Event study · 40,427 purchases
Benchmark
Does insider buying actually make money?
Corrected 26 Sep 2026: now on split- and dividend-adjusted prices, headline -0.62% to -0.11% vs the Russell 2000. What changed
Insider-bought stocks rose +2.69% on average over 90 days, about what the market did anyway. Against small caps the edge is nil to negative. Traded naively and equal-weighted, it is not an edge.
Verdict: 90-day excess vs
Russell 2000S&P 500
-0.11%-0.86%
Not distinguishable from zero.Statistically significant, in the wrong direction.
t = -0.8-5.5 ·
win rate 45.4%42.5%.
Raw 90d return+2.69%
Median excess-1.76%-2.68%
Events with full window40,427
We don't sell you a signal. We give you the evidence, so you can judge it yourself.
Curious how your own picks would have done?Test any tickers against the S&P 500, Nasdaq, Dow, Russell 2000 or the insiders themselves.
Backtest my portfolioThe benchmark choice decides the answer.
Insider buying is mostly small caps. Against the S&P 500 the same trades look far worse; most of that gap is company size, not the signal. We report both and treat the Russell 2000 number as the fair one.
vs S&P 500
Large caps, the wrong size match
-0.86%
vs Russell 2000
Small caps, the fairer number
-0.11%
Every strategy, 90-day excess vs
Russell 2000S&P 500
The famous filters do not rescue it
Every reported purchasen=40,427
-0.11%-0.86%Purchases ≥ $250Kn=10,208
-0.85%-1.49%C-suite ≥ $250Kn=2,970
-1.67%-2.07%Cluster buys (3+ insiders / 90d)n=14,376
-0.73%-1.48%Scale −4% ← 0 → +4%
Strategy breakdown
4 cuts · 3 horizons
90-day excess vs Russell 2000S&P 500
-0.11%
-0.86%
Horizon
Events
Raw
vs S&P
vs R2K
Win
t
30d
41,641
+0.68%
-0.21%
+0.13%
46.9%
+1.4
90d
40,427
+2.69%
-0.86%
-0.11%
45.4%
-0.8
180d
38,211
+3.88%
-2.71%
-0.51%
43.9%
-1.8
Median 90-day -1.76% · not significant at 90 days
By filing year, 90-day excess vs
Russell 2000S&P 500
2021-1.51%-5.09%
2022-0.12%+0.29%
2023-0.46%-3.06%
2024+2.85%+1.03%
2025-0.69%+0.47%
2026-2.09%-0.65%
90-day excess vs Russell 2000S&P 500
-0.85%
-1.49%
Horizon
Events
Raw
vs S&P
vs R2K
Win
t
30d
10,558
+0.48%
-0.32%
-0.08%
46.7%
-0.4
90d
10,208
+1.92%
-1.49%
-0.85%
44.2%
-2.3
180d
9,504
+2.31%
-3.96%
-2.12%
41.6%
-2.7
Median 90-day -2.28% · significant at 90 days
By filing year, 90-day excess vs
Russell 2000S&P 500
2021-3.28%-7.28%
2022-0.59%-0.12%
2023-0.83%-3.47%
2024-0.25%-2.08%
2025-0.20%+0.95%
2026-1.30%+0.13%
90-day excess vs Russell 2000S&P 500
-1.67%
-2.07%
Horizon
Events
Raw
vs S&P
vs R2K
Win
t
30d
3,091
+0.15%
-0.65%
-0.39%
46.8%
-1.5
90d
2,970
+1.50%
-2.07%
-1.67%
42.2%
-2.4
180d
2,737
+0.65%
-5.81%
-3.98%
40.1%
-5.4
Median 90-day -3.63% · significant at 90 days
By filing year, 90-day excess vs
Russell 2000S&P 500
2021-2.75%-5.51%
2022-2.37%-1.79%
2023-1.36%-3.89%
2024+1.38%-0.59%
2025-4.91%-3.54%
2026+1.94%+3.23%
Cluster buys (3+ insiders / 90d)
Backtest this as a portfolio →
Several insiders buying the same stock at once.
90-day excess vs Russell 2000S&P 500
-0.73%
-1.48%
Horizon
Events
Raw
vs S&P
vs R2K
Win
t
30d
14,682
+0.40%
-0.51%
-0.17%
47.6%
-1.3
90d
14,376
+2.18%
-1.48%
-0.73%
46.7%
-3.4
180d
13,639
+3.49%
-3.06%
-0.83%
44.7%
-2.6
Median 90-day -1.17% · significant at 90 days
By filing year, 90-day excess vs
Russell 2000S&P 500
2021-1.26%-5.44%
2022-0.97%-0.46%
2023-1.52%-3.90%
2024+3.60%+1.93%
2025-1.90%-0.95%
2026-3.65%-2.05%
01
The famous filters do not rescue it
The literature rates cluster and C-suite buys highest. Here cluster buys are roughly flat and C-suite buys do worst of all.
02
It is regime-dependent
2024 was positive for most strategies; 2026 was sharply negative, as small caps ran hard and raised the bar. A signal that works has to survive the year-by-year cut. This one does not.
03
Equal weighting is doing real work
Microcaps file far more insider buys, so an equal-weighted average is dominated by the smallest, least liquid names, before any trading costs.
04
What this is not
One naive rule: buy on publication, hold blind, sell on a fixed clock. It does not test insider data as one input among others, which is how institutions use it.
Method
Universe
Open-market purchases only (Form 4 code P, non-derivative). Grants, option exercises and sales are excluded.
Entry
The close on the first trading day after the SEC accepted the filing. You cannot trade on a filing before it is public, and the trade itself precedes the filing by up to two business days.
Exit
The close 30 / 90 / 180 calendar days later.
Benchmark
SPY and IWM over the identical window. The headline number is excess return.
Incomplete windows
Dropped, not truncated, so a short window cannot just record the latest market move.
No survivorship repair
Delisted and bankrupt names are missing from the price data, which biases every return upward.
Bias warning
Prices cover 76.6% of tickers with insider buys
(4,223 of 5,516). Delisted and
bankrupt names are missing, so every return here is biased upward:
the real numbers are somewhat worse.
Insider filings work as context, not as a trigger.
Read them next to a company's fundamentals and its own
history, or test a portfolio of your own.
Backtest your own portfolio
Monthly investing in any tickers we cover, against the same cash
in a benchmark or in the insiders' own buys.
Generated 2026-09-26 · prices through 2026-09-24 ·
equal-weighted, no transaction costs · past performance does not
predict future results
