An empirical examination of restructured electricity prices

We present an empirical analysis of restructured electricity prices. We study the distributional and temporal properties of the price process in a non-parametric framework, after which we parametrically model the price process using several common asset price specifications from the asset-pricing literature, as well as several less conventional models motivated by the peculiarities of electricity prices. The findings reveal several characteristics unique to electricity prices including several deterministic components of the price series at different frequencies. An "inverse leverage effect" is also found, where positive shocks to the price series result in larger increases in volatility than negative shocks. We find that forecasting performance is dramatically improved when we incorporate features of electricity prices not commonly modelled in other asset prices. Our findings have implications for how empiricists model electricity prices, as well as how theorists specify models of energy pricing. (c) 2005 Elsevier B.V. All rights reserved.