A k-factor GIGARCH process: estimation and application on electricity market spot prices

Some crucial time series of market data, such as electricity spot prices, exhibit long-memory, in the sense of slowly-decaying correlations, combined with heteroskedasticity. To be able to modelize such a behavior, we consider the k-factor GIGARCH process. The related parameter estimation problem is addressed using an extension of Whittle's estimation. We develop the corresponding asymptotic theory for estimation. We apply this approach to the electricity prices (spot prices) from the German energy market (European Energy eXchange). For these data, we propose two models of k-factor GIGARCH processes. To conclude the paper, we analyze in detail the forecasting performances of these models.