How robust are SVARs at measuring monetary policy in small open economies?

We study the ability of exclusion and sign restrictions to measure monetary policy shocks in small open economies. Our Monte Carlo experiments show that sign restrictions systematically overshoot inflation responses to the said shock, so we propose to add prior information to limit the number of economically implausible responses. This modified procedure robustly recovers the transmission of the shock, whereas exclusion restrictions show large sensitivity to the assumed monetary transmission mechanism of the model and the set of foreign variables included in the VAR. An application with Mexican data supports our findings.