Failure Prediction Models in the Nigerian Banking Industry - An Empirical Study (1986 - 1998)
Nwezeaku, Nathaniel C.
MetadataShow full item record
The purpose of this research is to investigate the possibility of developing a failure prediction model in the Nigeria economic environment with particular reference to the financial services industry and to compare the performance of such a model with the results of similar models developed in advanced economies. The choice of the banking industry is because of the recent distress that devastated it. Our sample consists of fifty banks divided into three groups. The experimental group of twenty (20) failed banks, the control group of twenty (20) non-failed banks and a hold-out sample of ten (10) banks. The technique of the Beaver Model was utilized for the analysis. It involved the Paired design, Paired analysis Profile Analysis and the Dichotomous Classification tests I and II. Twenty three financial ratios were developed and a total of three thousand four hundred and four (3,404) individual ratio were computed for the analysis. After the application of the last technique namely; the Dichotomous classification predictive tests, ten ratios emerged as the best predictors of distress in the Nigeria banking industry. The overall best predictor is total debt to total assets, which has a potential ability to predict distress four years ahead of the actual event with about ninety percent (90%) accuracy. Our final model was constructed with five of the ten best ratios as follows: