International Organizations and Structural Reforms
NBER Working Paper No. 21237
Different countries have been following different reform paths since the early 1990s. We develop a simple dynamic model of policy reform that captures some of the determinants that underlie these differences. The model emphasizes the interaction between domestic institutions and international organizations that promote reform, on the one hand, and the political incentives for reversing reforms, on the other. At equilibrium, there are three types of reform paths. A country can undergo a full-scale, lasting reform; it can undertake a partial but lasting reform; or it can go through cycles of reforms and costly counter-reforms. Domestic institutions, as well as the incentives provided by international organizations, determine the equilibrium path. Unless the cost of reversal is high enough, an international intervention that promotes reforms induces an increase in the probability of reversals. A benevolent international organization that is fully aware of the possibility and social cost of reversals will always increase social welfare if it embraces the following principle: promote the greatest partial reform that is compatible with no reversal, or induce cycles of full-scale reform and complete reversal, depending on which of those two paths will generate greater social welfare. A benevolent but politically myopic international organization, however, may reduce social welfare because it does not take the fact into account that an overly aggressive reform could trigger costly reversals that outweigh the benefits of the reform. Deliberately making the costs of reversal high could be a risky way of improving the trade-off between the extent of the reform and the probability of reversal. Our model suggests that international organizations should also consider the possibility of providing defensive funding for dealing with counter-reform shocks.
Document Object Identifier (DOI): 10.3386/w21237
Users who downloaded this paper also downloaded* these: