form4 cluster anomaly
What it checks
When 4+ insiders buy in a single month (a real cluster, not noise), the stock tends to outperform for the next few months.
Mechanism
Anomaly-tail Form 4 insider PURCHASE clusters (>=4 distinct insiders within 30 days, or >=3 insiders with aggregate dollar value >= $2M) strip routine-liquidity noise that dominates single-insider buys. Long the issuer 60-180d post-cluster.
Signal rule
rolling-30d distinct insider PURCHASES >= 4 (or >=3 with $>=2M) -> LONG 60/120/180d on T+1 of rising edge.
Data dependencies
daily_pricesAdjusted-close OHLCV for every US-listed ticker; primary price feed.
sec_insider_tradesForm-4 insider transactions with role, size, and trade direction.
Expected edge
- Paper alpha
- ~6.5% over 6mo (anomaly tail)
- Paper window
- T+1 to T+180d
Cohen-Malloy-Pomorski 2012 reports ~6.5% abnormal return over 6mo on clustered, non-routine purchase events.
Related families
insider form4Insider-ActivityLakonishok-Lee 2001, Cohen-Malloy-Pomorski 2012: corroborated insider purchases (Form 4 transaction_code='P') from multiple insiders within a short window are the single most-predictive insider signal — ~6%/year alpha over 6-12 months. Single-insider buys are noisy; cluster buys (3+ distinct insiders, or a single buy > $1M from CEO/CFO) are not. Long for 60 days on cluster buys; long for 30 days on a single $1M+ transaction. Sales are deliberately ignored — they're confounded by liquidity, diversification, and options expiry.
insider routine decoderInsider-ActivityEach insider has a 'routine' footprint (monthly/quarterly 10b5-1 clockwork buys) vs 'opportunistic' trades. Filter OUT routine insiders — the remaining cluster buys predict 82bps/month. Routine alone is zero.
Explore form4 cluster anomaly on alphactor.ai
See which tickers this family is currently firing on, with live signals and rankings.