352 B.E. Eckbo et al.
Ta b l e 2 1
(Continued)Study Issuer
type
Sample
sizeSample
periodHolding
periodαE. Convertible debt offerings
Spiess and Affleck-Graves (1999) All 400 1975–1989 5 yrs − 0. 31 ∗
Eckbo, Masulis, and Norli (2000) Ind 459 1964–1995 5 yrs − 0 .31%
The table reports the time-series estimate of the constant termαresulting from regressing the excess return on
a portfolio of issuing firms on a set of pricing factors in an empirical asset pricing model. The issuer portfolio
is formed using equal-weights. The issuer’s stock typically enters the portfolio in the month following the
issue month, and is held from three to five years. Superscript*indicates that theαis statistically significantly
different from zero at the 1% level.
aSample of non-venture-backed IPOs.
bSample of venture-backed IPOs.
cPricing model with Fama–French, momentum and liquidity factors.
dPricing model with macroeconomic risk factors.
ePricing model with Fama–French factors.
fPricing model with Fama–French, momentum and investment factors.
gPricing model with Fama–French and liquidity factors.
Eckbo, Masulis, and Norli (2000)examine these robustness issues for their sample of
SEOs and debt issues. Below, we discuss their approach, repeat their analysis using our
data, and draw qualitative inferences.^55 Overall, this discussion serves to illustrate that
the main conclusion of zero long-run abnormal performance following issue-activity is
robust.
5.4.1. Alternative and omitted risk factors
The matched-firm technique. The matched-firm technique produces evidence of abnor-
mal post-issue stock returns while the factor regression approach does not. This raises
the question of whether the characteristics-based matched-firm technique omits priced
risk factors. To check this,Eckbo, Masulis, and Norli (2000)andEckbo and Norli
(2005)estimate the abnormal return (alpha) to a zero-investment portfolio that is long
in issuer stocks and short in matched firms. This portfolio controls for any omitted risk
factor with identical factor betas across issuer and matched firm, effectively combining
the two standard matched-firm and asset pricing techniques.
To illustrate, suppose the true set of risk factors is given by the vectorF, and that
only a subsetF 1 of this vector is included in the regression model, with the comple-
ment vectorF 2 omitted. LetIdenote issuer andMmatched firm. The “issuer–match”
(^55) Detailed results are available upon request.