"Serving the energy market" - http://www.erasmusenergy.com
Understanding Electricity Price Volatility Within and Across Markets
http://www.erasmusenergy.com/articles/144/1/Understanding-Electricity-Price-Volatility-Within-and-Across-Markets/Page1.html
Mika Goto
Mika Goto, Research Economist,
Central Research Institute of Electric Power Industry
 
By Mika Goto
Published on 01/7/2008
 
Keywords: Electricity prices, volatility, market structure, international markets
Published in:
Publication year: 2003
Co-author 1: G. Andrew Karolyi

This study analyzes how electricity price volatility evolves over time for different electricity trading hubs in several deregulated markets around the world. The goal is to uncover common features across hubs within each market in the daily on-peak price volatility processes related to seasonality, mean reversion, conditionally autoregressive heteroskedasticity (ARCH) and possibly time-dependent jumps. We apply our analysis to markets in Australia, U.S., and the Nordic Power Exchange. We show that ARCH and time-dependent jumps are important statistical features of price volatility across all hubs in each market but with different levels of intensity. We also find that inferences about the role of seasonality components is sensitive to modeling of the ARCH and jump features.

Understanding Electricity Price Volatility Within and Across Markets

This study analyzes how electricity price volatility evolves over time for different electricity trading hubs in several deregulated markets around the world. The goal is to uncover common features across hubs within each market in the daily on-peak price volatility processes related to seasonality, mean reversion, conditionally autoregressive heteroskedasticity (ARCH) and possibly time-dependent jumps. We apply our analysis to markets in Australia, U.S., and the Nordic Power Exchange. We show that ARCH and time-dependent jumps are important statistical features of price volatility across all hubs in each market but with different levels of intensity. We also find that inferences about the role of seasonality components is sensitive to modeling of the ARCH and jump features.