A Study of the Risk Valuation Model of Review Report system on the Issuance of Bank Cash Card --The Case Analysis of One Domestic Bank

[[abstract]]本研究主要目的在建立一個銀行個人現金卡申請已核准但於發卡前所進行的授信回查機制的風險評估模型，係以國內某一銀行的全省信用貸款客戶為資料抽樣蒐集對象，樣本資料為2005年01月至2005年12月，資料內容為已核准之現金卡發卡客戶資料，總共250個樣本，樣本中60個為逾期戶，190個為正常戶。本研究是根據銀行已核准案件的申請書且目前尚未被銀行徵授信人員列為書面信用風險審核要素，但可能影響未來授信成敗，繳款正常與否的表外變數，運用SPSS統計套裝軟體，進行Logistic Regression迴歸分析，探討現金卡戶可能利用聯徵時間落差向同業申貸而發生逾期違約的重要顯著因素，並建構信用風險評估模型，利用回歸式可預測得知申辦者可能產生逾放的機率值。 本研究發現該表外變數在回查後同業額度增加者或回查後同業聯徵查詢次數增加者，信用風險愈易發生，尤其二種變數影響均發生時，信用風險發生的機率則大幅增加；代表客戶可能同時或之前已與其他銀行同業接洽過，核貸前應考量是否有信用過度膨脹擴充之虞，以及是否為其他同業所拒絕承作對象。故除了藉由制式的書面信用評分表中各項風險因子的判斷之外，亦需考慮本研究表外變數的增列，在瞭解各變數的重要性後，可用來作為案件准駁之之重要參考依據。 本研究成果為經由樣本t檢定與Logistic Regression迴歸的量化分析，導出最終的LR模型，此模型不僅適合於目前一般承作信用貸款銀行之應用性，可快速而客觀的處理現金卡貸款的申請資料，更能提供銀行徵信人員在主觀判斷之餘，另外作為客觀判斷的參考值，以提昇現金卡信用風險審核之效果及減少逾期違約之發生。[[abstract]]The purpose of this study was to establish a risk assessment model for credit re-check of approved cash cards before the cards are issued The research samples were credit loan clients of a domestic bank collected during Jan 2005~Dec 2005 A total of 250 samples with approved cash cards were selected 60 samples were non-performing and 190 were normal Based on the applications of approved cases and external variables not listed as credit risk assessment factors but influential to credit approval and normal repayment SPSS was applied to perform a Logistic Regression analysis in an attempt to investigate the significant factors of loan defaults caused when cash card applicants apply for loans from other banks during the time gap of joint credit check and further construct a credit risk assessment model Through the regression equation the over-loan rate could be derived It was discovered that the increase in loan amount approved by other banks after re-check or the increase in the joint credit check times after re-check indicates a higher credit risk Especially when the effects of the both variables are existent credit risk will be drastically increased indicating that the applicant may have simultaneously or previously submitted a loan application to other banks It is necessary to consider excessive credit expansion or denial of other banks before approving the cash card application Therefore in addition to various credit risks listed on the paper credit assessment it is also necessary to include the external variables proposed in this study After assessing the significance of each variable the assessment could be an important reference for case approval or denial Through sample t-test and the quantitative analysis of Logistic Regression the final LR model was derived This model could not only apply to general credit loans but also provide objective assessment of cash card applications In addition to subjective assessment credit staffs in banks could also use this objective assessment model as a reference to enhance the effect of cash card credit risk assessment and reduce non-performance and default