Post OBBB: What's next for the SAVE plan student loan borrowers?
Plus maps of where the mortgagors with SAVE plan student loans are located
The “Byrd Bathing” (last minute change) of some of the repayment provisions in the OBBB’s student loan overhaul muddied the waters on just when the 7.84m (4.5m of whom are likely mortgagors) still in forbearance and currently paying nothing SAVE plan student loan borrowers will be forced back into repayment, but it didnt change the reality that these loans will require repayment again, possibly soon.
Pre “Byrd Bath” OBBB impact on initial repayment of SAVE loans
All Income Driven Repayment student loans (SAVE, PAYE, ICR, IBR etc.) would have been forced into a new Income Based Repayment plan with payments at 15% of discretionary income (AGI minus applicable poverty line) within 9 months of bill passage, so by early April 2026. Given that concrete and relatively short timeframe, it seems likely the Dept. of Education and the courts that enjoined the SAVE plan would have let the SAVE loans stay in administrative forbearance until that switchover for the administrative ease of it. But that did not happen.
Post “Byrd Bath” OBBB impact on initial repayment of SAVE loans
The bill, as passed, requires SAVE borrowers to select (and start repayment under) a new plan by July 2028. While in theory that could allow for 3 more years of SAVE administrative forbearance, it seems incredibly unlikely that the Trump Department of Education/8th Circuit would allow that to be the case. Indeed, last week the Department of Education announced that SAVE student loans would start accruing interest again in August.
Thats pretty important, when interest wasnt accruing, there was little impetus for a borrower to switch from SAVE. Sure eventual forgiveness might be delayed, but its effectively an interest free loan for a while. Starting in August though, unless the borrower decides the Repayment Assistance Plan created in the OBBB, and set to become reality next summer, is their long term best option, it probably makes sense to switch to a different plan and have payments count towards forgiveness again. Unless, of course, the borrower is unable to manage making the payments under one of those student loan plans and needs to ride out “zero payment” for as long as possible, interest accrual be damned.
Accordingly, I dont expect the total 7.84m SAVE borrowers as of March 31st to decline much when studentaid.gov updates their data for end of June shortly, but the update for end of September, along with fall receipts into the TGA deposit category for the Department of Education will be more interesting to watch. It wont be dispositive on, but will provide some indication of how many of these SAVE borrowers were strategically hanging out in the plan vs. have a budgeting crisis on hand.
On the other hand, the Department of Education/8th Circuit could force transition into an existing plan at any time in the future prior to the mandatory July 2028 choice. If I had to guess, I think the SAVE plan borrowers who dont self select out due to interest accrual will be forced to choose (out of SAVE) by the time RAP is implemented next summer (again out of administrative ease), but it could be well before that (or after thru July 2028). Ironically, the “Bryd Bath” changes made everything much more muddied for the immediate future of SAVE.
Mortgagors with SAVE plan loans
The methodology I developed in previous posts to determine the intersection between student loans and mortgagors (most recent post here)
seems to have support from data presented by ICE (formally Black Knight) in their July Mortgage Monitor report (https://mortgagetech.ice.com/publicdocs/mortgage/imt-july-2025-mortgage-monitor-report-sMm33rhnRWDK.pdf).
ICE found that the share of mortgage holders with student loan debt was roughly 30% of FHA borrowers, 20% of VA borrowers, and 17% of GSE borrowers. This is very much in line with my methodologically derived estimations of 30% FHA, 24% VA and 17% for GSE borrowers presented in my linked piece.
ICE intersects their loan level mortgage data with credit report tradelines from Transunion to determine their percentages, so for their population of covered mortgages, they know exactly how many of them carry student loan debt. I derive it for each mortgage in the MBS pools by looking for unusually large non housing debt given the income of the borrower. For each such loan, student loan debt is very very likely the explanation for it. So I dont know with certainty for any given mortgage whether it has student loan debt and Im sure there are a few with it that I dont count, and a few without it that I do, but in a broad aggregate sense, the methodology seems to work given the agreement with ICE’s actual numbers.
What ICE does not go into is the number of mortgagors with the SAVE plan loans. These matter because the repayment shock on these mortgages hasnt hit yet. ICE probably could report this data by looking for zero payment on the tradelines and maybe they will in future Mortgage Monitors. For now however, my methodologically derived mortgagors with SAVE student loans is the only data out there. The methodology I use is to calculate what the borrowers payments would be under the various other existing plans using the borrowers income and estimated student loan balance, compare it to the payment under SAVE and figure that when SAVE was the lowest payment, those borrowers probably elected that plan. Again not perfect, but should in aggregate work fairly well. When I compare the range of student loan balances the methodology calculates for the mortgages in SAVE plan repayment, it lines up reasonably well with studentaid.gov’s breakdown of Income Driven Repayment plan (dominated by SAVE plans) balance ranges.
Because I have these MBS pool mortgages linked to their HMDA records (necessary to get the qualifying income to derive monthly non housing related debt from the borrowers DTI at origination), I also have the relatively precise (census tract) location of those mortgages.
Where are the mortgages with SAVE student loans located?
There is a lot in Texas, particularly in the outskirts of major metros. This image shows the Houston area.
There is a large variance between localities with some census tracts having a big concentration, while other census tracts have relatively few. As shown in the image that captures the details for 5429.02 in the western part of Harris County, I list the estimated number of mortgagors with SAVE loans from each portfolio (represents the loans my methodology identifies as likely in SAVE scaled to the full portfolio for the loans I cannot link to HMDA) along with the average payment shock if the borrower chose/was forced to switch to the Income Based Repayment plan (the 10% of discretionary version for borrowers only with loans post July 2014, borrowers with loans before that would be 15% so just multiply it by 1.5) and also the extended plan (amortized over a 25 year period). The point of this is to provide a general sense of what the payment shock will look like in each locality since it depends not only on loan size but also income levels of the borrowers in that locality. There are other plans like the graduated plans that would have a somewhat cheaper initial payment (the payments escalate over time in those plans) but its usually not too much less than the extended plan (which also tracks pretty closely to a consolidated standard plan that amortizes over a number of years that grows with the balance). There are also plans that would be higher, (like the 10 year standard plan) but I doubt may borrowers would choose that by will or by force.
I have scaled the “heat” in the maps by raw number of estimated mortgages with SAVE student loans in the census tract and chosen nearly 1000 levels of scaling. This comes with the downside that tracts with less than 100 mortgages with SAVE loans will all look fairly white, but allows the tracts with outlier numbers of these loans to pop more. Either way you can look at any census tract in the US (still minus Connecticut, Ill get to it soon I promise) and see the raw data for yourself.
The link to the map files are beneath the paywall.