Self-storage CMBS risk hides in the coupon reset

Trepp's latest read on securitized self-storage describes an asset class that has quietly given up occupancy without, so far, giving up cash flow, which is why the number worth pricing sits below the headline. Median occupancy across the $14.08 billion of balance with readings at both securitization and the most recent observation has fallen from 90.5% to 87.0%, and behind that median 62.7% of the occupancy-measurable balance is backed by properties below their securitization occupancy, 35.3% by properties down more than five percentage points, while 37.3% of the sample reported flat or higher occupancy — a broad decline and a shallow one at once, made tolerable only by coverage.

Within the $7.84 billion that carries both debt service coverage and coupon data, the median DSCR is 1.87x and only 2.3% of balance sits below 1.0x, a book that refinances on its own numbers in most rate environments. Restate that same balance at an illustrative 6.50% refinancing coupon and the share below 1.0x climbs to 13.2%, a multiple of roughly six. Trepp calls the result a concentrated coverage-risk pocket hidden in current DSCR, and the note flags the limits of the exercise: it is a debt service sensitivity rather than a complete test of refinancing proceeds, meaning no valuation mark and no proceeds constraint enter the math. Inside those limits, 13.2% of $7.84 billion is a little over $1 billion of balance crossing under 1.0x, against roughly $180 million today.

The distinction between a sensitivity and a proceeds test matters, and it cuts both ways. A loan restated at 0.95x can still refinance if the property's value supports the balance, a question outside Trepp's exercise entirely; what a coverage miss does is start the conversation, and on paper that has amortized nothing the conversation is about the entire balance rather than a shortfall against it, which is why a 13.2% share deserves more attention than its size suggests.

The sharp cliff owes less to storage demand than to how the loans were built. Fixed-rate paper runs from 88% to 96% of balance across the occupancy buckets, the median loan has paid down nothing since securitization, and interest-only structures run from 63.0% of balance in the deepest-decline bucket to 82.5% in the shallowest, so a loan that has paid interest only and amortized nothing reaches maturity at close to its original balance and takeout coverage becomes a function of the new coupon and little else, with no de-levering in the interim to absorb a repricing. What has kept current coverage presentable is the in-place coupon, and Trepp notes that the properties with the largest occupancy declines carry the lowest coupons — the cushion and the refinancing exposure sit on the same loans. Loans securitized in 2025 and 2026 carried median fixed coupons of 6.20% and 6.01%, and the note frames those against a 3.47% median in a comparison that ends before naming the cohort. Read together, a 6.50% test is close to a non-event for the newest paper and a full repricing for the oldest, which suggests the coverage pocket is a legacy, low-coupon, interest-only cohort, not a sector-wide problem.

Cash flow has refused to cooperate with the bear case: median net cash flow is above securitization levels in all four occupancy-change buckets, including the group down more than ten percentage points, and across the NCF-measurable population the median is up 4.9%, with 64.1% of balance backed by properties generating more cash than at securitization. The note does not show how fewer occupied units produce more cash, but the likeliest explanation is that street rates and expense control have carried the revenue line, making physical occupancy a lagging indicator for this asset class. If that reading holds, an allocator screening self-storage CMBS on occupancy is screening on the least informative variable in the file.

The concentration should shape how this paper trades. A 13.2% share of a $7.84 billion sample is a little over $1 billion, and Trepp's description of that exposure as a pocket rather than a spread turns it into a set of individual loans, not an allocation decision. This is the ground on which the refinancing wall is being resolved, and as this publication has argued, that wall has become a rescue-capital market whose clearing basis is set inside the debt stack before it reaches the closing table. Storage fits the pattern closely: a median 1.87x DSCR across the bulk of the book means most of these loans refinance themselves, and the minority that cannot is small enough for structured extensions, preferred equity, and fresh mezzanine to absorb without a distressed sale. A sponsor arriving at a 2027 maturity with a legacy in-place coupon well under today's market, interest-only terms, and occupancy off five points or more is negotiating against a lender holding the same table Trepp just published.

The variable that sizes the pocket is the takeout coupon. Trepp runs a single point at 6.50%, so the implication runs one way: more of that $7.84 billion sample falls below 1.0x at a higher coupon, less at a lower one. At a billion dollars, the exposure is small enough to underwrite loan by loan and specific enough that it is probably mispriced in both directions somewhere in the stack — the trade is in the collateral tape, loan by loan.