Estimation of Bankruptcy Probabilities Using Bayesian Approach

Yongdai Kim and Gunhee Lee
Department of Statistics, Hankuk University of Foreign Studies, Yongin, Korea
College of Business Administration, Sogang University, Seoul, Korea

Abstract
Bankruptcy probabilities of private companies are estimated via Bayesian approach. The proportional hazard models are used with time to bankruptcy as a dependent variable. Several financial ratios are included as independent variables. MCMC algorithm is used to obtain the posterior distribution. Finally, bankruptcy probabilities of companies are predicted using the predictive probabilities. 
Keywords: Proportional Hazard Model, Bayesian analysis, Markov Chain Monte Carlo, bankruptcy prediction model.