An 18-month exam cycle is defensible. Calling it burden relief is not
Widening eligibility for the longer examination cycle is a reasonable risk-based judgment. Selling it as a cut in cost, rather than a reallocation of scarce examiner time, means nobody ever has to prove it worked.
In the United States, banks are inspected on site by government examiners who grade their loans, controls and finances. Small, healthy banks have long been allowed a longer gap between inspections — eighteen months instead of twelve. The federal banking agencies have now widened the group of banks that qualifies, describing it as a reduction in regulatory burden.
Stretching the inspection gap for small, well-run banks is sensible, because frequency was never the same thing as vigilance. But the agencies sold it as cost relief and never said where the freed examiner time goes.
A visit, not an alarm
An examination is a scheduled visit, not a fire alarm. Examiners arrive, read the loan files, test the controls, grade the institution and leave. Whether that happens every twelve months or every eighteen tells you surprisingly little about how fast a bank in trouble gets found, because banks do not deteriorate on the examiner’s calendar. They deteriorate in the gaps.
That is the honest case for what the federal banking agencies did this week when they widened the group of community banks — small, locally focused institutions — eligible for the eighteen-month examination cycle rather than the annual one. It is a defensible judgment about where a fixed pool of examiner hours should be pointed. It is not the judgment the announcement actually made.
The word doing the work
The framing was burden. Reduced regulatory burden, fewer hours of disruption, less management attention diverted from running a bank to hosting people who are grading it. Those costs are real, and they fall hardest on institutions with two compliance staff rather than two hundred.
But burden is an input measure. It counts what supervision consumes, not what it produces. A rule change justified entirely by what it saves has no test attached to it: there is no number that would later show the decision went wrong, because the only number offered was the one that fell on purpose. Supervision is one of the few areas of public administration where the product is an absence — the failure that did not happen, the concentration caught before it metastasised — and measuring it by cost is how absences go unmeasured.
Frequency was never vigilance
Strip out the framing and the substance improves. On-site frequency is among the weaker instruments a supervisor holds. Between visits there are quarterly regulatory filings, off-site risk monitoring, funding and concentration data, consumer complaint flows, and the standing power to bring a scheduled examination forward when something looks wrong. A bank whose risk profile shifts in month seven is not supposed to be invisible until month eighteen.
So this is a bet on the quality of continuous monitoring. That is a respectable bet and the agencies should have made it out loud, because it is the bet that can be lost in public. Framing the same decision as paperwork relief converts a supervisory hypothesis into an accounting entry.
Where the hours go
The strongest argument for the change is the one least emphasised: reallocation. Examiner capacity is finite. An hour not spent grading a well-capitalised bank with clean ratings is an hour available for one drifting towards trouble, or for the resolution of an existing enforcement action. Supervision is a pipeline with an entry and an exit — the same agency announced terminations of enforcement actions against several institutions days earlier — and capacity moved from one end shows up at the other.
That is a claim with a shape. It says hours were moved, and it invites the question of where to. Burden relief says only that hours were removed, and leaves open the possibility that they were removed from the supervisory system altogether. Those are very different policies wearing the same press release.
The strongest case against me
Here is the argument I find hardest to answer. Community banks are concentrated in local commercial property and small business lending, portfolios that turn slowly and then all at once. The on-site examination is the only setting in which a supervisor reads individual credit files and challenges the judgment behind the grading. Everything in a quarterly filing is what management chose to classify; the examination is where classification gets argued with. Stretch the interval by half and an entire leg of a property cycle’s worth of optimistic grading can accumulate before anyone opens a file.
Worse, eligibility is a snapshot. A bank qualifies on the strength of its condition on a particular day and then carries that privilege forward through eighteen months in which the condition may quietly change. The bank most likely to be admitted to the longer cycle is precisely the confident, fast-growing one whose confidence has not yet been tested.
That argument is right, and it is why the eligibility criteria matter far more than the interval. The extended cycle should not be a status earned once and kept until revoked. It should lapse automatically on a ratings downgrade, on a material shift in concentration or funding mix, on any new enforcement action — and supervisors should use, and be seen to use, the power to pull an examination forward. Get the off-ramp right and eighteen months is not the danger. A supervisor who treats the gap as time off is.
What would settle it
One table settles this. Take the banks that moved onto the eighteen-month cycle, look at their condition at the next examination, and compare the downgrade rate with otherwise similar banks kept on the annual schedule. If the extended-cycle group comes back materially worse, the criteria are too loose and the interval is buying trouble. If it does not, the annual visit was buying repetition and the agencies should say so and go further.
Nobody has committed to producing that table. That, and not the eighteen months, is the failure here.
The federal banking agencies should commit now — before the first extended cycles finish — to publishing annually the number of institutions on the eighteen-month schedule, the number whose eligibility lapsed and why, the downgrade rate at their next examination against an annual-cycle comparison group, and a plain statement of where the freed examiner hours were redeployed.
Congressional oversight committees should ask for exactly that at the next supervision hearing and refuse to accept an answer denominated in hours saved. A regulator that cuts the frequency of inspection owes the public the evidence that nothing was missed, not an invoice showing what it cost less.
What would change this view
I am wrong if, within a year of the first extended cycles running their course, the agencies publish two numbers together: how many institutions moved onto the eighteen-month cycle, and the rate at which those institutions were downgraded at their next examination compared with similar banks kept on the annual cycle. If extended-cycle banks come back no worse than their peers, and the agencies can point to where the freed examiner hours went, the burden framing cost nothing and the eligibility criteria are doing their job.
If that happens, the correct conclusion is that frequency really was the loose screw in the system and the agencies should widen eligibility further, on the record, with the data attached. If instead the next announcement extends eligibility again with no such figures published, the pattern is a ratchet justified by cost alone, and the case for treating the change as unexamined gets stronger each time.
Worked from
- Agencies reduce regulatory burden for community banks, increase eligibility for 18-month exam cycle
US Federal Reserve · 2026-09-10