Public Disclosure Authorizeddocuments.worldbank.org/curated/en/...Introduction 1 1 Basics of Credit Reporting 5 2 Developing Credit Bureaus in Emerging Markets 21 3 Developing Value-Added

38857

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

ii

© 2006 International Finance Corporation2121 Pennsylvania Ave., N.W.Washington, D.C. 20433Telephone 202-473-3800Internet: www.ifc.org

All rights reserved.

The findings, interpretations, and conclusions expressed herein are those of the author(s) and do not necessarily reflect the views of the Board of Directors of the International Finance Corporation or the governments they represent.

The International Finance Corporation, part of the World Bank Group, does not guarantee the accuracy ofthe data included in this work.The boundaries, colors, denominations, and other information shown onany map in this work do not imply any judgment on the part of the International Finance Corporation orthe World Bank concerning the legal status of any territory or the endorsement or acceptance of suchboundaries.

RRiigghhttss aanndd PPeerrmmiissssiioonnssThe material in this work is copyrighted. Copying and/or transmitting portions or all of this work withoutpermission may be a violation of applicable law. The International Finance Corporation encourages dissemination of its work and will normally grant permission to reproduce portions of the work promptly.

For permission to photocopy or reprint any part of this work, please send a request with complete informa-tion to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA; telephone 978-750-8400; fax 978-750-4470; Internet: www.copyright.com.

All other queries on rights and licenses, including subsidiary rights, should be addressed to CorporateRelations, International Finance Corporation, 2121 Pennsylvania Avenue, N.W.,Washington, DC 20433, USA;fax: 202-974-4384.

Copyright 2006 International Finance Corporation

All rights reservedManufactured in the United States of AmericaFirst Printing September 2006

iii

Table of Contents

Preface v

Introduction 1

1 Basics of Credit Reporting 5

2 Developing Credit Bureaus in Emerging Markets 21

3 Developing Value-Added Services in Emerging Markets 45

4 Legal and Regulatory Framework 55

5 Case Studies 61

Compuscan, South Africa : Successfully Serving Microlenders 62

TransUnion Central America (TUCA): Building a Regional Solution 63

SIMAH, Saudi Arabia: Long-term Stakeholder Commitment 64

Estealam, Egypt : Building the First Private Credit Bureau in Egypt 65

Vietnam : Public Sector Support for the Establishment of a Private Bureau 66

Annex 1: Maps 67

Annex 2: Countries with Credit Information Sharing Laws 73

PrefaceIn 2001, the International Finance Corporation(IFC) launched the Global Credit Bureau Program.Since the launch of the program, the IFC has sup-ported the development of credit bureaus in over40 countries through technical assistance andinvestment, including support to the regional cred-it bureau in Central America and the first privatecredit bureau in Egypt, work on the legal and reg-ulatory framework in Kenya and Panama, andongoing assistance towards the development of aprivate bureau in Vietnam. The IFC, together withthe World Bank, began monitoring the creditreporting environment in over 100 countries, andthe results are included and disseminated throughthe Doing Business report.

The content of this Credit BureauKnowledge Guide reflects the IFC’s experiencewith credit bureau markets, and its purpose is toprovide a comprehensive overview of the develop-ment of credit bureaus. IFC’s experience in emerging markets indicates that global knowledgeon credit bureaus is fragmented, much like thecredit information sharing environment itself inmost emerging market countries.The objective ofthis guide is, therefore, to disseminate best practices in credit bureau development and fur-ther contribute to the development of creditbureaus in emerging markets.

The Credit Bureau Knowledge Guide wasprepared by a team led by Nataliya Mylenko and

v

comprising the members of the IFC Global CreditBureau Program team: Tony Lythgoe, OscarMadeddu, Colin Raymond, Shalini Sankaranarayan,Peter Sheerin,and Stefano Stoppani. The work wascarried out under the general direction of PeerStein. The authors would like to thank colleaguesin the World Bank Group for their continuous sup-port of the Global Credit Bureau Program workand preparation of this Guide.We are also gratefulfor the generous contributions of the creditbureaus around the world that made possible the development and publication of this Guide. Wewould particularly like to acknowledge the design,layout,and production support of Aichin Lim Jonesand editorial assistance of Madeline Nevins.

We would like to acknowledge the supportof our donors, without whom the Global CreditBureau Program’s activities would not have beenpossible. Specifically we would like to thank theItalian government for its support of our activitiesin Eastern Europe and Latin America and theCaribbean; the government of the Netherlands forits support of our activities in Africa; theNorwegian government for overall program support and support of our activities in Africa; theAustralian government for its support of our activ-ities in Vietnam; the government of New Zealandfor its support of our activities in Pakistan andIndonesia; and Visa International for global program support.

We hope this Guide will prove both informa-tive and useful.

Introduction

Credit bureaus are essential elements of the finan-cial infrastructure that facilitate access to finance.Today, less than 25 percent of the people living indeveloping countries have access to formal finan-cial services, compared to up to 90 percent indeveloped markets. Financial sector developmentunleashes the productive power of enterprisesand facilitates inclusion of the informal sector inthe formal economy. Access to savings and creditin rural areas allows farmers to smooth consump-tion and often survive the unpredictable risks ofdroughts and natural disasters. Obtaining a loan tosend children to school helps a family create bet-ter lives for their children and reduces the need forharmful child labor. Having long-term financing tobuild a proper home is the direct result of a complex interplay of different financial intermedi-aries within the right financial infrastructure andregulatory framework.

Banks play a central role in extending finan-cial services within an economy. In most markets,commercial banks began by focusing on largecompanies and select retail clients. Initially, theirorganizational structure made it too costly to servesmaller business clients and the mass markets.Accordingly, it has been primarily through infor-mal financial services and non-bank credit that atleast some of the needs of smaller entrepreneursand communities have been and are being met.These include money-lenders, supplier credit, andmany forms of mutual financial self-help groups,

such as the Rotating Savings and CreditAssociations or Tontines in Africa.The credit unionmovement that originated in the 19th century andhas since spread around the world is probably oneof the most prominent examples of the power ofthese mutual financial self-help groups.The rise ofcredit unions and the revival of banks’ social commitment towards their communities inspiredthe rise of microfinance in developing countriesover the past two decades. The approach to lending, however, has remained traditional: makingdecisions, based on subjective judgments, about aborrower’s propensity to repay supported by alternative risk-mitigating mechanisms such asgroup guarantees.

A true revolution in lending occurred withthe introduction of modern financial technologies,and this revolution has made access to creditalmost ubiquitous in developed markets.The tech-nologies allowed banks to move from the tradi-tional approach where credit is granted based onsubjective judgment to more automated processesbased on quantitative models. As a result, lendersare able to deliver financial services at significant-ly reduced costs and expand credit to broader seg-ments of the economy, thus further democratizingfinancial services. In particular, the introduction ofcredit scoring in the fifties in the United States—coupled with the automation of workflow andcredit underwriting—played a key role in therapid rise of consumer lending. Credit bureaus are

Introduction

1

critical in helping lenders make faster and moreaccurate credit decisions. Credit histories not onlyprovide necessary input for credit underwriting,but also allow borrowers to take their credit histo-ry from one financial institution to another, there-by making lending markets more competitive and,in the end, more affordable.

The first chapter of this Guide provides thebasis for understanding the operations of creditbureaus. It draws on the most recent empiricalresearch conducted by the World Bank on theindustry trends and the effects of using creditinformation on availability of financing andimproved risk management.

Although the first credit bureaus may betraced back to the early eighteen hundreds inLondon, modern credit bureaus have rapidlyevolved only since the fifties fueled by improve-ments in technology and expansion of credit.Among the developing and emerging markets,Latin America has some of the oldest creditbureaus in the world,but not until the nineties didcredit bureaus take off in most other developingand emerging markets. Between 1990 and 2005,the total number of private credit bureaus hasmore than doubled. In Asia, many emerging markets turned towards credit reporting after thefinancial crisis in the nineties. New credit bureaushave emerged at a rapid rate in Eastern Europeover the past five years, with many of the projectsthat started in the nineties eventually coming tofruition.The Middle East and North African regionhas only recently seen a growing interest in creditreporting, with new developments underway inMorocco, Egypt, and Pakistan. Sub-Saharan Africa,except for South Africa which is home to one ofthe oldest existing credit bureaus, is still laggingbehind, but many reform-minded countries are taking the lead to support their development inline with reforms for greater access to financing.

Chapter 1 also discusses the roles played byconsumer and commercial credit bureaus to support lending to small businesses.With the risein retail banking, small business lending hasbecome the latest frontier in innovation.Historically, small business borrowers represent a

difficult market to serve because of the traditionalhigh-cost approach of judgmental credit evalua-tion. Wells Fargo pioneered the adaptation of consumer lending technologies to small businesslending in the nineties in the United States.Although no dedicated small business creditreporting existed in the United States until a fewyears ago, consumer credit histories of the ownerof a business proved highly predictive of the credit performance of that business. The innova-tions in small business lending have since beenadopted widely in developed countries and havealso begun to find their way into developing countries. Microfinance institutions, which haverelatively high operating costs,have seen this inno-vation as an opportunity to reduce cost andbecome more competitive. As traditional retaillenders have started poaching their clients insome markets, such as Bolivia, it has also becomemore important for microfinance lenders to joinand support credit bureau initiatives.

Reporting on small and microbusiness seg-ments of the economy, have been neglected byboth consumer and commercial credit bureaus inthe past. Even in the United States, it took time foran industry consortium to launch small businesscredit reporting in 2002. Several developing mar-ket credit bureaus, such as in Thailand, India, andmore recently Turkey and the Kingdom of SaudiArabia (KSA), have already incorporated provisionof small business credit reporting into their busi-ness plans to avoid the mistakes of their moredeveloped counterparts.

The second chapter of the Guide summarizesthe experience of the IFC credit bureau expertteam in the development of private credit bureausin countries around the world. The chapter pres-ents analyses of the various approaches to thedevelopment of the bureau and discusses the tech-nology, financial, and staffing issues a developingbureau must address.

Development of a credit bureau takes a longtime, and requires a long-term commitment of allstakeholders. The entire process of setting up acredit bureau, from initial discussions to publiceducation and work on the legal and regulatory

CREDIT BUREAU KNOWLEDGE GUIDE

2

In many countries, information-sharing can-not begin because an adequate legal and regulato-ry framework is lacking. Chapter 4 presents anoverview of approaches to regulating sharing ofinformation.With the rise in retail lending and thecollection of data on individuals and small busi-nesses by credit bureaus, concerns about data pro-tection and consumer rights are also on the rise. Insome countries, this debate has been highly politi-cal; in others, the debate focuses more on recentabuses, such as identity theft. The latter hasbecome much more than a nuisance, especially inthe United States where people spend more andmore time protecting the integrity of their credithistories. This situation emphasizes the impor-tance of security measures that credit bureausmust take, but it also has more far-reaching impli-cations for the kinds of data that can be used forcredit decisions and the way bureaus ensure thequality of the data and value-added services theyprovide.

The difficulties in ensuring data quality,which many developing markets face, could delaythe startup of a new credit bureau. Developingmarkets are not the only ones that face this chal-lenge, however, as data quality concerns are alsopresent in more developed markets, including theUnited States. A recent study by the ConsumerFederation of America and the National CreditReporting Association revealed a significant varia-tion in credit score accuracy and in the quality ofunderlying credit history data among the leadingbureaus.Accordingly, the future of credit reportingwill not only require further consumer educationon the use, benefits, and risks of credit reportingbut also consistent endeavors by credit bureaus toensure data quality and consumer access.

Credit reporting legislation should carefullybalance the ability of creditors to share informa-tion with the individual’s right for privacy. Banksoften use their secrecy and confidentiality provi-sions as an excuse not to share information. Banksgenerally are willing to provide information ondefaults but not on good loans.This unwillingnessto share positive information, however, limits com-petition and does not allow a good borrower toleverage his/her good credit history to obtain bet-ter terms of credit. The borrower has a right to

framework, to actual implementation of thebureau’s systems, to uploading data and issuing thefirst credit report may take five years or longer.Active participation of creditors and the strongsupport of government are necessary in this effort.In many emerging markets, banks are the largestcreditor to individuals and firms in the formal sys-tem, hence credit bureau development oftenfocuses initially on facilitating information-sharingamong banks and then includes other creditors,such as telecom companies and retailers.

Credit bureaus are characterized byeconomies of scale, and coordination among cred-itors is critical for operations startup. In manycases, the strong support of bank supervisors aswell as the willingness of government to provideeasy access to public databases, are critical toenable credit bureau establishment. In some cases,the central bank opted to operate a credit registryand provide the data to lenders; more recentlyRussia and Kazakhstan chose a private sector solu-tion with strong encouragement from bank super-visors to share information.

Political support and willingness to shareinformation are essential, but challenges do notstop there. Once the creditors are ready to shareinformation, the bureau has to overcome multipletechnical challenges. In several countries, the infra-structure for data exchange is inadequate; uniqueIDs are unavailable; or other identifying informa-tion, such as names, addresses, dates of birth, arerecorded incorrectly and/or inconsistently. Allthese issue make the collection and merging ofinformation difficult, but they should not stop thedevelopment of a bureau. In many cases, the set-ting up of a bureau serves as a wake-up call forlenders to start capturing and storing necessaryinformation. Over time, it enables banks to bettermanage risks and optimize lending processes.

Basic information exchange is a first step.Thebureau uses this information to provide a compre-hensive analysis of borrower creditworthinessthrough such techniques as credit scoring. Thebureau can also use the information for portfoliomonitoring and fraud detection— just a few of thevalue-added services discussed in Chapter 3 that abureau can provide.

Introduction

3

have his or her credit history disclosed to anylender he or she may approach to obtain credit.The law should enable a credit bureau to facilitateinformation-sharing while at the same time ensuring data security and protection of data subject’s rights.

Credit bureaus are an important element inpromoting responsible lending. With the con-sumer lending crisis in Hong Kong (China) andSouth Korea just a few years ago, reckless creditcard lending in the absence of credit bureaus withpositive information led to the over indebtednessof individuals and subsequent rise in personalbankruptcies. Since then, Hong Kong (China)introduced positive credit reporting in order toreduce the risk of this happening again. As othermarkets struggle with predatory and reckless lend-ing, credit bureaus can play a central role in allow-ing lenders to evaluate indebtedness of clients andsetting prudent and responsible lending limits.

Rounding out the theoretical discussions andpractical guidelines are five case studies aboutcredit bureaus that have been established or arebeing established in recent years in different partsof the world: an example of a successful creditbureau serving microlenders in South Africa; aregional credit bureau in Central America,which isa promising solution for smaller markets wherelenders operate on a regional basis; a credit bureauin the KSA that demonstrates the importance ofthe long-term commitment of the stakeholders tothe bureau setup; the first credit bureau in Egypt,which demonstrates how a private credit bureaucan be set up in a relatively short time when allstakeholder interests are aligned and the projecthas the strong backing of the authorities; and aVietnamese bureau that shows the importance ofa comprehensive policy for the development ofthe private credit bureau and highlights the impor-tance of public sector support.


4

1Basics ofCredit Reporting

1.1 Definition of a CreditBureauA credit bureau is an institution that collects infor-mation from creditors and available public sourceson a borrower’s credit history. The bureau com-piles information on individuals and/or smallfirms, such as information on credit repaymentrecords, court judgments, and bankruptcies, andthen creates a comprehensive credit report that issold to creditors.

Credit bureaus differ from credit rating agen-cies, such as Standard & Poors (S&P), Moody’s, andFitch, which collect financial information on largecompanies; conduct detailed analyses of opera-tions, finances, and governance of such compa-nies; and then issue credit ratings. Credit bureausfocus on smaller creditors, mostly concentrate oncredit repayment records, and rely on statisticalanalyses of large samples of borrowers and not onin-depth analysis of individual companies.

Credit bureaus are essential to the success ofcredit markets. They serve as indispensable toolsused by financial institutions to support their retaillending business. Credit bureaus help address thefundamental problem in financial markets knownas “asymmetric information,”which means that theborrower knows the odds of repaying his or herdebts much better than the lender does.The inabil-ity of the lender to accurately assess the creditworthiness of the borrower contributes to higherdefault rates and affects the profitability of thefinancial institution.

Lenders address this problem by investigating aborrower’s ability to repay and/or by requiring col-lateral to cover the loss in case of a default.Requiring collateral is often problematic, especial-ly in developing countries and particularly in thecase of new firms, micro-entrepreneurs, and smalland medium-size enterprises (SMEs), which oftenlack significant assets for use as collateral. In addi-tion, the costs to lenders of seizing and liquidatingassets that were used as collateral can be signifi-cant and the process can take a long time.According to a World Bank survey,1 in most devel-oping countries it takes one to two years toenforce a contract and costs around 20-40 percentof the cost of the debt. In extreme cases, for exam-ple in the Congo, it takes on average three years toenforce a contract and may cost up to 250 timesthe cost of the debt.

Hiring investigators to check borrowers’ back-grounds is costly. Conducting in-depth backgroundchecks, while justifiable for larger loans, is not pos-sible for small loans. The unavailability of informa-tion at a low cost restricts the ability of lenders toexpand and profitably run retail lending operations.

Monitoring and screening borrower behavioroffers an alternative strategy to reduce the prob-lem of asymmetric information. Past behavior is anextremely reliable predictor of future behavior.Forexample, many countries commonly grant credit

Basics of Credit Reporting

5

1World Bank. 2005. Doing Business in 2005: Removing Obstacles to Growth.Washington, D.C.: World Bank, International Finance Corporation and OxfordUniversity Press.

to a firm only after the firm has had an accountwith the bank for at least six months to a year,which allows the bank to observe the firm’s cashflow. Another alternative, the group lendingapproach, mostly used by microfinance institu-tions, allows lenders to provide loans to individualborrowers who, through participation in thegroup, have developed a credit history with theinstitution. Only then does the lender extend indi-vidual loans. In these examples the credit historyof the borrower, sometimes referred to as “reputa-tional collateral,” enables an individual or a firm togain access to financing.

Credit bureaus also rely on monitoring andscreening of borrower behavior. Lenders shareinformation accumulated through their lendingoperations with a credit bureau, which then dis-seminates it to other credit providers.This allowsthem to better assess credit risks based on a givenborrower’s past payment behavior. Lenders, there-fore, can make better informed lending decisions.

1.2 Consumer versusCommercial ReportingThe private credit bureau industry is divided intotwo categories: consumer credit bureaus andcommercial credit bureaus (see Figure 1). Smallbusinesses can be covered by either side.

The business model of consumer credit report-ing consists of (a) receiving information for free,primarily from creditors and public sources andthen matching, cross-checking, and merging suchdata; (b) analyzing and interpreting the data; and(c) selling it back to the lenders. Historically, thismodel was applied to consumer reporting butnow increasingly bureaus also include informationon small-size loans to firms.

Commercial credit reporting, on the otherhand, relies less on reporting from lenders, andmore on company information available boththrough public sources and direct investigations,


6

FFiigguurree 11:: CClliieenntt BBaassee bbyy TTyyppee ooff CCrreeddiitt BBuurreeaauu//AAggeennccyy

Sources of Information for Private Credit Bureaus

Num

ber o

f P

riva

te C

redi

t Bur

eaus

FFiigguurree 22:: SSoouurrcceess ooff IInnffoorrmmaattiioonn ffoorr PPrriivvaattee CCrreeddiitt BBuurreeaauuss

The credit bureaus merge and cross-check thedata to produce a credit report for each individualborrower.This report constitutes a comprehensiveborrower profile and is then sold to lenders. Theseindividual credit reports generally contain person-al borrower information and information on bor-rower credit accounts.The personal section usual-ly captures the borrower’s name; identificationnumber, such as social security (if any); date ofbirth; former names; current and previous address-es; other forms of identification; employment his-tory; alerts, such as ID theft or security freezes; anddate of information update (see Figure 3). Thecredit summary section contains information onall credit accounts (both open and closed) that theborrower may have had, all accounts in goodstanding, past due accounts, negative account his-tory, and all inquiries made about the borrower forat least the past 12 months.

as well as payment behavior reported by suppliers.One of the best known commercial credit report-ing firms is Dun & Bradstreet. As mentioned earli-er,one has to distinguish this service from the serv-ices provided by credit rating agencies such asMoody’s, S&P, and Fitch.

1.2.1 Consumer CreditBureausConsumer credit bureaus collect information in astandardized format from several types of lenders,such as banks, credit card companies, retaillenders, other non-bank financial institutions, andutility companies. The World Bank survey, forexample, reported that roughly 60 percent of pri-vate credit bureaus included information fromretailers and merchants, and at least 43 percentincluded information from utility providers intheir databases (see Figure 2).2


7

2World Bank. 2005. Doing Business Database on Private Credit Bureaus.

Borrower credit history is often recorded interms of the number of missed payments providedin a format similar to the one in Figure 4.The cred-it report also provides information on collectionsmade on outstanding accounts and any availablepublic records, such as court judgments and bank-ruptcy rulings. The report usually outlines con-sumer rights and procedures for filing disputes.Finally, in many countries credit reports include acredit score.

FFiigguurree 44:: SSaammppllee HHiissttoorryy ooff PPaayymmeennttssReports are usually available electronically and

via the Internet, and large creditors have creditreports fed directly into their loan processing sys-

tems. Lenders pay the credit bureau for creditreports in the form of a subscription fee, a fee-per-query with significant volume discounts,or a com-bination of both.The price of a credit report mayrange from a few cents to US$5 or more depend-ing on the size of the lender, scope of information,and the country. Investigative reports prepared bycommercial reporting firms, on the other hand,have significantly more detail on businesses andcost anywhere between US$10-75 depending onthe size of the lender, scope of information, andthe country.

Historically, consumer credit bureaus only col-lected information on individuals. In recent years,however, with the expansion of small businesslending and advances in information technology(IT), more credit bureaus also include informationon small businesses. In a recent World Bank sur-vey4 approximately 76 percent of all private cred-it bureaus contained at least some information on


8



Types of Information

Per

cen

t o

f P

riva

te C

red

it B

ure

aus

FFiigguurree 33:: FFiirrmm--lleevveell IInnffoorrmmaattiioonn CCoolllleecctteedd bbyy PPrriivvaattee CCrreeddiitt BBuurreeaauuss33

firms. Collecting information on both individualsand firms in one bureau has an additional benefitof allowing the assessment of a business and itsowner to be combined. The credit history of abusiness owner is an important predictor of thecredit risk of a small business.Moreover,small busi-ness owners often mix personal and businessfinances, thus necessitating information on bothfor a more accurate assessment of risk.

1.2.2 Commercial CreditBureausCommercial credit bureaus provide informationon companies available through public sourcesand direct investigations and payment behaviorreported by suppliers. Commercial credit bureausreport on companies that are smaller in size andearnings than those corporations covered by rat-ing agencies. Commercial reporting may includevery small businesses although the information isoften limited because the reporting format is inap-propriate for small companies. In addition, the costof a report on a small (and micro) company is like-ly to be high in relation to the size of the loan. Forthis reason, small businesses are probably betterhandled within the framework of a consumer cred-it bureau.

The international leader in commercial creditreporting is Dun & Bradstreet, which traces itsroots back to the Mercantile Exchange, establishedin New York City in 1841.5 Formerly, Dun &Bradstreet delivered its reference books to sub-scribers under lock and key. Today, it transmitscredit information on more than 60 million busi-nesses worldwide.6 More recently, Coface, the sec-ond largest international credit risk insurer,entered the international market, building on itsdatabase of payment behavior of hundreds of thou-sands of medium-sized companies, which it builtthrough its credit risk insurance business. In mostcountries, local companies also operate creditinvestigation businesses. This Guide focuses onconsumers and small business credit bureaus.

1.3 Ownership StructuresGenerally the ownership structure of a creditbureau falls into one of the following categories:

� For-profit credit bureaus in which banks and/orother creditors are either majority or minorityshareholders.

� For-profit credit bureaus owned and operatedby a specialized firm with no ownership bycreditors.

� Not-for-profit credit bureaus, formed on thebasis of an association or chamber of com-merce, that mostly operate on membership feesof some kind.

A World Bank survey7 of 78 credit bureaus in 55countries around the world found that approxi-mately 46 private credit bureaus had no owner-ship by banks, financial institutions, or credit cardproviders; 22 were owned by banks, financial insti-tutions, or credit card providers, seven were heldby industry associations or chambers of com-merce; and only three were partially held by gov-ernments (see Figure 5).

FFiigguurree 55:: OOwwnneerrsshhiipp SSttrruuccttuurreess ooff PPrriivvaattee CCrreeddiitt BBuurreeaauuss


9

5Rowena Olegario. 2003. “Credit Reporting Agencies: A Historical Perspective,” inCredit Reporting Systems and the International Economy, ed. Margaret J. Miller.Boston: Massachusetts Institute of Technology.

6World Bank. 2004. Doing Business in 2004: Understanding Regulation.Washington, D.C.: World Bank, International Finance Corporation and OxfordUniversity Press.


Partial government ownership of credit bureausis rare, and only a few countries follow this model.For instance, in Sri Lanka, the Credit InformationBureau is a public-private partnership in whichthe central bank holds a 49 percent equity stake.8

The central bank, however, plans to eventuallydivest itself of its shares in the bureau. In otherexamples in this category, state-owned banksrather than government entities became majorityshareholders. In India for example, the State Bankof India and the Housing Development FinanceCorporation Ltd. (40 percent each) were majorityowners of CIBIL when it was first establishedalong with Dun & Bradstreet and TransUnion (10percent each).9 Over time, other banks joined asshareholders.

Independent credit bureaus with no ownershipby lenders, such as Equifax and TransUnion in theUnited States, are sometimes viewed as more effi-cient structures for operating a credit bureau.Credit reporting is the core business of such com-panies, and the shareholders’ main objective is tomaximize the value of the credit bureau byexpanding its operations and providing new serv-ices. Such an approach, however, is often not feasi-ble when establishing a credit bureau in countrieswhere credit bureaus do not yet exist. In mostcases, lenders are reluctant to share informationwith an independent credit bureau because of alack of trust.

A common solution, therefore, is to establish acredit bureau with ownership by banks and/orother lenders. Brazil, Germany, Italy, Romania,Turkey, and several other countries adopted thisapproach. The downside of this approach is thatlenders, even as shareholders of a credit bureau,may not always choose credit bureau growth as atop priority. For example, existing members maybe reluctant to allow new lenders to participate inthe bureau because newcomers, while unable tocontribute significant amounts of information,benefit greatly from information on existingclients.The fact that lenders own the bureau mayalso make them less likely to use the services of anindependent bureau, thus increasing barriers toentry in the credit information provider market. Incases when only a few banks are shareholders but

several other banks are members of the bureau, itis possible that shareholding banks may influencethe pricing policy in a manner that penalizes non-shareholder members. One way to mitigate suchproblems is to limit ownership of individuallenders in a credit bureau.

In summary, this approach has several down-sides when compared to the independent opera-tor option, but it is often the only feasibleapproach to gaining lenders’ trust. Recently, largeinternational credit bureaus have expanded intoemerging markets, and in many cases the newbureaus have the international bureau as well aslocal banks as shareholders as in Mexico, Russia,Kazakhstan, and several other countries. It is alsopossible that as lenders gain trust in the opera-tions of the bureau they may divest themselves oftheir ownership shares as was the case in HongKong (China) and the Dominican Republic.

The credit reporting business is characterizedby network externalities and economies of scalethat could potentially classify a credit bureau as anatural monopoly. On-going debate on the optimalnumber of credit bureaus in a market has not pro-duced any consensus thus far. On the one hand, asingle registry combining aggregated informationacross the entire system and including both bankand non-bank credit information would providelenders with the most complete set of informa-tion, including comprehensive inquiry informa-tion. On the other hand, the lack of competitioneliminates incentives for such a bureau to improvedata quality, provide value-added services, andlower prices.

Small markets are unlikely to support more thanone credit bureau,but many large countries have avery competitive credit information industry withtwo or more bureaus actively competing. Forexample, three major credit bureaus operate cur-


10

8World Bank. 2004. Credit Bureau Development in South Asia, Finance & PrivateSector Development, South Asia Region.. Washington, D.C.: Washington D.C.

9Credit Information Bureau (India) Ltd. (CIBIL) web site available at:http://www.cibil.com/web/promoters.htm

CCoommmmeerrcciiaall,, CCoommmmeerrcciiaall,, nnoo NNoonn--ccoommmmeerrcciiaall,,OOwwnneerrsshhiipp bbyy OOwwnneerrsshhiipp bbyy CCrreeddiittoorr

CCrreeddiittoorrss CCrreeddiittoorrss AAssssoocciiaattiioonn

PPrrooss � Often the only feasible way to � No conflicts of interest � The association provides establish a credit bureau and ensure exist in management. the cost-support.buy-in from lenders. � Commercial outlook

� Lender support implies strong ensures innovation and commitment and ensures bureau high-quality service.sustainability. � The bureau is open to

� Commercial outlook ensures broad market coverage.innovation and high-quality service.

CCoonnss � Conflicts of interest are possible, � Banks generally are � Limited incentives exist to where existing shareholders resist unwilling to share data innovate.the entry of new lenders to the without taking ownership � Usually service is of lower credit bureau or the introduction in a bureau. quality than in a for-profit of new services. � Capital is lacking. bureau.

� The decision-making process is slow � The decision-making as diverging views of large numbers process is slow.of shareholders need to be accommodated.

� Barriers to entry exist for new providers as well as new members.

� Government as shareholder creates conflict of interest between supervisory and shareholder functions.

EExxaammpplleess � CRIF (Italy) � Equifax (US) � Common in Latin America,� CIG (Iceland) � Experian (US, UK) where chambers of � SCHUFA (Germany) � TransUnion (US) commerce maintain lists � Serasa (Brazil) � CompuScan of bad debtors.� SIMAH (Kingdom of Saudi Arabia) (South Africa)

� Datacheck (Pakistan)

TTaabbllee 11:: CCoommppaarriissoonn ooff CCrreeddiitt BBuurreeaauu OOwwnneerrsshhiipp SSttrruuccttuurreess


11

rently in the U.S. market.This industry structure isthe result of consolidation in the financial servicesindustry over the past 15 to 20 years.While manybanks contribute data and obtain reports frommore than one credit bureau, the information con-tained in each report is not identical. To addressthis problem, another set of firms—aggregators—entered the market to provide comprehensiveaggregated reports. In South Africa, three creditbureaus all contain information from the same setof banks but compete on the quality of informa-tion and provision of value-added services. Othercountries with a competitive credit informationindustry include Italy, Chile, and Ecuador amongothers. On the other hand, Germany, Austria, andmost smaller European countries have only onemajor credit bureau.

Commercial credit bureaus are usually moreprevalent than non-commercial credit bureaus.Experience with non-commercial bureaus indi-cates that they are generally less innovative, lackthe ability to deliver top-quality services, and tendto be bogged down with bureaucratic procedures(see Table 1).While lenders’ associations or cham-bers of commerce may be a good platform tobegin the discussions on the need for sharinginformation and to build the consensus among thepotential bureau members, the bureau itselfshould be a commercial entity. See Chapter 2,Developing Credit Bureaus in Emerging Markets,for a more detailed discussion of credit bureaubusiness models.

Commercial, with

1.4 Types and Scope ofCollected InformationCredit history information can be broadly dividedinto two categories:

� NNeeggaattiivvee iinnffoorrmmaattiioonn:: credit history only con-tains information on defaults. The informationmay include amounts outstanding at default andthe date of last payment. When the debt isrepaid, information on delinquencies is deletedfrom the database.These types of databases arealso often referred to as black lists. Among allconsumer credit bureaus, 32 percent providenegative only information.

� PPoossiittiivvee ((aanndd nneeggaattiivvee)) oorr ffuullll--ffiillee iinnffoorrmmaattiioonn::credit history contains information on all openand closed credit accounts, including theamount approved, as well as the information onrepayment. If a borrower has defaulted on pay-ments, but eventually paid it off, the defaultinformation remains on file and is not deletedfor a defined period of time. Among all con-sumer credit bureaus, 68 percent provide bothnegative and positive information.

A report that includes positive informationallows the lender to more accurately assess thecreditworthiness of a borrower. A database withnegative-only information, excludes high-risk bor-rowers that have accumulated significant debtexposure without yet defaulting on any loans. Insuch instances, even a small shock to the borrow-er’s income could lead to cascading defaults on allof the accounts.

In recent years, Hong Kong (China) and SouthKorea have experienced a period of majorincrease in retail credit defaults, in an unfortunatecombination of reckless lending and unavailabilityof positive information. While both had negativeinformation registries, positive information wasnot shared. As competition in the credit card mar-ket increased and banks marketed credit cardsmore aggressively, many consumers accumulatedseveral credit cards. Borrowers would typicallyopen one credit card account and then anotherone to pay off the debt accumulated on the first

one.This borrowing was unsustainable and result-ed in a large number of credit card defaults. Forexample, in Hong Kong (China), where banksshared only negative credit information, on aver-age debtors who went bankrupt owed 42 timestheir monthly income in unsecured debt whereasdebtors in the United States,where banks suppliedall available information to the bureau, had unse-cured debts 21 times their monthly income.10

Since the credit card crisis, Hong Kong’s (China)bureaus have now migrated to a positive and neg-ative credit reporting system.

In South Korea between 1998 and 2003, creditcard issuance rose from a little over 40 millioncredit cards to approximately 90 million cards.Between 2002 and 2003 alone, delinquency ratioson credit cards rose from 12.8 percent to 43.3 per-cent. Following the crash of the credit card mar-ket, credit delinquents accounted for a staggering16.7 percent of the economically active popula-tion at the end of 2003. Personal delinquency wasa serious social issue because delinquents wereconcentrated among youth, women, and lowerincome population—the most fragile sector of theeconomy. This crisis could have been avoided ifthe lenders had had access to a credit bureau thatprovided positive and negative information, whichwould have helped them be more conservative intheir lending practices.

According to the World Bank survey, approxi-mately 68 percent of all private credit bureaus pro-vided both positive and negative information onindividuals, and roughly 50 percent provided bothpositive and negative information on firms.11 Thefact that several countries, including such finan-cially advanced markets as Brazil, do not have pos-itive information poses a significant threat to thesustainable growth of retail and SME credit inthese countries.


12

10Bailey, A., Suzi Chun and Jeffrey Wong. August 11, 2003. “Wanted: Asian cred-it bureaus.” The Bangkok Post. Available at: http://www.bangkokpost.net/mckinsey/McKinsey110803.html.


default rate for Argentine banks and a 45 percentdecrease in default rates for Brazilian banks (seeFigure 7). Having positive information improvesthe ability of lenders to separate good borrowersfrom bad ones and thus reduces their costs associ-ated with defaults. For a bank with a US$100 mil-lion loan portfolio, this translates into average sav-ings of US$830,000 in Argentina and US$1.5 mil-lion in Brazil.

FFiigguurree 77:: EEffffeeccttss oonn DDeeffaauulltt RRaatteess ooff IInncclluuddiinngg PPoossiittiivveeIInnffoorrmmaattiioonn ((AArrggeennttiinnaa aanndd BBrraazziill))

Including positive information also allowsbanks and other lenders to lend more or supplymore credit while keeping default rates at thesame level.Using the simulations with the U.S.datamentioned above and keeping the target defaultrate at three percent, the inclusion of positiveinformation almost doubles the percentage of bor-rowers approved from approximately 40 percentto 75 percent, indicating the importance of posi-tive information for improving access to credit(see Figure 8).14

Recent research studies have quantified theimpact of positive information on default and cred-it approval rates. One study simulated default rateson loans resulting from lending decisions, using acredit scoring model with only negative informa-tion and one with both negative and positive infor-mation.The simulations were based on data in oneof the largest U.S. credit bureaus.12

Many lenders use credit bureau information togenerate credit scores, which are statistical esti-mates of the probability of default of a borrowerbased on characteristics available in the creditbureau. Higher credit scores indicate higherexpected probability of repayment and can beused to set credit approval rules and procedures.The study generated credit scores using negativeonly and then both negative and positive informa-tion. Borrowers were then ranked by credit scoresand those with the highest scores (60 percent)approved for credit. According to the study, theresulting default rate from lending to borrowersbased solely on negative information was 3.35 per-cent. If,both positive and negative information hadbeen used, the default rate would have dropped to1.9 percent, a 43 percent decrease in default rates(see Figure 6).

FFiigguurree 66:: EEffffeecctt oonn DDeeffaauulltt RRaatteess ooff IInncclluuddiinngg PPoossiittiivveeIInnffoorrmmaattiioonn ((ssiimmuullaattiioonn uussiinngg UU..SS.. ddaattaa))

Another study13 conducted the same exerciseusing data from Brazil and Argentina and foundsimilar results. Inclusion of positive informationwould have produced a 22 percent decrease in the


13

12Barron, J.M. and Michael Staten. 2003. The Value of Comprehensive CreditReports: Lessons from the US Experience. Available at: http://www.privacyal-liance.org/resources/staten.pdf. Figures show the simulated credit defaultsassuming an acceptance rate of 60 percent.

13Powell, A., Nataliya Mylenko, Margaret Miller and Giovanni Majnoni. November2004. Improving Credit Information, Bank Regulation and Supervision: On theRole and Design of Public Credit Registries. World Bank Policy Research WorkingPaper 3443. Washington, D.C.: World Bank.

14Barron et. al., op cit.

Information sharing brings benefits to bothsmall and large institutions.The study, using infor-mation from Argentina,16 found that while smalllenders do benefit more than large lenders fromsharing information, large banks still experience asignificant drop in defaults if positive informationis used.Although the results may vary from coun-try to country and from lender to lender, bothanecdotal and available empirical evidence sug-gests that information sharing and use of creditscoring allow both large and small banks to signif-icantly reduce default rates and/or increase lend-ing volumes (see Figure 10).

FFiigguurree 1100:: EEffffeecctt oonn DDeeffaauulltt RRaatteess ooff IInnccrreeaassiinngg NNuummbbeerr ooffIInnffoorrmmaattiioonn SSoouurrcceess

In summary, credit bureau reports that have thehighest predictive power combine both positiveand negative information from both banks andnon-bank lenders. Bureaus or credit registries frag-mented by industry that provide negative informa-tion only deliver reports that have less predictivepower and often result in inaccurate credit riskassessment (see Figure 11).

FFiigguurree 88:: EEffffeecctt oonn AApppprroovvaallss ooff IInncclluuddiinngg PPoossiittiivvee IInnffoorrmmaattiioonn

Credit bureaus often include information notonly from traditional lenders, such as banks andcredit card companies, but also from other creditproviders, such as retailers, suppliers, and telecomand utility companies. In emerging markets whereinformation is scarce, such non-traditional sourcescan provide invaluable information. The studymentioned above15 found that in the United States,including information from non-bank lenders intoa credit scoring model allows lenders with a targetapproval rate of 60 percent to reduce default ratesby 38 percent. If the default rate is used as a target,the bank would be able to approve 11 percentmore clients before reaching the target three per-cent default rate (see Figure 9).Overall, the simula-tions show that sharing positive informationamong the broader category of lenders wouldallow significant operational improvements eitherthrough the lower costs of default or increasedlending volumes to new categories of borrowers.

FFiigguurree 99:: EEffffeecctt ooff IInncclluuddiinngg PPoossiittiivvee IInnffoorrmmaattiioonn oonn AApppprroovvaallssAAmmoonngg RReettaaiilleerrss aanndd OOtthheerr LLeennddeerrss


14

15Ibid.

16Powell, A. et al., op. cit.

FFiigguurree 1111:: EEffffeecctt ooff TTyyppeess aanndd SSoouurrcceess ooff IInnffoorrmmaattiioonn oonnPPrreeddiiccttiivvee PPoowweerr

1.5 Industry Overview andInternational TrendsThe private credit bureau industry has experiencedunprecedented growth over the past five years,especially in emerging markets (see Figure 12).

FFiigguurree 1122:: GGrroowwtthh ooff PPrriivvaattee CCrreeddiitt BBuurreeaauuss

Annex 1 provides maps showing private creditbureaus by region and by type of reporting (posi-tive and negative-only).

This growth was driven by two main factors:

� HHiigghh ggrroowwtthh ooff rreettaaiill ccrreeddiitt iinn eemmeerrggiinngg mmaarr--kkeettss.. Between 1985 and 1995, unfavorablemacroeconomic environments and structuralrestrictions in credit markets in developingcountries constrained credit growth. Over thisperiod, the private credit to GDP ratio for thedeveloping markets increased by 46 percent.17

Financial liberalization and a more stable macro-economic environment were associated with anincrease in credit growth, and the periodbetween 1996 and 2004 saw credit to the pri-vate sector increase by 62 percent.18 As lendersbegan to enter the retail credit market, the needfor credit information and for streamlining lend-ing processes resulted in the establishment ofcredit bureaus. This phenomenon was particu-larly pronounced in Eastern Europe, where thenumber of private credit bureaus grew from fivein 2002 to 12 in 2006.

� DDeevveellooppmmeennttss iinn iinnffoorrmmaattiioonn tteecchhnnoollooggyy.. Thecredit bureau industry is data-driven. Recentimprovements in database management softwareand decreasing costs of storing and processingdata,as well as decreasing costs of hardware,havereduced the start-up costs of a credit bureau.

According to the World Bank’s report, DoingBusiness in 2006, approximately 67 countries hada private credit bureau operating at the end of2005. Among developing countries, the LatinAmerica and the Caribbean region is the mostadvanced, where 16 out of 22 countries had a pri-vate credit bureau in operation.These bureaus alsocovered a higher percentage of the adult popula-


15

17World Bank’s World Development Indicators as of July 2006. Data based ondomestic credit to private sector as a percent of GDP for low-income, lower-mid-dle income, and upper-middle income countries. Includes 101 countries.

18Ibid. Includes 128 countries.

tion (31.2 percent) compared to all other regions(see Figure 13). All private bureaus in the regionprovided data electronically either via the Internetor via modem and phone lines.19 Bureaus in mostcountries in Latin America include both positiveand negative information with a few exceptions. InBrazil, for example, the credit bureau provides neg-ative-only information. Although Brazil’s creditbureau system is one of the oldest in the worldand highly sophisticated, it is unable to providelenders with positive information. Reforms areunder discussion that would provide the neces-sary legal and regulatory framework for sharingpositive information among private creditbureaus.The development of a credit bureau thatprovides positive information in Brazil, however,will require not only legal changes but also a broadbuy-in from lenders and a willingness to share pos-itive information.

FFiigguurree 1133:: AAvveerraaggee PPrriivvaattee BBuurreeaauu CCoovveerraaggee

The Middle East and Africa have the least devel-oped credit information infrastructure.The surveyreported that only five out of a total of 37 Sub-Saharan African countries had any private bureaucoverage. In countries where credit bureaus oper-ate, they mostly provide negative-only informationwith a few exceptions, such as the Kingdom ofSaudi Arabia (KSA) and South Africa.

In Asia and Eastern Europe,many countries haverecently established bureaus but experienceddelays in populating databases and issuing reports,which in part explains the low coverage ratio. InEastern Europe, the situation is changing fast andeven as this publication goes to print Russia andUkraine are in the process of operationalizingtheir credit bureaus and plan to issue creditreports by the end of 2006.

Another factor contributing to the low coverageratios in these regions is that the credit-active pop-ulation still only constitutes a small portion of thetotal population. As credit growth continues,how-ever, the scope of credit bureau coverage isexpected to expand as well.

In addition to the coverage ratio, the WorldBank’s Doing Business Survey also computes acredit information index for more than 120 coun-tries.The index measures credit information avail-ability based on six key factors that measurescope, access, and quality of available credit infor-mation (see Figure 14).


16

Perce nt o f Adu lts



17

FFiigguurree 1144:: CCrreeddiitt IInnffoorrmmaattiioonn IInnddeexx

The index attributes scores to countries andregions on a scale of 1 to 6, with 6 being the high-est. For each of the following six features, a coun-try or region receives one point and the points areadded to arrive at the total score:

� Both positive and negative credit information(for example on payment history, number andkind of accounts, number and frequency of latepayments, and any collections or bankruptcies)is distributed.

� Data on both firms and individuals are distrib-uted.

� Data from retailers, trade creditors, and/or utili-ties as well as financial institutions are distrib-uted.

� More than five years of historical data is pre-served.

� Data on loans above one percent of income percapita are distributed.

� By law, consumers have the right to access theirdata.

Using this indicator, the Latin American regionagain scores higher than all other regions with anaverage of 4.5. The Sub-Saharan African regionscores the lowest on this scale with a 1.5, followedclosely by the South Asia and East Asia & Pacificregions with 1.8 each. While bureaus in East Asiahave a higher coverage ratio, their collection ofnegative-only information contributes to their lowscore on the Credit Information Index.

The development of credit bureaus in manyemerging markets often means engaging one ofthe main international credit bureau operators.Asa result, several major players dominate the creditinformation industry globally. The three majorinternational credit bureau providers today areExperian, TransUnion, and Equifax. Their mainoperations are concentrated in the high-incomeOrganization of Economic Co-operation andDevelopment (OECD) countries,but all three haveactively expanded into emerging markets inrecent years. At least one of these three major

Cre

dit

In

form

atio

n I

nd

ex

lowed the French example and established PCRsafter the creation of the West African MonetaryUnion in 1962. Several Middle Eastern and NorthAfrican nations adopted PCRs in the 1950s and1960s: Egypt (1957), Tunisia (1958), Morocco(1966), and Jordan (1966). Among the EasternEuropean and Central Asian republics, Turkey wasthe first country to establish a PCR in 1951.

The Latin America region has the highest num-ber of PCRs. In this region, 15 countries haveestablished PCRs, including all of the largesteconomies such as Argentina, Brazil, Chile, andColombia. Mexico established Latin America’s firstPCR in 1964,with the primary aim to regulate poli-cies regarding the allocation of credit by sector atthat time. A few Latin American countries estab-lished PCRs in the 1970s and 1980s, but mostPCRs were set up in the 1990s, following the finan-cial crises that swept the region.

Since most public registries are set up with theprimary purpose of bank supervision, they usuallyonly include loans above a certain minimumamount. Of the 57 PCRs included in the WorldBank survey, 40 had minimum loan requirementsfor inclusion of a company or individual in thePCR. About 60 percent of PCRs had loan cutoffsthat were at least twice the average Gross NationalIncome (GNI) per capita, thus excluding mostretail and small business loans.

PCRs have lower coverage in general whencompared with private credit bureaus, which isnot surprising given their focus on larger loans andsupervised institutions. According to the WorldBank survey, the average coverage ratio of PCRs indeveloping countries was at 3.6 percent of theactive population, while private bureau coveragewas at 16.0 percent.21


18

international credit bureaus operates in 16 coun-tries in the Latin American & Caribbean region.The “big three” also have operations in six coun-tries in Sub-Saharan Africa, five in East Asia andSouth Asia, and four in the Europe & Central Asia.20

In recent years, several new credit bureaus withinternational operations have emerged includingCRIF, an Italian-based firm with a focus on theEastern European markets; Creditinfo, an Icelandiccredit information provider with operations inEastern European and Central Asian countries;D&B SAME with a target market in the Middle Eastand Africa; CompuScan, CRB, and XDS in two ormore African countries; and Baycorp Advantagebased in Australia and New Zealand with a previ-ous focus on Asian markets.The entry of the newoperators is a welcome development as morecompetition will result in a better product offeringand lower prices for the countries shopping for acredit bureau provider.

1.6 Public Credit RegistriesA public credit registry is defined as a databasemanaged by the public sector, usually by the cen-tral bank or the bank supervisor, that collects infor-mation on the creditworthiness of borrowers (per-sons or businesses) from supervised financial insti-tutions, makes such information available to finan-cial institutions, and is used primarily for supervi-sory purposes.

Participation in a Public Credit Registry (PCR) ismandatory for supervised financial institutions.Although the primary reason for establishing a PCRis to support banking supervision functions andmonitor systemic risks, many PCRs provide creditreports to lenders as part of their operations.

According to a World Bank survey in 2004, therewere 57 PCRs operating globally.Figure 15 illustratesthe growth in PCRs from before 1964 to 2002.

The first PCRs were established in WesternEurope—in Germany in 1934 and then in France in1946.By the mid-1960s, three other European coun-tries—Italy,Spain and Belgium—had also establishedPCRs. Former French colonies in Western Africa fol-

20Information drawn from web sites of Experian, Equifax, and TransUnion availableat www.experian.com, www.equifax.com and www.transunion.com, respectively.

21World Bank. Doing Business web site. Available at:http://www.doingbusiness.org/ExploreTopics/GettingCredit.

FFiigguurree 1155:: GGrroowwtthh ooff PPuubblliicc CCrreeddiitt RReeggiissttrriieess2222

Unlike private credit bureaus, few PCRs distrib-ute data electronically. Only 60 percent of PCRs inLatin America and 30 percent in Sub-Saharan Africadistribute data electronically. In both South Asiaand Sub-Saharan Africa, almost all PCRs providepaper copies of credit reports. The reliance ofPCRs on non-electronic data distribution compro-mises the quality of data and speed with whichsuch data are made available to PCR clients.Furthermore, non-electronic forms of data are lesslikely to be updated frequently and are, therefore,less likely to be accurate or current.


19

In countries with no cutoff requirements (i.e.,minimum loan requirements), PCRs provide creditreports to banks and perform functions similar tothose of private credit bureaus.The scope of PCRdata, however, is limited to fewer data items andexcludes information from non-banks. Figure 16indicates that PCRs rely largely on banks and finan-cial corporations for information. Less than fourpercent of all PCRs included information fromretailers, merchants, and other non-bank creditors.

PCRs usually provide their credit reports at lowor no cost to the lenders. Of the 57 PCRs surveyedby the World Bank,23 only four PCRs had a sub-scription fee. Owing to the limited availableresources and non-profit basis of operation, publicregistries often lack the necessary motivation andfunds to invest in data quality assurance systemsand provide value-added services, such as creditscoring and portfolio monitoring.

Development of a public registry with theobjective of serving the credit information needsof retail lenders might appear effective in fillingthe gap of available credit information by mandat-ing bank participation in a PCR. In the medium tolong term, however, establishing a PCR may under-mine private initiatives and thus hurt creditorswho will not have access to a broad set of creditdata, including the information from non-banksthat is usually collected by private bureaus.

22World Bank. 2004. Doing Business Database on Public Credit Registries.

23World Bank 2004. Doing Business Database on Public Credit Registries. Information was unavailable for two of the 57 PCRs surveyed.

24World Bank. 2004. Doing Business Database on Public Credit Registries.

FFiigguurree 1166:: SSoouurrcceess ooff IInnffoorrmmaattiioonn ffoorr PPCCRRss2244

Moreover, creditors who rely on public registrieswill be unable to benefit from value-added servic-es, such as bureau credit scores, which allow themto maximize the value of credit information.

Not all countries have a strong free market tra-dition, and in many cases the involvement of thepublic sector is essential for the establishment of aviable private credit information industry.The gov-ernment plays a critical role in creating an appro-priate legal and regulatory framework for privatecredit bureaus. Recently several countries, includ-ing Russia, Singapore, Kazakhstan, and Mexico,issued laws and regulations encouraging financial

institutions to share information through privatecredit bureaus. Another approach, successfullyimplemented in Ecuador, requires that all regulat-ed financial institutions report credit data to thecentral bank registry. The PCR, however, is notallowed to sell the credit reports to lenders;instead they must make the reports available tothe private credit bureaus that are operating in thecountry. Private bureaus combine this informationwith other data and compete on the quality ofservice.These examples demonstrate the success-ful collaboration of the public and private sectorto develop a sustainable credit reporting industry.


20

2.1 Prerequisites for CreditBureau DevelopmentTo establish a successful credit bureau in anemerging market, the founders must:

� Change perceptions of and build awareness inthe community.

� Ensure commercial viability.� Establish an appropriate legal and regulatory

framework.� Identify appropriate technology and software.� Ensure adequate data availability.� Specify staffing needs.

Change Perceptions and BuildAwareness

In emerging markets, the public is likely to resistthe concept of sharing information, particularlyfinancial information. For political reasons, author-ities unfamiliar or uncomfortable with sharingfinancial information may also be resistant to thisconcept. In such economies fragmentation, com-petition,and secrecy characterize the financial sec-tor, and lenders fear that by sharing positive infor-mation, their competitors will learn about andsteal their good customers. These perceptionsneed to be changed.

For a credit bureau to flourish and gain accept-ance among its users and subscribers in theseeconomies where credit is scarce, the public needsto be educated about the benefits of credit andcredit sharing, lenders must overcome their dis-

trust of their competitors, and consumers must beassured of the security of their information.

Consequently, the initial phase of building acredit bureau where such an institution does notexist should focus on building awareness amonglenders and their clients, the general public, gov-ernment officials, policy makers, regulators, andother potential participants in the credit bureau.

Tools that can be used to change perceptionsand build awareness of the benefits of informationsharing include the following:

� MMeeddiiaa.. Coverage of conferences and roundta-bles as well as publication of articles on the roleof credit information with expert opinions andreflections on the local debate is useful to pro-mote credit bureau development. For example,several conferences on the role of credit infor-mation were held in Russia and were well cov-ered in the press. As a result, awareness amongthe public on the need to build a credit historyand consent to submitting one’s credit recordsto a credit bureau was improved significantly.

� IInntteerrnneett.. A credit bureau’s web site should con-tain information for consumers that explains theimportance of building a positive credit history.The site must also explain the various channelsavailable to consumers to report and rectifyerrors in their credit reports.

� PPrrooaaccttiivvee ccoonnssuummeerr eedduuccaattiioonn.. The CreditBureau of Singapore offers one example of thistype of consumer education. Its senior execu-tives give presentations to diverse groups of bor-

2Developing CreditBureaus in Emerging

Markets

Developing Credit Bureaus in Emerging Markets

21

rowers. These meetings have succeeded inbroadening understanding of the activities ofthe bureau and explaining consumers’ rights.

� RRoouunnddttaabblleess aanndd ccoonnffeerreenncceess.. Several examplesexist of the efficacy of this type of awarenessbuilding. In Kazakhstan, for example,a large con-ference was organized with high-level participa-tion from the central bank, lenders, and speak-ers from several successful credit bureaus in theregion. The conference was used to jumpstartthe initial phases in establishing a credit bureauand to secure high level buy-in.As a result, nec-essary legal changes and consensus buildingamong lenders took place smoothly and withina reasonable time frame. Similar events, inVietnam, Russia, Kenya, and several other coun-tries were instrumental in promoting the estab-lishment of credit bureaus.

Issues to be addressed in these awareness build-ing activities differ according to the audience.

Campaigns targeted to the ggeenneerraall ppuubblliiccshould:

� Explain the role and nature of credit informa-tion-sharing to mitigate general concerns aboutsharing of personal information.

� Inform about the function of a credit bureauand its obligations to respect the privacy of con-sumer data.

� Discuss the obligations of a credit bureau to pro-tect all personal information and its duty to treatall such information as confidential.

� Discuss the measures that a credit bureau willtake to ensure the security of consumer dataand the way mistakes are to be corrected.

� Discuss how authorities and regulators willwork in conjunction with the credit bureau tocreate an environment conducive for sharingcredit information on a secure basis.

� Emphasize the importance of borrower consentto enable data sharing.

In addition to educating the public about creditbureaus, the campaign must educate the public onusing credit responsibly. In markets, where mem-bers of the public are not very credit savvy, the

public needs to understand the dangers of buyinginto various lender schemes and being overlyindebted.

Campaigns targeted to ffiinnaanncciiaall aanndd nnoonn--ffiinnaann--cciiaall iinnssttiittuuttiioonnss should:

� Address concerns about sharing of informationand the fear of losing market share due to suchinformation-sharing.

� Explain the different measures that could beenforced to prevent competitor institutionsfrom poaching customers.

� Emphasize the need for cooperation among acountry’s banking, financial, and non-financialinstitutions for the credit bureau to succeed.

� Assure lenders of the confidentiality of all information provided and discuss the obliga-tions of lenders to treat confidential informationappropriately.

� Explain the importance of sharing positiveinformation.

� Encourage broad participation by bank and non-bank lenders in the credit bureau.

� Promote the introduction of updated credit con-trol policies and procedures taking into accountthe information in the credit bureau.

� Highlight the need to educate staff about thecredit bureau.

For the target audience of ggoovveerrnnmmeenntt ooffffiicciiaallss,,ppoolliiccyy mmaakkeerrss,, aanndd rreegguullaattoorrss,, the campaignshould:

� Hold high-level meetings among various govern-ment officials, policy makers, and regulators toexplain the importance of a credit bureau andthe sharing of credit information.

� Explain their role and the need for laws con-ducive to credit information- sharing.

� Stress the importance of information sharing forfinancial stability and expansion of credit.

� Outline the value of obtaining access to thirdparty databases, particularly official registerssuch as the National ID or Company Registra-tion databases.


22

in the form of credit reports. In a second phase,the bureaus then leverage the value of the databy offering value-added products, such as creditscoring, portfolio monitoring, fraud alerts, andthe like.

� AApppprroopprriiaattee pprriicciinngg aanndd iinnvveessttmmeenntt ddeecciissiioonnss..The bureau must conduct market assessmentstudies and forecast demand that will enable itto price its reports accordingly. Pricing is one ofthe key factors in bureau sustainability, and cru-cial investment decisions such as softwareacquisitions should be aligned with the bureau’spricing strategy to avoid heavy losses.

� SShhaarreehhoollddeerr ccoommmmiittmmeenntt.. Establishing a creditbureau, especially bank-owned bureaus, takes along time and requires strong shareholder com-mitment. Several examples exist of creditbureaus that have struggled commercially dueto the lack of “total commitment”from the finan-cial community. Some bureaus struggle for yearstrying to build up membership. Local bankingcommunities, notwithstanding their initial com-mitment to support the credit bureau, often failto adhere to their commitment in the bureau’searly days.Other types of lenders may hold backuntil the banks join, thus adding to the problem.Ensuring strong commitment of the key share-holders is essential because a bureau may notbreakeven for three to five years, and it can sus-tain heavy losses in the early years.

� SSttrreeaammlliinneedd oorrggaanniizzaattiioonnaall ssttrruuccttuurree.. In its ini-tial stages, a bureau must pay particular atten-tion to its organizational structure and its humanand capital investments (see Specify StaffingNeeds below),and adopt a streamlined structurefrom the start to maintain profitability.

Establish an Appropriate Legaland Regulatory Framework

Without the appropriate legislation and regula-tions to enable the sharing of data and informa-tion, a credit bureau is, for all practical purposes,ineffective. For instance, in Uzbekistan, a creditbureau that was registered in 2000 was unable tooperate because it lacked the legal authority toshare data. In Slovakia,existence of laws restrictingthe collection of historical data delayed the

� Highlight the need for appropriate privacy anddata protection legislation.

Ensure Commercial Viability

A credit bureau needs to be commercially viable,regardless of the market in which it operates, andemerging markets are no exception to this rule. Inaddition to ensuring the sustainability and successof the credit bureau, commercial viability rein-forces the faith and changed perceptions of thepublic and all those involved in the operations ofthe credit bureau.

A commercially viable credit bureau requiresthe following:

� SSiizzeeaabbllee eeccoonnoommiiccaallllyy aaccttiivvee ppooppuullaattiioonn..Economies of scale underpinning the operationsof a credit bureau require a sufficiently largepopulation base that uses credit for a creditbureau to be commercially viable.The size of theeconomically active population dictates the levelof sophistication and complexity of the creditbureau system to be implemented. Without alarge borrower base, the bureau will have tocharge high fees for its credit reports,which mayreduce the demand from lenders. Credit bureausin emerging markets might face this challenge intheir initial years of operation if the populationdoes not use credit heavily. Developed countrieswith small populations, such as Iceland (popula-tion 300,000) and New Zealand (population 4million), operate small but profitable creditbureaus. In New Zealand, where the economical-ly active population is estimated at two million,the credit bureau receives about four and a halfmillion queries a year.

In emerging markets,however, the economicallyactive population may be too small to generatesufficient demand from lenders. In this case, thebureau may consider a regional solution.TransUnion Central America (TUCA), for exam-ple, operates a regional solution covering fivecountries.This approach lowers the fixed cost ofstarting a bureau and is particularly attractivewhen lenders also operate regionally.

� PPrroovviissiioonn ooff vvaalluuee--aaddddeedd sseerrvviicceess.. Most bureausbegin operations by providing lender raw data


23

bureau’s launch. Many other countries faced simi-lar challenges.

For credit bureaus in emerging markets, theenabling legislation must:

� Establish the rights and obligations of the creditbureau, its users, the organizations that supplyinformation to the bureau,and of public at large.

� Treat all stakeholders equally and favor no onestakeholder over others. The law should allowinformation-sharing among bank and non-banklenders.

� Provide clear guidelines on the kinds of datathat can be collected. Data accumulation can besimplified if the legislation does not restrict thetype of information that a credit bureau cangather as long as it is relevant, accurate, andmeets certain quality standards (for example, asin the U.S.Fair Credit Reporting Act and EU DataProtection legislation).

� Indicate how long data can be retained. In par-ticular, it must address the need to store andshare positive information, whereby all informa-tion on a particular client is saved, instead oferasing negative records once debts have beenrepaid, which significantly undermines the pre-dictive power of data. Historical data are greatindicators of a borrower’s future behavior and isinvaluable in making lending decisions.

A more detailed discussion of legal and regula-tory issues is provided in Chapter 4.

Identify Appropriate Technologyand Software The overriding duty of both a credit bureau and itsmembers is to maintain absolute security over sen-sitive personal information and to treat it appro-priately at all times.Failure to maintain this corner-stone function is a major breach of trust.As a trust-ed third party, a credit bureau must have the mostappropriate technology and software to providean adequate level of service while maintaining thesecurity of the data.

To function efficiently, a credit bureau mustinvest in appropriate software and technical sys-

tems.Technologies available in emerging marketsmay not be sufficiently sophisticated to meet mod-ern credit information-sharing needs. Some of theinfrastructure requirements are quite basic and aregenerally taken for granted in more sophisticatedmarkets. In many developing countries, however,essential infrastructure support, such as a reliableelectricity supply and adequate Internet and tele-phone connectivity, do not always exist and mayrequire additional investments.

In addition to infrastructure requirements, thesuccess of a bureau’s operations depends on theability to extract credit performance data fromfinancial institutions and other lenders and delivercredit reports in a format that is easy to use by oneand all. Experience shows that financial institu-tions in emerging markets are transitioning fromtraditional relationship lending models to morevolume-driven and standardized retail lendingbusinesses. In most cases, banks have not yet reor-ganized their processes to fit this new businessmodel.This has major implications in terms of:

� Availability and validity of historical information.

� Ability to automatically report data on assetportfolios.

� Flexibility of legacy systems to adapt to theinput/output requirements of a credit bureau.

� Availability/prioritization of IT resources thatcould be channeled to support adaptationsrequired for the new credit bureau.

� Existence of standardized formats or languageused within the systems.

In many cases,banks do not have adequate coresystems to store data on customers and their pay-ment histories. In countries as diverse as Russia,India, and Egypt, extracting data in a formatacceptable to the respective credit bureau was amajor challenge and required substantial ITresources to access old legacy systems. It hasproved much easier to extract credit card records,as these tend to be hosted on systems that aremore modern and store data in a much more logi-cal format. Large, local, often state-owned orrecently privatized banks with large branch net-works face a major challenge because records maybe paper-based and credit functions decentralized.


24

For bureaus in such situations, the most practicalsolution is often to start with credit portfolios thathave better quality information, with banks thatare able to provide such information, and withmore recent data, then gradually build up thebureau to include more lenders and more portfo-lio types. Often, however, the cost of uploadingcredit history data that are two or more years oldmay be too high compared with the benefit, giventhe varying levels of accuracy of lender files.

In such environments, credit bureaus functionless efficiently than their Western counterpartsand offer fewer products and services to their cus-tomers. For example, in Fiji, with a population ofonly one million, the bureau provides only nega-tive data on borrowers because banks and otherfinancial institutions are unwilling to commit timeand resources to provide comprehensive positivedata to the credit bureau, due to the small marketsize and lack of market sophistication. Conse-quently, the Fiji bureau is limited to confirmingwhether an individual is on the local body elec-toral roll and recording any previous applicationsfor credit in the credit report.

Ensure Adequate Data Availability

It is absolutely essential that banks and otherlenders store customer information in a formatthat allows the credit bureau to easily extract theinformation and upload it to its own system to further match and merge with other data. In mostemerging markets, financial institutions had developed their own unique database structureswell before a credit bureau was ever contemplated and, as a consequence, do not share standard data formats.

Identifying information, such as a unique IDnumber, name, address, and date of birth, is essen-tial to enable the merging of credit records for agiven borrower to form a complete credit historyof that borrower.

Countries that have a national ID system enjoyan advantage over those that do not, but only ifthe IDs are correctly recorded within databasesand can be verified against the registry

maintained by the government agency that issuesthem. Foreign residents that are not eligible for anational ID, can use a passport number insteadfor identification but this approach poses challenges because passports must be renewedon a regular basis and, as a consequence, the passports numbers change.

National IDs can cause issues if they are record-ed incorrectly. For example, in Malaysia, whenBank Negara, the Central Bank of Malaysia, estab-lished a credit bureau in the early 1990s, theassumption was that all Malaysian nationals had aunique national ID number. After all borrowerrecords had been centralized within one database,it was discovered that duplicate IDs existed for thesame person because previously various stateprovinces had held the power to issue IDs.Malaysia is not alone in the challenges that it facesin establishing unique identifiers. Many countriesencounter similar problems. These issues can beovercome if authorities set digital protocols thatcan be used to validate the numeric configurationof an ID. In the Kingdom of Saudi Arabia (KSA), forexample, a joint effort among lenders, borrowers,and the authorities in this regard has been particularly successful. It is important to note thatin all cases the establishment of the bureau helpedauthorities and lenders to identify the problemand start to solve it.

In jurisdictions without national IDs that areverifiable and suitable to use in the match-and-merge process, most bureaus use sophisticatedmatching algorithms to correctly match and mergeseparate records belonging to the same individual.These matching algorithms traditionally use a com-bination of name, address, and date of birth. NewZealand and Germany, for example, use suchsophisticated matching solutions because legisla-tion actually prevents the recording of uniqueidentifiers. The ability to use such algorithms is significantly restricted, however, in emerging markets where other crucial information, besidesunique IDs, such as names, addresses, and dates ofbirth are unreliable.

In addition to the issue of unique identifiers,consumers’ names and addresses may be stored in


25

various formats and styles. For instance, an institu-tion may choose to store data in single stringsrather than in a structured database,some with thesurname first, others with only initials, and somewith only the surname.Consumers also contributeto the confusion by using different names, particu-larly where names may be anglicized or a nick-name used (e.g., William may be known as Bill,and/ or Will).

Addresses are equally challenging. In the KSA,for example, residential addresses are not used andseveral unrelated families may share a single postbox. And in India, many banks record borroweraddresses as “care of” a bank branch.

Dates of birth can also be an issue, particularlywhere the data are stored in disparate databasesand in inconsistent formats. In some cultures, theexact date of an individual’s birth may not be asimportant as it is in Western society. Records maygive the date of birth as born in “the year of thegreat flood,” which from a bureau’s point of viewis somewhat difficult to use as part of a match-and-merge logic.Another frequent problem with datesof birth, and possibly with all other identifiers, isincorrect data entry: customers may appear to behundreds of years old or, in some cases, not evenborn. Incorporating simple validation rules intoboth the lenders’and credit bureaus’ input systemscan help overcome such inconsistencies.

These issues, while posing challenges forbureau operators, should not prevent a creditbureau from starting operations. Experienceshows that establishing a credit bureau serves as atrigger for financial institutions to start recordingand managing information in a more systematicmanner, which in turn improves their own abilityto measure risks. Postponing the establishment ofa credit bureau until banks have perfect informa-tion may only hasten financial crises as banksincrease lending without adequate informationand risk management tools.

Specify Staffing Needs

A credit bureau’s function and its employees’duties are to obtain account history data from

member institutions, sort and aggregate these datainto personal credit histories, and disseminate thisinformation to members at their request. In addi-tion to performing these functions, credit bureaustaff must be trained to use the available creditinformation to conduct credit analyses. Staffshould be current on market trends and needs andbe able to provide bureau users with advice andcustomized products that incorporate the mostupdated information and use the latest technolo-gy. Bureau employees must also be trained to sortout irrelevant information from all the informationthey compile from various sources. Data alonehave limited value—data need to be aggregated,analyzed, and converted into useful and reliableinformation on which critical lending decisionscan be based.

Integrity is an absolute prerequisite for allbureau staff because on a daily basis they willreceive requests from consumers to amendrecords on the basis of friendships or relation-ships, or simply to access the bureau’s databasewithout the consumers’ consent. Resisting suchrequests is particularly difficult in close-knit com-munities where bestowing favors is the norm anddenying requests is contrary to the norm. Allbureaus should establish strong rules that dictateimmediate dismissal of personnel should they failto resist such requests.

2.2 Business Models forCredit Bureau Operations The implementation of a private credit bureaurequires a business model where all key playershave clearly defined roles (see Figure 17). Eachcountry develops its own approach in establishingand operating a credit bureau to fit its cultural,economic,political, and regulatory environment. Itis possible, however, to identify several distinctbusiness models for setting up a private creditbureau, based on the principle of reciprocity,which lies at the core of any credit bureau’s busi-ness. In brief, this principle may be summarized asfollows - members of a bureau can obtain creditreports only if they provide data to the bureau.Allthe models described below are based on thisprinciple and vary primarily in ownership andmanagement structures.


26

FFiigguurree 1177:: KKeeyy CCrreeddiitt BBuurreeaauu SSttaakkeehhoollddeerrss

� Owners/Shareholders � Credit Bureau Operator� Members/Users� Data Subjects (Consumers/SMEs)� Regulator

OOwwnneerrss//SShhaarreehhoollddeerrss are the investors that provide seed capital to enable the credit bureau toacquire assets, including technology, the physicalpremises, office equipment, and so forth.They alsorecruit necessary personnel to manage thebureau’s operations. Potential owners/share-holders include:

� BBaannkkss aanndd ootthheerr ffiinnaanncciiaall iinnssttiittuuttiioonnss.They canbecome shareholders individually or as a groupthrough, for example, a banking association. The Association of Banks inSingapore, for instance, owns a share inSingapore’s credit bureau. Other countries inwhich a group of banks own credit bureausinclude Serbia, Turkey, Romania, Ukraine,Kazakhstan, and the KSA.

� TTeecchhnniiccaall ppaarrttnneerr.. This category refers to a quali-fied credit bureau operator, such as Experian,TransUnion, Equifax, Dun & Bradstreet, CRIF,Creditinfo, and others.The technical partner maybecome a minority shareholder, as in Kazakhstan,Russia, and Mexico, or a majority shareholder, asin Singapore and the Czech Republic.

� GGoovveerrnnmmeenntt.. In some countries, government enti-ties,such as the Central Bank in Sri Lanka,or publicsector financial institutions,as in India and Thailand,become major shareholders in the credit bureau.

� PPrriivvaattee iinnddiivviidduuaallss//eennttiittiieess.. Individuals or organ-izations that provide venture capital and individ-ual entrepreneurs may also become partners inthe credit bureau as is the case of Datacheck inPakistan, CompuScan in South Africa, CRB inKenya, and many others.

OOppeerraattoorrss of private credit bureaus run andoperate the business on a day-to-day basis.This cat-egory refers to the on site executive managementteam and staff charged with all operational respon-sibilities, such as ensuring that data are collected,managed, and dispersed to its users.This manage-ment team is responsible for a bureau’s sustainabil-ity and reports regularly to its shareholders.

UUsseerrss//MMeemmbbeerrss usually are creditors, includingfinancial institutions and non-bank lenders thatcontribute credit information on their customers’accounts on a regular basis, typically monthly, tothe credit bureau. In keeping with the principle ofreciprocity, these lenders can access credit infor-mation reports from the bureau to assess the cred-it worthiness of new credit applications and toreview accounts. Bureaus usually chargeusers/members a pay-per-use fee (per click) aswell as a membership/enrollment fee.

DDaattaa SSuubbjjeeccttss ((CCoonnssuummeerrss//SSMMEEss)) consist of theborrowers on whom the credit bureau collectsand disseminates credit information.They are thesubjects on whom lenders wish to assess the risksof default and non-payment before approving anynew loans or advancing further credit.

RReegguullaattoorr is responsible for setting the guide-lines, rules, and regulations under which the cred-it bureau, its users, and consumers will operate. Inmany jurisdictions, the regulator is also theenforcement authority and has the right to issue alicense to the bureau and penalize the bureau incase of severe violations. Once the credit bureau isestablished and users are sharing information, theregulator’s role is to ensure ongoing compliancewith the regulations by all parties.

A private credit bureau’s ownership and gover-nance structure is important as all participantsmust have confidence and trust in the bureau.Thistrust in turn motivates banks to contribute their


27

data and helps the bureau develop more rapidlyand effectively. Two major models for a privatecredit bureau in terms of ownership structure andoperations include the following:

� Creditors own and operate the bureau.� Creditors do now own and do not operate the

bureau. Independent operator.

Model 1. Creditors Own and Operate the Bureau

Many markets around the world follow this model.Lenders, most often banks, establish a creditbureau where they are the main shareholders ofthe credit bureau. In most credit bureaus using thismodel, shareholding is fragmented and mayinclude several banks as well as government enti-ties, international financial institutions such as theIFC, independent private investors, and a technicalpartner as a minority shareholder.

A credit bureau may operate either with thesupport of a technical partner as in Mexico,Kazakhstan, Russia, and many other countries, orwith in-house developed systems as in Germany,Austria, and Brazil.

The main benefit for bureaus that use thismodel is that it allows for a faster startup com-pared to a model without lenders as shareholders.Agreement among banks to become shareholdersof the bureau implies a strong commitment toshare data and ensures the sustainability of thebureau. Participation of a government authoritymay also add credibility to the venture. Including atechnical partner as a shareholder allows the cred-it bureau to better align its incentives and guaran-tees a long-term commitment.

A major downside of this model is that a possi-ble conflict of interest may exist between theobjectives of individual creditor shareholders andthe bureau: some banks, for example, may stronglyobject to the inclusion of new entrants in thebureau.Larger shareholders may also influence thepricing policy so that it favors some members overothers and is not entirely based on commercialconsiderations. If a government entity becomes a

shareholder, the possibility of conflict of interestexists between its supervisory function and itsrole as a credit bureau shareholder.

Several countries use a version of this model.Many bureaus in Europe, for example in Germany,Austria, and Switzerland, started with non-bankcreditors as shareholders and eventually expandedownership to include banks. Some of thesebureaus were originally non-profit entities buteventually were reorganized into commercialcompanies. These bureaus operate on technicalplatforms developed in-house.

As technology develops and related costsdecrease, more recently established credit bureaushave chosen to engage a technical partner insteadof developing the credit bureau platform in-house.In some cases the technical partner becomes aminority shareholder. Examples include Mexico,where TransUnion is a shareholder along with sev-eral banks, and Kazakhstan’s First Credit Bureau,which includes Creditinfo as a technical partnerand a minority shareholder. The Indian creditbureau is a joint venture where Dun & Bradstreetand TransUnion each hold a 10 percent share andlenders hold the rest.

A variation of this model may include a struc-ture,where lenders become shareholders in a hold-ing company that, in turn, owns a share in a creditbureau along with the technical partner. Thisapproach was used in setting up a credit bureau inSingapore. The shareholders of the Credit BureauSingapore include the Association of Banks inSingapore, local Singaporean entrepreneurs, andthe technical partner Baycorp Advantage.

Model 2. Creditors Do Not Ownand Do Not Operate the Bureau.Independent Operator

This model is also referred to as an independentbureau because lenders do not own the company.Major international credit bureaus follow thismodel, including TransUnion, Experian, Equifax,and Baycorp Advantage in Australia and NewZealand.While this model is not so widespread inemerging markets, some examples do exist includ-


28

ing Datacheck in Pakistan, CRB in Kenya,CompuScan and XDS in South Africa, CreditRegistry Corporation in Nigeria, and TUCA inCentral America.

The main advantage of such a model is that it isa purely commercial entity with strong incentivesto innovate and provide high-quality service to itscustomers, especially if the market of credit infor-mation providers is competitive. Conflicts of inter-ests are absent because ownership and operationsare separate and relationships with members andusers are driven by commercial interests.

Attracting lenders and convincing them to shareinformation are the main challenges facing suchbureaus. Due to economies of scale, lenders onlyhave an incentive to share information when theyare certain that other lenders will join the creditbureau and share their information. This is themain obstacle in countries with no credit bureausor where a bank-owned bureau already exists, asindependent providers struggle to attract credi-tors. Another potential challenge relevant to newindependent bureaus is the lack of capital, whichmay hinder the development of new products.

Major international credit bureaus are mostlyindependent operators. However, the lenders’ fail-ure to coordinate among themselves may hinderthe success of this model in some emerging mar-kets. Consequently, in many countries when thefirst credit bureau is established, the only feasibleapproach is for lenders to become bureau ownersso as to build trust and initiate the process of infor-mation exchange. In cases where a bureau hasbeen successfully operating for a period of time, itmay be in the lender-owners’ interest to dilutetheir shareholding. The credit bureau’s ownerswould then be commercially motivated to broadentheir customer base as well as to introduce newproducts and services as was the case in HongKong (China).

A structure combining the elements of the twomodels is operating in South Africa. Experian andTransUnion, as independent operators, both col-lect and distribute credit payment information for

the Consumer Credit Association (CCA) and theSouth African Banking Council (SABC).The associ-ation and the council operate as closed usergroups and can determine who can become amember and access data.The association and thecouncil retain ownership of data, meaning that thecredit bureaus collect,process, and distribute data,but on behalf of and with permission of the asso-ciation and council.

2.3 Practical Considerationsfor Setting up a Credit BureauAlthough this Guide attempts to present a stan-dard approach to establishing a credit bureau, theactual process of establishing a bureau may differsubstantially from the proposed approach becauseno two countries are alike in terms of demograph-ics, regulations, legislation, and needs.

An independent provider or an entity, such as abanking association, that is interested in establish-ing a credit bureau has to consider the following:

� Market Assessment and Strategy� Financial Projections� Technology Needs and Technical Partner� Organizational Structure and Staffing

Market Assessment and Strategy

Market assessment and strategy development for acredit bureau can generally be divided into threephases — Phase 1: Assess Prospects, Phase 2:Conduct a Feasibility Study, Phase 3: Develop aBusiness Plan. Depending on the market environ-ment, there will be a certain degree of overlap andvariation in these phases.

Phase 1: Assess Prospects

This phase involves conducting an initial assess-ment of the demand for and feasibility of develop-ing a credit bureau.

Making a decision to set up a credit bureaurequires consideration of the following marketcharacteristics:


29

� Population size, which indicates potential cus-tomer base for lenders.

� Size of existing retail and SME credit market andpotential for growth.

� Level of sophistication of the credit market interms of products and services.

� Size of the existing credit bureau(s) service interms of borrowers covered and personnel.

� Threat of competition from public credit reg-istries or other private credit bureaus.

� Legal and regulatory environment pertaining tocredit bureaus.

� Level of support for credit information-sharingfrom the government, central bank, and otherfinancial institutions.

� Possibility for regional expansion (as in the caseof TUCA, see Case Studies).

The next step is to conduct consultations withall key stakeholders including:

� Financial authorities such as the central bank,bank supervisory authority,and especially repre-sentatives of the public credit registry if oneexists.

� Legislative bodies including commissions work-ing on credit bureau laws.

� Government bodies when relevant, includingmembers of the ministries of finance and com-merce and other relevant ministries.

� Licensing bodies including, where they exist,regulatory agencies that oversee credit bureaus.

� Lenders/financiers, including regulated banks,non-bank financial institutions, microfinanceinstitutions, leasing companies, and other credi-tors such as utility and telecom companies.

� All existing data providers, including legal firmsfamiliar with credit information-sharing issues.

Phase 2: Conduct a Feasibility Study

A feasibility study is an in-depth analysis of the mar-ket and stakeholders to determine whether creditbureau establishment is feasible and in what form.

The components of the study are as follows:

� Market analysis� Technical scoping study

� Stakeholder analysis� Legal and regulatory assessment

MMaarrkkeett aannaallyyssiiss.. This component of the feasibil-ity study examines the demand for, and supply ofcredit information and the market risks. It address-es such questions as:

� What is the potential demand for credit informa-tion?

� What are the existing sources of information,including public information sources?

� What is the risk of competition from other pri-vate bureaus and public registry, if one exists?

� What are the credit market trends? � What are the broader economic and political

risks?

A useful framework for conducting a marketanalysis is to conduct demand and supply assessments.

The demand-side analysis looks at:

� Depth of the credit market in terms of numberand types of products offered and volume ofclients served.

� Historical rate of growth of the credit market.� Major players (financial and non-financial) in the

credit industry.

The supply-side analysis answers the followingquestions:

� Is the demand for credit information satisfied bythe existing providers?

� Are there credit bureaus already operating inthe country and, if so, what market do theyserve?

� Is there a public credit registry, and does itinclude or have plans to include information onretail borrowers?

� What are the existing sources of public informa-tion that the bureau can rely on? (Examples caninclude electoral rolls, telephone registers, courtjudgments, bounced check lists, vehicle registra-tion registries, collateral registration registries,and the like.)


30

TTeecchhnniiccaall ssccooppiinngg ssttuuddyy.. The objective of atechnical scoping study, a critical part of the feasi-bility study, is to assess the technical capacity andreadiness of the lenders to participate in thebureau. It is used to define the technical specifica-tions of the future bureau. Conduct of this studyinvolves sending detailed questionnaires on thenature and formats of available information to allpotential participants and following up with meet-ings with all lenders to discuss the results of thesurvey.The focus here is on issues such as:

� Types of consumer and SME credit productsoffered.

� Level and growth rates of retail and SME credit,by product.

� Current and expected number of credits issuedto inform projections about the potential vol-ume of inquiries.

� Availability of electronically stored historicalinformation.

� Borrower consent to disclose information to acredit bureau.

� Availability of unique ID numbers for individualsand firms.

� Level of sophistication of lenders’ internal infor-mation management systems.

� Technology and infrastructure constraints.� Level of awareness among lenders on issues

related to credit reporting.� Lender willingness to share negative and posi-

tive information.

Comprehensive analysis of these issues is usedto develop the technical specifications for thebureau as well as help lenders make the changesneeded in their lending processes to enable themto join the bureau.

SSttaakkeehhoollddeerr aannaallyyssiiss.. This component of the fea-sibility study addresses issues related to possiblestakeholders of the credit bureau. It asks questionssuch as:

� Is there a broad consensus among lenders onthe usefulness of credit information-sharing?

� Who are the potential members/users of thebureau?

� Are lenders willing to share positive and nega-tive information?

� Do lenders have the technological capacity toshare the information?

� Are the authorities supportive? � What level of technical and communication

infrastructure exists in the country,and will it beable to support the needs of the credit bureau?

� What is the potential business model for thebureau?

� Are consumer rights issues being addressed?

LLeeggaall aanndd rreegguullaattoorryy aasssseessssmmeenntt.. This compo-nent requires meetings with regulators and quali-fied legal experts to assess the legal landscape inthe country in question. Following are the mainissues on which providers should focus while talk-ing to regulatory agencies:

� Is information sharing allowed and, if not, whatare the limitations?

� What is the existing legislation relevant to infor-mation sharing and the proposed credit bureau?

� Who are the enforcement authorities for thelaws relevant to information sharing?

� What new rules/regulations are being pro-posed?

� What are the implications of the legal frame-work for the credit bureau’s operations?

� How organized are consumer groups,and how likelyare they to oppose information-sharing plans?

Experience shows that in many emerging mar-kets, the legal environment for the operation ofcredit bureaus is not clearly defined. In such cases,a formal legal opinion from a qualified, local legalcounsel should be obtained to inform decisionmakers, from a legal and regulatory viewpoint, onthe feasibility of establishing a credit bureau inthat particular market.

Phase 3: Develop a Business Plan

If the results of the feasibility study indicate that aviable credit bureau can be established, the nextstep consists of drawing up a viable business plan.When an independent company plans to enter thecredit bureau market, the process is straightforwardand includes preparation of a business plan thatbuilds on the findings of the feasibility study andrequires securing financing, and obtaining commit-ment from lenders to participate in the bureau.


31

If a credit bureau is expected to be a lender-owned entity, the process involves the establish-ment of the entity itself.The key stakeholders mustagree on the establishment of the credit bureauand its structure.This agreement can be achievedwith the support of the banking association,which can be used as the key platform to bringtogether all banks interested in the establishmentof the credit bureau and can provide project man-agement support until the credit bureau entity isregistered and funded. In some countries, such asNigeria, lenders with the most interest in establish-ing a credit bureau have contracted with a privateconsulting firm to manage the process by organiz-ing meetings of key stakeholders, helping themreach agreement on the structure of the bureau,and preparing a business plan in consultation withall key stakeholders.

Once the credit bureau company is registered andfunded, the next key steps involve the following:

� Identify and select the technology infrastructureto decide whether the bureau should developthe needed software in-house or collaboratewith a technical partner.

� Prepare financial projections to determinewhen the bureau will breakeven and the opera-tional implications in terms of pricing of servic-es and necessary volumes of operation.

� Decide on the organizational structure and iden-tify and hire staff accordingly.

These issues are discussed in more detail in thesections below.

Once the decision on a technical partner istaken and a financial plan is finalized, the bureaucan proceed with entering into agreements withits members. Because the principle of reciprocityis the basis for exchanging information, users ofdata are also suppliers of data. In some exception-al cases, a member, such as a public source of data,may have an agreement to only supply informa-tion. In most cases, however, the agreement willcover the rights and responsibilities of the mem-bers as both users and suppliers of data. Figure 18summarizes the key issues to be addressed in theagreement that members must sign with the cred-it bureau to enable information exchange.


32

FFiigguurree 1188:: KKeeyy CCoonnttrraaccttss//AAggrreeeemmeennttss wwiitthh UUsseerrss aanndd SSuupppplliieerrss

Protection of the credit bureau, users, and consumersCommercial conditions on external data provi-sions and usageRestrictions on the use of the dataStipulate frequency of updatesNotification of errors in the informationSpecify type, media and format of updatesNotification of changes to type, media, and/or format of the data Process to ensure high data quality standardsData protection implications with the data suppliedIn case of outsourcing, a Service Level Agree-ment defining, obligations, availability of data, access to the database, backup, etc.

�

�

�

�

�

�

�

�

�

�

Principles of reciprocityRules of data sharingData ownershipUsage protocolConfidentialityCost of services and availabilityAdherence to data protection and/or consumer credit legislationsUser obligations to provide accurate, timely dataCredit bureau responsibility to process data, maintain integrity, and securityOther clauses concerning claims, costs, damages and penalties for inaccurate data

�

�

�

�

�

�

�

�

�

�

Users Suppliers

Contracts/Agreements

2.3.2 Technology Needs andTechnical PartnerCredit bureaus require an adequate technical infra-structure to process data accurately, provide theappropriate communication network, and offer aneffective and secure delivery mechanism to itsclients. Credit bureau systems are not off-the-shelfsolutions that can be acquired and installed into acomputer hardware system. Each credit reportingsystem must be a customized and locally adaptedsolution, usually requiring six to eighteen monthsto develop, from analysis of available data fromdata suppliers through preparation of functionalspecifications, actual system development, andacceptance testing. Results of the feasibility andscoping studies described in the previous sectionwill help in designing this solution.

Identification of technology needs requires thebureau to specify the functions the system mustperform, and decide whether to build the systemin-house or to outsource it to a technical partner.

Functional Requirements for CreditBureau Technology

The credit bureau’s systems must perform the fol-lowing functions:

� Collect validate, and merge data.� Generate and distribute reports.� Provide data security and backup.

Collect, validate, and merge data

Credit bureaus collect data from many differentsources and in many different formats. Because ofthis variety in sources and formats, a bureau couldrun the risk of collecting inaccurate or insufficientinformation and thus providing erroneous reportsto its subscribers. As part of the process for identifying technology needs, therefore, thebureau should develop a standard data requisitionform to send to its subscribers that stipulates theminimum information requirements. Such formsshould meet all national and local legislativerequirements for fair credit reporting and disclosure.

As a further attempt to mitigate risk, somebureaus, as for example in Kazakhstan, build theircredit bureau system in a way that does not allowthe bureau to change the records supplied by thelender.While the bureau may accept or reject a filesupplied by the lender, it cannot make changes inthe file, thus limiting the bureau’s liability in caseof error.

In addition, the bureau platform must bedesigned to receive information in many differentforms even though now most data suppliers sub-mit data in electronic format on-line. Less sophisti-cated clients, including small banks and non-bank-ing financial institutions, may be unable to provideinformation on-line, thus making it necessary forthe credit bureau to accept data on DVDs, CDs,diskettes, magnetic tape, or other portable datastorage devices.

The bureau is responsible for validating all databefore uploading the data to the database.Accordingly, the bureau’s system must includeautomated processes to check for completion ofall mandatory fields, conformity to the standardformat,and quality of data.The system must also beable to reject files that have critical errors or miss-ing information and send them back to the dataprovider, who must resend a corrected file.

After the data have been validated, the bureaumust merge the new data into its database.The sys-tem must be able to locate the respective subject,be it an individual or a legal entity, using nationalunique identifiers, such as passport or identitycard numbers, tax IDs. In countries where uniqueidentifiers of individuals and/or legal entities arenon-existent or unreliable, the credit bureau usual-ly develops its own matching key and, if needed,cross-references the information with public data-bases containing identification data. Whatevermatching key is used, the borrower’s identificationdetails must be found and must be accurate, thebureau’s objective being to match the incomingdata with the single best possible match from allthe files held on the bureau database.

Once the correct subject file has been identi-fied, the system will update the existing record or,


33

if the information relates to a new borrower, cre-ate a new credit file in the database.

Generate and distribute reports

The system must also have the capability to allowsubscribers to access the bureau’s credit reports.Typical modes of access include the following:

� OOnn--lliinnee aacccceessss.. The most popular on-line accessis system-to-system (host-to-host), whereby theclient’s system is connected to the bureau via ahigh-speed leased line. Interaction with thecredit bureau is performed entirely by theclient’s system with no human interaction. Ahost-to-host connectivity solution may berequired for even a start-up bureau, as some cus-tomers with large volumes of data would wantto integrate their back office system with creditdata to eliminate data duplication and to stream-line internal workflow.

� DDiiaalluupp oorr wweebb aacccceessss:: In this case, access to thecredit bureau is routed either via an Internetbrowser or other PC software.Once connected tothe bureau, the client provides necessary authen-tication information (user name, password, etc.)to validate access to the credit bureau.This modeof access is preferred among users who are nottechnically ready or who make limited inquiriesto the credit bureau and for whom cost of a host-to-host solution cannot be justified.

� BBaattcchh aacccceessss.. This mode of access providesclients with a cost-effective means of processinglarge volumes of requests. Lenders deliver infor-mation to the bureau via DVD, CD, magnetictape or electronically and the bureau should beable to processes these requests off-hours witha quick turn-around. This mode of delivery isusually recommended for processing of riskmonitoring for large client portfolios.

Provide data security and backup

Security is a high priority for credit bureausbecause they manage highly confidential customerinformation. Secure systems protect the data andreports and in so doing protect the bureau’s

integrity and reputation and contributes to its suc-cess.The bureau must make sure that it has prop-er physical as well as system security in place toprotect the sensitive information it holds. Thebureau’s physical office must only be accessible byauthorized personnel who have been previouslyscreened. Information theft is as much a physicalthreat as a system threat, where printed reports orcomputer diskettes/CDs can be picked up by apassing unauthorized person.

The system’s security features should include:

� Strict control of access to the database via mech-anisms for identifying and authenticating users.

� Protection against hackers.

� Clear delineation of authority among networkadministrators.

� Procedures for backing up data.

� Automated data updates.

� Provisions for recovering information in theevent of a disaster or total system failure (e.g.,use of a secure site for storing data where every-thing needed to restart the system is available incase of a disaster).

� Continual updating of all items stored in therecovery site.

� Periodic testing of backup hardware and recov-ery plans and procedures.

Adequate policies and procedures are an inte-gral part of implementing the bureau’s technologysystem.Bearing in mind that credit bureaus handledata that are highly confidential and sensitive, thebureau staff must sign internal agreements toabstain from any misuse or breach of confidential-ity of the bureau information.

In short, it is ultimately the responsibility of thecredit bureau management to put in place a secu-rity policy and take the following actions:

� Develop and circulate a security policy thatapplies to all bureau staff and any technical con-tractors with access to data.


34

� Run regular audit checks to ensure enforcementand adherence to the policy.

� Apply appropriate disciplinary action in theevent of security breaches.

� Review and update the policy as relevantchanges occur in technology, data protectionlegislation, employment law, and the like.

� Keep members of staff up to date with the policy.

Failure to address the issues of security and dis-aster preparedness in the early stages of technolo-gy development puts bureaus at high risk and cre-ates a technical gap that might negatively affecttheir ability to attract lenders to their services.

Technology Acquisition Strategy

Once the functional requirements have been spec-ified and the hardware and software needs identi-fied, the bureau must decide whether to build itstechnology in-house or to outsource it to a techni-cal partner.

The emerging combination of micro/mid-rangeand web technology has dramatically changed thecredit bureau industry. Until a few years ago, theindustry mostly operated on heavy and costlymainframes. Now credit bureaus use web-basedapplications and are no longer as expensive tolaunch as they once were.The positive impact ofthe new technology, with its downward pressureon costs and lowering of entry barriers for newplayers, has made it more convenient for investorsand entrepreneurs in emerging and developingcountries to “buy” technical platforms from exter-nal sources—usually internationally reputabletechnology providers— rather than “build” themin-house.

Cost should not be the only driver in the deci-sion to buy a system. It may not be sufficient tohave a good local technical infrastructure, well-trained domestic computer programmers and ITspecialists. Credit bureaus require unique knowl-edge and experience to develop and operate thesystem because of their complexity and the high

sensitivity of the data held. Participation of aninternational credit bureau operator as a technicalpartner in a new credit bureau, therefore, not onlycan provide the required technology and technicalexpertise but also, and more importantly, can lendits reputation, crucial business know-how, and theexpertise to continuously develop new value-added services as the bureau grows and the mar-ket matures.

These added benefits explain why the majorityof emerging countries that have establishedlender-owned credit bureaus in the last few years,including Mexico, Czech Republic, Turkey,Romania, Kazakhstan, Russia, Ukraine, KSA, andEgypt have opted for the “buy” option. In somecases, the terms of the partnership may includeselling a share in the bureau to the technical part-ner as in Mexico, Kazakhstan, and Russia; in othercases, the contractual agreement covers only thesale of the software and support. Even when cred-it bureaus have opted to develop the software in-house, they often call upon international technolo-gy providers when their markets are ready to offervalue-added services—such as credit scoring—that require more than just technology.

When selecting a technical partner, the bureauneeds to evaluate a potential partner according tothe following criteria:

� Technical: Does the potential partner have thecapability to implement the system in accor-dance with the local technical specifications?Does it have a track record in implementingcredit bureau systems in similar markets?

� Strategic: Is the potential partner able to committo the credit bureau over the long term?

� Financial: Is the cost of the system in line withthe demand for services?

Figure 19 provides several key criteria that needto be applied from a technical and strategic pointof view.The decision regarding the financial impli-cations of choosing a technical solution must bemade in conjunction with overall financial plan-ning and is discussed in the next section.


35

TTaabbllee 22:: HHyyppootthheettiiccaall PPrriicciinngg MMaattrriixx ffoorr CCrreeddiitt BBuurreeaauuss

Inquiry volume Price Per Inquiry

<25000

25001 - 50000 1.00

50001 - 100000 0.95

100001 - 250000 0.85

250001 - 500000 0.80

> 500000 0.70

2.3.3 Financial ProjectionsForecasting financial outcomes of establishing thebureau requires an assessment of potential rev-enue and costs and an identification of the driversin each of these categories.

RReevveennuuee PPrroojjeeccttiioonnss.. The main revenue driverfor the credit bureau business is the number ofcredit reports or value-added services sold.Revenue projections are based on the estimateddemand for credit reports and the pricing ofreports. In most cases, the bureau charges a flatmembership fee plus a charge per inquiry (perclick). Volume discounts usually apply, and it iscommon to have a pricing matrix depending onthe volume of inquiries and the type of user.Table2 provides a hypothetical pricing matrix based onthe annual inquiry volume per user. The cutoffpoints for volume discounts are determined basedon projected demand and average expectedinquiries.

The inquiry-demand estimate is based on thesurvey of potential users.The financial projectionsfor revenue should allow for time between thelaunch of a credit bureau’s operations and thebreakeven point at which it actually achieves itstargeted inquiry volume. It is common to havemany technical issues related to connecting alender to the bureau and integrating the creditbureau’s information into the lending cycle of theinstitution. Resolving these issues may take at leastthree to six months. The growth rate for the vol-ume of inquiries is based on the projected creditgrowth rate for the economy and the expectednumber of new users joining the bureau. It is fea-sible to have growth rates of 50 percent and abovein the first three to five years of a bureau’s opera-tions in a country with stable credit growth andnew members joining the bureau.


36

FFiigguurree 1199:: QQuuaalliittiieess ooff aa SSttrroonngg TTeecchhnniiccaall PPaarrttnneerr

Willingness to add value to business plan and financial modelWillingness to take equity positionsFinancial strength of companyManagement profileAvailability of office/skilled resources in or near project countryUnderstanding of domestic banking/credit market and related issuesDirect relationship (no 3rd party)Willingness and proposal for know-how transfer

�

�

�

�

�

�

�

�

Experience (years)Track record/success with setting up bureaus in developing and developed economiesExpertise of personnel/management teamAbility to provide comprehensive solutions (products, software and value-added services)

�

�

�

�

Technical Strategic

Selection Criteria

CCoosstt PPrroojjeeccttiioonnss.. In large part, the costs are driv-en by the choice to acquire a credit bureau tech-nology system, also referred to as bureau platform,or to develop one in-house. In both cases, the pos-sible cost range is wide and depends on the levelof sophistication of the system and the types ofproducts it is expected to provide.

Cost projections based on the assumption thatan existing platform will be acquired must includethe following cost elements:

� Development/customization/installation fee forthe credit bureau platform (usually paid ininstallments).

� Maintenance fee, usually a flat fee paid monthly,quarterly, or annually.

� License and royalty fees paid to the technicalpartner based on the number of inquiriesreceived by the system in addition to fees tocover ongoing updates and enhancements tothe system, usually at an agreed-upon rate.

� Consultancy fees charged by the technical part-ner for any service over and above the servicesspecified in the development and maintenanceagreement.

The exact fee structure of the contract will varysignificantly in each case. For example, the licenseand royalty fees may not be applicable if the tech-nical partner is a shareholder in the bureau.

Other elements to be addressed in the cost pro-jections include hardware, such as database andnetwork servers; network equipment and worksta-


37

TTaabbllee 33:: HHyyppootthheettiiccaall PPrrooffiitt && LLoossss SSttaatteemmeenntt

YYEEAARR 11 YYEEAARR 22 YYEEAARR 33 YYEEAARR 44 YYEEAARR 55

TTOOTTAALL RREEVVEENNUUEE 00 550000,,000000 11,,000000,,000000 11,,775500,,000000 22,,662255,,000000% change in revenue 0 100% 75% 50%

CCoossttOOppeerraattiinngg ccoossttLabor 315,000 346,500 450,450 585,585 761,261Rent 50,000 52,500 55,125 57,881 60,775Utilities 1,500 1,800 2,160 2,592 3,110Office equipment, supplies 7,000 8,000 8,000 8,000 8,000Telecom 14,400 17,280 20,736 24,883 29,860Audit, legal, and other fees 12,000 12,000 12,000 12,000 12,000Insurance 13,000 13,000 13,000 13,000 13,000External data, marketing 20,000 25,000 30,000 37,500 46,250

Total Operating Costs 432,900 476,080 591,471 741,441 934,256% of Total Cost 52% 55% 54% 53% 53%

FFiixxeedd ccoossttRent, furniture, other 20,000 20,000 20,000 20,000 20,000fixed costsSystem HW & SW 75,000 75,000 75,000 75,000 75,000

Credit Bureau Platform 300,000 300,000 400,000 550,000 725,000% of Total Cost 36% 34% 37% 40% 41%

Total Fixed Cost 395,000 395,000 495,000 645,000 820,000

TTOOTTAALL CCOOSSTT 827,900 871,080 1,086,471 1,386,441 1,754,256% change in cost 5% 25% 28% 27%

NET INCOME BEFOREINTEREST & TAXES (827,900) (371,080) (86,471) 363,559 870,744

Tax 0 0 0 109,068 261,223NET INCOME AFTER TAXES (827,900) (371,080) (86,471) 254,491 609,521


38

tions; system software and other necessary soft-ware applications; office furniture and equipment;utilities and telecom expenses; and labor costs. Insome cases, an important cost component is thecost of the data that the bureau acquires fromexternal sources and then sells as part of its prod-uct offering.

Table 3 provides a hypothetical profit and lossstatement for a credit bureau’s five-year businessplan.

Based on this hypothetical financial plan, thecredit bureau would breakeven in the third year ofoperations. In most cases, credit bureaus reach thebreakeven point over a three-to-five year period(see Figure 20).

FFiigguurree 2200:: BBrreeaakkeevveenn PPooiinntt ffoorr aa NNeewwllyy EEssttaabblliisshheedd CCrreeddiitt BBuurreeaauu

It is important to assess the high and low sce-narios for the operation of the credit bureau as thesuccessful operationalizing of the credit bureaudepends on many external factors. For example,often the bureau faces delays in beginning its oper-ations due to the inability of banks to upload datato the bureau.Another example is underestimatedcosts. In many emerging markets, customizing andimplementing the system may require a signifi-cantly longer period of time than originallyplanned and contracted for. Usually, this meansthat the bureau may have to pay high consultingfees to the technology provider to finalize the sys-tem implementation, which is likely to delay thetiming of the breakeven point.

2.3.4 OrganizationalStructure and Staffing

Pre-operational Phase

During the startup or pre-operational phase, thebureau does not have any revenue and must, there-fore, keep staff numbers to an absolute minimum.Growth in market demand for its products andservices will determine subsequent staff growth.

Initially, staff members should cover more thanone role, whenever possible.The early phase of acredit bureau can essentially be run by a generalmanager/project manager, an office/communica-tion manager, and a technical coordinator. In gen-eral, employing a general/project manager who isknowledgeable, experienced, and well connectedin the financial sector is a critical factor for suc-cess. In addition to providing technical assistance,a reputable international technical partner canalso provide strategy and business developmentsupport to bureau management. Finance, adminis-trative and legal functions can be outsourced atthe beginning.

As with any other private business, staffing dur-ing the pre-operational phase of a credit bureauneeds to be tightly controlled.The credit bureau’smarket can be highly unpredictable, and a success-ful launch requires a combination of ingredientsthat could be difficult to obtain all at once, such asan enabling regulatory environment,willingness ofdata providers to contribute their data, availabledata of good quality,and pricing conditions accept-able to all parties involved. An over-staffed creditbureau in the pre-operational phase may fail evenbefore it starts selling its first credit report.

Operational Phase

Once the bureau becomes operational (i.e., thesystem goes live and starts selling its first reports)several factors affect the decision on how manypeople should be assigned to each role or whethermultiple roles could be assigned to a single posi-tion.The workload for each role guides these deci-sions.Among the factors to consider in determin-ing workloads are the following:


39

� Number of existing and potential subscribers.� Number of branches/workstations connected to

the bureau.� Inquiry volumes.� Competitors’ strength.� Consumer awareness of rights and legislation.� Projected and actual database size.� Growth plans for the bureau.� Complexity of operations (e.g., need for “off-

line” checks/updates overnight or on week-ends).

The main divisions of the operational creditbureau are: IT and Operations, BusinessDevelopment and Marketing and, Finance/Administration. Divisional heads in each area

report directly to the Chief Executive Officer(CEO) who manages the company’s activities and,in turn, reports to the Board of Directors. TheBoard, whose members are appointed by the par-ticipating investors of the bureau, is responsiblefor the overall corporate governance. Ideally, theBoard should include one or two members of theexecutive team (the Managing Director andOperations Director/representative of the techni-cal partner).The Board of Directors nominates oneof its members as Chairman of the Board. Figure21 presents a traditional organizational chart of anoperational credit bureau. Staffing requirementsand responsibilities for an operational creditbureau are outlined in Table 4.

FFiigguurree 2211:: OOrrggaanniizzaattiioonnaall SSttrruuccttuurree ooff CCrreeddiitt BBuurreeaauu

Legal/Compliance Unit

CEO / MD

AdministrativeOfficers

IT/SupportService

NetworkAdministrator

Database Officer

Customer ServiceSales & MarketingOfficers

Head of Finance/Administration

Head of IT & Operations

Head of Business Development& Marketing


40

RRoollee KKeeyy TTaasskkss

MMaannaaggiinngg Overall bureau strategy DDiirreeccttoorr//CCEEOO Marketing/business development activities

HHeeaadd ooff Financial and administrative operationsFFiinnaannccee aanndd Human resources functions (recruitment, compensation, performance AAddmmiinniissttrraattiioonn management, career development)

Finance/ Day-to-day administrative and bookkeeping operationsAdministration

Legal Counsel Overall legal supportInternal legal training

HHeeaadd ooff Market segmentation MMaarrkkeettiinngg aanndd Product development and brandingBBuussiinneessss AdvertisingDDeevveellooppmmeenntt Sales and promotion

Sales & Marketing Responsible for relationship with existing clients and for new clients Officers enrollment

Responsible for implementing sales & marketing plan and achieving business objectives Advertising, conferences/exhibitionsMarket researchMedia affairsdentify new data sources

HHeeaadd ooff Vendor relationsTTeecchhnnoollooggyy aanndd Data management OOppeerraattiioonnss Technology management

Network and security operationsCustomer service

Customer Consumer Help DeskService Officer

Database Officer Data quality checking proceduresData loadingEmergency updates

Network Network administrationAdministrator Subscriber communications interfaces

Network security

IT Support / HousekeepingTechnical Service System administration

Subscriber and internal Help Desk

TTaabbllee 44:: OOppeerraattiioonnaall PPhhaassee SSttaaffffiinngg


41

The bureau should operate a help desk staffedwith the qualified technical experts. Help desktechnical experts should assist credit bureau mem-bers who have problems connecting to the sys-tem, uploading data, and modifying some of thetheir data as well as new lenders that may needadditional help in enabling their internal systemsto “talk” to the bureau.

The help desk also should assist consumers andfirms whose information is stored in the creditbureau and who have the right to obtain their owncredit reports and challenge any erroneous infor-mation.The help desk staff should aid these clientsin obtaining their reports and registering com-plaints.The staff should also inform the responsi-ble team to determine whether a mistake resultedfrom bureau operations and the institution thathave submitted the data if necessary. This is animportant tool for monitoring the quality of thedata in the bureau.

To accommodate the needs of growing numbersof users and borrowers and their respectiverequests, most of the growth in staff in the creditbureau will occur in the Customer Service depart-ment. The Sales and Marketing group would alsoneed to grow to promote the products and servicesof the bureau as it tries to expand into new markets.

Last but not least, it is recommended that thecredit bureau set up a “Compliance Committee”(or Compliance Auditor) early on in the process.The Committee or Auditor would report directlyto the Board of Directors and regulators on issuesrelating to quality control. Its main duties are toensure compliance with the credit bureau’s Codeof Conduct and monitor any misuse of creditbureau data by users.

2.4 Measuring Effectivenessof a Credit BureauThe effectiveness of a credit bureau, as of anyother firm, can be measured in many differentways. A good performance measurement systemincludes multiple dimensions of performance,including financial, operational, and behavioral

characteristics. The key categories for measure-ment include: quality, quantity, timeliness of prod-ucts and services delivered, financial performance,and customer satisfaction (see Figure 22).

FFiigguurree 2222:: KKeeyy PPeerrffoorrmmaannccee IInnddiiccaattoorrss ooff aa CCrreeddiitt BBuurreeaauu

11)) QQuuaannttiittyy.. This category is a measure of thevolume of goods and services delivered. Relevantindicators may include:

� Number of queries received by the system overthe reporting period.This is the key measure ofthe demand for the credit bureau services.

� Number of credit reports sold. This is the keyoutput measure for the bureau. It can also betracked at the product level, for example howmany basic reports are sold, how many reportswith credit scores are sold, etc.

� Number of borrowers with credit records inthe system at the end of the reporting period.This measure can also be tracked for differentcategories of borrowers, such as firms andindividuals.

� Number of records in the system at the end ofthe reporting period. Each borrower may havemore than one credit line and the history oneach credit line is stored separately.

� Hit ratio. This is the ratio of the number ofreports issued to the number of queriesreceived and is an important indicator of theability of the bureau to satisfy lenders’ demandfor information.


42

� Number of products offered.This could includebasic reports, detailed reports, credit scores,portfolio monitoring, fraud detection, etc.

The World Bank Doing Business surveys esti-mates the depth of coverage by private creditbureaus at country level.The indicator used in thiscase is the number of borrowers covered by thesystem (see Figure 13).

A bureau’s objective is to simultaneouslyincrease its coverage ratio, defined as the numberof borrowers in the system divided by the activepopulation, and the hit ratio. Consideration of onlyone of these two measures does not provide anadequate understanding of the bureau’s perform-ance. For example, the hit ratio may be high in abureau with a very low coverage ratio.This situa-tion, often found in markets with underdevelopedcredit markets, indicates that the formal financialsystem serves a small group of individuals andmost lenders continue targeting the same groupfor new lending.

22)) QQuuaalliittyy.. This category refers to the accuracy,currency, completeness, and consistency of thebureau’s data. Information, the main asset of thecredit bureau, only has value if it is accurate andcurrent.Relevant indicators of quality may include:

� Number of complaints.The bureau must have amechanism to receive and log complaints of theborrowers about the accuracy of the informa-tion in their credit reports.

� The percentage of complaints with inaccuraciesdue to the actions of the bureau. Many com-plaints that bureau receives may be unjustifiedor result from mistakes made by the data suppli-er and not the bureau. Tracking the number ofcomplaints that can be attributed to thebureau’s actions allows the bureau to improvethe quality of its processes.

� Data quality reports. The bureau should run dataquality reports to analyze the completeness andconsistency of the data.Such reports produce tabu-lations of fields such as IDs and addresses, dates ofbirth, and other identifying information and allowthe credit bureau to determine whether there areduplicate or incomplete files in the system.

� Number of rejected files. When accepting thedata from the supplier, the bureau runs simpleconsistency checks on the data in the files. If thefile does not pass this test, is the system rejectsit and sends it back to the data supplier.Trackingthe number of rejected files allows the creditbureau to monitor the quality of data available inthe market.

Because bureaus are privately owned, ascertain-ing the quality of information contained in thebureau databases is difficult.The only publicly avail-able information is self-reported by the bureaus intheir annual statements. Several studies have beenconducted recently in the United States and Europeto examine this question. For example, a report bythe Consumer Federation of America and theNational Credit Reporting Association containedanalyses of the information in the credit reportsissued by the three large U.S. credit bureaus:TransUnion, Equifax, and Experian.The authors ana-lyzed a random sample of credit reports and foundthat 82.4 percent of the files contained inconsisten-cies in the category of the balance on revolvingaccounts or collections and 96.1 percent of the fileshad inconsistencies in the category of account’scredit limit. In 43.1 percent of the files, conflictingreports existed for the same accounts on how oftenthe consumer’s payments had been late by 30 days.The study estimates that,based on inaccuracies con-tained in credit reports, nearly 40 million peoplemay be mistakenly classified as sub-prime borrow-ers in the mortgage market.25

While it is true that credit bureaus manage high-ly complex and extremely large databases, therebymaking mistakes inevitable, it is essential that abureau implement all necessary procedures toassess the quality of data and take necessaryactions to ensure data consistency and accuracy.

33)) TTiimmeelliinneessss.. This category refers to the time ittakes for lenders to receive a credit report. The

25Consumer Federation of America and National Credit Reporting Association.December 2002. Credit Score Accuracy and Implications for Consumers.Available at: http://www.ncrainc.org/documents/CFA%20NCRA%20Credit%20Score%20Report.pdf.


43

World Bank’s Doing Business Survey26 indicatesapproximately 57 percent of private creditbureaus (in a survey of 78 credit bureaus) report-ed that data requests were met instantaneously.The majority of credit bureaus met all demandswithin seven days of receiving a request (seeFigure 23).

FFiigguurree 2233:: AAvveerraaggee TTiimmee bbeettwweeeenn RReeqquueesstt aanndd RReelleeaassee ooffDDaattaa

The key indicators for the timeliness of servicemay include:

� TTiimmee bbeettwweeeenn oobbttaaiinniinngg tthhee qquueerryy aanndd iissssuuiinnggtthhee rreeppoorrtt.. In many countries, the process isautomated, depends on the search capacity ofthe software, and takes seconds or less. In manydeveloping countries where the reports are notprovided on-line, the process may take hours orin some cases days.Minimizing the delivery timeis the key timeliness objective for the bureau.

� TTiimmee ttoo aassssiimmiillaattee iinnffoorrmmaattiioonn,, uuppddaattee rreeccoorrddss,,aanndd rreeccttiiffyy eerrrroorrss.. This is the time betweenreceiving information or updates from the datacontributors and integrating them into the data-base. Running checks on the files received fromlenders and merging the new records with exist-ing files may take anywhere from one day to onemonth27 depending on the quality of the infor-mation supplied by the lenders, the reliability ofthe unique identifiers,or the merging algorithm.This parameter is critical to ensure that the dataavailable to lenders are current.

This indicator also measures the time it takesthe credit bureau to correct any data errors that

have been reported. According to the WorldBank’s 2004 survey, 78 private credit bureaussurveyed, approximately 76 percent reportedtaking less than two weeks to rectify errors.Another eight percent reported taking betweentwo weeks and one month to correct errors.28

44)) FFiinnaanncciiaall PPeerrffoorrmmaannccee.. While return on equi-ty, profit margins, and operational costs are stan-dard indicators of financial performance, thebureau may also track more specific indicators,such as:

� Profit margin per product line.The services thatbureaus provide vary greatly and are bound tohave different levels of profitability and coststructure. For example, while the bureau maysell raw data at a relatively low cost, it may sellanalytical products, such as credit scoring andportfolio monitoring, at higher margins.

� Profit margin per client. Bureaus aim to attractlarge creditors by providing significant volumediscounts. Analysis of profit margins by clientallows a bureau to better tailor its pricing strategy.

55)) CCuussttoommeerr ssaattiissffaaccttiioonn.. Methods used tomeasure this category include customer surveysor actions taken by customer.

� Number of complaints. By tracking separatelycomplaints from lenders and those from datasubjects, the bureau can identify areas forimprovement.

� Average time to resolve complaint. Providing fastresponses to complaints is one way of improvingclient satisfaction.One approach to doing so is tooperate a help desk with staff available to answerquestions and complaints promptly.

Systematically tracking a set of key indicatorsenables the bureau to monitor its performanceand formulate a clear strategy to improve service.

26World Bank. 2004. Doing Business Database on Private Credit Bureaus as of2004.

27World Bank. 2004. Doing Business Database on Private Credit Bureaus. Basedon information from 62 private credit bureaus.

28Ibid. Information not available for 13 private credit bureaus.

3DevelopingValue-Added Servicesin Emerging Markets

Developing Value-Added Services in Emerging Markets

45

3.1 Definition and Importanceof Value-Added ServicesValue-added services comprise a broad class ofproducts that more sophisticated credit bureauscan offer. Such services entail the manipulation,processing,and analysis of existing raw data to pro-duce tools that can be easily integrated into banks’credit approval and risk management processes.The range of potential value-added services isquite extensive and includes, but is not limited to:

� Marketing services � Credit scoring � Application processing � Portfolio monitoring � Fraud detection � Collections

Raw credit data can be very useful in each ofthese areas, however, the lender needs significanttime, resources, and expertise to analyze and inter-pret the data.Value-added services use a variety oftechniques, ranging from simple data aggregationand cross-referencing to complex statistical algo-rithms, to provide an output that provides thelender with a simple interpretation of the informa-tion available (e.g., a risk score).

Given the volume of decisions often required tomanage a typical retail portfolio (e.g., grant/rejectfacility, over-limit authorization, cross sell/up sell,past due action required), many lenders haveturned to automation as a means of maintainingefficiency, but raw data in the form of a credit

report can be extremely difficult to integrate intosuch systems. Many types of value-added service(e.g., application processing systems and behav-ioral risk assessment) however, lend themselvesperfectly to inclusion within automated systems.

The major benefit of automated decision-mak-ing systems is that they allow users to managemany customer decisions on an exception basisrather than having to review each and every case.This helps contain the need for employing highlyexperienced, and often very expensive, individualsto make mundane or rudimentary decisions andallows lenders to channel this experience intomore productive tasks.

Larger financial institutions that operate in devel-oped markets typically develop customized value-adding tools, either using in-house analytical teamsor contracting with one of the numerous special-ized companies that have emerged to service thismarket, such as FairIsaac and Experian-Scorex(Global), PIC Solutions (Africa), and LISIM (LatinAmerica). Financial institutions in developing mar-kets, however, tend to be smaller and either havecustomer databases that are too small for such solu-tions to be statistically reliable or find it difficult tojustify the up-front capital cost of development.

In emerging markets, therefore, the creditbureau can play an important role in makingthese types of services available to a broaderaudience, by pooling data across a range of cus-tomers and spreading the cost of developmentacross its user base.

Although users still have to pay for these servic-es, typically on a “pay as you go” or “click” basis,they get immediate access to the benefits ofimproved lending methodologies, more cost-effec-tive processes, and increased operational efficien-cy that, under other circumstances, would only beavailable to larger institutions.

3.2 International IndustryTrends The range of value-added services offered by cred-it bureaus has broadened significantly over the last20 years. This growth has been fueled both fromthe demand side—users wanting more and moresophisticated products—and the supply side—bureaus trying to increase/maintain income mar-gins in an environment where there is consistentdownward pressure on commodity prices (thecost of the raw data).

The scope of the products offered is very mucha function of the environment in which the credit

bureau operates,that is, the extent to which the rawdata can be used.The trend in developed markets,however, has been to create a suite of value-addingproducts aligned to what is sometimes referred toas the “customer life cycle” (see Figure 24).

The customer life cycle effectively mirrors thecore business functions adopted by most lenderswhen managing customers: prospecting and mar-keting, new business acquisition (loan process-ing), customer relationship management, and col-lections.

The credit bureau, typically, builds products orsolutions that help its customers in each of thesebusiness functions make better or faster decisionsby using the predictive nature of bureau data. Ineffect, the bureau is re-cycling its databases so thatusers access the files more than just at the pointwhen they make an initial loan inquiry. For exam-ple, a behavioral scoring system may access a cus-tomer’s credit file monthly to identify updatesrather than just once at the point of application.


46

FFiigguurree 2244:: CCuussttoommeerr LLiiffee CCyyccllee:: OOffffeerriinngg VVaalluuee--aaddddeedd SSeerrvviicceess

SSttaaggee 11:: IInniittiiaall DDeeppllooyymmeenntt.. At inception, a newcredit bureau is intrinsically an empty box, a data-base with no data. Although able to back-fill thedatabase with some historical records, such as pastdefaults, the bureau does not have access to themost powerful data sources, such as the creditrepayment record and the inquiry footprint,whichneed to be built up over time.

SSttaaggee 22:: UUsseerr AAccqquuiissiittiioonn.. Although not neces-sarily the case in all countries, the trend in manyemerging markets is for the initial development ofcredit bureaus to take place within the bankingcommunity. The main driver of this approach isthat the banks are the major providers of creditand have one clearly defined supervisory entity.The first step then is to upload the data from theinitial members, i.e., the lenders.

SSttaaggee 33:: DDaattaa DDiivveerrssiiffiiccaattiioonn.. In parallel withStage 2, the bureau attempts to augment the basiccredit history data with other forms of informationthat may be beneficial to users, such as electoralrolls, identity records, court judgments, telephonenumbers, and company registration records. Thistype of data can be particularly useful to members:it may be predictive of future borrower behavioror it may make their processes simpler by provid-ing a portal to a “one-stop data shop”.The data alsoprovide a valuable source of information for datamining and modeling.

SSttaaggee 44:: UUsseerr DDiivveerrssiiffiiccaattiioonn.. Even where thebanks take a pro-active role in establishing thecredit bureau, it is often clear from the outset that,at some point in time, the user base should expandto include non-bank creditors, such as telecommu-nications companies, microfinance lenders, andthe like.

The introduction of new users can have a pro-found effect on the composition of the bureaudatabases and, therefore, the predictive nature ofthe data. In several countries,expanding to includetelecommunications providers has had a signifi-cant impact on the predictive power of the inquirydatabase as the pattern of telecom payments maybe indicative of future defaults on bank credits.

Some value-added services may be no morethan enhanced bureau reports, such as a watch-outservice that pro-actively advises a lender of achange to a customer’s file, and requires little inthe way of analytical expertise. Having introducedthese services at a relatively early stage, most cred-it bureaus aim to move up the value chain to addincreasingly more sophisticated tools, such as scor-ing and credit information management software.These more complex solutions have the dual ben-efit of generating greater revenues for the bureauand also locking in the clients to the bureau serv-ices, (i.e., making users more reliant on the supply-ing bureau and, therefore, less likely to transfer alle-giance to competitive sources of information).

In developed economies, bureaus tend to usespecialized internal analytical teams to developand maintain these services. In emerging countriesthat lack such specialists, bureaus typically rely onoutsourcing development, often to the same ven-dors that supply custom services directly to thelenders, such as FairIsaac, Experian, or LISIM, whorecently developed generic microfinance bureauscoring services for CompuScan in South Africa.The critical issue, however, is not who developsthe services, but when they can be deployed.

The credit bureau databases in most developedcountries have had many years to develop, are richin information, and for the most part offer highquality data, thus providing an ideal base for datamining and data modeling.

The credit bureau databases in many emergingmarkets,however, are considerably less developed:they may have information only from banks andmay not have been operational long enough tobuild the diversity of information sources requiredfor value-added products. Their inquiry/searchdatabases, for example, may not contain enoughhistorical information. In these circumstances, itmay be difficult, or indeed impossible, to buildsome of the more sophisticated solutions, such ascredit scoring.

Planning for the development of value-addedservices requires an understanding of the stagesrequired for a credit bureau to “mature.”


47

Adding new bureau members also has implica-tions in terms of reciprocity, namely access to theinformation on the basis of their level of data con-tribution. The rules of reciprocity extend to thedesign and delivery of value-added products. Abureau score that incorporates positive credit his-tory information, for example, should not be madeavailable to a member that provides only negativeinformation, even if the member never actuallysees the positive data.

SSttaaggee 55:: DDaattaabbaassee MMaattuurriittyy.. Credit bureau data-bases change over time as the availability of datasources and the number/type of users change.Databases typically tend to grow in both depthand breadth, but not always. Privacy restrictionscan result in changes to the availability of certaintypes of information as was seen in the UnitedKingdom in 2000 when restrictions were placedon using electoral roll information.

In general, however, the core bureau databaseneeds a period of time to mature from the abovestages of development in order for the data there-in to be predictive of a future outcome (see sec-tion below on Credit Scoring).

The ever-changing nature of the databaseexplains why value-added products and servicesrequire continuous monitoring and fine-tuning.Estimates based on today’s data may not apply 12months from now as the overall economic envi-ronment may change.

SSttaaggee 66:: SSeerrvviiccee EExxppaannssiioonn.. There are no hardand fast rules as to when value-added services canbe introduced. Simple services, such as expandedcredit reports, can be introduced at low cost at arelatively early stage, even during Stages 2 and 3.Bureaus typically develop more sophisticatedproducts, such as credit scoring, which are usuallymore expensive to build and maintain, when thedatabase and to some extent the user base havereached a level of maturity where the resultingproducts will be both robust and have a reason-able shelf life.This is most likely to occur once thebureau has reached Stages 3 or 4. It is only whenthe bureau has reached stage 5, however, that abroad suite of products, such as described inFigure 24, can be contemplated.

There are two other key factors that a bureauwould typically take into consideration when devel-oping value-added services: Return on Investment(ROI) and users’ capacity to adopt the service.

� RReettuurrnn oonn IInnvveessttmmeenntt.. A clear business casemust exist for the development of a value-addedservice.The projected revenue from the sales ofthe services must cover the investment cost andproduce positive return.The pricing and market-ing strategy often includes bundling value-added services with the sale of core data.

� TThhee ccaappaacciittyy ooff uusseerrss ttoo aaddoopptt tthhee sseerrvviiccee..Members will only demand a service if theyhave the capacity to use the service to improvesome element of their own processes.A bureauscore, for example, adds no value unless thelender is able to integrate it into its credit under-writing process to lower the costs of creditapproval.User-side constraints have a significantbearing, especially in emerging markets, on whowill use the services and in what quantities.

Even in developed markets, the uptake of newbureau products and services is not guaranteedand typically requires a highly pro-active sales andmarketing department/staff to promote the prod-uct to the market. Experian, for example, launchedDetect, its industry-leading fraud detection prod-uct, three times in the United Kingdom before theproduct received market acceptance.

In emerging markets, the problem of accept-ance of value-added products is even more pro-nounced. Except for the international banks, manylenders in emerging markets lack an understand-ing of the lending methodologies that can beimplemented using these services and of the ITinfrastructure needed to deploy them.

Credit bureaus in emerging markets should notunderestimate, therefore, the need for outreachtraining, market development, and sales functionswithin their organizations. As products becomemore sophisticated and more analytical, bureausshould also recognize the need to have internalspecialist resources to monitor and maintain theproducts and, perhaps more importantly, commu-nicate the benefits to potential users.


48

Developing value-added services can benefitboth the bureaus and their customers and ulti-mately may improve access to finance for thebroader community.The opportunities, challenges,and ensuing benefits, however, will vary consider-ably depending on a bureau’s individual circum-stances and the market in which it operates.

The following sections describe in detail someof the core value-added products credit bureausoffer.

3.3 ProductsThe following list, although not inclusive of all ofthe value-added products credit bureaus provide,serves as a guide to the key services typically avail-able. The accompanying examples indicate howthese products are deployed in certain marketsand may not be applicable to all circumstances.

3.3.1 Bureau ScoresA credit score is a number assigned to a borrowerbased on his ability and capacity to repay debt.This number falls within a range of scores, and ahigher score indicates a more creditworthy bor-rower.This score is computed from available cred-it history information using a statistical model ormathematical algorithm. Credit scores can be usedin the loan approval process for simpleaccept/reject rules or for more sophisticated risk-based pricing rules and credit limits.

Bureau score refers to credit scores developedon the basis of the credit bureau data and are dif-ferent from the credit scores developed on thebasis of the data supplied by an individual lender.Bureau scores are based on the informationpooled across many creditors as well as publicinformation sources and thus include characteris-tics otherwise unavailable to the individual lender,such as total exposure, number of outstandingloans, and previous defaults within the system.Allof these are highly predictive measures of futurerepayment.

Credit bureaus typically build scores using threehistorical data files that are unique to the creditbureau:

� Defaults on previous credit transactions.� Positive payment behavior (trade line data).� Previous searches/inquiries.

In certain circumstances, the models mayinclude other types of data, such as:

� Third party data, e.g., court judgments and bank-ruptcies.

� Demographic data, e.g., applicants’ personalattributes, such as age.

� Geodemographic data, aggregated informationat the geographic level.

Each of these components could potentially addpredictive power to a bureau score, but care mustbe taken to ensure that the resulting models do notconflict with a lender’s existing decision-makingprocess. For example, a credit score that incorpo-rates the customer’s age may be incompatible witha lender’s custom scorecard that also includes age.Typically, therefore, a credit bureau may choose todevelop a suite of models rather than just onemodel to accommodate as many different customerrequirements as possible. Examples follow:

� Positive bureau score for closed user groupmembers providing both positive and negativedata and typically used as a plug-in or addition toin-house custom scores.

� Negative bureau score for closed user groupmembers providing only negative information.

� Enhanced bureau score incorporating additionalcustomer demographic data and typically usedon a stand-alone basis by lenders with no otherscoring models.

� Industry-specific bureau scores using dataderived from specific industry sectors, such asbanking or telecommunications.

� Public domain bureau score using data availablein the public domain and, therefore, available toall customers.


49

Because different users can use the scores fordifferent purposes, the credit bureau typically usesa variety of different distribution channels. In itssimplest form, the credit score can be incorporat-ed into a credit report, usually with some explana-tion as to what the score means.Alternatively, thebureau may supply the score to the users electron-ically so that it can be incorporated into cus-tomized scoring solutions or automated softwareapplications.A third and increasingly popular serv-ice is the use of a regular batch service thatrescores complete portfolios periodically. Thecharging structure for each of these services alsovaries although most bureaus charge users on aper-score or per click basis.

When adequate quantities of reliable informa-tion are available, bureau scores can be statistical-ly derived, typically by using some form of multi-variate regression analysis.The techniques used todevelop the models are very similar to those usedfor any other type of customized model develop-ment. There are, however, several unique chal-lenges that can complicate the process of build-ing/deploying bureau models:

� RReettrroossppeeccttiivvee DDaattaa.. A key requirement of theanalysis is the ability to observe the transition ofa credit file from the point at which an applica-tion was made, through the observation period,to the outcome point. This requires that thebureau be capable of retrospectively recon-structing a credit file at various points in time.With adequate archiving of the database, recon-struction may not seem like a significant issue.Changes in customer name, address, ID num-bers, and the like can cause tracking problems,however, if not appropriately addressed.

� TThhiinn FFiillee.. The data files may range fromextremely detailed, as when a data subject has avariety of pre-existing credit facilities with vari-ous outcomes, to very thin data, as when thebureau has no pre-existing information on theapplicant. In cases when a bureau has only a lim-ited amount of data on borrower performanceand outcomes, standard statistical multivariateanalysis may not apply and other methodsshould be used.

� SSccoorriinngg MMooddeell CCaalliibbrraattiioonn.. The bureau buildsthe credit scores from a broad spectrum of cus-tomer histories found in its database. Thederived scores are typically calibrated for anaverage portfolio; that is, the distribution of cus-tomers across the range of scores reflects whatis seen across the whole spectrum of customersat the bureau.While probability of default at anygiven score should remain constant for all users,the cumulative good to bad odds will vary fromportfolio to portfolio depending on the risk pro-file of the applicant base. This can have a pro-found effect on the way lenders manage theircut-off strategies (the scores at which the lenderchooses to accept or decline applicants). It ishighly recommended, therefore, that individualportfolios be retrospectively tested before themodels are implemented.

In emerging markets where either the market istoo small or the credit bureau is insufficientlymature to have confidence in the data, the bureaumay consider offering models that rely more heav-ily on customer demographic characteristics thanon credit performance data.Although less predic-tive, these models often provide a useful introduc-tion to the methodology for lenders with little orno previous experience in credit scoring.

3.3.2 Software ApplicationsA key advantage of credit scoring is the bureau’sability to establish a quantifiable measure of risk inwhat is otherwise a highly subjective process.Having a numeric value (a measure of probabilityof default) for risk is valuable in its own right butbecomes increasingly powerful when integratedinto automated processes and used to pro-activelymanage strategy and a lender’s appetite for risk.

To help facilitate this process, many creditbureaus in mature economies have developed arange of software solutions that complement boththe raw bureau data and the scoring processadopted by sophisticated lenders.These solutionsare commonly provided either as software applica-tions—customized to specific user requirementsand maintained within the client’s own IT environ-ment—or as bureau solutions—more generic in


50

nature and hosted at the bureau.The available solu-tions are many and varied, but the following repre-sents a summary of the more popular applications.

� AApppplliiccaattiioonn PPrroocceessssiinngg

A key driver of profitability in mass market lend-ing environments, such as consumer loans andcredit cards, is the ability to keep the cost ofnew business acquisition to a minimum. Manyfinancial institutions have turned to automatedapplication processing systems as a means ofstreamlining the credit-granting process. Manyexamples of such systems exist, but the com-mon design incorporates several fundamentalfeatures:

EElleeccttrroonniicc DDaattaa CCaappttuurree.. Typically an applica-tion processing system has a series of standard-ized data capture screens. These screens allowthe operator to capture the information neces-sary to process the decision and, perhaps moreimportantly, store the customer data in a formatthat can later be used for analysis.

RRuullee//SSccoorriinngg EEnnggiinnee.. The system captures theapplication data electronically, then the softwareautomatically applies policy rules, such as mini-mum required lending criteria, and scoring algo-rithms, including score cut-off criteria.

DDeecciissiioonn OOuuttppuutt.. An automated application pro-cessing system assimilates all of the input data,including any available on-line information fromthe credit bureau; applies the rules and scoringmodels from the decision engine; and presentsthe operator with a recommended course ofaction, such as accept, refer, or reject. This out-put is then queued so that the final decision ispresented to an individual with the appropriatelevel of underwriting authority.

The degree of complexity of such software solu-tions varies depending on the technical sophis-tication of the user. Advanced decision systemsare capable of managing almost all aspects ofthe decision-making process, including cus-tomer segmentation and strategy allocation

(e.g., terms, limits, and product features) andeven champion/challenger strategy setting totest the lender’s appetite for risk.

� BBeehhaavviioorraall SSccoorriinngg ((CCaarrdd MMaannaaggeemmeenntt SSoolluuttiioonnss))For a variety of credit products, such as creditcards, charge cards, and overdrafts, the initialdecision whether or not to lend is only the firstof many decisions that must be taken during thelife of the lender-borrower relationship. Thesedynamic products require a greater degree ofmonitoring than term loan products as theexposure to risk increases over time.Additionalcredit decisions must be taken on a variety ofissues, such as limit management, over-limitauthorizations, and card reissue.

Behavioral credit scoring is an adaptation ofmore traditional scoring techniques specificallydesigned to observe and evaluate the paymentbehavior patterns of borrowers. The outputscore changes to reflect the changing risk pro-file over time and can be used either to auto-mate routine decisions or provide operatorswith an immediate assessment of current risk.

A range of powerful software solutions has beendesigned to host card management solutionsand provide strategic control over practically allaspects of customer relationship management.While the complexity of these systems has a cor-respondingly high price tag, these systems havebecome almost an integral part of mass marketcredit management.

� MMooddeell TTrraacckkiinngg aanndd PPeerrffoorrmmaannccee MMoonniittoorriinnggAn overlooked benefit of introducing credit scor-ing methodology into the lending process is theability to monitor customer risk in an objectiveand quantifiable manner. Undertaking this analy-sis requires an in-depth understanding of theway the models are performing. Several creditbureaus provide score diagnostic tools that mon-itor and report on the performance of scorecardcharacteristics in terms of their continuing abili-ty to discriminate and the way shifts in the appli-cant population may create misalignments thatwould affect the quality of the decisions.


51

3.3.3 Collections Services(Receivables Management)A long and often successful association has existedbetween credit bureaus and debt collection com-panies. In several instances, negative informationin credit bureaus has been derived directly frominformation gathered by debt collection compa-nies (e.g., Baycorp in New Zealand and CreditReference Bureau in East Africa).

Many different collections products and servic-es are available, but the following three are amongthe most common.

� TTrraacciinngg.. Tracing services products use the cred-it bureau data to identify the whereabouts of acustomer with whom a lender has lost contact(“Skips”). These products either trawl the cur-rent bureau databases to identify existing con-tact information of which the lender may beunaware (e.g., telephone numbers or a newaddress) or place a marker on the customer fileso that if the customer subsequently makesanother application for credit the previouslender can be informed.

� DDeebbtt MMaannaaggeemmeenntt.. Debt collection is an expen-sive and time-consuming function and typicallyrequires specially trained and dedicated person-nel. Some lenders, therefore, opt to outsourcethis function, and credit bureaus may performthis service. These services are usually per-formed on a fixed fee basis or on a performancebasis,where the collector gets to keep a propor-tion of any monies recovered.

� DDeebbtt PPuurrcchhaassee.. Credit bureaus that specializein receivables management may choose to takethe ultimate risk and buy distressed or non-per-forming accounts from the credit provider. Inthese circumstances, the bureau purchases theoutstanding balances from the lender at a dis-count, assumes responsibility for collecting thedebt, and keeps the proceeds once the debt hasbeen collected.

3.3.4 Asset RegistriesFor secured loans, a lender needs to establish thatthe collateral used for the loan actually exists andis unencumbered. Developed credit bureaus,therefore, often attempt to become more than justa source of credit data by providing customerswith access to associated lending information,such as asset registries. Bureaus can provide thisservice either by building an automated link to athird party database or by building and hosting theservice directly.Whether it is fixed assets, such asland and buildings, or moveable assets, such asmotor vehicles, these services typically providetwo basic functions:

� IInnqquuiirryy. This function allows users to ascertainthe bone fide nature of the asset and whether ornot there are any encumbrances prior to pur-chase or acceptance of the asset as collateral.

� RReeggiissttrraattiioonn ooff IInntteerreesstt.. This function allows thelender or individual to register a notice of acharge or lien on the asset.

3.3.5 Marketing ServicesThe use of credit bureau data, especially closeduser group data, for marketing purposes, is often ahighly contentious issue. In many countries,including Australia, use of such data is either pro-hibited by law or severely restricted to specificapplications. In many other countries, especially inemerging markets where lenders are already nerv-ous about sharing credit information, marketingapplications are intentionally excluded from thedefinitions of permissible purpose in either theindustry code of conduct or the membershipagreement between the bureau and its customers.

There are, however, several value-added mar-keting services that the bureau can provide thatdo not necessarily involve the use of creditbureau data.


52

The range of potential products/services thatcan be offered is extensive.The following list rep-resents a sample of the most common examples:

� CCuussttoommeerr pprrooffiilliinngg.. Historically, many financialorganizations have suffered from poor knowl-edge management systems (e.g., paper-basedcustomer records). Consequently, these organi-zations have relied heavily on branch distribu-tion channels to obtain comprehensive informa-tion about their customers. Customer profilingattempts to bridge this knowledge gap by pro-viding analytical services that help to profile theattributes of particular types of customers.Thisservice may include augmenting the lender’sexisting customer information with additionaldata from the credit bureau. The subsequentanalysis identifies homogeneous customer clus-ters or segments that have similar profiles, suchas young, credit-active high achievers, that canthen be used to help the financial institutioneither provide a more tailored relationship orbetter target cross-sell and up-sell promotions.

� MMooddeelliinngg.. As with credit scoring, the number ofapplications for modeling services is extensive.Among the more popular are propensity model-ing and response modeling. Propensity model-ing tries to predict the likelihood that a particu-lar prospect will take up a marketing offer;response modeling measures the effectivenessof particular marketing campaigns to increasethe responsiveness of customers in the futureand thereby optimize the cost of new businessacquisition. More complex forms of modelinginclude applications such as customer worth orcustomer life-time value. These techniques ana-lyze customer potential not only in terms ofactual, current contribution/profit but also interms of what a customer may contribute overthe lifetime of the relationship.

� GGeeooddeemmooggrraapphhiicc AAnnaallyyssiiss.. Geodemographicmodeling looks at the relationship between geo-graphical areas, indicated by zip codes or postalcodes, and the types of individuals/businessesthat live/work in a given area.The technique cre-ates similar customer profiles to those describedabove but does so using aggregated rather thanindividual data.

� LLiisstt SSeerrvviicceess.. In countries that have a maturedirect marketing industry, many credit bureaushave developed products and services to assistwith customer prospecting. These servicesrange from providing prospect lists (e.g., thenames and contact details of potential cus-tomers) augmented with credit bureau data orgeodemographic data, to the outsourced man-agement of a client’s customer relationship man-agement database.

� MMaaiill SSccrreeeenniinngg.. Again, in countries that usedirect mail extensively as a means of acquiringcustomers, the credit bureau can be useful inhelping ensure efficient targeting of potentialcustomers. Mail screening removes from a mail-ing list those applicants who are most likely tobe rejected for an offer of credit if they were toapply.This screening saves the lender time andeffort. This service also has positive customerbenefits in those countries that operate a do-not-mail database—a screening facility for con-sumers that would prefer not to receive unso-licited marketing offers.

Where marketing services are permissible (e.g.,in the United States and United Kingdom) and areused extensively, they have proven to be a highlylucrative form of added value for the credit bureauand a significant value-added proposition for theuser.These services also have a positive effect onthe risk management process of the bank by allow-ing the bank to pre-screen the offers.


53

3.3.6 Portfolio MonitoringMonitoring and maintaining credit quality is a taskthat all lenders undertake but one that has takenon more prominence in recent years with theupcoming introduction of Basel II.

Some credit bureaus have been providing serv-ices in this field for many years, using a range ofstandard reporting and bureau scoring products.

� PPoorrttffoolliioo MMoonniittoorriinngg SSeerrvviicceess.. These servicesadvise a lender of any significant change to acustomer’s credit file, such as a default regis-tered by another lender.

� BBaattcchh SSccrreeeenniinngg.. This service allows lenders toperiodically update the risk profile of entireportfolios by reviewing the current creditscores of its clients.

� MMoonniittoorriinngg aanndd RReeppoorrttiinngg.. These services typi-cally help smaller lenders with limited internalanalytical capacity to produce the managementinformation required to track credit quality.

The forthcoming introduction of the Basel IICapital Accord and the need for lenders to complywith the best practice risk management guidelineshave created an increased focus on the ability oflenders to monitor portfolio quality. Implementingthe Advanced Internal Ratings Based approachrequires all lenders to be capable of calculatingnot only Probability of Default but also Loss GivenDefault and Exposure at Default. Credit bureauswith developed analytical capabilities have seizedthis opportunity to use advanced modeling, soft-ware solutions, and consultancy to help theirclients with these compliance issues.

3.3.7 Fraud DetectionAs the retail credit market grows in an economy,sowill the incidence of fraudulent financial transac-tions. Fraudulent activity can range in severityfrom what is sometimes referred to as soft fraud—embellishing application information to obtaincredit—to more hard forms of fraud, such as iden-tity theft.

A variety of products and services can be devel-oped on the back of the bureau platform to helplenders identify and thereby prevent fraud.These prod-ucts include,but are not limited to, the following:

� FFiillee CCrroossss--rreeffeerreenncciinngg.. These relatively simpleproducts cross-reference various data files toidentify anomalies.

� KKnnoowwnn//SSuussppeecctt FFrraauudd CClloosseedd UUsseerr GGrroouuppss..These industry initiatives, such as the CreditIndustry Fraud Avoidance Scheme (CIFAS) in theUnited Kingdom, pool information aboutknown or suspected fraudulent activity.

� FFrraauudd SSccoorriinngg.. This product may be custombuilt models for individual institutions or gener-ic models developed by the credit bureau.

� FFrraauudd DDeetteeccttiioonn SSyysstteemmss.. These sophisticatedsoftware solutions use a combination of ruleslogic, scoring, and enhanced databases to identi-fy application fraud. A range of software solu-tions have also been developed specifically totrack card fraud by means of payment behavioranalysis.


54

4Legal andRegulatory Framework

Legal and Regulatory Framework

55

4.1 Key Principles for CreditInformation-SharingLegislation/RegulationThe legal framework for credit reporting differsfrom country to country and may include a combi-nation of the following legal acts:

� Credit Reporting Laws� Data Protection Laws� Consumer Protection Laws� Fair Credit Granting and Consumer Credit

Regulations� Personal and corporate privacy and secrecy

provisions

Bank secrecy provisions in the law or contractu-al confidentiality provisions are often cited as animpediment to the development of a credit bureau.In accordance with these provisions, banks are notallowed to disclose information related to theaccount and transactions of its client to a thirdparty.Analysis of 84 countries in all regions of theworld shows that in 62 countries with explicitbank secrecy provisions, roughly 63 percent or 39countries had an operating private credit bureau.Figure 25 shows the number of countries withoperating credit bureaus in the countries withexplicit bank secrecy provisions in the law. In high-income Organization of Economic Cooperationand Development (OECD) countries, 11 out of 12countries with bank secrecy provisions had a pri-vate bureau. In Latin America, seven out of eightcountries operated credit bureaus. In Eastern

Europe, to date 14 out of 22 countries have a cred-it bureau. Most bureaus in the region were estab-lished within the last five to seven years and all ofthem in countries with bank secrecy provisions.

Legal and regulatory approaches vary acrosscountries, but explicit borrower consent generallysuffices to enable a bank to share information.Debate surrounding the issue usually revolvesaround the format in which consent should besubmitted and the way banks should store suchconsent. Passing a credit reporting law or a regula-tion generally helps resolve these issues. In theabsence of such laws, however, a simple agree-ment among lenders to collect consent and shareinformation will suffice. In Switzerland, for exam-ple, a credit bureau has been operating since 1974merely on the basis of borrower consent.29

Although bank secrecy and confidentiality provi-sions do not prohibit the sharing of credit informa-tion, lack of clarity and agreement among stake-holders on the most appropriate way to addresssuch issues usually impedes the establishment ofcredit bureaus. In several countries, such as Russia,Kazakhstan, and Ukraine, credit bureaus were notset up until clarifying legislation was passed. Inother countries, such as Kenya, Tanzania, andGeorgia, the debate continues on the most appro-priate manner in which to share credit information.


The main objective of legislation to enable cred-it reporting is to balance the ability of institutionsto exchange credit information in the normalcourse of business while simultaneously protect-ing individuals’ rights to privacy. Two broadapproaches to the regulation of credit reportingcan be identified: 1) use of broad data protectionlaws and 2) use of specific credit bureau or creditreporting laws. The European Union and severalother European countries regulate credit report-ing activities under broad data protection lawsthat cover not only credit bureau activities but alsoany other relationships and transactions related todata management and exchange. Recently, severalemerging markets have followed this approach:Chile passed a data protection law in 1999, andArgentina passed a similar law in 2000.

Regulating credit reporting activities through aspecific credit bureau or credit reporting law isanother approach. Countries that have adoptedthis type of approach include the United States,Thailand, Russia, Kazakhstan, Peru, and Ukraine.See Annex 2 for further information on existingcredit reporting and data protection laws.

Good enabling legislation incorporates the fol-lowing characteristics:

� OOppeenn ssyysstteemm.. Reporting and access are open toboth financial institutions and non-bank credi-tors, such as retailers, telecom and utility compa-nies, and debt collectors.

� PPeerrmmiissssiibbllee ppuurrppoossee.. The legislation protectsthe rights of individuals and firms to ensure thatdata are not misused, while permitting the shar-ing of information.Typically, access to informa-tion is only allowed for a certain identified pur-pose, such as credit approval, portfolio monitor-ing, debt collection, and employment. Manycountries require borrower consent in order fora creditor to access information in a bureau.

� AAuutthhoorriizzeedd aacccceessss ttoo iinnffoorrmmaattiioonn.. Only author-ized parties may access information in the cred-it bureau, and information may only be used forpermissible purposes. The legislation mayrequire borrower consent or notification toallow the lender to access information.The leg-islation may also require that the names of alllenders that have accessed data be a part of thecredit report available to the borrower so as toensure that no information is accessed withoutborrower consent.


56

Nu

mb

er o

f C

ou

ntr

ies

1

1211

6

1

87

22

14

8

56

FFiigguurree 2255:: CCrreeddiitt BBuurreeaauuss iinn RReeggiioonnss wwiitthh BBaannkk SSeeccrreeccyy LLaawwss

� PPoossiittiivvee aanndd nneeggaattiivvee iinnffoorrmmaattiioonn--sshhaarriinngg..Lenders are generally hesitant to share positiveinformation for fear that their competitorsmight poach their best customers once positiveinformation is made available. Enabling legisla-tion, however, can include information-accessrules that restrict the ability of banks to poachother banks’ customers. For example, the legisla-tion may specify that a lender can only access abureau’s information if an individual or a firmhas applied to the lender for credit.

� CCoonnssuummeerr pprrootteeccttiioonn.. Regulations include pro-visions ensuring that individuals have the rightto check their own information and bureaushave mechanisms for correcting erroneousinformation. The regulation also creates griev-ance and dispute-resolution mechanisms thatinclude limits on the time the bureau may taketo respond to a borrower’s complaint. In mostcountries, this period ranges between 10 and 20business days. During this period, the bureaumust put a notice into the credit report indicat-ing the dispute.The law, however, should be rea-sonable and specify the penalties to be imposedon a credit bureau and/or lenders in case of vio-lations. In Thailand, a law was passed with verysteep penalties for violations such as data inac-curacies or the lack of borrower consent. Sincethe law did not clearly define the requirementsfor data accuracy and the procedures for obtain-ing consent, the two existing credit bureaus inthe country shut down for fear of having to facesevere liabilities. It took five months for theauthorities to issue clarifying regulations and forthe industry to adjust its processes before thebureaus could reopen.

� LLiicceennssiinngg aanndd rreeggiissttrraattiioonn.. In Mexico,Thailand,India, and Kazakhstan, credit bureaus arerequired to obtain a license to operate from thesupervisory authority. Licensing requirementsusually require credit bureaus to meet certainfinancial, security, and governance standards. Inmost countries where specific credit bureau

� CCoonnsseenntt.. Depending on the jurisdiction, explicitor implicit individual borrower consent may berequired to provide data to the bureau and toaccess a credit report.The objective of the con-sent is to enable the data subject to control theinformation flow. Several European Union coun-tries, Thailand, Kazakhstan, and many others,require explicit borrower consent to provideinformation to the bureau. The loan agreementusually includes standard clauses to cover suchconsent. Many countries also require borrowerconsent for the creditor or another eligible thirdparty to access a credit report. Loan or employ-ment applications usually contain consent clause.In the interest of maintaining operational efficien-cy,the onus of obtaining and maintaining a recordof borrower consent for data submission rests onlenders. In the event of a dispute, the lender mustbe able to demonstrate that it had obtained bor-rower consent in accordance with the law. Somecountries, including the United States, do notrequire explicit consent.The consent of the bor-rower is considered implicit if the borrower hasoriginated a transaction with the lender. Banksare required to inform their customers, however,on the use of the information that they obtainfrom their clients.

� LLeennggtthh ooff iinnffoorrmmaattiioonn rreetteennttiioonn.Legislation stip-ulates a specific length of time for allowinginformation to be stored. Although historicalinformation enables lenders to assess a borrow-er’s credit quality over a period of time, the leg-islation should specify a cut-off date for informa-tion storage, after which time information iserased to give borrowers a fresh start. Paymenthistory information is usually maintained for aminimum of five years.Rather than erasing infor-mation on defaults once loans have been repaid,this information should be stored with the restof the borrower’s file for the assigned period oftime. Public records relating to bankruptcy areusually retained for seven years or more.According to a World Bank survey, out of 78 pri-vate credit bureaus, 57 preserved historicalinformation for more than five years, and 34credit bureaus preserved data between five andseven years.30


57


bringing class action suits in case of systematicviolations.This type of enforcement mechanismis prevalent in the United Kingdom, Hong Kong(China),Australia, and South Africa.

To ensure the smooth implementation of creditbureau legislation, it is critical to build the capaci-ty of the supervisory authority. In several cases,implementation of the law was delayed or had asignificant negative impact on existing bureausdue to the lack of enforcement capacity. For exam-ple, in Russia a credit bureau law was passed inDecember 2004. The law required all financialinstitutions to submit information, to a registeredcredit bureau after obtaining borrower consent.Asupervisory authority, however, was appointedafter much delay and could not develop a registra-tion procedure in time.As a result, the implemen-tation of the law had to be postponed by morethan one year. Due to the lack of supervisorycapacity, the authority also was unable to provideguidance to financial institutions on compliancewith the law.

4.2 Self-Regulation PrinciplesIn Australia, the United Kingdom, South Africa, andHong Kong (China), credit bureaus operate on thebasis of a Code of Conduct (COC) under broad pri-vacy legislation. CoC is a binding agreementsigned by the members of the bureau. It providesthe rules that govern the operations of a creditbureau and mechanisms for resolving disputesamong its members. In countries where more thanone bureau exists, such as in South Africa, all cred-it bureaus operating in the country endorse theCoC, and an association of credit bureaus isresponsible for enforcement

The CoC usually covers the following key areas:1) bureau operation principles; 2) bureau rightsand obligations; 3) member rights and obligations;4) data subject’s rights and obligations; and 5) dis-pute resolution mechanisms.

The basic bureau operation principles include:

� Reciprocity—the requirement that all membersprovide data in order to obtain access to theinformation in the bureau.

laws have been passed recently, the centralbank, the bank superintendent, or the supervi-sor of a non-bank financial institution can per-form the licensing function. In Russia and someEuropean Union countries, the credit bureau isrequired to register with the supervisory author-ities but does not need to obtain a license tooperate. Other countries, such as the UnitedStates, require neither licensing nor registration.

In case of severe violations of data sharing regu-lations or laws, the supervisory authority canpenalize the bureau in question by revoking itslicense or registration. In the absence of licens-ing or registration procedures, any data viola-tions need to be resolved via the country’s courtsystem.This approach works well in countries,such as the United States, where the judicial sys-tems are well developed and consumers can filea class action law suit in case of systemic viola-tions by credit bureaus. Most emerging coun-tries, however, have weak judicial systems andopt for a strong regulatory agency instead.

� AAcccceessss ttoo ppuubblliicc iinnffoorrmmaattiioonn.. Public informa-tion access is open to all market participants atlow or no cost.

Enforcement

Enforcement is an essential element of the legaland regulatory framework necessary to enable theoperation of the credit information industry. Acountry may opt for one of the following twobroad approaches to enforcement depending onits legal traditions:

� AA ssttrroonngg ssuuppeerrvviissoorryy aauutthhoorriittyy with the powerto license, register, and control credit bureaus.The authority’s functions usually include issuingindustry regulations, granting licenses, conduct-ing or requesting audits, receiving and analyzingcomplaints, and imposing penalties. Mexico,Kazakhstan, and Thailand use this approach.

� IInndduussttrryy sseellff--rreegguullaattiioonn within an establishedlegal framework. Here the enforcement authori-ty’s role is limited to issuing clarifying state-ments, collecting complaints, and in some cases


58

� Data format—the specification of the format inwhich data are submitted and distributed.

� Frequency—the schedule for data submissionand updates. It is customary to have updates ona monthly basis with the exception of defaultrecords, which need to be updated only if thereis a change in the status of the credit.

� Quality—the requirement that the informationbe accurate, complete, current, and provided ontime.

Table 5 identifies the respective rights and obli-gations of a credit bureau and its members anddata subjects.

The CoC also specifies responsibilities ofbureaus and members in the areas of investigationand correction of erroneous information, investiga-tion and resolution of complaints, and impositionof penalties where appropriate.

Examples of CoC provisions related to investiga-tion into disputed information are as follows:

� A data subject or a member may notify thebureau in writing that the completeness oraccuracy of any item of information is disputed.

� The bureau shall investigate such disputed infor-mation and provide response to the informationsubject within a specified period of time, usual-ly between five and 15 business days.

� The bureau must ensure that informationregarding a consumer is updated and must veri-fy the information with the member that hassupplied the information.

� The bureau shall correct the information regard-ing the data subject if the complaint is substan-tiated in the course of investigation.

� For the duration of the investigation, the creditrecord in the bureau must indicate that theinformation is under dispute.


59

BBuurreeaauu’’ss rriigghhttss � Record, maintain, collate, synthesize, and/or process information properly aanndd oobblliiggaattiioonnss and accurately.

� Protect information against loss and damage.� Protect information against unauthorized access, use, modification, or

disclosure.� Retain and display information for the relevant periods.� Grant access to own credit reports to individuals who offer proof of identity.� Maintain a help desk.

BBuurreeaauu MMeemmbbeerrss’’ � Comply with reciprocity principles.rriigghhttss aanndd � Restrict inquiries to those allowed by law.oobblliiggaattiioonnss � Maintain records and be able to demonstrate that the query was made for a

permissible purpose.� Use information only for permissible purpose.� Disclose information obtained from a bureau only to authorized parties.� Secure information obtained from the bureau.� Appoint a bureau relationship manager.

DDaattaa ssuubbjjeeccttss’’ � Access own credit report.rriigghhttss aanndd � Dispute inaccurate information and obtain response within a set period oobblliiggaattiioonnss of time.

TTaabbllee 55:: RRiigghhttss aanndd OObblliiggaattiioonnss ooff CCOOCC PPaarrttiieess

With respect to breaches of the code and mech-anisms for resolving them, the CoC usually speci-fies that any member may submit a complaint ofviolation of the CoC by another member.Generally, a compliance committee comprisingthe members of the bureau and/or independentexperts investigates such complaints.This commit-

tee, appointed in consultation with the Board ofthe bureau, the members, and the supervisoryagency for the law governing the activities of thecredit bureau, has the right to impose penalties,such as suspending access to the registry by amember that systematically violates the Code.


60

5Case Studies

Case Studies

61

CompuScan, South Africa:Successfully ServingMicrolendersCompuScan was established in 1994 to providecredit bureau services to microlenders in SouthAfrica. Its owners founded the company inresponse to increasing levels of bad debts amongwhat appeared to be habitual nonpayers. WhileSouth Africa had a well-developed credit bureauinfrastructure at that time, the industry had tradi-tionally ignored the microfinance market becausethe cost to service an almost entirely paper-basedclient base was considered unviable.

The CompuScan bureau started life as a collec-tive of microfinance providers in the Cape Townarea that shared a negative list,updated on a week-ly basis and distributed as a simple Excel spread-sheet.As the value of the service became apparentto users, the geographical coverage expanded andthe bureau introduced a range of value-enhancingpropositions. A key component of CompuScan’sunique offer has been the provision of training tomicrofinance lending officers through the devel-opment of a specialized credit academy.

Today CompuScan offers a broad base of creditbureau products, incorporating both positive andnegative data, and serves more than 3500 creditofficers (branches) throughout South Africa. Allcredit data are accessed and loaded via a high-functionality Internet GUI, in batch format files orfully integrated into the in-house systems of themicrolenders and banks.CompuScan also providesa variety of other services, such as notifyinglenders when customers take loans or receivecourt judgments. These services are seen as an

integral part of allowing small lenders to buildcredit histories, thus facilitating their transition toformal borrowing and supporting broader accessto finance in the market.

In 2004, CompuScan undertook analyticalresearch into the effectiveness of the then availablegeneric bureau scores for the microfinance sector.Not surprisingly, given that these models had beenderived from activity within a completely differentmarket sector, the models appeared to be providingunacceptable levels of discrimination. CompuScan,with technical assistance from IFC, began work,therefore, on developing the first microfinance-spe-cific bureau scores for South Africa.

In 2005, CompuScan partnered with LISIM, aColombia-based credit scoring and analytical solu-tions company, to develop a suite of credit riskmodels using a combination of unique datasources from CompuScan’s database.These servic-es were launched in mid-2006 and are expected tohelp microlenders improve both the efficiency ofthe credit approval process and the quality ofcredit decisions.

CompuScan is an emerging regional player,withexisting operations in Namibia and Botswana andfurther plans to expand to other countries in theregion. The combination of using highly flexibletechnology and the focus on user developmentthrough training have clearly demonstrated thatservicing relatively low-value credit transactionscan be both profitable and sustainable.


62

cross-border private credit bureau enables thedelivery of standardized products and services thathave superior information quality.

With the support of IFC,TUCA’s primary focushas been on various educational initiatives, includ-ing demonstrating the importance of data contri-bution and full-file reporting from financial institu-tions and the positive effect they can have on theeconomy.In addition, through international confer-ences on credit information and credit scoring,along with roundtables and videoconferences,TUCA has raised the awareness and highlightedthe advantages of private credit bureaus, thus, facil-itating discussions among the various parties.

Going forward,TUCA intends to provide value-added services, such as credit scores and fraudalert systems, all of which will enhance the capac-ity of financial institutions to manage risk.

Case Studies

63

TransUnion Central America(TUCA): Building a RegionalSolutionDue to economies of scale in the credit informa-tion industry, small economies are often unable toattract state-of-the-art credit informationproviders.The lack of credit bureaus in Costa Rica,Guatemala,Honduras,and Nicaragua was imposinga constraint on credit growth among consumersand small businesses.

A regional private credit bureau can be a viablealternative to a national bureau in a small market.TransUnion in Central America (TUCA), estab-lished in 1999, is the private credit bureau thatprovides services to Guatemala, Honduras, ElSalvador, Costa Rica, and Nicaragua. TUCA’s busi-ness model is based on a hub-and-spokes system(with Guatemala as the hub and the other fourCentral American countries as the spokes). Thespokes leverage the more advanced technologicalsystem present in the hub, thus enablingeconomies of scale, improved efficiency, and high-er profitability. In addition, the creation of a single

In the absence of a legislative framework to sup-port the establishment of a credit bureau, SAMAand the participating banks agreed to operateunder a voluntary Code of Conduct until theappropriate legislation was drafted and passed bythe Sharia council. The relevant legislation hasbeen drafted and is expected to be enacted by theend of 2006.

Currently, SIMAH contains records relating toapproximately four million borrowers. This num-ber will grow substantially in the near future asmajor telecommunications providers have agreedto join the bureau and submit their customerrecords.The bureau provides in excess of 120,000individual credit reports per month.

To better serve the needs of lenders in the KSAand to expand credit to small businesses, SIMAH iscreating a commercial reporting business to com-plement its retail credit bureau operation.Expanding the bureau’s service to include com-mercial credit reporting will fill the existing infor-mation void. SIMAH is currently going through acomprehensive RFP process to select an appropri-ate technical partner to help with this expansion.The commercial phase of the credit bureau isexpected to be operational by mid-2007.


64

SIMAH, Saudi Arabia: Long-term StakeholderCommitment SIMAH, the Saudi Arabian credit bureau, beganoperations in 2004 and is jointly owned by 10banks. The Saudi Arabian Monetary Agency(SAMA), the Kingdom of Saudi Arabia’s (KSA) cen-tral bank, spearheaded discussions before thelaunch of the credit bureau.Although these discus-sions were launched in the mid-1990s, it was notuntil 1998 that SAMA facilitated the establishmentof a committee to work on the consumer creditbureau. The committee consisted of representa-tives from the banks as well as SAMA. In 1999, aRequest For Proposal (RFP) was issued to interna-tionally renowned technical vendors, but theprocess was soon abandoned. The process wasrevived in 2001 when SAMA issued a circular to allbanks outlining its position. SIMAH was estab-lished shortly thereafter.

SIMAH engaged the services of Accenture tohelp select the technical vendor. A Request ForInformation (RFI)/RFP process ensued, andBaycorp Advantage was selected as SIMAH’s tech-nical partner. SAMA was very supportive duringthe entire process, acting as a banking regulatorand appointing one of its key executives as man-aging director of the credit bureau.

Estealam recently contracted with Dun &Bradstreet International/D&B SAME to providesoftware solutions and operational know-how forthe creation of its database.

The creation of the bureau itself is divided intothree phases.The initial phase will focus on popu-lating the bureau’s database with credit card infor-mation in English, thus allowing bureau operationsto begin.At the same time, Estealam will work onthe data quality challenges identified in other assetportfolios. Examples of the issues to be addressedinclude the lack of unique IDs in Egypt, the record-ing of names and other information in English orArabic, and inconsistent formats across all lendersfor recording names, addresses, and dates of birth.Once the database has been populated, it will con-tain information from seven major banks and theSocial Fund for Development, which representapproximately 60 percent of national bank lending.

In phase two, Estealam will invite second-tierbanks and non-bank financial institutions to jointhe bureau once they are in a position to uploadtheir data in a format consistent with that of thebureau.

In the final phase of development, the bureauwill focus on the provision of value-added solu-tions, such as scoring, identity verification, frauddetection, and application processing.

Case Studies

65

Estealam, Egypt : Building theFirst Private Credit Bureau inEgyptEstealam, Egypt’s first private credit bureau, wasestablished in September 2005 and will includeinformation on both consumers and SMEs. Jointlyowned by 27 commercial banks and the SocialDevelopment Fund, each of which holds equalstakes, Estealam is wholly supported by theCentral Bank of Egypt.The bureau is still in its pre-liminary stages of operation and is expected to befully functional in the second quarter of 2007.

The Central Bank of Egypt was highly instrumen-tal in creating a legislative framework conducive tothe operation of a private credit bureau.The legisla-tion allows data sharing by banks and non-bankfinancial institutions. In addition, obtaining borrow-er consent is only mandatory for non-bank financialinstitutions. The legislation requires that all banksobtain a credit report for credit approval.

Estealam requested technical advisory supportfrom the IFC for the development of the bureau.IFC Global Credit Bureau assisted the bureau inthe following:

� Conducting a scoping and technical feasibilitystudy, which included a review of the legislativeenvironment and of the technical ability andwillingness of a selected number of banks andthe Social Fund for Development to contributedata to the bureau.

� Preparing and issuing an RFP to select potentialinternational vendors and providing credit bureauexpert assistance to the evaluation committee.

� Preparing a business plan for Estealam.

The World Bank and IFC recommended that thedevelopment of the consumer credit-reportinginfrastructure in Vietnam be undertaken as a pri-vate sector initiative with the strong support ofthe SBV. At the same time, CIC should be betterintegrated into the banking supervisory process.

IFC has worked closely with SBV to prepare adetailed strategy plan that lays out a roadmap fordeveloping the private credit bureau in Vietnam.The strategy envisages active participation of theSBV in developing an adequate legal and regulato-ry framework and assignment of the SBV as thesupervisor of credit bureaus. SBV’s role alsoincludes guiding and encouraging commercialbanks to actively participate in the establishmentof a credit bureau. To ensure broad buy-in andlimit fragmentation of the data in the early stagesof the credit bureau development, the Vietnameseauthorities have agreed that the best approach isto establish a private bureau owned by a consor-tium of banks and with participation of a techni-cal partner.

The SBV approved the strategy plan, and workon the legal and regulatory framework and thepreparation of the business plan for the creditbureau began in April 2005.


66

Vietnam : Public SectorSupport for the Establishmentof a Private BureauThe State Bank of Vietnam (SBV) operates a publicregistry called the Credit Information Center(CIC),which like most public registries is designedprimarily as a supervisory tool to identify systemicrisk in the banking industry. Initially the informa-tion captured by the CIC related predominantly tolarge exposures of banks to their corporate clients.More recently, the CIC expanded its role beyondsupervision to include sending information backto lenders by way of credit reports on potentialborrowers.

The increase in retail and SME credit in recentyears has created a need for an adequate informa-tion infrastructure to ensure that the growth ishealthy. SBV was requested to design a strategy forproviding the necessary information infrastruc-ture, including the future role of CIC and thepotential for private bureaus. SBV has asked theWorld Bank and IFC for support in defining thisstrategy.The key challenges identified through theanalysis of CIC operations and the credit market inVietnam include the following:

� CIC’s technical infrastructure cannot supportthe data volumes associated with a comprehen-sive consumer credit file.

� CIC does not have the requisite experience fordeveloping or operating a consumer creditbureau.

Annex 1: Maps

Annex 1: Maps

67


68

Annex 1: Maps

69


70

Annex 1: Maps

71


72

Annex 2: Countries with Credit

Information Sharing Laws

Annex 2: Countries with Credit Information Sharing Laws

73

CCoouunnttrryy NNaammee ooff RReelleevvaanntt LLaaww YYeeaarr PPaasssseedd

Albania Law 8517 on the Protection of Personal Data 1999Argentina Law 25326 on the Protection of Personal Data and Decree 2000, 2001

1558/2001Australia Australian Privacy Act 1998 1988Austria Austrian Data Protection Act 2000Belgium Law on Protection of Personal Data 1992Brazil Consumer Protection LawBulgaria Data Protection Law 2001Canada Privacy Act,The Personal Information Protection and Electronic 1983, 2000

Documents ActChile Supreme Decree No. 950 and Privacy Laws No. 19628 & 19812 1999Czech Republic Personal Data Protection Act 2000Denmark Act on Processing of Personal Data 2000Dominican Republic Credit Bureau Regulation and Protection of Individual Subject’s

Information Act 2005Ecuador Law on Credit Information Bureaus 2005Estonia Data Protection LawFinland Law on Personal Data (Data Protection) 1999France Act on Informatics, Data Files and Liberties 1978Germany German Data Protection Act 2002Greece Law 2472/97, Directive 95/46/EC, various decisions of the Various

Hellenic Data Protection Authority, Law 3259/2004 (Article 40)regarding the storage time of personal data in the CreditBehavior Databank

Hong Kong, China Personal Data (Privacy) Ordinance 1995Honduras CB incorporation and operations law 2005Hungary Act No. LXIII on the Protection of Personal Data and the 1992, 2003

Publicity of Data of Public InterestIceland Act Respecting Systematic Recording of Personal Data, 1989, 2001

Protection and Processing of Personal Data, No. 77/2000India Credit Information Companies (Regulation) 2004Ireland Data Protection Act 1988 and 2003 1988

Israel Law on Credit Data Service 2002Italy Protection of individuals and other subjects with regard to 1996

the Processing of Personal Data ActJapan Personal Information Protection Law 2005South Korea Act Relating to Use & Protection of Credit Information 1995Kuwait Royal Decree 2/2001 2001Latvia Law on Personal Data Protection 2000Lithuania Law on Legal Protection Of Personal Data of the Republic of 2000

LithuaniaMexico Law to Regulate Credit Information Societies 2002Netherlands Personal Data Protection Act 2000New Zealand New Zealand Privacy Act 1993 1993Nicaragua Private Credit Bureau Regulation 2005Norway Personal Data Act 2000Panama Law 24 that regulates credit information on consumers or clients 2002Paraguay Law 1682/01 and modified law 1969/02 on Data Protection 2001, 2002Peru Law 27863 for Private Credit RegistriesPoland The Act on Protection of Personal Data 1997Portugal Law 67/98 on Protection of Personal Data 1998Puerto Rico Fair Credit Reporting Act and the Fair and Accurate Credit 1992, 2003

Transactions ActRomania Law No 677/2001 on Protection of Individuals with regard to 2001

the Processing of Personal Data and the Free Movement ofSuch Data

Russian Federation Law on Information 1995Slovak Republic Personal Data Protection Act 1998Slovenia Personal Data Protection Act 1999Spain Organic Law 15/99 on the Protection of Personal Data 1999Sri Lanka Credit Information Bureau of Sri Lanka Act No.18 of 1990 1990Sweden The Credit Bureau Act 1998Switzerland Federal Data Protection Act 1992Taiwan, China Computer Processed Personal Data Protection Law 1995Thailand Credit Information Business Act, B.E. 2545 2002Ukraine Law on Information 1992United Kingdom Consumer Credit Act, Data Protection Act 1998United States Fair and Accurate Credit Transactions Act, Fair Credit Various

Reporting Act, Fair Credit Billing Act, Equal CreditOpportunity Act

Uruguay Law 17.838 on Protection of Personal Data 2004


74