BofA: Add GSE scores to models to predict delinquencies

Proprietary score data that influential government-related mortgage buyers recently revealed could give housing finance firms an edge in predicting loan performance, Bank of America researchers concluded in a new study.

"We think the providing of the agency score is a timely, robust addition to credit modeling, offering a consistent alternative," Jeana Curro, Chris Flanagan, Ge Chu and Ko-Hsiang Kao, wrote in a report on metrics that the GSEs and their oversight agency rolled out earlier this month.

Federal Housing Finance Agency Director Bill Pulte, who oversees the GSEs, said in a social media post that he welcomed analysis in the bank's study that shows the enterprises' scores "overall serve as better indicators of borrower delinquency risk" than some other measures.

Processing Content

While the researchers found the scores to be strong mortgage credit metrics, they also noted that the visibility they have into them is limited and other measures will still be needed. 

How external and internal scores compare

The characteristics and roles played by different credit score types used in the mortgage process follow.

Both the traditional credit scores the agencies use today and the newer models they are adopting run on a 300-850 scale, and both factor into the enterprises' pricing grids. The Classic FICO traditionally used represents a "snapshot" of borrower credit. Trends in data over time are part of newer metrics.

External scores get pulled at loan application, with initial visibility to borrowers and lenders before submission to the enterprises and disclosure to investors. These metrics play a role in loan pricing grids the GSEs have.

The GSEs' respective internal scores have different numerical scales than their external counterparts and get generated through automated underwriting, according to B of A. The enterprises traditionally get visibility into these at origination. Broader visibility occurs via mortgage-backed securities disclosures.

Fannie's scores span a 400-950 range. Freddie's scores start at 200 and go up to 1,000. Both enterprises' scores incorporate some limited trended and alternative data that are the hallmarks of the more modernized external metrics the GSEs are implementing to fulfill a legal mandate.

While external scores consider borrowers' broader credit profiles, the enterprises' focus on mortgage delinquencies.

"Due to their wider ranges and higher upper limits, both agency scores run higher than Classic FICO," the Bank of America researchers wrote.

Fannie's score tends to be around 10-20 points higher than FICO's classic metric on average and more or less in line with modernized external scores. 

Freddie's score generally runs 20-30 higher than the traditional external credit measure. It's typically 10-15 points above the newer FICO 10T and VantageScore 4.0.

The GSEs have begun buying VS4 loans from lenders with their scores needing to be around 20 points above the traditional FICO measure to get the same pricing to start. The enterprises also plan to add 10T.

GSE score benefits, uses and limits

While differences in Fannie's and Freddie's measures complicate comparison, the Bank of America researchers said they were able to see that the enterprises' scores beat other measures when it came to their predictiveness related to whether a mortgage would go 90-plus days delinquent after 24 months.

With these scores available to give lenders improved credit modeling, they may mitigate concerns about lender choice of external metrics leading to a gaming of the system that could hurt mortgage performance, the researchers said.

"The industry concern that lender choice among FICO and VS4 would meaningfully understate delinquency risk is not unfounded, however we find that the agency score largely helps to avoid this distortion," they wrote.

While an agency score is helpful in sizing up delinquencies, mortgage companies should use other measures to get a handle on prepayments.

"That score would do little to anticipate faster (or potentially slower) voluntary speeds and convexity changes," the researchers said.


For reprint and licensing requests for this article, click here.
Secondary markets Servicing Credit scores Underwriting Mortgages
MORE FROM ASSET SECURITIZATION REPORT
Load More