By Clifford Rossi
On May 1, 2023, a set of new, loan-level price adjustment (LLPA) grids for mortgages purchased by Fannie Mae and Freddie Mac mandated by the Federal Housing Finance Agency (FHFA) will go into effect. FHFA’s director stated that the rationale for these changes is “to increase pricing support for purchase borrowers limited by income or by wealth.”
Unfortunately, the FHFA has subverted the economically sound practice of risk-based pricing and in the process has undermined incentives for borrowers to improve their credit.
Imagine that as a safe driver over the years, you’ve enjoyed lower auto insurance premiums than riskier drivers. Then one day, you receive a notice that your premiums, with never having had an accident or moving violation, are going up 300%. Further, you find out that your new premiums are going to subsidize drivers with riskier driving habits and records. Essentially that’s what the FHFA is doing for mortgage borrowers.
Differential or risk-based pricing of key attributes describing the degree of credit risk in a mortgage has been in place for years. Both Fannie Mae and Freddie Mac charge ongoing guarantee fees to compensate the agencies for credit risk on mortgage loans purchased from lenders. Those guarantee fees are based on the risk attributes of those loans and are embedded in a borrower’s mortgage rate. In addition, upfront delivery fees, or LLPAs, are imposed on selected risk attributes such as credit score, loan-to-value (LTV) ratio and loan purpose (e.g., purchase of a home or refinance).
The new LLPA grids differentiate risk via a fee based on whether the borrower is purchasing the home, refinancing the mortgage with limited cash taken back out, or a cash-out refinance. The current LLPA grids are risk-based in the sense that higher fees are assigned to riskier FICO and LTV cells. However, the new grids will increase the cost of borrowing for a sizable borrowing cohort that presents very low credit risk while greatly lowering the cost of borrowing for borrowers that pose significant credit risk to Fannie and Freddie.
The changes between the current and new LLPA grid for purchase mortgages are shown in Table 1 below. The cells shaded in red depict increases in LLPAs while cells shaded green represent decreases. Borrowers with credit scores between 720-759 with LTVs between 80.01- 85% will go up by .75% from .25% to 1%, or a 300% increase on May 1, for example, while borrowers with credit scores less than 620 with LTVs above 95% will drop by 2%.
To put this in perspective, according to Fannie Mae historical credit performance data, borrowers in the low-risk group had a net loss rate of .29% while the high-risk group’s net loss rate was 2.09%, or more than seven times the low-risk cohort. Similarly, the high-risk group has a historical late-stage (i.e., more than 180 days past due ever) that is 6.5 times the rate of the low-risk group on loans originated between 1999-2022. Credit risk clearly does not increase in a linear fashion with credit score and LTV but rather results in a sharp acceleration when lower credit scores are combined with higher LTVs. Lowering fees invites more high-risk borrowers into the credit portfolios of both GSEs though they represent a very small portion of new GSE-eligible mortgages. More than 25% of prospective borrowers will face an increase in LLPAs while borrowers with credit scores less than 660 make up about 2% of new originations.
Figure 1: Changes (in percent) in LLPAs Between Current and New Grids
Note: numbers in red represent high risk credit scores or LTVs and blue represents high risk combinations of credit score and LTV.
Table 2 provides a sense of the impact of these changes on borrowers taking out a 30-year fixed-rate mortgage assuming a loan size of $300,000 and a mortgage rate of 6.4%. Borrowers with FICOs between 720-759 with LTVs of 80.01-85% would see an annual increase of about $360 or about a 1.6 percent increase in their payment overall. While this does not seem to be a substantial increase, on top of higher mortgage rates and inflation already embedded in the economy, it creates an additional financial headwind for these borrowers.
A countervailing argument can be made that higher risk borrowers will see significant reductions in their mortgage payments and that the risks to the GSEs from attracting more of these borrowers into their credit portfolios are offset by higher fees on the low credit risk borrowers. Still, such a policy puts many high-risk borrowers at risk given their risk profile at the wrong time of the economic cycle.
Table 2: Change in Monthly Mortgage Payment Under new LLPAs
The FHFA forced both GSEs to essentially flatten the actuarially fair pricing relationships of credit score and LTV to credit risk for the sake of improving housing affordability of borrowers with poor credit characteristics. However, that policy does nothing for higher credit risk borrowers to improve the long-term sustainability of retaining their home once the loan is made. We found out during the years leading up to the 2008 Global Financial Crisis that policies intended to help marginal borrowers become homeowners ultimately resulted in many heartbreaking stories of foreclosure. While supporting homeownership across all communities and incomes is a laudable objective, imposing affordable housing policy on risk-based pricing is ultimately an ineffective policy mechanism that comes at the expense of burdening a large segment of borrowers including those intended to fare better after May 1.
Clifford Rossi is a Professor of the Practice and Executive-in-Residence for the University of Maryland's Robert H. Smith School of Business. He has nearly 25 years of experience in the financial services industry where he held senior risk management positions at several of the largest financial institutions including Fannie Mae and Freddie Mac.
MEDIA CONTACTRegister for reporter access to contact details
Professor of the Practice & Executive-in-ResidenceUniversity of Maryland, Robert H. Smith School of Business