On the conditional effects of IMF program participation on output growth

The empirical evidence currently available in the literature regarding the effects of a country's IMF program participation on its output growth is rather inconclusive. In this paper we propose and estimate a panel data sample selection model featuring state dependence. As in this model the output growth effects of program participation can be conditional on the realization of a state variable (conditional pooling), our framework may reconcile previous empirical evidence based on models without state dependent effects. We find that the effects of IMF program participation on output growth vary systematically with an index reflecting a country's institutional record, and that output growth effects of program participation are significantly positive only if the IMF program is coupled with sufficient improvement of the institutional record.