Does Corporate Social Responsibility (CSR) Create .

1y ago
763.28 KB
52 Pages
Last View : 1m ago
Last Download : 1m ago
Upload by : Farrah Jaffe

Does corporate social responsibility (CSR) create shareholder value?Exogenous shock-based evidence from the Indian Companies Act 2013Hariom ManchirajuAssistant ProfessorDepartment of AccountingIndian School of BusinessHyderabad, Telangana, India 500032Email: Hariom [email protected] RajgopalSchaefer Chaired Professor of AccountingGoizueta Business SchoolEmory University1300 Clifton Road NE, Atlanta, GA 30030Email: [email protected] welcomeApril 27, 2015ABSTRACT:In 2013, a new law required Indian firms, which satisfied certain size and profitabilitythresholds, to spend at least 2% of their net income on CSR. We exploit this natural experimentto isolate the shareholder value implications of CSR activities. Using several identificationstrategies, including an event study, regression discontinuity design, difference-in-differencetests, and instrument variable approach, we find that the law caused a significant drop in thestock price of firms forced to spend money on CSR, consistent with the idea that firmsvoluntarily choose CSR levels to maximize firm value. Firms with greater agency costs andpolitical connections benefit from mandatory CSR. Our results potentially clarify the directionof causality underlying decades of mixed findings on the association between CSR and firmvalue.We thank Bernie Black, Sanjay Kallapur, Simi Kedia, Yongtae Kim, Suresh Radhakrishnan, SriniRangan, K.R.Subramanyam, Hao Zhang and workshop participants at the Indian School of Business forhelpful comments. We also thank Ranjan Nayak and Aadhaar Verma for providing excellent researchassistance. Finally, we thank our respective schools for financial support.Electronic copy available at: 2602960

Does corporate social responsibility (CSR) create shareholder value?Exogenous shock-based evidence from the Indian Companies Act 20131. INTRODUCTIONCorporate social responsibility (CSR) is now mandatory in India. According to theClause 135 of the Companies Act (mandatory CSR rule, hereafter), passed by the IndianParliament in 2013, a firm is required to spend 2% of its average net profits of the last three yearson CSR activities, if during any fiscal year, it 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. 167million) or more; or (3) a net profit of INR 50 million (about U.S. 0.83 million) or more.1 Alegislative mandate forcing corporations to spend funds on CSR activities is perhaps the first ofits kind in the world. We exploit this natural experiment to isolate the impact of the mandatoryCSR rule on shareholder value.There are two competing theoretical views on whether CSR affects firm value. The“shareholder expense” view, advocated most notably by Milton Friedman (1970), asserts that“the social responsibility of business is to increase its profits” and hence argues that CSRdestroys shareholder value, either because (i) it constitutes moral hazard in that managers’ selfinterest drives CSR spending at the expense of shareholders; or (ii) even absent moral hazard,CSR is a sacrifice of the firm’s profits in the social interest (Reinhardt, Stevins and Vietor 2008).The contrasting view, labeled here as the “stakeholder value maximization” view,following the “doing well by doing good” theory advanced in the management literature, arguesthat strategic CSR spending can increase firm value. The intuition is that a firm’s self-interestedfocus on stakeholders’ interests increases stakeholders’ willingness to support the firm’s1An exchange rate of 60INR 1US is assumed for these conversion of INR to US .1Electronic copy available at: 2602960

operations in several ways (Kitzmueller and Shimshack 2012). Following this stakeholder view,studies have documented that a high commitment to CSR activities is associated with attractingand retaining higher quality employees (Greening and Turban, 2000), improving theeffectiveness of the marketing of products and services (Fombrun, 2005), increasing demand forproducts and services (Navarro, 1988), providing superior access to valuable resources (Cochranand Wood, 1984), generating moral capital or goodwill that tempers punitive actions byregulatory agencies during a negative event (i.e. an insurance effect) and thereby preserves firmvalue (Godfrey 2005).Existing empirical evidence on whether CSR creates shareholder value is inconclusivedespite nearly four decades of research efforts, partly because these studies are clouded bymethodological concerns such as potential endogeneity, reverse causality or omitted variableproblems (Margolis, Elfenbein, and Walsh, 2009).2 For instance, the choice to conduct CSR isvoluntary and assuming firms spend their optimal level of CSR, on average, there ought to be noassociation between future firm performance and CSR in the cross-section. Hence, it is difficultto ascertain whether the observed associations between CSR and firm performance are causal innature or merely attributable to model misspecification due to the influence of unobserved firmlevel heterogeneity related to CSR (Himmelberg, Hubbard and Palia 1999). Second, ashighlighted by Hong, Kubik and Scheinkman (2012), reverse causality might drive the results asfirms that are doing well, and are hence less financially constrained, might be the ones spendingon CSR activities. Hence, firm performance could cause higher future CSR, as opposed to theother way around. Several extant studies suffer from this limitation, as pointed out by Margolis,Elfenbein and Walsh. (2009). Third, as Lys, Naughton and Wang (2015) suggest, CSR might2For reviews of the literature on CSR, see Griffin and Mahon (1997), Orlitzky and Benjamin (2001), Orlitzky,Schmidt and Rynes. (2003), Margolis and Walsh (2003), Margolis, Elfenbein, and Walsh (2009) and Kitzmuellerand Shimshack (2012).2

merely signal future profits, as opposed to causing them. In sum, the correlation between CSRand firm value or firm performance documented by hundreds of earlier studies, albeit interesting,does not necessarily warrant a causal interpretation.To overcome these inferential problems, one would ideally like to find an exogenousexperiment in which firms are randomly assigned to spend money on CSR or not. This wouldallow us to compare the outcomes of the treated firms with those of the non-treated firms and tohence attribute any differences in outcomes to their CSR spending. Remarkably, a fairly closeversion of that ideal experiment exists. The mandatory CSR rule of the Indian Companies Actrepresents an exogenous shock in the sense that it assigns firms in two groups: (i) firms that aremandatorily affected by these new CSR rule (treatment group); and (ii) firms that are notimpacted by the CSR rule (control group). Comparing the stock returns of treatment group tothose of control group (controlling for other firm attributes that are likely to affect returns)around the events that changed the probability of the passage of the Act can thus provide areliable basis for causal inference.3 Further, the numerical eligibility thresholds for the CSRprovision, based on reported net worth, sale or profits, also enable us to employ a regressiondiscontinuity design that compares the firms that were just above the rule cutoff and have tocomply with the CSR requirement with those that were just below the cutoff and did not have tocomply.If firms choose CSR to maximize their firm value, imposing binding legal constraints ontheir CSR choices will lead to declines in their values (Demsetz and Lehn 1985). In contrast, iffirms spent sub-optimally on CSR or if their CSR activities were not aligned with the stakeholder3Atanasov and Black (2014), in the survey of 863 studies examining the effect of governance on firm value, arguethat credible inference strategies usually rely on “shocks” to governance. In contrast, only a few recent papers relyon instrumental variables to address the causality problem (e.g., Cheng, Ioannou, and Serafeim, 2011; Deng, Kangand Low 2013).3

interests, the new mandatory CSR rule will lead to increase in firm value. Hence, if themandatory CSR rule has a positive (negative) effect on shareholder value, then all else equal, weexpect the treatment group to report higher (lower) announcement period stock returns relative tothe control group. Furthermore, among the treatment group, firms differ on the extent to whichthey spend on CSR and therefore vary on the extent to with they comply with the proposed CSRspending norms. If the CSR law has a positive (negative) effect on shareholder value, we expectthe firms that are less (more) compliant to report higher (lower) announcement period stockreturns than firms that are more (less) compliant.We test these predictions using a sample of approximately 2,100 firms listed on theNational Stock Exchange (NSE) of India. About 1,237 (58%) of these firms are affected by themandatory CSR rule and within this subsample, 458 firms currently spend on CSR activitieswhereas 779 firms do not currently spend on CSR activities. The remaining 883 firms (42%) ofour sample are not affected by the CSR rule. We label these three groups of firms as SPENDER,NONSPENDER, and UNAFFECTED, respectively. Our baseline results indicate that for thegroup of firms affected by the mandatory CSR rule, the median three-day cumulative abnormalreturns (CAR) around key events leading to the passage of the Companies Act ranges from 1.3% to -2.4%. Further, the negative returns are more pronounced for NONSPENDER firms,compared to SPENDER firms. In contrast, the average CAR around the same event dates for theUNAFFECTED firms is generally insignificant. These results suggest that the optimal amount ofCSR spending for NONSPENDER firms is indeed zero.These results are robust to a more rigorous regression discontinuity design (RDD). Whileapplying RDD, we use the INR 50 million of net income as the threshold because the CompaniesAct stipulates a firm with net income above INR 50 million to spend on CSR. Firms with net4

income between INR 50-75 million represent our treatment group and firms with net incomebetween INR 25-50 million constitute our control group. RDD assumes that the firms in thetreatment and control group are similar, expect for the variable that assigns firms in these twogroups and we confirm this assumption in our data. Therefore, any differences in the observedthree-day CAR for the SPENDER, NONSPENDER and UNAFFECTED firms can thus beattributed to the mandatory CSR rule that creates the discontinuity. Results of our RDD analysisindicate that the average three-day CAR around the event dates for the NONSPENDER firms arenegative, consistent with our earlier findings of a negative impact on CSR on shareholder value.Motivated by the “stakeholder value maximization” perspective of the managementliterature, we perform four cross-sectional analyses to identify conditions under whichmandatory CSR rule is likely to affect firm value differentially. Such an analysis also helps usmitigate omitted variable concerns that other macro shocks that possibly took effect during thekey event dates potentially drive our baseline results. Specifically, we examine whether theimpact of CSR on shareholder value varies depending on firm’s spending on agency costs facedby the firm, its political connections, its advertisement spending, and its affiliation to a highlypolluting industry.First, to the extent the increased reporting and governance requirements imposed by thelaw make firms redirect their CSR spending to maximize firm value, either by preventingmanagers from using CSR as an entrenchment device or as a means to satisfy their own socialpreferences, firm values will increase after the passage of the CSR rule as agency costs arereduced. To test this prediction we classify firms that belong to business group as the ones thatface high levels of agency costs. This classification is consistent with Bertrand, Mehta andMullainathan (2002) who view business groups as a structure that facilitates “tunneling” of funds5

from minority outside shareholders to group insiders. Consistent with our expectations, we findthat the average three-day CAR around the event dates for the SPENDER and NONSPENDERfirms is positive if such firms are affiliated to business groups.Second, political connections can be valuable to a firm in several ways (Faccio, 2006)However, firm value will be enhanced only when the marginal benefits of the connectionsoutweigh their marginal costs (rents extracted by politicians). CSR spending might constitute apotential mechanism via which a firm can satisfy the preferences of politicians and increase itsability to conduct business with the government and other entities with their politicalconnections. Therefore, we expect the positive (negative) effect of mandatory CSR rule on firmvalue to be higher (lower) among firms that are politically connected. We consider a firm to bepolitically connected if the firm or the business group to which a firm belongs has made acontribution of INR 20,000 or more to a political party in India during 2005-2012. Consistentwith our expectations, we find that the average three-day CAR around the event dates for theNONSPENDER firms is positive if such firms are politically connected.Third, Servaes and Tamayo (2013) find that CSR activities and firm value are positivelyrelated for firms with high customer awareness, but not for firms with low customer awareness.They argue that advertising expenditure enhances a firm’s information environment, therebyincreasing awareness among the firm’s potential customers. This, in turn, is likely to increasechances of a consumer being associated with the product. Therefore we expect the positive(negative) effect of mandatory CSR rule on firm value to be higher (lower) among firms withhigh levels of advertisement spending. We don’t find results consistent with our expectations.Finally, firms in polluting industries are less likely to enjoy the net benefits of CSRactivities because pressure from activists groups and the Government might force a firm to invest6

in green technologies. Such investments, while potentially beneficial to the environment, mightbe associated with limited firm-specific benefits. Thus we expect the positive (negative) effectof mandatory CSR rule on firm value to be lower (higher) among these firms. Again, we don’tfind results consistent with our expectations.We also conduct a difference-in-difference analysis as an alternative way to assesswhether the mandatory CSR rule predicts a decline in value for the affected firms, right at thetime when the mandatory CSR rule was adopted. If investors consider CSR activities asdetrimental to firm value, then the Tobin’ q ratio of the affected firms should decline more,relative to that of unaffected firms, in the years when the likelihood of the passage of mandatoryCSR rule increased. Consistent with our expectations, we find that in the years 2011 and 2013,the decline in Tobin’s q ratio was 13.8% and 9.5% more for the SPENDER firms, relative to thebenchmark UNAFFECTED firms. Similarly, during the years 2011 and 2013, the decline inTobin’s q ratio was 22.5% and 29.6% more for the NONSPENDER firms, relative to thebenchmark UNAFFECTED firms.These results relating to the effect of CSR on Tobin’s Q ratio are also robust to aninstrument variable (IV) based regression technique. We use pre-law variation in the CSRspending and low levels of financial flexibility as IVs for CSR spending. Whether a firm spendson CSR in the year 2008 (the year before the new law was introduced) is likely to be correlatedwith whether a firm spends on CSR during the years 2009-2013 but is not directly relatedchanges in firm value in the years 2009-2013. Similarly, low financial flexibility is likely toinfluence a firm’s CSR spending (Hong, Kubik, and Scheinkman, 2012) but is unlikely to affectchanges in firm value, especially because, in our second stage regressions, we control for thecontinuous effect of changes in cash holdings on changes in firm value. These IV regression7

results mitigate the possibility that firms might have anticipated the mandatory CSR rule, andmight have endogenously altered their CSR policies.Overall, our results suggest that, on average, the mandatory CSR rule imposesstatistically significant net costs on firms that are required to comply with this regulation, leadingto declines in shareholder value. These costs can arise due to variety of factors includingincreased pressure from the Government on the businesses to pick up the tab for social activitiesprescribed the new CSR law, increased compliance costs associated with reporting andmonitoring CSR activities or the use of scarce managerial time and effort in these activities thatdo not add shareholder value. However, under certain situations such as presence of agencycosts and political connections, mandatory CSR spending can be valuable to the shareholders.We make several contributions to the literature. Our primary contribution is to presentclean evidence on the impact of CSR spending on firm value using a natural experiment. Usingmultiple identification strategies such as an event study, regression discontinuity design,difference-in-difference analysis and instrument variable approach, we document a negativerelation between CSR and shareholder value. Second, what distinguishes our study from vastprior literature in this area is that while a majority of prior studies consider voluntary CSR, to thebest of our knowledge, we are the first to examine the impact of mandatory CSR spending.4 Ourresults suggest that mandating CSR activities, at least of the kind prescribed by the new CSR ruleof the Indian Companies Act 2013, internalizes the social costs to the firms and does not lead tovalue maximization for shareholders. These results thus suggest that firms voluntarily choselevels of CSR designed to maximize firm value. Consistent with Himmelberg, Hubbard andPalia (1999), studies that find an association between corporate outcomes (e.g., higher operating4Several papers (e.g. Hung and Wang 2014) examine the effect of mandatory disclosures of CSR activities (notspending) on shareholder value.8

performance, lower cost of capital or lower earnings management) and CSR need to grapple withthe possibility that are confounded by omitted variables. Our study suggests that there ought tobe no association between such outcomes and CSR given that firms, on average, pick CSR levelsoptimally to maximize firm value.Finally our study complements prior research that examines the effect of firm-specificcharacteristics on the relationship between CSR and shareholder value. Our study is related tothe work by Di Giuli and Kostovetsky (2014) who find that CSR activities motivated by thepolitical affiliation of stakeholders come at the expense of firm value. In contrast, we argue thatin developing economies like India where political connections can be valuable, CSR activitiescan be used to enhance political ties and therefore lead to increase in shareholder value.The rest of the paper is organized as follows. Section 2 discusses the institutionalbackground. Section 3 discusses related literature and our empirical predictions. Section 4presents our sample and data. Section 5 describes the empirical results and section 6 concludesthe paper.2. BACKGROUND OF THE COMPANIES ACT 2013India’s Companies Act, 2013 was enacted on 29th August 2013. This legislation attemptsto overhaul a nearly sixty year old Indian corporate law framework to bring it in line with globalbest practices and aims to provide a healthy regulatory environment for the businesses to grow.A unique feature of this Act is clause 135 that requires all firms of a certain size (measured interms of sales or net worth) or profitability are required to spend 2% of their average

Does corporate social responsibility (CSR) create shareholder value? Exogenous shock-based evidence from the Indian Companies Act 2013 1. INTRODUCTION Corporate social responsibility (CSR) is now mandatory in India. According to the Clause 135 of the Companies Act (mandatory CSR rule, hereafter), passed by the Indian