Two Business Days

What this is

On August 31 this series published The Form Split Closed. The Trust Split Did Not., which reported that the Form 1 versus Form 4 gap had gone and that what remained was a trust penalty of about twelve days applying to both forms. It ended by naming two things to watch in September: whether the individual lanes hold a three to five day median in a month where they have no backlog left to make progress against, and whether Form 4 Trust, the last lane to turn, finishes draining.

This is an interim check eleven days into that month, on the same pattern as the August 18 check. It covers 4,077 approvals reported by NFA filers, snapshotted September 11, 2026. Community data, not official ATF records, except where a figure is labelled as ATF's own.

Both questions have answers. A third thing turned up that neither question anticipated, and it is the one most likely to matter to somebody filing this week.

The individual lanes held

Over the fourteen complete days ending September 6, Form 1 Individual ran a median of 2 days on 37 approvals, with a 95% confidence interval of 2 to 3. Form 4 Individual ran 3 days on 39, interval 2 to 4. Both sit at or inside the three to five day range the August edition set as the thing to watch, and neither has drifted up.

Against January, the last month before the 2026 climb began, three of the four lanes are now faster on separated intervals: Form 1 Individual 13 days then against 2 now, Form 4 Individual 11 against 3, and Form 1 Trust 12 against 3. Form 4 Trust ran 10 days in January and 10 now, on intervals of 9 to 13 and 6 to 18, which is the same lane as far as this data can tell.

The trust gap: Form 1 closed it, Form 4 is closing it

Form 1 Trust is the clearer of the two. Over the fourteen days ending September 6 it ran a median of 3 days against Form 1 Individual's 2, with intervals of 2 to 5 and 2 to 3. Those overlap, so the twelve-day trust penalty the August edition described is not visible in the Form 1 lane any more.

⚠️ That rests on 15 approvals. It is a weak yes rather than a strong one, and a single thin week has fooled this series before.

Form 4 Trust is the lane with old work left. On the same fourteen days it ran 10 days on 38 approvals, interval 6 to 18, which does not overlap Form 4 Individual's 2 to 4. Read as a level, that is a trust penalty of about a week, and it would be the only gap left in the queue.

Read across the window rather than as a single number, it is a lane that finished draining inside it. Its weekly medians run 16 days in the week of August 17, 15 in the week of August 24, and 5 in the week of August 31. A fourteen-day window ending September 6 averages the last two of those.

The filing side says the same thing without the averaging. Form 4 Trust filings placed between August 14 and 28 cleared in a median of 5 days on 19 approvals, interval 3 to 8. The same window a month earlier, filings placed July 14 to 28, cleared in 24 days. For Form 1 Trust the two figures are 4 days and 22.

⚠️ A filing-side figure is a floor, not a level. Filings from that window that are still pending cannot appear in it, so it can only understate the wait. It is offered here as the other side of the approval-side number, not as a replacement for it.

The mechanism chart is the one to trust on timing. It shows the share of each week's approvals that were filed within the previous 30 days: a lane working a backlog runs low, a lane that has caught up runs near 100%. Form 1 Individual crosses in the week of July 27. Form 4 Trust reaches 88% in the week of August 10 and Form 1 Trust 89% a week later. In the week of August 31, all four lanes sit at 100%.

Two business days

Here is the thing neither September question anticipated.

Among individual filings placed between July 28 and August 28, the median wait was 2 business days. Half cleared within 2, and 96% within 5. Split by the day of the week the filing was placed, the calendar-day median runs from 2 days for a Monday filing to 5 for a Wednesday or Friday one. Counted in business days, every weekday lands on 2 or 3.

So the spread across the week is the weekend, and nothing else. A Friday filing is not treated differently from a Monday one; it simply has a Saturday and a Sunday sitting inside its wait.

This is a sentence that could not have been written at any earlier point this year. At 60-day waits the weekend is rounding error. At 2-day waits it is most of the variation between one filer's experience and another's, and it is the honest answer to why two people who filed the same week report different numbers.

⚠️ Federal holidays are not removed from the business-day count, so a filing that spans one reads a day slower than it worked out.

Form 1 stopped working weekends in June

From January through April, 26% to 34% of reported Form 1 approvals landed on a Saturday or Sunday. An operation spread evenly over seven days would put 29% there, so Form 1 was, for practical purposes, a seven-day operation. Form 4 sat at 12% to 18% the whole time.

In June that stopped. Form 1 has run 8% to 14% since, which is Form 4's range.

⛔ Note what this does not say. It is a share, not a volume, and it is the date filers report rather than a timesheet. What it rules out is narrow but worth having: the large throughput increase ATF's own ledger shows across July and early August did not arrive as a return to weekend-heavy approvals. Whatever produced it, it happened on weekdays.

An outside check that now reaches three lanes

ATF publishes average processing times by form for each month it finalizes. Its July 2026 table was still the current one when this was written, and it had not been recorded here before.

ATF's July averages were 57 days for eForm 1, 33 for Form 4 Trust and 9 for Form 4 Individual. The mean wait among approvals reported to this site in the same month was 54.0, 32.1 and 9.4 days, on 318, 89 and 143 approvals.

⛔ Means against averages, on purpose. Every headline figure on this site is a median, and comparing a median to ATF's average would be the kind of mismatch that produces a finding out of a definition. ATF does not split Form 1 by owner type, so neither does this comparison.

Previous editions could check this sample against ATF on one lane, individual eForm 4, because that is the only median ATF publishes. This is the first check that reaches Form 1 and Form 4 Trust, the two lanes that carried the weight of every edition since June, and they agree within three days.

The checks

Slower reporting is not manufacturing the trust figures. Recomputing the fourteen days ending September 6 on only records reported within three days of approval gives 11 days for Form 4 Trust against 10 on all records, and 4 days for Form 1 Trust against 3. Both are inside a day of the headline number.

No single state is carrying a lane. Across the most recent three weeks, the most common state is 16% of Form 1 Trust's records and 13% of Form 1 Individual's. A state cluster distorted a week of June, and this is the check that caught it then.

The item-type field still does not select for fast filings. Over the fourteen days ending September 6, labelled and unlabelled records differ by at most a day on the individual lanes. Among labelled Form 1 Individual approvals, short-barreled rifles ran 2 days and suppressors 3, which is what August showed.

What we still cannot see

The sample is thinning. Reported approvals ran about 148 a week in February and about 101 a week across August, and the week of August 31 brought 52. A fast queue produces less to report and fewer reasons to check, and the trust lanes are where that bites first: Form 1 Trust's fourteen-day figure rests on 15 approvals against 75 in January. Nothing here corrects for that, and the confidence intervals get wider as it continues.

We observe a sample, and only approvals. The filers who report are not a random draw from everyone who files, and nothing in this data shows denials or applications returned without action. ATF's finalized counts include those and ours cannot.

We cannot see the intake side this month. ATF's page overwrites itself and still showed July when this was written, so the third reading of the Form 1 backlog, the one that says whether the drawdown held, is not available yet.

What happens next

The queue has arrived somewhere it has not been all year: four lanes, no backlog in any of them, and waits short enough that the calendar matters more than the form. The question that follows is simply whether it stays there, and the honest answer is that nothing in this data can say.

Two of the three things that produced it were temporary in character. A backlog drains once. Intake fell about 41% between ATF's June and August readings, and the tax stamp is still zero, so nothing about that is obviously permanent. The throughput increase is the one that could persist and the one we can see least.

The end-of-September edition has three things to settle: whether the individual lanes hold two to three days through a full month, whether Form 4 Trust's last week was the level or a good week, and what ATF's August figures say about the backlog it published in August.

Method and limits

Figures come from nfa_watch.approvals as of September 11, 2026: 4,175 records carrying both a filing date and an approval date. One record with an implausible wait was excluded. Ninety-five were excluded as duplicates, on the rule that a duplicate requires both an identical (form, owner type, state, filing date, approval date) tuple and an identical masked control number, because our control numbers are truncated for privacy and matching on the mask alone would discard genuinely distinct filings. Two paper filings were excluded, leaving 4,077 records analysed.

Medians group by approval week or approval month, which is the output side of the pipe: what actually cleared in that period. Only complete weeks appear in the weekly charts. Confidence intervals are exact order-statistic intervals computed from the observed records, and they describe sampling uncertainty only, not the gap between this sample and the queue as a whole.

The business-day figures count Monday through Friday between the filing date and the approval date, treat federal holidays as ordinary working days, and are drawn from filings placed between 45 and 14 days before the snapshot. The window stops two weeks short of the snapshot because the newest filings have only had time to show their fastest approvals.

⚠️ One chart deliberately omitted. Earlier editions used the share of each week's approvals filed in an earlier calendar month as the backlog measure. Across a month boundary that measure jumps for a reason that has nothing to do with the queue: in the week of August 31, an August 28 filing cleared on September 2 counts as old work. The 30-day version above cannot do that and is the one this edition reads.

The ATF figures are ATF's own published statistics, read from its Current Processing Times page on September 11, 2026 and banked in this repository because that page overwrites itself monthly with no archive or API.

Charts on this page are locked to the September 11, 2026 snapshot and will not move. The live versions of these studies, which update continuously, are in the Lab. Current medians across all four lanes are on the wait times page.