Dissertation What drives Corporate Social Responsibility Performance? The Role of national Institutions.

(1)

What drives Corporate Social Responsibility Performance? The

Role of national Institutions.

Daniel Quarterman

RUG: S2867826 NUBS: 140002333

Dissertation

Advanced International Business Management Dual Award

Supervisors:

Dr. M.J. Klasing Dr. S. Reissner

7th_{December 2015}

Word Count: 14924

Faculty of Economics and Business

University of Groningen

Newcastle University Business School

(2)

ii

Abstract

In the absence of global legally binding Corporate Social Responsibility (CSR) regulation, understanding drivers of CSR performance is important. Therefore, this research investigates the extent to which firm-level variation in CSR performance is explained by country-level differences. Following recent insights, items of significant importance to firm-level variation in CSR performance are national institutions. Addressing the national institution and CSR performance link, this project draws on the comparative capitalism literature, the CSR literature and the direct research field comparative CSR to develop hypotheses. In contrast to preceding research in addition to CSR performance, the CSR performance dimensions social, environmental and governance are addressed individually. Findings based on cross-sectional data for 1231 firms from 20 countries proposes that market freedom is positively related to CSR governance performance. Moreover, the availability of skilled labor is negatively related to all CSR performance dimensions, while trade union strength is positively related to all CSR performance dimensions. Notably, across the CSR performance dimensions there are major differences in significance and effect size.

(3)

Acknowledgements

First and foremost, I would like to thank my supervisors Dr. Stefanie Reissner from Newcastle University and Dr. Mariko Klasing from the University of Groningen for their guidance, patience and valuable support during the last few months.

Furthermore, I am very grateful for the support provided by my parents. It enabled me to focus on my studies, since I did not have to worry about much else.

(4)

iv

List of Abbreviations

Adj = Adjusted Avg = Average

CME = Coordinated Market Economy Coef = Coefficient

CSR = Corporate Social Responsibility CV = Control Variable

D = Data

DV = Dependent Variable GDP = Gross Domestic Product

GICS = Global Industry Classification Standard GJ = Gigajoule

ICTWSS = Institutional Characteristics of Trade Unions, Wage Setting, State Intervention and Social Pacts

IV = Independent Variable LME = Liberal Market Economy M = Method

N = No

N/A = Not Available

OECD = Organization for Economic Cooperation and Development OLS = Ordinary Least Squares

PPP = Purchasing Power Parity ROA = Return on Assets T = Theory

USD = US Dollar

VoC = Varieties of Capitalism Y = Yes

(8)

viii

Table of Tables

Table 2.1: Overview CME and LME ... 9

Table 2.2: Literature Review Comparative CSR ... 13

Table 3.1: Distribution of Sample Firms across Countries ... 27

Table 3.2: Distribution of Sample Firms across Sectors ... 36

Table 3.3: Variable Overview ... 40

Table 4.1: Main Regression Results ... 41

Table 4.2: Overview Regression Results ... 46

Table 4.3: Robustness Test Regression Results Countries > 5 Companies ... 47

Table 4.6: Robustness Test Regression Results averaged Data ... 51

Table 4.7: Overview Main Results and Robustness Tests ... 52

Table 4.8: Effect Sizes by CSR Performance Dimension ... 56

Table 4.9: Effect Sizes by institutional Variable ... 57

Table 7.1: Illustrative Indicator Overview Bloomberg and Sustainalytics ... 75

Table 7.2: Overview descriptive Statistics ... 77

Table 7.3: Test for Multicollinearity ... 78

Table 7.4: Breusch-Pagan Test for Heteroscedasticity ... 79

(9)

Table of Figures

Figure 2.1: Conceptual Model ... 23

Figure 3.1: Bloomberg and Sustainalytics Methodology ... 31

Figure 7.1: Intra-class Correlation GICS Sector ... 77

Figure 7.2: Intra-class Correlation Country ... 78

Figure 7.3: Visual Test for Heteroscedasticity ... 79

Figure 7.4: Normality of Residuals via P-P Plot ... 80

(10)

Chapter 1: Introduction

This chapter starts by presenting Corporate Social Responsibility (CSR) and CSR

performance in the wider context of international business. Building on this contextualization, the research gap this project addresses is identified. Based on this research gap a research questions is introduced. Subsequently research methods and findings are briefly outlined. Finally, the structure of the research project is discussed in order to provide a stepping-stone for more detailed exploration of the remainder of the dissertation.

1.1 Background

Spurred by globalization, the way in which business is conducted is continuously changing. Referring to business practices of outsourcing and subsidiary firms, there is a general

tendency of disintegration of production (De Marchi, Di Maria and Ponte, 2014). Accordingly the production processes of a company are not limited to a single country anymore but

commonly happen in a transnational environment (De Marchi et al., 2014).

This change in the business environment has significant implications for the companies in question. Traditionally, it has been considered the role of the state to mediate the relationship between business and society by curbing distortions created by unregulated business (Jenkins, 2005; Merk, 2011). However, due to the change in business practices, the problem of global reach of corporations in contrast to the domestic reach of structuring regulation arises (Aguilera, Rupp, Williams and Ganapathi, 2007).

In the absence of a strong global legal framework and faced with demands by a heterogeneous group of stakeholders, corporations have turned to voluntary self-regulation as an alternative (Jenkins, 2005; Hack, Kenyon and Wood, 2014). Accordingly many firms invest into business practices related to CSR. Although CSR is not precisely defined, in commonly takes into account the social and environmental impact of a company as well as governance structures within the company (Fransen, 2012).

(11)

dimensions: CSR performance, CSR social performance, CSR environmental performance and CSR governance performance. Some companies by far outperform others regarding their CSR performance across these dimensions (see e.g. Thomson Reuters Datastream, (2010), Sustainalytics (2015); Sustainability-indices, 2015). While these rankings enable direct firm to firm comparison, the indicators hold no information about possible antecedents to the CSR performance achieved by companies. Considering the voluntary nature of CSR practices, the objective of this research is to investigate the extent to which firm-level variation in CSR performance is explained by country-level differences.

In order to engage with the research objective, this research follows insights that items of significant importance to variation in CSR performance, are national institutions (Ioannou and Serafeim, 2012; Peng, Dashdeleg and Chih, 2014; Thanetsunthorn, 2014; Hartmann and Uhlenbruck, 2015). In the context of this research, institutions refer to "a set of rules, formal or informal, that actors generally follow, whether for normative, cognitive, or material

reasons, and organizations as durable entities with formally recognized members, whose rules also contribute to the institutions of the political economy" (North, 1991, p.3)

To build a theoretical foundation as a point of departure, this research employs the Varieties of Capitalism (VoC) framework by Hall and Soskice (2001). The framework was developed to assess the implications of national institutions on firm level outcome variables. The framework links the national institutional environment by assessing the influence of institutions on a firm's capacity to build relationships. Following the logic of the VoC

framework, these relationships are essential for firm survival as they enable access to markets, suppliers and peers. In underscoring the importance of the firm to coordinate relationships across five different spheres (governance, inter-firm relations, capital–labor relations, education and training) the framework moves the firm to the center of the analysis (Hall and Soskice, 2001). Accordingly national institutions can directly be linked to the firm level outcome variable CSR performance.

(12)

relationships through market mechanisms; while firms in CMEs rely more heavily on non-market forms of coordination (Hall and Soskice, 2001).

1.2 Research Gap

Even as institutions have been identified as items of significant importance to variation in firm-level CSR performance across countries, the relationship between the institutional environment and CSR performance remains highly disputed. Regarding the VoC framework employed by this research project, there are two contradicting streams of literature, which have each produced empirical evidence. These streams run along the division the VoC draws between CMEs and LMEs. One stream of literature (see e.g. Middtun, Gautesen and

Gjolberg, 2006; Gjolberg, 2009b), claims that institutions related to the CME lead to higher CSR performance. While the second stream postulates that institutions related to the LME induce greater CSR performance (see e.g. Jackson and Apostolakou, 2010).

The debate regarding the influence of institutional variables is further complicated by the wide spread use of an aggregated CSR performance variable in this field of research. Because of the use of an aggregate CSR performance variable instead of addressing the underlying social, environmental and governance dimensions of CSR research lacks specificity regarding the dimensions of CSR. This lack of specifity, threatens the validity of previous research findings, as accordingly it is unclear which national institutions affect which CSR dimensions and if effects vary among dimensions. This has most likely also contributed to the mixed empirical evidence brought forward so far.

Contemporary CSR practices, are understood as exceeding what is legally required (Gjolberg, 2009b; Hack et al., 2014). Considering the broad interest in CSR among stakeholders, the question arises how companies can be convinced to invest into CSR. Accordingly addressing this research gap is relevant, as more detailed understanding of the relationship between institutions and CSR performance opens up avenues to influence companies in the absence of legal requirements.

1.3 Research Question

(13)

national institutions (according to the VoC framework) and individual CSR performance dimensions?"

1.4 Research Methods

Addressing the research question, this project builds on a positivist research philosophy and furthermore draws on a deductive research approach. Accordingly the literature relating to comparative capitalism, CSR and comparative CSR was reviewed and hypotheses were developed.

In line with similar work in the research field of comparative CSR (see e.g. Ioannou and Serafeim, 2012; Thanetsunthorn, 2014; Hartmann and Uhlenbruck, 2015), the sample used to test the hypotheses is determined by the availability of CSR performance data. This research relies on Bloomberg in collaboration with Sustainalytics as a data provider. Tackling the research question, this study draws on cross-sectional data from 1231 firms in 20 countries1. For these companies institutional variables were supplemented based on their headquarter location. Relevant institutional variables which have been identified for this research project include market freedom, the availability of skilled labor and trade union strength. In order to isolate the relationship of interest between national institutions and CSR performance, multi-level control variables were implemented. Control variables include economic development and level of internationalization at the country level, sector dummy variables at the sector level and visibility, short-termism, financial performance, size and age at the firm level.

For assessment of the data, Ordinary Least Squares (OLS) regression was utilized; owing to the multi-level structure of the data involved, clusters were implemented.

1.5 Findings

Following the analysis, this research project proposes that market freedom is not significantly related to CSR performance, CSR social performance and CSR environmental performance; while there is a significant and positive relation to CSR governance performance. Moreover, the availability of skilled labor is negatively related to CSR performance across all CSR performance dimensions. Finally, the strength of trade unions is positively related with all dimensions of CSR performance.

1_{Australia, Austria, Belgium, Canada, Denmark, Finland, France, Germany, Ireland, Italy, Japan, Netherlands,}

(14)

Concerning differences among CSR performance dimensions. CSR governance performance is the only significantly related CSR performance dimension to market freedom. Furthermore, the effect size of the availability of skilled labor variable for CSR environmental performance is roughly half compared to the other dimensions. Only regarding the strength of trade unions effect sizes and significance are similar across CSR performance dimensions.

In previously conducted research, the usage of an aggregate CSR variable was wide spread. Accordingly based on the findings regarding differences in CSR performance dimensions, the validity of previously conducted research can be questioned.

1.6 Structure of the Dissertation

To answer the research question, the dissertation begins with assessing the pertinent literature. Elaborating on the role of the individual streams of literature, the comparative capitalism literature in the form of the VoC framework serves as a foundation. The VoC framework brings together national institutions and firm level outcomes by elaborating on how individual firms are affected by institutions. This is where the CSR literature comes into play, it extends the VoC literature which has been focused on other firm level outcomes, and builds a

theoretical base to assess the relationship between national institution and the firm level outcome CSR performance. Naturally also the direct field which this dissertation adds to, comparative CSR, is addressed. Based on these three streams of literature, hypotheses were developed.

Next, in an effort to build a foundation for the testing of these hypotheses, the methodology this research project follows is presented. Here research philosophy, research approach, sample, data collection and data analysis are discussed.

Following this, results are presented and discussed. Finally, implications, limitations and suggestions for future research based on these results are proposed.

1.7 Introduction Chapter Summary

Based on the development in world business, CSR is a topic of rising importance.

(15)

only vaguely characterized by referring to voluntary business practices which in addition to economic gain also incorporate considerations regarding social, environmental and

governance impacts (Fransen, 2012).

Considering the voluntary nature of CSR practices and the resulting CSR performance, the research objective at the hearth of this project is to assess the extent to which firm-level variation in CSR performance is explained by country-level differences. In previous research, the institutional environment has been identified as a relevant driver of variation in firm-level CSR performance across countries (see e.g. Peng, Dashdeleg and Chih, 2014;

(16)

Chapter 2: Theoretical Framework and Hypotheses

Concerning the research question this research poses: "What is the relationship between national institutions (according to the VoC framework) and individual CSR performance dimensions?", three streams of literature were consulted. The question posed touches upon comparative capitalism in the form of the VoC framework, the CSR literature and the comparative CSR literature. Accordingly this chapter proceeds by introducing the

comparative capitalism literature, as it features the VoC framework adopted for this research project. The CSR literature is also scrutinized. Lastly the research field of comparative CSR which brings together comparative capitalism and CSR is assessed. Having consulted the different literature fields, finally a set of hypotheses is introduced.

2.1 Literature Review

This research is at the cross-section of three fields of literature, accordingly comparative capitalism, CSR and the interface comparative CSR were reviewed.

2.1.1 Comparative Capitalism

Different frameworks have been developed to analyze and explain how different national institutions influence firms (Quintanilla, 2000). One framework which has gained significant attention and is useful in the context of this work, is the VoC framework by Hall and Soskice (2001).

The VoC perspective has been specifically developed to understand institutional differences and similarities as well as their implications among economies (Hall and Soskice, 2001), for assumptions the VoC makes in regard to the institutional environment please see Appendix 1. The framework differentiates itself from other approaches by bringing firms to the center of its analysis (Gjolberg, 2009b).

(17)

The relationship a firm engages in are divided into five distinctive spheres including

governance, inter-firm relations, capital–labor relations, education and training. According to the logic of the VoC framework, these relationships are essential for firm survival as they enable access to markets, suppliers and peers. In underscoring the importance of the firm to coordinate across five different spheres and the influence institutions have on these

coordination processes, the framework moves the firm to the center of the analysis. This creates a theoretically grounded point of departure for assessing the relationship between national institutions and firm level outcomes.

In different national institutional contexts, however this ability of firms to coordinate

effectively across the range of different stakeholders varies (Hall and Soskice, 2001). Hence particular social, economic and political institutions provide for certain firm capabilities. Put into the words of Hall and Soskice (2001, p. 37):

"The basic idea is that the institutional structure of a particular political economy provides firms with advantages for engaging in specific types of activities there. Firms can perform some types of activities, which allow them to produce some kinds of goods more efficiently than others because of the institutional support they receive for those activities in the political economy, and the institutions relevant to these activities are not distributed evenly across nations."

Due to its applicability to researching national institutional influences on firm level outcome variables, which also applies to the case of CSR performance, many studies in the CSR field of research use this framework (see, for instance, Middtun et al., 2006; Jackson and

Apostolakou, 2010; Hartmann and Uhlenbruck, 2015). Furthermore the importance of

coordination and building relationships with various stakeholders corresponds very well to the CSR literature, where stakeholder coordination capabilities are also regarded as a key

component of success (Barnett, 2007).

Referring to Table 2.1, the VoC framework culminates in dividing national economies into LMEs and CMEs (Hall and Soskice, 2001). CME and LME refer to the primary way in which firms coordinate with each other and stakeholders in the respective economy (Hall and

(18)

firms build with stakeholders, based on underlying institutions the framework postulates that firms in LMEs primarily coordinate their relationships through market mechanisms; while firms in CMEs rely more heavily on non-market forms of coordination (Hall and Soskice, 2001).

Accordingly the framework divides economies into these two categories by assessing typical institutions and institutional configurations for the two types of economies. Considering previous literature, market institutions, social institutions and labor institutions have been identified as institutions relevant to CSR performance (Middtun et al., 2006; Ioannou and Serafeim, 2012; Hartmann and Uhlenbruck, 2015).

Elaborating on the differences between LMEs and CMEs, in LMEs institutions such as supply and demand guide the behavior of market actors (Hall and Soskice, 2001). In addition LMEs display characteristic features like a less interventionist state, individualized and adversarial capital–labor relations and more fluid labor markets (Cooke, 2007; Fransen, 2012). In CMEs, markets are also significant; however more so than in LMEs market mechanisms are

supplemented by non-market relationships which take the form of co-operation, networks, collaborative actions and government interference (Hall and Soskice, 2001; Middtun et al., 2006). Furthermore, characteristic features of CME approximating countries include

institutionalized dialogue between social partners, stringent rules and smaller inflexible labor markets (Hall and Soskice, 2001; Cooke, 2007).

Table 2.1: Overview CME and LME

Institutions Definition CME Implications LME Implications

Market Institutions: Constraints structuring economic interaction. More interventionist state (Middtun, Gautesen and Gjolberg, 2006); non-market modes of coordination (Hall and Soskice, 2001; Cooke, 2007) Less interventionist state (Middtun, Gautesen and Gjolberg, 2006); arm's length market

exchanges (Hall and Soskice, 2001; Cooke, 2007)

Social Institutions Constraints structuring social interactions.

Low share of tertiary graduates

(19)

(Schneider and Paunescu, 2012)

(Schneider and Paunescu, 2012) Labor Institutions Constraints structuring

the interaction between labor force and employer.

Rigid labor markets (Hall and Gingerich, 2004);

institutionalized dialogue between social partners (strong trade unions) (Hall and Soskice, 2001;

Schneider and Paunescu, 2012)

Fluid labor markets (Hall and Gingerich, 2004); individualized and adversarial capital–labor relations (Hall and Soskice, 2001; Schneider and Paunescu, 2012)

2.1.2 Corporate Social Responsibility

This research compares CSR performance across companies in different countries, accordingly to conduct an objective comparison it is important to have an underlying definition. Therefore, the review of the CSR literature focuses on defining CSR and highlighting the close relationship between CSR and CSR performance.

2.1.2.1 Defining Corporate Social Responsibility

Owing to the popularity of CSR as a topic with business, academia and the public it would seem that the term is well defined. However even as there seems to be a vague shared understanding of CSR, a unified concept of CSR is not firmly established (Dahlsrud, 2008). An item of significant importance, which has been identified throughout CSR related articles, are stakeholder expectations (Aguilera et al., 2007). These expectations constantly evolve and in the field of CSR have led from a focus on philanthropy to an increased focus on how profits are earned (Hack et al., 2014). This shift in expectations has major implications, as it directly puts the spot light on a company's core business activities; highlighting social, environmental and governance issues. Therefore in a way CSR is about internalizing externalities back into the corporation (Hack et al., 2014).

(20)

voluntary vs regulatory approaches and the roles of the state market and civil society (Gjolberg, 2009b).

The ongoing debates about CSR resulted in multiple differing high profile definitions (e.g. EU (EC Europa, 2015), UN (UNido, n.d.)) which respectively emphasize different aspects of CSR, however commonly CSR refers to the social, environmental and governance impacts of firm actions (Fransen, 2012). On the one hand this absence of a fixed definition, means that the concept can be used by larger groups of users and in a wider set of contexts; however regarding a comparative research project a fixed definition is paramount to making objective comparisons.

2.1.2.2 Corporate Social Responsibility Performance

The concepts of CSR and CSR performance are inherently linked. Owing to the ambiguity of CSR, there is a multitude of practices businesses utilize to address CSR related issues. Based on the heterogeneity of these CSR approaches, plenty of academics and professionals have made it their duty to assess the merits of the CSR initiatives of the respective companies (see e.g. Gjolberg, 2009a; Sustainability-indices, 2015). In the pertinent literature, CSR

performance is commonly assessed as the implementation of CSR conform practices

regarding the three CSR specific dimensions social, environmental and governance (Fransen, 2012; Hartmann and Uhlenbruck, 2015). Accordingly multiple indices using underlying individual indicators have been developed in order to assess performance of companies regarding CSR in general as well as regarding the respective dimensions (social,

environmental, governance). These indicators usually score, rank or give percentile scores to companies based on their performance regarding the indicators, accordingly CSR

performance of a company refers to its CSR related practices which are assessed against goals and thresholds set by academics and professionals.2

2.1.3 Comparative Corporate Social Responsibility

The research question of this project is at the intersection of the CSR and comparative

capitalism literature, accordingly the term comparative CSR describes this research field well. After having elaborated on the individual streams of literature, it is important to illuminate

(21)

recent work which has contributed directly to the understanding of CSR performance and cross-national institutional differences.

Concerning the general awareness, it is long established that companies are embedded in an institutional environment which significantly shapes the decision making of the respective companies (e.g. Campbell, Hollingsworth and Lindberg, 1991). Later on, the development of different frameworks to classify different national business systems has greatly advanced research. Through these frameworks it was demonstrated that different capitalist systems create divergent institutional environments; these different environments accordingly offer different comparative advantages for different processes and products (Hall and Soskice, 2001).

(22)

Table 2.2: Literature Review Comparative CSR

Author(s) Research Objective Variables Theory, Data & Method Findings

Waldman et al. (2006)

Explore how cultural and leadership variables are

associated with corporate social responsibility variables in decision making

IV: Leadership vision, leadership integrity, in-group collectivism, institutional collectivism, power distance DV: CSR values

T: Cultural differences D: 561 firms, 15 countries, cross-sectional

M: Hierarchical regression

Significant positive effect of leadership vision, leadership integrity, institutional collectivism; significant negative effect of power distance; No effect of in-group collectivism

Ringov and Zollo (2007)

Investigate the effect of differences in national cultures on CSR social performance and CSR environmental performance

IV: Masculinity,

individualism, power distance, uncertainty avoidance

DV: Social and environmental performance

T: Cultural differences D: 463 firms, 23 countries, cross-sectional

M: OLS regression

Significant negative effect of masculinity, power distance; no effect of individualism, uncertainty avoidance

Gjolberg (2009b)

Assess the importance of global forces and national political economic institutions for CSR engagement.

IV: Social models,

environmental sustainability index, corporatist integration, World Values Survey, foreign direct investments, proportion of TNCs

DV: CSR

T: Institutional and political economy approaches D: N/A

M: Fuzzy Set Analysis

Alternate pathways to strong CSR performance: Globalist and CME

Haxhi and van Ees (2010)

Analyze the cultural

determinants of the diffusion of codes of good governance

IV: Masculinity,

DV: Number of codes of good governance

T: Cultural differences D: 237 codes, 67 countries, cross-sectional

M: Poisson regression

Significant positive effect of individualism; no effect of masculinity, power distance, uncertainty avoidance Jackson and

Apostolakou (2010)

Examine how institutional factors influence CSR practices

IV: LMEs, CMEs, investor protection laws, employment protection laws

DV: CSR

T: Institutional theory D: 274 firms, 16 Western European countries, cross-sectional

M: OLS regression

Significant effect of market economy type, investor protection and

employment protection laws

Ho et al. (2012)

Examine how cultural values affect CSR performance

IV: Masculinity,

DV: CSR

T: Cultural differences D: 3680 observations, 49 countries, cross-sectional

(23)

M: 2-stage least squares regression

Ioannou and Serafeim (2012)

Explore the extent to which variation in CSR performance can be explained by national institutions

IV: Promotion of competition, , leftist ideology, shareholder protection, union labor density, corruption,

availability of skilled labor, socially responsible stock market, market-based financial system, individualism, power distance

DV: CSR

T: National business systems D: 2330 firms, 42 countries, longitudinal

M: OLS regression

Significant positive effect of union labor density, availability of skilled labor, individualism, power distance; significant negative effect of

promotion of market competition, leftist ideology, shareholder

protection, corruption, market-based financial system; no effect of socially responsible stock market Lim and

Tsutsui (2012)

Understand how the institutional environment drives commitment to socially responsible initiatives

IV: Number of NGOs, number of international treaties, past UN Global compact

participants, signature of UN Global Compact, level of education, democracy, public sector transparency, UN Global launch, bilateral investment, bilateral export DV: Baseline, ceremonial and substantive commitment to CSR

T: Institutional and political economy approaches

D: 99 countries, longitudinal M: Binomial regression

Significant positive effect of number of NGOs; significant negative effect of number of international treaties; no effect of past UN Global compact participants, level of education, democracy, public sector

transparency, UN Global Compact launch, bilateral investment

Toffel, Short and Quellet (2012)

Understand the impact of the institutional environment on adherence of private codes of conduct

IV: Number of NGOs,

protection of labor, number of labor treaties issue salience, press freedom DV: Adherence to codes of conduct T: Institutional theory D: 14,922/689 firms, 43/33 countries, cross-sectional M: Binomial regression

Significant positive effect of labor treaties, issue salience, press freedom; no effect of number of NGOs, protection of labor

Peng et al., (2014)

Explore the relationship between national culture and CSR

IV: Masculinity,

DV: CSR

T: Cultural differences D: 1,189; 29 countries, cross-sectional

M: Binary logistic regression

(24)

Thanetsunthorn (2014)

Examine the impact of national culture and geographic

environment on firms' CSR performance.

IV: Masculinity,

individualism, power distance, uncertainty avoidance DV: CSR T: Cultural differences D: 2,129, 14 countries, cross-sectional M: OLS Regression

Significant positive effect of uncertainty avoidance; significant negative effect of individualism Hartmann and

Uhlenbruck (2015)

Investigate the extent to which firm-level variation in CSR environmental performance is explained by country-level differences

IV: Environmental Treaties, market freedom, non-governmental organizations, press freedom DV: Corporate environmental performance T: Institutional theory D: 2724, 42 countries, longitudinal M: Multilevel modelling

Significant positive effect of market freedom, non-governmental

organizations, press freedom; significant negative effect of environmental treaties

(25)

Previously mentioned in Section "2.1.2.1 Defining CSR", often CSR is based on the three different dimensions social, environmental and governance. Nevertheless assessing the dependent variables used by studies featured in Table 2.2, the most popular measure is an aggregated conception of CSR performance (see e.g. Waldman et al., 2006;Gjolberg, 2009b; Jackson and Apostolakou, 2010; Ho et al., 2012; Ioannou and Serafeim, 2012; Lim and Tsutsui, 2012; Peng et al., 2014; Thanetsunthorn, 2014). This is in stark contrast to evidence pointing towards important differences between aggregated CSR performance, CSR social performance, CSR governance performance and CSR environmental performance (Ringov and Zollo, 2007; Hartmann and Uhlenbruck, 2015).

Due to the use of an aggregate CSR performance variable, there are inherent shortcomings; the literature on comparative CSR lacks specificity regarding the dimensions of CSR

performance (social, environmental, governance). This lack of specifity, threatens the validity of previous research findings, as accordingly it is unclear which national institutions affect which CSR dimensions (social, environmental, governance). Furthermore, it also remains unclear if the relationship with national institutions varies among the different CSR dimensions (social, environmental, governance).

Considering the shortcomings of an aggregated CSR variable, a major concern of this research project is to advance research in this regard. Accordingly for the purpose of this research project the concept of CSR is in addition to assessment of an aggregated variable split into the three different CSR dimensions social, environmental and governance.

Concerning the dimensions environmental, social and governance, admittedly each dimension still is an aggregate measure. However using such a lower level of aggregation, this research project is able to detect meaningful variance between the different dimensions.

(26)

Uhlenbruck, 2015). Similarly employees themselves are likely to influence firm behavior, for example through organization in trade unions (Ioannou and Serafeim, 2012).

Still relating to Table 2.2, the findings of the individual studies are to a large extent contradictory. This is concerning the popular independent variable culture, where the

relationship of individual dimensions is heavily disputed, as well as in regard to more formal institutions. Concerning formal institutions the debate essentially centers on whether

institutions related to a more coordinated or a more liberal market economy are conducive to CSR performance.

In summary, considering the main characteristics of the literature field, there is ongoing debate about which institutions are related to strong CSR performance. Further contributing to the debate, is whether these institutions have the same effects across the whole range of CSR dimensions (social, environmental, governance). Accordingly this research project contributes empirical evidence for the relationship between institutions and CSR performance by

conducting a more fine grained approach in terms of addressing the respective CSR dimensions (social, governance, environmental).

2.2 Hypotheses Development

The VoC framework enables a direct link between national institutions and CSR performance. Building on this link and assessing the shortcomings of the comparative CSR literature, hypotheses were developed. The shortcomings in the comparative CSR literature which the hypotheses relate to include the pervasive use of an aggregate CSR performance variable, furthermore the mixed empirical evidence regarding the institutional environment is addressed.

The hypotheses outlined in the following paragraphs all address a unified concept of CSR and do not address the individual CSR dimensions. This is done as previous literature has focused on using an aggregate measure of CSR (see e.g. Ioannou and Serafeim, 2012; Lim and

(27)

among the respective CSR dimensions are addressed in Section "4.3.4 Differences among Corporate Social Responsibility Dimensions".

2.2.1 Market Institutions

Market institutions concern the relationships a company forms with other enterprises, suppliers as well as its clients. Building and maintaining these relationships is of key importance to business success (Hall and Soskice, 2001).

Traditionally, the economy has been an arena of human interaction free of moral sentiment (Shamir, 2008). Accordingly markets have enjoyed and still enjoy "unprecedented ethical freedom from taking into account socio-moral considerations relating to the generalized harmful consequences of economic action, such as labor exploitation, environmental degradation, social disintegration or cultural decomposition." (Shamir, 2008, p. 375).

The exemption of business from moral obligations, is based on the notion that other social mechanisms, in particular governments, according to the logic of welfare and security were seen as responsible for managing populations (Szakolczai, Burchell, Gordon, Miller, Eribon, Wing and O'Farrell, 1993; Shamir, 2008).

Fulfilling this role, national governments, legitimized by the public and therefore acting on behalf of said public voiced collective claims of society. Pushing these societal claims, governments use means of market regulation and provision of social services as their instruments (Shamir, 2008). These instruments applied by governments are intended to correct distortions caused by ethically free markets and eliminate externalities generated by unregulated and underregulated business practices (Shamir, 2008).

(28)

In the contemporary sense of CSR higher conduct standards implemented by companies due to mandatory regulation from a national point of view would not qualify as CSR.

Nevertheless, when considering a global arena, such standards are not mandatory to firms and accordingly firms meeting such standards perform strong regarding their CSR from a global perspective (Gjolberg, 2009b).

Therefore:

Hypothesis 1a: Firms located in countries characterized by a higher degree of market freedom will have lower levels of CSR performance.

Hypothesis 1b: Firms located in countries characterized by a higher degree of market freedom will have lower levels of CSR social performance.

Hypothesis 1c: Firms located in countries characterized by a higher degree of market freedom will have lower levels of CSR environmental performance.

Hypothesis 1d: Firms located in countries characterized by a higher degree of market freedom will have lower levels of CSR governance performance.

2.2.2 Social Institutions

Social institutions concern the relationships firms build with society, workers and the labor market in order to attract and retain a qualified workforce.

Social institutions have been found to greatly influence an individual's education and career trajectory. This is as an individual's career choices within a society are, greatly dependent on the grand narrative of a career proposed by the respective society (Brown and Lent, 2013). Accordingly social structure, prescribes how individuals work lives should proceed. Through conditioning such essential drivers like the supply of a skilled workforce, which firms rely on, social institutions significantly shape the relationships firms build with individuals and the labor market.

(29)

personnel and also has implications for wage setting (Borjas, 2013). Accordingly the availability of skilled labor, as influenced by social institutions conditions the relationship between companies and potential candidates. In a labor market where labor is scarce,

employers have to increase their organizations attractiveness, by putting in greater effort and offering greater packages than in a markets with abundant skilled labor (Slack, Corlett and Morris, 2015). Accordingly the size of the labor market lays out informal rules for both the employer and employee.

Research has demonstrated that, companies can use CSR performance as a method to reach and attract potential candidates in terms of recruitment, employee morale, productivity and retention (Slack, et al., 2015). Assessing this relationship between CSR and employees, in previous research companies with strong CSR performance have been shown to be able to attract and retain higher-quality employees (Ioannou and Serafeim, 2012). This is as CSR performance can be used in order to give prospective candidates an insight of what it would be like to work for the focal firm (Ioannou and Serafeim, 2012). Recent insights have

concluded that a company with strong CSR practices sends out signals to employees and job seekers which ultimately affect organizational attractiveness and the willingness of suitable candidates to join the firm (Jones, Willness and Madey, 2014).

Key concepts concerning this signaling mechanism, which strong CSR performance of a company affects are "anticipated pride from being affiliated with the organization, their perceived value fit with the organization, and their expectations about how the organization treats its employees." (Jones, et al., 2014, p. 383).

Accordingly actively improving a company's CSR performance is a strategic action

companies take in order to recruit skilled human capital. Especially since for more and more talented employees CSR considerations matter and they prefer to work for a company with strong CSR performance (Ioannou and Serafeim, 2012). Further strengthening this argument is research which concluded that in industries where skilled labor is scarce, firms display greater proactivity in their CSR efforts (Siegel, 1999).

(30)

Therefore:

Hypothesis 2a: Firms located in countries characterized by higher availability of trained and skilled human capital will have lower levels of CSR performance.

Hypothesis 2b: Firms located in countries characterized by higher availability of trained and skilled human capital will have lower levels of CSR social performance.

Hypothesis 2c: Firms located in countries characterized by higher availability of trained and skilled human capital will have lower levels of CSR environmental performance.

Hypothesis 2d: Firms located in countries characterized by higher availability of trained and skilled human capital will have lower levels of CSR governance performance.

2.2.3 Labor Institutions

Based on the importance of labor to firm performance in general, naturally also the

institutions which control the terms to which labor is available are highly relevant (Whitley, 2005). Trade unions have been identified as having significant influence over "changing strategic priorities, technologies and markets as well as the decisions relating to labor management strategies" (Ioannou and Serafeim, 2012, p. 840)

Accordingly also for CSR related practices, trade unions have been identified as a key source of pressure for organizations to engage in CSR (Miller and Guthrie, 2007; Doshi and Khokle, 2012). Especially since work aspects related to site safety, health and other workers benefits are directly related to CSR. In addition employees typically also make up the "local"

stakeholder population, as often employees originate from local communities or at least reside within local communities (Ioannou and Serafeim, 2012). Accordingly trade unions in addition to work place improvements may also push for better community involvement environmental and social policies (Ioannou and Serafeim, 2012). Following this further, the strength of a trade union plays a significant role, as a strong trade union is capable of reducing the scope of influence of mangers and is able to push managers to seek legitimacy with their employees via CSR related activities (Miller and Guthrie, 2007).

(31)

progressive action in order to avoid the "threat" of actual unionization of their workforce (McWilliams and Siegel, 2001; Miller and Guthrie, 2007).

Therefore:

Hypothesis 3a: Firms located in countries characterized by higher levels of trade union power will have higher levels of CSR performance.

Hypothesis 3b: Firms located in countries characterized by higher levels of trade union power will have higher levels of CSR social performance.

Hypothesis 3c: Firms located in countries characterized by higher levels of trade union power will have higher levels of CSR environmental performance.

Hypothesis 3d: Firms located in countries characterized by higher levels of trade union power will have higher levels of CSR governance performance.

2.3 Theoretical Framework and Hypotheses Chapter Summary

First by assessing the comparative capitalism literature, the VoC framework was presented. This framework is well suited for this research project, as through it national institutions can be directly linked to firm level CSR performance. Second through assessing the literature on CSR it was highlighted that there are significant shortcomings in reaching a common

definition of CSR; however in contemporary use, CSR most commonly refers to the underlying dimensions of social, environmental and governance (Fransen, 2012).

Furthermore, the close relationship between CSR and CSR performance was elaborated on. Moving on from the CSR literature and the comparative capitalism literature, the direct research field which this project addresses, comparative CSR was introduced. Associated to shortcomings in the definition of CSR, comparative CSR research commonly uses an aggregate CSR performance variable instead of the social, environmental and governance dimensions which underlie CSR (see e.g. Ho et al., 2012; Ioannou and Serafeim, 2012). However due to the use of such an aggregated variable, the literature is not able to make inferences about differences regarding the underlying dimensions, while pioneering research gives reason to believe that there are substantial differences (see e.g. Hartmann and

Uhlenbruck, 2015). Furthermore, in regard to the relationship between CSR and the institutional environment there is conflicting empirical evidence.

(32)

hypotheses was developed. Owing to the previously identified shortcomings of the use of aggregated variables in CSR research, each of these sets of hypotheses addresses the social, environmental, governance and aggregate dimension of CSR performance. Accordingly the premise of this framework is that the CSR performance of a firm can be predicted by national level differences in market, social and labor institutions. For a graphic illustration of the conceptual model please see Figure 1.1 below. In addition to depicting the hypothesized relations, Figure 1.1 also displays the control variables considered throughout the research project.

(33)

Chapter 3: Methodology

This research project is based on underlying assumptions which are related to the methodology. To facilitate interpretation of results these assumptions are subsequently outlined.

This chapter starts by elaborating on the positivist research philosophy which guides this research project. Closely linked, the deductive research approach this project follows is also outlined. In addition, considerations regarding the sample and data sources are highlighted. Finally the method of analysis, regression analysis, is addressed.

3.1 Positivist Research Philosophy

Saunders, Lewis and Thornhill (2012) identify four different epistemological orientations; positivism, interpretivisim, realism and post-structuralism. The two most commonly used research philosophies are divided along the lines of qualitative and quantitative research. On the one hand qualitative research focuses on humans as social actors and mostly draws on the research philosophy of interpretivism (Saunders et al., 2012). Quantitative research on the other hand focuses on observing phenomena in a value-free way and mostly draws on the research philosophy of positivism (Saunders et al., 2012). Value free observations do not refer to worthless observations, but instead refer to the ideal situation that the researcher is external to the process of data collection. As a result the outcome of such value free observations is objective and can therefore be measured and predicted (Saunders et al., 2012).

The positivist research philosophy postulates that the world consists of regularities. These regularities are furthermore detectable and accordingly by observing these regularities a researcher can infer knowledge about the real world (Jonker and Pennink, 2014).

Keeping the research question in mind, "What is the relationship between national institutions (according to the VoC framework) and individual CSR performance dimensions?" this project seeks to make general inferences about relationships based on value free observations.

(34)

3.2 Deductive Research Approach

The research approach is a plan of action taken to answer the research question in the best possible way (Jonker and Pennink, 2014). Concerning research in the business and

management literature, typically two main research approaches can be found. One approach is known as the inductive approach, this approach starts with data and seeks to develop theory based on patterns found in the data (Ketokivi and Mantere, 2010). In contrast, the deductive approach starts with existing theories and concepts; from this existing material, hypotheses are derived (Saunders et al., 2012). Subsequently concepts and constructs featured in these hypotheses have to be operationalized, in order to enable the measurement of facts (Saunders et al., 2012).

This research project starts by addressing contemporary theories and concepts, through a literature review and subsequently develops hypotheses. The project then continues to

operationalize key concepts and draws on secondary data, accordingly a deductive approach is followed.

3.3 Sample

In accordance with the research field the sample for this research project is based on companies for which CSR data is available (see e.g. Thanetsunthorn, 2014, Hartmann and Uhlenbruck, 2015). Accessing the data provided by a collaboration between Bloomberg and Sustainalytics the initial sample for this research consists of 1656 companies from 31 countries3.

Accommodating to propositions brought forward in the VoC framework and in subsequent work, this initial sample was further trimmed down. First the VoC framework was developed with affluent democratic nations in mind, accordingly this limits the nation sample for this study to OECD nations4 (Hall and Soskice, 2001; Schneider and Paunescu, 2012).

3_{Australia, Austria, Belgium, Bermuda, Canada, Chile, China, Colombia, Denmark, Finland, France, Germany,}

Hong Kong, Ireland, Israel, Italy, Japan, Jersey, Luxembourg, Macau, Mexico, Netherlands, New Zealand, Norway, Portugal, Singapore, Spain, Sweden, Switzerland, United Kingdom,

United States

4_{OECD nations: Australia, Austria, Belgium, Canada, Chile, Czech Republic, Denmark, Estonia, Finland,}

(35)

Second regarding the companies suitable for assessment through the VoC framework, Hall and Soskice (2001) focused on production regimes. Nevertheless also non-industrial firms can be assessed through the framework (Hiscox and Rickard, 2002; Farndale et al., 2008; Faulconbridge, 2008). However the framework mostly ignores banks and other financial institutions, therefore in accordance with previous research in comparative capitalism the sample used for this research project excludes these (Kuznetsov and Jacob, 2015). These trimmings due to VoC framework specific requirements brought down the sample size from 1656 to 1244 companies.

Supplementing the sample firms with information concerning national institutions some assumptions had to be made. Many of the companies featured in the sample are large corporations with networks of subsidiaries which span the globe. Accordingly the question arises whether CSR performance is driven by centrally implemented practices and strategies or if subsidiaries face significant autonomy and accordingly shape CSR related outcomes. If the former is the case national institutional variables of a company's home country can be attributed. If the latter is the case, CSR related outcomes would be based on a network of different institutional influences from multiple countries. Addressing the potential issue of multiple underlying institutional environments, this research project draws on findings by Jamali (2010), who proposes that corporations adhere to a unified set of global CSR policies. These global CSR policies are designed by the respective firm's headquarter and are then subsequently diffused to subsidiaries. Although not explicitly addressed, authors in the research field of comparative CSR adhere to this view. Throughout all studies reviewed authors prescribe national institutions according to the headquarter countries of companies (see e.g. Ho et al., 2012; Thanetsunthorn, 2014, Hartmann and Uhlenbruck, 2015). Therefore, in line with previous work in this research field, this research project prescribes national institutions based on the home country of a company.

(36)

Table 3.1: Distribution of Sample Firms across Countries

Country Number of Companies Percentage %

Australia 48 3.90 % Austria 4 0.32 % Belgium 8 0.65 % Canada 72 5.85 % Denmark 11 0.89 % Finland 11 0.89 % France 58 4.71 % Germany 45 3.66 % Ireland 12 0.97 % Italy 12 0.97 % Japan 273 22.18 % Netherlands 20 1.62 % New Zealand 7 0.57 % Norway 6 0.49 % Portugal 4 0.32 % Spain 16 1.30 % Sweden 23 1.87 % Switzerland 33 2.68 % United Kingdom 85 6.90 % United States 483 39.24 % Total 1231 100 % 3.4 Data Collection

In regards to data collection, there are two main categories; primary data and secondary data. Primary data is collected by the researcher and is tailored to the research project and question at hand. Contrary, secondary data has been collected by others with possibly different

intended uses than for the research project at hand. Arguably it is desirable to have primary data to meet ones exact research needs (Saunders et al., 2012); however considering the scope across different institutions and worldwide scale of this research project it was not feasible to collect primary data. Therefore in accordance with the literature featured in Section "2.1.3 Comparative CSR", this projects draws on secondary data (for an overview see Section "3.7 Methodology Chapter Summary").

3.5 Operationalization and Data Sources

(37)

3.5.1 Corporate Social Responsibility Performance

In order to address the relationship between national institutions and firm-level CSR performance, it was important to first of all measure CSR performance. This is especially relevant as this study has a comparative nature and its reliability and validity are based on accurately gauging CSR performance.

3.5.1.1 Corporate Social Responsibility Performance Operationalization

Recalling Section "2.1.2 Corporate Social Responsibility", a general definition of CSR which underlies CSR performance is elusive. Accordingly, many researchers take a pragmatic approach in operationalizing CSR performance. It is common practice in the comparative CSR literature to operationalize CSR performance as an aggregate construct developed by businesses, analysts and stakeholders. Accordingly this group determines what falls inside and outside of the scope of CSR, regarding issues and practices (Fransen, 2012). Following this approach is widely popular and an accepted possibility among the research community to circumvent difficulty in defining CSR.

A further option to circumvent problems in defining CSR can also be found in the literature. In order to define CSR, Gjolberg (2009b) used formative indicators. Applied to Gjolberg's research, CSR practices and performance were determined by using sustainability indexes, business top lists, membership in business associations and management certification schemes as indicators.

Both methodologies which can be found in the literature offer significant shortcomings. For the formative index, difficulty in determining the validity of the index poses an issue; for the aggregated CSR performance variable the inability of researchers to make inferences with respect to differences between CSR performance dimensions (social, governance,

(38)

3.5.1.2 Measurement

The dataset for the dependent variables (CSR performance dimensions) of this research was constructed by Bloomberg in cooperation with Sustainalytics and concerns the year 2015. Its primary focus is on business users, nevertheless this dataset has also gained recognition in academic research (see e.g. Fernandez-Feijoo, Romero and Ruiz, 2013; Wolf, 2014).

3.5.1.2.1 Bloomberg and Corporate Social Responsibility

Bloomberg has decided to include CSR relevant information as part of their mainstream offering to analysts, as more and more investors consider CSR related aspects as part of their investment decision (Scalet and Kelly, 2010). This development is supported by a whole range of academic literature which implies that there is a connection between doing good for society and doing well in business (e.g. Griffin and Mahon, 1997; Camejo and Aiyer, 2002; Adam and Shavit, 2008; Poddi and Vergalli, 2009; Wang, Chen, Yu and Hsiao, 2015; Lins, Servaes and Tamayo, 2015). The concept of benefitting from doing good is based on the idea that even while initial CSR investment might have a negligible or even negative effect, long run performance will improve due to positive reputation effects, a reduction of long run costs and increased social responsible demand (Poddi and Vergalli, 2009).

Consequently prominent, well recognizable performance related CSR initiatives and rankings (like the scores given by Bloomberg and Sustainalytics) signal to investors that included companies have successfully passed a CSR related audit and accordingly comply with

internationally recognized CSR standards. This positively influences perceptions of corporate reputation (Cho, Guidry, Hageman and Patten, 2012), through which companies can attract potential investors (Ioannou and Serafeim, 2012; Raithel and Schwaiger, 2015).

Accordingly it is in the interest of firms to be part of such initiatives and rankings (Consolandi Jaiswal-Dale, Poggiani and Vercelli, 2009; Chatterji and Toffel, 2010). This greatly

strengthens the case for using indicator supplied data, as it is not only Bloomberg giving out scores, but furthermore companies demonstrating high CSR standards actively seek a rating to reap the benefits of their investment.

3.5.1.2.2 Bloomberg and Sustainalytics Methodology

(39)

social, governance and environmental scores relative to industry peers. For the top 1% of CSR performers the percentile is 99%, while for the bottom 1% of CSR performers the percentile is 1%.

Percentile ranks of the different CSR dimensions rely on four underlying pillars, for a graphic illustration of this methodology please see Figure 3.1 below. These pillars include the

preparedness of a firm. Preparedness is evaluated by assessing management systems and policies which are in place at a company in order to manage CSR related practices and issues (Sustainalytics, 2015). Second the disclosure of CSR relevant information is included into the score. This is done via assessing whether reporting by the company in question meets

international best practice standards and is transparent with respect to CSR issues

(Sustainalytics, 2015). Third quantitative performance regarding underlying CSR relevant indicators is assessed (Sustainalytics, 2015) (for an illustrative list of indicators see Appendix 2). The final pillar of CSR assessment by Bloomberg and Sustainalytics assesses performance in a qualitative matter, this is done via assessing individual incidents and or business practices the respective company may be involved in (Sustainalytics, 2015).

Specifying these assessments for the respective CSR dimensions, concerning the social dimension, the rank is built on the relative performance of companies regarding indicators which concern employees, suppliers, customers and society. Regarding the governance dimension, percentile ranks are determined by assessing underlying individual indicators, such as shareholder rights and compensation structures. Assessing the environmental dimension, the scores build on the ability of companies to meet specific environmental indicators measured relative to the performance of peers; relevant indicators include emissions, waste generation and membership in progressive initiatives.

Before any percentile scores are made available through Bloomberg, each company's CSR assessment undergoes a thorough internal peer review process (Sustainalytics, 2015). Following this some companies may be contacted in order to clarify outstanding issues (Sustainalytics, 2015).

(40)

Figure 3.1: Bloomberg and Sustainalytics Methodology (Adapted from Sustainalytics, 2015)

3.5.2 Institutional Variables

Regarding the relationships proposed in Section "2.2 Hypotheses Development", in order to accurately assess the hypotheses it was important to operationalize the variables which had been proposed.

Institutions have been describes as being very stable over time, they "do not respond in any rapid and fluid way to alterations in the domestic or international environment" (Krasner, 2009, p. 78). This is as once institutions have been established, there is a tendency of

reproduction in a stationary way (Florensa, 2004), as in a given environment the institutions and the patterns of behavior fostered create an equilibrium (Greif, 2003).

(41)

Lopez-de-Silanes, Shleifer and Vishny is also a popular measure featured in the literature review (see Haxhi and van Ees, 2010). Similarly work by Ferrera (1998) on social models around the world is also still considered to be a valid measure to assess the dominant social policy across countries (see Gjolberg, 2009b). Therefore although due the availability of data there is a slight discrepancy in time of measurement between the dependent variable CSR performance (2015) and the institutional variables (2012) utilized for this research project, this should not be of any concern. Especially since in comparative research such indicators which change only incrementally or not at all often do not stem from the same year (e.g. see Allen and Aldred, 2011).5

3.5.2.1 Market Freedom

Assessing market freedom is widely utilized in institutional research. A database which has been relied upon in research is the Index of Economic Freedom by the Heritage Foundation (see e.g. Herrera-Echeverri, Haar and Estévez-Bretón, 2014). This database has been assessed in dedicated research, yielding that it is precise and transparent in terms of the information it contains (Gwartney and Lawson, 2003). Data utilized is as of 2012, however again, this variable is commonly used by researchers even if there is a slight time difference between variables (see e.g. Evrensel, 2014).

3.5.2.2 Availability of skilled Labor

In terms of assessing the availability of skilled labor, this research project draws on

graduation rates in tertiary education (OECD, 2014). This special focus on tertiary education is justified by considering that the countries in the sample of this research project are all highly industrialized economies, being classified as knowledge economies (Worldbank, 2012).

Accordingly the rates reflecting the percentage of population having attained any kind of tertiary education degree reflects the relative availability of skilled labor for a knowledge economy. Again due to constraints on data availability, data utilized is as of 2012. However also this variable has been used in combination with variables from different years as it is considered to be rather stable (see e.g. Hall and Gingerich, 2009; Ioannou and Serafeim, 2012).

5_{For firm and sector level variables close attention was paid on gathering and assessing data according to the}

(42)

3.5.2.3 Trade Union Strength

Trade Union strength is assessed through assessing trade union density. Throughout the institutional research literature, trade union density is widely used to assess trade union strength. The measure is established by calculating the share of employed trade union members among all employees (Fitzenberger, Kohn and Wang, 2009).

The rational connecting trade union density to strength is that a higher number of paying trade union members results in better funding of a trade union. Continuing this further a financially well-funded trade union is able to better support trade union action like strikes and also is able to increase its representation and presence on the shop floor, which in turn increases

individual support for trade union action. Building up trade union support is essential for the ability to undertake strikes and inflict damage upon employers (Fitzenberger, Kohn and Lembcke, 2013).

Trade union density scores on a country level were taken from ICTWSS, a Database on Institutional Characteristics of Trade Unions, Wage Setting, State Intervention and Social Pacts in 34 countries between 1960 and 2012 (Visser, 2013). The database is collecting trade union data and makes it available for research purposes. Again, because of availability the data utilized is as of 2012. However referring to the stability of institutions it is common throughout institutional research to still use this data (see e.g. Benassi and Vlandas, 2015).

3.5.3 Control Variables

Assessing the literature featured in Section "2.1.3 Comparative CSR", the research field uses a rather homogeneous set of control variables. In previous research, on a firm level visibility, short-termism, financial performance, size and age have been suggested to be factors that affect CSR performance. Assessing the sector firms are in, some researchers also propose a relationship to CSR (Hull and Rothenberg, 2008). Finally considering country level

influences on CSR, economic development as well as the level of internationalization have been proposed to be relevant.

3.5.3.1 Firm Visibility

Dissertation What drives Corporate Social Responsibility Performance? The Role of national Institutions.

What drives Corporate Social Responsibility Performance? The

Role of national Institutions.

Dissertation

Abstract

Acknowledgements

Table of Contents

List of Abbreviations

Table of Tables

Table of Figures

Chapter 1: Introduction

Chapter 2: Theoretical Framework and Hypotheses

Chapter 3: Methodology