A survey of time-of-use (tou) pricing and demand-response (dr) programs july 2006

A SURVEY OF TIME-OF-USE (TOU) PRICING AND DEMAND-RESPONSE (DR) PROGRAMS July 2006
Submitted By: Energy & Environmental Economics
353 Sacramento Street, Suite 1700 San Francisco, CA 94111 (415) 391-5100 The U.S. Energy Environmental Protection Agency (U.S. EPA) commissioned this report to support state public utility commissions in their expanding use of energy efficiency (EE) in their states. The report is part of the EPA-State Energy Efficiency and Renewable Energy (EERE) Project index.htm). The EPA-State EERE Projects are a joint initiative between the U.S. EPA, the National Association of Regulatory Utility Commissioners, and individual state utility commissions designed to explore approaches that deliver significant energy cost savings and other benefits through greater use of energy efficiency, renewable energy, and clean distributed generation. TABLE OF CONTENTS Executive Summary . 1 1. Introduction. 6 1.1 Rate design basics . 6 1.2 Types of rates and programs . 9 1.3 Energy efficiency, conservation, and rate design . 16 1.4 Purpose of this survey. 18 2. A general assessment of time-varying pricing and DR programs . 21 2.1 Efficient pricing of electricity service. 21 2.2 Load shifting and conservation effects of time-varying pricing. 24 2.3 Efficient rationing of capacity . 28 2.4 Load relief for resolving system and/or local capacity constraint . 30 2.5 Revenue collection. 32 2.6 Mandatory vs. optional tariffs. 32 2.7 Implementation issues. 34 A. Information requirements. 34 B. Implementation cost . 36 C. Competing stakeholder interests . 37 3. Evaluating specific types of time-varying pricing and DR programs. 38 3.1 Program types . 38 A. Attributes of Price-Based DR Programs . 40 B. Attributes of Quantity-Based DR Programs. 42 C. Attributes of Hybrid DR Programs . 45 3.2 Program evaluation . 46 A. Criteria . 46 4. Survey of real-world designs . 49 4.1 Sample selection and data sources. 49 4.2 Findings. 50 A. Overview: residential DR programs and tariffs . 50 B. Overview: large customer DR programs and tariffs . 52 C. Residential TOU pricing . 54 D. Residential critical peak pricing. 56 E. Residential TOU/CPP hybrid programs . 59 F. Real time pricing. 61 H. Residential direct load control . 64 G. Residential demand subscription service . 67 I. Non-residential curtailable/interruptible service (CIS) . 69 J. Residential pricing/load control hybrid programs. 70 K. Non-residential pricing/load control hybrid programs. 71 5. Conclusion . 75 6. References. 77 Appendix 1: Source Material on Empirical Estimation of Customer Response to TOU Pricing and DR Programs . 83 Appendix 2: Tables: Details of Demand Responsive Rates and Programs . 87 A SURVEY OF TIME-OF-USE (TOU) PRICING AND DEMAND-RESPONSE (DR) PROGRAMS Executive Summary
Removing the barriers to energy efficiency investment requires tariff- and non- tariff-based solutions. Tariff-based solutions typically entail providing price incentives to electricity consumers to adopt energy saving measures (e.g., efficient lighting) and to alter consumption patterns (e.g., dynamically increasing thermostat settings in the summer). Non-tariff-based solutions are many, encompassing consumer education (e.g., public information on TV), tax incentives (e.g., tax credit for efficiency investment), financial assistance (e.g., efficient appliance rebate and low-cost financing), and partnership (e.g., a combined heat and power project developed by a utility, a local business and the city government). This report provides information on tariff-based solutions. It aims to assist regulators, utilities, customers, and other stakeholders who are deciding what if any time- varying pricing or other types of demand-response (DR) programs should be added to the utility's existing tariffs, and how these rate options might affect energy efficiency goals. To this end, this report describes the results of a survey of rate options used by electric utilities to flatten customer demand during peak hours. To aid understanding of these DR programs, the report introduces rate design basics (Section 1), discusses a number of key regulatory concerns in ratemaking that apply to these rates (Section 2), describes and evaluates each type of rate option in detail (Section 3), and presents the survey results The survey, conducted in 2005, examined the tariffs of 65 large utilities, including 50 in the United States and 15 overseas, and provides a snapshot of the prevalence, types, and characteristics of peak-shaving rates and programs, for both residential and commercial/industrial customers. The study includes tariffs that encourage customers to reduce on-peak usage by charging on-peak prices during high demand / cost hours that are higher than off-peak prices during low demand / cost hours. These tariffs are real- time pricing (RTP) with hourly prices that vary daily, time-of-use pricing (TOU) with time-differentiated rates that may vary seasonally, and critical peak pricing (CPP) with very high signals that are sent to a customer by the utility only during critical peak hours. It also include tariffs that allow utilities to limit the quantity of customer usage directly, such as curtailable/interruptible service (CIS), direct load control (DLC), and demand subscription service (DSS). Under these tariffs, a customer self-selects the portion of his/her kW loads that can be curtailed by the utility during hours of high cost or supply Rate design is a complex process that must take into account multiple objectives and multiple stakeholders. Typically, there are three dominant priorities in determining whether a rate or program will be adopted: (1) meeting utility revenue requirements; (2) fair apportionment of costs among customers; and (3) economic efficiency. Other important goals include revenue stability for the utility and rate stability for customers, transparency and ease of implementation, and affordability for low-income customers. Rates designed to encourage energy efficiency, reduced emissions, and peak load management must also be consistent with these other objectives to be successful and to achieve stakeholder acceptance. While both EE and DR can be complementary in an overall demand-side management (DSM) strategy, their fundamental goals are quite different. Where EE programs focus on reducing a customer's overall energy consumption, time-varying pricing and DR programs aim to reduce peak loads and the associated capacity requirement on the utility system. To be sure, conservation may occur as a result of peak load reduction; however it is usually a small component of a DR program's intent. In some cases the net effect of DR is actually a small increase in total energy consumption. In contrast, energy conservation tends to be largest in programs that raise customer energy awareness and provide feedback on usage and enabling technology to reduce consumption. This report's findings aim to assist regulators who are concerned with carefully coordinating EE and DR policies to avoid conflicting outcomes. The findings of the survey regarding rate options found in both standard tariffs and pilot programs are summarized below. Table 1 shows that for residential customers, the most common pricing program is TOU, offered by 82% of the U.S. utilities surveyed. The most common quantity-limiting rate is DLC, offered by 30% of the U.S. utilities surveyed. A major contrast found between U.S. and international utilities was that a substantial number of European and Japanese utilities offered DSS service as a default rate option, while among U.S. utilities surveyed it has been offered only once, as an experimental program that was subsequently discontinued. Given that DSS can offer substantial benefits from both DR and EE perspectives, this rate option may merit more attention from U.S. utilities and regulators. Table 1. Number of utilities offering residential DR programs (includes both current and historic programs) Utilities Abbreviations: TOU = time of use. RTP = real time pricing. CPP = critical peak pricing. DSS = demand subscription service. DLC = direct load control. CIS = curtailable or interruptible service. Hybrid includes demand bidding, capacity buyback, market-based load reduction, and quantity control with buy-through, in addition to combinations of the price-based and quantity-based programs in this table. Table 2 shows the results for large commercial and industrial customers, for which only U.S. utilities were surveyed. The most common pricing program was again TOU, offered by 96%, and the most common quantity-limiting program was curtailable/interruptible service, offered by 82%. Almost half of the utilities surveyed offered RTP as either a standard option or an experimental rate. Table 2 Number of US utilities with current or historic DR programs for large non­residential customers Utilities Abbreviations: TOU = time of use. RTP = real time pricing. CPP = critical peak pricing. DSS = demand subscription service. DLC = direct load control. CIS = curtailable or interruptible service. Hybrid includes demand bidding, capacity buyback, market-based load reduction, and quantity control with buy-through, in addition to combinations of the price-based and quantity-based programs in this table. To help situate the survey results in the larger ratemaking process, Section 5 of this report addresses the following policy questions: • Are time-varying pricing and DR programs conceptually sound from the standpoint of economic theory? Yes, they are. • What kind of benefits can they deliver to the utility and its customers? When properly designed and implemented, they can deliver benefits both to the utility's customers and to shareholders. • What are these programs' likely load shifting (i.e., flattening a customer's hourly load pattern) and conservation (i.e., reducing a customer's total kWh consumption) effects? While time-varying pricing and DR programs can reduce peak load and induce load shifting, their conservation effect is mild. • Do they offer load relief that can resolve system and/or local capacity constraints? • What real-world applications of these rates and programs are found in the electricity industry? Real-world applications of some types of time-varying pricing and DR programs are numerous, indicating that they are not excessively costly or difficult to Notwithstanding the above answers, the information presented in this report should not be interpreted as the final words on DR program design and evaluation. This is because the success of a DR program, as in any case of rate design, ultimately depends on the details of planning and implementation. 1. Introduction
1.1 Rate design basics
To assess the role of electricity rates in promoting public policy goals such as energy efficiency and demand response, it is helpful to start with a short review of the goals and methods of rate design. Rate design is a complex process that must take into account multiple objectives. There are usually three main priorities: (1) meeting utility revenue requirements; (2) fair apportionment of costs among customers; and (3) economic efficiency (Bonbright, 1961; Philips, 1988). Other important regulatory and legislative goals include provision of stable revenues for the utility and stable rates for customers; simplicity of understanding and ease of implementation; encouragement of effective load management; promotion of social equity in the form of lifeline rates for people with low incomes; and promotion of environmental sustainability in the form of rates that encourage reduced energy use and lower emissions. These various goals are weighted differently at different times and in different jurisdictions, and ratemaking often involves making tradeoffs among them. In a typical ratemaking process, the utility's total revenue requirement is determined, then costs are allocated to different customer classes, and finally a rate structure is developed to recover those costs. Until the early 1970s, rate structures were based on the principle of average-cost pricing, in which customer prices reflected the average costs of serving their customer class. Volumetric, time-invariant rates did not take into account the higher cost of serving customers whose usage was concentrated in peak periods. The most prevalent rate design was declining block pricing, in which the price of the initial increment of usage was set high enough to recover the utility's fixed costs, and successive increments were priced more cheaply, representing mostly fuel costs. This pricing scheme encouraged excessive peak usage that required utilities to continuously build new capacity; however, in a period with increasing economies of scale in plant construction and stable or falling fuel prices, system expansion led to declining average costs, and stakeholders remained content with average-cost based rates (Hirsh, 1999; Hyman et al., 2000). In the 1970s, the underlying economics of the electricity industry changed dramatically, with rising construction and fuel costs that were not offset by technological improvements. At the same time, the public became increasingly concerned with the environmental consequences of power production. These trends led to a sea-change in pricing philosophy, with marginal-cost pricing being widely adopted as the underlying principle of rate design. Marginal costs are the change in total costs required to produce one additional unit of electricity. According to economic theory, the most efficient outcome occurs when prices are equal to marginal costs, resulting in the maximum societal benefit. In a period of rising marginal costs, rates based on marginal costs more realistically reflect the cost of serving different customers, and provide an incentive for more efficient use of both economic and natural resources (Bonbright, 1961; Kahn, 1970; Huntington, 1975; Joskow, 1976; Joskow, 1979). Despite its theoretical advantages, there are significant barriers to fully implementing marginal-cost pricing in electricity, especially at the retail level. In contrast to other commodities, the necessity for generation to match load at all times means that outputs and production costs are constantly changing, and it can be very complicated and expensive to convey these costs as real time "price signals" to customers, especially residential customers. As a practical matter, application of marginal cost principles to retail rates frequently involves the use of approximations and proxy values, which can differ widely from jurisdiction to jurisdiction. Whether marginal or embedded costs are used, the determination of costs in ratemaking begins with the classification of cost components. Costs are classified according to function (generation, transmission, and distribution); type (fixed and variable); and category (capacity, energy, and customer). Capacity costs are based on the investment required to meet customer kilowatt demand in the generation, transmission, and distribution systems, and are typically measured in $/kW. Energy costs are the fuel and operation and maintenance costs used in supply energy, and are typically measured in $/kWh. Customer costs are those associated with providing customers access to the system, including service connections, metering, and billing, and are typically measured In the case of marginal-cost ratemaking, these costs are calculated for the provision of an additional unit of service – the next kilowatt of demand, kilowatt-hour of energy, or customer. Crucially, marginal costs can vary widely according to location and time. For example, the marginal capacity cost of adding an additional kilowatt of generation capacity may be very small (even zero) in the middle of the night, when demand is low and capacity is readily available, but it may be very high during a peak period on a summer day when system generation is operating at its limit. Marginal capacity cost of transmission or distribution may be low in locations where there is always available capacity, and high in other locations where capacity is often constrained. Similarly, marginal energy costs are typically low during off-peak periods when base- load generating units with low fuel and operating costs are in use, but high when peaking plants are employed. In states with wholesale electricity markets and locational prices, marginal energy and capacity costs are based on market prices; in states that still have vertically integrated systems, marginal costs must be based on other measures, such as production simulation computer models or proxies such as value of service or the cost of a combustion turbine. 1.2 Types of rates and programs
There are many possible rate designs that can fit within a utility's overall rate structure; the number of individual rates offered can range from a handful for smaller utilities to more than 100 for the largest. There is no simple correlation between the costing methodology employed and the types of rates offered to customers. For example, the most prevalent rate designs for residential customers today do not vary with location or time, even though the underlying marginal-cost calculations may take location and time into account. Rate designs can be categorized in different ways: • How does the customer price change as a function of time? Rates will be either time- varying or time invariant, with important variations of each (see further discussion • How many billing components does the rate have? Rates are generally either 1-part, 2-part, or 3-part. 3-part rates usually have separate capacity (demand), energy (volumetric), and customer components, which map to the comparable cost components for that customer or class. Many large commercial and industrial customer rates are 3-part. On the other hand, many residential rates are 1-part, and all cost components are folded in to a comprehensive volumetric energy charge; some are 2-part, with a flat customer charge plus a volumetric rate. An additional billing component of some rates is the customer "baseline," an expected monthly usage level that is charged at a certain price level, with increases above or decreases below this level being charged or credited at a different price level; • Is enrollment in the rate voluntary or mandatory, and if voluntary, is it the default option from which the customer may choose to "opt out", or must the customer deliberately choose to "opt in"? In some cases, enrollment is voluntary but some customer action once enrolled is mandatory. For example, customers may choose a curtailable rate option that provides lower prices at ordinary times but compels them to reduce their usage to a certain pre-arranged level when the utility announces a curtailment. Since this action is mandatory, failure to comply is severely penalized; • How are scarce resources such as peak capacity allocated to customers? This is done either through price-rationing, in which customers adjust their usage in response to price signals, or through quantity-rationing, in which the utility limits the amount of usage by pre-arrangement, as in the case of the curtailable rate described above. Three basic varieties of time-invariant rates are shown below in Figure 1.1. • Flat rates use the same per-unit price for all units consumed, either kWh of electricity or kW of demand. For example, in the figure the rate is $0.20/kWh for all usage. If the customer uses 300 kWh, the customer charge will be $60. • Declining block rates have per-unit prices that decrease for each successive block of units. For example, in the figure the rate is $0.30/kWh for the first block of 200 kWh, decreasing to $0.20/kWh for the next block of 200 kWh, and then to $0.10 for any additional usage. If the customer uses 300 kWh, the customer charge will be 200 x 0.30 = $60 for the first 200 kWh, plus 100 x 0.20 = $20 for the next 100 kWh, for a • Inverted block rates have per-unit prices that increase for each successive block of units. For example, in the figure the rate is $0.10/kWh for the first block of 200 kWh, increasing to $0.20/kWh for the next block of 200 kWh, and then to $0.30 for any additional usage. If the customer uses 300 kWh, the customer charge will be 200 x 0.10 = $20 for the first 200 kWh, plus 100 x 0.20 = $20 for the next 100 kWh, for a 1 A variant of inverted block pricing is tiered rates which are used in California for most residential rates. In this case, customers are assigned a monthly baseline usage level that varies with the customer's weather zone location. Usage at or below the baseline amount is charged at the base tier price, while usage above the baseline amount is charged at successively higher tiered prices. (b) Declining Block Figure 1.1. Three types of time-invariant rates. Time-varying rates are described extensively throughout the remainder of this paper. Three basic varieties of time-varying rates are shown in Figure 1.2: • Time-of-use (TOU) rates have different per-unit prices for usage during different blocks of time. The definition of TOU periods differs widely among utilities, based on the timing of their peak system demands over the day, week, or year. TOU rates sometimes have only two prices, for peak and off-peak periods; sometimes a shoulder or partial-peak rate is added. In some cases these prices apply year-round, and in others they differ by season. The figure shows a TOU rate with an on-peak price of $0.30/kWh for all usage during the hours 12 noon to 6 PM, and an off-peak price of $0.10/kWh for usage at all other times. TOU rates are the most prevalent time- varying rate, and are often mandatory for large commercial and industrial (C & I) customers. Note that TOU rates require meters that register usage during the different usage blocks. The additional cost of providing and operating these meters is often reflected in a separate customer charge. • Critical peak pricing (CPP) rates have high per-unit rates for usage during periods that are designated to be critical peak periods by the utility. Unlike TOU blocks, the days in which critical peaks occur are not designated in the tariff, but dispatched on relatively short notice as needed, for a limited number of days during the year. CPP rates can superimpose the critical peak price on other types of rates, for example on flat rates or TOU rates. The figure illustrates a critical peak rate (shown as a dashed line to indicate that it may or may not be dispatched on a given day) that is superimposed on a flat rate. A well-known CPP rate design used by Gulf Power is described in Section 4, but in general CPP rates are uncommon in the U.S. • Real-time pricing (RTP) rates vary continuously over time in a way that directly reflects the wholesale price of electricity, rather than at pre-set prices as in virtually all other rate designs. Several different varieties of RTP exist in the U.S. and elsewhere, as described in Section 4. Most frequently, RTP rates provide different prices for each hour of the day every day of the year, and these prices are made known to customers one day in advance. The figure illustrates hourly RTP prices. (b) Critical Peak Pricing (c) Real-time Pricing Figure 1.2. Types of time-varying pricing. In addition to the rate designs described above, other rate options are employed by utilities (or independent system operators) to reduce peak loads and improve system reliability, and to provide choices and cost control opportunities for customers. Often these are not distinct rate designs, but are in the form of riders or special incentives available to customers on existing rates. In this survey, the term "demand-response program" is used broadly to describe these options, as well as the time-varying rates described previously. DR programs considered here include direct load control, interruptible and curtailable rates, demand subscription, and demand bidding. They are described in Section 3. 1.3 Energy efficiency, conservation, and rate design
Here we briefly consider the interaction between the objectives of energy conservation and energy efficiency on the one hand, and those of time-varying pricing and DR programs on the other.2 The goal of energy efficiency is to reduce energy consumption per unit of useful output (for example, the electricity required to keep food in a refrigerator cold), and the goal of energy conservation is to reduce the absolute amount of energy consumption. Since emissions from electricity generation are directly 2 A useful review of this subject is York and Kushler (2005) whose principal findings are (1) the quantitative relationship of EE and DR is poorly studied and requires more research, and (2) while EE and DR can be complementary, they can also conflict, and require careful programmatic integration by utilities related to how much fuel is consumed, from an environmental standpoint energy efficiency and conservation are very important. Traditionally, energy efficiency and conservation have been promoted by regulators, legislators, and utilities in several ways, including: (1) tiered or inverted block rates that are structured to discourage high monthly energy usage; (2) directly subsidizing customer investments in improving the energy efficiency of buildings, appliances, and equipment, for example through tax incentives or energy efficiency programs; and (3) requiring utilities to invest in demand-side management (DSM) resources that are cost- effective in comparison with generation resources, typically as measured by cost effectiveness tests that consider customer, utility, and societal costs, avoided costs, and Where energy efficiency and conservation programs are focused on reducing energy consumption, time-varying pricing and DR programs have a different purpose, which is to reduce peak loads on the utility system; their focus is on capacity rather than energy. As discussed below in Section 2, demand reduction can result either from conservation or shifting of load to other hours, and both have potential benefits from the standpoint of reduced energy consumption and emissions. For example, building-shell improvements can reduce not only peak loads but total energy consumption. Shifting of emissions to other hours may also be beneficial where pollutant concentration thresholds are important, for example in the case of ground-level ozone formation. However, these potential benefits are limited by the fact that peak periods constitute only a relatively few hours out of the year, ranging from less than 1% of the year for most interruptible or direct load control programs to a maximum of about 15% for some TOU rates with wide year-round peaks. For this reason, energy conservation is usually only a small secondary effect of these programs.3 Moreover, it also possible that the net effect may be to increase total energy consumption and emissions. One reason for this is the "rebound" effect of load shifting, which can occur when energy consumption is increased in off-peak periods to compensate for conservation during peak periods. Another reason has to do with the utility's specific generation mix; if load shifting causes the utility to increase its dispatch of plants with higher heat rates, or that use more polluting fuels, the net effect may be negative. A related long-run effect can occur if the ability to dispatch demand response resources causes utilities to postpone the replacement of older, less efficient, and more polluting peaker plants with newer, cleaner, more efficient ones. 1.4 Purpose of this survey
This survey aims to provide practical information to assist a regulated electricity utility, its customers, its regulator, and other stakeholders (e.g., public interest groups and conservation advocates) to decide if time-varying pricing and other demand-response (DR) programs should be added to the utility's existing tariffs, which may be dominated 3 Conservation effects ranging from small negative to small positive impacts have been reported for demand-response programs of different kinds (King and Delurey, 2005). The strongest conservation effects were found when incentive programs were combined with aggressive information programs that increased overall energy awareness on the part of customers, leading to altered consumption behavior and increased investment in energy-saving equipment. A well-known case is the overall conservation effect in California of about 5% during the energy crisis as a by-product of statewide efforts to reduce peak loads, (Bartholomew, Van Buskirk and Marnay, 2002; Goldman, Eto and Barbose, 2002). by time-invariant rates.4 Here we define time-varying pricing as electricity rates that vary with the time of use (such as TOU, CPP, and RTP rates described above), for the purpose of clarifying the connection between customer consumption decisions and the costs of satisfying the resulting demand. We define a DR program as one that aims to induce customers to modify their demands in response to the program's attributes (e.g., time- varying rates or load curtailment request). Hence, time-varying pricing is one type of DR To enhance the survey's usefulness, we address the following substantive policy questions: (a) Are time-varying pricing and DR programs conceptually sound from the standpoint of economic theory? (b) Can they deliver benefits to the utility and its customers? (c) What are these programs' likely load shifting (i.e., flattening a customer's hourly load pattern) and conservation (i.e., reducing a customer's total kWh consumption) effects? (d) Do they offer load relief that can resolve system and/or local capacity constraints? and (e) What real-world applications of these rates and programs are found in the electricity industry? The answer to (a) helps to determine if time-varying pricing and DR programs should be considered in the first place. Programs with inherent conceptual flaws cannot be corrected by actual implementation. The answer to (b) is important because a conceptually sound program may not deliver benefits in reality due to implementation difficulties. The answer to (c) helps determine if time-varying pricing and DR programs 4 Time-invariant tariffs are common, developed mainly for the purpose of embedded cost recovery (Phillips, 1993). They can be highly non-linear, with a multi-part structure for stable cost recovery (Brown and can achieve the twin goals of energy conservation and peak load reduction. The answer to (d) helps determine if a utility can rely on time-varying pricing and DR programs for reliability reasons. The answer to (e) helps determine if time-varying pricing and DR programs are common in the industry, thus addressing concerns regarding practicability This survey is not a comprehensive review of the extensive literature on the economic theory of peak load pricing and capacity rationing, as already done by Crew, Fernando and Kleindorfer (1995). Nor does it fully review the empirical literature on customer response to time-varying pricing and DR programs. For the sake of completeness, however, a number of important papers in this area are listed in Appendix 1. These do present abundant evidence that electricity consumers respond to time- varying electricity rates: individual customer demand declines when electricity prices rise, The paper is organized as follows. Section 2 offers a general evaluation of time- varying pricing and DR programs from the standpoint of economic theory, and also discusses some implementation issues related to these programs. Section 3 evaluates the major program categories. Section 4 summarizes the results of the survey of real-world applications of time-varying rates and DR programs, for which the detailed results are found in Appendix 2. Section 5 draws some general conclusions based on the survey, which may be summarized as follows: (a) time-varying pricing and DR programs are conceptually sound; (b) when properly designed and implemented, they can deliver benefits to the utility's customers and shareholders; (c) while time-varying pricing and DR programs can reduce peak load and induce load shifting, their conservation effect is mild; (d) DR programs can reduce customer peak loads, which is useful for resolving system and/or local capacity constraints; and (e) real-world applications of some types of time-varying pricing and DR programs are numerous, indicating that they are not excessively costly or difficult to implement. 2. A general assessment of time-varying pricing and DR
programs
2.1 Efficient pricing of electricity service
Unlike other forms of energy (e.g., natural gas and oil) or public utility services (e.g., water and transportation), electricity cannot be economically stored and has few close substitutes for certain end-uses (e.g., lighting, refrigeration, and cooling). An immediate consequence is that a customer's short-run electricity demand tends to be price inelastic,5 fluctuates cyclically with time-of-day and season, and can surge randomly due to unpredictable events, such as extreme temperatures.6 Consider an integrated electricity utility obligated to provide reliable service to its customers. (We shall later consider the effect of wholesale market trading by an inter­ connected utility on TOU pricing and DR programs.) If the utility can readily react in 5 The papers listed in Appendix 1 indicate that own-price elasticity estimates by TOU have a magnitude 6 Residential demands tend to be highest in hot summer afternoons and cold winter evenings. Non­ residential demands are less weather sensitive than residential demands, but they are the highest during weekday afternoons. real time to demand changes at a stable and flat marginal cost, efficient electricity pricing to maximize the net benefits of consumption by diverse customers can be accomplished by setting the electricity energy rate ($/kWh) at that time-invariant marginal cost (Crew, Fernando and Kleindorfer, 1995; Woo, 1988; Chao, 1983). This is because a rate above (below) the marginal cost results in under-consumption (over-consumption), as the marginal benefits of consumption are above (below) the marginal cost of service Unfortunately, reliably supplying electricity to end-users is complex and potentially costly. Because electricity cannot be stored and must be supplied to meet random and geographically dispersed demand, the utility must have capacity for generation, transmission and distribution in place before demand occurs. As capital- intensive capacity investments are lumpy and require a long lead-time, the utility's short- run marginal cost curves are almost vertical when the utility is operating near full capacity. Fluctuating capacity availability implies that these marginal cost curves are random and that their uncertainty can be exacerbated by the utility's stochastic marginal costs for fuel and line losses.7 An upward shift of a local demand along a steep marginal cost curve implies that the efficient electricity rate for a specific location should spike, eliminating the local 7 For discussions on the topic of area-specific and time-varying marginal costs, see Bohn, Caramanis and Schweppe (1984), Hogan (1992), Baskette, Horiia, Kollman and Price (2005), and Heffner, Woo, Horii and Lloyd-Zannetti (1998). Examples of applying costs in integrated resource planning can be found in Baskette, Horiia, Kollman and Price (2005), Woo, Lloyd-Zannetti, Orans, Horii and Heffner (1995), and Orans, Woo and Horii (1994). excess demand that can jeopardize system reliability (Bohn, Caramanis and Schweppe, 1984; Hogan, 1992; Borenstein, 2005). If the electricity rate cannot respond to the demand shift, the utility may have to curtail supplies to customers, likely causing high customer outage costs (Munasinghe, Woo and Chao; Woo and Pupp, 1992). Similarly, an upward shift of a location-specific marginal cost curve can occur due to facility failure (e.g., generation plant outage or transmission line failure). Such a shift along a price- insensitive location-specific demand curve requires prices to increase in order to remove the excess demand. If the price change fails to occur, curtailment may again ensue. Suppose our assumed integrated utility is replaced by a local distribution company (LDC) that procures from wholesale electricity markets to meet its load obligations (Woo, Horowitz, Olson, Horii and Baskette, 2006). Wholesale spot market prices are time-varying and notoriously volatile (Woo, King, Tishler and Chow, 2006; Lafferty, Hunger, Ballard, Mahrenholz, Mead and Bandera, 2001). Nonetheless, the efficient pricing rule remains the same: the optimal electricity rates should track location- specific marginal costs in real time, resulting in a pricing mechanism commonly known as real time pricing (RTP). To be sure, the marginal cost computation must now reflect the LDC's market-based procurement costs (Baskette, Horii, Kolman and Price, 2006; Horowitz and Woo, 2006; Borenstein, 2005). 2.2 Load shifting and conservation effects of time-varying
Empirical evidence indicates that pricing electricity according to time-varying marginal costs leads to customers shifting their loads from high-cost hours to low cost hours.8 Hence, one can reasonably project that time-varying pricing can flatten a customer's An example illustrating the load shifting ability of time-varying pricing is the "Customer Choice and Control" (CCC) TOU rate option experiment with voluntary participation implemented by Central Power and Light in the City of Laredo, Texas (population 170,000) for the period February 1994 through May 1997 (Hartway, Woo and Price, 1999).9 Table 2.1 presents the option's design, which follows the pricing rule in Train (1991) and Mackie-Mason (1990).10 8 See the papers listed in Appendix 1, Hartway, Price and Woo (1999), Barbose, Goldman and Neenan (2004), and Taylor, Schwarz and Cochell (2005). 9 Laredo is located at the Texas-Mexico border on the Rio Grande River, about 160 miles southwest from San Antonio. It was chosen for the TOU pricing experiment because the area's characteristics make it a good candidate for a TOU rate option for the following reasons. First, Laredo's population is growing at 6­ 8% per year, thanks to the expanding cross-border trade due to North American Free Trade Agreement (NAFTA). Second, Laredo's electricity demand is growing at 15MW per year, causing substantial investment in transmission and distribution (T&D) in the next 10 years. Finally, Laredo's load peaks between 4-5 p.m. on weekdays, June-September, a result of the demand for air conditioning caused by southern Texas' hot climate. Hence, Laredo's marginal costs range from $0.273/kWh to $0.443/kWh during 1pm - 7pm on summer weekdays, greatly exceeding Laredo's off-peak marginal costs of Table 2.1. TOU rate option design for Laredo, Texas TOU period definition Marginal cost ($/kWh) TOU rate ($/kWh) On-peak: 4-5 p.m. (weekdays, Jun-Sep) Partial-peak: 1-4p.m. & 6-7 p.m., (weekdays, Shoulder: 1-7 p.m. (weekdays, Oct-May) Off-peak: Remaining hours Figures 2.1 and 2.2 below indicate that the TOU rate option induces load shifting by participating customers, with significant peak load reduction. The same figures also show that the option does not have a large conservation effect on total consumption. This corroborates the recent literature review by King and Delurey (2005) "that the average reduction ranges from about 4 percent for dynamic pricing programs, to a fraction of a percent for reliability programs" (p.60). The same review also indicates that programs with clear daily feedback on consumption and billing to customers can have an up to 10 percent conservation effect. ensure that a TOU option's implementation would not harm non-participating customers, a TOU option's on-peak rate should be above the default tariff's time-invariant energy rate but not exceed the on- peak marginal cost. The option's off-peak rate should be below the default tariff's time-invariant energy rate and above the on-peak marginal cost. Average Summer Shape with TOU Rate Option
Average Summer Shape without TOU Rate Option
Figure 2.1. Comparison of load shape with and without the TOU option. The lines with the diamonds indicate the weekday load shape (kW) and the squares indicate the weekend load shape (kW). The chart on the left shows the average summer shape for customers with the TOU rate option, and the chart on the right is the average of customers without the TOU rate option. Summer Weekday Impacts
Change of Consumption due to the CCC Rate Option
Figure 2.2. Load impacts and TOU rates per customer on average. The line indicates average load change of participants (kW) and corresponds to the left-hand axis, and the bars indicate TOU rate ($/kWh) on the right-hand axis. Load shifting offers cost savings because even if one on-peak kWh reduction is exactly offset by one off-peak kWh increase, the per kWh cost saving is the positive difference between on- and off-peak marginal costs. However, the per kWh cost saving may not reflect the shifted loads' effect on emissions from thermal generation. Consider the following two scenarios: • Scenario 1: On-peak generation is cleaner than off-peak generation. Suppose the on- peak generation uses natural gas with relatively low emission while the off-peak generation uses coal with relatively high emission. Load shifting in this case increases emissions, even though total output remains the same. • Scenario 2: On-peak generation is dirtier than off-peak generation. Suppose the last dispatched on-peak generation unit is a combustion turbine (CT) with relatively high emissions and the last dispatched off-peak generation unit is a new combined cycle gas turbine (CCGT) with relatively low emissions. Load shifting in this case reduces emissions, even though total output remains the same. If a pricing program can reduce consumption in both high- and low-cost hours, it achieves conservation, yielding an unambiguous reduction in emissions. However, the conservation effect of a pricing program can be minor because on-peak kWh reduction is often accompanied by an off-peak kWh increase, see Figure 2.1.11 11 To further illustrate this point, consider a household delaying his/her clothes washing and electric drying from day hours to night hours in response to TOU pricing. There is load shifting; but the total consumption remains unchanged, unless the household also lowers the drying temperature and shortens drying cycle. To induce a conservation effect, one may modify a TOU rate design with a single rate in each TOU period into one that has an inverted block structure in each TOU period. As an example, consider the following design: • On-peak rates. The on-peak rate is $0.10/kWh for the first monthly block of 200 kWh and $0.20/kWh for kWh in excess of the first block. • Off-peak rates. The off-peak rate is $0.03/kWh for the first monthly block of 200 kWh and $0.06 for kWh in excess of the first block. This design provides more incentive for both load shifting and conservation than a simple design consisting of an on-peak rate of $0.15/kWh for all on-peak kWh and an off-peak rate of $0.045/kWh for all off-peak kWh. 2.3 Efficient rationing of capacity
Real-time pricing of electricity by location for all customers of a utility is impractical and costly (Brennan, 2002).12 Absent location-specific rates that can continuously balance randomly fluctuating local demands and supplies, rationing during an emergency is necessary to preserve system integrity and prevent cascading blackouts. Having TOU rates in place does not completely remove the need for rationing because such rates may only reflect an ex ante expectation of short-run marginal costs, not ex post market-clearing values (Crew, Fernando and Kleindorfer, 1995). Random rationing – for example, curtailing loads via rotating outages – of customers with diverse outage costs is economically inefficient (Woo, 1988, 1990; Chao 12 As reported in Section 4 below, RTP implementation focuses on large customers. and Wilson, 1987; Spulber, 1992a, 1992b). Selectively curtailing service to customers with low outage costs is more efficient than those with high outage costs, and complements time-varying electricity rates (such as TOU) that do not truly follow location-specific real-time pricing. To implement efficient rationing, a utility must first establish a ranking of its customers for selective curtailment. However, the utility cannot expect its customers to voluntarily reveal, without any incentive, their private outage cost information. Bluntly put, why would any customer truthfully tell the utility his/her low outage cost so that he/she could be curtailed first in an emergency? To overcome the asymmetric information problem, the utility can offer reliability differentiation, with more reliable service commanding a higher average rate than less reliable service. There are two primary approaches to price reliability: priority service (Chao and Wilson, 1987; Woo, Horowitz and Martin, 1998) and demand subscription (Spulber 1992a, 1992b; Woo, 1990). Both approaches use financial incentive to induce a customer to self-reveal his/her private outage cost information. The two services differ in how a capacity shortage is resolved. Though slightly different in theory, these services are very similar in practice. Under priority service, a customer assigns priorities (e.g., high vs. low) to his/her demand segments and pays the corresponding priority charges set by the utility. High priority service has a higher charge than low priority service. During an emergency, low priority demands are first cut via service interruption to remove the excess demand. (We define "interruption" as a customer's demand of a particular priority is reduced to zero.) When necessary, interruption extends to high priority demands. If cutting service to all demand within a priority class results in too much load relief, the utility applies random interruption within the class. Under demand subscription, a customer specifies a firm service level (FSL) below which his/her demand cannot be curtailed. Hence, service curtailment is only limited to the loads above the self-selected FSL. In other words, the customer's loads above the FSL are curtailable. The FSL has a $/kW-month subscription charge that is higher than the one for the curtailable loads. During an emergency, curtailable loads are cut first. If cutting all curtailable loads results in too much load relief, the utility can apply proportional curtailment so that a customer's actual curtailed load is proportional to the customer's chosen FSL. 2.4 Load relief for resolving system and/or local capacity
TOU/RTP rates rely on time-varying electricity prices to flatten a customer's hourly loads. Since the rates are highest during the on-peak hours when a utility facing high aggregate demands, high on-peak rates can induce load relief to help resolve system and/or load capacity constraint. Despite relatively low price responsiveness as reported in the papers listed in Appendix 1, a sufficiently large rate differential between on-peaks hours and off- peak hours can produce significant load relief. For example, the CCC TOU rate option described by Table 2.1 provided significant load relief, see Figures 2.1-2.2. Similarly, RTP programs have been found to achieve up to 30% of load reduction of a participating customer (Barbose, Goldman and Neenan, 2004; Taylor, Schwarz and Cochell, 2005). Hence, if customer participation in a TOU/RTP program is mandatory, the program can be effective in providing sufficient load relief in an area with an imminent capacity shortage. If program participation is voluntary, however, enrolling sufficient customers can be a problem (Barbose, Goldman, Bharvirkar, Hopper, Ting and Neenan, 2005). To directly address the capacity shortage problem, a utility can offer incentives in the form of a bill reduction to a customer who gives the utility the right but not the obligation to curtail or interrupt its loads. The program can be offered as priority service, demand subscription service, or variants thereof, as reported in Sections 3 and 4 below. While customer participation rates rise with the incentive level, whether sufficient customer enrollment and enough non-firm loads can be obtained are not 100% certain. To see this point, consider the following examples: • Example 1: residential customers. If an area has many air-conditioner owners or electric water heater owners, a load control program may attract a large number of curtailable loads because these households can modify their appliance utilization relatively easily in response to a utility's load curtailment request. But if the area only has small users with critical end-use loads like computers and lighting, signing up sufficient curtailable loads may be a difficult challenge. • Example 2: non-residential customers. If an area has many industrial facilities such as oil refineries, chemical production plants, air separation plants, and cement factories, a curtailable service program can attract a large number of curtailable loads because these plants can modify their production schedule relatively easily in response to a utility's load curtailment request. But if the area only has high-tech firms, it becomes more difficult to sign up sufficient curtailable loads. 2.5 Revenue collection
It entails large fixed and sunk costs to generate, transmit, and distribute electricity to meet the stochastic cyclical demands of retail customers dispersed over a utility's service territory. Pricing electricity at locational real-time marginal costs, though economically efficient, may not collect sufficient revenue to cover the utility's fixed and sunk costs. To ensure adequate revenue collection, the utility may apply markup above marginal costs (Brown and Sibley, 1986; Woo, 1988; Spubler, 1992b).13 Alternatively, the utility can use a kW-month charge on a kW quantity that does not vary with current consumption (e.g., a customer's ratchet demand - the historic maximum kW over a 12­ month period or a customer connected kW load level) (Seeto, Woo and Horowitz, 1997; Woo, Horii and Horowitz, 2002). Since small customers may not be kW-metered, the utility may impose a monthly fixed charge ($/customer-month) that varies by historic consumption band, a proxy for ratchet demand and connected load.14 2.6 Mandatory vs. optional tariffs
Despite its theoretical attractiveness, RTP is seldom implemented in North America as a mandatory tariff; optional RTP is the norm (Barbose, Goldman and Neenan, 13 The markup for a TOU period is inversely related to the size of the price elasticity of demand in that period. This makes sense because a given markup causes less consumption distortion for an inelastic demand than an elastic one. 14 But if fixed charges are too high relative to variable charges based on usage, the customer's incentive to conserve is reduced. 2004; Moezzi, Goldman, Sezgen, Bharvirkar and Hopper, 2004; Taylor, Schwarz and Cochell, 2005; Horowitz and Woo, 2006). This is because RTP can have high implementation costs (e.g., interval meters and complicated billing) and onerous information requirements (e.g., customer tracking of frequently varying rates is necessary for rationally responding to those rates). However, an optional tariff can cause significant revenue loss to the utility, without offsetting gains (Woo, Orans, Horii and Chow, 1995). An example is a presumably "revenue-neutral" TOU rate option with on- and off-peak rates designed to collect the "same revenue" as the non-TOU default tariff from a hypothetical customer who has customer class' average on-to-off-peak consumption ratio (Mackie-Mason, 1990; Train, 1991; Hartway, Price and Woo, 1999). Customers with below average on-to-off­ peak consumption ratios are "free riders" who can reduce their bills by taking the TOU rate option without changing their consumption behavior. The revenue shortfall can have undesirable consequences. If the utility does not raise the default rate to cover the shortfall, its shareholders suffer a financial loss. Under cost of service regulation, the utility will likely raise the default rate level, harming the customers not taking the option. To mitigate a time-varying rate option's free-rider problem, a utility can implement a three-part design (Woo, Chow and Horowitz, 1996; Horowitz and Woo, 2006). The first part collects a participating customer's historic consumption profile at the default tariff. This ensures no revenue loss due to the option's implementation. The second part collects the option's program costs (e.g., meters, billing system and program administration), thus ensuring no program cost spill-over to non-participating customers. The third part prices the customer's consumption deviations from the historic consumption profile at TOU/RTP rates that track marginal costs. So long as customer demand curves are downward sloping and the default rates are between on-peak and off- peak marginal costs, the option reduces (encourages) consumption during high-cost (low­ cost) hours, yielding cost savings that can be shared among the utility, the participating customers and non-participating customers (Woo, Chow and Horowitz, 1996; Hartway, Price and Woo, 1999). Reliability differentiation can also be implemented using a three-part design (Woo, Horii and Horowitz, 2002; Seeto, Woo and Horowitz, 1997). The first part prices a participating customer's consumption below the FSL at the default tariff. The second part collects the option's program costs. The third part prices the customer's curtailable loads above the FSL at lower rates. 2.7 Implementation issues
If time-varying pricing and DR programs are conceptually sound and can deliver net benefits, one wonders why they are not more popular among utilities. Here we discuss some issues that limit these programs' adoption: (a) information requirements; (b) implementation costs; (c) stakeholder acceptance; and (d) inertia. A. Information requirements
Implementing time-varying rates and DR programs increases the information requirements for both utilities and customers, in such areas as metering, data communication, and customer education, and this in turn imposes costs on program implementation. For example, traditional metering typically records a customer's monthly kWh consumption, and in the case of larger customers, maximum kW demand. This type of metering is not suitable for time-varying pricing and DR programs because it cannot provide the price/quantity information necessary for utility and customer decisions. For instance, unless a utility can measure and bill consumption by TOU, it cannot place customers on TOU rates. The use of TOU rates requires the installation of TOU meters, which record usage during each TOU period separately. In the case of RTP, the billing information is in hourly or even finer increments (e.g., 15-minute intervals), and requires the installation of interval meters with the capacity to record this data. The utility must also be able to send the set of RTP rates daily to customers who can then respond to the price signals; this necessitates communication of prices via internet, pager, smart meter, or other specially installed devices. Similarly, a CPP program requires the utility to have the necessary communication/billing equipment to send and charge the high price to a CPP customer during critical peak hours. A utility must have load control and communication equipment in place for selective service interruption/curtailment in order to offer priority service or demand subscription service, and may require additional equipment if it desires real-time confirmation that the curtailment has actually occurred Finally, the effectiveness of programs and rates may depend on the extent to which customers have information about their consumption and means of controlling it. Lack of information also hampers a utility's program design prior to implementation. While the utility may know its own marginal costs, it may not know much in advance about customer response to a particular design. Absent the customer response information, a utility may be reluctant to offer a program in fear of substantial revenue loss. While it is possible to design a profitable program under imperfect information, the resulting design is often too conservative in its incentive structure, making it unattractive to customers. An example is a TOU rate option that only rewards a participating customer's actual shift of on-peak consumption to the off-peak period (Woo, Orans, Horii and Chow, 1995). B. Implementation cost
A program's implementation costs can be substantial for both the utility and its customers, in some cases to the point of preventing the program's adoption. Metering, billing, and program administration are some of the obvious costs that the utility may have to incur. For example, TOU rates require the installation of TOU meters, and RTP rates require the installation of interval meters, along with communication equipment to convey price signals to customers; in both cases, billing and administration are more complicated than with traditional volumetric kWh billing. Without a cost recovery mechanism, such as a monthly metering charge, the utility will be reluctant to pursue time-varying pricing and DR programs. Other examples of implementation costs range from the relatively simple radio-frequency devices that allow utilities to control loads remotely (e.g. wireless switches on air conditioners) to much more elaborate, internet- based statewide information and control systems linked to "smart" meters being considered in California and other jurisdictions (Gunther, 2005). There are also costs that a customer may bear in order to respond to the utility's price signals and curtailment requests. Such costs include behavior modification (e.g., higher temperature settings by an air-conditioner owner) and out-of-pocket costs (e.g., investments in altering the production process and equipment of a manufacturing plant). C. Competing stakeholder interests
Introducing a new program by a utility can affect the utility's shareholders, customers, and other interest groups (e.g., environmentalists and energy technology firms). These stakeholders have differing interests. For instance, utility customers are diverse (e.g., low- vs. high-income households, large vs. small users, and industrial/commercial vs. agricultural users). A TOU pricing program for small users can invite opposition from large users, if the latter are required to help pay for the program costs via a $/kWh surcharge. Similarly, a rate option may pit likely non-participants against potential participants, if the latter are expected to gain at the expense of the former. As well, environmentalists may question a program that may increase the utility's overall energy sales, when for example a pricing program like RTP encourages consumption at low-cost hours even while reducing it at peak hours. Finally, an energy technology firm may advocate one program (e.g., complicated RTP pricing) over another (e.g., simple TOU pricing), because it is motivated by its competitive advantage over other firms and by the profitability of its products. For regulators, distinguishing among competing claims begins with clarifying the objectives of rates and programs, quantifying their costs and benefits as seen from different stakeholder perspectives, and establishing the technical functionalities required for program implementation, which dictate technology requirements. D. Inertia
The slow adoption of time-varying pricing and DR programs is due in part to inertia of utility management, utility regulators, and customers. A utility subject to traditional rate of return regulation may have little profit incentive to pursue time-varying pricing and DR programs, even if these programs can help reduce its costs. However, this does not mean a utility will not respond to suitably designed incentive schemes via performance-based regulation. For instance, a utility may be rewarded for the additional customers placed under TOU pricing or in DR programs. Efficient pricing and capacity rationing assume customers are homo economicus who make decisions purely on the basis of dollars and cents (Thaler, 2000). But homo sapiens do not always behave like homo economicus (Thaler, 1994). A case in fact is the presence of status quo bias in household choices regarding electricity service reliability (Hartman, Doane and Woo, 1991). Imperfect information and risk aversion add to the "why bother?" attitudes of customers. 3. Evaluating specific types of time-varying pricing and
DR programs
Having evaluated the general characteristics of time-varying pricing and DR programs, here we describe and evaluate the specific categories of programs used by the electricity industry. To provide a relevant context, it is useful to first describe the program types based on real-world examples. This is followed by our evaluation. 3.1 Program types
For the purposes of this survey, DR programs (with time-varying pricing as a special case) are defined as any program or tariff offered by an electric utility to its customers, with either mandatory or voluntary enrollment, that seeks to modify the participating customers' consumption behavior and achieve demand reduction during hours that are costly to serve.15 Under this broad definition, there are three categories of DR programs: • Price-based programs. These programs use price signals to induce a change in consumption behavior, and usually focus on the customer's kWh energy consumption. Price-based programs are generally of three types: time-of-use (TOU) pricing, real- time pricing (RTP), and critical peak pricing (CPP). • Quantity-based programs. These programs directly limit a customer's coincident kW demand, either through a utility's physical control over a customer's load, or by sending a load curtailment request to a customer. Quantity-based programs are generally of three types: direct load control (DLC), curtailable and interruptible service (CIS), and demand subscription service (DSS). • Hybrid programs. These programs combine the attributes of both price- and quantity- based programs. For example, a TOU program can offer a bill discount to a customer's curtailable demand. Two important types of hybrid programs considered in this survey are demand bidding and quantity-control programs with customer override and buy-through. There are two versions of demand bidding: (1) a participating customer submits a schedule of curtailable load levels and the prices at which the customer is willing to accept to the utility that in turn decides how much curtailable load to buy from the customer; and (2) the utility posts the price that it is willing to pay and a participating customer submits the amount of curtailable load that he/she is willing to sell at that price. Customer override and buy-through refers 15 The survey does not include programs related to behind-the-meter generation that reduces a customer's metered load. to the provision in a quantity-based program that allows a participating customer not to comply with the utility's load curtailment request by paying the high buy-through rates posted by the utility that apply to the curtailment hours. A. Attributes of Price-Based DR Programs
Price-based DR programs vary according to the following key attributes: • Mandatory vs. voluntary enrollment. This refers to whether a customer must join the program in order to receive electricity service. Mandatory enrollment means that a customer must stay in a DR program in order to receive electricity service. Voluntary enrollment may mean (a) opt-in whereby a customer can decide to switch from a non- DR default service, typically at non-time-differentiated energy rates, to a DR optional program; or (b) opt-out whereby all customers are first placed in a DR program but a customer can later decide to quit the DR program and take a non-DR optional service, typically at non-time-differentiated rates. • Number of pricing periods. This refers to how many prices a participating customer may face over a billing period. A TOU program may only have on-peak and off-peak energy rates ($/kWh) at fixed hours during a whole season or year. An RTP program can have 24 different hourly energy prices ($/kWh) each day, transmitted by the utility to a participating customer. A CPP program may have TOU rates that vary by fixed periods, but with the addition of critical peak periods with much higher rates. CPP programs may also be structured to have a constant energy price ($/kWh) for almost every hour of the year, except for the relatively few critical hours. • Frequency of price updates. This refers to how often the program's rates are updated. A TOU program's seasonal (e.g., summer on-peak, partial-peak, and off-peak) rates tend to remain unchanged throughout the season. In contrast, a RTP program's rates may vary daily, reflecting the utility's daily cost and supply conditions. A CPP program's rates increase sharply in the critical hours. The critical hour rates can be preset and fixed in advance; or they can vary to reflect the utility's cost and supply conditions in those hours.16 A CPP program's critical hours are subject to preset terms and conditions that govern the maximum number of dispatch days per year, the maximum number of hours per dispatch, and the extent of advance notice (e.g., day- ahead vs. day-of notice). • Dispatchable vs. non-dispatchable price signals. This refers to whether the utility can dispatch a program's price signals to a participating customer. RTP and CPP programs have dispatchable price signals, but TOU programs do not. • One-part vs. two-part billing. This refers to whether a participating customer's bill is tied to his/her consumption pattern prior to joining the program. One-part billing depends only on a customer's currently metered consumption. Two-part billing requires current consumption to be measured against a customer baseline load (CBL), which is typically based on past consumption patterns. Two-part billing charges a customer's CBL usage at the otherwise applicable rate, and then adjusts the bill for the customer's usage above or below the CBL. 16 The six utilities in this survey that offer residential CPP rates only have fixed prices for critical hours. In principle, critical hour prices could be dynamic. Table 3.1 presents examples of the three main residential price-based programs to contrast key attributes. Section 4 below describes the variations found among the programs in the survey in greater detail. Table 3.1. Examples of residential price-based DR programs and some key attributes. TOU: Arizona Public CPP: Gulf Power's Service's ET-1 Tariff Good Cents Select Edison's RHEP Pilot Program Voluntary Voluntary voluntary enrollment Number of periods 8760 (each hour of 3 fixed periods plus with different prices variable number of critical peak periods17 Frequency of price Each rate cycle plus critical peaks called up to 20 times/year Dispatchable price One-part vs. two- One-part One-part B. Attributes of Quantity-Based DR Programs
Quantity-based programs vary according to the following key attributes: • Mandatory vs. voluntary enrollment. As with price-based programs, this refers to whether a customer must join the program in order to receive electricity service. • Firm service level (FSL). This refers to a customer's self-selected kW level, below which the customer's load is not subject to utility-activated curtailment. An 17 In this survey, the maximum number of critical peak hours per year ranged from 40 to 350, with a typical value of about 100 hours, or 20-25 CPP events per year. interruptible service program's FSL is zero kW, as the participating customer's entire load is subject to interruption.18 A curtailable customer's load above the customer's self-chosen FSL is subject to curtailment, hence the term "curtailable" load. A curtailable load service program is the same as demand subscription service (DSS) when the demand subscribed is the FSL during the curtailment hours. A variant of
the DSS is that the demand subscribed is a maximum kW that the customer must not exceed in any hour.
• Curtailment terms and conditions. This refers to the preset terms and conditions under which a curtailment can occur, including the maximum number of curtailment days per year, maximum number of hours per curtailment, and the extent of advance notice (e.g., day-ahead vs. day-of notice). • Incentive payment. This refers to a program's payment to participating customers, often an upfront payment in the form of a discount for the customer's curtailable load. It may also have a performance incentive to reward the customer's load reduction below the FSL on a curtailment day. Since the incentive payment is an adjustment to the customer's bill under the applicable standard tariff, a quantity-based program has two-part billing. • Penalty. This refers to a program's penalty for a participating customer who fails to reduce load below his/her FSL when requested by the utility. The penalty can be monetary, offsetting the incentive payment. It can also be non-monetary, for instance 18 This is the traditional definition of an interruptible program. The terminology is not used in a consistent way across all utilities, however. Some programs called "interruptible" have the features of a curtailable program as defined above. These distinctions are unimportant for the purposes of this survey, since interruptible and curtailable programs are grouped together. when a participating customer with repeated non-compliance is removed from the • Utility control vs. customer compliance. This refers to whether the offering utility can physically reduce a participating customer's load by "pushing a button." An example of utility control is an air-conditioning load-shedding program under which the utility can remotely shut off or cycle a customer's A/C unit during the curtailment hours. In contrast, programs based on customer compliance do not give the utility physical control over the customer's load or end-uses, even though non-compliance may result in a severe financial penalty. Table 3.2 presents real-world examples of the three main types of residential quantity-based programs in order to contrast a few key attributes. Section 4 below describes the variations found among the programs in the survey in greater detail. Table 3.2. Examples of demand subscription service (DSS) and direct load control (DLC) programs for residential customers, and curtailable/interruptible service (CIS) for commercial and industrial customers. Attribute DSS: ACEA-Electrabel's DLC: Northern States CIS: PPL Electric Power's "Saver's Switch" Utilities' IS-T(R) Enrollment Mandatory Eligibility Residential Residential customers Large customers with 69 with demand < 10 kW with air conditioners kV service and 1 MW of curtailable load Between 1.5 and 10 kW Entire AC load is Must be less than historic Curtailment terms Must not exceed at any Utility may turn Interruption may be customers' AC on and off called no more than 15 at 15 minute intervals up days per year, 5 days in to 300 hours per year any month, or 10 hours in any event Incentive payment None 15% off energy charges during summer months Penalty for non­ $24.95 per kW per event Utility control via load Utility direct control Customer voluntary customer control compliance, subject to C. Attributes of Hybrid DR Programs
A hybrid program has the attributes of a price-based and a quantity-based program. The following examples illustrate this point: • A TOU program can offer lower rates for a customer's curtailable demand and the non-firm energy associated with that demand. • An RTP program can have a CIS rider so that a participating customer's FSL (say, 10 kW) is below its CBL (say, 15 kW). On non-curtailment days, the customer receives RTP rates from the utility and can decide, without any restrictions, his/her demand response to the price signals. On a curtailment day, however, the customer must cut his/her load below the FSL of 10 kW. Thus, the CIS rider provides the utility with additional assurance of load reduction, beyond what the customer's potentially uncertain price response may imply. • A CIS program can have a buy-through feature under which a participating customer can consume above his/her FSL during curtailment hours by paying a charge that tracks the utility's marginal cost in those hours. The buy-through charge is not a penalty because a CIS program with buy-through can have a smaller incentive payment than one without. • A DLC program can have a buy-through feature under which the customer can choose to override the utility's load control signal to the customer's load. As with CIS programs, the customer is then exposed to dynamic prices during the period of • A demand bidding (DB) program may be used by a utility during high cost hours to invite a participating customer to offer varying amounts of curtailable load at prices that he/she is willing to accept. After receiving the offers, the utility can then decide how much curtailable load should be procured from each participating customer. 3.2 Program evaluation
A. Criteria
We apply the following criteria to evaluate the relative merits of each type of DR • Efficiency. This refers to (a) the program's ability to convey price signals based on the marginal cost pricing rule and to effect efficient capacity rationing during an emergency; and (b) the magnitude of the program's implementation costs that include the utility's program costs and a participating customer's costs (e.g., cost incurred in response to RTP or load curtailment). • Load shifting and conservation. This refers to the program's ability to (a) induce a customer to shift its loads from high-cost hours to low-cost hours; and (b) reduce the customer's overall consumption. • Load relief for resolving a system and/or local capacity shortage. This refers to the program's ability to provide enough load relief for a utility to resolve a shortage. • Rate impact. This refers to the program's effect on the utility's rate. If the program improves the utility's gross margin (i.e., the program's incremental revenue less incremental cost), the program's implementation does not increase the utility's rate, mitigating the opposition from non-participating customers. • Stakeholder acceptance. This refers to the program's acceptance by the utility and its customers. A conservative program design that protects the utility's revenue collection and cost recovery may not be attractive to the utility's customers. Table 3.3 summarizes the evaluation results. This table does not consider hybrid programs because their performance depends how a hybrid program is structured. To the extent that a hybrid program can incorporate the desirable attributes of a price-based and a quantity-based program, the hybrid may be superior to either program alone. What emerges from Table 3.3 are the following findings: • Mandatory TOU pricing is likely to be opposed by many customers because of the bill increase to customers with relatively more on-peak consumption. It can shift load and provide load relief, but may not achieve substantial conservation. While having low implementation costs, TOU pricing does not effect efficient rationing. Optional TOU pricing improves customer acceptance, but rate impact and revenue loss can be a concern to non-participating customers and the utility. • Mandatory RTP is likely to be opposed by the utility's customers because of its complexity and high implementation cost. It can shift load and provide load relief, possibly with a minor conservation effect. Customers with spiky load patterns will view the program unjustly raising their electricity bills. Optional RTP improves customer acceptance, but rate impact and revenue loss can be a concern to non­ participating customers and the utility. • Mandatory CPP is likely to be opposed by the utility's customers because of its high implementation cost. It can shift load and provide load relief on CPP days, but is unlikely to achieve substantial conservation. Optional CPP improves customer acceptance, unless there is significant cost spill-over to non-participating customers. • A mandatory quantity-based program is likely to be opposed by customers, many of whom simply desire reliable service at easy-to-understand and reasonable rates. The utility may not object to a mandatory program that can greatly help the utility to resolve system emergencies. It can provide load relief on curtailment days but does not achieve conservation. Making the program optional improves customer acceptance, unless the program's incentive structure is viewed by non-participants as being "excessive" and not warranted by the expected reliability improvement or cost Table 3.3. Evaluation results (a) Tracks expected (a) Can effect load Can achieve load (a) If mandatory, (a) If mandatory, short-run marginal shifting; (b) may likely opposed by costs by TOU and impact, but can be season, (b) Does not ration capacity customers; (b) If the utility; (b) If efficiently during emergency, (c) Low potentially large implementation cost impact due to free participants and riders, which can possibly the utility mitigated by a careful design (a) Tracks day-head (a) If mandatory, or day-of short-run likely opposed by marginal cost; (b) If the RTP rates are the utility due to set at correctly predicted marginal optional, likely efficiently; (c) High implementation cost participants and possibly the utility due to revenue loss (a) Tracks short-run (a) If mandatory, marginal cost shortly unless the utility likely opposed by before emergency; heavily discounts (b) If the CPP rates the rate for the the utility due to are set at correctly non-critical hours predicted marginal efficiently; (c) High implementation cost implementation cost spills over to other customer classes (a) Does not track (a) Can effect load (a) If mandatory, unless the utility likely opposed by marginal costs; (b) curtailment days; offers a curtailable customers; (b) If Implement efficient rate discount that participants may Potentially high utility's expected implementation cost "excessive" discount 4. Survey of real-world designs
To aid assessment of demand response (DR) programs by a utility, its customers and other stakeholders, this section reports the DR activities of 50 major US utilities and 15 major international utilities. 4.1 Sample selection and data sources
The survey sample contains 50 US utilities, selected to include (a) the 30 largest investor- owned and municipal utilities, measured by number of customers, as reported by the US DOE Energy Information Administration, and (b) 20 additional large utilities reported as undertaking either demand response or advanced metering activities. Table A.1, shown in Appendix 2, lists the US utilities included in this survey, along with their service areas, number of customers, sales (in MWh), and web site URLs. The sample also contains 15 international utilities, selected to include (1) the largest electric utilities in the 5 largest OECD electricity markets, as reported by the International Energy Agency, and (2) other large OECD utilities reported as undertaking DR or AMI activities. Table A.2 shows the international utilities included in this survey, along with their service areas, number of customers, sales (in MWh), and web site URLs. In order to present the full range of DR program designs that have been implemented by utilities, this survey is not restricted to current or standard tariffs. The data reported here also includes both pilot programs and programs that have been The data in this survey comes from (1) the relevant literature on DR, including academic and trade journal articles and national laboratory and industry association reports;19 (2) published electricity tariffs; (3) information from utility and public utilities commission (PUC) web sites; and (4) direct communications with utilities and PUCs. 4.2 Findings
A. Overview: residential DR programs and tariffs
Table 4.1 is an overview of the survey results for residential DR programs and tariffs in 50 US and 15 international utilities, including both standard and experimental tariffs and both current and historic tariffs.20 The main findings are the following: 19 Most surveys of DR programs to date have focused on large customer programs. Two exceptions that provide information on residential programs are Energy InfoSource, "Demand Response Programs," May 2002, and Peak Load Management Alliance, "Demand Response: Design Principles for Creating Customer and Market Value," November 2002. See References section for other literature referred to in the preparation of this report. 20 See Table A.3 in the Appendix for details. • Out of 65 utilities, 60 offered at least one type of residential DR program or tariff. Of the 50 US utilities, 48 had at least one DR option. Of the 15 international utilities, 12 offered at least one DR option. • Out of the 65 utilities, 17 offered two or more types of residential DR program or tariffs. Of the 50 US utilities, 13 offered two or more DR options. Of the 15 international utilities, 4 offered two or more DR options. • TOU is the most prevalent residential DR program. It is also the most prevalent form of price-based program. Out of the 65 utilities, 51 offered TOU tariffs. 10 of the 15 international utilities, and 41 of the 50 US utilities, offered TOU tariffs. • DLC is the second most prevalent residential DR option, a quantity-based program. Out of the 65 utilities, 15 offered DLC programs. All 15 of these were in the US. • Out of 65 utilities, 2 offered residential RTP tariffs. Both were in the US. • Out of 65 utilities, 6 offered residential CPP tariffs. Of these, 5 were in the US, and one was in Europe. The US utilities include the three California IOUs, Idaho Power, and Gulf Power. The international utility is Electricité de France. • Out of 65 utilities, 6 offered residential demand subscription rates. Of these, one was in the US, and 5 were international. • Out of 65 utilities, 6 offered hybrid programs. One program was in the US; it combined direct load control with a customer override and buy-through option. The other five programs were in Europe, and combined demand subscription with TOU or • None of the utilities in the sample were found to offer demand bidding or capacity buyback programs. Table 4.1. Number of utilities offering residential DR programs (includes both current and historic programs) Utilities Abbreviations: TOU = time of use. RTP = real time pricing. CPP = critical peak pricing. DSS = demand subscription service. DLC = direct load control. CIS = curtailable or interruptible service. Hybrid includes demand bidding, capacity buyback, market-based load reduction, and quantity control with buy-through, in addition to combinations of the price-based and quantity-based programs in this table. B. Overview: large customer DR programs and tariffs
Table 4.2 is an overview of the survey results for large customer (commercial and industrial) DR programs and tariffs offered by the 50 US utilities.22 It encompasses both standard and experimental tariffs, and both current and historic tariffs. The main findings are as follows: 21 The TOU, CPP, RTP, DLC, DSS, and/or CIS components of hybrid programs are counted separately, in addition to counting as a hybrid program. For example, the Enel SPA program that combines DSS with TOU is counted as a DSS program, a TOU program, and a hybrid program. Hybrid programs that do not contain the 6 price- and quantity-based components are counted as hybrid programs only.
22 See Table A.4 in the Appendix for details. • All 50 US utilities offered at least one type of DR program or tariff to large customers. 44 offered two or more programs, and 34 offered three or more. • The most prevalent DR program is TOU: out of the 50 utilities, 48 offered TOU • The second most prevalent DR program is curtailable and interruptible service: out of the 50 utilities, 41 offered CIS programs. • RTP is also popular. Out of 50 utilities, 24 offered RTP. • Out of 50 utilities, 4 offered CPP tariffs and they are all in California.23 • Out of 50 utilities, 6 offered large-customer DLC programs. • Out of 50 utilities, 15 offered some form of hybrid program that combines elements of a price-based program and a quantity-based program. The types of hybrid programs range from mandatory TOU pricing with CIS option to market-based load reduction (e.g., demand bidding and capacity buy-back). Table 4.2. Number of US utilities with current or historic DR programs for large non­residential customers Utilities Abbreviations: TOU = time of use. RTP = real time pricing. CPP = critical peak pricing. DSS = demand subscription service. DLC = direct load control. CIS = curtailable or interruptible service. Hybrid includes demand bidding, capacity buyback, market-based load reduction, and quantity control with buy-through, in addition to combinations of the price-based and quantity-based programs in this table. 23 In addition to the 3 California IOU CPP pilots, LADWP's Schedule XRT "Experimental RTP" pilot is also a CPP program in terms of its design features, rather than a RTP program. C. Residential TOU pricing
The 65 US and international utilities surveyed offered a total of 57 time-of-use pricing programs for residential customers. These are described individually in Table A.4 in the Appendix 2, including utility and program name, description of the program features, prices and customer participation cost, and tariff web site URL. Among the 57 TOU programs, 6 major variants were found. These are described in Table 4.3, with the number of instances found and one example of each. The most common type was an energy-only TOU design, which constituted 38 of the 57 programs. TOU pricing for demand only, with flat energy rates, was found in 2 cases. TOU pricing for both energy and demand was found in 4 cases. TOU energy pricing was combined with demand subscription in 5 cases, with block-rate energy pricing in 7 cases, and load management in one case. Table 4.3. Types of residential TOU tariffs based on 57 programs Type of TOU Tariff of Colorado (Xcel) Energy and demand TOU Power & Light Co Energy TOU plus demand 5 8.8% Electricidade Energy TOU with block Energy TOU with utility- installed load management Among the 57 TOU programs, several different features were found. These are described in Table 4.4, with the number of instances shown for each feature. Of the 57 programs, 50 required that customers pay to participate. Enrollment for 54 of the programs was voluntary only, while the remaining 3 programs were mandatory for some of the customers. All but one of the 57 programs required the customer actively opt-in to participate in the program, as opposed to setting TOU pricing as the default rate. Of the TOU programs surveyed, 52 were implemented as standard programs rather than pilot Table 4.4. Residential TOU participation features (based 57 programs) Feature Customer Pays to Participate Voluntary Enrollment Only Mandatory for Some Customers Opt-In (vs. Default) Standard (vs. Pilot) Table 4.5 describes other variations found among the 57 TOU pricing programs. The number of TOU periods ranged from 2 to 4, and the number of seasons from 1 to 4. The duration of on-peak periods ranged from 4 to 16 hours, with an average of 10 hours. On-peak prices ranged from 1.65¢/kWh to 39.37¢/kWh, with an average of 14.30¢/kWh. Off-peak prices ranged from 0.63¢/kWh to 13.81¢/kWh, with an average of 5.91¢/kWh. The ratio of on-peak to off-peak prices ranged from 1.0 to 29.0, with an average ratio of 3.5.24 Finally, customer charges for participating in TOU pricing ranged from zero to Table 4.5. Variations among residential TOU Tariffs based on 57 programs Variable Number of Periods Number of Seasons Duration of Peak Period On-Peak Price (¢/kWh) Off-Peak Price (¢/kWh) On Peak/Off Peak Ratio Monthly Charge ($/month) D. Residential critical peak pricing
Among the 65 utilities surveyed, 6 utilities offered a total of 9 critical-peak pricing (CPP) programs for residential customers. These are listed in Table A.5 in the Appendix, including utility and program name, description of the program features, prices and customer participation cost, and tariff web site URL. Table 4.6 summarizes CPP customer eligibility and participation features. In 5 cases, participating customers paid an additional monthly fee. Among the 9 CPP programs, 6 were pilots of the California investor-owned utilities, used in the Statewide Pricing Project (SPP). Altogether, 7 of the 9 CPP programs were pilots. All were voluntary and opt-in. The two standard-tariff CPP programs, those of Electricité de France and Gulf Power, are described individually in greater detail below. 24 For tariffs with different seasonal on-peak prices, the season with the highest value was used. 25 Customer charges were calculated as the difference between the charge for TOU service and the charge for default flat rate service. Table 4.6. Residential CPP participation features based on 9 programs Feature Customer Pays to Participate Voluntary Enrollment Only Mandatory for Some Customers Opt-In (vs. Default) Standard (vs. Pilot) Table 4.7 describes the rate design of the 9 CPP programs. Eight programs superimposed critical peak prices on a TOU background rate, while one program (Idaho Power) used flat-rate pricing as the background rate. Six programs (all in California) combined CPP with block pricing for usage above a location-specific baseline level. One program combined CPP with demand subscription, and 4 programs combined CPP with deployment of enabling technology (smart thermostats) for customer load management. Table 4.7. Residential CPP rate design based on 9 programs Type of TOU Tariff CPP with TOU rates CPP with flat rates 1 11% Electricité CPP + load management Good Cents Select CPP + block pricing Table 4.8 describes other variations found among the 9 CPP pricing programs. In 7 of the 9 programs, CPP days fall year-round, and in 2 programs they are restricted to a particular season. The maximum number of CPP days per year ranged from 10 to 22. The maximum number of hours per day ranged from 4 to 16, and the maximum number of hours per year ranged from 40 to 352. The number of background TOU periods ranged from 2 to 3, and the number of seasons from 1 to 2. CPP prices ranged from 20.6¢/kWh to 84.3¢/kWh, with an average of 50.5¢/kWh. TOU on-peak prices ranged from 5.1¢/kWh to 27.4¢/kWh, with an average of 15.8¢/kWh. Off-peak prices ranged from 3.0¢/kWh to 12.1¢/kWh, with an average of 6.6¢/kWh. The ratio of CPP to on-peak prices ranged from 2.4 to 8.5, with an average ratio of 3.7.26 The ratio of CPP to off-peak prices ranged from 4.1 to 15.5, with an average ratio of 8.8. Finally, customer charges for participating in TOU pricing ranged from zero to $4.95/month.27 Table 4.8. Variations among residential CPP tariffs based on 9 programs Variable CPP Months Per Year CPP Maximum Days Per Year CPP Maximum Hours Per Day CPP Maximum Hours Per Year CPP Event Advance Notice (hours) Number of TOU Periods 26 For tariffs with different seasonal CPP and/or on-peak prices, the season with the highest value was used. 27 Customer charges were calculated as the difference between the charge for CPP service and the charge for default flat rate service. Number of Seasons TOU On-Peak Price (¢/kWh) TOU Off-Peak Price (¢/kWh) CPP Price (¢/kWh) Ratio of CPP/On-Peak Price Ratio of CPP/Off-Peak Price Monthly Charge ($/month) E. Residential TOU/CPP hybrid programs
E.1 Gulf Power's GoodCents Select
Gulf Power's GoodCents Select program combines a programmable thermostat, a smart meter capable of communicating with the utility, and a CPP rate structure with TOU pricing for all non-CPP hours. The thermostat allows a participant to pre-program his/her price response by setting the desired amount of central heating and cooling, electric water heating and pool pumping. The program has approximately 9000 current residential participants with plans to expand eligibility to Multi-Family and Small Commercial customers in 2005.28 The service is offered on a voluntary basis with a participation charge of $4.95 per month. This charge includes the GoodCents Surge protection and outage notification service. Gulf Power's standard residential energy rate is 7.6 cents/kWh. The CPP rate is 32.1 cents/kWh and is limited to one percent of the hours in each year, or 87 hours per year. When the CPP rate is not in effect, participants are billed on a three-period TOU rate, 5.4 cents/kWh for the low price hours (28 percent), 28 Based on approximately 6000 participants in 2003 and a target of 3000 additional participants in 2004. 6.7 cents/kWh for the medium priced hours (59 percent) and 11.2 cents for the remaining high priced hours (12 percent). This breakdown is illustrated in a Gulf Power graphic: GoodCents Select
Participation Charge $4.95/month Standard Residential Rate 7.6 Price per kWh*
Low 5.4
*All prices are as of 04/01/05, excluding customer and/or participation charges and any applicable taxes. These prices are subject to change. E.2 EDF's Tempo CPP Tariff
EDF's "Tempo" tariff combines demand subscription, a TOU background rate, and critical peak pricing with utility-enabled customer load management. All EDF residential and small commercial tariffs require customers to subscribe to a demand level between 3 and 36 kVA, which cannot be physically exceeded. To be eligible for the "Tempo" rate, customers must subscribe to a demand level of 9 kVA or more. EDF communicates price levels by sending an intuitive signal indicating the "color" of a day. In a year, there are 300 blue or non-critical days, 45 white or semi-critical days, and 22

red or critical days; weekends are always blue, and within a given color day, there are two periods, peak and off-peak. The day's color signal is sent at 8:00 PM one day ahead to a custom-designed smart meter ("Le compteur électronique") that may be plugged into any electric socket, as signals are sent by power-line carrier. The smart meter provides current and cumulative consumption information, and allows customers to program space and water heating controls in response to the electricity price for a given color and time of day. Operation of the meter is demonstrated in the EDF graphic below: F. Real time pricing
F.1 Residential RTP
Our survey finds only two residential RTP pilot programs in the US. Commonwealth Edison's RHEP one-part RTP continues to be available today. Allegheney Power has discontinued its Electricity Price Response Pilot program, a hybrid with DLC and customer override. These programs are described in Table 4.9. Table 4.9. Two residential RTP pilot programs. UTILITY Utility and customer share Pilot program begun in Energy Info Source, control of heating and air 2001, with 300 customers conditioning loads via with central A/C. Program smart thermostat, with has been discontinued. customer having ability to override utility DLC in response to RTP signal. Day-ahead and day-of notification. 1-part RTP. All energy Pilot program begun in Barbose, Goldman, charged at hourly RTP 2001, with 750 customers rate, calculated based on as of 2003. Community index (LMP forecast price organization provides with PJM hourly load notification, load shape). Customers management training, and receive participation credit rebate to cap RTP rate at Notification at 7 PM day-ahead. F.2 Large customer RTP
Of the 50 US utilities surveyed, 24 offered RTP programs for large customers.29 Table 4.10 presents the customer participation features of the RTP tariffs found by our survey. Customer enrollment was voluntary and opt-in in all 24 cases, except for Niagara Mohawk's SC-3A tariff, which was the default rate. 11 out of the 24 programs were implemented as standard tariffs, and 13 as pilots. 29 These are described in Table A.7, including utility and program name, description of the program features, and tariff web site URL. Much of the detailed data in this section and in Table A.7 comes from Galen Barbose, Chuck Goldman, and Bernie Neenan, "A Survey of Utility Experience with Real Time Pricing," LBNL-54238, December 2004, which is the most comprehensive RTP survey to date. Table 4.10. Participation features based on RTP tariffs for large customers offered by 24 US utilities. Feature Voluntary Enrollment Only Default (opt-out) Table 4.11 describes the RTP tariffs. Of the 24 programs, 11 were one-part tariffs with all current consumption billed at the RTP rate. 12 were two-part tariffs with the CBL billed at the otherwise applicable tariffs and deviations from the CBL at RTP rates. Among the 12 two-part tariffs, 9 had adjustable CBL's, and 3 did not. Also, 9 of the 12 two-part tariffs had symmetric RTP rates for incremental and decremental usage, while 3 charged more for incremental usage than rebated for decremental usage. The most common pricing methodology, used in 14 of the 24 programs, was based on the utility's hourly marginal generation cost. The remaining 10 programs used power pool prices (4 cases), models based on temperature and other variables (2 cases), and indexes based on publicly reported forward prices (2 cases). 18 of the 24 programs have RTP rates containing adders that reflect marginal capacity/outage cost. 13 of the RTP programs had interruptible options, two of which permit buy-through without additional penalties. Finally, all 24 RTP programs offered day-ahead pricing, while one (Georgia Power) also offered an hour-ahead pricing option. Table 4.11. Large Customer RTP tariff types and characteristics based on 24 programs offered by US utilities Feature Two-Part RTP with adjustable CBL Two-Part RTP without adjustable CBL Hourly Price based on system lambda Hourly Price based on pool price Hourly Price based on model Hourly Price based on index Hourly Price based on other/not known Marginal outage or capacity cost adder Interruptible option Interruptible w/ no penalty buy-through Day ahead notification Hour ahead notification H. Residential direct load control
The 65 utilities surveyed offered a total of 15 direct load control programs for residential customers. These are described in Table A.8 in the Appendix, including utility and program name, description of the program features, customer incentive structures and participation cost, and tariff web site URL. Table 4.12 describes characteristics of customer eligibility and participation in the 15 DLC programs. In 14 cases, participation in the program entailed no additional customer charge; in one case (Detroit Edison), there was an additional monthly fee. Customer enrollment was purely voluntary and opt-in in all 15 cases. Table 4.12. Residential DLC Participation Features (from a sample of 15 programs) Customer Pays to Participate Voluntary Enrollment Only Opt-In (vs. Default) Standard (vs. Pilot) Table 4.13 describes the rate design of the 15 DLC programs. In 14 of the 15 programs, DLC was a rider attached to the otherwise applicable tariff. In one program, the Allegheny Power Electricity Price Response Pilot Program, DLC was combined with an RTP tariff. This program is described above in Table 4.9. Table 4.13. Residential DLC programs based on 15 programs Type of TOU Tariff DLC as rider on flat-rate or 14 93% Northern otherwise applicable tariff States Power (Xcel) Electricity Price Response Pilot Program Table 4.14 reports other features of the DLC programs. Seven of the 15 programs operated year-round, while the other 8 were seasonal only. DLC events were limited to a fixed maximum number of hours per day in 9 cases, but were limited in days per season in only 2 cases, and hours per season in only 3. Turning to customer incentives, 12 of the 15 programs provided customers a fixed monthly bill credit for participation, whether or not DLC was exercised. In 4 cases, customers were paid on a per event basis (for each day that DLC was exercised, regardless of how many times per day). 6 of the 15 programs focused only on central air conditioners, one on water heaters, and the remaining 8 on more than one customer load. In terms of control technology, 11 programs utilized remote switches on individual devices (such as air conditioners or water heaters), and 4 installed EMS/smart thermostat systems that controlled multiple Table 4.14. Variations among 15 residential DLC programs
Feature
Percentage
Year-Round Program Seasonal Program Limited Hours Per Day Limited Days Per Season Limited Hours Per Season Monthly Bill Credit Incentive Per Event Bill Credit Incentive A/C Cycling Only Water Heater Cycling Only Multiple Customer Loads Remote Switches on Individual Loads EMS System Controls Multiple Loads Table 4.15 shows the range of DLC designs. They varied from 4 months to year round operation. The number of curtailment hours per day ranged from 4 to 12. Monthly incentive bill credits ranged from zero to $36.00, and per event bill credits ranged from Table 4.15. Variations among 15 residential DLC programs Variable DLC months per year Maximum DLC hours per day Monthly incentive bill credit ($/month) Per event bill credit ($) Customer participation charge ($/month) G. Residential demand subscription service
The 65 utilities surveyed offered a total of 10 DSS tariffs for residential customers. These are described in Table A.9, including utility and program name, description of the program features, customer incentive structures and participation cost, and tariff web site Table 4.16 shows the main DSS program types. Two principal variants of DSS were found with regard to background rate structure. All but one of the 10 programs required customers to select a contractual maximum demand level; a demand-limiting device then prevented the customer's load from physically exceeding the subscribed level. All of these programs were in Europe or Japan. For the utilities in question, demand subscription was required even for flat-rate default tariffs, and for some the subscription level was linked to eligibility for energy rate options. For example, EDF's small customer tariff has mandatory demand subscription for all customers, at levels from 3 kVA to 36 kVA. At 3 kVA, customers have only a flat-rate tariff option. At 6 kVA, customers have the option of flat-rate or TOU tariffs. At 9 kVA and above, customers have the option of flat-rate, TOU, or CPP tariffs. The lone US DSS program found in the sample was Southern California Edison's Demand Subscription Service, a pilot program offered in the mid-1980s. In this program, customers subscribed to a level of demand that was less than their estimated maximum demand, in return for a monthly bill credit. During periods of high demand, the utility would require customers to reduce load to the subscribed level for a period of up to 4 hours, or all service was remotely curtailed. Customers could resume service by reducing their demand below the subscribed level. Among the 9 hardware-limited DSS programs, 4 had flat rates, 4 had TOU rates, and 1 had CPP rates. Table 4.16. Residential DSS based on 10 programs Type of DSS Tariff DSS with physical demand 9 90% Tokyo Electric Power Co DSS with curtailable 1 10% SCE Demand Subscription Service Pilot DSS + flat-rate tariff 4 40% Electricidade Tarifa bi-horaria 1 10% Electricite de Other features of the DSS programs are described in Table 4.17. In 4 cases out of 10, DSS was the default option, and in 6 it was opt-in. Customers paid to participate in 5 cases; although these payments were associated with the TOU or CPP component, rather than the DSS component. In the 9 international programs, DSS was a standard, year- round program. The SCE DSS pilot focused on reducing summer peak demand. Table 4.17. Features of 10 DSS programs Feature Voluntary Enrollment Only Mandatory/Default Option Customer Receives Incentive Payment Customer Pays to Participate Year Round Program Standard Program Requires Curtailment Notification I. Non-residential curtailable/interruptible service (CIS)
The most common quantity-based DR program for large customers is curtailable/ interruptible service (CIS), which was offered by 41 of the 50 US utilities surveyed. Nearly all CIS programs have large usage or connected load eligibility requirement, targeted to commercial and industrial customers. In fact, none of the utilities in the sample group were found to offer CIS to residential customers.30 Table 4.18 illustrates the range of incentives, penalties, minimum curtailable loads, curtailment events and 30 A few residential DLC programs bear the name "interruptible," but they are actually DLC programs. duration, and customer notification time, for a geographically mixed group of CIS Table 4.18. CIS program examples UTILITY OR ISO incentive for shorter notice share of utility deficiency payment hours/day, 10 days/month 15 days/year, 10 events/month, 120 hours/year J. Residential pricing/load control hybrid programs
At least six residential DR programs from the sample group can be considered hybrid programs, in that their design combined significant features of both price-based and quantity-based programs. Among the 6 hybrid programs there were three different types. One program combined RTP with DLC, 4 programs combined DSS with TOU, and one program combined DSS with CPP. An example of each type of residential hybrid is presented in Table 4.19. Table 4.19. Examples of residential hybrid programs found in survey of 65 utilities HYBRID TYPE NOTES/ELIGIBILITY Electricity Price Pilot, voluntary. Response Pilot Program Customer override of utility DLC and buy-through at RTP rates. Tariffa Bioraria "Due" Standard, voluntary. TOU rate option for customers with 3 kVA or more DSS. Electricite de France Standard, voluntary. CPP rate option for customers with 9 kVA or more DSS. In addition to the hybrid designs described in Table 4.19, a number of other hybrid combinations were possible if customers elected to combine available tariffs and riders. For example, a common type of residential hybrid results when a customer volunteers for a TOU rate option in combination with a DLC rider, which provides an additional bill discount for accepting the utility's control of the customer's air conditioner or water heater. K. Non-residential pricing/load control hybrid programs
The survey of large customer DR programs utilities identified at least 15 hybrid programs. Examples of the hybrid programs found are presented in Table 4.20. Table 4.20. Examples of large customer hybrid programs found in survey of 50 utilities HYBRID TYPE NOTES/ELIGIBILITY Pacific Gas & Electric E-20 (large customer TOU) + E-BIP (base demand, > 100 kW interruptible program) pot. reduction, 30 min notice, FSL, penalties RTP-DA-2 (day ahead > 250 kW peak demand, > 200 kW (interruptible demand pot. reduction, 30 min plus energy rider) notice, FSL, penalties Jacksonville Electric Schedule RTP (day ahead 2-part RTP) + Rider IS (interruptible) + penalties if buy- buy-through provision through plus adder Wisconsin Public Online Power Exchange > 500 kW pot. reduction, company or customer may post hourly bid for level, $ Market-based load Response (voluntary reduction utility & response, customer can initiate load reductions) savings on mkt price 4.3 Remarks
The survey results show that 60 out of the 65 utilities surveyed offered at least one form of residential DR program or tariff, but only 17 offered two or more. 51 utilities offered a residential TOU tariff. Only 2 utilities in our sample have offered optional RTP to their residential customers. 6 utilities from the sample have offered a form of CPP pricing. Of these, 5 were in the US and one was in Europe. A surprising number of utilities have offered or are currently offering quantity- based DR programs to their residential customers. 15 of the 65 utilities offered a DLC program and 7 utilities had a demand subscription (DSS) rate. All of the DLC programs were in the US. 6 of the 7 utilities with DSS are international. The international popularity of DSS is due in part to the fact that DSS allows a participating customer to decide which end-use load to cut at his/her discretion. In contrast, a DLC program's curtailment is often limited to control of a specific device, such as air conditioner or water heater. Hence, it is not surprising that some of the more advanced international utilities, such as EDF, have a more flexible program design based on DSS. The evolution of Gulf Power's GoodCents program provides a useful timeline for the early adoption of residential DR programs in North America. In the early 1980s, Southern Company began to experiment with combining RTP with energy management systems capable of facilitating an automatic response. This idea was tested at two subsidiaries of Southern Company: Georgia Power between 1985 and 1987 and subsequently at Gulf Power in 1991 and 1994. The Gulf Power demonstration project proved impractical because of difficult logistics. Moreover, residential customers and the Florida PSC questioned the acceptability of spot pricing to residential customers. These concerns led to the development of Gulf Power's residential service variable pricing (RSVP), which is, by and large, a CPP/TOU compromise version of RTP. Gulf Power's current GoodCents Select program incorporates the CPP/TOU design, along with a smart meter capable of receiving pricing signals as well as outage detection, and a programmable thermostat that allows the participant to control their air conditioning, heating, water heater and pool pump. This program is reported to have approximately 9,000 current residential participants, a 95% customer satisfaction rating, and a less than 2% churn rate. Although Gulf Power's residential DR experience is more extensive than that of most US utilities in our survey, DR is not a new concept in light of its extensive use in the commercial and industrial sectors. Most of the utilities we surveyed had experience with a more sophisticated DR design for their larger customers than a simple TOU tariff. As utilities begin to invest in more sophisticated DR enabling technologies, both international and domestic experience suggest that the number and variety of residential DR designs will increase. 5. Conclusion
We conclude by answering the questions posted in Section 1: (a) Are time-varying pricing and DR programs conceptually sound? The general answer is "yes" based on the analysis and findings in Section 2. To the extent that a utility's existing tariffs do not track locational time-varying marginal costs, one can design pricing programs such as TOU, RTP and CPP to improve economic efficiency. If a pricing program does not have continuously varying rates that reflect real-time system conditions, reliability differentiation (e.g., curtailable/interruptible service) can produce efficient rationing such that customers who value reliability most will receive service first during a capacity shortage. (b) Can they deliver mutual benefits to the utility and its customers? To the extent that a program can generate net benefits, these benefits can be shared by the utility and its customers. Good real-world examples include RTP programs for large customers and air conditioning load control programs for residential customers. (c) What are these programs' likely load shifting and conservation effects? Pricing programs can reduce peak load and induce load shifting from the high-cost hours to the low-cost hours. However, their conservation effect is mild. Reliability differentiation via quantity-based programs such as curtailable/interruptible service can effectively reduce loads during a capacity shortage. However, they typically do not have a large conservation effect, although DR programs in combination with customer energy awareness, frequent feedback on usage, and enabling technology can produce substantial conservation. Rates intended primarily to induce conservation include inverted block pricing, which is not-time varying and applies to the total consumption during the billing period (Weiss and White, 2005). Inverted block pricing is the default residential rate for many utilities, and it is worth considering the tradeoffs carefully if mandatory dynamic pricing is being considered as the alternative. Some utilities have combined inverted block rates with TOU rates to include both peak-shaving and conservation-inducing components.31 (d) Do they offer load relief that can resolve a system and/or local capacity constraint? The answer is "yes," especially price-based programs like RTP and CPP that dynamically track system capacity and cost conditions and quantity-based programs such as curtailable/interruptible service that permits the utility to control customer (e) What are the real-world applications of these rates and programs in the electricity industry? Section 4 provides numerous examples based on a survey of 65 utilities; see Appendix 2 for details. This evidence indicates that such programs are not necessarily costly and difficult to implement. 31 A recent example is the CPP experiment in California (Herter, McAuliffe and Rosenfeld, 2006). 6. References
Barbose, G., C. Goldman, R. Bharvirkar, N. Hopper, M. Ting and B. Neenan (2005) "Real Time Pricing as a Default or Optional Service for C&I Customers: A Comparative Analysis of Eight Case Studies," Report LBNL-57661, Lawrence Berkeley National Laboratory, University of California. Barbose G., C. Goldman and B. Neenan (2004) "A Survey of Utility Experience with Real Time Pricing," Report LBNL-54238, Lawrence Berkeley National Laboratory, University of California. Bartholomew, E., P. Van Buskirk, and C. Marnay (2002) "Conservation in California During the Summer 2001," Report LBNL-51477, Lawrence Berkeley National Laboratory, University of California. Baskette, C., B. Horiia, E. Kollman and S. Price (2006) "Avoided Cost Estimation and Post-Reform Fnding Allocation for California's Energy Efficiency Programs," Energy-The International Journal 31: 1084-1099. Bohn R.E., M.C. Caramanis and F.C. Schweppe (1984) "Optimal Pricing in Electrical Networks over Space and Time," Rand Journal of Economics 15(3):360-376. Bonbright, J.C. (1961) Principles of Public Utility Rates, New York: Columbia University Press Borenstein S. (2005) "The Long-Run Efficiency of Real-Time Electricity Pricing," Energy Journal 26(3): 93-116. Brennan, T.J. (2002) "Market Failures in Real-Time Metering: A Theoretical Look," Discussion Paper 02-53, Resources for the Future, Washington D.C. Brown, S.J. and D.S. Sibley (1986) The Theory of Public Utility Pricing, Cambridge, NY: Chao H-P (1983) "Peak Load Pricing and Capacity Planning with Demand and Supply Uncertainty," Bell Journal of Economics 14(1):179-190 Chao H-P and R. Wilson (1987) "Priority Service: Pricing, Investment and Market Organization," American Economic Review (5):899-916 Crew, M.A., C.S. Fernando and P.R. Kleindorfer (1995) "The Theory of Peak Load Pricing: A Survey," Journal of Regulatory Economics, 8: 215-248. Crew, M.A. and P.R. Kleindorfer (1986) The Economics of Public Utility Regulation, Cambridge, MA: MIT Press. Goldman, C., J. Eto, and G. Barbose, (2002) "California Customer Load Reductions during the Electricity Crisis: Did They Help to Keep the Lights On?" Report LBNL-49733, Lawrence Berkeley National Laboratory, University of California. Gunther E. (2005) Advanced Metering and Demand Responsive Infrastructure: A Summary of the PIER / CEC Reference Design, Related Research and Key Findings, California Energy Commission. Hartman, R.S., M.J. Doane and C.K. Woo (1991) "Consumer Rationality and the Status Quo," Quarterly Journal of Economics, February, 141-162. Hartway, R., S. Price and C.K. Woo (1999) "Smart Meters, Customer Choice and Profitable Time of Use Rate Option," Energy - The International Journal 24: 895­ Heffner, G., C.K. Woo, B. Horii and D. Lloyd-Zannetti (1998) "Variations in Area- and Time-Specific Marginal Capacity Costs of Electricity Distribution," IEEE Transactions on Power Systems 13(2): 560-567. Herter, K., P. McAuliffe and A. Rosenfeld (2006) "An Exploratory Analysis of California Residential Customer Response to Critical Peak Pricing of Electricity," Energy - The International Journal, forthcoming. Hogan W. (1992) "Contract Networks for Electric Power Transmission," Journal of Regulatory Economics 4(3):211-242. Horowitz, I. and C.K. Woo, (2006) "Designing Pareto-Superior Demand-Response Rate Options," Energy-The International Journal 31: 1040-1051. Huntington, Samuel (1975) "The Rapid Emergence of Marginal Cost Pricing in the Regulation of Electric Utility Rate Structures," Boston University Law Review 55: Hyman, L.S., A.S. Hyman, and R.C. Hyman (2000) America's Electric Utilities: Past, Present, and Future, 7th edition, Arlington, VA: Public Utilities Reports Joskow, P.L. (1976) "Contributions to the Theory of Marginal Cost Pricing," Bell Journal of Economics, 7(1):197-206. Joskow, P. (1979) "Public Utility Regulatory Policy Act of 1978: Electric Utility Rate Reform," Natural Resources Journal 19: 787-809 Kahn, A. (1970) The Economics of Regulation, New York: John Wiley & Sons Kahn, E. (1988) Electric Utility Planning and Regulation, Berkeley: U.C. Press King, C. and D. Delurey (2005) "Twins, Siblings, or Cousins? Analyzing the Conservation Effects of Demand Response Programs," Public Utilities Fortnightly143(3): 54-61. Lafferty, R., D. Hunger, J. Ballard, G. Mahrenholz, D. Mead and D. Bandera (2001) "Demand Responsiveness in Electricity Markets," Staff Paper, Office of Markets, Tariffs and Rates, Federal Energy Regulatory Commission, Washington D.C. Mackie-Mason J.K. (1990) "Optional Time-of-Use Pricing Can be Pareto Superior or Pareto Inferior," Economics Letters 33(4):363-367. Moezzi M., C. Goldman, O. Sezgen, R. Bharvirkar and N. Hopper (2004) "Real Time Pricing for the Real Live Firm," Report LBNL-54978, Lawrence Berkeley National Laboratory, University of California. Munasinghe, M., C.K. Woo and H.P. Chao, editors (1988) Special Electricity Reliability Issue, Energy Journal, 9. Orans, R., C.K. Woo and B. Horii (1994) "Targeting Demand Side Management for Electricity Transmission and Distribution Benefits," Managerial and Decision Economics 15, 169-175. Phillips, C.F. (1988) The Regulation of Public Utilities: Theory and Practice, Arlington, VA: Public Utilities Reports Seeto, D.Q., C. K. Woo and I. Horowitz (1997) "Time-of-Use Rates vs. Hopkinson Tariffs Redux: An Analysis of the Choice of Rate Structures in a Regulated Electricity Distribution Company," Energy Economics 19: 169-185. Spulber D.F. (1992a) "Optimal Nonlinear Pricing and Contingent Contracts," International Economic Review 33(4):747-772. Spulber D.F. (1992b) "Capacity-Contingent Nonlinear Pricing by Regulated Firms," Journal of Regulatory Economics 4(4):299-320. Taylor, T.N., Schwarz, P.M. and Cochell J.E. (2005) "24/7 Hourly Response to Electricity Real-Time Pricing with up to Eight Summers of Experience," Journal of Regulatory Economics 27: 235-262. Thaler, R.H. (1994) Quasi Rational Economics, Russell Sage Foundation, NY: New York. Thaler, R. H. (2000) "From Homo Economicus to Homo Sapiens." Journal of Economic Perspectives, 14(1): 133-141. Train K. (1991) Optimal Regulation, Cambridge, MA: MIT Press. Turvey, R. and D. Anderson, (1977) Electricity Economics: Essays and Case Studies, Washington, D.C.: World Bank Weiss, P.C. and M.W. White (2005) "Household Electricity Demand, Revisited," Review of Economic Studies 72: 853-883. Woo, C.K., I. Horowitz, A. Olson, B. Horii and C. Baskette (2006) "Efficient Frontiers for Electricity Procurement by an LDC with Multiple Purchase Options," OMEGA 34(1): 70-80. Woo, C.K., M. King, A. Tishler and L.C.H. Chow (2006) "Costs of Electricity Deregulation," Energy-The International Journal 31:747-768. Woo, C. K., B. Horii and I. Horowitz (2002) "The Hopkinson Tariff Alternative to TOU Rates in the Israel Electric Corporation," Managerial and Decision Economics 23: Woo, C. K., I. Horowitz and J. Martin (1998) "Reliability Differentiation of Electricity Transmission," Journal of Regulatory Economics 13: 277-292. Woo, C.K., P. Chow and I. Horowitz (1996) "Optional Real-Time Pricing of Electricity for Industrial Firms," Pacific Economic Review 1(1): 79-92. Woo, C.K., R. Orans, B. Horii and P. Chow (1995) "Pareto-Superior Time-of-Use Rate Option for Industrial Firms," Economics Letters 49: 267-272. Woo, C.K., D. Lloyd-Zannetti, R. Orans, B. Horii and G. Heffner (1995) "Marginal Capacity Costs of Electricity Distribution and Demand for Distributed Generation," Energy Journal 16(2): 111-130. Woo, C.K. and R.L. Pupp (1992) "Costs of Service Disruptions to Electricity Customers," Energy - The International Journal 17(2): 109-126. Woo, C.K. (1990) "Efficient Electricity Pricing with Self-Rationing," Journal of Regulatory Economics 2(1): 69-81. Woo, C.K. (1988) "Optimal Electricity Rates and Consumption Externality," Resources and Energy 10: 277-292. York, D. and M. Kushler (2005) Exploring the Relationship Between Demand Response and Energy Efficiency: A Review of Experience and Discussion of Key Ideas, ACEEE Report No. U052. Appendix 1: Source Material on Empirical Estimation of
Customer Response to TOU Pricing and DR Programs
Acton, P. and Park, R. "Projecting Response to Time-of-Day Electricity Rates," RAND Report N-2041-MD, December 1984. Aigner, D. and Chung, C. (1981) "Industrial and Commercial Demand for Electricity by Time-of-Day," Energy Journal 2(3): 91-110. Aigner, D. and Hirschberg, J. (1985) "Commercial/Industrial Customer Response to Time-of-Use Prices: Some Experimental Results," Rand Journal of Economics Aigner, D. and Leamer E. (1984) "Estimation of Time-of-Use Pricing Response in the Absence of Experimental Data," Journal of Econometrics, Vol. 26. Aigner, D. and Lillard, L. (1984) "Measuring Peak Load Pricing Response from Experimental Data: An Exploratory Analysis," Journal of Business and Economic Statistics, Vol. 2, No. 1. Atkinson, S. (1979) "Responsiveness to Time-of-Day Electricity Pricing: First Empirical Results," Journal of Econometrics 9(1-2): 79-95. Aubin, C. Fougere, E., Husson E. and Ivaldi, M. (1995) "Real-Time Pricing of Electricity for Residential Customers," Journal of Applied Econometrics 10. Brzycki, J. and Frederick, A. (1982) "Response of Industrial and Commercial Customers to Time-of-Use Rates," Energy Journal 3(2): 185-186. Caves, D. and Christensen, L. (1980) "Econometric Analysis of Residential Time-of-Use Rates Electricity Pricing Experiments," Journal of Econometrics, 14: 287-306. Caves, D. and Christensen, L. (1980) "Residential Substitution of Off-Peak for Peak Electricity Usage under Time-of-Use Pricing," Energy Journal 1(2): 85-142. Caves. D., Christensen, L. and Herriges, J. (1984) "Consistency of Residential Customer Response in Time-of-Use Electricity Experiments," Journal of Econometrics, 26: Caves, D., Christensen, L. and Herriges, J. (1987) "The Neoclassical Model of Customer Demand with Identically Priced Commodities: An Application to Time-of-Use Electricity Pricing," Rand Journal of Economics 18(4):564-580. Caves, D., Herriges, A. and Kuester, K. (1989) "Load Shifting Under Voluntary Residential Time-of-Use Rates," Energy Journal 10(4): 83-99. Caves, D. and Christensen, L., Schoech, P. and Hendricks, W. "A Comparison of Different Methodologies in a Case Study of Residential Time-of-Use Electricity Pricing," Journal of Econometrics 26: 17-34. Ham, J., Mountain, D. and Chan, M. (1997) "Time-of-Use Prices and Electricity Demand: Allowing for Selection Bias in Experimental Data," Rand Journal of Economics 28(1): 113-141. Hausman, J. and Trimble, J. (1984) "Appliance Purchase and Usage Adaptation to a Permanent Time-of-Day Schedule," Journal of Econometrics 26. Herriges, J., Baladi, S., Caves, D. and Neenan, B. (1993) "The Response of Industrial Customers to Electric Rates Based Upon Dynamic Marginal Costs," Review of Economics and Statistics 75(3): 446-454. Herriges, J., Caves. D. and Christensen, L. (1984) "Modeling Alternative Residential Peak-Load Electricity Rate Structures," Journal of Econometrics 26: 249-268. Herriges, J. and King, K. (1994) "Residential Demand for Electricity Under Bloke Rate Structures: Evidence from a Controlled Experiment," Journal of Business and Economic Statistics 12(4)4: 419-430. Hirschberg, J. and Aigner, D. (1983) "An Analysis of Commercial and Industrial Customer Response to Time-of-Use Rates" Energy Journal 4, Special Issue. Kohler, D. and Mitchell, B. (1984) "Response to Residential Time-of-Use Electricity Rates: How Transferable are the Findings?" Journal of Econometrics 26: 141-177. Lawrence, A. and Braithwait, S. (1979) "The Residential Demand for Electricity with Time-of-Day Pricing," Journal of Econometric 9: 59-77. Lillard, L. and Aigner, D. (1984) "Time-of-Day Electricity Consumption Response to Temperature and Air Conditioning," Journal of Business and Economic Statistics, Mountain, D. and Lawson, E. (1995) "Some Initial Evidence of Canadian Responsiveness to Time-of-Use Electricity Rates," Resource and Energy Economics, 17. Park, R. and Acton, J. (1984) "Large Business Customer Response to Time-of-Day Electricity Rates," Journal of Econometrics, 26: 229-252. Parks, R. W. and Weitzel, D. (1984) "Measuring the Customer Welfare Effects of Time- Differentiated Electricity Prices," Journal of Econometrics 26: 35-64. Spann, R. and Beauvais, E. (1984) "Econometric Estimation of Peak Electricity Demands," Journal of Econometrics 9: 119-136. Taylor, T.N., Schwarz, P.M. and Cochell J.E. (2005) "24/7 Hourly Response to Electricity Real-Time Pricing with up to Eight Summers of Experience," Journal of Regulatory Economics 27: 235-262. Tishler, A. (1983) "The Industrial and Commercial Demand for Electricity Under Time- of-Use Pricing," Journal of Econometrics 23: 369-384. Train, K. and Mehrez, G. (1994) "Optional Time-of-Use Prices for Electricity: Economic Analysis of Surplus and Pareto Impacts," Rand Journal of Economics 25: 263-283. Train, K. and Toyama, N. (1989) "Pareto Dominance Through Self-Selecting Tariffs: The Case of TOU Electricity Rates for Agricultural Customers," Energy Journal, Woo, C.K. (1985) "Demand for Electricity of Small Nonresidential Customers under Time-of-Use Pricing," Energy Journal 6(4): 115-127. Zarnikau, J. (1990) "Customer Responsiveness to Real-Time Pricing of Electricity," Energy Journal 11(4): 99-116. Appendix 2: Tables: Details of Demand Responsive
Rates and Programs
Table A.1. US Utilities Included in Survey
Utility Name
Service Area
Customers
Sales (MWh)
448,948 15,265,235 http://www.aep.com/ Michigan) Alabama Power Co 52,208,020 http://www.southernco.co Pennsylvania IOU 1,170,848 31,901,036 http://www.ameren.com/ Electric Arizona Public http://www.aps.com/home Service Baltimore Gas and 1,174,814 31,114,062 http://www.bge.com/ Electricity Bangor Hydro http://www.bhe.com/ Carolina Power & Light Co (Progress Energy) Cinergy (CG&E) 16,796,420 http://www.cinergy.com/ Illinois IOU 3,629,605 Edison Connecticut Light & http://www.cl-p.com/ Power Co Consolidated 23,517,194 http://www.coned.com/ Edison Consumers Energy 1,741,397 34,238,970 43,671,787 http://dteenergy.com Dominion Virginia 68,323,177 http://www.dom.com 53,024,862 http://www.dukepower.com Corporation Duquesne Light 439,155 9,654,461 http://www.duquesnelight.c om/ El Paso Electric 244,876 5,042,868 Florida Power and Florida IOU 4,117,229 http://www.fpl.com Light Florida Power Corp Florida IOU 1,510,494 (Progress Energy) 75,018,318 http://www.southernco.co 10,884,789 http://www.southernco.co Utility Name
Service Area
Customers
Sales (MWh)
12,351,079 http://www.idahopower.co Indianapolis Power http://www.ipalco.com/ and Light Co (IPALCO) Jacksonville http://www.jea.com/ Electric Jersey Central 18,786,247 http://www.firstenergycorp. Kansas City Power 265,829 8,256,870 http://www.kcpl.com and Light (KCPL) LADWP California http://www.ladwp.com Long Island Power 18,834,909 http://www.lipower.org/ Authority Massachusetts om/masselectric/ (National Grid) Niagara Mohawk 20,934,910 http://www.nationalgridus.c om/niagaramohawk/ 1,163,850 30,417,980 Power (Xcel) Ohio Edison (First Ohio IOU 713,508 Pennsylvania IOU (Exelon) PG&E California http://www.pge.com Potomac Electric http://www.pepco.com Pennsylvania IOU Utilities Public Service Co 1,277,525 25,845,962 of Colorado Public Service Elec 38,766,006 http://www.pseg.com & Gas Co Puget Sound 19,591,637 http://www.pse.com Energy (PSE) SCE California http://www.sce.com SDG&E California http://www.sdge.com http://www.smud.org Texas IOU 2,608,390 United Illuminating http://www.uinet.com/ (UI) Wisconsin Electric 1,033,818 24,858,918 Energies) Wisconsin Public 401,701 10,388,244 Table A.2. International Utilities Included in Survey: Basic Information
Utility Name
Service Area
Ownership
ACEA-Electrabel Italy Bewag (Vattenfall) http://www.bewag.de China Light and Power http://www.clpgroup.com Electrabel Belgium Electricidade de Portugal Portugal Private http://www.edp.pt Electricite de France (EDF) http://www.edf.fr http://www.enel.it http://www.enviam.de London Energy (EDF Energy) http://www.nuon.nl RAO-UES (United Energy http://www.rao-ees.ru System of Russia) Tokyo Electric Power Co http://www.tepco.co.jp Vattenfall Sweden Table A.3. Residential DR Rate Options Used by US and International Utilities
(includes current and historic programs)
Utility
Tariff or
AEP (Indiana Michigan) Alabama Power Co Allegheny Power (West Penn) Ameren Union Electric Arizona Public Service Baltimore Gas and Electricity Boston Edison (NSTAR) Carolina Power & Light Co (Progress Energy) Commonwealth Edison Connecticut Light & Power Co Consolidated Edison Consumers Energy Company Detroit Edison Dominion Virginia Duke Energy Corporation El Paso Electric Florida Power and Light Florida Power Corp (Progress Energy) Georgia Power Indianapolis Power and Light Co (IPALCO) Jacksonville Electric Jersey Central Power & Lt Co Kansas City Power and Light Long Island Power Authority Massachusetts Electric Co (National Grid) Niagara Mohawk (National Grid) Northern States Power (Xcel) Ohio Edison (First Energy) Tariff or
Pacific Power (PacifiCorp) PECO Energy (Exelon) Potomac Electric Power PPL Electric Utilities Public Service of Colorado (Xcel) PSE&G New Jersey Puget Sound Energy (PSE) TXU Electric Delivery United Illuminating (UI) Wisconsin Electric Power Co (WE Energies) Wisconsin Public Service (WPS) ACEA-Electrabel Uso Bewag (Vattenfall) China Light and Power Electricidade de Portugal (EDP) Electricite de France (EDF) London Energy (EDF Energy) NUON (Holland) RAO-UES (United Energy System of Russia) Tokyo Electric Power Co (TEPCO) Vattenfall Tidstariff Table A.4. Large-Customer DR Rate Options Used by US Utilities (includes
current and historic programs)
Utility
TOU RTP CPP DSS DLC CIS Hybrid
Alabama Power Co Allegheny Power (West Penn) Ameren Union Electric Arizona Public Service Baltimore Gas and Electricity Boston Edison (NSTAR) Carolina Power & Light Co (Progress Energy) Commonwealth Edison Connecticut Light & Power Co Consolidated Edison Consumers Energy Company Dominion Virginia Duke Energy Corporation El Paso Electric Florida Power and Light Florida Power Corp (Progress Energy) Indianapolis Power and Light TOU RTP CPP DSS DLC CIS Hybrid
Jacksonville Electric Jersey Central P&L Long Island Power Authority Massachusetts Electric Co (National Grid) Niagara Mohawk (National Grid) Northern States Power (Xcel) Ohio Edison (First Energy) Pacific Power (PacifiCorp) PECO Energy (Exelon) Potomac Electric Power PPL Electric Utilities Public Service of Colorado (Xcel) PSE&G New Jersey Puget Sound Energy United Illuminating Wisconsin Electric Power Co (WE Energies) Wisconsin Public Service Table A.5. TOU Tariffs Offered by US and International Utilities
Utility
Tariff Name Description
Peak/ Std
Opt- Cost Tariff Web Site
Peak Peak Off
in vs $/mon
Price Price Peak Pilot? Man Opt-
¢/kWh ¢/kWh Ratio
2-period, 1-season http://www.aepcustomer. 0.0105/kWh com/tariffs/Michigan/pdf/ credit for MISTD4-28-05.pdf off-peak only water and space heating 3 period summer, 2 13.00 http://www.southerncom period winter TOU pany.com/alpower/pricin g/bestpricing.asp?mnuO pco=apc&mnuType=res &mnuItem=ps 2 period, 2 season S V In 7.75 https://www2.ameren.co m/ACMSContent/Rates/ Rates_umbe28rt1M.pdf 2 period, 2 season http://www.aps.com/imag Service Co Arizona 2 period, 2 season http://www.aps.com/imag TOU for energy and es/pdf/ect-1r.pdf S V In 4.50 http://www.bge.com/CDA must have /Files/arschrl2.doc Electricity (BGE) Bangor 2 period, 2 season 9.36 4.14 2.3 S V/M* In 3.95 http://www.bhe.com/data *mandatory /pdf/rate_schedules/pg6r for > 2000 es_restou_0305.pdf Zeitzonen 2 period, year-round http://www.bewag.de/Pro 2 period, 2 season m/ss/customer_service/r ates/rates.asp#A5 2 period, 2 season TOU for energy and energy.com/aboutenergy /rates/01_01_05/NCSche duleR-TOUD-3.pdf 2 period, 2 season energy.com/aboutenergy /rates/01_01_05/NCSche duleR-TOUE-3.pdf 2 period, 2 season Commonwe Rate 1DR 2 period, 2 season TOU, with two-tier inverted block price 2 period, 2 season w1.coned.com/document s/elec/201-210.pdf Connecticut Rate 7 2 period, year round p.com/clpcommon/PDFs/ >1000 online/business/bill/rates/ kWh/month rate7.PDF 2 period, year round http://www.consumersen ergy.com/tariffs.nsf/ELE 5000 CTRIC_TARIFFS/67B0D customers, EF59F26908D85256F65 >750 005728BF/$FILE/elerate kWh/month s.pdf?Open 2 period, 2 season myAccount/pdfs/rates.pd f 2 period, 2 season http://www.dom.com/cust demand$6. TOU for energy and omer/pdf/va/vab1s.pdf $4.64/kW off peak Tariff Name Description
Peak/ Std
Opt- Cost Tariff Web Site
Peak Peak Off
in vs $/mon
Price Price Peak Pilot? Man Opt-
¢/kWh ¢/kWh Ratio
2 period, 2 season http://www.dom.com/cust omer/pdf/va/vab1t.pdf 2 period, 2 season http://www.dukepower.co $6.46/kw TOU for energy and m/aboutus/rates/ncrates/ peak, NCScheduleRT.PDF $3.22/kW off peak 2 period, 2 season http://www.dukepower.co NCScheduleRTE.PDF 2 period, year-round http://www.epelectric.co limited to m/internetsite/yhome.nsf/ 250 f5a02237b9a921fd87256 customers bc6005e40cf/d0a5b5aef 306646a87256bc600641 4d6/$FILE/Sch01.pdf Electricidade tarifa bi­ Demand subscription 17.13 de Portugal horaria (3.45 to 20.7 kW in 10 sp?CID=402300&LID=pt kVA to &MID=1&OID=3010000& calculate period, year-round PID=3000000&SESSID= customer TOU energy rate. o50B00I20s00x07F1W5 charge q4Ds Electricidade tarifa tri­ Demand subscription 27.33 de Portugal horaria (3.45 to 20.7 kW) + 3 sp?CID=402300&LID=pt kVA to period, year-round &MID=1&OID=3010000& calculate TOU energy rate. PID=3000000&SESSID= customer o50B00I20s00x07F1W5 charge q4Ds ENEL SPA Tariffa Demand subscription 15.28 http://www.enel.it/sportell used 3.0 (3-15 kVA) + 2 period, o_online/elettricita/tariffe kVA to year-round TOU, with _elettriche/tariffe_due_co calculate EnviaM EnviaM 2 period, year-round http://www.enviam.de/pr odukte/strom/privat/prod ukte_strom_privatkunden _enviam_basis.html 2 period, 2 season S V In 2.25 http://www.southerncom pany.com/gapower/pricin pricing in winter Idaho Power Time-of- 3 period summer, 1 http://www.idahopower.c period winter TOU om/aboutus/regulatoryinf o/tariffPdf.asp?id=264&. pdf Jacksonville Time-of- 4 period summer, 2 period winter TOU /pub/downloads/ElectricT ariff-LEGAL.pdf 2 period, 2 season p.com:80/customercare/c ache/_85256A17006827 9F_la_NJ+Part+III+2005 -0601 file_tariff_iii_eff06 0105.pdf Kansas City RTOD 3 period, 2 season http://www.kcpl.com/mot Light (KCPL) London Economy 7 2-period, year-round TOU; low period is energy.com/showPage.d o?name=homeenergy.s declining blocks witchBrand.prices.e7elec .til Long Island Rate 184 2 period, 2 season TOU, with two-tier fs/residential/resirates.pd >39,000 inverted block price by Los Angeles Time-of- 3 period, year-round http://www.ladwp.com/la dwp/cms/ladwp004844.js p Tariff Name Description
Peak/ Std
Opt- Cost Tariff Web Site
Peak Peak Off
in vs $/mon
Price Price Peak Pilot? Man Opt-
¢/kWh ¢/kWh Ratio
Massachuse R-4 2-period, http://www.nationalgridus must be .com/masselectric/home/ >2500 3 period summer and 17.06 winter, 1 period spring .com/niagaramohawk/ho me/rates/4_tou.asp http://www.nationalgridus .com/niagaramohawk/no n_html/rates_tou.pdf 2-period, year-round elijk/Images/61_15063.p df Ohio Edison Optional flat energy charge + demand charge; TOU p.com/customercare/cac periods are described he/_85256A170068279F _la_OE+Current file_O dependent rates are E_2005_PUCO_No11b.p 4 period, 2 season 2 period, 2 season m/peco/library/pdfs/s63_ complete_2005a.pdf Pennsylvani Time-of- 2 period, year-round m/NR/rdonlyres/875DF9 43-EC8E-4852-BF6A­BB28EF2168D4/0/ratertd .pdf 2 period, 2 season TOU, with customer 2 period, 2 season http://www.pge.com/tariff used in statewide pilot project 3 period, 2 season 10.41 1.1 S V/M* In 8.02 http://www.pepco.com/pd *mandatory f/dc_rate-schedules.pdf for > 2500 2 period, year round demand only TOU, m/docs/corpcomm/psco_ $6.58/kW energy is flat rate elec_entire_tariff.pdf Residential 2 period, 2 season http://www.pseg.com/cus Service Elec Load Mgt nding.jsp#anchor0 4-period, 2-season http://www.pse.com/acco 2 period, 2 season http://www.sdge.com/tm2 TOU, experimental 2 period, 2 season http://www.smud.org/co mmercial/rate_schedules 2 period, 2 season s/UnderstandingRT.pdf 2 period winter, 1 period summer TOU rivat/priser_och_avtal/el/t illsvidarepris/ 2 period, year-round 15.04 2.74 5.5 S V/M* In 2.95 energies.com/pdfs/etariff for >60,000 s/wisconsin/ewi_sheet23 kWh/yr Tariff Name Description
Peak/ Std
Opt- Cost Tariff Web Site
Peak Peak Off
in vs $/mon
Price Price Peak Pilot? Man Opt-
¢/kWh ¢/kWh Ratio
2 period, 2 season TOU; customers can icservice.com/home/help choose from 3 time options to define their TOU periods Table A.6. CPP Tariffs Offered by US and International Utilities
Utility
Tariff Name Description
Opt- Cost Tariff Web Site
Peak Peak Price vs
in vs $/mon
Price Price ¢/kWh Pilot? Man Opt-
¢/kWh ¢/kWh
Electricite de L'option Demand subscription + CPP with 2 period TOU for each of 3 categories of usage days, with day-ahead notification by utility. CPP days fall Nov 1­ Mar 31. Has load management enabling Gulf Power RSVP / CPP with 3 period, 2 http://www.southerncom Good Cents season TOU. CPP pany.com/gulfpower/prici thermostat days year round, with 30 min notification by utility, with limit of 1% CPP hours in year. Has load management enabling technology. Idaho Power Energy CPP with flat-rate http://www.idahopower.c tariff. Day-ahead notification of CPP events by utility. CPP days fall June 15-Aug 15. Has load management enabling technology. E-3, Rate A CPP with 2-period, 2­ days year round, day-ahead notification, up to 15 days/yr, 12 in summer, up to 3 consecutive days. Usage above CBL in each TOU period charged at inclining block rate. E-3, Rate B CPP with 2-period, 2­ days year round, day-ahead notification, up to 15 days/yr, 12 in summer, up to 3 consecutive days. No CBL, all usage charged at CPP/TOU rate. CPP with 2-period, 2­ http://www.pge.com/tariff Tariff days year round, day- ahead notification, up to 15 days/yr, 12 in consecutive days. Usage above 130% of CBL charged at inclining block rate. CPP with 2-period, 2­ http://www.pge.com/tariff Tariff days year round, day- ahead notification, up to 15 days/yr, 12 in consecutive days. CPP with 2-period, 2­ http://www.sdge.com/tm2 Variable starting hour and duration, notification 4 hours ahead. Limited to 90 hrs/yr total. Has load Tariff Name Description
Opt- Cost Tariff Web Site
Peak Peak Price vs
in vs $/mon
Price Price ¢/kWh Pilot? Man Opt-
¢/kWh ¢/kWh
management enabling technology. CPP with 2-period, 2­ http://www.sdge.com/tm2 days year round, day- ahead notification, up to 15 days/yr, 12 in summer, up to 3 consecutive days. Usage above CBL in each TOU period charged at inclining block rate. Table A.7. Large Customer RTP Programs (current and historic, 24 US utilities
from sample of 50)
Utility
Description of Program Eligib Std or Opt-
Adjus Source Interrupt
Time Tariff Web Site Notes
or Tariff
Pilot In or table of
Defau CBL? Hourly (Y/N)
2-part RTP: (1) fixed CBL > 1 http://www.aepc no customers charge (2) difference previo ustomer.com/tar as of 2003 between actual usage and us day iffs/Michigan/pdf CBL settled at RTP, symmetrically (3) has adders; also has interruptible option 1-part energy RTP. (1) http://www.sout 30 customers, all energy charged at herncompany.c 500 MW peak RTP, based on utility previo om/alpower/prici demand system lambda (2) has us day ng/bestpricing.a sp?mnuOpco=a pc&mnuType=re s&mnuItem=ps Ameren UE Rider RTP 2-part RTP: (1) fixed CBL all https://www2.a no customers charge (2) difference meren.com/AC had ever between actual usage and reside ahead MSContent/Rat enrolled as of CBL settled at RTP, es/Rates_umbe 2003 assymetrically (adder on incremental usage) (3) unbundled GTD adders (4) ancillary service charges 2-part RTP: (1) fixed CBL > 1 http://www.progr 85 customers charge (2) difference between actual usage and previo energy.com/abo enrollment cap) CBL settled at RTP, us day utenergy/rates/ assymetrically (adder on incremental usage) (3) ealTimePricing. pdf Path Wise 2-part RTP: (1) fixed CBL > 100 P In yes http://www.ciner 250 customers charge (2) difference ahead gycge.com/Busi originally, 140 between actual usage and ness_Services/p in 2003 CBL settled at RTP, assymetrically (adder on ervices/default_ incremental usage) (3) has adders (4) has unbundled T&D and ancillary service charges 1-part energy RTP. (1) http://www.exelo 9 customers. 12 all energy charged at ncorp.com/com MW peak RTP, calculated based on reside previo ed/library/pdfs/a demand; no index forecast with PJM us day dvance_copy_ta longer hourly load shape (2) has for day riff_revision6.pdf marketed adder (3) all non- commodity standard tariff charges apply 2-part RTP: (1) fixed CBL > 5 http://www.dom. 22 customers charge (2) difference com/customer/n and 513 MW in between actual usage and previo cbus_rates.jsp CBL settled at RTP, assymetrically (adder on incremental usage) (3) has adders; also has interruptible option 2-part RTP: (1) fixed CBL > 1 http://www.duke 53 customers in charge (2) difference power.com/abo 2003 between actual usage and previo utus/rates/ncrat CBL settled at RTP, us day es/NCSchedule symmetrically (3) has adders; also has interruptible option. Has optional pricing baseline separate from CBL. Description of Program Eligib Std or Opt-
Adjus Source Interrupt
Time Tariff Web Site Notes
or Tariff
Pilot In or table of
Defau CBL? Hourly (Y/N)
2-part RTP: (1) fixed CBL > 500 P In yes http://www.fpl.co no longer charge (2) difference m/about/rates/p exists; 20 between actual usage and previo df/electric_tariff_ customers at CBL settled at RTP, us day section8.pdf symmetrically (3) has adders RTP-DA-2, 2-part RTP: (1) fixed CBL > 250 S In yes http://www.sout 1,540 RTP-HA-2 charge (2) difference herncompany.c customers as of between actual usage and previo om/gapower/pri 2003, with 3250 CBL settled at RTP, us day cing/gpc_pricing MW of peak symmetrically (3) has for day _rates.asp?mnu demand; adders; also has ahead Opco=gpc&mnu largest in US interruptible option. 2 options: hour ahead and Gulf Power Schedule 1-part energy RTP. (1) http://www.sout 13 customers, all energy charged at herncompany.c 100-150 MW RTP, based on utility previo om/gulfpower/pr peak demand in system lambda (2) has us day icing/gulf_rates. 2003 multipliers and adders asp?mnuOpco= gulf&mnuType= com&mnuItem= er#rates Jacksonvill Schedule 1-part energy RTP: (1) all > 1 energy charged at RTP, based on NYISO LMP (2) customer-specific fixed charges based on forecast difference between RTP and std tariff hybrid RTP + CPP + TOU: > 10 http://www.firste not price (1) on-peak energy notifica nergycorp.com: responsive, charged at PJM hourly tion of 80/customercar since no RTP (2) off-peak energy custom e/cache/_85256 advance charged at PJM off-peak A170068279F_l notification of forecast (3) also has a_NJ+Part+III+2 price; created utility-designated critical peak periods, 208 0301 file_tariff customer in hours/year, additional _iii_eff030105.p 1992, and $0.34/kWh plus demand only this customer Kansas City Schedule 2-part RTP: (1) fixed CBL > 500 S In no lambda http://www.kcpl. 10 customers, charge (2) difference com/motariff.pdf 11.2 MW peak between actual usage and CBL settled at an average of RTP and standard tariff, symmetrically (3) has adders Long Island Voluntary 1-part energy RTP: (1) all > 145 P In n/a http://www.lipow 5 customers, 6 energy charged at RTP, er.org/pdfs/lipat MW peak based on NYISO LMP (2) previo ariff.pdf customer-specific fixed charges based on forecast difference between RTP and std tariff one-part RTP tariff http://www.natio nalgridus.com/ni > 2MW agaramohawk/n on_html/rates_p sc207.pdf Experiment 1-part energy RTP: (1) all > 1 http://www.xcele 2 customers, 90 energy charged at RTP nergy.com/XLW MW based on system lambd previo EB/CDA/0,3080, (2) demand charges (3) Description of Program Eligib Std or Opt-
Adjus Source Interrupt
Time Tariff Web Site Notes
or Tariff
Pilot In or table of
Defau CBL? Hourly (Y/N)
Experiment 2-part RTP: (1) fixed CBL > 30 http://www.firste 45 customers, charge (2) difference nergycorp.com/ 100-200 MW Ahead RTP between actual usage and ahead customercare/c peak demand; CBL settled at RTP, ache/_85256A1 not viewed as assymetrically (T&D adder 70068279F_la_ price on incremental usage) (3) OE+Current fil responsive, has adders; also has e_OE_2005_PU since interruptible option CO_No11b.pdf customers can withdraw on 3 days notice Pennsylvan PR-1(R), 2-part RTP: (1) fixed CBL > 2 http://www.pplel 12 customers, charge (2) difference ectric.com/NR/r 75 MW peak between actual usage and previo donlyres/2EB17 demand as of CBL settled at RTP, us day 0DE-FFF5­ assymetrically (adder on incremental usage) (3) unbundled GTD adders (4) ancillary service charges 1-part energy RTP: (1) all > 500 P In n/a http://www.pge. no longer energy charged at RTP, ahead com/tariffs/ERS. exists; max of originally based on system lambda, then on CALPX prices (2) demand charge based on std tariff (3) has PRTP,TRT 2-part RTP: (1) fixed CBL > 500 P In yes http://www.xcele expired 2004; Service Co P, SRTP charge (2) difference nergy.com/docs/ max 5 between actual usage and previo corpcomm/psco customers CBL settled at RTP, us day _elec_entire_tar symmetrically (3) has adders (4) has separate demand charges based on actual peak demand; also has interruptible option. 3 rates depending on voltage level. 1-part energy RTP. (1) > 500 S In n/a http://www.sce.c no longer open, all energy charged at ahead om/AboutSCE/R methodology RTP, based on synthetic, egulatory/tariffb obsolete; temperature based ooks/ratespricin currently 96 simulation model (2) has chargesstandard tariff 1-part energy RTP. (1) > 100 P In n/a all energy charged at RTP, based on index prices for SP15 (2) has unbundled T&D charges Wisconsin Experiment 1-part energy RTP: (1) all > 500 P In n/a energy charged at RTP energies.com/b customers based on system lambd previo usiness_new/el (2) demand charges (3) us day ec/elecrateswi.h Table A.8. DLC Programs Offered by US and International Utilities Utility Description Of Program And Customer Incentives Utility and customer share control of conditioning loads via internet-based smart thermostat, with customer having ability to override utility DLC in response to RTP signal. Customers paid monthly incentive + per hour for reductions. Utility direct control over A/C cycling using remote switch, 15 minutes on and 15 minutes off, at any time during year. Customer incentive bill credit Utility directly control over A/C and pool pumps using remote switch. Cycles for up to 8 hours per day during summer, a maximum of 20 days. 2 options for cycling time (50% and 100%), without and without pool pump. Customer receives bill credit of up to $12.50 per month. Utility directly control over A/C and water heaters using remote switch. Cycles 30 min on, 30 min off intervals for up to 8 hours per day, no limit on number of days. Pay-per-event interruptions ($4 per day), no monthly bill credit. Utility directly controls electric water heater with remote switch, cycles up to 8 hours per day year-round. Customer receives $4/month bill credit. Utility directly controls customer-specified devices through EMS system , with 2 levels of incentives based on cycling time (from 15 minutes to 4 hours) for each of 4 types of devices (AC, heating, water heating, pool pump) on seasonal basis. Combined incentives produce bill credit of up to Description Of Program And Customer Incentives Utility cycles heating and water heating during winter months through EMS system. Up to $11.50 per month bill credit, prorated for usage energy mgmt/E M_Flyer _RSL_2 .pdf Utility directly controls conditioner using remote switch, cycles up to 4 hours per day, 40 hours per month, 120 hours per season, which is June-August. Customer receives $7/month bill credit. Utility directly controls conditioner using unlimited days and September. Customer receives $5/month bill credit. Utility directly controls electric water heater and heat pumps using equipment installed at customer expense. Provides lower rates in return for unlimited days of interruption, up to 8 hours per day. 700682 79F_la_ OE+Cur rent fil e_OE_2 005_PU CO_No 11b.pdf Description Of Program And Customer Incentives Utility direct control over A/C cycling using remote switch, 15 minutes on and 15 minutes off, for a maximum of 300 hours per year; turns off water heater for up to 6 Customer receives 15% off of energy charges June through September for A/C, additional 2% for water Utility controls air conditioners, heat pumps, and water system. Two levels of incentives based on cycling time, with monthly bill credits May through September, plus per event payments. Utility directly controls A/C cycling using Customer receives monthly bill credit based on 3 levels of cycling, each with 2 options for number of interruptions (max of 15, or unlimited), and the A/C tonnage. Credits range from $0.10-$0.36/ton-day. SMUD Peak Utility controls S V In 5.00 3.00 customer A/C cycling using remote switch, June 1 to September incentive based on cycling time. Monthly credit + per event payment, ranging from $2.50/month + $1/cycle day for 27 min/hr, up to $5/month + $3/cycle day for 60 min/hr. conditioners and water heaters using remote switch. A/C cycling September, water heater cycling year-round. Bill credit $8/month for 100% A/C cycling, no credit for partial cycling; $2/month for water heater. Table A.9. DSS Programs Offered by US and International Utilities
Utility
Description Of Program Or
Customer
Tariff Web Site
Incentive
customer
Charge ($)
Customer-chosen demand level plus flat rate service. The chosen demand level tricita.it/aceael cannot be physically ectrabel_elettri exceeded due to demand cita/sportello/c ontratti.asp?na me=tariffe_uso _domestico.ht ml Customer-chosed demand level (3.45 to 20.7 kW in 10 increments) + 2 period, year- round TOU energy rate. 0&PID=30000 00&SESSID=o 50B00I20s00x 07F1W5q4Ds Customer chosen demand level (3.45 to 20.7 kW in 10 increments) + 3 period, year- round TOU energy rate. 0&PID=30000 00&SESSID=o 50B00I20s00x 07F1W5q4Ds Customer chosen demand http://particulie level (3 to 36 kVA) + flat rate rs.edf.fr/rubriq Customer-chosen demand http://particulie level (6 to 36 kVA) + 2 rs.edf.fr/rubriq period, year round TOU Demand subscription (9 to http://particulie 36 kVA) + CPP with 2 period rs.edf.fr/rubriq TOU energy rate for each of 3 categories of usage days: 300 blue days, 43 white days, 22 red days. White and red days called by utility day-ahead. Red days all fall within the period Nov 1-Mar 31. Customer chosen demand level (3-15 kVA) + flat tariff. l.it/sportello_o nline/elettricita/ tariffe_elettrich e/tariffe_due_c osti.asp Customer chosen demand level (3-15 kVA) + 2 period, l.it/sportello_o nline/elettricita/ tariffe_elettrich e/tariffe_due_c osti.asp Customer-chosen demand level plus flat rate service. The chosen demand level cannot be physically exceeded due to demand Description Of Program Or
Customer
Tariff Web Site
Incentive
customer
Charge ($)
Customers subscribe to a level of demand less than their calculated level, in return for monthly bill credit during summer months. During high load periods, customers were required to reduce load to subscribed level for a period of up to 4 hours or all service was remotely curtailed.

Source: http://smartenergydemand.eu/wp-content/uploads/2011/05/E3-Survey-of-Dynamic-Pricing-Programs-Jun-2006.pdf