The quality of GHG-disclosure

The quality of GHG-disclosure

and GHG-intensity in relation to financial performance regarding large listed

companies

Is GHG-disclosure value relevant for shareholders?

Master Thesis

Rijksuniversiteit Groningen Master Controlling

Supervisor: T. Marra

Second supervisor: D.A. de Waard _____________________________

Rienk Jan Hessels S2566400

rienkjanhessels@hotmail.com Tuinbouwstraat 103

9717JE, Groningen

_____________________________

Date:

04-06-2018

Word count:

9903

Abstract

The quality of GHG-disclosure and GHG- intensity in relation to financial

performance regarding large listed companies -

Is GHG-disclosure value relevant for shareholders?

This study assesses the association between the quality of GHG- disclosure and financial performance. Literature about the legitimacy and stakeholder theory is being used to form

hypotheses. These theories are addressing that the success of a company depends on its ability to manage the relationship with stakeholders and gaining legitimacy. This study looks for an answer to the question if providing a higher quality of GHG-disclosure is value relevant for shareholders. The sample in this research consists of worldwide large listed companies because they receive increased attention from stakeholders and perform more environmentally affecting activities. By using corporate reports from these

companies, the quality of these disclosures is measured. GHG-

intensity is taken into account as a moderator. In addition, the effect of GHG-intensity on financial performance is also tested by using multiple regression analysis. I found evidence that a higher quality of GHG-disclosure can improve a firm its financial performance measured in terms of Tobin’s q and is therefore value relevant for shareholders. This relation is not affected by the GHG-intensity of a company. Also, a higher GHG-intensity does not influence the

financial performance of a company. However, this study is useful for companies, investors and other users of GHG-disclosure

because it shows the value of high quality disclosures.

Key words: environmental disclosure, voluntary disclosure, Greenhouse Gas,

value relevance

Table of content

1. Introduction Blz 4

1.1. Background Integrated Reporting Blz 8

2. Literature review and hypotheses development Blz 10

2.1. Literature review Blz 10

2.2. Hypotheses development Blz 13

3. Methodology Blz 18

3.1. Sample selection Blz 18

3.2. The scopes of GHG-emissions Blz 19

3.3. Dependent variable: Financial Performance Blz 20

3.4. Independent variables Blz 21

3.4.1. Independent variable 1: Blz 21

Quality of GHG-disclosure

3.4.2. Independent variable 2: Blz 22

GHG-intensity

3.5. Control variables Blz 23

3.6. Regression model and dummy variables Blz 24

4. Results Blz 25

4.1. Descriptive statistics Blz 25

4.2. Correlation matrix Blz 27

4.3. Regression analysis Blz 28

5. Conclusion Blz 31

5.1. Conclusion Blz 31

5.2. Limitations and further research Blz 32

5.3. Practical implications Blz 33

6. References Blz 34

7. Appendix Blz 38

1. Introduction

The emission of greenhouse gas is a huge problem and serious threat for the world.

Therefore, countries undertake action. For example, at the end of 2015, the Paris

agreement was signed. This is an agreement between countries all over the world to slow down and stop the global average rise in temperature. The emission of Greenhouse Gas (GHG) is often related to a global rise in temperature. A key point of the Paris agreement is to reduce these GHG-emissions. To monitor this reduction, the countries agreed on rules to report in a consistent way about their emissions and plans (NRC Handelsblad, 2016). Basically, the reporting is not the problem, the emission generated activities are.

For example, companies in the utilities and oil & gas industry are expelling a lot of GHG-emissions. Basically, every company has some sort of direct or indirect GHG- emissions. Therefore, companies are receiving more interest from stakeholders regarding environmental activities (Rankin, Windsor & Wahyuni, 2011). To serve this interest, companies have started to disclose information about their GHG-emission.

This GHG-disclosure is still voluntary.

As stated before, the effects of climate change are becoming more and more important.

Rokhmawati, Sathye and Sathye (2015) argue that climate regulation and the more environmentally consciousness of stakeholders makes GHG-emission a growing significant risk, not only for the world, but also for companies. This could lead to a negative financial performance if companies do not take sufficient measurements to deal with those risk and do not respond to the increased pressure of stakeholders (Kolk, Levy and Pinske, 2008). GHG-disclosure is important to identify these risks and to communicate it towards stakeholders (The Greenhouse Gas Protocol Revised Edition).

Also, Iatridis (2013) is arguing that high quality disclosures are a signal of transparency

and allows investors to make informed judgements and eventually reduces uncertainty

which could lead to a competitive advantage for the company. In addition to stronger

regulations, companies face also another challenge. During the years, governments

introduced systems with the objective to reduce GHG-emission. An example in the EU is

a so-called cap and trade system. This cap and trade system is a set limit for companies

that refers to the maximum amount of emission that is allowed for an industry for a

certain period to emit. Due the fact that companies are allowed to buy and sell these

allowances, a lot of trading in allowances exist, causing a costs of carbon’. To meet future targets of the Paris Agreement, the CDP made a visualisation (Appendix 1) of the

expected price of carbon which is needed to meet those targets. Carbon dioxide is a gas that belongs to the group of Greenhouse Gases but is the most emitted. CDP stands for

‘Carbon Disclosure Project’, this organisation is there for institutional investors to motivate companies to disclose information about GHG.

As mentioned before, all companies have a certain amount of GHG-emissions. This is referred to as GHG-intensity. Over the last years, companies receive more attention regarding this topic which causes an increase in the need for information from stakeholders. In the basis, large companies receive more attention in general from stakeholders, which improves the need for information even more (Peters and Romi, 2014). In addition, large companies are carrying out more activities that affect the

environment and therefore also receive more attention form stakeholders (Rankin et al., 2011). Research indicates that environmental disclosure increases when companies are bigger, and that companies with an active strategy of stakeholder management make more extensive environmental disclosures (Magness, 2006). Freeman (1994) argued that the success of an organization depends on its ability to manage relationships with these stakeholders. Managing these relationships makes a good business sense, since managers are able to achieve their objectives in economic terms. However, these efforts of

publishing disclosure and managing the relationship with stakeholders are a costly activity for companies. Therefore, the effects on financial performance is a well-studied object. Beurden & Gössling (2008) performed a literature review about the relationship between Corporate Social Responsibility (CSR) and financial performance.

Environmental related disclosure can be seen as part of CSR. 23 studies found a significant positive relationship between CSR and financial performance, six studies found no significant relationship and two studies found a significant negative

relationship. In conclusion, acting ‘responsible’ pays off for a company.

However, a ‘problem’ with CSR is that there is no overall accepted definition. Dahlsrud (2006) performed a research on the definition of CSR used in the academic literature.

He divided CSR into five different dimensions. These dimension are economic

dimension, stakeholder dimension, social dimension, environmental dimension and

voluntariness dimension. The environmental dimension and voluntariness dimension

were significantly lower represented than the other dimensions. This is an important

finding because GHG-disclosure can be seen as part of the environmental and

voluntariness dimension which are lower represented within the literature. The results of Beurden & Gössling (2008) indicate a positive effect on financial performance, however CSR is broader than the environmental aspect. GHG-disclosure is even a more specific topic within the environmental aspect. Therefore, it is not necessarily proven that GHG-disclosure pays off.

There is also research available with a focus on environmental level. For example, Ness

& Mirza (1991) found that the environmental disclosure intensity in the oil industry was significantly higher compared to non-oil companies, however they did not research the effect on the financial performance of these companies. Iatridis (2013) performed research on the quality of environmental disclosure and its relation to shareholders value. The result was that indeed, the quality of environmental disclosure is viewed as value-relevant for shareholders. In his research, an index score for measuring the

quality was used. This index score was designed based on reporting guidelines that were broader than reporting about GHG only. Therefore, it is not proven that GHG related disclosure has value-relevance. Research with specific focus on GHG is not

overwhelming (Rokhmawati et al., 2015). In addition, they argue that regular studies about the environment are not applicable for GHG related issues given the unique characteristics and content considering carbon management. This means that GHG must be handled as a separate study objective.

The research of Iatridis’ (2013) researched only companies in Malaysia. In this research, Malaysia was classified as an advanced emerging market during the study. The need for capital is higher in emerging markets and higher quality of environmental disclosure reduces the difficulties in accessing capital markets (Iatridis, 2013). Well established companies do not have the same need for capital as companies in an emerging market.

Therefore, you could ask the question if this value relevance is applicable for established companies on a global level. Rankin et al. (2011) found that the energy and mining and industrial sector discloses more credible information compared to industries which have a lower GHG-intensity. Therefore, it is worth to research and consider the moderation effect of GHG-intensity on the relationship between the quality of GHG-disclosure and financial performance. In addition, the effect of GHG-intensity itself on financial performance will also be tested. As earlier mentioned, large companies have a higher GHG-intensity. Mainly because GHG-emission is a growing significant risk for companies they receive more attention from stakeholders and also provide more environmental disclosure. Due this high GHG-intensity in combination with more attention from stakeholders it is interesting to take large listed worldwide companies as a sample. The sample in this research consists of 500 large listed worldwide companies.

This study examines the quality of GHG-disclosure and its effect on financial

performance. Financial performance is measured in terms of Tobin’s q and Return on assets (ROA). A scoring index based on GRI-guidelines about GHG-reporting is used as a proxy for the quality of GHG-disclosure. Information about GHG-disclosure is obtained from reports published by the companies itself. GHG-intensity is measured in terms of’

total CO

and CO

equivalents in tonnes’.

The main research question of this paper is stated as follow:

‘’How value-relevant is GHG-disclosure for shareholders and what is the moderation effect of GHG-intensity on this relationship?

The remainder of the paper is organized as follows, in 1.1. a small background about integrated reporting is given. Section 2 gives an overview of the literature and

formulates the hypotheses. Section 3 describes the methodology of the research. The

results are presented in section 4 and the conclusion is presented in section 5.

1.1. Background Integrated Reporting

In this fast-changing world, there are concerns that the information provided to the stakeholder, in the form of traditional reporting methods, are not sufficient informative (Cheng, Green, Conradie, Konishi and Romi, 2014). During the years, attempts to

improve information consisted of reporting non-financial information next to traditional financial information. To report this information, a few different

mechanisms are used, for example stand-alone sustainability report, corporate social responsibility (CSR) report and even within the annual report (Cheng et al., 2014).

The International Integrated Reporting Council (IRRC) motivates companies to use a process “whereby an organisations’ value creation over time would be reported in a concise report, called integrated report, which would communicate an organization its strategy, governance, performance and prospects, in the context of its external

environment, in order to show value creation over the short, medium and long term”

(Cheng et al., 2014, p. 3). The IRRC is working together with the Global Reporting Initiative (GRI) to shape the future of corporate reporting were sustainability reporting play a vital role. GRI is a pioneer of sustainability reporting and helps businesses and governments worldwide to understand and communicate their impact on sustainability issues (website GRI, 2018

). Section 305 of the GRI Guidelines is about emissions. This section is also based on reporting and measuring requirement set by the Greenhouse Gas Protocol. The Greenhouse Gas Protocol Initiative is a multi-stakeholder partnership of businesses, non-governmental organizations (NGO’s) and others with their mission:

“to develop internationally accepted GHG accounting and reporting standards for business and to promote their broad adoption” (The Greenhouse Gas Protocol Revised Edition, p. 1).

There are also organizations initiated by institutional investors which are especially there to motivate companies to disclose (more) information about carbon related issues.

Kolk et al. (2008) recognized two prominent organisations. These organisations are the Carbon Disclosure Project (CDP) and the Investor Network on Climate Risk (INCR). The INCR is smaller compared to the CDP and is only US-based and launched in 2003.

Earlier in 2000 the CDP was launched. The CDP has more focus on the business

implications of climate change and is international orientated (Kolk et al., 2008). The

CDP is focussing on companies which are included in the FT500. The FT500 is yearly

composed by the Financial Times and is a yearly overview of the biggest 500 listed

companies worldwide. The CDP is a very successful project for institutionalizing carbon

reporting and during the years 77% of the companies included within the FT500 have

disclosed information through CDP. This high activity is not surprisingly because

research also concluded that the size of the company is positively related to the amount of disclosure (Beurden & Gössling, 2008) and the FT500 consists of the 500 largest worldwide companies.

The CDP also published a Carbon Disclosure Score of every company that disclosed the information. Kolk et al. (2008) concluded that in terms of response rate CDP is

successful, but the level of carbon disclosure that provides information which is particularly valuable for investors is questionable and differs among companies. This research will find out how valuable these disclosures can be.

1

Retrieved from URL:

https://www.globalreporting.org/Information/about-gri/Pages/default.aspx

2. Literature review and hypotheses development

This section describes theories that are useful for studying corporate disclosures . The legitimacy, stakeholder and agency theory will be shortly introduced.

2.1 Literature review

Two well-known central statements made by Freedman (1970) are saying that “the social responsibility of business is to increase profits” and “managers’ only responsibility is to increase shareholders’ wealth”. Due to these statements managers can be viewed as employees of the stockholders. So “their only responsibility is to conduct the business in accordance wither the owner’s desires to make as much money as possible conforming to the basic rules of society’’ (Beurden & Gössling 2008). But Freeman (1994) did not agree with these statements and argued that in order to keep business legitimacy, some social performance is needed, and furthermore, accountability of managers is not only towards shareholders but also towards stakeholders. The legitimacy theory is often used to study the relationship between voluntary social and environmental disclosure

(Rankin et al., 2011) and is the most dominant perspective (Mousa & Hassan, 2015).

Corporate disclosures are according to Owen (2008) a way to conform to societal expectations or can be seen as a way of maintaining or regaining legitimacy.

Summarized in a different view with the same meaning, the legitimacy theory is saying that “corporate social reporting is aimed at providing information that legitimises company’s behaviour by intending to influence stakeholders’ and eventually society’s”

(Hooghiemstra, 2000, p.57). Isomorphic forces also play a role in this theory because these forces can limit effectiveness, flexibility and innovativeness and eventually force companies in a certain direction to gain legitimacy, were the quest for legitimacy

sometimes outweigh economic reasoning (Brandau, Endenich, Trapp & Hoffjan, 2013).

Another theory closely related with the legitimacy theory is the socio-political theory.

Both, the legitimacy and socio-political theory support empirical evidence that corporate disclosures exist most of the time because of public pressure and increased media attention and the goal of the company is to gain a ‘license to operate’ (Qui, Shaukat & Tharyan, 2016).

These theories do not directly describe the relation between disclosure and financial performance, but it does explain the voluntary social and environmental disclosures

‘phenomenon’ (Mousa & Hassan, 2015). So basically, they answer the question why

companies voluntarily disclose information about environmental and social aspects and

describe what their motivations are. Hooghiemstra (2000) argued that communication is closely related to the legitimacy theory and ‘communication’ itself can be used as an overarching framework to study corporate social reporting. Both ‘corporate image’ and

‘corporate identity’ are central in this framework. The main purpose of voluntary disclosure is providing information that legitimises company’s behaviour by intending to influence stakeholders (Hooghiemstra, 2000). Appendix 2 provides an conceptual overview of the legitimacy theory. The legitimacy theory is basically about a social contract between society and companies. If companies do not respect the ‘contract’, society will evoke the contract and the company cannot continue its operations. This part is comparable to the earlier mentioned ‘licence to operate’ (Qui et al., 2016). This is the ‘breakdown’ side of the contract. There is also a ‘contribution’ side within the

contract. On this side companies using social and environmental reports to publish their social and environmental responsibility and try to gain certain objectives which will be introduced in 2.1.

The society itself can also be seen as a stakeholder. According to Freeman (1994) a stakeholder is “any group or individual who can affect or is affected by the achievement for the organization’s objectives’’. Both the legitimacy theory and stakeholder theory are different but do have some shared characteristics, according to Hooghiemstra (2000) they may be viewed as overlapping perspectives. According to Beurden & Gössling (2008), corporate social responsibility and the stakeholder theory are a fundament in studies of business and society. Furthermore, this theory was the most used theory in the papers they analysed for their literature review about CSR and financial

performance. Therefore, also the stakeholder theory should be taken into account together with the legitimacy theory. The stakeholder theory says that managers need to take into account all interests of all stakeholder of the firm. So not only financial

stakeholders but also employees, customers, communities and even under some

interpretations, the environment (Jensen, 2010). Making decisions which are in the best

interest for every stakeholder and managing these relationships will determine the

eventual or final success of the company. According to Gössling (2003) society can

determine which responsibilities companies have towards stakeholders, therefore it is

also expected that companies are accountable for their social and environmental

performance. This applies to actions as well to outcomes of these actions (Freeman,

1994).

The relationship between the company and his shareholders is often affected by the so- called agency-problem (Barako, Hancock & Izan, 2006). This problem describes the information asymmetry between the agent and the principal. This relationship is defined as “a contract under which one or more persons (the principal(s)) engage another person (the agent) to perform some service on their behalf which involves delegating some decision-making authority to the agent” (Jensen & Meckling, 1976, P.

308). So, the agency theory and legitimacy theory are having both a ‘contract’

perspective.

When looking at a company, the manager can be classified as the agent and the

shareholder as the principal, were the agent acts on behalf of the principal (Barako et al., 2006). Possible conflicts which can occur due to this information asymmetry are

addressed in formal contracts between the shareholder and manager. Shareholders basically create mechanisms to monitor managers which reduce opportunism and information symmetry and eventually maximizes shareholders’ wealth (Iatridis, 2013;

Fields et al., 2001; Meek et al., 1995). In addition to the legitimacy and stakeholder theory, the agency theory is also a theory which is suitable for developing and testing hypotheses about corporate social disclosure (Ness & Mirza, 1991). If the ownership and management are separated, agency costs can occur. These agency costs are according to Jensen and Meckling (1976) the sum of:

➢ The monitoring expenditures by the principal: costs incurred by the principal to limit the aberrant activities for the agent

➢ The bonding expenditures by the agent: costs incurred to ensure that the agent does not undertake actions that are not in the principles ‘interests

➢ The residual loss: due to sub-optimisation by the agent of the welfare-

maximisation objective

2.2 Hypothesis development

The main research question is; ‘’How value-relevant is GHG-disclosure for shareholders and what is the moderation effect of GHG-intensity on this relationship? Therefore, I will introduce certain objectives the company is trying to reach by providing disclosure and how these objectives can affect financial performance. The aim of this section is to support the conceptual (figure 1.) model with theoretical substantiation.

At the contribution side of the legitimacy theory companies using social and

environmental reports to publish their social and environmental responsibility and try to reach certain objectives (appendix 2). These objectives are:

➢ To create a good image or reputation

➢ To conform the legitimacy of their operations

➢ To demonstrate regulatory compliance

➢ To gain marketing benefits from reputation for environmental protection

➢ To differentiate themselves from their competitors

Considering the first objective, Qui et al. (2016) found that firms with better, hence higher quality social disclosures have higher market values. The researchers argued that this relation exists due to a strong reputation effect. In addition, Iatridis (2013) found that high quality environmental disclosures improve investor perceptions. Furthermore, these disclosures provide a signal of transparency and enhance managers’ reputation and social profile (Deegan et al., 2006; Patel et al., 2002; Simnett et al., 2009). These results are in line with the first objective and furthermore are evidence that a higher quality of disclosure can be value relevant.

The objective of demonstrating regulatory compliance holds together with the

regulatory risk. Although disclosure about GHG is still voluntary, in Australia companies

are obligated to report GHG-emissions towards the government. It is also obligated for

obtaining the ISO-14001 certificate (Rankin et al., 2011). Yusoff, Yatim & Nasir (2005)

also found that most companies in Malaysia disclose environmental information to

obtain the ISO-14001 certificate. This certificate is further used to improve the relation

between the company and its stakeholders (Iatridis, 2013). Regulations can differ among

countries and can change during the years. The change of regulation is a risk itself. It is named the ‘future regulation risk’. This is a risk because it is possible that companies are not sufficient prepared for the future regulations and face extra regulation costs or even fines which has a negative effect on the future cash flows. Information about GHG- disclosure then would appear to have value when a company is able to revise the

likelihood of paying fines for being non-compliance (Deegan, 2004). Increasing societal and regulatory pressure causes investors to be more interested in environmental (and social) disclosures (Friedman & Milles, 2001; Cormier & Magnan, 2007). Rankin et al.

(2011) argue that large firms receive more attention from stakeholders and therefore provide more GHG-disclosure to mitigate the future regulation risk. Therefore, demonstrating regulatory compliance is an important objective of disclosing

information and can eventually mitigate future regulation risks. By mitigating this risk the expected cash flow of a company can be larger and/or the cash flow itself can be perceived less risky.

There is another second risk regarding GHG-emissions. During the years governments introduced carbon regulations to reduce GHG-emissions, such as a carbon tax in Australia, a carbon trading scheme in the EU (EU-ETS) and energy management in Indonesia (Rokhmawati et al., 2015). Also, in North America a patchwork of systems exists, these are designed by states due the absence of federal regulation (Kolk et al., 2008). At this moment, 75% of the emissions which are covered by pricing regulation is priced

below $10 for 1 ton of CO

(CDP, 2017). But the price of carbon has significant price volatility during the years. For example, in 2008 prices under the EU Emission Trading System were €30 and €10 lower than a year later (CDP, 2017). This price volatility is also a risk for companies.

Furthermore, the CDP made a visualisation (Appendix 1) of the expected price of carbon which is needed to meet the targets of the Paris Agreement.

As can be seen in this image they expect huge price increases during the next decades. Earlier I mentioned companies face a future regulation risk, in addition, companies face also a ‘future price risk’. So GHG-emission itself is a growing significant risk for companies, GHG-disclosure is important to identify these risks and to communicate it towards

stakeholders (The Greenhouse Gas Protocol Revised Edition). By reducing emissions, the company can influence the future price risk. The perception of stakeholders can be influenced by reporting about the company's’ plans of reducing emissions. The reduction target is part of the quality

measurement of disclosure within this research.

2

World Bank and Ecofys, Carbon Pricing Watch 2017, May 2017

According to (Barako et al., 2006) environmental disclosures bring also competitive advantages and therefore economic benefits to the firm and also enhance firm’s reputation, therefore disclosing information can also be seen as a strategic tool. This is related to the objective that company tries to differentiate themselves from their competitors. But what exactly are these ‘economic benefits’? The attraction of less expensive capital and lower transaction and monitoring cost (agency theory) are associated benefits the company can expect by providing higher quality of disclosure.

According to Iatridis (2013) the lower agency costs are a result of having a better governance system. Were a higher quality of environmental disclosure is related to a better governance system. This lower costs can improve the cash flow of a company.

All together, when the provided information is relevant and valuable by providing information that describes the financial impact of GHG risks on the valuation of

corporate assets it can improve financial performance (Kolk et al., 2008). In addition, by providing better information, hence higher quality of disclosure risks can be better judged by stakeholders and companies can create a strong reputation. Furthermore, there are indications that a higher quality of disclosure leads to attracting less expensive capital and lower agency costs.

This leads to the first hypothesis;

H1: The quality of GHG-disclosure of a company is positively related to financial performance

As stated before GHG-emission leads to risks for the company and its shareholders. The amount of GHG-emissions is referred to as ‘GHG-intensity’. When GHG-intensity increases also the risk increases. The GHG-intensity between companies differs. But within certain industries, for utilities, oil & gas and financials this GHG-intensity is comparable. Due the importance of GHG-intensity and its possible effects on disclosure it is not treated as a separate control variable but as a separate independent variable and moderator within the model. First the moderator relationship will be explained and summarized in hypothesis 2. Thereafter, the relationship between GHG-intensity and financial performance will be explained.

Ness & Mirza (1991) found that that the environmental disclosure intensity in the oil industry was significantly higher compared to that of non-oil companies and that information that shareholders are likely to find relevant differs among industries.

Rankin et al. (2011) found comparable results with the addition that the energy and

mining and industrial sector provide more credible information, which is in line with the likelihood the company is exposed to higher regulatory and market risk concerning GHG emissions. Furthermore, they argue that companies within the energy and mining sector are more prepared for the future by meeting disclosure regulations in the future.

Hence, these companies are better prepared for handling the future regulation risk.

Patten (2002) also concluded that size and industry classification (related to the amount of pollution of the industry) are two of the main factors related to social disclosure.

Interesting is that they associate pollution with social disclosure and not with

environmental disclosure. This could lead to a discussion if disclosure about ‘GHG’ is a social or an environmental object. This discussion is not relevant in this paper because GHG-disclosure and intensity are treated as separate variables and are not further classified into environmental or social disclosure.

When GHG-intensity increases also the risk increases. This leads to higher pressure to report about these business risks (Kolk et al., 2008). In general, when the GHG-intensity of a company is high, there is an incentive to reduce this high risk by simply publishing and disclosing more information with a higher quality about these risks. Furthermore, Rankin et al. (2011) found that companies within the energy and mining and industrial sector also have stronger governance systems. Earlier I mentioned that firms with better governance systems also face lower agency costs. This means that the relation between GHG-disclosure and financial performance is positively moderated when GHG-

intensity rises. These considerations lead to the following hypothesis;

H2: The GHG-intensity of a company positively influences the relationship between GHG-disclosure and financial performance.

Beside looking at the possible moderating effect of GHG-intensity it is also interesting to see whether there is a relationship between GHG-intensity and financial performance.

This because companies pay nowadays a certain price for carbon. This ‘cost of carbon’ is

a result of the earlier mentioned systems governments introduced to reduce GHG-

emissions. When companies need to pay taxes for their expelled GHG-emissions or

need to buy certificates to offset their GHG-emissions, this can affect their financial

performance. It is the companies’ choice to make investments to reduce emissions or

buy certificates on the open markets. Companies will likely do this till the level of cost

effective equilibrium is reached (Clarkson et al., 2015). Another option to offset the cost

of carbon is to simple pass it on towards customers and society. Fullerton and Heutel (2007) made use of an equilibrium model to show that a carbon tax even can increase return on capital when customers have less choice to substitute between goods. Clarkson et al. (2015) found that firm valuation is affected by carbon emission, but this is

mitigated for firms with a better carbon performance (hence, lower GHG-intensity) compared to industry peers and companies within less competitive industries. This means that the impact of GHG-intensity is not the same across industries. The sample in my research consist of multiple industries. For the reason GHG-intensity can affect financial performance in a positive and negative way the direction of the hypothesis will be open. All together, this leads to the following third hypothesis;

H3: The GHG-intensity of a company is related to financial performance

The three-hypotheses together leads to the conceptual model presented in figure 1;

Figure 1; Conceptual model

3. Methodology

3.1. Sample selection

To find useful results, it is important to have a sample that includes companies which are providing GHG-disclosure. The probability of providing GHG-disclosure is higher for larger companies because they perform more environmentally affecting activities and receive more attention from financial and other stakeholders concerning the risk of an investment (Rankin et al., 2011). Kolk et al. (2008) conclude that there is a lot of disclosure and difference in the quality of this disclosure for companies who are

involved in the Carbon Disclosure Project (CDP). The focus of the CDP is on companies within the FT500. The FT500 is yearly composed by the Financial Times and is a yearly overview of the biggest 500 listed companies worldwide. Therefore, a sample which includes data from multiple years of companies who were included within the FT500 during that years will be optimal. However, the FT500 itself is not an exchange listed index. For that reason, data providers cannot create a FT500 list which includes unique company ID’s. By having this information, it would be possible to match the companies with the available data about the quality of GHG-disclosure (section 3.4.1). The database provider for this research is Thomson Reuters. The database itself is Asset4. Asset4 contains especially data about environmental, social and governance related objects. All data, except the data about the quality of GHG-disclosure, is from Asset4. This database is accessible via DataStream.

Due the missing of information about company ID’s, another way of selecting a sample with large stock listed companies was chosen. The available data about the quality of GHG-disclosure consists of a list with companies which all are scored on the quality of GHG-disclosure during the years 2013, 2014 and 2015. It includes also 2016, but due to the change of the DataStream licence I was not able to add some variables of my model.

First, companies that still lack any of the variables of my model were removed. This means automatically that only exchange listed companies were selected due to the fact that Tobin’s q was available for those companies. The residual companies were sorted on sales in ‘$’ from high to low. With the use of excel I built up the sample till it consist 500 different companies. This results in a sample of the 500 largest listed companies from the initial data. This sample is getting close to the ideal sample of large listed companies. Due the fact there are multiple years available for each company the amount of observations within the sample is 1333. This means that not for every

company every year was available. 38% of the observations is from 2013, 39% in 2015 and

23% in 2015. The companies within the sample are from different industries. Due the

fact that the FT500 consists of financials and the CDP also approaches financials to provide GHG-disclosure, financials are not eliminated from the sample. Table 2 provides an overview of the industry distribution of the sample.

3.2. The scopes of GHG-emissions

An important thing concerning GHG-emissions are the different scopes of emission.

The system of having three different scopes is designed by The Greenhouse Gas Protocol. By having these 3 types of scopes double-counting is being avoided (Boles, n.d.). A scope, “defines the operational boundaries in relation to indirect and direct GHG-emissions” (The Greenhouse Gas Protocol Revised Edition, p. 101). The following explanations are also based on The Greenhouse Gas Protocol Revised Edition (p. 25).

➢ Scope 1 emissions are emissions from sources which are owned or controlled by the company. Therefore scope 1 is al being called ‘Direct GHG-emission’.

➢ Scope 2 emissions are emissions from the generation of purchased electricity.

These emissions physically occur there where the electricity is generated. Scope 2 also being called ‘Electricity indirect GHG-emission’.

➢ Scope 3 emissions are emissions which are a consequence of the activities of the company. However, these sources are not owned or controlled by the company.

Scope 3 is a category for all other indirect emissions. Therefore scope 3 is also being called ‘Other indirect GHG-emissions’.

For independent variable 1 ‘Quality of GHG-disclosure’ and independent variable 2

‘GHG-intensity’, it is important to decide which scopes will be included. To measure the quality of GHG-disclosure Rankin et al. (2011) used the direct and indirect GHG-

emissions. They did not further specify this with ‘scopes’, but direct and indirect GHG-

emissions are comparable to scope 1, 2 and 3. This study used an environmental

disclosure index score as a proxy for the quality of disclosure which is also in line with

Clarkson, Overell & Chapple (2011) and Iatridis (2013). These index scores are based on

GRI guidelines. However, the GRI guidelines are broader than emissions only. GRI

section ‘305’ is about emissions and especially about GHG-emissions and covers all

scopes. The quality of GHG-disclosure in this research is measured based on the GRI

section 305 guidelines. For this reason, there is no evidence that only a certain scope can

affect financial performance. Therefore, all three scopes will be used to measure the

quality of GHG-disclosure.

For independent variable 2 ‘GHG-intensity’, Asset4 offers the possibility to take the direct, indirect or the sum of direct and indirect emissions. This is not further specified into ‘scopes’. There is no earlier research available which include the quality of GHG- disclosure and the intensity of GHG-emissions in one model. King & Lenox (2011) observed the relation between GHG emissions and financial performance (Tobin's q) and used ‘total emissions’ but did not further specify this with the use of the scope system. Regarding the definition of scope 2, this scope is about generated or purchased electricity. If, for example, the production process of a company is using a lot of

electricity, this will affect scope 2. In this situation the company can decide to invest in less electricity consuming assets to reduce emission in scope 2. This means that

company can also control emissions regarding scope 2. Therefore, it is fair to measure GHG-intensity as the sum of direct and indirect emission which is in line with the research of King & Lenox (2011). This means that independent variables 1 and 2 both include all three scopes.

3.3. Dependent variable: Financial Performance

The dependent variable financial performance will be measured as market value.

Market value can be measured in different ways. For example, Luo and Bhattacharya (2006) used Tobin’s q and stock return. Also Iatridis (2013) used Tobin’s q. Tobin’s q is basically a measurement of unrecorded asset value (Guidry & Patten, 2012). In other words, Tobin’s q presents what the market thinks how large the expected cash flow is from a dollar invested in assets. Tobin’s q is higher if the future cash flow is larger or if the cash flow itself is less risky (Clarkson et al., 2015). Risk is an important factor for companies regarding GHG. Therefore, Tobin’s q will be used to measure market value.

Stock prices are fluctuating on a daily base. Because the sample consists out of multiple years this effect is lowered. To measure the non-market value financial performance, an additional test based on the ROA will be conducted. The formula to derive Tobin’s q is:

share price x number of common stock outstanding + liquidating value of the firm’s preferred stock

+ short – term liabilities – short-term assets + book value of long term debt) / book value of total

assets.

3.4. Independent variables:

3.4.1. Independent variable 1: Quality of GHG-disclosure

The first independent variable is the quality of GHG-disclosure (GHG-Disclosure).

Information about the quality of GHG-disclosure can be obtained from an integrated report or from a separate sustainability report. This data is collected during the years by bachelor students from the university of Groningen who were performing their

bachelor thesis. Every student obtained an instruction how to interpret the information from a report and capture the data. With the use of a question list the following

dimensions are scored:

➢ Part A: General information

➢ Part B/C/D: Scope 1,2 & 3

➢ Part E: GHG-emissions general

➢ Part F: GHG-emissions target setting

➢ Part G: Board/firm variables The questions regarding part B/C/D are:

➢ Does the report separately disclose scope (1,2 or 3) GHG-emissions in metric tons of CO2 equivalent?

➢ Does the report separately disclose gases included in the calculation of scope (1,2 ore 3) GHG-emissions (whether CO2, CH4, N2O, HFC’s, PFC’s, SF6, NF3, or all)

➢ Does the report separately disclose scope (1,2, or 3) emissions’ (1) chosen base year, (2) the rationale for choosing the base year, (3) emissions in the base year, and (4) the context for any significant changes in emissions that triggered recalculations of base year (note: add 1 point for each sub-item, score ranges between 0-4)

➢ Doe the report disclose the chosen consolidation approach for emissions (equity share, financial control, operational control), with respect to scope (1,2 or 3).

Part E consist of the same questions as part B/C/D, however part E is only necessary

when a company does not provide information about separate scopes and reports only

emissions in total.

Part F is about emissions target setting. Two examples of in total 8 questions are:

➢ Does the report disclose a clear, quantitative target stating its ambitions regarding the reduction of GHG? (Note: No information = 0; one overall % = 1; at least 2 scopes separately = 2)

➢ Does the report disclose information on how the organization plans at (further) reducing its GHG-emissions and/or achieving targets (Note: No information on plans = 0; one overall description = 1; descriptions for at least 2 separately = 2) Although the questionnaire together with the instructions students receive does not provide much space for interpretation, there is still a possibility that students performed the data collection different from each other or made mistakes. Therefore, I personally performed a sample of 25 observations. Next, I compared my score with that of the students. Results indicated that the sample was for 88% identical. This means that the data does not have a large deviation what the reliability of further research results improves.

To use the collected data a quality index score is composed. Part A and G are not included in this index because the parts describe general information and board

variables which are not related to GHG. In total a company can earn 37 points related to the quality of GHG-disclosure. All three emission scopes are equally weighted. Dividing the scored points by the maximum score of 37 points represents the quality of GHG- disclosure.

3.4.2. Independent variable 2: GHG-intensity

The information about the variable GHG-intensity comes from Asset4. According to Asset4 the name of the variable is ‘CO2 Equivalents Emissions Total’. The description is

‘Total CO2 and CO2 equivalents in tonnes’. This basically means that all direct and indirect Greenhouse gases are rescaled into CO2 equivalents measured in tonnes.

Companies sometimes report different numbers according to different protocols. Such

protocols are the GHG-protocol, Kyoto-protocol and the EU-Trading Scheme. When

this is the case Asset4 takes the GHG-protocol as the reported value.

3.5. Control variables

Previous research emphasised some control variables which are important regarding financial and environmental reporting (Peters and Romi, 2014). An important variable is

‘relative environmental performance’. It can be assumed that companies which have a high environmental performance will communicate (disclose) this towards their stakeholders. By doing so, these companies create a good image or reputation. The Asset4 variable which is comparable with ‘environmental performance’ is

Environmental Score (ENVSCORE). According to Asset4, the Environmental Score measures a company's impact on living and non-living natural systems including the air, land and water, as well as complete ecosystems. It reflects how well a company uses best management practices to avoid environmental risks and capitalize on environmental opportunities in order to generate long term shareholder value.

Firm size is the second control variable. Approximately half of the studies about CSR and financial performance came up with the conclusion that firm size significantly influences the relationship between CSR and financial performance in a positive, as well a negative way (Beurden & Gössling, 2008; Magness, 2006). Although Orlitzky (2001) performed a meta-analysis about firm size in combination with the relation between CSR and financial performance. The conclusion of that research was that firm size does not confound this relationship. Another argument to include firms size as a control variable is that bigger companies are gaining more attention from stakeholders which improves the need for information (Peters and Romi, 2014). Due to the fact the

relationship is influenced in different ways it is wisely to control for firm size. Firm size is measured as Sales in ‘$’ (SALESUSD).

Leverage is the third and last control variable. A higher debt to equity ratio causes a higher demand for information of investors. This demand for information can be seen as an incentive for managers to voluntarily disclose more information (Peters and Romi, 2014). Companies with a higher debt to equity ratio also provide information with a higher quality to provide assurance that they respect the financial obligations (Iatridis, 2013). Leverage is measured as total debt divided by total assets.

3.6. Regression model and dummy variables

All variables together result in the following regression model:

Financial Performance t = α

+ β

(GHG-Disclosure) + β

(GHG-Intensity) + β

(GHG-

Disclosure*GHG-Intensity) + β

(ENVSCORE) + β

(SALESUSD) + β

(LEVERAGE) + β

(INDUSTRY D

) + β

(YEAR D

) + 𝜀 t

As shown above, also dummy variables are included to control for industry and year

effects.

4. Results

4.1. Descriptive statistics

Table 1 provides the descriptive statistics. All variables are winsorized based on 3 standard deviations. 1333 is the total amount of observations. 38,4% of this sample represents observations in 2013 (mean multiplied by 100), 38,5% in 2014 and 23,0% in 2015. The sample consists also of companies with a capital structure with a neglectable level of debt. The average score on GHG-disclosure is 0,281. The lowest score is 0 and the highest is 0,790.

Table 1 Descriptive Statistics

Variable N Minimum Maximum Mean Std.

Deviation

Financial Performance (Tobin’s Q)

1333 0,047 3,569 1,057 0,711

GHG-Disclosure 1333 0,000 0,790 0,281 0,168

GHG-Intensity* 1333 79,10 83.051.892 7.340.130 16.998.443

Environmental Score 1333 50,61 95,04 86,16 10,649

Leverage 1333 0,000 0,713 0,258 15,155

Sales (USD)** 1333 5.682.331 146.954.338 27.770.100 29.918.828

Dummie 2013 1333 0,000 1,000 0,384 0,487

Dummie 2014 1333 0,000 1,000 0,385 0,487

Dummie 2015 1333 0,000 1,000 0,230 0,421

*Total CO

and CO

equivalents in tonnes

**numbers in thousands

A more in-depth descriptive of the separate industries is given in Table 2. In addition, also the average is presented. GHG-disclosure, GHG-intensity, Tobin's’ q and sample percentage are presented all on industry levels. The industry Oil & Gas is the industry with the lowest quality score (0,235). The industry Technology is the one with the highest score (0,330). The industry Utilities has the highest GHG-intensity. Financials, not surprisingly, are showing the lowest GHG-intensity.

Table 2

Industry descriptives

Industries GHG-Disclosure GHG-

Intensity*

Tobin’s q % Sample

Basic Materials 0,268

16.243.524 0,99 11,78

Consumer Goods 0,302

1.688.886 1,42 13,65

Consumer Services 0,262

4.030.524 1,32 10,28

Financials 0,274

173.313 0,33 13,05

Health Care 0,312

630.913 1,81 4,58

Industrials 0,281

3.161.966 1,02 20,48

Oil & Gas 0,235

15.254.465 0,93 8,93

Technology 0,330

909.975 1,40 7,50

Telecommunications 0,271

2.234.065 1,17 3,53

Utilities 0,293

41.342.930 0,73 6,23

Average 0,281

7.340.131 1,06 x

Note: n=1333 *Total CO

and CO

equivalents in tonnes

4.2. Correlation matrix

Table 3 provides the correlation matrix. The variables GHG-intensity and sales are now transformed in logarithms based on the winsorized values.

There are certain significant correlations between the variables. First GHG-disclosure is positively significant related to financial performance (Tobin’s q) (r = 0,092, p < 0,05).

This is a first indication for hypothesis 1. Environmental score is also positively

significant related to GHG-disclosure (r = 0,0166, p < 0,05) as well to GHG-intensity (r = 0,072, p < 0,05). Leverage is positively related to GHG-intensity (r = 0,171, p < 0,05). Sales is significantly related to almost every variable except leverage. Which stands out is the relatively large (compared to other correlation coefficients) correlation between sales and GHG-intensity (r 0,309, p < 0,05). This is not obvious, sales is a measurement of the scale of the company and when a company is larger it is also logic that it has a higher GHG-intensity. There are no signals of multicollinearity due the fact there are no values higher or lower than +/- 0,7. In addition also the VIF values presented in table 4 are way below 10 (Sinan & Alkan, 2015).

Table 3 Correlations

Variable 1 2 3 4 5 6

1 Financial Performance (Tobin’s Q)

-

2 GHG-Disclosure 0,092* -

3 GHG-Intensity -0,15 0,005 -

4 Environmental Score 0,010 0,166* 0,072* -

5 Leverage 0,041 -0,12 0,171* -0,051 -

6 Sales -0,0145* 0,083* 0,309* -0,167* -0,043 -

Note: * = p < .05 (2-tailed), ** = p < .01 (2-tailed).

4.3. Regression Analysis

The results of the regression analysis are presented in table 4. Model 1 includes only the control and dummy variables. This model counts for 3 years. The first two years are taken into account in the regression analysis, where the year dummy of 2015 functions as a proxy. Model 2 includes the first independent variables GHG-disclosure related to hypothesis 1. Model 3 includes the second independent variable GHG-intensity which is related to hypothesis 3. Model 4 includes both independent variables. Finally, model 5 represents the whole conceptual model and includes the interaction effect between GHG-disclosure and GHG-intensity (H2).

Model 1 shows that sales is significantly negative. This effect is visible in all five models.

Also the dummy variable for years is significantly negative in all models. Results for hypothesis 1 indicates that the quality of GHG-disclosure significantly positive related to financial performance (Tobin’s q) in the models where this variable is integrated.

Therefore, the null hypothesis can be rejected in favour of the pre-set alternative hypothesis 1. This means that indeed the quality of GHG-disclosure leads to a better financial performance measured in terms of Tobin’s q. The adjusted R-square in model 2 is 0,037. This means that 3,7% of the dependent variable is influenced by the control variables and independent variable. In addition, the delta R-square is 0,010, which means that 1% of the variation in financial performance is explained by the independent variable GHG-disclosure.

Regarding hypothesis 3, model 2, 3 and 5 are showing that GHG-intensity is not significant related to financial performance (Tobin’s q). Therefore, the null hypothesis cannot be rejected in favour of the pre-set alternative hypothesis 3. This means that the GHG-intensity of a company has not a significant effect on the financial performance.

Model 5 is presented to test the moderating effect of GHG-intensity on GHG-disclosure and financial performance. This is called the two-way interaction. The conclusion is that there is no significant interaction effect. Therefore, the null hypothesis cannot be

rejected in favour of the pre-set alternative hypothesis 2.

TABLE 4

Regression results (dependent variable Tobin’s q)

Variable Model 1 Model 2 Model 3 Model 4 Model 5

Intercept 3,060 (0,405)*** 3,120*** (0,402) 3,073*** (0,405) 3,135 (0,403)*** 1,057 (0,019)***

Controls Environmental Score

0,002 (0,002) 0,001 (0,002) 0,002 (0,002) 0,001 (0.00) 0,009 (0,020)

Sales USD -0,289 (0,054)*** -0,300 (0,054)*** -0,305 (0,057)*** -0,319 (0,057)*** -0,115 (0,020)***

Leverage 0,173 (0,127) 0,174 (0,126) 0,150 (0,130) -0,148 (0,129) 0,023 (0,020) 2013 -0,163 (0,051)*** -0,166 (0,050)*** -0,164*** (0,051) -0,167 (0,050)*** -0,082 (0,025)***

2014 -0,131 (0,051)*** -0,140 (0,051)*** 0,132*** (0,051) -0,141 (0,051)*** -0,070 (0,025)***

Main effects GHG-Disclosure (H1)

0,438 (0,115)*** 0,441 (0,115)*** 0,072 (0,020)***

GHG-Intensity (H3)

0,020 (0,021) 0,022 (0,021) 0,020 (0,021)

Two-way interaction GHG-Disclosure x GHG-Intensity (H2)

-0,022 (0,023)

𝑅

0,031 0,042 0,032 0,043 0,043

Adjusted R-square 0,028 0,037 0,028 0,037 0,037

Δ R square 0,031 0,010 0,001 0,011 0,012

F-value 8,603*** 9,639*** 7,310*** 8,4138** 7,475***

Highest VIF 1,652 1,654 1,652 1,655 1,658

Note: * p < .10, ** p < .05, *** p < .01, two-tailed.

Parentheses contain standard error

Values are controlled for industry effects

Table 5 represents the same results as in table 4 except for the dependent variable, Tobin’s Q, which has been replaced by Return on Assets (ROA). ROA is also winsorized based on 3 standard deviations.

The quality of GHG-disclosure is also significant for the financial performance measured in terms of ROA. Notable is that the Adjusted R-square in model 2 is way lower (0,005) compared to the one in model 2 table 4 (0,037). This means that the effect of GHG-disclosure is stronger for Tobin's’ q compared to the effect for ROA.

TABLE 5

Regression results (dependent variable ROA)

Variable Model 1 Model 2 Model 3 Model 4 Model 5

Intercept 9,147 (3,448)*** 9,442 (3,446)*** 9,122 (3,452)*** 9,423 (3,450)*** 4,831 (0,163)***

Controls

Environmental Score

-0,003 (0,016) -0,009 (0,016) -0,003 (0,16) -0,008 (0,016) -0,088 (0,169)

Sales USD -0,465 (0,462) -0,552 (0,462) -0,434 (0,486) -0,500 (0,487) -0,179 (0,175) Leverage -2,704 (1,082)** -2,699 (1,080)** -2,660 (1,104)** -2,667 (1,102)** -0,400 (0,167)**

2013 -0,009 (0,432) -0,021 (0,432) -0,008 (0,432) -0,021 (0,432) -0,014 (0,210) 2014 -0,093 (0,432) 0,050 (0,432) -0,095 (0,432) -0,051 (0,432) 0,018 (0,211)

Main effects GHG-Disclosure (H1)

2,155** (0,861) 2,151 (0,988)** 0,350 (0,167)**

GHG-Intensity (H3)

-0,037 (0,182) -0,027 (0,182) -0,032 (0,176)

Two-way interaction

GHG-Disclosure x GHG-Intensity (H2)

-0,138 (0,169)

𝑅

0,005 0,009 0,005 0,009 0,009

Adjusted R-square 0,002 0,005 0,001 0,005 0,003

Δ R square 0,005 0,004 0,000 0,004 0,004

F-value 1,446 2,004* 1,211 1,719* 1,567

Highest VIF 1,652 1,654 1,652 1,653 1,658

Note: * p < .10, ** p < .05, *** p < .01 two-tailed.

Parentheses contain standard error

Values are controlled for industry effects

5. Conclusion

5.1. Conclusion

This research has examined the relation between the quality of GHG-disclosure and financial performance regarding large listed companies. In addition, the effect of the GHG-intensity was taken into account as an moderator. The reason for this research was that large companies face stronger attention from stakeholders and also have an higher GHG-intensity. The findings of this study show that the quality of GHG-disclosure is indeed positively related to financial performance in terms of Tobin’s q, which was in line with the first hypothesis. Also for financial performance measured in Return on Assets the results were significant. However, the effect of ROA was too weak, which makes it economic negligible. No significant interaction effect has been found between the GHG-intensity and GHG-disclosure of a company (hypothesis 2). This finding was not as expected. Ness & Mirza (1991) found that oil companies publish more

environmental disclosure compared to other industries and in addition Rankin et al.

(2011) found that the energy and mining and industrial sector provide more credible information. Iatridis (2013) argued that these firms also have better governance systems which can reduce agency costs. However, in this research, the oil & gas industry has the lowest score on the quality of GHG-disclosure. In addition, the sector utilities which has a huge GHG-intensity scored just above average. These specific results regarding GHG- disclosure are not in line with earlier findings regarding environmental disclosure overall and can be a reason why there was not an moderation effect of GHG-intensity has been found.

Regarding hypothesis 3, no significant relation has been found between the GHG-

intensity of a company and its financial performance. Main argument for a significant

relation was that nowadays companies have to deal with the cost of carbon. Companies

can choose to invest in emission reduction assets, pass the cost on towards the customer

or just accept the extra costs of carbon. Therefore, the direction how GHG-intensity

affects financial performance was unclear. Although the results indicate no link between

GHG-intensity and financial performance during the years 2013, 2014 and 215, the

future is uncertain. If the cost of carbon will rise during the years and will become more

volatile, the financial risk for companies will become larger. This could mean that the

GHG-intensity eventually can affect financial performance.

The average score on the quality of disclosure in my research was 0,28 were 1 was the maximum that can be obtained. Therefore, a general conclusion is that there is still a lot of room for improvement for companies regarding the quality of GHG-disclosure. This is in line with the findings of Kolk et al. (2008). They concluded that in terms of

response rate CDP (which has the focus on companies comparable to the sample in this research) is successful but the level of carbon disclosure which provides information is particularly valuable for investors, is questionable and differs among companies.

Although the overall quality of disclosure was not very high, the quality itself is indeed value relevant for shareholders. This also means that the quest for legitimacy in this case did not outweigh economic reasoning, as can be sometimes the case according to

Brandau et al. (2013). However, the improvement in financial performance was present for Tobin’s q and economic negligible for ROA. But even if this research concluded that there was no overall economic relevance of providing GHG-disclosure for shareholders, this will not mean that these disclosures have no relevance at all. It is a nice addition for the shareholder that these disclosures are value relevant. However, a shareholder is just one of the many stakeholders of a company. These days we should also encapture the environment as an stakeholder. Therefore, it is important that companies try reduce their GHG-intensity and keep disclose information about GHG to provide information about the possible risk towards stakeholders but also are becoming critical and aware about their own processes regarding GHG.

5.2. Limitations and further research

The difference in the relation between the quality of GHG-disclosure and Tobin's’ q compared to ROA can be an important finding. ROA is an financial measurement of realized financial performance and contains no component regadering risk. For the measurement Tobin's’ q, this is not the case. Tobin’s q is higher if the future cash flow is larger or if the cash flow itself is less risky (Clarkson et al., 2015). GHG-disclosure can affect the perception of a shareholder regarding the expected risk and/or cash flow. The difference between risk and cash flow perception is not visible in this research.

Therefore, further research can be performed were the risk and cash flow perception would be investigated separately. Expected is that this perception is more risk oriented, instead of cash flow. This because the relation between disclosure and ROA was really weak.

As in every research there are several limitation. The sample in this research consists of

worldwide large listed companies in different industries. GHG-regulations can differ

among countries. Therefore, the results cannot be generalized for small companies in countries were specific GHG-regulations are applicable. Furthermore, the sample consists only out of three consecutive years. Measurements regarding Tobin's’ q can be different in other periods of time, due economic expectations of investors regarding expected return and/or risk. Therefore, results can be different in other years. Another limitation of the data is regarding the quality of GHG-disclosure. This data is collected by multiple persons. This makes it possible that students performed the data collection in various ways. However, this risk is mitigated by comparing a students’ sample of observations with a sample of mine.

5.3. Practical implications

This study can be useful for companies, investors and other user of GHG-disclosure.

Because this study shows that the quality of these disclosures has an economic value. In

addition, they can also emphasize the risk regarding GHG which companies face in the

future. Furthermore, environmental groups can make use of this study to motivate

companies from an economic perspective, in addition to the environmental perspective,

to disclose information about GHG.