1
Staying on the radar: a behavioural perspective on how media
coverage influences the relationship between social and financial
performance
Jelmer Enhus 11146486
Final version – submitted 24-03-2017
MSc. Business Administration – Strategy Track University of Amsterdam
2
Statement of originality
This document is written by Student Jelmer Enhus who declares to take full responsibility for the contents of this document. I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it. The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
3 Table of contents
Abstract 4
Introduction 5
Literature Review 10
Social Performance and CSR-activities 10
Financial Performance 11
The general relationship between social and financial performance 12
Media Coverage 13
A behavioural perspective 15
The influence of CSR-activities: a consumer perspective 15 The influence of CSR-activities: a firm perspective 17
The influence of CSR-activities: media coverage 20
The influence of a firm’s CSR strategy 21
Method 22
Research design 22
Dataset 23
Sample 23
Deconstruction of components 23
Variables and models 24
Exploratory Analyses 26
Corporate Social Responsibility activities 26
Financial Performance 27 Media coverage 27 Control Variables 28 Results 29 Exploratory Analyses 34 Discussion 39
Limitations & directions for future research 43
Conclusion 45
References 48
Appendix 59
Appendix A: KLD-Strengths as from 2003 59
Abstract
How consumers react to a firm’s engagement in Corporate Social Responsibility (CSR) activities could be one of the leading causes of an increase in financial performance. However, the engagement in CSR-activities alone seems to be insufficient. This research tests how media coverage on a firm’s CSR could influences the relationship between a firm’s engagement in CSR-activities and its financial performance and how this influence could be explained using psychological and behavioural theories. Social and financial ratings, as well as media reports, were retrieved from several databases to test our hypotheses. Results from 94 firms active in business-to-consumer oriented industries suggest that in order for a firm’s CSR-activities to positively influence financial performance, media coverage on that firm’s CSR has to be substantial, and overwhelmingly positive. Public awareness of a firm’s CSR gives consumers additional motives to purchase that firm’s product, because the purchase or use of these products can spread status-increasing signals to the consumer’s surroundings. It would therefore be in firms’ best interest to receive as much positive media coverage as possible. Unfortunately, when firms increase the amount of CSR-activities it not lead to an increase in positive media coverage on their CSR. It therefore seems firms have little, direct or indirect, influence on their financial performance through the amount of engagement in CSR activities. However, we found evidence that supports arguments from both management and psychology literature, indicating that a CSR strategy focussing on a single category of CSR has more benefits for a firm than a generic CSR strategy.
Introduction
Modern society relies on firms to generate wealth and stimulate economic growth. Whether or not firms succeed in doing both could have a big influence on their appeal for society and its members. ‘Going the extra mile’ as a firm, by engaging in socially responsible activities, may even improve their financial performance (Orltizky, Schmidt, & Rynes, 2003). Considerable research has been done on the relationship between a firm’s social performance (often referred to as CSP) and financial performance (often referred to as CFP) (Griffin & Mahon, 1997; Margolis, Elfenbein, & Walsh, 2009; Orlitzky et al., 2003), which resulted in more positive (Brammer & Milington, 2008; Flammer, 2015; Luo, Wang, Raithel, & Zheng, 2015; Simpson & Kohers, 2002), than negative results (Griffin & Mahon, 1997). More recent research has not only confirmed initial findings about the existence of the (generally positive) relationship between social and financial performance, these studies made attempts at understanding and defining the numerous mechanisms that could explain the relationship (Andersen & Dejoy, 2011; Bhattacharya & Sen, 2004; Chih & Chih, 2014; Perrini, Russo, Tencati, & Vurro, 2011; Short, McKenny, Ketchen, Snow, & Hult, 2016).
These recent studies have revealed that variation in the business environment (often used to control the relationship) (Andersen & Dejoy, 2011; Short et al., 2016), the attention to a firm’s social performance by the media (Chi & Chi, 2014), stakeholder orientation (Berman, Wicks, Kotha, & Jones, 1999), changes in Corporate Social Responsibility (CSR) strategy (Ruf, Muralidhar, Brown, Janney, & Paul, 2001), and stakeholder’s reaction to Corporate Social Responsibility activities (Jones, Willness, & Madey, 2014; Sen & Bhattacharya, 2001) all influence the relationship between a firm’s social and financial performance. In the case of reactions to CSR-activities, a firm’s social performance could influence the behaviour of different stakeholder groups in different ways (Berman et al., 1999; Jones et al., 2014; Sen & Bhattacharya, 2001). For example: CSR-activities could both, depending on the situation,
6
increase purchase intention of consumers (Mohr & Webb, 2005), and/or increase firm attractiveness to prospective employees (Greening & Turban, 2000; Turban & Greening, 1997). These studies argue how and why several behavioural mechanisms influence the way stakeholder groups respond to CSR-activities (Bhattacharya & Sen, 2004; Jones et al., 2014 Turban & Greening, 1997). However, these studies pay little attention to how difference in information source (i.e., from the firm itself vs. from an independent source) and information intensity (e.g., how much attention a firm’s CSR receives in the media) could influence stakeholders’ reaction towards CSR-activities. Studies that do pay attention to information source suggest the amount of media coverage on a firm’s CSR-activities influences the relationship between social and financial performance (Chi & Chi, 2014; Deephouse, 2000; McWilliams & Siegel, 2001; Schuler & Cording, 2006). The studies focussing on the role of media coverage, in turn, do not consider the behavioural and psychological mechanisms behind the influence of media coverage. We propose that understanding the behavioural and psychological mechanisms behind the influence of media coverage on the relationship between social and financial performance could give important insights on the role stakeholders play in this relationship.
One particular group of stakeholders that appears to have both a large influence on a firm’s financial performance (Armstrong, Morwitz, & Kumar, 2000; Mohr & Webb, 2005; Morwitz, Steckel, & Gupta, 2007; Orlitzky, Louche, Gond, & Chapple, 2015) and is particularly susceptible to a firm’s CSR-activities, are consumers (Bayattacharya & Sen, 2004; Marin, Ruiz, & Rubio, 2009). Research suggests that a positive relationship exists between a firm’s CSR-activities and consumers’ reactions to that firm and its products (Creyer & Ross, 1997; Mohr & Webb, 2005). Consumers tend to base their purchase decision not only on the utility of a product, but also on alternative and additional factors (Griskevicius, Tybur, & Van den Bergh, 2010; Millet & Dewitte, 2007; Mukherjee & Hoyer,
7 2001; Nelissen & Meijers, 2011), provided that these factors are apparent and salient to both consumers and their surroundings (Griskevicius et al., 2010; Marom, 2006; Nelissen & Meijers, 2011; Schuler & Cording, 2006).
Both media coverage (Chi & Chi, 2014; Deephouse, 2000; McWilliams & Siegel, 2001) and consumer (purchase) behaviour (Bhattacharya & Sen, 2004; Mohr & Webb, 2005; Sen & Battacharya, 2001) appear to influence the relationship between social and financial performance. A better understanding of the interplay between these two factors could possibly help understand, and even partially predict, the relationship between a firm’s CSR-activities and its financial performance. We aim to give a better understanding of how media coverage on a firm’s CSR-activities influences the relationship between social and financial performance by using a behavioural approach. This aim leads to our research questions: (1) How does the portrayal of a firm’s CSR by the media influence the relationship between a firm’s CSR-activities and financial performance, and (2) how could this influence be explained using psychological theories of (consumer) behaviour?
To be able to answer our research question, we first need to take a closer look at the relationship between social and financial performance in general. Although more evidence supports than rejects the relationship between social and financial performance, several studies argue that the relationship has to be seen as contingent on a variety of contextual factors. Contextual factors include firm size, R&D ratio, risk, and most importantly industry environment (Andersen & Dejoy, 2011; Brammer & Millington, 2008; Margolis et al., 2009; McWilliams & Siegel, 2000; Short et al., 2016). The contingency on industry environment indicates the mechanisms that drive the relationship between a firm’s social and financial performance are most likely different across industries (Short, et al., 2016). As we aim to use a behavioural approach and theories of consumer behaviour to further understand and explain the relationship between CSR-activities and financial performance, this research focuses on
8
business-to-consumer oriented industries. Of all business-to-consumer oriented industries, the industries that both service directly to the consumers, and provide products and/or services that consumers find ‘non-essential’ will most likely offer the most insights to answer our research question. When consumers find products or services non-essential, their decision to purchase these products or services are subject to more elaborate decision-making processes and psychological mechanisms (Bagozzi, Gürhan-Canli, & Priester, 2002; Hoyer & MacInnis, 2008; Solomon, Bamossy, Askegaard, & Hogg, 2006). This focus gives us an environment in which we can redress the oversight of the influences information source (e.g., information provided by the firm itself vs. by independent media) and information intensity have on consumers’ reaction towards a firm’s CSR-activities.
We redress this oversight by examining the influence media coverage on a firm’s CSR has on the relationship between the firm’s CSR-activities and financial performance. We argue that a firm’s CSR-activities could become an additional motive for consumers to purchase that firm’s product or service, but only if the CSR-activities of the firm are apparent and salient to both consumers and their surroundings. These additional motives could be a reason for consumers to prefer products provided by firms that engage in CSR-activities, because the purchase or use of these products could send out signals that evoke an increase in (perceived) status among the consumer’s peers. Hence, the CSR-activities of a firm do not only have to be salient to the consumers, but also to their surroundings. Media coverage could be one of the leading factors that could make a firm’s CSR-activities salient to the public, which would explain its influence on the relationship between a firm’s CSR-activities and financial performance. Additionally, arguments from both management and consumer psychology propose that engagement in a single category, or similar categories of CSR-activities has a more positive influence on financial performance than a more generic engagement. Our findings suggest the influence of media coverage on a firm’s CSR is even
9 more important than initially suspected: in a sample of 94 firms active in business-to-consumer oriented industries, only when the media coverage of a firm’s CSR was overwhelmingly positive, a positive relationship between firm’s CSR-activities and financial performance was found. Additionally, we found that when looking at CSR-activities in individual categories, only environmental CSR-activities seem to have a positive relationship with financial performance. But again, only in firms the media had reported overwhelmingly positive on.
This research contributes to the CSR literature in two ways. First, it gives insight on why firms rely on independent sources of information to communicate about their CSR. Promoting their own CSR-activities could backfire for firms due to scepticism from stakeholders. The influence a firm’s CSR-activities has on its financial performance is contingent on the way independent sources of information such as the media report on and communicate about the firm’s CSR. As several studies already proposed how media coverage could play an important role in the relationship between social and financial performance, this research adds to these insights by explaining this role using psychological theories of consumer behaviour. Second, this research proposes how firms can potentially increase their financial performance by engaging in CSR-activities more strategically. We have found evidence that supports arguments from both management and psychology literature, suggesting that firms could benefit from a CSR strategy focussing on a single category of CSR, over a generic CSR strategy. Environmental CSR-activities alone had a positive influence on a firm’s financial performance, while CSR-activities on four other categories did not influence financial performance.
10
Literature Review
Social Performance and CSR-activities
Corporate Social Performance (CSP) has been present in the management literature for over 40 years. The CSP domain has always been controversial, fluid, ambiguous and difficult to research. As first described by Carroll (1979), CSP is seen as a three-dimensional construct of corporate social responsibility, corporate social responsiveness, and social issues. Wood (1991) builds upon the definitions by Carroll (1979) and Wartick and Cochran (1985) and defines CSP as: “a business organization’s configuration of principles of social responsibility, processes of social responsiveness, and policies, programs, and observable outcomes as they relate to the firm’s societal relationships” (Wood, 1991, p.693). Wood’s CSP-model contains three components: (1) principles of corporate social responsibility, (2) processes of corporate social responsiveness, and (3) outcomes of corporate behaviour (Wood, 1991).
Of the three components incorporated in Wood’s (1991) CSP-model, outcomes of corporate behaviour is the only component that is actually observable and open to assessment (Wood, 1991). As social activities are seen as the antecedent of social performance (Waddock & Graves, 1997), and outcomes of corporate behaviour is the only component that is observable and open to assessment (Wood, 1991), we argue a firm’s CSR activities are the main source of positive influence when it comes to consumers’ perception of a firm’s social performance. We combine the definitions used in the studies by Marquis, Glynn, and Davis (2007), and Schuler and Cording (2006) to define a CSR-activity as ‘a positive, voluntary, socially oriented business action, which is not directly made to make profit and produces social or third-party effects’. This definition makes researching the relationship between CSR-activities and consumer reactions and behaviour possible (Schuler & Cording, 2006), and refers to activities that extend beyond immediate profit maximization (Marquis, Glynn, & Davis, 2007). Also, it has similarities with the third component of Wood’s (1991) CSP-model,
11 which is arguably the only component of the CSP-model where any real performance exists (Wood, 1991; 2010).
Financial Performance
Financial Performance is a more straightforward construct than CSP. Financial performance can be more easily quantified than social performance, and is therefore more easily measured. The challenge in measuring financial performance lies with the definition of what financial performance actually means. Corporate performance itself, whichever area, depends on what goals firms set for themselves. What financial performance is for a firm that wants to increase its market share, or wants to penetrate other markets is different than for a firm that only wants to increase its profitability. Therefore, the selection of financial criteria to measure financial performance depends on what kind of (financial) ‘goal’ benchmarks performance. The variety of financial measurements used in CSR literature is confirmed by a meta-analysis by Griffin & Mahon (1997).
In our introduction we mention that our research will focus on the influence media coverage has on the relationship between social and financial performance and what role consumer (purchase) behaviour has in these relationships. As consumer behaviour has a large influence on a firm’s revenue (Armstrong et al., 2000; Morwitz et al., 2007; Solomon et al., 2006), we will use the accounting-based measurement net income as a measure for financial performance for two reasons. Firstly, net income is very closely related to a firm’s revenue, which is gives us a good representation of the influence consumer behaviour has (Schuler & Cording, 2006; Solomon et al., 2006), and secondly, net income has been used as a measurement of profitability (Griffin & Mahon, 1997), which has often been linked to the more general definition of financial performance (Griffin & Mahon, 1997; Margolis et al., 2009).
12
The general relationship between social and financial performance
For decades, the relationship of social and financial performance has been under scrutiny. This relationship was first established as a reciprocal relationship: financial performance was needed to be able to invest money into social activities, which resulted in social performance, and social performance was positively related to future financial performance (Waddock & Graves, 1997). Although a meta-analysis (Griffin & Mahon, 1997) from the same period could not find any significant relationship, they did find indicators that stand positively towards a relationship between social and financial performance (Griffin & Mahon, 1997). Their main implication for research was to attempt to standardize the measurement of social and financial performance, which is in line with the difficulties experienced during previous research on social performance (Rowley & Berman, 2000; Wood, 2010).
One of the most extensive analyses on the relationship of social and financial performance (Orlitzky et al., 2003) supports the bidirectional relationship Waddock and Graves (1997) proposed. Additionally, the general relationship of social and financial performance across a wide variety of industries was found (Orlitzky et al., 2003). Besides this general relationship, recent studies also support the relationship between social and financial performance in more specific situations (Chi, Chi, & Chou, 2010; Margolis et al., 2009; Orlitzky et al., 2015). These studies show that CSR-activities could positively influence the perception of stakeholders: it could build a positive image with customers, investors, bankers, suppliers, and prospective employees (Fombrun & Shanley, 1990; Turban & Greening, 1997). As CSR-activities improve stakeholders’ perception of the firm, they contribute to part of a firm’s total ‘good reputation’. Reputation is often used to explain why a firm’s CSR could have a positive influence in its financial performance (Fombrun & Shanley, 1990; Roberts & Downling, 2002). However, as reputation is generally defined as “a perceptual representation
13 of a firm’s past actions and future prospects that describe the firm’s overall appeal to all its key stakeholders when compared to other leading rivals” (Fombrun, 1996, p.72), CSR is only a small part of a firm’s reputation. Nevertheless, it seems the positive image CSR-activities evoke in external stakeholders could influence the willingness to cooperate and increase buying behaviour, which could improve financial performance (Mohr & Webb, 2005; Orlitzky et al., 2003). The collection of the unilateral results supporting the relationship between a firm’s social and financial performance leads to our first hypothesis:
Hypothesis 1: A firm’s engagement in CSR-activities is positively related to the firm’s Net Income.
Media Coverage
In a perfect market, where both firms and the media fully discloses relevant information and consumers react completely rational to the relevant information, media coverage should not have any impact on factors such as consumer behaviour, consumers’ perception of firms, or firm’s profitability. (Rumelt, 1997). However, media reports are neither fully objective, nor are consumers fully rational (Chih et al., 2010; Hamilton & Tilman, 1983; Lerner & Keltner, 2000; Rucker & Galinsky, 2009). Because consumers are aware firms do not fully disclose all information, media coverage can shape the perception consumers have of a firm by aggregating, certifying, and communicating relevant information to the public (Chih et al., 2010).
As media coverage could influence consumers’ perception of firms, this influence could be positive or negative, depending on the valence of the reports. As also described by Chih et al. (2010), readers in general are influenced by the valence of the news. Readers (unintentionally) rate the reports as favourable or unfavourable depending on the ‘tone’ of the report. Additionally, readers evaluate organizations, people, and events based on the favourableness of the news topics about them (Chi et al., 2010; Dean, 2003). This evaluation
14
process suggests that it is important for firms to appear favourable in the news, as this influences its stakeholders’ perception.
To assess media coverage on a firm’s CSR, we first need to define what we see as media coverage. To make an estimation of the general media coverage on a firm, we take what is argued to be the most influential and indicative media source: newspapers (Gaines-Ross, 2000; Robinson & Levy, 1996) As for media coverage on a firm’s CSR, there are two main factors that play a role in how the media could change consumers’ perception of firms: (1) the amount of news articles on a firm’s CSR, and (2) the valence of the news articles (positive, negative, or neutral tone) (Chi et al., 2010; Dean, 2003). Assuming that the media has the autonomy to choose what and how they cover their stories, firms can only partially control the two factors. However, following agenda-building theories, a well thought out CSR-strategy could positively influence the media coverage on a firm’s CSR-activities (Curtin, 1999; Kiousis, Popescu, & Mitrook, 2007). These theories tell us that by strategically choosing to engage/invest in CSR-activities, and how often to do so, a firm could potentially control part of the amount of media coverage their CSR receives, and if these reports are positive or not. As media coverage influences stakeholders’ perception of the firm, it would be valuable to know whether or not, and how CSR-activities could possibly influence the amount of positive media coverage of a firm’s CSR. We propose that by strategically investing and managing the engagement in CSR-activities, firms are able to (partially) influence how the media portrays its CSR:
Hypothesis 2: The firm’s engagement in CSR-activities is positively related to the amount of positive media coverage of the firm’s CSR.
15 A behavioural perspective
The behaviour of one particular stakeholder group appears particularly susceptible to a firm’s CSR-activities; 84% of American consumers say that they would be likely to switch brands to one associated with a good cause, if price and quality are similar (Bhattacharya & Sen, 2004). Additionally, consumers have one of the strongest influences on a firm’s revenue (Armstrong et al., 2000; Morwitz et al., 2007; Solomon, et al., 2006), which in turn appears to be one of the leading causes of a rise of financial performance after an increase in social performance (Ruf et al., 2001). These aspects make consumer behaviour a very interesting research subject when examining the relationship between social and financial performance, and is receiving more attention in the last decade.
Traditional literature of the strategy field would propose consumer behaviour could not be influenced by other factors than the utility of a product, due to the assumption of rationality (Barney, 2001; Simon, 1955). However, more recent studies incorporated more psychological and behavioural perspectives to the strategy, as well as the CSR literature (Schuler & Cording, 2006; Sen & Bhattacharya, 2001). Research by Marin, Ruiz, and Rubio (2009) proposes a similar effect as our research: salience of a firm’s CSR causes consumers to identify more strongly with the firm, resulting in a higher loyalty to the firm and/or brand. This research supports the more general psychological research that suggests consumers, beside utility, have additional motives to purchase a product (Griskevicius et al., 2010; Hoyer & MacInnis, 2008; Solomon et al., 2006).
The influence of CSR-activities: a consumer perspective
Consumers could have different additional motives when purchasing products. One of these additional motives is to spread what is called a ‘costly signal’ (Grafen, 1990; Millet & Dewitte, 2007; Mukherjee & Hoyer, 2001). Spreading a costly signal originates from the evolutionary biology. In nature, males need to find ways to show, or ‘signal’, their health and
16
strength to potential mates to have a better chance to reproduce. The potential mates, however, may have difficulty assessing the health and strength of males without investing precious time during their fertile period. Males therefore display ‘costly signals’: they show they are able to survive despite having a physical ‘handicap’. Costly signals in nature can vary from the huge tail of a male peacock to the giant antlers of a buck. Both animals show potential mates that despite the disadvantage these features could bring, they are still as strong and healthy as others who do not have the same amount of disadvantage (Grafen, 1990).
As animals spread costly signals with physical handicaps, modern consumers could spread costly signals when they purchase or use certain products (Nelissen & Meijers, 2011). A good example of spreading costly signals in modern consumerism derives from the additional motives of consumers to purchase ‘green’ products. Consumers often pay a price-premium for green products, not only because they are good for the environment, but also to signal that they are wealthy enough to pay a price-premium and still be able to provide all other necessities of themselves and/or their families (Griskevicius et al., 2010; Nelissen & Meijers, 2011). By spreading costly signals consumers could increase their status among their own social circle, but also their perceived status to unknown individuals (Griskevicius et al., 2010; Hardy & Van Vugt, 2006; Millet & Dewitte, 2007; Nelissen & Meijers, 2011).
Similar to the costly signals animals spread in nature, costly signals in modern consumerism need to be both observable and recognized by the sender’s surroundings, before they can be beneficial to the sender. The physical handicaps animals use are evolved for exactly that purpose: to be observable and recognized by the animal’s surroundings (Grafen, 1990). The characteristics of products on which consumers base whether or not the purchase or use of the product could spread costly signals (i.e., the additional characteristics of a product), may not always be that apparent. The purchase or use of a product only influences a consumer’s status if: (1) the additional characteristic(s) in the product are generally
17 recognized by others (e.g., it is generally known that a Tesla is an electric car, which is better for the environment than regular cars), and purchasing and/or using the product is visible to others (e.g., driving a Tesla is visible to others, but owning an eco-friendly mattress or regularly giving blood is not).
The likelihood that a consumer will consider a firm’s CSR-activities as an additional motive to purchase a product or service will therefore greatly depend on the public awareness of a firm’s CSR. It will be essential for consumers that (1) they see and recognize a firm’s CSR as an additional characteristic to the product or service, and (2) the public recognizes and understands what these additional characteristics mean in terms of spreading a costly signal (Griskevicius et al., 2010; Nelissen & Meijers, 2011; Schuler & Cording, 2006).
The influence of CSR-activities: a firm perspective
In all markets that serve the public, consumers are a key external stakeholder group when it comes to a firm’s financial performance. Consumers’ evaluation of a firm’s CSR could influence a firm’s financial performance, but only when the firm manages to (1) make information about its CSR-activities available to the consumers, and (2) make sure consumers are aware and able to comprehend this information (Schuler & Coding, 2006). Several studies suggest that CSR awareness among consumers is created through two communication channels: Corporate (i.e., CSR Reports, Corporate Website, PR, Advertising) and/or Independent (i.e., Media Coverage, Word-of-Mouth) (Du, Bhattacharya, & Sen, 2010; Fombrun & Shanley, 1990; McWilliams & Siegel, 2001). Spreading awareness about their engagement in CSR-activities through one, or both communication channels could therefore be crucial for firms to increase their financial performance.
The importance of CSR awareness could, as was the case with the origin of costly signals, be explained by evolutionary biology. Because not all ‘advantages of reproducing’ in nature are as observable as the physical handicaps of bucks and peacocks, animals have
18
developed systems to make their hidden advantages observable. Animals use signals to communicate their unobservable evolutionary advantages to potential mates (Kaplan & Gangestad, 2005; Scott-Phillips, 2008), which is similar to how firms engage in reputation building activities to signal their positive features (Fombrun, 1996; Gray & Balmer, 1998).
A good example of the similarities of signalling systems in nature and in business is how animals signal parasite resistance. Parasite resistance is, as is often the case with firm features such as CSR, sustainable business behaviour, or good overall management, an unobservable positive feature. These males ‘signal’ females of their high parasite resistance by giving off certain odours. If males could not use signals to make females aware of their evolutionary advantage over other males, the efforts these males have gone through to develop this resistance would have been lost (Kaplan & Gangestad, 2005).
Firms use similar mechanisms to signal the public of their unobservable positive features. When firms want to make their positive features more apparent to the public, they start using corporate communication in order to incorporate positive features into their reputation (Du et al., 2010; Gray & Balmer, 1998). In other words, firms allocate resources to activities that would build a reputation that ‘signals’ positive features to the public (Cable & Turban, 2003; Fombrun, 1996; Turban & Cable, 2003). For example, firms invest in acquiring certifications to signal sustainable behaviour to consumers (Connelly, Ketchen & Slater, 2011), or optimize prospective employees’ recruitment experience to signal good management in the rest of the firm (Rynes, Bretz, & Gerhart, 1991).
Comparing signalling systems in nature to modern consumerism and CSR, one can argue that firms represent the males that want to signal their positive features (i.e., engagement in CSR-activities) to the females (i.e., the consumers) to benefit from the reactions to their positive features. If firms do not find ways to effectively signal consumers about their CSR-activities, the resources invested in CSR-activities will most likely not have a
19 positive effect on their financial performance (Griskevicius et al., 2010; Hardy & Van Vugt, 2006; Millet & Dewitte, 2007; Nelissen & Meijers, 2011; Schuler & Cording, 2006; Scott-Phillips, 2008)
In general, firms try to build a good reputation by intentionally communicating about and allocating resource to activities that could signal their positive features to the public (Fombrun, 1996; Gray & Balmer, 1998). CSR is only a part of a firm’s reputation, but research suggests public awareness of CSR alone could bring large benefits for firms (McWilliams & Siegel, 2001; Martin et al., 2009; Schuler & Cording, 2006). However, providing information about CSR themselves could be a slippery slope: the notion that spreading awareness (i.e., signalling) of additional positive features results in a rise in interest based on the concept of ‘honest signalling’. Honest signalling in nature builds up a system where organisms ‘trust’ the signals of others: the female trusts that the odour spread by the male is a result of his higher resistance to parasites and is not a result of something else (Kaplan & Gangestad, 2005). Dishonest signalling, however, disrupts these systems and creates scepticism and distrust. If, for example, a signalling system of warning calls is present in a colony of seabirds, this system could be abused by these birds as well: a bird could send out a warning call to scare away competitor birds in his hunting area. When signals are abused more often, the system could collapse due to mistrust in the warning call (Dawkins & Krebs, 1978; Scott-Phillips, 2008).
This mistrust in the sincerity of the ‘signal’ is also a big risk for firms, especially when it comes to CSR. CSR-activities differentiate themselves from other reputation building activities by being characterized as ‘not made to directly make profit’. When firms promote their own CSR-activities, like they often do with other reputation building activities (Gray & Balmer, 1998), consumers are likely to challenge the sincerity of the activities because it could infringe upon the altruistic nature of the activities (Fein & Hilton, 1994). Consumers
20
could suspect a firm abuses the signalling system and has some ulterior motive to perform CSR-activities, which may even result in negative evaluations (Dawkins & Krebs, 1978; Scott-Phillips, 2008; Yoon, Gürhan-Canli, & Schwarz, 2006). Conversely, if information on a firm’s CSR-activities is provided by independent sources, such as the media, the likelihood average consumers are aware of a firm’s CSR increases (Fombrun, Gardberg, & Barnett, 2000; Schuler & Cording, 2006). The influential differences between internal and independent sources of information suggest that in the case of CSR-activities, firms’ signalling mechanisms (i.e., corporate communication channels) may not have the desired effect on stakeholders, like they have with other reputation building activities. Firms therefore need to carefully consider how the difference between corporate and independent communication could, more so than with other reputation building activities, alter the influence CSR-activities have on consumers.
The influence of CSR-activities: media coverage
As argued in the previous section, communication about CSR-activities cannot be seen identical to communication about other reputation building activities. As CSR-activities are altruistic in nature, consumers are more likely to be sceptic and suspect ulterior motives when firms promote or communicate about their CSR (Fein & Hilton, 1994; Yoon et al., 2006). Additionally, if a firm’s CSR-activities are provided by an independent source, such as the media, the CSR awareness among consumers is more likely to increase (Fombrun et al., 2000) and the information is perceived as more credible by the average consumer (Du et al., 2010). Consumers perceive independent information sources as essential when they evaluate a firm’s CSR and determine if its CSR can be seen as an additional characteristic to its products and/or services. The roles public awareness of a firm’s CSR, and information source play in the way consumers react to a firm’s CSR lead to our third hypothesis:
21 Hypothesis 3: The Net CSR Media Coverage on a firm’s CSR-activities positively moderates the effect of a firm’s CSR-activities on Net Income.
The influence of a firm’s CSR strategy
Firms have two choices to make when building a strategy concerning CSR: (1) the amount of CSR-activities they will engage in, and (2) what categories of CSR-activities they will engage in (e.g., CSR-activities focussed on environment, human rights, or employee relations). Tang, Hull, and Rothenberg (2012) argue that firms should try to engage in CSR-activities that fall in the same CSR category because related CSR categories have related requirements of resources. The more CSR categories a firm engages in simultaneously, the more different capacities and resources are required from a firm. As a result, the more difficult it is for a firm engaging in less-related or even unrelated CSR categories to improve its financial performance (Marcus & Anderson, 2006). For example; CSR-activities in the category environment needs complementary skills in innovating and implementing new processes (Christmann, 2000), and CSR-activities in the human rights category need complementary skills such as lobbying experience with international organizations (Windsor, 2006).
The relatedness of resources poses a strong argument for engaging in similar CSR-activities from a management perspective. From a consumer perspective, arguments can also be made in favour of focussing on similar CSR-activities. When consumers evaluate and compare products, the link between the product’s characteristics and their own social identity plays an important role (Zukin & Maguire, 2004). Consumers tend to purchase goods that will either confirm their social identity for themselves (internal identity), or confirm their social identity to others (collective identity) (Jenkins, 2008; Lamont & Molnár, 2001). Collective identities are formed on multiple levels. A collective identity can be as broad as one’s ethnicity or sexual preference, but can also be as specific as a fan base of a specific athlete or
22
users of a certain brand (He, Li, & Harris, 2012; Jenkins, 2008; Lamont & Molnár, 2001; Taylor, Whittier, & Morris, 1992). When consumers take their social identity as one of the factors in their evaluation and comparison of products, the link between their identity and the characteristics of a product is essential: it will have to be clear and generally known that the purchased product reflects and confirms their social identity (Lamont & Molnár, 2001; Taylor et al., 1992).
Findings from both management literature and consumer psychology pose strong arguments to limit a CSR strategy to a single CSR category. Restricting CSR-activities to one category or dimension could even have positive results according to CSP-literature: engaging in limited category of CSR-activities could increase the salience of a firm’s CSR, which is an important condition when influencing consumers (Marin et al., 2009; Schuler & Cording, 2006). It would therefore be interesting for firms to know if engagement in different CSR categories influences financial performance differently, and if so, which category has the most influence on financial performance.
Method
Research design
This research examines the relationship of a firm’s CSR-activities with their financial performance. This relationship is moderated by media coverage on a firm’s CSR, which suggests that the relationship is stronger or only exist when independent sources of information on a firm’s CSR are available. Data about CSR-activities, financial performance, and media coverage is gathered from multiple databases.
23 Dataset
Our dataset consists of 1103 firms that are active in 10 different sectors. All firms reported all necessary financial information to build up our constructs and were all evaluated by the Kinder, Lydenberg, Domini (KLD) firm. Financial information was gathered using the COMPUSTAT database, information about CSR-activities was gathered using the KLD database, and media coverage was gathered using the LexisNexis database. All data retrieved was from the (fiscal) year 2012, as the KLD database had significantly less data about firms’ CSR from 2013 onwards.
Sample
To examine the role of media coverage in the relationship between CSR-activities and financial performance, our sample exists of 94 firms that are active in similar industries. All industries are classified using the Global Industry Classification Standard (GICS) and are part of the ‘Consumer Discretionary’ sector. These firms’ products and/or services are all ‘available for use as needed or desired’, and seen as ‘non-essential’ by consumers, which ensures these firms’ financial performance relies on the behaviour of consumers. Industries such as ‘Automobiles’, ‘Textiles, Apparel & Luxury Goods’, and ‘Leisure Equipment & Products’ are part of the Consumer Discretionary sector.
Deconstruction of components
Even when examining a single industry, numerous different factors and aspects still influence Corporate Social Performance (CSP) and Corporate Financial Performance (CFP). The measurement and evaluation of CSP has been widely up for debate, mostly due to differences in definition (Wartick & Cochran, 1985; Wood, 1991) and measurement methods (Carroll, 1999; Clarkson, 1995; Griffin & Mahon, 1997). A lot of studies struggle with the complete measurement of CSP, due to the generally accepted definition by Wood (1991; 2010), stating that only a small portion of CSP is actually observable. CSP is most often
24
measured by subtracting the ‘negative CSR score’ from the ‘positive CSR score’ (Chatterji, Levine, & Toffel, 2009). This computation results in a ‘net’ score of a firm’s social performance. We propose that deconstructing CSP by viewing the positive and negative aspects separately, we can get a better understanding of what influence these aspects have on consumers. Especially from a psychological perspective positive and negative stimuli cannot be seen as equal in weight. They cannot be carelessly subtracted from each other to create a ‘net’ score (Peeters & Czapinski, 1990; Taylor, 1991). We therefore define that a firm’s activities are a representation of its positive CSR score. We will further elaborate on CSR-activities in the method section.
The construction of CFP has been contingent on how financial performance is defined. CFP has therefore been constructed differently across more than 200 studies (Griffin & Mahon, 1997). As our aim is to examine the role consumer behaviour has on the relationship between CSP and CFP, we want to have a measure for CFP that could give an indication of consumers’ reactions towards CSR-activities. For this reason, and reasons already mentioned in the introduction and literature review, we will use net income as a measurement for CFP.
The use of these simpler measurements, which are influenced by less factors than the original constructs of social and financial performance, allow us to more easily identify the behavioural factors that may play a role in the relationship between CSR-activities and financial performance.
Variables and models
Our first model consists of one dependent variable (net income) and one independent variable (activities). This model creates a baseline of the relationship between CSR-activities and financial performance and to test our first hypothesis about the relationship between CSR-activities and financial performance. Our second model also consists of one dependent variable (positive CSR media coverage) and one independent variable
(CSR- 25 activities). This model is used to test our second hypothesis about the relationship between CSR-activities and positive CSR media coverage. Our third model is similar to our first model, consisting of one dependent variable (Net Income) and one independent variable (CSR-activities). However, one moderator variable (Net CSR Media Coverage) is added. This model is created to test our third hypothesis about the influence of media coverage on the relationship between CSR-activities and financial performance. All three models contain three control variables: firm size, R&D ratio, and risk. Our models are depicted in Figure 1, Figure 2, and Figure 3. CSR-activities Net Income Figure 1. Model 1. Control Variables: Firm Size, R&D Ratio, and Risk H1 CSR-activities Positive CSR Media Coverage Figure 2. Model 2. Control Variables: Firm Size, R&D Ratio, and Risk H2 CSR-activities Net Income Net CSR Media Coverage Figure 3. Model 3. Control Variables: Firm Size, R&D Ratio, and Risk H3
26
Exploratory Analyses
For our exploratory analyses on (1) whether or not CSR-activities on individual categories influence financial performance differently, and (2) if so, CSR-activities on which category or categories influence financial performance the most, we ran regression analyses based on model 1 and model 3. For the exploratory analyses we replaced our independent variable (CSR-activities) with each one of the dimensional variables (e.g., environmental CSR-activities). The seven dimensional variables represent the CSR-activities on the seven corresponding CSR categories.
Corporate Social Responsibility activities
Corporate Social Responsibility activities (CSR-activities) were measured using ratings from the Kinder, Lydenberg, Domini (KLD) Database. The KLD database is described as “the largest multidimensional CSP database available to the public” (Deckop, Merriman, & Gupta, 2006, p. 344) and is widely used in CSP research (Chatterji et al., 2009). KLD is an independent rating service that focuses exclusively on assessment of corporate social performance across a range of dimensions related to stakeholder concerns (Waddock & Graves, 1997). The KLD dimensions are built up of multiple factors, which are scored using binary scores. The factors are seen as either positive or negative. All positive scores are characterized as ‘strengths’ and all negative scores are characterized as ‘concerns’.
To construct the variable CSR-activities, we gathered firm’s strengths scores on seven dimensions (i.e., environment, community, human rights, employee relations, diversity, product, and corporate governance) from the KLD database. We see the way the KLD-database made distinction between CSR dimensions as the way CSR literature uses categories. The total strengths scores on each of the seven dimensions is a representation of the level of a firm’s corporate social responsible action on that dimension (Marquis et al., 2007). To create the variable representing the total level of CSR-activities, we took the sum
27 all strengths and computed ‘CSR-activities’. For the seven dimensional variables (community CSR-activities, corporate governance CSR-activities, diversity CSR-activities, employee relations CSR-activities, environmental CSR-activities, human rights CSR-activities, and product CSR-activities) we used the total strengths of the corresponding dimension. Factors that build up the firm’s strengths in each of the seven dimensions are predominantly action-oriented (see Appendix A: KLD-strengths as from 2003), which is why the total number of firm strengths gives a good representation of a firm’s CSR-activities.
Financial Performance
To construct financial performance, we first need to define what aspect of financial performance we want to examine and explain. As mentioned in our introduction and literature review, this research aims to assess how consumer behaviour could further explain the relationship between social and financial performance. We chose Net Income as the measurement for financial performance for two reasons: (1) net income gives a good representation of consumers’ influence on financial performance through a increase or decrease in revenue (Armstrong et al., 2000; Morwitz et al., 2007; Solomon et al., 2006), and (2) net income has been as a measure of profitability, which is often linked to financial performance (Griffin & Mahon, 1997; Margolis et al., 2009).
Media coverage
To construct the variables concerning media coverage, we retrieved news articles from the LexisNexis database. According to research by Robinson and Levy (1996) and Gaines-Ross (2000), newspapers provide the most powerful media coverage. Therefore, we have selected ‘newspapers’ as our media source for each of our firms. LexisNexis provides each news article with ‘relevance scores’ to measure the quality of the match between an article and a firm. This relevance score is based on criteria such as the keywords’ frequency and its weight and location within the article. We have only used the news articles with a relevance
28
score of 90% and above, which LexisNexis describes as “Major Reference” (Fang & Peress, 2009).
In our models, we use the variables ‘positive CSR Media Coverage’ and ‘Net CSR Media Coverage’. Positive CSR media coverage refers to the total amount of CSR related articles with a positive tone towards the (CSR-)activity of the firm. Net CSR media coverage refers to the extend to which a firm’s CSR(-activities) have been in the news, and whether or not the articles were positive or negative. To construct the variables we first classified all retrieved news articles (N = 5900) as CSR related (N = 328), or non-CSR related (N = 5572). Subsequently, all CSR related articles were classified as ‘positive’ (N = 123), ‘neutral’ (N = 66), or ‘negative’ (N = 139) coverage (Chi & Chi, 2015). Positive indicates that the article’s content recognizes that the firm does good things for its stakeholders, employees, investors, or the environment; for example, a news article is considered positive if it indicates that the firm has decided to only work with products which are gathered in an environmental sustainable way. A negative news article indicates that the content is likely to be harmful to stakeholders, employees, investors, or the environment. An example of a negative report is the scenario in which (ex-)employees or customers bring a lawsuit to the firm. The news articles classified as neutral reflect neither a positive nor a negative tone towards the firm (Chi & Chi, 2014).
Control Variables
According to several studies (Andersen & Dejoy, 2011; Margolis et al., 2009; Short et al., 2016) it is important to control the relationship between social and financial performance for Firm Size, R&D ratio, and Risk. Data to construct all control variables was gathered using the COMPUSTAT database. For firm size, we used the number of employees in a company (Waddock & Graves, 1997). R&D ratio was constructed by dividing a firm’s R&D expense by its total sales volume. The R&D ratio indicates the priority of R&D for a firm: a higher R&D-Sales ratio indicates a firm has a high priority for innovation due to the larger portion of
29 resources invested in R&D (Margolis et al., 2009; McWilliams & Siegel, 2001). Risk was constructed by dividing a firm’s total debt outstanding by its total assets (Andersen & Dejoy, 2011; McWilliams & Siegel, 2000). To check the robustness of our results, we also constructed a control variable representing the negative aspects of a firm’s CSR. ‘Negative CSR score’ was constructed by summing up all ‘concerns’.
Results Table 1
Descriptive Statistics and Correlation Matrix of main variables
Variable M SD 1 2 3 4 5 6 7 1. CSR-activities 1.24 2.11 2. Net Income (CFP)a 263.76 942.76 .60** 3. Net CSR Media Coverage -0.17 4.386 .20* .27** 4. Positive CSR Media Coverage 1.31 3.55 .54** .50** .69** 5. Negative CSR Score 1.02 1.07 .48** .36** -.07 .24* 6. Firm Sizeb 30.31 65.30 .58** .50** -.29** .23* .40** 7. R&D ratio 0.02 0,07 .16 .01 .14 .13 .10 -.05 8. Risk 0.25 0.27 .14 -.05 .04 .10 .04 .08 -.03
Note: *Correlation is significant at the .05 level (2-tailed) **Correlation is significant at the .01 level (2-tailed)
aNet Income is measured in millions
bFirm Size is measured in thousands
Table 1 shows the descriptive statistics and the correlation matrix of the main variables. The firms’ CSR-activities score and net income were positively correlated. To give a better insight on the relationship between CSR-activities and net income (See Hypothesis 1 and Figure 1) we ran a hierarchical regression analysis with CSR-activities as independent variable, net income as dependent variable, and firm size, R&D ratio, and risk as control variables. In the first step, the control variables firm size, R&D ratio and risk were entered. This model was statistically significant (F(3, 90) = 10.61, p < .001), and explained 24% of the variance in net
30
income. After the entry of the independent variable CSR-activities at Step 2 the total variance explained by the model was 39% (F(4, 89) = 15.78, p < .001). The introduction of CSR-activities explained an additional 15% of the variance in net income, after controlling for firm size, R&D ratio, and risk. In the final model two out of four predictors were statistically significant, with CSR-activities having a higher Beta value than firm size. In other words, if CSR-activities increase by one point, net income increases by 223.5 million dollars, and if firm size increases by one employee, net income increased by 3.16 thousand dollars. Complete regression results are shown in Table 2.
These results seem to support hypothesis 1 about the direct relationship between CSR-activities and net income. However, when checking the main assumptions of regression analysis, results indicate the assumption of homogeneity of variance is violated. Due to heteroscedasticity the statistical power is of our regression model is largely reduced. Although the regression coefficients of the individual predictors under heteroscedasticity are not
Table 2
Regression Results Model 1 DV: Net Income R R2 R2 Change B SE β t p Step 1 .51 .24*** Firm Size 7.37 1.31 .51 5.61*** .000 R&D Ratio 390.77 1205.98 .03 0.32 .747 Risk -325.91 313.07 -.10 -1.04 .301 Step 2 .64 .39*** .15*** Firm Size 3.16 1.46 .22 2.16* .034 R&D Ratio -891.31 1111.44 -.07 -0.80 .425 Risk -497.95 282.44 -.15 .-1.76 .081 CSR-activities 223.50 46.24 .50 4.83*** .000 Step 3a .64 .42 Firm Size 3.16 6.81 - 0.46 .644 R&D Ratio -891.32 6851.97 - -0.13 .897 Risk -497.95 388.97 - -1.28 .204 CSR-activities 223.49 150.85 - 1.48 .142
Note. Statistic Significance: *p < .05, **p < .01, ***p < .001
31 different, the statistical significance of both the entire model and individual predictors cannot be assumed valid (Field, 2009). Unfortunately, increasing the sample size does not decrease the bias; it can actually aggravate the problems caused by heteroscedasticity (Hayes, 1996; Long & Ervin, 2000). When encountering problems of heteroscedasticity, multiple studies advise to use the HC3 estimator to create heteroscedasticity-consistent standard errors when using linear regressions (Hayes & Cai, 2007; Long & Ervin, 2000). Therefore, we used the HC3 estimator to run the regression model depicted in Figure 1 again. As shown in Table 2, no significant regression model was found when using the HC3 estimator to create heteroscedasticity-consistent SE’s. Thus, hypothesis 1 is rejected.
As shown in Table 1, CSR-activities and ‘positive CSR media coverage’ are positively correlated. To further examine the relationship between CSR-activities and positive CSR media coverage (See Hypothesis 2 and Figure 2) we ran a hierarchical regression analysis with CSR-activities as independent variable, positive CSR media coverage as dependent variable, and firm size, R&D ratio, and risk as control variables. In the first step we entered the control variables firm size, R&D ratio, and risk. This model was not statistically significant (F(3, 90) = 2.61, p = .056) and explained 5% of variance in positive CSR media coverage. After the entry of CSR-activities in Step 2 the total variance explained by the model was 24%. This model proved to be significant (F(4, 89) = 8.49, p < .001). The introduction of CSR-activities explained an additional 20% of variance in positive CSR media coverage, after controlling for firm size, R&D ratio, and risk. In the final model, only CSR-activities was a significant predictor for positive CSR media coverage. In other words, if a firm’s CSR-activities increased by one, its positive CSR media coverage increased by .95 articles.
As was the case in our regression analysis of model 1, the assumption of homogeneity of variance was violated in the regression model to test hypothesis 2. After using the HC3
32
estimator in Step 3, no significant regression model was found. Thus, hypothesis 2 is rejected. Complete regression results are shown in Table 3.
To further investigate the mechanisms behind the effect of CSR-activities on net income, we proposed that the relationship is influenced by the Media Coverage on a firm’s CSR (See Hypothesis 3 and Figure 3). To test Hypothesis 3 we ran a regression analysis with CSR-activities as independent variable, net income as dependent variable, net CSR media coverage as moderator, and firm size, R&D Ratio, and risk as control variables. As the direct relationship between CSR-activities and net income proved to have problems of heteroscedasticity, we used the HC3 estimator during the regression analysis. The model in Step 1 was statistically significant (F(7, 86) = 4.38, p < .001) and explained 58% of the total variance of net income. The regression coefficient of the interaction effect of CSR-activities and net CSR media coverage is 19.30 and is statistically different from zero. Thus, the effects of a firm’s CSR-activities on its net income depend on the net CSR media coverage. A closer Table 3
Regression Results Model 2
DV: Positive CSR Media Coverage
R R2 R2 Change B SE β t p Step 1 .28 .05 Firm Size 0.01 0.01 .23 2.26* .026 R&D Ratio 7.38 5.06 .15 1.46 .149 Risk 1.07 1.32 .08 0.81 .418 Step 2 .53 .24*** .20*** Firm Size -0.01 0.01 -.10 -0.89 .375 R&D Ratio 1.93 4.65 .04 0.41 .680 Risk 0.34 1.18 .03 0.29 .776 CSR-activities 0.95 0.19 .57 4.91*** .000 Step 3a .53 .28 Firm Size -0.01 0.01 - -0.38 .704 R&D Ratio 1.92 8.40 - 0.23 .819 Risk 0.34 0.88 - 0.38 .702 CSR-activities 0.95 0.60 - 1.60 .114
Note. Statistic Significance: *p < .05, **p < .01, ***p < .001
33 inspection of the conditional effects indicates that the relationship between CSR-activities and net income is significant only when the net CSR media coverage is high (+1SD), compared to when net CSR media coverage is average or low (-1SD). Complete regression results are shown in Table 4. The conditional effect of CSR-activities on net income is visualized in Figure 4. These results support Hypothesis 3.
Considering both a large portion of CSR-literature and the variable net CSR media coverage contain both positive and negative evaluations, but CSR-activities contains only positive factors, we ran our third model again, but added negative CSR score as an additional control variable. With this analysis, we wanted to check the robustness of our earlier results, considering a large portion of CSR-literature incorporates the negative factors of social performance as well. This model was statistically significant (F(7, 86) = 3.32, p < .01) and explained 59% of the total variance in net income. The regression coefficient of the interaction effect of CSR-activities and net CSR media coverage proved to be slightly more significant compared to the previous regression. The main results for Hypothesis 3 did not change with the addition of the additional control variable. Complete regression results are shown in Table 4. -300 -200 -100 0 100 200 300 400 500 600 700
Low CSR-activities High CSR-activities
N
et I
n
come Low Net CSR Media Coverage
High Net CSR Media Coverage
34 Table 4
Regression Results Model 3 DV: Net Income
R R2 R2 Change B SE t p
Step 1a .76 .58*
Intercept 214.16 189.23 1.13 .231
CSR-activities 114.38 104.53 1.09 .277
Net CSR Media Coverage -2.19 27.69 -0.08 .937
CSR-activities*Net CSR Media Coverage 19.30 8.82 2.18* .031 Firm Size 5.85 2.88 2.03* .045 R&D Ratio -837.26 5390.28 -0.16 .877 Risk -567.49 420.62 -1.35 .181
Conditional effect of CSR-activities on Net Income at levels of Net CSR Media Coverage
Effect SE t p
Low Net CSR Media Coverage (-1SD) 29.72 127.95 0.23 .817
Average Net CSR Media Coverage 114.38 104.53 1.09 .277
High Net CSR Media Coverage (+1SD) 199.03 92.06 2.16* .033
R R2 R2 Change B SE t p
Step 2a .77 .59** .01
Intercept 148.87 198.13 0.75 .455
CSR-activities 99.77 105.90 0.94 .349
Net CSR Media Coverage 1.74 28.29 0.06 .951
CSR-activities*Net CSR Media Coverage 18.90 8.03 2.35* .021 Firm Size 5.73 2.72 2.11* .038 R&D Ratio -900.04 5467.60 -0.16 .867 Risk -560.91 434.00 -1.29 .200 Negative CSR Score 67.99 68.92 0.99 .327
Note. Statistic Significance: *p < .05, **p < .01, ***p < .001
aAnalysis with HC3 estimator for Heteroscedasticity-consistent SE’s
Exploratory Analyses
We ran exploratory analyses to investigate (1) whether or not there was a difference between the CSR dimensions on financial performance, and (2) if so, which CSR dimension has the most influence on financial performance. Table 5 shows the descriptive statistics and correlations matrix of the variables used in the exploratory analyses.
Ta ble 5 De sc ript ive S ta tist ic s and C or re la ti on M at ri x explor ator y ana lyse s Va ri abl e M SD 1 2 3 4 5 6 7 8 9 10 11 1. Ne t I nc om e (C F P ) a 263. 76 942. 76 2. C om m un ity 0. 07 0. 26 .3 3 ** 3. C or po ra te Gov er na nc e 0. 00 0. 00 - - 4. Dive rsit y 0. 22 0. 51 .4 5 ** .6 7 ** - 5. Em ploye e R ela tio ns 0. 56 0. 96 .5 7 ** .3 9 ** - .4 9* * 6. Envi ron m ent 0. 26 0. 64 .4 0 ** .4 0 ** - .5 5* * .6 3* * 7. Hum an R ig hts 0. 00 0. 00 - - - - - - 8. P rod uc t 0. 13 0. 33 .4 1 ** .5 0 ** - .4 0* * .4 1* * .2 0 - 9. Ne t C S R M edi a C ove ra ge -0. 17 4. 386 .2 7 ** .1 0 - .2 2* .1 3 .1 7 - .1 8 10. F ir m S iz e b 30. 31 65. 30 .5 0 ** .5 3 ** - .5 0* * .5 2* * .4 2* * - .2 1* -.2 9* * 11. R & D R atio 0. 02 0, 07 .0 1 -.0 2 - -.0 3 .1 6 .3 3* * - -.0 1 .1 4 -.0 5 12. R isk 0. 25 0. 27 -.0 5 .0 6 - .1 3 .1 8 .0 8 - -.0 1 .0 4 .0 8 -.0 3 No te : *Co rr el at io n is s ig ni fi ca nt a t t he .0 5 le ve l ( 2-ta il ed ) **C or re la ti on is s igni fi ca nt a t t he .0 1 le ve l ( 2-ta il ed ) a Ne t In co m e is m ea su re d in m il li on s b Fi rm S iz e is m ea su re d in th ou sa nd s
The dimensions corporate governance and human rights were left out of the analyses, because both variables had a Mean and Standard Deviation of 0. This has left us with five dimensions to run our analyses with: community, diversity, employee relations, environment, and product.
To test whether CSR-activities on individual dimensions could predict Net Income we ran five regression analyses, each with the CSR-activities on one of the five dimensions as a independent variable, net income as the independent variable, and firm size, R&D ratio, and risk as control variables. In our exploratory analyses, CSR-activities on the dimensions diversity (β = .285, t(4, 89) = 2.82, p < .01), employee relations (β = .467, t(4, 89) = 4.72, p < .01), environment (β = .268, t(4, 89) = 2.55, p < .05), and product (β = .312, t(4, 89) = 3.58, p < .01) were significant predictors for net income, when the relationship was controlled for firm size, R&D ratio, and risk. Unfortunately, as was the case in our analyses testing our hypothesis, problems of heteroscedasticity were present in these analyses. After rerunning the analyses with the HC3 estimator for heteroscedasticity-consistent standard errors, none of the CSR-activities on individual dimensions could significantly predict net income (p > .144). Complete regression results are shown in Table 6.
To examine if the models were, as was the case with our main analysis of CSR-activities, dependent on media coverage, we ran five moderated regression models. These analyses were identical to the five analyses previously mentioned, with an addition of net CSR media coverage as a moderator. The models with CSR-activities on the dimensions community (F(6, 87) = 31.91, p < .001), diversity (F(6, 87) = 2.46, p < .05), and environment (F(6, 87) = 3.29, p < .01) were statistically significant, and explained respectively 59%, 57%, and 55% of the variance in net income. A closer inspection of the conditional effects indicates that the relation between environmental CSR-activities and net income is only significant when the net CSR media coverage is high (+1SD), compared to when net CSR media
