Are Insider Trades Informative?
Lakonishok and Lee document the purchase–sale asymmetry, small-firm concentration, and multi-insider intensity filter that underlie the site’s cluster definition.
Lakonishok and Lee examine insider activity for NYSE, AMEX, and Nasdaq firms from 1975 through 1995. They construct net purchase ratios by transaction count and dollar volume, compare the activity with past returns and valuation measures, and study both market-level and cross-sectional associations. The breadth of the sample and the explicit purchase–sale split make the paper a central bridge between the early event studies and modern Form 4 research.
Three findings shape Insider Atlas. First, insiders as a group behave contrarily, buying after weak past performance and selling after strength, yet simple contrarian controls do not absorb all of the purchase-side association. Second, the evidence is concentrated in smaller firms. Third, purchases carry the information in the sample while average sales largely do not. Their strongest purchase construction also requires at least three distinct insiders plus high net purchase intensity and relative dollar volume. The site’s three-insider, same-week event is inspired by that agreement filter, but it is a simpler descriptive implementation rather than a replication of the paper’s full DPL measure.
The limitations are equally important. There is little immediate market movement around the trade or report, dollar-volume measures can be dominated by one transaction, and historical small-firm results face liquidity and cost concerns. Those points explain why cluster pages show headcount, roles, window, and raw filings without converting them into a security-level outcome. They also explain the standing dead-cat-bounce caveat on sales displays.
Abstract excerpt
“In general, very little market movement is observed when insiders trade and when they report their trades to the SEC.”
Connections in Insider Atlas
Compare the opportunistic BGDE purchase cluster with the SUNB sales cluster and its company page. The caveat attached to the latter follows the paper’s buy–sell asymmetry.