The question which factors determine sovereign yield spreads is of particular importance to borrowers and investors in international bond markets. Whereas the former wish to minimize the costs of international capital and seek appropriate policy recommendations, the latter are primarily interested in risk-adequate pricing.
The empirical literature on country risk in general and that on yield spread determination in particular is large and has offered important insights. However, in some cases the findings are controversial and their practical applicability is therefore limited. In order to resolve doubt it is desirable to conduct further research in this area.
Country risk arises from a variety of factors. Among them are, for example, the incurrence of unsustainable debt levels, liquidity problems in response to external imbalances, market sentiment, and so on. In 1973, Henry J. Bitterman wrote in his book on the refunding of international debt:
- “Certain conditions appear to be characteristic of the countries obtaining debt relief. Almost all have heavy debt service in proportion to their earnings of foreign exchange. There has generally been an inflationary situation over a period of years. They have had prolonged balance of payments deficits and overvalued exchange rates. Crises have been precipitated by political events and by the cessation or diminution of the rate of capital inflow as a consequence of political events or shifts in policy, or the refusal of the creditors to continue extending additional credit.”
Most importantly, country risk is a product of many factors, which occur simultaneously and possibly enforce one another. Unfortunately, the measurement of these factors is difficult which leads to the situation that imprecise measures must be used to operationalize country risk. The theoretical contribution of this paper is to refine some existing concepts and to present new indicators in an attempt to ameliorate this situation. Indeed, it seems that four of them carry a remarkable explanatory power in the empirical analysis.
Looking at Bitterman’s description of countries obtaining debt relief, the question arises how the different determinants of country risk are connected. It is imaginable that the simultaneous occurrence of two or more of these phenomena leads to particularly high risks so that a systematic study of these interrelations seems promising. Therefore, the paper at hand tests whether interaction terms of country risk indicators carry a significant coefficient in a yield spread regression.
After the presentation of new indicators a comprehensive empirical analysis is performed which considers indicators suggested in previous research as well as those developed in this study. The majority of the estimation is based on a Bayesian model averaging algorithm in order to handle the large number of potential explanatory variables. This type of algorithm has proven helpful in similar situations, e.g. in the empirical growth literature. As the dependent variable this study uses the spread as measured by the JP Morgan EMBI+ index. The sample consists of 20 emerging market countries for which data are available over the period from 1993 until 2007.
Contents
Abstract
List of Figures
List of Tables
List of Abbreviations
List of Symbols
1 Introduction
2 Literature review
3 A model of the spread in competitive markets
4 Measures of country risk
4.1 An overview of past findings
4.2 Further indicators of country risk
- 4.2.1 The burden of debt
4.2.2 Sustainability of future debt incurrence
4.2.3 The compound liquidity indicator
4.2.4 Volatility of capital flows
4.2.5 Market sentiment
4.2.6 Global economic conditions
4.2.7 Reserves adequacy
4.2.8 Default of local currency and bank debt
4.2.9 Interaction terms
4.2.10 Summarizing the new indicators
5 Empirical evidence
5.1 The data
5.2 The model
5.3 The Bayesian framework
- 5.3.1 Motivation
5.3.2 Digression: An introduction to Bayesian methods
5.3.3 Prior, likelihood, and MC3
5.4 Results
- 5.4.1 First stage
5.4.2 Second stage
5.4.3 Third stage
5.5 Unit roots and normality
6 Conclusion
References
A Mathematical appendix
A.1 The indicator for the burden of debt
- A.1.1 Derivation
A.1.2 Partial derivatives
A.2 Sustainability of future debt incurrence
B Data and data sources
C Further results
Download
PDF Ebook Determinants of Sovereign Yield Spreads Revisited
