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 select the four firms for our preliminary study, we first collected information on the 3,724 firms that had at least one posting on The Motley Fool (TMF) stock message board during the period July 1, 1998 through January 31, 1999. (1) Then we classified each by the number of TMF messages. For the 504 firms in the TMF list that had at least 25 posts in the period July 1, 1998 through January 31, 1999, we collected the number of major news stories from Factiva. We defined a "major news story" as one in which the name of the company was either in the headline or was mentioned in the lead paragraph and appeared at least three times in the body of the article. We stratified the 504 firms into quintiles along two dimensions (number of posts and news stories) and selected one firm from each of the four extreme categories of the joint distribution. The four firms are shown in Table I. We did not select these four stocks to be representative of the average stock, but rather to help us to understand the extremes of information flow (Table I, Panel A).

Table I, Panel B provides summary statistics on the four firms. Delta Airlines is an old-economy company with a large work force, substantial institutional ownership, and positive earnings. Amazon is considered a flagship new-economy company. General Magic and Geoworks are small, not very profitable, firms attempting to serve the new economy, but General Magic--founded by former Apple Computer executives--has an extremely high level of posting activity for a firm its size. Each of these firms is unique, and each is characterized by quite different information flows, which is the dimension along which we stratify this sample.

Table I shows that these firms are different along other dimensions as well, particularly in levels of trading activity. The two firms with substantial discussion are those with active trading (Amazon and General Magic with 9.4% and 3.5% daily turnover, respectively) and the firms with less substantial discussion show less active daily trading (Delta's trading volume in shares and value exceeds General Magic, but its share turnover is a quarter of General Magic's).

Our sample period is characterized by a high degree of uncertainty about the future, during which perception and sentiment drove values. It is therefore a particularly appropriate period for this study.

II. The Nature of the Posters

Posting a message is a quasi-anonymous act. Posters select a screen-name (or several screen names) and select how much information to reveal. Posters may not be "representative" of the average or marginal investor. For Delta, where institutional investors hold 75% of the shares, posters--who by all accounts are individual investors--surely do not represent the average investor. For the smaller stocks (General Magic and Geoworks), where individuals hold 85% to 90% of all shares, posters are likely to be more representative.

A. Number of Posters and Posting Activity per Screen Name

Table II reports data on the posters for our four boards for the stocks we study. For example, over our study period, 12,169 unique screen-names post 102,820 messages on the four Amazon boards. Although some individuals might have posted under multiple names, we suspect that most of these names represent unique individuals. For the other three stocks, the number of unique posters ranges from 404 (for Delta) to 3,208 (for General Magic). We can compare these numbers with the number of holders of record for each of these stocks in our sample period. If all of the posters are investors, they would represent 2% of the registered holders of Delta, but 528% of the shareholders of Amazon. (2)