BANKRUPTCY PREDICTION INCORPORATING MACROECONOMIC VARIABLES USING NEURAL NETWORK

Since correct prediction of bankruptcy prediction of a company is very important for investors, lenders and managers, most efforts have been done to improve the predictive capability of corporate bankruptcy prediction models. Most previous studies use corporate financial statement to do bankruptcy prediction. However, corporate performance is always affected by macroeconomic conditions. This study explores the effect of macroeconomic variables on improving the predictive accuracy of corporate bankruptcy prediction with neural networks models. The experiments based on data from USA firms shows that neural networks models incorporating macroeconomic variables can made slight improvement on predictive accuracy with comparison to models without the macroeconomic information.