In February, the Mortgage Bankers Association released its 2026 commercial forecast built on a specific view of the Federal Reserve: that it was, in Mike Fratantoni's words, "close to end of its current rate-cutting cycle". One more cut in the federal funds target in 2026, and then a plateau. Every sponsor who chose to extend rather than refinance last year was underwriting some version of that same assumption.
By June, it had inverted. Nine of eighteen FOMC officials projected the federal funds rate finishing 2026 above its current 3.50%–3.75% range, and the median projection moved from 3.4% in March to 3.8%. The forecast that justified waiting is no longer the forecast.
That matters because of who is doing the waiting. A large and growing population of commercial loans has reached maturity, failed to size a par takeout, and is being carried current on interest while the sponsor waits for cheaper money. These are not distressed assets. In most cases the business plan worked. The property performs, coverage on the current debt service is comfortable, and the only thing wrong with the loan is that the new one won't be big enough to pay off the old one. That is a capital-structure problem with a knowable price, not a credit event — and it is materially cheaper to solve twelve months before maturity than sixty days after it.
The thesis, and it runs through everything below: size your takeout off the constant you can get today, not the one you underwrote.
What Lives in the Gap
We sized this cohort three weeks ago in Past Maturity, Still Paying, so the short version here. Trepp's June 2026 data put the headline CMBS delinquency rate at 7.35%, down 20 basis points on a large lodging cure. Fold back in the loans that are past maturity but current on interest and the rate registers 9.53% — up 36 basis points on the month and a multi-year high. That cohort is 2.18% of loans outstanding. The spread between those two numbers is the market's deferred-refinancing inventory.
One June figure is worth adding, because it describes the exit from that bucket when waiting stops working: non-performing matured balloons accounted for 65% of newly delinquent balances. The dominant failure mode in this cycle is not a borrower who stopped paying. It is a borrower who ran out of runway on a loan they were still servicing.
A note on how to read that. June is one month, and a headline decline driven by a single lodging cure is not a trend. The durable comparison is the year-over-year move in the matured-balloon-inclusive rate — up 82 basis points — not the 36 basis points it moved in June. The cohort is growing on a scale that survives any one month's cures.
The Assumption Underneath It
On July 29 the Committee held the target range at 3.50%–3.75%. The hold was expected. The vote is the part worth your attention: 9–3, with all three dissenters wanting to move in the other direction. Beth Hammack, Neel Kashkari, and Lorie Logan each preferred a quarter-point increase at this meeting. Three dissents pointing the same way is the first such alignment since September 2016.
The statement keeps the language that explains them: inflation "remains elevated relative to the Committee's 2 percent goal," set against economic activity the Committee still reads as expanding at a solid pace. Chair Warsh added no forward guidance in the press conference beyond allowing that higher rates could be part of the answer on inflation.
Put that next to the February baseline this post opened with. MBA's forecast assumed one more cut and then a plateau. Five months later, the live disagreement inside the Committee is not when the next cut arrives — it is whether the next move is a hike. Markets went into the meeting pricing roughly a one-in-three chance of one.
Now the part that actually sizes a takeout, because it is not the fed funds target. The long end sold off on the decision. The 10-year Treasury rose about 5 basis points to 4.657% on July 29 and traded near 4.70% on July 30; the 30-year moved further, up more than 9 basis points to 5.193%. A sponsor waiting on cheaper permanent money got the opposite of the news they needed — not because the Fed moved, but because the curve repriced the odds that it will, and it repriced at the long end — the part of the curve a permanent loan is priced against.
The Worked Example
Value-add multifamily, Northeast secondary market. Originated Q3 2024 on a 24-month initial term, maturing now. The sponsor executed — they hit their stabilized NOI. Everything below traces from two numbers: $315,000 of stabilized annual NOI against a $3,500,000 maturing balance.
Where the loan stands today
The loan floats at SOFR plus 350. SOFR was 3.65% for July 28, 2026 per the New York Fed — the most recent published value as of this writing — and it resets daily, so treat this as a snapshot, not a fixed input. The Fed's hold on July 29 leaves the short end where it was; a hike, if the three dissenters eventually carry the room, passes straight through to this leg.
- All-in rate: 3.65% + 3.50% = 7.15%
- Interest-only carry: $3,500,000 × 7.15% ÷ 12 = $20,854/month ($250,250/year)
- Coverage on current debt service: $315,000 ÷ $250,250 = 1.26×
- Debt yield on the maturing balance: $315,000 ÷ $3,500,000 = 9.0%
Say that plainly: the loan is performing comfortably. At 1.26× coverage, nothing in this file looks like distress, and no servicer is calling. Debt yield belongs on the maturing balance, as above — computing it on new DSCR-sized proceeds is circular, since it collapses to DSCR times the constant and tells you nothing the coverage test didn't.
Takeout sizing, 30-year amortization
Proceeds are capped by the coverage test: annual NOI divided by the required DSCR gives maximum annual debt service, and dividing that by the mortgage constant gives the loan.
| Takeout rate | Constant | DSCR 1.20 | DSCR 1.25 | DSCR 1.30 |
|---|---|---|---|---|
| 5.50% (as underwritten) | 6.81% | $3,852,664 | $3,698,557 | $3,556,305 |
| 6.25% (representative today) | 7.39% | $3,552,767 | $3,410,657 | $3,279,478 |
| 6.75% (stress) | 7.78% | $3,372,659 | $3,237,752 | $3,113,223 |
Against a $3,500,000 payoff. The 6.25% takeout is representative of where small-balance multifamily permanent pricing sits, re-checked against lender sheets published after the FOMC decision: apartment loans under $6 million quoted from 6.11%, larger multifamily starting near 5.70%, and CMBS from 6.63%, all as of July 29, 2026. Those sheets did not move on the decision, and the 10-year is within a basis point of where it closed a week ago — so 6.25% for a 7-year at 75–80% leverage still holds. It is not a quote. Your deal prices to its own sheet on its own day.
The finding. The sponsor underwrote a 5.50% takeout at 1.25× coverage and modeled $198,557 of cushion above par. At a representative current takeout, that same executed business plan sizes to $3,410,657 — $89,343 short of the payoff. They hit every operating assumption they made. The rate assumption alone moved them from a cash-out refinance to a cash-in one.
Which constraint actually binds
Sponsors routinely solve for coverage and get surprised by leverage, or the reverse. Both tests run, and the tighter one governs.
| Exit cap | Implied value | 65% LTV | 70% LTV |
|---|---|---|---|
| 5.50% | $5,727,273 | $3,722,727 | $4,009,091 |
| 5.75% | $5,478,261 | $3,560,870 | $3,834,783 |
| 6.00% | $5,250,000 | $3,412,500 | $3,675,000 |
| 6.25% | $5,040,000 | $3,276,000 | $3,528,000 |
At a 5.75% exit cap — inside the 5.5%–6.5% range CBRE data supports for secondary-market and value-add multifamily — the 65% LTV test allows $3,560,870, which is more than the $3,410,657 the coverage test permits. DSCR binds. Move the exit cap to 6.25% and the value falls enough that 65% LTV allows only $3,276,000, and now LTV binds and the coverage table is no longer your answer. A 50 basis point difference in an appraiser's cap rate silently changes which test is setting your loan amount.
What the Constant Actually Does
Here is the part worth internalizing, and it cuts both ways.
Proceeds are sized by the mortgage constant, not by the rate — and on a 30-year amortization the constant moves less than proportionally. Going from a 5.50% takeout to 6.25% is 75 basis points of rate but only 58 basis points of constant, 6.81% to 7.39%. Supportable proceeds at 1.25× fall from $3,698,557 to $3,410,657 — a 7.8% reduction. That is the reassuring direction: a 75 basis point miss on your rate assumption does not cost you 75 basis points of anything. Amortization absorbs part of it.
Now the same arithmetic, run the other way, and this is where I have to correct a piece of conventional wisdom I have repeated myself.
The instinct is that a 25 basis point move barely registers on a takeout this size. On this loan, that is wrong. From a 6.25% takeout, 25 basis points of coupon moves the constant from 7.39% to 7.19% and is worth $91,967 of proceeds. The gap is $89,343. One cut, fully passed through, would close it — with $2,624 to spare. The gap is almost exactly one rate cut wide.
That is the measurement, and it is why the gap looks so temptingly close to solving itself. But note what July 29 did to the hypothetical: the cut is no longer the scenario the market is pricing. Three voters wanted a hike at that meeting, and the odds going in on a hike were roughly one in three. The arithmetic below is unchanged. The direction you should read it in is not.
So why not wait? Because the sensitivity runs both ways, and only one of those ways is currently on the table:
| Takeout rate | Constant | Proceeds @ 1.25× | vs. $3.5M payoff |
|---|---|---|---|
| 5.75% | 7.00% | $3,598,522 | +$98,522 |
| 6.00% | 7.19% | $3,502,624 | +$2,624 |
| 6.25% | 7.39% | $3,410,657 | −$89,343 |
| 6.50% | 7.58% | $3,322,427 | −$177,573 |
| 6.75% | 7.78% | $3,237,752 | −$262,248 |
Twenty-five basis points the favorable way is worth $91,967. Twenty-five the unfavorable way costs $88,229. Fifty the unfavorable way costs $172,904 — 1.88× what a single cut would gain you. A sponsor who waits is not making a cautious choice. They are taking a levered, roughly symmetric position on the next 25 basis points, with a payoff date attached, on three assumptions stacked in a row: that a cut comes at all, that it passes through in full to a 30-year-amortizing commercial coupon, and that nothing else moves in the meantime. After July 29, the first of those is the one the Committee itself is arguing about, and three of its voters are arguing the other side.
The second assumption is independently weak, and it is the one sponsors misread most often. Fed funds is an overnight rate. Agency and bank permanent takeout prices off the long end. A cut in the federal funds target does not mechanically move your takeout coupon, and the two have spent this cycle moving independently — July 29 is the clean illustration, with the target unchanged and the 10-year up 5 basis points on the announcement. Meanwhile the short end, where a policy move does pass through, runs the other leg of your position: 25 basis points on $3.5 million of SOFR-based carry is $729 a month, $8,750 a year. In the cut case that is a saving; in the hike case it is a cost, arriving on the loan you are already trying to retire. Either way it is real money, and either way it is not $89,343.
The gap being one cut wide is an argument for sizing it precisely and closing it deliberately. It is not an argument for hoping.
The Decision, Framed as a Decision
Four paths out of an $89,343 gap. None of them is generically correct — the right one is a function of how large your gap is relative to your basis, and how much time you still control.
Paydown to par-sizing. Write the check, take the smaller loan, keep the asset and the rate. Cleanest when the gap is small relative to equity already in the deal — and $89,343 against a $3.5 million balance is a 2.6% paydown, which is very much in that territory. This is the path that gets quietly overlooked because it feels like a loss; on this file it is the cheapest option on the board.
Extension. Buys time, and costs an extension fee plus, in most cases, a replacement rate cap. Cap renewal is a real line item and sponsors routinely underestimate it — pricing depends on your strike, the term, and the forward curve on the day you buy it, so get a live quote rather than carrying last cycle's number in your model. The honest test for an extension: are you buying time to execute something specific, or time to be wrong for longer?
Sale. If the gap is large relative to your basis and the business plan is already fully harvested, the exit that clears the debt is a legitimate answer rather than a failure. Better executed as a marketed sale on your calendar than as a workout on a servicer's.
Recapitalization. Gap equity or preferred to bridge the difference between supportable proceeds and payoff. Expensive capital, but correct when the asset has genuine remaining upside that the current sizing simply cannot access.
The variable that governs all four is time. Every one of them is available on better terms twelve months before maturity than sixty days after it, because a current borrower controls the timing and the framing, and a borrower in special servicing controls neither. Which lender you take this to also matters more than it used to — bank appetite in particular is now a function of where the institution sits on CRE concentration, which is public and worth checking before you spend a month in an underwriting queue. Our NJ Bank CRE Lending Tracker computes it quarterly for every New Jersey-chartered bank.
The Counter-Case
Four things cut against the argument above.
The maturity wall is genuinely shrinking. This is the strongest objection and it deserves the top slot. MBA's survey puts $875 billion (17% of $5.0 trillion outstanding) maturing in 2026 — down 9% from the $957 billion scheduled in 2025 — with $652 billion in 2027, lower again. The aggregate is improving, and MBA's read is that the market is moving past the peak. Nothing above argues otherwise. The argument is that the composition of what remains is harder, not that the total is worsening.
MBA and Trepp are in genuine tension, and I am not going to paper over it. Fratantoni's read of the same survey was that lenders "were no longer simply extending loan terms" — which points at fewer stalled loans, not more. Trepp's matured-balloon population is simultaneously at a multi-year high. Both can be true: MBA surveys all capital sources, Trepp measures CMBS only, and CMBS is the channel with the least workout flexibility and the most rigid servicing mechanics. If you want a single reconciliation, it is probably that balance-sheet lenders resolved and CMBS accumulated. But that is my inference, not either firm's finding, and a reader should weigh it accordingly.
A performing matured balloon is not a default. The 2.18% is a population, not a loss estimate. Many of these loans resolve at par — through a paydown, a sale, or a takeout that eventually sizes. Nothing in the Trepp figure implies principal impairment, and reading it as a distress forecast overstates it badly.
Agency-backed multifamily is largely insulated from this. Just $39 billion — 4% — of the multifamily and health care balances held or guaranteed by Fannie Mae, Freddie Mac, FHA, and Ginnie Mae matures in 2026, against 21% of depository balances, 25% of CMBS/CLO/ABS, and 29% of credit-company and warehouse paper. The pressure is concentrated in bank, CMBS, and debt-fund lending. That happens to be exactly the lane we work in, which is why this reads as urgent from where we sit — but it does mean the sector-wide framing has a real limit, and a sponsor with agency paper is looking at a different 2026 than a sponsor with a debt-fund bridge.
FAQ
How much do proceeds change when takeout rates move? Less than sponsors expect, because proceeds are sized by the mortgage constant and the constant moves less than the rate. On a 30-year amortization, moving the takeout coupon from 5.50% to 6.25% is 75 basis points of rate but only 58 basis points of constant — 6.81% to 7.39% — and cuts DSCR-constrained proceeds about 7.8%. On a $3.5 million loan sized at 1.25x against $315,000 of NOI, that is roughly $288,000 of proceeds.
Is the refinancing gap on a maturing bridge loan really one rate cut wide? On the traced example, yes — and that is the argument against waiting, not for it. The gap is $89,343 and a single 25 basis point improvement in the takeout coupon is worth $91,967. But the sensitivity is nearly symmetric: 25 basis points the wrong way costs $88,229, and a 50 basis point rise costs $172,904. Waiting turns a solvable six-figure funding problem into a binary bet on a move the Committee is currently arguing about in the opposite direction — three voters dissented for a hike on July 29 — and on a coupon that prices off the long end, not off fed funds.
Does DSCR or LTV set my takeout proceeds? It depends on the exit cap rate, and the binding constraint can flip without warning. On the traced example at a 6.25% takeout and 1.25x coverage, the DSCR test caps proceeds at $3,410,657. At a 5.75% exit cap, a 65% LTV test would allow $3,560,870 — so DSCR binds. At a 6.25% exit cap the value falls to $5,040,000, 65% LTV allows only $3,276,000, and LTV becomes the tighter constraint. Solve for both before you assume which one governs.
Is the commercial maturity wall getting better or worse in 2026? Both, depending on which number you read, and the honest answer is that the aggregate is improving while the composition is getting harder. MBA's survey puts 2026 maturities at $875 billion, down 9% from $957 billion in 2025, with $652 billion in 2027 — the wall is shrinking. Trepp's CMBS data shows the population of loans past maturity but current on interest at a multi-year high. Different universes measuring different things: MBA covers all capital sources, Trepp covers CMBS only.
Send Us the Payoff and the NOI
If you are carrying a loan that has matured or matures inside the next eighteen months, the number that matters is not a rate quote — it is your gap. Send the maturing balance and the current NOI and we will size the takeout against real coverage and leverage tests, tell you which one binds, and show you what the shortfall is and the ways to close it. If the answer is that you are fine, that is a useful thing to know too.
We placed a $5.55M Newark bridge for a sponsor taking a stabilized asset toward permanent financing; the gap-sizing exercise behind it is the same one described here. Sizing the gap is arithmetic. Closing it is knowing which of 100-plus bank, agency, and private lenders will actually quote your file, and on what structure.
Dominick Prevete — Founder, Blue Sky Capital Advisors. 31 years in real estate finance; 100+ bank and private-lender relationships across the bridge, agency, and private-credit spectrum. 4 Sutton Ct, Hamburg, NJ 07419 · (908) 220-6404. NMLS information available upon request. Lending in all 50 states.
Delinquency figures are from Trepp's June 2026 CMBS delinquency report. Maturity volumes are from MBA's 2025 Commercial Real Estate Survey of Loan Maturity Volumes, released February 9, 2026. The rate decision, vote, and dissents are from the FOMC statement of July 29, 2026. SOFR is as published by the Federal Reserve Bank of New York for July 28, 2026 and resets daily. Takeout coupons, cap rates, and coverage requirements are indicative as of July 30, 2026 and every deal sizes to its own underwriting on the day it is quoted. This is market commentary, not a commitment to lend or investment advice.
Loans are for business purposes only and are not subject to TILA, RESPA, or HOEPA. Not for primary residences. Equal Housing Opportunity. All loans subject to underwriting approval. Rates and terms shown for illustration; actual rates depend on deal specifics. We do not lend to borrowers with credit below 600 or on owner-occupied properties.