FERC's proposed standard market design (SMD) has generated substantial interest in the topic of demand response in the US. Time-varying retail pricing provides a natural method for increasing the price responsiveness of customer loads, particularly given the growing acceptance of advanced metering as a cost-saving opportunity. However, recent focus has shifted to load reduction programmes operated by regional ISOs. We urge caution in estimating the benefits of DR programmes and setting DR payment levels.

Demand response, defined as changes in customer loads in response to time-varying retail prices or load reductions bid into wholesale markets, serves an essential role in well-functioning power markets. Load levels in the US have been largely non-price responsive because end-use customers typically face fixed retail prices. As a result, regional electricity markets have experienced wholesale price spikes, price volatility, and concerns about system reliability, economic inefficiency and market power abuses.


The US Federal Energy Regulatory Commission (FERC), recognizing the important role that demand response (DR) can play in keeping markets balanced, has advocated a standard market environment in which DR participates in energy and ancillary services markets on an equal footing with generation resources. While FERC envisions an important role for DR, its proposed standard market design provides few details. This is the case in part because retail customers are the source of demand response, which is delivered to the wholesale market by their energy supplier, or load serving entity (LSE). And, retail service and LSEs (most of which remain regulated utilities) come under the regulatory jurisdiction of the states rather than FERC in the U.S. Instead of directly influencing retail pricing design, FERC hopes to establish wholesale market infrastructures that will facilitate the development of DR mechanisms by the LSEs.

The proposed SMD contains key infrastructure elements that would encourage demand response, including the imposition of locational marginal pricing (LMP) and the establishment of centralized day-ahead and real-time markets for energy, ancillary services and transmission services. LMPs provide appropriate price signals that account for both marginal generation costs and transmission congestion at particular sites. Day-ahead energy markets would consist of voluntary, bid-based, security-constrained markets in which generators and energy suppliers are free to engage in bilateral transactions. Day-ahead markets create transparent market-clearing prices. Buyers and sellers can take actions appropriate to their changing expectations of high or low real-time prices, such as offering to reduce load in high-cost hours. The reaction of market participants to changing wholesale prices improves the liquidity of the market, encourages efficient use of resources, and may reduce price volatility in both the day-ahead and the real-time markets.


Most types of demand response mechanisms need the hourly interval data made possible by advanced metering systems, so that customers' actual energy usage can be measured and billed. The cost of providing such data has served as a historical barrier to wider adoption of advanced time-of-use (TOU) and real-time pricing (RTP) programmes, particularly for smaller customers. The recent growth in AMR installations for cost-saving reasons unrelated to retail pricing has reduced some of the existing barriers to more efficient pricing, and offers the potential for utilities to offer new services to their mass-market customers.

For example, Gulf Power Company in Florida is expanding its critical peak TOU pricing programme, known as GoodCents Select, finding that customers like both the opportunity to save money and the flexibility provided by the program's communication and pre-programmed control technology. During 85% of the hours of the year, Gulf Power customers see prices that are lower than under their standard tariff. Only in peak periods and the infrequent (1% of hours) critical periods do they face higher prices. Gulf Power has seen substantial load reductions during peak and critical peak periods, producing bill savings to customers and cost savings to the utility.


Economists have long argued that the frequent divergence between varying wholesale power costs and fixed retail prices provides substantial opportunity for cost-saving benefits to both consumers and their energy suppliers. The typical extent of cost-price differences is illustrated in Figure 1, using PJM data for the summer of 2000. The distribution of hourly costs is highly asymmetric, with many low-cost hours and relatively few high-cost hours. Resources are wasted in both low-cost and high-cost hours under fixed pricing. Consumers have no access to frequent low-cost power that suppliers could profitably offer at a price lower than the fixed retail price. And in the relatively few high-cost hours, suppliers incur costs far in excess of the fixed retail price to deliver power to consumers, some of whom would be willing to reduce consumption for a payment of even a fraction of the supplier's cost of power.

Distribution of hourly pricing

Dynamic pricing practices such as critical peak TOU and RTP, which set prices to reflect changes in marginal costs, provide incentives to customers to reduce usage in high-cost hours and increase usage in low-cost hours. Both types of load change produce cost savings to responding customers and their supplier. 

Georgia Power Company's RTP programme provides a notable example of successfully producing demand response by dynamic pricing. Approximately 1,700 commercial and industrial customers representing 5,000 MW of maximum demand currently participate in this 10-year old programme. Georgia Power estimates that it can achieve load reductions in the range of 750 MW during occasional high-cost summer hours. At the same time, RTP customers gain access to low-cost power during the majority of hours of the year.


Unfortunately, dynamic pricing programmes such as Georgia Power's RTP and Gulf Power's CP TOU are rare in the U.S. As a result, recent efforts to expand the amount of demand res-ponse in regional markets have focused on DR programmes run by the independent system operators (ISOs) such as PJM, New York ISO and ISO New England. In these programmes, end-use customers offer load reductions through their LSE at particular price levels in day-ahead and/or real-time markets. The LSEs are then paid by the ISO for load reductions below a baseline level at prices that reflect system costs, and they in turn pass on a portion of that payment to the end-use customers. A major issue in designing and operating DR programmes is the method used to calculate customers' baseline loads. In principle, they represent what the customers' loads would have been if they were not offering to provide load reductions. However, the process is subject to gaming possibilities and claims of unfairness by customers.

Furthermore, the process of shifting the focus from dynamic pricing to DR programmes appears to have deflected attention from the traditional source of cost-saving benefits that result from DR load reductions, to the perceived benefits associated with wholesale price spike reductions. These potential price reductions are claimed to imply lower costs for LSEs and their customers, thus leading some advocates of DR programmes to argue that DR should be subsidized by all consumers through payments for load reductions that effectively exceed the wholesale market price.

However, treating wholesale price reductions as immediate cost savings to LSEs is a fallacy, which can be seen by noting the prevalence of bilateral wholesale forward contracts at fixed prices. These contracts lock in LSEs' energy costs for most of the load that they expect to serve. Thus, reductions in wholesale prices in the current period have little direct effect on LSEs’ costs. That is the purpose of financial forward contracts – to manage the risk associated with price volatility. Subsidizing DR beyond market-based payments is a policy that could well lead to higher rather than lower market costs.


Increasing the price responsiveness of electricity demand by exposing more consumers to dynamic retail prices or market-based demand response payments can only improve the efficiency of wholesale electricity markets. Either approach can produce cost-saving benefits to the participating customers and their energy suppliers, both of whom have economic incentives to participate and to provide price-responsive load. However, ISOs should pay no more for load reductions than they do for generation – that is, the wholesale market price.

Finally, designers of demand response programmes should look to the logic of two-part RTP as a standard for market-based DR. First, consumers' baseline loads are established beforehand and are not subject to gaming during any particular price spike episode. Second, consumers face prices that signal the marginal cost of power, and have an economic incentive to reduce load whenever the cost exceeds their cost of curtailing. Third, consumers' risk from facing time-varying prices is managed through a fixed price on their baseline load.