DOI: 10.1111/1475-679X.12174
Journal of Accounting Research
Vol. 55 No. 5 December 2017
Printed in U.S.A.
Does Corporate Social
Responsibility (CSR) Create
Shareholder Value? Evidence from
the Indian Companies Act 2013
H A R I O M M A N C H I R A J U∗ A N D S H I V A R A M R A J G O P A L†
Received 6 October 2015; accepted 4 March 2017
ABSTRACT
In 2013, a new law required Indian firms, which satisfy certain profitability,
net worth, and size thresholds, to spend at least 2% of their net income on
corporate social responsibility (CSR). We exploit this regulatory change to
isolate the shareholder value implications of CSR activities. Using an event
study approach coupled with a regression discontinuity design, we find that
the law, on average, caused a 4.1% drop in the stock price of firms forced
to spend money on CSR. However, firms that spend more on advertising
are not negatively affected by the mandatory CSR rule. These results suggest
that firms voluntarily choose CSR to maximize shareholder value. Therefore,
∗ Department of Accounting, Indian School of Business; † Roy Bernard Kester and T.W.
Byrnes Professor of Accounting and Auditing, Columbia Business School.
Accepted by Douglas Skinner. We thank an anonymous referee, Yakov Amihud, Bernie
Black, Sanjay Kallapur, Simi Kedia, Yongtae Kim, Bill Kross, Stephanie Larocque (discussant),
Suresh Radhakrishnan, Srini Rangan, Sri Sridharan, K.R.Subramanyam, Stephan Zeume (discussant), Hao Zhang, and workshop participants at 2015 MIT Asia Conference in Accounting, 2015 AAA conference and the Indian School of Business for helpful comments. We also
thank Ranjan Nayak and Aadhaar Verma for providing excellent research assistance. Finally,
we thank our respective schools for financial support.
1257
C , University of Chicago on behalf of the Accounting Research Center, 2017
Copyright 1258
H. MANCHIRAJU AND S. RAJGOPAL
forcing a firm to spend on CSR is likely to be sub-optimal for the firm with a
consequent negative impact on shareholder value.
JEL codes: M14; M40; M41; M48
Keywords: CSR; shareholder value; Indian Companies Act 2013
1. Introduction
Spending on certain forms of corporate social responsibility (CSR) is now
mandatory in India.1 Clause 135 of the Companies Act (mandatory CSR
rule, hereafter) passed by the Indian Parliament in 2013 requires a firm,
which meets certain thresholds in any fiscal year, to spend 2% of its average
net profits of the prior three years on CSR activities.2 A legislative mandate
forcing corporations to spend funds on CSR activities is perhaps the first of
its kind in the world. We exploit this unique setting to isolate the impact of
CSR on shareholder value.
There are two competing views on whether CSR affects shareholder
value. The “shareholder expense” view, advocated most notably by Milton
Friedman [1970], asserts that “the social responsibility of business is to increase its profits.” He goes on to argue that “if businessmen do have a social responsibility other than making maximum profits for stockholders,
how are they to know what it is? Can self-selected private individuals decide
what the social interest is?” Under Friedman’s framework, CSR activities of
a firm are typically viewed as a manifestation of moral hazard toward the
shareholder. The contrasting view, labeled here as the “stakeholder value
maximization” view, following the “doing well by doing good” theory advanced in the management literature, argues that strategic CSR spending
can increase firm value. The intuition is that a firm’s self-interested focus
on stakeholders’ interests increases their willingness to support the firm’s
operations in several ways (Kitzmueller and Shimshack [2012]).
Existing empirical evidence on whether CSR creates shareholder value
is inconclusive, partly because these studies are clouded by methodological concerns such as potential endogeneity, reverse causality, or omitted
variable problems (Margolis, Elfenbein, and Walsh [2009]).3 The choice
to conduct CSR is voluntary. Assuming firms spend their optimal level of
1 While the term CSR itself is not defined in the Indian Companies Act 2013, the Act does
provide a list of activities that it considers to be CSR. We provide these details in section 3.1.
2 As per the mandatory CSR rule, if during any fiscal year a firm has either (1) a net worth
of Indian Rupees (INR) 5,000 million (about U.S. $83 million) or more; (2) sales of INR
10,000 million (about U.S. $167 million) or more; or (3) a net profit of INR 50 million (about
U.S. $0.83 million) or more, it is required to spend 2% of its average net profits of the last
three years on CSR related activities. An exchange rate of 60INR = 1US$ is assumed for the
conversion of INR to US$.
3 For reviews of the literature on CSR, see Griffin and Mahon [1997], Orlitzky and Benjamin
[2001], Margolis and Walsh [2003], Orlitzky, Schmidt, and Rynes [2003], Margolis, Elfenbein,
and Walsh [2009], and Kitzmueller and Shimshack [2012].
IMPACT OF CSR ON SHAREHOLDER VALUE
1259
CSR, on average, there ought to be no association between future firm
performance and CSR in the cross-section (Himmelberg, Hubbard, and
Palia [1999]). Hence, it is difficult to ascertain whether the observed associations between CSR and firm performance (i) are causal in nature;
or (ii) are merely attributable to model misspecification due to the influence of unobserved firm level heterogeneity related to CSR (Himmelberg,
Hubbard, and Palia [1999]). Second, as highlighted by Hong, Kubik, and
Scheinkman [2012], reverse causality can drive the results as firms that are
doing well, and are hence less financially constrained, are more likely to
spend resources on CSR activities. Hence, firm performance could cause
higher future CSR, as opposed to the other way around. Third, as Lys,
Naughton, and Wang [2015] suggest, CSR could merely signal and not
cause future profits. In sum, the correlation between CSR and firm value or
firm performance, documented by earlier studies, albeit interesting, does
not necessarily warrant a causal interpretation.
To partially overcome these inferential problems, we exploit numerical
thresholds specified in the mandatory CSR rule and employ a regression
discontinuity design (RDD) to identify a potentially causal link between
CSR and firm value.4 An RDD typically compares the outcomes just above
and just below a discontinuous threshold, and attributes any differences
in the outcome variable to the intervention that creates the discontinuity,
assuming that but for the intervention, firms above and below the threshold are similar.5 Because the intervention is exogenously imposed, firms
have no control as to whether they will receive the intervention or not.
Hence, there is an equal probability that the firm might be assigned on either side of the threshold, thereby making RDD a close approximation of
a randomized experiment near the threshold (Lee and Lemieux [2010]).
However, an important limitation of RDD is that it captures treatment effects localized around the threshold and hence the results are not always
generalizable to the entire population.
In this paper, the discontinuity arises because, around the INR 50 million profit threshold exogenously determined by the mandatory CSR rule,
a minor change in net income leads to a discrete change (i.e., a discontinuity) in the application of CSR rule. Such discontinuity classifies firms as
AFFECTED (those who are required to comply with CSR rule) and UNAFFECTED (those who are not required to comply with CSR rule). Intuitively,
there is no reason to expect systematic differences in a firm with a net income of INR 51 million and another with a net income of INR 49 million.
Thus, any differences in the cumulative abnormal returns (CAR) around
4 Gow, Larcker, and Reiss [2016] document an increasing trend in the use of quasiexperimental methods to draw causal inference in accounting research and believe that papers using these methods are considered stronger research contributions.
5 See Imbens and Lemieux [2008], Lee and Lemieux [2010], and Robert and Whited
[2012] for recent overviews on the RDD methodology and surveys of RDD applications in
the economics and finance literature.
1260
H. MANCHIRAJU AND S. RAJGOPAL
the significant CSR event dates (the outcome variable in our setting) for
firms that are just above the threshold and have to comply with the CSR
requirement and those that are just below the cutoff and did not have to
comply, can be attributed to the CSR rule. Hence, our setting enables us to
draw more reliable causal inferences on the impact of CSR on firm value.
Ex ante, it is difficult to predict the impact of the mandatory CSR rule on
shareholder value. In situations where firms’ CSR activities are not aligned
with their shareholder’s interests (as suggested by Friedman [1970], Tirole
[2001], Benabou and Tirole [2010], and Cheng, Hong, and Shue [2013]),
the new mandatory CSR rule, via its increased reporting and governance
requirements, will likely force firms to redirect their CSR spending to maximize firm value, leading thus to an increase in the shareholder value
of firms affected by the rule. For instance, the increased scrutiny under
mandatory CSR can force managers to cultivate relations with stakeholders
that would pay off in the long run. However, if firms already conduct CSR
to maximize their firm value before the law was passed, imposing binding
legal constraints on their CSR choices will lead to declines in their shareholder values (Demsetz and Lehn [1985]). For example, if at equilibrium,
a firm’s marginal benefits of CSR are lower than the marginal costs of CSR,
then the firm will not spend on CSR. In such a case, requiring such a firm to
spend 2% of its profits on CSR will only lead to a reduction of shareholder
value.
To test these predictions, we examine CAR of AFFECTED and UNAFFECTED firms around eight important event dates underlying the legislative passage of the mandatory CSR rule. Our main finding is that around
these significant events, CAR for the AFFECTED firms is lower than the CAR
for UNAFFECTED firms, suggesting an economically significant negative relation between CSR and shareholder value. We find that, on average, firms
that are forced to spend money on CSR experience a 4.1% drop in the stock
price around the eight events. Given that firms are required to spend 2% of
their profits on CSR, to the extent CSR-related activities are negative NPV
projects, the passage of this rule is likely to result in a 2% decline in shareholder value—corresponding directly to cash outflows of CSR activities. Although the 4.1% decline in shareholder value is not statistically different
from 2%, interviews with CFOs of some of our sample firms suggest that the
decline in shareholder value beyond 2% is plausible on account of factors
such as the stock market’s perception of increased Governmental interference in business, increased compliance costs associated with reporting and
monitoring CSR activities, and/or the diversion of scarce managerial time
and effort to CSR activities that do not add to shareholder value.6
We then perform three cross-sectional analyses to identify conditions under which the mandatory CSR rule is likely to affect firm value differentially.
6 We hasten to add that we cannot empirically disentangle the impact of these conjectures
on shareholder value.
IMPACT OF CSR ON SHAREHOLDER VALUE
1261
These predictions are motivated by the “stakeholder value maximization”
perspective of the management literature. Specifically, we examine whether
the impact of CSR on shareholder value varies depending on the firm’s political connections, its advertisement spending, and its affiliation to a highly
polluting industry. We find that the CAR for AFFECTED firms is less negative
if such firms advertise. This is consistent with Servaes and Tamayo [2013]
who argue that CSR and firm value are positively related for firms with high
consumer awareness, as proxied by advertising expenditure. However, we
find no consistent results for cross-sectional variation in stock price reactions to the mandatory CSR rule for AFFECTED and UNAFFECTED firms
when we partition firms on political connectedness or on their membership in a polluting industry.
We also examine whether the mandatory CSR rule impacts long-term
value of the AFFECTED firms using Tobin’s Q as a measure of firm value. If
investors consider CSR activities to be detrimental to firm value, then the
Q of the affected firms should decline more, relative to that of unaffected
firms, in the years when the likelihood of the passage of mandatory CSR
rule increased. Consistent with our expectations, we find a greater decline
in the Q of the AFFECTED firms in the years 2011 and 2013 relative to the
Q of the benchmark UNAFFECTED firms.
Overall, our results suggest that, on average, the mandatory CSR rule imposes significant net costs on firms that are required to comply with this
regulation, leading to declines in shareholder value. We perform several
additional tests to check the sensitivity of our results. First we test the basic
assumption of RRD that firms should not be able to manipulate the rating variable and thereby select themselves in treatment and control groups
(Lee and Lemieux [2010]). For this purpose, we examine the frequency
distribution of firms around the thresholds and find no abnormal jumps.
These results suggest that our RDD analysis is not affected by the possibility that potential treatment firms might have managed their profits/book
value/sales to fall right below the cutoff to escape having to comply the
law. Further, our results are robust to both nonparametric and parametric
techniques of estimating the treatment effect.
We make several contributions to the literature. First, our paper establishes a potentially causal link between CSR and firm value. Establishing
causality between CSR and firm value is difficult due to potential confounding factors that may also contribute to a firm’s decision to engage in CSR.
To overcome this hurdle, we use an RDD around regulatory change which
forced certain firms to spend on CSR, as a mechanism to examine whether
CSR indeed affects shareholder value. This research setting significantly
mitigates endogeneity concerns and makes it more likely that changes in
firm value around the event can be attributed to CSR.
Second, we document a negative relation between mandatory CSR and
shareholder value. Our finding suggests that if firms choose their CSR
spending optimally before the law was imposed, there ought to be no relation between voluntary CSR and firm value. Forcing these firms to spend
1262
H. MANCHIRAJU AND S. RAJGOPAL
on CSR therefore has a negative impact on shareholder value. A related
implication of our paper is that studies that find an association between
corporate outcomes (e.g., higher operating performance, lower cost of capital, or lower earnings management) and voluntary CSR spending need to
grapple with the possibility that such associations are confounded by omitted variables, consistent with Himmelberg, Hubbard, and Palia [1999].
Third, a majority of prior studies consider voluntary CSR. To the best of
our knowledge, we are the first to examine the impact of legislatively mandated CSR spending. Our results demonstrate that forcing firms to spend
on CSR, on average, is likely to lead to negative consequences for shareholders. Finally, our study complements prior research that examines the
effect of firm-specific characteristics on the relationship between CSR and
shareholder value (e.g., Servaes and Tamayo [2013], Di Giuli and Kostovetsky [2014]). We show that firms with greater advertisement spending are
less likely to be negatively affected by the mandatory CSR rule.
Like any study that exploits a legislative change set in a particular institutional context, our paper also suffers from certain limitations relating
to generalizability. It can be argued that the mandatory CSR activities prescribed by the Indian Companies Act 2013 are different in nature from
the voluntary CSR seen in the rest of the world. Although the findings are
relevant for India, it is unclear whether these findings can be applied to
other parts of the world which have different economic and regulatory environments. We submit that, despite these generalizability limitations, the
key findings of our paper have implications for efforts to encourage CSR
in other economies. In particular, our study suggests that giving the firm
flexibility to define what CSR means and letting it choose where it wants to
direct its CSR spending is preferable from the shareholder’s perspective.
2. Related Literature and Hypotheses
Does CSR affect firm value? The existing theoretical literature on this
question can be categorized under two opposing views: (i) shareholder expense view; and (ii) stakeholder value maximization view. The shareholder
expense view follows Milton Friedman’s [1970] assertion that “the social responsibility of business is to increase its profits.” Friedman argues that CSR
involves managers spending shareholders’ money and “in effect imposing
taxes, on one hand, and deciding how the tax proceeds shall be spent, on
the other.” He considers CSR to be a drain on the firm’s valuable resources
that should be utilized for shareholder value maximization.
In contrast to the shareholder expense view, Freeman’s [1984] stakeholder theory argues that a firm should consider the interests of everyone
who substantially affects (or is affected by) the welfare of the firm. Thus,
social, environmental or ethical preferences of stakeholders can induce
CSR activities (Baron [2001], McWilliams and Siegal [2001]). Such strategically motivated CSR activities can be profitable and the management
IMPACT OF CSR ON SHAREHOLDER VALUE
1263
literature terms this thesis as “doing well by doing good.” 7 Following this
stakeholder view, studies have documented that a high commitment to CSR
activities is associated with attracting and retaining higher quality employees (Greening and Turban [2000]), improving the effectiveness of the marketing of products and services (Fombrun [2005]), increasing demand for
products and services (Navarro [1988]), providing superior access to valuable resources (Cochran and Wood [1984]), generating moral capital or
goodwill that tempers punitive actions by regulatory agencies during a negative event (i.e., an insurance effect) and thereby preserving or increasing
firm value (Godfrey [2005]).
The economics-based reasoning for profitable CSR is that these activities
reduce transaction costs with stakeholders and provide net benefits to the
firm. The theory of the firm, as advanced by Coase [1937] and expanded by
Jensen and Meckling [1976], among others, views a firm as a nexus of contracts (both explicit and implicit) between shareholders and other stakeholders in which each group of stakeholders supplies the firm with critical
resources. Firms that undertake CSR activities tend to develop a reputation for keeping their commitments associated with the implicit contracts.
Consequently, stakeholders of these firms are more likely to contribute resources and effort to the firm and accept less favorable explicit contracts
(compared to stakeholders of firms with no or low levels of CSR activities).
The empirical evidence on the direct effect of CSR on firm value is
mixed. Margolis, Elfenbein, and Walsh [2009], in their influential metaanalysis of roughly 167 studies, find that some studies document a positive
effect when they regress firms’ financial performance (either accountingbased ROA or stock returns) on corporate goodness while others find a
negative effect. The average effect across these studies is a small positive
increase in firm performance. More recently, Dhaliwal, Li, Tsang, and Yang
[2011] conclude that the voluntary disclosure of CSR activities (i) attracts
institutional investors and analysts; and (ii) reduces the firm’s cost of capital. Servaes and Tamayo [2013] find that CSR activities and firm value are
positively related for firms with high customer awareness. In contrast, Di
Giuli and Kostovetsky [2014] find that Democrat-leaning firms are associated with more CSR activities than Republican-leaning firms, and increases
in firms’ CSR ratings are associated with negative future stock returns and
declines in ROA. This finding suggests that CSR activities that are motivated
by the political affiliation of stakeholders come at the expense of firm value.
While the discussion so far focuses on how voluntary CSR can be costly
or beneficial to shareholders, we have to also consider the implications
of mandatory CSR in India, as imposed by the Companies Act. There are
several reasons why mandatory CSR activities and their disclosure may not
benefit, and might even harm, shareholders. If Friedman’s view of CSR is
7 Benabou and Tirole [2010], Margolis, Elfenbein, and Walsh [2009], and Kitzmueller and
Shimshack [2012] are some papers that review the literature on “doing well by doing good.”
1264
H. MANCHIRAJU AND S. RAJGOPAL
indeed descriptively valid and firms optimally choose not to spend funds
on CSR, imposing binding legal constraints on their CSR choices will lead
to declines in their shareholder values. Once CSR spending and its reporting become mandatory, the Government could start prescribing how the
CSR money should be spent, thereby limiting a firm’s flexibility in coming
up with its CSR policies. Moreover, various interest groups may find it easier to lobby management to advance their environmental and social goals.
Finally, mandatory CSR also comes with compliance obligations such as administrative costs associated with reporting information and the need for
the board to monitor the firm’s CSR activities.
While a new regulation could often impose net costs on shareholders,
there can be circumstances where the regulation could be beneficial.8 In
the pre-mandatory CSR period, managers potentially expend a sub-optimal
level of CSR because of pressures to achieve short-term earnings targets.
However, under mandatory CSR, managers may be nudged to cultivate relations with various stakeholders that might pay off in the long run. Moreover, managers’ CSR activities could reflect their own social preferences. Increased reporting and governance requirements imposed by the law might
make firms redirect their CSR spending to maximize firm value. Finally, as
discussed earlier, firms could use CSR activities to signal their commitment
toward their implicit contracts. Mandatory CSR requires a firm to make its
CSR policies more formalized and visible, which, in turn, could strengthen
such a commitment. Hence, the terms of the firms’ explicit contracts could
become more favorable to the firm.
In summary, there are several ways in which CSR can either have a positive or negative impact on firm value. Based on this discussion our first
hypothesis is:
H1: The mandatory CSR rule affects shareholder value.
We go on to hypothesize conditions under which mandatory CSR rule is
likely to affect firm value differentially. Motivated by prior work on CSR, we
consider whether the impact of CSR on shareholder value varies depending on the firm’s political connections, its advertisement spending, and its
affiliation to a highly polluting industry.
First, we consider the impact of a firm’s political connectedness on the
relation between CSR and firm value. Faccio [2006] suggests that political connections can be of value to a firm in several ways including preferential treatment by government-owned enterprises (such as banks or raw
material producers), lower taxation, preferential treatment in competition
for government contracts, relaxed regulatory oversight of the company in
8 For example, Chhaochharia and Grinstein [2007] find that certain governance-related
provisions of Sarbanes Oxley Act led to increase in firm value. Black and Khanna [2007] find
that the governance reforms (Clause 49) introduced in India are associated with positive stock
market reaction. Similarly, Black and Kim [2012] find that the 1999 Korean governance reforms also led to increase in firm value for the affected firms.
IMPACT OF CSR ON SHAREHOLDER VALUE
1265
question, or stiffer regulatory oversight of its rivals. However, as pointed by
Shleifer and Vishny [1994], politicians themselves will extract at least some
of the rents generated by connections, and firm value will be enhanced only
when the marginal benefits of the connections outweigh their marginal
costs. CSR spending might constitute a potential mechanism via which a
firm can satisfy the preferences of politicians and increase its ability to conduct business with the government and other entities with their political
connections.
Political connections assume greater importance in the context of India where government regulation is high, corruption is accepted as a
ground reality, and enforcement mechanisms are unpredictable (Khanna
and Palepu [2000]). Consequently, CSR activities can become an important
part of corporate strategy to enhance these political ties. For example, in his
inaugural Independence Day speech, the Prime Minister of India recently
urged corporations to take up building toilets in schools as a priority under
their CSR policies. Within four days, companies announced over INR 200
crore (US$ 33.33 million) of contributions to the government’s “Swachh
Bharat” (Clean India) campaign.9 To the extent firms can use CSR as a
mechanism to enhance their political ties, CSR can be valuable. Therefore,
we hypothesize that:
H2a: The negative (positive) effect of mandatory CSR rule on shareholder
value is lower (higher) among firms that are politically connected.
The marketing and management literature suggest that CSR activities
of a firm can attract customers. Fisman, Heal, and Nair [2008], in their
model, argue that consumers realize that only firms that care about product
quality are willing to invest in CSR activities because purely profit-oriented
firms (that can compromise on quality) find these investments to be “too
expensive.” By engaging in CSR, firms are able to identify themselves as
the ones selling better quality products. Further, socially responsible consumers (e.g., “green” consumers) are more likely to buy and some are even
willing to pay more for products of CSR firms (Navarro [1998], Sen and
Bhattacharya [2001]). Servaes and Tamayo [2013] find that CSR and firm
value are positively related for firms with high consumer awareness, as proxied by advertising expenditure. They argue that a necessary condition for
CSR to modify consumer behavior and hence affect firm value is consumer
awareness of the firm’s CSR activities. Advertising expenditure increases the
potential customers’ awareness of the firm, its products, and practices (including its CSR activities). This, in turn, is likely to increase the chances of
a consumer identifying with the product, thereby enhancing revenue and
eventually firm value. Following this logic, we hypothesize that:
9 http://timesofindia.indiatimes.com/india/Modis-Swachh-Bharat-call-gets-Rs-200-crorefrom-TCS-Bharti/articleshow/40384230.cms
1266
H. MANCHIRAJU AND S. RAJGOPAL
H2b: The negative (positive) effect of mandatory CSR rule on shareholder
value is lower (higher) among firms with high levels of advertisement
expenditure.
Finally, we investigate whether the negative effect associated with the
mandatory CSR rule is relatively strong among firms in polluting industries. Once CSR becomes mandatory, environmental activist groups and local communities can pressure these firms to spend money on green technologies or environmental controls. While these investments can benefit
the overall environment, they typically tend to yield limited firm-specific
benefits. Hence, we hypothesize that:
H2c: The negative (positive) effect of mandatory CSR rule on shareholder
value is higher (lower) among firms that are in high polluting industries.
3. Research Design and Data
3.1
BACKGROUND OF THE COMPANIES ACT
2013
The Indian Companies Act, 2013 was enacted on August 29, 2013. This
legislation attempts to overhaul a nearly 60-year-old Indian corporate law
framework to bring it in line with global best practices. A unique feature
of this Act is clause 135 (referred to as the CSR rule in this paper) which
specifies that if in any given year, a firm has either (1) a net worth of Indian
Rupees (INR) 5,000 million (about U.S. $83 million) or more; (2) sales of
INR 10,000 million (about U.S. $167 million) or more; or (3) a net profit
of INR 50 million (about U.S. $0.83 million), it is required to spend 2% of
its average net profits of the last three years on CSR activities.
For example, the rule implies that if in the year 2012, a firm that meets
any of the three criteria has to spend 2% of profits averaged over the three
years, 2010, 2011, and 2012. Similar to Iliev [2010], in the context of section 404 of the Sarbanes Oxley Act, these numerical thresholds help us
disentangle the contribution of the CSR rule from other requirements of
the Companies Act 2013 (e.g., enhanced corporate governance and disclosure norms, greater accountability of management and auditors; stricter
enforcement, protection for minority shareholders, etc.) that apply to all
listed firms, regardless of numerical cutoffs.
The passage of Companies Act 2013 is a culmination of several years of
effort. We focus on eight event dates that represent major milestones in the
legislative passage of the Act. Figure 1 summarizes the timeline of events.
These events are as follows:
r Event # 1: The Companies Act was first introduced in the Lok Sabha
(the lower house of the Indian parliament) on August 3, 2009 as the
Companies Bill, 2009 (Bill, hereafter). The Bill in its original form had
no clause on CSR.
r Event # 2: The 2009 version of the Bill was referred to the Parliamentary Standing Committee on Finance, which submitted its report on
IMPACT OF CSR ON SHAREHOLDER VALUE
Event # 1
Aug 3, 2009
Event # 2
Aug 31, 2010
Event # 3
Feb 28, 2011
Event # 4
Dec 14, 2011
Introduction
of the
Company’s
Bill 2009 in
the Lok
Sabha
For the first
time,
Standing
Committee
on Finance
inserts the
notion of
Mandatory
CSR in its
report
Bowing to
corporate
pressure, the
Ministry of
Corporate
Affairs
announces
that it is
considering
making CSR
voluntary
and not
mandatory
Government
eventually
goes ahead
with the
mandatory
CSR and
introduces a
revised
Company’s
Bill in the
Lok Sabha
Event # 5
Jun 26, 2012
Standing
Committee
on Finance
endorses the
bill with no
new changes
relating to
mandatory
CSR
1267
Event # 6
Dec 18, 2012
Event # 7
Aug 8, 2013
Event # 1
Aug 29, 2013
The
Companies
Bill 2012,
containing
mandatory
CSR clause,
is passed in
the Lok
Sabha
The
Companies
Bill 2012 is
passed in
the Rajya
Sabha
The
President
signs the
bill and it
is enacted
FIG. 1.—Timeline of important events related to the passage of the mandatory CSR rule.
August 31, 2010. The Finance Committee introduced the notion of
mandatory CSR for the first time in its report dated August 31, 2010.
Anecdotal evidence and reports in the popular press suggest that the
Finance Committee, perhaps anticipating popular backlash that might
result from a very progressive pro-business bill, inserted several new
clauses to make the bill slightly more pro-development. The proposal
mandating CSR spending was among these clauses and noted that:
“every company having [(net worth of rupees 500 crore or more, or
turnover of rupees 1000 crore or more)] or [a net profit of rupees 5 crore
or more during a year] shall be required to formulate a CSR Policy to ensure
that every year at least 2% of its average net profits during the three immediately preceding financial years shall be spent on CSR activities as may
be approved and specified by the company.”10
r Event # 3: On account of protests from industry, the Ministry of Corporate Affairs announced on February 28, 2011 that it is considering
making only the disclosure, as opposed to the actual spending, of CSR
mandatory. The proposal for mandatory CSR attracted a lot of criticism from companies who argued that what they spend on community
welfare, education, health, development, and environmental activism
is for them to decide. Azim Premji, Chairman of Wipro Ltd., opposed
mandatory CSR and argued that “my worry is the stipulation should
not become a tax at a later stage ... Spending two per cent on CSR is a
lot, especially for companies that are trying to scale up in these difficult
times. It must not be imposed.” He also felt that a distinction should
be made between personal philanthropy and CSR, which is a company
activity.11 Murli Deora, the Minister of Corporate Affairs (the ministry
10 From a research design standpoint, as discussed later, it is important to note that these
numerical cutoffs introduced on August 31, 2010, remained unchanged through various iterations of the bill until the bill became law.
11 http://businesstoday.intoday.in/story/azim-premji-aima-convention-corporate-socialresponsibility/1/198960.html
1268
r
r
r
r
r
H. MANCHIRAJU AND S. RAJGOPAL
that crafted the Companies Bill, 2009), also acknowledged that there
was an argument as to whether the Government should mandate anything.
Event # 4: The Ministry of Corporate Affairs eventually resisted pressure from corporate houses and went ahead with the mandatory CSR
rule.12 Keeping in view the recommendations made by the Finance
Committee, a revised Bill was prepared and the original Bill was withdrawn. The new Bill was introduced in the Lok Sabha on December 14,
2011. The Bill was again referred to the Finance Committee as certain
new provisions, which had not been referred earlier to the committee,
were included in this new Bill.
Event # 5: The Finance Committee submitted its report on June 26,
2012.
Event # 6: The Lok Sabha subsequently approved the Bill on December
18, 2012 and labeled it as the Companies Bill, 2012.
Event # 7: The Companies Bill, 2012 was then considered and approved by the Rajya Sabha (the upper house of the Indian parliament)
on August 8, 2013 as The Companies Bill, 2013.
Event # 8: The Companies Bill, 2013 received the President’s assent on
August 29, 2013 and has now become law.
Under the hypothesis that CSR is value-destroying, we expect firms affected by the mandatory CSR rule to experience (i) negative stock market
reactions on Event # 2, 4, 5, 6, 7, and 8 because each of these events is associated with an increased probability that the mandatory CSR rule will pass;
and (ii) a positive stock market reaction on Event # 3 because it mitigates
any possible harmful effects of the mandatory CSR rule by decreasing the
probability that the legislation, in its proposed form, will pass.
While the term “CSR” itself is not defined in the Act, Schedule VII of the
Companies Act, quoted below, requires CSR activities of the firm to focus
on at least one of the following areas: (i) eradicating extreme hunger and
poverty; (ii) promotion of education; (iii) promoting gender equality and
empowering women; (iv) reducing child mortality and improving maternal
health; (v) combating HIV, AIDS, malaria and other diseases; (vi) ensuring
environmental sustainability; (vii) employment-enhancing vocational skills;
(viii) social business projects; (ix) contribution to the Prime Minister’s National Relief Fund or any other fund set up by the Central Government
or the state governments for socioeconomic development, and relief and
funds for the welfare of the scheduled castes, the scheduled tribes, other
backward classes, minorities and women; and (x) such other matters as may
be prescribed.13
12 Such resistance also suggests that reverse causality (from firm value to the implementation of the law) is highly unlikely.
13 Descriptive data on the magnitude and nature of CSR spending for one year, 2012, are
presented in panels B and C of table 1.
IMPACT OF CSR ON SHAREHOLDER VALUE
1269
The CSR rule also provides guidance on the enforcement mechanism
needed to achieve the CSR goals. Specifically, it requires that a firm create a CSR committee consisting of three or more directors, at least one of
which must be an independent director. The CSR committee is expected
to devise, recommend, and monitor CSR activities, and the amounts spent
on such activities. The new rule also requires that a firm must (i) publicly
disclose an official policy on its CSR activities and document CSR activities
implemented during the year in its annual report; and (ii) demonstrate a
preference for local areas where they operate. While a company is not subject to liability for failing to spend on CSR, a company and its officers are
subject to liability for not explaining such a failure in the annual report of
the board of directors.14
Overall, the mandatory CSR rule of the Companies Act, 2013 is a unique
regulation as it may be the first time in the world that a country has mandated expenditures for the public good, rather than simply tax companies
or leave them to conduct CSR activities on their own. According to Ernst &
Young, these provisions would generate over U.S. $2.5 billion of CSR spending annually. India is home to the largest concentration of poverty on the
planet and can certainly benefit from this large inflow of funds to socially
responsible activities. However, it is unclear whether such mandatory CSR
improves or hurts shareholder value.
3.2
EVENT STUDY APPROACH AND ASSUMPTIONS
To capture the effect of the CSR rule on firm value, we compare the
CAR of AFFECTED firms to those of UNAFFECTED firms (controlling for
other firm attributes that are likely to affect returns) around each of the
eight key legislative events related with the passage of the Companies Act
2013, as outlined in section 3.1. On each event date, we classify a firm as
AFFECTED or UNAFFECTED based on whether its profit/sales/book value
for the most recent fiscal year end meets the thresholds mentioned in the
CSR rule or not. For example, a firm whose profit equals INR 50 million
for the fiscal year ended March 31, 2010 would be classified as AFFECTED
when we consider the event date August 31, 2010. We measure abnormal
returns by estimating the market model using 200 trading days of return
data ending 11 days before the legislative event. The return on CNX 500
index is used as a proxy for market return.15 Daily abnormal stock returns
are cumulated to obtain the CAR from day t−1 before the legislative event
date to day t + 1 after the event date.
14 Failure to explain is punishable by a fine on the company of not less than INR 50,000
(about U.S. $833) and up to INR 25 lakhs (about U.S. $41,667). Further, officers who default
on the reporting provision could be subject to up to three years in prison and/or fines of not
less than INR 50,000 rupees (about U.S. $833) and as high as INR 5 lakh (about U.S. $8,333).
15 CNX 500 is a broad-based benchmark of the Indian capital market. It comprises
of 500 firms that represents about 96% of the free float market capitalization of the
stocks listed on National Stock Exchange of India (NSE). More details can be found at
http://www.nseindia.com/products/content/equities/indices/cnx 500.htm
1270
H. MANCHIRAJU AND S. RAJGOPAL
Before we proceed any further with the discussion of market reaction
tests, we lay out certain crucial assumptions that we need to make to interpret these results. It is important to emphasize that we do not assume that
the stock market has perfect foresight. Specifically, on any given event date,
the stock market does not know (i) if and when the CSR law will eventually
go in effect, and (ii) what the final form of the CSR law will look like when
it is eventually passed. Hence, the events we examine represent the natural
progression of a bill under the Indian parliamentary system. It is reasonable to assume that as we proceed along the time line of event dates (i) the
probability of eventual passage of the law increases, and (ii) the probability
of changes to the current version of the bill decreases. Therefore, the difference in the market reaction for the AFFECTED and UNAFFECTED firms
on any event date captures the market’s assessment of the eventual passage
of the law in its current form. We use such market reaction to proxy for the
effect of CSR activity on firm value.
It is important to note that our classification of firms into AFFECTED
and UNAFFECTED does not induce any perfect foresight bias. As mentioned earlier, we classify a firm as AFFECTED or UNAFFECTED based on
whether its profit/sales/book value for the most recent fiscal year end
meets the thresholds mentioned in the CSR rule or not. However, it is
not known whether that same company would continue to meet the profitability threshold at some point of time in future when the CSR rule is
actually passed and is effective. We assume that the market considers the
current level of profitability as an indicator of future profitability. Thus,
we assume that firms that are classified as AFFECTED on a given event
date before the passage of the law are also more likely to be classified as
AFFECTED even after the passage of the law. This assumption is further
strengthened by the fact that three firm characteristics (profitability, sales,
and book value) are used to classify a firm as AFFECTED or UNAFFECTED.
Because the numerical cutoffs related to these three firm characteristics
represent either or conditions, it is unlikely that a firm classified as AFFECTED on a given event date will not be classified as AFFECTED in the
future.16
16 These assumptions are reasonable because (i) the Companies Act with a CSR clause in it,
even though introduced in 2010, eventually was passed in 2013; (ii) no changes were made in
the threshold profit/sales/book value levels that would go in determining whether a firm will
be affected by the CSR rule or not; and (iii) on average, about 83% of firms that we classify
as AFFECTED in year t remain classified as AFFECTED in year t + 1. Similarly, 62% of firms
that we classify as UNAFFECTED in year t remain classified as UNAFFECTED in year t + 1. In
fact, with the benefit of hindsight, it can also be argued that the market fails to anticipate the
time needed to shepherd the Indian Companies Act 2013 into law. For example, by reacting
to Event 2 (figure 2b) as it did, the market is effectively acting as if the legislation will go into
effect immediately. As it turns out, the legislation took another three years, by which time,
whether a firm fell on one side of the cutoff in figure 2b is largely uncorrelated with whether
the firm fell into the AFFECTED category in 2013.
IMPACT OF CSR ON SHAREHOLDER VALUE
3.3
1271
BASIC SET UP OF A REGRESSION DISCONTINUITY DESIGN
We combine the event study approach with a regression discontinuity design (RDD) to document the effect of CSR rule on firm value. The RDD
technique was first introduced by Thistlethwaite and Campbell [1960] in
a study that examined the impact of merit awards on future academic outcomes of students. In their setting, students with test scores, R, greater than
or equal to a cutoff value c received an award, and those with scores below
the cutoff did not receive an award. The process of granting merit awards
generates a sharp discontinuity in the treatment variable (which is receiving the award, in this case) as a function of the rating variable (which is test
score, in this case).17 More formally, the receipt of treatment is denoted by
the dummy variable D ∈ {0, 1}, where D = 1 if R ࣙ c, and D = 0 if R < c. This
framework assumes that individuals with scores just below the cutoff (who
did not receive the award) are very much comparable to individuals with
scores just above the cutoff (who did receive the award). Therefore, a discontinuous jump in Y at c can be attributed to the causal effect of the merit
award because there is likely to be no reason, other than the merit award,
for future academic outcomes, Y, to be a discontinuous function of the test
score. The relationship between Y and R can therefore be estimated using
the following regression model:
Y = α + τ D + f (R) + ε,
(1)
where the term τ captures the effect of merit awards on future academic
outcomes. The inferences drawn under an RDD approach are considered
to be credible because the assignment of individuals in treatment and control groups is “as good as randomized” given that individuals cannot precisely control the assignment variable near the exogenously determined
cutoff (Lee and Lemieux [2010]).
Since the 1990s, the RDD technique has been used in a variety of economic contexts. Some recent applications of RDD in finance include: (i)
Akey [2015] who estimates the value of a firm’s political connections by
comparing the differences in abnormal stock returns of firms connected
to candidates who just won/lost a close election; (ii) Flammer [2015] who
estimates the effect of CSR on firm value by comparing the stock market
reaction to firms whose CSR proposals pass/fail by a small margin of votes;
(iii) Black and Kim [2006] who study Korean governance reforms, which
apply to firms with assets greater than 2 trillion Korean Won, but not to
smaller firms; and (iv) Iliev [2010] who studies the impact of section 404 of
Sarbanes-Oxley, which applies to U.S. firms with $75M public float but not
to the smaller firms.
Our setting of the mandatory CSR rule nicely fits the RDD framework.
The mandatory CSR rule imposes certain profitability/size cutoffs that
17 The rating variable is also referred to as forcing variable or the assignment variable in the
literature.
1272
H. MANCHIRAJU AND S. RAJGOPAL
classify firms as AFFECTED and UNAFFECTED. Assuming firms that are just
above and below the cutoffs are fundamentally similar, unobservable firm
characteristics are less likely to impact the relation between CSR and firm
value. Further, there is no reason to believe that one group is more likely
to undertake CSR activities than the other, thereby mitigating the concern that firms spending on CSR self-select themselves based on their private information about future profitability, as suggested by Lys, Naughton,
and Wang [2015]. Therefore, any differences in the stock market reaction around key events related to the passage of mandatory CSR rule, as
measured by CAR, for the two groups of firms (i.e., AFFECTED and UNAFFECTED) can thus be attributed to the mandatory CSR rule that creates the
discontinuity.
Our research setting differs from the basic RDD applications listed above,
in that the mandatory CSR rule relies on more than one rating score to
determine treatment status. A firm is affected by the mandatory CSR rule
if during any fiscal year it has either (i) a net worth of Indian Rupees
(INR) 5,000 million (about U.S. $83 million) or more; (ii) sales of INR
10,000 million (about U.S. $167 million) or more; or (iii) a net profit of
INR 50 million (about U.S. $0.83 million) or more. Researchers label this
special case of the RDD as multivariate RDD or MRDD. While there are
several methods to estimate treatment effects under MRDD, we follow a
binding-score regression discontinuity as described in Reardon and Robinson
[2012].18 Specifically, we create a new rating variable M, defined as the
minimum of the three individual rating scores based on profits, book value,
and sales (where each score is first centered around its cutoff score), that
perfectly determines the assignment. Continuing with the notation used in
model (1), receipt of treatment is denoted by the dummy variable D {0, 1},
where D = 1 if M ࣙ c, and D = 0 if M < c. We then fit the following model:
Y = α + τ D + f (M ) + ε.
(2)
As pointed out by Reardon and Robinson [2012], the binding-score regression discontinuity model is appealing because it is intuitively simple
and it parsimoniously reframes the multidimensional vector of rating scores
into a single dimension that alone determines treatment status, and hence
ensures minimal loss of data in estimation.
3.4
DATA
The data for our study are obtained primarily from Prowess, a database
maintained by the Centre for Monitoring Indian Economy (CMIE). The
Prowess database is widely used by scholars (e.g., Khanna and Palepu
[2000], Bertrand, Mehta, and Mullainathan [2002], Gopalan, Nanda, and
Seru [2007]) to conduct large sample studies on Indian firms.
18 Other methods to implement MRDD include response surface RD, frontier RD, fuzzy
frontier RD, and distance-based RD. Please refer to Reardon and Robinson [2012] and Wong,
Steiner, and Cook [2013] for more details.
IMPACT OF CSR ON SHAREHOLDER VALUE
1273
Our full sample comprises of firms with nonmissing data on total assets,
sales, net income, book value of equity, and market capitalization. We also
require a firm to have stock return data around key event dates related to
the passage of the Companies Act. These data restrictions yield a sample
of 9,921 firm-year observations relating to 2,120 unique firms for the period 2009–2013. There are 5,889 firm year observations (59% of the sample) that are classified as AFFECTED and the remaining 4,032 firm-year
observations (41% of sample) are classified as UNAFFECTED. Panel A of
table 1 presents the industry-wise distribution of our sample. Business services, chemicals, construction, textile, and wholesale industries are heavily
represented among the AFFECTED firms.
To assess whether the nature of CSR spending voluntarily incurred by
firms is consistent with the nature of CSR activities enumerated by the law,
we hand-collect information on the magnitude and characteristics of CSR
spending for just one year from the 2012 annual report or CSR reports of
firms. In panels B and C of table 1, we describe amounts spent on CSR and
the type of activities undertaken.19 Of the 1,237 AFFECTED firms for the
year 2012, only 458 firms (37%) spend money on CSR activities. Panel B
presents the descriptive statistics on the amount of CSR spent by the AFFECTED firms in the year 2012. The median amount spent on CSR is INR
3.02 million (approx. $50,333) which is roughly 0.37% of the average of last
three years’ net income. The 75th percentile for CSR spending as a proportion of net income is 1.31%. Hence, before the passage of the CSR law, a
majority of firms do not spend funds on CSR and those that do, spend well
below the proposed 2% level. Panel C provides details on the type of CSR
activity undertaken by these firms. The most common areas where firms
focus their CSR spending are related to community welfare, education, environment, and health care. These areas of spending are mostly in line with
the type of CSR activities listed in the mandatory CSR rule.
To implement RDD, we focus on firms that are just below and just above
the cutoff prescribed in the CSR rule. Specifically, we select firms whose
binding score rating variable M (as defined in section 3.3) ranges between
–0.50 and +0.50. Hereafter, we refer to this subsample as the RDD sample.
The RDD sample comprises of 2,664 firm year observations from 2009 to
2013. Of these total observations 1,633 firm year observations (61% of the
RDD sample) are classified as AFFECTED and the remaining 1,031 firm-year
observations (39% of RDD sample) are classified as UNAFFECTED. Because
the RDD analysis is conducted separately for each event date, and each
year can have multiple events, the exact number of observations used in
each test is different.
19 To be clear, we intend for the 2012 data on CSR magnitude and type of spending in panels
B and C to be a completely descriptive analysis. We emphasize that we do not use these CSR
magnitude data from 2012 to classify firms as treated and control firms throughout the paper.
In particular, we do not use the 2012 magnitude of CSR spending in any of the regression
analyses reported in the paper.
1274
H. MANCHIRAJU AND S. RAJGOPAL
TABLE 1
Sample Description
Panel A: Industry-wise distribution
Industry
UNAFFECTED
AFFECTED
TOTAL
%
135
167
380
263
161
69
87
89
134
121
98
207
133
294
348
462
884
4,032
41%
337
306
415
491
475
177
153
166
236
276
154
295
322
328
205
214
1,339
5,889
59%
472
473
795
754
636
246
240
255
370
397
252
502
455
622
553
676
2,223
9,921
5%
5%
8%
8%
6%
2%
2%
3%
4%
4%
3%
5%
5%
6%
6%
7%
22%
100%
Automobiles and Truck
Banking
Business Services
Chemicals
Construction
Construction Material
Consumer Goods
Electrical Equipment
Food Products
Machinery
Nonmetallic and Industries
Pharmaceutical Production
Steel Works etc.
Textiles
Trading
Wholesale
Other industries with < 2% frequency
Total
Panel B: Descriptive data on CSR spending
N = 458
Mean
Median
SD
P25
P75
CSR spending (INR million)
CSR spending /average last three years profits (%)
49.30
1.22
3.02
0.37
200.89
2.31
0.53
0.07
14.39
1.27
Panel C: Descriptive data on CSR activity undertaken
CSR activity
Community welfare
Education
Environment
Healthcare
Rural development
Women empowerment
Children health
Donations
Disaster relief
Sports
Support for physically challenged
N
%
242
212
191
178
44
37
35
31
25
12
9
52.84%
46.29%
41.70%
38.86%
9.61%
8.08%
7.64%
6.77%
5.46%
2.62%
1.97%
Our full sample consists of 9,221 firm year observations over the period 2009–2013. In panel A of this
table, we present the industry-wise distribution of the sample. In panels B and C of this table, we provide
descriptive statistics on amount spent on CSR, and the type of CSR activity undertaken by sample firms
in the year 2012. Note that the data in panels B and C are descriptive in nature and are not used in any
regression analysis reported in the paper.
IMPACT OF CSR ON SHAREHOLDER VALUE
1275
4. Results
4.1 VERIFYING ASSUMPTIONS UNDERLYING RDD
Consistent with the best practices recommended by Lee and Lemieux
[2008], we begin by testing two crucial assumptions of RDD. The first assumption is that individuals cannot precisely manipulate the rating variable
and hence cannot select themselves into the treatment and control groups.
For example, if firms systematically manage earnings downwards to report
earnings below INR 50 million to avoid compliance with the mandatory
CSR rule, then inferences based on RDD would be invalid because the assignment of firms in the treatment and control groups is not as good as
randomized. To examine this possibility, we graph the frequency of firms
around the profits, book value, and sales cutoffs prescribed in the CSR rule
for the year 2009 and the year 2013. In the year 2009, there was no information about the mandatory CSR rule, whereas in the year 2013, the guidelines for mandatory CSR were well known. Any abnormal jump (drop)
in the frequency of firms just to the left (right) of the cutoff in the year
2013, relative to the year 2009, would suggest that firms deliberately manipulated their profits/book value of equity/sales to avoid potential compliance with the mandatory CSR rule. Figures 2a–2f show no such jumps
or drops in the frequency of firms. Further, in our setting, three different
cutoffs determine the assignment into the treatment group. Hence, the
likelihood that a firm will simultaneously manipulate all the three different
thresholds to avoid compliance with the mandatory CSR rules is relatively
low.
Another important aspect of RDD is that for this approach to work, all
other factors that determine the dependent variable (CAR in our case)
must also evolve smoothly with respect to the rating variable. If the other
variables also show a discontinuity at the cutoff, then the estimated treatment effect of the intervention will be biased. In order to test the validity
of this assumption, we compare the mean and median values of the various
characteristics of AFFECTED and UNAFFECTED firms of the RDD sample.
These results are documented in table 2. The construction of these variables is explained in appendix A. AFFECTED firms are typically larger and
more profitable than the UNAFFECTED firms. This is to be expected because the cutoffs that divide firms as AFFECTED and UNAFFECTED firms
are based on size and profitability. However, AFFECTED and UNAFFECTED
firms are similar in terms of other firm characteristics that are likely to affect firm value such as leverage, sales growth, cash holdings, and board
independence.
AFFECTED firms are more likely to belong to a business group compared to UNAFFECTED firms. We adopt the Prowess database’s group classification to identify business group affiliation and government ownership.
Although we do not have an explicit hypothesis related to group affiliation,
that variable is important to control for in the Indian context, as done previously by Khanna and Palepu [2000], Bertrand, Mehta, and Mullainathan
1276
H. MANCHIRAJU AND S. RAJGOPAL
(a)
(b)
Distribution of firms around INR 500 million profit cutoff on 08/29/2013
0
0
50
50
Frequency
Frequency
100
100
150
150
Distribution of firms around INR 500 million profit cutoff on 08/03/2009
-50
0
50
Profit
100
150
-50
0
50
Profit
(c)
100
150
(d)
60
Distribution of firms around INR 5,000 million book value cutoff on 08/29/2013
0
0
20
20
Frequency
Frequency
40
40
60
Distribution of firms around INR 5,000 million book value cutoff on 08/03/2009
2000
3000
4000
5000
Book value
6000
7000
8000
2000
3000
4000
(e)
5000
Book value
6000
7000
8000
(f)
40
Distribution of firms around INR 10,000 million sales cutoff on 08/29/2013
0
0
10
10
Frequency
20
Frequency
20
30
30
40
Distribution of firms around INR 10,000 million sales cutoff on 08/03/2009
5000
7500
10000
Sales
12500
15000
5000
7500
10000
Sales
12500
15000
FIG. 2.—Did firms manipulate profits, book value, or sales to avoid mandatory CSR compliance? Graphs 2a and 2b (2c and 2d) plot the frequency distribution of firms around the
INR50 million profit cutoff (INR5,000 million book value cutoff). Graphs 2e and 2f plot the
frequency distribution of firms around the INR10,000 million sales cutoff. Firms with profit,
book value, or sales above these cutoffs are required by law to spend funds on CSR. Graphs
to the left (a, c, and e) are based on values on 08/03/2009, when the information about
mandatory CSR rule was unknown. Graphs to the right (b, d, and f) are based on values on
08/29/2013 when the information about the mandatory CSR rule was fully known.
[2002], Gopalan, Nanda, and Seru [2007], and other papers. AFFECTED
firms spend more on advertising than UNAFFECTED firms. Further, AFFECTED firms are more politically connected than UNAFFECTED firms. We
measure political connectedness using a dummy variable POLITICAL that
equals one if the firm or the business group to which a firm belongs has
made a contribution of INR 20,000 or more to a political party in India between 2005 and 2012, and zero otherwise. We obtain these data from the
Median
1.4889
0.0236
1.8766
0.5074
0.0203
0.0039
0.0259
0.4615
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
Mean
3.0191
0.0245
2.4263
0.4828
0.0505
0.0805
0.0574
0.4709
0.1009
0.1649
0.0010
0.0126
0.0016
0.0155
0.4578
SD
4.3100
0.0635
1.7854
0.2281
0.0761
0.2487
0.0882
0.1177
0.3013
0.3713
0.0311
0.1116
0.0076
0.1237
0.4985
12.3165
0.0492
1.9229
0.4887
0.0569
0.0780
0.0653
0.4611
0.2394
0.3356
0.0031
0.0276
0.0043
0.0790
0.4770
Mean
8.6899
0.0413
1.4392
0.5089
0.0315
0.0026
0.0351
0.4545
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
Median
SD
15.9263
0.0611
1.5815
0.2162
0.0714
0.1919
0.0848
0.1162
0.4269
0.4723
0.0553
0.1637
0.0132
0.2698
0.4996
AFFECTED (2) N = 1,633
[3] – [2]
7.2010∗∗∗
0.0177∗∗∗
−0.4374∗∗∗
0.0015
0.0113∗∗
−0.0013
0.0092
−0.0070
0.0000∗∗∗
0.0000
0.0000
0.0000
0.0000∗∗∗
0.0000∗∗∗
0.0000
9.2973∗∗∗
0.0247∗∗∗
−0.5034∗∗∗
0.0059
0.0064∗∗
−0.0025
0.0079
−0.0098
0.1386∗∗∗
0.1707∗∗∗
0.0021
0.0149
0.0026∗∗∗
0.0635∗∗∗
0.0192
Difference in median
[2] – [1]
Difference in mean
This table reports the mean, median, and standard deviation of various firm characteristics for the AFFECTED and UNAFFECTED firms for the RDD sample, which comprises of
2,664 firm year observations from 2009 to 2013. The RDD sample is constructed as follows. First, a binding score rating variable M is created which is defined as the minimum of the
three individual rating scores R1, R2, and R3, which are in turn defined as (Profit – 50)/50, (Book value – 5,000)/5,000, and (sales – 10,000)/10,000, and which perfectly determines
whether a firm is affected by the mandatory CSR rule or not. Profit, book-value and sales for the most recent fiscal year end before any event date are considered to eliminate any
perfect foresight bias. The INR 50 million profits, INR 5,000 million book value, and INR 10,000 sales are the three cutoffs specified in the mandatory CSR rule. If a firm has profit,
book value, or sale above these cutoffs, the firm is then required to spend on CSR. Firms with the variable M ranging between −0.50 and 0.50 comprise our RDD sample, where firms
with M > 0 are AFFECTED by the rule and firms with M < 0 are UNAFFECTED by the rule. The significance of differences in means and medians are evaluated based on the t-test and
Wilcoxon test, respectively (p-values for the t-statistics and Z-statistics are two-tailed). ∗∗∗ , ∗∗ and ∗ denote significance at 1%, 5%, and 10% level, respectively.
ASSETS (INR, billions)
ROA
BM
LEV
CAPEX
SGROWTH
CASH
BOARD INDPENDENCE
BIG4
BG
GOVT OWNED
MNC
AD
POLITICAL
POLLUTED
UNAFFECTED (1) N = 1,031
TABLE 2
Descriptive Statistics for the RDD Sample
IMPACT OF CSR ON SHAREHOLDER VALUE
1277
1278
H. MANCHIRAJU AND S. RAJGOPAL
Web site of the Association of Democratic Reforms, a not-for-profit organization working in the area of electoral and political reforms in India.20
Finally, AFFECTED and UNAFFECTED firms have equal representation in
industries identified by Ministry of Environment and Forests, Government
of India as heavily polluting industries.21 The dummy variable POLLUTED
captures these industries and is equal to one if a firm belongs to metallurgical, chemical, petrochemical, coal, thermal power, building materials,
paper, brewing, pharmaceutical, fermentation, textiles, leather, or the mining industry, and zero otherwise. For these four variables we find no jumps
in the covariates at the cutoffs.22 These results give us comfort to proceed
with the RDD estimation.
4.2
UNIVARIATE RESULTS
In table 3, we report median CAR for the AFFECTED and the UNAFFECTED firms in the RDD subsample, around the each of the eight key
legislative events related to the passage of the Companies Act 2013, as outlined in section 3.1.
There is no significant market reaction to Event # 1 on August 3, 2009,
that is, when the Bill was introduced for the first time in the Lok Sabha
(lower house of the Indian Parliament). This outcome is not surprising
because the August 3, 2009 version of the Bill contained no clause related to
CSR. The insignificant result (which is akin to a placebo test) mitigates the
possibility that some unobserved firm characteristics drive the differences
in CAR for the two sub-samples. 23
On August 31, 2010 (Event # 2), the Standing Committee on Finance
presented its report with a mandatory CSR clause. We find a significant
negative market reaction for the AFFECTED (−2.3%) firms but a slightly
positive reaction for the UNAFFECTED firms (0.3%), consistent with the hypothesis that CSR is value-destroying. This event is important in the context
of our study because it contained news related to only the CSR aspect of
the Companies Act. As discussed in section 2, the Parliamentary Standing
Committee on Finance vetted the initial version of the Companies Act and
inserted a mandatory CSR clause. This new rule was a totally unexpected
addition to the Act. For the first time profitability/size/net-worth cutoffs
were prescribed to determine which firms will be affected by the mandatory CSR rule. These cutoffs eventually remained unchanged in the final
version of the bill that was enacted.
20 http://adrindia.org/research-and-report/political-party-watch/combined-reports/2014
/corporates-made-87-total-donations-k
21 http://envfor.nic.in/legis/ucp/ucpsch8.html
22 These results and accompanying graphs are available upon request.
23 Note that we have used the numerical cutoffs prescribed in the second event date
(August 31, 2010) to classify firms into AFFECTED and UNAFFECTED for stock returns related
to the first event date (August 3, 2009). In one sense, this reflects hindsight bias. We prefer
to interpret the null result related to returns on the first event date on August 3, 2009 as a
placebo test.
IMPACT OF CSR ON SHAREHOLDER VALUE
1279
TABLE 3
Market Reaction – Univariate Results Based on RDD Sample
Event
#
1
2
3
4
5
6
7
8
Date
Aug 3, 2009
Aug 31, 2010
Feb 28, 2011
Dec 14, 2011
Jun 26, 2012
Dec 18, 2012
Aug 8, 2013
Aug 29, 2013
Median CAR for
a firm around
all eight events
Predicted sign on
AFFECTED firms (under
the hypothesis that CSR UNAFFECTED
is value-destroying)
(1)
Insignificant
−
+
−
−
−
−
−
−
0.003
0.003∗∗∗
−0.002
−0.001
0.002∗∗∗
0.006∗∗∗
−0.005∗∗
−0.004
−0.003
AFFECTED Difference
(2)
(2) – (1)
0.001
−0.023∗∗∗
0.004
−0.001
0.001∗
−0.016∗∗∗
−0.006
−0.003
−0.008∗
−0.001
−0.026∗∗∗
0.006
0.000
−0.001
−0.021∗∗∗
−0.001
0.001
−0.005
This table reports median 3-day CAR around the key legislative events for the two subgroups of our
sample, that is, UNAFFECTED, and AFFECTED firms in the RDD sample. We measure abnormal returns by
estimating the market model using 200 trading days of return data ending 11 days before key legislative
events related to the passage of the Companies Act 2013. The return on National Stock Exchange’s CNX
500 index is used as a proxy for the market return. Daily abnormal stock returns are cumulated to obtain
the CAR from day t − 1 before the legislative event date to day t + 1 after the event date. The RDD sample is
constructed as follows. First, a binding score rating variable M is created which is defined as the minimum of
the three individual rating scores R1, R2, and R3, which are in turn defined as (Profit – 50)/50, (Book value
– 5,000)/5,000, and (sales – 10,000)/10,000, and which perfectly determines whether a firm is affected by
the mandatory CSR rule or not. Profit, book-value and sales for the most recent fiscal year end before any
event date are considered to eliminate any perfect foresight bias. The INR 50 million profits, INR 5,000
million book value, and INR 10,000 sales are the three cutoffs specified in the mandatory CSR rule. If a
firm has profit, book value, or sale above these cutoffs, the firm is then required to spend on CSR. Firms
with the variable M ranging between −0.50 and 0.50 comprise our RDD sample, where firms with M > 0 are
AFFECTED by the rule and firms with M < 0 are UNAFFECTED by the rule. The significance of median and
the difference in median values for the two subgroups is tested using Wilcoxson test. ∗∗∗ , ∗∗ , and ∗ denote
significance at 1%, 5%, and 10% level (two-tailed), respectively.
For Event # 3, when the Ministry of Corporate Affairs announced that it is
considering making only the disclosure of CSR and not the actual spending
on CSR mandatory, we expect a positive market reaction for the AFFECTED
firms because this event is likely to mitigate any possible negative effects
of the mandatory CSR rule. However, we do not observe any significant
market reaction.
Events # 4–8 are associated with increased probability of the legislation
passing. Therefore, we expect a negative market reaction for the AFFECTED
firms under the hypothesis that CSR is value-destroying. However, we find a
statistically significant negative reaction for the AFFECTED firms only for
Event # 6 when the Lok Sabha (the lower house of the Indian Parliament) passed the Companies Act (–1.6%). The other events such as the
re-introduction of the Bill in the Lok Sabha (Event# 4), submission of a revised report by the Parliamentary Standing Committee on Finance (Event
# 5), the passage of this Bill in the Rajya Sabha, the upper house of the Indian Parliament (Event # 7), and Presidential assent to the Bill (Event# 8)
1280
H. MANCHIRAJU AND S. RAJGOPAL
are all associated with a statistically insignificant market reaction for the
AFFECTED firms.
The median CAR for the AFFECTED firms across all these eight events is
negative. Overall, these results provide initial evidence of a negative impact
of CSR on shareholder value.
4.3
GRAPHICAL ANALYSIS
Graphs provide a transparent way of showing how the treatment effect
is identified in an RDD framework. For each of the eight event dates outlined previously, we separately plot the stock market reaction for firms in
the RDD sample. Figures 3a–3h show a scatter plot that presents the stock
market reaction on significant CSR events for AFFECTED and UNAFFECTED
firms. On the X-axis is the binding score rating variable (M, in this case)
which ranges from –0.50 to 0.50, with 0 being the cutoff point. On the Y
axis is the three-day CAR around the event date. To enhance the visual impact of the graph, estimated CAR is superimposed on the plot, where CAR
is estimated as a function of the rating variable M separately on both sides
of the cutoff using local linear estimation, triangular kernel, and Imbens–
Kalyanaraman [2012] bandwidth, using the following equation:
CAR = α + τ AFFECTED + f (M ) + ε.
(3)
Because the only difference near the cutoff is that firms to the right of
zero are AFFECTED by the mandatory CSR rule whereas firms to the left of
zero are UNAFFECTED, any discontinuity in CAR at the cutoff can therefore
be attributed to the mandatory CSR rule because it imposes the threshold
and creates the discontinuity.
We use the RD command in STATA, developed in Nichols [2007], to generate this graph. Figures 3a through 3h correspond to each of the event
dates. For the sake of parsimony, we describe only three graphs. Graph 3a
relates to the Event # 1 on August 3, 2009. There is no significant jump
or drop at the cutoff point suggesting that the average CAR for the AFFECTED and UNAFFECTED firms on this date are statistically indistinguishable. Graph 3b relates to the Event # 2 on August 31, 2010. At the cutoff
point, the fitted curve on the right is significantly lower that the fitted curve
to the left. The only difference near the cutoff is that firms to the right of
zero are AFFECTED by CSR rule whereas firms to the left of zero are UNAFFECTED. Hence, this discontinuous drop in the fitted curve (that captures
firm value) at the threshold can be attributed to the mandatory CSR rule
which imposes the cutoff and creates the discontinuity. Based on this graph
3b, we infer a negative relation between CSR and firm value. Using the
same logic, we find a negative relation between CSR and firm value in Figure 3f that relates to the Event # 6 on December 18, 2012. However, all
the remaining graphs suggest that the market reaction for the AFFECTED
and UNAFFECTED firms on the remaining event dates are statistically
indistinguishable.
IMPACT OF CSR ON SHAREHOLDER VALUE
(b)
3-day CAR around Event# 2 - 08/31/2010
-.5
0
.5
-.06
-.04
-.04
-.02
-.02
0
0
.02
.02
.04
.04
.06
.06
(a)
3-day CAR around Event# 1 - 08/03/2009
1281
-.5
0
(d)
3-day CAR around Event# 4 - 12/14/2011
-.5
0
.5
-.06
-.06
-.04
-.04
-.02
-.02
0
0
.02
.02
.04
.04
.06
.06
(c)
3-day CAR around Event# 3 - 02/28/2011
.5
-.5
0
(f)
3-day CAR around Event# 5 - 06/26/2012
3-day CAR around Event# 6 - 12/18/2012
.5
-.06
-.06
-.04
-.04
-.02
-.02
0
0
.02
.02
.04
.04
.06
.06
(e)
-.5
0
.5
-.5
0
(h)
3-day CAR around Event# 7 - 08/28/2013
3-day CAR around Event# 8 - 08/29/2013
.5
-.06
-.04
-.04
-.02
-.02
0
0
.02
.02
.04
.04
.06
.06
(g)
-.5
0
.5
-.5
0
.5
FIG. 3.—CAR around significant CSR events for the RDD sample. Figures 3a–3h show the market reaction for the AFFECTED and UNAFFECTED firms around events related to the passage
of mandatory CSR rule for the RDD sample. The RDD sample is constructed as follows. First, a
binding score rating variable M is created which is defined as the minimum of the three individual rating scores R1, R2, and R3, which are in turn defined as (Profit – 50)/50, (Book value
– 5,000)/5,000, and (sales – 10,000)/10,000, and which perfectly determines whether a firm
is affected by the mandatory CSR rule or not. Profit, book-value and sales for the most recent
fiscal year end before any event date are considered to eliminate any perfect foresight bias.
The INR 50 million profits, INR 5,000 million book value, and INR 10,000 sales are the three
cutoffs specified in the mandatory CSR rule. If a firm has profit, book value, or sale above
these cutoffs, the firm is then required to spend on CSR. Firms with the variable M ranging
between −0.50 and 0.50 comprise our RDD sample, where firms with M > 0 are AFFECTED
by the rule and firms with M < 0 are UNAFFECTED by the rule. The variable M is plotted on
the X-axis. The vertical axis indicates 3-day CAR. The solid line plots predicted value of which
is estimated as a linear function of M, estimated separately to the left and right of the cutoff
point (M = 0). RD routine of STATA as described in Nichols (2017) is used to generate these
graphs.
1282
H. MANCHIRAJU AND S. RAJGOPAL
We ascertain the validity of these inferences by performing two robustness checks. First, as suggested by Lee and Lemieux [2010], if the true relation between the X and Y variables is nonlinear, then RDD using local
linear estimations will induce a bias in favor of finding a treatment effect
when there is none. To rule this possibility out, we assume the relationship
between CAR and M is a third-order polynomial function and then superimpose the fitted CAR on the scatter plot. In estimating local polynomial
regressions, the number of bins is determined using the approach outlined
in Calonico, Cattaneo, and Titiunik [2014a]. Despite changes in the estimation techniques, the graphs for Event # 2 and Event # 6, that is, on August
31, 2010 and December 18, 2012, display a sudden drop in the fitted CAR
at the cutoff point. Second, we rule out the possibility that certain extreme
values of CAR are driving these results. A sharp discontinuity at the cutoff
point is still seen in the graphs for these two event dates even after dropping
the top and bottom 1% of the observations.24
The magnitude of the treatment effect τ , as reflected in these graphs, is
summarized in table 4. Panel A documents the results following the same
estimation techniques as used in the graphs, that is, RDD analysis using local linear estimation, triangular kernel, and Imbens–Kalyanaraman [2012]
bandwidth. We use the RDROBUST command in STATA, developed in
Calonico, Cattaneo, and Titiunik [2014b] to perform RDD estimations. We
continue to find a significant negative market reaction around August 31,
2010, the date when the Parliamentary Standing Committee on Finance
inserted a mandatory CSR clause in the Companies Act. The CAR for AFFECTED firms is 2.4% lower than the UNAFFECTED firms. Similarly, there
is a significant negative market reaction around December 18, 2012, the
date when the Lok Sabha (lower house of the Indian Parliament) passed
the Companies Act. The CAR for AFFECTED firms is 1.6% lower than the
UNAFFECTED firms. We do not find a statistically significant difference
in the market reaction for AFFECTED and UNAFFECTED firms for other
events. In the panel B of table 4, we also report results from estimating the
RDD using third-order polynomial and optimal bandwidths as suggested in
Calonico, Cattaneo, and Titiunik [2014a]. The results are similar to those
documented in table 3.
4.4
PARAMETRIC ESTIMATION
We expand the basic RDD framework to incorporate other factors that
affect the stock market’s reaction and estimate the following regression
equation:
τ j EVEN T j ∗ AFFECTED + δM ∗ AFFECTED
CAR = α +
+
n
24 We
j
βn M ∧n +
i
γi Ri +
λk Contr ol j + ε.
k
do not tabulate these graphs in this paper, but these are available upon request.
(4)
0.004
(0.239)
480
3
26
−0.001
(−0.070)
480
1
26
EVENT 1
08/03/2009
−0.026∗∗∗
(−3.326)
517
3
28
−0.024∗∗∗
(−4.499)
517
1
28
EVENT 2
08/31/2010
−0.001
(−0.138)
556
3
28
0.001
(0.182)
556
1
28
EVENT 3
02/28/2011
0.002
(0.193)
543
3
26
0.003
(0.496)
543
1
26
EVENT 4
12/14/2011
0.003
(0.301)
514
3
28
0.001
(0.095)
514
1
28
EVENT 5
06/26/2012
−0.019∗∗
(−1.936)
536
3
29
−0.016∗∗∗
(−2.735)
536
1
29
EVENT 6
12/18/2012
0.004
(0.345)
491
3
25
0.002
(0.235)
491
1
25
EVENT 7
08/08/2013
−0.007
(−0.646)
481
3
25
0.000
(0.004)
481
1
25
EVENT 8
08/29/2013
This table summarizes the difference in the market reaction for the AFFECTED and UNAFFECTED firms around events related to the passage of mandatory CSR rule for the RDD
sample. The market reaction is measured as the 3-day CAR. We measure abnormal returns by estimating the market model using 200 trading days of return data ending 11 days
before key legislative events related to the passage of the Companies Act 2013. The return on National Stock Exchange’s CNX 500 index is used as a proxy for the market return.
Daily abnormal stock returns are cumulated to obtain the CAR from day t−1 before the legislative event date to day t + 1 after the event date. The RDD sample is constructed as
follows. First, a binding score rating variable M is created which is defined as the minimum of the three individual rating scores R1, R2, and R3, which are in turn defined as (Profit
– 50)/50, (Book value – 5,000)/5,000, and (sales – 10,000)/10,000, and which perfectly determines whether a firm is affected by the mandatory CSR rule or not. Profit, book-value
and sales for the most recent fiscal year end before any event date are considered to eliminate any perfect foresight bias. The INR 50 million profits, INR 5,000 million book value,
and INR 10,000 sales are the three cutoffs specified in the mandatory CSR rule. If a firm has profit, book value, or sale above these cutoffs, the firm is then required to spend on CSR.
Firms with the variable M ranging between −0.50 and 0.50 comprise our RDD sample, where firms with M > 0 are AFFECTED by the rule and firms with M < 0 are UNAFFECTED by
the rule. In panel A of the table the dependent variable CAR is estimated as a function of M, separately on the both sides of the cutoff (i.e., where M = 0) using local linear estimation,
triangular kernel, and Imbens–Kalyanaraman (2012) bandwidth and the difference in estimated CAR for the two sides is documented here. This difference in CAR corresponds to
the jump (or drop) in the fitted curve at the cutoff point as documented in figures 2(a)–2(g). In panel B, CAR is estimated as a function of M, using local third order polynomial
estimation, triangular kernel as per Calonico, Cattaneo, and Titiunik (2014a). Operationally, RDROBUST routine of STATA as described Calonico, Cattaneo, and Titiunik (2014b)
is used to generate these estimates. ∗∗∗ , ∗∗ , and ∗ denote significance at 1%, 5%, and 10% level (two-tailed), respectively.
Coefficient
z-statistics
Observations
Polynomial order
Bins
Panel B
Coefficient
z-statistics
Observations
Polynomial order
Bins
Panel A
Dependent variable = CAR
TABLE 4
Market Reaction – Summary of Graphical Analysis Using RDD
IMPACT OF CSR ON SHAREHOLDER VALUE
1283
1284
H. MANCHIRAJU AND S. RAJGOPAL
This pooled regression is estimated over all the event dates. The dependent variable CAR is the market reaction around the event dates, as
previously described. Our coefficient of interest is τ j which captures the
differential stock market reaction for the AFFECTED firms relative to the
UNAFFECTED firms on a given event date. The overall
differential market
τ j . The β j coefficients
reaction for the AFFECTED firms is measured by
j
in model (4) capture the nonlinearity in the relationship between CAR
and the rating variable M constructed under the binding-score regression
discontinuity design. The γ j coefficients capture the relationship between
CAR and the three rating scores R based on profits, book value, and sales
that separate the firms as AFFECTED and UNAFFECTED. The δ coefficient
in model (4) captures the differential relationship between the binding
score rating variable M and CAR in the AFFECTED firms. These terms are
included in the model following standard RDD practices, as suggested in
Lee and Lemieux [2008].
The regression model includes several control variables. Following Fama
and French [1992], we include firm size and book-to-market ratio in the
model and expect negative and positive coefficients on these variables, respectively. The mandatory CSR rule is a part of an overall Companies Act
2013. We control for three important provisions of this Act that relate to
board independence and audit quality (proxied by a Big 4 auditor), and
enhanced creditor rights (proxied by leverage). These regulatory requirements can either impose additional costs on firms or result in benefits due
to better monitoring. Hence, we do not make any predictions about the expected sign of coefficients on these variables. We acknowledge that capturing all the provisions of the Companies Act that are value relevant is difficult
and hence there are genuine concerns about omitted correlated variables.
However, it is important to emphasize that except for the mandatory CSR
rule, all the provisions of the Companies Act 2013 apply to all firms uniformly, regardless of whether they fall in the AFFECTED or UNAFFECTED
categories. Hence, the differential stock price reactions for these three categories of firms can therefore be uniquely attributable to the mandatory
CSR rule.
We also include indicator variables in the model that capture whether
a firm belongs to a business group, is foreign owned, or is Government
owned. Finally, we include industry dummy variables to overcome the problem associated with cross-sectional correlations in the individual firm returns that will arise given that the event period is common to all firms in our
sample. Event time dummies are also included to control for firm-invariant
time-specific effects.
Table 5 presents the results from estimating equation (4). Column (1)
shows results from estimating a reduced form model without control variables, column (2) reports results from estimating the full model with
control variables, and column (3) presents results from estimating the
full model with control variables and firm fixed effects. Across the three
EVENT 1 ∗ AFFECTED
EVENT 2 ∗ AFFECTED
EVENT 3 ∗ AFFECTED
EVENT 4 ∗ AFFECTED
EVENT 5 ∗ AFFECTED
EVENT 6 ∗ AFFECTED
EVENT 7 ∗ AFFECTED
EVENT 8 ∗ AFFECTED
M ∗ AFFECTED
M
M∗M
M∗M∗M
R1
R2
R3
SIZE
BM
LEV
ROA
CAPEX
SGROWTH
CASH
BOARD INDPENDENCE
BIG4
τ1
τ2
τ3
τ4
τ5
τ6
τ7
τ8
δ
β1
β2
β3
γ1
γ2
γ3
λ1
λ2
λ3
λ4
λ5
λ6
λ7
λ8
λ9
coef
0.003
−0.022∗∗∗
0.000
−0.003
−0.004∗∗
−0.016∗∗∗
0.001
−0.000
−0.014
0.011
0.009
−0.028
0.000
0.000
0.001
(1)
1.591
−14.174
0.258
−1.571
−2.479
−9.597
0.539
−0.048
−0.515
0.623
0.422
−1.189
0.849
1.412
1.207
t-stat
coef
0.002
−0.023∗∗∗
0.001
−0.003
−0.004∗
−0.015∗∗∗
−0.000
−0.001
−0.005
0.004
0.002
−0.021
0.000
0.000
0.001
0.001
−0.000
−0.004
−0.002
−0.000
−0.001
−0.001
0.010∗∗∗
0.000
(2)
0.984
−9.474
0.491
−1.090
−1.709
−6.275
−0.168
−0.272
−0.173
0.227
0.076
−0.741
0.671
0.364
1.013
0.762
−1.367
−1.440
−0.268
−0.051
−1.111
−0.141
5.459
0.071
t-stat
Dependent Variable = CAR
TABLE 5
Market Reaction –Parametric Estimation of RDD
coef
−0.000
−0.025∗∗∗
−0.004
−0.008
−0.006
−0.018∗∗∗
−0.004
−0.003
−0.061
0.039
0.063
−0.046
0.000
−0.001
−0.009∗∗∗
0.004
−0.000
−0.000
−0.004
0.004
−0.001
−0.011
−0.007
0.011
(3)
t-stat
(Continued)
−0.038
−5.301
−0.772
−1.508
−1.202
−3.494
−0.749
−0.571
−1.501
1.416
1.601
−0.878
0.470
−0.453
−2.625
0.785
−1.055
−0.005
−0.176
0.418
−0.506
−0.723
−0.586
1.207
IMPACT OF CSR ON SHAREHOLDER VALUE
1285
j =1
τ j = 2% F-stat (p-value)
λ10
λ11
λ12
λ13
λ14
λ15
j
τ j E V E N T j ∗ AF F E CT E D + δM ∗ AF F E CT E D +
9.73∗∗∗
(0.00)
2.66
(0.10)
Yes
No
No
4,118
0.050
coef
n
βn M ∧n +
i
t-stat
γi Ri +
coef
k
λk Contr ol j + ε.
Yes
Yes
No
3,807
0.068
0.003
−0.017∗∗∗
−0.003
−0.022
0.006
−0.001
4.72∗∗
(0.03)
1.36 (0.24)
(2)
1.159
−3.474
−1.098
−0.560
1.428
−0.109
t-stat
Dependent Variable = CAR
4.44∗∗
(0.03)
2.23
(0.13)
Yes
No
Yes
3,807
0.354
0.509
0.183
0.002
0.030
t-stat
−0.240
coef
−0.002
(3)
The dependent variable is the 3-day CAR. We measure abnormal returns by estimating the market model using 200 trading days of return data ending 11 days before key legislative
events related to the passage of the Companies Act 2013. The return on National Stock Exchange’s CNX 500 index is used as a proxy for the market return. Daily abnormal
stock returns are cumulated to obtain the CAR from day t – 1 before the legislative event date to day t + 1 after the event date. The estimation is restricted to the RDD sample.
The RDD sample is constructed as follows. First, a binding score rating variable M is created which is defined as the minimum of the three individual rating scores R1, R2,
and R3, which are in turn defined as (Profit – 50)/50, (Book value – 5,000)/5,000, and (sales – 10,000)/10,000, and which perfectly determines whether a firm is affected by
the mandatory CSR rule or not. Profit, book-value and sales for the most recent fiscal year end before any event date are considered to eliminate any perfect foresight bias.
The INR 50 million profits, INR 5,000 million book value, and INR 10,000 sales are the three cutoffs specified in the mandatory CSR rule. If a firm has profit, book value, or
sale above these cutoffs, the firm is then required to spend on CSR. Firms with the variable M ranging between –0.50 and 0.50 comprise our RDD sample, where firms with
M > 0 are AFFECTED by the rule and firms with M < 0 are UNAFFECTED by the rule. Control variables are defined in appendix A. The t-statistics are based on robust standard errors
clustered at firm and event level. ∗∗∗ , ∗∗ , and ∗ denote significance at 1%, 5%, and 10% level (two-tailed), respectively.
CAR = α +
This table presents the results from estimating model (4)
Event FE
Industry FE
Firm FE
Observations
R2
Test :
8
BG
GOVT OWNED
MNC
AD
POLITICAL
POLLUTED
8
Test : j =1 τ j = 0 F-stat (p-value)
(1)
T A B L E 5—Continued
1286
H. MANCHIRAJU AND S. RAJGOPAL
IMPACT OF CSR ON SHAREHOLDER VALUE
1287
columns, the coefficients τ2 , and τ6 are negative and statistically significant
suggesting that AFFECTED firms experienced a negative stock market reaction when the Parliamentary Standing Committee on Finance inserted
a mandatory CSR clause in the initial version of the Companies Act, and
the Bill was passed in the Lok Sabha. In column (3), the βn coefficients
are insignificant suggesting that there is no nonlinearity in the relationship between CAR and the binding score rating variable M. Further, the
rating score M is also not differentially associated with CAR in the AFFECTED firms (compared to UNAFFECTED firms), as seen by the insignificant δ coefficient. The γ1 and γ2 coefficients are also insignificant indicating
that the three individual rating scores R, based on profits and book value,
which separate firms as AFFECTED and UNAFFECTED, are unrelated with
CAR. However, the γ3 coefficient is significant and negative suggesting that
the farther the firm is from the sales-based cutoff for the applicability of
mandatory CSR rule, the more negative is the CAR. None of the control
variables are significantly associated with CAR. This is to be expected because RDD assumes that the treatment and control firms are all but similar
except for the intervention that creates the discontinuity.
Collectively, our results suggest that the mandatory CSR rule has a negative impact on the shareholder value of the AFFECTED firms. In terms of
economic significance of these results, overall, we
find that the CAR for the
AFFECTED firms is about 4.1% (as measured by τ j in column 1) lower
j
than the CAR for UNAFFECTED firms. Given that the mandatory CSR rule
directs that a firm has to spend 2% of its profits on CSR, diverting 2% of
profits away from the positive NPV projects of a firm should theoretically
reduce the firm value by around 2%.25 The point estimate of
τ j is not
j
statistically different from the theoretical prediction of a 2% reduction in
firm value at conventional significance levels. However, 4.1% is certainly
larger than 2%. Our interviews with Chief Financial Officers (CFOs) of certain Indian firms in the treatment sample reveal that in addition to the
direct cash flow effects of mandatory CSR rule, the stock market reaction
might potentially capture certain indirect costs such as (i) the diversion of
scarce managerial time and effort to comply with regulation; (ii) the possibility of greater government intervention in determining where the CSR
money should be spent; and (iii) the dilution of the benefits of strategic
CSR spending by channeling funds from CSR aimed at recruiting talented
employees to CSR that does not fetch benefits to the firm; and (iv) poten-
25 Tax deductibility of CSR expenses was unclear at the time when the mandatory CSR rule was being debated in the parliament (Chugh [2014]) http://archive.
financialexpress.com/story-print/1231302. If this expense is not tax deductible then under
the shareholder expense view of CSR, diverting 2% of profits away from the positive NPV
projects of a firm should theoretically reduce the firm value by around 2%. If this expense
is tax deductible then reduction in the firm value will be around 1.4% (assuming a 30% tax
rate).
1288
H. MANCHIRAJU AND S. RAJGOPAL
tial diversion of funds provisioned for CSR activities to a trust or foundation
owned by promoters or insiders of the firm.
4.5
CROSS-SECTIONAL VARIATION
Next, we focus on the cross-sectional variation in the market reaction to
the mandatory CSR rule for the AFFECTED firms. Specifically, we examine
three dimensions—political connectedness, consumer awareness, and environmental impact—hypothesized by prior work to affect the efficacy of CSR
spending. To test these predictions, we expand model (3) and estimate the
following specification:
τ j EVENT j ∗ AFFECTED +
μ j EVEN T j ∗ AFFECTED
CAR = α +
j
∗VARIATION + δM ∗ AFFECTED +
+
j
n
βn M ∧n +
γi Ri
i
λk Contr ol j + ε.
(5)
k
We estimate three different versions of this model where VARIATION
refers to POLITICAL (political connectedness), HIAD (high advertisement
intensity indicator variable), and POLLUTED (affiliation to a highly polluting industry), respectively. Other terms of the model are same as previously
described for model (4).
Results from estimating model (5) are tabulated in table 6. In column
(1), we consider the impact of political connections of a firm. While the
μ j coefficient is positive and significant on some event dates (especially,
EVENT # 2 and EVENT # 6, consistent with our expectations), it is negative and significant on other event dates. These inconsistent signs make
it difficult for us to draw meaningful inferences on the impact of political
connections on the CSR-firm value relation. In column (2), we generally
find positive μ j coefficients, suggesting that a firm’s advertisement spending mitigates the negative relation between CSR and shareholder value.
Finally, in column (3), the signs on the μ j coefficients are not always consistent with our expectations. These mixed results make it difficult to establish
variation in the CSR–shareholder value relation conditioned on affiliation
to a highly polluting industry.
For brevity, we do not report and discuss the coefficient estimates on control variables for this table. As a robustness test, we also perform this crosssectional analysis using the nonparametric RDD methodology adopted in
table 4. Specifically, we estimate equation (3) on each event date for each
subsample separately. These un-tabulated results are qualitatively similar to
results shown in table 6.
4.6
ANALYSIS USING TOBIN’S Q AS A MEASURE OF FIRM VALUE
While the previous sub-sections looked at the announcement period returns, in this section, we consider the effect of the mandatory CSR rule on
EVENT 1 ∗ AFFECTED
EVENT 2 ∗ AFFECTED
EVENT 3 ∗ AFFECTED
EVENT 4 ∗ AFFECTED
EVENT 5 ∗ AFFECTED
EVENT 6 ∗ AFFECTED
EVENT 7 ∗ AFFECTED
EVENT 8 ∗ AFFECTED
EVENT 1 ∗ AFFECTED ∗ VARIATION
EVENT 2 ∗ AFFECTED ∗ VARIATION
EVENT 3 ∗ AFFECTED ∗ VARIATION
EVENT 4 ∗ AFFECTED ∗ VARIATION
EVENT 5 ∗ AFFECTED ∗ VARIATION
EVENT 6 ∗ AFFECTED ∗ VARIATION
EVENT 7 ∗ AFFECTED ∗ VARIATION
EVENT 8 ∗ AFFECTED ∗ VARIATION
M ∗ AFFECTED
M
M∗M
M∗M∗M
τ1
τ2
τ3
τ4
τ5
τ6
τ7
τ8
μ1
μ2
μ3
μ4
μ5
μ6
μ7
μ8
δ
β1
β2
β3
0.003
−0.025∗∗∗
0.001
−0.002
−0.003
−0.017∗∗∗
0.001
−0.000
−0.015∗∗∗
0.034∗∗∗
−0.003
−0.007
−0.011∗∗∗
0.021∗∗∗
−0.014∗∗∗
−0.002
−0.006
0.004
0.002
−0.020
1.422
−10.810
0.618
−0.985
−1.559
−7.400
0.238
−0.212
−2.817
6.518
−0.555
−1.418
−2.632
4.815
−2.852
−0.404
−0.200
0.238
0.102
−0.733
tstat
0.002
−0.026∗∗∗
0.002
−0.005∗∗
−0.004∗
−0.017∗∗∗
0.001
−0.001
0.002∗∗
0.006∗∗∗
−0.001
0.006∗∗∗
0.000
0.004∗∗∗
−0.002∗∗∗
0.000
−0.005
0.004
0.002
−0.020
coef
0.754
−11.416
0.790
−2.117
−1.716
−6.923
0.351
−0.237
2.189
4.597
−0.944
7.075
0.379
4.509
−2.763
0.168
−0.176
0.229
0.072
−0.744
tstat
VARIATION = HIAD
VARIATION = POLITICAL
coef
(2)
(1)
Dependent Variable = CAR
TABLE 6
Market Reaction – Cross Sectional Analysis of the RDD Sample
0.004
−0.022∗∗∗
−0.000
−0.003
−0.001
−0.017∗∗∗
0.000
−0.001
−0.004∗∗∗
−0.004∗∗∗
0.003∗∗
0.000
−0.006∗∗∗
0.003∗∗∗
−0.001
−0.000
−0.004
0.004
0.001
−0.021
coef
(Continued)
1.491
−8.225
−0.045
−0.995
−0.541
−6.141
0.059
−0.235
−4.281
−4.832
2.376
0.127
−6.204
2.896
−1.143
−0.078
−0.150
0.215
0.049
−0.774
tstat
VARIATION = POLLUTED
(3)
IMPACT OF CSR ON SHAREHOLDER VALUE
1289
γ1
γ2
γ3
j
τ j E V E N T j ∗ AF F E CT E D +
j
tstat
0.716
0.382
0.954
coef
3,807
0.071
4.58∗∗ (0.03)
0.000
0.000
0.001
Yes
Yes
Yes
No
tstat
n
βn M ∧n +
0.658
0.285
0.994
μ j E V E N T j ∗ AF F E CT E D ∗ V ARI AT I ON + δM ∗ AF F E CT E D +
3,807
0.080
0.00 (0.96)
0.000
0.000
0.001
Yes
Yes
Yes
No
coef
VARIATION = HIAD
VARIATION = POLITICAL
i
γi Ri +
3,807
0.070
1.25 (0.26)
0.000
0.000
0.001
Yes
Yes
Yes
No
coef
k
λk Contr ol j + ε
0.653
0.396
1.083
tstat
VARIATION = POLLUTED
(3)
The dependent variable is the 3-day CAR. We measure abnormal returns by estimating the market model using 200 trading days of return data ending 11 days before key legislative
events related to the passage of the Companies Act 2013. The return on National Stock Exchange’s CNX 500 index is used as a proxy for the market return. Daily abnormal stock
returns are cumulated to obtain the CAR from day t – 1 before the legislative event date to day t + 1 after the event date. The estimation is restricted to the RDD sample. The RDD
sample is constructed as follows. First, a binding score rating variable M is created which is defined as the minimum of the three individual rating scores R1, R2, and R3, which are
in turn defined as (Profit – 50)/50, (Book value – 5,000)/5,000, and (sales – 10,000)/10,000, and which perfectly determines whether a firm is affected by the mandatory CSR rule
or not. Profit, book-value and sales for the most recent fiscal year end before any event date are considered to eliminate any perfect foresight bias. The INR 50 million profits, INR
5,000 million book value, and INR 10,000 sales are the three cutoffs specified in the mandatory CSR rule. If a firm has profit, book value, or sale above these cutoffs, the firm is then
required to spend on CSR. Firms with the variable M ranging between –0.50 and 0.50 comprise our RDD sample, where firms with M > 0 are AFFECTED by the rule and firms with
M < 0 are UNAFFECTED by the rule. VARIATION is an indicator variable that refers to BG (business group affiliation), POLITICAL (political connectedness of a firm), HIAD (high
advertisement spending indicator variable), and POLLUTED (affiliation to a highly polluting industry), in columns (1)–(4), respectively. The variables BG, POLITICAL, HIAD, and
POLLUTED are based on beginning of the year values and hence are not tainted by hindsight bias. Control variables are defined in appendix A. The t-statistics are based on robust
standard errors clustered at firm and event level. For brevity, coefficient estimates of the intercept, year dummies, and industry dummies are not reported. ∗∗∗ , ∗∗ , and ∗ denote
significance at 1%, 5%, and 10% level (two-tailed), respectively.
CAR = α +
This table presents the results from estimating model (5)
Observations
R2
j =1
R1
R2
R3
Controls
Event FE
Industry FE
Firm FE
8
Test : μ j = 0 F-stat (p-value)
(2)
Dependent Variable = CAR
(1)
T A B L E 6—Continued
1290
H. MANCHIRAJU AND S. RAJGOPAL
IMPACT OF CSR ON SHAREHOLDER VALUE
1291
firm value in the longer run. Following prior research on firm value and
governance (e.g., Ahern and Dittmar [2012], Black and Kim [2012]), we
use Tobin’s Q as our main measure of firm value.26 , where Q is calculated
as the sum of total assets and market value of equity less common book equity, divided by total assets. We consider natural logarithm of Q to address
the skewness in the variable. We assess whether the mandatory CSR rule
leads to a decline in value for the affected firms, right at the time when the
rule was adopted. If investors consider CSR activities as detrimental to firm
value, then the shareholder value of affected firms should decline, relative
to that of unaffected firms, in the years when the likelihood of the passage
of mandatory CSR rule increased, controlling for other factors that affect
shareholder value.
Table 7 presents results from our analysis of changes in Q.
Panel A reports the univariate results and indicates that there was a significant decline in Q of AFFECTED firms in the years 2011 and 2013, compared
to Q of UNAFFECTED firms. Panel B displays the results from the basic RDD
analysis. Specifically, Ln(Q ) is estimated as a function of the rating variable M separately on both sides of the cutoff using local linear estimation,
triangular kernel, and Imbens–Kalyanaraman [2012] bandwidth. We use
RDROBUST com mand in STATA, developed in Calonico, Cattaneo, and
Titiunik [2014b] to perform the RDD estimations. We find a statistically
significant difference in Ln(Q ) for AFFECTED and UNAFFECTED firms
in the year 2011 and 2013.
Finally, in panel C of this table, we show results from estimating the following model:
τ j YEAR j ∗ AFFECTED + δM ∗ AFFECTED
Ln Q = α +
+
n
j
βn M ∧n +
i
γi Ri +
λk Contr ol j + ε.
(6)
k
In this model we control for firm-specific characteristics such as size,
leverage, growth opportunities, profitability, liquidity, ownership structure,
board independence, and the presence of Big 4 auditor, which are known
to affect firm value. Column 1 shows the estimation results with industry fixed effects. Column 2 includes firm fixed effects. Across all these
columns, the coefficients on the interaction terms YEAR2011∗AFFECTED
and YEAR2013∗AFFECTED are negative and significant. These results are
consistent with our CAR results where we found that the most important
event dates in our study are in the fiscal year 2011 and 2013. Similar to
previously documented results, the economic magnitude of these coefficients is much larger than expected. The sign on coefficient estimates for
26 We obtain qualitatively similar results if we use one-year market adjusted buy hold returns
instead of Q ratio to measure the firm value in the longer run. These results are not tabulated
in this paper but are available upon request.
Coefficient
z-statistics
Observations
Polynomial order
Bins
Dependent Variable = ࢞Ln(Q)
0.017
(0.303)
481
1
25
YEAR 2009
−0.078
0.227
−0.030
−0.001
−0.062
2009
2010
2011
2012
2013
Panel B: Basic RDD analysis
Median
Mean
−0.033
0.203
−0.036
−0.010
−0.065
Year
UNAFFECTED (1)
Panel A: Univariate tests
Mean
−0.058
0.218
−0.074
0.025
−0.125
YEAR 2011
−0.100∗∗∗
(−2.729)
566
1
30
−0.027
(−0.870)
525
1
27
−0.025
0.015
−0.038∗∗
0.036∗∗
−0.061∗∗∗
(2) – (1)
0.021
(0.554)
546
1
26
YEAR 2012
Difference in Mean
YEAR 2010
−0.111
0.243
−0.128
0.042
−0.160
Median
AFFECTED (2)
࢞ Ln(Q)
TABLE 7
Changes in Tobin’s Q for the RDD Sample
(Continued)
−0.097∗∗∗
(−3.183)
546
1
29
YEAR 2013
−0.033
0.015
−0.098∗∗∗
0.044∗∗
−0.098∗∗∗
(2) – (1)
Difference in Median
1292
H. MANCHIRAJU AND S. RAJGOPAL
YEAR 2010 ∗ AFFECTED
YEAR 2011 ∗ AFFECTED
YEAR 2012 ∗ AFFECTED
YEAR 2013 ∗ AFFECTED
M ∗ AFFECTED
M
M∗M
M∗M∗M
R1
R2
R3
࢞SIZE
࢞LEV
࢞ROA
࢞CAPEX
࢞SGROWTH
࢞CASH
࢞BOARD INDPENDENCE
BIG4
BG
GOVT OWNED
Panel C: RDD analysis with controls
τ1
τ2
τ3
τ4
δ
β1
β2
β3
γ1
γ2
γ3
λ1
λ2
λ3
λ4
λ5
λ6
λ7
λ8
λ9
λ10
−0.001
−0.062∗∗∗
0.000
−0.072∗∗∗
−0.054
0.082
−0.039
−0.137
−0.000
−0.003
0.012
−0.073∗
0.377∗∗∗
0.062
−0.016
0.022
0.200∗∗
0.061∗∗∗
0.016
−0.012
−0.027
coef
(1)
T A B L E 7—Continued
−0.036
−3.395
0.029
−4.093
−0.426
0.879
−0.224
−1.135
−0.053
−0.987
1.479
−1.669
9.990
0.375
−0.453
0.943
2.110
2.652
1.221
−1.130
−0.334
tstat
−0.037
−0.115∗∗∗
0.001
−0.081∗∗
0.117
−0.074
−0.206
0.244
−0.001
−0.001
−0.004
−0.094∗∗∗
0.341∗∗∗
0.115
−0.019
0.014
0.210∗∗
0.082
0.023
−0.091∗∗
coef
Dependent Variable = ࢞Ln(Q)
(2)
(Continued)
−1.170
−3.412
0.030
−2.191
0.351
−0.362
−0.632
0.632
−1.032
−0.053
−0.142
−2.975
4.418
1.008
−0.346
0.650
2.035
0.982
0.488
−2.151
tstat
IMPACT OF CSR ON SHAREHOLDER VALUE
1293
j =1
8
τ j = 2% F-stat (p-value)
λ11
λ12
λ13
λ14
coef
0.034
1.234∗
−0.006
0.067
4.40∗∗
(0.04)
3.19∗
(0.07)
Yes
Yes
No
2,370
0.290
(1)
1.336
1.913
−0.331
0.823
tstat
5.09∗∗
(0.02)
4.25∗∗
(0.04)
Yes
No
Yes
2,370
0.616
0.086
2.015
coef
Dependent Variable = ࢞Ln(Q)
(2)
j
τ j Y E AR j ∗ AF F E CT E D + δM ∗ AF F E CT E D +
n
βn M ∧n +
i
γi Ri +
k
λk Contr ol j + ε.
tstat
0.515
1.397
Q is defined as book value of total assets plus market value of equity less book value of equity, divided by the book value of total assets. The estimation is restricted to the RDD
sample. The RDD sample is constructed as follows. First, a binding score rating variable M is created which is defined as the minimum of the three individual rating scores R1, R2,
and R3, which are in turn defined as (Profit – 50)/50, (Book value – 5,000)/5,000, and (sales – 10,000)/10,000, and which perfectly determines whether a firm is affected by the
mandatory CSR rule or not. Profit, book-value and sales for the most recent fiscal year end before any event date are considered to eliminate any perfect foresight bias. The INR
50 million profits, INR 5,000 million book value, and INR 10,000 sales are the three cutoffs specified in the mandatory CSR rule. If a firm has profit, book value, or sale above these
cutoffs, the firm is then required to spend on CSR. Firms with the variable M ranging between –0.50 and 0.50 comprise our RDD sample, where firms with M > 0 are AFFECTED by
the rule and firms with M < 0 are UNAFFECTED by the rule. Control variables are defined in appendix A and are measured at the end of the year. The significance of differences
in means and medians reported in panel A are evaluated based on the t-test and Wilcoxon test, respectively (p-values for the t-statistics and Z-statistics are two-tailed). The t-statistics
reported in panel B are based on robust standard clustered at firm and year levels. For brevity, coefficient estimates of the intercept, year dummies, and industry dummies are not
reported. ∗∗∗ , ∗∗ , and ∗ denote significance at 1%, 5%, and 10% level (two-tailed), respectively.
Ln(Q ) = α +
Panel A reports changes in the mean and median values of ࢞Ln(Q) for UNAFFECTED and the AFFECTED firms. Panel B reports results of estimating model (6):
Year FE
Industry FE
Firm FE
Observations
R2
Test :
MNC
࢞AD
POLITICAL
POLLUTED
8
Test : j =1 τ j = 0 F-stat (p-value)
Panel C: RDD analysis with controls
T A B L E 7—Continued
1294
H. MANCHIRAJU AND S. RAJGOPAL
IMPACT OF CSR ON SHAREHOLDER VALUE
1295
the firm-specific controls variables are consistent with prior research (e.g.,
Black and Kim [2012]). Specifically, we find that leverage and liquidity are
positively associated with Q ratio, whereas firm size and business group affiliation (proxy for agency costs) are negatively associated with Q.
Overall, the results of this section indicate that during the years 2011
and 2013, which coincide with a greater likelihood of the passage of the
mandatory CSR rules, AFFECTED firms experienced a decline in Q, relative
to UNAFFECTED firms. These results are robust to alternate measures of
long-run firm value such as annual buy hold market adjusted stock returns.
4.7
ADDITIONAL TESTS AND ROBUSTNESS CHECKS
In our additional analysis, we evaluate whether the impact of mandatory
CSR rule on shareholder value varies depending on the financial reporting quality of a firm. It can be argued that official guidance related to the
implementation of the mandatory CSR rule is still emerging and not fully finalized. For example, the Companies Act does not define CSR. Instead, the
Act lays out a list of possible activities that would be considered CSR. Further, the mandatory CSR rule is a “comply or explain” rule. Hence firms
might game the system and meet the requirements by changing their reporting techniques or by reclassifying existing activities as CSR activities. If
that is the case, the market reaction for such misreporting firms is likely
to be different from the market reaction for other firms. We classify firms
with higher value of discretionary accruals (either signed or unsigned)
calculated as per the modified Jones [1991] model and/or higher frequency of related party transactions (as suggested by Bertrand, Mehta, and
Mullainathan [2002]) as the ones that are more likely to adjust their reporting techniques. Our un-tabulated results show that there is no crosssectional variation in the stock price reaction for the AFFECTED and UNAFFECTED firms conditioned on the level of discretionary accruals and/or
frequency of related party transactions.
Next, given the comply-or-explain nature of the law, we examine whether
our results do indeed capture the “comply” part rather than the “explain”
part of the law. The treatment studied in our setting is a mandatory requirement to invest in CSR to the tune of 2% of profits and, failing that,
to explain noncompliance. Given the comply-or-explain nature of the law,
we may have effectively documented the effect of having to explain why the
firm does not comply. To address this possibility, we rely on data provided
by INDIA CSR, a leading CSR news portal in India.27 The portal’s report titled India CSR outlook (2016) provides initial evidence on how the mandatory CSR rule has played out in practice among the 250 Indian firms based
on market capitalization.28 The key findings of the report are that (i) on
aggregate, the prescribed CSR spending for these firms was INR 71.43 billion (or US$ 1.19 billion), whereas these firms actually spent INR65.78
27 http://indiacsr.in/about-us/about-india-csr/
28 http://indiacsr.in/wp-content/uploads/2016/09/CSR-Outlook-Report-2016-India.pdf
1296
H. MANCHIRAJU AND S. RAJGOPAL
billion (or US$ 1.10 billion), suggesting that firms, by and large, complied with the regulation; and (ii) education and healthcare attracted the
most CSR spending. These data suggest that the stock market’s perceptions
about the negative consequences of mandatory CSR are justified in that
firms were forced to spend 2% of their profits on social causes that are not
closely related to the firms’ operating activities.
We also conduct several robustness tests to assess the sensitivity of our results. In the event study/CAR-based analysis, we obtain similar results if we
(i) use buy-and-hold returns for the event window instead of summing daily
abnormal returns; (ii) use COSPI index and MSCI emerging market index
instead of CNX500 index as the proxy for market returns while estimating
the market model to calculate abnormal returns;29 (iii) use market adjusted
returns to capture abnormal returns instead of using the market model;
(iv) change the event window to (–5,+5), (–2,+2), or (0,+2) instead of
(–1,+1) while calculating CAR; and (v) exclude outliers instead of winsorizing them at 1% and 99%. We also obtain similar results in table 7 if we
estimate the regressions on a balanced panel where the same number of
firms are present throughout the event period.
5. Conclusion
This paper examines the effect of CSR activities on shareholder value.
Our identification comes from a unique regulatory change in India that
makes it mandatory for firms to spend 2% of their profits on CSR if their
profits/book value/sales exceed certain thresholds. This research setting
enables us to better isolate how stock market participants, in aggregate,
view the average effect of mandatory CSR activities on shareholder value.
We find that the CAR around key events leading to the passage of the
mandatory CSR rule is negative for firms affected by this regulation. On
average, the decline in the stock price of firms forced to spend money
on CSR accumulated over the eight important event dates is about 4.1%.
We also find that compared to firms unaffected by the mandatory rule,
firms affected by the mandatory CSR rule experience a greater decline in
Tobin’s q ratio in the years when the likelihood of the passage of mandatory
CSR rule increased. Overall, our evidence suggests that mandatory CSR activities can impose social burdens on business activities at the expense of
shareholders. Our findings also indicate that firms, left to their own devices, choose their optimal level of CSR spending designed to maximize
their firm value. Hence, we suggest caution in drawing causal inferences
from associations of corporate outcomes and CSR spending.
While the findings of this study are specific to India, they are potentially relevant to the policy debate of other emerging economies which are
29 COSPI is the CMIE’s Overall Share Price Index (COSPI) that comprises of all the stocks
in the CMIE universe.
IMPACT OF CSR ON SHAREHOLDER VALUE
1297
experiencing unprecedented CSR initiatives by their regulators due to concerns related to unequal economic growth and environmental abuses. An
important caveat of our study is that our analysis does not explore the social welfare implications of the mandatory CSR rule. Therefore, our study
is not a verdict on regulatory reforms related to CSR, which is inherently
a social welfare decision and involves numerous stakeholders, apart from
shareholders. Whether mandatory CSR achieved its objective of social welfare is an important question we intend to study in future work.
APPENDIX A
Variable Definitions
AFFECTED
UNAFFECTED
CAR
Q
SIZE
BM
LEV
indicator variable that equals one if the firm’s binding score
rating variable M > 0. The variable M is defined as the
minimum of the three individual rating scores R1, R2, and
R3, which are in turn defined as (Profit – 50)/50, (Book
value – 5,000)/5,000, and (sales – 10,000)/10,000, and
which determines whether a firm is affected by the
mandatory CSR rule or not. Profit, book-value and sales for
the most recent fiscal year end before any event date are
considered to eliminate any perfect foresight bias. The INR
50 million profits, INR 5,000 million book value, and INR
10,000 sales are the three cutoffs specified in the
mandatory CSR rule. If a firm has profit, book value, or
sales above these cutoffs, the firm is then required to spend
funds on CSR.
indicator variable that equals one if the firm’s binding score
rating variable M < 0.
cumulative abnormal returns where we measure abnormal
returns by estimating the market model using 200 trading
days of return data ending 11 days before key legislative
events related to the passage of the Companies Act 2013.
The return on National Stock Exchange’s CNX 500 index is
used as a proxy for the market return. Daily abnormal stock
returns are cumulated to obtain the CAR from day t – 1
before the legislative event date to day t + 1 after the event
date.
(book value of total assets + market value of equity –common
book equity)/ total assets
natural log of the market value of equity (market value of
equity = stock price at the end of fiscal year ∗ number of
shares outstanding)
ratio of book value of equity to the market value of equity at
the end of the year
long-term debt (including its current portion) divided by
total assets at the end of year
1298
H. MANCHIRAJU AND S. RAJGOPAL
SGROWTH
ROA
CAPEX
CASH
BG
MNC
GOVT OWNED
BOARD INDPENDENCE
BIG4
AD
HIAD
POLITICAL
POLLUTED
sales growth defined as (salest – salest –1 )/salest –1
income from continuing operations divided by the total assets
at the end of year
capital expenditure during the year/total assets at the end of
year
cash and marketable securities at the end of year/total assets
at the end of year
indicator variable that equals one if the firm belongs to a
business group as defined by the Prowess database, and
zero otherwise
indicator variable that equals one if the firm is a subsidiary of
multinational corporation, and zero otherwise
indicator variable that equals one if the firm is owned by the
central or a state Government, and zero otherwise
number of independent directors on the board/total number
of directors on the board
indicator variable that equals one if the firm engages a Big 4
auditor, and zero otherwise
advertisement expenses divided by the sales during the year
indicator variable that equals one if advertisement spending
AD is in the top quartile for the industry-year distribution,
and zero otherwise
indicator variable that equals one if the firm or the business
group to which a firm belongs has made a contribution of
INR20,000 or more to a political party in India between
2005 and 2012, and zero otherwise
indicator variable that equals one if the firm belongs to an
industry identified by Ministry of Environment and Forests,
Government of India as heavily polluting industries, and
zero otherwise
REFERENCES
AHERN, K. R., AND A. K. DITTMAR. “The Changing of the Boards: The Impact on Firm Valuation
of Mandated Female Board Representation.” The Quarterly Journal of Economics 127 (2012):
137–97.
AKEY, P. “Valuing Changes in Political Networks: Evidence from Campaign Contributions to
Close Congressional Elections.” The Review of Financial Studies (2015): Forthcoming.
BARON, D. P. “Private Politics, Corporate Social Responsibility, and Integrated Strategy.” Journal
of Economics and Management Strategy 10 (2001): 7–45.
BENABOU, R., AND J. TIROLE. “Individual and Corporate Social Responsibility.” Economica 77
(2010): 1–19.
BERTRAND, M.; P. MEHTA; AND S. MULLAINATHAN. “Ferreting out Tunneling: An Application to
Indian Business Groups.” Quarterly Journal of Economics 117 (2002): 121–48.
BLACK, B. S., AND V. S. KHANNA. “Can Corporate Governance Reforms Increase Firm Market
Values? Event Study Evidence from India.” Journal of Empirical Legal Studies 4 (2007): 749–96.
BLACK, B. S., AND W. KIM. “The Effect of Board Structure on Firm Value: A Multiple Identification Strategies Approach Using Korean Data.” Journal of Financial Economics 104 (2006):
203–226.
CALONICO, S.; M. D. CATTANEO; AND R. TITIUNIK. “Robust Nonparametric Confidence Intervals for Regression-Discontinuity Designs.” Econometrica 82 (2014a): 2295–326.
CALONICO, S.; M. D. CATTANEO; AND R. TITIUNIK. “Robust Data-Driven Inference in the
Regression-Discontinuity Design.” The Stata Journal 14(4) (2014b): 909–46.
IMPACT OF CSR ON SHAREHOLDER VALUE
1299
CHENG, H.; H. HONG; AND K. SHUE. “Do Managers Do Good with Other People’s Money?”
2013. Available at http://ssrn.com/abstract=1962120.
CHHAOCHHARIA, V. AND Y. GRINSTEIN. “Corporate Governance and Firm Value: The Impact of
the 2002 Governance Rules.” The Journal of Finance 62 (2007): 1789–1825.
CHUGH, R. “Tax Twists in CSR Spends.” Financial Express, 2014. Available at
http://archive.financialexpress.com/story-print/1231302.
COASE, R. H. “The Nature of the Firm.” Economica 4 (1937): 386–405.
COCHRAN, P. L., AND R. A. WOOD. “Corporate Social Responsibility and Financial Performance.” Academy of Management Journal 27 (1984): 42–56.
DEMSETZ, H., AND K. LEHN. “The Structure of Corporate Ownership: Causes and Consequences.” Journal of Political Economy 93 (1985): 1155–77.
DHALIWAL, D. S.; O. Z. LI; A. TSANG; AND Y. G. YANG. “Voluntary Nonfinancial Disclosure and
the Cost of Equity Capital: The Initiation of Corporate Social Responsibility Reporting.” The
Accounting Review 86 (2011): 59–100.
DI GIULI, A., AND L. KOSTOVETSKY. “Are Red or Blue Companies More Likely to go Green?
Politics and Corporate Social Responsibility.” Journal of Financial Economics 111 (2014): 158–
80.
FACCIO, M. “Politically Connected Firms.” The American Economic Review 96 (2006): 369–86.
FAMA, E. F., AND K. R. FRENCH. “The Cross-Section of Expected Stock Returns.” The Journal of
Finance 47 (1992): 427–65.
FISMAN, R.; G. HEAL; AND V. B. NAIR. “A Model of Corporate Philanthropy.” Working paper,
Columbia University, 2008.
FLAMMER, C. “Does Corporate Social Responsibility Lead to Superior Financial Performance?
A Regression Discontinuity Approach.” Management Science (2015): Forthcoming.
FOMBRUN, C. J. “Building Corporate Reputation Through CSR Initiatives: Evolving Standards.”
Corporate Reputation Review 8 (2005): 7–11.
FREEMAN, R. E. Strategic Management: A Stakeholder Approach. Boston: Pittman-Ballinger, 1984.
FRIEDMAN, M. “The Social Responsibility of Business is to Increase its Profits.” The New York
Times, September 13, 1970, pp. 32–33 & 122 & 126.
GODFREY, P. C. “The Relationship between Corporate Philanthropy and Shareholder Wealth:
A Risk Management Perspective.” Academy of Management Review 30 (2005): 777–98.
GOPALAN, R.; V. NANDA; AND A. SERU. “Affiliated Firms and Financial Support: Evidence from
Indian Business Groups.” Journal of Financial Economics 86 (2007): 759–95.
GOW, I. D.; D. F. LARCKER; AND P. C. REISS. “Causal Inference in Accounting Research.” Journal
of Accounting Research 54(2) (2016): 477–523.
GREENING, D. W., AND D. B. TURBAN. “Corporate Social Performance as a Competitive Advantage in Attracting a Quality Workforce.” Business & Society 39 (2000): 254–80.
GRIFFIN, J. J., AND J. F. MAHON. “The Corporate Social Performance and Corporate Financial
Performance Debate Twenty-Five Years of Incomparable Research.” Business & Society 36
(1997): 5–31.
HIMMELBERG, C. P.; R. G. HUBBARD; AND D. PALIA. “Understanding the Determinants of Managerial Ownership and the Link between Ownership and Performance.” Journal of Financial
Economics 53 (1999): 353–84.
HONG, H.; J. D. KUBIK; AND J. A. SCHEINKMAN. “Financial Constraints on Corporate Goodness.”
2012. Available at http://ssrn.com/abstract=1734164.
ILIEV, P. “The Effect of SOX Section 404: Costs, Earnings Quality, and Stock Prices.” The Journal
of Finance 65 (2010): 1163–96.
IMBENS, G. W., AND K. KALYANARAMAN. “Optimal Bandwidth Choice for the Regression Discontinuity Estimator.” The Review of Economic Studies 79(3) (2012): 933–59.
IMBENS, G. W., AND T. LEMIEUX. “Regression Discontinuity Designs: A Guide to Practice.” Journal of Econometrics 142 (2008): 615–35.
JENSEN, M. C., AND W. H. MECKLING. “Theory of the Firm: Managerial Behavior, Agency Costs,
and Ownership Structure.” Journal of Financial Economics 3 (1976): 305–60.
JONES, J. “Earnings management during import relief investigations.” Journal of Accounting Research 29 (1991): 193–228.
1300
H. MANCHIRAJU AND S. RAJGOPAL
KHANNA, T., AND K. PALEPU. “Is Group Affiliation Profitable In Emerging Markets? An analysis
of Diversified Indian Business Groups.” The Journal of Finance 55 (2000): 867–91.
KITZMUELLER, M., AND J. SHIMSHACK. “Economic Perspectives on Corporate Social Responsibility.” Journal of Economic Literature 50 (2012): 51–84.
LEE, D. S., AND T. LEMIEUX. “Regression Discontinuity Designs in Economics.” Journal of Economic Literature 48 (2010): 281–355.
LYS, T. Z.; J. P. NAUGHTON; AND C. WANG. “Signaling Through Corporate Accountability Reporting.” Journal of Accounting and Economics 60(1) (2015): 56–72.
MARGOLIS, J. D.; H. A. ELFENBEIN; AND J. P. WALSH. “Does it Pay to be Good . . . and Does it
Matter?” A meta-analysis of the relationship between corporate social and financial performance, 2009. Available http://ssrn.com/abstract=1866371.
MARGOLIS, J. D., AND J. P. WALSH. “Misery Loves Companies: Rethinking Social Initiatives by
Business.” Administrative Science Quarterly 48 (2003): 68–305.
MCWILLIAMS, A., AND D. S. SIEGEL. “Corporate Social Responsibility: A Theory of the Firm
Perspective.” Academy of Management Review 26 (2001): 117–27.
NAVARRO, P. “Why do Corporations give to Charity?” The Journal of Business 61 (1998): 65–93.
NICHOLS, A. “Causal Inference with Observational Data.” The Stata Journal 7 (2007): 507–41.
ORLITZKY, M., AND J. D. BENJAMIN. “Corporate Social Performance and Firm Risk: A MetaAnalytic Review.” Business & Society 40 (2001): 369–96.
ORLITZKY, M.; F. L. SCHMIDT; AND S. L. RYNES. “Corporate Social and Financial Performance:
A Meta-Analysis.” Organization Studies 24 (2003): 403–41.
REARDON, S. F., AND J. P. ROBINSON. “Regression Discontinuity Designs with Multiple RatingScore Variables.” Journal of Research on Educational Effectiveness 5(1) (2012): 83–104.
ROBERTS, M. R., AND T. M. WHITED “Endogeneity in Empirical Corporate Finance.” 2012. Available at http://ssrn.com/abstract=1748604.
SEN, S., AND C. B. BHATTACHARYA. “Does Doing Good Always Lead to Doing Better? Consumer
Reactions to Corporate Social Responsibility.” Journal of Marketing Research 38 (2001): 225–
43.
SERVAES, H., AND A. TAMAYO. “The Impact of Corporate Social Responsibility on Firm Value:
The Role of Customer Awareness.” Management Science 59 (2013): 1045–61.
SHLEIFER, S., AND R. W. VISHNY. “Politicians and Firms.” The Quarterly Journal of Economics 109
(1994): 995–1025.
THISTLETHWAITE, D. L., AND D. T. CAMPBELL. “Regression Discontinuity Analysis: An Alternate
to the Ex-Post Facto Experiment.” Journal of Educational Psychology 51(6) (1960): 309–17.
TIROLE, J. “Corporate Governance.” Econometrica 69 (2001): 1–35.
WONG, V. C.; P. M. STEINER; AND T. D. COOK. “Analyzing Regression-Discontinuity Designs with
Multiple Assignment Variables: A Comparative Study of Four Estimation Methods.” Journal
of Educational and Behavioral Statistics 38(2) (2013): 107–41.