This paper documents the effects of changes in US monetary policy on asset prices in 51 countries to evaluate the validity of the event-study approach. We find that the event-study estimates contain a significant bias. However, this bias is fairly small and the ordinary least squares approach tends to outperform in an expected squared error sense the heteroscedasticity-based estimator for both small and large sample sizes. Hence in general the event-study methodology should be preferred. Moreover, we show that US monetary policy has been an important determinant of global financial markets.