FICO, CoreLogic Partner to Map Future Credit Risk Scoring
A new collaboration between FICO and CoreLogic, two of the nation’s largest risk scoring data providers, plans to upgrade existing mortgage credit risk scoring models and consequently pave the way for broader access to multiscore rating benchmarks.
CoreLogic and FICO have agreed to develop new industry-standard mortgage credit risk scoring solutions that combine the borrower credit behavior analysis and future credit risk variables provided by the CoreLogic CoreScore Credit Report and the industry-standard FICO 8 Mortgage Score.
According to CoreLogic’s senior vice president of product management and analytics, Tim Grace, this new score is the first “in a series of new scoring solutions that FICO and CoreLogic will create for use in the mortgage industry and beyond.”
The collaboration marks a significant expansion of what has emerged as a national trend triggered and nourished by the housing crisis: The layering of traditional credit risk scoring models with nontraditional borrower-related datasets that are turning these solutions into the new customer rating standards.
Joanne Gaskin, the product management director for scores at FICO, told this publication the breadth of the data captured in the joint report “does not exist in the marketplace today.” And its primary benefit is that of packaging together the data offered by the three credit repositories with other data sources that are not regularly reported alongside the traditional credit repository reports.
Datasets include rental payments and other nontraditional information that “are married together to provide a more robust view of the customer,” she said, so lenders and serivers can make better, more predictive decisions. “We’re still finalizing all the data inputs into the model.”
The information input adds collateral, or property data, second-lien data, CoreLogic’s findings on payment trends if borrowers use a payday lender to pay their mortgage, past bills and mortgage payment history that illustrate borrower attitudes. Altogether, “a really wide array of new data is coming in,” she said. FICO is responsible for “understanding which of the data elements are most predictive” of customer behavior going forward.
Data are key, so lenders and servicers are trying to gather as much information as they can. What model users really want, according to Gaskin, is not a larger amount of data however valuable, as is the need to secure “actionable data” such as the three-digit FICO score that helps rank the risk and make a loan decision based upon that data.
Her description of actionable data: filtered basic data, information about predictable and nonpredictable behavior from the borrower of a given credit score, a full picture of the customer “and the likelihood of risk” around which to build both loan origination and collections or mortgage servicing strategies.
The addition of new data sources will be a continuous process. And the key differentiator to the types datasets selected is dependent on both the quality and quantity of information.
Three features make this initiative with CoreLogic different, Gaskin says. “The breadth of the data captured into a single credit report that does not exist in the marketplace today.” Lastly, creating a score “that replicates the same score range” that lenders and servicers are used to work with as it is part of the industry standard FICO score.
“Lenders know exactly” what is the risk level associated with the 300 to 850 FICO scores, they understand it, she argued, “so it is critical to make the new solution easy to adapt and useful.”
The goal is to provide scoring tools that deliver “additional loan-level insight,” that ultimately help the decision making process before and after loan origination, help grow loan volume and avoid future losses, said FICO’s VP of scoring and analytics, Greg Pelling.
As product use expands the partnership is looking forward to feedback from lender and servicer users, Gaskin said. And while there are no plans to bring new participants to the partnership, she said, FICO, Minneapolis, and CoreLogic, Santa Ana, Calif., “wouldn’t preclude new partners” and is interested in a “broader distribution” of the current and future tools.
Meanwhile, both companies are separately developing other products that could be added to a risk and loss mitigation package solution.
FICO is using predictive analytics to help servicers “identify borrowers at the greatest risk of strategic default.” It is a $20 billion problem that calls for immediate action, Gaskin said, so four of the country's top 10 mortgage servicers already have started to use the new tool.
The FICO product uses analytics to measure overall lending and servicing risk in the likelihood a borrower will become delinquent because of financial distress, compared to a strategic default—when borrowers have the means but make a conscious decision to default when “underwater.” Triggers of that behavior, she said, are the amount of value underwater, the velocity of that value loss, and the forecast for that property when it no longer is in the customer’s financial interest to pay the mortgage as agreed in the mortgage loan contract and they walk away.
FICO plans to prevent through education, by getting the right message to that risky customer as early as possible. FICO has designed a number of credit education materials specifically for servicers to help them communicate to a borrower what is the impacted score if they walk away, and what is the true cost of this decision, she said. “In part because this is such as new customer behavior there is not a broad amount of past experience and expertise about strategic defaults. If we work together to define these products we can find the most effective solution.”
New customer behavior traits are not easy to capture also because not all underwater customers would elect to strategically default. It is a very complex decision making process, she said, so the tool is “definitely analytical.” For instance, the LTV ratio is important, yet customers are complicated and it is a highly subjective decision in their part. The tool is based on the presumption that the customer continues to make all the other payments but does not pay the mortgage when the LTV is 100% or more.
Strategic default is predictable also because data show these borrowers are the riskiest group within the bottom 20% credit scoring population. “Servicers can find up to 80% of all the higher risk for strategic default people within that group,” she said. That information “is very actionable” because it allows servicers “to be very targeted in their communications and strategy.”