Traditional ways of analyzing the effects of monetary policy shocks via structural vector autoregressions require the use of unrealistic identifying assumptions: they either do not allow for a response of output and prices on impact of the shock, or they exclude contemporaneous values of these variables from the monetary authority's information set. This paper relaxes these incredible restrictions by exploiting a convenient natural setting, namely the fact that we can use data from dollarized countries. The fact that non-monetary US shocks do not seem to be transmitted to these countries, has the additional advantage that it makes the exercise less vulnerable to potential misidentification of the US monetary policy shock. The results obtained in this way suggest that prices fall quite rapidly after a monetary contraction. Consistent with this finding, the effects of monetary policy shocks on output seem to be small.
Monetary policy effects Price puzzle Structural VARs Identification Block exogeneity