Conditional loss estimation using a South African global error correcting macroeconometric model

Loading...
Thumbnail Image

Authors

De Wet, Albertus Hendrik

Journal Title

Journal ISSN

Volume Title

Publisher

University of Pretoria, Department of Economics

Abstract

Active credit portfolio management is becoming a central part of capital and credit management within the banking industry. Stimulated by the Basel II capital accord the estimation of risk sensitive credit and capital management is central to success in an increasingly competitive environment. If any risk mitigation or value-enhancing activity is to be pursued, a credit portfolio manager must be able to identify the interdependencies between exposures in a portfolio, but more importantly, be able to relate credit risk to tangible portfolio effects on which specific actionable items can be taken. This analysis draws on the macroeconometric vector error correcting model (VECM) developed by De Wet et al. (2007) and applies the proposed methodology of Pesaran, Schuermann, Treutler and Weiner (2006) to a fictitious portfolio of corporate bank loans within the South African economy. It illustrates that it is not only possible to link macroeconomic factors to a South African specific credit portfolio, but that scenario and sensitivity analysis can also be performed within the credit portfolio model. These results can be used in credit portfolio management or standalone credit risk analysis, allowing practical credit portfolio management and value enhancing applications.

Description

Keywords

Credit portfolio modelling, Macroeconometric correlation model, Economic capital, Scenario analysis, Default threshold

Sustainable Development Goals

Citation

De Wet, AH, Van Eyden, R & Gupta, R 2008, 'Conditional loss estimation using a South African global error correcting macroeconometric model', University of Pretoria, Department of Economics, Working paper series, no. 2008-26. [http://web.up.ac.za/default.asp?ipkCategoryID=736&sub=1&parentid=677&subid=729&ipklookid=3]