The early bird gets the worm? The stock returns and operating performance of quick SEOs.
Jiang, Yi ; Stohs, Mark ; Xie, Xiaoying 等
I. INTRODUCTION
Some firms decide to issue seasoned equity offerings (SEOs) very
quickly after an initial public offering (IPO). Such corporate decisions
are puzzling for the following reasons. First, it is known that the
average impact on firm value from an SEO is negative. So why issue SEOs
at all? Second, most firms issuing SEOs typically wait one to three
years after an IPO. Why then do some act so quickly, i.e., by issuing
SEOs within six months of an IPO, a very sensitive time period for
existing shareholders? (1) The primary objective of our research is to
investigate why some firms issue SEOs quickly following their IPOs.
While prior studies have improved our understanding of related
decisions, very few have examined the time between an IPO and the first
SEO. Our contribution fills this void.
The finance literature offers two primary explanations for quick
SEO decisions. First is the market feedback hypothesis, which states
that high stock returns after an IPO signal that the marginal return to
the project is high, encouraging managers to increase investment by
raising additional capital. Jegadeesh, Weinstein, and Welch (1993) find
that firms experiencing larger post-IPO returns tend to issue SEOs
within three years of their IPOs, and that the size of the SEOs is
larger, which they interpret as being consistent with the market
feedback hypothesis.
The second explanation points to market timing opportunities (also
labeled as the overvaluation hypothesis by Myers and Majluf (1984)).
According to this hypothesis, managers use quick SEOs to take advantage
of "open financing windows" to sell overvalued equity.
Overall, we find support for the market timing/overvaluation
hypothesis in explaining firms' SEO decisions shortly after their
IPOs. The support relies on studying publicly traded firms that issue
SEOs within six months of their IPOs. Specifically, we address the
following research questions: (1) which hypothesis best explains why
firms conduct quick SEOs? (2) How does the market react to the
announcement of a quick SEO? (3) What is the long-run stock performance
of firms conducting quick SEOs? (4) What is the operating performance of
our sample firms?
Jegadeesh, Weinstein, and Welch (1993) find that firms with larger
post-IPO returns are more likely to issue SEOs within three years of
their IPOs. (2) They interpret their results as being most supportive of
the market feedback hypothesis. Yet they overlook the overvaluation
hypothesis and some of their results do not support the market feedback
hypothesis. Their analysis employs a long (three-year) window. We
suggest that a short window is more likely to capture a firm's
equity issuance decision soon after its IPO. In addition, we examine the
post-issue performance of SEO firms to detect whether managers engage in
market timing.
DeAngelo, DeAngelo, and Stulz (2010) study the factors determining
a firm's decision to issue an SEO at a given year and find that
near-term cash need is the primary motivation for SEOs. They argue that
both market-timing opportunities and a firm's corporate lifecycle
(which is defined as the number of years listed) play a statistically
significant but only ancillary role in the decision. DeAngelo et al.
(2010) treat the time between an IPO and SEO as exogenous to an SEO
issue decision, using the number of years listed as a proxy for a
firm's lifecycle. Our analysis differs by treating the time between
an IPO and SEO as endogenously determined by firm characteristics and
market conditions. To control for the corporate lifecycle hypothesis,
following Loughran and Ritter (2004), we use the number of years since
the founding date of the firm as a proxy for a firm's lifecycle
stage. (3) Our results suggest that the market-timing hypothesis holds
after controlling for a firm's lifecycle stage.
There has been very little empirical work on the timing of a
firm's first SEO. Since firms issuing early SEOs tend to be
smaller, younger, and risker, the SEO timing decision may be extremely
valuable to them. In contrast, firms conducting SEOs at least one to
three years after their IPOs are larger, more mature, and are followed
by more analysts (e.g., Jegadeesh, Weinstein and Welch, 1993; DeAngelo,
DeAngelo, and Stulz, 2010; Krigman, Shaw, and Womack, 2001). Issuing
SEOs when the equity market is hot may not be as valuable for these more
mature firms. Jegadeesh, Weinstein and Welch (1993) interpret their
results as support for the market feedback hypothesis. DeAngelo,
DeAngelo, and Stulz (2010) argue near-term cash need is the primary SEO
motive. Krigman, Shaw, and Womack (2001) examine underwriter switching
for follow-on SEOs and find evidence consistent with signaling
hypothesis. Using these findings as a foundation, we reason that an
analysis of quick SEOs should provide more evidence about the proper
role of the market timing hypothesis. An analysis of quick SEOs is of
particular interest because of the high frequency of such SEOs in the US
stock market. In our sample, many firms use quick SEOs to reach a higher
level of capitalization soon after their IPOs. Thus, identifying factors
determining the time between firms' IPOs and their first SEOs is of
great importance to investors.
Our research approach is as follows. First, we address the question
of why some firms issue SEOs earlier than others. Results indicate that
quick SEO firms tend to be smaller and younger, have larger IPO issue
size and greater IPO underpricing, and larger post-IPO stock price
run-ups.
Next, we examine market reaction by analyzing the announcement
effect of SEOs. Prior research has generally demonstrated a negative
announcement effect upon an SEO announcement (e.g., Asquith and Mullins,
1986; Eckbo and Masulis, 1995). (4) Our study differs in that we compare
the announcement effect for firms issuing very early SEOs and late SEOs.
We find that the market is more surprised by quick SEOs and that the
price decline associated with the SEO announcement is more severe for
these firms, presumably because of stock price overvaluation. We argue
that such results support the overvaluation rather than the investment
opportunity hypothesis.
Third, we analyze whether the market properly values firms.
Specifically, if companies announce stock issues when their stock is
grossly overvalued, can the market reevaluate the stock appropriately,
or will the stock still be substantially overvalued when the issue
occurs? To address this question, we compare the long-term stock returns
of firms issuing SEOs in our sample against five alternative matching
benchmarks. Consistent with Loughran and Ritter (1995), we find strong
evidence of poor performance following equity issuance. The mean
three-year buy-and-hold abnormal-return (BHAR) for all SEOs in our
sample is -23.13%, while quick SEOs have a more negative BHAR of
-59.97%, compared with a BHAR of -17.54% for late SEOs. Our results
indicate current shareholders benefit from a quick SEO, while new
shareholders suffer a loss in the long-run. (5)
To evaluate the impact of the timing of SEOs and firm
characteristics on the firm's subsequent share performance more
thoroughly, we perform multivariate regressions of BHARs on the
logarithm of the time between the IPO and the first SEO (or early issue
dummy), pre-issue stock-price appreciation, and other control variables.
We find that firms' three-year BHARs are positively related to the
logarithm of the time between IPO and the first SEO (or negatively
related to the early issue dummy), which provides evidence for the
poorer long-run performance of quick SEOs.
Beyond the buy-and-hold returns approach, we employ two additional
procedures to examine the underperformance of firms conducting quick
SEOs. The first procedure uses a time-series of cross-sectional
regressions on monthly individual firm returns. The results suggest that
firms conducting new issues underperform by 41.5 basis points per month,
and firms conducting quick SEOs underperform by 111 basis points per
month. (6) This evidence suggests that firms conducting quick SEOs
experience more severe underperformance.
The second procedure is the calendar-time portfolio analysis. We
regress portfolio excess returns on Fama-French's three factors and
report the "alphas," which measure the monthly abnormal
returns associated with the SEO announcement. In the three-factor
regressions, the alphas of non-issuers are larger than the alphas of
issuers. For all issuers, the alpha of issuers conducting a late SEO
significantly exceeds that of issuers conducting an early SEO. These
results also support the overvaluation hypothesis.
We also consider the market feedback hypothesis as an alternative
to market timing. This hypothesis implies that investments increase with
aftermarket returns. Hence, firms issuing early SEOs should have higher
investment rates. Hovakimian and Hutton (2010) examine repeat SEOs and
document a positive relationship between the first year post-issue
returns and the likelihood of a follow-on equity issuance. They
interpret their results as most consistent with the market feedback
hypothesis: that a high post-issue return encourages managers to
increase the firm's investment because the marginal return to the
project is high. We test this hypothesis by estimating regressions of
investment on aftermarket returns, an SEO within six months of IPO
dummy, as well as the interaction variables between aftermarket returns
and the six months dummy. Our estimation results are inconsistent with
the market feedback hypothesis.
Finally, we examine whether the timing of an SEO affects post-issue
operating performance. We find that firms conducting quick SEOs exhibit
the most severe decline in operating performance among all the issuing
firms. As the inflated stock price cannot be sustained following the
IPO, the returns decline, reflecting poor operating performance. This
finding is also consistent with the overvaluation hypothesis.
The rest of the paper is organized as follows. Section II describes
our hypotheses and data, Section III discusses the methodology for
measuring SEO underperformance, Section IV presents the main results,
Section V provides robustness checks, and Section VI summarizes.
II. HYPOTHESES AND DATA
A. Hypotheses
The market feedback hypothesis states that high stock returns
signal high marginal returns to the projects, which in turn, encourage
managers to increase investment by raising additional capital. The
hypothesis therefore predicts:
H1: Firms with higher aftermarket returns are more likely to issue
SEOs earlier than firms with lower aftermarket returns.
Intuitively, firms with high aftermarket returns are high quality
firms with good investment opportunities. It is more costly for
high-quality firms to defer their investments in new projects than it is
for low quality firms.
H2: The market reacts less unfavorably to the announcement of quick
SEOs.
If firms with good investment opportunities are more likely to
issue quick SEOs, the market should be less surprised at SEO
announcements by these firms.
H3: Firms conducting quick SEOs exhibit relatively better long-run
stock performance.
If firms that issue quick SEOs are high-quality firms with good
investment opportunities, then these firms will exhibit better long-run
stock performance after the issue.
H4: Investment rates are higher for firms that issue SEOs shortly
after IPOs.
High aftermarket returns encourage managers to increase the
firm's investments because the marginal return to the project is
high. Thus, the investment rates should be higher for quick SEOs.
H5: Firms conducting quick SEOs exhibit stronger post-issue
operating performance.
The overvaluation hypothesis, in contrast, states that firms issue
equity when they believe their stock prices are overvalued relative to
management's private information. Thus, market timing hypothesis
predicts:
H1a: Firms with higher aftermarket returns are more likely to issue
quick SEOs than firms with lower aftermarket returns.
Under the overvaluation hypothesis, if managers believe their
stocks are overvalued, they tend to issue quick SEOs to exploit
"windows of opportunity" in ways that benefit existing
shareholders.
H2a: The market reacts more unfavorably to the announcement of
quick SEOs.
The market treats the SEO announcement shortly after an IPO less
favorably because such equity issuances might signal a greater degree of
stock price overvaluation.
H3a: Firms issuing quick SEOs experience poorer long-run stock
performance.
If the stock prices of firms issuing quick SEOs are even more
significantly overvalued than the others, then the poorer long-run
performance is merely a consequence of the market's failure to
incorporate all the information. The stock is still substantially
overvalued when the issue occurs.
H4a: Investment rates are not necessarily higher for firms that
issue quick SEOs.
If a firm's equity issuance decision is driven by
overvaluation rather than good investment opportunities, investment
rates may not be higher for firms issuing quick SEOs.
H5a: Firms conducting quick SEOs exhibit no better or even worse
post-issue operating performance.
The rationale behind this proposition is that after the issue, as
the inflated stock price cannot be sustained, the returns may decline,
reflecting poor operating performance.
B. Data
We use Thomson Financial's SDC Global New Issues database to
identify firms that conduct IPOs during 1970-2006, and then select the
first-time SEOs by these firms for the same time period. Our ending date
is restricted to 2006 so that we have available data from CRSP to
compute long-run returns. Our sample IPOs satisfy the following
criteria: (1) include only common share offers listed on NYSE (the New
York Stock Exchange), AMEX (the American Stock Exchange) or NASDAQ; (2)
exclude IPOs with offer price [less than or equal to] $5 (7); (3)
exclude IPOs with gross proceeds (in real 1984 dollar) less than $1
million; (4) exclude financial companies, such as banking, insurance and
REITs (SIC codes between 6000-6999) and utility companies (SIC codes
4900-4949); (5) exclude unit offers, spinoffs, carve-outs, rights, and
shelf offerings (8); (6) include only firms with stock return data
available in CRSP after the issue, and with financial data available in
COMPUSTAT, and (7) exclude firms with a market cap of less than $10
million at the time of issue during 1970-2006 to minimize the influence
of outliers in the analysis. The resulting sample consists of 1,610
first time SEOs.
Table 1 reports summary statistics of firm characteristics and
other main explanatory variables used in the paper, with more complete
definitions and the COMPUSTAT origins of data presented in the Appendix.
AT is the number of calendar days between IPO and the first SEO. The
median value of AT is about one and half years (496 days). Under IPO is
IPO underpricing, defined as the difference between the first day
post-issue price and the IPO offer price divided by the offer price,
with a median underpricing of 8.93%. AB RET 20 is the abnormal return
over the period from trading day 1 to trading day 20 after the IPO date,
with a median of 3.79%. AB RET 40 is the abnormal return over the period
from trading day 21 to trading day 40 after the IPO date. The median
abnormal return 20 days before SEO issues is 3.59% (not presented in
Table 1), indicating the fact that SEO firms experience strong price
run-ups prior to the issue. SEO AR is the SEO 3-day announcement period
abnormal return, calculated over the event days -1, 0, and +1. The
median SEO AR equals -3.40%.
Table 1 also reports firm characteristics traditionally used to
identify market timing. The median market value of equity for our SEO
sample is $277.73 million. The mean is larger at $694.49 million,
indicating skewness of distribution. Our sample firms have a median
Tobin's Q of 1.91, which suggests that the typical SEO firm is
profitable and has valuable growth opportunities. Finally, firms that
issue SEOs on average raise 2.03 times as much capital through SEOs as
they raise from their IPOs (measured by SEO/IPO).
We distinguish primary share offerings and secondary share
offerings for the sample. For SEO issues, the shares offered may include
pure primary shares (newly created shares that generate proceeds for the
firms), pure secondary shares (insider's shares that do not
increase the cash holdings of firms) or a mix of both. In the unreported
univariate analysis, we find that for early issuers, 18% (39 firms) are
pure secondary offerings, 69% (147 firms) are mixed offerings (60% of
which (89 firms) with primary shares less than 50% of total offers); for
late issuers, 13% are pure secondary offerings, 54% are mixed offerings
(32% of which with primary shares less than 50% of total offers). We
define pure secondary offerings or mixed offers with secondary shares
greater than 50% as "second". For early issuers, 60% of firms
are "second", and for late issuers, 30% of firms are
"second". The difference is significant at the 1% level. Table
1 shows that for the overall sample, 34 percent of the first time SEO
issues are dominated by secondary offerings.
Krigman, Shaw, and Womack (2001) examine underwriter switching for
follow-on SEOs and document that SEO firms often switch to underwriters
with a higher reputation and more analyst coverage. To control for the
potential issue of underwriter quality on very quick SEOs, we construct
several variables relating to the underwriters' reputation (IPO
underwriter reputation, SEO underwriter reputation, and whether firms
switch to better ranked underwriters in their first SEOs) to proxy for
information asymmetry in share offerings. We use the Carter-Manaster
(CM) ranking, which is based on an underwriter's relative position
in IPO/SEO tombstone announcements. This measure is developed by Carter
and Manaster (1990) and extended by Carter, Dark, and Singh (1998) and
Loughran and Ritter (2004). (9) We find that the reputation of SEO
underwriters (SEORANK) is on average higher than that of IPO
underwriters (IPORANK). About 18 percent of firms in our sample switched
to better ranked underwriters for their SEOs (SWITCHBETTER).
It is argued that the integer offer price of IPO or SEO may serve
as a measure of uncertainty as investment bankers are more likely to
settle for an integer price if they cannot discover the true intrinsic
price of an offer. Mola and Loughran (2004) argue that integer price
clustering provides evidence that reputable investment banks use their
influence to extract rents from issuing firms. Bradley et al (2004) find
that IPOs with integer offer prices experience higher underpricing.
Table 1 shows that 79 percent of our sample firms issued IPOs with
integer offer prices (INTEGER_IPO), and 41% of them issued SEOs with
integer offer prices (INTEGER_SEO).
Figure 1 presents the number of SEOs in our sample by year and the
proportion of quick SEOs. The volume of SEOs displays large variations
over time, with the period 1991-2000 being the "hot" issue
period, and we observe a higher proportion of quick SEOs during this
period as well. Quick SEOs account for 10%-30% of all SEOs during this
hot issue period.
[FIGURE 1 OMITTED]
III. MEASURING SEO UNDERPERFORMANCE
We use three procedures to examine the underperformance of seasoned
equity offerings. The first procedure is the BHAR analysis. The second
procedure uses a time-series of cross-sectional regressions on monthly
returns. The last procedure is the Fama-French three-factor regressions.
A. Buy-and-hold Abnormal Returns
Extensive literature exists about long-run stock performance
following corporate events, yet long-term studies on stock returns
remain controversial. We follow Billett, Flannery, and Garfinkel (2005)
in calculating the BHAR as (10):
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (1)
where [R.sub.i,t] is the daily return for firm i, T is the number
of trading days in the three-year window following the issue, and
[R.sub.i,T] is the cumulative holding period return.
For each issuing firm, we select five separate sets of matching
non-issuing firms, following Vijh's procedure (1999). (11) We
discuss the results based on the last set, where non-issuing firms are
matched by size, industry and book-to-market ratio.
B. Cross-sectional Regressions on Monthly Returns
Our second procedure for measuring SEO underperformance uses a
time-series of cross-sectional regressions based on monthly individual
firm returns. We run cross-sectional regressions on all firms listed on
NASDAQ, AMEX, or NYSE during 1970-2006 as follows:
[r.sub.it] = a + bln M[V.sub.it] + clnB/[M.sub.it] +
d[ISSUE.sub.it] + eISSUE6[month.sub.it] + [[epsilon].sub.it], (2)
where lnMV is the natural logarithm of the market value of equity
(MV EQ), lnB/M is the natural logarithm of the ratio of the book value
of equity to the market value of equity, and the book value is the book
value of equity for the most recent fiscal year end. ISSUE is a dummy
variable which equals one if a company conducted at least one public
equity offering (SEO or IPO) within the 60 months preceding a given June
30th. ISSUE6month is a dummy variable which equals one if a company
conducted an SEO within six months of its IPO. The dependent variable is
the monthly percentage of stock returns. This procedure allows us to
test whether there is an independent "new issues effect" and
whether firms conducting quick SEOs experience more severe
underperformance.
C. Fama-French Three-factor Regressions
Our third approach is to compute the calendar time abnormal return
and compare it with the buy-and-hold abnormal returns. Barber and Lyon
(1996), Kothari and Warner (1997), and Lyon, Barber, and Tsai (1999)
suggest that unbiased statistical significance levels are difficult to
compute using buy-and-hold returns. Consequently, long-run returns are
commonly computed using the Fama and French (1993) three-factor
timeseries model:
([R.sub.pt] - [R.sub.ft]) = a + b([R.sub.mt] - [R.sub.ft]) +
s[SMB.sub.t] + h[HML.sub.t] + [[epsilon].sub.t], (3)
where [R.sub.pt] is the equally weighted portfolio returns of
sample firms in month t; [R.sub.mt] is the return on the
equally-weighted index of NYSE, AMEX, and NASDAQ stocks in month t;
[R.sub.ft] is the three-month T-bill yield in month t; [SMB.sub.t] is
the return on small firms minus the return on large firms in month t,
and HMLt is the return on high book-to-market stocks minus the return on
low book-to-market stocks in month t. The intercepts from these
regressions are interpreted as abnormal returns. Abnormal returns will
be associated with the event studied if the intercepts in the
regressions are economically and statistically significant.
IV. RESULTS
A. Why Do Some Firms Return to the Equity Market Earlier Than the
Others?
We begin our analysis by examining why some firms return to the
equity issue market earlier than the others. Results are presented in
Table 2. We first focus on what kind of firms is more likely to issue
SEOs within six months of IPO. To address the concern that the "6
months" classification of "early" issue is
"arbitrary," we also use a continuous variable Ln[DELTA]T,
defined as the logarithm of the time between a firm's IPO and its
first SEO, as a dependent variable. To address the concern that the
IPO/SEO market may have changed over time (Loughran and Ritter, 2004)
regarding types of issuers, incentives of issues and market conditions,
we run the analyses for both the full sample (1970-2006) and subsample
(1990-2006). We also run the regression analyses for the subsample
(1990-2006) by excluding the period 1999-2000 to address the concern if
our results are driven by the internet bubble period in particular.
Our probit regression shows that firms with larger IPO underpricing
(Under IPO) are more likely to conduct an "early" issue. This
is consistent with the signaling hypothesis of IPOs by Chemmanur (1993)
and Welch (1989), which proposes that firms underprice their IPOs so
they can subsequently issue seasoned equity at a favorable price, and
can return more quickly to the equity market with SEOs. The coefficients
of AB RET 20 and AB RET 40 are also positively significant, suggesting
that firms experiencing larger stock price run-ups after the IPO tend to
return to the market with SEOs earlier than other firms. Since large
pre-issue stock price appreciation signals that the current stock price
is overvalued, the above results are consistent with the hypothesis that
management uses their private information to time equity offerings to
take advantage of the "windows of opportunity."
The associated coefficient estimates on firm size and firm age are
all negative, suggesting that larger firms and older firms are less
likely to conduct quick SEOs (t < 6 months following the IPO). We
also find that firms with dominating secondary share offerings are more
likely conduct a quick SEO. Regarding the impact of underwriter
reputation, the regression shows that firms with better ranked IPO
underwriters are less likely, while firms with better ranked SEO
underwriters are more likely to issue quick SEOs. In addition, a firm is
less likely to launch a quick SEO if the firm chooses to switch to a
better ranked underwriter for its SEO.
Similar results are found in regressions on the length of time
between an IPO and the first SEO. In addition to taking advantage of
overvalued stocks, the results also show that firms with higher
expenditure ratios and dominating secondary share offerings return to
the equity issuance market more quickly. The results for the subsample
(1990-2006) and the subsample excluding the internet bubble period are
mostly consistent with the full sample analyses. Overall, we conclude
that a firm's decision to issue a quick SEO is explained by market
timing rather than broader economic considerations. (12)
B. Market Reaction
Prior research generally shows a negative announcement effect upon
the announcement of an SEO (Loughran and Ritter, 1995; Houston and
Ryngaert, 1997; Jegadeesh, Weinstein and Welch, 1993). (13) We extend
the existing literature by examining the relations between the timing of
SEOs and SEO announcement effects. Specifically, we address whether or
not the market is more surprised by firms that conduct quick SEOs. If
the market timing hypothesis holds, we expect to find a less favorable
market response because an earlier SEO issue may signal a greater degree
of stock overvaluation. To address this question, we report the abnormal
returns around SEO announcements categorized by length of time since the
IPO at the date of the first SEO in Panel A of Table 3.
Consistent with the market timing hypothesis, we find that the
price decline associated with SEO announcement is more severe for the
group of firms conducting SEOs within six months of their IPOs (with an
SEO 3-day announcement period abnormal return of -5.79%). Panel B of
Table 3 shows the difference test between the groups of firms issuing
"quick SEOs" and firms whose SEO takes place more than six
months after an IPO. Based on the t-test and Wilcoxon test, the
differences in the SEO 3-day announcement-period abnormal return are
statistically significant at the 1% level. Table 3 also shows that among
the seven groups of firms classified by the timing of SEOs, the group of
firms conducting "quick SEOs" (t < 6 months) experience the
largest AB RET 20 (22.29%) and AB RET 40 (11.39%). This finding confirms
our results in Table 2, indicating that firms with higher stock price
run-ups after the IPO tend to return to the market with an SEO earlier
than the others.
To provide an additional test of the hypothesis that the market
might be more surprised by SEO announcements shortly after IPOs, we
conduct a regression analysis with the dependent variable being the SEO
3-day announcement abnormal returns, and the key independent variables
being either an "issuing within dix months" dummy or a
continuous variable representing the time difference between IPO and the
first SEO (lnAT). The results are presented in Table 4.
The coefficient of the six month dummy is negative and significant
for the full sample and the sub-sample excluding the bubble period,
suggesting that the market is more surprised by "quick SEOs"
and that the price decline associated with the SEO announcement is more
severe for these firms. The interaction between market overvaluation (AB
RET 20) and quick SEOs dummy is negative and significant for the full
sample period and the subsample period. The decision to issue equity for
these firms appears to be driven more by overvaluation rather than by
investment opportunities. Hence, there is a more negative market
reaction when good motivations are not apparent. The regression with
continuous days (ln[DELTA]T) also shows that firms waiting longer to
return to equity issue market generally experience higher announcement
abnormal returns. In addition, the SECOND dummy shows that SEOs with
dominating secondary share offerings tend to experience more negative
announcement abnormal returns.
C. Buy-and-hold Abnormal Returns Analysis
Table 5 reports the three-year buy-and-hold abnormal returns for
the sample firms between 1970 and 2006. Consistent with prior studies,
firms issuing SEOs underperform their size, industry and book-to-market
matched counterparts. The mean BHAR is -23.13% and is reliably different
from zero. Similar results are obtained for our sample firms based on
the other four alternative matching methods. The poor long-run
performance of issuers suggests that the market does not fully react to
the information implied by an equity issue announcement, because only
part of the overvaluation problem is corrected upon the announcement of
an equity issue.
Panel A of Table 5 reports the BHAR as categorized by the length of
time since IPO at the date of first SEO. Among the seven groups of firms
classified by the timing of SEOs, the group of firms conducting
"quick SEOs" (t < 6 months) experiences the most severe
long-run underperformance, as shown using size, industry and
book-to-market matched benchmarks. Panel B of Table 5 shows the
difference test between the groups of firms who use quick SEOs and those
with SEOs after six months of IPOs. Using size, industry and
book-to-market matched peer firms, firms returning to the equity market
within six months of an IPO experience a three-year BHAR of -59.97%,
while firms conducting SEOs more than six months of IPOs experience a
BHAR of -17.54%. The difference in the BHAR between the two groups of
firms is negative and statistically significant (p=0.035).
To evaluate the impact of SEO timing and firm characteristics on a
firm's subsequent share performance in more detail, we run
multivariate regressions of BHAR with the key explanatory variables
being the "quick SEO" dummy (the six months dummy) or the
logarithm of the time between a firm's IPO and its first SEO
(ln[DELTA]T). Table 6 reports the regression results.
We find that firms issuing "quick SEOs" experience lower
BHAR for the full sample period 1970-2006 and the subsample period. In
general, a longer waiting time between an IPO and the first SEO (lnAT)
is associated with higher ex post peer-adjusted, long-term stock
returns. This result provides further evidence about the poorer long-run
performance of firms using quick SEOs. In addition, we find that firms
with dominating secondary share offerings in their first SEO do not
perform better in the long run.
D. Cross-sectional Regressions on Monthly Returns
To test whether there is a "new issues effect"
independent of a more severe underperformance of firms conducting SEOs
within six months of IPOs, we perform a time-series of cross-sectional
regressions on monthly individual firm returns following Loughran and
Ritter (1995). Table 7 presents the multivariate analysis of monthly
firm returns under seven different model specifications
The key variables we examine are "ISSUE" and "ISSUE
6 month". In the full model (7), the coefficients of ISSUE and
ISSUE6month indicate that firms conducting new issues underperform by
41.5 basis points per month, and firms conducting quick SEOs
underperform by additional 69.5 basis points per month. In model (2), we
report the average coefficients for monthly regressions where the sole
explanatory variable is the new issue dummy variable. The mean parameter
value of -0.47 indicates that firms conducting new issues subsequently
underperform by 47 basis points. In model (3), where the only
explanatory variable is the ISSUE 6 month dummy, the coefficient
estimate of -0.968 implies that firms conducting quick SEOs subsequently
underperform by 96.8 basis points. In model (4), when we consider both
the new issues effect and the effect of quick SEOs, firms conducting
quick SEOs underperform by 64.9 basis points. The results in model (4)
and (7) imply that the underperformance of new issues cannot be solely
attributed to the size and book-to-market effects. Instead, a "new
issues effect" exists, demonstrating that issuing firms
underperform non-issuing firms, and that firms conducting quick SEOs
experience more severe underperformance.
E. Fama-French Three-factor Regressions
Table 8 reports the alphas from time-series regressions of monthly
portfolio excess returns on Fama-French three factors, as used in Fama
et al. (1993). The advantage of forming portfolios is that the
cross-sectional dependence problem in Table 7 is reduced while the
disadvantage is that power is sacrificed. We find that for all firms,
the alpha of non-issuers exceeds that of issuers by 0.50 on a monthly
basis, and the difference is statistically significant at the 1% level
(Panel A). We split the sample into large firms and small firms. Large
firms are those whose market capitalization is above the size of the
median NASDAQ, AMEX and NYSE firm in the sample. We find that for small
firms, the alpha of non-issuers is significantly higher than that of
issuers.
We also form portfolios of firms issuing an SEO both within six
months of an IPO (quick SEO) and more than six months following an IPO
(Panel B). The results show that for all issuers, the alpha of firms
conducting an SEO more than six months following an IPO exceeds that of
quick SEO issuers by 1.09 on a monthly basis. We find negative
differences when we split the sample into large and small issuers,
though the differences are not statistically significant. Overall, we
find underperformance of issuers and more severe underperformance of
those with "quick SEOs".
F. Aftermarket Returns and Investments
According to the market feedback hypothesis, high stock returns
signal that the marginal return to the project is high, which encourages
managers to increase investment by raising additional capital.
Therefore, firms issuing quick SEOs should have higher investment rates.
We test this hypothesis by estimating regressions of investment
(measured by total net property, plant and equipment, following
Hovakimian and Hutton (2010) (14) on aftermarket returns, six month
issue dummy or the logarithm of the time between a firm's IPO and
its first SEO (lnAT), and the interaction variables between aftermarket
returns and the 6 months dummy. We also include control variables such
as book-to-market, free cash flow and other firm characteristics.
Results reported in Table 9 are inconsistent with the market
feedback hypothesis. The coefficient estimates on AB RET 20 is negative
and significant and the coefficient on SEO within six months of the IPO
dummy is negative (though insignificant) for the full sample period and
the period of 1990-2006, instead of being positively significant. The
coefficients of the interacted terms are statistically insignificant
from zero. Overall, we find no evidence that capital expenditures
increase with the aftermarket returns for firms that conduct SEOs within
six months of IPOs. Interestingly, the SECOND dummy carries a
significantly negative sign, suggesting that SEOs with dominating
secondary share offerings actually decrease investment expenditures.
G. Changes in Operating Performance
Finally, we examine the operating performance of firms conducting
SEOs by addressing the questions: (1) does the post-issue operating
performance of issuers deteriorate relative to non-issuing firms? and
(2) is there more severe deterioration of operating performance among
the group of issuers who conducted SEOs shortly after IPOs? Table 10
presents the results.
Table 10 reports the median operating performance ratios for
issuers and non-issuers matched on industry, firm size and pre-issue
operating performance. (15) The matching procedure follows Barber and
Lyon (1996) and Bouwman, Fuller, and Nain (2009). We match the
non-issuance firms with SEO firms by: (1) same industry and year; (2)
size is within 70% and 130% range; (3) for those firms that satisfy
criteria (1) and (2), we select the firms with the closest pre-issue
performance (return on assets, ROA). We report the results based on the
median ROA (defined as Operating Income before Depreciation scaled by
assets). To examine whether the timing of an SEO affects post-issue
operating performance, we categorize the issuing firm by the length of
time since IPO at the date of the first SEO: early issuer (an SEO issued
within six months of an IPO) vs. late issuer (an SEO issued more than
six months following an IPO).
We find that the median late issuer on average demonstrates a
better post-SEO operating performance than the median non-issuer, but
the median early issuer does not. The difference between the two
categories (difference in adjusted) is statistically significant for
year +1, +2 and +3. In addition, when comparing matching firms adjusted
operating performance change from prior-SEO to post-SEO (-1 to 1, -1 to
2, and -1 to 3), we find median early issuers demonstrate more
deterioration than median late issuers. For example, for late issuers,
the median benchmark-adjusted ROA from year -1 to year 2 is around 0,
but the median benchmark-adjusted ROA for early issuers from year -1 to
year 2 is -3.25%. The difference between early issuers and late issuers
is statistically significant. Again, our results indicate that firms
conducting an SEO shortly after going public exhibit the most severe
decline in operating performance among all the issuing firms.
V. ROBUSTNESS CHECKS (16)
A. Impact of Lockup Period and Firm Performance
Existing literature has documented that many IPOs specify a lockup
period for future equity issues, and in general, most lockup
restrictions expire six months after the IPO (Field and Hanka, 2002).
Chen, Chen, and Huang (2012) find that insiders' (especially senior
executives) selling of shares has a negative impact on the long-run
stock returns subsequent to the lockup expiration. To test the impact of
lockup days, we perform univariate and regression analyses.
The median lockup period of early issuers is 180 days, not
different from that of late issuers. Figure 2 provides the lockup
patterns of IPOs and SEOs for our quick SEO sample. (17) To examine the
impact of lockup period on SEO firm performance, we further control for
the lockup period in the announcement return and buy-and-hold abnormal
return regression analyses, and find no impact of lockup period on the
more negative performance of quick SEOs.
B. Impact of Secondary Shares Offering, Venture-capital Backed
Offerings, and High-tech Industry
The shares offered with SEOs may include pure primary shares, pure
secondary shares or a mix of both. As motivations of firms issuing
secondary shares may differ significantly from those mainly issuing
primary shares, we perform a robustness check to test whether our
results are driven by secondary shares offerings. In particular, we
exclude pure secondary offerings from our sample and only keep firms
with at least some newly issued primary shares (Loughran and Ritter,
1995). Re-running all analyses confirms the original results.
It is also possible that if an IPO is venture capital (VC) based,
the existing owners may desire to exit via secondary shares as early as
possible (Brav and Gompers, 1997). To test this hypothesis, we tabulate
the percentage of VC-backed IPOs for early issuers and late issuers. The
key market timing opportunities variables, IPO underpricing, AB RET 20,
and AB RET 40, remain significant after controlling for VCs and
offerings type. (18)
Another concern is that high-tech/internet firms may behave
differently, as noted in the literature. (19) Loughran and Ritter (2004)
find that riskier IPOs offered by high-tech firms are more underpriced
than less-risky IPOs. Bartov et al. (2002) document differences in IPO
valuations between internet and non-internet firms. In our sample,
38.32% of early issuers are high tech/internet firms while 21.56% late
issuers are high tech/internet firms. In the subsample, about 26% of IPO
issuers are high tech firms during 1990-2006 (about 33% of IPO issuers
are high tech firms during 19952000). We add a high-tech firm dummy
variable to the regressions and the coefficient estimates for our key
market-timing variables are still significant after controlling for
high-tech firms. (20)
[FIGURE 2 OMITTED]
C. Cash Needs and the Possibility of Quick SEO Issue
As noted above, the results of DeAngelo, DeAngelo and Stulz (2010)
and McLean (2011) suggest that firms' near-term cash needs are the
primary motivation for SEOs. We address this concern by performing two
analyses, and find that our main results still hold. (21) We believe
near-term cash need around SEOs is a powerful factor determining whether
a firm issues an SEO, but it may not predict or explain why some firms
issue sooner than others.
VI. CONCLUSION
Our research investigates whether firms take advantage of
transitory "windows of opportunity" to time seasoned equity
issues when their equity is substantially overvalued with respect to
managers' private information. Our main results provide support for
the market timing hypothesis. First, we find that firms experiencing
larger IPO underpricing, larger stock price run-ups after the IPO, and
larger IPO offer size tend to return to the market with an SEO earlier
than the others. This implies that overvalued firms tend to time their
equity issues. Second, we find that firms issuing quick SEOs on average
earn a 2.69% lower three-day announcement excess return than those
issuing late SEOs, indicating that the market treats quick SEO
announcements less favorably because such equity issues might signal a
greater degree of stock price overvaluation.
Third, we show that a firm's three-year BHAR is positively
related to the logarithm of the time between its IPO and first SEO.
Using three different approaches (the buy-and-hold return analysis,
cross-sectional regressions, and calendar time portfolio analysis) we
document more severe underperformance for firms issuing quick SEOs. The
results hold after controlling for the effects of firm age, secondary
share offerings, lockup period and venture capital based nature of IPOs.
In addition, we find no evidence that investments increase with
aftermarket stock returns for firms conducting quick SEOs, which is
inconsistent with the market feedback hypothesis. Our results also
suggest that firms conducting SEOs shortly after their IPOs exhibit the
most severe deterioration in operating performance among all issuing
firms.
In general, the combined evidence is consistent with the
overvaluation hypothesis that managers with private information time
SEOs in ways that benefit existing shareholders. We find little support
for the market feedback hypothesis, which assumes that firms issuing
SEOs shortly after IPOs are high-quality firms with good investment
opportunities. Because their stock is more overvalued at the time of
issuance, firms conducting quick SEOs are worse off in terms of: (1) the
market's reaction to their SEO announcement, (2) their long-run
share returns, and (3) their subsequent operating performance. These
results are best explained by management's ability to time the
market by issuing overvalued equity to take advantage of the
"windows of opportunity."
ENDNOTES
(1.) See AT (days) in Table 1. The period is sensitive because
after an initial public offering, most existing shareholders are subject
to a lock-up period in which they cannot sell their shares for a
pre-specified time. In addition, asymmetric information is supposed to
be greater for new IPO firms, especially in the first few months after
the IPO.
(2.) They study 411 first SEOs issued during 1980-1986. They found
similar results when they used a five-year period window.
(3.) Loughran and Ritter (2004) define firm age as the year of the
IPO minus the year of founding.
(4.) Cooney and Kalay (1993) extend the Myers-Majluf framework by
introducing the existence of negative NPV projects. They show that an
announcement of SEO can contain favorable information about a firm and
that a positive price reaction upon the announcement of an SEO is
possible. Korajczyk, Lucas and McDonald (1991) report less of a negative
announcement effect when an SEO is conducted shortly after a favorable
earnings release.
(5.) With issuers and non-issuers matched by size, industry and
book-to-market.
(6.) 111 basis points = 41.5 basis points (new issue) + additional
69.5 basis points (issue within six months of an IPO) as shown in Table
7 Panel (7).
(7.) Analyses using the offer price<=$1 yield quantitatively
similar results.
(8.) A shelf SEO is defined as an SEO whose issue date is 60 days
after the filing date. Following Altinkilic and Hansen (2003) and Huang
and Zhang (2011), we exclude shelf registered offers.
(9.) CM rankings are obtained from Jay Ritter's website
(http://bear.warrington.ufl.edu/ ritter/ipodata.htm). For underwriters
that do not have a rank in Ritter's file, we assign rank value zero
to it.
(10.) The BHAR is over a three-year holding period following the
SEO issue. The BHAR is calculated from the first CRSP-listed post-issue
closing price to the appropriate anniversary date of the offering.
(11.) The five alternative sets of matching firms are constructed
as follows. The first set controls only for size. Each SEO firm is
matched with the non-issuing firm having the closest market
capitalization on the prior December 31. The second set controls for
size and book-to-market. We identify firms whose market value lies
between 70% and 130% of the sample firm value. Of those, we select the
firm with the closest book-to-market value. The third set controls for
size and industry effect. Each sample firm is paired with a peer firm
that has the closest market value and the same two-digit SIC code. The
fourth set controls for size and earnings-to-price effect. We identify
firms whose market value lies between 70% and 130% of the sample firm
value, and then select the firm that has the closest earnings-to-price
value. The last set controls for size, industry and book-to-market.
Following the existing literature, if a matching firm is delisted before
the three-year anniversary date of the offering, the next closest
matching firm's return is used. Up to four matching firms are kept
for each sample SEO firm. If sample firms are delisted, the BHAR is
calculated until the delisting date, and the corresponding matching
firm's return is used. The BHAR is the difference between the
holding period return for each sample firm and its matching firm.
(12.) To address the concern that firms may issue quick SEOs to
reflect their efforts to capture the hot stock markets, we control in
the regressions a market performance variable (named Mkt_ret, measured
by the NYSE/Amex value weighted cumulative return three months prior to
an SEO). We find that it is more likely and it takes shorter for firms
to go back to capital market during the period 1990-2006 (with bubble
period excluded) if the market is hot.
(13.) Literature has documented on average a -3% SEO announcement
abnormal return, followed by another -3% SEO issue day abnormal return.
Loughran and Ritter (1995) find that SEO firms underperform size and
industry matched non-issuance firms over the five years following SEOs.
(14.) We find similar results when using the measure of change of a
firm's capital expenditure ratio as the dependent variable.
(15.) In an earlier version, we matched non-issuers by industry and
pre-issue operating performance only. Our conclusion still holds.
(16.) Complete robustness results are available from the authors
upon request.
(17.) We follow Karpoff et al. (2013) and report five categories of
lockup patterns: <90, 90, 91-179, 180, and >180 days. For those
firms that issue quick SEOs, the majority specify a 180-day lockup
period, while in their SEOs, the majority have a lockup period shorter
than or equal to 90 days.
(18.) We find that 66.82% of early issuers are backed by venture
capital while 49.07% of late issuers are backed by venture capital. The
difference is significant at the 1% level. We then control for VCs, and
the interaction of VCs and secondary offering type in a probit
regression to assess the decision to conduct a quick SEO. We find that
VC dummy is not significant in the regressions, but VC backed secondary
offerings are conducted sooner than non-VC backed secondary offerings.
(19.) We follow Loughran and Ritter (2004) and Cliff and Denis
(2004) to categorize firms with the following SIC codes as tech firms:
2833, 2834, 2835, 2836, 3571, 3572, 3575, 3577, 3578, 3661, 3663, 3669,
3674, 3812, 3823, 3825, 3826, 3827, 3829, 3841, 3845, 4812, 4813, 4899,
7370, 7371, 7372, 7373, 7374, 7375, 7377, 7378, and 7379.
(20.) Regression results indicate: (1) high-tech firms do not
necessarily return to the equity issuance market earlier than others
(the coefficient estimate is not statistically significant) after
controlling for other firm characteristics; (2) high-tech firms do not
experience greater negative SEO 3-day announcement abnormal returns (the
coefficient estimate is not statistically significant); (3) high-tech
firms do not exhibit more negative long-run returns (the coefficient
estimate is not statistically significant). We also perform the
univariate analyses as shown in Table 5, with separate analyses for
non-high-tech firms and high-tech firms by early issue and late issue.
For both high tech and non-high-tech firms, early issuers demonstrate
more negative BHAR than late issuers. However, there is no significant
difference in BHAR between non-high-tech firms and high-tech firms
categorized by either early issuers or late issuers. Regression results
are available from the authors upon request.
(21.) First, following DeAngelo, DeAngelo and Stulz (2010), we
measure a company's near-term cash needs as: Pro Forma Cash/TA
ratio = (Cash t+1- SEO proceeds from primary shares)/(Total Assets t+1 -
SEO proceeds from primary shares), and add the variable to the probit
regression. As predicted, this variable carries no explanatory power
regarding early issue or late issue. The market-timing hypothesis still
holds after controlling for a company's cash saving needs. Second,
we consider the counterfactual condition that had there been no SEO
issue, would a firm run out of cash. We find that for early issuers,
about 68.4% of firms would have run out of cash without the issue, and
that 67.7% of the late issuers would have run out of cash. The
difference is statistically insignificant.
REFERENCES
Altinkilic, O., and R. Hansen, 2003, "Discounting and
Underpricing In Seasoned Equity Offers," Journal of Financial
Economics 69, 285-324.
Asquith, P., D.W. Mullins Jr., 1986, "Equity Issues and
Offering Dilution," Journal of Financial Economics 15, 61-89.
Barber, B.M., and J.D. Lyon, 1996, "Detecting Long-run
Abnormal Stock Returns: The Empirical Power and Specification of
Test-Statistics," Journal of Financial Economics 43, 341-372.
Bartov E, P. Mohanram, and C. Seethamraju, 2002, "Valuation of
Internet Stocks-An IPO Perspective," Journal of Accounting Research
40(2): 321-346.
Bouwman, Christa H.S., Kathleen Fuller, and Amrita S. Nain, 2009,
"Market Valuation and Acquisition Quality: Empirical
Evidence," Review of Financial Studies 22 (2) 2009: 634-679.
Bradley, Daniel J., John W. Cooney, Jr., Bradford D. Jordan, and
Ajai K. Singh, 2004, "Negotiation and the IPO Offer Price: A
Comparison of Integer vs. Non-Integer IPOs," Journal of Financial
Quantitative Analysis 39(3): 517-540.
Brav, Alon, and Paul A. Gompers, 1997, "Myth or Reality? The
Long-Run Underperformance of Initial Public Offerings: Evidence from
Venture and NonVenture Capital-Backed Companies," Journal of
Finance 52, 1791-1821.
Billett, M., M. Flannery, and J. Garfinkel, 2006, "Are Bank
Loans Special? Evidence on the Post Announcement Performance of Bank
Borrows," Journal of Financial and Quantitative Analysis 41, 4,
733-751.
Carter, R.B., and S. Manaster, 1990, "Initial Public Offerings
and Underwriter Reputation," Journal of Finance, 45, 1045-1068.
Carter, R.B., F. Dark, A. Singh, 1998, "Underwriter
Reputation, Initial Returns, and the Long-Run Performance of IPO
Stocks," Journal of Finance 53, 285-311.
Chen, H., S. Chen, and C. Huang, 2012, "Why Do Insiders Sell
Shares Following IPO Lockups?" Financial Management 41, 813-847.
Chemmanur, T.J., 1993, "The Pricing of Initial Public
Offerings: A Dynamic Model with Information Production," Journal of
Finance 48, 285-304.
Cliff, M., D.J. Denis, 2004, "Do IPO Firms Purchase Analyst
Coverage With Underpricing?" Journal of Finance 59, 2871-2901.
Cooney Jr., J.W., and A. Kalay, 1993, "Positive Information
from Equity Issue Announcements," Journal of Financial Economics
33, 149-172.
DeAngelo, H., L. DeAngelo, and R. Stulz, 2010, "Seasoned
Equity Offerings, Market Timing, and the Corporate Life-Cycle,"
Journal of Financial Economics 95, 275-295.
Eckbo, E., and R. Masulis, 1995, "Seasoned Equity Offerings: A
Survey," in: Jarrow, R., V. Maksimovic, and W., Ziemba, eds.,
Handbooks in operations research and management science, (Elsevier,
Amsterdam) Chapter 31.
Fama, E., and K. French, 1993, "Common Risk Factors in the
Returns of Stocks and Bonds," Journal of Financial Economics 33,
3-55.
Field, L., and G. Hanka, 2002, "The Expiration of IPO Share
Lockups," Journal of Finance 56, 471-500.
Houston, J., and M. Ryngaert, 1997, "Equity Issuance and
Adverse Selection: A Direct Test Using Conditional Stock Offers,"
Journal of Finance 52, 197-219.
Hovakimian, A., and I. Hutton, 2010, "Market Feedback and
Equity Issuance: Evidence from Repeat Equity Issues," Journal of
Financial and Quantitative Analysis 45, 3, 739-762.
Huang, R., and D. Zhang, 2011, "Managing Underwriters and the
Marketing of Seasoned Equity Offerings," Journal of Financial and
Quantitative Analysis 46, 141-170.
Jegadeesh, N., M. Weinstein, and I., Welch, 1993, "An
Empirical Investigation of IPO Return and Subsequent Equity
Offerings," Journal of Financial Economics 34, 153-175
Karpoff, J., G. Lee, and R. Masulis, 2013, "Contracting under
Asymmetric Information: Lockup Agreements in Seasoned Equity
Offerings," Journal of Financial Economics, forthcoming
Kothari, S.P., and J.B. Warner, 1997, "Measuring Long-Horizon
Security Price Performance," Journal of Financial Economics 43,
301-339.
Korajczyk, R.A., D.J. Lucas, and R.L. McDonald, 1991, "The
Effect of Information Releases on the Pricing and Timing of Equity
Issues," Review of Financial Studies 4, 685-708.
Krigman, L., W.H. Shaw, and K.L. Womack, 2001, "Why Do Firms
Switch Underwriters?" Journal of Financial Economics 60, 245-284.
Loughran, T., and J. Ritter, 1995, "The New Issues
Puzzle," Journal of Finance 50, 2351.
Loughran T., and J. Ritter, 2004, "Why Has IPO Underpricing
Changed Over Time?" Financial Management 33, 5-37.
Lyon, J.D., B.M. Barber, and C.L. Tsai, 1999, "Improved
Methods for Tests of LongRun Abnormal Stock Returns," Journal of
Finance 54, 165-201.
McLean, D., 2011, "Share Issuance and Cash Savings,"
Journal of Financial Economics 99 (3), 693-715.
Mola, S., and T. Loughran, 2004, "Discounting and Clustering
In Seasoned Equity Offering Prices," Journal of Financial and
Quantitative Analysis 39, 1-23.
Myers, S.C., and N.S. Majluf, 1984, "Corporate Financing and
Investment Decisions When Firms Have Information That Investors Do Not
Have," Journal of Financial Economics 13, 187-221.
Vijh, A., 1999, "Long-Term Returns from Equity
Carveouts," Journal of Financial Economics, 51: 273-308.
Welch, I., 1989, "Seasoned Offerings, Imitation Costs, and the
Underpricing of Initial Public Offerings," Journal of Finance 44,
421-449.
Yi Jiang (a) *, Mark Stohs (b), Xiaoying Xie (c)
(a) Department of Finance, Mihaylo College of Business and
Economics California State University, Fullerton yjiang@fullerton.edu
(b) Department of Finance, Mihaylo College of Business and
Economics California State University, Fullerton mstohs@fullerton.edu
(c) Department of Finance, Mihaylo College of Business and
Economics California State University, Fullerton xxie@fullerton.edu
* We thank Matt Billett, Philip Davis, Jon Garfinkel, Erik Lie,
Yiming Qian, Jay Ritter, Ashish Tiwari, and Anand Vijh for helpful
comments and suggestions. All errors are our own.
Appendix
Definitions of variables (COMPUSTAT items)
Phrase Used in Text Acronym Definition
Total Assets Total Assets
Operating Income OIBD
Before Depreciation
Number of Shares Shrs Out
Outstanding
Book Value of Equity BV EQ
Capital Expenditures Cap Exp
Share Price Price
Market Value of MV EQ Share Price x Shares
Equity Outstanding
Book-to-market B/M BV EQ / MV EQ
Tobin's Q Total Market Value of assets /
Total book value of Assets
Return on Assets ROA OIBD / Total Assets
Cap Exp Ratio Cap Exp - Total Assets
Free Cash Flow FCF Net Income After Tax +
Depreciation +
Amortization - Dividends -
Preferred Dividends
IPO SIZE Amount of Equity Capital
Raised by IPO
SEO SIZE Amount of Equity Capital
Raised by SeO
SEO/IPO SEO SIZE / IPO SIZE
SEO/MV EQ SEO SIZE / MV EQ
AT (days) [DELTA]T Number of calendar days
between IPO and first SEO.
Underpricing IPO Under IPO (1st post-issue price - IPO
offer price) / IPO offer price
Abnormal Return IPO AB RET 20 IPO abnormal return from
(1 - 20) trading day 1 to trading day
20 after IPO date.
Abnormal Return IPO AB RET 40 IPO abnormal return from
(21 - 40) trading day 21 to trading day
40 after IPO date.
SEO AR Equation 1 SEO Abnormal Return = SEO
3-day announcement-period
return; from event day -1 to
+1, where day 0 = filing date.
Phrase Used in Text COMPUTSAT XPF NAME
Total Assets AT
Operating Income OIBDP
Before Depreciation
Number of Shares CSHO
Outstanding
Book Value of Equity CEQ
Capital Expenditures CAPX
Share Price PRCC_F
Market Value of PRCC_F x CSHO
Equity
Book-to-market
Tobin's Q (AT-CEQ+MV EQ) / AT
Return on Assets
Cap Exp Ratio CAPX - AT
Free Cash Flow NI + DP - DVC - DVP
IPO SIZE
SEO SIZE
SEO/IPO
SEO/MV EQ
AT (days)
Underpricing IPO
Abnormal Return IPO
(1 - 20)
Abnormal Return IPO
(21 - 40)
SEO AR
Table 1
Descriptive statistics for sample SEO firms (first time SEOs)
Description Mean Median Min.
Market value ($M) 694.49 277.73 10.02
Total Assets ($M) 405.55 162.18 5.62
Book-to-market (B/M) 0.47 0.38 -0.06
Tobin's Q 2.58 1.91 0.66
ROA 0.06 0.12 -1.00
CAP EXP RATIO 0.09 0.06 0.00
FCF ($M) -0.03 0.13 -5.69
IPO SIZE ($M) 53.61 32.90 1.60
SEO SIZE ($M) 82.45 50.70 0.70
SEO/IPO 2.03 1.54 0.23
SEO/MV EQ 0.28 0.18 0.02
AT (days) 898.67 496.00 64.00
UNDER IPO 21.65% 8.93% -22.79%
AB RET 20 6.44% 3.79% -95.18%
AB RET 40 3.94% 2.43% -72.05%
SEO AR -3.46% -3.40% -40.00%
AGE 16.38 10.00 1.00
SECOND 0.34 0.00 0.00
IPORANK 2.98 0.00 0.00
SEORANK 3.26 0.00 0.00
SWITCHBETTER 0.18 0.00 0.00
INTEGER_IPO 0.79 1.00 0.00
INTEGER_SEO 0.41 0.00 0.00
Description Max Std. Dev. N
Market value ($M) 40098.59 2101.12 1610
Total Assets ($M) 22384.00 1040.07 1610
Book-to-market (B/M) 2.11 0.38 1610
Tobin's Q 11.08 1.89 1610
ROA 0.43 0.24 1598
CAP EXP RATIO 0.47 0.09 1589
FCF ($M) 1.42 0.78 1610
IPO SIZE ($M) 2745.50 96.16 1610
SEO SIZE ($M) 1292.20 110.39 1610
SEO/IPO 11.72 1.80 1610
SEO/MV EQ 2.42 0.36 1610
AT (days) 9290.00 1051.48 1610
UNDER IPO 458.41% 41.79% 1608
AB RET 20 176.09% 21.64% 1610
AB RET 40 119.10% 18.45% 1610
SEO AR 49.39% 7.48% 1610
AGE 166.00 19.45 1591
SECOND 1.00 0.47 1610
IPORANK 9.00 3.73 1610
SEORANK 9.00 3.85 1610
SWITCHBETTER 1.00 0.38 1610
INTEGER_IPO 1.00 0.41 1610
INTEGER_SEO 1.00 0.49 1610
Note: The sample consists of all 1,610 firms listed on NASDAQ,
AMEX, or NYSE that conducted both IPO and (first time) SEO during
calendar years from 1970-2006, after applying our sample screening
criteria. Market value is price multiplied by the number of shares
outstanding. B/M is the ratio of book value of equity to market
value of equity. Tobin's Q is the ratio of total market value of
assets to total book value of assets. ROA is the OIBD (Operating
Income before Depreciation) normalized by total assets. CAP EXP
RATIO is capital expenditure to total assets. FCF is the free cash
flow, defined as net income after tax plus depreciation less common
and preferred dividends, deflated by the firm's beginning-of-year
capital. IPO SIZE is the amount of capital raised in the IPO. SEO
SIZE is the amount of capital raised in the first SEO. SEO/IPO is
SEO size as a fraction of capital raised in the IPO. SEO/MV EQ is
SEO size as a fraction of market value of equity. AT is the number
of calendar days between IPO and the first SEO. UNDER IPO is IPO
underpricing, defined as the difference between the first
post-issue price and the IPO offer price divided by the offer
price. AB RET 20 is the abnormal return over the period from
trading day 1 to trading day 20 after the IPO date. AB RET 40 is
the abnormal return over the period from trading day 21 to trading
day 40 after the IPO date. SEO AR, the SEO 3-day announcement
period abnormal return, is calculated using market model over the
event days -1, 0 and +1, where day 0 is the filing date. AGE is the
number of years since the founding date of the firm to the year
issuing the SEO. SECOND is a dummy variable equal to one if the
percentage of secondary shares offered in an SEO is greater than 50
percent of total offerings. IPORANK is the underwriter reputation
ranking at the time of the IPO. SEORANK is the underwriter
reputation ranking at the time of the
Table 2
Firm characteristics and quick SEO
Ln [DELTA]T
Full Sample Subsample 1 Subsample 2
(1970- (1990-2006) (exclude
2006) bubble period)
UNDER IPO -0.267 *** -0.276 *** -0.503 ***
[0.062] [0.062] [0.120]
Ln IPO -0.214 *** -0.126 *** -0.184 ***
SIZE [0.039] [0.044] [0.047]
AB RET 20 -0.822 *** -0.745 *** -0.721 ***
[0.099] [0.103] [0.139]
AB RET 40 -0.544 *** -0.491 *** -0.813 ***
[0.110] [0.113] [0.138]
Tobin's Q 0.003 0.002 0.002
[0.008] [0.008] [0.008]
CAP EXP -0.674 *** -0.856 *** -0.671 **
RATIO [0.231] [0.266] [0.265]
FCF -0.013 0.003 -0.047
[0.016] [0.017] [0.034]
Ln(Total 0.173 *** 0.122 *** 0.184 ***
Assets) [0.027] [0.030] [0.032]
ROA -0.123 -0.101 -0.009
[0.111] [0.115] [0.138]
AGE 0.007 *** 0.006 *** 0.005 ***
[0.001] [0.001] [0.001]
SECOND -0.294 *** -0.347 *** -0.272 ***
[0.049] [0.054] [0.058]
IPORANK 0.052 *** 0.057 *** 0.060 ***
[0.009] [0.010] [0.010]
SEORANK -0.085 *** -0.091 *** -0.096 ***
[0.010] [0.011] [0.011]
SWITCHBE 0.703 *** 0.737 *** 0.724 ***
TTER [0.079] [0.090] [0.091]
INTEGER_I -0.048 -0.121 ** -0.107 *
PO [0.050] [0.059] [0.058]
Mkt ret -0.418 -0.298 -0.961 **
[0.375] [0.458] [0.482]
Intercept 7.074 *** 5.604 *** 5.489 ***
[0.969] [0.659] [0.641]
Industry and Not reported Not reported Not reported
year dummies
Sample 1,532 1,169 1,007
size
Adjusted 0.36 0.294 0.314
[R.sup.2]
p-value of
regression
Likelihood of a Quick SEO
Subsample 1 Subsample 2
Full Sampl (1990-2006) (exclude
(1970-2006) bubble period)
UNDER IPO 0.447 *** 0.390 *** 0.902 ***
[0.123] [0.124] [0.264]
Ln IPO 0.213 ** 0.191 * 0.304 **
SIZE [0.099] [0.108] [0.129]
AB RET 20 1.483 *** 1.427 *** 1.403 ***
[0.219] [0.225] [0.336]
AB RET 40 1.033 *** 0.947 *** 1.842 ***
[0.242] [0.248] [0.336]
Tobin's Q -0.011 -0.012 -0.022
[0.017] [0.017] [0.022]
CAP EXP 1.060 * 0.97 0.721
RATIO [0.546] [0.631] [0.683]
FCF 0.032 0.025 0.131
[0.039] [0.039] [0.132]
Ln(Total -0.081 -0.076 -0.181 **
Assets) [0.066] [0.070] [0.084]
ROA -0.375 -0.39 -0.472
[0.271] [0.280] [0.398]
AGE -0.010 *** -0.012 *** -0.011 ***
[0.003] [0.004] [0.004]
SECOND 0.600 *** 0.582 *** 0.462 ***
[0.115] [0.123] [0.144]
IPORANK -0.066 ** -0.060 ** -0.089 ***
[0.027] [0.029] [0.034]
SEORANK 0.084 *** 0.082 *** 0.103 ***
[0.029] [0.030] [0.036]
SWITCHBE -1.015 *** -0.918 *** -0.913 ***
TTER [0.254] [0.264] [0.291]
INTEGER_I 0.117 0.221 0.271
PO [0.139] [0.159] [0.175]
Mkt ret 1.208 0.696 2.338*
[0.946] [1117] [1.292]
Intercept -6.301 -6.194 -6.377
[135.642] [119.757] [123.586]
Industry and Not reported Not reported Not reported
year dummies
Sample 1,532 1,169 1,007
size
Adjusted
[R.sup.2]
p-value of 0.000 0.000 0.000
regression
Note: This table reports (1) the cross-sectional regression of the
logarithm of time between IPO and the first SEO. The dependent
variable is the logarithm of the time between the IPO and the first
SEO (Ln [DELTA]T), and (2) probit regression of the factors leading
to a quick SEO after IPO. The dependent variable is a dummy
variable with value equal to one when an SEO is issued within six
months of IPO and zero otherwise. The independent variables
include: UNDER IPO is IPO underpricing, defined as the difference
between the first post-issue price and the IPO offer price, divided
by the offer price. Ln IPO SIZE is the logarithm of IPO size (the
amount of equity capital raised in the IPO). AB RET 20 is the
abnormal return over the period from trading day 1 to trading day
20 after the IPO date. AB RET 40 is the abnormal return over the
period from trading day 21 to trading day 40 after the IPO date.
Tobin's Q is the ratio of total market value of assets to total
book value of assets. CAP EXP RATIO is capital expenditure scaled
by total assets. FCF is the free cash flow, defined as net income
after tax plus depreciation less common and preferred dividends,
deflated by the firm's beginning-of-year capital. Ln (Total Assets)
is the logarithm of the total assets. ROA is the OIBD (operating
income before depreciation) normalized by total assets. AGE is the
number of years since the founding date of the firm to the year
issuing an SEO. SECOND is a dummy variable equal to one if the
percentage of secondary shares offered in an SEO is greater than 50
percent of total offerings. IPORANK is the underwriter reputation
ranking at the time of the IPO. SEORANK is the underwriter
reputation ranking at the time of the SEO. SWITCHBETTER is a dummy
variable that equals one if the SEO underwriter is ranked higher
than the IPO underwriter. INTEGER_IPO is a dummy variable equal to
one if IPO offer price is an integer. Mkt_ret is NYSE/Amex value
weighted cumulative return in the prior three months before the
SEO. The independent variables also include dummy variables for
industry and the year of SEO. Standard errors are listed in
brackets. *, **, and *** denote significance level at the 10%, 5%,
and 1% levels.
Table 3
First time SEO three-day announcement abnormal returns and stock
price run-ups
Panel A: Abnormal returns ordered by length of time since IPO at
the date of first SEO
Time (t) between N SEO AB RET AB RET
IPO and SEO AR (%) 20 (%) 40 (%)
(1) t < 6 months 214 -5.79 22.29 11.39
(2) 6 months [less than or 428 -3.67 8.38 5.91
equal to] t < 1 year
(3) 1 year [less than or equal 379 -2.59 2.59 2.40
to] t < 2 years
(4) 2 years [less than or equal 185 -2.63 2.37 0.82
to] t < 3 years
(5) 3 years [less than or equal 115 -2.82 1.26 0.06
to] t < 4 years
(6) 4 years [less than or equal 82 -3.97 1.75 0.63
to] t < 5 years
(7) t [greater than or equal 207 -3.10 1.44 1.25
to] 5 years
Sample Size/Averages 1,610 -3.46 6.44 3.94
Panel B: Difference tests
< 6 months [greater thann
or equal to]6
months
Mean Median Mean Median
SEO AR (%) -5.79 -5.34 -3.10 -3.17
AB RET 20 22.29 14.63 4.01 2.43
AB RET 40 11.39 11.80 2.80 1.74
Difference Tests [p - value]
T- test Median Test
SEO AR (%) [0.00] *** [<0.001] ***
AB RET 20 [0.00] *** [<0.0001] ***
AB RET 40 [0.00] *** [<0.0001] ***
Note: The SEO AR (3-day announcement period abnormal return) is
calculated using standard market model over the event days -1, 0
and +1, where day 0 is the filing date. AB RET 20 is the abnormal
return over the period from trading day 1 to trading day 20 after
the IPO date. AB RET 40 is the abnormal return over the period from
trading day 21 to trading day 40 after the IPO date. They are
calculated by subtracting the market index from the returns at time
t. p-values are in the brackets. *, **, and *** denote significance
level at the 10%, 5%, and 1% levels.
Table 4
Regression of the first time SEO three-day
announcement period abnormal returns
6 Months Dummy
Full Sample Subsample 1 Subsample 2
(1970- (1990-2006) (exclude
2006) bubble period)
UNDER IPO 0.348 -0.092 1.33
[0.633] [0.687] [1.054]
Ln 0.061 0.171 0.183
(SEO/IPO) [0.217] [0.269] [0.284]
AB RET 20 0.451 -0.183 -0.186
[0.962] [1.111] [1.384]
AB RET 40 -0.284 -1.022 -1.231
[1.064] [1211] [1.352]
6 months -1.482 ** -1.08 -1.382 *
dummy or
Ln [DELTA]T [0.664] [0.735] [0.785]
AB RET 20 x -5.823 ** -5.762 ** -2.666
6 months
(Dummy) [2.758] [2.881] [3.733]
AB RET 40 x 3.838 2.227 2.27
6 months
(Dummy) [3.127] [3.373] [3.734]
Tobin's Q 0.022 -0.035 -0.065
[0.070] [0.073] [0.077]
ROA 1.147 1.331 1.491
[0.965] [1.085] [1.245]
CAP EXP 1.576 1.155 1.158
RATIO [1.587] [1.898] [1.923]
Ln(Total 0.502 *** 0.511 *** 0.424 **
Assets) [0.146] [0.172] [0.180]
FCF -0.08 -0.09 0.125
[0.165] [0.185] [0.339]
AGE -0.001 -0.005 -0.003
[0.007] [0.008] [0.008]
SECOND -0.787 ** -1.054 *** -1.182 ***
[0.326] [0.390] [0.405]
SEORANK -0.001 0.006 0.027
[0.045] [0.054] [0.056]
SWITCHBE 0.397 0.513 0.281
TTER [0.436] [0.557] [0.572]
INTEGER S -0.227 -0.361 -0.303
EO [0.296] [0.348] [0.359]
Mkt ret 3.668 4.505 4.858
[2.740] [3.581] [3.700]
Intercept -1.313 -4.698 -4.246
[6.537] [3.580] [3.605]
Industry and Not reported Not reported Not reported
year dummies
Sample size 1,532 1,169 1,007
Adjusted 0.034 0.042 0.035
[R.sup.2]
Ln [DELTA]T
Full Sample Subsample 1 Subsample 2
(1970- (1990-2006) (exclude
2006) bubble period)
UNDER IPO 0.061 -0.215 1.309
[0.629] [0.685] [1.054]
Ln -0.004 0.118 0.159
(SEO/IPO) [0.220] [0.270] [0.284]
AB RET 20 -0.306 -1.168 -0.562
[0.909] [1.023] [1.292]
AB RET 40 0.219 -0.637 -0.941
[1.008] [1.132] [1.261]
6 months 0.400 ** 0.535 ** 0.528 **
dummy or
Ln [DELTA]T [0.177] [0.230] [0.245]
AB RET 20 x
6 months
(Dummy)
AB RET 40 x
6 months
(Dummy)
Tobin's Q 0.034 -0.026 -0.061
[0.070] [0.073] [0.077]
ROA 1.101 1.399 1.39
[0.970] [1.085] [1.246]
CAP EXP 1.456 1.089 1.081
RATIO [1.592] [1.901] [1.921]
Ln(Total 0.490 *** 0.484 *** 0.411 **
Assets) [0.146] [0.172] [0.179]
FCF -0.077 -0.095 0.122
[0.166] [0.186] [0.339]
AGE -0.001 -0.005 -0.003
[0.007] [0.008] [0.008]
SECOND -0.786 ** -0.939 ** -1.054 **
[0.332] [0.405] [0.418]
SEORANK 0.006 0.013 0.033
[0.045] [0.054] [0.056]
SWITCHBE 0.323 0.48 0.225
TTER [0.443] [0.560] [0.574]
INTEGER S -0.27 -0.419 -0.356
EO [0.297] [0.348] [0.359]
Mkt ret 3.747 4.554 4.753
[2.749] [3.586] [3.698]
Intercept -3.764 -7.917 ** -7.462 *
[6.671] [3.909] [3.955]
Industry and Not reported Not reported Not reported
year dummies
Sample size 1,532 1,169 1,007
Adjusted 0.027 0.039 0.035
[R.sup.2]
Note: The dependent variable is the SEO 3-day announcement period
abnormal returns in percentages (SEO AR). The SEO 3-day
announcement period return is calculated over the event days -1, 0,
and +1, where day 0 is the filing date. 6 months dummy (SEO within
six months of IPO) is a dummy variable that takes on the value of 1
if the number of calendar days between IPO and the first SEO is
less than six months. Ln [DELTA]T is the logarithm of the number of
calendar days between IPO and the first SEO. UNDER IPO is IPO
underpricing. Ln (SEO/IPO) is the logarithm of the relative size of
the SEO and IPO. AB RET 20 is the abnormal return over the period
from trading day 1 to trading day 20 after the IPO date. AB RET 40
is the abnormal return over the period from trading day 21 to
trading day 40 after the IPO date. Tobin's Q is the ratio of total
market value of assets to total book value of assets. ROA is the
OIBD (operating income before depreciation) scaled by total assets.
CAP EXP RATIO is capital expenditure scaled by total assets. Ln
(Total Assets) is the logarithm of the total assets. FCF is the
free cash flow, defined as net income after tax plus depreciation
less common and preferred dividends, deflated by the firm's
beginning-of-year capital. AGE is the number of years since the
founding date of the firm to the year issuing an SEO. SECOND is a
dummy variable equal to one if the percentage of secondary shares
offered in an SEO is greater than 50 percent of total offerings.
SEORANK is the underwriter reputation ranking at the time of the
SEO. SWITCHBETTER is a dummy variable that equals one if the SEO
underwriter is ranked higher than the IPO underwriter. INTEGER_SEO
is a dummy variable equal to one if SEO offer price is an integer.
Mkt ret is NYSE/Amex value weighted cumulative return in the prior
three months before the SEO. The independent variables also include
dummy variables for industry and the year of SEO. Standard errors
are listed in brackets. *, **, and *** denote significance level at
the 10%, 5%, and 1% levels.
Table 5
The long-run performance of SEOs by length of time between
IPO and first time SEO
Panel A: BHAR
3-Year Mean Buy-and-Hold
Abnormal Returns %
Time (t) between Size Size and Size and
IPO and SEO alone book-to- SIC
market
(1) t < 6 months -51.80 -68.98 -37.89
(2) 6 months [greater -16.87 -32.27 -9.45
than or equal to] t
< 1 year
(3) 1 year [greater -20.04 -22.72 -19.34
than or equal to] t
< 2 years
(4) 2 years [greater -4.54 -23.55 -12.99
than or equal to] t
< 3 years
(5) 3 years [greater -17.39 -32.61 -25.17
than or equal to] t
< 4 years
(6) 4 years [greater -4.05 -3.90 4.79
than or equal to] t
< 5 years
(7) t [less than or -9.65 -9.87 -9.19
equal to] 5 years
(8) All SEOs -19.27 -29.53 -16.31
(1970-2006)
3-Year Mean Buy-and-Hold
Abnormal Returns %
Time (t) between Size and Size, SIC and
IPO and SEO earnings-to- book-to-market
price ratio
(1) t < 6 months -22.53 -59.97
(2) 6 months [greater -24.22 -24.81
than or equal to] t
< 1 year
(3) 1 year [greater -22.68 -11.85
than or equal to] t
< 2 years
(4) 2 years [greater -16.55 -23.86
than or equal to] t
< 3 years
(5) 3 years [greater -14.34 -25.42
than or equal to] t
< 4 years
(6) 4 years [greater -6.32 -19.01
than or equal to] t
< 5 years
(7) t [less than or -18.90 -2.44
equal to] 5 years
(8) All SEOs -20.46 -23.13
(1970-2006)
Panel B: Difference tests on BHAR
3-Year Mean BHR (%)
Time (t) between
IPO and SEO N SEOs Matching BHAR (%)
firms
(1) t < 6 months 208 -12.09 47.88 -59.97
(2) t [greater than 1370 21.48 39.01 -17.54
or equal to] 6 months
Difference Tests [0.035] **
(p-value)
Note: The sample of first seasoned equity offering during 1970 to
2006 is categorized by the length of time since IPO at the date of
SEO. BHAR is the abnormal return defined as the difference between
a sample firm's BHR and its matching firm's BHR. We use five sets
of matching firms. The first set controls for size. The second
controls for size and book-to-market. The third set controls for
size and industry effect. The fourth set controls for size and
earnings-to-price effect, and the fifth set controls for size,
industry, and book-to-market. Panel A reports BHAR using the five
alternative matching procedures. Panel B presents the difference
tests on BHAR (matched by size, SIC and B/M). p-values of t-tests
are in brackets. *, **, and *** denote significance level at the
10%, 5%, and 1% levels.
Table 6
Regression analysis of three-year buy-and-hold
abnormal return (BHAR) of first Time SEOs
6 Months Dummy
Full Sample Subsample 1 Subsample 2
(1970- (1990-2006) (exclude
2006) bubble period)
UNDER 1.852 2.244 0.131
IPO [14.006] [14.969] [33.925]
Ln -17.846 ** -18.864 ** -19.619 *
(SEO/IPO) [7.244] [8.743] [10.348]
AB RET 20 -2.931 -14.456 -23.434
[23.946] [26.430] [39.920]
AB RET 40 -3.277 -9.475 -14.131
[26.706] [29.397] [39.619]
6 months dummy -28.639 * -33.340* -37.349*
or Ln [DELTA]T [15.868] [17.293] [20.537]
Tobin's Q 6.397 *** 6.446 *** 7.068 ***
[1.889] [2.003] [2.354]
ROA 72.387 *** 63.569 ** 74.683 **
[25.731] [28.805] [37.313]
CAP EXP -68.902 -60.917 -67.597
RATIO [56.148] [69.109] [75.494]
Ln(Total 7.684 10.701 * 10.498
Assets) [5.260] [6.233] [7.190]
FCF 7.253* 3.214 1.831
[3.890] [4.474] [9.587]
AGE -0.322 -0.365 -0.308
[0.272] [0.317] [0.352]
SECOND -18.722 -16.546 -15.306
[12.692] [14.320] [16.709]
Intercept 0.648 -182.422 -193.515
[231.918] [145.161] [155.242]
Industry and year
dummies Not Not Not
reported reported reported
Sample size 1,532 1,169 1,007
Adjusted [R.sup.2] 0.047 0.0188 0.0127
Ln [DELTA]T
Full Sample Subsample 1 Subsample 2
(1970- (1990-2006) (exclude
2006) bubble period)
UNDER 2.983 3.709 4.547
IPO [13.955] [14.992] [33.993]
Ln -20.580 *** -20.716 ** -21.688 **
(SEO/IPO) [7.328] [8.768] [10.391]
AB RET 20 1.33 -13.143 -19.48
[23.951] [26.305] [39.949]
AB RET 40 0.618 -7.801 -10.717
[26.751] [29.384] [39.562]
6 months dummy 14.886** 17.907 ** 21.409 **
or Ln [DELTA]T [5.837] [7.619] [9.071]
Tobin's Q 6.647 *** 6.567 *** 7.169 ***
[1.889] [2.001] [2.351]
ROA 69.922 *** 64.356 ** 73.030 *
[25.738] [28.763] [37.286]
CAP EXP -70.761 -60.629 -68.863
RATIO [56.065] [69.050] [75.402]
Ln(Total 8.831 * 11.550 * 11.472
Assets) [5.287] [6.251] [7.201]
FCF 7.318 * 3.074 1.967
[3.886] [4.471] [9.576]
AGE -0.379 -0.403 -0.357
[0.273] [0.318] [0.353]
SECOND -15.539 -13.368 -11.497
[12.800] [14.492] [16.875]
Intercept -99.669 -300.727 * -334.522 **
[235.526] [154.752] [167.582]
Industry and year
dummies Not Not Not
reported reported reported
Sample size 1,532 1,169 1,007
Adjusted [R.sup.2] 0.0491 0.0203 0.015
Note: The dependent variable is the BHAR of SEOs in percentages,
computed as the difference between the BHRs of sample firms and the
matching firms selected by size, industry, and book-to-market over
a three- year holding period. 6 months dummy (SEO within six months
of IPO) is a dummy variable that equals 1 if the number of calendar
days between IPO and the first SEO is less than six months. Ln
[DELTA]T is the logarithm of the number of calendar days between
IPO and the first SEO. UNDER IPO is IPO underpricing, defined as
the difference between the first post-issue price and the IPO offer
price divided by the offer price. Ln (SEO/IPO)
Table 7
Monthly cross-sectional regressions
(1) (2) (3) (4)
Intercept 1.105 *** 1.460 *** 1.355 *** 1.46 ***
[2.69] [5.06] [4.45] [5.06]
Ln MV EQ 0.091 **
[198]
Ln B/M 0.25 ***
[5.25]
ISSUE -0.470 *** -0.466 ***
[-3.84] [-3.86]
ISSUE 6 Month -0.968 *** -0.649 *
[-2.62] [-195]
Avg [R.sup.2] 0.0 19 0.0 03 0.0001 0.004
# months 444 444 444 444
(5) (6) (7)
Intercept 1.171 *** 1.104 *** 1.171 ***
[2.94] [2.69] [2.94]
Ln MV EQ 0.097 ** 0.091 ** 0.097 **
[2.12] [199] [2.13]
Ln B/M 0.241 *** 0.25 *** 0.241 ***
[5.15] [5.25] [5.15]
ISSUE -0.42 *** -0.415 ***
[-3.7] [-3.72]
ISSUE 6 Month -0.974 *** -0.695 **
[-2.64] [-2.08]
Avg [R.sup.2] 0.022 0.019 0.022
# months 444 444 444
Note: The sample consists of all firms listed on NASDAQ, AMEX, or
NYSE during 1970-2006. Ln MV EQ is the logarithm of the market
value of equity. Ln B/M is the logarithm of B/M, using the book
value of equity for the most recent fiscal year end. ISSUE is a
dummy variable that takes the value of 1 if a company conducted at
least one public equity offering (SEO or IPO) within the 60 months
preceding a given June 30th. ISSUE 6 Month is dummy variable that
equals 1 if a company conducted SEO within six months of its IPO.
The dependent variable is the firm's monthly percentage stock
return. T-statistics are listed in brackets. *, **, and *** denote
significance levels at the 10%, 5%, and 1% levels. Model (2)
[r.sub.it] = a + bln[MV.sub.it] + clnB/[M.sub.it] + d[ISSUE.sub.it]
+ eISSUE[6month.sub.it] + [[epsilon].sub.it]
Table 8
Monthly alphas using Fama-French three-factor model
Panel A: Issuers vs. Non-issuers
Issuers Non-issuers Difference
All
firms -0.27 0.21 -0.50
[-1.66] * [2.03] * [-4.94] ***
Large
firms 0.17 0.29 -0.13
[1.01] [4.18] *** [-102]
Small firms -0.70 0.13 -0.86
[-3.62] *** [0.91] [-7.46] ***
Panel B: Early Issuers vs. Late issuers
< 6 months [greater than Difference
All or equal to]
6 months
issuers -0.89 0.12 -1.09
[-1.97] ** [1.06] [-2.62] ***
Large
issuers -0.13 0.28 -0.41
[-0.26] [3.02] *** [-0.86]
Small issuers -0.56 -0.01 -0.73
[-0.63] [-0.08] [-0.84]
Note: The sample consists of all firms listed on NASDAQ, AMEX,
or NYSE during 1970-2006. Large firms are those whose market
capitalization on June 30 of year t is greater than the market
capitalization of the median company in the sample. Small firms
are those whose market capitalization is below the median.
The monthly data for the market, size, and book-to-market
factor returns are obtained from French's website. Panel A
reports regression alphas for portfolios of issuers and
non-issuers and the difference in alphas. Panel B reports
regression alphas for portfolios of issuers conducting SEOs
within six months of IPOs (early issuers), issuers conducting
SEOs after six months of IPOs (late issuers), and the difference
in alphas. T-statistics are listed in brackets. *, **, and
*** denote significance level at the 10%, 5%, and 1% levels.
Model (3) ([R.sub.pt] - [R.sub.ft]) = a + b([R.sub.mt] - [R.sub.ft]) +
s[SMB.sub.t] + h[HML.sub.t] + [[epsilon].sub.t]
Table 9
Aftermarket returns and investments
6 Months Dummy
Full Sample Subsample 1 Subsample 2
(1970- (1990- (exclude
2006) 2006) bubble pd.)
AB RET 20 -7.863 *** -8.103 *** -9.955 **
[2.450] [2.771] [4.269]
AB RET 40 1.694 1.284 0.644
[2.672] [2.850] [4.413]
6 months -1.998 -2.337 -1.632
dummy or Ln AT [1.680] [1.969] [2.518]
AB RET 20 x 6 months 5.953 5.229 4.044
(Dummy) [4.463] [4.680] [8.332]
AB RET 40 x 6 months -1.226 0.449 -3.878
(Dummy) [5.446] [6.060] [8.983]
B/M -0.466 -0.625* 0.172
[0.426] [0.363] [1.364]
FCF 0.234 *** 0.202*** 0.213 ***
[0.045] [0.041] [0.045]
ROA 6.941** 3.878 4.35
[2.782] [2.725] [3.106]
Ln (Total 3.156 *** 3.653 *** 3.815 ***
Assets) [0.888] [0.965] [1.025]
SECOND -4.716 *** -4.813 *** -4.366 ***
[1.287] [1.357] [1.537]
Intercept 29.856 *** -4.466 -5.891
[5.034] [6.978] [7.151]
Industry and Not Not Not
year dummies reported reported reported
Sample size 1,451 1,115 958
Adjusted [R.sup.2] 0.256 0.275 0.247
Ln [DELTA]T
Full Sample Subsample 1 Subsample 2
(1970- (1990- (exclude
2006) 2006) bubble pd.)
AB RET 20 -5.556 ** -6.498 *** -8.692 **
[2.297] [2.504] [3.866]
AB RET 40 1.562 1.118 -0.021
[2.586] [2.616] [3.953]
6 months 1.030* 0.638 0.986
dummy or Ln AT [0.606] [0.795] [0.945]
AB RET 20 x 6 months
(Dummy)
AB RET 40 x 6 months
(Dummy)
B/M -0.416 -0.589 0.1 45
[0.433] [0.374] [1.373]
FCF 0.227 *** 0.198 *** 0.212 ***
[0.044] [0.040] [0.043]
ROA 6.362 ** 3.575 4.153
[2.740] [2.613] [3.003]
Ln (Total 3.278 *** 3.725 *** 3.884 ***
Assets) [0.887] [0.959] [1.018]
SECOND -4.377 *** -4.735 *** -4.155 ***
[1.331] [1.309] [1.473]
Intercept 22.694 *** -8.947 -12.719
[7.173] [8.955] [9.581]
Industry and Not Not Not
year dummies reported reported reported
Sample size 1,451 1,115 958
Adjusted [R.sup.2] 0.258 0.276 0.249
Note: The table reports an OLS regression estimating
the determinants of corporate investment. The dependent
variable is corporate investment measured by total net
property, plant and equipment scaled by total assets.
6 months dummy (SEO within six months of IPO) is a dummy
variable that takes on the value of 1 if the number of
calendar days between IPO and the first SEO is less than
six months. Ln AT is the logarithm of the number of calendar
days between IPO and the first SEO. AB RET 20 is the abnormal
return over the period from trading day 1 to trading day
20 after the IPO date. AB RET 40 is the abnormal return
over the period from trading day 21 to trading day 40 after
the IPO date. B/M equals the ratio of book value of equity to
market value of equity. FCF is the free cash flow. The cash
flow measure is scaled by the firm's beginning-of-year capital.
ROA is the OIBD (operating income before depreciation) normalized
by total assets. Ln (Total Assets) is the logarithm of the total
assets. SECOND is a dummy variable equal to one if the percentage
of secondary shares offered in an SEO is greater than 50 percent
of total offerings. Industry and year dummy variables are included
in the regression but results are not reported. Standard errors are
in brackets. *, **, and *** denote significance level at the 10%,
5% and 1% levels.
Table 10
Changes in operating performance: Median ROA (%)
Less than 6 months
Year N Unadjusted Adjusted
-1 218 8.66 0.08 **
+1 169 5.58 -0.89
+2 153 6.87 -0.27
+3 138 6.39 -4.26
-1 to 1 -5.21 **
-1 to 2 -3.25 **
-1 to 3 -2.74
More than 6 months
Difference
Year N Unadjusted Adjusted in Adjusted
-1 1,460 12.45 0.08 *** -0.01
+1 1,245 11.13 0.71 *** -1.60 *
+2 1,139 10.69 2.00 *** -2.27 **
+3 1,071 10.53 2.40 *** -6.66 **
-1 to 1 -0.92 ** -4.29 **
-1 to 2 0.00 -3.25 **
-1 to 3 0.25 -2.99
Note: This table reports the median operating performance for
issuers and non-issuers matched on industry, firm size and
pre-issue operating performance. The matching procedure follows
Bouwman, Fuller, and Nain (2009). The adjusted operating
performance is the paired difference between the ROA of the issuing
firms and the ROA of their matching non-issuing firms. We
categorize the issuing firms by the length of time since IPO at the
date of first SEO. The tables reports the median OIBD (operating
income before depreciation) scaled by assets. The ratio is
winsorized at the top and bottom 1%. Statistical tests are based on
the Wilcoxon signed-rank test. *, **, and *** denote significance
level at the 10%, 5%, and 1% levels.
