Form 4 by the Numbers: What 221,347 Insider Filings Reveal

InsiderAlpha ingests every SEC Form 4 as it is published. This study covers the deduplicated corpus as it stood on August 3, 2026: 221,347 primary Form 4 filings by 53,567 insiders across 6,114 companies, from December 2022 to that date. The live corpus is larger and grows every trading day; these figures are frozen so every percentage below stays computed on the same rows. Every number below is computed directly from those filings; the methodology, including every data-quality filter we applied, is at the end.

How fast do insiders really file?

Section 16 gives insiders two business days to report a trade. In practice, the median Form 4 arrives 2 calendar days after the earliest transaction it reports, and 90.5% arrive within 4 calendar days (our data records calendar days, so these figures bracket the business-day rule rather than measuring it exactly).

The tail is longer than most investors assume: 3.15% of Form 4s arrive more than 10 days after the trade, about 1 in 30, and 1.55% arrive more than 45 days late. The 99th percentile filing lag is 96 days. Late filings are legal filings - but a three-month old "insider signal" is a very different thing from a two-day-old one, which is why freshness matters when reading insider buying signals.

Insiders sold $4.32 for every $1 they bought in 2026

Across 2026 filings to date, insiders reported roughly $21.5 billion of open-market purchases against $93.0 billion of sales - $4.32 sold for every $1 bought, and 4.6 sale transactions for every purchase (50,755 vs 11,022). That asymmetry is structural, not bearish: insiders are paid in stock and sell to diversify, which is exactly why selling is weak evidence while buying - an insider choosing to concentrate personal wealth in their own company - carries signal.

Who does the buying: owners and directors, not the C-suite

By dollar value of 2026 purchases, 10% owners account for about 41% and directors about 40%, while officers (CEOs, CFOs and the rest of the C-suite) account for only about 13% (roles can overlap, so these do not sum to exactly 100%). The headline-grabbing "CEO buys own stock" story is real but rare; the big money in insider buying is large owners and board members.

96% of insider buying is discretionary

Only 4.02% of 2026 purchase transactions were marked as executed under a pre-scheduled Rule 10b5-1 trading plan. Nearly all insider buying is a deliberate, discretionary decision made in the moment - unlike plan-driven selling, which is scheduled months in advance.

One in three active companies sees cluster buying

Of the 1,890 companies with any insider purchase in 2026, 31% had at least two different insiders buy within a day of each other at least once. Multiple insiders reaching the same conclusion independently is one of the strongest patterns in the literature - the cluster buying signal - and it is far more common than most investors expect.

When filings actually hit EDGAR

Form 4s are a business-hours phenomenon with an after-close rush: Tuesday is the busiest filing day (23.0% of 2026 filings), Monday the lightest weekday (15.9%), and weekends are nearly silent (1.7% on Saturday, zero on Sunday). Within the day, filings spike in the two hours right after the US market close - timestamps in our feed are consistent with a 4pm-6pm Eastern peak. If you check filings once a day, check after the close; our own pipeline lands the median filing in the database 14 minutes after the SEC feed timestamp.

Methodology and data notes

• Corpus: 221,347 primary Form 4 filings (December 2022 to August 3, 2026), deduplicated so amendments and duplicate submissions are counted once. Source: SEC Form 4 filings on EDGAR.
• Filing lag is filing date minus the earliest transaction date on the filing, in calendar days; negative lags and lags over 10 years (52 filings) were excluded as data errors.
• Dollar totals exclude individual transactions above $1 billion to remove a small number of parse-error rows in source filings; this filter removes under 0.3% of transactions. Ratios are therefore conservative estimates.
• Role percentages can overlap (an insider can be both a director and a 10% owner) and are sensitive to the outlier filter; figures shown use the $1B filter.
• Hour-of-day timezone is inferred from feed timestamps, not stored explicitly.

Journalists and researchers: you are welcome to cite these figures with attribution to InsiderAlpha (insideralpha.ai) and a link to this page. For custom cuts of the data, contact us.

Browse the live Form 4 feed → · Recent insider purchases →

This article is informational and is not investment advice.