| According to the regulation by the CSRC, the listed companies will be subjected to smaller stock price limit and be capped with ST as a result of two consecutive years of losses. How to forecast the financial misfortune of a company as early as possible in order for any remediation measures has been an urgent problem to the company management, its investors and the watchdogs as well. The objective of this paper is to put forward a complex model, by using the neural network model, to forecast possible financial distress of a company.Based on the research of the experts at home and abroad, the author analyzes the occurance and development of the company financial distress from the point of system theory. And then three financial stages - sign stage, distress stage and crisis stage are also presented. 20 financial indicators are selected as initial factors for the forecasting model. In order to reduce the interference of some indicators that are not sensitive enough to the financial condition of a company, Independent Samples T-test and Factor Analysis are successively applied to the 20 indicators, and shortlist 8 final factors for the forecasting models, which include Current Ratio, ROA, EPS Growth Rate, Gross Margin, Turnover of Total Assets, Operation Cash flow per Share, Types of Auditor's Opinion and Revenue Growth Rate.The sample companies, totaling 122, are all of the newly added ST companies in 2007 and their reference companies. The data consist of the above 8 financial indicators of all of the sample companies ahead of being capped ST. The author establishes a neural network model through Matlab, and then inputs the company financial data of 3 years, 2 years and 1year ahead of company ST occurrence, together with the data from the reference company into the model. Empirical study has shown that the forecasting accuracy rises as the company being capped ST draws near, and the model based on the neural network can distinguish these three stages relatively clearly. |
PDF Full Text Request