Does political uncertainty influence the degree of tax avoidance of firms?

(1)

Does political uncertainty influence the

degree of tax avoidance of firms?

Thesis, MSc Accountancy, Research in Accounting and Control University of Groningen, Faculty of Economics and Business

16 January 2020 Word count: 9.323 ANDREA VISSER Studentnumber: S2692961 Schans 22 8531 DW Lemmer phone: +31 (0)6 24 39 14 94 e-mail: a.visser.30@student.rug.nl Supervisor:

(2)

ABSTRACT

Tax avoidance is a socially relevant topic. There are many multinationals who make billions of profits but pay zero income taxes. In this study, the influence of political uncertainty on the degree of tax avoidance is researched. Because the degree of tax avoidance can be influenced by opportunistic behavior of managers, also the interaction effect with earnings management (as proxy for opportunistic behavior) is included in the study. It is hypothesized, based on the agency theory and previous literature, that, the relationship of political uncertainty with tax avoidance can be either positive or negative. The traditional view of tax avoidance is that it adds value to the firm, which makes it more likely that firms perform tax avoidance. On the other hand, managers possibly face reputational risks, political risks, tax risks and other risks. Based on the results, there is no relationship found between political uncertainty and tax avoidance. To hypothesize the interaction effect of earnings management, opportunistic behavior is seen as conditional variable which influences the behavior of managers with respect to tax avoidance. Empirical results state that the relationship is significant when earnings management is included. So, opportunistic managers perform less tax avoidance in case of political uncertainty than less opportunistic managers. The research contains a sample of 37.443 observations representing 7.112 different firms from the period 2002-2016. Because the research shows significant results, it gives insights to politicians, investors and academics with respect to firm behavior in periods of political uncertainty (e.g. elections or new policies).

(3)

1. Introduction

In 2017, Trump introduced the Tax Cuts and Jobs Act, which sharply cuts federal corporate taxes from 35 percent to 21 percent. On the 16th_{of February 2019, an article in ‘The}

Guardian’ noted that Amazon did not pay any taxes in 2018, even though the company made a profit of $11.2bn. In that same year, Netflix reported a profit of $845m but neither paid federal or state income tax. In the article of ‘The Guardian’, the matter that multinationals are paying less corporate tax could be influenced by the Tax Cuts and Jobs Act. However, the adjusted corporate tax percentages of the new law are not able to explain the happening that firms pay less taxes on itself. The corporate tax paid by mentioned multinationals are not lowered by the tax rate but around zero. This could be the result of tax avoidance performed by the firms. Firms also react on certain events happening in the society. In this thesis, the focus is on political events. Besides the new Tax Cuts and Jobs Act, there have been other events in 2018 which may have affected behavior concerning tax avoidance of firms. Political events like the upcoming elections in the United States, the threats of the United States Congress to shut down the federal government and other events illustrate that firms face considerable political uncertainty. This political uncertainty potentially influences the degree of tax avoidance firms perform.

In this thesis, political uncertainty is defined as uncertainty about the government’s future actions (Pastor and Veronesi, 2013). For firms, this results in uncertainty about future implications of decisions made by different parties (e.g. managers, government, stakeholders, shareholders, investors and debtholders). Recent studies suggest that events as elections, a federal shutdown and other types of political events are likely to affect corporate behavior (e.g. Hassan, Hollander, van Lent and Tahoun, 2019; Pastor and Veronesi, 2013). For instance, Pastor and Veronesi (2013) find that, when risks occur where the future path of government policy is uncertain, businesses are more likely to delay spending money until the uncertainty has been resolved. This is consistent with the findings of Hassan et al. (2019). They found that when political uncertainty increases, it is associated with significant increases in stock volatility and with significant decreases in its investment, planned capital expenditures, and hiring. In this paper, I will add to this nascent literature by investigating the effects of political uncertainty on the degree of corporate tax avoidance.

(5)

Consistent with Dyreng, Hanlon and Maydew (2008), I define tax avoidance as the ability to pay a low amount of income taxes relative to corporate pre-tax earnings in a legitimate way. The traditional view on tax avoidance consists that it adds value to the firm (Desai and Dharmapala, 2009), because it maximizes after-tax profits. Performing tax avoidance can create a shield for managerial opportunism by hiding bad news and create more information asymmetry between the managers and the shareholders, which is favorable for the managers. The traditional view makes it likely that there exists a positive relationship between political uncertainty and the degree of tax avoidance, which means that a firm performs more tax avoidance in case of political uncertainty. On the other hand, there are several risks that occur when tax avoidance is performed (Christensen and Murphy, 2004). Because managers are risk-averse and respond conservative to political uncertainty, a negative relationship between political uncertainty and tax avoidance is expected. The more political uncertainty the firm experiences, the less tax avoidance it will perform.

The relation between political uncertainty and tax avoidance can be influenced by the opportunistic behavior of managers, which is proxied by earnings management. Earnings management is a purposeful intervention in the external financial reporting process, with the intent of obtaining some private gain. Earnings management is mostly performed by opportunistic and self-serving managers and thereby seen as proxy for opportunistic behavior (Jiraporn, Miller, Yoon and Kim, 2008). As stated in Dechow and Skinner (2000), earnings management can be performed through accounting choices (Accrual Earnings Management) and “real” cash flow choices (Real Earnings Management). While the first involves manipulation of accruals, by means of managerial discretion and judgement to alter financial reports within the possibilities of the country’s Generally Accepted Accounting Principles (Dechow, Sloan and Sweeney, 1995; Walker, 2013), “Real earnings management” involves decisions regarding the timing of investment or financing decisions in order to alter reported earnings (Schipper, 1989, p.92). In this thesis, the focus is on accrual earnings management, because prior research has used accrual earnings management as proxy for the degree of opportunistic behavior of managers (Jiraporn et al., 2008). When referred to earnings management in this paper, it means accrual earnings management. In case of opportunistic behavior of managers, the relationship between political uncertainty and tax avoidance can be different. Thus, earnings management potentially acts as contextual variable which influences the relation between political uncertainty and tax avoidance.

(6)

According to Li, Maydew, Willis and Xu (2018), the degree of tax avoidance is an outcome of a political process, and increased political uncertainty may be associated with changes in the extent of tax avoidance. Since the political events might influence the degree of tax avoidance of a company, it is interesting to find out if there is a relation between the political uncertainty a firm is experiencing and the degree of tax avoidance of organizations. The change in tax behavior can result in a change of the degree of tax avoidance. These possible relations will be more elaborated in the theoretical framework section. The research question that will be guiding the research is the following:

What is the effect of political uncertainty experienced by a firm on the extent of tax avoidance?

Because the pre-tax income can be influenced through the behavior of managers, also the influence of opportunistic behavior (proxied by earnings management) on the relation between political uncertainty and tax avoidance will be researched. This results in the following sub-question:

What is the effect of opportunistic behavior on the relation between political uncertainty the firm experiences and the extent of tax avoidance?

The research is expected to deliver a valuable scientific contribution to the literature focusing on the consequences of political uncertainty. Studies in this emerging stream of accounting and finance literatures find that political uncertainty is associated with a decline in investments during times of political uncertainty (Jens, 2017; Julio and Yook, 2012; Pastor and Veronesi, 2013) and affects cash flows, stock returns, option markets and international capital flows (Belo, Gala and Li, 2013; Guorio, Siemer, Verdelhan, 2015; Kelly, Pastor and Veronesi, 2016). To the best of my knowledge, the effects of political uncertainty on tax avoidance has not been researched. This study intends to fill that gap in the literature. The relation between political uncertainty and tax avoidance is a socially relevant relation, it affects many people, firms and nations.

While previous research about tax avoidance is based mainly on the relation between corporate tax avoidance and investors and creditors (e.g. Kim, Li and Zhang, 2011; Graham and Tucker, 2006; Goh, Lee, Lim and Shevlin, 2016; Shevlin, Urcan and Vasvari, 2013), this

(7)

research focuses on other stakeholders of the firm, namely on the relation between politics and corporate tax avoidance. There is a direct relationship between corporations and governmental institutions because of the corporate taxes that have to be paid by the firms, and an indirect relationship between organizations, governmental institutions and the community over the benefits of tax payment (Nurwanah, Sutrisno, Rosidi and Roekhudin, 2018). The degree of taxes paid by firms, influences the economy and affects the entire community. Tax avoidance is a form of behavior that is performed by a lot of firms. The effective corporate tax rates paid by firms significantly decreased the past twenty-five years, which is only explained for a small part by the decrease in statutory tax rates and changing firm characteristics (Dyreng, Hanlon, Maydew and Thornock, 2014). This effect over the past decades makes it important to perform research on tax avoidance. It gives insights for politicians about why and when the effective tax rates are lower than the statutory tax rate. Besides, it gives implications for politicians what the reaction of firms is on situations which firms see as uncertain (e.g. elections or new policies). For example, changes in the financial markets are associated with a certain party (democrats or republicans) winning the elections (Snowberg, Wolfers and Zitzewitz, 2007). This is consistent with the findings of Hassan et al. (2019), firms experience more political uncertainty around periods of elections. Besides, their measure of political uncertainty varies across industries that are highly influenced by governmental policies. So, elections and the dependence on governmental policies are determinants which increases political uncertainty.

The potential effect of earnings management on the extent to which political uncertainty and tax avoidance are linked has, to the best of my knowledge, not yet been researched. By involving earnings management in the relation, politicians can also get insights in which firms perform tax avoidance. Previous research on the relation between earnings management and tax avoidance stated that earnings management and tax (avoidance) are related topics (Desai and Dharmapala, 2009; Phillips, Pincus and Rego, 2003; Guenther, 1994). Desai and Dharmapala (2009) explored the direct relationships between earnings management and tax avoidance, Philips et al. (2003) used deferred tax expense to detect earnings management and Guenther (1994) found that earnings are managed as a result of the Tax Reform Act of 1986, when the taxes are reduced from 46 percent to 34 percent. Based on their findings and to extent the already performed research, it is interesting to involve earnings management as a moderating variable in this research.

(8)

Another contribution is that the research will be performed with new measures. The relation between tax avoidance and political uncertainty has been researched earlier by Li et al. (2018), but their paper used periods of elections worldwide as a proxy for political uncertainty. This research makes use of a different proxy of political uncertainty, namely the measurement which is recently developed by Hassan et al. (2019). This measure has not been used earlier for research in political uncertainty. The main difference between the measurement used by Li et al. (2018) and the measurement developed by Hassan et al. (2019), is that it is idiosyncratic. Other measurements suppose that all firms react in the same way on political uncertainty, while the measurement developed by Hassan et al. (2019) is based on the reactions of individual firms. Another difference is that it focuses on political decisions and conversations instead of elections as the proxy for political uncertainty, as it is used in previous research. What this measurement exactly implies will be discussed in the method section. Next to the measure of political uncertainty, a new measure for tax avoidance will be used. This measure is developed by Henry and Sansing (2018). Either this measurement will be elaborated further in the method section. Besides the new measures, also the traditional measures will be used as robustness checks to confirm that the new measures are reliable to use.

In the next section the theoretical background surrounding the main topics and the hypotheses development will be discussed. Section three addresses the research design and in section four the results of the research will be presented. In the fifth section the results of the robustness checks are described. The sixth section the results will be discussed and the conclusion, limitations and suggestions for future research is presented.

(9)

2. Theoretical background and hypotheses development

2.1 Agency theory

The starting point of the hypothesis development is agency theory. Agency theory starts with the view that a firm is a nexus of contracts (Jensen and Meckling, 1976) involving various parties that act rationally, driven by a desire to maximize their own economic utility. There exist various parties which have contracts with the firm due to the separation of ownership and control. This separation of ownership and control results in information asymmetry and possible moral hazard. The parties that are involved in a typical agency setting are the shareholders (principals) and the managers (agents). The two important contracts within agency theory are (1) the labor and compensation contracts of firms’ executives and (2) the contracts between the firm and the creditors/debtholders. Based on these contracts, agency theory addresses the potential conflicts between executives and creditors/debtholders that arise as their interests, preferences and goals may diverge (Walker, 2013). Agency theory especially highlights the potential conflicts of interests between the executives and shareholders and between the debtholders and shareholders. These conflicts result from the fact that shareholders retrieve the exceeding gains, after the executives, debtholders and other stakeholders have retrieved their rewards (Walker, 2013).

In the relationship between managers and shareholders, the managers act as agents, whereas the shareholders act as the principals (Yorke, Amidu and Agyemin-Boateng, 2016). The managers are hired by the shareholders to maximize the shareholder value as much as possible (Campbell, 2007; Christensen and Murphy, 2004). Because the ownership of U.S. firms is diverse, in the sense that listed firms typically involve a large number of minority shareholders each owning a small fraction of the firm’s stock, the agency problem between shareholders and managers is more severe in the United States than in countries where large shareholders are more common (Li, Liu and Ni, 2017). Several studies conclude that, due to the separation of ownership and control, frequently managers have personal goals that do not align with goals of shareholders and, hence, do not always act in the best interest of the shareholders (Jensen, 1994; Eisenhardt, 1989; Crutchley and Hansen, 1989; Donaldson and Davis, 1991; Jensen and Meckling, 1976).

Managers have, because they control the business, more information of the firm than shareholders. This usually results in information asymmetry between the managers and

(10)

shareholders. Managers have information (e.g. knowledge about current and future operations, resources, risks and earnings) that is not available for the shareholders (Scott, 2003). This information asymmetry increases the ability of the managers to make self-serving choices, which aggravates agency problems and induces a moral hazard risk for shareholders (Campbell, 2007; Ndofor, Wesley and Priem, 2015). Specifically, Arrow (1968) supposes that information asymmetry results in the potential for post-contract opportunism by managers which causes self-interested actions of managers that are harmful for, yet remain opaque to shareholders. The self-interested actions are a form of opportunistic behavior. A potential action of opportunistic behavior of managers is performing tax avoidance. Because tax avoidance helps to increase firm value and thus individual utilization, managers are more motivated to perform tax avoidance (Desai and Dharmapala, 2006; Watts and Zimmerman, 1978; Khurana and Moser, 2013).

2.2 Tax avoidance

As stated earlier, tax avoidance is the ability to pay a low amount of income taxes relative to corporate pre-tax earnings. In this paper, tax avoidance is limited to legal activities to lower the amount of income taxes paid (Dyreng et al., 2008). In a lot of industries, the amount of taxes paid are the second largest cost for firms (Houlder, 2010). Because of the magnitude of tax costs, tax consequences are a driving factor in many corporate decisions (Desai and Dharmapala, 2009). Previous research shows that tax avoidance already had been linked to the characteristics and incentives of firms and their managers. Tax avoidance is linked to managerial financial incentives, ownership structure, and governance (Cen, Tong and Sun, 2017; Phillips, 2003; Chan, Mo and Zhou, 2013). For instance, Cen et al. (2017) find that when governance is weak (strong), investors expect high (low) information asymmetry and thus more (less) concerns about the potential risks of tax avoidance. Another example, based on the research of Phillips (2003), is that compensation based on after-tax accounting-based measures, motivates managers to perform tax avoidance.

The traditional view is that tax avoidance increases firm value through maximizing the after-tax profits. An incentive for managers to perform tax avoidance is to create value for shareholders. When a firm engages in a lot of tax avoidance, the after-tax earnings and cash flow will increase, which can result in an increase in shareholders’ value and stock market rewards (Khurana and Moser, 2013; Frank, Lynch and Rego, 2009). On the other hand, tax avoidance can also be personally beneficial for managers and, hence, this view suggests tax

(11)

avoidance is a form of self-interested or opportunistic behavior. Tax avoidance can help managers in the agency problem by hiding bad news for investors and enable the managers to benefit at the expense of the shareholders to meet the benchmarks for their personal bonuses (Kim et al., 2011; Khurana and Moser, 2013). This creates information asymmetry and moral hazard as a result of opportunistic behavior of the managers by performing tax avoidance. However, there are at least three types of risks identified in the literature that occur when a firm involves in tax avoidance: tax risks, reputational risks, and political risks and other risks1_(e.g.

Christensen and Murphy, 2004; Guenther, Matsunaga and Williams, 2017; Gallemore, Maydew and Thornock, 2014).

The first risk that occurs when firms perform tax avoidance is a tax risk. A tax risk is the risk that regulators do not accept the magnitude of tax avoidance a firm engages in and penalize the firm with requiring it to repay the taxes with fines, penalties or interest. This risk is most of the time caused by several factors. First, the discretion firms have in measuring profitability and allocating costs. The second factor is the complexity in the business activities and the last factor is the complexity and lack of clarity in tax rules worldwide (Christensen and Murphy, 2004). Due to these factors, firms can face uncertainty about the legality of the way and degree of tax avoidance, especially because the legality is determined after the fact (Hanlon and Heitzman, 2010).

The second risk that is caused by performing tax avoidance is the reputational risk (Christensen and Murphy, 2004; Gallemore et al., 2014). Reputational risk can for example be realized in (digital) newspapers, television, internet and social media (Christensen and Murphy, 2004). But either stock price declines, recalling customers and being targeted by activist groups can harm the reputation of firm (Hanlon and Slemrod, 2009; McIntyre, Gardner, Wilkins and Philips, 2011). A well-known example is Starbucks’ legal tax avoidance in the United Kingdom in 2012. It became a target of, negative media publications, public outcries and political attacks after the reveal. The years afterwards, Starbucks decided to pay £20 million, voluntarily (The Guardian, 2012). Westphal and Deephouse (2012) suppose that a reputational risk can influence the behavior of managers. According to Gallemore et al. (2014), executives are making a

1_{There are also other risks with significant financial impact that tax avoidance involves. Firms can create costly}

structures or investments to change locations of factories or take advantage of transfer pricing arrangements. Besides, tax avoidance can result in more opportunity costs because it influences firms’ decision about how to manage cash, make investments, and pay dividends (Blouin and Krull, 2009; Foley, Hartzell, Titman, Twite, 2007). Next to opportunity costs, tax avoidance also influences the extent of agency costs (Desai and Dharmapala, 2009).

(12)

benefit-cost-analysis of the planned tax avoidance and decide not to engage in tax avoidance when the reputational costs are higher than the benefits. It can be concluded that the reputational risk has an important influence on the degree of tax avoidance a firm performs.

Another risk that Christensen and Murphy (2004) stated is the political risk that occurs when a firm performs a lot of tax avoidance. For example, in case of public outcries about that firms do not pay their fair share of taxes as a result of tax avoidance, it creates pressure on politicians to undertake actions. This potential risk can result in adverse legislative or regulatory actions, or changes in the tax laws. Politically sensitive firms pay more federal taxes to avoid losing political benefits because actions undertaken by the politicians can possibly harm firms economically (Christensen and Murphy, 2004; Mills, Nutter and Schwab, 2012). This argument is supported by Mills et al. (2012), who find that politically sensitive firms pay more federal taxes to avoid losing political benefits.

To conclude, tax avoidance exposes firms to tax risks and non-tax risks (Scholes, Wolfson, Erickson, Maydew and Shevlin, 2008). Based on the discussed arguments, even though tax avoidance increases current after-tax flows, it is not indubitably value-adding for a firm. This is supported by the argument of Desai and Dharmapala (2009), who found that firm value can also decrease through tax avoidance. Tax avoidance does not, as the traditional view states, always increase firm value (Desai and Dharmapala, 2009).

The influence of tax avoidance on the firm value can be unfavorable for shareholders, which results in an agency problem. Besides, managers are also able to create an agency problem by avoiding taxes. By performing complicated tax avoidance activities, managers can create a shield for managerial opportunism and the diversion of rents, which results in more information asymmetry. This is mainly the problem in poorly governed firms, because of the increased opportunities for managerial rent diversion (Desai and Dharmapala, 2009). Desai and Dharmapala (2009) use the level of institutional ownership as proxy for a good governed firm, because institutions are better able to monitor managerial performance. As discussed earlier, U.S. firms usually have a lot of diverse small shareholders, so are more likely to be poorly governed. The contrast between the traditional view of tax avoidance and the possible risks discussed, results in an agency problem between the shareholders and the managers.

(13)

2.3 Political uncertainty

Governments have a lot of influence on firms. Governmental institutions namely shape the conditions in which the private sector operates. The government influences the firms in certain ways: by levying taxes, providing subsides, enforcing laws, regulating competition, defining policies in various areas. Because governments change the rules continuously, it happens that managers of firms feel uncertain about future actions of the government (Pastor and Veronesi, 2013). These rules can affect businesses and result in different decisions and behavior by shareholders and managers. Because shareholders and managers are ignorant about the possible actions of the other party, this uncertainty results in more information asymmetry between the shareholders and the executives. Managers have more information about the firm and how the political uncertainty is influencing the business than the shareholders, information asymmetry between the managers and shareholders will occur. This information asymmetry will result in different behaviors for managers and shareholders.

On the one hand, political uncertainty results in several possibilities for managers, because of the information asymmetry. It is possible that managers apply an observing strategy in case of political uncertainty, which results in more conservative decisions. Managers will suspend decisions until the political uncertainty is over and the implications for the firm are clear (Bloom, Bond and van Reenen, 2007). Research has shown that an example of decisions that managers are delaying are less investments in time of political uncertainty (Jens, 2017; Julio and Yook, 2012; Pastor and Veronesi, 2013; Bloom et al., 2007). Political uncertainty also has an effect on cash flows, stock returns, option markets and international capital flows (Belo et al., 2013; Gourio et al., 2015; Kelly, Pastor and Veronesi, 2016). Besides postponing decisions, Hassan et al. (2019) find that managers manage the political uncertainty about relevant topics actively by performing lobbying activities, forging links to politicians and donating to political campaigns. Because of the information asymmetry, managers can also see political uncertainty as an opportunity to perform opportunistic behavior by tax avoidance. Thus, because of the information asymmetry that occurs in case of political uncertainty, the behavior of managers changes in order to prepare for the potential consequences of the political uncertainty.

Likewise, shareholders react on the information asymmetry caused by political uncertainty. Investors react to changes in political uncertainty a firm because of its association with the firm value (Pastor and Veronesi, 2013). Research has shown that stock volatility

(14)

increases in periods of political uncertainty (Hassan et al., 2019; Pastor and Veronesi, 2012). Shareholders experience uncertainty about how managers will react to the political uncertainty. When shareholders are experiencing political uncertainty, they demand more information about the policies of the firm to reduce the increased information asymmetry (Lev and Zarowin, 1999) and the risk premia are increased, because shareholders claim higher returns on their investments in case of more political uncertainty (Hassett, 2011; Pastor and Veronesi, 2013). Healey and Palepu (2001) state that the information exchange between managers and shareholders is contributing the accurate allocation of capital between managers and investors.

2.4 The effect of political uncertainty on tax avoidance

Based on the agency theory and the described literature about tax avoidance and political uncertainty, a hypothesis for the research can be formulated for the effect of political uncertainty on tax avoidance. Political uncertainty results in information asymmetry and thus, changing behaviors of managers and shareholders in times of political uncertainty. Because managers are uncertain about the implications of decisions, there is a non-directional hypothesis formulated for the relation between political uncertainty and tax avoidance.

On the one hand, managers are more conservative in case of political uncertainty because of the information asymmetry and the risk-aversity of managers will strengthen due to this information asymmetry (Wiseman and Gomez-Meija, 1998). Managers will perform a cost-benefit-analysis prior to performing tax avoidance and will not perform certain behavior if the costs exceed the benefits. Possible indirect (personal) benefits are reaching compensation targets and their reputation as a manager. The tax risk, reputation risk, politic risk and other risks described, will decrease the willingness for tax avoidance. In combination with the more-demanding attitude of shareholders in times of political uncertainty, I expect that there is a negative relationship between political uncertainty and tax avoidance.

On the other hand, there is also a possibility that there is a positive relationship between political uncertainty and tax avoidance. For instance, Li et al. (2018) concluded that political uncertainty will result in more tax avoidance. They stated that it might be possible that the period of political uncertainty results in different tax laws and the investment of firms in tax avoidance strategies will be rendered less effective after the change. The degree of tax avoidance is based on a cost-benefit-analysis. If the benefits will exceed the possible costs resulted by the risks described by Christensen and Murphy (2004) it is likely that firms will

(15)

perform more tax avoidance in case of political uncertainty. Besides, managers can use tax avoidance strategies to create a shield for managerial opportunism by hiding bad news for shareholders, which results in more information asymmetry between the managers and shareholders (Desai and Dharmapala, 2009). Even though tax avoidance is not always favorable for the firm value (and thus the shareholders), managers have more possibilities to do so, because of the increased information asymmetry. So, if managers think the (indirect) personal benefits will exceed the costs of tax avoidance in case of political uncertainty, the extent of the performed tax avoidance by managers will be increased in times of political uncertainty.

As a priori it is uncertain which of these views is likely to prevail, I refrain from making a directional hypothesis. Accordingly, I hypothesize:

H1: There is a relation between political uncertainty and tax avoidance

2.5 The moderating role of earnings management

In line with prior research (e.g. Ndofor et al., 2015; Jiraporn et al., 2008), in this study earnings management is used as a proxy for opportunistic behavior of management. It represents the degree to which the management is driven by self-serving behavior. The view on earnings management used in this thesis is that earnings management is opportunistic in nature as supposed by Jiraporn et al. 2008. Opportunistic motives for managers to engage in earnings management are for example, external motives, contracting motives and capital market motives (Walker, 2013; Healy and Wahlen, 1999; El Diri, 2018). In this paper, earnings management is the conditional variable which influences the relationship between political uncertainty and the degree of tax avoidance performed.

Managers whose performance is measured based on after-tax accounting measure, have incentives to perform tax avoidance. In this case, tax avoidance is a form of opportunistic and self-serving behavior (Desai and Dharmapala, 2006). If managers are more driven by opportunistic behavior, it is more likely that they will perform tax avoidance. Because of the hypothesized association between political uncertainty and tax avoidance, it is likely that this relation will become (more) positive when managers are driven by opportunistic behavior, in contrast to managers who are less driven by self-serving behavior.This results in the following hypothesis:

H2: If managers are driven by opportunistic behavior, the association between political

(16)

3. Research design and methods

3.1 Sample selection and data collection

To conduct the empirical research, I start with the data about political uncertainty of Hassan et al. (2019), who have analyzed the quarterly conference calls of listed firms in the United States. The data of the last quarter serves as a proxy for the whole year. The initial sample consists of 51.500 yearly observations from the period from 2002 to 2016. The data that is used for the dependent variable, moderating variable and control variables are retrieved from Compustat.

Because financial and insurance companies have different operating and financing structures, these companies are excluded from the data (4.210 observations). After that, the data is matched with the data about tax avoidance, which results in another deletion of 5.753 observations. After the matching with data about tax avoidance, the data also is matched to the data about accrual earnings management, which also causes a deletion of 2.750 observations. Afterwards, the matching with data about the control variables will take place and as a consequence 602 observations have to be deleted from the sample. Finally, data with less than two years of data about political uncertainty and is not usable for the research is removed from the sample, another 742 observations. This amount of observations is necessary to compare a period of political uncertainty with a period of political certainty. Thus, the final sample consists of 37.443 observations representing 7.112 different firms from the period 2002-2016. To prevent that applicable outliers influence the research, the highest and lowest 1% are winsorized. Table 1 gives an illustration of the sample.

Table 1. Sample used in analyses

Observations

Initial sample from Hassan et al. (2019) 51.500

Firms in financial sector (SIC code: 6000-6199) (2.615)

Firms in insurance sector (SIC code: 6300-6411) (1.595)

Missing data on tax avoidance (5.753)

Missing data on earnings management (2.750)

Missing data on control variables (602)

Companies with less than two observations (742)

(17)

3.2 Measuring political uncertainty

Based on the research of Hassan et al. (2019), political uncertainty (PUNC) is measured by the basis of the amount of words managers use to discuss political uncertainty in the quarterly conference calls with financial analysts. Hassan et al. (2019) made a distinction between political texts and non-political texts and analyzed the transcripts. Based on the transcripts, a pattern-based sequence-classification determines the amount that both sets are used in combination with synonyms for risk and uncertainty relative to the length of the conference-call. When in a conference call a higher share of political risk is discussed, it implies that political risk in that period is higher, which results in a higher score of PUNC. Prior research finds that conference-calls are usually focusing on uncertainties the firm is facing (Hollander, Pronk and Roelofsen, 2010; Bowen, Davis and Matsumoto, 2002; Huang, Lehavy, Zang and Zheng, 2018). Thus, the measure based on conference-calls is a suited measure for the research in this thesis. Besides PUNC, also PUNCTAX is used as a robustness test. PUNCTAX is the share of the conference-calls that centers on uncertainty concerning tax policy. The higher the score, the higher the share of the conversation concerns uncertainty about tax policies. This measurement is used because it is applicable to the subject of this thesis.

3.3 Measuring tax avoidance

The dependent variable (TAHS), tax avoidance, will be measured based on a developed measure of Henry and Sansing (2018). This relatively new measure of corporate tax avoidance is meaningful for all firm-year observations, including loss firms. The measure is the difference between a firm’s cash taxes paid, adjusted for tax refunds receivable, and the product of its pre-tax book income and the statutory pre-tax rate. Then the measure is scaled by the market value of a firm’s assets to make the measure comparable across firms. The higher the value of TAHS, the higher the degree of tax avoidance a firm engages in. The formula is as follows in Compustat items2_:

!"#$

_%&

=

((!*+,

%&

− .!*/

%&

− !*/

%,&12

3 − (0,35 ∗ +8

%&

))

=

+8,<?

%&

− @!*A=,

%&

0,35

B

C

"!

_%&

(2)

3.4 Measuring earnings management

_%&

(3)

In this model, ∆;"%& is the change in current assets for firm i in year t, ∆;EFℎ%& represents the change in cash and cash equivalents for firm i in year t, ∆;H_%& implies the change in current liabilities for firm i in year t, ∆,;H_%& is the change in short term debt included in the current liabilities for firm i in year t and ,=+%& is the depreciation and amortization expense of firm i in year t.

"

_%,&12

+ E

M

++=

_%&

"

_%,&12

+ N

%& (4)

Whereas !";;_%& is total accruals, "_%,&12 represents total assets in year t – 1 for firm i, ∆/=L is the change in revenues for firm i in year t, ∆/=;_%& means the change in net receivables for firm i in year t and ++=_%& implies the gross property, plant and equipment of firm i in year t. In this formula N%& gives the score of discretionary accruals (DACC) for firm i in year t, which is used in the research as proxy for earnings management.

3.5 Measuring control variables

To control for potential confounding factors, control variables are added to the model. The control variables are based on the firm-level control variables included in the research of Li et al. (2018). I start to include profitable firms (ROA) and firms with high leverage or complex finance structures (LEV) because those firms have higher incentives and more opportunities for tax avoidance. Besides, the natural log of total assets (SIZE) is included, to control for the size of the firm. Because Atwood, Drake, Myers and Myers (2012) find that revenue growth is positively associated with tax avoidance, also sales growth (SALESGR) is included as control variable. Because prior research (i.e. Koester, Shevlin and Wangerin, 2017) supposes that investments influences tax, PPE and INTANG are also considered in the model. Based on the research of Dyreng and Lindsey (2009) R&D is included in the research, because the research suggests that income generated by intangible assets is relatively simple to shift to tax havens. Because also foreign countries are used to avoid tax, FOREIGN is included in the research to capture the presence of operations in foreign countries. The indicator variable equals one if firms have non-missing or non-zero foreign income.

(20)

_%,&

∗ ,";;

_%,&

+ S OT ;<R!/<H$ + N

_%,&

[Model 2]

In this formula, +QR;_%,& ∗ ,";;_%,& indicate the moderating effect of earnings management. Depending on the outcomes of the first model, DACC strengthens or weakens the relationship between tax avoidance and political uncertainty, indicated by the sign of O_K.

(21)

4. Results

4.1 Descriptive statistics

Table 2 is presenting the descriptive statistics of all the variables for the full sample. The descriptive statistics show that firms in the sample discuss in conference calls on average for 30,8% of the time about political uncertainty (PUNC). An important part of these conference calls is about tax policy (PUNCTAX), namely 19,4% of the conference calls are on average about tax. Besides, the firms in the sample do on average not score high on both measures on tax avoidance (TAHS, TADD). Both scores are around zero. This is in line with the full sample of Henry and Sansing (2018), whom observed a mean of 0,015 and the research of Desai and Dharmapala (2009), who find a mean of 0,005 of the same measure. Also, the average degree of earnings management (DACC) performed by firms in the sample is relatively low, namely 0,180. The firms are not highly profitable and growing, because the mean of ROA and SALESGR are limited (respectively 0,266 and 0,128). The sample contains on average firms which are relatively big (SIZE) and have a lot of PPE, paid by debt (LEV). The extent of INTANG, FOREIGN and R&D expenses are relatively small. Based on the Pearson correlation matrix (table 3) of all pairwise combinations of the full sample, multi collinearity is not an issue. All scores are less than 0.7.

Table 2. Descriptive statistics

Full sample, N = 37.442

Variable Mean Std. Dev. Min Median Max

PUNC 30,751 27,893 0,000 21,820 97,296 PUNCTAX 19,398 25,888 0,000 10,036 95,407 TAHS 0,009 0,038 -0,042 -0,002 0,221 TADD -0,021 0,083 -0,444 0,000 0,147 DACC 0,180 0,268 0,000 0,106 12,849 ROA 0,026 0,190 -0,826 0,053 0,461 LEV 0,266 0,264 0,000 0,219 1,308 SIZE 6,893 1,935 2,585 6,808 11,706 SALESGR 0,128 0,365 -0,608 0,071 2,262 PPE 1,161 1,947 0,000 0,455 12,295 INTANG 0,208 0,242 0,000 0,120 1,194 FOREIGN 0,013 0,038 -0,102 0,000 0,563 R&D 0,053 0,101 0,000 0,001 0,182

(22)

Table 3. Pearson correlation matrix Variable (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (1) PUNC 1,000 (2) PUNCT 0,124* 1,000 (3) TAVOID -0,016* -0,023* 1,000 (4) TAVOIDD 0,010 0,010 -0,342* 1,000 (5) DACC -0,002 -0,020* 0,074* -0,014* 1,000 (6) ROA 0,016* 0,020* -0,699* 0,200* -0,010* 1,000 (7) LEV 0,001 0,016* -0,059* 0,077* 0,090* -0,059* 1,000 (8) SIZE 0,048* 0,109* -0,252* 0,168* -0,131* 0,292* 0,253* 1,000 (9) SALESGR 0,006 -0,006 -0,117* 0,017* 0,141* 0,033* 0,098* -0,151* 1,000 (10) PPE 0,007 0,011 0,095* 0,050* 0,113* -0,194* 0,227* 0,101* 0,220* 1,000 (11) INTANG 0,008 0,010 -0,107* 0,068* -0,073* 0,090* 0,230* 0,088* 0,100* -0,211* 1,000 (12) FOREIGN 0,013 0,024* -0,231* 0,151* -0,003 0,319* -0,061* 0,169* 0,012 -0,107* 0,058* 1,000 (13) R&D 0,003 -0,020* 0,263* -0,125* 0,133* -0,492* -0,173* -0,413* 0,171* 0,006 -0,073* -0,050* 1,000

(23)

4.2 Test of H1

Table 4 contains the results of the tests performed. In the second column, the test of hypothesis one is presented. The upper value is the coefficient, and the lower value the standard deviation. For the first hypothesis, it is examined if PUNC is significantly associated with TAHS. Thereafter, the sign is interpreted to conclude if it is a negative or a positive association. In the first model, after controlling for other explanatory variables, it is clear that there is no relationship between PUNC and TAHS. The coefficient of PUNC has a value of zero, which is not significant. So, there is neither a positive, nor a negative relationship between political uncertainty and tax avoidance. Hypothesis one can be rejected. This finding also potentially signals that the effects of political uncertainty on tax avoidance may be different depending on managerial characteristics.

The control variables have a significant impact in the model with a p-score less than 0,01. The F-score of the model is 0,000. This means that the model is validate. The R-squared of the model is 0,481. Which means that the correlation between the variables is moderate. The full sample is contained in the model (N = 37.442).

4.3 Test of H2

The third column of table 4 contains the results of the test of hypothesis two. Hypothesis two tests the influence of DACC on the relationship between PUNC and TAHS. In this model, PUNC has a significant influence on TAHS, with a small positive coefficient. Besides, DACC has a significant relationship with TAHS of 0,012. The interaction-effect of DACC on the relation between PUNC and TAHS, is significantly negative (with a coefficient of -0,00001 and a p-value < 0,05). This means that hypothesis two can be rejected. However, because the relation becomes significant when DACC is involved, DACC is, as expected, a conditional variable in the relationship between political uncertainty and tax avoidance. DACC weakens the effect between PUNC and TAHS.

Also, all the control variables have a significant relationship in the model. The R-square of the model is 0,488. This does not variate a lot from the first model and means that the correlation between the variables is moderate. Also, in the second model almost the full sample is tested (N = 36.930).

(24)

Table 4. Test of hypotheses

Model 1 Model 2

Dependent variable

Independent variable TAHS (1) TAHS (2)

PUNC -0,000 0,000 -0,000** 0,000 DACC -0,012*** 0,001 PUNC * DACC -0,000*** 0,000 ROA -0,174*** 0,001 -0,174*** 0,001 LEV -0,003*** 0,000 -0,003*** 0,001 SIZE 0,002*** 0,000 0,002*** 0,000 SALESGR -0,002*** 0,000 -0,003*** 0,000 PPE -0,001*** 0,000 -0,001*** 0,000 INTANG -0,017*** 0,001 -0,017*** 0,001 FOREIGN -0,036*** 0,005 -0,038*** 0,005 R&D -0,096*** 0,004 -0,102*** 0,004 CONSTANT -0,017*** 0,002 -0,013*** 0,002

INDUSTRY-FIXED EFFECTS YES YES

YEAR-FIXED EFFECTS YES YES

N 37.442 36.930

R-squared 0,481 0,488

Prob > F 0,000 0,000

*** = p-value < 0,01; ** = p-value < 0,05; * = p-value < 0,10

(25)

5. Robustness checks

5.1 Alternative regressions

Alternative regressions are performed to add robustness to the results. Firstly, on the both models of table 4 a robust panel-regression is performed. It is concluded that the results of the tests stay the same. Second, the same regression is done but with clustered robust standard errors on year and firm. Either this test does not result in different scores. So, the outcomes of both models are robust.

5.2 Alternative measure of tax avoidance

Because the measure of Henry and Sansing (2018) has not been used in a lot of research yet, also the measure of Desai and Dharmapala (2009) is tested, this measure is called TADD. Their measure contains the book-tax gap, divided by total assets. A high book-tax gap means that the firm undertakes a lot of tax avoidance. The results are presented in table 5. Model 1 presents the influence of PUNC on TADD and model 2 presents the influence of DACC on this relationship. Likewise, the relationship between TADD and PUNC is not significant. The influence of DACC on TADD is significant. However, the moderating influence is not significant. So, also with the measure of Desai and Dharmapala (2009), TADD, instead of the measure of Henry and Sansing (2018) TAHS, hypothesis one can be rejected. In comparison with the measure TAHS, also the moderating effect is not significant, which means that also hypothesis two has to be rejected.

5.3 Political uncertainty concerning tax

Likewise, the measure of Hassan et al. (2019) is tested in another way. Hassan et al. (2019) specified the topics where the conference calls are about and split the score in political uncertainty about tax policy, institutions, security defense, environment, trade, economic policy, health and technology. As an alternative measure, the political uncertainty about tax policy (PUNCTAX) is used, because it is applicable to the subject of this thesis. Based on the outcomes of model 1 presented in table 6, there is a slightly significant and negative (coefficient = -0,000008; p-value < 0,1) relationship between PUNCTAX and TAHS. Which means that hypothesis 1 can be accepted in the first model. In the second model, there is a significant negative influence of earnings management (DACC) on the relationship between political uncertainty concerning tax (PUNCTAX) and tax avoidance (TAHS).

(26)

5.4 Alternative measure of tax avoidance and political uncertainty concerning tax

The last test contains the both alternative measures for both models. The results are presented in table 7. Based on these measures, all the hypotheses can be rejected. Only model two gives a significant negative relation between earnings management (DACC) and political uncertainty.

Table 5. Test of hypotheses with TADD

Model 1 Model 2

Dependent variable

Independent variable TADD (1) TADD (2)

PUNC 0,000 0,000 0,000 0,000 DACC -0,008*** 0,002 PUNC * DACC 0,000 0,000 ROA 0,134*** 0,003 0,134*** 0,003 LEV 0,001 0,003 0,001 0,003 SIZE 0,003** 0,001 0,002** 0,001 SALESGR -0,004*** 0,001 -0,004** 0,001 PPE 0,003*** 0,000 0,003*** 0,000 INTANG 0,028*** 0,003 0,028*** 0,003 FOREIGN 0,087*** 0,015 0,081*** 0,015 R&D 0,072*** 0,015 0,071*** 0,015 CONSTANT -0,056*** 0,006 -0,055*** -7,680

N 37.442 36.930

R-squared 0,071 0,074

Prob > F 0,000 0,000

(27)

Table 6. Test of hypotheses with PUNCTAX

Model 1 Model 2

Dependent variable

Independent variable TAHS (1) TAHS (2)

PUNCTAX 0,000* _0,000 0,000 _0,000 DACC 0,011*** 0,001 PUNCTAX * DACC -0,000*** _0,000 ROA -0,174*** _0,001 -0,174*** _0,001 LEV -0,003** 0,001 -0,004** 0,001 SIZE 0,002*** _0,000 0,002*** _0,000 SALESGR -0,002*** 0,000 -0,003*** 0,000 PPE -0,001*** _0,000 -0,001*** _0,000 INTANG -0,017*** _0,000 -0,017*** _0,001 FOREIGN -0,036*** 0,005 -0,038*** 0,005 R&D -0,096*** _0,004 -0,102*** _0,005 CONSTANT 0,017*** 0,002 0,012*** 0,002

N 37.442 36.930

R-squared 0,481 0,488

Prob > F 0,000 0,000

(28)

Table 7. Test of hypothesis with PUNCTAX and TADD

Model 1 Model 2

Dependent variable

Independent variable TADD (1) TADD (2)

PUNCTAX -0,000 0,000 -0,000 0,000 DACC -0,008*** 0,002 PUNCTAX * DACC 0,000 0,000 ROA 0,134*** 0,003 0,137*** 0,003 LEV 0,001 0,003 0,001 0,003 SIZE 0,003*** 0,001 0,002** 0,001 SALESGR -0,004*** 0,001 -0,004*** 0,001 PPE 0,003*** 0,000 0,003*** 0,000 INTANG 0,028*** 0,003 0,028*** 0,003 FOREIGN 0,087*** 0,015 0,081*** 0,015 R&D 0,072*** 0,010 0,071*** 0,010 CONSTANT -0,056*** 0,006 -0,053*** 0,007

N 37.442 36.930

R-squared 0,071 0,074

Prob > F 0,000 0,000

(29)

6. Conclusion and discussion

In 2018, there are several multinational firms which have not paid any taxes, while they made billions of profits. Demonstrations of citizens made clear that this is a societal problem. Because tax is the second largest cost of firms in a lot of industries, it is likely that firms perform (legal) tax avoidance to reduce the taxes paid. In this thesis, the influence of periods of political uncertainty on the degree of political uncertainty is researched. Thereby, the influence of opportunistic behavior of managers on the relation between political uncertainty and tax avoidance is involved. This opportunistic behavior is illustrated by the degree of earnings management performed by managers.

To research this problem, two models are investigated. The first model contains the relationship between the degree political uncertainty firms experience and the extent a firm performs tax avoidance. Based on previous literature and the agency theory it was hypothesized that this relation could go either way. The traditional view of tax avoidance is that it adds value to the firm. But, managers expose themselves to tax risks, reputation risks, political risks and other risks. However, based on the empirical results, neither a positive nor a significant negative relationship is found. Hypothesis one is rejected. This potentially signals that the effects of political uncertainty on tax avoidance is different depending on managerial behavior.

The second model tested the moderating influence of opportunistic behavior of managers on the relationship between political uncertainty and tax avoidance. Based on previous literature, it was hypothesized that the association between political uncertainty and tax avoidance is likely to be more positive when managers are driven by self-serving behavior. The second model is significant, which means there is a positive significant relationship between political uncertainty and tax avoidance when opportunistic behavior is included in the model. Thus, earnings management is a conditional variable for the relationship. The interaction-effect of earnings management in the relationship is, in contrast to the hypothesis, negative significant. This means that more opportunistic managers perform less tax avoidance in case of political uncertainty in comparison to less opportunistic managers. A possible explanation for this effect can be that managers are more conservative and risk-averse in situations of political uncertainty (Wiseman and Gomez-Meija, 1998).

Besides the regular panel-data regressions, also some alternative regressions have been tested to determine the robustness of the results. These tests contained robustness tests and tests

(30)

with alternative measures for tax avoidance and political uncertainty. The results are different for the different measures. The relationships are more significant in case of the measure of Henry and Sansing (2018) when the measure of Desai and Dharmapala (2009) is used.

The research contains some practical and theoretical implications. Because the results indicate that political uncertainty does not influence the degree of tax avoidance, politicians have insight in an aspect of firm behavior in case of political uncertainty (e.g. elections or new policies). Moreover, for potential investors and shareholders, this research can give insights in the changing behavior of firms.

Besides the practical implications, this research contains also some theoretical contributions. The used measures of Hassan et al. (2019) and Henry and Sansing (2018) are relatively new measures to measure respectively political uncertainty and tax avoidance. The results indicate that the measure of Henry and Sansing (2018) gives different results than the generally used measure of Desai and Dharmapala (2009). This gives opportunities for other researchers to check the robustness of their results with this alternative measure. Furthermore, it contributes to the existing knowledge in the field of tax avoidance in relation to political uncertainty instead of tax avoidance in relation to investors or creditors.

Because every research contains limitations, there are some limitations identified for this paper. The first important limitation of this research is the scope. Because the sample only contains US-listed firms, it is not necessarily generalizable to other countries. For future research, this study could be replicated with another scope to compare the results and make the outcomes generalizable. The second identified limitation is that in previous literature it is already found that managers have more opportunities to perform earnings management in case of political uncertainty, because of the information asymmetry between the managers and the shareholders (Desai and Dharmapala, 2009). Even though the empirical results in this study do not show that the topics are related, this could be the case in practical situations. The last limitation contains the proxies for earnings management, because those might contain measurement errors. For future research, this study can be replicated with other measures, to determine the robustness of the results. An interesting topic for research in the future might be to include of awareness CSR in the relation between political uncertainty, opportunistic behavior and tax avoidance. Managers who are aware of CSR, are in my opinion less likely to perform opportunistic behavior because of their social awareness. This might change the results of the research.

(31)

7. References

Arrow, K.J. (1968). Economics of moral hazard – further comment. The American Economic Review, 58(3), 537-539.

Atwood, T., Drake, M., Myers, J., Myers, L. (2012). Home country tax system characteristics and corporate tax avoidance: international evidence. The Accounting Review, 87(6), 1831-1860. Belo, F., Gala V.D., Li, J., (2013). Government spending, political cycles, and the cross section of stock returns. Journal of Financial Economics, 107(2), 305-324.

Bloom, N., Bond, S., Van Reenen, J. (2007). Uncertainty and investment dynamics. Review of Economic Studies, 74(2), 391-415.

Blouin, J., Krull, L. (2009). Bringing it home: A study of the incentives surrounding the repatriation of foreign earnings under the American Jobs Creation Act of 2004. Journal of Accounting Research, 47(4), 1027-1059.

Bowen, R.M., Davis, A.K., Matsumoto, D.A. (2002). Do conference calls affect analysts’ forecasts? The Accounting Review, 77(2), 285-316.

Campbell, J.L. (2007). Why would corporations behave in socially responsible ways? An institutional theory of corporate social responsibility. Academy of Management Review, 32(3), 946-967.

Cen, W., Tong, N., Sun, Y. (2017). Tax avoidance and cost of debt: evidence from a natural experiment in China. Accounting and Finance, 57(5), 1517-1556.

Chan, K.H., Mo, P.L.L., Zhou, A.Y. (2013). Government ownership, corporate governance and tax aggressiveness: evidence from China. Accounting and Finance, 53(4), 1029-1051.

Christensen, S., Murphy, R. (2004). The social irresponsibility of corporate tax avoidance: taking CSR to the bottom line. Society for International development, 47(3), 37-44.

Crutchley, C.E., Hansen, R.S. (1989). A test of the agency theory of managerial ownership, corporate leverage, and corporate dividends. Financial Management, 18(4), 36-46.

Dechow, P.M., Skinner, D.J. (2000). Earnings management: reconciling the views of accounting academics, practitioners, and regulators. Accounting Horizon, 14(2), 235-250. Dechow, P. M., Sloan, R. G., and Sweeney, A. P. (1995). Detecting earnings management. The Accounting Review, 70 (2), 193-225.

Desai, M.A. and Dharmapala, D. (2009). Corporate tax avoidance and firm value. The Review of Economics and Statistics, 91(3), 537-546.

Desai, M.A. and Dharmapala, D. (2006). Corporate tax avoidance and high powered incentives. Journal of Financial Economics, 79(1), 145-179

Desai, M.A., and Dharmapala, D. (2009). Earnings Management, Corporate Tax Shelters, and Book-Tax Alignment, National Tax Journal, 62(1), 169-186.

(32)

Donaldson, L., Davis, J.H. (1991). Stewardship theory or agency theory: CEO governance and shareholder returns, Australian Journal of Management, 16(1), 49-64.

Dyreng, S.D., Hanlon, M., Maydew, E.L. (2008). Long-run corporate tax avoidance. The Accounting Review, 83(1), 61-82.

Dyreng, S.D., Hanlon, M., Maydew, E. L., and Thornock, J. R. (2014). Changes in corporate effective tax rates over the past twenty-five years. Working Paper. Retrieved from https://ssrn.com/abstract=2521497

Dyreng, S.D., Lindsey, B.P. (2009). Using financial accounting data to examine the effect of foreign operations located in tax havens and other countries on U.S. multinational firms’ tax rates. Journal of Accounting Research, 47(5), 1283-1316.

Eisenhardt, K.M. (1989). Making fast strategic decisions in high-velocity environments. The Academy of Management Journal, 32(3), 543-576.

El Diri, M. (2018). Motives of earnings management. In: Introduction to Earnings Management. Springer, Cham

Foley, C., Hartzell, J., Titman, S., Twite, G. (2007). Why do firms hold so much cash? A tax-based explanation. Journal of Financial Economics, 86(3), 579-607.

Gallemore, J., Maydew, E.L., Thornock, J.R. (2014). The reputational costs of tax avoidance. Contemporary Accounting Research, 31(4), 1103-1133.

Goh, B.W., Lee, J., Lim, C.Y., Shevlin, T. (2016). The effect of corporate tax avoidance on the cost of equity. The Accounting Review, 91(6), 1647-1670.

Graham, J., Tucker, A. (2006). Tax shelters and corporate debt policy. Journal of Financial Economics, 81, 563-594.

Guenther, D.A. (1994). Earnings management in response to corporate tax rate changes: Evidence from the 1986 Tax Reform Act. The Accounting Review, 69(1), 230-243.

Guenther, D.A., Matsunaga, S.R., Williams, B.M. (2017). Is tax avoidance related to firm risk? The Accounting Review, 92(1), 115-126.

Guorio, F., Siemer, M., Verdelhan, A. (2015). Uncertainty and international capital flows

(2014). Working Paper. Retrieved from:

https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2626635

Hassett, S.D. (2011). The risk premium factor: A new model for understanding the volatile forces that drive stock prices. Hoboken: John Wiley and Sons.

Hanlon, M., Heitzman, S. (2010). A review of tax research. Journal of Accounting and Economics, 50(2-3), 127-178.

Hanlon, M., Slemrod, J. (2009). What does tax aggressiveness signal? Evidence from stock price reactions to news about tax shelter involvement. Journal of Public Economics, 93(2), 126-141.

Does political uncertainty influence the degree of tax avoidance of firms?