Research On Financial Distress Prediction Based On Mixed Multinomial Logit Model

Financial crisis of 2008 highlighted the shortcomings of risk management practices within the lending environment and risk assessment at the micro level.Lenders and other investors in the corporate sector along with regulators require timely information on the default risk probability of corporations. The continual development of conceptually richer and more accurate forecasting models is of importance to regulators, practitioners, and academics.The thesis takes Chinese 201 listing corporations as the sample, based on the theoretical analysis and the actual situation of Chinese securities market, the companies is divided into three-state classification: financial distress company,sub-health company, healthy company. In the research, the mixed multinomial logit is the main method used, compared with traditional standard multinomial logit,mixed multinomial logit relaxes behaviorally questionable assumptions associated with independence from irrelevant alternatives(I.I.A)and allowed for observed and unobserved heterogeneity, the model shows the greater flexibility. Then multinomial logit and mixed multinomial logit model are employed to establish three-state classification financial distress prediction model using financial ratios data in the investigation respectively.The empirical results demonstrate that model-fits and out-of-sample forecasting accuracy of the mixed multinomial logit are much superior to standard multinomial logit. Also, mixed multinomial logit allows a meaningful interpretation of the role of the mean and variance influence of a particular variable through the structure of its parameter space. The mixed multinomial logit model that constructs in the research not only can be used to predict three-state classification, but also can apply to predict two-state classification. The model have valuable application in financial distress.