Early Warning Of Corporate Credit Risk That Considers The Personal Characteristics Of Executives

Most studies assess the credit risk of enterprises based on financial data,but in fact,there may be financial fraud in the external statements provided by enterprises;On the other hand,as the company’s leadership team,executives directly influence the formulation and implementation of corporate strategy,and may engage in financial whitewashing for their own benefit.Therefore,it is necessary to include the personal characteristics of executives in the index system for evaluating credit risk.This paper selects companies directly ST in China’s A-share market from 2014 to 2022 as research samples.On this basis,the Ridge Logistic,Lasso Logistic,Adaptive Lasso Logistic and Adaptive Sparse Group Lasso Logistic analyzes the impact of senior executive personal characteristics on corporate credit risk under different samples.The main research work and conclusions are as follows:(1)Establish an enterprise credit risk index system that considers the personal characteristics of executives.Eight indicators including gender,age and military experience were selected from the personal characteristics of executives,and the final research results show that the characteristics of corporate executives contain important information for evaluating corporate credit risk,which significantly improves the performance of the model.(2)The Lasso penalty in the feature selection method is used to establish an early warning model of enterprise credit risk.Considering that the enterprise credit risk index system has too many characteristics and serious multicollinearity problems,the Ridge Logistic model is used as the benchmark to construct the Logistic model based on Lasso,Adaptive Lasso and Adaptive Sparse Group Lasso for feature selection and classification of enterprise credit risk prediction.(3)The three feature selection methods selected in this paper all performed well in credit risk evaluation.By comparing the AUC value and F1 value of each model,it is found that compared with the full-variable Ridge Logistic model,the three penalty methods such as Lasso can select the features that play an important role in prediction from many independent variables,and the prediction effect is better.The research conclusions of this paper can provide a certain reference basis for the decision-making of individual investors and financial institutions such as banks.