Modelling and Forecasting Stock Returns: Exploiting the Futures Market, Regime Shifts and International Spillovers¤

A large empirical literature has reported that the futures market contains valuable information for explaining stock returns and that stock returns display signi ̄cant cross-correlations internationally. A parallel literature has recorded evidence that the distribution of stock returns is close to a mixture of normal distributions and that Markov switching models may therefore provide an adequate characterization of stock returns data. This paper ties together these strands of research in that we propose a vector equilibrium correction model of stock returns that exploits the information in the futures market, while also allowing for regime-switching behavior and international spillovers across stock market indices. Using data for three major stock market indices since 1988, we ̄nd that our model signi ̄cantly outperforms a number of alternative models in sample on the basis of standard statistical criteria. In an out-of-sample forecasting exercise, the model produces some of the highest R hitherto recorded in the literature and beats all of the competing models considered on the basis of density forecast accuracy. JEL classi ̄cation: G10; G13.