-
Journal of International Accounting, Auditing and Taxation 21
(2012) 127 144
Contents lists available at SciVerse ScienceDirect
Journal of International Accounting,Auditing and Taxation
Accounting conservatism and IPO underpricing: China evidence
Z. Jun Lin , Zhimin TianDepartment of Accountancy & Law,
Hong Kong Baptist University, Renfrew Road, Kowloon Tong, Kowloon,
Hong Kong Special Administrative Region
a b s t r a c t
This study examines how accounting conservatism impacts
underpricing of initial public offerings (IPOs) in the Chinese
stock market. Inaddition, we investigate how information asymmetry
affects the association of accounting conservatism with IPO
underpricing. Based onregression analysis of 674 A-shares companies
that went public through IPOs at both Shanghai and Shenzhen Stock
Exchanges in Chinaduring 20012009, we nd that (1) accounting
conservatism is negatively associated with the magnitude of IPO
underpricing; and (2)the relationship between accounting
conservatism and IPO underpricing is more pronounced when
information asymmetry is high. Thendings should shed a light on
what drives IPO underpricing and how it could be affected by
accounting conservatism in an emergingeconomy.
2012 Elsevier Inc. All rights reserved.
1. Introduction
Going public marks an important watershed in the life of a young
company that provides access to public equity capital andlowers the
cost of funding for the companys operation and investment. However,
many studies nd that when companies gopublic, their share prices
jump substantially on the rst day of trading, which is termed as
IPO1 underpricing. For instance,Stoll and Curley (1970), Logue
(1973), and Ibbotson (1975) document a systematic increase from the
offer price to the rst dayclosing price in the U.S. market;
Ljungqvist (2007) nds evidence of underpricing in a range of
countries including European,Asia-Pacic and Latin American
countries and the U.S. However, rms are worse off due to IPO
underpricing although insidersmay maximize their wealth in the IPO
process (Kennedy, Sivakumar, & Vetzal, 2006). Chi and Padgett
(2002) also report thatIPO underpricing is negatively related to
long-run performance of the issuing rms. These phenomena motivate a
stream ofresearch on IPO underpricing and its determinants and
consequences.
Among various theories of IPO underpricing, information
asymmetry models play a dominant role (Ljungqvist, 2007).Asymmetric
information models assume that one of the parties involved in an
IPO knows more about the issuing rm thanthe others. So underpricing
is necessary to reach equilibrium on the interests of all
participants. As information asymmetrytheory plays an extremely
important role in the underpricing of IPO rms, the veriability
criteria for the recognition of rmsoperating performance and
nancial reporting should matter. Accounting conservatism requires
more stringent standardsof verication for economic gains than
losses (Basu, 1997). Thus, the earnings and net assets reported
under conservatismwill be more reliable and veriable, and
accounting conservatism will constrain the managements
opportunistic behaviorsfor overstating income and understating
potential losses (Bushman & Piotroski, 2006). As a result,
information asymmetrybetween issuers and other stakeholders of IPO
rms under conservative accounting is relatively lower, which should
resultin less IPO underpricing.
The important role of accounting conservatism in reducing
information asymmetry and IPO underpricing leads to ourrst
hypothesis, i.e., accounting conservatism is negatively associated
with IPO underpricing. In addition, a higher degree
Corresponding author. Tel.: +852 3411 7537; fax: +852 3411
5581.E-mail address: [email protected] (Z.J. Lin).
1 An initial public offering (IPO), referred simply as an
offering or otation, is when a company (called the issuer) issues
its common stocks or sharesto the public for the rst time.
1061-9518/$ see front matter 2012 Elsevier Inc. All rights
reserved.http://dx.doi.org/10.1016/j.intaccaudtax.2012.07.003
-
128 Z.J. Lin, Z. Tian / Journal of International Accounting,
Auditing and Taxation 21 (2012) 127 144
of accounting conservatism is needed in a situation with greater
information asymmetry in order to reduce the likelihoodof managers
manipulation and enhance information quality. Thus, our second
hypothesis predicts that the association ofaccounting conservatism
with IPO underpricing is more pronounced for rms with high
information asymmetry. Using alarge sample of A-share companies2
that went public through initial public offerings on both Shanghai
and Shenzhen StockExchanges in China during 20012009, we obtain
empirical evidence to support the two hypotheses.
Our study contributes to the extant literature in several ways.
First, by revealing the impacts of accounting conservatismon IPO
underpricing in the Chinese context, this research supplements a
ourishing stream of empirical research on account-ing conservatism
showing that conditional conservatism is associated with positive
outcomes (Ahmed & Duellman, 2007;Ahmed, Billings, Morton, &
Stanford-Harris, 2002; Garca Lara, Garca Osma, & Penalva, 2009;
Lafond & Watts, 2008). Ourndings suggest that accounting
conservatism can help reduce IPO underpricing, which should have
positive implicationson the effective functioning of capital
markets. In addition, our within-country study will corroborate
that of Boulton, Smart,and Zutter (2011) who document a signicant
role of earnings quality in IPO underpricing in their cross-country
analyses.
Second, this study provides standard setters with additional
insight into the role of accounting conservatism in improvingthe
quality of accounting information in corporate nancial reporting.
China introduced a new set of Chinese AccountingStandards (CAS) in
February 2006, which became mandatory for all listed companies on
January 1, 2007. The new CASsubstantially changes the nations old
accounting system and covers nearly all topics under the current
IFRS.3 However,there is a debate in China on whether the
constrictive role of accounting conservatism can be relieved by
incorporating thefair value provisions of IFRS. In addition, the
institutional structure in China, especially the involvement of
governmentalequity holding and the underdeveloped asset appraisal
market, gives managers incentives to overstate earnings in
nancialreporting (Ball, Robin, & Wu, 2003). Both the new CAS
and institutional settings in China led to less conservative
accounting,which may have a negative impact on the quality of
accounting information (Ahmed, Neel, & Wang, 2010). By
exploringthe signicant role of accounting conservatism in the IPO
process, the results of this study suggest that the adoption
ofconservatism principle in accounting is necessary and important
in China where there is a large degree of informationasymmetry.
Third, this study contributes to the literature on the
determinants of IPO underpricing by controlling for both
deal-specicand rm-specic characteristics of IPO rms in China. As
indicated by our study results, IPO underpricing is
negativelyrelated to offering size, proportion of tradable shares,
integer offer price, location of the exchange listed, rm leverage
andprotability, while it is positively associated with government
(state) ownership, volume of new issues in the market, lapseperiod
of offering and listing, and underwriter reputation in the Chinese
market. The ndings should have both theoreticaland practical
implications on the development of the IPO mechanism in the Chinese
stock market.
Furthermore, this study helps address a void in the nance
literature on IPOs, i.e., how can the quality of
accountinginformation have a direct impact on IPO underpricing. It
is anticipated that this study should stimulate more research onthe
interactive effects of accounting conservatism and information
asymmetry, which will generate more convincing androbust empirical
evidence to explain IPO underpricing in both developed and less
developed markets around the world.
The rest of this paper is organized as follows. Section 2
outlines the study background and relevant literature, anddevelops
our hypotheses. Section 3 describes the research design and data.
Section 4 reports the empirical results, andSection 5 concludes
this study.
2. Study background and hypotheses
2.1. IPO market in China
In pace with the progress of wide-scaled economic reforms since
the beginning of the 1980s, the share capital system orstock
companies were restored in China to restructure the inefcient
state-owned enterprises (SOEs). The Chinese govern-ment reopened
capital markets by setting up the Shanghai Stock Exchange (SHSE)
and Shenzhen Stock Exchange (SZSE) in1990 and 1991, respectively,
to facilitate listing and trading of stocks in the country (Lin
& Chen, 2005). However, IPOs inChina have the following
characteristics of new issuances that are distinct from those in
most developed countries.
First, before mid-1999, a quota system for IPOs was implemented
by the Chinese authorities. Under the quota system,the State
Planning Commission, in conjunction with the Peoples Bank of China
(the central bank) and the China SecuritiesRegulatory Commission
(CSRC), sets up an annual quota for new stocks to be issued each
year, which is then allocated toprovincial governments or
ministerial authorities to select IPO rms under their jurisdiction.
The quota system was formallyabolished in mid-1999 and investment
banks or securities brokerage rms have assumed greater
responsibility in identifyingand developing listing candidates.
2 The Chinese stock market was segregated for domestic and
overseas investors when the two stock exchanges, i.e., Shanghai
Stock Exchange andShenzhen Stock Exchange, were set up in early
1990s. A-shares, denominated in RMB (Chinese currency), are
designated only for domestic investors whileB-share, denominated in
US$ and HK$ in Shanghai Stock Exchange and Shenzhen Exchange
respectively, are issued to overseas investors. A great majorityof
Chinese listed rms are A-share rms at present.
3 International Financial Reporting Standards (IFRS) are
principles-based standards, interpretations and the conceptual
framework adopted by theInternational Accounting Standards Board
(IASB).
-
Z.J. Lin, Z. Tian / Journal of International Accounting,
Auditing and Taxation 21 (2012) 127 144 129
Second, the shares of Chinese companies are split into
state-owned shares, legal-entity-owned shares, and tradableshares.
The state shares are those directly owned by the central or local
governments. Legal-entity shares are those held bydomestic entities
or nancial institutions (many of them are SOEs as well). Tradable
shares were the only class of shares thatcould be traded on stock
exchanges before 2006,4 and are further classied as A-, B-, H-, S-,
N-, and L-shares.5 Regardless ofthe share type, each share is
entitled to the same cash ow and voting right.
Third, before 2005, the offer price of an IPO rm was set months
before market trading started, and in a great majority ofofferings
there was no feedback mechanism through market demand that allowed
adjustment to the offer price. However,since 2005, the book
building mechanism has been adopted to replace the original xed
price mechanism. Book buildingmechanism involves underwriters in
eliciting indications of interest from informed investors, which
are then used in settingthe offer price (Sherman, 2005). Thus, the
information asymmetry among investors will be mitigated, which, in
turn, reducesthe level of underpricing. The implementation of book
building mechanism indicates Chinas adaptation to
internationalpractices. However, whether the book building
mechanism is efcient depends on the institutional features of
China. Forexample, Li (2009) argues that, because of weak market
regulation and surveillance in China, book building offers
issuersgreater discretion, gives institutional investors more
chances to pursue their own benets, and may result in a higher
levelof underpricing in China.
Therefore, Chinese IPOs provide an interesting case study for
several reasons. First, as stated earlier, the Chinese govern-ment
plays a dominant role in IPO processes. The majority of shares of
IPO rms are owned by the state or other legal entitiesand are not
available to the public. Public investors rush to submit
applications whenever there is a new issuing, creatinga huge demand
for new issues and pushing up stock returns on the initial trading
day. Second, in typical Chinese IPOs,the regulators and many
issuers (i.e., SOEs) are controlled by government, and both the
managers of issuing rms and IPOunderwriters have close ties with
the government. They have an incentive to hide inefciencies and
conduct rent-seekingbehaviors, which aggravates information
asymmetry between issuers and potential investors and leads to
substantial IPOunderpricing. Third, in the Chinese stock market, a
dominant portion of investors are individual investors who do not
haveaccess to sufcient information or do not have sufcient
knowledge or experience of investment. This situation
aggravatesinformation asymmetry between issuing rms and investors
and contributes to large underpricing of IPO stocks. Fourth,the
dominance of government ownership in Chinese rms (even with the
privatization of SOEs) has a signicant impacton Chinas transition
from a socialist country to a modern market economy (Cheung,
Ouyang, & Tan, 2009). Given that thestate-owned rms frequently
suffer severe governance problems (Rajan & Zingales, 2003;
Shleifer & Vishny, 1993), corporatemanagers and government
ofcials have an incentive to cover up governance-related
inefciencies and conduct rent-seekingbehaviors. Poor performance
resulting from these operating inefciencies and governance conicts
will impose a damage topersonal reputation or severe political cost
upon the managers and ofcials. This incentive of hiding information
or deferringthe disclosure of bad news increases information
asymmetry between issuers and investors in China.
Compared with the evidence from other countries, underpricing in
China is extremely obvious. Boulton et al. (2011)examine IPOs from
37 countries and show that IPO rms are underpriced in every
country, ranging from underpricing of120.7% in China to less than
2% in Argentina. Other markets in the U.S., Japan, and Malaysia
have average underpricingof 33.90%, 55.85%, 34.32%, respectively.
The rst-day returns of IPOs in the A-share market in China are well
above theaverage as similar evidence has been documented by prior
research that, on average, very large market-adjusted
rst-dayreturns exist in Chinese IPOs: 298% for the period 19921997
(Chen & Hui-Tzu, 2004); 129% for the period 19962000 (Chi&
Padgett, 2005); and 133.6% for the period 19922006 (Cheung et al.,
2009). A few studies have analyzed IPO underpricingin the Chinese
market with respect to the effect of information asymmetry. For
example, Mok and Hui (1998) report that theproxies for ex ante
uncertainty explain the pattern of A-share IPO returns during the
period 19901993. Chan, Wang, and Wei(2004) suggest that the
institutional setups in China could also magnify information
asymmetry and expand underpricingin the IPO process.
2.2. IPO underpricing and information asymmetry theory
There are three well-known asymmetric information models to
explain IPO underpricing in the literature.First, the best-known
asymmetric information model is the winners curse initially
presented by Rock (1986). Rock (1986)
assumes that some investors are better informed about the true
value of the shares on offer than investors in general, theissuing
rm, or its underwriting banks. Informed investors bid only for
attractively priced IPOs, whereas the uninformedinvestors bid
indiscriminately. This situation imposes a winners curse on
uninformed investors. Thus, the returns unin-formed investors
earned conditional on receiving a share allocation are below the
average underpricing returns. However,Rock (1986) argues that the
primary market is dependent on continuous participation of
uninformed investors, in the sense
4 Before 2006, state shares and legal-entities shares were not
permitted to be traded in the secondary market but could be
transferred only to anotherinstitution or person through
off-counter deals. They are permitted to be traded on stock
exchanges after 2006 when the government introduced the socalled
circulation of all shares reform.
5 A-shares are designated only for domestic investors and traded
on Shanghai and Shenzhen Stock Exchanges. B-, H-, S-, N-, and
L-shares are designatedonly for foreign investors to be traded on
securities exchanges in China (B-shares), in Hong Kong (H-shares),
in Singapore (S-share), on NYSE (N-shares), oron London Stock
Exchange (L-shares).
-
130 Z.J. Lin, Z. Tian / Journal of International Accounting,
Auditing and Taxation 21 (2012) 127 144
that the informed demand is insufcient to take up all shares on
offer even in attractive offerings. This case requires
thatconditional expected returns be non-negative so that the
uninformed investors can at least break even. In other words,
allIPOs must be underpriced in expectation.
Second, some researchers, Loughran and Ritter (2004) as the most
prominent ones, stress the dark side of underwritingbanks.
Investors may compete for allocation of underpriced stocks by
offering underwriters side-payments. Underwritersmay allocate
underpriced stocks to investors in order to induce them to reveal
their information truthfully in the bookbuilding process for IPOs.
Underwriters may also allocate underpriced stocks to the executives
at issuing companies inthe hope of winning their future investment
banking business (e.g., seasoned equity offering or SEO), a
practice known asspinning. In each case, underwriters stand to gain
from deliberately underpricing the issuers stocks.
The nal group of asymmetric information models emphasizes
information asymmetry between issuing rms and poten-tial investors
(Su & Fleisher, 1999). An issuing company has better
information about the present value or risk of its future cashows
than outside investors and underwriters do, so underpricing is used
to signal the companys true value. High-qualityrms have an
incentive to credibly signal their higher quality to the market in
order to raise capital on more advantageousterms (e.g., better
seasoned equity offerings). Low-quality rms will not send a signal
by underpricing because they do notexpect to recoup their
investment in underpricing through after-market SEOs. The best a
low-quality issuer can do is to takethe money and run when its
stocks are initially offered. Su and Fleisher (1999) also argue
that underpricing is a strategy forrms to signal their value to
potential investors.
2.3. Accounting conservatism and IPO underpricing
Since information asymmetry plays an extremely important role in
the IPO process, it is natural to ask whether thequality of
information matters. Boulton et al. (2011) investigate the
relationship between earnings quality and internationalIPO
underpricing and nd that IPOs are underpriced less in countries
where public rms produce higher quality earningsinformation.
However, as Boulton et al. (2011) stated in the limitation of their
study, for certain countries, it may be prob-lematic when using
earnings quality of existing public rms to proxy for earnings
quality in the same country rather thanusing earnings quality of
private rms prior to their IPOs.6 In fact, the earnings quality of
IPO rms is very different fromthat of the existing public rms due
to specic motivations for new issuers such as underpricing
incentives and reputationcosts. Besides, there are many studies
that investigate the relationship between IPO underpricing and
earnings quality fromthe perspective of earnings management
(Aharony, Wang, & Yuan, 2010; Ball & Shivakumar, 2008; Teoh
& Wong, 2002;Teoh, Welch, & Wong, 1998) and accounting
disclosures (Jog & Mcconomy, 2003; Leone, Rock, &
Willenborg, 2007; Schrand& Verrecchia, 2005). Few studies have
considered the direct effect of accounting conservatism, which is a
very importantattribute of earnings quality (Bushman and Piotroski,
2006).
Prior literature on accounting conservatism highlights the
benets of conservatism in contracting (Watts, 2003). Condi-tional
conservatism acts as a governance mechanism that benets both debt
and equity holders and increases rm value. Forexample, Kim and
Pevzner (2010) point out that conditional conservatism has
informational benets to shareholders andthey nd that higher current
conditional conservatism is associated with lower probability of
future bad news. In additionto these contracting benets, Lafond and
Watts (2008) argue that conservatism is expected to lower
information asym-metry between managers and outside investors.
According to Lafond and Watts (2008), the managers asymmetric
lossfunction provides them incentives to use their private
information to transfer wealth from investors to themselves by
over-stating nancial performance during their tenure at the rm.
Thus, information asymmetry also exists between managersand
investors. Conservative accounting could reduce information
asymmetry between managers and investors throughtwo potential
mechanisms. First, conservatism can provide investors of the best
possible non-stock-price hard summaryinformation on current
performance (Lafond & Watts, 2008). Hard information here
refers to veriable information. Asconservatism requires a higher
degree of veriability for gains than losses, the net result of
conservative accounting couldbe the provision of more veriable
information (Chaney, Faccio, & Parsley, 2011). Second, the hard
information provides abenchmark that makes it possible for
alternative soft sources (such as management forecasts and
voluntary disclosures) togenerate credible information. Investors
can compare different-source predictions to the hard numbers that
are eventuallyrealized, which could enable them to evaluate the
reliability of competing information sources. In this way, outside
investorswill know the true value of issuing rms and the
information asymmetry between informed and uninformed investors
andbetween issuers and underwriters. Thus the information asymmetry
between issuers and potential investors can be lessenedand the
level of IPO underpricing reduced. We, therefore, set the rst
hypothesis as:
H1. Accounting conservatism is negatively associated with IPO
underpricing in the Chinese market.
2.4. Accounting conservatism and information asymmetry
Our second hypothesis addresses the association of the magnitude
of information asymmetry with the impact of account-ing
conservatism on IPO underpricing. Information asymmetry aggravates
agency problems between two contractual parties
6 This problem exists in Boulton et al. (2011) because they
could not obtain the data on pre-IPO disclosures of private rms for
some countries.
-
Z.J. Lin, Z. Tian / Journal of International Accounting,
Auditing and Taxation 21 (2012) 127 144 131
(Haw, Lee, & Lee, 2010). Informed parties have more
incentives and opportunities to transfer wealth to themselves.
Thistendency is greater for pre-IPO rms owned by a small number of
large controlling shareholders (Burkart, Gromb, & Panunzi,1997;
Mueller & Inderst, 2001). Therefore, managers in these rms have
a greater latitude to indulge their own preferences,even at the
expense of shareholders interests (Francis & Martin, 2010).
Lafond and Watts (2008) argue that informationasymmetry between
insiders and outside investors lowers the rms stock price by
increasing the required rate of return onthe stock and generating
agency costs that reduce the rms expected cash ows. The more the
relatively private informa-tion, the larger the bid-ask spread and
the lower the returns to investors without private information
(Amihud & Mendelson,1986). The equilibrium return effect
provides managers and other insiders incentives to apply governance
mechanisms thatincrease public information, so reducing private
information and increasing stock price (Lafond & Watts, 2008).
However,managers have a motivation to use their private information
to maximize their own utilities, which may generate
deadweightlosses to shareholders and reduce the rms cash ows and
its stock price. Thus investors prefer rms to use veriable
infor-mation to restrict managements manipulation. As a result,
information asymmetry between insiders and outside
investorsgenerates incentives and demands for accounting
conservatism. There are, however, differences in corporate
governancesystems and managements control over accounting and
reporting processes among rms at varied development stages.Thus the
level of information asymmetry varies at different rms (Kennedy et
al., 2006; Rock, 1986). In rms with highinformation asymmetry,
conservative accounting can play a more signicant role in curbing
management manipulation ofaccounting information and improving the
credibility of nancial statements (Garca Lara et al., 2009), which
should con-tribute more substantially to the reduction of
information asymmetry between issuing rms and potential outside
investors(Ahmed et al., 2002). Therefore, for rms with greater
information asymmetry, accounting conservatism is likely to play
aneven more signicant role in reducing issuing rms incentives to
intentionally underprice their stocks. These argumentslead to the
second hypothesis as follows:
H2. The negative association between accounting conservatism and
IPO underpricing is more pronounced for rms withhigh information
asymmetry than for rms with low information asymmetry.
3. Research design and data
3.1. Measurement of accounting conservatism
Based on Givoly and Hayn (2000), we use the total accrual-based
measure of conservatism (CONS TA), which is the totalaccruals
deated by average total assets at the beginning of year, and
averaged over a 3-year period centered on the yearbefore the IPO,
multiplied by negative one. According to Givoly and Hayn (2000),
total accruals are dened as:
total accruals = (net income + depreciation) cash ow from
operations (1)
3.2. Measurement of IPO underpricing
Consistent with previous studies, we employ the methodology used
by Aggarwal, Leal, and Hernandez (1993) to mea-sure IPO
underpricing, i.e., we measure the market-adjusted abnormal returns
for the rst trading day to proxy for IPOunderpricing. The
measurement is described as follows.
The return of stock i at the end of the rst trading day is
calculated as:
Ri1 =(
Pi1Pi0
) 1 (2)
where Pi1 is the closing price of stock i on the rst trading
day, and Pi0 is the offering price and Ri1 is the total
rst-dayreturn on the issuing stock.
The return on market index for the corresponding time period
is:
Rm1 =(
Pm1Pm0
) 1 (3)
where Pm1 is the closing value of the corresponding Shanghai or
Shenzhen A-share market index on the rst trading dayand Pm0 is the
closing value of the corresponding Shanghai or Shenzhen A-share
market index on the offering day of thecorresponding stock, while
Rm1 is the rst-days comparable market return.
Using these two returns, the market-adjusted abnormal return for
each IPO on the rst trading day which we use tomeasure IPO
underpricing is computed as:
MAARi1 = 100 {[
(1 + Ri1)(1 + Rm1)
] 1
}(4)
In addition, we use the difference of market-to-book ratio
between that on the listing day and pre-offering (DMB)
toalternatively measure IPO underpricing (Chan et al., 2004). The
pre-offering MB ratio is calculated using the offering priceand net
assets per share one year before IPO. We assume that MB is higher
on the listing day than that in the pre-offeringperiod if IPO
underpricing exists, which will yield a positive DMB as its
magnitude denotes the level of underpricing.
-
132 Z.J. Lin, Z. Tian / Journal of International Accounting,
Auditing and Taxation 21 (2012) 127 144
3.3. Empirical model
We construct the following regression model to test our
hypotheses. IPO underpricing is the dependent vari-able and the
proxy of accounting conservatism is the key explanatory variable.
In addition, we incorporate severaldeal-specic and rm-specic
factors contextual to the distinct features in the Chinese IPO
market as controlvariables:
UPi = 0 + 1CONSi + 2BOOKi + 3CONSi BOOKi + 4LDAYi + 5OFF SIZEi +
6OWN Si + 7OWN Li + 8TSi+ 9VOLi + 10INTEGi + 11EXCHi + 12UWREPi +
13LEVi + 14ROAi + i (5)
where
UP (MAAR) measure of underpricing with market-adjusted initial
returns on the rst trading day of the IPO stockUP (DMB) measure of
underpricing with difference of market-to-book ratio between the
listing day and pre-offeringCONS (TA) proxy for accounting
conservatism with total accruals approach, which is measured by
income before extra-ordinary items
less cash ows from operations plus depreciation expense deated
by average total assets at the beginning of year, andaveraged over
a 3-year period centered on the year before IPO, multiplied by
negative one
BOOK dummy variable, equals to 1 if the rm goes public through
book building mechanism and 0 otherwiseCONS*BOOK the interaction
term of CONS and BOOKLDAY number of days between offering and
listing of an IPO stockOFF SIZE natural logarithm of the number of
offering shares multiplied by offering price, indicating the
offering size of individual IPO
stocksOWN S percentage of shares held by government and
state-owned legal entities after IPO, representing the proportion
of government
ownership in the equity of an issuing rmOWN L percentage of
shares held by non-state-owned legal entities after IPOTS
percentage of publicly tradable shares of each IPOVOL Number of
IPOs in each yearINTEG dummy variable, equals to 1 if the offer
price is an integer and 0 otherwiseEXCH dummy variable, equals to 1
if the new issue is listed on Shenzhen Stock Exchange and 0 if it
is listed on Shanghai Stock
ExchangeUWREP dummy variable, equals to 1 if the underwriter
belongs to Top 5 underwriters nationwide7 and 0 otherwiseLEV a rms
total debt divided by its total assets of the year before IPOROA
after-tax net income divided by total assets of the year before
IPO
Regarding the effect of accounting conservatism on IPO
underpricing, 1 is predicted to be negative for H1, thus
con-servative accounting should be able to reduce IPO underpricing.
In order to test H2, we partition all issuers into twosubgroups by
the extent of information asymmetry in pre-IPO period and analyze
whether the association betweenaccounting conservatism and IPO
underpricing is more pronounced when information asymmetry is high.
We mea-sure the magnitude of information asymmetry between issuers
and potential investors by four proxies that have beenused in prior
studies: sales growth, rm age, rm size, and corporate governance.
First, as stated by Khan and Watts(2009), information asymmetry
between managers and investors increases with growth options
because future cash owsfrom growth options are typically unveriable
and may produce more agency costs. Second, younger rms tend to
havemore growth options and less standardized accounting and
reporting systems relative to old rms and have a greaterscale of
information asymmetry (Stoll & Curley, 1970). Third, large rms
are more mature and have richer informationenvironments (e.g., more
analyst following), which contributes to less overall uncertainty
and information asymme-try associated with the realizability of
projected gains. Fourth, strong corporate governance rms will
better align theinterests of all stakeholders, which will mitigate
agency problems and result in relatively less information
asymmetry(Watts, 2003).
Sales growth is measured as the change in sales revenue divided
by one-year lagged sales revenue. Firm age is measured asthe number
of years passed since a rms establishment date before its IPO. Firm
size is measured as the total assets beforeIPO. Corporate
governance is measured by taking the unweighed average of
standardized composing variables includingownership concentration,
ratio of the number of independent directors in the Board,
separation of CEO/Chair positions, liti-gation risk, and leverage
risk (Burkart et al., 1997; Jog & Mcconomy, 2003). We use the
sample median of each partitioningproxy to divide the total sample
into high information asymmetry sample and low information
asymmetry sample. Firmswith high sales growth, young age, small rm
size and weak corporate governance are classied as high information
asym-metry rms, respectively. We run regressions (Eq. (5))
separately for the high- and low information asymmetry subsamplesto
test whether accounting conservatism has a varied impact on IPO
underpricing under different information asymmetryenvironments.
7 We rank underwriters according to the number of their IPO
deals annually. In general Top-5 underwriters represent
high-quality ones nationwide.
-
Z.J. Lin, Z. Tian / Journal of International Accounting,
Auditing and Taxation 21 (2012) 127 144 133
Table 1Sample selection.
Sample selection process Observations dropped Observations
remaining
Initial sample from 2001 to 2009 available from CSMAR 695After
eliminating rms in nancial industries 19 676After eliminating rms
that began their initial public offerings but have not been listed
on
stock exchanges2 674
6468
66
98
15
68
110
78
109
05
01
00
Nu
mb
er
of IP
Os b
y Y
ea
r
200920082007200620052004200320022001
Fig. 1. Number of IPOs in China by year (the nal sample includes
65 IPOs for 2003 and 77 IPOs for 2008. Two IPOs (each in 2003 and
2008 respectively)are eliminated from the sample because they have
not been listed on stock exchanges in the year although they began
their initial public offerings).
3.4. Data
Our sample includes all A-share companies that went public
through initial public offerings on both the Shanghai StockExchange
and Shenzhen Stock Exchange in China during 20012009.8 IPOs of
B-share companies are excluded because theirIPO behaviors and
performance are substantially different from those of A-share
companies due to varied regulatory require-ments (Chan et al.,
2004). We exclude companies in the nancial industry because the
nancial reporting environment fornancial institutions signicantly
differs from that of other industrial companies. Basic data about
IPOs, nancial perfor-mance, stock returns, and ownership
information are obtained from the database of Chinese Stock Market
and AccountingResearch (CSMAR). Following Chan et al. (2004), we
also exclude rms that have a lapsed time between offering date
andlisting date that exceeds 360 days.
Our initial sample extracted from the CSMAR database includes
695 IPOs. After excluding 19 companies in the nancialindustry and
two rms that began their initial public offerings but have not been
listed on stock exchanges, we have 674IPOs for the nal sample.9
Table 1 reports the rm-level sample selection process. The number
of new issues in each year isplotted in Fig. 1, which demonstrates
a considerable time-series volatility in the Chinese market. For
example, IPO activitypeaked in 2007 with a record of 110 rms going
public, but only 15 IPOs took place in 2005.
3.5. Descriptive statistics
Table 2 presents the descriptive statistics for the variables by
year. All variables are winsorized at the extreme 1% and 99%both to
mitigate the possible effect of outliers. The average magnitude of
IPO underpricing is 110.9% for the entire period,which is obviously
at a substantially high level. A breakdown of the IPOs by year
shows that IPO underpricing in the Chinesemarket undergoes two
stages. There is a decreasing trend in terms of IPO underpricing
from 145.0% in 2001 to 86.6% in 2006and from 192.7% in 2007 to
69.4% in 2009, respectively. In our sample, 54.7% of IPOs (369 of
674) went public through bookbuilding mechanism. The number of days
between offering and listing (LDAY) is 26.078 in 2001 and declines
to 13.982 in2009, indicating that the time lag between offering and
listing has become shorter and the Chinese IPO market has
becomemore efcient in recent years. The average offer size (OFF
SIZE) is RMB 831.465 million in 2001 (US$1 = RMB 6.5) andincreases
to RMB 1652.478 million in 2009, which indicates a negative
relationship with the level of underpricing. Thepercentage of state
shares (OWN S) or government ownership shows a steady decrease from
48.5% in 2001 to 12.0% in 2009.
8 In October 2000, Shenzhen Stock Exchange was temporarily
closed to IPOs, leaving Shanghai Stock Exchange as the only IPO
market from 2001 to 2003in China. Our sample starts from 2001 in
order to mitigate the effects arising from the suspension of IPO
activities on the Shenzhen that Stock Exchange.
9 All rms in the sample period have a lapsed time between
offering date and listing date of less than 360 days.
-
134Z.J.
Lin, Z.
Tian /
Journal of
International Accounting,
Auditing
and Taxation
21 (2012) 127 144Table 2Descriptive statistics for A-share IPOs
in China by year.
Year MAAR CONS BOOK LDAY OFF SIZE OWN S OWN L TS VOL INTEG EXCH
UWREP LEV ROA
2001
Mean 1.450 0.019 0.172 26.080 831.500 0.485 0.171 0.331 64 0.141
0.000 0.438 0.555 0.076Median 1.261 0.015 0 23.500 482.600 0.587
0.033 0.321 64 0.000 0.000 0.000 0.580 0.059SD 0.896 0.086 0.380
9.030 1494.000 0.243 0.234 0.064 0 0.350 0.000 0.500 0.112
0.045Obs. 64 64 64 64 64 64 64 64 64 64 64 64 64 64
2002
Mean 1.277 0.017 0.279 15.840 624.200 0.443 0.171 0.338 68 0.132
0.015 0.206 0.559 0.078Median 1.142 0.018 0 15.000 327.100 0.550
0.026 0.345 68 0.000 0.000 0.000 0.576 0.068SD 0.779 0.090 0.452
3.371 1419.000 0.254 0.231 0.072 0 0.341 0.121 0.407 0.113
0.045Obs. 68 68 68 68 68 68 68 68 68 68 68 68 68 68
2003
Mean 0.723 0.004 0.015 15.920 611.600 0.385 0.206 0.335 66 0.062
0.000 0.169 0.555 0.090Median 0.618 0.005 0 15.000 322.500 0.503
0.069 0.333 66 0.000 0.000 0.000 0.578 0.080SD 0.429 0.087 0.124
2.600 1273.000 0.283 0.261 0.069 0 0.242 0.000 0.378 0.121
0.050Obs. 65 65 65 65 65 65 65 65 65 65 65 65 65 65
2004
Mean 0.744 0.012 0 16.460 360.700 0.256 0.267 0.328 98 0.122
0.398 0.133 0.527 0.097Median 0.629 0.008 0 15.000 282.100 0.124
0.180 0.321 98 0.000 0.000 0.000 0.526 0.093SD 0.559 0.079 0 2.781
297.900 0.273 0.248 0.061 0 0.329 0.492 0.341 0.128 0.039Obs. 98 98
98 98 98 98 98 98 98 98 98 98 98 98
2005
Mean 0.505 0.018 0.267 16.200 384.200 0.251 0.436 0.258 15 0.000
0.800 0.200 0.549 0.093Median 0.487 0.017 0 15.000 285.700 0.048
0.580 0.265 15 0.000 1.000 0.000 0.516 0.086SD 0.329 0.060 0.458
4.004 443.000 0.301 0.295 0.039 0 0.000 0.414 0.414 0.153 0.056Obs.
15 15 15 15 15 15 15 15 15 15 15 15 15 15
2006
Mean 0.866 0.004 0.985 14.500 1056.000 0.245 0.364 0.235 68
0.132 0.824 0.324 0.567 0.079Median 0.674 0.004 1 14.000 309.700
0.007 0.264 0.205 68 0.000 1.000 0.000 0.593 0.071SD 0.709 0.071
0.121 3.239 2310.000 0.288 0.265 0.064 0 0.341 0.384 0.471 0.141
0.039Obs. 68 68 68 68 68 68 68 68 68 68 68 68 68 68
2007
Mean 1.927 0.014 0.909 13.440 2285.000 0.197 0.343 0.213 110
0.064 0.864 0.409 0.553 0.091Median 1.708 0.005 1 14.000 277.300 0
0.265 0.202 110 0.000 1.000 0.000 0.573 0.081SD 0.997 0.067 0.289
3.473 9329.000 0.273 0.286 0.061 0 0.245 0.345 0.494 0.135
0.040Obs. 110 110 110 110 110 110 110 110 110 110 110 110 110
110
2008
Mean 1.220 0.027 0.753 12.130 1382.000 0.158 0.326 0.201 78
0.013 0.922 0.416 0.569 0.100Median 0.901 0.027 1 12.000 321.600 0
0.264 0.201 78 0.000 1.000 0.000 0.579 0.086SD 0.904 0.070 0.434
3.254 4066.000 0.264 0.280 0.039 0 0.114 0.270 0.496 0.158
0.061Obs. 77 77 77 77 77 77 77 77 77 77 77 77 77 77
2009
Mean 0.694 0.054 1 13.980 1652.000 0.120 0.286 0.192 109 0.385
0.927 0.193 0.486 0.129Median 0.629 0.040 1 12.000 628.300 0 0.197
0.201 109 0.000 1.000 0.000 0.477 0.119SD 0.393 0.098 0 7.322
5364.000 0.230 0.256 0.033 0 0.489 0.262 0.396 0.155 0.069Obs. 109
109 109 109 109 109 109 109 109 109 109 109 109 109
Total
Mean 1.109 0.019 0.547 15.670 1166.000 0.265 0.278 0.264 85.240
0.138 0.556 0.280 0.543 0.095Median 0.858 0.011 1 15.000 349.100
0.110 0.171 0.235 78 0.000 1.000 0.000 0.561 0.086SD 0.854 0.083
0.498 6.075 4703.000 0.289 0.269 0.085 21.680 0.345 0.497 0.450
0.138 0.053Obs. 674 674 674 674 674 674 674 674 674 674 674 674 674
674
Note: Variables are dened as: MAAR, measure of underpricing with
market-adjusted initial returns on the rst trading day of IPO
stock; CONS, proxy for accounting conservatism, which is measured
by incomebefore extra-ordinary items less cash ows from operations
plus depreciation expense deated by average total assets at the
beginning of year, and averaged over a 3-year period centered on
the year before IPO,multiplied by negative one; BOOK, a dummy
variable, equals to 1 if the rm goes public through book building
mechanism and 0 otherwise; LDAY, number of days between offering
and listing of an IPO stock;OFF SIZE, natural logarithm of the
number of offering shares multiplied by offering price, indicating
the offering size of individual IPO stocks; OWN S, percentage of
shares held by government and state-ownedlegal entities after IPO,
representing the proportion of government ownership in the equity
of an issuing rm; OWN L, percentage of shares held by
non-state-owned legal entities after IPO; TS, percentage ofpublicly
tradable shares of each IPO; VOL, number of IPOs in each year;
INTEG, a dummy variable, equals to 1 if the offer price is an
integer and 0 otherwise; EXCH, a dummy variable, equals to 1 if the
new issueis listed on Shenzhen Stock Exchange and 0 if it is listed
on Shanghai Stock Exchange; UWREP, a dummy variable, equals to 1 if
the underwriter belongs to Top 5 underwriters nationwide and 0
otherwise; LEV, arms total debt divided by its total assets of the
year before IPO; ROA, after-tax net income divided by total assets
of the year before IPO.This table presents the description of
variables for the entire sample of 674 IPOs by year. All variables
are winsorized at the extreme 1% and 99%.
-
Z.J. Lin, Z. Tian / Journal of International Accounting,
Auditing and Taxation 21 (2012) 127 144 135
It should be noted that the shares owned by the state and legal
entities account for more than 50% of the total shares forthe
listed companies in the 20012009 period. This situation may
contribute to operational inefciencies and informationasymmetry
that aggravate IPO underpricing. Analysis of INTEG, EXCH and UWREP
shows that 13.8% of IPOs (93 of 674) havean integer offer price,
55.6% of IPOs (375 of 674) are listed on the Shenzhen Stock
Exchange during the sample period, and28.0% of IPOs (189 of 674)
are underwritten by the Top 5 underwriters in China.
The correlations between all variables are presented in Table 3.
These estimates do not reveal correlations that aresufciently high.
So the multicollinearity problem is moderate in our model and will
not have a signicant impact on therelationship between the
dependent and independent variables used in Eq. (3).
4. Empirical results
4.1. The association of accounting conservatism with IPO
underpricing
Table 4 shows the results of cross-sectional analyses for the
impact of accounting conservatism on IPO underpricing inthe Chinese
market based on Eq. (3). We use total accruals (TA) to proxy for
accounting conservatism10 and use both market-adjusted abnormal
return (MAAR) and difference of market-to-book ratios (DMB) to
measure IPO underpricing. In Model1, the dependent variable is the
market-adjusted abnormal returns (MAAR) on the rst trading day. The
main independentvariable is accounting conservatism measured by
total accruals (CONS TA). The model is reasonably well specied
withadjusted R2 of 17.9% at the signicance level of 1% (F = 11.45).
As we predict, the coefcient on CONS TA is 1.045 at the
10%signicance level, suggesting that accounting conservatism is
negatively and signicantly related to IPO underpricing in
theChinese market. This nding supports H1.
Results in Model 1 also reveal that the offering mechanism does
affect underpricing. The book building mechanism(BOOK) is found to
be positively related to underpricing (MAAR) at the 5% signicance
level (2 = 0.196, t = 2.35), which isconsistent with Ljungqvist,
Jenkinson, and Wilhelm (2003) and Li (2009). Li (2009) explains the
positive relationship betweenthe book building mechanism and the
level of underpricing by stating that book building offers issuers
greater discretionsand gives institutional investors more chances
to pursue their own benets. Book building mechanism allows shares
to beallocated preferentially and enables investors to call for
changes in offer price that will give everyone a fair chance
(Sherman,2005). Besides, we include an interaction term for book
building mechanism and accounting conservatism. Model 1 showsa
positive relationship between the interaction (BOOK*CONS) and the
level of underpricing (MAAR) at the 10% signicancelevel (3 = 1.378,
t = 1.85), indicating that the relationship between accounting
conservatism and IPO underpricing is lesspronounced after the
implementation of book building mechanism. This result further
conrms the substitute relationshipbetween BOOK and CONS as shown in
the correlation matrix.
Other results in Model 1 are broadly consistent with the extant
IPO underpricing literature. For example, LDAY has a pos-itive
relationship to the level of underpricing, although its coefcient
is not signicant at the conventional level (4 = 0.002,t = 0.30).
This result can be explained by the specic sample period. Our
regression uses IPOs that took place in 20012009,and as shown in
Table 2, the average lapsed time between offering and listing
during the test period is 15.67 days. The timegap is not as
unreasonably long as in previous periods. It can be interpreted
that the level of underpricing will not be signi-cantly affected by
the time elapsed between offering and listing if issuers can list
their offerings within a reasonable durationperiod. Regarding the
offering size of IPOs (OFF SIZE), we nd a signicantly negative
relationship to the level of underpricingat the 1% signicance level
(5 = 0.402, t = 10.41), which is in line with prior studies stating
that smaller offer size makesit easier for institutional investors
to control the issuing process and have more money left on the
table for themselves. Withrespect to the impact of shareholders
structure, the results show a positive relationship between the
proportion of stateshares (OWN S) and IPO underpricing at the 5%
signicance level (6 = 0.360, t = 2.23)11 and a negative
relationship betweenthe proportion of publicly tradable shares (TS)
and IPO underpricing at the 5% signicance level (8 = 1.258, t =
2.45),suggesting that a high proportion of state-owned shares give
managers more opportunities to cover up inefciencies,
whichaggravates information asymmetry and IPO underpricing. This
result is consistent with Chaney et al. (2011) who argue
thatpolitically connected rms have incentives to hide inefciencies
and end up with a large degree of information asymmetry.With a high
state ownership, managers are subject to less monitoring and
discipline of public investors and they are eagerto cover up
inefciencies because they tend to pursue their personal interests
such as their bureaucratic careers. However,a high proportion of
tradable shares may reduce the demand for new issues and push down
the stock returns on the initialtrading day.
We nd a signicantly positive relationship between IPO volume
(VOL) and initial IPO returns (MAAR) at the 1% level(9 = 0.008, t =
4.72), which is in line with the hot issues market phenomenon.
Consistent with Banerjee, Dai, and Shrestha(2009) and Boulton et
al. (2011), integer offer price (INTEG) is negatively related to
IPO underpricing (MAAR) at the 5% signif-icance level (10 = 0.210,
t = 2.37). Harris (1991) argues that the use of integer offer price
intends to reduce uncertainty instock pricing and the frequency of
integer offer prices should increase in offerings. In this respect,
integer offer price nega-
10 We also use discretionary accruals to proxy for accounting
conservatism in robustness tests with reduced sample size owing to
missing data of somerms and the result is qualitatively the same
(see Section 4.4.).11 Cheung et al. (2009) also nd similar evidence
on the relationship between state ownership and IPO underpricing in
the Chinese market.
-
136Z.J.
Lin, Z.
Tian /
Journal of
International Accounting,
Auditing
and Taxation
21 (2012) 127 144
Table 3Correlation matrix among variables.
MAAR CONS BOOK LDAY OFF SIZE OWN S OWN L TS VOL INTEG EXCH UWREP
LEV ROA
MAAR 1TA 0.0116 1BOOK 0.0542 0.0622 1LDAY 0.0102 0.0368 0.311***
1OFF SIZE 0.316*** 0.0353 0.156*** 0.0475 1OWN S 0.0128 0.19***
0.232*** 0.169*** 0.288*** 1OWN L 0.0359 0.0538 0.159*** 0.116***
0.124*** 0.613*** 1TS 0.0393 0.101*** 0.578*** 0.27*** 0.125***
0.304*** 0.263*** 1VOL 0.111*** 0.125*** 0.362*** 0.241*** 0.0913**
0.268*** 0.0637* 0.354*** 1INTEG 0.0869** 0.0628 0.0785** 0.0503
0.0455 0.0367 0.0373 0.0858** 0.147*** 1EXCH 0.0782** 0.158***
0.556*** 0.303*** 0.243*** 0.559*** 0.311*** 0.617*** 0.401***
0.0888** 1UWREP 0.0836** 0.0235 0.0433 0.0234 0.122*** 0.0593 0.012
0.0575 0.0137 0.0486 0.0322 1LEV 0.0235 0.085** 0.0349 0.0101
0.0592 0.0405 0.0656* 0.03 0.118*** 0.0973** 0.084** 0.0522 1ROA
0.0889** 0.248*** 0.0881** 0.0095 0.0161 0.283*** 0.114*** 0.251***
0.195*** 0.127*** 0.25*** 0.0048 0.591*** 1
Note: This table reports the correlation of variables used.
Variables are dened as in Table 2. All variables are winsorized at
the extreme 1% and 99%.* Statistical signicance at the 10%
level.
** Statistical signicance at the 5% level.*** Statistical
signicance at the 1% level.
-
Z.J. Lin, Z. Tian / Journal of International Accounting,
Auditing and Taxation 21 (2012) 127 144 137
Table 4Results for the impact of accounting conservatism on IPO
underpricing in 20012009.
Predictions MAAR DMB
Intercept 3.537*** 0.940(8.44) (0.29)
CONS TA 1.045* 7.676*(1.86) (1.77)
BOOK ? 0.196** 2.338***
(2.35) (3.64)TA*BOOK ? 1.378* 1.917
(1.85) (0.33)LDAY + 0.002 0.165***
(0.30) (4.00)OFF SIZE 0.402*** 1.115***
(10.41) (3.74)OWN S + 0.360** 1.766
(2.23) (1.42)OWN L + 0.068 1.130
(0.46) (1.00)TS 1.258** 2.660
(2.45) (0.67)VOL + 0.008*** 0.070***
(4.72) (5.68)INTEG ? 0.210** 0.495
(2.37) (0.72)EXCH ? 0.234** 1.498*
(2.28) (1.89)UWREP ? 0.232*** 1.057**
(3.44) (2.03)LEV 0.416 8.577***
(1.49) (3.98)ROA 1.936** 44.985***
(2.45) (7.39)N 674 674R2 0.196 0.247Adj. R2 0.179 0.231F 11.45
15.48
Note: This table reports the cross-sectional analyses for the
impact of accounting conservatism on IPO underpricing based on Eq.
(3). Denition of variablesis as specied in Table 2. All variables
are winsorized at the extreme 1% and 99%. t-Statistics are reported
in parentheses.
* Statistical signicance at the 10% level.** Statistical
signicance at the 5% level.
*** Statistical signicance at the 1% level.
tively affects the magnitude of underpricing. The coefcient on
EXCH reveals that the IPO underpricing in the Shenzhen
StockExchange is signicantly lower than that in the Shanghai Stock
Exchange at the 5% signicance level (11 = 0.234, t = 2.28),which is
consistent with Yu and Tse (2006). The regulations and investors
behaviors in the Shenzhen Stock Exchange aremuch closer to those in
neighboring Hong Kong where international standards and practices
prevail, which may result inless information asymmetry between
issuers and potential investors for the Shenzhen Stock Exchange.
Besides, it is notedthat some rms on the Shenzhen Stock Exchange
are joint ventures with foreign capitals, while those listed on the
ShanghaiStock Exchange are mostly SOEs (Yu & Tse, 2006). There
are more disclosures and less uncertainty in joint venture rmswith
foreign partners, so the IPO rms listed on the Shenzhen Stock
Exchange are relatively less underpriced.
The magnitude of underpricing is positively related to the
reputation of underwriters (UWREP) at the 1% signicance level(12 =
0.232, t = 3.44), which is consistent with Boulton et al. (2011)
and Ritter and Welch (2002). This result further shows
thatunderwriters in China still play a major role in underpricing
in the process of initial public offerings. Large underwriters
willdemand high underpricing to pursue their own benets. The
coefcient on rm leverage (LEV) is 0.416, which supports
themonitoring role of leverage in reducing managements
opportunistic behaviors and information asymmetry between
issuersand potential investors. However, the t-statistic of this
variable is not signicant at the conventional level, indicating
thatthe role of leverage in reducing information asymmetry and IPO
underpricing is rather weak in China. Consistent with Rock(1986)
who argues that the degree of information asymmetry is a decreasing
function of rm protability, ROA is negativelyrelated to the level
of underpricing at the 5% signicance level (14 = 1.936, t = 2.45),
suggesting that protability helpsreduce IPO underpricing in the
Chinese market.
Similar results are obtained when IPO underpricing is measured
alternatively as the difference of market-to-book ratiosbetween
pre-IPO and listing date (DMB). In Model 2, the coefcient on CONS
TA is 7.676 at the 10% signicance level, thusour empirical results
are robust in respect of the association of accounting conservatism
with IPO underpricing in the Chinesemarket.
-
138 Z.J. Lin, Z. Tian / Journal of International Accounting,
Auditing and Taxation 21 (2012) 127 144
Table 5Differences between high- and low information asymmetry
groups.
Partition proxy High information asymmetry Low information
asymmetry Mean variances
N Mean Std. dev. N Mean Std. dev. Variance t-Statistics
Growth 336 0.424 0.274 338 0.023 0.106 0.4003 25.01***
Firm age 337 2.412 0.859 337 7.400 2.814 4.989 31.13***Firm size
337 2.44e+08 8.54e+07 337 6.04e+09 1.92e+09 5.79e+09 5.55***Corp.
gove 337 1.696 1.399 337 1.696 1.295 3.393 32.66***
Note: Subgroups are divided based on the median value of the
four partition proxies respectively, so rms with high growth, young
age, small size and weakcorporate governance are dened as high
information asymmetry. Sales growth is measured as the change in
sales revenue divided by the one-year laggedsales revenue. Firm age
is measured as the number of years passed since a rms establishment
date. Firm size is measured as the total assets before
IPO.Corporate governance is measured by taking the unweighed
average of the standardized composing variables including
separation of CEO/Chair positions,the ratio of the number of
independent directors to the number of all directors, litigation
risk, ownership concentration, and leverage.*** Signicance of
1%.
4.2. Effect of information asymmetry on the association between
accounting conservatism and IPO underpricing
We examine the second hypothesis regarding the impact of pre-IPO
information asymmetry on the relationship betweenaccounting
conservatism and IPO underpricing in this section. As explained in
Section 3, we divide our sample into highvs. low information
asymmetry sub-groups based on four partition factors, i.e., growth,
rm age, rm size, and corporategovernance.12 Table 5 lists the group
means and between-group mean variances under the four partitions.
All show asignicant difference between the high and low information
asymmetry sub-groups at the signicance level of 1%. Forinstance,
the average age of the high information asymmetry group is 2.41
years while for the low asymmetry group itis 7.4 years. The average
total assets are RMB 244 million and RMB 6040 million,
respectively. Table 6 presents theanalyses for the second
hypothesis based on the partitioned sub-samples (e.g., high vs. low
information asymmetry). In eachcolumn, the left (right) sub-sample
represents high (low) information asymmetry rms. In the regression,
the coefcient onCONS TA is 1.115 for high growth sub-group (t =
1.58), which is more pronounced than the coefcient 0.525 for the
lowgrowth sub-group (t = 0.53). The coefcient on CONS TA for the
young rm sub-group is 1.113 at the 10% signicance level(t = 1.67),
which is more pronounced than the coefcient 1.001 for the old rm
sub-group (t = 0.85). The coefcient onCONS TA for the small rm
sub-group is 1.568 at the 5% signicance level (t = 1.98), which is
more pronounced than thecoefcient 0.206 for the large rm sub-group
(t = 0.26). The coefcient on CONS TA for the weak corporate
governancesub-group is 2.419 at the 1% signicance level (t = 2.85),
which is more pronounced than the coefcient of 0.278 for thestrong
corporate governance sub-group (t = 0.36). The results reveal that
the relationship between accounting conservatismand IPO
underpricing is more pronounced for high information asymmetry rms.
Furthermore, we evaluate the differencesin the coefcients between
high and low information asymmetry sub-samples. We nd that the
coefcient on CONS TA issignicantly more negative in rms with high
sales growth, young age, small size, and weak corporate governance
than intheir counterparts with low information asymmetry, thus
conrming that accounting conservatism has a greater impact onIPO
underpricing under a high information asymmetry environment.
4.3. Reverse causality test
Until now, we have not addressed the potential endogeneity
problem of accounting conservatism. An alternative expla-nation of
the negative relationship between accounting conservatism and IPO
underpricing is that rms with higher levelsof underpricing are more
likely to employ less conservative accounting. This reverse
causality explanation is difcult to ruleout since it is hard to
specify ex ante which rms are more underpriced and which are not.
We deal with this reverse causalityissue in the following ways.
First, we use the average of a three-year lagged accounting
conservatism measure to addressthe causality ow from accounting
conservatism to IPO underpricing. Second, we apply the two-step
Heckmans procedureto purge the endogenous component of accounting
conservatism. In the rst stage, we estimate a selection model
(probitmodel) to explain rms overall decisions to choose
conservative accounting or not. The dependent variable is an
indicatorvariable that gets the value of one if a rms accounting
conservatism measure at one year before the IPO is above thesample
median and zero otherwise. The explanatory variables include
leverage, return on assets, sales growth, total assets,the ratio of
independent directors to all directors, and board size. We use
two-year lagged data for explanatory variables asthe determinants
for accounting conservatism one year before IPO. Conditional on
this rst stage analysis, we analyze in thesecond stage the
relationship between accounting conservatism and IPO underpricing
by including other control variables.We construct an inverse Mills
ratio (IMR) based on the coefcient estimates from the rst step
probit model, and includeIMR as an additional explanatory variable
in the second-step model. Table 7 shows that IMR has a signicantly
negativecoefcient (0.992; t = 1.93) and the coefcient on CONS is
still signicantly negative (3.828; t = 3.31), consistent with
12 The partition is based on the group media for each factor
individually. Thus the high- and low information asymmetry
sub-groups have the samenumber of sample rms.
-
Z.J. Lin, Z. Tian / Journal of International Accounting,
Auditing and Taxation 21 (2012) 127 144 139
Table 6Results for the impact of information asymmetry on the
relationship between accounting conservatism and IPO
underpricing.
Sales growth Firm age Firm size Corporate governance
High Low Young Old Small Large Weak Strong
Intercept 3.823*** 3.334*** 4.677*** 2.121*** 6.101*** 2.012***
3.447*** 3.405***
(6.26) (5.30) (7.67) (3.47) (7.96) (3.74) (5.59) (5.54)CONS TA
1.115 0.525 1.113* 1.001 1.568** 0.206 2.419*** 0.278
(1.58) (0.53) (1.67) (0.85) (1.98) (0.26) (2.85) (0.36)BOOK
0.000 0.366*** 0.180 0.137 0.348*** 0.033 0.363*** 0.054
(0.00) (2.88) (1.48) (1.12) (2.61) (0.31) (3.01) (0.45)TA*BOOK
0.612 1.826 0.198 2.149 1.919* 0.089 2.913*** 0.230
(0.66) (1.33) (0.18) (1.63) (1.74) (0.09) (2.72) (0.21)LDAY
0.002 0.003 0.001 0.006 0.009 0.003 0.000 0.002
(0.22) (0.38) (0.12) (0.70) (1.12) (0.46) (0.01) (0.27)OFF SIZE
0.368*** 0.460*** 0.494*** 0.272*** 0.880*** 0.206*** 0.452***
0.328***
(5.66) (7.89) (8.40) (4.26) (7.36) (4.10) (7.84) (6.09)OWN S
0.473** 0.204 0.512** 0.199 0.245 0.559** 0.409* 0.337
(2.19) (0.84) (2.03) (0.90) (0.98) (2.58) (1.74) (1.47)OWN L
0.085 0.247 0.004 0.170 0.145 0.113 0.046 0.053
(0.44) (1.09) (0.02) (0.85) (0.72) (0.53) (0.23) (0.24)TS
2.106*** 0.402 2.078*** 0.338 2.622** 0.568 0.208 1.733**
(2.81) (0.55) (2.74) (0.45) (2.59) (0.98) (0.26) (2.53)VOL
0.004** 0.011*** 0.005** 0.011*** 0.008*** 0.008*** 0.007***
0.009***
(2.03) (4.70) (2.10) (4.84) (3.21) (3.85) (2.93) (4.03)INTEG
0.257** 0.119 0.224 0.249** 0.139 0.201* 0.125 0.257*
(2.28) (0.82) (1.53) (2.15) (1.03) (1.75) (1.06) (1.91)EXCH
0.196 0.164 0.280* 0.008 0.642*** 0.220* 0.422*** 0.004
(1.40) (1.07) (1.94) (0.05) (4.01) (1.73) (2.73) (0.03)UWREP
0.009 0.255** 0.166 0.024 0.139 0.021 0.236** 0.150
(0.08) (2.00) (1.35) (0.20) (0.95) (0.22) (2.53) (1.47)LEV 0.492
0.278 0.548 0.276 0.504 0.702** 0.161 0.807*
(1.20) (0.68) (1.31) (0.70) (1.05) (2.11) (0.41) (1.77)ROA 1.209
2.921** 2.315* 2.251* 0.324 2.370** 1.505 2.148
(1.17) (2.23) (1.94) (1.94) (0.25) (2.12) (1.51) (1.56)N 336 338
337 337 337 337 337 337R2 0.164 0.245 0.224 0.184 0.259 0.166 0.232
0.216Adj. R2 0.127 0.212 0.190 0.149 0.227 0.130 0.199 0.182F 4.487
7.471 6.636 5.187 8.045 4.573 6.965 6.351
Note: This table reports the regression results for test of
second hypothesis. Subgroups are divided based on the median value
of the four partition proxiesrespectively, so rms with high growth,
young age, small size and weak corporate governance are dened as
high information asymmetry. Sales Growthsis measured as the change
in sales revenue divided by the one-year lagged sales revenue. Firm
age is measured as the natural logarithm of the number ofyears
passed since a rms establishment date. Firm size is measured as the
natural logarithm of total assets. Corporate governance is measured
by takingthe unweighed average of the standardized composing
variables including ownership concentration, ratio of the number of
independent directors in theBoard, separation of CEO/Chair
positions, litigation risk, and leverage. Other variables are dened
as in Table 2. t-Statistics are reported in parentheses.
* Statistical signicance at the 10% level.** Statistical
signicance at the 5% level.
*** Statistical signicance at the 1% level.
the interpretation that rms with more conservative accounting
tend to have lower underpricing. Therefore, we concludethat our
results are robust to the correction of potential self-selection
bias that higher underpricing rms are more likely toemploy less
conservative accounting.
4.4. Results of robustness tests
We run several robustness tests to check the validity of our
study results. First, we use C SCORE developed by Khan andWatts
(2009) as an alternative conservatism measure because the
cash-based conditional measure of conservatism mayhave a better
governance role in reducing information asymmetry. We also
partition the full sample into two sub-samplesaccording to whether
or not the rm goes public through book building mechanism. As stock
returns are not available forrms before the IPO, C SCORE in this
study is based on the Ball and Shivakumar (2005) accruals model13
rather than the Basu(1997) model. The accruals model is specied
as:
ACCi = 0 + 1DCFOi + 2CFOi + 3DCFOi CFOi + i (6)
13 This approach presents the advantage of not relying on market
measures, thereby reducing the risk of drawing incorrect inferences
due to marketinefciencies (Garca Lara et al., 2009).
-
140 Z.J. Lin, Z. Tian / Journal of International Accounting,
Auditing and Taxation 21 (2012) 127 144
Table 7Results of reverse causality test using Heckman (1978)
two-step procedure.
Panel A: rst-stage probit regression Panel B: second-stage
regression
MAARIntercept 2.230** Intercept 6.006***
(2.22) (7.51)LEV 0.772* CONS TA 3.828***
(1.71) (3.31)ROA 4.423*** IMR 0.992*
(3.63) (1.93)GROWTH 0.028 BOOK 0.048
(0.20) (0.38)FIRM SIZE 0.139*** TA BOOK 6.787***
(2.92) (4.17)RINDE 0.417 LDAY 0.01
(0.90) (1.19)BOARD SIZE 0.047* OFF SIZE 0.599***
(1.66) (7.22)N 566 OWN S 0.178Pseudo-R2 0.057 (0.69)
OWN L 0.298(1.22)
TS 2.642***(3.43)
VOL 0.008***
(3.81)INTEG 0.200
(1.29)EXCH 0.453***
(2.82)UWREP 0.349***
(3.25)LEV 0.248
(0.50)ROA 0.355
(0.16)N 566
Note: This table reports the results for correcting the
self-selection bias using the Heckman (1978) two-step procedure.
Panel A presents the rst-stageprobit analysis of rms decision to
use conservative accounting or not. The dependent variable equals
to one if a rms accounting conservatism at oneyear before IPO is
above the sample median and zero otherwise. The explanatory
variables include leverage (LEV), return on assets (ROA), sales
growth(GROWTH), total assets (TA), the ratio of independent
directors to all directors (RINDE), and board size (BOARD SIZE). We
use two-year lagged data forexplanatory variables as the
determinants for accounting conservatism one year before IPO. Board
size is the number of directors. Panel B presents theresults of the
second-stage regression. The dependent variable is market-adjusted
abnormal returns. IMR is the inverse Mills ratio constructed based
onthe coefcients of the probit model in Panel A. Other variables
are as dened in Table 2. All variables are winsorized at the
extreme 1% and 99%. t-Statisticsare reported in parentheses.
* Statistical signicance at the 10% level.** Statistical
signicance at the 5% level.
*** Statistical signicance at the 1% level.
where i indexes the rm; ACC is the income before extra-ordinary
items less cash ows from operation plus depreciationexpenses deated
by average total assets at the beginning of the year before IPO;
CFO is cash ows from operations deatedby average total assets at
the beginning of the year before IPO, DCFO is a dummy variable and
equals to 1 if CFO < 0 and 0otherwise; and is the residual. In
this model, 2 is expected to be signicantly negative, showing the
negative correlationbetween accruals and cash ows. 3 is expected to
be positive in the presence of accounting conservatism showing a
positiveassociation between contemporaneous cash ows and accruals
in the bad news periods, suggesting that accrued losses aremore
likely to be reported in the periods of negative cash ows (Ball
& Shivakumar, 2005).
We refer to the timeliness of good news 2 as G SCORE and the
incremental timeliness of bad news 3 as C SCORE,which we use to
alternatively measure accounting conservatism. G SCORE and C SCORE
are linear functions of rm-speciccharacteristics (Watts, 2003):
G SCORE = 2 = 0 + 1SIZEi + 2MBi + 3LEVi (7)C SCORE = 3 = 0 +
1SIZEi + 2MBi + 3LEVi (8)
We substitute them into regression Eq. (6) to obtain Eq. (9)
below and get C SCORE as a rm-year measure of
accountingconservatism using 0, 1, 2 and 3:
ACCi = 0 + 1DCFOi + CFOi(0 + 1SIZEi + 2MBi + 3LEVi) + DCFOi
CFOi(0 + 1SIZEi + 2MBi + 3LEVi)+ (0SIZEi + 1MBi + 2LEVi + 3DCFOi
SIZEi + 4DCFOi MBi + 5DCFOi LEVi) + i (9)
-
Z.J. Lin, Z. Tian / Journal of International Accounting,
Auditing and Taxation 21 (2012) 127 144 141
Table 8Regression results for robustness tests.
Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7 Model
8
OTHERS BOOK 20012005 20062009
Intercept 5.802*** 2.584*** 4.285*** 3.995*** 3.902*** 3.541***
3.456*** 3.350*** 3.285*** 3.475***
(8.45) (4.08) (7.29) (6.04) (8.93) (8.29) (8.20) (8.96) (7.64)
(8.56)CONS TA 0.753* 0.521 1.064* 1.099* 1.046* 1.181** 1.049*
(1.73) (0.97) (1.87) (1.94) (1.86) (2.07) (1.87)CONS DA
1.569*
(1.80)C SCORE 0.126** 0.068
(2.01) (1.61)BOOK 0.306*** 0.927*** 0.185** 0.191** 0.199**
0.192** 0.267*** 0.183**
(2.64) (5.91) (2.18) (2.28) (2.37) (2.31) (3.23) (2.19)TA*BOOK
1.386* 1.436* 1.362* 1.543** 1.361*
(1.85) (1.92) (1.83) (2.07) (1.83)DA*BOOK 1.723
(1.45)LDAY 0.001 0.005 0.017*** 0.008 0.002 0.002 0.002 0.001
0.002 0.001
(0.07) (0.63) (2.95) (0.88) (0.35) (0.43) (0.40) (0.27) (0.46)
(0.27)OFF SIZE 0.313*** 0.566*** 0.538*** 0.479*** 0.434***
0.407*** 0.392*** 0.388*** 0.426*** 0.398***
(3.46) (9.72) (9.79) (7.25) (10.58) (9.67) (10.08) (9.64)
(10.83) (10.28)OWN S 0.106 0.407* 0.427* 0.435** 0.398** 0.391**
0.390** 0.358** 0.354** 0.385**
(0.40) (1.94) (1.76) (2.04) (2.42) (2.41) (2.40) (2.22) (2.18)
(2.41)OWN L 0.200 0.034 0.155 0.044 0.048 0.048 0.054 0.071 0.094
0.059
(0.75) (0.20) (0.62) (0.25) (0.33) (0.32) (0.37) (0.49) (0.64)
(0.40)TS 2.486*** 0.566 0.180 0.168 1.737*** 1.256** 1.195**
1.313** 1.156** 1.186**
(3.60) (0.71) (0.30) (0.20) (3.22) (2.43) (2.30) (2.55) (2.23)
(2.34)VOL 0.001 0.019*** 0.002 0.016*** 0.007*** 0.007*** 0.008***
0.008*** 0.008*** 0.008***
(0.62) (8.36) (0.90) (6.43) (4.56) (4.63) (4.65) (4.67) (5.03)
(4.71)INTEG 0.012 0.313*** 0.032 0.320*** 0.209** 0.231*** 0.224**
0.218** 0.192** 0.210**
(0.08) (2.72) (0.28) (2.66) (2.32) (2.59) (2.51) (2.45) (2.11)
(2.37)EXCH 0.073 0.628*** 0.434*** 0.601*** 0.315*** 0.198* 0.205**
0.228** 0.265** 0.226**
(0.57) (3.65) (3.98) (2.60) (2.91) (1.92) (1.99) (2.22) (2.58)
(2.21)UWREP 0.348*** 0.068 0.097 0.079 0.226*** 0.236*** 0.198***
0.227***
(3.30) (0.78) (1.08) (0.86) (3.28) (3.49) (2.93) (3.36)LEV
2.775*** 1.119* 0.668** 0.169 0.392 0.363 0.391 0.373
(3.17) (1.93) (2.01) (0.41) (1.36) (1.29) (1.39) (1.43)ROA
5.374*** 1.019 1.989** 0.197 1.949** 1.857** 1.914** 1.986***
1.160*
(3.47) (0.87) (2.01) (0.17) (2.44) (2.33) (2.41) (2.66)
(1.86)UWREP1 0.105
(1.24)UWREP2 0.097
(1.24)LEV1 0.064*
(1.89)SD 0.012
(0.06)OIA 1.650***
(2.68)N 299 358 310 364 653 674 674 673 656 674R2 0.204 0.326
0.337 0.314 0.203 0.183 0.183 0.198 0.207 0.197Adj. R2 0.171 0.303
0.308 0.289 0.185 0.166 0.166 0.180 0.189 0.180F 11.62 11.96 11.56
12.34 11.58 10.55 10.55 11.57 11.94 11.55
Note: This table reports the regression results for robustness
tests Model 1 use C SCORE as an alternative conservatism measure
and partitions the fullsample into two subsamples according to
whether a rm goes public through book building mechanism. Model 2
partitions the sample into two test periods(20012005 vs. 20062009)
to test the effect of circulation of non-tradable shares since 2006
on the association between accounting conservatism andIPO
underpricing. Model 3 use discretionary accruals as an alternative
measure of conservatism. Models 4 to 8 use alternative measures of
underwriterreputation, leverage and protability for robustness
tests. UWREP1 is a dummy variable and equals to 1 if the
underwriter belongs to Top 5 underwritersin China and 0 otherwise
as the underwriters are ranked according to their total
underwriting amounts. UWREP2 is a dummy variable and equals to 1
ifunderwriter belongs to Top 5 underwriters in China and 0
otherwise, as underwriters are ranked according to their rm size.
LEV1 refers to a rms totaldebt divided by its net assets of the
year before IPO. SD refers to a rms short-term debt divided by its
book value of total assets of the year before IPO. OIAis operating
income divided by total assets of the year before IPO. Other
variables are as dened in Table 2. All variables are winsorized at
the extreme 1%and 99%. t-Statistics are reported in
parentheses.
* Statistical signicance at the 10% level.** Statistical
signicance at the 5% level.
*** Statistical signicance at the 1% level.
-
142 Z.J. Lin, Z. Tian / Journal of International Accounting,
Auditing and Taxation 21 (2012) 127 144
Table 8 (Model 1) reports the regression results using C SCORE
as the independent variable to measure accounting con-servatism and
test H1. The coefcient on C SCORE is 0.126 (t = 2.01) at the 5%
signicance level for rms that go publicthrough other mechanisms
than book building mechanism, while the coefcient on C SCORE is
insignicant for rms that gopublic through book building mechanism.
This result is consistent with H1 and conrms the negative
relationship betweenaccounting conservatism and IPO underpricing
and the substitutive relationship between accounting conservatism
and bookbuilding mechanism.
As described earlier, the original non-tradable shares have been
permitted to be traded on stock exchanges in China since2006.
Increased volume of publicly tradable shares offers potential
investors more choices and reduces the high demand ofinvestors for
new issues. This fact may mitigate the demand for accounting
conservatism and lower the level of underpricing.We partition the
sample into two periods (20012005 vs. 20062009) to test the effect
of the circulation of originally non-tradable shares since 2006 on
the relationship between accounting conservatism and IPO
underpricing in the Chinese market.Results of Model 2 in Table 8
show that the coefcient on CONS TA in the 20012005 period is 0.753
(t = 1.73) at the10% signicance level but insignicant in the
20062009 period, indicating the negative relationship between
accountingconservatism and IPO underpricing is mitigated after the
circulation of non-tradable shares in 2006.
We also consider discretionary accruals (DA) as the measure of
accounting conservatism,14 because discretionary accrualsmay be
more directly affected by managers choice of accounting standards
or practices that may result in varied degrees ofaccounting
conservatism. We employ the modied Jones model (Dechow, Sloan,
& Sweeney, 1995) to measure discretionaryaccruals. The modied
Jones model is specied as follows;
TAitAit1
= 0 + 1(
REVit RECitAit1
)+ 2
GPPEitAit1
+ it (10)
where TAit is total accruals for rm i in year t, REVit is the
change in is total revenue from t 1 to t, and RECit is thechange in
accounts receivables for rm i from the previous year. GPPEit is the
gross acquisition cost of property, plants andequipments for rm i
in year t, and Ait1 is the value of total assets for rm i in year t
1.
The OLS residuals from Eq. (10) are the estimates of
discretionary accruals. The alternative proxy for accounting
conser-vatism (CONS DA) in this study is measured as discretionary
accruals multiplied by negative one. The regression results forH1
are similar when accounting conservatism is proxied by
discretionary accruals (CONS DA). In Model 3, the coefcient onCONS
DA is 1.569 at the 10% signicance level while coefcients on other
variables are generally the same as that shownin Table 4 when
discretionary accruals are used to measure accounting
conservatism.
Finally, besides ranking underwriters according to the number of
their IPO deals in the main test, we also rank under-writers in
terms of total underwriting amounts and their rm size for
robustness tests. In addition, we use a rms totaldebt divided by
its net assets at the year before IPO and a rms short-term debt
divided by its book value of total assets ofthe year before IPO to
alternatively proxy for leverage risk. We also use operating income
divided by total assets at the yearbefore IPO to proxy for rm
protability. Models 48 in Table 8 report the regression results of
the robustness tests and ourhypotheses are generally supported
after adopting alternative measures of those control variables.
5. Conclusion
The purpose of this study is to investigate whether and how
accounting conservatism impacts IPO underpricing in theChinese
stock market. We utilize the unique data from Chinese IPO rms
contextual to their inherent uncertainty andinformation asymmetry
to test two hypotheses. The results of our regression analyses
reveal that IPO underpricing in Chinais inversely associated with
accounting conservatism, after controlling for deal-specic and
rm-specic characteristics. Thisevidence is consistent with
asymmetric information theory underlying IPO underpricing in the
literature. Thus, accountingconservatism will help to reduce
information asymmetry facing IPO rms and mitigate IPO underpricing.
In addition, we ndthat the negative impact of accounting
conservatism on IPO underpricing is more pronounced for rms with
high informationasymmetry than for rms with low information
asymmetry, indicating that higher information asymmetry creates
moreincentives for conservative accounting by IPO rms. Our ndings
should have positive implications for the improvement ofaccounting
and reporting practices and the effective operation of capital
market in the emerging economies like China andother developing
countries.
Generalization of our study ndings must be made with caution as
there are some distinct characteristics for the develop-ing Chinese
stock market at present. A potential limitation is that this study
focuses mainly on the perspective of informationasymmetry to
investigate the impact of accounting conservatism on IPO
underpricing. However, there are other theoriesregarding the causes
of IPO underpricing in the literature. Factors other than
information asymmetry could be explored toinvestigate the
association of IPO underpricing with accounting conservatism in
future research. In addition, the validityof the corporate
governance variables (e.g., ownership concentration, proportion of
outside directors, duality of CEO/Boardchairman positions and
litigation risk) being used in our study may be further tested as
the Chinese stock market is oper-ating in a context with dominant
government ownership and weak rule of law. The environmental
differences should be
14 Our main test is based on the total accruals approach because
discretionary accruals data are not available for some rms, so the
proxy for accountingconservatism measured by discretionary accruals
is based on a smaller sample size.
-
Z.J. Lin, Z. Tian / Journal of International Accounting,
Auditing and Taxation 21 (2012) 127 144 143
taken into account, if data are available, if similar studies
with data from other capital markets are carried out. Finally,
morerigorous empirical models based on multivariate data analysis
may be developed to generate more powerful evidence onthe
relationship between accounting conservatism and IPO
underpricing.
Acknowledgements
The authors are thankful to two anonymous reviewers and
Professor Kathleen Sinning (the Journal Editor) for theirinsightful
comments and editorial support to the previous drafts of this
paper.
References
Aggarwal, R., Leal, R., & Hernandez, L. (1993). The
aftermarket performance of initial public offerings in Latin
America. Financial Management, 22(1), 4253.Aharony, J., Wang, J.,
& Yuan, H. (2010). Tunneling as an incentive for earnings
management during the IPO process in China. Journal of Accounting
and Public
Policy, 29(1), 126.Ahmed, A., & Duellman, S. (2007).
Accounting conservatism and board of director characteristics: An
empirical analysis. Journal of Accounting and Economics,
43(23), 411437.Ahmed, A. S., Billings, B. K., Morton, R. M.,
& Stanford-Harris, M. (2002). The role of accounting
conservatism in mitigating bondholder-shareholder conicts
over dividend policy and in reducing debt Costs. The Accounting
Review, 77(4), 867890.Ahmed, A. S., Neel, M. J., & Wang, D.D.,
2010. Does mandatory adoption of IFRS improve accounting quality?
Preliminary evidence (Working paper series).
Available at SSRN eLibrary.Amihud, Y., & Mendelson, H.
(1986). Asset pricing and the bid-ask spread. Journal of Financial
Economics, 17(2), 223249.Ball, R., & Shivakumar, L. (2005).
Earnings quality in UK private rms: Comparative loss recognition
timeliness. Journal of Accounting and Economics, 39(1),
83128.Ball, R., & Shivakumar, L. (2008). Earnings quality at
initial public offerings. Journal of Accounting and Economics,
45(23), 324349.Ball, R., Robin, A., & Wu, J. S. (2003).
Incentives versus standards: Properties of accounting income in
four East Asian countries. Journal of Accounting and
Economics, 36(13), 235270.Banerjee, S., Dai, L., & Shrestha,
K., 2009. Cross-country IPOs: What explain differences in
underpricing? (Working paper series). Available at SSRN
eLibrary.Basu, S. (1997). The conservatism principle and the
asymmetric timeliness of earnings. Journal of Accounting and
Economics, 24(1), 337.Boulton, T. J., Smart, S. B., & Zutter,
C. J. (2011). Earnings quality and international IPO underpricing.
The Accounting Review, 86(2), 483505.Burkart, M., Gromb, D., &
Panunzi, F. (1997). Large shareholders, monitoring, and the value
of the rm. Quarterly Journal of Economics, 112(3), 693728.Bushman,
R., & Piotroski, J. (2006). Financial reporting incentives for
conservative accounting: The inuence of legal and political
institutions. Journal of
Accounting and Economics, 42(12), 107148.Chan, K., Wang, J.,
& Wei, K. C. J. (2004). Underpricing and long-term performance
of IPOs in China. Journal of Corporate Finance, 10(3),
409430.Chaney, P. K., Faccio, M., & Parsley, D. (2011). The
quality of accounting information in politically connected rms.
Journal of Accounting and Economics,
51(12), 5876.Chen, C.-H., & Hui-Tzu, S. (2004). Initial
public offering and corporate governance in Chinas transitional
economy. In Governance, regulation, and privatization
in the Asia-Pacic region, NBER East Asia seminar on economics. ,
p. 12.Cheung, Y.-L., Ouyang, Z., & Tan, W. (2009). How
regulatory changes affect IPO underpricing in China. China Economic
Review, 20(4), 692702.Chi, J., & Padgett, C. (2002). The
performance and long-run characteristics of the Chinese IPO market.
SSRN eLibrary.Chi, J., & Padgett, C. (2005). Short-run
underpricing and its characteristics in Chinese initial public
offering (IPO) markets. Research in International Business
and Finance, 19(1), 7193.Dechow, P. M., Sloan, R. G., &
Sweeney, A. P. (1995). Detecting earnings management. The
Accounting Review, 70(2), 193225.Francis, J., & Martin, X.
(2010). Acquisition protability and timely loss recognition.
Journal of Accounting and Economics, 49(12), 161178.Garca Lara, J.,
Garca Osma, B., & Penalva, F. (2009). Accounting conservatism
and corporate governance. Review of Accounting Studies, 14(1),
161201.Givoly, D., & Hayn, C. (2000). The changing time-series
properties of earnings, cash ows and accruals: Has nancial
reporting become more conservative?
Journal of Accounting and Economics, 29(3), 287320.Harris, L.
(1991). Stock price clustering and discreteness. Review of
Financial Studies, 4(3), 389.Heckman, J. J. (1978). Dummy
endogenous variables in a simultaneous equation system.
Econometrica, 47, 153161.Haw, I. -M., Lee, J. J., & Lee, W. J.,
2010. Debt nancing and accounting conservatism in private rms.
(Working paper series). Available at SSRN eLibrary.Ibbotson, R. G.
(1975). Price performance of common stock new issues. Journal of
Financial Economics, 2(3), 235272.Jog, V., & Mcconomy, B.
(2003). Voluntary disclosure of management earnings forecasts in
IPO prospectuses. Journal of Business Finance & Accounting,
30(12), 125168.Kennedy, D. B., Sivakumar, R., & Vetzal, K.
R. (2006). The implications of IPO underpricing for the rm and
insiders: Tests of asymmetric information theories.
Journal of Empirical Finance, 13(1), 4978.Khan, M., & Watts,
R. (2009). Estimation and empirical properties of a rm-year measure
of accounting conservatism. Journal of Accounting and
Economics,
48(23), 132150.Kim, B. H., & Pevzner, M. (2010). Conditional
accounting conservatism and future negative surprises: An empirical
investigation. Journal of Accounting and
Public Policy, 29(4), 311329.Lafond, R., & Watts, R. (2008).
The information role of conservatism. The Accounting Review, 83,
443478.Leone, A., Rock, S., & Willenborg, M. (2007). Disclosure
of intended use of proceeds and underpricing in initial public
offerings. Journal of Accounting Research,
45(1), 111153.Li, Y. (2009). Empirical research on IPO
underpricing of A-share companies in China after the implementation
of book building mechanisms. Economist, 12,
8486.Lin, Z., & Chen, F. (2005). Value relevance of
international accounting standards harmonization: Evidence from A-
and B-share markets in China. Journal of
International Accounting, Auditing and Taxation, 14(3),
79103.Ljungqvist, A. (2007). IPO underpricing. In Handbook of
corporate nance (pp. 375422).Ljungqvist, A. P., Jenkinson, T.,
& Wilhelm, W. J., Jr. (2003). Global integration in primary
equity markets: The role of US banks and US investors. Review
of
Financial Studies, 16(1), 63.Logue, D. E. (1973). On the pricing
of unseasoned equity issues: 19651969. The Journal of Financial and
Quantitative Analysis, 8(1), 91103.Loughran, T., & Ritter, J.
(2004). Why has IPO underpricing changed over time? Financial
Management, 33(3), 537.Mok, H. M. K., & Hui, Y. V. (1998).
Underpricing and aftermarket performance of IPOs in Shanghai,
China. Pacic-Basin Finance Journal, 6(5), 453474.Mueller, H., &
Inderst, R., 2001. Ownership concentration, monitoring, and the
agency cost of debt. (Working paper series). Available at SSRN
eLibrary.Rajan, R., & Zingales, L. (2003). The great reversals:
The politics of nancial development in the twentieth century.
Journal of Financial Economics, 69(1),
550.Ritter, J. R., & Welch, I. (2002). A review of IPO
activity, pricing, and allocations. The Journal of Finance, 57(4),
17951828.Rock, K. (1986). Why new issues are underpriced. Journal
of Financial Economics, 15(12), 187212.
-
144 Z.J. Lin, Z. Tian / Journal of International Accounting,
Auditing and Taxation 21 (2012) 127 144
Schrand, C. M., & Verrecchia, R. E. (2005). Information
disclosure and adverse selection explanations for IPO underpricing.
SSRN eLibrary.Sherman, A. E. (2005). Global trends in IPO methods:
Book building versus auctions with endogenous entry. Journal of
Financial Economics, 78(3), 615649.Shleifer, A., & Vishny, R.
(1993). Corruption. The Quarterly Journal of Economics, 108(3),
599617.Stoll, H. R., & Curley, A. J. (1970). Small business and
the new issues market for equities. The Journal of Financial and
Quantitative Analysis, 5(3), 309322.Su, D., & Fleisher, B. M.
(1999). An empirical investigation of underpricing in Chinese IPOs.
Pacic-Basin Finance Journal, 7(2), 173202.Teoh, S. H., & Wong,
T. J. (2002). Why new issues and high-accrual rms underperform: The
role of analysts credulity. The Review of Financial Studies,
15(3),
869900.Teoh, S. H., Welch, I., & Wong, T. J. (1998).
Earnings management and the long-run market performance of initial
public offerings. The Journal of Finance,
53(6), 19351974.Watts, R. L. (2003). Conservatism in accounting.
Part II: Evidence and research opportunities. Accounting Horizons,
17(4), 287301.Yu, T., & Tse, Y. K. (2006). An empirical
examination of IPO underpricing in the Chinese A-share market.
China Economic Review, 17(4), 363382.
Accounting conservatism and IPO underpricing: China evidence1
Introduction2 Study background and hypotheses2.1 IPO market in
China2.2 IPO underpricing and information asymmetry theory2.3
Accounting conservatism and IPO underpricing2.4 Accounting
conservatism and information asymmetry
3 Research design and data3.1 Measurement of accounting
conservatism3.2 Measurement of IPO underpricing3.3 Empirical
model3.4 Data3.5 Descriptive statistics