The Information Content of Aggregate Insider Trading
Seyhun tests insider activity at the market level, a different unit of analysis from a company cluster and one that should not be used as a stock-level label.
This paper moves from individual issuers to the market-wide balance of insider purchases and sales. Seyhun aggregates approximately 60,000 open-market transactions from 1975 through 1981 and studies whether common economic information survives when firm-specific motives are combined. The design asks whether insiders, taken together, respond to economy-wide conditions as well as information unique to their employers.
The evidence is nuanced. Aggregate activity contains an economy-wide component, yet insiders do not always distinguish common forces from firm-specific ones. That makes the paper relevant to a market-level research page, but less direct for interpreting one company or one seven-day cluster. Aggregation can reduce idiosyncratic noise while introducing composition changes, valuation exposure, and timing questions of its own. It also combines purchases and sales whose motives differ materially.
Insider Atlas currently uses aggregates for descriptive counts and reported-dollar flows. Those totals are not a market-timing model. They do not reproduce Seyhun’s portfolio formation, return benchmark, or historical information set. The paper’s appropriate role here is to explain why a broad buy ratio is a separate research object with moderate support, below non-routine purchases and multi-insider purchase clusters in the synthesis ladder. The site therefore keeps aggregate displays apart from company pages and attaches the sales caveat wherever sell activity appears.
Abstract excerpt
“This study investigates the information content of aggregate insider trading by analyzing approximately sixty thousand open-market sales and purchases by insiders.”
Connections in Insider Atlas
The NTSK purchase event and DELL sales event are issuer-level observations. The NTSK company page illustrates why those records should not be mistaken for a market aggregate.