Ch. 6: Security Offerings 323
tion is small and insignificant in all of the subsequent studies byMikkelson and Partch
(1986), Eckbo (1986), Hansen and Crutchley (1990), Shyam-Sunder (1991), Chaplinsky
and Hansen (1993), Johnson (1995), Jung, Kim, and Stulz (1996)andHowton, Howton,
and Perfect (1998). The sample-weighted average across all of the studies is a statisti-
cally insignificantARsd=− 0 .22%.
Straight debt issues are to some extent predictable as the maturity date approaches
and the firm needs to refinance.Bayless and Chaplinsky (1991), Chaplinsky and Hansen
(1993)andJung, Kim, and Stulz (1996)develop models to predict whether an issuer will
choose to sell a public issue debt or equity.Chaplinsky and Hansen (1993)examine is-
suers of public debt and find that issues have substantial predictability and that issuers
have significantly lower earnings, significantly higher investment growth and debt refi-
nancing needs in the years immediately preceding and following the offering.Gomes
and Phillips (2005)examine private and public security issuance activity by publicly
listed firms. They find that firms with higher levels of asymmetric information mea-
sured by analysts’ earnings forecast errors or dispersion in earnings forecasts are less
likely to issue common stock or convertibles relative to debt in the public capital mar-
kets, but these firms are more likely to issue equity and convertibles over debt in the
private capital market. They also find that smaller public firms with higher risk, lower
profitability and good investment opportunities are more likely to issue equity and con-
vertibles privately, while firms experiencing stock price rise in the prior year relative to
a benchmark portfolio are more likely to issue equity in the public market.
Since announcement returns represents only the unanticipated portion of the total
price effect, this raises the question of whether partial anticipation explains the largely
insignificant market reaction to straight debt issues in Panel (a) ofTable 15. Eckbo
(1986)addresses this issue by partitioning his sample according to the stated purpose of
the issue (refunding versus finding of investment program), and according to risk (bond
ratings). Presumably, the degree of market anticipation is lower the riskier the debt is-
sue, and if the purpose is to fund new investment opportunities. However,Eckbo (1986)
reports that none of the subsample results sorted in this fashion indicate a significant
market reaction.Shyam-Sunder (1991)also find no effect of bond risk on announce-
ment returns, as measured by bond ratings.Bayless and Chaplinsky (1991)develop a
forecasting model for a firm’s debt versus equity issuance choice and find larger an-
nouncement effects when a security that is not expected is issued. For example, debt
issue announcements when an equity issue was expected have a positive 1% average
abnormal stock return (1 day).Chaplinsky and Hansen (1993)partition the debt sample
according to stated purpose of the issue, and find that the market reaction is insignificant
except in the sample of 68 issues with “no purpose specified” where it is a significantly
negative− 0 .63%. Overall, there are few indications that the evidence in Panel (a) is
significantly affected by partial anticipation. However, this remains a topic for future
research.
Finally, Panel (b) ofTable 15lists studies that report the stock market reaction to
convertible debt offerings. Since convertibles are a hybrid of straight bonds and war-
rants, the risk of overpricing (of the warrant) is greater than for straight bonds. They are