信用风险评级与商业银行信用风险管理

《信用风险评级与商业银行信用风险管理》由会员上传分享，免费在线阅读，更多相关内容在行业资料-天天文库。

1、对外经济贸易大学博士学位论文信用风险评级与商业银行信用风险管理姓名：张贵清申请学位级别：博士专业：国际贸易学指导教师：刘树林20050501对外经济贸易大学博士学位论文信用风险评级与商业银行信用风险管理ABSTRACTCreditriskisoneofthemostessentialrisksinfinancialinstitution.Creditriskmanagementhasevolveddramaticallyoverthelast20yearsinresponsetothespeed-upofglobalizationofeconomyandfinance.Thema

2、nagementsystemofcreditriskdevelopsquicklyandmanykindsofnewtechniqueareusedtoanalyzethecreditriskinsomedevelopedcountries.InChinathesimpletechniqueisstillusedtomanagethecredit,whichisnotsuitableforthefast-developingcommercialbank.Inthecreditmanagementsystemthekeyproblemisthecreditrating.Thenew

3、BaselAccordissuedin2002requiresthebanktoestablishthecreditratingsystemfirstly.Themainmotivationofthispaperistogiveasystematicintroductiontothelatestresearchresultincreditriskmanagementandthemanagementsituationintheactivebanks,comparethepracticeintheChinesebanks,andproposetobuildthecreditratin

4、gsystemfirst.Basedonthedatasetof1900loansthispaperbuildsthreemodelsofclustering,logisticregressionanddiscriminationanalysistoclassifythecreditriskrate.Theempiricaltestresultshowsthatthelogisticregressionmodel,theclusteringmodel,andthediscriminationanalysismodelcanpredictthecreditratewiththeov

5、erallaccuracypercentageof83.4%,72.05%,and68.14%,respectively.Thethreemodelsallpredictthebestandtheworstsampledatamorecorrectlythanthemiddlethreekindsofsampledata.Thus,themodelsbuiltinthispapercanbeusedinthepracticeoftheChinesecommercialbanks.Basedonthemodelsthispaperalsodesignsthecreditrating

6、andthecreditmanagementsystemfortheChinesecommercialbanksThispaperdiffersfromotherresearchaboutthistopicasshownbelow:(1)Thenumberofthesampledatais1900loanswhichismuchbiggercomparedwithotherresearchaboutthesametopic.(2)Theloansareallmorethan10millionRMBandincludealltheindustries.Thedatacanrepre

7、sentthetruesituationoftheloanssincethebanksevaluatetheseloansmorestrictlyandthemanagementintheenterprisesborrowingtheseloansismorestandard.(3)ThispaperisbasedonthepracticeoftheChinesecommercialbanksinchoosingthevariablesandthedata,buildingthe