The Extension Window Is Closing — $875 Billion in Commercial Real Estate Debt That Banks Can No Longer Defer

Two Numbers for One Asset Class

The delinquency rate on commercial real estate loans held at American commercial banks is 1.58 percent. The delinquency rate on office loans in the CMBS market, which prices the same underlying properties, was 12.34 percent as of January 2026, a record high per Trepp. Both numbers are current. One of them is wrong.

The gap between 1.58 percent and 12.34 percent is not a market pricing disagreement. It is an accounting structure.

Banks can modify a struggling loan. When a borrower cannot refinance at maturity, a bank can extend the deadline, adjust the terms, and continue to report the loan as performing. A commercial mortgage-backed securities trust cannot do this. Loans inside a CMBS structure that fail to pay off at maturity become delinquencies on the reporting date. There is no extension mechanism. The miss is immediate and public.

The result is two parallel records of the same deteriorating credit. One moves in real time. The other moves when banks choose to let it.

The End of the Extension Window

Banks chose extension for three years. The window is nearly shut.

CRE loan extensions peaked at $384 billion in 2024. In 2025, that figure fell to $200 billion, nearly a halving, as lenders demanded equity paydowns and higher spreads in exchange for any additional runway. A March 2026 Federal Reserve working paper found that undercapitalized banks disproportionately extended distressed loans and understated their default probabilities, a process the authors called evergreening. New York Fed Staff Report 1130 documented the same dynamic, finding that extension cycles create concentrated waves of future forced maturities when the deferral period ends.

That wave is the 2026 maturity wall.

The Mortgage Bankers Association estimated $875 billion in commercial mortgages would mature in 2026, more than double the 20-year annual average of roughly $350 billion. Embedded in that volume are loans extended once in 2024 and again in 2025. For those borrowers, there is no extension left.

For office specifically, the refinancing math is broken. Trepp's analysis found more than half of the approximately $100 billion in CMBS office loans maturing in 2026 are expected to fail to pay off at maturity. Among office loans that have already passed their maturity date and remain unresolved in CMBS trusts, default rates run above 80 percent in major markets. These buildings were underwritten at 3 to 4 percent financing costs. The current refinancing rate is 6 to 8 percent.

Who Gets Forced

Regional banks hold the concentrated exposure.

The SPDR S&P Regional Banking ETF fell 15 to 16 percent from its February 2026 peak. On March 2, the fund dropped 5 percent in a single session as geopolitical risk and the CRE maturity schedule converged in market pricing. KRE's equal-weight construction amplifies the exposure: smaller regional banks in the index carry higher CRE concentrations relative to risk-based capital than the large banks that dominate cap-weighted indices.

S&P Global projected that loan-loss provisions at regional banks could rise to 24 percent of net revenue in 2026, up from 20.8 percent in 2025. Provisions of that magnitude compress net interest income, reduce distributable earnings, and push earnings-per-share below consensus. The equity market reprices before the write-down formally appears.

Office REITs are the forward signal. SL Green Realty Corp reported a net loss of $84.4 million, or $1.20 per share, in the first quarter of 2026, compared to a net loss of $21.1 million in the same period in 2025. Vornado Realty Trust swung from net income of $86.8 million to a net loss of $22.8 million over the same stretch. Vornado's funds from operations fell from $126.2 million to $103.1 million in a single quarter. Wall Street downgraded Vornado in March 2026 as office REITs led the broader real estate sector lower.

The CMBS market has already repriced this. The regional bank equity market has not.

The Failure Path

The mechanism does not need a recession. It needs a calendar.

Here is the sequence: extended office loans hit revised maturity dates in the second and third quarters of 2026. Borrowers cannot refinance at current rates against collateral that CBRE's Q1 2026 report puts at a national office vacancy rate of 18.6 percent. Banks classify loans as non-performing. Allowances for loan losses increase.

As provisions rise, reported earnings fall. Banks that have been carrying modified and extended office loans as performing assets now recognize the credit deterioration that CMBS trustees have been recording for over a year. Regional bank earnings compress in Q3 and Q4 2026.

The CMBS channel reinforces the bank channel. As appraisals on office buildings decline, banks holding mortgages on those properties adjust their loan-to-value calculations. Assets that appear sound at 70 to 75 percent LTV on paper cross impairment thresholds when collateral values fall 20 to 30 percent. Losses deferred through extensions materialize as recognized provisions.

Flagstar Bank cut $4.7 billion in CRE exposure from its balance sheet in a single quarter. That is aggressive. Most regional banks have not moved at that pace. The ones that have not are now closer to forced action.

What I'm Watching

Three signals.

Regional bank provision ratios in Q2 and Q3 2026 earnings. S&P's projection of 24 percent provision-to-revenue puts a number on the threshold. Any bank reporting above that level has office exposure still on the books at values the CMBS market has not agreed with for months.

Trepp's monthly CMBS delinquency reports. The January 2026 record of 12.34 percent dipped to 11.4 percent in February. The volume of CMBS office loans maturing in the second half of 2026 will determine whether that dip was temporary or a genuine peak. A sustained move above 13 percent confirms the bank-held portfolio is further from fair value than any current filing reflects.

CRE extension volumes in quarterly bank regulatory filings. Extensions fell from $384 billion to $200 billion year-over-year. If Q2 disclosures show extension volumes declining further, the remaining runway for deferral is gone.

How This Thesis Fails

Three scenarios break this thesis.

The Federal Reserve cuts rates before the maturity wave peaks. A 150 to 200 basis point reduction brings financing costs back within refinancing range for some office borrowers. Default volumes fall short of projections. Bank provisions moderate below S&P's forecast.

Office demand recovers faster than vacancy data currently shows. CBRE's Q1 2026 report put overall vacancy at 18.6 percent, down from prior cycle peaks. If prime office absorption accelerates through late 2026, appraisal values stabilize, loan-to-value ratios improve, and fewer loans fail the refinancing test.

Banks front-load resolution rather than absorbing it over multiple quarters. Flagstar's approach of cutting $4.7 billion in a single quarter is an outlier. If more regional banks follow that path in 2026, the write-down cycle compresses into a short sharp event rather than a sustained earnings drag.

Closing Thoughts

Bank-held CRE delinquency at 1.58 percent and CMBS office delinquency at 12.34 percent cannot both be accurate descriptions of the same asset class for much longer. The extension mechanism that made the bank number possible is being withdrawn.

When it is gone, the 1.58 percent moves toward the 12.