Price Spike or Price Cap? An experimental investigation of the electricity market

During the winter 2009-2010, the Nordic electricity market experienced several price spikes, usually attributed to the cold weather but also to a low availability of nuclear capacity. These price spikes have been heavily debated in media and some commentators have raised the concern that these events might be taken as evidence of market power. Others claim that it is the natural result of capacity shortages during peak load periods and that price peaks constitute a “fair” reward for peak load capacities rarely in use and costly to build. It is therefore not obvious that price spikes call for intervention.

Policymakers have always stressed the importance of creating incentives for adequate investment in electricity generation but nevertheless also require low, non-volatile prices. Price caps are one of the regulatory tools to achieve the latter of these objectives but may inhibit adequate investment levels. Absent regulatory intervention, the market provides adequate investment but only thanks to price spikes. Price spikes may be efficient as they carry information about scarcity, signaling that generation capacity is in short supply and thereby encouraging investments.

This research project comprises experimental investigations in the laboratory of the electricity market. The aim of the analysis is a better understanding of the efficiency properties of price caps and of price spikes.

Description of the project:

To study the efficiency properties of electricity markets, and in particular the impact of price caps, it is necessary to have reliable cost data. Some empirical studies rely on historical data from the regulated era, but these data start to be old and are therefore not fully reliable. This project plans to design an experiment where the cost structure of the firms (also referred to as subjects) is perfectly known. This approach will enable us to compare market participants’ bids depending on the size of the price cap.

Several real world features of electricity markets inspire the experimental design. In our setup four identical generating subjects (playing the role of firms) interact in repeated multi-unit auctions competing on the supply side. The demand is perfectly inelastic and volatile, both with a deterministic element (with periods of peak load and off peak demand) and a stochastic element (as demand varies around the peak and off peak demand). Subjects have to make two types of decisions. Initially, they have to decide on how much capacity to make available to the market. Then they submit repeatedly multistep supply functions, i.e. schedules of quantities and prices specifying how much they are willing to supply for a given price. They cannot offer their units above a maximum price, exogenously given in the experiment. On the other hand there is no obligation to produce at full capacity. Subjects participate in different treatments with different price caps (from zero to well above the highest marginal cost). The above experimental setup will generate different groups of subjects. The pairwise comparison of these groups will enable us to address the distribution of surplus between producers and consumers as well as the performance of a price cap, both in terms of allocative and cost efficiency.

Expected results:

Below is a description of the results that can be obtained from the experimental design outlined above.

1. The option value of the price cap. Like in real world electricity markets, subjects decide on their costly capacity constraints when demand is uncertain. This is similar to a choice of irreversible investments under uncertainty, usually modeled with a real options approach. In this case one may expect price cap regulations to have an adverse effect on investment. With capped prices, subjects might indeed reduce the number of units that they make available on the market. Such an effect of the price cap can be estimated, by comparing treatments with different price cap levels.
2. The strategic effect of the price cap. If price caps limit the marginal return of each generating unit, firms might want to preempt their competitors by offering substantial available capacities before competing on the market. Allowing subjects to extend their capacity constraints during the auction game could test the presence of such a strategic effect. If subjects follow a preemption strategy they should choose their capacity at an early stage and at a significant higher level than in the basic setup.
3. Price caps as Focal points. The literature on price ceilings suggests that caps may lead to higher prices by providing a focal point for tacit collusion. To test the relevance of this concern, we will compare treatments with and without binding price caps. The concern would be validated if subjects are able to collude only in the treatment with a binding price cap.
4. Price volatility and price cap. Firms in each market encounter the same demand conditions several times. Thus we can ask how much prices fluctuate when the same set of firms are confronted with the same demand level repeatedly but with different price caps.
5. Students vs. Practitioners. Students but also practitioners will participate in our experiment, so we should able to compare the different outcomes. The project has established contacts with market participants in Singapore and in France.