J Bus Ethics

DOI 10.1007/s10551-013-1809-9

The Effect of Corporate Social Performance on Financial

Performance: The Moderating Effect of Ownership
Concentration
Chih-Wei Peng • Mei-Ling Yang

Received: 25 March 2012 / Accepted: 9 July 2013

Ó Springer Science+Business Media Dordrecht 2013

Abstract The purpose of this study is to extend prior relations between CSP and FP to be largely inconclusive
research on this topic by investigating whether the impact (Margolis and Walsh 2003). Some studies have found a
of ownership concentration moderates the link between negative relationship (e.g., Gollop and Roberts 1983; Smith
corporate social performance (CSP) and financial perfor- and Sims 1985) and suggest that firm investment in CSP
mance (FP). This study uses a set of unique, hand-collected diverts funds that could be used for productive investments.
pollution control data to measure CSP, based on a sample Other studies have found a positive relationship (e.g.,
of Taiwanese listed companies during the period from 1996 Simpson and Kohers 2002; Spicer 1978; Wu 2006; Chien
to 2006. The results of the empirical analysis provide firm and Peng 2012; Servaes and Tamayo 2013) and suggest
support for the idea that the divergence between control that not only may CSP be considered a business objective
rights and the cash flow rights of controlling owners neg- in and of itself but that it serves as a means to an end in
atively moderates the link between social and short- and regard to FP. However, still others have reported no rela-
long-run FP. tionship at all (e.g., Moore 2001; Seifert et al. 2003; Soana
2011). While some attention has been paid to the associ-
Keywords Corporate social performance  Financial ation between CSP and FP, the impact of corporate gov-
performance  Ownership structure  Control rights  ernance on CSP and FP has received even less attention.
Cash flow rights Friedman (1970) asserts that engaging in CSP is symp-
tomatic of an agency problem or a conflict between the
interests of CEOs and shareholders. In essence, agency
Introduction theory provides a theoretical basis with which to describe
the potential divergent and convergent interests between
An increasing number of scholars and practitioners today CEOs and other stakeholders and to predict how this
are paying attention to corporate social responsibility affects CSP.
(CSR). In particular, an association between corporate According to agency theory, separation of ownership
social performance1 (CSP) and financial performance (FP) and control leads to a divergence in the pursuit of mana-
has gained enthusiastic research interest (McGuire et al. gerial interests that are not in concurrence with owner
1988). However, previous studies have found empirical interests (Jensen and Meckling 1976), and thus monitoring
CEO decisions becomes essential for boards of directors in
order to assure that shareholder interests are protected
C.-W. Peng (&)
(Fama and Jensen 1983). In keeping with agency theory
Department of Accounting, National Changhua University
of Education, Changhua 500, Taiwan, ROC prescriptions, stronger managerial alignment will lead to
e-mail: [email protected]
1
Maron (2006) indicates that corporate social performance (CSP)
M.-L. Yang has multiple dimensional constructs that measure organizational
Department of Finance, National Sun Yat-Sen University, behavior across a wide range of dimensions, and the current study
Kaohsiung 804, Taiwan, ROC uses one of these measures, which is the extent that a firm invests in
e-mail: [email protected] pollution control equipment.

123
 C.-W. Peng, M.-L. Yang

maximization of owner interests (Jensen 1983). Jensen and This article extends the existing literature in two distinct
Meckling (1976) indicate that an increase in managerial ways. First, the impact of ownership concentration on the
ownership can help with regard to the alignment of the link between CSP and FP has become of great interest to
divergent interests of CEOs and shareholders. However, a shareholders, practitioners, and governance regulators.
controlling shareholder usually exercises control but owns There have been, however, relatively few empirical studies
only a small fraction of firm ownership as a result of the examining this issue (Jo and Harjoto 2011). Claessens et al.
use of pyramidal control structures and cross-stockholdings (2000) point out that most East Asian companies, including
in emerging markets (Claessens et al. 2000, 2002; La Porta those in Taiwan, have lower average ratios of cash flow to
et al. 1999, 2002). Thus, because of high control–owner- voting rights than those of Western European companies.2
ship disparities, these controlling shareholders have an Fan and Wong (2002) also argue that the actual degree of
incentive to pursue their private interests at the expense of ownership concentration in East Asia should be higher than
other shareholders. This study is an attempt to investigate the nominal statistics notwithstanding the data limitations.
whether the divergence between the control rights and cash Our study is an attempt to shed light on the mixed results of
flow rights of controlling owners (hereafter, control–cash previous studies related to the CSP–FP link by considering
flow divergence) moderates the link between CSP and FP. the control–cash flow divergence in emerging markets. It is
In this article, we expect control–cash flow divergence to anticipated that the results will be important as a means by
negatively moderate the link between CSP and FP. Our which to assess whether CSP, along with well-designed
argument is built on the following premise: First, when corporate governance systems, can enhance firm short- and
controlling shareholders have control rights in excess of long-run value.
their cash flow rights, this situation may exacerbate infor- Second, Margolis and Walsh (2003) summarize over
mation asymmetry (Claessens et al. 2002). Managerial 120 studies between 1971 and 2001 regarding the CSP–FP
decisions in the area of CSP are likely to reflect both high link. They suggest that previous studies have been subject
information asymmetry and low programmability (Deckop to various imperfections, such as measurement problems
et al. 2006). Thus, controlling shareholders will have greater related to the CSP, the FP, omitted variables and a lack of
incentive to pursue self-interest at the expense of society and methodological rigor. Our study is an attempt to develop an
other stakeholders (Amihud and Lev 1981; Jensen and appropriate statistical method that applies factor analysis to
Meckling 1976). Second, compensation systems may be integrate different FP measures into a single index, such as
ineffective with regard to providing an incentive for CEOs to return on assets (ROA), return on equity (ROE), earnings
improve firm CSP when control–cash flow divergence is per share (EPS), and cash flows to assets (CFA). This
high. Core et al. (1999) suggest that firms with weaker approach will strongly support the premise that an aggre-
governance structures have greater agency problems that gated indicator prevents the conflicting results character-
lead to the ability of CEOs to extract greater compensation. istic of different financial indicators. In addition, we also
In such an ineffective corporate governance system, CEOs focus on both short-term and long-term FP. This mea-
may focus on meeting short-term earnings goals, and they surement may prevent the conflicting results shown in prior
are reluctant to engage in CSP because issues related to studies regarding the association between CSP and short-
environmental neglect are not likely to be caught immedi- term FP.
ately (Short 2004). Finally, CSP also affects the reputation The rest of this article is organized as follows: second
of a firm, with good CSP improving investor perceptions of section presents the hypotheses, while third section
companies (e.g., Orlitzky et al. 2003), which in turn may describes the sample and descriptive statistics. Fourth
affect stock prices. Well-designed corporate governance section presents the research design and empirical results,
systems will align managers to engage in CSP actions that and fifth section concludes this study with recommenda-
can aid in improving firm reputation. tions for managers and regulators, as well as suggesting
The sample used in this study is composed of all of some of the limitations of this work.
Taiwan’s listed companies in five highly polluting indus-
tries: cement, plastics, chemicals, paper and pulp, and iron
and steel, during the period of 1996–2006. The empirical Hypotheses Development
results show that: (1) a weak relationship between CSP and
short-run FP and control–cash flow divergence negatively In general, CEOs are recognized as having extensive
moderates the link between CSP and short-term FP, and (2) decision-making power and the ability to significantly
there exists a positive relationship between CSP and long-
run FP. However, there is found to be a significant inverse 2
Claessens et al. (2000) report that the average ratio of cash flow to
function of control–cash flow divergence on the CSP and voting right in East Asian companies is lower than that of Western
long-run FP relationship. European companies (i.e., 0.75 \ 0.87).

123
Effect of CSP on FP

influence their firm’s CSP (Kochan 2002; Orlitzky and shareholders, prior studies have found divergence in con-
Swanson 2002). The next question is how CEOs as indi- trol rights and cash flow rights to be associated with lower
viduals consider the economic utility of CSP as it relates to market valuations and firm performance (Claessens et al.
their firms and to themselves. Agency theory provides a 2002) and reduced earnings quality (Fan and Wong 2002;
theoretical basis with which to describe the potential Francis et al. 2003; Haw et al. 2004).
divergent and convergent interests between CEOs and We posit that control–cash flow divergence might serve
other stakeholders and to predict how this affects CSP. as an important moderator on the link between CSP and FP.
According to agency theory, separation of ownership First, the ownership concentration of firms exacerbates
and control leads to a conflict of interest between principals agency costs. The influential role of the controlling
and agents and is commonly applied to the divergence shareholders generates information asymmetry caused by
observed in the pursuit of managerial interests (agents) their incentive to obtain the necessary information to
versus owner interests (principals) (Jensen and Meckling effectively control corporate policies (Attig et al. 2006).
1976). To eliminate self-interest behavior by an agent, a This perspective is documented by Claessens et al. (2002),
principal may monitor managerial decisions that shift some who showed that control–cash flow divergence leads to the
of the performance risk from the principal to the agent and agency problem and decreases firm value. Decisions in the
thus more closely align principal–agent interests (Fama and area of CSP are likely to reflect both high information
Jensen 1983). However, the fundamental agency problem asymmetry and low programmability (Deckop et al. 2006).
for listed companies in emerging markets, including Tai- McWilliams et al. (2006) indicate that asymmetric infor-
wan, is not a conflict of interest between managers and mation about CSP allows CEOs to hide the more practical
owners, but a conflict of interest between controlling motivations behind their CSR activities although they may
shareholders and minority shareholders (Shleifer and perceive that many external stakeholders view CSR activ-
Vishny 1997). Claessens et al. (2000) find that 80 % of the ity more favorably. Alternatively, the extraction of private
management in Taiwanese listed companies is derived control benefits, if detected, is likely to invite external
from the controlling family, whereas Yeh et al. (2001) intervention by minority shareholders, analysts, stock
report that 76 % of Taiwanese listed companies are con- exchanges, or regulators. Thus, we propose that control–
trolled by family shareholders. Thus, concentration of cash flow divergence increases the cost of monitoring and
ownership by controlling shareholders has been associated thus enhances agents’ opportunities to pursue self-interest
with the principal–principal agency problem in emerging at the expense of principals.
economies such as that of Taiwan. Second, compensation systems are ineffective in respect
The concentration of ownership by controlling share- to providing incentives for CEOs to improve firm CSP
holders has two major effects based on: (1) the conver- when firms have bad corporate governance systems. Core
gence-of-interests hypothesis and (2) the entrenchment et al. (1999) suggest that firms with weaker governance
hypothesis. Regarding the convergence-of-interests structures have greater agency problems and that CEOs are
hypothesis, when cash flow rights are concentrated in the therefore able to extract greater compensation. Both Chien
hands of a single shareholder, that controlling shareholder and Peng (2012) and Falck and Heblich (2007) indicate
has an incentive to optimize firm performance, as increases that CSR investments are most likely to pay off in the long
in firm wealth translate into increases in personal wealth run. As a result, CSR investment might actually be a bur-
(Jensen and Meckling 1976). Alternatively, the entrench- den for firms in the short run. Issues related to environ-
ment hypothesis argues that when ownership rights are mental neglect are not likely to be caught immediately
concentrated in the hands of a single shareholder, that (Short 2004) and are unlikely to show up in short-term FP.
controlling shareholder exercises greater discretionary As a result, CEOs may be concerned with meeting their
power and tends to be more self-interested, which leads to short-term goals and may not invest in CSP due to goal
even more severe agency problems (Jensen and Ruback conflicts related to time horizons (Bushee 1998). In con-
1983). However, a single large controlling shareholder trast, McGuire et al. (2003) indicate that corporate gover-
frequently possesses control rights in excess of cash flow nance effects regarding social performance may tend to be
rights and enjoys almost total control over managers by more reactive in nature as a result of deterring socially
using pyramidal control structures and cross-stockholdings risky behavior. We expect that firms with ineffective
(Claessens et al. 2000, 2002; La Porta et al. 1999, 2002). executive compensation mechanics may discourage CEOs
The high level of separation between cash flow rights and to pursue social objectives.
voting rights motivates the controlling shareholder to Finally, CSP also affects organizational reputation,
entrench him- or herself at the expense of outside minority thereby improving investor perceptions of firms (e.g., Or-
shareholders (Claessens et al. 2000; La Porta et al. 1999). litzky et al. 2003), which in turn may affect stock prices.
With regard to the expropriation incentive of controlling For example, KLD Research & Analytics dropped Coca-

123
 C.-W. Peng, M.-L. Yang

Cola Co. from its Broad Market Social Index in July 2006 information. For example, the current study carefully iden-
because of concerns about the company’s labor and envi- tifies the level of divergence between ownership and control
ronmental practices in the developing world. As a result, rights for controlling shareholders in order to test for nega-
TIAA-CREF, the largest U.S. retirement fund, subse- tive entrenchment effects. A binary summary of KLD’s
quently sold more than 50 million shares of Coca-Cola Co. positive and negative ratings may not fully catch this effect.
stock (Chatterji et al. 2009). Thus, we expect that good In addition, KLD rated companies are not globally focused
corporate structure is associated with better monitoring and only trade on the US stock exchange (Turker 2009).
because failure to comply with environmental regulations Literature on this topic has indicated that it is also important
can lead to penalties and bad publicity that effective to assess corporate CSR practices in places in the world
monitors would consider avoidable. other than in the U.S. (Chapple and Moon 2005; Matten and
We propose a negative relationship between CSP and Moon 2008). Therefore, this study hand-collected pollution
short-term or long-term FP for Taiwanese firms with a high control investment data were used to test CSR practices in
degree of control–cash flow divergence (i.e., the Taiwan in order to represent emerging markets.
entrenchment effect is expected to be found). Accordingly, In addition, the ultimate control and ownership structure
this study proposes the following hypotheses (in alternative and financial variable data of each firm were mainly col-
form): lected from company prospectuses and the Taiwan Eco-
nomic Journal (TEJ) database.
Hypothesis 1 Control–cash flow divergence negatively
moderates the relationship between corporate social per-
Descriptive Statistics
formance and short-term financial performance.
Hypothesis 2 Control–cash flow divergence negatively Table 1 shows the descriptive statistics for all the variables
moderates the relationship between corporate social per- included in this study. It can be seen that the short-term FP
formance and long-term financial performance. (Short-run FP) has a mean of -0.142; the 3-year window
long-term FP (Long-run FP1) has a mean of -0.151, and
the 5-year window long-term FP (Long-run FP2) has a
Sample and Descriptive Statistics mean of -0.111, suggesting, on average, that most firms
exhibited negative short-term and long-term FP during the
Sample sample year. In addition, the mean of Pollution (indicator
variable) is 0.210, suggesting that about 21 % of sample
This study’s sample is composed of all firms in the cement, firms mention pollution control investment disclosures in
plastics, chemical, paper and pulp, and iron and steel their annual reports. The mean of Divergence is 44.1 %,
industries with annual environmental capital expenditures in
Taiwan’s SFI (Securities and Futures Institute) database. By Table 1 Descriptive statistics
examining the annual reports of firms listed on the Taiwan Variable N Mean SD Minimum Maximum
Stock Exchange (TSE) for the fiscal years from 1996 to
Short-run FP 400 -0.142 0.626 -1.574 1.359
2006, we hand-collected data on their pollution control
investments in order to measure CSP. This data includes the Long-run FP1 364 -0.151 0.544 -1.657 1.240
date, amount, purposes, and expected benefits of each pol- Long-run FP2 333 -0.111 0.460 -1.299 0.969
lution control investment. As a result, a sample of 400 firms- Pollution 400 0.210 0.407 0 1
year observations was obtained for the purposes of this Divergence 400 0.441 0.246 0 0.973
study. 84 firm-year observations included pollution control MTB 400 1.503 0.954 0.068 6.358
investment data, and 316 firm-year observations did not SIZE 400 15.448 1.193 11.749 19.244
have this information, or only disclosed the amount that they DEBT 400 0.397 0.148 0.093 0.789
expected to spend for this purpose. Notably, previous studies Short-run FP denotes the difference in a firm’s factor score between
have commonly used the Kinder, Lydenberg, and Domini Year(t ? 1) and Year(t - 1). Long-run FP1 then denotes the differ-
(KLD) database to measure CSP (Graves and Waddock ence in the average factor score of Year(t ? 1) to Year(t ? 3) and
Year(t - 3) to Year(t - 1). Long-run FP2 then denotes the difference
1994; Hillman and Keim 2001; Johnson and Greening in the average factor score of Year(t ? 1) to Year(t ? 5) and
1999). However, KLD rated companies are not globally Year(t - 5) to Year(t - 1). Pollution is an indicator variable, which
focused and only trade on the US stock exchange based on is set to a numerical value of 1 for firms with pollution control
eight attributes of social activities (Turker 2009). As a result, investment, and 0 otherwise. Divergence represents the divergence
between the ultimate controlling shareholder’s voting rights and cash
they only disclose binary summaries for ENVIRONMENT flow rights. MTB is the ratio of a firm’s market value of equity to its
and GOVERNANCE. The KLD’s dichotomous variable for book value of equity. SIZE is the natural logarithm of a firm’s total
‘‘strength’’ and ‘‘concern’’ may ignore some valuable assets. DEBT is measured by the ratio of debt to assets

123
Effect of CSP on FP

suggesting that, on average, the ultimate controlling shareholder control and ownership on CSP and short-run
shareholders’ voting rights are substantially in excess of FP. All t-values are estimated based on standard errors
their cash flow rights, and thus they have greater incentive adjusted for heteroskedasticity (White 1980).
to expropriate minority wealth. Finally, the means of MTB, Shortrun FPi ¼ a0 þ a1 Pollution þ a2 Divergence
SIZE, and DEBT are 1.503, 15.448, and 0.397, respectively.
þ a3 Pollution  Divergence þ a4 MTB
Table 2 shows the Pearson correlation matrix of all the
variables. It can be seen that the Pollution 9 Divergence þ a5 SIZE þ a6 DEBT þ a7 PRIOR
interaction variable is significantly and negatively corre- þ a8 GDPGW þ a9 PRIOR  Pollution
lated with FP (the correlation between Short-run FP and þ ei;t ð1Þ
Pollution 9 Divergence is -0.124 with a p value \0.05;
the correlation between Long-run FP2 and Pollu- Short-run FP. Factor analysis with varimax rotation is
tion 9 Divergence is -0.094 with a p value \0.1). This performed to generate factor scores from four performance
finding is consistent with the negative entrenchment effect variables (ROE, ROA, EPS, and CFA) since prior studies
hypothesis, which suggests that the divergence in control- have used these financial variables (Jaggi and Freedman
ling shareholders’ control and ownership moderates the 1992). In this study, the factor analysis to identify the
link between CSP and FP. In addition, Table 2 shows that factor patterns yields only one factor (with an eigenvalue
the highest correlation between the dependent and inde- [1) that captures the commonality among these four
pendent variables is 0.502 (the correlation between Pollu- variables, and its factor score essentially represents an
tion and Pollution 9 Divergence). Gujarati (1995) suggest index that integrates these four performance measures.
that correlations in excess of 0.8 may indicate collinearity, Notably, the factor scores were standardized to a mean of
and thus this study does not expect collinearity to be a zero and a standard error of one, which ranged from
problem for the multivariate estimations. approximately -3.0 to ?3.0. Thus, employment of factor
scores to measure FPs can reduce the impact of an
extreme value, which may inadvertently distort the
Research Design and Empirical Results relationship between the variables of interest. Essentially,
the estimate of the factor score is based on the
Test of Hypothesis 1: The Effect of Short-Run contribution or weight (termed loadings) of each
Performance variable. In contrast, an average number of financial
variables assume that these variables are equally weighted.
The following empirical model (1) is used in this study to Thus, the employment of an integrated factor score may
test the effect of the divergence in regard to controlling avoid the inconclusive results characteristic of previous

Table 2 Pearson correlations

Short-run Long-run Long-run Pollution Divergence Pollution 9 MTB SIZE DEBT
FP FP1 FP2 Divergence

Short-run FP 1 0.292** 0.493** -0.059 -0.158** -0.124** -0.048 -0.136** 0.066

Long-run FP1 1 0.667** 0.048 -0.214** -0.054 -0.288** 0.242** 0.251***
Long-run FP2 1 0.038 -0.313** -0.094* 0.328** 0.209** 0.202***
Pollution 1 -0.022 0.502** 0.115* -0.046 0.059
Divergence 1 0.137*** -0.010 -0.133** 0.042
Pollution 9 Divergence 1 0.101** -0.027 -0.081
MTB 1 -0.082 -0.239
SIZE 1 -0.028
DEBT 1
Short-run FP denotes the difference in a firm’s factor score between Year(t ? 1) and Year(t - 1). Long-run FP1 then denotes the difference in
the average factor score of Year(t ? 1) to Year(t ? 3) and Year(t - 3) to Year(t - 1). Long-run FP2 then denotes the difference in the average
factor score of Year(t ? 1) to Year(t ? 5) and Year(t - 5) to Year(t - 1). Pollution is an indicator variable, which is set to a numerical value of
1 for firms with pollution control investment, and 0 otherwise. Divergence represents the divergence between the ultimate controlling share-
holder’s voting rights and cash flow rights. MTB is the ratio of a firm’s market value of equity to its book value of equity. SIZE is the natural
logarithm of a firm’s total assets. DEBT is measured by the ratio of debt to assets
***, **, and * represent significance at the 1, 5, and 10 % levels using the two-tailed test, respectively

123
 C.-W. Peng, M.-L. Yang

studies that have used individual financial indicators. Table 3 Definition of variables
Following the approach proposed by Megginson et al. Variables Definition
(1994) to measure short-term (1-year window) FP, letting
Year(t) be the year of investment, Short-run FP denotes Dependent variable
the difference in a firm’s factor score between Year(t ? 1) Short-run Factor analysis is performed to generate factor scores
and Year(t - 1). FPi from four accounting variables (ROE, ROA, EPS, and
CFA). Following the approach proposed by
Pollution is an indicator variable which is set to a Megginson et al. (1994), financial performance is
numerical value of 1 for firms with pollution control measured with three different windows. Let Year(t) be
investments, and 0 for firms without them. the year of investment; Short-run FP then denotes the
Divergence represents the divergence between the ulti- difference in a firm’s factor score between
Year(t ? 1) and Year(t - 1)
mate controlling shareholder’s voting rights and cash flow
Long-run As above, Long-run FP1 then denotes the difference in
rights. The core agency problem arising from a controlling FP1 the average factor score of Year(t ? 1) to Year(t ? 3)
shareholder is measured by the divergence of his or her and Year(t - 3) to Year(t - 1)
control and ownership. In previous research, it has been Long-run As above, Long-run FP2 then denotes the difference in
shown that control rights and ownership can be measured FP2 the average factor score of Year(t ? 1) to Year(t ? 5)
based on voting rights and cash flow rights, respectively and Year(t - 5) to Year(t - 1)
(Claessens et al. 2000; La Porta et al. 1999). Regarding Independent variables
Hypothesis 1, this study predicts a3 \ 0, indicating the link Pollution Pollution is an indicator variable, which is set to a
numerical value of 1 for firms with pollution control
between CSP and short-run FP to be negatively moderated investment, and 0 otherwise
by the high divergence in controlling shareholder control
Divergence Divergence represents the divergence between the
and ownership. ultimate controlling shareholder’s voting rights and
Fama and French (1992) use market-to-book ratio, firm cash flow rights
size, and firm risk as control variables. Thus, MTB denotes Control variables
the ratio of the market value of equity to the book value of MTB MTB is the ratio of a firm’s market value of equity to its
equity. SIZE denotes the natural logarithm of total assets. book value of equity
DEBT measured by the ratio of debt to assets. Prior studies SIZE SIZE is the natural logarithm of a firm’s total assets
have used DEBT to test the association between firm risk DEBT DEBT is measured by the ratio of debt to assets
and FP (Jensen 1986; Harris and Raviv 1991). PRIOR is PRIOR PRIOR represents the difference in factor scores
used to examine whether pre-investment performance between the investment year and the pre-investment
year
affects post-investment performance. GDPGW is measured
GDPGW GDPGW as measured by the annual growth rate of
by the annual growth rate of Taiwan GDP to control the Taiwan’s GDP
macroeconomic factor. Finally, with regard to the last
control variable, it is open to question whether firms with
better pre-investment performance tend to over comply
with regulations, in other words, whether firms with greater Seifert et al. (2003), and Soana (2011). The coefficient of
FP are more likely to engage in CSP because they can Divergence is -0.004, and this is significant at the 1 %
afford to, relative to investments in CSP providing a level, which means that a firm’s short-run FP will be
greater FP return. To investigate this in more detail, this adversely affected by the level of the controlling share-
study also includes an interactive pre-investment FP vari- holders’ divergence in regard to cash flow rights and
able and pollution control indicator (PRIOR 9 Pollution) control rights. This finding is consistent with the argument
in the regression models as control variables. Table 3 of Claessens et al. (2002), who suggest that divergence in
summarizes the definitions of the key variables employed control and cash flow rights is associated with worse firm
in this study. performance. The coefficient for Pollution 9 Divergence
Table 4 shows the results of regressing short-term FP is -0.024, which is significant at the 1 % level, and this
(Short-run FP) on Pollution, Divergence, Pollu- result is therefore consistent with the prediction of
tion 9 Divergence and on the control variables. All t-val- Hypothesis 1, indicating that the link between CSP and
ues are based on standard errors using White’s (1980) short-run FP is significant and negative when the control–
correction; the coefficient for Pollution is 0.018, and this is cash flow divergence is high. In addition, for the control
insignificant at the 10 % level, which indicates no direct variables, SIZE is 0.075, and DEBT is 0.559, both
relationship between CSP and firm short-run FP. This of which are significant at the 1 % level. However, MTB
result is similar to the stream of findings for Moore (2001), is -0.009; PRIOR0 is 0.007; GDPGW is 0.090, and

123
Effect of CSP on FP

Table 4 The effect of ownership concentration on the relationship Hypothesis 2, this study predicts the coefficients for Pol-
between corporate social performance and short-run financial lution 9 Divergence \ 0, indicating the link between CSP
performance
and long-run FP to be negatively moderated by the high
Independent variable Parameter Standard t-value p value level of divergence in controlling shareholders’ control and
estimates error ownership.
Dependent variable: short-run FP Table 5 shows the regression results regarding long-term
Intercept -1.312*** 0.410 -3.19 \0.01 FP, in which it can be seen that the coefficients for Pollution
Pollution 0.018 0.083 0.22 0.83 are 0.355 and 0.284, which are significant at the 1 % level.
Divergence -0.004*** 0.001 -3.98 \0.01 These results indicate that firms with higher levels of pol-
Pollution 9 -0.024*** 0.008 -3.00 \0.01
lution control investment tend to have higher long-run per-
Divergence formance. The coefficients for Divergence are -0.004 and
MTB -0.009 0.035 -0.26 0.79 -0.002, which are significant at the 10 and 5 % level,
SIZE 0.075*** 0.024 3.09 \0.01 respectively. This finding is consistent with the argument of
DEBT 0.559*** 0.205 2.72 \0.01 Claessens et al. (2002) that the divergence in control and
GDPGW 0.090 0.062 1.44 0.15 cash flow rights is associated with worse long-run firm
PRIOR0 0.007 0.030 0.25 0.80 performance. The coefficients for Pollution 9 Divergence
PRIOR0 9 Pollution 0.002 0.009 0.27 0.78 are shown to be -0.007 and -0.005, which are significant at
F-statistic (p value) = 6.12 (p \ 0.01), adjusted R2 = 9.60 %, the 1, 5 % level, respectively. These results confirm
N = 400 Hypothesis 2, indicating that the relationship between long-
run performance and pollution control investment is sig-
Definitions of variables appear in Table 3. PRIOR0 is the difference in
factor scores between the investment year and prior investment year nificant and negative when the control–cash flow divergence
t(-1). The maximum VIF is 1.40, which is \5, and thus multicol- is high. In addition, the control variables in Table 5 are
linearity is not a major concern similar to the findings in Table 4.
***, **, and * represent significance at the 1, 5, and 10 % levels using In addition, Tobin’s Q is also used as a measure of firm
the two-tailed test, respectively. The t-values based on standard errors market performance for a long-run focus. This is calculated
are adjusted for heteroskedasticity (White 1980)
as the market value of common stock plus the book value
of the preferred stocks and total liabilities deflated by total
assets (Chung and Pruit 1994). The results are reported in
PRIOR0 9 Pollution is 0.002. All of the above variables Table 6. The coefficient for Pollution is shown to be
are insignificant at the 10 % level. positive and significant (coefficient = 0.283, p \ 0.01),
while that for Pollution 9 Divergence is negative and
Test of Hypothesis 2: The Effect of Long-Run significant (coefficient = -0.005, p \ 0.05), consistent
Performance with the results reported in Table 5. Therefore, replacing
FP with Tobin’s Q does not influence the results reported
To test Hypothesis 2, the ordinary least-squares approach above.
was also used to estimate the regression parameters, and
standard regression diagnostics were performed to evaluate Robustness Tests
the reliability of the results (Greene 1997). All t-values are
estimated based on standard errors adjusted for heter- In this section, the following additional tests are discussed
oskedasticity (White 1980). to assure the above findings. First, we used the absolute
As discussed before, factor analysis using a varimax value of pollution control investment to investigate the
rotation was performed to generate factor scores from the association between CSP and FP. Chen and Metealf (1980)
four performance variables (ROE, ROA, EPS, and CFA). point out that the size effect will affect the above rela-
Following the approach proposed by Megginson et al. tionship since large firms tend to provide more voluntary
(1994), letting Year(t) be the year of investment, Long-run pollution control records than is the case for small firms. As
FP1 denotes the difference in the average factor score of a result, in this study, the natural log of the 84 firms’
Year(t ? 1) to Year(t ? 3) and Year(t - 3) to Year(t - 1), investment on pollution control is used to control for the
and Long-run FP2 denotes the difference in the average size effect and test the hypotheses. The first to third col-
factor score of Year(t ? 1) to Year(t ? 5) and Year(t - 5) umns in Table 7 show the empirical results, and the results
to Year(t - 1). It is also worth noting that the windows we are generally consistent with prior findings, while the result
use to measure long-run performance are those that have for Pollution 3 Divergence is negative and significant
been commonly adopted in the literature (for example, (coefficient = -0.002, -0.001, and -0.001; p \ 0.01,
Loughran and Ritter 1995, among others). Regarding p \ 0.01, and p \ 0.05).

123
 C.-W. Peng, M.-L. Yang

Table 5 The effect of ownership concentration on the relationship between corporate social performance and long-run financial performance
Dependent variable: long-run FP1 (N = 364) Dependent variable: long-run FP2 (N = 333)
Independent variables Parameter estimates t-value p value Independent variables Parameter estimates t-value p value

Intercept -1.693*** -5.10 \0.01 Intercept -0.750*** -2.70 \0.01

Pollution 0.355*** 2.85 \0.01 Pollution 0.284*** 2.86 \0.01
Divergence -0.004* -1.65 0.10 Divergence -0.002* -3.17 \0.01
Pollution 9 Divergence -0.007*** -2.73 \0.01 Pollution 9 Divergence -0.005** -2.40 0.02
MTB -0.114*** -4.04 \0.01 MTB -0.155*** -6.88 \0.01
SIZE 0.090*** 4.43 \0.01 SIZE 0.053*** 3.25 \0.01
DEBT 0.879*** 4.76 \0.01 DEBT 0.417*** 3.03 \0.01
GDPGW 0.004 0.55 0.57 GDPGW -0.001 -0.11 0.92
PRIOR1 0.238*** 9.06 \0.01 PRIOR2 0.153*** 6.92 \0.01
PRIOR1 9 Pollution -0.147** -2.43 0.02 PRIOR2 9 Pollution -0.069* -1.65 0.10
F-statistic (p value) 17.14 (p \ 0.01) F-statistic (p value) 19.00 (p \ 0.01)
Adjusted R2 22.97 % Adjusted R2 30.92 %
PRIOR1 is the difference in factor scores between the investment year and Avg(-3,-1). PRIOR2 is the difference in factor scores between the
investment year and Avg(-5,-1). Definitions of other variables appear in Table 3. The maximum VIF is 4.59, which is much \5, and thus
multicollinearity is not a major concern
***, **, and * represent significance at the 1, 5, and 10 % levels using the two-tailed test, respectively. The t-values based on standard errors are
adjusted for heteroskedasticity (White 1980)

Table 6 The effect of ownership concentration on the relationship Second, it is important to clarify the definition of firms
between corporate social performance and Tobin’s Q not providing relevant data or only disclosing the amount
Independent variable Parameter Standard t- p value that they expect to spend on pollution control. Thus, this
estimates error value study also includes an additional 46 firm observations in
Dependent variable: Tobin’s Q which expected spending on pollution control is assigned a
Intercept -0.205 0.261 -0.79 0.43 numerical value of 1 for firms with pollution control
Pollution 0.283*** 0.119 2.37 0.01
investments, and 0 for firms without this information. The
Divergence 0.006** 0.003 2.14 0.03
fourth to sixth columns in Table 7 show the empirical
Pollution 9 Divergence -0.005** 0.002 -2.20 0.02
results, and the results are consistent with prior findings,
while the result for Pollution 9 Divergence is negative and
SIZE -0.010 0.032 -0.34 0.73
significant (coefficient = -0.022, -0.008, and -0.004;
ROE -1.881*** 0.483 -3.89 \0.01
p \ 0.01, p \ 0.01, and p \ 0.05).
SALE 0.005 0.007 0.76 0.44
Finally, based on the type of moderator for ownership
GDPGW 0.004 0.010 0.46 0.64
concentration, subgroup analysis has been suggested to
PRIOR0 0.247*** 0.053 4.62 \0.01
measure the impact of a moderator variable (Sharma et al.
PRIOR0 9 Pollution 0.087 0.083 1.04 0.29
1981). In this study, the full sample is split in half, based on
F-statistic (p value) = 5.23 (p \ 0.01), adjusted R2 = 8.82 %,
N = 400
ownership concentration, in order to validate whether there
is a difference in CSP for low versus high concentration
This study follows Chung and Pruit (1994), who define Tobin’s Q as firms. The seventh to final columns in Table 7 show the
(market value of equity ? preferred stocks ? debt)/total assets. ROE
is defined as the net income to average equity. SALE is defined as a empirical results, and the results are consistent with prior
sales growth percentage from the prior 3 year’s sales. PRIOR0 is the findings, while there is negative association between pol-
difference in factor scores between the investment year and prior lution control investment and both short-term and long-
investment year t(-1). Definitions of other variables appear in term FP for high concentration firms (coefficient =
Table 3. The maximum VIF is 3.02, which is \5, and thus multicol-
linearity is not a major concern -0.221, -0.440, and -0.181; p \ 0.01, p \ 0.01, and
***, **, and * represent significance at the 1, 5, and 10 % levels using p \ 0.05). However, this study finds that there is insignifi-
the two-tailed test, respectively. The t-values based on standard errors cant association between pollution control investment and
are adjusted for heteroskedasticity (White 1980) both short- and long-run FP for low concentration firms.

123
 Effect of CSP on FP

Table 7 Additional test

Panel A: absolute value of pollution control investment Panel B: reclassification sample Panel C: split sample based on low versus high concentration firms

Dependent variable Short-run Long-run Long-run Short-run Long-run Long-run High concentration firms Low concentration firms
FP FP1 FP2 FP FP1 FP2
Short-run Long-run Long-run Short-run Long-run Long-run
FP FP1 FP2 FP FP1 FP2
Independent variable Parameter estimates

Intercept -1.147*** -1.196*** -0.647** -1.429*** -1.608*** -0.610** -1.916*** -1.830*** -1.020** -1.337** -1.892*** -1.002**
Poll_Indicator -0.008 0.016** 0.014** 0.008 0.361*** 0.256** -0.221** -0.440*** -0.181** -0.091 -0.132 -0.046
Divergence -0.004*** -0.002 -0.002** -0.004*** -0.004 -0.003***
Poll_Indicator 9 Divergence -0.002*** -0.001*** -0.001** -0.022*** -0.008*** -0.004**
MTB -0.002 -0.095*** -0.149*** 0.020 -0.107*** -0.151*** -0.053 -0.056 -0.246*** -0.030 -0.220*** -0.173***
SIZE 0.065*** 0.053*** 0.046*** 0.084*** 0.085*** 0.047*** 0.095** 0.097*** 0.073*** 0.077** 0.097*** 0.061**
DEBT 0.591*** 0.336* 0.190 0.459** 0.895*** 0.372*** 0.667** 0.700** 0.205 0.489 1.636*** 0.949***
GDPGW 0.001 0.116*** 0.118*** 0.044 0.239*** 0.136*** 0.026 0.287*** 0.343*** 0.122** 0.450*** 0.320***
PRIORi 0.008 0.001 -0.001 -0.008 0.003 -0.001 0.024 -0.159 -0.138 -0.013 0.023** -0.115
PRIORi 9 Poll_Indicator 0.003 0.005 -0.003 -0.005 -0.132** -0.052 0.153 -0.007 0.011 0.079 -0.003 -0.006
F-value 7.85 9.37 11.91 7.24 17.64 17.36 2.97 8.67 12.23 2.87 24.52 12.41
(p value) (p \ 0.01) (p \ 0.01) (p \ 0.01) (p \ 0.01) (p \ 0.01) (p \ 0.01) (p \ 0.01) (p \ 0.01) (p \ 0.01) (p \ 0.01) (p \ 0.01) (p \ 0.01)
Adj. R2 (%) 12.82 14.80 25.37 13.17 22.94 28.52 10.63 19.06 27.05 9.0 39.61 27.45
N 400 364 333 400 364 333 200 182 167 200 182 166

In Panel A, this study uses the absolute value of pollution control investment as Poll_Indicator. In Panel B, this study also includes additional 46 firm observations in which they disclose their expected spending on
pollution control records as a numerical value of 1 for firms with pollution control investments (Poll_Indicator = 1), and 0 for firms without this information (Poll_Indicator = 0). In Panel C, Poll_Indicator is an
indicator variable, which is set to a numerical value of 1 for firms with pollution control investment, and 0 otherwise. Definitions of variables appear in Table 3. The t-values based on standard errors are adjusted for
heteroskedasticity (White 1980). The maximum VIF is 4.60, which is \5, and thus multicollinearity is not a major concern
***, **, and * represent significance at the 1, 5, and 10 % levels using the two-tailed test, respectively

123
 C.-W. Peng, M.-L. Yang

Conclusions between the ultimate controlling shareholder’s voting

rights and cash flow rights. This study only includes listed
This study examines the role of the ownership concentra- firms in Taiwan, but unlisted firms may have direct and
tion on the link between CSP and FP from five highly indirect ownership links with those that are listed. There-
polluting industries in Taiwan. We demonstrate that the fore, the divergence level of ultimate ownership and con-
relationship between CSP and short-run FP is negatively trol may have been underestimated. However, the effects of
moderated by a high degree of control–cash flow diver- unmeasured variables are always unpredictable, and it is
gence. The link between CSP and long-run FP is also found possible that this limitation does not materially affect the
to be negative among firms with a high degree of control– results of this work.
cash flow divergence. As a result, the impact of ownership Our results suggest opportunities for future research.
concentration plays a critical monitoring role in the link First, this study finds a significant difference between CSP
between CSP and FP in emerging markets such as Taiwan. and FP for low versus high concentration firms. We
The results of this study have a number of important encourage future research to build rich, in-depth case
implications for business. First, the increasing focus on studies related to these findings in order to continuously
CSP has made it critical for owner interests (principals) to contribute to the practice of CSR in the business world. A
understand how specific agency problems may accomplish qualitative case study approach will enable the derivation
or deter social performance goals. Margolis and Walsh of in-depth understanding and explanations (Eisenhardt
(2003, p. 268) argued that, ‘‘we need to understand the 1989) about firm CSR activities and will contribute to a
conditions under which a corporation’s efforts benefit better understanding of how well-formed corporate gov-
society before we can understand the link between a firm’s ernance provides an assurance of addressing most CSR
social and financial performance.’’ Our findings highlight issues. For example, firms with well-formed corporate
this question by allowing us to better understand when governance programs will direct their attention toward the
agency problems caused by ownership concentration affect operation of productive and sustainable companies that
both the upsides and downsides of the social responsibil- benefit both society and their own FP as a whole.
ity–firm performance relationship. Second, our study uses ownership concentration as a
Second, our findings support the hypothesis of negative moderating variable to investigate the link between CSP
entrenchment and provide more nuanced insights concern- and FP in the Asia pacific region. In fact, different regions
ing the CSP–FP relationship. While a lot of attention has have different corporate governance features, environ-
been paid to the value creation aspect of corporate gover- mental policies, business practices, as well as their own
nance and the reduction of agency costs aspects of corporate specific cultures, etc. Thus, more studies from various
governance, limited research has actually considered how perspectives and different countries are essential to
good corporate governance systems can prevent managers enhance the understanding of the impact of corporate
from making poor social decisions. We find that firms with a governance on the link between CSP and FP. We leave this
high degree of control–cash flow divergence (i.e., the neg- important question to future research. For example, future
ative entrenchment effect) exacerbate the negative rela- researchers may include more corporate governance
tionship between CSP and FP. Thus, the results of this study mechanisms in their regions, e.g., institutional ownership,
are able to advance our understanding of managerial social outside blockholder ownership, or board characteristics. It
decisions through which expropriation of minority share- would be interesting to gain a better understanding of the
holders takes place in firms with concentrated ownership. impact of the interaction between governance mechanisms
Finally, our findings can help policy-makers and regu- and the link between CSP and FP.
lators better identify how the ownership concentration that Finally, most studies have focused on stakeholder theory
is common in emerging markets is associated with firm that assumes managers conduct CSR to fulfill their moral,
incentives to engage in pollution control. Regulatory ethical, and social duties for the benefit of their stake-
agencies can use this signal to assess, reframe and improve holders. CEO short- (e.g., salary and bonus) and long-term
CSR policies, in terms of their efficiency and effectiveness, (e.g., stock options) compensation payments rarely contain
by considering the governance that could lead to direct and explicit incentives to focus on CSR. The
enhancement of both social and financial performance. importance of CSR raises the question of whether CEOs
Some limitations of this work should be noted. The first are given incentives through pay packages to improve firm
is that this study uses only a single emerging country, CSP. For example, a long-term compensation plan will
Taiwan, to examine the role of ownership concentration on reduce CEO pressure to maximize short-term earnings,
the link between CSP–FP, and thus it may not be possible which provides a longer-term time frame wherein CSP
to generalize the results to other countries. Another limi- investments are unlikely to show up in short-term FP. We
tation of this study is the measurement of the divergence look forward to future research addressing these issues.

123
Effect of CSP on FP

Acknowledgments The authors thank two anonymous referees for Graves, S. B., & Waddock, S. A. (1994). Institutional owners and
their invaluable advice. We also thank Joëlle Vanhamme, JBE’s corporate social performance. Academy of Management Journal,
Editor, for her cordiality and encouragement through the review 37, 1034–1046.
process. Research support from the National Science Council, Taiwan Greene, W. H. (1997). Econometric analysis. New Jersey: Prentice
(grant # NCS 102-2410-H-018-040) is gratefully acknowledged. All Hall.
errors are our own. Gujarati, D. (1995). Basic econometrics (3rd ed.). New York:
McGraw-Hill.
Harris, M., & Raviv, A. (1991). The theory of financial leverage.
Journal of Finance, 46, 297–355.
References Haw, I. M., Hu, B., Hwang, L. S., & Wu, W. (2004). Ultimate
ownership, income management, and legal and extralegal
Amihud, Y., & Lev, B. (1981). Risk reduction as a managerial motive institutions. Journal of Accounting Research, 42(2), 423–462.
for conglomerate mergers. The Bell Journal of Economics, 12(2), Hillman, A. J., & Keim, G. D. (2001). Shareholder value, stakeholder
605–617. management, and social issues: What’s the bottom line?
Attig, N., Fong, W., Gadhoum, Y., & Lang, L. H. P. (2006). Effects of Strategic Management Journal, 22(2), 125–139.
large shareholding on information asymmetry and stock liquid- Jaggi, B., & Freedman, M. (1992). An examination of the impact of
ity. Journal of Banking and Finance, 30, 2875–2892. pollution performance on economic and market performance:
Bushee, B. (1998). The influence of institutional investors on myopic Pulp and paper firms. Journal of Business Finance and
R&D investment behavior. The Accounting Review, 73, Accounting, 19(5), 697–713.
305–333. Jensen, M. C. (1983). Organization theory and methodology. The
Chapple, W., & Moon, J. (2005). Corporate social responsibility in Accounting Review, 56, 319–338.
Asia: A seven country study of CSR website reporting. Business Jensen, M. C. (1986). Agency costs of free cash flow, corporate
and Society, 44, 415–441. finance, and takeovers. AEA Papers and Proceedings.
Chatterji, A. K., Levine, D. I., & Toffel, M. W. (2009). How well do Jensen, M. C., & Meckling, W. H. (1976). Theory of the firm:
social ratings actually measure corporate social responsibility? Managerial behavior, agency costs and ownership structure.
Journal of Economics and Management Strategy, 18(1), 125–169. Journal of Financial Economics, 3(4), 305–360.
Chen, K. H., & Metealf, R. W. (1980). The relationship between Jensen, M. C., & Ruback, R. S. (1983). The market for corporate
pollution control record and financial indicators revisited. The control: The scientific evidence. Journal of Financial Econom-
Accounting Review, 55, 168–177. ics, 11, 5–50.
Chien, C. C., & Peng, C. W. (2012). Does going green pay off in the Jo, H., & Harjoto, M. A. (2011). Corporate governance and firm
long-run? Journal of Business Research, 65(11), 1636–1642. value: The impact of corporate social responsibility. Journal of
Chung, K. H., & Pruit, S. W. (1994). A simple approximation of Business Ethics, 103, 351–383.
Tobin’s Q. Financial Management, 23(3), 70–74. Johnson, R. A., & Greening, D. W. (1999). The effects of corporate
Claessens, S., Djankov, S., Fan, J. P. H., & Lang, L. H. P. (2002). governance and institutional ownership types on corporate social
Disentangling the incentive and entrenchment effects of large performance. Academy of Management Journal, 42, 564–576.
shareholdings. Journal of Finance, 57(6), 2741–2771. Kochan, T. A. (2002). Addressing the crisis in confidence in
Claessens, S., Djankov, S., & Lang, L. H. P. (2000). The separation of corporations: Root causes, victims, and strategies for reform.
ownership and control in East Asian corporation. Journal of Academy of Management Executive, 16, 139–141.
Financial Economics, 58(1–2), 81–112. La Porta, R., Lopez-de-Silanes, F., & Shleifer, A. (1999). Corpo-
Core, J., Holthausen, R., & Larcker, D. (1999). Corporate governance, rate ownership around the world. Journal of Finance, 54,
chief executive officer compensation, and firm performance. 471–517.
Journal of Financial Economics, 51, 371–406. La Porta, R., Lopez-de-Silanes, F., Shleifer, A., & Vishny, R. (2002).
Deckop, J. R., Merriman, K. K., & Gupta, S. (2006). The effects of Investor protection and corporate valuation. Journal of Finance,
CEO pay structure on corporate social performance. Journal of 57, 1147–1170.
Management, 32(3), 329–342. Loughran, T., & Ritter, J. R. (1995). The new issues puzzle. Journal
Eisenhardt, K. M. (1989). Building theories from case study research. of Finance, 50(1), 23–51.
Academy of Management Review, 14(4), 532–550. Margolis, J. D., & Walsh, J. P. (2003). Misery loves companies:
Falck, O., & Heblich, S. (2007). Corporate social responsibility: Rethinking social initiatives by business. Administrative Science
Doing well by doing good. Business Horizons, 50, 247–254. Quarterly, 48, 268–305.
Fama, E. F., & French, K. R. (1992). The cross section of expected Maron, I. Y. (2006). Toward a unified theory of the CSP–CFP link.
stock returns. Journal of Finance, 47, 427–466. Journal of Business Ethics, 67(2), 191–200.
Fama, E., & Jensen, M. (1983). The separation of ownership and Matten, D., & Moon, J. (2008). ‘‘Implicit’’ and ‘‘Explicit’’ CSR: A
control. Journal of Law and Economics, 26, 301–325. conceptual framework for a comparative understanding of
Fan, J. P. H., & Wong, T. J. (2002). Corporate ownership structure corporate social responsibility. Academy of Management Review,
and the informativeness of accounting earnings. Journal of 33, 404–424.
Accounting and Economics, 33, 401–425. McGuire, J., Dow, S., & Archgeyd, K. (2003). CEO incentives and
Francis, J., Khurana, I., & Pereira, R. (2003). The role of accounting corporate social performance. Journal of Business Ethics, 45(4),
and auditing in corporate governance and the development of 341–359.
financial markets around the world. Asia Pacific Journal of McGuire, J., Sundgren, A., & Schneeweis, T. (1988). Corporate social
Accounting Economics, 10, 1–30. responsibility and firm financial performance. Academy of
Friedman, M. (1970). A Friedman doctrine: The social responsibility Management Journal, 31(4), 854–972.
of business is to increase its profits. New York Times Magazine, McWilliams, A., Siegel, D., & Wright, P. (2006). Corporate social
13, 33. responsibility: Strategic implications. Journal of Management
Gollop, F., & Roberts, M. J. (1983). Environmental regulation and Studies, 43, 1–18.
productivity growth: The case of fossil-fuel electric power Megginson, W. L., Nash, R. C., & Randenborgh, M. V. (1994). The
generation. Journal of Political Economy, 91(4), 654–674. financial and operating performance of newly privatized firms:

123
 C.-W. Peng, M.-L. Yang

An international empirical analysis. Journal of Finance, 49(2), Simpson, W. G., & Kohers, T. (2002). The link between social and
403–452. financial performance: Evidence from the banking industry.
Moore, G. M. (2001). Corporate social performance: An investigation Journal of Business Ethics, 35(2), 97–109.
in the U.K. supermarket industry. Journal of Business Ethics, Smith, J. B., & Sims, W. A. (1985). The impact of pollution charges
34(3–4), 299–315. on productivity growth in Canadian brewing. Rand Journal of
Orlitzky, M., Schmidt, F. L., & Rynes, S. L. (2003). Corporate social Economics, 16(3), 410–423.
and financial performance: A meta-analysis. Organization Stud- Soana, M. A. (2011). The relationship between corporate social
ies, 24, 403–441. performance and corporate financial performance in the banking
Orlitzky, M., & Swanson, D. L. (2002). Value attunement: Toward a sector. Journal of Business Ethics, 104, 133–148.
theory of socially responsible executive decision making. Spicer, B. H. (1978). Investors, corporate social performance and
Australian Journal of Management, 27, 119–128. information disclosure: An empirical study. The Accounting
Seifert, B., Morris, S. A., & Bartkus, B. R. (2003). Comparing big Review, 53(1), 94–111.
givers and small givers: Financial correlates of corporate Turker, D. (2009). Measuring corporate social responsibility: A scale
philanthropy. Journal of Business Ethics, 45(3), 195–211. development study. Journal of Business Ethics, 85(4), 411–427.
Servaes, H., & Tamayo, A. (2013). The impact of corporate social White, H. (1980). A heteroskedasticity-consistent covariance matrix
responsibility on firm value: The role of customer awareness. estimator and a direct test for heteroscedasticity. Econometrica,
Management Science, 59(5), 1045–1061. 48, 817–838.
Sharma, S., Durand, R. M., & Gur-Arie, O. (1981). Identification and Wu, M. L. (2006). Corporate social performance, corporate financial
analysis of moderator variables. Journal of Marketing Research, performance, and firm size: A meta-analysis. Journal of
18(3), 291–300. American Academy of Business, 8(1), 163–171.
Shleifer, A., & Vishny, R. (1997). A survey of corporate governance. Yeh, Y. H., Lee, T. S., & Woidtke, T. (2001). Family control and
Journal of Finance, 52(2), 737–783. corporate governance: Evidence from Taiwan. International
Short, J. (2004). Are we wasting time with the corporate social Review of Finance, 2(1&2), 21–48.
performance–financial performance link? Paper presented at the
national meetings of the Academy of Management, New
Orleans, LA.

123