An Empirical Analysis on Credit Risk Models and its Application

Abstract

This study intends to focus on introducing credit default risk with widely used credit risk models in an effort to empirically test whether the models hold their validity, apply to financial institutions which usually are highly levered with various types of debts, and finally reinterpret the results in computing adequate collateral level in the over-the-counter derivatives market. By calculating the distance-to-default values using historical market data for South Korean banks and brokerage firms as suggested in Merton model and KMV’s EDF model, we find that the performance of the introduced models well reflect the credit quality of the sampled financial institutions. Moreover, we suggest that in addition to the given credit ratings of different financial institutions, their distance-to-default values can be utilized in determining the sufficient level of credit support. Our suggested “smoothened” collateral level allows both contractual parties to minimize their costs caused from provision of collateral without undertaking additional credit risk and achieve efficient collateral management.