Bureaucrats, PC directors and capital structure: A comparative study in the US

Bureaucrats, PC directors and capital structure: A comparative study in the US

1 Abstract

This thesis distinguishes bureaucrats from other types of PC directors and examines their effect on the capital structure of firms. By analyzing distinct differences between bureaucrats and other types of PC directors I come up with a possible explanation as to how bureaucrat directors affect firm leverage. I find that bureaucrat directors are positively associated with firm leverage and not contemporaneously associated with the interest rate on debt. I also find that bureaucrat directors moderate the negative association between the financial crisis and leverage without

contemporaneously being associated with the interest rate on debt. My findings are in support of my hypothesized explanation for the relationship between bureaucrat directors and leverage, namely that bureaucrat directors decrease the perceived risk of a firm.

2 Introduction

Studies suggest that firms benefit from board members with political connections (Goldman et al., 2009; Cooper et al., 2010). These studies often follow the reasoning of the resource dependence theory, which argues that firms are dependent on influences from their environment. As the government is a major influence, firms try to manage their dependency on the government by appointing politically connected directors (or ‘PC directors’). The term PC director subsumes a variety of different governmental roles. Because these different roles have different characteristics, there remains ambiguity about how exactly PC directors add value to firms. In this thesis, I split the term PC directors into two subgroups, namely senior bureaucrat directors (or ‘bureaucrat directors’) and all other types of PC directors (e.g. politicians, ministers, etc.), to investigate whether differences in governmental backgrounds lead to differences in the capital structure of a firm. I posit that the different types of PC directors add value to firms in different ways. Specifically for bureaucrats as opposed to other PC directors I hypothesize that due to the nature of their role in government their added value comes from decreasing the perceived risk of the firm from the perspective of investors.

In their study, Hillman et al. (2009) describe three different ways in which PC directors add value to a firm. Firstly, they can add value through unique knowledge about procedures. Secondly, they can provide access to politicians and decision-makers, and thirdly they add legitimacy to a firm. Given these three ways in which PC directors add value, it is likely that some of these ways are more characteristic of politicians than bureaucrats and vice versa. Social capital as defined by Nahapiet & Ghoshal (1998) is “the sum of the actual and potential resources embedded within, available through, and derived from the network of relationships” (pp. 243). I posit that PC directors that are politicians have a higher social capital than bureaucrats through campaigning, building relationships, and increasing their public profile. As a result, it is more likely that their way to add value is through access to politicians and decision-makers. In contrast, given the experience of bureaucrats in the execution of laws and regulations, bureaucrat directors are more likely to add value to a firm through unique knowledge about procedures and through adding legitimacy (Mukherjee, 2020).

Bureaucrats are government officials that hold positions on a high level (i.e. directors of public institutes). They are employees in executive roles of government agencies who exercise the policies set by the politicians. This relationship between bureaucrats and politicians can be viewed as similar to the one between the agent and its principle (Moe, 2005). Bureaucrats hold high-ranking positions in agencies that implement the policies such as the FBI, FDA, or the Department of Agriculture, just to name a few. The main difference with politicians is that bureaucrats are not elected by the public and often have a longer tenure in their positions (Mukherjee, 2020). Furthermore, bureaucrats

3 cannot be removed from their positions by the public if they perform inadequately whereas

politicians can be. Senior bureaucrats have more information and experience with policies as they are the executive branch of regulation and policies whereas politicians are the makers of regulation and policy (Moe, 2006). Even though bureaucrats have more experience than politicians in the execution of policy, they might not have the same social capital that politicians have. Through the election process of campaigning and the public nature of their positions, politicians enjoy a greater public image than senior bureaucrats do.

When banks lend capital to corporations they protect their investments with restrictive covenants. These covenants aim to restrict the behaviour of corporations and thereby reduce the risks regarding information asymmetry and financial distress (Paglia & Mullineaux, 2006). Banks use these covenants to decrease the risk that their investment, in the form of debt capital, cannot be paid back or

recovered. I posit that, because of their long tenure in government positions implementing and enforcing rules and regulation, bureaucrat directors increase the confidence of banks that these covenants will be abided by and thereby decreasing the perceived risk of that firm. This would imply that firms with bureaucrat directors would have access to higher amounts of leverage without additional costs, as the higher leverage of firms with bureaucrat directors would be perceived as having less risk compared to firms with that same level of leverage but without bureaucrat directors.

Studies on PC directors have already shown several different effects that PC directors have on a firm. Firms with PC directors can have more tax benefits (Wu et al. 2012), easier access to financing from banks (Claessens et al. 2008) and are even more likely to be bailed out by the government (Faccio et al. 2006). Various studies have indicated a positive relationship between political connections and firm leverage (Dewenter & Malatesta, 2001; Liu & Tian, 2010; Liu et al. 2011). Yang et al. (2014) find that politically connected firms in China have higher leverage and Saeed et al. (2014) found the same results specifically for long-term debt for Pakistani firms. These studies tend to either focus solely on politicians or treat bureaucrats, ministers, and politicians as one group. In an early paper

investigating politically connected firms, Faccio (2006) defines a firm to be politically connected when “one of the companies’ shareholders or top officers is member of parliament, a minister, a head of state or closely related to a top official” (pp. 370). This definition, like many studies on political connectedness, encapsulates both politicians, ministers, and to some extent high-level (senior) bureaucrats causing the effects of bureaucrats as a group (if any) to be severely understudied (Conyon et al., 2015). This thesis will investigate whether bureaucrat directors affect firm leverage differently than other PC directors do. This study will contribute to the literature on PC directors by examining whether the different backgrounds of bureaucrats and other PC directors have different

4 effects on leverage and if bureaucrats should therefore be treated separately from other types of PC directors.

As PC directors affect the level of debt a firm has, they influence the capital structure of that firm. Following trade-off theory firms seek to optimize their capital structure, that is to optimize the balance between leverage and equity. Finding this optimum is crucial for maximizing enterprise value (Brusov et al., 2013). Leverage is appealing as it is a relatively cheap source of financing. However, increasing leverage too much also increases the financial risks of a firm. As the financial risks of a firm increase, equity investors will demand a higher return, driving the cost of equity up. The ‘optimal balance’ between equity and debt will result in the lowest Weighted Average Cost of Capital (WACC). A lower overall cost of capital makes a firm more profitable as the resources they use to generate revenue cost less on average. The idea that firms maximize their value by lowering their WACC is crucial for our hypothesis that bureaucrat directors, as opposed to other PC directors, add value to a firm by lowering the perceived risk of that firm. This implies that firms will always prefer debt over equity, which may not be always the case. I assume that firms always aim to maximize their capital structure and seek the lowest WACC, indicating an optimal balance between debt and equity. Given that this assumption holds when leverage increases this can be either a sign of banks more willing to lend capital or a sign of equity holders requiring less return. Either way, this can be viewed as a sign of lower perception from either banks or equity holders.

This thesis examines the relationship between bureaucrat directors and leverage in US-based firms. By creating separate identifications for bureaucrat directors and other PC directors respectively, I find that these have separate (different) effects on firm leverage. I find suggestive evidence to support the hypothesis that the effect of bureaucrat directors on firm leverage is associated with a decrease in perceived risk as results show that bureaucrat directors increase leverage and moderate the negative relationship between the financial crisis and leverage without increasing interest rates.

This thesis makes two contributions to the existing literature on PC directors. Firstly, this study shows that bureaucrats are different from other types of PC directors and that these differences result in different effects on leverage. Though prior research has focussed mainly on PC directors as a whole and especially politicians (Conyon et al., 2015), this study shows the importance of bureaucrats. The differences in backgrounds between bureaucrats and other PC directors mean that they both bring value to a firm in different ways. Results of this thesis show that bureaucrat directors matter as a group distinctly from other PC directors and should be treated separately. Future research into the effects of political connectedness should take note of this fact.

5 Lastly, this thesis provides a possible explanation as to why bureaucrat directors do allow firms to increase their leverage without additional costs whereas other PC directors do not. I hypothesize that the positive association between bureaucrat directors and leverage is because bureaucrat directors decrease the perceived risk of a firm. This is supported by the results that show that the increase in leverage is not contemporaneously accompanied by an increase in the interest rate. In this study, I use the three ways in which PC directors add value to the firm, as proposed by Hillman (2009), to explain why bureaucrat directors decrease perceived risk and other PC directors do not. Though this hypothesis should be researched further, this gives the first glimpse of one way where bureaucrat directors differ from other PC directors. Furthermore, this has practical implications as well. If firms want to secure their access to debt in times of higher risk they are better off appointing bureaucrat directors compared to other PC directors.

The next section will lay out the theoretical framework that underlies this research and the

motivation for our hypotheses. Thereafter I explain our selection of data, our variables, and empirical models that are used to test our hypotheses. This is followed by the results section where the results of our empirical tests will be interpreted followed by a conclusion on the thesis as a whole.

6 Theoretical background & hypotheses

Pecking order theory (Myers & Majluf, 1984) suggests that the cost of financing increases with the level of information asymmetry. Capital funding through debt is attractive for firms because the cost of capital for debt is cheaper. This is because interest can be deducted from your taxable income and because there is less information asymmetry in debt financing compared to equity financing (Leary & Roberts, 2010). The interest rate for debt capital is dependent on the business risk of the firm and debt holders require a greater return if the firm they lend to is deemed risky. The business risk of a firm is derived from its underlying assets and its capacity to produce stable cash flows (Amit & Wernerfelt, 1990). Because interest is paid before tax on the income statement, it creates value for a firm by decreasing a firm’s pre-tax income, causing them to have to pay fewer taxes. Too much debt, however, makes a firm overleveraged and increases the financial risk of a firm which, in turn, causes the cost of equity to go up (De Wet, 2006). As debt is cheaper but also increases the cost of equity, firms seek to find an optimal capital structure such that the WACC is as low as possible.

The capital structure of a firm can be influenced by many different factors, among which the firm’s environment, activities, and even the composition of its board. Research on Chinese firms by Yang et al. (2014) indicates that politically connected firms not only have, on average, more debt but

specifically have more long-term loans. This is supported by the study of Saeed et al. (2014) who find, in their study of Pakistani firms, that political connections only positively influence long-term debt but not short-term debt. Short-term debt is mainly used as working capital to finance business operations. As political connections give easier access to debt, literature seems to suggest that this effect is stronger for long-term debt.

Bureaucrats differ from politicians in various ways. Bureaucrats, unlike politicians, are not elected by the public but rather enter the government bureaucracy as a professional career (Mukherjee, 2020). The tenure of bureaucrats is not directly dependent on political cycles and their expertise comes from the execution of policy, rather than setting it. As stated by Cohen (2013), bureaucrats have more expertise in policies that cannot be matched by politicians. Politicians have larger social capital and are often better known to the public, mostly through elections and campaigns. Though I expect that the group of bureaucrats in our sample will have some social capital that could add value, it is on average much less than politicians have. Bureaucrats, on the other end, operate mostly outside the view of the public and are not as well-known publicly. The nature of their job, implementing and enforcing law and regulation behind closed doors, makes them less likely to add value through their social capital and more likely to add value through their expertise in policy and through adding legitimacy. Effects that bureaucrats have on leverage, if any, are therefore attributable to their long

7 tenure in public service and the experience gained in the bureaucracy of policy implementation (Mukherjee, 2020). Effects that other PC directors have on leverage, given their background, are more likely to be a result of their social capital. Therefore, even though bureaucrat directors and other PC directors may have a similar effect on leverage, this effect can be explained by different causal factors.

One of the most fundamental concepts of finance is the risk-return relationship as introduced by Fama & French (1992). This means that for any increase in risk, investors require an increase in the return of their investment. Therefore, if bureaucrat directors increase firm leverage this does not necessarily mean that this is due to a decrease in perceived risk because if this increase in leverage is accompanied by an increase in the interest rate, banks do perceive the additional capital loan as riskier. To test the hypothesis that bureaucrat directors decrease the perceived risk of a firm I examine whether any effects of bureaucrat directors on leverage are contemporaneously

accompanied by a change in interest rate. Additionally, I examine whether the change in leverage is through short-term debt or long-term debt.

Hypothesis 1:

Bureaucrat directors increase leverage without additional costs whereas other PC directors do not. Hypothesis 2:

Bureaucrat directors increase leverage through long-term debt and not through short-term debt. The risk perception of investors can be influenced by a great number of different factors, among which environmental ones. The global financial crisis had a major impact on markets around the world and many economists consider this crisis to be the worst since the Great Depression of the 1930s (Hodson & Quaglia, 2015). The financial crisis and the losses that banks suffered ran out to a decline in lending by financial institutions, particularly the lending of long-term debt declined (Gonzalez, 2015; Ivashina & Scharfstein, 2010). These institutions decreased their lending as they were increasingly worried about defaults which led to an overall economic downturn and less liquidity in the market (Cornett et al., 2011). In other words, the risk perception of financial institutions grew and capital lending declined thereby making it more difficult for firms to secure their access to capital.

If bureaucrat directors increase firm leverage by decreasing the perceived risk of a firm then this effect should be most notable during times when environmental factors cause an increase in the perceived risk for investors. Previous studies indicate a negative relationship between the financial crisis and leverage and that this is due to an increase in perceived risk of financial institutions. If bureaucrat directors decrease the perceived risk of a firm then firms with bureaucrat directors

8 should be able to borrow more during times of crisis as opposed to firms without bureaucrat

directors. Moreover, they should not only be able to borrow more debt capital but this extra capital should not be more expensive to acquire than other debt capital. I, therefore, expect that firms with bureaucrat directors had better access to debt capital during the financial crisis as opposed to firms without bureaucrat directors, which leads to the final hypothesis.

Hypothesis 3:

Bureaucrat directors moderate the relationship between the financial crisis and firm leverage without additional costs but other PC directors do not.

9

Research design

Sample & data

The data that is used for this thesis is comprised of merging various data from different sources. This study uses data of US-firms explicitly. Data for accounting and financial information is retrieved from Worldscope. This data is then merged with data regarding board members and their characteristics, which is obtained from BoardEx. The data is merged based on firm-id and year. A description of all variables used in this study can be found in appendix A.

Lastly and important for this thesis, I use hand-collected data about the governmental background of board members. This hand-collection was partly done by me and the rest of my team of fellow students under our supervisor dr. Mukherjee. Through internet searches and sites such as

Marketscreener, we validated whether board members have past experience working in government and subsequently categorized them based on their respective roles. We distinguished between politicians, ministers, senior bureaucrats, military, junior bureaucrats, and (political) advisors. Categorization was done based on the highest position to prevent overlap, whenever a board member had roles in multiple categories.

Dependent variables

The variable of interest for hypothesis 1 is leverage (LEV). Leverage is measured as total debt divided by total assets. This ratio is used because it makes the variable normalized and makes it clear to interpret. This ratio is often used in studies regarding leverage (Liu & Tian, 2010).

For hypothesis 2 I measure the effect of bureaucrat directors and other PC directors on long-term debt (LONG-TERM DEBT) and short-term debt (SHORT-TERM DEBT). These variables are measured by the amount of long-term debt (or short-term debt) divided by the total assets of the firm, thereby normalizing this variable. Lastly, I measure the effect of bureaucrat directors on the interest rate of debt (INTEREST RATE) for all hypotheses.

Explanatory variables

For the main effect, I aim to establish the effect of bureaucrat directors (BUREAUCRAT) on leverage. The variable BUREAUCRAT indicates the number of directors on a board who previously served as a bureaucrat in the US proportional to the total number of directors on that board. As data regarding the backgrounds of directors is scarce, the data for the bureaucrat directors is hand-collected. The data is established mostly through internet searches and verifying the previous positions of directors on sites such as Marketscreener. Their seniority was then determined not by the length of their tenure but rather the seniority of their position. Directors of governmental institutions and the like are categorized as senior bureaucrats. As this is a comparative study, I include a measure of other PC

10 directors (POLIT) in our regressions. This is partly used as a control variable but also as a comparative variable as the aim is to investigate whether different effects exist for bureaucrat directors as

opposed to other PC directors.

Further, a dummy for the crisis (CRISIS) is added as an explanatory and interaction variable. Here CRISIS is a dummy that is equal to 1 during the time of the global financial crisis between 2009 and 2010 and 0 otherwise.

Control variables

As the dependent variables are affected by more factors than our explanatory variables capture, I control for some of these factors in the regression models. By doing so the results are clearer to interpret as all these control variables can be held constant, thereby increasing the validity of the main effect as it is not influenced by the factors that are controlled for. In this research, the control variables are mainly accounting variables.

I control for lagged leverage (LEV) because leverage in a previous period has a large influence on how much firms can increase leverage. I control for foreign assets (FOREIGN ASSETS). This variable is calculated as total foreign assets divided by the total amount of assets and winsorized at the 1-99% percentile range. This variable measures to which extent a company has multinational operations, which is known to influence leverage (Baum et al., 2009). This study follows Faccio (2006) in

controlling for financial success ratios. In this study Faccio (2006) controls for ROA (ROA) and Tobin’s Q (TOBINS Q) as these signal the financial health of a firm which in turn impacts the leverage of a firm. I control for the level of inventory (INVENTORY) as larger inventories require more working capital and therefore more debt (especially short-term debt). A proxy for deferred taxes (DEFERRED TAX DIFFERENCE) is controlled for as deferred taxes influences the taxable income of a firm, which is similar to the way debt adds value to a firm, namely by reducing taxable income. Furthermore, I control for the number of business segments (BUSINESS SEGMENTS). This is done because more business segments signal a greater business risk as firms are more complex. This business risk in turn affects the leverage of the firm.

Board characteristics are known to impact firm-level outcomes and can therefore influence our dependent variables. Board independence (BOARD INDEPENDENCE) has shown to be associated with firm leverage (Almania, 2017). Additionally, Harford et al. (2008) indicated that differences in board size (BOARDSIZE) matter in predicting firm leverage. Lastly, the quality of corporate governance is negatively associated with firm leverage according to Nadarajah et al. (2016). To control for this I use the number of board affiliations of directors (BOARD AFFILIATIONS) as a proxy for corporate

11 governance quality. This variable is a measure of how many boards the directors of a firm sits on, which is an indication of how involved these directors are at the firm.

Empirical models

To test the hypotheses I use a firm-fixed effects regression model to control for time-invariant factors. The data is both cross-sectional as time-series for which panel data is used. For our main effect, I increase the robustness of the results with firm-fixed effects. The firm-fixed effects are included in all the models. Furthermore, the models include heteroskedasticity-robust standard errors. Below, for each of the three hypotheses, I explain the regression models that are used. I only explain the dependent and explanatory variables as the control variables are the same for all models. All explanatory and control variables are lagged by 1 year the effect that bureaucrat directors have on leverage is not a contemporaneous one.

Model 1

LEVi,t = αi + β1BUREAUCRATi,t-1 + β2POLIT,t-1 + β3CONTROLSi,t-1 + εi,t

Here β1 measures the effect of bureaucrat directors of firm i in t-1 on leverage in time t, compared to firms without bureaucrat directors given the controls. In this model, I included firm-fixed effects. Here β2 represents the coefficient for the effect of PC directors who are not bureaucrats on leverage and εi,t is the error term. Our expectations dictate that β1 and β2 are both positive. To test hypothesis 1 we test model 1 again but with INTEREST RATE as the dependent variable. Here, our expectations dictate that β1 is non-positive.

Model 2

Our second models measure the effect of bureaucrat directors on long-term- and short-term debt compared to firms that have no bureaucrat directors. These models are the same as model 1 only the dependent variable is LONG-TERM DEBT and SHORT-TERM DEBT instead of LEV. The goal is to

establish whether this effect is different for long-term debt than for short-term debt. Our

expectations indicate that β1 and β2 of the long-term model are significant and positive whereas β1 and β2 of the short-term model are not significant.

Model 3

LEVi,t = αi + β1BUREAUCRATi,t-1 + β2POLIT i,t-1 + β3CRISISt-1 + β4BUREAUCRATi,t-1 * CRISISt-1 + β5POLITi,t-1 * CRISISt-1 + β6CONTROLSi,t-1 + εi,t

The above model is equal to model 1 where an interaction term between bureaucrat directors and the crisis dummy is added. This makes β1

+ β

+

β

the effect of bureaucrat directors on firm

12 the global financial crisis given the controls. The partial effect of the crisis on leverage without the interactions is β3 which I expect to be negative. Following the third hypothesis, I expect that

13

Results

Descriptive statistics

Table 1 shows the descriptive statistics for all variables used in the regression models. All variables that are winsorized are winsorized at the 1-99% percentile range. The mean of LEV is .529 with a standard deviation of .26. The mean of LONG-TERM DEBT (SHORT-TERM DEBT) is .19 (.38) with a standard deviation of .045 (.08). The mean of BUREAUCRAT is .011 with a standard deviation of .038. This indicates that in our sample on average each firm-year has 1.1% of their board directors being bureaucrats. The variable POLIT has a mean 0f 0.012 with a standard deviation of 0.039, indicating that on average each firm-year observation has a board where 1.2% is a PC director.

| Insert table 1 here |

Pearson correlation table

Table 2 shows the Pearson correlation coefficients for all variables. I consider all correlation

coefficients larger than .7 as a sign of multicollinearity. As shown in the table no two variables have a correlation coefficient larger than .7. The largest coefficients are between SHORT-TERM DEBT and LONG-TERM DEBT (.594) and between DEFERRED TAX DIFFERENCE and ROA (.575).

| Insert table 2 here |

Regression results

This thesis studies the effect of bureaucrat directors and other PC directors on firm leverage. Our three hypotheses are tested by way of regression analysis. For each of the three hypotheses, the regression results are presented in tables 3 and 4. Important to note is that all significant

relationships that are discussed are made bold and all explanatory variables are lagged 1 year, as the effect is expected not to be contemporaneous.

In table 3 the results of the regression results for the first 2 hypotheses are presented. It is found that bureaucrat directors are positively related to leverage (β = .1739; p<.01) and this effect becomes stronger when we exclude the financial sector (β = .1862; p<.01). It is shown that while bureaucrat directors are positively associated with long-term debt (β = .0331; p<.01) this same effect is not present for short-term debt. Lastly, results indicate that bureaucrat directors are not significantly associated with the interest rate. PC directors who are not bureaucrats show no significant relationships for all models.

14 The results of the regression analysis for hypothesis three are displayed in table 4. The direct

relationship between BUREAUCRAT and LEV outside the financial crisis is significant and positive (β = .1754; p<.01). The relationship between the CRISIS and LEV is significantly negative (β = -.0078; p<.01). This effect is moderated if an interaction term with BUREAUCRAT is included (β = .0853; p<.10). All effects become larger and more significant when the financial sector is excluded, where the association between BUREAUCRAT and LEV is (β = .1702; p<.01), CRISIS and LEV (β = -.0110; p<.01), and the interaction between BUREAUCRAT and CRISIS on LEV (β = .0099; p<.10). The effect of CRISIS on the interest rate is significantly negative (β = -.0014; p<.01) and there is no significant effect of BUREAUCRAT on INTEREST RATE. Also, no significant relationship exists for the interaction

between BUREAUCRAT and CRISIS and the interest rate. Lastly, no significant results are found for any association between POLIT and LEV.

| Insert table 4 here |

Robustness

Heteroskedasticity robust standard errors are used in the models to allow heteroskedasticity among our residuals. I further test our hypotheses with and without the financial sector in our sample. I do this to check whether the financial sector influences our main results. Lastly, we included firm-fixed effects in all regression models.

15

Conclusion

This thesis aimed to establish whether bureaucrat directors should be treated as a distinct group from other PC directors. Where in other literature bureaucrat directors are understudied or treated as a part of PC directors as a group, this study shows that a distinction can and should be made between bureaucrats directors and other PC directors. This thesis indicates that significant differences between bureaucrat directors and other PC directors exist. This is done by creating separate variables for bureaucrat directors and other PC directors respectively.

I found support for our first and second hypothesis that bureaucrat directors increase firm leverage without increasing the interest rate. The results show that bureaucrat directors are positively associated with the leverage of a firm but not contemporaneously associated with the interest rate. This is particularly the case for term debt, as we found a significant relationship between long-term debt and bureaucrat directors but the same relationship is not present for short-long-term debt. While we found significant relationships between bureaucrat directors and leverage, we did not find these between other PC directors and leverage. These results support our hypothesis that bureaucrat directors increase leverage without increasing the interest rate.

Results show support for our third hypothesis that bureaucrat directors moderate the relationship between the global financial crisis and leverage without additional costs but other PC directors do not. I found that even though on average leverage went down during the crisis, firms with bureaucrat directors had less of a decrease in leverage. Moreover, this study indicates that this moderating effect did not come with a higher cost of debt, as interest rates were not significantly affected by bureaucrat directors during the crisis. Again, we found no significant results for PC directors who are not bureaucrats.

In the introduction, I hypothesized that the positive relationship between bureaucrat directors and leverage exists because bureaucrat directors decrease the perceived risk of a firm. For this

hypothesis, I found suggestive evidence as one crucial assumption was made, namely that firms at all times optimize their capital structure. I found that bureaucrat directors are associated with higher leverage during times of increased perceived risk, whereas other PC directors did not. Furthermore, this increased leverage was not accompanied by an increased interest rate. These results can be interpreted as evidence that firms with bureaucrat directors, as opposed to firms without bureaucrat directors, have an optimal capital structure that has proportionally more debt.

Limitations

This study uses a sample of US-based firms only. As political landscapes vary greatly internationally, this study is therefore not generalizable to other countries. Due to the scarcity of data, we

hand-16 collected the background information regarding bureaucrat directors. As we only differentiate between bureaucrat director and PC director, we do not account for any differences in background or experience that bureaucrat directors might have.

Practical implications & future research

Results of this study show that bureaucrat directors can be used to secure access to leverage, especially during times of crisis. Firms can appoint bureaucrat directors to achieve this end and by doing so making themselves less vulnerable to increasing financial uncertainty during a crisis. There are also academic implications. This thesis indicates that bureaucrats themselves are important and can be treated separately from other types of PC directors. As they are heavily understudied as a group, this study might denote their importance and spark more research into the effects of bureaucrat directors.

Though we show a significant relationship between bureaucrat directors and leverage, the

mechanism behind this is still not clear. As stated by Hillman et al. (2009), one of the advantages of PC directors is that they add legitimacy to a firm. Given their experience enforcing and following rules and regulations, I believe that bureaucrat directors especially add legitimacy and give lenders more confidence that any covenants that accompany debt contracts are upheld. This led to the hypothesis that bureaucrat directors decrease the perceived risk of a firm, for which I found suggestive

evidence. Future studies can research this further by exploring whether capital lenders such as banks view bureaucrat directors as indeed increasing legitimacy and whether bureaucrat directors increase adherence to debt covenants. On the whole, it should be further explored whether more different effects between bureaucrat directors and other PC directors exist and what the mechanisms behind these effects are.

17

References

Almania, O. (2017). Board of Director Independence and Financial Leverage in the absence of Taxes. International journal of economics and finance, 9(4): 90-94.

Amit, R., Wernerfelt, B. (1990), Why Do Firms Reduce Business Risk? Academy of Management Journal, 33 (3): 520-533.

Baum, C. F., Stephan, A., Talavera, O. (2009), The effects of uncertainty on the leverage of nonfinancial firms. Economic Inquiry, 47(2): 216–225.

Brusov, P.N., Filatova, T.V., Orekhova, N.P. (2013). Absence of an optimal capital structure in the famous tradeoff theory. Journal of reviews on global economics, (2): 94-116

Claessens, S., Feijen, E., Laeven, L. (2008). Political connections and preferential access to finance: the role of campaign contributions. Journal of Financial Economics, 88 (3): 554–580.

Cohen, N. (2013). The power of expertise? Politician–bureaucrat interactions, national budget transparency and the Israeli health care policy. Policy Studies, 34: 638–654.

Conyon, M. J., He, L., Zhou, X. (2015). Star CEOs or political connections? Evidence from China's publicly traded firms. Journal of business Finance & Accounting, 42: 412-443.

Cooper, M, J., Gulen, H., Ovtchinnikov, A. V. (2010). Corporate political contributions and stock returns. The Journal of Finance, 65(2): 687-724.

Cornett, M.M., McNutt, J.J., Strahan, P.E., Tehranian, H. (2011). Liquidity risk management and credit supply in the financial crisis. Journal of financial economics, 101 (2): 297-312.

De Wet, J.H.H. (2006). Determining the optimal capital structure: A practical contemporary approach. Meditari Accountancy Research, 14(2): 1–16.

Dewenter, K.L., Malatesta, P.H. (2001). State-owned and privately owned firms: The Effect of Ownership Structure on Leverage Decision. American Economic Review 91: 320–334.

Faccio, M. (2006). Politically connected firms. American Economic Review, 96: 369–386.

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

Fama, E.F., French, K.R. (1992). The cross-section of expected stock returns. The journal of finance, 47(2): 427-465.

Goldman, E., Rocholl, J., So, J. (2009). Do Politically Connected Boards Affect Firm Value? Review of Financial Studies, 22(6): 2331–60.

Gonzalez, V.M. (2015). The financial crisis and corporate debt maturity: the role of banking structure. Journal of corporate finance, (35): 310 – 328.

Harford, J., Li, K., Zhao, X. (2008). Corporate boards and the leverage and debt maturity choices. International Journal of Corporate Governance, 1(1): 3–27.

Hodson, D., Quaglia, L. (2009). European Perspective on the Global Financial Crisis: Introduction. Journal of common market studies, 47 (5): 939–953.

Hillman, A. J., Withers, M. C., Collins, B. J. (2009). Resource Dependence Theory: A Review. Journal of Management, 35(6): 1404–1427.

18 Ivashina, V., Scharfstein, D. (2010). Bank lending during the financial crisis of 2008. Journal of

Financial Economics, 97 (3): 319–338.

Leary, M.T., Roberts, M.R. (2010). The pecking order, debt capacity, and information asymmetry. Journal of financial economics, 95: 332-355.

Liu, Q., Tian, G. (2010). Political connections, founding family ownership and leverage decision of privately owned firms. 23rd Australasian Finance and Banking Conference, pp. 1-38.

Liu, Q., Tian, G., Wang, X. (2011). The effect of ownership structure on leverage decision: new evidence from Chinese listed firms. Journal of the Asia Pacific Economy, 16(2): 254-276.

Moe, T.M. (2005). Power and political institutions. American political science association, 3 (2): 215-233.

Moe, T.M. (2006). Political control and the power of the agent. The journal of law, economics, and organization, 22 (1): 1-29.

Mukherjee, S. (2020). The Obscure Directors: Why Bureaucrat Directors Matter. University of Groningen, Working Paper.

Myers, S., Majluf, N. (1984). Corporate financing and investment decisions when firms have information that investors do not have. Journal of Financial Economics, 13: 187–221.

Nadarajah, S., Ali, S., Liu, B., Haung, A. (2016). Stock liquidity, corporate governance and leverage: New panel evidence. Pacific-Basin Finance Journal, Forthcoming.

Nahapiet, J., Ghoshal, S. 1998. Social capital, intellectual capital, and the organizational advantage. The academy of management review, 23(2): 242-266

Paglia, J.K., Mullineaux, D.J. (2006). An empirical exploration of financial covenants in large bank loans. Banks and bank systems, 1(2): 103-122

Saeed, A., Belghitar, Y., Clark, E. (2014) Political connections and leverage: Firm-level evidence from pakistan. Managerial and decision economics, 36(6): 364-383.

Wu, W., Wu, C., Wu, J., Zhou, C. (2012). Political connections, tax benefits and firm performance: Evidence from china. Journal of Accounting & Public Policy, 31(3): 277- 300.

Yang, D., Lu, Z., Luo, D. (2014). Political connections, media monitoring and long-term loans. China journal of accounting research, 7: 165-177.

19

Table 1: Summary statistics

This table shows the summary statistics for all variables used in this study. For each variable the mean, standard deviation, minimum, maximum and number of observations are reported. A description of how these variables are constructed can be

found in Appendix A.

Variable Obs Mean Std. Dev. Min Max

LEV 49483 .529 .26 .008 1.44 BUREAUCRAT 51467 .011 .038 0 .5 POLIT 51467 .012 .039 0 .5 FOREIGN ASSETS 39652 .069 .158 0 1 LONG-TERM DEBT 45415 .019 .045 -.045 1.038 SHORT-TERM DEBT 48172 .038 .08 0 2.215 INTEREST RATE 47644 .023 .025 0 .142 TOBINS Q 49138 1.915 1.48 .491 9.555 BOARDSIZE 51467 2.095 .303 1.099 2.996 INVENTORY 40863 .099 .13 0 .617 ROA 47850 .018 .2 -1.291 .392 DEFERRED TAX DIFFERENCE 42854 -.047 .22 -.895 .693 BUSINESS SEGMENTS 51467 .638 .691 0 2.303 BOARD AFFILIATIONS 51467 .203 .101 0 .875 BOARD INDEPENDENCE 51467 .729 .152 0 1 CRISIS 51467 .146 .353 0 1

20

Table 2: Pearson correlation matrix

This table shows all correlation coefficient of each variable used in this analysis with all other variables. All correlation coefficients that are larger than .7 are deemed to be significant. A description of the variables can be found in Appendix A and the coefficients are discussed in the results section.

Variables (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16) (1) LEV 1.000 (2) BUREAUCRAT 0.051 1.000 (3) POLIT 0.074 0.038 1.000 (4) FOREIGN ASSETS 0.071 -0.005 0.018 1.000 (5) LONG-TERM DEBT 0.234 -0.004 0.018 0.005 1.000 (6) SHORT-TERM DEBT 0.281 0.017 0.006 -0.003 0.618 1.000 (7) INTEREST RATE 0.343 -0.005 0.019 -0.007 0.194 0.169 1.000 (8) TOBINS Q -0.235 -0.002 0.002 -0.048 -0.087 -0.105 -0.189 1.000 (9) BOARDSIZE 0.293 0.097 0.146 0.139 0.014 0.031 0.012 -0.092 1.000 (10) INVENTORY -0.049 -0.052 -0.072 0.052 -0.032 0.078 -0.053 -0.111 -0.036 1.000 (11) ROA 0.031 0.027 0.032 0.072 -0.038 -0.030 -0.092 -0.041 0.146 0.094 1.000

(12) DEFERRED TAX DIFFERENCE -0.124 -0.018 -0.024 0.104 -0.014 0.002 -0.058 -0.005 -0.011 0.170 0.571 1.000

(13) BUSINESS SEGMENTS 0.248 0.062 0.067 0.114 0.041 0.020 0.043 -0.173 0.304 -0.002 0.155 0.036 1.000

(14) BOARD AFFILIATIONS -0.064 -0.038 -0.015 -0.011 -0.020 -0.038 -0.003 0.070 -0.426 -0.062 -0.071 -0.003 -0.080 1.000

(15) BOARD INDEPENDENCE 0.075 0.062 0.033 0.064 -0.026 -0.056 -0.117 -0.030 0.162 -0.042 0.032 0.025 0.133 0.014 1.000

21

Table 3: Bureaucrat directors on leverage

This table includes the regression results of the first two hypotheses. The main effect of the first model measures the effects of bureaucrat- and other PC directors on leverage (LEV). Subsequent models use the same explanatory variables only the dependent variable changes. Column (2) shows results for the same regression as in (1) only in (2) I excluded the financial sector. In column (3) the dependent variable is

INTEREST RATE, in (4) LONG-TERM DEBT and in (5) SHORT-TERM DEBT. Bureaucrat directors are indicated by BUREAUCRAT and other PC directors by POLIT. All independent variables are lagged 1 year. Coefficients that are relevant to the hypotheses are bold.

Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1

(1) (2) (3) (4) (5)

VARIABLES (1 year lagged) LEV LEV INTEREST RATE LONG-TERM DEBT SHORT-TERM DEBT BUREAUCRAT 0.1739*** 0.1862*** 0.0015 0.0331*** 0.0272 (0.0545) (0.0564) (0.0072) (0.0125) (0.0220) POLIT 0.0105 -0.0168 -0.0006 -0.0050 -0.0103 (0.0598) (0.0606) (0.0077) (0.0102) (0.0186) FOREIGN ASSETS 0.0000 0.0029 -0.0001 -0.0020 -0.0018 (0.0108) (0.0108) (0.0013) (0.0026) (0.0037) LEV 0.0172*** 0.0152*** 0.0009** 0.0026*** -0.0004 (0.0037) (0.0038) (0.0004) (0.0008) (0.0012) TOBINS Q -0.0086*** -0.0092*** -0.0014*** -0.0008*** -0.0017*** (0.0015) (0.0015) (0.0002) (0.0003) (0.0004) BOARDSIZE 0.0117 0.0111 0.0021 0.0038 0.0069* (0.0124) (0.0129) (0.0016) (0.0028) (0.0040) INVENTORY 0.1383*** 0.1327*** 0.0013 0.0046 0.0266* (0.0399) (0.0402) (0.0052) (0.0069) (0.0140) ROA -0.0810*** -0.0831*** -0.0180*** -0.0104*** -0.0104** (0.0132) (0.0132) (0.0021) (0.0034) (0.0049) DEFERRED TAX DIFFERENCE -0.0925*** -0.0895*** 0.0056*** -0.0012 -0.0001 (0.0121) (0.0122) (0.0016) (0.0027) (0.0043) BUSINESS SEGMENTS 0.0117*** 0.0115*** -0.0003 0.0013 0.0008 (0.0035) (0.0037) (0.0004) (0.0012) (0.0012) BOARD AFFILIATIONS -0.0407 -0.0431 0.0117*** 0.0133** 0.0188* (0.0322) (0.0337) (0.0040) (0.0065) (0.0103) BOARD INDEPENDENCE 0.0378*** 0.0415*** -0.0175*** -0.0069** -0.0137*** (0.0129) (0.0134) (0.0017) (0.0029) (0.0041) Constant 0.3215*** 0.3283*** 0.0290*** -0.0036 0.0288** (0.0381) (0.0399) (0.0050) (0.0078) (0.0122) Observations 25,774 23,578 25,318 24,242 25,306 R-squared 0.0583 0.0592 0.0388 0.0044 0.0040 Number of firm_id 3,711 3,382 3,668 3,649 3,687

22

Table 4: Bureaucrat directors during the Global Financial crisis

This table shows the regression results that are related to the third hypothesis. The coefficients are a measure of the effect of bureaucrat- and other PC directors on leverage during the crisis. I measure hypothesis 1 while controlling for the GFC (1), I interact bureaucrat directors

and other PC directors with the GFC (2) and lastly, I repeat (2) but exclude the financial sector in (3). In column (4) I change the dependent variable to INTEREST RATE. Leverage is measured by the variable LEV. Bureaucrat directors are indicated by BUREAUCRAT and other PC

directors by POLIT. The dummy variable CRISIS equals 1 during the Global Financial crisis. All independent variables are lagged 1 year. Coefficients that are relevant to the hypotheses are bold.

Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1

(1) (2) (3) (4)

VARIABLES (1 year lagged) LEV LEV LEV INTEREST RATE BUREAUCRAT 0.1754*** 0.1604*** 0.1702*** 0.0024 (0.0545) (0.0549) (0.0569) (0.0078) CRISIS -0.0078*** -0.0098*** -0.0110*** -0.0014*** (0.0019) (0.0022) (0.0023) (0.0003) CRISIS * BUREAUCRAT 0.0853* 0.0990* 0.0023 (0.0498) (0.0537) (0.0079) POLIT 0.0123 0.0013 -0.0289 -0.0031 (0.0598) (0.0585) (0.0595) (0.0081) CRISIS * POLIT 0.0717 0.0930 -0.0004 (0.0641) (0.0669) (0.0076) FOREIGN ASSETS 0.0006 0.0007 0.0038 -0.0001 (0.0107) (0.0107) (0.0108) (0.0013) LEV 0.0173*** 0.0173*** 0.0154*** 0.0011** (0.0037) (0.0037) (0.0038) (0.0005) TOBINS Q -0.0087*** -0.0087*** -0.0093*** -0.0015*** (0.0015) (0.0015) (0.0015) (0.0002) BOARDSIZE 0.0106 0.0107 0.0102 0.0011 (0.0123) (0.0124) (0.0129) (0.0016) INVENTORY 0.1346*** 0.1344*** 0.1285*** 0.0003 (0.0398) (0.0398) (0.0400) (0.0053) ROA -0.0830*** -0.0832*** -0.0855*** -0.0185*** (0.0133) (0.0133) (0.0133) (0.0021) DEFERRED TAX DIFFERENCE -0.0911*** -0.0911*** -0.0880*** 0.0057***

(0.0122) (0.0122) (0.0123) (0.0016) BUSINESS SEGMENTS 0.0114*** 0.0113*** 0.0112*** -0.0004 (0.0035) (0.0035) (0.0037) (0.0004) BOARD AFFILIATIONS -0.0465 -0.0466 -0.0497 0.0102** (0.0321) (0.0321) (0.0336) (0.0042) BOARD INDEPENDENCE 0.0409*** 0.0408*** 0.0448*** -0.0168*** (0.0129) (0.0129) (0.0133) (0.0017) Constant 0.3242*** 0.3243*** 0.3312*** 0.0296*** (0.0381) (0.0381) (0.0400) (0.0051) Observations 25,774 25,774 23,578 23,154 R-squared 0.0590 0.0593 0.0605 0.0427 Number of firm_id 3,711 3,711 3,382 3,342 Robust standard errors in parentheses

23 Appendix A: Variable description

Variable Measurement Source

LEV Total debt / total assets Worldscope/BoardEx

BUREAUCRAT Number of bureaucrat directors on a board, proportional

to total board size Hand-collected POLIT Number of PC directors excluding bureaucrats on a

board, proportional to total board size Hand-collected FOREIGN ASSETS Foreign assets / total assets Worldscope/BoardEx LONG-TERM DEBT Long-term debt / total assets Worldscope/BoardEx SHORT-TERM DEBT Short-term debt / total assets Worldscope/BoardEx INTEREST RATE The average interest that is paid on debt Worldscope/BoardEx TOBINS Q (total assets – Shareholders equity + Market capitalization)

/ total assets Worldscope/BoardEx BOARDSIZE Total number of directors on a board Worldscope/BoardEx

INVENTORY Inventory / total assets Worldscope/BoardEx

ROA Operational income / total assets Worldscope/BoardEx

DEFERRED TAX

DIFFERENCE The difference between the income tax and deferred income tax, proportional to total assets Worldscope/BoardEx BUSINESS SEGMENTS The logarithm of the number of business segments of a

firm Worldscope/BoardEx

BOARD AFFILIATIONS Average number of current boards a directors sit on Worldscope/BoardEx BOARD INDEPENDENCE The number of non-executive directors proportional to

board size Worldscope/BoardEx CRISIS A dummy that equals 1 during the global financial crisis