The relationship between executive pay disparity and corporate social performance, and the impact of the financial crisis on this relationship

(1)

Julie van Lanschot | 10833757

MSc Business Administration: Strategy Final thesis

Supervisor: Dr. P. Vishwanathan

The relationship between executive pay disparity and corporate

social performance, and the impact of the financial crisis on this

(2)

Statement of originality

This document is written by Julie van Lanschot, 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)

List of figures

Figure 1. Conceptual model ... 18

List of tables

Table 1. Descriptive statistics and correlations ... 27

Table 2. Results of regression analysis ... 28

Table 3. Results of regression analysis with moderator ... 29

(4)

Abstract

Due to an increased awareness of the importance of corporate social performance (CSP) investments for firms, this study aims to provide more insights by investigating the relationship between pay disparity within the top management team (TMT) and CSP. It is argued that high disparity pay structures are reflective of individualistic firms fostering a shareholder-oriented view, conducive to lower levels of CSP. In contrast, low pay disparity structures are reflective of cooperative firms fostering a stakeholder-oriented view, conducive to higher levels of CSP. In addition, the current study extends existing research by examining whether the financial crisis has an impact on the aforementioned relationship. Findings show that executive pay disparities do not influence firms’ CSP investments. Furthermore, results partially support the assumption that the financial crisis influences the executive pay disparity – CSP relationship. Contributions to executive compensation and CSP research are discussed, and limitations and related suggestions for future research are given.

Keywords: Corporate social performance; executive compensation; tournament theory;

(5)

1. Introduction

In today’s business environment firms are under increasing pressure to engage in socially responsible activities because of emerging standards related to social performance and the fact that social performance ratings and evaluations are becoming more transparent to society (Deckop, Merriman & Gupta, 2006). The way a firm handles corporate social responsibility (CSR) initiatives in action is termed corporate social performance (CSP), which focuses on managing the interests of multiple stakeholders and creating value for them (Carrol, 1991; Goa, 2009; Wood, 2010). A firm’s CSP is important as it can influence the behavior of consumers (Sen & Bhattacharya, 2001), sell-side analysts and investors, and the firm itself (Chatterji & Toffel, 2010). Moreover, engaging in CSP activities can lead to long-term value and competitive advantages (Orlitzky, 2011).

An increasing number of firms have set themselves the aim of maximizing stakeholder value as primary goal of the firm rather than focusing solely on their shareholders (Bradley, Schipani, Sundaraman & Walsh, 1999). Researchers have investigated how executives can use this stakeholder perspective to manage their firm more effectively to increase long-term success (Freeman, 1984). This concept is impacted by a number of organizational policies, in which executive pay structures play a critical role (Hart, David, Shao, Fox & Westermann-Behaylo, 2015). The way pay structures motivate executives’ behavior and decision-making has been researched continuously in the field of strategy (Devers, Cannella, Reilly, & Yoder, 2007). Research shows that pay structures are used as an important channel through which corporate governance can incentivize executives to engage in CSP activities (Hong, Li & Minor, 2014).

Previous work examining the relationship between executive compensation and CSP has found mixed results when focusing solely on the amount or type of pay of the CEO

(6)

findings, researchers began examining other aspects of pay that might impact executives’ behaviors (e.g. Hart et al., 2015), and found that differences in pay among members of the broader top management team (TMT) also play an important role in motivating executives (Frederickson, Davis-Blake & Sanders, 2010). Accordingly, in order to successfully manage the interdependent (Rowley, 1997) and complex needs (Frooman, 1999) of all stakeholders it is incumbent that the TMT works together to accomplish this (Wong, Ormiston & Tetlock, 2011). Thus, when examining the executive compensation-CSP relationship it is necessary to look at the broader TMT and not just the CEO. In line with previous work (Hart et al., 2015), this study therefore focuses on the distribution of pay within the TMT (including the CEO), to investigate whether executive pay disparity fosters a stakeholder-orientation or shareholder-orientation conducive to respectively higher or lower levels of CSP.

Even though certain compensation structures are used as incentives to influence executives’ behavior towards increasing CSP, it is evident that desperate times can call for desperate measures. A desperate time that has gotten more attention in the field of strategy over the pas years is the financial crisis, which has put to test many firms and industries. In order to cope with the crisis, firms needed to change or redefine their business objectives (Porter & Kramer, 2002; Fernández & Souto, 2009). For many firms this meant a change in both executive compensation structures and CSP investments (Karaibrahimoglu, 2010). Taking this into account, the current study examines whether the financial crisis influences the relationship between pay disparity within the TMT and the overall CSP of a firm. This research makes several important contributions to both existing literature and practice. First, the current study contributes to the little research that has been done on the role that the pay distribution (rather than solely the level of pay) of the broader TMT (rather than solely the CEO) plays in CSP investments (e.g. Hart et al., 2015). In addition, this study focuses on all executives within the TMT rather than the five highest paid executives, which

(7)

was considered a limitation in previous studies. By doing so, results provide more reliable and generalizable insights for literature discussing the consequences of pay disparities within the TMT. Third, it extends existing studies by taking the financial crisis into account when examining the executive pay disparity – CSP relationship. It is interesting to reinvestigate this relationship in bad times, such as the financial crisis, to see how firms respond to the crisis regarding executive compensation and CSP, and how this influences the relationship. Fourth, by analyzing the relationship between executive pay disparity and CSP within the framework of the tournament theory (Lazear & Rosen, 1981) and equity theory (Adams, 1963), findings of the current study add to the legitimacy of the aforementioned theories and provide more understanding for the executive pay disparity – CSP relationship in both good and bad times. Lastly, results of this study regarding the subsequent compensation incentives for CSP can help firms in creating efficient pay structures for their TMT. This research provides an in-depth understanding of executive pay disparity – CSP relationship, and allows for more insights by including the financial crisis into this relationship. The research question is:

To what extend does executive pay disparity influence CSP, and how does the financial crisis influence this relationship?

The structure of this paper is as follows. First, a literature review explains the constructs of the current study and their expected relationships based on detailed theoretical background and findings of previous research. Second, the method section describes and explains the collection, measurements and analysis of the data used for the database research. After that, the results of the research are presented and discussed. The paper concludes with a discussion and conclusion of the findings, including contributions, limitations and directions for future

(8)

2. Literature review

For many years researchers have studied the importance and benefits of CSP. CSP is often described as a set of descriptive categorizations of business activity that focuses on the impact and outcomes for society, stakeholders and the firm itself (Wood, 1991; Wood; 2010). It concerns the harms and the benefits that result from the interaction of a company with its larger environment, including the social, cultural, legal, political, economic and natural dimensions (Wood, 2010). CSP includes CSR responsiveness and socially beneficial activities of businesses, and is necessary for a company in order to appropriately adapt to its environments (Carrol, 1991; Wood, 2010). Investment in CSP improves relationships with stakeholders, which results in increasing satisfaction of stakeholders and an improved overall performance (Waddock & Graves, 1997).

In the field of strategy two types of firms are distinguished: shareholder-oriented firms and stakeholder-oriented firms. The first type of firms have set themselves the aim to benefit one stakeholder group, the shareholders, whereas the latter is structured to benefit all stakeholder groups (Donaldson & Preston, 1995). The goal of shareholder-oriented firms is primarily maximizing shareholder value (Sundaram & Inkpen, 2004). These firms are associated with a contract-based, transactional set of calculative and individualistic relationships (Hart et al., 2015). Stakeholder-oriented firms seek to find a balance between the interests of all stakeholders (Freeman, 1984), and are associated with a relational, communitarian, reciprocal and collaborative TMT (Hart et al., 2015). Some researchers have argued that the stakeholder-orientation will pass while the shareholder-orientation of a firm must fail, as it merely focuses on the interests of shareholders (e.g. Bradley et al., 1999). They find the shareholder-orientation a normatively unacceptable expression of an immoral monism of interests at the expense of a moral pluralism of interests (Donaldson & Preston, 1995). Recent work shows a shift in focus from firms with a shareholder-orientation to a

(9)

more stakeholder view, where the interests of all stakeholders are taken into consideration by the TMT (Adams, Licht & Sagiv, 2011; Lorsch & MacIver, 1989).

The on-going debate on differences between the shareholder and orientation has led to the discussion whether shareholder-oriented firms or stakeholder-oriented firms can achieve higher levels of CSP (Hart et al., 2015; Jones & Felps, 2013; Sundaram & Inkpen, 2004). It can be argued that shareholder-oriented firms generate high levels of CSP as all stakeholders can benefit from improved financial performance (Orlitzky, et al., 2003; Sundaram & Inkpen, 2004). However, due to the fact that their relationships with stakeholders are more transactional (Brickson, 2007), shareholder-oriented firms will “allocate just enough resources to stakeholders to induce contributions needed for shareholder wealth maximization (McWilliams & Siegel, 2001) and prevent those stakeholders from exiting their relationships with the firm (Coff, 1999)” (Hart et al., 2015, p.6). More favorable arguments can be made for stakeholder-oriented firms. Firms with a stakeholder-orientation focus more on fostering the development of long-term, mutually beneficial relationships amongst all stakeholder groups and the firm, and seek to maximize the return of stakeholders’ investments rather than the bare minimum required (Brickson, 2007; Coff, 1999). Studies have shown that firms that create and maintain greater reciprocity within and outside the firm, take the interests and needs of all stakeholders into consideration, which in turn will lead to higher levels of CSP in relation to firms that focus solely on shareholders (Bosse, Phillips & Harrison, 2009; Hart et al., 2015).

Whether firms execute and enact a shareholder or stakeholder-orientation, lies within the decision-making process of the executives (Hambrick & Mason, 1984). They are authorized to make decisions regarding the allocation of managerial, financial and other resources to stakeholders (Hambrick & Mason, 1984; Hill & Jones, 1992). The question rises

(10)

behaviors (Devers et al., 2007). Studies show that corporate governance can influence firm’s social performance by using compensation as clear mechanism to motivate executives (Hong, Li & Minor, 2014). Research examining executive compensation as incentives for CSP, focused almost exclusively on the amount and type of CEO pay and found mixed results (Devers et al., 2007; McGuire, Dow & Argheyd, 2003). Both long-term and short-term pay amounts resulted in a diversity of results (negative, positive and mixed) in relation to CSP, making conclusive statements on this relationship impossible (Deckop et al., 2006; Manner, 2010; McGuire et al., 2003). Hart et al. (2015) try to explain these inconclusive findings by arguing that scholars might have “inherently adopted an agency theory perspective that attempts to provide the right mix of incentives to align the interests of owners with managers.” (Hart et al., 2015, p.7). This shareholder-orientated view has led them to ignore the fact that other aspects of pay can serve to align the TMT more closely with stakeholders, thereby failing to bring implications for CSP for other stakeholders to the surface.

In order to bring more clarity and to help illuminate these inconclusive previous results, researchers investigated other aspects of pay that might impact attitudes and behaviors towards stakeholders (e.g. Hart et al., 2015). Results show that not only the amount or type of pay, but also differences in pay among members of the TMT play an important role in motivating executives (Frederickson, Davis-Blake & Sanders, 2010; Henderson & Fredrickson, 2001). When investigating these differences, a distinction is made between vertical pay disparity and horizontal pay disparity. In this study, vertical pay disparity is regarded as the disparity in pay between the CEO and the rest of the TMT (Henderson & Frederickson, 2001), whereas horizontal pay disparity is the disparity of pay amongst the non-CEO executives (Frederickson et al., 2010).

(11)

2.2 Vertical pay disparity and CSP

Vertical pay disparity is often discussed together with the tournament theory, which argues that executives are motivated to work hard because they are striving to win successive tournaments (Lazear & Rosen, 1981). In line with the tournament theory, it can be said that the aforementioned pyramid compensation structure can create an environment within firms in which employees are motivated to, and rewarded for, outperforming their fellow employees (Hart et al., 2015). Scholars argue that tournament structures for executive pay compensation increases individual work effort (Siegel & Hambrick, 2005), and discourages shirking and free riding (Gibbons & Murphy, 1990; Hambrick, 1995), which effectively aligns the interests of managers with shareholders (Hart et al., 2015). Research on financial incentives shows that employees are indeed motivated to put forth greater effort in high disparity pay structures (Lazear & Rosen, 1981). Jones and Felps (2013) claim that this type of research conducts a shareholder-oriented view, which focuses on maximizing financial profits to create the most total value that will ultimately benefit all stakeholders.

Although tournament structures have been theorized to increase work effort, researchers have also found negative consequences such as aggressive and competitive behavior (Dye, 1984). In addition, vertical pay disparity contributes to a hierarchical distance that can create perceptual and substantive barriers between levels, which in turn can reduce efficient information processing and coordination of effort (Miles & Snow, 1978; Eisenhardt & Bourgeois, 1988). Furthermore, Cowherd and Levine (1992) argue that great pay disparity between lower-level employees and senior management can provoke feelings of inequality and relative deprivation, resulting in a decrease of employee effort and cooperation.

Research focusing on high pay disparity within groups such as the TMT (including the CEO), also found negative effects on behaviors and attitudes of individuals (Harbring &

(12)

can encourage employees to take reckless, selfish or other subversive actions to make themselves appear and perform better than their fellow employees (Eisenhardt & Bourgeois, 1988) even if these actions hurt the firm. Within the TMT this could lead to actions such as preventing fellow executives from obtaining important information, manipulating information for individual benefit, and secretly trying to influence decision-makers (Eisenhardt & Bourgeois III, 1988; Milgrom & Robert, 1988). In addition, when members of the TMT (non-CEO executives) believe that the CEO has been awarded compensation that he/she does not deserve or if the CEO’s pay excessively exceeds their own compensation, they may experience relative deprivation (Crosby, 1976) and feelings of resentment towards the CEO and possibly the firm itself (Siegel & Hambrick, 2005). This can lead to non-CEO executives becoming less willing to collaborate with others in accomplishing their designated tasks (Siegel & Hambrick, 2005).

High vertical pay disparity structures foster individualistic (Brickson, 2007), self-regarding (Bridoux & Stoelhorst, 2014), and competitive behavior (Bradley et al., 1999), which is linked with a profit-maximizing, shareholder-orientation. This shareholder-oriented view can negatively affect the behaviors from the TMT, subsequently leading to potential harmful consequences for CSP (Hart et al., 2015). This can be explained by the fact that the subversive and destructive behaviors of executives make it difficult to balance competing executive claims (Sundaram & Inkpen, 2004), and prevent them from exhibiting stakeholder-centric attitudes to manage complex stakeholder issues (Rowley, 1997; Wong et al., 2011). It is therefore expected that firms with high vertical pay disparity adopt a shareholder-orientated view rather than a stakeholder-shareholder-orientated view, ultimately resulting in lower levels of CSP. The first hypothesis is therefore:

(13)

2.3 Horizontal pay disparity and CSP

Previous research examining the role that financial incentives play in motivating executives, has found that both vertical and horizontal pay disparity can affect executives’ behavior (Frederickson, Davis-Blake & Sanders, 2010; Siegel & Hambrick, 2005). In other words, the behaviors of non-CEO executives are influenced not only by perceptions of pay fairness as compared with the CEO but also with their fellow non-CEO executives (from now on ‘executives’) (Hart et al., 2015). Horizontal pay comparisons are often explained on the basis of the equity theory (Adams, 1963), which argues that individuals compare their input-to-output ratio with others whom they view as similar to themselves (Gupta, Conroy & Delery, 2012). When looking at the TMT, the input-to-output ratio consists of the inputs that executives put into their work (e.g. effort, time, etc.) and the outputs they receive from that work (e.g. salary, bonuses, etc.). Due to individual’s intrinsic need to evaluate own abilities and outputs (Festinger, 1954), it is expected that executives will engage in a series of (social) comparisons to determine whether the rewards for their performance are justified and fair (Adams, 1965). Executives will experience equity when they feel that their input-to-output ratio is in balance with other executives’ ratios. However, inequity exists when executives perceive that their input-to-output ratio is worse than other similar executives, resulting in executives altering their behavior and attitudes (Frederickson et al., 2010).

Scholars argue that individuals in general, and executives in particular, tend to overestimate the value of their own contributions and underestimate and/or diminish the value of the contributions of others (Frederickson et al., 2010; Kruger & Dunning, 1999). In this way, executives typically perceive themselves as similar to their fellow peers regarding superficial and observable characteristics, and superior to them regarding less observable but more meaningful contributions to the firm (Hiller & Hambrick, 2005).

(14)

Horizontal pay disparity has proven to have both beneficial and detrimental outcomes, depending on the level of need for collaboration and interdependence (Siegel & Hambrick, 2005). Researchers found that settings in which individuals are motivated to work for individual outcomes rather than group or organizational outcomes, benefit from high levels of horizontal pay disparity as it helps attracting and retaining high performers within the firm, and increases individual effort (Cadsby, Song & Tapon, 2007; Carnahan, Agarwal & Campbell, 2012). Thus, in situations where the need for collaboration and interdependence is low, pay disparity structures can lead to positive results. However, in situations where cooperation and collaboration are important for achieving group or organizational outcomes, researchers found negative results of horizontal pay disparity (Lazear & Rosen, 1981). Negative results of horizontal pay disparity were found both in situations where executives receive far more and far less pay than their fellow peers (Siegel & Hambrick, 2005). Executives making more than their fellow peers may show condescending and aloofness behavior, and social distancing towards their seemingly inferior counterparts (Siegel & Hambrick, 2005). In contrast, executives that receive less pay than their fellow peers may respond with an increase in invidious comparisons (Deutsch, 1985), feelings of jealousy (Barnard, 1938), lower job satisfaction (Pfeffer & Langton, 1993), and resentment towards fellow executives that receive more pay (Siegel & Hambrick, 2005). In both cases, joint or collective effort may be minimized, resulting in an increase of independent behavior and a decrease in collaboration and integration within the TMT (Frederickson et al., 2010; Siegel & Hambrick, 2005). In addition, scholars found that higher horizontal pay disparity is linked with lower research productivity (Pfeffer & Langton, 1993), higher executive turnover (Bloom & Michel, 2002), and lower financial performance across multiple industries (Frederickson et al., 2010).

(15)

individual ambition and competition (Bradley et al., 1999) indicative of a shareholder-orientated view, in contrast to collaborative and cooperative efforts indicative of a stakeholder-oriented view (Hart et al., 2015). Research shows that egalitarianism (O'Brien and David, 2014), trust (Harrison, Bosse & Phillips, 2010), cooperation (Bradley et al., 1999), and reciprocity (Bosse, Phillips & Harrison, 2009) are characteristics needed by the TMT to consider various viewpoints and work collaboratively in order to balance competing stakeholder claims (Wong et al., 2011). These characteristics are associated with a stakeholder-oriented view, which is conducive to higher levels of CSP (Hart et al., 2015). Since firms with high horizontal pay disparity structures foster a shareholder-oriented view rather than a stakeholder-oriented view, it is argued that these firms will lack collaboration and cooperation that is needed to generate high levels of CSP. Hence, the second hypothesis is:

Hypothesis 2: Higher horizontal executive pay disparity leads to lower CSP.

2.3 The financial crisis as moderator of the executive pay disparity – CSP relationship

The financial crisis is likely to be characterized by an uncertain business environment (Karainbrahimoglu, 2010), which had led firms to face different and more intense market conditions and competition. Firms respond to these uncertain times by taking remedial actions, such as cutting costs by laying off employees, postponing investments, reducing budgets and consuming less (Karaibrahimoglu, 2010; Yelkikalan & Köse, 2012). In line with the current study, the question therefore rises of which actions firms take as precaution and saving tendencies for their survival in terms of executive compensation and CSP investments.

(16)

The financial crisis has caused sharp decreases of both firm performance and value of assets for many firms (Bhagat & Bolton, 2014). As a result, shareholders became more aware of the risks of the crisis and were less willing to pay their executives with large salaries (e.g. Murphy, 2012). On top of that, media and public paid more attention to executive compensation levels, focusing on whether executives were paid too much or if they should pay a part of it back (Bührer, 2010). The aforementioned developments led to a change in executives’ compensation structures, whereby the current study will focus on annual salary and bonuses.

According to the agency theory, most executives act in their own interests at the expense of shareholders’ interests (Berle & Means, 1932). In order to align the interests of both parties, shareholders set targets for bonuses on top of executives’ basis salary such that executives run the firm in the interests of the shareholders (Eisenhardt, 1989). However, maximizing shareholder-value in times of crises is difficult and since bonuses depend on firm performance, a decrease in firm performance will lead to a decrease in bonuses as well (John, Mehran & Oian, 2009). Also, due to budget restrictions and limited resources, shareholders are obligated to shift their focus to firm’s survival rather than personal short-term interests. Decreased salaries, lower amounts of bonuses and stricter targets are used as means to reallocate and safe financial capital to cope with the threats of the crisis (Zhu & Dong, 2013). Also, in the face of public anger, a number of executives have voluntarily renounced their bonuses for 2008 (OECD, 2009). Taking the increased scrutiny and cost reductions of executive compensation into account, it is expected that pay disparities within the TMT may have decreased as well due the financial crisis.

The Economist (2009) refers to the financial crisis as “a stress test for good intentions”, and argues that executives nowadays think carefully about CSP investments. On the one hand, cutting back on CSP activities would seem to be a quick and easy way to save

(17)

money and limit the threat for firms’ survival (Economist, 2009; Frankental, 2001). On the other hand, investing in CSP is more important than ever as high levels of CSP can help firms “build or sustain their brand name, consumers’ trust and redefine the relationship between companies and society” (Giannarakis & Theokas, 2011, p.5).

Literature on the relationship between the financial crisis and CSP found that in some cases CSP investments are considered a starting point for improving business operation (e.g. Arevalo & Aravind, 2010). Firms that maintain or increase CSP investments during the financial crisis will not only gain the common benefits such as economic performance (Schiebel & Pochtrager, 2003), employee satisfaction (Fafaliou, Lekakou & Theotokas, 2006) and increased sales (Weber, 2008), but also gain back the lost trust between the firm and society (Decker & Sale, 2009). In addition, research indicates that CSP investments are needed in times of crisis as it allows for better risk management and strengthens the sense of confidence in firms (Yelkikalan & Köse, 2012).

However, even though CSP is increasingly viewed as an important business outcome (Deckop et al., 2006), research shows that investment in CSP is often seen as a threat for firms’ survival (Giannarakis & Theotokas, 2011). Firms that focused on maximizing profit reduced CSP initiatives in times of the financial crisis (Karaibrahimoglu, 2010), and many firms moved away from socially responsible behavior due to high additional efforts and costs of satisfying stakeholders’ expectations (Giannarakis & Theotokas, 2011). It is therefore expected that the financial crisis has led to a decrease in firms’ overall CSP.

Considering the influence that the financial crisis has had on both executive compensation and CSP, it is argued that the executive pay disparity – CSP relationship might also be influenced by the crisis. The current study assumes that cutting back costs on executive compensation may have led to lower pay disparities within the TMT. However,

(18)

2015), it is expected that in bad times lower executive pay disparity will not lead to a strong increase of CSP due to firms’ budget restrictions and limited resources. In addition, since the crisis resulted in overall lower pay disparities within the TMT, it is expected that ‘high’ pay disparities will lead to a relatively less strong decrease in overall CSP. Hence, the third hypothesis is:

Hypothesis 3: The relationship between a) vertical and b) horizontal executive pay disparity and CSP is less negative during times of the financial crisis.

Figure 1. Conceptual model

H1 + 2 H3 Executive Pay Disparity Corporate Social Performance Financial Crisis

(19)

4. Method

4.1 Sample and data collection

In this study, research is conducted on the relationship between executive pay disparity and CSP, and the influence of the financial crisis on this relationship. In order to gain more insights into the research question several hypotheses were drawn up. These hypotheses were tested by means of a database research, from which data were collected and merged for every U.S.-based, publicly traded firm for which firm-level financial data from COMPUSTAT, executive compensation data from EXECUCOMP, and governance variables from RISK METRICS were all available over the time period 2005-2010. In line with previous research, all financial companies (such as insurance companies and banks) were excluded from the sample due to increased government regulation of these industries that create certain compensation structures that are less likely to be generalizable to other firms (Hart et al., 2015; Wright, Lado & Elenkov, 2005). After collecting the data, the data were merged with firms for which CSP data from the Kinder, Lydenberg and Domini, & Co. (KLD) database was available. Due to the fact that CSP is measured at time t +1, the chosen time period for the KLD database was 2006-2011. After these steps were taken, the final sample existed of 792 firms across 56 industries (2-digit SIC codes).

4.2 Measures

Dependent Variable. The CSP measure was constructed using the KLD database. Despite the

fact that there are other CSP rankings available (e.g. FTSE4Good and Innovest), the KLD ratings are among the oldest, and most influential and frequently used by academics when measuring CSP in social performance research (Chatterji, Levine & Toffel, 2009; Deckop, Merriman & Gupta, 2006; Waddock, 2003). Over the past twenty years it has been used

(20)

previous research, seven different categories were used in this study (Hart et al., 2015): human rights, corporate governance, employees, products, environment, community, and diversity. It is assumed that all seven categories are deemed equally important for CSP, as there is no study that provides objective weights to the categories. Furthermore, studies found a lag between the behavior of executives and firm-level outcomes of the company (Wong, Ormiston & Tetlock, 2011). Due to the fact that compensation disparity structures are used as incentives to motivate executives towards certain behaviors, the current study follows previous studies and measures CSP as t + 1, allowing the effects of these behaviors in year t to reveal themselves a year later in the firm’s subsequent outcomes (Hart et al., 2015; Hong, Li & Minor, 2014).

The KLD ratings indicate the presence or absence of strengths and concerns regarding the aforementioned categories (Mattingly & Berman, 2006). Although some academics have subtracted the CSP strengths from the CSP concerns, others have argued that this can lead to the scores cancelling each other out or some of the scores overshadowing the others (Mattingly & Berman, 2006; Strike, Gao & Bansal, 2006). When the focus of the research lies on how strong the CSP strengths and CSP concerns are, then it is appropriate to examine the strengths and concerns separately. The current study focuses on how stakeholder management affects overall CSP, whereby trade-offs can frequently be made that affect strengths and concerns separately or jointly. A specific focus on either strengths or concerns is therefore unnecessary. Thus, in line with prior research, the CSP variable is utilized as indicator of the CSP of the firm (Hart et al., 2015; Wong et al., 2011). CSP is computed as the overall mean over the seven CSP strengths and the seven CSP concerns (van der Laan, van Ees & van Witteloostuijn, 2008).

(21)

Independent Variable. To calculate executive pay disparity, data was collected from the

EXECUCOMP database. The calculation of executive pay disparity follows a two-step process. First, compensation is determined as the sum of annual salary and bonuses, the Black Scholes value of stock options and restricted stock grants (Devers et al., 2007; Yanadori & Marler, 2006). However, since the Black Scholes value of stock options and restricted stock grants did not provide sufficient data to use for research, the current study solely uses the sum of annual salary and bonuses to calculate compensation. The sum of annual salary and bonuses is regarded as short-term compensation, in contrast to long-term compensation that includes the sum of the Black-Scholes value of stock options and restricted stock grants (Devers et al., 2007; Yanadori & Marler, 2006). Hence, the compensation measure in this study focuses on short-term compensation. Second, two types of actual disparity values are distinguished: vertical and horizontal disparity (Hart et al., 2015). In this study both vertical disparity and horizontal disparity are used. In order to calculate the vertical pay disparity values the log of difference was used between the compensation of the CEO and the average compensation of the rest of the TMT (Carpenter & Sanders, 2004). To create the horizontal pay disparity values the coefficient of variation (CV) of pay for the non-CEO executives was calculated. These values are calculated by dividing the standard deviation (SD) by the arithmetic mean of the TMT pay (Frederickson et al., 2010).

With the hypotheses regarding the relationship between executive pay disparity and CSP it is assumed that pay disparity will have negative effects on CSP, regardless of the type of compensation. However, to employ a vigorous test of the hypotheses and to provide more in-depth results, compensation is divided in salary (i.e. annual salary compensation), bonuses, and total compensation (i.e. the sum of the annual salary and bonuses). Hence, the effect of salary, bonuses and total vertical pay disparity on CSP is examined to test hypothesis 1, and

(22)

hypothesis 2. These measures were also used when examining the moderation effect of hypothesis 3.

Moderating Variable. To test the assertions for the moderator ‘financial crisis’, the same data

was used from COMPUSTAT, EXECUCOMP and RISK METRICS over the time period 2005-2010, and from KLD over the time period 2006-2011 (because CSP is measured at time

t+1). This time range captures the period before (2005-2007) and during/after (2008-2010)

the financial crisis, which may show differences in firms’ CSP level (Giannarakis & Theotokas, 2011) and executives’ compensation (Zhu & Dong, 2013) conducive to a difference in strength of their relationship. The financial crisis variable is calculated as ‘1’ during/after the financial crisis, and ‘0’ for before the financial crisis.

Control Variables. Several control variables are included in the analysis. First, there is

controlled for firm size, as firms with more employees are likely to have more hierarchical compensation structures, which in turn may influence the vertical pay disparity. In line with previous work, the number of employees is used as a measure of firm size (Siegel & Hambrick, 2005).

Previous work has found that the composition of the TMT is linked with CSP (Johnson & Greening, 1999) and can influence corporate strategies (Canella, Park & Lee, 2008). Therefore, board size (the total number of directors on the board) and the number of

female executives are also used as control variables.

In addition, whether a CEO also serves as chairman of the board may influence the CEO’s ability to influence CSP investments and corporate strategies (Zajac & Westphal, 1996). Also, a CEO with long tenure may have greater influence on corporate strategies, which in turn can influence CSP as well (Henderson & Fredrickson, 2001). Thus, there is

(23)

controlled for both CEO duality (calculated as ‘1’ if the CEO is also chairman and ‘0’ if the CEO is not the chairman) and CEO tenure (measured as the number of years since the CEO took office) (Fredrickson et al., 2010; Sanders & Carpenter, 1998).

Previous studies have found that accounting-based measures (more than market-based measures) of firm performance can correlate with CSP (e.g. Orlitzky, Schmidt & Rynes, 2003). Also, if strong firm performance is disproportionately attributed within the TMT, it could increase the CEO’s pay more than the pay of other members, resulting in higher levels of pay disparities (Finkelstein & Hambrick, 1989). Therefore, this study controls for firm performance using return on assets (ROA) (Frederickson, et al., 2010; Wright et al., 2005). Lastly, as industries have different norms regarding executive compensation, executives may compare their pay to industry-peers as well as firm-peers (Finkelstein & Hambrick, 1989; Hart et al., 2015). Because of this, there is controlled for industry pay norms by including a measure of industry salary, bonuses and total vertical disparity and industry

salary, bonuses and total horizontal disparity, which were calculated as the average of

vertical and horizontal pay disparity in 2-digit primary industries (Coombs & GiIley, 2005). Due to a large amount of missing values, it was not possible to calculate the industry bonuses vertical/horizontal disparity. The industry total vertical/horizontal disparity is therefore used in analyses regarding the bonuses.

4.3 Analysis

The data used in this study include multiple observations of firms over several years. Because of this, the dataset consists of multiple unbalanced panels of observations (not every firm is represented in every year of the panel), which makes an ordinary least squares regression not appropriate to analyze the data. Therefore, to test the hypotheses, a cross-sectional time series

(24)

errors. In line with previous work (Hart et al., 2015), a Breusch-Pagan test was used to detect whether heteroskedasticity was present in the data. Results show that the null hypotheses could be rejected (p < 0.001), which indicates the presence of heteroskedasticity. Furthermore, to test for autocorrelation the Woolridge test was used (Woolridge, 2002). Results from the Woolridge test showed the existence of first-order autocorrelation. Taking both the presence of heteroskedasticity and first-order autocorrelation into account, the ‘vcu robust cluster’ procedure was utilized in Stata. This procedure allows standard error estimates to be robust to disturbances that are heteroskedastic and autocorrelated (Woolridge, 2002). After taking the aforementioned steps, the influence of vertical and horizontal pay disparity on CSP was tested by means of a regression analysis. This was done for each of the three measures of vertical and horizontal pay disparity. For all analyses, control variables were included. A Hausman test (Hausman & Taylor, 1981) indicated that the fixed-effects model should be rejected in favor of the random effects model. The Wald chi test replaces the F-test in random effects structural models to test the goodness of fit for the model on the data. It can be interpreted the same way as the F-test; a significant result in this test indicates that the model approximates the data distribution.

To test the moderation effect of the financial crisis on the executive pay disparity – CSP relationship, a regression analysis with moderator is used. The moderating interaction term was a dichotomous variable, scoring ‘1’ for years 2008-2010, and ‘0’ for the years 2005-2007 (for CSP t + 1). Again, this was done for all three measures of vertical and horizontal pay disparity, and included the control variables. A Hausman test indicated that for this model the random-effects specification should be used as well. A significant result of the Wald chi test indicates that the model approximates the data distribution.

(25)

5. Results

Table 1 reports the means, standard deviations, and correlations of the dependent, independent, moderating and control variables that were used in this study. Statistics show that both vertical pay disparity and CSP have a strong correlation with firm size, number of female executives and board size. Moreover, horizontal pay disparity has a strong correlation with CEO tenure.

The first two hypotheses predict that executive pay disparity negatively influences CSP. A regression analysis is used to test whether these relationships are indeed negative and significant. Hypothesis 1 predicts that vertical pay disparity is negatively related to CSP. When testing this hypothesis, a distinction is made between the effect of total, salary and bonuses vertical disparity on CSP. As shown in table 2, the relationship between vertical pay disparity and CSP is neither negative nor significant. All three compensation measures, total (b = 0.029; p = 0.902), salary (b = 0.535; p = 0.113), and bonuses (b = 0.107; p = 0.353) were not significant. Hence, these results provide evidence that high levels of vertical pay disparity do not lead to lower overall CSP. Hypothesis 1 is therefore rejected.

Hypothesis 2 predicts that horizontal pay disparity is negatively related to CSP. Again, a distinction is made between total, salary and bonuses horizontal disparity. As shown in table 2, the relationship between horizontal pay disparity and CSP is not significant. All three compensations measures, total (b = 0.377; p = 0.485), salary (b = 0.338; p = 0.133) and bonuses (b = -0.051; p = 0.929) were not significant. High levels of horizontal pay disparity do not result in lower overall CSP. Thus, hypothesis 2 is also rejected.

Hypothesis 3 predicts that the relationship between executive pay disparity and CSP is influenced by the financial crisis. To test this assumption, a regression analysis with interaction (moderator analysis) is conducted with the financial crisis as moderator, executive

(26)

distinguished in a) vertical and b) horizontal pay disparity. Furthermore, to test this hypothesis all three compensation measures (total, salary and bonuses) are examined for both vertical and horizontal pay disparity.

In table 3 it is shown that the financial crisis influences the relationship between total (b = 2.235; p < 0.01), salary (b = 3.581; p < 0.01) and bonuses (b = 0.986; p < 0.01) vertical disparity. After including the financial crisis as moderator, significant results are found for the negative effect of total (b = -0.733; p < 0.01) and salary (b = -0.877; p < 0.01) vertical disparity on CSP. However, the relationship between bonuses vertical disparity and CSP remains insignificant (b = -0.202; p = 0.130). Due to the fact that the moderation effect for total vertical disparity is significant, an assertion can still be made that the financial crisis moderates the relationship between vertical pay disparity and CSP.

Furthermore, results in table 3 show that the financial crisis does not influence the relationship between total (b = -0.724; p = 0.258), salary (b = -0.182; p = 0.771) and bonuses (b = 0.109; p = 0.894) horizontal disparity. The effect of total (b = 0.323; p = 0.607), salary (b = 0.182; p = 0.763) and bonuses (b = -0.053; p = 0.930) horizontal disparity on CSP is still not significant after including the financial crisis as moderator. Thus, the financial crisis does not moderate the horizontal pay disparity – CSP relationship. Hypothesis 3 is therefore partially supported. Table 4 reports the means and standard deviations of the vertical and horizontal pay disparity per year to provide more understanding for the current findings of hypothesis 3, which will be discussed in the next section.

(27)

Table 1. Descriptive statistics and correlations

Notes: Correlation coefficients greater than 0.04 are significant at p<0.05; those greater than 0.06 are significant at p<0.01.

Variables Mean S.D. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

1. CSP 0.55 3.35 1.00

2. Salary vertical disparity 5.66 0.22 0.16 1.00

3. Bonuses vertical disparity 5.50 0.81 0.09 0.37 1.00

4. Total vertical disparity 5.73 0.29 0.10 0.70 0.78 1.00

5. Salary horizontal disparity 1.86 0.12 -0.02 0.16 0.23 0.22 1.00

6. Bonuses horizontal disparity 2.14 0.22 0.02 0.03 -0.03 -0.01 0.55 1.00

7. Total horizontal disparity 1.86 0.12 -0.02 0.16 0.23 0.22 1.00 0.55 1.00

8. Board size 9.32 2.14 0.29 0.38 0.24 0.31 0.03 -0.06 0.03 1.00

9. Female executives 0.12 0.10 0.38 0.21 0.01 0.09 0.01 0.11 0.01 0.31 1.00

10. CEO duality 0.59 0.49 0.07 0.20 0.07 0.17 -0.02 -0.07 -0.02 0.06 0.09 1.00

11. CEO tenure 18.00 10.52 -0.12 0.05 0.09 0.05 0.35 0.18 0.35 -0.05 -0.14 0.10 1.00

12. Industry salary vertical disparity 5.67 0.09 0.06 0.35 0.21 0.29 0.01 -0.04 0.01 0.24 0.09 0.05 0.09 1.00

13. Industry total vertical disparity 5.73 0.11 0.00 0.31 0.33 0.39 0.05 -0.10 0.05 0.23 -0.00 0.04 0.11 0.86 1.00

14. Industry salary horizontal disparity 1.87 0.03 -0.04 0.13 0.15 0.18 0.30 0.14 0.30 0.07 0.04 -0.09 0.09 0.10 0.24 1.00

15. Industry total horizontal disparity 1.87 0.03 -0.04 0.13 0.15 0.18 0.30 0.14 0.30 0.07 0.04 -0.09 0.09 0.10 0.24 1.00 1.00

16. ROA 0.04 0.13 0.01 0.03 -0.02 0.04 -0.08 -0.17 -0.08 0.04 -0.01 0.02 -0.03 -0.03 -0.01 -0.04 -0.04 1.00

17. Firm size 3.82 0.73 0.30 0.48 0.31 0.44 0.16 0.05 0.16 0.51 0.25 0.14 -0.01 0.11 0.10 0.18 0.18 0.17 1.00

(28)

Table 2. Random effects regression of vertical and horizontal pay disparity on CSP

Notes: Unstandardized regression coefficients, robust standard errors in parentheses. p<0.05*, p<0.01**.

Dependent variable (CSP)

Effect of total vertical & horizontal disparity on CSP

Effect of salary vertical & horizontal disparity on CSP

Effect of bonuses vertical & horizontal disparity on CSP Model 1 Control Model 2 Main effect Model 3 Control Model 4 Main effect Model 5 Control Model 6 Main effect

Independent variables b (s.e.) b (s.e.) b (s.e.) b (s.e.) b (s.e.) b (s.e.)

Total vertical disparity 0.029 (0.232)

Salary vertical disparity 0.535 (0.337)

Bonuses vertical disparity 0.107 (0.115)

Total horizontal disparity 0.377 (0.541)

Salary horizontal disparity 0.338 (0.530)

Bonuses horizontal disparity -0.051 (0.750)

Board size 0.137** (0.038) 0.136** (0.038) 0.136** (0.038) 0.132** (0.038) 0.137** (0.038) 0.078 (0.082)

Female executives 6.512** (0.817) 6.495** (0.815) 6.475** (0.817) 6.406** (0.813) 6.512** (0.817) 8.642** (1.692)

CEO duality 0.011 (0.109) 0.011 (0.109) 0.009 (0.109) -0.010 (0.110) 0.011 (0.109) -0.157 (0.186)

CEO tenure -0.048** (0.005) -0.050** (0.006) -0.048** (0.005) -0.049** (0.006) -0.048** (0.005) -0.031** (0.008)

Industry total vertical disparity 0.944 (0.747) 0.924 (0.767) 0.944 (0.747) -0.007 (1.084)

Industry salary vertical disparity 1.485 (0.938) 1.021 (0.982)

Industry total horizontal disparity -4.782 (2.790) -5.048 (2.807) -4.782 (2.790) -7.487 (4.564)

Industry salary horizontal disparity -4.616 (2.712) -4.767 (2.730)

ROA -0.039 (0.415) -0.045 (0.414) -0.038 (0.418) -0.061 (0.414) -0.039 (0.415) -0.045 (0.413)

Firm size 1.025** (0.153) 1.021** (0.155) 1.023** (0.153) 0.968** (0.160) 1.025** (0.153) 0.990** (0.259)

Constant -1.195 (5.398) -1.386 (5.406) -4.494 (6.150) -4.942 (6.192) -1.195 (5.398) 8.884 (9.660)

Observations (N) 5069 5069 5069 5069 1013 1013

Wald chi2 (Wald chi²) 334.83** 338.62** 340.25** 352.55** 334.83** 80.18**

(29)

Table 3. Random effects regression of vertical and horizontal pay disparity on CSP with the financial crisis as moderator

Notes: Unstandardized regression coefficients, robust standard errors in parentheses. p<0.05*, p<0.01**.

Dependent variable (CSP)

Effect of total vertical & horizontal disparity on CSP

Effect of salary vertical & horizontal disparity on CSP

Effect of bonuses vertical & horizontal disparity on CSP Model 1 Control Model 2 Conditional effect Model 3 Control Model 4 Conditional effect Model 5 Control Model 6 Conditional effect

Independent variables b (s.e.) b (s.e.) b (s.e.) b (s.e.) b (s.e.) b (s.e.)

Total vertical disparity -0.733** (0.251)

Salary vertical disparity -0.877* (0.367)

Bonuses vertical disparity -0.202 (0.133)

Total horizontal disparity 0.323 (0.628)

Salary horizontal disparity 0.182 (0.605)

Bonuses horizontal disparity -0.053 (0.608)

Board size 0.129** (0.037) 0.133** (0.037) 0.121** (0.037) 0.215** (0.037) 0.087 (0.079) 0.093 (0.079)

Female executives 6.195** (0.796) 6.230** (0.794) 6.050** (0.789) 6.115** (0.787) 8.666** (1.666) 8.629** (1.658)

CEO duality -0.048 (0.108) -0.025 (0.108) -0.081 (0.108) -0.054 (0.108) -0.139 (0.189) -0.134 (0.189)

CEO tenure -0.043** (0.005) -0.044** (0.005) -0.043** (0.005) -0.045** (0.005) -0.033** (0.008) 0.032** (0.008)

Industry total vertical disparity 0.249 (0.757) 0.822 (0.764) -0.216 (1.060) 0.066 (1.068)

Industry salary vertical disparity 0.350 (0.954) 1.026 (0.986)

Industry total horizontal disparity -4.777 (2.802) -5.141 (2.810) -8.440 (4.588) -8.373 (4.540)

Industry salary horizontal disparity -4.727 (2.735) -4.970 (2.742)

ROA -0.106 (0.407) -0.075 (0.409) -0.126 (0.400) -0.104 (0.405) -0.067 (0.403) -0.102 (0.392)

Firm size 0.951** (0.162) 1.021** (0.155) 0.915** (0.153) 0.995** (0.160) 0.912** (0.256) 0.947** (0.256)

Financial crisis -9.683** (1,632) -11.116** (1.835) -18.234** (2.204) -19.640** (2.448) -4.945** (1.654) -5.561** (1.910)

Total vertical disparity * financial crisis 2.235** (0.279)

Salary vertical disparity * financial crisis 3.581** (0.392)

Bonuses vertical disparity * financial crisis 0.986** (0.201)

Total horizontal disparity * financial crisis -0.724 (0.640)

Salary horizontal disparity * financial crisis -0.182 (0.626)

Bonuses horizontal disparity * financial crisis 0.109 (0.818)

Constant 2.974 (5.454) 3.657 (5.443) 2.586 (6.281) 3.495 (6.223) 12.551 (9.549) 11.800 (9.793)

Observations (N) 5069 5096 5069 5069 1013 1013

Wald chi2 (Wald chi²) 386.22** 400.41 436.63** 439.37** 98.16** 102.48**

(30)

Table 4. Descriptive statistics of vertical and horizontal pay disparity by year Salary vertical disparity Bonuses vertical disparity Total vertical disparity Salary horizontal disparity Bonuses horizontal disparity Total horizontal disparity

Fiscal year Mean S.D. Mean S.D. Mean S.D. Mean S.D. Mean S.D. Mean S.D.

2005 5.64 0.23 5.46 0.79 5.72 0.31 1.86 0.13 2.14 0.21 1.86 0.13 2006 5.64 0.23 5.43 0.86 5.72 0.30 1.87 0.12 2.16 0.22 1.87 0.12 2007 5.79 0.35 5.55 0.89 5.97 0.34 1.86 0.12 2.13 0.22 1.86 0.12 2008 6.65 1.03 5.56 0.89 6.73 1.07 1.86 0.12 2.14 0.22 1.86 0.12 2009 6.66 1.03 5.56 0.89 6.74 1.07 1.86 0.12 2.14 0.23 1.86 0.12 2010 6.66 1.03 5.57 0.90 6.74 1.07 1.86 0.12 2.15 0.23 1.86 0.12

(31)

6. Discussion

6.1 Summary results

The aim of this study was to provide an in-depth understanding of the relationship between executive pay disparity and CSP, and to test whether the financial crisis moderates the aforementioned relationship. The results from the tested hypotheses have lead to interesting findings, which are discussed below.

First, hypothesis 1 and 2 are both rejected. Hypothesis 1 expected that high vertical pay disparity leads to lower overall CSP for firms. In line with the tournament theory, research found that individuals are motivated to put forth higher levels of individual effort to achieve greater personal rewards (Lazear & Rosen, 1981). However, in order to achieve these rewards, individuals may engage in subversive, individualistic and competitive behaviors. Such behaviors foster a shareholder-orientation that can lead to lower levels of CSP. Hypothesis 2 expected that high horizontal pay disparity leads to lower overall CSP for firms. Based on the equity theory, scholars argue that when executives compare their pay with their fellow non-CEO executives, “feelings of condescension are generated in those making comparatively more than their peers and feelings of resentment are stirred up in those making comparatively less” (Hart et al., 2015, p. 24). These feelings result in more independent behaviors and less collaborative effort, making executives less capable of managing the complexity of stakeholder management (Wong et al., 2011). Hence, high horizontal pay disparity fosters a shareholder-oriented view that can lead to lower levels of CSP.

Findings of the current study do not provide support for the expected relationship between executive pay disparity and CSP. A possible explanation for these findings can be the enactment of the Sarbanes-Oxley Act (SOX) of 2002, a United States (U.S.) federal law that set new and/or expanded requirements for all U.S. public company boards, management

(32)

(Jahmani & Dowling, 2008). Derived from the law, an independent board is assigned to monitor firms’ fulfillment of this new corporate governance regime. Several major elements of SOX have such strong influence on executive compensation and investment actions that it may have abolished the executive pay disparity – CSP relationship shown in previous studies (e.g. Hart et al., 2015). For example, the law consists of multiple sections and mandates that state that executives have an individual responsibility for the accuracy, validity and completeness of corporate financial reports to increase the effectiveness of a firm’s internal control (Jahmani & Dowling, 2008). As a result, firms become more transparent, which leads to executives being obligated to justify the appropriateness of investment decisions (Zong, 2004), and more engagement and greater accountability of the TMT to demonstrate a strong control environment (Wagner & Dittmar, 2006).

In addition, the law encourages more effective coordination and cooperation within the firm (Wagner & Dittmar, 2006). The inclusion of a code of ethics promotes honest and ethical conduct, and benefits the ethical handling of actual or apparent conflicts of interest between employee relationships (Jahmani & Dowling, 2008). As a result, selfish and subversive behaviors and attitudes of executives may have diminished.

Lastly, the law prohibits publicly traded firms from making or arranging loans (including salary advances) to their directors and executive officers, and forfeitures bonuses and profits due to financial reporting restatement as a result of misconduct (Zong, 2004). Executives can only receive loans that are no more generous than what is commonly offered to the public and need to disclose all their stock purchases and sales (Zong, 2004). Adding to the scrutiny of executive compensation is the Securities and Exchange Commission (SEC) that adopted extensive and far reaching amendments to the disclosure requirements for executives’ compensation in 2006 (McGuireWoods, 2007, p.1) to clarify and demystify executive compensation. These amendments focus on whether executive compensation

(33)

structures are ‘right’ or ‘fair’, place limits on executive compensation, and provides information for better monitoring and regulating executive pay practices (Salley, 2009). Due to SOX’s provisions and SEC’s new executive compensation disclosure rules, disparities between executive pay compensation may have decreased.

Recognizing the aforementioned consequences of SOX and SEC, it can be argued that this might have led to the absence of significant support for the first two hypotheses in this study. The full, fair, accurate, timely and understandable periodic reports make it more important for firms to meet the requirements of the law. These requirements focus on both shareholders and stakeholders, making it difficult for firms to solely embrace a shareholder- or stakeholder-oriented view conducive to respectively lower and higher levels of CSP (Hart et al., 2015). Moreover, due to the corporate governance regime changes and strong monitoring from independent boards, the TMT is obligated to deliver solid performance for the pay they receive and constrain their self-interested behaviors (Wang & Yang, 2014). The new compensation rules may have added to a decrease in executive pay disparity. Taking all these changes into account, it is plausible to assume that firms’ executive pay disparities are not high enough to and/or do not play a significant role in changing executives’ behavior and attitudes in a way that it (negatively) influences a firm’s overall CSP.

Hypothesis 3 is partially supported. Hypothesis 3 expected that the financial crisis influences the relationship between vertical and horizontal pay disparity and CSP, resulting in a less negative relationship. The financial crisis has put many firms and industries under pressure, forcing them to take actions to minimize or avoid the negative effects of the crisis (Yelkikalan & Köose, 2012). These actions often include a change in executive compensation structures (Zhu & Dong, 2013) and CSP investments (Giannarakis & Theotokas, 2011). As a result, both pay disparities within the TMT and firms’ CSP levels may have decreased.

(34)

is assumed that this relationship is less strong because of the smaller overall pay disparities and low overall CSP investments. However, due to the fact that the executive pay disparity - CSP relationship was not found in the current study before adding the financial crisis as moderator, it was interesting to test whether a relationship between executive pay disparity and CSP would be established during times of the crisis.

Results show that there is a negative relationship between vertical pay disparity and CSP as a result from the crisis. However, still no relationship is found between horizontal pay disparity and CSP. Within this context, findings support the assumption that the financial crisis can influence the executive pay disparity – CSP relationship. Table 3 and 4 provide a strong possible explanation for the current findings. In line with the expectations of this study, table 3 provides support for the fact that there is a decrease in firms’ CSP level during times of the financial crisis. Statistics of table 4 show that horizontal pay disparity levels did not change during and after the crisis, which can explain the absence of the horizontal pay disparity – CSP relationship. In contrast, table 4 shows that vertical pay disparities increase during and after the crisis. An increase in vertical pay disparity may have led to less cooperation and collaboration and more individual and competitive behaviors within the TMT, which leads to lower CSP levels (Brickson, 2007). Thus, higher vertical pay disparity and less CSP investments can explain the strong negative relationship between vertical pay disparity and CSP.

6.2 Theoretical and practical implications

The results of the current study have some important implications. First, although previous research has found a relationship between executives’ compensation incentives and CSP (e.g. Hart et al., 2015), the current study provides evidence that the governance of these incentive systems often fail. The question therefore rises of whether the executive pay disparity – CSP

(35)

relationship is truly causal or whether corporate governance regarding compensation incentives is truly effective. It is argued that these systems often fail because decisions and negotiations are not carried out at arm’s length, and executives have too much influence over the level and conditions of performance-based compensation (OECD, 2009). Therefore, this study suggests that in order to create effective compensation structures, steps must be taken within the governance process where the roles and responsibilities of those involved are clearly defined and separated, and in which executive board members cannot participate due to the inherent conflict of interest.

Second, the assumptions of the current study are build on the tournament theory and the equity theory, which both provide arguments for the expectation that compensation incentives can motivate executives towards certain behaviors, conducive to lower of higher levels of CSP. Findings of the current study provide evidence that the arguments of the aforementioned theories can be applied when a certain level of pay disparity is established within a firm, which however not always the case. Therefore, current research contributes to existing literature by adding to the legitimacy of the use of these theories within this field of study.

Lastly, results from the current study provide evidence that most firms decrease their CSP investments in bad times. However, it is argued that the benefits that arise by implementing CSP initiatives are even more important for firms’ survival during the financial crisis as they assist in relocating firms’ business to a better position (Fernández & Souto, 2009). Current study stresses the importance of analyzing the costs and benefits of CSP investments during crises for firms. In addition, results show that the financial crisis has influenced pay disparity within the TMT, which leads to lower levels of CSP. Taking these results into consideration, firms can anticipate on this and change their

(36)

6.2 Limitations and future directions

Like all research, the current study contains limitations that confine the validity and generalizability of the findings and related conclusions. First, due to the fact that data was collected from publicly traded firms in the U.S. for which executive compensation data, governance data, financial data and CSP data were available, it is questionable whether or to what degree these findings are generalizable to firms not having these characteristics. Although this is a common limitation discussed in similar research (e.g. Coombs & Gilley, 2005; Hart et al., 2015), it is encouraged to continue searching for (unique) data that enable similar issues to be examined in dissimilar ways and contexts.

Second, this study focuses on overall CSP rather than any particular dimension of CSP when collecting data for the analysis. As a result, this study fails to account for firms that may not have a high overall CSP score but do excel in specific dimensions. The same applies for firms with high overall scores but that are brought down by one bad CSP dimension, resulting in lower overall scores. Utilizing overall CSP instead of just CSP strengths or concerns, or any specific CSP dimension, limits the findings of this study to inferences made from overall CSP scores. For future research it is therefore recommended to also look at CSP strengths and concerns separately, and specific CSP dimensions.

Another limitation of the current study is that because of limited data the total compensation was measured as solely the sum of annual salary and bonuses, therefore neglecting the influence of the Black Scholes value of stock options and restricted stock grants (Devers et al., 2007; Yanadori & Marler, 2006) on executive pay disparities. In doing so, the current findings are limited to short-term compensation disparities, which are likely to be smaller than if Black Scholes value of stock options and restricted stock grants were included (Frederickson et al., 2010). To increase validity it is therefore recommended that future studies seek data that provides all measurements to calculate both short-term and

(37)

long-term compensation.

Additionally, findings of the moderation effect focused on the influence of the financial crisis in general. However, it can be argued that the impact of the financial crisis is different for different industries. It is interesting for future research to investigate whether the current findings apply for different industries separately. Moreover, it would be interesting for future research to investigate how compensation was determined during the crisis to identify what caused the increase of vertical pay disparity. Also, for extension of this study, future research is encouraged to examine narrower settings and specific context in which the current findings are strengthened or weakened. For example, studying how technological intensity (Siegel & Hambrick, 2005) or other contextual variables influence the executive pay disparity – CSP relationship can be worthy fields of inquiry that provides for more insights for studies on these topics. In addition, examining firms with national cultures other than the U.S. can lead to interesting findings as well; in countries that highly value collectivism, the effects of executive pay disparities on CSP may be more severe than the current study implies. Lastly, results show that gender diversity correlates with both pay disparities and CSP. Future research is therefore encouraged to examine whether other aspects of board diversity influences executive compensation and CSP.

(38)

7. Conclusion

Database research in the current study was conducted to provide an (detailed) answer to the following question: To what extend does executive pay disparity influence CSP, and how

does the financial crisis influence this relationship?

It is argued that firms with high executive pay disparity encourage competition and individual ambition, which fosters a shareholder-oriented view conducive to lower levels of CSP (Bradley et al., 1999; Brickson, 2007; Hart et al., 2015). In contrast, firms with low executive pay disparity encourage collaboration and cooperation, and are linked to a stakeholder-oriented view conducive to higher levels of CSP (Hart et al., 2015; Siegel & Hambrick, 2005). Previous studies examining these topics found that internal factors, such as executive pay disparities, indeed influence a firm’s overall CSP (e.g. Hart et al., 2015). However, they often neglect external factors that might impact the aforementioned relationship as well. Therefore, this study extends existing literature by taking the financial crisis into account as an external factor that influences firms from the outside. The financial crisis of 2008 has led to drastically changed conditions that have brought upon new challenges for many firms. A decrease in firms’ CSP levels is expected due to threats of the financial crisis. In addition, compensation structures change due to cost reductions, limited resources and increased attention from the media, public and shareholders.

Studies investigating the relationship between executive compensation and firm’s CSP found mixed results. Scholars argue that these mixed findings are the result of focusing solely on the level of pay and the role of the CEO in CSP investments (e.g. Hart et al., 2015; Wong et al., 2011). Thus, by taking the role of pay distribution within the broader TMT into account, this study contributes greatly to the previous studies. In addition, all members of the TMT rather than the five highest paid members were considered when examining the aforementioned relationship, increasing the validity of the current findings (Hart et al., 2015).

(39)

Moreover, several theories were used as framework for the expected relationships, therefore providing more empirical evidence for the legitimacy of existing literature. Also, by including the financial crisis as moderator of the executive pay disparity – CSP relationship, this study extends previous research and establishes a basis for future longitudinal studies that investigate the impact of the financial crisis from a long-term perspective.

A database research was conducted to find support for the formulated assumptions in this study. Data were collected from COMPUSTAT, EXECUCOMP, and RISK METRICS over the time period 2005-2009, and from the KLD database over the time period of 2006-2010. After that, a regression analysis is utilized to test hypothesis 1 and 2, and a regression analysis with moderation to test hypothesis 3. Results did not provide sufficient support for the expected vertical and horizontal pay disparity – CSP relationship of the first two hypotheses. However, results show that the financial crisis influences the relationship between vertical pay disparity and CSP. No support was found for the horizontal pay disparity – CSP relationship.

The current study explains the absence of the relationship between executive pay disparity and CSP by the impact that the enactment of SOX and rules of SEC have on firms. Due to law provisions and strong monitoring of independent boards, the TMT is obligated to justify their decisions, increase responsibility and act more in common interests. Furthermore, the provisions of SOX and rules SEC might have resulted in smaller pay disparities. As a consequence, different incentives and penalties influence executives’ decisions regarding CSP investments and the lower executive pay disparity might not be sufficient to motivate executives towards certain behaviors.

A negative relationship between vertical pay disparity and CSP was found for firms during and after the financial crisis. Findings of the current study explain this by showing an