The Demonstrative Research On The Early Warning Model For Financial Crisis Of Enterprise

With the great progress of the China's economic system reform and fast development of capital market, more business enterprises have fallen into finance crisis or applied bankruptcy. The research on the finance crisis at home starts late and is limited in the scholarship research field, the result of research is seldom applied in the business enterprise.Thus ,designing the compact and efficient early warning model for Financial crisis of enterprise is of important and realistic meaning.Based on analyzing the early warning model for financial crisis of enterprise systematically, this paper designs and compares three early warning models for financial crisis of enterprise,. which are short-term, general and applied strongly. The thesis is made up of four parts: Part one: based on reviewing the research status about financial crisis of enterprise at home and abroad, put forward the research idea of this thesis.Part two: introduce the relative base theories of the early warning for financial crisis, make contrasting analysis on the current the early warning model for Financial crisis of enterprise. Part three: from the two parts above, bring forward the design project of early warning model for financial crisis of enterprise in this thesis.Part four: the demonstrative research on the early warning model for financial crisis.Selecting 35 the first time deficit enterprises and 35 profitable enterprises on Shanghai and Shenzhen A-Stock Market of 2004 in china as the estimated samples, 30 the first time deficit enterprises and 30 profitable enterprises of 2005 in china as the tested samples. This paper makes three early warning model which are logit regression analysis model, linear discriminate analysis model, principal component analysis model via Wilcoxon test and principal component analysis. As a result of the test samples, the accuracy of discriminate analysis is about 85%.According to the contrastive analysis on two kinds error probability, this paper draws a conclusion about three models:the whole estimate accuracy of the logit regression analysis model is tallest, suited for the government organization; linear discriminate analysis model is simple calculated, easy comprehend, operational, suited for the Small-and-mediumscale investor; the discriminate analysis accuracy to financial crisis enterprise of principal component analysis model is the tallest,which is calculated complicatedly and not used for primitive finance data, thus suited for the professional investment consults organization.