers perceived that they had greater control over their electric bills and
better understood when and how they use power.22 Likewise, for several years Georgia Power’s
real-time pricing program for commercial and industrial customers has been viewed as
successful for both the utility and customers. 23
Although dynamic pricing can have compelling benefits for both consumers and grid
operators, the merits of dynamic pricing may not be apparent to customers who have always
been on traditional, time-invariant rates.24 Getting customers to opt into these programs can be
difficult and requires attention to:
• 	strategies to inform customers about price changes and ways to respond to them;

25 Barbose et al., “A Survey of Utility Experience with Real Time Pricing,” supra note
23. 26 See Severin Borenstein and Stephen Holland, Center for the Study of Energy Markets,
Univ. of Cal. Energy Inst., Working Paper #106R, “Investment Efficiency in Competitive
Electricity Markets With and Without Time-Varying Retail Prices” (revised July 2003); Severin
Borenstein, “The Long-Run Efficiency of Real-Time Electricity Pricing,” supra note 18.
17
• 	the way customers think about risks and price changes;
• 	program implementation; and
• 	marketing.
Although utilities’ carefully implemented commercial programs have succeeded, others
have failed to attract customers. One-third of available, commercial real-time pricing programs
have zero participants.25 Low participation often reflects inadequate implementation or
promotion. Dynamic pricing programs tend to fail when they stem from a regulatory edict that
the utility opposes and consequently implements with little attention to marketing, user
friendliness, or other details crucial to attracting and retaining customers. Programs that exist
only on paper squander opportunities to temper market power, to reduce distortionary regulation,
and to save billions of dollars for customers. Thus, regulators need to look for ways to give
utilities (and other firms offering these programs) the incentives and flexibility to devote
resources to program implementation and refinement. Moving to dynamic pricing can benefit
utilities, customers who choose dynamic pricing, and even customers who remain on time-
invariant pricing.26 Offering utilities a share of the benefits from dynamic pricing programs may
be an appropriate way to offer them a stake in the programs’ success.
Experience shows that many customers will not take action to change away from
whatever rate the regulator establishes as the default. Regulators and utilities may wish to

27 See, e.g., Lisa Wood, “The New Vanilla: Why Making Time-of-Use the Default Rate
for Residential Customers Makes Sense,” Energy Customer Mgmt. (July/Aug. 2002); John
Beshears, James J. Choi, David Laibson, and Brigitte C. Madrian, “The Importance of Default
Options for Retirement Savings Outcomes: Evidence from the United States,” in Lessons from
Pension Reform in the Americas (2008).
28 See, e.g., James M. Lacko and Janis K. Pappalardo, Bureau of Econ., Fed. Trade
Comm’n, “Improving Consumer Mortgage Disclosures: An Empirical Assessment of Current
and Prototype Disclosure Forms” (June 2007), available at
http://www.ftc.gov/os/2007/06/P025505MortgageDisclosureReport.pdf; James M. Lacko and
Janis K. Pappalardo, Bureau of Econ., Fed. Trade Comm’n, “The Effect of Mortgage Broker
Compensation Disclosures on Consumers and Competition: A Controlled Experiment” (Feb.
2004), available at http://www.ftc.gov/os/2004/01/030123mortgagefullrpt.pdf.
18
consider making dynamic pricing the default rate, as several states have already done with
respect to large commercial and industrial customers.27
Residential and Small Commercial and Industrial Customers: The FTC staff
researches how individual consumers understand marketing materials and mandatory
disclosures. This research – as well as our experience with disclosure regulation – shows that
people with legal, engineering, or policy analytic expertise often write materials that consumers
have difficulty understanding. 28 Communications expertise, testing, and revision can improve
dynamic pricing materials’ effectiveness at attracting customers and equipping them to respond
to dynamic prices.
Careful and innovative design of programs and marketing efforts are important. Well-
designed residential dynamic pricing programs have gotten low sign-up rates, often on the order
of 1 percent. Identifiable flaws in simplified customer decision-making patterns may bias
customers against signing up for dynamic pricing programs described in straightforward but ill-
chosen ways. Letzler describes these decision patterns and suggests the use of incentive-

29 Robert Letzler, Center for the Study of Energy Markets, Univ. of Cal. Energy Inst.,
Working Paper #162, “Applying Psychology to Economic Incentive Design: Using Incentive
Preserving Rebates to Increase Acceptance of Critical Peak Pricing” (Nov. 2006).
30 Frank A. Wolak, “Residential Customer Response to Real-Time Pricing: The Anaheim
Critical-Peak Pricing Experiment” (May 24, 2006), available at
ftp://zia.stanford.edu/pub/papers/anaheim_cpp.pdf.
19
preserving rebates to present critical peak pricing in a more appealing manner without changing
incentives, total annual bills, or the opportunity cost of power.29
Care also must be taken in designing dynamic rate demand response programs to prevent
some customers from manipulating reward levels. For example, there is a longstanding
argument about the relative merits of dynamic pricing and baseline-rebate programs. Baseline-
rebate cases calculate a personalized “baseline” demand level from each customer’s consumption
history and then pay customers rebates when they consume less than their baseline amount
during a critical period. Many customers in an Anaheim, California, baseline-rebate field
experiment exploited these flawed incentives to raise their own baselines and earn larger rebates
at the expense of other ratepayers.30 Consuming more power during the baseline-setting period
allows customers to earn a larger rebate without reducing their critical period consumption.
Thus, baseline-rebate programs present dynamic pricing well, but can create flawed incentives.
By contrast, conventional dynamic pricing can amount to a poor marketing presentation of good
incentives. Letzler summarizes evidence that both incentives and presentation matter, and
suggests the use of incentive-preserving rebates that achieve an appealing rebate-based
presentation of critical peak pricing incentives.
Large Commercial and Industrial Customers: The challenges of getting larger
enterprises to participate in dynamic pricing are different but no less important. Dynamic

31 Severin Borenstein, “Wealth Transfers Among Large Customers from Implementing
Real-Time Retail Electricity Pricing,” 28:2 Energy J. 131 (2007).
32 See Borenstein, “Customer Risk from Real-Time Retail Electricity Pricing: Bill
Volatility and Hedgability,” supra note 20.
33 If the dynamic pricing design reduces revenues but has no effect on consumption
patterns, the regulatory body may wish to consider adopting a different design with greater
demand response.
20
pricing can significantly lower the cost of some accounts’ consumption patterns, letting them
save even if they exhibit little price sensitivity in the timing of their consumption. It can raise
the cost of other consumption patterns, meaning that those customers would have to respond
significantly to prices before they saw any net savings. For example, offering customers the
right to buy power at the time-invariant price – and then billing only deviations from this prepaid
power consumption schedule at the dynamic price – can keep bill levels stable while offering a
more accurate price for deviations.31 Further, dynamic pricing increases bill volatility, but
simple hedges can reduce this risk. 32 Moreover, large customers may be more likely to
participate if they have access to infrastructure to reduce the transaction cost of realizing savings,
such as consulting and documentation of practices that have saved similar customers money.
D. Removing Financial Penalties Against Utilities that Allow or Foster Increased
Price Sensitivity of Demand
Dynamic electricity pricing has the potential to generate billions of dollars in social
savings that customers and utilities or power marketers can share. Regulators should work with
utilities and power marketers to seek out deals that benefit both customers and suppliers.
Regulators should commit to the principle that rational customer response to a well-designed
dynamic pricing program approved by the regulator should not financially penalize the utility.33
Regulators may need to commit to frequent rate adjustments during the first few years of the

34 In “The Long-Run Efficiency of Real-Time Electricity Pricing,” supra note 18, Severin
Borenstein estimates that universal implementation of real-time pricing could reduce the cost of
operating the electric system by 5 to 10 percent. Total U.S. purchases of electricity were $342
billion in 2007 (see http://www.eia.doe.gov/cneaf/electricity/epm/table5_2.html), so universal
deployment of real-time pricing could save between $17 billion and $34 billion. Those estimates
do not take into account the potentially large benefits of making the system robust to unexpected
events, such as the combination of poor hydroelectric conditions, natural gas supply problems,
and a thriving economy that set the stage for California’s crisis in the summer of 2000. Further,
Borenstein and Holland show that putting some customers on real-time pricing benefits those
customers who remain on time-invariant pricing, and that the first customers who switch to
dynamic pricing have the greatest impact. See Borenstein and Holland, “Investment Efficiency
in Competitive Electricity Markets With and Without Time-Varying Retail Prices,” supra note
26.
35 A similar revenue shortfall could occur if customers who are currently using an above-
average proportion of their power during inexpensive periods – and are thus paying a cross-
subsidy to other customers – were to flock to dynamic pricing but not change their consumption
patterns. That dynamic pricing program would lower their bills by offering them a lower price
for their off-peak consumption, which could reduce the utility’s revenue. Utility executives who
are worried about this problem may seek to set inefficient dynamic rates that recover their costs
off-peak or may be inclined to oppose meaningful dynamic pricing. Decoupling can solve this
problem by ensuring that the utility earns its regulated rate of return regardless of the quantity of
power it sells during any time period.
36 A switch from traditional, fixed retail prices to real-time retail prices can be engineered
to cause or prevent shifts in costs among customer classes (i.e., residential, commercial,
industrial) because it does not facilitate arbitrage among classes or make it more difficult to
determine the class to which a customer belongs. A move to real-time pricing typically will
reduce cross-subsidies from customers with flat demand to those with “peaky” demand. For
21
implementation of dynamic pricing to ensure that, in the event the program leads to
unanticipated changes in consumption patterns, there are not prolonged, important deviations
from the rate of return and incentive scheme that the regulators set for the utility.
Unfortunately, many ratemaking systems make it unprofitable or risky for utilities to
offer dynamic pricing.34 Utilities that operate regulated, natural monopoly distribution systems
for electric power often have been reluctant to offer dynamic pricing for electric power because
they fear financial losses if consumption declines below the level for which they have planned 35
(and on which the regulated rates are based).36 The restructuring of utilities’ financial incentives

example, Ramsey pricing principles (minimizing deadweight loss by charging higher prices to
customers with less elastic demand) could be used in conjunction with either pricing system to
allocate joint and common costs among customer classes.
37 Sheryl Carter, “Breaking the Consumption Habit: Ratemaking for Efficient Resource
Decisions,” 14:10 Electricity J. 66 (Dec. 2001).
22
to which we refer can increase the feasibility of quickly implementing real-time (or other time-
varying) retail pricing arrangements.
Traditional utility price regulation typically sets a price based on the total of fixed and
variable costs at the time of the previous rate case, divided by the quantity of electricity (or
natural gas) consumed at that point in time. If the volume declines after the rate case, the
revenue will fall by more than the decline in costs (because only the variable, but not the fixed,
costs will decline). The resulting revenue may not be sufficient for a normal rate of return on the
utility’s investment until an adjustment is made in the volume data used in setting rates. 37
Hence, in comparison to utilities that do nothing or even discourage conservation, utilities that
foster conservation could be penalized by traditional, fixed retail rates. The PA PUC may wish
to investigate whether existing rates and regulations maintain disincentives for suppliers to
promote demand response and conservation by their customers.
Regulators might use one or more of a number of methods to avoid this result (and thus
avoid creating a disincentive to pursue energy conservation and efficient price signals).
“Decoupling” – one method used in several states, with differing specific elements – adjusts
prices to cover fixed costs. Another way is to separate the portion of customer charges for fixed
costs from the portion designed to recover variable costs. Several states – including Idaho and

38 Dovra Bachrach, Sheryl Carter, and Sarah Jaffe, “Do Portfolio Managers Have an
Inherent Conflict of Interest with Energy Efficiency?,” 17:8 Electricity J. 52 (Oct. 2004).
39 Residential onsite electric power generation (other than backup generators) currently
consists primarily of solar cell arrays installed on rooftops. Hot water solar panels also are
relevant to the extent that they displace the use of electricity to heat water. Small-scale wind
generators are being developed for residential use, and residential-scale fuel cells may attract
considerable interest as prices of these generators decline and their reliability is established. The
considerable research and development regarding small-scale fuel cells for use in the
transportation sector could be transferable to onsite generation products.
40 At the same time, onsite generation can benefit the customer/investor, the distribution
utility, and other ratepayers – especially in transmission-constrained areas – by increasing supply
and easing reliability concerns inside the constraint. The Massachusetts Department of
Telecommunications and Energy opened an inquiry into policies affecting onsite generation,
including incentives to exclude onsite generation, recovery of potential stranded costs of utilities
caused by onsite generation, and whether utilities should offer more than one type of standby
service. See Commonwealth of Massachusetts, Dep’t of Telecomm. and Energy, Investigation of
Standby Rates and Alternative Rate Structures that Will Promote Efficient Deployment of
Distributed Generation, D.T.E. 7-06 (Mar. 23, 2007), available at http://masstech.org/2007-03-
23-DG-DTE-07-6-Order.pdf.
23
New York – are considering or seeking proposals designed to encourage utilities to work with
consumers to conserve energy through variations on these two methods. 38
E. Reforming Standby Electric Power Service Regulations
Technical improvements in small-scale electricity generators have made it increasingly
attractive for some commercial (and even some residential) electricity customers to consider
investing in onsite generation – i.e., building their own electric generation facilities on the site of
an industrial manufacturing facility or at home.39 Onsite generation (also known as “distributed
generation”) represents a form of competition – a substitute for power from the grid. Because
onsite generation can result in reduced (or more price-sensitive) demand for incumbent utilities,
utilities can have incentives to prevent customers from making such investments. 40 In the FTC
investigation of the DTE/MichCon merger, for example, the staff obtained documents indicating

41 Fed. Trade Comm’n, Analysis of the Proposed Consent Order and Draft Complaint to
Aid Public Comment in DTE Energy Company and MCN Energy Group Inc., File No. 001 0067,
available at http://www.ftc.gov/os/2001/03/dteanalysis.htm. For an example of government
modeling the penetration of onsite generation, see Erin Boedecker, John Cymbalsky, and Steven
Wade, “Modeling Distributed Electricity Generation in the NEMS Buildings Models” (2002),
available at http://www.eia.doe.gov/oiaf/analysispaper/electricity_generation.html.
42 Dynamic (ideally, real-time) pricing is a particularly important tool to create incentives
for efficient investment in intermittent generation technologies (e.g., wind and solar) and in
technologies that store power or change load shapes. Real-time prices capture changing patterns
of electricity scarcity or abundance and also harness market forces to help integrate technologies
such as wind turbines, solar generation, plug-in hybrid vehicles, and other energy storage
devices into the grid.
24
that the incumbent electric utility considered onsite generation to be a competitive threat.41 A
potential policy concern in this situation is that a utility not covered by a revenue-decoupling
arrangement may manipulate regulatory policies to impede socially efficient onsite generation to
avoid financial losses by the utility. Environmental concerns and technological developments
are transforming the characteristics of cost-effective generation technologies, making it
particularly important to design technologically neutral regulatory approaches that enable
emerging technologies to enter based on their merits.42
1. 	Nondiscriminatory Onsite Generation Policies
One scenario of concern involves onsite generators that are able to, and often do, operate
full-time. Onsite gas turbines (e.g., microturbines) and other technologies that typically run
essentially nonstop are vulnerable to regulatory flaws that can differ from those that would
discriminate against intermittent technologies, such as wind or solar onsite generation. The
scenario starts with the recognition that customer interest in onsite generation depends on a
variety of factors, such as fuel and equipment costs relative to the price of power obtained from
the grid. Reliability preferences also are likely to be a factor. One potential problem for a

43 Most customers who invest in intermittent generation technologies such as wind and
solar generation will need power from the grid on a regular basis.
44 Such a strategy on the utility’s part would be profit-maximizing solely by dint of its
effect in preventing entry by competing onsite generators.
45 Under this approach, the utility would claim the need to retain more generation
capacity than necessary to maintain an acceptable level of reliability and would blame the excess
costs on customers with onsite generation. In reality, the utility could reduce capacity reserves
with an acceptable level of reliability, because it is extremely unlikely that onsite generators
would simultaneously suffer mechanical breakdowns. James Mulligan, “The Economies of
25
customer considering an onsite generation investment is that the customer occasionally may need
to receive “standby” power from the utility (if, for example, the onsite generator has a
mechanical breakdown or needs maintenance).43 Utilities, which generally are allowed to charge
special rates to such intermittent customers, can frustrate competitive inroads of onsite
generation by charging inefficiently high prices for standby service. Supracompetitive standby
rates exceed opportunity costs (including a market-based risk premium) and may provide
utilities with a rate of return above the competitive level, or even above the short-term profit-
maximizing price.44 For example, the utility might set standby service charges so high that they
offset any savings the customer might expect from generating power onsite. If the price of
standby service exceeds the efficient price, some customers are likely to be deterred from
undertaking efficient onsite generation projects and competing.
There may be a number of ways in which an incumbent utility could persuade the
regulatory body to authorize a price for standby service that exceeds the efficient price and thus
deters the entry of non-intermittent onsite generation. One key way to do so is to posit an
unrealistic scenario in estimating the costs of providing such service – e.g., a situation in which
all onsite generators simultaneously break down when demand from other utility customers is at
its peak.45 A utility should not be allowed to block the efficient entry of onsite generation by

Massed Reserves,” 73 Am. Econ. Rev. 725 (1983); Walter Y. Oi, “Productivity in the
Distributive Trades: The Shopper and Economies of Massed Reserves,” in Zvi Griliches (ed.),
Output Measurement in the Service Sectors 161 (U. of Chicago Press, 1992), available at
http://nber15.nber.org/bookcv_chicago/9780226308852_web.pdf.
46 By contrast, if a utility were unable to manipulate government regulatory proceedings
in order to block entry by onsite generators, its only recourse in responding to the challenge of
onsite generation might be to improve its service and reduce its costs and prices, just as
incumbent suppliers respond to increased competition in other markets. Such improvements in
economic performance would redound to the benefit of all electricity consumers.
26
positing an unrealistic scenario as the basis for setting the price of standby service.46 Instead, the
pricing of standby service should reflect calculations – such as those of reliability organizations
– that relate reserve levels to the probability of blackouts or other reliability incidents.
In the case of intermittent forms of onsite generation (or of onsite generation that could
generate full-time but nonetheless operates intermittently depending on the relative costs of
making or buying power), a utility without decoupling protection could seek to block entry of
onsite generation by applying excessive standby rates to all power purchases by customers with
onsite generators. This scenario could be inappropriate for solar generators in power systems
where peak loads occur in the summer and demand response is well developed. Solar generators
operate during afternoon peak demand periods but not overnight. Hence, onsite solar generators
could flatten the load profile of the customer and the system. This kind of onsite solar
generation, combined with demand response programs that reduce load during morning, evening,
and cloudy peak periods, could allow the power system to increase capacity utilization and
reduce the need to build and operate costly peakers.
Even if a utility does not establish a standby service price in excess of the efficient level,
it might be able to block competition and increased demand response from onsite generators by
offering only an “unlimited” form of standby service. This strategy in and of itself can

47 For example, “as-available” standby service could be conditioned on whether power is
readily available to serve the standby power customer at or below a pre-set wholesale price.
Another variation on the theme would offer power at the real-time wholesale price plus a
distribution charge (unless power was so scarce that the system was in an emergency status and
running with reduced reserves). Another alternative to unlimited standby service could be
standby service that is capped at a specific quantity. (A cooperative utility in Hawaii allows
customers to specify the amount of standby service for which they are willing to pay. Under this
tariff, the utility operates a circuit breaker (paid for by the customer) to ensure that the customer
draws no more than the specified amount. See
http://www.kiuc.coop/anne/IRP_public_site/Tariff/Rate_Rider%20S.pdf.) Both of these
alternatives would involve lower costs for the utility – and presumably lower prices to the onsite
generator – than unlimited standby service.
More generally, customers with onsite generation, like other customers, may have
varying preferences for the reliability of their electric service. Many utilities offer lower prices
to customers who will accept a lower level of reliability (known as “interruptible service”). The
same range of reliability and price tradeoffs could apply to standby service.
48 One form of alternative standby service involves the utility’s supply of the additional
power only if generation and transmission capacity are readily available. Some states require
utilities to offer this type of contingent standby service, and to price it below the price of
unlimited standby service. Contingent standby service is conceptually very similar to the
“interruptible service” (referenced in note 47, supra) that is routinely offered to industrial and
commercial customers at a rate lower than the rate for standard service. Under interruptible
service, customers get a discount on all of their power in return for an agreement to waive their
right to demand as much power as they want at the predetermined price.
49 If the concern is limited to utilities’ decisions to offer only one form of standby service,
regulators may wish to evaluate the benefits and costs of requiring utilities to offer a choice
among levels of standby service.
27
discourage entry by onsite generation, because some onsite generation investments may be
financially viable only if standby service is available in a limited form that costs less than
unlimited standby service.47 If the utility persuades the regulatory body to allow it to offer only
“unlimited” – and thus more expensive – standby service, then competition from onsite
generators that require nothing beyond limited (and less expensive) standby service might not
develop.48,49 One solution worth considering would be to put standby customers on the

50 So long as competition in the supply of standby service is feasible, maintaining such
competition is a potentially attractive solution. If customers wanted (and were permitted) to buy
standby capacity, presumably they could enter into arrangements under which other parties
would build generators for this purpose, attach the new generators to the existing network (or a
new network), and sell standby capacity under long-term contracts to willing buyers.
In light of the severity of the regulatory challenges, allowing entry and competition in the
provision of standby service may well benefit customers more than attempts to regulate the price
while blocking entry. If competition in the supply of standby service is not allowed, then a
market power problem may be present that may be difficult to address with price regulation.
Clearly the price has to be high enough for the seller to anticipate earning a normal rate of return.
If there are economies of scale in the provision of standby capacity, however, marginal-cost
pricing alone will not raise enough revenue. Moreover, a regulator that mandates standby
service at a price that is set ex ante is forcing the utility to assume risks, which raises the
problem of determining what the compensation should be for assuming such risks. An
alternative would be to shift the risk to customers, by charging prices determined ex post to raise
the right amount of revenue. But this raises the concern among standby service customers that
the risk premium will be set above the market level as part of the utility’s strategy to protect
itself from competition. In short, this is not an easy matter for regulation to remedy efficiently or
effectively.
51 Jay Apt, M. Granger Morgan, and Carnegie Mellon Electricity Industry Center staff,
Critical Electric Power Issues in Pennsylvania: Transmission, Distributed Generation and
Continuing Services When the Grid Fails (2005), at 33 et seq., available at
http://wpweb2.tepper.cmu.edu/ceic/pdfs_other/Critical_Electric_Power_Issues_in_Pennsylvania.
pdf. A micro-grid can improve reliability for member customers by providing power even if the
utility grid is not functioning and the customer’s onsite generator is not operable.
28
wholesale, real-time price plus the same, appropriately chosen distribution and connection
charges applicable to other, similarly-situated, real-time pricing customers.
One method to ensure that utilities do not use distortions in standby service to impede
onsite generation is to allow the entry of alternative standby service providers.50 For example,
one or more owners of onsite generating capacity could function as sources of standby service
for other onsite generators if they are linked. This linking of onsite generators in order to
provide alternative standby services into what is known as a “micro-grid” 51 is likely to be most
practical when onsite generators are clustered in an industrial park or another type of commercial
development area. In order for micro-grids to be formed, state or federal regulations may need

52 If the concern is limited to utilities’ decisions to offer only one form of standby service,
regulators may wish to evaluate the benefits and costs of requiring utilities to offer a choice
among levels of standby service.
53 The challenge for regulators becomes complicated when a significant fraction of rates
goes to pay costs that onsite generation does not affect. Onsite generation reduces fuel costs and
variable operating and maintenance costs. Onsite generation can reduce fixed operating and
maintenance costs by allowing facilities to be retired or by avoiding construction. When onsite
generation avoids construction, it also avoids such facilities’ fixed costs.
29
to be amended to allow the currently prohibited entry of competing transmission facilities
(“overbuilds”).52
2. 	Sunk Costs Complicate the Analysis of Onsite Generation Policy
The question for regulators about expanding onsite generation – aside from removing
regulatory barriers that exist for purely anticompetitive reasons – becomes more complex if the
opportunity to avoid paying markups that cover the utility’s sunk costs (i.e., its irreversibly
expended fixed costs) drives the interest in onsite generation. 53 Most facility construction costs
are sunk, as are the yet-to-be-recovered losses from past wholesale purchases of expensive
power that then was resold at lower, regulated retail rates. When regulators guarantee utilities’
ability to recover costs, the purchase of onsite generators often lets a customer shift the
responsibility to pay to other customers who continue to buy all of their power from the utility.
When such markups make onsite generation attractive, regulators should consider how their
onsite generation policies affect the distribution of sunk costs.
Customer actions to avoid paying sunk costs can create costs to society that fully or
partially offset the benefits. For example, consider a situation in which an onsite generator’s
marginal cost is greater than the grid’s marginal cost, but the utility’s fixed-cost-recovery
markups make grid power more expensive than onsite power. The customer comes out ahead by

54 If customers can come out ahead by building their own generator, other customer
choices are likely to be distorted. For example, they are likely to site their production facilities
in less convenient places than they would if the price of power were closer to its marginal cost.
55 Onsite generation entry combined with slow regulatory action to adjust rates can
reduce utility profits, as noted above. The threat of such a result may provide utilities with an
incentive to avoid investment cost overruns.
30
using the onsite generator, but only because he or she shifted to someone else the responsibility
to pay an amount greater than his or her savings. This result is socially wasteful. If a rate allows
ratepayers to benefit by making socially wasteful choices about onsite generation, then changing
the rate’s allocation of fixed costs may improve onsite generation investment and facility siting
choices.54,55
If utilities were purely profit-maximizing firms providing a product in the marketplace,
consumer choice about whether to make or buy power would be beneficial. If utilities were
purely public service entities that taxed electricity to cover sunk costs, however, then customers’
substitution of untaxed electricity for taxed electricity to make other customers pay their share of
the sunk costs could be unfair and unproductive free riding. Regulated utilities with substantial
sunk costs can create a situation that is a mix of the two.
F. 	Facilitating the Participation of Customers as “Suppliers” in Wholesale Markets
An increase in demand response participation can produce a number of beneficial results:
it increases ratepayer savings, can be an effective substitute for distortionary market power
mitigation policies, reduces the risk of exercise of generators’ market power, and improves
system reliability. As with designing an appealing dynamic pricing rate structure, attracting
participation in demand response aggregations is a task that deserves careful attention. The PA
PUC, PJM Interconnection, and Midwest ISO may wish to consider whether letting customers

56 For a general discussion and framework for considering customization of demand
response offers, see Electric Power Research Inst., New Principles for Demand Response
Planning (Mar. 2002), EP-P6035/C3047, available at
http://www.goodcents.com/info/New%20Principles%20for%20Demand%20Response.pdf.
31
customize their demand response offers can increase participation rates because some potential
participants face constraints on how quickly and for how long they can cut back their power
consumption. 56 Some useful insights about this design task may be garnered from efforts to
encourage generators to provide capacity reserves.
A key concept in attracting generators to supply reserves is to recognize that different
generators face different technological constraints. A good example is ramping speed. Some
generation technologies can be ramped up from low power levels to full power levels in a
relatively short period, while others require a longer time. Rather than setting a single price and
a single requirement for ramping speed, market designers and operators encourage participation
by more generators by customizing contracts to supply reserves. Generally, system operators
pay more for reserve capacity that can ramp up quickly than for slower-ramping reserves.
Generators that cannot ramp up output quickly require the ability to specify a minimum
run time during which the system operator dispatches the unit. Some retail customers
contemplating offering demand response may have the opposite concern. For example, a
grocery store may be willing to reduce electric load temporarily by postponing refrigerator
operation for a few minutes, so long as this does not risk exposing the food to unsafe

57 Not all load from refrigeration equipment is necessarily devoted to cooling of food.
For example, display cases may also have heating elements that keep the doors from collecting
condensation when the air is moist. Hence, demand response may involve no change in cooling
of the food, but instead may involve turning off the anti-condensation heating elements for a
period of time. The store might wish to avoid leaving condensation at higher levels for an
extended period, but this would be less likely to raise health and safety concerns than decreasing
refrigeration. Cal. Energy Comm’n, “Enhanced Automation Case Study 7: Lighting and
Equipment Controls/Grocery Store” (2005), available at
http://www.energy.ca.gov/enhancedautomation/case_studies/CS07_Albertsons_w2.pdf.
58 Charles Goldman, Nicole Hopper, and Ranjit Bharvirkar (Lawrence Berkeley Nat’l
Lab.) and Bernie Neenan and Peter Cappers (Utilipoint Int’l), “Estimating Demand Response
Market Potential Among Large Commercial and Industrial Customers: A Scoping Study,” Paper
LBNL-61498, § 3.4.3 (Jan. 2007), available at
http://www.energetics.com/electricity_forum_2007/pdfs/61498.pdf.
59 Sila Kiliccote and Mary Ann Piette (Lawrence Berkeley Nat’l Lab.) and David Hansen
(U.S. Dep’t of Energy), “Advanced Controls and Communications for Demand Response and
Energy Efficiency in Commercial Buildings” (Jan. 2006), paper for the Second Carnegie Mellon
Conference in Electric Power Systems: Monitoring, Sensing, Software and Its Valuation for the
Changing Electric Power Industry, available at
32
temperatures.57 Absent an ability to specify a maximum duration of refrigeration curtailment,
however, the food store is unlikely to offer to postpone its cooling load.
A manufacturer with an energy-intensive batch process may be willing to offer demand
response so long as it is given enough notice to complete safely the processing of the current
batch or to postpone processing the next batch in an orderly manner.58 Similarly, a manufacturer
or retailer may be willing to consider bidding to supply demand response only if it is assured that
there will be sufficient spacing between the instances when the system operator asks the firm to
trim its consumption – spacing that may be necessary to meet the firm’s existing obligations to
supply its own customers or to maintain adequate inventories.
In the case of large commercial buildings, the magnitude of demand response offers may
be contingent on the time of day when dispatch occurs, or on how early notice was provided of a
pending dispatch of the building’s offer to reduce power consumption.59 During the early

http://www.osti.gov/energycitations/servlets/purl/889248-7DjwKn/889248.PDF.
33
evening, residential demand rises for cooking, climate control, and lighting, while commercial
flexibility also increases. Dispatch of demand response near the end of the business day could
be larger and longer because occupancy of office buildings will be low at the time of the
dispatch and natural cooling will help bring interior temperatures within acceptable limits before
the next morning. Demand response by office buildings could be even larger and longer if the
dispatch occurred just before a weekend. Early notice can facilitate larger dispatch at a
subsequent time if the building is pre-cooled to the low end of the acceptable interior
temperature range prior to the expected dispatch period, because pre-cooling reduces the need to
air condition the building to stay within the acceptable zone during the dispatch period.
In general, the PA PUC, Conservation Service Providers, PJM Interconnection, and
Midwest ISO may wish to consider a wide range of customized demand response specifications,
so long as the benefits are likely to exceed the costs of administering the customized offers. The
PA PUC also may wish to urge FERC, PJM Interconnection, and Midwest ISO to consider ways
to ensure that wholesale markets can accommodate the range of innovative, customer-friendly
varieties of demand response that Conservation Service Providers may develop.
IV. 	Conclusion
We commend the PA PUC for seeking to increase demand response and energy
efficiency. Dynamic pricing and demand response programs can be powerful tools to empower
customers to help manage peak and overall load. Good programs can empower customers to
manage load shapes to enhance reliability, reduce peaking costs, and complement unpredictable,
intermittent generators with a combination of flexible demand and flexible supply. Advanced
metering that both provides energy consumption data to customers and allows dynamic pricing is

34
a key facilitating technology. This comment has recommended that the PA PUC: (1) encourage
real-time or other dynamic pricing programs that increase economic efficiency; (2) urge utilities
to design and market dynamic pricing programs that appeal to customers; (3) eliminate
regulatory provisions that financially penalize power suppliers if they facilitate efficient dynamic
pricing; (4) offer fair standby pricing policies for customers with onsite generation investments;
and (5) advocate for demand response bid flexibility.