Bid‐ask Spread Components Around Anticipated Announcements

Abstract

In this paper I re‐examine spreads around dividend and earnings announcements and provide new evidence on patterns by examining the components of the bid‐ask spread. Transaction data are examined through a recently developed spread decomposition model that decomposes the bid‐ask spread into a fixed (execution) component and an adverse selection component. In addition, this model does not rely on a constant spread as previous spread decomposition models require. The results show that around earnings announcements, the bid‐ask spreads and spread components have significant changes indicating that the anticipated announcement is informative. However, the actual public announcement of a dividend does not alter the bid‐ask spread and spread components of actively traded securities.

Merger Announcements and Trading

“…In addition, they show that at the announcement day, the spread declines significantly and persists at its lower level thereafter. Brooks (1994) finds that only one day before the earning announcements, quoted bid‐ask spread and the estimated adverse‐selection component are significantly higher than those of previous day. In addition, he shows that on and after the announcement days, bid‐ask spread and the estimated adverse‐selection cost decrease significantly.…”

An Intraday Examination of the Components of the Bid–Ask Spread

“…Barclay and Smith (1988) find an increase in spreads around acquisitions and stock repurchase announcements. Brooks (1994) examines the spread around earnings and dividend announcements and finds the changes in the spread and AI component to be consistent with the specialist anticipating and discouraging informed trading. Lee, Mucklow, and Ready (1993) find an increase in spreads around earning announcements.…”

Effect of Regulation FD on Asymmetric Information

“…Masson (1993) proposed an alternative estimation (first suggested by Glosten and Milgrom) and used the bid and ask quote revisions following a trade to estimate the adverse-selection component of the spread. The Masson approach requires only one simple step, as opposed to the three-step approach of Stoll, and it is free of the autocovariance assumptions that plague the methods of Roll (1984) and Stoll. In the study reported here, we used Masson's model (as detailed in Brooks 1994 andBrooks andPatel 2000), mainly for the model's lack of bias, reasonable results, and simplicity in application:…”