International Journal of Accounting & Information Management

Ownership Structure and Tax Aggressiveness of Chinese Listed Companies

Tingting Ying, Brian Wright, Wei Huang,
Article information:
To cite this document:
Tingting Ying, Brian Wright, Wei Huang, "Ownership Structure and Tax Aggressiveness of Chinese Listed Companies",
International Journal of Accounting & Information Management, https://doi.org/10.1108/IJAIM-07-2016-0070
Permanent link to this document:
https://doi.org/10.1108/IJAIM-07-2016-0070
Downloaded on: 27 June 2017, At: 23:32 (PT)
References: this document contains references to 0 other documents.
To copy this document: [email protected]
The fulltext of this document has been downloaded 3 times since 2017*
Access to this document was granted through an Emerald subscription provided by emerald-srm:333301 []
For Authors
If you would like to write for this, or any other Emerald publication, then please use our Emerald for Authors service
Downloaded by Cornell University Library At 23:32 27 June 2017 (PT)

information about how to choose which publication to write for and submission guidelines are available for all. Please
visit www.emeraldinsight.com/authors for more information.
About Emerald www.emeraldinsight.com
Emerald is a global publisher linking research and practice to the benefit of society. The company manages a portfolio of
more than 290 journals and over 2,350 books and book series volumes, as well as providing an extensive range of online
products and additional customer resources and services.
Emerald is both COUNTER 4 and TRANSFER compliant. The organization is a partner of the Committee on Publication
Ethics (COPE) and also works with Portico and the LOCKSS initiative for digital archive preservation.

*Related content and download information correct at time of download.

Ownership Structure and Tax Aggressiveness of Chinese Listed Companies

Abstract

This paper investigates the influence of state shareholding and control versus institutional
investors on tax aggressiveness of Chinese listed firms. By exploring recently available tax
reconciliation data required under 2006 Accounting Standards for Business Enterprises on a
sample of Chinese A-share listed firms, we calculate a direct measure of tax aggressiveness
and find that state ownership and control are positively associated with corporate tax
Downloaded by Cornell University Library At 23:32 27 June 2017 (PT)

aggressiveness. A positive link between the collective shareholding by the top ten
shareholders and firm tax aggressiveness is also found. In contrast, institutional share
ownership is negatively associated with corporate tax aggressiveness. Our results indicate
that political connections and ownership concentration empower firms to pursue aggressive
tax planning whereas institutional investors partially mitigate such influences.

JEL classification: G3; M4; H2

Key words: Ownership Structure; Book-Tax Differences; Tax Aggressiveness; China

1
1. Introduction
Prior research has documented significant cross-sectional differences in tax aggression
among firms domiciled in the U.S (Dyreng et al. 2008; Hanlon and Heitzman 2010) and
suggests that book-tax differences (BTDs) generated by the different reporting rules for book
and tax purposes are related to earnings manipulation and tax sheltering activities (e.g. Mills
and Newberry, 2001; Phillips et al., 2003; Hanlon, 2005; Frank et al, 2009; Wilson, 2009;
Muhmud et al., 2011; Tang and Firth, 2011; Lee and Chio, 2016). This strand of literature
generally posits that the strength of corporate governance should be negatively related to tax
sheltering activity, and/or should mitigate financial reporting aggressiveness (Xu et al., 2015;
Eraheem, 2016). This view of tax planning behaviour implies that tax aggressiveness can be
Downloaded by Cornell University Library At 23:32 27 June 2017 (PT)

viewed on balance as undesirable management behavior, because of the concomitant lack of

financial transparency that it may induce. Moreover, tax aggressiveness may facilitate
managerial resource diversions (Desai and Dharmapala, 2006, 2009), rent seeking and bad
news hoarding activities (Kim et al., 2011) in the absence of strong corporate governance.

However, it is not clear a priori that the relationship between corporate governance
mechanisms and tax aggressiveness should be so unambiguously negative. This is due to the
fact that tax sheltering activities themselves provide earnings benefits and real cash flow
advantages to the firm, but may also incur associated costs that may outweigh these benefits
if tax avoidance activities are too aggressive, or if they arise as a result of managerial
incentives relating to the creation and exploitation of information asymmetries to the
detriment of shareholders. These potential costs include increased information asymmetry
due to managerial opacity, and the resulting agency costs, as well as the potential for
increased tax compliance costs due to an increased likelihood of tax audits, revised tax
assessments and litigation. It is therefore possible that the net effect of more effective
corporate governance mechanisms on tax reporting aggressiveness could be positive or
negative, depending on the relative preference for tax sheltering of firm managers and
shareholders (Moore, 2012). Firms with strong corporate governance structures should be
able to minimize their agency problems with respect to tax position and achieve the optimal
level of tax aggressiveness 1 by alignment of the interests of managers with those of

1
A firm’s level of tax aggressiveness can be jointly determined by corporate governance structure, managerial
discretion and the changes in tax regime. The optimal level of tax aggressiveness can be viewed as the profit-
maximization level of tax aggressiveness which balances the benefits and costs associated with tax aggressive
2
shareholders. Firms with weak corporate governance could unwittingly provide managers
with opportunities to take advantage of uncertainty within the tax system, and their
informational advantage, to engage in tax aggressiveness that provides them with personal
gain at the expense of shareholders' wealth. This point of view has been evident in recent
studies (Desai and Dharmapala, 2006, 2009; Moore, 2012), and is particularly relevant in the
Chinese context, given the relative weakness of financial reporting oversight.

The significance of political connections as a determining factor of tax aggressiveness has

attracted growing research interest (Adhikari et al, 2006; Zeng, 2010; Wu et al., 2012b; Wu et
al., 2013; Chan et al., 2013), due to its perceived importance as an influence on firms'
Downloaded by Cornell University Library At 23:32 27 June 2017 (PT)

behavior. Recent studies have investigated the effects of the separation of ownership and
control on tax aggressiveness, which provides managerial incentives to pursue self-serving
behavior and potentially increases agency costs. It has been shown, for example, that family
ownership (Chen et al. 2010; Ebraheem, 2016), dual-class stock ownership (McGuire et al.
2012), and public versus private ownership (Mills and Newberry, 2001) are associated with
corporate tax aggressiveness. However, only a very limited number of studies have examined
the effects of political connections on the tax benefits received by firms with different
ownership structure (e.g. Adhikari et al. 2006; Wu et al. 2012b). This paper aims to fill this
research gap by examining the tax sheltering activities of Chinese listed firms.

The emerging Chinese stock market is a particularly suitable environment for conducting this
research. A notable difference between China and developed Western economies is that
business relationships in the former tend to be characterized as 'relationship-based' rather than
'market-based'' (Adhikari et al. 2006). Prior studies show that politically connected firms
receive preferential treatment from the government, including bank loans and favorable tax
treatments (Adhikari et al. 2006; Wu et al. 2012b). It is suggested by Faccio (2006) that the
benefits associated with political associations are more pronounced in countries with highly
interventionist governments and weaker property rights protection compared to those of other

positions by the interest alignments of managers and shareholders and induce managers to take tax positions to
enhance wealth of shareholders. Shareholders can implement incentives and controls through corporate
governance mechanisms and should be able to minimize their agency conflicts related to the tax aggressive
transactions in order to induce managers to achieve firms’ optimal level of tax aggressiveness given prevailing
tax environment. See Desai and Dharmapala (2006, 2009); Moore (2012) for details.

3
countries. Shleifer and Vishny (1997) and La Porta et al. (2002) emphasize that the
expropriation of minority investors by shareholders with a controlling interest is the primary
agency conflict in firms outside of the U.S. and U.K. Corporate governance and principal-
agent relationships in Chinese firms are often different from those in developed markets
which may lead to managerial entrenchment and tunneling by controlling shareholders (Jiang
et al. 2010; Liu and Tian, 2012). In particular, the majority of Chinese listed firms are
privatized and partially-privatized SOEs that sold their shares in the domestic stock markets
as a result of the government’s privatization strategy over recent decades (Jiang et al. 2010;
Huang and Wright, 2015; Guo et al. 2016). The state has maintained control in many of the
listed firms, and the ownership structure of the listed firms exerts strong influence on the
Downloaded by Cornell University Library At 23:32 27 June 2017 (PT)

nature of firm leadership and leads to principal-principal conflicts in addition to principal-

agent conflicts, with politicians and managers with political background being appointed as
directors, supervisors, and executives in many listed firms. State ownership represents a
strong form of political connection and a more direct tie with the government than having
connected managers (Wu et al. 2012b).

Therefore, China's institutional environment provides a unique research setting conducive to

observing the effects of political connections intertwining with market-oriented forces in the
adoption of tax reporting practices during the economic transition period. In the aftermath of
various corporate scandals and the credit crisis in the Western world, there is now tangible
discomfit at the current status, and even underlying rationale, of corporate governance. It is
thus timely to consider the interplay between corporate governance attributes and tax
reporting aggressiveness in the context of a country engaged in active reform and
marketization of its corporate practices. The contribution of this study is twofold. Firstly, we
extend the tax-effect BTDs proposed by Tang and Firth (2011) by taking advantage of
information available in the notes to tax reconciliations to construct a new, refined measure
of corporate tax aggressiveness. By applying recently available tax reconciliation data
required under Accounting standard 12 Income Taxes (ASBE, 2006) on a sample of Chinese
A-share listed firms, we calculate the differences between income for financial reporting
purposes and income for tax reporting purposes and non-conforming tax planning activities.
Secondly, This paper complements recent studies on tax aggressiveness in the U.S. (e.g.
Armstrong et al. 2012; McGuire, et al. 2012) by analysing tax planning activities of Chinese
listed firms on which there is very limited evidence in the prior literature (e.g. Zeng, 2010;

4
Tang and Firth, 2011,Wu et al. 2012a.b; Wu et al. 2013; Chan et al. 2013). We highlight firm
ownership and control factors that encourage aggressive tax planning in China. Our study has
important implications for both public policy and corporate governance in emerging markets
similar to China.

We find that state ownership and control are positively associated with corporate tax
aggressiveness of Chinese listed firms. A positive link between the collective shareholding by
the top ten shareholders and firm tax aggressiveness is also found. In contrast, institutional
share ownership is negatively associated with corporate tax aggressiveness. Our results
indicate that political connections and ownership concentration empower firms to pursue
Downloaded by Cornell University Library At 23:32 27 June 2017 (PT)

aggressive tax planning whereas the presence of institutional investors partially mitigates
such influences.

The remainder of the paper is structured as follows. Section 2 discusses the institutional
background and develops our hypotheses. Section 3 outlines our tax aggressiveness measure,
sample data, and our models. Section 4 discusses our results followed by a conclusion in
section 5.

2. Institutional background and hypothesis development

2.1 State ownership and tax aggressiveness
A special feature of the corporate ownership structure in China is the existence of non-
tradable shares including shares owned by the state to retain control over listed firms, which
are classified as state shares and legal person shares (Jiang et al. 2010; Huang and Wright,
2015; Xu et al., 2015; Guo et al. 2016). The non-tradable shares give entitlement to exactly
the same cash flow and voting rights but are not publicly tradable. Despite the non-tradable
share reforms over the last decade, the majority of Chinese listed firms are still closely linked
to the government. The largest and the controlling shareholder for a Chinese listed firm is
most likely a large SOE or central/city government, which can significantly influence
managerial decisions through its shareholding and political power (Liu and Lu, 2007; Lo et al.
2010;).

Empirical evidence suggests state ownership is not an effective governance mechanism in

China due to inefficient monitoring and operation, low managerial incentives and higher

5
agency costs as a result of higher information asymmetry (Huang and Wright, 2015). Several
studies have examined the influence of state shareholding on tax strategies, and these studies
offer divergent conclusions. On the one hand, it is argued that managers of government-
controlled firms may have different tax objectives compared with their non-government
controlled counterparts (Zeng, 2010; Wu et al., 2012b; Chan et al., 2013). As managers in
government-controlled firms are appointed and evaluated by the government, state ownership
provides them with incentives to pursue social and political objectives that sideline the
objective of maximization of after-tax profits. Zeng (2010) finds that management in firms
with higher state shareholding are less aggressive in tax reporting in order to obtain a good
reputation through remittance of higher tax payments, presumably to facilitate promotion or
Downloaded by Cornell University Library At 23:32 27 June 2017 (PT)

smooth the path to a promising political career. Wu et al. (2007) and Wu et al. (2012a) do not
find a significant relationship between state shareholding and tax burdens; however, the
effect of firm size on effective tax rate is mediated by the nature of the controlling
shareholder. In contrast, some studies document that managers of government-controlled
firms are more likely to maximize corporate resources under their control through aggressive
tax planning strategies by taking advantage of their political connections (Faccio, 2006;
Zimmerman, 1983), which may in turn increase managerial compensation through the impact
on firm value (Wang et al., 2008). For example, Adhikari et al. (2006) find that firms in
Malaysia with higher state shareholding pay less tax due to benefits accruing from their
political connections. Kim and Zhang (2015) also show that political connections are
positively associated with aggressive tax planning, due to lower cost of tax aggressiveness
such as through reduction of the probability of tax behavior being investigated by the tax
administration, and a resulting lesser need for financial transparency via political connections
(Yu and Yu, 2011). Therefore, it is possible that the cost of tax planning is lower and the
benefit is higher for politically connected firms than for their counterparts. These arguments
lead to our first hypothesis:

H1: Tax aggressiveness of Chinese listed firms is positively associated with state ownership
and control.

2.2 Ownership concentration and tax aggressiveness

Large controlling shareholders can typically exert influence over major corporate decisions,
including tax strategies. There are at least two competing effects of ownership concentration

6
on corporate governance (Claessens et al. 2002). First, an incentive alignment effect.
Concentrated ownership can serve to focus managerial activity to the benefit of shareholders,
with a resulting alignment of the interests of controlling and minority shareholders in
countries with a less developed legal and institutional environment (Lins, 2003). Secondly, an
entrenchment effect. The controlling shareholders are provided with incentives to transfer
cash flows out of the firm at the expense of minority shareholders (known as ‘tunneling’)
(Claessens et al. 2002), or to pursue objectives inconsistent with value maximization. In view
of the unique institutional environment in China, the widespread concentration of state
ownership in Chinese listed firms is more likely to induce an entrenchment effect (Jiang et al.
2010; Liu and Tian, 2012) at the expense of incentive alignment, which in turn may lead to
Downloaded by Cornell University Library At 23:32 27 June 2017 (PT)

weak corporate governance and low transparency. This effect may be evidenced in part by a
firm's level of tax aggression, with lowered tax aggressiveness being symptomatic of
entrenchment and alteration of the corporate objective function.

Prior literature provides mixed results on the association between ownership concentration
and the level of tax aggressiveness. On the one hand, firms in countries with higher
ownership concentration are provided with incentives for tax aggressiveness as they have
lower non-tax costs and may be more tax aggressive because large shareholders can
effectively monitor and incentivize managers to generate more tax savings (Desai and
Dharmapala, 2009; Zeng, 2010;). On the other hand, firms in countries with higher ownership
concentration may be less tax aggressiveness due to the implementation costs and agency
costs involved (Chen et al. 2010). In particular, large shareholders have more incentive to
monitor managers' behaviors, including their tax-saving activities, due to the significant costs
associated with risky tax planning activities and informational opacity, and to ensure that
managers behave in ways that benefit shareholders (Badertscher et al. 2013). Although the
prior literature offers inconclusive evidence, in line with our first hypothesis, we further
hypothesize that:

H2: Tax aggressiveness is positively associated with the level of ownership concentration of
Chinese listed firms.

2.3 Institutional ownership and tax aggressiveness

7
It has recently been argued that institutional shareholding has a beneficial effect on overall
corporate governance practices, due to the active role of monitoring and disciplining
managerial opportunism as well as improvement of information efficiency through
information provided to the capital market (Bushee, 2001; Wei et al., 2005). The presence of
mutual funds, in particular, in emerging markets may strengthen the bargaining power of
minority shareholders and provide beneficial monitoring of a firm’s large and dominant
shareholders as well as managers and directors in the corporate decision process. More
recently, Huang and Boateng (2016) document that institutional investors have significant
impact on the value of Chinese firms. Prior empirical studies generally find that tax
aggressiveness is related to the representation of active institutional shareholders, although
Downloaded by Cornell University Library At 23:32 27 June 2017 (PT)

the results are mixed (Koh, 2003). Moore (2012) analyzes the impact of institutional
ownership on the level and time-series variability in book-tax differences among US firms
and shows that institutional ownership is negatively associated with book-tax differences. In
contrast, Khurana and Moser (2013) find that U.S. firms with higher institutional ownership
which also have higher ownership concentration are generally more tax aggressive. However,
this finding is driven by firms with higher levels of short-term institutional investment, which
is claimed to typically influence firms to promote short-term market value as opposed to
long-term profitability, making them more likely to push managers to invest in projects with
greater expected near-term earnings (Bushee, 2001), whereas firms with a higher presence of
institutional investors with longer-term investment horizon are more concerned about long-
term consequences of tax aggressiveness owing to a long lag between the
design/implementation of tax transactions and detection by the IRS (Khurana and Moser,
2013) and thus are less aggressive . It is, moreover, unclear whether findings from the US
market are applicable to an emerging market such as China. We therefore test our third
hypothesis:

H3: Tax aggressiveness is negatively associated with institutional ownership of Chinese

listed firms.

3. Data and methods

We measure the extent to which a firm is able to avoid paying tax on its accounting income
by utilizing the Book-Tax Differences (BTDs) construct that has been operationalized in the
recent literature. Studies such as Desai and Dharmapala (2006, 2009) and Frank et al. (2009)

8
on US firms and Tang and Firth (2011) on Chinese B-share listed firms have attempted to use
a ‘residual’ approach to decompose BTDs into 'normal' BTDs (mechanical differences
between tax rules and financial accounting standards) and 'abnormal' BTDs (residual from
total BTDs resulting from opportunistic managerial choices in accounting and tax treatments).
We follow an approach similar to Tang and Firth (2011) by applying recently available tax
reconciliation data required under 2006 Accounting Standards for Business Enterprises
(ASBE 2006) on a sample of Chinese A-share listed firms. ASBE 2006 Article 18 Income
Taxes provides guidelines for various disclosures in relation to the tax expense in “notes to
financial statements”. Under ASBE 2006, book-tax differences will arise principally as a
result of the following common categories: (1) Income not taxable; (2) Non-deductible
Downloaded by Cornell University Library At 23:32 27 June 2017 (PT)

expenses for tax purposes; (3) The effects arising from differences in effective tax rate of
subsidiaries, particularly those operating in foreign jurisdictions; (4) Prior year adjustments to
tax payable. The reported values of these BTD categories are manually collected from the
'additional notes to financial statements' in the firm's annual report for all firms that disclose
the relevant information for the financial years 2006-20122. We then calculate the sum of
these categories as the total book-tax difference BTD value. Appendix A shows the
categories, drivers, and proxy variables for BTD under ASBE 2006.

(Insert Appendix A here)

Our measure of tax aggressiveness is therefore a direct measure based on the detailed tax
expenses disclosures, in contrast to more common proxy variables for tax avoidance such as
the effective tax rate.3 We then, we follow a residual approach similar to Tang and Firth

2
The hypothesized drivers for these BTDs take into account of the tax adjustment items on tax forms of Chinese
income tax laws and accounting standards, which are shown in tax reconciliation disclosed in annual financial
reports. For example, in terms of expenses not deductible, it is hypothesized that the normal level of non-
deductible expenses will differ by industries but should be constant as proportion of total expenses within an
industry. Therefore, it is hypothesized that industry membership and operating expenses will be the main drivers
of non-deductible expenses, and operating expenses is included to control for different expenses recognition in
accounting rules and tax laws.

3
Chinese listed firms are subject to very different statutory tax rates and often receive different local
government tax rebate. Effective tax rate therefore is a noisy measure of tax avoidance.

9
(2011) to calculate the abnormal component of BTD as our measure of tax aggressiveness.
We adopt a fixed-effects regression model as follows.

BTD୧୲ = α୧ + βଵ INVINC୧୲ + βଶ INTINC୧୲ + βଷ OPEXP୧୲ + βସ FSALE୧୲ + βହ LOG(ASSETS୧୲ )

+ β଺ OPBIT୧୲ + β଻ PBT୧୲ + β଼ PBT୧୲ିଵ + βଽ PBT୧୲ିଶ + τ୲ + ε୧୲
--- Equation (1)
In Equation (1), the dependent variable is total BTD scaled by the percentage of prior year
end total assets, and the independent variables on the right-hand side are a set of variables
considered as drivers for the aforementioned BTD categories under the 2006 ASBE as
follows: current period investment income INVINC୧୲ , interest income INTINC୧୲ , operating
Downloaded by Cornell University Library At 23:32 27 June 2017 (PT)

expenses OPEXP୧୲ , the percentage of overseas sales FSALE୧୲ , the log of total assets
LOG(ASSETS)୧୲ , operating profit before interest and tax OPBIT୧୲ , net profit before tax PBT୧୲ ,
and the net profit before tax in previous periods PBT୧୲ିଵ and PBT୧୲ିଶ . We use fixed firm
effects α୧ and year effects τ୲ to control for unobservable influences on BTDs including
industry membership, geographical tax policy differences, and time variations in tax
regulations and enforcement. We scale all continuous variables in the model except
LOG(ASSETS)୧୲ and FSALE୧୲ by prior year-end total assets and winsorize them at 1% and 99%
to run the regression. The predicted BTD by equation (1) captures the BTD due to tax
reporting and financial reporting under the 2006 ASBE. We then use the regression residuals
ε୧୲ as our tax aggressiveness measure denoted as TAXAGG୧୲ , calculated as the actual total
BTD value minus the fitted BTD value from the Equation (1) regression to capture
opportunistic managerial choices in accounting and tax treatments.

Our sample includes Chinese firms (excludes firms in financial and insurance industries) that
were listed on the Shenzhen and Shanghai Stock Exchanges Markets during 2006-2012. we
retain firms for which we are able to compute the tax aggressiveness measure. To have
complete data for the measurements of tax aggressiveness, we delete firm-year observations
for missing data on the selected variables. All the data for the institutional variables unrelated
to the tax planning, control variables and corporate governance variables are matched with
the resulting tax aggressive firm observations, which were obtained from CSMAR database,
with WIND database as a supplement. Our efforts result in a final sample of 958 firm/years
with all necessary data for the analysis and among which around 50% are manufacturing
sector firms. The size of our sample compares favourably to 525 firm-year observations

10
adopted in Tang and Firth (2011), and thus provides a more up-to-date and robust assessment
of the tax incentives facing Chinese listed firms. Listed firms ownership, accounting, and
financial data are collected from the China Stock Market and Accounting Research (CSMAR)
database for this sample. Having isolated the tax aggressiveness variable TAXAGGit, we then
test an OLS model including industry dummies IND, and year dummies YEAR as in equation
(2) to examine the relationship between the various aspects of corporate ownership structure
and the level of tax aggressiveness.

TAXAGG୧୲ =α + β Ownership୧୲ + γ Controls୧୲ + φ YEAR ୲ + η IND୧ + ε୧୲

---Equation (2)
Downloaded by Cornell University Library At 23:32 27 June 2017 (PT)

The dependent variable TAXAGG୧୲ represents the residuals from the BTD model in Equation
(1), measuring firm tax aggressiveness. Ownership୧୲ represents a vector of ownership
variables including: the percentage of state shares (including state-owned legal person shares)
STA; a dummy variable GOV which equals to 1 if the firm is under control of the
government or a government agency and zero if it is under the control of a private investor;
the total percentage of top 10 shareholder shareholdings as ownership concentration measure
OC; the percentage of institutional investor shareholding INST and the percentage of mutual
fund shareholding FUND. 4 These ownership and control variables are tested in separate
regressions where appropriate corresponding to our hypotheses to avoid multicollinearity
problems induced by their correlations (Wei et al., 2005). We include a set of control
variables Controls୧୲ in this model as follows: LOSS to capture a firm's current loss-making
status, in order to examine whether loss-making firms have differential incentive to engage in
aggressive tax strategies (Tang and Firth 2011; Badertscher et al, 2013), being a dummy
variable set to 1 when a consolidated entity has a loss in the current year t and 0 otherwise. In
line with Wilson (2009), Frank et al. (2009), and Armstrong et al. (2012), we control for firm
financing risk and the extent of the tax shield of debt using the market value financial
leverage ratio LEV which equals the book value of debt divided by the total of market equity
capitalization and book value of debt. The log of market capitalization as a measure of firm
SIZE is added to capture differences in the scale or size of the firm and also as a proxy for the

4
Institutions include mutual funds, insurance companies, securities companies, wealth management products,
QFII, pension funds, financial companies, trust companies, and banks.

11
benefits of tax sheltering (Wilson, 2009; Armstrong et al., 2012; Tang and Firth, 2012;
Khurana and Moser, 2013). The return on equity ROE is added to control for a firm's
profitability. As tax aggressiveness can be indicative of both earning management and tax
management (Mills and Newberry, 2001; Phillips et al. 2003; Hanlon, 2005; Frank et al, 2009;
Tang and Firth, 2011), we also control for earnings management EM calculated as the value
of discretionary accruals which is the prediction error when regressing total accruals against
change in sales, fixed assets, and industry and year fixed effects. Capital intensity (CAPINT)
is added to control for the opportunities related to investments in fixed assets, which impacts
BTDs via the differential financial reporting and tax treatment of depreciation (Frank et al.
2009; Armstrong et al. 2012). In light of Minnick and Noga, (2010) and Lanis and
Downloaded by Cornell University Library At 23:32 27 June 2017 (PT)

Richardson (2011), we further control for the influences of board independence BODIND
and board size BODSIZE on tax aggressiveness in the model. Finally, we control for the
influence of auditors using the log of audit fees AUDFEE, auditor opinion AUDOP, and
“BIG 4”audit firm, following Wang et al. (2008) and DeFond et al. (1999).

Table 1 reports the descriptive statistics of our sample variables. As TAXAGG is the
prediction error from model (1), its mean equals zero. We note that the percentage of state
shares has a mean of 19% which is slightly higher than that of institutional shareholding
(17%) and that of mutual fund shareholding (13%). A high level of ownership concentration
is evident across the sample with a mean of 63% collective shareholdings by top 10
shareholders. Moreover, the majority (85%) of the sample firms are under government
control.

(Insert Table 1 here)

Pearson correlations among these variables are reported in Table 25. The majority of these
correlations are significantly different from zero at the 0.05 level or above. TAXAGG is
positively correlated with the state shareholding (STA) and the government control dummy

5
The Variance inflation factors (VIFs) are calculated when estimating the regression model to test for the issue
of multi-collinearity between the independent variables, the results shows that VIFs are relatively low and none
of the VIFs exceed 10 for any of our independent variables, therefore, the serious multi-collinearity problems is
unlikely to be present for our study, The results are not reported to conserve space.

12
(GOV), and negatively correlated with institutional shareholding (INST) and mutual funds
shareholding (FUND).

(Insert Table 2 here)

4. Analysis and results

In Table 3, we conduct several tests of tax aggressiveness TAXAGG sample means between
subsamples classified according to our ownership variables. We note that government
controlled firms have slightly higher ownership concentration and much higher state
shareholding compared to private investor controlled firms. The sample mean tests of
Downloaded by Cornell University Library At 23:32 27 June 2017 (PT)

TAXAGG by firm control type, high-low ownership concentration, and high-low institutional
ownership reveal that firms under government control or exhibiting high ownership
concentration are more tax aggressive than their counterparts. Firms with high mutual fund
shareholdings are less tax aggressive than those with low mutual fund shareholdings.

(Insert Table 3 here)

Table 4 reports our basic regression results on firm ownership and tax aggressiveness.
Models 1 and 2 test our hypothesis 1. Model 1 shows that the variable STA has a significant
positive coefficient suggesting a strong significant relationship between state shareholding
and the level of tax aggressiveness. This is consistent with the model 2 result when we
replace STA with the dummy variable for government control GOV which suggests that
TAXAGG is significantly higher for government-controlled firms than for private investor
controlled firms. Our results here are consistent with Adhikari et al. (2006) and Tang and
Firth (2011) and Wu et al. (2012b) and support the view that government association
empowers firms to pursue more aggressive tax avoidance strategies. This is also grounded in
the multi-faceted perspective of political connections in China’s 'relationship-based' economy
(Adhikari et al. 2006; Wu et al., 2012b). Government-associated firms have more incentives
to be tax aggressive than their counterparts as a result of with lower possibility of tax audits
and penalties being imposed for tax evasions (Faccio, 2006; Faccio, 2010; Wu et al. 2012a.b).
This contrasts with a competing characterization of state-dominated firms as entities whose
profit orientation may be somewhat subordinated to governmental objectives, for example the
generation and capture of tax revenues. Our results appear to differ from earlier results

13
reported by Ding et al. (2007) and Wu et al (2012a.b), which document a negative
relationship between state ownership and tax planning in China. Our results therefore bring
into question the proposition that preferential treatment due to firm-government associations
dictates tax revenue maximization for the purpose of the supporting political goals and social
objectives among China’s listed SOEs. From the perspective of marketization of Chinese
corporations, this finding is perhaps encouraging. Private investor controlled firms in
comparison are at a disadvantage in terms of tax enforcement treatment.

Model 3 of Table 4 tests hypothesis 2 and shows that ownership concentration (OC) is
significantly and positively associated with the level of tax aggressiveness, in line with
Downloaded by Cornell University Library At 23:32 27 June 2017 (PT)

earlier studies such as (Liu and Lu, 2007; Zeng, 2010). It does support the argument that the
entrenchment effect is expected to dominate the alignment effect in the Chinese context, and
indicates that large shareholders in firms with highly concentrated ownership have sufficient
incentives to exercise control and power over managerial process including their tax-saving
activities that they are concerned, thus increasing monitoring costs, which is an inefficient
corporate governance strategy in the Chinese market with weak legal protection of minority
shareholders. Meanwhile, it also suggest that large shareholders in Chinese listed firms have
ties with government through political connections, which can be utilized to help firms to
obtain government-related resources and supports such as tax-benefits (Zeng, 2010; Wu et al.,
2012). Finally, this finding supports the results on state shareholdings, and further suggests
that ownership concentration brings focus to managerial incentives to act in the interests of
shareholders.

Models 4 and 5 test our hypothesis 3. The results show that the coefficient on institutional
ownership (INST) is negative and highly significant. Consistent with Moore (2012) and
Khurana and Moser (2013), we find that firms with larger institutional ownership are less tax
aggressive. This appears to support the monitoring role of institutional investors on
managerial reporting behaviors, but contrasts with the findings for ownership concentration
and government ownership. Model 5 further focuses on mutual funds as the most important
institutional investors and shows a consistent positive sign but larger coefficient on FUND.

(Insert Table 4 here)

14
Regarding the control variable in our models, we note that board composition and auditors
appear to have rather strong influences on firm tax aggressiveness. Board size and audit fees
are positively and significantly related to firm tax aggressiveness. However, board
independence, audit opinion and “Big 4” auditors do not affect tax aggressiveness. The
coefficients on LEV, EM, and CAPINT are positive and significant. Firm size is negatively
associated with tax aggressiveness, suggesting that ‘political costs’ (Zimmerman, 1983) may
dominate increased institutional capacity for tax planning, as firm size increases..

In Table 5, we further include the interaction variables STAit*INSTit, GOVit*INSTit,

OCit*INSTit, STAit*FUNDit, GOVit*FUNDit, OCit*FUNDit where appropriate as an
Downloaded by Cornell University Library At 23:32 27 June 2017 (PT)

independent variables in our model to examine the relative strength of government

association versus institutional investors influence on firm tax aggressiveness. We note that,
with the presence of the interaction variables, the effects of variables STA, GOV, and OC
now appear statistically insignificant although the coefficients on these variables are still
positive. In contrast, the coefficients on INST and FUND are still negative and highly
significant. In addition to the regression coefficients reported, we also note that when
interactions are added into our models, the marginal effects of INST and FUND on the
dependent variable TAXAGG are conditional on the values of STA, GOV, and OC where
appropriate. 6 We therefore compute the mean of these marginal effects test if they are
significantly different from zero. Results, also reported in Table 5, confirm that the mean
marginal effects of INST and FUND are all negative. All in all, our results generally indicate
that the institutional investors exert effective constraint on firm tax aggressive strategies and
can mitigate the associated risks of a firm utilizing its political connections for aggressive tax
avoidance. This finding also supports the promotion of institutional investors as monitoring
agents for corporate governance and minority interest protection in China.

(Insert Table 5 here)

A caveat attaching to the preceding model results is the potential endogeneity problems
caused by reversed causality from firm tax aggressiveness to firm ownership due to investor
clienteles. Most prior studies use static models to test firm ownership structure and tax

6
For instance, the marginal effect of INST on TAXAGG in model (1) of Table 5 is -0.923+1.444*STA.

15
reporting practices (e.g. Chen et al. 2010; Zeng, 2010; Moore, 2012; Wu et al. 2012a.b;
Badertscher et al. 2013). One may expect their association to be lagged due to slower
adjustments of tax strategies responding to ownership changes. Although limitations of hand
collected sample prohibit us from using more complex dynamic models, we use the lagged
firm ownership variables to replace their contemporaneous terms used in Table 4 and 5 to
rerun the regressions in table 6. Our findings are unaffected across all models which suggest
that endogeneity is unlikely to be a severe problem in our empirical models.

(Insert Table 6 here)

Downloaded by Cornell University Library At 23:32 27 June 2017 (PT)

5. Conclusion
This paper seeks to understand the effect of ownership structure on firm tax sheltering
activities in the Chinese institutional environment. We utilize available information contained
in notes to financial reports of Chinese listed companies to construct a new measure of
corporate tax aggressiveness. Our results suggest that state ownership and control are
associated with more aggressive tax avoidance strategies among our sample firms. Ownership
concentration also appears to have a positive influence on tax aggressiveness. This is in line
with the fact that large shareholders of Chinese listed firms are often the government and its
agencies at central and local levels. Our results suggest political connections empower firms
to pursue aggressive tax management at the cost of public welfare funded by taxation. Private
firms appear to bear heavier tax burdens in the absence of government preferential treatment.
These findings offer some evidence that the Chinese regulatory authorities’ attempts at
further marketization of the state-controlled corporate sector are meeting with a certain level
of success, in the sense that such companies are increasingly free to pursue market-based
objectives. However, the finding that institutional investor presence leads to less tax
aggressiveness, presumably as a result of the monitoring role of such investors, further
suggests that state-dominated firms, rather than being subservient to their state constituencies,
are able to turn such connections to the advantage of their shareholders. This provides some
indication that further progress needs to be made in separating state-controlled companies
from their state connections, if the objective is to create a flourishing private market with a
level playing field.

16
This study contributes to existing literature that explores the relationship between tax and
corporate governance characteristics in both developed and emerging markets. Our findings
provide a better understanding of the issues concerning listed firms in China during its
transition from a centrally planned to a market economy in terms of separation of control and
tax aggressiveness and have important theoretical as well as policy implications. Transitional
economies are often characterized by weak legal environment and poor corporate governance
systems, the evidence suggests that political connections are a significant determinant of
corporate tax practices in Chinese listed firms relative to market forces. Therefore, the
function of market forces in Chinese listed firms is limited by political connections. This may
serve as a caution to regulators and policymakers.
Downloaded by Cornell University Library At 23:32 27 June 2017 (PT)

This paper also opens avenues for further study in the area of corporate governance and tax
planning. It calls for improvement of disclosure and transparency of tax reporting. Financial
accounting disclosures in China have not yet reached the international standards met in many
developed economies. Mandatory disclosure of tax-related notes in annual reports may also
allow future studies to draw more reliable inferences from a larger sample.

17
References

Adhikari A, Derashid C, and Zhang H. 2006. Political connections, and effective tax rates:
longitudinal evidence from Malaysia. Journal of Accounting and Public Policy 25, 574-595.

Armstrong CS, Blouin JL, and Larcker DF. 2012. The incentives for tax planning. Journal of
Accounting and Economics 53, 391-411.

Badertscher BA, Katz SP, and REGO SO. 2013. The Separation of Ownership and Control
and Corporate Tax Avoidance. Journal of Accounting and Economics, 56, 228-250.

Bushee B. 2001. Do institutional investors prefer near-term earnings over long-run value?
Contemporary Accounting Research 18, 207-246.

Chan KH, Mo PLL, and Zhou AY. 2013. Government ownership, corporate governance and
Downloaded by Cornell University Library At 23:32 27 June 2017 (PT)

tax aggressiveness: evidence from China. Accounting and Finance, 1-23.

Chen SP, Chen X, Cheng, Q. and Shevlin T. 2010. Are family firms more tax aggressive than
non-family firms? Journal of Financial Economics, 95, 41-61.

Claessens S, Djankov SJPH, and Lang LHP. 2002. Disentangling the incentive and
entrenchment effects of large shareholdings. Journal of Finance, 57, 2741-2771.

Defond M and Wong TJ. 1999. The impact of improved auditor independence on audit
market concentration in China. Journal of Accounting and Economics 28, 269–305.

Desai MA, and Dharmapala D. 2009. Corporate tax avoidance and firm value. Review of
Economics and Statistics, 91, 537-546.

Desai MA and Dharmapala D. 2006. Corporate tax avoidance and high-powered incentives.
Journal of Financial Economics, 79, 145-179.

Dyreng S, Hanlon M. and Maydew EL. 2008. Long-run corporate tax avoidance. The
Accounting Review 83, 61-82.

Ebraheem SSA, 2016, Ownership structure and earnings management: evidence from Jordan,
International Journal of Accounting & Information Management, 24 (2), 135-161

Faccio M. 2010. Differences between politically connected and non-connected firms: A cross
country analysis. Financial Management, 39, 905-928.

Frank MM, Lynch LJ, and Rego S. 2009. Tax reporting aggressiveness and its relation to
aggressive financial reporting. Accounting Review, 84, 1-47.

Guo L., Dai Y., and Lien D. 2016. The effects of China’s split-share reform on firms’ capital
structure choice. Applied Economics 48 (27), 2530-2549.

Hanlon M. 2005. The persistence and pricing of earnings, accruals, and cash flows when
firms have large book-tax differences. Accounting Review, 80, 137-166.

18
Hanlon M. and Heitzman S. 2010. A review of tax research. Journal of Accounting and
Economics, 50, 127-178.

Huang W. and Boateng, A. 2016. On the value relevance of analyst opinions and institutional
shareholdings in China. International Journal of Accounting & Information Management 24
(3), pp. -

Huang W. and Wright, B. 2015. Analyst earnings forecast under complex corporate
ownership in China. Journal of International Financial Markets, Institutions and Money 35,
69-84.

Hyun-Ah Lee , Won-Wook Choi , 2016, Allowance for uncollectible accounts as a tool for
earnings management: Evidence from South Korea, International Journal of Accounting &
Information Management, 24 (2), 162-184.
Downloaded by Cornell University Library At 23:32 27 June 2017 (PT)

Jiang G, Lee CMC, and Yue H. 2010. Tunneling through Inter-corporate Loans: The China
Experience. Journal of Financial Economics, 98, 1-20.

Khurana I. and Moser WJ. 2013. Institutional Shareholders' Investment Horizons and Tax
Avoidance. The Journal of the American Taxation Association, 35, 111-134.

Kim J, Li Y, and Zhang L. 2011. Corporate tax avoidance and stock price crash risk: Firm-
level analysis. Journal of Financial Economics 100, 639–662.

Kim C and Zhang L. 2015. Corporate Political Connections and Tax Aggressiveness.
Contemporary Accounting Research forthcoming doi:10.1111/1911-3846.12150.

Koh PS. 2003. On the association between institutional ownership and aggressive corporate
earnings management in Australia. British Accounting Review, 35, 105–128.

La Porta R, Lopez-De-Silanes F, Shleifer A, and Vishny R. 2002. Investor protection and

corporate valuation. Journal of Finance 57, 1147-1170.

Lanis R. and Richardson G. 2011. The effect of board of director composition on corporate
tax aggressiveness. Journal of Accounting and Public Policy, 30, 50-70.

Lins KV. 2003. Equity ownership and firm value in emerging markets. Journal of Financial
and Quantitative Analysis, 38, 159-184.

Liu Q. and Lu Z. 2007. Corporate governance and earning management in the Chinese listed
companies: a tunneling perspective. Journal of Corporate Finance, 13, 881-906.

Liu Q. and Tian G. 2012. Controlling shareholder, expropriations and firm's leverage
decision: Evidence from Chinese Non-tradable share reform. Journal of Corporate Finance,
18, 782-803.

Lo A, Wong R. and Firth M. 2010. Can corporate governance deter management from
manipulating earnings? evidence from related-party sales transactions in China. Journal of
Corporate Finance 16, 225-235.

19
Mahmud H, Santanu M, Zabihollah R, 2011, Incremental information content of option‐
related excess tax benefit under FASB Statement No. 123R: A research note, International
Journal of Accounting & Information Management, 19 (2), 146-168.

McGuire ST, Omer TC, and WANG D. 2012. Tax avoidance: does tax-specific industry
expertise make a difference? The Accounting Review, 87, 975-1003.

Mills LF, and Newberry KJ. 2001. The influence of tax and non-tax costs on book-tax
reporting differences: Public and private firms. Journal of the American Taxation Association
23, 1-19.

Minnick K. and Noga T. 2010. Do corporate governance characteristics influence tax

management? Journal of Corporate Finance, 16, 703-718.

Moore JA. 2012. Empirical evidence on the impact of external monitoring on book–tax
Downloaded by Cornell University Library At 23:32 27 June 2017 (PT)

differences. Advances in Accounting, 28, 254-269.

Phillips J, Pincus M, and Rego S. 2003. Earnings management: New evidence based on
deferred tax expense. Accounting Review, 78, 491-521.

Shleifer A. and Vishny RW. 1997. A survey of corporate governance. Journal of Finance, 52,
727-783.

Tang YH. and Firth M. 2011. Can book-tax difference capture earnings management and tax
management? empirical evidence from China. The International Journal of Accounting, 46,
175-204.

Tang YH. and Firth M. 2012. Earnings Persistence and Stock Market Reactions to the
Different Information in Book-Tax Differences: Evidence from China. The International
Journal of Accounting, 47, 369-397.

Wang Q, Wong TJ. and Xia L. 2008. State ownership, the institutional environment and
auditor choice: evidence from China. Journal of Accounting and Economics, 46, 112-134.

Wei Z, Xie F, and Zhang S. 2005. Ownership Structure and Firm Value in China’s Privatized
Firms: 1991-2001. Journal of Financial and Quantitative Analysis, 40, 87-108.

Wilson R. 2009 An examination of corporate tax shelter participants. Accounting Review, 84,
969-999.

Wu LS, Wang YP, Gills P. and Luo W. 2012a. State ownership, tax status and size effect of
effective tax rate in China. Accounting and Business Research, 42, 97-114.

Wu LS, Wang YP, Lin BX, LI C. and Chen, S. 2007. Local tax rebates, corporate tax burdens,
and firm migration: Evidence from China. Journal of Accounting and Public Policy 26, 555-
583.

Wu WF, Rui OM. and Wu CF. 2013. Institutional Environment, Ownership and Firm
Taxation. Economics of Transition, 21, 17-51.

20
Wu WF, Wu CF, Zhou CY, and Wu J. 2012b. Political connections, tax benefits and firm
performance: Evidence from China. Journal of Accounting and Public Policy, 31, 277-300.

Xu_D.J , Kamran A and Wei L, 2015, The impact of corporate governance and ownership
structure reforms on earnings quality in China, International Journal of Accounting &
Information Management, 23 (2), 169-198.

Yu F. and Yu X. 2011. Corporate lobbying and fraud detection. Journal of Financial and
Quantitative Analysis, 46, 1865-1891.

Zeng T. 2010. Ownership concentration, state ownership and effective tax rate: evidence
from China's listed firms. Accounting Perspective, 9, 271-289.

Zimmerman, J. L. 1983. Taxes and firm size. Journal of Accounting and Economics 5, 119-
149.
Downloaded by Cornell University Library At 23:32 27 June 2017 (PT)

21
Appendix A: Categories, drivers, and proxy variables of BTD

Category of Permanent BTD Hypothesized Drivers of Variables

Category Database

1 Income not taxable Investment income and CSMAR-P/L

Finance income account and
WIND database

2 Expenses not deductible Industry membership; CSMAR-P/L

operating expenses. account

3 The effect of the application of a different Profit before exceptional CSMAR-B/S

tax rate to income, either because it is items; foreign sales; total account and P/L
generated abroad or because it is subject to assets all by geographical account
a different domestic tax rate segment.

4 Prior year adjustments Prior two years’ lagged CSMAR-P/L

Downloaded by Cornell University Library At 23:32 27 June 2017 (PT)

pre-tax profit. account

5 Utilization of brought-forward tax losses Current period pre-tax CSMAR-P/L

(Recognition of previous unrecognized profit and two lags of pre- account
losses) tax profit.

6 Current period tax losses carried forward Current period pre-tax CSMAR-P/L
(Current period unrecognized losses) profit and two lags of pre- account
tax profit.

7 Taxation of capital gains and losses Accounting gains on sale CSMAR-Cash

of fixed assets flow statement

8 Other permanent differences Try a combination of the CSMAR and

previous drivers WIND database

Descriptive statistics for the BTD model

Variables Obs. Mean Std. Dev Min Max
BTD 1080 0.014 0.288 -1.568 7.800
OPEXP 1080 0.673 0.489 -0.014 3.543
OPBIT 1080 0.061 0.067 0.354 0.339
PBT 1080 0.059 0.069 -0.037 0.665
PBT_(it-1) 1052 0.051 0.072 -0.837 0.763
PBT_(it-2) 1024 0.087 1.469 -0.899 46.981
INVINC 1022 0.010 0.022 -0.029 0.279
LOG(ASSETS) 1080 -1.532 1.675 -7.973 3.738
INTINC 1074 0.002 0.002 0.000 0.016
FSALE 1024 7.688 16.241 0.000 87.084

22
Table 1: Summary statistics

Variables Obs. Mean Std. Dev Min Max

TAXAGG 958 0.00 0.01 -0.06 -0.03
OC 957 0.63 0.18 0.15 0.98
STA 958 0.19 0.24 0.00 0.86
INST 935 0.17 0.18 0.00 0.82
FUND 865 0.13 0.16 0.00 0.77
GOV 917 0.85 0.36 0.00 1.00
LEV 958 0.35 0.21 0.03 0.81
LOSS 958 0.07 0.26 0.00 1.00
SIZE 953 10.02 0.59 8.49 12.35
ROE 949 0.08 0.35 -8.89 0.66
EM 934 0.10 6.61 -19.22 20.65
CAPINT 958 0.30 0.21 0.00 0.86
BODIND 950 0.37 0.06 0.09 0.80
Downloaded by Cornell University Library At 23:32 27 June 2017 (PT)

BODSIZE 950 10.00 2.32 4.00 18.00

AUDFEE 795 14.29 1.15 12.10 18.37
AUDOP 958 0.97 0.17 0.00 1.00
BIG4 958 0.57 0.50 0.00 1.00

Variable definitions: TAXAGG is the measure of firm tax aggressiveness calculated as the
regression error from a book-tax difference prediction model; OC is the ownership
concentration measured as the collective percentage shareholding of top 10 shareholders;
STA is the state percentage shareholding; INST is the percentage of institutional investors’
shareholding; FUND is the percentage of mutual funds’ shareholding; GOV is a dummy
variable which equals to 1 when the firm controlling shareholder is the government or its
agency, and 0 if it is a private investor or firm; LEV is the market value financial leverage
ratio which equals to the book value of debt divided by the total of market capitalization and
book value of debt; LOSS is a dummy variable which equals to 1 if firm reports a net loss for
the year, and 0 if otherwise; SIZE is the log of the market value of equity at the fiscal year-
end; ROE is the return on equity; EM is earnings management calculated as the value of
discretionary accruals which is the prediction error when regressing total accruals against
change in sales, fixed assets, and industry and year fixed effects. CAPINT is the capital
intensity ratio calculated as the fixed assets divide by total assets. BODIND refers to the
percentage of board members that are independent. BODSIZE is the size (number) of board
members. AUDFEE is the log of annual audit fees. AUDOP is a dummy which equals to 1 if
the auditor opinion is standard or 0 if it is non-standard. BIG4 is a dummy which equals to 1
if the firm’s auditor is one of the “Big-4” accounting firms.

23
Downloaded by Cornell University Library At 23:32 27 June 2017 (PT)

Table 2: The correlations matrix

TAXA CAPIN BODIN BODSI AUDF AUDO

OC STA INST FUND GOV LEV LOSS SIZE ROE EM
GG T D ZE EE P
OC 0.16*
STA 0.11* 0.32*
INST -0.10 0.10* 0.21*
FUND -0.13* 0.07* 0.25* 0.88*
GOV 0.15* 0.17* 0.31* -0.01 -0.08*
LEV 0.13* 0.01 -0.06* -0.03 -0.01 -0.01
LOSS -0.09* -0.07* 0.03 -0.12* -0.12* 0.03 0.17*
SIZE 0.17* 0.49* 0.24* 0.29* 0.28* 0.21* 0.04 -0.15*
ROE 0.08* 0.08* 0.00 0.14* 0.13* -0.04 -0.13* -0.37* 0.13*
EM 0.19* -0.04* -0.08* -0.08* -0.07* -0.06 0.1* -0.19* -0.11* 0.11
CAPINT 0.05 0.23* 0.16* 0.02 -0.01 0.19* -0.09* 0.13* 0.16* -0.04* -0.35*
BODIND 0.08* 0.07* 0.02 -0.06 -0.07* 0.05 0.16* 0.03 0.16* -0.06 0.11* -0.15*
BODSIZE 0.09* 0.09* 0.14* 0.13* 0.14* 0.23* 0.00 -0.03 0.22* 0.03 -0.11* 0.15* -0.34*
AUDFEE 0.28* 0.46* 0.19* 0.08* 0.06 0.13* 0.27* -0.03 0.73* 0.09* -0.06 0.11* 0.29* 0.08*
AUDOP 0.07* 0.08 0.05 0.08* 0.06 0.00 -0.16* -0.12* 0.15* 0.02 0.00 0.00 0.03 0.00 0.09*
BIG4 0.14* 0.34* 0.13* 0.12* 0.08* 0.07* 0.13* -0.03 0.48* 0.03 -0.07* 0.09*3 0.17* 0.13* 0.63* 0.07*
*denotes 5% significance or above.

24
Table 3: Mean differences tests

Government-related Non-government-related
Variables Difference
firms (GOV=1) firms (GOV=0)

TAXAGG 0.121 -0.073 -0.194***

OC 0.651 0.568 -0.083***
STA 0.237 0.019 -0.218***

High ownership Low ownership

Variables concentration concentration Difference
(OC>64%) (OC<64%)

TAXAGG 0.095 -0.091 -0.186***

High institutional Low institutional

Downloaded by Cornell University Library At 23:32 27 June 2017 (PT)

Variables shareholding shareholding Difference

(INST>11%) (INST<11%)

TAXAGG -0.036 0.032 0.068

High mutual fund Low mutual fund

Variables shareholding shareholding Difference
(FUND>7%) (FUND<7%)

TAXAGG -0.027 0.061 0.088***

High and low ownership concentration and institutional shareholdings are defined by being
above or below their median value. ***1% significance

25
Table 4: Ownership structure and firm tax aggressiveness

Model (1) (2) (3) (4) (5)

STA 0.329**
(2.49)
GOV 0.263***
(3.18)
OC 0.377*
(1.94)
INST -0.606***
(-3.15)
FUND -0.473*
(-1.83)
BODIND 0.471 0.403 0.592 0.404 0.099
Downloaded by Cornell University Library At 23:32 27 June 2017 (PT)

(1.07) (0.96) (1.34) (0.92) (0.24)

BODSIZE 0.040*** 0.038*** 0.045*** 0.043*** 0.031**
(2.90) (2.71) (3.24) (3.16) (2.32)
AUDFEE 0.226*** 0.242*** 0.216*** 0.196*** 0.164***
(5.39) (5.56) (5.06) (4.66) (4.31)
AUDOP 0.400 0.248 0.407 0.372 0.088
(1.57) (0.96) (1.58) (1.40) (0.29)
BIG4 0.031 0.024 0.019 0.010 0.070
(0.45) (0.35) (0.27) (0.15) (0.98)
LEV 0.448*** 0.383** 0.422*** 0.466*** 0.477***
(2.73) (2.34) (2.59) (2.81) (2.75)
EM 2.047*** 2.078*** 2.125*** 1.971*** 1.993***
(4.84) (4.87) (4.99) (4.59) (4.89)
ROE 0.285 0.303 0.260 0.340 -2.362***
(0.49) (0.54) (0.44) (0.62) (-3.78)
LOSS -0.043 -0.028 -0.019 -0.029 -0.626***
(-0.21) (-0.14) (-0.09) (-0.15) (-2.82)
CAPINT 0.304* 0.330* 0.298* 0.293* -0.063
(1.80) (1.95) (1.77) (1.76) (-0.41)
SIZE -0.232*** -0.239** -0.230*** -0.076 0.085
(-2.73) (-2.55) (-2.80) (-0.85) (1.01)
Constant -2.386*** -2.394*** -2.385*** -3.196*** -3.565***
(-4.24) (-3.91) (-4.30) (-5.41) (-5.55)
Observations 781 752 780 763 705
R-squared 0.210 0.224 0.207 0.217 0.280
Notes: All models are OLS regressions controlling for (CSRC) industry dummies and year
dummies. The dependent variable is the measure of firm tax aggressiveness TAXAGG. Results
on industry dummies and year dummies are omitted to conserve space. *10% significance,
**5% significance, ***1% significance. Variable definitions follow Table 1.

26
Table 5: Further evidence on ownership structure and firm tax aggressiveness

Model (1) (2) (3) (4) (5) (6)

STA*INST 1.444**
(1.97)
STA*FUND 1.100
(1.30)
GOV *INST 1.037**
(2.29)
GOV* FUND 0.427
(0.88)
OC*INST 2.886***
(2.59)
OC*FUND 4.382***
Downloaded by Cornell University Library At 23:32 27 June 2017 (PT)

(3.80)
STA 0.085 0.234
(0.50) (1.42)
GOV 0.064 0.174*
(0.62) (1.82)
OC -0.053 -0.182
(-0.24) (-0.83)
INST -0.923*** -1.399*** -2.446***
(-3.65) (-3.25) (-3.40)
FUND -0.842** -0.646 -3.275***
(-2.44) (-1.42) (-4.32)
Marginal effect -0.641*** -0.627*** -0.573*** -0.306*** -0.649*** -0.547***
Observations 763 705 748 691 763 705
R-squared 0.230 0.293 0.241 0.302 0.230 0.300
Controls YES YES YES YES YES YES
Notes: All models are OLS regressions controlling for (CSRC) industry dummies and year dummies. The
dependent variable is the measure of firm tax aggressiveness TAXAGG. Marginal effect refers to the
marginal effects of INST and FUND which are conditional on the values of STA, GOV, and OC where
appropriate in the above models. We compute the mean of the coefficients conditional on the values of
STA, GOV, and OC and test if these means are significantly different from 0. Results on industry dummies
and year dummies are omitted to conserve space. *10% significance, **5% significance, ***1%
significance. Variable definitions follow Table 1.

27
Downloaded by Cornell University Library At 23:32 27 June 2017 (PT)

Table 6: Robustness tests using lagged ownership variables

Model (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11)
L.STA 0.362*** 0.123 0.150
(2.91) (0.67) (0.88)
L.GOV 0.263*** 0.084 0.138
(3.18) (0.78) (1.33)
L.OC 0.456** 0.217 0.021
(2.36) (0.79) (0.08)
L.INST -0.586*** -0.915*** -1.058** -1.756***
(-2.91) (-3.28) (-2.29) (-2.82)
L.FUND -0.830*** -1.220*** -1.007** -2.771***
(-3.64) (-3.59) (-2.22) (-4.49)
L.STA*L.INST 1.296*
(1.82)
L.STA*L.FUND 1.223
(1.55)
L.GOV*L.INST 0.656
(1.39)
L.GOV*L.FUND 0.316
(0.67)
L.OC*L.INST 1.823*
(1.81)
L.OC*L.FUND 2.994***
(3.04)
Observations 778 752 778 623 692 623 692 610 679 623 692
R-squared 0.212 0.224 0.210 0.287 0.289 0.303 0.302 0.307 0.299 0.301 0.304
Controls YES YES YES YES YES YES YES YES YES YES YES
Notes: All models are OLS regressions controlling for (CSRC) industry dummies and year dummies. The dependent variable is the measure of firm tax aggressiveness
TAXAGG. Results on industry dummies and year dummies are omitted to conserve space. *10% significance, **5% significance, ***1% significance. Variable
definitions follow Table 1.

28