Thin DSCR cushions explain the 2023-vintage multifamily CMBS delinquency spike

At 32 months of seasoning, the 2023 conduit multifamily vintage's private-label delinquency rate is 29.01%. Trepp's research is careful to flag the context alongside the number: the vintage carried only about $2.8 billion in private-label multifamily collateral at securitization. Small pools produce dramatic percentages. This one is a warning flag, not yet a loss event.

The bigger collateral sits behind it. Trepp's July 2026 snapshot puts the 2023 vintage at 23.8% conduit multifamily delinquency, and the 2024 vintage at 10.60%. Delinquency counts loans at least 30 days behind. The 2023 vintage climbed from 11.13% after 18 months to 13.71% after 24, then to 29.01% after 32. Trepp calls that headline rate a clear underwriting and surveillance signal while noting dollar exposure remains comparatively contained. The 2024 vintage holds $20.1 billion of private-label multifamily collateral, about seven times the 2023 base. The 2025 vintage holds $16.6 billion and carries the largest multifamily share at 23%.

The collateral mix in new conduit deals intensifies the signal. Multifamily represented about 7% of conduit securitization balance in 2017. Trepp puts it between 20% and 23% from 2024 through July 2026, with the 2025 vintage at the high end. Office, which filled about a third of conduit balance in most vintages from 2017 through 2022, has fallen to 15% of the 2024 and 2025 vintages. A multifamily underwriting error used to touch a modest slice of a pool. It now touches the largest slice.

The operating data point to underwriting rather than to collapsing properties. The 2022 vintage recorded larger declines in DSCR and occupancy than the later vintages did, yet its current DSCR remains 2.17x. The 2023 vintage is at 1.36x and the 2024 vintage at 1.28x. The difference is the starting point: 2022 was underwritten at a 2.51x debt service coverage ratio, 2023 at 1.56x, 2024 at 1.40x.

A DSCR near 2.5x gives a property room to absorb vacancy, a rent rollback, or a rate shock. A DSCR near 1.4x does not. The implied erosion since issuance is modest: about 0.20x for the 2023 vintage and 0.12x for the 2024 vintage. Modest erosion from a thin base is the problem. At a current DSCR of 1.28x, the next soft quarter can push coverage toward 1.0x. The distance from 1.28x to 1.0x is 0.28x. A loan underwritten with that little room is priced for perfect operations, not for a vacancy dip.

Trepp's mix argument deserves emphasis. The shift does not make new conduit pools inherently riskier, the firm argues; it moves the risk. A multifamily-specific underwriting error now lands across 20% to 23% of pool balance rather than the 7% to 15% range of older vintages. The newest pools are also the least seasoned. The old hedge, that a multifamily mistake would hit only a small slice of a diversified pool, no longer holds. For the first-loss buyers in a 2024-vintage conduit deal, that combination is effectively a concentrated bet on multifamily underwriting discipline.

The 2024 vintage will put that bet to the test. Its starting DSCR was thinner than the 2023 vintage's, and its collateral base is seven times larger. A repeat of the 2023 curve would drive the delinquency rate from 10.60% through the teens toward 30% on a base seven times larger, and the loss severity would not stay contained. The 10.60% reading is early, and the next two quarters of delinquency data will show whether it is the start of the same curve. At a 1.28x cushion, there is no room for a bad quarter, and the pool is too large to treat as noise.