The Model And Empirical Study On Pre-warning Of Financial Crisis For Listed Company In China

The Financial Crises is world problem. Since 1960s,more and more researchers try to predict bankruptcy through quantitative analysis. In the recent 50 years, many models such as Multivariate Discriminant Analysis (DMA) and Neural Network (NN) come out. However, in China, Financial Crises pre-warning has just begun. The main reason is the lack of uniform Accountant Rule before July, 1,1993. Thus, It is urgent for us to establish Financial Crises pre-warning model in China.On the basis of suduying the domestic and international Financial Crises pre-warning model, this paper regards listed companies received "special treatment"(ST) as a signal of financial crises, and try to predict financial crises of listed companies in China. Firstly, This paper disigns a set of pre-warning index system that includes 7 inspects and 30 indexes. Secondly, based on the analysis and evaluation of the traditional methods in pre-warning financial crises, the article puts forwards the nonlinear combination pre-warning model based on BP neural network. Finally, regarded A-share listed company in Shenzhen and Shanghai stock market of our country as the research objects, selecting 92 fistly "ST" companies and 92 financial health companies in 2002-2004 as the samples, the empirical analysis have been carried on.The result of the empirical analysis indicates that: (1) the indexes in the listed companies don't follow the Nomal Distribution, thus we can't use the DMA model to establish financial crises pre-warning system in China and can't use T-test to select indexes, too. (2) Nonlinear combination forecasting model integrates advantages of statistical models and artificial intelligence models. It has 85.9%, 86.5% and 87.0% accuracy of discrimination respectively in three, two and one years before "ST" and thus takes on stronger superiority and application value.