The Influence of CEOs’ Political Ideology on
Corporate Political Activities
Author: Lisa Kremer ID: 11746173
Supervisor: Dr. Pushpika Vishwanathan
Thesis submitted in partial fulfillment of the requirements of a M.Sc. Degree in
Business Administration - Strategy
University of Amsterdam
Graduate School of Business and Economics June 2018
i
Statement of Originality
This document is written by student Lisa Kremer who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document is 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.
ii
TABLE OF CONTENTS
ABSTRACT ………... 1
1. INTRODUCTION ……….……..………...………....…. 2
2. LITERATURE REVIEW AND HYPOTHESES DEVELOPMENT………..……. 8
2.1. Executives’ Political Ideologies ………..…….... 8
2.2. Agency Theory & Corporate Political Activity …….…………..………. 11
2.3. Political Action Committees (PACs) ………..……….. 13
2.4. Governance Mechanisms Moderating PACs ……….……... 15
Monitoring Mechanisms: CEO Duality ………..…….… 16
Bonding Mechanism: CEO Compensation Tied to Shareholder Returns ……… 18
2.5. Managerial Risk Aversion & Corporate Political Activity ………..…..…...… 20
2.6. Lobbying ……….………..…….... 22
2.7. The “Revolving Door” ……….………...….. 24
3. RESEARCH DESIGN ………...…… 26
3.1. Data and Sample Selection ……….…………..…………...……. 26
3.2. Methodology ……….………..…….. 27
4. STATISTICAL ANALYSIS ………...………... 30
4.1. Preliminary Analysis & Assumption Testing .…....………...……... 30
5. RESULTS ………..………...………... 32
5.1. Descriptive Statistics ..………... 32
5.2. Regression Results ………...………. 34
6. DISCUSSION ………...……….. 39
7. SUMMARY AND CONCLUDING REMARKS ………..… 45
7.1. Limitations & Future Research ……….……….... 47
- 1 -
ABSTRACT
This paper extends recent research on the influence of executive attributes and behavioral consistencies on corporate political activities. The analysis is based on the premise that political orientations of executives, along with behavioral variances in risk aversity attributable to political preferences, stand in a causal relationship to corporate PAC contributions, lobbying expenditures, and the involvement with ‘revolvers’ of their respective firms. Empirical results were based on a sample of Fortune 500 firms throughout the years 2010-2016. Evidence demonstrates that Republican CEOs make larger donations to corporate Republican PACs, while liberal-leaning CEOs make significantly greater contributions to corporate Democratic PACs. The study further explores how this situation leads to an agency conflict between principals and agents if corporate funds are directed towards political activities that are not increasing firm value. Two moderators were proposed as a solution to the agent-principle dilemma – and have proven to be partially significant. Systematic variation in risk perception demonstrates that risk-averse Republican managers are more likely to choose safer bets by establishing valuable relationships with policy-makers. This firm-level decision manifests itself in higher corporate lobbying expenditures but did not reveal a correlation with the number of revolvers a firm engages with.
Key words: Political Ideology; CPA; Corporate Political Activity; Agency Theory; Behavioral
Consistency Theory; Upper Echelons Theory; Corporate Governance; Lobbying; PAC; Revolving Door
- 2 - ACKNOWLEDGEMENTS
I would like to cordially thank Dr. Pushpika Vishwanathan for her exceptional guidance and enlightening discussions; and the anonymous reviewer for his or her helpful comments throughout the revision process. I gratefully acknowledge support by the University of Amsterdam and its School of Business and Economics.
______________________________________________________________________________
1. INTRODUCTION
Increasing complexity and volatility in the competitive environment have made it challenging for firms to exclusively rely on market-oriented strategies. As a result, during the past two decades, firms have started to increasingly acknowledge the benefits of non-market tactics, including the engagement in corporate political activities. Throughout this study, the terms ‘corporate political activity’ and CPA will be used interchangeably in accordance with the practice of research in this field. These CPA engagements may take shape in corporate
contributions to political action committees (PACs), intensive lobbying, as well as the engagement with so-called ‘revolvers’ - political appointments from the private sector or corporate appointments of former government officials.
The link between firms’ involvement in the political arena and its contribution to firm value has long been part of an ever-evolving discussion in the strategic management and political science literature (for review, see Getz, 1997; Lord, 2000; Vogel, 1996). From a shareholder primacy perspective, a firm’s political engagement may provide economic opportunities in favor of the firm, thereby increasing firm value, whether through increased financial assets or through reduced operational liabilities and risks. Responsible for the endeavor to construct and authorize firms’ political activities in the United States are corporations’ CEOs. Several insightful scholars have pointed out that if we want to understand why organizations do the things they do, or why
- 3 -
they perform the way they do, we must consider the biases and dispositions of their most powerful actors — their top executives (Crossland & Hambrick, 2007; Andrews, 1971; Child, 1972). Considering that managerial attributes have shown to directly impact corporate strategy and firm-level decision-making (e.g. England, 1967), one may suggest that executives’ political ideologies and their inclination towards precarious corporate investment decisions have the potential to affect the direction of a firm’s political strategy. This notion bears a close
resemblance to the theory of behavioral consistency, as proposed by Epstein (1979), Colvin and Funder (1992), which postulates that individuals display constant mentalities across a variety of domains and situations; it holds that peoples’ attitudes are predictive of their behavior (Epstein, 1979).
If political conservatism translates into conservative attitudes and political liberalism translates into liberal attitudes in the corporate domain, executives might be tempted to misuse their firms’ economic power by converting it into political influence in line with their own dogmatic agenda (Hutton, Jiang & Kumar, 2014). While some may argue that corporations purposefully chose CEOs in line with the firms’ ideologies, the covert nature of CPA, exacerbated by low transparency and information asymmetries between managers and shareholders, leads this scholarship to avoid the assumption that owners and managers have similar views regarding CPA (Freed, 2007; Hadani, 2012; Ansolabehere et al., 2003). Hereby, the increased influence that close relations to lawmakers provide may engender additional personal benefits to executives that stand independent from the firms’ economic value and the shareholders’ interests. As a result, an agency problem arises; the cornerstone of agency theory is the notion that the interests of principal and agent diverge (Jensen & Meckling, 1976).
- 4 -
need to understand corporate governance systems, predominantly in the context of financial contributions to political action committees.
In addition, a core difference between identified Republican and Democratic CEOs is personal risk tolerance. Converging lines of evidence (e.g. Schreiber et al., 2013; Kanai et al., 2011; Lieberman, Schreiber & Ochsner, 2003) suggest that biology significantly influences different cognitive processes in brain structure linked to risk-taking behavior. Any strategic decision in the corporate realm has the potential to either increase or decrease the risk of a business in its environment; it is therefore of ultimate importance to explore the link between executives’ political values, varying degrees of executive risk perception, and the potential resulting impact on the non-market strategies of lobbying and collaboration with revolvers.
The majority of prior research on corporate political activity has been categorized into macro-level and micro-level issues. The former concerns the societal and industry perspective; at this key level, a large group of studies (e.g., see Wats & Zimmermann, 1978; Hill et al., 2013; Agrawal & Knoeber, 2001; Brasher & Lowery, 2006; Drope & Hansen, 2006) has addressed the instrumentality and influence of exogeneous factors on firms’ decision to participate in CPA. On the other hand, the micro-level perspective concerns the private governance of CPA (Dahan, Hadani & Schuler, 2013). Mathur, Singh and Thompson (2013) extended research on corporate governance on a micro-level but exclusively emphasized lobbying as a non-market strategy. While a portion of this scholarship examines shareholder control, the micro-level perspective does not only regard private governance challenges, but also the attributes of those steering the firms, their respective CEOs. This interpretation is in part grounded on Hambrick and
Abrahamson’s (1995) understanding of managerial discretion as a micro-level matter
- 5 -
on managerial value traits (Finkelstein & Hambrick, 1997). Although researchers (e.g. Getz, 1997; Vogel, 1996) have defined several corporate motives for political influence, knowledge on executives’ impact on corporate political activities is still relatively scarce.
Several studies have explored the influence of various managerial attributes, such as overconfidence (Malmendier & Tate, 2002), narcissism (Chatterjee & Hambrick, 2007), life and job experiences (Malmendier, Tate & Yan, 2011; Gupta & Govindarajan, 2000), self‐awareness (Wohlers & London, 1989) and military background (Benmelech & Frydman, 2015), to
determine the form and intensity of non-market strategies in the political realm. In addition, the influence of CEOs’ conservatism versus liberalism has been discussed by Chin, Hambrick and Treviño (2013) who, however, exclusively fixated on firms’ corporate social responsibility (CSR) practices. Similarly, Hutton et al. (2014) have embraced research on financial investment decisions and debt levels specific to Republican managers yet failed to consider other partisan orientations. Schreiber et al. (2013) have studied the differences in cognitive and evaluative processes in liberal and conservative managers; however, their research utterly neglected the link to non-market strategies and firm-level decision-making. Notions of risk tolerance and its
interplay with CEOs’ political preference have been virtually absent in CPA research despite its importance in managerial decision-making.
Taking prior literature into consideration, the aim of this research is to offer a venue to discuss an aspect of CPA that management scholarship has little explored so far, the influence of top-managements’ own political ideology and its associated degree of risk aversion on CPA. Hereby, it is central to point out that this scholarship has partially been inspired by Chin et al. (2013) who proposed the influence of executives’ personal ideologies on PAC spending as a future research avenue. However, this scholarship aims to adopt a more holistic angle by
- 6 -
connecting prior literature and extending the study to three distinctive types of corporate political activity; donations to corporate political action committees, corporate lobbying, as well as the number of ‘revolvers’ the firm collaborates with. The revolving door phenomenon in particular has been barely examined in the management field, with the exception of Etzion and Davis (2008) who directed their research towards former government makers becoming board
members. To my knowledge, this is the first study that directly measures political orientation and risk aversity of Fortune 500 executives to help explain three specific types of CPA. This is surprising, given the suggestion that “CEOs can be quite autonomous in their political activities, […] in the corporate hierarchy, CEOs control so many resources that their subordinates are unlikely to object to such behavior” (Hart, 2010, p. 8).
This scholarship makes three theoretical contributions to the corporate political activity literature. First, this scholarship endorses agency theory by extending research on how the political attitudes of agents may impact CPA. The link between managers and firm strategies has fundamental consequences for CPA because firm legitimacy is central to investor relations (Hillman & Hitt, 1999). In addition, this research recognizes that principals may not have ample resources or homogeneous impetus to apply potent corporate governance control systems (Johnson & Greening, 1999). Varying ability to monitor portfolio firms infers that even though institutional ownership at the corporate level can counteract self-serving behavior of executives, the nature of such control mechanisms, as well as principal and agent characteristics, may de
facto encourage it - and potentially contribute to shareholder wealth reduction.
Secondly, this study serves to incorporate the subject of executives’ political values into
upper echelons theory, the notion that demographics and values of top management can
- 7 -
1984). Among the many studies of upper echelons theory (outlined in Finkelstein, Canella & Hambrick, 2009), a limited number of studies has considered the impact of CEOs’ personality (e.g. Peterson et al., 2003) and even fewer have explored the role of CEOs’ values (Simsek et al., 2005). Even within these studies, upper echelons research has emphasized the characteristics of a firm’s top management team (TMT) as a whole (for review, see Lee & Park, 2006; Camelo-Ordaz, Hernandez-Lara, & Valle-Cabrera, 2005; Bantel & Jackson, 1989; Dwyer, Richard, & Chadwick, 2003). Consequently, this scholarship aims to extend and supplement the existing theory by countering the vast focus on macro-level forces and the treatment of firms as a single coalition attending work collectively.
Lastly, this empirical work contributes to a small but growing finance literature on
behavioral consistency (for review, see Epstein, 1979; Colvin & Funder, 1992). Au contraire to
current behavioral consistency research which up until now has largely focused on patterns between corporate financial decisions and managers’ behavioral dispositions in their private lives (Chyz, 2013; Cronqvist, Makhija & Yonker, 2012), this scholarship aims to pave out a new avenue for research by exploring if the concept equally applies to firms’ engagements in the political arena.
The sample analyzed in this study consists of data from Fortune 500 firms between the years 2010 and 2016. The main analysis estimated a number of bootstrap regressions and found a strong causal relation between executives’ political orientation and the direction of PAC
contributions. These relationships did not seem to be moderated by CEO duality; tying executive compensation to shareholder returns, however, did restrain Democratic CEOs from opportunistic behaviors. My results suggest that CEOs’ political ideology directly influences lobbying
- 8 -
A subsequent empirical test also points out that executives’ party ideology did not impact the number of revolvers that the company chose to work with.
In order to efficiently pursue this study’s research quest, the academic work is organized as follows; the subsequent section will discuss CEOs’ personal political ideologies. Next, agency theory, the primary theoretical framework that this research draws from, will be explored;
hereby, I will outline arguments as to why agency theory and CPA are related, and why this is particularly relevant to the study of political action committees. The discussion on agency theory will further lead to a review of bonding and monitoring practices influencing corporate PACs. Furthermore, I will offer a theoretical overview of executive risk-taking behavior and explore its impact on both lobbying and engagements with revolvers. This theoretical development will ultimately give rise to the framework underlying the tested hypotheses. The study’s research design, data collection and methodology are then outlined, leading to the statistical results attained in the empirical analysis. Following a thorough discussion on the outcome of the
empirical testing, I will offer a brief summary and concluding remarks in the final section of this paper. Lastly, a closing sub-section will take into account limitations of this study and offer directions for future research.
2. LITERATURE REVIEW AND HYPOTHESES DEVELOPMENT
2.1. Executives’ Political Ideologies.
Personal values have been systematically studied by several behavioral scientists since the mid-1930s (for review, see Lewin, 1935; Allport, 1937; Murray, 1938) More prominently, England’s (1967) research has provided empirical evidence that personal values directly impact managers’ decision-making and performance behavior, and operate at the level of corporate
- 9 -
strategy. More recent evidence of behavioral consistency in corporate contexts (Epstein, 1979; Funder & Colvin, 1992) has confirmed that executives do in fact translate their personal beliefs into other states of affairs, including corporate policies (Di Giuli & Kostevetsky, 2014). This is in line with Converse (1964, p.3) who described belief systems as a “configuration of ideas and [specified] attitudes” that is predictive of subsequent and interdependent ideas and attitudes. According to upper echelons theory (Hambrick & Mason, 1984; Hambrick, 2007), strategic decision making is not so much an objective venture as an interpretive one, and executives perceive their circumstances and options through highly personalized lenses, which are formed by their values, experiences, and personalities; these, in turn, shape executives’ courses of action (Chin, Hambrick & Treviño, 2013).
Political psychology literature provides sufficient evidence to suggest that CEOs’
political orientations reflect their values (for review, see Rosenberg, 1956; Layman, 1997;
Barnea & Schwartz, 1998; Goren, Federico & Kittilson, 2009); it helps to rationalize why people do what they do. In fact, Erikson and Tedin (1987, p. 64) defined political ideology in terms of values; “the term political ideology is normally defined as a set of attitudes and values about the proper goals of society and how they should be achieved […].” According to Goren and
colleagues (2009, p. 805), particularly in the United States where politics and law are extremely polarized, ‘‘party identification represents the most stable and influential political predisposition in the belief system of ordinary citizens.’’ Poole and Rosenthal (1984) reasoned that in the United States, most political and societal matters can be depicted in liberal-conservative terms, with Democrats claiming primarily liberal positions and Republicans arguing from a largely conservative standpoint - a pattern observed by additional scholars as well (e.g. Diaz-Veizades et
- 10 -
al., 1995; Dunlap & Allen, 1976). Schwartz (1996) concluded that this liberal-conservative spectrum is one of the most substantial for comprehending core beliefs.
Bridging this discussion to firms’ corporate political activities, Ansolabehere, de Figueiredo and Snyder (2003) specified that ’corporate’ contributions are essentially made by individuals, and not by firms themselves. Thus, they believe any sort of monetary contributions or political collaboration can be accounted for as a personal consumption directed to advance a personal ideology, rather than a manifestation of nonpartisan corporate involvement. For instance, executives identifying with the Republican party have been found to enforce and maintain more financially conservative corporate policies (Ansolabehere, de Figueiredo & Snyder, 2003). Those firms have lower levels of corporate debt, lower capital as well as research and development (R&D) expenditures, less risky investments, but higher profitability (Hutton et al., 2014); the firms’ executives also place more emphasis on respect for authority, stability and order (Erikson, Luttbeg & Tedin, 1988; Jost et al., 2003). Conversely, Di Giuli and Kostovetsky (2014) found that Democrat managers are more socially responsible; they are more sensitive to issues as the environment, human rights and social change (Schwartz, 1996). To illustrate, Chin
et al. (2013) examined the influence of managerial conservatism versus liberalism on
organizational outcomes and found that, compared to conservative CEOs, liberal CEOs exhibit greater advances in CSR. Although the authors focused largely on corporate social responsibility practices, the overall study is of fundamental importance as it brings into focus, both for
corporate boards and overall society, the potential importance of executives’ values in shaping the actions and outcomes of large corporations.
The central conclusion of cited academic work is that while firms engage in CPA
- 11 -
manifestations of their personal values and belief systems. Hereby, corporate political activities have the potential to become the personal and political instruments of CEOs. Considering this vast body of evidence, I view CEOs’ political ideology as a determinant of political activities with respect to corporate PAC contributions, strategic lobbying efforts and employment of and/or collaboration with ‘revolvers’ in an effort to promote favorable political and regulatory outcomes, gain and maintain access (Schuler, Rehbein & Cramer, 2002); and overall, to combine these various activities to implement a coherent strategy that matches firms’ resources and objectives within a given political environment (Hart, 2010; Hillman & Hitt, 1999).
2.2. Agency Theory & Corporate Political Activity.
One of the theoretical frameworks underlying this empirical study is Stakeholder-Agency Theory. Agency theory is a dominant paradigm in the financial economics literature and is primarily concerned with the relationship between managers and stockholders (Jensen & Meckling, 1976). It posits that one person (the principal) delegates authority to another person (the agent) to perform a service on his or her behalf (Jensen & Meckling, 1976). Hereby, agency theory assumes that the interests between principal and agent diverge, which leads to a conflict of interest between the two parties. In a utopian corporate world with no agency conflict between principal and agent, and absent other market friction, corporate political activities would lead to higher shareholder value (Mathur et al., 2013). From a shareholder perspective, therefore, corporate political activities should comprise activities that, through potential legislative and regulatory actions, reduce operational risk and increase firm value through increased cashflow (Faccio, 2006). As value-maximizing entities, one expects for executives to engage in CPA as a strategy to favorably influence public policy and create and appropriate firm value.
- 12 -
When executives’ own political ideology is aligned with the objectives of the firm and the interests of shareholders, management conducts optimal political action that has the potential to benefit owners; the risks for principal-agent conflict would be low (Skaife et al., 2013). In contrast, CEOs’ personal political preferences, especially when not aligned with the political agendas of the firms they head, may lead to suboptimal business-level decisions, thereby causing firm value to decrease at the expense of shareholders. The very nature of corporate political activities, characterized by low transparency, exacerbates this issue and raises significant
concerns among institutional investors. As a matter of fact, studies demonstrate that some CEOs decide to engage in corporate political activities even though firm value is not maximized (Skaife et al., 2013).
This key finding confirms that CPA may be a manifestation of executives’ personal convictions – and provides evidence of a potential agency conflict, whereby managers might make the decision to become involved in the political arena to achieve personal utility and pursue their own biased objectives at the detriment of shareholders’ interests. Hereby, it is crucial to specify that, as opposed to lobbying and involvement with revolvers, this scholarship views the agency problem as particularly pertinent to the discussion on political action
committees. This is due to the fact that among most other forms of corporate political activities that often take place largely independent of political party, corporate PAC contributions are administered directly by a CEO with the intention to support selective senate or presidential candidates of either the Republican or Democratic party. Conversely, a professional lobbyist or revolver might be paid by a client or business to meet with legislators and push the client's political agenda; lobbyists and revolvers are hereby merely costly agents who aim to influence Congress and federal agencies and their way of thinking in regard to an issue or specific piece of
- 13 -
legislation - without the personal affiliation to a political party or the objective to support a candidate’s campaign. Consequently, the direction of donations, and the consequential agency problem, will be particularly pertinent for corporate PAC contributions; while the later
discussion on lobbying and revolvers will focus on the partisan-related risk tolerance revolving the decision (and intensity) to engage in CPA.
2.3. Political Action Committees.
Several authors have made an effort to define political action committees (PACs). While campaign contributions from PACs are often portrayed in the media as the functional equivalent of bribes (Milyo, Primo & Groseclose, 2000), a political action committee is an organization that pools campaign contributions from members and donates those funds to campaigns (Janda, Berry, Goldman & Hula, 2014). To be more specific, the OpenSecret database defines political action committees as a popular term for a political committee organized for the purpose of raising and spending money to elect and defeat candidates (“What is a PAC? | OpenSecrets", 2018); hereby, most PACs represent business, labor or ideological interests. Donations distributed by PACs to political campaigns must be voluntary, and managerial employees can legally contribute no more than $5,000 per calendar year to a PAC (Clawson, Neustadl & Scott, 2018). PACs, on the other hand, can give $5,000 to a candidate committee per election, give up to $15,000 annually to any national party committee, and $5,000 annually to any other PAC (Lord, 2000) – making them highly powerful instruments.
Burris (2001) found that campaign contributions by individual capitalists using their own financial capital follow a logic different from those of corporate PACs; firms are generally more interested in buying influence with incumbents; campaign contributions from corporate-affiliated
- 14 -
PACs are hereby considered to be a more direct form of pressure-oriented influence activity than other forms of CPA. Through their affiliated PACs, companies may either reward or deny
financial support to legislators’ election campaign organizations (Stratmann, 2005). The logic for the influence of corporate PAC contributions on legislative decision-making is based on a fairly simple economic rationale. PAC contributing firms have shown to exhibit better stock
performance (Cooper, Gulen & Ovtchinnikov, 2010), while allowing corporations to develop a relationship with politicians (Goldman et al., 2009). Although any explicit benefit would be illegal, PAC contributions are considered to be means by which corporations may purchase favorable public policy, legislative decisions or actions (Ryan, Swanson & Buchholz, 1987). Individuals, on the other hand, are more concerned with boosting the election prospects of favored candidates or policy makers representing their own moral doctrines (Burris, 2001). This distinction between individuals’ and corporate motivations is critical as it provides an indicator for executives’ impetus to steer corporate PAC funds. As a matter of fact, since the 1980s, CEOs have long come to realize that they can translate their firms’ economic power into political influence (Burris, 2001), as top officers of a corporation tend to play a key role in deciding how corporate PAC money will be spent (Clawson, Neustadtl & Scott, 1992). Corporate PAC contributions may therefore become the personal instruments of CEOs to influence politics according to their values.
Consequently, a potential inconsistency between executives’ personal ideologies and the optimum political direction of the firm points towards an agency conflict; corporate PAC contributions allow agency-driven CEOs to financially support government laws that create a power balance in line with their own dispositions, creating benefits for a happy few at the expense of most shareholders (Dahan et al, 2013). Gompers et al. (2003) acknowledge that such
- 15 -
self-serving behaviors may in fact lead to poor corporate value at the cost of owners and are often exacerbated by being performed behind closed doors, making it opaque at best (Butler & Ribstein, 1995; Hillman & Hitt, 1999). PAC contributions in particular are a high-discretion managerial activity that commonly drives abuse and is further aggravated by owners’ inability to limit agents’ opportunistic and self-serving behaviors (Hadani, Dahan & Doh, 2015). Hence, the key challenge for principals is to impose the proper governance structures to prevent the
injection of managerial preference into donations to corporate PACs, as discussed in the succeeding discussion.
Taking prior literature into consideration, I wish to augment existing research by comparing the personal political contributions of CEOs to the PAC spending of the businesses these individuals are aligned with. Assuming that CEOs are motivated to donate to a PAC to get like-minded individuals into office, I assert the following hypothesis:
Hypothesis 1: Executives’ political ideologies will influence firms’ PAC contributions. Hereby,
(1a) Republican CEOs will make significantly higher contributions to PACs known for
supporting Republican candidates, while (1b) Democratic CEOs will make significantly higher contributions to corporate PACs known for supporting Democratic candidates.
2.4. Governance Mechanisms Moderating PACs.
Considering the self-serving influence and potentially hidden agendas of executives’ political ideologies, there is a strong need to explore how executives’ involvement with a
corporate PAC should be governed. Consequently, part of the purpose of this study is to explore the challenges of governance created by the increasing, and often dubious, engagements of
- 16 -
businesses in politics. At the most general level of agency theory, the principal can discourage the divergence from his or her interests through monitoring and bonding mechanisms (Jensen & Meckling, 1976). Monitoring involves observing the behavior and/or the performance of agents. Bonding refers to arrangements that align both parties’ objectives by penalizing agents for acting in ways that violate the interests of principals or rewarding them for achieving principals’ goals (Jensen & Meckling, 1976). The contracts between agents and principals – or management and owners – specify these governance arrangements. These interest alignment mechanisms, or governance structures, do not come without a price; firms must incur monitoring costs intended to limit opportunistic behavior on behalf of the agent (Jensen & Meckling, 1976). The agent, on the other hand, may have to bear bonding costs that guarantee that he or she will not take actions that will harm the principal. Two of these monitoring and bonding systems are CEO duality and tying executive compensation to shareholder returns.
Monitoring Mechanism: CEO Duality. Monitoring practices aim to alleviate the information
asymmetry between managers and stakeholders and allow owners to make judgements regarding the agent’s underlying behavior. Monitoring activity through boards of directors can take many forms; however, the focus of this study is on CEO duality, a formal board structure which arises whenever the same person holds both the CEO and board chairperson positions in a corporation (Rechner & Dalton, 1991). The combination of the positions of CEO and board chairperson creates an ambiguous firm leadership structure that results in contrasting effects that boards must attempt to balance. On one hand, the consolidation of the two most senior management positions establishes a clear direction, unity of command at the top of the firm and faster response to external events; on the other hand, duality can firmly entrench a CEO at the top of an
- 17 -
organization, challenging a board's ability to effectively monitor and discipline (Mallette & Fowler, 1992).
The focus on CEO duality is of key importance because it may allow the CEO him- or herself to directly limit board members’ allocation of attention to monitoring, which may in fact represent the most essential function of a corporate board (Tuggle et al., 2010). To expatiate, the board of directors plays a critical role in the enactment of monitoring mechanisms, being
fiduciarily responsible to monitor and sanction the decisions of the CEO (Moldoveanu & Martin, 2001). CEO duality, however, complicates the execution of this role. This form of managerial entrenchment enables managers to attain so much power that they are able to use the firm to further their own interests rather than the interests of shareholders. In order to obtain this dual stance, CEOs acquire shares within a range that is high enough to give them influence, but sufficiently low to make other shareholders bear the brunt of their non-value maximizing actions (Pan, Yue Wang & Weisbach, 2017). Consequently, more so than any other monitoring
practices, the dual positioning of a manager as both the CEO and chairman may impose a limit to enforcing compliance regulations and fiduciary responsibilities, even in large and prominent companies.
While this scholarship recognizes that CEO duality may not always be dysfunctional, I propose that granting a CEO an influential role on the board of directors may directly act as an impediment to the board’s monitoring abilities of top executives and lead to opportunistic and inefficient behavior that may ultimately reduce shareholder wealth. This proposition follows prior research by Meckling and Jensen (1976) on CEO duality, where the CEO also holds the title of Chair of the Board. Worrell, Nemec and Davidson III (1997) show empirical evidence that when one person obtains dual titles, stock market performance suffers. Considering that the
- 18 -
board is the primary internal control mechanism for aligning the different interests of shareholder and top management (Mizruchi, 1983; Walsh & Seward, 1990), reduction in board control may facilitate the pursuit of CEOs’ agenda, including the realization of political ideologies. Tuggle et
al. (2010) point out that corporate CEOs, with or without duality, generally have incentives to
direct board attention away from monitoring; a dual position as CEO and chairman therefore grants them additional command (Cannella & Lubatkin, 1993) to steer attention away from monitoring and pursue their own (political) interests. Based on this logic, I propose the following hypothesis:
Hypothesis 2: The relationship between CEOs’ political ideology and corporate PAC
contributions will be positively moderated by CEO duality on corporate boards.
Bonding Mechanism: Executive Compensation Tied to Shareholder Returns. The presence of
agency costs implies that owners have an impetus to monitor agents. With that in mind, agents also have a reason to guarantee that they are acting in ways consistent with the shareholders’ interests. In more general terms, a bonding mechanism is an illustration of what Williamson (1985, p. 20) refers to as a “credible commitment,” a concept that requires agents to post a hostage forfeitable upon malperformance. While bonding manifests itself in several ways, including non-financial rewards like promotions and social recognition, executive compensation schemes in particular have long attracted a great deal of attention from financial economists (for review, see Bebchuk & Fried, 2003; Murphy, 1999; Guay, Core & Larcker, 2001) and have consistently been recognized as a prominent factor in the strategic decision-making process. Executive compensation can include cash, stock options, benefits, and perks. In all these cases,
- 19 -
the underlying goal is to create a mutual dependence between CEOs and shareholders so that interests are more closely aligned. Citing Clegg, Hardy and Nord (2013, p. 127), “if agent compensation depends […] upon behaving and performing consistent with a principal’s interest, then – assuming that agents value financial rewards – they will behave in ways consistent with those interests.”
Among the previously cited scholars, the prevalent approach to the study of executive pay arrangements considers compensation as a remedy to the agency conflict between managers and owners. In consonance with agency theory, aligning the incentives of CEOs with the
interests of stockholders via compensation leads to a direct increase in the sensitivity of CEO wealth to stock price (Hall & Liebman, 1998; Jensen & Murphy, 1990); as a result, CEOs will work harder and/or more effectively because they now share gains and losses with stockholders. Consequently, corporate managers tend to be significantly more risk-averse than principals, who are risk neutral. This is based on the logic that, compared to equity investors, managers cannot diversify their labor “investment” (Jensen & Meckling, 1976, p. 134) so as to eliminate firm-specific risk.
Following the free cash flow managerial-shareholder argument made by Jensen (2010), executives whose financial incentives are tied to firm performance are motivated to grow firms beyond the optimal size as growth in sales is positively related to changes in compensation. The dominant view in the strategy literature suggests that firms’ political engagements provide economic opportunities in favor of the firm, increase firm value, and help firms improve their financial bottom line (for review, see Chen, Parsley & Yang, 2010; Cooper, Gulen &
Ovtchinnikov, 2010, Bonardi, Hillman & Keim, 2005; Hillman, Keim, & Schuler, 2004). A recent exhaustive analysis of the political activities and financial performance of 943 S&P 1500
- 20 -
firms over an 11-year period, however, provides evidence that firms’ political investments are negatively associated with market performance (Dahan, Hadani & Schuler, 2013). Other current studies confirm that the relationship between CPA and firm performance is mixed (e.g. Rajwani & Liedong, 2015); it is therefore feasible to suggest that making executives’ financial incentives dependent upon firm value will make CEOs more cautious in their political pursuits. In line with this novel empirical evidence and traditional understandings of agency theory, I argue that executive compensation schemes that are tied to firms’ financial performance will have a weakening effect on the original relationship between CEOs’ political orientation and corporate PAC donations. I therefore develop the following hypothesis:
Hypothesis 3: The relationship between CEOs’ political ideology and corporate PAC
contributions will be negatively moderated by tying executive compensation to shareholder returns.
2.5. Managerial Risk Aversion & Corporate Political Activity.
Repetto (2006) found that CEOs see much of their corporate political activity as defensive efforts to prevent and fend off potential threats and preserve a beneficial context of government law within which they run their business. In compliance with upper echelons and behavioral consistency theory, the attitudes of executives as individuals are predictive of their behavioral dispositions. Accordingly, the second theoretical assertion as to why CEOs’ political ideology is expected to influence corporate political activities is in terms of managerial risk aversion in line with partisan orientation, as supported by Grinstein and Hribar (2004). As
- 21 -
lobby and collaborate with revolvers; as opposed to purposefully steered PAC contributions, lobbying and employment of revolvers is partially about the reinforcement of social capital and relationships between business executives and public officials that allows for the reduction of risks and uncertainty (Blumentritt, 2003) by maintaining beneficial business conditions (Clawson, Neustadtl & Weller, 1998).
Several insightful scholars (Fellows, 2004; Schonberg, Fox & Poldrack, 2011; Slovic, 2011) have defined risk-taking as the propensity to undertake an activity that comprises the uncertain prospect for a valuable outcome, but also the potential for a contrary result. From an evolutionary standpoint, a vast body of evidence (e.g. Adorno, Frenkel-Brunswik, Levinson & Sanford, 1950; Jost, Glaser, Kruglanski & Sulloway, 2003; Bechara, 2001; Vorhold, 2008) suggests that liberals and conservatives exhibit different cognitive styles that bring about and accentuate different psychological characteristics including attitudes towards risk and uncertainty (Webster & Kruglanski, 1994). These variances in brain function can be attributed to different neural processes in brain structures implicated in risk-taking and decision-making; liberals have shown increased gray matter volume in the anterior cingulate cortex, while conservatives have demonstrated increased gray matter volume in the amygdala, both of which lead to different responses in situations of threat and conflict (Jost et al., 2003). Evidence suggests that
Republicans exhibit stronger attitudinal reactions to situations of risk and animosity; in contrast, Democrats tend to seek out “novelty and uncertainty” (Jost et al., 2003, p. 358). Moreover, Kam and Simas’ research (2010) demonstrated that liberals are more risk accepting than
conservatives. Consistent with this argument, Republicans have shown to have more intense physical responses to threatening stimuli than Democrats (Amodio, Jost, Master & Yee, 2007).
- 22 -
If patterns of brain activity during the assessment of potential risk differs significantly between liberals and conservatives, then this is additional evidence for the link between neural processes and political ideologies. In line with this reasoning, Hutton et al. (2014) found that Republican-leaning managers are more likely to adopt financially conservative policies. Furthermore, although there are Republican managers who are known for riskier corporate policies, on average, Republican executives have more conservative corporate policies and are more risk-averse than non-Republican managers (Hutton et al., 2014). Adopting an identification strategy similar to Hong and Kostovetsky (2012), I consequently suggest that executives’
political orientation, amplified by their risk aversion, will translate into their proclivity to engage in risk-reducing non-market strategies at the firm-level, in particular lobbying and the number of
revolvers the firm choses to engage with.
2.6. Lobbying.
Lobbying activities are defined as any direct and explicit attempt of communicating with lawmakers on behalf of a client in order to influence their opinions for a desired legislative goal to the manager’s or firm’s advantage (Mathur & Singh, 2010). Lobbying is the predominant way in which firms engage with the American political system (de Figueiredo & Richter, 2014) and constitutes a major element in the political strategy of firms for securing favorable outcomes; it is therefore perceived as a source of firms’ competitive advantage (Sadrich & Annavarjula, 2005). Existing research has pointed out that corporate lobbying is the most significant channel, in terms of volume and effectiveness, to influence political decisions in the United States (Mathur & Singh, 2010). Compared to PAC contributions, there is no restriction on the amount that firms can spend on lobbying (de Figueiredo & Richter, 2014). While the Federal Election
- 23 -
Campaign Act of 1974 (FECA) states that organizations are not allowed to allocate funds from the corporate treasury to donate directly to politicians, lobbying can be sourced from the corporate treasury controlled by firms’ executives (Fleishman, 1975). As a result, lobbying is a corporate expense that is accounted for, reported and incurred at managerial volition (Mathur & Singh, 2010).
Repetto (2006) pointed out that executives turn to lobbyists because, despite being deeply immersed in their industries, CEOs are often times less knowledgeable about public affairs and the political economy. According to Richter, Samphantharak and Timmons (2009), the main motivation behind lobbying, however, is the potential to persuade politicians, who in turn, may be willing to modify legislative acts, provide tax policy benefits and changes in trade regulations. Other academics have further highlighted that lobbying can persuade policy-makers to modify market structures to a firm’s competitive benefit (Peltzman, 1976; Stigler, 1971) and strengthen relationships between senior management and government officials that sustain promising circumstances for the respective firm (Clawson, Neustadtl & Weller, 1998).
Consequently, several studies concluded that managers enter the political playing field to achieve a reduction in firm risk and improve the prospects of their businesses through legislative and regulatory shifts (Unsal, Hassan & Zirek, 2016). One line of reasoning has suggested that management may engage in lobbying purely as a risk-minimizing strategy (Grinstein & Hribar, 2004). Lord (2000) asserted that by actively attempting to shape public policy, many companies seek to reduce uncertainty and eliminate threats. Similarly, Repetto (2006) maintained that CEOs see much of their lobbying efforts as defensive acts to ward off potential threats. As explored in the discussion on CEOs’ political ideology, Republican managers have been found to enforce reducing corporate policies while Democratic managers are more apt to engage in
risk-- 24 risk--
taking behaviors (Hutton et al., 2014; Di Giuli & Kostovetsky, 2014; Kowert & Hermann, 1997). Taking these findings into consideration, I assert the following hypothesis:
Hypothesis 4: CEOs’ political ideology will influence corporate lobbying expenditures. Hereby,
Republican CEOs will lobby more intensively than Democratic CEOs.
2.7. The “Revolving Door”.
One longstanding focus of research in the United States has been the extent to which the lobbying industry is dominated by the so-called ‘revolving door’ between government and industry (for review, see Blanes i Vidal, Draca, M., & Fons-Rosen., 2012; LaPira & Thomas, 2014, 2017; Luechinger & Moser, 2012; Lucca et al., 2014). The revolving door has traditionally been defined as the phenomenon of political appointments from the private sector as well as corporate appointments of former government officials, including members of Congress, staffers and members of the executive branch, leaving their jobs in government and becoming corporate lobbyists (Lazarus, McKay & Herbel, 2016). Many political scientists and economists view government employees in the lobbying industry as valuable; after all, “Washington is all about connections,” and former government officials can exploit their network of colleagues and friends on behalf of their clients (Blanes i Vidal, Draca & Fons-Rosen, 2012, p.2). Burt (1995), Kale, Singh and Perlmutter (2000) hereby point to the accumulation of human capital, stating that mutual trust, personal links of friendship and politician-specific knowledge allow lobbyists to be more effective in their endeavors. This access to insider information about government operations gives the private sector an unfair advantage in its lobbying efforts (Holman, 2007; Leech, 2013).
- 25 -
Former political appointees bring valuable connections and information about processes, upcoming decisions and competitors to firms hiring them upon leaving government service (Luechinger & Moser, 2012). Favored treatments manifest themselves in several ways: directly in “procurement, regulation and oversight, and supervision of mergers and acquisitions;
indirectly in strategic planning and preferential access to decision makers and information” (Luechinger & Moser, 2012, p. 94). Consequently, similar to strategic lobbying, the political benefits associated with political revolvers can be considered a means for managers to reduce risks in their endeavors. In their book, LaPira and Thomas (2017) point out that the partnership and collaboration with former government employees is a distinct form of enhanced political insurance against the risk that government policies will harm their firms’ interests. Consequently, I propose that the revolving door system in Washington can be thought of as a political economy of risk reduction. The authors further argued that the rise of revolving door lobbyists has allowed former government employees to differentiate themselves in order to serve organizations with specific risk-reduction needs. Revolvers therefore provide executives with strategic process insights that lessen the risks associated with uncertainty in policy making (LaPira & Thomas, 2017).
As previously discussed, a large number of academic studies have provided strong evidence that those who gravitate towards conservative or right-wing ideologies are, in general, less tolerant of uncertainty, fear and environmental ambiguity (e.g., Adorno, Frenkel-Brunswik, Levinson, & Sanford, 1950; Frenkel-Brunswik, 1954) as compared with those who gravitate towards liberal or left-wing ideologies. Based on this academic research, I propose that risk-averse Republican CEOs will show higher incentives to work with revolvers than their Democratic counterparts in an effort to reduce uncertainty as perceived by executives.
- 26 -
Hypothesis 5: CEOs’ political ideology will influence firms’ employment of and/or collaboration
with ‘revolvers’. Hereby, Republican CEOs will engage with a higher number of revolvers than their Democratic counterparts.
Figure 1. Overall Theoretical Framework
3. RESEARCH DESIGN
3.1. Data and Sample Selection.
Based on multiple considerations, this study’s original sample consisted of 635 Fortune 500 firms potentially involved in U.S. American politics during the years 2010, 2012, 2014 and 2016. This time frame was selected on the basis of availability and completeness. The
- 27 -
the country's most valuable businesses as measured by revenue and published annually by Fortune Magazine ("Fortune 500 Companies 2018: Who Made the List", 2018). Data on Fortune 500 corporations have been used in earlier analyses of domestic corporate political activity (for review, see Boies, 1989; Humphries, 1991; Masters & Keim, 1985). However, after accounting for all companies with invalid or insufficient data in the predictor variable, my criteria yielded a final sample of 202 companies.
The data on the independent variable and all three dependent variables - CEOs’ personal contributions to Republican or Democratic political parties, the amount and direction of
corporate PAC contributions, companies’ expenditures on lobbying, as well as the number of revolvers the company engages with – have been obtained from the Center for Responsible Politics’(CRP) OpenSecrets (OpenSecrets.org) database, which collects data from mandatory political activity disclosures filed with the U.S. Senate and executive agencies. In addition, I attained information regarding the two moderator variables, CEOs’ potential position on the board of directors and executive compensation schemes, from the Institutional Shareholder Services (ISS), the world’s leading provider of corporate governance data services. Financial information regarding the chosen control variables firm size and slack resources was obtained from the Compustat database.
3.2. Methodology
IV: CEOs’ political ideology. In the U.S., the liberalism/conservatism distinction has been
associated with the two major political parties, the Democratic Party (more liberal) and the Republican Party (more conservative). In estimating parameters in my empirical model, the political ideology of a firm’s CEO (CPID) is the main independent, or explanatory, variable.
- 28 -
CEOs’ political ideology will be identified by measuring his or her personal contribution to either Republican or Democratic political electives as indicated by the total of donated funds to PACs, political parties, and political candidates. In other words, CEOs’ political liberalism can be measured by examining the degree to which he or she supported the Democratic Party, as opposed to the Republican Party, using data on individual political contributions. Following Hutton et al. (2014), the political ideology of a CEO is measured by:
CPID = 𝐶𝑜𝑛𝑡𝑟𝑖𝑏𝑢𝑡𝑖𝑜𝑛 𝑡𝑜 𝑅𝑒𝑝𝑢𝑏𝑙𝑖𝑐𝑎𝑛 𝑃𝑎𝑟𝑡𝑦−𝐶𝑜𝑛𝑡𝑟𝑖𝑏𝑢𝑡𝑖𝑜𝑛 𝑡𝑜 𝐷𝑒𝑚𝑜𝑐𝑟𝑎𝑡𝑖𝑐 𝑃𝑎𝑟𝑡𝑦
𝑇𝑜𝑡𝑎𝑙 𝐶𝑜𝑛𝑡𝑟𝑖𝑏𝑢𝑡𝑖𝑜𝑛
Hereby, a positive CPID score will be taken as an indicator for a Republican CEO, while a negative CPID score will be interpreted as a Democratic CEO. To reach more conclusive results, donations to independent political parties were excluded from the analysis.
DV1a: Corporate PAC contributions to Republicans. In this empirical study, firm-level PAC
contributions to Republicans is a dependent variable and will be measured by the total amount of corporate campaign contributions to Republican PACs.
DV1b: Corporate PAC contributions to Democrats. In this empirical study, firm-level PAC
contributions to Democrats is a dependent variable and will be measured by the total amount of corporate campaign contributions to Democratic PACs.
DV2: Lobbying Expenditures. Firms’ lobbying spending is a dependent variable and is
measured by the annual lobbying expenditures reported by firms as stated on OpenSecrets.org. This lobbying data is based on lobbying disclosure reports filed with the Secretary of the Senate's Office of Public Records (SOPR). OpenSecrets.org calculates annual lobbying expenditures by adding the four quarterly totals. In cases where both a parent and its subsidiary organizations lobby or hire lobbyists, lobbying spending is attributed to the parent organization ("Lobbying | OpenSecrets", 2018).
- 29 -
DV3: Number of Revolvers. Employment and/or collaboration with revolvers reveals the
relationships between those who represent special interests and those in government who regulate those interests, and acts as a dependent variable. Engagement with revolvers is measured by taking into consideration the total number of revolvers as indicated on the OpenSecrets database ("Revolving Door: Methodology | OpenSecrets", 2018).
M1: CEO Duality. CEO duality, proposed as a monitoring mechanism, is defined by whether or
whether not the CEO simultaneously chairs the board, and/or whether or whether not the
chairman of the board has been the chief executive of the company in the past. Such differences in CEOs’ power relative to that of their boards are measured in a categorical ‘yes’ for CEO duality and ‘no’ for the absence of a dual position.
M2: Executive Compensation Tied to Shareholder Return. Executive compensation, proposed
as a bonding mechanism, is defined by whether or whether not CEOs’ incentive schemes are dependent on shareholder returns and measured in a categorical ‘yes’ for the presence of a tie and ‘no’ for its absence.
In order to ensure internal validity and eliminate potential extraneous factors, this
empirical analysis includes a comprehensive array of control variables; the reasons for which are outlined below.
C1: Firm Size. Firm size has been found to be a critically important factor for driving results in
CPA research. Theory predicted on an economic approach (for review, see Brasher & Lowery, 2006; Hill et al., 2013) proposes that larger firms are more likely to participate politically and to a greater degree as they have greater capacity and incentive to do so (Watts & Zimmermann, 1978), may gain more benefits from strategic lobbying and are more effected by government
- 30 -
regulations (Agrawal & Knoeber, 2001). In other words, greater firm size creates greater political visibility and therefore a greater salience of politics. Therefore, corporations with larger
economies of scale in the production of influence might lobby on more issues or spend more on the matters on which they do lobby (Brasher & Lowery, 2006). For the purpose of this study, firm size is measured by a firm’s annual revenue.
C2: Slack Resources. A firm's excess resources are sometimes referred to as organizational slack
(Bourgeois, 1981; Cyert & March, 1963; Singh, 1986) - the cushion of resources that allows an organization to alter strategies in response to internal or external pressures (Bourgeois, 1981). Political lobbying is expensive (Schuler, 1996); the availability of resources may therefore directly affect a firm's ability to engage in politics (Singh, 1986). Keim and Baysinger (1988) add that securing and maintaining organizational resources are critical to successfully implement political strategies. Yoffie (1987), Lenway and Rehbein (1991) extended these findings; firms that are active in the political process are likely to have sufficient slack resources, and firms lacking slack will take a lesser role. For the aim of this scholarship, slack resources are measured by a firm’s current assets divided by its existing liabilities.
4. STATISTICAL ANALYSIS
4.1. Preliminary Analysis & Assumption Testing.
In accordance with stated methodologies, the political ideology index CPID was computed, specifying the scores of each CEO on the liberalism-conservatism scale throughout the four years. As this study is not of longitudinal nature, the means for all variables were calculated across different years. The binary variable CEO Duality was coded into a dummy variable set to 1 for executives holding a dual position as CEO and board chairman, and 0
- 31 -
otherwise. Similarly, Executive Compensation Tied to Shareholder Returns was coded into a dummy variable set to 1 for compensation tied to shareholder returns, and 0 otherwise. Rare inconclusive values were considered in favor of CEO duality and executive incentive schemes linked to shareholder returns as throughout the four years at least two cases fulfilled the criteria. To achieve substantial results and work with a legitimate focus, only CEOs with distinct partisan scores were included in the analysis. Errors and missing data, particularly in the critical
independent variable, were cleaned; listwise deletion method was applied to attain a complete case analysis and produce candid standard errors that mirror the actual amount of information used. However, in order to account for internal consistency, a second analysis was conducted taking all scores into the analysis and comparable results were attained in all tests.
All predictor and outcome variables are of continuous nature; due to one single predictor variable, the assumption of perfect multicollinearity became impertinent. As the variance of residual terms was constant, heteroscedasticity was of trivial concern. All variables showed a normal distribution; only Lobbying Expenditures and Number of Revolvers were moderately skewed at 2.270 and 2.039, respectively. Applying an inverse transformation failed to alleviate the results; consequently, bootstrapped regression was assessed; bias-corrected and accelerated (BCa) bootstrap interval accounted for abnormally distributed errors and skewness. Stratified sampling was exchanged for a simple sampling approach (Field, 2013). For all other empirical relationships, simple regression was performed to investigate the ability of CEOs’ political ideology to predict diverse forms of CPA. To confirm regression estimates, PROCESS, as per Andrew F. Hayes (2015, p. 2), was established to “estimate and probe a moderation model in which one variable’s nonlinear effect is linearly moderated, or in which one variable’s linear effect is nonlinearly moderated;” in the subsequent analysis it was used to test the interacting
- 32 -
relationship between corresponding variables. To test the moderation effects, CEOs’ political ideology and the control variables were standardized. Lastly, all findings were treated as statistically significant at α = .05.
5. RESULTS
5.1. Descriptive Statistics.
Table 1 provides descriptive statistics and the bivariate correlation analysis of the
variables for the final dataset. The relationships were examined using Pearson’s product moment correlation coefficient, ranging from +1 to -1. First, descriptive statistics are discussed to offer an overview of the dataset. Subsequently, the correlations between the dependent variables and control variables are outlined; the correlations between the dependent variables and the independent variable are then presented. Finally, some of the correlations between the moderating variables and independent variable, and the correlations between the moderating variables and the control variables are discussed briefly.
After accounting for all missing and invalid data, 202 company CEOs were considered in the analysis, out of which all 202 engaged in some form of political donation clearly addressed to either the Republican or Democratic party and/or candidate. In regards to corporate political activity of those executives’ firms, engagement/collaboration with revolvers was the most common – 99% of examined firms stepped foot into the revolving door. Roughly 98.52%, or 199 corporations, have corporate PACs and 97.53% firms have lobbyists or representatives in Washington or retain counsel there. Overall, CEOs scored significantly higher as Republican than Democrat, at an average, positive score of 0.341. Firms’ lobbying expenses resulted in an average figure of USD 2,740,298.70. Over the years 2010-2016, firms donated an average of
- 33 -
around USD 346,858.00 to corporate PACs, out of which roughly half (USD 171,278.48) went to Republicans – over 20% more of what corporate PACs distributed to Democrats. For the number of revolvers firms worked with, data showed an average score of 16 (15.50). Most CEOs held dual positions as chief executive and board member at a mean score of 0.81; equally, most executives’ compensation was tied to shareholder returns at a mean score of 0.83.
In regards to correlations between the dependent variables and control variables, firm size correlated moderately with PAC donations to Republicans (r = .304) and PAC donations to Democrats (r = .241) and showed a strong and positive correlation with lobbying expenditures (r = .582) and number of revolvers (r = .589); slack resources, however, did not proof to have a significant correlation with the outcome variables. As anticipated, the predictor variable CEO political ideology showed a positive and moderate correlation with PAC donations to Republican parties (r = .217) and a negative, and fairly strong correlation with Democratic PACs (r = - .314). Both CEO duality and executive compensation tied to shareholder returns did not correlate with CEO political ideology at r = .051 and r = .062, respectively. Correlations between control variables and moderators did not proof to be significant. For additional information on the diverse variables and the full set of correlations, please refer to Table 1.
- 34 - 5.2. Regression Results.
Tables 2 through 5 present a summary of regression results for firms’ Republican PAC donations, Democratic PAC donations, lobbying expenditures and number of revolvers - respectively. Model 1 displays a base model consisting only of control variables. The
independent variable is added in model 2. Tables 2 and 3 additionally display the moderating effects in model 3 and model 4, respectively, as well as the interaction terms in model 5 and 6. Relationships were tested with robust bootstraps regression, confirmed with PROCESS, and a statistical significance at the 5% level.
Hypothesis 1 suggests a positive relationship between CEOs’ political ideology and corporate PAC contributions, controlled for by firm size and slack resources. With respect to the control variables, results indicate that larger firms have a greater tendency to make PAC
donations to Republican parties at β = .303, p = .000; in contrast, slack resources did not make a significant impact considering the regression coefficient β = .019, p = .823 (Table 2, Model 1). For contributions to PACs known for supporting Republican candidates, the value of the F-ratio is significant at F (3, 128) = 6.287, p = .001, implying a linear relationship between the predictor and outcome variable, and confirming the utility of the model as a whole. The output indicates an R =.358, and because there is only a sole predictor, this value represents the simple correlation between CEO political ideology and corporate PAC contributions to Republicans; this finding has been confirmed by a positive and highly significant Pearson’s correlation coefficient in the descriptive statistics table (Table 1). In addition, the adjusted 𝑅2 = .108 exhibits a shared
variance of almost 11% between the two variables, adjusted for the unstandardized coefficient B and the number of predictors. For Republican PACs, the results denote a positive, moderately strong and statistically significant (β = .189, p = .024) relationship, whereby the standardized
- 35 -
regression coefficient β [ranging from -1 to 1] indicates the gradient of the regression line and the strength of the relationship between the predictor variable and corporate PAC contributions to Republicans (Table 2, Model 2). The bootstrap interval for Republican PACs falls between 1.359 and 14.060 which confirms the significance of the relationship.
With respect to control variables, for donations to Democratic PACs neither firm size nor slack resources proved to have a significant effect at β = .149, p = .090 and β = .042, p = .631, respectively. When including the independent variable, results reveal a highly significant value F (3, 128) = 9.745, p = .000, indicating a regression coefficient deviant from 0 and confirming the validity of the model as a whole. The output further shows an R = .431 and an adjusted 𝑅2 of .186, indicating a shared variance of almost 19%. For Democratic PACs, the results imply a negative and statistically significant (β = - .404, p = .000) relationship (Table 3, Model 2). The bootstrap confidence interval falls between - 17.945 and - 6.717, confirming the likelihood of a truly significant relationship. Thus, Hypothesis 1 is supported.
Hypothesis 2 proposes that CEO duality on corporate boards enhances the relationship between CEOs’ own political ideology and corporate PAC contributions. For donations to Republican parties, the model consisting of CPID and CEO duality is extremely significant without the moderating or interaction effect of the two variables at F (4, 123) = 4.331, p = .007 (Table 2, Model 3); when including the interaction effect of CEO duality, the relationship’s significance sustains at F (5, 122) = 3.468, p = .006. Change statistics, however, indicate no increase or decrease in variation explained by the addition of the interaction term at 𝑅2 Change = .001 and a statistical significance of p = .711. The regression coefficient of the interaction
variable, depicted in Table 2, Model 5, is positive and insignificant at β = .069, p = .297. Out of curiosity, and to confirm these findings, the same tests were performed testing the impact of no
- 36 -
duality held; for donations to Republican parties, the moderating effect is statistically
insignificant at 𝑅2 Change = .001; and β = - .038, p = .711.
Similar results were attained for corporate donations to Democratic parties; both CPID and CEO duality are significant as concurrent variables without and with the interaction effect, respectively, at F (4, 123) = 7.974, p = .000; and F (5, 122) = 6.355, p = .000. The interaction or moderating effect indicates an 𝑅2 Change = .001; the regression coefficient of the interaction variable results in β = .058, p = .740, indicating a statistically insignificant result (Table 3, Model
5). These outcomes have been validated both by a PROCESS analysis and a manual linear
regression analysis. To follow up on these findings, identical tests were performed challenging the impact of no duality held; for donations to Democratic parties the moderating effect is statistically insignificant at 𝑅2 Change = .001, β = - .332 at p = .740. The relationship between CEOs’ ideology and corporate PAC donations maintains strongly significant both in the present, and even in the absence of CEO duality. Overall, the results confirm that CEO duality does not significantly moderate the relationship between CEOs’ political ideology and corporate PAC donations to either political party; thus, hypothesis 2 is not supported.
Hypothesis 3 proposes that linking executive compensation to shareholder return weakens the relationship between CEO’s political ideology and corporate PAC contributions. For PAC donations to Republican parties, Model 4 (Table 2) without the moderating interaction effect is significant at F (4, 120) = 6.629, p = .000; with the interaction term between CEO political ideology and executive compensation tied to shareholder return, the configuration remains significant at F (5, 119) = 5.320, p = .000 (Table 2, Model 6). The fact that the original relationship between CEOs’ political ideology and PAC contributions to Republicans remains statistically significant even after introducing the moderator indicates that the amount of variance
