Corporate Bankruptcy Prediction: a Penalized Semiparametric Index Hazard Model Approach

• Activity Number: 621 - Portfolio Choice, Stock Returns, Bankrupcty, and Default
• Type: Contributed
• Date/Time: Thursday, August 3, 2017 : 8:30 AM to 10:20 AM
• Sponsor: Business and Economic Statistics Section
• Author(s): Shaobo Li* and Shaonan Tian and Yan Yu
• Companies: University of Cincinnati and San Jose State University and University of Cincinnati
• Keywords: Accounting Ratios ; Nonparametric ; Single-index model ; SCAD ; Hosmer-Lemeshow

Abstract

We introduce a flexible yet easy-interpretable index hazard model for corporate bankruptcy prediction under a semiparametric modeling framework. Motivated by the long debate between accounting and finance researchers, we propose a penalized double-index hazard model with automatic variable selection. The two indices are naturally constructed by separate market and accounting based bankruptcy predictors. The unknown functions are estimated by polynomial splines. In order to identify important predictors, a nonconcave penalty function, SCAD, is adopted due to its attractive statistical properties. We develop a comprehensive database of the publicly traded firms in North America manufacturing sector and focus our empirical studies on this largest sector among all industries. We show that the proposed index hazard model reveals a novel nonlinear relationship. The two newly constructed composite indices: market and accounting index may be of great potential interest in practice. In addition, we find that the accounting index would consist of more accounting based predictors as the prediction horizon increases, while market index would include fewer market based variables.