The observed predictability of excess returns in equity and foreign exchange markets has largely been attributed to the presence of time-varying risk premiums in these markets. For example, excess equity returns were found to be explained by various financial and economic variables. Similarly, in the foreign exchange market, the forward rate was found not to be an unbiased predictor of the future spot rate, and excess foreign exchange returns were shown to be partially explained by other variables of the foreign exchange market, notably the forward premium. However, notwithstanding the extensive empirical evidence on the above, theoretical models of international asset pricing have not been entirely successful in producing equilibrium conditions that replicate the actual behaviour of the different asset moments in empirical tests for reasonable parameter values. In fact, these models had limited success despite either rich preference structures or general driving processes for the exogenous environment of the model.

In this paper, we evaluate excess asset returns in equity and foreign exchange markets by combining generalized preferences to a heteroscedastic driving process in the same model. We do so by extending the international asset-pricing model of Bekaert, Hodrick, and Marshall (1997) in which the authors adopt disappointment-aversion-type preferences and a homoscedastic exogenous environment. We show that our very general framework, with plausible parameter values, is fairly successful in generating predictability and moment levels of excess returns that are consistent with the sample data.