[教育]銀行風(fēng)險(xiǎn)監(jiān)管能力的工具

1、Credit Registers: a Tool for Improved Bank Risk Regulation and Supervision. 信用登記：提高銀行風(fēng)險(xiǎn)監(jiān)管能力的工具 Jerry Caprio The World Bank世界銀行 ….. BejingSeptember 29, 2004北京2004年9月29日,Outline of the presentation講演大綱,Credit

2、 Registers are tools intended to broaden the access to credit markets … 信用登記著眼于拓寬信貸市場進(jìn)入渠道…2. … making possible better risk assessments on the part of the banks ... 使銀行部門可能更好地進(jìn)行風(fēng)險(xiǎn)評估…… in addition they may

3、 be a tool for more advanced risk based supervision… … 此外還有利于開展高級風(fēng)險(xiǎn)性監(jiān)管 …… as in the case of the assessment of bank capital requirements... ... 同銀行資本金要求的評估相似...…but important challenges lie ahead!! ...

4、 但面臨著重大的挑戰(zhàn)!! ...,CR & Access to credit market信用登記與信貸市場準(zhǔn)入,Banks often do not lend to individual customers lacking “physical collateral” … 銀行通常不向缺少實(shí)物擔(dān)保的個(gè)人提供貸款… and therefore restrict their credit to their wealth

5、iest customers. …因此他們向最富有客戶的信貸受到了限制 A good credit record represents an alternative and “wealth free” collateral: a “reputation collateral”… 良好的信用記錄反映了另一種非財(cái)產(chǎn)擔(dān)保的方式：信譽(yù)擔(dān)保 ... that allows banks to lend to all customers i

6、ndependently from their wealth. ...這就使銀行可向所有客戶貸款而不需考慮他們的財(cái)產(chǎn)狀況,CR & Access to credit market信用登記與信貸市場準(zhǔn)入,World Bank research shows that lending may increase substantially, without increasing the risk exposure, when ba

7、nks supplement their own information with that provided by Credit Registers (Majnoni et al. 2004). 世行研究表明，若銀行將已有信息同信用登記所提供的信息結(jié)合起來，信貸大幅增長的同時(shí)不會帶來風(fēng)險(xiǎn)敞口的擴(kuò)大(Majnoni et al. 2004) 。 Evidence from Argentina and Brazil shows:

8、 阿根廷和巴西的經(jīng)驗(yàn)證明： that a small bank, when adding CR data in its creditworthiness assessment, could increase its lending up to 30% without increasing its risk; 小型銀行若在對信用可靠度進(jìn)行評估時(shí)添加信用登記提供的數(shù)據(jù)，可以增加高達(dá)30%的貸款而不至增加風(fēng)險(xiǎn)； gains are

9、higher for smaller banks. 對小型銀行而言收益更高。,From Access to Credit to Credit Risk Management 從信貸市場準(zhǔn)入到信貸風(fēng)險(xiǎn)管理,Credit information sharing by improving banks’ risk assessment helps banks: 通過改善銀行風(fēng)險(xiǎn)評估的信貸信息共享有助于銀行： to extend cre

10、dit to new previously rationed but creditworthy customer. By affecting credit origination CRs improve the access to bank credit.向新的客戶提供信貸,這些新客戶之前得不到足夠的貸款但卻具有良好信用度。通過影響信用機(jī)構(gòu)，信用登記可以提高銀行信貸的可得性。 to measure the aggregate ris

11、k of their loan portfolios improving risk management. 提高風(fēng)險(xiǎn)管理能力，衡量貸款組合的總體風(fēng)險(xiǎn)? Market Access and Risk Management are in fact two faces of the same coin. ?市場準(zhǔn)入和風(fēng)險(xiǎn)管理是一枚硬幣的正反兩面。,CR & Credit Risk Management信用登記與信用風(fēng)險(xiǎn)管理,

12、World Bank research shows that risk can decrease substantially, at any level of lending, when banks supplement their own information with that provided by Credit Registers (Majnoni et al. 2004). 世行研究表明，若銀行將已有信息同信用登記所提

13、供的信息結(jié)合起來，在任意給定的信貸水平上，風(fēng)險(xiǎn)可以大幅降低。(Majnoni et al. 2004) 。 Evidence from Argentina and Brazil shows that a banks who on average accepts 60% of its loan applicants can decrease its default risk by 30% by using CR data in addi

14、tion to his own information. 阿根廷和巴西的經(jīng)驗(yàn)證明，一家平均放貸率為60%的銀行，若在對信用可靠度進(jìn)行評估時(shí)添加信用登記提供的數(shù)據(jù)，可以減少30%的違約風(fēng)險(xiǎn)。,CR & Credit Risk Management信用登記與信用風(fēng)險(xiǎn)管理,Empirical evidence shows that the design of CRs affects the contribution that C

15、R data may offer to risk management. Notably:經(jīng)驗(yàn)證據(jù)表明，信用登記的設(shè)計(jì)會影響到信用登記數(shù)據(jù)對風(fēng)險(xiǎn)管理的貢獻(xiàn)程度。主要表現(xiàn)在： the presence of positive information about borrowers increases precision of default estimates reducing the amount of risk for any l

16、ending level (to what extent banks may be willing to share positive information, though, is an open issue). 借款人的正面信息提高了違約估計(jì)的準(zhǔn)確性，也減少了任一借貸水平上的風(fēng)險(xiǎn)（但銀行愿意在何種程度上分享正面信息的問題仍未得到解決） appropriate historical length of credit record

17、s should be achieved. 應(yīng)有足夠長的信貸歷史記錄。,From Risk Management to Risk Based Supervision從風(fēng)險(xiǎn)管理到風(fēng)險(xiǎn)性監(jiān)管,What proves useful for Banks’ risk assessment is also useful for Supervisors’ control of banks’ risk exposures. Supervisor

18、s have used CR data for:被證明對銀行風(fēng)險(xiǎn)評估有用的資料同樣對監(jiān)管者控制銀行風(fēng)險(xiǎn)敞口有用。監(jiān)管者已經(jīng)在如下方面采用了信用登記的數(shù)據(jù)： on-site supervision (53 countries out of the 57 countries reviewed by the World Bank); 現(xiàn)場監(jiān)測（世行review的57個(gè)國家中的53個(gè)國家采用）； off-site supervisio

19、n (49 countries out of 57); 非現(xiàn)場監(jiān)測(57個(gè)國家中的49個(gè)采用）； control credit risk concentration. 控制信貸風(fēng)險(xiǎn)的集中。In addition, CRs favor competition and may help preventing the establishment of dominant positions in the banking marke

20、t and in that of credit information.此外，信用登記鼓勵競爭并將有助于防止銀行市場和信貸信息上支配力量的形成。,CR & Risk Based Supervision信用登記與風(fēng)險(xiǎn)性監(jiān)管,The international trend toward a risk based bank supervision (i.e. an approach less based on administra

21、tive rules and more on a close monitoring of banks’ risk exposures) opens up a new role for CRs. 向風(fēng)險(xiǎn)性監(jiān)管轉(zhuǎn)變的國際潮流使信用登記有了新作用。 Adequate credit risk databases for banks and bank supervisors are in fact a precondition for r

22、isk based regulation and supervision (together with adequate skills among bank supervisors, new effective monitoring procedures and sanctions). 向銀行和銀行監(jiān)管者提供的足夠的信貸信息數(shù)據(jù)實(shí)際上是風(fēng)險(xiǎn)性監(jiān)管的前提條件（還包括：監(jiān)管者一定的能力，新的有效監(jiān)控程序和sanction)。,CR &a

23、mp; Risk Based Supervision信用登記與風(fēng)險(xiǎn)性監(jiān)管,The ongoing debate about the desirable bank capital regulation approach offer a useful example of the role that CR data may play in the definition of a risk based regulation in emerg

24、ing countries. CR data may in fact help to solve a:一直以來關(guān)于理想的銀行資本監(jiān)管方法的討論為信用登記數(shù)據(jù)在新興國家對風(fēng)險(xiǎn)性監(jiān)管的定義方面所起的作用提供了有用的例子。信用登記數(shù)據(jù)實(shí)際上將有助于解決一下問題： measurement problem: quantifying the VaR for a banking system without using complex model

25、s;測算問題：不使用復(fù)雜的模型而能確定銀行體系風(fēng)險(xiǎn)價(jià)值的大小； policy problem: indicating how VaRs can be split among Capital and Loan Loss Reserves (an area where Basel 2 does not provide guidance).政策問題：指出風(fēng)險(xiǎn)價(jià)值在資本和貸款損失準(zhǔn)備之間是如何劃分的（新巴賽爾協(xié)議未涉及此問題）,CR &a

26、mp; Bank Capital Regulation信用登記與銀行資本金監(jiān)管,Recent research at the World Bank (Majnoni, Miller, Powell, 2004) shows that countries with CRs, may apply simple statistical techniques to:世界銀行最近研究(Majnoni, Miller, Powell, 2004

27、)表明，實(shí)行信用登記的國家可能采用簡單的統(tǒng)計(jì)技巧進(jìn)行如下操作： Visualize the distribution of Credit Losses and the VaR without incurring into model or estimation errors.將信貸損失和風(fēng)險(xiǎn)價(jià)值的分布形象化，而無需使用模型或產(chǎn)生估計(jì)錯(cuò)誤。 Split the VaR among Capital and LLR requiremen

28、ts.Selected results, presented in the following slides, show how accurate measures of the distribution of credit losses have been obtained from CR data with simple statistical procedures.在資本和貸款損失準(zhǔn)備要求之間劃分風(fēng)險(xiǎn)價(jià)值。接下來的幻燈片中我挑

29、選了一些結(jié)果，它們表明了利用信用登記數(shù)據(jù)和簡單的統(tǒng)計(jì)程序，得到的信貸損失的分布是相當(dāng)準(zhǔn)確的。,CR & Bank Capital Regulation信用登記與銀行資本金監(jiān)管,Step 1: extract from the CR all the loans at time t and check how many have defaulted at t+12.第一步：從信用登記中選出在t時(shí)的所有貸款，然后檢查在t+12時(shí)有

30、多少起違約。Example: default frequency on 70017 loans from a CR例：某信用登記中70017筆貸款的違約頻率,CR & Bank Capital Regulation信用登記與銀行資本金監(jiān)管,Step 2: Simulate a bank portfolio by drawing randomly 500 loans for many (20,000) times.第二步：

31、通過隨機(jī)抽取500筆貸款模擬一種銀行貸款組合，重復(fù)多次（20,000）。Example: distribution median value is US$ 215 mill.例：分布中位數(shù)為215百萬美元,CR & Bank Capital Regulation信用登記與銀行資本金監(jiān)管,Step 3: find the loans that defaulted over 1 year and compute the los

32、ses for the 20,000 portfolios.第三步：找出違約超過1年的貸款，并計(jì)算20,000種組合的損失。Example: EL=4.8% and the UL(99%) =14.8%例： EL=4.8% ,UL(99%) =14.8%,CR & Bank Capital Regulation信用登記與銀行資本金監(jiān)管,The previous chart (Step 3) deserves close

33、r consideration for the following reasons:上圖（第三步）值得進(jìn)一步考慮，因?yàn)椋?it is a snapshot of an “invisible” entity: the distribution of credit losses for loan portfolios;它是一種“不可見”實(shí)體的直觀印象：貸款組合的信用損失分布； it allows to compute the val

34、ue of the VaR at any level of confidence (95%, 99%, 99.9%)它可以計(jì)算任意置信度下（95%, 99%, 99.9%）的風(fēng)險(xiǎn)價(jià)值； it shows how to split the VaR into loan loss reserves (equal to the value of EL) and capital (equal to the difference VaR-EL

35、) thereby insuring consistency of LLR and capital rules.它顯示了如何將風(fēng)險(xiǎn)價(jià)值分為貸款損失準(zhǔn)備（等于EL值）和資本金（等于風(fēng)險(xiǎn)價(jià)值與EL的差值），因而確保了貸款損失準(zhǔn)備和資本金規(guī)則的一致性。,,CR & Bank Capital Regulation信用登記與銀行資本金監(jiān)管,Parameters from the distribution of credit losse

36、s from Mexico, Argentina, and Brazil CRs follows.以下是墨西哥、阿根廷和巴西信用登記中的信用損失分布的參數(shù)。Re-sampling UL and Basel UL (Year 2001)重復(fù)抽樣UL和巴塞爾UL（2001年）,,CR: Open Challenges信用登記：充滿未知的挑戰(zhàn),CRs may continue to contribute to support risk

37、 based regulation and supervision but face important challenges. CRs need to consider:信用登記能夠持續(xù)支持風(fēng)險(xiǎn)監(jiān)管，但是面臨重大的挑戰(zhàn)。信用登記需要考慮： borrowers consolidated positions;借款人合并頭寸； exposures with foreign lenders;外國貸款人的敞口； informatio

38、n of loans after default (such as data on losses given defaults);違約后的貸款信息（如既定違約下的損失數(shù)據(jù)）； integration with private Credit Bureaus;與私人征信局相結(jié)合； loan classifications coherent with Basel regulatory schemes.貸款分級與巴塞爾規(guī)則一致。,CR

39、: Open Challenges信用登記：充滿未知的挑戰(zhàn),I have highlighted with specific examples and figures how CR can offer a significant contribution, not only to widen the access to credit market, but also to improve risk monitoring by bank

40、s, bank regulators and supervisors.我用詳細(xì)的實(shí)例和圖表著重指出了信用登記如何對銀行監(jiān)管者在拓寬信用市場的準(zhǔn)入和改善銀行風(fēng)險(xiǎn)監(jiān)控方面的重要貢獻(xiàn)。 More to the point, I have illustrated how “wider market access” and “more effective risk monitoring” can be seen largely as dif

41、ferent aspects of the same problem.需要強(qiáng)調(diào)的是，我舉例說明的“更寬泛的市場準(zhǔn)入”和“更有效的風(fēng)險(xiǎn)監(jiān)控”能夠在很大程度上被看作是同一問題的不同方面。,CR: Open Challenges信用登記：充滿未知的挑戰(zhàn),This perspective opens new important challenges a) for the development of new application of C

42、R data and b) for the design of CR information databases.這一觀點(diǎn)對于信用登記數(shù)據(jù)新應(yīng)用的發(fā)展和信用登記信息數(shù)據(jù)庫的設(shè)計(jì)是全新的重大的挑戰(zhàn)。 These perspectives add new and fascinating dimensions to the work that the World Bank, jointly with the national author