Multifamily Prepayment Penalty: Repricing the Mid-Hold Exit

Published October 6, 2026

The 5-year Treasury par yield was 5.01% on October 1, 2026 and the 10-year was 5.24%, per the U.S. Treasury's daily yield curve. For an owner holding an agency loan written when the 10-year sat well below 2%, that single fact moves the exit arithmetic more than any rent comp does. The yield maintenance premium on that loan has likely collapsed to its contractual minimum, and most models are still carrying the number that was true at origination.

The prepayment line is unusual among underwriting inputs because it is a formula, not an estimate. It can be computed almost exactly from the note and today's curve. It is also one of the few lines that got cheaper while everything else in the capital stack got more expensive.

Why the Multifamily Prepayment Penalty Inverted This Year

Yield maintenance exists to compensate the lender for reinvesting prepaid principal at a lower rate than the note carries. The calculation is, in simplified form, the present value of the differential between the note rate and the yield on a Treasury of comparable remaining term, applied to the scheduled payment stream. When the Treasury yield climbs above the note rate, that differential goes negative — there is no lost yield to make whole, because the lender can reinvest at more than the loan was paying.

That is why the contract carries a floor rather than a refund. Per Fannie Mae's DUS Program Overview, the prepayment premium during the yield maintenance period is "the greater of: (a) 1% of the amount of principal being prepaid" or the yield maintenance calculation, and the same document notes a 1% premium may apply after the yield maintenance period ends. In a rising-rate market the greater-of test stops being a formality and becomes the whole answer. The multifamily prepayment penalty on a low-coupon note is then approximately 1% of the balance, regardless of how many years remain.

Nothing about this is a forecast. It is a reading of the curve on a specific date, and the curve could retrace. The point is that the input is observable today and most models are not observing it.

The 1% Floor Is Doing the Work on 2021-Vintage Paper

The gap is the whole story. The 10-year Treasury yield stayed well below 2% throughout 2021, per Treasury par yield data, and agency fixed-rate loans priced off that base. Agency 10-year fixed quotes sat roughly in the 6.30% to 6.70% range as of October 1, 2026, per Northmarq's published rates and spreads, against a 30-day average Secured Overnight Financing Rate near 3.75% on the same page. A note written in the low 3s now sits roughly 300 bps or more below where a replacement loan would price.

Run that through the yield maintenance differential and the result is not a smaller premium — it is no economic premium at all, floored at the contractual 1%. The structural detail that matters more than the arithmetic is the yield maintenance period, which is a term of the note rather than a market variable. Freddie Mac's conventional fixed-rate product sheet describes yield maintenance until securitization followed by a two-year lock-out and defeasance thereafter, with "no prepayment premium for final 90 days," while Fannie Mae and Freddie Mac differ in structure in ways worth reviewing against the actual documents rather than by program reputation.

A model that carries a blanket prepayment reserve — three points, five points, whatever the sponsor used in 2021 — is wrong in both directions depending on the note. Reading the prepayment section is a 10-minute job that moves net proceeds by more than most of the revenue assumptions that get debated for a week.

What the Prepayment Line Actually Costs: A Worked Example

Example — illustrative. A 240-unit Class B property in a Sun Belt secondary market, acquired in mid-2021 for $35M with $26M of Fannie Mae fixed-rate debt and about $9M of equity. Note rate of 3.40%, 10-year term, 30-year amortization, yield maintenance running nearly the full term. By October 2026 the balance has amortized to roughly $23.2M with about five years remaining. The owner is evaluating a sale.

The simplified differential: the note pays 3.40%, a 5-year Treasury yields about 5.01%, so the reinvestment rate exceeds the coupon and the computed yield maintenance amount is zero. The floor governs.

The 2021-era row is a counterfactual, not a quote: approximately 2.40% on a $23.2M balance is about $557,000 a year, and discounting five years of that at roughly 1% produces something near $2.7M. Actual agency formulas discount the scheduled principal and interest stream rather than a flat differential, and they vary by note, so treat both rows as directional.

The difference is approximately $2.5M of net sale proceeds on about $9M of original equity — close to 27 points of equity, arriving entirely from the prepayment section of a loan document. For a sponsor whose Internal Rate of Return (IRR) promote hurdle sits anywhere near the base case, a swing of that size is the difference between a promote and a return of capital. It is also the line most likely to be carried over from the original acquisition model untouched, the same way property taxes get trended instead of reset.

Defeasance Moves the Same Way, With Different Friction

Defeasance substitutes a portfolio of government securities for the property as collateral, with a successor borrower assuming the debt, and it responds to rates in the same direction. Per J.P. Morgan's agency lending commentary, rising rates lower the cost because "higher Treasury yields mean borrowers need to purchase fewer securities to generate cash flow to cover the remaining mortgage payments." The same commentary notes that falling rates do the reverse.

The difference is that defeasance is a transaction, not a wire. It involves a securities purchase, a successor-borrower entity, consultants, and documentation, and J.P. Morgan's note describes the process as complex enough to warrant specialist help. Those costs are process-driven rather than rate-driven, which means they do not shrink when the premium does, and they are quoted rather than computed. A model that replaces a yield maintenance estimate with a clean 1% and nothing else may understate a defeased payoff.

Timing is the other asymmetry. Securities pricing happens close to closing, so a rate move between underwriting and payoff flows straight through, and a lock-out window can make the cheapest theoretical cost unavailable in practice. Verify the available route and current pricing directly with the servicer or a defeasance consultant for the specific loan.

Where This Breaks

The floor cuts both ways. If the long end retraces below a note's coupon, yield maintenance comes back, and the October 2026 curve is a reading rather than a trend — recent moves at the long end have been attributed to fiscal rather than policy factors in press coverage, which is not a stable base for a five-year assumption. A model that hard-codes 1% is making the same mistake as one that hard-codes five points.

Lock-out is the larger practical obstacle. A Freddie Mac loan in its post-securitization lock-out cannot be prepaid at any price, per the product sheet's own terms, so a cheap theoretical premium may be irrelevant for the window in which a sponsor actually wants to transact. Partial prepayments and curtailments are frequently restricted outright, which rules out using the cheap window to deleverage incrementally.

The strongest counter-argument is that the low-rate loan is the asset. Prepaying a 3.40% note into a market quoting 6.30% to 6.70% destroys far more value than a 1% premium saves, which generally argues for marketing the assumption instead and treating the cheap premium as a fallback rather than a plan. That calculus is a separate exercise from the one above, and it depends on assumption fees, supplemental availability under the agency caps, and whether a buyer can be found who values the paper.

Scope is worth stating plainly too. The Mortgage Bankers Association reported that $875B of commercial and multifamily mortgage balances, or 17% of the $5.0 trillion outstanding, were scheduled to mature in 2026 — down approximately 9% from the $957B scheduled in 2025. Loans reaching maturity are generally open or nearly open to prepayment anyway, so this analysis applies to mid-hold decisions rather than to the bulk of the volume moving through the maturity wall.

The Underwriting Takeaway

The prepayment premium is one of the few exit inputs that can be calculated rather than assumed, and it is routinely assumed. Three things follow. The note's prepayment section belongs in the model as a live calculation against the current curve and the remaining term, not as a reserve percentage inherited from the acquisition file. The yield maintenance period and any lock-out deserve their own dates, because a premium that is cheap in year six is unavailable in year two. And where the prepayment cost has fallen, the honest next question is whether prepaying is the right route at all, given what a sub-4% note is worth to a buyer in a 6%-plus market.

A deal that pencils on a prepayment assumption nobody re-read is not mispriced by a rounding error. On the illustrative loan above, it is mispriced by roughly a quarter of the equity.

UWmatic is an AI-powered underwriting platform built for multifamily and mobile home park investors. Price the prepayment premium off the current curve and the actual note terms, and see what the corrected exit cost does to net proceeds and sponsor IRR. Try the free underwriting calculator →

This analysis reflects current market interpretations as of the publication date and may evolve as new data becomes available. Figures cited are drawn from public sources including the U.S. Treasury's daily par yield curve, Fannie Mae's DUS Program Overview, Freddie Mac's conventional fixed-rate product sheet, J.P. Morgan's agency lending commentary on defeasance, Northmarq's published rates and spreads, and Mortgage Bankers Association maturity data, and are subject to revision. Nothing in this post is investment advice; readers should conduct their own diligence and consult qualified professionals before making investment decisions.