Insider Buying vs Selling: Which Signal Matters

Insider transaction data is one of the few genuinely informative public disclosures available to retail investors. It is also the one most consistently misread, because people treat buying and selling as symmetric signals.

They are not, and the asymmetry is the entire point.

The asymmetry

An executive sells for many reasons. A tax bill on vesting shares. Diversifying a net worth that is 80% one stock. A house. A divorce. A pre-arranged plan set up eighteen months ago. None of these say anything about the company’s prospects, and most selling is mechanical.

An executive buys for essentially one reason. They are spending their own after-tax money, increasing an exposure that is already concentrated, in the one company where they know most about the outlook. There is no tax or diversification story that explains buying.

Reading a Form 4

In the US, insiders must file a Form 4 within two business days of a transaction. It is free and public on the SEC’s EDGAR system.

What to look at, in order of importance:

Transaction code. P is an open-market purchase, the one that matters. S is a sale. M is an option exercise, and A is an award or grant. Option exercises and grants are compensation rather than conviction, and are routinely miscounted as “insider buying” by summary sites.

Whether it was a 10b5-1 plan. Pre-scheduled sales were decided months earlier and carry almost no information. The filing normally indicates this.

Size relative to holdings. A director buying $50,000 while already holding $40 million is noise. An officer increasing their position by 40% is a statement.

Who. A CFO signal generally outweighs a CEO’s, because CFOs see the numbers first and are typically more conservative about the optics of buying. Independent directors carry less weight than operating executives.

The patterns worth tracking

Cluster buying. Several insiders buying independently within a few weeks. One person can be wrong, over-optimistic or acting on personal circumstances. A CFO, two directors and the CEO all buying within a fortnight is much harder to explain away, and this is the strongest form of the signal.

Buying into weakness. Purchases while the stock is falling. This is insiders disagreeing with the market in public, with their own money, at a moment when it would be far easier to wait.

A first purchase in years. An executive who has only ever sold, suddenly buying, is a change in behaviour rather than a continuation of it, and changes in behaviour are more informative than levels.

Unusual selling. The only selling pattern that carries real weight: multiple insiders selling large proportions of their holdings, outside scheduled plans, in a compressed window. That is different in kind from routine vesting sales.

What it cannot tell you

Timing. The information is about value, not about when the market will agree.

Magnitude. A purchase says “I think this is undervalued”. It does not say by how much.

The whole picture. Insiders are subject to the same biases as everyone else, plus a structural one: they are institutionally optimistic about the company they run. Plenty of executives have bought aggressively on the way to zero.

Using it in practice

Insider data works best as a screening input rather than a trade signal.

1. Filter for open-market cluster buys: code P, multiple insiders, non-trivial size.
2. Run the fundamentals. Does the cash flow and margin picture support the insiders’ apparent view, or contradict it?
3. Check for a value trap. Insiders buying a structurally declining business is a falling knife with an endorsement attached.
4. Wait for price confirmation. Because insiders are early, entering when price stops falling costs you some upside and removes much of the drawdown.
5. Size normally. The signal does not justify an oversized position. Apply the same position sizing you would to any other idea.

The realistic value of insider data is that it narrows a universe of thousands of companies to a shortlist worth examining. That is genuinely useful, and considerably less than the way it is usually sold.