Sovereign rescheduling probabilities in emerging markets: a comparison with credit rating agencies’ ratings

Abstract

This study estimates default probabilities of 124 emerging countries from 1981 to 2002 as a function of a set of macroeconomic and political variables. The estimated probabilities are then compared with the default rates implied by sovereign credit ratings of three major international credit rating agencies (CRAs) – Moody's Investor's Service, Standard & Poor's and Fitch Ratings. Sovereign debt default probabilities are used by investors in pricing sovereign bonds and loans as well as in determining country risk exposure. The study finds that CRAs usually underestimate the risk of sovereign debt as the sovereign credit ratings from rating agencies are usually too optimistic.

Keywords:

• sovereign debt
• default probabilities
• credit rating agencies
• credit ratings

JEL Classification :

• F33
• F34

Notes

1. For the purpose of this study, we have selected two out of four forecasting models developed in our previous study based on their superior level of forecasting ability. In addition, we have reduced the data sample–from 127 to 124 emerging economies – to match the country coverage with credit rating agencies for comparison purposes.

2. The higher the imports in relation to the size of the economy, the more open the country is and, thus, more vulnerable to foreign shocks, and more likely to external debt rescheduling (e.g. 11).

3. Some studies define a country as in default if there is a debt rescheduling agreement or negotiations. Other studies consider sovereign default if there are arrears on principal or interest payments, or a country concludes an upper-tranche IMF agreement.

4. Examples of political variables used in some studies are: democracy index, political instability index, long- and short-term armed conflict, changes of the finance minister and/or the minister of the economy.

5. Private or official debt in relation to the capacity of repayment.

6. External debt service/reserves or External debt service/exports.

7. E.g. Money/Gross International reserves, Exchange rate devaluation, as per IMF research.

8. Real growth, exchange rate, inflation, etc.

9. Variables that explain the country's ‘willingness to pay’, e.g. government stability, socioeconomic conditions, external conflict, internal conflict, corruption, military in politics, religion in politics, democratic accountability, etc.

10. The cost of obtaining and processing this information is marginal providing the analytical set-up in a financial institution is already in place. For lower net worth individual investors, there may be a cost advantage of utilising the existing credit rating agencies’ default probabilities.

11. For details about Principal Component Analysis, interested readers may refer to Solberg 36.

12. The PCA analysis is presented and discussed in detail in Georgievska et al. 12.

13. In addition, the calculation of marginal effects at mean in Stata software, which we utilize, is more convenient and output is numerically equally stable as the average marginal effect at every observation for a large data sample.

14. Similarly, Balkan 2 finds the political factors and proxies of political factors very significant variables in explaining a country's rescheduling/default and the country's risk exposure faced by international lenders.