STOCHASTIC MODELING OF ELECTRICITY SPOT PRICES WITH MEAN REVERSION AND JUMP DIFFUSION

Abstract

The electricity market is one of the most established financial markets today with enormous trading volumes. It is also a crucial mechanism in ensuring the smooth transaction and delivery of electricity from power plants to households and factories. We look at some of the relatively new models that have been proposed for electricity markets, and compute corresponding forward prices. The most basic models are more generic commodities pricing models which capture the mean-reversion observed in commodity prices. These models do a reasonable job of representing the price data. However, we see many sharp spikes in electricity markets that very rarely occur in other commodity markets. Hence, we believe that a more complex model is required to model the spikes. We also study one such jump-diffusion model which attempts to capture the spikes in electricity prices. Following which, we look at PJM hourly price quotes for the past 15 years, between 2001 and 2015, and fit each model to the prices. This study will attempt to find the model that best describes our data set, by using the Maximum Likelihood Estimator as the basis for comparison. We also seek to improve on the current jump-diffusion model by incorporating an additional factor in modeling prices. The additional factor will add one more dimension of predictability for modeling future prices, and can be useful for hedging in electricity markets.