Specification Risk and Calibration Effects of a Multifactor Credit Portfolio Model

Abstract

This article examines a crucial source of specification risk when calibrating a typical industry-type, Merton-based credit portfolio model. It emerges from the necessity of having to choose a proxy for creditworthiness. In addition to equity prices and asset values, which are the classical choices, the authors consider credit default swap (CDS) spreads and expected default frequencies (EDF, from Moody’s KMV) as alternatives. Based on 40 large European companies from different industries, the authors calibrate a macroeconomic factor model with an OLS regression analysis for each specification and calculate the corresponding economic capital. Eighteen macroeconomic and financial variables are considered as risk factors. Their findings are: a) on average, two to three risk factors are needed to adequately model creditworthiness on the obligor level, b) stock market variables are the most important risk factors, c) model-implied credit correlation is extremely sensitive to the choice of the proxy for creditworthiness, and d) only the EDF specification leads to less economic capital compared with regulatory capital, according to Basel II, while it is exceeded substantially by all other specifications. In particular, credit correlation in the CDS specification by far exceeds any estimate mentioned in the literature. Most important, the authors show that the economic capital of their sample portfolio can be reduced by 78%, depending on which variable is chosen as a proxy for creditworthiness.