Real-time pricing has been hindered by the misperception that a shift to RTP will create new types of risks, without creating benefits for utilities or customers.
During the past two decades, real-time pricing (RTP) has spawned a cottage industry of experts who continue to wax eloquent on its benefits. RTP can indeed provide substantial benefits to energy customers and utilities. However, only a handful of utilities offer such programs, and only a few thousand customers receive RTP service. This paradox resolves itself once we realize that there are significant barriers to RTP, many of them having to do with perceptions and not reality.
The overarching barrier to widespread application of RTP is a misperception that a shift to RTP will create new types of risks for utilities and regulators, without creating commensurate benefits for either utilities or customers. Both utilities and regulators have become risk averse in experimentation with new policies, having been burned first in California and then in the Enron crisis. The challenge is to convince them that no such failures await them with implementing RTP.
Another barrier is a misperception that a prerequisite for RTP is competition between retail energy service providers. However, as the examples of California and Georgia illustrate, retail competition is neither sufficient nor necessary for RTP.
The vast majority of customers have a natural reluctance to participate in RTP because they equate higher price volatility with higher prices, which in their minds translates into higher bills. They do not realize that higher price volatility often means that prices will be very high during a certain number of hours, but very low during a greater number of hours-potentially with lower annual bills. Of those customers who realize that higher price volatility may well translate into lower expected bills, a large number are risk averse. These customers are not inclined to "play the market." Other customers believe that the effort involved in any type of load shifting outweighs any potential benefits. And then there are some paranoid customers who think that RTP is just another way for their electric utility to gouge them.
These perceptions are borne out in a series of market research studies that have been conducted over the past five years. In these studies, researchers interviewed customers about their preferences for a range of pricing options, based on their stated intent to buy or not buy one or more of these products ().
For example, customers are willing to pay real money to avoid being placed on a time-dependent rate structure, such as seasonal, time-of-use (TOU), or RTP.1 In fact, large commercial and industrial (C&I) customers would be willing to pay a premium of .33 cents per kWh in a flat rate, rather than be placed on a two-part RTP. Small and medium C&I customers would be willing to pay a substantially higher premium, of 3.9 cents per kWh, to avoid being placed on hourly prices. These perceptions constitute a real barrier to RTP.
As the example of utilities such as Georgia Power illustrates, it is possible to overcome this barrier through successful program design. Examples from other industries indicate that customers do respond to the opportunity to lower costs by shifting their usage patterns (airlines) or by taking on time-varying products (adjustable rate mortgages).
Utility customers are attracted to RTP on the premise that it will save them money. However, when prices spiked in various markets during the past few summers, many customers dropped out. For example, Duke Energy had 100 customers on its RTP program, but now it has only 59. BC Hydro had 25 customers; now it has none. Researchers need to study and analyze the following inter-related set of issues:
- If customers were offered RTP on a voluntary basis, how many would take it?
- Are customers more likely to take a two-part design than a one-part design because it provides a measure of price insurance?
- What types of customers are drawn to RTP?
- At what rate are customers willing to trade off a lower expected value of price against a higher standard deviation of price?
- How many more customers would take RTP if it were to be combined with some type of price protection product, such as a price cap or a price collar?
- What is the demand for other types of market-based pricing roducts, such as occasional RTP?
Another issue that needs to be tackled by researchers is the amount of load clipping or shifting that would be induced by RTP. Experience has shown that only a few customers either reduce load or shift it from on-peak to off-peak periods. And of those that do respond to RTP, there is considerable day-to-day variation in response patterns.
Customers may not sign up for RTP because they do not know how to lower costs by reducing usage during high-cost hours and increasing usage during low-cost hours. Or they may lack the capability to shift load. This contention, often advanced by skeptics, has been negated by research conducted over a number of years and over several geographical regions. This research finds that customers do shift load, but the magnitude of shifting varies across business types.
The Electric Power Research Institute's (EPRI) StatsBank contains the measured responses of about 1,000 customers in the United States and the United Kingdom. Each of these customers has been on some form of a time-differentiated or RTP for several years, while others have been on some type of curtailable or interruptible rate. EPRI has estimated the between hours.2 Across all business segments, the estimated hour-to-hour elasticity of substitution within a day ranges from zero for some segments to values in excess of .30 for other segments.
Within the manufacturing sector, the highest elasticities are observed for electrically intensive customers. These include firms in the pulp and paper and primary metals industries. These customers have an average elasticity of .09. The lowest elasticities are observed for non-electrically intensive customers, such as firms in furniture manufacturing, printing, and publishing. These firms have an average elasticity of .04. The elasticities rise significantly if the customers have on-site generation. For example, the elasticity for electrically intensive customers with on-site generation is .15, compared with .09 for customers without on-site generation. The elasticity for the least electrically intensive customers is .07, compared with .04 for customers without on-site generation. For firms in the pulp and paper industry, the presence of on-site generation doubles the elasticity from .15 to .30.
It would be useful to know what types of customers are likely to shift more load. It also would be useful to know how much load relief can be expected from RTP, and whether the amount of load relief varies with a one-part design versus a two-part design. What segments are likely to shift more load? How applicable is the load shifting information contained in EPRI's StatsBank to utilities that are not included in that database? Is load-shifting information stable and reliable over time?
A related issue deals with market segmentation and targeting. Prior work suggests that customers with on-site generation, discrete production processes, and previous experience with interruptible tariffs are more likely to benefit from RTP. What is the best recruitment strategy for signing up these customers? How much education is needed to get customers acclimatized to the incentives provided by RTP?
A final issue relates to implementation strategy. Some utilities have only a handful of customers on RTP. Can better results be obtained? Some have argued that customers face real transactions costs when switching to a new rate structure, and those costs become a barrier to their joining RTP programs. Are these costs real or perceived? It would be useful to conduct a pilot program before proceeding with full-scale implementation.
Utility Barriers: Rates & Revenue Loss
Utilities have several issues regarding RTP. First, they are concerned about the revenue loss that can arise from RTP. This problem is acute with one-part designs, where revenue loss can arise if the rates are offered on a voluntary basis and customers who have inverse load shapes self-select themselves onto the real-time rate. The customers would lower their bills without shifting any load from peak to off-peak hours. They would benefit, but the utility and non-participating customers would lose. There would be a loss of revenue to the utility, without any reduction in its costs-resulting in a loss of earnings. The lost earnings then would have to be made up by charging other customers a higher price.
This concern can be addressed in three ways: by offering a two-part design; by making the rates mandatory; or by offering a true-up mechanism, ensuring that forecast revenues are recovered. It would be useful to conduct research on how serious is the potential revenue loss associated with RTP. Also, to what extent can it be offset by following a two-part design?
A second issue relates to the potential for gaming associated with two-part designs that require the establishment of a customer baseline usage (CBL). There is some anecdotal evidence that customers may have gamed the selection of their base load when signing up for market-based load curtailment programs. A similar concern may also apply to two-part RTP designs. This could be addressed through researching whether there is any empirical evidence from other states that customers have gamed the selection of their CBLs. If so, can better educational programs offset this problem?
A third issue deals with billing and settlement systems. Most existing systems are not capable of calculating bills based on hourly usage patterns. Modifications must be made by the IT staff, which often is overburdened with other duties. The only practical solution is to outsource this capability, and that often comes with a large price tag. It would be useful to research the cost of implementing billing and settlement systems that would enable RTP, and how these costs can be managed most effectively.
A fourth issue relates to the lack of integration of demand-side responses with system dispatch. Oftentimes, utilities are skeptical that the pricing program will produce real and tangible savings that they can count on. Some have more confidence in traditional load management programs involving direct load control of specific appliances. This concern is not limited to the United States. For example, ESKOM, the state-owned utility in South Africa, has 1,400 MW on RTP with a simultaneous load response capability of 350-400 MW for up to three consecutive hours. While RTP is set up on a day-ahead basis, customer response is not used to optimize the dispatch of the power system. Electricity prices are based on the Pool Output Price, and do not change in response to changes in customer demand that may be induced by RTP. The utility is not aggressively marketing the program for this reason. It hopes that once a competitive energy market has been created, with a functioning Power Exchange, RTP then will be able to play its proper role in system operations3
Fifth, there is a perception that RTP makes sense only during periods of wholesale price spikes. Thus, if wholesale prices are low (as they have been during the past year in the western states), RTP is not needed. What is often overlooked is that wholesale prices were high not too long ago, and that the sequential existence of low and high prices implies high price volatility. Customers who sign up for RTP will benefit when prices are low, and the existence of low prices during several hours of the year can be an inducement to participation in RTP. When a utility has a large number of customers on RTP, it creates flexibility for itself during high-price periods, when it can transmit a high-price signal to the customers and get customers to cut back on usage.
Finally, there is a perception that customers will not like RTP and will complain to the public utility commission. Utilities are concerned they may be trading off customer satisfaction for some questionable efficiency gains. Many cite the consumer revolt that was triggered in San Diego when electric bills doubled and tripled during the summer of 2000. However, it would be incorrect to regard the San Diego experience as a fair test of RTP. Customers were neither educated nor prepared for a doubling or tripling of their bills. Their prices did not vary hourly, as they would in an RTP setting, but on a monthly basis. No hourly meters were in place, and load profiling was used to compute their monthly bill.
Regulatory Barriers: The Morality of Pricing
- The concept of RTP originated with William Vickrey in 1971, when he wrote a groundbreaking article on "responsive pricing." Vickrey, who went on to win the Nobel Memorial Prize in Economic Sciences in the late 1990s, wrote that "the main difficulty with responsive pricing is likely to be not just mechanical or economic, but political." He felt that people shared the medieval notion of a just price as an ethical norm, and that prices that varied according to the circumstances of the moment were intrinsically evil. He opined prophetically:
In a similar vein, veteran energy analyst Eric Hirst noted recently, "the greatest barriers are legislative and regulatory, deriving from state efforts to protect retail customers from the vagaries of competitive markets."5
One of the key barriers among regulators relates to fairness and distributional concerns. Not every customer would benefit from a switch to RTP, and some customers would be at a disadvantage. Those who consume large amounts of energy during peak times would be disadvantaged because they would lose their subsidy from the other customers. Recently, Puget Sound Energy was unable to get regulatory approval for its tracker rate, which would have given customers a daily price signal layered on top of a traditional four-period time-of-use rate. There was a concern that many customer segments comprised of elderly customers or low-income customers would be harmed by the real-time component of the rate. Unfortunately, a policy that seeks to make no one worse off will imprison us in the status quo.
We need to search for a more forward-looking way of thinking. Policy makers have to devise a framework for balancing the competing demands of greater efficiency against the political pressures of special interest groups. What insights can be derived from the vast literature on welfare economics? According to the well-known Kaldor-Hicks criterion, if the gainers from a public policy can compensate the losers, the policy may be worth pursuing, even if the losers are not compensated.6
A related topic is whether market-based load curtailment programs can co-exist with RTP. Should customers who volunteer for a load curtailment program be excluded from receiving service on a real-time basis? Would this constitute double dipping? Or would it be a cost-effective way to obtain additional load shifting without having to make any additional investment in control technologies?
A third topic relates to a perception that RTP will seriously inconvenience customers because they cannot reduce peak usage or shift load from on-peak to off-peak periods. What is the best way to convince regulators that customers can indeed be trained to shift their loads from on-peak to off-peak hours? It would be useful to conduct a series of seminars and workshops for regulators, utilities, and prospective customers, in which customer case studies from other parts of the country would be featured.
Finally, a fourth issue arises from California's situation, in which the state has purchased power under long-term contracts, thereby seemingly eliminating the need for RTP. There appears to be no hourly variation in power prices under these contracts, but there is variation by pricing period. For example, blocks of peak power are much more expensive than blocks of off-peak power. Some have argued that RTP is now irrelevant in California because there is no hourly price variation in the wholesale price of power.
The argument that RTP does not apply in this situation has two weaknesses. First, it ignores the fact that the existence of long-term contracts has not eliminated the wholesale spot market for power. According to some sources, during peak periods, as much as 30 to 40 percent of the power may be traded in this market. During such times, RTP at the retail level would provide customers with the appropriate signal to conserve power usage. This would benefit the state, and, if customers can reduce peak load, it would also benefit the participating customers.
Second, it overlooks the fact that during times when the state has surplus power at long-term contracts, it is forced to dump this power on the wholesale market at below-cost prices. If customers were on RTP, they could be offered this power at the state's cost, which would be lower than the customer's average price. This lower price may stimulate growth in customer usage during off-peak hours, especially if the customers have been trained in how to increase power usage by rescheduling operations. It would lead to even greater usage if customers have enabling technologies on their premises. Customers and taxpayers would be better off, and so would the state.
Technological Barriers: The Cost of IT
Many technological barriers also impede the introduction and diffusion of RTP. The barriers are not intrinsically technological because the required technologies exist in today's marketplace. However, the market penetration of these technologies has been very limited due to their high capital costs. This in turn is due at least in part to their limited market penetration and to the barriers discussed in the previous sections. Technological barriers include the:
- lack of hourly metering equipment,
- lack of digital communication equipment to transmit hourly prices in real-time to customers,
- limited penetration of sophisticated energy management and control systems,
- even more limited penetration of time-flexible energy-using equipment that allows the energy to be stored during off-peak periods and released during on-peak periods, and
- small penetration of distribution energy resource systems.
An independent study of four industrial firms finds that self-generation significantly enhances customer responsiveness to RTP.7 Duke Power's experience suggests that customer response increases over time. Customer elasticities grew from .20 in 1995 to .25 in 1999.8 These findings illustrate the role of enabling technologies and provide some preliminary evidence on whether customer responses increase over time. They raise an important research issue: As customers learn how to take advantage of RTP and invest in new enabling technologies, do they display increasing responses, suggesting that long-run elasticities of substitution would be higher than short-run elasticities?
Of course, both utilities and state and federal regulatory commissions should make research on the barriers identified in this article a priority. Otherwise, like King Tantalus in Greek mythology, tomorrow's customers will continue to be starved of these benefits, whether they bend low to drink the water from the stream or reach high to eat the fruit from the tree.
- These estimates were derived from data gathered through customer interviews. Customers ranked various pricing products, and the resulting rankings were subjected to conjoint analysis. One important caveat is that, in all but one interview, the data were derived from a standard model of customer choice that enforces the same preference functions on all customers. Once that assumption is relaxed, by using a more advanced model known as the mixed logic model, important information on the variation in customer preferences within segments is revealed. Thus, on average, customers may display a preference away from RTP and be willing to pay a higher flat price than go on a real-time price. However, several customers within each segment may have a preference toward RTP, and these customers can form the target population for the RTP program.
- This elasticity measures the percentage change in the ratio of usage in two hours that is induced by a one percent change in the corresponding price ratio. The elasticity is a negative number, but in popular writing the negative sign is dropped. Thus, a higher elasticity number indicates that customers are more easily able to substitute usage between hours.
- Personal correspondence with Pieter Brand, RTP Product Manager and Gold Sector Specialist, Johannesburg, South Africa, April 5, 2002.
- Vickrey, William, "Responsive pricing of public utility services," Bell Journal of Economics and Management Science, 2, 1971, pp. 337-346.
- Eric Hirst, "Price-Responsive Demand in Wholesale Markets: Why Is So Little Happening?" The Electricity Journal, May 2001.
- Of course, if the compensation is actually paid, then no one is worse off, and the program is Pareto Optimal.
- Nainish K. Gupta and Albert L. Danielsen, "RTP: Ready for the Meter? An Empirical Study of Customer Response," Public Utilities Fortnightly, November 1, 1998.
- Peter M. Schwarz, Thomas N.Taylor, Matthew Birmingham, and Shana L. Dardan, "Industrial Response to Real-Time Prices for Electricity: Short-Run and Long-Run," Economic Inquiry, forthcoming 2002.
Industry research yields new insights on RTP.
The benefits of real-time pricing (RTP) arewell known, and include the mitigation of price volatility and market power in wholesale markets.1 RTP promotes economic efficiency by giving customers a strong incentive to lower usage when hourly prices are high. This serves to reduce the threat of power outages. As Californians found out the hard way in the winter of 2001, blackouts are an inefficient way of rationing customer demand because they affect all customers equally, regardless of the value they place on electricity. They can impose significant economic costs, since customers often place a substantially higher value on electric service than the amount they pay in electric rates. The Electric Power Research Institute (EPRI) has estimated that the annual cost of power outages for California businesses ranges between $12 billion and $18 billion, and the corresponding cost to U.S. businesses ranges between $104 billion and $164 billion.2
In another study, several researchers quantified the benefits of RTP to California by conducting a counterfactual experiment (see Figure 2).3 They showed what would have occurred in the year 2000 had the state implemented RTP. The study assumed that all customers above 200 kW load were eligible for RTP, and assessed impacts under three scenarios of customer participation, ranging from 25 percent to 50 percent to 100 percent. Figure 2 shows what would have occurred if customers had displayed a moderate amount of price responsiveness: half of them would have displayed no price responsiveness, and the other half would have displayed a degree of responsiveness consistent with experience elsewhere in the United States and the United Kingdom. If all customers had chosen to participate, peak demand would have fallen by 2.8 percent, peak prices by 20.4 percent, and seasonal electricity costs by 6 percent.
Even higher results would be obtained if customers had displayed a greater degree of price responsiveness. To quantify these benefits, the study examined another scenario in which two-thirds of all customers displayed a degree of price responsiveness equal to the average responsiveness that has been observed in other utilities, and the other third of them displayed a greater degree of price responsiveness than has been observed with customers who are equipped with enabling technologies such as self-generation. If all customers participated in RTP in this scenario, peak demand would have fallen by 6.9 percent, peak hourly prices by 56 percent, and seasonal electricity costs by 13.2 percent.
While examining the effects of RTP in the Carolinas, Tom Taylor of Duke Power and researchers at the University of North Carolina have found that RTP has induced load reductions of about 70 MW, which translates into a long-term savings of some $2.7 million per year.4 In a similar vein, MIT Professor Schweppe's definitive work demonstrated the economic efficiency that would result from RTP.5 Schweppe and his co-authors conjectured that if the various "publics" involved in RTP could be convinced of its many benefits, implementation would flow automatically. California's inability to implement RTP in the summer of 2001 shows that, in the real world, implementation never can be expected to flow automatically.
Lack of interval metering traditionally had been considered the biggest barrier to implementing RTP. To overcome this barrier, the California Assembly passed a bill, AB 29X, allocating $35 million for installation of real-time meters on the premises of all customers with a demand in excess of 200 kW. About 15,600 real-time meters were expected to be installed or upgraded by June 2002, affecting approximately 30 percent of peak demand in the state.6 This should have paved the way for implementation of RTP, but it did not.
The California Energy Commission (CEC) filed an RTP tariff with the California Public Utilities Commission (CPUC) in June 2001.7 This tariff was modeled after Georgia Power's approach. Georgia Power runs what may be the world's largest and most successful RTP program. Using a two-part design, Georgia Power has been able to attract about 1,650 customers to RTP, representing 40 percent of the 4,100 customers larger than 250 kW. These customers represent about 5,000 MW of total load, and 80 percent of them are engaged in industrial activities. Through its RTP program, Georgia Power has been able to achieve a peak demand reduction of up to 17 percent on critical days.
The CPUC rejected the CEC's proposal in a draft decision in July 2001. The CEC revised its proposal, and submitted a simpler approach to the CPUC. The following month the CPUC rejected this approach, and directed the utility distribution companies to make their own filings. These were submitted later in August. As of this writing, the CPUC still has not ruled on these filings, but the CPUC is considering a new Order Instituting Rulemaking (OIR) on the issue.
California's experience with RTP draws attention to the very real barriers that face utilities and states seeking RTP implementation. Last August, we conducted a series of telephone interviews to assess the experience of utilities around the country with RTP. Nine important lessons emerged from our interviews.
Lessons Learned from Utility RTP Programs
RTP programs have been around for more than 15 years, and over 30 utilities have implemented some type of program, though many of these have been experimental. However, relatively few utilities have sizable RTP programs. Figure 3 lists some of the leading RTP programs. The following are lessons we have identified from our utility interviews:
-AF and MM
- RTP programs can offer significant load shifting benefits, but most of the load response comes from relatively few customers.
Customers pay attention to prices, and shift loads in response to price movements. It is not unusual to find examples of peak load shifts of 15 to 20 percent or more. However, most of the load shifting comes from a relatively small group of customers, and many customers do not shift load at all.8 One utility found that it had roughly the same response in 2001 with 59 customers as it did in 2000 with over 100, because the non-responsive customers dropped off the rate. Another found that only three of its 14 customers have done any "significant" shifting. Some utilities found that almost no customers are shifting. Customers had to be allowed to drop off their market-based RTP program by a California utility after PX prices shot up, because customers didn't know how to shift. A Canadian utility has found that even though it has roughly 25 customers on RTP, only one ever has really shifted load, and even that customer could not sustain this shifting.
Even among price-responsive customers, response can vary significantly over time.9 Presumably, customers have more scheduling flexibility at certain times than at others.
- Certain types of customers are more likely to respond to RTP.
The following groups of customers repeatedly have been shown to be more responsive:
- customers with on-site generation, such as hotels and large office buildings in the commercial sector, and pulp and paper mills in the industrial sector;
- customers with non-continuous (discrete) production processes, such as municipal water pumping and cement production; and
- customers who have previously been on interruptible rates, such as universities.
- A variety of customers can respond to prices.
Customers with the incentive to shift load can find innovative ways to do so. Therefore, RTP programs should not exclude customers simply because they are not in the groups most likely to respond to prices. Price responsive customers at one utility include office buildings and grocery stores. Another utility has a price responsive hospital that changes its chiller use in response to hourly prices.
- Customers join RTP to save money.
Customers join real-time pricing programs to save money. While it may seem obvious that customers with a choice of rates would choose the one in their best interest, it is also true that customers joining an RTP program have certain expectations about the rate-producing savings. This appears to be true even for customers who do not plan to shift load significantly in response to price variation. Thus, when the overall level of RTP prices (not just the price volatility) increases, customer satisfaction with the program decreases. In some cases, customers return to embedded cost-based rates in response to higher overall prices. For example, all of BC Hydro's customers dropped off its RTP program within a year after market prices increased.10
This finding also holds true for residential customers. Electricit‚ de France (EdF) has had over 120,000 residential customers on a simplified RTP rate called Tempo since 1996. Its surveys show that customers join the rate, and are satisfied with it, because of bill savings. EdF has spent a significant amount of time finding customers "suitable" to the rate-e.g., those with the ability to shift and save money and to whom EdF can offer peak reductions.11 The program features two daily pricing periods, on-peak and off-peak. It also features day-of-the-year pricing. The year is divided into three types of days. The blue days are the most numerous (300) and least expensive; the white days are the next most numerous (43) and mid-range in price; and the red days are the least numerous (22) and the most expensive.12
EdF does not offer a fixed calendar of days, but customers can know what color will take effect the next day by checking a variety of different sources:
Commercial and industrial customers at PG&E, Southern California Edison, Virginia Electric Power Company (Vepco), and Niagara Mohawk also indicated that bill savings, as well as the related "control over costs," are a major reason for joining RTP programs.13
- consulting the Tempo Web site, www.tempo.tm.fr,
- subscribing to an email service that alerts them of the colors to come,
- using Minitel (a data terminal particular to France, sometimes called a primitive form of Internet),
- using a vocal system over the telephone, or
- checking an electrical device (Compteur Electronique) provided by EdF that can be plugged into any electrical socket.
While not as well documented, the implication that RTP programs have "free riders" that join to save money is a serious issue. One utility has a one-part rate that is designed to be revenue neutral with its time-of-use (TOU) rate. Since one-part rates are often designed to be revenue neutral for the class as a whole, they produce "winners" and "losers"-e.g., customers who can switch to RTP and expect to save money without necessarily shifting load. Customers on this utility's market-based rate did not seem to know how to respond to prices (because they had not planned to shift) when PX prices increased significantly. As a result, the utility ended up allowing customers to move back to the TOU rate, despite the fact that customers had a contractual obligation to remain on the RTP rate.
Even customers who can and do respond to prices become less satisfied when the overall level of prices increases. When prices increased on average by a penny per kWh, Georgia Power's customers asked the state PUC for relief. The utility modified the rate by slightly lowering hourly prices for these customers.
- Customers do not like unmitigated price volatility.br /> Customer satisfaction with RTP programs increases if they feel that their price exposure is limited.14 Greater price volatility can lead to higher bills. This is true even for price responsive customers, because they may have periods when their response is constrained because of business conditions. EdF has found that customers on the RTP rate generally postpone laundry and other tasks on high-priced days, but if three high-priced days come in a row, customers-especially those with young children-will do their laundry and pay the high prices.15
To guard against such situations, many utilities have integrated some type of risk mitigation feature into their RTP programs. For instance, Vepco's program sets an upper limit on the number of high-priced days and a minimum limit on the number of low-priced days.16 A study at Long Island Lighting Company also found that limiting risk was likely to increase program participation.17
To limit price risk for its RTP customers, Southern California Edison limits its program's highest-priced hour to $3.00/kWh. In the United Kingdom, customers can purchase contracts for differences that essentially provide them with fixed prices over a specific period of time.
In part to reduce customers' exposure to extreme price volatility, most U.S. utilities with RTP programs recently have developed two-part RTP rates. These rates allow a portion of customers' loads to be protected from price volatility. The Tennessee Valley Authority (TVA), a utility that has had a one-part rate for 15 years, has developed a two-part rate and has two very large industrial customers on that rate. Some utilities with two-part RTP rates also offer risk protection products similar to those offered in the United Kingdom. These products are intended to provide customers with relief in the event that high prices coincide with periods in which their ability to shift is limited. These products generally apply to a specific period of time, and the utilities that offer them still offer incentives to reduce load during high-cost hours.
- RTP programs create revenue stability issues for utilities and bill stability issues for customers.
Because RTP programs encourage decreased usage during high-priced hours, utilities face the risk of under-collecting revenues. This is particularly true with one-part rates, which generally include fixed costs in the hourly energy price. The pilot RTP programs have shown that it is actually not hard for utilities to lose money on RTP. The reason most RTP programs are still "experimental," and even permanent ones often have very few customers on them, is at least in part due to this issue.18
Utilities are choosing two-part RTP rates in part because they have lower risk of under-collecting fixed costs. One company has chosen to apply its "adder," which provides a contribution to fixed costs, only on net incremental (vs. decremental) energy to help avoid this problem. Other utilities with two-part rates, such as Aquila and Public Service of Oklahoma, use variable adders, which tend to result in smaller adders being paid to customers for when they reduce usage below their customer baseline usage (CBL) than those paid by customers for incremental use.
- With two-part RTP rates, utilities and customers often prefer simpler designs.
Georgia Power and Duke, which have among them the oldest and most successful two-part RTP programs in the country, began setting CBL with 8,760 hour load profiles. Both utilities have moved to simpler CBLs for most customers. The utilities report that they believe the simpler CBLs make sense for most customers, and that customers tend to find them less confusing.
- RTP programs can be successfully combined with interruptible programs.
In some cases, such as TVA, the utility's RTP rate applies only to interruptible power. While high prices alone would in theory encourage customers to decrease load at critical times, the interruptible nature of the power ensures that TVA has a certain load management resource. TVA does offer a firm RTP rate, but as of this writing, it had no takers for that rate.19 More typically, utilities allow interruptible customers on RTP programs, and require them to interrupt to the level of firm demand during interruptible periods. Customers not interrupting may have to pay a penalty for non-compliance, in addition to purchasing energy at the hourly price. Some utilities offer customers the option of buying energy during interruption periods, but reduce the size of the interruptible discount for customers choosing this option.
Typically, utilities offer interruptible customers RTP because these customers have demonstrated an ability to shift load, and may be able to provide valuable price response during periods of high prices that are outside interruptible periods. None of the utilities that offered RTP to its interruptible customers felt that RTP negatively impacted responses from the interruptible rate. As one respondent explained it, "The rates are so different; interruptions only occur a few times a year, [but customers can respond to prices every hour]."
- Customer education is key.
This is the single most important lesson. Customers need to be educated repeatedly for a couple of reasons. First, there is turnover at their companies, so the person who best understands the RTP program might not be there in the future. Second, customers tend not to focus on RTP when prices are low, and begin to pay attention only as prices increase. To address its educational needs, Georgia Power holds annual meetings of RTP customers all over the state. These meetings are very well attended, and the utility believes the education program has definitely paid off in terms of customer satisfaction.
- Steven Braithwait and Ahmad Faruqui, "The Choice Not to Buy: Energy Savings and Policy Alternatives for Demand Response," Public Utilities Fortnightly, March 15, 2001, 48-60.
- EPRI's Consortium for Electric Infrastructure for a Digital Society (CEIDS), "The Cost of Power Disturbances to Industrial and Digital Economy Companies," Palo Alto, California, July 29, 2001.
- Ahmad Faruqui, Hung-po Chao, Vic Niemeyer, Jeremy Platt, and Karl Stahlkopf, "Getting out of the dark," Regulation, Fall 2001, pp. 58-62.
- Peter M. Schwarz, Thomas N.Taylor, Matthew Birmingham, and Shana L. Dardan, "Industrial Response to Real-Time Prices for Electricity: Short-Run and Long-Run," Economic Inquiry, forthcoming 2002.
- Fred C. Schweppe, Michael C. Carmanis, Richard D. Tabors, and Roger E. Bohn, Spot Pricing of Electricity, Kluwer Academic Publishing, 1988.
- Michael R. Jaske and Arthur H. Rosenfeld, "Developing Demand Responsiveness in California's Energy Markets," 76th Annual WEA Conference, July 2001.
- "Petition of the California Energy Commission for Modification of Decision 01-05-064 By Proposing a RTP Tariff," June 21, 2001.
- Kathleen King, "The Impact of Real-Time Pricing: Evidence from the British Experience," Proceedings: 1994 Innovative Electricity Pricing, EPRI TR-103629, 257-267. At the time of the study, all of the customers on Midlands Electricity's rate were 1 MW or larger.
- According to BC Hydro's Allan Chung, the RTP rate was negotiated with customers when market prices were lower than tariffs. It was not symmetrical, and some customers could apply for reduced customer baseline usages (CBLs). From the onset, customers viewed the program as a way to lower bills. Thus, when prices went above market, many customers had price exposure arising from their reduced CBLs and could not benefit from the rate because of the lack of symmetrical pricing. They all left the rate. But the results could have been different, had the rate been structured differently.
- J. Cubille and P. Valentin, "Tempo Customers: Their Reaction to a New Tariff Option," presented at the Unipede Conference on Customers and Markets, Lisboa, June 1998.
- Aubin et al. (1995).
- Juliet C. Mak and Bruce Chapman, "A Survey of Current Real-time Pricing Programs," The Electricity Journal, August/September 1993, page 62.
- EPRI, Real-Time Pricing QuickStart Guide, TR-105045, August 1995, page 24.
- Cubille and Valentin, at section 3.3.
- Juliet C. Mak, and Bruce Chapman, "A Survey of Current Real-time Pricing Programs," The Electricity Journal, August/September 1993, page 62.
- Yannis Takos, Mitchel Horowitz, and Ellen Ford, "Gauging Customer Acceptance for Various Real-Time Pricing Configurations," Proceedings: 1994 Innovative Electricity Pricing, EPRI TR-103629, 138-144. 18 Glen Weisbrod and Ellen Ford, "Market Segmentation and Targeting for Real Time Pricing," Proceedings: 1996 EPRI Conferences on Innovative Approaches to Electricity Pricing, EPRI, TR-106232, 14-1. 19 Personal correspondence with Tom Miraldi, Cost and Price Analysis, TVA, April 4, 2002.
Articles found on this page are available to Internet subscribers only. For more information about obtaining a username and password, please call our Customer Service Department at 1-800-368-5001.