 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
       IntroductionCredit Score BasicsCredit Scores - AdvancedRe-ScoringObtaining Your Credit ReportReading Your Credit ReportDisputing ErrorsIdentity Theft BasicsMaking & Mixing Credit ReportsReinvestigations (or not)HistoryCredit RepairDebt CollectionAuto InsuranceHomeowners InsuranceMortgage InsuranceThe Color of Credit ScoresSpecial ChallengesOpting OutImpermissable AccessDamage & DamagesThe 2003 FACTA BattleMissing Credit LimitsConclusion |
|
Credit Scores and Credit Reports, by Evan HendricksChapter 16 The Color Of Credit Scores
We must rapidly begin the shift from a
Is there a connection between credit scoring and racial discrimination? Clearly, "the jury is still out" on this very important question, but by 2004, all sides were sharpening their arguments and preparing studies to support their viewpoint. For example, in a 2004 study, Missouri Insurance Commissioner Scott Lakin published the first research indicating that credit scores were disproportionately harsher on minorities.246"thing-oriented" society to a "person- oriented" society. When machines and computers, profit motives and property rights are considered more important than people, the giant triplets of racism, materialism, and militarism are incapable of being conquered. — Rev. Martin Luther King "Beyond Vietnam: A Time To Break Silence" (April 1967) In a 1996 report, Freddie Mac, the giant mortgage underwriter, acknowledged in a report that African-American borrowers were about three times as likely than white borrowers to have high-risk credit scores — defined as FICO scores below 620. Hispanic borrowers were twice as likely as White borrowers to have high-risk scores.247 246 Kabler, Brent Ph.D., "Insurance-Based Credit Scores: Impact on Minority and Low Income Populations in Missouri," Missouri Dept. of Insurance Statistics Section, January 2004 247 "Automated Underwriting: Making Mortgage Lending Simpler and Fairer for America's Families," Report By Freddie Mac, September 1996 (Herein cited as "Freddie Mac Report") In a footnote, the Freddie Mac was quick not to blame it all on Fair Isaac. "This pattern, while not well understood, seems to reflect less about credit markets and more about the general economic condition of many minority families. African-American and Hispanic households tend to have higher unemployment rates, less job security, and significantly lower levels of wealth. For example, in times of financial difficulty, minority households may be less able to get help from parents or other family members and more likely to fall behind in their payment obligations."248 Using Freddie Mac's data from its 1999 National Consumer Credit Survey, Birny Birnbaum, executive director of the Council for Economic Justice in Austin, Texas, compared overall scores to those of minorities. "When combining African-Americans, Hispanics, and Whites into one group, he estimated that:
For African-Americans he estimated that:
For Hispanics he estimated that:
248 Id. 249 Birnbaum, Birny, "Insurers' Use of Credit Scoring for Homeowners Insurance In Ohio: A Report For the Ohio Civil Rights Commission," January 2003 250 Id. For Whites, in contrast, he estimated that:
Like SAT Scores When this author was in college, a group of African-American students invited black and white students to take a multiple-choice test patterned after the Student Achievement Tests (SATs) that have become a rite of passage for high school students preparing to enter college. But the test they gave was heavily weighted with questions about African-American history and culture. After struggling through so many unfamiliar topics, many of the white students would ask, "How am I supposed to know the answer to these questions?" The African-American organizers of the event made their point: they who create and give the test, likely will influence who succeeds. Credit scores have been favorably compared to SAT tests because both systems rate achievement according to where the individual's score falls within the percentile, as compared to the entire group. Credit Scores and SAT tests share another commonality: they reflect the assumptions of a narrow group of Caucasian-Americans who created them and have grown to be a widely accepted standard that touches the lives of millions of Americans. As one observer quipped, "Credit scores are just one more reminder that it's still a White Man's world." Freddie's 'Seal of Approval' FICO's defenders, on the other hand, insist the scoring system is both colorblind and the most effective predictor of credit risk. In its 1996 study, for instance, Freddie Mac found that regardless of whether the borrower was white, black or Hispanic, loans to borrowers with scores above 660 performed better than loans for borrowers with scores between 620 and 660, which in turn performed better than loans with FICO scores below 620.251 Freddie Mac also cited an August 1996 "Discussion Paper" by Fair Isaac that sought to determine whether minorities were adequately represented in credit bureau data. Fair Isaac found that residents of "high-minority areas" accounted for 7.8 of adults, and for 6.7 percent of consumers with credit reports maintained by the three major CRAs. Fair Isaac conceded that the figures indicated a "slight under-representation of high-minority-area residents" in credit bureau files. But the company said the "data clearly cover a significant number of minority households."252 Looking at Fair Isaac's data, Freddie Mac said it was satisfied: "The fact that credit-bureau scores are powerful predictors for minority borrowers confirms that they are sufficiently represented."253 Freddie Mac and Fair Isaac separately examined whether heavy use of finance companies, as opposed to major banks or other mainstream creditors, push their credit scores into a high-risk category. Fair Isaac "detected little variation in the number of finance company accounts used by consumers regardless of racial composition... (and) Freddie Mac reached a similar conclusion."254 Freddie Mac noted that more lenders were turning to alternative means for documenting credit history of low-income people, including rent and utility payments and other recurring obligations. 251 Freddie Mac Report (see Footnote 2) 252 Fair Isaac and Company, Inc., "Low to Moderate Income and High Minority Area Case Studies," Discussion Paper, August 1996, as cited in the Freddie Mac Study. 253 Freddie Mac Study 254 Id. It concluded that the system would welcome with open arms those consumers that were willing to conform their behavior to its standards. "Fortunately, consumers are able to go from no credit history to an acceptable one relatively quickly, in perhaps one or two years. To do this, they need to open and use several credit accounts and make timely payments without running up large balances."255 Insurance Scoring and Discrimination? Throughout 2003 and into 2004, the debate intensified over the alleged discriminatory effects of insurance scoring. As one insurance industry official recently said, "We live in a racist society. Scoring is not at fault; it simply reflects what's there, whether we like it or not." But before more closely examining the Missouri research, it is useful to examine, albeit briefly, the evolution of one insurer's practices and how that insurer historically discriminated against minorities. The 'Negro' Factor For nearly 100 years, the Metropolitan Life Insurance Company had a policy of selling sub-standard insurance policies to "Negroes" and other non-whites.256 Some of these policies were known as "burial" policies, as they were touted as a means for poor people to have enough money to pay for a decent burial. MetLife agents would go door-to-door to collect weekly premiums. In some cases, the premiums exceeded the policy's payoff. 255 Id. 256 The following information about MetLife is from the opinion and order of U.S. District Judge Harold Baer, Jr., in Karl Thompson, et al. v. Metropolitan Life Insurance Co.: USDC - Southern District of New York — 00 civ. 5071; June 21, 2001. Christa Collins was co-counsel on the case; she currently is partner at James, Hoyer, Smiljanich & Newcomer, in Tampa, Florida. The Insurance Library in New York City is also a useful resource. In its manual entitled, "Instructions To Agents," dated June 1940, MetLife advised agents, "In designating race, state the race definitely; for example: White; colored; Indian; Mexican; Hawaii; Porto Rican (sic); West Indian; etc. Colored includes not only the Negro, but the Mulatto and all other persons having colored blood. Only persons of pure Caucasian blood are designated as white; all other races should be designated by their proper term or distinction." 257 As blatant discrimination became less acceptable, MetLife found other ways to continue its practice of selling sub-standard policies to African-Americans and other non-whites. One approach was to classify policyholders by job classification. Since blacks typically were the only ones who worked as "porters," or "black boots," this approach temporarily succeeded at separating whites and non-whites. A 1964 memo showed that MetLife could distinguish between blacks and whites by classifying them by neighborhood, sometimes called "geo-coding," other times, "red-lining." "This memo and attached map identified 108 neighborhoods in Manhattan that had 'the least favorable population and housing characteristics,' including the Lower East Side; south of 14th Street; in east and central Harlem; and in the lower parts of Washington Heights,'" wrote Judge Harold Baer, Jr., in his 2001 order.258 Judge Baer quoted from the 1964 memo to underscore MetLife's preoccupation with race. "About two-fifths of the residents in these areas are nonwhites and an additional one-fifth are Puerto Rican. 257 Id. 258 Id. This compares with one-fourth nonwhite and one-eighth Puerto Rican for Manhattan as a whole," said the memo.259 As time passed, MetLife needed to delve more into personal details of applicants to continue its discriminatory practices. This trend was reflected in the "Mercantile" reports, which were the product of questionnaires only required of non-whites. Some of the questions included: Number of rooms in the home? Number of persons living in the home, including lodgers? Does home contain toilet facilities? Running water? Is the home in good repair? Clean? Does the applicant or premium-payer associate with criminals or gamblers such as those in the policy number game? Does the applicant or premium-payer get into fights? Have the applicant's or premium payer's drinking habits been criticized?260 MetLife believed that blacks had higher mortality rates than whites, even though some experts said the company's actuarial tables were flawed. A 1950 memo showed that MetLife was concerned that some state anti-discrimination laws might be unfair to whites: "Since the laws of several states do not allow us to take race into account in appraising an applicant for insurance, we have had to adopt other means of avoiding unfair discrimination against white policyholders in the cost of their insurance."261 Allstate In California In the previous chapter on automobile insurance, we noted how California Insurance Commissioner John Garamendi consistently voiced concerns about the use of credit scoring. 259 The memo was authored by "Paul H. Jacobsen Ph.D.". 260 Thompson, et al. v. Metropolitan Life, op cit. 261 Id. In 2002, he brought a formal action against Allstate for its use of credit scoring in homeowners insurance. Garamendi charged that Allstate's proprietary score "in underwriting or tier placements results in excessive, inadequate, unfairly discriminatory, or unreasonable rates." His department made the move after an Allstate motion for a rate increase confirmed that the company was using credit scores in rating homeowner policies.262 In possibly the first order of its kind, Administrative Law Judge Lisa Williams ordered Allstate to reveal its scoring algorithms to prove they were not discriminatory or unreasonable. Soon thereafter, Allstate settled, agreeing to stop using credit scores in setting homeowners insurance. It also agreed it would not try to re-introduce credit scoring unless it first revealed its credit scoring methodologies with the department. Allstate said the debate on the use of credit scoring, which it referred to as "insurance financial stability" in underwriting, "would best be served in a public policy forum such as the legislature, or a formal rulemaking by the department, rather than in the context of a rate hearing."263 In a subsequent filing, Garamendi charged that Allstate's use of credit scoring, or "insurance financial stability (IFS)," in auto insurance, violated California insurance law. 262 Order Adopting Proposed Decision, In The Matter of a Rate Application by Allstate Insurance Company, Calif. Dept. of Insurance, No. PA02026201; Admin. Law Judge Lisa Williams, April 25, 2003 263 Id. Noting that IFS was used to "predict the likelihood of future losses" based upon credit reports, the complaint said that Allstate's credit scoring program used a "scorecard approach where a consumer credit report's value for a particular attribute or combination of attributes receives a score. The total score for the credit report is then an aggregation of all the individual attribute scores."264 "The use of an aggregate score, such as that generated by [Allstate's] IFS program, which combines various credit characteristics that may or may not have a direct substantial relationship to loss exposure, developed without disclosure of the components... does not allow the [Calif. Dept. of Insurance] to determine compliance with the applicable statutes and regulations," the complaint alleged. The complaint also charged that Allstate "failed to maintain reasonable records and has failed to make available information regarding the IFS program to the Commissioner so that a determination can be made whether its use of the IFS program complies with the California Insurance Code." The case was pending at the time this book went to press. The Allstate Class Action A pending credit score-discrimination lawsuit charged that, "Over the years Allstate has employed various artifices to conceal its discriminatory conduct. Initially, Allstate established geographical districts to identify and target non-Caucasian neighborhoods for the purpose of charging the non-Caucasian residents of these districts higher premiums because of their race."265 264 Fifth Amended Notice of Noncompliance Pursuant to CIC 51858, In the Matter of the Rates, Rating Plans, or Rating Systems of Allstate Insurance Company et al., Calif. Dept. of Insurance, File No.NC01017552 265 Jose C. DeHoyos v. Allstate Corp., et al.: U.S. District Court for the Western Dist. Of Texas; class action complaint filed Nov. 2, 2001. In the early 1990's, however, Allstate began "making extensive use of credit reports to assess 'financial stability'" so it could continue its racially discriminating practices under a more objective veneer, the lawsuit charged.266 Next came credit scoring. "In recent years Allstate has intensified its use of credit scoring through development of the Insurance Financial Stability score and the Strategic Risk Management system. Allstate designed each of these artifices for the purpose of using them as a pretext to discourage sales to non-Caucasians and to charge non-Caucasians higher premiums." Jose DeHoyos and Eva Perez-DeHoyos, Mexican-Americans living in Texas, had maintained good driving records since 1984, the year they took out auto and homeowners insurance policies with Allstate. They also had good credit histories. But in July 2001, Allstate said their premiums for both their homeowners and automobile policies were being substantially increased. When the DeHoyos asked why, they were informed that Allstate's new use of Insurance Financial Stability Scores caused the increase in their premiums. Four other plaintiffs, African-Americans and other non-Caucasians from Texas or Florida, recounted similar experiences. The complaint indicated that a major motivating factor for Allstate was to use credit scoring to identify "High Lifetime Value" customers (i.e. the kind that have more money and the greatest potential to buy additional policies). In court, Allstate argued that 1930s-era law267 barring federal involvement in insurance issues prevented plaintiffs from suing under federal civil rights laws. But a federal court in the Western District of Texas disagreed, allowing the suit to go forward. 266 Id. 267 McCarran-Ferguson Act, 15 U.S.C. § 1012(b) On September 3, 2003, in a 2-1 opinion, a three-member panel of the U.S. Court of Appeals for the Firth Circuit also rejected Allstate's argument, stating, "Every circuit that has considered the question has determined that federal anti-discrimination laws may be applied in an insurance context, even where the state insurance agencies have mechanisms in place to regulate discriminatory practices."268 The case was pending at the time this book went to press. The Missouri Study Missouri law prohibits sole reliance on credit scoring to determine whether to issue a policy. However, there are no limits on price increases that can be imposed due to credit scores so long as such increases can be actuarially justified. Accordingly, auto insurers use credit scores widely. To understand how minority populations were affected, the Missouri Insurance Department collected scoring data on 330,000 applicants from the top-20 auto insurers. The study found a major gap between "all-minority" and "no-minority" neighborhoods. The average credit score in the "all-minority" neighborhood fell within the 18.4 percentile, meaning the average score was in the lowest one-fifth of Missouri residents. Meanwhile, average credit scores in the "no-minority" neighborhoods were in the 57.3 percentile (close to the top two-fifths) — a gap of 38.9 points.269 Even after eliminating a broad array of socioeconomic variables, such as income, educational attainment, marital status and unemployment rates, the relationship between below-average credit scores and minority concentration in a ZIP Code remained. "Indeed, minority concentration proved to be the single most reliable predictor of credit scores," it said.270 268 Jose C. DeHoyos v. Allstate Corp., et al.: 345 F.3d 290 (5th Cir.) Dissent by Judge Edith Jones. The Fifth Circuit rejected Allstate's en banc appeal, raising the possibility of an appeal to the U.S. Supreme Ct. 269 Kabler, op cit. 270 Id. Moreover, the study examined the percentage of minority and white policyholders in the lowest three-fifths of credit score ranges and found that minorities were over-represented in this worst credit score group by 26.2 percentage points.271 In response, Missouri Governor Bob Holden said he would introduce legislation to ban the use of credit scoring in auto or homeowners insurance. "The concern is that credit scoring is unfairly penalizing low income citizens with inflated insurance prices, with much of the burden falling on African-Americans and Hispanics. This places unnecessary obstacles in the way of many people and many communities that are struggling to move forward." Holden expressed concerns about credit scoring techniques, indicating they are secretive, not well understood by the public, and they may not be an accurate indicator of a person's financial responsibility. Scott Lakin, director of the Missouri Insurance Department, said that Missouri would head a national study of credit scoring that will test whether similar patterns exist in other states. Ten states already had signed up for the project; another ten have expressed interest in joining the effort.272 Maryland & Industry Differ The Maryland Insurance Administration subsequently criticized the study, as "inconclusive," partly because insurers don't collect data about race. Joe Thesing, Missouri manager for the National Association of Mutual Insurance Companies (NAMIC), said the Missouri study was based on "faulty methodology." "Most professional researchers think ZIP codes are far too large for accurate analysis," he said. The study also failed to focus on actual purchasers of insurance and claims experience, he said, adding, "All previous evidence indicates that insurance scoring actually allows companies to insure more people at a lower cost."273 271 Id. 272 Id. 273 NAMIC Press Release, Jan. 30, 2004 www.namic.org/newsreleases04/040130nr1.asp © 2005 Evan Hendricks and Privacy Times, Inc. All rights reserved. |
|
 |
       |
 |
 |
 |