The most important number in Trepp's latest office refinancing study is not the overall failure rate. It is the inversion. Approximately 75% of maturing Class A office loans in the sample would need fresh equity to refinance, against 61% of Class B and C loans. The market has spent three years treating "flight to quality" as the dividing line between survivors and casualties in office. The loan-level data says the trophy assets were levered hardest, on the most fragile terms. What separates them now is the capital stack, not the building.
That matters because the consensus framing of commercial real estate risk has quietly migrated from a credit story to a property story. Once regional bank failures faded and the Federal Reserve backstopped the system, the conversation narrowed to vacancy, remote work and the obsolete Class C tower. Class A was the safe end of the trade. The servicer data disagrees, and it is the servicer data, not the narrative, that determines who has to write a check by September 2027.
Interest-Only Debt Is the Real Fault Line
Trepp examined 309 office loans in commercial mortgage-backed securities, totaling $18.2 billion, all scheduled to mature by September 2027. Roughly 80% of the maturing Class A loans were interest-only, compared with about 55% of the Class B and C loans. Among interest-only loans, 84% faced a refinancing shortfall. Among amortizing loans, the figure fell to 40%.
That 44-point gap is the cleanest causal evidence in the study. Building quality and amortization are separable variables, and amortization explains more of the outcome. A Class A tower with full occupancy and a creditworthy anchor tenant can still fail a refinance if the borrower spent a decade paying only interest on a loan sized against a 2016 or 2017 cap rate. Nothing about the asset's physical quality reduces the principal balance. Only amortization or equity does that.
The reason Class A borrowers chose interest-only terms is not mysterious. Lenders competed for the safest collateral, and the concession they offered was structure: higher proceeds, no principal paydown, a maturity date that assumed a refinancing market would exist on similar terms. The quality premium on these assets was partly a leverage premium. Investors who bought office exposure on the assumption that quality equals safety bought the leverage along with it.
Why Extensions Cannot Close the Gap
The standard rebuttal to maturity-wall arguments is that lenders extend. Loans get modified, maturities roll, and losses are absorbed slowly. Trepp's income arithmetic shows why that defers the problem without resolving it.
To refinance just half of a maturing loan balance at par after a two-year extension, Trepp calculated that a property would need annual net operating income growth of approximately 18%. Median office NOI growth from 2020 to 2025 was 0.6% a year. The required growth rate is roughly thirty times the realized one. An extension buys time, but time only helps if income compounds toward the debt, and office income has been essentially flat for five years.
The debt-yield distribution tells the same story from the lender's side. Nearly $22 billion of the $78.1 billion private-label CMBS maturity pool coming due between October 2026 and the end of 2027 carries a debt yield of 8% or less. Debt yield is net operating income divided by loan balance, and it converts directly into coverage. At an 8% debt yield and an illustrative 6.5% interest-only coupon, debt service coverage is about 1.23 times. Put that same loan on a 30-year amortization schedule at 6.5% and the annual constant rises to roughly 7.6%, pushing coverage down to about 1.05 times. Those coupon assumptions are this article's illustration rather than Trepp's figures. The direction holds across any plausible rate: a loan at an 8% debt yield cannot be refinanced at the same balance on conventional amortizing terms in the current rate environment. Either the balance shrinks or the lender accepts a structure it would not write on a new deal.
This is where the rate channel does real causal work. The source reports that the 10-year Treasury yield recently moved above 5%. Commercial mortgage coupons price as a spread over that benchmark, so the long end sets the floor for every refinancing in the pool. Fed cuts at the front end do nothing for a borrower whose replacement loan prices off the 10-year. A cutting cycle that coexists with sticky long yields, driven by inflation risk and federal borrowing needs, is the worst combination for this cohort, because it removes the policy relief the market has been counting on.
Delinquencies Are Already Moving Ahead of the Wall
The maturity schedule is forward-looking. The delinquency data is not. As of September, CMBS loans at least 30 days delinquent totaled approximately $48.7 billion, or 8.02% of the $606.6 billion Trepp tracks. That balance has risen nearly 12% since the start of the year, before the heaviest stretch of maturities arrives. Of the $78.1 billion maturing through 2027, about $61.1 billion falls in the next calendar year, including $22.2 billion tied to office.
Rising delinquencies ahead of a maturity spike are a meaningful sequence. Loans normally go delinquent at maturity, when a balloon payment cannot be refinanced. Rising delinquency before maturity points to cash-flow failure, which means some borrowers are already unable to cover current debt service, not just the future refinance. That stress compounds. A property that misses payments today enters its refinancing with impaired leverage and no negotiating position.
Two Manhattan loans in the report show how wide the spread of outcomes has become within one market and one asset class. At 667 Madison Avenue, an approximately $254 million refinancing closed despite weakening debt-service coverage. At 60 Madison Avenue, a $100 million loan moved to special servicing after occupancy fell to 58% and operating income deteriorated sharply. The location did not decide these outcomes. Occupancy, income and the willingness of capital to show up did.
The Strongest Case Against the Thesis
The Class A inversion has a fair counterreading, and it deserves to be stated plainly. "Requires additional equity" is not the same as "will default." Class A sponsors tend to be the best-capitalized owners in the market: institutional funds, large REITs and sovereign capital with the liquidity to inject equity and keep the asset. A 75% equity-need rate among owners who can afford to pay may produce fewer losses than a 61% rate among Class B and C owners who cannot. On this reading, the inversion measures the size of the check, not the probability of a loss, and 667 Madison is the template for how premium buildings get through.
The sample has limits too. It covers 309 securitized office loans, a slice of the market with its own vintage and structure characteristics. It is not a census of bank-held commercial mortgages, insurance-company loans or private credit. Bank-held office loans may amortize more often, which would make the 40% amortizing failure rate the more relevant benchmark for regional lenders.
Both points refine the thesis rather than refute it. Sponsor equity is a transfer of loss from lender to owner, not the absence of loss. Fund vehicles that write those checks are recapitalizing assets whose equity value has already been impaired, and the returns that institutional investors underwrote on office as a stable income asset do not survive a forced capital call. The loss still exists. It moves to a different line on a different balance sheet. The more important open question is how many Class A sponsors will decide the equity check is not worth writing and hand the keys back. That decision depends on each sponsor's view of the building's terminal value, and nothing in the current data observes it directly.
What Would Confirm or Break the Read
The thesis is specific enough to test. If capital structure rather than property quality drives distress, Class A loans should begin appearing in special servicing transfers in larger numbers over the next 12 to 18 months. That would happen where sponsors decline cash-in refinancings, especially in assets with large interest-only balances and debt yields near or below 8%. A Class A share of new special-servicing transfers that rises toward the equity-need rates in this study would confirm the read. A Class A cohort that refinances cleanly through cash-in deals, with delinquencies concentrated in Class B and C, would show that sponsor liquidity was deep enough to absorb the inversion.
The rate condition is just as concrete. A sustained 10-year yield above 5% raises the coupon floor on the entire $61.1 billion due next year, while a retreat in the long end narrows the debt-yield gap without any help from NOI. The delinquency rate is the third gauge. It reached 8.02% in September and rose nearly 12% year to date; continued increases through the first half of 2027 would show that payment stress is outrunning the extension machinery.
Until one of those conditions turns, the cleaner reading of the evidence is that office distress has not ended. It has shifted toward the assets the market stopped worrying about. The buildings are fine. The loans written against them were sized for a rate regime that no longer exists, and someone has to fund the difference.
