Table of Contents
List of Tables
List of Appendices
List of Abbreviations
2 Literature Review and Hypotheses Development
3 Empirical Approach and Data
3.2 Regression Models
3.2.1 Baseline Regression
3.2.2 Regression with Country Characteristics
3.3 Sample Selection and Descriptive Statistics
4 Empirical Results
4.1 Baseline Results
4.2 Results with Country Characteristics
4.3 Additional Analyses
5 Robustness Test
List of Tables
Table 1: Descriptive statistics
Table 2: Short- and long-term tax avoidance
Table 3: Tax avoidance and statutory tax rate
Table 4: Tax avoidance and book-tax conformity
Table 5: Additional results for tax avoidance and statutory tax rate
Table 6: Additional results for tax avoidance and book-tax conformity
Table 7: Long-term tax avoidance and country characteristics
Table 8: Additional results for long-term tax avoidance and statutory tax rate
Table 9: Additional results for long-term tax avoidance and book-tax conformity
List of Appendices
Appendix 1: Variable Definitions
Appendix 2: Descriptive Statistics Country-level
Appendix 3: Distribution of Tax Avoidance
List of Abbreviations
Abbildung in dieser Leseprobe nicht enthalten
On January 22, 2019, the German political party “Bündnis 90/Die Grünen” published a study about the extent of tax avoidance in the European Union, revealing a gap between the effective tax rate paid by companies and the statutory tax rate in their home country. This gap between statutory tax rate and effective tax rate is one way of considering tax avoidance (Atwood, Drake, Myers & Myers, 2012). There are other definitions as well1 (Gebhart, 2017; Guenther, 2014; Salihu, Obid & Annuar, 2013) but overall tax avoidance can be seen as not paying the share of pre-tax income as taxes as intended by the government. In other words, tax avoidance is the attempt to legally reduce the tax burden paid to the government as much as possible. When talking about tax avoidance, it is imperative to differentiate it from the illegal form called tax evasion. Tax avoidance refers to the exploitation of discontinuities of the tax law while having an “economic substance” or a “business purpose” (Lisowsky, 2010), whereas tax evasion refers to the illegal use of these tax shelters2 by for instance not disclosing assets hidden abroad, also considered as offshore tax evasion in tax havens (Alstadsaeter, Johannesen & Zucman, 2018; Johannesen & Zucman, 2014).3
These legal tax shelters share the basic idea to increase the after-tax profit of a firm by efficiently reducing its tax burden paid to the government through earnings management or “state tax planning” (Gupta & Mills, 2002; Klassen & Shackelford, 1998). The Department of the Treasury (1999) formulates three methods of reducing corporate income tax liabilities: (1) excluding/deducting income from taxation transactions, (2) deferring income or losses into a later period, or (3) converting income into a different, lower-taxed form or jurisdiction.
The first refers to the idea of exploiting the deduction of several tax shelters (Graham & Tucker, 2006). For example, Germany’s thin capitalization rules state that interest expenses are deductible to the extent of interest income, and the remaining net interest expenses are “deductible only up to 30 percent of tax Earnings Before Interest, Taxes, Depreciation, and Amortization” (EU Tax Centre, 2015).4 The second method consists of the process of loss carry forwards or accruals management. For example, firms making a loss in a certain period can postpone this loss to a certain extent to a later period to offset their taxable income in that period, subsequently reducing their tax burden (Langenmayr & Lester, 2017). The third mechanism of reducing the taxes paid in a multinational company is to strategically choose the best suitable country for those transactions in order to leverage their difference in statutory tax rates. By shifting profit to unprofitable firms (De Simone, Klassen & Seidman, 2017)5 or low tax rate countries, also called tax havens (Dyreng & Lindsey, 2009; Dyreng, Lindsey & Thornock, 2013) or shifting debt to high tax rate jurisdictions6 the global effective tax rate (ETR) can be reduced.
In context with examining the mechanisms how entities avoid taxes, another critical issue that comes up is the question why companies do not pay the required share of taxes but rather put a lot of effort in complicated tax avoidance strategies and simultaneously take the risk of sliding into tax evasion. Generally, taxes have always been an important mechanism to save money as they decrease the profit of a firm and as a result further financial numbers (Wilde & Wilson, 2018). Since managers are more or less incentivized to maximize the after-tax profit rather than just minimizing the taxable income to pay fewer taxes (Dyreng, Jacob, Jiang & Müller, 2019), tax departments are often considered profit centers aiming to efficiently handle the trade-off between reducing the tax base and maximizing after-tax profits (Robinson, Sikes & Weaver, 2010) and thereby balancing the cost and benefits of tax avoidance. The extent of the managers’ incentive and subsequently the firm’s aggressiveness in tax avoidance decisions depends on several components. For instance, the ownership structure of a company, the extent to which the compensation of a manager is connected to the equity of a company, the degree of taxes ultimately taken by the shareholders (Dyreng et al., 2019), and potential reputational concerns encourage (discourage) firms to pursue an aggressive tax avoidance strategy.
The high numbers of articles in top-line journal articles regarding taxes show the prevalence of tax research (Wilde & Wilson, 2018)7, particularly focussing on tax avoidance. Broadly investigated areas in the literature are the above-mentioned methods of tax avoidance, factors that impact its degree, and the historical development of tax avoidance. This thesis contributes to the large literature of drivers of tax avoidance, for example tax system characteristics (Atwood et al., 2012) or the degree of tax incidence (Dyreng et al., 2019). By further expanding their, mostly on an American sample based findings, to an international context including 55 countries, I am investigating which kind of firms drive the discovered difference in tax avoidance between countries. Thereby, I try to answer the question which kind of firms benefit the most from the above-mentioned tax mechanisms and thus use them the most. By doing so, I focus on firm characteristics that are not part of the tax system itself, but rather are common characteristics associated with a difference in tax avoidance. Moreover, I complement my research by investigating some tax related country characteristics to answer the question whether the observed firm-tax avoidance correlations differ depending on certain country-specific factors. I therewith further complete the explanation for the difference between the degree of tax avoidance in different countries. By investigating the implications of known firm-characteristic correlations in an international setting, I am refining findings of previous literature.
This thesis can serve in politics as an indicator regarding which companies drive the degree of tax avoidance in a respective country compared to other nations. Furthermore, the results can be used in the ongoing debate about efficient taxation by giving additional insight into which firms avoid the most taxes and thus could be taxed differently.8 It may provide further studies a starting point on which they can build international, country-level tax avoidance comparison. For example, Thomsen and Watrin (2018) could use my findings to further elaborate their comparison of country tax avoidance on an even more international level. Additionally, this paper could be interesting for investors because, as pointed out by prior research, knowing which kind of firms avoid the most taxes can be a good estimator about their current and future earnings and thus could be more attractive to invest in (Atwood et al., 2012; Hanlon & Heitzman, 2010).
The remainder of this paper will be structured as followed: Section 2 introduces the specific research question and related literature in a more detailed way and leads to the development of the research hypotheses. Section 3 describes the empirical approach, including the measures and models used, and gives some descriptive statistics, followed by empirical results in Section 4. A robustness test in Section 5 and a short conclusion in Section 6 conclude this thesis.
2 Literature Review and Hypotheses Development
Going back to the study from the beginning, the “Bündnis 90/Die Grünen” party states that the majority of big companies does not pay the statutory tax rate. They point out that “the bigger the company, the lower the effective tax rate is”9, indicating that there are certain firm characteristics that drive the different degree of tax avoidance on a country-level.
Consistent with the explanations in Section 1 about the different tax mechanisms, firms with a higher leverage ratio pay fewer taxes since, by taking the corporation’s level of the effective tax rate as an indicator for tax avoidance, they have a higher portion of tax deductible interest expenses, subsequently reducing the global Cash ETR10 (Atwood et al., 2012; Dyreng et al., 2019). Depending on the measure of tax avoidance other literature find contrary relations suggesting that firms with less leverage engage more in tax shelters because they do not sufficiently benefit from the reduction in the taxable income through interest expenses and are consequently more incentivized to find an alternative way of reducing their tax burden (Lisowsky 2010; Wilson, 2009).11 This is supported by DeVito and Jacob (2018), who state that an increase in leverage and thus an increase in deductible interest expense decreases the marginal benefit of tax avoidance. Again, this shows the importance of understanding the measurement of tax avoidance and the rationale applied by the researcher.
Following a similar line of argumentation, the literature suggests that firms relying more on accrual-based earnings management avoid more taxes than firms not relying on this accounting practice. DeVito, Jacob & Müller (2019) indicate that an increase in reported accruals results in a decrease in the GAAP ETR, corresponding to an increase in tax avoidance. The significance of this finding is even persistent to other measures of tax avoidance. On first sight, the opposite coefficient sign of the accrual’s component variable in the regression of Atwood et al. (2012) seem to contradict DeVito et al. (2019), but by recognizing that Atwood et al. (2012) use a different measure of tax avoidance. this finding turns out to be in line with DeVito et al. (2019). By defining tax avoidance as the ability to pay a smaller effective tax rate than the one determined by the government, an increase in accrued earnings increases this difference (ceteris paribus) and thus the degree of tax avoidance. The reason for this is that accruals are i.a. also considered as tax deductible (DeVito et al., 2019) and reduce the tax burden.
As explained in Section 1, in many countries, entities are allowed to postpone their losses to a later period which enables them to offset their tax burden to a certain extent. In Germany, for example, this loss offset is possible to the amount of one million euros, and beyond this, 60% of profits that are above one million euros can additionally be offset from the tax burden. Hence, it is not surprising that many researchers expect a positive relation between tax loss carryforwards and the degree of tax avoidance, even though there is often no significant relation observable (Graham, Hamlon, Shevlin & Shroff, 2014; Lisowsky, 2010). However, the significance does exist as proven by Dyreng et al. (2019). They find that firms having a tax loss carryforward have a 4.4 percentage points smaller Cash ETR than other firms. This corresponds to a decrease in Cash ETR of 15.7%12 compared to the sample mean.
These findings are more or less straight-forward since all described characteristics are directly related to the tax mechanism itself.13 They all incorporate a component of determining the taxable income, concurrently the tax burden, and thus directly influence the degree of tax avoidance. However, there are further firm characteristics that are not part of the tax calculation directly, but rather of describing nature, characterizing the kind of entities that pursue more tax avoidance. In prior literature, they are often just used as control variables, still showing that they have an impact on tax avoidance. For my analyses, three firm attributes are selected that often function as control variables in other regressions. Thereby, an implication-based approach has been used, meaning that I investigate conclusions that I draw from the findings of prior research and apply them in an international context.
One frequently used control variable within the regression of tax avoidance is the sales growth of a company.14 Several papers indicate that there is a significant positive correlation between the growth of a company and its tax avoidance aggressiveness. For instance, Graham et al. (2014) find that an increase in the sales growth of a company by its sample standard deviation of 0.223 leads to a decrease of 2.45 percentage points15 in the three-years cash ETR, being tantamount to an increase in tax avoidance. Supplemented by the findings of Dyreng et al. (2019) and Atwood et al. (2012), I derive the first hypothesis of this thesis. If firms with a higher growth avoid more taxes, an increase in the degree of tax avoidance between countries can partially be explained by a higher growth mean of the entities within these countries, subsequently. The rationale behind this is that firms with a higher sales growth, being equivalent with an increase in the tax base (ceteris paribus), have a higher incentive to engage in more aggressive tax avoidance strategies (Kim and Im, 2017). Thus, I expect a positive relation between sales growth and tax avoidance.
In other words, my first hypothesis is:
H1: Higher firms’ sales growth in the mean is associated with a high degree of tax avoidance.
Similarly, Graham et al. (2014) indicate that more profitable firms try to avoid more taxes since, due to their pre-tax profitability, they have a higher tax base (ceteris paribus) and the incentive to keep their profitability even after they have paid the taxes as a high ETR would reduce their profitability (Zhu, Mbroh, Monney & Bonsu, 2019). Again, complemented by Dyreng et al. (2019) and Atwood et al. (2013) and following the implication-driven rationale as above, this leads to expect that if firms are more profitable in the mean this leads to a higher degree of tax avoidance (positive relation between profitability and tax avoidance). This results in my second hypothesis:
H2: The high degree of tax avoidance in specific countries is driven by the fact that the firms in these countries are more profitable in the mean.
One very frequently used control variable, and my last firm characteristic, is the size of a company. As mentioned in Section 1, particularly big companies are said to not pay their fair share of taxes implying that they pursue a quite aggressive tax avoidance strategy. In the literature, there are contradicting findings about the relation between the size of a company and its degree of tax avoidance. The reason for these ambiguous results is the competition between two common theories in the literature; the political power and the political cost theory. Belz, von Hagens & Steffens (2019) point out that in the last two decades 20 out of 49 studies support the political cost and 9 the political power theory, eleven find evidence for both theories and nine did not find any clear relation at all, illustrating the ambiguity of this relation.
On the one hand, the political power theory states that the bigger a company is, the more taxes it avoids since, due to their size, they have more political power, enabling them to influence the political process in their favor16 (Siegfried, 1972). Evidence for this positive relation can be found in i.a. Guha (2007), Mills, Erickson & Maydew (1998), and Richardson & Lanis (2007). On the other hand, the political cost theory indicates a negative relation between the size of a firm and its degree of tax avoidance. The reason for this is the fact that larger firms are more visible for the public, exposing them to greater observation through the government on the one hand (Boynton, Dobbins & Plesko, 1992) and more customer awareness on the other hand, leading to an increase in the expectation of social responsibility (Jensen & Meckling, 1976; Watts & Zimmerman 1986; Zimmerman, 1983). While the former increases the risk of state actions against the company (Aichian & Kessel, 1962), the latter bears the risk of reputational damage (Graham et al., 2014) if a company is perceived as a “poor corporate citizen”17 (Hanlon & Slemrod, 2009, p. 127).
This prevalence of literature regarding the impact of the size of a company on its tax avoidance behavior shows the importance of this firm characteristic. Hence, I expect a significant correlation between these two terms does exist. Due to the dissension about the course of the relation, I do not want to make a hypothesis about the sign of the regression coefficient. Consequently, my third hypothesis states:
H3: The degree of tax avoidance in specific countries is driven by the average size of firms in these countries.
In their paper, Atwood et al. (2012) find that the level of the statutory tax rate determines the degree of tax avoidance. Firms in countries with a higher statutory tax rate are more incentivized to avoid taxes since their marginal effect of tax avoidance on the after-tax profit is higher than for firms in low tax rate jurisdictions. This intuitive thought implies that e.g. more profitable firms should be even more aggressive in their tax avoidance strategy if they are located in a high tax jurisdiction compared to being located in a country with lower statutory tax rate. This leads to my fourth hypothesis.
H4: The effect of stronger growing, more profitable and smaller/larger firms on tax avoidance is higher in countries with a high level of statutory tax rate compared to low tax countries.
Concurrently, Atwood et al. (2012) point out that the degree of tax avoidance depends on the book-tax conformity (BTaxC)18 required by the government. Firms working under a higher required BTaxC have fewer opportunities of avoiding taxes, resulting in a negative correlation between BTaxC and the degree of tax avoidance. This is consistent with Chan, Lin & Tang (2013), Desai (2005), and Tang (2015) and thus, I expect the opportunities for e.g. more profitable firms being fewer in countries with high BTaxC leading to a smaller effect of this firm characteristic on tax avoidance in high BTaxC countries compared to low ones. My last hypothesis can be written as:
H5: The effect of stronger growing, more profitable and smaller/larger firms on tax avoidance is higher in countries with low book-tax conformity compared to high BTaxC countries.
3 Empirical Approach and Data
As mentioned at the beginning, there are different ways of measuring tax avoidance. Gebhart (2017) point out that there are several “Effective Tax Rate based measures, Book-Tax-Differences based measures, […] Tax Shelter Scores, [and] Unrecognized Tax Benefits” (p.43). In my measure, I am combining the effective tax rate component with the book-tax rationale. Therefore, I define tax avoidance as the difference between the statutory tax rate and the GAAP ETR. In other words, the difference between the taxes that should have been paid and the actual tax expense scaled by the pre-tax income. Thus, a higher value for this measure implies an increasing ability to undercut the statutory tax rate and not just reducing the taxable income. According to Dyreng, Hanlon & Maydew (2008), trying to measure tax avoidance just within one year would be ambiguous since fewer taxes in one year are mostly offset by higher taxes paid in a later period. Thus, a one-year tax avoidance measure would not be a good predictor of long-run tax avoidance.19 Keeping this in mind, I am still going to use a measure for tax avoidance that is based on annual financial numbers. I am following this rationale because I am interested in analyzing the country differences in the degree of tax avoidance per year and not the characteristics of a firm being able to avoid taxes over a long time period. For this purpose, I slightly modify the measurement used by Atwood et al. (2012) by using just the one-year tax expense and pre-tax income. Following this argumentation, the measure of the degree of tax avoidance for firm i in year t looks the following:
Abbildung in dieser Leseprobe nicht enthalten
where Pi is the pre-tax income, the statutory tax rate in which the firm’s headquarter is located and txt the tax expense of the firm.
To not neglect the findings of Dyreng et al. (2008) and to show the robustness of my results even over a longer period of time, I will use the original measure of Atwood et al. (2012) as robustness test within a second regression.
Abbildung in dieser Leseprobe nicht enthalten
1 More detailed descriptions of further tax avoidance measures are given in Section 3.1.
2 While tax shelters in general are specific transaction structures that aim to reduce the tax burden of a company, legally and illegally, the term refers only to the legal ones in this thesis (Brown, 2011).
3 Since tax evasion is not part of this study, the mechanisms of tax evasion are not further elaborated.
4 There are some exceptions to these rules which are not further discussed in this paper.
5 The authors find that unprofitable firms (cost > revenue) have a marginal tax rate of zero, thus shifting profit to these firms does not cost any taxes and reduces the tax base of the initial profit-making firm.
6 Deducting the interest expense in high tax jurisdiction has a higher marginal tax saving effect than the same process in a low tax jurisdiction.
7 In the last decade, on average, 15 tax articles were yearly published in the top accounting journals (p.72).
8 There is some research investigating the question if firms that are not paying their fair share of taxes, should be really taxed higher because research suggests that some of them are able to pass the taxes on e.g. their consumers (Dyreng et al., 2019) which would result in inefficient taxation.
9 Translated from the German website stuttgarter-nachrichten.de (2019).
10 While Dyreng et al. (2019) use the global Cash ETR as an indicator of tax avoidance and thus find a negative correlation between the independent variable Leverage and the dependent variable (ceteris paribus) through e.g. shifting the higher debt to high tax jurisdictions, subsequently reducing ETR, Atwood et al. (2012) suggest a positive coefficient of their Leverage variable. The rationale behind it is that they define tax avoidance as the difference between statutory tax rate and ETR. Consequently, an increase in leverage leads to an enlargement in this difference through the same process as just explained, resulting in an increase in the dependent variable TaxAvoid.
11 Lisowsky (2010) and Wilson (2009) differentiate between the reduction of the taxable income through expense deductions and other tax shelter transactions. Thereby, they define tax avoidance as the latter followed by the finding that less prior interest deductions, resulting in a higher taxable income after these deductions, lead to higher engagement in other tax shelter transactions.
12 This decrease is calculated by dividing the coefficient of the LCF variable by the sample mean of LCF (-0.0442/0.2815=-0.1570).
13 There are more characteristics, like the extent of international operations (Dyreng et al., 2019; Graham et al., 2014; Robinson et al., 2010) but since they are not part of my analysis, I did not further discuss them.
14 A definition of this term can be found in section 3.
15 This is calculated by multiplying the coefficient of the Sales Growth variable (-0.11) with its standard deviation.
16 One example therefor are lobbying activities (Belz et al., 2019).
17 This term refers to a firm not paying its fair share of taxes by e.g. engaging in tax shelters.
18 A detailed description of this measure can be found in section 3.
19 Dyreng et al. (2008) showed that annual Cash ETRs are not good predictors of multiple years’ ETR.
- Quote paper
- Simon Falcke (Author), 2019, A Countries' Level of Tax Avoidance. Empirical Investigation of Exemplary Drivers, Munich, GRIN Verlag, https://www.grin.com/document/540163