Blog
The Day the File Froze
Every loan in your book was a good deal on the day it closed.
The file was complete. The pro forma was reviewed. Committee asked hard questions and got good answers. Approval was earned, not given.
Then the loan funded — and the file froze.
The market didn’t. Rents moved. Expenses moved. The tenant roster moved. The refinancing math moved. But the assumptions that justified the deal — the rent growth, the exit value, the sponsor’s plan — those were written once, on the most optimistic day in the life of the loan, and most of them have never been re-read since.
Payment history gets watched. Assumptions don’t.
Here’s a question worth asking about any performing loan more than a year old: when was the last time someone compared the original underwriting — not the payment record, the assumptions — against what has actually happened?
For most books, the honest answer is the day it closed.
Annual Review, Quarterly Risk
Annual review. Quarterly risk.
Most CRE loan surveillance is built on two mechanisms:
The calendar — the annual review, scheduled by anniversary date, not by anything happening at the property.
The payment — the watchlist trigger, which fires when a borrower goes delinquent.
Both are honest, disciplined processes. And both share the same flaw: they respond to the schedule and the symptom, not to the risk.
By the time a payment breaks, the story is usually two or three years old. The tenant that didn’t renew. The expense line that kept creeping. The rent growth that never showed up. The refinance math that stopped working eight rate meetings ago. All of it was visible — sitting in the gap between the original assumptions and the actual performance — long before anything hit the past-due report.
Delinquency is not an early warning. It’s a late confirmation.
An honest question for credit and risk officers, and we’re genuinely interested in the answers: what is your earliest reliable signal today that a performing loan is quietly becoming a problem loan?
The Fed's Receipt
Banks just told the Fed why they’re easing CRE credit standards. The #1 answer wasn’t confidence in the market.
Every quarter, the Federal Reserve asks senior loan officers why they tightened or eased lending standards. Two findings from the most recent surveys deserve more attention than they’re getting:
First — per the January 2026 survey, banks began easing CRE underwriting standards for the first time since rates started rising in 2022.
Second — per the April 2026 survey, the most cited reason for easing CRE credit policies over the past year was more aggressive competition from other banks and nonbank lenders.
Not improved fundamentals. Not property-level conviction. Competition.
Read those together and the picture is uncomfortable: standards are loosening again, and by the banks’ own account, the primary driver is the auction — the pressure to win deals against other capital.
Competition doesn’t just compress your spread. It shapes the assumptions under the deals you win. The concession isn’t always visible on the term sheet; sometimes it lives in the rent growth you accepted, the exit cap you allowed, the reserve you didn’t require.
The Most Dangerous Day
The most dangerous day in the life of a loan is the day it closes.
Not because anything goes wrong that day. Because of what stops that day.
Up to closing, a deal gets more scrutiny than it will ever get again — underwriter, appraiser, counsel, committee, all focused on the same file at the same time. Then the wire goes out, and something predictable happens:
Everyone’s attention transitions to the next transaction.
The originator has a pipeline. The committee has a calendar. The appraisal is filed. The lawyers close the folder. The deal — along with every assumption baked into it — enters the quietest phase of its life at exactly the moment the real world starts testing those assumptions.
Performance unfolds for years. Scrutiny doesn’t.
This is the second force, and it’s the mirror image of the first. The first force means the assumptions started optimistic. The second means nobody is systematically re-examining them as reality reports in — until the calendar says so, or the payment breaks.
Optimism at formation. Inattention after. Either one alone is survivable. Together, they compound.
Fifty Years, One Pattern
We’ve spent a combined 50 years inside the institutions that hold CRE debt. The same pattern shows up in almost every troubled credit we’ve ever touched.
Between us — across Prudential, GE Capital, BNY, GlassRatner, and SitusAMC; across $27B+ in CRE debt and assets; across origination desks, servicing shops, surveillance platforms, and workout tables
— the post-mortem almost always reads the same way:
The assumptions were optimistic at formation. Not fraudulently — structurally. Every role in the chain pulled the same direction, and the contested deals pulled hardest.
And then nobody re-examined those assumptions while there was still time to act. Attention had moved to the next deal. The calendar review confirmed “performing.” The payment record looked fine right up
until it didn’t.
Two forces. One at the deal’s birth, one for the rest of its life. Every troubled credit is a different story on the surface, and underneath, it’s this story.
A few years ago we stopped treating that as an observation and started treating it as a framework — something that could be named, made explicit, and built into how a portfolio is actually watched.
Interrogate the Baseline
Your stress tests shock the economy. They never question the baseline.
Rates up 200. Vacancy up 500 bps. Cap rates out. Every good credit shop runs these — and they all share one silent assumption: the pro forma being shocked is taken as given.
DRIFT says that’s exactly the wrong thing to take as given. The baseline is where the bias lives.
So we built a second axis. It’s called Bias-Adjusted Return (BAR), and it interrogates the baseline itself: who produced each assumption, what their role’s incentive was, and how contested the deal was when those assumptions were set. Then it makes every incentive-driven assumption explicit and adjustable — and shows you what the deal looks like across a defensible range of bias, not just a range of economic shocks.
A deal can pass every stress test you own and still be mispriced — because the pro forma it started from was the most favorable of five bids. You need both axes. Almost nobody is giving you the second one.
One thing BAR is not, and we’re deliberate about this: it is not a truth machine. It doesn’t claim to know the exact size of the optimism in your book. We don’t claim to know exactly how large the optimism is. We claim you should never again underwrite — or watch a portfolio — as if it were zero.
BAR is a transparency and sensitivity lens. The mechanics are something we walk through with clients on a real loan — usually one of theirs.
The Reckoning
The extend-and-pretend era is ending. The problems it deferred are not.
Roughly $875 billion of commercial real estate debt matures in 2026 — and this year, far less of it will be quietly kicked down the road. Extensions have slowed sharply. Loans that were modified in 2023 and 2024 are coming back to the table, and this time they need a real answer: payoff, refinance, restructure, or workout.
At the same time, the supervisory posture is shifting. As of January 2026, the OCC eliminated mandatory policy-based exam requirements for community banks. Whatever you think of that change, its practical meaning is simple: the examiner is no longer the early-warning system.
You are.
Independent analysis used to be something lenders did partly for the exam. In 2026, it’s something lenders do because no one else is going to catch the weakening credit first.
The institutions that come through this cycle well won’t be the ones with the fewest problem loans. They’ll be the ones who found their problem loans earliest — while options still existed.
Nobody Lied to You
Nobody lied to you. And the numbers were still optimistic.
Before a CRE loan ever funds, its numbers pass through a chain of hands:
The borrower, who genuinely believes in the business plan. The broker, whose package is built to get to yes. The appraiser, who needs the value to support a closing. The underwriter, working under real volume pressure. The credit committee, balancing institutional caution against the pipeline.
No one in that chain intends to mislead. Most are unaware their role is shaping the assumptions at all. But every role pulls — a shade of rent growth here, a slightly tighter exit cap there, a business plan taken a bit more at face value than it would be in a quieter market.
Each pull is small. They all point the same direction. And they compound — invisibly — before the deal is ever funded.
This isn’t a people problem. Replacing everyone in the chain with equally honest professionals produces the same result, because the pulls come from the roles, not the individuals.
Which raises an uncomfortable question: if the bias is structural, what does it do to an entire book of loans, built deal by deal, over years?
Not a Random Sample
Your loan book is not a random sample of your underwriting.
Think about how a deal actually gets onto your books in a contested market. You weren’t the only lender looking at it. Three, four, five credit shops ran the same property through their models. Each produced a view. The borrower took the best one.
You won the deals where your view was the most favorable at the table.
Which means your originated book isn’t a fair cross-section of your credit judgment — it’s a portfolio systematically tilted toward the deals where your assumptions were rosier than everyone else’s. The deals where your caution prevailed? Those closed across the street.
Sit with that for a second, because it changes what “our underwriting is strong” means. It can be true — your underwriting can be genuinely excellent — and the selection effect still tilts the book. Your underwriters didn’t fail. They won auctions.
There’s a cost baked into every contested win. Internally, we’ve taken to calling it the optimism tax: the gap between the assumptions that won the deal and the assumptions a disinterested party would have written.
Nobody ’s at fault. Everybody pays it. And almost nobody measures it.
The Whole Recipe
Small optimism at close. Years of inattention after. That’s the whole recipe.
Here’s a composite — details changed, pattern real, drawn from 50 combined years of files:
A multifamily deal closes after a contested process. Five lenders bid; ours won. Rent growth penciled a shade above the submarket’s history. Exit cap a notch tighter than the last three comparable trades. Each assumption defensible on its own. Committee noted the aggressiveness, approved with conditions, moved on.
Year one: rent growth comes in below plan. Not alarming — one data point. Nobody maps it back to the original pro forma, because no process exists to do that.
Year two: an expense line creeps. The gap between underwritten and actual widens. The loan pays like clockwork. The annual review confirms: performing.
Year three: rates have moved. The refinance math that worked at close no longer does — a fact visible to anyone who re-ran the exit assumptions, which is to say, visible to no one.
Maturity: the payoff doesn’t come. Everyone is surprised.
But here’s the thing — the information was never hidden. Every gap was sitting in plain sight, in the space between the original assumptions and the actual performance, from year one forward. What was missing wasn’t data. It was attention with a method.
DRIFT, Named
DRIFT: Distributed Role-Influenced Focus Transition.
That’s the name we gave the pattern this series has been building toward. The theory, in three sentences:
Distributed, role-influenced bias. Every deal passes through a chain of parties — borrower, broker, appraiser, underwriter, committee — each bringing an unconscious, role-driven pull to the numbers they touch. No one intends to mislead. The pulls point the same direction, and they compound before funding — hardest on the deals you fought to win.
Focus transition. Once the deal closes, the same parties’ attention transitions to the next transaction. The original deal — and the assumptions baked into it — receives diminishing scrutiny at exactly the
moment real-world performance begins testing those assumptions.
The claim. A deal’s true risk is best understood not as a fixed underwriting output, but as the product of these two compounding forces: bias at formation, and the routine-driven loss of focus that follows.
That’s DRIFT. If the past three weeks of posts felt familiar — the frozen file, the chain of hands, the Fed’s competition finding, the book that isn’t a random sample, the attention that leaves at closing — that’s because each one was a piece of it.
The theory is free. We think it should be — naming the problem helps everyone who holds CRE debt. What you do about it is the work.
Attention That Never Leaves
DRIFT says attention leaves the deal at closing. Surveillance built on DRIFT is attention that never leaves.
Picture a credit officer’s quarter running this way:
Every loan’s original assumptions — not just its payment status — are re-examined against actual performance. The gaps are mapped, quantified, and trended, starting in year one, when options are still wide open.
The watchlist is triaged by a different logic: not just “who’s late,” but which credits carry the most embedded optimism from formation and the least attention since — the DRIFT profile of the book. The contested wins get watched hardest, because they earned it.
And every quarter produces a documented, written record — analysis a board, an LP, or an examiner can read and see that this institution systematically re-examines its own assumptions. In a year when
examiners are stepping back, that record is the early-warning system.
That’s the practice. No heroics, no crystal ball — just the two DRIFT forces, run in reverse, on a schedule.
If four weeks of this series has you wondering what it would show on your own paper, here’s the way in: the BAR Snapshot — a fixed-fee, single-loan pilot. You pick one loan. We return the conventional analysis side-by-side with the same deal through the BAR lens. You judge the method on your own paper, at defined cost, before committing to anything larger.
One loan. Fixed fee. Your paper, your verdict.
Thank you for reading along — this series was the theory. The work is
where it gets interesting.