The impact of mandatory on-financial reporting on ESG
performance
Name: Adriana Botman Student number: 10094245 Thesis supervisor: dr. A. Sikalidis Date: June 24, 2018
Word count: 15791
MSc Accountancy & Control, specialization Control
2
Statement of Originality
This document is written by student Adriana Botman 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.
3
Abstract
As of July 2017, the non-financial reporting directive has been put into law in 22 countries in the EU, however there are countries that already have mandatory non-financial regulation put in pace. This thesis examines the effects of implementing non-financial reporting regulation on the ESG performance (environmental, social, corporate governance) of firms. Furthermore, this paper also adds to existing literature by examining the benefits of improving the ESG scores of a firm, mainly its profitability. The findings of this paper are mixed. Treated countries show no improvement in ESG scores after implementation, however the ESG scores of treated countries are significantly higher than that of non-treated countries. Furthermore, the results of this research show that the improvement of ESG scores has a positive effect on the profitability of the firm. The results are proven through empirical research using firm data from France, Denmark, South Africa and the USA from 2006-2017.
4
Contents
1 Introduction ... 6
2 Literature review ... 8
2.1 Defining sustainability, non-financial reporting and ESG ... 8
2.2 History of non-financial disclosure, ESG practices and regulations ... 9
2.2.1 Voluntary regulation ... 10
2.2.2 Denmark, South Africa & France ... 10
2.3 Motivations for improving ESG practices ... 11
2.4 Motivations for non-financial disclosure ... 13
2.5 Consequences of mandatory disclosure ... 13
3 Hypothesis development ... 15
3.1 ESG performance ... 15
3.2 Environmental and social performance ... 16
3.3 Governance performance ... 17
3.4 Economic performance ... 18
4 Research design ... 20
4.1 Data and sample ... 20
4.2 ESG Performance... 21 4.2.1 Dependent Variable ... 21 4.2.2 Independent Variable ... 21 4.2.3 Control Variables ... 21 4.2.4 Models... 23 4.3 Economic performance ... 24 4.3.1 Dependent Variable ... 24 4.3.2 Independent Variable ... 24
5 4.3.3 Control Variables ... 24 4.3.4 Models... 25 5 Empirical Results ... 27 5.1 ESG Performance... 27 5.1.1 Descriptive statistics ... 27 5.1.2 Multicollinearity ... 28 5.1.3 Tests of hypotheses ... 29 5.2 Economic performance ... 36 5.2.1 Descriptive statistics ... 36 5.2.2 Multicollinearity ... 37 5.2.3 Tests of hypothesis ... 38
6 Conclusion and Discussion ... 42
6.1 Conclusion ... 42 6.2 Interpretation of results ... 43 6.2.1 Validity of research ... 43 6.2.2 ESG Performance... 43 6.2.3 Economic performance ... 46 6.3 Contributions... 46 6.4 Limitations ... 47 6.5 Recommendations ... 47 References ... 49
Appendix 1: Predictive validity framework ... 53
6
1 Introduction
The main purpose of this paper is to investigate whether the implementation of regulation for the disclosure of non-financial information like corporate social responsibility (CSR) reporting will lead to an in increase in the environmental, social and governance (ESG) performance of companies impacted. Additionally, this paper will research whether a higher ESG performance will lead to an increase in firm performance, giving this paper a political and practical context.
The companies in countries who have implemented non-financial disclosure regulations (South Africa, Denmark, and France) will be studied. Their ESG performance before and after the regulation was implemented (treated countries) will be compared and ESG performance with similar companies in countries without this regulation will also be compared with the sample from South Africa, Denmark, and France. This will be done through a quantitative, multivariate analysis. The results will be used to attempt to answer the following research question:
RQ: Does the implementation of mandatory non-financial disclosure have an impact on the ESG performance of companies?
As of July 2017, the non-financial reporting directive has been put into law in 22 countries in the EU (“The Ultimate Guide to the Non-Financial Reporting Directive | Datamaran,” n.d.). This research question has importance because it will provide information concerning the consequences of implementing non-financial disclosure regulation and how this affects company performance, specifically ESG performance.
Evidence has been found that by making non-financial reporting mandatory, companies are more likely to disclose more information voluntarily so as to differentiate themselves from other companies and for comparability (Ioannou & Serafeim, 2017). It has also been theorized that firm size, visibility, resource access, and scale of operations are all firm attributes that influence a company’s decision to participate in CSR reporting (Udayasankar, 2008). However, there has been little to no prior research on whether making non-financial reporting mandatory will motivate a company to change their ESG strategy, despite the fact that one of the aims of implementing these regulations are to incentivize companies to improve their practices (Ioannou & Serafeim, 2017).
7
A reason for why there has been little research done on the effects of making non-financial reporting mandatory could be because very few countries have implemented regulation before the EU ‘Non-financial Reporting Directive’ and the countries that already had regulation have not had it for a very long time, so there has been little opportunity to compare results.
The answer to this research question is especially important for lawmakers who are considering making non-financial reporting mandatory as well as activists who are lobbying for non-financial reporting disclosures to be put into law, because this will help prove (or disprove) the impact of such regulation.
This paper will also add to the current literature by determining whether a higher ESG score will have an impact on firm financial performance. This will also be done with a multivariate analysis. Researching this is relevant because even though the relationship between ESG performance and financial performance has been researched for decades, there is still no definitive answer on its impact (Tang, Hull, & Rothenberg, 2012; Bassen & Kovacs, 2008). To research this impact is important because if results show that and improvement in ESG scores benefits firm performance, firms will also gain practical motivations to improve their ESG performance.
Therefore, should it be concluded that mandatory non-financial reporting improves ESG scores and that improved ESG scores improve firm performance, it will make the implementation of mandatory non-financial reporting even more beneficial towards both the environment and society, as well as the organization, making it a win-win situation.
This research paper consists of five sections. The first section provides a literature review concerning past academic research concerning mandatory non-financial reporting and its history, and information about ESG scores, corporate sustainability, and its determinants. The next section will describe the hypothesis development of this paper. Afterwards, the research design will be described, followed by the next section providing empirical results of the research. Finally, the results will be summarized and discussed.
8
2 Literature review
2.1 Defining sustainability, non-financial reporting and ESG
When referring to non-financial reporting, this paper is generally referring to corporate social responsibility reporting. The following paragraphs will explain what is considered corporate social responsibility.
There is much overlap and congruence between CSR (corporate social responsibility) and CS (corporate sustainability) in management literature, with no definitive difference between the two (Montiel, 2008). The European Commission defines CSR as “the responsibility of enterprises for their impacts on society” (European Commission, 2011). However, one of the most cited definitions of CSR is from Carroll (1979): “the social responsibility of business encompasses the economic, legal, ethical, and discretionary expectations that society has of organizations at a given point in time” (Carroll, 1979).
In 1987, the World Commission on Environment and Development defined corporate sustainability, using the term sustainable development, as when companies’ needs could be met “without compromising the ability of future generations to meet their own needs” (Brundtland, 1987). As the term grew more popular in the 1990s, the definition of corporate sustainability grew into two separate strains by scholars, one of them focused on ecological
sustainability, while the other broadened corporate sustainability into three dimensions:
environmental, economic, and social (Montiel, 2008).
In his paper defining the concepts of CS and CSR, Montiel (2008) explains that the main overlap between the two expressions are that both concepts have three dimensions: economic, social, and environmental. However, the interpretation of these dimensions is also how CS and CSR differ most. With CS, these dimensions are usually interpreted as being interconnected, while with CSR, each dimension is interpreted as being an independent component. Montiel (2008) further explains that despite there being many quantitative studies concerning the correlation between the dimensions, no definitive conclusion can be made (Montiel, 2008).
This difference in the interpretation of dimensions becomes especially evident in the interpretation of the economic dimension. When defining CSR, Carroll (1979) describes social responsibility as being a supplement to the economic dimension, whereas Bansal
9
(2005) describes in his paper that the social, environmental, and economic elements are complementary, and therefore that these elements integrated together create sustainability (Bansal, 2005; Carroll, 1979).
Montiel (2008) argues that over time, the differences between CSR and CS has become trivial, seeing as both of the concepts claim that the three elements of economic, environmental and social all are part of the definition of sustainability, and that whether these elements ought to be integrated or separated are unimportant, considering that both concepts share a common goal of achieving sustainability with these elements (Montiel, 2008).
This theory is further confirmed in the study by Hahn & Kühnen (2013). They claim that the definitions of corporate sustainability and corporate social responsibility are slowly merging to become one homogenous approach (Hahn & Kühnen, 2013). Therefore, similarly to the paper of Hahn & Kühnen (2013), this paper will consider there to be no difference between CS and CSR, thus will use the terms interchangeably.
Sustainability reporting is a natural progression of corporate sustainability. Due to stakeholder theory, which will be further explained in a later section of this paper, organizations will want to formally disclose their progress to their stakeholders.
When measuring the sustainability of a company, a firm’s environmental, social, and governance performance (ESG) can be analyzed (Bassen & Kovacs, 2008). For each aspect of ESG, certain KPIs are measured to determine the level of success a firm has in ESG performance. For instance, an element of scoring environmental performance is to measure a firms CO2 emissions or total waste. To score social performance, diversity within the firm can be measured. To score governance performance, litigation payments can be analyzed (Bassen & Kovacs, 2008).
2.2 History of non-financial disclosure, ESG practices and regulations
The thesis will give a brief history of the development sustainability reporting and the development of government reporting regulations in the world and specifically the countries that will be examined. Academic papers will be used to explain the developments of non-financial disclosure practices and regulations. The thesis will focus specifically on the development of these practices in the countries that are being studied.
10
2.2.1 Voluntary regulation
Starting in the 1960s, people became more aware of the impact and responsibility organizations had towards the environment and society, however at this point there was no governmental pressure or regulation for organizations to respond. Nevertheless, because of this increased awareness, organizations became pressured by outside stakeholders to disclose certain sustainability practices and results. These disclosures became the first voluntary sustainability reports, which mainly started in France and the Netherlands (Ioannou & Serfeim, 2017).
Starting in the 1980s, 'negative screening' was put into practice. 'Negative screening' is an investment approach where investment firms excluded certain firms from their investment portfolios based on the firm's ethical and social performance.
In 1989, environmental reporting guidelines were first introduced by the Coalition for Environmentally Responsible Economies (CERES), called the 'CERES/Valdez Principles. In 1997, CERES also introduced the Global Reporting Initiative (GRI) together with the United Nations Environmental Program (UNEP). The GRI gave reporting guidelines for economic, environmental, and social performance of firms, called the 'triple bottom line'. Voluntary sustainability reporting practices grew further in the 1990s due to an increased demand from society for more transparency and accountability.
Recently, due to the global financial crisis and increased publicity of corporate scandals, as well as the increasing social and environmental challenges, society has become even more demanding of corporations to take part in sustainability practices and reporting, but society is also less trusting of corporations to self-regulate. At the same time, investors have recently been implementing ESG scores to value companies, which also led to an increased demand for sustainability reporting (Ioannou & Serafeim, 2015).
Because of these developments, some countries began to make the disclosure of ESG data required by implementing laws, regulations, or stock listing requirements.
2.2.2 Denmark, South Africa & France
In 2008, Denmark implemented an amendment to the Danish Financial Statements Act. This act required that large companies needed to report on their corporate social responsibility, starting in the year 2009, if they satisfied two out of the three following criteria: 1. have total
11
assets of more than DKK 143 million, 2. have net revenue of DKK 286 million, or 3. have 250 full-time employees. In the amendment, CSR was defined as:
"voluntary include considerations for human rights, societal, environmental and climate conditions as well as combatting corruption in their business strategy and corporate activities."
Although it is not required in Denmark to implement these policies, it is required to disclose that they do not adopt these practices in their report.
In 2010, South Africa implemented regulation that made the disclosure of sustainability performance of companies required and also implemented regulation that required the disclosure of integrated reporting starting in the year 2011. Integrated reporting differs from a sustainability report from Denmark. In an integrated report, sustainability information is mixed with financial information, instead of separately reported (Ioannou & Serafeim, 2017). This is done by reporting the value creation process of a company, meaning that companies will have to disclose their impact on the environment and various stakeholders.
Similarly to Denmark, France implemented regulation that companies with over 500 employees must issue a yearly report concerning social and environmental issues, starting in the year 2013. There are 42 categories concerning social, environmental and societal issues that the companies must report on (Kaya, 2016).
What the different regulations from these countries all have in common is that there are no set guidelines and standards given on how to report ESG policies. However, if no sustainability data is disclosed, these countries must explain why this information is withheld.
2.3 Motivations for improving ESG practices
There are many different motivations for a firm to choose to invest in improving ESG practices in their firm strategy. According to Udayasankar (2008), motivations for firms to improve their ESG practices can either be strategic or altruistic. Strategic motivations for companies to improve their ESG will be explained.
One of the reasons that companies would want to improve their ESG practices, is that many investors nowadays use ESG performance indicators to determine the longevity of a business, because they believe that a good sustainability and corporate governance score leads to the creation of long-term value for a company (Kocmanová & Dočekalová, 2012). Studies have
12
shown that participation in sustainability practices often results in an improved company performance, one of these results being improved stakeholder relations and therefore the reduction of the company’s business risk (Udayasankar, 2008).
Artiach et al. (2010) goes further into determinants for why firms would invest in improving their corporate sustainability performance. One of their determinants is to gain competitive advantage. This relates to the stakeholder theory. Because of public awareness of the impact that firms have on society, firms want to ensure that they maintain a good relationship with stakeholders, which is anyone who can affect or is affected by a firm obtaining their objectives, so as to keep having access to resources. Artiach et al. (2010) argues further that evidence has shown that these pressures do result in firms acting in a more sustainable way. Other arguments that Artiach et al. (2010) give for why firms would invest in sustainability practices are to maintain or enhance the reputation of a firm, to forestall stakeholder action and to appeal to the socially responsible investors.
The legitimacy theory, which is related to the stakeholder theory, also could explain why certain firms choose to improve their ESG practices. According to Suchman (1995), legitimacy can be defined as followed:
“Legitimacy is a generalized perception or assumption that the actions of an entity are desirable, proper, or appropriate within some socially constructed systems of norms, values, beliefs, and definitions (Suchman, 1995, p. 574)”
According to Patten (1992) legitimacy theory is closely related to the concept of a social contract, which entails that all social institutions operate via a social contract, if survival and growth are based on the delivery or distribution of desirable benefits to society or groups from which it gets its power. Neither of these are permanent, which is why an institution must constantly prove that society needs these benefits, and those benefiting have society’s approval.
Environmental performance is becoming a large aspect of gaining legitimacy. This is because society finds the impact that companies have on the environment increasingly important. Organizations face more pressure to show the efforts they make in sustainability, otherwise they risk losing organizational legitimacy (Mobus, 2005). Therefore, keeping, gaining and repairing legitimacy is an important motivation in improving and disclosing ESG performance.
13
Visibility of a firm is likely to be another determinant of ESG performance (Mobus, 2005; Udayasankar, 2008). Firms with a high visibility face more scrutiny from outside parties and regulators. If firms with a high visibility wish to keep their legitimacy, they might choose to improve their ESG performance.
2.4 Motivations for non-financial disclosure
Here, the motivations for companies to disclose non-financial information will be explained. By using articles from Gamerschlag, Möller, & Verbeeten (2011) and Hahn & Kühnen, (2013), the determinants of why companies would choose to implement voluntary sustainability reporting will be summarized.
Legitimacy theory and signaling theory are likely motivations for a firm to want to disclose non-financial information. To gain, keep, or repair their legitimacy, a firm will want to disclose their non-financial information to prove that they are keeping up with societal norms (Kaya, 2016; Mobus, 2005). This disclosure is related to signaling theory, which occurs when there is information asymmetry between two parties (society and the company), and one party wishes to disclose information to gain a certain advantage (legitimacy) (Hahn & Kühnen, 2013). Society usually doesn’t know how a firm performs on CSR aspects, therefore, a company will want to ‘signal’ (disclose) this information, so as to show that they are performing well or better than is expected, thus gaining or keeping its legitimacy.
Stakeholder theory could also explain why firms would want to voluntarily disclose non-financial information. Under the stakeholder theory, companies respond to pressures and desires from their most powerful stakeholders (Kaya, 2016). Companies are most likely to listen to these powerful stakeholders because the most powerful stakeholders have power of the resources that companies require. Therefore, companies respond to these stakeholders as to retain access to the necessary resources. Recently, sustainability has because important to society and powerful stakeholders, which has caused companies to disclose more sustainability information to appease them (Hahn & Kühnen, 2013).
2.5 Consequences of mandatory disclosure
There are some arguments against the implementation of mandatory non-financial reporting. According to Brown, Jong, & Levy (2009) there are those that fear that there are not enough enforcement mechanisms for credible assurance practices and standards. Also, there are fears
14
that mandatory non-financial reporting would lead to this kind of reporting becoming a routine checklist, and therefore innovative potential would be lost (Brown, Jong, & Levy, 2009).
However, Mobus (2005) uses the legitimacy theory to argue that mandatory non-financial are a source of pressure for firms to comply with ESG performance norms from society.
Additionally, Ioannou & Serafeim (2017) find that the implementation of mandatory sustainability reporting will lead to an increased level of disclosure, meaning increased transparency and comparability between companies. Furthermore, according to their research, the implementation of this regulation has a positive economic effect. Therefore, the implementation of mandatory non-financial reporting has many positive effects regardless of whether it improves ESG performance or not.
15
3 Hypothesis development
3.1 ESG performance
According to Ioannou & Serafeim (2017), the main goals of implementing mandatary nonfinancial disclosure regulation is to improve transparency and thus decrease information asymmetry as well as improving environmental, social, and governance standards.
As mentioned earlier, there is already evidence that mandatary nonfinancial reporting will lead to more disclosure from these companies, which in turn leads to an increase in transparency.
Although no empirical evidence has been found that these regulations will lead to an improved ESG performance from the affected companies, it has been theorized that one of the motivations behind improving ESG performance is pressures from business drivers, like customers, suppliers and competitors (Delmas & Toffel, 2008).
With most regulations implemented so far, companies must disclose certain nonfinancial information or explain why they refuse to disclose (Ioannou & Serafeim, 2017). This will lead to increased transparency and will increase the customers’ and investors’ ability to compare the affected firm with other firms’ ESG performance. One of the key motivators for increased disclosure is the desire for firms to differentiate themselves with other firms and social pressure.
There are a few theories that are linked to motivating a company to invest in increasing their ESG performance, according to Hahn & Kühnen (2013). First of all, there is the legitimacy theory. A company will try to gain legitimacy for a license to operate. This legitimacy is threatened when a firm is determined not to be operating in an acceptable manner according to society (Hahn & Kühnen, 2013). When a company is required to disclose certain non-financial information, it may be revealed that they are not operating in an acceptable manner concerning ESG aspects and by refusing to disclose certain information it may be interpreted by society that they have something to hide and therefore lose their legitimacy. To avoid this, firms will want to retain or improve their legitimacy by improving their ESG performance. The second theory that is a determinant of improving ESG performance is stakeholder theory, which is linked to legitimacy theory. Summarized, stakeholder theory states that organizations should take the needs of all their stakeholders, and not only their shareholders,
16
into account (Artiach et al., 2010; Hahn & Kühnen, 2013). Companies are most likely to adhere to the demands of their most powerful stakeholders (Kaya, 2016). As mentioned in section 2.3, to retain access to resources, organization need to cater to these powerful stakeholders. This is of importance to ESG performance because nowadays, sustainability trends are quite important to society (Hahn & Kühnen, 2013). For organizations to retain or gain access to necessary resources they need to show that they are adhering to these stakeholder desires. This leads to the next relevant theory; signaling theory.
Signaling theory is about how in a situation with information asymmetry, the information-holding party will try to convey this information to obtain some sort of advantage (Hahn & Kühnen, 2013). As explained with legitimacy theory and stakeholder theory, to appease society and to gain competitive advantage by attracting sustainability focused investors, organizations will want to ‘signal’ that they have a superior sustainability performance. When mandatory disclosure comes into play, all of their ‘competitors’ will disclose certain sustainability information. To compete and differentiate themselves companies may want to improve their ESG performance so that they can ‘signal’ that their performance is even better than other organization, and that they are the right investment.
Lastly, Hahn & Kühnen (2013) mention institutional theory as being a determinant for improving ESG performance. This theory states that organizations choose their corporate activities based on institutionalized expectations of the environment and not necessarily from business strategy. Therefore, ESG performance could increase after mandatory non-financial disclosure measures are put into place because it is part of the institutionalized expectations of the environment, and not necessarily because it is logical on a business sense.
These four strategies put together leads to the following hypothesis:
H1: The implementation of mandatory non-financial reporting will lead to an increased ESG performance of affected firms.
3.2 Environmental and social performance
The hypothesis, H1, has been split up into three hypotheses, because ESG can be split up into three elements: environmental, social, and governance performance. Although there is much overlap between the theory for why mandatory non-financial reporting would lead to an improvement for each element, since they are connected, some theory behind each element is unique, as substantiated below, and it may be that mandatary non-financial disclosure
17
measures have a different effect on each of those aspects, which would be interesting to research. Therefore, it is relevant to study each of these elements separately and as a whole.
H1a: The implementation of mandatory non-financial reporting will lead to an increase in environmental performance of affected firms.
H1b: The implementation of mandatory non-financial reporting will lead to an increase in social performance of affected firms.
3.3 Governance performance
Mandatory non-financial reporting may directly and indirectly lead to improved corporate governance performance.
First of all, how mandatory non-financial reporting may indirectly improve corporate governance performance will be explained. Because the responsibility of the board of the directors is to ensure that the organization meets their objectives, it is often assumed that the shareholder’s interests are prioritized over other stakeholders. However, in their paper, Blair & Stout (1999) explain that the board of directors do not solely serve the interests of the shareholders, but act as ‘independent hierarchs’, who are responsible for the interests of the entire corporation, which include all stakeholders, like employees, creditors, and local communities (‘joint welfare’). According to them, this is especially prevalent during situations where ‘team production’ is required. ‘Team production’ occurs when members of an organization need to make company specific investments to improve the collective results of the entire organization. In such a situation, shareholders and stakeholders will be willing to relinquish control to a ‘mediating hierarchy’, which is often the board of directors. In such cases, the board of directors might choose to prioritize the interests of the other stakeholders, not only because this might eventually lead to long-term benefits for the shareholders, but because it is important to the ‘joint welfare’, which is arguably the responsibility of the board (Blair & Stout, 1999).
Eccles, Ioannou, and Serafeim (2014), support the ‘team production’ theory in their paper. They further argue that sustainability is an important aspect of ‘joint welfare’, and therefore under those circumstances, the board of directors are more likely to follow sustainability goals. In their paper, they further theorize that when environmental and social objectives are important to an organization, they are more likely to have responsibility over these issues,
18
meaning that companies that perform highly on sustainability, tend to have a governance structure that matches benefits the interests of all stakeholders (Eccles & Ioannou, 2014). This means that if the board of directors choose to follow a more sustainable corporate strategy, it is likely that the cause of this is because of the team production theory. Therefore, under those conditions the board of directors are not only more likely to follow a sustainability corporate strategy, but also support the interests of the ‘joint welfare’, and thus follow best practices to improve these conditions. Therefore, if mandatory reporting would lead to an improvement of environmental practices, that would mean that the company has enlarged their CSR strategy, indicating that the governance structure is more focused on the interests of all stakeholders, leading to an improved corporate governance.
This leads to the following hypothesis:
H1c: The implementation of mandatory non-financial reporting will lead to an increase in governance performance of affected firms.
3.4 Economic performance
In their paper, Al-Tuwaijri, Christensen, & Hughes (2004) found that there was a positive relation between the improvement of environmental performance and economic profitability. This is likely due to and improvement in environmental performance leads to an improvement in efficiency (Al-Tuwaijri, Christensen, & Hughes, 2004). According to them, this increase in efficiency is caused because a bad environmental performance, for instance large amounts of pollution, is due to resources not being used completely, inefficiently or ineffectively. Therefore, if environmental performance improves, it is likely that this is because those resources are used more efficiently. This increase in efficiency thus affects firm performance, because improved efficiency and less waste is cost effective. So, if mandatory non-financial reporting would improve ESG performance, it would in turn also improve economic performance.
An improved ESG performance leading to a more effective use of resources, could mean that ESG performance can be interpreted as an intangible asset (Surroca, Tribo, & Waddock, 2010). The research from Surroca et al. (2010) argues that intangible assets are related to financial performance as well ESG performance, claiming that intangible assets are the mediating variable. This is because many firms react to societal pressures to improve their ESG performance by developing intangible assets that will lead to competitive advantage
19
(Surroca et al., 2010). Interestingly, Surroca et al. (2010) theorized that the mediating effect of intangible resources between ESG performance and financial performance is multi-directional, meaning that an improved financial performance could lead to more intangible resources and thus an improved ESG performance, but an improved ESG performance could lead to more intangible resources, which in turn lead to an improved financial performance. This theory is partly applied in this paper, because firm financial performance is used as a control variable for ESG performance, and ESG performance will be an independent variable used for firm financial performance.
Furthermore, an increased focus on sustainability performance will potentially lead to better relationships with stakeholders and therefore better access to capital, which in turns leads to a better firm performance (Artiach et al., 2010; Surroca et al., 2010).
The positive accounting theory states that companies would only partake in the disclosure and practice of certain sustainability activities if it should have positive economic effects for the company (Kaya, 2016). Therefore, if a company improves its ESG practices due to mandatory disclosure, it is likely that it does so because the firm predicts there would be economic advantages. If this prediction is accurate, this would mean that an increase in ESG scores would lead to an increase in economic performance.
This leads to the following hypothesis:
H2: An improvement of ESG scores will lead to an increase in the economic performance of affected firms.
20
4 Research design
In the appendix, there is a ‘Libby box’ (predictive validity framework) to show how the theory of the research question will be operationalized.
4.1 Data and sample
ESG performance will be measured by an ESG rating provided by Thomson Reuters DataStream, using the ASSET4 database, which scores ESG performance aspects of large companies. Various previous academic papers have used ASSET4 for their research and considered it a reliable source for ESG data (Cheng, Ioannou, & Sarafeim, 2014; Eccles & Ioannou, 2014; Ioannou & Serafeim, 2012).
The sample will be taken from three countries with non-financial disclosure regulation: Denmark, South Africa and France. These countries are chosen specifically because their disclosure requirements specifically targets social, environmental and governance areas (Kaya, 2016). The companies selected from these countries are among the largest companies in their respective countries, which would mean that these companies likely have the largest impact on environmental, social, and governance factors, making the results from the sample economically relevant (Ioannou & Serafeim, 2017).
Data from these countries will be taken from 2006, which is at least three years before the non-financial disclosure regulation was implemented and 2017, which is at least four years after the regulation was implemented. In contrast, Ioannou & Serafeim (2017) used only two years after the regulation was implemented, so in this paper two extra years of change will be measured. So, for each country, data will be taken from the following years:
Table 1 - Data collection overview Country Year of non-financial reporting regulation
implementation
Years used for this research
Observations Source
Denmark 2009 (Ioannou & Serafeim, 2017) 2006 - 2017 350 Datastream South Africa 2010 (Ioannou & Serafeim, 2017) 2006 - 2017 1099 Datastream France 2013 (Kaya, 2016) 2006 - 2017 1034 Datastream USA None (King, Bartels, Fogerlberg, Hoballah, &
Van der Lugt, 2016)
2006 - 2017 9052 Datastream
To determine the effect of mandatory non-financial disclosure implementation, treated countries will be compared as well as a comparison between the treated countries Denmark,
21
South African and France, and the United States will be taken as a control country because the USA is a good example of a country where most of the non-financial reporting is voluntary (King, Bartels, Fogerlberg, Hoballah, & Van der Lugt, 2016). The effect of ESG scores on economic performance will be similarly measured, however, data from all the countries (France, Denmark, South Africa, and the USA) will be taken. This paper will use a multivariate analysis to determine these effects.
4.2 ESG Performance
4.2.1 Dependent Variable
ESG performance will be the dependent variable. This will be measured using the ESG ratings collected by the Thomson Datastream ASSET4 database. As mentioned in section three, the average rating for each of the three elements (environmental, social, and governance) will be looked at separately and the combined score will also be used. In the appendix, the data that Thomson Datastream collects and measures to determine ESG scores and its definitions are included.
4.2.2 Independent Variable
The independent variable will be the presence of mandatory non-financial reporting. This will be done through a dummy variable, 1 being the presence of mandatory non-financial reporting (starting from the year of implementation), and 0 being the absence of mandatory non-financial reporting (the years before implementation).
4.2.3 Control Variables
Firm visibility is one of the factors that can influence ESG scores (Udayasankar, 2008). According to her research, firms with high visibility are more likely to implement a sustainability strategy, because if they do not, they will suffer from effects due to a loss of reputation and legitimacy. Firm size is often used as a proxy for firm visibility, because the larger a company is, the higher it’s visibility (Mobus, 2005; Udayasankar, 2008). Therefore, firm size will be used as one of the control variables. Firm size will be measured by taking the natural logarithm of total sales (Ioannou & Serafeim, 2017).
Leverage is another factor that may impact ESG scores (Ioannou & Serafeim, 2017). The stakeholder theory states that a company is responsible to various stakeholders. Companies are most likely to be influenced by their most powerful stakeholders. A stakeholder is
22
powerful if they have control over certain resources that a company needs. Because debtholders supply capital, which is a necessary resource, to the firm, having a large amount of debt means these debtholders have a lot of power over the organization. If that is the situation, it is more likely the organization will focus on the demands of these debtholders, than those of the employees and the community (Artiach et al., 2010), therefore likely having a negative relationship with the ESG score. Leverage will be measured by calculating the total liabilities divided by the total assets (Ioannou & Serafeim, 2017).
Profitability will also be used as a control variable for ESG performance. Opposite to leverage, it is theorized that should an organization have a high economic performance, the debtholders will hold less power over the organization. Furthermore, the company will likely have the financial capacity to listen and act to the demands of stakeholders such as the employees and the community (Artiach et al., 2010). Therefore, if a company performs economically well, the ESG score should go up, while when a company has a low profitability there will be less focus on ESG scores. Profitability will be measured by taking the return-on-assets (Ioannou & Serafeim, 2017; Tang et al., 2012).
Another control variable that will be used to measure the effect of mandatory non-financial regulation is the level of growth options. This variable is relevant because it is theorized that firms who have a relatively high investment in tangible assets are less incentivized to use strategies that include sustainability values because they have no financial room to invest in such strategies (Artiach et al., 2010). Growth options will be measured as the price-to-book ratio (Artiach et al., 2010).
The firm’s liquidity will also be used as a control variable. According to Kocmanová & Dočekalová (2012), a firm’s cash flow is an indicator of a firm’s ESG performance. Liquidity will be measured by taking the cash flow/revenues ratio (Surroca et al., 2010).
First the summary statistics will be presented, where the average difference in means between the variables before and after the implementation of mandatory non-financial disclosure is presented. Afterwards, a correlation matrix will be made to determine whether there is any risk of multicollinearity.
A multiple regression analysis will be used to determine whether mandatory nonfinancial reporting is correlated with a positive change in ESG performance.
23
4.2.4 Models
To determine which models to use to test the hypotheses, the type of data that has been collected needs to be determined. Since the data collected is a combination of cross-sectional and time-series data, the type of data is panel data. The type of panel-data is unbalanced short panel-data because the amount of years of data take in different per cross-section (for some firms only 6 years of data has been collected, for other firms 10 years) and the number of cross-sections is more than the amount of years.
To determine the appropriateness of an ordinary least squares (OLS) regression, tests such as the Hausman test and F-test are performed to determine the goodness of fit of the model (Hausman, 1978). Furthermore, this model is also used in academic research comparable to the research of this paper, such as Bhagat & Bolton (2008), Gamerschlag et al. (2011), and Ioannou & Serafeim (2017). These papers also studied the effects of regulation or ESG data. Therefore, following models will be tested:
Model 1
ESG-performance = + Mandatory + Firm size + Leverage + Profitability + Growth Options + Liquidity +
Model 2
Environmental performance = + Mandatory + Firm size + Leverage + Profitability + Growth Options + Liquidity +
Model 3
Corporate governance performance = + Mandatory + Firm size + Leverage + Profitability + Growth Options + Liquidity +
Model 4
Social performance = + Mandatory + Firm size + Leverage + Profitability + Growth Options + Liquidity +
Model 5
ESG-performance = + Mandatory + Firm size + Leverage + Profitability + Growth Options + Liquidity +
Model 6
Environmental performance = + Mandatory + Firm size + Leverage + Profitability + Growth Options + Liquidity +
Model 7
Corporate governance performance = + Mandatory + Firm size + Leverage + Profitability + Growth Options + Liquidity +
24
Model 8
Social performance = + Mandatory + Firm size + Leverage + Profitability + Growth Options + Liquidity +
Models 1-4 are excluding the USA as a control group and Models 5-8 are including the USA as a control group.
The prediction of H1-H1c is that the coefficient on the Mandatory (the presence of mandatory non-financial disclosure implementation) variable will be positive.
4.3 Economic performance
4.3.1 Dependent Variable
For the second part of this research, the effect that ESG scores have on economic performance will be measured. There are various indicators that can be used to measure financial performance. Those include indicators of liquidity, profitability, indebtedness, financial and asset structure other activities (Chvatalová, Kocmanová, & Dočekalová, 2011). According to Artiach et al. (2010) as well as Al-Tuwaijri et al. (2004), return-on-assets (ROA) can be used to determine a firm's profitability. This ratio is better than using net income because net income doesn't take firm size into consideration. There are many other financial performance measures that can be used, but for the sake of consistency in this research paper, ROA will be used to measure economic performance. It will be interesting to compare the effect that ROA has on ESG performance as well as the effect that ESG performance has on ROA.
4.3.2 Independent Variable
The independent variable will be the ESG score. Similarly to H1, the effect of ESG scores will be looked at with each element separated (environmental, social, and governance) and as a whole (combined score).
4.3.3 Control Variables
R&D is an important factor that could influence ESG-performance and financial performance. Research has shown that R&D investments and ESG investments compete with each other for the organization's limited resources (Tang et al., 2012). Therefore, if either one is high, it is likely the other variable would be low, therefore influencing each other. The
25
R&D investments are calculated by taking the firm's total R&D investments and weighting them by the firm's total assets (Tang et al., 2012).
According to Surroca et al. (2010), size, risk, industry, country, and year can all impact the ROA, and therefore can be used as control variables for measuring the effect on financial performance. The same calculation to measure firm size will be used as with H1, the natural logarithm of the total sales. Financial risk will be measured using the firm's beta. Industry, country, and year can be controlled for by taking the mean values of the dependent variable, which is the ROA, for the corresponding industry, country, and year.
Furthermore, there are tangible resources that can have an effect on firm performance, which will also be controlled for. The tangible resources that will be used as control variables are physical resources, leverage, and financial resources (Surroca et al., 2010).
Physical resources can be measured by taking the current assets over the total assets, which is the proportion of long-term assets (Surroca et al., 2010). Leverage will once again be measured by dividing the total liabilities by the total assets. Financial resources, which is the firm's liquidity, will be calculated by taking the cash-flow-to-revenues ratio (Surroca et al., 2010).
First the summary statistics will be presented, where the average difference in means between the variables before and after the implementation of mandatory non-financial disclosure is presented. Afterwards, a correlation matrix will be made to determine whether there is any risk of multicollinearity.
A multiple regression analysis will be used to determine whether ESG performance is correlated with a positive change in financial performance.
4.3.4 Models
Similarly to testing hypotheses H1-c, the regression model that will be chosen will be based on the data collected. The data used for measuring the effect on economic performance is from the same dataset as for measuring the effect on ESG performance, except there are newly added variables to these models. Therefore, the type of data presented is once again unbalanced short panel-data.
Once again, to determine the appropriateness of an ordinary least squares (OLS) regression, tests such as the Hausman test and F-test are performed to determine the goodness of fit of the model(Hausman, 1978). Moreover, same arguments for comparable academic research
26
can be used for these models; comparable academic literature, like that from Bhagat & Bolton (2008), Gamerschlag et al. (2011), and Ioannou & Serafeim (2017) (see above), also use this method.
Therefore, following models will be tested:
Model 9
ROA = + ESG-score + R&D + Firm size + Industry/country/year + Leverage + Risk + Physical resources + Liquidity +
Model 10
ROA = + Environmental-score + R&D + Firm size + Industry/country/year + Leverage + Risk + Physical resources + Liquidity +
Model 11
ROA = + Corporate governance-score + R&D + Firm size + Industry/country/year + Leverage + Risk + Physical resources + Liquidity +
Model 12
ROA = + social-score + R&D + Firm size + Industry/country/year + Leverage + Risk + Physical resources + Liquidity +
The prediction of H2 is that the coefficient on the ESG-score (ESG, environmental, social, corporate governance) variables will be positive.
27
5 Empirical Results
5.1 ESG Performance
5.1.1 Descriptive statistics
In Table 2, the descriptive statistics can be found. The descriptive statistics have been split up into three categories, the statistics for France, Denmark and South Africa from before mandatory non-financial reporting was implemented, the statistics for these countries after mandatory non-financial reporting has been implemented, and the statistics for the USA, which is used as a control country. For the first setting, there are 760 observations, except for the explanatory variable for the liquidity ratio, which has 759. This is due to the deletion of outliers. For the second setting, there are 1723 observations, except for the liquidity ratio and the price to book ratio also due to the deletion of outliers. There are more observations for the second setting because data has been collected for more years in this setting (see Table 1). Using the descriptive statistics, the means between the two settings can already be looked at (treated countries). Compared to a non-mandatory setting, in the mandatory setting there is only a slight increase in the mean of the overall ESG score. This is mostly due to the increase in the mean Corporate Governance score and the mean Social score. Interestingly, the mean Environmental score has decreased in the mandatory setting. In the regression it can be seen whether these differences are significant or not.
Looking at the descriptive statistics of the USA setting, it has by far the largest amount of observations, which is 9052, except for the liquidity ratio and the price to book ratio due to the deletion of outliers. This is due to the fact that the USA has overall more listed companies as well as than France, Denmark, and South Africa put together, as well as observations being used for the entire year range of 2006 to 2017, and not split between mandatory and non-mandatory settings, since the USA only has a voluntary setting.
Interesting, the difference in the mean of the ESG score between the USA setting and the France, Denmark, and South Africa setting is larger than the difference in mean comparing before and after. The Environmental and Social score in the USA setting is smaller than in the France, Denmark, and South Africa setting, however the Corporate Governance score is larger in the USA setting. Therefore, from the descriptive statistics it looks like the comparison between the USA setting and the France, Denmark, and South Africa setting has an opposite effect for the Environmental and Corporate Governance score as when comparing
28
treated countries. From the regression it can be analyzed whether these differences are significant.
Table 2 - Descriptive statistics for H1-c
variable observations mean median sd min p25 p75 max Setting 1 Dependent variables
ESG score 760 69.61 81.75 27.42 3.14 51.40 91.69 97.94 CGV score 760 52.94 55.18 26.55 2.09 30.10 76.49 96.56 ENV score 760 73.11 84.99 25.41 9.82 62.34 92.47 96.94 SOC score 760 73.40 84.31 25.51 5.17 59.34 93.43 98.55 Explanatory variables Firm Size 760 15.86 15.93 1.61 6.76 14.76 17.00 19.56 Leverage 760 0.66 0.63 0.34 0.08 0.50 0.78 3.81 Profitability 760 0.05 0.04 0.09 -0.38 0.01 0.07 1.21 Growth Options 760 2.73 1.62 5.73 -9.50 0.93 2.62 67.83 Liquidity 759 15.19 13.48 14.86 -114.92 7.83 21.54 127.21 Setting 2 Dependent variables
ESG score 1723 70.83 80.67 24.86 3.14 58.86 89.73 96.60 CGV score 1723 61.98 65.45 22.18 1.92 48.51 79.79 97.55 ENV score 1723 66.13 74.14 26.44 8.58 46.84 89.85 95.31 SOC score 1723 74.63 85.23 24.25 4.02 64.61 92.03 97.46 Explanatory variables Firm Size 1723 15.94 16.06 1.65 6.02 14.86 17.03 19.65 Leverage 1723 0.56 0.54 0.24 0.00 0.40 0.70 3.39 Profitability 1723 0.06 0.05 0.09 -0.46 0.02 0.09 0.66 Growth Options 1721 2.10 1.06 3.73 -9.39 0.20 2.31 42.96 Liquidity 1710 16.71 13.40 16.36 -111.84 6.64 23.62 129.20 Setting 3 Dependent variables
ESG score 9052 53.26 51.66 29.60 2.96 25.46 83.49 97.96 CGV score 9052 70.57 75.24 19.55 1.43 59.61 85.60 97.82 ENV score 9052 43.96 33.54 31.43 8.27 13.87 76.83 96.85 SOC score 9052 46.12 42.12 28.89 3.55 18.73 72.89 98.59 Explanatory variables Firm Size 9052 15.26 15.21 1.40 10.13 14.34 16.16 20.03 Leverage 9052 0.63 0.62 0.23 -0.48 0.50 0.76 3.66 Profitability 9052 0.04 0.04 0.09 -2.40 0.01 0.08 1.02 Growth Options 8996 2.82 2.06 6.63 -98.11 1.27 3.43 99.35 Liquidity 9041 19.22 15.31 16.80 -187.41 9.24 26.14 173.79 Setting 4 Dependent variables
ESG score 11535 56.96 59.92 29.66 2.96 28.37 86.66 97.96 CGV score 11535 68.13 73.31 21.10 1.43 56.21 84.64 97.82 ENV score 11535 49.19 45.49 32.00 8.27 15.11 83.19 96.94 SOC score 11535 52.17 53.08 30.33 3.55 22.40 82.44 98.59 Explanatory variables Firm Size 11535 15.40 15.36 1.48 6.02 14.42 16.36 20.03 Leverage 11535 0.62 0.61 0.24 -0.48 0.48 0.75 3.81 Profitability 11535 0.04 0.04 0.09 -2.40 0.01 0.08 1.21 Growth Options 11477 2.71 1.90 6.23 -98.11 1.10 3.25 99.35 Liquidity 11510 18.58 14.84 16.66 -187.41 8.85 25.48 173.79
Setting 1 = Non-mandatory setting for Denmark, France and South Africa. Setting 2 = Mandatory setting for Denmark, France, and South Africa. Setting 3 = USA setting. Setting 4 = total descriptive statistics for all setting combined.
5.1.2 Multicollinearity
Before running the regression, a correlation matrix is made to determine the collinearity between the explanatory variables. As is evident from Table 3, the highest Pearson’s correlation is 0,1734 between Leverage and Firm Size. Since this number is lower than the
29
cutoff of 0.5 and the VIF is less than 2.5, it can be determined that there is no multicollinearity between the control variables, meaning they can all be used and included as separate unique variables in the regression.
Table 3 - Correlation Matrix H1-H1c
Firm Size Leverage Profitability Growth Options Liquidity
Firm Size 1 Leverage 0.173*** 1 Profitability 0.092*** -0.173*** 1 Growth Options -0.012 -0.045*** 0.094*** 1 Liquidity -0.247*** -0.084*** 0.092*** 0.006 1 * p<0.05, ** p<0.01, *** p<0.001 5.1.3 Tests of hypotheses
The first four models are used to determine the significance of the effect of implementing mandatory non-financial reporting on ESG performance when comparing the selected countries, Denmark, France and South Africa before and after the implementation (treated countries). In Table 4 the results of the regressions can be found. Considering that a large dataset with panel-data is involved, the results from the regressions are corrected for heteroscedasticity and unwanted correlation by clustering at the firm-level.
Table 4 – Regression results treated countries Models 1-4 (H1-H1c)
Model 1 Model 2 Model 3 Model 4
Variable ESG score ENV score CGV score SOC score
Mandatory 0.452 -7.423*** 8.554*** 0.675 (1.572) (1.531) (1.585) (1.488) Firm Size 7.506*** 6.958*** 3.693*** 6.741*** (0.528) (0.580) (0.639) (0.556) Leverage -3.443 -2.766 -4.685 -3.499 (2.606) (2.544) (3.899) (2.352) Profitability -7.440 -27.828*** -30.602*** -8.861 (7.679) (8.003) (7.204) (7.669) Growth Options 0.156 0.234 -0.187 0.069 (0.172) (0.198) (0.158) (0.158) Liquidity -0.017 0.009 -0.021 -0.066 (0.064) (0.067) (0.053) (0.066) Constant -46.980*** -34.861*** -0.315 -29.960*** (8.935) (10.013) (10.194) (9.613) Observations 2,467 2,467 2,467 2,467 R-squared 0.202 0.184 0.104 0.186 Regression coefficients are shown in the table. Robust standard errors in parentheses. *** p<0.01, ** p<0.05, * p<0.1. OLS regressions are made with ESG-, ENV-, CGV-, and SOC-scores as the dependent variable (Model 1 to 4 respectively). The variable Mandatory is a dummy variable where 1=after implementation of non-financial regulation and 0=before, excluding data from USA.
Firm Size=log sales, Leverage=total liabilities/total assets, Profitability=ROA, Growth Options =
30
To determine the results of H1-H1c, four regressions have been made. The first model is to determine the impact that the implementation of non-financial reporting has on the overall ESG score. The second, third, and fourth model measures the impact on the individual factors Environmental, Social, and Governance.
The first model measures the impact that the implementation of mandatory non-financial reporting has on the overall ESG score, controlled for by Firm Size, Leverage, Profitability, Growth Options, and Liquidity, and whether this impact is statistically significant. When looking at the first model, it can be found that the implementation of mandatory reporting doesn’t have a significant effect on the ESG score (β = 0.452; p>0.1). Neither do the predictor variables of Leverage (β = -3.443; p>0.1), Profitability (β = -7.440; p>0.1), Growth Options (β = 0.156; p>0.1) and the Liquidity (β = -0.017; p>0.1). This indicates that in this model, these variables do not have a statistically significant impact on the difference in ESG scores.
The predictor variable Firm Size (β = 7.408; p<0.01) does have a statistically significant impact on the change in ESG score, which is according to this paper’s predictions. This indicates that when a Firm Size increases, it will positively impact either the Environmental, Governance, Social or all three of the dependent variables. Overall, the findings for model 1 show that the implementation of mandatory non-financial reporting does not have a significant (positive) relationship with the total ESG score, therefore showing no support for Hypothesis 1. However, the models (2, 3, and 4) concerning individual scores (environmental, social, and governance) and the models including the USA as a control group show different results (see Table 5).
Model 2 measures the impact that the implementation of mandatory non-financial reporting has on the Environmental score, controlled for by Firm Size, Leverage, Profitability, Growth Options, and Liquidity, and whether this impact is statistically significant.
The findings from Model 2 show that there is a statistically significant negative relationship between the implementation of mandatory non-financial reporting and the environmental score (β = -7.423; p<0.01). This indicates that the implementation of mandatory non-financial reporting in Denmark, France and South Africa has led to a decrease in environmental scores. This is contradictory to hypothesis H1a, which predicted that this would have a positive effect on environmental scores.
31
Model 2 further shows that there are statistically significant relationships between the predictor variables Firm Size and Profitability and the Environmental Score. Firm size has a positive statistically significant relationship (β = 6.958; p<0.01), indicating that an increase in Firm Size will lead to an improvement of the Environmental score. Profitability has a negative statistically significant impact (β = -27.828; p<0.01), indicating that as the return-on-assets increases, the environmental score decreases.
Overall, the findings of Model 2 do not support the hypothesis of H1a that mandatory non-financial reporting will have a positive impact on the environmental score.
Model 3 measures the impact that the implementation of mandatory non-financial reporting has on the Corporate Governance Score, controlled for by Firm Size, Leverage, Profitability, Growth Options, and Liquidity, and whether this impact is statistically significant.
The findings of Model 3 show that there is a statistically significant positive relationship between the implementation of mandatory non-financial reporting and the corporate governance score (β = 8.554; p<0.01). This indicates that the implementation of mandatory non-financial reporting in Denmark, France and South Africa has led to an increase in environmental scores. This supports hypothesis H1c, which predicted that this would have a positive effect on corporate governance scores.
Model 3 further shows that, similarly to Model 2, there are also statistically significant relationships between two of the five explanatory variables, Firm Size and Profitability and the Corporate Governance Score. Firm size has a positive statistically significant relationship (β = 3.693; p<0.01), indicating that an increase in Firm Size will lead to an improvement of the Corporate Governance score. Profitability has a negative statistically significant impact (β = -30.602; p<0.01), indicating that as the return-on-assets increases, the corporate governance score decreases.
Overall, the findings of Model 3 support the hypothesis of H1c that mandatory non-financial reporting will have a positive impact on the corporate governance score.
Model 4 measures the impact that the implementation of mandatory non-financial reporting has on the Social Score, controlled for by Firm Size, Leverage, Profitability, Growth Options, and Liquidity, and whether this impact is statistically significant.
The findings of Model 4 show that there is no statistically significant positive relationship between the implementation of mandatory non-financial reporting and the social score (β =
32
0.675; p>0.1). This indicates that the implementation of mandatory non-financial reporting in Denmark, France and South Africa hasn’t led to a significant increase in social scores. This does not support hypothesis H1b, which predicted that this would have a positive effect on social scores.
Model 4 further shows that there is only a statistically significant relationship between the predictor variable Firm Size and the Social Score. Firm size has a positive statistically significant relationship (β = 6.741; p<0.01), indicating that an increase in Firm Size will lead to an improvement of the Social performance of the firm.
Overall, the findings of Model 4 do not support the hypothesis of H1b that mandatory non-financial reporting will have a positive impact on the social score.
When looking at the overall results of Model 1 through 4, mandatory non-financial reporting only seems to have a significant impact on the environmental and corporate governance scores. Furthermore, using these models comparing treated countries, only the model with the corporate governance score supports one of this paper’s hypotheses (H1c).
However, there are also four regression models comparing the sample countries of Denmark, France, and South Africa where mandatory non-financial reporting is implemented, to data from the country USA, which is used as a control group, where there is no mandatory non-financial reporting. The results from these regressions are quite different.
Similar to the first four models, in Table 5 the results of the regressions measuring the effect of the implementation of mandatory non-financial reporting on the overall ESG score, the Environmental score, the Corporate Governance score and the Social score, using the same explanatory variables of Firm Size, Leverage, Profitability, Growth Options and Liquidity to control, is presented. The main difference is now the effect is measured using the USA as a control group for a non-mandatory setting.
Considering that a large dataset with panel-data is involved, the results from the regressions are corrected for heteroscedasticity and unwanted correlation by clustering at the firm-level.
33 Table 5 – Regression results treated countries and USA Models 5-8 (H1-H1c)
Model 5 Model 6 Model 7 Model 8
Variable ESG score ENV score CGV score SOC score
Mandatory 8.135*** 12.357*** -10.906*** 18.882*** (1.179) (1.340) (0.981) (1.189) Firm Size 11.862*** 11.476*** 5.427*** 11.319*** (0.331) (0.405) (0.258) (0.362) Leverage -5.370** -3.560 -5.055*** -2.478 (2.256) (2.536) (1.793) (2.215) Profitability 17.565*** -2.184 -3.048 9.289** (4.219) (4.014) (3.124) (4.113) Growth Options 0.235*** 0.251*** 0.083* 0.252*** (0.050) (0.057) (0.045) (0.054) Liquidity 0.036 0.003 0.035* -0.029 (0.025) (0.030) (0.020) (0.027) Constant -125.686*** -127.829*** -11.434*** -123.966*** (5.023) (6.114) (4.076) (5.460) Observations 11,452 11,452 11,452 11,452 R-squared 0.370 0.313 0.142 0.394 Regression coefficients are shown in the table. Robust standard errors in parentheses. *** p<0.01, ** p<0.05, * p<0.1. OLS regressions are made with ESG-, ENV-, CGV-, and SOC-scores as the dependent variable (Model 5 to 8 respectively). The variable Mandatory is a dummy variable where 1=after implementation of non-financial regulation and 0=before, including data from USA.
Firm Size=log sales, Leverage=total liabilities/total assets, Profitability=ROA, Growth Options =
price/book ratio, Liquidity = cash flow/sales ratio.
Considering that a large dataset with panel-data is involved, the results from the regressions are corrected for heteroscedasticity and unwanted correlation by clustering at the firm-level. In Model 5 the effect of the implementation of non-financial reporting on the overall ESG score, and whether this effect is statistically significant is measured. This model shows that mandatory non-financial reporting has a highly statistically significant positive effect on the ESG score (β = 8.135; p<0.01). Four of the five explanatory variables also have a statistically significant effect on the ESG score. Firm Size (β=11.862; p<0.01), Profitability (β=17.565; p<0.01) and Growth Options (β=0.235; p<0.01) all have a positive effect on ESG, meaning that should any of these variables increase, so would the ESG score. On the other hand, Leverage has a negative relationship with the ESG score (β = -5.370; p<0.01), meaning that should the debt percentage over assets increase in a firm, the ESG score would decrease. Overall, the findings of Model 5 support the hypothesis H1 that the implementation mandatory non-financial reporting will have a statistically significant positive impact on the ESG score.
In Model 6 the effect of the implementation of non-financial reporting on the environmental score and whether this effect is statistically significant is measured. Model 6 shows similar
