CORPORATE TAX STRATEGIES
…and corporate social responsibility?
BY A.L.C Stumpel
Master Thesis Business Studies –International Management First Supervisor: E. Dirksen, MSc.
Second Supervisor: Dr. A.E. Kourula. By A.L.C Stumpel (10510311) Date: 03-03-2014
2
A.L.C. (Guus) Stumpel
Student Number: 10510311
MSc Business Studies – International Management
Corporate tax strategies and corporate social responsibility?
Master ThesisFirst supervisor: E. Dirksen, MSc. Second supervisor: Dr. A.E. Kourula.
3
COLOPHON
Submitting this thesis: ‘Corporate tax strategies and corporate social responsibility’ fulfills all the requirements for a Master of Science Degree in the field of Business Studies (specialization International Management) for the University of Amsterdam .
The title refers to the possible relationship between corporate tax strategies and corporate social responsibility.
4
ABSTRACT
Corporate tax strategies (CTS) are very topical these days. In the past few years CTS have become world-wide phenomena, which Multinational enterprises (MNE’s) operate in a rather discrete manner. The objective of this thesis is to examine a relationship between corporate tax strategies and corporate social responsibility (CSR). The research question of this thesis is: “How are corporate tax strategies associated with corporate social performance?”. To answer this question three hypotheses are developed, and two different statistical tests (multiple regression and two-way anova) are used to examine the relationship. The dependent variable is CSR – Score of the Reputation Institute. The first independent variable is CTS, operationalized in effective tax rate, and the second variable is ETR in MNEs financial statement. The control variables are Company size and Industry. The results show that the first hypothesis, showing the negative association between CTS and CSR – score, is insignificant. The second hypothesis is accepted which argues that the transparency of CTS in the financial statement will lead to a higher CSR –score. The last and third hypothesis argues a positive mediator effect of publishing the CTS in financial report on the relationship between CTS and CSR – score, this third hypothesis is accepted. Because the relationship between CTS and CSR – Score is not confirmed, more research is needed on this subject. A managerial recommendation can be drawn that being transparent about CTS will lead to higher CSR – Scores.
ACKNOWLEDGEMENTS
First of all, I would like to thank the Reputation Institute for providing me the necessary data to do this study. Furthermore, without the help of Erik Dirksen this thesis would not succeed, thanks! Last I would like to thank my family (Louk, Gita & Lisa), and friends (special thanks to Lieselotte) for their support along the process!
5
INDEX
COLOPHON ... 3 ABSTRACT ... 4 ACKNOWLEDGEMENTS ... 4 INDEX ... 5 LIST OF TABLES ... 8 LIST OF FIGURES ... 8 LIST OF EQUATIONS ... 8 1. INTRODUCTION ... 9 2. LITERATURE REVIEW ... 11Evolution of the CSR concept ... 11
Less definitions, more research... 14
From CSR to CSP ... 17 Relating CSP to CTS ... 22 Hypotheses ... 23 3. RESEARCH DESIGN ... 26 Quantitative research ... 26 Dependent variable ... 26 Independent variables ... 27 Control Variables ... 29
Validity and Reliability ... 30
Method of analyses ... 31
Conceptual model of hypothesis ... 32
Limitations ... 33 4. RESULTS ... 34 Descriptive statistics ... 34 Statistical analyses ... 34 Correlation analysis ... 34 Regression Analysis ... 37
Assumptions for the regression model ... 39
Two – way anova ... 42
5. DISCUSSION ... 44
Strategical implications for managers ... 46
6 Further recommendations ... 49 7. LITERATURE ... 51
7
ABBREVIATIONS
BTD – Before Tax Deductions CEO – Chief Executive Officer
CFP – Corporate Financial Performance CSP – Corporate Social Performance CSR - Corporate Social Responsibility
CSRI-Score – Corporate Social Responsibility Index CTS – Corporate Tax Strategy
ETR – Effective Tax Rate
FDA – Food and Drug Administration HQ – Head Quarters
MNE – Multi National Enterprise US – United States
8
LIST OF TABLES
Table 1: Level of Analysis ... 20
Table 2: Typical Corporate and Stakeholder Issues ... 21
Table 3: Descriptive Statistics ... 34
Table 4: Summary Statistics and Correlation Matrix ... 36
Table 5: Collinearity Statistics ... 37
Table 6: Regression Analysis ... 39
Table 7: F-tests and Sum of Squares ... 42
Table 8: Two-way Anova ... 43
LIST OF FIGURES
Figure 1: The Four Faces of Corporate Responsibility ... 15Figure 2: Conceptual model ... 32
Figure 3: Histogram with normal curve for CSR. ... 40
Figure 4: Histogram with normal curve for Company Size ... 41
LIST OF EQUATIONS
Equation 1: Operationalization of CTS ... 28Equation 2: Standard Equation by Berry and Fieldman (1985) ... 31
Equation 3: Equation of study ... 32
9
1. INTRODUCTION
Over the last few years corporate tax strategies (CTS) have increasingly received the attention of the media. CTS are tax planning activities of Multinational enterprises (MNE’s) to avoid or minimize corporate income tax. In recent hearings of the House of Commons in England it became clear that Google, Amazon and Starbucks are examples of MNEs who use legal tax minimization or tax avoidance schemes to avoid high corporate taxes in the country of origin (Parliament.uk 2012). Starbucks tax expenses were approximately 9 million pounds over 14 years of operating in the UK, while generating over 3 billion pounds in revenue (BBC 2012). Over the years Facebook paid 238,000 pounds to the UK treasury, while their turnover was 20,4 million pounds (Cairns 2012). The literature is still developing on CTS: the first contributing to this concept was Slemrod (2001). However, scholars have identified the increasing importance of this concept. Now, the effects of CTS are questioned for MNE (e.g. Christensen and Murphy 2004), society (e.g. Freedman 2013) and government (e.g. Devereux, Lockwood et al. 2008). Furthermore scholars try to identify ways to discover MNEs using CTS, this leads to many problems because MNEs using CTS are very discrete (Heitzman 2010). A few popular and well known strategies are used for CTS, for example: transfer pricing (this could be intellectual property rights) and the use of offshore accounts in tax havens, all resulting in a low effective tax rate (ETR).
Another trend within MNE’s is the rising importance of corporate social responsibility. According to Kotler and Lee, this can be seen as (2008): “Doing the most good for your company and your cause”. CSR is an old concept, which has evolved over the years in terms of definitions and themes. An old definition of CSR is provided by Bowen (1953): “It refers to the obligations of businessmen to pursue those policies, to make those decisions, or to follow those lines of action which are desirable in terms of the objectives and values of our society” (p.6) to more modern definitions of CSR by Carroll (1979): “The social
10 responsibility of business encompasses the economic, legal, ethical, and discretionary expectations that society has of organizations at a given point in time (p. 500).” From 1980 onwards the focus on the formulation of definitions shifted towards the identification of various themes within CSR. Examples of these themes are: corporate social performance (CSP), business ethics, public policy and many more (Carroll and Carroll, 1999).
This thesis studies the relation between the concepts of corporate social responsibility and tax avoiding strategies. The theoretical contribution of this study is the relationship between CTS and CSP/CSR. The justification towards society is obvious as well: it’s important for MNE’s to know whether these strategies can back-fire, or if it has no significant impact on the perception of society on the operating MNE. Clearly this thesis does not aim to justify or to judge MNE’s who are using tax-avoidance strategies. This thesis solely examines whether there is a significant relationship between those concepts.
The central question which will be addressed by this thesis is: “How are tax-avoidance strategies (CTS) associated with corporate social performance (CSP)”. In chapter 2, the major contribution to the literature of CSR and CTS will be discussed. In chapter 3, the research design will be presented. In chapter 4 the results found in this study will be addressed. In chapter 5 these results will be analyzed and discussed. Chapter 6 contains a summary of the findings and the conclusion.
11
2. LITERATURE REVIEW
This literature review will address the concept of corporate social responsibility as a historical event. In this time frame the paper will try to identify all the major contributions to the literature of CSR. Hereafter, the paper will identify the different themes related to CSR. The last part of this literature review discusses the gap between the current literature in CSR and the relation with Corporate Tax Strategy (CTS).
Evolution of the CSR concept
The major contributions to the concept of CSR will be discussed in this paper in a historical time frame. The first attenuated forms of CSR can be found in the literature dating back to the 1930’s and 1940’s (Carroll and Carroll 1999). At this point in time, many references addressed the concept as ‘social responsibility’ (Clark, 1939; Barnard, 1938; Kreps, 1940).
The 1950’s are seen as the start of the modern period on this topic, initiated by the publication of Howard R. Bowen’s book: ‘Social Responsibilities of the Businessman’ (1953) (Carroll and Carroll 1999). In his book, Bowen (1953) argued that the businesses are the center of power and decision making, and therefore the actions of these businesses affect day-to-day life of the citizens. The first definition of the attenuated form of ‘CSR’, the social responsibility of businessmen, is presented by Bowen (1953): “It refers to the obligations of businessmen to pursue those policies, to make those decisions, or to follow those lines of action which are desirable in terms of the objectives and values of our society” (p. 6).
The next decade, the 1960’s, was characterized by developing the definition of CSR. One of the most prominent leaders in this debate was Keith Davis (1960) (Carroll and Carroll 1999). Davis (1960) sets forth a new definition of social responsibility in a managerial context: “businessmen’s decisions and actions taken for reasons at least partially beyond the
12 firm’s direct economic or technical interest”(p.70). Keith Davis’s definition (1960) implies that social responsibilities have two faces: First, businessmen are managing an economic unit in society, thus they have a broad obligation to the community regarding economic development, which in turn affects public welfare. Second, businessmen have the obligation to nurture, maintain and develop human values, such as moral, motivation and cooperation (Davis, 1960). In the next few years of this decade several other definitions of social responsibility are given by major contributors to this subject. Frederick (1960) defined social responsibilities as: “businessmen should oversee the operation of an economic system that fulfills the expectations of the public. And this means in turn that the economy’s means of production should be employed in such a way that production and distribution should enhance total socio-economic welfare. Social responsibility in the final analysis implies a public posture toward society’s economic and human resources and a willingness to see that those resources are used for broad social ends and not simply for the narrowly circumscribed interests of private persons and firms. (p. 60)”.
Another definition comes from Davis and Blomstrom (1966) who defined social responsibility as: “Social responsibility, therefore, refers to a person’s obligation to consider the effects of his decisions and actions on the whole social system. Businessmen apply social responsibility when they consider the needs and interest of others who may be affected by business actions. In doing so, they look beyond their firm’s narrow economic and technical interests (p. 12)”. These contributions are still using the notion of ‘businessmen’, the first reference addressing the relation between a corporation and society is Clarence C.Walton (1967) (Carroll and Carroll 1999). Walton (1967) argues that this relationship should be kept in mind by the top managers when pursuing their goals.
The peak of the proliferation of the CSR concept was reached in the 1970’s. Harold Johnson (1971) presented a variety of definitions for CSR in his book. The first view of CSR
13 by Johnson (1971) was called ‘conventional wisdom’ (Johnson 1971). In this view Johnson (1971) hints toward the stakeholder approach. Conventional wisdom is defined as: “A socially responsible firm is one whose managerial staff balances a multiplicity of interests. Instead of striving only for larger profits for its stockholders, a responsible enterprise also takes into account employees, suppliers, dealers, local communities, and the nation” (p. 5). The second view of CSR by Johnson (1971) is described as ‘long-run maximization’. In this view businesses should carry out social programs to be socially responsible to their organizations. Thirdly, ‘utility maximization’ is presented by Johnson as a view which assumes a firm is socially responsible when the prime motivation of an organization is utility maximization. This addresses that a firm should have more goals than just profit maximization. The fourth and last view of Johnson (1971) is ‘lexicographic view of social responsibility’, this view mentions strongly profit-oriented firms. Johnson (1971) argues that these firms engage in socially responsive behavior when they have reached their profits target. Manne and Wallich (1972) opened the debate on the definition of CSR, both authors argued that a definition of CSR should contain three elements: “(1) the setting of objectives, (2) the decision whether to pursue given objectives, and (3) the financing of these objectives” (p. 41). Manne and Wallich (1971) argued that CSR could be possible in certain circumstances, but they were both in favor of the stockholder instructions to the organization. In 1975, Backman was the first to write about the relation between CSR and corporate social performance (CSP). Backman’s (1975) definition of CSR was thus: “Social responsibility usually refers to the objectives or motives that should be given weight by business in addition to those dealing with economic performance (e.g., profits)” (p. 2). The dimensions of CSP are discussed by Sethi (1975); this resulted in three sorts of corporate behavior. The first corporate behavior is ‘social obligation’ it refers to organizations who respond to market or legal constrains. Second: ‘social responsibility’ goes beyond ‘social obligation’, in this form of corporate behavior the firm
14 acts “up to a level where it is congruent with the prevailing social norms, values, and expectations of performance” (p. 62). The third form of corporate behavior is social responsiveness. In this stage a firm is preventive and anticipatory of social needs (Sethi, 1975). Carroll (1979) argued that CSR was embedded in CSP, for top-managers to engage in CSP they need to have “(a) a basic definition of CSR, (b) an understanding/enumeration of the issues for which a social responsibility existed (or, in modern terms, stakeholders to whom the firm had a responsibility, relationship, or dependency), and (c) a specification of the philosophy of responsiveness to the issues.” (p. 499). Overall Carroll and Carroll (1999) identified that the definitions of CSR contained (1) business responsibilities to make profit; (2) obey the law; (3) go beyond these activities. Therefore Carroll (1979) proposed a new definition which embraced these three criteria: “The social responsibility of business encompasses the economic, legal, ethical, and discretionary expectations that society has of organizations at a given point in time (p. 500).”
Less definitions, more research
The interest in CSR did not fade away after the 1970’s. From 1980 and onwards the focus in the field of CSR made way for research and writings into alternative concepts. Alternative concepts, models and themes were presented, examples of these are: corporate social responsiveness, CSP, Public Policy, Business ethics, etcetera (Carroll and Carroll 1999).
A model proposed by Tuzzolino and Armandi (1981) presented a different mechanism to assess CSR: need-hierarchy framework. Tuzzolino and Armandi (1981) discussed that the definition of Carroll (1979) should be operationalized by this framework. Dalton and Cosier (1982) proposed a two by two matrix, on one axis ‘legal’ and ‘illegal’ and on the other axis ‘irresponsible’ and ‘responsible’. The four cells indicate the four faces which CSR could have. Figure 1 presents the matrix of Dalton and Cosier (1982).
15
Figure 1: The Four Faces of Corporate Responsibility
(Dalton & Cosier 1982, p. 20).
Dalton and Cosier (1982) first argue about the illegal / irresponsible cell (cell A). When operating an illegal and irresponsible business operation it’s about the risk of getting caught, and the consequences connected with these operations. Dalton and Cosier (1982, p.20) give the example of a car company which is obliged by law to install in every car a new pollution device, which costs 500 dollars. When the company does not install this device it has to pay an extra fine of 50 dollars. But the risk of getting caught is 1 out of 20. Dalton and Cosier (1982) argue that most, or all, companies facing this dilemma would not install the device. This is because the risk of getting caught, and the consequences, are very low. This is argued to be illegal and irresponsible behavior of MNEs.
The second cell (B) argues that the illegal / responsible cell raises interesting issues according to Dalton and Cosier (1982). As an example, they give the Cycle-Safe incident of
16 Monsanto1. Dalton and Cosier (1982) argue that the one potential strategy dealing with this problem is to challenge the law, laws can deliberately be violated for no other reason than to challenge the application. But this can be criticized by the public as well, because companies could use court battles as a form of delay – strategies (Dalton & Cosier 1982).
The third cell (C), legal and irresponsible, has shown many excesses in the history, some more serious than others; for example, a ‘diet pill’ which promised that a person taking this pill could eat anything at any time and still lose weight. This was true, but the pill contained so much tapeworm larvae that the person would starve to death taking these pills. Thus, criticizing companies who operate as a law-abiding citizen in cell B, society demands a higher standard than obliged by law.
The fourth and last cell (D) represents the responsible and legal cell. Even if a company operates a legal business and is responsible, it still can be criticized. Dalton and Cosier (1982) following the reasoning of Friedman, which argues that nowadays managers do not ‘own’ the business, but are primarily responsible for wishes of the stockholders. Managers thus act within the wishes of the stockholder, and any social actions beyond these wishes are involuntary redistribution of assets.
Finally Dalton and Cosier (1982) conclude that three fundamental principles should be considered when choosing a strategy. First ‘primum non necore’, this means “above all, knowingly do no harm” (Dalton & Cosier 1982, p.26). The second principle ‘organizational accountability’, holds that a organizations should be responsible for its impact. The third and final principle, ‘double standard’, which might have a negative connotation, refers to the power-responsibility equation. The larger a company gets the greater the influence it has on
1
Monsanto invested 47 million USD in developing plastic bottles, which could be recycled. Monsanto felt that this was social responsible behavior. In 1977 the Food and Drug Administration (FDA) banned this product. According to the FDA it was unsafe to use when the bottles where heated above 120 degrees for an extended period of time.
17 society. Dalton and Cosier (1982) argue that the MNEs should adopt the strategy of cell D, thus being responsible and legal.
Drucker (1984) shows an excellent example of the increasing interest of operationalization of CSR concept, and whether this has any relation to the financial performance. In 1984, Drucker presented a new meaning of CSR, with the new idea that profitability and responsibility were compatible notions. This trend was continued by Cochran and Wood (1984), who did an empirical study showing how profitable firms could also be socially responsible firms. Cochran and Wood used the Mokowitz index to measure the social performance of a firm (Carroll & Carroll 1999). Aupperle and Carroll et al (1985) used the four-part definition of Carroll (1979) to understand the relation between profitability and CSR. The definitional components where separated as ‘concern for economics’ and on the other hand ‘concern for society’. The results of the empirical study shows that not everyone sees the ‘concern for economics’ as a ‘concern for society’. Furthermore the societal orientation can be assessed by the importance it has for three non-economic components compared to economics (Carroll & Carroll 1999).
From CSR to CSP
According to Carroll and Carroll (1999) CSP is a major theme within CSR. From 1980 onwards the quest to go beyond ‘CSR’ increased. The models of CSP grew rapidly and earlier references have written about this notion (e.g. Carroll, 1977, 1979; Preston 1978; Sethi, 1975). Carroll and Carroll (1999) identify a trend to build a model for CSP. The first model is developed by Preston (1975). The focus of Preston’s matrix is on the management of social issues by corporations. Based on the assumption that managers follow stages of a process identified as corporate social involvement (Clarkson 1995). Preston (1975) defined the four stages of this process as follows: (1) awareness or recognition of an issue; (2) analysis and
18 planning; (3) response in terms of policy development; (4) implementation. However, the framework provided by Preston (1975) lacked a definition of a social issue. Furthermore there was no guidance for a researcher or manager to determine whether a social issue was one which the company should become concerned or involved with (Clarkson 1995). In 1979 a new model was introduced by Carroll, an advancement of the framework by Preston and a new conceptualization of CSP. In this model Carroll (1979) tried to reconcile corporate social and economic objectives, corporate responsibility with corporate responsiveness and to put a focus on the important element CSP (Clarkson 1995). Carroll (1979) argues that the three views: basic definition, an enumeration of the issues, and the specification of the philosophy of response, should be incorporate in corporate social performance. For the first part, the definition of corporate social responsibility it should embody the: legal, economical, ethical and discretionary categories of business performance (Carroll 1979). This four-part framework provides the definition: “The social responsibility of business encompasses the economic, legal, ethical and discretionary expectations that society has of organizations at a given point in time” (Carroll 1979, p500). For the second part of the model, the issues of social responsibility, there is no agreement upon because it varies in industries. Carroll (1979) uses the survey by Holmes (1976) which examined the important factors of social involvement by businesses. The top five factors were: “(1) Matching a social need to corporate need or ability to help. (2) Seriousness of social need. (3) Interest of top executives. (4) Public relations value of social action. (5) Government pressure” (Carroll 1979, p501). The third and final component of the three dimensional model is the philosophy mode, or the strategy behind business response to social responsibility and social issues. This can be placed on a continuum from ‘no action’ to ‘proactive action’. The assumption behind the philosophy mode is that businesses does have a social responsibility and the major focus is not on management accepting this moral obligation but on the degree of managerial action (Carroll
19 1979). The model proposed by Carroll (1979) was comprehensive and integrative, but the three-dimensional matrix did not lend itself for development of a methodology because it was complex and hard to test. Building on the model of Carroll (1979), Wartick and Cochran (1985) further developed the three-dimensional integration: responsibility, responsiveness and social issues (Carroll & Carroll, 1999). Wartick and Cochran (1985) identify an integrative model, to further operationalize social responsibility and performance. The major contribution was to recast the three aspects of Carroll (1979) into a framework of principles (of social responsibility), processes (of social responsiveness), and policies (of issue management). Wood revised this model in 1991. Wood’s model (1991) was based on the three dimensional model of Carroll (1979) and the Wartick and Cochran (1985) model. Wood reformulated the three dimension of the CSP model of Wartick and Cochran into three principles. First Wood (1991) took four domains of (Carroll 1979): economic, legal, ethical and discretionary and identified how these principles where related to CSR principles of social legitimacy, public responsibility and managerial discretion. Furthermore, Wood identified processes of corporate social responsiveness which Wartick and Cochran (1985) had formulated as policies, Wood highlighted these processes as: environmental assessment, stakeholder management and issues management. Lastly, Wood (1991) took Wartick and Cochran’s (1985) policies, and reorganized them under a new topic of concern-outcomes of corporate behavior (Carroll & Carroll, 1999).
According to Clarkson (1995) there is now a general agreement on the definition of CSP and its objective. This definition is formulated by Preston (1988): “corporate social performance was intended to suggest a broad concern with the impact of business behavior on society. The concern is with ultimate outcomes or results, not simply with policies or intentions; moreover there is some implication that these outcomes are to be evaluated, not simply described (p. 12).”
20 Wood (1991) observed that the concept of CSP has received serious theoretical and empirical attention, but the concept has not moved further than the framework of Wartick and Cochran (1985). According to Wood (1991) the reason for this is the lack of an appropriate level of analyses. In Table 1 the level of analyses is presented by Clarkson (1995). This table shows the level for Corporate social responsibility and responsiveness (institutional level), CSP (organizational level) and stakeholder management (individual level).
Table 1: Level of Analysis
Level of Analysis
Corporate social responsibility and
responsiveness (CSR1 and CSR2) Institutional Business Society
Corporate Social Performance (CSP) Organizational Corporations Stakeholder groups Stakeholder management Individual Managers Issues/relationships (Clarkson 1995, p.104)
To facilitate data collection, it was necessary to distinguish between stakeholder issues and social issues, because managers manage their relations with stakeholders and not with society. Table 2 represents the typical corporate and stakeholder issues.
21
Table 2: Typical Corporate and Stakeholder Issues
Typical Corporate and Stakeholder Issues
1. Company 1.1. Company history
1.2. Industry background 1.3. Organization structure 1.4. Economic performance 1.5. Competitive environment 1.6. Mission or purpose 1.7. Corporate codes
1.8. Stakeholder and social issues management systems
2. Employees 2.1. General policy
2.2. Benefits
2.3. Compensation and rewards 2.4. Training and development 2.5. Career planning
2.6. Employee assistance program 2.7. Health promotion
2.8. Absenteeism and turnover 2.9. Leaves of absence 2.10. Relationships with unions 2.11. Dismissal and appeal
2.12. Termination, layoff, and redundancy 2.13. Retirement and termination counselling 2.14. Employment equity and discrimination 2.15. Women in management and on the board 2.16. Day care and family accommodation 2.17. Employee communication
2.18. Occupational health and safety
2.19. Part-time, temporary, or contract employees 2.20. Other employee or human resource issues
3. Shareholders 3.1. General policy
3.2. Shareholder communications and complaints 3.3. Shareholder advocacy
3.4. Shareholder rights 3.5. Other shareholder issues
4. Customers 4.1. General policy
4.2. Customer communications 4.3. Product safety
4.4. Customer complaints 4.5. Special customer services 4.6. Other customer issues
5. Suppliers 5.1. General policy
5.2. Relative power 5.3. Other supplier issues
6. Public Stakeholders 6.1. Public health, safety, and protection 6.2. Conservation of energy and materials 6.3. Environmental assessment of capital projects 6.4. Other environmental issues
6.5. Public policy involvement 6.6. Community relations
6.7. Social investment and donations
22
Relating CSP to CTS
CTS and avoidance is entrenched in the corporate cultures of MNEs in most western economies (Taylor and Richardson 2012). These strategies of MNEs take advantages of gaps or loopholes in national tax legislation to reduce the effective corporate tax (Taylor and Richardson 2012).
Agrawal (2007) identified several areas of CTS: choice of country of residence, avoidance of double taxation, choice of organization, nature of income, use of own assets, choice of capital structure, buying and making of products, choice of accountancy system, filling of income return, corporate restructuring, foreign collaboration, or joint venture.
The literature on CTS is relatively young; the first author discussing CTS practices by MNE was Slemrod in 2001. This article claims that MNEs use an inter-related and globally orientated CTS to reduce effective corporate tax rates (Slemrod, 2001). Other authors such as Mills and Newberry (2004) identified that the taxable income of United States MNEs vary significantly by the tax rate of foreign countries where the subsidiaries operate. If operating in a country where the tax rate is low, the taxable income of the headquarters (HQ) in the US reports a low taxable income. In 2008 the authors Dyreng, Hanlon, and Maydew (2008) discussed several variables which could influence the level of tax avoidance culture of a MNE. In 2010 Dyreng, Hanlon, and Maydew complement this with the variable: Corporate executives, according to the article, it has a significant impact on the level of tax avoidance culture. Frank, Lynch, and Rego (2009) relate the tax aggressiveness and financial reporting aggressiveness. Book-tax differentials and aggressive use of tax avoidance show the relation between tax avoidance and financial reporting. Hanlon and Heitzman (2010) discussed twelve commonly used measures to identify tax avoidance2. The commonly used tax mechanisms
2
Measure of tax avoidance are: GAAP ETR, Current ETR, Cash ETR, ETR Differential, DTAX, Total BTD, Temporary BTD, Abnormal total BTD, Unrecognized tax benefits, Tax shelter activity, Marginal tax rate. These measures capture the average rate of tax per dollar of income or cash flow (Hanlon & Heitzman 2010).
23 that have recently gained major media attention are: transfer-pricing, re-invoicing, offshore ‘special purpose vehicles’, corporate inversions, dubious charitable trusts and other vehicles for tax abuse (Christensen and Murphy 2004). These are tax mechanism which can help to avoid corporate taxes in the country of origin, or where the business is exploited.
Several studies, for example Cochran and Woods (1984), have pointed out the direct relationship between CSP and financial performance. According to PWC (2003) CSP is not yet related to tax strategies, corporate governance is seen as an area for institutional investors with issues as: “reduction of corruption, collusion, nepotism; inadequate disclosure and insufficient transparency of financial statements; inadequate enforcement of existing rules, and a lack of clear separation of company ownership and management, all being seen as key areas of institutional concern” (Christensen and Murphy, 2004). In terms of relating CSP with corporate tax strategy, the agency or stakeholder theory is being scrutinized. In informal studies done by Christensen and Murphy (2004) CSR statements published by multinational companies, the companies did not include tax strategies in their CSR – agenda. This gap is recognized and researched in this study and has led to the following research question:
Research Question: How is corporate tax strategy (CTS) associated with corporate social performance (CSP)?
Hypotheses
The current debate in CSR-agendas is dominated by the CTS – issues. How MNEs should act in terms of their CSR or if CTS should be incorporated in the CSR – statements of MNEs. The research objective of this thesis is to identify if there is a possible negative or positive relation between CSP and CTS. Does CTS influence, positively or negatively, the CSP of MNEs?
24 upon the aggressiveness of the activity (Hanlon and Hetizma 2010). Tax avoidance is defined by Hanlon and Heitzman (2010) as: “the reduction of explicit taxes” (p.137). This is a very broad definition of Hanlon and Hetizma (2010). It incorporates the definition of Dyreng et al. (2008) which argue that tax avoidance: “reflects all transactions that have any effect on the firm’s explicit tax liability” (Hanlon and Heitzman 2010, p.137) . This thesis argues that CTS are tax avoidance mechanisms, thus reduction of corporate income taxes by MNEs resulting in a lower effective tax rate.
According to Lanis and Richardson (2011) linking the CSR to corporation taxation is currently encapsulated in the assumption of agency theory. This theory assumes the most important relationship is between managers and shareholders. This thesis relies on the views of Avi-Yonah (2008) and Schön (2008), who argue that a corporation exerts a significant influence on society which goes beyond the relationship between managers and shareholders. A corporation should develop policies, strategies and operations that provide favorable outcomes for both the company and the environment. Freedman (2003) argues that when a company does not pay its “fair-share” of taxes, it would generate hostility and can damage reputations, which negatively affect the corporation’s CSP. Thus, pursuing a CTS should have a negative impact on the corporate social performance of that company. To test this relationship the first hypothesis is:
H1: CTS of a MNE has a negative effect on its CSRI – score
The second hypothesis is to identify whether incorporating the ETR in their financial statement should have a positive effect on the CSR-score of that company. Operating a corporate tax strategy is usually done in a discrete manner. MNEs try to avoid the public,
25 because according to Freedman (2003) it will lead to reputational damage which is reflected in their corporate social performance and CSR-score. Jenkins and Newell (2013) argue that publishing ETR would have a positive effect on the CSR-score. This is because MNEs which publish their corporate tax strategy or ETR take the lead in the CSR-agenda, which should incorporate the transparency on tax strategies. On the CSR-agenda transparency is key for being a good corporate citizen. This transparency of MNEs on their corporate tax strategies will lead to a higher CSR-score. The second hypothesis is stated as follows:
H2: Incorporating ETR in a MNE's financial statement has a positive effect on its CSRI- score
The third hypothesis tries to identify the effect when the company both uses CTS and publishes its ETR in the financial statement. This hypothesis is based on a study of PWC (2013), in this CEO survey one of the key conclusion is that CEO’s do not see CTS as part of the CSR-agenda. Taking this in consideration, Jenkins and Newell (2003), argue that being transparent on CTS has a positive effect on the CSR-score. Thus, when operating CTS and publishing this CTS in the financial statement of a MNE, the reporting of CTS in financial statement should have a moderate positive effect on the relation between CTS and CSR-score. Thus the third hypothesis is:
H3: Incorporating CTS in a MNE's financial statement has a positive effect on the relationship between ETR and its CSRI- score.
26
3. RESEARCH DESIGN
The intended domain of generalization of this thesis is to study the effect of CTS on the CSP of a MNE. This research design elaborates on the study, and how this is conducted to be reproducible. The nature of the study is quantitative. First, this thesis will elaborate on the concept of quantitative research. Second, it discusses the dataset. Third, it discusses the (in)dependent variables and operationalization of these variables. Fourth, this chapter discusses the internal, external and construct validity. The fifth part explains the statistical analyses. The last chapter concludes with limitations of the study.
Quantitative research
The research objective is to find an association between the concepts of CSP and CTS. This is accomplished through quantitative research. According to King et al. (1994) many studies do not fit in the characterization of qualitative or quantitative research. Often, studies combine both. The concepts studied here could be researched through a case-study, which is amenable to statistical analyses. However, in this study it is preferred to do a quantitative research since the thesis is identifying an effect between two concepts.
This quantitative study uses statistical analyses. This allows the thesis to test whether there is an effect between the (in) dependent variables, and how strong or weak this effect is. To test the hypotheses this study uses multiple-regression analysis and two-way anova statistical test.
Dependent variable
CSP
According to Dodge (2006), the dependent variable is also known as the response variable, the measured variable or outcome variable. In this research, the outcome variable is the CSP
27 of a MNE.
The data for the dependent variable is collected at the Reputation Institute3 (Reputation-Institute 2014a). Every year this institute publishes Corporate Reputation and Social Responsibility Rankings. For both rankings it rates the corporate social performance of MNEs. This sample consists of 193 MNEs for the years 2011 and 2012. The commercial activities are not engaged in investment trusts, holding companies or wholly owned subsidiaries from foreign MNEs.
The CSRI-index is not based on self-administered questionnaires from the world-wide public, but each respondent had to rate five MNEs based on their CSR experience with the MNE. This index incorporates almost 24.000 ratings, in January and February 2011 and 2012. A minimum of 100 respondents was needed to have a significant effect on the CSRI- index (Reputation-Institute 2013).
Independent variables
CTS
The independent variable is the variable according to Dodge (2006): the predictor variable, manipulated variable or explanatory variable. The dataset shows whether the MNEs use CTS and whether these are published in their CSR-statements. For this reason the CTS is the independent variable.
CTS are difficult to study, when tax planning MNEs minimize effective tax rates, this is mostly done in a discrete manner. Hanlon and Heitzman (2010) discuss the most commonly
3
Reputation Institute is the world’s leading reputation-based advisory firm. Founded in 1997 by Dr. Cees van Riel. The expertise of the Reputation institute is to help organizations to improve their relationship with stakeholders, signature analyses are RepTrak® model for analysing the reputations of companies and institutions. Best known by the Forbes-published Global RepTrak®, the world’s largest study of corporate reputations (Reputation-Institute 2014b).
28 used measures for tax avoidance. Not all the measures are appropriate for all research questions. In the effective tax rate measures this study uses the Current ETR of a MNE to determine whether the MNEs use CTS to avoid income taxes. This measures the current tax expense per dollar of pre-tax book income. It reflects the deferral strategy of an MNE, it shows the non-conforming avoidance of an MNE and is computable per jurisdiction, and it could reflect the impact of an accountant’s earnings. This measure is derived by the following equation:
Equation 1: Operationalization of CTS
Because Hanlon and Heitzman (2010) discusses that the actual effective tax rate is a few percent above or beneath what the calculated tax rate could be, this thesis takes a standard error measure of 5%. The calculated tax rate will be compared with the corporate income tax rate of the country of origin of the MNE. When a MNE uses CTS, it will be operationalized by ‘1’ and when it’s not pursuing CTS it will be given a ‘0’.
Financial Statement
This independent variable is created for the second hypothesis to understand the effect on CSRI - score when a MNE publishes its ETR in their financial statement. For the whole sample the financial statement is collected, and looked for its ETR. When a company publishes its ETR in their financial statement it will be given a ‘1’ and if not a ‘0’.
29
Control Variables
Industry
This thesis controls for certain industry types, because Sweeney and Coughlan (2008) argue that certain industry types benefit from being more or less socially responsible. This thesis uses the Industry Classification Benchmark 2011 which consists of the following industries:
1) Basic Materials (Basic Resources, Commodities, Chemicals);
2) Consumer Goods (Automobiles and Parts, Food & Beverage, Personal and Household Goods);
3) Consumer Services (Media, Retail, Travel & Leisure); 4) Financials (Banks, Financial Services, Insurance);
5) Health Care (Health Care Equipment, Pharmaceuticals and Biotechnology); 6) Industrials (Construction and Materials, Industrial Goods and Services); 7) Oil & Gas (Oil and Gas Production and Distribution);
8) Technology (Software and Computer Services and Hardware and Equipment); 9) Telecommunications (Fixed Line and Mobile);
10) Utilities (Electricity, Gas, Water). (ICB 2010)
A dummy variable will be used for this classification. A MNE will be coded ‘1’ when it operates in one of these industries and ‘0’ when it cannot be classified as a particular industry. This classification excludes Energy, because this can be perceived as the industry Oil & Gas.
Firm Size
This thesis also controls for firm size, because according to Udayasankar (2007) smaller firms have less financial resources to spend on CSR. This will lead to less visibility on the CSR continuum. Jenkins (2006) argues that small- and medium-sized enterprises are more dynamic
30 in responding to changing environments. Thus this thesis controls for company size, and uses the Fortune Global 500 for data to rank the companies.
Validity and Reliability
In order to measure the quality of a study, the concept’s validity and reliability are introduced. When considering quantitative research, reliability “pertains to the consistency of the research findings” (Kvale and Brinkmann 2009). And validity, according to Messsick (1990) is the “integrated evaluative judgment of the degree to which empirical evidence and theoretical rationales support the adequacy and appropriateness of interpretations and actions based on test scores or other modes of assessment.” (Messisk, 1990). Usually, there are three types of validity discussed in academic research: content validity, predictive and concurrent criterion-related validity, and construct validity (Messick 1990).
Content validity evaluates how well the content of the samples represent all facets of a social construct. The social construct in this study is the effect on CSRI score, the samples CTS are investigating to determine if there is an effect. Thus, the content in this case is the CSRI - score and is studied by looking at the direct effect of CTS. The study is about the role and effect of operating CTS.
Criterion validity, according to Messick (1990), “is evaluated by comparing the test scores with one or more external variables (called criteria) considered to provide a direct measure of the characteristic or behavior in question”. Criterion validity refers to predictive validity, which determines if the test is generalizable and confirms the theory. The study tries to identify a new variable which influences the CSRI-score of a MNE. This study can generalize for sector and firm size, because it incorporates a control variable. Since this research studies a direct effect on the performance of an index, and the ceteris paribus, the influence of CTS on CSRI score can be measured directly.
31 Finally, construct validity according to Messick (1990): “investigating what qualities a test measures, that is, by determining the degree to which certain explanatory concepts or constructs account for performance on the test”. In order to obtain high construct validity, the research should incorporate all control variables. This thesis has controlled for the following: firm size and industry. Doing so increases the construct validity because these external variables can influence the effect studied.
Method of analyses
This study uses the ‘multiple regressions’ method of analysis, to test the hypothesis stated in the literature review. This method of analysis is regarded as the most relevant because the study incorporates two independent variables, next to two control variables. According to Berry and Fieldman (1985) the use of a multiple regression analysis is an important tool for non-experimental data. In general, according to Berry and Feldman (1985), “the linear regression model assumes that for each set of values for the k independent variables (…) there is a distribution of Y j values (…)”. The following linear equation will be used; it includes control variables and the standard error:
Equation 2: Standard Equation by Berry and Fieldman (1985)
32
Equation 3: Equation of study
Furthermore, for hypothesis 3 this thesis will run a two-way anova statistical test, because it is studying the moderation effect of financial statement on the relationship between CTS and CSP. This two-way anova is chosen because Financial Statement and CTS are categorical, but the dependent variable is numerical.
This study uses the open source software of IBM SPSS statistics 21.
Conceptual model of hypothesis
33
Limitations
All academic research suffers from limitations (Bryman 1984). The following limitations are recognised in this study.
First, the information provided by MNEs on CTS is very scarce and limited. MNEs operating tax planning activities will do this in a discrete manner. Thus, it is difficult to identify MNEs who operate CTS.
As said, MNEs will operate CTS in a discrete manner. Thus, CTS should be tracked and traced. The Hanlon and Heitzma (2010) measurements to track CTS, which are commonly used, are not 100% accurate. The effective tax rate measurement should be careful when making inferences about tax avoidance: “If samples under consideration contain firms with differing levels of importance placed on financial accounting earnings” (Hanlon & Heitzma, 2010).
34
4. RESULTS
Descriptive statistics
Table 3 provides the descriptive statistics results for the dependent and independent variables, as well for the control variable company size. For this sample, 61% of the MNEs use a corporate tax strategy, while 66% publish their ETR in their financial statement. This means that MNEs, while pursuing a CTS, publish their ETR in their financial statement (O’brien 2007).
Table 3: Descriptive Statistics
Descriptive Statistics
N Range M SD Minimum Maximum
CSR Score 193 17.50 69.23 3.89 61.53 79.03
CTS 193 1.00 0.61 0.49 0.00 1.00
Company Size 193 480.00 191.80 135.06 8.00 488.00
Financial Statement 193 1.00 0.66 0.48 0.00 1.00
The lowest CSR score is 61.53 and the maximum is 79.03. This sets the range at 17.5. The lowest score for company size is 8, the biggest company in this sample size is the 8th biggest company of the world. The smallest company in the sample is ranked at 488 in the Fortune Global 500 ranking.
Statistical analyses
Correlation analysis
In table 4 the correlation matrix is showed for all the independent variables. This matrix shows the strength of the relationship between the variables, using Pearsons correlation coefficients. Overall, the strength of the positive relationship between the variables is very weak, all less than 0.23. This highest relationship is between financial statement and CTS.
35 The negative relationship between the independent variables is less weak, the highest negative relationship is -0.38. This is the strength of the relation between Consumer Service and Company Size
36
Table 4: Summary Statistics and Correlation Matrix
Summary Statistics and Correlation Matrix
M SD 1 2 3 4 5 6 7 8 9 10 11 Independent Variable 1. CTS 0.61 0.49 1 2. Financial Statement 0.66 0.48 .23** 1 Control Variables 3. Company Size 2.35 0.44 .18 .07 1 4. ENERGY 0.01 0.10 -.02 -.14* -.13 1 5. MATERIALS 0.03 0.16 -.14 .05 .05 -.02 1 6. INDUSTRIALS 0.14 0.35 .11 -.09 .18* -.04 -.07 1 7. CONSUMERDIS 0.22 0.42 -.08 -.06 -.38** -.06 -.09 -.22** 1 8. CONSUMERGOOD 0.17 0.38 -.06 .10 .11 -.05 -.07 -.18* -.24** 1 9. HEALTHCARE 0.05 0.21 .75 .16 .13 -.02 -.04 -.09 -.12 -.10 1 10. INFORMATION 0.17 0.37 .18* -.06 .06 -.05 -.07 -.18* -.24* -.20** -.10 1 11. TELECOMMUNICATION 0.01 0.10 -.02 .07 -.04 -.01 -.02 -.04 -.06 -.05 -.02 -.05 1 * Correlation is significant at the 0.05 level (2-tailed).
** Correlation is significant at the 0.01 level (2-tailed). .
37 It is important to rule out the multicollinearity between the variables. This is shown in table 5, all the variables should have a VIF- score of less than 5, and a tolerance score bigger than -.10 (O’brien 2007). This means that none of the independent variables show signs of multicollinearity.
Table 5: Collinearity Statistics
Collinearity Statistics Stand Coefficients Beta t Tolerance VIF CTS .07 -0.69 .80 1.25 Financial Statement .27 2.87** .86 1.17 Company Size -.13 -1.32 .83 1.21 ENERGY -.07 -0.72 .87 1.15 MATERIALS -.003 -0.03 .75 1.33 INDUSTRIAL .05 -0.37 .41 2.42 CONSUMER .18 1.19 .31 3.23 CONSUMERGOOD .06 0.48 .41 2.46 HEALTHCARE -.08 -0.68 .54 1.86 INFORMATION .23 1.58 .35 2.87 TELECOMMUNICATION -.20 -2.08** .80 1.22
Dependent Variable: CSR2012a **p< .05
Regression Analysis
Table 6 show four regression models. Model 1 in table 6 only incorporated the control variables. Model 2 added the independent variable CTS, model 3 added financial statement and model 4 incorporated all the variables. Model 1, 3 and 4 showed to be significant at a confidence level of 95%. Model 2 showed to be not significant. The following paragraphs will elaborate on the models of table 6.
38 The adjusted R square of model 1 and 3 are 6.4% and 13.7%, respectively. It shows the amount of variation in the model explained by the selected set of variables. Model 2 is not significant. The variation of model 4 which incorporated all the selected independent variables is 13.3%.
Model 1 only incorporated the control variables. The control variables show the study that a higher ranking on company size, thus a smaller firm, leads to a higher score on the CSR-ranking. Furthermore, operating in telecommunication and energy industry, leads as well to a higher score on the CSR-ranking.
Model 2 is designed to give an answer to the first hypothesis which predicted the negative relationship between the CTS and the CSR-score of the company. Model 2 shows to be insignificant at the confidence level of 95%. Thus, it fails to prove that CTS has a negative effect on the CSR-score of that company.
Model 3 shows the regression model including the control variables and the independent variable financial statement. This independent variable is significant and shows that publishing the CTS of a company in their financial statement has a positive effect on the CSR-score of that company.
Model 4 shows to be significant. The adjusted R square, and thus the variation in the model, is explained by the selected set of variables, 13.3%. Publishing the ETR in their financial statement and operating in telecommunication has a positive effect on the CSR-score of a company. Table 6 sums up all the scores for the models.
39
Table 6: Regression Analysis
Regression Analysis with Dependent Variable CSR score
Models 1 2 3 4 Sig. F Change .06 .15 .003** .005** R2 .14 .15 .21 .21 Adj. R2 .07 .08 .14 .13 Constant 0 0 0 0 Independent Variable CTS .15 .50 Financial Statement .002*** .005*** Control Variables Company Size .27 .21 .21 .19 INDUSTRIALS .85 .87 .70 .72 CONSUMERSERVICE .24 .16 .28 .24 CONSUMERGOODS .59 .46 .71 .63 HEALTHCARE .83 .91 .45 .50 INFORMATION .14 .13 .12 .12 TELECOMMUNICATION .07* .10 .03** .04** ENERGY .26 .38 .41 .48 MATERIALS .93 .86 .86 .97 1. Control variables
2. Model one + independent variables * p < .1
** p < .05 *** p < .01
Assumptions for the regression model
To be able to use and interpret these results, a set of assumptions should be checked before using a multiple regression analysis. The first condition which should be met for a multiple regression, is that all variables should be normally distributed. All variables have met this condition, except for the dependent variable CSR-score and independent control variable Company Size.
40 Both the dependent variable CSR-score as well the independent control variable Company Size failed for the normal distribution requirement. Both variables should be transformed. This is accomplished using SPSS to compute the variables into the numeric expression, LG(CSR)10 and LG(CompanySize)10. Figures 3 and 4 show the transformed variables and histogram which clearly indicate a normal distribution. All of the other requirements for the regression analysis are met for all variables.
41
Figure 4: Histogram with normal curve for Company Size
Furthermore, this thesis should check the sample on heteroskedasticity by using the Breusch-Pagan tests (Breusch and Pagan 1980). This test is calculated as follows:
Equation 4: Breusch – Pagan tests
Before this is done, a new variable, g, should be created using “unstandardized residuals” and the “unstandardized predicted values”. This is done by using a Macro of Withley in a new syntax of SPSS (SPSTOOLS.net 2002). Table 7 shows the sum of squares of this regression.
42
Table 7: F-tests and Sum of Squares
F-test and Sum of Squares of Model 1
Sum of Squares df Mean Square F P
Regression 24.29 11 2.22 .62 .81
Residual 383.82 107 3.59
Total 408.20 118
In the Breusch-Pagan tests the null hypothesis stands for homoscedasticity. To calculate the p-value first the B-p-value should be calculated. This is done for the regression and to calculate B the Sum of Square should be multiplied with 0.5, this results in the B-value of 12.19. To test the null hypothesis the p-value of chi square tests is used and the two-tailed p-value is 0.349. This p-value is greater than the allowed p-value of 0.05. The null hypothesis cannot be
rejected. Thus the sample and regression does not suffer from heteroskedasticity (Breusch and Pagan 1980).
Two – way anova
For the third hypothesis a second test is used to test the moderation effect. According to De Vocht (2009) this moderation effect should be done by a two-way anova. Because this test only allows a maximum of 10 variables, the control variable industry is transformed in a categorical variable, from 1-10. Table 8 shows the following results for the two-way anova.
43
Table 8: Two-way Anova
Two-way Anova with Dependent Variable CSR
df F p Partial Eta Squared CTS 1 4.26 .048* .12 Financial Statement 1 2.71 .110 .08 Company Size 95 1.49 .106 .82 Industries 7 1.78 .127 .29 CTS*Financial Statement 1 6.90 .013* .18 a. Dependent variable: CSR
b. R Squared = .860 (Adjusted R Squared = .305) *p<0.05
The results show that the interaction of CTS on the dependent variable is less strong (partial eta squared = 0.12) than when the interaction effect of CTS and Financial Statement on the dependent variable (partial eta squared = 0.18). Both of the control variables are insignificant and no conclusion can be drawn from the provided information. Hypothesis three can be accepted.
44
5. DISCUSSION
For this study, the relationship between ETR and a higher CSR-score showed to be not significant. Although the literature claims to view the firm embedded in society, as Freedman (2003) argues is the case in the “the real world”, this thesis could not base any conclusion on the regression model. The “real world” holds that the firms are interdependent. Firms are affected by society, and should see society as a major stakeholder. It should not only consider the relationship between managers and stakeholders. This theory holds that when a firm negatively affects society it backfires on the firm, making the firm more difficult to operate. Thus, following Freedman’s (2003) arguments, a firm should develop policies, strategies or operations which pursue the companies organizational goals and do not neglect important stakeholders, such as shareholders or the embedded society.
The study could not base any conclusion on the results. This could be due to a relatively small sample (n=193). Furthermore because MNEs operate CTS in a very discrete manner, the variable which operationalizes whether a MNE is using CTS could not have been 100% accurate. Christensen and Murphy (2004) argue that paying corporation tax is one of the most fundamental ways to comply with social contract, therefore major stakeholders such as shareholders must see paying tax as their prime duty. CTS could not be associated with a lower CSR-score. Hypothetically, when this thesis could identify a negative relationship between both concept, it would be argued that operating CTS could backfire on a MNE in the long run for their financial performance. This is theoretically grounded in the very early work of scholars such as Drucker (1954), but also more recent work by Orlitzky, Schmidt et al. (2003) which indicated that there is a significant relationship between CSP and corporate financial performance (CFP). The final conclusions of the first hypothesis: ‘CTS of a MNE has a negative effect on their CSRI – score’, is rejected on the ground that the statistical tests where not significant in the regression model (3).
45 The second hypothesis theorized that when a company is transparent on its CTS, this transparency will lead to a higher CSR-score. This is tested in a multiple regression, the adjusted r-square shows the thesis that including the variables statement and all control variables, the predictor variables influence the dependent variable for 13.7%. The regression model shows that the financial statement is significant and the theory is confirmed. Jenkins and Newell (2013) argue that transparency is one of the key factors of CSR. Thus when MNEs are transparent in their financial report on CTS, this will lead to higher CSR-score. The statistical significance (p<0.05) shows that indeed when a company incorporates its CTS, it will have the effect that the company obtains a higher CSR-score. The second hypothesis is therefore accepted and the theory is tested and confirmed.
The third hypothesis emphasized that according to a survey of PWC (2013) the CEO’s of MNEs do not relate CTS as part of the CSR-agenda, this leads to the question whether when companies pursue and are transparent on their CTS what this will imply for the relationship between CTS and a higher CSR-score. This last relationship has shown to be significant, thus publishing the CTS in financial statement has a significant effect on the relationship between CTS and CSR (p=0.013). More transparency thus leads to a positive effect on the relationship between CTS and CSR. The theory of Jenkins and Newell (2013) is correct because they argue that being more transparent will lead towards a better CSR-score. This moderating effect is tested by a two-way anova, and showed to be significant as well. The Partial Eta Squared (=0.18) shows the study the amount of variation to the dependent variable CSR. Being transparent on which CTS a MNE operates will have a positive effect on the relationship between CTS and CSR, so for MNEs it is important to understand that CTS is definitely part of CSR. The third hypothesis: ‘Incorporating CTS in a MNE's financial statement has a positive effect on the relationship between ETR and its CSRI - score, is accepted.
46 This thesis used several control variables, because these could have significant influence on the dependent variables. Udayasankar (2007) argued that small companies have less means to spend on CSR-activities and thus will be less visible on the CSR-spectrum. For all regression models, company size did not identify a significant relationship on CSR-score. The second control variable industry type showed that the companies which operate in the industry information will be likely to obtain a higher CSRI - score. Because in the regression models the industry information showed to be significant (p<0.05).
Strategical implications for managers
The findings of this study could have implication for the decisions of managers regarding policies, strategies or operations. Using CTS could have a negative effect on the corporate social performance, but it’s beyond the scope of thesis to give any directions on the possible relationship between CTS and the financial performance of this company.
What this study identifies for managers, is that being transparent on whether a MNE uses CTS or not will lead to a higher CSR- score. The moderation effect of publishing the CTS in financial report would have on the relationship between CTS and CSR-score is clear as well. The theory and study of this thesis indicates that it would have a positive impact and therefore should be taken in consideration. It is important for managers to moderate the potential negative effect of CTS on CSR.
47
6. CONCLUSION
In the conclusion of this thesis, the research question will be discussed and answered. The intended domain of generalization of this thesis is what the effect of CTS is on its CSR-score, and thus corporate social performance. Furthermore, it is the objective to understand and find a relationship between publishing their ETR in the financial reporting and their CSR-score and what the moderation effect is of this published ETR on the relationship between CTS and CSR. These questions are developed to answer the research question: ‘How is corporate tax strategy (CTS) associated with corporate social performance (CSP)?’
The introduction of this thesis showed that the debate on operating CTS is very topical. Google, Apple, Facebook and Amazon are a few MNEs who received global media attention on their CTS behaviour. CTS, or tax avoidance, tax shelter, tax planning etc., is focused on creating strategies to avoid corporate income taxes in the country where headquarters is settled. Though MNEs increasingly operate these strategies, academics merely question what the effects could have for their CSR and CSP. Nonetheless, no empirical study has been done to find a negative relationship between CTS and CSR.
The literature review of this study introduced the main concepts of this thesis. CSR is an old concept, with many scholars discussing definitions and how to measure it. The other concept, CTS, is a more modern concept, and increasingly academics have shown interest in understanding and developing the definitions of CTS. MNEs will operate CTS as secretive as possible, which could make it more difficult to study this concept. The research gap this thesis recognises is the lack in studies which focus on the relationship between CTS and CSR-score. Therefore, this thesis research question is: “How is corporate tax strategy (CTS) associated with corporate social performance (CSP)?”. As stated in the literature review, three hypotheses were developed to answer this research question.
