Research On Financial Distress Prediction Of Listed Companies Based On Cox Proportional Hazards Model

Enterprise falling into financial distress is a continuous worsing and relative long process. If the financial distress symptoms can be discovered and its worsen process can be dynamically showed at it's primary period, this would be very important to the enterprise and its profit chain. On the base of domestic and overseas classic literatures review and summary, this paper try to construct a prewarning model of financial distress for Chinese listed companies and dynamically showed its developing process.First, this paper analysed applicability and superiority of survival analysis method in financial prewarning, and state the theoretical frame in detail. Then it thoroughly analysed the current financial distress in listed companies and the problems existing in its prewarning modle application, all these provided this paper the powerful support in sample choosing and research method. On the basis of predecessors'research results, this paper choose finance prewarning initial matrix and deal with its missing values and abnormal value to maintain their data as much as possible. At last, with 2004-2006 financial ratios data of listed company, prewarning financial distress model was built based on a Cox proportional risk model, analyze the hazards factors and protective factors in finance distress and validate the model.Furthermore, in order to avoid the impacts capacity and stability of the sensitivity of the model, we tested by choosing 1 to 4,1 to 2 and 1 to 1 of sample rates and two cut-off points as well as its different sample observation period of the model The results showed that the discriminant accuracy is relatively high and over 80% when taking the proportion of the normal samples in total samples as cut-off point, and sample size is greater as the prediction accuracy is higher; model's predictability is not subject to the impact of the different sample observation period. It is worth mentioning that the model can not only provide a more accurate forecast, but also has the ability to estimate the company's future survival time and provides the point forecasts of occurrence of financial distress, with the dual function of "judging" and "defusing" risk is the other traditional cross-section warning model can not be realized, which is a more intuitive, dynamic and accurate new method of financial distress prediction and offers important reference and practical value.