The Influence of Authority on the Valuation
of Corporate Social Responsibility
Initiatives
Alexander de Groot
10078797
Amsterdam, July 15, 2014
Bachelor’s Thesis Supervisor: Dr. K.J.P. Quintelier Business Studies 2013-2014Foreword
This thesis was written for my Bachelor degree in Economics and Business, specialization Business Studies, at the University of Amsterdam. I would like to thank some people who helped me during the preparation and writing of my thesis. First of all, I want to thank my supervisor K. Quintelier for her guidance during this period. She was a very helpful supervisor, who was able to trigger me not just to be satisfied with any result. Furthermore, she helped me find useful articles, collect data and was able to get in contact with a professional translator who translated our survey to Castilian. All her dedication to my thesis was quite inspiring and I will benefit from this during the rest of my studies. Secondly, I would like to thank David Estape for distributing the survey among his employees, and for his hospitality during my short stay at the factory near Barcelona, Spain. Finally, I would like to thank Joshua de Groot. During the last few months we spent a lot of time together at the library. Not only studying for our last Bachelor degree exams, but also discussing and exchanging ideas on the thesis topics we were assigned.
Thanks, and enjoy reading my thesis!
Alexander de Groot
Table of Contents
Foreword ... 2
Introduction ... 5
2 Literature review ... 8
2.1 Stakeholder reactions towards corporate social responsibility ... 8
2.2 Moral orientation ... 9
2.3 Authority ... 10
2.4 Types of corporate social responsibility initiatives ... 12
2.4.1 Community ... 12 2.4.2 Diversity ... 12 2.4.3 Employment ... 13 2.4.4 Environment ... 13 2.4.5 Human rights ... 13 2.4.6 Products ... 13 2.4.7 Alcohol ... 13 2.4.8 Gambling ... 14 2.4.9 Firearms ... 14 2.4.10 Military ... 14 2.4.11 Nuclear ... 15 2.4.12 Tobacco ... 15 2.4.13 Corporate Governance ... 15
2.5 Use of moral orientation as Mediator ... 17
2.5 Visual representation of the conceptual framework ... 17
2.5.1 Mediation model 1 ... 17
2.5.2 Mediation model 2 ... 18
3 Methodology ... 19
3.1 Research design ... 19
3.2 Measurement of authority ... 21
3.2.1 Formal hierarchical position ... 21
3.2.2 Sanctioning authority ... 21 3.2.3 Decision-making authority ... 21 3.3. Measurement of CSR valuation ... 22 3.2 Sample ... 24 3.3 Analyses ... 25 4 Results ... 28 4.1 Reliability ... 28 4.1.1 Moral orientation ... 28 4.1.2 CSR initiatives ... 29 4.2 Descriptive statistics ... 30
4.3 Model estimation and interpretation ... 31
4.3.1 Mediation model 1 ... 31
4.3.2 Mediation model 2 ... 33
5 Discussion ... 35
5.1 Key findings ... 35
5.2 Theoretical implications ... 37
5.3 Managerial implications ... 37
5.4 Limitations and suggestions for further research ... 38
6 Conclusion ... 40
Bibliography ... 41
Appendix I ... 44
Abstract
The aim of this study is to investigate the influence of authority on the valuation of corporate social responsibility initiatives (CSR), mediated by moral orientations. The data were collected by self-administrated questionnaires from 89 employees working at two different organizations. In contrast to what was expected, authority does not seem to have a direct effect on the valuation of CSR initiatives. Moreover, the findings of this study do not indicate that moral orientations mediate the relationship between authority and the valuation of CSR. However, the findings of this study indicate that certain moral orientations tend to have an effect on the valuation of CSR initiatives. These results indicate that further research into moral psychology in a business context is needed to be able to provide useful managerial implications.
Introduction
Improving stakeholder relationships is an important reason for organizations to invest in corporate social responsibility (CSR) (Aguinis & Glavas, 2012). However, there is no straightforward relationship between more socially responsible firm behavior and stakeholder reactions toward CSR. On the one hand, Turban and Greening (1997) state that firms with a stronger (perceived) corporate social performance are more attractive to prospective employees and they have a lower turnover among employees. Not only does a stronger social performance have influence on employees, it also has a positive impact on consumer’s buying behavior (Mohr, Webb, & Harris, 2001; Sen & Bhattacharya, 2001). On the other hand, CSR can actually reduce customer
satisfaction levels, and this can result in harming the market value of a firm (Luo & Bhattacharya, 2006). Studies find that this relationship is moderated by stakeholders’ personal values. Every stakeholder group has different characteristics. For instance, shareholders attach value to increasing dividends and information duty, while suppliers put emphasis on long-term relationship between the firm and supplier (Freeman, 1984). A company should therefore not only focus their socially responsible initiatives on one stakeholder group, as this could result in negative perceptions from other stakeholder groups (Öberseder, Schlegelmilch, & Murphy, 2013). One stakeholder characteristic that is likely to influence CSR evaluations is the kind of moral issues that are deemed important by the stakeholder. This has not yet been investigated. Moral orientations can be defined as: the extent to which specific moral issues are deemed important. On the one hand, people could find equality between individuals important. On the other hand, people could find a certain
hierarchy and cultural traditions important. Because of this, moral orientations might influence how people react to CSR initiatives.
According to Jost (2006), moral orientations of an individual predict a lot of outcomes. For example, voting can be closely linked to moral values but also religion can be connected to personal moral orientation (Graham, Haidt, & Nosek, 2009). This is relevant in the context of CSR because different CSR issues concern different kind of moral issues. For instance, while diversity concerns equality, hierarchy in a firm concerns authority and respect. If companies were able to get more insight in the moral orientations of their stakeholders, it might be possible for them to maximize the
value of their chosen corporate social initiatives. For instance, according to Winterich, Zhang, and Mittal (2012), the moral orientation of a charity and of the stakeholder play an important role in the reaction of people’s willingness to donate more or less money to the specific charity. In addition, following Graham et al. (2011), one could predict that stakeholders with certain moral orientation will react more positively than stakeholders with other moral orientation towards firms that consciously do not invest in the alcohol industry. However, there has not been much research into moral
orientations within a business context. As there is not much known about individual differences in moral orientations and how this influences individual stakeholder’s reaction to the firms’ CSR initiatives. This thesis will help to fill this void.
Moreover, studies have mainly focused on external stakeholders. Internal stakeholders are also important for the performance of the firm. In addition, just like other people, internal stakeholders are likely to differ in their moral orientations. Moral orientations have been found to vary across age, gender, and socioeconomic status (Graham et al., 2013). One factor that can influence one’s moral orientation is an employee’s authority. Studies suggest that people with more authority tend to value the hierarchy within a firm. Thus, it is important to investigate the moral orientations of employees, as resulting from their level of authority, and how these influence the employees’ reactions towards specific socially responsible activities.
In order to fill this gap in the literature and provide more information about the role of moral psychology in the valuation of CSR initiatives, this study will examine how individual differences in employees’ moral orientation influences valuation of specific CSR activities. This study investigates individual differences arising from the level of authority an employee has. Therefore, the main research question is: How does the level of authority an employee has play a role in the valuation of a variety of corporate social responsibility initiatives?
To provide an answer to these two questions a survey will be distributed among employees working at two different firms. With the results of the survey an overview will be given on the extent to which moral orientation plays a role in the valuation of CSR activities. Not only will these findings fill a gap in the literature, it may enable managers to predict the reception of specific CSR initiatives among employees with different authority ranks within the organization.
This thesis will begin by reviewing the relevant existing literature on the topic. By doing this, an extensive overview will be given on the three constructs used throughout the thesis: authority, moral orientation, and CSR initiatives. At the end of the literature review a visual presentation of the hypotheses will be given. In the third section the research design and methodology will be discussed. This part will explain and justify the methods used for answering the research questions. After the method has been explained, the data collection will be introduced and this will be followed by the analysis and results. The discussion section will try to create a relationship between the results and the research question. Practical and managerial implications will be provided on which type of corporate social responsible initiatives firms should engage in. The last paragraph will consist of the possible limitations of the research. Opportunities for future research will also be considered.
2 Literature review
After having introduced the relevance of the topic and having stated the main research question in the previous section, this section will address the related literature and present a framework on which this research is based. An appropriate definition of CSR will be given and used throughout the research. Secondly, the two major moral orientations known in the literature will be discussed and related to the level of authority within a firm. Also this literature review will introduce the six hypotheses that are tested. Finally, a visual representation of the framework will be given.
2.1 Stakeholder reactions towards corporate social responsibility
As mentioned in the introduction, improving relationships with stakeholders is becoming more important for firms by engaging in a variety of CSR initiatives. But what exactly is CSR? Half a century ago researchers in the United States started to define corporate social responsibility. An important question that surfaced was: “What responsibilities to society may businessmen reasonably be expected to assume?” (Carrol, 1999). During the second half of the 20th
century the literature on CSR expanded rapidly and an all-inclusive definition of CSR was formulated by Crane and Matten (2010). They stated that CSR consists of four different dimensions, namely: economic, legal, ethical, and philanthropic expectations. These four
dimensions indicate that CSR is multidimensional. Therefore, companies should try to focus on all four dimensions to reach a high corporate social performance. The
definition of CSR used throughout the thesis is: “Corporate social responsibility includes the economic, legal, ethical, and philanthropic expectations placed by society at a given point in time (Crane & Matten, 2010, p.53).”
In order to make a distinction between different kinds of CSR initiatives, in this study the KLD Stats database, constructed by KLD Research and Analytics Inc., is employed. KLD is an investment firm, situated in Boston, which specializes in tracking firm’s CSR activities. Since 1991 this firm has been collecting data on various CSR initiative aspects for a large number of U.S. companies. KLD divides CSR activities in 13 broad categories: community, diversity, employment, environment, human rights, products, alcohol, gambling, firearms, military, nuclear, tobacco, and corporate governance (Servaes & Tamayo, 2013). The decision to use
the criteria formulated by KLD is based on the fact that many companies participate in CSR within the criteria constructed by KLD.
Different employees can have different valuations of CSR involvement. A firm’s managers may seek to explicitly overinvest in CSR. According to Barnea and Rubin (2010), overinvesting in CSR will not result in a better firm performance. Overinvesting in CSR will benefit the reputation of the managers themselves. Focusing on prospective employees, according to Turban & Greening (2007) firms with a stronger (perceived) corporate social performance are more attractive to prospective employees and they have a lower turnover among employees. Referring back to the different CSR categories used in this research, Backhaus, Stone, & Heiner (2002) state that jobseekers consider the CSR dimensions of environment, community, employee relations, diversity, and product more important than the other dimensions. This suggests that employees prefer certain CSR initiatives to others. The preference for each of these categories may not only depend on the stakeholder group. CSR initiative preference may also depend on individual differences, such as moral orientations. The two major moral orientations are discussed in the following paragraph. The moral orientations will be reflected to the level of authority an individual has within a firm.
2.2 Moral orientation
For decades scientists were debating on the history of moral judgments. Eventually the main question being debated was: How many moral mental systems are there? This is where the Moral Foundations Theory (MFT) comes into stake. According to (Graham et al., 2013), the MFT explains different aspects of the moral domain. The theory states that there are five major moral foundations that each correspond to a mental system: Harm/care, Fairness/reciprocity, Ingroup/loyalty, Authority/respect, and Purity/sanctity. The five moral foundations describe what is considered an aspect of human morality (Graham et al., 2013, p.56).
The Harm/care foundation is about caring, nurturing, and protecting sensitive individuals. For example, if a baby is crying and needs help, the Harm/care foundation is triggered. Fairness/reciprocity, the second moral foundation, concerns equality issues. ‘Everyone needs to be treated equally’ is one of the leading statements embraced by this moral foundation. A practical example provided by Graham et al. (2013) is buying a soda from a vending machine: You put your money
in the machine and it fails to deliver the desired product. This can trigger the Fairness/reciprocity foundation: the defect of the machine is perceived as unfair. The first two moral foundations are considered individualizing foundations (Graham et al., 2009).
The third moral foundation is called the Ingroup/loyalty foundation. Are you as a person willing to be part of a group? This foundation can be illustrated by an example provided by Graham et al. (2013): Sports fans instantly trigger the third moral foundation. They are all willing to give up their own identity for the team. Universities in the United States even name their students after the university mascot to create a feeling of trust and support. The Authority/respect foundation refers to the social order people live in (Graham, Haidt, and Nosek, 2009). Especially at work a social order is visible. A manager of an organization clearly has more rights (and salary) than an employee fabricating the products. Also the employees are expected to respect certain cultural traditions, usually developed by the managers higher up in the company. The last moral foundation, the Purity/sanctity foundation, refers to the way people treat their bodies. Nowadays smoking can be seen as a dangerous threat for health, this refers to the Purity/sanctity foundation. The last moral foundation can also be related to people who are offended by other people (not caring for themselves). Graham et al. (2009) refer to the last three moral foundations as the binding foundations.
Research has shown that a political distinction can be made between the two moral orientations. According to Haidt and Graham (2007), on the one hand, liberals would find issues of Harm/care and of Fairness/reciprocity more important than conservatives. Conservatives, on the other hand, would prioritize the Ingroup/loyalty, Authority/respect, and the Purity/sanctity foundations compared to the liberals.
People differ in their moral orientation depending on their social status. Studies show that socioeconomic status positively relates to liberals (Bock, Beeghley, & Mixon, 1983). It is, therefore, not unthinkable that authority plays an important role in the distinction between the two major moral foundations, therefore the next section of this study will cover the authority employees have within a firm.
2.3 Authority
compared to lower-level employees within a firm. This level of authority can be associated with more power. Authority has three distinct dimensions. Firstly, there is the formal hierarchical position, secondly sanctioning authority, and lastly decision-making authority (Wright et al., 1995). In the measurements section (p. 21) of this study a more extensive overview of the three dimensions is given.
Current research on the relationship between power and political preference (either being a liberal or a conservative) states that individuals in a powerful position tend to make more conservative decisions (Maner, Gailliot, Butz, & Peruche, 2007). (Clapham & Schwenk, 1991)In a business context, this means that there might be a relationship between authority on the work floor and moral orientation.
According to Weber (1993), managers tend to maintain the conventional social order within a firm. For people with more authority it would be self-serving to prefer other people to respect authority and act in accordance with the given social order. It is the manager’s main task to uphold the social order. Hence managers will score higher on the authority foundation than people with less authority. It is self-serving to expect that managers will also score higher on the loyalty foundation. Managers will be responsible for the performance of the firm; therefore, people with more authority expect other lower-leveled employees to contribute to the group or firm. For people with more authority, it will be self-serving to expect that other people (in the firm) stay healthy, because managers will often be held accountable if their employees are often ill. Thus they will score higher on the Purity/sanctity foundation. Combining these self-serving expectations result in the assumption that people with more authority will score higher on the binding foundations.
Also this study holds the assumption that people with more authority will score lower on the individualizing foundations: namely, Harm/care and Fairness/reciprocity. People with more authority will often care less about vulnerable individuals because they are not very unsusceptible themselves. Lastly, it is self-serving to expect that people with more authority will care less about justifying their behavior to others, this may be because they are usually not obligated to justify their actions. These arguments lead toward the first two hypotheses.
Hypothesis 1: Employees with less authority will score higher on the individualizing foundations than employees with more authority.
Hypothesis 2: Employees with less authority will score lower on the binding foundations than employees with more authority.
After having stated the first two hypotheses, the prediction is that employees with less authority will score higher on the individualizing foundations and will score lower on the binding foundations.
This study is not limited to whether an employee with a certain level of authority will score higher or lower on the two moral orientations. It is a lot more interesting and relevant for firms to seek the relationship between the thirteen different types of social activities and the two leading moral orientations. Therefore, the next section will briefly cover all the thirteen types mentioned above and categorize them into either individualizing or binding initiatives.
2.4 Types of corporate social responsibility initiatives
As mentioned above, for this study the thirteen types of CSR initiatives provided by KLD Research and Analytics Inc. have been chosen. All thirteen corporate social responsible items are categorized as either an individualizing or a binding corporate social responsible initiative. Specifically, if engaging in a certain initiative reveals that one finds individualizing foundations important, then this initiative is categorized as an individualizing initiative.
2.4.1 Community
When firms give notable amounts to charities, they are caring for other people than themselves. Finding this action important or morally relevant would correspond to higher scores on the individualizing foundations. Therefore, community activities are categorized as individualizing initiatives.
2.4.2 Diversity
The second type of CSR initiative can also be categorized as individualizing initiatives. For example, when a firm nominates a woman for the position of CEO of a firm, this initiative follows the requests of individualizing foundations. Diversity within a firm refers to the moral foundation Fairness/reciprocity: every individual has the same opportunity to become the CEO of the firm, both men and women.
2.4.3 Employment
An example of an employment initiative is when a firm decides to provide a stable safety plan for their employees. In this case the individualizing moral foundations relate the closest to the chosen activity. It is fair and caring towards employees that employment initiatives such as proper safety measures are installed. Therefore, any employment initiative can be categorized as an individualizing initiative.
2.4.4 Environment
When a firm engages in environmental benefits, they conform to the values that are prescribed in the binding foundations. An example of an environmental initiative is using separate garbage cans for recyclable and non-recyclable items, or attempting to reduce the pollution of their plant. The firm is not trying to be more profitable themselves, they are rather providing an effort to make the environment cleaner as a whole. The two binding foundations, Purity/sanctity and Ingroup/loyalty, relate the closest to environmental social initiatives. Therefore, activities related to the environment are categorized as binding initiatives.
2.4.5 Human rights
There are multiple examples of corporate social initiatives that enhance human rights; for instance, donating notable amounts to non-governmental organizations (NGO) such as Amnesty International. Another example, related more to the employees, could be treating the unionized workforce fairly. The human rights aspect within a firm refers most closely to the ‘fairness’ foundation. Therefore, human rights initiatives are categorized as individualizing initiatives.
2.4.6 Products
Creating innovative products makes the society better as a whole. The binding foundations correspond the best with this. Therefore, initiatives associated with products are categorized as binding initiatives.
2.4.7 Alcohol
When firms maintain a no-alcohol policy within their company they are enhancing their authority, sustaining the social order, and preserving the purity of their employees. Authority and purity are two of the three foundations that are associated
with the binding foundations. Therefore, alcohol restrictions on the work-floor can be seen as binding initiatives.
2.4.8 Gambling
Social responsible activities categorized as gambling can be seen as binding initiatives. When a company decides to license its company or brand name to gambling products, that is an example of a socially responsible gambling initiative. Gambling is often related to other ‘sins’ such as drinking or using drugs, and therefore gambling relates the closest to the ‘purity’ foundation.
2.4.9 Firearms
An example of a CSR initiative related to firearms, is not being involved in the production of firearms. The CSR category firearms can be closely related to two of the five moral foundations. First of all, a gun can harm an individual. Secondly, people can hold the belief that every individual has the right to own a firearm. The first example refers to the Harm/care foundation, and in the second case it reflects the Fairness/reciprocity foundation. Both of these moral foundations are categorized as individualizing foundations. However, there is to be noted that scoring high on the Harm/care foundation will result in the person preferring the firm not to be involved in producing firearms, while scoring high on the Fairness/reciprocity foundation will result in the person preferring the firm to be involved in producing firearms. Considering this study is done in Western Europe, this category is more likely to be perceived as belonging to the Harm/care foundation, since the discussion about firearms in Western Europe mainly revolves around the harm of weapons, rather than about the right to own a gun.
2.4.10 Military
When questions are asked about the military, people usually relate that to their loyalty towards their country in order to protect it. The moral foundation that fits the closest to military initiatives would be Ingroup/loyalty. This foundation is categorized as a binding foundation, and therefore CSR initiatives that relate to the military are seen as binding initiatives.
2.4.11 Nuclear
Nuclear materials do not improve the environment. ‘Purity’ enhances this thought and relates the closest to nuclear initiatives. CSR initiatives meant to avoid nuclear companies or materials can be seen as binding initiatives.
2.4.12 Tobacco
For this study there is chosen to link the tobacco category to the Purity/sanctity foundation. The tobacco category has been measured by using two questions. Firstly, the employees were asked to what extent they would agree to the fact that their company would engage in the production of products necessary for the production of tobacco products. Secondly, the employees were asked to what extent they would agree to have a smoking room provided by the employer. Scoring high on both of these items means that the employee has actually scored low on the Purity/sanctity foundation. To solve this issue there is chosen to reverse code both of these items.
2.4.13 Corporate Governance
The examples used in the survey that are linked to corporate governance are closely related to the ‘fairness’ foundation. Awarding notably low levels of compensation to the top management, and transparent decision-making within a firm are typical examples of individualizing initiatives. Therefore, corporate governance initiatives are categorized as individualizing activities.
Table 1: Categorization CSR initiatives
Category Examples Reverse coded? Moral foundation
Community Giving notable amounts to charity No Individualizing
Diversity Nominating a woman for the
position of CEO
No Individualizing
Employment Having a strong safety plan for
employees
No Individualizing
Environment Reducing pollution during
production process
No Binding
Human rights Donating notable amounts to
NGO
No Individualizing
Products Creating innovative products No Binding
Alcohol Restricting alcohol for employees No Binding
Alcohol Own another company that
produces alcohol
Gambling License its company or brand name to gambling products
Yes Binding
Firearms Firearms can harm individuals No Individualizing
Military Loyalty towards your country No Binding
Nuclear Investing in a company that
produces nuclear power material
Yes Binding
Tobacco Providing a smoking room for
employees
Yes Binding
Tobacco Producing products necessary for
the production of tobacco products
Yes Binding
Corporate Governance
Awarding notably low levels of compensation to top management
No Individualizing Corporate Governance Transparent decision-making within a firm No Individualizing
After having categorized the thirteen types of CSR initiatives into two groups (see Table 1 above), it is possible to make a distinction between initiatives preferred by employees who score higher on the binding foundations and between employees who score higher on the individualizing foundations. This leads toward to the third and fourth hypotheses.
Hypothesis 3: Employees who score higher on the binding foundations will be more attracted to the following CSR initiative categories: environment, products, alcohol, gambling, military, nuclear, and activities related to corporate governance.
The three binding foundations Authority/respect, Ingroup/loyalty, and Purity/sanctity characterize all of the initiatives mentioned in hypothesis 3.
Hypothesis 4: Employees who score higher on the individualizing foundations will be more attracted to the following CSR initiative categories: community, diversity, employment, human rights, and activities related to firearms.
The two individualizing foundations Harm/care and Fairness/reciprocity characterize all of the initiatives mentioned in hypothesis 4.
2.5 Use of moral orientation as Mediator
In the last two hypotheses the moral orientation will be used as mediator. As
discussed before, the moral orientation of individuals could affect the personal values of internal stakeholder regarding the choice of certain CSR initiatives. This leads towards the last two hypotheses.
Hypothesis 5: The positive relationship between authority and binding CSR initiatives will be partly mediated by the score on binding moral foundations.
Hypothesis 6: The negative relationship between authority and individualizing CSR initiatives will be partly mediated by the score on individualizing foundations.
2.5 Visual representation of the conceptual framework
As discussed above, two mediation models are going to be tested. Both of them will be shortly discussed and visualized below. To run all regressions in one model at once, a custom model, developed by Hayes (2013) was used. Model 4 (mediation model) was used; this model is most suitable for testing the six hypotheses. In the next section the two mediation models are visualized and described.
2.5.1 Mediation model 1
The first mediation model will test the relationship between authority and the valuation of binding CSR initiatives, mediated by the individualizing moral foundations. The hypotheses 1,4, and 6 will be tested.
Figure 1: Mediation Model 1
2.5.2 Mediation model 2
The second mediation model will test the relationship between authority and the valuation of individualizing CSR initiatives, mediated by the binding moral foundations. The hypotheses 2,3, and 5 will be tested.
3 Methodology
This paragraph will explain how the previously stated hypotheses are tested, how the research is designed, what data were collected, how the data were collected, and finally how the three main constructs are measured.
3.1 Research design
The first decision to make is whether to choose for a quantitative, qualitative or multiple methods research design. For this research there was chosen for a quantitative research design. The dataset was collected through a questionnaire-based survey. The survey was sent electronically using the Intranet of the firm and self-completed by the employees; this is called an intranet-mediated questionnaire (Saunders, Lewis, & Thornhill, 2012, p. 419). By having spread it through the Intranet, a higher return rate was expected because the questionnaire actually reached every employee of a firm and the employees felt more involved. A self-completed survey saved a lot of time in collecting the data. It also reduced the participant bias, since there was no interviewer present. The participants’ tendencies were not to act in the way that they assumed the researcher wanted them to. Therefore, the reliability of the data improved (Saunders et al., 2012, p.192). Using a survey made it easier to achieve a large sample size, which was needed to draw any conclusions from the collected data.
Besides the advantages of a survey, there are limitations to a survey design. The questions asked in the survey must not be too complicated. Every respondent should be able to understand the questions or statements. Besides the type of questions, the amount of questions asked must also be considered (Saunders et al., 2012, p.421). According to Saunders et al. (2012), the number of questions that can be included in a survey is a major limitation. Respondents generally are not willing to fill out a large, time-consuming questionnaire. Another limitation of an online survey tool for this specific research is that it excludes a large group of people who have no access to the Internet. Considering most employees in a management or supervisory function have access to the Internet and the lower ranked employees might not have computer access, limits the representativeness of the sample.
To solve this issue a paper version of the survey also was constructed. The online and paper versions were identical to each other. The major advantage of having provided the respondents with a paper version is that not only employees with Internet access could participate in the research but also the lower ranked employees. To be able to answer the main research question it was necessary to find participants ranked from the highest level of the firm to the lowest level of the firm.
The survey was constructed using an online survey tool called Qualtrics. This tool made it very easy to develop a survey; questions could be easily added or modified, also the survey flow could be edited. This meant that a respondent could be redirected to certain follow-up question based on answers on previous questions.
The survey used for this thesis consisted of five parts. The questionnaire started with an introduction. After this the participant was asked to answer demographic questions such as ‘what is your gender?’ and ‘in what year were you born?’ These variables are called control variables. After successfully answering the demographical questions, the participant was asked to answer questions on the organization he or she works for. These questions were focused on the different levels of authority within a firm. The questionnaire was set up in a way that the participant was redirected to the appropriate question after answering the previous question. For example, after stating that he or she is part of the top management team, the participant was redirected to follow-up questions on the specific level of authority he or she can exercise within the firm. In the next part of the survey the respondents had to answer to what extent they would like their organization to participate in 28 different initiatives. The 28 statements are all examples of the 13 different types of CSR initiatives. However, the participant did not have any knowledge of this fact. For the rating of the statements a Likert-style rating was used in which the respondent was asked to what extent they would like their organization to engage in these initiatives (Saunders, Lewis, & Thornhill, 2012). The last part of the questionnaire consisted of a version of the Moral Foundations Questionnaire (MFQ). The main objective of this part of the questionnaire was to identify whether an individual employee scored high on individualizing and binding foundation. The MFQ can be seen as a reliable, valid, and easy-to-use tool to examine whether the respondents score high or low on the individualizing and binding foundations (Graham et al., 2011).
In the following sections, the authority measurement and the measurement of valuation of CSR initiatives are discussed. This is followed by a section about the sample.
3.2 Measurement of authority
Wright et al. (1992) state that authority can be measured using the following three dimensions: formal hierarchical position, sanctioning authority, and decision-making authority. Firstly, in the next section the three dimensions will be explained and operationalized. Finally, a conclusion will be provided how to compute the ‘authority’ variable used throughout this study.
3.2.1 Formal hierarchical position
According to Wright et al. (1995), a clear distinction can be made between management positions. Managerial, supervisory or non-management positions are the three main hierarchical positions that can be held within a firm. The management positions can be divided into four different segments. Categorized from high to low the segments are: top, upper, middle, and lower management. The formal hierarchical position is only one of the conditions of authority.
3.2.2 Sanctioning authority
The second dimension of authority is questioned from a different angle. Sanctioning authority refers to the capacity of an individual to enforce a reward and punishment system on subordinates (Wright et al., 1995, p. 9).
3.2.3 Decision-making authority
The last distinct dimension of authority is called decision-making authority. According to Wright et al. (1995,) this variable concerns the direct participation in organizational policy making decisions.
All three of these dimensions combined generate the ‘authority’ variable used throughout the study. If an individual scores positive on two of the three dimensions the individual can be seen as an individual with authority (see the paper survey in Appendix I for all the questions covering the authority variable).
3.3. Measurement of CSR valuation
The measurement of the thirteen CSR categories is based on the KLD Stats database. However, since two versions of the survey were distributed to employees of two different companies the questions covering each CSR category slightly differ.
In Table 2 an overview of all the questions covering the thirteen CSR categories is given.
Table 2: Question overview CSR categories
CSR category Questions Company
Community - Giving notable amounts to charity
- Supporting local cultural institutions
- Both companies - Both companies
Diversity - Nominating a woman for the
position of CEO
- Undertaking efforts to have more women or minority members on the board of directors
- Both companies
- Belgium company
Employment - Deciding to be an employer of
disabled people
- Having a strong safety plan for employees
- Both companies
- Both companies
Environment - Reducing pollution during
production process
- Use of separate garbage cans for recyclable and non-recyclable items - Use of recycled or environmentally
friendly materials
- Both companies
- Both companies
- Belgium company
Human rights - Giving money to a global NGO like
Amnesty International
- Implementing hiring practices that reduce discrimination of applicants - Treating its unionized workforce
fairly
- Both companies
- Belgium company
- Belgium company
Products - Investing a substantial amount in
R&D
- Initiating a Product Quality Program
- Both companies
- Investing in the safety of its products or services
- Belgium company
Alcohol - Owning another company that
produces alcohol
- Having a no-alcohol policy
- Spanish company
- Belgium company
Gambling - Licensing its brand name to
gambling products
- Collaborating with an overseas company that produces gambling products
- Spanish company
- Belgium company
Firearms - Engaging in the production of
ammunition or firearms
- Taking assignments from companies that produce firearms
- Spanish company
- Belgium company
Military - Being owned for more than 50% by
a company that is involved in military activities
- Being a spin-off of an organization that is involved in the military industry
- Spanish company
- Belgium company
Nuclear - Being involved in the transport of
nuclear material
- Investing in an organization that produces nuclear power material
- Spanish company
- Belgium company
Tobacco - Producing products that are
necessary for the production of tobacco products
- Providing employees with an indoor smoking room
- Spanish company
- Belgium company
Corporate Governance
- Awarding notably low levels of compensation to its top management - Having a transparent
decision-making process.
- Both companies
- Both companies
Most of the questions created to measure the thirteen CSR categories were used for both the Spanish firm and the company in Belgium. However, Table 2 shows that
changes were made in the questionnaire mostly in the following categories: alcohol, gambling, firearms, military, nuclear, and tobacco. After revising and collecting feedback on the first version of the survey distributed in Spain, it became clear that the questions used to measure the CSR categories mentioned above were not appropriate for the firm situated in Belgium. Therefore, a second version was developed with questions that would be appropriate for the Belgium firm.
3.2 Sample
Sending out the survey to random individuals would not have been sufficient enough to answer the research question, since it was necessary to collect data from both managers and lower leveled employees. The first thought was to send the survey out to one large organization and hope that all of the employees would participate. Statistically seen the response rate of online surveys is relatively low, namely 11% or lower (Saunders et al., 2009, p. 364), but by adding the paper version the response rate was expected to be significantly higher.
For this research two companies were selected. The first company, a Spanish company situated in Barcelona, was approached. The company in Spain, called Estoli (www. estoli.com), specializes in the fabrication of toilet seats and other bathroom accessories. To make it easier for the respondents to actually comprehend the survey, there was chosen to make two versions: One in English and one in Castilian. By doing this the response rate would not only rise, but also the collected data would be more trustworthy. Only sending out an English version of the questionnaire would have resulted in data lost due to the language barrier. A professional translator, specialized in Spanish, English, French, and Italian, did the translation of the survey. Also a Spanish student, studying in the United States, checked the questionnaire for mistakes. No translation mistakes were found.
The second firm selected for this research is a Belgium company called Isabel (www.isabel.eu). Isabel is the main provider in Belgium of Internet banking for the professional user. The company consists of 170 employees and, since it is an IT company, most employees have access to the Internet (Isabel, 2014).
In the Spanish company, consisting of only 53 employees, 34 employees responded. This was not a lot of participants, but the response rate was significantly higher, namely 64%. More than half, 19, of the respondents preferred the paper
of employees working for the company. By adding another, larger firm to the data collection more reliable conclusions can be made.
Within the Belgium firm a lot more data were collected. Fifty-six employees completed the online survey. This results in a response rate of 31%. Saunders et al. (2009, p. 364) states that the response rate of online surveys is in general relatively low, namely 11% or lower. However, in this particular research, the average response rate is significantly higher.
The research was eventually limited to 89 participants, differing in function, and firm. The demographics of the participants can be found in Table 3.
Table 3: Demographics respondents
Variable Factor Percentage
Gender Male 76,7% Age 10-20 1,1% 21-30 14,4% 31-40 38,9% 41-50 30% 51-60 13,3% 61-70 2,2%
Educational level Secondary education 7,1%
University Bachelor’s degree 50% University Master’s degree 35,7% Post-university degree 7,1%
As you can see in Table 2 almost all of the respondents have completed at least a bachelor’s degree.
3.3 Analyses
To test the hypotheses the Statistical Package for the Social Sciences program was used.
First the reliability analysis was conducted to test the internal consistency of the variables. By conducting the internal consistency test it was possible to see if it was necessary to delete an item from the variables ‘binding CSR initiatives’, ‘individualizing CSR initiatives’, and ‘moral orientation’. Then the descriptive
statistics (mean, standard deviations) were computed. Thereafter a custom dialog developed by Hayes (2013) was used. This was done to run all regressions in one model at once. By doing this it is possible to simply include the mediator in this study. The mediator used for this research was the variable ‘moral orientation’. By using the custom dialog developed by Hayes (2013) it was not necessary to do a stepwise mediation analyses. The model chosen to use for this study is the process Model 4 developed by Hayes (2013). This model included a mediator. In this study there was chosen not to perform a factor analysis. First of all, the size of the sample is not large enough. For a factor analysis the sample should consist of at least 100 respondents (Child, 1990); unfortunately, this study only included 89 participants. Secondly, insufficient definitions of factors could lead to a conflict between exactness and practicality (McNemar, 1951). Lastly, when conducting a factor analysis the researcher could tend to “see what is not really there”, which could lead the researcher to force the factors to correspond with her or his hypotheses (Child, 1990).
Table 4 provides an overview of each hypothesis, the methods used to test them, and the measurements used.
Table 4: Hypotheses overview
Variables Measurements Statistical tests H1 - Authority
- Individualizing moral foundations
- 3 dimensions
- MFQ Hayes regression,
Model 4
H2
- Authority
- Binding moral foundations
- 3 dimensions - MFQ
Hayes regression, Model 4
H3 - Binding moral foundations - Binding CSR initiatives - Mean binding moral foundations - Mean binding CSR initiatives Hayes regression. Model 4 Cronbach’s Alpha
H4 - Individualizing moral foundations
- Individualizing CSR initiatives - Mean moral foundations - Mean individualizing CSR initiatives Hayes regression. Model 4 Cronbach’s Alpha H5 - Authority - Binding CSR initiatives - 3 dimensions - Mean binding CSR initiatives Hayes regression, Model 4 H6 - Authority - Individualizing CSR initiatives - 3 dimensions - Mean individualizing CSR initiatives Hayes regression, Model 4
4 Results
The results of the research discussed in the previous section will be provided in this section. First, the reliability of scales is shown by means of the Cronbach’s alpha for the variables ‘binding foundations’, ‘individualizing foundations’, ‘binding CSR initiatives’, and ‘individualizing CSR initiatives’. Thereafter, some descriptive statistics will be presented. Finally, the results of the regression model 4 will be discussed.
4.1 Reliability
In the survey many scales were used to probe underlying constructs for this study. The scores on the items of these scales were used to calculate the mediators and dependent variables. The mediator variables in this research are ‘individualizing foundations’ and ‘binding foundations’. The dependent variables are ‘individualizing CSR initiatives’ and ‘binding CSR initiatives’. To test if the items questioned on the survey are reliable and actually measure the same construct needed to answer the main research, question the statistic coefficient Cronbach’s Alpha is used. The Cronbach’s Alpha can be accepted when it is at least 0,7 (Bland & Altman, 1997).
4.1.1 Moral orientation
This construct consists of two variables used for this study, namely ‘binding
foundations’ and ‘individualizing foundations’. The questions asked on the survey to measure these variables are directly implemented from the already existing Moral Foundations Questionnaire (MFQ). The Cronbach’s Alpha value for the
individualizing foundations variable measured α=0,693, which is not sufficient enough. However, α= 0,715 can be reached by removing one question that asked the respondent to what extent she or he agrees on the statement “one of the worst things a person could do is hurt a defenseless animal”. This item was therefore removed. The Cronbach’s Alpha for the other moral orientation variable was measured using the same method. This resulted in α=0,725, which is high enough, but also this one could be improved by removing the question that asked the respondent to what extent she or he agrees on the statement “I would call some acts wrong on the grounds that they are unnatural”. By doing this the reliability of the summated scale improved to α=0,742.
4.1.2 CSR initiatives
The items used in the questionnaire to measure the following two variables are all based on the definition of CSR provided by KLD Research and Analytics Inc. Moreover, the thirteen CSR categories were divided into two groups either
individualizing or binding CSR initiatives. This process has not been performed in the literature before and therefore, cannot be supported by previous research.
Since participants from two different companies (and countries) participated in this research, a slightly different version of the questionnaire was spread. Simply because some CSR questions were not applicable for one of the companies, it is not possible to calculate one reliability coefficient that applies to both companies. Therefore, there is chosen to calculate the Cronbach’s Alpha for both companies and compare the possible differences.
For the company situated in Belgium α=0,792 for the variable ‘individualizing CSR initiatives’ and α=0,596 for the variable ‘binding CSR initiatives’. The
Cronbach’s Alpha measuring the individualizing CSR initiatives is sufficient enough, but still could be improved to α=0,833 by removing the statement to what extent she or he agrees on the company engaging in the production of ammunition or firearms. The reliability of the binding CSR initiatives (α=0,596) is not high enough to reliably measure the construct. Removing an item from the survey will still not significantly improve the Cronbachs’s Alpha, but it would slightly improve to a value of α=0,633 if the item that refers to any Product Quality Programs initiated within the firm was removed. Thus, these two items were removed.
The results concerning the questionnaires filled out by employees of the company in Spain are similar for the variable ‘individualizing CSR initiatives’, α=0,719. This is the highest reachable value after removing, surprisingly, the same question from the survey concerning to what extent she or he agrees on their company engaging in the production of ammunition or firearms. The Cronbach’s Alpha
measured for the binding CSR initiatives, however, is slightly lower than the reliability coefficient noted for the Belgium firm, α= 0,522. After removing the question that accounts for any Product Quality Programs initiated within the firm, the Alpha value improves to α=0,569. Thus, these two items were removed.
The conclusion that can be made after calculating the Cronbach’s Alpha for the appropriate variables is that all of the Alpha’s measures, except the reliability
coefficient covering the moral orientations, are low. The reason for this could be because the items that measure the thirteen different CSR initiatives are not based on existing literature.
4.2 Descriptive statistics
In this section essential statistical descriptive findings will be presented. In Table 5 the mean and standard deviation are stated. It is interesting to view these results before implementing the variables and testing them with regression Model 4, developed by Hayes.
Table 5: Descriptive statistics
Minimum Maximum Mean Std. Deviation Authority (n=89) 0,00 1,00 0,69 0,47 Individualizing foundations (n=89) 1,92 5,25 4,30 0,63 Binding foundations (n=89) 2,33 4,83 3,65 0,51 Individualizing CSR initiatives (n=89) 2,00 5,33 3,45 0,56 Binding CSR initiatives (n=89) 2,50 5,09 3,62 0,57
There are a few notable observations to be made. First of all, comparing the two major moral orientations results in a significant higher value for the individualizing moral orientation, 4,30 compared to 3,65. On a 0-5 scale this is quite significant. Secondly, looking at the CSR initiatives, both individualizing and binding initiatives, the values measured are almost equal. This means that the participants equally value CSR initiatives focused on an individualizing moral orientation as a CSR activities on a binding moral orientation. Furthermore, Table 6 states that the participants with authority are slightly overrepresented in this study. The ideal situation would have been to have an authority mean of 0,5. This would mean that exactly half of the
participants have notable power within a firm (see p.22 for the variable authority used in this study).
4.3 Model estimation and interpretation
In this chapter the significance of the models will be presented and both models will be interpreted in order to conclude whether or not the hypotheses are either accepted or rejected. To start, for both models in this study, the process regression model was used to test whether the two main variables (authority and CSR initiatives) are mediated by the third variable (moral orientation). Andrew Hayes developed this custom dialog providing SPSS with an additional tool that measures mediation effects without doing separate statistical tests as it shows all relevant results at once. The method presented below makes use of unstandardized B-coefficients, to make sure the results are easy to interpret.
4.3.1 Mediation model 1
The process regression analysis of the first model tested whether authority within a firm predicted the valuation of a variety of binding CSR initiatives and if the relationship was mediated by binding moral orientation. Hypotheses 1,4, and 6 are combined in this first mediation model.
Hypothesis 1 stated that employees with less authority would score higher on the individualizing foundations than employees with more authority. The model shows that this hypothesis could be rejected since β=-0,1838; p=0,4319. This means that if authority goes down by one, the individualizing moral orientation goes up by 0,1838, so an employee with less authority has a higher individualizing moral orientation than an employee with more authority. However, since the p-value is larger than 0,05 this result is not significant. Therefore, the hypothesis is rejected. Suggestions to improve the significance of the result will be given in the last part of this study.
Hypothesis 4 stated that employees who score higher on the individualizing foundations will be more attracted to the following CSR initiative categories: community, diversity, employment, and human rights. This hypothesis could be accepted with β=0,2473; p<0,05. This means that the higher an employee scores on the individualizing moral foundations the more an employee will be attracted to individualizing CSR initiative categories.
The last hypothesis in this research stated that the negative relationship between authority and individualizing CSR initiatives would be partly mediated by the score on individualizing foundations. The results present that the negative relationship between authority and individualizing CSR initiatives are not significantly mediated by the score on the individualizing moral foundations. The hypothesis could be rejected. All the results are further presented in figure 3.
To further explain the (lack of) mediation of an individualizing moral orientation, the output in SPSS also shows a 95% confidence interval of the stated indirect effect. According to Hayes (2013), if the confidence interval does not contain zero, it is likely that there is a mediation effect. The results show a mediation value close to zero, so there is a small indirect effect, but not significant. Since a 95% confidence interval is used, it might not be unthinkable that with a slightly smaller confidence interval there would be a significant mediation effect. The findings are presented in Table 6.
Table 6: Mediation model 1 indirect effect 4.3.2 Mediation model 2
The process regression analysis of the second mediation model tested whether authority within a firm predicted the valuation of a variety of binding CSR initiatives. In addition, the analysis tested if the relationship between authority and the binding CSR initiatives was mediated by the individualizing moral orientation. Mediation model 2 covers hypotheses 2,3, and 5; these hypotheses will be discussed in the next section together with the appropriate results.
Hypothesis 2 stated that employees with less authority would score lower on the binding foundations than employees with more authority. The model shows that this hypothesis can be rejected with β=-0,3370; p=0,0514. This means that if the authority of an employee goes up by one the valuation on the binding moral orientation goes down with 0,3370. This result contradicts the stated hypothesis.
Hypothesis 3 stated that employees who scored higher on the binding foundations will be more attracted to the following CSR initiative categories: environment, products, alcohol, gambling, military, nuclear, and activities related to corporate governance. The model shows that also this hypothesis could be rejected presenting a β=-0,0316; p=0,8154. This means that if the binding foundations score goes up by one the score on binding CSR initiatives goes down by 0,0316, so the higher an employee scores on the binding foundations the less attracted an employee will be to binding CSR initiatives. These results contradict any expectations formulated in the hypothesis.
Hypothesis 5 stated that the positive relationship between authority and binding CSR initiatives would be partly mediated by the score on binding moral foundations. The mediation model shows that there is no indirect effect. This means that the binding moral orientation variable does not mediate the relationship between the authority of an employee and the valuation of a variety of binding CSR initiatives. Indirect effect of authority on individualizing CSR initiatives
Effect
Confidence interval (95%) Lower bound Upper bound Individualizing foundations -0,0413 -0,1394 0,0067
Comparable to the result of hypothesis 6 the indirect effect found is simply too small to be a significant mediator. All the results are further presented in Figure 4 and Table 7.
Figure 4: Mediation Model 2 including results
Table 7: Mediation model 2 indirect effect
4.4 Conclusion
After having tested all six hypotheses, a conclusion can be reached that only one hypothesis could be accepted. Employees who score higher on the individualizing moral foundations are more attracted to individualizing CSR initiatives. Unfortunately, the suggested mediation effects were not supported either. However, these results do lay the basis for further research on the relationship between the moral orientation of employees and the valuation of CSR initiatives.
Indirect effect of authority on binding CSR initiatives
Effect
Confidence interval (95%) Lower bound
Upper bound Binding foundations 0,0306 -0,0125 0,1506
5 Discussion
This section discusses the findings of this study and compares them to the existing literature and the findings of previous research. Furthermore, based on the findings some theoretical and managerial implications are presented. Finally, the limitations of the study are discussed and suggestions for future research are made.
5.1 Key findings
This study served the purpose of providing a deeper insight into the influence of authority on a variety of CSR initiatives. However, this study also focused on the underlying psychological mechanisms that would mediate the relationship between authority and the valuation of CSR initiatives. The research started with two
hypotheses that predict that authority and moral orientation are directly related to each other. The first hypothesis stated: Employees with less authority will score higher on the individualizing foundations than employees with more authority. Based on the results this hypothesis could be rejected. Employees with less authority do not score significantly higher on the individualizing moral foundations than employees with more authority. Actually, the findings show that employees with less authority score higher on the individualizing foundations. Unfortunately, this result is not significant and the hypothesis still is rejected. The reason for the lack of support for this
hypothesis could be the fact that the sample was small.
The second hypothesis, in line with the first one, stated: Employees with less authority will score lower on the binding foundations than employees with more authority. The second hypothesis could also be rejected. The results here show that employees with more authority will score lower on binding foundations than employees with less authority. An explanation for the lack of support could be that previous research argues that a high socioeconomic status positively relates to liberals who have lower scores on binding foundations, and that employees with more
authority might have a higher socioeconomic status (Bock, Beeghley, & Mixon, 1983; (Graham et al., 2009). Thus, future studies should control for socioeconomic status.
Furthermore, the next two hypotheses focus on the direct relationship between authority and the valuation of a variety of CSR initiatives. Hypothesis 3 predicting
‘employees who score higher on the binding foundations will be more attracted to binding CSR initiative’ could be rejected. The findings state that employees who score high on the binding foundations will be less attracted to binding CSR initiatives. An explanation for the lack of support could be that the questions asked in the
questionnaire to measure the variable binding CSR initiatives were not totally reliable. Unfortunately, there was no previous research that categorized a variety of CSR initiatives as either an individualizing or a binding initiative. However, the fourth hypothesis ‘employees who score higher on the individualizing foundations will be more attracted to individualizing CSR initiatives’ could be supported. The study shows us that employees who score higher on the individualizing foundations are more attracted to certain individualizing CSR initiatives than employees who score low on the individualizing foundations. An interesting question that can be asked is: why could hypothesis 4 be supported while hypothesis 3 had to be rejected? A possible explanation for this unexpected result might be the reliability of variables. On the one hand, the individualizing CSR initiatives variable has a sufficient
Cronbach’s Alpha (α=0,833). On the other hand, the binding CSR initiatives variable was not reliable enough. Blant and Altman (1997) argue that for a summated scale to be reliable the Cronbach’s Alpha has to have a minimum of α=0,7, which was not the case. Trying to improve the Cronbach’s Alpha values in this study would not only enhance the reliability but also the results might be more significant.
Unexpectedly, the relationship between the authority of an employee in a firm and the valuation of a variety of binding CSR initiatives is not mediated by binding foundations. The fifth hypothesis ‘the positive relationship between authority and binding CSR initiatives will be partly mediated by the score on binding moral foundations’ was not supported. An explanation for the lack of support could be that there is no significant positive relationship between authority and binding CSR initiatives. Furthermore, a 95% confidence interval was used to reveal any mediation effects. By using a slightly smaller confidence interval the indirect effect might be significant, since the values of the interval are situated very close to zero.
As is the case with the previous hypotheses the last hypothesis ‘the negative relationship between authority and individualizing CSR initiatives will be partly mediated by the score on individualizing foundations’ is not supported. A reason for the lack of support for this hypothesis could be the 95% confidence interval used. It is
