Forecasting of time series model with stable errors

In this paper we present a Power GARCH model with stable errors and apply the general theory of volatility forecasting to it. The use of the new model is illustrated with an application to the volatility of stock and exchange rate returns. In general, standard GARCH is outperformed by more sophisticated Power GARCH model, but use of imperfect volatility proxies leads to loss of precision in evaluating forecasts.