Corporate Social Responsibility and Cost Stickiness

Abstract

This article examines the effects on cost stickiness of firms’ involvement in corporate social responsibility (CSR) activities. Cost stickiness represents asymmetric cost behavior whereby the magnitude of cost increases in response to an increase in the activity level is greater than the magnitude of cost decreases with a decrease in the activity level. We hypothesize that CSR involvement requires ongoing investments in value-creating activities; hence, it is difficult to scale down committed resources instantly even when the activity declines. We use two different CSR proxies and find support for the CSR-related cost stickiness hypothesis. We further decompose CSR into strategic and tactical CSR and find that cost stickiness is more pronounced for strategic CSR. Finally, we examine the CSR-related cost behavior pattern across business cycles and find some evidence of cost stickiness during an expansionary phase of the economy and cost anti-stickiness during a recessionary phase but only for the tactical CSR component.

Keywords

corporate social responsibility (CSR)cost stickiness KLD strategic CSR tactical CSR

Companies are allocating significant portions of their expense budgets to CSR—$28 billion on sustainability and $15 billion on corporate philanthropy spent by large U.S. firms in 2010.

—Di Giuli and Kostovetsky (2014)

Corporate social responsibility (CSR) has become an integral part of conducting business around the world. Companies are allocating significant portions of their expenses to CSR-related activities. Because of its strategic and practical implications in the business world, a wide body of academic literature has emerged around CSR (Margolis & Walsh, 2003; Mattingly, 2017; Orlitzky, Schmidt, & Rynes, 2003; Wang, Dou, & Jia, 2016). Although research on the revenue side of strategic CSR abounds, there is a paucity of research on firm-level CSR investments and importantly on how such investments change with changes in the activity level. We attempt to fill this void in the literature.

Stakeholder theory argues that firms perform in a socially responsible manner by looking after the interests of their stakeholders (McGuire, Dow, & Argheyd, 2003). At the institutional level, firms engage in CSR to satisfy institutional pressures, particularly from stakeholders (Agle, Mitchell, & Sonnenfeld, 1999; Stevens, Kevin, Harrison, & Cochran, 2005). Studies based predominantly on agency and stakeholder theory claim that employing valuable firm resources to engage in CSR provides better access to valuable resources (Cochran & Wood, 1984; Waddock & Graves, 1997), attracts and retains higher quality employees (Greening & Turban, 2000; Turban & Greening, 1997), and allows for better marketing of products and services (Fombrun, 1996; Moskowitz, 1972). The managerial opportunism hypothesis, however, argues that managerial self-serving interests might lead to CSR overinvestment: an action detrimental to the interests of the stakeholders, creating a competitive disadvantage and affecting a firm’s value negatively (Bénabou & Tirole, 2010; Friedman, 1970; Preston & O’Bannon, 1997).

Both these views allude to long-term CSR investment, which is as much a response to external pressure as it is to firm-level resources. The level of resources that will be devoted to CSR activities in the short term depends mainly on the accessibility of resources that are not required for other purposes. Even though firms’ involvement in CSR is voluntary in nature, it entails significant costs.

From a cost behavior perspective, it is interesting to explore whether managers adjust resources devoted to CSR activities in response to changes in the firm-level activity levels, the so-called cost stickiness phenomenon. The theoretical perspective on cost stickiness relies on the notion that many costs, including investments in CSR-related activities, arise from managers’ deliberate resource commitment decisions. Once committed, it is not easy to scale back resources without incurring some kind of adjustment costs¹, defined as “economic sacrifices, social, contracting or psychological costs which emerge during the resource-adjustment process” (Venieris, Naoum, & Vlismas, 2015, p. 55). Therefore, to the extent that managers recognize the trade-offs arising because of adjustment costs, they will reduce CSR investments to a lesser extent when activity decreases than they will expand CSR investments when activity increases, thus generating cost stickiness (Anderson, Banker, & Janakiraman, 2003; Banker, Byzalov, & Chen, 2013; and for a comprehensive review of the cost stickiness literature, see Banker & Byzalov, 2014).

Considering that not all CSR investments share the same characteristics, we categorize CSR investments into strategic versus tactical CSR investments (Bansal, Jiang, & Jung, 2015). Strategic CSR is costly, long term in nature, and often irreversible. The irreversibility characteristic, in particular, is theoretically aligned with cost stickiness because of the stakeholder demand for continuous support for the environment, employees, and customers. Tactical CSR, however, requires fewer resources and is often reversible and, hence, less amenable to cost stickiness. We, therefore, expect that investments in strategic CSR will exhibit cost stickiness whereas investments in tactical CSR will generate cost anti-stickiness (reducing costs at a faster rate than the change in the activity level).

Given that CSR investments require resources, it is important to understand the implications for CSR-related cost stickiness conditional on the prevailing state of the economy. It is natural to expect that managers would be more inclined to cut back resources, including investments in CSR, during periods of economic recession (Campbell, 2007). We test whether the cost stickiness associated with CSR investments decreases during periods of economic recession. As long as the managerial resource adjustment costs are negligible, we would expect a proportionate decrease in CSR with decreases in revenue. However, as argued before, CSR, particularly strategic CSR, is not easily reversible and, hence, may continue exhibiting stickiness during recessionary periods as well but to a lesser extent than in expansionary periods.

We use the firm-level CSR scores from the KLD Research & Analytics, Inc. database (KLD) as our proxy for CSR activities and operating costs reported by firms as a crude proxy for CSR investments. We find that CSR investments, on average, exhibit cost stickiness that is more pronounced for strategic CSR than for tactical CSR. We also evidence a variation in cost stickiness related to CSR investments across environmentally sensitive versus other industries. We then include economic recession as a contextual variable and find some evidence of cost stickiness during an expansionary phase of the economy and cost anti-stickiness during a recessionary phase but only for the tactical CSR component.

Our study extends the CSR literature in three ways. First, we enrich the CSR literature by documenting the managerial trade-offs regarding resource adjustments. To the best of our knowledge, ours is the first study to use CSR as a contextual variable affecting cost stickiness. Di Giuli and Kostovetsky (2014) examine the association between CSR and selling, general, and administrative (SG&A) expense and document a significantly positive association between the two. However, they do not investigate the cost behavior pattern associated with CSR investments.² Second, we follow recent research on CSR to distinguish strategic versus tactical CSR theoretically and empirically (Bansal et al., 2015) and investigate the CSR-related cost behavior patterns for these two aspects separately. Our study suggests that strategic CSR is stickier than tactical CSR, which provides valuable implications for managerial CSR-related resource adjustment. Third, we extend the study by Bansal et al. (2015), who document a decrease in both strategic and tactical CSR during an economic recession but do not investigate the cost behavior pattern across CSR types and economic conditions. Our study also contributes to the cost stickiness literature by expanding our understanding of cost behavior in the light of firms’ involvement in CSR activities. Our findings demonstrate that resource allocation decisions regarding socially responsible behavior trigger the sticky cost phenomenon. It is plausible that the significant managerial benefits associated with CSR may motivate managers to maintain CSR investments even in the face of a decline in the activity level.

The remainder of this article consists of five additional sections. The “Literature Review and Hypotheses Development” section reviews the related literature and develops the hypotheses. The “Data and Method” section describes the data and methodology. The empirical results are described in the “Results” section. Finally, the “Discussion and Conclusion” section explains the implications and limitations of the study.

Literature Review and Hypotheses Development

CSR, commonly defined as voluntary firm actions to improve social or ecological conditions (McWilliams & Siegel, 2001), is a response to stakeholder demands (Freeman, 1984; Hillman & Keim, 2001; Rowley & Berman, 2000). CSR demands might include ensuring pollution-free environments, workplace diversity, and good working conditions for employees, support for education and housing, and high-quality products. CSR activities are often referred to as an effective tool to obtain support from stakeholders, ensure the effective use of an organization’s resources, gain favorable coverage from the media, signal legitimacy to the community, and lessen investors’ and employees’ scrutiny (Orlitzky et al., 2003).³ CSR can have a positive impact on firm performance through the provision of better access to valuable resources (Cochran & Wood, 1984; Waddock & Graves, 1997), attracting and retaining higher quality employees (Greening & Turban, 2000; Turban & Greening, 1997), allowing for better marketing of products and services (Fombrun, 1996; Moskowitz, 1972), and contributing to the gaining of social legitimacy (Alexander & Buchholz, 1978).

The concerns about CSR, however, have grown considerably in the last two-and-a-half decades among the business press, business and political leaders, customers, suppliers, community groups, and government (McWilliams & Siegel, 2001). The dramatic growth in the number of institutes and mutual funds screening stocks based on positive CSR behavior encourages corporations to be socially responsive. To cope with the increased attention paid to corporations’ impact on society, more than half of the Fortune 1000 companies in the United States issue CSR reports regularly, and nearly 10% of U.S. investments are screened to ensure that they meet CSR-related criteria (Galema, Plantinga, & Scholtens, 2008).⁴ Moreover, a growing number of firms worldwide have undertaken serious efforts to integrate CSR into various aspects of their business (Harjoto & Jo, 2011; Jo & Harjoto, 2011). According to the Sustainable and Responsible Investing (SRI) report published in 2014 (The Forum for Sustainable and Responsible Investment, 2014), SRI assets had grown by 76% since the beginning of 2012 to a total US$6.57 trillion, which manifests the dramatic increase in CSR investment as well as CSR-related disclosures in recent years.

Although highly desirable, CSR investments are costly, and divergent views remain on the desirability of CSR investments. The proponents argue that CSR investments generate a number of benefits, as discussed above, and that inadequate engagement in CSR may endanger organization legitimacy (Kondra & Hinings, 1998; Salancik & Pfeffer, 1978). The opponents, however, argue that investment in CSR is a waste of scarce resources and hence detrimental to shareholders’ interests (Bénabou & Tirole, 2010; Friedman, 1970; Preston & O’Bannon, 1997). McWilliams and Siegel (2001) argue that managers conduct a cost/benefit analysis to determine the level of resources to devote to CSR activities, that is, simultaneously assessing the demand for CSR and the cost of satisfying this demand, in determining the optimal CSR investment.

The costs associated with CSR investments can be categorized into sunk costs and recurrent costs. Generally speaking, environment-related CSR activities cause costs mainly in terms of capital (e.g., new equipment, improvement of energy use) and cause only minor recurrent costs (such as equipment updating and maintenance). By contrast, the recurrent costs of CSR activities that aim to improve the social aspects of business operations often exceed the capital costs. Given the increasing stakeholder demand to be socially responsible, firms are required to invest resources in undertaking CSR activities. However, once resources have been committed to CSR activities, it is not easy to scale back the investments, because such an act might jeopardize social legitimacy. For example, the social costs emanating from labor unemployment are non-trivial and, hence, adjusting labor resources downward is a costly exercise. Therefore, CSR investment is likely to increase with an increase in the activity level but less likely to decrease proportionally to a decrease in the activity level, thereby generating cost stickiness. According to Anderson et al. (2003), “. . . costs are sticky if the magnitude of the increase in costs associated with an increase in volume is greater than the magnitude of the decrease in costs associated with an equivalent decrease in volume” (p. 48).

The theoretical perspective on cost stickiness relies on the notion that many costs, including investments in voluntary CSR activities, arise from managers’ deliberate resource commitment decisions. A conventional model of cost accounting would suggest a proportional increase (decrease) in costs with an increase (decrease) in the activity level (Noreen, 1991). However, managers incur adjustment costs in reducing CSR investments, and a high level of adjustment costs prevents managers from reducing discretionary resource consumption, including CSR investments that are long term and irreversible in nature, in proportion to the reductions in the firm’s level of economic activity. For example, a “hybrid” version of an automobile would be more environmentally friendly than a standard automobile and might command a price premium in the market, given the social value of reduced pollution. If the sales of hybrid automobiles fall because of an economic crisis, the management will not necessarily abandon or cut down resources to the same extent as the decline in sales, as long as the demand shock is considered to be temporary. Therefore, to the extent that managers recognize the trade-offs that arise because of adjustment costs, they will reduce CSR investments to a lesser extent when activity decreases than they will expand CSR investments when activity increases, generating cost stickiness (Anderson et al., 2003; Banker et al., 2013). The following hypothesis tests this proposition:

Hypothesis 1 (H1): Costs related to CSR activities are sticky.

We expect the above hypothesis to be more pronounced for companies demonstrating strategic as opposed to tactical CSR. Following Bansal et al. (2015), strategic CSR is defined as “CSR [that] comprises activities with long time horizons, large resource commitments and significant structural adjustments” (p. 69). Strategic actions are more difficult to implement and often non-reversible. Some examples of strategic CSR include increasing employee diversity, ensuring product safety, and implementing environmental management systems, all of which have profound cost implications and are long term in nature. Tactical CSR, however, often requires fewer resources and shorter implementation times and is often reversible.⁵

Because strategic CSR requires resource commitments and is non-reversible in nature, the concept of cost stickiness is more appropriately linked to this category of CSR than to tactical CSR, which entails fewer resource commitments and is reversible in nature. Importantly from a cost stickiness perspective, tactical CSR investment might exhibit cost anti-stickiness, whereby costs fall proportionately more than a decline in the activity level. The following hypotheses are developed.

Hypothesis 2A (H2A): Strategic CSR exhibits cost stickiness.

Hypothesis 2B (H2B): Tactical CSR exhibits cost anti-stickiness.

We further expect CSR-related cost stickiness to be more pronounced for firms operating in environmentally sensitive industries (e.g., automobiles, pharmaceuticals, and oil and gas). Prior research finds industry, together with size, to be the most common variables in explaining the extent of CSR involvement (Adams, Hill, & Roberts, 1998; Cowen, Ferreri, & Parker, 1987; Gray, Kouhy, & Lavers, 1995). The results from these studies show that corporations from environmentally sensitive industries are associated with greater CSR involvement than corporations from other industries. In contrast, other industries, particularly newer manufacturing industries and the service sector, have significantly lower environmental impacts and are associated with fewer visible environmental issues. As a result, we would expect firms in the latter industries to be subject to significantly less stakeholder pressure regarding their environmental performance and to be more pressed to satisfy the interests of their financial stakeholders as opposed to other stakeholders (Reverte, 2009). Based on the preceding discussion, we develop the following hypothesis:

Hypothesis 2C (H2C): Firms from environmentally sensitive industries exhibit more CSR-related cost stickiness than those in other industries.

Because CSR involvement entails the commitment of valuable resources, it is important to consider the availability of resources to managers for committing to CSR investments. The extent to which the availability of economic resources determines the level of CSR investments has not been studied extensively. However, a few recent studies examine the effect of the business cycle on real economic decisions (Braun & Larrain, 2005; Covas & Haan, 2011, 2012; Karabarbounis, Macnamara, & McCord, 2014).

During an economic expansion, many lucrative investment opportunities allow firms to capitalize on the favorable economic conditions, expand their operations, and boost their earnings and desired growth. Thus, an economic expansion may reduce the uncertainty of future cash flows and, therefore, might prompt managers to increase their CSR investments with a concomitant increase in the activity level. Contrary to this, during a recession corporate sales decline, profits shrink, and dividends may slump or disappear entirely. Firms with above-average profits in one period might not maintain the same level of profits in subsequent periods, because the profits will be eroded by competitors (Glen, Lee, & Singh, 2001). As a result, during a recession, the cash flow uncertainty increases and managers may cut back resources for CSR as neither consumers nor shareholders feel that the marginal utility of social expenditure outweighs its marginal costs (Campbell, 2007; Kitzmueller & Shimshack, 2012). Bansal et al. (2015) provide empirical support for this proposition by documenting that managers reduced both tactical and strategic CSR during the recent global financial crisis but that the magnitude was greater for tactical than for strategic CSR.⁶ A direct implication for cost stickiness of Bansal et al.’s (2015) study is that the stickiness will decrease for both strategic and tactical CSR during an economic recession. However, given the greater degree of resource adjustment costs associated with strategic CSR, we expect the decrease in cost stickiness to be smaller for strategic than for tactical CSR. The following hypothesis tests this proposition:

Hypothesis 3 (H3): Cost stickiness will decrease for both strategic and tactical CSR during an economic recession but more so for tactical CSR than for strategic CSR.

Data and Method

Measurement of CSR

We use information from one of the most widely adopted CSR scoring standards, that is, the KLD data set. KLD compiles annual ratings of more than 3,000 publicly traded U.S. firms, including Standard and Poor (S&P) 500 firms and 150 firms from the Domini Social Index. KLD rates companies on a wide range of activities that reflect how well companies perform in social responsibility and in building relationships with various stakeholders. It captures more than 94 measurement items along seven social dimensions: community, diversity, employee relations, environment, corporate governance, human rights, and product safety. For each measure, KLD offers “strength” and “concern” (Waddock & Graves, 1997; Waldman, Siegel, & Javidan, 2006) for each firm year (the appendix).

Following prior studies (Attig, Cleary, El Ghoul, & Guedhami, 2014; Y. Kim, Li, & Li, 2014), we exclude the corporate governance dimension from our CSR score. Corporate governance is a set of mechanisms that allow the shareholders to reward and exert control on the managers. CSR, however, deals with social objectives and stakeholders other than shareholders (Servaes & Tamayo, 2013). We calculate a net score for each of the remaining six dimensions of CSR as the number of strengths minus the number of concerns. Thus, our primary dependent variable, CSR_NET, is the sum of the net score from each of the six CSR dimensions. To test H2, we sum the CSR strength scores of the environment, employee, product, and diversity categories to obtain Strategic CSR and use the CSR strength score of the community category as Tactical CSR, following Bansal et al. (2015). We obtain firm financial data from Compustat Fundamentals Annual database and stock return data from the Center for Research in Security Prices (CRSP).

Measurement of Cost Stickiness

The following model, proposed by Anderson et al. (2003), is used to capture cost stickiness:

\begin{array}{l} \ln \frac{O C_{t}}{O C_{t - 1}} = γ_{0} + γ_{1} \ln [\frac{R e v e n u e_{t}}{R e v e n u e_{t - 1}}] \\ + γ_{2} D e c r e a s e_d u m_{t} \times \ln [\frac{R e v e n u e_{t}}{R e v e n u e_{t - 1}}] + ε, \end{array}

where OC is operating costs. We use two versions of OC. OC1 is operating costs defined as sales revenue minus income before extraordinary items. OC2 is defined as sales revenue minus operating income after depreciation for firm i in year t. Given the qualitatively similar results, we report the results based on OC1 only. We use sales revenue as our activity level, as we assume that operating costs include CSR-related expenditures and vary in response to changes in sales. Decrease_dum_t takes the value of one when the sales revenues in year t are less than those in year t − 1 and zero otherwise. Coefficient γ₁ measures the percentage increase in OC with a 1% increase in sales revenue. The sum of the coefficient (γ₁ + γ₂) measures the percentage decrease in OC with a 1% decrease in sales revenue. A significantly positive coefficient for γ₁ and a significantly negative coefficient for γ₂ would confirm cost stickiness.

Because we are interested in analyzing the cost behavior pattern of CSR investments, ideally we would have obtained the actual dollar amounts associated with CSR investments. However, we are constrained by the availability of such data and hence use operating costs, which are likely to include some investments associated with CSR activities. For example, many KLD categories (especially KLD “strengths”) are programs that the firm can institute by spending money: extra spending that would show up in higher levels of operating costs. Examples of such programs include charitable giving, work/life benefits such as child care, pollution prevention, employee health and safety programs, and quality control. The amount disbursed for philanthropic reasons has to be expensed within the operating cost section of the income statement. Another example relates to R&D investments. Accounting standards require immediate expensing of research costs. Hence, R&D investments for environmental protection, for example, have to be reported within the operating costs section of the income statement. Yet another example could be employee welfare expenses (e.g., the existence of cash profit sharing and employee work safety programs) (Di Giuli & Kostovetsky, 2014). Therefore, we would expect a positive correlation between KLD scores and operating costs, all else being equal. However, in the absence of data on the actual spending on CSR activities, our choice of operating costs as a proxy variable needs to be interpreted with a certain degree of caution. Despite the difficulties associated with estimating the actual contribution of CSR costs to the firm-level operating costs, we attempt to infer a rough estimate of the proportion of CSR costs vis-à-vis the total operating costs. However, such an approach is far from perfect, and some of the limitations of the estimated proxy are discussed later.

The extant research, too, adopts different approaches regarding the choice of the appropriate cost variable. Di Giuli and Kostovetsky (2014) use a scaled measure of SG&A expense (SG&A / (SG&A − advertising expenses)) instead of the actual CSR costs to determine the increase in CSR-related costs. J. B. Kim and Wang (2014) find that state-level unemployment insurance aimed at protecting labor unemployment costs reduces cost stickiness in SG&A. They model SG&A as the cost proxy, although this includes costs other than labor costs. Banker et al. (2013) investigate the effects of employment protection legislation but consider operating costs instead of labor costs as the dependent variable. Dierynck, Landsman, and Renders (2012), however, use labor costs as the cost proxy, because they specifically test for the labor cost stickiness when managers want to meet or beat a zero-earnings benchmark.

Empirical Model

We estimate the following comprehensive model that incorporates cost stickiness, CSR effects on cost stickiness, and other firm-specific determinants of cost stickiness (Anderson et al., 2003; Chen, Lu, & Sougiannis, 2012).

\begin{array}{l} \ln \frac{O C_{t}}{O C_{t - 1}} = γ_{0} + γ_{1} \ln [\frac{R e v e n u e_{t}}{R e v e n u e_{t - 1}}] + \\ γ_{2} D e c r e a s e_d u m_{t} \times \ln [\frac{R e v e n u e_{t}}{R e v e n u e_{t - 1}}] + γ_{3} {C S R}_{t} + \\ γ_{4} C S R_{t} \times \ln [\frac{R e v e n u e_{t}}{Re v e n u e_{t - 1}}] + γ_{5} C S R_{t} \times D e c r e a s e_d u m_{t} \times \ln [\frac{R e v e n u e_{t}}{R e v e n u e_{t - 1}}] + \\ \sum_{m = 6}^{9} γ_{m} E c o n_V a r_{m},_{t} \times D e c r e a s e_d u m_{t} \times \ln [\frac{R e v e n u e_{t}}{R e v e n u e_{t - 1}}] \\ + \sum_{n = 10}^{13} γ_{n} E c o n_V a r_{n, t} + ε . \end{array}

Our variable of primary interest is γ₅: the interaction between CSR and cost stickiness. A negative and significant γ₅ would confirm cost stickiness associated with CSR investments. We use four different specifications of CSR: net CSR (CSR_NET), industry-adjusted CSR (CSR_IND), strategic CSR (Strategic CSR), and tactical CSR (Tactical CSR). ECON_VAR are economic variables and include asset intensity (ASST_INTEN), measured as the total assets divided by the sales revenue for year t; employee intensity (EMP), the ratio of the total number of employees over sales; successive decrease (SUC_DEC), which is an indicator variable that is equal to 1 if the revenue in year t − 1 is less than the revenue in t − 2 and 0 otherwise; and stock performance (RETURN), measured as the raw stock return (from CRSP). All these stand-alone economic variables are interacted with γ₂.

We estimate the above regression for environmentally sensitive and other industries separately to test H2C. If the cost stickiness is greater in the former industries, then the interactive coefficient, γ₅, would be more negative for the sensitive industries. We adopt a similar procedure to test H3, whereby we estimate the above regression for economic expansion and for economic recession periods separately.

Sample and Summary Statistics

We began with an initial sample of 40,518 firm-year observations from 1991 to 2013 available from the KLD database. We matched these data with COMPUSTAT and lost 4,033 firm-year observations. We then excluded 9,203 firm-year observations pertaining to utility (two-digit SIC codes 48 & 49) and financial institutions (two-digit SIC codes 60-69). Finally, we lost another 5,325 firm-year observations because of missing values for the regression variables. Our final sample for conducting the regressions contains 21,957 firm-year observations.

Table 1 provides descriptive statistics for the variables used for the regression analyses. The mean (median) of the log of both the OC and the revenue ratio is 0.09 (0.08). The average CSR_NET is −0.03 with a large standard deviation, although the overall median is zero, suggesting a relatively balanced distribution of firms with negative and positive CSR performance. The mean Strategic (Tactical) CSR is 1.24 (0.16). The average CSR components range from a low of −0.02 (EMP_NET) to a high of 0.11 (COM_NET). The sample firms use US$1.33 million (median = 0.71) of assets to support each million dollars in sales revenue. The median firm has not experienced two consecutive years of sales decreases in the past 2 years (median = 0.00, M = 0.24), and the mean (median) raw stock return in the year is 0.19 (0.12). The firm-year observations come from a wide variety of industries, with the two-digit SIC codes 35 to 39 and 70 to 79 commanding the largest industry representation in our sample, as is evident from Table 2. Finally, Table 3 provides the correlation analysis. Most of the correlations are significant at the conventional level. The correlations between LN_OC and different CSR proxies are significantly negative (e.g., the correlation is −.06 for Strategic CSR). However, the negative correlation is due to our use of the changes in operating costs. The correlation becomes significantly positive when changes in CSR instead of the CSR level are used (untabulated).

Table 1.

Descriptive Statistics.

Variables	M	SD	First quartile	Median	Third quartile
Ln [OC_t/OC_t−₁]	0.09	0.24	0.00	0.08	0.18
Ln [Revenue_t/Revenue_t−₁]	0.09	0.21	0.00	0.08	0.18
Decrease_dum	0.25	0.43	0.00	0.00	0.00
CSR_NET	−0.03	2.41	−1.00	0.00	1.00
CSR_IND	0.39	0.25	0.20	0.33	0.55
Strategic CSR	1.24	2.01	0.00	0.00	2.00
Tactical CSR	0.16	0.50	0.00	0.00	0.00
ASST_INTEN	1.33	1.10	0.71	1.05	1.59
EMP	8.38	1.73	7.16	8.39	9.54
SUC_DEC	0.24	0.42	0.00	0.00	0.00
RETURN	0.19	0.55	−0.12	0.12	0.40

Note. OC are defined as sales minus income before extraordinary items. Decrease_dum takes the value of one when sales revenues in year t are less than those in year t − 1 and zero otherwise. CSR_NET, the net CSR score, is estimated as the total strengths minus total concerns across the main six social rating areas: community, diversity, employee relations, environment, human rights, and product. CSR_IND is the industry-level CSR score that ranges from zero to one and is calculated using the formula below:

$C S R_I N D_{i, t} = \frac{C S R_N E T_{i, t} - M I N_C S R_N E T_{j, t}}{M A X_C S R_N E T_{j, t} - M I N_C S R_N E T_{j, t}},$

where, i, j, and t denote firm i, industry j (two-digit SIC codes), and year t, respectively. Moreover, MIN and MAX refer to the minimum and maximum CSR_NET for firm i’s industry in year t, respectively. Strategic CSR is the sum of the strengths in the diversity, employee, environment, and product characteristics whereas Tactical CSR is the total CSR strength score from the community category following Bansal, Jiang, and Jung (2015). Asset intensity (ASST_INTEN) is total assets divided by sales revenue for year t; employee intensity (EMP) is the ratio of total number of employees over sales; successive decrease (SUC_DEC) is an indicator variable that is equal to 1 if revenue in year t − 1 is less than revenue in t − 2, and 0 otherwise; and stock performance (RETURN) is the raw stock return (from CRSP).

Table 2.

Industry Distribution.

Code	Industry	Observations	% observations
1-14	Agriculture and mining	1,397	0.06
15-17	Building construction	331	0.02
20-21	Food and kindred products	844	0.04
22-23	Textile mill products and apparels	349	0.02
24-27	Lumber, furniture, paper, and printing	1,177	0.05
28-30	Chemical, petroleum, and rubber and allied products	2,373	0.11
31-34	Metal	986	0.04
35-39	Machinery, electrical, computer equipment	6,454	0.29
40-47	Railroad and other transportation	1,180	0.05
50-51	Wholesale goods, building materials	841	0.04
53-59	Store merchandise, auto dealers, home furniture stores	2,168	0.10
70-79	Business services	3,212	0.15
80-99	Others	645	0.03
	Total	21,957	1.00

Table 3.

Correlation Analysis.

	LN_OC	ΔSALE	STICKY	CSR_NET	CSR_IND	Strategic CSR	Tactical CSR
LN_OC	1.00
ΔSALE	.74***	1.00
STICKY	.49***	.66***	1.00
CSR_NET	−.03***	−.04***	.04***	1.00
CSR_IND	−.03***	−.04***	−.02**	.66***	1.00
Strategic CSR	−.06***	−.08***	.001	.77***	.54***	1.00
Tactical CSR	−.04***	−.05***	.02**	.51***	.38***	.53***	1.00

Note. Variable definitions are in Table 1.

p < .05. ***p < .01.

Regression Results

CSR Investments and Cost Stickiness

Table 4 presents the regression results for the effect of CSR investments on cost stickiness. We estimate the asymmetrical adjustment of operating costs at the firm level with firm-clustered standard errors (Gow, Ormazabal, & Taylor, 2010; Petersen, 2009). Industry and year dummies are included in all our regression models. Column (1) shows that the coefficient for γ₁ is 0.87 (t statistic = 56.96, p < .001). This indicates that operating costs increase by about 0.87% for a 1% increase in sales revenue. The estimated value of γ₂ is −0.08 (t statistic = −1.64). The combined value of γ₁ + γ₂ = 0.79 indicates that operating costs decrease by about 0.79% per 1% decrease in sales revenue, reflecting cost stickiness. Our reported coefficients vary significantly from those of Anderson et al. (2003) probably because of our use of operating income before extraordinary items instead of SG&A expense as the cost measure and our use of a different sample period. The coefficient for CSR itself is negative and significant, although intuitively it should have been positively related to costs. This finding is explained by the fact that, although the dependent variable is expressed in a change version, the CSR variable is not (the correlation between LN_OC and ΔCSR is .03, significant at p < .01).

Table 4.

Regression Results: CSR Investments and Cost Stickiness.

$\begin{array}{l} l n \frac{O C_{t}}{O C_{t - 1}} = γ_{0} + γ_{1} l n [\frac{R e v e n u e_{t}}{R e v e n u e_{t - 1}}] + γ_{2} D e c r e a s e_d u m_{t} \times l n [\frac{R e v e n u e_{t}}{R e v e n u e_{t - 1}}] + γ_{3} C S R_{t} + γ_{4} C S R_{t} \times l n [\frac{R e v e n u e_{t}}{R e v e n u e_{t - 1}}] \\ + γ_{5} C S R_{t} \times D e c r e a s e_d u m_{t} \times l n [\frac{R e v e n u e_{t}}{R e v e n u e_{t - 1}}] + \sum_{m = 6}^{9} γ_{m} E c o n_V a r_{m, t} \times D e c r e a s e_d u m_{t} \times l n [\frac{R e v e n u e_{t}}{R e v e n u e_{t - 1}}] + \sum_{n = 10}^{13} γ_{n} E c o n_V a r_{n, t} + ε . \end{array}$ (2)

Variables	Predicted sign	(1)	(2)	(3)	(4)
Variables	Predicted sign	CSR_NET	CSR_IND	Strategic CSR	Tactical CSR
γ₁: $l n [\frac{R e v e n u e_{t}}{R e v e n u e_{t - 1}}]$	+	0.871*** [56.96]	0.815*** [26.51]	0.845*** [44.46]	0.818*** [53.94]
γ₃: CSR_t	?	−0.003*** [−4.01]	−0.019** [−2.38]	−0.003*** [−4.79]	−0.007*** [−4.82]
Two-way interaction terms
γ₂: $D e c r e a s e_d u m_{t} \times l n [\frac{R e v e n u e_{t}}{R e v e n u e_{t - 1}}]$	−	−0.08* [−1.64]	0.021 [0.33]	−0.036 [−0.70]	0.035 [0.88]
γ₄: CSR_t × $l n [\frac{R e v e n u e_{t}}{R e v e n u e_{t - 1}}]$	?	0.020*** [3.17]	0.130** [1.98]	0.022*** [3.28]	0.074*** [5.07]
Three-way interaction terms
γ₅: CSR_t × γ₂	−	−0.037*** [−3.73]	−0.202* [−1.83]	−0.03*** [−2.67]	−0.05 [−1.44]
γ₆: ASST_INTEN_t × γ₂	−	−0.024* [−1.76]	−0.023* [−1.73]	−0.062** [−2.49]	−0.066*** [−3.28]
γ₇: EMP_t × γ₂	−	0.060*** [6.02]	0.062*** [6.32]	0.048*** [3.01]	0.033** [2.34]
γ₈: SUCDEC_t × γ₂	+	0.160*** [5.02]	0.164*** [5.11]	0.077** [2.15]	0.056** [1.96]
γ₉: RETURN_t × γ₂	?	0.005 [0.27]	0.008 [0.39]	0.095*** [3.39]	0.043* [1.81]
Main terms
γ₁₀: ASST_INTEN_t	?	0.006** [2.30]	0.006** [2.35]	−0.000 [−0.14]	0.000 [0.07]
γ₁₁: EMP_t	?	0.001* [1.74]	0.001* [1.81]	0.003*** [3.63]	0.002*** [2.69]
γ₁₂: SUCDEC_t	?	−0.034*** [−11.84]	−0.034*** [−12.03]	−0.029*** [−9.42]	−0.024*** [−10.72]
γ₁₃: RETURN_t	?	−0.023*** [−8.29]	−0.023*** [−8.11]	−0.047*** [−14.07]	−0.033*** [−14.65]
Constant		0.005 [0.35]	0.016 [0.97]	0.011 [1.03]	0.018** [2.07]
Industry Fixed Effects		Yes	Yes	Yes	Yes
Year Fixed Effects		Yes	Yes	Yes	Yes
Observations		21,957	21,957	21,957	21,957
Adj. R²		.56	.58	.59	.74

Note. Robust t statistics in brackets. Variable definitions are in Table 1.

p < .10. **p < .05. ***p < .01.

We incorporate different specifications of CSR variables into the regression and an interactive variable, $C S R_{t} \times D e c r e a s e_d u m_{t} \times$ $\ln [\frac{R e v e n u e_{t}}{R e v e n u e_{t - 1}}] .$ We should expect a negative and significant coefficient for the interactive variable to support our hypothesis that CSR-related investments generate cost stickiness. We find support for our prediction, as the coefficient for $C S R_{t} \times D e c r e a s e_d u m_{t} \times \ln [\frac{R e v e n u e_{t}}{R e v e n u e_{t - 1}}]$ is negative and significant in column 1 (coefficient = −0.037, t statistic = −3.73, p < .001). Cost stickiness is also evident for CSR_IND (column 2; coefficient = −0.20, t statistic = −1.83, significant at p < .10). Columns 3 and 4 report the results for the cost behavior of Strategic CSR and Tactical CSR, respectively. The coefficient for $C S R_{t} \times D e c r e a s e_d u m_{t} \times \ln [\frac{R e v e n u e_{t}}{R e v e n u e_{t - 1}}]$ is negative and significant for the former (coefficient = −0.03, t statistic = −2.67, p < .01) but insignificant for the latter group (coefficient = −0.05, t statistic = −1.44). This supports H2B: It would be more costly to adjust resources devoted to strategic CSR than tactical CSR.⁷

With respect to the economic determinants of cost stickiness, we observe a significantly negative coefficient for ASSTINT × γ₂, implying a greater degree of cost stickiness for firms that require relatively more assets to support their operations. The coefficient for SUC_DEC × γ₂ is significantly positive, suggesting a lower degree of operating cost asymmetry in firms experiencing negative demand shocks in two consecutive years. However, unlike Anderson et al. (2003), we find a significantly positive coefficient for employee intensity EMP × γ₂, suggesting a lower degree of cost asymmetry in firms that require relatively more employees to support their operations.⁸ Our finding is consistent with Chen et al. (2012) and Dierynck et al. (2012). The coefficient for RETURN × γ₂ is insignificant. The explanatory power of the four models ranges from 56% to 59%.⁹

Table 5 reruns regression Model 2 after replacing OC with other cost components that make up operating costs. We include the cost of goods sold (COGS), SG&A, R&D costs, and labor costs as the four cost components and report the results. Although investments in CSR are a component of OC, CSR costs could also be included in different cost components. For example, R&D costs incurred to manufacture innovative products can be reflected in the company-wide R&D expenditure. Similarly, CSR investments with respect to employee relations could be incorporated into company-wide labor expenses. J. B. Kim and Wang (2014) follow a similar procedure. We find that CSR costs exhibit stickiness across all but labor cost components. For example, the coefficient for $C S R_{t} \times D e c r e a s e_d u m_{t} \times \ln [\frac{R e v e n u e_{t}}{R e v e n u e_{t - 1}}]$ is −0.032 with a t statistic of −3.12 (p < .01) when we use SG&A instead of OC. For the COGS category, the corresponding coefficient is −0.026, with a t statistic of −2.55 (p < .05). The insignificant coefficient for labor cost categories is most likely to be due to the small number of valid observations.

Table 5.

CSR Investments and Cost Stickiness: Cost Component Analysis.

Variables	Predicted sign	(1)	(2)	(3)	(4)
Variables	Predicted sign	CSR_NET	CSR_IND	Strategic CSR	Tactical CSR
SG&A
γ₁: $l n [\frac{R e v e n u e_{t}}{R e v e n u e_{t - 1}}]$	+	0.615*** [39.83]	0.55*** [55.17]	0.579*** [33.47]	0.596*** [37.77]
γ_5:CSR_t × γ₂	−	−0.032*** [−3.12]	−0.22** [−1.97]	−0.040*** [−2.60]	−0.079 [−0.95]
Economic determinants × γ₂		Yes	Yes	Yes	Yes
Economic determinants		Yes	Yes	Yes	Yes
Industry Fixed Effects		Yes	Yes	Yes	Yes
Year Fixed Effects		Yes	Yes	Yes	Yes
Observations		21,877	21,821	21,877	21,877
Adj. R²		.49	.49	.49	.49
COGS
γ₁: $l n [\frac{R e v e n u e_{t}}{R e v e n u e_{t - 1}}]$	+	0.96*** [53.83]	0.91*** [27.26]	0.82*** [50.49]	0.84*** [58.83]
γ₅: CSR_t × γ₂	−	−0.026** [−2.55]	−0.23* [−1.78]	−0.043*** [−4.62]	0.02 [0.33]
Economic determinants × γ₂		Yes	Yes	Yes	Yes
Economic determinants		Yes	Yes	Yes	Yes
Industry Fixed Effects		Yes	Yes	Yes	Yes
Year Fixed Effects		Yes	Yes	Yes	Yes
Observations		21,877	21,821	21,877	21,877
Adj. R²		.61	.61	.61	.61
R&D
γ₁: $l n [\frac{R e v e n u e_{t}}{R e v e n u e_{t - 1}}]$	+	0.56*** [16.81]	0.41*** [8.13]	0.52*** [12.92]	0.53*** [15.23]
γ₅: CSR_t × γ₂	−	−0.04** [−2.17]	−0.54** [−2.21]	−0.03* [−1.71]	−0.18 [−1.46]
Economic determinants × γ₂		Yes	Yes	Yes	Yes
Economic determinants		Yes	Yes	Yes	Yes
Industry Fixed Effects		Yes	Yes	Yes	Yes
Year Fixed Effects		Yes	Yes	Yes	Yes
Observations		11,756	11,724	11,756	11,756
Adj. R²		.16	.16	.16	.16
LABOR COST
γ₁: $l n [\frac{R e v e n u e_{t}}{R e v e n u e_{t - 1}}]$	+	0.73*** [10.94]	0.54*** [4.74]	0.66*** [6.87]	0.66*** [7.78]
γ₅: CSR_t × γ₂	−	0.04 [0.60]	−0.51 [−0.76]	−0.06 [−0.60]	0.25 [1.47]
Economic determinants × γ₂		Yes	Yes	Yes	Yes
Economic determinants		Yes	Yes	Yes	Yes
Industry Fixed Effects		Yes	Yes	Yes	Yes
Year Fixed Effects		Yes	Yes	Yes	Yes
Observations		1,280	1,279	1,280	1,280
Adj. R²		.26	.26	.26	.26

Note. Robust t statistics in brackets. Variable definitions are in Table 1.

p < .10. **p < .05. ***p < .01.

Table 6 reruns regression Model 2 but separately for environmentally sensitive (columns 1 and 2) versus other industries (columns 3 and 4). We report the results for the CSR_NET, Strategic CSR, and Tactical CSR categories. Firms operating in environmentally sensitive industries would be more concerned about CSR investments in the environment, employee, and product safety categories because of greater stakeholder, including government, scrutiny emanating from the risks posed by these industries on the external environment. However, companies that do not operate in environmentally sensitive industries may experience a greater need to satisfy the interests of their financial stakeholders as opposed to other stakeholders to ensure continued access to financial resources and, thus, the survival of their business.

Table 6.

CSR Investments, Industry Orientation, and Cost Stickiness.

	Predicted sign	(1)	(2)	(3)	(4)	(5)	(6)
		CSR_NET		Strategic CSR		Tactical CSR
		Sensitive industries	Other industries	Sensitive industries	Other industries	Sensitive industries	Other industries
γ₁: $l n [\frac{R e v e n u e_{t}}{R e v e n u e_{t - 1}}]$	+	0.85*** [41.14]	0.81*** [31.75]	0.91*** [39.59]	0.90*** [35.40]	0.828*** [35.68]	0.892*** [34.49]
γ₃: CSR_t	?	−0.003*** [−3.34]	−0.004*** [−4.17]	−0.003*** [−3.08]	−0.007* [−1.78]	−0.012*** [−5.63]	−0.006 [−1.37]
Two-way interaction terms
γ₂: $D e c r e a s e_d u m_{t} \times l n [\frac{R e v e n u e_{t}}{R e v e n u e_{t - 1}}]$	−	−0.15** [−2.19]	−0.10 [−1.45]	0.05 [0.60]	0.073 [0.88]	−0.080 [−1.16]	0.100 [1.21]
γ₄: CSR_t × $l n [\frac{R e v e n u e_{t}}{R e v e n u e_{t - 1}}]$	?	0.03*** [2.76]	0.026*** [2.90]	0.012* [1.81]	0.06** [2.33]	0.118*** [5.32]	0.058** [2.24]
Three-way interaction terms
γ₅: CSR_t × γ₂	?	−0.04*** [−3.05]	−0.038** [−2.66]	−0.035*** [−2.71]	−0.10 [−0.87]	−0.128*** [−3.23]	−0.099 [−0.87]
Economic determinants_t × γ₂		Yes	Yes	Yes	Yes	Yes	Yes
Economic determinants_t		Yes	Yes	Yes	Yes	Yes	Yes
Year Fixed Effects		Yes	Yes	Yes	Yes	Yes	Yes
Observations		11,955	11,955	11,955	10,002	11,955	10,002
Adj. R²		.63	.63	.55	.55	.63	.55

Note. Wolf (1996) notes that all companies with manufacturing facilities under Standard Industrial Classification codes 20 through 39 are subject to both the Environmental Protection Agency’s Toxic Release Inventory (TRI) program and the Occupational Safety and Health Administration’s Hazard Communication Standards. Hence, sensitive industries are defined as observations belonging to these industries. Robust t statistics in brackets. Variable definitions are in Table 1.

p < .10. **p < .05. ***p < .01.

We find support for H2C, as firms operating in environmentally sensitive industries exhibit greater cost stickiness than firms operating in other industries. The coefficients for $C S R_{t} \times D e c r e a s e_d u m_{t} \times \ln [\frac{R e v e n u e_{t}}{R e v e n u e_{t - 1}}]$ are negative and significant for the CSR_NET (coefficient = −0.04, t statistic = −3.05, p < .01), Strategic CSR (coefficient = −0.04, t statistic = −2.71, p < .01), and Tactical CSR categories (coefficient = −0.13, t statistic = −3.23, p < .01). Cost stickiness is evident for other industries, too, but only for the CSR_NET category (coefficient = −0.04, t statistic = −2.66, p < .01).

CSR Investments and Cost Stickiness: The Business Cycle as a Moderating Variable

We now present evidence of differential cost stickiness behavior related to CSR investments during different economic cycles. During a recession, cash flow uncertainty induces managers to cut back resources, including CSR, as the marginal utility of social expenditure outweighs its marginal costs (Campbell, 2007; Kitzmueller & Shimshack, 2012). Bansal et al. (2015) find that managers reduced both tactical and strategic CSR during the 2008-2009 global recession but the magnitude was greater for tactical than for strategic CSR. We extend Bansal et al.’s (2015) study by investigating the changes in cost stickiness behavior for these two categories of CSR during expansionary and recessionary periods.

We use a well-accepted measure based on the National Bureau of Economic Research (NBER) business cycle classification for identifying different states of economic activity (i.e., expansion and contraction). The results are presented in Table 7. The coefficients for $C S R_{t} \times D e c r e a s e_d u m_{t} \times \ln [\frac{R e v e n u e_{t}}{R e v e n u e_{t - 1}}]$ are negative and significant for both the CSR_NET (column 1) and Strategic CSR (column 3) categories during an expansionary period (coefficients = −0.04 and −0.031, t statistic = −3.69 and −2.40, respectively). We hypothesized in H3 that both Strategic CSR and Tactical CSR would decrease during a recession but that the magnitude of the decrease would be greater for tactical than strategic CSR. However, the coefficient for strategic CSR continued to be negative, albeit insignificant (coefficient = −0.038, t statistic = −1.20; column 4). This finding may suggest that managers consider strategic CSR to be a strong signal of firms’ commitment to CSR initiatives even during a period of economic crunch. Reducing tactical CSR during a recession, therefore, seems to be the natural response of managers. A positive and significant coefficient would support this hypothesis. Column 6 reveals that the coefficient for $C S R_{t} \times D e c r e a s e_d u m_{t} \times \ln [\frac{R e v e n u e_{t}}{R e v e n u e_{t - 1}}]$ is indeed positive and marginally significant (coefficient = 0.12, t statistic = 1.77, p < .10).

Table 7.

Business Cycle, CSR, and Cost Stickiness.

Variables	Sign	(1)	(2)	(3)	(4)	(5)	(6)
		NBER		NBER		NBER
		Expansion	Recession	Expansion	Recession	Expansion	Recession
		CSR_NET	CSR_NET	Strategic CSR	Strategic CSR	Tactical CSR	Tactical CSR
γ₁: $l n [\frac{R e v e n u e_{t}}{R e v e n u e_{t - 1}}]$	+	0.862*** [56.73]	0.826*** [15.65]	0.844*** [42.82]	0.792*** [13.06]	0.591*** [34.86]	0.766*** [17.67]
γ₃: CSR_t	?	−0.002*** [−3.18]	−0.004** [−2.21]	−0.003*** [−4.57]	−0.005** [−2.29]	−0.016*** [−3.35]	0.006 [1.60]
Two-way interaction terms
γ₂: $D e c r e a s e_d u m_{t} \times l n [\frac{R e v e n u e_{t}}{R e v e n u e_{t - 1}}]$	−	−0.050 [−0.70]	−0.019 [−0.20]	−0.014 [−0.20]	−0.028 [−0.26]	−0.216*** [−3.30]	0.128* [1.75]
γ₄: CSR_t × $l n [\frac{R e v e n u e_{t}}{R e v e n u e_{t - 1}}]$	?	0.021*** [3.12]	0.016 [0.79]	0.022*** [3.07]	0.030 [1.31]	0.102** [2.23]	0.058 [0.96]
Three-way interaction terms
γ₅: CSR_t × γ₂	−	−0.043*** [−3.69]	−0.04 [−1.54]	−0.031** [−2.40]	−0.038 [−1.20]	−0.143 [−1.35]	0.118* [1.77]
Economic determinants_t × γ₂		Yes	Yes	Yes	Yes	Yes	Yes
Economic determinants_t	?	Yes	Yes	Yes	Yes	Yes	Yes
Industry Fixed Effects		Yes	Yes	Yes	Yes	Yes	Yes
Observations		17,932	3,945	17,932	3,945	17,932	3,945
Adj. R²		.59	.55	.59	.55	.47	.50

Note. Robust t statistics in brackets. Variable definitions are in Table 1.

p < .10. **p < .05. ***p < .01.

Endogeneity Test

Our analysis so far suggests that CSR involvement is positively associated with cost stickiness. However, the sign, magnitude, and statistical significance of these estimates may be biased if the cost stickiness associated with CSR activities, $C S R_{t} \times D e c r e a s e_d u m_{t} \times \ln [\frac{R e v e n u e_{t}}{R e v e n u e_{t - 1}}],$ and the error term (ϵ) are correlated. To alleviate this concern, we adopt a two-stage instrumental variable approach to re-examine the findings reported in Table 4. Motivated by prior studies (Harjoto & Jo, 2011; Zhang, Zhu, Yue, & Zhu, 2010) showing that firm advertising intensity is positively associated with both the probability and the amount of CSR, we use advertising intensity as an instrumental variable. Moreover, we use industry-level mean CSR cost stickiness as an instrument, because industry and firm-level cost stickiness are highly correlated. Table 8 (Section I) reports the first-stage regression results in which the endogenous variable, $C S R_{t} \times D e c r e a s e_d u m_{t} \times \ln [\frac{R e v e n u e_{t}}{R e v e n u e_{t - 1}}],$ is regressed on the selected instruments and the exogenous variables from our analysis in Table 4. As expected, the coefficients of the instrumental variables are significant at the conventional level, suggesting that $C S R_{t} \times D e c r e a s e_d u m_{t} \times \ln [\frac{R e v e n u e_{t}}{R e v e n u e_{t - 1}}]$ is associated significantly with the selected instrumental variables.

Table 8.

Endogeneity Test.

Section I: First-stage regressions	(1)	(2)
Section I: First-stage regressions	CSR_NET	CSR_IND
Explanatory variable	CSR_t × γ₂	CSR_t × γ₂
Instrument
Advertising intensity (Adv_Int)	0.059* [1.75]	0.012*** [2.77]
Industry mean CSR_t × γ₂	0.944*** [11.70]	0.053*** [14.72]
Unreported control variables included in regression
All variables in main test	Yes	Yes
Year Fixed Effects	Yes	Yes
Industry Fixed Effects	Yes	Yes
Observation (N)	21,780	21,731
Adjusted R²	.634	.633
Underidentification test
Kleibergen–Paap rk Lagrange multiplier(LM) statistic	139.093	201.910
p	.000	.000
Weak identification test
Corrected Cragg–Donald Wald F statistic	87.81	113.613
Stock and Yogo (2005) 10% maximal IV size (Critical value)	19.93	19.93
Test of overidentifying restrictions
Hansen J statistic	1.308	1.496
p	.253	.221
Section II: Second-stage regressions	Ln(OC_t/OC_t _{− 1})	Ln(OC_t/OC_t _{− 1})
Potentially endogenous instrumented variable
CSR × γ₂	−0.039** [−2.23]	−0.502** [−2.07]
All variables in main test	Yes	Yes
Year Fixed Effects	Yes	Yes
Industry Fixed Effects	Yes	Yes
Hausman test for the effect of CSR (coefficient 2SLS = coefficient OLS)
Cluster-robust F statistic	0.000	2.331
p	.998	.127

Note. 2SLS = two-stage least squares; OLS = ordinary least squares.Robust t statistics in brackets.

p < .10. **p < .05. ***p < .01.

In support of the instruments, we also conduct underidentification, weak identification, overidentifying restrictions, and Hausman’s endogeneity tests. In Table 8, the underidentification test results (Lagrange multiplier (LM statistic)) reveal that the excluded instruments are “relevant.” The weak instrument test results show that the excluded instruments are correlated with the endogenous regressors, because the Cragg–Donald Wald F statistic is greater than Stock and Yogo’s (2005) critical value (i.e., 19.93) at 10%. Thus, the Cragg–Donald Wald F statistic shows that a weak instrument is not a concern with our estimates. The results from Hansen’s overidentifying restrictions tests do not reject the null hypothesis (p > .10), suggesting that the instruments are uncorrelated with the error term and are correctly excluded from the second-stage regressions, a finding that reflects the validity of the instruments used for the 2SLS regression. Finally, Hausman’s (1978) test suggests that endogeneity is not a concern for our analysis. Importantly, the coefficients for $C S R_{t} \times D e c r e a s e_d u m_{t} \times \ln [\frac{R e v e n u e_{t}}{R e v e n u e_{t - 1}}]$ are negative and significant for both CSR_NET and CSR_IND, suggesting that our results remain qualitatively similar even after controlling for endogeneity.

Additional Analysis

Actual CSR Costs and Cost Stickiness

We use operating costs as a proxy for CSR costs, which is far from a perfect proxy for CSR investments because operating costs include many other costs that are unrelated to CSR investments. We explored the possibility of accessing any machine readable database with actual CSR cost data. We first checked BLOOMBERG Int., which provides around 440 data points on ESG (environmental, social, and governance). Unfortunately, very few items provide dollar amounts of CSR expenditure. Furthermore, most of the data points do not contain any values.

We then accessed the ASSET4 database, which provides more than 500 data points on ESG. This database was of some assistance, as we could identify firm-year observations with reported CSR dollar amounts in the environmental expenditures, donations, and environmental R&D expenditure categories. However, the number of firm-year observations with actual dollar amounts for these categories is small as well. For example, there are only about 1,100 firm-year observations in the environmental expenditure category. If both donations and environmental expenditures are considered, then the sample reduces to 610 firm-year observations. Using this sample, we calculated the proportion of CSR costs in the total operating costs ((environmental expenditure + donations) / operating costs). We did not consider environmental R&D, because very few observations have dollar values. We found that the share of CSR costs (from the above categories) in OC is approximately 7%. We then used this as a proxy for average CSR costs and used (7% × operating costs) in our main data set as the dependent variable. Using this approximation, we find the coefficient for $C S R_{t} \times D e c r e a s e_d u m_{t} \times \ln [\frac{R e v e n u e_{t}}{R e v e n u e_{t - 1}}]$ to be −0.028 (t statistic = −3.12, p < .01) for CSR_NET and −0.022 (t statistic = −2.23, p < .05) for Strategic CSR. The negative and significant coefficients imply the presence of cost stickiness. However, the results should be interpreted cautiously, because the reported 7% could have been much different had a representative sample with dollar values been used.

Additional Control for Agency Variables: Free Cash Flows (FCFs), Institutional Ownership, and Executive Compensation

Anderson et al. (2003) propose that part of cost asymmetry may be attributable to agency costs. Chen et al. (2012) test this proposition empirically and find support for it. However, they also find that strong corporate governance mechanisms attenuate cost asymmetries. In our additional test, we control for some of the agency and governance variables to minimize the omitted variables problem. We include free cash flows (FCF), measured as the cash flow from operating activities less common and preferred dividends, scaled by total assets (FCF accounts for 9% of the total assets for our sample firms), institutional shareholdings (the number of shares held by institutional investors divided by the total shares outstanding), and executive compensation (the natural log of the total compensation during year t, where the total compensation consists of salary, bonus, value of restricted stocks and options, and all other annual payouts retrieved from EXECUCOMP). Jian and Lee (2015) document a positive association between the CEO compensation and the normal level of CSR. We interact these stand-alone variables with γ₂ and rerun Model 2. Untabulated results show that the coefficient for $C S R_{t} \times D e c r e a s e_d u m_{t} \times \ln [\frac{R e v e n u e_{t}}{R e v e n u e_{t - 1}}]$ is negative and significant for CSR_NET (coefficient = −0.024, t statistic = −2.23, p < .05) and for Strategic CSR (coefficient = −0.029, t statistic = −2.03, p < .05).

Non-Zero CSR Observations

Our main analysis is based on a sample that includes missing CSR data (coded as zero in the database). We rerun our main regression Model 2 retaining only non-zero CSR observations. Untabulated results provide results that are qualitatively similar to those in the main analysis. For example, the coefficient for the interactive variable for CSR_NET is −0.036 (t statistic = −3.71, p < .01).

Decile Ranking of CSR and Cost Stickiness

A potential concern relating to the KLD database is that KLD has added and deleted item ratings over time. For instance, reporting on South African CSR strengths and concerns in the area of qualitative human rights was stopped in 1995, whereas labor rights strengths in the same area were added in 2002 and a volunteer program pertaining to a community strength was added in 2005. As a result, the CSR scores may not be comparable over time. To address this concern, we transform the respective CSR issues’ area scores and the CSR_NET score into decile ranks for each year. A higher value of this decile rank indicates a higher level of CSR performance. The regression results using the decile rank of CSR corroborate our earlier findings. More clearly, we find the coefficient for $C S R_D E C I L E \times D e c r e a s e_d u m_{t} \times \ln [\frac{R e v e n u e_{t}}{R e v e n u e_{t - 1}}]$ to be negative and significant (coefficient = −0.018, t statistic = −2.03, p < .05).

Alternative Scalar for the Dependent Variable

Balakrishnan, Labro, and Soderstrom (2014) suggest scaling the dependent variable with lagged sales rather than with the lagged total cost. The rationale is to avoid the non-constant cost response to activity changes, which is useful under varying proportions of fixed costs across firms. In the following specification, γ₁ is interpreted as the variable cost ratio.

\begin{array}{l} \frac{C o s t_{t} - C o s t_{t - 1}}{R e v e n u e_{t - 1}} = γ_{0} + γ_{1} \frac{R e v e n u e_{t} - R e v e n u e_{t - 1}}{R e v e n u e_{t - 1}} \\ + γ_{2} R e v_D e c_{t} \times \frac{R e v e n u e_{t} - R e v e n u e_{t - 1}}{R e v e n u e_{t - 1}} + γ_{3} C S R_{t} + γ_{4} C S R_{t} \times γ_{1} \\ + γ_{5} C S R_{t} \times γ_{2} + O t h e r c o n t r o l s + I n d u s t r y F E + Y e a r F E . \end{array}

The variables are defined as before. If the CSR-related costs are sticky, then we would expect the coefficient γ₅ to be negative and significant in this specification, too. Untabulated results confirm our prediction (coefficient = −0.02, t statistic = −3.67, p < .01) for CSR_NET. For Strategic CSR, the coefficient is again negative and significant (coefficient = −0.019, t statistic = −2.46, p < .05). We, therefore, conclude that our results are robust to an alternative cost stickiness specification.

Annual Regressions

Finally, we conduct the Fama and MacBeth (FM; 1973) cross-sectional regression. Untabulated results reveal comparatively weaker evidence of cost stickiness using the FM regression in contrast to the pooled regression. The coefficient for $C S R_{t} \times D e c r e a s e_d u m_{t} \times \ln [\frac{R e v e n u e_{t}}{R e v e n u e_{t - 1}}]$ is negative and significant for CSR_NET (coefficient = −0.046, t statistic = −3.34, p < .01) and Strategic CSR (coefficient = −0.048, t statistic = −2.51, p < .05). However, the coefficients are not significant for CSR_IND and Tactical CSR.

Discussion and Conclusion

This article explores the cost behavior of CSR investments and investigates whether firms’ CSR involvement gives rise to cost stickiness. Although research on the value implications of CSR abounds, controversies remain regarding the true benefits of CSR activities. One view holds that CSR exerts a positive impact on firm value, because it provides firms with competitive advantages by reducing the conflicts among different stakeholders and streamlining the operational and financial activities. The latter may include enhancing employee morale and productivity, charging a premium price for differentiated products, attracting socially responsible consumers and investors, lowering litigation costs, and so on (Cochran & Wood, 1984; Greening & Turban, 2000; Turban & Greening, 1997; Waddock & Graves, 1997). The other view holds that CSR activities destroy value because of the managerial incentives to overinvest in CSR activities (Friedman, 1970).

Given corporations’ increasing focus on CSR and the substantial cost associated with such activities, it is important to understand the cost behavior pattern of CSR investments. Although a conventional cost behavior model would suggest a proportionate change in CSR investments with a change in activity, in reality a more complex cost behavior pattern is likely to exist. Because CSR involvement is long term in nature and firms engage in CSR activities to satisfy diverse stakeholders, we argue that managers may be reluctant to downsize CSR investments, even when the activity levels decrease. However, when the activity levels increase, managers may make additional investments in CSR-related activities. Thus, owing to the fact that managers expand CSR investments when the activity increases but are less likely to reduce CSR investments when the activity decreases, CSR investment is likely to be sticky.

Using four different specifications of CSR (CSR_NET, CSR_IND, Strategic CSR, and Tactical CSR), we show that CSR-related costs exhibit cost stickiness, because they decrease less with a decrease in firm revenue. When the CSR score is decomposed into strategic and tactical CSR, we find that cost stickiness is more pronounced for strategic than tactical CSR. We also find some evidence that the CSR cost exhibits stickiness during an expansionary phase of the economy and cost anti-stickiness during a recessionary phase for tactical CSR, suggesting that an economic recession influences managerial decisions to adjust CSR investments. These findings are shown to be robust using a series of sensitivity tests.

Our study has important implications for understanding the CSR overinvestment hypothesis. The agency theory-based overinvestment hypothesis suggests that insiders tend to overinvest in CSR at the cost of shareholders to enhance their personal reputation as socially responsible executives and that such overinvestment adversely affects firm value (Barnea & Rubin, 2010; Friedman, 1970). The findings from our study suggest that stakeholders need to evaluate managerial decisions regarding CSR investments in the light of the fact that, once committed, investment in CSR is not always easy to adjust downwards in the event of a shock. This is because decisions about resource adjustment should be based on long-term comprehensive cost–benefit analysis of resource reduction, for example, the economic adjustment costs borne by the firm, loss of morale, or loss of reputation (Mahlendorf, 2009).

Our study also implies that, while evaluating CSR-related cost stickiness, stakeholders should consider the nature of CSR: tactical verses strategic CSR. Tactical CSR is short term in nature and requires the commitment of few organizational resources and fewer structural adjustments, whereas strategic CSR is long term in nature and requires large resource commitments and significant structural adjustments (Bansal et al., 2015). Therefore, when making a trade-off between long-term CSR commitments and short-term profit maximization, managers need to consider whether the CSR in question is tactical or strategic. In particular, if managers want to discontinue CSR to deliver short-term profit goals, then the natural response would be to reduce tactical CSR investments. Taken from this viewpoint, our study implies that CSR-related cost stickiness is a reflection of “good” cost stickiness, which is associated with optimal resource planning arising from long-term managerial orientation regarding the firm’s economic performance (Anderson et al., 2003; Banker & Byzalov, 2014).

Despite the importance of the research, we highlight a number of limitations due to which the findings from this study should be interpreted with caution. One area of concern relates to the use of operating costs as a proxy for investments in CSR, and another concern relates to the use of the KLD database for constructing the CSR index. In this study, we assumed operating costs to be a reasonable proxy for CSR investments. We reasoned that many KLD categories (especially KLD “strengths”) are programs that firms can undertake by spending monetary resources, which would be captured by operating costs. However, this approach does not allow us to quantify the actual contribution of CSR costs to the total operating costs. We, therefore, conducted an additional test whereby we retrieved the reported CSR costs (retrieved from a commercial database) as a proportion of the total operating costs, which we found to be approximately 7%. However, the number of firm-year observations with reported CSR costs is only about 5% of the total sample used in this study. This arguably questions the generalizability of our estimates. Furthermore, the amount of actual CSR investment is unlikely to be constant across industries, as environmentally sensitive industries are likely to make more CSR investments given the nature of their business operations. It would have been a more reasonable estimate if we could show that CSR investment is a certain fraction of the total operating costs for each industry. However, that assumption is difficult to validate empirically.

In this study, we used data from the KLD, which is the most comprehensive and prominent source of CSR data and is extensively used in scholarly research to operationalize the social performance of firms (Choi & Wang, 2009; Coombs & Gilley, 2005; Graves & Waddock, 1994; Hillman & Keim, 2001). An oft-cited argument for using KLD scoring as a proxy for actual CSR investments is provided by Waddock and Graves (1997), who note that “Where possible, KLD uses quantitative criteria to determine the rating (e.g., dollar amount paid in fines or penalties; percentage of employees receiving certain kinds of benefits)” (p. 367). Kempf and Osthoff (2007) also use KLD data to explore the effect of socially responsible investing on portfolio performance. However, a significant concern remains regarding the use of the KLD database for CSR research. Some of the inherent limitations of the KLD include the lack of connection with corporate values (Johnson & Greening, 1999), the lack of systematic theory for multidimensional sets of CSR categories (Mattingly & Berman, 2006; Rowley & Berman, 2000), the “crude” and “equal-weight” nature of the CSR score (Wood & Jones, 1995), the use of binary variables to indicate a “strength” or a “weakness” of a firm concerning a certain social issue that makes it impossible to differentiate between inferior and superior performers gradually (Schreck, 2011), the lack of an industry-specific weighting scheme to aggregate the CSR score (Surroca, Tribó, & Waddock, 2010), and the lack of an explanation for why certain aspects of social categories are selected and others excluded in the database (Wood & Jones, 1995). Entine (2003) argues that the KLD ratings are biased toward larger and growth companies concentrated in technology and financial sectors whereas they are biased against firms in the chemicals, natural resources, and energy industries owing to the difficulty of measuring the positive aspects of these businesses. Finally, the KLD measure is in fact a blend of CSR performance and CSR disclosure, which, to some extent, complicates the interpretation of the score. Despite these limitations, this study contributes to the better understanding of cost behavior regarding CSR investments.

Footnotes

Appendix

KLD Description of CSR Strengths and Concerns Across CSR Categories.

Community strengths • Charitable giving: The company has consistently given more than 1.5% of trailing 3-year net earnings before taxes to charity. The company has a notably innovative giving program that supports non-profit organizations. • Support for housing: The company is a prominent participant in public/private partnerships that support housing initiatives for the economically disadvantaged, for example, the National Equity Fund or the Enterprise Foundation. • Support for education: The company has either been notably innovative in its support for primary or secondary school education, or the company has prominently supported job-training programs for youth. • Non-U.S. charitable giving: The company has made a substantial effort to make charitable contributions abroad, as well as in the United States. To qualify, a company must make at least 20% of its giving, or have taken notably innovative initiatives in its giving program, outside the United States.	Community concerns • Investment controversies: The company is a financial institution whose lending or investment practices have led to controversies. • Negative economic impact: The company’s actions have resulted in major controversies concerning its economic impact on the community. These controversies can include issues related to environmental contamination, water rights disputes, and plant closings. • Disputes: The company has recently been involved in major tax disputes involving Federal, state, local or non-U.S. government authorities, or is involved in controversies over its tax obligations to the community.
Employee relations strengths • Cash profit sharing: The company has a cash profit-sharing program through which it has recently made distributions to a majority of its workforce. • Employee involvement: The company strongly encourages worker involvement and/or ownership through stock options available to a majority of its employees, gain sharing, stock ownership, sharing of financial information, or participation in management decision making. • Health and safety strength: The company is noted by the U.S. Occupational Health and Safety Administration for its safety programs	Employee relations concerns • Health and safety concern: The company recently has either paid substantial fines or civil penalties for willful violations of employee health and safety standards, or has been otherwise involved in major health and safety controversies. • Workforce reductions: The company has reduced its workforce by 15% in the most recent year or by 25% during the past 2 years, or it has announced plans for such reductions. • Retirement benefits concern: The company has either a substantially underfunded defined benefit pension plan, or an inadequate retirement benefits program.
Environmental strengths • Beneficial products and services: The company derives substantial revenues from innovative remediation products, environmental services, or products that promote the efficient use of energy, or it has developed innovative products with environmental benefits. • Pollution prevention: The company has notably strong pollution prevention programs including both emissions reductions and toxic-use reduction programs. • Recycling: The company either is a substantial user of recycled materials as raw materials in its manufacturing processes, or a major factor in the recycling industry. • Clean energy: The company has taken significant measures to reduce its impact on climate change and air pollution through use of renewable energy and clean fuels or through energy efficiency. • Property, plant, and equipment: The company maintains its property, plant, and equipment with above-average environmental performance for its industry.	Environmental concerns • Hazardous waste: The company’s liabilities for hazardous waste sites exceed US$50 million, or the company has recently paid substantial fines or civil penalties for waste management violations. • Regulatory problems: The company has recently paid substantial fines or civil penalties for violations of air, water, or other environmental regulations. • Ozone depleting chemicals: The company is among the top manufacturers of ozone depleting chemicals. The company’s legal emissions of toxic chemicals from individual plants into the air and water are among the highest of the companies followed by KLD. • Climate change: The company derives substantial revenues from the sale of coal or oil and its derivative fuel products, or the company derives substantial revenues indirectly from the combustion of coal or oil and its derivative fuel products.
Product strengths • Quality: The company has a long-term, well-developed, company-wide quality program, or it has a quality program recognized as exceptional in U.S. industry. • R&D/innovation: The company is a leader in its industry for research and development (R&D), particularly by bringing notably innovative products to market. • Benefits to economically disadvantaged: The company has as part of its basic mission the provision of products or services for the economically disadvantaged.	Product concerns • Product safety: The company has recently paid substantial fines or civil penalties, or is involved in major recent controversies or regulatory actions, relating to the safety of its products and services. • Marketing/contracting concern: The company has recently been involved in major marketing or contracting controversies, or has paid substantial fines or civil penalties relating to advertising practices, consumer fraud, or government contracting. • Antitrust: The company has recently paid substantial fines or civil penalties for antitrust violations such as price fixing, collusion, or predatory pricing, or is involved in recent major controversies or regulatory actions relating to antitrust allegations.

Note. This appendix provides detailed definitions of KLD strengths and concerns indicators in four major KLD CSR qualitative issue areas. CSR = corporate social responsibility.

Acknowledgements

We appreciate the helpful comments and suggestions of Barry Mitnick, the associate editor, Bryan Husted, and two anonymous reviewers.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Notes

Author Biographies

Ahsan Habib is an associate professor in accounting at Massey University, New Zealand. His areas of research interest include capital market–based accounting research, the interface between corporate governance and accounting information, corporate social responsibility, and audit quality. His articles have appeared in such journals as ABACUS, Accounting and Business Research, British Accounting Review, Business & Society, Corporate Governance: An International Review, Journal of Accounting Literature, and Journal of Contemporary Accounting & Economics.

Mostafa Monzur Hasan is a lecturer in the School of Economics and Finance, Curtin University, Australia. His research focuses on corporate finance, corporate governance, and corporate sustainability. He has published in Accounting and Business Research, Business & Society, Corporate Governance: An International Review, European Accounting Review, and Journal of Contemporary Accounting & Economics.

References

Adams

C. A.

Hill

W. Y.

Roberts

C. B.

(1998). Corporate social reporting practices in Western Europe: Legitimating corporate behaviour. British Accounting Review, 30, 1-21.

Agle

B. R.

Mitchell

R. K.

Sonnenfeld

J. A.

(1999). Who matters to CEOs? An investigation of stakeholder attributes and salience, corporate performance, and CEO values. Academy of Management Journal, 42, 507-525.

Alexander

G. J.

Buchholz

R. A.

(1978). Corporate social responsibility and stock market performance. Academy of Management Journal, 21, 479-486.

Anderson

Banker

R. D.

Janakiraman

(2003). Are selling, general and administrative costs “sticky”? Journal of Accounting Research, 41, 47-63.

Attig

Cleary

S. W.

El Ghoul

Guedhami

(2014). Corporate legitimacy and investment–cash flow sensitivity. Journal of Business Ethics, 121, 297-314.

Balakrishnan

Labro

Soderstrom

N. S.

(2014). Cost structure and sticky costs. Journal of Management Accounting Research, 26, 91-116.

Banker

R. D.

Byzalov

(2014). Asymmetric cost behavior. Journal of Management Accounting Research, 26(2), 43-79.

Banker

R. D.

Byzalov

Chen

L. T.

(2013). Employment protection legislation, adjustment costs and cross-country differences in cost behavior. Journal of Accounting & Economics, 55, 111-127.

Bansal

Jiang

G. F.

Jung

J. C.

(2015). Managing responsibly in tough economic times: Strategic and tactical CSR during the 2008-2009 global recession. Long Range Planning, 48, 69-79.

10.

Barnea

Rubin

(2010). Corporate social responsibility as a conflict between shareholders. Journal of Business Ethics, 97, 71-86.

11.

Barnett

M. L.

Darnall

Husted

B. W.

(2015). Sustainability strategy in constrained economic times. Long Range Planning, 48, 63-68.

12.

Bénabou

Tirole

(2010). Individual and corporate social responsibility. Economica, 77, 1-19.

13.

Braun

Larrain

(2005). Finance and the business cycle: International, inter-industry evidence. Journal of Finance, 60, 1097-1128.

14.

Campbell

J. L.

(2007). Why would corporations behave in socially responsible ways? An institutional theory of corporate social responsibility. Academy of Management Review, 32, 946-967.

15.

Carroll

A. B.

(1979). A three-dimensional conceptual model of corporate performance. Academy of Management Review, 4, 497-505.

16.

Chen

C. X.

Sougiannis

(2012). The agency problem, corporate governance, and the asymmetrical behavior of selling, general, and administrative costs. Contemporary Accounting Research, 29, 252-282.

17.

Cho

C. H.

Freedman

Patten

D. M.

(2012). Corporate disclosure of environmental capital expenditures. Accounting, Auditing & Accountability Journal, 25, 486-507.

18.

Choi

Wang

(2009). Stakeholder relations and the persistence of corporate financial performance. Strategic Management Journal, 30, 895-907.

19.

Clarkson

M. E.

(1995). A stakeholder framework for analyzing and evaluating corporate social performance. Academy of Management Review, 20, 92-117.

20.

Clarkson

P. M.

Richardson

G. D.

(2004). The market valuation of environmental capital expenditures by pulp and paper companies. The Accounting Review, 79, 329-353.

21.

Cochran

P. L.

Wood

R. A.

(1984). Corporate social responsibility and financial performance. Academy of Management Journal, 27, 42-56.

22.

Coombs

J. E.

Gilley

K. M.

(2005). Stakeholder management as a predictor of CEO compensation: Main effects and interactions with financial performance. Strategic Management Journal, 26, 827-840.

23.

Covas

Haan

W. J. D.

(2011). The cyclical behavior of debt and equity finance. American Economic Review, 101, 877-899.

24.

Covas

Haan

W. J. D.

(2012). The role of debt and equity finance over the business cycle. Economic Journal, 122, 1262-1286.

25.

Cowen

Ferreri

Parker

L. D.

(1987). The impact of corporate characteristics on social responsibility disclosure: A typology and frequency-based analysis. Accounting, Organizations and Society, 12, 111-122.

26.

Dierynck

Landsman

W. R.

Renders

(2012). Do managerial incentives drive cost behavior? Evidence about the role of the zero earnings benchmark for labor cost behavior in private Belgian firms. The Accounting Review, 87, 1219-1246.

27.

Dhaliwal

D. S.

O. Z.

Tsang

Yang

Y. G.

(2011). Voluntary nonfinancial disclosure and the cost of equity capital: The initiation of corporate social responsibility reporting. The Accounting Review, 86, 59-100.

28.

Di Giuli

Kostovetsky

. (2014). Are red or blue companies more likely to go green? Politics and corporate social responsibility. Journal of Financial Economics, 111, 158-180.

29.

El Ghoul

Guedhami

Kwok

C. C.

Mishra

D. R

. (2011). Does corporate social responsibility affect the cost of capital? Journal of Banking & Finance, 35, 2388-2406.

30.

Entine

(2003). The myth of social investing: A critique of its practice and consequences for corporate social performance research. Organization & Environment, 16, 352-368.

31.

Epstein-Reeves

(2010, December 15). Consumers overwhelmingly want CSR. Forbes. Retrieved from http://www.forbes.com/sites/csr/2010/12/15/new-study-consumers-demand-companies-implement-csr-programs/#4bf672135e1d

32.

Fama

E. F.

MacBeth

J. D.

(1973). Risk, return, and equilibrium: Empirical tests. The Journal of Political Economy, 81, 607-636.

33.

Fombrun

C. J.

(1996). Reputation: Realizing value from the corporate image. Boston, MA: Harvard Business School Press.

34.

The Forum for Sustainable and Responsible Investment. (2014). Annual report, US SIF and US SIF foundation. Retrieved from http://www.ussif.org/files/Publications/Annual_%20Report_14_FINAL.PDF

35.

Freeman

R. E.

(1984). Strategic management: A stakeholder approach. Boston, MA: Pitman.

36.

Friedman

(1970, September). The social responsibility of business is to increase its profits. The New York Times Magazine, 13, 32-33.

37.

Galema

Plantinga

Scholtens

(2008). The stocks at stake: Return and risk in socially responsible investment. Journal of Banking & Finance, 32, 2646-2654.

38.

Glen

Lee

Singh

(2001). Persistence of profitability and competition in emerging markets. Economics Letters, 72, 247-253.

39.

Gow

I. D.

Ormazabal

Taylor

D. J.

(2010). Correcting for cross-sectional and time-series dependence in accounting research. The Accounting Review, 85, 483-512.

40.

Graves

S. B.

Waddock

S. A.

(1994). Institutional owners and corporate social performance. Academy of Management Journal, 37, 1034-1046.

41.

Gray

Kouhy

Lavers

(1995). Corporate social and environmental reporting: A review of the literature and a longitudinal study of UK disclosure. Accounting, Auditing & Accountability Journal, 8(2), 47-77.

42.

Greening

D. W.

Turban

D. B.

(2000). Corporate social performance as a competitive advantage in attracting a quality workforce. Business & Society, 39, 254-280.

43.

Harjoto

M. A.

(2011). Corporate governance and CSR nexus. Journal of Business Ethics, 100, 45-67.

44.

Hausman

J. A.

(1978). Specification tests in econometrics. Econometrica, 466, 1251-1271.

45.

Hillman

A. J.

Keim

G. D.

(2001). Shareholder value, stakeholder management, and social issues: What’s the bottom line? Strategic Management Journal, 22, 125-139.

46.

Jian

Lee

K. W.

(2015). CEO compensation and corporate social responsibility. Journal of Multinational Financial Management, 29, 46-65.

47.

Harjoto

M. A.

(2011). Corporate governance and firm value: The impact of corporate social responsibility. Journal of Business Ethics, 103, 351-383.

48.

Johnson

R. A.

Greening

D. W.

(1999). The effects of corporate governance and institutional ownership types on corporate social performance. Academy of Management Journal, 42, 564-576.

49.

Johnston

(2005). An investigation of regulatory and voluntary capital expenditures. Journal of Accounting and Public Policy, 24, 175-206.

50.

Karabarbounis

Macnamara

McCord

(2014). A business cycle analysis of debt and equity financing. Economic Quarterly, 100(1), 51-85.

51.

Karpoff

J. M.

Lott

J. R.

Jr. Wehrly

E. W.

(2005). The reputational penalties for environmental violations: Empirical evidence. Journal of Law and Economics, 48, 653-675.

52.

Kempf

Osthoff

(2007). The effect of socially responsible investing on portfolio performance. European Financial Management, 13, 908-922.

53.

Kim

J. B.

Wang

(2014, January). Labor unemployment risk and sticky cost behavior (Working paper). City University of Hong Kong, Retrieved from http://www.af.polyu.edu.hk/files/jiar2014/cc017%20full%20paper-2_final.pdf

54.

Kim

(2014). Corporate social responsibility and stock price crash risk. Journal of Banking & Finance, 43, 1-13.

55.

Kim

Park

M. S.

Wier

(2012). Is earnings quality associated with corporate social responsibility? The Accounting Review, 87, 761-796.

56.

Kitzmueller

Shimshack

(2012). Economic perspectives on corporate social responsibility. Journal of Economic Literature, 50, 51-84.

57.

Kondra

A. Z.

Hinings

C. R.

(1998). Organizational diversity and change in institutional theory. Organization Studies, 19, 743-767.

58.

Mahlendorf

M. D.

(2009). Sticky cost issues—Kostenremanenz bei Nachfrageschwankungen. Controlling & Management, 53, 193-195.

59.

Margolis

J. D.

Walsh

J. P.

(2003). Misery loves company: Rethinking social initiatives by business. Administrative Science Quarterly, 48, 268-305.

60.

Mattingly

J. E.

(2017). Corporate social performance: A review of empirical research examining the corporation–society relationship using Kinder, Lydenberg, Domini Social Ratings data. Business & Society, 56, 796-839.

61.

Mattingly

J. E.

Berman

S. L.

(2006). Measurement of corporate social action discovering taxonomy in the Kinder Lydenburg Domini ratings data. Business & Society, 45, 20-46.

62.

McGuire

Dow

Argheyd

(2003). CEO incentives and corporate social performance. Journal of Business Ethics, 45, 341-359.

63.

McWilliams

Siegel

(2001). Corporate social responsibility: A theory of the firm perspective. Academy of Management Review, 26, 117-127.

64.

Moskowitz

(1972). Choosing socially responsible stocks. Business and Society Review, 1, 71-75.

65.

Muoghalu

M. I.

Robison

H. D.

Glascock

J. L.

(1990). Hazardous waste lawsuits, stockholder returns, and deterrence. Southern Economic Journal, 57, 357-370.

66.

Noreen

(1991). Conditions under which activity-based cost systems provide relevant costs. Journal of Management Accounting Research, 3, 159-168.

67.

Orlitzky

Schmidt

F. L.

Rynes

S. L.

(2003). Corporate social and financial performance: A meta-analysis. Organization Studies, 24(3), 403-441.

68.

Petersen

M. A.

(2009). Estimating standard errors in finance panel data sets: Comparing approaches. Review of Financial Studies, 22, 435-480.

69.

Preston

L. E.

O’Bannon

D. P.

(1997). The corporate social-financial performance relationship. Business & Society, 36, 419-429.

70.

Reverte

(2009). Determinants of corporate social responsibility disclosure ratings by Spanish listed firms. Journal of Business Ethics, 88, 351-366.

71.

Rowley

Berman

(2000). A brand new brand of corporate social performance. Business & Society, 39, 397-418.

72.

Salancik

G. R.

Pfeffer

(1978). Who gets power—And how they hold on to it: A strategic-contingency model of power. Organizational Dynamics, 5(3), 3-21.

73.

Schreck

(2011). Reviewing the business case for corporate social responsibility: New evidence and analysis. Journal of Business Ethics, 103, 167-188.

74.

Servaes

Tamayo

(2013). The impact of corporate social responsibility on firm value: The role of customer awareness. Management Science, 59, 1045-1061.

75.

Stevens

J. M.

Kevin

S. H.

Harrison

D. A.

Cochran

P. L.

(2005). Symbolic or substantive document? The influence of ethics codes on financial executives’ decisions. Strategic Management Journal, 26, 181-195.

76.

Stock

J. H.

Yogo

(2005). Testing for weak instruments in linear IV regression. In Andrews DWK Identification and inference for econometric models: Essays in honor of Thomas Rothenberg (pp. 80-108). New York, NY: Cambridge University Press.

77.

Strugatch

(2011). Turning values into valuation: Can corporate social responsibility survive hard times and emerge intact? Journal of Management Development, 30, 44-48.

78.

Surroca

Tribó

J. A.

Waddock

S. A.

(2010). Corporate responsibility and financial performance: The role of intangible resources. Strategic Management Journal, 31, 463-490.

79.

Turban

D. B.

Greening

D. W.

(1997). Corporate social performance and organizational attractiveness to prospective employees. Academy of Management Journal, 40, 658-672.

80.

Venieris

Naoum

V. C.

Vlismas

(2015). Organisation capital and sticky behaviour of selling, general and administrative expenses. Management Accounting Research, 26, 54-82.

81.

Waddock

S. A.

Graves

S. B.

(1997). The corporate social performance-financial performance link. Strategic Management Journal, 18, 303-319.

82.

Waldman

D. A.

Siegel

D. S.

Javidan

(2006). Components of CEO transformational leadership and corporate social responsibility. Journal of Management Studies, 43, 1703-1725.

83.

Wang

Dou

Jia

(2016). A meta-analytic review of corporate social responsibility and corporate financial performance: The moderating effect of contextual factors. Business & Society, 55, 1083-1121.

84.

Wolf

S. M.

(1996). Fear and loathing about the public right to know: The surprising success of the Emergency Planning and Community Right-to-Know Act. Journal of Land Use & Environmental Law, 11, 217-324.

85.

Wood

D. J.

Jones

R. E.

(1995). Stakeholder mismatching: A theoretical problem in empirical research on corporate social performance. International Journal of Organizational Analysis, 3, 229-267.

86.

Zhang

Zhu

Yue

Zhu

(2010). Corporate philanthropic giving, advertising intensity, and industry competition level. Journal of Business Ethics, 94, 39-52.