Forecasting the Vulnerability of Industrial Economic Activities: Predicting the Bankruptcy of Companies

Katsuyuki Tanaka, Takuo Higash

Abstract

This study introduces a framework for forecasting the vulnerability of industrial economic activities based on the financial statements of companies. We consider the task of identifying the bankruptcy of a company as a classification problem and apply the random forest method to build a highly accurate bankruptcy prediction model. Based on the predictions of our model, we summarize the vulnerability of the economic activities of companies in different industries. We also apply the predicted probability of bankruptcy of our model as risk management of industries. Finally, we demonstrate that our bankruptcy model can predict credit default swaps to a significant degree. The high accuracy of the bankruptcy model and usage of more than 500,000 pieces of company data enable us to provide more precise and reliable forecasts of industrial economic activity.