The 2025 Commercial MortgageBacked Securities (CMBS) reappraisal cohort cleared $23 billion of collateral at a median 53% discount to origination.
The distress is still heavily concentrated in the property types the market has worried about most: office first, retail second, and mixed-use close behind. These three property sectors accounted for roughly 86% of the reappraised collateral, with office alone accounting for over half.
What matters here is not just that values are down, but where forced re-underwriting is occurring. Reappraisals tend to happen only after credit stress has become hard to ignore. That makes this dataset especially useful for identifying where deterioration has moved beyond soft weakening and into the part of the cycle where losses, appraisal reductions, and resolution pressure become more tangible.
Before diving into the analysis, two characteristics of this dataset matter before the numbers do. First, every loan in the
sample is a single-property loan. Portfolio loans have been excluded, which means each valuation change traces
unambiguously to one asset. Second, reappraisals are a filtered sample, not a representative one.
A servicer only orders a new appraisal when something has gone wrong. The valuations in this cohort reflect loans already in distress rather than the broader CMBS population. That filter is the analytical starting point, and it shapes every conclusion that follows.
A new appraisal is ordered only when one of several distress events crystallizes the servicer's need for a current mark. The
standard trigger set includes a 90-day payment default, a voluntary borrower bankruptcy, an involuntary bankruptcy that is not dismissed within 60 days, appointment of a receiver on the property or loan, transition to Real Estate Owned (REO), or a maturity event where the loan cannot refinance cleanly, and a reappraisal is required to frame the workout. This trigger
structure has an analytical implication that is easy to overlook.
Reappraisal cohorts are not random samples. They are, by construction, the set of loans where distress has advanced far
enough to require a repricing of collateral. The -53% median across 2025 is not a statement about the CMBS market. It is a
statement about the subset of CMBS that has already broken. What makes the cohort informative is that once a loan hits one of these triggers, the valuation decline is no longer hypothetical. It becomes the basis against which losses are calculated and the starting point from which any subsequent resolution, whether modification, note sale, Debtor-in-Possession, or REO disposition, is negotiated. Reading the reappraisal data tape is, in effect, reading the market's floor price on distressed collateral.
For the 2025 cohort, maturity events are a rising share of triggers relative to earlier years. The post-COVID wave of lodging and retail delinquencies drove much of the 2021–2023 reappraisal activity, and those loans reached the tape through term default and special servicing transfer. The 2025 book is increasingly populated by loans that reached scheduled maturity, could not refinance at current rates against current values, and were forced into a mark when the extension conversation gave way to a workout conversation. The shift in trigger composition matters for how the cohort should be interpreted, as 2025 is less a delinquency story and more a refinance-failure story, and the valuation declines reflect that evolution.
Across all property types, 2025 reappraisals covered $23.4 billion in loan balance over 495 single-property loans. Office dominates at $12.2 billion, which is more than half the total balance and 42% of the property count. Retail and mixed-use together add another $8.0 billion across 169 properties. Lodging ($1.7 billion across 61 properties) and multifamily ($1.2 billion across 39 properties) sit in the middle of the distribution. Industrial ($221 million), "other" ($59 million), and mobile home ($17 million) round out the tail.
The median value change tells a cleaner story than the balance weights. Office, retail, and mixed-use clusters tightly in the -53% to -56% range, which at first glance reads as a uniform problem across the three property types most exposed to post-COVID demand rotation. It is not. Lodging comes in at -42% and multifamily at -34%, which are meaningfully better, consistent with the operational recovery in those sectors and the narrower valuation adjustment from rate-driven cap rate expansion rather than a structural demand reset.
Industrial prints -23%, which at this stage of the cycle qualifies as resilience: the assets being reappraised in industrial are typically loans that hit a specific refinance or sponsor issue rather than a collateral-value problem.
The distribution is bifurcated. Four property types, including office, retail, mixed-use, and the "other" bucket, show median declines deep enough to restructure recovery economics.
Three property types, including industrial, multifamily, and lodging, sit in a range where modification and extension remain plausible resolution paths because the collateral can still, at the new mark, support a plausible refinance at reasonable leverage. That distinction is the first-order conclusion from the aggregate cut, and it frames the sections that follow.
The more informative cut is the comparison between median and weighted-average declines within each property type. Where weighted averages pull significantly below medians, as they do in office, concentration is doing the work. A small number of very large loans with very deep write-downs are driving the aggregate, which signals that the "office problem" is not a broad office problem but a concentration problem within a specific office subtype. The subtype cross-section in the next two sections confirms this directly.
The vintage cross-section is where the cycle gets interesting. Office loans securitized in 2017 and 2018 reappraised in 2025 at median declines of -71% and -67%, respectively. Older vintages, such as the 2014 and 2015 cohorts that have been working through office distress since before COVID, printed -50% and -51%. The 2012 and 2013 vintages, deeper still into their workout cycles, came in at -53% and -50%. The 2019 cohort, a smaller sample, marked at -50%.
At first glance, this inverts what a practitioner would expect. Older vintages should, in theory, be the troubled book, given that they include the post-financial-crisis legacy loans that have been grinding through distress the longest. Instead, the 2017–2018 cohort is printing worse marks. The explanation is mechanical once the details are laid out.
The more informative cut is the comparison between median and weighted-average declines within each property type. Where weighted averages pull significantly below medians, as they do in office, concentration is doing the work. A small number of very large loans with very deep write-downs are driving the aggregate, which signals that the "office problem" is not a broad office problem but a concentration problem within a specific office subtype. The subtype cross-section in the next two sections confirms this directly.
Loans securitized in 2014 and 2015 have had, by 2025, roughly a decade to build an amortization cushion, accumulate net operating income history across multiple economic cycles, and, in many cases, absorb prior equity contributions from sponsors attempting to defend the collateral. By the time they hit a reappraisal trigger, the economic loss has been
partially worn down by interim performance and prior equity support. The 2017–2018 cohort has had none of that. These are loans that were underwritten at peak-cycle office valuations, as 2017 and 2018 sat at or near the top of the pre-COVID office capital markets, and then hit the reappraisal window without the softening influence of years of debt service or operational performance. The valuation shock lands cold on the 2017–2018 book, and the tape shows it. This is the cohort where the extend-and-pretend strategy has the shortest runway, because there is neither accumulated amortization nor accumulated sponsor equity to draw against.
Retail tells a broader, more distributed story. Every vintage from 2011 through 2023 shows material
declines, ranging from -3% in 2023 (a small sample still early in its workout) to -78% in the 2011 cohort.
The shading on the 2014–2015 retail vintages indicates material property counts, and the -55% to
-65% prints on those years reflect the slow grind of legacy mall and power center distress that has
defined the retail sector's credit performance since before the pandemic. Retail's vintage pattern is
closer to what intuition would suggest: longer seasoning, deeper marks, broader distribution across
the calendar.
Multifamily and industrial show minimal vintage dispersion, consistent with their status as the sectors where the 2025 reset has been shallow and selective rather than broad. The 2022 and 2023 multifamily cohorts, where floating-rate financing from a peak valuation environment is meeting materially higher debt service, are beginning to print meaningful declines (-44% and -40%), but the number of properties reaching reappraisal remains contained. Industrial's scattered marks across 2016–2024 vintages reflect property-specific workouts rather than any sector-wide repricing.
The practitioner takeaway on vintage is concrete. The 2017–2018 office paper is the cohort that will define workout economics for the next 18 to 24 months. It is the largest concentrated block of severely-impaired collateral that has not yet worked through resolution, its marks are measurably worse than the legacy book's, and the structural features that distinguish it, such as peak underwriting, limited seasoning, and minimal amortization cushion, will continue to produce deep reappraisal declines as more of that book reaches the trigger set.
Pulling the cross-section down to subtype resolves the ambiguity that the property-type aggregate leaves open. The headline number is urban office: $9.0 billion in balance across 97 single-property loans, reappraised at a median 64%. That is the largest concentrated block in the 2025 cohort by a significant margin. It represents approximately 74% of the office balance and carries a valuation decline twelve percentage points deeper than the office property-type aggregate. The "office problem" is, with precision, an urban office problem.
Suburban office, by contrast, marked at -52% across $2.5 billion and 94 properties. That is a severe outcome by any normal standard, but against the urban office, it reads as a different regime. The gap between -64% and -52% is the gap between assets being revalued as structurally obsolete and assets being revalued as cyclically impaired. Medical office at -29% across $207 million sits in a different sector altogether, as it is operating real estate tied to healthcare demand fundamentals rather than office demand fundamentals, and the tape reflects that.
The subtype evidence argues strongly against treating "office" as a unified analytical category. The aggregate median, the weighted-average spread, and the vintage dispersion all point to the same conclusion: office credit performance is now a bifurcated story with an urban central business district (CBD) problem on one side and a more manageable suburban and specialty story on the other.
Retail shows the same within-type dispersion. Regional mall and superregional mall print -67 and -66% respectively, which is deeper than urban office on a percentage basis, though at a smaller aggregate balance of $3.6 billion combined across 30 properties. Urban/street retail marked -60% across $315 million. Community shopping centers and outlet centers clustered around -46% to -47%.
Neighborhood centers came in at -39%, and garden/low-rise multifamily at -27%. As with office, the retail story is not uniform. It is a concentration story in enclosed mall formats and urban high-street retail, with grocery-anchored and community retail absorbing a visible but structurally manageable hit.
Four subtypes, including urban office, regional mall, superregional mall, and urban/street retail, together account for roughly $12.9 billion of the 2025 balance and carry median declines between -60% and -67%. This is the analytical core of the 2025 reset.
These are the assets where collateral values are being repriced as structurally rather than cyclically impaired, and where recovery analysis needs to proceed from the new basis rather than the origination basis. For investors positioning against this cohort, treating these four subtypes as a distinct analytical group, separate from suburban office, community retail, and the broader commercial real estate (CRE) book, is the most useful framing the data supports.
The implication of a 60%+ valuation decline is not merely accounting. It is a change in the set of viable
resolution paths. When appraised value falls by half or more from origination, the unpaid principal balance
necessarily exceeds the new collateral value, often by a wide margin. The debt yield on the original loan
amount is no longer the relevant refinance metric.
For example, a property that was financed at a 9% debt yield at origination and has since lost 60% of its value
would need to support a debt yield north of 22% on the unpaid balance to clear the market at similar leverage,
which simply doesn’t make sense. Urban office assets, particularly Class B stock in CBDs, are still working
through the demand impact of hybrid work, are not producing those debt yields, and the refinance market
is not making exceptions.
This has three practical consequences for workout economics. First, the modification becomes less
productive. A borrower offered a rate reduction or term extension on a loan that sits 60%+ above the current collateral value has no economic incentive to commit additional equity into the deal.
The modification preserves the servicer's book position and defers loss recognition, but it does not restore economic
viability. The loan simply returns to distress at the next stress point, whether that is the next maturity, the next
tenant rollover, or the next rate reset. This is the mechanism behind maturity recidivism, as loans extended
through one maturity and re-defaulting at the next, and the 2025 reappraisal cohort is the set of loans where recidivism
becomes mathematically unavoidable rather than merely likely.
Second, the resolution mix shifts toward note sales, discounted payoffs, and REO dispositions. Each of these
resolution paths prices off the new appraised value, not the original loan amount, and accepts the economic loss in the
current period rather than deferring it.
For special servicers, this means the 2025 cohort will contribute to realized losses on a faster timeline than the slow-rolling legacy book. For bondholders, it accelerates loss recognition but reduces the tail of cost overruns interest advances, and protective advances that accompany prolonged workouts. The trade-off is cleaner than it has been in
earlier phases of the cycle, since you can now take the loss at a known amount or carry it as an uncertain workout liability for another two to four years.
Third, bidder economics on distressed paper becomes more legible. At the 2025 mark, buyers underwriting to the new
appraised basis, rather than the stale origination number, have a clean starting point. The information asymmetry between the special servicer and the secondary bidder narrows when a fresh appraisal is in place. This is part of the reason distressed CRE debt trading activity has picked up materially in 2025 and is likely to continue through 2026 as the reappraisal pipeline keeps delivering marked assets into the bid set.
The overall shift the 2025 cohort signals is from extend-and-pretend to resolve-and-realize. Extension works when the
collateral can plausibly grow back into the loan over a workable time horizon. At -64%, that thesis does not survive
underwriting. At the subtype level, including urban office, malls, and urban street retail, it has become analytically
unavoidable that resolution economics need to start from the new mark.
The 2025 reappraisal cohort is not the apex of the reset cycle. It is a working level, and the 2026 pipeline suggests the cohort will look similar in size or larger next year. Three forces are driving continued flow.
The 2026 maturity calendar remains the dominant source. A significant block of 2016-era ten-year loans and 2021-era five-year loans reaches scheduled maturity in 2026, and the refinance math on the 2016 office book in particular is not materially different from what the 2017–2018 cohort is experiencing today, where peak-cycle underwriting meets post-COVID demand destruction, with limited amortization and limited sponsor equity support. Each of those loans that cannot refinance cleanly generates either a term extension conversation or a maturity default, and a meaningful share of those defaults will produce reappraisal triggers within the next twelve months.
Term defaults remain elevated in office, independent of the maturity calendar. Each term default that advances to 90 days delinquent generates a new reappraisal trigger, and the backlog of loans sitting in special servicing without a current appraisal is still being worked through. Some of the 2025 marks are being taken on loans that transferred to special servicing in 2023 and 2024 and are only now reaching the appraisal stage of the workout. That backlog alone suggests continued reappraisal flow through 2026 even if no new defaults occur.
Finally, the 2017–2018 office vintage has not finished working through resolution. The vintage data argues that the worst marks in this cycle will come from that cohort, and only a fraction of it has reached the reappraisal tape so far. As additional 2017–2018 office paper transitions from modification to resolution over the next eighteen months, the
concentration of deep write-downs in that vintage band is likely to extend rather than diminish.
Investors pricing distressed CRE off origination values are pricing off the wrong number. The 2025 marks are the relevant basis, and the 2026 marks will extend the dataset. For allocators evaluating CRE debt strategies, the reappraisal cohort provides a cleaner valuation signal than any forward-looking model of cap rate expansion or discount rate
shifts, as it is the realized number, filed by a state-licensed appraiser, against which real recovery is measured.
For special servicers, bondholders, and sponsors, the 2025 cohort is the point in the cycle where the path forward for urban office and mall collateral became analytically unavoidable: workout against the new basis, loss crystallization, and eventual redeployment of capital at clearing prices. The cyclically impaired categories, including suburban office, grocery-anchored retail, most multifamily, and industrial, retain optionality. The structurally impaired ones do not, and the 2025 tape makes that distinction visible at scale.