Where is the CRE distress, really?
This is the public conduit CMBS tape — the SEC-registered, multi-borrower deals that disclose every loan, every month, on EDGAR. We read all of it, weight by balance, and split it by what each loan is actually secured by. Distress means a loan 60+ days delinquent or in special servicing, located at the property. You will see higher CRE distress figures elsewhere; further down we show exactly what they include that this does not, and why we publish only the part we can show our work on.
Office is the story everyone is telling, and it earns it: conduit office distress has gone from under 2% to 11.2%, and it is still rising. That is a genuine, slow-motion downturn, and it deserves the attention.
But while the attention was on office, something quieter happened next door. Multifamily — the asset class everyone files under "defensive" — re-rated from roughly 2% to about 8% over a single year, and then it stopped climbing and simply held there. It is not creeping up; it already moved. It is sitting at a level nobody associates with apartments, and almost no one is looking, because the conduit blend (6.2%) averages it back down toward calm.
This is not two buildings. The multifamily distress is spread across roughly a hundred and fifty separate properties in Texas, California, New York and Florida — remove the five largest and the rate barely moves. It is a broad re-rating of the asset class, hiding in plain sight inside a headline that stays low precisely because office and multifamily are averaged together with retail, industrial and storage, which are fine.
The same quarter, by what the loan is secured by
One headline rate, seven very different books. This is 2026Q2, each type as a share of the CMBS balance and its own distress rate — the distribution the blend flattens into a single line.
2026Q3 is held back — only ~18.0% of a normal quarter has filed so far, too little to read as the state of the market.
Why our number is lower than the ones you've seen
Headline CRE distress figures elsewhere run higher — often above 11%. The difference is not a disagreement about the loans we both see. It is a whole universe we deliberately leave out: single-borrower (SASB) deals — one giant loan on one trophy tower, mall or resort, packaged privately under Rule 144A. Those deals file no public loan tape. Their monthly performance goes to bondholders, not to EDGAR, and the vendors who do have it license it under terms that forbid showing you the loans. We can't verify those numbers, so we don't republish them.
Here is the tell that the conduit number is right, not merely low. Take a type with almost no private-deal presence — self-storage. It reads 0.1% here, essentially identical to the market-wide figures that include the private deals. Where the two universes overlap, we match. We sit below only in the types where single-borrower deals concentrate — office, lodging, bridge multifamily. The gap is not measurement error; it is precisely the private slice, and we would rather name it than guess at it.
The shape survives the exclusion: office is still the worst, multifamily still re-rated hardest. Folding the private deals back in would lift the levels, not change the story — if anything it sharpens it, since the most distressed single-borrower office is exactly what's missing here.
What this does not do
It is not a forecast. It states where the distribution sits right now and shows the series behind the statement. Nothing here says what any type does next quarter — only where each one already is.
It is not a per-loan verdict. This is an aggregate read by property type. It cannot tell you anything about a particular building or borrower, and a high type-level rate is entirely compatible with individual loans in that type that are perfectly fine.
It is conduit CMBS, not banks or private deals. This is the securitized, publicly-registered tape — loans in conduit CMBS trusts, located at the property. It is a different lens from the bank-call-report cycle clock, and it excludes the private single-borrower universe above. Each lens can disagree with the others, and all can be true at once.
Where this comes from
Every conduit CMBS trust files monthly asset-level servicer reports with the SEC — the Reg AB II ABS-EE filing — naming, loan by loan, the property, its type, its balance, how many days past due it is, and whether it has gone to special servicing. We read all of it off EDGAR, take the latest filing for each loan in each quarter, weight by balance, and group by what the loan is actually secured by.
Nothing here is licensed, scraped, or proprietary. The filings are statutorily public and you could rebuild this yourself — the work is in making thousands of loans across dozens of servicers comparable, and in refusing to let the average speak for the distribution. That is also why we stop at the public tape: the number we publish is one you can check.