Construction And Empirical Analysis Of The Financial Crisis Early Warning Model Based On Cash Flow

Enterprises usually from normal financial conditions gradually developed to the financialcrisis occurred, until the final bankruptcy is a gradual process, the financial crisis earlywarning research is the indicator of the business, effective health assessment of the financialposition of the enterprise, scientific warning financialcrisis is necessary for the protection ofthe objective requirements of the system of market competition and enterprise survival anddevelopment.Early warning system based on the existing financial crisis accrual system based onfinancial indicators, accrual financial indicators of the system vulnerable to manipulation ofthe operators and lead to unrealistic, to establish the financial crisis early warning in the aboveunreal warning indicators, its accuracy is questionable. And cash flow information is authentic,based on authentic cash flows of the financial crisis early warning model building andempirical analysis has important theoretical and practical significance.On the basis of a review of domestic and international financial crisis early warningResearch and demonstration of cash flow financial early warning function, select1472011A-share market, the financial status of the exception listed companies as samples of thefinancial crisis, the same year, the same asset size in the same industry. paired147normallisted companies as a sample of the non-financial crisis, as a study sample of294listedcompanies, selected17cash flows of financial ratios as research variables, with the SPSS13.0statistical analysis software, the use of multiple logistic regression, multivariate probabilityregression and multivariate the three methods of linear discriminant financial crisis earlywarning model were constructed, using receiver operating characteristic curve ROC test ofthree models of early warning effect. The study concluded: early warning of a multiplelogistic regression model ST-2discrimination correct rate highest, reaching79.6%, followedby the the multivariate probability regression model and multivariate linear discriminantmodel ST-3years,79.3%and78.6%, respectively; ST-1, ST-2years and ST-3years, financingcash inflows of the total cash inflow ratio "discriminant strongest in the ST-4ST-5," investingcash outflow of the total cash outflow ratio and financing cash inflow and outflow ratesdiscriminant ability is the strongest.