ESSAYS ON GRADUATION

Abstract

This dissertation attempts to address the elusive concept of "graduation", that is the emergence from frequent crisis suffering status. It contains two chapters. The first uses a data set covering over two hundred years of sovereign debt, banking and inflation crises to explore the question of how long does it take a country to "graduate" from the typical pattern of serial crises that most emerging markets experience. We find that for default and inflation crises, twenty years is a significant period, but the distribution of recidivism has extremely fat tails. In the case of banking crises, it is unclear whether countries ever graduate. We also examine the more recent phenomenon of IMF programs, which sometimes result in "near misses" but sometimes end in default even after a program is instituted.

The second chapter investigates the impact of countries' institutions on their likelihood of sovereign default from both an empirical and theoretical perspective. By employing a dataset of more than 80 countries, two facts emerge: 1) high institutional quality is associated with a low frequency of sovereign default crisis, and 2) in particular, polarized governments tend to default more often. To explain these facts, we developed a model that establishes a link between institutions, government polarization and sovereign default crises. Countries that lack rules and institutional settings to limit the pressure of powerful groups on a central government's policies default more often than countries that do have good institutions. Given that there are no barriers to limit the influence of powerful groups, a more polarized government defaults more because groups do not coordinate, giving rise to a negative externality. Simulations of the model succeed in matching the cross-country differences in sovereign default frequencies, given their institutional quality and degree of government polarization in the data.