The Commercial Real Estate Maturity Wall Is a Repricing Schedule, Not a Cliff

A maturity wall is one of the few risks in credit markets whose timing is known years in advance. A commercial mortgage written in 2021 on a five-year term matures in 2026 because that is what the note says. Nothing has to go wrong for the date to arrive, and that property separates this exposure from the shocks that usually dominate macro coverage.

The Mortgage Bankers Association reported on 9 February 2026 that $875 billion of commercial and multifamily mortgage balances, or 17 percent of the $5.0 trillion outstanding, is scheduled to mature during 2026 — roughly 9 percent below the $957 billion that came due in 2025. The headline is large. It is also smaller than last year's, and that direction rarely survives into the coverage.

The Current Position

The debt is not concentrated where most readers assume. Federal Reserve H.8 data for the week ending 5 August 2026 put commercial real estate loans at all commercial banks at roughly $3,124 billion, of which small domestically chartered banks held about $2,096 billion against roughly $908 billion at large ones in late June. The release defines "large" as the largest 25 domestically chartered commercial banks by domestic assets, so "small" is the regional and community banking system — carrying two-thirds of bank-held CRE credit on a far smaller share of total banking assets. Exposure is a ratio, and it runs opposite to headline bank size.

Collateral has stopped falling but has not recovered. The Federal Reserve's May 2026 Financial Stability Report noted that inflation-adjusted transaction-based price indexes for commercial property "have further stabilized following significant declines," while flagging refinancing needs as the live vulnerability. Stabilised is not restored: a property that repriced lower and then held flat still meets its maturity date at the lower number.

Aggregate stress is measurable and moderate. The FDIC's 2026 Risk Review put the industry's CRE past-due-and-nonaccrual ratio at 1.45 percent in the fourth quarter of 2025, with office vacancy at 14.0 percent at year-end, highest of the four major property types. Bank-held CRE loans still grew 3.1 percent in 2025 while construction and land development balances fell 5.8 percent. Those are the numbers of a slow adjustment.

The Transmission Mechanism, Test by Test

A maturing commercial mortgage does not come due the way a consumer loan does. The borrower must repay, refinance, sell or negotiate; four tests decide which.

One: the amortization gap

Commercial mortgages are commonly written with a term shorter than the amortization schedule, leaving a balloon balance. The 2023 interagency Policy Statement on Prudent Commercial Real Estate Loan Accommodations and Workouts, published in the Federal Register on 6 July 2023, works through examples built exactly that way: an office loan with "payments based on an amortization of 20 years," a "balloon payment of $13.6 million at the end of year five," and "75 percent loan-to-value (LTV) based on an appraisal reflecting a $20 million market value" with a "debt service coverage (DSC) ratio of 1.30x." The statement notes that "the lender expected to renew the loan when the balloon payment became due."

That expectation is the mechanism: the loan was never designed to be repaid at maturity, only replaced.

Two: coverage at the new coupon

Debt service coverage is net operating income divided by annual debt service. Take a property producing $1.2 million against a $15 million loan on a 25-year amortization schedule. At a 4.25 percent coupon, annual debt service is about $975,000 and coverage is 1.23x. Hold income constant and reset the coupon to 6.75 percent: debt service rises to roughly $1.24 million and coverage falls to about 0.96x. The building has not changed and the tenants have not left; the loan simply no longer covers itself.

Restoring 1.25x coverage at that coupon requires annual debt service of $960,000, implying a loan of roughly $11.6 million. The gap — $3.4 million, or 23 percent of the original balance — is the equity cheque required to refinance, and that cheque, not the maturity date, is the event.

Three: value, and the cap rate that sets it

Income-approach valuation divides net operating income by a capitalization rate, so the discount rate does the work. The same $1.2 million is worth about $20.9 million at a 5.75 percent cap rate and $16.0 million at 7.5 percent. Against the $15 million loan, leverage moves from 72 to 94 percent on repricing of income alone.

The reference points are public. The Interagency Guidelines for Real Estate Lending Policies set loan-to-value limits of 85 percent for improved property, 80 percent for commercial and multifamily construction, 75 percent for land development and 65 percent for raw land, with exceptions permitted in aggregate up to "100 percent of total capital." A 94 percent LTV loan is not prohibited; it is capital-expensive, which is what constrains renewal.

Four: the lender's own balance sheet

The last test has nothing to do with the property. Under the 2006 interagency guidance on CRE concentrations, supervisors apply further scrutiny where construction, land development and other land loans reach 100 percent or more of total risk-based capital, or where total CRE loans reach 300 percent or more of total risk-based capital and the portfolio has grown 50 percent or more over the prior 36 months.

The FDIC's 2026 Risk Review reported a median CRE concentration ratio of 200 percent for the industry, but 311 percent for banks with $1 billion to $10 billion in assets and 289 percent for the $10 billion to $100 billion cohort. A lender above the second criterion weighs a marginal renewal differently from one at 120 percent, however sound the loan. Identical properties in one submarket can receive opposite answers.

Extension Is Sanctioned, and Contested

A maturity wall converts into a schedule because extension is a sanctioned supervisory outcome. The 2023 policy statement distinguishes an accommodation — a short-term modification before a loan reaches workout status — from a workout, meaning renewals, extended terms, additional credit or restructuring. Institutions implementing prudent arrangements "will not be subject to criticism," and modified loans to borrowers "who have the ability to repay their debts according to reasonable terms will not be subject to adverse classification solely because the value of the underlying collateral has declined" below the loan balance.

That second clause is the most important line in the debate: a fall in appraised value, by itself, does not force loss recognition. Cash flow does. What the latitude produces is disputed inside the Federal Reserve System, and both sides of that dispute are set out below.

The fragility case. Crosignani and Prazad, in New York Fed Staff Report No. 1130 (October 2024, revised June 2026), argue that weakly capitalized banks granted extensions and payment relief to distressed borrowers to preserve capital, producing credit misallocation and reduced new lending. Extensions, on their account, "increased the stock of CRE mortgages maturing in the near term, raising the risk of large losses materializing over a short period." As of the fourth quarter of 2024 they report CRE loans expiring within three years equal to 37 percent of marked-to-market capital at less-capitalized banks against 27 percent at better-capitalized banks.

The stability case. David Glancy, in Federal Reserve Board FEDS working paper 2026-025 (4 May 2026), reaches the opposite conclusion from supervisory loan-level data. Extension rates ran in the mid-to-low 40s before the pandemic, peaked just under 60 percent in the year after its onset, and returned to roughly 50 percent through 2023 and 2024. Terms tightened too: non-recourse loans were 4 percentage points less likely to be extended during the stress period, extensions were 5.5 percentage points more likely to require paydowns above 5 percent, and among low-debt-yield office loans the share requiring material paydowns rose to 30–40 percent from 15–20 percent. Extensions, he concludes, "predominantly address temporary payment frictions."

Both can hold at once: practice at well-capitalized lenders can be disciplined while practice at thinly capitalized lenders is not. Dispersion, not the average, is where the risk sits — which is why national aggregates predict poorly which institutions get into trouble.

The Property Types Do Not Move Together

Treating commercial real estate as one asset is the most common error in coverage of this topic. Trepp's June 2026 CMBS delinquency readings put the overall rate at 7.35 percent, components spanning an order of magnitude.

Property type	CMBS delinquency, June 2026
Office	11.57%
Multifamily	7.23%
Retail	6.91%
Lodging	5.22%
Industrial	1.20%

Office and industrial differ by roughly 10.4 percentage points, close to a tenfold ratio, and the reason is structural. Logistics assets reprice leases on short cycles against demand largely independent of office attendance; office assets carry long leases that roll slowly, so a 2021 demand change is still arriving in 2026 income statements. Retail, written off a decade ago, now shows the lowest vacancy of the four major types in the FDIC's Risk Review, years of minimal new construction having removed the supply overhang. A national maturity total mixes segments with almost nothing in common.

The Historical Parallel: 1989 to 1995

The FDIC's History of the Eighties records that during the 1980s, while total bank real estate loans more than tripled, commercial real estate loans nearly quadrupled. Tax treatment did much of that work: breaks in the Economic Recovery Tax Act of 1981 "greatly enhanced the after-tax returns on real estate investment," and the downturn "was aggravated by the Tax Reform Act of 1986, which removed tax breaks for real estate investment." Buildings financed on one set of assumptions matured into another.

Resolution followed years later. FIRREA was enacted on 9 August 1989, creating the Resolution Trust Corporation, which operated until December 1995 and resolved 747 thrifts holding roughly $402.6 billion in assets. Between 1980 and 1994 the FDIC separately handled 1,617 failed or failing banks with $302.6 billion — together some 2,912 institutions and roughly $924 billion, close to one failure every other day for fifteen years.

Two features carry forward. It took about six years from enabling legislation to wind-down, so losses did not arrive in one quarter. And disposition method mattered as much as the losses: the RTC's bulk-sale programme drew criticism for discounting assets heavily enough to depress the very local markets in which remaining collateral had to be valued. Speed of recognition and speed of disposal are separate choices, and the second can do more damage.

What differs matters equally. The supervisory LTV framework, the 2006 concentration criteria and the 2023 accommodations statement did not exist in current form in the 1980s. The comparison is a template for the shape of the adjustment, not a forecast of its depth.

Conditions That Would Turn the Schedule Into an Event

A slow-motion reading is a base case, not a certainty. Four conditions would invalidate it, each observable in advance.

• Funding stress at the lender, not the borrower. Extension requires a lender with capacity to wait. A deposit outflow or funding-cost shock at a bank above the 300 percent criterion converts a patient renewal into a forced one, and the liability side then governs timing.
• Correlated maturity within one submarket. If a disproportionate share of a metropolitan office submarket's loans came from the same two vintages, they mature together, and simultaneous distressed sales set comparables for every remaining appraisal there. That is a local cliff inside a national schedule.
• A second leg down in net operating income. Every calculation above holds income constant, and the FDIC's 2026 Risk Review already notes decelerating rent and income growth. If income falls rather than flattens, coverage and value fail together and no coupon path rescues the loan.
• Recognition rules changing faster than the assets. The 2023 policy statement is guidance, not statute. A supervisory shift toward faster classification would compress the adjustment regardless of fundamentals — low probability, but the path that most directly turns a schedule into an event.

The frame could equally prove too pessimistic. If prices continue the stabilisation the Federal Reserve described in May 2026 and financing costs fall enough to close the coverage gap, a meaningful share of 2026 and 2027 maturities refinances conventionally with no extension at all.

What to Watch Next Week

• The weekly H.8 release. Watch the CRE line for small domestically chartered banks. A flat or falling balance alongside growth at large banks would signal smaller lenders shrinking exposure rather than renewing.
• Monthly CMBS delinquency prints. The office-to-industrial spread matters more than the headline. Narrowing driven by industrial deterioration would be broad-based; widening driven by office alone stays a sector story.
• Bank disclosures on modifications. Track extensions granted with principal paydowns against those without; Glancy's finding turns on that distinction.
• Appraisal-driven provisions. Provisions moving on reappraisal rather than missed payments signal marking to a market, not a payment history.
• Any revision to interagency guidance. Changes to accommodation or concentration criteria act on the whole loan stock at once.

Concrete Framework

1. Establish the denominator first. The 2026 figure is 17 percent of $5.0 trillion, and smaller than 2025's. Ask both of every such headline.
2. Split by property type. A spread of 1.20 to 11.57 percent means the aggregate describes no actual property. If a source does not disaggregate, treat the number as unreadable.
3. Locate the holder. Banks, CMBS trusts, life insurers and agency lenders face different capital rules and renewal incentives. Roughly two-thirds of bank-held CRE sits outside the largest 25 domestically chartered banks.
4. Recompute the two ratios. Coverage at the current coupon, LTV at a current cap rate. Both are estimable from public disclosure and both routinely differ from the originated figures.
5. Ask what the lender's capital permits. The 100 and 300 percent criteria are public and concentration ratios reportable. A lender below those thresholds has options a constrained one does not, for identical collateral.
6. Separate recognition timing from economic loss. Extension defers the accounting event; it neither creates nor destroys value. The question is whether income covers debt service, at what coupon, for how long.
7. Name the break condition before the data moves. Write down which break condition would change the reading, then monitor that rather than the headline.

Stated narrowly: the 2026 total is knowable, smaller than 2025's, and concentrated in a lender segment whose capacity to extend is variable and imperfectly observable. Whether that produces an orderly repricing or a compressed episode depends on funding conditions at the lenders, not on the maturity schedule — and that variable is not yet determined.

This article is analysis of publicly available economic data and is not investment or financial advice.