Insider Buying Is the Signal Retail Loves and the Data Mostly Ignores

The Setup

Insider buying is one of the most-watched signals in retail investing, and it is easy to see why. Corporate insiders have to disclose open-market purchases of their own company's stock within two business days on an SEC Form 4. Those filings are public, free, and seductively simple to read. An executive spending real money on their own shares feels like the purest form of conviction there is. Entire screeners, newsletters, and social feeds exist to surface the moment it happens.

The premise of this article is narrow and specific. It is not a claim that insiders are clueless or that their trades are random. It is a claim about what retail thinks the signal does versus what it actually does. The popular belief is that a cluster of executive buying is a forward buy signal: it tells you the stock is about to go up, so you should follow them in. We put that belief through a regime-by-regime backtest. The belief did not survive contact with the data, with one exception that turned out to be more useful than the rule it broke.

The Noise

Start with what is happening right now. As of June 12, the stock with the heaviest concentrated insider buying on the Wealth Engine Pro platform is United Therapeutics (UTHR): 10 different insiders buying, roughly $258 million in open-market purchases, with the CEO personally accounting for about $1.4 million of it. Below it sits Associated Banc-Corp (ASB), where 30 separate insiders bought, the CEO included. Ventas (VTR) shows a single CEO purchase of roughly $4.6 million. Accenture (ACN) and Salesforce (CRM) round out the cluster. This is exactly the kind of activity a retail screener lights up on: senior people, multiple buyers, real dollars.

In a bull or sideways market, our backtest says the correct response to that list is to do nothing with it. Concentrated executive buying in those regimes did not lead to outperformance you could trade. Stocks with heavy insider buying went on to do roughly what comparable stocks without it did. The signal that retail treats as a catalyst was, in the two regimes that describe most of market history, indistinguishable from statistical noise. The screener was telling you something true (insiders were buying) that was also useless (it did not change what happened next).

What the Research Actually Found

This is not a lonely finding. The most-cited academic work on the subject points the same direction once you read past the headline. The landmark study by Josef Lakonishok and Inmoo Lee, published in the Review of Financial Studies, examined every insider trade on the NYSE, AMEX, and Nasdaq across two decades. Their summary is blunter than most people remember: very little price movement happens when insiders trade or when they report it, and insider selling carries little predictive power. Whatever signal exists comes from purchases, not sales.

And the purchase signal that does exist is narrow. Lakonishok and Lee found the predictive effect was concentrated in smaller companies, and that it strengthened when multiple insiders bought at once rather than just one. Later work by Lauren Cohen and co-authors sharpened it further: only opportunistic trades, the irregular, information-driven ones, predicted returns, while routine calendar-driven buying predicted nothing. So the honest version of the academic consensus is not "insiders know everything." It is that a specific, narrow slice of insider purchases carries a modest edge, mostly in small caps, and the broad act-on-it buy signal retail imagines is largely not there. Our regime split is consistent with that, and adds the piece the literature spends less time on: when the edge shows up at all.

The One Regime That Changes Things

Here is the exception, stated precisely as a correlation, because that is what it is:

In bear markets, stocks with concentrated open-market buying by senior executives held within 5% of their price about 70% of the time over the following quarter.

Read that carefully, because the useful part is in what it does not say. It does not say these stocks went up. It does not say insiders timed the bottom. It says they tended to hold their ground while the broader market was falling. In a bear market, not dropping is itself a meaningful outcome. The signal flips from an upside-catalyst story, which the data never supported, to a downside-resilience story, which the data in this one regime actually does. When the people closest to a company's books are willing to step in with their own money while everything around them is selling off, that willingness lined up with prices that subsequently stayed anchored more often than not.

Why a Put-Seller Pays Attention

A stock that holds within 5% of its price is not an exciting story for someone trying to buy low and sell high. It is a very interesting story for someone selling put options. A put-seller does not need the stock to rise. The position makes money if the stock stays flat, drifts up, or even falls modestly, as long as it does not break decisively below the strike. The entire edge is built on a stock holding its ground. That is precisely the behavior the bear-market insider signal lined up with.

This is the difference between an interesting data point and a usable one. Most market signals tell you a direction. This one tells you about stability, and stability is the raw material of premium-selling income. It does not turn a bad setup into a good one on its own, and it is not a substitute for the strike-and-delta discipline that any put sale lives or dies on. But in a falling market, a screen of concentrated senior buying becomes a reasonable place to start hunting for names where assignment risk may be lower than the option premium is implying. We work through that kind of evaluation in detail in The Options Desk, where the entire question is whether a put is worth selling.

What This Does Not Prove

The honest caveat here makes the finding stronger, not weaker, so it is worth stating plainly. Bear markets are rare. That is the whole reason they matter, and it is also the reason the 70% figure rests on a relatively small number of episodes compared to the years of bull and sideways data behind the noise finding. A correlation that holds across a handful of downturns is suggestive, not bankable. It is the kind of relationship that can hold for one decade and soften in the next as the mix of who is buying and why shifts.

And to be exact about the word: this is correlation, not causation. We are not claiming executive buying causes a stock to hold its level. The more likely reading is that both the buying and the resilience are downstream of the same thing, a management team that genuinely believes its shares are mispriced and has the balance sheet to act on it during a panic. That is a reason to treat the signal as one input among several, not a green light. Which, conveniently, is the entire point of the piece.

The Bottom Line

So where does that leave the list of insider buyers stacking up this week? Right now, the Wealth Engine Pro market regime read is Neutral, not bear. The Fear & Greed gauge sits around 31, the VIX is above 22, and market breadth is split roughly evenly at 50% of stocks above their 200-day average. By the finding in this article, that means UTHR and the rest of the current buying leaders fall in the noise bucket today. Not a reason to act. The regime has not tipped.

But the readings underneath that Neutral label are leaning the other way: breadth and junk-bond demand are both flashing the kind of fear that precedes a regime change. That is the thing to watch. The same insider data that is noise in a Neutral tape would start to carry information if the market rolls into a true bear phase. The signal does not change. The regime does.

That is the discipline Wealth Engine Pro is built around. Insider buying is a perfect example of a narrative that feels like data: it is concrete, it is public, and it confirms a story we already want to believe about conviction and skin in the game. The numbers say that story is mostly wrong, and then they say something more interesting than the story ever did. We would rather show you the regime where a signal works and the regimes where it does not than hand you a rule that sounds good and fails three years out of four. Data over narrative, especially when the narrative is this comfortable.