支持有效的中小企业贷款风险管理——《商业银行小企业贷款与授信业务》管理创新研修班李师刚合伙人企业风险管理部德勤华永会计师事务所13801077672jasonlishigang@.中小企贷款——危险还是机遇？SME Lending –Threat or Opportunity?2.大企业贷款与中小企贷款的区别Differences between Corporate Lending and SME Lending议程3.理想的中小企信用评估流程信息系统框架Ideal IT AGENDAInfrastructure for SME Credit Rating Process4.好的内部评级模型的特征Characteristics of a Good Internal Ratings Model5.实施内部评级系统的目标Objectives of Implementing Internal Ratings System6.中小企组合的挑战和机遇SME Portfolio Challenges and Opportunities7.结论Conclusion8.问答Q&A2COPYRIGHT ©Deloitte, ERS,. ALL RIGHTS
第一节中小企贷款——危险还是机遇？SECTION1SME Lending –Threat or Opportunity?3COPYRIGHT ©Deloitte, ERS,. ALL RIGHTS RESERVED.中小企贷款——危险还是机遇？SME Lending–Threat or Opportunity?�中小企产值占GDP很大比例SMEsform a large part of the GDP�中小企通常不十分透明SMEsare usually not transparent�中小企数量众多SMEsare large in numbers�中小企涵盖许多行业SMEsare widespread amongst various industry sector�单一中小企风险暴露相对较小Relatively small exposure per SME client �缺乏充足内部数据及外部风险对比基准No pooled data or external risk benchmarks available�银行可以对发生的这些高风险贷款收取更高利率Banks could charge high price for the higher risk they pose�中小企通常缺乏银行融资渠道SMEsare generally under banked �中小企融资方面，社区/中小银行通常比大银行具有优势Community/Small banks generally have competitive advantage over large banks in SME lending�相对大客户而言, 通常需提供不同金融产品组合以满足其需要A different mix of banking products vis-à-vis Corporate Clients, are needed to satisfy their needs�银监会与国际金融公司积极推进中小企信贷CBRC and IFC are promoting SME lending4COPYRIGHT ©Deloitte, ERS,. ALL RIGHTS
中小企贷款——危险还是机遇？小企业对中国国民经济的重要性�工商注册企业中90%以上为小企业�小企业的最终产品和服务占GDP的1/3以上—个别省份更高�75%以上的城镇新增就业岗位由非公有经济提供�70%以上的农村转移劳动力在非公有经济中就业�而非公有经济中的大多数是小企业或微型企业5COPYRIGHT ©Deloitte, ERS,. ALL RIGHTS RESERVED.中小企贷款——危险还是机遇？我国各省中小企业对国民经济的贡献率20中小企业贡献率大企业贡献率产出增长率1612%840北京天津河北山西内蒙辽宁吉林黑龙江上海江苏浙江安徽福建江西山东河南湖北湖南广东广西四川贵州云南陕西甘肃青海宁夏新疆资料来源：World Bank6COPYRIGHT ©Deloitte, ERS,. ALL RIGHTS
中小企贷款——危险还是机遇？我国小企业融资情况�小企业贷款余额万亿，比年初增加1412亿元。�授信小企业数量为万户，比年初增加万户�但是，贷款余额的增长与小企业发展的融资需求相比，仍然存在较大的差距。主要银行业金融机构小企业贷款余额占各项贷款余额仅为%。7COPYRIGHT ©Deloitte, ERS,. ALL RIGHTS RESERVED.中小企信贷是一个巨大机遇!! We say it’s a huge Opportunity!!中小企贷款所带来的危险如何才可转化为机遇？How can the threat imposed by SME lending be converted into an Opportunity?中小企贷款的最新发展是使其越来越象零售信贷。高效的、高度一致的、基于数据的、并能产生风险和收益特征优异的组合。这意味着：The state of the art in SME lending is to make it more and more like retail efficient, very consistent, data based and able to produce portfolios with excellent characteristics of risk and return. This implies;�给信贷员/信贷专家一个现代工具（基于网络/软件的），以便他们能够输入每笔贷款申请的关键信息giving every lender/credit expert a modern (internet/software based) tool so that they can enter the critical details of each loan application�该软件必须是基于数据的，这样可以保留所有申请记录，同时对所有已完成贷款也建立数据库。这为银行提供了解那些是好的经验以及那些地方需要进一步提高的机会。the software must then be data based so that there is a record of all applicationsas well as a database of those loans that are done, this gives the opportunity for each bank tolearn about the things they do well as well as ways they can improve.�使每一信贷员的决策过程有章可循并持续一致的，这样每一客户申请都被平等的对待并遵循通常的决策标准making the decision process of each lender disciplined and consistent so that each client application is treated with equal care and with common decision criteria�允许信贷员记录用于决策的定量（财务信息）和定性关键因素allowing the lender a means to record the key elements of the decision both quantitative (financial information) and qualitative �产生对客户的一致的无偏见的风险估计producing a consistent and unbiased estimate of the risk of the ©Deloitte, ERS,. ALL RIGHTS
第二节大企业贷款与中小企贷款的区别SECTION2Differences between Corporate and SME Rating Methodology9COPYRIGHT ©Deloitte, ERS,. ALL RIGHTS RESERVED.为什么大企业信贷方法不适用于中小企�大企业贷款——使用基于交易的信贷技术Corporate Lending –Uses Transaction Lending Technology�中小企贷款——使用关系型信贷技术SME Lending –Used Relationship Lending Technology10COPYRIGHT ©Deloitte, ERS,. ALL RIGHTS
基于交易的信贷技术Transaction Lending Technology�诸如财务报表信贷、基于资产信贷，及中小企信用打分卡等方法，主要基于“确凿的”定量数据Such as financial statement lending, asset-based lending, and small business credit scoring –are primarily based on “hard”quantitative data. �这些交易信贷技术主要依据信息源类型划分——These transactions lending technologies are distinguished primarily by the source and type of information used –1) 财务比率用于财务报表信贷financial ratios for financial statement lending; 2) 可用抵押品——通常是应收帐款和存货——的数量和质量，用于基于资产的信贷；及the quantity and quality of the available collateral–usually accounts receivable and inventory –for asset-based lending; and 3) 适用于小企业信用评分的财务状况和公司主要所有人的历史the financial condition&history of the principal owner of the firm for small business credit scoring�交易类型信贷所使用信息一般来说相对透明并且非常容易在信贷发起时观察The information used in transactions lending is generally relatively transparent and is easily observable and verified at the time of the credit origination. �只有那些信息相对透明、财务报告质量很高、抵押品质量很好，及/或所有者过往业绩优秀的中小企可以从银行那里获得交易型贷款Only relatively informationallytransparent SMEs that have high-quality financial statements, high-quality collateral, and/or owners with high-quality past performance generally receive transactions credit from banks. 11COPYRIGHT ©Deloitte, ERS,. ALL RIGHTS RESERVED.关系型信贷技术Relationship Lending Technology�建立在大量“软”信息基础之上based in large part on “soft”information, �例如公司所有人的特点及可靠程度such as the character and reliability of the firm’s owner, �该公司与供货商的过往关系the history of the firm’s relationships with its suppliers, �公司运营地周边地区业务前景and the business prospects in the local neighborhood in which the firm operates. �信息通常经由银行信贷员与该公司、所有人、供货商、客户及当地社区多个层面长时间收集而来The information is generally gathered through contact over time by the bank loan officer with the firm, its owner, itssuppliers, its customers, and its local community on a variety of dimensions. �采用关系型信贷分析的中小企，其信贷决策所依据信息通常不透明、没有足够的、高质量的、定量的、可核实的信息The SMEs with relationship credit are generally informationallyopaque, without sufficient high-quality quantitative and verifiable hard information on which to base the credit. 12COPYRIGHT ©Deloitte, ERS,. ALL RIGHTS
第三节理想的中小企信用评估流程信息系统框架SECTION3Ideal IT Infrastructure for SME Credit Rating Process13COPYRIGHT ©Deloitte, ERS,. ALL RIGHTS RESERVED.逐渐脱离当前手工流程Moving away from current manual process“通常作法分卡批准文件纸面信贷文件CommExcel数据表Excel打on=EapprovalPaper thexcel spreadscorecardPracdocumentCredit filetice”没有总行对贷款部门的控制No control from head office over lending units没有全行统一的评级过程No bankwideconsistent rating process没有组合视角…没有集中数据库No portfolio view …no centralized data没有用于资本分配的预期损的失计算No expected loss computations for capital allocation没有为管理目的而共享数据No data sharing for management needs14COPYRIGHT ©Deloitte, ERS,. ALL RIGHTS
过渡到基于信息技术的流程To IT based process“最佳实务=数据采集风险模型定价BestDatacaptureriskmodelspricing核准/暴露Practice”整合approval /组合视角检测和再校准预警exposureportfointegrationliotesting &earlyviewrecalibrationwarning“数据、模型驱动的评级实现平台”“Data, model driven rating delivery platform”15COPYRIGHT ©Deloitte, ERS,. ALL RIGHTS RESERVED.预警E信贷评估arly warningC拨备Provisioning动态额度监控redit reviewDynamic Limits monitoring监管资本经济资本Regulatory capitalEconomic capital评级数据库Ratings Database中小企SME大企业Large corporate房地产企业CRE内评法打分内评法打分内评法打分IRB scoreIRB scoreIRB score外部观点－MKMV-预期违约频率external opinion –MKMV -EDFs16COPYRIGHT ©Deloitte, ERS,. ALL RIGHTS 客户customer组合portfolio
资本分配框架的基本构件基础输入经配对违约概率巴塞尔计算Base InputsBasel CalculationsMapped•EL, UL PD•Provisions•Regulatory Capital银行数据和看内部风险评估借贷/风险管理参数法BankInternal Risk ’s data Lending / Risk ManagementandAssessment viewsParametersPD输出output市场信号、标尺外部风险评估监管资本组合结构MaRegulatory rket signals, ExPortfolioternal Risk PDbenchmarkingACapitalssessment风险整合计算ConsoleRisk LGDCombination违约损失率模型经济资本Calculations组合管理报告LGDPortfolio LGD modelEconomic Management CapitalReportingEAD输出output违约风险额模型EAD modelEAD•Economic Capital(or EAD = 100%)风险为基础的定•EL, UL价Risk Based •Risk Contribution•SPricingtress Testing•EVA其他数据经济计算OEconomic Calculationsther data分析ANALYSIS执行EXECUTION17COPYRIGHT ©Deloitte, ERS,. ALL RIGHTS RESERVED.新巴塞尔协议概览Basel II Overview内部评级法Internal Ratings Based风险构成R标准法isk Components基础法高级法AdvancedStandardisedFoundation违约概率评级机构银行自行估计银行自行估计Bank Probability of Default (PD)Agency RatingsBank estimaestimatete违约风险暴露巴塞尔协议规定巴塞尔协议规定银行自行估计Bank Exposure at Default (EAD)Basel estimateBaseestimatel estimate违约损失率巴塞尔协议规定巴塞尔协议规定银行自行估计Bank Loss Given Default (LGD)Basel estimateBasel estimateestimate“数据、模型驱动的评级平台Data, model driven rating delivery platform”随时间而迁移、提升Migrate over time18COPYRIGHT ©Deloitte, ERS,. ALL RIGHTS 市场风险成本业务操作风险成本Market Risk Operational Risk ChargesCharges
第四节一个好的内部评级模型的特征SECCHARACTERISTICS OF A GOOD TION4INTERNAL RATING MODEL19COPYRIGHT ©Deloitte, ERS,. ALL RIGHTS RESERVED.一个好的评级系统的特征Characteristics of a good rating system一个好的评级系统的特征Characteristics of a good rating system:1.为管理信用风险而设Designed for managing credit risk2.把数据/事实和判断结合起来Combines facts with judgment内部评级系统Internal Rating3.把借款人的风险从债项风险分开分析Separates Systemborrower risk from facility risk4.要为整个机构内有关人员所理解和掌握Understood across the organization5.支持好的银行管理操作实务Supports good banking practices20COPYRIGHT ©Deloitte, ERS,. ALL RIGHTS
1为管理信用风险而设Designed for managing credit risk�“有效的Powerful”用信用风险正确地排列商业借款人风险高低的能力Ability to correctly rank commercial borrowers by credit risk-公司甲比公司乙的风险更高吗Is Company A riskier then Company B?21COPYRIGHT ©Deloitte, ERS,. ALL RIGHTS RESERVED.有效Power“最佳做法Best 无效Not Practice”Powerful好1AAGood2BJ3CD4DC违约概率标尺5EEDefault Probability6FFscale7GB8HH9II差10JGBad22COPYRIGHT ©Deloitte, ERS,. ALL RIGHTS
1为管理信用风险而设Designed for managing credit risk�有效的Powerful�“经校准Calibrated”给出正确违约概率水平的能力Ability to give the correct levelof default probability-就定价而言，风险的顺序排列并不适当For pricing, ordinal ranking of the risk is inappropriate-公司乙比公司甲的风险究竟要高多少How muchriskier is Company B compared to Company A?-公司乙的违约概率是1%还是%IsCompany B’s probability of default 1% or is it %?23COPYRIGHT ©Deloitte, ERS,. ALL RIGHTS RESERVED.校准Calibration“最佳做法Best 未经校准Not Practice”Calibrated好1AGoodA2BB3CC4D违约概率标尺D5EDefault EProbability6FscaFle7GG8HH9I差I10JBadJ24COPYRIGHT ©Deloitte, ERS,. ALL RIGHTS
风险定价：客户挽留和提升盈利Pricing mapped to risk ratings对信用质量一般般的客户过低定价—以损失为代价留住客户Underpricingthe average credits --retained at a loss对好信用质量的客户过高定价–导致最佳客户流失Overpricing good quality credits --lose best 两层定价曲线customersTwo-tieredpricing curve风险评级Risk Rating25COPYRIGHT ©Deloitte, ERS,. ALL RIGHTS RESERVED.一个好的评级系统的特征Characteristics of a good rating system1为管理信用风险而设Designed for managing credit risk�有效的Powerful �经校准Calibrated�精细的Granular能分辨细小信用风险单位的差异Ability to distinguish small units of credit risk�更高的精确度有助更好的决策Higher precision leads to better decision making�新巴塞尔协议加强了对精确性的要求Basel II reinforces the importance of greater granularity�德勤最新的研究也支持这一看法RecentMKMV research supports this view26COPYRIGHT ©Deloitte, ERS,. ALL RIGHTS 息差Pricing Spread违约概率Probability of Default (EDF)
精细度Granularity“最佳做法不够精细Not Best Practice”Granular好AGoodBA, B, C1CD内部风险F评级ED, E, 2Internal FRisk RGatingsHG, H, I3差IbadJJ27COPYRIGHT ©Deloitte, ERS,. ALL RIGHTS RESERVED.一个好的评级系统的特征Characteristics of a good rating system1为管理信用风险而设Designed for managing credit risk�有效的Powerful �经校准Calibrated�精细的Granular�“透明的和前瞻的Transparent & forward-looking”评级的驱动因素易于理解和掌握The drivers of the rating are easily understood-评级的方法经过验证吗Is the rating methodology validated?预测将来的能力Ability to predict into future-我的评级系统能捕捉到早期恶化的信号吗Will my rating system catch early signals of deterioration?28COPYRIGHT ©Deloitte, ERS,. ALL RIGHTS
分析一个内部评级模型的预测力Analyze the predictive power of an internal modelM管理流动性盈利能力业务活动杠杠比率anagementLiquidityProfitabilityActivityLeverage行业公司规模增长变动因素债务覆盖IndustryCompanySizeGrowth VariableDebt Coverage内部评级级别Internal rating grades最高风险Highest risk最低风险Lowest risk29COPYRIGHT ©Deloitte, ERS,. ALL RIGHTS RESERVED.分析一个内部评级模型的预测力10个借款人已违约90个借款人表现好10Borrowers defaulted90Borrowers performed wellxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx内部评级级别Internal rating grades最高风险Highest risk最低风险Lowest risk30COPYRIGHT ©Deloitte, ERS,. ALL RIGHTS
用内部风险评级依序排列借款人109876543210020406080100Borrowers in risk grade order已违约的借款人Borrowers按风险评级顺序排列的借款人 that defaultedxxxxxxxxxxxxxxxxxxxx最高风险Highest risk最低风险Lowest risk31COPYRIGHT ©Deloitte, ERS,. ALL RIGHTS RESERVED.随意性模型The random model1098765432100102030405060708090100Borrowers in risk gradeorder按风险评级顺序排列的借款人xxxxxxxxxxxxxxxxxxxx最高风险Highest risk最低风险Lowest risk32COPYRIGHT ©Deloitte, ERS,. ALL RIGHTS 违约者Defaults违约者
最佳模型The best model1098765432已违约的借款人B1orrowers that defau0lted0102030405060708090100xxBxxxx按or风ro险we评rs级 in顺 r序is排k g列ra的d借e o款rd人erxxxxxxxxxxxxxx最高风险Highest risk最低风险Lowest risk33COPYRIGHT ©Deloitte, ERS,. ALL RIGHTS RESERVED.最差的模型The worst model109876543210已违约的借款人0102030405060708090100Borrowers that Borrowers in risk grade orderdefaulted按风险评级顺序排列的借款人xxxxxxxxxxxxxxxxxxxx最高风险Highest risk最低风险Lowest risk34COPYRIGHT ©Deloitte, ERS,. ALL RIGHTS 违约者Defaulters违约者
准确度比率Accuracy Ratio10109988776655432÷43= 65%2110001020304050607080901000102030405060708090100Borrowers in PD OrderBorrowers in PD Order银行内部评级结果理想的评级模型结果IdealBank’s Internal Ratings35COPYRIGHT ©Deloitte, ERS,. ALL RIGHTS RESERVED.一个好的评级系统的特征Characteristics of a good rating system:1.为管理信用风险而设Designed for managing credit risk内部评级系统Internal Rating 2.把数据/事实和判断结合起来Combines facts with Systemjudgment36COPYRIGHT ©Deloitte, ERS,. ALL RIGHTS
2把数据/事实和判断结合起来Combines facts with judgment�客观事实（以财务报表为基础的数据）Objective (financial statement based data) �主观判断（以专家经验为基础的因素选择）Judgments (expert based factor selection)�综合外部和内部的数据信息来源Combines external & internal data sources�为持续性的验证和改善收集数据Data capture for continuous validation & improvement37COPYRIGHT ©Deloitte, ERS,. ALL RIGHTS RESERVED.一个好的评级系统的特征Characteristics of a good rating system一个好的评级系统的特征Characteristics of a good rating system:1.为管理信用风险而设Designed for managing credit risk2.把数据/事实和判断结合起来Combines facts with judgment内部评级系统Internal Rating3.把借款人风险与债项风险分开分析Separates Systemborrower risk from facility risk38COPYRIGHT ©Deloitte, ERS,. ALL RIGHTS
3把借款人的风险从债项风险分开分析Separates borrower risk from facility risk�借款人风险Borrower risk借款人对某一债务违约的可能性What is the likelihood that the borrowerwill default on an obligation?-由财务数据和专家判断驱动Driven by financial data and expert judgment�债项风险Facility risk一旦借款人违约，我们预期损失有多大In the event of a default, how much do we expect to lose?-由债权的优先序、抵押品和担保的质量驱动Driven by seniority, the quality of collateral, and guarantees39COPYRIGHT ©Deloitte, ERS,. ALL RIGHTS RESERVED.一个好的评级系统的特征Characteristics of a good rating system一个好的评级系统的特征Characteristics of a good rating system:1.为管理信用风险而设Designed for managing credit risk2.把数据/事实和判断结合起来Combines facts with judgment内部评级系统Internal Rating3.把借款人的风险从债项风险分开分析Separates Systemborrower risk from facility risk4.要为整个机构内有关人员所理解和掌握Understood across the organization40COPYRIGHT ©Deloitte, ERS,. ALL RIGHTS
要为整个机构内有关人员所理解和掌握4Understood across the organization�它是一个评级系统而不是一个排序系统It is a ratingsystem not a ranking system�评级是预期结果联系的Ratings are linked to expected outcomes�信贷专家对有关信用风险的深入见解要传达至有关用户，支持他们作出连贯一致的信贷决策Credit insight from experts is distributed to users to support consistent credit decisions41COPYRIGHT ©Deloitte, ERS,. ALL RIGHTS RESERVED.一个好的评级系统的特征Characteristics of a good rating system一个好的评级系统的特征Characteristics of a good rating system:1.为管理信用风险而设Designed for managing credit risk2.把数据/事实和判断结合起来Combines facts with judgment内部评级系统Internal Rating3.把借款人的风险从债项风险分开分析Separates Systemborrower risk from facility risk4.要为整个机构内有关人员所理解和掌握Understood across the organization5.支持好的银行管理操作实务Supports good banking practices42COPYRIGHT ©Deloitte, ERS,. ALL RIGHTS
5支持好的银行管理操作实务Supports good banking practices�风险定价Pricing for risk �监管资本和经济资本Regulatory & economic capital�作为迈向积极主动组合管理的基础Foundation toward active portfolio management43COPYRIGHT ©Deloitte, ERS,. ALL RIGHTS RESERVED.内部风险评级Internal Risk Ratings例示债务人评级及其对应的违约概率Example Obligor Grade and Associated Default Probabilities内部贷款评级12345678910InternalLoan Grade平均违约概率Average Probability %%%%%%%%%100%of Default好good平均average劣bad1234567891044COPYRIGHT ©Deloitte, ERS,. ALL RIGHTS
债务人评级及其对应的违约概率Obligor Grades & Default Probabilities •1•2•3•4•5•6•7•8•9•%•%•%•%•%•%•%•%•%••••••••••••••••••借款人评级Borrower Rating••••••••••••••••••�借款人对某一债务违约的可能性有多高What is the ••••••••• that the borrower will default on an obligat• by financials and non-financia•2l. 0i6nfor••••••• ©Deloitte, ERS,. ALL RIGHTS RESERVED.债务人评级及其对应的违约概率Obligor Grades & Default Probabilities %%%%%%%%%债A10%%%0债项评级Facility 项D40%�在遇到违约时，我们预期的损失会是多少In the E50% of a default, how much do we expect to lose?评F60%%�1取.60决于2有.32关债4.权31地位6.的49高低9.、12抵押12品.86和担19保.91的质28量.33好H80%坏. by .,4 t2he 1o4f. , 8级guaranteesI90% ©Deloitte, ERS,. ALL RIGHTS
债务人评级及其对应的违约概率Obligor Grades & Default Probabilities •1•2•3•4•5•6•7•8•9•%•%•%•%•%•%•%•%•%债•A•10%自动核准•B•20%Autoapprove•C•30%项•D•40%•E•50%须进行详尽分析评•F•60%Expanded•G•70%analysis自动拒绝•H•80%级Auto•I•90%reject47COPYRIGHT ©Deloitte, ERS,. ALL RIGHTS RESERVED.解决方案概要Solution Schema内部看法Internal View准备、装载数据和运用组合管理系统组合报告PrepaPortfolio Reportingre, load, and run PM�经济资本Economic 组合数据组合管理CapitalPortfolio ManagementPortfolio Data�监管资本Regulatory Capital外部看法External View�预期损失、意外损失数据仓库其他模型、风险贡献、错误定价DOther Models aEL, UL, RC, Mispricingta Mart(RAROC, etc)�组合报告Portfolio Reporting�经济增加值EVA�其他分析Other Analyses市场价格数据、模型参数估计M积极主动组合管理、arket Price Data, 价格数据风险为基的定价PrModel Parameter Estimatesice DataActive Management,Risk-Based Pricing估值Valuation贷款发起Origination盯市估值Mark-to-MarketValuation贷款发放人、交易员等Originators, Traders, ©Deloitte, ERS,. ALL RIGHTS
第五节实施内部评级系统的目标SECTION5OBJECTIVES OF IMPLEMENTING INTERNAL RATING SYSTEMS49COPYRIGHT ©Deloitte, ERS,. ALL RIGHTS RESERVED.目标一：让数据收集和管理标准化Objective 1: Standardized Data Collection & Management�财务数据输入Financial data entry�非财务信息定义和收集Non-financial information definition and collection�适当的评级因素选择及权重比例设定Appropriate weighting of factors�将评级结果校准到对应的银行评级级别Calibration of output to bank scale�适当的更替处理Appropriate handling of overrides�数据存储的集中化Centralized data storage50COPYRIGHT ©Deloitte, ERS,. ALL RIGHTS
目标二：数据要求覆盖借款人和所涉及的贷款的表现情况Objective 2: Data requirements cover both the borrower and the associated loan performance借款人信息贷款信息Bo informationrrower informaLoantion客户评级数据库贷款会计系统Customer Ratings Loan AccountingDatabaseSystem�为建模目的而跟踪收集违约信息Track defaults for modeling purposes�校准评级模型Calibrate rating model�更准确地定价Price more accurately51COPYRIGHT ©Deloitte, ERS,. ALL RIGHTS RESERVED.目标三：连贯一致地使用所有能获得的信息Objective 3: Using all available information consistently输评级RRaatingstings借款人评级borrower rating出非财务因素N公司管理层的质素、公司地位、行业风险分析Noonn-Financial-Financialmanagement quality, company standing, FFaacctorstorsindustry risk analysis输财务因素业务运作、流动、资本结构和偿债比率分析FFinaoperations, liquidity, capital structure, and debt inannccial Factorsial Factors入service ratio analysis.外部评级EEx构评级、预期违约频率agency xternternaalScores外部评级机lScoresratings, Expected Default Frequencies52COPYRIGHT ©Deloitte, ERS,. ALL RIGHTS
目标四：把借款人违约风险和债项回收风险分离开来Objective 4: Separation of borrower default risk and facility recovery risk借款人评级债项评级Borrower RatingFacility Rating�借款人出现违约的可能性怎样�若借款人违约，我们预期会损What is the likelihood that 失多少In the event of a default, the borrower will default on how much do we expect to lose?an obligation?�取决于债权地位高低、抵押和�由财务和非财务因素驱动D担保质量Driven by seniority, the riven by financials and non-quality of collateral, and financial informationguarantees�结合起来便可产生双维评级Come together to produce a Dual Rating�决定预期损失计算和用以的补偿损失的资本分配Drives Expected Loss calculations and capital allocation to cover loss53COPYRIGHT ©Deloitte, ERS,. ALL RIGHTS RESERVED.目标五：在同一信贷管理程序使用多个评级模型Objective 5: Using multiple rating models inside one credit process客户财务财务分析//风险评级模型信信贷贷审审批批和报表评估SSpprreeaadd//R检讨和检讨CCreredditRitisiskk R Raattiningg M MooddeelslsRReevviewiewCustomerRReevviewiewF大型公司工商中小企金融理inancialsFFininaannccialsialsCPfolio Coor企业Poorrttfolio rppCC&&IISME机构组合管FFIIMMggtt受控制的数据环境经标准化的风险衡量经扩展后的风险管理Controlled dataStandardized riskEnhanced riskenvironmentmeasurementmanagement=改善盈利能力Improved Profitability54COPYRIGHT ©Deloitte, ERS,. ALL RIGHTS
第六节中小企业信贷业务所面对的挑战与机遇SECTION6SME PORTFOLIO –CHALLENGES AND OPPORTUNITIES55COPYRIGHT ©Deloitte, ERS,. ALL RIGHTS RESERVED.中小企贷款–信用风险评估所面对的挑战SMEPortfolio –Credit Risk Assessment Challenges�数量大Large number�没有财务数据No financial accounts�财务数据差Poor financial data�财务数据未经审计/稽核Un-audited financial data�财务数据不可靠Unreliable financial data借款人评级Borrowerratings分析Analyze数据收集Datacapture56COPYRIGHT ©Deloitte, ERS,. ALL RIGHTS
德勤建议采取的方法MKMV Proposed Approach�把信用风险因素分离出来（评级打分卡）isolate key indicators of credit risk (scorecards)�让决策的作出基于对关键因素的判断enable decisions to be made with judgment applied only to critical factors �依据过去历史经验情况衡量有关评级的准确性（好和坏）以改进决策质量measure performance ex post to improve decision quality (goods and bads)�在低波动的情况下获得连贯一致的盈利have consistent profitability with low volatility 57COPYRIGHT ©Deloitte, ERS,. ALL RIGHTS RESERVED.采取该等方法的优势Advantages of such an Approach�其表现能被评估并能持续提升改进can be evaluated for performance and continuously enhanced�不存个人偏好和偏见are dispassionate and unbiased�提供一个能产生连贯一致结论的手段provide a means to produce consistency�帮助确定连贯一致的定价help to produce consistent pricing�得出用以持续改善评级准确性的非常宝贵的数据库produce databases that will be invaluable in continuous improvement58COPYRIGHT ©Deloitte, ERS,. ALL RIGHTS
信用风险评估框架-通用评级打分卡�德勤开发了两个中小企评级打分卡模板MKMVhas developed two SME Scorecard Template�一个用于能提供可靠财务数据的客户One for customers with reliable financial data�另一个用于没有财务数据或财务数据不可靠的客户Another one for customers with none or unreliable financial data�评级打分卡考虑因素包括Scorecard Attributes借款人评级�财务分析FBorrowerinancial Analysisratings�行为评估Behaviour Assessment分析�稳定性StabilityAnalyze�业务评估Business Assessment数据收集Datacapture59COPYRIGHT ©Deloitte, ERS,. ALL RIGHTS RESERVED.小/中规模企业Small / Medium Sized Enterprises打分财务分析行为稳定性业务报表日期报表类型审计方法净销售增长营业利润率应收账款天数表外科目杠杆比率对财务信息的信心度财务信息及时、完整和可靠确认取消澄清帮助60COPYRIGHT ©Deloitte, ERS,. ALL RIGHTS
小/中规模企业Small / Medium Sized Enterprises打分财务分析行为稳定性业务企业的操守好偿债历史没有先前借贷信用历史没经验坦诚度和合作性连贯一致的坦诚和合作确认取消澄清帮助61COPYRIGHT ©Deloitte, ERS,. ALL RIGHTS RESERVED.小/中规模企业Small / Medium Sized Enterprises打分财务分析行为稳定性业务企业开业年限20年以上关系年限2年以上额外的股东支持现有债项的50-75%风险胃口风险寻找业务结构融入变化中确认取消澄清帮助62COPYRIGHT ©Deloitte, ERS,. ALL RIGHTS
小/中规模企业Small / Medium Sized Enterprises打分财务分析行为稳定性业务客户集中度（头帮助3大客户）本地市场前景帮助产品多元化帮助策略方向帮助财务报告和计划帮助行业管理经验帮助产品资源适应性帮助确认取消澄清帮助63COPYRIGHT ©Deloitte, ERS,. ALL RIGHTS RESERVED.小/中规模企业Small / Medium Sized Enterprises评级摘要财务分析行为稳定性评级对等违约概率业务合计更替约概率更替理由评级更替违授权违约风险恶违约损失率预期损失率授信额债项评级承诺建议债项摘要关闭帮助64COPYRIGHT ©Deloitte, ERS,. ALL RIGHTS
德勤信用风险评估框架-通用评级打分卡（续）Example Scorecard Contd..q财务分析Financial Assessment�净销售增长Net Sales Growth�营业利润率Operating Profit Margin�净债务人天数Net debtor days�表外科目杠杆Off balance sheet gearingq行为分析Behaviour Assessment�账户行为操守Conduct of account借款人评级orrower�贷款偿还LoanB Repaymentsratings�信用历史Credit History分析�坦诚度和合作Openness and CooperativenessAnalyze数据收集Datacapture65COPYRIGHT ©Deloitte, ERS,. ALL RIGHTS RESERVED.德勤信用风险评估框架-通用评级打分卡（续）q稳定性Stability�开业年限Years in relationship�股东的额外支持Additional Owner’s Support�风险偏好胃口Risk Appetite �业务结构Business Structureq业务分析Business Assessment�产品多元化Product diversity借款人评级�策略方向SBorrowertrategic directionratings�管理经验Management experience分析�产品资源适应性Production Resources SuitabilityAnalyze数据收集Datacapture66COPYRIGHT ©Deloitte, ERS,. ALL RIGHTS
“今天的做法“最佳做法Today’sBestPractice”Practice”6全行化数据共享Bank-wide data sharing5与审批系统整合Integration with approval system4将输入/评级存档于中小企数据库Archive inputs / ratings in SME database 37-10个中小企借款人评级级别7-10SME borrower grade buckets2概括客户行为规律（收集统计数据而不是意见）Profile account behavior (capture statistics not opinions)1开始收集财务报表数据Begin collecting financial statement data67COPYRIGHT ©Deloitte, ERS,. ALL RIGHTS RESERVED.第七节结论ConcluSECTION7sion68COPYRIGHT ©Deloitte, ERS,. ALL RIGHTS
一个好的评级系统的特征Characteristics of a good rating system一个好的评级系统的特征Characteristics of a good rating system:1.为管理信用风险而设Designed for managing credit risk2.把数据/事实和判断结合起来Combines facts with judgment内部评级系统Internal Rating3.把借款人的风险从债项风险分开分析Separates Systemborrower risk from facility risk4.要为整个机构内有关人员所理解和掌握Understood across the organization5.支持好的银行管理操作实务Supports good banking practices69COPYRIGHT ©Deloitte, ERS,. ALL RIGHTS RESERVED.李师刚问答@ ©Deloitte, ERS,. ALL RIGHTS
