An insider files a Form 4 when they trade their own company's stock, and they have to do it fast, within two business days of the transaction. Officers, directors, and anyone owning more than 10% of the shares all count as insiders here. So the filing itself is not exotic. What makes it useful is that the person doing the trade knows the business better than you do, they are legally on the hook for what they file, and the clock is short enough that by the time it hits EDGAR the trade is still fresh.
The catch is that most Form 4s are noise. A CEO exercising options and immediately selling, a director unloading shares on a preset schedule, a routine grant that lands in someone's account without them spending a dime. All of that shows up as a Form 4. If you read every one you will drown, and worse, you will start treating sells as signal when they almost never are. The whole game is learning to throw away 95% of what you see and keep the small pile that actually tells you something.
The one transaction code that carries weight
Every line on a Form 4 has a transaction code, a single letter that tells you what actually happened. You care about one of them more than all the others combined: P, an open-market purchase. That is an insider going out and buying shares at the market price, with their own money, when they had every other option in the world for what to do with that cash. Here are the codes worth knowing and how I weigh them:
- P, open-market purchase. This is the signal. Someone chose to buy.
- S, open-market sale. Usually noise. People sell for a mortgage, a divorce, a tax bill, a boat. One reason to buy, a hundred reasons to sell.
- A, a grant or award. The company gave them stock. No conviction in that, they didn't reach into their own pocket.
- M, exercise of an option. They converted a derivative they already held. Often paired with an immediate S to cover the cost and taxes.
- F, shares withheld to pay tax on a vesting event. Pure mechanics, ignore it.
- G, a gift. Also not a market view.
When you see an M followed by an S on the same day, that is not an insider selling out of worry. It is someone monetizing compensation they earned years ago, and it tells you close to nothing about what they think of the stock today. The reflex to read "insider dumping shares" into an M/S pair is where a lot of people go wrong. Strip those out and the filing gets a lot quieter.
Why a CFO's cash purchase beats a splashy grant
The old line is still the most useful thing anyone has ever said about this: insiders sell for a hundred reasons, they buy for exactly one. They think the stock is going up. There is no diversification story, no tax story, no lifestyle story that ends with a person voluntarily concentrating more of their net worth into the company they already work for. So a real P purchase is a piece of information you can lean on.
And not all buyers are equal. A CFO buying is worth more to me than almost anyone else on the org chart, because the CFO sees the numbers before the market does. They know the receivables, the covenant headroom, whether the quarter is tracking. When a CFO puts down two hundred grand of personal cash at the current market price, that is a person with the clearest possible view of the near-term reality voting with their wallet. A board director buying is meaningful too, though a director sitting on eight boards has less of an edge than an operating officer who lives in the P and L every day.
Contrast that with the grant that gets a press release. A big equity award looks like the company believes in itself, but the executive did not choose it and did not pay for it. It was negotiated into a comp package. I would rather see a quiet fifty-thousand-dollar open-market buy by the head of a division than a seven-figure grant with a photo attached.
10b5-1 sales, and why you can mostly ignore them
A lot of insider selling happens under a 10b5-1 plan, which is a mechanism that lets executives schedule trades in advance, at set dates or price triggers, so they can sell even while sitting on material nonpublic information. The whole point of the plan is to remove discretion, and to remove exactly the kind of "I know something bad is coming" timing you might otherwise read into a sale. A Form 4 will usually flag when a transaction was made under a 10b5-1 plan, sometimes with a footnote referencing the plan adoption date.
So when you see a scheduled sale under a plan that was set up eight months ago, there is no fresh information in it. The insider decided to sell before they knew anything about this quarter. Treating that as bearish is reading tea leaves. The reforms that took effect a couple of years back added cooling-off periods and other guardrails, which made these plans even cleaner as a source of noise you can filter out. This is another reason sells are almost always something to discard rather than study.
Filtering the firehose down to what matters
EDGAR gets flooded with Form 4s every day. To make it usable I run everything through a few filters, and this is basically the same logic we bake into the insider signal on Blockcircle:
- Keep only P (open-market purchases). Drop everything else on the first pass.
- Filter out anything flagged 10b5-1, whether buy or sell.
- Weight by who filed. CFO and CEO buys rank above other officers, officers rank above directors, and 10% owners get judged case by case since a fund rebalancing is different from a founder adding.
- Weight by size relative to the person's known holdings and pay. A buy that meaningfully increases someone's stake matters more than a token purchase to hit an ownership guideline.
- Look for clustering. One insider buying is interesting. Three different officers buying the same stock in the same two-week window is a much stronger read.
Clustering is the part people underrate. Any single person can be wrong or can be buying for an idiosyncratic reason. When several insiders who don't coordinate their personal finances all step in around the same time, the odds that they are collectively seeing something the market has not priced go up a lot.
This is not the same thing as congressional trading
It is easy to lump these together, but they are separate worlds with separate rules. Form 4 covers corporate insiders trading their own company under the SEC's Section 16 regime, with a two-business-day deadline and single-letter transaction codes. Congressional trading runs under the STOCK Act, where members of Congress and senior staff disclose trades in any stock, not just one they run, and the reporting window is far more generous, up to about 45 days. The filings look different, they land in different places, and a lawmaker buying a defense contractor tells a completely different story than that company's own CFO buying it.
Both are worth watching, and we track both, but do not read one through the lens of the other. Congressional disclosures are about who might have policy or contract visibility. Form 4 is about the people who actually run the business betting on it with their own money. When you strip a day's filings down to the genuine open-market purchases by operating insiders, weight them by seniority and size, and watch for a cluster, you are looking at one of the cleaner signals available, and you can be looking at it a day or two before most people bother to.