A coalition of trade organizations — including the American Bankers Association (ABA), the Housing Policy Council (HPC), the Independent Community Bankers of America (ICBA), the Mortgage Bankers Association (MBA) and U.S. Mortgage Insurers — sent a letter on Monday to the Federal Housing Finance Agency (FHFA) to express concerns about the FHFA’s intent to publish the VantageScore 4.0 historical dataset.

Announced in February, the agency had originally planned to publish the dataset next year, but it chose instead to advance the timeline to third-quarter 2024 in order to “better support market participants with this transition” to a new bi-merge credit reporting requirement that is also being challenged by trade groups and lawmakers.

In the joint letter, the organizations contend that the data to be published under the plan is insufficient for their needs.

“Based on our understanding of what Fannie Mae and Freddie Mac (the GSEs) anticipate  including in the release, our members have expressed concerns that the information will be  insufficient to meet their data analytics needs,” the letter stated. “We hope that by providing early, proactive, and constructive feedback to FHFA and the GSEs, there will be adequate time to augment the data made available in the planned 3Q 2024 data release as well as in any subsequent data release.”

The VantageScore dataset that is scheduled to be made available includes information going back to 2013, which is not far enough, the groups assert.

“FHFA was comfortable approving VantageScore 4.0 for use by the GSEs based on limited historical data,” the letter reads. “Many industry participants have raised concerns about this, given that the data was generated during a time when the housing market was appreciating, and unemployment was declining to historic lows. Our members continue to believe that it is imperative to have credit reporting data ‘through the cycle’ back to 2003 given the sensitivity of mortgage default and prepayment to origination credit scores.”

FHFA’s current lack of a plan to include “through the cycle” data increases the importance of the data itself, the letter explained. Because of that, the groups want to make sure there is specific criteria that is included when the dataset is released.

These include “unrestricted use of data contained in the mortgage-backed security (MBS), credit risk transfer (CRT), and historical performance disclosures, which currently prohibit using the data for developing, analyzing, or calibrating models (among other restrictions), which prevents users from comparing the accuracy of the new scores’ performance and accuracy to that of the Classic FICO, since the origination and updated Classic FICO scores are contained in the disclosures,” the letter reads.

To allow users to compare and conduct the modeling they need for new credit scores to the Classic FICO model, “the GSEs must provide a legally binding waiver of these restrictions or provide a separate database with exactly the same set of fields as are currently available in each of the disclosures, but that is not subject to the unworkable contract terms,” the groups said.

The groups also request “sufficient data to evaluate the impact” of the new representative score calculations, as well as “sufficient detail” to evaluate the accuracy “and fair lending analysis” of bi-merge combinations.

The groups also request a sample dataset under the new format to be provided before the third quarter, which would provide an “opportunity to identify and address any additional gaps or systemic issues before the historical archives are available,” the letter explained.

“These recommendations for additional data are critical to the successful implementation of the VantageScore 4.0 and FICO 10T credit score models as well as the bi-merge credit reporting policy,” the letter added. “For this reason, we also recommend the delay of any scheduled data release until these recommendations can be accommodated.”

In late 2022, FHFA announced that it would replace the Classic FICO credit model — which Fannie Mae and Freddie Mac have relied on for roughly 20 years — with new models (FICO 10T and VantageScore 4.0) that use trended credit bureau data.

Since then, there have been a number of delays and challenges in getting these changes over the finish line, as mortgage companies, trade groups and lawmakers have increasingly scrutinized the implementation timeline.

Related