Stepwise Discriminant Analysis Applied To Credit Risk Measurement Of Publicly Listed Companies

The credit risks usually have another name called thedefault risk, mean that debtor, securities publisher or theother side of trade are unwilling or unable to fulfill theterms of the contract and form breaking a contract becauseof various reasons. Theses will cause bank, investor or theother side of trade to suffer lost; More generally, creditrisks include because of change of credit rating of debtorand honor an agreement change of ability cause market valuechange of debt their and cause loss also possibility. Thecredit risks are the oldest financial risks, it is financialinstitution and most important financial risks that thegovernment face, such as enterprise owner and manager,investor, commercial bank, etc. too. So, tolerance of thecredit risks has all been key subjects that theseorganizations face all the time. This text is devoted tothe tolerance of quantization of the credit risks of listedcompany of our country, thus settle solid foundation forestablishing the system of credit in our country.After the eighties of the 20th century, the developmentof the credit market and the changes of the credit risksmake the tolerance of risk and management research fieldbegin to present a lot of new quantization analytical methodand tolerance Model. By the look of research currentsituation of foreign countries, quantization measure anduse quantization not to measure model mode to manage creditrisks acknowledge generally already, academia andfinancial circle develop a series of technology and methodwith can attempt tolerance accurate and manage credit risksalready too. Watch from what publish already at home issuealready about enterprise and financial institution creditrisks work and thesis of tolerance, there is still fewquantitative analysis that involves, have not seen twokinds of latest methods to measure the credit risks yet upuntil now --The reasoning analytic approach of the optionand application in the real example of VaR method. And makefrom assets up angle is it measure to go on to credit risksand research of management see.The rate of violation, retrieving rate and thedependence in breach is three main parameters of using themodern model to measure the credit risks, among them therate of violation is the keys of these parameters andstarting point. The financial market of our country andcredit system are still at the new stage of going throughtransition and developing, credit risks administrativeskill relatively backward, relevant credit risks assessapplied less likely achievement that study, there are noenough sufficient data as the foundation of structuring themodern model. The results of study of Altman, etc. (1997)indicate, ZETA model has made quite good result indeveloping countries, such as Mexico, etc. Therefore, itaccords with our country's conditions even more to use ZETAmodel to measure the credit risks of listed company. Becauseof the reason of the intellectual property right, detailsof ZETA model have not been totally disclosed. So, thesimulation ZETA model method of this text, adopt pluraldiscrimination to analyse that assess the credit risks ofthe listed company. In addition, we can also use Bayes todifferentiate analysing to predict to the thing that thelisted company break a contract, thus realize the toleranceof substitution to the rate of violation, this generallyuse the realistic meaning undoubtedly with importance tomodern models in our country. In our country, because ofthe influence of the factors of various fields, a lot ofenterprises in bankrupt state are unable to be bankruptaccording to the set procedure, but afford to struggle still,so this text implement into " ST " as the standard ofbreaking a contract with company. " ST deal with especially(Special Treatment) ", list with is it list first kindduring especially deal with the measure to delayingnormally in Limited Company. As one securities marketsurpervision means, " ST " system apply to stock market ofour country begin with 1998 formally. On March 16, 1998,China Securities Regulatory Commission was in the form of" department's rule ", issue " especially deal with thenotice of the way about listed company stock during beingstate unusual ", require Shanghai Stock Exchange andShenzhen Stock Exchange according to the stipulations ofstock listing regulations of stock exchange, the stockexchange of listed company unusual to the state implementsespecially dealing with. The state unusually refers to theunusual or other unusual states of financial situation. Theformer meant and " lost for two years in succession " or" the net assets per share have been lower than the facevalue of the stock ", the latter means the natural calamity,serious accident ,etc. cause production and operatingactivities of company stop or involve possible compensationamount of money not to exceed our company lawsuit situationof net assets basically. According to reason of ST, " thefinancial situation is unusual " has certain determinacies,easier to predict; And it have suitable uncertainty that" other states are unusual ", difficult to predict. In theresearch of this text, we, according to the currentsituation of the listed company of our country and researchcondition, draw lessons from the method and proceduredocuments properly, try one's best to remedy the defect thatexisted in the achievement in the past, use some we thinkappropriate method, " ST " Company predict that analysesto of our country security market.We will choose and announce it is not samples as thepositive research of ST listed company as 47 listedcompanies of ST and 57 in 2004. Meanwhile, combine actualconditions one of our country according to some real exampleresearch results of foreign countries, this text has widelychosen 22 financial rates to analyse. On the choice of themethod, this text will use stepwise diagnostic method toanalyse about the sample data mainly, and carry on therational explanation to the result that the income reach.As to method, is it stepwise diagnostic method to adoptfirst, the result of which woll have a compare with thatof model differentiate then . All statistics analysisresults run and get on SPSS10.0.We use stepwise method and get 6 variable, build themodel of setting up finally.Then we use data carry onparameter estimate.After building the model,we will choose30 ST company and 32 between Sep 2004 and May 2005 as testsample,getting a satisfactory forecast level.Place of innovation of this text lie in choice ofvariable. In recent years, credit risks tolerance andmanagement method has already make considerable progressin the world, develop a lot of new models and new methods.