<![CDATA[The modeling of heteroskedasticities and kurtosises of electricity prices is crucial to forecast the future distribution of electricity prices, to understand the behavior of derivatives pricing and to quantify the risk in electricity markets. A GARCH model with t-innovations, which is solved by maximum likelihood estimation, is proposed. The model can explicitly address the relationship with system loads, seasonalities, heteroskedasticities, and kurtosises of electricity prices. The empirical analysis based on the historical data of the PJM electricity market shows that the system load squares have a significant effect on the average daily electricity prices, there exist volatility clustering and weekly, semi-monthly, monthly, bimonthly, quarterly and semi-annual periods, and the variances and kurtosises of electricity prices manifest clearly time-varying characteristics. The model holds parsimonious scale of estimated parameters, less computational costs, easy to select the orders and high practical application value. DOI: http://dx.doi.org/10.11591/telkomnika.v11i10.3393]]>
Copyrights © 2013
