Predicting Bank Failure in Kenya

Abstract

Predicting default or probability of failure is a concept that has brought out a lot of interest since 1960’s, the banking sector has not been left out as models have been developed to predict bank failure. In this study, we estimate the probability of bank failure in a Kenyan context based on data between 2013 & 2017, inclusive. We build a model using binary panel data regression to predict probability of bank failing twelve months in advance for Kenyan banks based on eleven micro factors. Empirical results show that early warning signal developed produces a robust result. In addition, we have established shareholders’ funds, non-performing loans, insider loans ratio, return on equity, prudential liquidity ratio, average deposit market share, largest shareholder shareholding, years bank in existence, and weighted average cost of funds are the factors that contribute to bank failure in Kenya.