Experimental Economics and Retail Electricity Deregulation: Demonstrating the Benefits of Choice
Regulatory change in electricity moves slowly, in part because of human dislike of change and aversion to risk. Convincing people that regulatory change is worth undertaking would be a lot simpler if we could demonstrate the possible outcomes of a change, and thereby that such change need not be disastrous and could even create value and opportunity for many people. But electricity networks are so complex, and so expensive to construct, that real-world experiments are very costly and unlikely to occur. It’s not like opening a new restaurant and seeing if consumers perceive value in the offerings of food, wine, service, and environment. In this context, the methodology of experimental economics is incredibly useful and can help us place bounds on what may occur if we implement fuller retail electricity deregulation. Experimental economics simulates environments having different rules, with real economic agents facing real choices, and with the potential to earn real money payoffs. Because rules influence incentives and therefore shape outcomes, rules and the institutions that create and enforce them matter, including regulatory institutions. Experimental economics can inform us about the relationships between different rules and institutions and different outcomes.
An experimental approach to analyzing electricity deregulation enables the abstraction of some features of real-world activity to focus on specific features. The recent electricity experiments of Vernon Smith, Bart Wilson, and Steve Rassenti, of George Mason University and the Mercatus Center, illustrate the power of experimental methodology to create information about what is likely, and what we cannot predict, from different features of electricity deregulation. They have created a portable laboratory setup, complete with laptops, wireless networking, and well-designed software, to perform electricity market experiments in a variety of locations, with diverse participants. This description of experimental economics from the Interdisciplinary Center for Economic Science website provides a good overview of experimental economics and how it helps us understand how markets work. Smith’s entry on experimental economics in the Palgrave Dictionary of Economics is also a useful description.
Recently they have been running electricity experiments with groups ranging from college students to Congressional staffers to federal energy regulators. The people in the experiment are electricity generators, participating in a wholesale generation market. They are the supply side of the market, and they can own different types of generation capacity — baseload (low cost), intermediate cost “load followers”, and high-cost peaking units. These levels realistically reflect a typical supply curve, in which generators run their least expensive units until they hit capacity, then move to the intermediate-cost plants, and only run the expensive peaking plants during periods of peak load. Generators bid by submitting schedules of asking prices for their capacity in a given period, and all generators receive the market-clearing price. These rules mirror those found in many wholesale electricity markets. Generators can also have market power, depending on what kind of capacity they own and how concentrated that ownership is.
The experimental design can also vary the rules governing the expression of demand, either to reflect the current fixed “must serve” demand, or to test possible regulatory changes in how consumers are allowed to express their demand. Since state regulation of the electricity industry commenced in 1907, retail customers have faced average rates that change infrequently. Retail electric service is provided on a guaranteed-price basis, under the regulatory “obligation to serve” remit. In terms of consumer expression of their demand, these regulated rates have meant that consumer demand signals are metered, aggregated, and transmitted to suppliers on a monthly basis. With such unchanging rates, the demand or typical aggregate load profile fluctuates greatly across the day.
The response of the typical consumer may be “so what?” Because of the “obligation to serve” requirement facing utilities as part of traditional retail regulation, utilities must be able to generate or buy enough power to satisfy peaks throughout the day. Put another way, the responsibility for satisfying all consumers, whenever they want power, rests with the suppliers. Fixed, regulated rates mean that consumers have no incentive to take on any of that responsibility. The result of this supply-focused approach is lots and lots of generation capacity, because suppliers are required to serve all demand, whenever it occurs, without changing prices to reflect the different costs of serving that demand at different times. Retail prices cannot change even though costs do change, as captured in the three different types of plants used by the industry and in these electricity experiments.
In presenting the economics underlying the power of consumer demand in electricity markets, Smith analogizes between the electricity industry and other industries, particularly the airline and hotel industries. All three are service industries, facing peak demand that fluctuates and that determines capacity, with substantial capital investment requirements to satisfy demand. In competitive markets for airline travel and hotel rooms, where both consumers and producers can provide and respond to price signals, rates typically go up in peak demand periods and plummet in off-peak periods. The high rates in peak demand periods, rates that certainly exceed marginal cost, pay for the capital that is necessary to satisfy the peak, and the interaction of these price signals lead to optimal capacity investment.
That doesn’t mean, though, that all customers who want a seat or a room at peak will get it at a price they are willing to pay. So what do we all do when facing high airfares or hotel rates? We time-shift, traveling on a different day or at a different hour. Hotels and airlines do not operate under a regulatory obligation to serve, yet consumers deal with the fact that they might not be able to consume the flight or the hotel room they want when they want at the price they want. They deal by shifting their demand to different times, trading off convenience for cost depending on their individual preferences. Thus the comparison with the airline and hotel industries reveals exactly the extent to which the “electricity cannot be stored” rationale for regulation is a canard — airline travel and hotel service cannot be stored either, yet no one is arguing that these industries should operate under “must serve” obligations like those in the electricity industry.
In this article in Regulation in Fall 2001, Rassenti, Smith and Wilson compared two bidding systems in a wholesale electricity market — one with only supply-side bidding, and one with both supply-side and demand-side bidding. The demand side of the experiment proceeds as follows: take a very simple rule by which consumers can choose whether or not to let the retail electricity supplier interrupt their service at a couple of different points, and see what effect that rule could have on the outcomes in the wholesale market. They then divided the demand into four types: must-serve demand, off-peak demand, shoulder demand, and peak demand. Under fixed retail rates, all demand is essentially must-serve demand, including the high peaks. One of the important things to learn in this experiment is whether allowing consumers to choose to have their demand interrupted at two different prices would lead to increased consumer benefit, increased supplier profits, and any change in the ability of suppliers to exercise market power in the experiments when they have it. As Rassenti, Smith and Wilson say in their Regulation article, “our small simplifications enabled us to focus on the key issues we wanted to study while still capturing the essence of the daily natural cycle in demand in all electrical delivery systems.”
The generators then have to choose prices at which they bid into the wholesale market; in some experiments the generators face perfectly inelastic must-serve demand, and in some experiments they face consumers who can choose to have their service interrupted. In their experiments the timeframe is several days (compressed into a few hours), so the generators experience bidding over the fluctuating demand cycle. And, at the end of the day, the participants get to keep their profits (with some modifications when performed with government employees), so the incentives are real.
So what happened? When there was no demand-side responsiveness, suppliers with market power were more able and more likely to exercise it by withholding capacity. In the experiments with both demand-side and supply-side bidding, suppliers with market power were not as able to exercise it, and price fluctuations were smaller. Not only were average prices lower, but the variance of prices was also lower; demand responsiveness reduced price levels and price volatility, even in the face of supplier market power. When suppliers did not have market power, demand responsiveness still led to lower and less volatile electricity prices.
In their Regulation article, Rassenti, Smith and Wilson report these results from experiments performed at the University of Arizona. I have also seen similar results from experiments performed at the Federal Energy Regulatory Commission — yes, when facing profit incentives and no demand response, even the regulators exercised market power when they had it.
Experimental economics methodology improves upon “blackboard economics” in reflecting what Michael Polanyi calls tacit knowledge — when we make choices, including social interaction choices like market exchange, we are not always conscious of all of the information and knowledge that we bring to bear in making these choices. Experiments with real people facing real incentives create an environment in which the effects of tacit knowledge are not assumed away to solve the equations on the blackboard.
Market-based retail pricing is a crucial component of the ability to deliver choice and value to customers. Fixed, regulated average rates are an obsolete relic of a regulatory approach that, if it persists, will stifle the application of creativity in this industry. If utilities, regulators, and politicians consider the possibility that utilities can offer different value propositions to their customers than just “juice coming through the wall”, utilities can benefit from using market-based pricing as a tool for offering an attractive portfolio of service options to their customers. Creating value from this change, though, requires vision, and getting the transitions and the institutions right can be extremely tricky. Consumers will change how they think about buying electric service, and what that service is, exactly. For that change to occur, politicians and regulators will have to act on the leadership and vision that would allow consumers to take responsibility for their individual purchasing choices.