CFPB: The “Credit Invisibles” and Alternative Data

On May 5, 2015, the CFPB published a report, Data Point: Credit Invisibles, finding that 26 million Americans (as of December 2010, and out of 189 million Americans with scorable credit records), or one in every ten, do not have any credit history with a nationwide credit reporting agency (Equifax, Experian or TransUnion). As a result, these so called “credit invisibles” (also referred to in the industry as “no file” consumers) have limited, if any, access, to traditional credit, which may impede their ability to pay for college, start a business, or buy a home. This report highlights two major regulatory and industry issues which will continue to generate debate and drive industry innovation in the coming year: (1) How to reach underbanked and credit invisible consumers; and (2) How to incorporate alternative data into credit scores in a reliable, accurate and predictive fashion. The report also could signal an alignment of interests between the CFPB and the big three national credit bureaus. Namely, the bureaus have an interest in gaining traction in the alternative data market, and the CFPB has an interest in making invisibles scorable, but within a regulatory framework the CFPB controls.

The report found that African-American consumers, Hispanic consumers, and consumers in low-income neighborhoods are more likely to be “credit invisible.” Furthermore, the report attempts to quantify the population of “unscored” consumers (also referred to in the industry as “thin file” consumers), which includes consumers who do not have enough credit history to generate a credit score or who have credit reports that contain “stale” or outdated information. The CFPB found that 19 million consumers have unscored credit records (~8% of the adult population), with a near even split between consumers with “insufficient” credit history (9.9 million) and consumers with “stale” credit histories (9.6 million).

The report, which analyzed the data with respect to age, income level and race/ethnicity, contained the following specific findings:

• The problem of limited credit history affects all racial or ethnic groups, but African-Americans and Hispanics are more likely than Whites or Asians to be credit invisible or to have unscored credit records
• There is a strong relationship between income and having a credit record, with almost 30% of consumers in low-income neighborhoods classified as credit invisible and an additional 16% having unscored records
• Most consumers who are credit invisible or have an unscored credit record are young – over 10 million of the 26 million credit invisibles are younger than 25 and 80% of 18 or 19 year olds are credit invisible or have unscored records

National credit reporting agencies generate credit scores using highly proprietary “secret sauce” algorithms, which they apply to positive and negative credit file data in their databases (most of which is furnished from myriad financial institutions), including the number and types of open accounts, duration of open accounts, payment histories, and balances. As anyone who has ever applied for a mortgage or student loan can attest, the three digit credit score generated has a major impact on consumers’ access to credit and the rates at which consumers can obtain such credit. The national consumer reporting agencies argue that their scores offer the best available mixture of predictive, accurate and timely data upon which to assess the risk of non-payment of the individuals applying for credit.

The challenges that credit invisibles and “unscorables” face in gaining access to traditional credit has received a great deal of attention from researchers, regulators and industry stakeholders. In fact, this discussion has extended beyond no file or thin file consumers’ limited access to credit, to an examination of underbanked consumers’ access to basic banking products, such as prepaid or debit accounts. With respect to reaching both the underbanked and the credit invisible (who often are one in the same), one of the major impediments lies in the fact that banks and other financial institutions lack reliable predictive data to assess the “risk” of these individuals, thus making them unable or unwilling to extend credit or banking services to these consumers. Many argue that the solution to this is the incorporation of so called “alternative data” into the “secret sauce” algorithms for credit scores.

Proponents of the inclusion of alternative data into credit score calculations argue that incorporating predictive and non-traditional factors – such as assets that we own, utility and telecommunication services payments, rental payments, remittance transactions, and demand deposit account activity – would not only make credit scores more predictive due to the inclusion of more viable data points, but would also make previously “thin file” or “invisible” consumers scorable. Accordingly, utilizing alternative data should allow the national credit reporting agencies to generate a predictive score on a previously credit invisible consumer using his or her past utility or telecommunications payment records as data points. Financial institutions would then be able to use this score to evaluate the creditworthiness of the individual and calculate acceptable terms upon which that institution may extend credit to that individual.

Opponents argue that alternative data may be furnished outside of the existing channels established by the federal Fair Credit Reporting Act, which may reduce its accuracy as well as consumers’ rights to identify and correct inaccuracies. Opponents also cite high costs, the diffuse nature of reporting alternative data, and certain state and local requirements which limit reporting of utility information to credit bureaus as barriers to the utility of alternative data.

Despite the ongoing debate, the use of alternative data in credit scores has gained traction in the industry. Most recently, on April 2, 2015, FICO, LexisNexis Risk Solutions and Equifax announced an ongoing pilot program which allows 12 of the largest credit card issuers in the US to use alternative data to identify creditworthy individuals who otherwise would not have access to the traditional credit system. In its press release, FICO stated that FICO’s data analysts had determined that using alternative data (obtained from LexisNexis and Equifax) such as property records, telecommunications and utility records “can reliably be used to score 15 million consumers who do not have enough credit data to generate FICO scores.” In support of the program, a FICO executive stated, “Working with Equifax and LexisNexis, we set out to help unbanked, under-banked and disadvantaged people gain equal access to the standard credit products enjoyed by millions of Americans.” The new FICO score utilizing alternative data has been engineered to work alongside the existing FICO scores, which reportedly will enable credit card issuers to utilize the sore almost immediately.

As discussed above, the CFPB report could signal an alignment of interests between the CFPB and the big three national credit bureaus. Accordingly, we may see a push for the CFPB to “bless” or create a safe harbor for the national credit bureaus’ use of alternative data as a means to score otherwise unscorable consumers and thus facilitate their access to traditional credit. Such action could have far reaching effects, both in terms of expanding access to the credit system and adversely impacting non-traditional lenders whose market advantages to date have included scoring methodologies that don’t depend on the big three – for example, methodologies that depend on data taken from public sources, avoiding the heavily regulated consumer reporting agency system.