If you look at insider trading data without understanding Rule 10b5-1, you will misread most of what you see. It is the single largest source of noise in insider selling.
The problem it solves
Insiders are almost always in possession of something the market does not have. A CFO knows the quarter is soft weeks before anyone else. If trading while holding material non-public information is illegal, and insiders essentially always hold some, when can they ever sell?
They have to sell sometimes. Most senior pay is delivered in stock, and an executive who could never convert it to cash would effectively be paid in something they cannot spend.
Rule 10b5-1, adopted in 2000, resolves this. An insider can set up a written plan in advance — a schedule specifying the amount, price and timing of future trades — at a moment when they do not hold inside information. Trades later executed under that plan get an affirmative defence against insider trading liability, because the decision was made before the information existed.
A typical plan: sell 5,000 shares on the first trading day of each quarter for the next two years.
Why this matters for reading the data
Consider two sales, both $2 million, both by the same CFO, both reported identically on Form 4:
- Sale A was executed under a plan adopted eighteen months ago. The CFO did nothing this week. A pre-set instruction fired.
- Sale B was a decision made on Tuesday.
Sale A tells you what that person thought a year and a half ago. Sale B tells you something about now.
Headlines that read "CFO dumps $2M of stock" almost never distinguish between them. Most large insider sales are Sale A.
The 2022 rules changed things
In December 2022 the SEC tightened Rule 10b5-1 substantially, after years of research showing the defence was being gamed — plans adopted days before good news, plans cancelled when convenient, overlapping plans that let insiders pick whichever performed best.
The main changes:
Cooling-off periods. Directors and officers must now wait the later of 90 days, or two business days after the next quarterly results are disclosed, before the first trade under a new plan. Everyone else waits 30 days. You can no longer adopt a plan on Monday and sell on Tuesday.
No overlapping plans. Generally one plan at a time for open-market trades, which kills the strategy of running several and using whichever worked out.
Single-trade plans limited. Only one single-trade plan in any twelve-month period.
Good faith throughout. The insider must act in good faith for the life of the plan, not just when adopting it.
Disclosure. Companies must disclose plan adoptions and terminations by directors and officers in their quarterly reports, and — the change that matters most here — Form 4 gained a checkbox.
The checkbox, and the gap before it
Since EDGAR release 23.1 in March 2023, Form 4 carries a machine-readable field declaring that a transaction was made pursuant to a plan intended to satisfy the Rule 10b5-1(c) affirmative defence.
This is genuinely useful. Before it, determining whether a trade was planned meant reading footnotes — free text, inconsistently worded, sometimes absent.
That creates an awkward split in the historical data:
- Filings from March 2023 onwards have a definitive machine-readable answer.
- Everything older has, at best, a sentence in the footnotes.
We also check per-transaction footnotes on modern filings, because the checkbox applies to a whole filing while a single filing can mix planned and unplanned lines.
What to actually do with this
Filter them out. The interesting subset of insider selling is the discretionary part — trades somebody chose to make recently. It is a much smaller list.
Do not treat a planned sale as bearish. It usually means an executive is converting compensation into cash on a schedule set long ago.
Watch for plan adoptions around news. The cooling-off rules make the old abuses harder, but the timing of when an insider chooses to set up a plan still carries information. That is disclosed in company filings rather than on Form 4.
Remember purchases are rarely planned. Almost all open-market insider buying is discretionary. This is part of why buying carries more signal than selling — there is no equivalent noise source.
The short version
A 10b5-1 plan is a legitimate mechanism that lets insiders sell without breaking the law. It also means the large majority of insider selling you see reported carries very little information about what that insider believes today.
Any source that shows you insider selling without telling you which trades were planned is showing you noise and calling it signal.
Related: is insider buying a bullish signal? · every Form 4 transaction code