Title: Risk insolvency predictive model maximum expected utility

Authors: Daria Marassi, Valetino Pediroda

Addresses: Eu-Ra Europe Rating S.p.A., L.go Don Francesco Bonifacio 1, Trieste 34125, Italy. ' Dipartimento di Ingegneria Meccanica, Universita degli Studi di Trieste, via Valerio 10, Trieste 34127, Italy

Abstract: This paper presents a new approach to develop the probability of default for private firms. This work provides a global perspective regarding the credit risk prediction, starting from the work of the Basel Committee on Banking Supervision, with a deep study of the more predictive variables for default prediction and, finally, building a new mathematical model based on machine learning. The used method is called Maximum Expected Utility (MEU) and represents the most promising methodology for the default prediction. The main idea is to use the interaction between variables to improve the final model efficiency. The development model has been tested on complex analytical function (in which the classical models fault) and finally has been developed to assess the distress of industrial companies according to Basel II guidelines. The evidence was related to Italian industrial enterprises and took into consideration, the situation of the Italian economy both from a micro and macro perspective.

Keywords: rating; default probability; default risk; credit risk prediction; Basel II; default prediction; risk insolvency model; default models; predictive models; variables interaction; predicted default variables; self organizing maps; statistical analysis; credit rating; risk model development; cumulative accuracy profile; machine learning; Italy.

DOI: 10.1504/IJBPM.2008.016637

International Journal of Business Performance Management, 2008 Vol.10 No.2/3, pp.174 - 190

Published online: 11 Jan 2008 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article