For 37 years, the leading publication covering payment systems worldwide.
SEPTEMBER 2007
Issue 887
Posted with permission rom The Nilson Report, Carpinteria, Caliornia.
scoring thin credit files The three major credit bureaus – Equiax, Experian, and TransUnion — maintain records on 172 million adults that contain enough data on fle to create a credit score using automated systems. Those adults represent 75% o the total adult population in the U.S. There are another 24 million adults about whom the credit bureaus have no inormation at all. And alling between those
c ca s – u.s. FICO Score
Market Segments
Unscored No Records Unscored Thin Files Unscored Subtotal 500-649 Subprime 651-749 Prime 750-800+ Superprime Scored Subtotal Bureau Files Adult Population
Adult Population as of July 2007
Market Negative Saturation Risk*
24,000,000
<5% <12%
? ?
55% 92% 86%
40% 7% 2%
33,000,000 57,000,000 61,000,000 99,000,000 12,000,000 172,000,000 202,000,000 229,000,000
Note: This chart contains industry estimates and pooled results that vary
between bureaus, scorers, and databases. *Likelihood o becoming 90 days or more delinquent within the frst two years. ©2007 The Nilson Report
categories are 33 million adults about whom the bureaus have fles that contain insufcient inormation or automated underwriting. These people are said to have “thin” fles. In terms o credit risk, thin fles range rom unbanked subprime customers to the very wealthy superprime. No common defnition o a thin fle is agreed upon by the bureaus, but typically the term means a fle with less than fve current tradelines or credit events such as auto loans, credit cards, mortgages, student loans, personal loans, installment loans, as well as credit inquiries — requests
by a lender or a consumer ’s credit report or credit score, which are included in consumer fles even i the person has not yet established a tradeline. Full fles by comparison average 13 to 14 tradelines that have been updated in the last six months. About 15% o all credit card applications are rejected by issuers, not because the consumer has a history o bad credit, but or lack o enough positive credit inormation or no inormation at all. This has card issuers pressuring bureaus to develop methods or reaching these new prospects. Over the past two years, bureaus have tested the predictive value o nontraditional data related to payment histories rather than credit histories — data generated by checking accounts or debit cards, utility bill payments, rent payments, rent-to-own accounts, payday loans, etc. Sources o this data include ARC Systems, Austin Logistics, ChoicePoint, ChoicePoint, Credco Credco,, DataX, eBureau, eFunds, First Advantage, Innovis, and Teletrack. Out o the 24 million Americans with no credit fles as many as hal might qualiy or credit. These people include divorcees and widows who relied on their husbands or credit, college students, people who rely exclusively on checking accounts, and oreign nationals who may or may not be living in the country legally. Foreign nationals are also included in the hal who would not be likely to qualiy, along with the unemployed, welare recipients, the institutionaliz institutionalized, ed, and recent bankrupts. Eorts by risk management companies to generate enough inormation to apply credit scoring technology to the 25% o U.S. adults with thin or no fles includes Fair Isaac’s Expansion Score credit risk model, which searches databases holding 654 million payment records including checking accounts, membership clubs, cell phones, and utilities. The company claims it can generate credit scores or over 70% o the people with thin or no fles, and that typically 35% o
© HSN Consultants Inc. 2007 THE NILSON REPORT Reproducing or allowing reproduction or dissemination o any portion o this newsletter in any manner or any purpose is a copyright violation subject to substantial fnes. Yearly Subscription or 23 print and electronic issues is $1,295. ISSN 1087-8718 THE NILSON REPORT: 1110 Eugenia Place, Suite 100, Carpinteria, CA 93013 USA • PHONE (805) 684-8800 • FAX (805) 684-8825 •
[email protected]
For 37 years, the leading publication covering payment systems worldwide.
SEPTEMBER 2007
Issue 887
Posted with permission rom The Nilson Report, Carpinteria, Caliornia. these receive scores between 640 and 850, representing a range rom the upper end o the subprime credit market to the superprime market.
contActs Equifax John Carter is Sr. VP, Data Acquisitions in Atlanta, Georgia, (404) 885-8300,
[email protected]. t
Experian Zaydoon Munir is Sr. VP, Marketing in Costa Mesa, California, (714) 830-7849,
[email protected]. t
t
Fair Isaac Lisa Nelson is VP, Global
Scoring in Minneapolis, Minnesota, (612) 758-5469, lisanelson@ fairisaac.com. t
LexisNexis Tom Brown is VP,
Financial Services Market in Boca Raton, Florida, (561) 999-4422, thomas.brown@ lexisnexis.com.
TransUnion Chet Wiermanski is VP, Analytical and Decision Services VP, in Chicago, Illinois, (312) 466-7827,
[email protected]. t
Equiax’s MarketMax, developed by LexisNexis, generates a risk assessment score by applying statistical techniques that draw inormation rom nontraditional databases. MarketMax also identifes emerging credit customers with thin or no fles, paying particular attention to recent college graduates and newly arrived immigrants. MarketMax can score prospects or auto loans, general purpose credit cards, retail credit cards, and wireless telecom accounts. LexisNexis markets its own
product called RiskView that lets card issuers identiy and predict the risk o consumers with thin or no credit fles based on 300 attributes. RiskView uses databases o public records to consider three actors: stability (based on LexisNexis’s LexisNexi s’s history o address records), ability to repay (based on the prospect’s asset profle), and willingness to repay (based on an analysis o bankruptcy, liens and judgments, criminal history, and utility inormation).
sources including payday loans, rental payment histories, and debit card data to help score underserved consumers. Other data sources include utility bills and telecom payments, which TransUnion has been collecting since the 1980s. Prior issues: 885, 883, 880, 873, 869, 868, 867, 865
Experian’s soon-to-be-released Experian’s emerging market risk model combines traditional credit data sources along with nontraditional sources it describes as “creditlike.” The company claims that its research has validated that monthly payment data rom utility and energy companies, property rental, retail payment, and purchase data has the greatest impact on scoring thin fles. Experian fnds that credit-like data
About 15% of All credit cArd ApplicAtions Are rejected due to A lAck of positive credit informAtion.
moves 10% o an analysis sample rom unscorable to scorable. TransUnion’s L2C Thin File Model uses nontraditional data
© HSN Consultants Inc. 2007 THE NILSON REPORT Reproducing or allowing reproduction or dissemination o any portion o this newsletter in any manner or any purpose is a copyright violation subject to substantial fnes. Yearly Subscription or 23 print and electronic issues is $1,295. ISSN 1087-8718 THE NILSON REPORT: 1110 Eugenia Place, Suite 100, Carpinteria, CA 93013 USA • PHONE (805) 684-8800 • FAX (805) 684-8825 •
[email protected]