A Study on Recursive Defaults for Taiwan Industries:A Competing Risk Approach
Keywords:
Competing Risk Analysis; Cox PH Regression; Default Risk Analysis; Survival Analysis.Abstract
In default risk analysis, the discriminate analysis for default and probit or logit models for default probability are common used. For these analyses, we only consider the effects of financial variables without time effects. In this study, we want to compare the default factors in four different industries, Traditional industry, Metallurgic industries and Electrical industries, Electronic industry, and others. We use two approaches in survival analysis to find the effects of defaults with time effect. First we consider the competing risk model to find whether there is a difference in survival rates between industries. Then we are using Cox proportional hazard rate (PH) model to model default probability. Risk factors of Capital structure analysis, Liquidity analysis, Operating performance analysis, Return on investment analysis and Cash flow are considered. We find it is necessary to separate the industrial types for the analyses since the survival rate for each industry are different. Also the risk factors are different. Financial institute such as Banks can use this model for predict default probability and reduce the risk for holding default customers. The empirical analysis of TEJ data gives the evidence for model adequacy.