Changes in Corporate Social Responsibility and Stock Performance

Abstract

We study the relationship between corporate social performance and financial performance by comparing the portfolio returns of firms with changes in corporate social responsibility (CSR) intensity. Using an extensive US sample from the MSCI ESG database, we find that improvement in the overall CSR is generally value enhancing. The relationship varies with CSR dimensions. More importantly, the relationship shifts differently for various CSR dimensions during the crisis period when trust in the society is low and financial resource is limited. Improvement in environment, human rights, and product characteristics shows higher financial returns during the financial crisis period, whereas the value enhancement of improvement in employee relations is more pronounced during the non-crisis period.

Introduction

Business ethics is an important topic that has received much attention from business practitioners and academic researchers. According to Lewis (1985), business ethics is “moral rules, standards, codes, or principles which provide guidelines for right and truthful behavior in specific situations.”

One central question in this field is how managers can manage a firm with appropriate attention to ethical concern and corporate social responsibility (CSR). CSR can be regarded as an obligation of the business society to all its stakeholders (Gössling and Vocht 2007), and has become increasingly important. Today more and more managers incorporate CSR into their management agenda and provide corporate responsibility report to the public. A recent study conducted on over 1000 top business executives in the world reveals that 93% of the CEOs believe that sustainability is crucial to the success of their businesses, and 81% of the CEOs agree that the sustainability reputation of their companies is important to consumer purchasing decisions.¹ KPMG reports that 75% of the 4900 global companies that were surveyed undertake the practice of corporate responsibility reporting.² Meanwhile, due to the increasing awareness of CSR issues by investors, socially responsible investing has become popular practice in the investment society (Renneboog et al. 2008).³

The literature has two opposing views on whether managers should take CSR into account in making managerial decisions. Freeman (1984) argues that the management of a corporation should be held accountable for the welfare of its stakeholders, which include not only its stockholders, but also its employees, customers, suppliers, communities, etc. According to this so-called stakeholder theory, a positive relationship between firms and their stakeholders helps improve firm value, which ultimately benefits their stockholders. Proponents of the stakeholder theory believe that the goal of firm value maximization can be consistent with the practice of CSR, such as environmental protection, employee relations improvement, and human rights promotion (Macintosh 1999; Deng et al. 2013). However, this view contradicts the argument made by Friedman (1970) that the sole goal of corporate management is to maximize stockholders' interests and thus CSR activities essentially cost stockholders instead of benefiting them (see also Aupperle et al. 1985; Waddock and Graves 1997). Jensen (2001) argues that the stakeholder theory lacks a precise objective function. Because of the inherent conflict of interest among different stakeholders, this theory empowers managers to use valuable firm resources to be engaged in CSR that helps to build up their personal reputation instead of maximizing firm values (see Jensen and Meckling 1976; Benabou and Tirole 2010; Masulis and Reza 2015).

The opposing views predict different relationships between corporate social performance (CSP) and financial performance. Preston and O’Bannon (1997) propose a typology of these relationships. The social impact (tradeoff) hypothesis proposes that better CSP leads to improvement (deterioration) in financial performance. The available funding (managerial opportunism) hypothesis, on the other hand, claims that good financial performance causes better (worse) CSP. The positive and negative synergy hypotheses argue that CSP and financial performance are synergistic. Many studies in the business ethics or finance literature focus on whether CSP has a financial payoff. One main concern is that CSR activities may negatively impact stockholders’ wealth. Although stockholders’ welfare includes more than pecuniary benefits, financial return is still of most importance to investors. On the other hand, being socially responsible may result in a win-win situation that not only enhances other stakeholders’ welfare but also benefits stockholders. The relationship between CSP and financial performance can be complex and can change with CSR dimensions or external environments. Empirically, the literature has reported mixed evidence. Some authors find a positive relation (Orlitzky 2001; Flammer 2015; Ferrell et al. 2016). Others demonstrate a negative relationship (Hillman and Keim 2001; Moore 2001; Kruger 2015). Still other authors report no significant relationship (Seifert et al. 2003; Makni et al. 2009). Bowman and Haire (1975) show an inverted “U” relationship, i.e., an intermediate level of CSP will maximize firm value.

This paper contributes to the business ethics and finance literature by examining the relationship between CSP and financial performance from a different perspective. Existing studies compare financial returns of firms with different levels of CSP in the cross-section. If a firm maintains a certain level of CSP, its impact should have been reflected in the firm value. Thus, evidence obtained by comparing the performance of firms with different levels of social performance in the cross-section may not identify the effect of CSP on financial performance. In this paper, we examine the relation between CSP and financial performance by comparing the portfolio returns of firms that have experienced changes in CSP over time. Specifically, we study the portfolio returns of firms that experience changes in different CSR dimensions, including community, diversity, employee relations, environment, human rights, and product characteristics. If CSP is positively related to financial performance, we should expect the portfolio of firms with improvement in CSP in the past to have better financial performance in the future. On the other hand, if managerial engagement in CSR reflects an agency problem between managers and stockholders, we should anticipate the portfolio of firms which commit more resources to CSR to subsequently underperform financially.

Most studies on the valuation of CSR assume a stationary relation between CSP and financial performance. Some researchers, however, suggest that the value of CSP is likely to be revealed during adverse market conditions or when the firm suffers a negative event. Godfrey et al. (2009) and Shiu and Yang (2017) indicate that CSR activities offer an insurance-like protection, so firms engaging more in CSR activities receive less negative judgments from their shareholders in the face of adverse events. Lins et al. (2017) find that stocks of high-CSR firms performed better than those of low-CSR firms during the 2008–2009 financial crisis, and that firm-specific social capital, which is built up via the engagement in CSR, pays off when trust in the society is low. Consistent results are also found in Nofsinger and Varma (2014). On the other hand, Koh et al. (2014) indicate that the value enhancement of CSR is constrained if firms do not obtain pragmatic legitimacy, e.g., when firms are financially unhealthy. Wang and Qian (2011) show that firms’ engagement in prosocial activities may result in punishment from stakeholders if firms have a poor financial position. These studies predict different impacts of CSR practice on financial performance during the crisis period when financial resource is limited and trust in the society is scarce and suggest that the relationship may change with market conditions. Our paper contributes to the business ethics literature by studying the relationship between CSP and financial performance during the crisis and non-crisis periods. We investigate whether companies can establish their social capital by engaging in CSR activities during the crisis period when financial resource is limited and obtain better financial performance. Furthermore, we separately examine the value of each of the CSR dimensions, including environment, employee relations, community, diversity, human rights, and product characteristics.

Using risk-adjusted stock returns as a measure of financial performance, we find that in general firms with improvement in CSR have better financial performance. However, the relation changes with different CSR dimensions. Specifically, firms with improvement in environment or human rights earn positive risk-adjusted returns, which are significantly higher than those with deterioration in the corresponding CSR dimensions. Firms with improvement in employee relations or product characteristics show positive risk-adjusted returns. However, we find mixed evidence that firms with deterioration in these two dimensions also earn positive risk-adjusted returns.

Splitting the sample into financial crisis and non-crisis periods provides much interesting insight. Improvements in environment or human rights are value enhancing during both the crisis and non-crisis periods. On the other hand, firms with improvement in product characteristics outperform those with deterioration during the crisis period, but during the non-crisis period, firms with deterioration in product characteristics have higher risk-adjusted returns. Firms with deterioration in employee relations outperform those with improvement during the financial crisis, while these two groups of firms do not perform differently during the non-crisis period. This finding supports Koh et al.’s (2014) argument that the value generation of CSR is constrained if firms perform poorly financially, and is also consistent with Wang and Qian’s (2011) results that firms’ engagement in prosocial activities may result in punishment from stakeholders if firms have a poor financial position. Our evidence shows that the relation between financial performance and CSP in employee relations and product characteristics changes with market conditions. Furthermore, we demonstrate that improvement in environment, human rights, and product characteristics during the crisis period is better rewarded than during the non-crisis period, whereas the value enhancement of improvement in employee relations is more pronounced during the non-crisis period. Our overall results suggest that the relationship between CSP and financial performance shifts with market conditions and depends on particular CSR dimensions.

Our findings are robust to a battery of sensitivity tests, including an alternative measure of financial performance (Tobin’s Q), a potential reverse causation problem, alternative sample periods, controls for potential confounding effects from other dimensions, and controls for changes in the definition of strengths and concerns in the CSR dimensions in the database over time.

The remainder of this paper is organized as follows. “Literature Review and Hypothesis Development” reviews literature and develops our hypotheses. “Data” describes the data. “Empirical Methodology” presents the empirical methodology. Empirical results are presented in “Empirical Evidence”. Robustness checks are conducted in “Robustness Checks” and conclusions are provided in “Conclusions”.

Literature Review and Hypothesis Development

The literature has reported mixed empirical evidence on the relationship between firm value and CSR. A number of authors show a positive relation (Orlitzky 2001; Simpson and Kohers 2002; van Beurden and Gössling 2008). Contrary to the argument of agency theory, Ferrell et al. (2016) demonstrate that firms with less concern of agency problem are engaged more in CSR practices. Furthermore, researchers document a positive relation between financial performance and specific CSR dimensions, such as corporate governance (Bebchuk et al. 2009), environmental performance (Derwall et al. 2005; Statman and Glushkov 2009), and employee satisfaction (Jiao 2010; Edmans 2011). Wang et al. (2018) show that mandatory CSR disclosure helps to improve the quality of financial information disclosure and constrain earnings management. Several studies indicate that firms with higher CSR performance face lower cost of capital and are financially less constrained (Sharfman and Fernando 2008; Goss and Roberts 2011). A number of researchers also find that the practice of CSR offers an insurance-like protection that reduces firm risk (Godfrey et al. 2009; Jo and Na 2012). Chih et al. (2010) indicate that financial firms engage more CSR activities to enhance their competitive advantages when the market becomes more competitive.

By contrast, consistent with the agency theory, several researchers show that firms engaged in CSR suffer more from agency problems and are associated with lower shareholder value (Hillman and Keim 2001; Kruger 2015). Moore (2001), for instance, finds a negative contemporaneous relationship between social and financial performance in the U.K. supermarket industry. Cheng et al. (2014) show that investment in CSR activities declines as managerial ownership or monitoring increases. Hong et al. (2012) argue that instead of doing well by doing good, firms do good only when they do well, and they show that financially less constrained firms are more likely to be engaged in CSR activities. A related argument is made by Lys et al. (2015), who indicate that firms are more likely to be engaged in CSR activities in the current period when managers anticipate better financial performance in the future and better CSR performance does not itself lead to good financial performance. Di Giuli and Kostovetsky (2014) find that Democratic-leaning firms spend more money on CSR than Republican-leaning ones and investment in CSR is negatively related to future stock returns and ROA. Hong and Kacperczyk (2009) demonstrate that sin stocks offer positive abnormal returns. Meanwhile, other authors do not find a significant relation between financial performance and CSR (McWilliams and Siegel 2000; Seifert et al. 2003).

Most existing research studies the relationship between CSP and financial performance by comparing firms with different levels of CSP. However, if a firm has a stable CSR intensity, the market should have incorporated the effect of CSP into its stock price. Therefore, evidence obtained by comparing the performance of firms with different CSP levels may not be able to identify the effect of CSP on financial performance. In this paper, we investigate the relation by comparing the portfolio returns of firms that have experienced changes in CSP over time. Khan et al. (2016) study future financial performance of firms that experience changes in residual CSR scores. They compare the impact of changes in material and immaterial sustainability investments on firm performance. We examine the relationship between financial performance and changes in CSR for different CSR dimensions. If the agency theory provides the primary explanation, we should find a negative relation between change in CSR intensity and firm value. On the other hand, a firm’s financial performance should be positively related to improvement in CSR if the stakeholder theory plays a dominant role. Therefore, we test the following:

Hypothesis 1:

Under the stakeholder theory, we predict a positive association between corporate social performance and financial performance.

Hypothesis 2:

Under the agency theory, we predict a negative association between corporate social performance and financial performance.

Godfrey et al. (2009) and Shiu and Yang (2017) indicate that CSR activities offer an insurance-like protection, so firms engaging more in CSR receive less negative judgments from their shareholders in the face of adverse events. Similarly, Lins et al. (2017) argue that the value of social capital created through CSR practices matters the most during the crisis period when trust in the society is low. Nofsinger and Varma (2014) show that socially responsible mutual funds outperform conventional funds during the crisis period, but not during the non-crisis period. They indicate that investors of socially responsible mutual funds are willing to accept lower returns during non-crises in exchange for higher returns during market crises. Similar evidence is also found in socially responsible bond funds (Henke 2016). On the other hand, Koh et al. (2014) show that the value enhancement of CSR is constrained when firms are not financially healthy. If companies can quickly establish social capital by engaging in CSR activities, we may anticipate a stronger relationship during the crisis period when trust in the society is scarce. However, CSR practices may be less rewarded during crisis period because firms’ financial resource is limited. Therefore, the relationship between CSP and financial performance may change during crisis and non-crisis periods. Hence, we test the following:

Hypothesis 3:

The relationship between corporate social performance and financial performance during the crisis period may change from that during the non-crisis period.

Data

We use the MSCI ESG (formally known as KLD Research & Analytics) database to determine a firm's social performance. MSCI ESG has been widely used in the literature to measure a firm’s CSR intensity (e.g., Di Giuli and Kostovetsky 2014). The number of firms covered in the MSCI ESG database has significantly increased over time. From 1991 to 2000, MSCI ESG covered all firms of the S&P 500 and Domini Social index. In 2001, MSCI ESG expanded to cover firms of the Russell 1000 index and in 2002 added firms of the Large Cap Social index. In 2003, all firms of the Russell 2000 and the Broad Market Social index are added into the MSCI ESG database. The average number of firms covered in the database is 653 during 1991–2000, 1107 in 2001, 1108 in 2002, and 2,941 during 2003–2012.⁴ MSCI ESG measures the performance of CSR in thirteen dimensions. The first seven dimensions are community, corporate governance, diversity, employee relations, environment, human rights, and product characteristics. Within each dimension, MSCI ESG establishes a list of strengths and concerns (see the Appendix for the list for year 2012 as an example) and uses them as criteria to measure a firm's social performance. For each strength (concern), MSCI ESG assigns a score of one for the presence of that strength (concern) in the evaluated firm and zero otherwise. For instance, employee involvement is one of the strength criteria listed in employee relations. The evaluated firm will get a score of one in employee involvement if the firm fulfills the criterion and zero otherwise. Similarly, employment health & safety is one of the concerns listed in employee relations. The evaluated firm gets a score of one if MSCI ESG sees the presence of that concern in the company and zero otherwise. The last six dimensions indicate if firms are involved in controversial businesses, including alcohol, gambling, tobacco, firearms, military, and nuclear power. For each dimension, a firm gets a score of one if its operation is involved in the indicated controversial business and zero otherwise.

To measure CSP, we follow El Ghoul et al. (2011) and Kempf and Osthoff (2007) by considering the following six dimensions: community, diversity, employee relations, environment, human rights and product characteristics. We exclude corporate governance from the CSR analysis. According to Servaes and Tamayo (2013), “corporate governance is about the mechanisms that allow the principals (shareholders) to reward and exert control on the agents (the managers),” whereas CSR “deals with social objectives and stakeholders other than shareholders.” Because this study focuses on the relationship between CSP and financial performance, we do not include corporate governance in our measure of CSP. The impact of corporate governance on firm value is also well researched in the corporate governance literature. Firms involved in the six dimensions of controversial businesses are fundamentally different from other firms in terms of CSR and thus are excluded from our study. Hong and Kacperczyk (2009) find that stocks involved in the controversial businesses usually outperform the market because they have been neglected by norm-constrained investors. Studies on the relationship between CSR and firm performance usually separate regular businesses from firms involved in the controversial businesses (e.g., Galema et al. 2008; Lins et al. 2017).⁵ In this study, we focus on the general relationship between CSR and financial performance, so we exclude firms involved in the controversial business from our analysis. Because MSCI ESG may have a firm on its list but does not examine its performance in all CSR dimensions, we exclude those firms from our analysis if they do not receive any strength and concern indicator in the dimension that we examine (Statman and Glushkov 2009). To control for the variation in the number of strengths and concerns considered by MSCI ESG over the sample years, for each CSR dimension, we standardize each firm’s strength (concern) scores by dividing its strength (concern) scores by the total number of strengths (concerns) considered by MSCI ESG in the same year. The standardized strength and concern scores range between zero and one. We then subtract the standardized concern score from the standardized strength score to obtain the firm’s social performance.

Table 1 summarizes the number of strengths (S) and concerns (C) considered by MSCI ESG from 1991 to 2012. We use the notations “COMˮ, “DIVˮ, “EMPˮ, “ENVˮ, “HUMˮ, and “PROˮ to represent the dimensions of community, diversity, employee relations, environment, human rights, and product characteristics, respectively. As shown in Table 1, the number of strengths and concerns considered by MSCI ESG varies over the sample period. Taking the dimension of community as an example, in 2012, there are 2 (1) strengths (concerns) considered by MSCI ESG, compared to 7 (4) strengths (concerns) in 2009. The number of strengths and concerns also differs among dimensions. In 2012, while there are 9 (9) strengths (concerns) considered in environment, only 2 (4) strengths (concerns) are examined in human rights.

Table 1 Summary of MSCI ESG data

Table 1 also presents the average performance of CSR of the sample firms in each dimension. The performance of CSR ranges between − 1 and 1 by construction, with − 1 (1) meaning the firm receives scores of one in all concerns (strengths) and zero in all strengths (concerns). We find that the average social performance changes over time and in most cases does not deviate from zero greatly. For example, the average performance of employee relations in 1991 and 2012 is equal to 0.02 and 0.16, respectively, and in 2010 it is − 0.2. Among these dimensions, we do not find a significant change in the average social performance during the financial crisis of 2008. Table 1 shows some improvement in the performance of community, diversity, employee relations, human rights, and product characteristics during the last three years. We also find a great variation in CSP among the dimensions considered. For instance, while firms overall have positive performance in community, the average performance in environment is negative in most years in the sample.

Eccles and Serafeim (2013) show that the importance of different CSR dimensions varies with industries. Thus, it is possible that firms in certain industries pay attention and engage more in certain CSR dimensions. Due to the variation in CSR across industries, we adjust for heterogeneity by constructing an industry-adjusted MSCI ESG performance measure. The industry-adjusted scores allow us to identify firms that improve or deteriorate relative to their peers of the same industry in different CSR dimensions. Specifically, for each year and each dimension, we compute the average performance for each industry and subtract it from the performance of companies with the same SIC code.⁶ We obtain industry-adjusted MSCI ESG performance for sample firms from 1991 to 2012. Then, for each year from 1992 to 2012, we measure changes in CSP by subtracting the industry-adjusted MSCI ESG score in the previous year from the industry-adjusted score in the current year and rank the differences in performance from high to low. Firms with the highest change in performance are those with the most improvement in the engagement of CSR activities. Likewise, firms with the lowest performance difference are those with the most deterioration in CSP. For each dimension and each year, we construct the up (down) portfolio with firms whose performance differences are ranked within the top (bottom) 20% among all firms. We construct the up-down portfolio by going long in the up portfolio and short in the down portfolio. Furthermore, we construct the up (down) portfolio of overall social responsible performance by identifying firms that are in at least two up (down) portfolios but not in any of the down (up) portfolios. “ALL” is used to denote the overall social performance.

Using annual data from COMPUSTAT, we estimate the mean of total assets, market value, net income, revenue, debt ratio, and return on equity (ROE) of firms in the up and down portfolios for each year during 1992–2012.⁷ That is, we examine the financial data of the companies in the year when the portfolios are constructed.

Table 2 presents the average, standard deviation, and median of the aforementioned mean of financial variables over the period from 1992 to 2012. We conduct a paired t-test to see if there are significant differences in these financial variables between firms in the up and down portfolios. We find that for most dimensions there is no significant difference in the total assets, market value, sales revenue, financial leverage, or profitability between firms in the up and down portfolios. However, firms in the up portfolio of “diversity” show significantly higher ROE than the corresponding firms in the down portfolio. The average ROE’s of the up and down portfolios of diversity are equal to 0.13 and 0.04, respectively. The total asset of the up portfolio of “diversity” is also significantly higher than that of the down portfolio by $11.80 billion. Furthermore, the up portfolio of “product characteristics” has an average ROE of 0.16, which is significantly higher than that of 0.09 for the down portfolio. We also find that the up portfolio of “human rights” has significantly higher net income than the down portfolio, with an average value of $3349 million and $2073 million, respectively.

Table 2 Summary statistics of firms in the up and down portfolios

Empirical Methodology

Social Performance and Financial Performance

We use stock returns to measure a company’s financial performance because market-based measurements relate more closely to stockholders’ wealth. We follow Kempf and Osthoff (2007) by using a portfolio approach. Specifically, after constructing the up and down portfolios, we use monthly stock returns in the following year to compute the equal-weighted portfolio returns. Monthly stock returns are obtained from the Center for Research in Security Prices. We then examine if the risk-adjusted returns of the up, down, or up-down portfolio are significantly different from zero. The up-down portfolio is formed by the strategy that goes long in the up portfolio and short in the down portfolio. If engagement in CSR has a positive impact on financial performance, we should expect to see the up portfolio to have a positive risk-adjusted return. On the other hand, if engagement in CSR reflects an agency problem between managers and stockholders, an improvement in CSR, which requires commitment in firm resources, should lead to deterioration in financial performance. A superior performance of up-down portfolio suggests that firms with improvement in CSR outperform those with deterioration in CSR.

To estimate abnormal portfolio returns, we run regressions of monthly portfolio returns based on the four-factor model of Carhart (1997) as follows:

$$R_{t} - R_{t}^{f} = \alpha + \beta_{1} \left( {R_{t}^{m} - R_{t}^{f} } \right) + \beta_{2} SMB_{t} + \beta_{3} HML_{t} + \beta_{4} MOM_{t} + \varepsilon_{t} ,$$

(1)

where $R_{t}$ represents the portfolio return in month t, $R_{t}^{f}$ stands for the risk-free rate, and $R_{t}^{m}$ represents the return of the market portfolio. SMB, HML, and MOM denote the size, book-to-market, and momentum factors, respectively. The risk-free rates, excess market returns, and SMB, HML, and MOM factors are obtained from Kenneth French's website. Our focus is on the α coefficient, which is a measure of abnormal return.

Social Performance and Financial Performance during Crisis and Non-crisis Periods

Several studies suggest that social capital matters predominately during adverse market conditions or when a firm suffers from a negative event (Godfrey et al. 2009; Nofsinger and Varma 2014; Lins et al. 2017). On the other hand, Koh et al. (2014) indicate that the value enhancement of CSR is constrained when firms are not financially healthy. These arguments predict different relationships between CSP and financial performance during a crisis period. We employ the following specification to study the relationship separately during the non-crisis and crisis periods:

$$R_{t} - R_{t}^{f} = \left[ {\alpha^{NC} + \beta_{1}^{NC} \left( {R_{t}^{m} - R_{t}^{f} } \right) + \beta_{2}^{NC} SMB_{t} + \beta_{3}^{NC} HML_{t} + \beta_{4}^{NC} MOM_{t} } \right] \times I_{t}^{NC} + \left[ {\alpha^{C} + \beta_{1}^{C} \left( {R_{t}^{m} - R_{t}^{f} } \right) + \beta_{2}^{C} SMB_{t} + \beta_{3}^{C} HML_{t} + \beta_{4}^{C} MOM_{t} } \right] \times I_{t}^{C} + \varepsilon_{t} ,$$

(2)

where $I_{t}^{C}$ ($I_{t}^{NC}$) is a dummy variable equal to 1 if month t is in a crisis (non-crisis) period, and 0 otherwise. According to the National Bureau of Economic Research, the recession during the recent financial crisis is from December 2007 to June 2009. Since the portfolio is updated on a yearly basis, we compare the returns of the up and down portfolios from January 2008 to December 2009 to examine the relation during the crisis period.

Empirical Evidence

Baseline Regression Results

We run regressions on the monthly returns of the up and down portfolios on the Carhart (1997) four factors. We apply the Newey and West (1987) correction to all regressions in order to control for possible conditional heterogeneity and autocorrelation. To investigate the difference in performance between the up and down portfolios, we form an up-down portfolio by going long in the up portfolio and short in the down portfolio.

Table 3 presents the regression results of the Carhart four-factor model. In the top panel for the overall social responsible performance (ALL), the up portfolio has a significantly positive risk-adjusted return of 0.40% per month, whereas the risk-adjusted return of the down portfolio is not significantly different from zero. The evidence provides some support for the argument that improvement in CSR is positively related to financial performance in the future. Since the overall CSR measure is composed of several dimensions, each of which may have separate implications for a firm’s financial performance, we take a closer look at the relation by comparing the performance of the up, down, and up-down portfolios of each social performance dimension.

Table 3 Performance of equally weighted returns of the up, down, and up-down portfolios based on Carhart four-factor model

We indeed find the relation to vary with social performance dimensions. For environment, the risk-adjusted return of the up-down portfolio is 0.32% per month, which is significantly different from zero at the 10% level, suggesting a positive relation between environmental investment and financial performance. This is consistent with findings in Derwall, et al. (2005) and Jiao (2010). Similarly, improvement in human rights is found to be positively related to financial performance. The up portfolio of “human rights” has a positive risk-adjusted return of 0.49% per month, which is significant at the 10% level, and the up-down portfolio has an excess return of 0.98% per month, which is significant at the 5% level. When companies are involved in human rights violation, such as child labor, human exploitation or bad working conditions, their reputations and brand images are damaged. As consumers become concerned about buying products or services from firms involved in human rights violation, more and more companies incorporate human rights as part of their corporate social responsibility. Socially conscious consumers may be more willing to buy products or services from firms with good human rights reputation, which can lead to better financial performance as our result indicates.

As for employee relations, both the up and down portfolios have significantly positive risk-adjusted returns, equal to 0.28% and 0.29% per month, respectively. The up-down portfolio has a negative alpha but not significantly different from zero. The evidence that the up portfolio has significantly positive risk-adjusted returns is consistent with the results in the literature that companies with high employee satisfaction have better financial performance (Edmans 2011). However, we find that deterioration in employee relations does not necessarily decrease company value. Similar to employee relations, both the up and down portfolios in product characteristics have significantly positive risk-adjusted returns, but their difference is insignificant. For community and diversity, the risk-adjusted returns of the up, down, and up-down portfolios are not significantly different from zero.

In summary, we find some support that an improvement in the overall CSR is positively associated with financial performance in the future. However, the relation varies with CSR dimensions. CSR activities in environment and human rights are value enhancing. In the dimensions of employee relations and product characteristics, while an improvement in CSP is positively associated with financial performance, deterioration does not decrease company value. We do not find a significant relation between financial performance and CSP in the community and diversity dimensions.

Non-Crisis and Crisis Periods

Table 4 presents the regression results based on the Carhart four-factor model when the non-crisis and crisis periods are separated. The risk-adjusted returns of the up portfolios of the overall social responsible performance (ALL) are equal to 0.29% and 0.79% per month during the non-crisis and crisis periods, respectively, both of which are significantly different from zero at the 5% level. The down portfolio, on the other hand, does not outperform the market in either period. This suggests that improvement in CSR is value enhancing regardless of the market condition. Furthermore, the risk-adjusted return of the up portfolio during the crisis period is higher than that during the non-crisis period.

Table 4 Performance of equally weighted returns of the up, down, and up-down portfolios based on CARHART four-factor model with separation of non-crisis and crisis periods

Table 4 also presents returns of the up, down, and up-down portfolios for each CSR dimension. We find that the relation between CSP and financial performance changes with market conditions. For employee relations, for instance, during the non-crisis period, neither the up nor the down portfolio has superior performance. During the crisis period, however, the risk-adjusted return of the down portfolio is significantly higher than that of the up portfolio. This finding accords well with economic intuition. During the financial crisis, some companies may reduce their commitment to employee relations in order to improve financial performance. For example, firms may lay off employees or cut employee benefits during the financial crisis in order to reduce expenses or improve the bottom line. Our results suggest that firms, which are less committed to employee relations, have higher risk-adjusted returns during economic downturns. This finding is also consistent with Koh et al.’s (2014) argument that the value generation of CSR is constrained if firms have poor financial performance. Similarly, Wang and Qian (2011) show that firms’ engagement in prosocial activities may result in punishment from stakeholders if firms have a poor financial position.⁸

As for product characteristics, the up portfolio has better financial performance than the down portfolio during the crisis period. The risk-adjusted returns of the up-down portfolio is 0.87% per month, which is significant at the 5% level. A reverse pattern is found during the non-crisis period. The up-down portfolio yields a risk-adjusted return of -0.28% per month, which is significant at the 10% level. Furthermore, we find that the risk-adjusted return of the up-down portfolio during the crisis is significantly higher than during the non-crisis period. During the crisis period when society trust is low, firms with more reliable and safe products or services may be perceived as more trustworthy. These firms’ products or services therefore may be in greater demand during the crisis period, thereby leading to better financial performance. The non-unidirectional relation between financial performance and CSR at the employee relations and product characteristics dimensions as reported in the preceding section can be attributed to the changing relations in response to market conditions.

For environment, the risk-adjusted return of the up-down portfolio is positive in both periods, but is significant at the 10% level only during the crisis period. The abnormal return of the up-down portfolio during the crisis period is also higher than that during the non-crisis period at the 10% significance level. It suggests that improvement in environment is value enhancing, especially during the financial crisis. For human rights, the up-down portfolio earns a significantly positive risk-adjusted return of 1.61% at the 5% level during the crisis period, which is higher than the corresponding return of 0.90% during the non-crisis period. This result suggests that improvement in human rights may enhance firm value especially during the financial crisis. As for community and diversity, the risk-adjusted returns of the up-down portfolios are positive during the non-crisis period and negative during the crisis period, but they are statistically insignificant in either period.

In sum, improvement in CSR is in general value enhancing, regardless of the market conditions. However, the relation between CSR and financial performance changes with CSR dimensions and market conditions. The practice of CSR in environment, human rights, and product characteristics matters predominantly during the crisis period. Our finding is consistent with the argument made by Godfrey et al.(2009) and Shiu and Yang (2017) that CSR activities offer an insurance-like protection and firms engaging more in CSR activities receive less negative judgments from their shareholders in the face of adverse events. The pattern for employee relations is, however, reversed. We find that firms with deterioration in employee relations have better financial performance during the crisis period.

Robustness Checks

This section reports additional tests for the robustness of our findings to an alternative financial measure, potential reverse causation problems, an alternative sample period, possible confounding effects from other dimensions, and potential problems caused by changes in the number of strengths and concerns used in the MSCI ESG database over time.

Alternative Measure of Financial Performance

One concern about our results is that the risk-adjusted returns may reflect companies’ cost of equity capital instead of future cash flows due to CSR practices. That is, a positive (negative) risk-adjusted return can reflect a higher (lower) cost of equity capital which is not captured by the Carhart risk factors, instead of a value enhancement (deterioration) caused by social performance. The literature, however, does not support this explanation. For example, Derwall et al. (2011) find that the high risk-adjusted returns of socially responsible stocks are caused by the market’s slow reaction to the positive impact of CSR activities on firms’ future cash flows (see also Borgers et al. 2013). Their evidence is also consistent with Edmans (2011) that strong employee relations are associated with positive earnings surprises and abnormal returns surrounding earnings announcements. Also, El Ghoul, et al. (2011) show that firms with higher CSR scores have lower cost of equity capital.

However, to address this concern, we use Tobin’s Q as an alternative measure of financial performance and run the following model with year-fixed effects on firms in the up, down, and middle of the portfolios (Cai et al. 2012):

$$\Delta TobinQ_{i,t + 1} = \beta_{0} + \beta_{1} {\Delta }Asset_{i,t} + \beta_{2} \Delta DebtRatio_{i,t} + \beta_{3} \Delta \frac{CapExp}{{Asset}}_{i,t} + \beta_{4} \Delta \frac{R\& D}{{Asset}}_{i,t} + \beta_{5} \Delta SalesGrowth_{i,t} + \beta_{6} {\Delta }ROA_{i,t} + \beta_{7} UP_{i,t} + \varepsilon_{i,t + 1}$$

(3)

where Δ denotes change from base year t − 1, TobinQ denotes Tobin’s Q, Asset is book value of total assets, DebtRatio is determined by total debt over total assets, CapExp/Asset is capital expenditure over book value of total assets, R&D/Asset is R&D expenses over book value of assets, SalesGrowth denotes the growth rate of sales revenue, ROA is determined by operating income over total assets, and UP is a dummy variable that equals one if the firm is in the up portfolio and 0 otherwise.

Estimation results are reported in Table 5. We can see that the coefficient of the UP variable in the “ALL” regression is 0.1026, which is significantly greater than zero. This suggests that firms with an improvement in the overall CSR perform better financially as measured by Tobin’s Q than other firms. The coefficients of the UP variable are significantly positive for the employee relations and environment dimensions, positive but insignificant for community, human rights, and product characteristics, and negative without statistical significance for diversity. These results suggest that for all dimensions except diversity, firms with an improvement in CSP in general perform better financially than other firms, and the evidence is statistically significant in two dimensions (employee relations and environment). This finding is broadly consistent with our results reported in the preceding section when stock returns are used as a measure of financial performance (see Table 3).

Table 5 Tobin’s Q and corporate social performance

We further investigate the relationship between CSP and Tobin’s Q during the non-crisis and crisis periods by running the following model with year-fixed effects on all firms in the up, down, and middle of the portfolios:⁹

$$\Delta TobinQ_{i,t + 1} = \beta_{0} + \beta_{1} {\Delta }Asset_{i,t} + \beta_{2} \Delta DebtRatio_{i,t} + \beta_{3} \Delta \frac{CapExp}{{Asset}}_{i,t} + \beta_{4} \Delta \frac{R\& D}{{Asset}}_{i,t} + \beta_{5} \Delta SalesGrowth_{i,t} + \beta_{6} {\Delta }ROA_{i,t} + \beta_{7} Crisis_{t} + \beta_{8} UP_{i,t} + \beta_{9} UP_{i,t} \times Crisis_{t} + \varepsilon_{i,t + 1} ,$$

(4)

where Crisis_t is a dummy variable that equals one if year t + 1 is during the crisis period and zero otherwise.

Estimation results are presented in Table 6. We are interested in $\beta_{7}$, $\beta_{8}$, and $\beta_{9}$, namely, the coefficients of the variables Crisis, UP, and $UP \times Crisis$. Economically, $\beta_{7}$ measures how a firm’s Tobin Q changes during the crisis period relatively to the non-crisis period, $\beta_{8}$ shows the impact of CSP on Tobin’s Q for firms in the up portfolio during the non-crisis period, and $\beta_{9}$ captures the additional effect of CSP on Tobin’s Q for firms in the up portfolio during the crisis period relative to the non-crisis period. Notice that the full impact of CSP on Tobin’s Q for firms in the up portfolio during the crisis period is measured by the sum of the coefficients $(\beta_{8} + \beta_{9}$). The last row of Table 6 presents the estimate of this sum and its associated t-statistic.

Table 6 Tobin’s Q and corporate social performance with separation of non-crisis and crisis periods

We find that the coefficient of the Crisis variable is negative for all regressions, and is statistically significant for the overall social performance (ALL) and for the dimensions of diversity, employee relations, and human rights. These results imply that firms during the crisis period in general underperform those during the non-crisis period. The coefficient of the UP variable is positive for all regressions except that for the diversity dimension. Furthermore, this parameter is statistically significant for the overall social performance (ALL) and for the environment dimension, indicating that firms in the up portfolio for the overall social performance and for environment have better financial performance than other firms during the non-crisis period.

The coefficient of UP × Crisis is negative for the overall social performance, and for the community, diversity and employee relations dimensions, and is positive for the environment, human rights and product characteristics dimensions, but without statistical significance. We investigate the full impact of CSP on Tobin’s Q for firms in the up portfolio during the crisis period, which is captured by the sum of the coefficients of UP and UP × Crisis $(\beta_{8} + \beta_{9}$). The last row of Table 6 reports a positive estimate for all regressions except the diversity dimension. Furthermore, the full impact is significantly positive for the environment, human rights and product characteristics dimensions, implying that firms with improvements in these three dimensions financially outperform other firms during the crisis period, consistent with the findings in Table 4.

In summary, we find significant evidence that improvements in environment, human rights, and product characteristics are better rewarded during the crisis period. On the other hand, the value enhancement of improvement in community, diversity and employee relations, is insignificant during the crisis period. Our evidence may suggest that the positive impacts of social capital on financial performance when trust in the society is low can be offset by the negative impacts of CSR practices during the crisis period when financial resource is limited. The dominating factor can change such that the relationship between CSP and financial performance shifts differently for different CSR dimensions during the crisis period.

Reverse Causation

“Empirical Evidence” shows that firms with improvement in CSR are more likely to have higher risk-adjusted returns in the future. In other words, firms do well by doing good. However, it is also possible that firms with more engagement in CSR are already relatively more successful ones, or firms may be more likely to do good when they do well (e.g., Hong et al. 2012; Lys et al. 2015).

To check for such a possibility, we run regressions of monthly portfolio returns in the same year when the up, down, and up-down portfolios are constructed. If firms indeed do good when they do well, we should expect that firms in the up (down) portfolio to have superior (inferior) financial performance in the same year, and the up portfolio to outperform the down portfolio. In Table 7 Panels A and B, we report the risk-adjusted returns of the Carhart four-factor model for the full sample and the model with the separation of crisis and non-crisis periods, respectively.

Table 7 Performance of the up, down, and up-down portfolios in the year when the portfolio is constructed

In Panel A, the risk-adjusted return of the up portfolio is not significantly different from zero, whereas the risk-adjusted return of the down portfolio is significantly positive. This is in contrast to the pattern found in Table 3 where future financial performance of firms with improvement in CSR is significantly positive and higher than that of firms with deterioration in social performance. Compared to the results in Table 4 that future risk-adjusted returns of the up portfolio are significantly positive and higher than that of the down portfolio, the risk-adjusted returns of the up and down portfolios in Panel B are not significantly different from zero and the returns of the up portfolio are not higher than that of the down portfolio. Therefore, we do not find significant evidence that firms do good when they do well in our sample.

Subsample Analysis

To check if our findings hold in alternative sample periods, we consider the subsample period of 2004–2013 and conduct the same regression analysis on the monthly portfolio returns as we did previously.¹⁰ We select the subsample period of 2004–2013 for the following reasons. First, in 2001 MSCI ESG expanded to cover firms of the Russell 1000 index and in 2002 added firms of the Large Cap Social index. In 2003, all firms of the Russell 2000 and the Broad Market Social index are added into the MSCI ESG database. Due to the expansion of the database during the period 2001–2003, the observations in the early years are less comparable to those in the later years. Second, the database after 2003 covers the largest number of firms, allowing us to draw conclusions based on a larger dataset. Finally, the period before 2004 does not include the financial crisis period, thus preventing us from separating the effects for crisis and non-crisis periods. Table 8 presents the regression results.

Table 8 Performance of the up, down, and up-down portfolios based on Carhart four-factor model for subsample period from 2004 to 2013

In Table 8, the risk-adjusted return of the up portfolio of ALL is positive and significantly different from zero, which is consistent with the preceding finding. Similarly, improvement in both environment and human rights is value enhancing. Consistent with the preceding evidence when the full sample period is used, the relation between CSP at employee relations and financial performance is not unidirectional. When the non-crisis and crisis periods are separated, the risk-adjusted returns of the up-down portfolios of environment, human rights, and product characteristics are significantly positive during the crisis period. On the other hand, the portfolio of firms with deterioration in employee relations outperforms that of firms with improvement during the financial crisis. Overall, our finding is robust to alternative sample periods.

Possible Confounding Effect from Other Dimensions

To examine if the preceding results are influenced by possible correlations in the changes of different CSR dimensions, we conduct a robustness check by excluding the effects from other dimensions. Specifically, for each dimension, we construct the up (down) portfolio by including firms whose social responsibility changes are ranked within the top (bottom) 20% among all firms but excluding firms that are also in the up (down) portfolios of any other dimensions. The regression results are presented in Table 9. The patterns are in general similar to what we show previously. CSP at environment and human rights is value enhancing. Improvement in employee relations increases firm value, although mixed evidence is found that firm value is also positively related to deterioration in employee relations. The results for the other CSR dimensions, however, are not statistically significant, which may be caused by the smaller sample sizes due to the exclusion of firms that are in the up or down portfolios of multiple dimensions.

Table 9 Performance comparison with adjustment of confounding effects from other dimensions

Adjustment for Factor Changes

As shown in Table 1, the number of strengths and concerns considered by MSCI ESG for each social responsibility dimension changes over time. We conduct a robustness test by controlling for the changes in the strength and concern factors. Specifically, we construct portfolios as in “Data” but exclude monthly returns during those years if their portfolios are constructed in the years with a change in the number of strengths or concerns. The regression analysis based on the four-factor model is presented in Table 10. Consistent with the preceding results, both the up and down portfolios of employee relations and product characteristics have superior performance. Firms with improvement in human rights have positive risk-adjusted stock returns, which are significantly higher than the returns of firms with deterioration. Overall, our basic conclusions remain the same after the consideration of changes in the number of strengths and concerns in the MSCI ESG database over time.

Table 10 Performance comparison with adjustment of changes in strengths/concerns

Conclusions

We study the relationship between corporate social performance and financial performance by comparing the portfolio returns of firms with changes in corporate social responsibility (CSR) intensity. Overall, we provide supporting evidence that corporate social performance is positively related to financial performance, but the relationship varies with CSR dimensions and shifts during non-crisis and crisis periods. Improvement in environment, human rights, and product characteristics is better rewarded during the crisis period, whereas the value enhancement of good employee relations is stronger during the non-crisis period. Our results are robust to a battery of sensitivity tests, including an alternative measure of financial performance, a potential reverse causation problem, an alternative sample period, additional control for potential confounding effects from other CSR dimensions, and control for the changes in the number of strengths and concerns used in MSCI ESG database over time.

Appendix

See Table

Table 11 Strengths and concerns considered in MSCI ESG data in year 2012

11.