Title: Evaluating electricity price volatility risk in competitive environment based on ARMAX-GARCHSK-EVT model

Authors: Jin Wang; Ruiqing Wang

Addresses: College of Physics & Electrical Engineering, Anyang Normal University, No. 185, Huanghe Road, Anyang 455000, Henan Province, China ' School of Computer & Information Engineering, Anyang Normal University, No. 185, Huanghe Road, Anyang 455000, Henan Province, China

Abstract: Effectively evaluating volatility of price risk is the foundation of risk management in competitive environments. Considering the natures of electricity prices, a two-stage model for estimating value-at-risk (VaR) based on ARMAX-GARCHSK and extreme value theory (EVT) is proposed. First, an ARMAX-GARCHSK model, which can capture the most important characteristics of electricity price series, such as seasonalities, heteroscedasticities, skewnesses and lepkurtosises, is used to filter electricity price series. In this way, an approximately independently and identically distributed normalised residual series is acquired. Then EVT is adopted to explicitly model the tails of the standardised residuals of ARMAX-GARCHSK model, and accurate estimates of VaR in electricity market can be yielded. The empirical analysis suggests that the ARMAX-GARCHSK-EVT model can rapidly reflect the recent and relevant changes of electricity prices and produce more accurate forecasts of VaR at all confidence levels, showing better dynamic characteristics. These results present several potential implications for electricity market risk quantifications and hedging strategies.

Keywords: electricity markets; VaR; value-at-risk; ARMAX-GARCHSK model; EVT; extreme value theory; GPD; generalised Pareto distribution; Gram-Charlier series expansion; electricity price volatility; volatility risk; electricity prices; risk management; modelling; risk quantification; hedging strategies.

DOI: 10.1504/IJCAT.2014.066722

International Journal of Computer Applications in Technology, 2014 Vol.50 No.3/4, pp.174 - 179

Published online: 07 Feb 2015 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article