National cultural impact of MNEs´
home country on human rights violations
Kathrin Fürk – 11085681
MSc Business Administration: International Management University of Amsterdam
Supervisor: Dr. Michelle Westermann-Behaylo Second Reader: Dr. Niccolò Pisani
Statement of Originality
This document is written by Student Kathrin Fürk 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.
Abstract
The role of multinational enterprises has changed during the last centuries due to a constantly growing degree of international activities, a certain shift of power from the state to MNEs and increasing scale of global corporate networks (Ruggie, 2013). This study addresses this fact and challenges arising for the MNE itself, as well as for the state and international organisations to establish rules of the game in order to protect the less fortunate and their human rights. The focus of this work is to investigate the impact of national culture of the MNE´s home country, measured by Hofstede´s dimensions, on the presence of human rights violations. Culture constitutes one of the most widely used concepts in international management according to Caprar et al. (2015). Additionally, Franke and Nadler (2008) as well as Christie et al. (2003) both find significant support for the negative impact of power distance, individualism and uncertainty avoidance on ethical behaviour. Furthermore, human rights awareness, measured by the presence/absence of a human rights policy, is introduced as a moderator, as codes are hypothesised to reduce the occurrence of incidents and increase the ethical performance of a company (Erwin, 2011; Kaptein, 2004). In order to test the hypotheses, a sample was selected based on Sustainalytics what resulted in a final sample size of 58 MNEs whose data was collected from 2010 to 2013 for all applied variables. To my knowledge, no scholar has investigated the impact of national culture on human rights violations before. Although the results of the statistical tests are not significant or opposed the hypothesised effect, this study provides different insights and gives valuable suggestions for future research.
Keywords: multinational enterprises; culture; Hofstede; human rights; human rights awareness; codes
Table of Content
Introduction ... 1
Literature Review ... 4
Conceptualisation and Measurement of Culture ... 4
Human Rights and Guiding Principles ... 7
Role of Multinational Companies and Human Rights Violations ... 9
MNEs´ Human Rights Awareness and Codes of Conduct ... 11
Identifying the Research Gap and Research Questions ... 14
Theoretical Framework ... 16
Human Rights Violations ... 16
Culture ... 17
Culture and Human Rights Violations ... 17
Companies´ Human Rights Awareness as a Moderator ... 19
Data and Methods ... 22
Sample and Data Collection ... 23
Measures ... 24
Dependent Variable ... 24
Independent Variable ... 25
Moderating Variable ... 27
Control Variables ... 27
Statistical Analysis and Results ... 28
Descriptive Statistics ... 28
Testing of the Main Effects ... 30
Testing of the Moderation Effects ... 33
Discussion ... 35
Academic Relevance ... 36
Limitations and Suggestions for Future Research ... 37
Conclusion ... 39
References ... 41
Appendices ... 46
Appendix 1: Complete list of MNE sample applied in this study ... 46
Appendix 2: Classification tables of binary logistic regression tests ... 47
Appendix 3: Results of PROCESS Model 1 for power distance and uncertainty avoidance ... 48
List of Figures Figure 1: Conceptual Model ... 22
Figure 2: PROCESS Model 1 (Hayes, 2013)... 33
List of Tables Table 1: Sample Overview ... 24
Table 2: Hofstede´s Cultural Dimensions ... 26
Table 3: Correlation Matrix ... 29
Table 4: Descriptive Statistics ... 29
Table 5: Model 1; Variables in the Equation ... 31
Table 6: Model 2 (Hypothesis 1); Variables in the Equations ... 31
Table 7: Model 3 (Hypothesis 2); Variables in the Equation ... 31
Table 8: Model 4 (Hypothesis 3); Variables in the Equation ... 31
Table 9: Model 6 (Hypothesis 5); PROCESS Model 1 Results... 34
Introduction
Multinational enterprises (MNEs), “defined as a firm that owns and controls activities in two or more different countries” (Buckley & Casson, 2009, p. 1564), face different challenges resulting from the complexity of managing their multiple embeddedness (Meyer et al., 2011). Multiple embeddedness in this context refers to different nation states which all have their own institutional framework and national culture in place (Meyer et al., 2011). Being globally active in multiple host locations creates complex interdependencies of culture that can result in cross cultural ethical conflicts (Hamilton & Knouse, 2001; Meyer et al., 2011). Additionally, Franke and Nadler (2008) conclude in their studies that organisations, which operate in areas with different national culture dimensions may encounter ethical conflicts.
Given its importance and relevance for international active companies, culture as the “the collective programming of the mind” (Hofstede, 1994, p. 1) is one of the most persistent concepts of international business (IB) and international management (IM) (Caprar et al., 2015). It is discussed as standalone concept (e.g. Hofstede, 1980, 1994; Kirkman et al., 2006) but also set in relation to other main topics of research like expansion behaviour of companies, internationalisation patterns of MNEs or leadership style (Caprar et al., 2015).
Another topical and widely discussed theme in relation to multinational enterprises are human rights. In the course of globalisation cross-border activities and networks are expanding, and raise the challenge of accountability and the balancing act of interference in domestic affairs of host countries (Ruggie, 2013). Consequently, human rights, according guidelines and principles, MNE impact and resulting problems are examined in many papers (e.g. Fasterling & Demuijnck, 2013; Giuliani & Macchi, 2014; Harrison & Sekalala, 2015; Ruggie, 2013; Wettstein, 2010).
Franke and Nadler (2008) and Christie et al. (2003), amongst others, combine both topics in their studies and investigate the impact of national cultural on the ethical attitude of
managers and nations. As ethical attitudes and decision making are possible antecedents of ethical conflicts and abuses (e.g. exploitation of workers and human health – McDonald´s; sweatshop conditions – Nike; sexual harassment – Mitsubishi), Christie et al. (2003) and Franke and Nadler (2008) do not study the direct impact of national culture on human rights violations but on general ethical issues. This study, however, aims to investigate the direct relationship of national culture and human rights violations while developing hypotheses based on the findings of Christie et al. (2003) and Franke and Nadler (2008), to ensure comparability.
Furthermore, Erwin (2011) and Preuss and Brown (2012) set the topic “human rights” in relation with “corporate social responsibility” and “codes of conduct” and examine the impact of code content on ethical performance. Preuss and Brown (2012) focus in their work on the prevalence of codes on human rights and the analysis of their content, while Erwin (2011) measures the direct relationship between the quality of codes and ethical performance. Although this relationship is very interesting to measure, the assessment of code effectiveness is very sensitive to poor quality codes. As a consequence, future research should focus on the assessment of presence or absence of codes according to Erwin (2011).
This study contributes to the literature threefold. First, Christie et al. (2003) limit their research of cultural impacts on ethical behaviour on India, Korea and the United States. In order to attain different insights, this study focuses on 19 different MNE home countries based on the selected sample provided by Sustainalytics (see Table 1). Second, Franke and Nadler (2008) and Christie et al. (2003) both measure the influence of culture on human rights violations indirectly as their focus of research is on ethical attitudes and not on reported abuses. This study aims to investigate the direct influence of national culture on human rights violations to generate additional insights. Finally, Preuss and Brown (2012) focus in their study of code prevalence solely on companies listed with the FTSE 100 index; resulting in a potential UK bias of their results. Thus, this study is following their call for further research to remove the
potential UK bias and investigate cross-cultural differences in MNEs´ approaches to human rights while comparing the prevalence of codes on human rights across regions.
To investigate these relationships and impacts, data from five different sources are collected and evaluated; namely the Corporations and Human Rights Database (CHRD), official webpage of Hofstede, Sustainalytics Global Platform, Orbis and company webpages. In order to test the formulated hypotheses, binary logistic regression as well as PROCESS Model 1 are run in SPSS. Following the results of these tests, none of the hypotheses is supported. Nevertheless, the study provides valuable insights as it sheds the light on different aspects as Franke and Nadler (2008) and Christie et al. (2003) what provides various suggestions for future research.
The structure of this work is as follows. First, a main overview is provided of the main concepts and theories applied in this study, resulting in the identification of the research gap and questions. This section is followed by the derivation and formulation of the hypothesis based on the discussed literature. Before the analysis is conducted, the sample selection, as well as the measures applied for each variable are described. The results of the analysis run in SPSS are summarized in section 5 of this study. Where needed, tables and figures are inserted to illustrate the findings in order to summarize the key information for the reader. The discussion, including limitations of this study as well as suggestions for future research, and the conclusion are based on these findings and are presented in section 6 and 7 of this work.
Literature Review
“Culture” is one of the most widely used concepts in international management and international business according to Caprar et al. (2015), Drogendijk and Slangen (2006) and Franke and Nadler (2008). Scholars devote their attention not only to the description and measurement of culture (e.g. Hofstede, 1980, 1994; Kirkman et al., 2006), but also set the concept in relation to other main topics of research (Caprar et al., 2015).
Another topical theme in the literature are human rights, especially in turn of the increasing golablisation of markets (Ruggie, 2013). Human rights themselves and related topics like according guidelines and principles, MNE impact and resulting problems are examined in many papers (e.g. Fasterling & Demuijnck, 2013; Giuliani & Macchi, 2014; Harrison & Sekalala, 2015; Ruggie, 2013; Wettstein, 2010).
Franke and Nadler (2008) and Christie et al. (2003), amongst others, combine these two topics and investigate the impact of national cultural on the ethical attitude of managers and organisations. As ethical attitude can be considered as one possible antecedent of ethical conflicts, both measure indirectly the influence of culture on human rights violations as their focus of research is on ethical attitude and not on reported abuses directly. This study addresses this gap and aims to investigate the direct relationship between the culture of MNEs´ home country and human rights violations and if this link is in any way impacted by the company´s human rights awareness.
In this section, I will first give an overview over the main concepts used in this study, resulting in the identification of the research gap and formulation of the research questions of this thesis.
Conceptualisation and Measurement of Culture
According to Caprar et al. (2015) the concept of culture, defined by Hofstede (1994, p. 1) as “the collective programming of the mind”, is one of the most persistent concepts of international
business (IB) and international management (IM). It is often connected to other key topics, such as internationalisation, multinational enterprises (MNEs), and cross-country activities; and it is used in different forms – as an antecedent, moderator, mediator independent or dependent variable (Caprar et al., 2015). Culture is applied to explain different patterns of international expansion, knowledge transfer, leadership style or MNE corporate social responsibility (Caprar et al., 2015). Furthermore, scholars used the concept of culture to analyse its impact on ethical attitudes of business managers, organisations and nations (e.g. Christie et al., 2003; Franke & Nadler, 2008). The findings of Christie et al. (2003) and Franke and Nadler (2008), support the idea that cultural characteristics influence ethical behaviour, although they apply different sample sizes and research methods.
Christie et al. (2003) investigate the relationship between culture and ethical attitudes,
inter alia, towards gift-giving, compliance with superior´s order or nepotism of managers based
on a cross-cultural comparison of India, Korea and the United States. The impact of culture is assessed as it is one of the main sources for attitude formation and an important determinant of ethical decision making according to Christie et al. (2003). Due to the growing degree of globalisation and an increasing number of ethical conflicts (e.g. exploitation of workers and human health – McDonald´s; sweatshop conditions – Nike; sexual harassment – Mitsubishi) studies about business ethics get increasing attention (Christie et al., 2003). Christie et al. (2003) test empirically if national culture, measured by Hofstede´s dimensions, explains differences in ethical behaviour. Based on the evaluation of primary survey data, the analysis of Christie et al. (2003) reveals a strong relationship between individualism and power distance and ethical attitudes; whereas the results for masculinity, uncertainty avoidance and long-term orientation are mixed.
Franke and Nadler (2008) apply a bigger sample in their study, consisting of 44 different countries, to assess the impact of culture and economic development on national ethical attitudes. Ethical attitudes are described by Franke and Nadler (2008) as moral standards
influencing ethical judgment and assessment what behaviour is appropriate. Secondary data is used in this study to measure culture (Hofstede´s dimensions) and economic development (gross domestic product (GDP)). Their findings show that power distance and uncertainty avoidance have a negative impact on the national ethical attitude, while individualism and masculinity are found to have no influence. An important conclusion of Franke and Nadler (2008) to highlight is that organisations, which operate in areas with different uncertainty avoidance and power distance scores may encounter ethical conflicts resulting from incompatible principles and varying perceptions.
As culture is used in different settings, there are different types of measurement; at country or individual level, as a single value, or as an index (Caprar et al., 2015). Hofstede´s five-dimensional model of national culture constitutes a popular method for studying cultural differences at a national level and is used in a variety of empirical and theoretical studies (Kirkman et al., 2006; Minkov & Hofstede, 2011). Hofstede identifies five dimensions to explain cross-national differences – power distance (PD), individualism (IND), uncertainty
avoidance (UA), masculinity (MAS) and long-term orientation (LTO) (Hofstede, 1994). Power distance identifies the degree to which less powerful members of an institution,
organisation or society expect and accept the inequality of power distribution (Hofstede, 1994). A low score on power distance indicates that people question authority and rather expect consultancy by supervisors than an autocratic leadership style (Franke & Nadler, 2008).
Individualism versus collectivism describes the degree of group-integration and therefore how
loose/tight the ties between individuals are (Hofstede, 1994). Individualistic characteristics are competitiveness, self-assurance, and initiative; whereas collectivistic traits include empathy, self-control, self-sacrifice, and conformity (Franke & Nadler, 2008). Masculinity, with its counterpart femininity, refers to a society where assertive and competitive behaviour is dominant (Hofstede, 1994). The masculinity score describes, if the society is more self-centred (high score) or more focused on relationships and modesty (low score) (Franke & Nadler,
2008). Uncertainty avoidance measures the tolerance of a society for uncertainty and ambiguity and indicates to what extent members of a society feel comfortable with unknown, novel or surprising situations (Hofstede, 1994). In societies with a high uncertainty avoidance score, religion, regulations and laws play a bigger role to cope with ambiguity (Franke & Nadler, 2008). Long versus short-term orientation describes, if people in a society focus more on the future versus the present (Franke & Nadler, 2008). Thrift and perseverance are values associated with long-term orientation; whereas protecting one’s “face” and respect for tradition are associated with short term orientation (Franke & Nadler, 2008).
Human Rights and Guiding Principles
“Human rights are universal legal guarantees protecting individuals and groups against actions and omissions that interfere with fundamental freedoms, entitlements and human dignity” (UNHRC, 2012, p. 10), which are characterized by universality, inalienability, interrelation,
interdependence and indivisibility. The aforementioned rights, together with the according
principles and guidelines, are formulated by state organisations, such as the United Nations (UN), Organisation for Economic Co-operation and Development (OECD) or International Labour Organisation (ILO), and also by nongovernmental organisations (NGOs). In order to ensure the compliance of the basic rights of mankind, almost every state has its own human rights policy and instruments in place (UNHRC, 2012). Certain documents form the foundation of international human rights law, such as the “Universal Declaration of Human Rights”, while others represent legally non-binding guiding principles, which help to implement consistent international standards of human rights (UNHRC, 2012).
The Universal Declaration of Human Rights represents the most widely accepted document of human rights (Cassel, 2001). It is formulated and designed by the Commission on Human Rights, which consists of members with various cultural, political and religious backgrounds. The definition process started in 1946 chaired by Eleanor Roosevelt, as a result
of the experiences in World War II. The aim of this document is to ensure that such human rights violations and atrocities, which took place in this war, would never happen again (UN, 2015). The final document, which was published on the 10th of December 1948, consists of 30 articles covering civil, political, economic, collective social and cultural rights. Those 30 articles are defined to be common standard for people all over the world, regardless of sex, nationality or religion. Over the years, it has become a yardstick and benchmark to measure what is right and what is wrong, as well as a worldwide tool for fighting against human rights violations (Ki-moon, 2015).
In addition to these basic rights, the ILO has specifically formalised international standards towards labour and working environment. These principles cover, inter alia, child and forced labour, gender equality and the setting of maximum working hours (Cassel, 2001). In this regard, the ILO drafted the document “Tripartite Declaration of Principles Concerning Multinational Enterprises and Social Policy” to deal with the increasing important role of multinational enterprises in the world economy and to ensure that the principles set by the ILO are respected and met (ILO, 2006).
Due to the changing role of MNEs over the last years and the fact that the UN Declaration of Human Rights is primarily addressed to governments, organisations, such as the UN and OECD, guiding principles especially towards MNEs and human rights getting more attention (Preuss & Brown, 2012; Ruggie, 2013). The OECD formulated “Guidelines for Multinational Enterprises” in 1976. The UN assigned the task to define guidelines, which set the roles and responsibilities of state and business regarding human rights, to UN Special Representative, John Ruggie, in 2005. The final document, “Guiding Principles on Business and Human Rights” (UNGP), was published in 2011 and provides guidance for implementing the UN “Protect, Respect and Remedy” framework (Fasterling & Demuijnck, 2013). These three pillars of the framework are (1) the state´s duty to protect human rights; (3) the business
to respect and comply to human rights and; (3) the need for effective remedies in case of violations and define the roles of state and non-state actors (UNHRC, 2011).
The centrepiece of this work is the human rights due diligence (HRDD) process, which should include the assessment, analysis, measuring, reporting and tracking of potential human rights impacts by companies (UNHRC, 2011). The HRDD process should help companies to identify, prevent and mitigate potential human rights impacts and therefore address the risk of legal claims (Ruggie, 2013; UNHRC, 2011). It targets a company´s own activities, but also all activities linked to its operations and to its business relationships; like joint venture partners or security forces (Ruggie, 2013). The HRDD process applies “to all enterprises regardless of their size, sector, operational context, ownership and structure” (UNHRC, 2011, p. 15), although it can be adjusted relatively, among other factors, to a company´s size.
Despite this important achievement of this UNGP document, Fasterling and Demuijnck (2013) as well as Harrison and Sekalala (2015) formulated their criticism towards the HRDD process. As scholars and governments have to rely on self-reporting data of companies, which hardly contains any information and insights, the effectiveness of this process cannot be tested reliably (Fasterling & Demuijnck, 2013; Harrison & Sekalala, 2015). For this reason, the risk exists, that companies are using HRDD as a strategic instrument with superficial human rights legitimation. Therefore, this empirical research is using non-company sources for collecting data regarding human rights violations and remedies to generate additional insights, increase accuracy and ensure objectivity.
Role of Multinational Companies and Human Rights Violations
Most scholars who discuss topics and trends around human rights, highlight the changing role of multinational enterprises (e.g. Giuliani & Macchi, 2014; Monshipouri et al., 2003; Muchlinski, 2001; Preuss & Brown, 2012; Ratner, 2001; Ruggie, 2013; Wettstein, 2010). Ratner (2001) and Ruggie (2013) build their arguments around the state limits of protecting
human rights sufficiently. MNEs became powerful global actors and states often lack resources, or even the willingness, to control corporate actions (Ratner, 2001). In some cases, governments are reliant on supplies by corporations, and in most cases, inward foreign direct investment (FDI) has a positive impact on developing countries´ economy and is highly welcome (Giuliani & Macchi, 2014; Ratner, 2001).
MNEs´ direct influence on the economy, politics and social services, influences society both positively and negatively (Monshipouri et al., 2003). Thus, corporations have significant impact on the wellbeing and life quality of individuals in the host country and should act and conduct their business accordingly. Nevertheless, some firms shift their activities on purpose to countries with a weak institutional framework in order to facilitate their business and to avoid increasing institutional pressure at home (Surroca et al., 2013; Ratner, 2001). Cross-country variations of stakeholder expectations impact the costs of obtaining legitimacy and allow, inter
alia, doing business at an unrealistically low minimum wage level (Monshipouri et al., 2003;
Surroca et al., 2013). What seems to be unacceptable in the home country is possible in the host country (Surroca et al., 2013). As a result, the net effect of globalisation and FDI in developing countries on human rights is questionable.
Some of these positive and negative effects of MNEs on human rights are summarised by Giuliani and Macchi in their work of 2014. Optimists claim that corporations can improve poor living conditions in developing countries by social investments and initiatives, while actively contributing to peace-building processes in conflict-zones (Giuliani & Macchi, 2014). Critics, however, claim the negative impacts on labour, environmental and social conditions due to increasing global competition for FDI and the lack of international binding norms for MNEs (Giuliani & Macchi, 2014).
Due to higher human rights awareness and greater media scrutiny, Giuliani and Macchi (2014) argue that most companies do not engage in potentially corrupt contracts and host countries on purpose. MNEs are primarily profit seeking for their shareholders, and avoid
activities that could interfere with this goal. Therefore, it is unlikely that MNEs engage in fundamental human rights violations knowingly and risk damaging brand reputation or even financial loss (Muchlinski, 2001). This is especially true as some consumers pay more attention whether corporations comply with guidelines and principles, and adjust their consuming behaviour accordingly (Castaldo et al., 2009). These “ethical consumers” boycott particular brands and products, what increases the pressure for compliance and makes social responsibility a more important point on MNEs´ agendas (Anderson, 2000; Weissbrodt, 2003). This is why the HRDD process is described as the centrepiece of the “Guiding Principles on Business and Human Rights” as it helps to uncover the unknown and unintended human rights violations.
MNEs´ Human Rights Awareness and Codes of Conduct
Not only do consumers put pressure on MNEs to comply to human rights, as Erwin (2011), Weissbrodt (2003) and Anderson (2000) argue in their work, but also other stakeholders of the corporation such as investors and media apply pressure (Monshipouri et al., 2003). Financial institutions, investors and mutual funds are asking for more social conscious activities due to increasing international attention on human rights violations committed by corporations (Anderson, 2000; Monshipouri et al., 2003). As a consequence, the internal and external, as well as national and international pressure on multinational enterprises to be compliant with guidelines and respect human rights is significant.
To cope with this pressure and to follow the request of governments and world organisations, including the UN or OECD, an increasing number of MNEs adopt codes of conduct, which is gaining momentum internationally (Anderson, 2000; Monshipouri et al., 2003; Ruggie, 2013). According to Kaptein (2004, p. 13), “a business code is a policy document that defines the responsibilities of the corporation towards its stakeholders and/or the conduct the corporation expects of employees”.
Following the work of Preuss and Brown (2012), multinational enterprises address human rights in different ways and cover them to a different degree. According to their study of 2012, corporations have either a standalone human rights policy in place, a section, or a reference in the code of conduct dedicated to human rights. But almost half (42 out of 98) of the investigated firms had no reference and did not address human rights at all. Moreover, Preuss and Brown (2012) found industry patterns in their study. Alcoholic beverages, oil and gas and tobacco are the top three industries regarding human rights awareness and consequently tend to have a standalone human rights policy in place.
Despite the varying degrees of MNEs´ human rights awareness, the content of codes and policies differs across corporations as well. To assess the content and its validity, many scholars use benchmarking against the UN Declaration of Human rights as it represents the most widely accepted document regarding human rights (Preuss & Brown, 2012). The content of codes is assessed against the 30 articles of the Declaration to check how extensively they are covered by the corporation´s guidelines. In addition, the frequency of use of certain topics, gives an indication on how important one specific article is for a company or, depending on the study, for an industry (Kaptein, 2004; Preuss & Brown, 2012).
As MNEs are profit-seeking and maximising corporations, codes are often used for public relations, and not to actively protect human rights (Monshipouri et al., 2003; Scherer et al., 2006). Hence, not only the content of codes is subject of research amongst scholars, but also the code performance. According to Erwin (2011), most empirical studies measure the presence or absence of codes. To measure the effectiveness of codes and their performance, the fit between MNEs´ behaviour and the ethical expectations standards of society and stakeholders has to be assessed (Erwin, 2011). Furthermore, Erwin (2011) argues that the quality of codes directly influences its performance. High quality codes are more effective in establishing a socially responsible culture and ensuring compliant behaviour of employees (Erwin, 2011).
Despite this trend of implementing codes and pressure of governments and organisations for guideline implementation, scholars also formulate their critics about their effectiveness (e.g. Anderson, 2000; Giuliani & Macchi, 2014; Monshipouri et al., 2003; Scherer et al., 2006). The most widely discussed point of criticism is that corporations use codes of conduct and social corporate responsibility as a public relation tool and gimmick (Anderson, 2000; Giuliani & Macchi, 2014; Monshipouri et al., 2003). Corporations apply guidelines and principles issued by international organisations, such as the UN, ILO and OECD. However, these are not followed and applied in a consistent way, and the comparison is especially difficult as international consistent valuation standards are missing (Monshipouri et al., 2003).
The principles and guidelines, formulated by the UN in 2011, towards human rights due diligence are one starting point. But how companies implement this HRDD process and to what extent they follow the UN document can hardly be assessed due to companies´ self-reporting data (Harrison & Sekalala, 2015). Another point of criticism is that company codes lack enforceability as they are self-imposed rules which are non-binding and can consequently be circumvent very easily (Anderson, 2000). Moreover, corporate codes are neither uniform nor consistent with other codes and sometimes not even within the MNE (Anderson, 2000; Kaptein, 2004). Sometimes subsidiaries have their own guidelines in place on top of the parent´s policy, which can lead to confusion among employees and decreasing code performance. Overall it is questionable if a well-defined code of conduct leads to a higher degree of human rights compliance as the moral commitment of a MNE itself is decisive (Fasterling & Demuijnck, 2013). Given these varying arguments and scepticism regarding codes, this study is investigating if the existence of a dedicated policy on human rights has an effect on human rights violations committed by a MNE to generate additional insights.
Identifying the Research Gap and Research Questions
Reviewing the literature, “culture” is one of the most widely used concepts in international management and international business (Caprar et al., 2015; Drogendijk & Slangen, 2006; Franke & Nadler, 2008). It is used as standalone framework (Hofstede 1980, 1994) or in relation to other dominant topics (Caprar et al., 2015). As culture is a prevailing theme in the literature, there are many different definitions and types of measurement; although the measurement framework and definition of culture as “the collective programming of the mind” by Hofstede (1994, p. 1) are most dominant.
Human rights as another topical theme of the recent literature is investigated by state and nongovernmental organisations, as well as by scholars. Organisations like the UN, OECD and the ILO set definitions, guidelines and principles, to increase the worldwide consistency towards human rights and provide tools for implementation, like the UNGP. Given the changing role of the multinational enterprises in the global economy and their gaining power, the human rights due diligence process is especially tailored to increase the human rights compliance of companies (Fasterling & Demuijnck, 2013; Ruggie, 2013; UNHRC, 2011).
The influence of MNEs on human rights is questionable. On the one hand, FDI is mostly associated with positive impacts as knowledge spillover and higher standards of living; but on the other hand, the government´s dependency on MNEs´ supply and its strong interest on inward FDI often leads to sweatshop-like work conditions and disregard of human rights (Giuliani & Macchi, 2014; Monshipouri et al., 2003; Ruggie, 2013). But latterly consumers pay more attention if corporations comply with guidelines and principles, human rights awareness is increasing and greater media scrutiny force MNEs to be more compliant (Giuliani & Macchi, 2014). To cope with this pressure, an increasing number of MNEs adopt codes of conduct, which is gaining momentum internationally (Anderson, 2000; Monshipouri et al., 2003; Ruggie, 2013). These codes address a company´s responsibilities towards its stakeholders and should increase the international acceptance of MNEs´ activities (Anderson, 2000). Despite this
trend, there is some criticism. Code of conducts are prone to be used as public relation tools and they lack enforceability as they are considered to be self-imposed rules which are non-binding and can consequently be circumvent very easily (Anderson, 2000).
This study is based on the arguments and findings of these scholars and organisations and aims to investigate, if the culture of the MNE´s home country influences the presence of human rights violations. Furthermore, it is measured, if this relationship is moderated by the company´s human rights awareness.
Christie et al. (2003) and Franke and Nadler (2008) investigate in their work the cultural impact on ethical behaviour. They measure in their studies how national culture influences the judgment and assessment what behaviour is appropriate and how culture impacts business ethics. Both studies find strong support for the proposed relationship while applying different research methods and using different sample sizes. Christie et al. (2003) and Franke and Nadler (2008) apply ethical behaviour as their dependent variable; meaning that they do not address the direct impact of culture on human rights violations but on ethical decision making in general. As ethical behaviour and decision making are possible antecedents of ethical conflicts and abuses, it can be assumed that culture has a similar impact on human rights violations as on ethical behaviour. This study addresses this gap and aims to investigate, if the culture of the MNE´s home country influences the presence of human rights violations.
Further, Erwin (2011) and Preuss and Brown (2012) investigate the impact of codes of conduct on ethical performance and conclude that companies´ approach towards human rights compliance varies in depth, quality and performance. Erwin (2011) suggests researchers to first focus on measuring the presence and absence of ethical codes before assessing the effectiveness.
Following these scholars, this study contributes to the literature threefold. First, Christie et al. (2003) limit their research of cultural impacts on ethical behaviour on India, Korea and
the United States. In order to attain different insights, this study focuses on 19 different MNE home countries based on the selected sample provided by Sustainalytics. Second, Franke and Nadler (2008) and Christie et al. (2003) both measure the influence of culture on human rights violations indirectly as their focus of research is on ethical attitudes and not on reported abuses. This study aims to investigate the direct influence of culture on human rights violations. Finally, Preuss and Brown (2012) focus in their study of code prevalence solely on companies listed with the FTSE 100 index; resulting in a potential UK bias of their results. Thus, this study is following their call for further research to remove the potential UK bias while comparing the prevalence of codes on human rights across regions.
Theoretical Framework
Human Rights Violations
Various scholars discuss the impact of multinational enterprises (MNEs) on human rights (e.g. Clapham & Jerbi, 2001; Giuliani & Macchi, 2014; Harrison & Sekalala, 2015; Ruggie, 2013; Wettstein, 2010). Giuliani and Macchi (2014) review both, the positive and negative potential impacts of MNEs on human rights and conclude that the literature on this topic is highly controversial. Further they find that the relation between MNEs and human rights violations is too complex to be described as a universal and linear relationship (Giuliani & Macchi, 2014). Contrasting, Clapham and Jerbi (2001) and Wettstein (2010) anticipate in their work a negative impact of MNEs` on human rights, and focus on the classification of human rights violations; especially on the concept of complicity and its different forms and characteristics (Clapham & Jerbi, 2001; Wettstein, 2010).
In this study, a potential negative impact of multinational enterprises on human rights is anticipated as well like in the studies of Clapham and Jerbi (2001) and Wettstein (2010). The focus is of this research, however, is not on the classification of human rights abuses, but on the measurement of presence or absence of human rights violations committed by the MNE.
Culture
Culture defined by Hofstede (1994, p. 1) as “the collective programming of the mind”, is one of the most persistent concepts of international business (IB) and international management (IM) according to Caprar et al. (2015). The concept is used by various scholars to explain key topics of the literature (e.g. internationalisation, the existence of multinational enterprises, patterns of international expansion) and it is applied in different forms in empirical research; as antecedent, moderator, mediator independent or dependent variable (Caprar et al., 2015).
Culture and Human Rights Violations
Christie et al. (2003) and Franke and Nadler (2008) are investigating the relation between national culture and ethical attitude of managers and nations. They employ different samples in their empirical research, but come up with similar findings while applying Hofstede´s framework. Both introduce ethical attitudes as dependent variable with the reasoning that culture is one of the main sources for attitude formation, and an important determinant of ethical decision making (Christie et al., 2003). As ethical behaviour and decision making are possible antecedents of ethical conflicts and abuses, it can be assumed that culture has a similar impact on human rights violations as on ethical behaviour. Therefore, this study aims to investigate, if the national culture of the MNE´s home country influences the presence of human rights violations, while developing hypotheses based on the findings of Christie et al. (2003) and Franke and Nadler (2008), to ensure comparability.
Two dimensions of Hofstede´s framework are neglected in this study – long-term orientation and masculinity – as there is no relation found with ethical attitudes neither by Christie et al. (2003) nor Franke and Nadler (2008). Thus, the focus of this research is on the dimensions power distance, uncertainty avoidance and individualism.
Research has consistently hypothesized that power distance has a negative effect on ethical attitude (Franke & Nadler, 2008). Both, Christie et al. (2003) and Franke and Nadler
(2008) hypothesise and prove a strong negative relationship in their studies. According to their research, high power distance has a negative impact on ethical attitudes of managers and nations. Following these findings and the argument of Christie et al. (2003) that ethical attitudes are an important determinant of ethical decision making, it can be assumed that high power distance has a negative effect on human rights violations as well. Furthermore, a society with high power distance expects employees to follow orders without questioning authority (Franke & Nadler, 2008). This may result in a blind following of orders without scrutinising business decisions and activities by employees. Therefore, the hypothesis is formulated:
H1: Increasing power distance results in higher probability of human rights violations committed by the multinational enterprise.
Individualism versus collectivism describes the degree of group-integration and how loose/tight
the ties between individuals are (Hofstede, 1994). Individualistic characteristics are competitiveness, self-assurance, and initiative; whereas collectivistic traits include empathy, self-control, self-sacrifice, and conformity (Franke & Nadler, 2008). Thus, in a society with a high score of individualism, employees tend to be more egoistic and look more after themselves instead of following group interest (Christie et al., 2003). To reach their goals, individualists may be as a consequence more willing to neglect rules, laws and principles, what potentially increases the probability of human rights abuses. Therefore, Christie et al. (2003) as well as Franke and Nadler (2008) propose a negative relationship between individualism and ethical attitudes. While Christie et al. (2003) find support for this relationship, the negative impact of individualism is not significant following the findings of Franke and Nadler (2008). Based on these findings as well as on theory, the following hypothesis is proposed:
H2: Increasing individualism results in higher probability of human rights violations committed by the multinational enterprise.
Uncertainty avoidance measures the tolerance of a society for uncertainty and ambiguity and
indicates to what extent members of a society feel comfortable with unknown, novel or surprising situations (Hofstede, 1994). In societies with a high uncertainty avoidance score, religion, regulations and laws play a bigger role to cope with ambiguity (Franke & Nadler, 2008). Although societies with high individualism are expected to have more rules in place, it does not imply a low level of risk avoidance (Franke & Nadler, 2008). “Unethical behaviours may be seen as acceptable risks because they often bring about immediate desired outcomes” (Franke & Nadler, 2008, p. 257), as well as breaking unjust laws. Consequently, the probability of human rights abuses in countries with high uncertainty avoidance may increase as employees see unethical behaviour as acceptable risk. Furthermore, following regulations and laws, does not imply that the resulting actions and activities are compliant with human rights (Christie et al., 2003). For example, the export and shipment of harmful products can be permitted by law, but may not be in line with human rights regulations (Christie et al., 2003). Christie et al. (2003) and Franke and Nadler (2008) investigate a negative impact of uncertainty avoidance on ethical attitudes. Whereas Christie et al. (2003) obtain mixed results in their study, Franke and Nadler (2008) find a strong negative relationship. Following existing research, in this study a negative impact of uncertainty avoidance is proposed as well:
H3: Increasing uncertainty avoidance results in higher probability of human rights violations committed by the multinational enterprise.
Companies´ Human Rights Awareness as a Moderator
Baron and Kenny (1986) as well as Frazier et al. (2004) define a moderator as a quantitative or qualitative variable, which influences the strength and/or direction of the relation between an independent and a dependent variable. “Thus, a moderator effect is nothing more than an interaction whereby the effect of one variable depends on the level of another” (Frazier et al., 2004, p. 116). Many scholars introduce a moderator variable, when they are interested, if the
relationship between an independent and a dependent variable is stronger in some occasions (Frazier et al., 2004).
Sower and Sower (2004) as well as Caligiuri (2000), amongst others, introduce various moderating variables in their studies which they argue impact the effect of national culture. Sower and Sower (2004) for example introduce in their work various demographic moderating variables (gender, age, work experience, religiosity, etc.), which they argue have an effect on the relationship between national culture and individual ethical behaviour. Caligiuri (2000) however, propose a moderating effect of personality traits, openness and sociability specifically, on cross-national adjustments.
In this study, however, a different moderating variable is introduced. Based on the argumentation and findings of Erwin (2011), the company´s human rights awareness is introduced as a potential moderator of the relationship between national culture and human rights violations (Figure 1). Erwin (2011) investigates in his study the effects of code content and quality on ethical performance what results in mixed findings. Although a correlation is found between high quality codes and high ethical performance, the causality of this relationship is uncertain (Erwin, 2011). This is, as the quality of codes is very difficult to assess and poor quality codes may skew results (Erwin, 2011). Consequently, Erwin (2011) recommends to assess the presence or absence of ethical codes to measure ethical performance. Further, Erwin (2011) describes the adoption of codes of conduct as a symbol for corporate social responsibility (CSR) and human rights awareness. Therefore, in this study the potential moderating effect of human rights awareness, measured by the presence or absence of codes, is investigated.
Increasing power distance is hypothesised to result in a higher probability of having human rights violations committed due to a higher degree of blind obedience by employees. As Franke and Nadler (2008) argue, in societies with high power distance, employees are expected to follow orders without questioning; possibly also those which lead to human rights violations.
Codes, however, affect organisational culture according to Erwin (2011) as they provide guidelines for actions and ethical business activities. Further codes contain the values and norms of a company which are targeted to decrease the amount incidents (Kaptein, 2004). Consequently, it can be assumed that companies which have a human rights policy in place commit less human rights violations as employees are provided with according guidelines to prevent abuses.
H4: A company´s human rights awareness negatively moderates the relationship between power distance and the presence of human rights violations.
A high individualism score signals certain predominant characteristics of society like competitiveness, self-assurance and egoism (Franke & Nadler, 2008). Such traits increase the probability that individuals follow their goals without respecting laws, guidelines or principles. Putting the outcome of interest above compliance may as a consequence lead to a higher probability of having human rights violations committed. Codes are put in place by companies to regulate employees´ behaviour and to steer their activities into the desired direction (Erwin, 2011; Kaptein, 2004). Furthermore, codes can lead to organisational change and thus have the power of influencing actions and conduct of employees (Erwin, 2011). As companies´ codes can impact the activities of employees to be more compliant with certain guidelines and principles, the probability of human rights violations may be decreasing.
H5: A company´s human rights awareness negatively moderates the relationship between individualism and the presence of human rights violations.
On the one hand, societies with high uncertainty avoidance tend to have a higher amount of rules, regulations and principles in place to avoid ambiguity and uncertainty (Hofstede, 1994). On the other hand, following regulations and laws, does not imply, that the resulting actions and activities are compliant with human rights (Christie et al., 2003). Laws may permit the
production and shipment of certain goods, or business activities in highly controversial regions, but this may not be in line with human rights regulations (Christie et al., 2003). Companies therefore put codes in place to provide guidelines for employees to signal what is expected from stakeholders and considered as acceptable business practices (Erwin, 2011; Kaptein, 2004). Hence, codes put pressure on employees to be compliant with human rights regulations to reduce the occurrence of incidents and to increase stakeholder confidence (Kaptein, 2004).
H6: A company´s human rights awareness negatively moderates the relationship between uncertainty avoidance and the presence of human rights violations.
Figure 1: Conceptual Model
Data and Methods
This study aims to investigate the relationship between the national culture of a MNEs´ home country and human rights violations; and whether this relationship is moderated by the company´s human rights awareness. To explore these relationships, this work follows a deductive approach as the formulated hypotheses are based on existing theory and empirical research. Additionally, this study is of explanatory nature, as its aim is to investigate the relationship between key variables (Saunders & Lewis, 2012).
In order to study the relationships and impacts proposed in this work, quantitative research is conducted with data from multiple sources. Panel data from the Corporations and Human Rights Database (CHRD) is gathered, as well as information from Sustainalytics, Orbis, Hofstede´s webpage on national culture (http://geert-hofstede.com/national-culture.html) and MNEs´ company webpages to operationalise the introduced variables.
Sample and Data Collection
The sample selection of this study is based on Sustainalytics and the availability of Hofstede cultural dimensions. Sustainalytics Global Platform is an independent consulting company which periodically gathers and evaluates data of more than 2,000 MNEs based in 27 different countries regarding their social, governance and environmental performance (Garcia-Castro & Francoeur, 2014). Further, Sustainalytics is used by many scholars as source for their data, like Garcia-Castro and Francoeur (2014) or Surroca et al. (2013), to conduct their research.
The sample of this study is limited to the oil and gas companies available in Sustainalytics, as the human rights attention in this industry is above-average and the public pressure towards corporate social responsibility (CSR) is especially high (Preuss & Brown, 2012). Furthermore, the extractive sector accounts for 28% of all reported human rights violations what provides lots of relevant data and makes it an interesting and highly topical industry to investigate (Ruggie, 2013). In addition, data on human rights abuses of the oil and gas industry is most complete in the Corporations and Human Rights Database (CHRD) compared to other sectors, what allows sound research and data analysis.
Additionally, the sample is limited to companies, where data in Sustainalytics is available for measuring the presence and absence of human rights policies. This is resulting in a sample of 60 MNEs whose sustainability report has been checked for human rights between 2010 and 2013. Checking further for the availability of the Hofstede dimensions for the MNE home countries, two MNEs have to be dropped as the Hofstede values are not available. The
final sample applied in this study is 58 MNEs with 19 different home country backgrounds in total (Table 1). For a full overview of the applied sample please refer to Appendix 1.
Country Region MNEs
Australia Asia & the Pacific 3
Belgium Europe 1
Brazil Latin America 1
Canada North America 9
China Asia & the Pacific 2
Colombia Latin America 1
France Europe 2
Great Britain Europe 8
India Asia & the Pacific 2
Ireland Europe 1
Italy Europe 2
Japan Asia & the Pacific 2
Netherlands Europe 1
Norway Europe 2
South Korea Asia & the Pacific 3
Spain Europe 1
Sweden Europe 1
Thailand Asia & the Pacific 2
United States North America 14
Total 58
Table 1: Sample Overview
Measures
Dependent Variable
Based on the selected sample in Sustainalytics panel data is collected from the Corporations and Human Rights Database (CHRD) to measure the dependent variable, human rights violations, from 2010 till 2013. The CHRD contains various data regarding claims of corporate human rights abuses, like information about who reports these claims, how do companies respond and how violations are addressed from 2000 through today (CHRD, 2016). It is still a pilot project as it requires a lot of financial resources and manpower to code all reported human rights violations in all industries. Therefore, the database is gradually growing and covering more countries and industries over time. As the extractive sector makes up almost 30% of all
human rights violations reported according to Ruggie (2013), the CHRD project started to code all reported claims in this industry Thus, the CHRD constitutes a valid source to collect data regarding the presence of human rights violations in the oil and gas industry, as it is one of the industries which is completely coded in the database.
To operationalise this variable, a dichotomous variable is applied which takes on the value “1” if the MNE has violated human rights in the given period, and ”0” if there are no MNE specific human rights abuses reported. This means, if the MNE has committed one or more human rights violations in one of those four years, the dependent variable takes on the value of “1”. Just if there is no human rights abuse reported during all four years, the variable takes on the value of “0”.
Before getting access to the CHRD, we had to gather information and code human rights violations ourselves. In my case, I was responsible for coding 21 cases which were reported in South Africa and occurred in the mining industry. Coding one case took between 2,5 and 3 hours, resulting in a total work load of approximately 58 hours.
Independent Variable
To understand and measure national culture, different methods and frameworks have been developed like Hofstede´s dimensions, Schwartz´s value survey, the GLOBE project or the World Value Survey (Drogendijk & Slangen, 2006; Haxhi & van Ees, 2010). As above-mentioned, this study applies the framework of Hofstede – power distance, uncertainty avoidance, individualism, masculinity and long-term orientation – to measure national culture.
Although some points of criticism have been formulated on Hofstede's framework – such as reducing culture down to five dimensions, limiting the sample to one single MNE, ignoring cultural evolution over time and within-country heterogeneity – it is still most widely used amongst scholars and accepted in the international business field, what ensures credibility and appropriateness of this measurement (Christie et al., 2003; Kirkman et al., 2006).
Furthermore, using Hofstede´s framework allows the comparison of this study with existing similar empirical research on the impact of culture on human rights violations; e.g. the work of Christie et al. (2003) and Franke and Nadler (2008).
Hofstede analysed survey data of more than 117,000 IBM employees, which were located in 40 different countries (Drogendijk & Slangen, 2006). Based on the employees´ survey responses, Hofstede formulated his framework for evaluating national culture which consists of five dimensions and each takes on a score between 0 and 100 whereas a score of 50 can be considered as midlevel (Drogendijk & Slangen, 2006; Hofstede, 1994; Hofstede et al., 2010).
The relevant data for this measurement is collected from the official webpage of Hofstede; (http://geert-hofstede.com/national-culture.html). The values for each cultural dimension available on this webpage are based on the publications
“Culture’s Consequences: Comparing Values, Behaviors, Institutions, and Organizations Across Nations” (Hofstede, 2001) and “Cultures and Organizations: Software of the Mind” (Hofstede
et al., 2010) and can be considered as valid over time, as culture changes very slowly (Hofstede, 2001). Further, Williamson (2000) describes in his study that norms, mores and traditions – each components of culture – change very slowly, on
the order of centuries or millennia, what supports validity of Hoftede´s values over time. Consequently, the Hofstede dimensions available on the official webpage of Hofstede in June 2016, can be applied to measure national culture from 2010 to 2013. Table 2 gives an overview
Country PD IND UA Australia 36 90 51 Belgium 65 75 94 Brazil 69 38 76 Canada 39 80 48 China 80 20 30 Colombia 67 13 80 France 68 71 86 Great Britain 35 89 35 India 77 48 40 Ireland 28 70 35 Italy 50 76 75 Japan 54 46 92 Netherlands 38 80 53 Norway 31 69 50 South Korea 60 18 85 Spain 57 51 86 Sweden 31 71 29 Thailand 64 20 64 United States 40 91 46
of the 19 different MNEs´ home countries and the according values of the three cultural dimensions applied in this study – power distance, individualism and uncertainty avoidance.
Moderating Variable
The data for measuring MNEs´ human rights awareness is gathered from Sustainalytics. Between 2010 and 2013 the sustainability report of various companies has been checked for human rights providing the necessary data on the presence/absence of a human rights policy.
The human rights awareness is measured by the presence/absence of a standalone human rights policy of the MNE as recommended by Erwin (2011) instead of evaluating code quality and effectiveness. The operationalisation of this variable follows the same logic as the dependent variable. A dichotomous variable is introduced which takes on the value “1” if the MNE has a human rights policy in place in the given time period, and ”0” if there are is no reference to human rights found in the company´s sustainability report during the four years.
Control Variables
The first variable that is controlled for is firm size, which is a common control variable in the literature and can be linked to the ownership of both, financial and non-financial assets (Gubbi et al. 2010; Vermeulen & Barkema, 2002). As firms which are bigger in size tend to have more capacity, they can devote more of their available resources and expertise to code formulation, tracking and implementation of human rights due diligence (HRDD) (UNHRC, 2011). Therefore, the firm size in this work is measured by the logarithm of the company´s average total assets over the 4 years (Gubbi et al. 2010). Data for this measurement is obtained from Orbis for the respecting companies and timeframe.
The second control variable is the firm´s experience in doing business abroad. The experience a company has acquired while doing business abroad and engaging in overseas activities, has been proven to add positively to the company´s accumulated knowledge (Gaba et al., 2002). Having more experience in doing business abroad helps the MNE to be more
sensitive and thoughtful about local conditions and potential sources for human rights violations. The measurement therefore is the average number of countries a MNE is active in over the given time period from 2010 till 2013 (Chang & Rhee, 2011). The according data has been gathered from the annual reports available on the companies´ webpages for the given time period.
Statistical Analysis and Results
Descriptive Statistics
The descriptive statistics of all variables applied in this study – dependent, independent, moderating and control – are displayed in Table 3 & 4 and they give an overview of all means, standard deviations and correlations. As a first step, the predictor variables are checked for multicollinearity to ensure the accuracy of estimates of coefficients and standard errors (Alin, 2010; Grewal et al., 2004). Grewal et al. (2004) mention various methods how multicollinearity can be detected like computing a correlation matrix of predictor variables or path coefficients, or the inspection of the variance inflation factors (VIF). Alin (2010) states, however, that it is not enough to just check the correlation matrix of predictor variables but one should also inspect the variance inflation factors (VIF) what constitutes a very popular method of multicollinearity diagnostic. Following these two authors, first the correlations of the independent variables are checked in Table 3 followed by the analysis of the VIFs of the predictors in SPSS.
The correlation coefficients of power distance, individualism and uncertainty avoidance in Table 3 show a high relationship between the three dimensions which might indicate multicollinearity. The correlation between PD and IND is highest with r = -.78, followed by PD and UA with r = .49, and IDN and UA with r = -.49; all at a significance level of p < 0.01. This is not completely surprising as all three dimensions measure facets of national culture although they signal different potential characteristics. Checking further for the variance inflation factors, however, shows that multicollinearity is not problematic in this model. The
VIF values are with 1, 1 and 3 beneath the threshold of VIF < 4; and the tolerance levels are with 0.8, 0.8 and 0.4 above the according threshold of 0.25 (O´Brien, 2007).
Variable 1 2 3 4 5 6 7
1. Human rights violations 1
2. Power distance .21 1
3. Individualism -.06 -.78** 1
4. Uncertainty avoidance -.08 .49** -.49** 1
5. Human rights awareness .06 -.08 -.05 .26* 1
6. Firm size .46** -.02 -.01 .09 .41** 1
7. Firm experience .30* .20 -.13 .41** .60** .46** 1
** p < 0.01, * p < 0.05
Table 3: Correlation Matrix
Variable N Min Max Mean Std. Dev.
1. Human rights violations 58 0 1 .38 .49
2. Power distance 58 28 80 46.24 14.09
3. Individualism 58 13 91 71.17 25.18
4. Uncertainty avoidance 58 29 94 53.29 18.43
5. Human rights awareness 58 0 1 .55 .50
6. Firm size 55 8.81 20.12 16.71 1.94
7. Firm experience 50 1 84 20.02 19.77
Table 4: Descriptive Statistics
The descriptive statistics in Table 4 show that 38% of the companies in the applied sample have committed human rights violations during the time period of 2010 till 2013 and 55% of the 58 MNEs have a human rights policy in place. Furthermore, the values for power distance, individualism, and uncertainty avoidance have each a mean of 46, 71 and 53 respectfully. This implies that on average the MNE home country culture shows moderate characteristics of power distance and uncertainty avoidance, and a strong form of individualism. Given a minimum of 1 and a maximum value of 84 for the number of countries a company is active in, we can see that there is quite a difference in size of the selected MNEs. Excluding the correlations coefficients of the predictors, the strongest relationship in our model can be found between the moderator and firm experience (r = .60; p = 0.01), followed by the relation between the control
variables (r = .46; p = 0.01) and the relation of human rights violations and firm size (r = .46;
p = 0.01).
Testing of the Main Effects
To test Hypothesis 1, 2 and 3 binary logistic regression is conducted in SPSS as the dependent variable in this model is dichotomous. Logistic regression is applied to model dichotomous outcome variables while the predictors can be dichotomous, ordinal, or numerical. For all models in the analysis, the classification tables (displayed in Appendix 2) and the Nagelkerke
R2 values are analysed in order to get information about model accuracy, predictive capacity and variance explained. In a next step, the odds ratio with the according significance are interpreted to see, if the formulated hypotheses are supported or not. First the baseline model, and the model including the control variables are investigated. As a next step, each independent variable – power distance, individualism and uncertainty avoidance – is separately added to Model 1. Table 5, 6, 7 and 8 show the according results of the binary the logistic regression tests which are now successively discussed.
In Model 1, solely the control variables are added to the analysis what increases the overall accuracy from 64.6% to 77.1%. The 64.6% predictive capacity of the baseline model (please refer to Appendix 2: Block 0) can be interpreted as the accuracy obtained without adding any predictors to the model; also meaning, by chance or flipping a coin 64.6% of the cases would be correctly assigned to 1 or 0. Further, the Nagelkerke R2 = .396 indicates a moderate fit of the data sample to Model 1. Table 5 shows the significance of the two control variables in Model 1. Firm size has a significant effect on human rights violations (Exp(B) = 2.54; p < 0.01), whereas the effect of firm experience is not significant (Exp(B) = 1.00; p = .90). By each unit increase in firm size, the likelihood that a firm is committing human rights violations increases by 2.54., controlling for all other variables in the model.
95% CI for Exp(B)
Variable B S.E. Wald df Sig. Exp(B) Lower Upper
Firm size .93 .33 7.98 1 .01 2.54 1.33 4.86
Firm experience .00 .02 .02 1 .90 1.00 .96 1.04
Constant -16.72 5.61 8.88 1 .00 .00 - -
Table 5: Model 1; Variables in the Equation
95% CI for Exp(B)
Variable B S.E. Wald df Sig. Exp(B) Lower Upper
Firm size .95 .35 7.54 1 .01 2.59 1.31 5.12
Firm experience -.00 .02 .00 1 .97 1.00 .96 1.04
Power Distance .02 .03 .41 1 .52 1.02 .96 1.08
Constant -17.85 6.21 8.27 1 .00 .00 - -
Table 6: Model 2 (Hypothesis 1); Variables in the Equations
95% CI for Exp(B)
Variable B S.E. Wald df Sig. Exp(B) Lower Upper
Firm size .93 .32 8.24 1 .00 2.54 1.34 4.79
Firm experience .00 .02 .04 1 .84 1.00 .97 1.04
Individualism .01 .02 .25 1 .62 1.00 .98 1.04
Constant -17.28 5.66 9.32 1 .00 .00 - -
Table 7: Model 3 (Hypothesis 2); Variables in the Equation
95% CI for Exp(B)
Variable B S.E. Wald df Sig. Exp(B) Lower Upper
Firm size .91 .33 7.76 1 .01 2.48 1.31 4.71
Firm experience .02 .02 .76 1 .38 1.02 .98 1.07
Uncertainty Avoidance -.04 .03 2.45 1 .12 .96 .92 1.01
Constant -14.60 5.55 6.93 1 .01 .00 - -
