Earnings management and long-run stock underperformance of private placements.
He, Daoping ; Yang, David C. ; Guan, Liming 等
INTRODUCTION
This study seeks to accomplish two goals regarding the issuance of
seasoned private placements of common equity (hereafter, private
placements): (1) to investigate managers' earnings manipulation
behavior of U.S. issuers around the time of the issuance; (2) to examine
whether such earnings manipulation behavior helps explain the long-term
post-issue stock underperformance.
Private placements, together with seasoned public offerings of
common equity (SEO), are two important vehicles by which public firms
obtain equity financing. Contrary to seasoned public offerings of common
equity, which issue new equity to the general public, private placement
issuers sell new equity to a restricted number of investors. More than
30 percent of seasoned equity financing from external investors in
recent years has come from private placements (Federal Reserve Bulletin,
see Appendix A). While there is a sizeable body of literature on
earnings management around the time of seasoned public offerings and on
the issuers' post-issue stock underperformance, research on these
issues related to private placements is scarce.
Studies on earnings management constitute an important research
stream in the literature about the quality of earnings. Typical research
methodology on earnings management first identifies a firm-specific
event around which managers' incentives to engage in opportunistic
earnings manipulation appear to be strong, then employs various accrual
models to test the researchers' earnings management hypotheses
(Healy and Wahlen 1999). In response to the call of Dechow and Skinner
(2000) for further research on earnings management incentives around the
time of securities issuance, this study examines, among other things,
earnings management around the time of private placement, an important
corporate event.
Earnings have been widely used by investors to assess firm value
and security offerings provide a direct incentive to manipulate earnings
(Dechow and Skinner 2000). Managers of an issuing firm could use the
accounting methods allowed under generally accepted accounting
principles (GAAP) to inflate reported earnings at the time of the
issuance in an attempt to portray a rosy picture of the firm's
prospects. To the extent that such income-increasing accounting choices
are not detected by investors, managers may obtain more favorable terms
when selling new shares, thus gaining direct monetary benefits for
themselves and the firm. Existing literature provides an abundance of
evidence of earnings management around the time of various types of
security issues, particularly initial public offerings (Aharony et al.
1993; Friedlan 1994; Teoh et al. 1998a; DuCharme et al. 2001), seasoned
public equity offerings (Teoh et al. 1998b; Rangan 1998), convertible
bond issues (Margetis 2004), and stock-for-stock mergers (Erickson and
Wang 1999). Since private placement is one of the most important sources
of corporate financing, this study predicts that managers of the issuing
firms have strong economic incentives to inflate reported earnings
around the time securities are issued.
This study finds that managers of U.S. private placement issuers
tend to engage in income-increasing earnings management around the time
of the issuance. The mean and median of the discretionary accruals, the
proxy for earnings management, of 348 sample firms from 1989 to 2001 are
3.27 percent and 2.49 percent of total assets in the year prior to the
issue year. To eliminate the impact of other influencing factors, the
study employs a control sample consisting of firms matched on size and
leverage in the same industry of the issuing firms. In the year prior to
the private placements, the discretionary accruals of the issuing firms
significantly exceed their non-issuing peers by 3.99 percent in mean and
1.98 percent in median.
Issuing private placements could be an endogenous choice. To
mitigate this self-selection bias, the study also adopts the propensity
score matching method to form the control sample. Consistent with those
of size and leverage matching, discretionary accruals of issuing firms
significantly exceed their non-issuing peers by 4.27 percent in mean and
1.87 percent in median in the year prior to the issuance. This result
suggests that the findings on earnings management are not sensitive to
alternative matching methods in selecting the control sample.
Among various anomalies in the semi-strong form of the market
efficiency hypothesis, private placement has recently been found to be
mispriced at the time of the issuance. In particular, Hertzel et al.
(2002) observe that firms conducting private placements experience
negative abnormal stock returns during the post-issue period. They
postulate that the reason for the post-issue stock underperformance is
that, at the time of the issuance, investors are over-optimistic about
the issuing firms' future performance. However, the source of the
over-optimism is not identified clearly.
Thus, the second purpose of this study is to examine whether
earnings management at the time of private placements serves as a likely
source of investor over-optimism. If investors misinterpret the
manipulated earnings around the time of private placements, the stock
price would be temporarily overvalued. However, when the inflated
earnings do not persist in the future and/or the income-increasing
accruals made at the time of issuance reverse in the subsequent periods,
investors become disappointed and beat down the stock price of the
issuing firms. To test this projection, this study proposes that the
income-increasing accounting accruals made at the time of private
placements predict the post-issue stock underperformance.
To explore the relationship of discretionary accruals at the time
of private placement issuance with the post-issue stock performance, the
sample is divided into four groups according to the discretionary
accruals made in the year prior to the issuing year. The quartile group
with the smallest discretionary accruals is called the conservative
group and the one with the largest discretionary accruals is called the
aggressive group. The study finds that the aggressive group has
consistently poorer post-issue stock performance than the conservative
group. For example, the three-year post-issue market excess return of
the aggressive group lags that of the conservative group by 13.29
percent, and the three-year post-issue abnormal return of the aggressive
group lags that of the conservative group by 35.23 percent. This result
suggests that firms that inflate their earnings more aggressively around
the time of private placements experience poorer stock performance
subsequent to the issuance.
The study also runs regressions of the post-issue stock performance
on discretionary accruals at the time of private placements, along with
control variables such as size, book to market ratio, and industry
dummies. Three-year market excess returns and the abnormal returns of
issuing firms are both significantly negatively correlated with the
discretionary accruals made in the year prior to the issuance. This
evidence supports the hypothesis that the discretionary accruals around
the issuance of private placements predict the post-issue stock
underperformance. Thus, earnings manipulation around the time of private
placements could be a source that causes investors' over-optimism
at the time of the issuance.
This study makes a number of contributions to the literature.
First, the study adds evidence to the sizeable body of research on
earnings management around the time of security issuance by documenting
income increasing earnings management around the time of private
placements. Therefore, it complements the findings of prior studies on
other types of security issuances. Second, the study sheds new light on
the capital market anomaly related to private placements. While existing
literature attributes post-issue stock underperformance to investor
over-optimism at the time of the private placement, the findings of this
study suggest that a possible source of this over-optimism is earnings
management around the time the securities are issued.
BACKGROUND AND LITERATURE REVIEW
Public offerings and private placements, focusing on raising equity
capital from external investors, are two important ways for public
companies to conduct seasoned equity offerings. Both individuals and
institutional investors can participate in public offerings, which are
usually conducted via a managing investment bank as the underwriter or
underwriting syndicate. U.S. companies must register the issue with the
Security and Exchange Commission (SEC) when they conduct public
offerings.
Companies can avoid this costly process if they conduct equity
offerings privately. Private placements refer to the direct issuance of
equity securities to a restricted number of investors. Most private
placement investors are large institutional investors such as mutual
funds and pension funds. The price of the issue is determined by
negotiation between the issuer and the investors (Ross et al. 2002;
Keown et al. 2003; Marciukaityte 2001).
There are several other advantages of private placements over
public offerings. First, the issuers are exempted from the registration
and disclosure requirements of various securities statutes. Second,
investment dealers' fee for a private placement is much less than
that for a public offering. Third, private placements can also help
firms raise capital quickly. Fourth, obtaining private placements
significantly improves a firm's ability to attract additional
capital, research partners, and commercial partners (Srivastava 1989;
Folta and Janney 2004; Janney and Folta 2006).
Most studies on seasoned public offerings in the mid-1980s examine
the stock market reaction to the issue announcements. These studies
generally document a negative market reaction to the announcement of the
issuance. On average, the two-day abnormal stock returns on announcement
of public offerings are -3.14 percent for industrial companies and -0.75
percent for utility companies (Smith 1986), suggesting that stock price
is overvalued when firms conduct public offerings. A theory developed by
Myers and Majluf (1984) to explain this negative market reaction argues
that managers know a good deal about the company they manage, so when
the company's stock price is undervalued, managers will be less
likely to issue equity to new investors because it would let them take
advantage of existing shareholders. It is more likely that managers
decide to issue new equity when the stock price is overvalued, so stock
price falls when firms announce the public offerings because it sends a
signal to the market that managers believe the company's stock is
overvalued.
Loughran and Ritter (1995) conduct the first study to examine the
long-run post-issue stock performance of firms that conduct public
offerings. They find that after public offerings, firms experience
negative abnormal stock returns for up to five years compared to similar
size firms in the same industry which do not issue new equity. Spiess
and Affleck-Graves (1995) find similar results, so both studies conclude
that managers take advantage of a firm's specific information to
issue equity when the firm's stock is overvalued. This explanation
was called the windows of opportunity hypothesis. Considering the
negative market reaction to the announcement of the new issues, the
underreaction hypothesis is developed to explain the poor post-issue
stock performance. The underreaction hypothesis maintains that the stock
market reflects only part of the information about the share price when
new issues are announced. Daniel et al. (1998) develop a model to
explain the underreaction phenomenon from the behavioral theory approach
and conclude that investors tend to overestimate their ability to
generate information and tend to underestimate their forecast errors.
Unlike firms making public offerings, firms conducting private
placements experience positive market reaction at the announcement of
issuance. Wruck (1989) documents a 4.5 percent average abnormal return
during the announcement period and Hertzel and Smith (1993) report a
similar result. However, the two papers differ in their explanations for
this phenomenon. First, Wruck proposes an ownership structure hypothesis
to interpret her findings, explaining that the higher the level of
ownership concentration, the easier it is for a small group of
shareholders to influence managers' behavior to align
managers' and shareholders' interests. To support this
explanation, Wruck finds that the total holdings of those investors
reported in proxy statements increase from an average of 31 percent to
an average of 37 percent of the firm's voting shares and the change
in firm value at the announcement of a private placement is strongly
correlated with the resulting change in ownership concentration. Hertzel
and Smith (1993), on the other hand, propose the information hypothesis
to explain the positive market reaction. Following Myers and Majlufs
(1984) assumption, Herzel and Smith conclude that undervalued firms will
not likely issue equity publicly to avoid releasing negative signals
about the firms' value. In addition, they argue that the
willingness of private placement investors to commit funds to the
issuers conveys a signal to the market that the issuers are undervalued.
To support this hypothesis, they find a correlation between the positive
abnormal returns at the announcement time and the potential
undervaluation.
The findings of Wruck (1989) and Hertzel and Smith (1993) tend to
support the view that the involvement of large investors that purchase
private placements increases the issuer's value by providing either
a monitoring role or a certification role. Given that most private
placement investors are large institutional investors such as mutual
funds and pension funds, this is consistent with many studies
documenting that institutional investors have an effective monitoring
effect on management behavior (In a review of corporate governance
studies, Shleifer and Vishny (1997) conclude that institutional
investors in the U.S. reduce agency cost in firms and pressure managers
to improve their true economic performance).
If the institutional investors that purchase private placements do
enhance the monitoring role and constrain managers' opportunistic
behavior, different from the public offerings, we may not find earnings
management behavior around the time of private placement issuance.
However, recent findings on private placements have suggested that most
of those institutional investors involved in private placements are
passive and that they bring no more of a monitoring role than do
investors in public offerings (Barclay et al. 2005; Wu 2004). Therefore,
it is still likely that managers engage in earnings management to
mislead investors at the time of private placements.
Hertzel et al. (2002) find that, along with positive market
reaction to the announcements of issuance, public firms conducting
private placements experience poor post-issue stock performance, which
is not consistent with the underreaction hypothesis drawn from public
offering studies. Under the underreaction hypothesis, the positive
announcement effect should cause firms conducting private placements to
experience positive abnormal returns in the long run. Hertzel et al.
conclude that investors are overoptimistic about the prospects of firms
that issue equity, publicly and privately. The importance of this
finding is that, contrary to the traditional belief, firms conducting
private placements are overvalued, possibly because investors are
overoptimistic about the monitoring role of new institutional investors.
Recent studies have challenged the anticipated monitoring effect by
the involvement of new institutional investors. Larcker et al. (2005)
find that fourteen corporate governance factors, including institutional
ownership, explain only 0.6 percent to 5.1 percent of the
cross-sectional variation of a wide set of dependent variables,
including abnormal accruals. This finding suggests that institutional
ownership has very limited ability to explain managerial behavior and
organizational performance. Barclay et al. (2005) provide evidence that
supports the entrenchment hypothesis, which proposes that managers
consider not only the interests of shareholders but their own interests
as well when they conduct private placements. The entrenchment
hypothesis also maintains that managers dislike being monitored (Brennan
and Franks 1997; Field and Sheehan 2004), and are, therefore, likely to
place the equity with passive institutional investors who will not
interfere with managerial decisions. Barclay et al. find that, after the
issuance, most private placement purchasers remain passive, that firm
value declines, and that there are few acquisitions. Wu (2004) examines
the monitoring role of managers on the choice between public offerings
and private placements and finds that private placement investors do not
engage in more monitoring than public offering investors do. In the
absence of the monitoring roles brought by the new passive institutional
investors, managers may act opportunistically when they conduct private
placements if there are strong incentives for them to do so.
HYPOTHESIS DEVELOPMENT
Earnings are among the most important measures investors use to
assess a firm's future performance (Healy and Wahlen 1999). Dechow
and Skinner (2000) suggest that, around the time of new securities
issuance, managers have strong incentives to manipulate earnings to
portray a rosy picture of the firm's future performance and,
consequently, may sell the securities on more favorable terms and
therefore reduce the cost of financing. Empirical evidence on certain
types of securities issue appears to support this argument. For example,
both the initial public offering issuers (Aharony et al. 1993; Friedlan
1994; Teoh et al. 1998a; DuCharme et al. 2001) and seasoned public
equity issuers (Teoh et al. 1998b; Rangan 1998) tend to make
income-increasing accounting choices around the time of issuance in an
attempt to increase the selling prices of the new equity. In corporate
stock for stock mergers, the acquiring firms manage earnings upward in
the periods prior to the merger agreement to increase their stock prices
in order to reduce the cost of buying the target firms (Erickson and
Wang 1999). Unlisted firms also tend to manipulate earnings upward prior
to receiving venture capital financing in order to show a better picture
of their company's prospects, thereby increasing the chances of
being funded by venture capitalists (Beuselinck et al. 2005). Similarly,
firms conducting private placements may also have incentives to report
inflated earnings prior to the issuance in order to attract more
investors since the manipulated earnings may lead the investors to
believe that the reported earnings could continue into the future and
therefore become overly optimistic about the issuers' future
performance. This would allow private placement issuers to boost their
images and sell their new equity on more favorable terms.
It would be pointless for managers to manipulate earnings if
private placement purchasers could see through it. However, Healy and
Wahlen (1999) argue that investors may not fully see through earnings
management that is reflected in accruals; even for underwriters, fully
adjusting for accounting choices may be difficult and costly (Friedlan
1994). Since most private placement investors are institutional
investors, managers' opportunities to manage earnings around the
issuance may be restrained because of the active involvement of
institutional investors (Chung et al. 2002). However, Barclay et al.
(2005) find that managers issuing private placements deliberately select
passive institutional investors, and Wu (2004) provides evidence showing
that private placements investors do not provide more monitoring roles
than public investors. Therefore, it is likely that private placement
issuers still have the opportunity to manipulate earnings around the
issuance and that such behavior may go undetected by private placement
investors.
Engaging in earnings management at the time of private placement is
not without downside risks. First, subsequent discovery of the earnings
management around the time of private placement may lead to lawsuits by
investors if the earnings overstatement leads to investor losses
(DuCharme et al. 2004). Second, such a discovery will undoubtedly reduce
the credibility of the issuing firms' financial statements and
impair their ability to raise additional capital at favorable terms in
the future. Finally, U.S. firms identified by the Securities and
Exchange Commission (SEC) as violators of GAAP will face an increase in
their future costs of capital.
Even so, because the economic benefits of reporting inflated
earnings prior to private placement are substantial, if managers do not
think they are likely to be discovered or if the costs of discovery are
perceived less than the potential benefits, they are likely to adopt the
discretionary accounting choices that increase the reported earnings
around the time of private placement issuance. Therefore, the first
hypothesis of the study is:
[H.sub.1]: Managers of U.S. private placement issuing firms
manipulate reported earnings upward around the time of issuance.
Hertzel et al. (2002) recently pinpoint the long-term stock
underperformance subsequent to private placements. Relative to control
firms matched by size and book to market ratio, the mean three-year buy
and hold abnormal return is -23.8 percent, which is similar to that
found for initial public offerings and seasoned public equity offerings.
Hertzel et al. argue that the post-issue stock underperformance is
likely due to the investors' over-optimism about the issuers'
future performance around the time of private placements, although they
do not identify clearly the source of the over-optimism.
This study proposes that earnings management around the time of the
issuance of private placements is a likely source of investor
over-optimism. If investors expect the reported (but manipulated)
earnings around the time of private placements to persist into the
future, stock price will be temporarily overvalued. Then, when the
income-increasing accruals reverse in subsequent periods and the
earnings trend does not persist, investors may become disappointed and
beat down the stock price. Thus, if earnings management at the time of
issuance is a source of investor over-optimism, the aggressive
accounting accruals at the time of the private placement issuance will
cause post-issue stock underperformance. Several studies examining
seasoned public offerings have made a similar prediction and found
evidence supporting this prediction. For example, Teoh et al. (1998b)
and Rangan (1998) find that earnings management around the issuance of
seasoned public offerings explains a portion of the post-issue stock
underperformance. Teoh et al. (1998a) find that the manipulated earnings
around the issuance of initial public offerings are correlated with
post-issue stock underperformance. The higher the discretionary accruals
around the issuance of initial and seasoned public offerings, the lower
the abnormal post-issue stock returns. Following this line of reasoning,
the second hypothesis is:
[H.sub.2]: Earnings management around the time of private
placements conducted by U.S. issuers predicts post-issue stock
underperformance.
SAMPLE AND METHODOLOGY
The initial U.S. sample of private placement issues is obtained
from the New Issues Database from Securities Data Corporation. The
issuers' financial data are obtained from Standard and Poor's
Research Insight database, and stock returns from the Center for
Research in Security Prices (CRSP). The New Issues Database contains 831
private placement common stock issues from 1989 to 2001 in the New York
Stock Exchange (NYSE), the American Stock Exchange (AMEX), and the
NASDAQ. To qualify for the study sample, firms issuing private
placements must have the necessary financial data to allow a calculation
of discretionary accruals in the year prior to the issue, matching
measurements, and stock returns for at least one month after the issue.
Financial and utility firms are eliminated from the sample because these
firms are subject to special regulations. In order to reduce the
confounding effects on earnings management from public equity offerings,
firms conducting both public offerings and private placements in the
same year are also excluded from the sample. If firms issued multiple
private placements within three years, only the first issue is kept in
the final sample.
The final sample contains 348 observations. Table 1 presents the
sample size and gross proceeds by year and industry classification for
the U.S. private placements issues. Because the study adopts the cash
flow approach to calculate discretionary accruals for the U.S. firms,
the sample starts from the year 1989. The study also needs to test the
stock performance five years after the issuance, so sample data ends in
the year 2001. Four years (1992, 1993, 2000, and 2001) are very active
and contain more than 25 issues each year. Chemical products,
instruments and related products, and service industries each carry more
than 10 percent of the sample.
Identifying of the timing of earnings management a priori is
critical in any earnings management study. In keeping with the earnings
management research on IPOs and SEOs, a company's annual financial
statements are defined as being for the issuing year (year 0) if the
fiscal year-end of the financial statements is within 12 months after
the private placement date. Based on this definition, it is possible
that the annual financial statements of an issuing firm for year 0 will
cover some months prior to the private placement date because the fiscal
year-end can be less than 12 months after the private placement date.
Once year 0 is defined, financial statements for other years surrounding
the private placement date (i.e., year -2, -1, +1 and +2) can be defined
accordingly.
Extant earnings management literature on IPOs and SEOs documents
that earnings management is most prominent in the year preceding the
issuing year (year -1) and/or the year of the issuing year (year 0).
Several studies (Friedlan 1994; Aharony et al. 1993; DuCharme et al.
2000) have examined earnings management prior to making IPOs and find
that IPO firms tend to inflate earnings in the year prior to the IPO
(year -1). Teoh et al. (1998a) define the issue year (year 0) as the
fiscal year in which the IPO occurs (which is the same definition this
study uses) and includes both pre- and post-IPO months, arguing that IPO
firms have incentive to manipulate both pre- and immediate post-IPO
earnings. Teoh et al. (1998b) study earnings manipulation around the
issue of SEOs and find that discretionary current accruals for SEO firms
are positive in year -1 and more prominent in year 0.
While the same arguments about earnings management for IPOs and
SEOs can be made for private placements, the relative strength of these
arguments and the effects of some other factors also need to be
considered in determining the timing of earnings management by private
placement issuers:
1. Most IPO and SEO studies maintain that earnings management made
in year -1 helps to inflate stock prices and to increase the proceeds
from the issuance.
2. Since the reverse of accruals made in earlier years increases
the litigation risk for the issuers, managers have incentives to
continue to manage earnings upward after issuing securities.
3. Prior to issuing securities, firms have strong incentive to
release optimistic earnings forecasts and to announce good news in order
to boost the stock price or investors' confidence. Ruland et al.
(1990) find that firms issuing earnings forecasts are more likely to
finance externally in the subsequent three months than are the control
firms that did not issue earnings forecasts. Frankel et al. (1995) also
find that firms are more likely disclose earnings forecasts if they plan
to raise capital. When managers conduct voluntary disclosures, they tend
to disclose information favoring them or existing shareholders. For
example, Aboody and Kasznik (2000) find that CEOs opportunistically
manage the timing of their information disclosures to increase the value
of their stock option awards by delaying announcements of good news and
rushing forward bad news before the awards. Again, to reduce the risk of
litigation or to increase the credibility of managers' voluntary
disclosures, the issuing firms have incentive to inflate earnings in the
annual reports that cover the time period when the earnings forecasts
and good news are released.
4. The ability of managers to boost earnings in consecutive years
is limited by a number of factors. First, the current accrual accounting
system provides a limited set of methods by which to manipulate earnings
(Watts and Zimmerman 1986). Second, since the balance sheet accumulates
the effect of previous accounting choices, managers' ability to
manipulate earnings decreases with how much net asset values have been
already overstated on the balance sheet (Barton and Simko 2002).
5. Since all accruals will reverse in the future, the external
auditor bears a higher risk of litigation if income-increasing earnings
management occurs in consecutive years. Therefore, if earnings
management occurs in one year, the external auditor has incentive to
curtail earnings management in the following year.
6. The issuing firms may also face litigation and reputation damage
if a large amount of earnings management is detected by the investors.
7. Since private placement buyers are mostly institutional
investors, the ability of these investors, even though passive, to
initiate lawsuits against the issuer, if earnings management is
detected, is likely to be greater than the ability of investors in IPOs
or SEOs to do so. This is because the free-rider problem is less severe
among the private placement buyers.
Thus, the timing and extent of earnings management around the time
of private placements are an empirical issue. It is possible that the
issuers manage earnings upward in both year -1 and year 0, while it is
also possible that the issuers manage earnings upward only in year -1
and not in year 0, or vice versa. Since this study examines the
discretionary accruals in each of the five years surrounding the issuing
year, the timing and extent of the earnings management made by U.S.
issuing firms will likely be captured by this time span.
If earnings management has occurred, it is likely that there is
evidence in measures that reflect accounting policy choices. It would be
informative not only to find evidence of earnings management, but also
to identify the accounting choice or choices that have been employed to
achieve the desired earnings. However, since most accounting choices are
not observed by outsiders, researchers rely on various aggregate
measures of earnings management, assuming that methods used by managers
to manipulate earnings would be spread over a portfolio of accounting
choices. Most recent studies on earnings management have used
estimations of discretionary accruals as a measure of earnings
management, so this study will adopt the same measurement as the proxy
for earnings management.
Dechow et al. (1995) assess the relative performance of five
alternative discretionary accrual models for detecting earnings
management and conclude that a modified version of the Jones (1991)
model provides the most powerful tests of earnings management. Further,
Subramanyam (1996) finds that the cross-sectional variation of modified
Jones (1991) model provides better estimates of the normal accruals than
the times-series model does. Bartov et al. (2000) find that the
cross-sectional modified Jones model outperforms the time-series
modified Jones model in identifying firms with qualified audit opinions.
Therefore, this study will adopt the modified Jones (1991) model and the
cross-sectional estimation method to measure discretionary accruals.
Also, since firms' past performances could also affect the level of
their accruals (Kathori et al. 2005), the lagged return on assets (ROA)
is also included in the regressions to estimate nondiscretionary
accruals for each firm.
Total accruals are measured using the cash flow approach:
[TA.sub.t] = [NI.sub.t] - [CFO.sub.t] ] / [A.sub.t-1] (1)
where:
[TA.sub.t] = total accruals
[NI.sub.t] = income before extraordinary items and discontinued
operations (Research Insight data item #123)
[CFO.sub.t] = cash flow from operations (Research Insight data item
#308 minus data item #124)
To calculate the discretionary accruals, the non-discretionary
portion of total accruals must be estimated. The expected
nondiscretionary accruals for firm i in year t (N[DA.sub.it]) are
measured as:
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (2)
where:
[DELTA] [REV.sub.it] = change in revenue for firm i in year t
[DELTA] [REC.sub.it] = change in net receivables for firm i in year
t
[PPE.sub.it] = gross property, plant, and equipment for firm i at
the end of year t
[ROA.sub.t-1] = Return on average assets in year t-1
[[beta].sub.0it], [[beta].sub.1it], [[beta].sub.2it],
[[beta].sub.3it] = firm-specific parameters for firm i in year t.
In equation (2), the firm-specific parameters, [[beta].sub.0it],
[[beta].sub.1it], [[beta].sub.2it], and [[beta].sub.3it], are estimated
cross-sectionally using the two-digit SIC code for firm j's data (j
[not equal to] i):
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (3)
Once the nondiscretionary accruals are estimated, the discretionary
accruals for firm i in year t ([DA.sub.it]) are calculated as the
prediction error:
[DA.sub.it] = [TA.sub.it] - [NDA.sub.it] (4)
Managers of issuing firms may have managed earnings for reasons
other than trying to induce investor optimism. Two prominent reasons
relate to firms' desire to reduce the political cost and the debt
default cost (Watt and Zimmerman, 1986). Empirically, a firm's
political cost is usually proxied by firm size, and debt default cost is
proxied by the leverage of the firm. To mitigate the impact of these two
factors on the measurement of earnings management, this study employs a
matched control sample of non-issuers. If the discretionary accruals of
the issuing firms are significantly different from those of the
non-issuing control firms in year t, then there is evidence of earnings
management among the issuing firms in year t and the results of earnings
management of the issuers are not likely due to incentives other than
inducing investors to accept terms more favorable to the issuers.
For the matching procedure, one control firm is selected for each
private placement firm by matching the two-digit SIC code, total assets,
and debt-to-asset ratio at the end of year -1. The control firm should
not have conducted private placements in the following three years.
Because there are two continuous variables in the matching, this study
employs the procedure proposed by Murray (1983). Thus, for each
potential matched firm, the Mahalanobis distance is calculated as:
[D.sup.2] = ([M.sub.b] - [M.sub.c])'[W.sup.-1] ([M.sub.b] -
[M.sub.c]), (5)
where:
[D.sup.2] = the distance measure of firm b from private placement
firm c,
[M.sub.b] = a vector of matching variables for firm b,
[M.sub.c] = a vector of matching variables for private placement
firm c,
W = the covariance matrix of the cross-section of matching
variables.
[D.sup.2] is considered as a univariate measure of multidimensional
differences and is calculated for all non-issuing firms within the same
two-digit SIC code as the issuing firm. The control firm is the one that
has the smallest [D.sup.2]. The Mahalanobis distance approach provides a
more precise matching measurement than the Euclidean distance approach
in that it considers the variance and covariance of each matching
variable (Murray 1983).
Three measures of stock performance are used in this study: raw
return, market excess return, and abnormal return. All the measures are
the buy-and-hold returns over a three-year period. The raw returns and
market excess returns do not adjust for firms' risk and, therefore,
are biased estimates of stock performance. The purpose of including
these two measures is so they can serve as benchmarks for the abnormal
returns.
The method used to calculate the abnormal returns is similar to
Hertzel et al. (2002). Specifically, the abnormal return of an issuing
firm is calculated as the difference between the buy-and-hold raw return
of the issuer and that of a matched non-issuer. Finding the correct
matching firm is critical in this process because measures of long-term
abnormal stock returns could be subject to greater measurement errors
than measures of short-term returns (Kothari and Warner 1997). Barber
and Lyon (1997) analyze the empirical power and specification of test
statistics in event studies designed to investigate long-run abnormal
stock performance and conclude that the control firm approach, in which
sample firms are matched by similar size and book to market ratios of
industry peers, yields well specified test statistics. The matching
procedure in this study will employ the Mahalanobis distance approach
using two continuous variables--firm size and book to market ratio--for
a firm within the same two-digit SIC code as the issuer. The firm with
smallest [D.sup.2] in the same two-digit SIC code is the control firm
for the issuer.
With the issuing year defined as year 0, discretionary accruals of
issuing firms in years -2, -1, 0, +1, +2 are calculated. The most
important measures used to test Hypothesis 1 are discretionary accruals
in years -1 and 0, since discretionary accruals for other years are not
expected to be significantly different from zero. A statistically
significant positive measure of discretionary accruals in either year -1
or year 0 (or both) will lend support to Hypothesis 1.
Testing Hypothesis 2 involves two steps. In the first step, four
portfolios of issuers are formed based on the quartiles of discretionary
accruals around the issue year; the discretionary accruals for the year
that are significantly positive are used to form the portfolios, and the
two extreme portfolios are called the aggressive (quartile 4) and the
conservative (quartile 1) private placement issuers. The abnormal
returns for each portfolio are calculated as the cumulative buy-and-hold
return on sample firms less the simple cumulative buy-and-hold return on
control firms over the three-year post-issue period. Hypothesis 2
predicts that the most aggressive issuers will exhibit the most negative
post-issue abnormal returns and that the most conservative issuers will
exhibit the least negative (in terms of magnitude) post-issue abnormal
returns. This first step provides a view of the relationship between the
earnings management around the time of private placements and the
post-issue stock performance.
Building on the results of the first step, the second step is a
formal statistical test of Hypothesis 2. Specifically, OLS regressions
are run using individual issuer's three-year post-issue raw return,
market access return, and abnormal return as the dependent variable. The
independent variable of primary interest to the study is the abnormal
accruals for year -1 or year 0, whichever is significantly positive. The
regressions also include an industry dummy, firm size, and book to
market variables as control variables. The industry dummy accounts for
post-issue stock performance across industries and firm size and book to
market variables control for firm characteristics. A significantly
negative estimate of the coefficient of the discretionary accruals
variable will lend support to Hypothesis 2.
The OLS regression model is specified as:
[R.sub.i] = [[beta].sub.0] + [[beta].sub.1] ([DA.sub.i] ) +
[[beta].sub.2] ([Size.sub.i]) + [[beta].sub.3] ([BtoM.sub.i]) +
[summation] [gamma] ([Industry_dummies.sub.i] ) + [[epsilon].sub.i] (6)
where:
[R.sub.i] = issuer's raw return, market excess return, or
abnormal return
[DA.sub.i] = issuer's discretionary accruals around issuance
[Size.sub.i] = issuer's market value of equity
[BtoM.sub.i] = book to market ratio
[Industry_dummies.sub.i] = industry dummy variables
EMPIRICAL RESULTS
Table 2 reports five years of asset-scaled discretionary accruals
around the issue of private placements for the U.S. issuers and those
for the control firms. The mean and median for year -2 are not
significantly different from zero; but for year -1, the year prior to
the private placements, the discretionary accruals of the issuing firms
have a mean of 3.27 percent and a median of 2.49 percent of total
assets, both of which are significantly positive. For years 0, +1, and
+2, the means and medians are greater than zero; however, they are not
statistically significant at the conventional levels (except for the
median for year +1, which is significant at ten percent level).
Since managers' incentives to manage earnings could also be
due to firm size and leverage, the results from the modified Jones model
should be compared with those of the control group to draw an overall
inference about earnings management around the time of private
placements. Table 2 also reports the mean and median for those control
firms over the same time period, none of which is significantly
different from zero. The study conducts pair-wise comparison tests on
the differences in discretionary accruals between private placement
issuers and their control firms, and the result shows that in the year
-1, private placement issuers have significantly greater discretionary
accruals in both mean and median than their non-issue peers, suggesting
that the observed abnormally high magnitude of earnings managements in
year -1 cannot be attributed to firm size, leverage or the
industry-specific categories of the sample. The comparison of the mean
and median of discretionary accruals between the issuing firms and the
control firms does not reveal any statistically significant difference
in other years. Thus, the empirical results support the first hypothesis
that managers tend to manipulate earnings upward in the year prior to
the issue of private placements.
Existing literature documents that after issuing private
placements, firms experience negative abnormal stock performance
(Hertzel et al., 2002). This study examines the sample firms and finds
results consistent with existing literature. The average three-year
buy-and-hold abnormal stock returns for the sample are -32.83 percent.
The existing literature has postulated that the negative abnormal stock
return is due to investors' over-optimism about these issuing
firms' future performance. This study argues that earnings
management around the private placement issue could be a factor for the
investors' over optimism, because the inflated earnings around the
issuance could mislead investors about the issuer's future
performance. The reversal of accruals in the following years results in
a drop in the issuer's stock price. Thus, the more the earnings
management, the poorer the post-issue stock performance.
To test this hypothesis, the study first classifies the issuing
firms according to the level of their discretionary accruals in year -1
to derive four portfolios. The quartile group with the lowest
discretionary accruals is called the conservative group, and the group
with the highest discretionary accruals is called the aggressive group.
Buy-and-hold raw returns are developed for each portfolio, and the
portfolio is rebalanced every year. Market excess returns for each
portfolio are also calculated. In addition, the study adopts the
Mahalanobis distance approach to develop a control firm for each sample
firm with a similar size and book to market ratio in the same industry.
The buy-and-hold excess returns over their control firms for each
portfolio are also developed as abnormal returns.
Table 3 reports the raw returns, the market excess returns, and the
abnormal returns for the conservative and aggressive portfolio for each
year over a three-year post-issue period. The market excess returns for
the conservative portfolio over the three-year period are 3.54 percent
and they are -9.75 percent for the aggressive portfolio. The abnormal
returns for the conservative portfolio over the period are -15.98
percent and they are -51.21 percent for the aggressive portfolio. Figure
1 depicts the size and book-to-market value adjusted abnormal returns
for each quartile, showing that the aggressive quartile performs more
poorly than the conservative quartile and suggesting that the higher the
level of discretionary accruals prior to the private placement issuance,
the poorer the post-issue stock performance.
Footnote:
The total sample is classified into four groups by the
issuer's discretionary accruals in the year preceding the issuing
year. The conservative group is the quartile group with the smallest
discretionary accruals and the aggressive group is the one with the
largest discretionary accruals.
[FIGURE 1 OMITTED]
The study runs OLS regressions of three-year post-issue stock raw
returns, market excess returns, and abnormal returns on discretionary
accruals in year -1, as well as on the control variables of market
value, book-to-market, and industry dummies. The regression is run at
the firm level, and Table 4 reports the regression results. The variable
of most interest is the discretionary accruals in year -1. The
coefficient of the discretionary accruals in year -1 on the three-year
market excess returns is -0.3485 and significant at the 5% level (t =
-1.96); the coefficient of discretionary accruals in year -1 on the
three-year abnormal returns is -0.4343 and also significant at the 5%
level (t = -1.83).
Thus, the results in table 4 support the second hypothesis. For
private placement issuers in the U.S., the higher the discretionary
accruals in year -1, the poorer the three-year post-issue market excess
returns and abnormal returns. Therefore, the level of earnings
management is associated with the three-year post issue stock
underperformance, suggesting that earnings management could be a factor
causing investors' over optimism prior to the issue of private
placements.
SENSITIVITY ANALYSIS
The selection of control firms is crucial in many empirical
studies. To test the first hypotheses, a control firm for each private
placement-issuing firm is chosen to mitigate the influence of factors
such as industry, size, and leverage. The discretionary accruals of
private placement firms in the year prior to the issues are
significantly greater than those of control firms, so the detected
earnings management is not due to factors other than the private
placement. To test the second hypotheses, control firms are also
developed for the calculation of the issuers' post-issue abnormal
returns. Compared to raw returns and market excess returns, abnormal
returns are a more accurate measure of stock performance in that they
mitigate some systematic and firm specific factors that can influence
the stock returns.
Empirical studies have used a variety of matching methods to derive
the control firms. When two or more continuous variables are used in the
matching, many studies have adopted the Euclidean distance approach and
have chosen the firm with the closest Euclidean distance to the
experimental firm as its control firm. Murray (1983) is the first to
apply the Mahalanobis distance approach to accounting and finance
empirical research. The Mahalanobis distance approach, as used in this
study, improves the Euclidean distance approach by considering the
variance and covariance of these control variables when calculating the
distance and, thus, provides more accurate matching. However, the
application of the Mahalanobis distance approach is seldom used in
accounting and finance studies.
An alternative matching approach, the propensity score matching
approach that structures non-experimental data to look like experimental
data, has gained popularity in economics research in recent years.
Rosenbaum and Rubin (1983) advocate the use of propensity scores, which
measure the probability that firms receive treatment, to reduce the
dimensionality of the matching. By matching on the scalar variable,
sample firms could be matched with the nearest non-treated firms having
a similar treatment condition on covariates. Since the propensity score
matching method is a significant improvement in matching techniques, it
becomes a rapid growing method in accounting and corporate finance as a
sensitivity analysis to address self-selection issues (Li and Prabhala
2007).
The sensitivity analysis conducted in this study uses the
propensity score matching method to select the control sample. A
logistic regression is processed to calculate propensity scores for the
sample firms and the potential control firms. Following existing
literature, the independent variables include trading system, industry,
issue year, firm size, leverage, book to market ratio, sales, return on
assets, and research and development expenses (Spiess and Affleck-Graves
1995, Schultz 2003, Mclaughlin et al. 1996, Loughran and Ritter 1995,
and Jung et al. 1996). The propensity scores, the probability that a
firm may conduct private placements, are derived after the regression.
The logistic regression results are shown in Table 5.
Overall, the logistic model for U.S. firms reports a 70 percent
accuracy for the issuing firms and a 75 percent accuracy for non-issuing
firms. To be qualified into the pool of non-issuing firms, a firm must
have necessary data to calculate discretionary accruals and stock
returns. Once the propensity score is calculated for each firm, the
control sample can be derived. A non-issuing firm with the nearest
neighbor of propensity score in the same industry is chosen as the
control firm for each issuing firm.
The annual discretionary accruals for the five-year period
surrounding the issue year are calculated for the control firm for each
private placement issuer. Table 6 reports the five year discretionary
accruals for the issuers and their control firms. Similar to the results
from using the Mahalanobis distance matching to select the control
sample, the mean and median of discretionary accruals in year -1 of the
sample firms are both significantly greater than those of their control
firms. There is no significant difference in discretionary accruals
between the issuing firms and control firms for years -2, 0, and +1.
Although year +2's the mean of discretionary accruals for sample
firms is significantly and marginally greater than that for the control
firms, the median is not significantly different between the two groups.
Thus, the results are consistent with the Mahalanobis distance approach,
suggesting that the empirical results in support of the first hypothesis
are not sensitive to alternative matching procedures in selecting the
control sample.
Raw returns are derived for the control firms and three years
post-issue abnormal returns are developed as buy-and-hold excess returns
over these control firms. Table 7 reports the results of the regression
of post-issue stock performance on discretionary accruals for year -1.
The coefficient of discretionary accruals in year -1 on three-year
abnormal returns is -0.5082 and is significant at the 10 percent level.
The results are qualitatively the same as those from the Mahalanobis
distance approach in selecting the control sample, suggesting that the
findings on the association between post-issue stock performance and
earnings management around the time of private placements are not
sensitive to alternative matching methods.
CONCLUSION
Private placements provide direct incentive to managers to
manipulate earnings. In doing so, managers may portray a rosy picture of
the firms' prospects to attract new investors and obtain more
favorable terms for selling new shares. This study investigates whether
private placement issuers manipulate their earnings around the time of
issue, and the results indicate that managers tend to income
increasingly manage their earnings around the private placement.
The study also examines the effect on stock performance of earnings
management around the issue of private placements. Investors could be
misled by the manipulated earnings and become over-optimistic about the
issuers' future performances. When the income-increasing accruals
reverse in subsequent periods, investors become disappointed and beat
down the stock price to the firms' fundamental values. Thus, the
study finds that post-issue stock underperformance is associated with
earnings management prior to the private placement and that, the higher
the level of earnings management before the issue of private placements,
the poorer the post-issue stock performance will be.
The study also investigates whether an alternative matching
technique could have influence on the findings. A popular matching
method in economics, propensity score matching, is used to replace the
Mahalanobis distance approach and the results are similar, suggesting
that the findings are not sensitive to alternative matching methods.
Testing earnings management using accrual models is a simultaneous
test of earnings management and the validity of the accrual models
(Kothari et al., 2005), so this study is limited by the accuracy of the
accrual model that is adopted to capture the existence and level of
earnings management. Although the study adopts the best performing
accrual model, to the extent that the model could fail to correctly
extract the discretionary portion from the total accruals, the results
should be interpreted with caution.
This study is also limited by the effectiveness of the matching
models in dealing with self-selection bias since issuing private
placements could be an endogenous choice. To deal with the
self-selection problem, this study conduct a sensitivity test using
propensity score matching method. Issuing firms are matched with control
firms by the similar probability to issue private placements. Although
the consistent results using the propensity score matching method with
the dimension by dimension matching using Mahalanobis distance are
conforming, the robustness of this study is based on the assumption that
unobserved private information should not explain outcome differentials
between firms choosing to issue private placements and those choosing
not to.
The findings of this study document that on average firms issuing
private placements have tendency to manipulate their earnings around the
issuance. However, the magnitude of the manipulations is different
across the issuing firms. It will be interesting to investigate what
factors affect the firms' decision to manage earnings or the
magnitude of the manipulations, whether these factors have similar
impact on private placement issuers with other equity issuers, such as
initial public offerings, seasoned public offerings, right offerings and
convertible bond offerings, etc and whether these factors have similar
impact across countries.
APPENDIX A: NEW U.S. SEASONED SECURITY ISSUES OF CORPORATION
BY TYPE OF OFFERING, 1996-2003
1996 1997 1998 1999
Public offerings * 114.6 110.9 121.3 125
Private placements * 43.2 61.9 84.7 112.7
Total * 157.8 172.8 206 237.7
% of Private placements 27.4% 35.8% 41.1% 47.4%
2000 2001 2002 2003
Public offerings * 134.9 120.9 110.4 123.3
Private placements * 177 100 60.5 58.8
Total * 311.9 220.9 170.9 182.1
% of Private placements 56.7% 45.3% 35.4% 32.3%
Footnotes:
Source: Board of Governors of the Federal Reserve System, Federal
Reserve Bulletin.
REFERENCES
Aboody, David and Ron Kasznik (2000). CEO Stock Option Awards and
the Timing of Corporate Voluntary Disclosures. Journal of Accounting and
Economics 29: 73-100.
Aharony, Joseph, Chan-Jane Lin, and Martin P. Loeb (1993). Initial
Public Offerings, Accounting Choices, and Earnings Management.
Contemporary Accounting Research 10: 61-81.
Barber, Brad M. and John D. Lyon (1997). Detecting Long-Run
Abnormal Stock Returns: The Empirical Power and Specification of Test
Statistics. Journal of Financial Economics 43: 341-372.
Barclay, Michael J., Clifford G. Holderness, and Dennis P. Sheehan
(2005). Private Placements and Managerial Entrenchment. Working paper,
University of Rochester.
Barton, Jan and Paul J. Simko (2002). The Balance Sheet as an
Earnings Management Constraint. Accounting Review 77: 1-27.
Bartov, E., F. Gul, and J. Tsui (2000). Discretionary-Accruals
Models and Audit Qualifications. Journal of Accounting and Economics 30:
421-452.
Beuselinck, Christof, Marc Deloof, and Sophie Manigart (2005).
Private Equity and Earnings Quility. Working paper, Ghent University.
Brennan, M. J. and J. Franks (1997). Underpricing, Ownership and
Control in Initial Public Offerings of Equity Securities in the UK.
Journal of Financial Economics, 45: 391-413.
Daniel, Kent, David Hirshleifer, and Avanidhar Subrhmanyam (1998).
Investor Psychology and Security Market under- and Overreactions.
Journal of Finance 53 (6): 1839-1885.
Dechow, P., R. Sloan, and A. Sweeney (1995). Detecting Earnings
Management. Accounting Review 70: 193-225.
Dechow, Patricia and Douglas J. Skinner (2000). Earnings
Management: Reconciling the views of Accounting Academics,
Practitioners, and Regulators. Accounting Horizons 14 (2): 235-250.
DuCharme, Larry L., Paul H. Malatesta, and Stephan E. Sefcik
(2001). Earnings Management: IPO Valuation and Subsequent Performance.
Journal of Accounting, Auditing & Finance 16: 369-396.
DuCharme, L. L., P. H. Malatesta and S. E. Sefcik (2004). Earnings
Management, Stock Issues, and Shareholder Lawsuits. Journal of Financial
Economics 71: 27-49.
Erickson, Merle and Shiing-wu Wang (1999). Earnings Management by
Acquiring Firms in Stock for Stock Mergers. Journal of Accounting and
Economics 27: 149-176.
Field, Laura Casares and Dennis P. Sheehan (2004). IPO Underpricing
and Outside Blockholdings. Journal of Corporate Finance 10 (2): 263-280.
Folta, Timothy B. and Jay J. Janney (2004). Strategic Benefits to
Firms Issuing Private Equity Placements. Strategic Management Journal
25: 223-242.
Frankel, Richard, Maureen McNichols, and G. Peter Wilson (1995).
Discretionary Disclosure and External Financing. Accounting Review 70:
135-150.
Friedlan, John (1994). Accounting Choices of Issuers of Initial
Public Offerings. Contemporary Accounting Research 11 (1): 1-31.
Healy, Paul M. and James M. Wahlen (1999). A Review of the Earnings
Management Literature and Its Implications for Standard Setting.
Accounting Horizons 13 (4): 365-383.
Hertzel, Michael and Richard L. Smith (1993). Market Discounts and
Shareholder Gains for Placing Equity Privately. Journal of Finance 48
(2): 459-485.
Hertzel, Michael, Michael Lemmon, James S. Linck, and Lynn Rees
(2002). Long-Run Performance Following Private Placements of Equity.
Journal of Finance 57 (6): 2595-2617.
Janney, Jay J. and Timothy B. Folta (2006). Moderating Effects of
Investor Experience on the Signaling Value of Private Equity Placements.
Journal of Business Venturing 21: 27-44.
Jones, J. (1991). Earnings Management during Import Relief
Investigations. Journal of Accounting Research 29: 193228.
Jung, Kooyul, Young-Cheol Kim, and Rene M. Stulz (1996). Timing,
Investment Opportunities, Managerial Discretion, and the Security Issue
Decision. Journal of Financial Economics 42: 159-185.
Keown, Arthur, David F. Scott, Jr., John D. Martin, and J. William
Petty (2003). Foundations of Finance. Prentice Hall.
Kothari, S.P., and J. Warner (1997). Measuring Long-Horizon
Security Price Performance. Journal of Financial Economics 49:
1541-1578.
Kothari, S.P., Leonge, Andrew J. and Wasley, Charles E. (2005).
Performance Matched Discretionary Accrual Measures. Journal of
Accounting and Economics 39: 163-197
Larcker, David F., Scott A. Richardson, and Irem Tuna (2005). How
Important is Corporate Governance? Working paper presented in 2005 AAA
Annual Conference, University of Pennsylvania.
Li, Kai and N.R. Prabhala (2007). Self-Selection Models in
Corporate Finance. Hand Book of Corporate Finance, chapter two.
Loughran, Tim and Jay R. Ritter (1995). The New Issues Puzzle.
Journal of Finance 50: 23-51.
Marciukaityte, Dalia (2001). Issues in Public and Private Equity
Offerings. Unpublished doctoral dissertation, Drexel University.
Margetis, Speros L. (2004). Two Essay on Earnings Management:
Earnings Management Surrounding Straight and Convertible Debt. And the
Intra-Industry Contagion Effect of Earnings Restatements. Unpublished
doctoral dissertation, University of South Florida.
McLaughlin, Robyn, Assem Safieddine, and Gopala K. Vasudevan
(1996). The Operating Performance of Seasoned Equity Issuers: Free Cash
Flow and Post-Issue Performance. Financial Management 25 (4): 41-53.
Murray, D. (1983). The Effect of Certain Research Design Choices on
the Assessment of the Market's Reaction to LIFO Changes: a
Methodological Study. Journal of Accounting Research 21 (1): 128-140.
Myers, Stewart C. and Nocholas S. Majluf (1984). Corporate
Financing and Investment Decisions When Firms Have Information That
Investors Do Not Have. Journal of Financial Economics 13: 187-221.
Rangan, Srinivasan (1998). Earnings Management and the Performance
of Seasoned Equity Offerings. Journal of Financial Economics 50:
101-122.
Rosenbaum, Paul R. and Donald B. Rubin (1983). The Central Role of
the Propensity Score in Observational Studies for Causal Effects.
Biometrika 70: 41-55.
Ross, Stephen A., Randolph W. Westerfield, and Jeffrey Jaffe
(2002). Corporate Finance (Fifth edition), Irwin McGraw-Hill.
Ruland, William, Samuel Tung, and Nashwa E. George (1990). Factors
Associated with the Disclosure of Managers' Forecasts. Accounting
Review 65: 710-721.
Schultz, Paul (2003). Pseudo Market Timing and the Long-Run
Underperformance of IPOs. Journal of Finance 58 (2): 483-518.
Shleifer, Andrei and Robert W. Vishny (1997). A survey of Corporate
Governance. Journal of Finance 52 (2): 737-783. Smith, Clifford W. Jr.
(1986). Investment Banking and the Capital Acquisition Process. Journal
of Financial Economics 15: 3-29.
Spiess, D. Katherine and John Affleck-Graves (1995).
Underperformance in long-run stock returns following seasoned equity
offerings. Journal of Financial Economics 38: 243-267.
Srivastava, Ashwani Kumar (1989). Private Equity Placements: An
Investigation of Discounts, Price Effects, and Floatation Costs.
Unpublished doctoral dissertation, University of Toronto.
Subramanyam, K. (1996). The Pricing of Discretionary Accruals.
Journal of Accounting and Economics 22: 249-282.
Teoh, Siew Hong, Ivo Welch, and T.J. Wong (1998a). Earnings
Management and the Long Term Market Performance of Initial Public
Offerings, Journal of Finance 53: 1935-1974.
Teoh, S.H., I. Welch and T.J. Wong (1998b). Earnings Management and
the Underperformance of Seasoned Equity Offerings. Journal of Financial
Economics 50: 63-99.
Watts, Ross L., and Jerold L. Zimmerman (1986). Positive Accounting
Theory. Prentice Hall.
Wruck, Karen Hopper (1989). Equity Ownership Concentration and Firm
Value--Evidence From Private Equity Financings. Journal of Financial
Economics 23: 3-28.
Wu, YiLin (2004). The Choice of Equity Selling Mechanisms. Journal
of Financial Economics 74: 93-119.
Daoping He, San Jose State University
David C. Yang, University of Hawaii at Manoa
Liming Guan, University of Hawaii at Manoa
Table 1: Summary of U.S. private placements sample size and gross
proceeds
Panel A: Sample size and gross proceeds by calendar year
Gross
proceeds
Sample (million
Year size Percentage $)
1989 17 4.89% 1,214.65
1990 13 3.74% 688.74
1991 14 4.02% 213.50
1992 25 7.18% 275.50
1993 25 7.18% 324.25
1994 19 5.46% 272.27
1995 16 4.60% 474.40
1996 15 4.31% 199.65
1997 16 4.60% 464.96
1998 7 2.01% 169.33
1999 18 5.17% 872.46
2000 40 11.49% 1,710.40
2001 123 35.34% 2,675.25
Total 348 100.00% 9,556.08
Panel B: Sample size and gross proceeds by industry classification
Gross
proceeds
Sample (million
Industry classification size Percentage $)
Mining 9 2.59% 221.76
Oil And Gas 6 1.72% 438.42
Food Products 6 1.72% 368.10
Chemical Products 94 27.01% 2,218.40
Computer Equipment 16 4.60% 431.52
Electronic Equipment 25 7.18% 299.00
Transportation 18 5.17% 470.88
Instruments And Related Product 38 10.92% 677.54
Communications 12 3.45% 294.48
Wholesale 6 1.72% 69.90
Retail 21 6.03% 893.97
Panel A: Sample size and gross proceeds by calendar year
Gross
proceeds
Sample (million
Year size Percentage $)
Financial Services 12 3.45% 482.88
Services 67 19.25% 1,886.05
Others 18 5.17% 801.54
Total 348 100.00% 9,556.08
Table 2: Discretionary accruals for U.S. issuing firms versus
control firms
Std.
Variable Mean Dev.
Year -2 (N=289)
PP firms -0.0021 0.3535
(t=-0.10, pr=0.920)
Control firms 0.0044 0.2920
(t=0.24, pr=0.811)
Test of difference (t=-0.23, p=0.815)
Year -1 (N=348)
PP firms 0.0327 ** 0.2816
(t=2.15, pr=0.032)
Control firms -0.0072 0.2125
(t=-0.64, pr=0.521)
Test of difference (t=2.11, pr=0.035)
Year 0 (N=321)
PP firms 0.0054 0.2466
(t=0.39, pr=0.694)
Control firms -0.0018 0.1991
(t=-0.16, pr=0.874)
Test of difference (t=0.40, p=0.687)
Year +1 (N=291)
PP firms 0.0127 0.1805
(t=1.20, pr=0.231)
Control firms -0.0020 0.2189
(t=-0.14, pr=0.886)
Test of difference (t=0.83, p=0.405)
Year +2 (N=260)
PP firms 0.0193 0.1963
(t=1.58, pr=0.114)
Control firms -0.0130 0.1855
(t=-0.95, pr=0.343)
Test of difference (t=1.76, p=0.079)
Variable Median Min Max
Year -2 (N=289)
PP firms 0.0104 -1.7926 1.2482
(pr=0.281)
Control firms 0.0045 -1.2817 1.6583
(pr=0.444)
Test of difference (pr=0.672)
Year -1 (N=348)
PP firms 0.0249 *** -1.4318 1.5114
(pr=0.001)
Control firms 0.0051 -0.8906 0.8037
(pr=0.782)
Test of difference (pr=0.059)
Year 0 (N=321)
PP firms 0.0128 -1.1477 1.5460
(pr=0.307)
Control firms 0.0055 -1.4221 0.7863
(pr=0.281)
Test of difference (pr=0.865)
Year +1 (N=291)
PP firms 0.0143 * -0.6056 0.8132
(pr=0.066)
Control firms -0.0080 -1.1885 1.5788
(pr=0.627)
Test of difference (pr=0.157)
Year +2 (N=260)
PP firms 0.0068 -1.2727 1.1556
(pr=0.108)
Control firms 0.0023 -0.8277 0.8948
(pr=0.553)
Test of difference (pr=0.187)
Footnotes:
Paired-sample t-test is used to evaluate difference in means, and
Wilcoxon rank-sum test is used to evaluate the difference in
medians.
***, **, *: Significant different from zero at 0.01, 0.05, and 0.10,
respectively, two-tailed test.
Table 3: Post-issue stock returns for extreme discretionary accruals
quartiles: U.S. issuers
Years after Raw returns Market excess returns
issuance
Conservative Aggressive Conservative Aggressive
1 -11.73 -6.72 -8.49 -3.78
2 19.25 -5.03 12.91 -13.08
3 18.22 3.54
Years after Abnormal returns
issuance
Conservative Aggressive
1 -23.16 -27.67
2 -21.49 -41.66
3 -51.21
Table 4: Regressions of post-issue stock performance on
discretionary accruals in year -1 and control variables for U.S.
issuers
Three-year Three-year Three-year
raw returns market abnormal
excess returns
returns
Discretionary Coef -0.2327 -0.3485 ** -0.4343 **
Accruals (t) (-1.22) (-1.96) (-1.83)
Market Value (t) -0.12 (-0.06) -0.48
Book to Market (t) -1.98 -1.76 -1.25
Industry dummies Not reported Not reported Not reported
Obs 293 293 293
R-square 3.31% 3.81% 4.17%
Footnotes:
***, **, *: Significant different from zero at 0.01, 0.05, and 0.10,
respectively , one-tailed test
Table 5: Logistic analysis of private placement decision
Independent variables Coefficient z-stat
Return on Assets -0.0906 ** -2.1
Leverage 0.0694 0.5
Size -0.0855 *** -2.83
R & D/ Assets 0.4763 ** 2.54
Book to Market -0.1761 *** -3.88
Sales/Assets -0.5164 *** -5.06
Traded on American Stock Exchange 0.3466 1.52
Traded on NASDAQ 0.2589 * 1.89
Industry Dummies 3 are significant
Year Dummies 5 are significant
Intercept -4.5894 *** -10.82
Number of Obs 36942
Pseudo R-square 11.59%
Footnotes:
***, **, *: Significant different from zero at 0.01, 0.05, and 0.10,
respectively.
The dependent variable is 1 if a firm issues private placement in a
certain year and 0 otherwise.
Table 6: Discretionary accruals for U.S. private placement firms
versus control firms--propensity score matching
Variable Mean Std. Dev.
Year -2 (N=289)
PP firms -0.0021 0.3535
(t=-0.10, pr=0.920)
Control firms 0.0081 0.2920
(t=0.52, pr=0.601)
Test of difference (t=-0.39, p=0.694)
Year -1 (N=348)
PP firms 0.0327 ** 0.2816
(t=2.15, pr=0.032)
Control firms -0.0100 0.2125
(t=-0.56, pr=0.577)
Test of difference (t=1.82, pr=0.069)
Year 0 (N=321)
PP firms 0.0054 0.2466
(t=0.39, pr=0.694)
Control firms 0.0050 0.1991
(t=-0.39, pr=0.695)
Test of difference (t=0.02, p=0.983)
Year +1 (N=291)
PP firms 0.0127 0.1805
(t=1.20, pr=0.231)
Control firms -0.0102 0.2189
(t=-0.78, pr=0.435)
Test of difference (t=1.36, p=0.173)
Year +2 (N=260)
PP firms 0.0193 0.1963
(t=1.58, pr=0.114)
Control firms -0.0072 0.1855
(t=-0.78, pr=0.434)
Test of difference (t=1.73, p=0.083)
Variable Median Min Max
Year -2 (N=289)
PP firms 0.0104 -1.7926 1.2482
(pr=0.281)
Control firms 0.0102 * -1.7113 1.2321
(pr=0.059)
Test of difference (pr=0.869)
Year -1 (N=348)
PP firms 0.0249 *** -1.4318 1.5114
(pr=0.001)
Control firms 0.0062 -1.7070 1.5687
(pr=0.957)
Test of difference (pr=0.042)
Year 0 (N=321)
PP firms 0.0128 -1.1477 1.5460
(pr=0.307)
Control firms 0.0060 -1.4365 1.1002
(pr=0.420)
Test of difference (pr=0.662)
Year +1 (N=291)
PP firms 0.0143 * -0.6056 0.8132
(pr=0.066)
Control firms -0.0030 -0.9299 1.1316
(pr=0.996)
Test of difference (pr=0.171)
Year +2 (N=260)
PP firms 0.0068 -1.2727 1.1556
(pr=0.108)
Control firms -0.0032 -0.6795 0.5421
(pr=0.970)
Test of difference (pr=0.526)
Footnotes:
Paired-sample t-test is used to evaluate difference in means, and
Wilcoxon rank-sum test is used to evaluate the difference in
medians.
***, **, *: Significant different from zero at 0.01, 0.05, and 0.10,
respectively, two-tailed test.
Table 7: Regressions of post-issue stock performance on
discretionary accruals in year -1 and control variables for U.S.
issuers--propensity score matching
Three-year Three-year Three-year
raw returns market abnormal
excess returns
returns
Discretionary Coef -0.2327 -0.3485 ** -0.5082 *
Accruals (t) (-1.22) (-1.96) (-1.43)
Market Value (t) -0.12 (-0.06) (-0.05)
Book to Market (t) -1.98 -1.76 -0.04
Industry dummies Not reported Not reported Not reported
Obs 293 293 293
R-square 3.31% 3.81% 4.28%
Footnotes:
***, **, *: Significant different from zero at 0.01, 0.05, and 0.10,
respectively, one-tailed test.
