The influence of environmental factors on a firm's environmental performance

0

UNIVERSITY OF AMSTERDAM BUSINESS SCHOOL - EXECUTIVE PROGRAM IN MANAGEMENT STUDIES - STRATEGY TRACK

The influence of

environmental factors on a

firm's environmental

performance.

1

Statement of Originality

This document is written by Lizzy Buisman who declares to take full responsibility for the contents of this document. I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it. The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.

2

Table of Content

Governments and the environment ... 7

The firm and the environment ... 8

3. Theoretical Framework ... 11 Hypothesis 1 ... 11 Hypothesis 2 ... 12 Hypothesis 3 ... 14 Hypothesis 4 ... 15 4. Methodology ... 17 4.1 Sample ... 17

4.2 Independent variable: Air pollution ... 17

4.3 Dependent variable: Environmental performance ... 18

4.4 Moderator variables: Energy consumption, freedom of press and NGO density ... 18

4.5 Control variables: firm size, return on assets, industry and geographical region ... 19

5. Results ... 21

5.1 Descriptives ... 21

5.2 Normality ... 24

3

5.4 Regression analysis ... 29

6. Discussion ... 34

6.1 Academic relevance ... 34

6.2 Implications for practice ... 35

7. Limitations of the research ... 37

8. Conclusion ... 38

References ... 39

Appendices ... 44

Appendix 1: Theoretical model ... 44

Appendix 2: Normality analysis ... 44

Appendix 3: Process model 1 Hayes ... 58

List of figures

Figure 2. Descriptive statistics for continuous variables ... 24

Figure 3. Correlation matrix ... 28

Figure 4. Result of hierarchical linear regression analysis for hypothesis 1 ... 30

Figure 5. Result of simple moderation for hypothesis 2 ... 31

Figure 6. Result of simple moderation for hypothesis 3 ... 32

4

Abstract

In the past few years numerous environmental disasters have happened all over the world. These environmental disasters cause interference by governments, for example by means of regulation, and awareness among firms. The environment should be seen in the broadest sense of the word and is impacted by anything and everything that happens in the world.

In this research we measure the influence of air pollution in a country on the environmental performance of a firm, and test if energy consumption, environmental NGO density and press freedom have a moderator effect on this relationship. The data is controlled for by the

variables firm size, return on assets, geographical region and type of industry.

A sample of 458 firms from the FT Global 500 is used as a starting point of the research, furthermore data has, among others, been retrieved from the Thomson Reuters Asset4 database, BP research and Reporters Without Borders.

The results of this study show no significant relationship. The direct relationship between air pollution in a country and environmental performance of a firm, as well as the moderator relationships, main type of energy consumption, NGO density and freedom of press in a country, have proven not to be significant.

Nevertheless, the research might show a different outcome when the relationships are measured over a longer period of time due to the fact that some changes in the environment are not immediately accounted for and visible in a one year time span.

5

1. Introduction

Currently there are a lot of environmental problems in the world, examples one could think of are the tsunami at the coasts of Sri Lanka and Thailand, destroying countless lives and

people's habitat, and later the tsunami at the coast of Japan, causing a nuclear disaster on the Fukushima plant. Other events are, to a greater extent, influenced by human behaviour such as oil leakage which happened in the Gulf of Mexico or the ground water pollution in North America due to fracking for shale gas or oil. On a daily basis, everyone is contributing to CO2 emissions and energy waste. All these events, which are referred to as environmental disasters (Scherer & Palazzo, The Oxford Handbook of Corporate Social Responsibility, 2009), have an impact on the world as we know it.

Traditionally governments and non-governmental organisations (NGO’s) try to develop regulation on how to address the issues regarding climate change and the limitation of environmental damage, both globally and on a national level. Several initiatives have been introduced. Within Europe, the Europe2020 targets have been introduced in order to reduce greenhouse gas emissions by 20% in 2020. On a more global level, there are the Kyoto

protocol and the recently signed COP21. All of these are examples of such initiatives focussed on the improvement or sustainability of the climate.

However, it is among others, the firms who, for example with their factories, cause such damage. Although the private sector (e.g. firms) becomes more aware of their share towards environmental damage, and start to act to some extent more towards environmental friendly behaviour. In this thesis focus lies on exploring the relationship between national

environmental characteristics and a firm’s environmental performance. It is argued that the level of environmental performance of a firm is influenced by national environmental characteristics of a country in terms of air pollution.

6 The report is structured as follows, in the next section we will start with a literature review describing the relevant concepts such as the current theory on environmental performance and its relation to firms followed by the hypotheses. Second, the data and the methods that are used for analysis are illustrated by explaining how the data is collected and analysed. The section that follows will present the results, followed by a section that discusses the implications of the findings and the final conclusion.

7

2. Literature review

The environment

Environmental disasters (Chernobyl, global warming, overfishing of oceans, loss of bio-diversity, etc) ... do not halt at national borders but affect the life of people who become aware that their traditional nation state institutions have become unable to protect them from harm (Scherer & Palazzo, 2009). Examining the meaning of environmental cohesion uncovers that the understanding of this definition is often not as straight forward. In organisational and strategic literature the definition often refers to the organisational environment. The definition of environment can best be understood as the economic concept of "externality" (Berchicci & King, 2007). This comes down to all activities of a firm which are not internally focussed. However, another topic which is often discussed when referred to environment is the natural environment. The interaction between those two elements - natural environment and

organisational environment - as well as between each of them and various constituencies is what makes research complex (Etzion, 2007). The natural environment has no voice of its own (Prasad & Elmer, 2005). It is important to realise that the "needs" of the natural environment are never determined by itself but are derived from external factors such as groups and collective entities. The organisational environment, reverting to the natural

environment, consists of various key-actors such as: customers, the media, investors, activists, boards of directors and the public (Etzion, 2007).

Governments and the environment

Over the past decade governments have become drivers adopting public policy to promote and encourage businesses to behave in a responsible and sustainable manner (Aaronson & Reeves, 2002)). In trying to work on the prevention of environmental disasters and increase

8 environmental performance governments have introduced regulations such as the Kyoto protocol and COP21. Both focussing primarily on the reduction of air pollution. According to Whitley (1999), an essential feature of any political system is the power of the state: "the extent to which states dominate the economy and share risks such that businesses become dependent on state policies and actions". Given significant diversity across countries in the power of the state, it follows that the extent to which states "directly or indirectly regulate market boundaries, entry and exit, as well as set constraints on the activities of economic actors" (Whitley, 1999) through laws and regulations is also an important determinant of organisational outcomes in general, and social performance in particular (Ioannou &

Serafeim, 2012), such as environmental performance. Broadly, laws and regulations play an important role in facilitating the corporations engagement with the state, as well as with its key stakeholders (Aguilera & Jackson, 2003).

The firm and the environment

It is argued, between economists, what is best for firms in terms of market regulation. According to accepted economic theory regulation should hurt firms because, it increases costs and constrains the choices available to managers (Berchicci & King, 2007). According to an article of Porter and Van der Linde (1995) it is argued however that market regulation is not a bad thing at all, and that it can enhance competitiveness. In the case of environmental performance a firm’s reduction of its consumption and (environmental) resources can lead to profitable opportunities. This might positively influence firm performance as well as public welfare, the latter can in that case be seen as an additive.

Another claim which is made by Porter and Van der Linde (1995) is that, even though there are regulations, it does not mean that those cannot be profitable. They state that it is

9 impossible for a manager to foresee and discover all profitable opportunities, meaning that perfect information does not exist, and therefore regulations cannot be considered as negative per se. Regulation is a very influential driver for improving environmental practices (Sharma & Henriques, 2005). It is assumed that competitive advantage can be achieved if a firm should improve its environmental performance. This has proven to be the case especially for leading firms (Etzion, 2007) (i.e. firms in the FT500).

As could be derived from the previous paragraphs environmental performance is the outcome of a firm's strategic activities that manage (or not) its impact on the natural environment (Walls, Phan, & Berrone, 2011). The activities however can vary per firm and are dependent on what is needed for the firm to acquire specific capabilities and resources in order to create a competitive advantage.

Considerable research has been conducted into the relationship between the environment and the firm. However most of these studies focus on a combination of the environment or the environmental performance of the firm in relation to the financial performance of the firm (Berchicci & King, 2007). Another focus is a firm’s environmental performance and the relation to significant environmental events e.g. earth quakes or a tsunami (Berchicci & King, 2007). Adjacent to these remarks most of the research published in English, is primarily focussed on the United States (Etzion, 2007). Limited research is conducted on the relation of national environmental aspects and environmental performance of firms, but can be found in the article of Ioannou and George (2012), who study the impact of nation-level institutions on firms and corporate social responsibility. In other related studies, such as from Maignan and Ralston (2002) or Chapple and Moon (2005) corporate websites are researched on corporate social behaviour in respectively several Western countries and Asia. Hart and Ahuja (1996) researched the relationship between air pollution and a firm’s environmental performance with regards to the potential upside for firms to be green in terms of profit maximisation. For

10 future research one of their recommendations has been to examine the relation between lower emissions and the pollution prevention and emission reduction activities of firms. They questioned whether more profitability for firms leads to more investment in improving environmental performance.

A great deal of research has been done on the subject of firms and their environmental

performance. However often this research has a global or American focus. This research aims to uncover what national environmental aspects have an influence on the environmental performance of a firm. Focus lies on uncovering if there is a relation between air pollution in a country and environmental performance of a firm and whether this is moderated by the factors main type of energy consumption, environmental NGO density and press freedom on a national level.

11

3. Theoretical Framework

Hypothesis 1

The state, it is assumed, is capable of setting the rules in such a way that the consequences of market exchange contribute to (or at least do not harm) the well-being of society (Scherer & Palazzo, 2007). However, as the rules cannot cover all aspects, due to imperfection or

incompleteness, there remains a part where an appeal is done on the social responsibilities of a firm to do what is right. Therefore, it is argued that the reduction of air pollution can be achieved through two primary means: control and prevention (Forsch & Gallopoulos, 1989). The latter approach prevents the pollution, for example, during the manufacturing. This leads towards more efficiency and waste reduction, which is beneficial to the firm and the

environment. Whereas the control approach entails expensive, non-productive pollution control equipment to achieve compliance with existing regulation (Hart & Ahuja, 1996). Not only do prevention and control of air pollution save costs and increase productivity and efficiency, it could also enhance focus on a firms environmental objectives (Hart & Ahuja, 1996). Firms with a focus on minimising waste were found to make higher investments in pollution prevention which is related to environmental performance improvement (Berchicci & King, 2007). Pollution prevention strategies offer the potential to cut emissions well below the levels required by law (Rooney, 1993). It could be argued that, despite legislation, in countries where there is a low level of air pollution firms will have a higher focus on environmental performance, for example due to waste reduction in order to minimise costs, meaning that when the level of air pollution is higher in a country firms potentially have limited focus on environmental performance. Whiteman, Walker, & Parego (2012) even argue that while CO2 emissions (air pollution) reduction is unquestionably valuable, the overall resilience of the planet depends upon corporate sustainability initiatives being eco-efficient. However, if firms are aware that improving their environmental performance could lead to

12 profit maximisation, and to take it even further, that their improvements could seriously lead to a more sustainable planet, it should be researched if firms would act upon this. Therefore it would be interesting to test if the following hypothesis is true:

H1: In countries with a high level of air pollution (by means of CO2 emission per capita) firms will score lower on environmental performance.

Hypothesis 2

In most countries the production of electricity (for consumption) is one of the main sources of pollution (Fare, Grosskopf, & Hernandez-Sancho, 2004). Globally, around 87% of our total energy is generated (in order to be consumed) by fossil fuels (coal, oil and natural gas), of which 28% is generated from coal, 21% is generated from natural gas and the remaining 38% is generated from oil. About 6% of our total asset of energy in generated in nuclear plants, and the remaining 7% is generated from renewable resources, such as hydro (a major part), wind, solar, geothermal and bio fuels (Bose, 2010). According to Milieu Centraal, by means of the Lijst Emissiefactoren (2016), coal is the most polluting type of energy measured in terms of CO2 emissions followed by oil, natural gas, nuclear and hydroelectric. As it turns out, the global reserve of the fossil fuels is distributed such as that the most polluting fossil fuel has also the highest availability meaning, coal is the largest available fossil fuel and the dirtiest fuel on environmental pollution (Bose, 2010). This causes environmental problems. Over 75% of all energy-project support from international financial firms between 2008 and 2011 went to fossil fuel projects in 12 of the top developing-country emitters and twice as much in fossil fuel projects of developed-country emitters (Rashchupkina, 2015). Once in place, fossil fuel subsidies are extremely difficult to remove (Laan, Beaton, & Presta, 2010). Governments might be able to encourage improvement by publicising information about the costs and

13 benefits of certain activities (Berchicci & King, 2007). Whereas stakeholders wish to

encourage firms to improve their performance, but are impeded from doing so by minimising transaction costs. High availability also leads to lower prices for fuel and therefore lower production costs. One way to overcome this problem is for stakeholders to invest in innovations that are both valuable, in terms of profit maximisation, and better for the

environment (Berchicci & King, 2007). Government programs (i.e. subsidies such as SDE+ in the Netherlands or the German Energiewende) could have a positive influence on the firm’s choice of energy. A substantial portion of global energy demand can be met by promoting environmentally clean renewable energy sources, and the current trend in the world is to explore them vigorously (Bose, 2010). Although the share of renewable energy in the total fuel mix is marginal their potential is considerable. For example in the 1990s the global stock of wind turbines increased by an average of 27% per annum it should however be taken into account that the energy market is huge (Jacobsson & Bergek, 2004). Even with continued high growth rates, wind and solar power may only begin to replace the stock of conventional energy technologies well after 2020 (Jacobsson & Bergek, 2004). Taking this into account one could argue that in countries where the main energy source is renewable energy there is a high focus on (the improvement of) environmental performance. Combining the energy programs, and the renewable energy developments we therefore, in this research, assume that, in countries that focus on renewable energy sources the main type of energy source for

consumption is relatively green, and that this has an effect on the relationship between air pollution and the environmental performance of the firm. Leading to the following

14

H2: A county’s main energy source for consumption moderates the negative relationship between the air pollution in a country and the environmental performance of a firm, such that the cleaner the main energy source consumed in a country, the weaker the relationship between the air pollution in a country and the environmental performance of a firm.

Hypothesis 3

Non-governmental organisations who were once focussed on influencing governments have shifted towards targeting business firms in order to make them more responsive to social and environmental concerns (den Hond & de Bakker, 2007) (Doh & Guay, 2006). Environmental NGO's usually target larger firms (and not necessarily the worst polluters), because these firms are often the most likely to respond in order to avoid damage to their reputations (Bianchi & Noci, 1998) (Greve, 1989). NGO's that manage to garner public support for their campaigns are more likely to succeed in changing corporate behaviour (Eesley & Lenox, 2006). Reflecting on the peculiar status of NGO's in relation to companies, some authors suggest that they should be seen as "stake seekers", claiming to have a stake in the

corporation's decision making (Holzer, 2008). Indeed, groups such as social movements might "declare" themselves as stakeholders, even if a company is reluctant to grant them this status, because it believes they are not representative. The company could also dislike their methods or prefer to deal with other groups. Ignoring such groups carries risks, as has been

documented by the case of Shell with the Ogoni (Wheeler, Fabig, & Boele, 2002). In the CSR context, of which environmental performance is a part, there has been an increasing

institutionalisation of NGO activity (Arenas, Lozano, & Albareda, 2009). Many companies have included NGO's in their stakeholder dialogue, which is clearly related to the

15 participation in global networks and interest in global issues, NGO influence and relationship with business also depend heavily on nation and regional contexts (Doh & Guay, 2006). Therefore we assume the following hypothesis:

H3: The density of environmental NGO’s in a country moderates the negative relationship

between air pollution in a country and the firm environmental performance, such that this relationship is weaker for firms from countries with a higher density of environmental NGO’s.

Hypothesis 4

Firms could focus on limiting costs and waste which could lead to a higher environmental performance driven from the inside of the organisation. Studies also show that environmental performance has value from an external perspective. More specifically, studies have found that independent company ratings and ranking schemes (e.g. Consumer Reports, Moody's) can significantly influence the behaviour of consumers (Sen & Bhattacharya, 2001) as well as the organisations that are being rated (Chatterji & Toffel, 2010). In other words, increasing environmental awareness among consumers and the comparison between firms could lead to an enlarged focus on a firm's environmental performance. This is most likely to occur in countries where there is a high level of transparency, freedom of press.

In a setting where there is a higher freedom of press the environmental damage a firm causes shall be noticed and act up on faster by the public. Corporate leaders of environmentally damaging firms shall especially be concerned about the fact that they would be exposed to the local press or civil society actors (King, 2008) (Campbell, 2005). According to Hoffman and Ocasio (2001) managers typically infer general public awareness of an environmental issue by the manner and scope of its coverage in mainstream media outlets. The same thing applies on

16 a country level where governments have to justify why certain measures are (not) taken in order to decrease air pollution. As mentioned by Aguilera et al. (2007) business organisations are embedded in different national systems, they will experience divergent degrees of internal and external pressures to engage in social responsibility initiatives. We assume that the level of air pollution and the relation we expect it to have with the firms’ environmental

performance is also influenced by external factors such as freedom of press.

This leads to the following hypothesis:

H4: The freedom of press in a country moderates the negative relationship between air pollution in a country and firm environmental performance, such that the relationship is weaker for firms from countries with a higher level of freedom of press.

17

4. Methodology

To test the hypotheses introduced, cross sectional data was gathered through database research. The data is not derived from one single source but from multiple sources. In this section, the selection of the data and the method shall be explained by describing the sample size and the different variables.

4.1 Sample

The firms that are examined in this research reflect the Financial Times (FT) Global 500. This is an annual snapshot of the world's largest firms in terms of market capitalisation. For this research the FT Global 500 published in March 2015 with annual data of 2014 is used.

The data concerning the independent, dependent, moderator and control variables are collected for the year 2014. The FT 500 firms are checked for the availability of the

environmental performance from ASSET4. A total of 42 firms are eliminated due to missing information, resulting in the final list of 458 firms.

4.2 Independent variable: Air pollution

In this research air pollution is defined as the level of CO2 emission in a country. The CO2 emission reported in the dataset reports the level of CO2 emission per capita of fossil fuel use and industrial processes for every country. The data is derived from the Emission Database for Global Atmospheric Research (EDGAR) a joint research centre by the European

18 4.3 Dependent variable: Environmental performance

The data on environmental performance is derived from the Asset4 database on

Environmental Social and Governance (ESG) metrics. This score is based on 70 different key performance indicators determined and measured for each firm by Thomson Reuters, with three different overarching subjects: emission reduction, product innovation and resource reduction. A high score indicates a high level of environmental performance.

4.4 Moderator variables: Energy consumption, freedom of press and NGO density

There are three moderator variables analysed in this research in order to research if a firms’ environmental performance is influenced by environmental factors.

The data which provides more detail on the energy consumption per country is derived from the Statistical Review of World Energy, the research department of BP which conducts and publishes annual researches on energy, both on global and national levels. The research indicates, per country, the distribution of energy sources in terms of consumption. In this research, the energy source with the highest consumption in a country is set to be the main energy source. The main energy sources found, in order of most to least polluting in terms of CO2 emissions, are coal, oil, natural gas, nuclear and hydroelectric (Lijst emissiefactoren, 2016). In the analysis the most pollution energy source is coded null against the least polluting energy source which is coded four.

Freedom of press is measured by the Press Freedom Index which is an annual ranking

published by Reporters Without Borders resulting in a global ranking of countries. This index is based on questionnaires which are sent on a yearly basis to partner organisations and their

19 150 correspondents around the world (Borders, 2016). The variable is a ranked from a high level of press freedom (1) to countries with little to no freedom of press (175).

In order to determine the density of environmental NGO's per country based on the FT

Global500 list we look at the International Union for Conservation of Nature (IUCN) which is an international environmental organisation with over a 1,000 member organisations. The number of NGO's per country based on this list shall be divided by the total population of a country in order to determine the density (Esty, Levy, Srebotnjak, & de Sherbinin, 2005). The higher the outcome, the higher the NGO density in a country.

4.5 Control variables: firm size, return on assets, industry and geographical region

In this research, four variables have been chosen to control for any variations in outcome namely, firm size, return on assets, industry and geographical region. All of these variables have been shown in previous literature to affect the environmental performance of a firm. All the control variables are gathered for the year 2014 from FT Global 500 (firm size and return on assets) and Thomson Reuters (industries).

Firm size is controlled for, because even though the firms in the sample are part of the largest 500 firms globally, there are still sizeable differences which may have an effect on their environmental performance. Larger firms might have more budget available to invest in environmental measures that influence their environmental performance, then smaller firms might have. Firm size is measured by means of the total assets of the firm.

Another control variable is return on assets. Return on assets indicates how many dollars of earning a firm can derive from each dollar of assets that they control. The return on assets is calculated by the net income divided by the average total assets of the firm. The number is

20 used for numerical comparison of the firm.

Industry is also a control variable which is included. Industries are compared because firms in the same industry might have similarities. The distinction in industries is based on the

Thomson Reuters Business Classifications which is an industry classification for global companies entailing 10 overarching industry classifications namely: financials, basic materials, cyclical consumer goods, energy, healthcare, industrials, non-cyclical consumer good, technology, telecommunication and utilities. In order to control for this variable, various dummy variables are created, the reference group is financials.

The environmental performance might also be influenced by the geographical region of the firm, because different continents might have different environmental standards to which firms need to adhere. By incorporating the continent from which the firm originates these differences are controlled for. Dummy variables are created with Europe as the reference group.

21

5. Results

5.1 Descriptives

In this study, data is collected from all the Financial Times 500 firms in 2015 for 2014. However, due to missing data for some of the variables, the final number of firms which are used for the research is brought down to a total 458 firms.

Three categorical variables are used in this research, which are, the moderator variable the country of origin of the firm, and the control variables type of industry and geographical region, the continents. For the latter three variables dummy variables are created. The majority of the firms in the sample originate from the United States (42.4%) followed by Japan (7.6%) and the UK (5.9%), while Israel, Qatar and Thailand are some of the least represented countries (0.2%). Figure 1 shows a complete overview of countries represented in the sample.. The majority of the firms operate in the financial sector (26.6%) followed by the industrial sector (15.1%) and the cyclical good sector (11.6%) the smallest sector in this sample is the utilities sector (1.1%). Finally, in terms of geographical regions most firms are located in North America (46.9%) while the least of the firms in this sample are located in the continent Africa (0.7%).

22 Figure 1. Descriptive statistics for categorical variables

Variable Level N % Country Australia 10 2.2 Belgium 2 0.4 Brazil 6 1.3 Canada 18 3.9 China 23 5.0 Denmark 3 0.7 Finland 2 0.4 France 21 4.0 Germany 18 3.9 Hong Kong 18 3.9 India 14 3.1 Indonesia 3 0.7 Israel 1 0.2 Italy 6 1.3 Japan 35 7.6 Mexico 3 0.7 Netherlands 6 1.3 Norway 3 0.7 Qatar 1 0.2 Russia 5 1.1 Singapore 4 0.9 South Africa 3 0.7 South Korea 4 0.9 Spain 5 1.1 Sweden 10 2.2 Switzerland 10 2.2 Taiwan 2 0.4 Thailand 1 0.2 UK 27 5.9 US 194 42.4 Industry Financials 122 26.6 Basic Materials 31 6.8

Cyclical consumer goods 53 11.6

Energy 48 10.5

Healthcare 34 7.4

Industrials 69 15.1

Non-cyclical consumer goods 41 9.0

Technology 32 7.0

23

Utilities 5 1.1

Geographical region Europe 114 24.9

Asia 110 24.0

Africa 3 0.7

North America 215 46.9

South America 6 1.3

Australia 10 2.2

Next to the three categorical variables, seven continuous variables have been used in this research namely, the dependent variable environmental performance and the independent variable air pollution, the moderator variables main type of energy consumed in a country, press freedom and NGO density and the control variables firm size and return on assets. The descriptives of these variables are shown in figure 2.

The dependent variable, environmental performance, ranges between 8.46 and 94.91, with a mean of 72.8 and a standard deviation of 27.17. The values for air pollution range between 1.8 and 39.13 with a mean of 1.61 and a standard deviation of 5.32.

Subsequently, concerning the moderator variable, main type of energy consumed in a country the vulues range between 0 and 4 with a mean of 1.08 and a standard deviation of 0.65. For press freedom, the values range between 1 and 175 with a mean of 54.69 and a standard deviation of 40.83. NGO density has a minimum value of 0.01 and maximum value of 1.78 with a mean of 0.29 and a standard deviation of 0.3.

The control variables, at last, show a range of 2,264.1 and 3,316,893 for firm size with a mean of 218,417.7 and a standard deviation of 463,803.4, the return on assets is ranged between -0.03 and 0.59 with a mean of 0.06 as well as a standard deviation of 0.06.

24 Figure 2. Descriptive statistics for continuous variables

Variable N Min Max Mean Std. Deviation

Environmental performance 458 8.46 94.91 72.8 27.17 Air pollution 458 1.8 39.13 11.61 5.32 Energy 458 0 4.00 1.0808 0.65156 Press freedom 458 1 175 54.69 40.83 NGO density 458 0.01 1.78 0.29 0.3 Firm size 458 2264.1 3316893 218417.7 463803.4 Return on assets 458 -0.03 0.59 0.06 0.06 5.2 Normality

In order to measure the normality of the seven continuous variables a normality test was conducted for environmental performance, air pollution, main type of energy consumed in a country, press freedom, NGO density, firm size and return on assets. The descriptives of the normality tests can be found in appendix 2 figure 1. The output of the Kolmogorov-Smirnov test can be found in figure 2 of appendix 2.

The dependent variable environmental performance has a substantial negative skewness of -1.21 and a kurtosis of 0.03 meaning that the variable clustered from the left to the centre and peaked to the right. The variable was transformed using the formula Y=Log10 (K-X),

improving skewness to 0.143 and kurtosis to -1.302. Even though the data is transformed the variable Log Environmental performance still has a significant different distribution than normal, as shown in the Kolmogorov-Smirnov test. However, the skewness and kurtosis are within the limits of +2 and -2 which is acceptable for a normal distribution and therefore should not have a drastic effect on the results (George & Mallery, 2010). This is supported by the Q-Q plot (graph 2 of appendix 2) which shows a slightly skewed distribution but still close to a straight line.

-25 0.03. A log transformation does not improve the normal distribution of the data or the

significance for the Kolmogorov-Smirnov test. The Q-Q plot on the air pollution distribution (Appendix 2, graph 5) shows however that the values are close to a straight line. Taking into account these factors the assumption is that the distribution of the data is close to normal. The moderator variable main type of energy consumed in a country has a skewness of 1.968 and a kurtosis of 5.559. A Log or Sqrt transformation does not lead to an improvement of these factors or the Kolmogorov-Smirnov test. In appendix two graph 8 shows that the values are not completely, but close to a straight line. While air pollution and type of energy consumed in a country did not respond well to the transformations, moderator variable press freedom did. The transformation to Sqrt Press freedom, using the formula Y=√(X), improved the skewness from 1.8 to 0.809 and the kurtosis from 2.57 to 1.21 which now are in line with George and Mallery (2010). The Q-Q plot in appendix 2 graph 11 shows that the distribution is almost in a straight line, even though the Kolmogorov-Smirnov test is not significant. The expectation based on this data is however that the data is close to normal.

The moderator variable environmental NGO density also responds well to transformation, when using the formula Y=Log(X) the skewness and kurtosis improve from 1.8 and 2.57 towards 0.0809 and 1.21 are currently within the recommended range of -2 to +2 and with an improvement of the Q-Q plot they are close to a straight line. Regardless the Kolmogorov-Smirnov test did not change towards a significant outcome so we still assume normality. Finally the control variables, firstly return on assets, before transformation the skewness (3.67) and the kurtosis (15.42) were not in line with the recommended range by George and Mallery (2010) transforming the data by applying the Y=Log(X) formula the skewness (0.541) and the kurtosis (0.022) changed to a respectable level. The histogram in appendix 2 graph 16 starts to look like a normal distribution. The normality test of Kolmogorov-Smirnov however does not show significance for normality. The Q-Q plot (appendix 2 graph 17) looks

26 like the data is normally distributed and we therefore do not expect a drastic effect on the result.

Finally, the control variable return on assets was tested. Before transformation there was a strong positive skew of 2.47 and a kurtosis of 11.62. The only transformation leading to a slight improvement is the logarithmic transformation. Log Return on assets had a slightly better skewness of 1.96 and a kurtosis of 6.69. This indicates a significantly different distribution than previously supported by the Q-Q plot in appendix 2 graph 20 and the Kolmogorov-Smirnov test. However due to the transformation, Log is still the best representation of return on assets.

5.3 Correlation

Correlations between the various variables are investigated using the Pearson correlation coefficient. The variables industry and geographical region are dummy variables. The result of the Pearson’s correlations test is presented in figure 3.

The independent variable air pollution is positively and significantly correlated (r=0.17, N=458, Sig. = 0.000) to log environmental performance. The correlation contrasts with the hypothesised direction of the relationship between the level of air pollution in a country and the log environmental performance.

The moderator variable energy consumption is significantly negatively correlated to the independent variable air pollution (r=-0.107, N=458, Sig.=0.022) as well as to the dependent variable log environmental performance (r=-0.179, N=458, Sig.=0.000). This entails that the dirtier the energy consumed the higher the air pollution and the lower the log environmental performance.

27 (r=-0.192, N=458, Sig.=0.000) but significantly positively correlated to log environmental performance (r=0.310, N=458, Sig.=0.000).

Log NGO density is significantly correlated to air pollution (r=0.178, N=458, Sig.=0.000) and log environmental performance (r=-0.219, N=458, Sig.=0.000). This entails that the higher the NGO density the lower the air pollution and the higher the environmental performance. Because none of the correlations are above 0.8, multicollinearity is unlikely to be a problem. In terms of control variables log total assets is negatively correlated towards both air pollution (r=-0.163, N=458, Sig.=0.000) and log environmental performance (r=-0.219, N=458,

Sig.=0.000) however only the correlation with log environmental performance is significant. The correlation towards the moderator variables energy consumption (r=0.030 N=458, Sig.=0.518) sqrt press freedom (r=0.041, N=458, Sig.=0.381) and log NGO density (r=0.000, N=458, Sig.=0.998) are all positive however all these relationships are not significant. Furthermore, the control variable for geographical regions except for Africa (r=-0.064, N=458, Sig.=0.170) all variables have a significant relationship towards air pollution. Asia (r=-0.403, N=458, Sig.=0.000) and South America (r=-0.197, N=458, Sig.=0.000) have a negative correlation meaning that there is more air pollution in these countries compared to Australia (r=0.160, N=458, Sig.=0.001) and North America (r=0.825, N=458, Sig.=0.000) where the influence would be positive and therefore lower the influence on the level of air pollution.

28 Figure 3. Correlation matrix

Variables Mean SD 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

1 Log Envi ronmental performa nce 1,0129 0,58501 1 2 Ai r pol l ution 11,606 5,32411 0,170** 1

3 Energy 1,0808 0,65156 -0,179** -0,107* 1

4 Sqrt Pres s freedom 6,966 2,48691 0,310** -0,192** -0,452** 1 5 Log NGO dens i ty -1,7194 1,05306 -0,267** 0,178** 0,448** -0,750** 1 6 Log Total As s ets 11,1272 1,43035 -0,219** -0,163** 0,03 0,041 0 1 7 Log Return on As s ets 0,0601 0,05735 0,02 0,073 -0,31 -0,037 0,015 -0,576** 1 8 Europe vs North Ameri ca 0,4694 0,49961 0,196** 0,825** -0,117* -0,121** 0,121** -0,230** 0,116* 1 9 Europe Vs As i a 0,2102 0,42766 0,112* -0,403** -0,313** 0,664** -0,705** 0,07 -0,073 -0,529** 1 10 Europe vs Afri ca 0,0066 0,08076 0,046 -0,064 -0,135** -0,016 0,04 -0,049 0,005 -0,076 -0,046 1 11 Europe vs South Ameri ca 0,0131 0,11383 -0,009 -0,197** -0,014 0,166** -0,075 0,031 0,01 -0,108* -0,065 -0,009 1 12 Europe vs Aus tra l i a 0,0218 0,1463 -0,52 0,160** -0,019 -0,101* 0,214** 0,061 0,017 0,141** -0,084 -0,012 -0,017 1 13 Fi na nci a l s vs Ba s i c ma teri a l s 0,0677 0,25148 -0,057 -0,009 -0,033 0,04 -0,027 -0,088 0,017 -0,044 -0,009 -0,022 0,045 0,079 1 14 Fi na nci a l s vs Cycl i ca l cons umer goods 0,1157 0,32024 0,109* 0,095* 0,008 0,029 0,04 -0,200** 0,151** 0,097* -0,044 0,055 -0,042 0,039 -0,097* 1 15 Fi na nci a l s vs Energy 0,1048 0,30663 0,053 0,006 0,034 0,067 -0,054 0,036 -0,061 0,035 -0,009 -0,028 0,023 -0,051 -0,092* -0,124** 1 16 Fi na nci a l s vs Hea l thca re 0,0742 0,26244 0,140** 0,098* -0,022 -0,07 0,075 0,150** 0,101* -0,081 -0,023 -0,033 0,015 -0,076 -0,102* -0,097* 1 17 Fi na nci a l s vs Indus tri a l s 0,1057 0,3581 -0,201** 0,004 0,023 -0,043 -0,053 -0,094* -0,005 0,017 0,092* -0,034 -0,049 -0,063 -0,113* -0,152** -0,144** -0,119* 1 18 Fi na nci a l s vs Non cycl i ca l cons umer goods0,089 0,2858 -0,04 -0,06 0,055 -0,042 0,117* -0,190** 0,167** -0,04 -0,105* -0,025 0,031 -0,047 -0,084 -0,113* -0,107* -0,089 -0,132** 1 19 Fi na nci a l s vs Technol ogy 0,699 0,2552 -0,084 0,118* -0,034 -0,048 -0,01 -0,181** 0,0227** 0,137** -0,054 -0,022 -0,032 -0,041 -0,074 -0,099* -0,094* -0,078 -0,115* -0,086 1 20 Fi a nci a l s vs Tel ecommuni ca tion 0,0502 0,21863 -0,059 -0,101* 0,018 -0,039 0,049 0,001 0,028 -0,116* 0,058 0,105* -0,026 0,034 -0,062 -0,0831 -0,079 -0,065 -0,097* -0,072 -0,0631 1 21 Fi a nci a l s vs Util i ties 0,0109 0,10403 -0,06 -0,017 0,052 -0,047 0,038 0,031 -0,082 -0,015 -0,059 -0,009 -0,012 -0,016 -0,028 -0,038 -0,036 0,03 -0,044 -0,033 -0,029 -0,024 1

**. Correlation is significant at the 0.01 level (2-tailed). *. Correlation is significant at the 0.05 level (2-tailed).

29 5.4 Regression analysis

In order to test the stated proposition and answer the research question, a hierarchical linear regression analysis is conducted. In the regression log environmental performance is added as the dependent variable. The control variables, industry, geographical region, firm size and return on assets are entered into the first model of the regression. Then, in the second model the independent variable air pollution is included. For hypothesis 2 to 4, concerning the moderating effect, the PROCESS application of Hayes (2012) is used. Thus, model 1 for simple moderation is applied, as shown in appendix 3, where Y exhibits the dependent

variable, X the independent variable and M the moderator variable, XM shows the interaction effect.

The result of the regression regarding hypothesis 1 environmental performance is presented in figure 4 at the end of this paragraph. The results show that 29.1% of the variance concerning environmental performance was explained by the control variables. Model 1 is significant for environmental performance (F=11.318, Sig.=0.000). When the independent variable air pollution is added in model 2, no additional variance within the model is explained. The variable has no significant influence on the model (F=10.642, Sig.=0.687) meaning, that there is no relationship between air pollution and environmental performance and the first

30 Figure 4. Result of hierarchical linear regression analysis for hypothesis 1

Environmental performance

Variable Model 1 Model 2

B Sig. B Sig.

Control variables

Log Firm size -0.173 0 -0.173 0

Log Return on Assets -1.125 0.03 -1.116 0.032

Europe Vs Asia 0.393 0 0.389 0

Europe vs Africa 0.372 0.211 0.368 0.217

Europe vs North America 0.355 0 0.318 0.004

Europe vs South America 0.169 0.424 0.184 0.392

Europe vs Australia 0.071 0.674 0.03 0.878

Financials vs Basic materials -0.484 0 -0.487 0

Financials vs Cyclical consumer goods -0.28 0.004 -0.281 0.004

Financials vs Energy -0.221 0.014 -0.211 0.014

Financials vs Healthcare -0.179 0.109 -0.184 0.105

Financials vs Industrials -0.673 0 -0.676 0

Financials vs Non-cyclical consumer goods -0.444 0 -0.443 0

Financials vs Technology -0.652 0 -0.655 0 Financials vs Telecommunication -0.422 0 -0.422 0 Financials vs Utilities -0.5 0.321 -0.501 0.031 Independent variable Air pollution 0.004 0.687 R² 0.291 0.291 R² change 0.291 0 F 11.318 10.642 Sig. F Change 0.000 0.687

31 The results for hypothesis 2 are shown below in figure 5. The regression coefficient for XM concerning Log Environmental performance is b3=-0.0081 and is not significant, t (458)=-0.6154, Sig.=0.5386. Thus, the effect that the air pollution has on environmental performance does not depend on the type of energy consumed. Hypothesis 2 remains unsupported due to lack of statistical support.

Figure 5. Result of simple moderation for hypothesis 2

Coefficient SE T P

Intercept i1 3,0314 0,3451 8,7849 0

Air pollution (X) b1 0,0154 0,0203 0,7592 0,4482

Energy (M) b2 0,0203 0,0906 0,2245 0,8225

Energy x Air pollution (XM) b3 -0,0081 0,0132 -0,6154 0,5386

Log Total Assets -0,1756 0,0243 -7,2374 0

Log Return on Assets -1,1132 0,5211 -2,1365 0,0332

Europe Vs Asia 0,3677 0,0816 4,5056 0

Europe vs Africa 0,3064 0,3082 0,9943 0,3206

Europe vs North America 0,2683 0,1229 2,1832 0,0296

Europe vs South America 0,1829 0,2174 0,8415 0,4005

Europe vs Australia -0,0207 0,2035 -0,1018 0,919

Financials vs Basic materials 0,4943 0,1099 -4,497 0

Financials vs Cyclical consumer goods -0,284 0,0965 -2,9461 0,0034

Financials vs Energy 0,2146 0,0903 -2,3768 0,0179

Financials vs Healthcare -0,1875 0,1121 -1,6718 0,0953

Financials vs Industrials -0,6796 0,0854 -7,9599 0

Financials vs Non-cyclical consumer goods -0,4462 0,1055 -4,2275 0

Financials vs Technology -0,6592 0,1148 -5,7415 0 Financials vs Telecommunication -0,4186 0,1195 -3,5027 0,005 Financials vs Utilities -0,5038 0,2322 -2,1692 0,0306 R² 0,29 F (19,438)=9,5397 Sig. F Change <0.001

32 Figure 6, below, presents the result for hypothesis 3. According to this hypothesis the NGO density should have a positive effect on the relationship between air pollution and

environmental performance. The results show, however, that regression coefficient for XM is b3=-0.0079 and is not significant, t (458)=0.007, Sig.=0.2601. Meaning the moderator

variable NGO density does not influence the relationship between air pollution and environmental performance.

Figure 6. Result of simple moderation for hypothesis 3

Coefficient SE T p

Intercept i1 3,0272 0,344 8,8 0

Air pollution (X) b1 -0,0019 0,016 -0,1175 0,9065

Log NGO density (M) b2 -0,0897 0,0649 -1,3813 0,1979

Log NGO density x Air pollution (XM) b3 -0,0079 0,007 -1,1277 0,2601

Control variables

Log Total Assets -0,1795 0,0238 -7,5273 0

Log Return on Assets -1,2491 0,5113 -2,4432 0,015

Europe Vs Asia 0,0711 0,1121 0,6338 0,5265

Europe vs Africa 0,2917 0,2939 0,992 0,3215

Europe vs North America 0,1442 0,1218 1,1843 0,2369

Europe vs South America 0,0253 0,2184 0,1156 0,908

Europe vs Australia 0,1671 0,228 0,7326 0,4624

Financials vs Basic materials 0,524 0,1083 -4,8391 0

Financials vs Cyclical consumer goods 0,2741 0,0952 -2,8792 0,0042

Financials vs Energy 0,2375 0,0885 -2,6828 0,0076

Financials vs Healthcare 0,1697 0,1106 -1,5349 0,1255

Financials vs Industrials 0,6749 0,0838 -8,0547 0

Financials vs Non-cyclical consumer goods 0,4251 0,1042 -4,0811 0,0001

Financials vs Technology 0,6598 0,1128 -5,8515 0 Financials vs Telecommunication 0,3792 0,1181 -3,2115 0,0014 Financials vs Utilities 0,5417 0,2282 -2,3738 0,018 R² 0,3184 F (19,438)=10,7664 Sig. F Change <0.001

33 The final hypothesis, hypothesis 4, states that the press freedom has a positive moderator influence on the relationship between air pollution and environmental performance. The result on the regression analysis, in figure 7, shows a regression coefficient of XM of b3= 0.0039 and lacks significance (t(458)=0.0034, Sig.=0.255). This means that the hypothesis is not supported.

Figure 7. Result of simple moderation for hypothesis 4

Coefficient SE T P

Intercept i1 2,9615 0,4257 6,9575 0

Air pollution (X) b1 0,0492 0,0357 1,378 0,1689

Sqrt Press freedom (M) b2 -0,0161 0,0357 1,378 0,1689

Sqrt Press freedom x Air pollution (XM) b3 0,0039 0,0034 1,1398 0,255

Control variables

Log Total Assets -0,1854 0,0234 -7,919 0

Log Return on Assets 1,2395 0,504 -2,4592 0,0143

Europe Vs Asia -0,033 0,1224 -0,2448 0,8067

Europe vs Africa 0,1924 0,2919 0,6592 0,5101

Europe vs North America 0,1018 0,1739 0,5854 0,585

Europe vs South America -0,1536 0,2288 -0,6712 0,5025

Europe vs Australia -0,0205 0,2556 -0,08 0,9363

Financials vs Basic materials -0,5267 0,1064 -4,9499 0

Financials vs Cyclical consumer goods 0,2722 0,093 -2,9257 0,0036

Financials vs Energy -0,2449 0,872 -2,8094 0,0052

Financials vs Healthcare -0,1663 0,1083 -1,5364 0,1252

Financials vs Industrials -0,6293 0,0825 -7,6293 0

Financials vs Non-cyclical consumer goods -0,4692 0,1021 -4,5954 0

Financials vs Technology -0,6174 0,1111 -5,5583 0 Financials vs Telecommunication -0,3675 0,113 -3,1688 0,0016 Financials vs Utilities -0,53 0,2246 -2,3593 0,0187 R² 0,34 F (19,438)=11,08091 Sig. F Change <0.001

34

6. Discussion

This research is conducted to examine the influence of air pollution in a country on a firm's environmental performance, more specific the focus lies on the air pollution in a firm’s home country and measures this against the firm’s environmental performance. Subsequently it is tested how this relationship might be influenced by national factors such as the main type of energy consumption i.e. coal, oil, natural gas, wind and hydroelectricity from dirtiest to cleanest, environmental NGO density and press freedom in a country. The firms' home country is chosen as the basis and is measured against the statistics of that particular country. To answer the research question multiple hypothesis are created, which stated that the higher the air pollution in a country the lower the firms environmental performance. The main

energy source, NGO density and press freedom are expected to moderate the relationship. The results of the regression analysis, however, do not support the propositions and are all

rejected. This means that the independent variable and none of the moderator variables seems to play a part in the proposed relationship. The result will be further discussed and analysed in the discussion part in order to find a possible explanation for the overall absence of significant outcomes.

6.1 Academic relevance

The result indicates that none of the hypothesised relationships are significant, this might indicate that the theoretical arguments are not valid. Another option could be that there are other elements in this research design that have an influence on the study result. This study has set the stage for future research concerning the exploration of environmental pollution-related aspects in a firms' home country, such as air pollution in a country on a firm’s environmental performance. We will now further discuss a few possible reasons for the

35 insignificance of the hypothesis in order to create a better understanding of how to proceed in the future.

The first possible explanation for the insignificant findings of this research is that the time frame of the research might have an effect on the outcome. It is possible that when the relations between the variables are studied for a longer period of the time, the research shows a different result. It might be that firms only started acting on environmental performance recently or that there have been recent changes in the level of air pollution which cannot be captured in a sample of a one year period. Therefore it would be interesting to research if the same results apply in if a dataset with, for example, a tenor of ten year is analysed.

Secondly, the data shows that a large part of the sample is from the geographical areas North America, Europe and Asia. The fact that there are a few under represented areas might have influenced the outcome of the research since there are simply too few firms in some regions to make a valid comparison which might have affected the diversity.

6.2 Implications for practice

Firms increasingly focus on improving their environmental performance and it is important to know what drives this. It could be argued that regulations focussed on the environment play a part in this, but the awareness and willingness of the firm to improve their environmental performance also plays a role. Besides, over the past years, a number of environmental catastrophes took place and the dialogues regarding the environment, in specific the decreasing air-pollution on a national level, are taking place more than ever.

This study responds by creating an insight in to the relationship between air pollution on a national level and a firm’s environmental performance in its home country. To a certain extend a firm is able to influence its own level of air pollution influencing its environmental

36 performance. However, it is important to put the implication in perspective, because in a country with a very high level of air pollution one firm with a high environmental performance will not make the difference. This could be researched by focusing on one country while researching a larger sample of firms in that country and comparing environmental performance scores.

37

7. Limitations of the research

The main strength of the current research design is the use of mixed variables, the ones that have been used extensively in environmental research and have proven to impact the firm and the addition of new variables. Regardless there are certain limitations to the research that need to be addressed and can be used for future research.

The first limitation has already been discussed in the previous section, implicating that an extended time frame of the research might benefit to the significance of the hypotheses and thus on the research. The impact of air pollution on the environmental performance of the firm might take some time to become visible and therefore a period of one year might not be sufficient for the topic of this research. As mentioned before this might be more visible in a research with, for example, a ten year period since variations from year to year can be measured and compared.

A second option for future research could be to replace the independent variable in order to test if the moderators have a significant effect when there is a different main relationship. The current research has shown no significant relationships between the dependent and

independent variable and the dependent and independent variables in combination with the moderators. The moderators also fail to have a significant main effect on the dependent variable. A possible new independent variable could, for example, be cultural related. It might be interesting to research if the score of on environmental performance of a firm might have a dependency related to the culture of the country.

38

8. Conclusion

This research is focused on exploring if there is a relationship between air pollution in a firm's home country and the environmental performance of that firm. Next to that it is investigated if there is a moderator effect to this proposed relationship for main type of energy consumption, NGO density and press freedom.

Over the years, the environment has become more and more important due to events such as tsunamis or nuclear catastrophes. There are global measures that should reduce air pollution and firms are measured for the environmental behaviour. It is expected therefore that the level of air pollution in a country would have a negative effect on the environmental behaviour of a firm. Regarding the moderator variables (main type of energy consumption, NGO density and press freedom) it is expected that there is a moderating relationship. The type of energy consumption of a country is a choice that can be made and the high representation of NGO's and press freedom could make or break a firm’s image. The moderator variables are therefore expected to have a negative influence environmental performance of a firm.

In order to analyse the proposed relationships, data was collected from the FT Global 500 firms, the Thomson Reuters Asset4 database, BP studies, IUCN, EDGAR and Reporters Without Borders.

The results of the study showed that the level of air pollution in a country has no significant influence on the environmental performance of the firm. This was also not influenced by the results of the moderator variables main type of energy consumption, NGO density and press freedom. This means that according to this research none of the variables have a significant impact on a firm’s environmental performance. The outcome might have been affected by the time frame of the research or might have been significant if another independent variable was investigated. However, due to the time frame of this research it is not possible to test if these changes are feasible. This might be options for future research on this topic.

39

References

Aaronson, S., & Reeves, J. (2002). The European response to public demands for global corporate responsibility. National policy assiciation , 1-52.

Aguilera, R., & Jackson, G. (2003). The cross-national diversity of corporate governance: Dimensions and determinats. Academy of management Review , 28 (3), 447-465.

Aguilera, R., Ganapathi, J., Rupp, D., & Williams, C. (2007). Putting the S back in corporate social responsibility: A multilevel theory of social change in organizations. Academy of

Management Review , 32 (3), 836-863.

Arenas, D., Lozano, J., & Albareda, L. (2009). The role of NGO's in CSR: Mutual Perceptions among stakeholders. Journal of Business Ethics , 88 (1), 175-197.

Berchicci, L., & King, A. (2007). Postcards from the Edge: A review of the Business and Environment Literature. ERIM Report Series Reseach In Management , 1-60.

Bianchi, R., & Noci, G. (1998). Greening SMEs' competitiveness. Small Business Economics

, 11, 269-281.

Borders, R. w. (2016). World press freedom index. Opgeroepen op 2016, van Reporters without borders for freedom and information: https://rsf.org/en/ranking/2014

Bose, B. (2010). Energy, environmental pollution, and the impact of power electronics. IEEE

Industrial Electronics Magazin , 6-17.

Campbell, J. (2005). Social Movements and Organizations Theory (Vol. Where do we stand? Common mechanisms in organizations and social movements research.). New York:

40 Chapple, W., & Moon, J. (2005). Corporate social responsibility (CSR) in Asia: A seven-country study of CSR web site reporting. Busness and Society , 44 (4), 415-441.

Chatterji, A., & Toffel, M. (2010). How firms respond to being rated. Strategic Management

Journal , 31 (9), 917-945.

den Hond, F., & de Bakker, F. (2007). Ideologically motivated activism: how activist groups influence corporate social change activities. Academy of Management Review , 32, 901-924.

Doh, J., & Guay, T. (2006). Corporate social responsibility, public policy, and NGO activism in Europe an dthe United States: an institutional-stakeholder perspective. Journal of

Management Studies , 43, 47-73.

Eesley, C., & Lenox, M. (2006). Firm reponses to secondary stakeholder action. Strategic

Management Journal , 27, 765-781.

Esty, D., Levy, M., Srebotnjak, T., & de Sherbinin, A. (2005). Environmental Sustainability

Index: Benchmarking National Environmental Stewardship. New Haven, CT.: Yale Center

for Environmental Law and Policy.

Etzion, D. (2007). Research on Organizations and the Natural Environment, 1992-Present: A Review. Journal of Management , 33 (4), 637-664.

Fare, R., Grosskopf, S., & Hernandez-Sancho, F. (2004). Environmental performance: an index number approach. Elsevier , 26, 343-352.

Forsch, R., & Gallopoulos, N. (1989). Strategies for manufacturing. Scientific American , 144-152.

George, D., & Mallery, M. (2010). SPSS for Windows Step by Step: A Simple Guide and

41 Goldman, M., & Schurman, R. (2000). Closing the "great devide": New social theory on society and nature. Annual Review of Sociology , 26, 563-584.

Greve, M. (1989). Environmentalism and bounty hunting. Public Interest fall , 15-19.

Hart, S. L., & Ahuja, G. (1996). Does it pay to be green? An emperical examination of the relationship between emission reduction and firm performance. Business Strategy and the

Environment , 5, 30-37.

Hart, S., & Ahuja, G. (1996). Does it pay to be Green? An emperical examiniation of the relationshiop between emission reduction and firm performance. Business Strategy and the

Environment (5), 30-37.

Hoffman, A., & Ocasio, W. (2001). Not all events are attended equally: Toward a middle-range theory of industry attention to external events. Organization Science , 12 (4), 414-434.

Holzer, B. (2008). Turning stakeseekers into stakeholders a political coalition perspective on the politics of stakeholders influence. Business and Society , 47, 643-651.

Ioannou, I., & Serafeim, G. (2012). What drives coporate social performance? The role of nation-level institutions. Journal of International Business Studies , 43, 834-864.

Jacobsson, S., & Bergek, A. (2004). Transforming the energy sector: the evolution of

technological systems in renewable energy technology. Industrial and Corporate Change , 13 (5), 815-849.

King, B. (2008). A political mediation model of corporate response to social movement activism. Admin. Sci, Quart. , 53 (3), 395-421.

Laan, T., Beaton, C., & Presta, B. (2010). Strategies for reforming fossil-fuel subsidies: practical lessons form Ghana, France and Senegal. Global Subsidies Initiative , 3-40.

42

Lijst emissiefactoren. (2016, 05). Opgeroepen op 06 23, 2016, van CO2Emissiefactoren:

http://co2emissiefactoren.nl/lijst-emissiefactoren/

Maignan, I., & Ralston, D. (2002). Corporate Social Responsibility in Europe and the US; Insights from businesses' self-presentations. Journal of International Business Studies , 33 (3), 497-515.

Porter, M., & van der Linde, C. (1995). Toward a new conception of the environment-competitiveness relationship. The journal of economic perspectives , 9 (4), 97-118.

Post, C., Rahman, N., & McQuillen, C. (2014). From board composition to corporate to corporate environmental performance through sustainabiity-themed alliances. Journal of

Business Ethics , 423-435.

Prasad, P., & Elmer, M. (2005). In the name of the practical: Unearthing the hegemony of pragmatics in the discourse of environmental managment. Journal of Management Studies ,

42 (4), 845-876.

Rashchupkina, Y. (2015). The European Union's role in networks on removal of fossil fuel subsidies and disclosure of climate change information. Contemporary politics , 23 (3), 354-366.

Rooney, C. (1993). Economics of pollution prevention: how waste reduction pays. Pollution

Prevention Review , 261-276.

Scherer, A., & Palazzo, G. (2007). Toward a political conception of corporate responsibility. Business and society from a Habermasian perspective. Academy of Management Review , 32, 1096-1120.

43 Scherer, G., & Palazzo, G. (2011). The new political role of business in a globalized world: A review of a new perspective on CSR and its implications for the firm, governance and

democracy. Journal of Management Studies , 48, 899-931.

Scherer, G., & Palazzo, G. (2009). The Oxford Handbook of Corporate Social Responsibility (Vol. Globalization and corporate social responsibility). Oxford: University Press.

Sen, S., & Bhattacharya, C. (2001). Does doing good always lead to doing better? Consumer reactions to corporate social responsibility. Journal of Marketing Research , 38 (2), 225-243.

Sharma, S., & Henriques, I. (2005). Stakeholder influences on sustainability practices in the Cnadian forest products industry. Strategic Management Journal , 26, 159-180.

Walls, J., Phan, P., & Berrone, P. (2011). Measuring environmental strategy: construct development, reliability and validity. Business & Society , 50 (1), 303-319.

Wheeler, D., Fabig, H., & Boele, R. (2002). Paradoxes and dilemmas for stakeholders responsive firms in the extractive sector: Lessons from the case of Shell and the Ogoni.

Journal of Business Ethics , 39, 297-318.

Whiteman, G., Walker, B., & Parego, P. (2012). Planteraty boundaries: Ecological

foundations for corporate sustainability. Journal of Management Studies , 50 (2), 307-336.

Whitley, R. (1999). Divergent capitalisms: The social structuring and change of business

44

Appendices

Appendix 1: Theoretical model

Appendix 2: Normality analysis Figure 1. Normality Analysis - Descriptives

Variable Descriptives Statistic Std. Error

Log10 Environmental

Performance Mean 1.01 0.027

95% Confidence

Interval for Mean Lower Bound 0.96 Upper Bound 1.07 5% Trimmed Mean 1.01 Median 0.96 Variance 0.34 Std. Deviation 0.59 Minimum 0 Maximum 1.94 Range 1.94 Interquartile Range 1.07 Skewness 0.143 0.11 Kurtosis -1.30 0.23

Air Pollution Mean 11.61 0.25

95% Confidence

Interval for Mean Lower Bound 11.12 Upper Bound 12.09

5% Trimmed Mean 11.79

Median 11.8

45 Std. Deviation 5.32 Minimum 1.8 Maximum 39.13 Range 37.33 Interquartile Range 9.62 Skewness -0.03 0.11 Kurtosis -0.03 0.23 Energy Mean 1.08 0.03 95% Confidence

Interval for Mean Lower Bound 1.02 Upper Bound 1.14 5% Trimmed Mean 1.02 Median 1 Variance 0.43 Std. Deviation 0.65 Minimum 0 Maximum 4 Range 4 Interquartile Range 0 Skewness 1.97 0.11 Kurtosis 4.56 0.23

Sqrt Press Freedom Mean 6.97 0.12

95% Confidence

Interval for Mean Lower Bound 6.74 Upper Bound 7.19 5% Trimmed Mean 6.88 Median 6.78 Variance 6.19 Std. Deviation 2.49 Minimum 1 Maximum 13.23 Range 12.23 Interquartile Range 1.39 Skewness 0.809 0.11 Kurtosis 1.21 0.23

Log NGO Density Mean -1.72 0.05

95% Confidence

Interval for Mean Lower Bound -1.82 Upper Bound -1.62 5% Trimmed Mean -1.7 Median 1.66 Variance 1.11 Std. Deviation 1.02 Minimum -4.61 Maximum 0.58 Range 5.18 Interquartile Range 0.98 Skewness -0.62 0.11

46

Kurtosis 0.44 0.23

Log Firm size Mean 11.13 0.066

95% Confidence

Interval for Mean Lower Bound 10.99 Upper Bound 11.26 5% Trimmed Mean 11.09 Median 10.84 Variance 2.05 Std. Deviation 1.43 Minimum 7.72 Maximum 15.01 Range 7.29 Interquartile Range 1.6 Skewness 0.541 0.11 Kurtosis 0.22 0.23

Log Return on Assets Mean 0.06 0

95% Confidence

Interval for Mean Lower Bound 0.05 Upper Bound 0.07 5% Trimmed Mean 0.05 Median 0.05 Variance 0 Std. Deviation 0.06 Minimum -0.03 Maximum 0.46 Range 0.49 Interquartile Range 0.07 Skewness 1.96 0.11 Kurtosis 6.96 0.23

Figure 2. Normality Analysis - Kolmogorov-Smirnov Statistics

Kolmogorov-Smirnov Shapiro-Wilk

Variable Statistic df Sig. Statistic df Sig.

Environmental

performance 0.085 458 0.000 0.934 458 0.000

Air pollution 0.277 458 0.000 0.823 458 0.000

Energy 0.462 458 0.000 0.536 458 0.000

Sqrt Press freedom 0.254 458 0.000 0.847 458 0.000

Log NGO density 0.234 458 0.000 0.88 458 0.000

Log Firm size 0.089 458 0.000 0.971 458 0.000

Log Return on assets 0.143 458 0.000 0.843 458 0.000

47 Graph 2. Q-Q plot for Log Environmental performance

48 Graph 3. Boxplot for Log Environmental performance

49 Graph 5. Q-Q plot for Air pollution

50 Graph 7. Histogram for Energy

51 Graph 9. Boxplot for Energy

52 Graph 11. Q-Q plot for Sqrt Press freedom