Financial Distress Prediction Models Of Pharmaceutical Listed Companies Based On GEE

The study of financial distress prediction(here-in-after called “FDP”)has been an important research topic for a long time,and the establishment of early-warning mechanism is the key link to avoid enterprises from falling into financial crisis.According to the relevant research,the study of the FDP is focused on universality,and the differences are becoming more obvious with the development of economy and society,so the construction of FDP system for each industries has become a hot spot.The pharmaceutical industry,which is known as “the eternal sunrise industry”,has high technical value and industrial potential.However,this industry has “three importants” characteristics about high investment,high return and high risk,and now,some listed pharmaceutical companies have declined in operating income due to COVID-19,and even financial crisis.As a barometer of economic operation and enterprise operation,how to establish a scientific and reasonable FDP model for listed pharmaceutical companies is the main problem to be settled.By reviewing literatures,we find that most of the studies on the FDP in the pharmaceutical industry are based on cross-sectional data analysis,while few on longitudinal data.Therefore,we mainly focus on the FDP from the perspective of longitudinal financial data,to enrich empirical analysis of enterprise FDP in pharmaceutical industry.Based on the development status of listed companies in China’s pharmaceutical industry and the basic characteristics of companies in financial crisis,we collect the financial data of 16 listed companies that are ST for the first time(2017-2021)and the same number of paired ones in the pharmaceutical industry of Shanghai and Shenzhen Stock Exchange,to analysis crisis warning,and select four major categories of financial indicators to construct a financial warning system.Specifically,after combing the literatures and theories related,the data is pre-processed with standardization,nonparametric testing and factor dimensionality reduction,to obtain a system of financial indicators reflecting the ability of debt servicing,profitability,etc.;and then,three methods involving GEE,LR and SVM are established under the cleaned data,and deriving the parameter estimation and statistical significance;finally,the confusion matrix and some indicators are used to compare and analyze the above methods to find a better FDP model to help identifying financial risks.We find that,whether a pharmaceutical company is in financial crisis or not is related to relevant financial indicators such as profitability,solvency,etc.;the financial crisis prediction rates of above methods established are relatively high,indicating that the selection of financial indicators is effective;the prediction accuracy of the GEE is higher than that of LR and SVM,because it establishes on longitudinal data,which takes into account the intra-group correlation.In conclusion,we conclude that the FDR of listed companies in the this i ndustry by using the GEE is effective,which can identify crisis ones,and play the role of early-warning of such companies to a certain extent.At the end,rational proposals for reversing crisis are given to each companies and stakeholders,hoping to help to improve the management level,optimize the capital structure,reduce the risk and the loss of stakeholders;and provide a little help for the future development of listed companies in the pharmaceutical industry.