Faculty of Economics and Business
The effect of CSR performances on the relationship between a company’s assets and CFP A natural-resource-based view in the North American oil and gas industry
Author: Martin Mathot Student number: 10818391
Date of submission & version: July 1, 2016, final version
Qualification: MSc. in Business Administration – Strategy Track Name of institution: University of Amsterdam
Statement of originality
This document is written by Martin Mathot who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.
The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
Abstract
Purpose – The purpose of this study is twofold: first, to explore whether a linkage between Corporate Social Responsibility (CSR) and Corporate Financial Performance (CFP) exists; and second, to investigate whether the relationship between a company’s assets and CFP will be moderated by highly-CSR engaged companies.
Design/methodology/approach – The study adopts a multiple linear regression analysis of a sample comprising of 69 North American oil and gas companies to examine the effect of CSR on relationship between a company’s assets (including three measures: intangible assets, innovation, & physical assets) and CFP.
Findings – This study finds a significant positive relationship between CSR and CFP (measured by EBIT). Regarding the interaction effect, companies that are highly-CSR engaged, will face a negative effect on the physical assets and CFP relationship (measured by EBIT, ROA, & ROE).
Research limitations/implications – Future research may investigate the relationship between different/more assets and CFP across industries and countries. Furthermore, a qualitative approach is suggested to investigate CSR performances more accurate.
Practical implications – Although it was not possible to find strong relationship between the dependent, independent, and moderator, CSR engagement remains an important variable within controversial industries.
Originality/value – The paper incorporates CSR into the resource-based view of the firm, resulting in a natural-resource-based view of the firm. While past studiesm mainly focused on the relationship between CSR and CFP, the natural-resource based view of the firm in this study focused on a company’s assets in order to create sustainable competitive advantage and superior financial performance.
Keywords: corporate social responsibility, corporate financial performance, natural-resource-based view of the firm, (in)tangible assets, North America, oil & gas industry
Table of Contents Statement of Originality Abstract Table of Contents List of Tables 5 List of Figures 5 1. Introduction 7 1.1 Background 7 1.2 Research Motivation, Research Objective, Research Contributions 9 1.2.1 Scientific point of view 9 1.2.2 Managerial relevance 10 1.3 Structure of the Paper 11 2. Theory and hypotheses 12 2.1 Controversial industries 12
2.1.2 Oil and gas industry 13
2.2.2 CSR - CFP relationship in the oil and gas industry 15
2.3 Natural-resource-based view of the firm and CSR 17
2.4 (In)tangible assets for North American oil & gas companies 19
2.4.1 Intangible assets 20
2.4.2 Innovation 21
2.4.3 Physical assets 22
2.5 Conceptual framework 24
3. Data and methodology 25
3.1 Sample 25
3.1.1 Dow Jones Sustainability Index North America 27
3.2 Data collection 29
3.3 Variables 30
3.3.1 Dependent variable (CFP) 31
3.3.2 Independent variables (company’s assets)) 32
3.3.3 Moderator (CSR) 32
3.3.4 Control variables 33
3.4 Empirical models 36
4. Results 38
4.1 Descriptive statistics and Pearson correlation matrix 38
4.2 Statistical analyses 39
4.3.1 Earnings Before Interests and Taxes (EBIT) 40
4.3.2 Tobin’s Q 42
5. Discussion 46
5.1 The impact of high-CSR engagement on CFP 46
5.2 The effect of CSR on the relation between a company’s assets and CFP 47
5.3 Theoretical implications 49
5. 4 Managers implications 50
5.5 Limitations and suggestions for future research 50
6. Conclusion 52
References 54
Appendices
Appendix 1 Linear regression analysis for ROA Appendix 2 Linear regression analysis for ROE
List of tables
Table 1 Economic, social, and environmental criteria
Table 2 Distribution of selected companies in the North American oil and gas industry Table 3 Operationalization of the central variables
Table 4 Descriptive statistics and Pearson correlation matrix
Table 5 Summary of linear regression analysis for variable predicting EBIT Table 6 Summary of linear regression analysis for variable predicting Tobin’s Q Table 7 Overview significant relationships
List of figures
1. Introduction
1.1 Background
“Oil Companies Face Boycott Over Sinking of Rig” was the headline of the New York Times illustrating the stakeholder pressure concerning Shell’s social responsibility of its sinking oil rig the Brent Spar (Nash, 1995). Nowadays, corporations worldwide are paying more attention to corporate social responsibility (henceforth, CSR), meaning that “businesses contribute to achieving economic, social and environmental sustainability, by responding to not only their shareholders but also other stakeholders” (Ranängen & Zobel, 2014, p. 299). High polluting industries, like the oil and gas industry, face these concerns by their stakeholders more or in a different way than others. Such industries are commonly known as controversial industries because of their “extensive extraction of natural resources without provision for their renewal” (Stevenson, 2010, p. 620). Regarding the Brent Spar case, oil and gas companies could no longer afford to ignore their social responsibility.
CSR engagement has been a unit of analysis for more than three decades in relation to corporate financial performance (henceforth, CFP), resulting in a small, but positive relationship (e.g. Aguinis & Glavas, 2012; Waddockk & Graces, 1997). What CSR is about, and why it exists seems to be clear in the literature, whereas the how deserves more attention (Yuan et al., 2011)
Based on the analysis of 20 years of CSR research across different industries by Dabic et al., (2016), an unevenly distributed collection has been identified showing the incompleteness of industry-specific studies on CSR which hampers comparisons to make. In addition, concerning controversial industries like the oil and gas sector, focus have predominantly been on qualitative measurements of CSR engagement (Dabic et al., 2016). This quantitative omission is surprising, given that corporations within this industry deliver
about fifty percent of the world’s energy (Key world energy statistics, 2015) and the pressure on (short-term) profit motives and (long-term) sustainability requirements can especially be found within the oil and gas industry (Coburn et al., 2012). Collier (2008) points out the negative extras for the society which come along with the oil and gas production, such as (air) pollution, oil spills, and even deaths. After for instance the Deepwater Horizon oil spill in the Gulf of Mexico in April 2010, it clearly shows the risks to companies that are related to this industry. Therefore, oil and gas companies try to integrate CSR-activities within their strategy to compensate these public critics, increase their reputation, and to create legitimacy which is required in the long-term (Du & Vieira, 2012).
From a resource-based view, CSR-activities need to be integrated into the company’s business strategy, as this view ignores environmental and social aspects (Hart, 1995). Regarding the negative effects on the society, due to oil and gas production, it seems necessary to integrate CSR into the RBV in order to internalize the challenges created by the natural environment: natural-resource-based view of the firm (Hart, 1995).
In order to provide more empirical support on the natural-resource-based view regarding the oil and gas industry, the following research question is formulated:
“What is the moderating effect of CSR performances on the relationship between a company’s assets and CFP?”
1.2 Research Motivation, Research Objective, Research Contributions 1.2.1 Scientific point of view
Understanding the relationship between CSR and CFP has been much debated and well-documented in previous decades (e.g. Brammer, Brooks, & Pavelin, 2006; Griffin & Mahon, 1997; Marom, 2006; Horváthová, 2010; Ramchander, Schwebach & Staking, 2012). However, even in recent years, research has not arrived at a consensus. Focusing on the North American oil and gas industry, the objective of this database research is to answer the research question in order to examine whether companies engage in CSR activities only for the sake of ‘doing good’ or also to contribute to their financial results. Stemmed from the natural-resource-based view, the relation and value of a company’s assets and CFP have been examined many times by scholars (e.g. Barney, 1991; Peteraf, 1993). Therefore, this study focuses on the impact of CSR engagement on the aforementioned relationship. This research contributes to CSR research in several ways.
Firstly, although there has been a great deal of research regarding the relation between CSR and CFP, this study focuses on a single industry which provides opportunities to develop an in-depth understanding of variables influencing the CFP. Griffin & Mahon (1997) reviewed the CSR - CFP debates of twenty-five years of incomparable research and called for new research into this relationship that focuses on a single industry (to lead to increased knowledge). In addition, a more recent study (Horváthová, 2010) also points out that the majority of the studies on the CSR - CFP relationship deals with more than one industry which makes it hard to distinguish its effect. Furthermore, as mentioned by Griffin & Mahon (1997) and Horváthová (2010), studies on the CSR - CFP issue deal with multiple industries and therefore results cannot be generalized to other industries which make further research within specific industries interesting (Pätäri et al., 2014). For instance, regarding this research, the North American oil and gas industry belongs to the controversial industries in
which stakeholders are more critical, so returns on CSR performances will be lower as well compared to other (non-controversial) industries (Sen et al., 2006; Du & Vieira, 2012). Secondly, according to the perspectives from which CSR research has been evolved, the resource-based view of the firm received little attention towards CSR - CFP relationship (Jones & Bartlett, 2009). In line with the increasing pressure by stakeholders and civil society on companies in the controversial industry to act socially responsible and environmentally ‘friendly’ (Campbell, 2007), Hart (1995) stated that for the resource-based view of the firm to remain relevant, natural environmental challenges and social problems need to be adopted. This research, therefore, contributes to the existing literature as it will add an extra dimension, CSR, in order to create a balance between a company’s resources and its external environment: natural-resource-based view.
1.2.2 Managerial relevance
CSR has become increasingly important in the corporate world, due to the increase of corporate scandals (Cedillo Torres et al., 2012). Sponsor managers of companies with bad reputations seem to be interested in changing their negative image through CSR activities (Arnold, 2001). Earlier research (Simmons & Becker-Olsen, 2006; Yoon, Gürhan-Canli, & Schwarz, 2006) has pointed out that sponsorship fit and perceived sincerity of the company’s motives in determining the success of CSR campaigns is of influence on the firm’s financial performance for well-liked firms. For sponsorship management, these findings indicate that sponsorship fit and the sincerity of the company’s motives play a key role in sponsorship alliances. This study will be of importance for managers within the controversial oil and gas industry because the independent effect of CSR will be determined. Furthermore, this impact of CSR will be demonstrated by a moderating (interaction) variable. Through this interaction, the effect on the existing relation can easily be identified.
1.3 Structure of the Paper
The structure of the remainder of this paper is as follows. Section 2 “Theory and hypotheses” gives deeper insight into prior literature about important subjects surrounding this study such as the North American oil and gas industry context, the natural-resource-based view of the firm and the relation between CSR and CFP. This literature review will lead to the creation of hypotheses. Section 3 “Data and methodology” describes the sample selection and data collection, which together create the research design in which variables are operationalized. Section 4 “Results” interprets the outcome of the analyses and this study ends with Section 5 “Discussion and conclusion”, in which will be elaborated on the discussion of the findings in order to provide an answer to the research question, followed by the limitations of this study and brief suggestions for future research streams.
2. Theory and hypotheses
2.1 Controversial industries
Nowadays, globalization is one of the biggest drivers of the global economy (Forbes, 2013) and also of the CSR as the impact of globalization needs to be compensated (Smith, 2008). Smith (2008) argues that multinational corporations see globalization as an opportunity to increase their business in several (positive) ways and CSR as “a means of attendant legal and reputational risks” (p.2). On the other side, stakeholders (i.e. non-governmental organizations, investors, environment etc.) pay attention to the negative effects (corporate) globalization can have according to the potential exploitation of the environment, global inequalities, and human rights (Newell & Frynas, 2007).
Compared to manufacturing corporations which are able to choose their business area, corporations within controversial industries are not provided with these possibilities such as where to find oil, gas, gold, and diamonds (Smith, 2008). Smith (2008) points out in his speech that sometimes corporations, in extractive industries, have to do their operations in fragile states/conflict zones where there is not an active institution checking this activity. However, the required resources could be situated in for instance rainforests where CSR is highly required. This type of industry can also be seen as a controversial (Du & Vieira, 2012; Cai et al., 2012). However, there are still cross-cultural differences about what is actually considered controversial (Waller et al., 2005; Reast et al., 2013).
These legal industries are explained by Cai et al. (2012) as industries “characterized by social taboos, moral debates, and political pressures, include sinful industries, such as tobacco, gambling, alcohol, and adult entertainment as well as industries involved with emerging environmental, social, or ethical issues, i.e., weapons, nuclear, oil, cement, and biotech” (p. 2). Because companies within controversial industries are operating in a way that may be unethical due to failure to act accordance with stakeholder interests or the negative
environmental impact, implications for maintaining legitimacy arise (Palazzo & Scherer, 2006). Due to the industry's’ negative reputation and ethical, social, and environmental issues, companies try to invest in maintain their ‘social license to operate’ (Inkpen & Moffett, 2011; Parsons et al., 2014). After the crash of the American titan Enron in October 2001, social responsibility has become more important to corporations and especially those within the controversial industry (due to their bad reputation) such as Enron in the oil & gas industry, to recover this negative image (Yoon et al., 2006).
2.1.2 Oil and gas industry
On the evening of 20 April 2010, the Deepwater Horizon oil rig in the Gulf of Mexico started releasing gas which subsequently caused an explosion (BP, 2010). The explosion created a fire, which burned for 36 hours, and made the oil rig sank. Due to the leak of 134 million gallons of oil in the Gulf, responsible for the death of 11 people and 17 injured people, it resulted into one of the worst oil spill and environmental disasters in the United States of America (The Telegraph, May 09, 2015).
The oil and gas industry impacts everyone’s lives through petrochemical products which are used for clothing and carpets as well as fuels for heating, electricity, and of course transportation (Inkpen & Moffett, 2011). In addition to transportation fuels, crude oil and natural gas prices are globally two of the most closely watched commodities and therefore affect national issues, such as politics and security, and sometimes cause international conflicts (Inkpen & Moffett, 2011). As oil and gas play an important role in the global economy, demand of energy will probably increase, accordance to the International Energy Agency, by yearly 1.5% to 2030 (Inkpen & Moffett, 2011). However, in comparison to companies represented in uncontroversial industries, controversial industries like the oil and gas industry have higher standards of scrutiny and companies therefore challenge difficulties in obtaining organizational legitimacy (Reast et al., 2013; Byrne, 2007). Certainly, these
companies do not all meet the requirements of their institutional environment but they neither are able to effectively improve their legitimacy status strategically (Miller & Michelson, 2013). For instance, Winn et al. (2008) mentioned in their article that this is partly affected by the asymmetrical effect between ‘good’ and ‘bad’ actions, in which the latter will often have a more harmful impact. Due to high profile disasters, together with an increasing competition through globalization, made the management environment more difficult and raised more criticism towards oil companies (Colbert & Kurucz, 2007; Du & Vieira, 2012). Oil and gas companies have been criticized by governmental and non-governmental organizations, but also media pay attention to issues ranging from the impact on local communities to environmental violations (i.e. Telegraaf 2016; New York Times, 2016, The Guardian, 2015).
Although debates about international climate emerged already around the 1980s, corporate strategic change got implemented into the Kyoto Protocol in 1997 (Kolk & Pinkse, 2004). This international agreement mainly aimed at reducing the level of greenhouse gasses to support the overall development of regulation with respect to climate change (Kolk & Pinkse, 2004). For companies within the oil and gas industry to improve their negative image, several social and environmental initiatives are being supported such as the Shell foundation which focuses on game-changing solutions to market failures in Africa and Asia and BP’s targeted health program in order to eliminate malaria in Bintuni Bay Indonesia (Shell Foundation, 2016; BP Global, 2016). There can thus be stated that oil and gas companies, in response to stakeholder sensitivity to social and environmental issues and widespread negative publicity, are engaging in CSR in order to build reputational capital and gain legitimacy for their actions (Du & Vieira, 2012).
2.2.2 CSR - CFP relationship in the oil and gas industry
The question whether companies in controversial industries are able to create wealth in a way that their (social) environment will not be harmed, has busied many scholars by focusing on doing well while doing bad (Margolis, Elfenbein, & Walsh, 2007). Within this report, Corporate Social Performance (CSP) and CSR will be used interchangeably, with equivalent meaning.
An increasing amount CSR literature is focusing on the important role in controversial industries (Cai, Pan, & Jo, 2012). Regarding the oil and gas industry within the Organisation for Economic Co-operation and Development (OECD) countries, for instance, the United States of America, the government implemented regulations of emissions and oil and gas activities (Spence, 2011). As this industry is facing numerous challenges due to highly visible (negative) effects in terms of accidents (e.g oil spills), stakeholders require companies to engage in CSR more than other industries (Frynas, 2009; Spence, 2011). Therefore, this industry has been among the leading industries in championing CSR (Frynas, 2009). According to the literature, research in controversial industries conducted by Cai et al., (2012) and Moura-Leita et al. (2014), found contradictory results regarding the CSP-CFP relationship (Rodrigo, Duran, Arenas, 2016). These contradictory results might stem from the fact that, according to Hart and Ahuja (1996), the effect of emission reduction on CFP will be caused by cost reduction due to the lower costs of solving inefficiencies and wastes. However, they state that it is getting more difficult to improve CFP in the long term, as these investment costs may exceed (physical asset replacement) the savings generated. As the controversial oil and gas industry, compared to other industries, is more associated with critical stakeholders who are conscious of their interests, the returns on CSP will also directly/indirectly be lower/less (Sen et al., 2006; Du & Vieira, 2012). However, the positive effect on CFP through CSR-engagement is considered to be bigger in controversial
industries, which are heavy-polluting, in comparison to more neutral industries. Schreck (2009) explains this by the positive effect on cost reduction and reputation improvements. Thus, the relationship between CSR and CFP seems to differ across types of industry. A longitudinal study conducted by Clakson et al. (2011), focused on the CSR-CFP relationship among the four most controversial industries in the United States - Oil and Gas; Metals and Mining; Pulp and Paper; and Chemicals. Their study pointed out that CSR engagement positively influences CFP (proxied by Tobin’s Q). Furthermore, Baird et al. (2012) found a positive effect of CSP in relation to CFP among companies in polluting industries.
H1: Highly-CSR engaged companies in the oil and gas industry are positively related to CFP.
Testing this hypothesis is essential for this research as it represents the starting point of the study. Hereafter, the further hypotheses will be narrowed down to the company’s tangible and intangible assets. Several theoretical perspectives focused on CSR like for instance the stakeholder theory, agency theory perspective, and the theory of the firm perspective (Stoian & Gilman, 2016). However, some limitations can be identified within these approaches such as the lack of CSR activities specification which develop company performance and limited depth of practical implications (Stoian & Gilman, 2016). According to McWilliams, Siegel, and Wright (2006), relating strategic CSR performances to performance improvements enables CSR to be studied from the resource-based view of the firm. The resource-based view of the firm has been examined many times by scholars across industries and the acknowledged the important positive role of intangible assets in gaining competitive advantage and superior CFP (e.g. Hall, 1993; Barney, Wright, & Ketchen, 2001) However, the resource-based view of the firm focuses on a company’s internal assets without taking external effects into account such as social and environmental dimensions (Hart, 1995;
Stoian & Gilman, 2016). Therefore, integrating CSR into the resource-based view of the firm would be an interesting extension: the natural-resource-based view of the firm.
2.3 Natural-resource-based view of the firm and CSR
The resource-based view of the firm specifies the nature of a company’s resources and capabilities, to explain performance heterogeneity in terms of profit and value at the company level (Fahy, 2002; Hoopes et al., 2003; Curado & Bontis, 2006). As explained by Barney (1986; 1991), differences in resources are caused by factor market imperfections. When the tangible and intangible resources are valuable, rare, inimitable, and non-substitutable, they could be a source of sustainable competitive advantage (Galbreath, 2004).
Due to the omission of social and environmental elements within the resource-based view of the firm, Hart (1995; Hart & Dowell, 2010) created the natural-resource-based view of the firm. Especially in the case of companies performing in controversial industries, like the oil and gas industry, the natural environment could certainly influence the creation of sustainable competitive advantage (Hart & Dowell, 2010). In other words, strategy and competitive advantage could derive from ‘capabilities that facilitate environmentally sustainable economic activity’ (Hart, 1994, p. 991). The natural-resource-based view of the firm focuses on three strategic assets: pollution prevention, product stewardship, and sustainable development (Hart & Dowell, 2010). Pollution prevention puts pressure on companies to eliminate or at least reduce emissions from their business practices (Hart, 1995). For instance, in the United States in 1986, companies were required to publicly report their emission levels, the so-called toxic release inventory, which caused management in for instance the oil and gas industry to revise their pollution control (Hart, 1995). Pollution prevention reduces financial and environmental costs resulting in a competitive (cost) advantage and increasing financial performances (Hart, 1995). Product stewardship is an environmental management strategy in which the impact of every life-cycle stage needs to be
minimized (Hart, 1995). Product stewardship could be the source of a company’s reputation building and product differentiation by creating the first-mover advantage (Hart, 1995). The main goal of pollution prevention is to diminish emissions, whereas product stewardship mainly focuses on raw materials selection in order to diminish their environmental impact (Hart, 1995). Besides the focus on the relationship between the company and its environment, sustainable development strategy, however, pays attention to developing countries in relation to economic activity and environment (i.e. social element) (Hart, 1995).
Hart & Dowell (2010) mention in their article that there has been significant research on the positive relationship between pollution prevention and CFP. However, they stated that more research is required in terms of the company’s assets and the effect on CFP. Foss (1997) points out that there are several cases in which physical assets create sustainable competitive advantage for a company. Additionally, it may be empirically beneficial for the natural-resource-based view of the firm to include both tangible and intangible assets (Foss, 1997; Andersen & Kheam, 1998; in Galbreath, 2005). In order to assess the natural-resource-based view of the firm’s support in the empirical literature, Newbert (2007) assessed 55 empirical articles on RBV. Of these 55 articles in the sample, 33 articles tested the relation between specific resources and CFP. However, empirical support is found for only 37% of the tests(Newbert 2007). The overall findings seem to suggest that while single core competencies and capabilities positively affect CFP and SCA (67% and 71% respectively), assets do not (Newbert, 2007). In addition, Newbert (2007) states that the effect on CFP is likely to be caused by a combination of capabilities and/or resources.
Within the last decades, the increasing social awareness of CSR by various stakeholders put pressure on companies (Stoian & Gilman, 2016). Besides the question about which assets for companies to priorities, the challenge for companies is to exploit those that are more likely to contribute to their competitive advantage and financial performance (Stoian &
Gilman, 2016). As the natural-resource-based view of the firm pays attention to the influence of social and environmental elements on CFP, the omission within the traditional resource-based view of the firm can be supplemented and extended by a company’s CSR engagement. However, based on the literature provided, companies in the oil and gas industry would benefit from the further empirical investigation of CSR related performances which can, according to the theory, develop superior firm performance and sustainable competitive advantage.
2.4 (In)tangible assets for North American oil & gas companies
In the oil and gas industry, companies face difficulties in responding to environmental pressures while at the same time to the shareholders who are pressuring for a return on their investment (Garcia, Lessard, & Singh, 2014). Such controversial industry is typically characterized by people in a way in which tangible assets play the most important role in relation to firm performance. However, the value is generally created by a combination of tangible and intangible assets (Garcia, Lessard, & Singh, 2014).
In the present study, the strategic natural-resource-based view of the firm will be applied to investigate whether the effect a company’s assets on CFP will be influenced by the engaging in CSR activities. Because this study stems from the natural-resource-based view of the firm, intangible assets have an important role in establishing competitive advantage and will, therefore, be incorporated (e.g. Barney, 1991). Moreover, due to measurement difficulties and the barrier to imitation and substitute, make them interesting to study (Padgett & Galan, 2010). Oil and gas companies have to deal more with uncertainties than other types of firms, which make intangible assets seem to be more (positively) influential on competitive advantage and CFP (Casault, Groen, & Linton, 2015). Based on the three strategic assets related to the natural-resource-based view, “sustainable development strategy” pays attention
to the social element. Therefore, this strategic asset is linked to the first variable chosen: “intangible assets”, as it contains relevant statistical data in Datastream Professional regarding the oil and gas industry (e.g goodwill). However, “innovation”, is not taken into account by Datastream Professional’s measure of intangible assets, so for this study “product stewardship” could create first-mover advantages and supports reputation building. This concept is used as the second intangible asset. Contrarily, “tangible assets” are represented by the company’s amount of physical assets. Regarding the natural-resource-based view, “pollution prevention” focuses on environmental costs which can be realized through specialized machinery/equipment. In the existing literature, the aforementioned variables have been examined by many scholars regarding the effect on CFP (Culver, 2008). However, from a natural-resource-based perspective, the aforementioned relationship should be tested on the influence of CSR-engagement (McWilliams & Siegel, 2000).
By operationalizing and digging deeper into the literature, more relevance will be explained resulting into hypotheses.
2.4.1 Intangible assets
According to the controversial industry, the increasing amount of information about intangible assets in annual reports shows the importance of disclosing for companies (Kristandl & Bontis, 2007). Since no generally accepted definition exists in the literature (White, 2006), intangible assets can be defined as “a claim to future benefits that does not have a physical or financial (a stock or a bond) embodiment” (Lev, 2001, p.5). Andanova & Ruíz-Pava (2016) complement that intangible assets, in the oil and gas industry, are internally (e.g. blueprints, software) as well as externally (e.g. patents, licenses) acquired assets. According to the natural-resource-based view, it acknowledges that intangible assets seem more likely to generate superior performances and sustainable competitive advantage when they are valuable, rare, inimitable, and non-substitutable (e.g. Barney, 1991; Connor, 2002;
Galbreath, 2004;) Nevertheless, the relationship between CSR, intangible assets, and financial performance is statistically and quantitatively difficult to create (Parisi & Hockerts, 2008).
The value creation process seems to rely on the company’s non-financial (intangible) assets which are supported by CSR strategies (Parisi & Hockerts, 2008). According to CSR strategies, many elements are intangible which cannot be measured traditionally (Vilanova, Lozano, & Arenas, 2009). Several companies claim to be highly engaged with CSR performances through, for instance, sustainability reports and environmental policies.
Based on natural-resource-based view, oil and gas companies can be expected to legitimize themselves to increase their brand image through contribution to the environmental system (Popoli, 2015). With the brand image or corporate reputation being one of the company’s most important intangible assets, especially in controversial industries, CSR is seen as interrelated and supportive subject (Fombrun & Gardberg, 2000; Popoli, 2015). In the past decade, the reputation of companies within the oil and gas industry has been negatively affected due to accidents (e.g. spills) and bad media attention (Van Halderen et al., 2016). When companies start to invest in CSR activities, this can benefit firm-stakeholder relation and the overall firm reputation (Sen & Bhattacharya, 2001). In 2012, the world’s three biggest oil and companies (Halliburton, Baker Hughes, and Schlumberger) adopted precautionary practices to contribute to the ecological environment, resulting in more than 1000 patents (to protect innovation) (intangibles). The information provided results in the following hypothesis:
H2: The relation between a company’s intangible assets and CFP, will be positively moderated by highly-CSR engagement.
2.4.2 Innovation
Nowadays, for companies to be innovative and successful, attention needs to be paid to the effects of operational processes on social and environmental surroundings (McWilliams & Siegel, 2000; MacGregor & Fontrodona, 2008). Based on the natural-resource-bases view of the firm, it emphasizes the importance of innovation, in terms of R&D investment, to improve long-term financial performance. Because innovation plays an important role in efficiency matters and climate change issues, this has led to increased attention towards R&D activities (Costa-Campi, García-Quevedo & Trujillo-Baute, 2015). CSR can, therefore, be seen as a way of investing in order to differentiate in terms of processes improvement and regulate the impact of products (Padgett & Galan, 2010). As one of the biggest challenges for innovation goes most likely hand in hand with environmental improvements, like for instance emissions reduction, this can also be seen as an opportunity with win-win outcomes (Frynas, 2015). CSR is, therefore, able to be a valuable interplay between business and society. Regarding the aforementioned challenges, as companies develop innovation strategies for R&D investments, CSR strategies need also be created in order to create a level of differentiation (e.g Boehe & Cruz, 2010; McWilliams & Siegel, 2000) According to McWilliams & Siegel (2000), this level of differentiation, demonstrates the company’s social and environmental involvement. The oil and gas sector has been steadily redefining production possibilities and uses ‘green’ marketing which turned out to positively influence product image (Ko et al., 2003). Furthermore, as innovation became one of the main drivers of gaining competitive advantage and CSR has been acknowledged within a company as the strategic asset, both concepts need to be combined to generate superior financial performances (MacGregor & Fontrodona, 2008).
H3: High-CSR engagement significant positively moderates the relationship between high innovative companies and CFP.
2.4.3 Physical assets
Derived from the first industrial revolution, physical assets can be seen as the engine of growth, especially the power of steam engines and the availability of energy from fossil fuels have played an important role in the global economic growth (Ayres & Warr, 2005). Physical assets in the oil and gas industry generally contain associated plant and equipment, above ground and sub-seafloor pipelines, drilling rigs, and oil production platforms (Jeeva, 2014). From a research perspective, Foss(1997) acknowledges that in multiple cases the company’s physical (tangible) assets generate competitive advantage. Therefore, scholars recommend the inclusion of tangible resources in order to empirically test the resource-based view (Galbreath, 2005). For companies in the oil and gas industries, it is difficult to maintain the capacity and capability regarding physical assets and to simultaneously operate (cost) efficiently. However, effective and efficient physical assets will support the company’s profitability (Jeeva, 2014). According to CSR performances, controversial companies which pretend to maintain an environmental policy within their production processes, mostly redesign these processes and physical assets in order to establish operational efficiency and waste minimization (Branco & Rodrigues, 2006). In order to embrace pollution prevention, these end-of-pipe treatments influence only a company’s physical assets (Russo & Fouts, 1997). By using for instance filtration equipment in a company’s physical assets, specialized expertise is not required (Russo & Fouts, 1997). Regarding the oil and gas industry, CSR-engagement could be established by investing in, for instance, a smokestack scrubber, an air pollution control device, in order to create an economy of scale (McWilliams & Siegel, 2001). Moreover, BP and Royal Dutch/Shell became leaders in renewable energy and
managed to decrease their contribution to global warming (Frynas, 2005), which is partly done in order to influence stakeholders’ behavior and attitudes. Sustainable technologies can be used to use resources more effectively and additionally to reduce the gas emissions (Bohnsack et al.,2014). The information provided results in the following hypothesis:
H4: The relation between a company’s physical assets and CFP, will be positively moderated by their CSR performances.
2.5 Conceptual framework Figure 1 Conceptual framework
Where:
H1: H1: Highly-CSR engaged companies in the oil and gas industry are positively related to CFP (EBIT, ROA, ROE, and Tobin’s Q).
H2: H2: The relation between a company’s intangible assets and CFP, will be positively moderated by highly-CSR engagement.
H3: The relation between high innovative companies and CFP, will be positively moderated by their CSR performances.
H4: The relation between a company’s physical assets and CFP, will be positively moderated by their CSR performances.
3. Data and methodology
In this section, the data and methodology that have been used in order to answer the aforementioned hypotheses are presented. Focusing on CFP and physical assets since they are measured numerically, a quantitative method was considered most suitable for this study. Although intangible assets are also included, which are difficult to measure numerically, due to the limited timeframe for conducting this study, a qualitative approach was not realizable. Thus, the study uses a completely reproducible quantitative method, a database research, to examine the effect of CSR-engagement on the relation between a company’s assets and CFP. Firstly, a description of the sample is given, followed by an explanation of the data collection.
3.1 Sample
This study undertakes a database analysis of North American listed companies in the oil and gas industry, for the sample period from January 2011 to December 2014. This period was chosen because RobecoSAM could not provide the 2010-2011 list with 2014 as the most ultimate year available. Originally, the research was based on two databases - Compustat and Dow Jones Sustainability Index (DJSI). However, due to missing data, annual and
sustainability reports, as well as Datastream Professional have also been employed.
During the data collection process, two companies (Venoco Inc. & Zion Oil & Gas Inc.) were excluded due to missing balance sheet values. For this study, the total sample consists of 69 companies from 2011 to 2014, resulting in 276 firm-years concerning low-CSR (44) and high-low-CSR (25) engagement among companies in the North American oil and gas industry. However, data were checked in terms of outliers and missings to obtain a normal distribution. In Appendix 3, all boxplots are provided in which all outliers removed are marked (*). As all missing data were hand-collected using annual reports, no more missings were included. To counteract the multicollinearity, centered variables were created for all independent variables and the moderating variable. Secondly, due to skewness to the right, a normal distribution was obtained by eliminating (not deleting) the outliers, preventive, among all variables (to decrease Kurtosis). In order to meet the (normality) assumptions of parametric statistical tests, the variable “firm size” was transformed into logarithms. This logarithmic transformation squeezes together data from the large oil and gas companies, whereas the data of smaller companies will partly be stretched. From a theoretical perspective, a negative aspect of using logarithms is that real-world examples are being reshaped as they fundamentally alter the nature of the variable (e.g. Litchfield & Wilcoxon, 1949; Bloom, 1999; Osborne 2010). From a statistical point of view, however, logarithmic transformations improve normality of data and meet the assumptions of parametric tests (e.g. Zimmerman, 1998; Osborne,2010). Finally, according to the other variables, outliers were excluded by trimming the top and bottom numerically. In other words, regarding the order of the percentages of the variables included, 50% of the lower and 2% of upper were deleted for CFP proxy EBIT (e.g. EBIT >.50 & EBIT < 98). Moreover, in order for the regression analysis to be able to predict the future, independent variables need to be lagged. Based on the CSP-CFP research of Waddock & Graves (1997) and McWilliams & Siegel (2000), this
study uses lagged CFP proxies and lagged control variable year. Thus, the independent and control variables are lagged by one year. 65 cases were excluded due to the lagging control variable year 2014, together with four outliers. However, as almost 25% of the cases were deleted by lagged variables, both regression analyses have been compared and did not differ considerably. After the data cleaning process was conducted, 165 firm years (69 companies) will be used to conduct the examination of the empirical models to test the hypotheses. Regarding the descriptive statistics in Table *, 39% (.39) of these 165 firm years are included in the Dow Jones Sustainability Index North America (DJSINA), which are considered to be highly engaged in CSR.
Since the natural-resource-based view of the firm does not incorporate social and environmental aspects into its theory, this study uses CSR performances to moderate the relationship between a company’s assets and CFP. According to the existing literature, this study deduces the level of a company’s CSR engagement based on at least two years of inclusion, from 2011 to 2014, in the Dow Jones Sustainability Index North America. This index objectively evaluates companies on several dimensions over time, and is, therefore, able inclusion in the Dow Jones Sustainability Index North America, to serve as the proxy for CSR (e.g. Lopez et al., 2007; Consolandi et al., 2009; Artiach et al., 2010).
3.1.1 Dow Jones Sustainability Index North America
As stated in section 2.2.2, CSR has been defined as ‘a balanced approach for organizations to address economic, social and environmental issues in a way to benefit people, communities, and society’ (Leonard & McAdam,2003, p. 27). In collaboration with S&P Dow Jones Indices, investment specialist RobecoSAM composed the globally used databases of financially sustainability information: Dow Jones Sustainability Index (RobecoSAM, 2016). The Dow Jones Sustainability Index tracks and assesses around 3,800 listed companies in terms of economic, social, and environmental criteria, addressing issues such as
climate change mitigation, controversial issues, and risk & crisis management (RobecoSAM, 2016). Additionally, the Dow Jones Sustainability Index was identified as the most relevant sustainability ranking in 2014 (RobecoSAM, 2016). The number of companies invited for the DJSINA was 600 in 2015 (RobecoSAM, 2016). RobecoSAM uses the Corporate Sustainability Assessment, a questionnaire featuring 100 economic, social, and environmental questions, to identify whether a company is able to respond to world’s sustainability trends (RobecoSAM, 2016). These economic, social, and environmental criteria and weightings are stated in Table 2 below. Companies which gain the highest score will be included in the Dow Jones Sustainability Index, continuously checked, and thus stimulated to continue investing in sustainability practices (RobecoSAM, 2016). Due to this yearly review, Dow Jones Sustainability Index North America focuses on best-in-class companies of which 20% of the best companies per industry will be selected (DJSI North-America Index Guide, 2015).
Table 1 Economic, social, and environmental criteria
3.2 Data collection
To be able to answer the hypotheses, a multiple linear regression analysis will be conducted, based on secondary panel data available from several databases: RobecoSAM, COMPUSTAT, and Datastream Professional.
According to the sample period from 2011 to 2014, the companies (for at least two years) included in the Dow Jones Sustainability Index North America were collected by requesting data from RobecoSAM. Since 2011, the Dow Jones Indices data are no longer available online, but every Master student (only) who is interested in obtaining data from RobecoSAM needs to fill out an academic request and send this to RobecoSAM’s headquarter in Zurich. By signing a nondisclosure agreement, the information requested has been provided by e-mail. RobecoSAM only provides data whether a company is included in the DJSI or not, which can be seen as a weakness for scholars.
Given the sample, first the COMPUSTAT database is used, within the Wharton Research Database Service, as it contains accurate data from companies in different segments such as North-America, Global, Bank, & Historical segments (Wharton, 2016). According to the aim of this study, data will only be obtained from COMPUSTAT North America. To determine whether CSR engagements influence the effect of tangible and intangible assets on CFP, COMPUSTAT North America was used to create the second sample group consisting of companies with low-CSR engagement. Using the oil and gas industry SIC code (Ticker), 1311, COMPUSTAT North America provided all companies presented in the industry. Comparing the company lists from Dow Jones Sustainability Index North America and COMPUSTAT North America, resulted in two groups: high-CSR and low-CSR engagement, respectively.
In order to test the aforementioned hypotheses empirically, this study gathered data from two different databases. After it was determined which companies were included and excluded in the Dow Jones Sustainability Index North America (from 2011 to 2014), the data regarding the variables were obtained. Firstly, COMPUSTAT was requested in order to identify several CFP indicators: return on equity (ROE), return on assets (ROA), and earning before interests and taxes (EBIT). Secondly, provided by the University of Amsterdam, the financial database of businesses, stock prices, and macroeconomic data “Datastream Professional” has been accessed. The first independent variable “intangible assets” and control variable “Tobin’s Q” have been obtained by Datastream Professional. Datastream Professional contains financial data for more than 175 countries worldwide over the past 50 years. Finally, data for the remaining independent variables “innovation”, “physical assets”, and control variables “firm size’’ and “firm risk”were obtained from COMPUSTAT. Some data from years, in particular, were missing, which have been hand-collected from annual and sustainability reports online.
Within this study, the Pearson correlation and multiple linear regression analysis are designated as the most appropriate method for investigating the interaction effect of CSR on the relationship between a company’s assets and CFP. The data collected will be tested using the analysis software IBM SPSS Statistics in which the regression analysis tests the hypotheses whereas the Pearson correlation explains the extent and direction of the possible relationships.
Table 2 Distribution of selected companies in the North American oil and gas industry
Robustness check
Before conducting the main analysis, residual analysis were conducted to check the multivariate linear regression model assumptions. As stated by Gujarati (2003), the assumptions of linear regressions include normality, data linearity, homoscedasticity, independence of error, and multicollinearity. Regarding the regression output, VIF seems to be <10 which suggests no multicollinearity (Table 5). Based on the histogram and scatterplot, these show the normal distribution and homoscedasticity. When error terms do not correlate and are consistent, this confirms the homoscedasticity within the study (Petersen, 2009). Furthermore, in contribution to the robustness of this analysis, the linearity of the partial plots and the created year dummies contribute to establish independent errors.
3.3 Variables
To test if engagement in CSR performances has impact on the relationship between a company’s tangible and intangible assets and CFP, several data related to CSR and the
natural-resource based view of the firm and financial data is collected. The variables used to research the effect on CFP are explained in paragraph 3.3.1, the measures for a company’s tangible and intangible asset are explained in paragraph 3.3.2 and finally the control variables will be presented in paragraph 3.3.3.
3.3.1 Dependent variable (CFP)
The financial performance indicators used in this study derived from the scientific literature and have been used in examining CFP relationships (e.g. Choi, Kwak, & Choe, 2010; Albertini, 2013). Mostly accounting-based measures will be used to account for the CFP of the companies investigated. In the existing literature, accounting-based measures are often used because CFP indicators concern mostly growth and profitability (Griffin & Mahon, 1997). Margolish and Walsh (2001) found that ROA and ROE have been used most often in examining company performances. However, as this study also focuses on a company’s ability to generate profit (Zahra, 1995, p.235), EBIT has been incorporated into the analysis as well. Finally, Tobin’s Q, a financial market-based measure of firm performance, (e.g. Hall, 1993; Bharadwaj, Bharadwaj, & konsynski, 1999).
ROE, ROA, EBIT & Tobin’s Q
Derived from the scientific literature, return on equity (ROE) is calculated by dividing the net income by average shareholders’ equity. To account for the return on assets (ROA), the net income of the companies is divided by their average total assets. The data regarding a company’s earnings before interest and taxes (EBIT) have been derived from COMPUSTAT data. Tobin’s Q has been incorporated because of the inclusion of intangible assets as variable. Tobin’s Q has been widely accepted by other scholars focusing on the CSR-CFP relationship (Chen & Wang, 2011). The proxy used for Tobin’s Q in this research is the market to book value ratio, obtained by Datastream Professional.
3.3.2 Independent variables (company’s assets related to natural-resource-based view) According to the literature provided in this study and explained by Barney (1991), companies can create a sustainable competitive advantage when their tangible and intangible assets are valuable, rare, inimitable, and non-substitutable and managed properly. However, as explained before, the value is generally created by a combination of tangible and intangible assets (Garcia, Lessard, & Singh, 2014). Therefore, this study uses both intangible assets (i.e. intangible assets & innovation) and tangible assets (physical assets), which will be operationalized below.
Intangible assets, innovation, & physical assets
Data for intangible assets have been collected by Datastream Professional. According to the oil and gas industry, intangible assets contain for instance goodwill, patents, copyrights, trademarks, capitalized software development costs/computer programs (Datastream Professional, 2016).
As stated by Siegel & Vitaliano (2007), product innovation and differentiate business strategies are directly linked to company’s level of CSR-engagement. Regarding company’s innovation capacity, the ratio of in-process R&D Expense to company’s (total) number of employees has been used as innovation proxy.
Physical resources are measured by the ratio of total assets minus the firm’s current assets and divided by total assets.
3.3.3 Moderator (CSR)
As explained in section 3.1.1, CSR will be measured by inclusion in the Dow Jones Sustainability Index North America for at least two years from 2011 to 2014. Companies that are listed in the Dow Jones Sustainability Index North America, integrated social and
environmental aspects into their business practices, which can, therefore, be considered as highly-CSR engaged.
Dow Jones Sustainability Index North America inclusion
Like above mentioned, a company’s level of social and environmental behavior is based on the inclusion in the Dow Jones Sustainability Index North America. As this index maintains high criteria to get included, company inclusion will be treated as exceptional. Therefore, companies included are expected to be highly engaged in CSR performance whereas companies, that are not included, are considered as low-CSR engaged. This study uses a dummy variable for CSR to moderate the relationship between company’s assets and CFP: 1 if listed on the Dow Jones Sustainability Index North America and 0 otherwise. According to Beleo et al. (2004), in their study, they confirm the measurement usage of the company’s level of CSR-engagement by being included in the Dow Jones Sustainability Index. Furthermore, to provide accurate data regarding CSR research, RobecoSAM is highly recommended.
3.3.4 Control variables
Two control variables, firm size and firm risk, will be used in this study which is based on previous empirical research on similar topics (McWilliams et al., 2000; Galbreath, 2005).
Firm size, firm risk, and year
It is suggested that the size of the firm will positively influence a company’s financial and sustainability performance (Trotman & Bradley, 1981; Artiach et al, 2010). Moreover, as bigger companies have the possibility to be more vulnerable to stakeholder activism or that there may be economies of scale in CSR performance, size is integrated in order to control for this possibility (Siegel & Vitaliano, 2007). In other words, larger companies are generally more visible to society so they will face more pressure for CSR engagement (Mousiolis,
Zaridis, Karamanis, & Rontogianni, 2015). According to Weber (2008), companies in controversial industries would like to maintain their ‘license to operate’. When company size increases, also will the amount of stakeholders which results in even more pressure to invest in CSR. There are many ways in which to measure size, however in this study firm size will be measured by using the natural logarithm (ln) of sales volume. Due to the natural log transformation, the skewness of the data has been reduced. Furthermore, the size of a firm can affect the quantity of resources available and the perceived risk (Kirby & Kaiser, 2003).
Firm risk (i.e. financial risk) of the firm is also seen as an important control variable when studying the CSR-CFP relationship (e.g Ortlitzky & Benjamin, 2001; Andersen & Dejoy, 2011). In this study, the ratio of total debt to total assets is a proxy often used in the literature, and will therefore be used in this study.
Because this study contains data from several years, a so-called longitudinal study, the result may vary over time. A dummy variable for each year, 2011 - 2014, has been created with 2014 as the base category. In Table 3 summarizes the operationalization of the central variables. Ta ble 3 Op erat ion aliz atio
n of the central variables
3.4 Empirical models
In order to test the hypotheses, a Pearson correlation analysis will be conducted, followed by a multiple linear regression with interaction effects will be run. For the multiple linear regression analysis an appropriate empirical model specification is needed, therefore the following models have been constructed:
Equation 1 (Base model 1):
Equation 2 (Model 2): 𝐶𝐶𝐶𝐶𝐶𝐶 = 𝛼𝛼 + 𝛽𝛽0𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷 + 𝛽𝛽1𝐼𝐼𝐷𝐷𝐼𝐼𝐷𝐷 + 𝛽𝛽2𝐼𝐼𝐷𝐷𝐷𝐷𝐼𝐼 + 𝛽𝛽3𝐶𝐶𝑃𝑃𝑃𝑃𝐷𝐷 + 𝛽𝛽5𝐷𝐷𝐼𝐼𝑆𝑆𝑆𝑆 + 𝛽𝛽6𝑅𝑅𝐼𝐼𝐷𝐷𝑅𝑅 + 𝛽𝛽7𝐼𝐼𝐷𝐷𝐼𝐼𝐷𝐷 ∗ 𝐷𝐷𝐷𝐷𝐷𝐷𝐼𝐼𝐷𝐷𝐷𝐷 + 𝛽𝛽8𝐼𝐼𝐷𝐷𝐷𝐷𝐼𝐼 ∗ 𝐷𝐷𝐷𝐷𝐷𝐷𝐼𝐼𝐷𝐷𝐷𝐷 + 𝛽𝛽3𝐶𝐶𝑃𝑃𝑃𝑃𝐷𝐷 ∗ 𝐷𝐷𝐷𝐷𝐷𝐷𝐼𝐼𝐷𝐷𝐷𝐷 + 𝜀𝜀 Equation 3 (Model 3): 𝐶𝐶𝐶𝐶𝐶𝐶 = 𝛼𝛼 + 𝛽𝛽0𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷 + 𝛽𝛽1𝐼𝐼𝐷𝐷𝐼𝐼𝐷𝐷 + 𝛽𝛽2𝐼𝐼𝐷𝐷𝐷𝐷𝐼𝐼 + 𝛽𝛽3𝐶𝐶𝑃𝑃𝑃𝑃𝐷𝐷 + 𝛽𝛽5𝐷𝐷𝐼𝐼𝑆𝑆𝑆𝑆 + 𝛽𝛽6𝑅𝑅𝐼𝐼𝐷𝐷𝑅𝑅 + 𝛽𝛽7𝐼𝐼𝐷𝐷𝐼𝐼𝐷𝐷 ∗ 𝐷𝐷𝐷𝐷𝐷𝐷𝐼𝐼𝐷𝐷𝐷𝐷 + 𝛽𝛽8𝐼𝐼𝐷𝐷𝐷𝐷𝐼𝐼 ∗ 𝐷𝐷𝐷𝐷𝐷𝐷𝐼𝐼𝐷𝐷𝐷𝐷 + 𝛽𝛽3𝐶𝐶𝑃𝑃𝑃𝑃𝐷𝐷 ∗ 𝐷𝐷𝐷𝐷𝐷𝐷𝐼𝐼𝐷𝐷𝐷𝐷 + 𝑣𝑣 ∑4 𝛿𝛿𝛿𝛿𝑃𝑃𝑆𝑆𝐷𝐷𝑅𝑅 𝑗𝑗=1 𝛿𝛿 + 𝜀𝜀 Where:
- CFP: is the dependent variable for corporate financial performance, proxied by EBIT, ROA, ROE, and Tobin’s Q.
- 𝛼𝛼 & 𝛽𝛽i are regression coefficients, and 𝜀𝜀 is the residual.
CSR related variables:
- DSJINA is the dummy variable for the company’s inclusion in the Dow Jones Sustainability Index North America
Company’s assets related variables: - INTA is a proxy for intangible assets - INNO is a proxy for innovation - PHYS is a proxy for physical assets - SIZE is a proxy for firm size
- RISK is a proxy for firm (financial) risk
- YEAR is the dummy variable for the years 2011 - 2014 (2014
The proposed regression model is consistent with previous studies that were investigating the relationship between CSR and CFP (Waddock & Graves, 1997; López et al., 2007, McWilliams & Siegel, 2000), however this research includes additional variables to extend the scope of the research to company’s tangible and intangible assets, to incorporate the natural-resource based view of the firm.
4. Results
The previous chapter provided the methodology and research design, which enabled the empirical testing of the hypotheses. In this section, the results of the empirical analysis will be discussed. A multiple linear regression analysis with interaction effects was conducted in order to test the hypotheses derived from the theory. This chapter is structured as follows. First, to provide insight into the data, the descriptive statistics and Pearson correlation analysis are presented. Then, the results of the multivariate linear regression analyses are
provided in order to examine the existing relationships between the variables.
4.1 Descriptive statistics and Pearson correlation matrix
Table 4 represents the descriptive statistics for all variables included in this study. The total amount of firm-year observations (N) is the same, 234 from 2011 to 2014, as missings were eliminated from the analysis. Furthermore, the Pearson correlation matrix is also integrated into Table 4 and describes the correlation between the dependent, independent and control variables included in this study.
Table 4 Descriptive statistics and Pearson correlation matrix
As the dependent variables, EBIT, ROA, ROE are in millions, the means for the sample period are 951.6364, .0276, .0602 respectively. The descriptive statistics show that the CSR measurement (DJSI inclusion) is a dummy variable which results in that 39% (.39) of the companies is included in the Dow Jones Sustainability Index North America (i.e. highly-CSR engagement). According to the normality assumptions of parametric statistical tests, some centered variables have been created due to missings, and outliers were preventively eliminated and transformed into logarithms in order to maintain a normal distribution. There will be elaborated on that further in this report.
According to the correlation test results, companies that are highly engaged in CSR, seem to have a positive significant correlation with financial performance based on EBIT (r = .531 at the 0.1 significance level). This is in line with the expectations of this study.
Furthermore, CSR-engagement seems to have a positive significant correlation with a company’s intangible assets and innovation (r = .475 & .620 at the 0.1 significance level respectively) However, regarding the independent variables, only the level of innovation is significantly correlated with EBIT at a 0.1 level. The independent variables are not exceptionally related amongst each other, only intangible assets show a significant correlation with all other variables but CFP measures.
In order to check for the multicollinearity issue, the variance inflation factor (VIF) has been measured. According to multiple scholars (e.g Stine, 1995; O’brien, 2007; Venta, Velez, & Lau, 2016) the rule of thumb associated with VIF, is less than 10. Being presented in Table 5, all values lie within the range of 10 with a maximum VIF of 6.39 and a mean VIF of 2.27.
4.2 Statistical analyses
In the previous paragraph, as a first indication, a significant correlation is found between all CSR-involvement and CFP. Nevertheless, a multivariate linear regression analysis was conducted to study the interaction effect on the relationship between the independent and dependent variables. Firstly, model 1 was created by the control variables (firms size, firm risk, and years) which were entered in the analysis to control for the effect on CFP other than from the independent variables. Secondly, model 2 focused on the relation between the independent and dependent variables. Thirdly, all variables were included in model 3 together with the interaction effects. In all analyses for the different CFP measures, the last model (3) had the highest amount of explained variance, which will therefore mainly be used to discuss the results. Furthermore, in order to test the hypotheses developed, model 3 contains the data needed, which can then be compared with model 2 (to determine the interaction effect). However, in order to create a clear picture, it has been chosen to focus on one accounting-based measure for CFP and one market-based measure for CFP: EBIT and
Tobin’s Q respectively. The regression analyses based on return on assets (ROA) and return on equity (ROE) can be found in Appendix 1 and Appendix 2, respectively.
4.3.1 Earnings Before Interests and Taxes (EBIT)
In order to test the proposed hypotheses, first a regression analysis for variables predicting EBIT was conducted. According to Table 5, adding more variables into the regression analysis improved the amount of explained variance. In model 3, all independent variables explain 75.4% of the variance in the dependent variable, CFP.
Table 5 Summary of linear regression analysis for variable predicting EBIT (N = 234)
Generally stated, there seems to be a significant relation, albeit moderately, between firm size, intangible assets, physical assets, CSR, and CFP. According to these results, based on model 2 and model 3, the strongest significant positive relationship is found between CSR and CFP, in which is β = .185 & β = .691, p = .000. Thus, there can be stated that proxied by EBIT in this analysis, high-CSR engagement will positively influence CFP. Moreover, when looking at the unstandardized coefficient of CSR, 2541.002, this means that companies included in the Dow Jones Sustainability Index North America, on average have around $2541 billion higher CFP (based on EBIT).
When looking to independent variables, there seems to be a significant positive relation between the company’s intangible assets and CFP. However, when paying attention to the relationship between a company’s assets and CFP, the moderating effect of highly-CSR engagement shows that the significant correlation between intangible assets and CFP is still positive (β = .409, p = .000). In other words, when companies are engaged in CSR practices, this will improve the effect of intangible assets on CFP (proxied by EBIT). In contrary, physical assets are significant negatively related to a company’s financial performance and this will be even worse when companies are involved with CSR
According to all three models, firm size is also significant positively related to CFP (β = .277/.602/.277, p = .000), so this influences the overall relationship between highly-CSR engaged companies and increase in CFP as bigger companies might be able to spend more money on CSR-related activities. In addition, the unstandardized coefficient suggests that an increase of 1 (factor e, natural logarithm), approximately $280.396 billion will be added to the EBIT amount.
A significant regression equation for model 3 was found (F(11,153) = 18.282, p < .000), with an R2 of .537. The R2 changes significantly positive in every model (.375, .440, and .537 respectively).
As Table 5 represents the regression analysis of CFP proxied by EBIT, the other accounting-based measures (return on assets and return on equity) can be found in Appendix 1 (return on assets) and 2 (return on equity).
4.3.2 Tobin’s Q
When using Tobin’s Q as the dependent variable, table 6 shows that there is no significant relationship between high-CSR engaged companies and CFP proxied by Tobin’s Q (β = .685/ .176, p= .003), indicating that oil and gas companies that are involved in social and environmental awareness, do not have better financial performances.
Table 6 Summary of linear regression analysis for variable predicting Tobin’s Q (N = 234)
The final regression model with respect to the prediction of Tobin's Q was based on all above mentioned variables plus interaction effects in order to examine the influence of CSR-engagement on the effect of a company’s tangible and intangible assets and CFP. The effect of the tangible and intangible assets on CFP does not result in any significant relationship. However, regarding the focus of this study, the moderating effect of CSR engagement has been integrated to see whether a significant interaction effect can be identified. Based on model 3, the effect of a company’s tangible and intangible assets is not influenced by the company’s involvement in CSR. A significant regression equation of model 3 was found (F(11,153) = 1.347, p < .003), with an R2 of .088. This last model had with 8.8% the highest
amount of explained variance.
4.3.3 Return on Assets (ROA) & Return on Equity (ROE)
According to ROA as a dependent proxy for CFP, the regression analysis was also performed first to test whether the hypothesis was indicative of a positive relationship between CSR and CFP. The result for in CSR-engaged companies is like proxied by ROE, no significant relationship with CFP. The firm (financial) risk of a company is significant negatively related to CFP measured by ROA and ROE. In order to test whether high-CFP engaged companies
will positively influence the effect of a company’s physical assets on CFP, the interaction effect was created. The first regression analysis showed a significant negative effect on ROA and ROE (β = -.253 and β = -.187, significantly). According to the R-squares of both proxies, they both have the lowest amount (ROA: R2 of .168 & ROE: R2 of .121) of explained variance
even after adding all independent variables to model 3. Based on these two variables, only a significant relation can be identified regarding the negative influence of CSR engagement upon the impact of a company’s physical assets on CFP. Both regression analyses can be found in Appendix 1 (ROA) and 2 (ROE).
Based on the analyses provided, the results are summarized in Table 7.
