The Queue That Split: Six Months From One Line to Four

What this is

This edition follows Free Stamps, Frozen Form 1s, which reported three weeks ago that the Form 1 freeze had not released. It has still not released — but "freeze" turns out to be the wrong word, and that is the finding here.

A freeze implies something stuck. What the data shows is something moving, steadily and in one direction, at a rate that can be measured and stated. All figures below come from 3,318 community-reported approval records with both a filing and an approval date, snapshotted 21 July 2026. That is a sample of filer reports, not ATF's internal ledger, and the limits of that are discussed at the end.

In January there was one queue

The clearest way to see what happened in 2026 is to look at where the year started. Grouping every approval by the month ATF cleared it, the January medians read 11 days for Form 4 Individual, 10 for Form 4 Trust, 13 for Form 1 Individual and 12 for Form 1 Trust.

Three days separate the fastest lane from the slowest. In January, for practical purposes, there was no such thing as a Form 1 wait or a Form 4 wait or a trust wait. There was an NFA wait, and it was about a week and a half.

Five months later those same four lanes read 6, 29, 69 and 75 days.

The lines do not cross and they do not wobble. Form 4 Individual drifts slightly down across the period. Form 4 Trust steps up once in February and then holds nearly flat near four weeks. Both Form 1 lanes climb every single month without exception.

The freeze is a slope

Fitting an ordinary-least-squares trend to the complete monthly medians from January through June gives the rate directly:

• –Form 1 · Trust — 11.4 days added per calendar month
• –Form 1 · Individual — 10.7 days added per calendar month
• –Form 4 · Trust — 2.8 days per month
• –Form 4 · Individual — −0.9 days per month

Form 1 is not stuck. It is accruing wait at roughly eleven days for every month that passes, and it has done so with enough regularity that a straight line describes it well. July is not yet complete, but the partial month is consistent: Form 1 Individual is running at 74 days and Form 1 Trust at 83.

The timing is worth stating plainly, and so is its limit. The $0 tax stamp took effect on 1 January 2026, and January is the last month in which the lanes were indistinguishable. Every month after it, the Form 1 lanes separate further. That timing is consistent with a demand shock arriving on the Form 1 side and the queue absorbing it as wait rather than as refusals. It is not proof of causation — this dataset can see approvals, not applications, so it cannot observe the intake that would demonstrate the mechanism. What it can say is that the divergence begins precisely where the policy change does.

Where the four lanes stand now

Over the last 90 days of approvals:

Form 4 Individual runs a median of 7 days, with the middle half of filers landing between 4 and 9 (368 approvals). Form 4 Trust sits at 30 days, middle half 21 to 35 (238). Form 1 Individual is at 64 days, middle half 59 to 71 (365). Form 1 Trust is slowest at 71 days, middle half 65 to 77 (228).

The slowest lane now takes about ten times as long as the fastest. Note also how tight each band is: Form 4 Individual's middle half spans five days, Form 1 Individual's twelve. These are not noisy distributions with overlapping tails. They are four separate, fairly predictable processes that happen to share a name.

For anyone planning around a wait, the practical consequence is that a single "NFA wait time" figure has stopped being useful. In January it was a reasonable thing to quote. It no longer is.

The queue is still fair

A queue that fragments this hard invites the suspicion that something has gone wrong inside it — that filings are being worked out of order, or that some filers are being favoured. The data does not support that.

Comparing every pair of filings placed at least fourteen days apart within a lane, and counting how often the later filer was approved first, Form 4 Individual comes in at 1.1% across 322,842 compared pairs, Form 4 Trust at 2.2% across 211,215, Form 1 Individual at 2.9% across 206,471, and Form 1 Trust at 2.6% across 67,164.

Across roughly 800,000 comparisons, every lane sits under 3%. Given that our filing dates are self-reported and a residual error rate is expected from that alone, this is close to as clean as a measurement like this can look. Whatever is making Form 1 slow, it is not queue-jumping and it is not discretion. The line is orderly; it is simply long.

This matters for how the slowdown should be read. A fair queue that is getting longer is a capacity story, not an integrity story.

Form 1 works Sundays. Form 4 does not.

One further difference separates the two forms, and it is not about speed.

Over the last 180 days, Form 1 approvals land on Sunday 14.2% of the time — very nearly the 14.3% an evenly spread seven-day operation would produce, and Form 1's second-busiest day. Form 4's Sunday share is 5.9%.

The obvious objection is that this is an artifact of reporting rather than approving: filers notice an email on Sunday and log it as Sunday. That objection is testable, and it fails. Sorting the same records by when they entered our database rather than by their approval date produces a completely different shape — 53.6% Monday, 20.8% Tuesday, and only 1.9% Sunday. Ingestion is an early-week phenomenon and is uncorrelated with the approval-date pattern. Nor is it a batch release: Form 1's Sunday approvals are spread across 24 distinct Sundays at an average of 6.7 each, statistically indistinguishable from Form 1's Wednesdays (25 days, 7.0 each).

The decisive point is the comparison between forms. Both are reported by the same people through the same channel and imported by the same pipeline. A quirk of reporting behaviour cannot produce a 2.4× difference between Form 1 and Form 4. The difference has to be in the processing.

The natural reading is that Form 1 clearance is substantially system-driven and runs when the calendar does not matter, while Form 4 depends on steps that keep business hours. We are not in a position to confirm the mechanism, and one part of the pattern remains genuinely unexplained: Form 1's Saturday share is 9.2%, well below its Sunday. A purely automated process should not care which weekend day it is. We report the asymmetry rather than explain it away.

Geography is still a myth

For completeness, one thing that has not changed. Restricting to Form 4 Individual filings so that form type cannot confound the comparison, and to states with at least fifteen reported approvals in the last 180 days, the fastest and slowest state medians are separated by seven days. Twenty-four states cluster inside that.

NFA processing is federal. State-level differences in this dataset remain noise rather than policy, and there is no evidence that where you file changes what you wait.

What this means if you are waiting

• –Which form you filed now matters more than anything else about you. The gap between Form 1 and Form 4 dwarfs the gap between filing as an individual and filing through a trust.
• –On Form 4, the trust decision costs about three weeks at current medians — 30 days against 7. On Form 1 it is worth roughly a week.
• –If you are in the Form 1 queue, plan on the number continuing to move. Eleven days per month is the rate at which recent history has been adding to it. That is an observed trend and not a forecast, and it will stop when capacity or demand changes — but nothing in this data shows it stopping yet.
• –Your position in line is being respected. That is worth knowing while waiting a long time.

Method and limits

Figures come from nfa_watch.approvals as of 21 July 2026: 3,318 records carrying both a filing date and an approval date, with implausible waits excluded. Monthly medians group by approval month — the output side of the pipe, what ATF actually cleared each month — which is a different view from filing-cohort curves and should not be read as one. Only complete months appear in the divergence chart; July is in progress and would read as a dip.

Three limits deserve stating. The data is self-reported and skews toward filers who participate in online communities, so it is a sample rather than a census. It observes approvals only, never applications, which is why the demand-side explanation for the Form 1 slope stays an inference. And a rising median in the approval-month view partly reflects which cohorts happen to be clearing, so it should be read alongside the filing-cohort view rather than instead of it.

Charts on this page are locked to the 21 July 2026 snapshot and will not move. The live versions of these studies, which update continuously, are in the Lab.