eInformation: a clinical study of investor discussion and sentiment

Financial Management (Financial Management Association), Autumn, 2005 by Sanjiv Das, Asis Martinez-Jerez, Peter Tufano

To analyze response, we measured the speed from the press release to the first discussion on the message boards, the time series of subsequent discussion, and the nature of the discussion. We find the following patterns, which we think of as empirically driven "hypotheses" about the different functions of the boards:

A. Message Boards Provided Factual Foreshadowing of Subsequent Press Releases

In quite a few instances, posters provided readers with advance warning of subsequent news events. For example, one of our tracer events is General Magic's spin-off of its DataRover division. Nine days before the DataRover spin-off, someone reported on the Yahoo! message board that they had found a new DataRover website that did not mention General Magic. This site was apparently taken off-line in a few hours by General Magic, which was probably testing the URL for the imminent spin-off. The board's readers could not only read the message, but also confirm it by going to the site.

In another example, one day before General Magic announced an agreement with Microsoft, someone posted that the two looked like they would share a booth at the Consumer Electronics Show, a tip-off to some closer relationship. A third General Magic poster alerted readers to a local radio broadcast that had suggested that the firm would enter into an agreement with Intuit. The agreement was not publicly announced until a few hours later.

In other instances, posters speculated about upcoming stock splits and bond issues. In two others, we see advance discussion of upcoming earnings numbers, or so-called "whisper numbers" as studied by Bagnoli, Beneish, and Watts (1999).

This anecdotal evidence suggests that posters provide active surveillance, especially of smaller companies, well before the traditional press picks up news events. In these instances, the posters seemed to "stumble across" non-public information, rather than being privy to inside information. Having stumbled across it, they then shared it, possibly to get others to validate it or to make sense of it.

B. Rapid Postings Disseminate Company Information Quickly

Companies tend to issue press releases either before markets open (sometimes in the middle of the night) or after markets close. Table III shows the number of minutes between the time-stamps on each press release and the first posting of the news on one of the stock chat boards. For many, but not all, announcements, the first post is incredibly soon after the news is posted, and in a few cases, prior to the time stamp of the first major news wire story. First posts tend to contain a short notice of the news, often with a URL directing readers to the press release. The boards are apparently serving to disseminate information to interested investors quite quickly. This observation is consistent with those of the poster, Glenn, who mentioned that it was his routine to scan the press about Amazon and post links to new and important stories.

C. Extensive On-Point Discussion Is Sustained for Eight Hours After News Releases