The float you can't check

Public float decides whether a company carries the full weight of SEC reporting, including the audit of its internal controls. It is one number on the cover of the 10-K, and half of it cannot be verified from the outside.

Public float is the number that decides how much of the SEC's requirements apply to a company. Clear the threshold and you are a large accelerated filer: faster deadlines, and the auditor's attestation on your internal controls under Section 404(b). Fall below it and that audit requirement falls away. It is one dollar figure, printed on the cover of the 10-K. I went looking at how solid that figure is, and the honest answer is that half of it cannot be checked from the outside, and for the companies sitting closest to the line, that half is quietly gameable.

What the number actually is

Public float is not market capitalisation. It is the market value of the shares held by non-affiliates: take the company's total shares, remove the ones held by affiliates, and multiply what is left by the share price on the last business day of the second fiscal quarter (Exchange Act Rule 12b-2). So the figure has two moving parts. One is price. The other is the count of shares the company has decided are held by non-affiliates.

Price you can check, shares you cannot

Price is the easy part. It is public, it is time-stamped, and anyone can pull the closing price for that date and confirm it. The share count is the problem.

Whether a given block of stock counts toward float depends on whether the holder is an affiliate, meaning someone who controls, is controlled by, or is under common control with the company (Rule 12b-2).

The lever is bigger than the line

This would not matter much if affiliate holdings were trivial. They are not. Across the companies in our calendar year 2024 population where the figures compare cleanly, affiliates account for at least 20% of market value at about three in four filers, and the median is closer to two fifths. The measurement-date convention adds noise to any single comparison, so I would treat these as a range rather than a point, but even the low end is large.

Set that against the companies sitting near the threshold. Of the 5,272 filers with a positive public float, 124 are within 10% of the proposed $2 billion large accelerated filer line, and 67 are within 5%. The discretionary affiliate exclusion routinely moves far more value than the gap that decides a company's status. A borderline filer does not need to misstate anything. It needs only to decide, defensibly, that one large holder is an affiliate, or is not.

Directors, officers and holders of 10% or more are usually treated as affiliates, but the determination is the company's to make, and companies say so directly. The standard cover-page hedge is that directors and officers can be assumed to be affiliates, but that the statement shall not be construed as an admission for any other purpose. In plain terms, the company tells you it has drawn the line where it chose to, and reserves the right to have drawn it elsewhere.

How crowded the line is

The same crowding shows up at the current thresholds. 130 companies sit within 10% of the $700 million line that separates large accelerated from accelerated filers, and 112 within 10% of the $75 million line below that. These are not rare edge cases. In any given year a few hundred companies are close enough to a filer-status boundary that a modest, defensible change in the affiliate determination would move them across it.

Why you cannot disprove it

Here is what makes the number so hard to police: there is no clean way to check it from the outside. The closest public record is the beneficial ownership table in the proxy, required under Item 403 of Regulation S-K, which lists holders of more than 5%, the directors and the officers. However that table is measured differently. It reports beneficial ownership under Exchange Act Rule 13d-3, as of the most recent practicable date, often lifted straight from a holder's Schedule 13G. The float calculation needs the affiliate share count as of the last business day of the second fiscal quarter. Different definition, different date. So even the disclosure that looks like it should let you reconstruct the affiliate split does not reconcile to the float. You can confirm the price. You cannot confirm the shares.

And no one audits it

The number is not audited either. The public float and the filer-status checkboxes sit on the cover page, outside the financial statements.

It is thinner still in a quarterly report. A 10-Q's financial statements are reviewed, not audited, under PCAOB Auditing Standard 4105, which gives limited assurance and explicitly no opinion. The number that decides whether a company must obtain an audit of its internal controls is itself a number the auditor does not verify.

What our own data shows

Our own work makes the point a different way. To use the as-reported float at all, we first had to clean it. Of the 5,272 filers with a public float, 142, about 2.7%, had a demonstrable scale or decimal error in the figure they tagged in at least one of the two years the proposed rule looks at, and on the most recent cover-page number alone the rate was 1.6%. These were not subtle. All 83 current-year corrections were decimal shifts of roughly a thousandfold or more. The right number was obvious only because the wrong one was absurd.

A scale error that lands a company at a plausible figure, a float tagged 20% high, say, would sail straight through, because there is nothing to check it against. So 2.7% is a floor, not a ceiling, and that is before you reach the affiliate judgement, which leaves no footprint to catch at all.

Why this matters for the reform

The SEC's proposal raises the large accelerated filer line to $2 billion in public float (Release No. 33-11419). Whatever the merits of the threshold, raising it does not fix the input. It still rests on a share count the company determines, that outsiders cannot reconstruct, and that no one audits. For the couple of hundred companies near the line, that is an invitation, not a safeguard.

If the goal is to decide who carries the 404(b) audit on the basis of size, the determination of size needs structure: a documented, consistent affiliate methodology, and ideally an independent check on the share count, not just the price. Standardising float and validating it against a company's own multi-year history is the kind of discipline that would make the number mean what everyone already assumes it means. I came into this expecting to argue about where the line should sit. I came out more worried about the number we are drawing the line with.

Note on the population. We use calendar year 2024 because it is the same basis the SEC used in its own reform projection (Release No. 33-11419), which keeps our numbers directly comparable. Starting from every domestic annual report for the year, Forms 10-K and 10-K/A, we kept Exchange Act filers, set aside foreign private issuers and asset-backed issuers, and reduced multiple filings by the same company to a single record each, which leaves 5,930 registrants. The figures here that rest on public float use the 5,272 of those that reported a positive float. The remaining 658 reported zero or left it blank, almost all of them debt-only or wholly-owned issuers with no public equity to value.