Using demand response to mitigate rate shocks.
Ahmad Faruqui is a principal with The Brattle Group, located in the firm’s San Francisco office. This paper was inspired by conversations with Roger Levy and has benefited from comments on a draft version by Bob Borlick, Phil Hanser, Bob Lieberman, Johannes Pfeifenberger, and Lisa Wood. Contact Faruqui at firstname.lastname@example.org. Ryan Hledik provided background research for this article.
By protecting customers from price spikes during a few hours in the year, existing rate-design regulations also are preventing them from lowering their average rates throughout the entire year. That is the paradox of utility regulation.
Responding to the directives of the Energy Policy Act of 2005 (EPACT), two recent reports by the U.S. Department of Energy and the Federal Energy Regulatory Commission (FERC) make a strong case for dynamic pricing of electricity.1 These reports pick up on a theme that was first articulated by the California Public Utilities Commission (CPUC) in 2002. As it began its deliberations on dynamic pricing, advanced metering, and demand response (DR), the CPUC instituted Rulemaking (R.) 02-06-001 “to provide the forum to formulate comprehensive policies that will develop demand flexibility as a resource to enhance electric system reliability, reduce power purchase and individual consumer costs, and protect the environment.”2
Five years after the nation’s biggest power crisis, the percentage of Americans on advanced metering and dynamic pricing in the United States continues to be in the single digits. Prices for basic electricity service—i.e., standard rates in regulated markets and default rates in restructured markets—do not vary by time of use, let alone vary dynamically with changing market conditions. In the minds of many policy-makers, dynamic pricing has become associated with rate shocks, rate volatility, unpredictability, and loss of control over energy costs. This is ironic, since it is designed to overcome precisely these very problems. How has this occurred and what can be done to change it? These questions are taken up here.
The Game Plan
Because electricity cannot be stored easily in large quantities, and because the cost of providing power varies across the day and season, wholesale electricity prices exhibit significant temporal variability. For markets to function efficiently, this temporal variation needs to be passed through to retail customers. Otherwise, customers will over-consume electricity during peak hours, necessitating the installation of expensive peaking capacity.
There are three generic ways in which such time variation could be provided to customers (see Table 1).
Within each category, several varieties exist. Essentially, there is almost no limit to the choices that can be given to customers so they can make a well-informed tradeoff between average electricity costs and price volatility.
There is strong empirical evidence that during critical peak hours, when the power system is stressed by a shortage of supply relative to demand, reducing customer loads by a few percentage points can lower the wholesale cost of electricity significantly. As shown in California’s recent statewide pricing pilot, customers do not have to make drastic adjustments in order to drop their load during these critical hours to achieve this goal.4
• The average residential customer dropped usage in the critical peak period by 13 percent in response to a price signal that was about five times higher than the standard tariff price.
• Customers with enabling technologies dropped their load by twice as much. In most cases, customers simply raised their thermostat setting by a couple of degrees. In other words, customers responded to higher prices without making any drastic changes in their lifestyle.
• More important, empirical evidence suggests that the majority of the load drop came from a minority of customers. Several customers choose to make no load reductions and were content with buying through at the higher prices.
A drop of 13 percent in critical peak loads can have a substantial impact on wholesale energy costs in the near term, benefiting all customers (see sidebar, p. 49 which discusses the benefits of a 3 percent reduction). But by lowering the need for investing in peaking generation (and transmission and possibly distribution) capacity, demand response can have a greater impact on the long-run price of electricity.5
During the 2000-2001 power crisis, California’s retail customers unknowingly exacerbated the demand-supply imbalance by not lowering their electricity usage. Then Gov. Gray Davis famously said that had he been able to raise electric rates, he would have solved the problem in a few minutes. Customers finally cut their usage in response to blackouts and a barrage of appeals by the state of California to conserve power. In the process, one utility was bankrupted and two others were brought to the verge of financial ruin. Ultimately, these costs were recovered from retail customers who were hit by price hikes in the 20 percent range.6 In addition, the state’s multi-billion dollar budget surplus became a multi-billion dollar deficit as government officials signed expensive long-term contracts for the supply of electricity to Californians.
While these facts are only too well known to utilities and regulators, dynamic pricing has yet to find its way into the mainstream of utility ratemaking. One reason is that customers are not really aware that the price they pay for electricity today includes a significant risk premium for insuring them against wholesale price volatility. Effectively, policy-makers have made this choice for them. What the policy-makers have failed to communicate to customers is that the average price of electricity will go down if dynamic pricing is introduced.
The State of Play
EPACT requires utilities to provide individual customers with advanced interval meters upon request and “time-based” rate schedules that reflect the variance in wholesale prices. It is unclear whether time-based prices are synonymous with dynamic pricing or whether static, time-of-use (TOU) rates also qualify. And, while EPACT encourages DR devices, it does not encourage utilities to attract customers to time-based rates. According to the FERC staff report cited earlier, while 7 percent of retail customers are on advanced metering systems that would allow dynamic pricing signals to be conveyed, a very small percentage of customers are actually on dynamic pricing rates.7
In April of last year, Gordon van Welie, president and CEO of the New England ISO, spoke at the ISO’s Annual Demand Response Summit about the need for states to implement innovative retail-rate designs that encourage price-responsive demand by letting prices rise during peak periods, especially on critical supply shortage days. He said only then can customers make informed decisions about electricity use. Positing that states face a choice between continuing to incur the high costs of overbuilding the system that reinforces consumption on the few highest demand days or, instead, becoming more efficient and getting more from the existing infrastructure, he called for making dynamic pricing the basis for default pricing for large customers.8
The atmosphere for dynamic pricing was similar at the national town meeting and symposium on DR that was held in Berkeley in June 2006.9 Jackie Pfannenstiel, the chair of the California Energy Commission, noted that when she was just beginning her career in Connecticut, a time-of-use pricing experiment was conducted in that state showing that customers did respond to time-varying rates and that they would accept them as a means of lowering their bills and improving system reliability. But even though 30 years had come and gone, and numerous other experiments had been conducted across the country, decision makers in key positions at utilities and commissions across the nation continue to be skeptical.
Four summers ago, just months after the California energy crisis ended, the time seemed right for DR. Specifically, one analyst noted:10
“There is a timely and renewed national interest in price-responsive demand among utilities, independent system operators, policy-makers, and regulators. The Federal Energy Regulatory Commission and Department of Energy co-sponsored a conference on demand response in February. Independent system operators in California, New York, and New England initiated widely publicized price-responsive retail-load programs last summer that are still in place. And individual utilities have recently introduced a variety of price-responsive demand programs.”
That year, then FERC Chairman Pat Wood had touted price-responsive demand as a way to potentially increase the effectiveness of wholesale markets. In FERC’s standard market design white paper issued in March 2002, DR was cited as an important design requirement.
But outside of the developments in California, and a few pilots in other states, little progress on dynamic pricing has taken place. One way to break the logjam is to make dynamic pricing part of the conditions of basic electricity service. Last summer’s heat wave is yet another reminder that we still have not progressed very far in terms of sending dynamic-price signals to customers. Demand was kept in check and blackouts remained localized by cutting supplies to customers on curtailable and interruptible rates and through voluntary appeals for conservation. Absent such emergency efforts, electricity could have gone out for large portions of the country.
The Blocking Action
The inability of the state regulatory process to include dynamic pricing as part of the standard electric-rate offering to retail customers in non-restructured states, and as part of the default service in restructured states, demands some attention. This reluctance seems to be coming from the desire of regulators to “protect” consumers by not offering them choices. This is akin to “killing the market with kindness.”
Admittedly, the process of making choices in the marketplace is complex and burdensome. Nobel Laureate Dan McFadden has found in his new research on consumer behavior that because consumers are “suspicious of trading partners, and fearful of deception, exploitation, or unfair treatment … [they] … exhibit various degrees of agoraphobia, a term that means literally ‘fear of the marketplace.’”11
However, in the case of electricity, especially for small commercial and residential customers, the issue is not so much about making a complex choice as making a simple choice. Do they want to pay higher average rates in return for price stability or do they want to pay lower rates in return for some price instability?
The time-honored Bonbright criteria suggest that prices for basic electricity service should be fair, simple, acceptable, effective, equitable, non-discriminatory, and efficient.12 In today’s environment, this no longer translates into “flat” rates. Electricity prices vary, not only within a day, but throughout the entire year. This cost variation is especially pronounced in transmission-constrained locations. Yet, even in California, six years after the crisis, the only customers who see such time-variation in their basic rates are large commercial and industrial customers and even they are on static TOU rates, not dynamic rates that change based on demand-supply conditions.
It is economically inefficient to mask the time-variation in the cost of producing electricity. Typically, gas-fired combustion turbine plants are used as peaking plants because they are cheap to build but expensive to operate. Such peaking units have low capacity factors and are idle for most hours of the year. The issue is, in part, a cost-effectiveness one—when and whether to use price-responsive demand to replace the expensive power generated by a peaking plant. In other words, dynamic pricing (as well as other forms of DR) needs to be part of the resource-adequacy process.
Some argue that it is incorrect to introduce time-variation in default electric rates because customers prefer rate stability. But this preference for stability masks important cross-subsidies that often are ignored in the rate-setting process. Non-time varying rates subsidize customers with relatively peaky load shapes at the expense of customers with relatively flat load shapes.
Customers face time-varying rates for other products and services, such as cellular phone services, bridge tolls, airline tickets, and vacation packages, so why not for electricity? A likely reason is that a transition from non-time varying rates to dynamically time-varying rates, even though it will be cost-based, will create winners and losers. The losers will be those customers whose load shapes are peakier-than-average, since their average rates would rise, and the gainers will be those customers with flatter-than-average customers whose average rates would fall. Is it that the regulatory process has not been able to make well-informed tradeoffs between gainers and losers, especially when the losers are vocal and well-connected?
No, because there are other cases when such tradeoffs have been made. A case in point is inverted block rates for electricity, where the more you use, the more you pay per kilowatt-hour. (Some will remember that not-so-long-ago, the inverse was true, and customers faced declining block rates—a practice designed to encourage more consumption of product, and common to many industries.) Even if they are cost-based, reflecting the rising marginal cost of electricity, they do make some consumers better off at the expense of others. And sometimes similar tradeoffs between winners and losers have been made even when they are not cost based.
The state of California, in the aftermath of the energy crisis, concluded that rates had to go up by roughly 20 percent to make the state whole in its role as the supplier of last resort. But this increase was not applied uniformly across the board.
There are encouraging signs that utilities and states are moving to offer dynamic pricing on an elective basis, with Southern California Edison being the latest large utility to make such a commitment.13 However, such elective offerings by a handful of utilities are unlikely to prevent the recurrence of a power crisis in the years to come, since only a small number of customers are likely to enroll in such programs. During an electricity crisis, the usual remedies of curtailing and interrupting large customers and making voluntary appeals for conservation to small commercial and residential customers are likely to be the only means of defense against blackouts and brownouts.
While reserve margins are still in the 20 percent-plus range, it is time to rethink the anti-DR bias inherent in default service. Regulators and utilities seriously should rethink their current policies and consider offering some type of dynamic pricing as the default rate for all customers and allowing customers to “opt out” of static, non-time varying rates if they so choose. The specific type of dynamic-pricing rate can vary by customer class. For example, critical-peak pricing might be best suited to residential and small commercial and industrial customers, while real-time pricing might be best suited for larger commercial and industrial customers.
Will that lead to a ratepayer revolt? No. These very same customers have learned to live with adjustable rate mortgages where the risks and corresponding benefits are much higher than those associated with monthly electricity purchases.
1. U.S. Department of Energy, Benefits of Demand Response in Electricity Markets and Recommendations for Achieving Them, A Report to the United States Congress Pursuant to Section 1252 of the Energy Policy Act of 2005 and Federal Energy Regulatory Commission, Assessment of Demand Response & Advanced Metering, Staff Report, Docket Number AD-06-2-000, August 2006.
2. California Public Utilities Commission, “Order Instituting Rulemaking on Policies and Practices for Advanced Metering, Demand Response, and Dynamic Pricing,” Ruling 02-06-001, filed June 6, 2002, San Francisco, Calif.
3. FERC Staff Report, Assessment of Demand Response & Advanced Metering, Docket Number AD-06-2-000, August 2006. Figure IV-2.
4. Ahmad Faruqui and Stephen S. George, “Quantifying Customer Response to Dynamic Pricing,” The Electricity Journal, May 2005.
5. To date, generation planning has considered the costs of alternative power plants and the mix of fuel sources. Typically, gas-fired combustion turbine plants are used as peaking plants, coal plants are used for base load, and combined-cycle plants are used for some amount of base load and some amount of peaking. The issue is, in part, a cost-effectiveness one—when and whether to price responsive demand to replace the expensive power generated by a peaking plant.
6. Every residential customer saw a price hike of a penny per kilowatt-hour in January 2001. In May, customers who used more than 130 percent of their baseline allowances saw increases that ranged from 5 to 10 pennies per kilowatt-hour.
7. FERC Staff Report, “Assessment of Demand Response & Advanced Metering,” Docket Number AD-06-2-000, August 2006. Page 29.
8. See http://www.iso-ne.com/pubs/pubcomm/pres_spchs/2006/dr_speech.pdf. Also, Gordon van Welie, The Providence Journal, May 13, 2006.
9. National Town Meeting and Symposium on Demand Response, Berkeley, Calif., June 2006.
10. Lisa Wood, “The New Vanilla: Why Making Time of Use Rates the Default Rate for Residential Customers Makes Sense,” Fortnightly’s Energy Customer Management, July/August 2002.
11. Daniel McFadden, “Free markets and fettered consumers,” Presidential address to the American Economic Association, Jan. 7, 2006.
12. James C. Bonbright, Albert L. Danielsen, and David R. Kammerschen, Principles of Public Utility Rates, 2nd ed. (Arlington, VA: Public Utilities Reports, 1988).
13. Southern California Edison, Advanced Metering Infrastructure: Conceptual Feasibility Report, August 2006.