Stochastic modelling of volatility and inter-relationships in the Australian electricity markets

Abstract

To model the price, price volatilities and their inter-relationships of the Australian wholesale spot electricity markets, the multivariate generalized autoregressive conditional heteroskedasticity (MGARCH) models have been applied (Higgs, Energy Economics 31 (5):748–56, 2009) and these models were extended to the sophisticated GARCH-vine-copula model (Apergis et al., Energy Economics 90:104834, 2020) recently. Stochastic volatility (SV) models, as flexible alternatives to GARCH models, have demonstrated their superiority in many financial applications. However, the use of SV models in the modeling of electricity markets is still quite limited. This paper investigates existing multivariate SV models and proposes efficient SV models with bilateral Granger causality and skew error distributions, to model the price and price volatilities of three pairs of markets, selected from four regional electricity markets in Australia, which are shown to be highly correlated in the Higgs (Energy Economics 31 (5):748–56, 2009) study. Bayesian approach using Markov chain Monte Carlo (MCMC) method is adopted and models are implemented using the software OpenBUGS. Based on Deviance Information Criterion, the models with skew error distributions perform better than those with symmetric distributions.