A decade-long insider trading backtest of scored SEC Form 4 purchases returned over 7× the S&P 500: $10,000 grew to approximately $283,000 by investing $100 in every high-conviction signal, compared to $38,000 for the index. Here's how the simulation worked and what the data shows about insider buying returns.

Insider buying is one of the most discussed and most misunderstood signals in investing. Headlines announce "massive insider buy!" without distinguishing between a CEO writing a personal check and an options exercise that happens automatically. Academic papers confirm insiders outperform, then bury the methodology in footnotes.

We wanted a clear answer, so we ran a straightforward simulation: start with $10,000, invest $100 in every high-conviction insider signal over a decade, and compare the outcome to simply buying the S&P 500. It's the detailed version of the question of whether insider buying beats the market, answered with every rule on the table.

The setup

The rules were simple and consistent across over a decade of data:

• Starting capital: $10,000
• Investment per signal: $100 for every Elite or Strong tier insider purchase
• Entry price filter: Only trades where the stock price was $5 or above at entry
• Entry timing: Buy at the filing-date adjusted close (the price when the SEC filing becomes public)
• Exit strategy: Sell after 63 trading days (approximately 3 months)
• Deduplication: One trade per ticker per filing date, so no double-counting when multiple insiders file on the same day
• Period: Over a decade (2015–2026)

No cherry-picking. Every qualifying signal was allocated $100, subject to available capital: of 12,976 eligible trades, the portfolio took 11,803. The remaining 9% were skipped on days when capital was fully deployed in open positions.

The result

Over the full period, the simulation took 11,803 trades.

Fixed 63-day exit: The portfolio grew from $10,000 to approximately $283,000.

Over the same period, $10,000 invested in the S&P 500 (SPY buy-and-hold) grew to approximately $38,000.

That's roughly 7× the benchmark return.

For context, we also measured a "peak exit" scenario: selling at the best price within the 63-trading-day window rather than at the fixed endpoint. Because peak exits return capital faster, this simulation took 12,586 trades (vs. 11,803 for fixed exit) and reached approximately $498,000, or about 13× the S&P 500.

The peak exit is not a realistic trading strategy, since you'd need perfect timing. But it shows the magnitude of the upside that existed in these signals. Even the conservative fixed-exit approach dramatically outperformed.

What this means (and what it doesn't)

Before going further, the most important caveat: this is a historical backtest, not a tradable execution model.

It does not account for:

• Trading commissions and fees
• Bid-ask spreads (particularly relevant for smaller-cap stocks)
• Market impact from entering and exiting positions
• Slippage between the filing-date close and actual execution
• Tax implications
• The practical challenge of monitoring and executing 11,803 trades over a decade

Together, trading costs, spreads, and slippage would reduce returns by an estimated 4–5%.

It does not mean:

• "You will make 7× the S&P 500 by following insider signals"
• "Elite-tier trades always go up"
• "InsiderSignals is recommending you buy these stocks"

What it does show:

• Across a large sample of trades over a long time period, high-conviction insider purchases, as scored by our model, were followed by returns that significantly exceeded the broad market
• The signal is persistent, not based on a handful of lucky picks
• Scoring and filtering matters, because not all insider trades carry the same informational weight

How the tiers compare

Our model scores every insider equity purchase and assigns it to one of five tiers: Elite, Strong, Positive, Neutral, and Caution.

To see how the tiers compare, we scored all 231,210 qualifying equity purchases across five tiers. Only Elite and Strong were allocated capital in the portfolio simulation.

*Not every eligible trade was taken. The simulation started with $10,000 and allocated $100 per signal, so on days when capital was fully tied up in open positions, trades were skipped regardless of tier.

**Median return over the 63-trading-day holding window across the backtest period; each of the top three tiers landed in the 17–19% range, versus roughly 2–3% for the S&P 500 over a comparable window.

The top three tiers don't represent dramatically different return magnitudes from each other. Instead, they represent confidence and selectivity.

Elite-tier signals are rarer and more concentrated. Elite and Strong together represent the top ~5.6% of all scored purchases in this backtest, and Elite alone is just 0.85%: trades where the insider's role, trade size, holding increase, and historical patterns all align. Strong-tier signals are the next level of conviction.

The difference between tiers is not "Elite makes more money than Strong." It's that Elite trades meet a stricter set of criteria, making them a more curated subset of the already-filtered universe of discretionary insider purchases.

Why 85% of filings don't qualify

This backtest only included the signals that passed our scoring model's filters. The raw universe of insider filings is much larger, and much noisier.

Over 4 million SEC Form 4 filings from the last decade break down roughly as follows:

• 39.3% are sales (dispositions)
• 35.0% are stock grants and awards
• 9.8% are options exercises
• Only 15.5% are actual open-market equity purchases

Most "insider buying alert" services treat all filings the same. A CEO spending $2 million of personal money on open-market shares gets the same treatment as an automatic options exercise that happens on a predetermined schedule.

These are fundamentally different events. The backtest results reflect what happens when you focus exclusively on the 15% that represents discretionary conviction, and then score within that subset for the signals with the strongest historical profiles.

The methodology, in detail

Transparency matters. Here's exactly how the simulation worked:

1. Universe: All SEC Form 4 filings classified as open-market equity purchases from 2015 through 2026
2. Scoring: Each purchase was scored by our logistic regression model, which evaluates insider quality (role, track record), market context (timing, conditions), and transaction strength (size, conviction patterns)
3. Tier assignment: Based on the model's output, each trade was assigned to Elite, Strong, Positive, Neutral, or Caution
4. Trade selection: Only Elite and Strong tier trades with an entry price of $5 or above were included
5. Entry: Adjusted closing price on the SEC filing date
6. Exit (fixed): Adjusted closing price 63 trading days after entry
7. Exit (peak): Highest adjusted closing price within 63 trading days of entry
8. Capital allocation: $100 per qualifying trade, starting from a $10,000 base
9. Deduplication: When multiple insiders filed for the same ticker on the same date, only one $100 allocation was made
10. Benchmark: S&P 500 (SPY) buy-and-hold over the same period, starting from the same $10,000

One thing to be upfront about: the scoring model didn't exist during most of the backtest period. We built it on 2015–2018 data, validated it on 2019–2021, and then scored the full decade retroactively. This backtest shows what the model would have identified, not what it did identify at the time. The 2019–2026 portion is genuine out-of-sample performance. The 2015–2018 portion is in-sample.

What the academic research says

The finding that insider purchases predict positive returns is not new. It's one of the most replicated results in financial economics:

• Lakonishok and Lee (2001) found that insider purchases predict future stock returns, particularly for smaller firms
• Jeng, Metrick, and Zeckhauser (2003) documented that insider purchases earn abnormal returns of more than 6% per year
• Seyhun's extensive research across multiple decades shows that aggregate insider buying correlates with broad market returns
• European research using the AFM's insider trading database confirms that purchases by Dutch executives are followed by significant abnormal returns

Our backtest is consistent with this body of research. What InsiderSignals adds is a scoring layer that distinguishes between the insider purchases that are historically informative and the ones that are routine.

The chart

The chart shows steady separation over the decade, with the gap widening during periods of high insider buying activity.

What this means for investors

The backtest answers the fundamental question: historically, following high-conviction insider signals produced significant outperformance over a large sample and a long time horizon.

But a backtest is a starting point, not a guarantee. Markets change. Past patterns don't always repeat. And the practical realities of executing trades introduce frictions that a simulation doesn't capture.

What the data does provide is a foundation for treating insider filing data as a serious input to investment research: not a curiosity, not a get-rich-quick trick, but a statistically validated signal that has persisted across market cycles, sectors, and economic conditions for over a decade.

Start your free 30-day trial at insidersignals.io and see today's highest-scoring insider filings.

Related Reading

• Does Insider Buying Beat the Market?: The short answer to the question this backtest tests in depth.
• Not All Insiders Are Equal: Which Roles Produce the Strongest Signals: The role-level analysis behind the insider quality score used here.
• WSJ Analyzed 1,400 Insider Buys. Here's What They Missed.: Why unfiltered studies of insider buying find weaker results than this one.

For informational purposes only. Not investment advice.