The influence of governments and NGOs on animal testing in companies

(1)

0

THE INFLUENCE OF GOVERNMENTS AND

NGOS ON ANIMAL TESTING IN COMPANIES

Elise H.J. de Reijer 11388765

Master thesis Business Administration: Strategy Supervisor: dr. D.A. Wäger

(2)

1

Statement of originality

This document is written by Elise de Reijer, 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.

Acknowledgement

With the writing of this thesis I close a period of my life that has been fun, informative and sometimes lead to weariness. I realize that this thesis could not have come to a successful end without the help and support of some people.

I would like to thank my parents for their indefinite support and for always being there for me even if I thought I did not needed them.

Furthermore, I would like to thank Nils and Marieke, two of my close friends, for all their feedback, tips and time.

I want to acknowledge and thank my thesis supervisor Dr. Daniel Wäger, of the University of Amsterdam, for his valuable feedback and guidance. Even though we did not meet regularly, I knew I could always send an e-mail with all of my questions. I greatly appreciate the speed with which he responded to my calls for help.

(3)

Master thesis Elise de Reijer, Business Administration, 2017

2

3 Abstract

Background: Corporations try to behave more socially responsible in order to meet societal calls. These demands are voiced through NGOs and other parties which try to influence governments and CSR reports. The purpose and uniqueness of this study is to provide empirical quantitative evidence on how governments and NGOs can influence the use of Animal Testing in companies. Governmental influence is measured by studying Governmental Legislation and Governmental Commitment

regarding the protection of animals. NGO influence is measured through partnerships between companies and environmental NGOs and through NGO presence.

Methods: A Hierarchical Logistic Regression analysis was performed on data retrieved from the databases Thomson Reuters ASSET 4 and WorldScope, from the Animal Protection Index and by calculating a variable based on the International Union for Conservation of Nature . Results: Both Governmental Legislation (p < .05) and Governmental Commitment (p < .01) have a positive main effect on Animal Testing. Partnerships between corporations and environmental NGOs have a negative main effect (p < .01), meaning that Partnerships decrease the amount of animal testing. Environmental NGO presence has a positive main effect (p < .01), more NGOs relate to less Animal Testing. NGO presence also strengthens the relationship between Legislation and Animal Testing (p < .01).

Conclusions: On a societal level these results imply that NGOs can have an influence on animal testing, and most likely on other areas of animal welfare, by simply being around and by partnering with corporations. Recommendations for governments for the protection of animals are difficult to give and more research is needed.

(5)

4 1. Introduction

In the last decade businesses have engaged in so-called corporate social responsibility (CSR) behavior. Nowadays, companies do not only set financial and strategic goals, but also strive to show socially responsible behavior in areas such as the environment, ethics and health (Marquez & Fombrum, 2005; Moir, 2001). CSR is stimulated by calls and criticism from society (McDonnell & King, 2013). These calls are regularly initiated by a minority of people, and often voiced through Non-Governmental Organizations (NGOs). NGOs try to mobilize people worldwide through the use of internet in order to boycott or criticize companies that use practices which they believe should be altered or banned (Baur & Arenas, 2014). By uttering criticism, NGOs aim to solicit moral and responsible behavior from companies. Often firms facing a boycott do respond to this criticism by engaging in CSR (McDonnell & King, 2013). Criticism can damage a company’s reputation and financial performance (Marquez & Fombrum, 2005; McDonnell & King, 2013). Therefore, companies want to regulate these unintended interactions with society (Baur & Arenas, 2014; Murphy & Dee, 1992). One way to do so, is by forming partnerships with NGOs (Jamali & Keshishian, 2009; Murphy & Dee, 1992). Aforementioned partnerships show the willingness of companies to be socially responsible, and they allow NGOs to protect the needs of the stakeholders that they represent, for instance the environment or animals (Betsill & Corell, 2001).

A wide variety of CSR topics exist. In 2012, the Human Research Council found that of all CSR issues, people rated animal welfare and worker rights as their main priority in the U.S. (Human Research Council, 2015). This shows that animal welfare is a relevant topic, especially in industries such as the food and the animal testing industry where animals are a part of the business. Even though on average a Dutch person eats more than 727 animals in their lives (Verhoog, Wijsman, & Terluin, 2015), consumers are more aware now than a decade ago of animal welfare, and are more influenced by food labels and animal welfare criteria, shown on food packaging (Forsman‐Hugg et al., 2013). With a rise of consumer interest in food with ethical and social attributes, firms in the food

(6)

5

supply industry are obliged to be more socially responsible (Costanigro, Deselnicu, & McFadden, 2016).

A partnership between the NGO World Wide Fund of Nature (WWF) and the multinational Unilever, has affected the fishing industry greatly, and improved marine life (Fowler, Penny & Heap, Simon, 1998). Their partnership introduced the Marine Stewardship Council (MSC) label that ensures better fishing conditions to ensure marine life survival. The label is given to fish caught according to MSC conditions (Elkington, 1998).

NGOs not only influence companies directly by criticizing them and by partnering with these firms, but also by influencing governments. NGOs lobby, and sometimes even partner with, political parties in order to try and stimulate the introduction of new laws and regulations (Rosenberg, Hartwig, & Merson, 2008). Governments can influence companies via legislation and by showing commitment towards an issue. Calls from society regarding animal welfare led to the accountability of animal welfare as a fundamental value of the European Union (EU) in article 13 (Treaty on the Functioning of the European Union, 2009). The EU has herewith introduced multiple trade regulations regarding the protection of animal welfare. All eggs imported, for example, must be from hens with a cage size of at least 750cm2 (Van Horne & Achterbosch, 2008).

Another topic of animal welfare is animal testing. NGOs have been fighting to ban animal testing for years. Millions of animals are used for scientific studies, for practice exercises in the health sector and for testing cosmetics and other products (Dignon, 2016). Strikingly, there is a lot of evidence indicating that medicines effective for animals often do not work for humans and these animal tests are frequently unsuccessful. More importantly, there are multiple and cheaper alternatives to animal testing (Balls, 1994; Bottini, 2009). At the moment, two NGOs, namely Animal Rights and An Animal A friend (EDEV), are organizing protests and are raising awareness on social media against the largest ape testing center in Europe, located in Rijswijk, the Netherlands. The goal of these protests is to influence government to introduce laws that forbid animal testing, close the

(7)

6

facility, and to simultaneously influence companies to stop investing in animal testing.

Evidently, animal testing is a very contemporary issue. Although there has been a lot of research on the determinants of general CSR, there is almost no research on the determinants of a firm’s engagement in animal welfare activities, such as animal testing. More specifically, no research has combined the influence of governments and of NGOs on animal testing in companies. Moreover, the effect of the presence of NGOs on companies behavior has also rarely been studied. This study tries to fill parts of that literature gap. First will the literature be reviewed, after which the methodology of this study will be assessed. Then the results will be shown and discussed.

2. Literature review

2.1 Corporate Social Responsibility

There is still no consensus on the definition of Corporate Social Responsibility (Sheehy, 2015). This paper will rely on one of the most popular and the most frequently cited definition of CSR, constructed by the Commission of the European Communities (2001): “A concept whereby companies integrate social and environmental concerns in their business operations and in their interaction with their stakeholders on a voluntary basis” (Employment, 2001). This definition encompasses an umbrella definition incorporating all processes and actions towards social causes. CSR can be separated into five dimensions: the environmental, the social, the economic, the stakeholder and the voluntariness dimension (Dahlsrud, 2008). Regarding animal welfare, this topic is often put as a subcategory of the environmental dimension (Moir, 2001).

There is an enormous amount of literature on the social investments of corporations, many of these studies have examined the effect of CSR on financial performance (Cavaco & Crifo, 2014; Jitaree, Lodh, & Bhati, 2014; Saeidi, Sofian, Saeidi, Saeidi, & Saaeidi, 2015). Presently, pressure and critique from NGOs, discontent employees or distraught consumers, can easily lead to boycotts and negative campaigns against corporations. As mentioned before, these actions can harm the reputation of the firm greatly and thereby they can (also) reduce the revenues (Marquez &

(8)

7

Fombrum, 2005). Regularly, those parties (NGOs, discontent employees and distraught consumers) demand open and transparent reporting from firms about their corporate and social behavior (Forsman‐Hugg et al., 2013).

2.2 Animal welfare, in particular animal testing, in CSR

Forsman‐Hugg et al., (2013) investigated CSR themes that consumers rate as important, and found that animal welfare should be one of the key dimensions of CSR in the food industry. Current research has looked at animal welfare in CSR primarily by performing case studies. These were done mostly in the industries of food and animal testing (Carrero & Valor, 2012; Lindgreen & Hingley, 2003).

Regarding the food industry, multiple studies found that one of the main concerns is animal welfare. An example hereof can be found in the German Pork industry where animal welfare matters. Transport time for the animals was considered to be a significantly more important issue than environmental or social issues such as environmental footprint or employees payments (Hartmann, Heinen, Melis, & Simons, 2013). Nevertheless, research shows remarkable resistance against animal welfare considerations in the food supply chain if the costs of production increase when taking into account the welfare of animals (Bornett, Guy, & Cain, 2003).

Considering the animal testing industry, this industry has to deal with many concerns from society about animal welfare and about the necessity to use animals. Animal testing is no longer necessary for laboratory research due to replacement alternatives such as null mutations, and cosmetics can also be tested differently or not at all if natural ingredients are used (Wolfer, Crusio, & Lipp, 2002). Yet millions of animals are still bred and used for laboratory studies. This is also shown by one study that found that especially for the pharmaceutical industry it is important and beneficial to engage in CSR and to, among other things, be more responsible considering the welfare of testing animals (Smith, 2003). Companies active in animal testing therefore gain a lot of criticism from society. This criticism is sometimes mouthed through NGOs and primarily directed at the company in

(9)

8

an attempt to reduce the number of testing animals (Stoddart & Brown, 2014). For the business, giving in to these whishes from society holds the potential of improving reputation and profitability. However, little research has been done on animal testing and animal testing methods used by companies.

Most of these studies on animal welfare in general but also those specific on animal testing, were performed only in one specific country, such as studying the economic costs of animal welfare in the UK pig production industry (Bornett et al., 2003). Other studies only investigated one specific company, for example one study investigated the influence of improving animal welfare in Tesco’s supply chain (Lindgreen & Hingley, 2003). This current study will investigate a large number of companies, from multiple countries that are also active in different industries in order to study influential factors of the use of animal testing by companies.

2.3 Governmental Legislation and Governmental Commitment

As stated previously, multiple studies have investigated factors that influence CSR. One of these factors is the influence of legislation, such as the EU legislation (Employment, 2001). Governmental Legislation is defined as the combined set of laws and regulations which is agreed upon by society (Campregher & Jeglic, 2016; Detomasi, 2008). Another influence is governmental commitment which has been investigated in several studies (Detomasi, 2008; Dharmaratne, Christopher, Cullen, & others, 2011; Gjølberg, 2012). Governmental commitment exists when a government shows support to a specific issue, and shows willingness to change and improve this issue by providing resources (Jackson, 2015). Detomasi (2008) found that the overall political structure in the home country of a company influences whether that company will pursue a CSR initiative and if so, what kind of initiative. Another study of Dharmaratne et al., (2011) found that when a society faces problems and needs to act as one society, governmental commitment and legislation can lead to more disclosure of CSR, as well as to more consideration about the CSR of individual companies.

(10)

9

shown before with the European regulations regarding cage sizes of egg hens (Van Horne & Achterbosch, 2008). Legislation for all animals is a challenge since there are not only different types of animals (wildlife, farm-, domestic and entertaining animals) to be protected, but also different usages for those animals (food, testing, entertainment, raw materials such as fur and leather) that all need individual legislation and regulations. Each government has a different approach towards the protection of animals and the use of animals for testing purposes. Therefore this study takes on an international approach and studies companies that have their headquarters in different countries across the globe, in order to investigate the influence of governmental legislation and governmental commitment on the use of animal testing by companies. One could assume that with more legislation surrounding the protection of animals, fewer animals are used for testing.

2.4 Governmental Legislation and Commitment on Animal Testing

Governments have the ability to introduce laws and procedures that protect non-human beings. The European Union has made guidelines and schemes trying to stimulate farmers to be more responsible towards their animals, as was shown by the introduction of European regulations regarding cage sizes of egg hens (Van Horne & Achterbosch, 2008). The EU aims to stimulate national governments to install regulations regarding animals (Veissier, Butterworth, Bock, & Roe, 2008). However, there are still plenty of areas that are not covered. It is up to national governments to ensure laws that protect animals. Notably, the degree in which governments are committed to protect animals and to implement legislation varies across countries (World Animal Protection, 2013).

The focus of this thesis is on animal testing. The primary reason to focus on animal testing is due to nature of the subject and the intensity and duration of the debates surrounding animal testing. Millions of animals have been used and are still being used on an annual basis for testing all kinds of products like shower gel, perfume, medicine and medical treatments. The deployment of animals for this purpose, receives a lot of criticism from society and from NGOs like PETA (People for

(11)

10

the Ethical Treatment of Animals) who absolutely dislike this matter (Stoddart & Brown, 2014). Next to this, there has been a lot of literature on alternative options for animal testing in an attempt to stop this practice (Balls, 1994; Schuppli, Fraser, & McDonald, 2004; Wolfer et al., 2002). Moreover, animal testing is a multimillion dollar industry, the costs of experiments on animals are high, higher than costs of using alternative testing methods (Bottini, 2009). Regardless of criticism, the high costs, and the availability of alternatives that do not involve animal cruelty, animal testing is still a common practice (Bottini, 2009).

Some governments have shown to be committed to improving the issue of animal testing. Thereby introducing legislation that improved the situation of testing animals, such as the Cosmetic Products Regulation that forbids cosmetic products of which even a single element is tested on animals after 11 march 2009 (Heringa, de Wit-Bos, Bos, & Hakkert, 2015). Simultaneously, other legislation has increased the amount of animal testing, for example the European Law that all chemicals used in households must be tested (Heringa et al., 2015; Hofer et al., 2004). One study found that even the majority of researchers that are involved in animal testing dislike the matter and would prefer other methods (Dignon, 2016). Some scientists have even formed an NGO called ECOPA, the European Consensus Platform on Alternatives, that strives to fight the introduction of legislation that is not in the favor of testing animals and of the introduction of alternatives to animal testing (Hofer et al., 2004).

Governments can introduce legislation that either impair or improve the condition of animal testing. However, research on whether Governmental Legislation and Commitment of governments towards a certain issue also affects the behavior of companies regarding animal testing has not been performed yet. In order to study the influence of governments, an international approach is necessary to compare and contrast the various degrees in which governments are committed towards the protection of animals and have introduced legislation regarding animal testing.

(12)

11 2.5 Partnerships between companies and environmental NGOs and specialized parties

Nowadays, companies get more involved with stakeholders. It is important to notice that not only consumers are stakeholders of companies who are directly affected by the behavior of these companies but, the environment and animals are stakeholders as well (Smith, 2003). The interests of the environment and animals are, as previously stated, represented by NGOs and organizations specialized in environmental issues (referred to after this as specialized parties). As said before, protest-based partnerships initiated by NGOs and specialized parties can harm the legitimacy of a company (Guay, Doh, & Sinclair, 2004). Therefore, partnerships initiated by the company with NGOs and specialized parties can be far more beneficial for a company (Nijhof, de Bruijn, & Honders, 2008). Such partnerships are also beneficial for NGOs since these partner companies will take into account the concerns of the NGO and help to protect the stakes defended by the NGO (Jamali & Keshishian, 2009; Elkington, 1998).

These partnerships are seen as an influential factors of CSR (Wikström, Nilsson, & Agnesson Franzén, 2010). One study examined the influence of the World Wide Fund for Nature (WWF), on the carbon footprint of companies in India (Wikström et al., 2010). Others studies investigated how these partnerships can benefit both the company and society (Jamali & Keshishian, 2009; Dahan, Doh, Oetzel, & Yaziji, 2010; Damlamian, 2006; Elkington, 1998; Perez-Aleman & Sandilands, 2008). They found that by partnering with NGOs and specialized parties, companies are able to show cooperation and willingness to improve social causes. Such partnership decreases the risk of negative publicity for companies (Nijhof et al., 2008).

Possibly even more important is the impact that a partnership between companies and NGOs and specialized parties can have on an entire industry. For example, one study looked at the partnership between Starbucks and the Conservation International (CI), involving the use of Fairtrade coffee beans. Starbucks agreed with the CSR initiative of the CI to only buy sustainable and ethical coffee beans. As a result of this partnership other multinationals and companies followed. Hence,

(13)

12

this partnership has a positive effect on social ethics and the environment in this coffee industry (Perez-Aleman & Sandilands, 2008). Consequently, partnerships between companies and NGOs and specialized parties can have an effect on the behavior of the company, as well as on an industry. Therefore, this study will also look at the effect of such partnerships.

2.6 Environmental NGO presence

Having either protest-based relationships or partnerships with a firm, might not be the only way in which NGOs can have influence on animal testing procedures of a company. NGOs might also have indirect and unintended effects on the CSR of companies. This might be through simply being present as an NGO. With this is meant that companies active in countries where there are a lot of environmental NGOs might be influenced more by NGOs than companies active in countries with only a few NGOs. Several studies found out that activism, also in the form of NGOs, affect non-targeted companies, by creating investment uncertainty and by altering perceptions of market opportunity (Briscoe & Gupta, 2016; Briscoe, Gupta, & Anner, 2015). If an NGO targets one company and negatively influences the material and reputational resources of that company, other companies might feel the threat that this will also happen to them (Briscoe & Gupta, 2016). Furthermore, one study found out that the mere presence of environmental NGOs affects the transparency and scrutiny of corporate disclosure (Marquis, Toffel, & Zhou, 2016).

As said before, NGOs represent stakeholders of companies, such as the environment and animals. Often NGOs gain awareness by initiating actions and addressing the public, government and companies (Arenas, Lozano, & Albareda, 2009). Therefore, it is likely that companies incorporate the wishes and interests that these NGOs defend, even by not directly partnering with these NGOs, most likely out of fear (Arenas et al., 2009; Briscoe & Gupta, 2016; D. Jamali & T. Keshishian, 2009). In order to prevent protests and negative actions from these NGOs, firms might improve their CSR beforehand. Even though several studies have investigated the indirect effect of activism on companies, almost no study incorporated the influence of NGO presence on CSR topics such as

(14)

13

Animal Testing. Filling this gap in literature, could provide suggestions and support for the effectiveness of NGOs, both on government and on companies.

2.7 Conclusion

To review, animal welfare has been found in research to be an important topic of CSR, especially in industries closely related to animal welfare like animal testing. Governmental commitment and legislation has been linked to CSR. In particular on animal testing governments can influence animal welfare by introducing new laws and by showing their commitment towards an issue. NGOs influence companies by starting protest-based relationships, but also by partnering with them. Additionally, NGOs also have indirect and sometimes unintended effects on companies. Companies surrounded by multiple NGOs might be, even though they are not targeted by these NGOs, affected by these NGOs more than companies that are not surrounded by NGOs.

There is an gap in literature regarding these topics. First of all, no quantitative study has been done on the relationship between governmental influence and animal testing in corporations, and on partnerships between companies and NGOs and animal testing. Very little research has been done on the effect of NGO presence on governments and on companies. This study, will attempt to fill these literature gaps by studying governmental and NGO influence on animal testing in companies. Afterwards recommendations will be provided on a societal level.

This study attempts to answer two research questions. The first question is: What is the influence of governmental commitment and governmental legislation on animal testing and protection in companies? The second research question is: Can NGOs and specialized parties influence the relationship between governmental influence (commitment and legislation) and animal testing, by partnering with companies and by being present?

(15)

14 3. Theoretical Framework

This section will provide the four hypotheses that are formed to answer the two research questions stated in the previous paragraph. The first research question respecting the relationship between governmental influence and animal testing in companies will be theorized into two hypotheses. After this, the second research question of this paper, regarding the moderating influence of NGOs on the relationship between governmental influence and Animal Testing, will be explained, by also reframing it into two hypotheses.

As said before, this study will fill the gap in current literature towards governmental influence on animal testing. Governments are influenced by demands from society. These demands often lay the foundation for Governmental Commitment towards a new issue. Governmental Commitment refers to a government that is dedicated towards an issue and provides resources to the protection and benefit of that certain issue. Governmental Commitment is useful for interpreting and forecasting future legislation (Dharmaratne et al., 2011). Governmental Commitment is especially important when the law is not explicit and corporations have to figure out their own moral guidelines in relation to the law. In these situations, companies are addressed by society on their own moral authenticity (Adams, 2013). As a result, for companies it might be a good strategy to anticipate future legislation based on Governmental Commitment and on demands from society (Dharmaratne et al., 2011). Governmental Legislation influences companies by providing laws and regulations, which companies have to obey. Consequently, companies that are aware of the areas in which their government is committed will most likely be in a better position to alter their strategy and CSR plans to gain a benefit and meet social demands. With regard to animal testing, governments across the globe show different degrees of commitment and have introduced different legislation (Animal Protection Index, 2015).

It is expected in this study that there is a negative relationship between Governmental Legislation, with this is meant legislation that regulates animal welfare and welfare conditions of

(16)

15

testing animals, and the number of companies that use Animal Testing. Hence, it is expected that situations in which governments protect animals with legislation, meaning laws and regulations, will relate to lower numbers of companies involved in Animal Testing. Moreover, it is also expected that there is a negative relationship between Governmental Commitment towards the protection of animals and Animal Testing in companies. Similar to Legislation, it is expected that situations in which governments express their commitment towards the protection of animals are related to less companies involved in Animal Testing. Therefore the first two hypotheses are as follows:

Hypothesis 1. Governmental Legislation on animal welfare is negatively related to reported Animal Testing in companies , in such a way that more Legislation is related to lower numbers of companies using Animal Testing.

Hypothesis 2. Governmental Commitment towards the protection of animals is negatively related to reported Animal Testing in companies, in such a way that more Commitment is related to lower numbers of companies using Animal Testing.

Moreover, to answer the second previously given research question, this thesis also looks at the possible moderating effect of NGOs on the relationship between governmental influence (Commitment and Legislation) and Animal Testing in firms. More specifically, the partnering of firms with NGOs and specialized parties and the presence of NGOs will be considered in their possible moderating effect.

Firstly, the effect of partnerships between companies and environmental NGOs and specialized parties. As described in the previous section, firms constantly face the risk of being negatively judged by the public. This critique is often voiced through NGOs and specialized parties (Baur & Arenas, 2014; Jamali & Keshishian, 2009; Murphy & Dee, 1992). A good strategy for companies and especially for companies which are active in multiple countries, is to partner with NGOs and specialized parties in order to meet moral demands from society (Dahan et al., 2010;

(17)

16

Nijhof et al., 2008). The environment and animal welfare are often interdependent and related to each other. Those two topics are regularly combined goals for NGOs and improvements in the environment are almost always accompanied by improvements in animal welfare. This is shown for example in a study that aims to improve the welfare of one species of fish, namely the finfish, by improving environmental conditions of its habitat (Bergqvist & Gunnarsson, 2013). Another example is the extinction of animals due to deforestation (Brook, Sodhi & Ng, 2003). Notably, as stated in the introduction, animal welfare is rarely defined as an independent topic in CSR, but often seen as a subtopic of environment (Moir, 2001). Similarly, most NGOs and specialized parties that strive for animal welfare are registered and known as environmental NGOs. Examples of such NGOs are: the World Wide Fund for Nature (WWF), African Wildlife Foundation (AWF), and Greenpeace. These NGOs protect animals by protecting the environment animals live and exist in. For example, the attempt to reduce poaching in African national parks by AWF, but also by directly addressing animal welfare topics, such as the ‘No animal testing’ campaign of Greenpeace. Therefore, partnering with environmental NGOs and specialized parties is very likely to also address animal welfare issues. Consequently, companies willing to partner with NGOs about environmental issues and animal welfare, are expected to include these topics in their CSR reports (Arenas et al., 2009; Wilkins et al., 2005). Therefore, it is expected that partnerships with environmental NGOs and specialized parties leads to less animals are used for testing. Therefore the third hypothesis is as follows:

Hypothesis 3a. Partnership with environmental NGOs and other specialized parties positively moderates the relationship between Governmental Legislation and Animal Testing in companies, such this relationship is stronger for higher levels of Partnerships.

Hypothesis 3b. Partnership with environmental NGOs and other specialized parties positively moderates the relationship between Governmental Commitment and Animal Testing in companies, such this relationship is stronger for higher levels of Partnerships.

(18)

17

As said previously, NGOs might not only have a direct influence through partnering with firms but might also have an influence on animal testing by simply being present. Very little research has been done on the influence of NGO presence, however multiple studies were performed on the indirect influence of activism. As mentioned before, activism indirectly affects companies by shaping perceptions of market opportunities and by affecting resources of companies (Briscoe & Gupta, 2016). One study was conducted to determine the best practices for NGOs in order to have the most influence on the UN. They found that being close as an NGO to where decisions are made provides the opportunity to influence the final decision making (Tallberg, Dellmuth, Agné, & Duit, 2015). Another study done by Marquis, Toffel and Zhou (2016) looked at NGO presence when it comes to the disclosure of companies and the level of scrutiny in disclosing information from those companies. They included NGO presence per million of population for each country. They found that firms who cause more environmental damage, are less prone to selective disclosure when there is a greater NGO presence in their country (Marquis et al., 2016). All these studies suggest that NGO presence has an indirect influence and effect on a firm’s behavior. It is interesting to see whether the presence of environmental NGOs has an influence on the relationship between governmental influence and Animal Testing. If so, it might be beneficial for social activists to initiate more NGOs in close vicinity to companies. This leads to the fourth and final hypothesis:

Hypothesis 4a. Environmental NGO presence positively moderates the relationship between Governmental Legislation and Animal Testing in companies, such that this relationship is stronger for higher levels of environmental NGO presence.

Hypothesis 4b. Environmental NGO presence positively moderates the relationship between Governmental Commitment and Animal Testing in companies, such that this relationship is stronger for higher levels of environmental NGO presence.

(19)

18

Figure 1. This framework will be studied and the results will be stated and discussed in the results and discussion section.

Figure 1. Conceptual model of the relationship between the two independent variables and the two moderator variables on the independent variable animal testing in companies.

(20)

19 5. Methods

5.1 Sample and data collection

The overall sample is based on international companies gathered from the Thomson Reuters ASSET4 database. Information about this database will be provided later. The data for the dependent variable, one control variable and the moderator variable Partnership is retrieved from this database. This database investigated the same companies as the Thomson Reuters WorldScope database. Two other control variables were selected from the WorldScope database. WorldScope and ASSET4 were matched to each other based on the DataStream ID number that is given to each company. The sample was filtered by deleting all cases with missing values. Afterwards, the sample was filtered even further by retrieving the home country of headquarters of every case from WorldScope, and matching those to the 50 countries selected by the Animal Protection Index (API). These 50 countries were selected by the API because these are the 50 largest producers of meat, milk and eggs; based on the Statistical Yearbook of World and Agriculture 2012, published by the Food and Agriculture Organization of the United Nations. The API is used for data for the independent variables Governmental Legislation and Governmental Commitment. All datasets were from the year 2015. The final sample consists of 2715 companies who were selected and measured in both ASSET4 and WorldScope and all have their companies’ headquarter in 29 of the 50 countries on which API provides data. Ideally the sample mirrors societal demography, in order to be able to provide recommendations on societal level.

5.2 Independent variables: Governmental Legislation and Commitment

The independent variables in this study are the degree in which (testing) animals are protected by Governmental Legislation and the degree in which there is Governmental Commitment towards the protection of animals. The Animal Protection Index (API), obtained by the Non-profit Organization World Animal Protection, has provided data for the independent variables Governmental Legislation and Governmental Commitment is. As said before, the API measured animal protection for those 50

(21)

20

countries that are the biggest milk, egg and meat producers. The World Animal Protection (WAP) organization is active on a global level striving for animal welfare by providing direct help to animals, by providing education to people and by lobbying for Governmental Legislation. The WAP indicates with its Animal Protection Index those countries that still need a lot of lobbying by social activists to get better legislation and policies. The API has given each of the 50 countries scores on 15 indicators that were grouped into five separate themes. Two of these themes, namely ‘recognizing animal protection’ and ‘governance structure and systems’, are used for this study to collect data for the independent variables Governmental Legislation and Governmental Commitment. The API measured the relevant items on a 7-point Likert scale, where G is very little protection or commitment towards of animal welfare and testing animals from government and A represents the best protection or commitment for animals and testing animals. In order to be able to analyze the data, these letters are converted to scores of 1 to 7 with A being 7 and G being 1

The first theme, ‘recognizing animal protection’ has three items that measure Governmental Legislation regarding the protection of animals. These three items are used in this study to get a summary score for Governmental Legislation. The first item is whether animal sentience, e.g. the notion that animals have feelings and are able to feel pain, is formally recognized in legislation and/or policy. The second indicator measures the degree in which there are animal protection laws that prohibit causing animal suffering either by a deliberate act of cruelty or by a negligence to act. The third indicator is the degree to which animals that are used in scientific research are protected by Governmental Legislation. This is a relevant item since the dependent variable is whether companies use animal for testing. For this third item, the World Animal Protection looked at the amount of suffering and pain and the incorporation of the three R’s principle of ‘replacement refinement and reduction’ by animal testing laboratories. These three R’s represent the intention to provide a better life to lab animals after they have finished their task (replacement), to improve their life conditions (refinement) and to reduce the number of lab animals used (reduction). These three

(22)

21

items will be made into one summary score by adding the scores up per country and dividing them by three, as was done by Creti, Fichten, Amsel, Brender, Schover, Kalogeropoulos and Libman, (1998). This will be the independent variable Governmental Legislation.

The second independent variable used is governmental commitment, obtained from the second theme ‘governance structure and systems’. This item is also measured on a 7-point Likert scale. This one item measures whether government has assigned responsibility and accountability for improving animal protection through Legislation and has provided resources for the protection of animals, including testing animals.

5.3 Dependent variable: Animal Testing

For the dependent variable of this empirical quantitative study, data obtained from Thomson Reuters ASSET4 is used. Thomson Reuters measures 900 evaluation points of environmental, societal and governance (ESG) performance. Thomson Reuters ASSET4 primarily provides objective investment analysis tools by analyzing portfolios and ESG data, rather than let the companies insert surveys themselves. The evaluation points needed for this study are found under the section “Environmental factors” of the Thomson Reuters ASSET4. The environmental section consists of items on CO2 emission, waste and two items on Animal Testing. These two items are used for this study. The two items are ‘whether the company mentions animal testing in their CSR report’, and ‘whether the company has established animal testing reduction initiatives in the CSR report’. Both were measured in Yes and No. These two indicators are combined into an index number to get one dependent variable named Animal Testing. All companies that scored ‘yes’ on the animal testing reduction initiatives also score a ‘yes’ on whether they use animal testing. Therefore, if a company scored ‘yes’ on either one of the items, or on both items, they are given a 1; otherwise they are given a 0, making it a binary outcome variable.

(23)

22 5.4 Moderator variables: NGO partnerships and NGO presence

Two moderator variables are used in this study: partnership with NGOs and specialized parties; and Environmental NGO presence. The first moderator variable, Partnership with NGOs and specialized parties, is also provided by the previously described Thomson Reuters ASSET4 dataset. It is the item “Does the company report on partnerships or initiatives with specialized NGOs, industry organizations, governmental or supra-governmental organizations that focus on improving environmental issues?” The item is measured as a binary variable, and is answered by companies with either ‘Yes’ or ‘No’. The variable was dummy coded, and Yes = 1 and No = 0 .This variable is used to answer the third hypothesis, whether a partnership between companies and NGOs and other specialized parties moderates the relationship between Governmental Influence (Commitment, Legislation) and Animal Testing.

The second variable, Environmental NGO presence, was derived from dividing the population of each of the 29 home countries of the headquarters of all companies in this sample, for the year 2015 (retrieved from World bank), by the number of Environmental NGOs associated with the International Union for Conservation of Nature (IUCN) in that country. The outcome of this is the number of environmental NGO per million of population per country. Some countries have a high number, meaning that companies headquartered in those countries are surrounded by relatively more environmental NGOs than companies headquartered in countries with low number of NGO presence. This method was based on the study of Marquis et al., (2016). This variable is used to answer the fourth hypothesis, about the moderation effect of NGO presence on the relationship between Governmental Influence (Commitment & Legislation) and Animal Testing.

All continuous predictor variables, Governmental legislation, Commitment and environmental NGO presence, were mean-centered prior to computing their interaction term and prior to model estimation to improve interpretation of regression coefficients (Fairchild & MacKinnon, 2009). The computed interaction terms are Legislation*NGO presence,

(24)

23

Legislation*Partnership, Commitment*NGO presence and Commitment*Partnership. This mean-centering allows for a better interpretation of the results.

5.5 Control variables: Size, Profitability, Industry & Location

Multiple items have been used as control variables in this study. All of them were extracted from a third database namely Thomson Reuter’s WorldScope. WorldScope measuresthe same companies as ASSET4 but measures more financial variables. The control variables extracted from this database are Industry, Profitability (Measured in Return of Assets (ROA)) and Size (measured as Total Assets). The fourth control variable is Location, derived from the location of headquarters of each company.

Industry is used as a control variable, since some industries, such as food and drug retailing, are more involved in animal welfare than others, such as banking. The companies used in this study are active in 54 different industries, see table 1, panel C. Before analysis, industry was dummy coded. Return on Assets (ROA) is chosen to measure the profitability of the company (Juan García-Teruel & Martinez-Solano, 2007). Companies that are more profitable are more likely to have more money and resources to spend on CSR themes, such as animal welfare, than companies that are less profitable or not at all. Previous research has already identified profitability as both an indicator and an outcome of CSR (A. McWilliams & Siegel, 2001; van Beurden & Gössling, 2008).

Total Assets is used as a measure for company size. Total Assets is an indication of all assets that are bought and in control of a company. Size is chosen as a control variable since larger companies might be more influenced by governmental legislations and by the public opinion and criticism than smaller companies.

Finally, the continents in which a company is headquartered is chosen as a measure for the fourth and final control variable; Location. Companies are settled across the globe. Location can have an influence on Animal Testing, on Governmental Legislation and on Governmental Commitment. Therefore, it is important to control for location. This variable was dummy coded. As can be seen in Panel B, from table 1, all continents are represented by at least 100 companies. North-America is

(25)

24

represented by the most companies (965), followed by Europe (608) and Asia (607).

Country N Country N

(AT) Austria 8 (IT) Italy 31

(AU) Australia 304 (JP) Japan 328

(BR) Brazil 63 (MA) Morocco 2

(CA) Canada 209 (MX) Mexico 27

(CH) Switzerland 39 (MY) Malaysia 37

(CL) Chile 16 (NL) Netherlands 19

(CN) China 71 (NZ) New Zealand 18

(DE) Germany 58 (PH) Philippines 20

(DK) Denmark 23 (RU) Russia 25

(EG) Egypt 8 (SE) Sweden 33

(ES) Spain 27 (TH) Thailand 26

(FR) France 71 (TR) Turkey 19

(GB) Great Britain 280 (US) United States 756

(ID) Indonesia 30 (ZAF) South Africa 97

(IN) India 70

Total 2175

Panel B. Sample distribution across Continents

Continent

N

Continent

N

Europe

608 North-America

965 Oceania

322 Asia

607 South America

106 Africa

106

Panel C. Sample distribution across industries

SIC code - Industry N SIC code -Industry N

10 Aerospace & Defence 10 23 33 Insurance 104

11 Automobiles & Auto Parts 59 34 Investment Banking & Investment Services

79

2 Banking Services 186 35 Leisure Products 17

Table 1

Descriptive statistics

(26)

25

12 Beverages 37 9 Machinery, Equipment &

Components

119 13 Biotechnology & Medical Research 24 36 Media & Publishing 79

8 Chemicals 82 4 Metals & Mining 180

5 Coal 23 37 Multiline Utilities 18

14 Collective Investments 14 38 Natural Gas Utilities 13

15 Communications & Networking 18 39 Office Equipment 9 16 Computers, Phones & Household

Electronics

15 1 Oil & Gas 138

17 Construction & Engineering 62 40 Oil & Gas Related Equipment and Services

65 18 Construction Materials 24 41 Other Specialty Retailers 76 19 Containers & Packaging 20 42 Paper & Forest Products 11 20 Diversified Retail 39 43 Passenger Transportation Services 38 21 Diversified Trading & Distributing 10 44 Personal & Household Products &

Services

42

22 Electric Utilities & IPPs 73 45 Pharmaceuticals 64

23 Electronic Equipment & Parts 9 46 Professional & Commercial Services 98 24 Food & Drug Retailing 35 3 Real Estate Operations 68

25 Food & Tobacco 81 6 Renewable Energy 9

26 Freight & Logistics Services 37 47 Residential & Commercial REITs 111 27 Healthcare Equipment & Supplies 41 48 Semiconductors & Semiconductor

Equipment

36 28 Healthcare Providers & Services 32 49 Software & IT Services 110

29 Holding Companies 7 50 Telecommunications Services 66

30 Homebuilding & Construction Supplies

43 51 Textiles & Apparel 24

31 Hotels & Entertainment Services 76 52 Transport Infrastructure 27

7 Household Goods 17 53 Uranium 6

32 Industrial Conglomerates 15 54 Water utilities 6

Total 2175

5.6 Statistical Analysis

The aim of this study is to find out if there is a relationship between Governmental Legislation and Governmental Commitment and whether companies are involved in Animal Testing, and if so, whether this relationship is moderated by the willingness of the company to partner with NGOs and organizations specialized in environmental issues and by companies being in close vicinity of environmental NGOs. Therefore, regression analyses are performed, in order to determine a relationship. The data will be analyzed with the Statistical software Package for Social Sciences (SPSS). More specifically, a binary logistic regression analysis will be performed in order to study the hypotheses. A binary logistic model is a type of logistic regression in which the dependent variable is binary (Bachand-Marleau, Lee, & El-Geneidy, 2012).

(27)

26 6. Results

6.1 Summary statistics

The descriptive statistics for the entire sample are represented in Table 2. As can be seen in Table 2, the dependent variable Animal Testing is binary, as well as the moderator variable Environmental Partnership. Around 38 percent of the sample has environmental partnerships (N = 1023). Additionally, half of the companies is located in a country where there is half an environmental NGO per million of inhabitants (M = .34). This indicates that 1358 companies are located in a country with little environmental NGO presence.

Two control variables were right skewed, namely Profitability and Size. A q-plot indicated a non-normal distribution for Profitability. The Kurtosis test confirmed this since the kurtosis value is 63.94. After a lognormal transformation, using the formula using Y = Log10 (X + 1), Profitability was normally distributed with a kurtosis value of .22. Although there is still an outlier, this outlier was inspected and found appropriate since it reflects a very profitable company. After log transformation, Profitability has a mean of -1.31 (SD = .39), see Table 2. Although this implies that many companies are making a loss, this should be interpreted carefully since most of the companies had profitable number before transformation of this variable.

Moreover, Size did not follow a normal distribution with a kurtosis value of 876.85 and a skewness of 27.81, indicating a right skewed distribution. After a lognormal transformation, using again the formula using Y = Log10 (X + 1), Size was normally distributed (Kurtosis = -.06, skewness = .29). After lognormal transformation, the mean of Size is 7.14 (SD = 1.24). A 1-point difference, is on average a difference of *10 (before log transformation). This means that the mean of 7.14 equals a Total Assets of € 13,752,000.00. Evidently, the sample contains quite large companies since these companies have assets worth millions of dollars.

(28)

27

companies namely; Banking services, Metals & Mining and Oil & Gas. However, they represent only approximately 23 percent of the companies.

The control variables Size and Industry are only correlated to other control variables, see Table 3. Location is also correlated to all the independent variables. Logically Governmental Legislation and Commitment and NGO presence are all computed based on countries, hence this is related to continents. Some locations have higher Governmental Influence and NGO presence than others. Finally, Profitability is correlated to almost all variables.

When looking at the scores on the three independent variables Governmental Legislation, Commitment and the interaction term, all three items were scored high. This means that the most of the sample is headquartered in countries of which government has implemented several laws and regulations in order to protect the welfare of animals (M Legislation = 5.39, SD = .86), and most of these governments are committed to the protection of animals (M = 5.72, SD = 1.01). Therefore, it would be logic to assume that the interaction term of Commitment and Legislation would also be on the high side. This is supported by the findings of a mean of 31.38 (SD = 9.12).

Table 2

Summary statistics

Variable N Mean Median Std. Dev. Min. 25th Perc. 75th Perc. Max.

Animal Welfare 2715 .04 0 .20 0 0 0 1 Company Size 2715 7.14 6.89 1.24 3.61 6.25 7.89 12.00 Profitability 2715 -1.31 -1.30 .39 -2 -1.52 -1.05 0.37 NGO Partnership 2715 .38 0 .48 0 0 1 1 NGO presence 2715 .54 .37 .47 .02 .20 .69 2.31 Governmental Legislation 2715 5.39 5 .86 2 5 6 7 Governmental Commitment 2715 5.72 6 1.01 1 6 6 7

(29)

28

Table 3

Pearson Correlations between all variables

Variables 1 2 3 4 5 6 7 1. Animal Testing 1 2. Profitability .11** 1 3. Size -.02 -.06** 1 4. Governmental Legislation .06** .09** -.15** 1 5. Governmental Commitment .08** _.05* _-.08** _.63** ₁ 6. Environmental Partnership(1) .09** -.09** .02 .02 .07** 1 7. NGO Presence .02 .08** -.16** .58** .27** -.01 1

6.2 Binary Logistic Regression

In order to test the hypotheses, a hierarchical binary logistic regression analysis was performed in SPSS (Peng, Lee, & Ingersoll, 2002). This test was selected since the outcome variable of this test, Animal Testing, is a binary variable; measured as either ‘1’ (indicating that a company uses Animal Testing) or ‘0’ (indicating that a company does not use Animal Testing). An binary logistic regression analysis allows to measure relationships and effect of categorical or continuous variables on a binary outcome variable (Peng et al., 2002). The control, independent and moderator variables were entered into the model in seven steps. The first step entailed entering the four control variables, Size, Profitability, Industry and Location. Step 2 involved entering the control variables and the first independent variable, namely Governmental Legislation. This step is used to analyze hypothesis 1. Step 3 also included the first moderator variable which is Partnerships between companies and environmental NGOs and its interaction term with the independent variable Governmental Legislation. This step investigates hypothesis 3a. The second moderator variable, environmental NGO presence, is entered in step 4, along with its interaction term with the independent Governmental Legislation, while simultaneously the first moderator variable is removed. Therefore, step 4 studies

(30)

29

hypothesis 4a. Step 2, 3 and 4 were repeated for the second independent variable, Governmental Commitment. Thus step 5, 6 and 7 study respectively hypotheses 2, 3b & 4b. See Figure 2 for an overview of the steps.

Figure 2. Representation of the steps in which variables were entered into the logistic regression. The results of the regression analysis are presented in Table 4. To start off with the control variables, the control variable profitability has a significant effect on Animal Testing, B = 1.75, SE B = .29, p < .01. The Odds ratio indicates that profitable firms are 5 times more likely to have Animal Testing than less profitable firms. Therefore, firm profitability is positively related to animal testing, suggesting that bigger firms are related to more animal testing, which is reported in CSR reports.

The other control variables have no effect on the dependent variable.

The first hypothesis is supported by the binary logistic regression analysis. This means that Governmental Legislation has an effect on Animal Testing in companies

(B = .49, SE B = .24, p <

.05). Nagelkerke’s pseudo R

2

changes by .01 ( p < .01) when Governmental Legislation is added to the control variables, see model 2.

This result indicates that more Governmental Legislation

regarding the protection of animals is related to more companies using Animal Testing. This

relationship was expected to be negative rather than positive. Therefore, the first hypothesis is

(31)

30

Table 4

Results logistic regression analysis

(32)

31

The second hypothesis investigates the relationship between Governmental

Commitment and Animal testing in companies. Governmental Commitment has a main

effect on Animal Testing, B = .53, SE B = .19, p < .01, see model 5 in Table 4. This indicates

that more Governmental Commitment regarding the protection of animals is related to

more companies using Animal Testing. Also the Odd Ratio supports this, see Table 4,

depicting that high Governmental Commitment is related around 1.7 times more to Animal

Testing than low Governmental Commitment. This was not expected, since it is a positive

rather than a negative relationship. Hence, the second hypothesis is rejected.

In steps 3, 4, 6 and 7 the moderator variables were introduced, along with their

interaction term with the independent variables. Both moderator variables have a main

effect on Animal Testing. First of all, Partnering with environmental NGOs and other

specialized parties, has a main effect on Animal Testing, B = -1.23, SE B = .24, p < .01. This

indicates that when a firm partners with environmental NGOs and other specialized parties

this is related to a lower number of companies using Animal Testing. Also when looking at

the Odds Ratio in model 7, partnerships between companies and NGOs is related to 2.81

times less Animal Testing then when there is no such partnership.

Next to this, the results indicate an absence of a moderator effect of Partnership with

Governmental Commitment as well as with Governmental Legislation. This means a rejection

of hypothesis 3a and hypothesis 3b. The variable Partnerships between NGOs and

companies only act as an independent variable with a main effect but does not have any

moderation effects in this model.

Secondly, when entering NGO presence in model 7, a main effect of environmental

NGO presence on Animal Testing, B = 1.03, SE B = .40, p < .01 was found. This means that

(33)

32 Additionally, the interaction term between NGO presence and Governmental Legislation,

entered in step 4 and displayed in model 4, is significant B = .44, SE B = .50, p < .05.

Hypothesis 4a is thereby accepted, meaning that the number of environmental NGO’s in a

country in which a company is active strengthens the relationship between Governmental

Legislation that protects animals and the number of companies involved in Animal Testing.

In order to interpret this moderation effect, an interaction plot was made, see Figure 3.

Figure 3 reveals that the strongest positive association between Governmental Legislation

and Animal Testing occurs when there is a lot of NGO presence. The weakest association is

seen for companies surrounding by only little environmental NGOs. Therefore, the presence

of environmental NGOs strengthens the relationship between Governmental Legislation and

the use of Animal Testing by companies.

Hypothesis 4b is rejected however, due to a non-significant moderator effect of NGO

presence on the relationship between Governmental Commitment and the number of

companies using Animal Testing.

Figure 3. Simple Slope analysis, where NGO presence acts as a moderator on the

(34)

33 7. Discussion

7.1 Answering research questions

This study investigated the relationship between Governmental Legislation and Commitment

and Animal Testing in corporations. Furthermore, this study looked at the effect of the

presence of environmental NGOs and of partnerships between companies and

environmental NGOs and specialized parties. It was expected to see lower numbers of

Animal Testing in situations where there is a high degree of Legislation and in situations in

which there is a high degree of Commitment, in comparison to situations with a low degree

of Legislation or a low degree of Commitment. Moreover, it was expected that both

moderator variables, NGO presence and Partnership, would strengthen the relationship

between Governmental influence and Animal Testing. Not all of these expectations were

fully supported by the findings of this study.

First of all, this study found a main effect of Governmental Legislation on Animal

Testing. However, it was contrary to expectations since this study shows that Legislation is

related to more Animal Testing in companies, instead of less. Nevertheless, this study

supports the findings of Van Horne & Achterbosch (2008) and of Veissier et al., (2008), that

EU regulations can impact and affect CSR behavior in companies.

Next to this, Governmental Commitment also has a positive main effect on Animal

Testing. This can be interpreted in such a way that more commitment from a government

towards animal welfare is related to more Animal Testing.

A likely explanation for this positive rather than negative relationship between

Governmental Legislation, Governmental Commitment and Animal Testing could be that

(35)

The influence of governments and NGOs on animal testing in companies