The Application of Time-dependent Covariates Cox Model for Predicting Bankruptcy

• Main Authors: Jr-Jung Shr, 施智中
• Other Authors: Lyinn Chung
• Format: Others
• Language: en_US
• Published: 2008
• Online Access: http://ndltd.ncl.edu.tw/handle/51382051642616086945

碩士 === 國立臺北大學 === 統計學系 === 96 === Management business is closely associated with society, since if a large company went bankrupt it could have a negative effect on the society. It is crucial to establish a financial warning system to prevent financial institutions from blindly going down the path of financial distress. This paper investigates the forecasting accuracy by employing time-dependent covariates Cox model, and stratified time-dependent covariates Cox model in survival analysis using U.S. firms data during 1989-2006. There are two main findings in this research. First, the time-dependent covariates Cox model with Shumway's variables is used to incorporate the relation between covariates and the survival duration of each firm at each point in time for predicting bankruptcy. The forecasting ability of this model is as accurate as discrete-time hazard model. Second, the stratified time-dependent covariates Cox model stratified with confounded factor, firm size, can further improve the forecasting accuracy for predicting bankruptcy.