Default rate modelling of a bank in Ghana using Cox Regression Model.

Loading...
Thumbnail Image
Date
May, 2016
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
The subject of non-performing loans which in e ect a ect all banks including small market enterprises' ability to prevent ultimate ruin. \For a credit bureau to function e ectively, however, it must be possible to uniquely identify individuals with reasonable certainty. Identi cation is necessary in order to retrieve a current loan applicant's past credit history from a credit database" (Gine et al., 2011). There arises a fundamental need to device an easy strategy to reduce the oc- currences of non-performing which would be widely accepted. Cox proportional- hazard model with time-varying covariates was used to evaluate certain nancial sensors that would de nitively predict the bank's loan failures during the years of 2012 to 2015 of Zenith Bank Ahodwo branch. The Breslow approximation to the partial likelihood of the overall cox relations of client is used to estimate the correlation coe cient parameter, betas through di erentiation. The models resultant outputs were compared to the actual defaults and it is worthy to note that the model presents a well above average predictor quality. The signi cance of each covariate is tested statistically and the 4th covariate (the number of de- pendents) was found to be signi cant there were insu cient data reject the null hypothesis for all the other covariates. A new way of coarse-classifying of covari- ates enlisting the help of survival analysis methods is proposed. Also, a good number of diagnostic methods (Cox-Snell, Martingale amongst other residuals) are used to check the aptitude of the model's tness for applicability to any loan data. These results can be extended to a general case of bank failures under the volatile economic environments.
Description
A thesis submitted to the Department of Mathematics, Kwame Nkrumah University of Science and Technology in partial fulfillment of the requirement for the Degree of Master of Science in Actuarial Science,
Keywords
Citation