Bankruptcy Prediction Model Conventional Bank Rural Banks in Indonesia

Abstract

This study aims to form a prediction model for the bankruptcy of rural banks in Indonesia. The analytical method used logistic regression by first analyzing the factors followed by testing the validation of the model based on new data. The study population is the Rural Credit Bank in Indonesia. The sample used was 229 banks consisting of 29 bankrupt banks and 200 non-bankrupt banks. The data used is the quarterly financial statement data of the bank from 2006 to 2016 as a design sample and 2017 data as a validation sample. A total of 120 were used as design samples to form a bankruptcy prediction model and 109 banks as validation samples to test the accuracy of the model formed. The results showed that of the four prediction models that were successfully built, it turned out that only MP3s were eligible to be used as a prediction model for Bank Credit in Indonesia. At the level of MP3, modeling has a classification accuracy of 100% with a cut-off point of 0.2, and at the level of MP3, validation has a classification accuracy of 88.99% with a cut-off point of 0.09. The interest rate of the LPS guarantee is the dominant financial risk factor that predicts the probability of bankruptcy of rural banks in Indonesia. Keywords: Financial Risk, Prediction, Bankruptcy, Logistic Regression.