Trade organizations express VantageScore data set concerns to FHFA

Data in planned release might be too limited to thoroughly prepare for credit model transition

Doubling down about trepidations regarding the move to replace traditional credit score models with new credit bureau data, five mortgage trade organizations have co-signed a letter to the Federal Housing Finance Agency (FHFA), expressing concerns about the upcoming release of a VantageScore 4.0 data set.

The release is related to the upcoming credit score transition at the government-sponsored enterprises (GSEs), Fannie Mae and Freddie Mac. The FHFA on October 2022 approved two new credit score models, FICO 10T and VantageScore 4.0, for use by Fannie and Freddie, replacing the Classic FICO model that the GSEs have used for about two decades. The historical data set, which is planned to be released early in the third quarter, was planned by the FHFA to make for a smoother transition for market stakeholders.

But the groups, including the Mortgage Bankers Association, American Bankers Association, Housing Policy Council, Independent Community Bankers of America and U.S. Mortgage Insurers, are wary that the information included in the release won’t meet their data analytics needs.

Foremost among the groups’ apprehensions is that the data in the release might encompass too limited of a timeframe to be useful.

“It is our collective understanding that the VantageScore dataset will include information going back to 2013,” the letter reads. “Moreover, we understand that FHFA was comfortable approving VantageScore 4.0 for use by the GSEs based on limited historical data. 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.”

It has been proven that the correlation between credit scores and mortgage performance has varied significantly over economic cycles, the groups asserted. This has caused trepidation that using models calibrated only for the period after Global Financial Crisis would fall short of prudential regulatory standards.

The letter also expresses concerns that the upcoming data set release will not include Classic FICO scores for comparison. Rather, users of the data set will be instructed to gather that information from data contained in mortgage-backed security, credit risk transfer and historical performance disclosures. Those disclosures, however, are under strict prohibitions against using the information contained therein for certain data analysis models, effectively preventing users from comparing the accuracy of the new scores to those of the longstanding Classic FICO models.

The groups, then, are calling either for the FHFA to provide a separate database of Classic FICO information matching the data available in the aforementioned disclosures, or for a legally binding waiver allowing them to use disclosure information for data analysis.

According to the text of the letter, the groups hope that by expressing their concerns proactively and early, there will be adequate time for the FHFA to augment the data currently planned for release. The provision of additional data that the groups deemed “critical” in this initial release is important for the industry to meet the FHFA’s target implementation date of fourth-quarter 2025 for the VantageScore 4.0 system.

“For this reason, we also recommend the delay of any scheduled data release until these recommendations can be accommodated,” the letter reads. “While we are certainly committed to working with FHFA and the Enterprises on this project, it is also critical that FHFA understand and consider the broader implications and potential negative impacts on the pricing of mortgage credit if industry stakeholders are unable to obtain the data necessary to complete their analysis.”


More Headlines