Decoding Inside Information
Cohen, Malloy, and Pomorski make routine-versus-opportunistic classification a first-class requirement: most activity is routine, while documented associations concentrate in the residual group.
Cohen, Malloy, and Pomorski introduce the modern routine-versus-opportunistic split. Their classifier uses an insider’s historical calendar pattern: an insider is routine when trading in the same calendar month across prior years, while activity outside that pattern is opportunistic. More than half of the transaction universe is routine under their approach.
The distinction changes the unit of analysis. Average insider trading combines mechanical liquidity and diversification with a smaller residual that has stronger associations with later firm news and returns in the authors’ sample. Routine portfolios have essentially zero abnormal return in their framework; the documented result resides in the opportunistic subset. Classification, rather than the mere presence of a Form 4, is the central finding.
Insider Atlas treats this as first-class context but cannot reproduce the paper exactly. A full classifier needs several years of clean insider identity and transaction history. The current pipeline uses a deliberately labeled code-based proxy: grants, dispositions to the issuer, tax withholding, exercises, and similar Rule 16b-3 codes are routine; open-market P and S codes remain eligible for opportunistic cluster detection. That proxy identifies transaction mechanics, not a person’s recurring month. The site excludes routine-only windows from headline clusters while retaining them in a drill-down, making the approximation inspectable rather than invisible.
Abstract excerpt
“Using a simple empirical strategy, we decode the information in insider trades.”
Connections in Insider Atlas
Compare the opportunistic BGDE purchase cluster with the LOW routine-only window, then inspect the LOW company history. The badges disclose the site’s narrower proxy.