Title: The Value of Comprehensive Credit Reports: Lessons from the U.S. Experience
1The Value of Comprehensive Credit
ReportsLessons from the U.S. Experience
- Michael E. Staten
- McDonough School of Business
- Georgetown University
The Legal and Regulatory Environment for Credit
Reporting Systems, sponsored by the World Bank
Washington, DC June 18, 2001
2Credit Products Have Achieved Deep Penetration of
U.S. Households with Relatively Low Delinquency
Rates
- Percent of Households Percent of Borrowers
- with Product 30 days Delinquent
- 1998 1st Quarter 2000
Mortgage 43.1 2.8 Closed-end inst
allment 43.7 6.2 Bank Credit Card
67.6 4.6
Source Federal Reserve Boards 1998 SCF and
Trans Union LLP.
3Bankcard Ownership by Household Income
Percent of households with at least one
bankcard.
Source Federal Reserve Board.
4How Did U.S. Creditors Achieve Deep Penetration
with Low Delinquency Rates?
- Legal rules which permit the collection of
personal credit data, both positive and negative
- The development of statistical scoring techniques
to predict borrower risk
- Risk-based pricing made possible by the repeal of
legislated interest rate ceilings
- Use of credit bureau pre-screening to target
offers of credit
5Simulations to Demonstrate How More Complete
Credit Histories Improve A Scoring Models
Predictive Power
- Begin with sample of U.S. credit reports
- Build scoring model to predict delinquency on new
accounts
- Drop selected data fields to simulate what is
missing from credit files in other countries
- Re-build a restricted model on remaining data
to predict delinquency on new accounts
- Compare results
6Simulation of a Negative-Only Environment
- Full model built under U.S. reporting rules
- positive and negative data, plus inquiries
- Negative-only model built under Australian
rules
- negative data, plus inquiries
- Sample of 312,000 borrowers with new accounts
opened in May 1997
- Models estimate probability of 90 day
delinquency within 24 months
7Caveats Regarding Model Construction
- Data and analytical advice provided by Experian
- Our model contained 56 of 400 Experian
variables
- Variables were selected from 4 major categories
- Outstanding debt and types of credit, all
sources
- Length of credit history
- Applications for new credit (Inquiries)
- Payment behavior (delinquency, chargeoff,
bankruptcy, etc)
8Effects on Default Rates of Adopting a
Negative-only Credit Scoring Model for Various
Approval Rates
9Effects on Credit Availability of Adopting a
Negative-only Credit Scoring Model for Various
Default Rates
10Implications for a Negative-data-only Market
(relative to markets with full bureau data)
- Consumer credit will be generally less available
(smaller percent of households owning product for
each socioeconomic category)
- Smaller consumer borrowing will impair growth of
consumer spending and durable goods industries
- Higher price for credit as penetration rises
11Fragmented-Reporting ScenarioBureau Data
Restricted By Type of Lender
- Typical scenario in many Latin American countries
and Japan.
- Credit reporting historically driven by either
bank or retailer alliances.
- Mostly negative data but some positive data are
present in credit files.
- Simulations measure the impact of having data
available for only one type of lender the
borrowers total debt exposure is unknown
12Simulation to Compare Models Built under Full
Reporting vs. Retail-only
- Full model (U.S.) includes positive and
negative data, all lenders
- Retail-only model includes only retailer
account experience, positive and negative (no
data about loans from banks)
- Sample of 67,000 borrowers with new retail
accounts
- Model estimates probability of 90 day
delinquency on these new retail accounts within
24 months
13Effects on Retail Loan Default Rates of a
Retail-only Credit Scoring Model for Various
Retail Loan Approval Rates
14Effects on Credit Availability of a Retail-only
Credit Scoring Model for Various Retail Loan
Default Rates
15Conclusion Privacy Restrictions that Put Limits
on Credit Reporting Impose a High Cost
- The ability of U.S credit bureaus to report
positive as well as negative credit data has
meant
- Greater access to credit products for all
socioeconomic groups
- Reduction in loan losses that would result from
such market penetration without positive bureau
data
- Creditors can monitor accounts and use scoring to
adjust credit lines and prevent excessive debt
buildup