征信机构产品在信用卡业务中的应用 Roles of Credit Bureau in Credit Card Market
Experian – Scorex
邹台虹 博士
Dr. Robert Tsou
提纲 Agenda
征信机构简介 Bureau Introduction
信用卡生命周期 Credit Card Life Cycle
征信机构工具 Bureau based Tools
预先筛选名单提供者 Pre-screening List Provider
信贷政策/战略决策 Credit Policy/Strategy Decisions
信用卡申请欺诈甄别 Application Fraud Detection
记分卡 Scorecard
监控服务 Monitoring Services
通知服务——触发因素 Notification Services – Trigger
商业智能服务 Business Intelligence Services
总结 Summary
各国征信机构实例 Some Credit Bureau Profiles
负面信息 Negative
Baycorp
澳大利亚 Australia
正面信息 Positive
JCIC
中国台湾 Taiwan/China
负面信息、纳税信息 Negative, tax
UC
瑞典 Sweden
负面信息、纳税信息 Negative, tax
Experian, D&B
挪威 Norway
负面信息 Negative
Experian
丹麦 Denmark
部分正面信息 Limited positive
BKR, Experian
荷兰 Netherlands
负面信息 Negative
Experian, Equifax
西班牙 Spain
正面信息 Positive
CRIF, Experian
意大利 Italy
负面信息、欺诈信息 Negative, fraud
B de F, Experian
法国 France
正面信息 Positive
Schufa, Burgel, CEG
德国 Germany
所有正面信息 Full Positive
Experian, Equifax
英国 UK
所有正面信息 Full Positive
Experian, Equifax, TU
美国.
国家/地区 Country/Region
征信机构名称 Bureau
信息类型 Type of Information
波兰 Portland
BIK
正面信息 Positive
征信机构数据的效力 Power of Bureau Data
坏账减少程度：无征信机构数据与包括了正负面征信机构数据的比较 Bad debt reduction: No bureau vs. full bureau
正／负面征信机构数据 Positive & negative bureau data
+ 只有负面征信机构数据 + Negative bureau data
无征信机构数据，内部评分 No bureau data, Internal scoring
18个月后的表现 Performance after 18 months
拒绝个人贷款40% Personal Loans Reject 40%
拒绝信用卡33% Credit Cards Reject 33%
拒绝支票账户15% Cheque Accounts Reject 15%
8% 坏账 bad
7% 坏账bad
9% 坏账 bad
7% 坏账 bad
6% 坏账 bad
45+%
5% 坏账 bad
4% 坏账bad
4% 坏账 bad
5% 坏账bad
40+%
38%
以英国为例 UK Example :
美国征信局的数据 . Bureau Data
征信局数据 Bureau Data
每家征信局都有自己的“黑匣子”来进行数据处理 Each bureau has its own “black box” for data processing
信息来自于银行、零售、按揭贷款、汽车、金融（财务公司）、保险、公用事业、企业／个人名录服务和医疗等领域 Information across banking, retail, mortgage, auto, finance, insurance, utility, catalog, medical etc.
数据由查询部门、贸易往来部门、公共记录等部分组成 The data composed with inquire sectors, trade line sectors, public records sectors
每个数据主体都有多条记录 Multiple records for each individual
反映一个特定时间的记录快照 One time snapshot
无支付信息 No payment information
反映24个月的整体网格状态 24 month grid status
信用卡生命周期管理 Credit Card Life Cycle Management
征信机构数据提供了风险、市场营销、运营等工作所需要的信息，以帮助用户管理消费者的信贷生命周期 Credit bureau data provide information for Risk, Marketing, Operation to manage the consumer’s credit life cycle.
时间 TIME
0
利润 Profit
营销（寻找潜在客户） Prospecting
营销对象 Whom to solicit?
提供何种产品？ Which offer?
信用卡种类 Card type
利率/价格 Rate/price
数量 Amount
条款 Terms
市场研究 Market research
业务发起 Origination
信用卡申请欺诈 Application Fraud
授信额度 Credit Limit?
利率/价格 Rate/Price?
条款 Terms?
激活 Activate?
账户管理 Account Management
降低损失 Loss Mitigation
哪种产品/服务 Which products/ services?
风险定价 Risk base re-pricing
授信额度增加/减少 CLI/CLD
授权 Authorize?
重发卡 Reissue?
资本金要求 Capital requirement?
保留 / 退出 Retention/
Attrition
保留对象 Whom to retain?
保留成本 At what cost?
收款 Collections
处理策略 Which treatment strategy?
对象 Whom to collect?
外包 Outsourcing?
商业决策
Business Decisions
交叉销售 / 追加销售 Cross-sell/Up-sell
营销对象 To whom?
产品/服务 What products/ services?
一些征信机构工具（即基于征信数据开发的工具） Some of Bureau Based Tools
预先筛选名单提供者 Pre-screening List Provider
信贷政策/战略决策 Credit Policy/Strategy Decision
信用卡申请欺诈甄别 Application Fraud Detection
记分卡 Scorecard
通用模型 Generic models
定制模型 Custom models
监控服务 Monitoring Services
通知服务——触发因素 Notification Services - Trigger
商业智能服务 Business Intelligence Services
预先筛选（市场推广执行）过程 PRE-SCREENING CAMPAIGN EXECUTION
风险和市场营销方面的筛选标准
Risk & Marketing Selection Criteria
推广活动数据集市
Campaign Marts
按姓名和地址因素进行最后排除
Name & Address Append
Final Suppressions
筛选实施过程
Campaign Execution Process
执行过程
Fulfillment Process
产出结果
Process Output
得出结果以进行推广
Output Files for Campaign Execution
DB2
征信机构数据
Bureau Data
征信机构作为名单提供者 Bureau as List Provider
信息汇总 Information
Aggregated
一些征信机构工具 Some of Bureau Based Tools
预先筛选名单提供者 Pre-screening List Provider
信贷政策/战略决策 Credit Policy/Strategy Decision
信用卡申请欺诈甄别 Application Fraud Detection
记分卡 Scorecard
通用模型 Generic models
定制模型 Custom models
监控服务 Monitoring Services
通知服务——触发因素 Notification Services - Trigger
商业智能服务 Business Intelligence Services
信贷政策和战略决策 Credit Policy and Strategy Decision
征信机构产生的属性被各类金融机构用于制定信贷政策和战略 Credit bureau derived attributes have been used to define the credit policy and strategies on various financial industry.
征信机构属性在信用卡方面的典型应用包括 Typical application on the credit card were
业务发起——贷款发放决定、额度分配、风险定价、产品分配等 Origination - underwriting decision, line assignment, risk based pricing, product assignment etc.
组合管理——授权、额度管理、根据风险重新定价、重发卡等 Portfolio Management – authorization, line management, risk based re-pricing, reissue etc.
客户保留/客户消耗（退出）——交叉销售、追加销售、常规联系、定价等 Retention/Attrition – cross sell, up sell, call routine, pricing etc.
收款——前期、后期、收款渠道选择等 Collections – early stage, late stage, channel selection etc.
市场营销——BT、潜在销售机会选择、销售机会优化 Marketing – BT, lead selection, lead optimization
信贷政策和战略决策 Credit Policy and Strategy Decision
发放 Origination
组合管理
Portfolio Management
保留/消耗 Retention/Attrition
收款 Collection
信贷生命周期
Credit Life Cycle
属性举例
Example of Attributes
提早收款 Early Collection
过去12个月出现报损交易的数量 Number of derogatory trades in the last 12 months
距上次公共记录的时间 Time since last public-record
目前当期未完结交易的百分比 Percent of presently current open trades
免收费用 Fee Waiver
信用卡开卡总数 Total number of open credit cards
其它信用卡的利用率 Utilization ratio on other credit cards
可用额度 Available credit (OTB)
发生严重还款拖欠的次数 Number of major delinquencies
提高授信额度 Credit Line Increase (CLI)
可用额度 Available credit (OTB)
一张信用卡的最高信贷额度 Maximum credit limit on a credit card
平均帐龄 Average age of accounts
销售机会选择 Lead Selection
信用卡开卡总数 Total number of open credit cards
信用卡查询总数 Total number of credit card inquires
距上次信用卡还款拖欠的时间 Time since last delinquency on a credit card
一些征信机构工具 Some of Bureau Based Tools
预先筛选名单提供者 Pre-screening List Provider
信贷政策/战略决策 Credit Policy/Strategy Decision
信用卡申请欺诈甄别 Application Fraud Detection
记分卡 Scorecard
通用模型 Generic models
定制模型 Custom models
监控服务 Monitoring Services
通知服务——触发因素 Notification Services - Trigger
商业智能服务 Business Intelligence Services
欺诈解决方案 Fraud Solution
数据来源： APACS 2005
Source APACS 2005
欺诈造成的损失（百万欧元）
遗失和被盗 没有收到邮件 假信用卡 身份盗窃 未见卡交易
信用卡申请时的欺诈防范 Application Fraud Solutions
欺诈类型 Type of Fraud
篡改申请信息 Application manipulation
邮件寄出但没有收到 Mailed not received
身份盗窃 Identity Theft
假冒已故使用者 Deceased Impersonation
首次付款违约 First Payment Default
结合规律与模型 Combination of Rule based and models
争取把欺诈扼杀在摇篮中 Stopping fraud before it starts
持续的监控能保证欺诈甄别的质量和准确性 Constant monitoring maintains quality and accuracy of fraud detection
当前申请
以往申请
欺诈——申请表筛选 Fraud – Application form screening
将申请表内容与以往的申请表内容进行仔细对照检查 Cross-check application details with previous applications
信用卡申请欺诈Application Fraud
仔细审核申请表各细节，力图发现自相矛盾的地方 Check details on the application form to highlight inconsistencies
当前申请
以往申请
一些征信机构工具 Some of Bureau Based Tools
预先筛选名单提供者 Pre-screening List Provider
信贷政策/战略决策 Credit Policy/Strategy Decision
信用卡申请欺诈甄别 Application Fraud Detection
记分卡 Scorecard
通用模型 Generic models
定制模型 Custom models
监控服务 Monitoring Services
通知服务——触发因素 Notification Services - Trigger
商业智能服务 Business Intelligence Services
征信机构记分卡 Bureau Based Scorecards
征信机构数据被广泛应用于创建统计模型，以帮助信贷生命周期决策 Credit Bureau data is widely used in developing statistical models to aid the decisions of credit life cycle
典型的通用记分卡 Typical generic scorecards
风险 Risk
破产 Bankruptcy
回应 Response
负债 Indebtness
细分行业记分卡和其它市场细分记分卡 Industry specific/Segmentation scorecards
信用卡风险 CC Risk
按揭贷款 Mortgage
汽车 Auto
次级客户 Sub-prime
不同行业的征信机构通用评分 Generic Bureau Score on Various Industry
英国征信局通用评分的表现 Performance results of Generic bureau scores at .
次级 Sub-prime
保险 Insurance
电信 Telcos.
信用卡和零售银行业务 Credit Card & Retail Banking
零售 Retail
基尼系数
Gini Co-efficients
垂直市场
Vertical Markets
围绕信贷生命周期的各类定制模型 Various Custom Models for Credit Life Cycle
Basel II Accord Part H – Solution for Requirements
增强活跃程度和使用次数 Increase activation and usage
行为记分卡 Behaviour Scorecard
触发因素/事件记分卡 Trigger/Event Scorecard
盈利性记分卡 Profitability Scorecard
交易欺诈记分卡 Transaction Fraud Scorecard
客户价值最大化 Maximise customer value
偏好记分卡 Propensity Scorecard
回应记分卡 Response Scorecard
提高收款和坏帐回收额 Improve collections and recovery values
收款记分卡 Collection Scorecard
坏帐回收记分卡 Recovery Scorecard
使保留收入最大化 Maximise retained revenues
保留/消耗记分卡 Retention / Attrition Scorecard
再激活记分卡 Reactivation Scorecard
通过吸引新客户来实现市场份额和盈利的最大化 Maximise market share and profitability by attracting new customers
预先筛选记分卡 Pre-Screen Scorecard
回应/收入记分卡 Response / Revenue Scorecard
新客户登记 Registering new customers
贷款申请记分卡 Application Scorecard
破产记分卡 Bankruptcy Scorecard
欺诈记分卡 Fraud Scorecard
营销
新业务
收款 回收
账户管理
保留 再激活
交叉销售 追加销售
二维风险级别 Two-Dimensional Risk Grading
根据通用和定制风险模型的二维矩阵形成战略和客户机会选择 Strategy and lead selection based on 2-dimensional matrix of Generic and Custom Risk model
高风险High Risk
边际风险 Marginal Risk
低风险 Low Risk
以上：申请和征信机构通用评分的好/坏几率表
Above: Good/Bad odds tabulation of application and generic bureau score
Delphi 评分
未达到
申请评分
用于客户管理的征信机构评分 Credit Bureau Scores for Customer Management
交叉销售和追加销售 Cross-sell & Up-sell
信贷偏好评分 Credit Propensity Score
个人贷款偏好评分（%至20%）
Personal Loan Propensity Score (% to 20%)
%
1
%
2
%
3
%
4
%
5
%
6
%
7
%
8
%
9
%
10
在之后6个月内余额显著增加的百分比 % with Significant Balance Increases in next 6m
得分 Score 十分位 Deciles
一些征信机构工具 Some of Bureau Based Tools
预先筛选名单提供者 Pre-screening List Provider
信贷政策/战略决策 Credit Policy/Strategy Decision
信用卡申请欺诈甄别 Application Fraud Detection
记分卡 Scorecard
通用模型 Generic models
定制模型 Custom models
监控服务 Monitoring Services
通知服务——触发因素 Notification Services - Trigger
商业智能服务 Business Intelligence Services
组合管理的典型数据来源 Typical Data Sources for Portfolio Management
信用卡数据库 Credit Card Data Warehouse
信用卡分析数据集市 Credit Cards Analytical Data Mart
征信机构原始数据 Credit Bureau Raw data
征信机构汇总程序 Bureau Aggregation program
征信机构属性 Bureau Attributes
Others
MC
VISA
信用卡交易信息 CC Transactional Information
决策引擎 Decision Engine
战略 Strategies
行动 Actions
产品线管理
Line Management
根据风险重新定价
Risk Base Re-pricing
客户关怀
Customer Care
交叉销售/追加销售
Cross-Sell/Up Sell
市场促销
Marketing Promotion
1
2
1
2
3
代表进行账户管理的主要输入信息 Represents the key inputs for Account Management
风险和运营信息 Risk and Operational Information
风险管理平台 Risk Management Platform
其它产品数据库 Other Product Data Warehouse
其它产品分析数据集市 Other Product Analytical Data Mart
其它产品关系Other product Relationship
6
4
5
6
征信机构信息Bureau Information
申请资料Application Information
3
支付信息 Payment Information
4
促销资料Promotional Information
5
记分卡 Scorecards
维萨卡
其它卡
万事达卡
一些征信机构工具 Some of Bureau Based Tools
预先筛选名单提供者 Pre-screening List Provider
信贷政策/战略决策 Credit Policy/Strategy Decision
信用卡申请欺诈甄别 Application Fraud Detection
记分卡 Scorecard
通用模型 Generic models
定制模型 Custom models
监控服务 Monitoring Services
通知服务——触发因素 Notification Services - Trigger
商业智能服务 Business Intelligence Services
通知服务 Notification Services
账户审核和预先筛选——每日传送数据 Account review and prescreening - data delivered daily
根据客户确定的“触发因素”产生结果，这些触发因素是可以改变和管理的 Output based on “Triggers” established by the client, which can be changed and managed
用途 Used for the purposes of:
降低损失 Loss mitigation
保留现有帐户 Retaining Existing Accounts
交叉销售/追加销售 Cross-Selling/Up-selling
营销 Prospecting
通知服务——日触发因素解决方案 Notification Services - Daily Trigger Solutions
通过及时发现事件和早期预警来控制风险 Control risk via immediate identification and early warning
降低损失 Loss Mitigation
为正在积极从你的竞争对手处寻求更多信贷的现有客户提供新产品 Offer new products to existing customers that are actively looking for additional credit from your competitors
交叉销售 Cross-Sell
通过识别那些正在寻找新信贷产品的现有客户，来应对这些现有客户不断变化的新需求 Address the evolving needs of existing customers by identifying those shopping for new credit
追加销售 Up-Sell
在客户接受其它竞争对手的产品之前，发现这些潜在的客户耗减情况 Identify potential attriters before they accept a competing offer
保留 Retention
营销 Prospecting
及时发现市场中适合本银行某类产品的潜在客户 Immediately identify prospects who are in the Market for one of your product lines
通知服务的测试结果 Notification Services test results
在6个月内对168,763名客户实施了监控 168,763 consumers monitored over 6 month period
个人银行卡账户 Personal Bank Card Accounts
70,043 每日发出通知数 Total Daily Notices produced
60,268 与客户保留相关的“触发因素” Retention Triggers
9,775 与风险相关的“触发因素” Risk Triggers
比较两个可能发生不良行为时的余额：即帐户审核日和收到通知日 Compared balances at account review date for adverse action vs. date notification received by NS for adverse action
明确每个帐户根据风险触发服务而可能产生的节余（规避的损失额） Determined potential savings for each account by risk trigger
从168,763个监控账户中共规避损失2,179,174美元 Total overall savings of $2,179,174 for the 168,763 accounts monitored
触发因素和帐户表现恶化之间的时间差 Timing between trigger and degraded performance
在那些有触发通知服务且出现账户表现恶化的银行D帐户中，61%的帐户是在通知后的3个月内就出现了账户表现恶化的现象 Of the triggered accounts that have degraded performance on their Bank D account 61% do so within 3 month of the trigger date
触发因素的通知时间与银行账户表现恶化之间的时间（月份）
存在触发通知服务且出现银行账户表现恶化的比例
一些征信机构工具 Some of Bureau Based Tools
预先筛选名单提供者 Pre-screening List Provider
信贷政策/战略决策 Credit Policy/Strategy Decision
信用卡申请欺诈甄别 Application Fraud Detection
记分卡 Scorecard
通用模型 Generic models
定制模型 Custom models
监控服务 Monitoring Services
通知服务——触发因素 Notification Services - Trigger
商业智能服务 Business Intelligence Services
商业智能服务 Business Intelligence Services
商业智能服务是定制的、全面的顾问式服务，它通过以下方式使金融机构能够更好地认识其客户基础 Business Intelligence is a customized, comprehensive, consultative offering that allows organization to better understand their customer base from a variety of perspectives:
使金融机构能将自身与整个市场和某一特定竞争群体进行对比 Ability to contrast your organization to the overall market and specific competitor groups
基于分析而得出结论 Analytically derived conclusions
将新客户与现有客户进行对比 Contrast new versus existing customers
根据分析找出长期趋势 Trend the analysis over time
按一个构建的客户形态来识别潜在客户机会 Profile customers versus prospects
商业智能服务内容 Business Intelligence Flavors
示例分析包括 Example analyses include:
市场份额分析 Market Share analysis
形态分析 Profiling analysis
拖欠分析 Delinquency analysis
产品组合分析 Product Mix analysis
交叉销售分析 Cross-sell analysis
债务移动/消耗分析 Debt Movement / Attrition analysis
评分范围
客户
同类人群比较
排除
总计
总结 Summary
征信行业展望 Industry’s expectation:
提供增值产品，而不仅仅是信用报告 Value added products instead of credit reporting
从批量到（接近）或实时市场营销 Move from batch to (near) or real-time marketing
提供针对不同行业的量身定制的解决方案 Industry specific solution
更频繁地监控贷款组合 Monitor the portfolio more frequently
一体化数据解决方案 Integrated data solution
以事件为基础的决策和相关行动 Event-based decisions and resulting actions
客户可以从自己的桌面电脑直接获得重要的战略和市场推广开发工具 Client desktop access to critical strategy and campaign development tools
提供针对性强的定制产品／服务 Deliver high targeted and tailored offers
从风险缓释战略转向为客户的整个生命周期提供增值服务 Evolve from risk mitigation strategies to life-time value focus
关注新兴的消费者 Focus on emerging consumers
Questions
????
??
??
??
??
??
问答 Q & A
Not all fraudsters have a masterminded plan to deceive lenders so the solution also looks at previous applications made by the same person. Here there might also be other inconsistencies…….for instance on this application they claim that they have been with their current bank for 6 years…..but on a previous application made 6 months ago they might have stated that they have been with their bank for 3 years……..again it is an inconsistency that is worth checking……
ES solution provides full compliance for Basel II requirements and it’s based on best practice risk management.
The different software components include:
Integrated decision system
And the Basel II components
Covering the customers origination and management throughout the customer decision cycle (from origination to loss).
Use this slide to reference the time line. This client compared each point in time when a triggered occurred to 30 days (basically the difference 30-x) to determine a cost. This total cost ended up being the savings of $2m. For instance, if a BK occurred on Day 3 then they would calculate how much they would have lost if they found out about the BK on Day 30 (a normal account review). The difference would be added with all the other results which would be ultimately the amount they’d be saving by using Triggers. This is the power of speed!
This slide emphasizes the importance of Risk Triggers and clearly shows how a bank can put itself in risk by not taking immediate action.
