Description:
In this paper we use data from Mexico to identify Dornbusch's (1976) exchange rate overshooting hypothesis. We specify and estimate a structural cointegrated VAR that considers explicitly the presence of a set of long-run theoretical relations on macroeconomic variables (a purchasing power parity, an uncovered interest parity, a money demand, and a relation between domestic and U.S. output levels). We then impose a recursiveness assumption to identify the response of domestic variables to a monetary policy shock. The long-run restrictions embedded in the model are themselves identified, estimated, and tested using an ARDL methodology that is robust to the degree of persistence of the time series and, in particular, to whether they are trend- or first-difference stationary. With this approach, we are able to find that the response of the exchange rate to monetary policy shocks is consistent with Dornbusch's model.