Research On Early Warning Of The Financial System For State-owned Groups In China

The economy has been developed so fast in the socialist market economic system. As to the betterment on effectiveness and the quality on the industry arrangement, composite has played a more important role on the economics. If the management of composite caused any issues, the society will be greatly impacted. In that way, how to find the potential problem and get the backup plan first is crucial to the management and the investor. If one single corporation in the composite has the financial crisis, the situation will get worse because it may trigger the connection problem and lead to serious bad effect. So, as the core of the composite, the parent company should set up an effective risk warning system on financing. Because the shareholder equity for composite is much more complex than the others, this dissertation imported the management related index for upgrading the effectiveness of the risk warning system.Domestic and foreign financial warning from the empirical study of the circumstances, whether from the use of methods, variable selection, or the use of indicators are gradually improving. More diverse methods of early warning models, from the beginning of the discriminant analysis, Logistic regression, factor analysis, the statistical model to the later application of neural network model; index is more comprehensive, from the beginning was the added financial indicators to indicators of corporate governance non-financial indicators; more perfect, research scholars from different angles to establish early warning.According to the deficient usage for the risk warning system in the state owned corporation, Chen Chuxuan (2008) set up one based only on transportation industry and only considered the financial index instead of the other management index. Because composite is consisted of a lot of corporations and more complex, using ordinary model for single company is not accurate. China is a public-owned and socialist-market country. State owned corporations have significant effect on the social market because they are directly controlled by the state and crucial to the economics in China, This dissertation selected State owned corporations as the research object, analyzed the related theories of financial crisis and risk warning models, using the financial, nonfinancial indexes, logistic regression model and judgement analysis to put forward the risk warning model on financial crisis.The samples are based on the state owned corporations from 2003 to 2008. We screened out 56 corporations with financial crisis and compared these samples with other 56 corporations without financial crisis. First, I did analysis on 24 main financial indexes and decided 10 most significant financial indexes afterwards. Then,I used the Logistic regression model with 10 financial early-warning indicators of the principal component and the M1 from M9 the nine non-financial indicators of the early warning model(model 1),followed in its basis of the M10 , M11 and M12 three conglomerates-specific non- financial indicators to establish Logistic early warning model(model 2), and I compared model I and model II accuracy rate of discrimination, again applied discriminant analysis model robustness test, and with the Logistic model more predictive, the last application of the 2008 sample predictive good model for testing. The results show that:(1) The financial crisis is not likely to come if a company has enough cash flow, gets more income, has higher turnover ratio on assets, has higher AR ratio compared to the current assets and sales revenue, the higher degree relationship of diversificationand,has higher ratio of 5 biggest shareholders;(2) ST occurred in the first 3 years than model III than the other three models could in financial difficulties before they occur and make relatively accurate predictions, in financial difficulty before the 3 years the right to judge the rate of more than 76%;(3)Adding group company-specific non-financial indicators of the Logistic regression model and discriminant analysis models than just adding the general non-financial indicators of the early warning model accuracy for higher;(4) for the same information set, membership conglomerates specific non-financial Index of Logistic forecasting model false positive rate low.The dissertation contained 6 parts: 1.Reserach background, motivation, objectives and definitions. 2. Quotation from the researches of foreign and domestic scholars. 3. Theory analysis based on ownership theory, owner financing theory, corporation management theory, management control theory. 4. Research analysis according to the data and sample selection criteria. 5. Practical analysis using the method of logistic regression and judgement method. 6. Conclusion, restriction, and future analysis.There are two creativenesses for this dissertation: 1.The research object is state owned corporation which makes the model more complete and exact. 2. Using both financial and non financial indexes and include specific indicators of enterprises belonging to the group the number of subsidiaries, group companies and their industries to predict whether related diversification to made the system more complete.This dissertation still has restrictions due to the limited capability of the author. For example, the samples are all selected after 2003. Most of the samples are from manufacturing industry. The pair method is 1:1 instead of 1:3(Shi Xiaojun, Xiao Yuanwen, Ren Ruoen, 2005). The dissertation didn't include all the financial and non financial indexes. The system didn't use both the statistics and non statistics method. Otherwise, different risk warning system should be used in different corporations based on the difference of management basis, employee capability, corporation culture, competitiveness in the industry. Therefore, the company should continuously modify the system according to diverse backgrounds. I hope this dissertation can provide some new thoughts on the utilization of the risk warning system. I'd love to modify the research process in risk warning system with other management, scholars, and practitioners.