Evaluating Alternative Ways of Mitigating High Prices in Electricity Markets Using Multi-Agent Simulations

The main objectives of this paper are 1) to determine which characteristics of a deregulated wholesale market for electricity make price spikes likely to occur, and 2) how effectively these price spikes can be mitigated by different modifications to the structure of the market. The analysis employs a Multi-Agent Simulation (MAS) to replicate a spot market with six supply firms, represented by adaptive autonomous agents. These firms submit offers to maximize their own expected profits, and an Independent System Operator (ISO) clears the market for a predetermined load in a uniform price auction. The firms learn about the structure of the market and the behavior of their competitors by comparing actual market outcomes with predicted outcomes based on an estimate of their own residual demand curve. This estimated demand curve is updated each period using a Kalman filter. An additional objective of the paper is to show that a MAS provides a useful analytical framework for evaluating the effects of specified modifications to a deregulated electricity market.
The two most important characteristics of the market for creating price spikes are 1) uncertainty about the system load is the primary determinant of the observed speculative behavior (i.e. submitting offer curves shaped like a hockey stick), and 2) all firms eventually become speculators, and it is unrealistic to expect firms to behave like price takers in a market with only six suppliers. In other words, given the intrinsic uncertainty about the load (and the possible outages of generators), speculative behavior by suppliers is entirely rational.
With six suppliers controlling the same amount of generating capacity in a MAS, high prices can eventually be sustained in the market, and the resulting price behavior does not replicate the relatively infrequent price spikes seen in a market like PJM. Using the actual pattern of daily peak load in the summer of 1999 for PJM, observed price behavior can be replicated if 1) two of the six firms are vertically integrated, and 2) two of the firms are each replaced by four small firms that behave more like price takers. Since this modified structure of firms is roughly the same as the actual structure of firms in PJM and the simulated price volatility is very similar to the actual volatility, this modified structure of firms is used as the base case for evaluating different ways of mitigating high prices.
The main results are that it is impractical to rely on any one of the following policies to eliminate speculative behavior by suppliers: 1) more suppliers, 2) vertically integrated firms, 3) capacity payments (to discourage withholding of capacity), 4) requiring suppliers to hold forward contracts (if the contracts have to be renewed), 5) additional fixed contracts for imported power, or 6) increasing the base load capacity of each supplier. Individually, each policy helps to mitigate high prices, but not enough to ensure that prices are competitive unless an unrealistically extreme form of the policy is implemented. In contrast to the other policies, price-responsive load is shown to be an effective way to mitigate high prices, particularly if there is some response at relatively low prices. Overall, the results show that the analytical framework of a MAS can replicate actual behavior in a deregulated electricity market, can provide new insights into how suppliers behave, and can evaluate a wide range of policy options for mitigating high prices.