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

Using new data on the term in office of central bank governors for a large set of countries for 1970-2005, we estimate a model for the probability that a central bank governor is replaced before the end of his legal term in office. We formulate hypotheses based on the literature on the determinants of central bank independence that are tested using conditional logit models and the robustness approach of Sala-i-Martin (1997). We conclude that, apart from the share of the legal term in office that has elapsed, political and regime instability, the occurrence of elections, and the ratio of private credit to GDP increase the probability of a turnover.