Predicting output growth at long horizons: stock return volatility and the monetary policy influence

Abstract

In this paper, we investigate the presence of multiple horizon causation from stock return volatility to the growth rates of industrial production in terms of a causal chain system, where monetary policy instruments and inflation operate as auxiliary processes. Multiple horizon non-causality is tested by implementing the test procedure proposed by Dufour et al. (2006) on data from four economies, namely USA, Germany, Japan and Italy. Our results reveal a large number of highly significant direct and indirect causality links of significant size running from stock return volatility to output growth at both short- and long-horizons in all four economies. A pseudo out-of-sample forecasting evaluation also shows how conditioning on such information yields better output growth predictions at different forecast periods.