Heavy tails and electricity prices: Do time series models with non-Gaussian noise forecast better than their Gaussian counterparts?

This paper is a continuation of our earlier studies on short-term price forecasting of California electricity prices with time series models. Here we focus on whether models with heavy-tailed innovations perform better in terms of forecasting accuracy than their Gaussian counterparts. Consequently, we limit the range of analyzed models to autoregressive time series approaches that have been found to perform well for pre-crash California power market data. We expand them by allowing for heavy-tailed innovations in the form of α-stable or generalized hyperbolic noise.