1
The Importance of CEO Political Ideology in CSR
Practices: Monitoring Incentives of the Board of Directors
Student: Martyna Lubińska
MSc Business Administration – Strategy Amsterdam Business School
Student number: 10418091
Thesis supervisor: Dr. Pushpika Vishwanathan
2
Statement of Originality
This document is written by Student Martyna Zuzanna Lubińska 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.
3
Abstract
This paper highlights the relevance of the CEO political ideologies and corporate governance characteristics in the CSR-CFP relationship. Building on the idea that CEO’s political values influence a firm’s engagement in CSR activities, the study examines the CSR activities of the firm from the corporate governance perspective. Therefore, CEOs’ political ideologies, which are reflections of their values, are specifically investigated in terms of political liberalism vs. conservatism. To test this idea, the study shows an analysis on 350 companies from the list of Fortune500 within the time frame of 2012-2015. The results show that CSR-CFP relationship is not moderated by the ratio of independent directors. This indicates that the monitoring incentives of independent directors are not influenced by the political ideology of the CEO.
Keywords: Corporate Social Responsibility, Corporate governance, Board of Directors, Board Independence, CEO characteristics, Political Ideology, Agency Theory, Upper Echelons Theory
4 Table of Contents Abstract ... 3 1. Introduction ... 5 2. Literature review ... 8 2.1 CSR-CFP relationship ... 9
2.2 Agency Theory and Upper Echelons Theory ... 12
2.3 CEO’s political ideologies ... 14
2.4 The moderating effect of the CG... 17
3. Methodology ... 23
3.1 Sample and Data Sources ... 23
3.2 Dependent Variable ... 25 3.3 Independent Variable ... 26 3.4 Moderating Variable ... 27 3.5 Control Variables ... 31 3.6 Statistical model ... 33 4. Results ... 35 4.1 Descriptive statistics ... 35 4.2 Regression analysis ... 41 5. Discussion ... 43 5.1 Major findings ... 44
5.2 Contribution of this study ... 46
5.3 Limitations and future research ... 48
6. Conclusion ... 50
5
1. Introduction
Multiple studies have taken into account different moderators when investigating the CSP-CFP relationship. However, limited amount of the literature has been dedicated to the characteristics of the top management teams. Hambrick and Mason (1984) were one of the first ones who showed that strategic choices of the executives can be predicted by their characteristics. In their research, they show that the demographic characteristics of the executive can be used as the proxy for his cognitive frames. Moreover, they suggest to use other characteristics relevant for the CEO background, which might add relevance to predictions of the strategic decisions. For that reason, the further research have focused more on the CEO characteristics, which might have an influence on the firm’s strategic decisions (Miller & Shamsie, 2001; Peterson, Smith, Martorana, & Owens, 2003).
Accordingly, Chin, Hambrick and Trevino (2013), examine the CEO characteristics that might have an influence on the CSR-CFP relationship. The authors show in their research that CEOs make the CSR decisions partly based on their political ideologies. According to the authors, liberal CEOs exhibit more CSR initiatives compared to the conservative CEOs. Additionally, their research shows that CEOs’ political ideologies are significantly related to their corporate political action committee (PAC) allocations. These findings raise the question about the CEOs’ decisions, since according to the agency theory the CEO should make decisions on behalf of the shareholders.
The existing research has also explored the characteristics of the board of directors. Specifically, the board composition (Hermalin & Weisbach, 1991), the board size (Yermack, 1996) and the board independence (Harjoto & Jo, 2010). Among these, the board independence has gained particular attention. That is due to the fact that board of directors serves as an internal
6
monitoring device to ensure that the CEO makes the decisions based on what is best for the shareholders. This represents one of the principles of the agency theory, which states that effective monitoring can improve the performance of the firm by reducing the agency costs. Accordingly, Harjoto and Jo (2010) examine if the governance characteristics are positively related to the firm operations. Authors use the conflict-resolution theory in their hypothesis to show that under effective corporate governance, managers use CSR activity to resolve conflicts with various stakeholders, in order to maximize the wealth for the shareholders. Their results confirm the agency theory assumptions and the idea that outside directors reinforce managers to act in the best interest of the shareholders.
However, the previous research shows that CEO injects his biases in the strategic decisions, specifically that the CEO political ideology has an influence on the firm’s engagement in the CSR (Chin et al., 2013). This implies that the board of directors who monitor the CEO’s actions might show different incentives for the CEO with liberal and conservative view. This paper will therefore, focus on the CEO political ideology and corporate governance characteristics, specifically the board independence, as moderators in the relationship between CSR and CFP. It builds on the idea of Chin et al. (2013) that CEO political values have an influence on the company CSR activities and examines the relationship between the characteristics of the board of directors and the CEO political preference. Similarly to the previous research, I develop the idea that CEO’s political ideology – conservatism vs. liberalism has an influence on the firm’s engagement in the CSR activities. Implementing this idea in the agency theory would therefore suggest that the board of directors might have different incentives to monitor the CEO depending on his political preference. On one hand, the board of directors can see more liberal CEOs as unfavorable since they invest in CSR regardless of the company
7
financial performance. On the other hand, if it is actually good for the company performance to invest in the CSR, more conservative CEOs might be seen as unfavorable since they invest less in CSR compared to the liberal CEOs (Chin et al., 2013). Based on this assumption I expect to find that the incentives to monitor, in the form of independent directors on the board, are different for the CEO with more liberal preferences and for the CEO with more conservative preferences. The research question for this paper is as follows:
What is the role of the CEO political preference and the monitoring role of the board of directors in the CSR-CFP relationship?
In order to answer this research question, the literature review examines the CSR-CFP relationship from the agency theory perspective. Afterwards, the moderating effect of the political ideology of the CEO and board independence on the CSR-CFP relationship is examined. The hypotheses are tested based on the quantitative research by applying the collected data containing the top management team’s characteristics and political contributions between 2012 and 2015 for the companies listed on the Fortune 500 list. The CSR is measured based on the CSR annual profile of the firm, derived from the Kinder, Lydenberg and Domini (KLD) database. The financial performance is measured by four different proxies (ROA, ROE, ROI, EBITA). The information about the financial performance indicators is retrieved from the COMPUTAT database and 10-K firm’s reports. The moderating variable, CEO political ideology, is constructed based on the political donations made to the Democratic or Republican Party. Data is obtained from the U.S. Federal Election Comission (FEC). The independence of the board of directors is expressed as the ratio of the independent directors on the board to the total number of the board members. This data is retrieved from the ESG database. In addition, there are several control variables used in this study. These include the industry type, firm size,
8
board size, number of non-executive members on the board, corporate governance score, presence of corporate governance committee and CEO duality.
The results of this paper show that the political view, specifically political conservatism vs. liberalism of the CEO, do not reflect the company engagement in the CSR initiatives. Furthermore, it reveals that the monitoring incentives of the independent directors do not change according to the political ideology of the CEO. This shows the important contribution to the agency and upper echelons theory. Based on both theories, I compute predictions about the monitoring incentives of the board of directors. The findings show that there is no empirical evidence that one of the theories is dominating in explaining the firm’s incentives to engage in CSR activities.
The paper has the following structure. First, the literature review briefly discusses the relevant academic literature and theories, which have been described till now in the field of CSR and corporate governance. The section ends with highlighting the hypotheses constructed in this research. Following part includes description of how the study will be executed and it presents the results of this study. The last part provides the conclusion with an answer to the research question. It also contains suggestions for further research as well as the limitations of this paper.
2. Literature review
This section gives an overview on the CSR-CFP relationship as well as presents the contrasting views on the social responsibilities of a business. Next, it provides deeper theory on corporate governance related moderators, specifically the top managers characteristics. These are then divided between the CEO and board of directors characteristics.
9 2.1 CSR-CFP relationship
The interest in the corporate social responsibility has its beginnings in 1970 when the academics in the CSR field showed that there exist a relationship between the CSR and CFP. Since then, the development of CSR has been defined based on the two contrasting approaches. The first economic approach was introduced by Friedman (1970), who claims that the only social responsibility of organization is to increase its profits. According to the author, investments in CSR are considered as theft and political subversion. Moreover, he states that it is the role of the state to address the social problems. This field of CSR, which supports the Friedman’s view, represents the shareholder approach to social responsibility and argues in favor of a negative association between the CSR and CFP. This negative relationship reflects the assumption of an agency problem or a conflict between the interests of managers and shareholders. Freidman argues that managers use CSR as a means to further their own social, political, or career agendas, at the expense of shareholders. According to this view, resources devoted to CSR would be more wisely spent, from a social perspective, on increasing firm efficiency.
This theory has been tested by Wright and Ferris (1997). Authors examine the impact of the public announcements of divestment of South African operations on the stock return of publicly traded firms in relation to the principal-agent relationship. Their results show that announcement of corporate divestment of South African business units are associated with significant negative excess returns. It is reflected by the fact that managers decided to divest the business units due to the pressure of society and perceived loss of their public image. This proves that the actions of senior managers might be motivated by their self-interest and noneconomic forces. These results show the contrary principle to the traditional shareholder theory which says that top managers act in the best interest of the shareholders. Instead, it shows that top managers
10
act based on their own interest, which has negative consequences for the shareholders. It is therefore, the typical illustration of the agency conflict between the shareholders and executives. In contrast to the shareholder view, Freeman (1984) supports the stakeholder view, which states that all stakeholders should be treated as an important factor in determining the objective of the firm. According to him, firms that recognize and take into consideration the interests of the relevant stakeholders will perform better than the other companies. The field of CSR which supports this view suggests that there is a positive relationship between the CSP-CFP(Margolis & Walsh 2003; Margolis, Elfenbein, & Walsh 2009; Orlitzky, Schmidt, & Rynes, 2003). Paper of Barnett and Salomon (2012) brings an important contribution to this is relationship. In their study, authors argue that the positive relation depends on how well the firm capitalize on its social responsibility efforts. They build their arguments on the concept of SIC, which was developed by Barnett (2007), defined as “the ability of the firm to identify, act on, and profit from opportunities to improve stakeholder relationships through CSR” (p. 803). According to his view, the stakeholder response to their actions varies with SIC, therefore the firms with differing levels of SIC will receive different returns to social responsibility. Based on this argument, Barnett and Salomon conclude that there is a U-shaped relationship between CSP and CFP, which explains why firms with low CSP have higher CFP than firms with moderate CSP. These findings support Barnett’s (2007) argument that as firms engage in social responsibility, they accumulate stakeholder influence capacity that improves their ability to transform social investment into financial returns.
The alternative perspective which describes the social aspect of the stakeholder theory of the CSR-CFP relationship is introduced by the McWilliams and Siegel (2001). In their article, authors try to answer the question about how much firm should spend on CSR. The research is
11
based on the theory of the firm, which assumes that firms make optimal choices to maximize their profit. According to the authors, the investment in CSR activities should be treated just as any other type of the investment. Based on that, authors derive the supply and demand on the CSR attributes, which can be determined by the cost-benefit analysis. Using this framework allows to identify the level of CSR investments that maximizes profit and at the same time satisfy the stakeholder demand for CSR. That is because based on the demand and the cost of providing the attribute, the firm chooses how much of the attribute it will use, in order to maximize the firm performance, conditional to the assumption that it will maximize the shareholder wealth. This theory therefore, presents the neutral relationship between the CSR activity and firm financial performance.
The paper of Margolis and Walsh (2003) explains the tension in the CSR-CFP relationship due to the existence of two problems. First concern is about the misappropriation of the corporate resources by the firm owners or employees. Second concern is the misallocation of the firm resources by using them for a poorly suited purpose. According to the authors the companies should engage in CSR “by doing what they do best: employing workforce to provide goods and services to create a marketplace, and by doing so fulfilling the people’s needs and creating wealth”(p. 272). In this view, authors show that the shareholder view of maximizing profits is consistent with the interest of people in well-being. The paper offers a new starting point for organization theory and research. It says that instead of finding a link between CSR-CFP we need to understand the conditions under with firm’s efforts benefit society. This posts the questions to the CSR activity and competing theories of the firm.
This continuing discussion have raised questions about the CSR initiatives and self-interest actions. That is because on one hand, the CSR initiatives reflect the stakeholder theory,
12
according to which the company invest in CSR in order to improve its relations with the stakeholders, which can improve its performance. On the other hand, however, engagement in the CSR might be a reflection of the self-interest of the individual who aims to obtain his own social, political, or career agenda, at the expense of shareholders. For that reason, increasing amount of literature investigates the mechanisms underlying the CSR-CFP relationship, however the limited studies examine the moderators related to the corporate governance. There is a significant amount of the research which examines the influence of CEO experience on the decisions he makes (Miller & Shamsie, 2001; Westphal & Zając, 2013). However, the largely missing is the attention to how the CEO’s political ideology might shape the CSR initiatives and what influence does it have on the board of directors. Additionally, aside of the executives’ role in shaping the organizations’ strategy, there is an issue of effective corporate governance practices. Harjoto and Jo (2011) examine this relation by investigating the impact of the internal and external governance systems on the firm’s CSR engagement. The authors define this relationship as the “governance-CSR nexus”. Authors propose four hypotheses in order to explain why there is an increasing trend of companies engaging in CSR and how does the CSR relate to the company performance. They use the conflict-resolution explanation to show that effective corporate governance enforces managers to act in the interest of the shareholders and therefore, predict that those firms will more actively engage in CSR. In their results, authors show that when managers engage in the CSR to solve the conflicts with the stakeholders, it results in the better performance of the firm.
2.2 Agency Theory and Upper Echelons Theory
The agency theory informs us about the relationship between managers and stakeholders. It is based on the four assumptions – bounded rationality, self-interest, information asymmetry and
13
higher risk aversion of the agents (Barney & Hesterly, 1996). According to it, the relationship occurs whenever one member (the principal) of transaction delegates to another member (the agent) the task, which involves decision-making authority to the agent. The assumption of the theory is that the interest of the agent and principal is different. In order to prevent it, the principal has to incur the agency costs, which cover the monitoring costs and incentives for the agent. Agency theory is therefore, concerned with resolving the problem that exist in the agency relationships when there is divergence of goals between the principal and agent (Fama & Jensen, 1983). In order to lower the agency problems and the costs of monitoring, large organizations have separated the decision management and decision control between the CEO and the board of directors. The task of the CEO is to make the decisions about the operations and management of a business on the behalf of the shareholders. However, due to the bounded rationality assumption, it is expected that CEO might not always be rational when making the decisions about the firm strategy.Instead, such decision might be madee with the rule of thumb, indicating that the executive based his decision on the cognitive maps, which capture the past success, or experience (Levinthal & March, 1993). Therefore, it is the role of the board of directors to monitor the CEO’s actions and make sure that his decisions reflect the interest of the shareholders.
The assumption of bounded rationality is also present in the upper echelons theory of Hambrick and Mason (1984). In their research authors showed that top management characteristics influence the firm’s outcomes. In particular, they focused on the CEO values and cognition, which have an impact on the strategic choices and performance of the company. According to the writers, the executive in the situation of making the strategic choice for the organization, make his decision based on the cognitive base and values. This shows that strategic
14
decision making is more of an intuitive than a technical decision process, which is affected by the individual values. The upper echelons theory informs us that CEOs’ values, experiences and personalities substantially affect their interpretations of the situations and options they face, which again, influence their choices (Hambrick, 2007).
Both theories are therefore, based on the bounded rationality assumption. However, the upper echelons theory differs from the agency theory. The agency theory explains the different business performance between the companies due to the different governance arrangements. While the upper echelons theory explains the differences between the company performance and cognitive skills and characteristics of the top management. More precisely, it suggests that leaders can greatly vary in their decisions due to the injection of their personal biases. For that reason, the characteristics of executives need to be taken into account when building an understanding why some companies act different than others (Hambrick, 2007).
Hambrick and Mason (1984) showed that the demographic characteristics can be used as the proxy for the cognitive frames of the CEO. In their research they suggest to use other characteristics relevant for the CEO background, which might add relevance to predictions of the strategic decisions. Accordingly, the recent research have focused more attention on the CEO characteristics, which might have an influence on the firm’s strategic decisions (Miller & Shamsie, 2001; Peterson, Smith, Martorana, & Owens, 2003). However, there is still a limited amount of studies which focus in particular on the personal values of the CEO’s that might have direct influence on their choices.
2.3 CEO’s political ideologies
Building on the findings of Hambrick and Mason (1984), recent research has focused on the political ideologies of the CEOs (Chin et al., 2013). The authors based their results on the
15
political science and political psychology literature, which shows that the majority of individuals will base their core values according to their political ideologies (Poole & Rosenthal, 1984; Schwartz 1996; Knight, 2006; Jost, Federico, & Napier, 2009). Indeed, according to the definition of Tedin (1987), political ideology can be described as “an interrelated set of attitudes and values about the proper goals of society and how they should be achieved ” (p. 65). In discovering the relationship between political ideologies and personal values, it was Freedman (2003), who pointed out on the fact that the individual values are indeed a reflection of the political ideology.
Among many of the political dimensions, the political ideology can be divided between the two most contrary dimensions – liberalism vs. conservatism. This division can be illustrated with two major political parties – Democratic Party, which support more liberal beliefs and the Republican party, which support more conservative view. According to Joost (2006), this continuum has been identified as most useful in defining the political views. That is because these two parties differ from each other significantly with the values they represent. Accordingly, individuals with liberal ideologies will more in favor of social justice, environment friendly policy (Jost et al., 2003; Joost, 2006). In contrast, conservatives will vote for free markets, property rights and individualism (Murtha & Lenway, 1994; Roe, 2003; Detomasi, 2008).
Based on these findings, Chin et al. (2013) constructed the political ideology as political liberalism vs. conservatism and demonstrated that CEOs make the CSR decisions partly based on their political ideologies. The authors show that the CEO who is more liberal has different effect on the CSR than the conservative CEO. Precisely, liberal CEOs when compared to the conservative ones, exhibit more CSR initiatives. Additionally, their research shows that CEO political ideologies are significantly related to their corporate political action committee (PAC)
16
allocations. With respect to the company performance, the researchers also showed that more liberal CEOs, in comparison to conservative ones, are less dependent on recent performance of the firm, which results in engagement in the CSR activity regardless of the financial performance. This is based on the argument that, since politically liberal CEOs believe that CSR matters to the bottom line, they are predicted to be relatively insensitive to recent financial performance when encouraging CSR (Chin et al., 2013). Thus, liberal CEOs will promote CSR irrespective of how their firms are performing, as they view it as an integral part for improving business (Chin et al., 2013). Politically conservative CEOs, in comparison, whose personal values do not necessarily support CSR outcomes, tend to promote CSR only when they think that they can afford it (Chin et al., 2013).
Similar research which builds on the upper echelons theory, was conducted by Gupta and Wowak (2016), who show that the CEO remuneration is partially determined by the political ideologies of board of directors. Based on the similar arguments regarding the political ideology, they found that the board members who are more conservative tend to believe that the CEO actions have a strong influence on the company performance which translates to his higher remuneration. Additionally, their findings show that the difference between the CEO pay granted by conservative and liberal boards of directors is only significant when the performance of the company has increased as well. In their paper, authors show that the conservative boards pay their CEO 18 percent more than the liberal board (Gupta & Wowak, 2016, p. 22), concluding that the conservative boards tend to react stronger to the pay-for-performance principle of the agency theory.
Based on the upper echelons theory and findings of Chin et al. (2013) it is therefore, expected that political ideology of the executives, which reflects the individual values, has an
17
influence on the company pursuit of CSR. However, at the same time there are various reasons why CEO might invest in the CSR. According to the stakeholder theory, the engagement in the CSR improves the firm performance, thus decision of the CEO to invest in the CSR strategy might simply reflect acting along the value maximization theory. While according to the agency theory the CEO pursuits the CSR activities in order to pursuit his own social, political or career agenda, at the expense of the shareholders. Therefore, with the current emergence of companies being green, which makes the business to pay more attention to the CSR activities, it becomes a question if the individual values of the CEO reflected in the political preference has an influence on the CSR-CFP relationship. In order to measure it, first it is proposed that:
Hypothesis 1 (H1): There is a positive relationship between the company’s CSR activity and
its financial performance.
Next in order to examine the effect of the CEO political ideology on CSR-CFP relationship, the moderating effect is tested. Due to the differences in the characteristics between liberal and conservative CEO, with politically liberal CEOs showing more CSR initiatives when compared to the conservative ones (Chin et al., 2013), I expect that the CEO ideology has a reflection in the CSR initiatives. Thus:
Hypothesis 2 (H2): The CSP-CFP relationship is positively moderated by the CEO’s
political liberalism.
2.4 The moderating effect of the CG
Although there is a significant amount of research concerning the social responsibility of the business, there is still no clear explanation on how the board structure influence the CSR-CFP relationship in firms with CEOs with different political ideologies.
18
According to the theory of ownership (Hansmann, 1988) the optimal assignment of ownership belongs to the party involved in the firm that minimizes the costs of market contracting with non owners and the costs of ownership with owners. However, in publicly listed corporations, the ownership is dispersed among the general public in many shares of stock which are freely traded on the stock exchange. Therefore, one of the main features of these firms is the separation of the ownership and control, which comes back to the principal-agent problem (Fama & Jensen, 1983). As a result, the core governance challenge of the PLFs is to ensure that managers and executives take the decisions in line with the shareholders’ interest. In order for the principal to ensure that the actor is acting in the best interest of the company, he has to incur the agency costs. These costs include costs for structuring, monitoring and bonding a set of contracts among agents with conflicting interests. One of the tools used for controlling the agency problems is the board of directors (Fama & Jensen, 1983).
The board of directors is most often described as the formal link between the shareholders of the company and the managers who run its day-to-day activities (Mintzberg, 1983). Consistent with the notion of separation of ownership and control, the board of directors serves as a control system of the firm’s decisions. Therefore, they are responsible for the giving adivse on the strategic decisions and monitoring of the decisions (Fama & Jensen, 1983). The first one, advisory role, refers to the resources that individual directors bring to corporate boards which are largely a function of their human and social capital and refer to the directors' expertise, knowledge, skills and social networks (Lester, Hillman, Zardkoohi, & Cannella, 2008). In their advisory role, directors use their resources and skills to provide advice and counsel to the executive management and helps with the formulation of the business strategy.
19
of the shareholders. This task is performed by the independent directors, sometimes called outside directors. An independent member of the board is the one who is not related to the any senior managers as well as has no relationship with the company besides his compensation fee. The independent directors are chosen due to the fact that they are not directly engaged into the company business activity. Therefore, they are expected to perform their job with more objective view and taking the shareholder view as the main intention. Since it is the task of the independent directors to perform the monitoring role, the board effectiveness is mainly measured by the link between the proportion of outside directors and performance of the company (Bhagat & Black, 2000).
This is confirmed by the agency theory, since due to the separation of the ownership and control, the independent directors are believed to effectively help the company to achieve the superior performance. This notion has been researched by a wide amount of academics. Baysinger and Butler (1985) showed that across U.S. companies, the higher proportion of outside directors corresponds with higher return on equity. Similar research has been done by Ezzamel and Watson (1993), on the sample with U.K. corporations. Authors found that the amount of the outside directors is positively related to the firm’s profitability. The positive relationship has been also confirmed by Zahra and Pearce (1992), who show that the board size and number of outside directors have a positive influence on the firm’s financial performance.
While there is a substantial amount of the literature dedicated to the CSP-CFP relationship and to the corporate governance influence on the company financial performance, there is relatively small amount of the research on how corporate governance affects the CSP-CFP relationship. According to the agency theory, the board of directors serves as a monitoring mechanism to protect the interests of shareholders. Consequently, higher number of independent
20
directors on the board results in the better performance of the company due to the wealth maximization. The argument which suggests that independent directors on the board represent the effective corporate governance mechanism, which reinforce managers to act in the best interest of the shareholders can be supported by the research of Harjoto and Jo (2011). According to the authors the company CSR activity is positively related to its governance characteristics including board independence, institutional ownership and analyst following. Authors use the definition of the “governance - CSR nexus” to explain how different governance systems engage in the CSR. They use the conflict-resolution theory in their hypothesis to show that under effective corporate governance, managers use CSR activity to resolve conflicts with various stakeholders, in order to maximize the wealth for the shareholders. Based on that, the authors predict that firms with effective corporate governance, represented as independent directors on the board, will more actively engage in CSR. These results confirm the agency theory and idea that outside directors reinforce managers to act in the best interest of the shareholders.
Since the effectiveness of the board of directors is measured by the proportion of independent directors on the board, they might have a different response to the liberal or conservative CEO, in the context of monitoring. That is due to the differences in the CSR initiatives between more liberal and conservative CEOs (Chin et al., 2013). When analyzing the CEOs’ characteristics, the liberal CEOs exert more CSR initiatives and are less dependent on the financial performance of the firm. Consequently, the board of directors can see more liberal CEOs as unfavorable. For that reason, it can be expected that the monitoring incentives for the these directors are higher compared to the conservative directors. Therefore, the board of directors will have more independent directors if the CEO has more liberal preferences. This might also work the other way around. Since conservative CEOs exert less CSR initiatives
21
regardless of its influence on the financial performance, the board of directors can see more conservative CEOs as unfavorable.
Following this reasoning, in the context of the CSR-CFP relationship, the board of directors might show different incentives for the monitoring. According to Chin et al. (2013), CEOs with conservative ideological preferences are more sensitive to the financial performance of the firm. Additionally, their results show that these CEOs view the CSR activities as an attribute that should be taken only if the financial situation of the firm allows it. According to the authors the conservative CEOs decide to engage in CSR only if there is substantial amount of the resources, which can be used for that. Given the fact that there is a significant amount of the research which shows the positive relationship on CSR-CFP (Margolis & Walsh 2003; Margolis, Elfenbein, & Walsh 2009; Orlitzky, Schmidt, & Rynes, 2003), this fact shows that it might be conservative CEOs who are seen as unfavorable. That is because if it is actually good for the company performance to invest in the CSR, more conservative CEOs might be seen as unfavorable since they invest less in CSR compared to the liberal CEOs (Chin et al., 2013). Accordingly, the board of directors might have higher monitoring incentives for the conservative CEOs. As a result, the board of directors will have more independent directors if the CEO has more conservative preferences.
This shows that CEOs political ideologies might have an interesting implications for the corporate governance practices. If the CEO injects his personal values into the CSR activities, that might have various consequences for the monitoring incentives of the board of directors. On one hand the board of directors can see more liberal CEOs as unfavorable since they invest in CSR regardless of the company financial performance. For that reason, the board might have higher incentives to monitor the actions of liberal CEO and have a higher ratio of independent
22
directors on the board. On the other hand, if it is actually good for the company performance to invest in the CSR, more conservative CEOs might be seen as unfavorable since they invest less in CSR compared to the liberal CEOs (Chin et al., 2013). Similarly, the board might have higher incentives to monitor the actions of conservative CEO and have a higher ratio of independent directors on the board. Based on the existing theories and developed arguments it is not determined which side should be chosen. Therefore, two alternative hypotheses are created and tested. In order to measure the difference between the monitoring incentives for the CEO political ideology, the sample is divided into two separate samples with liberal and conservative CEOs. Hypotheses 3a and 3b test if the board of directors views the more liberal CEO as unfavorable. Hence:
Hypothesis 3a (H3a): In firms with CEOs with liberal preferences the CSP-CFP relationship
is positively moderated by board independence.
Hypothesis 3b (H3b): In firms with CEOs with liberal preferences the CSP-CFP relationship
is negatively moderated by board independence.
Hypotheses 4a and 4b test if board of directors views conservative CEO as unfavorable. Thus:
Hypothesis 4a (H4a): In firms with CEOs with conservative preferences the CSP-CFP
relationship is positively moderated by board independence.
Hypothesis 4b (H4b): In firms with CEOs with conservative preferences the CSP-CFP
23
3. Methodology
The primary goal of this paper is to measure the CSR-CFP relationship. Additionally, the research investigates how the political ideology of the CEO and independence of the board members moderates the CSR-CFP relationship. To test the moderating effect of the political ideology on this relationship, the sample with the all CEOs is constructed. However, in order to test the moderating effect of the board independence, there are two separate samples consisting of the liberal and conservative CEOs.
The analysis consists of two steps. Fist, in order to measure the CSR-CFP relationship I run four regressions with four different measures of the financial performance of the company. Next, I choose the proxy which shows the most significant results and check the moderating effect in order to test hypothesis 2. Afterwards, I divide the sample according to the political ideology of the CEO and run the regressions to test hypothesis 3 and 4. Thus, in the second step there are 3 additional regressions. Furthermore, the collected data about the companies was constructed as a cross-sectional data for four years. Therefore, the average was calculated for each variable. The reason for that is the usage of data for one year across four years might show more random results. Instead, taking the average will result in more stable outcome since it avoids the single events that might influence the data across one year.
3.1 Sample and Data Sources
Several different sources where used in order to collect the data. First, to obtain the data about the CSR activities, the Kinder, Lydenberg and Domini, & Co. (KLD) database is used. The KLD database, measures annually CSR profile of the individual firms (described below). As already mentioned before, the average CSR score of four years is used in this paper. The information about the financial performance indicators is retrieved from the COMPUTAT database and 10-K
24
firm’s reports.
The focus of this paper are the companies listed on the Fortune 500 list during the period 2012-2015. The Fortune 500 list is an annual list published by Fortune Magazine and consists of the 500 largest U.S. companies listed by total revenue. The list contains both public and private companies, with publicly available revenue data. The timeline was chosen due to House of Representatives elections, which are held every two years. Therefore, all companies that were on the list of Fortune500 within the timeframe 2012-2015 are the part of the study. In total there are 593 companies selected. The companies that changed their status from private to public or other way around in that period are still part of the study. The same applies to the companies, which were acquired. Therefore, as an example, if the company was acquired in 2014, I collected the lobbying expenditures of this company from 2012, 2013 and the lobbying expenditures from 2014, 2015 of the company that acquired this company. The companies with the missing values for the CSR chosen categories and for the financial indicators were deleted from the research. As a result there are 350 companies across 48 industries.
As described further, the political ideology of the CEO was measured based on the executives’ political donations and PAC contributions for the years 2012-2015. This methodology is consistent with Chin et al. (2013) and Gupta and Wowak (2016) approach.Two sources were used to collect the data about the CEOs and political contributions. The first source is the website of Fortune500, which provides information about the CEOs. The second source is the Center for Responsive Politics (www.opensecrets.org), which contains information about the federal campaign contribution, lobbying data and movements between government institutions and private sector. To provide a clear relation in the data sample, there is always only one CEO per financial year. If there was a change of the CEO in the company, the executive with a longer
25
tenure within a year was chosen. That is due to the assumption that longer tenure year gives more information about donation of given individual.
In addition, I will collect data on CG variables such as the board size, corporate governance score, number of non-executive members on the board, monitoring and CEO duality. This data is retrieved from the ESG database, which contains information about the global companies concerning the environmental, social and governance issues. Regarding the corporate governance practices, the ESD provides analysis on the governance structure and corporate behavior, which helps to identify the issues of the excessive pay, corporate misconduct and accounting irregularities.1
3.2 Dependent Variable
Financial performance. In this research, the dependent variable is the financial performance of
the firm. There are multiple proxies used to determine the financial performance. In this model, firm financial performance is gauged by four accounting-based measures: Return on Equity (ROE), Return on Assets (ROA), Return on Investments (ROI) and Earnings before interest, taxes and amortization (EBITA).
The first proxy used for company performance is Return on Equity. It measures the profitability of the firm by looking at the revenue which company generates with the money invested by shareholders. It is calculated as a ratio of net income and shareholder equity. The next proxy is the Return on Assets, which is an easy way to show how well management is using assets to generate the profit. Return on assets is an indicator of how well company is doing relative to its total assets. It is supported by Chin et al. (2013) and Gupta and Wowak (2016) to use both of the measures for the assessment of the financial performance. Moreover, these two
1
26
financial measures are repeatedly used in the history of the research of CSR-CFP relationship. In his research Griffin and Mahon (1997) showed that ROE and ROA are one of the most widely used financial measures in the research.
Additionally, Return on Investment and EBITA are used as determinants of the financial performance. Although, they are not the common measures used in the CSR-CFP literature, they belong to the traditional financial performance indicators. ROI is calculated by subtracting the cost from the total income and dividing it by the total cost. It shows how much money was made on the investment as a percentage of the purchase price. EBITA is calculated as the sum of the earnings before tax, interest expense and amortization expense. It is used to gauge the operating profitability of the company.
3.3 Independent Variable
Corporate social performance. In order to measure the company’s CSR profile, the KLD ratings
are used. The KLD data, constructed by the financial advisory firm Kinder, Lydenberg, Domini, and Company, have been frequently used and considered as one of the best database available for measuring CSR (Griffin & Mahon, 1997; Chin et al., 2013; Choi & Wang, 2009). The KLD ranking is constructed by the independent analysts, who use multiple sources in order to construct the firms’ annual CSR score based on eight categories (e.g., human rights, employees, products, environment). Based on the categories, every year, KLD rates all companies from the Standard and Poors500, based on the presence or absence of the specific strengths and concerns (using a binary indicator).
In this paper, in order to measure the CSR score, there are six categories included: community relations, corporate governance, employee relations, environment, human rights and product quality. After deleting the components with the missing data there are 8 strengths and 18
27
concerns for each category. Although some academics have focused on the specific strengths or concerns, this research analyses the CSR overall. Therefore, the CSR score was calculated as the sum of all the strengths minus all the concerns. This methodology is consistent with Chin et al. (2013) and Gupta and Wowak (2016) approach. Following their arguments, this approach is based on the fact that more liberal CEO will be more in favor of social justice, environment friendly policy.
Despite the wide use of the KLD rating, there are some concerns about subjectivity of the KLD database as well as masking the industry effect (Chiu & Sharfman, 2011). Additionally, the index does not have a weighting scheme for the different dimensions of CSR, which results in all dimensions treated equally (Waddock & Graves, 1994).
3.4 Moderating Variable
CEO political ideology. The political ideology in the US can be divided between the two major
parties – Democratic Party, which supports more liberal beliefs and the Republican Party, which supports a more conservative view. At the same time, these political parties reflect different political ideologies. Accordingly, democrats supporting liberal beliefs, and republicans supporting more conservative beliefs (Poole & Rosenthal, 1984; Goren, Federico, & Kittilson, 2009). Moreover, since the distinction has been made between these parties, it was found that the division between them has been increasing across the years (McCarty, Poole, & Rosenthal, 2006; Hetherington, 2009). For that reason, the previous research uses the information about donations made to the one party over another, as a reflection of the individuals’ ideology (Chin et al., 2013). The corporate political investments, which include funding campaigns or PAC contributions, are made by the company for a various strategic reasons (Hillman & Hitt, 1999). One of the main findings in the political science research shows, that the companies often engage
28
in different forms of CPA in order to support the policy that will have a positive effect on the financial performance of the firm (Hillman, Keim, & Schuler, 2004). At the same time, the political contributions on the individual level have been identified with the individuals’ willingness to support the particular ideology (Ansolabehere, de Figueiredo, & Snyder, 2003; Francia, Green, Herrnson, Powell, & Wilcox, 2005). This supports the argument that the contributions towards Democratic or Republican party, might reflect the political ideology of the CEO. This is also consistent with the prior research, which shows that the CEOs’ ideologies might be examined by the comparison of the donations to the two major parties in the U.S – Democratic Party and Republican Party (Chin et al., 2013; Gupta & Wowak, 2016; Christensen, Dhaliwal, Boivie, & Graffin, 2015; Hutton, Jiang, & Kumar, 2014).
In order to measure the political ideology of the CEO, the data was collected about the CEO’s contributions to the representatives of Republican and Democratic party as well as total value of PAC contributions. This was supported by the previous research, which showed that the CEO’s political preferences are significantly related to their corporate political action committee allocations (Chin et al., 2013) and CEO’s contributions (Mathur, Singh, Thompson, & Nejadmalayeri, 2013). The data is collected for the year 2012 and 2014. The reason for that is that 2012 is the only year which provides the results from presidential election. This data was retrieved from the Center for Responsive Politics (www.opensecrets.org), which was built by an independent group of researches in order to track the money used in politics for federal campaign contributions and lobbying. Center for Responsive Politics is also used to show the spending patterns of the major industries, companies, as well as individuals, for every federal candidate. All data was examined very carefully. In order to determine the right person as the CEO, his first and last name was checked, as well as the state where the firm is headquartered and the name of
29
the employer. When the right individual was identified, all the contributions to Democrats were added and all the contributions to the Republicans were added. The amount donated to neither of the parties was identified as unknown.
Based on the information on political contributions, including the CEO’s contributions and PAC allocations, the political liberalism index was constructed. It was calculated as the sum of the CEO’s contributions and PAC contributions to Democrats divided by the amount of donations to both parties. This financial commitment is constructed based on the study of Chin et al. (2013). Following their methodology, in order to deal with the zero values, 0.1 was added to all numerators and 0.2 to all denominators. The scores varied between the values 0 and 1. In order to construct the political liberalism index, the values below 0.50 represented conservatism, and values above 0.50 indicated liberalism. The CEO score is calculated respectively:
Political liberalism in dollar value of contributions
=
29,295+0.173,800+0.2
= 0.397
The calculations show that the CEO’s political liberalism score is 0.397, or relatively liberal. The distribution of the score is represented on the figure 1 below. In the sample with 350 companies, the average score of the political liberalism is equal to 0.44. The data is left-skewed, which shows that on average the CEOs are tended towards conservatism. The mid-point of the scale represents around 29 percent of the sample. On the conservative side there is about 37 percent of the CEOs and around 33 percent is on the liberal side. This results are consistent with the Chin et al. (2013) analysis, which showed that the sample on average tends towards conservative side.
30 Figure 1. Distribution of CEO's political liberalism scores.
Board members independence. Independent board members are the outside directors, who are
not affiliated with the top management of the firm and have no business dealings, except the sitting fees. According to the literature in the corporate governance field, the presence of independent directors on the board serves for a purpose of aligning the interest of the managers and the shareholders. Therefore, the independent directors serve as a monitoring device of the board of directors and are considered to be an important element of the practice of effective corporate governance (Byrd & Hickman, 1992; Hermalin & Weisbach, 1998). According to the authors, the effectiveness of the directors depends on the on board’s independence from the management.
In order to gain the information about the independent board members, the data is retrieved from the ESG database. The value assigned to each company represents the percentage of the independent board members. Following the research of Harjoto and Jo (2011), the independent directors serve as a measure of the effective corporate governance since they are
0 5 10 15 20 25 30 35 0 0.2 0.4 0.6 0.8 1 Per ce n t
31
used as a tool for the internal monitoring.
3.5 Control Variables
The control variables used in this research can be divided into two categories, the company factors and the corporate governance characteristics. The first category includes the firm size and type of the industry sector. That is due to the fact that both factors affect the CSR activity and financial performance (Ullman, 1985). Using them as control variables has been supported in the previous studies in the CSR field (Waddock & Graves, 1997; Brammer & Millington, 2008). The firm size variable is expressed as natural logarithm in order to adjust for standard normal distribution. The second type of control variables used in the model refer to the corporate governance conditions. The following variables are used - board size, corporate governance score, CEO duality, number of non-executive members on the board and monitoring.
Industry. To test if the certain industries might have an influence on the financial performance,
the dummy variables are included for the following sectors: financial, consumer goods and utilities. The reason for that is the fact the CSR-CFP relationship depends on the type of the goods or services sold (Siegel & Vitaliano, 2007). That is due to the fact that each sectors can gain different benefits from the CSR activity. Additionally, the companies from financial sectors often have a higher financial performance when compared to other industries. To account for the industry effect, each company is categorized based on three mentioned sectors based on the four-digit SIC code. This results in creation of three dummy variables (financials, utilities, consumer goods) for companies which belong to one of these industries. Thus, the companies which do not belong to any of these categories are labeled with 0 while the companies which belong to one of these categories are labeled with 1.
32
Size is considered to be an important control variable, since bigger firms might show more evidence for the socially responsible behavior. The reason for that is, that they affect more stakeholders and therefore, they need to be more transparent with their actions. At the same time, it makes them a more attractive target for the environmental activists (Kourula & Halme, 2008). Moreover, bigger firms operate on the bigger scale and interact with multiple external parties. Therefore, they have more extensive network of stakeholders, compared to the smaller firms (Burke & Logsdon, 1986).
Board size. The size of the board is represented by the total number of the directors on the board.
It is often used as an important characteristic, which is likely to have an important influence on the board functioning. As already mentioned, one of the roles of the board of directors is monitoring, which prevents the actions of self-interested managers. When size of the board increases, there is also more opportunity for the free-riding problems due to which the monitoring services become less effective (Raheja, 2005; Harris & Raviv, 2007). Therefore, bigger board size implies that there should be higher incentives for monitoring.
Corporate governance score. The corporate governance score is retrieved from the ESG
database. It measures the company's systems and processes, with regard to the adoption of corporate governance practices, which reflect action in the interest of the shareholders. In other words, the score represents how well the company is using the management practices.The scale of the corporate governance score varies from 0 till 100, with 0 representing the lowest score, and 100 representing the highest score.
CEO duality. The CEO duality refers to the situation when the CEO is at the same time the
chairman of the board. According to the research of Zając and Westphal (1996), the fact that the CEO is also a chairman of the board, account for the higher amount of the board power, which
33
may result in the CEO ability to manipulate the corporate strategies, including the CSR. In the model it is represented by the dummy variable labeled with 1 when the CEO was also a board chair and labeled with 0 if he was not.
Non-executive board members. Non-executive directors or external director are the members of
the board of directors, who are not the part of the executive management team. At the same time, they are not employees of the firm and therefore, are not involved in the day-to-day management of the company. They are chosen due to their independence and their task is to provide the objective advice about the firm strategy. The information about non-executive directors is retrieved from the ESG database, and it represents the percentage of the non-executive directors relatively to the board size.
Monitoring. Monitoring refers to the ability of the company to monitor the board functions
through the establishment of a corporate governance committee. The committee is responsible for developing the set of corporate governance rules, including the code of conduct and ethics. Additionally, it helps to establish the remuneration policy for the directors. In the model it is represented by the dummy variable that is coded as 1 when there is a corporate governance committee within the firm and as 0 if there is no committee.
3.6 Statistical model
The data used in this study include multiple observations across the time period 2012 – 2015. However, using the records for each year might show more random results, due to the fact that a single even might influence the data across one year. In order to avoid that and obtain a more stable data, the average of each variable is calculated, which results in creation of cross-sectional data for four years. In order to test the hypothesis, the ordinary least square model is used. The financial performance variables are transformed to the logarithms to limit the influence of the
34
outliers. Furthermore, due to the heteroscedasticity in the data, the robust standard errors are used. To prevent multicollinearity, all the variables are mean-centered before testing the moderating effect. In order to examine if there is a moderating effect, first the CSR-CFP relationship is examined. Four different financial performance indicators were regressed on the CSR score of the firm. Next, the indicators which showed the significant results are used in the further analysis. After testing the moderating effect of the political ideology of the CEO, I examine the moderating effect of the board independence. In order to make a distinction between the monitoring incentives towards liberal and conservative CEOs, the sample has been divided into two separate samples. First one with firms with liberal CEOs and second with firms with conservative CEOs. This can be illustrated on the model below.
Figure 2: Whole sample including all the firms
35 Figure 3: Sample divided into two separate samples with firms with liberal and conservative CEOs
4. Results
This section presents the main findings of this research and has the following structure. First, the descriptive statistics of the variables are presented to provide an overview of the data. Second, a correlation analysis is showed to study the relationship between the variables. Afterwards, the results of the direct impact of CSR on CFP are presented. Third, the section presents the moderating impact of the CEO political ideology and the independence of the board of directors on the CSR-CFP relationship.
4.1 Descriptive statistics
Table 1 presents descriptive statistics and the bivariate correlation analysis for all variables used in this research. The correlation between the variables is measured by the Pearson correlation coefficient. In the table, first the correlation between the dependent and independent variables is presented. Afterwards, the relation between the dependent and moderating variables as well as independent and moderating variables is outlined. Finally, the correlation between the dependent and control variables is analyzed.
36
Across the different type of industries, there are 53 companies from the financial industry, 8 from consumer goods industry and 27 from utilities. The remaining 262 companies did not belong to any of the controlled sectors. The proxies used for measuring financial performance as well as the size of the company are all expressed as logarithms.
When analyzing the descriptive statistics, we observe that the correlation between the corporate social responsibility and financial performance varies depending on the proxy used. The results show that there is a positive and significant correlation between the CSR and ROA and between the CSR and ROI (r = 0.149 and r = 154 respectively, p < 0.01). In the table, it can further seen that the correlation between the CSR and EBITA shows significant but negative correlation (r = - 0.214, p < 0.05), while the CSR and ROE has no significant correlation. These results are consistent with the research, which finds that there is a positive correlation between the CSR and financial performance. The surprising result is the significant but negative correlation between CSR and EBITA. This suggests that in contrast to the predictions and existing research the additional investment in CSR results in lower EBITA.
Regarding the correlation between the financial performance and moderating variables, none of the variables show significant correlation. Also, almost all of the correlations are negative. The correlation between the ROA and political ideology of CEO as well as the correlation between ROI and the political ideology both negative and not significant (r = - 0.006 and r = - 0.05 respectively). At the same time, both proxies have the negative correlation with the number of independent board members (r = - 0.059 and r = - 0.056 respectively). The correlation between the EBITA and CEO political ideology is also low and negative (r = - 0.067). The only positive correlation is between EBITA and number of independent board members on the board (r = 0.069). It is also considered to be the highest correlation between the dependent and
37
moderating variable. In addition, there is no significant correlation between the CSR and both moderating variables, showing the negative correlation with CEO political ideology (r = - 0.047) and positive correlation with independent board members (r = 0.076). This shows that in contrast to the predictions from the upper echelons theory the political ideology of the CEO is not correlated to his strategic decisions. Furthermore, lack of significance in the correlation with the independent board members shows that there is no support for the agency theory explaining the monitoring incentives of the board.
Considering the correlation between the financial performance and control variables, we can observe that only few variables show significant correlation. ROA and EBITA have significant correlation for financial industry (r = - 0.138 and r= 0.205 respectively). Also ROA and ROI show significant but negative correlation with utilities industry (r = - 0.144 and r= 0.121 respectively). Moreover, there is a significant correlation between ROA and EBITA and number of non-executive directors (r = - 0.108 and r = 0.170 respectively). Additionally, EBITA has significant and positive correlation with sales and board size (r = 0.413 and r = 0.146 respectively, p < 0.01). In the comparison to all financial proxies, EBITA shows significant and positive correlation with most of the control variables.
41 4.2 Regression analysis
Table 2 presents a summary of the hierarchical multiple regression, which was performed to investigate the relationship between the CSR and different proxies of the financial performance of the firm. Models 1,3,5,7 contain only the control variables. Models 2,4,6,8 test the main effect of CSR on the financial performance, holding all other variables constant.
In the regression with ROE and EBITA used as a proxy, the results show no significance for the CSR-CFP relationship. For that reason, these results are not discussed in the paper. In the regression with ROA and ROI used as a proxy, the results show significance for the CSR-CFP relationship. For the model 1 with ROA, the model is statistically significant F (11, 338) = 3.908 (p < 0.001) and explained 11.3% of variance in ROA. After entry of the CSR, the total variance explained by the model as a whole was 12.4%, F (14, 335) = 3.387 (p < 0.001). Therefore, the introduction of the CSR explained only additional 1.1%, after controlling for all the control variables (R2 Change = 0.011; F (3, 335) = 1.425. As shown in model 2, the relationship between CSR and ROA is significant (p < 0.05). The coefficient of CSR has a value of 0.05 which indicates that for every unit increase in CSR, a 0.05% increase in ROA is predicted, holding other variables constant. Additionally, in the model 2 with ROA, there are three control variables, which are statistically significant, out of which sales shows the highest Beta value (β = 0.619, p < 0.05). Other variables which show significance for the model are financial and utilities industry (p < 0.05).
Model 5 and 6 show the regression results for the ROI. In the model 5, which includes only control variables, the model is statistically significant F (11, 338) = 3.026 (p < 0.001) and explains 9% of variance in ROI. After entry of the CSR and moderating variables in Model 6, the total variance explained by the model as a whole is 10.7% F (14, 335) = 2.864 (p < 0.001). The
