When is a central bank governor fired? Evidence based on a new data set

This paper uses a new data set on the term in office of central bank governors in 137 countries covering the period 1970-2004 to estimate a model for the chance that a central bank governor is replaced. We formulate a number of hypotheses based on the literature on the determinants of central bank independence that are tested using conditional logit models and the Extreme Bounds Analysis. We conclude that, apart from the share of the current term in office elapsed, high levels of political and regime instability, the occurrence of elections, and high inflation increase the probability of a turnover.