Modeling the Electricity Spot Price with Switching Regime Semi-Nonparametric Distributions

Abstract

Spot prices of electricity in liberalized markets feature seasonality, mean reversion, random short-term jumps, skewness and highly kurtosis, as a result from the interaction between the supply and demand and the physical restrictions for transportation and storage. To account for such stylized facts, we propose a stochastic process with a component of mean reversion and switching regime to represent the dynamics of the spot price of electricity and its logarithm. The short-term movements are represented by semi-nonparametric (SNP) distributions, in contrast to previous studies that traditionally assume Gaussian processes. The application is done for the Colombian electricity market, where El Nino phenomenon represents an additional source of risk that should be considered to guarantee long-term supply, sustainability of investments and efficiency of prices. We show that the switching regime model with SNP distributions for the random components outperforms traditional models leading to accurate estimates and simulations, and thus being a useful tool for risk management and policy making.