Relationship between block holders and aggressive tax planning

(1)

1

Relationship between block holders and aggressive

tax planning

Author: Olaf Berendsen

Master 2014/2015: Accountancy and Control (track Accountancy) Student number: 10003675

First Supervisor: Dr. W.H.P. Janssen Submission date: 22th of June 2015 Number of words: 12.772

(2)

2

Abstract

Purpose

This thesis aims to gain knowledge about the determinants regarding tax avoidance. The purpose of this research is to examine the impact of major shareholders with respect to the agency theory in relation of tax avoidance.

Design/methodology/approach

This research is conducted with data of the database Bureau van Dijk / Osiris. For this research 367 US listed banks are examined.

In order to measure the association between block holders and tax avoidance I created a formula/model to give each of the five largest shareholder a weight and compute the block holders spread for each bank. With this measure block holders spread it is possible to compare different levels of large shareholders in banks. For the level of tax avoidance I use the GAAP ETR (Total income tax expense divided by total pre-tax accounting income).

Findings

Based on the population of 367 banks I found with the measure block holders spread and GAAP ETR that there is a positive association between block holders spread and GAAP ETR. The results of this thesis is supports the hypothesis that banks with a low block holders spread have higher levels of tax avoidance. This implicates that major shareholders in a firm make a difference in respect to tax avoidance. In addition to the block holders spread I found that leverage is negative associated with tax avoidance. This indicates that for the examined population lower levels ETR GAAP are associated with higher levels of leverage.

Research limitations

The data used for this research exist of only US banks and listed. The research is conducted for the year 2013, this could have limitations to extrapolate the results. Level of tax avoidance is measured with GAAP ETR and the level of major shareholder influence is measured with block holders spread, if both variables should be measured in another way, the outcome could differ.

Originality/Value

To my knowledge there is no other research that examines directly the effect of major

shareholders with respect to tax avoidance. This research complement a broad academic trend of examining the effect on ownership structure and policy in a firm.

Key words

(3)

3

This document is written by student Olaf Berendsen 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.

(4)

4

I. Introduction

This paper examines the factor of major shareholders and their influence on the tax policy of a firm. In particular, this is done by the measure block holders spread, with this measure it is possible to compare the different levels of presence of major shareholders in a firm.

Motivation to engage into aggressive tax planning is the pressure of showing high profits. The most mentioned claim of the society and the media is the excessive rewards for CEO’s and especially the compensation plans with bonuses. These bonuses and pressure should encourage forms of tax avoidanceand would financially harm the society.

Although tax avoidance could be profitable for the firm on short term, there are serious risks involved with aggressive tax planning. The general risks associated with aggressive tax planning could be reputational risks, penalties/fines of governments, too opportunistic view on fiscal claims at current moment and extra negative attention of governments on the company.

There is a growing interest among investors for risks associated with tax avoidance. In November 2015 the Financial Times published an article were investors explained that in their investment policies, tax matters become a more important factor. Some of the investors claim that only companies with a ‘normal’ tax policy should be eligible. This implicates that investors have serious doubts about the long term value of boosting profits with aggressive tax planning. The literature is contradicting about the long term effects of tax avoidance. E.g. Gallemore et al (2014) examined the reputational cost, and they conclude that there is only a temporary (negative) effect. However Kim et al (2011) investigated tax avoidance and stock-price crash risk, their results are that tax avoidance policies are positively associated with future crash risk of firm-specific returns.

Hanlon & Slemrod (2009) examined the effect of the stock price when news come out with respect to engagement of a firm with tax avoidance. Their results show that stock prices declines when firms are publicly linked in the news to tax avoidance, especially firms in the retail sector. Hanlon and Slemrod suggest that investors are afraid of the possible negative reaction of consumers and the influence of this negative attention for the profitability on the long term. Penalties of enforcement is in a lot of cases a fraction of the real cost of reputation risks. Karpoff et al. (2009) examined the effect of legal penalties in relation with reputation costs. His results showed that the market penalty (lower present value due to higher finance and contract

(5)

5 costs) is more than 7,5 times the cost of the penalty imposed of the government or enforcement body.

Through monitoring the managers who make the decisions on behalf of the shareholders could be restrained to avoid too risky decisions. “Monitoring is necessary to induce managers to work hard; only concentrated share ownership induces shareholders to monitor.” Quote of Huddart (1993). This quote suits the thought that concentrated ownership have more ability to affect the managers. Weisbach (2001) wrote a paper with ten propositions about tax shelters and tax avoidance. His aim with his propositions is to contribute to the discussion of tax avoidance in a broader way then only legalistic thinking of governments and sceptic people towards tax avoidance. His main question is why is there not more tax avoidance? This thesis is an attempt to form a part of an answer. In my opinion the owner structure of a firm could play an role.

For the research in this thesis I use 367 US banks listed at the NYSE or NASDAQ. I use the five largest shareholders and the remaining part of the shareholders to compute the measure of block holders spread. For the level of tax avoidance I use the measure of GAAP ETR.

This thesis links the agency theory to tax avoiding for the banking industry. With the active monitoring role of a major shareholder he has the ability to use this influence for the tax policy of a firm. We hypothesize that banks with a low block holders spread (relatively more small shareholders) have a higher level of tax avoidance. This because influence and possibility to monitor will decrease when there are a lot of different shareholders, this will lead to more discretion for the managers of the firm. I found evidence based on a population of 367 banks that banks with a low block holders spread are associated with higher levels of tax avoidance. In addition I found that leverage is negative associated with the level of tax avoidance.

This thesis contributes the academia and governments to gain knowledge about the factors that

aredeterminative for aggressive tax planning of firms. Second this research helps to understand

the influence that (major) shareholders have on the policy of firm based on the agency theory. This thesis contributes the research of tax avoidance because we measured ownership structure in a new manner.

This thesis proceeds as follows. I elaborate the theory in section II and explain why the agency is important in relation to tax policy and discuss earlier research with regards to risks and tax avoidance. The research design and description of the examined sample will be in Section III. Section IV contains the results of the empirical tests. The last section V concludes the research in the form of a summary.

(6)

6

II.

Literature

The crisis of 2008 and the aftermath changed the (financial) world in several ways. One of them, is the way how the society thinks about tax avoidance by large companies. Since governments face difficulties with their government budget the opinion about tax avoidance became a tougher and firmer debate. Several elements have contributed to a broader public debate for example several whistleblowers (luxleaks), the attention of governments for fiscal evasion and extra media attention.

Tax paid by companies is a large part of the income of the (local) government. Governments are dependent of these kind of income to provide the society with public services and goods. Recent years the criticism of governments and society increased towards firms that engage in aggressive tax planning. E.g. European Commission conducts research into various countries (Luxembourg, Ireland and the Netherlands) with regards to the tax policy and tax deals with multinationals. Tax avoidance is a sensitive topic in the business world and in the politics, however in the academic world there is not so much known about tax avoidance related to the ownership structure. A lot of researchers name the topic of tax avoidance in relation to possible determinants but real research is limited and it is a topic that became recently a more researched topic. Because of this limited research Desai and Dharmapala (2009) call for further research with tax related to the agency theory. The limited research could also be attributed to the difficulties with appropriate measuring tax avoidance.

Tax avoidance

Every company have to pay tax about their pre-tax income. Governments enacted many laws to which all companies must comply. Only there are a lot differences between tax-laws in various countries and it is not always clear how to interpret the laws in a specific context. This “grey” area is used through some companies to reduce tax payments or to pay less as possible taxes. This implies not necessarily that this kind of actions are illegal, it could be legal but it is not always sure to in which extent. This thesis is not aimed to provide a clear definition of tax avoidance or to explain all the manners companies attempt to avoid tax. I will use the definition of Hanlon and Heitzman (2010) to interpret tax avoidance “tax avoidance is the reduction of explicit tax payments”. In this thesis I will define tax avoidance as a plain concept wherein companies try to pay less as possible tax for whatever reason.

(7)

7 Tax avoidance could be possible in several ways, if a company operates in different countries there is the simple ability to use the most beneficial tax regime and shift the profits towards a specific country with a low tax rate. Shifting of income is usually performed by rent payments (give a loan from a tax friendly regime subsidiary to another subsidiary in another country) and the profit shifts from one country to the other. Another much-tested method is the movement of intangible assets like brand names, recipes or patents to a country with low tax rate and charge subsidiaries with invoices in other countries for using these intangibles assets. Especially banks have a lot of loans and in general have a high leverage (debt to equity ratio) and therefore these banks have the ability to engage in tax avoidance. For services that banks provide it is easily to operate global and use the advantages of operating global. Therefore the banking industry is interesting to examine for tax research.

Despite that there is evidence for the risks regarding tax avoidance, still research shows that tax avoidance is general practice for large firms. Tax avoidance is seen by managers as a way to boost the profits and dividends of a firm. In the U.S. the prescribed tax rate is about the 40%. It could differ because each state have their own autonomy for a part of the tax rate. Therefore in general a firm should pay about 40% tax over their income. The perception of the media and the society is that a lot of large companies pay much less than the prescribed 40%. The academic research underpins that claim. Dyreng et al. (2008) examined 2.400 US firms during the years 1995-2004 and their effective tax rates. The results of Dyreng et al. show that 26% of the firms are able to get their ETR (Effective Tax Rate) lower than 20% for all these years. If I interpret this result I conclude that some firms systemically avoid tax payments. If I look in depth to the results, it is also true that a lot of the companies have a higher ETR than 30%. This could be due the policy of a firm or the ability that a firm has to engage in tax avoidance. In the context of the agency theory it could be that some managers choose explicitly to stay away from tax avoidance because of the risks or choose thoughtful their level of tax avoidance dependent on the kind of business, ability and the pressure from shareholders and other stakeholders.

One fact of the Dyreng research in combination of other research is that the tax policy expressed in ETR varies a lot over the companies and industries. Therefore research towards determinants of aggressive tax policies is important. Furthermore it is important to note for this thesis that in the Dyreng et al. (2008) research the mean firm still have the tax rate of 30% which is far below the prescribed tax rate of 40%. In the essence there is a gap between the rate that governments mark as reasonable and the tax rate that companies actually report.

(8)

8 Bryan Cloyd et al. (1996), show that managers are an important factor of decisions regarding the tax policy. They found that managers choose for accounting methods which are in line with their aggressive tax positions. The reason for this kind of decisions has two main elements, in a discussion with the tax authority, firms could substantiate their (aggressive) tax position with the accounting methods. And more importantly managers align accounting methods with the aggressive tax policy to decline the potential critic and cover the risks associated with the gap between the earned income and the paid taxes. This conclusion implicates that managers feel that they are doing something that is not fully accepted or desired. Because otherwise why should you cover your tax practices with accounting methods? And it is remarkable that it seems that managers want to hide their tax policy and their minimal disclosure about this topic. That behavior increases the information gap between the managers of the firm and the investors. An growing information gap could lead to risks for the investors (shareholders).

D. Phillips (2003) examined the impact of after-tax accounting-based performance measures for business unit-managers. His results show that such after tax- accounting based performance measures helps the company to decrease the effective tax rate. Thus if a company want to decrease their effective tax rate, they could use after-tax accounting bases performance measures as an effective instrument. This contributes to the thought that the managers have the ability to steer the effective tax rate to a level that is optimal for them, namely to create personal wealth. This findings suits the taught that if shareholders have large influence through establish the compensation plans, shareholders could use this power to manage the level of tax rate towards an “optimal or save” percentage. This “optimal or save” tax rate should be seen as, a level whereby the firm pays less as possible tax, but also mitigate risks associated with tax avoidance. Tax avoidance is of all times, nowadays there is a debate going on about the desirability of aggressive tax planning of the large firms. Firms avoid taxpaying to increase their profits. Including federal, state and the local taxes the tax rate could be more than one-third share of the (pre-tax) profits of a company. Because of this high percentage managers have an incentive to avoid these taxes and instead of contribute to the society, they prefer to pay dividend to their shareholders and get their bonuses based on their ‘profit’ performance. A lot of research focus on the power of the shareholders and their influence on corporate management. Research of the power of large shareholders regarding taxes isn’t extensively examined.

(9)

9 Based on earlier research I am very curious about the following; use major shareholders their influence to manage the risk of extreme aggressive tax planning? I wonder if managers of companies with a lot of different shareholders have more freedom and as result of their autonomy they have more the ability to increase their bonuses with aggressive tax planning. It not unreasonable to think that an opportunistic behavior is more present without a strong monitoring role of a large shareholder. The risks regarding tax avoidance is mostly related to reputation, especially in the banking sector reputation is very important. Banks have the characteristic that they are high leveraged because of their social function (lending money to consumers and companies in order to support the economy). Solvency ratio’s for banks are very low compared to other industries. This implicates that despite the low ratio people and companies trust their money to them. Reputation is therefore a large key factor of this system. The reason for this thesis is based on literature regarding to the impact of large shareholders, determinants of ownership structure in management decisions, myopic behavior of managers to increase their bonuses and the (crash) risk on the stock price of aggressive tax planning.

Monitoring and stewardship role of shareholders

To understand the literature of financial accounting it is important to understand the agency theory. Jensen and Meckling (1976) described the agency relationship as the contract between the principal (owner) and the agent (manager). Both want to maximize their wealth, the principal wants a high profit and the agent want a high salary (for less as possible effort). Both interests could harm each other. To solve this problem, they could make an agreement to align their interest. A lot of companies use bonuses based on financial performances, if the agent perform well with the company he earns a high bonus and the principal have a higher profit. If there are large shareholders they could use their power to influence the management. Large shareholders have a large influence on the management because they have the ability to sell their shares in the company and in the end, shareholders are the owners of the company and have voting rights in the shareholders meeting. Salaries for managers are in most cases directly or indirectly linked to the share price are inclined to listen carefully to their block holders (Edmans, 2009). Jensen and Warner (1988) conclude in their paper that the ownership structure definitely affect the corporate behavior in important ways. They mention the importance of; shares held by managers, institutional investors or other investors.

(10)

10 People are driven by incentives, because also managers are humans, their behavior is also for a major part explainable by incentives. In the economic and business world parties use bonuses to get the desired results. Compensation plans usually contain one or more rewards for specific results to encourage managers to meet specific targets. Research underpin the importance of incentives and compensation plans with bonuses. If it is the fact that incentives work, it is definitely not said that the outcome is the preferred outcome.

Based on the theory of Healy (1978) there is support for the thought that managers with income-reporting incentives in their bonus contracts are more likely to use accruals in order to meet their targets. Accruals could lead to too opportunistic reporting about profits, due to the incentive to report a profit that meets the targets in an attempt to cash the bonus (maximizing the wealth of the agent). In principle, this could be in the interest of the principal and the agent, however accrualsbearing the risk of profits that reverse over time.

If I extract the accruals research to the research with regards to tax avoidance there are a lot of similarities, there is discretion for the managers (interpretation of tax rules, accounting methods), enforcement is difficult for shareholders (the government isn’t able to audit all the tax statement in detail), managers use their information advantages, less tax payments increase the profit of the firm and also too opportunistic assessment of tax payments could reverse over time.

Investment decisions on behalf of the firm are part of managing a firm. Bushee (1998) showed that managers are less likely to cut R&D investments when institutional ownership is high. The results could be interpret in the way that institutional investors have an effective monitoring role for the managers and act more as an owner than a trader. However he divided institutional investors and individual investors and showed that the incentives and behavior of managers are different. The Bushee study is relevant for this research in the sense that ownership structure could be an important factor in decisions of managers and the shareholders influence these decisions towards their interests. The Bushee’s results fits in the thought that large shareholders have a better ability and incentive for an active monitoring role.

Another study that is interesting in the context of incentives and the agency theory is the research of Clifford (2008) in the field of hedge funds. Hedge funds are nowadays a normal phenomenon in the market. Hedge funds attempt to gain a significant influence on the managers and try with this influence to achieve more efficiency in the firm and create a more profitable firm. Mostly this hedge fundsare so-called “activist shareholders”. They use active their voting rights and influence to steer the firm. Clifford found that hedge funds earn larger investment

(11)

11 returns for their ‘active’ blocks than their ‘passive’ blocks. For his dataset, this could mean that active monitoring and stewardship role is better rewarded than a passive role of the hedge fund. Clifford mention also in his research the free-rider problem for a block holder, if the block holder / major shareholder puts a lots of effort to improve the firm, other shareholders gets the benefits too. The free-rider problem declines when the shareholder has a larger part of the total shares, in the way the largest part of the future profits is for also for the large shareholder. This implicates that the incentive to put effort in the firm increase, when the stake in the company is larger for the block holder. Consequently the ownership structure (proportion of large versus small shareholders) could affect the level of activism of shareholders, and therefore could decline the decision freedom of the managers.

Academic literature note the link between the characteristics and the capital providers. The research of Brailsford et al. (2002) demonstrate the link between block holders (ownership structure) and leverage (capital structure). They build their research on the monitoring and stewardship role of shareholders in accordance with the previously mentioned agency theory. Brailsford et al. found that if a (small) part of the shares are held by the management the leverage ratio is low due to the convergence of the interest namely investor (principle) and manager (agent) are the same. But if the managers held more than 50% of the shares also managers become too opportunistic what results in a higher debt ratio. In my opinion this research implicates that if managers have a too high stake in the company they transform towards a ‘normal’ shareholder and will act like a normal shareholder.

Furthermore Brailsford examined also to the relationship of block holders with leverage, leverage could be a risk for the risk with regards to possibility of a future bankruptcy. Leverage (capital structure) is relevant because block holders have an incentive to monitor (there is a lot in stake) and they have the power (shares give voting rights) to steer the management to the interest of the shareholders. The outcome of their research is that the level of external block holders is positively associated with leverage.

The results of Brailsford et al. could be interpret in the way that shareholders, if they have an incentive (a lot of shares that could grow in value or decrease in value) and have the power/influence through the voting rights to influence the board policy on several issues, especially something important like tax avoidance. They could use their power to prevent the firm for a risky tax policy.

(12)

12

Link between tax avoidance and the monitoring role

There are some writers in the academic literature who did research to the link of the manager and the level of tax avoidance. There could be a link because of the decision authority of the manager regarding to the accounting matters, external reporting and the information advantage that a manager has. Managers who govern the firm have a large influence of the culture in the firm and that also could reflect in the outcome like external reporting and tax policy. Chyz (2010) examined the effect of an aggressive tax manager to the tax rate of a firm. His definition of an aggressive tax manager is a manager who in personal matters act pro-active in avoiding tax. Chyz examined if this kind of manager is associated with the level of tax avoidance of the firm. The results of Chyz lead to the conclusion that the presence of an aggressive managers is associated with tax avoidance in the company. This supports the thought that managers do have an impact on the level of tax avoidance. Additional he found that the presence of aggressive managers in a company is positively associated with the firm value. This result could be interpret as contrary to Hanlon et al. (2009) research because they conclude that in general stock prices decrease if companies are in the news with regards to tax avoidance. It is contrary in the way that the results of Chyz seems to lead to the conclusion that it could be profitable to engage in tax avoidance and the results of Hanlon seems to suggest that it is a risk to engage in tax avoidance.

In addition to the Chyz research there is a research of Dyreng et al. (2010) that researched the influence of the CEO and the level of tax avoidance. This research focus on the ‘tone at the top’. The culture in the boardroom (tone at the top) could play an significant role on the behavior in the whole organization and also e.g. tax policy. For his research Dyreng followed more than 590 executives of companies during the period 1992 - 2006 and examined the effect of their characteristics on the level of tax avoidance. The results document that the person of the CEO is an important factor in the level of tax avoidance. And a characteristic of a CEO that is associated with tax avoidance is an MBA degree or law degree.

McGuire et al. (2014) examined the effect of dual class ownership and the effect on tax avoidance. Dual class ownership firms have two different kind of shares, namely normal shares with the right to vote once and other ‘superior’ class shares which are held mostly of the time by firms’ officers and directors and give them more voting power than the normal share, which implicates they have more power and mostly control the majority of the firm. McGuire et al. hypothesize that firms with dual class ownership have a lower level of tax avoidance, because of the benefits (greater tax savings and potential higher dividends) are for the largest part for the

(13)

13 normal shareholders, and the superior shareholders have less dividend payments. Superior shareholders are mostly employed by the firm and want less as possible scrutiny and intervention of regulators or governments. Looking at the results of the research they found evidence that supports this hypothesis. This could implicate that there are indeed high risks involved because otherwise the firm’s officer should also choose for the tax avoidance option to maximize their wealth. In addition McGuire et al. examined the level of tax avoidance of dual class firms compared to the publicly traded companies. Their results show that the tax avoidance decrease when the level of dual class (difference between the normal shares and superior shares) increase. This result underpins that the shareholder structure have influence of the level of tax avoidance. The research of McGuire is very important for the hypothesis of this research because, he shows with his hypothesis and outcome that there are risks involved with tax avoidance. Because a manager avoid tax when he has no stake in the company (objective mind) and become more profit minded when he has more stake in the company (subjective mind).

For this thesis there must some support for the idea that there is a clear link between ownership structure and tax avoidance. Cheng et al. (2012) examined the effect of hedge funds and the level of tax avoidance in companies. They found that companies after a hedge fund intervention increased their level of tax avoidance. This finding indicates that there is an effect after the intervention of a hedge fund, this could be due several reasons such as bad tax planning of the firm before the intervention, extra leverage after the intervention etc. Whatever the reason is, due changes of the ownership structure the level of tax avoidance differs significantly for the studied group. Therefore the research of Cheng contributes to the idea of a link between tax policies and ownership structure.

The study of Chen et al. (2012) examined the effect of family held firms and tax avoidance compared to non-family held firms measured by the Effective tax rate (ETR) and Book-Tax Difference. Their results show that family firms are less involved with tax avoidance than non-family held firms. Chen et al. suggest that there is a link between the long term horizon (value creation for long term) that most of the family held firms have. Tax avoidance could be associated with future reputation risk and furthermore the reputation of the family is strong linked to the reputation of the firm. This could be an extra factor that deter especially family held firms to exhibit a tax aggressive policy. The results of the Chen et al. research is contrary to the idea of the society, most people think that especially family firms engage in tax avoidance. However the results of this research strengthen the idea that due risks in long term, some kind of shareholders avoid low tax levels.

(14)

14 To my knowledge there is no evidence of increased firm value by aggressive tax planning in general. Although for well-governed firms there is some evidence for increased firm value (Desai & Dharmapala, 2009). However Desai and Dharmapala suggest that this is due the governance of a firm, firms which are marked as well-governed, benefits from tax avoidance. This because investors are reservedly with not well-governed firms, because of the potential information gap, possibility of too opportunistic behavior of the management and potential risk reputation. To draw a careful conclusion for the Desai and Dharmapala research, a firm does not increase in firm value in general, except well-governed firms, however the risks could grow if a firm apply an aggressive tax policy due to reputation risks and the information gap. For well-governed firms investors believe that the risks aren’t very large and tax avoidance will be in the benefit of the firm.

In addition of the findings of Desai and Dharmapala, Kim et al. (2011) examined the future crash risk related to tax avoidance. Kim et al. examined a large sample of US firms during a period of 13 years and studied if companies with larger book-tax differences and long run low tax rates are more likely to experience future stock price crash. Their results show that indeed firms that have an aggressive tax policy, which is reflected in a large book-tax difference or low tax rate are more likely to face future stock price crash. This outcome is in accordance with the theory that tax avoidance is associated with risks and that an aggressive tax policy isn’t always is in favor of the shareholders.

This result of Kim et al. should have implications in the way that (large) shareholders should (active) monitor the managers for the tax policy in their own interest. This monitoring may entail that block holders should achieve a ‘safe or optimal’ tax rate level that reduces the risks. Because if managers could make the decisions with regards to the tax policy it could harm the stock price enormously. In accordance with Desai and Dharmapala (2009) Kim et al. showed that the future crash risks declines when the monitoring mechanisms around a firm are better constructed.

(15)

15

Long term risk does it matter?

In the media, society and also in the academic world there are a lot of people that have the opinion that shareholders only care about short-term profits and therefore shouldn’t care about the (longer term) risk regarding tax avoidance. That could be the truth in some cases, that I won’t deny, however if an investment face less risks an (potential) investor would pay more and the risk premium is lower. This is the reason why I argue that long term risks also could affect the short term profits of a firm and it is relevant for every shareholders, doesn’t matter what his horizon is. To substantiate the thought that investors always should have an eye on the tax policy, I will use anecdotal evidence of Starbucks.

Starbucks is fully scrutinized because of their tax deductions the last few years. There was a parliamentary hearing held by the British Parliament to gain insight in tax avoidance of British companies. Also Starbucks had to explain why they pay so little tax in the United Kingdom despite that they earn billions with their coffee bars. As results of the critic from the British Parliament, media and consumers Starbucks promised to pay in the future a fair tax rate. Important in this case is the critic of consumers, because that is an important factor for the reputation of Starbucks, what if customers ignore Starbucks because of their involvement in tax avoidance? That could really affect the firm value directly and shareholders should be aware of these developments and should require a reaction of the management. Starbucks act after the critic in order to mitigate the reputation scandal.

The anecdote above of Starbucks shows that reputation of a firm could be harmed in a quick moment due the tax policy and could affect in that sense the short term as well the long term firm value. To support this view Hanlon et al. (2009) examined the effect of news about corporate tax aggressiveness for firms and the reaction of stock prices. They searched for news articles and announcements about tax avoidance related to a company. Their results show that companies who were associated with tax avoidance in the news have to deal with a stock price decline. The results show a stronger effect for firms that operate in the retail sector, this could be due that these kind of companies are more consumer focused and are more susceptible for the argument of risks regarding reputation and boycotts of consumers. This leads to the conclusion that the long term does matter. Reputational risks could have a large impact for the shareholders. And the risk of a growing information gap is present.

(16)

16 Figure 1. Visualizing the disadvantages of aggressive tax planning for (major) shareholders. These risks should be an incentive for shareholders with influence to monitor active the tax policy of their investments.

Tax avoidance could lead to less future proftitability due to reputational risk through negative media attention, negative attention of governments and the possibility of penalties

To show profits and dividents to their shareholders, the managers could engage in tax avoidance Managers get an incentive with bonuses to increase the profits and dividends on behalf of the

shareholders

Shareholders want to maximize their return on investments

Less disclosure of information means an increase of the information gap and create a risk premium Decreasing the disclosure of the tax policy by managers because of fear for the tax authorities and

potential critic about the gap between earned income and paid taxes

Managers use accounting methods to substandiate their tax position and to decrease the gap between earned income and paid taxes

(17)

17

Block holders spread

For this thesis I created a measure ‘block holders spread’ to compute the level of major shareholders in the company. With block holders spread I refer to the spread of large shareholders, and their ability to get influence relatively compared to others. If there is one large shareholder he is better able to monitor the company than in the situation where there are several even large shareholders. Contrary to that situation, if there is one large shareholder and a lot of other small shareholders, there is more power for the largest shareholder. An effective monitoring role for shareholders in a firm with a lot of small shareholders is more difficult and I assume that the management have the ability go its own way by maximizing their own wealth by using tax avoiding. Associated risks of tax avoidance could harm the shareholders, especially the major stockholders (most affected through the risks).

This study contributes the literature in the first place with more insight in determinants of aggressive tax planning. Second, it gives more information about the role and influence of major shareholders with regard to tax avoidance and the agency theory. Lastly, this study is important for governments in their battle against the still growing gap between pre-income and the tax expense.

Hypothesis

Banks with low block holders spread have higher levels of tax avoidance

The hypothesis is strong based on the risks of tax avoidance and the theory that major shareholders could influence the tax policy. Based on the theory I predict that for firms with several small shareholders, managers have more abilities to engage in these risky tax policies in order to create maximum personal wealth for the manager.

(18)

18

III. Research methodology

This thesis aims to understand the factors that determine tax avoidance, especially the ownership structure. With this thesis I want to know if there is a relationship between the level of tax avoidance and the spread of shareholders. This could be relevant in the context of the agency theory, monitoring role and the effect of shareholders' influence on the policy of a firm.

I will examine only the banking industry in the United States and only listed firms in the United states (Nasdaq / NYSE). There are several reasons why I choose for this specific sector and country. It is important to note here that tax regulations is a totally national matter. International standards like IFRS or ISA lacking the scope of tax rates or regulations, mostly these standards have only the disclosure of tax in their scope. I choose for the United States because first of all, European countries have different tax frameworks, different tax rates and most important, enforcement of the tax rules differ enormously. In the United States the tax framework is for the most part equal, there is a small tax difference between the states but overall the enforcement of the government for large companies is the same for all companies.

Shareholders must fill out a form if they own a certain percentage shares in the company by the SEC (United States Securities and Exchange Commission). Databases that track all the forms for several years aren’t public or accessible for academics. The choice of the banking sector is partly made because of the availability of data with regards to shareholders and these forms. Therefore I have to put my focus on the data that is available for academics.

The banking sector isn’t representative for the whole population (all large firms in different industries), however by choosing a specific sector the comparability will increase. Banks in general have more or less the same balance sheet. The characteristics that bank balances share is that there are a lot of loans and derivatives on the balance sheet and in general less tangible assets. Therefore I assume that banks are overall in the same position to engage in tax avoidance. My consideration for the choice of only listed firms is simply made by the availability of the data, and the consideration that it is for shareholders most relevant because their shares are publicly traded, however this has limitations to extract the results for all banks or non-listed firms.

For collecting the data with respect to dependent variable and the independent variables I use the database of Bureau van Dijk / Osiris. The same database I will use for data with respect to control variables; Return on Assets, Leverage, Cash Holdings, Size and Intangible Assets. For this thesis I use IBM SPSS to exercise the regressions with the dataset.

(19)

19

Tax Measurement and the dependent variable

In the literature there is a discussion going on the topic which manner is the best to measure tax avoidance. Some academics have the opinion, that with the current disclosure requirements, it is impossible to measure tax avoidance in an appropriate way. Other researchers attempt to capture tax avoidance with the measure such as deferred tax liabilities (postpone tax payments could also be a form of tax avoidance and tax policy) or the paid taxes compared to the revenue of the firm or some academics use totally different measures to examine tax avoidance. Hanlon and Heitzman (2010) described a review of tax research and give information and insight about this subject, and the different possibilities to undertake research in this area. Based on their paper and the availability of data I will use the tax expense and pre-tax income (named as GAAP ETR). Deferred taxes have the disadvantage that it change constantly due to new estimations and change with the circumstances. I am aware of the fact that also GAAP ETR is only a moment in time and could differ over the years. And more importantly it could be affected by accounting procedures by the managers. Although the GAAP ETR in relation to the pretax income is informative about the tax gap that a company sends to the society, governments and investors. If a company has in general the policy to send the message to the outside world that they are responsible with tax, they will avoid a low GAAP ETR. Tax planning strategies that defer taxes will not have influence on the GAAP ETR. Therefore I conclude that the GAAP ETR as reported in the annual report is an appropriate measure for the purpose of this research.

I prefer to undertake a research for several years but the data with regards of shareholders is the bottleneck. Because this research is only for the year 2013, the outcome and conclusion of this research has his limitations.

Another important factor is that research regarding tax avoidance isn’t always simple because companies are reserved with disclosure about this topic (their internal policy etc.), mostly due to confidentially and fear that competitors and governments will take advantage of this information what could result in stricter enforcement and higher tax payments. Companies fear also the potential media attention and reputational risks related to tax. The opinion of the society could be an obstacle for firms to publish this kind of information, however I rely on published data of annual reports with regards to tax.

(20)

20 Based on the outline of Hanlon and Heitzman (2010) I will use for calculation of the dependent variable GAAP ETR (Effective Tax Rate). Hereby the risk of not reporting or not fully disclosure is nonexistent because every firm have to disclose this information. The computation for GAAP ETR is as follows:

Creating the variable “Block holders spread”

With regards to the shareholder information, shareholders of a firm could change overtime. With my dataset it wasn’t possible to choose one particular moment, therefore I only took companies that have shareholder information available for year end 2013 and begin 2014. My data assumption is that shareholders held the share at the reporting date of 31-12-2013. I collect the top five shareholders of every bank. I choose the top five shareholders because I assume that the monitoring role with more than five large shareholders is less possible due to the level of influence. I only use only banks were all five top shareholders were given, in order to have a fair comparison.

Formula for main analysis

V = Block holders spread

q = Shareholder percentage of total shares

q_{6 = Remaining shareholders percentage}q_{6 = 100% – (}q₁₊q₂₊q₃₊q₄₊q₅₎ δ = Weight factor of the blocks (0.25)

I use the above formula to calculate the “Block holders spread” as independent variable. In order to give firms with relative more large shareholders a higher block holders spread, I use a δ in a range of 0-1 and for each subsequent shareholder the power will be higher. Firms with a very large shareholder get a higher block holders spread as result than a firm with several relatively small shareholders. The shares that are not held by the largest five shareholders, are merged in q6. This because with more than five large shareholders it will be difficult to have an effective monitoring role. q6 is 100% minus the sum of the five largest shareholders.

(21)

21 The hypotheses predict that firms with a low block holders spread will have higher level of tax avoidance. If I translate this hypothesis to the formula; Firms with a low block holders spread have a low GAAP ETR percentage. This is accordance the literature of chapter two, where I describe that tax avoidance could lead to significant risks and I suppose that companies with a (few) major shareholder(s) have a better opportunity to monitor the firm and have influence on the tax policy.

For the main analysis I use a δ of 0.25, a low δ strengthens the position of the largest shareholders. In order to test the hypotheses with different influence levels I add several robustness tests with the δ of 0.5 and 0.75. These robustness tests meets the demands of more skeptical people with regards of the influence of these major shareholders on the tax policy and these robustness tests will contribute to substantiate the conclusions of this research.

In order to substantiate the outcome of the main analysis I add an extra analysis for the three largest shareholders and the remaining shareholders. For this extra analyses I use the same formula, beside I use three shareholders and the remaining shareholders. This could give more information about the impact of the (three) largest shareholders, this because you could reasonable argue that the fourth and fifth largest shareholders have less or no influence and therefore hardly affected the tax policy.

Formula for the three largest shareholders

V= Block holders spread

q = Shareholder percentage of total shares

q_{4 = Remaining shareholders percentage}q_{4 = 100% – (}q₁₊q₂₊q₃₎ δ= Weight factor of the blocks

(22)

22

Control Variables

In prior literature authors controlled for factors that could also affect the level of tax avoidance (Dyreng et al. 2008; Chen et al. 2008). The conclusion and the results of this paper will be more valid, more robust and put the results in perspective with several variables. In the first place I add the control variable for leverage. Leverage plays a role due interest payments in tax avoidance. Also Return on Assets could affect the level of tax avoidance, this because high profitable firms have to pay more taxes relative to firms with low ROA. Same as Dyreng et al. I choose to add the control variable Cash Holding, since firms with little cash might be inclined to postpone tax payments. Firm size is could also be a factor for tax avoidance with the ability of the firm to engage in complex and complicated tax avoidance with expensive consultants. Finally I add intangible assets as control variable because in the literature intangible assets is often named as a possibility to structure the tax avoidance

Table 1. Variables

Control Variable Computation

Leverage Debt to equity ratio=

Total liabilities --- Shareholders equity

Return On Avg Assets

(ROAA) Return On Average Assets (ROAA) =

Worldwide Income --- Average Assets

Cash Holding Cash Holding =

Cash and Due from Banks --- Total assets

Size Size = Total assets

Intangible of total assets Intangible of total assets =

Intangible assets --- Total assets This table shows the variables used in this research.

All of the control variables are susceptible for critic, because the way companies structure their tax avoidance differ, all of the control variables aren’t fully suited for all companies. However with the nature of the dataset (only banks) I use five control variables and the block holders spread I have the opinion that it is robust enough for the purpose of this research.

(23)

23 Table 2. Data selection

Total banks in the database 47576

Total banks listed at NYSE or NASDAQ at 31-12-2013 773 Information available for the top five shareholders 742 Tax information available for the top five shareholders 580

(Check) Pre-tax -/- tax = post-tax income 415

Deleting Negative GAAP ETR 367

The above table shows the selection of the examined population. In the end I use 367 banks for this research. Only firms that have tax information at 31-12-2013 I will use for the population.

The dataset I use originally exist of 773 banks listed on the NYSE or NASDAQ. After filtering for the completeness of the top five shareholders there remain 742 banks. For these 742 banks I searched for the Pre-tax profit with fiscal year end 31-12-2013, tax and post-tax profit. With regards to the tax data not all the data was useful mainly due to different fiscal years. With regards to the tax data that was available with the appropriate fiscal year I should be sure that it is right measured. In order to obtain certainty I checked the requested data with formula: Pretax profit -/-Taxes should be Post-tax profit. Only banks were data with regards to tax was available and the verification with the formula was fine I will use, therefore 415 banks meet these requirements and remain.

For the remaining 415 firms I take only the banks with a positive GAAP ETR, this results in the end of a dataset that contains 367 banks.

I haven’t deleted any outliers therefore it could be that the dataset contain extreme values. Before I start with regressions I interpret the plain data, there were very few outliers. Some of the outliers were extreme, however with a dataset of 367 the impact won’t be large.

(24)

24

IV. Analyses and results

In this section I will discuss the results and explain the analyses in depth. In the literature there is some evidence that firms have as average GAAP ETR about 30%. For example Dyreng (2010) found in his research a mean of the GAAP ETR of 30.9 %. This is according to our findings of a mean GAAP ETR of 31.01 % in table 3 and the median is 30.64%. Beside the average and the median we see that the minimum and maximum (0.01 % and 279 %) differ very much. This confirms the difficulty of a proper tax measure related to the comparability. Although the mean and the median of table 3 compared to the outcome of earlier research seems to lead to the conclusion that the population is a reasonable representative of the whole population and therefore is appropriate for this research.

Main analysis

Table 3 presents the descriptive statistics for our sample size of 367 banks. We see that the leverage (Debt to equity) is relatively high, the mean 8.36 implicates that banks have on average 8 times more debt than equity. The (high) ratio of 8.36 confirms the high leverage in the banking industry. This is consistent with the literature in the way that banks rely on trust and have a social function to provide consumers and companies with loans and therefore it is conventional that banks have a higher debt ratio. In addition of the high leverage table 3 shows that cash holdings is also relatively low compared to other industries.

The profitability of the banks is embedded in the variable ROAA. Profitability could be an factor for tax avoidance because, if a company has a lot of profit he should pay more taxes. The mean of the ROAA is about 1.2%. To my knowledge this isn’t very high. An possible explanation could be that the aftermath of the financial crisis of 2008 is a reason for this relatively low ROAA level for the banking sector. Table 3 shows that intangible assets isn’t an important element on the balance sheet of banks. The mean of the intangible assets related to the total assets is less than 2.2%. My explanation for the low presence of intangible assets is simply because of the kind of services that banks provide. In general we see for the variables ROAA, cash holdings and intangible assets little spread, this could affect the results, because if the population nearly differs for these variables the chance of significant become smaller.

(25)

25 Table 3. Descriptive statistics of variables

Variable Mean Minimum Maximum Percentile

0.25 Percentile 0.50 Percentile 0.75 Tax rate 31.01% 0.01% 278.87% 24.60% 30.64% 34.73%

Debt to equity ratio 8.36 0.04 24.66 6.87 8.26 9.82

Return On Average Assets

(ROAA) (%) 1.20% -1.63% 38.08% 0.64% 0.86% 1.10%

Cash holdings (%) 2.66% 0.10% 43.90% 1.10% 1.65% 2.52% Size ( Total Assets )

($ x 1000)

73,955 75 8,717,198 1,124 2,347 7,763 Intangible of total assets (%) 2.17% 0.00% 50.64% 0.19% 1.12% 2.66% This table examines the spread and gives an overview of the characteristics of the 367 banks in the population.

My primary research question relates to the low block holders spread and the higher level of tax avoidance. Based on the agency theory and the ability of large shareholders to use their influence for lowering the risk of tax avoidance. I use a δ of 0.25 in order to give firms with a few large shareholder more weight and firms with a lot of different smaller shareholders get a lower block holders spread. Table 4 shows the descriptive statistics of the main analysis. The table shows that the minimum and the maximum differ very much, however in general the values of the first quartile and the third quartile are relatively closer to each other. However it seems that the spread really differs between the population of banks. For this population I conclude that there is serious difference in presence of large versus small shareholders.

Table 4. Descriptive Statistics for the variable block holders spread

δ 0.25 Mean 0.87 Percentiles 0.25 0.58 Percentiles 0.50 0.71 Percentiles 0.75 0.82 Minimum 0.18 Maximum 7.59

The table presents the descriptive statistics for the analysis with regards to main analysis. The minimum and maximum differ a lot, although the Q1 and Q3 are both under 1.

(26)

26 In the theory in section II based on the literature I hypothesize that the spread of shareholders could play an important role for the tax policy of the firm. Therefore I hypothesize that a low block holders spread is associated with a low GAAP ETR ratio. Table 5 presents that with a δ of 0.25 and the block holders spread formula the block holders spread is a significant variable. The variable block holders spread is associated to the level of tax avoidance measured as GAAP ETR. This implicates that for the examined sample it is the case that when the block holders spread become higher we expect that the GAAP ETR ratio become also higher. If I formulate it the other way around, based on the results of table 5, for a low block holder spread we expect a lower GAAP ETR. This is in accordance with the hypotheses, based on the population of 367 banks in table 5, the formulated hypothesis is true.

Prior literature assumed that leverage (Gupta & Newberry, 1997 and Brailsford, 2002) is a potential important factor (and ability) for the level of tax avoidance. Table 5 seems to confirm the view of the relation between leverage (total liabilities divided by shareholders equity). Leverage is a significant variable at a significance level of 90%. It is less significant for this population than the block holders spread, nevertheless it is significance. It is according to prior literature that firms that are high leveraged are negatively associated with GAAP ETR ratio. A potential explanation for this result could be that firms that take risks with leverage this risk policy also applies on their tax policy. And maybe a better explanation is that debt in combination with interest payments is often used to avoid tax, therefore the results that leverage is significant is logical. Table 3 with the descriptive statistics of the variables shows that leverage levels differ a lot for the banks.

For the other variables Cash Holdings, intangible assets and ROAA there is in table 5 no evidence for significant effect on the dependent value GAAP ETR. In combination with table 3 (descriptive statistics) there is a possibility that these variables are not significant because of the little difference between the quartiles therefore these variables have less the ability to explain the relationship.

Size measured as total assets on the balance isn’t significant, this is partly contrary to the literature (Zimmerman, 1983). Zimmerman find in his result that for the largest companies (fifty largest companies at that moment) the ETR was higher than other firms. He suggest that a higher ETR for the largest firms is because of there is a political issue, namely lobbying and potential government scrutiny and stricter regulation. However for our sample of 367 banks firm size is not significant.

(27)

27 Table 5. Results of the main analyses

Dependent variable: GAAP ETR Coefficient

(T-Value)

δ 0.25

Block holders spread 0.185***

(-3.528)

Leverage (Debt to equity ratio) -0.092*

(-1.537)

Return On Average Assets (ROAA) -0.013

(-0.174)

Cash holdings -0.067

(-0.954)

Size (Total Assets ) -0.062

(-1.205)

Intangible of total assets _(-0.757)-0.043

* Significant at 90 % confidence level ** Significant at 95 % confidence level *** Significant at 99 % confidence level

Table 5 examined the relation of several independent variables on the dependent variable GAAP ETR. The table shows that there are two significant variables namely block holders spread and leverage.

Robustness checks

I assume with the main analysis and a delta of 0,25 a large influence difference between the largest shareholders relative to the other shareholders. Because this only my assumption based on the theory I add robustness checks. These tests could contribute to persuade critics of my conclusion therefore I add several additional analyses to gain a better overview for the association between the block holders spread and tax avoidance. For the analysis as showed in table 6 I use the same formula as earlier mentioned. The only thing I modify is the δ from 0,25 to 0.5 and 0.75. Set δ higher implicates that the weight of major shareholders decline compared to the weight of smaller shareholders. Furthermore I adjust the formula, for a regression with only the 3 largest shareholders, which results are presented in the fourth, five and sixth column of table 6.

(28)

28 Table 6 shows the descriptive statistics of the additional analysis and compared to table 4 (Descriptive Statistics main analysis) we see a substantially higher block holders spread. This is a logical result of a higher δ. Table 6 shows that with a higher δ the variance decline. In table 4 for the smallest block holders spread for a firm is 0.18 and the highest 7.59, so the highest is 42 times as large than the smallest. And for δ 0.75 in table 6 the highest value for block holders spread is 5.5 times as large as the smallest value. For the model and the theory behind the model less variance implicates that there is less difference in influence for the largest shareholders relative to the other shareholders. The assumption of influence and the ability for the major shareholders to affect the tax policy erodes slowly with increasing the delta in the formula. Table 6. Descriptive Statistics of block holders spread (robustness checks)

Block holders spread Top 5 shareholders Top 3 shareholders δ 0.5 0.75 0.25 0.5 0.75 Mean 4.94 22.29 0.87 4.80 21.61 Percentiles 0.25 3.62 19.65 0.58 3.49 19.04 Percentiles 0.50 4.26 21.01 0.70 4.11 20.31 Percentiles 0.75 4.95 22.67 0.81 4.74 21.81 Minimum 1.67 15.34 0.18 1.63 15.15 Maximum 36.17 85.07 7.58 36.02 84.34

Table 6 examined the extra analyses for different δ and examined if the formula contain three shareholders instead of five large shareholders. The table shows that an increase of the δ results in an increase of the block holders spread and an decrease of the variance. The descriptive statistics for the top three shareholders are in general conform the values showed in the top five.

The results of the additional analyses with higher a delta and with only the top three shareholders are presented in table 7. The table shows in general the same results as the main analyses in table 5. The significance of block holders spread declines with a higher delta. The other variables in general stay constant. Still the only other variable that is significant is the leverage variable.

(29)

29 Table 7. Results of additional analyses

Dependent variable:GAAP ETR Coefficient _(T-Value)

Top 5 shareholders Top 3 shareholders

δ 0.5 0.75 0.25 0.5 0.75

Block holders spread 0.153** _(2.893) 0.127** _(2.399) 0.185*** _(2.399) 0.154** _(2.920) 0.131** _(2.464) Leverage (Debt to equity ratio) _(-1.515)-0.091* _(-1.512)-0.091* -0.091* _(-1.512) _(-1.517)-0.091* _(-1.516)-0.092*

Return On Average Assets

(ROAA) (-0.104) -0.008 (-0.047) -0.003 (-0.047) -0.008 (-0.110) -0.008 (-0.060) -0.004

Cash holdings _(-.959)-0.068 -0.068 _(-.954) _(-.954)-0.068 _(-.959)-0.068 -0.068 _(-.955) Size (Total Assets ) _(-1.204)-0.063 _(-1.204)-0.063 _(-1.206)-0.063 _(-1.208)-0.063 _(-1.212)-0.063 Intangible of total assets _(-0.741)-0.043 _(-0.743)-0.042 _(-0.756)-0.043 _(-0.739)-0.043 _(-0.731)-0.042 * Significant at 90 % confidence level

** Significant at 95 % confidence level *** Significant at 99 % confidence level

The above table examines the association of the independent variables for the dependent variable GAAP ETR. The results shows that there are for several delta two significant variables namely block holders spread and leverage. The results are nearly the same as table 5, with the remark that the significance level decrease.

The additional analyses of table 7 strengthen the results of table 5 that the block holder spread has a significant positive association with the level tax avoidance. Also the results of table 7 is in accordance the hypothesis that low block holders spread is associated with lower levels of GAAP ETR.

With the formula used for computing the block holders spread there is a possibility that the remaining part of the shareholders (q6), even if they get a delta to the power 7 weight too much compared to the other shareholders. Because of the less ability for them to use their influence, it is arguable based on theory that you shouldn’t use them at all. In order to test if this is the case, I run an extra regression. The formula for the block holders spread without the remaining shareholders is as follows;

Formula without the remaining shareholders, this makes sure that we now only look at the major shareholders and the structure between these five shareholders.

Table 8 presents the results of this additional regression without the remaining shareholders. The results confirm again the hypothesis that the variable block holders spread is significant for each delta that is chosen. Also in table 8 the variable block holders spread become less significant with a higher delta. In the table the results for the other variables are less or more the same. Only the

(30)

30 leverage is significant, all the other variables are not significant for the dependent variable GAAP ETR. Table 8 is very valuable for our conclusion based on the spread holders formula. Because for the remaining part of the shareholders q6, it could be that they have too much weight and therefore influence the outcome too much. With the part in the formula q6. δ7 the remaining shareholders got a low value, but if remaining part of the shareholders has although a large percentage their influence could be very high in the formula. Now table 8 shows nearly the same outcome as table 5, we see that the formula was correct for our purpose.

Table 8.

Dependent variable: GAAP ETR Coefficient _(T-Value)

Top 5 shareholders

δ 0.25 0.5 0.75

Block holders spread 0.185*** 0.149** 0.113**

(3.528) (2.816) (2.113)

Leverage (Debt to equity ratio) -0.092* -0.093* -0.094*

(-1.537) (-1.555) (-1.559)

Return On Average Assets (ROAA) -0.013 -0.045 -0.040

(-0.147) (-0.644) (-0.569)

Cash holdings -0.067 -0.021 -0.020

(-.954) (-.304) (-.283)

Size (Total Assets ) -0.062 -0.061 -0.061

(-1.205) (-1.172) (-1.171)

Intangible of total assets -0.043 -0.048 -0.048

(-0.757) (-0.840) (-0.834)

* Significant at 90 % confidence level ** Significant at 95 % confidence level *** Significant at 99 % confidence level

Table 8 shows the results for a different computation of the block holders spread. For the independent variable block holders spread are only the 5 largest shareholders included. This confirms the view of table 4 and table 6 that only the block holders spread and the leverage are significant in all the regressions.

Relationship between block holders and aggressive tax planning