This paper examines the ability of vector autoregressive (VAR) models to properly identify the transmission of monetary policy in a controlled experiment. Simulating data from an estimated small open economy DSGE model for Australia, we find that sign-restricted VAR models do reasonably well at...