Test of asymmetry effect of demand on spot price using MCMC methods

Analysing and forecasting spot price has become a task that cannot be overlooked since the deregulation of the electricity market in Australia. There are many factors that affect electricity spot prices. Demand, especially, as one of the fundamental drivers, plays an important role as prices are decided according to the balance of the supply and demand. Financial economists have studied the temporal relationship between price and the trading volume in most commodity markets. The relationship implies the validity of a number of hypotheses, including sequential information arrival hypothesis and the market inefficiency, in the sense that volume can be used to predict future price changes. Such research finding can be used to design trading strategy for market players. Previous studies were either mainly concentrated on price itself or the direct relationship between demand and price, if they incorporate demand into spot price forecasting. In this paper, we attempt to rind out whether or not positive and negative changes in demand have the same effect on spot price. As electricity spot prices possess the characteristics of mean-reversion, stochastic volatility and occasional jumps, we considered a mean-reverting asset pricing model that allows the conditional variance to incorporate log-stochastic volatility and a Poisson-directed jump process. Due to the complexity of the model, conventional estimation methods, such as Ordinary Least Squares, Maximum Likelihood and Quasi Maximum Likelihood, do not suit anymore. To accommodate this, Markov Chain Monte Carlo (MCMC) method, combined with Bayesian inference, is chosen to estimate the chosen model. MCMC methods have been reported to outperform the conventional estimation methods especially in estimating asset pricing models with stochastic volatility. The data we used in this study is the peak 30-minute spot price and demand from 2001-2005 in the Victoria region, Australia. The findings should help predict future electricity spot prices and strengthen any trading strategy based on price-demand relationship.