Quarterly evidence on participation in IMF programs for the period 1974-2003 is examined statistically using the techniques of hazard analysis and error-correction estimation. Three hypotheses are proposed and tested, and the results support a quot;revolving doorquot; explanation of...