Prediction Research Of Financial Distress For Chinese Listed Companies Based On Lasso-Logistic Model

Prediction of financial distress has been a topic of interest over the decades because of itsgreat importance to listed companies,interested stakeholders and even the economy of a country.If the prediction of financial distress is reliable,managers of firms can initiate remedial measures to avoid deterioration before the crisis,and investors can grasp the profitability situation of the listed companies and adjust their investment strategies to reduce anticipated investment related losses.Therefore,it is very important to establish an effective early warning system to monitor the financial crisis of listed companies.This paper starts with the analysis of the financial distress situation of manufacturing enterprises in several typical regions of China.Based on the characteristics of manufacturing industry,this paper constructs a financial distress risk assessment index system for listed manufacturing enterprises,and establishes a set of corporate financial distress prediction models to measure The enterprise will be in the financial distress probability,assess the business operation and development status of the enterprise,to reflect the possibility of the company's future default behavior,and provide financial institutions such as commercial banks with credit evaluation and loan decision-making basis for manufacturing companies.This paper proposes the probability prediction models of financial stress occurrences of Chinese listed manufacturing companies through introducing firm-idiosyncratic financial factors,macroeconomic factors and the interaction terms of firm-idiosyncratic and macroeconomic factors to a logistic regression one by one.Our main contribution in this work is three-fold.First,we explore the influence of firm-idiosyncratic factors on the probability of financial distress at 4 consecutive time windows t,t-1,t-2,t-3,by using panel data of a collection of financial factors listed manufacturing companies from 2014 to 2017.We find that a firm-idiosyncratic factor at different time windows has different and significant effects on the probability of financial distress of the selected samples of companies.Second,we select 43 factors from 578 firm-idiosyncratic factors by using Lasso-Logistic model,which helps solving data over-fitting problem and improving the efficiency of prediction.Third,we detect the heterogeneous influence of some of firm-idiosyncratic factors under different macroeconomic conditions on the probability of financial distress by introducing interaction terms to the models.The empirical results indicate that introduce of the interaction terms has improved the prediction accuracy of the financial distress occurrence of selected companiescompared with the model without interaction term and the existing techniques.Therefore,we may provide a way to illustrate the unobservable heterogeneity factors that significantly affect the probability of financial distress of manufacturing Chinese listed manufacturing...