Retail Loans & Basel II: Using Portfolio Segmentation to Reduce Capital Requirements

ECRI Research Report No. 8

36 Pages Posted: 27 Feb 2007

See all articles by Daniel Kaltofen

Daniel Kaltofen

Ruhr University of Bochum - Faculty of Economics; BiTS Business and Information Technology School

Stephan Paul

Ruhr University of Bochum - Faculty of Economics

Stefan Stein

University of Bochum - Department of Finance and Banking

Date Written: August 2006

Abstract

This paper presents a new technique for grouping retail loans into homogenous risk pools, which adheres to the provisions of Basel II. We use recursive partitioning and test it on a data set of approximately 413,000 auto loans. By classifying loans according to selective predictors of default, we find that banks can achieve significant savings in terms of a lower regulatory capital requirement. Alternatively, this provides the opportunity to increase lending capacity.

Keywords: Basel II, retail portfolio, credit risk, classification algorithms, portfolio segmentation

JEL Classification: C14, C25, C53, G21, G28

Suggested Citation

Kaltofen, Daniel and Paul, Stephan and Stein, Stefan, Retail Loans & Basel II: Using Portfolio Segmentation to Reduce Capital Requirements (August 2006). ECRI Research Report No. 8, Available at SSRN: https://ssrn.com/abstract=958204 or http://dx.doi.org/10.2139/ssrn.958204

Daniel Kaltofen

Ruhr University of Bochum - Faculty of Economics ( email )

Ruhr University of Bochum
Faculty of Economics
Bochum, DE 44780
Germany

BiTS Business and Information Technology School ( email )

Reiterweg 26b
Iserlohn, D-58636
Germany

Stephan Paul

Ruhr University of Bochum - Faculty of Economics ( email )

Ruhr University of Bochum
Faculty of Economics
Bochum, DE 44780
Germany

Stefan Stein (Contact Author)

University of Bochum - Department of Finance and Banking ( email )

Bochum, DE 44780
Germany

Do you have a job opening that you would like to promote on SSRN?

Paper statistics

Downloads
913
Abstract Views
4,311
rank
32,076
PlumX Metrics