巴塞尔新资本协议下的内部评级法研究

《巴塞尔新资本协议下的内部评级法研究》由会员上传分享，免费在线阅读，更多相关内容在学术论文-天天文库。

1、2006年5月本研究的计量经济模型为我国商业银行识别违约企业提供了定量化标准，也使得全面风险管理时代的商业银行通过建立稳健的投资组合获得超额收益成为可能。关键词：信用评级巴塞尔新资本协议内部评级法逻辑回归因子分析AbstractThisdissertationmainlystudiestheInternalRatingsBasedApproach(IRB)oftheNewBaselCapitalAccord(theBaselII)．Theresearchreviewedthemainprogressandachievementindomest

2、icandinternationalcreditanalysisandmanagement．thenfocusedonthestudyofIRBanditskeytechniques．Itselected7881istedcompaniesinmanufacturingindustryintheendof2004asthesamplegroup，andchoose14variablesfrom97mainfinancialratios，withtheapproachoffactoranalyses，theresearchsynthesized

3、6independentcomplex—factoras：StructureofAssets，Growing—upAbility，AssetsEfficiency，DevelopInertia，Debt-payingAbiIity，CompetitivePower，etc．Onthisbasis，theresearchbuiltupa6-factorLogitModeltopredicttheprobabilityofdefault．Theresultshows，thehitrateofthemodelis91．8％，and92％totheh

4、itrateofnodefault．thehitrateonpredictionofdefaultis89．2％．ThisdissertationanalyzedthefactorsthatmayinfluencetheLossGiventheDefault(LGD)，andhasproposedthetheorymodelofLGD．Accordingtotherealityofourcountry，thisresearchputtheinternalratingandriskmanagementundertheframeofrelatio

5、nshiplending，buiItacreditdecisionmodelandwithdrawgamingmodelbetweenbanksandtheirenterprisesundertheframeofrelationshiplending，accordingtotheresultofthemodelanalysis，thedissertationproposedthepracticalthinkingfromtwoaspectsoftechnologyandinstitutionalimprovement．Comparedwith

6、otherresearch，theinnovationofthisthesisisclicked：Comparedwithtraditionalstaticmodel，thehazardmodelinthisresearchhassuccessfullyovercomethepossiblyinfluenceofotherresearcheswhichusuallyusethemethodofmatesample：Intheaspectoffactorselection，itstressedthekeycashflowratios，andha

7、smadesurethatthestaticfactorsanddynamicfactorsareunifiedwhichcorrespondstothemanagementenvironmentandmanagementrealityofcommercialbanksinourcountry．HavingintroducedthemaincompositionanalyticapproachintheprocessofLogisticanalyze，ithaveguaranteedtheintegralityofcustomer’sinfo

8、rmationwhilereducingthevariables，thismadethemodelmoresteadyandmorereliable：Thishas