Country Risk Ratings and Financial Crises 1995--2001: A Survival Analysis

Article excerpt

Executive Summary

The health of the financial system is a sign of economic growth and a key indicator for investors. As a consequence, one of the main purposes for policymakers is to guarantee its stability and to shield it from foreign disturbances. Both financial and economic activities are susceptible to crises. As soon as a financial crisis happens, a country may face a default risk, which can be measured in the long term through the country's debt risk rating. Even though, with the recent crisis, the ability of ratings to predict a weak debtor has been questioned, in this paper we propose that the survival analysis methodology should be used to analyze falling rating duration. It has usually has been used in labor economics and with few exceptions in financial economics. Additionally we test the capability of macroeconomic variables to predict that event in 78 countries between 1995 and 2001. From the analysis, important differences between developed and emerging economies are indicated in exchange rate risk and economic indebtedness.


As globalization grows, financial system stability is a key signal to investors and a bad behavior of the system will not contribute to economic growth. An indicator of the domestic capital market's health is the credit rating given to the Long Term Debt, which gives information of short term macroeconomic stability and long term payment capability.

This paper presents an effort to analyze financial crises through sovereign risk ratings. Our analysis presents two important aspects of downgrading which have not been considered before; they are the timing of the crisis and the impact of neighboring countries in crisis. In order to approach these points, we propose the use of a survival model to compute the risk function of a downgrading, controlling by macroeconomics and exchange monthly variables which reflect the economy's health at a short and mid term. We exclude from the analysis real sector variables, as they may be endogenous to the country's risk rating.

This methodology also allows forecasting crisis length and contagion by geographic and economic regions. We use a semi parametric methodology, which is better suited to the analysis of non-monotonic risk functions, given the persistence and the contagion effects.

Related Literature

As financial crisis are not new, we present literature that follows these events chronologically, as well as key indicators. The First Generation Models explain crises in the early 1980s, which were characterized by the macroeconomic imbalances. Krugman (1979, p. 318) uses a simple model to show how attempts to defend a fixed exchange rate can collapse in the face of a speculative attack. He shows how the persistent balance of payments' deficits (1) can create a run on the authorities' stock of international reserves and destroy the country's capacity to defend its exchange rate by limiting its ability to intervene in the foreign-exchange market. The central purpose of the models of this generation is to demonstrate how an attack and collapse of the exchange rate can occur before reserves have gotten exhausted and can also speed up the timing of the crisis. The end result is that the government uses up its reserves and cannot replenish them by borrowing abroad. The leading indicators of this generation are budget deficits, excessive rates of growth of money supply, dwindling reserves, excessive inflation, real exchange rate overvaluation and high interest rates.

With the crisis of early 1990s, the theory above was questioned, since not all the countries that succumbed displayed large fiscal and current account deficits. It is even more questionable considering that capital controls were lifted and international markets got wider and perhaps deeper. Obstfeld (1997, p. 68) and Ozkan and Sutherland (1998, p. 345) added the assumption that governments tend to balance the costs and benefits of defending the currency, through tight monetary policies and high interest rates. …