Do Chinese government subsidies affect firm value?

Description
Consistent with the prevailing socio-political ideology of China, the Chinese government
offers financial assistance to firms, including many listed companies. Government subsidies
are provided for several reasons including support for investment, support to enable
firms to pursue social objectives, and support to prop up ailing firms in order to protect
jobs.
We examine the value relevance of government subsidies for Chinese listed companies
and structure our study around three questions. First, whether the subsidies received by
Chinese listed companies are value relevant consistent with their time-series properties.
Second, whether the value relevance of subsidies depends on the purpose for which they
are used. Third, whether the value relevance of subsidies depends on the channel through
which they are granted.

Do Chinese government subsidies affect ?rm value?
Edward Lee
a
, Martin Walker
b,?
, Cheng Zeng
c
a
Manchester Accounting and Finance Group, Manchester Business School, Crawford House, Oxford Road, Manchester M13 9PL, United Kingdom
b
Manchester Accounting and Finance Group, Manchester Business School, Room 6.21, Harold Hankins Building, Booth Street West, Manchester M13 9QH,
United Kingdom
c
School of Economics, Finance, and Management, University of Bristol, Social Sciences Complex, 8 Woodland Road, Clifton BS8 1TN, United Kingdom
a b s t r a c t
Consistent with the prevailing socio-political ideology of China, the Chinese government
offers ?nancial assistance to ?rms, including many listed companies. Government subsi-
dies are provided for several reasons including support for investment, support to enable
?rms to pursue social objectives, and support to prop up ailing ?rms in order to protect
jobs.
We examine the value relevance of government subsidies for Chinese listed companies
and structure our study around three questions. First, whether the subsidies received by
Chinese listed companies are value relevant consistent with their time-series properties.
Second, whether the value relevance of subsidies depends on the purpose for which they
are used. Third, whether the value relevance of subsidies depends on the channel through
which they are granted.
We motivate these research questions through interviews of accountants, managers, aca-
demics, government of?cials and ?nancial analysts. Through large sample analyses, we
con?rm that subsidies are positively related to ?rm value, but less so for distressed ?rms
and subsidies granted through non-tax channels. Our study contributes to improved
understanding of Chinese-style capitalism.
Ó 2014 Elsevier Ltd. All rights reserved.
Introduction
China is playing an increasingly important role in the
global economy, and is expected to surpass the US to be-
come the world’s largest economic power (Hawksworth
& Tiwari, 2011). An interesting feature of Chinese listed
?rms, that distinguishes them from their Western peers,
is the pervasiveness of their access to ?nancial subsidies
from the government. Despite over three decades of trans-
formation from a centrally planned to a market-oriented
economy (Ezzamel, Xiao, & Pan, 2007), the Chinese govern-
ment (both central and local levels) maintains a high de-
gree of in?uence on listed ?rms through such practices
(Allen, Qian, & Qian, 2005). In recent years, however, there
is increased international concern that state subsidies are
giving Chinese ?rms an unfair advantage over foreign com-
petitors (Capital Trade Incorporated, 2008). These reasons
call for a better understanding of the effects of Chinese
state subsidies.
The main purpose of this paper is to examine the link
between ?rm value and Chinese government subsidies.
Since state subsidies capture the government’s ‘‘visible
hand’’, a study of the value relevance of subsidies for Chi-
nese listed ?rms provides an opportunity to examine the
direct role of government in the economy. Although sev-
eral studies have investigated the value relevance of
accounting numbers in China (e.g. Bao & Chow, 1999;
Chen, Chen, & Su, 2001; Sami & Zhou, 2004), there has been
little systematic examination of how subsidies are re-
?ected in the market value of Chinese listed companies.http://dx.doi.org/10.1016/j.aos.2014.02.002
0361-3682/Ó 2014 Elsevier Ltd. All rights reserved.
?
Corresponding author. Tel.: +44 0161 275 4008; fax: +44 0161 275
4023.
E-mail address: [email protected] (M. Walker).
Accounting, Organizations and Society 39 (2014) 149–169
Contents lists available at ScienceDirect
Accounting, Organizations and Society
j our nal homepage: www. el sevi er. com/ l ocat e/ aos
Moreover, most existing studies on the value relevance of
accounting information in China fail to model the underly-
ing information dynamics, which prevents them from con-
sidering whether the accounting numbers are rationally
valued. Thus, we explore this issue by using large scale sta-
tistical data to estimate accounting based models of ?rm
value based on the ideas in Ohlson (1995, 1999).
However, before undertaking this statistical analysis we
?rst engaged in a number of interviews of people actively
involved in the granting and receiving of subsidies (e.g.
government of?cials and entrepreneurs) as well as people
with expertise on this issue (e.g. accountants, academics,
and ?nancial analysts). The interviews were used to gain
a better understanding of why the market value of subsi-
dies might vary across different contexts. Based on the re-
sult of the interviews, we structure our econometric
analyses of the valuation consequence of state subsidies
around the following three questions.
First, we investigate whether the subsidies received by
Chinese listed companies are value relevant and, if so, is
the value relevance of subsidies consistent with their
time-series properties? Following the approach of Barth,
Hand, and Landsman (1999) and Landsman, Miller, and
Yeh (2007), which extend the Ohlson (1999) model, we
?nd that subsidies are value relevant on average, consis-
tent with the results of our analyses of their information
dynamics. This is also consistent with subsidies being
viewed as bene?cial from the perspective of outside equity
investors.
Second, we investigate whether subsidies are equally
value relevant across all ?rm contexts. In particular, does
the value relevance of subsidies differ if they are used to
boost ?rms’ competitiveness or to bailout ?rms in trouble?
We partition the sample into ?nancially healthy and dis-
tressed ?rms and our analyses indicate that subsidies are
value relevant for the former group, but not the latter
group. This is consistent with subsidies not adding value
to the ?rm if they are likely to be used to assist earnings
management or impede natural selection against weak
?rms.
Third, we consider if the value relevance of subsidies
depends on the channel through which they are granted.
We classify subsidies into those based on tax rebates and
those provided by non-tax approaches such as direct cash
payments, loan guarantees, and debt forgiveness. Our anal-
yses reveal that the valuation consequence of subsidies dif-
fers considerably between the two different channels. In
particular, we ?nd that the value relevance of tax based
subsidies is substantially greater than that of non-tax
based subsidies, which is consistent with the former being
granted on a more persistent and transparent basis.
Our ?ndings have implications for academics, policy
makers, and ?rms. From an academic point of view, we
show that it is possible to use accounting based valuation
models to demonstrate the valuation consequence of state
subsidies. This research methodology also enables us to
show that value relevance varies in predictable ways
according to the purpose and channel of the subsidies.
For policy makers, our evidence implies that subsidies
are an essential input to the valuation decisions of inves-
tors in Chinese listed ?rms and policies that ensure their
clear disclosure should be promoted. As such, our study
also offers policy implications for Chinese capital market
regulators, who are keen to enhance transparency and
information ef?ciency.
1
For the ?rms, our study suggests
that the Chinese capital market recognizes that state subsi-
dies provide net bene?ts to corporate value. Thus we have
identi?ed a valuation driver speci?c to the context of China,
which ?rms should consider as part of their effort to maxi-
mize their shareholders’ wealth.
The remainder of the paper is organized as follows: ‘Lit-
erature review and institutional background’ brie?y re-
views the economic literature on the role of subsidies in
the development of the economy and introduces the insti-
tutional environment in China. ‘Interviews and hypotheses
development’ presents the ?ndings of our interviews, and
develops our main research hypotheses. ‘Modeling the va-
lue relevance of government subsidies’ describes the re-
search design and sample selection process. ‘Empirical
results’ presents the econometric evidence and the last
section concludes.
Literature review and institutional background
Related economic literature
Subsidies are economic intervention tools used by gov-
ernments around the world to offset market imperfections,
to exploit economies of scale, and to pursue social policy
objectives (Schwartz & Clements, 1999). For instance, a
government may offer subsidies to ?rms to in?uence re-
search and development (R&D) (González, Jaumandreu, &
Pazo, 2005; Gorg & Strobl, 2007; Irwin & Klenow, 1996),
export competitiveness (Desai & Hines, 2008; Ishikawa &
Spencer, 1999; Shivakumar, 1993), production ef?ciency
(Azzimonti, de Francisco, & Krusell, 2008; Bagwell & Stai-
ger, 1989, 2006), wages and employment (De Mel, McKen-
zie, & Woodruff, 2010; Phelps, 1994; Snower, 1994), and
environmentally friendly practices (Conrad, 1993; Conrad
& Wang, 1993; Kohn, 1997). However, existing empirical
research on the economic effects of state subsidies yields
a mixed picture. On the one hand, some studies suggest
that subsidies can stimulate R&D activities (Almus & Czar-
nitzki, 2003; Davidson & Segerstrom, 1998; González et al.,
2005), improve ?rm pro?tability (Desai & Hines, 2008; Gir-
ma, Gong, Gorg, & Yu, 2009), enhance competitiveness
(Bagwell & Staiger, 1989; Brander & Spencer, 1985), and
reduce the capital constraints of ?rms (Claro, 2006; Li,
2002). On the other hand, some studies suggest that subsi-
dies result in overproduction or ef?ciency losses (Larsen &
Shah, 1992; Lopez & Galinato, 2007; To, 1994), and invite
international disputes (e.g. Neary, 1994). In recent years,
the increased outcry of the US government against Chinese
state subsidies to ?rms (Capital Trade Incorporated, 2008)
suggests that subsidies may distort competitiveness and
disadvantage US ?rms against their Chinese counterparts.
1
These issues were mentioned, for instance, in the Chinese Capital
Markets Development Report issued by the China Security Regulatory
Committee (CSRC) in 2007 to discuss the future of China’s stock market
development:http://www.csrc.gov.cn/pub/csrc_en/Informations/publica-
tion/200911/P020091103520222505841.pdf (in Chinese).
150 E. Lee et al. / Accounting, Organizations and Society 39 (2014) 149–169
The role of government in transitional economies,
which are evolving from centrally planned socialism to
market-oriented capitalism, is widely debated in the eco-
nomic literature. In terms of process, the literature puts
forward the ‘‘big bang’’ and sequential reform approach.
The ‘‘big bang’’ approach argues that reforms should be
carried out quickly (Balcerowicz, 1995; Lipton & Sachs,
1990) so as to create irreversibility (Boycko, Shleifer, &
Vishny, 1995) and exploit the window of opportunity. This
was adopted by the Eastern European and former Soviet
Union countries through mass privatization to overcome
political constraints (Shleifer & Treisman, 2000). The
sequential reform approach suggests a gradual process to
build upon the successes of previous stages in order to gen-
erate incentives for further changes (Dewatripont & Ro-
land, 1995; Litwack & Qian, 1998; Wei, 1997). This was
implemented in China through a gradual liberalization
process (Lau, Qian, & Roland, 2000). In terms of govern-
ment behavior in transitional economies, the existing liter-
ature studies the ‘‘grabbing hand’’ and ‘‘helping hand’’
approaches (e.g. Frye & Shleifer, 1997). The former portrays
government as controlled by of?cials that do not maximize
social welfare and pursue their own sel?sh objectives
(Shleifer & Vishny, 1998), and is associated with local gov-
ernments such as Russia and Poland (Shleifer & Vishny,
1993). The latter is often associated with the case of China
(Che & Qian, 1998; Walder, 1995), where the government
seeks to guide or stimulate economic growth by reallocat-
ing ?nancial resources. In terms of outcome, most transi-
tional economies experience an initial drop in output,
which is then followed by a recovery of varying paces
(Hellman & Shankerman, 2000). China is an interesting
exception to this pattern since it began with high growth
immediately without the early stage recession (Roland,
2002).
Inter-jurisdictional competition among local govern-
ments is an important feature of state intervention in
the economy. Earlier literature suggests that competition
among regions is bene?cial because it helps satisfy regio-
nal demand (Oates, 1972), improve governance (Brennan
& Buchanan, 1980), and reduce resource allocation
distortions (Qian & Roland, 1998). More recent literature
suggests competition among regions may cause problems
such as the depletion of local government revenues
(Keen & Marchand, 1997) and the ?xation of local gov-
ernment on business interests (Bardhan & Mookherjee,
2000). Empirical studies report mixed results, with the
in?uence of local government competition on growth in
China being positive but in Russia being negative (Jin,
Qian, & Weingast, 2005; Zhuravskaya, 2000). One possi-
ble reason for this contrast is the difference in the degree
of political centralization between the two transitional
economies (Enikolopov & Zhuravskaya, 2007). In Russia,
the central government was ineffective in disciplining
local government of?cials against self-serving policies
that would impose negative externalities on the rest of
the country (Shleifer & Treisman, 2000). In China, local
government of?cials are more disciplined due to political
career advancement concerns as a result of the effective
control of a single political party across the entire
country.
Institutional setting
The growing interest among academics, policy makers,
and practitioners in Chinese-style capitalism stems from
China’s increasingly important position in the world econ-
omy.
2
China’s stock market began with the opening of the
stock exchanges in Shanghai and Shenzhen back in the early
1990s to facilitate economic reform. It subsequently grew to
become the largest developing economy by 2001, and the
world’s 5th largest economy, in terms of market capitaliza-
tion, by 2010.
3
The overwhelming majority of the listed
?rms are former state-owned enterprises (SOE) in which
the government retained a majority ownership following
their initial public offering (IPO) (O’Connor et al., 2004).
For Chinese ?rms having a stock market listing is important
both for capital acquisition purposes, and for the purpose of
international visibility.
Unlike Western developed economies, the Chinese
stock exchanges impose delisting rules that in?uence the
allocation of market capital. The China Security Regulatory
Committee (CSRC) mandates that if a listed ?rm reports
losses in two consecutive years, its stock will be classi?ed
as specially treated (ST) and face many trading and ?nan-
cial restrictions. For such ?rms, the daily stock price move-
ment is restricted within a range of 5%, the semi-annual
report must be audited, and raising additional capital from
the stock market is prohibited. If an ST ?rm reports one
more year’s loss, then its stock is suspended from trading
on the stock exchanges, and after a fourth annual loss the
stock is delisted. To avoid these restrictions or delisting,
?rms manipulate earnings in order to avoid reporting
losses and state subsidies may be used to achieve this pur-
pose (Chen, Lee, & Li, 2008).
Subsidies from the government are an important re-
source for Chinese ?rms.
4
Based on Chinese Accounting
Standards (2006), government subsidies are de?ned as mon-
etary or non-monetary assets obtained freely by an enter-
prise from the government, but excluding the capital
invested by the government as the partial owner of the
enterprise.
5
According to state policy, the primary goal of
subsidies is to develop national priority areas such as agri-
culture, public utilities, and high-tech industries (Chen
et al., 2008; Girma et al., 2009). In addition, the governments
may also provide subsidies to help ?rms overcome capital
constraints and to aid ?rms in ?nancial dif?culties (Claro,
2006).
6
Chinese SOEs receive more subsidies on average rel-
ative to privately controlled ?rms, because the government
makes use of them to pursue socio-political objectives such
2
Appendix A compares the economy of China as of 2009 against
developed G7 countries (Panel A) and other major emerging economies
(Panel B).
3
The statistics can be found at:http://am-articles.blogspot.co.uk/2011/
09/worlds-top-5-stock-exchanges.html.
4
Allen et al. (2005) suggest that state subsidies are one of the four most
important ?nancial sources of Chinese ?rms, along with bank loans, ?rms’
self-fundraising, and foreign direct investment.
5
Chinese ?rms report state subsidies as one of the below-the-line items
in ?nancial statements. State subsidies were reported as an independent
item before 2007, and included as part of the ‘‘other revenues’’ item
afterwards.
6
We use the word ‘‘governments’’ here because both central and
regional governments provide subsidies.
E. Lee et al. / Accounting, Organizations and Society 39 (2014) 149–169 151
as creating job opportunities and stabilizing local economies
(e.g. Hung, Wong, & Zhang, 2008; Lee, 2001; O’Connor et al.,
2006). Subsidies can be granted by both central and local
governments (Thomas, 1994), although such powers are
increasingly delegated to local government of?cials who
have a better understanding of their regional development
needs. This ?scal decentralization also induces competition
among local governments to offer subsidies since the num-
ber of ?rms under their jurisdiction is viewed as a perfor-
mance indicator of the of?cials.
State subsidies can be broadly categorized into tax and
non-tax based. Tax based subsidies are mainly granted
through rebates of value added tax and export tax to
encourage prioritized regions, sectors, and projects. Firms
become eligible for tax rebates if they are located in special
economic zones or if they invest in projects and/or operate
in sectors prioritized by government policies. As such, the
receipt of tax base subsidies may signal that a ?rm has bet-
ter prospects or reduced risks. The non-tax based subsidies
are offered as direct cash payments, loan guarantees, and
debt forgiveness. Government of?cials grant non-tax based
subsidies on a more discretionary and subjective basis. As
such, the receipt of non-tax based subsidies is more a
re?ection of political favoritism rather than a ?rm’s poten-
tial. These differences in the implication of tax and non-tax
based subsidies are expected to in?uence investors’ deci-
sions on the valuation of the recipient ?rms.
The value relevance of accounting information has been
widely studied in the Western accounting literature.
7
For
instance, Landsman et al. (2007) ?nd that earnings and book
values explain around 64.3% of the variation in market value
of US ?rms over the period of 1990–2000. Barth et al. (2001)
argue that value relevance evidence is important to account-
ing standard setters since one of the primary purposes of
?nancial reporting is to provide information useful for valu-
ing ?rms. In the emerging economies such as China, the lit-
erature on the value relevance of accounting information is
more limited. Existing studies on this issue in China (Chen
& Wang, 2004; Chen et al., 2001; Haw, Qi, & Wu, 1999) con-
sistently show that the earnings information is incorporated
into investors’ valuation of listed ?rms.
8
For instance, over
the sample period of 1995–1998, Chen et al. (2001) ?nd that
earnings and book value explain around 25% of the variation
in market value of their sample. To the best of our knowl-
edge, Chen and Wang (2004) is the only prior study to con-
sider the effect of government subsidies on the valuation of
Chinese listed ?rms. Using a sample for 1997–2000, they
?nd no evidence that subsidies are value relevant despite
their ?nding that it is the most persistent below-the-line
item. However, given the importance of state subsidies in
the context of Chinese listed ?rms and the rapid change of
Chinese economy, their value relevance deserves to be stud-
ied using a longer and more recent period sample.
Interviews and hypotheses development
Interview design
To gain a better understanding of Chinese government
subsidies policy and to motivate our hypotheses for the
large sample econometric analyses, we carried out 12 inter-
views of people familiar with this issue. The interviews
were held over the telephone and varied in length, ranging
from 20 min to 50 min. Based on existing literature (e.g.
Chen & Li, 2001; Chen et al., 2008; Tang & Luo, 2007; Zhu
& Chen, 2009) and anecdotal evidence, a pre-determined
interview protocol was designed which included a set of
open-ended questions. Appendix B provides further details
on these questions. We were aware that some of our ques-
tions are politically sensitive, and that some of the intervie-
wees may have been reluctant to respond to such questions
because of their position. We addressed this issue in several
ways: First, we reserved the ‘‘more controversial’’ questions
for later in the interview, consistent with Sudman and
Bradburn (1983). Second, we tried to ask the questions tact-
fully. Third, we assessed some answers through the tone of
the interviewees such as ‘‘very likely’’, ‘‘hard to say’’, or ‘‘the
possibility cannot be excluded’’. Finally, to ensure the accu-
racy of the responses, we conducted additional checks to
con?rm the answers until the interviewees veri?ed our
understanding of their responses.
Our selectedinterviewees consist of twelve peoplewhoare
familiar with government subsidies in China. They include
government of?cials, accountants, entrepreneurs, academics,
and analysts, who are based in cities like Beijing and Chongq-
ing, or provinces such as Shandong and Shaanxi. Appendix C
Panel A presents the details of interviewees and the date of
their interviews. The interviews themselves were undertaken
by our Chinese co-author. The analysis of the interviews was
conducted by all three authors. The interviewquestions were
initially designed in English. Because the interviewees are all
native Chinese speakers, the questions were askedandthe an-
swers were recorded in Chinese (Mandarin). Finally, the Chi-
nese co-author translated the Chinese answers into English.
Interview results
What conditions must ?rms meet to obtain government
subsidies?
The most commonly noted reason for Chinese govern-
ments to provide subsidies is to encourage the develop-
ment of speci?c businesses or industries. This was cited
by nine of interviewees. In other words, the businesses or
industries that the Chinese government prefers to promote
growth are more likely to receive subsidies. The response
from one of the accountants illustrated this point:
‘For ?rms to qualify for subsidies, their research and
development expenses must exceed a given threshold,
or their products must be within the high technology
?elds as stipulated in of?cial guidelines.’
Another important consideration that in?uences
whether or not ?rms receive subsidies is the personal con-
nections (or guanxi) between entrepreneurs and of?cials.
7
For more literature review on this topic, see Barth, Beaver, and
Landsman (2001).
8
These studies examine the relevance of accounting information to the
value of domestic shares (also known as A-shares), which are issued by vast
majority of Chinese listed ?rms. Other studies such as Bao and Chow (1999)
and Sami and Zhou (2004) focus on a smaller group of Chinese listed ?rms
that also issue foreign shares (also known as B-shares), and produce
broadly similar ?ndings that earnings are value relevant.
152 E. Lee et al. / Accounting, Organizations and Society 39 (2014) 149–169
This is especially true for the non-tax based subsidies such
as direct cash grants, loan guarantees, and debt forgive-
ness. For instance, one of the entrepreneurs fromShandong
province suggested that:
‘If the local government has budget limitations, then the
politically connected producers are more likely to be
subsidized than their non-connected peers.’
How often do ?rms receive subsidies from governments?
A majority of interviewees stated that the frequency of
receiving subsidies varies across industries as well as the
type of projects the ?rms used for application. This was ci-
ted by six of the interviewees. One of the entrepreneurs
from Shandong province suggested that the frequency
may depend on the channel of subsidies granted:
‘When the subsidies are received in a form of non-tax
rebates, it is likely to be less persistent because such
form of subsidies is often based on occasional budgets,
which in turn is affected by the availability of ?nancial
resource of the government.’
As to the question whether central or local government
provides subsidies more frequently, most of the intervie-
wees who answered this question chose the former. This
was cited by ?ve of the six interviewees, while the remain-
ing four did not provide a clear answer. The of?cial from
Chongqing explained this issue from the perspective of
governments:
‘The central government provides subsidies more fre-
quently than local governments because the former
manages much more ?scal resources.’
Which kind of subsidies is more likely to be fair and less
subjective?
Most of the participants of our interviews agreed that
tax based subsidies are granted on a more objective basis
and are less likely to be in?uenced by the discretion of gov-
ernment of?cials. This was cited by eight of the intervie-
wees. For instance, the entrepreneur from Chongqing also
suggested that:
‘Tax based subsidies are relatively more transparent
and the application procedure is also easier than for
other forms of subsidies.’
Interestingly, all three of?cials we interviewed did not
deny that non-tax based subsidies can be less fair and
more obscure. The comments from the Chongqing of?cial
were especially insightful:
‘Because the ?scal resources of local governments can
be limited, ?rms eligible for non-tax subsidies must
be further selected and this process can involve a cer-
tain degree of subjectivity.’
What are the costs and bene?ts of subsidies?
From the ?rms’ point of view, government subsidies are
an extra source of funding that helps them increase their
earnings. Moreover, subsidies also bring along reputational
bene?ts to the recipient ?rms. For example, the Chongqing
entrepreneur commented that:
‘Receiving subsidies imply that the ?rms have promis-
ing projects endorsed by the government. The subsidies
not only assist the ?rms ?nancially when developing
the products, but the reputational bene?t they bring
along also helps the ?rms promote the sales of these
products.’
Of course, to qualify for government subsidies, the crite-
ria often require ?rms to make initial investments such as
equipment, staff, and other expenses. Furthermore, as
mentioned earlier, personal networking and political con-
nections often improve the chances of ?rms receiving sub-
sidies. Thus, initial investment and human capital are
prerequisites of qualifying for subsidies. One of the entre-
preneurs from Shandong province weighed the bene?t
against the cost of subsidies:
‘To qualify for subsidies, ?rms may have to bear the cost
of initial investment as well as human capital. However,
bene?ts of government support should outweigh these
costs in the long run.’
What are the governments’ objectives in providing subsidies?
There are mixed responses across the interviewees over
the question of whether or not the rescue of distressed
?rms is one of the objectives of the government to offer
subsidies. The Chongqing of?cial suggested that:
‘One reason for aiding distressed ?rms is that their
bankruptcy may have substantial in?uence on local
economy or social stability through employment. How-
ever, this is carried out on a case by case basis.’
When asked if distressed ?rms are able to receive tax
based subsidies, the of?cials from Shandong province and
Chongqing gave positive answers. For example, the former
noted that:
‘Whether or not a ?rm is eligible for tax rebates has lit-
tle to do with whether it is pro?table or not. The key
point is that the ?rm must satisfy the government pol-
icies designed for tax based subsidies.’
Do subsidies really matter in analysts’ earnings forecast,
target price forecast, and stock recommendations?
Both of the analysts agreed that subsidies information is
useful for them to anticipate ?rms’ earnings, target prices,
or make stock recommendations. One of the analysts men-
tioned that:
‘We pay more attention to subsidies if it accounts for a
higher percentage of a ?rm’s revenue. We also forecast
the changes in national policies on subsidies.’
As to the question of whether subsidies are more useful
for analysts to forecast earnings, target prices, or make rec-
ommendations for ?nancially healthy ?rms than distressed
?rms, both analysts provided consistently positive an-
swers. One of them explained that:
E. Lee et al. / Accounting, Organizations and Society 39 (2014) 149–169 153
‘For loss making ?rms, subsidies information is not use-
ful unless the magnitude received is signi?cant enough
to turn overall losses into gains. In such circumstances,
we also consider the effect on a case by case basis.’
When it comes to the question of which type of subsi-
dies, i.e. tax or non-tax based, is more useful for analysts
to make forecasts or recommendations, one of the analysts
noted that:
‘Tax based subsidies could be more important, because
it is typically more predictable than non-tax based sub-
sidies, and has greater in?uences on ?rms’ future cash
?ows.’
Hypotheses development
Based on the interview results in section ‘Interview re-
sults’, we formulate three testable hypotheses for our large
sample econometric analyses of the valuation conse-
quences of state subsidies. First, according to the ?nancial
analysts’ responses in section ‘Do subsidies really matter in
analysts’ earnings forecast, target price forecast, and stock
recommendations?’, the information on state subsidies re-
ceived by the ?rms they follow serves as a useful input to
their earnings forecasts, target price forecasts, and stock
recommendations. Financial analysts are sophisticated
users of accounting information and they serve as interme-
diaries between ?rms and investors. The fact that analysts
incorporate state subsidies in their forecasts suggests that
this information is systematically relayed to investors in
the capital market. It also implies that state subsidies in-
deed capture the attention of informed investors and
may be widely considered in their investment decisions.
This is not surprising since state subsidies directly increase
the receiving ?rm’s cash ?ows, and may also indirectly
contribute to a decrease in the cost of further capital acqui-
sition. Furthermore, state subsidies also signals to inves-
tors that the receiving ?rm has good prospects and less
uncertainty since it has government backing. These are
broadly consistent with the interview ?ndings in section
‘What are the costs and bene?ts of subsidies?’. As a result,
based on the classical discounted cash ?ow (DCF) model,
subsidies should increase a ?rm’s value. Given these argu-
ments, we hypothesize that:
H1. On average, government subsidies are value relevant
for Chinese listed ?rms.
Next, based on our interview ?ndings in sections ‘What
conditions must ?rms meet to obtain government subsi-
dies?’ and ‘What are the governments’ objectives in pro-
viding subsidies?’, subsidies are granted to encourage
?rms to focus on sectors, regions, or projects that are prior-
itized by the government, or to rescue distressed ?rms in
order to stabilize the local economy and increase job secu-
rity. However, our interview ?ndings in section ‘How often
do ?rms receive subsidies from governments?’ also reveal
that the budget limitations of local governments can affect
their ability to grant subsidies. This in turn could affect the
ability of local government to aid distressed ?rms. In other
words, government is less likely to provide subsidies to
distressed ?rms on a persistent basis. As such, the subsi-
dies granted to rescue ?rms in distress is less likely to
in?uence the long term performance of such ?rms. Aside
from that, distressed ?rms are less likely to have growth
opportunities to make use of the additional capital from
subsidies. The fact that ?rms are in distress in the ?rst
place implies a fundamental reduction in their earnings
generating capacity (Schleicher, Hussainey, & Walker,
2007). Prior empirical literature also supports this by
showing that distressed ?rms reporting a pro?t by virtue
of government subsidies often move back to losses again
in the near future (Cheng, Aerts, & Jorissen, 2010). Further-
more, state subsidies to distressed ?rms are more likely to
be used to arti?cially in?ate reported earnings to avoid
delisting (Chen et al., 2008), which undermine the credibil-
ity of reported earnings to outside investors in the stock
market. As a result, we expect government subsidies to
have lower consequences for ?rm value for distressed
?rms compared to ?nancially healthy ?rms. Given these
arguments, we hypothesize that:
H2. Government subsidies are less value relevant for the
?rms that are in distress.
Finally, our interview ?ndings in section ‘Which kind of
subsidies is more likely to be fair and less subjective?’ also
highlighted the differences between the two main chan-
nels of subsidies, i.e. tax and non-tax based. Tax based sub-
sidies are driven by established policies and guidelines
while non-tax based subsidies are more dependent on
the discretion of of?cials. The transparency of tax based
subsidies renders them more predictable to investors,
which in turn facilitates the incorporation of this informa-
tion into their valuation decisions. The criteria for such
subsidies also implies that a ?rm is associated with sectors,
regions, and projects encouraged by government, which
provides a positive signal about the ?rm’s prospects. In
contrast, although non-tax based subsidies may signal that
political connections are enjoyed by the recipient ?rms,
the subjectivity of their allocation by of?cials could also
make such subsidies less predictable to investors. Also,
the frequency of non-tax based subsidies may be reduced
by ?scal constraints, which tends to be greater among local
governments. A lower frequency of grants implies a lower
persistence for non-tax based subsidies. This in turn re-
duces the impact of such subsidies on the long-term value
of the recipient ?rms. Given these arguments, we hypoth-
esize that:
H3. Subsidies through tax rebates are more value relevant
than other forms of government subsidies.
Modeling the value relevance of government subsidies
Research design
To test our three hypotheses, we apply the well-estab-
lished accounting based valuation models that originated
from Ohlson (1995, 1999). Such models are based on the
assumption that stock prices fully re?ect publically avail-
able information. Existing studies (e.g. Firth, Fung, & Rui,
154 E. Lee et al. / Accounting, Organizations and Society 39 (2014) 149–169
2006; Long, Payne, & Feng, 1999) suggest that the Chinese
stock market had achieved a reasonable degree of informa-
tional ef?ciency before the early 2000s. Since that time the
Chinese stock market has developed at great pace, which is
in line with the country’s rapid economy growth and the
government’s continued reforms toward a market-based
economy. For instance, listed ?rms were mandated with
new governance requirements in 2001 and tighter disclo-
sure rules in 2004 (Jing, 2009). Furthermore, reforms on
the state ownership of listed ?rms were instigated in
2006 (Chen, Li, Xin, & Yeung, 2012), and listed ?rms were
mandated to employ new accounting standards that were
largely converged to International Financial Reporting
Standards (IFRS) from 2007 onward (Lee, Walker, & Zeng,
2013). Meanwhile, there has been continued growth in for-
eign institutional investors, the stock broking industry, and
the ?nancial media (e.g. Chen, Du, Li, & Ou-Yang, 2013; Gu,
Li, & Yang, 2012). Based on these arguments, we believe
that China’s stock market is suf?ciently ef?cient to justify
the accounting based valuation model we apply to exam-
ine the relationship between ?rm value and the subsidies
they receive. Nevertheless, it is possible that the Chinese
stock market is not as informationally ef?cient as the US
stock market on which most of the work on these models
has been carried out. However, it is important to note that,
to the extent Chinese stock market lacks informational ef?-
ciency, this would work against us ?nding statistically sig-
ni?cant results consistent with our hypotheses. In
particular, if the market failed to take notice of the infor-
mation content of subsidies that is available in ?nancial
statements, then we should ?nd no relation between ?rm
value and subsidies.
9
Thus, the extent to which we ?nd that
subsidies are value relevant is, if anything, likely to be an
underestimate of their true value relevance.
Total value relevance tests
The methodology we use is based on Barth et al. (1999)
and Landsman et al. (2007), which model earnings compo-
nents by extending the linear information system devel-
oped in Ohlson (1999). The linear information system
comprises four equations:
NI
a
j;t
¼ x
0
þx
1
NI
a
j;tÀ1
þx
2
S
j;tÀ1
þx
3
BV
j;tÀ1
þe
1j;t
ð1Þ
S
j;t
¼ h
0
þh
1
S
j;tÀ1
þh
2
BV
j;tÀ1
þe
2j;t
ð2Þ
BV
j;t
¼ b
0
þb
1
BV
j;tÀ1
þe
3j;t
ð3Þ
P
j;t
¼ a
0
þa
1
NI
a
j;t
þa
2
S
j;t
þa
3
BV
j;t
þe
4j;t
ð4Þ
Eqs. (1)–(3) capture the information dynamics of the three
variables that are assumed to be the main drivers of ?rm
value. These are abnormal earnings, state subsidies, and
book value respectively. Eq. (4) is the valuation equation
that relates share price to the three main valuation drivers.
A novel feature of this study is that we explicitly model the
information dynamics of the three valuation drivers for
Chinese ?rms, and we relate the estimated coef?cients of
the information dynamics to the coef?cients in the valua-
tion equation.
Eq. (1) is the abnormal earnings prediction equation,
where abnormal earnings NI
a
t
_ _
is de?ned as earnings
(NI
t
) less a normal return on equity book value
(r  BV
tÀ1
).
10
S
tÀ1
is subsidies in year t À 1. The coef?cient
on subsidies (x
2
) re?ects the incremental ability of subsi-
dies to forecast abnormal earnings. If all earnings compo-
nents have the same ability to forecast future abnormal
earnings, then x
2
should be zero. Since subsidies are part
of abnormal earnings, its total forecasting relevance coef?-
cient should be x
1
þx
2
. Thus, subsidies are forecasting
irrelevant only if x
1
þx
2
¼ 0. Eq. (2) describes the time ser-
ies properties, or persistence, of subsidies. If subsidies are
not entirely transitory, then the coef?cient h
1
should be sig-
ni?cantly different from zero. Following Barth et al. (1999)
and Landsman et al. (2007), Eq. (3) is included to preserve
the triangular information structure of the generalized ver-
sion of Ohlson’s (1999) model. This triangular structure en-
sures that, in theory, parameters relating to book value have
no effect on the valuation coef?cients on abnormal earnings
and subsidies in Eq. (4) (Barth et al., 1999). Like previous
studies, we do not report its regression results for brevity.
Finally, Eq. (4) is the valuation model based on the
information dynamics in Eqs. (1)–(3). The value relevance
coef?cient on subsidies is a
2
. In parallel with the interpre-
tation of x
2
in Eq. (1), a
2
indicates whether the subsidies
have an incremental valuation effect. If subsidies do not
have incremental explanatory power for equity value (P
jt
)
beyond other earnings components, then a
2
should be
zero. Also analogous to Eq. (1), a
1
þa
2
re?ects the total va-
lue relevance of subsidies. If a
1
þa
2
¼ 0, the subsidies
would be value irrelevant.
A comprehensive list of the de?nitions of the variables
used in our econometric analyses is provided in Appendix
D. We estimate Eqs. (1)–(4) using Seemingly Unrelated
Regressions, which allows regression errors to be corre-
lated across equations. Moreover, we do this with industry
and year ?xed-effects controlled. The standard errors in
the regressions are also corrected for heteroskedasticity.
Finally, following Dechow et al. (1999) and Landsman
et al. (2007), we estimate all models using per share data.
9
To evaluate whether subsidies are systematically mispriced by inves-
tors, we carried out an additional test to examine the relationship between
current subsidies over book value and subsequent returns. The ?nding from
this analysis (untabulated for brevity) suggests that investors do not
misprice the information contained in the state subsidies variable. This
supports the argument that the market is ef?cient with regard to the
information content of subsidies since it is quickly impounded into stock
prices.
10
The variable r denotes required return (or cost of capital). Following the
consistent research design of prior literature (e.g. Barth et al., 1999;
Dechow, Hutton, & Sloan, 1999; Landsman et al., 2007) we apply a uniform
level of cost of capital for all ?rms, and we did not separately estimate it on
a ?rm-speci?c basis. In practice, the estimation of ?rm-speci?c implied cost
of capital requires analysts’ earnings forecast data, which is not available
for a large cross-section of ?rms in China (Chen, Chen, Lobo, & Wang, 2011).
To avoid reduction in sample size and to prevent potential selection bias
issue, our analyses do not use ?rm-speci?c implied cost of capital. In our
study, we apply a 12% cost of capital across Chinese ?rms, and this follows
previous research on China’s stock markets (e.g. Shen, 2007). As sensitivity
analyses, we replicated our main analyses by using different level of cost of
capital, including 5%, 10% and 15%, and our inferences are unaffected.
E. Lee et al. / Accounting, Organizations and Society 39 (2014) 149–169 155
Value relevance tests for ?nancially healthy and distressed
?rms
To test hypothesis H2, we partition the sample into
?nancially healthy and distressed ?rms by using a version
of the Z-score as adapted to the Chinese market by Altman,
Zhang, and Yen (2007). Since the pioneering research of
Altman (1968), a number of studies have used the Z-score
to identify potentially distressed ?rms (e.g. Hillegeist,
Keating, Cram, & Lundstend, 2004; Summers & Sweeney,
1998). Despite the popularity of the Z-score used in failure
prediction, most of the extant research focuses only on
developed markets, which might not be applicable to
emerging markets such as China due to its unique institu-
tional context. To address this issue, Altman et al. (2007)
develop a particular model to support identi?cation of po-
tential distress ?rms in China. The modi?ed Z-score model
is of the form:
Z-score ¼ 0:517 À0:460x
1
þ9:320x
2
þ0:388x
3
þ1:158x
4
ð5Þ
where x
1
is the liability to asset ratio (total liabilities/total
assets), x
2
is the rate of return on total assets (net pro?t/to-
tal assets), x
3
is working capital to total asset ratio ((cur-
rent asset minus current liabilities)/total assets), and x
4
is
retained earnings to total assets ratio (retained earnings/
total assets). We calculate the Z-score for Chinese listed
?rms following Eq. (5) and classify a ?rm as a distressed
?rm in the current year if its Z-score in the previous year
is negative.
In order to increase con?dence in our results, we use
two other proxies for distress as a sensitivity check. First,
we de?ne Low as one if the ?rm’s return on equity (ROE)
is below the 20th percentile in each year, and zero other-
wise. Since the regulatory requirements in China are
mainly based on accounting earnings, out of a fear of being
delisted, the listed ?rms that have relatively lower pro?ts
are more likely to use subsidies to help them window dress
their underperformance. Second, we de?ne ST as one if the
?rm is specially treated (see discussion in section ‘Institu-
tional setting’) in that year, and zero otherwise. Firms with
delisting avoidance motives are more likely to use subsi-
dies to manipulate earnings.
Value relevance tests of government subsidies components
So far we have assumed that the value relevance of dif-
ferent types of subsidies is identical. As mentioned earlier,
the government can grant subsidies through either tax or
non-tax based channels. To test hypothesis H3, we separate
subsidies into these two components and replicate Eqs.
(1)–(4) by replacing total subsidies with these compo-
nents. Speci?cally, tax based subsidies include the rebates
of various taxes such as value-added tax, consumption tax,
export tax, and income tax. Among them, value-added tax
rebates account for the largest proportion. This is consis-
tent with of?cial statistics disclosed by the Ministry of Fi-
nance as of 2010.
11
Prior literature (e.g. Tang & Luo, 2007)
also documents a similar ?nding. Non-tax based subsidies
are largely provided through direct cash grants, loan guaran-
tees, and debt forgiveness.
Sample selection and data
The data used in this study other than subsidies is ob-
tained from the China Securities Markets and Accounting
Research Database (CSMAR). The subsidies data, including
the amount of total subsidies, as well as both tax and non-
tax based subsidies, are hand-collected from annual re-
ports. We begin our sample selection with all manufactur-
ing ?rms listed on the Shanghai and Shenzhen Stock
Exchanges from 2002 to 2008.
12
We focus on the manufac-
turing sector for two reasons. First, a single sector reduces
the heterogeneity of ?rm characteristics in the sample,
which may introduce noise into our analyses. Second, man-
ufacturing is by far the largest and most important sector in
the Chinese economy.
13
Appendix C Panels B to D present the sample we use for
the regression analyses. Panel B presents the sample selec-
tion process. Of the 5646 observations available for the
2002–2008 sample period, we eliminate 748 observations
with insuf?cient data to test value relevance, leaving us
with a sample of 4898 ?rm-year observations to test
hypotheses H1 and H2. In order to test hypothesis H3, we
had to delete a further 582 observations, because they do
not have suf?cient information to distinguish tax rebates
from other sources of subsidies. The ?nal sample for test-
ing hypothesis H3 therefore consists of 4316 ?rm-year
observations. Panel C presents the distribution of observa-
tions across the sample years. It shows that around 70% of
the ?rm-years are subsidized, which reveals the pervasive-
ness of this practice in China. Panel D reports the distribu-
tion of ?rm-years across sub-industries, which are based
on the CSRC’s classi?cation. Among the ten sub-industries,
machinery accounts for the largest proportion with 1306
total ?rm-years, of which 918 are subsidized.
Empirical results
Summary statistics
Table 1 presents summary statistics on the variables
used in our econometric analyses. Panel A is based on the
main sample used to test hypotheses H1 and H2. The mean
and median of stock price are 8.199 and 5.990 respectively.
11
Please see:http://www.mof.gov.cn/zhuantihuigu/czjbqk2010/2cztz/
201110/t20111031_603357.html (in Chinese).
12
Our subsidies data begin for 2000 after the data for this variable
becomes more available. Because the calculation of abnormal earnings
requires the book value of the previous year, the observations in 2000 had
to be excluded. Moreover, we had to exclude observations in 2001 because
the information dynamics test of abnormal earnings cannot be carried out
without the data for abnormal earnings in 2000. Thus the years examined
in this study are 2002–2008.
13
According to a report released by IHS Global Insight, one of the world’s
largest economic advisory ?rms, China accounted for a 19.8% share of
global manufacturing in 2010, and became the largest manufacturing
nation in the world by surpassing the US (http://www.?nfacts.i.e./irish?-
nancenews/article_1021835.shtml). Over the period 2000–2005, manufac-
turing accounted for 32% of China’s GDP (Hanson & Robertson, 2008). Also
manufacturing ?rms accounted for around 50% of the market value of
China’s stock market, making it more specialized in that sector than any
other large emerging economy.
156 E. Lee et al. / Accounting, Organizations and Society 39 (2014) 149–169
On the other hand, the mean and median of book value per
share are only 2.948 and 2.808 respectively. This suggests
that book value alone is insuf?cient to explain the stock
price and therefore other value relevant variables are re-
quired. Both the mean and the median of abnormal earn-
ings per share are negative, i.e. À0.213 and À0.184
respectively. This result indicates that during our sample
period, on average, Chinese manufacturing ?rms failed to
generate a normal rate of return on capital. The mean
and median of subsidies per share are positive, i.e. 0.017
and 0.002 respectively. Panel B is based on the sub-sample
used to test hypothesis H3. Unreported results show that
the mean and median of stock price, book value, and
abnormal earnings are not materially different between
Panels A and B. This suggests that the main sample and
the sub-sample are comparable and that sample selection
bias is not an issue in testing hypothesis H3. Panel B shows
that the mean and median of non-tax based subsidies are
higher than tax based subsidies.
Simple correlations
Table 2 presents the bivariate correlations of our study
variables. Panel A is based on the main sample used to test
hypotheses H1 and H2. Book value, abnormal earnings and
subsidies are all signi?cantly positively correlated with
stock price. This is consistent with all three variables being
possible value drivers for Chinese listed ?rms. There is a
negative correlation between stock price and Zscore. Panel
B is based on the sub-sample used to test hypothesis H3.
The results for stock price, book value, and abnormal earn-
ings are broadly similar to those reported in Panel A. Both
tax based and non-tax based subsidies are signi?cantly
positively correlated with stock price. Finally, since the
correlations among the independent variables are less than
0.7, multicollinearity should not be a concern in this
study.
14
Test of hypothesis H1
Table 3 presents our regression analyses using the main
sample. Panel A reports the ?ndings for our forecast equa-
tion, which is based on Eq. (1). We observe the coef?cient
on abnormal earnings (x
1
) is positive and statistically sig-
ni?cant at the 1% level. This is consistent with our expecta-
tion as well as prior research (e.g. Barth, Beaver, Hand, &
Landsman, 2002; Dechow et al., 1999; Landsman et al.,
2007). In the same regression, the incremental forecasting
coef?cient of subsidies (x
2
) is insigni?cantly different
from zero (À0.036, Z-stats = À0.28) while the total fore-
casting coef?cient of subsidies, i.e. x
1
þx
2
, is signi?cantly
positive (0.321, Z-stats = 2.43). Taken together, these ?nd-
ings suggest that subsidies are helpful in forecasting future
abnormal earnings, although their forecasting relevance is
similar to other earnings components.
Panel B of Table 3 presents our analyses of Eq. (2). It
indicates that subsidies are signi?cant and highly persis-
tent across time. This speci?c result is broadly consistent
with the ?ndings of Chen and Wang (2004) over an earlier
sample period (1997–2000) that subsidies are the most
persistent among all the below-the-line items.
Panel C of Table 3 presents the value relevance analyses
for our main sample based on Eq. (4), which is our direct
Table 1
Summary statistics.
Obs. Mean Median Sd. Min Max
Panel A: Variables to test hypotheses H1 and H2
Dependent variable
P
t
4898 8.199 5.990 7.203 1.830 48.020
Independent variables
BV
t
4898 2.948 2.808 1.748 À2.022 8.736
NI
a
t
4898 À0.213 À0.184 0.424 À1.977 1.084
S
t
4898 0.017 0.002 0.039 0.000 0.249
Zscore
t
4898 0.152 0.000 0.359 0.000 1.000
ST
t
4898 0.101 0.000 0.301 0.000 1.000
Low
t
4898 0.200 0.000 0.400 0.000 1.000
Panel B: Variables to test hypothesis H3
Dependent variable
P
t
4316 8.069 5.900 7.096 1.800 47.930
Independent variables
BV
t
4316 2.947 2.799 1.753 À2.017 8.741
NI
a
t
4316 À0.211 À0.185 0.420 À1.979 1.084
S
T
t
4316 0.006 0.000 0.020 0.000 0.135
S
NT
t
4316 0.010 0.000 0.026 0.000 0.166
This table presents the summary statistics of the variables used in our main analyses. Panel A is based on the main sample to test hypotheses H1 and H2.
Panel B is based on the sub-sample to test hypothesis H3. The variables are de?ned as follows: P
t
is stock price at the end of year t. BV
t
is book value per
share for year t. NI
a
t
is abnormal earnings per share for year t, de?ned as NI
t
À r
Ã
BV
tÀ1
, where NI
t
is net income per share for year t and r is cost of equity
capital. S
t
is subsidies per share for year t. S
T
t
is tax based subsidies per share for year t. S
NT
t
is non-tax based subsidies per share for year t. Zscore
t
takes a
value of one if the ?rm’s Z-score in year t À 1, calculated based on Altman et al. (2007), is positive, and zero otherwise. ST
t
is a dummy variable that equals
one if the ?rm is specially treated in year t, and zero otherwise. Low
t
is a dummy variable that equals one if the ROE of the ?rm-year is below the 20th
percentile of the industry it is operating in, and zero otherwise. All variables except dummies are winsorized at the 1% and 99% levels.
14
Lind, Marchal, and Mason (2002) point out that multicollinearity may
exist if the correlation coef?cients exceed 0.7, which is a typical threshold
to identify the presence of multicollinearity.
E. Lee et al. / Accounting, Organizations and Society 39 (2014) 149–169 157
test of hypothesis H1. It shows that a
1
þa
2
(11.618,
Z-stats = 6.24) and a
2
(6.596, Z-stats = 3.54) are both
signi?cantly positive. These results suggest that subsidies
are not only value relevant, but also have an effect
Table 2
Correlation matrices.
P
t
BV
t NI
a
t
S
t
Zscore
t
ST
t
Low
t
Panel A: Variables to test hypotheses H1 and H2
P
t
0.414
***
0.396
***
0.143
***
À0.252
***
À0.226
***
À0.285
***
BV
t
0.392
***
0.037
**
0.187
***
À0.502
***
À0.433
***
À0.339
***
NI
a
t
0.420
***
0.201
***
0.054
***
À0.004 0.071
***
À0.414
***
S
t
0.134
***
0.164
***
0.096
***
À0.089
***
À0.104
***
À0.001
Zscore
t
À0.168
***
À0.504
***
À0.075
***
0.0003 0.632
***
0.258
***
ST
t
À0.145
***
À0.462
***
0.009 0.001 0.632
***
0.156
***
Low
t
À0.211
***
À0.288
***
À0.423
***
0.049
***
0.258
***
0.156
***
P
t
BV
t NI
a
t
S
T
t
S
NT
t
Panel B: Variables to test hypothesis H3
P
t
0.419
***
0.391
***
0.054
***
0.119
***
BV
t
0.393
***
0.033
**
0.134
***
0.134
***
NI
a
t
0.423
***
0.201
***
0.011 0.054
***
S
T
t
0.094
***
0.160
***
0.060
***
0.141
***
S
NT
t
0.111
***
0.102
***
0.087
***
0.049
***
This table provides the correlation matrices of the variables used in our main analyses. Panel A is based on the main sample to test hypotheses H1 and H2.
Panel B is based on the sub-sample to test hypothesis H3. Below (above) the diagonal are the Pearson (Spearman) correlations respectively. The variables are
de?ned as follows: P
t
is stock price at the end of year t. BV
t
is book value per share for year t. NI
a
t
is abnormal earnings per share for year t, de?ned as
NI
t
À r
Ã
BV
tÀ1
, where NI
t
is net income per share for year t and r is cost of equity capital. S
t
is subsidies per share for year t. S
T
t
is tax based subsidies per share
for year t. S
NT
t
is non-tax based subsidies per share for year t. Zscore
t
takes a value of one if the ?rm’s Z-score in year t À 1, calculated based on Altman et al.
(2007), is positive, and zero otherwise. ST
t
is a dummy variable that equals one if the ?rm is specially treated in year t, and zero otherwise. Low
t
is a dummy
variable that equals one if the ROE of the ?rm-year is below the 20th percentile of the industry it is operating in, and zero otherwise. All variables except
dummies are winsorized at the 1% and 99% levels.
Ã
Indicate signi?cance at the 10% level.
***
Indicate signi?cance at the 1% level.
**
Indicate signi?cance at the 5% level.
Table 3
Value relevance analyses of subsidies: test of hypothesis H1.
Panel A: Regression of abnormal earnings on lagged abnormal earnings, subsidies, and book value
NI
a
t
¼ x
0
þx
1
NI
a
tÀ1
þx
2
S
tÀ1
þx
3
BV
tÀ1
þe
1t
NI
a
tÀ1
S
tÀ1
BV
tÀ1
Total forecasting relevance test Obs. Adj. R
2
Coef?cient Z-stats Coef?cient Z-stats Coef?cient Z-stats x
1
þx
2
(Z-stats)
0.357
***
30.41 À0.036 À0.28 À0.046
***
À13.95 0.321
**
(2.43) 4898 0.218
Panel B: Regression of subsidies on lagged subsidies, and book value
S
t
¼ h
0
þh
1
S
tÀ1
þh
2
BV
tÀ1
þe
2t
S
tÀ1
BV
tÀ1
Obs. Adj. R
2
Coef?cient Z-stats Coef?cient Z-stats
0.651
***
49.44 0.001
***
3.69 4898 0.368
Panel C: Regression of stock prices on abnormal earnings, subsidies, and book value
P
t
¼ a
0
þa
1
NI
a
t
þa
2
S
t
þa
3
BV
t
þe
t
NI
a
t
S
t
BV
t
Total value relevance test Obs. Adj. R
2
Coef?cient Z-stats Coef?cient Z-stats Coef?cient Z-stats a
1
þa
2
(Z-stats)
5.022
***
29.00 6.596
***
3.54 1.340
***
32.13 11.618
***
(6.24) 4898 0.532
This table presents the results for the test of hypothesis H1 based on our main sample. Panel A reports the forecast equation, which is the regression of
abnormal earnings of year t on abnormal earnings, subsidies and book value of year t À 1. Panel B reports the persistence equation, which is the regression
of subsidies of year t on subsidies and book value of year t À 1. Panel C reports the valuation equation, which is the regression of price on book value,
abnormal earnings and subsidies. The variables are de?ned as follows: NI
a
t
is abnormal earnings per share for year t, de?ned as NI
t
À r
Ã
BV
tÀ1
, where NI
t
is net
income per share for year t and r is cost of equity capital. S
t
is subsidies per share for year t. NI
a
tÀ1
is abnormal earnings per share for year t À 1. S
tÀ1
is
subsidies per share for year t À 1. P
t
is stock price at the end of year t. All equations are estimated using Seemingly Unrelated Regressions (SUR) with year
and industry ?xed effects. The standard errors in the regressions are corrected for heteroskedasticity. All variables are winsorized at the 1% and 99% levels.
Ã
Indicate signi?cance at the 10% level.
***
Indicate signi?cance at the 1% level.
**
Indicate signi?cance at the 5% level.
158 E. Lee et al. / Accounting, Organizations and Society 39 (2014) 149–169
incremental to the other independent variables in the
regression. The ?nding that subsidies are value relevant
is consistent with hypothesis H1. Our evidence differs from
Chen and Wang (2004), which shows that subsidies are
persistent but not value relevant. The most likely reason
for this inconsistency in ?nding is the difference in sample
period between the two studies. We apply a longer and
more recent sample period than Chen and Wang (2004).
The Chinese stock market was established only since the
early 1990s and therefore the more recent sample period
we apply studies the market in a more mature stage of
development, where subsidies information is more likely
to be ef?ciently impounded into stock prices. Another
interesting observation that stems from Panel C is that
subsidies have an incremental valuation effect, since we
show that a
2
> 0, despite them not having incremental
forecasting ability for abnormal earnings since we show
that x
2
¼ 0. One possible reason why subsidies have an
incremental effect on valuation but not abnormal earnings
is that they may in?uence the discount rates that investors
attach to ?rms. Evidence from existing literature supports
the argument that government subsidies to ?rms reduce
their cost of equity and/or debt capital (e.g. Callahan,
Vendrzyk, & Butler, 2012; Chen et al., 2011). We do not ex-
plore this possibility in this paper, but we believe that it
may be worthy of consideration in future research.
Test of hypothesis H2
Table 4 presents our regression analyses after partition-
ing the main sample into ?nancially healthy and distressed
?rms. We carry out this partition based on three different
proxies, i.e. Zscore, ST, and Low. Panel A is based on Eq.
(1) and x
1
indicates that subsidies do not have incremen-
tal ability to forecast abnormal earnings in either ?nan-
cially healthy or distressed ?rms. However, the results
for x
1
þx
2
indicate that the total forecasting relevance
of subsidies is signi?cantly positive only for ?nancially
healthy ?rms but not for their distressed counterparts.
For example, using the Zscore as a proxy for distress, the
forecasting relevance of subsidies is 0.696 (Z-stats = 5.57)
for healthy ?rms and À0.243 (Z-stats = À0.70) for dis-
tressed ?rms. This is consistent with the expectation that
healthy ?rms are more likely to make use of subsidies to
fund positive Net Present Value (NPV) projects or growth
opportunities while their distressed counterparts are more
likely to use subsidies to recover losses.
Panel B of Table 4 is based on Eq. (2) and it shows that
the persistence coef?cient (h
1
) is signi?cantly positive and
higher among ?nancially healthy ?rms than the distressed
?rms. One possible reason is that distressed ?rms are more
likely than healthy ?rms to receive one-off rescue packages
through non-tax based subsidies, as suggested through our
interview results. Thus, subsidies received through this
channel are less persistent.
Panel C of Table 4 presents value relevance analyses
based on Eq. (4), which is our direct test of hypothesis
H2. Across different proxies for ?nancial distress, we con-
sistently ?nd that the total effect (a
1
þa
2
) and the incre-
mental effect (a
2
) of subsidies on valuation are
signi?cantly positive only among the healthy ?rms and
not among the distressed ?rms. For instance, using the
Zscore as a distress proxy, a
1
þa
2
and a
2
are 14.335
(Z-stats = 6.68) and is 6.781 (Z-stats = 3.168) respectively
for healthy ?rms, and 3.598 (Z-stats = 1.50) and 2.748
(Z-stats = 1.136) respectively for distressed ?rms. The differ-
ence in a
2
between the healthy and distressed sub-samples
is also consistently statistically signi?cant. For instance,
the Z-statistic is 1.95 for the sub-sample difference when
Z-score is used as distress proxy. This implies that our ?nd-
ing that the value relevance of subsidies is weaker in dis-
tressed ?rms is not caused by the possibility of higher
standard errors due to the smaller sample size of this
group.
15
These ?ndings are consistent with hypothesis H2
and robust to the different proxies for ?nancial distress that
we apply. This suggests that investors attach a higher mar-
ket value to subsidies received by ?nancially healthy ?rms
than subsidies received by ?nancially distressed ?rms. The
asymmetric value relevance of subsidies between ?nancially
healthy and distressed ?rms indicated in Panel C is consis-
tent with underlying value drivers revealed in Panels A
and B.
Test of hypothesis H3
Table 5 presents the analysis that distinguishes be-
tween tax based and non-tax based subsidies. Panel A is
based on Eq. (1) and compares the ability of the two kinds
of subsidies to forecast abnormal earnings. The total fore-
casting coef?cient (x
1
þx
2
) is signi?cantly positive only
for the tax based subsidies and statistically insigni?cant
for the non-tax based subsidies. In other words, only tax
based subsidies are forecasting relevant for future abnor-
mal earnings. One possible reason is that ?rms that qualify
for tax based subsidies are likely to be in sectors or regions
or investing in projects encouraged by the government
through of?cial policy guidelines. These sectors, regions,
and projects prioritized by the government are more likely
to have growth opportunities, which the subsidies could
help fund. In contrast, non-tax subsidies are more likely
to have been granted as a result of political connections
or as distress bailouts.
Panel B of Table 5 reports the persistence tests based on
Eq. (2). It reveals that tax subsidies have a higher degree of
persistence than non-tax based subsidies. For instance, the
persistence coef?cient (h
1
) is 0.840 (Z-stats = 77.79) in the
case of tax based subsidies and 0.492 (Z-stats = 26.85) for
non-tax based subsidies. This ?nding is consistent with
the evidence from our interviews, which suggest that
non-tax based subsidies are less predictable and more sub-
ject to the discretion of of?cials. One issue worth noting is
that non-tax subsidies are statistically signi?cantly persis-
tent, albeit to a lower degree than tax based subsidies.
Thus, neither tax nor non-tax based subsidies can be con-
sidered as purely transitory items. This affects our choice
of methodology and makes the approach of Barth et al.
(1999) and Landsman et al. (2007) more suitable than
15
Table 2 correlation matrices reveals mixed evidence on the association
between distress and state subsidies. Therefore, the lower value relevance
of subsidies among distressed ?rms is unlikely to be caused by distressed
?rms receiving systematically lower state subsidies.
E. Lee et al. / Accounting, Organizations and Society 39 (2014) 149–169 159
direct implementation of the Ohlson (1999), which is de-
signed to examine purely transitory items.
Panel C of Table 5 presents our direct test of hypothesis
H3 through the value relevance analyses based on Eq. (4).
For the tax based subsidies, we observe that the total effect
(a
1
þa
2
) and incremental effect (a
2
) on valuation are both
signi?cantly positive, i.e. 12.987 (Z-stats = 3.51) and 7.974
(Z-stats = 2.15) respectively. For the non-tax based
subsidies we observe relatively weaker valuation effects,
i.e. a total effect of only 8.825 (Z-stats = 2.97) and a statis-
tically insigni?cant incremental effect. The observation
that tax-based subsidies have more pronounced value
Table 4
Value relevance of subsidies between healthy and distressed ?rms: test of hypothesis H2.
Distress proxy = Zscore Distress proxy = ST Distress proxy = Low
Healthy Distressed Healthy Distressed Healthy Distressed
Panel A: Regression of abnormal earnings on lagged abnormal earnings, subsidies, and book value
NI
a
t
¼ x
0
þx
1
NI
a
tÀ1
þx
2
S
tÀ1
þx
3
BV
tÀ1
þe
1t
NI
a
tÀ1
0.698
***
0.083
***
0.524
***
0.069
***
0.356
***
0.110
***
(Z-stats) (41.608) (4.005) (38.508) (2.748) (30.593) (4.690)
S
tÀ1
À0.002 À0.326 0.061 À0.391 0.094 0.274
(Z-stats) (À0.012) (À0.931) (0.471) (À0.963) (0.743) (1.042)
[0.62] [1.12] [À0.96]
BV
tÀ1
À0.030
***
À0.176
***
À0.035
***
À0.176
***
À0.038
***
À0.196
***
(Z-stats) (À8.490) (À12.622) (À9.799) (À10.382) (À12.516) (À23.766)
Constant À0.146
***
0.031 À0.115
***
À0.085 À0.007 0.120
(Z-stats) (À3.941) (0.222) (À3.021) (À0.519) (À0.197) (1.316)
forecasting relevance test 0.696
***
À0.243 0.586
***
À0.322 0.450
***
0.384
x
1
þx
2
=0 (Z-stats) (5.57) (À0.70) (4.48) (À0.80) (3.56) (1.46)
Obs. 4153 745 4404 494 3916 982
Adj. R
2
0.313 0.258 0.290 0.256 0.251 0.404
Panel B: Regression of subsidies on lagged subsidies, and book value
S
t
¼ h
0
þh
1
S
tÀ1
þh
2
BV
tÀ1
þe
2t
S
tÀ1
0.726
***
0.279
***
0.764
***
0.014 0.670
***
0.575
***
(Z-stats) (57.462) (6.081) (61.344) (0.284) (50.908) (15.626)
[4.64]
***
[14.06]
***
[1.06]
BV
tÀ1
0.002
***
0.001 0.001
***
À0.000 0.002
***
0.000
(Z-stats) (5.156) (0.942) (3.968) (À0.226) (5.695) (0.003)
Constant 0.003 0.012 0.005
*
0.001 0.007
**
0.007
(Z-stats) (1.046) (0.995) (1.694) (0.047) (2.194) (0.664)
Obs. 4153 745 4404 494 3916 982
Adj. R
2
0.482 0.068 0.497 0.030 0.442 0.230
Panel C: Regression of stock prices on abnormal earnings, subsidies, and book value
P
t
¼ a
0
þa
1
NI
a
t
þa
2
S
t
þa
3
BV
t
þe
t
NI
a
t
7.554
***
0.850
***
6.252
***
1.412
***
7.510
***
0.051
(Z-stats) (36.041) (4.360) (31.578) (6.346) (30.664) (0.316)
S
t
6.781
***
2.748 7.237
***
À1.066 9.669
***
0.205
(Z-stats) (3.168) (1.136) (3.509) (À0.396) (4.053) (0.131)
[1.95]
*
[2.91]
***
[3.17]
***
BV
t
1.372
***
0.626
***
1.329
***
0.539
***
1.381
***
0.407
***
(Z-stats) (25.873) (7.654) (26.204) (5.777) (28.377) (6.825)
Constant 4.904
***
4.173
***
4.461
***
4.061
***
4.261
***
3.763
***
(Z-stats) (8.572) (5.008) (7.974) (4.527) (7.182) (6.871)
value relevance test 14.335
***
3.598 13.490
***
0.345 17.179
***
0.255
a
1
þa
2
¼ 0 (Z-stats) (6.68) (1.50) (6.52) (0.13) (7.23) (0.16)
Obs. 4153 745 4404 494 3916 982
Adj. R
2
0.556 0.487 0.544 0.533 0.560 0.470
This table presents the results for the test of hypothesis H2, based on our main sample, which is partitioned into ?nancially healthy and distressed ?rms.
Panel A reports the forecast equation, which is the regression of abnormal earnings of year t on abnormal earnings, subsidies and book value of year t À 1.
Panel B reports the persistence equation, which is the regression of subsidies of year t on subsidies and book value of year t À 1. Panel C reports the
valuation equation, which is the regression of price on book value, abnormal earnings and subsidies. The variables are de?ned as follows: NI
a
t
is abnormal
earnings per share for year t, de?ned as NI
t
À r
Ã
BV
tÀ1
, where NI
t
is net income per share for year t and r is cost of equity capital. BV
t
is book value per share for
year t. NI
a
tÀ1
is abnormal earnings per share for year t À 1. BV
tÀ1
is book value per share for year t À 1. S
t
is subsidies per share for year t. S
tÀ1
is subsidies per
share for year t À 1. P
t
is stock price at the end of year t. Zscore
t
takes a value of one if the ?rm’s Z-score in year t À 1, calculated based on Altman et al.
(2007), is positive, and zero otherwise. ST
t
is a dummy variable that equals one if the ?rm is specially treated in year t, and zero otherwise. Low
t
is a dummy
variable that equals one if the ROE of the ?rm-year is below the 20th percentile of the industry it is operating in, and zero otherwise. All equations are
estimated using Seemingly Unrelated Regressions (SUR) with year and industry ?xed effects. The numbers reported in the brackets are Z-statistics for the
differences in coef?cients on subsidies between the two sub-samples. The standard errors in the regressions are corrected for heteroskedasticity. All
variables are winsorized at the 1% and 99% levels.
***
Indicate signi?cance at the 1% level.
**
Indicate signi?cance at the 5% level.
*
Indicate signi?cance at the 10% level.
160 E. Lee et al. / Accounting, Organizations and Society 39 (2014) 149–169
Table 5
Value relevance analyses of tax and non-tax based subsidies: test of hypothesis H3.
Panel A: Regression of abnormal earnings on lagged abnormal earnings, subsidies components, and book value
NI
a
t
¼ x
0
þx
1
NI
a
tÀ1
þx
2
S
T
tÀ1
þx
3
BV
tÀ1
þe
1t
NI
a
t
¼ x
0
þx
1
NI
a
tÀ1
þx
2
S
NT
tÀ1
þx
3
BV
tÀ1
þe
1t
NI
a
tÀ1
S
T
tÀ1
BV
tÀ1
Total forecasting
relevance test
Obs. Adj.
R
2
NI
a
tÀ1
S
NT
tÀ1
BV
tÀ1
Total forecasting
relevance test
Obs. Adj.
R
2
Coef. Z-
stats
Coef. Z-
stats
Coef. Z-stats x
1
þx
2
(Z-stats) Coef. Z-
stats
Coef. Z-
stats
Coef. Z-stats x
1
þx
2
(Z-stat)
0.361
***
28.19 0.071 0.28 À0.044
***
À12.71 0.432
*
(1.67) 4316 0.224 0.367
***
28.90 À0.441
*
À1.79 À0.044
***
À12.72 À0.074 (À0.30) 4316 0.225
Panel B: Regression of subsidies components on lagged subsidies components, and book value
S
T
t
¼ h
0
þh
1
S
T
tÀ1
þh
2
BV
tÀ1
þe
2t
S
NT
t
¼ h
0
þh
1
S
NT
tÀ1
þh
2
BV
tÀ1
þe
2t
S
T
tÀ1
BV
tÀ1
Obs. Adj. R
2
S
NT
tÀ1
BV
tÀ1
Obs. Adj. R
2
Coef. Z-stats Coef. Z-stats Coef. Z-stats Coef. Z-stats
0.840
***
77.79 0.0002
**
2.03 4316 0.603 0.492
***
26.85 0.001
***
3.47 4316 0.200
Panel C: Regression of stock prices on abnormal earnings, subsidies components, and book value
P
t
¼ a
0
þa
1
NI
a
t
þa
2
S
T
t
þa
3
BV
t
þe
t
P
t
¼ a
0
þa
1
NI
a
t
þa
2
S
NT
t
þa
3
BV
t
þe
t
NI
a
t
S
T
t
BV
t
Total value relevance test Obs. Adj. R
2
NI
a
t
S
NT
t
BV
t
Total value relevance test Obs. Adj. R
2
Coef. Z-
stats
Coef. Z-
stats
Coef. Z-
stats
a
1
þa
2
(Z-stats) Coef. Z-
stats
Coef. Z-
stats
Coef. Z-
stats
a
1
þa
2
(Z-stats)
5.013
***
27.39 7.974
**
2.15 1.315
***
30.23 12.987
***
(3.51) 4316 0.535 5.001
***
27.26 3.824 1.28 1.324 30.64 8.825
***
(2.97) 4316 0.534
This table presents the results for the test of hypothesis H3, based on our restricted sample. Panel A reports the forecast equation, which is the regression of abnormal earnings of year t on abnormal earnings,
subsidies and book value of year t À 1. Panel B reports the persistence equation, which is the regression of subsidies of year t on subsidies and book value of year t À 1. Panel C reports the valuation equation, which
is the regression of price on book value, abnormal earnings and subsidies. The variables are de?ned as follows: NI
a
t
is abnormal earnings per share for year t, de?ned as NI
t
À r
Ã
BV
tÀ1
, where NI
t
is net income per
share for year t and r is cost of equity capital. NI
a
tÀ1
is abnormal earnings per share for year t À 1. BV
t
is book value per share for year t. BV
tÀ1
is book value per share for year t À 1. S
T
t
and S
T
tÀ1
is tax-based subsidies
per share for year t and t À 1, respectively. S
NT
t
and S
NT
tÀ1
is non-tax based subsidies per share for year t and t À 1, respectively. P
t
is stock price at the end of year t. All equations are estimated using Seemingly
Unrelated Regressions (SUR) with year and industry ?xed effects. The standard errors in the regressions are corrected for heteroskedasticity. All variables are winsorized at the 1% and 99% levels.
***
indicate signi?cance at the 1% level.
**
indicate signi?cance at the 5% level.
*
indicate signi?cance at the 10% level.
E
.
L
e
e
e
t
a
l
.
/
A
c
c
o
u
n
t
i
n
g
,
O
r
g
a
n
i
z
a
t
i
o
n
s
a
n
d
S
o
c
i
e
t
y
3
9
(
2
0
1
4
)
1
4
9
–
1
6
9
1
6
1
Table 6
Joint effect of ?rm characteristics and subsidies channel.
Panel A: Regression of stock prices on abnormal earnings, subsidies components, and book value for healthy ?rms
P
t
¼ a
0
þa
1
NI
a
t
þa
2
S
T
t
þa
3
BV
t
þe
t
P
t
¼ a
0
þa
1
NI
a
t
þa
2
S
NT
t
þa
3
BV
t
þe
t
NI
a
t
S
T
t
BV
t
Total value relevance test Obs. Adj. R
2
NI
a
t
S
NT
t
BV
t
Total value relevance test Obs. Adj. R
2
Coef. Z-stats Coef. Z-stats Coef. Z-stats a
1
þa
2
(Z-stats) Coef. Z-stats Coef. Z-stats Coef. Z-stats a
1
þa
2
(Z-stats)
7.415
***
29.02 9.840
**
2.32 1.375
***
27.27 17.255
***
(4.07) 3455 0.564 7.378
***
28.73 7.202
*
1.78 1.378
***
27.38 14.580
***
(3.62) 3455 0.563
Panel B: Regression of stock prices on abnormal earnings, subsidies components, and book value for distressed ?rms
P
t
¼ a
0
þa
1
NI
a
t
þa
2
S
T
t
þa
3
BV
t
þe
t
P
t
¼ a
0
þa
1
NI
a
t
þa
2
S
NT
t
þa
3
BV
t
þe
t
NI
a
t
S
T
t
BV
t
Total value relevance test Obs. Adj. R
2
NI
a
t
S
NT
t
BV
t
Total value relevance test Obs. Adj. R
2
Coef. Z-stats Coef. Z-stats Coef. Z-stats a
1
þa
2
(Z-stats) Coef. Z-stats Coef. Z-stats Coef. Z-stats a
1
þa
2
(Z-stats)
0.047 0.28 11.855
***
2.89 0.344
***
5.67 11.902
***
(2.91) 861 0.469 0.085 0.51 À1.847 À0.83 0.381
***
6.34 À1.762 (À0.80) 861 0.466
This table presents the results of further analyses that examine the joint effect of ?rm characteristics and subsidies channel. A ?rm is classi?ed as a distressed ?rm if the ROE of the ?rm-year is below the 20th
percentile of the industry it is operating in, and healthy ?rm otherwise. Panels A and B are based on healthy and distressed ?rms respectively. Each panel reports the valuation equation, which is the regression of
price on book value, abnormal earnings and subsidies components for healthy ?rms. Panel B reports the regression of price on book value, abnormal earnings and subsidies components for distressed ?rms. The
variables are de?ned as follows: NI
a
t
is abnormal earnings per share for year t, de?ned as NI
t
À r
Ã
BV
tÀ1
, where NI
t
is net income per share for year t and r is cost of equity capital. NI
a
tÀ1
is abnormal earnings per share
for year t À 1. BV
t
is book value per share for year t. BV
tÀ1
is book value per share for year t À 1. S
T
t
and S
T
tÀ1
is tax-based subsidies per share for year t and t À 1, respectively. S
NT
t
and S
NT
tÀ1
is non-tax based subsidies
per share for year t and t À 1, respectively. P
t
is stock price at the end of year t. All equations are estimated using Seemingly Unrelated Regressions (SUR) with year and industry ?xed effects. The standard errors in
the regressions are corrected for heteroskedasticity. All variables are winsorized at the 1% and 99% levels.
***
Indicate signi?cance at the 1% level.
**
Indicate signi?cance at the 5% level.
*
Indicate signi?cance at the 10% level.
1
6
2
E
.
L
e
e
e
t
a
l
.
/
A
c
c
o
u
n
t
i
n
g
,
O
r
g
a
n
i
z
a
t
i
o
n
s
a
n
d
S
o
c
i
e
t
y
3
9
(
2
0
1
4
)
1
4
9
–
1
6
9
relevance than non-tax based subsidies is consistent with
our hypothesis H3. This can be interpreted as evidence that
investors either attach a higher market value to subsidies
granted through tax based than non-tax based channel
and/or that they are more likely to make use of information
about the receipt of tax based instead of non-tax based
subsidies to formulate their valuation decisions. The asym-
metric degree of value relevance between the two catego-
ries of subsidies observed in Panel C also corresponds to
the pattern of abnormal earnings performance forecast in
Panel A and subsidies persistence in Panel B. In this case,
we demonstrate that the type of subsidies that is more
likely to be granted to ?rms with growth prospects and
on a persistent basis is also more likely to contribute to
the generation of ?rm value.
Further analyses
Joint effect of ?rm characteristics and subsidies channels
In hypotheses H2 and H3, we have separately consid-
ered the conditioning effect of ?rm characteristics and sub-
sidies channels on the valuation consequences of state
subsidies. In Table 6 we carry out additional analyses to
examine the joint effect of ?rm characteristics and subsi-
dies channels. For brevity, we only report analyses that
identify distressed ?rms through the Low measure, which
classi?es ?rms as distressed if ROE is below the 20th per-
centile each year. The Low measure enables us to classify
more ?rms into the distressed group relative to the other
two measures (Zscore and ST) that we applied in our tests
of hypothesis H2.
16
Table 6 reveals that the value relevance of tax based sub-
sidies are more pronounced than non-tax based subsidies,
and that this pattern is consistent in both healthy and dis-
tressed ?rms. For example, the incremental value relevance
indicated by a
2
for tax based subsidies is 9.840
(Z-stats = 2.32) in healthy ?rms and 11.855 (Z-stats = 2.89)
in distressed ?rms. On the other hand, for non-tax based
subsidies, the incremental value relevance is 7.202
(Z-stats = 1.78) in healthy ?rms but insigni?cant in dis-
tressed ?rms. This suggests that non-tax subsidies do not
necessarily contribute to the equity value particularly for
those inneed of bailout. This is broadly inline with interview
?ndings based on the responses from the ?nancial analysts.
Effects of growth opportunities and international orientation
Throughout the paper, we have observed that the value
relevance of state subsidies is more pronounced whenever
it is received on a more persistent basis, i.e. in healthy
Table 7
Effects of growth opportunities and international orientation.
High growth Low growth
Coef. Z-stats Coef. Z-stats
Panel A: Effect of ?rm growth on the valuation of tax-based subsidies
NI
a
t
5.951
***
18.69 2.909
***
11.70
S
T
t
25.879
***
4.12 3.722 0.62
[1.93]
*
BV
t
1.242
***
16.12 0.863
***
13.71
Constant 2.955 1.16 4.651
***
5.73
a
1
þa
2
¼ 0 (Z-stats) 31.830
***
(5.06) 6.631 (1.11)
Obs. 1428 1443
Adj. R
2
0.588 0.495
International ?rms Domestic ?rms
Coef. Z-stats Coef. Z-stats
Panel B: Effect of international orientation on the valuation of tax-based subsidies
NI
a
t
5.253
***
9.98 5.050
***
22.72
S
T
t
35.288
***
3.81 À0.674 À0.16
[2.30]
**
BV
t
1.441
***
10.81 1.335
***
25.69
Constant 6.202
***
4.28 3.552
***
5.12
a
1
þa
2
¼ 0 (Z-stats) 40.541
***
(4.37) 4.375 (1.03)
Obs. 701 2893
Adj. R
2
0.546 0.542
This table presents the results of further analyses of the conditioning effect of growth opportunities (Panel A) and international orientation (Panel B). A ?rm
is classi?ed as a high (low) growth ?rm if its average sales growth ratio over the past ?ve years is above (below) the industry and year median. Based on
prior literature (e.g. Knight & Cavusgil, 2004; O’Connor et al., 2011) a ?rm is classi?ed as an internationally oriented ?rm if its ratio of sales from
international business divided by total sales is more than 25%, and a domestically oriented ?rm otherwise. The variables are de?ned as follows: NI
a
t
is
abnormal earnings per share for year t, de?ned as NI
t
À r
Ã
BV
tÀ1
, where NI
t
is net income per share for year t and r is cost of equity capital. BV
t
is book value
per share for year t. BV
tÀ1
is book value per share for year t À 1. S
T
t
is tax-based subsidies per share for year t. P
t
is stock price at the end of year t. All
equations are estimated using Seemingly Unrelated Regressions (SUR) with year and industry ?xed effects. The numbers reported in the brackets are
Z-statistics for the differences in coef?cients on subsidies between the two sub-samples. The standard errors in the regressions are corrected for heter-
oskedasticity. All variables are winsorized at the 1% and 99% levels.
***
Indicate signi?cance at the 1% level.
**
Indicate signi?cance at the 5% level.
*
Indicate signi?cance at the 10% level.
16
When we identify distressed ?rms using other measures such as Zscore
and ST, our analyses provides similar inferences to those based on the Low
measure.
E. Lee et al. / Accounting, Organizations and Society 39 (2014) 149–169 163
?rms and through tax based channels. In addition to the
likelihood of receiving further subsidies, we conjecture
that the valuation consequence of subsidies may also be af-
fected by a ?rm’s growth opportunities or international
orientation. Firms with greater growth opportunities are
more likely to make use of the subsidies received on posi-
tive NPV projects, which contributes to value creation for
shareholders.
17
Firms that are internationally-oriented can
make use of subsidies to boost their competitiveness against
foreign contenders and help them access larger markets
abroad, which also enhances their pro?tability.
Panel A of Table 7 presents additional tests to examine
the effect of growth on the valuation consequence of sub-
sidies. We ?rst classify ?rms into high and low growth sub-
groups based on their sales growth rate over the past ?ve
years. Speci?cally, a ?rm is classi?ed as a high (low)
growth ?rm if its average sales growth ratio over the past
?ve years is above (below) the industry median. We then
carry out the full set of analyses based on Eqs. (1)–(4) for
the two groups of ?rms separately, and for brevity only
tabulate the value relevance results of Eq. (4). We now fo-
cus only on tax based subsidies since our tests of hypothe-
ses H3 reveals that they drive value relevance. Panel A
indicate that the total and incremental value relevance
effects of tax based subsidies are 31.830 (Z-stats = 5.06)
and 25.879 (Z-stats = 4.12) respectively among the high
growth ?rms and only 6.631 (Z-stats = 1.11) and 3.722
(Z-stats = 0.62) respectively among low growth ?rms. The
difference in the coef?cients of tax subsidies is statistically
signi?cant at the 10% level. This ?nding is consistent with
our conjecture that the growth opportunities of ?rms con-
tribute to the value relevance of subsidies.
Panel B of Table 7 presents additional tests to investigate
the in?uence of international orientation on the value rele-
vance of subsidies. We followprior literature (e.g. O’Connor,
Vera-Muñoz & Chan, 2011; Knight & Cavusgil, 2004) and
classify ?rms as internationally-oriented if the foreign reve-
nue divided by total sales is more than 25%, and as domesti-
cally-oriented if otherwise. For brevity, we again focus on
taxbasedsubsidies andonlytabulatethe value relevancere-
sults of Eq. (4) despite implementing the full set of analyses
based on Eqs. (1)–(4). Panel B indicates that the total and
incremental value relevance effects of tax based subsidies
are 40.541 (Z-stats = 4.37) and 35.288 (Z-stats = 3.81)
respectively among internationally-oriented ?rms and only
4.375 (Z-stats = 1.03) and À0.674 (Z-stats = À0.16) respec-
tively among domestically-oriented ?rms. The difference
in the coef?cients of tax subsidies is signi?cant at the 5%
level. This ?nding is consistent with the expectation that
internationally-oriented ?rms bene?t more from subsidies
than their domestically-oriented counterparts.
Conclusions
This study investigates the value relevance of govern-
ment subsidies in the Chinese stock market. We combine
insights from interviews of informed players along with
formal econometric tests that broadly con?rm these
insights.
The interviews were carried out to improve under-
standing of subsidies and to motivate the empirical re-
search questions. The interviews reveal three general
?ndings. First, ?nancial analysts consider ?rm-speci?c
subsidies in their research and disseminate this informa-
tion to investors in the market through forecasts of earn-
ings and target prices, as well as stock recommendations.
Second, the Chinese government grants subsidies either
to promote the speci?c sectors, regions, or projects it
encourages, or to rescue distressed ?rms to avoid job losses
and social instabilities. Finally, tax based subsidies are
more transparent, objective and predictable than non-tax
based subsidies.
The large sample econometric analyses highlight three
main results. First, subsidies are value relevant on average
in the Chinese stock market. Second, the value relevance of
subsidies is materially greater for healthy ?rms than dis-
tressed ?rms. This implies that investors anticipate greater
bene?ts from government subsidies for ?rms with pros-
pects to convert subsidies into future pro?tability. Finally,
subsidies granted through the more transparent and fair
channel, i.e. tax based subsidies, are more value relevant
than those granted through more obscure and more discre-
tionary channels, i.e. non-tax based subsidies.
From a policy maker perspective, our ?ndings suggest
that the transparency and fairness of subsidies matter. In
other words, the Chinese government should consider
focusing its resources on providing tax based subsidies
as opposed to non-tax based subsidies. From an invest-
ment practitioner perspective, our results reveal that
investors in Chinese listed ?rms can see through the pur-
pose of the subsidies received by the ?rms and price this
information accordingly. For example, whilst the use of
subsidies to boost the short term reported earnings of
distressed ?rms may help such ?rms to avoid delisting,
they do not materially increase the market values of such
?rms.
The results will also be of interest to potential interna-
tional investors in Chinese listed ?rms, the international
competitors of Chinese listed ?rms, and international pol-
iticians and regulators interested in the regulation of free
trade. Investors will be interested to know that access to
government subsidies is a signi?cant contributor to the
market valuation of Chinese ?rms. The competitors of
Chinese ?rms need to be aware that these ?rms bene?t
from subsidies, making it more dif?cult to compete with
them, and international trade regulators need to know that
Chinese state subsidies potentially ‘‘load the dice’’ in favor
of Chinese ?rms in competition with Western ?rms
(Capital Trade Incorporated, 2008). However, it needs to
be noted that other states also ?nd ways of implicitly sub-
siding their ?rms, for example the US Defense budget is
arguably a major form of state subsidies to US industries.
Of course, our study is limited to the impact of state
subsidies on the stock valuation of listed ?rms. Future
studies could evaluate whether subsidies help promote
other aspects of economic development such as the effect
on unlisted ?rms or social bene?ts. For instance, we are
17
In Appendix E we provide a more formal depiction of the relationship
between the expected value of subsidies that a ?rm receives and the ?rm’s
growth as well as likelihood of receiving further subsidies.
164 E. Lee et al. / Accounting, Organizations and Society 39 (2014) 149–169
Appendix A. Comparison of economic development and growth.
Rank GDP in 2009 GNP in 2009 using PPP Annual growth rate of GDP
(2001–2009)
Country GDP (US $ bn) Country GNP (Int’l $bn) Country Growth rate (%)
Panel A: China vs. G7 countries
1 US 14,119 US 14,011 China 10.49
2 Japan 5069 China 9170 Canada 1.76
3 China 4985 Japan 4265 US 1.56
4 Germany 3330 Germany 3017 UK 1.50
5 France 2649 UK 2217 France 1.17
6 UK 2175 France 2191 Germany 0.59
7 Italy 2113 Italy 1919 Japan 0.50
8 Canada 1336 Canada 1258 Italy 0.18
Panel B: China vs. Other major emerging economies
1 China 4985 China 9170 China 10.49
2 Brazil 1594 India 3786 India 7.56
3 India 1377 Russia 2599 Russia 4.97
4 Russia 1232 Brazil 1968 Pakistan 4.69
5 Mexico 875 Mexico 1506 Argentina 4.05
6 Argentina 307 Argentina 568 South Africa 3.57
7 South Africa 285 South Africa 496 Brazil 3.17
8 Pakistan 162 Pakistan 455 Mexico 1.39
This table presents the comparisons of economic development and growth between China and the G7 countries (Panel A), as well as between China and
other major emerging countries (Panel B). All the economic ?gures, including Gross Domestic Product (GDP), Purchasing Power Parity (PPP) adjusted Gross
National Product (GNP), and GDP growth rate, are obtained from the World Bank Development Indicator. All of China’s economic ?gures exclude Hong
Kong.
Appendix B. Interview questions.
Q1 What conditions must ?rms meet to obtain government subsidies?
(a) Do these conditions differ across ?rm type (e.g. industry, region, state/private controlled ?rms)?
(b) Do these conditions differ between central and local government?
(c) What do the governments expect of the ?rms that get subsidies?
(d) Do these expectations differ between central and local government?
Q2 How often do ?rms receive subsidies from government?
(a) What kind of ?rm is more likely to get subsidies more frequently or routinely?
(b) Is central or local government more likely to give subsidies frequently or routinely?
Q3 Which kind of subsidies is more likely to be fair and less subjective?
(a) Are tax based subsidies more likely to be fair and less subjective than non-tax based subsidies?
(b) Among tax based subsidies, are income tax rebates more likely to be fair and less subjective than other components?
Q4 What are the costs and bene?ts of subsidies?
(a) From ?rms’ point of view what are the costs and bene?ts of receiving subsidies?
(b) Do the costs and bene?ts of receiving subsidies depend on ?rm characteristics (e.g. industry, region, state/private controlled)?
(c) From governments’ point of view what are the bene?ts and costs of providing subsidies?
(d) Do the costs and bene?ts of subsidies depend on level of government, i.e. central and local?
Q5 What are the governments’ objectives in providing subsidies?
(a) Are governments more willing to provide subsidies to loss or distress ?rms for the purpose of rescue?
(b) Are governments more willing to provide subsidies to well-performing or growth ?rms to maximize return on government investment?
(c) Do these objectives differ between central or local government?
Q6 Do subsidies really matter in analysts’ earnings forecast, target price forecast, and stock recommendation?
(a) Is information on government subsidies useful for analysts to forecast listed ?rms’ earnings, target prices or make stock recommendations?
(b) Is information on government subsidies of pro?table ?rms more useful for analysts to forecast earnings, target prices or make stock recom-
mendations than that of loss ?rms?
(c) Is information on tax based subsidies more useful than that of non-tax based subsidies for analysts to forecast listed ?rms’ earnings, target
prices or make stock recommendations?
E. Lee et al. / Accounting, Organizations and Society 39 (2014) 149–169 165
Appendix C. Sample selection.
Date of Interview Occupations Regions Units
Panel A: List of interviewees
June 13th, 2011 Of?cial Shandong Fiscal bureau
September 28th, 2011 Of?cial Shaanxi Fiscal bureau
October 4th, 2011 Of?cial Chongqing Economic development authority
June 15th, 2011 Accountant Beijing A state-owned petro-chemical enterprise
July 15th, 2011 Accountant Beijing A listed state-owned manufacturing company
June 13th, 2011 Entrepreneur Shandong A cotton processing ?rm
June 13th, 2011 Entrepreneur Shandong A food production ?rm that also provides hotel service
October 10th, 2011 Entrepreneur Chongqing A privately-owned high-tech ?rm
June 4th, 2011 Professor Beijing A well-known university in China
October 6th, 2011 Assistant Professor Beijing A well-known university in China
October 31st, 2012 Analyst Beijing A leading securities company in China
October 30th, 2012 Analyst Beijing A leading securities company in China
Observations
Panel B: Sample selection process
Initial CSMAR sample from 2002 to 2008
Less: 5646
Observations with insuf?cient data to test value relevance (748)
Main sample to test hypotheses H1 and H2 4898
Less:
Observations with missing data on composition of government subsidies (582)
Sub-sample to test hypothesis H3 4316
Years Number of ?rm-years Number of subsidized ?rm-years % of subsidized ?rm-years
Panel C: The distribution of ?rm-years
2002 583 326 55.92
2003 625 364 58.24
2004 669 403 60.24
2005 706 443 62.75
2006 770 511 66.36
2007 758 594 78.36
2008 787 690 87.67
Total 4898 3331 68.01
Sub-industry CSRC
code
Number of ?rm-
years
Number of subsidized ?rm-
years
% of subsidized ?rm-
years
Panel D: Distribution of ?rm-years by sub-industry classi?cation
Food and drink C0 360 244 67.78
Textiles and apparel C1 379 252 66.49
Timber and
furnishings
C2 19 14 73.68
Paper and printing C3 164 93 56.71
Petrochemicals C4 941 629 66.84
Electronics C5 273 218 79.85
Metals and non-
metals
C6 786 498 63.36
Machinery C7 1306 918 70.29
Pharmaceuticals C8 556 387 69.60
Other manufacturing C99 114 78 68.42
166 E. Lee et al. / Accounting, Organizations and Society 39 (2014) 149–169
currently conducting a separate study on the impact of
government subsidies on the corporate social responsibil-
ity reporting of Chinese listed ?rms.
Appendix E. Drivers of the value of state subsidies
We demonstrate below that, in steady state growth, the
expected present value of government subsidies is deter-
mined by the ?rms’ growth rate, the persistence of subsi-
dies and the cost of capital. In general, the present value
of government subsidies at time t, based on discount rate
r, is:
PVS
t
¼

1
j¼tþ1
E
t
ðS
j
Þð1 þrÞ
tÀj
ðA1Þ
De?ne X
t
as the state subsidies driver (e.g. sales or prof-
its); h represents the persistence of state subsidies (i.e. the
probability that the subsidies will continue for one more
year); g denotes the ?rm’s constant growth rate. Assume
that the ?rm grows at a constant rate. Then we have:
EðS
tþ1
Þ ¼
S
t
X
t
_ _
ÂEðX
tþ1
Þ Âh
¼
S
t
X
t
_ _
ÂX
t
ð1 þgÞ Âh ¼ f ÂX
t
ð1 þgÞ Âh ðA2Þ
where f ¼
St
Xt
_ _
.
From Eq. (A2), EðS
tþ1
Þ ¼ S
t
ð1 þgÞ Âh.
Furthermore E
t
ðS
tþj
Þ ¼ S
t
ð1 þgÞ
j
Âh
j
j ¼ 1; . . . 1.
Also, assume that
ð1þgÞh
ð1þrÞ
_ _
< 1. This will be the case
provided that the ?rm’s growth rate is less than the cost
of capital, as is required for the present value to be less
than in?nity.
Substituting for EðS
j
Þ in Eq. (A1), we have:
PVS
t
¼
S
t
ð1 þgÞh
1 þr Àð1 þgÞh
ðA3Þ
Thus, Eq. (A3) implies that the value relevance of state sub-
sidies is positively associated with ?rms growth rate (g)
and the persistence of receiving subsidies (h), and nega-
tively associated with the cost of capital (r).
References
Allen, F., Qian, J., & Qian, M. J. (2005). Law, ?nance, and economic growth
in China. Journal of Financial Economics, 77(1), 57–116.
Almus, M., & Czarnitzki, D. (2003). The effects of public R&D subsidies on
?rms’ innovation activities: The case of eastern Germany. Journal of
Business and Economic Statistics, 21(2), 226–236.
Altman, E. I. (1968). Financial ratios, discriminant analysis and the
prediction of corporate bankruptcy. Journal of Finance, 23, 189–209.
Altman, E. I., Zhang, L., & Yen, J. (2007). Corporate ?nancial distress
diagnosis in China. Working paper. New York University.
Azzimonti, M., de Francisco, E., & Krusell, P. (2008). Production subsidies
and redistribution. Journal of Economic Theory, 142(1), 73–99.
Bagwell, K., & Staiger, R. W. (1989). The role of export subsidies when
product quality is unknown. Journal of International Economics, 27,
69–89.
Bagwell, K., & Staiger, R. W. (2006). Will international rules on subsidies
disrupt the world trading system? American Economic Review, 96(3),
877–895.
Balcerowicz, L. (1995). Socialism, capitalism, transformation. Budapest:
Central European University Press.
Bao, B. H., & Chow, L. (1999). The usefulness of earnings and book value
for equity valuation in emerging capital markets: Evidence from
listed companies in the People’s Republic of China. Journal of
International Financial Management & Accounting, 10(2), 85–104.
Bardhan, P., & Mookherjee, D. (2000). Capture and governance at local and
national levels. American Economic Review, 90(2), 135–139.
Barth, M. E., Beaver, W. H., & Landsman, W. R. (2001). The relevance of the
value relevance literature for ?nancial accounting standard setting:
Another view. Journal of Accounting and Economics, 31(1–3), 77–104.
Barth, M. E., Beaver, W. H., Hand, J. R. M., & Landsman, W. R. (2002).
Constraints on accrual components of earnings in equity valuation.
Working paper.
Barth, M., Hand, J., & Landsman, W. (1999). Accruals, cash ?ows, and
equity values. Review of Accounting Studies, 3, 205–229.
Boycko, M., Shleifer, A., & Vishny, R. (1995). Privatizing Russia. Cambridge,
Mass.: MIT Press.
Brander, J. A., & Spencer, B. J. (1985). Export subsidies and international
market share rivalry. Journal of International Economics, 18, 83–100.
Brennan, G., & Buchanan, J. (1980). The power to tax: Analytical foundations
of a ?scal constitution. New York: Cambridge University Press.
Appendix D. De?nitions of variables for large sample econometric analyses.
Variables Notations De?nitions
Price P
jt
Stock price of ?rm j at the end of ?scal year t
Book value BV
jt
Book value per share of ?rm j for year t
Net income NI
jt
Net income per share of ?rm j for year t
Abnormal
earnings
NI
a
jt
Abnormal earnings per share of ?rm j for year t, de?ned as NI
jt
ÀrBV
j;tÀ1
where NI
jt
is net
income per share of ?rm j for year t and r is cost of equity capital
State
subsidies
S
jt
Government subsidies per share of ?rm j for year t
Tax based
subsidies
S
T
jt
Tax rebates excluding income tax rebates per share of ?rm j for year t
Non-tax
subsidies
S
NT
jt
Government subsidies other than tax rebates per share of ?rm j for year t
Z-score Zscore
t
Dummy variable that takes a value of one if the ?rm’s Z-score in year t À 1, calculated
based on Altman et al. (2007), is negative
Specially
treated
ST
t
Dummy variable that equals one if the ?rm is specially treated in year t, otherwise zero
Low pro?t Low
t
Dummy variable that equals one if the ROE of the ?rm-year is below the 20th percentile of
the industry it is operating in for year t, and otherwise zero
E. Lee et al. / Accounting, Organizations and Society 39 (2014) 149–169 167
Callahan, C. M., Vendrzyk, V. P., & Butler, M. G. (2012). The impact of
implied facilities cost of money subsidies on capital expenditures and
the cost of debt in the defense industry. Journal of Accounting Public
Policy, 31, 301–319.
Capital Trade Incorporated (2008). An assessment of China’s subsidies to
strategic and heavyweight industries. Report submitted to US–China
Economic and Security Review Commission.
Che, J., & Qian, Y. (1998). Insecure property rights and government
ownership of ?rms. Quarterly Journal of Economics, 113(2), 467–496.
Chen, H., Chen, J. Z., Lobo, G. J., & Wang, Y. (2011). Effects of audit quality
on earnings management and cost of equity capital: Evidence from
China. Contemporary Accounting Research, 28(3), 892–925.
Chen, C. J. P., Chen, S. M., & Su, X. J. (2001). Is accounting information
value-relevant in the emerging Chinese stock market? Journal of
International Accounting, Auditing and Taxation, 10(1), 1–22.
Chen, Z. A., Du, J. M., Li, D. H., & Ou-Yang, R. (2013). Does foreign
institutional ownership increase return volatility? Evidence from
China. Journal of Banking and Finance, 37, 660–669.
Chen, D. H., Li, O. Z., Xin, F., & Yeung, B. (2012). Five-year plans and China
?nance. Working paper.
Chen, X., Lee, C. W. J., & Li, J. (2008). Government assisted earnings
management in China. Journal of Accounting and Public Policy, 27,
262–274.
Chen, X., & Li, J. (2001). Exploring the role of local government ?scal
supports in accounting performance of the listed companies. Journal
of Accounting Research, 12, 20–28 (in Chinese).
Chen, S. M., & Wang, Y. T. (2004). Evidence from China on the value
relevance of operating income vs. below-the-line items. The
International Journal of Accounting, 39, 339–364.
Cheng, P., Aerts, W., & Jorissen, A. (2010). Earnings management,
asset restructuring, and the threat of exchange delisting in an
earnings-based regulatory regime. Corporate Governance: An
International Review, 18(5), 438–456.
Chinese Accounting Standards Committee (2006). Accounting standards
for enterprises No. 16: Government subsidies.
Claro, S. (2006). Supporting inef?cient ?rms with capital subsidies: China
and Germany in the 1990s. Journal of Comparative Economics, 34(2),
377–401.
Conrad, K. (1993). Taxes and subsidies for pollution-intensive industries
as trade policy. Journal of Environmental Economics and Management,
25(2), 121–135.
Conrad, K., & Wang, J. M. (1993). The effect of emission taxes and
abatement subsidies on market structure. International Journal of
Industrial Organization, 11(4), 499–518.
Davidson, C., & Segerstrom, P. (1998). R&D subsidies and economic
growth. RAND Journal of Economics, 29(3), 548–577.
De Mel, S., McKenzie, D., & Woodruff, C. (2010). Wage subsidies for
microenterprises. American Economic Review, 100(2), 614–618.
Dechow, P. M., Hutton, A. P., & Sloan, R. G. (1999). An empirical
assessment of the residual income valuation model. Journal of
Accounting and Economics, 26, 1–34.
Desai, M. A., & Hines, J. R. (2008). Market reactions to export subsidies.
Journal of International Economics, 74(2), 459–474.
Dewatripont, M., & Roland, G. (1995). The design of reform
packages under uncertainty. American Economic Review, 85(5),
1207–1223.
Enikolopov, R., & Zhuravskaya, E. (2007). Decentralisation and political
institutions. Journal of Public Economics, 91, 2261–2290.
Ezzamel, M., Xiao, Z., & Pan, A. (2007). Political ideology and accounting
regulation in China. Accounting, Organizations and Society, 32(7–8),
669–700.
Firth, M., Fung, P. M. Y., & Rui, O. M. (2006). Corporate performance and
CEO compensation in China. Journal of Corporate Finance, 12(4),
693–714.
Frye, T., & Shleifer, A. (1997). The invisible hand and grabbing hand.
American Economic Review, 87(2), 354–358.
Girma, S., Gong, Y. D., Gorg, H., & Yu, Z. H. (2009). Can production
subsidies explain China’s export performance? Evidence from ?rm-
Level data. Scandinavian Journal of Economics, 111(4), 863–891.
González, X., Jaumandreu, J., & Pazo, C. (2005). Barriers to innovation
and subsidy effectiveness. RAND Journal of Economics, 36(4),
930–950.
Gorg, H., & Strobl, E. (2007). The effect of R&D subsidies on private R&D.
Economica, 74(294), 215–234.
Gu, Z., Li, Z., & Yang, Y. G. (2012). Monitors or predators: The in?uence of
institutional investors on sell-side analysts. The Accounting Review,
88(1), 137–169.
Hanson, G. H., & Robertson, R. (2008). China and the manufacturing
exports of other developing countries. NBER working paper.
Haw, I. M., Qi, D. Q., & Wu, W. (1999). Value relevance of earnings in an
emerging capital market: The case of A-shares in China. Paci?c
Economic Review, 4(3), 337–348.
Hawksworth, J., & Tiwari, A. (2011). The world in 2050 – The accelerating
shift of global economic power: Challenges and opportunities.
.
Hellman, J., & Shankerman, M. (2000). Intervention, corruption and
capture: The nexus between enterprises and the state. Economics of
Transition, 8(3), 545–576.
Hillegeist, S. A., Keating, E. K., Cram, D. P., & Lundstend, K. G. (2004).
Assessing pro?tability of bankruptcy. Review of Accounting Studies, 9,
5–34.
Hung, M. Y., Wong, T. J., & Zhang, T. Y. (2008). Political relations and
overseas stock exchange listing: Evidence from Chinese state-owned
enterprises. Working paper.
Irwin, D. A., & Klenow, P. J. (1996). High-Tech R&D subsidies: Estimating
the effects of Sematech. Journal of International Economics, 40(3–4),
323–344.
Ishikawa, J., & Spencer, B. J. (1999). Rent-shifting export subsidies with an
imported intermediate product. Journal of International Economics,
48(2), 199–232.
Jin, H., Qian, Y., & Weingast, B. (2005). Regional decentralisation and ?scal
incentives: Federalism Chinese style. Journal of Public Economics,
89(9–10), 1719–1742.
Jing, L. (2009). Governance and ?nancial reform in China’s transition
economy. Hong Kong: Hong Kong University Press.
Keen, M., & Marchand, M. (1997). Fiscal competition and the pattern of
public spending. Journal of Public Economics, 66, 33–53.
Knight, G. A., & Cavusgil, S. T. (2004). Innovation, organizational
capabilities, and the Born-Global ?rm. Journal of International
Business Studies, 35(2), 124–141.
Kohn, R. E. (1997). The effect of emission taxes and abatement subsidies
on market structure: Comment. International Journal of Industrial
Organization, 15(5), 617–628.
Landsman, W., Miller, B., & Yeh, S. (2007). Implication of components of
income exclude from pro forma earnings for future pro?tability and
equity valuation. Journal of Business Finance and Accounting, 34,
650–675.
Larsen, B., & Shah, A. (1992). World fossil fuel subsidies and global carbon
emissions. Working paper.
Lau, L., Qian, Y., & Roland, G. (2000). Reform without losers: An
interpretation of China’s dual-track approach to transition. Journal
of Political Economy, 108(1), 120–143.
Lee, C. W. J. (2001). Financial restructuring of state owned enterprises in
China: The case of Shanghai Sunve Pharmaceutical Corporation.
Accounting, Organizations and Society, 26, 673–689.
Lee, E., Walker, W., & Zeng, C. (2013). Does IFRS convergence affect
?nancial reporting quality in China? ACCA Research Report.
Li, W. L. (2002). Entrepreneurship and government subsidies: A general
equilibrium analysis. Journal of Economic Dynamics and Control,
26(11), 1815–1844.
Lind, D. A., Marchal, W. G., & Mason, R. D. (2002). Statistical techniques in
business and economics. Irwin: MaGraw-Hill.
Lipton, D., & Sachs, J. (1990). Creating a market economy in Eastern
Europe: The case of Poland. Brookings Papers on Economic Activity, 1,
75–133.
Litwack, J., & Qian, Y. (1998). Balanced or unbalanced development:
Special economic zones as catalysts for transition. Journal of
Comparative Economics, 26(1), 117–141.
Long, D., Payne, J., & Feng, C. (1999). Information transmission in the
Shanghai equity market. Journal of Financial Research, 22, 29–45.
Lopez, R., & Galinato, G. I. (2007). Should governments stop subsidies to
private goods? Evidence from rural Latin America. Journal of Public
Economics, 91(5–6), 1071–1094.
Neary, J. P. (1994). Cost asymmetries in international subsidy games:
Should governments help winners or losers? Journal of International
Economics, 37(3–4), 197–218.
Oates, W. (1972). Fiscal federalism. New York: Harcourt Brace Jovanovich.
O’Connor, N. G., Chow, C. W., & Wu, A. (2004). The adoption of ‘‘Western’’
management accounting/controls in China’s state-owned enterprises
during economic transition. Accounting, Organizations and Society,
29(3), 349–375.
O’Connor, N. G., Deng, J., & Luo, Y. D. (2006). Political constraints,
organization design and performance measurement in China’s state-
owned enterprises. Accounting, Organizations and Society, 31,
157–177.
O’Connor, N. G., Vera-Muñoz, S. C., & Chan, F. (2011). Competitive forces
and the importance of management control systems in emerging-
168 E. Lee et al. / Accounting, Organizations and Society 39 (2014) 149–169
economy ?rms: The moderating effect of international market
orientation. Accounting, Organizations and Society, 36(4), 246–266.
Ohlson, J. A. (1995). Earnings, equity book values, and dividends in equity
valuation. Contemporary Accounting Research, 11, 661–688.
Ohlson, J. A. (1999). On transitory earnings. Review of Accounting Studies,
3, 145–162.
Phelps, E. S. (1994). Low-wage employment subsidies versus the welfare
state. American Economic Review, 84(2), 54–58.
Qian, Y., & Roland, G. (1998). Federalism and the soft budget constraint.
American Economic Review, 88(5), 1143–1162.
Roland, G. (2002). The political economy of transition. Journal of Economic
Perspective, 16(1), 29–50.
Sami, H., & Zhou, H. Y. (2004). A comparison of value relevance of
accounting information in different segments of the Chinese stock
market. The International Journal of Accounting, 39(4), 403–427.
Schleicher, T., Hussainey, K., & Walker, M. (2007). Loss ?rms’ annual
report narratives and share price anticipation of earnings. The British
Accounting Review, 39, 153–171.
Schwartz, G., & Clements, B. (1999). Government subsidies. Journal of
Economic Survey, 13(2), 119–147.
Shen, H. B. (2007). Market segmentation, cross listing and the expected
capital cost: Evidence using Ohlson-Juettner model. Journal of
Financial Research, 2, 146–155 (in Chinese).
Shivakumar, R. (1993). Strategic trade policy: Choosing between export
subsidies and export quotas under uncertainty. Journal of International
Economics, 35(1–2), 169–183.
Shleifer, A., & Treisman, D. (2000). Without a map: Political tactics and
economic reform in Russia. Cambridge, Mass.: MIT Press.
Shleifer, A., & Vishny, R. (1993). Corruption. Quarterly Journal of Economics,
108(3), 599–671.
Shleifer, A., & Vishny, R. W. (1998). The grabbing hand: Government
pathologies and their cures. Harvard University Press.
Snower, D. J. (1994). Converting unemployment bene?ts into
employment subsidies. American Economic Review, 84(2), 65–70.
Sudman, S., & Bradburn, N. M. (1983). Asking questions: A practical guide to
questionnaire design. San Francisco, CA: Jossey-Bass.
Summers, S. T., & Sweeney, J. T. (1998). Fraudulently misstated ?nancial
statements and insider trading: An empirical analysis. The Accounting
Review, 73(1), 131–146.
Tang, Q. Q., & Luo, D. L. (2007). Empirical study on the motivation and the
effect of government subsidies: Evidence from Chinese listed
companies. Journal of Financial Research, 6, 149–163 (in Chinese).
Thomas, G. R. (1994). Chinese industrial reform: Accomplishment,
prospects, and implications. American Economic Review, 84(2),
271–275.
To, T. (1994). Export subsidies and Oligopoly with switching costs. Journal
of International Economics, 37(1–2), 97–110.
Walder, A. (1995). China’s transitional economy: Interpreting its
signi?cance. China Quarterly, 144, 963–979.
Wei, S. (1997). Gradualism versus big bang: Speed and sustainability of
reforms. Canadian Journal of Economics, 30(4), 1234–1247.
Zhu, S., & Chen, Y. S. (2009). Government subsidy decisions, earnings
management incentives, and loss reversals of listed ?rms. China
Accounting and Finance Review, 11(3), 115–140.
Zhuravskaya, E. (2000). Incentives to provide local public goods: Fiscal
federalism Russian style. Journal of Public Economics, 76(3), 337–368.
E. Lee et al. / Accounting, Organizations and Society 39 (2014) 149–169 169

doc_670570589.pdf