Credit Information Sharing in Australia: A Roadmap for Transitioning

POLITICAL & ECONOMIC RESEARCH COUNCIL

by Michael Turner, Ph.D.

D&B Consumer Credit ConferenceMelbourne, Australia – October 14, 2008

Credit Information Sharing in Australia: A Roadmap for Transitioning

2

Introduction

Basics

Surprises, Choices and Pitfalls: 7 Lessons

Conclusion

Agenda

3

Introduction:Considerations for a Transition

Benefits of positive, comprehensive reporting over negative, fragmented well established

• Studies (statistical, simulations)• Experience (cases)

IFC’s Credit Bureau Knowledge Guide provides a vast wealth of information on credit bureaus, technical considerations,

Missing: what to expect in the transitions from non or negative reporting to positive reporting

4

Introduction

Basics


Conclusion

Agenda

5

The Importance of the Legal & Regulatory Framework

Laws and regulations over procedures:

•what are the rights of data subjects;• Access, notification, dispute resolution, redress

•what information can be collected;•what are acceptable uses of information sharing;•what are the data security requirements; and •obligations of credit bureaus, data furnishers, and data users

• Data quality, data verification, notification of adverse action, ID theft, fraud

They reflect a societal consensus about the underlying rationale for credit reporting

6

The Importance of the Legal & Regulatory Framework

They reflect a societal consensus about the underlying rationale for credit reporting

• Important for stability of system in the eyes of the public at large

• Consensus necessary for future changes in regulations that may

arise owing to changes in practices

• E.g. expansion of reporting to new categories of information new

permissible purposes, or the inclusion of new sectors.

7

The Importance of Education to Build a Societal Consensus

Education should occur at an institutional level

• Draft legislation should be vetted through the appropriate venues

• Quality and depth of consumer education will impact probability

of success of regime change

• Credit reporting to affect individual behavior

• Negative-only regimes have poor public perception

• Data furnishers must be educated to explain how

implementation will affect their operations

• Goal of education to provide tools to stakeholders to implement

operational changes in order to facilitate smooth transition.

8

Element of an Effective Legal & Regulatory Framework

Requirements regarding information contained in credit files:

•protection of consumer rights;•information privacy, limited and regulated access to consumer information;• use of public record information;• periodicity of reporting;•data expiration regulation;•provisions for the sharing of both positive and negative information;• noting disputed information or suspected fraudulent activity; and,•equal treatment of financial and non-financial industries that report

How each of these factors is addressed will vary by economy.

9

Elements of an Effective Legal & Regulatory Framework

Data Subject rights and protections

Data furnisher obligations

Credit bureau obligations

Control over third-party access to credit file

Regularly report accurate data

Data quality standards

Right to access credit file

(free disclosure)

Respond to consumer disputes in timely fashion

Disclosures to data subject

Procedures for consumer disputes and reverification

Duty to correct and update inaccurate information

Dispute resolution procedures

Redress for harms Duty to report status of account (open, closed, delinquent)

Data privacy and security, including third-party verification

Notification of adverse actions and dispute rights

Duty to respond to suspected ID fraud

Inclusion of financial, governance and security standards

10

Key Data Subject Rights: An Overview

The OECD’s fair information principles specifies some key rights. Limitations on:

o What is Collected o The Means of Collection o From Whom The Data Is Collected o Knowledge or Consent of the Data Subject o Scope of the Collection Limitation Principle

Data Quality Use Limitation (permissible purposes)

o Control Against Original Purposes o Exceptions o The Mechanism of Disclosure

Security Safeguards Openness Individual Participation

o The Right of Subject to Know of the Existence of Data o The Right to Access Data o The Right of Challenge, and Provision of Reasons for Refusal o Subject Challenge to Data

Accountability

11

Key Data Subject Rights: Dispute and Grievance

Four basic phases of grievance resolution: •Personal Information:

Credit bureaus must be able to immediately release information to consumers All information in the consumer file must be released

•Receipt of Grievance: a consumer contests the information Credit bureaus must provide consumers easy access to customer service Clear framework for the resolution of each case should be in place

•Credit bureau must have a system to verify data•Consumer rights in notice and follow up of grievance procedure

Notification of data subject of result of their case. Institute system of appeals when consumer refutes the resolution

In addition to safeguarding rights, improves data quality and system legitimacy

12

Technical Issues

Four domains: •Data formatting standard: common standards of reporting make it easier to collect and use information•Identity verification: helps match information, improve accuracy and protect against fraud•Data security: key to data integrity and fraud protections•Disaster recovery: preservation of the information and thus financial infrastructure

Methods and technical wherewithal varies by society and changes in technology, but the 4 issue areas are constant

13

Introduction

Basics


Conclusion

Agenda

14

Lesson 1: Changes in lending patterns, yes, but first the “valley of transition”.

15

Greater access: a host of studies indicate that widened credit access follows from greater information sharing, when information is used properly

Evidence fromUSCanadaBrazilColombiaArgentina

Expected Changes in Lending Patterns

16

Fairer access: these studies also indicate that credit access widens to under privileged groups (ethnic minorities, low-income segments, women)

Evidence fromUSColombia

Expected Changes: Distributional Fairness

Table 2: Change in the Acceptance Rate with Reporting Regime Change

US Full-File Colombia Full-File

(Neg.-Only = 1.00) (Neg.-Only = 1.00)

Ethnicity

Black 1.28

Hispanic 1.37

White 1.22

Age

18-251.47 18.31 (a)

26-35

36-45 1.22 6.48 (b)

46-55 1.21 4.54 (c)

56-65 1.203.85 (d)

>65 1.19

HH Income (000)

<20 1.36 (a)

20-29 1.3 (b)

30-49 1.24

50-99 1.21

>99 1.18

Gender

Female 12.39

Male 5.91(a) Actual Range is 18-32; (b) Actual Range is 32-42, (c) Actual Range is 42-50; (d) Actual Range is > 57.

17

Wider access & fair distribution are indicators of a better functioning lending system, resulting from information sharing…

…in the long run

Short run may involve potential surprises

Expected Changes as Monitoring Devices

18

Sometimes information sharing is followed by a contraction of credit access, i.e., it “gets worse before it gets better.” Logic:

When only negative information is shared, overextensions are hard to observe when borrowers utilize multiple lenders to service debt.

Only negatives reported here means over-indebtedness is NOT observed

Information sharing shows overextended customers. Potential consequences:

Banks reduce lending because of an uncertainty about the risk associated borrower

Banks reduce lending because of the need to cover defaults results from withdrawal of credit (can no longer rely on credit to service debt)

The Valley of Transition

19

Hong Kong: information sharing a response to rising credit card bankruptcies Bankruptcy filings rose from 893 in 1998 to 4,606 in 2000 to 25,328 in 2002. Opened credit bureau with positive information 2002. Info sharing reveals large numbers servicing debt with debt. Bankruptcies spiked as lenders became increasingly aware of borrowers

using loans to service other loans. Two-year recovery time--but lending is now broader, deeper, and more

sound.

BUT not seen in Brazil or Russia possibly used awareness to educate lenders Quickly reoriented lending to those NOT overextended Made adjustments to prepare for possibility (reserves, risk-based price)

The Valley of Transition: Hong Kong and Others

20

Lenders should be prepared for these contingencies

Near-term credit contraction should not be a “surprise”

Unhealthy lending results from extending more and more

accounts to existing borrowers

Lenders should consider lending to under-served but low-

risk segments of consumers.

The Valley of Transition: How to Prepare

21

Lesson 2: Small business benefits from positive reporting

22

Examination of impact of information sharing on (1) SMME lending, and (2) SMME loan performance based on bank estimations.

World Bank on SMME lending--examined more than 5,200 firms in 50 countries found private bureaus increased lending far more than public ones--nearly twice as much. Firm report (business profile) less of a constraint with information sharing.

IADB examined data from 170 Latin American banks to measure impact on loan performance. Banks that loaned to both consumers and SMMEs average non-

performance rate 7.75% points lower than banks that did not, encouraging small business lending.

US Federal Reserve study found that large banks enter the SMME financing space, especially small business loans (<$100,000 US) when information sharing and scoring are available tools.

I

Information Sharing & SMME Lending

23

Lesson 3: Gradually change most of the time, but not on all dimensions

24

Gradual v. Rapid apply to at least two dimensions of credit report contents:

First--breadth and depth of positive information that is collected. List of positive data fields is extensive (loan amount, outstanding

balance, interest rate, type of collateral, value of collateral, loan rating, maturity)

Few economies collect all of these fields Inclusion of “timeliness of payment” sufficient to be “full-file”

Second-- scope of sectors that report to credit bureau Data may not be collected from all types of credit providers

(segmented into retail, bank, money lenders--e.g. Japan) Data may not be collected for all credit instruments (in many

economies mortgage payment data not collected).

I

Gradual v. Rapid Implementation

25

Examples of Gradual v. Rapid

At a practical level, almost all expansion is gradual—larger furnishers in sophisticated sectors come in first--except when reporting is mandated.

I


Inclusion of Positive Data

Gradual Rapid

Inclusion of Sectors

Gradual I (US, EU, Aus, Brazil, Japan, South Africa)

II (India, Indonesia)

Rapid III (Sri Lanka) IV (Colombia)

26

Gradual v. Rapid expansion of practice and of reporting regulations

Expansion of practice, some considerations: Data furnisher comfort levels (e.g., skills infrastructure, competition) Consumer protection, practices must emerge in tandem

Expansion of laws and regulations, some considerations:Concerns for societal norms of privacyConcerns about “fit” of other regulations in the system

Bankruptcy and collateral Creditor rights

Concerns for development of lending markets (structure of secondary markets, e.g.)Concerns for mis-application of information tools, i.e., learning curve


27

Gradual v. Rapid expansion of practice and of reporting regulations

To the extent that regulatory changes--e.g. the reporting of data fields--require societal support, and to the extent that it’s lacking, a gradual effort can serve as a series of experiments in which all social segments--consumers, lenders, regulators--become progressively comfortable with credit reporting. The structure of the credit reporting system matters.

If already segmented, then comprehensive becomes difficult If only one or a few bureaus, gradual approach optimal to set example in one sector

that can be applied to others


28

Lesson 4: More data, the cause of, but even more, a solution to identity-based crime

29

All else being equal, the fact of more data being “out there,” that is, in more databases, increases the chances of data in unauthorized hands.

BUT all else is NOT equal: more data also available to fight identity theft.

Data Security Issues

30

Greater information sharing is optimal for lenders and borrowers, and outperforms “negative-only” system. Negative only is events-based. For most borrowers,

triggers are rare, or are never experienced. Little information on vast majority.

Positive information sharing means data is regularly reported on all data subjects.

Payment patterns available to them to identify fraudulent activity.

Truth databases to verify identity and simply to identify fraud.

Data Security Issues

31

The reporting of new account information to a centralized third-party allows a data subject to review regularly what, if any, new accounts have been opened in her/his name.

Free disclosures serve as instruments to reduce ID theft and ID fraud. Key is to engage the data subject

Recent declining number of ID theft cases and in total losses in US correlates with free disclosure (FTC studies)

Data Security Issues: The Role of Disclosures

32

Lesson 5:

More data means better data

33

Positive systems allow for data quality improvements that negative-only systems do not Positive data permits for the creation of data quality tests that do not rely entirely on consumer engagement

Tests on tradeline data (payment on revolving account showing zero balance)

ID missing fields, ID duplicate files

Positive information sharing involves automated reporting systems, brining relatively higher level of data accuracy than manual

Data Quality Issues: Importance of Data Quantity to Quality

34

Market structure and policy act to shape data quality: Credit bureaus, in a competitive market, have strong incentive to improve data quality. Policy that punishes for willful or negligent maintenance of inaccurate data helps.

Data furnishers can control customer service costs, and maintain customer satisfaction, by investing in data quality. Policy must involve furnishers in dispute resolution and make them culpable for knowing or negligent maintenance of inaccurate data.

Data subjects have incentive to engage contents of report. Access to credit file should be relatively simple and affordable (free for limited number of disclosures). Policy must incentivize data subject to engage credit file.

Data Quality Issues: Stakeholder Incentives

35

Lesson 6: Participation in credit reporting system always in a lender’s best interest

36

Independent studies using credit file data from Asia, North America, and South America show:

strong positive correlation between growth in lending to private

sector and participation in credit reporting system;

lenders experience sustainable growth in lending to new markets;

loan portfolios perform better (impacts profits, reserves)

Why Lenders Share Information with Credit Bureaus

37

Positive credit reporting system has network characteristics:

value of credit reporting system increases as number of participants

increases; and,

cost of sitting on sidelines increases as growing number of competitors

report customer data to credit bureau.

Lenders must overcome fear of competition. Requires leadership with vision and courage.

Why Lenders Share Information with Credit Bureaus

38

Lesson 7: One more “valley of transition,” this time in value-added products

39

Positive reporting witnesses new value added services opportunities emerge. More consumer information means more potential products.

New data takes time to gather, test and deploy.

Disincentive to continue developing tools based on older

limited information, leads to transition phase.

The new products may take some time.

Testing with retrospective data may reduce transition

Another Valley of Transition: Value-Added Products

40

Introduction

Basics

Surprises: Choices and Pitfalls

Conclusion: Two Further Considerations

Agenda

41

Dysfunctions in the US mortgages Consumers who sought to borrow more than they

can pay (either out of ignorance, speculation, or hope that rising asset prices would offset crunch) Many did convert (switch to fixed rate, low rate) once

info history was build Others thought they could sell if they couldn’t meet

payments (…but what is true at an individual level is not true at a global level)

Information Sharing and the US Subprime Meltdown

42

Moral hazard in the lending systems—a string of principle-agent problems (greed/arrogance all the way down)

Regulators failed to curb risky loans Risky loans made possible by

Brokers had incentives to make the sale commissions were on sales not performance, likely result of regulatory changes that relaxed lender

supervision over brokers Lenders securitized them for sale on secondary markets (CDO

boom made the market), not “holding the bag” if the loan goes belly up

Investment houses and hedge funds, convinced that structured securities based on rising home values could offset any defaults

Ratings agencies that were paid by the issuers


43

BUT one key shift in system—reduced incentive to use information to assess risk.

o Availability of information is one key component in stable, expanded lending.

o The incentives to use information to extend loans to those who can afford it and not extend loans to those who can’t depend on other factors

o Origins of crisis: changes removed incentive to use information, i.e., NOT the use of information BUT the failure to use information


44

Confluence of circumstances exacerbate situation: Pro-business not pro-markets administratoinTurtles the whole way down Election year--lame duck president 30 years of uneven wealth distribution

Paul Krugman--Age of Diminished Expectations Lesther Thurow--Steps and Ladders Cannot buy way out of crisis Cannot service debt burden--house of cards collapses


45

For reform to work, must build societal consensus. Cannot let the perfect be the enemy of the good.

o Engage in stakeholder educationo Explore setting data formatting standardo Explore developing consumer-friendly dispute resolution

procedureo Agree to monitor impacts

o Risk an institutional decisiono Breadth and depth of lending better metrics

Never been better time for credit reporting reform in Australia than now. Must not miss this opportunity.

Considerations for Australia

POLITICAL & ECONOMIC RESEARCH COUNCIL

100 Europa Drive, Suite 403 Chapel Hill, NC 27517

www.infopolicy.org

Phone: +1.919.338.2798

http://www.infopolicy.org/

Credit Information Sharing in Australia: A Roadmap for Transitioning

Economy & Finance

Credit Information Sharing in Australia: A Roadmap for Transitioning