AbstractThis article is concerned with modeling the dynamic and distributional properties of daily spot and forward electricity prices across European wholesale markets. Prices for forward contracts are extracted from a unique database from a major energy trader in Europe. Spot and forward returns are found to be highly non normally distributed. Alternative densities provide a better fit of data. In all cases, conditional heteroscedastic models are used with success to specify the data generating process of returns. We derive implications from the relation between spot and forward prices for the evaluation of hedging effectiveness of bilateral contracts. The relation is parametrized by the mean of multivariate GARCH models possibly allowing for dynamic conditional correlation. Because correlation between spot and forward returns is very low on each market, derived optimal hedge ratios are insignificant. We conclude to a great inefficiency for forward markets at least for short-term horizon. Hedging effectiveness is not improved, for our data, through the use of dynamic correlation models.
Electricity; multivariate GARCH; dynamic correlation models; non Gaussian densities; optimal hedging; cross-hedging;