Accuracy and convergence of the Asymptotic Single Risk Factor formula in a large credit portfolio

Date
2016
Journal Title
Journal ISSN
Volume Title
Publisher
University of Delaware
Abstract
According to international banking standards, all financial institutions must classify the risks associated with the credit portfolios they hold. Mathematical approximations of default distributions are among the most common ways of assessing such risk. Understanding the errors of these approximations is crucial for generating reliable credit portfolio default risk calculations. We consider the error associated with Vasicek’s Asymptotic Single Risk Factor model for the cumulative distribution of losses in a portfolio of N companies. We analytically and numerically verify the scaling of the error to be O(N–1), scaling as the reciprocal of the number of companies. Our results provide insight into the error associated with one of the most commonly used credit risk models and serve as a model for future work in examining the errors of more complex, hierarchical structural models.
Description
Keywords
Citation