By Dan Smith, President and CEO, Consumer Data Industry Association
On July 1, Fannie Mae and Freddie Mac released historical FICO Score 10T and additional VantageScore 4.0 data covering more than a decade of mortgage performance. The release is an important milestone in the ongoing transition to modern credit scoring models, and it gives lenders, investors, and other stakeholders the information they need to evaluate how these newer models perform in mortgage lending. CDIA welcomes the adoption of these updated scoring models because they have the potential to make mortgage underwriting more accurate, more predictive, and more inclusive.
For consumers, credit score modernization holds real promise. By incorporating more predictive data and expanding the number of borrowers who can be evaluated, it can help lenders make more accurate decisions and broaden access to homeownership. But those benefits depend on one critical foundation: complete and accurate credit data.
This is why the tri-merge credit report is the industry’s gold standard. It provides the most complete and accurate picture of a borrower’s credit history, supporting better lending decisions, stronger competition, and greater confidence throughout the mortgage market. More data, not less, is required to protect lenders and taxpayers and open more opportunities for borrowers. Therefore, a system built on more complete data helps lenders see borrowers more clearly. That is a good day for consumers.
As the industry works to put these models to use, some stakeholders have renewed calls to pull a mortgage applicant’s credit from a single bureau instead of all three. That argument is not new, and in its most recent form, a supporter of a single-pull relies on an analysis of roughly 105,000 applications, which found that about two-thirds of borrowers would stay in the same pricing bucket under a single file and about 90 percent would stay within one bucket.
The conclusion offered is that moving away from the tri-merge would change very little. But that framing skips the harder question – what’s the impact on the 30 percent of borrowers who would be impacted?
A borrower’s credit score helps determine which pricing bucket they fall into, and that bucket influences the loan-level pricing adjustments that ultimately affect the interest rate and cost of a mortgage. If two-thirds of borrowers land in the same bucket, roughly one-third do not. That is roughly a third of mortgage decisions made on a less complete view of a borrower’s credit, depending on which single file a lender happened to pull. For the family at that closing table, a different bucket is not a statistic. It could mean paying a higher interest rate and a higher monthly mortgage payment for a reason that has nothing to do with how they actually manage their credit.
Those buckets, the credit score ranges Fannie Mae and Freddie Mac use to set loan-level price adjustments, determine the fees built into a mortgage, so a move from one bucket to another may change what a borrower pays.
One bucket can mean real money for a household. Andrew Davidson & Co. found that within the common 640 to 779 score range, a ten-point score difference moves a borrower into a different pricing bin more than 80 percent of the time, and on a $350,000 loan that can mean $3,000 to $5,000 over the life of the loan. A single file can move a borrower in either direction. Some would land lower and may pay more than their true risk warrants, simply because the one file that was pulled did not tell their whole story. That uncertainty is really the heart of it. Furnishing data to the nationwide credit reporting agencies is voluntary, so there are legitimate differences in the information each consumer report contains, and those differences can grow as new data sources such as rental payment history are added to credit files. Only a tri-merge gives lenders a complete picture. Using a single file risks omitting predictive data from the decision. the
That same uncertainty carries through to the market participants who price and ultimately hold the risk, including the GSEs and investors. When pricing is built on a less complete picture, no one downstream knows whether a borrower would belong in a different bucket if all the data were considered, and that uncertainty is passed through as a premium on rates. The market is not waiting for this debate to be settled. Changes away from the status quo introduce risks to the system which some economists argue will make lending more expensive for all borrowers.
That is the real cost of a single-bureau standard. It is not simply that some borrowers would see a different price. It is that no one, not the borrower, not the lender, not the investor, can say in advance who those borrowers are or whether the price they received reflects their actual risk.
Credit score modernization and the tri-merge are working toward the same goal, which is an accurate picture of each borrower. The tri-merge exists for a reason. It ensures that a borrower’s full credit history is considered. It promotes data accuracy, supports market competition, and strengthens investor confidence, and above all it protects the person with the most at stake in getting the price right, which is the homebuyer.