수정 KMV 모형 기반 부도 위험 예측 개선방안 연구

초록

This study utilized data from companies listed on China’s A-share market from 2018 to 2024, aiming to propose a novel approach to improving the accuracy and flexibility of corporate default prediction. A total of 276 companies were selected as the sample, including 138 firms that underwent bankruptcy reorganizations. Gradient Descent was employed to optimize the weights of short-term and long-term liabilities, allowing for the redefinition of the Default Point Threshold (DPT) in the KMV model. By integrating the optimized weights, the study introduced a revised Default Point Threshold (DPT) and an optimized Distance to Default (DD), addressing the limitations of the traditional KMV model. The findings reveal that the revised KMV model significantly enhances the precision and adaptability of default risk prediction compared to the original model. By incorporating firm-specific debt structures, the revised model enables more accurate assessments of default risk and demonstrates flexibility in responding to changes in financial environments. This approach not only overcomes the limitations of the traditional model but also establishes the revised KMV model as a dynamic and robust tool for evaluating default risk with higher precision. This study is expected to expand the practical applicability of the Merton model and contribute to improving the reliability of corporate credit risk evaluation in complex financial environments.

키워드