Methodology
How ClusterDesk works.
What is a cluster buy?
A cluster buy occurs when two or more insiders at the same company purchase shares on the open market within a 5-trading-day window. This pattern is considered a stronger signal than a single insider buying, because it suggests independent conviction across multiple people with direct knowledge of the company.
Methodology
- —Source: SEC Form 4 filings via EDGAR
- —Universe: U.S. micro-cap stocks ($50M to $500M market cap). Institutional funds face liquidity constraints below this range and cannot take meaningful positions without moving the price against themselves, leaving retail investors on equal footing with the same public SEC filing data
- —Open market purchases only (transaction code P)
- —Minimum $25,000 per insider, $100,000 combined cluster value
- —Excludes: 10b5-1 plans, RSU vesting, option exercises, ESPPs
- —Scored 0–100 on cluster size, value, seniority, and recency
- —Only clusters scoring 60+ are published
Scoring
Each cluster is scored on four dimensions: number of insiders (up to 30 pts), total purchase value as a percentage of market cap (up to 25 pts), seniority of roles where CEOs and CFOs score higher than Directors (up to 25 pts), and how tightly clustered the trades are in time (up to 10 pts), plus trade size (up to 10 pts). Maximum score is 100.
Academic research
Insider cluster buying is not a novel idea. It has been studied in peer-reviewed finance journals for decades. The evidence consistently shows that coordinated insider purchases, especially in smaller firms, carry significant predictive power for future returns.
Insider purchases made within 2 days of a colleague's purchase generate 2.1% abnormal returns over the next month, nearly double the 1.2% from solitary insider purchases. Over 90 days, cluster buys yield 5.8% vs. 3.3% for non-cluster buys. The effect is strongest among insiders who work closely together and during periods of high information asymmetry.
Vol. 42(2), pp. 331–360
Insider purchases, not sales, reliably predict future stock returns. The effect is concentrated in smaller firms where informational advantages are greatest. Purchases by multiple insiders amplify the signal. This paper established why micro-cap cluster buys are particularly information-rich.
Vol. 14(1), pp. 79–111
Distinguishes "routine" insider trades (predictable, calendar-based, uninformative) from "opportunistic" ones (irregular, conviction-driven). Opportunistic purchases, the kind that form cluster events, generate 82 basis points of monthly alpha. Routine trades generate none. Won the Chicago Quantitative Alliance Academic Paper Competition First Prize.
Vol. 67(3), pp. 1009–1043
Aggregate net insider purchases predict 60% of one-year-ahead stock market returns across a 15-year sample (1975 to 1989). The more insiders buying simultaneously, across or within companies, the stronger the forward-looking signal. One of the earliest papers quantifying the information content of coordinated insider activity.
Vol. 107(4), pp. 1303–1331
ClusterDesk does not claim to replicate or guarantee the returns described in any academic study. Past research findings do not guarantee future results. This section is provided for informational context only.
Disclaimer
ClusterDesk is for informational and educational purposes only. It is not financial advice, investment advice, or a recommendation to buy or sell any security. Information is derived from publicly available SEC Form 4 filings. Past insider trading patterns do not guarantee future stock performance. Always conduct your own research and consult a licensed financial advisor before making investment decisions. ClusterDesk is not a registered investment adviser.