1
CSR Motives and Firm Performance
To what extent do different motives behind CSR have a different
impact on a firm’s financial and CSR performance?
Suzanne Sweers
Student number: 10460470 Date of submission: 29-6-2015 Academic year 2014-2015 Project 4: Bachelor thesis First supervisor: K. Quintelier Second supervisor: F. Bridoux
2 Statement of originality
This document is written by Suzanne Sweers 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
Abstract
In this research the motives behind the incorporation of CSR practices within firms is being investigated. It is hypothesized that firms with a higher percentage of profit-driven motives will have a higher financial performance in the short run and that, in turn, firms with a value-driven motive will have a lower financial performance in the short run. Furthermore, it is tested whether firms with a value-driven motive have a better CSR performance and an increased financial performance in the long run. After performing linear regressions, the findings do not provide significant evidence that these hypotheses are true. Several explanations of these results are given. This study contributes to existing theory by showing the complexity in the relation between CSR motives and performance of a firm. Managerial implications can be found looking at the influence of stakeholders have on the manager and the firms’ CSR decisions made. The insignificant results of the regressions may be due to several limitations of this study, such as a limited number of firms in the dataset with valid information and lacking information concerning the control variables. Future research should look into the limitations of this research.
4
Contents
Statement of originality………2 Abstract………..………3 1. Introduction………...6 2. Literature review………...82.1 Influence of CSR on Financial performance……….…..…..…...8
2.2 What influences CSR performance………...……….……9
2.3 CSR Motives………..………9
2.3.1 Influence of CSR motives on CSR performance……….….11
2.3.2 Influence of CSR motives on financial performance ………..11
2.4 Conclusion……….………..12
3. Theoretical framework………...………14
3.1 Firms vary in the CSR motives………14
3.2 Influence CSR motives on CSR performance………..………15
3.3 Influence stakeholder culture on financial performance………..…..…16
3.4 Profit-driven motive and stakeholder behaviour………17
3.5 Motives and short term financial performance………...18
4. Methodology………....19
4.1 Data gathering, dependent & independent variable………...19
4.2 Control variables……….…20
5. Results………..22
5.1 Descriptive statistics………...22
5
5.1.2 Motives ……….….…..………22
5.1.3 KLD average……….……….………..24
5.1.4 Net income & ROA (Quarterly)………...…………25
5.1.5 Net income & ROA (Annually)……….………..28
5.2 Correlation………..31 5.3 Regression………...35 6. Discussion……….………43 6.1 Discussion findings……….43 6.2 Theoretical implications……….44 6.3 Managerial implications……….45
6.4 Limitations and suggestions for future research...45
7. Conclusion………...47
Bibliography………48
6 1. Introduction
Corporate Social Responsibility (CSR) is a phrase that is often used in business literature, as it is a very relevant term in today’s business world. According to Yilmaz it can be defined as " business decision-making linked to ethical values, compliance with legal requirements, and respect for people, communities, and the environment" (2013, p.54). Nowadays, nearly every company feels the need to incorporate CSR to show that they care about social and
environmental issue, but they also engage with CSR activities in order to stay competitive (Falkenberg & Brunsael, 2011). This shows that different motives can be detected to
incorporate CSR activities, with different intentions linked to these motivations. Some firms can focus on their financial performance, while other firms may be more concerned about their CSR performance. In this thesis both financial performance and CSR performance will be regarded as dependent variables, and it is tested if the motivations are suitable independent variables.
Much research has already been done concerning CSR, showing the relevance of this subject. However, academics have not been consistent in their findings whether businesses’ performance benefits from CSR practices or not (Cavaco & Crifo, 2014). Or as Barnett states: “After more than thirty years of research, we cannot clearly conclude whether a one-dollar investment in social initiatives returns more or less than one dollar in benefit to the
shareholder.” (2007, p.794). This quotation shows the link between CSR practices and financial performance of a firm. Besides the financial aspect of performance, also the CSR performance of a firm can be investigated (Barnett & Salomon, 2012). However, little attention in academic literature has been given to the different motivations behind practicing CSR and what those differences imply for performance. Barnett (2007) also stated in his research that future research should continue to uncover elements that allow managers to determine whether CSR practices are beneficial. The different motives behind CSR could be one of these elements and will be investigated in this study.
The practical relevance of this paper can be seen in the influence of the different motives behind CSR on performance. Not only can firms (and managers) take this
information into account in deciding what stance to take towards CSR, also it is beneficial for them to know whether a value driven motive that might seem to have a negative influence on the short term, can perhaps be beneficial in the long term. This might change the perception of managers and extend their vision to longer terms.
7
apparel industry are taken as a sample. Derived from their annual reports, their division in motivations behind CSR are established, which is used as an independent variable in a linear regression analysis. The dependent variables are considered to be CSR performance and financial performance, of which data is derived from KLD and Compustat. The results show that no significant results can be obtained from the regressions performed, the reasons behind this will be elaborated upon in the discussion part.
This thesis will continue with a literature review. After this, a theoretical framework will be established, leading to the methodology and the actual analysis of the data. The aim of the linear regressions performed, is to answer the following research question: “To what extent do different motives behind CSR have a different impact on a firm’s financial and CSR performance?”. The results will be elaborated upon in the discussion, before ending the thesis with a conclusion.
8 2. Literature review
2.1 Influence of CSR on Financial performance
Looking at the concept of CSR, many researches have focused on the particular influence CSR has on financial performance. The quotation about the one dollar return from Barnett (2007) in the introduction already touched upon this link, showing that researchers have not find an ultimate agreement on the influence of CSR on financial performance. Aiming to give a clearer view of this relationship and to solve this discussion, Orlitzky, Schmidt, & Rynes (2003) provided a meta-analysis of 52 previous observations on the link between CSR and financial performance. With the 33.878 observations as a total sample size, the findings suggest that there is a positive relation between corporate social responsibility of a company and its financial performance.
A mediating role in the link between CSR activities and financial performance could be given to CSR performance (CSP), as is shown by the research of Yilmaz (2013). He explored the relationship between CSR activities and financial performance and the direction of this relation. His research investigates the influence of CSR activities in the form of CSR performance. From the definition given in his paper, it is clear that CSR performance could be seen as a mediator between CSR activities and financial performance. It is stated that: “ The term corporate social performance reflects the impact of a corporation's activities on society. This embodies the performance of its economic functions and other actions taken to
contribute to the quality of life.”(Porwal, L.S. 1993; In: Yilmaz, 2013). The CSR activities of a firm are reflected in the CSP of that company, showing the mediating role of this measure.
The research by Yilmaz (2013) confirmed the link between CSR activities and financial performance. He found that CSP has a significant relationship with particular financial performance indicators such as the Liquid Assets/Total Assets ratio, Net Operating Profit and Return On Equity. Since CSR can be linked to CSP and CSP in turn can be linked to financial performance, this shows that indeed a valid link between CSR activities and financial performance can be made, with a mediating role of CSP. Also Molteni (2004, In: Patrizia, 2012) confirms that CSP can be seen as a mediator between CSR and financial performance. He indicated that CSR influences economic performance of a firm through intangible resources, such as human, organizational and relational capital. Linked to these intangible resources are (among others) ecological/social business, corporate culture, and motivating and involving employees. These are elements that reflect the CSR performance of a company, establishing the mediating role of CSP.
9
2.2 What influences CSR performance
It was indicated in the last paragraph that CSR performance can be viewed as a mediator between CSR activities and financial performance of a firm. Therefore, it is valuable to understand what influences and drives CSR performance. Vlachos (2012) studied the moderating role of individual traits between CSR performance and the consumer-retailer emotional attachment. He found with his study that variations in consumers’ individual traits may impact the effectiveness of CSR initiatives. A link can be made between this
effectiveness and performance; the more useful the CSR initiatives, the better the CSR performance of a firm. Since personality traits have an impact on effectiveness, it can be argued that personality traits may thus influence the CSR performance of a company.
This study by Vlachos focused on the effect of consumers, which is an external influencer of CSR performance of a firm. When looking at internal elements that could affect the CSR performance of a firm, the employees of a firm can be discussed as one of the elements that influences CSR performance. Story & Neves (2015) look at the attributions of employees towards CSR practices of an organization and its influence on employee
behaviour. They thereby made a distinction between intrinsically and extrinsically motivated attributions. Intrinsic motivation, on one hand, can be defined as coming from within the firm, because they sincerely care about the goal (Story & Neves 2015; Vlachos, 2012). Extrinsic practices on the other hand, are external motivators, such as the satisfaction of shareholders, or other stakeholders (Story & Neves, 2015).
Story & Neves (2015) find, among other results, that employees work with more effort when they perceive that their organization engages with a CSR practice both from an intrinsic and extrinsic motivation. It shows that the perception of the employees on the motivation behind CSR activities are important for a firm. Looking at employees is a more individual approach of the internal elements that can influence CSR performance. However, Story & Neves (2015) did also indicate that companies themselves can also perform from an intrinsic or extrinsic motivation and how that motivation behind CSR can be observed to understand CSR performance This study thus shows that the different motives behind the incorporation of CSR activities is one element that could influence CSR performance.
2.3 CSR Motives
Researchers that have paid some attention to the motivational part behind CSR recognized various drivers that motivate companies. As was indicated in the previous paragraph, Story &
10
Neves (2015) found an intrinsic motivation and an extrinsic motivation. The first entailed motivation from within the firm, the latter would work as a motivator to satisfy external parties, such as stakeholders. Their observation can be linked to Kim’s study (2014). He also looked at motivations behind socially responsible practices and makes a division between a self-serving motive and a society-serving motive, whereby the first motive is more directed at the benefits of society and the second more directed at the benefits of the business. The first motive would represent the intrinsic motive, the latter an extrinsic motive.
Maignan & Ralston (2002) also payed attention in their study to the motives behind CSR and according to them not two, but three motivations can be classified. The motives of Maignan & Ralston overlap with the motivations identified by Kim (2014) and Story & Neves (2015). First of all according to them, the profit-driven motive can be established, whereby CSR is viewed as an instrument to achieve its profitability objectives. This connects with the self-serving motive of Kim (2014), as both are aimed at the benefits of the own firm.
Secondly, Maignan & Ralston (2002) defined the value-driven motive, which suggests that firms are self-motivated to contribute to CSR. Kim (2014) defined this as society-serving motive. Lastly, Maignan & Ralston (2002) identify the motive to conform to stakeholder norms that value the socially responsible practices: the stakeholder-driven motive. Kim (2014) does not take this perspective of separate stakeholders into account and views the self-serving motive as a part of stakeholder communication, thereby not acknowledging that stakeholders might influence the motives of a company. Maignan & Ralston (2002) do identify this possibility of conformation towards stakeholder expectations.
Ellen, Webb & Mohr (2006) do not identify only three motivations, but four motives. They investigated the influence of consumer attributions for corporate socially responsible programs and found that consumers’ attributions reflected many of the motivations that can be identified within companies. According to them the self-centred motive to attribute or
participate in CSR activities can be differentiated into two other motives: a strategic self-centred motive and an egoistic self-self-centred motive. The other-self-centred motive can also be divided: in a value driven and a stakeholder driven motive, adding up to four different motives. This division of motives is similar to Maignan & Ralston (2002) for the other-centred motive, identifying the value-driven and stakeholder-driven motive. However, where Maignan & Ralston (2002) observe a profit-driven motive (or self-serving motive as
identified by Kim (2014)), Ellen et al. (2006) find both a strategic and an egoistic self-centred motive. This extra differentiation was not incorporated in the study by Maignan & Ralston (2002). Perhaps because a strategic self-centred motive is focused at better (strategic)
11
performance of a company, which might ultimately lead to better pay-offs to the individual too, satisfying also the egoistic motive.
2.3.1 Influence of CSR motives on CSR performance
The paragraph showed that many authors have investigated the several motives to incorporate CSR, identifying serval classifications. The incorporation of CSR is implemented to get results from it. This can be in terms of financial performance, or it can be in terms of CSR performance. The CSR performance of a firm is relevant in the current business environment, as can be seen in the fact that CSR rating agencies assess the corporations according to social and environmental performance indicators. This information is checked by not only
companies themselves, but also by stakeholders and broader public (Scalet & Kelly, 2010). However, according to Maignan & Ferrell (2001), it is difficult to measure the
performance outcomes, as the relation between CSR and performance outcomes is correlated and thus complex to assess. However, the study of Lindgreen, Swaen & Johnston (2009) aimed to give a more clear insight of the actual returns that derive from CSR programmes that are implemented by firms. Their study mainly takes into account the stakeholder-driven motivation and society serving motivation, as was elaborated upon in the paragraph above. Using a database of 401 organizations in the US, they performed an empirical investigation, measuring (among others) various performance outcomes. They found that organizations that had more developed CSR, appeared to have better performance in terms of impact on social and economic health and corporate image (Lindgreen et al., 2009). These elements are indicators of CSR performance and thus show that more developed CSR activities within a company, influence the CSR performance.
2.3.2 Influence of CSR motives on financial performance
Several motives can be observed when looking at the motivation behind practicing CSR. The profit-driven motive has been explained in the paragraph about CSR motives. It suggests that companies establish CSR practices to gain profits. According to Swanson (1995), CSR can indeed be viewed as a means to achieve the financial performance objectives of a firm, such as profitability, sales volume or return on investment (In; Maignan & Ralston, 2002). Also Armstrong & Green (2013) state that CSR is an option for firms to enhance profits. They give the example that customers are willing to pay higher prices for products that are more
12
company. The higher prices can lead to higher profits for the firm. This shows the link between a profit-driven motivation and financial performance.
Similarly, Wu & Shen (2013) investigated the motives of banks to engage in CSR and linked it to financial performance. They identify three motives: strategic motive, altruism and greenwashing. The strategic and altruism motive have already been elaborated upon in this research. Greenwashing is defined as an attempt to improve the corporate image, while the business remains relatively unchanged (Frankental, 2001; In: Wu & Shen, 2013). According to them the influence of CSR on financial performance is suggested to be positive for the strategic motive, uncertain for the altruism motive and non-existent for the greenwashing motive. Since they find the dominant presence of the strategic motive, they conclude that financial performance is positively influenced by CSR activities (Wu & Shen, 2013).
The altruism motive as identified by Wu & Shen (2013), overlaps with the value-driven motive of a firm. Wu & Shen (2013) state that for an altruism motive the financial performance is uncertain, or non-negative. Contrasting to this, Barnett & Salomon (2012) find in their research that practising CSR can have a negative financial consequence. They state in their research that as firms take on additional social programs from a value-driven
perspective, they get involved with higher costs and consequently also a lower financial performance. However, Barnett and Salomon (2012) also state that after having invested in social programs, firms might earn financial returns that offset and even exceed the initial costs. They hypothesize in their research that as firms invest more in corporate social programs, they will first lose money, before it neutralizes and evolves into a positive
relationship between expenditures on social programs and financial performance. This shows that several links between the different motives and performance of a firm can be identified.
2.4 Conclusion
This literature review looked at the motives behind the incorporation of CSR and firm performance. Findings of the meta-analysis by Orlitzky et al. (2003) suggest that a positive relation between a CSR practices of a company and its financial performance can be found. A mediator between CSR activities and CSR performance could be CSR performance, as
confirmed by Yilmaz (2013) and Molteni (2004). Elements that in turn could influence CSR performance are indicated to be: the consumers’ individual traits, the perception of employees on the motivation behind CSR activities and the different motives behind incorporation in the firm itself.
13
and extrinsic motivation, which overlaps with the self-serving motive and society-serving motive observed by Kim (2014). Maignan & Ralston (2002) identify three motives: profit-driven, value-driven and stakeholder-driven motive. Ellen et al. (2006) make a division of four motives and divide the self-centred motive in a strategic self-centred motive and egoistic self-centred motive.
The influence of CSR motives on CSR performance is confirmed by the study of Lindgreen et al. (2009). Looking at financial performance, Armstrong & Green find that a profit-driven motive has a positive influence on financial performance. Wu & Chen (2013) find that for motives other than the profit-driven motive, the influence of CSR motives is uncertain. However Barnett & Salomon (2012) find that it can have a negative financial consequence at first, but show that it might evolve in a positive financial performance. The following section will build upon this literature found by constructing a theoretical
14 3. Theoretical Framework
3.1 Firms vary in the CSR motives
In the literature review the different motives behind CSR practices of firms were introduced. To investigate which motives are present in the firm, it useful to look at the stakeholder culture that dominates the company. According to Jones, Felps & Bigley (2007) a stakeholder culture involves the beliefs, values and practices that are linked to the stakeholder
relationships and is a part of the organizational culture. The group of stakeholders of a firm can be very divers and can be made up of customers, employees, shareholders, the
government, members of the community and suppliers of which each can have their own level of influence (Maignan & Ralston, 2002). Due to this heterogeneity, various firms can have a different stakeholder culture with different ideas and motives behind particular investment choices. Jones et al. (2007) identify in their research five different types of stakeholder cultures, each having another level of ethical involvement and concern for others. Within this involvement the different motivations behind CSR are embedded. A company that is more involved and concerned for others for example, might practice CSR from a more value driven perspective.
According to Jones et al. (2007),these are the five stakeholder cultures can be detected: agency, corporate egoist, instrumentalist, moralist and altruist. They ascend in their level of ethical consciousness and rank from individually self-interested to other-regarding type of culture. An agency type of stakeholder culture is the self-interested type, with managerial egoism and no concern for others, as stated by Jones et al. (2007). This can be most closely linked with the profit-driven motivation behind CSR, where the main goal of practising CSR activities is to increase the profitability of the firm. With such a purpose, ethicality is not priority and thus this can be seen as a self-regarding type of stakeholder culture.
The other two types, corporate egoist and instrumentalist both are defined as cultures with limited morality, whereby concern for the shareholders is present, but stakeholders receive less attention. The last two stakeholder cultures are the moralist and altruist. They have the highest level of morality and are involved with all the stakeholders. The latter two can be linked with a driven motive, as defined by Maignan & Ralston (2002). A value-driven motive is an altruistic motivation to practice CSR, therefore it overlaps with those two types of stakeholder culture with the highest level of morality: the moralist and altruist.
15
to the cultures with limited morality. This might seem contradicting, as Jones et al. defines these two types of stakeholder culture (corporate egoist and instrumentalist) as paying
attention to only the shareholders and not the entire group of stakeholders. Therefore, it might not seem compatible with the stakeholder-driven motive. However, this motivation is in both cases imposed by an external group, to satisfy either the shareholders, or the entire group of stakeholders. Whereas the profit-driven motive is mostly self-interested and the value-driven motive is other-regarded, the motivation behind stakeholder-driven is derived neither from egoist purposes nor the from highest level of morality from within the firm itself, This can also be found in the stakeholder culture of limited morality. The stakeholder-driven motivation and stakeholder cultures with limited morality do overlap in their level of morality, but are different in their scope of stakeholder attention.
3.2 Influence CSR motives on CSR performance
The previous paragraph showed that the stakeholder cultures from Jones et al. (2007) with their levels of morality and levels of regarding others can be matched with the different motives as identified by Maignan & Ralston (2002). The presence of a particular CSR motive is therefore expected to influence CSR performance, as the CSR motives are embedded in the stakeholder culture. This type of culture present in a company is important, as managerial decisions are often guided by stakeholder wishes and expectations. Company policies and decisions are results of actions and ideas of top managers (Jones, 1995).
To look at the influence of stakeholder culture, stakeholder theory can be taken into account. Stakeholder theory can be divided in three types of theory (Donaldson & Preston, 1995). First there is the descriptive/empirical type, whereby the theory attempts to describe and explain how firms and the managers within behave. Secondly, Donaldson and Preston (1995) define instrumental stakeholder theory, which focuses on the effect of the behaviour of managers and firms. Lastly, they find the normative stakeholder theory, which takes the moral perspective of firms and managers’ behaviour into account.
According to the instrumental stakeholder theory, CSR can be practiced by relatively large firms to maintain a good relationship with their stakeholders (Donaldson & Preston 1995; Russo & Perrini, 2010). To optimize this relationship, managers feel the obligation to conform to stakeholder expectations and their claims (Russo & Perrini, 2010).Jones et al. (2007) also state that the stakeholders can influence managers. He states that stakeholder culture can be used to understand firm-stakeholder relationship from an ethical point of view. Managers will act upon the expectations from stakeholders, as expressed in the different types
16
of stakeholder cultures. Managerial decision making is thus influenced by the type of
stakeholder culture, which can influence CSR performance directly. Particular decisions can be made in favour of CSR performance, such as the financing of particular projects
concerning CSR, because the manager is affected by the level of morality coming from the stakeholder culture. In this research the influence of value-driven motives on CSR
performance will be tested with the following hypotheses:
H1: Firms that have a higher percentage of value-driven motives as stated in their annual reports will have a higher CSR performance than other firms, in the longer term.
3.3 Influence stakeholder culture on financial performance
As could be read in the paragraph above, stakeholder culture can influence the managers of a firm and the decisions they make in relation with their CSR practices. Bridoux & Stoelhorst (2014) took the instrumental stakeholder theory into account and investigated how a firm approaches its stakeholder management influences its economic value creation. In their research they identify two types of stakeholder motivations to contribute to the value creation: self-regarding and reciprocal stakeholders, whereby the latter values fairness on a higher level than the former. These can represent the different stakeholder culture types as identified by Jones et al. (2007). The different stakeholder cultures can relate to CSR practices and are influenced by the motivations, as was developed in the paragraph above.
In their study Bridoux & Stoelhorst (2014) state that self-regarding stakeholders are usually connected to a more distant type of relationship and that reciprocal stakeholders are more closely involved with the firm. The result of their study showed that for the benefit of the firm it does not matter which type of stakeholder management is applied. Both types of stakeholder motivations (and the management related to that) can be a source financial performance, through the creation of sustained economic value and under the condition that they are supported by suitable organizational practices (Bridoux & Stoelhorst, 2014).
Also Lindgreen et al. (2009) state that the initial benefits achieved by the incorporation of CSR, such as social welfare in the firm, can eventually evolve in profits and sales return in the long run. This initial incorporation of CSR can also occur via stakeholder cultures,
influenced by different motivations, as was developed before. From this theory and its link with financial performance, the next hypothesis can be established:
H2: Firms that have a higher percentage of value-driven motives as stated in their annual reports will have a higher financial performance in the long term.
17
3.4 Profit-driven motive and stakeholder behaviour
The research of Bridoux & Stoelhorst (2014) touched upon the influence of motives and the financial performance of a firm. In the previous paragraph, value-driven motive and the stakeholder culture and the behaviour linked to it was elaborated upon. Since value-driven motive can be linked to stakeholder culture, it is likely that profit-driven motive, as it was identified to be one of the other classified motivations, can also be linked to it.
Motives behind a practice of the individual are criticized on moral basis (Hoffman, Yoeli & Nowak, 2014). This shows that people are sensitive about the intentions behind actions of others. According to Martinko, Harvey, & Dasborough (2011) this is part of the attribution theory, which entails that people look at attributions to determine causes of success or failure. Martinko et al. (2011) state that the attribution theory has much unused potential to illustrate organization behaviour. From a more practical perspective this can be an explanation of how stakeholders perceive the investment in CSR of a firm and react to this. Especially since humans do not only value and judge motives among themselves, as was identified by Hoffman et al. (2014), but also in the interaction with companies. The reason behind this is the fact that individuals tend to react to organizations as if they are human too (Davies, Chun & Silva, 2003; In: Story & Neves, 2015), showing the importance of decisions made by managers.
According Friedman (1970) the main social responsibility of managers is to maximize the firm value to, thereby establishing the classical view of business responsibility (Pava, & Krausz, 1996; Globerman, 2011). Implementing CSR can be an addition to their operations to enhance this financial performance (Maignan & Ralston, 2002). These authors shows that practicing CSR can be implemented from a profit-driven motivation. Such a motivation is likely to lead to increased financial performance. In order to conform to stakeholder and their (moral) expectations towards organizations as if they were human, the particular motives of a firm can be influenced. If the stakeholders expect good financial results, it is likely that managers perform according to a CSR motive that drives financial performance.
From this the following hypothesis was derived:
H3: Firms that have a higher percentage of profit-driven motives as stated in their annual reports will have a higher financial performance in the short term (first quarter after annual report).
18
3.5 Motives and short term financial performance
This hypothesis shows the assumption that firms with a profit-driven motive aim for short term financial benefits for the company. Graham, Harvey & Rajgopal (2005) found with their study that CFO's of a firm are willing to sacrifice long-term firm value in order to achieve the short-term financial objectives, motivated by their own wealth, career and also external reputation. It is assumed that companies will also participate in CSR activities if these are beneficial for the financial position of the firm with a short, direct effect, also because firms then immediately see direct result of their investment. However, on longer term investments are more difficult to predict and managers are thus more hesitant in their investment decisions on longer term (He & Tian, 2013). That is why companies with a profit-driven motive mostly focus on the operations on a short term basis.
According to Brammer & Millington (2008) firms with low corporate social performance have high financial performance in the short run. The low corporate social performance can occur because firms are not paying attention to corporate social programs and thus focusing on another goal. Armstrong & Green (2013) also state that when firms are chasing a profit objective, code of ethics should be implemented that minimize the possibility of irresponsible decisions to take place. This shows that the pursuing of profit might cause goals in the CSR sphere to be neglected.
However, there are not only firms operating from a profit-driven motive, also the value-driven motive can be observed. These companies are not operating from a mainly profit-driven focus, but aim to give back more value to the company, its stakeholders and society than just profit. It was argued in the paragraph above that firms aiming for financial benefits have a better financial performance in the short run. The opposite of this can thus also be argued: firms stepping aside of this profit-driven focus will not have a beneficial financial position in short term. From this the last hypothesis can be derived:
H4: Firms that have a higher percentage of value-driven motives as stated in their annual reports will have a lower financial performance in the short term.
This fourth hypothesis ends the theoretical part of this study. In the following section it will be explained how these four hypotheses will be tested
19 4. Methodology
4.1 Data gathering, dependent & independent variable
In this research 104 firms in the food industry and in the textile and apparel industry will be used for analysis. In order to test the hypotheses stated, the percentage of value-driven motives and the percentage of profit-driven motives are regarded as independent variables. Following the research by Maignan & Ralston (2002), thethree motives of value-driven, profit-driven and stakeholder-driven are identified and categorized, using an elaborated coding scheme, which can be found in the appendix. The information about the motivation of CSR practices in each firm will be checked in their annual reports of the companies in the year of 2011. This year is chosen, as financial data had to be gathered for the dependent variable of which most was only available until 31-12-2012. This way the most recent annual reports can be taken into account, while still being able to link the motives to the financial data one year later. The motives found in the annual reports are based upon particular words/phrases linked to one of the three motivations as established in the coding scheme. Each motivation is labelled and eventually leads to the percentage value-driven, profit-driven or stakeholder-driven of each firm, completing the independent variable of the research.
Data for the dependent variable are gathered, building upon the research Barnett and Salomon (2007). They investigated the relationship between corporate social performance and corporate financial performance and collected data of the CSR performance of the firms from the database Kinder, Lydenberg, and Domni (KLD). It is the largest multidimensional CSP database that is available to the public and is extensively used in academic research (Barnett & Salomon, 2012).Therefore it provides an adequate source for this research. The database provides information about the CSR performance of firms on basis of 13 items that are evaluated with a 0 (negative CSR performance) and 1 (positive CSR performance). By calculating the average of all elements, a result concerning CSR performance is obtained.
For financial performance, besides the use of annual reports, the database Compustat is used, which is accessed through the same channel as KLD. To determine a firm’s financial performance quarterly and annual variables are taken into account, depending on which hypothesis (short-term financial performance and long-term financial performance) is tested. Barnett & Salomon (2012) test financial performance with return on assets (ROA, defined as net income divided by total returns) and Net income (defined as earnings after interest, taxes, depreciation and amortization), as only returns might not give a correct interpretation of the results. Gregory, Tharyan & Whittaker (2014) also state that focusing only on returns to
20
indicate the financial impact of CSR is too limited. Each will be separately tested in a linear regression, since these are two dependent variables that both are impacted differently. Firstly, Net income is assumed to be influenced by how consumers value CSR in the industry. This is because the information regarding CSR influences the consumer's attitude towards the brand and thus on their purchase intentions (Schuler & Cording, 2006). A positive attitude might lead to more purchases, thus higher revenue and ultimately to a higher net income. Secondly, the ROA of a company is assumed to be influenced via stakeholders, as stakeholders can collectively influence the financial performance of the firm by deciding to invest in the firm or not (Neville, Bell & Mengüç, 2005). This impacts the equity and thus the total assets of the company, showing that stakeholders can influence the ROA of a company
The different independent and dependent variables will be used in SPSS to perform a linear regression analysis, testing the four hypotheses. For the hypotheses related to financial performance this thus means two regressions per hypothesis, one testing the hypothesis on Net income and the other testing on ROA.
4.2 Control variables
In order to test the relative influence of the independent variable, the motive of the firm, control variables are added to the regression in this study. A firm’s level of CSR provision depends on its size, level of diversification, advertising, the percentage of government sales, research and development (R&D), consumer income, labour-market conditions and the stage in the industry life cycle they are in (McWilliams & Siegel, 2001). It is not possible to take into account all these elements in this research, as the list is too
elaborate. Following the research by Barnett & Salomon (2007) on the influence of corporate social responsibility performance and financial performance, in this study there will be controlled for: firm size, risk, advertising and R&D expenditure. It is logical to take into account their control variables, as a CSR motive is directly linked to the CSR performance of a firm, as can be derived from the literature review.
Firm size can be controlled for by adding the amount of employees present in the company measured in thousands. It is important to control for this, as larger organizations have more resources to implement CSR activities than smaller organizations (Lindgreen et al., 2009). According to Orlitzky & Benjamin (2001) risk is also an element that influences CSR. They showed with their meta-analysis that prior financial risk is negatively related to CSR performance of a firm. In order to control for this influence, the debt ratio was added as control variable. Debt ratio (Total Debt/Total Assets) is a common measure used to assess the
21
financial risk of a company and to control for CSR influence (Orlitzky & Benjamin, 2001; Barnett & Salomon, 2007).
Also the technological capabilities of a firm and marketing efforts of a firm can be a source of value creation, which might give a troubled view on the influence of CSR motives behind activities on the performance of a firm (Cavaco & Crifo, 2014;McWilliams & Siegel, 2001). Therefore advertising intensity (advertising expenditures/sales) and R&D intensity (R&D expenditure/sales) will be used as control variables in this research, following the research of Barnett & Salomon (2007). Data for the control variables will be extracted from the annual reports. Industry will also be controlled for, as this study will take into account both the food & beverage industry and the apparel & textile industry. Food & beverage industry will be labelled as 0, apparel & textile industry as 1.
Lastly, another control variable will be added in relation to the CSR motive. When a firm has a higher percentage of value-driven motive, the profit-driven motive percentage of the firm will be added as control variable, as they are dependent on each other. For a higher percentage of profit-driven motive, the value-driven motive percentage will serve as control variable.
22 5. Results
5.1 Descriptive statistics
5.1.1 Sample characteristics
For this research companies were gathered from the food & beverage industry and the textile & apparel industry. These industries were chosen, as they are relevant for CSR practices (Maloni & Brown, 2006). One criteria that was taken into account when looking for
companies was the size of the company, as bigger companies have more shareholders to relate to, and give information to (source). Several companies from the sample thus come from the SP500. In the end 134 firms were gathered, of which 66 operate in the food & beverage industry and 68 in the textile & apparel industry. However, when looking for annual reports of 2011 and 2012 of these companies, several firms had to be removed from the list. Either because the annual report could not be found, or because a validation code had to be
used/payment had to be done in order to access the report. Other firms had to be taken out of the sample, because their company was not present in the data base (KLD or Compustat) used. In the end, the sample consisted of 103 companies. Of these companies 57 are present in the food & beverage industry and 47 in the textile & apparel industry, which is a quite equal divide of 54% versus 46%, creating a sample that is not unevenly divided. Furthermore, the sample of 103 is above the statistical norm of 100, contributing to the validity of this research.
5.1.2 Motives
As was elaborated in the methodology, the coding of the companies motivations as reported in their annual report of the year 2011 was done using a coding scheme, based on a screening of 20 annual reports. The coding scheme that was used can be found in the appendix. Of all companies the motives in the annual reports were labelled. In the end there were 4 companies that reported no motivations behind CSR in their annual reports at all, creating a minimum of 0 motives. The highest amount of calculated motives that was present in the annual report was 61, which appeared twice in the sample. Of all motives, 30 of the 103 companies had 5 or less total motives. This can also be seen in the fact that smallest mode found, is reported at 3. The median was stated at 9 motives. In the appendix an overview of the companies and their accompanying division of motives can be found. The histogram below gives an overview of the division of the motives, with a mean of 12.33 and a standard deviation of 12.321. Also the outlier of 61 motives is clearly visible there.
23 Figure 1: Histogram and normal distribution on motives
The independent variables used in this research are the percentage profit-driven and the percentage value-driven a company is. For the profit-driven motives the mean is a percentage of 22.8, with a standard deviation of 21.08. The minimum that occurred is 0%, the maximum 75%, showing that the companies with the highest value of profit-driven motives had 75% of their motives in the profit-driven category. The skewness of the data is 0.466, with a kurtosis of -1.035. Since these values are around 1, this means that the distribution is rather normally divided. However, this is not reflected in the significance of the Kolmogorov – Smirnov test, which tests for the normality of the data. With a significance of 0.000 (p-value <0.001) it shows that the data is not normally divided. This is because the null-hypothesis states that the sample is not drawn from a normal distribution, which is accepted with the significant value of 0.000. After removing the outliers, the skewness and kurtosis of the data had improved, but the Kolmogorov – Smirnov test still remained significant. Even after transforming the dataset, the data still was not normally distributed according to the Kolmogorov – Smirnov test and the transforming of data also had little effect on improving the skewness and kurtosis.
For the value-driven motive the mean was stated to be a percentage of 35.1 with a standard deviation of 25.2. The minimum again was 0% and the maximum 100%, which shows that in this dataset at least one company was fully dedicated to the value-driven motivation. The skewness is 0.484 and the kurtosis is -0.121. While these values are close to the absolute value of 1, the significant outcome of the Kolmogorov – Smirnov test (p-value =
24
0.000), means that it can again be concluded that the data is not normally distributed. Therefore, it was at first decided to take the outliers out of the data set. However, the Kolmogorov - Smirnov test still reported to have a value of 0.000 after this removal. After transforming the data, again a p-value of 0.000 was found, showing that the dataset was still not normally distributed. Since the removal of outliers and transformation of data did not have an effect on the normal distribution of the data set and the skewness and kurtosis (values -0.850 and -0.342 after transforming the value-driven motives respectively), these outcomes for both value-driven and profit-driven motives suggest to continue the regressions with the original data.
5.1.2 KLD average
To see if the motives reported in the annual reports of the companies in the year 2011 had influence on CSR performance, a time lag was implemented of one year (Barnett & Salomon, 2012). This meant that in KLD data was extracted from the year 2012. Using the ticker codes of companies, information about their CSR performance based on several elements, such as community, corporate governance, environment and human rights, was gathered. The number of strengths and number of concerns were not included as these are accumulated items of the rest of the elements. It was decided not to include the controversial section (military, tobacco, firearms, nuclear power, gambling), because this is not very relevant in the industries of food & beverage industry and the textile & apparel industry. Including them might give a distorted image of the CSR performance of a company. Of the sections that were included all boxes were ticked with data available in 2012 to give a complete overview.
The data set that started with was 103 companies, in the appendix an overview of all companies and their KLD score can be found.However, some companies were present in KLD when searched for, but did not give data when retrieving information about the CSR performance in 2012. Of the 93 companies 24 had a reported CSR performance of 0 average, the minimum in the dataset. It indicates that they scored a 0 on every element selected, thus have a negative CSR performance. Not surprisingly, the mode of the dataset was 0.00. The highest CSR performance score, the maximum, that was obtained was a 0.4 average, which appeared only once. The median was set at 0.0404. In the figure below, which gives an overview of the division of CSR performance as measured between 0 and 1, it can be seen that the measure of 0.4 is an outlier within the dataset. The dataset is divided with a mean of 0.077 and a standard deviation of 0.093.
25 Figure 2: histogram and normal distribution on CSR performance
The skewness of the KLD averages found was stated at 1.515, which can also be seen looking at the normal curve in the figure above. It shows a right tail, which means there is a positive skewness present. The kurtosis is found to be 1.798, meaning that the distribution is not flat. Since these values are around 2, it could be assumed that the data is normally distributed. However, when performing the Kolmogorov – Smirnov test, it shows a significance of 0.000, indicating a dataset that is not normally distributed. Also after taking out the outliers, the p-value remained 0.000. A transformation to the data was performed in order to see whether this influenced the normal distribution of the data. However, this was not the case, as the outcome of the Kolmogorov – Smirnov test remained significant. The value after transformation of skewness is 0.362 and the value of the kurtosis became -0.780. Since these values are closer to a normal distribution, it suggests to take the transformed values of the KLD average in the regression analysis.
5.1.3 Net income & ROA (Quarterly)
To test for the financial performance of the firm, Net income and ROA are taken into account, as was elaborated upon in the methodology. Data was retrieved at first from Compustat, but since a lot of data was found to be missing, annual reports of the fiscal year 2012 were used in addition to complement the dataset and find the ROA and Net income of the 103 companies.
26
and Net income of the first quarter of fiscal year 2012 was taken into account, taken from the Compustat quarterly announcements and the financial reports in 10-Q form, which provide quarterly financial information. It was decided to look at the first quarter of 2012, taking 3 months after the annual report of 2011 as short term financial performance. For one company (Unilever), data on the first quarter of fiscal year 2012 was not found in the 10-Q reports, Compustat and other online available reports.
Net income
When looking at the net income in the dataset, 12 companies were found to have a negative income in the first quarter, of which the minimum set at a Net loss of -215 million. The maximum Net income was set at 2220 million and the mode at a Net income of 15 million. The median of the dataset on quarterly results on Net income was set at 37,9 million. In the figure below an overview of the division of Net income in the first quarter of fiscal year 2012 is presented, with a mean of 128,8 million and a standard deviation of 312, 4 million. In the figure also the outlier of 2220 million can be observed. This is the company of Anheuser- Busch INBEV, but as their total assets are also high, this had no consequences for their ROA to turn into an outlier.
27
The skewness and kurtosis of the data are 4.170 and 21.722 respectively. These are high values (>2), indicating that the sample is not normally distributed. This is also shown in the Kolmogorov – Smirnov test, as the value of p is 0.000, thereby supporting the null-hypothesis of having a non-normally distributed sample. The outliers were removed from the sample to see whether this would improve the dataset. This was not the case and thus the data was transformed. The skewness after transformation was 3.827 and the kurtosis 23.312. These high values indicate a not normally distributed dataset, which is also reflected in the
significance of the performed Kolmogorov – Smirnov test with the transformed data. Since the values of the skewness and kurtosis did not improve significantly by transforming the data, the regression might better be performed with the original dataset.
ROA
To test the short term financial performance, ROA was also checked, as an addition to Net income. The reasoning behind this was given in the methodology. In the dataset used, the minimum ROA was set at -0.72. This is the firm New York & Company, Inc., which had a quarterly loss of -210 with limited assets, resulting in a high negative ROA. The maximum of ROA found was 0.07. Furthermore, the median was reported to be 0.0185. The figure below shows the division of ROA on quarterly financial results, with a mean of 0.01 and a standard deviation of 0.076. In the figure the minimum value of -0.72 is visible as an outlier.
28
The skewness and kurtosis of this data set were reported to be -8.850 and 84.970. The
corresponding outcome of an significance of the Kolmogorov – Smirnov test, shows that this dataset is not normally distributed. The removal of the outliers did not change this outcome. Therefore, it was decided to transform the dataset. After this transformation the Kolmogorov – Smirnov test gave a value of 0.200, indicating the rejection of the null-hypothesis and thus a normal distribution of the dataset. Also the skewness and kurtosis of this dataset reflect this normal distribution, with values of 0.212 and -0.523 respectively. Therefore this outcome suggests to continue working with the transformed dataset of ROA of the first quarter of fiscal year 2012.
5.1.4 net income & ROA (Annually)
To assess the long term financial performance the net income and ROA of the firms were calculated. The long term was set on one year after the annual reports were published on the fiscal year of 2011, in which motivations were stated. As the research focused on the long-term financial effect of these motivations, it was decided to gather data for the end of fiscal year 2012. From all companies the data needed was found, either in Compustat or in the annual reports of fiscal year end 2012.
Net income
In the dataset 9 companies had a Net loss at the end of fiscal year 2012, of which the
minimum found in the data set was a loss of 405 million. The maximum Net income reported was 7243 million and the median reported was 164.68 million. The figure below shows the division of Net income annually, with a mean of 602,8 million and a standard deviation of 1222.9 million. This high standard deviation shows that Net income is not steadily divided in the sample, as also can be seen in the figure itself, where a few observations occur outside of the normal distribution.
29 Figure 5: histogram & normal distribution on net income (annual)
The figure above shows that a right tail can be observed in the dataset. This is supported by the positive value of 3.649 for skewness. The value of kurtosis was reported to be 14.487, indicating a non-normally distributed dataset. This is also seen when looking at the
Kolmogorov – Smirnov test, with a significant result for the null-hypothesis it shows that the dataset is not normally distributed. After the outliers were removed, the skewness and kurtosis values improved, but the testing for normality still showed a dataset that was not normally distributed. A transformation was performed on the dataset, after which the Kolmogorov – Smirnov test still reported a significant value of 0.000. However, the skewness and kurtosis did improve when comparing it with the original dataset, reporting values of 0.836 and 0.109 respectively. Therefore, the transformed values of the Net income on annual basis is perhaps a better option to use in the regression analysis.
ROA
Lastly, the ROA on fiscal year end of 2012 was also calculated. With a minimum of -0.60 of a ROA and a maximum of 0.50, also this financial variable has a rather broad range. This minimum value was accounted for the company of Chiquita Brands International., which also had the minimum value of Net income of 2012, as reported above. The median in the sample was set at 0.0783. The figure below gives an overview of the division of annual ROA of the sample, with a mean and standard deviation of 0.085 and 0.113 respectively. The broad range is reflected in the value of the standard deviation, which is quite high.
30 Figure 6: histogram & normal distribution on ROA (annual)
Also the skewness and kurtosis were checked for this dataset, as their values illustrate whether the sample is normally distributed or not. The values reported were -1,483 for skewness and 14.726 for kurtosis, indicating a not normally distributed dataset. This was also supported by the outcome of the Kolmogorov – Smirnov test, with a significance value of 0.000. In order to check whether the dataset could become normally distributed, the outliers were removed. However, this removal and also the transformation of data did not improve the value of the skewness and kurtosis, nor did it change the outcome of the Kolmogorov – Smirnov test.
Because only three out of the seven variables tested for their normality showed an improvement when they were transformed, it was decided to perform the regressions with the original data. While this meant that for three values (KLD average, ROA (Q1) and Net income (annual)) the normal distribution within the dataset could have been better, this does mean that for the bigger part of the variables (four out of seven) the original data provides a better presentation of the data with better skewness and kurtosis than the transformed variables. In order to perform a regression that is not blemished by the transformations of data, original data is used. Before performing a regression, the correlations of this original data were tested. The next part of this thesis will elaborate on this.
31
5.2 Correlation
The Pearson Correlation was used to test the linear dependence between the variables used in this research. The correlation between variables indicates the relatedness of two variables whereby a value of 1 means positive correlation and -1 negative correlation. All variables were used in testing the correlation, of which an overview can be seen in table 1 below. Significant relations are notated with one asterisk (*) for significance at 0.05 level and with two asterisks (**) at the 0.01 level.
The first significant relationship can be found between profit-driven motives and industry, positively influencing each other (Pearson correlation coefficient = .347, N=99 , Sign. = .000). Also between industry and value-driven motives an significant relation can be found, in this case a negative relationship (Pearson correlation coefficient = -.339, N=99, Sign =.001). While for both motives a significant relation with industry can be found, this is
different for the type of motive detected. The category of profit-driven motives has a positive relation with industry, while value-driven motives have a negative relation with the industry. Not surprisingly the variables of value-driven motive and profit-driven motives are also significantly related (Pearson correlation coefficient = -.257, N=99, Sign.=.010). A negative relation indicates that the higher the profit-driven motives, the lower the value-driven motives (or the other way around) which is logical in this case. Firms normally have either a value-driven motive of profit-value-driven motive when taking a decision.
Another significant relation was found between the variables of R&D intensity and firm size. With a positive relation (Pearson correlation coefficient=.375, N=103, Sign.=.000) this indicates that the higher the firm size, the higher the R&D expenditure of a firm. This is a logical reasoning, as bigger firms attach more value to R&D than smaller companies and have more resources to support this. Also a positive and significant relation was found between advertising intensity and debt ratio (Pearson correlation coefficient=.296, N=103,
Sign.=.002). A firm with higher advertising intensity might thus indicate a higher expenditures, which are financed by higher debts.
A significant relation was found between KLD average, indicating the CSR performance of a firm and Industry (Pearson correlation coefficient=-.350, N=94,
Sign.=.001), Size of company (Pearson correlation coefficient=.507, N=94, Sign=.000) and also for R&D expenditure (Pearson correlation coefficient=.546, N=94, Sign=.000). These results may indicate that KLD average of a firm depends on what industry a firm is present in and that the bigger the company and the higher the R&D expenditure, the higher the CSR
32
performance. The last relationship (R&D intensity and CSR performance) can be explained by the positive correlation present between the two variables of R&D intensity and firm size. A company with higher R&D intensity is suggested to be a bigger company and thus might have more resources to spend on their CSR performance.
Furthermore, net income annually and quarterly both have a significant positive relation KLD average (Pearson correlation coefficient = .636 and .736 respectively, N =94 and 93 respectively and Sign.=.000 for both). This could indicate that the higher the KLD average and thus CSR performance of a firm, the higher the net income of the firm. It can be linked to the hypothesis two, that hypothesizes that firms with a higher percentage of value-driven motives have a better financial performance in the long run. Firms that have more value-driven motives can be linked to better CSR performance (Lindgreen et al., 2009), as was elaborated upon in the literature review.
Besides that, also a positive and significant relationship was found for both size (Pearson correlation coefficient=.754, N=103, Sign.=.000) and R&D intensity (Pearson correlation coefficient=.582, N=103, Sign.=.000) with net income annually. The correlation for net income annually and size of the firm is quite high (>0.7), indicating that there is a strong linear relationship between the two. This could mean that the bigger the firm, the higher annual net income. For R&D intensity there is a moderate strength and again this can be explained by the fact that there is also a correlation between R&D intensity and firm size. Firms with bigger size have more R&D expenditures and also have higher net income, creating a correlation between R&D intensity and net income. It can also be explained by the fact that the higher the R&D intensity of a firm, the more innovative they are, creating a higher net income for the company. The same explanations can be given for the significant and positive relations found for net income quarterly and firm size (Pearson correlation coefficient=.665, N=102, Sign.=.000) and R&D expenditure (Pearson correlation coefficient=.495, N=102, Sign.=.000).
The last significant relations observed are between net income annually and net income quarterly (Pearson correlation coefficient=.963, N=102, Sign.=.000) and ROA annually and ROA quarterly (Pearson correlation coefficient=.239, N=102, Sign.=.016). These correlations are not surprisingly, as these are the same financial measures and the quarterly results are expected to have an impact on the annual results of that particular
measure.
Since for profit-driven motives and value-driven motives little correlation was found with the variables (beside each other and industry) this might be an indication that these
33
independent variables are not much related to the dependent variables. This assumes a limited predictability of the motives on the dependent variable, which might influence the regressions that will be performed.
34 Table 1 – Correlations Variable 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 1. Industry 1 2. Profit % .347** 1 3. Value % -.339** -.257* 1 4. Size -.148 -.042 .049 1 5. Debt Ratio -.121 -.136 .187 .057 1 6. R&D intensity -.161 .097 .004 .375** .090 1 7. Advertising intensity -.099 -.089 .007 -.141 .296** .096 1 8. KLD average -.350**-.109 .137 .507** .037 .546** .177 1
9. Net income (annual) -.176 .041 .084 .754** .052 .582** -.049 .636** 1
10. ROA (annual) .124 .143 .040 .074 -.002 .047 .125 -.002 .157 1
11. Net income (Q1) -.181 -.048 .120 .665** .039 .495** -.041 .0736** 0.963** .093 1
12. ROA (Q1) -.097 -.096 .012 .078 -.058 .053 .037 .061 .100 .239** .169 1
35
5.3 Regression
After analysis of the sample and the data used in the research, an overview of the regressions performed will be given. Four hypotheses were constructed through the theoretical
framework, that were tested using a linear regression. The first hypothesis was as follows:
H1: Firms that have a higher percentage of value-driven motives as stated in their annual reports will have a higher CSR performance than other firms, in the longer term.
The independent variable was the percentage of value-driven motives present in the annual reports and the dependent variable was the CSR performance of the firms, which was the average of data found in the KLD database. The control variables that were used were: industry size, debt ratio, R&D intensity, advertising intensity and the percentage of profit-driven values.
The table below gives an overview of the variables used. It shows that the control variables of industry, size and R&D intensity are significant in the model, with p-values of .020, .000 and .000 respectively. For these variables the null-hypothesis that the coefficient has no effect, can be rejected, thus showing that these are valid predictor variables. For the independent variable of value-driven motive this is different. With a significance value of .436 (>0.05), it can be stated that the percentage of value-driven motives does not
significantly impact the CSR performance of a firm and therefore hypothesis 1 is not
supported. Since the F change also does not show a significant result when the independent
variable is added, it means that the addition did not significantly improve the prediction. Thus adding the variable of percentage of value-driven motives, besides the control variables, does not add to the prediction of a better CSR performance. The R Squared of this regression is 0.506, which indicates a moderate goodness of fit. This means that a moderate proportion of the variance is accounted for by the model with the independent variable.
Table 2 - linear regression on the influence of value-driven motives on KLD average
Variable Model 1 Model 2
B Sign. B Sign.
36 Industry -.043 .008 -.040 .020 Size .001 .000 .000 .000 Debt ratio -.027 .152 -.030 .123 R&D intensity 3.066 .000 3.080 .000 Advertising intensity .473 .039 .487 .034 Profit-driven motive % .074 .466 .000 .515 Value-driven motive % .000 .436 R Squared .502 .506 F Change 13.967 .613 Sign. F Change .000 .436
It is also important to take the tolerance of the variable into account, as it checks for multicollinearity. The tolerance of this independent variable has a value of .854. Since the tolerance value is higher than .2 (and VIF smaller than 5, being 1.18), it shows that the independent variables (including the control variables) are independent from each other. This means that the independent variable of percentage of value-driven motives is not superfluous for the outcome of the regression.
The second hypothesis that was tested for, was the following:
H2: Firms that have a higher percentage of value-driven motives as stated in their annual reports will have a higher financial performance in the long term.
At first, the long term financial performance was measured using net income. The table below shows more information concerning the coefficients. It shows that the control variables of size and R&D intensity have significant outcome, indicating that they have a predictive value concerning the dependent variable. The result of the regression shows that the influence of a higher percentage of value-driven motives does not have significant positive influence on financial performance in the long term, measured with net income. The value of p = .349, which is higher than 0.05.This is confirmed by the significant F change, which is not significant for the addition of the independent variable with a value of .349. The control
37
variables of size and R&D intensity show to have a significant effect (at 0.01 level) on Net income annually, indicating that they are valid predictor variables. The R squared found was .677, which is quite high, showing that a large part of the variability in de dependent variable is explained by the independent variable.
Table 3 - linear regression on the influence of value-driven motives on Net income annually
Variable Model 1 Model 2
B Sign. B Sign. (Constant) 63.691 -65.530 Industry -125.129 .443 -81.433 .630 Size 13.335 .000 13.348 .000 Debt ratio -50.793 .797 -81.433 .685 R&D intensity 37722.479 .000 37929.688 .000 Advertising intensity 293.579 .899 470.356 .839 Profit-driven motive % 3.007 .435 3.516 .366 Value-driven motive % 3.027 .349 R Squared .673 .677 F Change 31.615 .886 Sign. F Change .000 .349
Secondly, the long term financial performance was measured using the ROA for fiscal year end 2012. The results of the linear regression show are shown below in table 4. With a p-value of .245 it can be stated that the p-value-driven motives of a firm to incorporate CSR performance, does not influence the long term financial performance significantly, as
measured for ROA. This is confirmed by the insignificant value of the F change (.245). The R Squared has a value of .072, which is a very low value, indicating that variables are limited related to each other. Taking together the results of the regression of value-driven motives on both net income and ROA annually, it can be stated that, with the not significant outcomes (p=.349 and p=.245), hypothesis 2 is not supported.
