Insider buying, tested on 17,345 purchases: it didn’t beat similar stocks
Officers’ and directors’ open-market buys at US companies worth $300M or more trailed stocks of the same size and trend by 0.5% over the next year. Big-company CEO and CFO buys trailed by 3.7%.
- Insider buys vs similar stocks, 1 year
- −0.5%
- Purchases
- 17,345
- CEO/CFO, $2B+
- −3.7%
- Its t-statistic
- −3.5
- Peers matched on
- Size, trend
What we found
Open-market purchases by officers and directors at US companies worth $300M or more, from 2013 to 2025, did not beat stocks of the same size and recent trend. A year later they were 0.5% behind (t = −0.9, no edge either way).
CEO and CFO purchases at companies worth more than $2B did worse: 3.7% behind after a year (t = −3.5). The only positive result is small and short-lived: CEO and CFO buys at $300M–$2B companies gained about 1% in the first three months (t ≈ 2), then gave it back.
The full results
- All buys, $300M+
- CEO, CFO or president, $300M–$2B
- CEO, CFO or president, over $2B
- Cluster buys, $300M–$2B
| Group | Events | 1 month | 3 months | 6 months | 12 months | t (12 mo) | Beat peers |
|---|---|---|---|---|---|---|---|
| All buys, $300M+ | 17,345 | +0.0% | +0.0% | −0.1% | −0.5% | −0.8 | 46% |
| CEO, CFO or president, $300M+ | 5,962 | +0.2% | +0.3% | −0.5% | −1.5% | −1.6 | 44% |
| Cluster buys (3+ insiders), $300M+ | 2,575 | +0.3% | −0.1% | −0.4% | −1.9% | −1.3 | 44% |
| All buys, $300M–$2B | 8,762 | +0.1% | +0.3% | +0.6% | +0.8% | +1.0 | 46% |
| CEO, CFO or president, $300M–$2B | 3,373 | +0.6% | +1.1% | +0.4% | +0.5% | +0.4 | 45% |
| Cluster buys, $300M–$2B | 1,534 | +0.6% | +0.2% | −0.2% | −1.2% | −0.7 | 45% |
| All buys, over $2B | 8,583 | −0.1% | −0.1% | −0.7% | −1.7% | −3.3 | 46% |
| CEO, CFO or president, over $2B | 2,589 | −0.2% | −0.3% | −1.3% | −3.7% | −3.4 | 44% |
| Cluster buys, over $2B | 1,041 | −0.2% | −0.7% | −1.3% | −3.6% | −2.4 | 44% |
Events are those with a full 12 months of returns. t is Newey–West. With 36 tests here (9 groups × 4 horizons), a result needs |t| above about 3.2 to survive a Bonferroni correction. Beat peers: the share of events that did better than their peers.
| Group | Events | 1 month | 3 months | 6 months | 12 months | t (12 mo) | Beat peers |
|---|---|---|---|---|---|---|---|
| All sales, $300M+ | 93,469 | −0.1% | −0.1% | −0.1% | −0.1% | −0.1 | 46% |
| Under a 10b5-1 plan, over $2B (2023+) | 8,260 | +0.3% | +0.6% | +0.3% | +2.3% | +2.1 | 43% |
| Discretionary, over $2B (2023+) | 7,206 | +0.2% | −0.4% | −0.6% | −1.6% | −4.1 | 44% |
The 10b5-1 checkbox on Form 4 exists only since April 2023, so the plan split covers about 30 months.
Why the obvious story fails
Insiders know their companies, so their purchases should be informative. But they are also long-term holders who buy after their stock has fallen, and stocks that have fallen tend to keep lagging for a while. Compared with all stocks, insider buys can look good or bad depending on the period; compared with stocks of the same size and recent trend, the edge disappears.
The larger the company, the worse the result. Big-company executives buying their own stock may be signalling confidence more than acting on information, and the market has usually priced what they know.
What is still worth watching
Selling is more telling than buying. Most insider sales are routine (taxes, diversification, pre-set 10b5-1 plans) and say nothing, but discretionary sales at large companies were followed by a 1.6% shortfall over a year. The data is short, and it will be rechecked as it grows.
On a QuantLabs stock page the Insiders tab separates plan sales from discretionary ones, marks cluster buys and shows Form 144 notices of planned sales, so you can see which kind of activity you are looking at.
How it was measured
- Form 4 filings with transaction code P (open-market purchase), $10,000 or more, price at least $1; one event per company per month.
- Bought at the close the day after the filing, when the trade became public.
- Each purchase is compared with peers in the same market-cap decile and prior-six-month-return quintile, over the same days.
- Returns winsorised at 1% and 99%; t-statistics are Newey–West, with the lag matched to the horizon.
Caveats
- Survivors only: companies that later delisted are missing, and those are disproportionately ones insiders bought into, so the true result is likely weaker still.
- An early, uncharted result: discretionary sales (no 10b5-1 plan) at companies over $2B trailed by 1.6% over a year (t = −4.1), but on only 30 months of data, since the plan checkbox began in April 2023.
Questions
- Does insider buying predict stock returns?
- Not on average. Open-market purchases by officers and directors at US companies worth $300M or more (2013–2025) trailed matched peers by 0.5% over a year, a result indistinguishable from zero.
- Are CEO purchases a better signal?
- No. CEO and CFO purchases at companies worth more than $2B trailed peers by 3.7% over a year (t = −3.5). At $300M–$2B companies they gained about 1% in three months, then gave it back.
- Is insider selling a signal?
- Most sales say nothing. Discretionary sales (outside a 10b5-1 plan) at companies over $2B trailed by 1.6% over a year, but on only 30 months of data.
More research
- Data studyWe tested six SEC red flags. Only one predicted returns: filing late
- Market factInvesting all at once beat spreading it over a year 77% of the time
- Market factBuying at an all-time high did no worse than buying any other month
Historical analysis. Past performance does not guarantee future results. Not investment advice. Published Nov 2026. Data: Prices to Sep 25, 2026.