Reputation and corporate tax avoidance in the FTSE 100

BEARING THE BURDEN

Reputation and corporate tax avoidance in the FTSE 100

J. P. Rasch 13 March 2016

ABSTRACT

Encouraged by recent events and research in the U.K., this thesis strives to establish relationships between tax avoidance, corporate social responsibility (CSR) and reputational concerns in the context of the FTSE 100. While thoroughly researched individually, the hypothesized common ground in between aforementioned themes has not been studied extensively and constitutes a research gap worthy of investigation. By searching the LexisNexis news database for cases of tax shaming of firms in the media, the approach is to measure whether firms’ perceive this pressure as problematic enough to change their behaviour. Empirical results from observing if firms bear a higher tax burden or propagate CSR after being pressurized succeed in confirming that these firms increase their tax payments. Another matter is whether CSR also applies to tax strategy, by studying if companies propagating responsible conduct also pay the public’s perceived fair share of tax. The results suggest that companies which issue a CSR report appear to actually pay less tax than companies who do not.

This implies that tax avoidance is not viewed as conflicting with CSR and constitutes a different topic.

What these outcomes suggest is that firms which traditionally issue CSR reports bear a lower tax burden than other companies. This could be explained by the “insurance” hypothesis from Godfrey (2005) and legitimacy theory (Dowling & Pfeffer, 1975). By creating a perception of being a good corporate citizen, companies appear to earn goodwill from the public, which will decrease the chances of being scrutinized by the public. The findings in this thesis extend prior research by finding evidence suggesting that companies can use symbolic management to mitigate chances of public scrutiny while reaping the benefits of actions which can be perceived by the public as problematic.

KEYWORDS

Corporate social responsibility, corporate tax avoidance, legitimacy theory, reputational costs,

shareholder value maximization, stakeholder theory

BEARING THE BURDEN

Reputation and corporate tax avoidance in the FTSE 100

Master’s Thesis Accountancy

Name : J. P. (Jan) Rasch

Student number : S2518872

Date of birth : 7 October 1991

E-mail address : j.p.rasch@student.rug.nl | j.rasch@outlook.com University : University of Groningen (Rijksuniversiteit Groningen)

Faculty : Economics and Business

Study Program : MSc Accountancy & Controlling (track: Accountancy) Supervisor (university) : dr. S. (Soojin) Lee

Co-supervisor (university) : dr. R. B. H. (Reggy) Hooghiemstra

Internship organization : Ernst & Young Accountants LLP (Amsterdam) Supervisor (organization) : J. H. (Johan) Hennipman MSc

Word count (main body) : 11.181

TABLE OF CONTENTS

I. INTRODUCTION ... 1

1.1 STATEMENT OF THE PROBLEM ... 2

1.2 RELEVANCE ... 3

1.3 STRUCTURE ... 3

II. THEORETICAL BACKGROUND AND HYPOTHESES DEVELOPMENT ... 4

2.1 THEORETICAL BACKGROUND ... 4

2.2 HYPOTHESES DEVELOPMENT ... 8

III. RESEARCH DESIGN ... 11

3.1 SAMPLE SELECTION ... 11

3.2 DEPENDENT VARIABLES ... 13

3.3 INDEPENDENT VARIABLES ... 13

3.4 CONTROL VARIABLES ... 14

3.5 MULTIVARIATE MODELS... 15

IV. RESULTS ... 17

4.1 DESCRIPTIVE STATISTICS... 17

4.2 MULTIVARIATE RESULTS ... 23

V. DISCUSSION AND CONCLUSION ... 29

5.1 CONCLUSION ... 29

5.2 IMPLICATIONS FOR PRACTICE AND THEORY ... 30

5.3 LIMITATIONS AND AVENUES FOR FUTURE RESEARCH ... 31

ACKNOWLEDGEMENTS ... 33

BIBLIOGRAPHY ... 34

1

I. INTRODUCTION

“When there is an income tax, the just man will pay more and the unjust less on the same amount of income.”

- Plato, Philosopher (Plato, Republic I, 343d).

In current times, various tax avoidance scandals have surfaced, which attracted increasing amounts of attention from the public, especially in the United Kingdom (Barford & Holt, 2013). The term “tax shaming” has been coined, indicating publicly shaming and criticizing companies for their tax strategy (Dastyari, 2015). Although tax avoidance is not illegal, companies practicing it can invoke the public’s anger by not paying the public’s perceived fair share of taxes. Public outrage is fuelled by the

dissatisfaction that companies can use methods unavailable to the common citizen to legally avoid taxes (Brown, 2012). The public feels cheated as austerity measures mandated by the financial crisis press on their living environment while they might have to carry a disproportionate tax burden because large multinational companies free ride on the tax money they pay (Slemrod, 2004). These circumstances indicate that the public demands more information about tax avoidance and how the dissatisfaction surrounding it can be managed.

One practical example of public pressure altering firm’s tax strategy is the pledge of Starbucks U.K. to pay more tax than legally obliged (Neville & Treanor, 2012). Starbucks U.K.’s managing director explicitly states that their company is heeding their customer’s demands by paying more corporation tax than legally obliged. This company agent will probably have valid reasons for why paying

additional tax will benefit shareholders. The question is whether this incident is unique (Shah, 2013).

In the aforementioned British setting, Dyreng, Hoopes and Wilde (2016) study the relationship between public pressure and the tax strategy of the U.K. FTSE 100. Dyreng et al. (2016) use an event to explain rising effective tax rates of FTSE 100 firms. The event which they study involves ActionAid, a non-profit activist group which tries to combat poverty and who holds corporate tax avoidance as a large contributor to global poverty. In one of their campaigns ActionAid exposed unpublicized

subsidiaries of British FTSE 100 firms in 2010. Since 2006 these companies are legally obliged to disclose all subsidiaries regardless of size and materiality in their company register filings. ActionAid believed that companies not conforming to the disclosure law did so on purpose to hide entities used for tax avoidance activities (ActionAid UK, 2011). In their tests Dyreng et al. (2016) find consistent significant results between public scrutiny (being labelled as a disclosure law violator) and rising effective tax rates in the years following public scrutiny. The conclusion from Dyreng et al. (2016) suggests increased reputational costs of excessive tax avoidance might have caused companies to increase their tax burden to mitigate their chances of being labelled a bad corporate citizen.

To demonstrate the uniqueness of the British example, recent research on the effects of public pressure on tax avoidance in the U.S. setting has not been very successful at inferring relationships.

Gallemore, Maydew and Thornock (2014) observe no significant relationships between firms’

reputation, stock price, executive turnover, market value and income after these firms are labelled as

excessive tax avoiders. In their conclusion, the authors are puzzled why not every large multinational

firm excessively avoids tax if the net trade-off is positive because of the absence of reputational

costs. Perhaps the British FTSE 100 setting constitutes a unique socio-economic climate in the

2 context of reputational costs. The unobserved reputational costs may also be due to the timeframe as Gallemore et al. (2014) study the years 1995-2005, which are before the financial crisis of 2007.

Public outrage could be fuelled because the public faces austerity measures mandated by the financial crisis while companies might not pay the perceived fair share of taxes. Perhaps higher tension makes the reputational costs rise. Lastly, another explanation can be the intensification of internet communication since 2005, making the organization of boycotts and information exchange easier and faster (PWC, 2013).

The observed British shift in tax strategy of large multinational enterprises might indicate that reputational costs in the U.K. have become significant enough to affect corporate tax strategy. The results from Dyreng et al. (2016) provide an opportunity to expand the understanding of the determinants of tax avoidance. Perhaps there are more relationships to be inferred about how the effect of public pressure on a firm’s tax strategy is moderated. Godfrey, Merrill and Hansen (2009) infer that a corporate social responsibility (CSR) image can function as “insurance” for negative publicity ramifications about company actions being deemed immoral. This view is supported by organizational legitimacy theory, which states that companies can position themselves ex-ante as legitimate companies through symbolic management (Ashforth & Gibbs, 1990). These companies are less vulnerable to public scrutiny because they are already established as being legitimate. This study further explores whether this holds for tax avoidance scandals in the British setting.

The contribution of this study is to seek out how reputational concerns of large multinational companies affect the relationship between public scrutiny and tax avoidance in the United Kingdom.

Different proxies will be used to measure how corporate tax strategy reacts to reputational concerns as the contrasting results imply an academic research gap (Gallemore, Maydew, & Thornock, 2014).

1.1 STATEMENT OF THE PROBLEM

By adhering to the shareholder value approach too strictly, company managers can be incentivized to be desensitized. By pursuing the mitigation of corporate tax paying to the absolute minimum as long as the net trade-off between the company’s costs and benefits of tax avoidance is positive can have harmful effects on society (Sikka, 2010). If corporate tax avoidance becomes an important

competition instrument between companies, a “race to the bottom” can ensue (Gunther, 2014).

When corporate tax avoidance grows too rampant collectively, stakeholders might stand up to correct the perceived norm-violating companies (Dyreng, Hoopes, & Wilde, 2016).

Excessive tax avoidance is not punishable by the judiciary power as long as the ambiguous border distinguishing tax avoidance from tax evasion is not crossed. Still, the public can view tax avoidance as being immoral (Slemrod, 2004) and try to punish the companies for it (Grilled over tax avoidance, 2012). This thesis will refrain from taking an ethical standpoint and will look at reputational concerns affecting the relationship between public scrutiny and tax strategy. The contribution of this thesis is to gain deeper understanding of the this relationship. Thus, the main research question is formulated as:

Do reputational concerns influence multinational firms’

tax avoidance behaviour?

3

Before developing the central research question further into hypotheses, the relevance of the problem is advocated in the next paragraph.

1.2 RELEVANCE

As will be apparent from the theoretical background paragraph, much has been studied on the determinants of tax avoidance alone. The avenue of the effects of reputational concerns on tax avoidance is relatively new however. The observed significant effect of the ActionAid event on corporate tax strategy is the first one where public pressure explains differing tax avoidance trends (Dyreng, Hoopes, & Wilde, 2016). This implies that for the first time in tax avoidance literature, stakeholders apart from the shareholders have the power to limit firms’ perceived immoral

behaviour. The research gap here is adding a CSR component which could influence the relationship between reputational concerns and tax strategy. This provides a research opportunity to expand the knowledge about how companies react to reputational concerns and whether CSR influences this relationship.

The choice for the British FTSE 100 setting in this thesis is natural as it capitalizes on the country setting with the first significant relationship between stakeholder pressure and tax strategy (Dyreng, Hoopes, & Wilde, 2016). Choy, Lai and Ng (2015) add the dimension of corporate governance and find positive associations between the perceived corporate governance standards and corporate tax strategy. Choy et al. (2015) find that firms who positioned themselves as having high corporate governance standards prior to the ActionAid event suffered more if they are labelled as a tax avoider.

The increased attention to corporate tax strategy has not escaped the attention of corporations themselves according to surveys by two Big-Four accounting firms. For example, EY surveys aimed at employees in executive or high financial functions of large companies show that concern about media coverage of perceived low tax rates has risen sharply since 2011 (Ernst & Young, 2014). PWC reports similar findings in a survey of CEOs of large companies. PWC (2013) states that corporate taxes are subject to public debate and companies should understand their stakeholders, this notion should then be carefully integrated into their tax strategy communication strategy.

The results in this thesis are mixed in establishing whether public scrutiny of corporate tax strategy will lead to firms correcting their behaviour. What does seem to consistently hold is that companies issuing CSR reports bear a lower tax burden than other companies. This suggests that companies propagate responsible conduct to shield themselves from being labelled a bad corporate citizen. The test results have important implications for company managers concerned with the public’s

perception of their company and also for institutions combating tax avoidance.

1.3 STRUCTURE

The remainder of this thesis proceeds as follows. Section II summarizes earlier research concerning

these subjects and elaborates on the hypotheses developed and tested in this thesis. Section III

introduces the key data and develops models that test the hypotheses. Section IV presents the

results. Finally, section V discusses the research limitations, possible future research avenues and

concludes.

4

II. THEORETICAL BACKGROUND AND HYPOTHESES DEVELOPMENT

“Over and over again courts have said that there is nothing sinister in so arranging one's affairs as to keep taxes as low as possible. Everybody does so, rich or poor; and all do right, for nobody owes any

public duty to pay more than the law demands: taxes are enforced exactions, not voluntary contributions.”

- Learned Hand, Judge (Commissioner v. Newman, 1947).

An important distinction to make when studying corporate tax avoidance, because sometimes the terms are used interchangeably, is the difference between tax avoidance and tax evasion. Taxpayers can utilize certain means which limit the tax revenues of the state whilst doing nothing wrong in the legal sense.

A common way to divide lowering payable tax into two areas is the following:



Tax avoidance is lowering the amount of tax one has to pay through legal means.



Tax evasion is lowering the amount of tax one has to pay through illegal means.

Although seemingly simple, the border between legal and illegal can be vague. To make this difference more distinct established definitions are used. Prior literature defines tax evasion as:

“when a taxpayer fails to declare all or part of his or her income or makes a claim to offset an expense against taxable income that he or she did not incur or was not allowed to claim for tax purposes” (Palan, Murphy, & Chavagneux, 2010, p. 9), and tax avoidance as “the gray [sic] area between tax compliance and tax evasion” (Palan, Murphy, & Chavagneux, 2010, p. 10). The Organisation for Economic Co-operation and Development (OECD), is dedicated to combat tax avoidance. The OECD describes tax avoidance as: “[..] the arrangement of a taxpayer's affairs that is intended to reduce his tax liability and that although the arrangement could be strictly legal it is usually in contradiction with the intent of the law it purports to follow” (OECD, n.d.).

2.1 THEORETICAL BACKGROUND

Tax avoidance

When legislators draft a tax statute, they strive to formulate the text in a way it will encompass the indistinct activities that are targeted by it. Unfortunately for the legislators, these broad definitions may be interpreted differently than the original spirit behind the law. Alternative interpretations according to the letter of the law may be exploited by taxpayers to minimize their tax liabilities (Brown, 2012). Large multinational companies can use experts to design legal ways to pay lower amounts of tax. These designs usually involve complex transactions and legislation of various countries, combining legal loopholes and tax incentives to minimize tax expenses for the company overall (Dyreng & Lindsey, 2009). These legal loopholes can be found between countries without substantial operational opportunities apart from avoiding tax (Roberts, 1995), as well as between countries with operational market potential (Bartelsman & Beetsma, 2003). Modern technologies and terminology make income or capital value continually more difficult to measure. In the current world royalties, transfer pricing, complicated corporate structures and patents can determine the

1 Alternative literature is available (Frey, 1997) about why individuals commit to avoiding taxes, which is beyond the scope of this thesis.

5

profitability of a multinational’s domestic division (Bartelsman & Beetsma, 2003). Organized experts like accountants, lawyers, and tax-specialists try to interpret the law to the advantage of the

company that pays them (Sikka & Hampton, 2005).

Governments and taxpayers are in a constant struggle of debating how (or if) something should be taxed. While the current legal distinctions can be implicitly known, it is hard to determine whether different forms of lowering of taxes owed should be legal. To further complicate this field, metrics for measuring degrees of tax avoidance can contain noise. These metrics contain tax cuts that were intended for stimulating certain activities by implementing certain tax credits, as are unintended tax shelters which can be seen as immoral but not illegal.

The general public has varying opinions on this matter, but in the end the policymakers set the rough boundaries from which they will work.

Hanlon and Heitzman (2010) provide an extensive review of tax avoidance literature until 2010. In their review various perspectives on tax avoidance from various disciplines are summarized. In another article of note by Dyreng, Hanlon and Maydew (2008), they present statistical evidence that tax avoidance is not a reflection of temporary postponement of tax. Literature has seen much use of ETR (the effective tax rate according to the annual report) and Cash ETR (the actual corporation tax paid related to profit before tax) as proxies for tax avoidance. An important remark is that these may contain temporary tax credits caused by earnings management, which will be reversed at a later stage. By measuring long-term effective tax rates, Dyreng et al. (2008) infer that companies not only defer tax, but can permanently and structurally avoid corporate taxation. In their following paper, Dyreng, Hanlon and Maydew (2010) provide compelling evidence that individual executives appear to be instrumental in influencing their firm’s effective tax rates. Following up on executive influence, researchers studying compensation incentives of tax executives find a negative relationship between the degree of performance-based pay and the ETR (Armstrong, Blouin, & Larcker, 2012). These results are consistent with the notion that executives are incentivized to decrease corporate tax payments.

An alternative line within tax avoidance research continues on the relationship between incentive compensation and tax avoidance. This line of research is alternative in the way that they do not examine tax avoidance as a transfer of wealth between state and shareholders, but from an agency theory perspective (Desai & Dharmapala, 2009). Their research implies that the negative reaction to tax avoidance is caused by an agency theory problem. Based on U.S. observations, they suggest that shareholders link tax avoidance to managerial opportunism and rent extraction and therefore encourage agents to abstain from tax avoidance. They carefully present this as a possible explanation for why not every firm engages in excessive tax avoidance if the net effects are perceived to be positive. In their earlier work they find empirical evidence that in weak corporate governance settings where executives receive incentive pay, tax avoidance is much less observed (Desai &

Dharmapala, 2006). The researchers explain these results by suggesting that shareholders or investors can associate corporate tax avoidance with rent extracting by the agent (immoral behaviour could lead to other immoral behaviour). In performance based pay situations they find

2 In the quote mentioned at the start of the introduction, the ancient philosopher Plato is cited, who associates paying the right amount of taxes with the justness of the individual (where the word right is deliberately italicized as the definition of right could spark a lengthy philosophical, ethical and economical debate). The answer to the question of what the equilibrium of taxation is, is beyond the scope of this thesis, apart from being near impossible to ascertain and finally not a static one (Agranov & Palfrey, 2015).

6 significantly more tax avoidance in companies with high corporate governance. They relate this to shareholders and investors having trust in the internal control systems preventing agents from using tax avoidance to cover up managerial opportunism. Desai and Dharmapala (2009) expand this view in a later study by inferring that corporate tax avoidance only significantly improves firm value under the condition of a strong corporate governance setting.

Shareholder value approach

The main goal of commercial businesses is to increase the company’s value (Friedman, 1970). As the chief executives of large multinational companies are usually judged by the amount of income their corporate strategy generates, they are incentivized to maximize corporate income (Eisenhardt, 1988). Managers who meet or beat their principal’s demands for profitability can generally count on continued employment or increasing remuneration (Garen, 1994). According to classic agency theory (Jensen & Meckling, 1976), incentives between agent and principal are aligned by linking executive compensation to shareholder value (Fama & Jensen, 1983).

According to the shareholder approach, taxes are expenses that do not contribute to future growth of the company directly. Even though taxes pay for infrastructure and services with a valuable function for the tax avoider, this is irrelevant for the avoider’s trade-off, as the tax avoider can free ride on the tax payments of others (Slemrod, 2004). Hence, taxes paid can be seen as a loss by agents, as this money could have instead been invested for future profits in the company or paid out as dividends to the shareholders (Eisenhardt, 1989). Agents are incentivized from making an

economic trade-off between the benefits and costs of restructuring the firm’s transactional structure to avoid certain taxes (Slemrod, 2004). Benefits are almost exclusively a lower tax burden, while costs can comprise legal costs, costs of advisory services, fees for restructuring or reputational costs like decreased revenues following public scrutiny (Dyreng, Hoopes, & Wilde, 2016). If the agent concludes that this trade-off will result in a net benefit, the agent is theoretically making the right decision when adhering to the shareholder value approach (Slemrod, 2004).

Following a shareholder value approach, companies will minimize their tax burden to increase profitability. This is consistent with literature which reports tax departments being managed as profit centres, where effective tax rate minimization is more appreciated than tax compliance (Slemrod, 2004). Based on U.S. results, the popular belief that large multinational companies benefit from economies of scale regarding tax avoidance appears to hold (Rego, 2003).

Stakeholder value approach and corporate social responsibility

Researchers have criticized the shareholder approach mentioned above, which led to the

development of the stakeholder approach which is rising in popularity (Freeman, 2010). Contrary to

shareholder theory, this field addresses that company management should carefully weigh the

interests of more parties rather than just the company owners. It adds a dimension of morals and

values to the way business success is monitored as the company management also has to consider

the effects of business strategy on stakeholders other than the shareholders. These stakeholders are

the government, employees and society for example. The government being a stakeholder is further

stressed by the following quote: “The state, thanks to its tax claim on cash flows, is de facto the

largest minority shareholder in almost all corporations” (Desai, Dyck, & Zingales, 2007, p. 592). The

stakeholder approach has gained in popularity the last decade and the theory can be applied to

corporate tax avoidance (Smith, 2003). The consequences of corporate tax avoidance might have a

7

net positive effect on the shareholders. By including the interests of all stakeholders affected by tax avoidance, the trade-off will shift as the negative consequences of paying less tax will probably be larger and might make management conclude that tax avoidance has a net negative effect on the stakeholders, and thus should be refrained from. Dhaliwal, Li, Tsang and Yang (2011) look at stakeholder valuation theory from a shareholder perspective. Their view advocates listening to stakeholders in an extrinsic way, namely to the degree in which it ultimately adds to shareholder value maximization (Shank, Mannulang, & Hill, 2005). Other researchers, usually from fields other than economics and business, believe stakeholder valuation should be practiced for intrinsic reasons (Carroll, 1999).

Related to stakeholder valuation is CSR which states that businesses should not maximize

shareholder value at all costs. Important for this thesis, is that other research finds that having high CSR standards can benefit companies. A history of being a good corporate citizen can function as

“insurance” by protecting them from being labelled a bad corporate citizen and thus mitigating the chances of costly public scrutiny (Godfrey, 2005). He provides the theoretical framework, and more recently, an empirical test that infers that under certain conditions, a good CSR image can function as

“insurance” for negative publicity (Godfrey, Merrill, & Hansen, 2009). These conditions appear to be that the CSR efforts should not be concentrated on supply chain partners but on more distant stakeholders, like society as a whole, to attain scrutiny “insurance”.

When relating corporate actions to mission and vision statements, Sikka (2010) exposes contradictions between words and actions. Mission and vision statements appear to be quite standard in their underlying values and are not competitive, merely a condition. CSR image and tax avoidance do appear to be related. In a U.S. setting, companies with a bad CSR image are associated with more tax avoidance behaviour (Hoi, Wu, & Zhang, 2013). This is consistent with other literature that advocates treating tax avoidance as a CSR issue (Fisher, 2014).

Organizational legitimacy theory

To achieve organizational goals, organizational legitimacy theory states that companies have to be perceived as being legitimate (Dowling & Pfeffer, 1975). The theory states that organizations viewed as being legitimate by society will not incur financial sanctions from stakeholders and gain a

competitive advantage over competitors. A company can be perceived as legitimate when conforming to the social standards of the society in which it operates. Dowling and Pfeffer (1975) posit that companies try to operate in an equilibrium of economically viable, legal and legitimate behaviour. Furthermore, an important division in corporate strategies aiming to achieve legitimacy is between substantive management and symbolic management (Ashforth & Gibbs, 1990). Symbolic management is defined as companies portraying themselves as being legitimate, while actual actions do not necessarily follow this (Ashforth & Gibbs, 1990). The claim of legitimacy can thus, only be in external appearance. Substantive management is regarded as actually changing corporate behaviour to more legitimate standards (Ashforth & Gibbs, 1990). The article further states that companies with a track record of low legitimacy have more difficulty with attaining greater legitimacy because the perceived mismatch in image is greeted with scepticism. This notion is also applied the opposite way,

3 Extrinsic and intrinsic are established adjectives from a field of motivational theory (Ryan & Deci, 2000).

Extrinsic motivation refers to motivation incited by external factors, like a reward for performing an activity, in the broad sense. Intrinsic motivation indicates motivation incited by internal factors, like the personal interest or enjoyment of a certain task. These definitions are also applied in explaining corporate behaviour contexts.

8 where organizations with high legitimacy are less sensitive to criticism on their perceived legitimacy because their established reputation conflicts with the criticism (Ashforth & Gibbs, 1990).

External pressure effects

An extensive survey about which concerns press on tax executives provides results supporting the notion that public opinion matters, especially in publicly traded firms. In determining tax strategy, executives deem reputational risk concerns more important than Internal Revenue Service (IRS, U.S.

tax authority) litigation or financial accounting concerns (Graham, Hanlon, Shevlin, & Shroff, 2014).

In an empirical setting, Pruitt and Friedman (1986) study the effect of organized boycotts on stock prices and meeting corporate targets. The results are mixed with around a quarter of boycotts substantially hurting a firm, the outcomes are attributed to individual case characteristics. More recently, Hanlon and Slemrod (2009) infer that news about tax shelter involvement relates to a declining stock price. Consistent with the notion that public pressure sensitivity depends on certain firm characteristics, firms servicing consumers suffer larger impacts, presumably because they are more sensitive to public boycotts (Hanlon & Slemrod, 2009). Although share prices determine shareholder value, this may only reflect a panic reaction from traders. Monitoring the reaction of sales patterns (sensitive to boycotts) and advertising expenses (proxy for reputational repair expenses) to negative tax avoidance publicity in the U.S., other researchers do not find a significant relationship (Gallemore, Maydew, & Thornock, 2014). They also find that the negative stock reaction they observe fully reverses within a month, casting doubt on the reputational costs in trade-offs.

Next, the theories mentioned above will be developed into hypotheses.

2.2 HYPOTHESES DEVELOPMENT

The study by Dyreng et al. (2016) provides an opportunity to gain more understanding of the effect of reputational concerns on corporate tax strategy. If companies perceive negative coverage of their tax strategy as being harmful to shareholder value maximization, it would mean that companies change their tax strategy toward socially acceptable norms (Dowling & Pfeffer, 1975). Perhaps tax strategy is not strongly associated with responsible societal conduct in the geographic area and timeframe which Gallemore et al. (2014) investigate. For that reason the hypotheses will examine the timeframe after tax strategy might have become publicly relevant to an extent that negative coverage can damage shareholder value maximization. Thus, it would imply that firms perceive negative media coverage as a problem and they will correct their tax strategy to conform to socially acceptable norms. This leads to the first hypothesis:

H1:

FTSE 100 companies which underwent negative media coverage of their tax strategy since 2010 will exhibit decreasing tax avoidance.

When reasoning along the lines of the “insurance” or risk management hypothesis (Godfrey, 2005),

one could expect that to shield a company from negative publicity, it will try to establish itself as

being a good corporate citizen. When applying this perspective to CSR, one might expect that

companies’ efforts to better society give them some positive image goodwill, making them less

vulnerable to public scrutiny (Godfrey, Merrill, & Hansen, 2009). To improve the way a company is

perceived by their stakeholders, one could expect a company to issue a CSR report (or an equivalent)

9

(Ashforth & Gibbs, 1990). In such a document the company discloses how they contribute to society and advocate good corporate citizenship.

This can support the expectation that the pressurized firms were generally less CSR-motivated companies and will start advocating CSR-motivated practices for extrinsic reasons. An explanation could be that pressurized firms now perceive reputational costs as exceeding the costs of building up a CSR image. An established positive CSR image might enable the firm to continue reaping the benefits of avoiding tax, without being scrutinized (Godfrey, Merrill, & Hansen, 2009). Thus, the second hypothesis is formulated as:

H2:

FTSE 100 companies which underwent negative media coverage of their tax strategy are more prone to issue their first CSR report that year than other FTSE 100

companies.

Depending on the answer of the first hypothesis, test results will suggest whether public monitoring has structurally changed the tax conduct of businesses. Regardless of the answer to the first

hypothesis, the recent British events have pointed out that stakeholders care about a company’s tax strategy (Dyreng, Hoopes, & Wilde, 2016). It is up to the companies to value this notion and how they weigh this in defining corporate strategies. Continuing on aforementioned expectations, the third hypothesis will focus on understanding whether CSR is translated into a company’s internal tax strategy, as advocated in certain articles (Fisher, 2014). Whether a particular class of companies considers tax as a CSR issue or that the corporate culture these companies have highlights compliance and being a good corporate citizen in general, cannot be said. As shown in the theory section it has become clear that certain firms do not maximize shareholder value through excessive tax avoidance. Weisbach (2002) is the first to describe this phenomenon, which is later labelled as the “undersheltering puzzle” within the research community. No academic literature has yet given an ultimate answer to what motivates the abstinence of certain companies from tax avoidance while these profit maximizing opportunities are present and legal.

The time span studied for this hypothesis starts in 2010, when companies are presumably aware of the public’s attention to tax strategy which arguably has changed the companies’ perspective on tax strategy. Irrespective of whether actual company behaviour has changed, their perspective on organizational legitimacy probably has changed.

In the U.S. setting the literature mentioned earlier suggests that non-disclosure under a voluntary disclosure regime is associated with tax avoidance (Hope, Ma, & Thomas, 2013). In the same setting, companies with less CSR-motivated ways of running their affairs also appear to be more prone to excessive tax avoidance (Hoi, Wu, & Zhang, 2013). This suggests that through intrinsic or extrinsic motivation of agents and principals, excessive tax avoidance can be a forbidden tool for value maximization. One could expect that these findings from the U.S. can be generalized to the U.K. FTSE 100 (Hoi, Wu, & Zhang, 2013). This would mean that CSR-motivated companies would intrinsically or extrinsically abstain from excessive tax avoidance as they do not want to be perceived as a bad corporate citizen. Thus, the expectation emerges that a group of highly CSR-motivated companies would have higher effective tax rates than companies who are more motivated by other matters.

In contrast, the hypothesis from Godfrey (2005) about “insurance” or risk management suggests the

opposite. Reasoning along these lines, their CSR image and efforts can generate positive image

10 goodwill which might partially insure the company from negative publicity (Godfrey, Merrill, &

Hansen, 2009). The aforementioned “bad corporate citizen” label might not be applicable in this case as the company already contributes their fair share to finance society through other channels. This way, the general public might pardon their perceived free riding (Ashforth & Gibbs, 1990). Lastly, arguing from the shareholder value approach (Friedman, 1970), company agents could try to compensate for their “corporate charity costs” from CSR initiatives with gains from lower effective tax rates. The CSR image could also be a brand image building strategy enforced separate from company tax strategy, which combined, aims to maximize shareholder value. These contradictory arguments lead to formulating the third hypothesis without an expected direction.

To measure CSR motivation, two indicators are utilized. One being the issuance of a CSR report in a given year, which as explained earlier, pertains disclosure about CSR being weighted into corporate strategy. An important limitation of this indicator is that it denotes CSR motivation voiced by the company. These plans and reports are made by the company itself and are not always audited and might include only symbolic management (Ashforth & Gibbs, 1990). Thus, they can contain a positive image which will not perfectly translate in company actions (Sikka, 2010). CSR can be perceived challenging to measure at a glance and it is very subjective for (potential) investors to assess how CSR-motivated a company is. To provide for this need, the British FTSE Group, a provider of stock market indices, has launched the FTSE4Good indices in 2001. These indices comprise companies which are identified as demonstrating care for the environment, social stakeholders and governance (Inman, 2001).

One of these indices is the FTSE4Good U.K. 50 index, which as the name implies, encompasses fifty companies from the U.K. FTSE 100. The important aspect is that company actions are taken into account, which Ashforth and Gibbs (1990) classify as substantive management. The variants of the third and final hypothesis are formulated as:

H3A: FTSE 100 companies who issue CSR reports practice different degrees of tax

avoidance than companies who do not issue CSR reports.

H3B:

FTSE 100 companies who are also constituents of the FTSE4Good U.K. index practice different degrees of tax avoidance than non-FTSE4Good U.K. companies.

4 The inclusion criteria are elaborated in a report issued by the group (FTSE, 2015).

11

III. RESEARCH DESIGN 3.1 SAMPLE SELECTION

The sample which is analysed comprises the FTSE 100 companies listed on 30 June 2015. The time span which will be observed encompasses five years, being from the start of 2010 until the end of 2014. This way the sample covers the years after public pressure presumably changed companies’

perceptions of negative publicity trade-offs.

The sample can provide a maximum of 500 firm-year observations, the amount of observations depends on the minimal data requirements per hypothesis. All annual financial company data is collected from Thomson Reuters’ Worldscope database and missing data is manually added when reliably available.

Table 1 presents the amount of firms which were tax shamed by the public by year. The data clearly indicates that most tax shaming occurred in 2013. This could be explained by Starbucks yielding to public pressure in 2012, which might have encouraged the public to scrutinize more firms for excessive tax avoidance. The count of tax-shamed firms differs between hypotheses because the minimal data requirements lead to dropping some of these firms.

Hypothesis 1 sample

The minimal data requirements for the first hypothesis are non-missing values for ETR and Cash ETR.

Additionally, positive pre-tax income is required for interpretable (Cash) ETRs. Leverage and total

assets need to be non-missing for control purposes. Control variables regarding intensity dimensions

which are ultimately missing are adjusted to zero. Consequently, 105 firm-year observations are

dropped due to negative pre-tax income , the dataset which remains consists of 395 firm-year

12 observations which correspond to 79 unique firms, consisting of 100 firm-year observations which correspond to 20 unique firms (treatment group). The other 59 unique firms form the control firms, corresponding to 295 firm-year observations, which will be matched. Tax-shamed (treatment) firms and control firms are matched using industry (ICB supersector) as a primary measure and secondarily to the one closest in size (Gallemore, Maydew, & Thornock, 2014). Because it is not possible to match every treatment firm to a control firm, an additional treatment firm is dropped leading to a final sample of 95 matched treatment firm-years for testing the first hypothesis (see table 2, panel A).

Hypothesis 2 sample

For the second hypothesis, leverage and total assets need to be non-missing and missing intensity dimensions are adjusted to zero. The maximum of 500 firm-year observations holds for testing this hypothesis as all data is available. From the FTSE 100 firms, 26 have been labelled a corporate tax avoider by the media. The treatment firms and control firms are matched using industry (ICB supersector) primarily and size secondarily (Gallemore, Maydew, & Thornock, 2014). As every treatment firm can be matched, the final sample for testing the second hypothesis consists of 130 firm-year observations, corresponding to 26 unique firms (see table 2, panel B).

Hypothesis 3 sample

For testing hypothesis 3, the minimal data requirements for variants A and B are equal. These are non-missing values for ETR and Cash ETR. For interpretable (Cash) ETRs, positive pre-tax income is required. Leverage and total assets need to be non-missing for control purposes and control variables regarding intensity dimensions which are ultimately missing are adjusted to zero.

Due to negative pre-tax income, 105 firm-year observations are dropped. The dataset which remains is the final sample for testing the third hypothesis, comprising 395 firm-year observations

corresponding to 79 unique firms (see table 2, panel C).

13

3.2 DEPENDENT VARIABLES

Tax avoidance proxies for the first hypothesis

For testing the first and third hypotheses, proxies for tax avoidance are used. The first hypothesis utilizes two proxies, ETR and Cash ETR. As these are annual observations, they capture short-term trends. The ETR is calculated as the tax expense divided by pre-tax income according to the income statement defined by International Financial Reporting Standards (IFRS). A limitation is that ETR does not account for tax deferral strategies. By deferring tax long term or practically indefinitely, these obligations are recorded on the balance sheet but not in the profit and loss statement and thus distort the power of the proxy. Regardless, the proxy is one of the most popular in tax avoidance research (Hanlon & Heitzman, 2010).

To counter this, Cash ETR is the second proxy measuring tax avoidance trends. This is calculated by dividing the cash taxes paid by pre-tax income after eliminating special items according to the profit and loss statement defined by IFRS (Dyreng, Hanlon, & Maydew, 2008). By using the cash taxes paid as a variable, tax deferral effects are filtered out of the tax avoidance proxy. The reason for using both, is that this proxy also has its own disadvantages, as the cash taxes paid are usually a lagging measure, meaning that (part of) the actual payment will take place in the year after they are owed.

The sample which is tested is statistically small, to mitigate the risk of influential extreme observations within the tax avoidance proxies, these variables are winsorized at 0 and 100.

Issuance of the first CSR report for the second hypothesis

The second hypothesis involves a dependent variable coding whether a company issued their first CSR report in a certain year. As one of the largest proponents of CSR reporting, the Global Reporting Initiative (GRI) continually records CSR reports of worldwide companies. The GRI has been recognized by the OECD as an important party in promoting corporate social responsibility and transparency. For this reason OECD and GRI initiated a partnership in 2010 to jointly promote this cause by developing standards for CSR reporting together (OECD, 2010). The GRI database is relied on in coding the dummy variable whether a FTSE 100 company issued a CSR report for the first time in a certain year.

Tax avoidance proxies for the third hypothesis

Both variants of the third hypothesis will include the same proxies for tax avoidance as the first hypothesis, being ETR and Cash ETR.

Regarding the nature of the third hypothesis, additional proxies aim to capture structural measures of tax avoidance. By summing up cash taxes paid as a nominator and pre-tax income (after

eliminating special items) as denominator for five consecutive years and subsequently dividing them, the variable becomes less volatile (Dyreng, Hanlon, & Maydew, 2008). The same five-year measure will be also be used for the ETR measure. The same concerns regarding influential extremes in the sample lead to winsorizing these variables at 0 and 100.

3.3 INDEPENDENT VARIABLES

Public pressure proxy for the first and second hypotheses

This variable is a dummy variable and will be true when a company has been subject to negative

14 publicity in major news publications in the years 2010-2014. LexisNexis Academic is used to scan the major world news publications using the company names in combination with the keywords: tax, avoid or shelter. The search results are examined to ascertain that the article was explicitly criticizing the actual company and that the tax type was corporation tax and not any other tax.

CSR motivation proxies for the third hypothesis (variant 3A and 3B)

The independent variable which is utilized for testing the third hypothesis is derived from the GRI database, like the independent variable of the second hypothesis. CSR motivation through symbolic management is measured through the issuance of a CSR report in a certain year. The variable not necessarily records the first issuance, as opposed to hypothesis two. This dummy variable is coded one when a FTSE 100 company issues a CSR report in a certain year and zero otherwise.

Being a FTSE4Good constituent proxies for CSR motivation through substantive management as a dummy variable. The advantage of this proxy is that by dummy coding FTSE4Good U.K. 50

constituency the FTSE 100 are evenly split in fifty above mean and fifty below mean CSR-motivated companies. It also relates to actual company actions rather than the promises and figures contained in CSR reports (FTSE, 2015). The companies which are identified as a FTSE4Good U.K. constituent are those which are a constituent on 30 June 2015. The rationale behind this date comes from the premise that it takes time for strategic action to be executed and subsequently being noted by the FTSE. By using the most recent constituents in the time of writing, this offers the most recent evaluation of company actions. Because of this, it is most likely that strategic changes have been executed and noted after possible changes of strategy.

3.4 CONTROL VARIABLES

Prior research ascertained associations which need to be controlled for when ETR is a dependent variable (Chen, Chen, Cheng, & Shevlin, 2010). The control variables utilized are capital expenditures, inventory intensity, leverage, profitability (return on assets) and research and development intensity (Hoopes, Mescall, & Pittman, 2012). Furthermore, size (natural log of assets) appears to be

associated with higher chances of reputational costs (Zimmerman, 1983) and opportunities for tax avoidance (Rego, 2003).

The dependent variable of the second hypothesis, issuing a CSR report for the first time, has different predictive variables than those mentioned for the tax avoidance measures. According to Dowling and Pfeffer (1975) companies which are perceived as legitimate will avert incurring financial sanctions from stakeholders condemning illegitimate behaviour. Larger and thus more visible firms generally have more stakeholders and will bear higher reputational costs when targeted by public scrutiny. To mitigate reputational costs, larger firms will be more prone to issue a CSR report to mitigate these costs and thus, size (natural log of assets) is controlled for. Juholin (2004) supports the notion of size as predictor of CSR motivation and furthermore asserts that more profitable firms are more inclined

5 Two companies have been labelled as being scrutinized although they were not scrutinized within the designated timeframe. The reason for this is that both companies were scrutinized just before this timeframe starts and thus the effects of coverage should still apply within this timeframe. Untabulated sensitivity analyses conclude that the results for hypothesis one only change in the way that the p-value from the effect of public pressure on Cash ETR goes from significant to weakly significant. Hypothesis two demonstrates no change in results at all.

15

towards being more CSR motivated. In summary, size and profitability (return on assets) are the control variables for the second hypothesis. For the full description of the control variables table 3 can be consulted. All continuous control variables are winsorized at the 1

and 99

percentiles to mitigate the risk that extremes influence the test excessively.

3.5 MULTIVARIATE MODELS

For testing the first hypothesis, an adaptation of the empirical model from Gallemore et al. (2014) is used. The model used is of difference-in-differences and ordinary least squares (OLS) methodology for the dependent variables. This results in the following model:

𝑇𝐴𝑋 𝐴𝑉𝑂𝐼𝐷𝐴𝑁𝐶𝐸

= 𝛽

+ 𝛽

𝑃𝑈𝐵𝐿𝐼𝐶𝐼𝑇𝑌 𝐹𝐼𝑅𝑀

+ 𝛽

𝑃𝑈𝐵𝐿𝐶𝐼𝑇𝑌 𝑌𝐸𝐴𝑅

+ 𝛽

𝑃𝑈𝐵𝐿𝐼𝐶𝐼𝑇𝑌 𝐹𝐼𝑅𝑀

∗ 𝑃𝑈𝐵𝐿𝐼𝐶𝐼𝑇𝑌 𝑌𝐸𝐴𝑅

+ ∑ 𝐶𝑂𝑁𝑇𝑅𝑂𝐿𝑆

𝑘

+ 𝜀

(𝟏) The two dependent variables are ETR and Cash ETR, as described in the paragraph 3.2. The

independent variable PUBLICITY FIRM is a dummy variable set to one if the firm has been named in major news media in conjunction with corporate tax avoidance. Firms which do not constitute this group are set to zero and function as a control group. The dummy variable PUBLICITY YEAR is set to one for both treatment and control firms in the year that the treatment firm was covered. The focus for the test results is on the interaction in β

as it attempts to capture the effects of negative media coverage on tax avoidance.

For the second hypothesis, an adaptation of the Dyreng, Hoopes and Wilde (2016)model is used:

𝐹𝐼𝑅𝑆𝑇 𝐶𝑆𝑅

= 𝛽

+ 𝛽

𝑃𝑈𝐵𝐿𝐼𝐶𝐼𝑇𝑌 𝐹𝐼𝑅𝑀

+ 𝛽

𝑃𝑈𝐵𝐿𝐼𝐶𝐼𝑇𝑌 𝑌𝐸𝐴𝑅

+ 𝛽

𝑃𝑈𝐵𝐿𝐼𝐶𝐼𝑇𝑌 𝐹𝐼𝑅𝑀

∗ 𝑃𝑈𝐵𝐿𝐼𝐶𝐼𝑇𝑌 𝑌𝐸𝐴𝑅

+ ∑ 𝐶𝑂𝑁𝑇𝑅𝑂𝐿𝑆

𝑘

+ 𝜀

(𝟐)

16 The dependent variable is FIRST CSR, which captures the whether a company issued their first CSR report (or equivalent) in a given year. The approach is to establish whether publicly criticized firms are more prone to start issuing CSR reports, measured in the β

coefficient. The test approach follows difference-in-differences methodology in the logistic form for the dependent dummy variable.

For the variations of the third hypothesis, the following models are used:

𝑇𝐴𝑋 𝐴𝑉𝑂𝐼𝐷𝐴𝑁𝐶𝐸

= 𝛽

+ 𝛽

𝐶𝑆𝑅 𝑅𝐸𝑃𝑂𝑅𝑇𝐸𝑅

+ ∑ 𝐶𝑂𝑁𝑇𝑅𝑂𝐿𝑆

𝑘

+ 𝜀

(𝟑) 𝑇𝐴𝑋 𝐴𝑉𝑂𝐼𝐷𝐴𝑁𝐶𝐸

= 𝛽

+ 𝛽

𝐹𝑇𝑆𝐸4𝐺𝑈𝐾

+ ∑ 𝐶𝑂𝑁𝑇𝑅𝑂𝐿𝑆

𝑘

+ 𝜀

(𝟒)

The dependent variable TAX AVOIDANCE definition is equal to the first model but with addition of

the five year ETR and Cash ETR variables. The independent variable CSR REPORTER is a dummy

variable coding whether the company issued a CSR report in a given year. The independent variable

FTSE4GUK relates to a company being a FTSE4Good U.K. constituent, which is a dummy variable

equal to one if the company is a constituent on 30 June 2015. The approach is of multiple regression,

following OLS methodology.

17

IV. RESULTS 4.1 DESCRIPTIVE STATISTICS

Hypothesis 1

Table 4 presents the descriptive statistics for the sample of the first hypothesis. The mean effective tax rate is 23,36% for all sample firms. It is important to note that the means of the treatment firms for both tax measures increase (ETR from 23,18% to 24,63% and Cash ETR from 19,82% to 22,27%) when comparing the periods before scrutiny and after scrutiny. This contrasts the firms in the control group which show a slight decrease in effective tax rates. The treatment firms appear to be higher leveraged than control firms and appear to be less inventory intensive.

Measures of multicollinearity between variables are presented in table 5. Some independent

variables appear to be significantly correlated. The highest significant coefficient however, is below

0,500 correlated (Pearson correlations: Leverage and Size). Although some multicollinearity is

observed, it is deemed that it does not corrupt the regression model, judging by the coefficients.

18

19

Hypothesis 2

Table 6 shows the descriptive statistics of the sample used for testing hypothesis 2. The treatment group has approximately the same sample mean for First CSR as the control group. This indicates that if hypothesis 2 is to be accepted, a strong relationship between scrutiny year and first CSR year should exist. Examining the First CSR year means, most treatment firms appear to have issued their first CSR report in 2012 (20%). Keeping in mind that most public scrutiny occurred in 2013 (with no new issuers in the treatment group), this bodes poorly a priori for affirming the second hypothesis.

Furthermore, treatment firms tend to be less inventory intensive, and higher leveraged than control

firms. Size appears to be quite close, but this is in part artificially caused because the firms were

secondarily matched by size.

20

The correlation matrix in table 7 shows that some variables are significantly correlated. The highest

coefficient (Pearson correlations: Leverage and Size) does not exceed 0,500. Thus it is deemed that

multicollinearity is indeed observed, but not enough to invalidate the logistic model results.

21

Hypothesis 3 (variant H3A and H3B)

In table 8 the descriptive statistics for the sample used for testing hypothesis 3 are presented. It shows that the sample means of tax expenses (ETR 24,36% and ETR 5Y 30,60%) are higher than the actual tax payments (Cash ETR 20,18% and Cash ETR 5Y 20,93%), which suggests that in general there is a time delay between tax expensing and the moment of payment caused by long-term or indefinite tax deferral strategies.

The mean of FTSE 4 Good U.K. constituents is not evenly split due to dropped observations which did not meet the minimal data requirements, which shifts the symmetric distribution slightly. Also, ETR 5Y data is less available than the other variables which negatively affects the statistical power this sample can produce.

In table 9 the correlation matrix is presented. It indicates that certain independent variables are

significantly correlated. The highest significant coefficients do not exceed 0,500 correlation however

(both types: size and profitability), which mitigates multicollinearity concerns. Thus it is judged that

the regression models are not impaired by multicollinearity.

22

23

4.2 MULTIVARIATE RESULTS

Hypothesis 1

The main results for testing the first hypothesis are presented in table 10. The positive coefficients

observed in the interaction between Publ. firm x Publ. year and (Cash) ETR supports the expectations

from hypothesis 1, suggesting public scrutiny increases a company’s tax burden. While both being in

the hypothesized direction, the p-value of ETR is completely insignificant (p-value > 0,10). The

variable Cash ETR is significant (p-value < 0,05). The table shows a mean coefficient increase of Cash

ETR by 7,171, which translates into a more than two billion increase in tax cash flow. The variable

ETR appears to increase with 2,315 following public scrutiny. However, given the standard deviation

of 5,065 for this coefficient, this measure is too volatile to be considered significant. For these

reasons, the first hypothesis can only be accepted with regard to Cash ETR.

24

25

Hypothesis 2

The prediction of the second hypothesis is that scrutinized companies will react by issuing a CSR report to repair their perceived reputation as a bad corporate citizen. Being scrutinized in a certain year should lead to the issuance of a CSR report for the first time for a company. The results of the logistic regression model are found in table 11.

The variable of interest, the interaction term Publ. firm x Publ. year, points in the negative direction, which contradicts the hypothesis. Because of a very large standard deviation of 8.096,474,

significance is almost perfectly absent (p-value > 0,10). This sample gives absolutely no evidence that the second hypothesis can be accepted. These results actually support the null hypothesis (p-value <

0,01), implying that issuing a CSR report for the first time is not related to negative media coverage of

a company’s tax strategy. This suggests that CSR is not tied to tax strategy and that CSR reports are

not used to polish a company’s image post-event. An alternative explanation can be found when

reasoning according to legitimacy theory from Ashforth and Gibbs (1990). Companies which are

already perceived as being less legitimate will be met with scepticism when they suddenly start

propagating responsible conduct. This could explain that company management will not bother

incurring the expenses associated with starting a CSR project.

26

The third hypothesis predicts that CSR-motivated companies practice different degrees of tax avoidance than other companies.

Testing hypothesis 3A, the test results listed in table 12 provide inference that companies which issue CSR reports actually avoid more tax than companies which do not. Three of the four measures of tax avoidance suggest that companies which express themselves through CSR reports as being good corporate citizens actually pay less corporation tax than other multinational companies. The findings are significant for ETR (p-value < 0,05), significant for Cash ETR (p-value < 0,05), weakly significant for ETR 5Y (p-value < 0,10), and insignificant for Cash ETR 5Y (p-value > 0,10).

The results consistently show a negative association between issuing CSR reports and effective tax

rates which leads to accepting hypothesis 3A in the negative direction. The weaker significance in ETR

27

5Y and insignificance of Cash ETR 5Y might be explained because some values for these variables are missing, weakening the statistical power which can be obtained from the sample. Accepting

hypothesis 3A in the negative direction leads to the notion that tax is not treated as a CSR issue, as

companies which issue CSR reports bear a lower tax burden than companies who do not. Another

explanation from previous literature which might be applicable is that a CSR report merely functions

as a legitimization instrument, to attain “insurance” from public scrutiny (Godfrey, Merrill, & Hansen,

2009). It could be suggested that companies engage in symbolic management with the goal of

making themselves less susceptible to public criticism (Ashforth & Gibbs, 1990). If the external

stakeholders treat tax as a CSR issue, the gap between words (CSR report) and actions (observed

effective tax rates) is hypocritically wide in stakeholder perception.

28

Table 13 lists the regression results for testing hypothesis 3B. The coefficients all point in the negative direction, which is consistent with the direction of variant A of the CSR effect on tax avoidance. The p-values however, are completely insignificant for ETR (p-value > 0,10), ETR 5Y (p- value > 0,10), Cash ETR (p-value > 0,10) and Cash ETR 5Y (p-value > 0,10). Concluding, the test results of regression model (4) provide no basis for accepting hypothesis 3B.

Contrasting hypothesis 3A, the test results do not imply that CSR actions and paying more or less taxes are intertwined. Thus, it seems again that tax is not treated as a CSR issue by companies.

Noteworthy is also the contrast between results for both variants of the third hypothesis, companies use symbolic management to try to attain “insurance” from scrutiny, while the actual actions

(substantive management) do not necessarily follow.

29

V. DISCUSSION AND CONCLUSION 5.1 CONCLUSION

This thesis demonstrates that the robust findings by Dyreng et al. (2016) have external validity within the British FTSE 100. Linking tax shaming in major news publications to firms’ tax avoidance

behaviour following this scrutiny event produces the hypothesized results to some extent. While the results point in the direction that public scrutiny decreases tax avoidance, the results only provide enough significance to accept the hypothesis with regard to cash effective tax rates. The

insignificance observed concerning regular effective tax rates might be caused by the statistical power the relatively small sample can produce. Thus the “undersheltering puzzle” referred to by Gallemore et al. (2014), is still not convincingly solved. Individual characteristics and the magnitude of the perceived social offence by a company can determine the intensity of public scrutiny.

According to Pruitt and Friedman (1986), attempts at shaming and boycotting an organization are not homogenous occurrences. The ActionAid event was a carefully coordinated effort which

simultaneously confronted multiple companies with the same offence. This similarity within scrutiny lends to a better event observation than the distinct tax shaming scandals used within this study. This variation in scrutiny intensity and individual characteristics can explain the unobserved significance regarding firms’ effective tax rates.

Another notion studied in this thesis is whether companies react to public scrutiny by issuing CSR reports to repair their negative public image. Test results provide no support for this hypothesis, which can be explained by the relatively small sample size and the chosen scrutiny event as stated in the preceding section. If reputational costs from public scrutiny generally do not appear to make a tax avoidance trade-off result in a net negative effect. Which could be inferred from the insignificant effects of public scrutiny on effective tax rates, this would also dissuade firms from engaging in reputational repair strategies. Because the multinational companies still reap the benefits of a net positive effect, bearing the administrative costs of CSR reports appears less tempting. Additionally, according to Ashforth and Gibbs (1990), issuing a CSR report after public scrutiny also has a reduced effect because the contrast between actual reputation and portrayal in the CSR report will be met with scepticism. Thus it would seem rash to waste company resources on reputational improvement strategies when the effect would be at its weakest.

The most promising results from this study come from the relationship between CSR and tax avoidance. Companies which issue CSR reports appear to bear a significantly lower tax burden than firms who do not issue CSR reports. Another CSR proxy, being a FTSE4Good U.K. 50 constituent, does not relate to significant differences between constituents and non-constituents in terms of tax avoidance. The implications of these test results will be interpreted in conjunction with the answer to the original research question from this study.

The main research question inquires whether reputational concerns influence multinational firms’

tax avoidance behaviour. Taking together all results it can be suggested that reputational concerns are important in determining tax behaviour as public scrutiny can pressurize firms to pay more tax.

Firms appear to engage in symbolic management to protect themselves from instances of public