New research finds mortgage intermediaries kept over 60% of GSE fee waivers meant for lower-income first-time buyers, with pass-through varying sharply by loan officer.
Market power in consumer finance is often discussed in terms of big institutions setting prices. New research argues that another source can be just as consequential: the individual loan officer sitting across the table from a borrower.
In a working paper, economists Neil Bhutta, Benjamin J. Keys, and Adithya Raajkumar examine what happened after Fannie Mae and Freddie Mac introduced a temporary waiver of certain upfront fees. The policy was aimed at helping lower-income, first-time homebuyers by reducing charges embedded in mortgages that are ultimately sold to, or guaranteed by, the two government-sponsored enterprises (GSEs).
If a fee is waived at the GSE level, the straightforward expectation is that lenders and intermediaries would reflect that lower cost in the pricing offered to borrowers—delivering cheaper loans to the households the program was designed to assist. But the authors find that the savings often did not fully reach consumers.
Using mortgage market data from 2023 and 2024, the researchers estimate that more than 60% of the program’s roughly $1.5 billion in total fee waivers was effectively retained by mortgage intermediaries, rather than being passed through in the form of lower borrowing costs. In other words, a majority of the intended subsidy appears to have been absorbed within the origination channel—reducing the extent of relief experienced by targeted borrowers.
The paper also documents wide differences in how that pass-through played out across individual loan officers. Some intermediaries appear to have delivered much of the benefit to borrowers, while others transmitted little of the waiver at all. The starkest pattern is tied to loan officers’ past behavior: those with histories of originating relatively costly mortgages—after accounting for differences in lenders’ baseline pricing—passed through less than 20% of the waiver, the authors estimate.
That relationship suggests the “pricing power” at issue is not only institutional. It can be personal and situational: loan officers may be able to capitalize on information gaps, consumer confusion, or shopping frictions, keeping a larger share of a subsidy even when the policy is designed to reduce borrowers’ costs.
The findings also raise questions about the effectiveness of post–financial crisis rules intended to curb steering—practices that push borrowers toward higher-rate or higher-fee loans that increase profits for originators or intermediaries. Despite those regulations, the authors conclude that loan officers still have meaningful incentives and ability to direct borrowers into more expensive mortgages, particularly when consumers have limited ability to observe how upstream fee changes should affect the offered rate.
The researchers note that the paper benefited from comments and discussions with housing and mortgage market experts, as well as feedback from seminars and conferences including sessions at Rice University, Freddie Mac, the Wharton School, and meetings of real estate and finance research groups. Keys reports support from research sponsors at Wharton’s Zell/Lurie Real Estate Center. The authors emphasize that their conclusions are their own and do not necessarily reflect the views of the Federal Reserve or the National Bureau of Economic Research.
Why it matters
The fee waiver was designed to make homeownership more attainable for lower-income first-time buyers, but this evidence suggests a large share of the subsidy may have been diverted before reaching households. If intermediaries can systematically capture benefits from targeted policies, future affordability initiatives may need stronger guardrails, clearer disclosure, or mechanisms that make pass-through easier to verify. Policymakers and regulators will likely watch whether similar programs deliver measurable reductions in borrower costs—or mainly increase margins within the mortgage origination process.
