Political connections and corporate debt: Evidence from two U.S. election campaigns

Political connections and corporate debt: Evidence from two U.S. election campaigns

G Model ARTICLE IN PRESS QUAECO-1263; No. of Pages 11 The Quarterly Review of Economics and Finance xxx (2019) xxx–xxx Contents lists available at...

728KB Sizes 0 Downloads 5 Views

G Model

ARTICLE IN PRESS

QUAECO-1263; No. of Pages 11

The Quarterly Review of Economics and Finance xxx (2019) xxx–xxx

Contents lists available at ScienceDirect

The Quarterly Review of Economics and Finance journal homepage: www.elsevier.com/locate/qref

Political connections and corporate debt: Evidence from two U.S. election campaigns Imed Chkir a,∗ , Mohamed Imen Gallali b , Manara Toukabri b a b

Telfer School of Management, University of Ottawa, 55 Laurier E. (DMS 7107), Ottawa, ON, K1N 6N5, Canada École Supérieure de Commerce de Tunis, La Manouba, Tunis, 2010, Tunisia

a r t i c l e

i n f o

Article history: Received 6 February 2018 Received in revised form 10 March 2019 Accepted 4 May 2019 Available online xxx JEL classification: D72 G32 Keywords: Political connections Election campaign Financial leverage Cost of debt

a b s t r a c t In this paper, we analyze the effect of political connections on the level and cost of debt of U.S. firms. We identify politically connected firms through manually collecting data on their financial contributions to the two major U.S. political parties, the Democratic Party and the Republican Party, during the 2008 and 2012 U.S. election campaigns. Our main results indicate that during the 2009 to 2015 period, politically connected firms had a significantly higher debt ratio and experienced a lower cost of debt than did unconnected firms. Additional tests show that these results are robust to different model specifications and suggest that politically connected firms benefit from a higher indebtedness level than do unconnected firms and that their political connections give them an advantage that offsets the negative effect of high leverage on their cost of debt. © 2019 Board of Trustees of the University of Illinois. Published by Elsevier Inc. All rights reserved.

1. Introduction To protect themselves and their ownership rights from governments’ excessive control, regulatory rigidity, high taxation, and dysfunctional markets, many entrepreneurs seek political coverage by creating formal or informal political networking (Li, Meng, & Zhang, 2006). Ling, Zhou, Liang, Song, and Zeng (2016) consider a close connection with the government to be valuable social capital for a company. The origins of the notion of corporate political connection go back to the resources dependency theory proposed by Pfeffer and Salancik (1978). The theory stipulates that firms are naturally dependent on external organizations and structures. This dependence allows firms to both mitigate volatility shocks that may affect their environment and reduce transaction costs (Hillman, 2005). Nevertheless, it is only recently that the notion of corporate political connection has gained solid theoretical and empirical grounds, particularly after the seminal paper of Faccio (2006), who identified politically connected firms (PCFs) in 47 countries. Indeed,

∗ Corresponding author. E-mail addresses: [email protected] (I. Chkir), [email protected] (M.I. Gallali), [email protected] (M. Toukabri).

in the past decade financiers and political scientists have shown a growing interest in understanding the relationship between firms and politicians, a relationship that rests on an exchange of benefits agreement between these two parties (Tahoun, 2014). This agreement is most often implicit and unwritten in order to avoid accusations of favoritism, malfeasance, or other illegal undertakings. According to Khwaja and Mian (2005), political connections materialize through exchanges of gifts and privileges between politicians and their friends who work in the firms. Claessens, Feijen, and Laeven (2008) show that for firms, the benefits of political connections manifest through several channels, including easier access to debt financing. Faccio, Masulis, and McConnell (2006) also show that PCFs are more likely to receive donations and government bailouts during a financial hardship. Relatedly, Chaney, Faccio, and Parsley (2011) and Leuz and Oberholzer-Gee (2006) show that PCFs, which receive lesser pressure from the financial markets, can afford to be opaque in the quality of their financial reporting. In the same vein, Tahoun and van Lent (2019) suggest that political connections played a significant role in the bailout decisions of the 2008 financial crisis. Indeed, the authors found that government support to financial institutions under the 2008 Emergency Economic Stabilization Act (EESA) was biased by the personal wealth of politicians who owned stocks in these institutions.

https://doi.org/10.1016/j.qref.2019.05.003 1062-9769/© 2019 Board of Trustees of the University of Illinois. Published by Elsevier Inc. All rights reserved.

Please cite this article in press as: Chkir, I., et al. Political connections and corporate debt: Evidence from two U.S. election campaigns. The Quarterly Review of Economics and Finance (2019), https://doi.org/10.1016/j.qref.2019.05.003

G Model QUAECO-1263; No. of Pages 11 2

ARTICLE IN PRESS I. Chkir et al. / The Quarterly Review of Economics and Finance xxx (2019) xxx–xxx

On the other side, political connections can also adversely affect firms. Shleifer and Vishny (1994) develop a model of bargaining between politicians and managers in which politicians naturally extract resources from firms to serve their own political and social interests. Brockman, Rui, and Zou (2013) find that the negatives of this rent extraction may overweight the positive effect of political connections. Indeed, exploring the stock and operating performance of a sample of merger and acquisition (M&A) deals across 22 countries, the authors find that in contexts of a strong rule of law and low level of corruption, politically connected acquirers underperform their matched unconnected bidders for 3 years after the deal. In an international sample of 31 countries, Boubakri, El Ghoul, and Saffar (2013) find that PCFs are more likely to hold cash. The authors explain this result using the incentives politicians have to use these firms as “cash cows” and by the fact that PCFs suffer from an accentuated agency problem. Bertrand, Kramarz, Schoar, and Thesmar (2006) find that, during election years in France, PCFs inefficiently increase their plant and jobs creation rate at a high cost, to help their incumbent politicians be reelected. Fan, Wong, and Zhang (2007) find that newly privatized Chinese firms with politically connected CEOs underperform those without politically connected CEOs and are more likely to appoint bureaucrats, rather than competent directors, to the board. Boubakri, Cosset, and Saffar (2008) report similar results using an international sample covering 41 countries. The authors find that newly privatized firms with a politician or an ex-politician on their board of directors exhibit poor accounting performance compared with nonconnected firms. Whether the negative or positive effects of political connections on firms dominate may depend on different factors, such as the nature and/or the level of the connection, or the macroeconomic environment of the firm. In this sense, Brockman et al. (2013) find that in contexts of a strong rule of law and a low level of corruption, politically connected acquirers underperform unconnected ones. In contrast, when acquirers are located in a corrupt system with a weak level of regulation, politically connected bidders outperform their counterpart. Wang (2015) finds that the presence of an independent director on the board of Chinese firms with political ties has a different impact depending on the nature of the controlling shareholders. On one hand, privately controlled, connected firms outperform their nonconnected counterparts because of firms’ preferential access to external debt financing and more subsidies from the government. On the other hand, state-controlled, connected firms suffer from an accentuated agency conflicts, the expropriation of minority investors’ interest, and overinvestment decisions. In an international sample from 43 countries, Borisova, Fotak, Holland, and Megginson (2015) find that government-owned firms (an extreme form of political connection) pay a higher cost of debt compared with their nongovernment peers during noncrisis years, but this finding reverses during years of financial and other banking crises. A number of studies have explored the effect of corporate political connections on firms’ financial decisions and performance, such as the default probability of PCFs (Faccio et al., 2006), PCFs’ stock returns (Cooper, Gulen, & Ovtchinnikov, 2010; Fisman, 2001; Goldman, Rocholl, & So, 2009), PCFs performance (Ding, Jia, Wu, & Zhang, 2014), PCFs’ cost of equity (Boubakri, Guedhami, Mishra, & Saffar, 2012), post-privatization performance of PCFs (Boubakri et al., 2008; Fan et al., 2007), and the trading decisions of investors (Shen & Lin, 2015). However, few studies have examined the effect of corporate political connections on firms’ level and cost of debt (Bliss & Gul, 2012a, 2012b). In addition, and excepting the Houston, Jiang, Lin, and Ma (2014) study on the cost of private bank loans of U.S. PCFs, most studies have focused on emerging markets when considering this issue. Countries with emerging markets are often characterized by a weak level of regulation, dysfunctional financial markets, and a high level of corruption. Such weaknesses tend to

root political connections (Faccio, 2006; Faccio et al., 2006). Testing the effects of political connections in the United States is more difficult. In fact, it is easier to assess the benefits of this connection in countries with interventionist governments, a high level of corruption, a weak law system, underdeveloped financial systems, and prevalent state ownership of media and banks (Beck, Demirgüc¸Kunt, & Levine, 2006; Borisova et al., 2015; Brockman et al., 2013; Dinc¸, 2005; Houston, Lin, & Ma, 2011; La Porta, Lopez-de-Silanes, & Shleifer, 2002). Political systems of developed countries like that of the United States stand out for their strong democratic roots, fairness, and government independence. In addition, the United States is quite an exception to the state-owned banks phenomenon largely widespread around the world (Dinc¸, 2005). As a result, international studies on the effect of political connections that have used the Faccio (2006) measure of this connection have identified very few U.S. PCFs (13 of 458 in Faccio, 2010; 2 of 234 in Boubakri, Cosset, & Saffar, 2012).1 Nevertheless, a scan of recent history shows that the American context has experienced several episodes of political corruption (Goldman et al., 2009). In addition, more and more U.S. firms compete to support and fund the election campaigns of political parties (Cooper et al., 2010). In this context, this study examines corporate political connections in the United States to determine whether U.S. PCFs have preferential access to public debt compared with politically unconnected firms (PUFs). To this end, we manually collected data provided by the U.S. Federal Election Commission and the Center for Responsive Politics on the participation of firms in the financing of political parties during the 2008 and 2012 U.S. presidential election campaigns, through the creation of political action committees (PACs).2 Thus, we have created an original database of U.S. PCFs. To the best of our knowledge, this is the first study that examines the impact of political connections on the debt level of U.S. firms. We also contribute to the extant literature by analyzing the cost of the total debt of U.S. PCFs and by using the financing of political parties during the years of presidential elections as a measure of political connection. This paper is structured as follows: the second section reviews the literature on political connections. The third section presents our research methodology. The fourth section discusses the obtained results. The fifth section concludes the paper.

2. Review of the literature 2.1. Political connection: definition and measures The notion of corporate political connection is well known and documented in the literature, yet it remains difficult to identify and measure. Indeed, it goes back to the resources dependency theory of Pfeffer and Salancik (1978). The corporate political connection denotes an implicit benefits exchange contract between managers and politicians. On the one hand, managers are interested in profits, subsidies, and protections to their firms, and, on the other hand, politicians are interested in different forms of financial benefits, funding, and votes during elections (Shleifer & Vishny, 1994). Zhou (2013) believes that a political connection is a form of informal contract that gives a company some privileges for its functioning and performance.

1

See Section 2.1 for more details about this measure. The Federal Election Campaign (FEC) defines these funds as an entity separate from the political party committees or candidate committees. PACs can be funded by companies or unions using independent and separate funds to support a political party and/or candidate in any election campaign 2

Please cite this article in press as: Chkir, I., et al. Political connections and corporate debt: Evidence from two U.S. election campaigns. The Quarterly Review of Economics and Finance (2019), https://doi.org/10.1016/j.qref.2019.05.003

G Model QUAECO-1263; No. of Pages 11

ARTICLE IN PRESS I. Chkir et al. / The Quarterly Review of Economics and Finance xxx (2019) xxx–xxx

The financial literature identifies PCFs in different ways, which are summarized below: - A PCF is a firm with a major shareholder (owning more than 10% of company’s shares) or a member of the board of directors who holds an influential political position (president, minister, Member of Parliament) (Faccio et al., 2006). - PCFs can refer to close relationships (family or friends) between a member of the board of directors or a major shareholder and an influential member of government (president or a minister) or a member of parliament (Faccio et al., 2006). - PCFs can refer to a major shareholder or a member of the board of directors who is a member of a political party (Faccio et al., 2006). - PCFs can refer to the company funds used for presidential or parliamentary election campaigns (Claessens et al., 2008). In the United States, political life is centralized on two main parties—the Republican Party and the Democratic Party—whose ideologies diverge considerably. Their differences affect the perception of the role of the state, corporate governance policies, taxation policies, and economic aggregates. Under the U.S. campaign financing legislation, corporate contributions to an election campaign funding must be made through funds created voluntarily and separately from the firm’s general treasury. Contributions to these funds can be made by shareholders, directors, officers, and employees and are limited in amounts. These funds are known as PACs. Stratmann (1998) indicates that PACs are forms of friendly contracts between the funding company and the political beneficiary, as no third party is present to sanction PACs’ malfunction. Thus, PACs help to create a solid mechanism for managing the relationship between the firm and politicians, while also setting up a permanent and a repetitive interaction between them (Tahoun, 2014). Chen, Parsley, and Yang (2014) note that firms’ involvement in U.S. politics has increased after the Citizens United v. Federal Election Commission decree of 2010, which lifted the restriction on donation limits and required PACs’ to disclose donors’ names.3 2.2. Political connection: Effect on the company As noted above, most studies of corporate political connections have focused on emerging countries. Many of these studies have explored the effect of corporate political connections on firms’ financial decisions and performance. For example, Fisman (2001) found that stock returns of Indonesian PCFs significantly dropped each time rumors about President Suharto’s deteriorating health reached the news. Ding et al. (2014) found that operational performance, measured by return on assets, of Chinese PCFs is higher than that of PUFs. Johnson and Mitton (2003) show that Malaysian firms that had connections with the prime minister were immune to shocks from the Asian economic crisis, and those that were slightly affected benefited from restructuring policies. On the other hand, Benjamin, Zain, and Abdul Wahab (2015) found that Malaysian investors avoid PCFs, which are risky because the likelihood of policy change following an election is high. In an international setting covering 47 different countries, Faccio (2010) finds that, in spite of their higher leverage and higher market shares, PCFs underperform PUFs. In addition, she shows that PCFs enjoy lower marginal tax rates and a higher market power compared with PUFs and that these findings are more pronounced in countries with a higher degree of corruption.4 Using data from the

3 However, that individual donations by the firm’s management outside of the firm’s PAC are not recorded could potentially underestimate the level of the firm’s political connection. 4 The sample included only 13 U.S. PCFs.

3

Faccio (2006) political connection measure, Boubakri, Cosset et al. (2012) find that PCFs from 23 countries benefited from a lower equity cost and achieved better operational performance than did PUFs during the 1989 to 2003 period.5 In the United States, Goldman et al. (2009) report positive abnormal stock returns for U.S. firms after firms announce the appointment of a member with political connection to their board of directors. In the same U.S. context, Cooper et al. (2010) find a positive relationship between firms’ financial contributions to political parties and their stock returns. Santa-Clara and Valkanov (2003) report that election results significantly affect stock market returns depending on the winning party (Republican or Democrat). Hutton, Jiang, and Kumar (2014) find that the conservative culture of Republican managers reflects in the policies of the firms they manage. Such firms are characterized by low equity costs, low investment in research and development, fewer risky investments, and high profitability. With regard to debt, many studies have sought to examine the effect of political connections on the level and cost of debt in emerging markets. Leuz and Oberholzer-Gee (2006) show that Indonesian PCFs have easier access to debt. The authors point out that firms with political connections have preferential access to domestic financing and, therefore, do not need access to international capital markets. The same result is reported by Chan, Dang, and Yan (2012) and Cull, Li, Sun, and Xu (2015) for Chinese PCFs. The authors find that firms whose CEOs have political ties have no financing constraints and have easier access to external credit. In the same vein, Fan, Rui, and Zhao (2008) use an event study approach to investigate 23 corruption scandals in China and show that firms connected to corrupt government officials significantly reduced their debt level following the arrest of corrupt officials. This finding indicates that these firms were receiving preferential financing treatments. Claessens et al. (2008) finds that most Brazilian firms contributing to election campaigns enjoy preferential access to debt financing. The authors find that firms that have donated to political parties experience a significant increase in their debt levels in the postelection period if donations are made to the winning party. Belghitar, Clark, and Saeed (2016) believe that political connections reinforce the position of firms and strongly influence their financial policies. They also show that Pakistan’s PCFs have, on average, a higher level of indebtedness, which decreases when politicians with whom they are connected step down from office. In the same Pakistani context, Khwaja and Mian (2005) show that firms with political ties receive preferential treatment from financial institutions despite a 50% higher default rate than that of PUFs. The authors point out that this preferential treatment enjoyed by PCFs mainly comes from government-owned banks, which could partly explain the failure of state-owned banks. Bliss and Gul (2012a, 2012b), Fraser, Zhang, and Derashid (2006) and Johnson and Mitton (2003), and study the relationship between political connection and leverage in Malaysian firms. Their main findings reveal that PCFs, compared with PUFs, have a higher debt level, while bearing a higher cost of debt. On the other hand, Bunkanwanicha and Wiwattanakantang (2009) find no significant effect of political connections on access to debt financing by Thai PCFs, which do not enjoy preferential treatment from creditors. To the best of our knowledge, Houston et al. (2014) is the only study that focuses on U.S. PCFs financing. The authors only focus on the cost of private credit of U.S. firms in which at least one board member holds a senior political position. Their results show that U.S. PCFs were less vulnerable to the repercussions of the 2007–2008 financial crisis, thanks to access to several financial

5 The United States was among the countries surveyed; however, the sample included only two U.S. PCCs.

Please cite this article in press as: Chkir, I., et al. Political connections and corporate debt: Evidence from two U.S. election campaigns. The Quarterly Review of Economics and Finance (2019), https://doi.org/10.1016/j.qref.2019.05.003

G Model QUAECO-1263; No. of Pages 11

ARTICLE IN PRESS I. Chkir et al. / The Quarterly Review of Economics and Finance xxx (2019) xxx–xxx

4

Table 1 Description of the sample of politically connected firms.

Number of PCFs PCFs exclusively donating to the Democratic Party PCFs exclusively donating to the Republican Party Total amount of donations to the Dem. Party(in million $) Total amount of donations to the Rep. Party(in million $)

2008

2012

2008 et 2012

403 17 22 27.818 31.614

358 17 23 27.035 40.646

478 19 30 54.853 72.260

This table describes the sample of politically connected firms (PCFs) as well as the amounts donated to the Democratic Party and the Republican Party during the two U.S. election campaigns of 2008 and 2012.

privileges, including greater flexibility in banking credit clauses, which allowed PCFs to meet the required liquidity threshold. The authors report that U.S. PCFs had access to bank credit at lower costs than those charged to PUFs. In light of these results, consensus among scholars seems to indicate that political connections favor firms by allowing them to access higher levels of debt financing in emerging countries. We hypothesize that this is also true in the United States, and we test whether U.S. PCFs, compared with PUFs, have a higher debt financing ratio. Thus, we test the following hypothesis: H1.

U.S. PCFs have a higher debt ratio than do PUFs.

The results of previous studies are, however, mixed regarding the cost of debt of PCFs. In this paper, we test whether U.S. PCFs, compared with PUFs, are also favored by facing a lower debt cost. H2.

U.S. PCFs, compared with PUFs, have a lower cost of debt.

Or we test whether U.S. PCFs’ higher debt level translates to a higher cost of debt given an increasing financial risk. H2’.

U.S. PCFs, compared with PUFs, have a higher cost of debt.

3. Research methodology 3.1. Sample selection Our sample includes all U.S. firms covered by the Compustat database during the period 2009–2015, except for firms in the financial and utility sectors. PCFs are identified through their donations during the 2008 and 2012 U.S. election campaigns. The former elections resulted in a transition from a Republican administration under the leadership of George W. Bush to a democratic administration under the leadership of Barack Obama, who was reelected in 2012. First, we manually collected information on firms with PACs during the 2008 and 2012 election campaigns. These data are provided by the Federal Election Commission (FEC) website.6 Then we manually collected data on the funding and the proportions allocated to the Democratic Party and the Republican Party from the “Center for Responsive Politics” website.7 Finally, we performed a manual cross-checking of the names of the PACs and the firms in the Compustat database, from which we obtained accounting and financial data. Table 1 describes the sample of PCFs as well as the amounts donated to the Democratic and Republican parties during the two election campaigns. We were able to identify 403 donor firms during the 2008 election campaign and 358 during the 2012 campaign.8 Most PCFs in our sample donated in both election campaigns. Indeed, we identified 478 different firms that formed a PAC

6

http://www.fec.gov/. Last accessed December 18, 2017. https://www.opensecrets.org/. Last accessed December 18, 2017. Goldman et al. (2009) identify 315 donor firms during the 2000 election campaign, whereas Cooper et al. (2010) report that, on average, only 9.49% of firms covered by Compustat donate through PACs. 7 8

in 2008 and/or 2012. Moreover, and in line with previous studies exploring political money in the U.S. context (see, for example Cooper et al., 2010; Snyder, 1990), we note that most firms donated to both the Democratic Party and the Republican Party. Indeed, in the two election campaigns, only 19 firms exclusively donated to the Democratic Party and 30 to the Republican Party. In addition, the Republican Party has benefitted from the highest amount of donations. 3.2. Model and variables Drawing on Bliss and Gul (2012a; 2012b) and Boubakri, Cosset et al. (2012) we first test the effect of political connections on firms’ debt level using the following model: LEV = ˇ0 + ˇ1 PCN + ˇ2 SIZE + ˇ3 TAN+ ˇ4 CDT + ˇ5 CUR +ˇ6 ROA + ˇ7 GRW + ˇ8 REA + ˇ9−18 Sector + ˇ19−25 Year + ε. (1) This model includes the control variables suggested by the previous literature on capital structure, the cost of debt, and the effect of political connections. These variables are defined as follows: V: Total debt ratio, which is measured by the ratio of long-term and short-term debt to total assets PCN: Political connection, which is a binary variable that takes the value of 1 in for the years 2009 to 2011 for firms with a PAC during the 2008 election and for years 2013 to 2015 for those with a PAC during the 2012 election, and 0 otherwise SIZE: Firm size, which is measured by the natural logarithm of total assets TAN: Ratio of property, plant, and equipment to total assets CDT: Cost of debt, which is measured by the ratio of interest expenses to the sum of long-term and short-term debt CUR: Ratio of current assets to current liabilities ROA: Return on assets, which is measured by the ratio of net income to total assets GRW: Growth potential, which is measured by the ratio of market to book value of equity REA: Ratio of retained earnings to total assets Moreover, and in line with Bliss and Gul (2012a, 2012b); Boubakri, Cosset et al. (2012), and Boubakri, Guedhami et al. (2012), we introduce binary variables to control for years and industrial sector fixed effects. Next, we test the effect of political connections on the cost of debt using the following model: CDT = ˇ0 + ˇ1 PCN + ˇ2 SIZE + ˇ3 TAN+ ˇ4 LEV + ˇ5 LEV × PCN +ˇ6 CUR + ˇ7 ROA + ˇ8 GRW + ˇ9 REA +ˇ10−19 Sector + ˇ20−26 Year + ε,

(2)

where variables are defined in Model (1). Drawing on Chaney et al. (2011), who analyzed the effect of accounting information quality on the cost of debt of PCFs, we also include an interaction factor between political connection and leverage to test whether the association between the cost of debt and leverage is different for connected firms, compared with unconnected firms.

Please cite this article in press as: Chkir, I., et al. Political connections and corporate debt: Evidence from two U.S. election campaigns. The Quarterly Review of Economics and Finance (2019), https://doi.org/10.1016/j.qref.2019.05.003

G Model

ARTICLE IN PRESS

QUAECO-1263; No. of Pages 11

I. Chkir et al. / The Quarterly Review of Economics and Finance xxx (2019) xxx–xxx

5

Table 2 Correlation coefficients.

CDT SIZE TAN CUR ROA GRW PCN REA

LEV

CDT

SIZE

TAN

CUR

ROA

GRW

PCN

0.014*** 0.069*** 0.024*** −0.017*** −0.0264*** 0.0013 0.039*** −0.394***

−0.035*** −0.085** 0.017*** −0.055*** −0.002 −0.019** −0.035***

0.062*** −0.034*** 0.200*** −0.004 0.353*** 0.333***

−0.023*** 0.017*** 0.006 −0.019*** 0.085***

0.002 0.000 −0.011** −0.05

0.068*** 0.041*** 0.370***

0.001 0.015**

0.065***

This table presents the correlation coefficients between the variables used in models (1) and (2). The variables are defined as follows: LEV: Total debt ratio measured by the ratio of long-term and short-term debts to total assets; CDT: Cost of debt measured by the ratio of interest expense to the sum of long-term and short-term debt; SIZE: Natural logarithm of total assets; TAN: Ratio of property, plant and equipment to total assets; CUR: Ratio of current assets to current liabilities; ROA: ratio of net income to total assets; GRW: Ratio of market to book value of equity; REA: ratio of retained earnings to total assets; REA: ratio of retained earnings to total assets; PCN: Political connection is a binary variable that takes the value of one in years 2009 to 2011 for firms with a PAC during the 2008 election, and in years 2013 to 2015 for those with a PAC during the 2012 election, and 0 otherwise. (*), (**), (***) denote respectively degree of significance at the 10%, 5% and 1% levels.

Table 3 Descriptive statistics. Entire Sample

LEV CDT SIZE TAN CUR ROA GRW REA

PUFs

PCFs

Mean

Stand.Dev

Median

Mean

Stand.Dev

Median

Mean

Stand.Dev

Median

0.266 0.046 5.336 0.287 5.230 −0.230 2.990 −2.321

0.754 0.609 2.523 0.172 4.629 1.762 6.799 10.025

0.141 0.020 5.288 0.285 1.971 0.010 1.990 −1.211

0.263 0.047 5.103 0.288 5.451 −0.249 2.972 −2.492

0.776 0.631 2.396 0.289 4.881 1.820 7.642 10.344

0.128 0.020 5.083 0.117 2.017 0.006 1.751 −0.168

0.300 0.025 8.741 0.267 3.940 0.043 3.078 0.136

0.241 0.019 1.762 0.214 2.480 0.119 5.466 0.789

0.250 0.022 8.920 0.198 2.607 0.057 2.315 0.246

t-stat

z-stat

−4.001*** 1.840* −79.584*** 1.733 1.342 −8.576*** −0.144 −13.510***

−26.119*** −5.237*** −66.033*** −6.405*** −15.202*** −35.811*** −13.404*** −42.417***

This table reports the descriptive statistics of the variables used in models (1) and (2). The total sample is divided into politically connected firms (PCFs) and politically unconnected firms (PUFs) according to the possession of a PAC during the 2008 and 2012 election years. The variables are defined as follows: LEV: Total debt ratio measured by the ratio of long-term and short-term debts to total assets; CDT: Cost of debt measured by the ratio of interest expense to the sum of long-term and short-term debt; SIZE: Natural logarithm of total assets; TAN: Ratio of property, plant and equipment to total assets; CUR: Ratio of current assets to current liabilities; ROA: ratio of net income to total assets; GRW: Ratio of market to book value of equity; REA: ratio of retained earnings to total assets. (*), (**), (***) denote respectively degree of significance at the 10%, 5% and 1% levels.

In line with the financial theory of capital structure (see for example Harris & Raviv, 1991; Jensen, 1986; Jensen & Meckling, 1976), we expect the effect of the variables SIZE, TAN, and GRW to be positive on the debt level (i.e., a positive sign in Model (1)) and negative on the cost of debt (i.e., a negative sign in Model (2)), because size and tangible fixed assets are synonymous with additional guarantees for creditors, while growth potential indicate additional need for financing. We also expect that the effect of the CUR, ROA, and REA variables will be negative on the level and the cost of debt, because the profitability and liquidity levels are synonymous with a lower need for financing by debt and additional guarantees for creditors. We also expect a positive sign for the LEV variable in Model (2) because higher leverage induces a higher risk, which results in a higher cost of debt. Finally, we introduce the variable CDT in Model (1) to control for whether firms choose less leverage if debt becomes expensive. In addition, and in both models, we compute clustered standard errors to account for the correlations between firms’ leverage ratios within industrial sectors.

comparison tests of these variables between the two samples are reported in the last two columns. Like with the results of Bliss and Gul (2012b) for Malaysian firms, this table shows a significant difference between the mean debt ratio of U.S. PCFs and U.S. PUFs. In fact, PCFs have a mean debt ratio of 0.30 against 0.26 for PUFs, and the difference is significant at the 1% level. In addition, the mean cost of debt is significantly lower for PCFs (2.5% vs. 4.7% for PUFs). For the control variables, the results in Table 3 show that PCFs are, on average, significantly larger, more profitable, and have a higher retained earnings ratio than PUFs. In addition, the two samples are not significantly different in terms of the mean of the other control variables. The results in Tables 2 and 3 provide preliminary evidence to support H1 and H2 and reject H2’, suggesting that U.S. PCFs firms have higher debt ratios at a lower cost compared with PUFs. We now extend the univariate analysis by testing Models (1) and (2). Doing so allows us to control for the impact of other variables that may influence the level and the cost of debt. 4.2. Multivariate analysis

4. Results and discussion 4.1. Descriptive statistics and univariate analysis Table 2 presents the correlation coefficients between the variables used in our models. In particular, this table shows that political connection positively correlates with the debt level and negatively correlates with the cost of debt. Table 3 reports the descriptive statistics of the variables used in our models for the entire sample as well as the two samples of PCFs and PUFs. The t- and z-statistics of the mean and the median

4.2.1. Debt level and political connection Table 4 reports the results of the regression of the debt ratio on the binary variable representing PCFs and the different control variables. The results confirm those of the univariate analysis and show that, after controlling for the main factors considered in the literature as affecting debt level, PCFs are significantly more indebted than PUFs during the 2009 to 2015 period (panel A). Indeed, the PCN variable is positive and significant at the 1% level. This result supports H1 and is consistent with those results found in earlier studies about emerging countries (Belghitar et al., 2016; Bliss &

Please cite this article in press as: Chkir, I., et al. Political connections and corporate debt: Evidence from two U.S. election campaigns. The Quarterly Review of Economics and Finance (2019), https://doi.org/10.1016/j.qref.2019.05.003

G Model

ARTICLE IN PRESS

QUAECO-1263; No. of Pages 11

I. Chkir et al. / The Quarterly Review of Economics and Finance xxx (2019) xxx–xxx

6 Table 4 Effect of political connections on debt level.

PCN SIZE TAN CDT CUR ROA GRW REA Intercept Sector Effect Year Effect N Adj. R2

Panel A (Full Sample)

Panel B (2009 to 2011)

Panel C (2013 to 2015)

Panel D (2009)

Coefficient

t-stat

Coefficient

t-stat

Coefficient

t-stat

Coefficient

t-stat

Panel E (2013) Coefficient

t-stat

0.091 0.012 0.155 −0.009 −0.038 −0.119 −0.001 −0.009 −5.578 yes yes 32,622 0.156

5.723*** 7.578*** 9.554*** −0.773 −5.323*** −15.183*** −0.109 −0.785 −7.607***

0.043 0.056 0.062 −0.018 −0.059 −0.243 −0.002 −0.019 −4. 500 yes yes 17,163 0.145

3.433*** 8.021*** 5.446*** −1.633 −4.011*** −21.221*** −1.009 −1.914* −6.653***

0.123 0.094 0.118 −0.002 −0.022 −0.407 0.001 −0.056 −9.864 yes yes 15,456 0.138

3.979*** 6.741*** 9.850*** −0.189 −4.493*** −22.585*** 0.115 −2.102** −3.722***

0.098 0.013 0.128 −0.022 −0.176 −0.288 −0.052 −0.038 0.318 yes no 5,762 0.153

2.751*** 3.453*** 6.418*** −1.140 −8.870*** −14.181*** −1.735 −1.934 10.998***

0.109 0.091 0.141 −0.001 −0.109 −0.422 0.019 −0.056 0.116 yes no 6,005 0.185

2.111** 3.843*** 6.965*** −0.075 −5.580*** −9.776*** 0.980 −2.109** 3.785***

This table reports the results of model (1). Dependent variable is leverage (LEV) measured by the ratio of long-term and short-term debts to total assets. Control variables are: PCN: Political connection, a binary variable that takes the value of one in years for firms with a PAC during the 2008 election, and in years 2013 to 2015 for those with a PAC during the 2012 election, and 0 otherwise; SIZE: Natural logarithm of total assets; TAN: Ratio of property, plant and equipment to total assets; CDT: Cost of debt measured by the ratio of interest expense to the sum of long-term and short-term debt; CUR: Ratio of current assets to current liabilities; ROA: ratio of net income to total assets; GRW: Ratio of market to book value of equity; REA: Ratio of retained earnings to total assets. Robust t-statistics are reported based on industry-clustered standard errors. (*), (**), (***) denote respectively degree of significance at the 10%, 5% and 1% levels. Table 5 Effect of the level and type of political connections on debt level.

CONTDEM CONTREP

Panel A: Democrats in control (2009 and 2010)

Panel B: Divided congress (2011 and 2013)

Model 1.a

Model 1.a

DEM REP SIZE TAN CDT CUR ROA GRW REA Intercept Sector Effect Year Effect N R2

Model 1.b

0.011 (1.604) 0.014 (1.416)

0..010 (2.712)*** 0.120 (2.718)*** −0.089 (−0.901) −0.039 (−4.924)*** −0.215 (−2.101)** 0.001 (0.172) −0.008 (−1.918)* −1.087 (−0.258) yes yes 756 0.090

Panel C: Republicans in control (2014 and 2015) Model 1.b

0.005 (0.109) 0.007 (1.307) 0.152 (3.015)*** 0.012 (0.558) 0.009 (2.683)*** 0.123 (2.775)*** −0.092 (-0.924) −0.037 (−5.107)*** −0.219 (−2.128)** 0.001 (0.180) −0.009 (−1.9140)* −0.946 (−0.237)

0.088

Model 1.a

Model 1.b

0.004 (0.893) 0.017 (1.471)

0.007 (2.757)*** 0.093 (2.287)** −0.118 (−0.748) −0.032 (−5.197)*** −0.477 (−5.012)*** 0.001 (0.235) (−0.018) (−2.233)** −2.455 (−1.247) yes yes 710 0.107

0.033 (0.915) 0.067 (1.115) 0.008 (2.785)*** 0.101 (2.296)** −0.099 (−0.585) −0.036 (−5.228***) −0.511 (−5.204)*** 0.001 (0.312) (−0.017) (−2.197)** −1.744 (−1.043)

0.095

0.014 (2.125)** 0.103 (2.804)*** −0.101 (−0.974) −0.029 (−4.892)*** −0.348 (−5.718)*** 0.001 (0.271) −0.024 (−2.108)** −1.673 (−1.119) yes yes 620 0.110

0.002 (0.087) 0.094 (2.794)*** 0.010 (2.107)** 0.139 (2.954)*** −0.089 (−0.781) −0.033 (−5.015)*** −0.339 (−5.202)*** 0.001 (0.282) −0.030 (−2.110)** −1.055 (−1.008)

0.101

This table reports the results of the regression of leverage (LEV), measured by the ratio of long-term and short-term debts to total assets on the following variables: CONTDEM: Ratio of donations to the Democratic Party to total assets; CONTREP: Ratio of donations to the Republican Party to total assets; DEM: Binary variable that equal to one for firms exclusively donating to the Democratic Party and 0 otherwise; REP: Binary variable equal to one for firms exclusively donating to the Republican Party and 0 otherwise; SIZE: Natural logarithm of total assets; TAN: Ratio of property, plant and equipment to total assets; CDT: Cost of debt measured by the ratio of interest expense to the sum of long-term and short-term debt; CUR: Ratio of current assets to current liabilities; ROA: ratio of net income to total assets; GRW: Ratio of market to book value of equity; REA: Ratio of retained earnings to total assets. Robust t-statistics are reported in parentheses based on industry-clustered standard errors. (*), (**), (***) denote respectively degree of significance at the 10%, 5% and 1% levels.

Gul, 2012b; Claessens et al., 2008; Fraser et al., 2006; Johnson & Mitton, 2003; Khwaja & Mian, 2005; Leuz & Oberholzer-Gee, 2006). The finding further suggests that the political connections of U.S. firms provide their creditors with additional collateral. This more favorable perception of default risk by these firms allows them, in accordance with the financial theory of capital structure, to access a higher debt level, and thus benefit from a greater leverage effect. In addition, the results in Table 4, panel A, show that the main con-

trol variables used in Model (1) have the expected signs. On the one hand, the debt level significantly increases with the size and tangible assets ratio. These two factors are indeed sources of guarantees for creditors (Bliss & Gul, 2012b). On the other hand, firms with high profitability and current assets ratios are significantly less indebted, suggesting that these firms have a lower need for debt financing. The market-to-book ratio does not significantly affect the debt ratio of the firms in our sample.

Please cite this article in press as: Chkir, I., et al. Political connections and corporate debt: Evidence from two U.S. election campaigns. The Quarterly Review of Economics and Finance (2019), https://doi.org/10.1016/j.qref.2019.05.003

G Model

ARTICLE IN PRESS

QUAECO-1263; No. of Pages 11

I. Chkir et al. / The Quarterly Review of Economics and Finance xxx (2019) xxx–xxx

7

Table 6 Effect of political connections on the cost of debt. Exhibit 1: Without leverage interaction factor

PCN SIZE TAN LEV CUR ROA GRW REA Intercept Sector Effect Year Effect N Adj. R2

Panel A (Full Sample)

Panel B (2009 to 2011)

Panel C (2013 to 2015)

Panel D (2009)

Panel E (2013)

Coefficient

t-stat

Coefficient

t-stat

Coefficient

t-stat

Coefficient

t-stat

Coefficient

t-stat

−0.301 −0.791 −1.517 0.401 −0.001 −2.513 0.003 −0.002 4.554 yes yes 32,622 0.021

−2.350** −2.255** −1.915* 2.196** −0.174 −5.717*** 0.225 −0.358 0.613

−0.245 −0.304 −1.124 0.578 −0.003 −3.612 −0.002 −0.093 5.887 yes yes 17,163 0.018

−2.008** −3.578*** −1.866* 1.987** −0.145 −5.008*** −0.103 −0.676 0.662

−0.312 −0.293 −2.145 0.814 −0.001 −2.106 0.006 −0.035 4.153 yes yes 15,456 0.015

−2.404** −2.907*** −1.017 1.820* −0.084 −4.983*** 0.403 −0.458 1.308

−0.198 −0.378 −2.012 0.886 −0.007 −3.121 −0.001 −0.048 4.229 yes no 5,762 0.010

−1.894* −2.944*** −0.548 2.116** −0.182 −2.106** −0.062 −0.358 1.202

−0.177 −0.355 −2.429 0.618 −0.003 −3.335 −0.008 −0.071 2.818 yes no 6,005 0.010

−1.761* −3.020*** −0.839 2.121** −0.117 −6.114*** −0.708 −0.987 1.171

Exhibit 2: With leverage interaction factor

PCN SIZE TAN LEV LEVxPCN CUR ROA GRW REA Intercept Sector Effect Year Effect N Adj. R2

Panel A (Full Sample)

Panel B (2009 to 2011)

Panel C (2013 to 2015)

Panel D (2009)

Coefficient

t-stat

Coefficient

t-stat

Coefficient

t-stat

Coefficient

t-stat

Coefficient

t-stat

−0.157 −0.830 −1.488 0.425 −0.320 −0.001 −2.487 0.003 −0.001 4.275 yes yes 32,622 0.023

−0.440 −2.264** −1.835* 2.202** −2.147** −0.161 −5.647*** 0.225 −0.216 0.550

−0.209 −0.298 −1.141 0.641 −0.502 −0.005 −3.488 −0.002 −0.088 5.739 yes yes 17,163 0.019

−0.380 −3.490*** −1.904* 1.990** −2.195** −0.164 −4.823*** −0.103 −0.447 0.507

−1.437 −0.151 −2.953 0.861 −0.688 −0.001 −1.855 0.006 −0.025 3.893 yes yes 15,456 0.017

−0.340 −2.746*** −1.080 1.842* −1.926* −0.099 −4.738*** 0.403 −0.253 1.210

−1.558 −0.348 −1.663 0.956 −0.812 −0.009 −2.820 −0.001 −0.037 4.232 yes no 5,762 0.010

−0.389 −2.753*** −0.179 2.161** −2.252** −0.223 −1.846* −0.062 −0.246 1.211

−0.607 −0.313 −2.130 0.669 −0.554 −0.002 −3.237 −0.008 −0.056 2.740 yes no 6,005 0.010

−1.039 −2.960*** −0.445 2.136** −1.907* −0.047 −6.553*** −0.708 −0.964 1.039

Panel E (2013)

This table reports the results of model (2). Dependent variable is the cost of debt (CDT) measured by the ratio of interest expense to the sum of long-term and short-term debt. Control variables are: PCN: Political connection, a binary variable that takes the value of one in years 2009 to 2011 for firms with a PAC during the 2008 election, and in years 2013 to 2015 for those with a PAC during the 2012 election, and 0 otherwise; SIZE: Natural logarithm of total assets; TAN: Ratio of property, plant and equipment to total assets; LEV: Total debt ratio measured by the ratio of long-term and short-term debts to total assets; CUR: Ratio of current assets to current liabilities; ROA: ratio of net income to total assets; GRW: Ratio of market to book value of equity; REA: ratio of retained earnings to total assets. Robust t-statistics are reported based on industry-clustered standard errors. (*), (**), (***) denote respectively degree of significance at the 10%, 5% and 1% levels.

Given that the 2008 and 2012 American election campaigns resulted in a move from a Republican administration under the leadership of George W. Bush to a Democrat administration under the leadership of Barack Obama in 2008, and the reelection of the latter in 2012, we reestimated Model (1) over two subperiods: from 2009 to 2011 and from 2013 to 2015. Dividing the entire sample into two equal periods of 3 years also allows us to exclude the year of each election campaign, to account for the fact that PCFs probably do not adjust their debt ratio in the election year, but that they do so gradually over a few years after. The results of these tests are reported in panels B and C of Table 4 and are statistically similar to those of panel A. In particular, PCFs are significantly more indebted than are PUFs in both subperiods, and this result is significant at the 1% level. In panels D and E of Table 4, we report the results of Model (1) for the first postelection year of the two subsamples (i.e., 2009 and 2013), to avoid the confounding effects of any other events in years 2 and 3 and to test whether the political connection effect on debt is more or less pronounced in year 1. Once more, the main results remain statistically similar to those in panels A, B, and C. The results of the previous multivariate analysis show that the debt level of PCFs is significantly higher than that of PUFs, even after controlling for the other variables. In what follows, we propose testing whether the level and the type of the political connection, estimated by the total amount of each company’s donations, and the party to which the company is connected, affect the debt level of PCFs. To this end, we reestimated two different versions of Model

(1) on the sample of PCFs only over different subperiods. First, we added four new variables to Model (1): CONTDEM: Ratio of donations to the Democratic Party to total assets (Model (1.a)) CONTREP: Ratio of donations to the Republican Party to total assets (Model (1.a)) DEM: Binary variable equal to 1 for firms exclusively donating to the Democratic Party and 0 otherwise (Model (1.b)). REP: Binary variable equal to 1 for firms exclusively donating to the Republican Party and 0 otherwise (Model (1.b)). Second, donations to both parties or exclusive donations to a single party may be affected by which party had control after the elections (Claessens et al., 2008). As such, even though 2009–2015 was under a Democratic presidency, Democrats were in control of both Congress chambers in 2009 and 2010, and Republicans had control of both chambers in 2014 and 2015. In 2011 and 2013, Congress was divided. Thus, we redesigned our subsamples accordingly. Table 5 reports the results of this analysis. First, the table shows that the contribution amounts made by firms to the PACs do not affect the debt level regardless of the party receiving the contributions and no matter which party is in control. Indeed, both CONTDEM and CONTREP are not significant during the three subperiods. This suggests that the level of the political connection, as estimated by the level of the company’s involvement in the financing of political parties, does not affect its debt ratio. In other words,

Please cite this article in press as: Chkir, I., et al. Political connections and corporate debt: Evidence from two U.S. election campaigns. The Quarterly Review of Economics and Finance (2019), https://doi.org/10.1016/j.qref.2019.05.003

G Model

ARTICLE IN PRESS

QUAECO-1263; No. of Pages 11

I. Chkir et al. / The Quarterly Review of Economics and Finance xxx (2019) xxx–xxx

8

Table 7 Effect of political connections on the cost of debt: Altman Z-Score. Exhibit 1: Without leverage interaction factor

PCN LEV Z Intercept Sector Effect Year Effect N Adj. R2

Panel A (Full Sample)

Panel B (2009 to 2011)

Panel C (2013 to 2015)

Panel D (2009)

Panel E (2013)

Coefficient

t-stat

Coefficient

t-stat

Coefficient

t-stat

Coefficient

t-stat

Coefficient

t-stat

−0.188 0.010 −0.015 4.913 yes yes 32,622 0.016

−1.959* 2.109** −1.433 0.480

−0.195 0.013 −0.013 2.992 yes yes 17,163 0.020

−1.917* 2.288** −1.382 0.367

−0.190 0.010 −0.035 3.019 yes yes 15,456 0.037

−1.912* 3.005*** −1.872* 0.711

−0.201 0.022 −0.010 3.908 yes no 5,762 0.025

−1.984* 2.214** −1.000 1.118

−0.095 0.010 −0.038 2.449 yes no 6,005 0.037

−1.532 2.186** −1.994* 0.315

Exhibit 2: With leverage interaction factor

PCN LEV LEVxPCN Z Intercept Sector Effect Year Effect N Adj. R2

Panel A (Full Sample)

Panel B (2009 to 2011)

Panel C (2013 to 2015)

Panel D (2009)

Panel E (2013)

Coefficient

t-stat

Coefficient

t-stat

Coefficient

t-stat

Coefficient

t-stat

Coefficient

t-stat

−0.024 0.011 −0.230 −0.021 6.008 yes yes 32,622 0.017

−1.339 2.125** −2.147** −1.796* 0.550

−0.027 0.011 −0.311 −0.019 3.330 yes yes 17,163 0.022

−1.000 2.017** −2.195** −1.804* 0.537

−0.021 0.012 −0.472 −0.032 3.646 yes yes 15,456 0.041

−1.151 3.443*** −1.426 −1.799* 0.875

−0.050 0.020 −0.688 −0.012 4.232 yes no 5,762 0.027

−0.634 2.091** −2.252** −1.054 1.211

−0.021 0.011 −0.317 −0.030 3.054 yes no 6,005 0.040

−1.048 2.312** −1.907* −1.959* 1.133

This table reports the results of the regression of cost of debt (CDT) measured by the ratio of interest expense to the sum of long-term and short-term debt on the following variables: PCN: Political connection is a binary variable that takes the value of one in years 2009 to 2011 for firms with a PAC during the 2008 election, and in years 2013 to 2015 for those with a PAC during the 2012 election, and 0 otherwise; LEV: Total debt ratio measured by the ratio of long-term and short-term debts to total assets; Z: The Altman Z-Score. Robust t-statistics are reported based on industry-clustered standard errors. (*), (**), (***) denote respectively degree of significance at the 10%, 5% and 1% levels.

that the company is politically connected seems to give it an advantage over PUFs in terms of financial leverage. This advantage does not seem to be proportionally related to the company’s level of political connection. Second, the results in Table 5 show that being exclusively connected to one party gives firms an additional advantage in terms of financial leverage when that party is in control of Congress. In fact, both DEM and REP are positive and significant at the 1% level during the period in which Democrats and Republicans, respectively, were in control of the chambers. However, both these variables are not significant during the period in which Congress was divided. However, this result should be taken with caution. Indeed, and as previously discussed, most PCFs in our sample donated to both the Democratic Party and the Republican Party. It is worth noting that the results on the control variables in the three panels of Model (1) are qualitatively similar to those found in the original version of the model.

4.2.2. Cost of debt and political connection Table 6 reports the results of the regression of cost of debt on the binary variable representing PCFs as well as the different control variables (Model (2)). The results of Table 6, Exhibit 1, show that after controlling for the main factors determining the cost of debt, PCFs still face a lower cost of debt than do PUFs. In fact, PCN is significant at the 5% level during the entire period of 2009 to 2015 (panel A) and the two subperiods of 2009 to 2011 (panel B) and 2013 to 2015 (panel C) and at the 10% level during the years 2009 and 2013 (panels D and E). These results corroborate the preliminary findings in Tables 2 and 3. They are also in line with those of Houston et al. (2014), with regard to the cost of bank loans for U.S. PCFs. In fact, the authors found that U.S. PCFs, compared with PUFs, benefit from lower bank interest rates. Therefore, these results support H2, suggesting that the political connections of U.S. firms provide them

with privileges, such as receiving debt financing at a lower cost than that charged to PUFs. On the other hand, these results are inconsistent with those of Bliss and Gul (2012a) in the Malaysian context, who found that PCFs pay a higher interest rate on debt than do PUFs. Because the cost of debt increases with firms’ level of indebtedness, because of increased default risk, and because the results of Model (1) show that PCFs have a higher debt level than do PUFs, one would expect PCFs to incur a higher average cost of debt than do PUFs (H2’). Exhibit 1 of Table 6 and our preliminary results in Tables 2 and 3 seem to reject this hypothesis. The results in Table 6 also show that the main factors that affect cost of debt are size and profitability ratio, both of which have a negative effect, and debt ratio, which has a positive effect. Exhibit 2 of Table 6 adds the variable LEVxPCN to our Model (2) to test for the interaction effect between leverage and political connections on the cost of debt and help understand these a priori conflicting results on the effect of political connections on the cost of debt. The results in Exhibit 2 show that LEV is still—as is expected—positive and statistically significant in all panels, confirming that the cost of debt increases with leverage. PCN is no longer significant; however, its interaction with leverage is negative and significant in all panels. Moreover, the sum of the coefficients of LEV and LEVxPCN is insignificantly different from zero. In a nutshell, our findings suggest that PCFs benefit from a higher indebtedness level than do PUFs, and their political connection gives them an advantage that offsets the negative effect of high leverage on their cost of debt. Relatedly, Chaney et al. (2011) find that lower accruals quality results in a higher cost of debt but that the interaction factor of accruals’ quality with political connection completely offsets this effect. As for our preliminary results, PCFs being larger and more profitable than PUFs explains why the univariate analysis showed that, on average, they have a significantly lower cost of debt.

Please cite this article in press as: Chkir, I., et al. Political connections and corporate debt: Evidence from two U.S. election campaigns. The Quarterly Review of Economics and Finance (2019), https://doi.org/10.1016/j.qref.2019.05.003

G Model

ARTICLE IN PRESS

QUAECO-1263; No. of Pages 11

I. Chkir et al. / The Quarterly Review of Economics and Finance xxx (2019) xxx–xxx

9

Table 8 Effect of political connections on debt level: Instrumental variable estimation. (Full Sample)

Panel A: First-stage results DIS SIZE TAN CDT CUR ROA GRW REA Intercept Sector Effect Year Effect N Adj. R2 Wu-Hausman F-statistics F-statistics

(2009 to 2011)

(2013 to 2015)

(2009)

Coefficient

t-stat

Coefficient

t-stat

Coefficient

t-stat

Coefficient

t-stat

Coefficient

t-stat

−0.006 0.033 −0.028 −0.009 0.005 0.008 0.001 0.015 −3.766 yes yes 32,622 0.139 18.321*** 398.715***

−5.818*** 7.901*** −6.587*** −0.255 0.268 3.881*** 0.727 1.565 −3.128***

−0.006 0.037 −0.040 −0.011 −0.003 0.009 0.001 0.019 5.408 yes yes 17,163 0.125 16.955*** 277.520***

−3.541*** 7.812*** −6.568*** −0.165 −0.274 4.717*** 0.451 1.909* 1.807

−0.007 0.033 −0.022 −0.008 0.007 0.008 0.002 0.009 5.261 yes yes 15,456 0.112 16.019*** 208.843***

−4.318*** 8.761*** −3.272*** −0.157 0.129 3.348*** 0.777 1.021 1.124

−0.006 0.038 −0.027 −0.012 0.006 0.007 −0.001 0.014 −1.309 yes no 5,762 0.144 12.983*** 129.123***

−2.350** 18.969*** −4.157*** −0.216 0.118 2.928*** −0.195 1.508 −2.995***

−0.006 0.030 −0.022 −0.008 0.005 0.013 −0.001 0.013 −0.574 yes no 6,005 0.141 11.847*** 109.993***

−2.757** 7.522*** −2.411** −0.198 0.233 3.010*** −0.204 1.488 −2.521**

4.070*** 4.783*** 7.013*** −1.427 −3.008*** −9.157*** −0.427 −1.179 −4.255***

0.050 0.023 0.152 −0.025 −0.044 −0.278 −0.002 −0.023 −5.012 yes yes 17,163 0.142

3.754*** 5.118*** 3.820*** −1.528 −5.896*** −13.519*** −1.554 −1.691 −3.355***

0.155 0.022 0.182 −0.020 −0.039 −0.094 0.001 −0.056 −7.324 yes yes 15,456 0.155

2.988*** 2.344** 5.861*** −0.939 −5.023*** −7.441*** 0.090 −1.899* −3.814***

0.101 0.018 0.150 −0.009 −0.020 −0.098 0.001 −0.017 −3.844 yes no 5,762 0.125

2.194** 2.289** 2.110** −0.883 −4.707*** −7.904*** 0.435 −1.004 −3.128***

0.091 0.021 0.088 −0.008 −0.019 −0.087 −0.001 −0.038 −3.891 yes no 6,005 0.124

2.207** 4.353*** 1.704 −0.059 −4.136*** −5.915*** −1.209 −2.221** −3.997***

Panel B: Second-stage results Predicted PCN 0.109 0.020 SIZE 0.160 TAN CDT −0.010 −0.022 CUR −0.201 ROA −0.001 GRW −0.012 REA Intercept −9.573 yes Sector Effect Year Effect yes 32,622 N 0.139 Adj. R2

(2013)

This table presents the two-stage estimation results for the debt level. In the first stage, we regress the political connection binary variable PCN (which takes the value of one in years 2009 to 2011 for firms with a PAC during the 2008 election, and in years 2013 to 2015 for those with a PAC during the 2012 election, and 0 otherwise) on the instrumental variable; DIS (the natural logarithm of one plus the distance from a firm’s headquarter to Washington D.C. in km). In the second stage, we regress leverage (LEV, measured by the ratio of long-term and short-term debts to total assets, model 1) on the predicted PCN from the first stage and the following control variables: SIZE: Natural logarithm of total assets; TAN: Ratio of property, plant and equipment to total assets; CDT: Cost of debt measured by the ratio of interest expense to the sum of long-term and short-term debt; CUR: Ratio of current assets to current liabilities; ROA: ratio of net income to total assets; GRW: Ratio of market to book value of equity; REA: Ratio of retained earnings to total assets. Robust t-statistics are reported in parentheses based on industry-clustered standard errors. (*), (**), (***) denote respectively degree of significance at the 10%, 5% and 1% levels.

4.3. Robustness tests In this section, we perform some additional tests to check the robustness of our results. First, we rerun Model (2) by replacing the control variables with the Altman Z-score, which is defined as Z = 1.2 (working capital/total assets) + 1.4 (retained earnings/total assets) + 3.3 (earnings before interest and taxes/total assets) + 0.6 (market value of equity/book value of total liabilities) + 1.0 (sales/total assets). With this alternative control variable, we obtain the same results (see Table 7) for the effect of leverage and the interaction between leverage and political connection on the cost of debt. Second, we check whether our results suffer from any endogeneity bias. In fact, one might suspect firms with financial difficulties, due to high leverage, to be opting for lobbying through creating PACs as a way to obtain guarantees. Therefore, we repeat our analysis in Models (1) and (2) using a two-stage instrumental variables estimation method, to address any potential reverse causality between political connection, debt level, and cost. In the first stage, and following Houston et al. (2014), we regress the political connection binary variable PCN on the instrumental variable DIS, which measures the distance from the firm’s headquarters to Washington DC, as well as the control variables. In the second stage, we regress leverage (Table 8) and cost of debt (Table 9) on the predicted PCN from the first stage and the control variables. Various tests suggest that our selected instrumental variable is valid. The Wu-Hausman test rejects the null hypothesis that the political connection variable PCN is exogenous, and the high F-statistic indicates that our

estimation results do not suffer from the problem of weak instrument. Thus, our selected instrumental variable satisfies the relevant conditions of being a valid instrument. The results of the instrumental variables estimation method are qualitatively similar to those reported in Tables 4 and 6. They confirm that PCFs, compared with PUFs, have a significantly higher debt level without enduring the higher cost of debt that PUFs do. 5. Conclusion In this paper, we have examined the effect of political connections of U.S. firms on the level and the cost of their debt. We have focused on two major political events in the United States, namely, the 2008 and 2012 election campaigns, to identify PCFs through their funding of the two main U.S. political parties: the Democratic Party and the Republican Party. Our results show that during the 2009 to 2015 period, and after controlling for the main factors suggested by the financial literature on capital structure, PCFs were significantly more indebted than were PUFs, whereas the cost of debt was not significantly different between the two groups of firms. The results regarding the debt level are consistent with those found in earlier studies of emerging markets (Belghitar et al., 2016; Bliss & Gul, 2012b; Claessens et al., 2008; Fraser et al., 2006; Johnson & Mitton, 2003; Khwaja & Mian, 2005; Leuz & Oberholzer-Gee, 2006). The findings indicate that the political connections of U.S. firms provide their creditors with additional collateral. By reducing the perception of default risk of PCFs, political connections grant

Please cite this article in press as: Chkir, I., et al. Political connections and corporate debt: Evidence from two U.S. election campaigns. The Quarterly Review of Economics and Finance (2019), https://doi.org/10.1016/j.qref.2019.05.003

G Model

ARTICLE IN PRESS

QUAECO-1263; No. of Pages 11

I. Chkir et al. / The Quarterly Review of Economics and Finance xxx (2019) xxx–xxx

10

Table 9 Effect of political connections on the cost of debt: Instrumental variable estimation. (Full Sample)

Panel A: First-stage results DIS SIZE TAN LEV CUR ROA GRW REA Intercept Sector Effect Year Effect N Adj. R2 Wu-Hausman F-statistics F-statistics

(2009 to 2011)

(2013 to 2015)

(2009)

Coefficient

t-stat

Coefficient

t-stat

Coefficient

t-stat

Coefficient

t-stat

Coefficient

t-stat

−0.007 0.036 −0.033 0.006 0.001 0.007 0.001 −0.001 −4.463 yes yes 32,622 0.130 19.146*** 487.227***

−5.993*** 6.742*** −2.003** 1.003 0.209 2.321** 0.805 −3.459*** −3.901***

−0.006 0.039 −0.042 0.008 −0.002 0.010 −0.001 −0.002 6.821 yes yes 17,163 0.139 15.228*** 276.236***

−3.708*** 9.752*** −6.018*** 1.433 −0.381 5.670*** −0.352 −6.483 1.582

−0.007 0.033 −0.024 0.005 0.001 0.006 0.001 −0.01 4.822 yes yes 15,456 0.122 13.813*** 215.092***

−4.754*** 4.357*** −3.409*** 1.317 0.105 2.294** 0.130 −5.167*** 1.031

−0.006 0.041 −0.037 0.005 0.001 0.006 0.001 −0.020 −0.105 yes no 5,762 0.142 10.9918*** 105.453***

−2.064** 8.940*** −3.039*** 0.453 0.628 1.513 1.106 −4.108 −4.698

−0.007 0.032 −0.022 0.004 −0.001 0.005 −0.001 −0.010 −0.056 yes no 6,005 0.120 10.134*** 102.471***

−2.788*** 7.331*** −2.3156** 0.730 −0.088 2.673*** −0.266 −2.573** −2.913***

−0.298 −0.243 −0.433 0.015 −0.004 −0.312 −0.002 0.002 3.357 yes yes 17,163 0.015

−2.031** −3.717*** −3.404*** 1.881* −0.113 −3.248*** −0.119 0.308 0.398

−0.288 −0.085 −0.011 0.011 0.001 −0.010 0.001 −0.001 2.003 yes yes 15,456 0.028

−1.908* −2.382** −0.057 3.527*** 0.098 −5.115*** 0.987 −0.893 1.138

−0.611 0.236 −0.060 0.010 0.001 −0.101 0.001 −0.001 −0.109 yes no 5,762 0.031

−1.820* 3.716*** −1.477 1.898* 1.082 −3.300*** 0.028 −0.094 −0.205

−0.441 −0.071 −0.018 0.010 −0.001 −0.021 0.001 −0.001 0.009 yes no 6,005 0.035

−1.819* −2.281** −1.525 1.897* −1.008 −4.008*** 0.198 −0.735 0.045

−0.242 −0.260 −0.510 0.018 −0.422 −0.005 −0.376 −0.003 0.003 4.165 yes yes 17,163 0.015

0.416 −3.965*** −3.602*** 1.899* −2.199** −0.176 −4.175*** −0.132 0.331 0.416

−0.264 −0.093 −0.010 0.013 −0.688 0.001 −0.011 0.001 −0.001 2.873 yes yes 15,456 0.032

−1.436 −2.438** −0.050 3.749*** −1.905* 0.123 −6.316*** 1.004 −1.085 1.582

−0.938 0.236 −0.058 0.011 −0.698 0.001 −0.112 0.001 −0.001 −0.123 yes no 5,762 0.040

−0.266 3.716*** −1.396 1.903* −2.172** 1.152 −3.341*** 0.037 −0.100 −0.266

−0.370 −0.088 −0.020 0.010 −0.600 −0.001 −0.026 0.001 −0.001 0.023 yes no 6,005 0.049

−1.165 −2.398** −1.752* 1.890* −1.912* −1.363 −4.334*** 0.248 −0.855 0.664

Panel B: Second-stage results Exhibit 1: Without leverage interaction factor Predicted PCN −0.331 −2.205** SIZE −0.158 −3.018*** TAN −0.013 −1.907* LEV 0.020 3.118*** CUR −0.001 −0.255 ROA −0.101 −5.028*** GRW 0.001 0.400 REA −0.001 −0.715 6.125 1.455 Intercept Sector Effect yes yes Year Effect 32,622 N 2 0.015 Adj. R Exhibit 2: With leverage interaction factor Predicted PCN −0.290 −0.609 −0.162 −3.269*** SIZE TAN −0.015 −1.923* 0.022 3.271*** LEV LEVxPredicted PCN −0.219 −2.208** CUR −0.001 −0.271 ROA −0.115 −6.944*** GRW 0.002 0.436 REA −0.001 −0.863 Intercept 7.829 1.890* Sector Effect yes Year Effect yes N 32,622 2 Adj. R 0.016

(2013)

This table presents the two-stage estimation results for the cost of debt. In the first stage, we regress the political connection binary variable PCN (which takes the value of one in years 2009 to 2011 for firms with a PAC during the 2008 election, and in years 2013 to 2015 for those with a PAC during the 2012 election, and 0 otherwise) on the instrumental variable; DIS (the natural logarithm of one plus the distance from a firm’s headquarter to Washington D.C. in km). In the second stage, we regress cost of debt (CDT) model (2) on the predicted PCN from the first stage and the following control variables: SIZE: Natural logarithm of total assets; TAN: Ratio of property, plant and equipment to total assets; LEV: Total debt ratio measured by the ratio of long-term and short-term debts to total assets; CUR: Ratio of current assets to current liabilities; ROA: ratio of net income to total assets; GRW: Ratio of market to book value of equity; REA: ratio of retained earnings to total assets. Robust t-statistics are reported in parentheses based on industry-clustered standard errors. (*), (**), (***) denote respectively degree of significance at the 10%, 5% and 1% levels.

these firms access to a higher debt level, thus giving them higher leverage. Houston et al. (2014) find that U.S. PCFs pay lower interest rates than PUFs. Our results for the cost of debt are in line with those findings. However, our findings differ from those of Bliss and Gul (2012a), who find that PCFs in a Malaysian context pay a higher interest rate on debt than do PUFs. Our findings suggest that PCFs benefit from a higher indebtedness level than do PUFs and that their political connections give them an advantage that offsets the negative effect of high leverage on their cost of debt. Our results do not suffer from any endogeneity. They are robust when the analysis is conducted separately over different subperiods and when alternative measures of the cost of debt and the control variables are used. The results of this study have implications for both corporate executives and portfolio managers. At the firm level, donations to

political parties can benefit firms in ways other than tax exemptions for such donations. These additional benefits do not depend on the amount of the donations. From a portfolio management perspective, shareholders should be interested in accounting for political connections, while forming their portfolios and assessing the risk of the firms they are interested in, because PCFs seem to convey a lower risk signal than do PUFs, all other things being equal. Although acceptable for statistical and econometric analysis, the number of PCFs in our study remains small compared with the total sample of U.S. firms. To mend this, extending this study with a different measure of political connection, such as affiliations and/or family and other relationships, between senior company executives and influential members of the government could be envisioned. In addition to the possibility of increasing the number of PCFs in the sample, this type of study could possibly test the effect of

Please cite this article in press as: Chkir, I., et al. Political connections and corporate debt: Evidence from two U.S. election campaigns. The Quarterly Review of Economics and Finance (2019), https://doi.org/10.1016/j.qref.2019.05.003

G Model QUAECO-1263; No. of Pages 11

ARTICLE IN PRESS I. Chkir et al. / The Quarterly Review of Economics and Finance xxx (2019) xxx–xxx

changing the status of a PCF to a PUF, for example, following a government reshuffling. This kind of analysis was not possible in our study because most PCFs in our sample financed the two election campaigns and thus kept the status of PCF throughout the study period. References Beck, T., Demirgüc¸-Kunt, A., & Levine, R. (2006). Bank supervision and corruption in lending. Journal of Monetary Economics, 53(8), 2131–2163. Belghitar, Y., Clark, E., & Saeed, A. (2016). Political connections and corporate financial decision making. Working paper SSRN Available at:https://ssrn.com/ abstract=2821123[Last accessed: 10 June 2018]. Benjamin, S., Zain, M., & Abdul Wahab, E. (2015). Political connections and dividends payout in Malaysia: The influence of institutional investors. Working paper SSRN Available at:https://ssrn.com/abstract=2557944[Last accessed: 10 June 2018]. Bertrand, M., Kramarz, F., Schoar, A., & Thesmar, D. (2006). Politicians, firms and the political business cycle: Evidence from France. Unpublished working paper. University of Chicago. Bliss, M., & Gul, F. (2012a). Political connection and cost of debt: Some Malaysian evidence. Journal of Banking & Finance, 36(5), 1520–1527. Bliss, M., & Gul, F. (2012b). Political connection and leverage: Some Malaysian evidence. Journal of Banking & Finance, 36(8), 2344–2350. Borisova, G., Fotak, V., Holland, K., & Megginson, W. L. (2015). Government ownership and the cost of debt: Evidence from government investments in publicly traded firms. Journal of Financial Economics, 118(1), 168–191. Boubakri, N., Cosset, J. C., & Saffar, W. (2008). Political connections of newly privatized firms. Journal of Corporate Finance, 14(5), 654–673. Boubakri, N., El Ghoul, S., & Saffar, W. (2013). Cash holdings of politically connected firms. Journal of Multinational Financial Management, 23(4), 338–355. Boubakri, N., Cosset, J.-C., & Saffar, W. (2012). The impact of political connections on firm’s operating performance and financing decisions. Journal of Financial Research, 35(3), 397–423. Boubakri, N., Guedhami, O., Mishra, D., & Saffar, W. (2012). Political connections and the cost of equity capital. Journal of Corporate Finance, 18(3), 541–559. Brockman, P., Rui, O. M., & Zou, H. (2013). Institutions and the performance of politically connected M&As. Journal of International Business Studies, 44(8), 833–852. Bunkanwanicha, P., & Wiwattanakantang, Y. (2009). Big business owners in politics. Review of Financial Studies, 22(9), 2133–2168. Chan, K. S., Dang, V. Q., & Yan, I. K. (2012). Chinese firms’ political connection, ownership, and financing constraints. Economics Letters, 115(2), 164–167. Chaney, P., Faccio, M., & Parsley, D. (2011). The quality of accounting information in politically connected firms. Journal of Accounting and Economics, 51(1), 58–76. Chen, H., Parsley, D., & Yang, Y. (2014). Corporate lobbying and firm performance. Journal of Business, Finance, and Accounting, 42(3-4), 444–481. 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. Cooper, M., Gulen, H., & Ovtchinnikov, A. (2010). Corporate political contributions and stock returns. The Journal of Finance, 65(2), 687–724. Cull, R., Li, W., Sun, B., & Xu, L. C. (2015). Government connections and financial constraints: Evidence from a large representative sample of Chinese firms. Journal of Corporate Finance, 32, 271–294. Dinc¸, S. (2005). Politicians and banks: Political influences on government-owned banks in emerging markets. Journal of Financial Economics, 77, 453–479. Ding, S., Jia, C., Wu, Z., & Zhang, X. (2014). Executive political connections and firms performance: Comparative evidence from privately-controlled and state-owned enterprises. International Review of Financial Analysis, 36, 153–167. Faccio, M. (2006). Politically connected firms. American Economic Review, 96(1), 369–386. Faccio, M. (2010). Differences between politically connected firms and non-connected firms: A cross-country analysis. Financial Management, 39, 905–927.

11

Faccio, M., Masulis, R., & McConnell, J. (2006). Political connections and corporate bailouts. The Journal of Finance, 61(6), 2597–2635. Fan, J. P., Wong, T. J., & Zhang, T. (2007). Politically connected CEOs, corporate governance, and Post-IPO performance of China’s newly partially privatized firms. Journal of Financial Economics, 84(2), 330–357. Fan, J. P., Rui, O. M., & Zhao, M. (2008). Public governance and corporate finance: Evidence from corruption cases. Journal of Comparative Economics, 36(3), 343–364. Fisman, R. (2001). Estimating the value of political connections. American Economic Review, 91(4), 1095–1102. Fraser, D., Zhang, H., & Derashid, C. (2006). Capital structure and political patronage: The case of Malaysia. Journal of Banking and Finance, 30(4), 1291–1308. Goldman, E., Rocholl, J., & So, J. (2009). Do politically connected boards affect firm value? Review of Financial Studies, 22(6), 2331–2360. Harris, M., & Raviv, A. (1991). The theory of capital structure. The Journal of Finance, 46(1), 297–335. Hillman, A. (2005). Politicians on the board of directors: Do connections affect the bottom line? Journal of Management, 31(3), 464–481. Houston, J., Lin, C., & Ma, Y. (2011). Media ownership, concentration and corruption in bank lending. Journal of Financial Economics, 100, 326–350. Houston, J., Jiang, L., Lin, C., & Ma, Y. (2014). Political connections and the cost of bank loans. Journal of Accounting Research, 52(1), 193–240. Hutton, I., Jiang, D., & Kumar, A. (2014). Corporate policies of republican managers. Journal of Financial and Quantitative Analysis, 49(5–6), 1279–1310. Jensen, M. (1986). Agency costs of free cash flow, corporate finance, and takeovers. The American Economic Review, 76(2), 323–329. Jensen, M., & Meckling, W. (1976). Theory of the firm: Managerial behavior, agency costs and ownership structure. Journal of Financial Economics, 3(4), 305–360. Johnson, S., & Mitton, T. (2003). Cronyism and capital controls: Evidence from Malaysia. Journal of Financial Economics, 67(2), 351–382. Khwaja, A., & Mian, A. (2005). Do lenders favor politically connected firms? Rentprovision in an emerging financial market. Quarterly Journal of Economics, 120(4), 1391–1411. La Porta, R., Lopez-de-Silanes, F., & Shleifer, A. (2002). Government ownership of banks. The Journal of Finance, 57(1), 265–301. Leuz, C., & Oberholzer-Gee, F. (2006). Political relationships, global financing, and corporate transparency: Evidence from Indonesia. Journal of Financial Economics, 81(2), 411–439. Li, H., Meng, L., & Zhang, J. (2006). Why do entrepreneurs enter politics? Evidence from China. Economic Inquiry, 44(3), 559–578. Ling, L., Zhou, X., Liang, Q., Song, P., & Zeng, H. (2016). Political connections, overinvestments and firm performance: Evidence from Chinese listed real estate firms. Finance Research Letters, 18, 328–333. Pfeffer, J., & Salancik, G. (1978). The external control of organizations: A resource dependence perspective. New York: Harper & Row Publisher. Santa-Clara, P., & Valkanov, R. (2003). The presidential puzzle: Political cycles and the stock market. The Journal of Finance, 58(5), 1841–1872. Shen, C.-H., & Lin, C.-Y. (2015). Betting on presidential elections: Should we buy stocks connected with the winning party? The Quarterly Review of Economics and Finance, 56, 98–109. Shleifer, A., & Vishny, R. (1994). Politicians and firms. Quarterly Journal of Economics, 109(4), 995–1025. Snyder, J. (1990). Campaign contributions as investments: The U.S. house of representatives, 1980–1986. Journal of Political Economy, 98(6), 1195–1227. Stratmann, T. (1998). The market for congressional votes: Is the timing of contributions everything? Journal of Law and Economics, 41(1), 85–114. Tahoun, A. (2014). The role of stock ownership by U.S. members of congress on the market for political favors. Journal of Financial Economics, 111(1), 86–110. Tahoun, A., & van Lent, L. (2019). The personal wealth interests of politicians and government intervention in the economy. Review of Finance, 23(1), 37–74. Wang, L. (2015). Protection or expropriation: Politically connected independent directors in China. Journal of Banking & Finance, 55, 92–106. Zhou, W. (2013). Political connections and entrepreneurial investment: Evidence from China’s transition economy. Journal of Business Venturing, 28(2), 299–315.

Please cite this article in press as: Chkir, I., et al. Political connections and corporate debt: Evidence from two U.S. election campaigns. The Quarterly Review of Economics and Finance (2019), https://doi.org/10.1016/j.qref.2019.05.003