The Impact of Corporate Environmental Performance on Financial Performance; an Image-based Approach

(1)

The Impact of Corporate Environmental Performance on Financial Performance;

an Image-based Approach

Elsemieke Debora Veenema

Abstract

This thesis examines whether the environmental performance of firms impacts the financial performance of firms. Assumed is that environmental measures of businesses in current society create returns by improving the image and credibility rather than improving resources. Since this view is more social in nature, this research will examine whether culture and perception play a role by comparing two continents, and whether communication and customer contact are of influence by comparing the impact between consumer-oriented industries and industries which are less consumer-oriented. This research, performed over different groups of firms from 2009 to 2014, finds weakly significant evidence of a negative relation between corporate environmental and financial performance, contradicting the image-based argument.

Field Key Words: environmental performance, financial performance, United States, Europe, consumer-oriented industries, image-based view, ASSET4

(2)

I. Introduction

On the 12th_{of December 2015, after two weeks of negotiations, 195 countries signed the United}

Nations-climate agreement to counteract the harmful effects of climate change during the COP21 in Paris. It is a historical treaty, since it is the first time in history that so many countries have jointly signed a climate agreement. This agreement is not only an endeavour, like many previous agreements, but it is legally binding for all countries (van Dijk, 2015). One important aspect of the agreement is the role given to businesses (Forbes, 2016). Together with the involved governments, they have an important share in combatting the effects of global warming. Already hundreds of companies have demonstrated leadership, such as making the decision to switch to 100% renewable energy. Moreover it is to be expected that this will become the norm for numerous companies. The consequences of the agreement go beyond the measures in the field of energy that companies have to take, there will be a shift in among others, the banking sector, investment world, stock exchanges and research. Therefore, this agreement proves that the environmental issues for businesses are now more important than ever before. Since the focus of the world’s citizens shifts towards the environmental policies and performances of companies and since the development of worldwide information- and communication technology, companies have to be very cautious when communicating their environmental pursuits or failures.

Environmental communication might become a part of business strategy. An example of "fatal" effects of negative environmental news on a company is well-renowned and publicly traded company Volkswagen Aktiengeschellschaft, hereafter referred to as Volkswagen. In September of 2015, it was made public that Volkswagen manipulated their software when assigning energy labels to their cars. This resulted in energy-efficient labels, while in fact the cars were polluting and not suitable for this category (Jessayan, 2015). When the news broke, of course consumers, suppliers, employees and other stakeholders all were shocked. Though, reactions of the stock market did the most, short-term, harm to Volkswagen. In not more than four days, their shares nosedived with approximately 37,5% from 169,65 to 106 euro (Yahoo!Finance, 2015). The conclusion that can be drawn from this, is that environmental performances of firms are not only important for gaining competitive advantages through innovation and adjusting resources but also through image, reputation and trust. This image-based approach will be the main argument in this research.

(4)

21st_{century. Furthermore, there has been research in the past about the effects on firm}

performance generated by environmental actions. The most important theory in those studies is the theory of Porter (1991), stating that environmental regulation can lead to competitive advantages, since it requires innovation. This argument for gaining financial returns can be defined as a resource-based perspective, since it is focused on an optimal use of resources. In the past 24 years, however, a lot has happened in the field of information and communication technology and economic globalization, which has caused a shift of perception on the environmental issue. The conception in this research is that in current society firms will gain structural profits through environmental performance because environmental concerns will lead to improving the image, strengthening the brand, gaining trust and improving the firm’s credibility. This will lead to more long-term commitments of consumers, suppliers, employees and shareholders, an argument supported by Paul Polman, Chief Executive Officer of Unilever N.V. (Forbes, 2015). This conception on environmental performance is new in the field of financial studies. Because the image-based approach, in contrast to the research-based approach, takes into account the more social and behavioural aspects, this research wants to analyse whether continental differences, such as culture and perception, and industry differences, such as strong or little branding and communication, will influence the effects of environmental performance. A combination of these elements leads to the main research question answered in this study;

Will a focus on corporate environmental performance have social consequences that lead to better corporate financial performance?

(5)

II. Literature review

A. Corporate Social Responsibility and Corporate Environmental Responsibility

Introducing the concept of Corporate Environmental Responsibility in this research, it is necessary to first understand the meaning of Corporate Social Responsibility (CSR). Many academics have attempted to define the concept of CSR, this however has proven to be somewhat of a challenge. In the literature focusing on corporate social responsibility, not one universally accepted definition is used. Dahlsrud (2006) compared 37 different definitions, including the dimensions they contain, and concluded that it is not as important to define CSR, as it is to understand how CSR is socially constructed within a context and taken into account when business strategies are developed. The most counted definition, and therefore the most commonly used one, is the definition used by the Commission of the European Communities, that reads as follows: “A concept whereby companies integrate social and environmental concerns in their business operations and in their interaction with their stakeholders on a voluntary basis” (Dahlsrud, 2006, p7). Determining CSR is most complete when the description contains the dimensions Voluntariness, Stakeholder, Social, Environmental and Economic (Dahlsrud, 2006). Here, Voluntariness indicates that businesses should perform their social responsible activities above regulatory requirements. With Stakeholders is meant any group or individual who can affect or is affected by the achievement of an organization's objectives, such as suppliers, customers, employees or communities. In CSR, a trusting and cooperative relationship between a firm and its stakeholders is essential. The three dimensions left are often used to define the nature of CSR-actions (Uddin, 2008). With the Social dimension is meant the relationship between business and society. This implies not only the relationships with stakeholders, but also preventing or ameliorating the social problems such as poverty and hunger. With the Economic aspect is meant the direct and indirect economic impacts that the organization’s operations have on the surrounding community and on the company’s stakeholders (Uddin, 2008). Examples of this can be; seeing taxes not as a cost that needs to be avoided but rather as a part of their social contract with society, or corporations being financially accountable (e.g. transparency in their reports). The last aspect is the Environment, which focuses on the long-term perspective on nature and environment, dealing with issues such as global warming and depletion of natural resources. As indicated previously, this last aspect will be the focus in this thesis.

(6)

This study focuses on the environmental nature of actions that businesses take, which henceforth will be called Corporate Environmental Responsibility (CER). Examples of environmental actions by businesses can be reducing emission of harmful gases or recycling of waste. The extent to which companies undertake these environmentally oriented actions is referred to as Corporate Environmental Performance (CEP). A high degree of environmental performance is designated by environmental-friendliness.

This research examines the impact of Corporate Environmental Performance on Corporate Financial Performance. With a firm’s financial performance is meant a measurement to determine the results of a firm’s policies and operations in monetary terms over a given period of time. In other words, with financial performance is meant to what extent a company can generate revenues using their assets. Corporate Financial Performance can be expressed in accounting variables such as return on assets and return on equity or market-based variables such as stock returns. All those variables imply a firm’s financial performance.

B. The rise of Corporate Environmental Responsibility - Classical view

After the world wars in the 20th century environmental concerns rose considerably. Especially the

concerns about air pollution as a consequence of the Industrial Revolution increased after some serious smog incidents in England and the United States (American Meteorological Society, 1999). With this rise of environmental concerns in the mid-20th_{century, the volume of environmental}

(7)

regulations comes from the situation where the power of innovations in technology to change all assumptions is not included. In a static model, where all factors remain the same, environmental regulations may indeed lead to higher costs. However, Porter and van der Linde (1995) said that innovation is a dynamic process, shifting all factors, and thereby changing environmental regulation to competitive advantages is possible. They stated that the orientation of businesses should shift from pollution control to resource productivity, a so-called resource-based view, and that success must involve innovation-based solutions that promote both environmentalism and industrial competitiveness. Porter’s resource-based view is supported and reinforced by Hart’s (1995) natural-resource-based perspective. This view is based on the idea that businesses in the future will be constrained by and dependent on nature. Hart (1995) supposed that a firm’s strategy and competitive advantage is rooted in capabilities that facilitate environmentally sustainable economic activity.

Summarizing, the resource-based perspective of Porter (1991) shows that environmental actions taken by companies, could lead to productivity improvement, resource-efficiency and innovation. These resource-based improvements lead to competitive advantages, which in turn lead to increased financial returns. In this research, this view of Porter is considered as the classical view on gaining returns through environmentally friendly actions.

C. Corporate Environmental Responsibility - Renewed view

(8)

opportunity since it means that the actions of an organization have a significant influence on individuals. This means that companies have to take environmental action and communicate about this in order to strengthen their credibility, integrity and societal consciousness, a so-called image-based approach.

According to Fombrun and Shanley (1990), people’s perceptions of a firm’s concern for society may influence their judgement of the firm. The social responsiveness of a firm demonstrates that it achieved beneficial relationships with different kinds of stakeholders. Because social responsibility will increase the goodwill of employees, consumers and other parties that enhance long-term profitability, a good reputation is to be expected. Fombrun and Shanley (1990) found a positive, significant result when testing whether a firm’s contribution to social welfare would positively influence a firm’s reputation. Eberle et al. (2013) in addition found that CSR can improve a firm’s reputation by communicating about it. Communicating interactively about CSR improves message credibility, and thereby company credibility. Moreover, communicating interactively about CSR makes stakeholders identify better with the company because it establishes a relationship between them. This feeling of identification has a positive effect on the company’s reputation and causes stakeholders to engage in positive word-of-mouth behaviour about the company. This result implies that using interactive channels to communicate about CSR can improve corporate reputation (Eberle et al., 2013).

The question then remains how this improved reputation can generate financial returns. Freeman’s (1984) stakeholder theory already demonstrated that corporate social performance of a company leads to better financial performance. He states that the enhanced reputation results in employees spending more energy on their work, in creditors charging less interests, in suppliers requiring less processing charge and in customers remaining loyal when companies are faced with difficult times. Waddock and Graves (1997) support these arguments by stating that corporate social performance increases morale, productivity and satisfaction of employees and collaboration of local government. All these effects are expected to yield better financial performances. Waddock and Graves (1997) found scientific evidence that these expectations are in fact reality.

(9)

increasingly base their purchases on a firm’s environmental performance and avoid or reduce using products that generate negative environmental impact.

Orlitzky (2003) tested, by means of a meta-analysis, the mediating role of reputation in the relationship between corporate social responsibility and corporate financial responsibility but found that, in contradiction to the literature, CSR disclosure has a low reputational impact on the financial performance. However, the studies analysed by Orlitzky (2003) supported the image-based approach relatively more strongly than the resource-image-based approach.

Summarizing, this renewed view on Corporate Environmental Responsibility, referred to as the image-based approach, states that a firm that expresses environmental concerns and behaves environmentally friendly improves firm reputation and stakeholder confidence. This improved reputation leads to increased sales and stronger long-term relationships with stakeholders, which in turn lead to better financial performance.

Whether or not this image-based view is matter at issue in practice, generally, all previous arguments indicate higher financial returns for firms engaging in environmental issues and acting environmentally responsible.

Hypothesis 1:

High corporate environmental performance has a positive influence on a firm’s financial performance in the long-run.

(10)

D. Continental differences

The impact that the environmental policy of a business can have on its financial performance may differ geographically, due to differences in view on the subject. The prevailing view on the environmental topic within a country or continent can be influenced by the government, the businesses or the consumers. In this research the impact of environmental performance in the United States (US) will be compared with the impact in the European Union (EU). The comparison between the United States and the European Union is made because both continents are seen as global environmental leaders (Kelemen and Vogel, 2009). Since companies in both continents are strongly involved in the environmental issue and to a great extent act on their environmental responsibility, the strongest impact of environmental performance is expected in comparison to the rest of the world. Therefore, these continents are of greatest interest for this research.

When analysing the development of the environmental issue in the United States for the past 25 years, striking is that the focus on the environmental issue was lost for a few decades (Kelemen and Vogel, 2009). Despite early laws and regulations on environmental issues, the US stagnated environmental development the past 25 years and lost their leadership role (Kelemen and Vogel, 2009). Concerns about the climate seemed to be re-increased only recently. While the US signed the Kyoto protocol on reducing greenhouse gas emissions in 1997, no US president submitted it to the Senate for ratification. President Bush, who held office between 2001 and 2009, declared he would not support the Kyoto Protocol and refused to propose any regulations for carbon dioxide emissions (Vogel et al., 2010). In Europe, the environmental actions occurred gradually and actively since the rise of environmental regulations during the 20th_{century. In 1989,}

the political dynamics of the international environmental policy have shifted and the European Union became the new global environmental leader (Kelemen and Vogel, 2009). Unlike in the US, the Kyoto protocol changed the climate issue in the European Union fundamentally (Arnold, 2000). The implementation of regulations started at the national level, with the use of taxes as an incentive (Arnold, 2000; Vogel et al., 2010), but shifted towards the European level from the Single European Act in 1987 (Vogel et al., 2010). The European Commission was originally solely motivated to prevent the divergence of national regulation, but later on the environmental issue became a theme of the European Commission in its own right.

(11)

Li, 2014). Since scholars, businesses and households are all linked together, European companies are more inclined to realize personal needs, provide high welfare and emphasize personal satisfaction and quality of life (Hou and Li, 2014).

Since the topic of climate change has only been supported again by the US government since a few years, and businesses mainly act responsibly because of regulations, expectations are that it is still too little considered an influential factor for consumers, society and businesses. Furthermore, since the environmental issue is an important focus of the European Union, European countries see the climate as a scale problem where the whole community is involved. Hypothesis 2:

European companies gain higher returns by acting in a corporate environmentally responsible way compared to US companies.

E. Industrial differences

According to the image-based approach, society’s perception on a firm is essential when wanting environmental actions to yield financial returns. Consumers increasingly expect companies to care about and be involved in society and the environmental issues (Margolis and Walsh, 2003). These expectations induce companies to have significant influence on individuals (Margolis, 2001). Hence firms can gain competitive advantages by letting consumers know that they are concerned with environmental issues and produce environmentally friendly and fair trade products (Prahalad and Hamel, 1994). Communication is an essential tool for firms to influence a customer’s perception of the company. Communicating interactively about a firm’s social responsibility, through advertisement, leads to more credibility and a stronger feeling of identification with the brand and therefore an improved reputation (Eberle et al., 2013; Fombrun and Shanley, 1990). Since consumers expect more of companies and brands, they adjust their buying behaviour to it (Leonidou et al., 2010). Webb, Mohr and Harris (2007) found that responsible consumption is increasing; consumers increasingly base their purchases on a firm’s environmental performance and avoid or reduce using products that generate negative environmental impact.

(12)

yielding returns, the difference in impact of environmental actions is examined between consumer-oriented industries and industries where consumers are not considered as important stakeholders. A good example of a strong consumer-oriented industry is the Fast Moving Consumer Goods industry. Fast Moving Consumer Goods (FMCG) are non-durable goods that are sold quickly. The products have a short life, either because of high consumer demand and rapid developments or because the product is perishable. The FMCG-industry has the largest market share in advertising expenditures, it accounts for more than 25% of all advertisement spending and this share continues to grow (Nielsen, 2013). As an industry that does not heavily invest in consumer relations, the construction industry is tested. The construction industry comprises of activities involving construction, alteration and repair of, for example, buildings, roads and bridges (Occupational Safety and Health Administration, 2002).

Concluding, it can be said that the Fast Moving Consumer Goods industry will benefit more from an environmental policy than the Construction industry, since the industry's main focus is on consumers and it spends generally more on advertising.

Hypothesis 3:

(13)

III. Data and Methodology

A. Data collection

In order to test the hypotheses that have been set in the literature, firm-specific information is needed of firms worldwide, split into continents and industries.

A.1 Companies

The data used in the regressions will be extracted from the ASSET4 database from Datastream by Thomson Reuters. Although Thomson Reuters acquired the ASSET4 database in 2009, ASSET4 comprises environmental, social and governmental (ESG) information since 2003 (Thomson Reuters, 2013). It currently contains ESG information for more than 4,600 publicly listed firms worldwide. Ratings are derived by company comparisons for a total of 226 Key Performance Indicators (KPI). The 226 KPIs are derived from over 500 separate data points to facilitate accurate and transparent ESG screening. The indicators are divided into three pillars; the environmental, social and governmental pillar. In this study only the Environmental pillar of ASSET4 is taken into account. Environmental Ratings are derived from a total of 70 KPIs (Thomson Reuters, 2013). Although opinions about the reliability of the ASSET4 ESG database differ (van den Heuvel, 2012; Chatterji et al., 2014; Rees and Mackenzie, 2011), it is an interesting database to use, since limited CSR and CER research is done based on this data. Data will be used from 2009 until the last available data in 2014, therefore six years will be captured in this research.

As in this research three different hypotheses will be tested, three different datasets will be used. For the first hypothesis, worldwide information is used. Data is extracted from constituent list Global LA4RGNGL in ASSET4. It consists of 4,248 companies over six years, so 25,488 observations, Table 1. Because of omitting variables and errors, eventually 20,210 observations of 4,206 firms are included in the regression. The descriptive statistics of the dataset are shown in Table 2.

For the second hypothesis, the database is built up from two constituent lists, all US companies (LA4CTYUS) and all European companies (LA4RGNEU). Datastream bases the origin of a firm on the location of its headquarters. The American list consists of 999 firms of which five firms became delisted between 2009 and 2014, which leaves 994 eligible firms. The European list consists of 976 firms of which two firms became delisted, stopped existing, leaving 974 eligible firms. The total amount of observations is 11,808, 1,968 firms over six years, but due to missing values, 10,539 observations are used in the research (Table 1). Descriptive statistics of the dataset can be seen in Table 3.

(14)

associated Standard Industrial Classification (SIC) codes (Appendix, Table AI). There are 666 FMCG-companies included in the ASSET4 database and 163 firms from the construction sector, which leaves a total of 829 firms and 4,974 observations. Without missing values, 3,917 observations remain. The descriptive statistics are shown in Table 4.

Table 1. Number of firms used in this research.

This table shows how many firms are used for performing the regressions and up to how many observations this leads.

Firms Over six years Used observations

H1 World 4,248 25,488 20,210

H2 Continents 1,968 11,808 10,539

H3 Industries 829 4,974 3,917

A.2 Variables

Of the companies included in the ASSET4 database, various company-specific variables are needed to generate reliable outcomes and to answer the hypotheses.

A.2.1. Dependent variable; financial performance

As explained in the literature section, Corporate Financial Performance is a measurement to determine the results of a firm’s policies and operations in monetary terms over a given period of time. There are different variables used in research for determining financial performance. Generally, accounting variables such as return on assets, return on equity or return on investments are used for researching past performance, which implies backward-looking. It is used to capture a firm’s relative efficiency of asset utilization (Cochran and Wood, 1984) and consequently reflect internal decision-making capabilities and managerial performance (Cochran and Wood, 1984; Orlitzky, Schmidt and Rynes, 2003). Another way of considering firm performance is looking at the external market responses to organizational actions (Edmans, 2011). Stock market returns can be observed or the market value of a firm can be compared with their book value to indicate whether growth is expected by the investors. This is a more future-based approach since it represents investors’ evaluation of the ability of a firm to generate future economic earnings (Cavaco and Crifo, 2014).

(15)

A.2.2. Independent variable; environmental performance

Corporate Environmental Performance was defined in the literature as the extent to which companies undertake environmentally oriented actions. In this research, the Environmental score from the ASSET4 database will be used to quantify corporate environmental performance. The Environmental Pillar Score (ENVSCORE) is composed of 70 different environmental measurements, such as resource usage and reduction, emissions and emissions reductions, environmental activism and product or process innovation (Thomson Reuters, 2013). A business can be allocated a score between zero and hundred, whereby zero means incredibly bad environmental performances and hundred means acting perfectly towards the environment. In this research the environmental score is lagged by one year, EnvScorei,t-1, since lagging the

independent variable tackles the problem of reversed causality (Waddock and Graves, 1997) and environmental measures need time to generate financial returns. Considering the image-based view in this research as source of financial returns, it can be assumed that it takes time for firms to create competitive advantage through an enhanced reputation and credibility. Lagging the responsibility score is in line with existing CSR and CER literature (Oh and Park, 2015; Surroca, Tribo and Waddock, 2010; Russo and Fouts, 1997; Jo, Kim and Park, 2013).

A.2.3. Control variables

For the determination of control variables, numerous comparable CSR and CER articles are consulted, concluding that controlling for Size, Risk, Industry, Research and Development expenditures and Advertising intensity is especially relevant.

Size

Most research controls for size because size is recognized as a determinant of social and financial performance (Ullman, 1985). Waddock and Graves (1997) state that taking size into account is important because smaller firms may not exhibit as many overt socially responsible behaviours as larger firms do. Size is usually added in the regression as total assets, total sales or number of employees (Cavaco and Crifo, 2014; Surroca, Tribo and Waddock, 2010). In this thesis, Size will be expressed in total assets and net sales. Since total assets and net sales are expressed in absolute numbers, in this research for all firms the natural logarithm of the USD is used.

Leverage

(16)

be seen as the internal risk (Schwartz, 1959). This thesis calculates Leverage by the debt to asset ratio.

Industry

According to Ullman (1985), an industry is likely to act as intervening variables in the relationship between CSR and financial performances. The extent to which firms act responsibly depends on the industry they operate in. Just like with size, stakeholders rely on conspicuous industries and firms, given their limited amount of information available to them (Ullman, 1985). Controlling for industry makes sure that industry-level factors that have been shown to explain variation in firm performance, are included in the research (McWilliams and Siegel, 2000). In this research, controlling for Industries is done by inserting a dummy variable based on the four-digit Standard Industrial Classification (SIC) code.

R&D expenditures and Advertising expenditures

McWilliams and Siegel (2000) state that controlling for investment in Research and Development is necessary when drawing a conclusion about the impact of Corporate Social Responsibility, since R&D investments have shown to be an important determinant of firm performance. Moreover, there is a long standing theoretical link in the literature between R&D investments and economic performance and taking R&D into account can give some insides to the underlying argument for the impact of CSR (McWilliams and Siegel, 2000). A significant influence of R&D expenditures on firm performance would implicitly support Porter’s (1991) resource-based argument. Since this research focuses on the image-based instead of the resource-based argument, adopting the Advertising expenditures would be of great interest to examine the role of branding and communication in the impact of Corporate Environmental Responsibility. However, after collecting the data from ASSET4, it soon became clear that the database lacks information about R&D expenditures and advertising expenditures, whereby controlling for these variables becomes impossible.

(17)

Table 2. Descriptive statistics and correlations of dataset 1.

Characteristics of the variables that will be used in the regression are given in the column Mean and Standard Deviation. Next to the characteristics, the correlation matrix shows. The N stands for the number of observations.

N=20,210

Mean _DeviationStandard Return on _assets Environmental _Score _AssetsTotal _SalesNet Debt to _Assets

Return on assets 5.606 10.961 x

Environmental Score 49.941 32.052 -0.001 x

Total Assets 15.679 1.700 -0.080 0.388 x

Net Sales 14.937 1.767 0.075 0.491 0.771 x

Debt to Assets 24.776 19.376 -0.085 0.034 0.120 0.076 x

Characteristics of the variables that will be used for testing hypothesis 2 are given in the column Mean and Standard Deviation. Next to the characteristics, the correlation matrix shows. The N stands for the number of observations.

N=10,539

Mean _DeviationStandard Return on _assets Environmental _score _AssetsTotal _SalesNet Debt to _Assets

Total Assets 15.915 1.620 -0.170 0.356 x

Net Sales 15.240 1.520 0.006 0.472 0.763 x

Debt to Assets 26.072 20.294 -0.099 0.004 0.055 -0.025 x

Characteristics of the variables that will be used for testing hypothesis 3 are given in the column Mean and Standard Deviation. Next to the characteristics, the correlation matrix shows. The N stands for the number of observations.

N=3,917

Mean _DeviationStandard Return on _assets Environmental _Score _AssetsTotal _SalesNet Debt to _Assets

Total Assets 15.488 1.376 -0.003 0.436 x

Net Sales 15.247 1.563 0.070 0.468 0.854 x

(18)

B. Methodology

When measuring the impact of environmental performance on financial performance, one can choose between measuring short-term or long-term effects. Measuring short-term effects can be done by an event-study (Flammer, 2013; Krüger, 2015; McWilliams, Siegel and Teoh, 1999), measuring long-term effects can be done by regression (Ullman, 1985). As mentioned before, this study wants to test the long-term effects of environmental performance and therefore will use a regression. Using a regression is very common when testing for long-term effects of corporate social responsibility or environmental responsibility (Cochran and Wood, 1984; Waddock and Graves, 1997; Russo and Fouts, 1997; McWilliams and Siegel, 2000; Surroca, Tribo and Waddock, 2010; Jo, Kim and Park, 2013; Oh and Park, 2015). Since this research uses panel data, three multiple regression analyses will be run. Eviews will be used to perform the regressions.

Hypothesis 1:

As discussed in the data collection section, the composed database lacks information about the R&D and advertising expenditures of the firms, so these variables will not be included in the regression. In addition, before using all available variables, there must be considered whether multicollinearity is an issue. Multicollinearity is the situation wherein the explanatory variables correlate too much [>50%]. It can lead to high standard errors for the individual coefficients (Brooks, 2008) which means the regression becomes very sensitive to small changes in the specification and wrong assumptions are quickly drawn. The correlation matrix of Table 2 shows that Net Sales correlate highly with Total Assets [0.771], which is not entirely surprising, since they both indicate the size of the firms, so Net Sales will be excluded and taken out of the dataset.

Implementing the remaining variables in Eviews without adjustments will give a pooled ordinary least squares (OLS) regression on all observations together. Since this research uses data of multiple firms over multiple years, there is bound to be heterogeneity in these units (Brooks, 2008). Consequence can be that the standard errors are inappropriate, they can be either too large or too small, and therefore, wrong inferences can be made. To test how the panel data should appropriately be implemented in Eviews, the Redundant Fixed Effects test and Hausman test are performed (Appendix, Table BI - BIII). The use of both cross-section fixed effects and period fixed effects prove to be applicable. Since controlling for industries in this research is necessary, cross-section fixed effects will be applied on industries using 4-digit SIC codes, and period fixed effects will be applied on the years.

The regression for hypothesis 1 will look like;

(19)

Hypothesis 2:

For hypothesis 2, two regressions need to be run; the effects in Europe and in the United States. Table 3 shows that the variables Total Assets and Net Sales correlate too strong [0,763] so Net Sales will be excluded from the regression to solve the problem of multicollinearity. To test for heterogeneity in this dataset and to select the appropriate regression technique, Redundant Fixed Effects test and the Hausman test will be performed. Again, a pooled regression is not appropriate and both industry and period fixed effects have to be applied (Appendix, Table CI-CVIII). This causes the regression of hypothesis 2 for both Europe and the United States to look the same as that of hypothesis 1;

ROAi,t = β0 + β1EnvScorei,t-1 + β2Sizei,t + β3Leverage i,t + IndustryEffects + YearEffects + ɛi,t (2)

To test whether the impact of corporate environmental performance in the US is different relative to the impact in Europe, a dummy for US firms is added, together with a variable as a product of US firms and the environmental impact. This use of continent-dummy makes it impossible to control for industry differences so only period fixed effects will be applied.

ROAi,t = β0 + β1EnvScorei,t-1 + β2Sizei,t + β3Leverage i,t + β4Continent + β5Continent*EnvScore i,t-1 +

YearEffects + ɛi,t (3)

Hypothesis 3:

Testing the differences between the impact in the FMCG-industry and Construction industry is related to testing differences between continents. Although this database contains less data than for the other two hypotheses, again a high correlation between Total Assets and Net Sales is found [0.864], leading to a third exclusion of Net Sales. This third hypothesis focuses on the differences between two industries, meaning that controlling for industries is irrelevant. However, the Hausman and Redundant Fixed Effects tests show that applying cross-section fixed effects is appropriate (Appendix, Table DI, DII, DV and DVI). Therefore this regression controls for firm-specific effects.

ROAi,t = β0 + β1EnvScorei,t-1 + β2Sizei,t + β3Leverage i,t + FirmEffects + YearEffects + ɛi,t (4)

Examining possible differences between the impact of corporate environmental performance in the FMCG-industry and in the Construction industry will be done in accordance with the continental comparison. The firm-specific effects are replaced by a dummy for the FMCG-industry and a variable of FMCG-firms times the environmental impact.

ROAi,t = β0 + β1EnvScorei,t-1 + β2Sizei,t + β3Leverage i,t + β4Industry + β5Industry*EnvScore i,t-1 +

(20)

IV. Results

The results of the regressions as they are presented in the Methodology section will be presented. First, the analyses for the environmental impact worldwide will be discussed. After this, the difference in impact between Europe and the United States will be reviewed, followed by a discussion of the impact in the fast moving consumer goods industry and the construction industry.

A. World

In running the regression analysis of environmental and financial performance as presented in Table 5, first the dataset of firms worldwide is used. In terms of control variables, Table 5 shows strong significant results for both Firm Size and Leverage, which means that the Return on Assets for a large extent can be explained by the size and capital structure of a firm. The correlation between Firm Size and ROA is positive, so in general, larger companies yield higher returns. The opposite is true for Leverage, indicating that a high debt-ratio generates worse financial performance. The one-year lagged independent variable Environmental Performance, the most interesting variable, shows a t-statistic significant at a 10% level, which means that the environmental performance of a firm is of weak significant influence on its return on assets. The statistic is negative, meaning that companies that to a great extent undertake environmentally oriented actions generate significant lower return on assets. This contradicts the introduced image-based theory as well as the resource-based theory of Porter. The image-based approach, introduced in this thesis, states that firms obtain competitive advantages by involving in environmental actions, since environmental friendliness improves social aspects such as reputation and identification. The resource-based theory of Porter (1991) states that environmental actions taken by companies lead to competitive advantages through productivity improvement, resource-efficiency and innovation. The weak significant negative correlation between environmental and financial performance contradicts both theories. It is particularly striking that Porter’s theory is rejected since this theory has been proven to be true in several different studies (Russo and Fouts, 1997; Porter, 1990). This negative correlation can be stated with any certainty since this research is based on a large amount of observations and the R2_{, the}

(21)

Table 5. Results of the global impact of corporate environmental performance on corporate financial performance.

Regression performed over 4,428 firms worldwide in the period 2009 to 2014. NT indicates the number of observations taken into the regression and R2 and adjusted R2 indicate the proportion of explained variance of the regression model.

Dependent variable: Return on Assets World

t-Statistic Independent variable NT 20,210 Environmental score (-1) -1.780* R2 0.576 Adjusted R2 0.471 Control variables Size 2.896*** Leverage -31.001*** Constant 0.088 *** significant at 1% level ** significant at 5% level * significant at 10% level

B. Continents - Europe and the United States

To draw a conclusion on hypothesis 2, regressed on solely firm in Europe and the US, Table 6 must be considered. Again, the control variables Size and Leverage seem of significant influence on the Return on Assets in both Europe and the United States. Looking at the one-year lagged independent variable Environmental performance, it shows a negative significant result for Europe (<5%) but no significant result for the US. The last column shows that the impact of the environmental performance in the US is significantly higher than the impact in Europe.

In this study, the impact of environmental performance in Europe and the United States is examined to test whether cultural aspects are of influence on the effects of environmental activity. The literature implies a different view on the environmental subject since the environmental issue in Europe has been much more discussed in the past 25 years than in the US. Moreover, in Europe, the environment is seen as a societal problem which encourage firms to take environmental measures above and beyond environmental regulatory requirements. In the US, regulations seem to be the main driver for responsible actions. The significant negative effect in Europe rejects the theory that environmental friendly companies gain competitive advantage through positive conception and rather implies that an environmental policy leads to costs and decreasing returns. For US firms, a firm’s environmental actions do not seem to have any correlation with its financial performance. This result can be considered likely since it is carried out on large scale and has a R2

(22)

the US with Europe [Continent*Environmental score] implies a significantly stronger positive impact in the US relative to the EU, which indicates significant cultural differences. This assumption, though, needs to be put in perspective. As described in the methodology section, to compare continents with each other, dummy variables are used and no cross-section fixed effects applied. This resulted in a R2_{of 5% which is drastically lower than the individual regressions.}

Concluding can be said that strong corporate environmental performance again does not lead to increased financial returns and that cultural differences regarding the environment are no influential factor.

Table 6. Regression output of Europe and the United States.

Regression performed over 974 European firms and 994 American firms in the period 2009 to 2014. NT indicates the number of observations taken into the regression and R2 and adjusted R2 indicate the proportion of explained variance of the regression model.

Dependent variable: Return on Assets Europe United States Both

t-Statistic t-Statistic t-Statistic

(23)

C. Industries – The Fast Moving Consumer Goods- and Construction Industry

The third regression of this research is performed over the Fast Moving Consumer Goods (FMCG) industry and the Construction industry, as can be seen in Table 7. In contrast to the other two hypotheses, this time the Size of the firm, tested by using the firm’s total assets, has no significant influence on the firm’s Return on Assets. Within these two industries, only the Leverage ratio seems to explain the firm’s financial performance. The independent variable Environmental performance shows neither a significant result for the worldwide fast moving consumer goods-industry and the Construction industry. This means that within these two industries, the environmental performance of a firm is of little influence on its operating income.

The difference in impact between the FMCG and construction industry is examined to find out whether the customer focus of a firm is of influence on the effects of environmental activity. Since the fast moving consumer goods industry has strong focus on customers and is the highest spender on advertising, building a strong reputation and relationship with the consumer is essential in gaining financial returns. This thesis presumed that disclosure of a company’s environmental friendliness would improve a firm’s reputation and credibility and thereby gain competitive advantages. Therefore, the impact of environmental performance was considered more positive for FMCG firms than for construction firms, who to lesser extent consider branding as competitive resource. The two insignificant effects of environmental performance imply that the environmental focus of a firm has no effect on the financial performance of a firm and that there are other sources that explain profitability within these industries. Especially for the FMCG-industry, this result can be considered plausible with an R2_{of 70%. This outcome is conflicting}

with the conception of Unilever’s Chief Executive Officer Paul Polman that environmental performance generates financial profits by expanding the long-term commitments with stakeholders, as appointed in the introduction. When examining the impact of environmental actions in the FMCG-industry relative to the impact in the construction industry [Industry*Environmental Score], a significant negative result is found. The statistic indicates that the environmental performance yields significantly lower returns in FMCG industry than in the construction industry. This contradicts the hypothesis, however, here too applies that cross-section fixed effects should have been added for reliable results. The model has an R2_{of 2%,}

(24)

Table 7. Regression output of the Fast Moving Consumer Goods- and Construction Industry

Regression performed over 666 FMCG firms and 163 Construction firms in the period 2009 to 2014.

Dependent variable: Return on Assets FMCG Construction Both

t-Statistic t-Statistic t-Statistic

Independent variable Environmental score (-1) -0.081 -0.890 -2.565** Control variables Size 1.147 0.706 2.734*** Leverage -12.002*** -8.016*** -3.674*** Industry 1.687* Industry*Environmental Score(-1) 1.406 Constant 0.105 -0.125 0.425 NT 3,178 739 3,810 R2 0.700 0.368 0.020 Adjusted R2 0.624 0.193 0.018 *** significant at 1% level ** significant at 5% level * significant at 10% level

Overall, based on the resource-based and image-based approach set out in this research, a positive impact of strong environmental performance on the financial performance of a company was hypothesized. Subsequently, the possible underlying source of improved returns was presumed to be image-based and this was tested by comparing continents and industries. Since already no evidence could be found for the first hypothesis, it was very unlikely that the source of competitive advantage would be found with the subsequent regressions. The weak negative significant effects of hypothesis 1 and 2 is not in line with existing literature. This outcome does, however, correspond with the inconclusive results found of the relationship between corporate social responsibility and financial performance (McWilliams and Siegel, 2000; Orlitzky et al., 2003). Despite specific environmental-arguments as those of Porter (1991), it seems like the same inconclusive results can be applied to environmental performance.

(25)

V. Conclusion and limitations

A. Conclusion

This thesis has examined whether a focus on corporate environmental performance has social consequences that lead to better corporate financial performance. Based on the resource-based theory of Porter (1991) and the image-based arguments of Margolis and Walsh (2003) and Prahalad and Hamel (1994), it was expected to find a positive relationship between the two. Conducting a worldwide research of 4,248 firms over six years, a significant negative impact of environmental performance was found. This contradicts both theories and existing literature, which may imply that the outcomes of research around Corporate Environmental Responsibility are in line with the inconclusive results found on the topic of Corporate Social Responsibly.

In this study, it was assumed that environmental performances of businesses in current society create returns through improving the image, reputation and credibility rather than obtaining competitive advantages by adjusting or improving resources. Since this view is more social and behavioural in nature, differences were expected between continents and industries. Since different continents imply different cultures and perceptions, the difference in impact of environmental performance on financial performance was tested between Europe and the United States of America. Because of different governments and different business ideas, it was hypothesized that the impact would be more positive for European firms. This research, performed over 974 European and 994 American firms from 2009 to 2014, found a significant negative impact of environmental responsibility on financial returns in Europe and no significant impact in the US, which rejects the expectations that culture affects the effect of environmental involvement. Last, to again test the impact of environmental performance but at the same time take into account aspects such as communication, branding and identification, the consumer-oriented Fast Moving Consumer Goods-industry was compared to the business-to-business Construction industry. The regression, performed over 666 FMCG-businesses and 163 Construction companies, does not point out any impact. This means that no evidence could be found for the third hypothesis and that also customer contact could not be proven to be an influential factor.

(26)

B. Limitations and future research

Some attention has to be given to the limitations of this research. First, the environmental performance of a firm in this research is indicated by the ASSET4 pillar score, which is measured as an aggregated score of 70 environmental indicators. ASSET4 assigns a score between 0 and 100 to all firms in the database. This score is based on the firm’s individual performance of various indicators, without taking into account the other firms in the database. Since the scores are not assigned relative to each other, it does not indicate whether a firm performs more or less environmentally friendly than the comparative firms. Moreover, using an aggregated score to express environmental performance means that environmental weaknesses can be substituted by environmental strengths, leaving the score not be completely objective.

Second, the global environmental concerns are a relatively new topic for businesses to cope with. Since the recent emergence of the environmental aspect within corporate social responsibility, the data around this subject is relatively minimum. The focus on collecting environmental data will enhance in the next decades, as in the ASSET4 database an upward trend can be observed over the years. This makes it possible to enlarge research and come to more specified conclusions. As discussed in the Data section, the resource and development expenditures and advertising expenditures should have been included in the regression to get a more complete understanding of the effects of environmental policy (McWilliams and Siegel, 2000). Since this research considers social consequences of environmental friendliness as the resource for competitive advantages, it would have been of great interest to see what the regression would say about the role of advertising. Adopting advertising in the model would give a better insight to the role of image-based arguments such as communication, branding and customer contact.

Adding data of R&D and advertising expenditures would not be enough to draw conclusions on social consequences of environmental policy. Since aspects like reputation and credibility are intangible assets (McWilliams and Siegel, 2000), their impact cannot be examined directly through a regression. To scientifically include the social effects of environmental policy, a different method should be applied.

(27)

References

American Meteorological Society, 1999. History of the Clean Air Act. [ONLINE] Available at: https://www.ametsoc.org/sloan/cleanair/. [Accessed 13 January 16].

Arnold, R. (2000). A changing climate for carbon services in the EU. Academic Journal Corporate Environmental Strategy, 7(3):298-303.

Brooks, C. (2008). Introductory Econometrics for Finance. 2nd ed. New York: Cambridge University Press: 1-674.

Cavaco, S. and Crifo, P. (2014). CSR and financial performance: complementarity between environmental, social and business behaviours. Applied Economics, 46(27): 3323– 3338. Chatterji, A., Durand, R., Levine, D. and Touboul, S. (2014). Do Ratings of Firms Converge?

Implications for Strategy. Working Paper Institute for Research on Labour and Employment.

Cochran, P.L. and R.A. Wood (1984). Corporate social responsibility and financial performance. Academy of Management Journal, 27: 42–56.

Dahlsrud, A. (2006). How corporate social responsibility is defined: an analysis of 37 definitions. Corporate Social Responsibility & Environmental Management, 15(1): 1-13.

Eberle, D., Berens, G. and Li, T. (2013). The impact of interactive corporate social responsibility communication on corporate reputation. Journal of Business Ethics, 118(4): 731-746. Edmans, A. (2011). Does the stock market fully value intangibles? Employee satisfaction and

equity prices. Journal of Financial Economics, 101: 621–640.

Environmental Protection Agency, 1985. EPA History (1970-1985). [ONLINE] Available at: http://www.epa.gov/aboutepa/epa-history-1970-1985. [Accessed 13 January 16]. Flammer, C. (2013) Corporate Social Responsibility and shareholder reaction: the environmental

awareness of investors. Academy of Management Journal, 56(3): 758–781.

Fombrun, C. and Shanley, M. (1990). What’s in a name? Reputation building and corporate strategy. Academy of Management Journal, 33(2): 233-258.

Forbes. 2015. Unilever's Paul Polman: CEOs Can't Be 'Slaves' To Shareholders. [ONLINE] Available at: http://www.forbes.com/sites/andyboynton/2015/07/20/unilevers-paul-polman-ceos-cant-be-slaves-to-shareholders/#2d365f9140b5. [Accessed 08 December 15]. Forbes. 2016. Five CSR trends that will blossom in 2016. [ONLINE] Available at:

http://www.forbes.com/sites/susanmcpherson/2016/01/08/5-csr-trends-that-will-blossom-in-2016/#5f5800f742a9. [Accessed 13 January 16].

Freeman. R. E. (1984). Strategic management: A stakeholder approach. Boston: Pitman.

(28)

Hartman, R.S., Bozdogan, K. and Nadkarni, R.M. (1979). The Economic Impacts of Environmental Regulations on the U.S. Copper Industry. The Bell Journal of Economics, 10(2): 589-618. Hou, S. and Li, L. (2014). Reasoning and Differences between CSR Theory and Practice in China,

the United States and Europe. Journal of International Business Ethics, 7(1): 19-30. Jessayan, H. (2015). ‘Sjoemelsoftware stort VW in diepe crisis’. Het Financieele Dagblad 21

September. Available from: <fd.nl>. [11 December 2015].

Jo, H., Kim, H. and Park, K. (2015). Environmental Costs and Firm Performance in the Financial Services Sector. Journal of Business Ethics, 131(2): 257-284.

Keleman, R.D. and Vogel, D. (2009). Trading Places: The Role of the United States and the European Union in International Environmental Politics. Comparative Political Studies, 10(5): 1–30. Leonidou, L.C., Leonidou, C.C. and Kvasova, O. (2010). Antecedents and outcomes of consumer environmentally friendly attitudes and behaviour. Journal of Marketing Management, 26(13-14): 1319–1344.

Margolis, J.D. and Walsh, J.P. (2003). Misery Loves Companies: Rethinking Social Initiatives by Business. Administrative Science Quarterly, 48(2): 268-305.

Margolis, J.D. (2001). Responsibility in Organizational Context. Business Ethics Quarterly, 11(3): 431-454.

McWilliams, A. and Siegel, D. (2000). Corporate social responsibility and financial performance: correlation or misspecification. Strategic Management Journal, 21: 603-609.

McWilliams, A., Siegel, D. and Teoh, S.H. (1999). Issues in the Use of the Event Study Methodology: A Critical Analysis of Corporate Social Responsibility Studies. Organizational Research Methods, 2(4): 340-365.

Mohr, L.A., Webb, D.J. and Harris, K.E. (2001). Do Consumers Expect Companies to be Socially Responsible? The Impact of Corporate Social Responsibility on Buying Behaviour. The Journal of Consumer Affairs, 35 (1): 45-72.

Nielsen, 2013. Ad spend by sector: consumer goods marketers spend big. [ONLINE] Available at: http://www.nielsen.com/us/en/insights/news/2013/ad-spend-by-sector-consumer-goods-marketers-spend-bigger.html. [Accessed 13 January 16].

Oh, W. and Park, S. (2015). The Relationship Between Corporate Social Responsibility and Corporate Financial Performance in Korea. Emerging Markets Finance & Trade, 51: 85–94 Orlitzky, M., Schmidt, F.L. and Rynes, S.L. (2003).Corporate Social and Financial Performance: A

Meta-analysis. Organization Studies, 24(3): 403-411.

Porter, M. (1991). America’s green strategy. Scientific American 264, no. 4.

(29)

Prahalad, C.K. and Hamel, G. (1994). Strategy as a Field of Study: Why Search for a New Paradigm? Strategic Management Journal, 15: 5-16.

Rees, B., and Mackenzie, C. (2011). Corporate Social Responsibility and the Open Society. Working Paper, 1-44.

Russo, M.V. and Fouts, P.A. (1997). A resource-based perspective on corporate environmental performance and profitability. Academy of Management Journal, 40(3): 534-559.

Schwartz, E. (1959). Theory of the Capital Structure of the Firm. Journal of Finance, 14(1):18-39. Surroca, J., Tribo, J. and Waddock, S. (2010). Corporate responsibility and financial performance:

The role of intangible resources. Strategic Management Journal, 31 (5): 463-490.

Thomas, L.M. (1992). The business community and the environment: an important partnership. Business horizons, 35(2): 21-24.

Thomson Reuters Corporate Responsibility Ratings (TRCRR) Rating and Ranking Rules and Methodologies. August 2013.

Uddin, M.B., Hassan, R. and Tarique, K. (2008). Three Dimensional Aspects of Corporate Social Responsibility. Journal of Business and Economics, 3(1): 199-212.

Ullman, A.A. (1985). Data in Search of a Theory: A Critical Examination of the Relationships Among Social Performance, Social Disclosure, and Economic Performance of U.S. Firms. Academy of Management Review, 10(3): 540-557.

United States Department of Labor. 2013. Occupational Safety and Health Administration. [ONLINE] Available at: https://www.osha.gov/doc/index.html. [Accessed 13 January 16]. Van den Heuvel, R. (2012). How Robust are CSR Benchmarks? Comparing ASSET4 with

Sustainalytics. Tilburg University, department of Economics. Paper not published.

Van Dijk, B. (2015). ‘Wereld omarmt ‘historisch’ klimaatakkoord’. Het Financieele Dagblad 13 December. Available from: <fd.nl>. [15 December 2015].

Vogel, D., Toffel, M., Post, D. and Uludere Aragon, N.Z. (2010). Environmental federalism in the European Union and the United States. Harvard Business School Technology and Operations. Working paper.

Volkswagen AG (VOW3.DE) (2015). Interactive chart share price. Yahoo!Finance. Retrieved from http://finance.yahoo.com/echarts?s=VOW3.DE+Interactive#{"range":"6mo","allowChart Stacking":true}.

Waddock, S.A. and Graves, S.B. (1997). The corporate social performance – financial performance link. Strategic Management Journal, 18(4): 303-319.

(30)

Appendix

A. Data collection

Table AI. Assigned SIC-codes to industries.

Fast Moving Consumer Goods Construction

Group SIC-code Group SIC-code

Major Group: Major Group:

Food and kindred products 20

Building

Construction 15

Tobacco products 21 Heavy Construction 16

Apparel 23 Construction Special 17

Electronic equipment 36

Wholesale trade non-durable goods 51 Industry Group:

Drugs 283

Personal products, cleaning preparations 284

Computer and office equipment 357

Grocery stores 541

B. Output hypothesis 1

Table BI. Redundant fixed effects test

Redundant Fixed Effects Tests Equation: Untitled

Test cross-section and period fixed effects

Effects Test Statistic d.f. Prob.

Cross-section F 5.285261 (4031,16170) 0.0000 Cross-section Chi-square 16986.775768 4031 0.0000 Period F 31.316069 (5,16170) 0.0000 Period Chi-square 194.759772 5 0.0000 Cross-Section/Period F 5.330624 (4036,16170) 0.0000 Cross-Section/Period Chi-square 17099.448641 4036 0.0000

Table BII. Hausman cross-section random effects

Correlated Random Effects - Hausman Test Equation: Untitled

Test cross-section random effects

(31)

Table BIII. Hausman period random effects

Test period random effects

Test Summary Chi-Sq. Statistic Chi-Sq. d.f. Prob.

Period random 12.759471 3 0.0052

Table BIV. Output regression hypothesis 1

Dependent Variable: ROA Method: Panel Least Squares Date: 01/19/16 Time: 21:42 Sample (adjusted): 2009 2014 Periods included: 6

Cross-sections included: 4032

Total panel (unbalanced) observations: 20210

Variable Coefficient Std. Error t-Statistic Prob. ENV_SCORE(-1) -0.011040 0.006202 -1.780024 0.0751 TOTAL_ASSETS_LN___$ 0.776080 0.267962 2.896236 0.0038 DEBT_TO_ASSETS -0.264290 0.008525 -31.00083 0.0000 C 0.370931 4.201110 0.088294 0.9296

Effects Specification Cross-section fixed (dummy variables)

Period fixed (dummy variables)

R-squared 0.576596 Mean dependent var 5.470198 Adjusted R-squared 0.470836 S.D. dependent var 10.84043 S.E. of regression 7.885726 Akaike info criterion 7.144768 Sum squared resid 1005526. Schwarz criterion 8.726771 Log likelihood -68157.88 Hannan-Quinn criter. 7.662088 F-statistic 5.451957 Durbin-Watson stat 1.928224 Prob(F-statistic) 0.000000

C. Output Hypothesis 2

Table CI. Europe - Redundant fixed effects test

(32)

Table CII. Europe - Hausman cross-section random effects

Test Summary Chi-Sq. Statistic Chi-Sq. d.f. Prob. Cross-section random 161.273820 3 0.0000

Table CIII. Europe – Hausman period random effects

Period random 11.083808 3 0.0113

Table CIV. Europe - Output

Dependent Variable: RETURN_ON_ASSETS Method: Panel Least Squares

Date: 12/14/15 Time: 12:39 Sample: 2009 2014

Periods included: 6

Variable Coefficient Std. Error t-Statistic Prob. ENV_SCORE -0.027408 0.011348 -2.415239 0.0158 TOTAL_DEBT_ASSETS -0.287398 0.016475 -17.44425 0.0000 TOTAL_ASSETS_$___LN 1.704215 0.503961 3.381641 0.0007 C -12.29714 7.945434 -1.547699 0.1218

(33)

Table CV. United States of America - Redundant fixed effects test

Table CVI. United States of America – Hausman cross-section random effects

Table CVII. United States of America – Hausman period random effects

(34)

Table CVIII. United States of America – Output

Dependent Variable: ROA Method: Panel Least Squares Date: 12/14/15 Time: 14:58 Sample: 2009 2014

Periods included: 6

Variable Coefficient Std. Error t-Statistic Prob. ENV_SCORE 0.008652 0.008160 1.060248 0.2891 TOTAL_ASSETS_$___LN 1.058555 0.400512 2.643005 0.0082 DEBT_ASSETS -0.155508 0.011076 -14.04012 0.0000 C -6.744268 6.356636 -1.060981 0.2888

R-squared 0.659182 Mean dependent var 6.346591 Adjusted R-squared 0.583324 S.D. dependent var 8.690941 S.E. of regression 5.610040 Akaike info criterion 6.450260 Sum squared resid 138856.9 Schwarz criterion 7.651584 Log likelihood -16416.58 Hannan-Quinn criter. 6.869688 F-statistic 8.689727 Durbin-Watson stat 2.001721 Prob(F-statistic) 0.000000

Table CIX. Comparing US with Europe

Dependent Variable: ROA Method: Panel Least Squares Date: 01/27/16 Time: 16:25 Sample: 2009 2014

Periods included: 6

Variable Coefficient Std. Error t-Statistic Prob. ENV_SCORE 0.014615 0.004827 3.027840 0.0025 TOTAL_ASSETS_$___LN -1.149535 0.061787 -18.60481 0.0000 DEBT_ASSETS -0.043909 0.004540 -9.670942 0.0000 COUNTRY 0.333067 0.394722 0.843801 0.3988 ENV_SCORE*COUNTRY 0.016585 0.006113 2.713248 0.0067 C 24.15877 0.937656 25.76507 0.0000 Effects Specification Period fixed (dummy variables)

(35)

D. Output Hypothesis 3

Table DI. FMCG industry - Redundant fixed effects test

Table DII. FMCG industry – Hausman cross-section random effects

Table DIII. FMCG industry – Hausman period random effects

The Impact of Corporate Environmental Performance on Financial Performance; an Image-based Approach