The Early Warning Model Of Listed Companies Financial Distress Based On Logistic Regression Method

With the constant development and perfection of the market economy of our country, the competition between enterprises is becoming more competitive. Our enterprises have competed with enterprises from other countries after China joined the WTO. The corporate finance confronts with greater risk and complexity. Because enterprises falling into financial distress will bring huge economic losses to the shareholders, creditors, employees and government, the accurate prediction of financial distress can play a role in warning stakeholders. The accurate early warning model provides the decision-making activities of the stakeholders with a very practical and effective tool.In the modern economy which takes the companies as center, high quality and efficiency of the corporate governance structure has become a deciding factor of competitive power of companies. If a company has good corporate governance structure, its financial performance is usually healthy. But if a company has weak corporate governance structure, its financial performance tends to deteriorate. The weakening of corporate governance is an important reason for financial distress. In other word, financial distress is the performance and final result of the weakening of corporate governance.This paper uses logistic regression model to analyze how the corporate governance variables influence the prediction accuracy, and applies Granger causality test to show the lead-lag relation between financial and corporate governance variables. The chapters are arranged as follows:Chapter 1 is the introduction of this paper, and introduces the research background, significance and framework.Chapter 2 first explains the meaning of financial distress. Because the domestic and foreign scholars don't have a unified definition of financial distress, based on the previous studies, this paper regards the specially treated companies due to unusual financial situation as the symbol of financial distress. After defining the meaning of financial distress, this chapter illustrates the existing research methods of financial distress. The traditional research methods of financial distress have single ratio analysis, MDA, Z score model and Probit model. In the 1980s, some new approaches appeared, such as survival analysis, neural network model and so on. The research of the relationship between corporate governance and financial distress also began in this period. The domestic and foreign scholars have conducted in-depth study of the relationship between them.Chapter 3 first introduce the method of this research, then 13 variables are selected from many financial indicators which reflect the profitability, liquidity, operational capacity, growth capacity and cash flow of the listed companies. We generally consider that the level of corporate governance often affects the company performance. Corporate governance includes two aspects, namely internal governance and exterior governance. Because our companies'exterior governance is inefficient, we only consider internal governance. Internal governance includes board governance, equity structure and principal-agent problem, so we select 6 variables to reflect the level of corporate governance. The sample of this paper is 100 listed companies which had been specially treated in the stock markets of Shenzhen and Shanghai from 2006-2009. In order to eliminate the adverse effects caused by different industries and asset scale, we select 100 listed companies whose industries and asset scales are similar to the ST companies as a matching sample. This research randomly divides the sample into two groups: one group is used for building the early-warning model of the financial distress; the other one is used for measuring the model's predictive ability.Chapter 4 is the empirical study of this paper. Firstly, we conduct the normality test of the previous year's financial variables and corporate governance variables. The results show that the sample data does not follow normal distribution. So we should select the non-parametric methods for mean difference test. The results of test show that only one financial variable—Inventory Turnover is non-significant. And just one corporate governance variable—Administrative Expenses Ratio is significant. The variables which are non-significant are removed. We use factor analysis to select variables with the highest factor loading from the remaining variables. We choose Net Margin, Quick Ratio, Total Assets Turnover and Cash Flow based on the previous year's data. The total prediction accuracy of the model built with these variables is 93.5%. After adding the corporate governance variable, the goodness of fit, the significance level of regression coefficients and prediction accuracy are improved. Moreover, the model including corporate governance variable has less chance of committing Alpha Error. Because Alpha Error will bring more losses to stakeholders (such as creditors, shareholders, employees and government) than Type 2 Error, corporate governance variables can reflect the existing problems in the business management, adding corporate governance variables in the model is necessary.Logistic regression only explains the relationship between financial and corporate governance variables and probability of financial distress. It can not determine the lead-lag relation between financial and corporate governance variables. We choose 5 variables which are used to build the previous year's model. And we select quarter data of 5 variables of 95 ST companies. ADF test shows that all the variables are significant at 5% level, so we believe that there is no unit root in each series. According to AIC and SC criterions, we consider that the lag of one period is suitable. The results of Granger Causality Test show that Administrative Expenses Ratio and Quick Ratio are leading indicators; Total Assets Turnover and Net Margin are intermediate indicators; Cash Flow is a lagging indicator. When a company is loosely managed, managers concentrate on position consume and M&A regardless of the benefits of shareholders. Agent conflict and loose supervision mechanism are bound to reduce the operational efficiency and profitability. When a company's profitability is being eroded, the company's cash flow will be serious. Then the financial distress may happen.The conclusion is summary of the empirical analysis. It point out the shortage of this research and make suggestions for future research.