The effect of communicated CSR motives
on firm financial performance
Name: Lotte Horikx
Student Number: 10340300
First supervisor: Katinka Quintelier
Second supervisor: Flore Bridoux
Faculty: Economics and Business, University of Amsterdam
Study program: BSc Business Administration
Statement of originality
This document is written by Student Lotte Horikx 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.
Table of contents
Abstract………4
1. Introduction………5
2. Literature review………6
3. Theoretical framework………...8
1. Value-driven motives………8
2. Performance-driven motives………..9
3. Stakeholder-driven motives………...11
4. A combination of value- and performance-driven motives
………..12
5. A combination of performance- and stakeholder-driven motives
……….14
6. A combination of value- and stakeholder-driven motives
………15
4. Methodology………..16
1. Sample……….16
2. Independent variables………17
3. Dependent variables………..18
4. Control variables………...18
5. Results………19
1. Descriptives………..19
2. Normality……….20
3. Correlations………..22
4. Regression analysis………...25
6. Discussion………...36
7. Limitation and future research………39
8. Conclusion………...40
9. Bibliography………....41
Abstract
Over the last decades, sustainability has gained importance in the business world, mainly due to increased pressure from society. Firms have responded by investing in social and
environmental initiatives, hoping to improve society’s perception of them and reap the rewards. Even though, there has been quite some debate on the financial benefits of engaging in CSR (Corporate Social Responsibility), there is a large body of research which has found a positive relationship between CSR and a firm’s financial performance. Yet, what part of CSR, be it processes or motives, facilitates this positive relationship, remains unclear. What,
however, has become clear over the years is that stakeholders find the reasons for engaging in CSR increasingly important. Therefore, this research has endeavoured to better understand the effect of a firm’s communicated motives on stakeholders, and the impact this has on the firm’s financial performance, specifically in the food/drink and textile/apparel industry. The application of hierarchical regression analysis, showed there to be no significant relationship between CSR motives and a firm’s financial performance, while controlling for industry, size, risk and R&D expenditure. However, this might be due to the small sample size or exclusion of variables such as social performance and fit. Nevertheless, this research has taken an important step in understanding the relationship between CSR and financial performance, and has opened up new avenues to explore in terms of future research.
1. Introduction
Over the years sustainability has become an increasingly hot topic in the world of business. This is due both to shifts in values and preferences from key stakeholders, particularly in the developed Western world, as well as to the rapid evolution of communication technology (Hartmann, 2011), the latter creating an increasing demand for transparency and sustainability within businesses and their supply chains (Robertson, 2015, Freeman, Harrison, Wicks, Parmar & De Colle, 2010). Not surprisingly, businesses have responded by communicating more fully their ideas about sustainability and their own CSR (Corporate Social
Responsibility) efforts. CSR is a company’s ideas about its social and environmental responsibilities, which are often expressed through various programs, such as pollution reduction or contributing to better education within the community. Firms engage in CSR because of the conviction that this will help their firm in one of multiple ways, be it by increasing their reputation or by lowering operating costs. The end result is then an improved financial performance, or so management believes. Yet, even though Horváthová (2010) has shown that most research has proven there to be a positive relationship between CSR and financial performance, the reason for this is still unclear.
Barnett and Salomon (2012) argue that firms use CSR as a tool for creating and maintaining trusting stakeholder relationships, resulting in an improved stakeholder
influencing capacity (SIC). Barnett (2007) defines SIC as the ability of a firm to act on and profit from opportunities to strengthen stakeholder relationship through CSR. How well a firm does, in terms of their CSR activity, is expressed in its CSP (Corporate Social Performance). Barnett and Salomon (2012) found a U-shaped relationship between CSP and financial performance, but perhaps more importantly the firms with the highest CSP also had the highest financial performance (p. 1316). They see this as support for their view that the perception that stakeholders have of the company’s CSR efforts is paramount in creating a trusting relationship, and in extension the firm’s ability to reap the profits. Stakeholders base these perceptions for a large part on the communications of a firm.
It is, in addition, believed that the motives behind a firm’s CSR engagement are a powerful tool in driving CSR perceptions and stakeholder relationships (Ellen, Web & Mohr, 2006, Barnett, 2007, Barnett & Salomon, 2012, Becker-Olsen & Hill, 2006). The three main motives a firm has to engage in CSR are value-driven, performance-driven and stakeholder-driven (Ellen et al., 2006, Maignan & Ralston, 2002).
The abovementioned CSR motives can elicit various responses by consumers and other stakeholders (Ellen et al., 2006). The question which immediately comes to mind is whether these responses are also expressed in the financial success of organizations. This question has not yet been answered in the current literature (Ellen et al., 2006, p. 156), at least not directly. Research has been done on overall CSR and financial performance (Horváthová, 2010, McWilliams & Siegel, 2000), as well as the effect of CSR motives on purchase intent and consumer responses (Ellen et al., 2006, Groza, Pronschinske, & Walker, 2011), but no research has focussed yet on the direct effect of CSR motives on firm financial performance.
This thesis contributes to both theory and practice. In terms of practice, for instance, a firm may be more inclined to express their performance-driven motives if this will result in higher financial rewards, because consumers believe the motive to be legit (Forehand & Grier, 2003). While in terms of theory, this thesis contributes by shedding light on the role CSR motives play on firm financial performance, which may lead to new research areas. Therefore, this thesis will focus on researching the possible relationship between CSR motives and a firm’s financial performance.
Next, a short review of the relevant literature will be discussed, which will lead to the proposed research question and propositions. Afterwards the research design is presented in the methodology, endeavouring to provide an overview of how an answer to the research question will be found. Following, the results will be presented and linked to the stated propositions. This results in a conclusion and possible ideas for further research.
2. Literature review
Research endeavouring to find and explain the relationship between CSR and firm financial performance has been plagued with inconsistent results (Horváthová, 2010, p. 52).
Nevertheless, Horváthová (2010), has come to the conclusion that most of the literature so far has found that CSR has a positive effect on financial performance. However, CSR is a broad topic, so the actual, practical use for managers is limited when they are unaware what part of CSR, be it principles, processes or communications, causes this positive effect on financial performance.
Many researchers have tried to rectify this problem, by addressing different aspects of CSR. The explicit focus of this research lies with the motives behind CSR, which are the communicated motivations of a firm in why they engage in CSR (Maignan & Ralston, 2002, Hooghiemstra, 2000). As found by Maignan and Ralston (2002), there are three different
motives an organization can act on in addressing CSR issues, namely value-driven,
performance-driven and stakeholder-driven motives. Value-driven motives refer to the firm’s motivation to have a positive impact on stakeholders, communities and the environment, regardless of social pressures. Performance-driven motives, on the other hand, are motives to use CSR as a tool for furthering the firm’s (financial) performance, in terms of profitability, return on investment and sales volume. And finally, the stakeholder-driven motives to respond to stakeholder pressures by conforming to certain CSR norms and behaviour (Maignan & Ralston, 2002, p. 498, Swanson, 1995, Du, Bhattacharya & Sen, 2010, p. 10).
Maignan and Ralston have found a difference in how firms in the US and Europe communicate their CSR principles, processes and stakeholder issues to consumers. Firms in the Netherlands, for instance, are more performance- and stakeholder-driven than value-driven compared to the US (2002, p. 511). Ellen et al. (2006) and Groza et al. (2011), on the other hand, have focussed particularly on the impact of CSR motives on consumer responses and purchase intent. Both articles conclude that value- and performance-driven CSR motives have a positive effect, and stakeholder-driven motives have a negative effect on purchase intent. What’s more, a combination of values is possible as well, which again has another effect. For instance, consumer responses are particularly positive when a firm communicates both value- and performance-driven motives (Ellen et al., 2006, p. 154). This might be due to the fact that consumers believe the company to be more truthful, and therefore trustworthy, when it highlights its personal gain as well (Forehand & Grier, 2003, Ellen et al., 2006).
Next to consumers, a firm is faced with many other stakeholders, such as suppliers, shareholders, communities, the government and so forth. A stakeholder is someone who can affect or be affected by the actions of an organization as a whole (Freeman, 1984, p. 25). Different stakeholders can have different needs and place different demands upon a firm. This applies to the level and direction of CSR engagement as well. However, one thing
stakeholders are in agreement about is the fact that a firm must attend to and correct its social wrongdoings (Murray and Vogel, 1997, as mentioned in Groza et al., 2011, p. 641).
Now in terms of how stakeholder sentiments can be used to gain a financial advantage. In the study conducted by Barnett and Salomon (2012), the relationship between CSP and CFP (corporate financial performance) seems to be U-shaped. This entails that firms with both a low as well as a high CSP have a higher CFP compared to firms with a moderate CSP. Nonetheless, they have found that a firm with a high CSP also has the highest CFP. CFP is another name for a firm’s financial performance. These results are based on the idea that in
performance, in order to build stakeholder influencing capacity (SIC) (Barnett & Salomon, 2012). According to Barnett (2007), firms accrue SIC through consistently engaging in socially responsible acts. SIC offers financial rewards by enabling a firm to assimilate and exploit knowledge and stakeholder favours (Barnett & Salomon, 2012, p. 1304). However, when a firm only has a moderate CSP, the firm is mainly faced with the high costs of
investing in CSR, while not being able to reap the benefits, due to an insufficient SIC (Barnett & Salomon, 2012).
The abovementioned literature suggests that there might be a relationship between CSR motives and firm financial performance. This idea, however, has yet to be tested, that is why the research question of this thesis is as follows:
How do firms’ communicated CSR motives affect their respective financial performance?
Next, the propositions will be discussed in the theoretical framework.
3. Theoretical framework
1. Value-driven motives
To answer the above-stated research question, nine propositions concerning the three different CSR motives have been made using the available literature. Firstly, in terms of value-driven motives, the expectation is that these have a positive effect on firm financial performance. Namely because, value-driven firms can be seen as being pro-active, and go beyond what is expected from a firm by its stakeholders (Murray & Vogel, 1997, p. 144). The firm is often seen as altruistic, since they do not immediately expect something in return. When one views it in terms of the game theory presented by Hoffman, Yoeli and Nowak (2015), value-driven firms are expected to make their choices, regardless of their expected costs and benefits, but based on their principles. Since these principles are not expected to change rapidly
(DiMaggio, 1994), value-driven firms are seen as more trustworthy than firms with self-serving motives, such as performance- or stakeholder-driven, because they will act in
accordance with their values even if it is not in the firm’s best interest (Hoffman et al., 2015). This relationship of mutual trust and cooperation expresses itself in lowering costs, for instance contract costs or transaction costs.
Due to the above stated pro-active and trustworthy nature of value-driven firms, a stock of SIC (Barnett, 2007) is built, and general goodwill towards the company is created.
This will provide the company with a kind of insurance against the effects of negative
publicity in the future, since stakeholders are more resistant to the idea that such a firm could be negatively implicated (Hartmann, 2011, p. 304). As a result the costs of insurance will be lower, but more importantly the risk of losing a part of the customer base is significantly diminished.
Moreover, based on image theory, Jones (1995) argues that companies attract
stakeholders, such as employees and suppliers, who have the same (moral) values and ideas as the company itself. This in turn creates an environment of trust and cooperation, resulting again in lowering costs, for instance by reducing contract costs with suppliers. When values do not align, stakeholders will leave, because they will be less likely to achieve their personal goals (Bridoux & Stoelhorst, 2014). Hence, moral firms will attract moral stakeholders and opportunistic firms will attract opportunistic stakeholders. This alignment will magnify the image embodied by the firm, and will give the moral company a competitive advantage over its opportunistic competitors (Jones, 1995, pp. 419-422, Hartmann, 2011, p. 308).
It is not just a question of costs which create the positive relationship between value-driven motives and firm financial performance. Ellen et al. (2006) and Groza et al. (2011) have shown that the motives consumers attribute to a firm have a significant effect on their purchase intent. Since the value-driven motives are associated with altruism and sincerity, consumers’ purchase intent increases, resulting in a better financial performance. What’s more, the relationship of mutual trust between consumers and the company creates loyalty, which creates the opportunity to ask price premiums (Barnett & Salomon, 2012, p. 1304, Creyer, 1997, p. 428). All in all, value-drive motives will boost sales and lower costs. Therefore, the first proposition is as follows.
Proposition 1: Firms which have value-driven motives as their most prevalent CSR motives, have a higher financial performance than firms with either (a) performance- or (b) stakeholder-driven motives as their most prevalent CSR motives.
2. Performance-driven motives
In terms of the performance-driven motives, research by Ellen et al. (2006, p. 154) found a positive relationship with purchase intent. One of the reasons for this positive attitude of consumers towards strategic CSR decisions, is based on the simple fact that in order to survive a firm has to earn money. This notion is widely accepted as legitimate (Ellen et al.,
2006, p. 150). Yet, while consumers accept performance-driven motives, they favour value-driven ones (Groza et al., 2011, p. 642). This is due to a variety of factors.
First of all, again using the game theory described by Hoffman et al. (2015), a performance-driven firm can be viewed as self-serving. Which means that when the firm believes it is in its best interest to cooperate, it will do so, however, if the firm is not of this opinion, it will default. The same idea applies to CSR investment. It is therefore difficult for stakeholders to know what actions they can expect from the firm, which makes it hard to trust them. This is unlike value-driven firms who are clearly guided by their principles. However, as mentioned earlier (Ellen et al. 2006), when the strategic decision is concerned with the survival of the firm, stakeholders are more forgiving.
What’s more, this scepticism towards performance-driven firms can be mitigated by the simple inclusion of firm-serving benefits in the communication from the firm towards its stakeholders (Forehand & Grier, 2003, p. 353). This is based on the idea, that stakeholders become sceptical of a firm’s actions when, in communications, it hides its ulterior motives (Forehand & Grier, 2003, p. 351). So, when the firm is honest about the self-serving aspects of CSR, scepticism is removed and a form of acceptance or even a form of trust is created with stakeholders. Therefore, corporate credibility remains intact (Becker-Olsen et al, 2006).
Yet, despite the removal of scepticism, when a company mentions that their own interest are what really guided their decision in taking part in CSR, one cannot help but notice the paradox. The firm is acting only for self-serving reasons, while the CSR action in itself is other-serving. When this is shown in an example it makes sense. For instance, when a
company invests in creating awareness among employees about healthier eating habits, the employees will live longer and healthier, thus work more years at the firm. This reduces costs, because of fewer days sick-leave, lower turnover rate and lower training costs. Hence, the communication of performance-driven motives is based on a logical sequence of events, and is therefore seen by stakeholders as legitimate (Ellen et al., 2006).
So, the expectation is that while performance-driven motives reduce costs and increase sales, the relationship is not as strong as the value-driven one, because it does not increase the mutual trust and loyalty. It is, however, expected to be stronger than the stakeholder-driven motive, since it at least mitigates the prevalent scepticism among stakeholders. All in all, this suggests that a better financial performance can be achieved, when the firm expresses its performance-driven motives, leading to the second proposition.
motives, have a (a) higher financial performance than firms with stakeholder- driven motives, but a (b) lower financial performance than firms with value driven motives, as their most prevalent CSR motives.
3. Stakeholder-driven motives
However, contrary to the positive influence of both value- and performance-driven motives on purchase intent, stakeholder-driven motives are often believed to have a negative influence (Groza et al., 2011, p. 643, Ellen et al., 2006, p. 154). This is probably due to the negative duty (Swanson, 1995), which entails that the firm only meets the CSR requirements, because they are pressured to do so by their stakeholders (Ellen et al., 2006, p. 149). This can be seen as a active response by the company, and is often viewed as forced and insincere. A re-active response suggests that before stakeholders should ask or demand some sort of action from the firm, a specific (irresponsible) event should have occurred to elicit this (Groza et al., 2011, p. 641). According to Becker-Olsen et al. (2006), these events often have a negative effect on consumers’ thoughts and attitudes towards the firm. This may lead to reduced corporate credibility, due to an increase in feelings of corporate self-interest, decreased corporate legitimacy and loss of feelings of trust and honesty (Kernisky, 1997).
In terms of game theory (Hoffman et al., 2015), stakeholder- and performance-driven motives have in common that they both are self-serving, and thus are often viewed with scepticism. Only the difference is that performance-driven motives have some form of strategic goal, such as making profit, whereas stakeholder-driven motives are just based on pleasing relevant stakeholders. This does not create a relationship of cooperation and trust, but one of scepticism (Du et al., 2010). This scepticism is not removed, unlike with value- and performance-driven motives (Forehand & Grier, 2003).
However, according to McGuire, Sundgren and Schneeweis (1988), stakeholder theory suggests that a firm is faced with both explicit as well as implicit claims by stakeholders. Explicit claims are demands that are clearly expressed to the firm. Implicit claims, on the other hand, are unspoken demands, which are not clearly communicated to the firm, and are based on what stakeholders would prefer to see the firm do instead of demanding a certain behaviour. It further suggests that explicit goals, such as wage contracts, are costlier to the firm than implicit claims, such as product quality. However, lower CSR engagement may lead to doubts on the ability of the firm to honour its stakeholders’ implicit claims. This, in turn, could lead to an increase in the firm’s number of more costly explicit claims (McGuire et al.,
1988, pp. 854-855). Thus, the expectation is that stakeholder-driven firms are faced with more distrust from stakeholders, resulting in a surge of explicit claims, heightening a firm’s costs.
What’s more, when prospect theory is applied, stakeholder-driven motives can be seen as just meeting the stated reference point. This reference point can be based, for instance, on enforced legislation or animal welfare demands made by stakeholders, but more importantly it is the accepted minimum. This compliance with demands or laws does not provide the
company with the insurance, with which value-driven firms are safe-guarded, in order to mitigate negative effects caused by an unfortunate event (Barnett & Salomon, 2012, Barnett, 2007). Since losses are weighed more heavily than gains (Hartmann, 2011, p. 304), the negative effect of such an event will be detrimental to the firm, due to decreases in perceptions and purchase intent, and thus resulting in a lower financial performance.
Lastly, the lack of strategic direction and simple adherence to stakeholder demands, may lead the firm to invest but a small to moderate amount in CSR. Based on the research by Barnett and Salomon (2012), the continuum of CSR and financial performance takes a U-shaped form. This entails that when a firm invests a moderate amount, such as stakeholder-driven firms are expected to do, they will have a markedly lower financial performance than either value- or performance-driven firms, which are expected to invest more heavily. This is probably due to the cost of actually investing in CSR, and yet not being able to reap the benefits, like value- and performance-driven firms can (Barnett & Salomon, 2012, Barnett, 2007).
All in all, the expectation is that the communication of stakeholder-driven motives results in lowering consumer perception and purchase intent, and increasing costs, while at the same time failing to remove the scepticism felt by relevant stakeholders. This leads to the third proposition.
Proposition 3: Firms which have stakeholder-driven motives as their most prevalent CSR motives, have a lower financial performance than firms with either (a) value-
or (b) performance-driven motives as their most prevalent CSR motives.
4. A combination of value- and performance-driven motives
Next, previous research suggests that a combination of CSR motives may have an even stronger effect on purchase intent than a single motive (Ellen et al., 2006, pp. 150-154). This suggests that firms with a combination of value- and performance-driven motives have a better financial performance than firms with either motive on its own. This is probably due to
the fact that lone value motives are sometimes viewed with a certain amount of scepticism, because the possible financial benefit of CSR for the firm is not communicated (Forehand & Grier, 2003, Du et al., 2010). Likewise, a lone performance motive lacks the mutual trust and altruistic reputation of the value motive. Furthermore, previous research has shown that stakeholders can accept the duality in persuasive or motivational communications (Williams & Aaker, 2002, Ellen et al., 2006, p. 155). This, in turn, suggests that the mentioning of self-serving/performance-driven motives eliminates almost all the scepticism of consumers towards the firm, and makes a combination with value-driven motives more trustworthy and legitimate (Forehand & Grier, 2003, p. 353, Du et al., 2010, p. 17). This creates general goodwill (Groza et al., 2011). So, compared to both lone value- and performance-driven motives, the combination of the two is expected to have a greater positive effect on financial performance, forming the basis for proposition four.
Proposition 4: Firms which have a combination of value- and performance-driven motives, as their most prevalent CSR motives, have a higher financial performance than firms with either lone (a) value- or (b) performance-driven motives as their most prevalent CSR motives.
Furthermore, in comparison to the other two possible combinations, value- and stakeholder-driven or and stakeholder-stakeholder-driven, the combination of value- and performance-driven motives is expected to have a stronger positive effect on financial performance than the other possible combinations (Ellen et al, 2006, p. 154). This is because, performance- and stakeholder-driven motives do not increase trust and goodwill. While, a combination of value- and stakeholder-driven motives, is expected to be viewed with even more scepticism than any other combination. This is probably due to the implausible nature of such a combination. More specifically, that a firm invests in CSR because it believes it to be the right choice, while at the same time the firm only invests because it is pressured to do so by its stakeholders. In conclusion, the combination value- and performance-driven motives is expected to generate a higher financial return than the other two possible combinations, thus proposition five.
Proposition 5: Firms which have a combination of value- and performance-driven motives, as their most prevalent CSR motives, have a higher financial
performance than firms with either a combination of (a) value- and
stakeholder-driven or (b) performance- and stakeholder-driven motives, as their most prevalent CSR motives.
5. A combination of performance- and stakeholder-driven motives
In terms of the combination performance- and stakeholder-driven motives, the expectation is that the combination of these motives will have a more positive effect on firm financial performance, than either motive could achieve on its own. This is based on the earlier mentioned concept, that when performance-driven motives are communicated, stakeholders will be less sceptical of the firm’s hidden agenda (Forehand & Grier, 2003, p. 353, Du et al., 2010). Which in turn will legitimize the company. This concept also applies to the
combination of performance- and stakeholder-driven motives. When the motives are viewed for their own merits and demerits, performance-driven motives mitigate scepticism. Only they lack the acknowledgement that other stakeholders matter to the firm aside from shareholders. For it is the shareholders who receive the profits, and can exert pressure on the firm’s
management for new initiatives. While, on the other hand, stakeholder-driven firms put their focus on the needs and, mainly, demands of all their stakeholders. However, they lack the necessary credibility (Groza et al., 2011). Thus, as suggested, a combination will be more financially rewarding, as expressed in proposition six.
Proposition 6: Firms which have a combination of performance- and stakeholder-driven motives, as their most prevalent CSR motives, have a higher financial performance than firms with either lone (a) performance- or (b) stakeholder-driven motives as their most prevalent CSR motives.
As stated previously, the combination of performance- and stakeholder-driven motives, is expected to have a lower financial performance in comparison to a combination of value- and performance-driven motives. Nonetheless, it is expected to yield a higher financial
performance than the combination of value- and stakeholder-driven motives, based on the lack of credibility concerning the latter, as argued previously. Hence, proposition seven.
Proposition 7: Firms which have a combination of performance-driven and stakeholder- driven motives, as their most prevalent CSR motives, have a (a) higher financial performance than firms with a combination of value- and
stakeholder-driven motives, but a (b) lower financial performance than firms with a combination of value- and performance-driven motives, as their most prevalent CSR motives.
6. A combination of value- and stakeholder-driven motives
Finally, the combination of value- and stakeholder-driven motives is expected to be the least financially rewarding of all possibilities, because the scepticism brought on by a stakeholder focus is not eliminated (Ellen et al. 2006, Forehand & Grier, 2003). More specifically, communicating that the firm believes in CSR without social pressure and at the same time mentioning that CSR is demanded by their stakeholders, is contradictory, and therefore implausible. As argued previously, this will increase the scepticism and doubt of stakeholders even more than is the case with the lone stakeholder motive. The mutual trust of the value-driven motive, which is based on honesty, will be diminished, thereby mitigating the positive offset of this motive. Thus, the expectation is that a value- and stakeholder-driven
combination has an even stronger negative effect on firm financial performance than a lone value- or stakeholder-driven motive, as stated in proposition eight.
Proposition 8: Firms which have a combination of value- and stakeholder-driven motives, as their most prevalent CSR motives, have a lower financial performance than firms with either lone (a) value- or (b) stakeholder-driven motives as their most prevalent CSR motives.
As reasoned previously, the combination value- and stakeholder-driven motives is also expected to have a lower financial performance in comparison to both the value- and
performance-driven as well as performance- and stakeholder-driven motive combinations, due to scepticism and lack of trust. Which is the basis for the ninth, and last, proposition.
Proposition 9: Firms which have a combination of value- and stakeholder-driven motives, as their most prevalent CSR motives, have a lower financial performance than firms with either a combination of (a) value- and performance-driven or (b) performance- and stakeholder-driven motives as their most prevalent CSR motives.
4. Methodology
In order to answer the research question and corresponding propositions, a quantitative, empirical research design was chosen. Secondary data sources, specifically the database COMPUSTAT and the annual reports of the selected companies, were used for collecting all relevant data. These sources were used, because they have all the necessary information for answering the research question, demonstrated by the repeated use in relevant literature. What’s more, the data is easy to access, which is an important factor to keep in mind when the research is constrained for time, as this thesis was (Saunders, Lewis & Thornhill, 2012, pp. 318-323). With the collected data a hierarchical linear regression analysis was performed, as well as normality tests, in order to make sure that all the requirements for parametric testing were met (Saunders et al., 2012, pp. 524-525).
1. Sample
The research focuses on companies within the food and drink industry and the textile and apparel industry, because over the last decade, increasing attention has been paid to their social and environmental policies (Hartmann, 2011). This attention is primarily due to the nature of these industries, namely that they cater to basic human needs and thus have a high impact. Therefore, they are more critically scrutinized in terms of meeting certain
environmental and social requirements (Hartmann, 2011, pp. 297-298). This scrutiny has led to numerous scandals exposing firms and their failures to comply with the requested
standards, think for instance of the apparel brand Nike which admitted to using child labour (Boggan, 2001). Therefore, the expectation is that firms in these two industries have acted accordingly by increasing their CSR efforts and communications. This facilitates a possible relationship between CSR motives and firm financial performance that is more pronounced than in other industries.
Furthermore, the use of just two industries removes most of the possible discrepancies in results between industries, based on, for instance, their differences in stakeholder
relationships or the amount of resources allocated to different activities (Fu & Jia, 2012, p. 131). So, the use of two industries ensures that the sample is a fair representation of the population. A sample of firms was drawn from the S&P 500 and yahoo finance on the biggest and best performing firms in North America in the food/drink and textile/apparel industry. This led to a total of 134 firms, namely, 66 food/drink and 68 textile/apparel firms. Every firm on the list was then checked for the availability of annual reports from 2011, and relevant
financial data from 2012 in COMPUSTAT. A total of 36 firms was then eliminated due to missing data, resulting in a final list of 98 firms, namely, 54 food/drink and 44 textile/apparel firms.
2. Independent variables
The annual reports of the selected firms were used for coding the CSR motives, specifically, value-, performance- and stakeholder-driven motives. Research conducted by Groza et al. (2011) showed that when communications came from internal sources (websites, annual reports, etc.) instead of external ones (newspapers, Forbes, etc.), the motives attributed to a firm’s CSR engagement were amplified (Groza et al., 2011). Therefore, the expectation is that the communication of motives through annual reports will have a pronounced effect on a firm’s financial performance.
The coding was done as follows. First, twenty firms were randomly chosen and analysed for words relating to CSR, which could be placed under one of the three motives. This resulted in a coding scheme (Appendix, section 2) consisting of 21 key terms for performance-driven motives, 26 key terms for value-driven motives, and 16 key terms for stakeholder-driven motives. This coding scheme was then used to code all annual reports, noting how many times a certain motive was mentioned, resulting in an overall tally of each motive.
Afterwards, a categorical variable was made, based on predetermined parameters for dividing the different motives. For instance, when a certain motive, such as value, had been mentioned so often that it made up more than fifty percent of the overall motives mentioned, and the other two motives, performance and stakeholder, were each mentioned less than thirty percent of the time, the firm is labelled as value-driven. The same parameters were used for defining the other lone motives. A combination of motives was deduced when two motives were present in almost equal measure within one firm. The chosen parameters for a
combination were as follows. When, for instance value and performance motives are each mentioned over thirty percent of the time, and the stakeholder value is mentioned less than twenty percent overall, the firm is labelled as value- and performance-driven.
The categorical variable was then used to create different dummy variables in order to answer the stated propositions. For example, the dummy variable Value had value 1 when the firm was labelled as only value-driven, and 0 when the firm was defined by any other motive or motive combination. When the regression analysis was performed, it was possible to use
the other dummy variables as control variables, while excluding the dummy variable with which the independent variable was being compared.
3. Dependent variables
The firm’s financial performance is measured by their ROA (return on assets) and net income, because these measures have been shown, in various studies, to represent a firm’s financial performance (Fu & Jia, 2012, p. 134, Berman et al., 1999). The financial data is collected from the COMPUSTAT database and includes financial data for the year 2012. This entails that the chosen time lag in this research was one year. Since there is, as of yet, no consensus among researchers what the exact time frame should be (Hartmann, 2011, p. 304), a year provides the stakeholders with time to process and act on the expressed CSR motives, while making sure that the motives are still relatively fresh in their memories. What’s more, the use of a one-year time lag by Barnett and Salomon (2012, p. 1310) has shown it to be as valid as it can possibly be, within a field with so little consensus on the matter.
4. Control variables
Four variables have been chosen to control for any variations in possible outcomes, namely industry, size, risk and R&D expenditure. The selection of the aforementioned control variables is based on their ability to influence the relationship between CSR (motives) and firm financial performance, as shown in previous literature, thus making them more reliable (Andersen & Dejoy, 2011, Fu & Jia, 2012, Barnett & Salomon., 2012). All control variables are collected through the database COMPUSTAT (McWilliams & Siegel, 2000). Firstly, industry is believed to be an important control variable, because the level of differentiation within an industry is believed to impact financial performance (Fu & Jia, 2012, p. 131). Even though only two industries were taken into account, there might still be a difference in impact which needs to be controlled for. Industry is measured based on the economic sector
(Andersen & Dejoy, 2011, p. 242). Secondly, the size data is based on the firm’s total sales figures (Andersen & Dejoy, 2011, p. 241, Fu & Jia, 2012, p. 132). The expectation is that a firm’s size impacts its financial performance due to, for instance, economies of scale or scope (McWilliams & Siegel, 2001, pp. 123-124). Another control variable used in many researches is risk, which is measured by the debt ratio (total debt/total assets) (Orlitzky & Benjamin, 2001, Fu & Jia, 2012, Barnett & Salomon, 2012). Risk is used as control variable, because scholars have pointed out its potential impact of influencing managers’ behaviour. Either by disciplining managers in making the right decisions for the firm, or by decreasing managerial
latitude and thereby limiting exploration of opportunities, thereby impacting profit (Barnett & Salomon, 2012, p. 1310). And finally, R&D expenditures, are used to control the relationship between CSR motives and firm financial performance. R&D expenditure is an important control variable, because literature has found it linked to improved long-run economic performance (McWilliams & Siegel, 2000, Fu & Jia, 2012).
5. Results
1. Descriptives
In this research, data was collected from 98 firms. However, the normality analysis required an elimination of extreme cases, which reduced the number of firms ready to use down to 79. Furthermore, the control variable R&D expenditure was often not listed or non-existent for firms in the COMPUSTAT database, which resulted in only 45 firms with this variable. In order to run the regression with 79 firms, the missing R&D expenditure values were replaced with the variable’s mean. The regression was also performed with the exclusion of missing values listwise, to ascertain that this had no significant effects on the results. This was not the case, therefore, the regression was performed with 79 firms instead of 45.
Two categorical variables were used in this research, namely the control variable industry, and the independent variable communicated motive. Both variables have been recoded into dichotomous variables for further analysis. The descriptives of these variables are shown below in table 1. In terms of industry, firms were present in almost equal measure, with 43 firms working in the food and drink industry (54.4%), and 36 working in the textile and apparel industry (45.6%). However, the variable communicated motive showed more variation. Most firms, 18 (22.8%), communicated lone stakeholder-driven motives, while only 3 firms (3.8%) communicated lone performance-driven motives. The combination most often expressed by firms was the value/stakeholder-driven motive, specifically by 13 firms (16.5%). The least expressed motive combination, by just 7 firms (8.9%) was performance/stakeholder-driven. As shown in the table there were 15 firms (19%), with no discernible motive, which means that they could not be placed in any of the pre-determined categories. However, they were not removed from the data in order to serve as another control measure. This was done by including them into the dummy variables.
Table 1. Descriptive statistics for categorical variables
Variable Levels N %
Industry Food and Drink 43 54.4
Textile and Apparel 36 45.6
Communicated motive Value 15 19.0
Performance 3 3.8
Stakeholder 18 22.8
Value-Performance 8 10.1
Performance-Stakeholder 7 8.9
Value-Stakeholder 13 16.5
Other (belonging to no discernible category) 15 19
Besides the two categorical variables, five continuous variables have been used in this research, namely the dependent variables net income and ROA, and the control variables sales, risk and R&D expenditure. The descriptives of these variables are shown below in table 2. The values of net income ranged between -405.02 and 1383.80, with a mean of 249.58 and a standard deviation of 312.42. ROA on the other hand, ranged from -0.24 to 0.34 with a mean of 0.079 and a standard deviation of 312.42. In terms of sales, values ranged between 399.64 and 25878.37 with a mean of 4499.67 and a standard deviation of 5389.67. The values of risk were ranging from 0.00 to 10.82 with a mean of 0.86 and a standard deviation of 1.75. Lastly, the control variable R&D expenditure showed a range of values from 0.00 to 259 with a mean of 27.98 and a standard deviation of 5389.67.
Table 2. Descriptive statistics for continuous variables
Variable N Min Max Mean Std. Dev.
Net Income 79 -405.02 1383.80 249.58 312.42
ROA (return on assets) 79 -.24 .34 .079 .091
Sales 79 399.64 25878.37 4499.67 5389.67
Risk (Debt ratio) 78 .00 10.82 .86 1.75
R&D Expenditure 45 .00 259 27.98 5389.67
2. Normality
A normality test was conducted to measure the adequacy of the five continuous variables, namely net income, ROA, sales, risk and R&D expenditure. The descriptives of the normality analysis can be found in the appendix, specifically section 1 table 1. The results of the
Kolmogorov-Smirnov test can be found in appendix section 1 table 2.
First of all, net income had a positive skewness of 1.338 and a positive kurtosis of 1.990, after eliminating the extreme cases from the data. The rest of the outliers visible in the
boxplot (appendix section 1, graph 3) were not eliminated, because too much valuable data might have been lost. To create a closer to normal distribution, transformations on net income were attempted, only to result in even higher skewness and kurtosis. So, net income has not been transformed. The Kolmogorov-Smirnov test still showed that net income was
significantly different from a normal distribution. However, in the Q-Q plot of net income (appendix section 1, graph 2), the values are believed to be reasonably close to a straight line, thus a normal distribution. Moreover, George and Mallery (2010) argue that skewness and kurtosis between -2 and +2 is within the boundaries of an acceptable normal distribution. So, even though everything has been done to minimize skewness and kurtosis, net income is still somewhat different from a normal distribution. However, it is believed that this will not drastically affect the results.
ROA had similar problems with its distribution as net income. The kurtosis was 3.427 and thus above the norm of George and Mallery (2010). On the other hand, skewness was low with a value of 0.12. As shown in the histogram, the distribution was barely skewed but somewhat peaked in the centre (appendix section 1, graph 4). The Kolmogorov-Smirnov test also showed a significant difference from a normal distribution. However, the elimination of extreme values and attempted transformations did nothing to improve the distribution. Therefore, the distribution, using all the original data, is expected to be as close to a normal distribution as it can be. This belief is supported by the relatively straight line shown in the Q-Q plot (appendix section 1, graph 5).
Where net income and ROA, did not respond well to transformation, sales greatly improved. After a logarithmic transformation and the elimination of extreme values, sales (Log Sales) showed a positive skewness of 0.309 and a negative kurtosis of -0.623. Which means that the distribution is slightly skewed to the left and somewhat flat in the centre (appendix section 1, graph 7). However, the distribution is not significantly different from normal, as shown by the Kolmogorov-Smirnov test. The Q-Q plots supports this finding, showing a relatively straight line (appendix section 1, graph 8). Thus, the distribution of logarithmic sales (Log Sales) is not significantly different from a normal distribution.
After transforming risk (Transf. Risk), using the formula X*=1/X, and eliminating extreme cases, the distribution of Transf. Risk showed a skew to the right (-1.021) and a slightly flat centre (-0.391), as shown in the histogram (appendix section 1, graph 10). Despite the transformation, risk still had a significantly different distribution than normal, as shown by the Kolmogorov-Smirnov test. However, the skewness and kurtosis are within the boundaries
set by George and Mallery (2010). Thus, the slightly abnormal distribution of risk (Transf. Risk) is not expected to have a drastic effect on the results.
Lastly, the elimination of extreme cases and the logarithmic transformation markedly improved the distribution of R&D expenditure (Log R&D Expenditure). With a skewness of 0.897 and a kurtosis of -0.868 the distribution is clustered to the left and somewhat flat in the centre (appendix section 1, graph 13). The Kolmogorov-Smirnov test reveals a significant difference from a normal distribution, and the Q-Q plot is somewhat different from a straight line (appendix section 1, graph 14). However, the skewness and kurtosis are not outside the boundaries set by George and Mallery (2010), and thus results are not expected to be greatly affected.
3. Correlations
Correlations between the variables used in this research, were investigated using the Pearson correlation coefficient. The control variable industry, was transformed into a dummy variable, where the food and drink industry was assigned 0 and the textile and apparel industry was assigned 1. This was done based on the fact that the food and drink industry was the largest group of firms, and according to Hardy (2013) should therefore serve as the reference group. The communicated motives, were also changed to dummy variables, which entails that for the dummy variable value, lone value-driven motives were assigned 1 and all the other motives or combinations were assigned 0. This process was repeated for the other motive variables. The results of the correlation analysis are presented in table 3 at the end of this section.
Only the significant correlations will be discussed in this section. First of all, there was a significant positive correlation between Net income and ROA (r = 0.512, N = 79, Sig. =0.000). This entails that when net income increases so does ROA. This is not surprising, considering that ROA is net income divided by total assets. The correlation was just higher than 0.5, which is seen as the criterion for a statistically significant relationship (Field, 2013, p. 275).
In terms of control variables, a significant positive correlation was found between Transf. Risk and Net income (r = 0.376, N = 78, Sig. = 0.001). Which means that when risk, the debt ratio, increases so does net income. Furthermore, significant relationships were found between Transf. Risk and Log Sales (r = -0.543, N = 78, Sig. = 0.000), Log R&D expenditure and Log Sales (r = 0.616, N = 45, Sig. = 0.000), and Log R&D expenditure and Transf. Risk (r = -0.720, N = 44, Sig. = 0.000). This might present a multicollinearity problem, which entails that two or more predictor variables are so strongly correlated that they measure the
same construct. The threshold for possible multicollinearity, presented by Field (2013, p. 325), is 0.8. Not any of the correlations between control variables are above 0.8, these scales are therefore not likely to cause multicollinearity. The independent variable Value shows a significant correlation with Industry (r = -0.313, N = 79, Sig. = 0.005), Log Sales (r = 0.262, N = 79, Sig. = 0.020) and Log R&D expenditure (r = 0.302, N = 45, Sig. = 0.044). This means that when a firm communicates that it is solely value-driven it is more likely to be in the food and drink industry, sales will rise and more money will be invested in R&D. The combination of value- and performance-driven motives is positively correlated with Industry (r = 0.367, N = 79, Sig. = 0.001). This entails that a firm which communicates a combination of value- and performance-driven motives is more likely to be in the textile and apparel industry. The combination of value- and stakeholder-driven motives is significantly, and negatively correlated with Industry (r = -0.269, N= 79, Sig. = 0.017). This means that when a firm communicates both its value and stakeholder motives, it is more likely to be in the food and drink industry.
Between motive variables, there was a significantly negative correlation between value and stakeholder motives (r = -0.263, N = 79, Sig. 0.019), as well as between a combination of value-performance and value-stakeholder motives (r = -0.241, N = 79, Sig. = 0.032).
However, correlation did not exceed 0.8, and is therefore not likely to cause multicollinearity (Field, 2013, p. 325).
4. Regression analysis
A hierarchical linear regression analysis was conducted to test the stated propositions and answer the research question. Net income and ROA were added as the dependent variables in the regression. The control variables, industry, sales, risk and R&D expenditure, were entered into the first model of the regression. What’s more, the communicated motive variables were also added to the first model to serve as an additional control variable. The independent, predictor variable was added to the second model. Only the motive variable to which the predictor variable was being compared was not added to any model. So, for example, proposition 1a states that value-driven firms have a higher financial performance than performance-driven firms. This means that the control variables, as well as the variables stakeholder, value-performance, performance-stakeholder and value-stakeholder are added to model 1, and the dummy variable value is added to model 2. This way, when model 2 shows a significant improvement over model 1, value-driven firms have a higher financial
performance than performance-driven firms, because the variable performance is the only variable not added, and thereby the one to which value is compared.
The results of the regression regarding proposition 1a/b, for both net income and ROA, are presented in table 4 and 5 at the end of this section. Concerning proposition 1a, the results show that 29% of the variance concerning net income was explained by the control variables, and only 7.8% for ROA. However, model 1 was significant for net income (F = 3.644, Sig. = 0.001), yet not for ROA (F = 0.739, Sig. = 0.657). Even though the variable value was added, no additional variance of net income was explained. For ROA an additional 29% variance was explained by model 2, only this was insignificant (F = 0.662, Sig. = 0.729). Therefore, no support has been found for proposition 1a. Regarding proposition 1b, the control variables explained 30% of the variance of net income and 7.3% of ROA, where model 1 was significant for net income (F = 3.743, Sig. = 0.001), it was insignificant for ROA (F = 0.686, Sig. = 0.703). Model 2 did not improve the amount of variance which could be explained for either net income (F = 3.291, Sig. = 0.790), or ROA (F = 0.636, Sig. = 0.590). So, value-driven firms do not have a higher financial performance than stakeholder-driven firms, and thus no support has been found for proposition 1b.
Table 4. Results of hierarchical linear regression analysis for proposition 1a
Net Income ROA
Variable Model 1 Model 2 Model 1 Model 2
Transf. Risk 648.985 .000 649.966 .000 .054 .290 .055 .283
Log Sales 250.866 .005 252.814 .006 .009 .744 .012 .690
Log R&D expenditure 40.472 .537 41.307 .536 .006 .776 .007 .742
Stakeholder -47.000 .569 -51.511 .598 -.041 .142 -.047 .155 Value-Performance -73.293 .536 -76.311 .539 -.008 .830 -.012 .763 Performance-Stakeholder -123.418 .303 -127.116 .320 -.055 .165 -.060 .157 Value-Stakeholder 73.686 .430 68.917 .526 -.032 .306 -.038 .292 Value -9.779 .930 -.013 .729 𝑅2 .294 .294 .078 .079 𝑅2change .294 .000 .078 .002 F 3.644 3.194 .739 .662 Sig. F change .001 .930 .657 .729
Table 5. Results of hierarchical linear regression analysis for proposition 1b
Net Income ROA
Variable Model 1 Model 2 Model 1 Model 2
B Sig. B Sig. B Sig. B Sig.
Industry -48.288 .509 -42.208 .583 .001 .960 .005 .836
Transf. Risk 645.651 .000 642.431 .000 .053 .297 .051 .321
Log Sales 255.324 .004 251.353 .006 .010 .728 .007 .802
Log R&D expenditure 41.082 .529 39.068 .555 .006 .773 .005 .823
Value-Performance -35.135 .760 -31.712 .786 .017 .660 .019 .621 Performance-Stakeholder -94.149 .417 -88.644 .455 -.035 .370 -.031 .432 Value-Stakeholder 95.167 .290 103.932 .281 -.016 .595 -.010 .755 Performance 162.398 .350 164.536 .347 .078 .184 .079 .179 Value 25.189 .790 .017 .590 𝑅2 .300 .300 .073 .077 𝑅2change .300 .001 .073 .004 F 3.743 3.291 .686 .636 Sig. F change .001 .790 .703 .590
The results for the regressions regarding proposition 2a/b are shown below in tables 6 and 7. Proposition 2a, stating that performance-driven firms have a higher financial performance than stakeholder-driven firms, is not supported by the results, either in terms of net income or ROA. While the control variables explain 29.1% of the variance in net income (F = 3.596, Sig. = 0.002) and 5.2% in ROA (F = 0.479, Sig. = 0.867), model 2 does not significantly improve the ability to explain the variance of either net income (F = 3.291, Sig. = 0.347) or ROA (F= 0.636, Sig. = 0.179). Proposition 2b, which states that performance-driven firms have a lower financial performance than value-driven firms, also remains unsupported by the results. Model 1 explains 29.4% of the variance in net income (F= 3.644, Sig. = 0.001) and 7.8% of the variance in ROA (F = 0.739, Sig. = 0.657). Model 2, however, does not
significantly improve the explanatory power for either net income (F = 3.304, Sig. = 0.404) or ROA (F = 0.786, Sig. = 0.286).
Table 6. Results of hierarchical linear regression analysis for proposition 2a
Net Income ROA
Variable Model 1 Model 2 Model 1 Model 2
B Sig. B Sig. B Sig. B Sig.
Industry -26.639 .723 -42.208 .583 .013 .613 .005 .836
Transf. Risk 648.848 .000 642.431 .000 .054 .295 .051 .321
Log Sales 243.663 .007 251.353 .006 .004 .900 .007 .802
Log R&D expenditure 38.277 .563 39.068 .555 .005 .837 .005 .823
Value-Performance -52.007 .650 -31.712 .786 .010 .805 .019 .621 Performance-Stakeholder -102.025 .386 -88.644 .455 -.038 .343 -.031 .432 Value-Stakeholder 95.556 .319 103.932 .281 -.014 .662 -.010 .755 Value 21.097 .823 25.189 .790 .015 .635 .017 .590 Performance 164.536 .347 .079 .179 𝑅2 .291 .300 .052 .077 𝑅2change .291 .009 .052 .025 F 3.596 3.291 .479 .636 Sig. F change .002 .347 .867 .179
Table 7. Results of hierarchical linear regression analysis for proposition 2b
Net Income ROA
Variable Model 1 Model 2 Model 1 Model 2
B Sig. B Sig. B Sig. B Sig.
Industry -27.796 .697 -44.007 .554 .013 .597 .006 .816
Transf. Risk 648.985 .000 644.391 .000 .054 .290 .052 .307
Log Sales 250.866 .005 257.288 .004 .009 .744 .012 .675
Log R&D expenditure 40.472 .537 41.338 .529 .006 .776 .007 .763
Value-Performance -73.293 .536 -49.891 .682 -.008 .830 .001 .971 Performance-Stakeholder -123.418 .303 -107.090 .378 -.055 .165 -.049 .230 Value-Stakeholder 73.686 .430 84.317 .372 -.032 .306 -.027 .383 Stakeholder -47.000 .569 -33.044 .695 -.041 .142 -.035 .218 Performance 148.636 .404 .063 .286 𝑅2 .294 .301 .078 .093 𝑅2change .294 .007 .078 .015 F 3.644 3.304 .739 .786 Sig. F change .001 .404 .657 .286
Regression results concerning proposition 3a/b are presented in tables 8 and 9, respectively. According to proposition 3a, stakeholder-driven firms have a lower financial performance than value-driven firms. The results show that model 1 explains 30% of the variance in net
Model 2 does not significantly improve the ability to explain variance in either net income (F = 3.304, Sig. = 0.695) or ROA (F = 0.786, Sig. = 0.218). Therefore, proposition 3a is not supported. Proposition 3b states that stakeholder-driven firms have a lower financial
performance than performance-driven firms. While model 1 explains 29.1% of the variance in net income (F = 3.596, Sig. = 0.002), the inclusion of the variable stakeholder does not
significantly improve the explanatory power for net income (F = 3.194, Sig. = 0.598). The same goes for ROA, where model 1 explains 5.2% in variance (F = 0.479, Sig. = 0.867) and model does not significantly improve the explanatory power of the model (F = 0.662, Sig. = 0.155). Thus, proposition 3b is not supported by the results.
Table 8. Results of hierarchical linear regression analysis for proposition 3a
Net Income ROA
Variable Model 1 Model 2 Model 1 Model 2
B Sig. B Sig. B Sig. B Sig.
Industry -48.288 .509 -44.007 .554 .001 .960 .006 .816
Transf. Risk 645.651 .000 644.391 .000 .053 .297 .052 .307
Log Sales 255.324 .004 257.288 .004 .010 .728 .012 .675
Log R&D expenditure 41.082 .529 41.338 .529 .006 .773 .007 .763
Value-Performance -35.135 .760 -49.891 .682 .017 .660 .001 .971 Performance-Stakeholder -94.149 .417 -107.090 .378 -.035 .370 -.049 .230 Value-Stakeholder 95.167 .290 84.317 .372 -.016 .595 -.027 .383 Performance 162.398 .350 148.636 .404 .078 .184 .063 .286 Stakeholder -33.044 .695 -.035 .218 𝑅2 .300 .301 .073 .093 𝑅2change .300 .002 .073 .020 F 3.743 3.304 .686 .786 Sig. F change .001 .695 .703 .218
Table 9. Results of hierarchical linear regression analysis for proposition 3b
Net Income ROA
Variable Model 1 Model 2 Model 1 Model 2
B Sig. B Sig. B Sig. B Sig.
Industry -26.639 .723 -29.846 .693 .013 .613 .010 .694
Transf. Risk 648.848 .000 649.966 .000 .054 .295 .055 .283
Log Sales 243.663 .007 252.814 .006 .004 .900 .012 .690
Log R&D expenditure 38.277 .563 41.307 .536 .005 .837 .007 .742
Value-Performance -52.007 .650 -76.311 .539 .010 .805 -.012 .763 Performance-Stakeholder -102.025 .386 -127.116 .320 -.038 .343 -.060 .157 Value-Stakeholder 95.556 .319 68.917 .526 -.014 .662 -.038 .292 Value 21.097 .823 -9.779 .930 .015 .635 -.013 .729 Stakeholder -51.511 .598 -.047 .155 𝑅2 .291 .294 .052 .079
𝑅2change .291 .003 .052 .028
F 3.596 3.194 .479 .662
Sig. F change .002 .598 .867 .155
No support has been found for propositions 4a and 4b. Regression concerning proposition 4a (table 10), shows model 1’s ability to explain net income’s variance as 29.9% (F = 3.741, Sig. = 0.001), and ROA’s variance as 9.3% (F = 0.897, Sig. = 0.524). Adding the variable value-performance does not significantly improve the models of net income (F = 3.304, Sig. = 0.682) and ROA (F = 0.786, Sig. = 0.971). Therefore, no support has been found for
proposition 4a. In terms of proposition 4b (table 11), the control variables explain 29% of the variance in net income (F = 3.578, Sig. = 0.002) and 7.8% of the variance in ROA (F = 0.743, Sig. = 0.654). Model 2 does nog significantly improve on model 1, for both net income (F = 3.194, Sig. = 0.539) and ROA (F = 0.662, Sig. = 0.763). This means that proposition 4b is not supported. So, firms with a combination of value- and performance-driven motives do not have a higher financial performance than firms with either lone value or performance motives.
Table 10. Results of hierarchical linear regression analysis for proposition 4a
Net Income ROA
Variable Model 1 Model 2 Model 1 Model 2
B Sig. B Sig. B Sig. B Sig.
Industry -55.527 .417 -44.007 .554 .006 .789 .006 .816
Transf. Risk 645.546 .000 644.391 .000 .052 .304 .052 .307
Log Sales 255.311 .004 257.288 .004 .012 .671 .012 .675
Log R&D expenditure 44.056 .498 41.338 .529 .007 .763 .007 .763
Performance-Stakeholder -95.014 .417 -107.090 .378 -.049 .209 -.049 .230 Value-Stakeholder 90.709 .328 84.317 .372 -.028 .370 -.027 .383 Performance 165.348 .338 148.636 .404 .063 .273 .063 .286 Stakeholder -22.358 .779 -33.044 .695 -.035 .188 -.035 .218 Value-Performance -49.891 .682 .001 .971 𝑅2 .299 .301 .093 .093 𝑅2change .299 .002 .093 .000 F 3.741 3.304 .897 .786 Sig. F change .001 .682 .524 .971
Table 11. Results of hierarchical linear regression analysis for proposition 4b
Net Income ROA
Variable Model 1 Model 2 Model 1 Model 2
B Sig. B Sig. B Sig. B Sig.
Log Sales 245.008 .007 252.814 .006 .011 .717 .012 .690
Log R&D expenditure 43.592 .511 41.307 .536 .008 .727 .007 .742
Performance-Stakeholder -104.667 .390 -127.116 .320 -.057 .163 -.060 .157 Value-Stakeholder 85.929 .412 68.917 .526 -.035 .309 -.038 .292 Stakeholder -29.158 .747 -51.511 .598 -.043 .155 -.047 .155 Value 9.219 .931 -9.779 .930 -.010 .783 -.013 .729 Value-Performance -76.311 .539 -.012 .763 𝑅2 .290 .294 .078 .079 𝑅2change .290 .004 .078 .001 F 3.578 3.194 .743 .662 Sig. F change .002 .539 .654 .763
The next proposition to be tested was 5(a/b). This proposition states that firms with a
combination of value- and performance-driven motives have a higher financial performance than firms with either a combination of (a) value- and stakeholder-driven motives or (b) performance- and stakeholder-driven motives. The results from the regression regarding 5a, as presented in table 12, shows that model 1 explains 29.1% of the variance in net income (F = 3.585, Sig. = 0.002) and 8.3% of the variance in ROA (F = 0.797, Sig. = 0.607). Model 2 does not improve the variance explained of either net income (F = 3.204, Sig. = 0.529) or ROA (F = 0.705, Sig. = 0.810). Hence, proposition 5a is not supported. The results of the regression for proposition 5b is shown in table 13. The control variables explain 29.4% of the variance in net income (F = 3.653, Sig. = 0.001) and 7.3% of the variance in ROA (F = 0.686, Sig. = 0.702). The model is not significantly improved by adding the predictor variable value-performance in terms of net income (F = 3.202, Sig. = 0.911) as well as ROA (F = 0.620, Sig. = 0.696). Thus, just as 5a, proposition 5b was not supported by the results.
Table 12. Results of hierarchical linear regression analysis for proposition 5a
Net Income ROA
Variable Model 1 Model 2 Model 1 Model 2
B Sig. B Sig. B Sig. B Sig.
Industry -75.340 .279 -57.982 .439 .012 .597 .010 .689
Transf. Risk 654.633 .000 653.065 .000 .049 .333 .049 .335
Log Sales 252.660 .006 259.240 .005 .013 .661 .012 .685
Log R&D expenditure 49.580 .452 46.036 .488 .005 .826 .005 .811
Performance-Stakeholder -120.710 .307 -140.013 .254 -.041 .295 -.039 .343 Stakeholder -51.197 .537 -69.241 .432 -.026 .341 -.024 .414 Value -27.755 .769 -39.527 .683 .009 .782 .010 .751 Performance 146.031 .398 120.920 .497 .069 .232 .072 .226 Value-Performance -77.627 .529 .010 .810 𝑅2 .291 .295 .083 .084
𝑅2change .291 .004 .083 .001
F 3.585 3.204 .797 .705
Sig. F change .002 .529 .607 .810
Table 13. Results of hierarchical linear regression analysis for proposition 5b
Net Income ROA
Variable Model 1 Model 2 Model 1 Model 2
B Sig. B Sig. B Sig. B Sig.
Industry -45.401 .538 -42.965 .578 .007 .764 .005 .860
Transf. Risk 661.536 .000 662.595 .000 .063 .212 .062 .225
Log Sales 253.332 .006 254.595 .006 .014 .636 .013 .676
Log R&D expenditure 45.944 .486 45.787 .490 .009 .670 .010 .666
Stakeholder 6.879 .937 3.479 .970 -.025 .381 -.021 .487 Value 42.005 .682 39.386 .710 .007 .843 .010 .781 Performance 185.839 .281 181.497 .308 .072 .215 .077 .199 Value-Stakeholder 121.093 .233 118.687 .256 -.018 .591 -.015 .658 Value-Performance -13.673 .911 .016 .696 𝑅2 .294 .295 .073 .075 𝑅2change .294 .000 .073 .002 F 3.653 3.202 .686 .620 Sig. F change .001 .911 .702 .696
Tables 14 and 15, show the results of propositions 6a and 6b, respectively. Firstly, in terms of proposition 6a, the results show that the control variables explain 28.4% of the variance in net income (F = 3.468, Sig. = 0.002) and 5.2% of variance in ROA (F = 0.482, Sig. = 0.865). Model 2 improved neither the variance explained of net income (F = 3.194, Sig. = 0.320) nor of ROA (F = 0.662, Sig. = 0.157). So, proposition 6a was not supported. Neither was
proposition 6b. While model 1 explained 29.5% of net income’s variance (F = 3.655, Sig. = 0.001) and 6.8% of ROA’s variance (F = 0.641, Sig. = 0.741), the inclusion of the predictor variable performance-stakeholder did not significantly improve the model for either net income (F = 3.291, Sig. = 0.455) or ROA (F = 0.636, Sig. = 0.432).
Table 14. Results of hierarchical linear regression analysis for proposition 6a
Net Income ROA
Variable Model 1 Model 2 Model 1 Model 2
B Sig. B Sig. B Sig. B Sig.
Industry -26.766 .723 -29.846 .693 .011 .654 .010 .694
Transf. Risk 675.881 .000 649.966 .000 .067 .186 .055 .283
Log Sales 250.252 .007 252.814 .006 .011 .722 .012 .690
Value 27.031 .797 -9.779 .930 .005 .896 -.013 .729 Value-Stakeholder 103.239 .317 68.917 .526 -.022 .527 -.038 .292 Value-Performance -40.814 .731 -76.311 .539 .004 .912 -.012 .763 Performance-Stakeholder -127.116 .320 -.060 .157 𝑅2 .284 .294 .052 .079 𝑅2change .284 .010 .052 .027 F 3.468 3.194 .482 .662 Sig. F change .002 .320 .865 .157
Table 15. Results of hierarchical linear regression analysis for proposition 6b
Net Income ROA
Variable Model 1 Model 2 Model 1 Model 2
B Sig. B Sig. B Sig. B Sig.
Industry -43.044 .575 -42.208 .583 .005 .844 .005 .836
Transf. Risk 663.161 .000 642.431 .000 .058 .248 .051 .321
Log Sales 255.217 .005 251.353 .006 .009 .766 .007 .802
Log R&D expenditure 46.129 .480 39.068 .555 .007 .734 .005 .823
Value 37.514 .686 25.189 .790 .021 .492 .017 .590 Value-Stakeholder 117.076 .216 103.932 .281 -.005 .864 -.010 .755 Value-Performance -15.174 .894 -31.712 .786 .025 .512 .019 .621 Performance 180.191 .298 164.536 .347 .085 .146 .079 .179 Performance-Stakeholder -88.644 .455 -.031 .432 𝑅2 .295 .300 .068 .077 𝑅2change .295 .006 .068 .008 F 3.655 3.291 .641 .636 Sig. F change .001 .455 .741 .432
The regression results of proposition 7a and b are presented in tables 16 and 17, respectively. The control variable explain 28.1% of the variance in net income (F = 3.423, Sig. = 0.002) and 7.2% of the variance of ROA (F = 0.680, Sig. = 0.707). Model 2 includes the independent variable performance-stakeholder, but does not significantly improve the model for net
income (F = 3.204, Sig. = 0.254) or ROA (F = 0.705, Sig. = 0.343). This entails that no support has been found for proposition 7a. The results regarding proposition 7b show an explained variance for net income of 29.4% (F = 3.653, Sig. = 0.001), and 7.3% for ROA (F = 0.686, Sig. = 0.702). Adding the variable performance-stakeholder, did not significantly improve the variance explained for net income (F = 3.253, Sig. = 0.471) or ROA (F = 0.790, Sig. = 0.383). Therefore, proposition 7b is not supported by the results.
Table 16. Results of hierarchical linear regression analysis for proposition 7a
Net Income ROA
