Interval forecasting of spot electricity prices

Abstract

In this paper we assess the short-term forecasting power of different time series models in the electricity spot market. However, instead of evaluating point predictions we concentrate on interval forecasts. The latter are specifically important for risk management purposes where one is more interested in predicting intervals for future price movements than simply point estimates. We find evidence that non-linear regimeswitching models outperform their linear counterparts and that an additional GARCH component significantly improves interval forecasts of linear time series models. 1.