Credit Risk Transfer and the Forecasting of Bank Defaults

Abstract

The current financial crisis has brought into sharp focus the need for robust empirical analysis of bank default prediction models. The contagion currently affecting the banking sector has its roots in traditional banking crises, i.e. inflated asset valuations and poor risk management. The difference, however, between past crises and that which appears to have began in earnest in August 2007 is the presence of the credit derivatives (CDs) market. The transmission of credit risk via these types of instruments appears, according to international financial regulators, to have amplified the current global financial crisis by offering a direct and unobstructed mechanism for channeling defaults between financial institutions of a variety of types. This paper proposes a default risk model, in the vain of the classic Merton-type, which incorporates forward-looking measures of market and credit risk using the credit default swaps (CDS) market as a latent variable for forecasting asset volatility. We present a robust set of empirical tests to illustrate the effectiveness of this model over alternative specifications.