Deconstructing This New Debate On Credit Scoring Models
In 1989 Fair Isaac introduced the first credit scoring model. Prior to the introduction of the score, known as FICO, lenders depended on individual underwriters to manually review credit reports, establishing rules for loan approval based on guidelines that included (among other things) requirements for a minimum number of open accounts, the age of the accounts and number of late payments.
The credit score brought important concepts to the industry:
- Credit Consistency: By adopting a common model, credit investors could develop better default modeling, scalable across the national marketplace, avoiding the inconsistencies of individual underwriter judgement. This development was also a way to avoid discrimination in the credit evaluation process, as it was model-driven, and avoided the bias of human intervention.
- Convexity: For investors, a score improves predictability about duration. Individual lender or underwriting bias – or simple variability in underwriting guidelines – by an originator could result in different default patterns, or prepayment risk. A standard credit model can help provide more predictability from a modeling standpoint.
- Automation: Simply put, a scoring model embedded in an AUS has helped increase the speed of underwriting, decreasing the numbers of hands-on underwriters needed in loan organizations.
IF YOU CAN NOW GET A SCORE FOR PEOPLE WHO HAVE BEEN OTHERWISE DISENFRANCHISED FROM HOMEOWNERSHIP OPPORTUNITIES, WOULDN’T YOU AT LEAST TRY?
In the years since its inception, FICO has served the industry well, upgrading its models based on learned-performance outcomes. There is, however, an increasing concern about access to credit, which has led to more importance being put on innovations in underwriting. For some, the concern is that the current model limits access to homeownership from those without established traditional credit, or perhaps too short a history, or too little of it.
This has led to specific calls from advocates for Hispanic and other minority homeownership concerns to change the way credit is evaluated.
This began with attempts to promote alternative, or thin-file, credit evaluation. Rent payments, utilities, mobile phone, or other payments not consistently reported to the bureau could count in the establishment of a credit history.
In 2017 the Progressive Policy Institute (PPI) issued a lengthy research report on the subject stating, “Millions of Americans lack access to valid credit scores. Sitting outside the mainstream credit market can restrict their personal economic growth and potentially lock them into a cycle of borrowing from predatory lenders in order to meet their credit needs.”
I get it. I remember when one of my children graduated from college and, after getting a job, wanted to buy a car. The problem: he had no credit. Lessons learned by this parent – he didn’t qualify for a car loan. I stepped in and co-signed, but even with my high score, they essentially paired the two together and divided in half. This led to a higher loan rate, due to the mid 500s blended score. This was not an issue of bad credit, it was simply one of no formal credit. His financial history, like rent payments through college and cell phone payments, were not considered. We ended up getting the car under my score alone, but this was for a college-degreed and salaried young man. Imagine a self-employed parent with solid income, but one who has always paid in cash. The challenge here is that the unbanked and un-borrowed may show up as “poor credit” in the current traditional models.
In comes this upstart Vantage Score. With a new model that claims of being able to “score” tens of millions more in this country, they quickly caught the eye of regulators, members of congress, and advocates, with many calling for Fannie Mae and Freddie Mac to implement a pilot to test the viability of the model. You’d think this was a pretty easy ask. After all, if you can now get a score for people who have been otherwise disenfranchised from homeownership opportunities, wouldn’t you at least try?
Presidential Candidate Julian Castro had called for new Credit Scoring Models while he was the HUD Secretary stating, “I appreciate this focus today on out-of-the-box thinking.” The National Association of Hispanic Real Estate Professionals’ Gary Acosta has demanded the same, articulating the need: “Hispanics tend to have thin credit profiles. Not necessarily bad, but very little credit. Seventy-percent of Hispanic home-buyers are first-time home buyers that don’t have a lot of wealth, and don’t have a long history working with banks.”
Despite the need for change, progress was quashed when the FHFA nixed the idea, under the team of Obama nominee and Senate-confirmed Mel Watt. Despite an open period where they requested comment and views from all stakeholders, they ended up determining that they would not move forward.
There are varied concerns about introducing a second model to the world of mortgage banking.
- Cost: Some analysis determined it to be extremely costly, from an IT change management perspective, to introduce a new model with little increase in eligibility rates – estimated to be as low as 4-5 million eligible After all, simply getting a score does not infer credit worthiness. This pushback came mainly from the GSEs, but also some lenders as well.
- Adverse Selection: There was concern that some might “score shop” depending on how a pilot, or a full implementation, might work. It was viewed as a risk that some might take a low FICO and then try a Vantage Score (or vice versa) and simply submit the higher score for approval and best price. While controllable under a variety of recommended alternatives, it weighed on the decision makers.
- Predictability: For better or worse, investors and many risk managers do not like change. Adding a new score brings unclear/uncertain outcomes for loan performance. Between adverse selection risk in MBS pooling or simply the unknown of how the scores would perform, even a test could be cause for TBA pool exclusion or at worse a separate prefix during the pilot.
This past week, FHFA Director Mark Calabria reversed the decision made by Watt. The FHFA Fact Sheet sets in place a process for application, review, and approval off credit scoring models.
“One of my priorities is to ensure that the American people have a safe and sound path to sustainable homeownership, which requires tools to accurately measure risk,” Calabria said in a written statement. The new rule “is an important step toward achieving that goal.”
Charles Gabriel Of Capital Alpha Partners was among several analysts to quickly point out the negative implications to Fair Isaac and their stock was quick to reflect that adding, “We see FHFA Director Mark Calabria as a forceful believer in competition to achieve optimal scoring accuracy, and also not one to flout the obvious intentions of bipartisan lawmakers on the Senate Banking Committee, where he once served.”
As reported by the Wall Street Journal, Vantage Score CEO Barrett Burns stated, “Competition is critical for markets to operate efficiently and we are confident this decision will benefit consumers, lenders and the economy at-large.”
While the pathway ahead to introduce new models will be long and tedious, the direction should be credit-positive for many in this country who have been denied any opportunity, simply because they lack a 3-digit number from FICO.
David H. Stevens, CMB, has held various positions in real estate finance, including serving as SVP of Single Family at Freddie Mac, EVP at Wells Fargo Home Mortgage, President and COO of the Long and Foster Realty Companies, Assistant Secretary of Housing and FHA Commissioner, CEO of the Mortgage Bankers Association.