Multifamily delinquencies are falling for the wrong reason
As bank delinquencies ease, rising charge-offs and lender-run sales describe a transfer of duration, not a cure.
According to CRED iQ's analysis of FDIC data from all insured institutions, bank multifamily delinquencies fell in the second quarter to 1.41% of loans from a multiyear high of 1.47%, and the dollar amount of delinquent loans slipped from $9.78 billion to $9.41 billion, on a book that grew 3.6% year over year to $667.6 billion. Expect those numbers to be quoted all month as evidence that apartment credit is mending.
But the same release cuts against that reading: loans 30 to 89 days past due improved while 90-plus delinquencies and net charge-offs both climbed, the short bucket curing even as the long bucket and realized losses worsened. CRED iQ calls that combination—easing delinquency alongside rising realized losses—"consistent with a workout-driven cycle rather than a resolving one." It reads as a rate improving partly because losses are being taken rather than because loans are being earned back, and anyone pricing bank apartment exposure off the headline is watching the wrong line.
CMBS carries the same ambiguity in softer form: Trepp's August numbers, as Multifamily Dive reported, had the multifamily CMBS servicing rate declining while delinquencies stayed flat, so both datasets show the top-line number moving the right way without evidence that severity has been worked through. A delinquency count tells you a loan is late; it does not tell you what the collateral fetches when the loan is finally sold.
CRED iQ puts the 1.41% at roughly 6.7 times the 2019 low of 0.21% and still well below the 5.90% peak of the Global Financial Crisis. Sitting that far above a floor sounds worse than it is: the rate is nowhere near crisis levels in absolute terms, and it is still improving, the reassuring half of the quarter doing most of the talking.
Run the two figures against each other and they reconcile—$9.41 billion is 1.41% of $667.6 billion—so the improvement is not a composition effect hiding inside the rate; the loss line is what the check exposes. The $370 million decline in delinquent balances landed in the same quarter that net charge-offs rose, suggesting part of the improvement was written off rather than worked out, and a bank book 3.6% larger than a year ago means new multifamily credit is still being originated while the old vintages are being cleared.
The 90-day bucket is the tell
Investors told Multifamily Dive that lenders are becoming more aggressive about resolving distressed properties, and the trade tape has started to show it. This week, Machine Investment Group and RPM Living Investments purchased 75 West, a 490-unit multifamily community in Dallas, through what a Sept. 14 press release described as an off-market, lender-driven process—the trade PRED covered on Sept. 14 as a lender-sourced reset. RPM Living Investments had already surfaced on the buy side in August, buying a 358-unit Arlington community with PCCP, which makes this a buyer group that has decided the basis is there.
What changed, according to Shimon Greenspan, CFO of Long Beach-based Westland Real Estate Group, is arithmetic rather than appetite. Lenders initially let borrowers "extend and pretend," he told Multifamily Dive, and the past 90 days have made clear that rates are not coming down fast enough to rescue owners who bet on them. "I think we're going to start to see a lot of that stuff work out in ways that it hadn't previously been forced to work out," he said. That is a shift in the lender's math: every added month of extension on a loan that cannot refinance is carry an institution absorbs for a borrower who cannot pay it.
The mechanics are straightforward: an extension moves a maturity date, not principal; a sale converts the loan into cash for the bank and a discount for the buyer, and the shortfall lands in the loss line instead of the delinquency bucket. That is how one quarter ends with a lower rate, fewer delinquent dollars and higher realized losses at the same time, and it is the strongest argument that this is a transition rather than a plateau.
PRED has argued that the refinancing wall is less a distress event than a duration transfer from banks to private credit, and the second-quarter data supplies the mechanism: extensions moved duration to borrowers through 2024 and 2025, while workouts move it to buyers, and the buyers taking it are small. Machine Investment Group runs 13 employees and $624 million in registered assets, and it is acquiring 490-unit properties through processes the lender controls.
That narrows the bid and stretches the timeline: when the price of a troubled apartment loan is set by whichever institution needs it off the balance sheet and whichever specialist answers, price discovery happens one asset at a time on the seller's clock, and the comps that come out of it are thinner than an auction's. The recovery in apartment credit will be marked by the buyers of Dallas deals before it shows up in a servicing report.
Two lines are worth tracking into the fall: the first is the pair that moved together in the second quarter, 90-plus delinquency and net charge-offs, against a book still 3.6% larger than a year ago; a rate can fall because loans cure or because the denominator outgrows the numerator, and the loss figures are the honest read on which is happening. The second is the buyer pool: Machine and RPM Living's lender-driven acquisition is the template, a 13-employee shop taking 490 units off a lender's hands without a marketed process. If more capital shows up at that size, the workout cycle clears quietly and the headline rate keeps falling. If it doesn't, the 1.41% holds its ground while the 90-plus bucket does the telling.
It reads as a rate improving partly because losses are being taken rather than because loans are being earned back, and anyone pricing bank apartment exposure off the headline is watching the wrong line.