Temporal aggregation of multivariate GARCH processes

This paper derives results for the temporal aggregation of multivariate GARCH processes in the general vector specification. It is shown that the class of weak multivariate GARCH processes is closed under temporal aggregation. Fourth moment characteristics turn out to be crucial for the low frequency dynamics for both stock and flow variables. The framework used in this paper can easily be extended to investigate joint temporal and contemporaneous aggregation. Discussing causality in volatility, I find that there is not much room for spurious instantaneous causality in multivariate GARCH processes, but that spurious Granger causality will be more common however numerically insignificant. Forecasting volatility, it is generally advisable to aggregate forecasts of the disaggregate series rather than forecasting the aggregated series directly, and unlike for VARMA processes the advantage does not diminish for large forecast horizons. Finally, results are derived for the distribution of multivariate realized volatility if the high frequency process follows multivariate GARCH. A numerical example illustrates some of the results