Ebook Sovereign Credit Risk with Endogenous Default
This paper provides a model, and an accompanying empirical analysis, of sovereign default risk. The model helps to explain the variation across time in Emerging Market Bond Index (EMBI+) spreads to a degree not offered by prior empirical models. I also generate theoretical predictions of the relationship between credit risk and the macro'variables considered in the model that are consistent with the empirical literature.
The importance of understanding sovereign credit risk is clear in the light of the role played by sovereign debt in financial markets. Sovereign foreign debt has been at the center of a number of international lending crises. It is currently the largest asset class in emerging markets, representing approximately $5,500bn of principal in 2007 (FT, 2007).
I generate estimates of daily and quarterly credit spreads for Brazil, Mexico, Peru, and Russia over the period 1998'2006, comparing these estimates to observed EMBI+ spreads. I rely on government revenue information to predict credit spreads at the quarterly frequency, while daily credit spreads are predicted from information on stock market performance made available daily. Government revenues and stock market indices present severe non'linear relationships with EMBI+ spreads.
The main contribution of the paper resides in developing a model that breaks down the non'linearity in the data to predict estimates of credit spreads that are linearly related to EMBI+spreads. In a panel analysis with fixed effects, the credit spreads predicted by the model explain about 92% of the variation across time in daily EMBI+ spreads. The explanatory power rises only slightly, to 93%, when accounting for the VIX option'implied volatility index as an additional time'varying factor. This finding may change interpretations of the results of Longstaff, Pan, Pedersen, and Singleton (2007) and Pan and Singleton (2008). These authors show that the VIX index is a key factor in explaining credit risk movements. However, they do not include stock market information which seems to nearly eliminate the explanatory power of the VIX in the analysis.
This paper offers the first structural model explaining the dynamics of daily EMBI+ spreads. Prior studies interested in daily spreads have considered either a reduced'form ap ne structure model1 or a reduced' form contingent'claims analysis. Studies analyzing the dynamics of quarterly or annual credit spreads have used a dynamic stochastic equilibrium model and a panel'based approach. An advantage of the proposed model is that it provides an intuitive theoretical framework for the determinants of credit spreads. The model can then be used to motivate empirical specifications, one of the aims of this paper.
The theoretical predictions of the relationship between credit risk and the macro'variables analyzed in the model are consistent with the empirical literature. First, within the model, credit risk decreases with economic performance because the sovereign country is more likely to default in a recession. This is consistent with previous empirical works. Second, higher macroeconomic volatility leads to greater credit risk, also consistent with prior empirical findings. Third, the model suggests that credit risk first increases and then decreases with the risk'free interest rate.
A positive is also in the data. Fourth, the model implies that sovereign credit risk is negatively affected by the severity of economic losses upon default, as has been confirmed empirically. A default crisis reduces the sovereign countryns access to the international trade market, thereby reducing future sovereign revenues. If the sovereign country relies heavily on trade, then it is less inclined to default because the impact of such economic costs is large. Fifth, as domestic investment generates high returns relative to the risk'free rate, the sovereign wealth is predicted to increase. Credit risk then decreases as has been documented empirically.
Download
PDF Ebook Sovereign Credit Risk with Endogenous Default
Posted in :
