1
Amsterdam Business School
The relation between CEO characteristics, focused on individual
cultural aspects, and corporate tax avoidance
Name: Lorenzo Baars Student number: 10365583
Thesis supervisor: prof. dr. L.R.T. van der Goot Date: June 20th 2016
Word count: 17,422
MSc Accountancy & Control, Specialization Control
2
Statement of Originality
This document is written by student Lorenzo Baars who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.
The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
3
Abstract
This research addresses the increasingly relevant topic of corporate tax avoidance. Prior research demonstrated that corporate tax avoidance is widespread and institutions such as the European Union and the International Monetary Fund have expressed their concerns. However, literature mainly focuses on the relation between firm characteristics and tax avoidance, while little attention is being paid to the incremental influence of individual executives. Therefore, this study contributes to earlier research by examining the relation between corporate tax avoidance and individual CEO characteristics, while mainly focusing on cultural background. In order to attempt to answer this question, quantitative research has been conducted through making use of three publicly available databases. This includes one database containing data on four managerial cultural dimensions, being power distance, individualism, masculinity and uncertainty avoidance. As a result, I provide early evidence for a relation between corporate tax avoidance and individual cultural CEO characteristics, as CEOs from less masculine cultures, thus feminine, tend to engage in tax avoidance to a greater extent. Besides that, I also find evidence for the relevance of biographical CEO characteristics and fixed firm characteristics. As such, older and newly appointed CEOs are more likely to engage in such activities, just like firms which report higher profitability and lower leverage and firms which are publicly quoted, newly incorporated or larger in size. Concluding, this research proves not only that firm characteristics are related with corporate tax avoidance, but it also demonstrates the importance of individual biographical and cultural CEO characteristics.
4
Contents
1. Introduction 6
2. Theory and hypothesis development 9
2.1. Corporate tax avoidance 10
2.2. Consequences of corporate tax avoidance 11
2.3. Agency theory 12
2.4. Prior research 14
2.5. Individual cultural characteristics 17
2.5.1. Power distance 18
2.5.2. Individualism 19
2.5.3. Masculinity 20
2.5.4. Uncertainty avoidance 21
2.6. Summary 22
3. Methodology and descriptive statistics 23
3.1. Data on Chief Executive Officers 24
3.2. Data on the four cultural dimensions 28
3.3. Data on corporations 31
3.3.1. The measurement of corporate tax avoidance 31
3.3.2. Financial corporate control variables 33
3.3.3. Non-financial corporate control variables 34
3.4. Regression model 36
4. Regression results 38
4.1. Analysis of the regression results before adjusting for outliers 38 4.2. Analysis of the regression results after adjusting for outliers 40 4.2.1. Analysis of the results on cultural CEO characteristics 41 4.2.2. Analysis of the results on biographical CEO characteristics 43 4.2.3. Analysis of the results on fixed firm characteristics 44
4.3. Robustness check on the company size measure 46
5. Conclusion 47
5
List of tables
Table 1 Data collection and sample size development 24
Table 2 CEO gender distribution 25
Table 3 CEO year of birth distribution (in decades) 25
Table 4 Variable definitions 26
Table 5 CEO title years distribution 27
Table 6 CEO nationalities and cultural values 29
Table 7 Corporate tax rates and local exchange rates to EUR 32
Table 8 Descriptive statistics on the financial independent variables 33 Table 9 Company size distribution as defined by Bureau van Dijk
AMADEUS Managers 34
Table 10 Corporate ownership distribution 35
Table 11 Corporate status distribution 35
Table 12 Corporate incorporation year distribution (in quarter centuries) 35 Table 13 Pearson correlation coefficients for the independent variables 37
Table 14 OLS regression results before adjusting for outliers 39
Table 15 OLS regression results after adjusting for outliers 41
6
1.
Introduction
Over the past few years, tax avoidance has become an increasingly relevant topic. Already during the 1990s and the early 2000s an increase has become visible in the amount of firms which engage in tax avoidance activities through different tax shelter strategies (Rego & Wilson, 2009). Also in later years, for example during the financial crisis which affected Greece, tax avoidance and tax evasion have been of great concern to regulators such as the European Union and the International Monetary Fund (IMF) (Unknown, 2012). In Greece, an amount of 22 billion euros was sent abroad by a total of 54,000 taxpayers between 2009 and 2011 and as a result, no taxes were paid on this amount. This led to the Greek Ministry of Finance naming tax evasion a national sport. Also after the financial crisis, tax avoidance remained a public concern, as can be observed by looking at the in 2016 leaked Panama Papers (Akkerman, 2016). According to Jeroen Smit, financial writer and former chief editor of the Dutch Financial Times, companies and its CEOs do not seem to fully realize their position in the society in which they operate. For example, on the one hand, a company like Starbucks uses sustainable coffee beans, but on the other hand, it exercises innovative company structures to prevent having to pay taxes. As the latter also goes for Apple, its CEO Tim Cook was asked for the reason for this behaviour, as Apple did benefit from American facilities such as education and infrastructure, but it did not pay American taxes. He answered to this by stating they did not pay taxes because they have the opportunity to do so.
As a result of the above mentioned concerns, research has focused on corporate taxation. For example, Dyreng, Hanlon and Maydew (2008) point out that the mean effective tax rate for American firms is 29.6 percent, measured over a ten year period from 1995 to 2004. Besides that, the median is 28.3 percent. However, they also find that 26.3 percent of the companies is able to keep its effective tax rate below twenty percent. Furthermore, another 9.2 percent manages to keep this rate below ten percent over this ten year period. Hence, this suggests that a considerable amount of companies engages in tax avoidance as effective rates stay low for a long period of time.
However, when reading prior literature, it appears that the field of corporate tax avoidance has not been fully explained yet (Hanlon & Heitzman, 2010). Mainly, the focus in literature is at the relationship between the extent to which firms manage to pay a lower amount of taxes and the characteristics of those companies, both financial and non-financial. For example, Wilson (2009) finds that bigger firms are more likely to be accused of tax
7
sheltering. Also firms which have an aggressive way of financial reporting and which operate more internationally tend to pay fewer taxes. Lisowsky (2010) confirms this by stating that firms are more likely to have tax shelters when they earn a higher income abroad and when they possess subsidiaries in countries identified as tax havens. As other indicators, he also names inconsistent reporting and size and he also finds positive relationships for litigation losses and profitability and a negative relationship with the amount of leverage a firm carries. Furthermore, Hanlon and Heitzman (2010) conclude out of literature that ownership may also have an effect on tax avoidance, because whether or not a company is publicly quoted may predict the amount of taxes that company will pay. However, this relationship is not clear and can be argued both ways.
Continuing, Dyreng, Hanlon and Maydew (2010) confirm the above stated current state of research, writing that research has mainly focused on firm characteristics and also pointing out the relations with company size, operations abroad and reporting tactics. However, it appears from their study that tax avoidance cannot be explained solely by these firm characteristics. In their results, they find evidence that top executives individually play a significant part in the firm’s act of engaging in tax avoidance. As for this, an eleven percent gap can be observed by looking at the interquartile range (third quartile minus the first quartile) of the effective tax rates realized by a sample of top executives, calculated from the financial statements. This sample included not only CEOs, but also CFOs and other executives. However, the results for CEOs appeared most significant. This contributes to prior evidence that the extent of tax aggressiveness is highly dependent on CEO and CFO compensation, as Rego and Wilson (2009) found a strong positive relationship on this. Also Phillips (2003) found a similar result. Despite observing that CEO compensation based on after-tax performance measures does not directly result in decreasing effective tax rates, he does find a negative relationship for unit managers. However, as CEOs are responsible for the compensation of these unit managers, they do have indirect influence, as they are more likely to grand these compensations if they are compensated on after-tax measures themselves too.
Besides the above stated findings, Dyreng et al. (2010) have also looked at the relation between tax avoidance and biographical characteristics specific to executives, but they were only able to find little evidence for this. One of the reasons for this is missing data. They therefore can only yet conclude a general relation between executives and tax avoidance. As they are the first to report on this relationship, further research is required on this topic. This is of particular importance, because of the societal effects tax avoidance can have (Hanlon & Heitzman, 2010). For example, not only are managers affected by it, also shareholders,
8
creditors and government can face negative effects. One way to verify the existence of such effects, is looking at the market reaction of the company’s shares after tax avoidance activities become publicly known. As such, Hanlon and Slemrod (2009) observed a 1.04% decrease in share prices after the first appearance in press regarding tax avoidance activities. It is therefore important to gain an understanding of which managers are most likely to engage in tax avoidance and why they behave in this way while other managers do not, just like this is important for firms as a whole (Rego & Wilson, 2009). Then, clarity regarding this question provides us with the opportunity to compose optimal contracting on performance compensation and employee motivation in order to prevent tax avoidance.
Given the above mentioned current state of literature, I contribute to research by extending our knowledge on the relationship between individual CEO characteristics and the amount of corporate tax avoidance, as a research gap still exists regarding this particular topic. In this article, I therefore address this by using the four cultural dimensions applying to individuals as formulated by Hofstede (1980), being power distance, individualism, masculinity and uncertainty avoidance. Hence, I look at the extent to which not only simple biographical, but also cultural characteristics are predictive for the amount of tax avoidance an individual CEO is likely to engage in. To be able to do this, I make use of the agency theory where CEOs, shareholders and also lower managers serve as the key participators within a firm. As a result, I am able to explain the behaviour of these individual CEOs and shareholders regarding compensations based on after-tax measures which might stimulate tax avoidance. This is of particular importance, because a thorough understanding of this behaviour provides relevant stakeholders with the opportunity to address these problems by composing for example proper regulations and proper contracting regarding performance compensation for individual CEOs. Taken together, the purpose of this article can be formulated concisely as investigating whether CEO characteristics, focused on individual cultural aspects, are related to corporate tax avoidance.
To address this question, I perform a quantitative research in which I use three publicly available databases. Firstly, I use Bureau van Dijk AMADEUS Managers to retrieve a set of CEOs and several relevant biographical characteristics such as gender, age and nationality. Secondly, I retrieve the tax data and several other financial control variables of the companies at which these CEOs are employed. In order to collect these matching data, I use Bureau van Dijk AMADEUS Financials. Both these Bureau van Dijk databases contain a wide range of fourteen million different European companies. However, most valid data retrieved for this research originates from companies located in Italy. Lastly, in order to
9
construct measurable variables for the four cultural characteristics as formulated by Hofstede (1980), I use a database set up by Hofstede himself in which he scores different nationalities for the characteristics he formulates (Hofstede & Hofstede, 2014). As such, I am able to match these values with the nationalities retrieved from the first database. Then, by combining these three databases, I run an OLS regression after which I draw conclusions about the relation between CEO characteristics and corporate tax avoidance.
As a result, I find that individual cultural CEO characteristics are indeed related to the extent to which companies engage in corporate tax avoidance activities. Foremost, this research provides evidence that CEOs from more feminine countries, thus less masculine, are more likely to engage in tax avoidance activities. For the other three cultural dimensions however, such evidence has not been found. Despite this, the result on masculinity does provide a first indication for the relevance of the cultural backgrounds of individual CEOs regarding the tax avoidance topic. Moreover, this study also provides evidence for biographical CEO characteristics and fixed firm characteristics to be of significant influence on the extent of corporate tax avoidance. As such, CEOs from earlier birth years and CEOs who have recently been appointed engage in corporate tax avoidance to a greater extent. Also, regarding fixed firm characteristics, high profitability, low leverage, large company size, being publicly quoted and recently being incorporated contribute to a higher likelihood of engagement in tax avoidance as well.
Finally, in the remainder of this text, the second chapter contains an elaboration on the theoretical background of corporate tax avoidance, prior research on this, agency theory and CEO characteristics, mainly focused on Hofstede (1980). Also, the second chapter contains the hypotheses I formulate based on this theory. Then, in the third chapter I continue by explaining my research methodology, after which I will discuss the results of this research in the fourth chapter. Finally, I come to a conclusion in the fifth chapter, in which I also discuss limitations and suggestions for future research.
2. Theory and hypothesis development
To be able to gain a thorough understanding of the relationship between CEO characteristics and the corporate tax avoidance in which these CEOs engage, it is necessary to give a clear definition of all relevant concepts. Therefore, in this chapter I first give a definition of tax
10
avoidance, after which I name several positive and negative consequences. Then, by using agency theory, I explain the behaviour of CEOs and shareholders regarding tax avoidance. Following this, I elaborate on the current state of research regarding corporate tax avoidance. Lastly, as this research mainly focuses on cultural characteristics of individual CEOs, I will discuss the four cultural dimensions formulated by Hofstede (1980), being power distance, individualism, masculinity and uncertainty avoidance. Also, by relating these concepts to tax avoidance and using the agency theory for this, I derive hypotheses for each of these four characteristics, on top of two hypotheses on firm and biographical CEO characteristics.
2.1. Corporate tax avoidance
Firstly, looking at existing literature, it is important to acknowledge that no single definition of tax avoidance yet exists. Researchers use different measures and also multiple phrasings are used, such as tax avoidance, tax evasion, tax aggressiveness and tax sheltering (Hanlon & Heitzman, 2010). Mostly, these terms can be used interchangeably, except for tax evasion, which is perceived as the illegal and intentional practice of reducing tax duties. As a result of these various phrasings, people generally derive different meanings from it. To take this indistinctness away, Dyreng et al. (2008) simply define tax avoidance as every practice that reduces the effective tax rate a company eventually pays. Hence, they imply that tax avoidance is not necessarily wrong, as it can come forth from not only illegal practices and grey areas, but also from legal tax reductions. For example, tax law can provide tax reducing provisions or it can be precarious in certain situations, such as complex transactions. Therefore, large companies may be able to benefit from this unclearness. Rego and Wilson (2009) share this vision, as they also define tax aggressiveness as a practice which can either be fraudulent, but can also be legal regarding to tax law.
Continuing, based on the above stated visions, Hanlon and Heitzman (2010) extend this by developing a continuum for tax avoidance, stating that on the one hand, one can recognize legal tax reducing activities such as tax deductible investments. However, moving to the other side of the continuum, we find tax aggressiveness, tax sheltering, noncompliance with tax law and tax evasion. As stated above, tax evasion is perceived illegal and should therefore be placed at the outer end of the continuum. This distribution can be confirmed by looking at a sample of fiscal officers, business lawyers and business students (Kirchler, Maciejovsky, & Schneider, 2001). Generally, this sample perceives tax evasion as a rather negative practice, while perceiving tax avoidance positively, despite both having the same
11
economic consequences. However, fiscal officers have a greater tendency to perceive both as unfair. So concluding, each tax reducing activity can be placed anywhere on the continuum, based on the extent to which it is perceived either legal or fair, or illegal or unfair. However, this is therefore dependent on personal believes, as Hanlon and Heitzman (2010) argue: “much like art, the degree of aggressiveness (beauty) is in the eye of the beholder; different people will often have different opinions about the aggressiveness of a transaction” (p. 137).
Secondly, literature is also inconsistent regarding the measurement of the effective corporate tax rate. Above however, tax avoidance is defined as any activity that reduces this effective tax rate. Therefore, it is important to clarify this measurement. As Hanlon and Heitzman (2010) observe, there are several different measures which can be found commonly in research. However, mostly these measures are retrieved from financial statements, because real tax data, as a property of the government, are often not available to the public. Therefore, one of these measures is the effective tax rate based one the Generally Accepted Accounting Principles (GAAP). According to these standards, companies report their pre-tax income, tax expenses, deferred tax assets and deferred tax liabilities. For example, also doing research on the relation between executives and corporate tax avoidance, Dyreng et al. (2010) also make use of this measure. Hence, they measure the effective corporate tax rate by dividing the total tax expenses by pre-tax GAAP income. As a result, they are able to include all activities which lead to a decrease in the effective corporate tax percentage. By doing this, tax avoidance in general is being addressed rather than one particular part of the continuum, for example illegal tax evasion. Consequently, this measure succeeds in capturing the reach of the above formulated definition of tax avoidance.
2.2. Consequences of corporate tax avoidance
Tax avoidance involves both positive and negative consequences for different stakeholders. To start, governments can suffer great tax income losses as a result of tax shelter strategies that involve the use of foreign subsidiaries (Gravelle, 2015). For example, firms operating in multiple countries might use the opportunity to reallocate their income from countries applying high corporate tax rates to countries applying low ones. Also, losses, debts and tax credits are shifted the other way around to lower taxable income. This way, very little tax can be collected on multinationals having foreign activities. For example, yearly losses with respect to this are estimated to be one hundred billion dollars for the United States only.
12
Contrary to this, the reduction of taxes payable increases the net income for the company and thus the return to shareholders (Hanlon & Heitzman, 2010). However, as tax avoidance may also contain an illegal part, companies face the risk of being detected by tax authorities. As a result, these companies are at risk of being sentenced through for example fines and interest on top of their regular taxes. Therefore, tax avoidance could also result in a decrease of firm value. Another reason for this is the market reaction when information about the firm engaging in tax avoidance goes public, as the market value of a firm on average decreases by 1.04 percent after the first media attention on a tax shelter (Hanlon & Slemrod, 2009). Moreover, this result can be amplified by consumer or taxpayer opinion. Taken together, firms should seek for the optimal balance regarding tax avoidance, taking into account both benefits and costs. However, according to Weisbach (2001), firms do not succeed in this. He argues that firms engage in less tax avoidance than what would be optimal given the many opportunities and low detection risk. Desai and Dharmapala (2009) then explain this phenomenon by stating that a firm’s engagement in tax avoidance is also positively related to the extent to which shareholders are in control over the behaviour of their CEOs, in other words the presence of good corporate governance.
2.3. Agency theory
Having earlier defined the concept of tax avoidance and how to measure this, it is also important to understand why firms do or do not engage in the practice of tax avoidance. Looking at prior literature, this behaviour can be explained by the agency theory (Hanlon & Heitzman, 2010). This theory is defined by Jensen and Meckling (1976) as a contractual relationship between two participators. This contract serves to make sure that the first party, called the agent, acts on behalf of the second party, called the principal. In order to realize this, the principal has to delegate his decision rights to the agent, which thus involves giving authority to the agent. Furthermore, in this model Jensen and Meckling assume both participators to maximize their utility. However, as conflicts of interest may arise, the agent might not always behave in the way his or her principal demands. To solve for this, the principal then sets up a contract in which he attempts to motivate the agent through incentives, either monetary or non-monetary, such that the agent will comply with the principal’s best interest.
Likewise, this theory can also be applied to the behavioural part of tax avoidance. In this case, the shareholders of a firm act as the principal, while the CEO acts as the agent
13
(Hanlon & Heitzman, 2010). Both thus attempt to maximize their utility. As research shows, tax avoidance activities can create value for the company. Therefore, shareholders are likely to steer the company in this direction (Rego & Wilson, 2009). However, because the CEO has the authority to decide upon this, a conflict of interest arises if the respective CEO is not able to extract any benefits out of it. Therefore, shareholders have to resolve this conflict by compensating the CEO for engaging in tax avoidance (Hanlon & Heitzman, 2010). They then do this by incentivizing the CEO through performance measures based on after-tax profit. Consequently, avoiding taxes leads to higher after-tax profit figures, which then result in a higher utility for the CEO.
However, Rego and Wilson (2009) recognize that companies do not incentivize CEOs to engage in tax avoidance to an unlimited extent. They explain this by not only naming the benefits of tax avoidance, but also the costs. Besides paying these incentives to CEOs, costs may arise in several variations. For example, developing new tax avoidance strategies can be very costly as a result consultancy costs or attorney costs. Besides that, by engaging in illegal forms of tax avoidance, the company is also at risk of being detected. Detection can then result in sentences such as a higher tax obligation or interest on the avoided taxes (Hanlon & Heitzman, 2010). Besides this, Rego and Wilson (2009) also name costs regarding audits by the tax authorities and legal costs. On top of that, they also mention that costs with respect to the company’s reputation can be incurred. Hanlon and Slemrod (2009) mention some examples for this, such as increasing transaction costs due to customers and suppliers becoming more cautious, or even the loss of these customers and suppliers. Finally, detection also renders a signalling effect for other possible corporate malefactions, inaccurate financial reporting or disability of the CEO.
Adding to the before mentioned costs resulting from tax avoidance engagement, Chen and Chu (2005) also find an efficiency loss in internal control for the firm. They discuss compensating CEOs ex ante for the risk of being detected engaging in illegal activities. They explain that contracting ex post incentives for these illegal tax avoidance activities is not considered as an option, because of the inability of the CEO to enforce such an illegal contract in court. Therefore, shareholders should compensate their CEO ex ante for the risk of being sentenced when detected. However, by paying ex ante incentives the motivation of the CEO will not be optimal anymore as the payment is no longer dependent on the outcome. So, this causes firms engaging in an illegal variant of tax avoidance to lose some extent of its control over its CEO, which is considered an efficiency loss in internal control. This corresponds with results found by Desai and Dharmapala (2009), suggesting that companies
14
with greater control over its CEO, thus having better corporate governance, benefit to a greater extent of tax avoidance.
Furthermore, as not only the shareholders of a firm face benefits and costs when engaging in tax avoidance, the same considerations exist for CEOs. As mentioned above, they benefit from this by receiving incentives, either monetary or non-monetary, from the shareholders, partly as a compensation for risk. However, according to Crocker and Slemrod (2005), costs also apply to CEOs, as they thus face a risk of being detected as well. Moreover, they argue that penalizing CEOs is more effective than penalizing the firm when contesting tax avoidance, because the same punishment is perceived more severe by the CEO, causing conflicts of interest with shareholders to increase. So, using agency theory, I conclude both shareholders and CEOs have to take into account all costs and benefits when deciding on engaging in tax avoidance.
Lastly, agency theory applied to corporate tax avoidance is not restricted to the relationship between the shareholders and the CEO, but also applies to lower layers of the organization. As to this, Dyreng et al. (2008) argue that CEOs most likely have no expertise with respect to taxes. Therefore, a CEO is expected not to be able to individually steer upon tax avoidance. However, they state that they are by setting the tone at the top. Bertrand and Schoar (2003) confirm this, as they mention that managers are capable of implementing their personal style into the company. With respect to this, a CEO can then incentivize the tax director and other divisional managers to engage in tax avoidance activities. As such, also an agency relation exists between the CEO and lower managers. This reasoning is confirmed by Phillips (2003), who does not find a direct relationship between the after-tax performance compensations awarded to CEOs and corporate tax avoidance. However, he does find such a relation for business unit managers in lower layers of the organization. Hence, this suggests an indirect relationship between after-tax performance compensation of CEOs and corporate tax avoidance does exist. Moreover, Phillips recognizes that these compensations of CEOs serve as an additional incentive on top of just job retention. As for this, CEOs on their turn compensate their business unit managers for after-tax income figures. Thus, this confirms that regarding tax avoidance, agency theory also applies to lower layers of the organization.
2.4. Prior research
When taking a first glance at prior literature, it stands out that evidence exists for systematic corporate tax avoidance (Dyreng et al., 2008). The average firm from the United States pays
15
an effective tax rate of 29.3 percent. Opposed to this, the official rate is set to be forty percent (KPMG, 2016). Nonetheless, 26.3 percent of the companies manages to effectively pay less than twenty percent, while 9.2 percent effectively pays less than ten percent.
In order to be able to give an explanation for these findings, research first focused on the relation between firm characteristics and corporate tax avoidance. Several publications report on such a relationship. For example, Lisowsky (2010) finds multiple firm characteristics which positively relate to tax avoidance, such as company size, magnitude of foreign income, the amount of subsidiaries in tax havens, profitability, inconsistent financial reporting and litigation losses. Lastly, he also names a negative relationship with leverage. These relations found by Lisowsky are consistent with Wilson (2009), who uncovers similar results for size, international operations and aggressive financial reporting. Besides this, Hanlon and Heitzman (2010) emphasize ownership as well. They argue that firms with a rather concentrated structure of ownership, such as family owned businesses, might engage in tax avoidance to a greater extent. As an explanation they bring up the owners being in full control and thus fully benefiting from the proceeds. This is in accordance with the agency theory described in the preceding section, as no agency problems exist in this situation. However, the opposite can also be argued: firms with concentrated ownership engaging in less tax avoidance. For this, the long-term vision of these owners can be brought up, as they care more about long-term costs such as a loss of reputation. Taking all this together, my first hypothesis is formulated as follows:
H1: Firm characteristics are related to the extent of corporate tax avoidance in which a company engages.
Then, as research on corporate tax avoidance continues to develop, focus also shifts to individual executives. While Dyreng et al. (2010) once more report similar relations as Lisowsky (2010) and Wilson (2009), they also show empirical evidence for individual executives having additional influence on the extent to which companies engage in corporate tax avoidance. As such, they find an eleven percent difference in effective tax rates when comparing the first and third quartile of their executives sample. This sample consists of CEOs, CFOs and other top executives as well. Finding significant results for all subsamples, the relation for CEOs appears to be the strongest.
Continuing, these findings contribute to earlier research looking at the relation between compensations based on after-tax performance measures and the amount of corporate
16
tax avoidance. For example, Armstrong, Blouin and Larcker (2012) show evidence that effective tax rates are significantly lower as a result of incentivizing the responsible tax managers. However, they add to this that this result does not imply that CEOs and CFOs have no influence on decision making regarding corporate tax avoidance. Nevertheless, the relationship for CEOs and CFOs is weaker, indicating that they delegate such responsibility to their tax managers. Also, Rego and Wilson (2009) demonstrate a positive relation between this kind of after-tax measures and the degree of tax aggressiveness initiated by CEOs and CFOs. As opposed to this, Phillips (2003) shows contradictory findings as he does not observe a direct relation between tax avoidance and incentives based on after-tax performance measures which are awarded to CEOs. Nevertheless, as business unit managers do engage in tax avoidance more severely when incentivized by after-tax performance measures, he does find an indirect relationship for CEOs. This assertion holds, because of the fact that CEOs hold responsibility for incentivizing these business unit managers. Again, the results from these three researches are in line with the agency theory elaborated upon in the preceding section, as conflicts of interest are resolved by making use of monetary incentives. This not only applies for the relation between shareholders and CEOs, but also for lower managers, as CEOs are able to set the tone at the top.
Finally, Dyreng et al. (2010) were the first to also do research on the extent to which the characteristics of individual executives can serve as a predictor for whether a firm will engage in tax avoidance or not. Looking at several simple biographical characteristics such as age, gender and education, they find some evidence for a relation with the amount of tax avoidance a firm engages in. This appears to be in accordance with earlier research conducted by Bertrand and Schoar (2003), who find that management style has its foundation in the fixed characteristics of the respective manager, such as age, gender and educational background. As such corporate decisions tend to be more conservative in cases of older CEOs. Also, CEOs holding an MBA degree tend to take more aggressive decisions. Despite of these CEOs affecting the behaviour and performance of the corporation, the evidence from Dyreng et al. (2010) appears to be little. This is mainly caused by missing data. However, as the above results clearly do indicate a relation between tax avoidance and biographical CEO characteristics, I express the following hypothesis:
H2: Individual biographical executive characteristics are related to the extent of corporate tax avoidance in which a company engages.
17
2.5. Individual cultural characteristics
Taken together, the text above explains that CEOs are to a great extent responsible for taking decisions on whether or not to engage in corporate tax avoidance. According to agency theory, they perform a cost-benefit analysis, taking into account the relevant incentives they are awarded by the shareholders. Besides that, on their turn they motivate lower managers of the company to engage in tax avoidance activities as well, for example by imposing their management styles on the company by setting the tone at the top and awarding incentives based on after-tax performance measures. Furthermore, research also confirms this theory as there indeed exists a positive relation between corporate tax avoidance and executive compensation. On top of that, literature shows an incremental influence of executives on tax avoidance. Nevertheless, it thus far fails to address the individual characteristics of the executives which play an important role when taking these decisions.
However, Bertrand and Schoar (2003) do conclude that top executives show several differences in their style of management, which consequently translates in corporate decisions. They also draw a parallel with agency theory, stating that shareholders are more likely hiring CEOs which will implement a value enhancing management style, for example through tax avoidance activities. On top of this, according to the theory of Allingham and Sandmo (1972) on individual taxes, the decision to engage in tax avoidance is one which is based on multiple factors, such as the magnitude of benefit, the probability of detection, the magnitude of corresponding penalties, risk-aversion and also moral considerations such as civic duty. Adding to this, Hanlon and Heitzman (2010) state that these factors are also taken into consideration for corporate taxes. As agency theory applies, the decision on corporate tax avoidance is therefore also based on the individual considerations of the responsible CEO (Slemrod, 2004).
Continuing, Bertrand and Schoar (2003) explain that these individual considerations appear in the heterogeneous ways managers act within a firm. Managers not only tend to differ on skill levels, but they also differ on risk-aversion, personal preferences and opinions. Mainly these last two properties are consistent with Kirchler et al. (2001) finding that moral concerns play an important role in how individuals judge tax avoidance. So taking the considerations from the above stated researches together, such as civic duty, personal preferences, moral concerns and personal opinions, this very well connects to research on cultural differences between managers conducted by Hofstede (1980; 1983; 1984), as they are all subject to cultural background. This can be explained by looking at the definition of
18
culture (Hofstede, 1984). Culture can be described as a mind-set which determines the way of thinking of a certain group of people, for example a country. It is a phenomenon which transfers from generation to generation and it defines how people think of their lives, of the world and how they interact which each other. This then results in the development of certain values and beliefs, which enables people to draw lines between good and bad or true and false.
As Hofstede (1980) mentions, heterogeneities in management style, as well as setting the tone at the top for the corporation and motivating and incentivizing lower layers of management, are all dependent on differences in culture. Furthermore, consistent with this statement, he also claims in later research (1984) that people, thus the primarily responsible CEOs, shape corporations by taking into account their values. Therefore, a certain form of management style or philosophy which fits a CEO from one culture, does not have to fit a CEO from another culture (Hofstede, 1984). Hence, in order to gain an understanding of which cultural characteristics cause managers to exercise management styles involving corporate tax avoidance activities, the following subsections elaborate on the four cultural dimensions on which managers differ according to Hofstede (1980), being power distance, individualism, masculinity and uncertainty avoidance.
2.5.1. Power Distance
To start with the first cultural dimension of the framework formulated by Hofstede (1980), called power distance, it is important to underscore that different cultures are placed along a continuum instead of being labelled either a large or small power distance culture. This concept then ranks cultures to the extent to which both more and less powerful people tend to feel comfortable with the distribution of power in organizations, that is either equal or unequal (Hofstede, 1984). Hence, cultures which contain much of hierarchy where this is not subject to any debate are placed along the large power distance side of the continuum. Contrary to this, a culture is positioned along the small power distance side if it satisfies the requirement of having relative equality regarding power. Therefore, this cultural dimension can be of great influence for the structure of a company.
Furthermore, Hofstede (1983) explains that the power distance score is dependent on several aspects. Firstly, cultures where managers make decisions by themselves, without consulting their employees, are perceived as having large power distance. Secondly, the extent to which employees tend to prefer this kind of behaviour also positively relates to the
19
power distance score. And lastly, large power distance cultures also are denoted by employees experiencing fear when disagreeing with their managers. Applying the agency theory as explained earlier to this measure and also taking into account prior research on tax avoidance, CEOs from a high power distance culture are more likely to agree with everything demanded by their principal, in this case the shareholders. As such, when dealing with a high power distance CEO, I expect the company to engage in corporate tax avoidance to a greater extent, as this is assumed to be the desire of the shareholders. I therefore hypothesize:
H3a: The power distance score of the culture to which a CEO belongs is positively related to the extent of corporate tax avoidance in which the company he or she works for engages.
2.5.2. Individualism
Secondly, Hofstede (1980) formulates the extent to which a culture is individualistic or collectivistic as his second cultural dimension. Again, different cultures are positioned along a continuum and thus are not either purely individualistic or purely collectivistic. He then explains that cultures on the individualistic side of the continuum contain people who prefer to be responsible for themselves and their closest family (1984). On the contrary, people from cultures which score relatively low on individualism, that is cultures which are collectivistic, tend to prefer society to take care of them. However, this is a reciprocal process, as they will also take care of others within the society. Therefore, a high score on individualism implies relative little interdependence between the people that are part of the society.
The before mentioned also applies to the relationship people have with companies (Hofstede, 1983). As such, people from individualistic cultures are likely to act as an independent agent, while people from collectivistic cultures tend to feel more comfortable when they can depend to a greater extent to the firm they work for. This includes getting training, working under excellent conditions and receiving additional fringe benefits. Hence, individualism results in attaching little value to these aspects, while greatly appreciating personal spare time, freedom and experiencing challenges in work. Putting this together and taking into consideration that engaging in tax avoidance is incentivized upon by the shareholders according to the agency theory, I expect CEOs to engage in more tax avoidance when ranked relatively high on individualism. The main reason for this is an individualistic CEO being an independent agent and attaching much value to working on challenges like
20
after-tax performance measures. Moreover, individualism causes CEOs to prioritize looking after themselves and their families first instead of society as a whole, implicating that the purpose of taxes, being common welfare, forms less of a constraint when deciding upon engagement in tax avoidance. As for this, my hypothesis on individualism is as follows:
H3b: The individualism score of the culture to which a CEO belongs is positively related to the extent of corporate tax avoidance in which the company he or she works for engages.
2.5.3. Masculinity
Thirdly, Hofstede (1980) defines his next dimension as the masculinity within a culture. Again formulating a continuum, cultures are therefore either placed along the masculinity side of the continuum, or they are placed along the opposite feminine side. He then describes these masculine cultures as societies in which materialism, performance, assertiveness and bravery are important factors (1984). On the opposite however, more feminine cultures tend to appreciate relationships, taking care of each other, unpretentiousness and quality of life. As these concepts are based on traditional task distribution, it describes how a certain society assigns social tasks to both men and women. For example, masculine cultures prefer to draw a clear line between both sexes by assigning masculine tasks to the men and feminine tasks to the women. Contrary to this, feminine cultures tend to ignore this traditional distribution by also assigning feminine tasks to men and masculine tasks to women. It should therefore be clear that distinguishing between masculinity and femininity differs from distinguishing men and women. A man can still be described feminine, while a woman can be described masculine.
Likewise, Hofstede (1983) explains that when measuring masculinity, cultures in which recognition, the quantity of the earnings, promotion and challenging work are considered important, will receive a higher rating. Also contributing to a high score is attaching relative little value to good relationships with superiors, job security, cooperating with others and living or working in an area of desire. Again, when applying agency theory and the incentivizing of CEOs through after-tax performance measures to these properties, it is possible to set up a hypothesis on the relationship between masculinity and corporate tax avoidance. In this case however, the direction of this relationship is not clear, as CEOs from masculine cultures are expected to appreciate recognition and both monetary and
non-21
monetary incentives, but appear to attach little value to a good cooperation with the shareholders. Furthermore, attaching little value to this also becomes evident in not caring much about job security. As a result, the earlier stated implicit motivation of job retention to engage in corporate tax avoidance deteriorates (Phillips, 2003). Hence, this contradictory reasoning causes the direction of the relationship between masculinity and corporate tax avoidance to remain unclear. As such, I am forced to formulate a two sided hypothesis:
H3c: The masculinity score of the culture to which a CEO belongs is related, either positively or negatively, to the extent of corporate tax avoidance in which the company he or she works for engages.
2.5.4. Uncertainty avoidance
Finally, the fourth and last cultural dimension Hofstede (1980) formulates is uncertainty avoidance, assigning cultures with a strong tendency to avoid uncertainties a high ranking, while assigning low rankings to cultures which have a weaker tendency with regard to this. Thus, different cultures are again positioned along a continuum. This uncertainty avoidance continuum is then defined as the extent to which people within a culture feel uneasy when experiencing uncertainties and ambiguities (Hofstede, 1984). Certain feelings then result in creating certainties and maintaining the status quo, therefore having implications on how people tend to structure organizations. A strong uncertainty avoidance culture also implies people to be suspicious regarding progressive and innovative initiatives. Contrary to that, certain initiatives are more accepted when uncertainty avoidance is weak. Also, people from such cultures tend to be better able to accept the future as being unknown and to accept that the present will turn into history in considerable time.
Furthermore, Hofstede (1983) explains that the uncertainty avoidance measure is dependent on three factors. Firstly, a culture is placed along the strong uncertainty avoidance side of the continuum when its people indicate that they expect to be working for a long period of time for the same firm. In this case, he defines a long period of time as longer than five years. Secondly, unconditionally believing that the rules set by the firm should not be violated by its employees also leads to a high score on uncertainty avoidance. And thirdly, when people point out that they often feel nervous, tense or anxious at work, this also contributes to a higher uncertainty avoidance score. Moreover, this anxiety can be increased by any behaviour mentioned before, such as considering quitting one’s job or violating
22
company rules. Once more, when applying agency theory to this measure, it should be clear that CEOs from strong uncertainty avoidance cultures are expected to be more likely to engage in corporate tax avoidance. This is the case, as this will be demanded by the shareholders of the company. As such, the relevant CEO believes he or she should not violate this demand or rule and is also not likely to consider leaving the company after such a request. Therefore, the only option remaining is engaging in corporate tax avoidance. As mentioned earlier, this is also a result of avoiding such behaviour as it increases feelings of anxiousness at work. So, taken together, regarding the relationship between uncertainty avoidance and corporate tax avoidance, I hypothesize the following:
H3d: The uncertainty avoidance score of the culture to which a CEO belongs is positively related to the extent of corporate tax avoidance in which the company he or she works for engages.
2.6. Summary
So, summarizing, when looking at the literature on corporate tax avoidance, it appears that still no general agreement exists regarding the definition and the measuring. By reading through several researches however, it can be stated that tax avoidance contains both legal and illegal forms as it can be defined as every practice that reduces the effective tax rate paid by a company. Also regarding the measurement exists no single solution, but the effective tax rate based on financial reporting appears to be the best option when doing research on CEO characteristics. Furthermore, corporate tax avoidance has several consequences for multiple stakeholders. As such, governments suffer losses of income, while companies and its shareholders benefit from a higher after-tax profit. It has to be stressed however, that companies also have to keep in mind possible disadvantages such as the risk of being detected and penalized. According to the agency theory, with managers as the agents and shareholders as the principal, the same goes for CEOs and other lower layers of management. As to this, they also benefit from engaging in corporate tax avoidance when compensated for after-tax performance measures, but they are subject to the possibility of detection and punishment as well. Continuing, prior research indeed finds that companies engage in corporate tax avoidance. This phenomenon can be partly explained by looking at firm characteristics. Also, when applying agency theory, a positive relationship with incentivizing CEOs and other managers based on after-tax performance becomes clear. Then, later research found the
23
individual CEO to be of incremental significant influence as well. However, evidence on simple biographical characteristics is still weak. As a result, only evidence for a general relationship remains. Lastly, the decision of the CEO on whether or not to engage in corporate tax avoidance appears to be dependent on the management style, which comes forth out of cultural background. Therefore, four cultural managerial characteristics are expected to affect this management style, being power distance, individualism, masculinity and uncertainty avoidance.
3. Methodology and descriptive statistics
As I stated earlier in this text, I address the question whether CEO characteristics are related to corporate tax avoidance by conducting a quantitative research in which I make use of several publicly available databases, being Bureau van Dijk AMADEUS Managers, Bureau van Dijk AMADEUS Financials and the Hofstede database on cultural dimensions (Hofstede & Hofstede, 2014). Firstly using Bureau van Dijk AMADEUS Managers yielded an initial sample of 57,143 CEOs. However, further developing the dataset by matching these CEOs to the tax data of the companies they work for, adding the data on the four cultural dimensions and adding several control variables results in a loss of observations which leaves a total sample size of 21,783. Table 1 gives a more thorough elaboration on this process. Based on this data collection process, this chapter provides an explanation on how the dataset is constructed, which variables are included and how these variables are determined. Besides this textual part, I also provide relevant descriptive statistics on the data. Beginning in the next section, I elaborate on the CEOs within the dataset, after which I cover the data on the four cultural dimensions. Thirdly, the corporate data, both financial and non-financial, is explained. Thus, this also includes the corporate tax data regarding the measurement of the extent of tax avoidance. Concluding, this chapter also provides information on multicollinearity, after which I provide the linear OLS regression model I use to analyse the data.
24
TABLE 1
Data collection and sample size development
Data Sample Size
Valid data on Chief Executive Officers 57,143
Loss of observations due to invalid tax data (24,669)
32,474 Loss of observations due to negative GAAP income before taxes (8,999) 23,475
Loss of observations due to missing cultural data (19)
23,456
Loss of observations due to invalid control data (23)
23,433
Loss of observations due to deleting outliers (1,650)
Total sample size (n) 21,783
3.1. Data on Chief Executive Officers
When retrieving data on managers from Bureau van Dijk AMADEUS Managers, several considerations are of particular importance. Firstly, when taking into account prior research on corporate tax avoidance, it becomes evident that several top managers, such as CEOs, CFOs, tax managers et cetera, have an influence on the decisions regarding engagement in corporate tax avoidance (Armstrong et al., 2012; Dyreng et al., 2010; Phillips, 2003; Rego & Wilson, 2009). However, Dyreng et al. (2010) are the only ones who also have reported on the influence of individual executive characteristics on corporate tax avoidance decisions. While results are consistent for CEOs, CFOs and other executives, they find stronger results when looking at CEOs separately. Therefore, as I am also doing research on the relation between corporate tax avoidance and individual executive characteristics and as I am the first one to the best of my knowledge who is also taking into account cultural characteristics, I decide to lay particular focus on CEOs in this research. Moreover, based on the application of agency theory on this subject as explained in the preceding chapter, I have theoretical evidence that this stronger effect for CEOs holds. Therefore, as Bureau van Dijk AMADEUS Managers contains a strongly varying set of managers, I initially filtered down the dataset on CEOs, also retrieving their full names. On top of that, I am also in possession of identification codes and names of the companies they work for, in order to be able to match other variables to this data.
25
Furthermore, a second consideration exists in biographical information regarding the CEOs. Building upon previous research on corporate tax avoidance conducted by Dyreng et al. (2010), it appears that age, education and gender are taken into consideration regarding biographical characteristics. This contributes to earlier research by Bertrand and Schoar (2003), who found the age and education of managers to be of significant influence on financial corporate decisions: older managers apparently act more conservative, while on average managers with a higher educational degree act more aggressive. Hence, building upon these results, I filter the set of CEOs, remaining only with observations having valid data on gender and date of birth. Unfortunately, the database does not include information regarding educational degrees. Therefore, I create a dummy variable GENDER, assigning a value of 1 if male, and 0 if female, and also a variable called BIRTH_YEAR, containing the year of birth of the relevant CEO. Concluding, Table 2 and Table 3 give an overview of the
TABLE 2 CEO gender distribution
Gender Number of observations Percentage
Male 17,413 79.94%
Female 4,370 20.06%
Total (n) 21,783 100%
TABLE 3
CEO year of birth distribution (in decades)
Decade Number of observations Percentage
1910-1919 1 0.00% 1920-1929 74 0.34% 1930-1939 749 3.44% 1940-1949 2,716 12.47% 1950-1959 4,845 22.24% 1960-1969 7,499 34.43% 1970-1979 4,779 21.94% 1980-1989 1,071 4.92% 1990-1999 49 0.22% Total (n) 21,783 100% Mean 1961.88 Standard Deviation 11.67 Minimum 1917 Maximum 1994 1st Quartile 1954 Median 1963 3rd Quartile 1970
26
distributions regarding respectively gender and year of birth, the latter sorted by decade. Table 4 also provides a summary of all variables used in this research, thus including all variables on which I elaborate later in this text as well.
TABLE 4 Variable definitions Variable Definition
TAX_AVOIDANCE Corporate tax rate minus the effective tax rate, the latter being calculated by dividing tax expenses by pre-tax GAAP income
PDI Power distance measure on a 0 to 100 scale as defined by Hofstede (1983)
IDV Individualism measure on a 0 to 100 scale as defined by Hofstede (1983)
MAS Masculinity measure on a 0 to 100 scale as defined by Hofstede (1983)
UAI Uncertainty avoidance measure on a 0 to 100 scale as defined by Hofstede (1983)
GENDER Assigned a value of 1 if the CEO is male, assigned a value of 0 if the CEO is female
BIRTH_YEAR Year of birth of the CEO
TITLE_YEARS The number of years a CEO carries his or her title, based on the date of appointment
VERY_LARGE Assigned a value of 1 if the company satisfies the Very Large category as defined by Bureau van Dijk AMADEUS Managers and assigned a value of 0 otherwise
LARGE Assigned a value of 1 if the company satisfies the Large category as defined by Bureau
van Dijk AMADEUS Managers and assigned a value of 0 otherwise
MEDIUM Assigned a value of 1 if the company satisfies the Medium Sized category as defined by
Bureau van Dijk AMADEUS Managers and assigned a value of 0 otherwise
SMALL Assigned a value of 1 if the company satisfies the Small category as defined by Bureau
van Dijk AMADEUS Managers and assigned a value of 0 otherwise
LN_SIZE The natural logarithm of total assets
INCORP_YEAR The year of incorporation of a firm
OWNERSHIP Assigned a value of 1 if a firm is publicly quoted and assigned a value of 0 if it is not
LEGAL_STAT Assigned a value of 1 if a firm carries an active status and assigned a value of 0 if a firm carries an active status but being in default of payments
ROA Pre-tax GAAP income divided by total assets
LEVERAGE A ratio calculated by dividing total liabilities by total assets
PP&E Property, plants and equipment measured by total tangible fixed assets
INTANGIBLES A ratio calculated by dividing total intangible fixed assets by total assets
Then, a third consideration relates to the date on which a CEO receives his title within the company. This is of particular interest for this research, as I match the data on CEOs to the tax data of the corporations they work for. Therefore, CEO observations are only valid when they are already employed by the company during the entire year for which the tax data are
27
calculated. For example, as the tax data are from 2014, on which I elaborate later in this chapter, CEOs are required to having received their title on the December 31st 2013 or earlier. This is the case, because when a manager is appointed somewhere throughout 2014, it is not possible to determine whether the tax data are a result of the new CEO’s policy, or of his or her predecessor. Also should it be clear that CEOs appointed in 2015 or 2016 are unsuitable for this research as well.
On top of this, besides filtering the observations for CEOs who are appointed at December 31st 2013 at the latest, I also use the time of CEOs having their title within the firm as a control variable, as prior research indicates this is an important factor when looking at corporate decision making (Bertrand & Schoar, 2003; Dyreng et al., 2010). As for this, literature suggests that being employed for a longer time enables a CEO to have greater influence on the company’s policy. Therefore, as agency theory suggests that CEOs are incentivized to engage in corporate tax avoidance, I expect CEOs who carry their title for a longer time are more likely to be able to implement these tax avoidance activities in the company. Hence, I create a variable TITLE_YEARS, based on the date of appointment, in which I determine the number of years, measured in integers, for which the CEOs carry their title up until the start of 2014. Table 5 gives an overview on the distribution of this variable, sorted within ranges of 3 years.
Lastly, a fourth consideration regarding CEO observations is valid nationality data. As mentioned earlier, it is possible to attach values to individual CEOs for power distance,
TABLE 5
CEO title years distribution
Decade Number of observations Percentage
0-2 years 9,198 42.23% 3-5 years 4,291 19.70% 6-8 years 3,245 14.90% 9-11 years 2,181 10.01% 12-14 years 1,438 6.60% 15 years or longer 1,430 6.56% Total (n) 21,783 100% Mean 5.09 Standard Deviation 5.23 Minimum 0 Maximum 34 1st Quartile 1 Median 4 3rd Quartile 8
28
masculinity, individualism and uncertainty avoidance, because differences in these characteristics come forth out of different cultures. However, in order to determine the cultural background of an individual CEO, it is necessary to determine its nationality, as the Hofstede dataset uses different nationalities to attach values to these four cultural dimensions (Hofstede & Hofstede, 2014). Therefore, I also filtered the CEO observations for valid data on nationality. However, as can be made up from Table 1, some observations are lost due to missing cultural values. As for this, despite that the Hofstede dataset contains cultural values for most nationalities, data on some countries, mostly underdeveloped ones, is unavailable. In this case, the nineteen lost observations were caused by data lacking for Albania, Benin, Bosnia and Herzegovina, Burundi, Egypt, Eritrea, Jordan, Lebanon, Libya, Liechtenstein, Moldova and Ukraine. Therefore, Table 6 displays the distribution of the nationalities of the remaining CEOs. From this, it can be seen that a vast majority of the observations is Italian, which is caused by limited availability of tax data for companies from other countries than Italy. On this, I elaborate further later in this chapter.
3.2. Data on the four cultural dimensions
As can be made up from Table 6, it is possible by using the Hofstede dataset to attach values to different nationalities for the four cultural dimensions as formulated by Hofstede (1980). This then enables me to match these values with the data observations for CEOs as discussed in the preceding section. However, these values can be difficult to interpret, which can then result in raising several questions about the way of measurement. Therefore, this section elaborates on the measurement technique as explained by Hofstede (1984).
As a start, all values for these four cultural dimensions are derived from a survey based research conducted by handing out questionnaires to individual employees working in a large variety of countries (Hofstede, 1983). These questionnaires contain about 150 questions and are provided in a varying set of languages, ensuring them to be understandable. In order to extract cultural differences between countries, answers on a specific set of those questions are used. These questions focused on the values to which people live and on the way people view the world they live in.
For instance, the power distance variable, labelled PDI, is determined by adding the percentage of participants who indicate that their superior decides in an autocratically and in a persuasive manner to a value of 135 (Hofstede, 1983). Subsequently, the degree to which participants are afraid of disagreeing with their managers, measured by a one to five scale and
29
TABLE 6
CEO nationalities and cultural values
Nationality Number of Observations PDI IDV MAS UAI
Argentina 2 49 46 56 86 Australia 6 36 90 61 51 Austria 20 11 55 79 70 Belgium 8 65 75 54 94 Brazil 2 69 38 49 76 Bulgaria 1 70 30 40 85 Canada 1 39 80 52 48 China 7 80 20 66 30 Czech Republic 1 57 58 57 74 Denmark 24 18 74 16 23 Finland 5 33 63 26 59 France 75 68 71 43 86 Germany 62 35 67 66 65 Greece 3 60 35 57 100 Ireland 5 28 70 68 35 Israel 12 13 54 47 81 Italy 21,393 50 76 70 75 Japan 2 54 46 95 92 Luxembourg 2 40 60 50 70 Malaysia 2 100 26 50 36 Netherlands 8 38 80 14 53 New Zealand 1 22 79 58 49 Norway 9 31 69 8 50 Peru 1 64 16 42 87 Portugal 3 63 27 31 99 Romania 1 90 30 42 90 Russia 2 93 39 36 95 Slovenia 1 71 27 19 88 Spain 16 57 51 42 86 Sweden 10 31 71 5 29 Switzerland 40 34 68 70 58 Taiwan 1 58 17 45 69 Turkey 3 66 37 45 85 United Kingdom 30 35 89 66 35 United States 21 40 91 62 46 Uruguay 1 61 36 38 98 Venezuela 2 81 12 73 76 Total (n) 21,783
30
multiplied by 25, is subtracted, after which the percentage of participants who indicate not preferring their manager to consult them is added. Despite having theoretically possible values of -90 and 210, all values found fit in a range of 11 till 104. However, these values are then rescaled between a 0 to 100 range (Hofstede & Hofstede, 2014). Furthermore, the other three cultural dimensions, being individualism labelled as IDV, masculinity labelled as MAS and uncertainty avoidance labelled as UAI, are measured in a similar way, using questionnaire items relating to the factors already explained in the preceding chapter. As a result, IDV renders values between 6 and 91, MAS fits within a range of 5 and 95, while measuring UAI delivers values between 8 and 112. Again, the ranges of these three cultural dimensions are rescaled between a 0 to 100 scale. As a result, all four dimensions generate values within the same scale.
Concluding, when taking a look at these cultural values for the observed nationalities displayed in Table 6, it stands out that Austria scores lowest on PDI with 11, while Malaysia is assigned the highest possible value of 100. Besides that, Scandinavian countries, being Denmark, Sweden, Norway and Finland obtain relatively low values: respectively 18, 31, 31 and 33. On the other side however, non-European countries such as Malaysia, Venezuela, China and Brazil, and also countries from Eastern Europe such as Russia, Romania, Slovenia and Bulgaria, appear to score highest on PDI. Furthermore, with respect to IDV, Venezuela appears to be ranked lowest with 12 and the United States are ranked highest with 91. Another notable fact exists in developed Western countries scoring particularly high on this cultural dimension, while less developed countries such as Venezuela, Peru and Taiwan tend to obtain lower values. Then turning to MAS, the lowest value can be found at Swedish people, scoring a value of 5. Again, the other Scandinavian countries as well as the Netherlands obtain comparable ratings. On the other end however, with a value of 95, Japan ranks highest. Other high scores can be found in countries surrounding the Alps, such as Austria, Switzerland and Italy. Finally, when observing the values relating to UAI, it can be concluded that Western countries, with China as an exception, receive low values. In particular, Scandinavian countries relatively tend to score lowest again, with Denmark getting rating the lowest with 23 and Sweden following up with 29. On the other side however, a varying set of countries obtain the highest scores, with Greece, Portugal and Uruguay respectively having UAI values of 100, 99 and 98.
