Do former ministers enable higher levels of corporate tax avoidance? Evidence from the US

(1)

Do former ministers enable higher levels of corporate tax

avoidance? Evidence from the US

Jesse Blaauw

Student number: S3465500

MSc accountancy – University of Groningen

Dr. S. Mukherjee

Supervisor – University of Groningen

Date: 21-6-2019

Word count: 7.722

(2)

1 Abstract: This study examines the influence of former ministers who are now a director on the level of corporate tax avoidance in the US. I argue that, based on the resource dependency theory, former ministers enable a firm to get away with higher levels of corporate tax avoidance by providing regulatory cover over a firm. Further, I argue that this relationship is strengthened by dual CEOs, because both former ministers on the board and a dual CEO position increase the opportunities for a firm to engage in higher levels of corporate tax avoidance. My findings are consistent with my predictions. This study contributes to the literature on corporate tax avoidance and how significant political connections enable such practices.

Keywords: * ministers * political connections * corporate tax avoidance * CEO duality * US * book-tax difference * board of directors * resource dependency theory

(3)

2

Table of content

Introduction ... 3

Literature and hypotheses development ... 6

Methodology... 9

Results ... 13

Robustness test ... 15

Conclusion ... 16

(4)

3

Introduction

United States has a long-standing tradition of appointing former politicians to corporate boards. Prior research shows that on-balance political connections are valuable for a firm, since firms can use the former politicians’ network in their favor (Faccio, 2006; Hillman, 2005). Prior research also shows that boards have a significant influence on the levels of a firm’s corporate income tax (corporate tax) avoidance (Armstrong, Blouin, Jagolinzer and Larcker, 2015). Yet, the relationship between former politicians on a board and the level of corporate tax avoidance a firm undertakes is largely uncertain except for the study by Kim and Zhang (2016). In this study I examine whether former politicians, especially ministers, on a board enable higher levels of corporate tax avoidance. In addition, I will examine the influence of CEOs on the firm’s level of corporate tax avoidance when they are also chair of the board (CEO duality) especially when coupled with former ministers on the board.

The Kim and Zhang (2016) study shows that there is a positive and significant relationship between former government officials on a board and the level of corporate tax avoidance a firm undertakes. Their study used a wide range of government officials to determine whether a firm is politically connected. Faccio (2016) has identified some shortcomings of the study of Kim and Zhang (2016). She mainly criticized the treatment effects model used to estimate the impact of political connections on tax aggressiveness. Furthermore, she argued that the mechanisms behind the documented results are not thoroughly explained, since the Kim and Zhang (2016) study is explorative in nature. My study differentiates in multiple ways from the Kim and Zhang (2016) study. First, I will specifically focus on former politicians in ministerial positions (ministers) who are now a director1_{. I will focus on former ministers who are now a}

director for two reasons. First, politically connected boards have been less researched than politically connected executives. Second, former ministers had an active employment and a high position in the government (Bache and Flinders, 2004), suggesting that they had more influence than bureaucrats. Also, the study of Faccio (2007) suggests that corporate political connections are of greater value when they are with a minister, rather than with a member of Congress or a bureaucrat.

The resource dependency theory (RDT) developed by Pfeffer and Salancik (1978) offers an explanation for why firms appoint former ministers to their boards. According to the RDT, a firm is dependent on its external environment, which includes the government. By appointing a former minister to the board, the firm is linking itself with the government and can thereby reduce the risk and uncertainty caused by potential scrutiny of the government. Following the RDT, I argue that former ministers can help a firm to engage in higher levels of corporate tax avoidance by using their network in government to provide a regulatory cover over the firm (Faccio, 2006). Essentially, I posit that former ministers enable a firm to undertake higher levels of corporate tax avoidance by preventing potential scrutiny by the government.

CEOs are often the driving force in determining the level of corporate tax avoidance their firms undertake (Dyreng, Hanlon and Maydew, 2010). A study of Steijvers and Niskanen (2011) even suggested that CEOs are the decision makers regarding the level of corporate tax avoidance.

1_{Within this study the following positions are considered as ministerial positions: national or federal minister,}

deputy minister, assistant minister, shadow minister, state secretary, minister of state, assistant secretary of state, under-secretary, diplomats and governors.

(5)

4 Prior research has shown that firms with dual CEOs engage in higher levels of corporate tax avoidance due to the reduced interference of the board (Zhou, 2011). I posit that the relationship between former ministers on the board and the level of corporate tax avoidance will be strengthened by CEO duality, because a firm with former ministers on the board and a dual CEO position has even greater opportunities to engage in higher levels of corporate tax avoidance.

To investigate my predictions, I use US panel data compromising of industrial firms in the period 2005-2015 that were listed on the S&P 1500. I obtained the data from three sources, namely Worldscope, BoardEx and hand collection. The Worldscope database contains financial statement data. The BoardEx database contains data about the directors and whether they previously served as a minister. In addition, I hand-collected data on the former ministers to supplement the BoardEx database and to ensure that I have the current identification of the former ministers. After eliminating the firm-year observations that did not match the selection criteria, I am left with 18.651 firm-year observations to test both my predictions.

To test my predictions, I used the ordinary least squares regression with fixed effects. I used two variants of the book-tax difference as measures for the level of corporate tax avoidance. I focus on measures of the book-tax difference because prior studies have shown that the book-tax difference effectively captures the level of corporate tax avoidance a firm undertakes (Blaylock, Shevlin and Wilson, 2011). I used two measures of former ministers on the board, which are the percentage of former ministers on the board and a dummy variable which takes the value of 1 when there was at least one former minister on the board. In addition to the former minister measures, I included a measure for non-minister former government officials (public officers), which is the variable percentage public officers on the board. I included this variable in an apart model in the regression to examine whether the results differ between former ministers and former public officers. The public officers measure contains former senior government officials and former senior military officials who are now a director. To specifically examine the effect of former senior government officials and former senior military officials, I excluded former ministers from this variable. In the regression, I included control variables relating to CEO characteristics, firm characteristics and board characteristics. Consistent with my predictions, I found a positive and significant relationship between former ministers on the board and the level of corporate tax avoidance. Hereby providing evidence which shows that a firm with former ministers on the board engages in higher levels of corporate tax avoidance. In contrast, the relationship between former public officers on the board and the level of corporate tax avoidance is insignificant. This result suggests that former ministers can help a firm to engage in higher levels of corporate tax avoidance and that former public officers cannot. This result is consistent with the suggestion of Faccio (2007), who suggested that political connections are of greater value when the connection is with a minister rather than with a bureaucrat. To rule out the risk that my results are subject to endogeneity, I matched firms based on firm characteristics using a Propensity Score Matching (PSM) technique using its nearest-neighbor approach and found that the relationship between former ministers on the board and the level of corporate tax avoidance is still positive and statistically significant. Also, the relationship between former public officers on the board and the level of corporate tax avoidance remains insignificant in the PSM test. Overall, these

(6)

5 results provide strong evidence that only firms with former ministers on the board engage in higher levels of corporate tax avoidance and that not all political connections can help a firm to engage in higher levels of corporate tax avoidance. Second, I found that the relationship between former ministers on the board and the level of corporate tax avoidance is positive and significantly strengthened when the CEO is also chair of the board. This is evidence in support of the view that dual CEOs utilize former ministers’ political cover over the firm for corporate tax avoidance purposes. Result for the public officers are again insignificant. My research contributes to the literature regarding corporate tax avoidance and to the literature about corporate political connections. To the best of my knowledge, this is the first study that specifically focuses on former ministers who are now a director and their effect on the level of corporate tax avoidance a firm undertakes. Only the study of Kim and Zhang (2016) has empirically studied the relationship between political connected directors and the level of corporate tax avoidance, where one proxy for political connected directors was whether a director has previously served in a governmental position. As discussed, Kim and Zhang (2016) also found a positive and significant relationship between former government officials on the board and the level of corporate tax avoidance. I have extended the study of Kim and Zhang (2016) by specifically looking at former ministers on the board, by grounding my expectations based on the RDT and by showing that my results still hold in a propensity score matching test. I also contributed to the discussion of CEO duality and the consequences this might have for firms. I showed that former ministers on a board in combination with CEO duality results in even higher levels of corporate tax avoidance.

A practical contribution of my research results is that present governments should focus more on firms that have former ministers on their board, since these firms are more likely to engage in higher levels of corporate tax avoidance than firms without former ministers on the board. Governments should especially focus more on firms that have dual CEO positions and former ministers on the board, since these firms are likely to engage in even higher levels of corporate tax avoidance.

The remainder of this study is structured as follows. The second section will review the literature and develop hypotheses. The third section discusses the methodology of this study. The fourth section reports the results of the tests of my hypotheses. The fifth section presents the additional test performed in this study. The sixth and final section of this study contains the discussion of the results and the conclusion.

(7)

6

Literature and hypotheses development

Political connections and corporate tax avoidance

Corporate tax is a significant cost to firms. When firms engage in corporate tax avoidance activities, their after-tax earnings and cash flow will increase, which in turn can increase shareholders’ wealth (Khurana and Moser, 2012). But corporate tax avoidance activities can be risky for firms, Francis, Hasan, Wu and Yan (2014) argued that governments’ tax authorities are the most direct risk associated with corporate tax avoidance activities, since they can impose large penalties on firms. Therefore, when engaging in corporate tax avoidance activities, firms should reduce the risk faced from governments’ tax authorities. One possible way for firms to reduce this risk could be by having political connections through its corporate board.

Prior research focusing on the relationship between political connections of a firm and the government suggests that political connections can help a firm to reduce the risk faced from the government. For example, the study of Correia (2014) showed that political connected firms are less involved in enforcement actions of the Security and Exchange Commission (SEC) and face lower penalties when they are prosecuted by the SEC. Her research suggests that political connections can help influence the SEC in a way that is beneficial for a firm. Also, the cross-country study of Faccio, Masulis and McConnell (2006) document that politically connected firms are more likely to be bailed out when facing financial distress, thereby suggesting that politically connected firms get preferential government treatment regarding corporate bailouts.

A theory that might explain how political connections can bring value to a firm is the RDT. The RDT, developed by Pfeffer and Salancik (1978), argues that firms are dependent on external organizations, and that this dependency creates risk and uncertainty regarding the firms’ performance. According to the RDT, firms can reduce this risk and uncertainty resulting from this dependency by linking themselves with important sources of external dependency. Prior research argued that the government is a critical external dependency for firms (Hillman, 2005). The government can affect a firm’s competitive position, influence the firm’s future performance, they can establish entry and exit barriers, provide a firm with special tax treatments or subsidies, and can impact costs on employees and safety regulations (Lester, Hillman, Zardkoohi and Cannella, 2008). At the same time, many firms view the government as a very difficult external dependency to control because of the heterogeneous interests of different agencies within the government (Aharoni, Maimon and Segev, 1981). A connection with the government can help firms to buffer themselves against negative influences from the government (Hillman, 2005).

Politically connected boards and corporate tax avoidance

A convenient way for firms to link themselves with the government is by appointing former government officials to the board (Lester et al., 2008). Kim and Zhang (2016) have empirically studied the relationship between former government officials on the board and the level of corporate tax avoidance a firm undertakes. They found that firms with former government officials on their board engage in higher levels of corporate tax avoidance than firms that do not have former government officials on their board in the US. They argued that politically connected firms engage in higher levels of corporate tax avoidance because they have lower expected costs of tax enforcement, lower market pressure for transparency, an information

(8)

7 advantage about tax law enforcement changes and greater risk-taking tendencies because of politically network connections. Faccio (2016) discussed the Kim and Zhang (2016) study, she argued that the treatment effects model that was used to estimate the impact of political connections on the level of corporate tax avoidance could lead to endogenous results and that the mechanisms behind their results are not explained.

I posit that the RDT can help explain the effect former government officials could have on the firm’s tax avoidance practices through their presence on the board. The RDT posits that the board has, besides its monitoring role (Adams and Ferreira, 2007; Harris and Raviv, 2006; Raheja, 2005), an important function in providing resources to the firm (Ruigrok, Peck, Tacheva, Greve and Hu, 2006). Where a resource could be anything that can be thought of as a strength for a firm (Hillman and Dalziel, 2003). Former government officials who are now a director are likely to have valuable connections (resources) with current government officials. According to the RDT, former government officials who are now a director can use their network in government to reduce the regulatory scrutiny on the firm, and thereby reduce the risk and uncertainty faced from government departments such as the tax authorities. This argument is supported by Faccio (2006) who stated that political connections in a firm are associated with a more relaxed regulatory oversight. When there is a reduced regulatory scrutiny, a firm might undertake higher levels of corporate tax avoidance. This argument is consistent with the suggestion of Pfeffer and Salancik (2003) that firms attempt, through political mechanisms, to create an environment that is best for their own interests as well as to build political power in the external environment, which can be used in the future for the firm’s interests (Sun, Hu and Hillman, 2016).

Especially former ministers who are now a director can help a firm to reduce the regulatory scrutiny, since they had an active employment and a high position in the government (Bache and Flinders, 2004). So, they are likely to have a greater network in government. This argument is supported by Faccio (2007) who suggested that corporate political connections are of greater value when the connection is with a minister, rather than with a member of Congress or a bureaucrat. Hence, my first hypothesis is:

H1: Firms with former ministers on their board engage in higher levels of corporate tax avoidance.

Influence of dual CEOs

The relationship between former ministers on the board and the level of corporate tax avoidance might be influenced by influential CEOs. Research has shown that CEOs have incentives to engage in corporate tax avoidance activities (Dyreng et al., 2010; Rego and Wilson, 2012). Firms with dual CEOs might engage in even higher levels of corporate tax avoidance. Dual CEOs are charged with leading the firm in their CEO function and monitoring the firm’s activities in their board chair function. CEO duality is in contrast with the agency theory view, which suggests that boards should be independent of management to prevent managerial entrenchment (Krause, Semadeni and Cannella, 2014). Research focusing on the agency relation between dual CEOs and the board, showed that boards are less independent when CEOs are also chair of the board, since such CEOs have a strong influence on all aspects of the board’s decision-making (Ruigrok et al., 2006). Also, a study of Bliss (2011) argued that

(9)

8 CEOs can use their board chair position to influence other directors, which hampers their independent evaluations of corporate activities.

Previous research focusing on the relation between the degree of board independence and the level of corporate tax avoidance showed that more independent boards result in a firm engaging in lower levels of corporate tax avoidance (Armstrong et al., 2015; Lanis and Richardson, 2011; Richardson, Taylor and Lanis, 2013). This is consistent with the agency theory view, which suggests that boards should be independent of management to prevent managerial entrenchment (Krause et al., 2014). A study by (Zhou, 2011) has specifically focused on the relationship between dual CEOs and corporate tax avoidance. He argued that firms with dual CEOs engage in higher levels of corporate tax avoidance for three reasons. First, the dual CEO position will strengthen the relationship between the executives and the board and therefore the aggressive tax strategy could be more easily approved by the board. Second, the dual CEO role leads to a reduced oversight by the board, consistent with the agency theory view, which in turn results in higher levels of corporate tax avoidance. Third, since leadership of the firm is concentrated in one person, board members are unable or less likely to challenge proposals of corporate tax avoidance if they do not have sufficient knowledge about tax law.

In the previous section I have argued that, based on the RDT, former ministers on the board enable a firm to get away with higher levels of corporate tax avoidance by reducing the regulatory scrutiny on the firm. Building further on this prediction, I posit that CEO duality strengthens this relationship. Because CEO duality reduces the boards’ ability to challenge management about their corporate tax avoidance activities. Firms with both CEO duality and former ministers on their board can engage in even higher levels of corporate tax avoidance. Since the dual CEO position ensures that the board does not interfere with CEO’s incentives to engage in higher levels of corporate tax avoidance, which are enabled by the former ministers who provide the regulatory cover over the firm. Hence, my second hypothesis is: H2: The relationship between former ministers on the board and the level of corporate tax avoidance (H1) is strengthened if the CEO is also chair of the board.

(10)

9

Methodology

Sample and data

I use S&P 1500 panel data focusing on the period 2005–2015. I used three data sources for the empirical analysis. The first data source is a database obtained from Worldscope, which contains the financial statement data of the firms. The second data source is a database obtained from BoardEx, which contains the board data. As third data source, I used hand collected data to complement the BoardEx database. After merging both the datasets, I excluded observations with missing variables for the empirical analysis, and consistent with Kim and Zhang (2016), I excluded firms in the financial services and utilities industries. After cleaning the data, I am left with 18.651 firm-year observations. The observations per year are presented in table 1.

--- Insert table 1 here ---

Table 2 provides an overview of the proportion of observations with and without former ministers as director. As can be seen in the table, there are 207 observations in which a former minister is a director of a firm, which represents 1,11% of the total sample.

--- Insert table 2 here --- Variables

This subsection describes the dependent variables, independent variables and control variables.

Dependent variables

The focus of the dependent variable within this study is corporate tax avoidance. I focus on the book-tax difference with two variants. I use the book-tax difference because research has shown that the book-tax difference is strong and positively related with the level of corporate tax avoidance (Blaylock et al., 2011). The book-tax difference is defined as the difference between book income and taxable income (Goh, Lee, Lim and Shevlin, 2016). The book-tax difference represents corporate tax avoidance activities that generate both temporary and permanent differences (McGuire, Omer and Wang, 2012). Therefore, the book-tax difference captures a wide range of corporate tax avoidance activities.

The first variant of the book-tax difference I use is the book-tax difference used by Goh et al. (2016), which is presented in the formula below.

BTD = Pretax income – ( income taxes / statutory tax rate )

Where BTD represents the book-tax difference. For comparison purposes, the book-tax difference is scaled by lagged total assets. As can be seen in the equation, the pretax income represents the book income, and (income taxes / statutory tax rate) represents the taxable income.

The second variant of the book-tax difference I use focuses only on the income tax paid in the home country, which is the US in this study. So, it excludes income tax paid in foreign jurisdictions. The formula of this second variant of the book-tax difference is presented below.

(11)

10 BTD home country = Pretax income – ( income taxes paid in US / statutory tax rate )

Like the first variant of the book-tax difference, this second variant is also scaled by lagged total assets for comparison purposes.2

Independent variables

The independent variable in this study focuses on the former government officials who are now a director. This study will specifically focus on former government officials in ministerial positions. I considered ministerial positions as positions in which individuals were responsible for implementing government policies. So, I considered ministerial positions as: national or federal minister, deputy minister, assistant minister, shadow minister, state secretary, minister of state, assistant secretary of state, under-secretary, diplomats and governors. I manually checked the BoardEx dataset for correctness and consistency. I use two different measures for the ministers. The first measure for the ministers is the percentage ministers on the board. This variable is calculated by dividing the number of ministers on a board by the board size. The second measure for the ministers is the minister dummy, which takes the value of 1 if there was at least one former minister on the board and takes the value of 0 when there were no former ministers on the board.

Besides the measures of former ministers, I included the variable percentage public officers on the board to examine whether results are different for former ministers and former public officers. The public officers represent former senior government officials and former senior military officials. I excluded former ministers from the variable percentage public officers on the board, so that I am left with a variable that only captures former public officers and not ministers.

Control variables

In the regression model, several organizational control variables are included that may affect the level of corporate tax avoidance a firm undertakes. I included three different categories of control variables. First, control variables relating to CEO characteristics. Second, control variables relating to firm characteristics and performance. And third, control variables relating to board characteristics.

To control for CEO characteristics, I included three control variables in the model. The first control variable is CEO duality, since CEO duality is associated with higher levels of corporate tax avoidance (Halioui, Neifar, and Abdelaziz, 2016). Second, I included a control variable for CEO change to control for individual CEO effects, since research has shown that the effective tax rate can change significantly around the year of executive hire and departure (Dyreng et al., 2010). Third, I control for CEO gender, to rule out gender differences between males and females (Blaufus, Hundsdoerfer, Jacob, and Sünwoldt, 2016; Dyreng et al., 2010).

2_{Consistent with Goh et al. (2016), I will focus on the book-tax difference because it captures a broad spectrum}

of corporate tax avoidance activities. I do not use the effective tax rate (ETR) as a measure of corporate tax avoidance, since research has shown that the ETR can vary significantly per year (Wahab and Holland, 2015), which could negatively influence my results because I measure corporate tax avoidance over a period of 11 years.

(12)

11 To control for firm characteristics and performance, I included eleven control variables in the model. First, I control for the market valuation of the firm by using the control variable Tobin’s-q, because Desai and Dharmapala (2009) showed that firms that are well-governed experience a positive relationship between corporate tax avoidance and firm value. The second control variable is total assets, which controls for the firm size. Third and fourth, I control for firm performance by including the variables return on assets (ROA) and a loss dummy, where the loss dummy takes the value of 1 if ROA is below zero. I include these two variables because research has shown that a firm’s performance does influence the level of corporate tax avoidance (Rego and Wilson, 2012). Fifth, I included the variable total liabilities divided by total assets, which controls for the way a firm is financed. The sixth control variable is the debt interest rate, I included this variable to control for excessive interest rates, since interest payments are tax deductible and could provide a way to avoid corporate tax (Lanis and Richardson, 2015). Seventh, I included the ratio of capital expenditures divided by total assets, because other studies focusing on corporate tax avoidance also included this control variable (Dyreng et al., 2010; Rego, 2003). Eight, I control for the ratio research and development (R&D) divided by total assets, since firms can use the capitalization of R&D expenses as a way to avoid corporate tax (Goh et al., 2016). Ninth, consistent with McGuire et al. (2012) I control for the cash needs of a firm, which might motivate tax deferral activities. Tenth, I included the variable closely held shares to control for differences in ownership structure. And eleventh, I control for the firm’s foreign assets, since many firms use foreign entities to avoid corporate taxes in the US (Rego, 2003).

I also included six control variables relating to board characteristics. The first control variable is board size, since research has shown that board size can influence the amount of corporate tax avoidance (Zemzem and Ftouhi, 2013). The second control variable focuses on the amount of non-executive directors, who are defined as more independent than executive directors, I included this control variable because research has shown that more independent boards attenuate extreme levels of corporate tax avoidance (Armstrong et al., 2015). Third, I control for the amount of female directors, since females are generally associated with a lower risk taking preference than men (Dwyer, Gilkeson, List, 2002), and since a study of Francis et al. (2014) showed that female chief financial officers are associated with lower levels of corporate tax avoidance than males. Fourth, I control for the amount of foreign directors on the board, since foreign directors can weaken the ability of a board to monitor a firm due to the use of non-native languages by directors (Hooghiemstra, Hermes, Oxelheim, Randøy, 2019). Fifth, I control for director tenure, since short tenured directors have less knowledge about the firm, and long tenured directors may become too friendly with management, they are likely to have an influence on board effectiveness (Brown, Anderson, Salas and Ward, 2017). And sixth, I control for the number of other quoted board affiliations of directors, since monitoring management’s actions takes time of a director, other board affiliations may increase time pressure on individual directors and thereby reducing his monitoring ability (Lanis and Richardson, 2011).

Empirical model

To test my first hypothesis, I use the following fixed effects regression model. Model 1: TAXAVOIDi,t, = a0 + β1FMi,t, + βnCONTROLSi,t, + Ɛi,t,

(13)

12 Where TAXAVOIDi,t, are the two variants of the book-tax difference of firm i in year t, β1FMi,t,

represents the three measures of former government officials on the board of firm i in year t, which are the percentage ministers on the boards, a minister dummy which takes the value of 1 if at least one former minister was on the board in a given year and 0 otherwise, and the percentage of public officers on the board. I expect a positive relationship between former ministers on the board and the level of corporate tax avoidance, therefore I expect the variables percentage ministers on the board and the minister dummy to be positive. βnCONTROLSi,t represents all the control variables of firm i in year t, and Ɛi,t, is the error term

of firm i in year t.

To test my second hypothesis, I use the following fixed effects regression model.

Model 2: TAXAVOIDi,t, = a0 + β1FMi,t, + β2CEODUALITYi,t, + β3FMi,t, * CEODUALITYi,t, +

βnCONTROLSi,t, + Ɛi,t,

Where β3FMi,t, * CEODUALITYi,t, is the interaction variable between the three measures of

former government officials on the board and CEO duality of firm i in year t. I expect this interaction variable to be positive in the interaction with the percentage ministers on the board and the minister dummy, because I expect that CEO duality strengthens the relationship between former ministers on the board and the level of corporate tax avoidance.

I also included year fixed effects and industry fixed effects in the regression, the industry fixed effects is a dummy variable for the 48 industries specified by Fama and French (1997).

In the analysis all variables, except the independent variables, are winsorized at 1% to reduce the effect of outliers. The independent variables are not winsorized because they are either a dummy variable or not normally distributed. All independent and control variables are lagged one year in the regression model.

(14)

13

Results

Descriptive statistics

Table 3 below presents the descriptive statistics. This table shows that both means of the book-tax difference variables are negative. The Means of the two variants of the book-tax difference are close to other studies in the area (Frank, Lynch and Rego, 2009; Goh et al., 2016). Furthermore, the mean of CEO duality is 49%, which means that 49% of the firm-year observations have dual CEOs.

Table 4 presents the Pearson correlation matrix. The two variants of the book-tax difference are highly correlated, which was expected. Moreover, the two measures of former ministers, the percentage ministers on the board and the minister dummy, are highly correlated, which was also expected.

--- Insert table 4 here --- Multivariate analyses

This section shows the empirical results of the tests of my hypotheses. Both hypotheses are tested with a firm fixed-effects panel data regression and I included year and industry dummies. All p-values in the table are two-tailed, unless stated otherwise, and are based on robust standard errors. For both tests of the hypotheses, I included six models. Model 1 focuses on the first variant of the book-tax difference and the percentage ministers on the board. Model 2 focuses on the second variant of the book-tax difference and the percentage ministers on the board. Model 3 focuses on the first variant of the book-tax difference and the minister dummy. Model 4 focuses on the second variant of the book-tax difference and the minister dummy. Model 5 focuses on the first variant of the book-tax difference and the percentage public officers on the board. Model 6 focuses on the second variant of the book-tax difference and the percentage public officers on the board.

Test of H1

Table 5 shows the regression results of hypothesis 1, which tests the association between former ministers who are now a director and the level of corporate tax avoidance a firm undertakes. As can be seen in table 5, all four models focusing on former ministers show a significant and positive relationship between former ministers on the board and the book-tax difference. Model 1 is significant at the 10% level using a one-tailed p-value, model 2 and model 4 are significant at the 5% level, and model 3 is significant at the 10% level. These results strongly support my first hypothesis that firms with former ministers on their board engage in higher levels of corporate tax avoidance than firms without former ministers on their board, since a higher level of the book-tax difference represents higher levels of corporate tax avoidance. The variable percentage public officers on the board is insignificant in model 5 and model 6, suggesting that former public officers on the board do not have a strong influence on the level of corporate tax avoidance a firm undertakes. Further, some control variables show a significant relationship with the book-tax difference. First, Tobin’s-q shows a very significant and positive relationship with the book-tax difference, which suggests that firms with a relatively high market value have higher book-tax differences (i.e. higher levels of corporate tax avoidance). Second, total assets are significant and negatively associated with

(15)

14 the book-tax difference, which suggests that bigger firms have a lower book-tax difference, which is consistent with prior research (Goh et al., 2016; McGuire et al., 2012). Third, ROA is strongly significant and positively associated with the book-tax difference, suggesting that firms with better financial performance have bigger book-tax differences, which is also consistent with prior research (Kim and Zhang, 2016; McGuire et al., 2012). Fourth, the loss dummy is strongly significant and negatively associated with the book-tax difference. The loss dummy takes the value of 1 when ROA is below zero, so the strong significant and negative relationship coincides with the strong significant and positive relationship between ROA and the book-tax difference. Fifth, the ratio of total liabilities divided by total assets (i.e. leverage) shows a significant and positive relationship with the book-tax difference, which is also consistent with prior research (Kim and Zhang, 2016; McGuire et al., 2012). Sixth, the debt interest rate shows a significant and negative relationship with the book-tax difference, suggesting that firms with lower interest rates have higher book-tax differences. Seventh, closely held shares is weak significant and negatively associated with the book-tax difference in three of the six models. Eight, board size is very significant and negatively associated with the tax difference, suggesting that firms with larger boards have a lower level of book-tax difference. Ninth, I find a weak significant and positive relationship between the amount of females on the board and the book-tax difference. This is contrary to prior research, which suggests that females are more risk averse than males (Dwyer et al., 2002). And lastly, the number of other quoted board affiliations of directors shows a significant and negative relationship with the book-tax difference, suggesting that directors with multiple board appointments reduce the level of the book-tax difference. 3

Test of H2

Table 6 shows the regression results of hypothesis 2, which tests whether the relationship between former ministers on the board and the level of corporate tax avoidance is strengthened when the CEO is also chair of the board. The results show a significant and positive relationship between the interaction variable percentage ministers on the board with CEO duality and a significant and positive relationship between the interaction variable minister dummy with CEO duality, all four models at the 5% significance level. Thereby providing strong evidence in support of hypothesis 2. Notable is that the variable CEO duality is insignificant by itself, thereby suggesting that dual CEOs cannot avoid corporate tax by themselves, but that that they need the regulatory cover provided by the former ministers to engage in higher levels of corporate tax avoidance. Contrary to these results, the interaction variable percentage public officers on the board with CEO duality is insignificant. The control variables in the regression show roughly the same outcomes as in the test of H1.

3_{In addition to the regression in table 5, I also performed a difference-in-difference test. Where a firm is in the}

treatment sample if it ever had at least one former minister on their board during the sample period 2005-2015. The results of the difference-in-difference test were insignificant. Thereby suggesting that firms with and without former ministers on their board do not significantly differ with respect to the level of corporate tax avoidance. Therefore, the significant result of H1 is robust and can be attributed to the time where a former minister was part of the board.

(16)

15

Robustness test

The results in the previous section showed that there is a significant and positive relationship between former ministers on the board and the level of corporate tax avoidance. This result was found by using a multivariate regression analysis. To rule out the risk that the relationship between former ministers on the board and the level of corporate tax avoidance is endogenous, I performed a nearest-neighbor PSM test, where firms that have similar characteristics are matched based on their propensity scores.

I used the PSM test to match firms with former ministers on their board with firms that do not have former ministers on their board based on firm-specific characteristics. I used a logit regression, with the variables total assets, board size, Tobin’s-q, CEO duality, debt interest rate and capital expenditures divided by total assets, to estimate the propensity score of each firm in the sample. Then I matched the firms in my treatment sample (firms with former ministers on their board) with firms in the control sample (firms without former ministers on their board) using the nearest-neighbor matching method without replacement and within a caliper of 0,01. The results of the PSM logit regression are based on a sample of 408 observations and are presented in table 7.

To assess whether the significant and positive relationship between former ministers on the board and the level of corporate tax avoidance will still hold after firms are matched based on their propensity scores, I performed the same multivariate regression as for hypothesis 1, but this time the regression is based on the propensity scores of the firms. The results of the propensity score regression are presented in table 8 below.

Table 8 shows that there still is a significant and positive relationship between former ministers on the board and the level of corporate tax avoidance after applying the PSM test. If I compare table 5 to table 8, I see that model 1 loses its significance, model 2 is significant at the 5% level in both tables, model 3 is significant in table 5 at the 10% level and insignificant in table 8, and model 4 is significant in table 5 at the 5% level and significant in table 8 at the 10% level. Therefore, I find that the relationship between former ministers on the board and the level of corporate tax avoidance is still significant in the PSM test, but that there is a lower level of significance compared to the results in the multivariate regression in table 5.

Quality of the match

To ensure the results of the PSM test are not due to inadequate matching of variables, I performed a propensity score t-test to assess the balancing between the treatment and control sample. The results of the propensity score t-test are presented in table 9.

The outcomes of the t-test in table 9 show that no matching variable is significant, which means that the quality of the match was good. Which in turn increases the validity of the results presented in table 8.

(17)

16

Conclusion

This study investigated the relationship between former ministers on the board and the level of corporate tax avoidance a firm undertakes. I found that former ministers on a board are associated with a higher book-tax difference. So, I found support for my hypothesis which states that firms with former ministers on their board engage in higher levels of corporate tax avoidance than firms without former ministers on their board. I argued, based on the RDT, that former ministers can help a firm to link itself with the government, and that the former ministers can help to provide a regulatory cover over the firm by using their network in government. This reduced regulatory scrutiny could help the firm with getting away with higher levels of corporate tax avoidance. This result remained significant after a difference-in-difference test and a PSM test. In contrast, I found that public officers on the board are not associated with higher levels of corporate tax avoidance.

Further, I found that the relationship between former ministers on the board and corporate tax avoidance is strengthened when the CEO is also chair of the board. I argued that firms with both former ministers on their board and a dual CEO position have even better opportunities to engage in higher levels of corporate tax avoidance. The former ministers are used to provide the regulatory cover over the firm and the dual CEO is used to ensure the board does not interfere with the corporate tax avoidance activities.

The findings in this study advance the literature about corporate tax avoidance and politically connected boards in three ways. First, I contribute to the discussion about corporate tax avoidance and the characteristics of firms that engage in corporate tax avoidance. My study shows partial support for the study of Kim and Zhang (2016), because I showed that specifically the former ministers enable a firm to get away with higher levels of corporate tax avoidance, since the public officers showed an insignificant relationship with the level of corporate tax avoidance. Second, I showed that CEO duality strengthens the relationship between former ministers on the board and the level of corporate tax avoidance. This result contributes to the ongoing discussion about CEO duality by providing evidence of the possible consequences of CEO duality. Third, I grounded my expectations based on the RDT. No prior research has integrated the RDT in the relationship between former senior government officials on the board and corporate tax avoidance. I integrated the RDT in the relationship and argued that former ministers provide a regulatory cover over firms which in turn enables them with getting away with higher levels of corporate tax avoidance.

My study has some implications for practice. Given the research results, tax authorities should more extensively focus on firms that appoint former ministers to their board, since firms with former ministers on their board are associated with higher levels of corporate tax avoidance. Second, shareholders should critically evaluate the power of the CEO when he is also chair of the board. Since dual CEOs in combination with former ministers on the board results in even higher levels of corporate tax avoidance.

My research has some limitations. First, my study only focuses on US firms, so the results might not be generalizable to other countries. Second, this study only focuses on S&P 1500 firms, therefore my results might not be generalizable to US firms who are not in the S&P 1500.

(18)

17 Further research should more closely focus on the mechanisms, through relationships, that underly the relationship between former ministers on the board and the level of corporate tax avoidance a firm undertakes. Research could for example investigate through which connections former minsters can provide the regulatory cover over the firm. Also, research should focus on the influence of the former minister’s political orientation and the level of corporate tax avoidance a firm undertakes, since differences in political beliefs may has an influence of the level of corporate tax avoidance. In addition, research should use more measures for corporate tax avoidance than this study, so that results are stronger. Since a variety of corporate tax avoidance measures captures different types of corporate tax avoidance activities (Desai and Dharmapala, 2009).

(19)

18

References

Adams, R. B., & Ferreira, D. (2007). A theory of friendly boards. The journal of finance, 62(1), 217-250.

Aharoni, Y., Maimon, Z., & Segev, E. (1981). Interrelationships between environmental dependencies: A basis for tradeoffs to increase autonomy. Strategic Management Journal, 2(2), 197-208.

Armstrong, C., Blouin, J., Jagolinzer, A., & Larcker, D. (2015). Governance, Incentives, and Tax Avoidance. Working Paper, University of Pennsylvania, University of Colorado, Stanford University.

Bache, I., & Flinders, M. (2004). Multi-level governance and the study of the British state. Public policy and administration, 19(1), 31-51.

Blaufus, K., Hundsdoerfer, J., Jacob, M., & Sünwoldt, M. (2016). Does legality matter? The case of tax avoidance and evasion. Journal of Economic Behavior & Organization, 127, 182-206. Blaylock, B., Shevlin, T., & Wilson, R. J. (2011). Tax avoidance, large positive temporary book-tax differences, and earnings persistence. The Accounting Review, 87(1), 91-120.

Bliss, M. A. (2011). Does CEO duality constrain board independence? Some evidence from audit pricing. Accounting & Finance, 51(2), 361-380.

Brown, J. A., Anderson, A., Salas, J. M., & Ward, A. J. (2017). Do investors care about director tenure? Insights from executive cognition and social capital theories. Organization Science, 28(3), 471-494.

Correia, M. M. (2014). Political connections and SEC enforcement. Journal of Accounting and Economics, 57(2-3), 241-262.

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

Dwyer, P. D., Gilkeson, J. H., & List, J. A. (2002). Gender differences in revealed risk taking: evidence from mutual fund investors. Economics Letters, 76(2), 151-158.

Dyreng, S. D., Hanlon, M., & Maydew, E. L. (2010). The effects of executives on corporate tax avoidance. The Accounting Review, 85(4), 1163-1189.

Faccio, M., Masulis, R. W., & McConnell, J. J. (2006). Political connections and corporate bailouts. The Journal of Finance, 61(6), 2597-2635.

Faccio, M. (2006). Politically Connected Firms. The American Economic Review, 96(1), 369-386.

(20)

19 Faccio, M. (2007). The characteristics of politically connected firms.

Faccio, M. (2016). Discussion of “Corporate Political Connections and Tax Aggressiveness”. Contemporary Accounting Research, 33(1), 115-120.

Fama, E. F., & French, K. R. (1997). Industry costs of equity. Journal of financial economics, 43(2), 153-193.

Francis, B. B., Hasan, I., Wu, Q., & Yan, M. (2014). Are female CFOs less tax aggressive? Evidence from tax aggressiveness. The Journal of the American Taxation Association, 36(2), 171-202.

Frank, M. M., Lynch, L. J., & Rego, S. O. (2009). Tax reporting aggressiveness and its relation to aggressive financial reporting. The Accounting Review, 84(2), 467-496.

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.

Halioui, K., Neifar, S., & Ben Abdelaziz, F. (2016). Corporate governance, CEO compensation and tax aggressiveness: Evidence from American firms listed on the NASDAQ 100. Review of Accounting and Finance, 15(4), 445-462.

Harris, M., & Raviv, A. (2006). A theory of board control and size. The Review of Financial Studies, 21(4), 1797-1832.

Hillman, A. J., & Dalziel, T. (2003). Boards of directors and firm performance: Integrating agency and resource dependence perspectives. Academy of Management review, 28(3), 383-396.

Hillman, A. J. (2005). Politicians on the board of directors: Do connections affect the bottom line?. Journal of Management, 31(3), 464-481.

Hooghiemstra, R., Hermes, N., Oxelheim, L., & Randøy, T. (2019). Strangers on the board: The impact of board internationalization on earnings management of Nordic firms. International Business Review, 28(1), 119-134.

Khurana, I. K., & Moser, W. J. (2012). Institutional shareholders' investment horizons and tax avoidance. The Journal of the American Taxation Association, 35(1), 111-134.

Kim, C., & Zhang, L. (2016). Corporate political connections and tax aggressiveness. Contemporary Accounting Research, 33(1), 78-114.

Krause, R., Semadeni, M., & Cannella Jr, A. A. (2014). CEO duality: A review and research agenda. Journal of Management, 40(1), 256-286.

Lanis, R., & Richardson, G. (2011). The effect of board of director composition on corporate tax aggressiveness. Journal of Accounting and Public Policy, 30(1), 50-70.

(21)

20 Lanis, R., & Richardson, G. (2015). Is corporate social responsibility performance associated with tax avoidance?. Journal of Business Ethics, 127(2), 439-457.

Lester, R. H., Hillman, A., Zardkoohi, A., & Cannella Jr, A. A. (2008). Former government officials as outside directors: The role of human and social capital. Academy of Management Journal, 51(5), 999-1013.

McGuire, S. T., Omer, T. C., & Wang, D. (2012). Tax avoidance: Does tax-specific industry expertise make a difference?. The Accounting Review, 87(3), 975-1003.

Pfeffer, J., & Salancik, G. R. (1978). The external control of organizations: A resource dependence perspective. New York: Harper & Row.

Pfeffer, J., & Salancik, G. R. (2003). The external control of organizations: A resource dependence perspective. Stanford University Press.

Raheja, C. G. (2005). Determinants of board size and composition: A theory of corporate boards. Journal of financial and quantitative analysis, 40(2), 283-306.

Rego, S. O., & Wilson, R. (2012). Equity risk incentives and corporate tax aggressiveness. Journal of Accounting Research, 50(3), 775-810.

Rego, S. O. (2003). Tax‐avoidance activities of US multinational corporations. Contemporary Accounting Research, 20(4), 805-833.

Richardson, G., Taylor, G., & Lanis, R. (2013). The impact of board of director oversight characteristics on corporate tax aggressiveness: An empirical analysis. Journal of Accounting and Public Policy, 32(3), 68-88.

Ruigrok, W., Peck, S., Tacheva, S., Greve, P., & Hu, Y. (2006). The determinants and effects of board nomination committees. Journal of Management & Governance, 10(2), 119-148. Steijvers, T., & Niskanen, M. (2011). Tax Aggressive Behaviour in Private Family Firms-the effect of the CEO and board of directors. Proceedings of the 7th European Conference on Management, Leadership and Governance, 379-386.

Sun, P., Hu, H. W., & Hillman, A. J. (2016). The dark side of board political capital: Enabling blockholder rent appropriation. Academy of Management Journal, 59(5), 1801-1822.

Wahab, N. S. A., & Holland, K. (2015). The persistence of book-tax differences. The British Accounting Review, 47(4), 339-350.

Zemzem, A., & Ftouhi, K. (2013). The effects of board of directors’ characteristics on tax aggressiveness. Research Journal of Finance and Accounting, 4(4), 140-147.

(22)

21 Zhou, Y. (2011). Ownership structure, board characteristics, and tax aggressiveness. Working paper, Lingnan University

(23)

22 Table 1:

Observations per year

year Freq. Percent Cum.

2005 1,691 9.07 9.07 2006 1,831 9.82 18.88 2007 1,798 9.64 28.52 2008 1,609 8.63 37.15 2009 1,686 9.04 46.19 2010 1,661 8.91 55.10 2011 1,795 9.62 64.72 2012 1,822 9.77 74.49 2013 1,537 8.24 82.73 2014 1,595 8.55 91.28 2015 1,626 8.72 100.00 Total 18,651 100.00

(24)

23 Table 2:

Observation with and without former ministers

Minister dummy Freq. Percent Cum. No minister director 18,444 98.89 98.89 Minister director 207 1.11 100.00

(25)

24 Table 3:

Descriptive statistics

Variable Obs Mean Std. Dev. Min Max

Book-tax difference 18,651 -0.039 0.175 -0.999 0.268

Book-tax difference home 18,651 -0.027 0.180 -0.997 0.283

Percentage ministers on board 18,651 0.001 0.012 0 0.167

Minister dummy 18,651 0.012 0.108 0 1

Percentage public officers on board 18,651 0.163 0.156 0 1

CEO duality 18,651 0.492 0.500 0 1

CEO change 18,651 0.088 0.283 0 1

CEO gender dummy 18,651 0.031 0.173 0 1

Tobins-q 18,651 2.073 1.486 0.491 9.555

Total assets 18,651 6.378 2.020 0.810 12.169

ROA 18,651 0.042 0.177 -0.863 0.352

Loss dummy 18,651 0.238 0.426 0 1

Total liabilities / Total assets 18,651 0.459 0.217 0.008 1

Debt interest rate 18,651 0.024 0.024 0 0.142

Capital expenditures / Total assets 18,651 0.052 0.062 0 0.349

R&D / Total Assets 18,651 0.046 0.097 0 0.753

Cash / Total assets 18,651 0.210 0.224 0 0.936

Closeheldshares 18,651 0.220 0.216 0.001 0.934

Foreign assets / Total assets 18,651 0.091 0.180 0 1

Boardsize 18,651 2.073 0.269 1.099 2.890

Non-executive directors / Boardsize 18,651 0.747 0.135 0 0.944

Female director / Boardsize 18,651 0.089 0.099 0 0.429

Foreign director / Boardsize 18,651 0.035 0.081 0 0.857

Director board tenure / Boardsize 18,651 1.069 0.706 0 3.536

Other quoted board affiliations / Boardsize 18,651 0.217 0.077 0 0.875

(26)

25 Table 4:

Pearson correlation matrix

1 2 3 4 5 6

(1) Book-tax difference 1.000

(2) Book-tax difference home 0.993 1.000

(3) Percentage ministers on board 0.012 0.016 1.000

(4) Minister dummy 0.016 0.022 0.967 1.000

(5) Percentage public officers on board 0.059 0.067 0.113 0.124 1.000

(6) CEO duality 0.089 0.093 0.037 0.042 0.086 1.000

(27)

26 Table 5:

Test of H1

Model 1 Model 2 Model 3 Model 4 Model 5 Model 6

BTD BTD home BTD BTD home BTD BTD home

Percentage ministers on board 0.215§ 0.275**

(0.133) (0.137)

Minister dummy 0.023* 0.029**

(0.012) (0.013)

Percentage public officers on board -0.029 -0.028

(0.018) (0.019)

CEO duality -0.003 -0.001 -0.003 -0.001 -0.003 -0.001

(0.004) (0.004) (0.004) (0.004) (0.004) (0.004)

CEO change 0.001 0.002 0.001 0.002 0.001 0.002

(0.004) (0.004) (0.004) (0.004) (0.004) (0.004)

CEO gender dummy -0.017 -0.017 -0.017 -0.018 -0.017 -0.018

(0.012) (0.012) (0.012) (0.012) (0.012) (0.012) Tobins-q 0.013*** 0.014*** 0.013*** 0.014*** 0.013*** 0.014*** (0.002) (0.002) (0.002) (0.002) (0.002) (0.002) Total assets -0.022*** -0.022*** -0.022*** -0.022*** -0.022*** -0.022*** (0.004) (0.004) (0.004) (0.004) (0.004) (0.004) ROA 0.296*** 0.303*** 0.296*** 0.303*** 0.296*** 0.303*** (0.027) (0.027) (0.027) (0.027) (0.027) (0.027) Loss dummy -0.015*** -0.015*** -0.015*** -0.015*** -0.014*** -0.015*** (0.004) (0.004) (0.004) (0.004) (0.004) (0.004)

Total liabilities / Total assets 0.171*** 0.174*** 0.172*** 0.174*** 0.172*** 0.174***

(0.014) (0.014) (0.014) (0.014) (0.014) (0.015)

Debt interest rate -0.202* -0.231** -0.202* -0.231** -0.205* -0.234**

(0.107) (0.110) (0.107) (0.110) (0.107) (0.110)

Capital expenditures / Total assets 0.004 -0.000 0.004 -0.000 0.004 -0.000

(0.033) (0.033) (0.033) (0.033) (0.033) (0.033)

R&D / Total Assets 0.050 0.052 0.050 0.052 0.051 0.053

(0.057) (0.056) (0.057) (0.056) (0.057) (0.056)

Cash / Total assets 0.012 0.010 0.012 0.010 0.011 0.009

(0.017) (0.017) (0.017) (0.017) (0.017) (0.017)

Closeheldshares -0.014 -0.015* -0.014 -0.015* -0.014 -0.016*

(0.009) (0.009) (0.009) (0.009) (0.009) (0.009)

Foreign assets / Total assets -0.009 -0.003 -0.009 -0.003 -0.009 -0.003

(0.009) (0.009) (0.009) (0.009) (0.009) (0.009)

Boardsize -0.043*** -0.041*** -0.043*** -0.041*** -0.041*** -0.039***

(0.014) (0.014) (0.014) (0.014) (0.014) (0.014)

Non-executive directors / Boardsize -0.018 -0.017 -0.018 -0.017 -0.017 -0.016

(0.018) (0.018) (0.018) (0.018) (0.018) (0.018)

Female director / Boardsize 0.044* 0.050* 0.044* 0.050* 0.044* 0.050*

(0.026) (0.026) (0.026) (0.026) (0.026) (0.026)

Foreign director / Boardsize 0.021 0.030 0.020 0.029 0.021 0.031

(0.032) (0.032) (0.032) (0.032) (0.032) (0.032)

Director board tenure / Boardsize -0.001 0.000 -0.001 -0.000 -0.000 0.001

(28)

27

Other quoted board affiliations / Boardsize -0.070** -0.059* -0.070** -0.059* -0.065* -0.055

(0.035) (0.036) (0.035) (0.036) (0.035) (0.035) Business segments 0.004 0.004 0.004 0.004 0.003 0.004 (0.003) (0.003) (0.003) (0.003) (0.003) (0.003) Geographic Segments 0.004 0.007* 0.004 0.007* 0.004 0.007* (0.004) (0.004) (0.004) (0.004) (0.004) (0.004) Constant 0.098* 0.097* 0.099* 0.098* 0.096* 0.095* (0.054) (0.053) (0.054) (0.053) (0.054) (0.053)

Year fixed effects Yes Yes Yes Yes Yes Yes

Industry fixed effects Yes Yes Yes Yes Yes Yes

Observations 18,651 18,651 18,651 18,651 18,651 18,651

R-squared 0.110 0.113 0.110 0.113 0.110 0.113

Number of firm_id 3,330 3,330 3,330 3,330 3,330 3,330

Robust standard errors in parentheses Two-Tailed: *** p<0.01, ** p<0.05, * p<0.1;

(29)

28 Table 6:

Test of H2

Percentage ministers on board -0.064 -0.016

(0.156) (0.153)

Percentage ministers on board * CEO duality 0.403** 0.420**

(0.186) (0.182)

Minister dummy -0.007 -0.002

(0.017) (0.016)

Minister dummy * CEO duality 0.040** 0.041**

(0.020) (0.019)

(0.023) (0.024)

Percentage public officers on board * CEO duality -0.015 -0.008

(0.019) (0.019)

CEO duality -0.003 -0.002 -0.003 -0.002 -0.000 0.000

(0.004) (0.004) (0.004) (0.004) (0.006) (0.006)

CEO change 0.001 0.002 0.001 0.002 0.001 0.002

(0.004) (0.004) (0.004) (0.004) (0.004) (0.004)

CEO gender dummy -0.017 -0.017 -0.017 -0.017 -0.017 -0.018

(0.012) (0.012) (0.012) (0.012) (0.012) (0.012) Tobins-q 0.013*** 0.014*** 0.013*** 0.014*** 0.013*** 0.014*** (0.002) (0.002) (0.002) (0.002) (0.002) (0.002) Total assets -0.022*** -0.022*** -0.022*** -0.022*** -0.022*** -0.022*** (0.004) (0.004) (0.004) (0.004) (0.004) (0.004) ROA 0.296*** 0.303*** 0.296*** 0.303*** 0.295*** 0.303*** (0.027) (0.027) (0.027) (0.027) (0.027) (0.027) Loss dummy -0.014*** -0.015*** -0.014*** -0.015*** -0.014*** -0.015*** (0.004) (0.004) (0.004) (0.004) (0.004) (0.004)

Total liabilities / Total assets 0.172*** 0.175*** 0.172*** 0.175*** 0.171*** 0.174***

(0.014) (0.014) (0.014) (0.014) (0.014) (0.015)

Debt interest rate -0.202* -0.232** -0.202* -0.232** -0.204* -0.234**

(0.107) (0.110) (0.107) (0.110) (0.107) (0.110)

Capital expenditures / Total assets 0.003 -0.001 0.004 -0.001 0.004 -0.000

(0.033) (0.033) (0.033) (0.033) (0.033) (0.033)

R&D / Total Assets 0.050 0.052 0.050 0.052 0.051 0.053

(0.057) (0.056) (0.057) (0.056) (0.057) (0.056)

Cash / Total assets 0.012 0.010 0.012 0.010 0.011 0.009

(0.017) (0.017) (0.017) (0.017) (0.017) (0.017)

Closeheldshares -0.014 -0.015* -0.014 -0.015* -0.014 -0.016*

(0.009) (0.009) (0.009) (0.009) (0.009) (0.009)

Foreign assets / Total assets -0.009 -0.003 -0.009 -0.003 -0.009 -0.003

(0.009) (0.009) (0.009) (0.009) (0.009) (0.009)

Boardsize -0.043*** -0.041*** -0.043*** -0.041*** -0.041*** -0.039***

(0.014) (0.014) (0.014) (0.014) (0.014) (0.014)

Non-executive directors / Boardsize -0.018 -0.017 -0.018 -0.017 -0.017 -0.016

(30)

29

Female director / Boardsize 0.044* 0.050* 0.044* 0.050** 0.044* 0.050*

(0.026) (0.026) (0.026) (0.026) (0.026) (0.026)

(0.032) (0.032) (0.032) (0.032) (0.032) (0.032)

Director board tenure / Boardsize -0.001 0.000 -0.001 -0.000 -0.000 0.001

(0.004) (0.004) (0.004) (0.004) (0.004) (0.004)

Other quoted board affiliations / Boardsize -0.070** -0.059* -0.070** -0.059* -0.065* -0.055

(0.035) (0.036) (0.035) (0.036) (0.035) (0.035) Business segments 0.004 0.004 0.004 0.004 0.003 0.004 (0.003) (0.003) (0.003) (0.003) (0.003) (0.003) Geographic Segments 0.004 0.007* 0.004 0.007* 0.004 0.007* (0.004) (0.004) (0.004) (0.004) (0.004) (0.004) Constant 0.099* 0.098* 0.100* 0.099* 0.094* 0.093* (0.054) (0.053) (0.054) (0.053) (0.054) (0.053)

Observations 18,651 18,651 18,651 18,651 18,651 18,651

R-squared 0.110 0.114 0.110 0.114 0.110 0.113

Number of firm_id 3,330 3,330 3,330 3,330 3,330 3,330

Robust standard errors in parentheses

(31)

30 Table 7:

PSM logit regression

Minister Dummy Coef. Std. Err. z P> [Z] [95% conf. Interval]

Total assets 0.407 0.050 8.14 0.000 0.309 0.505

Boardsize 0.913 0.377 2.42 0.015 0.175 1.651

Tobin's-q 0.148 0.056 2.62 0.009 0.0372 0.258

CEO duality 0.687 0.156 4.42 0.000 0.382 0.992

Debt interest rate 5.724 3.943 1.45 0.147 -2.004 13.453

Capital expenditures / Total assets 3.584 0.991 3.62 0.000 1.642 5.527

(32)

31 Table 8:

PSM regression results H1

Percentage ministers on board 0.194 0.387**

(0.175) (0.175)

Minister dummy 0.009 0.027*

(0.016) (0.016)

(0.118) (0.113)

CEO duality 0.059 0.068* 0.058 0.066* 0.057 0.065*

(0.037) (0.037) (0.037) (0.037) (0.038) (0.038)

CEO change 0.040* 0.042** 0.041* 0.043* 0.043* 0.047**

(0.022) (0.021) (0.023) (0.022) (0.022) (0.021)

CEO gender dummy -0.044 -0.049 -0.044 -0.049 -0.046 -0.052

(0.032) (0.032) (0.032) (0.031) (0.033) (0.032) Tobins-q 0.015 0.022 0.017 0.024 0.018 0.026* (0.015) (0.015) (0.015) (0.015) (0.014) (0.014) Total assets -0.004 0.007 -0.001 0.010 0.001 0.015 (0.030) (0.032) (0.030) (0.033) (0.033) (0.035) ROA -0.083 0.006 -0.082 0.007 -0.067 0.016 (0.144) (0.143) (0.144) (0.144) (0.154) (0.155) Loss dummy 0.011 0.015 0.012 0.015 0.012 0.016 (0.027) (0.025) (0.027) (0.025) (0.027) (0.025)

Total liabilities / Total assets 0.052 0.051 0.053 0.052 0.054 0.055

(0.071) (0.070) (0.072) (0.071) (0.072) (0.072)

Debt interest rate -0.928 -0.735 -0.924 -0.720 -0.933 -0.745

(0.934) (0.903) (0.938) (0.910) (0.935) (0.910)

Capital expenditures / Total assets 0.128 0.276 0.124 0.269 0.120 0.267

(0.208) (0.244) (0.207) (0.246) (0.203) (0.244)

R&D / Total Assets 1.779 1.500 1.786 1.508 1.747 1.516

(1.257) (1.176) (1.261) (1.182) (1.237) (1.180)

Cash / Total assets -0.017 -0.015 -0.019 -0.018 -0.023 -0.026

(0.150) (0.145) (0.149) (0.145) (0.149) (0.144)

Closeheldshares -0.143* -0.132* -0.151* -0.143* -0.155* -0.156**

(0.078) (0.073) (0.079) (0.075) (0.079) (0.075)

Foreign assets / Total assets 0.036 0.044 0.036 0.042 0.036 0.044

(0.038) (0.037) (0.038) (0.037) (0.037) (0.038)

Boardsize -0.032 0.045 -0.041 0.027 -0.033 0.026

(0.072) (0.068) (0.071) (0.067) (0.079) (0.076)

Non-executive directors / Boardsize 0.041 0.027 0.040 0.025 0.048 0.025

(0.157) (0.156) (0.158) (0.156) (0.158) (0.158)

Female director / Boardsize 0.034 0.057 0.023 0.038 0.033 0.037

(0.086) (0.090) (0.085) (0.087) (0.095) (0.093)

(0.173) (0.176) (0.176) (0.180) (0.180) (0.183)

Director board tenure / Boardsize -0.010 0.018 -0.012 0.015 -0.012 0.014

(33)

32

Other quoted board affiliations / Boardsize -0.129 -0.093 -0.117 -0.065 -0.096 -0.068

(0.201) (0.203) (0.207) (0.210) (0.247) (0.247) Business segments -0.036 -0.026 -0.037 -0.026 -0.040 -0.029 (0.032) (0.032) (0.032) (0.032) (0.031) (0.029) Geographic Segments -0.071 -0.072* -0.069 -0.069 -0.068 -0.065 (0.043) (0.043) (0.043) (0.043) (0.043) (0.043) Constant 0.074 -0.243 0.080 -0.226 0.047 -0.247 (0.288) (0.317) (0.285) (0.316) (0.320) (0.350)

Observations 408 408 408 408 408 408

R-squared 0.269 0.284 0.268 0.280 0.268 0.275

Number of firm_id 215 215 215 215 215 215

Robust standard errors in parentheses

(34)

33 Table 9:

Quality of PSM test

Mean t-test V (T) /

Variable Treated Control %bias t p> I t I V (C)

Total assets 8.456 8.468 -0.6 -0.06 0.951 0.89

Boardsize 2.271 2.248 8.6 0.86 0.393 0.89

Tobin's-q 2.037 2.061 -1.8 -0.19 0.849 1.02

CEO duality 0.691 0.706 -3.1 -0.32 0.747 .

Debt interest rate 0.025 0.024 3.7 0.41 0.680 0.89

Capital expenditures / Total assets 0.068 0.070 -2.2 -0.20 0.839 1.23

Ps R2 LR chi2 p>chi2 MeanBias MedBias B R %Var