Order flow, trading costs and corporate acquisition announcements - Market Microstructure and Corporate Finance Special Issue

Financial Management (Financial Management Association), Winter, 1992 by Jennifer Conrad, Cathy M. Niden

Jennifer Conrad is an Associate Professor of Finance at the Kenan-Flagler Business School, University of North Carolina, Chapel Hill, North Carolina. Cathy M. Niden is an Assistant Professor of Finance, College of Business Administration, University of Notre Dame, Notre Dame, Indiana.

Many recent papers in finance focus on components of the bid-ask spread. The literature decomposes the spread into adverse selection, inventory cost, and order processing cost components. Copeland and Galai |10~ first argued that an uninformed specialist will increase the spread to compensate for expected losses to privately informed traders. Glosten and Milgrom |19~ termed this increase in the spread the "adverse selection component." Stoll |41~ and George, Kaul, and Nimalendran |17~ estimate the components of the bid-ask spread; both studies conclude that adverse selection costs account for a significant portion of the spread.(1)

In this paper, daily spreads of NYSE firms are examined around acquisition announcements in order to test for evidence of increased adverse selection. The typically large price effect of acquisition announcements gives insiders a strong motive to trade on their information.(2) Consequently, the effects of adverse selection on daily bid-ask spreads may be measured more easily around acquisition announcements than at other times.

Studies using daily data typically have limited success in detecting spread changes around informational events (Morse and Ushman |33~, Skinner |39~, and Venkatesh and Chiang |44~). A notable exception is the recent study by Conrad, Mandelker, Niden, Rosenfeld, and Shastri |8~, hereafter "CMNRS", which examines daily spreads of NASDAQ targets around corporate acquisition announcements and finds a significant spread increase coincident with the announcement. Studies that use intraday spread data document significant increases in the spread prior to important disclosures (Foster and Viswanathan |15~, Lee, Mucklow, and Ready |30~, and Jennings |26~).

We also examine the effect of acquisition announcements on measures of trading activity and on the relation of these measures to the bid-ask spread. Copeland |9~, Verrecchia |45~, Hakansson, Kunkel, and Ohlson |20~, Kyle |27~, Easley and O'Hara |14~, and Admati and Pfleiderer |1~ all suggest that, under most conditions, information will induce increased trading. Sanders and Zdanowicz |36~ and Jarrell and Poulsen |24~ document increased pre-announcement volume associated with corporate acquisition announcements. But the use of a single measure of trading volume may miss important features of information-related trading. For example, if informed volume cannot be distinguished from liquidity (uninformed) volume, then a positive relation between informed volume and spread -- the adverse selection effect -- may be difficult to detect in the negative relation between total volume and spread documented in prior work (e.g., Tinic and West |43~). Easley and O'Hara |14~ suggest that informed traders will choose to trade larger amounts and therefore, that order size may proxy for informed trading. Consistent with this conjecture, both Madhavan |32~ and Laux |28~ document significantly positive relations between spread and average order size during normal trading. If informed trading is associated with larger orders, an increase in order size should be observed around the acquisition announcement and a stronger positive relation between spread and order size should be observed in the event period than in normal periods.

Using a sample of NYSE takeover targets, we examine the time series of changes in spread, volume, average order size and number of transactions around the acquisition announcement. Consistent with the large price effects associated with acquisition announcements, we find substantial changes in spread, volume and number of transactions. We find little evidence of increased adverse selection, i.e., increased spreads, prior to the announcement. A persistent decline in the level of the spread, which coincides with a dramatic increase in trading activity, occurs at and after the announcement.

Significant increases in aggregate trading volume begin three trading days before the first acquisition announcement. The volume increase appears to be due to an increase in the number of transactions; order size does not change significantly in the pre-announcement period. In the framework of Easley and O'Hara's |14~ model, this evidence suggests that the pre-announcement volume run-up is broad-based and not due to the presence of a small number of privately informed investors making abnormally large trades. However, given that the bidder and other informed investors have strong incentives to both acquire target shares prior to the announcement and avoid detection, we cannot discount the possibility that the increase in the number of transactions observed prior to the announcement reflects informed trading.

We also postulate a cross-sectional model in which spread is a function of both order size and order frequency; the sensitivity of spread to these two variables is allowed to shift in the event period. We find that changes in the spread are positively related to changes in average order size. The sign of this relation is consistent with adverse selection during normal trading. This is an important confirmation of the work of Easley and O'Hara |14~, who argue that order size may proxy for informed trading. However, we find no evidence of increased adverse selection around the acquisition announcement.