Title: Optimising a bank's credit portfolio

Authors: Annshirley Aba Afful; Michael Kofi Asare; Raymond Benjamin Afful

Addresses: Ghana Institute of Management and Public Administration (GIMPA) Business School, P.O. Box AH 50, Achimota, Accra, Ghana ' Department of Mathematics and Statistics, Graduate of Kwame Nkrumah University of Science and Technology, Private Mail Bag, University Post Office, Kumasi, Ghana ' Takoradi Polytechnic, Mathematics and Statistics Department, P.O. Box 256, Takoradi, Ghana

Abstract: The purpose of this paper is to show the practical application of linear programming and logistic regression models in the formulation of an optimal bank credit policy. Firstly, we formulate a linear programming model and develop a solution (using the simplex algorithm) that optimally allocates funds, where a financial institution is facing the problem of allocation of limited funds among different types of loans/advances at different markup/interest rates with varying degree of risk (bad debts). We go further, after optimal allocation of funds, to propose a binary logistic regression model (BLRM) to discriminate loan defaulters from non-defaulters. The study revealed that the available funds of GH¢166 million for credit facilities will yield a return of GH¢35.25 million after allocation. Four important influences were identified and the LR proposed predicts that about 80% of prospective customers are likely not to default.

Keywords: optimisation; bank credit portfolio; loans; optimal fund allocation; linear programming; logistic regression modelling; credit policy; limited funds; credit risk; loan defaults; default risk; financial institutions; banking industry.

DOI: 10.1504/IJAMS.2016.075466

International Journal of Applied Management Science, 2016 Vol.8 No.1, pp.68 - 82

Received: 08 Oct 2014
Accepted: 01 Aug 2015

Published online: 23 Mar 2016 *

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