Forecasting returns and volatilities in GARCH processes using the bootstrap

Abstract

We propose a new bootstrap resampling scheme to obtain prediction densities of levels and volatilities of time series generated by GARCH processes. The main advantage over other bootstrap methods previously proposed for GARCH processes, is that the procedure incorpora tes the variability due to parameter estimation and, consequently, it is possible to obtain bootstrap prediction densities for the volatility process. The asymptotic properties of the procedure are derived and the finite sample properties are analysed by means of Monte CarIo experiments, showing its good behaviour versus altemative procedures. Finally, the procedure is applied to estimate prediction densities of retums and volatilities of the Madrid Stock Market index, IBEX-35.