Implied asset correlation in retail loan portfolios
Abstract
Credit risk arises from the interaction of multiple connected factors, but the most frequently-used models designed to measure it assume only one. These models — which, inter alia, fit distributions to loss data — are heavily influenced by the common correlation between loan values and the single factor (commonly assumed to be some gauge of economic health). Scarce and shoddy loss data for retail loan classes hamper the estimation of this correlation. A technique is proposed to calculate asset correlations embedded in empirical loss data. These values are then compared with those stipulated by the Basel II Accord for minimum capital requirements.
The full article is available to subscribers to the journal.
Author's Biography
Marius Botha was employed as a hedge fund risk manager and portfolio analyst at Old Mutual Asset Managers in South Africa before moving to London in 2005 where he worked as a risk manager at Threadneedle Investments for several years. He recently joined Threadneedle’s commodity hedge fund group as a quantitative analyst and portfolio manager. He delivers lectures on risk management in hedge funds at the School of Management, University of the Free State, South Africa, where he is a visiting professor.
Gary Van Vuuren began his career in nuclear physics before moving to market risk at ABSA bank in South Africa. He headed the Quantitative Analysis Team at Old Mutual Asset Managers before moving to London in 2002. He worked as a risk manager at Standard Bank, headed the Quantitative Group at Ernst & Young and then the Product Control Group at Merrill Lynch before moving to Fitch Ratings. Gary is also a visiting professor at the University of the Free State School of Management in South Africa.
Citation
Botha, Marius and Vuuren, Gary Van (2010, March 1). Implied asset correlation in retail loan portfolios. In the Journal of Risk Management in Financial Institutions, Volume 3, Issue 2. https://doi.org/10.69554/YCKW8811.
Publications LLP