Nonlinear time series analysis on the offer behaviors observed in an electricity market

Abstract

In electricity markets where supply and demand drives the price for the purchase and sale of electricity, generating firms change capacity for various reasons including load level, policy, and varying market conditions. These types of fluctuating production patterns can result in the reduction of market efficiency. In an inefficient market, where the price for electricity exceeds marginal cost, the locational marginal price (LMP) is often used to measure market efficiency. Stochastically driven changes in the market are captured by this approach, however, these random changes (frequently observed in efficient markets as well) do not affect market efficiency in the long run. Conversely, a slow, consistent change is not captured by the snap-shot approach and affects the efficiency significantly. Therefore, it is necessary to construct an algorithm that captures only consistent changes that truly affect market efficiency. Fractal analysis can characterize a price behavior in the electricity markets because the price exhibits a self-similarity. 1 1 A self-similar object is exactly or approximately similar to a part of itself. Once a system undergoes a change, the fractal dimension of the system reflects the change. In this paper, an approach using nonlinear time series analysis is proposed and tested on actual offer behavior observed in the electricity markets in the United States.