324 B.E. Eckbo et al.
Ta b l e 1 6
Summary of sample-weighted average market reaction (AR, %) to security offerings (aggregate sample size
and sample period in parentheses)
Type of offering U.S. Foreign
A. SEOs
Uninsured rights ARur=− 0. 59 ARur= 0. 70
(53; 1963–1981) (484; 1980–1999)
Standby rights ARsr=− 1. 33 ∗ ARsr=− 1. 32 ∗
(349; 1963–1998) (1,201; 1980–1999)
Private placements ARpp= 2. 45 ∗ ARpp= 3. 12 ∗
(2,830; 1979–2000) (691; 1974–1999)
Firm commitments ARfc=− 2. 22 ∗ ARfc= 1. 10 ∗
(15,017; 1963–2001) (1,064; 1974–1997)
Shelf offerings ARsh=− 0. 66 ∗ n.a.
(1,851; 1980–2003)
B. Debt offerings
Straight debt ARd=− 0. 24 n.a.
(2,615; 1964–1993)
Convertible debt ARcd=− 1. 82 ∗ n.a.
(307; 1964–1982)
TheARreported in this table also appear in the panels headings inTables 13, 14 and 15. The reportedAR
weighs each individual study in the panel with its sample size. Superscript*indicates statistical significance
at the 1% level.
also less predictable than straight debt offerings. So, the expectation is that convertibles
will be met with a stronger market reaction than straight debt issues.Dann and Mikkel-
son (1984), Mikkelson and Partch (1986), Eckbo (1986), andHansen and Crutchley
(1990)all report negative and statistically significant market reactions to convertible
debt offerings. The sample-weighted average abnormal return is a statistically signifi-
cantARcd=− 1 .82%.
4.5. Implications of the announcement-return evidence
For convenience, the sample-weighted averages reported in these tables are summa-
rized inTable 16. The significant price reaction to security offerings leaves little doubt
that these corporate events typically convey significant new information to the market.
As such, the evidence provides generic support for models of the issue decision that
presume some form of asymmetric information between the issuer and the market.
What is more difficult to determine, of course, is the precise content of the new in-
formation that the market is reacting to. We discuss some possible inferences below.
These are the result of cross-sectional analysis of the announcement effect, often per-
formed using multivariate regressions with the announcement effectARas dependent
variable. The expected profits from issuing and investing shown in equation(1), and