Heterogeneity of agents and exchange rate dynamics: Evidence from the EMS

We develop and estimate a dynamic heterogeneous agent model for the EMS period. Our empirical results suggest that the existence of heterogeneous interacting agents is indeed a possible explanation for the dynamics of exchange rates during the EMS. We find strong evidence of heterogeneous boundedly rational beliefs, and the fact that agents switch between these beliefs. Moreover, we show that the dynamic heterogeneous agent model outperforms the random walk and the static heterogeneous agents' model in out-of-sample forecasting in the large majority of country-horizon combinations.