MASTER THESIS
___________________________________________________________________________
Political contribution: always an advantage?
A quantitative research into the effect of corporate political activity on
corporate financial performance, taking the role of political preferences in the
States of the U.S. into account
Student Name: Michael Bos
Student Number: 11843926
Date of Submission: 21
thof June 2018
Thesis MSc. Business Administration - International Management Track
University of Amsterdam – Faculty of Economics and Business
2
Statement of Originality
This document is written by Michael Bos who declares to take full responsibility for the contents of this document. I declare that the text and the work presented in this document are original and that no sources other than those mentioned in the text and its references have been used in creating it. The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
3
Abstract
Due to the fact that an increasing amount of empirical researchers state the value of using political non-market strategies to bring forth beneficial policy outcomes and financial performances, the interest in the topic of Corporate Political Activity, CPA, increased largely over the years. Despite the rise of political investments and all conducted studies among the topic of this relationship, a clear answer to the relationship being either positive or negative has not been found yet. In the United States of America, firms engaging in CPA is more the rule than the exception. With the ongoing political discussion and disunity within this country, it is interesting to examine the role of firms and the role of political parties in the U.S.. As each state differs in political preference, this study is going to analyze not only the influence CPA has on the Corporate Financial Performance, CFP, in general, but also the role the Headquarter location of a firm has in terms of the political preferred party in the located State in the U.S.. To test the formulated research questions and hypotheses, a data set of Fortune 500 firms is used to run multiple linear regression analyses. Although the expectation is that financial CPA’s have a positive relation with CFP, the analysis shows no significant results, just a weak positive one. The political State diversity on the other hand shows significant results at some points, as firms located in Democrat States obtain more financial performance than firms located in Republican States when making CEO contributions. This research contributes both to theory and practice, as the new findings on financial CPA’s and political State diversity adds more to the existing literature, and managers can use these findings when determining their political non-market strategy.
Keywords: Corporate Political Activity (CPA), Corporate Financial Performance (CFP), CEO contributions, PAC contributions, Political State diversity in the United States (U.S.), Republican Party, Democrat Party.
4
Table of contents
Introduction………... 5
Literature review………... 9
Corporate Political Activity – Corporate Financial Performance………….. 10
Political Connectedness………. 10
CEO Political Behaviour………... 12
PAC Contributions………. 12
Theoretical framework………. 13
The effect of CPA on CFP………. 14
The moderating effect of State Diversity on the Main Relationship………. 15
Conceptual Model……….. 16
Methodology……….. 17
Sample………... 17 Data collection………... 18 Variables……… 18 Method………... 19Results……… 21
Descriptive Statistics………. 21 Correlations……… 22 VIF Table………... 25 Regression Analyses……….. 25Discussion and Conclusion………... 31
Limitations………. 32
Further Research……… 33
5
Introduction
In 1999, Hillman et al. already noted: “Even the best competitive strategies accompanied by superior products and unique firm resources will not survive without attention to the government” (Carroll & Hall, 1987). Ever since, an increasing amount of empirical researchers encourage the view that engagement in expending resources as for instance lobbying and political contributions to persuade government bodies in making decisions favorable to the contributing firm, is a valuable strategy firms use to bring forth beneficial outcomes (Graziano, 2001; Kaiser, 2010; Shaffer, 1995). The effect these so called corporate political activities, CPAs, have on firm performance is since the significant increase of investments in CPA an even more interesting issue to investigate than before (Bonica, 2016). This increase derives from the positive look of investors towards CPA, which comes forward by politicians willing to offer revenue growth opportunities to firms (Hillman, Keim, & Schuler, 2004). Next to this, changes in policies which could affect a firm’s attractiveness in terms of joint ventures or acquisitions and the influence one has over regulations also prosper investments (Baron, 1995 & Oliver & Holzinger, 2008).
In the United States, federal regulations cost firms over $2.000 trillion in 2012, making firms really attend on above mentioned activities as it can saddle firms with a substantial amount of costs (Crain & Crain, 2014). Consequently, taken the large governmental influence into account, it is no shock firms invest resources to influence government bodies into favorable decision making (Baysinger, 1984). Large firms as Boeing and Lockheed Martin both invested a great amount of money for example. For every $1 spent in lobbying, the aircraft manufacturer secured more than $7000,- in tax breaks (Hallman, 2014) and Lockheed Martin generated around $90 billion in government contracts, just by spending $55 million at lobbying expenditures (Miller, 2006). These financial returns transcend average returns seen nowadays. Therefore, it is of no surprise that firms acting in lobbying and contributing to political entities is omnipresent around the world, which releases major attention from researchers and media (Choi, Jia & Lu, 2014; Hillman, Keim & Schuler, 2004).
The view of the broader society towards CPA is on the other hand generally negative, as the majority thinks firms should not associate themselves with politics (Smith, 2000). In the 2010 Minnesota gubernatorial election for example, Target Corporation saw its brands heavily damaged after donating to a group opposite of their target audience (Torres-Spelliscy, 2016). The societal constraint deriving from this leads to investors making contributions preferably and if possible nameless.
6
Nevertheless, scholars in CPA reason that CPA engagement provides firms both with directly visible benefits as the security of government contracts, as with benefits not directly shown like regulatory borders and tax minimization (Blumentritt, 2003; Hart, 2001; Chen et al., 2010; Hillman et al., 1999). Obviously, both ensues firms’ financial performances. Performance measures taken by empirical studies are conceptualized with use of accounting-based measurements, where politicians prefer tactics beneficial for firms but rather undetectable (Kim, 2008; Lux et al., 2011).
History however, has shown situations in which politicians openly express policy changes, especially when competition rises to its highest point. The higher the rivalry between certain political candidates or parties, the more frequently suppliers are willing to provide policy favors to firms in return for support to their campaign, as they want to maximize their chances of winning the election (Baron, 2001). In 2002 for example, G.W. Bush, the 43th president of the United States of America, had a large voice in the run for House. Steel producers within the country were looking for tariffs on steel imports and increased their lobbying substantially. As the race between the Republican candidates and Democrat candidates for the House of Representatives was extremely tight in some states with steel industries, the bargaining power of those firms was very strong. As expected, some months before the election Bush enforced a 30 percent tariff (The Politics of Steel, 2002). Therefore, it could be said with certainty that the performance of the nonmarket strategy of a firm is positively related to rivalry in politics (Bonardi, Holburn, Vanden Bergh, 2006).
Anterior research paid attention to several aspects of CPA. The political marketplace is conceptualized by Bonardi et al. (2006). In their paper they discuss the most attractive circumstances of a political market in which a particular firm can best engage in. The government influence and dependence of firms towards these governments is studied by Shaffer (1995). Holburn and Vanden Bergh (2002) took a side path and focused on regulated firms and their political strategy and Baysinger (1984) looked in general towards firm’s objectives at the moment they make transactions with government bodies and political actors. Despite all conducted studies among the topic of firms acting in the political nonmarket place, research did not pay much attention to the level of engagement into financial CPAs such as i.e. PAC contributions and CEO contributions, and the effect it has on corporate financial performance. Therefore, the following question is going to be answered in this research:
What is the effect of the Corporate Political Activity of a firm in terms of political related contributions on its Corporate Financial Performance?
7
Hillman & Hitt (1999) provided three types of political strategies in their research. One of them is the financial strategy, in that case firms perform activities in which they are making contributions to political actors to gain benefits. Past researchers state undoubtedly that there is a relationship between CPA and Corporate Financial Performance, CFP. However, none has found a clear answer to the direction of the relationship between CPA and CFP, being either positive or negative (Hillman et al., 2004). With the abovementioned research question this study breathes new life into this topic, examining the relationship between a financial political strategy and the financial performances of a firm.
Subsequently, as shortly mentioned above, the United States of America is the precursor of smart political investments for own benefit. Therefore, it is surprising that no past research has taken the United States into account while discussing this field of research. Especially with the ongoing political discussion and disunity in the country, the influence of firms engaging in political activities is interesting to investigate (Foran, 2017). Within the United States of America, each state differs in which political party is preferred and therefore supported. No study has examined if the amount contributed to a particular association of the political party in favor in that state significantly affects the firm’s corporate financial
performance compared to the effect contributions have on a firm’s financial performance at first. Consequently, this study’s second research question is formulated as:
How does the political preference in the State of the United States where a firm’s Headquarter is located moderates the relationship between the Corporate Political
Activity and its Corporate Financial Performance?
Crain & Crain (2014) pointed out the enormous costs firms in the United States pay as a result of federal regulations. Added to this comes the huge rivalry between the two political parties in the country (Bonardi, Holburn, Vanden Bergh, 2006). The largest reason for these facts can be explained by the difference in opinion, thinking and political view (Hutton, Jiang & Kumar, 2014). Two variating parties, with two more or less differing views towards firms engaging in political activities. These views influence policy making, also in the 50 States of the United States. Taking both the opinion of the people and the differing policies in each state into account, there is a good reason to expect diverse results out of an analysis.
The hypotheses formed in the theoretical framework later on, are tested by running a multiple regression analyses. In this regression analyses a sample from two databases is used. Both databases consist information of Fortune 500 firms from 2012 till 2015. The data in these
8
sets show amounts of contributions firms made to associations of either the Democrat or the Republican party, the CEO’s made to the parties and other expenditures concerning lobbying activities. Furthermore, financial data as sales records, income states and other resources are stated over these four years as well. Based on the location of each firm’s Headquarter, a list of States which are present from these firms comes forward.
By investigating these relationships, this paper will contribute to both theory and practice. Theoretically this will add broad insights on the matter both vertically and horizontally, as these results will add on current literature on the effect CPA has on corporate financial performance and new insights are provided on the very discussed topic of political preferences in the States of the United States. From a practical point of view, U.S. firms can base their nonmarket strategies on the outcome of this paper, just by looking at the state their Headquarter is located in, together with the preferred political party in that particular state. Based on this paper’s results, a firm can or cannot decide to contribute to either the Republican or Democrat party, depending on their preference.
Former management scholars, experienced economists and political scientists highlighted great heterogeneity both on firm and industry level as political activity predictors. Related to that, industry concentration, firm size and the extent of regulations within a certain industry are for instance factors to take into account (Grier, Munger, and Roberts, 1994; Hillman, 2003; Macher, Mayo, and Schiffer, 2011; Macher and Mayo, 2012; Weymouth, 2012). In this research, industry heterogeneity and firm size will be captured and analyzed to obtain a clear overview of potential differences.
To do this in a structured approach, we start off in the next chapter with the literature review. This section holds an overview of the most important and most relevant articles among this topic. The literature derived from this section will be put into a theoretical framework, where own expectations are formulated through hypotheses and a well-developed conceptual model will be applied. After this section, the method of analysis will be described, in which the sample will be explained more into detail, variables are introduced and data is collected. The results evolving from the conducted research are shown in the next section, followed by a discussion in which the results meet the literature and hypotheses, to eventually answer the research question. Finally, the limitations of the study and options for further research are considered.
9
Literature Review
Prior research stated several points which are going to be discussed in this section. Besides going more into detail about CPA and corporate financial performance, the importance of being politically connected as a firm is being discussed. Following up on that, the role a Chief Executive Officer, CEO, or manager plays within a firm has great influence on its political strategy. Literature does not only show that firms which have board members with a political background, and thereby experience at political institutions obtain a better general firm performance (Hillman et al., 1999), their political preference and the effects of those preferences will also being highlighted. Lastly, this research examines the results deriving from contributions of firms’ Political Action Committees, PACs, so a clear explanation will be provided.
First of all, a general overview of the political marketplace will introduce the topic more into detail. The corporate political marketplace consists of suppliers, which are politicians and governmental entities valuing votes and resources which can provide votes, and demanders, which are i.e. firms looking to gain policy decisions in their favor (Hillman & Keim, 1995). Firms design their own nonmarket strategies, in collaboration with other firms or just individually. Their objective is to be effectively present in political markets, which can be achieved by providing votes, either through campaign contributions, through general financial support, or through information sharing regarding policy alternatives and consequences (Hillman & Hitt, 1999). Figure 1 shows a drawing of a political marketplace.
Figure 1: An example of a political market, involving a focal firm that wishes to influence a particular public policy (Bonardi, Holburn, Vanden Bergh, 2006).
10
Corporate Political Activity – Corporate Financial Performance
Former studies about performance measurements deriving from political nonmarket activities already contributed to the literature with accurate measures as corporate financial profitability (Hillman, Zardkoohi, & Bierman, 1999; Shaffer, Quasney, & Grimm, 2000). Firms can obtain those profits from CPA in two ways, which are campaign contributions and lobbying (Hillman et al., 2004; Lux et al., 2011; Tahoun, 2014). Firms do campaign contributions to help a particular political candidate in an election, reciprocally for the political candidate acting in the interests of the firm (Kroszner & Stratmann, 2000; Milyo, Primo, & Groseclose, 2000; Tahoun, 2014). Firms which contribute to a campaign are relatively limited as only those who are seeking to be elected can be reached. With lobbying on the other hand, firms and political actors communicate information to influence certain actions (Chen et al., 2010; Nownes, 2006). This form of influence can be carry out on a broad basis, as all government entities are reachable. The amount firms spent on lobbying is often linked to new income, a higher market share and higher equity returns (Kim, 2008; Shaffer, Quasney, & Grimm, 2000). Firms can enforce several government related activities, i.e. apply for and secure permits (Nownes, 2006), comply with laws and regulations (Peltzman, 1976; Stigler, 1971), secure government contracts (Blumentritt, 2003) and maintain and build connections with officials publically known (Clawson, Neustadtl & Weller, 1998; Goldman, Rocholl & So, 2009). In this case, performance is reached when the firm is able to effectively exploit the policy decisions made in their favor. This can either derive from import tariff policies, from regulated rates, from regulatory and legislative support for environmental emissions standards, or from antitrust decisions. In relation to this, performance increases when the firm is able to come closer to these favorable policies, or by making sure proposals that go against these policies will be blocked (Bonardi, Holburn & Vanden Bergh, 2006).
Political Connectedness
Past literature already observed that firms with established political connections benefit more as taxation is lighter, regulatory oversight is more relaxed, government owned enterprises treat them as a preferential client and government is more likely to sign their contracts with them (Faccio, 2006). Moreover, firms connected to politics have a significant higher chance to be bailed out if economic distress appears, than similar firms with no connections with political instances (Faccio, Masulis and McConnell, 2006). In China for example, firm-government relationships are seen as even more important, if not critical, as obtaining bank loans and gaining institutional support is key (Peng & Luo, 2000; Xin & Pearce, 1996; Bai et al., 2006).
11
For some industries, studies have shown some major correlations between corporate activities deriving from the political point of view and actual firm performance. Example of these are the utilities (Bonardi, Holburn, & VandenBergh, 2006) and the airline industry (Shaffer, Quasney, & Grimm, 2000). These firms, acting in the industries with higher government dependence are more likely to benefit from those political activities, as they are in greater need of gaining political legitimacy (Hillman, 2005; Meznar & Nigh, 1995; Peng & Luo, 2000). Same holds for firms which are privately owned, managers of these firms should actively contribute to political activities, as they could use the available resources and political support even more (Wang & Qian, 2011).
Firms with a large governmental network of connections tend to have the ability and influence to act upon government actions favorable to the firm. These connections are of major importance both in terms of regulatory processes and government contracting (Goldman et al., 2009; Goldman, Rocholl, & So, 2013; Hillman, 2005; Nownes, 2006; Vidal et al., 2012). People without such connections do not have access to so called powerbrokers, people with the influence, leverage and power to make certain important decisions. Not only are firms building up trustworthy relations while engaging in government contact, they also increase their reputation in the market, learn a lot from experience and develop capabilities specifically for performing in these environments (Dean & Brown, 1995). Moreover, direct experience in terms of political contact empower firms in understanding preferences of politicians and recognizing patterns of behavior (Holburn & Vanden Bergh, 2002; Ring, Lenway, & Govekar, 1990). Therefore, it could be said that the number of governmental relationships, also named political connectedness, relates to the benefits a particular firm receives, which subsequently results in a positive outcome for the firm (Ridge, Ingram & Hill, 2017).
There are a couple of ways to increase the firms’ political connectedness. First of all, firms can increase connections by hiring former government officials as lobbyists, as they have access to information and former colleagues still working there (Hillman, 2005; Kim, 2008; Vidal et al., 2012; Levine, 2009; Rosenthal, 2001; Santos, 2006). Secondly, which is following up on the first way, firms can employ former or current government officials as new employees. They will bring both experience and an advanced network of contacts to the firm (Faccio, 2006; Faccio, Masulis, & McConnell, 2006; Hillman et al., 1999). Lastly, firms can improve the connectedness through electoral campaign donations. As mentioned before, this will help candidates securing their position in return for beneficial regulations (Kroszner & Stratmann, 1998; Snyder, 1990; Tahoun, 2014; Tripathi, 2000). With an increasing network of political connections and better established relationships, government officials are likely to share
12
information a firm can use to obtain a greater firm performance, mostly by increasing the value of government contracts (Ridge, Ingram & Hill, 2017).
CEO political behavior
CEO’s or managers of companies in the United States often have specific political preferences, leaning more to either the Republican side, or the Democrat side. This comes forward in their personal ideologies and corporate behavior, which can be both intended or unintended. Republican – leaning managers tend to be, on average, more conservative and are more likely to set financially conservative policies (Hutton, Jiang & Kumar, 2014). In the rule, firms with Republican managers have therefore less debt than firms with Democrat – leaning leaders, comparable in size and industry. These characteristics also influence the way managers, and with them firms choose the level of investments made. Where Democrat managers invest more in research and development, tangible assets and risky investments, Republican managers choose for profitability in the short run, and invest less in the above-mentioned divisions. Although this choice of lower innovations may increase profits, it could dampen innovation and can cost shareholders a lot of money in the long run (Hutton, Jiang, Kumar, 2014).
Thus, conservative people are more likely to favor the Republican Party, in particular since the 1970’s (Abramowitz & Saunders 2006). Hutton et al. (2014) especially stress in their research that, “the political contributions of managers reflect their personal political orientation and can serve as a good proxy for the level of their financial conservatism”. Moreover, this will influence their choice of political contributions as well. Corporate managers have the option to do a direct contribution to candidates or party committees on personal behalf, or to do an indirect contribution through their own company political action committees (PACs), explained more detailed in the following section. PAC contributions are normally going to both parties at an election, 71% of managers acting on personal behalf tend to contribute to just one party (Hutton, Jiang & Kumar, 2014).
PAC Contributions
A Political Action Committee (PAC) is an organization that supports political candidates through donations made by members, which on firm-level mostly results in positive abnormal results (Cooper, Gulen, & Ovtchnikov, 2010). PAC’s can be associated with a firm, from where contributions are made. However, the contributions are eventually made by shareholders and employees of the firm. These contributions are not only made for political reasons, they can be made to raise the image and reputation of a company as well, thus strategically (Brammer &
13
Millington, 2005; Godfrey, 2005; Porter & Kramer, 2002; Saiia, Carroll, & Buchholtz, 2003). In that case, risks of reputational losses are mitigated and critical resources from stakeholders are secured (Fombrun, Gardberg, & Barnett, 2000; Godfrey, 2005; Williams & Barrett, 2000). Corporate contributions might help to gain approval from government officials or political legitimacy, which gives a particular firm access to political resources. In economies of many nations it is very hard for firms to enjoy policy benefits and to gain political resources, despite of the critical need for it to prosper development (Wang, Qian, 2011).
It is beyond dispute that the height of demand a particular firm is able to make, rests with the amount of resources this firm possesses. Therefore, more votes, information and money results in more positive policy outcomes (Dahan, 2005). Other literature however, doubts about the best amount of contributions and provides another view with no linear growth of more resources resulting in a better outcome. According to Brammer and Millington (2008), firms could optimally contribute if they offer unusual high or low amounts of money, as financial performance is respectively best in the long and in the short term. Taking the middle road in this case will have no or a negative effect both in the short and in the long term strategy of the firm. Next to that, firms usually contribute money to both political parties, making it interesting which elective party obtains more financial support (Hutton, Jiang & Kumar, 2014).
Theoretical Framework
Without any doubt, it could be said that a lot of researchers analyzed effects of Corporate Political Activity on Corporate Financial Performance. However, a clear and precise image of the outcome of these effects and the drivers affecting the outcomes is lacking. Not only does prior research not provide either a straight positive or negative effect of CPA on CFP (Dahan, 2005; Faccio, 2006; Hillman et al., 2004;Lux et al., 2011; Shaffer, Quasney, & Grimm, 2000), lots of studies are also conducted rather broadly, with the focus on CPA as a unified term (Hadani & Schuler, 2013). Aggarwal et al., (2012), state in their research that some political activities performed by firms are independent of each other as well. The aspect of the precise effect of financial CPA as i.e. contributions and lobbying on a firm’s financial performance is not followed up either (Hillman & Hitt, 1999). Taken the role of a firm’s political preference or CEO of a particular firm’s political preference into account has not been done at all. Therefore, this research explores at first the dominant relationship between CPA and CFP. As a second and last relationship, the political preferences within the States in the U.S. are examined as a moderating effect. This means in theory that each form of political state
14
preference and diversity in political contributions made to either one of the two large political parties in the U.S. will lead to a divergent effect on the dominant relationship CPA – CFP. This section outlines the concepts and works towards the relationships shortly explained above. The two possible effects are described more into detail and a framework based on the available theories is presented. Moreover, these concepts and theoretical framework will lead towards a set of different hypotheses, which are tested later in this study.
The effect of CPA on CFP
In total, there is plenty of literature on Corporate Political Activity aspects, some also written in context with effects on Corporate Financial Performance. Moreover, with politicians keep on willing to offer growth opportunities to firms, supply and demand for this effect will undoubtedly stay (Hillman, Keim, & Schuler, 2004). Firms can be politically active in a couple of ways. First of all, lobbying is seen as a political activity to get information from government officials and change their strategy upon the tips and insights they have received (Chen et al., 2010; Nownes, 2006). Secondly, firms can engage in political contributions by using a Political Action Committee (PAC). In this way, shareholders and employees of a certain firm contribute to political associations in return for regulation changes in benefit of the firm (Kroszner & Stratmann, 2000; Milyo, Primo, & Groseclose, 2000; Tahoun, 2014). Lastly, contributions can be made by the CEO of a firm. CEO’s making contributions are doing it with the same targets and to achieve the same goals as firms contributing throughout PAC’s, to let politicians act positively for the firm. Dahan (2005) argued, the more money politicians receive, the better policies are changed in the interest of those firms. However, the amount of money is an important issue to take into account, as a linear growth curve is absent. Taking the middle road is usually killing for both long and short term performance and doing no, little or a lot of contributions help firms achieving the political assets they want (Brammer and Millington, 2008).
Financial performance can be obtained both directly and indirectly. Direct visible advantages are i.e. the security of government contracts, what normally originates after a long and strong connection between the firm and government officials (Blumentritt, 2003; Ridge, Ingram & Hill, 2017), the application and security of permits (Nownes, 2006) and image building by linking and associating publically known politicians to your firm (Goldman, Rocholl & So, 2009). Indirect advantages are i.e. tax minimization and a more relaxed regulatory oversight (Blumentritt, 2003; Faccio, 2006; Hart, 2001; Chen et al., 2010; Hillman et al., 1999). Import tariff policies could be changed in interest, just as that negative proposals
15
could be blocked by politically known officials (Bonardi, Holburn, Vanden Bergh, 2006). Lastly, a firm can become more interesting for other firms to take over or to start a partnership with, which prospers a firm’s financial performance as well (Baron, 1995 & Oliver & Holzinger, 2008).
Based on these theories, both the ways to perform political activity and the gained financial performance provide a positive possible effect. Although the height of contributions is another aspect to take into account when targeting the optimal result, all other theories argue the importance of CPA for the CFP. Moreover, as Baron (2001) examined in his study, especially when political rivalry is at its top, politicians are looking for contributions and are more than willing to change policies openly in the interest of the firms. It could be said that the nonmarket strategy performance of firms increases through rivalry (Bonardi, Holburn, Vanden Bergh, 2006). Taking the theories and the aforementioned example into consideration, the next hypotheses are formulated:
H1: There is a positive relationship between the amount of CEO contributions of a firm and the Corporate Financial Performance.
H2: There is a positive relationship between the amount of PAC contributions of a firm and the Corporate Financial Performance.
The moderating effect of State Diversity on the Main Relationship
Consequently, political contributions are expected to affect financial firm performance positively. However, the United States political system contains two major parties, the Republican Party and the Democrat Party, who hold relatively different ideologies against each other. Subsequently, firms usually contribute money two both political parties, which makes it even more interesting which of the two parties receives more support. For CEO’s only, it is somehow different. Namely, 71% of the CEO’s tend to contribute to just one party and automatically expresses his or her political preference in public (Hutton, Jiang & Kumar, 2014). Society on the other hand does not react that well on both firm and CEO contributions most often. Scholars argue the negative view on political interference from business associations which frequently results in brand depreciation of the concerning firm (Smith, 2000; Torres-Spelliscy, 2016). As an effect, some companies but mostly CEO’s prefer to contribute anonymous and by doing so stay unknown for the public (Smith, 2000). Other researchers indicate that firms and CEO’s contribute deliberately as some kind of marketing business
16
strategy, to raise the image and reputation (Godfrey, 2005; Porter & Kramer, 2002; Saiia, Carroll, & Buchholtz, 2003).
These types of strategies and ways to engage in CPA can be analyzed by looking at both political parties and their characteristics. Republicans tend to have a more conservative attitude, as they hold less debt, make lower investments and are more focused on short term rather than long term profits. Regarding the type of contributions and the effects that come with it, Republican firms, managers and people prefer to make PAC contributions over CEO contributions. Reasons for this could be i.e. the fear of negative opinion from the crowd and image depreciation, or the feeling that with PAC contributions the politicians are more supported, as PAC contributions are normally higher than just CEO contributions (Hutton, Jiang & Kumar, 2014). Democrats are known as more risky, making more investments in i.e. innovation, R&D and in tangible resources and assets. They lean more towards long term profits and are more progressive. In general, Democrats do not hold a specific preference for a certain type of contribution, either PAC or CEO contributions (Hutton, Jiang & Kumar, 2014). In terms of influence of the states in the United States, the House of Representatives elections provide a clear picture of the political party disunity in the country. Based on the theories mentioned above and the examples given from former Republican and Democrat behavior, the following is hypothesized:
H3: The positive relationship between either the amount of CEO contributions or PAC contributions and Corporate Financial Performance is weaker when the firm’s HQ is located
in a Republican State..
H4: The positive relationship between either the amount of CEO contributions or PAC contributions and Corporate Financial Performance is stronger when the firm’s HQ is
located in a Democrat State.
Conceptual Model
Based on the theoretical framework explained above, a model is conducted which illustrates the hypothesized relationships of this study. As can be seen in figure 2, both the relationship between CPA and CFP is shown and the moderating effect of each firm’s Headquarter location in combination with contributions to a political preferred party in each United States’ state.
17
Corporate Political Activity
Corporate Financial Performance
Most of the amount contributed to an association of the political
party in favor in the State of the U.S. where the HQ is located
Figure 2: Theoretical Model
Methodology
SampleTo investigate the above mentioned hypotheses on the effects of PAC and CEO contributions on the firm’s financial performance and the role of the political party in favor in several states in the United States, a sample from two databases is constructed. Both databases consist information of Fortune 500 firms from 2012 till 2015. The data in these sets show amounts of contributions firms made to associations of either the Democrat or the Republican party, the CEO’s made to the parties and other expenditures concerning lobbying activities. Furthermore, financial data as sales records, income states and other resources are stated over these four years as well. Based on the location of each firm’s Headquarter, a list of States which are present from these firms comes forward. From the 50 States, 38 States of the United States of America have one or more Headquarters of firms from the Fortuna 500 list located over these four years. One state more packed with these firms than other states obviously.
Moreover, data concerning industries and political parties are added from other sources. From each firm, the SIC code of the industry they are operating in is available. The North American Industry Classification System (NAICS) provides a clear overview of industries matching the SIC codes. Britannica, the oldest English encyclopedia, explains the role of the House of Representatives in the United States. This governmental pile is responsible for approving bills set in states. The political party with most members in the House of Representatives per state has an advantage in determining policy. Elections from the 112th, 113th
and 114th United States Congress show the members of the House of Representatives with a
position in each state (History, Art & Congress). Based on this data, the political party in favor, either the Democrat or the Republican party, could be determined. This choice of sample is very appropriate for the research questions I am investigating, while data of these 500 firms is
18
clear and accessible. Moreover, the amount of contributions and expenses of a firm and the firm’s financial performance exactly show the correlations I am looking for, to answer the hypotheses and research questions.
Data Collection
The measurements of each of the variables below directly shows what is intended to measure. For each firm, I am going to examine the correlations between the Corporate Political Activity by amount of PAC contributions and CEO contributions in total Dollars, and the Corporate Financial Performance by the ratio of Return on Assets and the Tobin’s Q ratio to show potential relationships. I am going to do this by using the statistical software program SPSS. Thereafter, with SPSS I will process my data and measure the relationships between the states. This is done by checking which of the two political parties are in favor in a state and correlating the total contributed Dollars of a firm’s HQ located in that state to that favorable party, with again the corporate financial performance in Return on Assets and the Tobin’s Q ratio. Some variables have less number of observations than others, as there were some missing values within the datasets. Eventually, it is of most interest to see the effect of contributions to the favorable party, in relation to each of the firm’s financial performance.
Variables
Corporate Financial Performance is the dependent variable. For this variable, two measures were employed. One way to measure is by utilizing return on assets (ROA), where net income is calculated over total assets. This is a common measure of Corporate Financial Performance in the accounting world (Wang & Qian, 2011). Secondly, the market value of each company is divided by the replacement value of the firm’s total assets, generally known as the Tobin’s Q ratio. A high Q ratio, higher than 1, indicates an outstanding financial performance, where a low Q ratio, between 0 and 1, indicates a poor financial performance. Tobin’s Q is commonly used as a market measure of a firm’s value. Some advantages of this measure is that the ratio is risk-adjusting, forward-looking and an indicator of the intangible value of a firm (Montgomery & Wernerfelt, 1988; Hall, 1993; Hirschey, 1982; Megna & Klock, 1993). Most interesting is that the Tobin’s Q ratio is used to study effects of types of market power advantages on financial performance, which is applicable to this study. Both measures, ROA and Tobin’s Q, are widely used in management research (Hillman, 2005; Tuschke & Sanders, 2003).
The independent variable in this study is Corporate Political Activity. The political activity a given firm engages in, is provided by the amount of the firm’s CEO contributions and
19
the amount of the firm’s PAC contributions (Hillman & Hitt, 1999). Besides the total amount of money contributed, both measurements give the exact amount of money contributed to either a Republican Party association or a Democrat Party association. Unknown contributions are shown as well.
As a moderator, the influence of a particular firm’s Headquarter location per State in the United States of America is examined. Politically, each state has either the Republican or the Democrat Party in favor, based on House of Representatives elections over the years 2012 till 2015. According to the hypotheses in this study, firms who contribute more to the political party in favor of where their Headquarter is located in, generate more positive financial results than firms who contribute more to the other political party. For this variable a dummy variable is created. If the relationship of the conceptual model is tested within Republican States, all ‘Republican firms’ are coded with a 1, and all ‘Democrat firms’ with a 0. If the relationship is tested within Democrat States, coding is the other way around.
Firm Size, Slack Resources and Type of Industry are the measures taken as control variables. For firm size, the natural logarithm of total assets per firm is measured. Previous studies show that firm size is an important variable to analyze the influence of a firm’s political activity on financial performance, as larger firms normally carry more resources than smaller firms and may utilize advantages as economies of scale (Orlitzky, 2001; Roberts & Dowling, 2002). Another important antecedent of political contributions are slack resources. Slack resources are measured as the total cash flow from i.e. the financing, investing and operating activities of a firm, scaled by total assets (Buchholtz, Amason & Rutherford, 1999; Seifert, Morris & Bartkus, 2004). Furthermore, the 10 different industry categories identified by the North American Industry Classification System (NAICS) are controlled for. Studies have shown that the type of industry affects how a firm is politically active (Galaskiewicz & Burt, 1991). 10 dummies are created to check for any differences in political behavior and financial performance across industries (Seifert, Morris & Bartkus, 2004).
Method
The hypotheses are tested by running a multiple regression analyses with the equations below. For the first and second hypothesis, the main effect of this study is tested. Therefore, the following regression equation is used:
20
𝐶𝑜𝑟𝑝𝑜𝑟𝑎𝑡𝑒 𝐹𝑖𝑛𝑎𝑛𝑐𝑖𝑎𝑙 𝑃𝑒𝑟𝑓𝑜𝑟𝑚𝑎𝑛𝑐𝑒𝑖
= 𝛽0+ 𝛽1𝐶𝐸𝑂𝑐𝑜𝑛𝑡𝑟𝑖𝑏𝑢𝑡𝑖𝑜𝑛𝑠𝑖 + 𝛽2𝑃𝐴𝐶𝑐𝑜𝑛𝑡𝑟𝑖𝑏𝑢𝑡𝑖𝑜𝑛𝑠𝑖 + 𝛽3𝐹𝑖𝑟𝑚𝑠𝑖𝑧𝑒𝑖
+ 𝛽4𝑆𝑙𝑎𝑐𝑘𝑅𝑒𝑠𝑜𝑢𝑟𝑐𝑒𝑠𝑖+ 𝛽5𝐼𝑛𝑑𝑢𝑠𝑡𝑟𝑦𝐷𝑢𝑚𝑚𝑖𝑒𝑠𝑖 + 𝜀
For the moderating effect, several regressions will be run. For every regression analysis, the basic model will be adjusted. The following basic regression equation will be used:
𝐶𝑜𝑟𝑝𝑜𝑟𝑎𝑡𝑒 𝐹𝑖𝑛𝑎𝑛𝑐𝑖𝑎𝑙 𝑃𝑒𝑟𝑓𝑜𝑟𝑚𝑎𝑛𝑐𝑒𝑖
= 𝛽0+ 𝛽1𝑅𝑒𝑝𝑢𝑏𝑙𝑖𝑐𝑎𝑛𝐶𝐸𝑂𝑐𝑜𝑛𝑡𝑟𝑖𝑏𝑢𝑡𝑖𝑜𝑛𝑠𝑖
+ 𝛽2𝑅𝑒𝑝𝑢𝑏𝑙𝑖𝑐𝑎𝑛𝐶𝐸𝑂𝑐𝑜𝑛𝑡𝑟𝑖𝑏𝑢𝑡𝑖𝑜𝑛𝑠 ∗ 𝑅𝑒𝑝𝑢𝑏𝑙𝑖𝑐𝑎𝑛𝑆𝑡𝑎𝑡𝑒𝑖+ 𝛽3𝐹𝑖𝑟𝑚𝑠𝑖𝑧𝑒𝑖
+ 𝛽4𝑆𝑙𝑎𝑐𝑘𝑅𝑒𝑠𝑜𝑢𝑟𝑐𝑒𝑠𝑖+ 𝛽5𝐼𝑛𝑑𝑢𝑠𝑡𝑟𝑦𝐷𝑢𝑚𝑚𝑖𝑒𝑠𝑖 + 𝜀
Corporate Financial Performance represents the dependent variable in this equation measured in Return on Assets and Tobin’s Q, 𝛽0 is the intercept. For the first equation, the independent variables are 𝛽1 and 𝛽2. These beta’s represent the total amount of CEO contributions and PAC contributions and are controlled for 𝛽3, 𝛽4 and 𝛽5. 𝛽5 hold nine of ten industry dummies of the database, as one is excluded from the regression. 𝜀 is the error term, the residual.
For the second equation, the basic model consists of Corporate Financial Performance as the dependent variable, again measured in Return on Assets and Tobin’s Q. 𝛽0 is the intercept and 𝛽1 represents the CEO contributions made to associations of the Republican Party. To measure the moderating effect, 𝛽2 is stated as the CEO contributions made to associations of
the Republican Party, multiplied by the dummy variable of the states, to get the interaction effect. In this case, firm’s which have a Headquarter located in a Republican State hold a 1, firm’s which have a Headquarter located in a Democrat State hold a 0. Equal to the first equation, 𝛽3, 𝛽4 and 𝛽5 hold the control variables and 𝜀 is the residual term.
The first regression is performed to test the effect of control variables on the dependent variables. Subsequently, the first independent variable, the total amount of CEO contributions per firm, is added. Thirdly, the total amount of PAC contributions is added as the second independent variable. The forth model consist of both independent variable, as it could be interesting as a certain indication. Now, the second equation is used to measure the moderating effects. First of all, Republican CEO contributions are added to the regression analyses, together
21
with the interaction effect of these Republican CEO contributions made in Republican States, coded with a 1.
Then, as an adjustment on the basic equation illustrated above, instead of the dummy variable of the Republican States, the dummy variable of the Democrat States is used. Next, same is done for the Republican PAC contributions, with both dummy variables. The last models are performed with the Democrat contributions. Similar to the Republican contributions, the first regression is run with the Democrat CEO contributions, added by the interaction effect of the Democrat CEO contributions multiplied by the dummy variable of the Democrat States. The model hereafter is done with the dummy variable of the Republican States. The same procedure is done for the PAC contributions. Eventually, the total regression analysis results in a total of 12 models, as is shown in figure 4 and figure 5 in the next section of this research.
Results
This section provides a clear overview of the statistical analyses. First, a table is presented with the descriptive statistics of the variables included. Secondly, a correlation test is shown, followed by a Variance Inflation Factor multicollinearity test. Lastly, detailed analyses of the regression models are presented in two tables.
Descriptive Statistics
In table 1, the descriptive statistics of all variables included in this research are displayed. With an average Tobin’s Q ratio of 1,084, an indication is given of the relatively high performing position of the 549 firms in the dataset, some being superior with even a Tobin’s Q ratio of over 5. For the independent variables, the statistics show that firm’s engaging in contributions differ a lot from each other. With a mean of almost 84000 U.S. Dollars, a maximum of 31 million U.S. Dollars and a standard deviation of 1.3 million U.S. Dollars it could be said one CEO is definitely more generous than other CEO’s. For both CEO and PAC contributions, the Republican Party receives more contributions than the Democrat Party. In terms of PAC contributions it is almost twice as much, for CEO contributions it shows even an average amount which is more than three times as much. However, the mean of Republican and Democrat CEO contributions do not add up to the total, because there are also a lot of contributions made, in which the political party is unknown. For the PAC contributions this is
22
not the case, as means of all three variables is clearly shown. This could explain the larger difference in Republican and Democrat CEO contributions, although it could also be even more. From the 557 firm’s with a Headquarter location which is known, 311 are located in Republican States, and 246 in Democrat States. Based on theory and the hypotheses, this could declare why there are more CEO and PAC contributions made to associations of the Republican Party, as 56% of the firms are located in Republican States and could be more inclined to contribute to that party. The control variables are divided in Firm size, Slack resources and 10 industries. With a mean of 4,224 for Firm size and a standard deviation of 0,615, there is evidence that the firms in this analysis are in general large of size. The variable Slack resources states large difference, having a mean of 435 and a standard deviation of 3383. Therefore, there are definitely a lot of outliers with an abnormal distribution. About the industry variable it could be said that the manufacturing industry hold most firms, followed by the transportation, communication, electric & gas industry, respectively 193 and 84, a large difference.
Correlations
A Pearson correlation test is conducted to look into the effects of each variable on all other variables used in the analyses. From these effects, the coefficients and significance levels are stated. Table 2 holds the correlation test, divided in dependent variables, independent variables, moderating variables and the control variables. The first and the second hypotheses expect a positive relationship between the dependent variables ROA and Tobin’s Q and the independent variables CEO and PAC contributions. The dependent variables correlate high and significant and although all effects between the dependent and independent variables are also positive, the correlation is rather weak and insignificant. The other hypotheses expect positive effects between contributions made to an association of a party which is in favorite in the state where the firm’s Headquarter is located. The variable Democrat CEO contributions shows a positive, but weak significant effect on the Tobin’s Q ratio: ,085 with p<.05. The Republican and Democrat States are both correlating significant with the dependent variables. However, based on a correlation table and the dummy variables of the moderating effect, no conclusions can be drawn. Finally, the control variables Firm size and most of the Industry types correlate significantly with the dependent variables and the independent variable PAC contributions. As expected, larger firms contain more resources, employees and have the possibility to contribute more. Slack resources has a weak and insignificant correlation towards the dependent variables, although they are positive. With the multicollinearity test in the next part, the variables should
23
not be predicted by one another. This needs to be done to run a strong as possible regression analysis with the right coefficients.
Table 1: Descriptive Statistics. This table presents the number of observations, minimum, maximum, sum, mean and standard deviation for each variable. The sample consists of data from Fortune 500 firms over the years 2012 till 2015.
N Minimum Maximum Sum Mean Std. Deviation
Return on Assets 549 -,667 ,264 ,046 ,066
Tobin's Q 549 -,011 5,462 1,084 ,931
Total CEO contributions 559 ,00 31294447 46731128 83597,724 1325037,809
Democrat CEO contributions 559 ,00 356150 2396057 4286,327 18015,249
Republican CEO contributions 559 ,00 583650 8209794 14686,573 39453,465
Total PAC contributions 556 ,00 3087564 118821976 213708,590 380167,513
Democrat PAC contributions 556 ,00 1310678 43995354 79128,335 152139,152
Republican PAC contributions 556 ,00 1776886 74758770 134458,218 238437,799
Republican States 557 0 1 311 ,56 ,497
Democrat States 557 0 1 246 ,44 ,497
Firm size 549 2,018 6,511 4,224 ,615
Slack resources 421 ,645 61418,994 435,258 3383,810
Agriculture, Forestry & Fishing Industry 568 0 1 1 ,00 ,042 Mining Industry 568 0 1 21 ,04 ,189 Construction Industry 568 0 1 9 ,02 ,125 Manufacturing Industry 568 0 1 193 ,34 ,474 Transportation, Communication, Electric & Gas Industry
568 0 1 84 ,15 ,355
Wholesale Trade Industry 568 0 1 36 ,06 ,244
Retail Trade Industry 568 0 1 62 ,11 ,312
Finance, Insurance & Real Estate Industry
568 0 1 77 ,14 ,343
Services Industry 568 0 1 51 ,09 ,286
24
Table 2: Pearson Correlations.This table expresses the pair wise correlation matrix for all variables in the sample. The dependent variables are Return On Assets and Tobin’s Q.
+ Correlation is significant at the 0.1 level (2-tailed). * Correlation is significant at the 0.05 level (2-tailed). ** Correlation is significant at the 0.01 level (2-tailed).
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 1. Return on Assets 1 2. Tobin’s Q ,529** 1 3. Total CEO contributions ,035 ,065 1 4. Democrat CEO contrib. -,040 ,085* ,007 1 5. Republican CEO contrib. ,023 ,059 ,241** ,005 1 6. Total PAC contributions ,041 ,018 -,011 ,026 ,100* 1 7. Democrat PAC contrib. ,053 ,016 -,017 ,040 ,063 ,957** 1 8. Republican PAC contrib. ,032 ,019 -,007 ,016 ,119** ,983** ,887** 1 9. Republican States -,141 ** -,089* ,036 -,076+ ,078+ -,031 -,092* ,009 1 10. Democrat States ,141** ,089* -,036 ,076+ -,078+ ,031 ,092* -,009 -1,00** 1 11. Firm size -,076+ -,180** ,015 ,011 ,112** ,414** ,401** ,404** -,152** ,152** 1 12. Slack resources ,029 ,021 -,007 -,013 -,030 -,035 -,031 -,037 ,079 -,079 -,061 1 13. Agri, Fore & Fish Indus.
,044 ,058 -,002 ,010 -,011 ,026 ,013 ,034 ,038 -,038 ,007 -,005 1 14. Mining Industry -,364 ** -,035 -,003 -,032 ,109* -,007 -,072+ ,035 ,138** -,138** ,069 ,057 -,008 1 15. Construct. Industry ,005 -,080+ -,006 ,045 -,030 -,034 -,036 -,030 ,057 -,057 -,080+ ,436** -,005 -,025 1 16. Manufact. Industry ,210** ,190** -,035 -,082+ -,045 ,005 -,019 ,021 -,035 ,035 -,098* -,054 -,030 -,141** -,091* 1 17. Trans, Comm, Elec & Gas Indust.
-,096* -,002 -,017 ,018 ,019 ,135** ,147** ,121** ,057 -,057 ,111** -,051 -,017 -,082+ -,053 -,299** 1 18. Whole. Trade Indust. -,015 -,091* -,016 -,048 -,083+ -,106* -,098* -,106* -,001 ,001 -,245** -,027 -,011 -,051 -,033 -,187** -,108** 1 19. Retail Trade Indust. ,154** ,218** -,016 -,047 ,031 -,083+ -,081+ -,080+ ,092* -,092* -,174** ,037 -,015 -,069 -,044 -,251** -,146** -,091* 1 20. Fin, Insur & R.E Ind.
-,113** -,411** -,016 ,002 ,047 -,004 ,020 -,019 -,163** ,163** ,406** -,018 -,017 -,078+ -,050 -,284** -,165** -,103* -,139** 1 21. Services Industry ,014 ,131 ** ,142** ,205** ,008 -,065 -,039 -,078+ -,037 ,037 -,038 -,028 -,013 -,062 -,040 -,225** -,131** -,082+ -,110** -,124** 1 22. Public Admin. Indu. -,005 -,066 -,004 ,061 -,031 ,112** ,125** ,100* -,030 ,030 ,096* -,009 -,004 -,018 -,012 -,068 -,039 -,025 -,033 -,037 -,030 1
25
Table 3: Tolerance Statistics and Variance Inflation Factors. This table presents the outcomes of the tolerance statistics and the variance inflation factor for all variables. The dependent variable are Return on Assets and Tobin’s Q.
VIF table
In this table, three variables are excluded. First, the variable of Total PAC contributions, which is argumentative as the Republican and Democrat PAC contributions add up to this variable. Secondly, the variables of Republican States and Manufacturing Industry are excluded. This is a logical effect, as both variables are dummies so coded 1 or 0, and within their group both variables are highest. Compared to Democrat States there are more
Republican and compared to all other Industries, the Manufacturing Industry is represented best. To detect a possibility of multicollinearity, a check has been done using a Variance Inflation Factors table with tolerance statistics. As is shown in table 3 above, no problem of multicollinearity arises. First of all, not one VIF is larger than 10, just the Democrat and Republican PAC contributions are higher than 5. Moreover, the average VIF is 1,7 which is not significantly higher than 1. The next part contains the regression analyses.
Regression Analyses
To check the hypotheses set in the theoretical framework, regression analyses are used. In total, 14 different models are computed to note the effect on the dependent variables ROA and
Tolerance VIF
Total CEO contributions ,870 1,149
Democrat CEO contributions ,927 1,078 Republican CEO contributions ,809 1,236 Democrat PAC contributions ,188 5,318 Republican PAC contributions ,183 5,453
Democrat States ,857 1,167
Firm size ,550 1,819
Slack resources ,795 1,258
Agriculture, Forestry & Fishing Industry
,991 1,009
Mining Industry ,859 1,164
Construction Industry ,789 1,267
Transport., Communication, Electric & Gas Industry
,772 1,295
Wholesale Trade Industry ,827 1,209
Retail Trade Industry ,837 1,195
Finance, Insurance, Real Estate
Industry ,950 1,053
Services Industry ,808 1,237
26
Tobin’s Q. In table 4 and table 5, β values show the effect of each variable in every model. The adjusted R2 is included to test for the fitness of the model, as it indicates variance in the
dependent variables. When the model is improved after a certain variable is added, the R2
adjusts. Similar to the correlation table, a coefficient is significant when p < .01, p < .05 and when p < .1. As is mentioned before, table 4 and table 5 each show 12 models. These 12 models can be divided in two groups. The first 4 models are related to Hypothesis 1 and Hypothesis 2, measuring the direct effect of CPA on CFP. The other 8 models contain information for the other hypotheses, in which the effect of the moderating variables are included.
Model 1 starts with a regression of the control variables. For ROA, there is a significant positive effect on firm size with β= - .016, and p < .05. This significant effect holds for all the twelve models. Three of the nine industries are significant and for Tobin’s Q five from the nine industries are significant. These industries keep on providing a significant effect throughout the 12 models. In the second model, the total CEO contributions are added. For both dependent variables however, the variable CEO contributions holds a positive but weak and insignificant effect. Model 3 has been conducted with the total PAC contributions instead of CEO contributions and model 4 contains both. Both models do not state any
significant effects, not in the ROA table and neither in the table of Tobin’s Q. The adjusted R2
between the four models does not change and the first two independent variables are therefore not an improvement of the model. The independent variables total CEO and total PAC
contributions both have positive effects with ROA and Tobin’s Q. On the other hand, both effects are weak and insignificant.
From model 5 on, the moderating effects are stated. The variable with the Republican CEO contributions are taken in model 5 together with the Republican amount of CEO
contributions made in Republican States. Hereafter, the Republican CEO contributions are taken together with the contributions made in Democrat States. As explained in the method part above, this is done with the Republican PAC contributions as well. Model 9 till model 12 express the Democrat contributions, both CEO and PAC. For the ROA table, only Republican CEO contributions contain significant results. The Republican CEO contributions illustrate β= 4,527E-7, and p < .01. The interaction between Republican CEO contributions multiplied by the Republican States show an effect of β= -3,210E-7, and p < .1. The interaction effect between Republican CEO contributions multiplied by Democrat States is also significant with β= -3,210E-7, and p < .1. All other models do not contain any significant effects. In this case
27
both Hypothesis 3 and Hypothesis 4 are in line with the regression models looking at the CEO contributions.
For the Tobin’s Q table, the variable which contains the total of Republican CEO contributions is again significant. Difference with the ROA is that the Democrat CEO contributions are significant as well. Total Democrat CEO contributions are significant in model 10 with β= 5,809E-6 and p < .01. The interaction between Democrat CEO
contributions and Republican States gives β= -1,209E-5 and p < .1. Same holds for the interaction with the Democrat States, although the effect is positively significant with β= 1,209E-5 and p < .1. Again, both hypotheses are in line with the regression models when looking just at the CEO contributions. The PAC contributions do not hold any significant results. The adjusted R2 in the ROA table only shows an improvement in model 5 and model
6 towards ,219. In the Tobin’s Q model is a similar improvement visible. Model 9 and model 10 contain an improved adjusted R2 of ,092.
+ Correlation is significant at the 0.1 level (2-tailed). * Correlation is significant at the 0.05 level (2-tailed). ** Correlation is significant at the 0.01 level (2-tailed).
Model 1 Model 2 Model 3 Model 4
Constant -,004 -,006 -,002 -,001
Firm size ,016* ,016* ,015+ ,015+
Slack resources 1,597E-6 1,597E-6 1,600E-6 1,603E-6
Agriculture, Forestry & Fishing Industry
,047 ,047 ,047 ,047
Mining Industry -,159** -,159** -,158** -,158**
Construction Industry -,051 -,051 -,051 -,052
Transportation, Communication, Electrics &
Gas Industry
-,036** -,036** -,036** -,036**
Wholesale trade Industry -,012 -,012 -,012 -,012
Retail Trade Industry ,008 ,009 ,009 ,009
Finance, Insurance & Real
Estate Industry -,017 -,017 -,017 -,017
Services Industry -,017+ -,018+ -,016 -,018+
Public Administration
Industry ,016 ,017 ,013 ,012
Total CEO contributions 1,777E-9 1,793E-9
Total PAC contributions 2,982E-9 3,191E-9
Adjusted R2 ,209 ,208 ,207 ,206
Table 4: Regression Results for Return on Assets. This table represents the outcomes of the first four models of the regression analyses, concerning Hypothesis 1 and Hypothesis 2. The dependent variable is Return on Assets. Model 1 contains only control variables and Models 2 and 3 contain the effect of either Total CEO or Total PAC contributions. Model 4 includes both independent variables together with all control variables.
28 + Correlation is significant at the 0.1 level (2-tailed).
* Correlation is significant at the 0.05 level (2-tailed). ** Correlation is significant at the 0.01 level (2-tailed).
Model 5 Model 6 Model 7 Model 8 Model 9 Model 10 Model 11 Model 12
Constant ,007 ,007 ,000 ,000 -,006 -,006 -,001 -,001
Firm size ,013+ ,013+ ,015+ ,015+ ,017* ,017* ,015+ ,015+ Slack resources 1,691E-6+ 1,691E-6+ 1,608E-6 1,608E-6 1,618E-6 1,618E-6 1,609E-6 1,609E-6 Agriculture,
Forestry & Fishing Industry ,051 ,051 ,049 ,049 ,051 ,051 ,048 ,048 Mining Industry -,164** -,164** -,157** -,157** -,159** -,159** -,158** -,158** Construction Industry -,052 -,052 -,051 -,051 -,052 -,052 -,051 -,051 Transportation, Communication, Electrics & Gas Industry -,035** -,035** -,035** -,035** -,035** -,035** -,035** -,035** Wholesale trade Industry -,011 -,011 -,012 -,012 -,012 -,012 -,012 -,012 Retail Trade Industry ,008 ,008 ,008 ,008 ,008 ,008 ,008 ,008 Finance, Insurance & Real Estate Industry -,016 -,016 -,017 -,017 -,015 -,015 -,017 -,017 Services Industry -,016 -,016 -,016 -,016 -,013 -,013 -,016 -,016 Public Administration Industry ,020 ,020 ,004 ,004 ,016 ,016 ,006 ,006 Republican CEO contributions 4,527E-7** 1,317E-7 Republican CEO contributions * Republican States -3,210E-7+ Republican CEO contributions * Democrat States 3,210E-7+ Republican PAC
contributions 1,558E-8 -1,383E-9
Republican PAC contributions * Republican States -1,696E-8 Republican PAC contributions * Democrat States 1,696E-8 Democrat CEO contributions -4,433E-7 -1,346E-7 Democrat CEO contributions * Democrat States 3,087E-7 Democrat CEO contributions * Republican States -3,087E-7 Democrat PAC
contributions -1,733E-9 1,652E-8
Democrat PAC contributions * Democrat States 1,825E-8 Democrat PAC contributions * Republican States -1,825E-8 Adjusted R2 ,219 ,219 ,205 ,205 ,207 ,207 ,205 ,205
Table 5: Regression Results for Return On Assets. This table represents the outcomes of the last eight models of the regression analyses, concerning Hypothesis 3 and Hypothesis 4. The dependent variable is Return on Assets. Models 5 and 6 contain Republican CEO contributions in both Republican and Democrat States. Models 7 and 8 contain Republican PAC contributions in both Republican and Democrat States. Models 9 and 10 contain Democrat CEO contributions in both Democrat and Republican States and Models 11 and 12 contain Democrat PAC contributions in both Democrat and Republican States.
