Econometric Models of Credit Spreads

Abstract

In its document “Basel III: A global regulatory framework for more resilient banks and banking systems”, the Basel Committee set a CVA methodology for the trading book, at a time the determination of the credit spread for those entities interested in advanced models in their risk management is vital. In a later document, the Basel Committee reaffirmed the idea of requiring the financial entities to estimate the credit spread curves considering the different factors of rating, sector and region of each counterparty. In the aftermath of the financial and economic crisis, is the OLS regression the optimal method to estimate credit spread curves? Least squares estimates can be influenced by the presence of outliers, so that more robust estimators may be preferred. Additionally, from the point of view of a financial institution, it would be desirable to reduce the fluctuations in risk estimates, which may need of nonlinear specifications. Using single name CDS data for the 2006-2012 period, we compare the performance of hierarchical models and exponential models with that of the linear least squares approach to provide credit risk estimates.

Keywords: Credit Spread Curve, Asset Pricing, Hierarchical Regression, Quantile Regression

JEL Classification: C58, C52, G012