Bank risk and deposit insurance

Abstract

The increased number of bank failures in recent years has emphasized the need to assess bank risk. The prediction of these bank failures through the development of an effective early warning system can allow for sufficient allocation of resources by supervisors and regulators hence it is important to determine the factors that affect early warning systems in the identification of weak banks. The aim of the study is to determine how a reliable, efficient, and sustainable deposit insurance scheme should carry out surveillance in order to efficiently predict the probability of bank failure and therefore maintain a stable banking industry. I assessed the effectiveness of the CAMELS framework use in Kenya by testing the statistical adequacy of the variables used and comparing trends across the period under review to check whether the framework does signal weakening banks beforehand. The study used PCA to determine which principal component factors are able to account for the variance in the data. These factors were further used to build the PCA-CAMELS factor linear discriminant model and logit model. The two models were then used to predict the failure or non- failure of banks given the selected PCA factors over the years 2010 to 2015