The relationship between integrated reporting and corporate tax avoidance

(1)

The relationship between integrated reporting and

corporate tax avoidance

Name:

Wouter Kempers

Student number:

10422676

Thesis supervisor:

Dhr. Dr. A. Sikalidis

Date:

26-06-2017

Word count:

MSc Accountancy & Control, specialization: Accountancy

Faculty of Economics and Business, University of Amsterdam

(2)

2 Statement of Originality

This document is written by student Wouter Kempers 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)

3 Abstract

This paper researches the possible relationship between engaging in integrated reporting and engaging in tax avoidance. By using agency theory and stakeholder theory I explain that firms want to balance the interests of different stakeholders. By engaging in integrated reporting, a firm shows how it creates value for stakeholders. Firms only disclose voluntary information if that means they show they are not acting sub optimally. Tax avoidance can have a negative influence on the government, other stakeholders, and shareholders, which firms would want to hide. Thus, firms that are engaging in integrated reporting are less likely to engage in tax avoidance. This is tested by (1) looking at firms that mention the international integrated reporting council in their annual report, or are influenced by it, through engaging in an integrated reporting network, and by (2) looking at firms that prepare their annual report according to the guidelines of the IIRC framework. The sample that is used consists of 327 and 385 firms respectively, for the two hypotheses that are being tested. I find no statistically significant evidence indicating a correlation between integrated reporting and tax avoidance. These findings are contradictory with the expectations based on the literature study. This could be explained by the limited amount of data that is yet available on integrated reporting because of its relative novelty. Possible other explanations could be that the interest of managers of more transparent firms (i.e. engaging in integrated reporting) is more aligned with the interest of shareholders that have incentive to pay as little taxes as possible. Furthermore, the IIRC shifts away from its objective of sustainable reporting which can be seen by their definition of value as value for investors instead of value for society. This means that firms that engage in integrated reporting are not necessarily focused on providing value for society, but merely for investors that have an incentive to avoid taxes. The paper is subject to a few limitations. First, the measure of ETR might not correctly depict tax avoidance. Second, because of the relative novelty of integrated reporting, there is no abundance of data, which limits the potential of the study. Third, the model does not take into account ownership structure of a company as a control variable, which could have an influence on tax avoidance.

(4)

4 Table of contents Abstract ... 2 1. Introduction: ... 4 2. Literature review ... 8 2.1 Integrated reporting ... 8 2.2 Tax avoidance... 9

2.3 Stakeholder theory and agency theory ... 11

3. Research questions and designs: ... 12

3.1 Hypotheses development: ... 12

3.2 Empirical basis:... 14

3.3 Research model: ... 15

4. Empirical results: ... 17

4.1 Results for hypothesis 1: ... 17

4.1.1 Correlation matrix: ... 18

4.1.2 Regression results: ... 20

4.2 Results for hypothesis 2: ... 23

4.2.1 Correlation matrix: ... 24

4.2.2 Regression results: ... 25

5. Conclusion: ... 28

6. References: ... 31

7. Appendices ... 35

Appendix A: Sample selection... 35

Appendix B: Variable measurement ... 36

1. Introduction:

Recent years have shown that corporate tax avoidance has increasingly become a major issue (Campbell, 2016). The European Council has adopted a new directive that addresses the main ways in which large companies attempt to reduce their tax liability (EC, 2016). Also in the United states of America there has been critique on tax avoidance, as Oxfam (Ratcliff & Rusu, 2016) claims that US listed companies stash $1.3 trillion in offshore accounts. Tax avoidance not only has an effect on government, that gains less from taxes, but also on

(5)

5 inhabitants of a country, that pay more taxes relative to companies engaging in tax avoidance, and therefore on the reputation of a company. Also, Kim et al (2011) find strong evidence that tax avoidance is positively associated with bad news hoarding. This in turn leads to future crash risk, which has a negative effect on the shareholders of the company. Thus, it is important from both a public and shareholder viewpoint to know what the drivers of tax avoidance are.

Furthermore, reporting on corporate, social, and responsibility (CSR) has been on the rise over the last couple of years. Where 80% of the G250 companies reported on CSR in 2008, this number has risen to 93% in 2013 (KPMG, 2008; KPMG 2013). The European commission (2017) defines CSR reporting as: the responsibility of enterprises for their impact on society. Traditional CSR reporting mostly takes place in the form of an addendum to traditional annual reports (Jensen & Berg, 2012). However, separating financial and nonfinancial information does not make sense as they depend on each other in a company. Therefore, the question is raised on how traditional CSR is actually embedded in a corporation and how much value CSR reporting has if it does not comprise information on the strategic intent of the company and does not display historic business (Jensen & Berg, 2012).

To overcome this problem, advocates have called for a new form of reporting, known as integrated reporting. An integrated report is defined by the international integrated reporting council, which is a global coalition of regulators, investors, companies, standard setters, the accounting profession and NGOs to promote communication about value creation as the next step in corporate reporting, as “a concise communication about how an organization’s strategy, governance, performance and prospects, in the context of its external environment, lead to the creation of value in the short, medium and long term” (IIRC, 2013).

Following the formation of the International integrated reporting council in 2010, the IIRC has been working on a breakthrough phase since 2013. This phase enhances the move from the creation of the framework and market testing, to the development and adoption by reporting organizations (IIRC, 2013). The IIRC is not a regulator or standard setter, so companies merely report on a voluntary basis.

To date there have been studies that investigate the impact that CSR reporting has on corporate tax avoidance, however to my knowledge there have been no studies looking at the impact between integrated reporting and tax avoidance. Therefore, that is what this study looks at.

As will be explained in more detail in the following chapter, integrated reporting could have an effect on tax avoidance because of the following. As stakeholder theory shows, a major objective of firms is to attain the ability to balance the conflicting demands of various

(6)

6 stakeholders of the organization (Ansoff, 1965). Thus, firms have to take into account all stakeholders. By engaging in integrated reporting a firm shows how it creates future value for its stakeholders (IIRC, 2016). From agency theory, it is known that firms want to show that they are acting optimally by disclosing voluntary information, i.e. engaging in integrated reporting (Jensen & Meckling, 1976). A company will not give out a report, if it shows that the firm will affect stakeholders by creating substandard value. Tax avoidance means the reduction of explicit taxes in a legal way (Hanlon & Heitzman, 2010). This enhances that firms withhold money that would otherwise go to governments. This has a negative impact on society as there is less money available for the government. So, either less money is available to invest in society, or society has to pay extra tax to fill up this shortage. Also, shareholders are negatively affected as tax avoidance is positively associated with bad news hoarding, which in turn could lead to future crash risk. This means that companies create less value for their shareholders, other stakeholders, directly less value for governments, and indirectly for society. It is expected that when firms engage in integrated reporting, they do so because they do not have to hide their sub optimal behaviour (engaging in tax avoidance), and therefore are less likely to engage in tax avoidance. Accordingly, the following research question is examined in this paper:

What is the effect of using integrated reporting on engaging in corporate tax avoidance?

To test this effect a database study is conducted. A distinction is made between firms that engage in integrated reporting and firms that do not. The sample contains firms that are listed around the world. Tax avoidance is measured by using the proxy of effective tax rate. I construct two hypotheses to examine the possible effect that engaging in integrated reporting could have on engaging in tax avoidance. As a first criterion for engaging in integrated reporting, I look at companies that refer to the IIRC in their annual report, or that are influenced by integrated reporting by participating in integrated reporting networks. I find that there is no link between integrated reporting and tax avoidance when this is the criterion for integrated reporting. For the second hypothesis I look at whether companies prepared their annual report by following the IIRC framework, which is a stricter definition of integrated reporting. Also for this hypothesis I do not find any statistically significant supporting evidence. Putting the outcomes of these two hypotheses together, I do not find any statistically significant evidence that supports the research question that integrated reporting has an effect on engaging in tax avoidance.

This paper contributes both to the inside and outside of the academic world. The contribution to the academic world is twofold. First, there is an increasing amount of

(7)

7 literature available on tax avoidance. This paper adds to the amount of literature on tax avoidance. To really understand tax avoidance it is important to know what the underlying factors of that subject are. Second, as integrated reporting is a fairly new topic of research, not a lot of research has been conducted on it. By looking at the possible effect it has on tax avoidance, the advantages and disadvantages of integrated reporting are examined.

Furthermore, the paper contributes to outside of the academic world, or the ‘real world.’ As governments and shareholders stand to lose money from tax avoidance, it is important for them to know the drivers of it. Christensen and Murphy (2004) find that tax revenue are the necessary ‘lifeblood' of a democratic government. Furthermore, as Hanlon and Slemrod (2009) find that firms suffer stock price declines following a public revelation of tax avoidance, shareholders could lose a great deal of money because of the reputational costs that tax avoidance could enhance. Kim et al. (2011) show that corporate tax avoidance is positively associated with future crash risk of firm-specific returns, because it facilitates opportunistic behaviour of managers. The crash risk increases because managers get the tools to hide bad news or information, which is being stockpiled in this way. This could also have a negative impact on shareholders as there is a potential for them of losing value. By looking at integrated reporting, which may have an impact on tax avoidance, this paper gives shareholders an understanding of the drivers of it. Governments especially have much to gain from this paper, as they have the means of influencing companies’ decisions to engage in voluntary integrated reporting by for instance promoting it.

Complementary, it is important to stress that the influence of corporate social responsibility on tax avoidance has been tested by other authors, but the influence of integrated reporting has not been tested. Because integrated reporting is seen as solving the problems of traditional CSR, it is important to test this relationship.

This paper is structured as follows. First the literature that is used in explaining the relationship between integrated reporting and tax avoidance is explained. This paper unites these two streams of research. The effect of the latter on the former can be shown by using two streams of theory, being agency theory and stakeholder theory. Next, the hypotheses are developed, followed by the empirical basis. Then the research model is explained, followed by the empirical results of the two hypotheses. The last chapter contains the conclusion.

(8)

8 2. Literature review:

In this chapter the literature underlying this paper is being outlined. First, integrated reporting is explained. Next, the definition along with some key concepts of tax avoidance are being outlined. In the third paragraph both stakeholder theory and agency are being explained. This paper unites two streams of research, namely tax avoidance and integrated reporting. The effect of the latter on the former can be shown by using two streams of theory, being agency theory and stakeholder theory. The hypotheses are developed based on these two theories.

2.1 Integrated reporting

In this paragraph, the definition as well as some key concepts of integrated reporting are explained. Integrated reporting is defined by the IIRC (2013) as “a concise communication about how an organization’s strategy, governance, performance and prospects, in the context of its external environment, lead to the creation of value in the short, medium and long term.”

Social and environmental reporting has a long history. From the 1990s onwards, the social and environmental disclosures have increasingly been made in stand-alone reports next to financial reports. More recently however, there has been a move towards integrating social and environmental information with the financial statements. This has come to be known as integrated reporting (De Villiers et al, 2014). Integrated reporting has been on the rise since the first integrated report in 2002, and is seen as an alternative to traditional sustainability reporting (Hespenheide 2010). Specifically, Jensen and Berg (2011) use the institutional theory to explain why companies are choosing integrated reporting more in comparison to separated sustainability reporting. The authors hypothesize and find that in countries in which companies are seen as responsibility bearing parts of the society, integrated reports are produced more than in other countries. Thus, companies feel the urge to commit more to integrated reporting when more people are seeing them as the responsible parts of society. In other words, firms listen to what society thinks or wants from them.

In 2010 the international integrated reporting council (IIRC) was established. The IIRC’s aim is to integrate financial, social, environmental and governmental information in a clear, concise, and comparable format (IIRC, 2013a). According to the IIRC (2013b), integrated reporting promotes the reporting company’s access, use and the degree of dependency on social, environmental, and economic resources, its relation with different

(9)

9 sorts of capital, and the impact the organisation has on these sorts of capital. Thus, by engaging in integrated reporting a company shows its additional impact on social, environmental, governmental and economic issues. This realisation is important to keep in mind later on in the paper when the stakeholder theory and agency theory are discussed. Furthermore, Kim et al (2011) study whether firms that exhibit corporate social responsibility (CSR) reporting, also behave in a responsible manner to constrain earnings management. Even though CSR reporting is not the same as Integrated reporting, an integrated report also contains a part of which is based on corporate and social responsibility. Therefore, managers that work for a firm that engages in integrated reporting also report on corporate and social responsibility. Carroll’s (1979) delineation of a firm’s social responsibilities suggests that CSR firms should strive to make a profit, obey the law, be ethical, and, further, be a good corporate citizen by financially supporting worthy social causes. Numerous theoretical studies on the ethical view of CSR (e.g., Carroll 1979; Donaldson and Preston 1995; Jones 1995; Phillips 2003) argue that there is a moral imperative for managers to ‘‘do the right thing.” Thus, in other words, according to the above studies, managers of firms that engage in integrated reporting, are motivated to ‘do the right thing.’

2.2 Tax avoidance

In this paragraph, the definition, along with some key concepts of tax avoidance are outlined. Research on tax avoidance calls for a more elaborate investigation of the subject. The primary reason for this is that tax avoidance does not have a universally accepted definition (Hanlon & Heitzman, 2010). Therefore, in this paper I use a broad definition, provided by Hanlon and Heitzman as the reduction of explicit taxes in a legal way.

Chen et al (2010) show that the government usually takes around one third of regular companies’ pre-tax profits. This significance makes it understandable why shareholders and owners would want companies to use tax avoidance to lower the costs of taxes, and heighten the after tax profit. Auerbach (2006) substantiates this view, as they find that shareholders bear a significant part of the tax burden that is imposed on companies.

On the other hand, Christensen and Murphy (2004) are critical of the practices as described above. They argue that tax revenue are the necessary ‘lifeblood' of a democratic government. The authors pose that businesses should have corporate, social and responsibility standards on taxation. The former practices of shareholders wanting companies to pay fewer taxes further leave out the assumption that tax avoidance also comes with non-tax costs. These costs are for example litigation costs, brand costs and

(10)

10 damage to corporate image costs (Scholes et al, 2005). Graham et al (2014) find that managers are hesitant in engaging in tax avoidance as they are afraid that they will suffer reputational costs. Austin and Wilson (2015) argue that this concern is well-grounded, as they find that firms can suffer reputational damage from engaging in tax avoidance which can have a negative effect on the company’s share price which is disadvantageous from a shareholder’s viewpoint.

Kim et al (2011) provide strong evidence that tax avoidance is positively associated with the future crash risk of firm-specific returns. According to Desai and Dharmapala (2006), tax avoidance activities provide managers with opportunistic behaviour. Kim et al argue that hiding earnings in tax shelters creates tools for managers to hide negative operating outcome for an extended period of time. This means that bad information is likely to stockpile within the firm, which results in an unexpected crash of an asset at the moment that a certain threshold is passed. This in turn diminishes firms’ results. The evidence from the Kim et al study can be seen as providing evidence that aggressive tax planning facilitates bad news hoarding, and that it increases future crash risk.

Recent years have shown that corporate tax avoidance has increasingly become a major issue (Campbell, 2016). The European Council has adopted a new directive that addresses the main ways in which large companies attempt to reduce their tax liability (EC, 2017). Also in the USA, there has been critique from a public viewpoint on tax avoidance, as Oxfam claims that US listed companies stash $1.3 trillion offshore. Feldstein (1999) shows in his paper that deadweight losses of income tax are ten times as high when tax avoidance is taken into account than when it is not accounted for. This shows that tax avoidance has a bad impact on the finances of the government and therefore a bad governmental influence. Therefore, it is important from a governmental and public viewpoint to know what the drivers of tax avoidance are.

Furthermore, Sikka (2010) argues that there is a conflict in view on tax avoidance between shareholders and the public. Shareholders’ concern is whether a firm is reducing taxes to increase shareholder value, these may therefore benefit from a reduction of taxes. On the other hand, public’s concern is whether a firm pays its share of taxes, therefore avoidance of taxes may be at the expense of the society as a whole. If a firm avoids taxes, the firm’s profitability may rise, but it may also affect the support for governmental infrastructure or social programs. Therefore tax avoidance may be seen as socially irresponsible.

Thus, tax avoidance is seen as both bad from a governmental, and from a societal viewpoint. It could have disadvantageous consequences for shareholders, and there is strong evidence indicating that tax avoidance is strongly associated with bad news hoarding, which in turn increases future crash risk.

(11)

11

2.3 Stakeholder theory and agency theory

In explaining the influence that integrated reporting can have on whether a company engages in tax avoidance, both agency theory and stakeholder theory are used. Therefore, these two are theories are explained below.

Stakeholders are defined by Freeman (1984) as any group who can affect or is affected by the achievements of a firm’s objectives. These can therefore be among others society, government and shareholders. In the stakeholder theory a major objective of firms is to attain the ability to balance the conflicting demands of various stakeholders of the organization (Ansoff, 1965). The stakeholder concept was divided into three models by Freeman (1984) of which the corporate planning and business policy model of the stakeholder concept focuses on getting the approval of corporate strategic decisions by groups that are necessary for the firm to continue to exist. The third model of corporate social responsibility includes external influences that may assume adversarial positions. It is concerned with regulatory and special interest groups with social issues. A major role of corporate governance is to meet stakeholder demands. The more powerful the stakeholders get, the more important it gets to meet their demands for the company. Thus, stakeholder theory is important for the subject as it explains that a company must take into account the demands that stakeholders have.

On top of the stakeholder theory, the agency theory is used in this paper in explaining the influence that integrated reporting can have on whether a company engages in tax avoidance.

Agency theory talks about the agency problem in which there are a principal and an agent. The principal and the agent have to cooperate, but have a different goal and view on labour. The principal delegates work to the agent, who performs that work. There is an information asymmetry between the two players. The agent has all the information on the work, but the principal does not, as he only knows what the agent is telling him. Agency theory tries to explain this relationship by using the concept of a contract (Jensen & Meckling, 1976). There are two problems that can arise in this theory. The first is the risk sharing problem, in which the agent and principal have a different appetite for risk. The second problem is the agency problem which arises when the agent and principal have different desires and when it is difficult for the principal to verify in which actions the agent engages. The agent has an incentive to consume perks because he has the full marginal utility, but does not experience the full costs. The principal on the other hand has the desire to make as much profit (which means the agent must consume as few perks as possible).

(12)

12 This contradiction of interests means that the principal must make monitoring costs to keep track of what the agent is up to (Jensen & Meckling, 1976). The part of agency theory that is important for this paper is that agents can disclose voluntary information, thus making the firm more transparent, to show that they are acting optimally. Think of this as a glass house, one would not live in a glass house (i.e. being transparent), if one would have something to hide (i.e. acting optimally). The same holds for firms. One would not disclose information if one would have something to hide. In the next chapter the hypotheses are developed.

3. Research questions and designs:

3.1 Hypotheses development:

In this chapter the hypotheses are developed. As explained before, integrated reporting and tax avoidance could be related through the usage of the agency theory and the stakeholder theory. This argument is outlined below.

As stakeholder theory shows, firms have to take into account all stakeholders (Ansoff, 1965). By engaging in integrated reporting a company shows how it creates future value for its stakeholders (IIRC, 2016). Agency theory is about the relationship between an agent and a principle (Jensen & Meckling, 1976). Problems arise because the agent has information that the principal does not have, and has different interests than the principal. It is known from this theory that firms (which are the agents in this case) want to show that they are acting optimally to stakeholders (the principal) by disclosing voluntary information, i.e. engaging in integrated reporting. Furthermore, it is known from this theory that a company will not give out a report, if it shows that the firm will affect stakeholders negatively by creating substandard value. As the paragraph about tax avoidance shows, tax avoidance means that firms (1) withhold money that would otherwise go to governments, (2) is considered bad from a societal viewpoint, (3) and could have disadvantageous consequences for shareholders. Thus, tax avoidance can be considered as having a negative influence on stakeholders. Therefore, it is expected that when firms engage in integrated reporting, they are less likely to engage in tax avoidance.

Furthermore, as is explained above in the literature section, Kim et al (2011) study whether firms that exhibit corporate social responsibility (CSR) reporting, also behave in a responsible manner to constrain earnings management. Even though CSR reporting is not the same as Integrated reporting, an integrated report also contains a part of which is based on corporate and social responsibility. Therefore, managers that work for a firm that engages

(13)

13 in integrated reporting also report on corporate and social responsibility. Carroll’s (1979) delineation of a firm’s social responsibilities suggests that CSR firms should strive to make a profit, obey the law, be ethical, and, further, be a good corporate citizen by financially supporting worthy social causes. Numerous theoretical studies on the ethical view of CSR (e.g., Carroll 1979; Donaldson and Preston 1995; Jones 1995; Phillips et al. 2003) argue that there is a moral imperative for managers to ‘‘do the right thing.” Thus, the reporting on integrated financial on the one hand, and corporate, social and responsibility on the other hand, is argued to have managers do the ‘right’ thing. So, they are less likely to engage in tax avoidance, as this is seen as the wrong ‘thing.’

As can be seen from what is stated above, companies that are engaging in integrated reporting are less likely to engage in tax avoidance. As a criterion for engaging in integrated reporting I use the following two norms. This could be either that the company is influenced by integrated reporting through participation in integrated reporting networks, or is influenced by it through engaging in an integrated reporting network on the one hand, or that a company refers to the IIRC framework in their annual report on the other hand. It is necessary to stress here that this test contains companies on which no quality assessments of the engagement in integrated reporting have been conducted. This leads to the following hypothesis 1 that is investigated in this paper which is as follows:

H1: Companies that mention the influence of the IIRC in their annual report or are influenced by the IIRC by engaging in an IR network are less likely to engage in tax avoidance than other companies.

As is explained for the first hypothesis, the criterion of integrated reporting does not contain any assessment whatsoever of the quality of the integrated report. This enhances that under the former hypothesis, companies are less strictly held to what they have to report. In the second hypothesis, I study whether it makes a difference when companies report according to the IIRC guidelines as drawn up by the international integrated reporting council then when they just mention it or are influenced by it. When using this framework, companies are not just randomly providing voluntary information, but are held to the boundaries of a framework. Thus, companies have to provide valuable voluntary information. The threshold for doing so is higher than for providing random voluntary information. As is explained above, a company does not give out a report if it does not show that they act optimally to the stakeholders of the company (Jensen & Meckling, 1976). Thus, it is more likely that firms that are held more strictly to providing voluntary information (i.e. following the IIRC guidelines), are less likely to engage in creating substandard value (i.e. tax avoidance).

(14)

14 Therefore, I predict that when firms report according to the IIRC guidelines, they are less likely to engage in tax avoidance.

H2: Firms that engage in integrated reporting according to the IIRC framework are less likely to engage in tax avoidance

The data that is used in investigating the hypothesis is drawn up in the next paragraph. This is followed by the operationalization of the variables.

3.2 Empirical basis:

This chapter explains the empirical basis that the tests are being based upon. To test the first hypothesis I drew up, I require data both from companies that did and did not engage in integrated reporting. The IIRC has posted a list of companies worldwide that either referred to the framework or were influenced by the framework by engaging in integrated reporting networks in their annual report. This sample contains 554 firms, that are used for checking the correlation. I use 2015 and 2016 as years in which is tested. the variable of firms that refer to the framework or are influenced by it is seen as a dummy variable, only able to contain the values 0 and 1. The WRDS database has been used to find information on companies from the sample. Firms about which it was not possible to obtain sufficient information have been deleted from the sample. The sample that has been used contains 327 firm-years of which approximately 67% refer to the IIRC or are influenced by it, and 33% of the companies are not influenced by it, nor refer to it. Further information on the sample selection can be found in appendix A.

For the second hypothesis the same list on the IIRC website has been used when checking for integrated reporting. The companies’ websites on the list have been consulted to check whether or not firms engage in integrated reporting and from what year onwards. This led to a sample of firms that engage in integrated reporting. The sample ranges from 2013 (the first operating year of the IIRC) until 2016 (the latest financial report) in which there is a difference in the initial year that firms started to engage in integrated reporting. Firms about which there was no sufficient available information have been deleted from the sample. In the WRDS database information has been found about the companies that has been used to determine the effective tax rate of the company. Firms that did not have available information about the cash tax paid, the pre-tax income, and/or the special items have also been deleted from the sample. Since integrated reporting is a relatively new subject, and because of the other restrictions, the sample of firms engaging in integrated

(15)

15 reporting contains 385 firm-years. Of this sample 47 % of the firms engage in integrated reporting according to the IIRC framework. Further information on the sample selection can be found in appendix A.

Due to the fact that integrated reporting has only recently started to grow, there is an absence of data availability for firms engaging in integrated reporting. Firms only have the opportunity to refer to the IIRC or adjust their report according to the IIRC framework from 2013 onwards. Especially for the second hypothesis, there is no abundance of information on firms engaging in integrated reporting. This results in the relatively small samples that the hypotheses consist of.

The engagement in integrated reporting has been measured as a dummy variable, in which firms that do engage in integrated reporting are assigned a 1 and firms that do not engage in this type of reporting get assigned a 0. The next chapter explains the research model that is used to test the relationship between integrated reporting and tax avoidance.

3.3 Research model:

This chapter outlines the research model that is used to test the hypotheses. In this paper, an archival database study is conducted, using a cross sectional model, for most firms for each year since 2014. For the first hypothesis, the independent variable of the model that will be used to assess the research question, is the dummy variable of whether or not a firm refers to the IIRC in their annual report, or is influenced by it through engaging in an integrated reporting network. For the second hypothesis, the variable will be whether the annual report of the company is prepared according to the guidelines of the IIRC framework. Further investigation on my part will show whether integrated reporting has an influence on tax avoidance.

The dependent variable for both hypotheses is whether firms engaged in tax avoidance. Tax avoidance is impossible to measure directly. Therefore, a proxy is needed for this variable. Tax avoidance can be measured with either book-tax differences (BTD), or effective tax rate (ETR) measures. ETR measures are used in this case, as Guenther argues that BTD measures do not provide incremental information over ETR measures (2014). If the ETR is lower than average for a comparable company, the company engages in tax avoidance.

There are several measures for ETR that can be used, such as GAAP effective tax rate (ETR), or cash ETR (Dyreng et al, 2008). As argued by Dyreng et al. a disadvantage of using GAAP measure is that it both includes current and deferred taxes. A great part of tax avoidance involves accelerating deductions and deferring income for tax purposes relative to

(16)

16 book purposes. Dyreng et al. argue that this will be missed when using GAAP ETR, and looking at this reason, they argue that the cash effective tax rate should be used. This is therefore what is used in this paper, and it can be calculated as follows:

To control for other influences, some control variables must be used. Since the second hypothesis merely tests the same relationship as the first hypothesis does, only with a more strictly defined criterion for integrated reporting that a report must meet, approximately the same model can be used for the two hypotheses. Thus, the same control variables can be used under the different hypotheses, as explained below.

First, leverage (LEV) is added to the control variables. Firms that use more debt financing in comparison to equity financing may need less tax avoidance, as they have a higher deduction of interest expense (Desai & Dharmapala, 2009). Secondly, according to Chen et al (2010), firms that have a higher profit could have more incentives to engage in tax avoidance. Therefore, the proxy of return on assets (ROA) is controlled for. Third, companies that have more property, plant and equipment are likely to have higher depreciation expenses, and hence pay lower taxes than companies that have lower property, plant, and equipment. Therefore, the former are likely to pay less taxes than the latter companies. This means that (PPE) is also controlled for. Fourth, as R&D (RD) and intangible assets (INTA) are also being controlled for because of the different treatments under book and tax of intangibles and consolidated earnings according to the equity method (Rego & Wilson, 2012). To test the relationship between tax avoidance and integrated reporting under the first hypothesis, the following empirical model is used:

ETR = 𝜷0 + 𝜷1 IR + 𝜷2 ROA + 𝜷3 PPE + 𝜷4 RD + 𝜷5 INTA + ε

The ETR is measured by the effective tax rate as is explained above and is the dependent variable. A higher ETR means lower tax avoidance. A positive relationship between the independent variables and the ETR means that the independent variables have a positive influence on the effective tax rate. Mind you however, the positive influence of the independent variables on the ETR, means a negative influence on the tax avoidance of the firm. The variable of integrated reporting (IR) is measured as a dummy variable and can therefore only have two values. The value is 1 if a company engaged in integrated reporting in a year, but the value is 0 if the company has not engaged in this type of reporting in a year. Under H1, a company engages in integrated reporting if it is influenced by the IIRC by

(17)

17 participating in an integrated reporting network, or if it refers to it in its annual report. If the 𝜷1 is statistically significant different from 0, the null hypothesis can be rejected, meaning that there is a relationship between integrated reporting as defined under this hypothesis, and tax avoidance.

To test the relationship between integrated reporting and tax avoidance as defined under the second hypothesis, the following model is used, which is similar to the model under the first hypothesis:

ETR = 𝜷0 + 𝜷1 IIRC + 𝜷2 ROA + 𝜷3 PPE + 𝜷4 RD + 𝜷5 INTA + ε

The only difference between this model and the one under H1 is the dummy variable which has a value of 1 when a company engages in integrated reporting along the guidelines of the IIRC framework and 0 if it does not. If the 𝜷1 is statistically significant different from 0, the null hypothesis can be rejected, meaning that there is a relationship between integrated reporting as defined under this hypothesis, and tax avoidance. The next chapter explains the empirical results that are derived in this study.

4. Empirical results:

This chapter outlines the tests that have been conducted on the two hypotheses that I drew up. As explained above, the ETR of the company is used as a proxy for tax avoidance. Therefore, both correlating and descriptive statistics have been used to test this hypothesis. First, I review the first hypothesis fully, then I investigate the second hypothesis.

4.1 Results for hypothesis 1:

The first hypothesis was formulated as follows:

The following table shows the descriptive statistics for the first hypothesis. As can be observed in the table, there are 327 observations. These stand for the 327 firm-years that

(18)

18 have been measured when firm years about which there was no sufficient information were eliminated and outliers were taken out over 2015 and 2016. The table includes the dependent variable, dummy variable and the dummy variable.

Table 1: Descriptive statistics for H1

Variable Obs Mean Std. Dev. Min Max A 327 0.211794 0.199107 -0.53731 0.97976 B 327 0.657492 0.475276 0 1 C 327 0.015551 0.209593 -2.72278 0.388819 D 327 76468.57 570971.8 0.094 7048000 E 327 395.281 2658.486 0 40508 F 327 27842.04 251407.8 0 3213360

It can be seen from the table above that the mean of the ETR is 0.211794. This means that the effective tax rate is just above 21%. When comparing this to Dyreng et al (2008), it can be seen that the percentage I find is lower. However, the researchers in Dyreng et al eliminate negative ETR’s, possibly resulting from loss carry forwards. Because I let these kind of tax payments in the sample, it can be explained that the ETR that I find is lower. The mean of the dummy variable (IR) is 0.657492 meaning that 65.75% of the sample has released an integrated report. The mean of the ROA is 1.6%, comparable to what I find in hypothesis two. The minimum of the ROA is exceptionally low. This ROA would mean that the net loss of the firm in question is 2.7 times greater than the value of the assets of the company. It is possible that this is a measuring error. However, there is a possibility that it is not, and therefore I leave it in the sample. The average PPE is 76468.57, the average RD is 395.281, and the average of the intangible assets is 27842.04. The Low mean of the PPE, RD, and INTA in comparison to the maximum can be explained by the high presence of lower values for these assets.

4.1.1 Correlation matrix:

In this paragraph, the pairwise correlation between the independent variable, the dependent variable and the dummy variable is outlined. The numbers that are depicted directly behind the name of the variable represents the correlation between that variable and the variable that it is depicted beneath. The number that is at the line below the correlation represents the significance of the correlation.

(19)

19

Table 2: Correlation matrix for H1

Column1 ETR IR ROA PPE RD INTA

ETR 1

IR 0.0333 1

0.5482

PPE 0.153 0.1849 1

0.0055 0.0008

ROA 0.0649 0.087 0.0074 1

0.2415 0.1163 0.8934

RD 0.0592 0.0927 0.0726 0.271 1

0.2859 0.0942 0.1904 0

INTAN 0.0771 0.0657 0.0068 0.5312 0.5948 1

0.1642 0.2358 0.902 0 0

The first thing that is noticeable from this matrix is that the correlation between IR and ETR is low, and is statistically not significant when using a p-value of 0.05. The correlation between ROA and ETR, and IR respectively is larger and also statistically more significant. this means that ROA is better correlated with ETR and IR, than IR and ETR are correlated amongst themselves. Especially the PPE has a strong correlation with ETR, and is also statistically significant different from zero. This means that this is the dependent variable that explains most of the ETR of a firm, from the variables that I have chosen. The correlation among the control variables is also not particularly high.

This means that there is a low chance of any existing multi-collinearity. However, the chance of multicollinearity must be tested, and therefore I use the variance inflation test as can be seen in the table below. The mean VIF is 1.13, and the highest VIF is 1.28. This shows that there is no multicollinearity in the sample.

(20)

20

Table 3: VIF for H1

Variable VIF 1/VIF INTAN 2.01 0.497325

RD 1.57 0.636568

PPE 1.41 0.711525

IR 1.05 0.954907

ROA 1.04 0.960431

Mean VIF 1.42

4.1.2 Regression results:

The first hypothesis that I drew up is the following:

To test this hypothesis, I use a regression model that controls for other influences on tax avoidance. As is explained above, the following formula is used to test this effect:

ETR = 𝜷0 + 𝜷1 IR + 𝜷2 ROA + 𝜷3 PPE + 𝜷4 RD + 𝜷5 INTA + ε

On the following page the regression results for the first hypothesis are presented in

a table and explained thereafter.

(21)

21

Table 4: Results for H1

Source SS df MS

Model 0.387789 5 0.077558 Prob > F= 0.0804

Residual 12.536 321 0.039053 R-squared= 0.03

Adj R-_squared= 0.0149

Total 12.92379 326 0.039644 Root MSE= 0.19762

A Coef. Std. Err. t P>t [95% Conf. Interval]

B -0.00082 0.023566 -0.03 0.972 -0.04718 0.045548 C 0.144665 0.053285 2.71 0.007 0.039832 0.249497 D 1.16E-08 2.27E-08 0.51 0.611

-3.31E-08 5.63E-08 E 5.19E-07 5.16E-06 0.1 0.92

-9.63E-06 1.07E-05 F 4.31E-08 6.17E-08 0.7 0.485

-7.83E-08 1.65E-07 _cons 0.20779 0.018786 11.06 0 0.170831 0.244748

The R squared equals 0.0300 for this model, so 3%. This means that 97% of the variation in ETR is not explained by the model. The adjusted R-squared is 1.49% and is a measure to show how well the regression model fits the observations. It shows what part of the total variation is explained by the variation that is being tested.

The coefficient of the dummy variable (IR) is negative, meaning that when a report is an integrated report, the etr is lower. This means that there is more tax avoidance when companies refer to the IIRC or are influenced by it, according to this sample. However, it is highly insignificant, as the p-value equals 0.972 which is higher than the highest statistically significant p-value of 10%. Only the ROA is positively correlated with the ETR, and is statistically significant when applying a p-value of 1%. The RD, PPE and INTA are negatively associated but are also statistically insignificant.

Based on this evidence the null hypothesis can be accepted: companies that mention

the IIRC in their annual report, or that are influenced by the IIRC by engaging in an IIRC network do not engage less in tax avoidance than do other companies. The result that the

dummy variable is not statistically significant correlated with ETR is not what was predicted by the hypothesis. This can have different causes. First, it could be due to the small sample size. As is explained above, because integrated reporting is a relatively new phenomenon,

(22)

22 there is a lack of data availability on companies that engage in integrated reporting. Second, the proxy of effective tax rate for tax avoidance could be an additional explanation of the result. As explained above, there are multiple proxies that can be used for tax avoidance. It might be a direction for future research to look at other proxies for tax avoidance. Another reason for the result could be that tax avoidance is not seen as something that is ‘bad.’ Wang (2010) finds that more transparent firms, so firms that engage in integrated reporting and have less agency problems, are more likely to engage in tax avoidance. Income taxes namely represent a significant cost of doing business and they reduce the profit after taxes for the firm. As firms are more transparent, it is easier to see for shareholders what is happening on the inside. Francis and Martin (2010) argue that this means that interests are more aligned with shareholders of transparent firms because it is easier for investors to monitor the firm. As it is in the shareholders advantage to engage in tax avoidance, managers of firms that are more transparent are also engaging more in tax avoidance. To the same effect, Desai and Dharmapala (2006) find that high powered incentives lead to less tax avoidance. However, this effect could be neglected for firms that are well governed, and more transparent (i.e. engaging in integrated reporting). Thus, in case that there are some firms with high-powered incentives in this sample, this could well be a reason for the lack of statistically significant evidence that I find. Next to that, Flower (2015) argues that the IIRC is a story of failure. The principal objective of the commission was the promotion of sustainability accounting. However, he argues, the IIRC has abandoned this objective, and has defined value as value for investors, and not value for society. This means that companies engaging in integrated reporting, are not necessarily taking into account society, which means they could be only focused on creating value for investors, which is the same as is the case under normal financial reporting. This could be another reason for the results that got out of this model. Also, Badertscher et al (2013) examine the variety in corporate tax avoidance for firms that are privately owned versus firms that are publicly traded. They do this as a way of looking at tax avoidance from an agency theory perspective. When the owners of a firm are more concentrated in a smaller group, they will probably be more risk averse, and therefore willing to invest less in risky projects according to the researchers. As tax avoidance is a risky activity, the researchers hypothesize and find that companies which have less agency costs due to the private ownership of the company, engage less in tax avoidance. It could be the case that the sample that was picked here contains a lot of firms with a lot of owners explaining why I did not get any results that are statistically significant from this sample. Future research should control for this variable by putting it in the research model.

(23)

23 4.2 Results for hypothesis 2:

The second hypothesis is formulated as follows:

The following table one shows the descriptive statistics for the hypothesis. As can be seen below there are 385 observations, which represent the 385 firm years from the sample as explained above. The table includes the effective tax rate, the binary variable of integrated reporting and the control variables.

Table 5: Descriptive statistics for H2

Variable

Obs Mean Std. Dev. Min Max

ETR 385 0.2184427 0.7187956 -0.8236515 13.57547 IIRC 385 0.4701299 0.4997564 0 1 ROA 385 0.0246762 0.1660795 -1.306986 0.6780942 PPE 385 21230.22 94811.12 0 979858 RD 385 56.99064 285.0874 -0.8 3044 INTAN 385 3885.1 15017.94 0 204108

As can be seen in table 1 above, the sample mean of the effective tax rate is 0.218443, meaning that the 385 firms on average pay a little less than 22% taxes. This is lower in comparison to the amount of taxes paid that Dyreng et al (2008) find. However, this can be explained by the fact that they take out negative taxes, where I did not take the negative taxes out of the sample. This is the same reason as in the first hypothesis. Of this sample there are 204 firm years in which is being engaged in integrated reporting and 181 firms in which is not being engaged in integrated reporting. This leads to a mean of the dummy variable (IIRC) of 0.47013. The mean of the (ROA) is 0.024676 meaning that the average of the ROA of the firms is approximately 2.5%. The average (PPE) is 21230.22, (RD) on average equals 56.99, and (INTA) have a mean of 3885.1. The sample contains a lot of firm

(24)

24 years that have zero for their RD. This can explain the low mean in comparison to the maximum.

4.2.1 Correlation matrix:

In this paragraph, the pairwise correlation between the independent variable, the dependent variable and the dummy variable is outlined. The numbers that are depicted directly behind the name of the variable represents the correlation between that variable and the variable that it is depicted beneath. The number that is at the line below the correlation represents the significance of the correlation.

Table 6: Correlation matrix for H2

ETR IIRC ROA PPE RD INTAN

ETR 1 IIRC 0.0818 1 0.1092 ROA 0.0418 0.1393 1 0.4137 0.0062 PPE -0.0099 -0.0494 0.2145 1 0.8458 0.3342 0 RD 0.0064 -0.0031 0.0316 0.04 1 0.9005 0.9516 0.5362 0.4335 INTAN -0.0138 -0.0707 0.0109 0.0583 0.1868 1 0.7868 0.1665 0.8306 0.254 0.0002

The first thing that is noticeable from this matrix is that the correlation between IR and ETR is low, and is statistically not significant when using a p-value of 0.1. The correlation is positive however, meaning that when a company engages in integrated reporting, it on average engages more in tax avoidance. However, compared to the first hypothesis, the correlation is higher, on top of the higher significance of the correlation. The correlation

(25)

25 between of both ROA and RD on the one hand and ETR on the other hand is lower, accompanied with a lower statistical significance. The PPE and intangibles both are negatively correlated to ETR, meaning that higher PPE or intangibles would be associated with a lower ETR. These two variables also have exceptionally low statistical significance. In this sample, the IIRC is the variable that correlates most with ETR. The correlation among the control variables furthermore is not particularly high.

This means that there is a low chance of any existing multi-collinearity. However, the chance of multicollinearity must be tested, and therefore I use the variance inflation test as can be seen in the table below. The mean VIF is 1.13, and the highest VIF is 1.07. This shows that there is no multicollinearity in the sample.

Table 7: VIF for H2

Variable VIF 1/VIF

ROA 1.07 0.930932 PPE 1.06 0.944521 INTAN 1.04 0.957922 RD 1.04 0.96362 IIRC 1.03 0.969382 Mean VIF 1.05

4.2.2 Regression results:

The second hypothesis that has been drawn up is the following:

To test this hypothesis, I use the same regression model as has been used to test the first hypothesis.It controls for other influences than integrated reporting under the IIRC framework on tax avoidance. As is explained above, the following formula is used to test this effect:

ETR = 𝜷0 + 𝜷1 IIRC + 𝜷2 ROA + 𝜷3 PPE + 𝜷4 RD + 𝜷5 INTA + ε

(26)

26 In the table below, the results of the regression are presented.

Table 8: Results for H2

SS df MS Model 1.5721972 1 5 0.3144394 4 Number of obs 385 Residual 196.82795 4 379 0.5193349 7 F(5, 379) 0.61 Prob > F 0.6958 Total 198.40015 2 384 0.5166670 6 R-squared 0.0079 Adj R-squared -0.0052 Root MSE 0.72065

ETR Coef. Std. Err. t P>t

[95% Conf. Interval] IIRC 0.1089362 0.074739 9 1.46 0.146 -0.0380205 0.255893 ROA 0.1468195 0.229500 2 0.64 0.523 -0.3044336 0.598072 6 PPE -1.00E-07 3.99E-07 -0.25 0.802 -8.85E-07 6.84E-07

RD 0.0000198

0.000131

4 0.15 0.88 -0.0002385

0.000278 2 INTAN -4.57E-07 2.50E-06 -0.18 0.855 -5.38E-06 4.46E-06

_cons 0.1663814

0.052960

5 3.14 0.002 0.0622482

0.270514 6

The R squared equals 0.0079 for this model, so 7,94%. This means that 98.21% of the variation is not explained by the model. The adjusted R-squared is -0.52% and is a measure to show how well the regression model fits the observations. This means that the model is a poor representation of the data. The R-squared and adjusted R-squared are both a worse representation of the data than they were under the former hypothesis.

The coefficient of the dummy variable (IIRC) is positive, meaning that when a report is an integrated report, the etr is higher. This means that there is more tax avoidance, when

(27)

27 a firm engages in integrated reporting according to the sample. However, it is insignificant, as the p-value equals 0.146 which is higher than the highest statistically significant p-value of 10%. The ROA and RD are positively correlated with the ETR, but are both not statistically significant when applying a p-value of 10%. Both the PPE and INTA are negatively associated and are also statistically insignificant.

Based on this evidence the null hypothesis can be accepted: engaging in integrated

reporting according to the IIRC framework does not have an effect on tax avoidance. The

result that the dummy variable is not statistically significant correlated with ETR is not what was predicted by the hypothesis. This can have different causes. First, as is also explained for H1, it could be due to the small sample size, or to the proxy that has been used for tax avoidance. It might be a good direction for future research to look at other proxies for tax avoidance. Furthermore, as is explained under hypothesis 1, Wang (2010) finds that more transparent firms (i.e. firms that engage in integrated reporting) are more likely to engage in tax avoidance. Another reason could be that, as is also explained under the first hypothesis, Flower (2015) argues that the IIRC’s primary objective is not sustainable reporting anymore, which can be seen from the fact that it defines value as value for investors, instead of value for society. This means that companies engaging in integrated reporting do not necessarily have to take into account society when preparing an annual report in an integrated way. This enhances that they could be only focused on creating value for shareholders, similar to the case under regular financial reporting. To the same effect, Desai and Dharmapala (2006) find that for firms with high powered incentives, more transparent firms (i.e. ones engaging in integrated reporting) nullify the, otherwise present, effect that firms engage in less tax avoidance. This could potentially also be a reason for the lack of statistically significant data that I got. Also, Badertscher et al (2013) examine the variety in corporate tax avoidance for firms that are privately owned versus firms that are publicly traded. They do this as a way of looking at tax avoidance from an agency theory perspective. When the owners of a firm are more concentrated in a smaller group, they will probably be more risk averse, and therefore willing to invest less in risky projects according to the researchers. As tax avoidance is a risky activity, the researchers hypothesize and find that companies which have less agency costs due to the private ownership of the company, engage less in tax avoidance. This could also explain the results that I got, and therefore future research must control for this variable. The next chapter contains the conclusion of this study.

(28)

28 5. Conclusion:

This chapter contains some concluding remarks and suggestions for future research. Recent years has shown that tax avoidance has increasingly become a major issue. Tax avoidance has its merits for shareholders, as it increases profits after tax. However, it also has downsides. It not only puts a burden on governments that have less tax income, but also on the reputation of a company. Next to that, tax avoidance is found to be positively associated with bad news hoarding. This leads to future crash risk, which in turn is disadvantageous for shareholders. On the other side, corporate, social, and responsibility reporting has been on the rise. However, the problem with this kind of reporting is that it separates financial and non-financial information. These two kinds of information depend on each other, and to overcome this problem integrated reporting was brought to life. This kind of reporting shows how a firm creates value over the short, medium and long term, and integrates both financial and non-financial information.

In this thesis I investigate the link between integrated reporting and tax avoidance by looking at the following research question: What is the effect of engaging in integrated

reporting on engaging in tax avoidance. The link between the two can exist because of the

following. According to stakeholder theory, a major objective of a firm is to attain the ability to balance the conflicting demands of various stakeholders of the organization. From agency theory it is known that firms want to show that they act optimally by disclosing voluntary information (i.e. engaging in integrated reporting). Companies engaging in tax avoidance can be seen as not acting optimally because of the following. It means less money is available to the government, leaving less money available to invest in society. Furthermore, tax avoidance is found to be associated with bad news hoarding which could lead to future crash risk of the company. Therefore, a firm can be seen as not acting optimally to its stakeholders when it is engaging in tax avoidance. Accordingly, I predict that engaging in integrated reporting has a diminishing effect on engaging in tax avoidance.

To answer the research question, two hypotheses are being identified. First, I look at companies that mention the IIRC in their annual report, or are influenced by integrated reporting through participation in integrated reporting networks. This leads to the following hypothesis:

(29)

29 To test this hypothesis, I use effective tax rate as a proxy for tax avoidance. The higher the ETR, the lower the tax avoidance. I use a dummy variable to indicate whether a company engages in integrated reporting or not. On top of that, I use four variables to control for any other influences on the effective tax rate, namely return on assets (ROA), property plant and equipment (PPE), research and development (RD), and Intangible assets (INTAN).

Second, I look at a more strict definition of companies engaging in integrated reporting. Namely, companies that engage in integrated reporting according to the IIRC framework. This leads to the following hypothesis:

To test this hypothesis, I use effective tax rate as a proxy for tax avoidance. I use the same model for the second hypothesis as I did for the first hypothesis.

I find for the first hypothesis that there is no statistically significant relationship between the fact that a company is influenced by the IIRC, or mentions the IIRC in their annual statement, and the increase in effective tax rate. This indicates the absence of a relationship between the fact that a company is influenced by the IIRC, or mentions the IIRC in their annual statement, and the fact that a company engages in tax avoidance. The second hypothesis reveals similar results. There is no statistically significant relationship between the fact that a company engages in integrated reporting according to the IIRC framework and the effective tax rate, and therefore tax avoidance.

The reason for the lack of a statistically significant relationship could be due to the small samples that I have used to test both hypotheses. Due to the fact that integrated reporting is a relatively new phenomenon, there is a lack of data for companies that engage in integrated reporting. Furthermore, the proxy that is used for tax avoidance could provide different results than other proxies would provide. This could explain the lack of statistically significant evidence for the two hypotheses.

The contribution of this study is twofold as it contributes both to the academic world and outside the academic world. First, this paper adds to the amount of literature on tax avoidance. To have a thorough understanding of tax avoidance it is important to understand what phenomenon do or do not drive of it. Second, because of the relative novelty of integrated reporting, there is no abundance of research on this topic. By looking at the possible effects it has on tax avoidance, the advantages and disadvantages of integrated reporting are investigated. Furthermore, this study contributes to what is not the academic world. Third, as governments and other stakeholders have the possibility of losing money