By Ahmad Faruqui
Electric Power Research Institute
Palo Alto, California, USA

A real-time pricing agreement ensures that consumers pay a variable price for power that changes in proportion to its price in a wholesale spot market. Could this be the answer to California’s ongoing power problems?

California, a state with a population of 34 million and the world’s sixth largest economy, is dealing with a power crisis that has seen a four-fold increase in wholesale electricity costs in just one year, from $7 billion in 1999 to $28 billion in 2000. The costs are expected to rise to $70 billion this year. Stage three alerts, in which reserves fall below 1.5 per cent and blackouts become likely, have become a common occurrence. According to the National Electricity Reliability Council, the state can expect about 260 hours of blackouts this summer.

There is no easy solution to this very complex problem. The underlying fundamentals of demand and supply are out of balance, since demand has been rising by about 1000 MW each year over the past decade, and supply has been essentially stagnant. Since the power market was restructured in 1998, no new supply has come on-line. On the other hand, wholesale costs have mushroomed due to a variety of reasons, including rising natural gas prices and low hydro conditions. Until recently, electricity rates were frozen at their 1997 levels, as part of the state’s deregulation law. Thus, the normal mechanism for bringing supply and demand into balance, a price increase, has been unavailable.


The effect of RTP on wholesale market prices: the impact of enabling technology.
Click here to enlarge image

In the near term, the best solution is to raise retail rates to reflect some portion of the increase in wholesale power costs. California’s public utilities commission has moved in that direction, but the rate increase – while large in comparison to consumer expectations – may be insufficient to bring demand and supply into balance or to restore the state’s two leading investor-owned utilities to financial solvency. Politically, however, there are real limits to how much retail rates can be raised.

There is however, one solution that may provide some relief in the short term that is not politically infeasible. That is real-time pricing (RTP) of electricity. Under this arrangement, customers pay a variable price for power that moves in proportion to its price in the wholesale spot market. Often, this pricing scheme focuses on large power customers who have the necessary interval metering systems already in place, or where it can be put in place at relatively low costs. For example, the state of California has estimated that real-time meters can be put in place for about $35 million for the state’s largest 18 000 customers.

Demand response

RTP lets customers respond by reducing their usage during expensive periods and increasing their usage during inexpensive periods. This concept, called demand response, deals with a key deficiency in California’s market design – the disconnection between wholesale and retail markets. Regardless of how high wholesale prices climb, most customers see the same fixed retail price, and therefore have no incentive to reduce consumption. If some consumers – particularly the large commercial and industrial customers who already have interval meters – faced hourly prices that reflected wholesale market prices, then their demand response during periods of tight capacity and high prices would help relieve resource constraints and hold down market prices.

EPRI has conducted a study on the market potential for RTP in California. The results were based on demand response data from existing utility real-time pricing (RTP) programmes in the US and the UK, and actual California data from summer 2000. These results show that customer demand response to hourly, market-based retail prices could generate load reductions of 1000 to 2000 MW, reduce summer peak prices by six to 19 per cent, and produce energy cost savings ranging from $0.3-$1.2 billion. Demand response would bring at least short-term relief to next summer’s likely problems while other efforts are put in place to solve the longer term financial and resource issues.


Georgia Power Company’s RTP programme has achieved load reductions of 750 MW on very high price days.
Click here to enlarge image

While unresponsive demand is often cited as a contributing factor to high prices in times of capacity constraints, some ask whether customers will actually respond to changes in hourly prices. Fortunately, ample evidence is available from recent studies of existing RTP programmes to confirm that both commercial and industrial customers respond in a consistent and predictable manner to changing hourly prices, particularly at extremely high levels.

Georgia Power Company operates the largest RTP programme in the US, with more than 1600 commercial and industrial customers accounting for as much as 5000 MW of demand. Georgia Power estimates that it achieved load reductions ranging from 400 to 750 MW on moderate to very high price days in 1999. One group of the most responsive Georgia Power customers reduced load by 30 per cent during periods of moderately high prices ($.30/kWh), and 60 per cent in the few hours in which prices exceeded $1.00/kWh. On average, customers can be expected to reduce loads by about 17 per cent.

Data analysis

EPRI’s analysis used data for the summer of 2000. However, the results can serve as an indication of summer 2001. To conduct the analysis, EPRI used its customer demand modelling software, along with empirical parameters drawn from EPRI’s StatsBank database (a storehouse of information compiled by EPRI over the last five years that summarizes findings on hourly price elasticities from RTP programmes from a number of utilities in the US and UK). They combined this information with publicly available data on hourly customer loads in California to estimate demand response in the California market. EPRI applied these hourly demand response estimates for the summer months to an econometric model that relates California’s wholesale market (PX) prices to total load levels in the state and other relevant factors, thus producing estimates of price reductions and cost savings.


Price-responsive demand should be a key element of mending California’s broken energy market.
Click here to enlarge image

It is useful to illustrate the effect of demand response by evaluating it in one prototypical high-price hour, such as an hour in which prices hit the June price cap of $750/MWh. Because of the uncertainty in such analyses, EPRI performed simulations for three alternative scenarios, representing different combinations of hourly pricing market share and price responsiveness assumptions.

In the medium scenario, demand response produced a load reduction of approximately 1000 MW, or 2.5 per cent of the total ISO load. Applying this load reduction to a wholesale price model based on the PX cost curve, EPRI estimates that a load reduction of that magnitude would have reduced prices by $160/MWh, or 24 per cent. The resulting energy cost savings, applied to the entire ISO load in that hour, would have reached $6.7 million.

It is also useful to study the total effect of demand response for the entire summer period. In the medium scenario, EPRI estimates that peak period prices would fall by an average of nearly $25/MWh, or 12 per cent, implying that energy costs would drop by $700 million, or almost ten per cent. Under the high scenario, with a larger market share and greater price responsiveness, peak prices fall by nearly 20 per cent, and costs are reduced by $1.2 billion, or 16 per cent.

Real-time technology

These results illustrate the potential benefits of hourly pricing in a tight power market.

Market-based interruptible programmes, as well as spot pricing programmes, will benefit from the application of new monitoring, energy control, and communication technologies. Technologies that improve the ability of customers to respond to hourly prices in an automated fashion, with predetermined strategies that cause minimal disruption, will enhance demand response and customer benefits. For example, EPRI has developed a prototype Automated Energy Control System that will offer customers energy management flexibility in the presence of hourly pricing.

The Automated Energy Control System is a striking example of the effect of enabling technology. The programme included a communication and control device that allowed customers to pre-programme their response to both traditional TOU prices, as well as occasional “critical” prices. The customers’ price responsiveness was found to be twice as great as in most previous studies of TOU price response, where no such control technologies were involved.

Other hourly pricing incentives

There is a potential incentive for many more customers to respond as if they faced hourly prices, even if they do not do so contractually. This incentive, which has not been appreciated to date, arises as more customers have their energy usage metered on an hourly basis. When hourly metered data is available, suppliers will be able to calculate the cost to serve customers for a recent historical period, and adjust their price offers accordingly (i.e. charge higher prices to those customers whose usage tends to be high during high-cost hours).

When customers understand this process they will have an incentive to reduce load during high-price periods even if they do not face hourly prices explicitly. In a recent step in this direction, Puget Sound Energy is installing hourly meters for nearly half of its customers, as well as providing information on hourly energy costs, though not billing at hourly prices.

With all of the benefits to be gained from hourly pricing, and nearly everyone agreeing that price-responsive demand is a key part of the process of mending California’s broken energy market, a natural question is: Why hasn’t hourly pricing been adopted to any degree? A number of barriers and misconceptions appear to be delaying the move toward implementing hourly pricing.

The most notable problem arises from lack of experience with hourly pricing. Operating an hourly pricing programme requires hardware and software for communicating prices to customers (e.g. on a day-ahead basis), metering customers’ energy consumption on an hourly basis, and billing customers based on hourly prices and usage values. The major California utilities have run pilot RTP programmes in the past, but have not expanded them. However, some of the key features are in place (e.g. most large commercial and industrial customers already have hourly meters installed), and expertise on implementing RTP is available.

What does it all mean?

EPRI has argued that ample evidence is available to project how customers will respond to hourly electricity prices, and that their demand response can help connect California’s wholesale and retail power markets, and in the process reduce prices.

In addition, EPRI has shown that concern about rate caps and potential bill increases does not have to stand in the way of implementing hourly pricing. Instead, available financial risk management mechanisms can allow customers to lock in regulated price levels on their baseline level of usage. This will limit (though not eliminate) bill changes, while at the same time providing the desired incentive for demand response during periods of high prices.

EPRI analyzed a full-time hourly pricing approach similar to the ones already implemented by a number of utilities in RTP programmes, and focused only on commercial and industrial customers

However, there are other types of price structures that can provide hourly price signals during certain periods of high wholesale prices, and can be applied to smaller customers and involve new communication and control technologies. Examples of this include demand-side bidding programmes, and California’s ISO’s new discretionary load curtailment programme, which may effectively replace the utilities’ current traditional interruptible programmes once customers are allowed to leave them.

It should also be recognized that even non-hourly market-based pricing targeted at smaller customers could produce needed general demand response. For example, increasing block-rate pricing with a price cap for initial baseline consumption, and market-based prices on higher levels of usage would provide a strong conservation incentive.

The exact form of pricing that is implemented is of less importance than the fact that some form of demand response is added to the California power markets. Price-responsive demand will not solve California’s power problem completely; additional generating capacity is clearly needed, and the extent of price manipulation by market power is still under study. However, demand response can help matters in the short-term, and can be implemented more quickly than needed capacity can be added, and it can provide continuing long-term market efficiencies.