110 S. Dasgupta and R.G. Hansen
3688 acquisitions, the average two-day abnormal return around the announcement of
an acquisition was 16% (statistically significant at the 5 percent level) for the target;
−0.7% for the acquirer (statistically insignificant); and the combined gain was 1.8%
(statistically significant at the 5% level).Boone and Mulherin (2003)further update the
recent evidence; they find for a sample of acquisitions between 1989 and 1999 that target
returns were on average 21.6%, and that the return to acquirers was an insignificant
−0.7%.
Further cuts on the data provide interesting results on the returns to bidders. Re-
turns to bidders are generally more negative the more is the competition from other
bidders (although seeBoone and Mulherin, 2006b, discussed below). All-stock of-
fers generally yield lower returns to bidders than do all-cash offers (see discussion
below). Returns to bidders are generally more positive when the acquisition is large
relative to the acquirer’s size (Loderer and Martin, 1990; Eckbo and Thorburn, 2000;
Moeller, Schlingemann and Stulz, 2004). One strong empirical regularity is that the to-
tal profit to bidders and targets (as measured by the event studies) is greater for auctions
than for merger negotiations. This is true for both bidders and targets. This may point
to a particular measurement problem: merger bids are often a more drawn-out and par-
tially anticipated takeover process than auctions—which means profits in auctions are
more easily measured. It is also possible that tender offers are more profitable because
they tend to remove old management (to a greater extent than mergers).
The most recent evidence come from the large-sample studies ofBetton, Eckbo and
Thorburn (2005, 2006). They study more than 12,000 publicly traded targets of merger
bids and tender offers over the period 1980–2004. Following the approach ofBetton and
Eckbo (2000), bids are organized sequentially to form contests for a given target, and
they focus in particular on the first and on the winning bidder (which need not be the
same). Since the surprise effect of the initial bid is greater than that of subsequent bids,
and since the initial bidder starts the contest, studying abnormal returns to the initial
bidder yields additional power to test hypotheses concerning the sign and magnitude
of bidder gains. Moreover, since bids are studied sequentially in calendar time, they
present a natural laboratory for testing auction-theoretic and strategic bidding proposi-
tions (toehold bidding, bid preemption, bid jumps, target defenses, etc.).
Initially, Betton, Eckbo and Thorburn follow the tradition and report average abnor-
mal returns for samples of offer outcomes, including “successful” and “unsuccessful”
bids. In the traditional analysis, abnormal returns to “success” (ARs) is found by cu-
mulating abnormal returns from the first bid announcement through completion of the
takeover process which may take several months. The lengthy cumulation adds noise to
this estimate ofARs. Therefore, Betton–Eckbo–Thorburn also report ex-ante estimates
ofARsusing the more precisely measured market reaction to the initial bid announce-
ment only. To illustrate, letxdenote a set of offer characteristics (e.g., bid premium, the
payment method, toehold purchases), andp(x)the probability that the bid will succeed
as a function ofx. The market reactionΓin response to the initial announcement of
bidiis
Γi(xi)=ARsp(xi)+ARu (27)(
1 −p(xi)