The Study On The Financial Forewarning Model For Chinese Public Companies Based On The Principal Component Analysis

As our economy market becomes global and the economy in the world is integrative gradually, the corporations are in face of mutative market. It is risky for the managers of the corporations on account of their limited sense and the asymmetry of the information. If the corporations not do their best, then, capital backbone (financial management) will be infracted, which will possibly result in financial distress. So, constructing financial forewarning model could avoid financial distress, which is very important in financial management. In this aspect, there are lots of researches in our country, and many financial forewarning models were constructed. But the models should be perfected ceaselessly, what's more, we should update the data, too, which could better forecast the financial distress.Basing on the circumstances, we research into 120 public companies (60 special treatment companies and 60 non-special treatment companies) which are in Shanghai and Shenzhen securities business until the end of the year in 2006 in virtue of the principal component analysis. Although the method was used in research previously, the coefficients of the principal components were computed variously, some of which are wrong, for instance, some people use factor score coefficient, other people use the coefficients in component matrix. What's more, the standard according to which the principal components are extracted is different, for example, the principal components are extracted according to cumulative percentage of variance (>85%) or total (>1 or >the mean of all totals). After we consult some datum, we extracted the principal components according to the standard of the giving attention to two or more things, that is: first, total is bigger than one; secondly, cumulative percentage of variance reaches 70% around. Then, we begin to do the demonstration research.Chapter 1 gives a brief introduction to the background,significance,the retrospect and actuality of domestic and overseas research in this aspect and the structure of this research.Chapter 2 mainly introduces some research summarization about corporation financial distress forewarning models. This chapter sets out from the basic concept, describes the concept of financial distress in domestic and overseas research, meanwhile, it describes our comprehension. In the next place, we point out five characteristics about financial distress, e.g. external cumulation,outburst etc. and some factors inside and outside which result in financial distress. Eventually, we give emphasis to financial distress forewarning theories of domestic and overseas research, of which overseas research includes three parts, there are initiating research,improving research variables and improving research means.Chapter 3 introduces some research designing steps based on the principal component analysis financial distress forewarning model. In the first place, we fetch in the concept,the basic idea and the mathematic model of the principal component analysis. Afterwards, we list its steps concretely, first, extract samples and financial ratios; Then, analyze the financial ratios of the estimation group depending on the principal component analysis method; Thirdly, select the principal components and interpret them, point out their economic meanings; fourthly, construct forecasting model; Fifthly, estimate forecasting numerical values of the estimation group, sort by the values and select division; At last, test the model by the testing group.Chapter 4 is the core part of this dissertation, we select 120 public companies (60 special treatment companies and 60 non-special treatment companies, not including B stock) which are in Shanghai and Shenzhen securities business from 2004 to the end of the year in 2006. Owing to the research demand, we select three years'financial ratios before the companies were judged as special treatment (thereinafter for short ST) ones. First, 60 ST companies are detached and become two groups at random, the other 60 non ST ones are matched with them correspondingly, hence, estimation group includes 60 samples (30 ST and 30 non ST), testing group also has 60 samples (30 ST and 30 non ST); Next, we select 11 financial ratios (including return on net assets,return on assets,inventory turnover and so on) according to the current financial evaluation system internationally, then we do one-sample T test again, ultimately, 8 financial ratios are selected, e.g. return on assets,receivable turnover ratio,current ratio and so on; Whereafter, we begin to analyze these financial ratios in virtue of SPSS14.0 and EXCEL; In the end, we obtain better testing effect about three years'forecasting results before financial distress: the forecasting exactness percentage before one year when the companies were judged as special treatment ones is 98.33%, 1 of 60 companies is judged wrong. So, this model could forewarn the next year's finance situation of the public companies very well; And the forecasting exactness percentage is 85% before two years, this result is also better; Moreover, the forecasting exactness percentage is 68.33% before three years, it approaches 70%, thus, this result has passed, too.As a result, we could draw a conclusion, this model forecasts very well, it is significant. For instance, it is useful to the corporation manager,the investor and even the government. In addition, we could discover a rule: the forecasting exactness percentage is worse along with time lengthens which before the companies were judged as special treatment ones obviously, which is in accord with the theory.In the last chapter of this dissertation, we evaluate the model impersonally (involving four advantages and six shortages) and give five suggestions to the subsequent pursuers. For instance: adjusting the false financial rates, adding non-financial variables, doing some research by each industry (or by other taxonomies, or adding non-public companies), enlarging sample capability, at last, we hope that the public companies could throw daylight on true finance information and so on, at the same time, we hope that some departments could perfect the systems on special treatment companies. Basing on the true information of the companies, the financial forewarning model could forecasts very well.