- Email: [email protected]
Does community environment matter to corporate social responsibility?
ARTICLE IN PRESS
JID: FRL
[m3Gsc;May 25, 2016;16:32]
Finance Research Letters 0 0 0 (2016) 1–9
Contents lists available at ScienceDirect
Finance Research Letters journal homepage: www.elsevier.com/locate/frl
Does community environment matter to corporate social responsibility? Dejun Wu a,1,∗, Chen Lin b, Sibo Liu b a b
School of Accounting, Zhongnan University of Economics and Law, Wuhan, China School of Economics and Finance, The University of Hong Kong, Hong Kong
a r t i c l e
i n f o
Article history: Received 20 November 2015 Revised 18 March 2016 Accepted 3 April 2016 Available online xxx JEL classification: G02 G12 G30 G32
a b s t r a c t How does the preference of local stakeholders shape firm’s corporate social responsibility (CSR)? Using a twenty-year U.S. sample, this paper examines the effects of the community demographic and cultural environment on CSR. We find that firms located in counties with more senior residents and/or higher levels of religiosity display higher degree of corporate social responsibility. In addition, we find the positive effects of local seniors and religiosity are stronger for less financial constrained firms. The result is robust after controlling for firm fixed effects, operating state trends and incorporation state trends. Taken together, our evidence supports a stakeholder theory of corporate social responsibility. © 2016 Published by Elsevier Inc.
Keywords: Corporate social responsibility Stakeholders Community Seniors Religiosity
1. Introduction Corporate social responsibility (CSR) has been embraced by a growing number of firms and gained much public attention in recent years. Most US firms nowadays practice corporate social responsibility in some forms. According to a report1 by KPMG, 86% of US firms issue social responsibility reports. Some large firms invest hundreds of millions of dollars annually on altruistic endeavors related to employee or community development programs (Hong et al., 2012). Practitioners are assessing the possible channels that CSR can add value to the firm. A survey2 conducted by McKinsey & Company suggests most of CFOs perceive that CSR increases firms’ value. In particular, meeting society’s expectations is an important channel. CSR have been discussed in academic studies for decades. The debate focuses on why firm would invest significant resources on CSR. According to some studies, CSR is considered as a crucial corporate behavior that deals with firms’ relationship with their stakeholders, such as employees, communities and etc. The pro-social behavior that provide value to firms stakeholders (Benabou and Tirole, 2010) and therefore gain the support from them. It has been documented that firm’s CSR ∗
1 1 2
Corresponding author. E-mail addresses: [email protected] (D. Wu), [email protected] (C. Lin), [email protected] (S. Liu). Dejun Wu acknowledges financial support from the Education Ministry of China (Project No. 14YJA630067). The KPMG Survey of Corporate Responsibility Reporting 2013 Valuing corporate social responsibility: McKinsey Global Survey Results.
http://dx.doi.org/10.1016/j.frl.2016.04.010 1544-6123/© 2016 Published by Elsevier Inc.
Please cite this article as: D. Wu et al., Does community environment matter to corporate social responsibility? Finance Research Letters (2016), http://dx.doi.org/10.1016/j.frl.2016.04.010
JID: FRL 2
ARTICLE IN PRESS
[m3Gsc;May 25, 2016;16:32]
D. Wu et al. / Finance Research Letters 000 (2016) 1–9
level matters for debt financing (Goss and Roberts, 2011), equity financing (El Ghoul et al., 2011), credit rating (Attig et al., 2013) and so on. The stock prices also display different patterns for firms with more CSR concerns (Chen and Gavious, 2015; Erragraguy and Revelli, 2015; Lam et al., 2015). In spite of these extensive discussions, the question on how the preference of stakeholders in the community affects firms’ CSR behavior is still less touched. The demographic and cultural features are fundamental elements of the society reflecting the preference of stakeholders in the community. As shown in the existing literature (Becker et al., 2011), firm’s corporate decision, such as dividend payment is influenced by the characteristics of local demographics. Specifically, firms located in area with more senior residents tend to pay more dividends. On the other hand, the cultural environment in the community, frequently proxied by religion, could influence corporate behaviors in many ways. For instance, cultural environment can be used to predict the variation of the investor protection, such as credit rights (Stulz and Williamson, 2003; Baele et al., 2014), the formation of international trade (Cavalcanti et al., 2007; Lewer and Van den Berg, 2007), stock trading behaviors (Grinblatt and Keloharju, 2001) and etc. Building on these important studies, in this paper, we investigate how community demographic and cultural environment shape corporate social responsibility. In this paper, we examine how the preference of the community affects firm’s CSR behavior. We adopt two proxies to quantify the effect of community preference, local seniors and religious residents, as both seniors and religious people are important stakeholders in the community. On the one hand, as documented in the literature (Peterson et al., 2001; Forte, 2004), people’s age appears to be positively correlated with ethical behavior. The local seniors have shown to have strong effect on corporate policy, such as payout policy (Becker et al., 2011). On the other hand, religiosity is the common proxy for culture used in existing literature. It has been documented that religion influences the governance quality (La Porta et al., 1999), firm’s behavior in earnings management (Du et al., 2014) and etc. And religiosity has also been shown as a strong predictor for ethical business behaviors (Conroy and Emerson, 2004). Therefore, local seniors and religious residents are likely to exhibit a stronger preference for CSR behaviors. And importantly, they are likely to exhibit less tolerance for unethical activities or behaviors harmful for the society. For example, people with strong religious belief are less likely to accept the practices of accounting manipulation (Conroy and Emerson, 2004; Longenecker et al., 2004). And McGuire et al., (2012) document that religious social norms are correlated with lower accounting risk and accounting practices challenged by the shareholder via lawsuit. Building on these arguments, we thus posit the communities with a high degree of religiosity naturally impose more demand for corporate social responsibility and are less tolerant about the corporate activities that will create social concerns. To mitigate the potential endogeneity issues in our study, we estimate a panel regression model with different combinations of fixed effects. We include firm fixed effect to account for firm specific time-invariant characteristics. The geographical or legal shocks that happen in different states may affect our results. To mitigate this concern, we consider operating stateby-year and incorporation state-by-year fixed effects to control for the possible contemporaneous effects at both operating state level and incorporation state levels. Using a 20-year panel sample with 3104 unique US firms and 19,580 firm-year observations, we find that the fraction of senior people and religious adherents in the community are both positively correlated with firms’ level of corporate social responsibility. In addition, we find that the effects are stronger for less financially constrained firms. It suggests only less constrained firms are able to meet the expectations of the communities. This paper contributes to the existing literature along several dimensions. First, this paper is among the first to provide the evidence consistent with the notion that CSR as a delegated prosocial behaviors influenced by the preference of local stakeholders. In a related study, Di Giuli and Kostovetsky (2014) document that firms score higher on CSR when they are headquartered in Democratic rather than Republican-leaning states. In our study, we focus on the preferences of stakeholders at county-level and show that firms located in counties with more senior residents and religious adherents have higher CSR scores. Our study adds important texture in this line of research. By proposing local seniors and religiosity to proxy for the effect of community environment, our research is also directly related to the studies on the determinants of corporate social responsibility. Second, this paper contributes to the studies on the influential effect of regional characteristics, such as demographics and culture on corporate policy. As documented in the literature (Becker et al., 2011; Stulz and Williamson, 2003), differences in demographics and culture, can partially explain investor behaviors. In contrast, the evidence in our study reveals that the influence of demographics and culture seems to be critical channels for stakeholder protection. The evidence help us understand the specific mechanisms how the effect of stakeholders interact with corporate behaviors. 2. Hypothesis Benabou and Tirole (2010) consider the nature of CSR as a delegated pro-social behavior. Firms might consider the stakeholders’ preference in deciding investment in CSR. In our analysis, we specifically consider the preference of stakeholders in the community where the firms are located. Our analysis builds on two important reasons. The first one is that some important stakeholders in the community indeed have been shown prefer socially responsible firms. The second one is related to the stakeholder value maximization view (Deng et al., 2013). Specifically, it is beneficial for a firm to invest in CSR because focusing on the interests of stakeholders may increase their willingness to support a firm’s operation and improve a firm’s business environment. To jointly satisfy these two arguments, we propose local seniors and local religiosity to measure the effect of community environment. First, senior citizens and religious adherents are important stakeholders for firms operating in the community Please cite this article as: D. Wu et al., Does community environment matter to corporate social responsibility? Finance Research Letters (2016), http://dx.doi.org/10.1016/j.frl.2016.04.010
ARTICLE IN PRESS
JID: FRL
D. Wu et al. / Finance Research Letters 000 (2016) 1–9
[m3Gsc;May 25, 2016;16:32] 3
(Becker et al., 2011). According to US 2010 Census data, senior citizens (age at 65 and over) on average account for 12% of the population and hold about 25% of total household wealth (Becker et al., 2011). Senior citizens influence the allocations of various resources that local firms may rely on for development, for example, the access to capital and labor. Religious adherents on the other hand account for more than half of the population (Hilary and Hui, 2009). Their attitudes to the firms are therefore important elements in corporate decisions. Second, as documented in the existing literature (Peterson et al., 2001; Forte, 2004; Conroy and Emerson, 2004), both senior citizens and religious adherents tend to prefer socially responsible behaviors and dislike social irresponsible behaviors. Therefore, firms that display higher level of strength and lower level of concerns with regard to corporate social responsibility, such as environmental concerns, are more likely to be supported by them. Firms with social irresponsible behaviors are less tolerated by the communities with more senior citizens and religious adherents. We thus have our first two hypotheses. Hypothesis 1. Firms operating in counties with more seniors are more likely to engage in social responsible activities and less likely to engage in social irresponsible activities, which result in a higher net level of corporate social responsibility. Hypothesis 2. Firms operating in counties with religious adherents are more likely to engage in social responsible activities and less likely to engage in social irresponsible activities, which result in a higher net level of corporate social responsibility. Investing in CSR is financially costly. And firms’ investment decisions are closely related to financial constraints (Cleary, 1999). As discussed in Hong et al. (2012), firms’ CSR behavior is more sensitive to financial constraints than capital or R&D investments. Therefore, less constrained firms are more likely to invest in CSR in order to meet the expectation of the community. We consider a battery of proxies for measuring firm-level financial constraints, such as firm size, firm age, and HP index following Hadlock and Pierce (2010). We have our third hypothesis. Hypothesis 3. For firms with less financial constraints, percentage of seniors and religious adherents in the operating county are positively correlated with the net level of corporate social responsibility. 3. Sample and variables Our analysis is based on a sample consisting of 19,580 firm-year observations with 3104 unique US firms from 1991 to 2010. We follow Becker et al. (2011) and construct the variable measuring the effect of local seniors. We collect the county characteristics Intercensal population estimates from US Census Bureau, Population Division. We identify the number of people not less than 65 in each county year and define the variable Senior as follows.
Seniori,t =
# of people with age equal and higher than 65 in the county i at year t total population in the county i at year t
Following Hilary and Hui (2009), we rely on the data of religiosity collected from the American Religion Data Archive (ARDA). We construct a variable at the county-level measuring the degree of religiosity. Religiosity is defined as follows.
Religiosityi,t =
# of religious adherents in the county i at year t total population in the county i at year t
The detailed number of religious adherents in the county is only available for 1990, 20 0 0 and 2010. We use linear interpolation method to fill the value in other years (Alesina and La Ferrara, 20 0 0). The firm’s county information is inferred from firm’s address post code from Compustat. Following existing studies on corporate social responsibility (as in Huseynov and Klamm (2012) and Gao and Zhang (2015)), we collect the CSR data from Kinder, Lydenberg, and Domini (KLD) database, which provides the most accepted CSR score in CSR literature. In KLD database, the firm’s best practices and the most serious challenges are ranked in different dimensions. For example, in the strengths category of employee relations, KLD database ranks managers’ practices in improving union relations, adopting cash profit-sharing program, having a strong health and safety programs and so on. In the concerns category, in contrast, KLD database considers whether the company has a history of notably poor union relations, whether recently has either paid substantial fines or civil penalties for willful violations of employee health and safety standards, has made significant reductions in its workforce and so on. Following the literature, we consider the strength ratings and concern ratings in seven dimensions: (1) community, (2) diversity, (3) corporate governance, (4) employee relations, (5) environment, (6) human rights, and (7) product. Stakeholders in the community are concerned not only firms’ advantageous practices with regards to CSR. Some serious concerns such as product safety, human rights and poor relationship with labor union are also stakeholders’ major concerns. Therefore, following prior studies (such as Deng et al., 2013, Di Giuli and Kostovetsky, 2014, Kim et al., 2014 and etc), we construct the net score of each dimension by subtracting CSR concern score from CSR strength score. Then, we aggregate the seven net score and define the overall CSR rating (CSR). We include several accounting variables to control for the possible effects from time-varying fundamentals. The data is collected from Standard & Poor’s Compustat. The control variables include Size (logarithm of market value of equity), Leverage (book value of liability over total assets), MTB (market value of equity and book value of liability over total assets), ROA (EBIT over total assets), R&D intensity (R&D expense over total assets), Ad (advertising expense over total assets), FAT Please cite this article as: D. Wu et al., Does community environment matter to corporate social responsibility? Finance Research Letters (2016), http://dx.doi.org/10.1016/j.frl.2016.04.010
ARTICLE IN PRESS
JID: FRL 4
[m3Gsc;May 25, 2016;16:32]
D. Wu et al. / Finance Research Letters 000 (2016) 1–9
Table 1 Variable definitions. Variable
Definition
CSR
The score of strength ratings minus concern ratings for corporate social responsibility on 7 dimensions: (1) community, (2) diversity, (3) corporate governance, (4) employee relations, (5) environment, (6) human rights and (7) product. Source: KLD database The number of people with the age not less than 65 divided by the total population in a county; Source: US Census Bureau The number of religious adherents divided by the total population in a county; Source: ARDA Natural logarithm of market value of equity (Compustat variable prcc_f∗ csho) at last fiscal year-end; Source: Compustat Book leverage (Compustat (dltt+dlc)/at) at last fiscal year-end; Source: Compustat Market value of equity and book value of liability over book value of assets (Compustat (csho∗ prcc_f+lt)/at) at last fiscal year-end; Source: Compustat Operating income before depreciation (Compustat variable oibdp) scaled by total assets (Compustat variable at); Source: Compustat R&D expenses (compustat variable rdx) scaled by total assets; Source: Compustat Advertising expenses (compustat variable adx) scaled by total assets; Source: Compustat Fixed assets ratio defined as plant, property, and equipment (compustat variable ppent) scaled by total assets; Source: Compustat Log value of population of a county in a fiscal year; Source: US census data The number of years since the first year when the firm appears in Compustat; Source: Compustat Financial constraint index constructed following Hadlock and Pierce (2010) with the following equation: HP index = −0.737 × Size + 0.043 × Size^2−0.040 × Age, where Size is the natural log of total assets capped at $4.5 billion and Age is the total number of years a firm has been on Compustat and is capped at thirty-seven years; Source: Compustat
Senior Religiosity Size Leverage MTB ROA R&D Ad FAT Ln_pop Age HP index
Table 2 Summary statistics. Panel A. Descriptive statistics Variable
N
Mean
S.D.
P25
Median
P75
CSR Senior Religiosity Size Leverage MTB ROA R&D Ad FAT Ln_pop
18,332 18,332 18,332 18,332 18,332 18,332 18,332 18,332 18,332 18,332 18,332
−0.298 0.117 0.521 7.444 0.22 1.966 0.075 0.034 0.013 0.255 13.662
2.315 0.026 0.109 1.615 0.212 1.396 0.154 0.08 0.043 0.237 1.082
−2 0.102 0.439 6.263 0.048 1.141 0.031 0 0 0.06 13.125
0 0.116 0.525 7.305 0.189 1.503 0.08 0 0 0.181 13.716
1 0.132 0.591 8.474 0.326 2.244 0.134 0.034 0.006 0.392 14.277
Panel B. Variance decomposition Standard deviation Group
# Groups
Full sample
Between groups
Within groups
Operating states Operating states X year Incorporation states Incorporation states X year
50 834 49 768
2.32 2.32 2.32 2.32
0.93 1.59 0.93 0.95
2.25 2.17 2.25 2.28
This table reports summary statistics of variables used in the regressions. The sample period is from 1991 to 2010. All variables are defined in Table 1. Panel A presents descriptive statistics and the variance decomposition is reported in Panel B.
(tangible assets over total assets). In addition, we include the log of total population in a county as a control variables. The variables definitions are detailed in Table 1. Table 2 reports the summary statistics. As shown in the Panel A, the average rating for all the seven CSR dimensions (CSR) is about −0.3 with standard deviation of 2.3. On average, senior citizens account for 11.7% of the total population with very low standard deviation (0.03). The number of religious adherents over total population in the county is on average 52%. For other control variables, the values are very close to the existing studies. The detailed information about the CSR score for firms in each operating states and incorporation states are presented in the Appendix 1. Our identification strategy is based on the variations in the county-level. Appendix 1 shows a great heterogeneity across counties within the same state, which facilitates a valid analysis. And the state with best CSR is Vermont, while Nebraska has firms with worst CSR. In addition, we conduct a variance decomposition approach to further explore Please cite this article as: D. Wu et al., Does community environment matter to corporate social responsibility? Finance Research Letters (2016), http://dx.doi.org/10.1016/j.frl.2016.04.010
ARTICLE IN PRESS
JID: FRL
[m3Gsc;May 25, 2016;16:32]
D. Wu et al. / Finance Research Letters 000 (2016) 1–9
5
Table 3 Local seniors and corporate social responsibility.
Senior
(1) CSR
(2)
(3)
(4)
15.481∗∗∗ (5.79)
15.667∗∗∗ (5.95)
Yes Yes No 18,332 0.660
Yes Yes Yes 18,332 0.664
19.033∗∗∗ (5.83) −0.091∗∗∗ (0.03) −0.014 (0.12) −0.022 (0.02) 0.477∗∗∗ (0.15) 0.306 (0.34) −0.580 (0.66) −0.467∗ (0.24) −2.122∗∗∗ (0.60) Yes Yes No 18,332 0.662
19.741∗∗∗ (5.95) −0.080∗∗ (0.03) 0.001 (0.12) −0.031 (0.02) 0.469∗∗∗ (0.15) 0.144 (0.35) −0.877 (0.62) −0.415∗ (0.25) −2.521∗∗∗ (0.65) Yes Yes Yes 18,332 0.665
Size Leverage MTB ROA R&D Ad FAT Ln_pop Firm State-year Incorp State-year N Adj. R-sq
This table presents the estimation of the effect of local seniors on corporate social responsibility. The results are estimated using fixed effects panel regressions. Definitions of all variables are contained in Table 1. Different combinations of firm, operating state-by-year, and incorporation state-by-year fixed effects are included. The heteroscedasticity-robust errors are reported in parentheses. ∗ Denote 10% significance level. ∗∗ Denote 5%, significance level. ∗∗∗ Denote 1% significance level.
whether the within-states variation is large. Panel B presents the result. We consider four group definitions, operating state group, operating state-by-year group, incorporation state group and incorporation state-by-year group. Consistent with the result in Appendix 1, the variation within state or state-by-year groups are quite large for a reasonable analysis. 4. Results We discuss our findings in this section. To quantify the effect of the community on corporate social responsibility, we estimate the following equation.
CSRi,t = β0 + β1Communit yit + γ Zit + αi + θit where CSRi, t denotes the aggregated CSR net score. Communityit is represented by two proxies, Senior or Religiosity. The variations in these two variables are at county-year level. Zit includes several control variables. α i is the firm fixed effects to control for the firm-level time-invariant characteristics. These firm-specific characteristics include some observable features such as the industry effect and all other unobservable traits that are not time-varying. θ it are operating state-by-year and incorporation state-by-year fixed effects to control for the operating state level and incorporation state level trends. This econometric method is to mitigate the endogeneity concerns with regard to omitted variable problems and follows standard approaches in studies using panel data (such as Bertrand and Mullainathan, 2003). Table 3 presents the baseline result for the effect of local seniors on corporate social responsibility. The significantly positive coefficients suggest firms operating in counties with high local seniors display higher level of corporate social responsibility. One standard deviation increase in percentage of local seniors is correlated with 0.4 to 0.51 increase in CSR. The result is robust after controlling for standard control variables, operating state-by-year fixed effects or incorporation state-by-year fixed effects. The evidence supports our prediction in Hypothesis 1. We then investigate the effect of local religiosity. Table 4 indicates that local religiosity gauged by fraction of religious adherents can predict the level of CSR. One standard deviation increase in religiosity correlates with 0.23 to 0.25 increase in CSR net rating. The evidence on the effect of seniors and religiosity is robust after controlling for firm-level characteristics, operating state-level trends and incorporation state trends. The results still hold when two variable, Senior and Religiosity are included in the same model as shown in Table 5. The evidence is consistent with our Hypothesis 2 that local religiosity indeed influences firms’ investment CSR. Please cite this article as: D. Wu et al., Does community environment matter to corporate social responsibility? Finance Research Letters (2016), http://dx.doi.org/10.1016/j.frl.2016.04.010
ARTICLE IN PRESS
JID: FRL 6
[m3Gsc;May 25, 2016;16:32]
D. Wu et al. / Finance Research Letters 000 (2016) 1–9 Table 4 Religiosity and corporate social responsibility.
Religiosity
(1) CSR
(2)
(3)
(4)
2.282∗∗∗ (0.82)
2.154∗∗∗ (0.83)
Yes Yes No 18,332 0.660
Yes Yes Yes 18,332 0.664
2.289∗∗∗ (0.82) −0.098∗∗∗ (0.03) 0.009 (0.12) −0.022 (0.02) 0.480∗∗∗ (0.15) 0.295 (0.34) −0.485 (0.65) −0.430∗ (0.24) −1.754∗∗∗ (0.60) Yes Yes No 18,332 0.661
2.116∗∗ (0.82) −0.085∗∗ (0.03) 0.021 (0.12) −0.031 (0.02) 0.474∗∗∗ (0.15) 0.141 (0.35) −0.767 (0.62) −0.366 (0.25) −2.155∗∗∗ (0.65) Yes Yes Yes 18,332 0.665
Size Leverage MTB ROA R&D Ad FAT Ln_pop Firm State-year Incorp State-year N Adj. R-sq
This table presents the estimation of the effect of religiosity on corporate social responsibility. The results are estimated using fixed effects panel regressions. Definitions of all variables are contained in Table 1. Different combinations of firm, operating state-by-year, and incorporation state-by-year fixed effects are included. The heteroscedasticity-robust errors are reported in parentheses. ∗ Denote 10% significance level. ∗∗ Denote 5% significance level ∗∗∗ Denote 1% significance level.
Table 5 Local seniors, religiosity and corporate social responsibility.
Senior Religiosity Controls Firm State-year Incorp State-year N Adj. R-sq
(1) CSR
(2)
(3)
(4)
14.591∗∗ (5.76) 2.145∗∗∗ (0.82) No Yes Yes No 18,332 0.661
14.831∗∗ (5.90) 2.007∗∗ (0.82) No Yes Yes Yes 18,332 0.664
18.071∗∗∗ (5.79) 2.108∗∗∗ (0.81) Yes Yes Yes No 18,332 0.662
18.864∗∗∗ (5.91) 1.917∗∗ (0.81) Yes Yes Yes Yes 18,332 0.665
This table presents the estimation of the effect of local senior and religiosity on corporate social responsibility in one regression model. The results are estimated using fixed effects panel regressions. Definitions of all variables are contained in Table 1. Different combinations of firm, operating state-by-year, and incorporation state-by-year fixed effects are included. The heteroscedasticity-robust errors are reported in parentheses. ∗∗ Denote 5% significance level. ∗∗∗ Denote 1% significance level.
Next, we empirically test whether firm-level financial constraints play a role in the relationship between community and CSR. We use firm size, firm age and HP index (Hadlock and Pierce, 2010) to measure the degree of financial constraints. Larger firm and more mature firms are theoretically less financially constrained. HP index is calculated using firm size and firm age. Larger value of HP index indicates more severe financial constraint. We split the sample according to these three variables. The results in Table 6 exhibit the consistent pattern that the positive effects of seniors and religiosity appear to be stronger for less financially constrained firms. The evidence is consistent with the notion that firms seem to be able to meet the local demand for CSR when they are in good financial conditions. Hypothesis 3 is therefore supported. Please cite this article as: D. Wu et al., Does community environment matter to corporate social responsibility? Finance Research Letters (2016), http://dx.doi.org/10.1016/j.frl.2016.04.010
ARTICLE IN PRESS
JID: FRL
[m3Gsc;May 25, 2016;16:32]
D. Wu et al. / Finance Research Letters 000 (2016) 1–9
7
Table 6 The effect of financial constraint. Panel A. The effect of local seniors (1) CSR Size > Median Senior Controls Firm State-year Incorp State-year N Adj. R-sq
35.188 (9.38) Yes Yes Yes Yes 9166 0.670
∗∗∗
(2)
(3)
Size < Median
Age > Median
−7.290 (7.70) Yes Yes Yes Yes 9166 0.740
25.251 (8.01) Yes Yes Yes Yes 9131 0.678
(2)
(3)
∗∗∗
(4)
(5)
(6)
Age < Median
HP index < Median ∗∗∗
−0.728 (7.75) Yes Yes Yes Yes 9198 0.705
32.909 (8.23) Yes Yes Yes Yes 9165 0.681
(4)
(5)
Age < Median
HP index < Median
HP index > Median 1.236 (8.05) Yes Yes Yes Yes 9164 0.703
Panel B. The effect of religiosity (1) CSR Size > Median Religiosity Controls Firm State-year Incorp State-year N Adj. R-sq
∗∗∗
3.450 (1.33) Yes Yes Yes Yes 9166 0.669
Size < Median ∗∗
2.336 (1.07) Yes Yes Yes Yes 9166 0.740
Age > Median 2.233 (1.17) Yes Yes Yes Yes 9131 0.677
∗
(6)
∗∗
1.036 (1.13) Yes Yes Yes Yes 9198 0.705
2.687 (1.20) Yes Yes Yes Yes 9165 0.680
HP index > Median 1.195 (1.12) Yes Yes Yes Yes 9164 0.703
This table presents the heterogeneous effect of community on corporate social responsibility according to the firm-level financial constraints. The results are estimated using fixed effects panel regressions. Definitions of all variables are contained in Table 1. Different combinations of firm, operating state-by-year, and incorporation state-by-year fixed effects are included. The heteroscedasticity-robust errors are reported in parentheses. ∗ Denote 10% significance level. ∗∗ Denote 5% significance level. ∗∗∗ Denote 1% significance level.
5. Conclusion This paper documents the effect of local stakeholders in the community on firms’ behavior in corporate social responsibility. Using a comprehensive US sample, we find that both local senior and local religiosity have positive effects on firm’s CSR level. In addition, we uncover that the link between the preference of stakeholders in the community and CSR appears to be stronger for firms that are less financially constrained. Based on evidence obtained from a battery of tests, our study contributes to the literature by supporting the stakeholder theory of CSR. Specifically, this study first provides new stakeholder-based determinants for CSR. And the main evidence suggests the firm’s local stakeholder is an important consideration when firm making decision on CSR.
Appendix 1 Summary of corporate social responsibility by operating state and incorporation state. Operating State
N
Mean
SD
Incorporation State
N
Mean
SD
AK AL AR AZ CA CO CT DC DE FL GA HI
13 103 145 238 2800 330 475 84 77 551 558 31
0.46 −0.23 −1.41 −1.09 0.12 −0.67 0.28 0.02 −1.7 −0.75 −0.63 1.32
0.78 2.12 2.97 2.17 2.26 2.07 2.66 2.53 2.53 1.83 1.97 1.9
AK AL AR AZ CA CO CT DC DE FL GA HI
7 27 25 27 391 45 75 − 11,242 312 225 20
0.14 0.85 0.24 −0.22 0.35 0.38 0.61 − −0.45 −0.4 −0.05 0.45
0.38 1.99 1.2 1.65 2.03 1.99 1.66 − 2.31 1.8 1.54 1.1
(continued on next page)
Please cite this article as: D. Wu et al., Does community environment matter to corporate social responsibility? Finance Research Letters (2016), http://dx.doi.org/10.1016/j.frl.2016.04.010
ARTICLE IN PRESS
JID: FRL 8
[m3Gsc;May 25, 2016;16:32]
D. Wu et al. / Finance Research Letters 000 (2016) 1–9 Appendix 1 (continued) Operating State
N
IA ID IL IN KS KY LA MA MD ME MI MN MO MS MT NC ND NE NH NJ NM NV NY OH OK OR PA RI SC SD TN TX UT VA VT WA WI WV
135 67 969 250 102 136 157 874 273 15 432 538 397 45 26 434 13 106 40 676 7 117 1475 804 165 173 922 78 93 24 357 1648 109 495 37 366 341 31
Mean 0.42 0 −0.45 −0.1 −1.1 −1 −0.64 −0.07 −0.31 1.07 0.19 0.57 −0.88 −0.71 1.23 −0.54 −1.31 −1.72 1.45 −0.13 0.29 −0.76 0.07 −0.48 −0.76 0.8 −0.54 0.04 −0.54 −0.67 −0.59 −1.02 −0.1 −0.96 3.76 0.41 −0.22 0.19
SD
Incorporation State
N
Mean
SD
1.73 2.36 2.28 1.97 2.28 2.6 1.3 1.55 2.21 0.88 2.84 2.51 1.86 1.18 0.95 2.33 1.65 2.75 2.93 2.1 1.89 1.62 2.34 2.57 1.89 2.77 2.02 2.02 2.04 2.08 2.3 2.54 2.06 1.97 5.67 2.26 1.75 1.01
IA ID IL IN KS KY LA MA MD ME MI MN MO MS MT NC ND NE NJ NM NV NY OH OK OR PA RI SC SD TN TX UT VA VT WA WI WV WY
78 19 70 227 48 49 68 277 401 8 207 315 211 33 11 159 8 31 306 7 231 557 541 75 144 498 34 52 22 130 240 71 298 19 232 234 19 6
0.09 0.47 −1.37 0.29 −1 −0.78 0.06 0.22 −0.44 1.13 0.79 0.1 −0.53 −0.45 0.82 −0.8 −0.5 0.61 −0.16 0.29 −1.06 0.34 −0.11 −0.32 1.12 −0.26 1.56 −0.46 −0.36 −0.69 −0.15 −0.99 −0.46 4.47 0.66 −0.12 0.63 −2.17
1.25 1.5 2.63 2.24 2.47 2.26 1.09 1.63 1.74 0.83 2.91 2.11 1.78 1.15 1.08 2.44 1.6 1.48 2.61 1.89 1.71 3.28 2.68 1.92 2.69 1.84 1.52 1.82 1.89 1.95 2.32 2.39 2.3 4.11 2.75 1.82 0.83 1.72
Reference Alesina, A., La Ferrara, E., 20 0 0. Participation in heterogeneous communities. Q. J. Econ. 115 (3), 847–904. Attig, N., El Ghoul, S., Guedhami, O., Suh, J., 2013. Corporate social responsibility and credit ratings. J. Bus. Ethics. 117 (4), 679–694. Baele, L., Farooq, M., Ongena, S., 2014. Of religion and redemption: evidence from default on Islamic loans. J. Bank. Financ. 44, 141–159. ´ Z., Weisbenner, S., 2011. Local dividend clienteles. J. Financ. 66 (2), 655–683. Becker, B., Ivkovic, Bénabou, R., Tirole, J., 2010. Individual and corporate social responsibility. Economica 77 (305), 1–19. Bertrand, M., Mullainathan, S., 2003. Enjoying the quiet life? Corporate governance and managerial preferences. Journal of political Economy 111 (5), 1043–1075. Cavalcanti, T.V., Parente, S.L., Zhao, R., 2007. Religion in macroeconomics: a quantitative analysis of Weber’s thesis. Econ. Theor. 32 (1), 105–123. Cleary, S., 1999. The relationship between firm investment and financial status. J. Financ. 54 (2), 673–692. Chen, E., Gavious, I., 2015. Does CSR have different value implications for different shareholders? Financ. Res. Let. 14, 29–35. Conroy, S.J., Emerson, T.L., 2004. Business ethics and religion: religiosity as a predictor of ethical awareness among students. J. Bus. Ethics. 50 (4), 383–396. Deng, X., Kang, J.K., Low, B.S., 2013. Corporate social responsibility and stakeholder value maximization: evidence from mergers. J. Financ. Econ. 110 (1), 87–109. Di Giuli, A., Kostovetsky, L., 2014. Are red or blue companies more likely to go green? Politics and corporate social responsibility. J. Financ. Econ. 111 (1), 158–180. Du, X., Jian, W., Lai, S., Du, Y., Pei, H., 2014. Does religion mitigate earnings management? Evidence from China. J. Bus. Ethics. 131 (3), 1–51. El Ghoul, S., Guedhami, O., Kwok, C.C., Mishra, D.R., 2011. Does corporate social responsibility affect the cost of capital? J. Bank. Financ. 35 (9), 2388–2406. Erragraguy, E., Revelli, C., 2015. Should Islamic investors consider SRI criteria in their investment strategies? Financ. Res. Let. 14, 11–19. Forte, A., 2004. Business ethics: a study of the moral reasoning of selected business managers and the influence of organizational ethical climate. J. Bus. Ethics. 51 (2), 167–173. Gao, L., Zhang, J.H., 2015. Firms’ earnings smoothing, corporate social responsibility, and valuation. J. Corp. Financ. 32, 108–127. Goss, A., Roberts, G.S., 2011. The impact of corporate social responsibility on the cost of bank loans. J. Bank. Financ. 35 (7), 1794–1810. Grinblatt, M., Keloharju, M., 2001. How distance, language, and culture influence stockholdings and trades. J. Financ. 56 (3), 1053–1073. Hadlock, C.J., Pierce, J.R., 2010. New evidence on measuring financial constraints: moving beyond the KZ index. Rev. Financ. Stud. 23 (5), 1909–1940. Hilary, G., Hui, K.W., 2009. Does religion matter in corporate decision making in America? J. Financ. Econ. 93 (3), 455–473. Hong, H., Kubik, J.D., Scheinkman, J.A., 2012. Financial constraints on corporate goodness (No. w18476). National Bureau of Economic Research. Huseynov, F., Klamm, B.K., 2012. Tax avoidance, tax management and corporate social responsibility. J. Corp. Financ. 18 (4), 804–827. Kim, Y., Li, H., Li, S., 2014. Corporate social responsibility and stock price crash risk. J. Bank. Financ. 43, 1–13. La Porta, R., Lopez-de-Silanes, F., Shleifer, A., Vishny, R., 1999. The quality of government. J. Law. Econ. Organ. 15 (1), 222–279.
Please cite this article as: D. Wu et al., Does community environment matter to corporate social responsibility? Finance Research Letters (2016), http://dx.doi.org/10.1016/j.frl.2016.04.010
JID: FRL
ARTICLE IN PRESS D. Wu et al. / Finance Research Letters 000 (2016) 1–9
[m3Gsc;May 25, 2016;16:32] 9
Lam, S.S., Zhang, W., Jacob, G.H., 2015. The mispricing of socially ambiguous grey stocks. Financ. Res. Let. 13, 81–89. Lewer, J.J., Van den Berg, H., 2007. Religion and international trade: does the sharing of a religious culture facilitate the formation of trade networks? Am. J. Econ. Sociol. 66 (4), 765–794. Longenecker, J.G., McKinney, J.A., Moore, C.W., 2004. Religious intensity, evangelical Christianity, and business ethics: an empirical study. J. Bus. Ethics. 55 (4), 373–386. McGuire, S.T., Omer, T.C., Sharp, N.Y., 2012. The impact of religion on financial reporting irregularities. Account. Rev. 87 (2), 645–673. Peterson, D., Rhoads, A., Vaught, B.C., 2001. Ethical beliefs of business professionals: a study of gender, age and external factors. J. Bus. Ethics. 31 (3), 225–232. Stulz, R.M., Williamson, R., 2003. Culture, openness, and finance. J. Financ. Econ. 70 (3), 313–349.
Please cite this article as: D. Wu et al., Does community environment matter to corporate social responsibility? Finance Research Letters (2016), http://dx.doi.org/10.1016/j.frl.2016.04.010
JID: FRL
[m3Gsc;May 25, 2016;16:32]
Finance Research Letters 0 0 0 (2016) 1–9
Contents lists available at ScienceDirect
Finance Research Letters journal homepage: www.elsevier.com/locate/frl
Does community environment matter to corporate social responsibility? Dejun Wu a,1,∗, Chen Lin b, Sibo Liu b a b
School of Accounting, Zhongnan University of Economics and Law, Wuhan, China School of Economics and Finance, The University of Hong Kong, Hong Kong
a r t i c l e
i n f o
Article history: Received 20 November 2015 Revised 18 March 2016 Accepted 3 April 2016 Available online xxx JEL classification: G02 G12 G30 G32
a b s t r a c t How does the preference of local stakeholders shape firm’s corporate social responsibility (CSR)? Using a twenty-year U.S. sample, this paper examines the effects of the community demographic and cultural environment on CSR. We find that firms located in counties with more senior residents and/or higher levels of religiosity display higher degree of corporate social responsibility. In addition, we find the positive effects of local seniors and religiosity are stronger for less financial constrained firms. The result is robust after controlling for firm fixed effects, operating state trends and incorporation state trends. Taken together, our evidence supports a stakeholder theory of corporate social responsibility. © 2016 Published by Elsevier Inc.
Keywords: Corporate social responsibility Stakeholders Community Seniors Religiosity
1. Introduction Corporate social responsibility (CSR) has been embraced by a growing number of firms and gained much public attention in recent years. Most US firms nowadays practice corporate social responsibility in some forms. According to a report1 by KPMG, 86% of US firms issue social responsibility reports. Some large firms invest hundreds of millions of dollars annually on altruistic endeavors related to employee or community development programs (Hong et al., 2012). Practitioners are assessing the possible channels that CSR can add value to the firm. A survey2 conducted by McKinsey & Company suggests most of CFOs perceive that CSR increases firms’ value. In particular, meeting society’s expectations is an important channel. CSR have been discussed in academic studies for decades. The debate focuses on why firm would invest significant resources on CSR. According to some studies, CSR is considered as a crucial corporate behavior that deals with firms’ relationship with their stakeholders, such as employees, communities and etc. The pro-social behavior that provide value to firms stakeholders (Benabou and Tirole, 2010) and therefore gain the support from them. It has been documented that firm’s CSR ∗
1 1 2
Corresponding author. E-mail addresses: [email protected] (D. Wu), [email protected] (C. Lin), [email protected] (S. Liu). Dejun Wu acknowledges financial support from the Education Ministry of China (Project No. 14YJA630067). The KPMG Survey of Corporate Responsibility Reporting 2013 Valuing corporate social responsibility: McKinsey Global Survey Results.
http://dx.doi.org/10.1016/j.frl.2016.04.010 1544-6123/© 2016 Published by Elsevier Inc.
Please cite this article as: D. Wu et al., Does community environment matter to corporate social responsibility? Finance Research Letters (2016), http://dx.doi.org/10.1016/j.frl.2016.04.010
JID: FRL 2
ARTICLE IN PRESS
[m3Gsc;May 25, 2016;16:32]
D. Wu et al. / Finance Research Letters 000 (2016) 1–9
level matters for debt financing (Goss and Roberts, 2011), equity financing (El Ghoul et al., 2011), credit rating (Attig et al., 2013) and so on. The stock prices also display different patterns for firms with more CSR concerns (Chen and Gavious, 2015; Erragraguy and Revelli, 2015; Lam et al., 2015). In spite of these extensive discussions, the question on how the preference of stakeholders in the community affects firms’ CSR behavior is still less touched. The demographic and cultural features are fundamental elements of the society reflecting the preference of stakeholders in the community. As shown in the existing literature (Becker et al., 2011), firm’s corporate decision, such as dividend payment is influenced by the characteristics of local demographics. Specifically, firms located in area with more senior residents tend to pay more dividends. On the other hand, the cultural environment in the community, frequently proxied by religion, could influence corporate behaviors in many ways. For instance, cultural environment can be used to predict the variation of the investor protection, such as credit rights (Stulz and Williamson, 2003; Baele et al., 2014), the formation of international trade (Cavalcanti et al., 2007; Lewer and Van den Berg, 2007), stock trading behaviors (Grinblatt and Keloharju, 2001) and etc. Building on these important studies, in this paper, we investigate how community demographic and cultural environment shape corporate social responsibility. In this paper, we examine how the preference of the community affects firm’s CSR behavior. We adopt two proxies to quantify the effect of community preference, local seniors and religious residents, as both seniors and religious people are important stakeholders in the community. On the one hand, as documented in the literature (Peterson et al., 2001; Forte, 2004), people’s age appears to be positively correlated with ethical behavior. The local seniors have shown to have strong effect on corporate policy, such as payout policy (Becker et al., 2011). On the other hand, religiosity is the common proxy for culture used in existing literature. It has been documented that religion influences the governance quality (La Porta et al., 1999), firm’s behavior in earnings management (Du et al., 2014) and etc. And religiosity has also been shown as a strong predictor for ethical business behaviors (Conroy and Emerson, 2004). Therefore, local seniors and religious residents are likely to exhibit a stronger preference for CSR behaviors. And importantly, they are likely to exhibit less tolerance for unethical activities or behaviors harmful for the society. For example, people with strong religious belief are less likely to accept the practices of accounting manipulation (Conroy and Emerson, 2004; Longenecker et al., 2004). And McGuire et al., (2012) document that religious social norms are correlated with lower accounting risk and accounting practices challenged by the shareholder via lawsuit. Building on these arguments, we thus posit the communities with a high degree of religiosity naturally impose more demand for corporate social responsibility and are less tolerant about the corporate activities that will create social concerns. To mitigate the potential endogeneity issues in our study, we estimate a panel regression model with different combinations of fixed effects. We include firm fixed effect to account for firm specific time-invariant characteristics. The geographical or legal shocks that happen in different states may affect our results. To mitigate this concern, we consider operating stateby-year and incorporation state-by-year fixed effects to control for the possible contemporaneous effects at both operating state level and incorporation state levels. Using a 20-year panel sample with 3104 unique US firms and 19,580 firm-year observations, we find that the fraction of senior people and religious adherents in the community are both positively correlated with firms’ level of corporate social responsibility. In addition, we find that the effects are stronger for less financially constrained firms. It suggests only less constrained firms are able to meet the expectations of the communities. This paper contributes to the existing literature along several dimensions. First, this paper is among the first to provide the evidence consistent with the notion that CSR as a delegated prosocial behaviors influenced by the preference of local stakeholders. In a related study, Di Giuli and Kostovetsky (2014) document that firms score higher on CSR when they are headquartered in Democratic rather than Republican-leaning states. In our study, we focus on the preferences of stakeholders at county-level and show that firms located in counties with more senior residents and religious adherents have higher CSR scores. Our study adds important texture in this line of research. By proposing local seniors and religiosity to proxy for the effect of community environment, our research is also directly related to the studies on the determinants of corporate social responsibility. Second, this paper contributes to the studies on the influential effect of regional characteristics, such as demographics and culture on corporate policy. As documented in the literature (Becker et al., 2011; Stulz and Williamson, 2003), differences in demographics and culture, can partially explain investor behaviors. In contrast, the evidence in our study reveals that the influence of demographics and culture seems to be critical channels for stakeholder protection. The evidence help us understand the specific mechanisms how the effect of stakeholders interact with corporate behaviors. 2. Hypothesis Benabou and Tirole (2010) consider the nature of CSR as a delegated pro-social behavior. Firms might consider the stakeholders’ preference in deciding investment in CSR. In our analysis, we specifically consider the preference of stakeholders in the community where the firms are located. Our analysis builds on two important reasons. The first one is that some important stakeholders in the community indeed have been shown prefer socially responsible firms. The second one is related to the stakeholder value maximization view (Deng et al., 2013). Specifically, it is beneficial for a firm to invest in CSR because focusing on the interests of stakeholders may increase their willingness to support a firm’s operation and improve a firm’s business environment. To jointly satisfy these two arguments, we propose local seniors and local religiosity to measure the effect of community environment. First, senior citizens and religious adherents are important stakeholders for firms operating in the community Please cite this article as: D. Wu et al., Does community environment matter to corporate social responsibility? Finance Research Letters (2016), http://dx.doi.org/10.1016/j.frl.2016.04.010
ARTICLE IN PRESS
JID: FRL
D. Wu et al. / Finance Research Letters 000 (2016) 1–9
[m3Gsc;May 25, 2016;16:32] 3
(Becker et al., 2011). According to US 2010 Census data, senior citizens (age at 65 and over) on average account for 12% of the population and hold about 25% of total household wealth (Becker et al., 2011). Senior citizens influence the allocations of various resources that local firms may rely on for development, for example, the access to capital and labor. Religious adherents on the other hand account for more than half of the population (Hilary and Hui, 2009). Their attitudes to the firms are therefore important elements in corporate decisions. Second, as documented in the existing literature (Peterson et al., 2001; Forte, 2004; Conroy and Emerson, 2004), both senior citizens and religious adherents tend to prefer socially responsible behaviors and dislike social irresponsible behaviors. Therefore, firms that display higher level of strength and lower level of concerns with regard to corporate social responsibility, such as environmental concerns, are more likely to be supported by them. Firms with social irresponsible behaviors are less tolerated by the communities with more senior citizens and religious adherents. We thus have our first two hypotheses. Hypothesis 1. Firms operating in counties with more seniors are more likely to engage in social responsible activities and less likely to engage in social irresponsible activities, which result in a higher net level of corporate social responsibility. Hypothesis 2. Firms operating in counties with religious adherents are more likely to engage in social responsible activities and less likely to engage in social irresponsible activities, which result in a higher net level of corporate social responsibility. Investing in CSR is financially costly. And firms’ investment decisions are closely related to financial constraints (Cleary, 1999). As discussed in Hong et al. (2012), firms’ CSR behavior is more sensitive to financial constraints than capital or R&D investments. Therefore, less constrained firms are more likely to invest in CSR in order to meet the expectation of the community. We consider a battery of proxies for measuring firm-level financial constraints, such as firm size, firm age, and HP index following Hadlock and Pierce (2010). We have our third hypothesis. Hypothesis 3. For firms with less financial constraints, percentage of seniors and religious adherents in the operating county are positively correlated with the net level of corporate social responsibility. 3. Sample and variables Our analysis is based on a sample consisting of 19,580 firm-year observations with 3104 unique US firms from 1991 to 2010. We follow Becker et al. (2011) and construct the variable measuring the effect of local seniors. We collect the county characteristics Intercensal population estimates from US Census Bureau, Population Division. We identify the number of people not less than 65 in each county year and define the variable Senior as follows.
Seniori,t =
# of people with age equal and higher than 65 in the county i at year t total population in the county i at year t
Following Hilary and Hui (2009), we rely on the data of religiosity collected from the American Religion Data Archive (ARDA). We construct a variable at the county-level measuring the degree of religiosity. Religiosity is defined as follows.
Religiosityi,t =
# of religious adherents in the county i at year t total population in the county i at year t
The detailed number of religious adherents in the county is only available for 1990, 20 0 0 and 2010. We use linear interpolation method to fill the value in other years (Alesina and La Ferrara, 20 0 0). The firm’s county information is inferred from firm’s address post code from Compustat. Following existing studies on corporate social responsibility (as in Huseynov and Klamm (2012) and Gao and Zhang (2015)), we collect the CSR data from Kinder, Lydenberg, and Domini (KLD) database, which provides the most accepted CSR score in CSR literature. In KLD database, the firm’s best practices and the most serious challenges are ranked in different dimensions. For example, in the strengths category of employee relations, KLD database ranks managers’ practices in improving union relations, adopting cash profit-sharing program, having a strong health and safety programs and so on. In the concerns category, in contrast, KLD database considers whether the company has a history of notably poor union relations, whether recently has either paid substantial fines or civil penalties for willful violations of employee health and safety standards, has made significant reductions in its workforce and so on. Following the literature, we consider the strength ratings and concern ratings in seven dimensions: (1) community, (2) diversity, (3) corporate governance, (4) employee relations, (5) environment, (6) human rights, and (7) product. Stakeholders in the community are concerned not only firms’ advantageous practices with regards to CSR. Some serious concerns such as product safety, human rights and poor relationship with labor union are also stakeholders’ major concerns. Therefore, following prior studies (such as Deng et al., 2013, Di Giuli and Kostovetsky, 2014, Kim et al., 2014 and etc), we construct the net score of each dimension by subtracting CSR concern score from CSR strength score. Then, we aggregate the seven net score and define the overall CSR rating (CSR). We include several accounting variables to control for the possible effects from time-varying fundamentals. The data is collected from Standard & Poor’s Compustat. The control variables include Size (logarithm of market value of equity), Leverage (book value of liability over total assets), MTB (market value of equity and book value of liability over total assets), ROA (EBIT over total assets), R&D intensity (R&D expense over total assets), Ad (advertising expense over total assets), FAT Please cite this article as: D. Wu et al., Does community environment matter to corporate social responsibility? Finance Research Letters (2016), http://dx.doi.org/10.1016/j.frl.2016.04.010
ARTICLE IN PRESS
JID: FRL 4
[m3Gsc;May 25, 2016;16:32]
D. Wu et al. / Finance Research Letters 000 (2016) 1–9
Table 1 Variable definitions. Variable
Definition
CSR
The score of strength ratings minus concern ratings for corporate social responsibility on 7 dimensions: (1) community, (2) diversity, (3) corporate governance, (4) employee relations, (5) environment, (6) human rights and (7) product. Source: KLD database The number of people with the age not less than 65 divided by the total population in a county; Source: US Census Bureau The number of religious adherents divided by the total population in a county; Source: ARDA Natural logarithm of market value of equity (Compustat variable prcc_f∗ csho) at last fiscal year-end; Source: Compustat Book leverage (Compustat (dltt+dlc)/at) at last fiscal year-end; Source: Compustat Market value of equity and book value of liability over book value of assets (Compustat (csho∗ prcc_f+lt)/at) at last fiscal year-end; Source: Compustat Operating income before depreciation (Compustat variable oibdp) scaled by total assets (Compustat variable at); Source: Compustat R&D expenses (compustat variable rdx) scaled by total assets; Source: Compustat Advertising expenses (compustat variable adx) scaled by total assets; Source: Compustat Fixed assets ratio defined as plant, property, and equipment (compustat variable ppent) scaled by total assets; Source: Compustat Log value of population of a county in a fiscal year; Source: US census data The number of years since the first year when the firm appears in Compustat; Source: Compustat Financial constraint index constructed following Hadlock and Pierce (2010) with the following equation: HP index = −0.737 × Size + 0.043 × Size^2−0.040 × Age, where Size is the natural log of total assets capped at $4.5 billion and Age is the total number of years a firm has been on Compustat and is capped at thirty-seven years; Source: Compustat
Senior Religiosity Size Leverage MTB ROA R&D Ad FAT Ln_pop Age HP index
Table 2 Summary statistics. Panel A. Descriptive statistics Variable
N
Mean
S.D.
P25
Median
P75
CSR Senior Religiosity Size Leverage MTB ROA R&D Ad FAT Ln_pop
18,332 18,332 18,332 18,332 18,332 18,332 18,332 18,332 18,332 18,332 18,332
−0.298 0.117 0.521 7.444 0.22 1.966 0.075 0.034 0.013 0.255 13.662
2.315 0.026 0.109 1.615 0.212 1.396 0.154 0.08 0.043 0.237 1.082
−2 0.102 0.439 6.263 0.048 1.141 0.031 0 0 0.06 13.125
0 0.116 0.525 7.305 0.189 1.503 0.08 0 0 0.181 13.716
1 0.132 0.591 8.474 0.326 2.244 0.134 0.034 0.006 0.392 14.277
Panel B. Variance decomposition Standard deviation Group
# Groups
Full sample
Between groups
Within groups
Operating states Operating states X year Incorporation states Incorporation states X year
50 834 49 768
2.32 2.32 2.32 2.32
0.93 1.59 0.93 0.95
2.25 2.17 2.25 2.28
This table reports summary statistics of variables used in the regressions. The sample period is from 1991 to 2010. All variables are defined in Table 1. Panel A presents descriptive statistics and the variance decomposition is reported in Panel B.
(tangible assets over total assets). In addition, we include the log of total population in a county as a control variables. The variables definitions are detailed in Table 1. Table 2 reports the summary statistics. As shown in the Panel A, the average rating for all the seven CSR dimensions (CSR) is about −0.3 with standard deviation of 2.3. On average, senior citizens account for 11.7% of the total population with very low standard deviation (0.03). The number of religious adherents over total population in the county is on average 52%. For other control variables, the values are very close to the existing studies. The detailed information about the CSR score for firms in each operating states and incorporation states are presented in the Appendix 1. Our identification strategy is based on the variations in the county-level. Appendix 1 shows a great heterogeneity across counties within the same state, which facilitates a valid analysis. And the state with best CSR is Vermont, while Nebraska has firms with worst CSR. In addition, we conduct a variance decomposition approach to further explore Please cite this article as: D. Wu et al., Does community environment matter to corporate social responsibility? Finance Research Letters (2016), http://dx.doi.org/10.1016/j.frl.2016.04.010
ARTICLE IN PRESS
JID: FRL
[m3Gsc;May 25, 2016;16:32]
D. Wu et al. / Finance Research Letters 000 (2016) 1–9
5
Table 3 Local seniors and corporate social responsibility.
Senior
(1) CSR
(2)
(3)
(4)
15.481∗∗∗ (5.79)
15.667∗∗∗ (5.95)
Yes Yes No 18,332 0.660
Yes Yes Yes 18,332 0.664
19.033∗∗∗ (5.83) −0.091∗∗∗ (0.03) −0.014 (0.12) −0.022 (0.02) 0.477∗∗∗ (0.15) 0.306 (0.34) −0.580 (0.66) −0.467∗ (0.24) −2.122∗∗∗ (0.60) Yes Yes No 18,332 0.662
19.741∗∗∗ (5.95) −0.080∗∗ (0.03) 0.001 (0.12) −0.031 (0.02) 0.469∗∗∗ (0.15) 0.144 (0.35) −0.877 (0.62) −0.415∗ (0.25) −2.521∗∗∗ (0.65) Yes Yes Yes 18,332 0.665
Size Leverage MTB ROA R&D Ad FAT Ln_pop Firm State-year Incorp State-year N Adj. R-sq
This table presents the estimation of the effect of local seniors on corporate social responsibility. The results are estimated using fixed effects panel regressions. Definitions of all variables are contained in Table 1. Different combinations of firm, operating state-by-year, and incorporation state-by-year fixed effects are included. The heteroscedasticity-robust errors are reported in parentheses. ∗ Denote 10% significance level. ∗∗ Denote 5%, significance level. ∗∗∗ Denote 1% significance level.
whether the within-states variation is large. Panel B presents the result. We consider four group definitions, operating state group, operating state-by-year group, incorporation state group and incorporation state-by-year group. Consistent with the result in Appendix 1, the variation within state or state-by-year groups are quite large for a reasonable analysis. 4. Results We discuss our findings in this section. To quantify the effect of the community on corporate social responsibility, we estimate the following equation.
CSRi,t = β0 + β1Communit yit + γ Zit + αi + θit where CSRi, t denotes the aggregated CSR net score. Communityit is represented by two proxies, Senior or Religiosity. The variations in these two variables are at county-year level. Zit includes several control variables. α i is the firm fixed effects to control for the firm-level time-invariant characteristics. These firm-specific characteristics include some observable features such as the industry effect and all other unobservable traits that are not time-varying. θ it are operating state-by-year and incorporation state-by-year fixed effects to control for the operating state level and incorporation state level trends. This econometric method is to mitigate the endogeneity concerns with regard to omitted variable problems and follows standard approaches in studies using panel data (such as Bertrand and Mullainathan, 2003). Table 3 presents the baseline result for the effect of local seniors on corporate social responsibility. The significantly positive coefficients suggest firms operating in counties with high local seniors display higher level of corporate social responsibility. One standard deviation increase in percentage of local seniors is correlated with 0.4 to 0.51 increase in CSR. The result is robust after controlling for standard control variables, operating state-by-year fixed effects or incorporation state-by-year fixed effects. The evidence supports our prediction in Hypothesis 1. We then investigate the effect of local religiosity. Table 4 indicates that local religiosity gauged by fraction of religious adherents can predict the level of CSR. One standard deviation increase in religiosity correlates with 0.23 to 0.25 increase in CSR net rating. The evidence on the effect of seniors and religiosity is robust after controlling for firm-level characteristics, operating state-level trends and incorporation state trends. The results still hold when two variable, Senior and Religiosity are included in the same model as shown in Table 5. The evidence is consistent with our Hypothesis 2 that local religiosity indeed influences firms’ investment CSR. Please cite this article as: D. Wu et al., Does community environment matter to corporate social responsibility? Finance Research Letters (2016), http://dx.doi.org/10.1016/j.frl.2016.04.010
ARTICLE IN PRESS
JID: FRL 6
[m3Gsc;May 25, 2016;16:32]
D. Wu et al. / Finance Research Letters 000 (2016) 1–9 Table 4 Religiosity and corporate social responsibility.
Religiosity
(1) CSR
(2)
(3)
(4)
2.282∗∗∗ (0.82)
2.154∗∗∗ (0.83)
Yes Yes No 18,332 0.660
Yes Yes Yes 18,332 0.664
2.289∗∗∗ (0.82) −0.098∗∗∗ (0.03) 0.009 (0.12) −0.022 (0.02) 0.480∗∗∗ (0.15) 0.295 (0.34) −0.485 (0.65) −0.430∗ (0.24) −1.754∗∗∗ (0.60) Yes Yes No 18,332 0.661
2.116∗∗ (0.82) −0.085∗∗ (0.03) 0.021 (0.12) −0.031 (0.02) 0.474∗∗∗ (0.15) 0.141 (0.35) −0.767 (0.62) −0.366 (0.25) −2.155∗∗∗ (0.65) Yes Yes Yes 18,332 0.665
Size Leverage MTB ROA R&D Ad FAT Ln_pop Firm State-year Incorp State-year N Adj. R-sq
This table presents the estimation of the effect of religiosity on corporate social responsibility. The results are estimated using fixed effects panel regressions. Definitions of all variables are contained in Table 1. Different combinations of firm, operating state-by-year, and incorporation state-by-year fixed effects are included. The heteroscedasticity-robust errors are reported in parentheses. ∗ Denote 10% significance level. ∗∗ Denote 5% significance level ∗∗∗ Denote 1% significance level.
Table 5 Local seniors, religiosity and corporate social responsibility.
Senior Religiosity Controls Firm State-year Incorp State-year N Adj. R-sq
(1) CSR
(2)
(3)
(4)
14.591∗∗ (5.76) 2.145∗∗∗ (0.82) No Yes Yes No 18,332 0.661
14.831∗∗ (5.90) 2.007∗∗ (0.82) No Yes Yes Yes 18,332 0.664
18.071∗∗∗ (5.79) 2.108∗∗∗ (0.81) Yes Yes Yes No 18,332 0.662
18.864∗∗∗ (5.91) 1.917∗∗ (0.81) Yes Yes Yes Yes 18,332 0.665
This table presents the estimation of the effect of local senior and religiosity on corporate social responsibility in one regression model. The results are estimated using fixed effects panel regressions. Definitions of all variables are contained in Table 1. Different combinations of firm, operating state-by-year, and incorporation state-by-year fixed effects are included. The heteroscedasticity-robust errors are reported in parentheses. ∗∗ Denote 5% significance level. ∗∗∗ Denote 1% significance level.
Next, we empirically test whether firm-level financial constraints play a role in the relationship between community and CSR. We use firm size, firm age and HP index (Hadlock and Pierce, 2010) to measure the degree of financial constraints. Larger firm and more mature firms are theoretically less financially constrained. HP index is calculated using firm size and firm age. Larger value of HP index indicates more severe financial constraint. We split the sample according to these three variables. The results in Table 6 exhibit the consistent pattern that the positive effects of seniors and religiosity appear to be stronger for less financially constrained firms. The evidence is consistent with the notion that firms seem to be able to meet the local demand for CSR when they are in good financial conditions. Hypothesis 3 is therefore supported. Please cite this article as: D. Wu et al., Does community environment matter to corporate social responsibility? Finance Research Letters (2016), http://dx.doi.org/10.1016/j.frl.2016.04.010
ARTICLE IN PRESS
JID: FRL
[m3Gsc;May 25, 2016;16:32]
D. Wu et al. / Finance Research Letters 000 (2016) 1–9
7
Table 6 The effect of financial constraint. Panel A. The effect of local seniors (1) CSR Size > Median Senior Controls Firm State-year Incorp State-year N Adj. R-sq
35.188 (9.38) Yes Yes Yes Yes 9166 0.670
∗∗∗
(2)
(3)
Size < Median
Age > Median
−7.290 (7.70) Yes Yes Yes Yes 9166 0.740
25.251 (8.01) Yes Yes Yes Yes 9131 0.678
(2)
(3)
∗∗∗
(4)
(5)
(6)
Age < Median
HP index < Median ∗∗∗
−0.728 (7.75) Yes Yes Yes Yes 9198 0.705
32.909 (8.23) Yes Yes Yes Yes 9165 0.681
(4)
(5)
Age < Median
HP index < Median
HP index > Median 1.236 (8.05) Yes Yes Yes Yes 9164 0.703
Panel B. The effect of religiosity (1) CSR Size > Median Religiosity Controls Firm State-year Incorp State-year N Adj. R-sq
∗∗∗
3.450 (1.33) Yes Yes Yes Yes 9166 0.669
Size < Median ∗∗
2.336 (1.07) Yes Yes Yes Yes 9166 0.740
Age > Median 2.233 (1.17) Yes Yes Yes Yes 9131 0.677
∗
(6)
∗∗
1.036 (1.13) Yes Yes Yes Yes 9198 0.705
2.687 (1.20) Yes Yes Yes Yes 9165 0.680
HP index > Median 1.195 (1.12) Yes Yes Yes Yes 9164 0.703
This table presents the heterogeneous effect of community on corporate social responsibility according to the firm-level financial constraints. The results are estimated using fixed effects panel regressions. Definitions of all variables are contained in Table 1. Different combinations of firm, operating state-by-year, and incorporation state-by-year fixed effects are included. The heteroscedasticity-robust errors are reported in parentheses. ∗ Denote 10% significance level. ∗∗ Denote 5% significance level. ∗∗∗ Denote 1% significance level.
5. Conclusion This paper documents the effect of local stakeholders in the community on firms’ behavior in corporate social responsibility. Using a comprehensive US sample, we find that both local senior and local religiosity have positive effects on firm’s CSR level. In addition, we uncover that the link between the preference of stakeholders in the community and CSR appears to be stronger for firms that are less financially constrained. Based on evidence obtained from a battery of tests, our study contributes to the literature by supporting the stakeholder theory of CSR. Specifically, this study first provides new stakeholder-based determinants for CSR. And the main evidence suggests the firm’s local stakeholder is an important consideration when firm making decision on CSR.
Appendix 1 Summary of corporate social responsibility by operating state and incorporation state. Operating State
N
Mean
SD
Incorporation State
N
Mean
SD
AK AL AR AZ CA CO CT DC DE FL GA HI
13 103 145 238 2800 330 475 84 77 551 558 31
0.46 −0.23 −1.41 −1.09 0.12 −0.67 0.28 0.02 −1.7 −0.75 −0.63 1.32
0.78 2.12 2.97 2.17 2.26 2.07 2.66 2.53 2.53 1.83 1.97 1.9
AK AL AR AZ CA CO CT DC DE FL GA HI
7 27 25 27 391 45 75 − 11,242 312 225 20
0.14 0.85 0.24 −0.22 0.35 0.38 0.61 − −0.45 −0.4 −0.05 0.45
0.38 1.99 1.2 1.65 2.03 1.99 1.66 − 2.31 1.8 1.54 1.1
(continued on next page)
Please cite this article as: D. Wu et al., Does community environment matter to corporate social responsibility? Finance Research Letters (2016), http://dx.doi.org/10.1016/j.frl.2016.04.010
ARTICLE IN PRESS
JID: FRL 8
[m3Gsc;May 25, 2016;16:32]
D. Wu et al. / Finance Research Letters 000 (2016) 1–9 Appendix 1 (continued) Operating State
N
IA ID IL IN KS KY LA MA MD ME MI MN MO MS MT NC ND NE NH NJ NM NV NY OH OK OR PA RI SC SD TN TX UT VA VT WA WI WV
135 67 969 250 102 136 157 874 273 15 432 538 397 45 26 434 13 106 40 676 7 117 1475 804 165 173 922 78 93 24 357 1648 109 495 37 366 341 31
Mean 0.42 0 −0.45 −0.1 −1.1 −1 −0.64 −0.07 −0.31 1.07 0.19 0.57 −0.88 −0.71 1.23 −0.54 −1.31 −1.72 1.45 −0.13 0.29 −0.76 0.07 −0.48 −0.76 0.8 −0.54 0.04 −0.54 −0.67 −0.59 −1.02 −0.1 −0.96 3.76 0.41 −0.22 0.19
SD
Incorporation State
N
Mean
SD
1.73 2.36 2.28 1.97 2.28 2.6 1.3 1.55 2.21 0.88 2.84 2.51 1.86 1.18 0.95 2.33 1.65 2.75 2.93 2.1 1.89 1.62 2.34 2.57 1.89 2.77 2.02 2.02 2.04 2.08 2.3 2.54 2.06 1.97 5.67 2.26 1.75 1.01
IA ID IL IN KS KY LA MA MD ME MI MN MO MS MT NC ND NE NJ NM NV NY OH OK OR PA RI SC SD TN TX UT VA VT WA WI WV WY
78 19 70 227 48 49 68 277 401 8 207 315 211 33 11 159 8 31 306 7 231 557 541 75 144 498 34 52 22 130 240 71 298 19 232 234 19 6
0.09 0.47 −1.37 0.29 −1 −0.78 0.06 0.22 −0.44 1.13 0.79 0.1 −0.53 −0.45 0.82 −0.8 −0.5 0.61 −0.16 0.29 −1.06 0.34 −0.11 −0.32 1.12 −0.26 1.56 −0.46 −0.36 −0.69 −0.15 −0.99 −0.46 4.47 0.66 −0.12 0.63 −2.17
1.25 1.5 2.63 2.24 2.47 2.26 1.09 1.63 1.74 0.83 2.91 2.11 1.78 1.15 1.08 2.44 1.6 1.48 2.61 1.89 1.71 3.28 2.68 1.92 2.69 1.84 1.52 1.82 1.89 1.95 2.32 2.39 2.3 4.11 2.75 1.82 0.83 1.72
Reference Alesina, A., La Ferrara, E., 20 0 0. Participation in heterogeneous communities. Q. J. Econ. 115 (3), 847–904. Attig, N., El Ghoul, S., Guedhami, O., Suh, J., 2013. Corporate social responsibility and credit ratings. J. Bus. Ethics. 117 (4), 679–694. Baele, L., Farooq, M., Ongena, S., 2014. Of religion and redemption: evidence from default on Islamic loans. J. Bank. Financ. 44, 141–159. ´ Z., Weisbenner, S., 2011. Local dividend clienteles. J. Financ. 66 (2), 655–683. Becker, B., Ivkovic, Bénabou, R., Tirole, J., 2010. Individual and corporate social responsibility. Economica 77 (305), 1–19. Bertrand, M., Mullainathan, S., 2003. Enjoying the quiet life? Corporate governance and managerial preferences. Journal of political Economy 111 (5), 1043–1075. Cavalcanti, T.V., Parente, S.L., Zhao, R., 2007. Religion in macroeconomics: a quantitative analysis of Weber’s thesis. Econ. Theor. 32 (1), 105–123. Cleary, S., 1999. The relationship between firm investment and financial status. J. Financ. 54 (2), 673–692. Chen, E., Gavious, I., 2015. Does CSR have different value implications for different shareholders? Financ. Res. Let. 14, 29–35. Conroy, S.J., Emerson, T.L., 2004. Business ethics and religion: religiosity as a predictor of ethical awareness among students. J. Bus. Ethics. 50 (4), 383–396. Deng, X., Kang, J.K., Low, B.S., 2013. Corporate social responsibility and stakeholder value maximization: evidence from mergers. J. Financ. Econ. 110 (1), 87–109. Di Giuli, A., Kostovetsky, L., 2014. Are red or blue companies more likely to go green? Politics and corporate social responsibility. J. Financ. Econ. 111 (1), 158–180. Du, X., Jian, W., Lai, S., Du, Y., Pei, H., 2014. Does religion mitigate earnings management? Evidence from China. J. Bus. Ethics. 131 (3), 1–51. El Ghoul, S., Guedhami, O., Kwok, C.C., Mishra, D.R., 2011. Does corporate social responsibility affect the cost of capital? J. Bank. Financ. 35 (9), 2388–2406. Erragraguy, E., Revelli, C., 2015. Should Islamic investors consider SRI criteria in their investment strategies? Financ. Res. Let. 14, 11–19. Forte, A., 2004. Business ethics: a study of the moral reasoning of selected business managers and the influence of organizational ethical climate. J. Bus. Ethics. 51 (2), 167–173. Gao, L., Zhang, J.H., 2015. Firms’ earnings smoothing, corporate social responsibility, and valuation. J. Corp. Financ. 32, 108–127. Goss, A., Roberts, G.S., 2011. The impact of corporate social responsibility on the cost of bank loans. J. Bank. Financ. 35 (7), 1794–1810. Grinblatt, M., Keloharju, M., 2001. How distance, language, and culture influence stockholdings and trades. J. Financ. 56 (3), 1053–1073. Hadlock, C.J., Pierce, J.R., 2010. New evidence on measuring financial constraints: moving beyond the KZ index. Rev. Financ. Stud. 23 (5), 1909–1940. Hilary, G., Hui, K.W., 2009. Does religion matter in corporate decision making in America? J. Financ. Econ. 93 (3), 455–473. Hong, H., Kubik, J.D., Scheinkman, J.A., 2012. Financial constraints on corporate goodness (No. w18476). National Bureau of Economic Research. Huseynov, F., Klamm, B.K., 2012. Tax avoidance, tax management and corporate social responsibility. J. Corp. Financ. 18 (4), 804–827. Kim, Y., Li, H., Li, S., 2014. Corporate social responsibility and stock price crash risk. J. Bank. Financ. 43, 1–13. La Porta, R., Lopez-de-Silanes, F., Shleifer, A., Vishny, R., 1999. The quality of government. J. Law. Econ. Organ. 15 (1), 222–279.
Please cite this article as: D. Wu et al., Does community environment matter to corporate social responsibility? Finance Research Letters (2016), http://dx.doi.org/10.1016/j.frl.2016.04.010
JID: FRL
ARTICLE IN PRESS D. Wu et al. / Finance Research Letters 000 (2016) 1–9
[m3Gsc;May 25, 2016;16:32] 9
Lam, S.S., Zhang, W., Jacob, G.H., 2015. The mispricing of socially ambiguous grey stocks. Financ. Res. Let. 13, 81–89. Lewer, J.J., Van den Berg, H., 2007. Religion and international trade: does the sharing of a religious culture facilitate the formation of trade networks? Am. J. Econ. Sociol. 66 (4), 765–794. Longenecker, J.G., McKinney, J.A., Moore, C.W., 2004. Religious intensity, evangelical Christianity, and business ethics: an empirical study. J. Bus. Ethics. 55 (4), 373–386. McGuire, S.T., Omer, T.C., Sharp, N.Y., 2012. The impact of religion on financial reporting irregularities. Account. Rev. 87 (2), 645–673. Peterson, D., Rhoads, A., Vaught, B.C., 2001. Ethical beliefs of business professionals: a study of gender, age and external factors. J. Bus. Ethics. 31 (3), 225–232. Stulz, R.M., Williamson, R., 2003. Culture, openness, and finance. J. Financ. Econ. 70 (3), 313–349.
Please cite this article as: D. Wu et al., Does community environment matter to corporate social responsibility? Finance Research Letters (2016), http://dx.doi.org/10.1016/j.frl.2016.04.010