Research On The Financial Early-Warning Model Of Listed Companies

In recent years, with the competition in the domestic market becoming increasingly fierce, challenges faced by enterprises are also being more and more severe. As a result, enterprises will easily fall in financial trouble by any carelessness. Considering this, more and more scholars commit themselves to the research of business finance, hoping to build a more perfect warning system to effectively supervise the corporate finance so that when enterprises are caught in financial trouble, the early-warning system will be able to find out the abnormity to ensure a sound development.In previous researches on early-warning models, scholars only choose paired samples from listed companies without "ST" in a certain P-eriod. In this thesis, however, the author will choose paired samples different from theirs to establish early-warning models and through analyzing and comparing them to improve their discrimination effect. Combining with the characteristics of domestic listed companies, the thesis will respectively match the listed companies with "ST" with the listed companies without "ST" and the listed companies that have been removed "ST" shortly. The paired samples the thesis chooses to match with belong to the same categories and are also been listed in similar periods. In the last part, the thesis will use the method combining the factor analysis with the logistic regression to establish early-warning models, and use other samples and data to test them.The result shows the model that uses the normal listed companies without "ST" as paired samples is much better, whose discrimination rate is95.7%and predictive rate for checking samples is82.5%. To further prove the rationality of the selected samples, the thesis compares the financial data of the companies before and after being removed "ST", and finds that the differences between them are not obvious.