Estimating Bank Failure Using Survival Analysis

Cheong, Sin Yong (2018) Estimating Bank Failure Using Survival Analysis. Final Year Project (Bachelor), Tunku Abdul Rahman University College.

Text Cheong Sin Yong_FULL TEXT.pdf Restricted to Registered users only Download (1MB)

Abstract

Failure of a bank could affect the entire financial market and the economy as whole. The purposes of this study is to present a Cox proportional hazards model and apply this model on prediction of bank failures, as well as identify important covariates that affects affecting these failures. The Cox model has been used widely in biomedical studies and only in the recent years researchers have started experimenting using the Cox model in the finance literature. The study identifies financial indicators from financial statements of banks that might lead to bank failure by employing the Cox model. The data set consist of banks in the United States from 2013 to 2017. This study shows that survival of banks in the US is mostly influenced by 3 predictor variables which are return on assets, return on equity and capital. It is advised that bank authorities should watch out for, to forestall financial distress. Also, another significant finding is that the average survival time for banks in the US varies significantly according to their specialty. For instance, banks categorized by area or by charter status. In this study, the procedures and results are obtained using statistical software that are readily available. At the end of this study, it is also proven that the accuracies of classification of this model in overall accuracy of classification are 87.5%.