Potentially endogenous borrowing and developing country sovereign credit ratings

Abstract

This paper studies the determinants of the foreign currency credit ratings assigned to about 30 developing country national governments by the two leading rating agencies using two different econometric techniques. First, an equation relating the average of the two agencies' ratings assigned to a particular national government to a number of potential determinants, including measures of the country's external debt burden, is estimated by ordinary least squares (OLS) under the assumption that all of the explanatory variables are exogenous. It is shown that, under this assumption, a very large share of the cross-country dispersion of the ratings studied is explained by a small number of variables. The highly statistically significant variables include a corruption perceptions index, the time elapsed since the resolution of the government's most recent default on its foreign currency debt, debt service on public and publicly guaranteed external debt relative to exports and a measure of the maturity structure of international banking claims against both private and public sector entities in the country. Collectively, these four variables explain 87% of the dispersion in average foreign currency credit ratings. The highly statistically significant variables remain statistically significant under an alternative estimation strategy (instrumental variables estimation) that permits one measure of the country's external debt burden - the ratio of total external debt to gross national income - to respond to the values of these ratings as would be the case if the demand for external financing by developing country entities responded systematically to borrowing costs. No evidence for such feedback is found, suggesting that it may be appropriate to estimate the parameters of rating equations that include measures of the external debt burden by OLS.