Financial distress prediction for educational organizations under variable interest entity structure

This paper intends to address the gap in the research field on financial distress prediction for educational companies with Variable Interest Entity (VIE) structure, which is a comparatively new and unique business phenomenon in response to China’s policy on prohibiting foreign investment on such organizations. The study utilizes machine learning methods to perform binary classification tasks for healthy and unhealthy companies based on selected financial and non-financial ratios and examined the individual feature’s contribution to the best performing machine learning models.
The results show that traditional financial distress prediction models using financial ratios as independent variables still work for educational companies with VIE structure, and adding efficiency variables can enhance the models’ performance.