University of Iowa Study Examines Anonymity of Insider Trading
Garfinkel and M. Nimalendran from the University of Florida, examine this visibility of insiders' trades in their paper "Market Structure and Trader Anonymity: An Analysis of Insider Trading," recently published in the University of Washington's Journal of Financial and Quantitative Analysis (JFQA).
The research focused on trades by company officers and/or directors in the firm's shares, and asked whether other stock market participants reacted to such trades. The authors chose to focus on insiders' trades because these individuals are often viewed as "informed," and because data on their trading is readily available.
"If other market participants are concerned about trading against informed individuals, they may react differently when insiders trade, as long as there is some detection," Garfinkel says. "Essentially, we set out to examine whether anonymity holds true on the NYSE and NASDAQ trading systems, by looking at the reactions of other market participants on days when insiders traded."
To ascertain whether market participants behaved as if they suspected the presence of an informed trader, the authors focused on changes in effective spreads between days when an insider traded and control days. Effective spreads capture reactions by market makers (specialists on NYSE and dealers on NASDAQ) to trading activity in the stock in which they make a market. Essentially, they measure how much above the midpoint price (between the bid and ask quotes) a trader pays on a buy order and how much below the midpoint price a trader receives on a sell order. If, for whatever reason, market makers feel a need to protect themselves, they widen the effective spread. In other words, the market maker who takes the other side of the transaction "buys lower" and "sells higher."
Garfinkel and Nimalendran found that effective spreads are wider in NYSE stocks on days when insiders trade medium-sized quantities (500-9,999 shares), than on other days. They found no such difference in NASDAQ stocks. Their findings were consistent with the study's joint hypothesis that medium-sized trades are more likely based on private information, and insiders' trades are more transparent on the NYSE than on NASDAQ.
The professors also noted that there is a theoretical basis for the difference in market maker behavior on the NYSE versus NASDAQ. Prior academic work suggests mechanisms by which the specialist might elicit some information about traders' motives from floor brokers. In particular, frequent interaction between these floor brokers and specialists is crucial to such information flows. On the other hand, NASDAQ dealers do not deal frequently with a particular individual, at least not knowingly, since trading is electronic. Thus, they would be unlikely to behave differently on informed or uniformed trades, as they are less likely to have information regarding trader type.
After comparing hundreds of insider trading days, Garfinkel and Nimalendran discovered that the effective spread was larger than usual on insider trading days on the NYSE, but not on the NASDAQ. This finding suggests that the unique relationship between floor brokers and specialists does affect market responses to insider trades on the NYSE, even before they are officially announced.
To receive copies of the paper, please contact the Journal of Financial and Quantitative Analysis at (206) 543-4598 or by email to firstname.lastname@example.org. More information on Garfinkel is available at: www.biz.uiowa.edu/jgarfinkel.
Contact: George McCrory, UI News Service, 319-384-0012