A Comparative Study On Financial Early-warning Model Of Listed Company

With the accelerating of global economic integration, enterprises are facing many opportunities as well as facing more intense competition in the market. Under the dual influence of the complex external environment and internal environment, the risks are increasing, the enterprise may, at any time, get involved in financial trouble, and even the financial crisis. However, the emergence of financial crisis for enterprises will experience a gradual process, if the enterprises before involving in the financial crisis can have an effective financial early-warning mechanism, which timely find enterprise abnormal financial position, and take the corresponding appropriate remedial measures.Based on a lot of literature review concerning the financial early-warning model developed by domestic and foreign scholars, and citing previous conclusion of the study and related financial knowledge, the paper first introduced the research background and importance, reviews the research status of the financial early-warning model of research, then make a judgment on above literature review; then looked back the definition of financial crisis made by domestic and foreign scholars, and redefine the concept of financial crisis according to research question of paper. Dissertation also detailed the internal and external reasons which got corporate involved in financial crisis, on the basis of above analysis it defines the data source and the selection criteria of data. Through qualitative analysis, we selected 23 indicators including ratio of enterprise financial and non-financial indicators, the first one of which contains financial indicators of corporate profitability, solvency, operation ability, growth ability and the public Company size, and the second one of which include corporate ownership structure, audit report. By analyzing the preliminary selection of 23 indicators in the manner of profile analysis, paper selected variables which have significant differences in at least consecutive two years, on this basis, and did correlation analysis to exclude all the variables which have strong correlations with others, finally carried on the principal component analysis for the rest of the indicators and converted them into six principal components. Using the data of the sample group, we used the principal component analysis and the logistic regression analysis to model the 6 principal components, in order to get the two models, finally reuse test sets of data samples to test of the two models obtained to choose the better model.