What insider selling actually means

The headline problem

"CEO sells $40 million in stock" is a story that writes itself, and it is almost always missing the information that would make it meaningful.

Consider what that number does not tell you. What fraction of their holdings was it — 3% or 80%? Was it a discretionary decision made last week, or the twelfth automatic execution of a plan adopted last October? Was it a sale at all, or the disposal of shares withheld to cover the tax on a vesting grant? Did other insiders sell in the same window, or was this one person?

Without those, the headline number is close to meaningless. With them, it occasionally says something.

Why insiders sell

Insider transactions look like decisions about the company and are usually decisions about a paycheck. Executive compensation at most public companies is dominated by equity. A senior executive may take a salary that is a small fraction of a stock grant vesting over four years. That structure has a mechanical consequence: to convert compensation into money, they must sell.

The routine reasons account for most reported selling:

• Getting paid. Equity compensation is only compensation once it is liquidated.
• Tax. Restricted stock units are taxed as income when they vest. Shares are commonly sold — or withheld by the company — to cover the bill. These "sell to cover" transactions appear as sales.
• Diversification. An executive's salary, bonus, equity, and career are all bets on one company. Reducing that concentration is ordinary financial prudence, and their advisors tell them so.
• Life. Houses, divorces, tuition, philanthropy.

None of these carries information about the company's prospects. They are noise that looks exactly like signal.

The distinction that matters: scheduled vs. discretionary

Rule 10b5-1 lets an insider adopt a written plan — while not in possession of material non-public information — that specifies in advance what will be sold, at what prices, on what dates. Trades then execute automatically, whatever the insider thinks that week.

The rule exists to solve a real problem: insiders are in possession of inside information most of the time, and without a mechanism like this they could barely trade at all. The SEC tightened it in 2022, adding cooling-off periods (generally 90 to 120 days for directors and officers), restricting overlapping plans, and requiring the Form 4 to indicate when a transaction was made under a plan.

That checkbox is the single most useful field on the form.

A scheduled sale tells you what the insider decided months ago. It reflects a diversification decision, not a view on this quarter.

A discretionary sale tells you what they decided recently. It is not proof of anything — people have lives — but it is a decision made with current knowledge, and that is a different category of event.

Treating these as the same thing is the central error in reading insider data.

What actually deserves attention

A sale is worth a second look when several of these hold at once:

1. It is discretionary, not plan-executed.
2. It is large relative to the insider's own position — not large in dollars, and not relative to market cap. A $50m sale by someone holding $2bn is a rounding error; a $2m sale representing most of a CFO's stake is not.
3. Several insiders sell in the same window. One person has private reasons. Five people rarely have the same private reason at the same time.
4. The timing is unusual — shortly before an earnings release, or after a long period of no selling.
5. A plan was terminated early, or a new one adopted at a conspicuous moment.

Even then, this is a prompt to look harder at the business. It is not a conclusion.

Buying is the stronger signal

The asymmetry is the useful part. There are many reasons to sell a stock and essentially one reason to buy it with your own after-tax money.

Research on insider transactions has generally found modest but persistent excess returns following insider purchases — concentrated in smaller companies, where insiders hold a genuine informational advantage over the market, and strongest when several insiders buy independently within a short window.

Caveats that matter in practice. Insiders are frequently early; a purchase can precede another year of decline. Purchases are far rarer than sales, so there is less data. And the effect documented in the literature is an average across many names, not a property of any individual trade you are looking at.

Purchases by a CEO or CFO tend to carry more weight than those by a director, on the straightforward grounds that operating executives see the numbers first.

Reading a Form 4

The form is short and the fields worth reading are few:

• Transaction code. P is an open-market purchase. S is an open-market sale. M is an option exercise, A a grant or award, F shares withheld for taxes. Much of what gets counted as "selling" in aggregate summaries is M and F — mechanical events, not decisions.
• The 10b5-1 checkbox. Scheduled or discretionary.
• Shares owned following transaction. This is what makes the sale size meaningful. It is the denominator the headline always omits.
• Date of transaction versus date of filing. Filings are due within two business days, so the gap is small, but it is not zero.

Filings are free and searchable on the SEC's EDGAR system. Synoptiv surfaces the same data alongside insider sentiment and congressional trading on each analyzed stock, so the clusters are visible without reading filings one at a time.

The honest summary

Insider data is genuinely useful and routinely oversold. It is a public window into the decisions of people who know more than you do, and it is heavily contaminated by compensation mechanics that have nothing to do with prospects.

The practical rules:

• Ignore the dollar headline. Look at the fraction of holdings.
• Check whether the trade was scheduled before reading anything into it.
• Weight buying far more heavily than selling.
• Look for clusters, not individuals.
• Treat any of it as a reason to investigate, never as a reason to trade.

An insider selling stock is, most of the time, an employee being paid. Treating that as a warning will have you selling good businesses for no reason — which is a more expensive error than the one you were trying to avoid.