By JoAnne Gaskin
The vast majority of U.S. consumers can access affordable credit as a result of the introduction of automated underwriting systems and the widespread adoption of credit scoring by financial institutions during the past 25 plus years. In fact, 92% of the consumer credit applicant population today receives a FICO® Score that leverages the information contained in the credit files at the three largest U.S. consumer reporting agencies (CRAs).
With that said, there is still a sizable opportunity for the financial industry to expand access to consumer credit. This is an important issue as access to affordable credit products helps consumers’ ability to build wealth. There are approximately 50 million consumers that don’t receive a score today, many of whom would be considered good credit risks by lenders. These consumers do not receive a score for one of two reasons. The first is that there are 25 million consumers who have not used any form of traditional credit, and as a result, have no record at the CRAs and are often referred to as “credit invisible.” The remaining consumer population does have some information at the CRAs but have insufficient information to generate a reliable credit score. In conducting research (see graphic below), FICO found that these 28 million unscorable consumers that have existing credit bureau files profile very differently, as outlined below:
- Credit Retired. Consumers with no recent trade line updates in 4.5 years, on average. This group has a median age of 71 and most have voluntarily stopped using credit accounts. Their appetite for new credit is very low; fewer than 5% of this population recently applied for credit.
- Lost Access to Credit. Consumers with no recently active/updated trade lines, and with historical signs of serious delinquency or other negative behavior in their credit file (e.g., collections and public records). This segment consists of approximately 18 million of the 28 million unscorable consumers who have credit files at the three largest CRAs. These consumers have in many cases experienced a negative event and subsequently lost access to credit. As such, their traditional credit files are “frozen” at their moment of financial distress. Scoring these consumers solely based on this ‘frozen’ data will result in many of them receiving low scores with no opportunity to positively impact their credit standing in a favorable way.
- New to Credit. Consumers with 1+ recently active/updated trade lines but with less than 6 months of credit history. These consumers have a median age of 24 and typically have recently opened their first credit card with a low (< $1,000) credit limit. As these consumers maintain the credit account, they will receive a FICO Score after 6 months history.
In light of the above findings, FICO chose to turn to data sets outside of the three CRAs to develop a new alternative data scoring model, FICO® Score XD. This score is designed to identify creditworthy consumers in the traditionally unscorable population and its intended use is for unsecured lending.
“Alternative data” is a term that is widely used but frequently misunderstood. At FICO, we define alternative data as information that cannot be found in the files maintained by the three major credit reporting agencies. FICO® Score XD uses consumer report data, such as telco, utility and property record information, which is not captured in the credit files maintained by the three largest CRAs, but is demonstrably and statistically predictive of the individual’s creditworthiness.
As a side note, the credit files maintained by the three largest CRAs contain trace amounts of telco, utility and rental data that may be referred to by some as alternative data. Just how sparsely reported is this information? An estimated 92% of US adults have a cell phone, however, just 2.5% of all consumer credit bureau files contain telco account information. The story is similar on the utility account side, where over 60% of American adults pay for utilities, but just 2.4% of consumer credit bureau files contain utility (non-telco) payment information.
Much like with utility and telco, rental data is rarely encountered in consumer credit files. The US Census bureau estimates that there are 240 million adults in the US, with 35% living in rental housing. Of the roughly 80 million US adults who live in rental housing, we found that just 270,000 (or 0.3%) of those consumers have a rental trade line reported in their credit file. There simply isn’t an opportunity to safely and soundly score millions more Americans using credit bureau data alone. To score more people, we need to use truly alternative data.
Assessing Alternative Data
A key reason there’s an opportunity to score more consumers today is the growing number of alternative data providers with national scale that have entered the market in recent years.
Telecommunications payment data, for instance, has many of the same qualities as data reported in traditional credit files. While very little of this data resides at the CRAs, more complete telecom data is available from alternative sources that include positive payment history as well as negative information. Property and public record data is another good example of useful alternative data. These data sources tend to be inclusive of large groups of consumers, and can provide both positive as well as negative information relating to credit risk.
FICO’s six point test for alternative data sources include:
- Regulatory compliance. Any data source must comply with all regulations governing consumer credit evaluation. It’s also critical to be able to clearly explain the role the data plays to consumers and regulators.
- Depth of information. The more detailed the data, the greater the value it may offer.
- Scope and consistency of coverage. Since the goal is to score as many consumers as possible, useful databases must cover a broad percent of the population. For instance, with more than 90 percent of U.S. adults using cell phones, mobile companies are a data source with broad coverage.
- Accuracy. Inaccurate data compromises its predictive value. Data repositories must ensure accuracy.
- Predictiveness. The data should predict future consumer repayment behavior. For example, consumers who have been at their addresses for longer periods are more likely to pay credit obligations than those more transient.
- Additive value (aka “orthogonality”). Useful data sources should be supplemental to credit bureau reports.
One of the data sources that met all the guidelines detailed above in FICO’s six point test was The National Consumer Telecom and Utilities Exchange, Inc. (NCUTE). NCTUE is a member-owned database through which its member companies exchange source-anonymous information which include new connect requests, payment history and historical account status for telecommunications (e.g., landline and mobile), pay TV (e.g., cable and satellite) as well as a small amount of utility information. The database contains information on more than 210 million consumers and includes both positive and negative payment history. The database is housed and managed by Equifax.
FICO also discovered a number of public record data sources which were not present today in the credit bureau files and that met FICO’s rigorous six point test. Public records data sources from LexisNexis Risk Solutions include such types of data as length in residence, property ownership, frequency of address changes and how long it has been since a consumer has been evicted.
FICO’s Financial Inclusion Initiative Outside the U.S.
FICO’s efforts to expand the scorable population of creditworthy individuals are not limited to the United States. In 2016, FICO announced the FICO Financial Inclusion Initiative, a global effort to increase access to affordable credit for consumers with limited or no credit history. Through this initiative, FICO is leveraging alternative data outside the U.S. that includes a broader set of data sources than in the U.S. – such as psychometric surveys, social network, retail purchase and mobile phone information to identify creditworthy consumers among the 3 billion consumers globally that are unbanked or underbanked. These projects test an approach to obtain alternative data through consumer consent, which can allow for a broad array of alternative data sources to be considered for a lending decision at the discretion and control of the consumer.
Expanding Access to Credit Isn’t About Lowering Standards
By incorporating alternative data, lenders can effectively identify good risks and safely expand access to credit in a controlled, responsible and predictive way while ensuring a quality score and quick access for consumers. The goal is not to just generate more credit scores, but to ensure those scores enable lenders to safely and responsibly extend credit to a broader group of people. Leveraging alternative data is a critical tool to generate meaningful, predictive scores that reveal many creditworthy individuals who would otherwise still be left out of mainstream credit. In doing so, we can ensure the integrity of our scoring standards as well as empower lenders with the credit scoring tools necessary for expansion of credit to a broader and more diverse group than ever before.
Joanne Gaskin is the Senior Director, Scores and Analytics at FICO. She oversees the FICO Scores regulatory practice and is responsible for the strategic direction of FICO’s analytic solutions and partnerships serving the mortgage industry.