The Financial Crisis Prediction Study On Introducing Longitudinal Financial Information

As Chinese economy has entered the new normal economy,the market economic system reform of our country has continuously been deepening,and the Security Market has constantly improved itself.Enterprise Financial Risk increasingly become an important factor affecting social stability and restricting economic development.How to prevent and control financial risks also increasingly become a highly valued problem for even every company.Therefore,it is quite necessary to build a sound forewarning system of financial distress for the listed companies and the government part in the Security Market,as well as the stakeholders such as investors creditors and so on.Current research on the financial crisis early warning is mostly static analysis based on cross-sectional data,with little dynamic analysis.Financial crisis is a continuous dynamic process of financial situation's deteriorating,while the cross-sectional data can not reflect the dynamic trend.Thus it is necessary to introduce the longitudinal financial information reflecting the developing trends,on the basis of traditional static index,so as to build a more efficient and dynamic financial crisis warning model.Which is also the meaning of this study.At first,this paper describes the relatively classic research achievement on the financial crisis forewarning model of the domestic and foreign scholars,and make a commentary on it.The article also highlighted the applicability and efficiency of the binary logistic regression analysis in predicting the financial crisis.Secondly,combining the domestic and foreign scholars' definition of financial crisis and our specific national conditions,this paper defines the concept of the financial crisis and longitudinal financial information in its empirical research.Besides,it explains the importance of longitudinal financial information in the financial crisis early warning model.It also interprets causes of the manufacturing listed companies' financial crisis in three aspect respectively,macro environment,manufacturing and their own factors.What's more,it discusses the performance characteristics of financial crisis combined the relative theory.Which also confirms the great significance of the longitudinal financial information for financial crisis warning.In the empirical research,the article selects financial crisis enterprises from the ones that were first specially treated by the SFC(Securities Commission)in 2014 or 2015,and selects non-ST enterprises at a paired ratio of 1:3.Also,based on the causes and characteristics of financial crisis,this paper soundly chose 54 financial indicators and 13 non-financial indicators.Through normality test,significance test,multi-collinearity test and treatment,we can filter some variable indexes,and based on which the paper constructs a binary logistic regression model(named original model)of T-2 year as a stepwise regression way.After that,the article introduce the longitudinal financial indicators,and then process them as the same way as above.And,the paper constructs a binary logistic regression model(named improved model)of T-2 year based the processing result above.At last,this paper draws conclusion through testing the goodness of fit and the actual effect of the two prediction models.The empirical results show that,compared to the original model,the improved model has better goodness of fit and predictive accuracy,especially in avoiding the second type of error,the letter owns more apparent advantage.Finally,the paper gives advice in financial crisis management based on the financial crisis early warning study.What's more,based on the limitations of the empirical research,it provides several recommendations for the future study of financial crisis early warning.