MSc Thesis A&C Controlling
CSR-related performance targets and the amount of CEO compensation
Master‟s thesis Controlling (EBM870B20) Rijksuniversiteit Groningen
Abstract
My study investigates the existence of a relationship between targets in CEO performance measurement related to Corporate Social Responsibility (CSR) and CEO compensation amounts. Previous CSR-related research in accounting literature has mainly focused on CSR reporting and the implementation of CSR in management control (systems). My study tries to add understanding of the effects of the implementation of CSR-related targets. Drawing from agency- and stakeholder theory, CSR-related targets are expected to lead to superior CSR performance and higher CEO compensation. Cross-sectional analysis was performed on British FTSE100 index constituents‟ 2014 fiscal years, scanning their annual reports for CSR-related targets in their CEO‟s performance evaluation. CSR-CSR-related targets were found to lead to superior environmental performance, but not overall CSR performance. Environmental performance was also found to create an indirect, fully mediated relationship between CSR-related targets and CEO compensation amounts.
Keywords: Corporate social responsibility; CSR; executive compensation; management control; performance measurement.
Name: Lars Koldeweij
Student number: S3015378
Track: MSc Accountancy & Controlling (A&C) Supervisor: dr. R.C. Trapp
Co-assessor: dr. C. A. Huijgen
Date: June 20th, 2017
1
1. Introduction
Corporate Social Responsibility (CSR) programs and investments are of increasing importance to companies in the 21st century (Hong et al., 2016). CSR focuses on the broader environmental and social issues firms are facing, instead of the traditional focus on shareholder returns. Society demands for CSR are expanding and show a degree of permanence (Mahony & Thorn, 2005) and arguments have been made for CSR to be a source of opportunity for competitive advantages (Porter & Kramer, 2006). Tension has arisen concerning the effects (positive- or negative effects) of CSR on organizations: on one hand, CSR seems to have a positive effect on firm value, with unexpected CSR disclosures showing to be related to firm value (Cahan et al., 2015) and firms showing strong CSR performance enjoying higher fundamental value estimates (Elliott et al., 2014). On the other hand, empirical evidence remains ambiguous (Hong et al., 2016). However, a majority of findings indicate the existence of a positive relationship between companies‟ socially responsible conduct and financial performance (Margolis & Walsh, 2003). Besides, recent findings by Hong et al. (2016) found “consistent evidence that CSR is likely to be financially beneficial
for firms, and for shareholders” (p. 208). Even with no clear, uncontested evidence of the
effects of CSR actions by organizations, this indicates that CSR actions appear to be a valuable source for organizations, for example due to higher customer loyalty through improved reputation because of CSR or because of other stakeholders‟ demands regarding the implementation of CSR (Huang & Watson, 2015). As such, this research takes the notion of CSR having a positive effect on organizational performance.
Management accounting- and -control systems can be expected to reflect CSR activities. Agency theory suggest that in order for CEOs to act in the best interest of the firm, incentives can be implemented to reduce the risk of CEOs preferring their own interests over the interests of the firm. One of these incentives can be based around compensation schemes (Baker et al., 1988). When CSR is regarded as a strategic objective and a source of competitive advantage, CSR (performance) will be important to the firm. Hence, incorporating CSR performance targets and –measurements will provide incentives for CEOs to pursue CSR targets which are to the best interest of the firm, since the interests of CEO and the firm will then be aligned and provide motivation for CEOs to exert increased effort towards the set targets (Sprinkle, 2000). Besides, CSR is increasingly seen as a strategic objective (Arjaliès & Mundy, 2013). Management control literature suggests that management control systems (MCSs) – of which performance measurement systems are part
2 of (Malmi & Brown, 2008) – should be aligned with (strategic) objectives of the firm, resulting in goal congruence (Merchant, 2006). Therefore, management control literature provides a second reason why MCSs can be expected to have incorporated CSR-related targets (Gond et al., 2012) as these targets will be aligned with corporate strategy and -objectives. These CSR-related targets will focus on non-financial performance concerning topics such as (among others) safety, (employee) well-being, environmental aspects, sustainability and ethical compliance. Incorporation of CSR-related targets does not say anything about potential effects of this incorporation, except for (hopefully) creating goal congruence. However, when looking at MCSs more broadly, part of MCSs is devoted to rewarding employee performance (Malmi & Brown, 2008; Ferreira & Otley, 2009), most notably through compensation. Thus, incorporating CSR-related targets in the performance measurement systems of CEOs can be expected to affect subsequent reward/compensation policies, leading to compensation for the CEOs if the CEO‟s targets are met.
In this light, this study investigates the relationship between CSR-related targets and CEO compensation and will contribute to the scarce literature on the implications and effects of organizations‟ CSR activities. Previous CSR research in accounting has mainly focused on CSR reporting (Cormier & Magnan, 2003; Clarkson et al., 2008; Cahan et al., 2016), while research into the use of CSR in management accounting so far has been limited. In recent years, research into the use of CSR in management control has mainly focused on implementation of CSR in performance measurement systems. Main contributions include building on Simons‟ Levers of Control (LOC) framework by focusing on the integration of CSR and sustainability within firms‟ strategy (Gond et al., 2012) and the identification of external pressures firms are facing regarding the implementation of CSR into strategic processes (Arjaliès & Mundy, 2013). With these studies focusing on the implementation of CSR in performance measurement targets, the effects of the implementation and use of CSR-related performance measurements remain largely unexplored. Hong et al. (2016) investigated the link between CSR-related targets and corporate governance structures, but did not investigate the effects of CSR-related targets either. Specifically, it is yet unknown whether CSR-related targets are related to CEO compensation amounts, which management control literature suggests should be the case. My study tries to add knowledge into this unexplored field of research by investigating the relationship between CSR-related targets in performance measurement systems and CEO compensation amounts, intending to provide a better base for understanding implications and effects of CSR activities by organizations on their
3 performance measurement system. My findings will also provide new insights into the governance literature, as well as help (investors) form expectations when organizations are starting to guide their management control- and compensation systems towards CSR-related targets. Using a British sample will also give insights into the use of CSR targets in an environment with a large focus on shareholders (García-Castro et al., 2008). This leads to a somewhat paradoxical situation in the sense that while firms showing strong and explicit CSR performance are believed to increase shareholder value (Cahan et al, 2015; Hong et al., 2016), achieving strong CSR performance asks for cooperation with- and considerations of other stakeholders beside the shareholders (Buysse & Verbeke, 2003; Madsen & Bingham, 2014). The „traditional‟ stakeholder- and shareholder-focused views seem to meet each other: this research will give insights into the use of a stakeholder-oriented concept (CSR) in a shareholder-focused business culture. Besides, this study also brings together two subjects having two very different sentiments in society: whereas CSR is considered to be positive and of increasing importance to society, CEO compensation is a fiercely debated topic creating tensions between firms and society (Connelly et al., 2016). Especially the rising gap between CEOs and average employees (so-called pay-dispersion) has grabbed the attention of (business) press and watchdog groups (Connelly et al., 2016). Coupling CSR and CEO compensation into one study may give interesting insights into a trade-off between two subjects with almost opposing sentiments in public opinion. Are the positive effects of CSR performance in the public eye mitigated if it results in CEO compensation being (too) high(er)? Questions like this could be an avenue of future research.
The main research question which will be addressed in this study is as follows:
“What is the effect of CSR-related performance targets of CEOs on the amount of compensation received by CEOs?”
Two competing explanations regarding the expected effect of CSR targets on CEO compensation amounts are formed. The first notion is grounded in agency theory and expects to find an indirect positive relationship between CSR targets and CEO compensation amounts through CSR performance. The presence of CSR targets is expected to lead to improved CSR performance, as these targets will offer motivation for CEOs to exert increased effort towards those targets, whilst this increase in effort will lead to improved performance (Bonner & Sprinkle, 2000). This increase in CSR performance in turn is expected to lead to an increase in CEO compensation, as CEOs get rewarded for achieving the predetermined CSR targets.
4 This expectation is supported by Madsen & Bingham‟s (2014) findings, which demonstrate a positive effect of strong Corporate Social Performance (CSP) on initial (first year) executive compensation.
The second notion for the expected effect of CSR targets on CEO compensation amounts focuses on a direct positive link between CSR targets and CEO compensation amounts, neglecting CSR performance as a potential mediator. CSR brings into play additional stakeholders (Ferrell et al., 2016), who will likely be pursuing different and often conflicting interests. As a result, managing these stakeholders will be challenging but important for firms looking to implement CSR and/or use CSR to their advantage. Considering this, firms can compensate their CEOs for the complex task of aligning and managing these stakeholders and exploiting CSR as a strategic tool and source of competitive advantage. This explanation of the hypothesized relationship has a resemblance to studies in the field of stakeholder theory (Benson & Davidson, 2010; Coombs & Gilley, 2005). Madsen & Bingham (2014) combined stakeholder theory and human capital theory and found that the positive effect between CSP strengths and executive compensation varies between different executive roles, depending on stakeholder-management responsibilities. Notably, they found this positive effect to become even stronger for CEOs compared to other executives, as CEOs will need to acquire more human capital to balance the interests of many different stakeholders.
Studying British FTSE100 firms‟ 2014 fiscal years, no relationship between CSR-related targets and CEO compensation was found, nor with CSR performance. However, if the mediating variable CSR performance is substituted by environmental performance, CSR-related targets were found to lead to superior environmental performance. Environmental performance was found to fully mediate the relationship between CSR-related targets and CEO compensation, creating an indirect relationship between CSR-related targets and CEO compensation.
The remainder of this study is organized as follows. Section 2 gives a broader theoretical background regarding executive compensation and CSR research, after which my hypotheses are formed. Methodology is discussed in section 3, providing detailed information about the sample, variables and data collection. Results are presented in section 4, while section 5 provides discussions and conclusions.
5
2. Literature review & hypothesis development
2.1 Agency theory and executive compensation
Executive compensation is mainly related to agency theory. Agency theory assumes individuals to be entirely self-interest and striving for the maximization of two utilities: wealth and leisure (Bonner & Sprinkle, 2002). Related, agency theory also describes the potential hazards which emerge when dealing with separation of ownership („principals‟) and management („agents‟) in organizations. When there is a conflict of interest between agents and principals, agents may take actions that are not in the best interest of the principals (Eisenhardt, 1989). Information asymmetry arises when agents have access to more (detailed) information than the principals and contributes to a moral hazards risk for shareholders (the „principals‟) (Ndofor et al., 2015). Since agency theory assumes entirely self-interested agents, this causes a potential risk of agents indulging in actions which are to their own benefit, but not to the benefit of the principals. This risk can be mitigated by providing incentives which lead to the alignment of the interests of the agent and the principals (goal congruence), like compensation sums and –structures (Baker et al., 1988).
2.2 CSR research in accounting
Many studies from differing streams of literature have taken CSR as their subject of study. In accounting literature, these studies have mainly focused on CSR reporting (Cormier & Magnan, 2003; Clarkson et al, 2008; Cahan et al., 2015). Aside from studies focusing on CSR reporting, only limited research has been conducted regarding the use of CSR in management control (systems). Building on Simons‟ Levers of Control (LOC) framework, Gond et al. (2012) came up with theorized roles of management control systems in the integration of sustainability within strategy of firms. Using the same LOC framework, Arjaliès & Mundy (2013) identified the need for CSR to be implemented into management control systems since firms are facing increasing pressure in relation to CSR management (for a full review of CSR research in accounting, see Huang & Watson, 2015). One aspects of this pressure concerns the implementation of CSR into strategic processes, in order for improvements in social and environmental performances to translate into long-term shareholder value (Arjaliès & Mundy, 2013). The results of their study indicate a confirmation of the strong role and influence of external stakeholders in setting strategic direction, as well as the establishment of objectives regarding CSR performance measurements. However, because of the large number of stakeholders, different views must be incorporated into processes and plans, exposing a risk of
6 prioritizing views of powerful stakeholders at the expense of other stakeholders (Arjaliès & Mundy, 2013). These findings provide a solid base for the further development of my study. Hong et al. (2016) studied the link between corporate governance and executive compensation for CSR. They found evidence suggesting CSR to be financially beneficial to firms and their shareholders. As a result, firms with shareholder-friendly corporate governance policieswere found to provide more incentives to their managers to engage in CSR. Shareholder-friendly corporate governance is achieved by the implementation of rules, practices and incentives to align the shareholders‟ and the firms‟ managers interests (Hong et al., 2016). Madsen & Bingham (2014) studied a sorting effect instead of an incentive effect and found initial, first-year compensation for executives to be positively related to strong Corporate Social Performance (CSP). The sorting effect of executive compensation suggests prospective executives have a certain demand about the level of compensation they want to receive when accepting to take an executive role within a firm. Subsequently, sorting effects can act as a device used to attract the most capable employee/executive (Cadsby et al., 2007). While the incentive effects of executive compensation can be considered an ex-post tool, sorting effects can be considered an ex-ante tool (Goldmanis & Ray, 2015). Madsen & Bingham (2014) also found compensation to vary across different executive roles, with CSP showing a stronger positive relationship with CEO compensation compared to other executives. According to Madsen & Bingham (2014), this can be explained by the need for a greater degree of human capital a CEO needs to acquire to successfully balance the interests of multiple stakeholders. Mahony & Thorn (2006) focused on compensation structures of Canadian firms and reported evidence suggesting effectiveness of compensation structures to focus executives‟ efforts on positive CSR actions. They found that both the value of stock options and bonuses are positively related to strong CSR performance, while salary amounts showed a positive relationship with weak CSR performance.
2.3 Hypothesis development
When CSR is considered to be a strategic objective and a source of competitive advantage creation, CSR can be assumed to be of importance for organizations and its principals, e.g. owners/shareholders. Increasing evidence is found supporting the claim that CSR conducts have a positive effect on firms´ financial performance and firm value (Coombs & Gilley, 2005; Cahan et al, 2013; Elliott et al., 2014; Hong et al., 2016; Ferrell et al., 2016). Reasoning from agency theory, this in turn should lead to CSR targets and goals being implemented in CEO performance measurement systems and subsequent compensation terms, in order to
7 ensure that the CEO (acting as the „agent‟) acts in the interest of the shareholders (acting as „principles‟) by aligning the interests of both parties. This should create a level of goal congruence between the CEO and the shareholders. Agency theory states that agents (or more generally: individuals) try to maximize two separate utilities: leisure and wealth (Bonner & Sprinkle, 2002). By setting specific targets the motivation of agents (CEOs) to act towards those targets can be increased, as long as one of these utilities can be satisfied without this satisfaction of one utility negatively influencing the other. As such, providing monetary incentives like compensation will increase the wealth utility when targets are met and result in increased effort towards the targets for which compensation is paid (Bonner & Sprinkle, 2002). This is supported by Sprinkle (2000) who found that reliance on incentive-based compensation improves performance of individuals by motivating them to increase the duration and intensity of their effort (Sprinkle, 2000). Considering this, the inclusion of CSR targets in the performance evaluation of CEOs will act as an incentive for CEOs to engage in CSR actions. The result of these CSR targets can be expected to be reflected in improved CSR performance of the firms as a whole, as increased effort will result in improved performance (represented with arrow „a‟ in figure 1) (Sprinkle, 2000; Bonner & Sprinkle, 2002). Empirical findings by Hong et al. (2016) further support this notion, showing that providing compensation linked to CSR leads to higher (average) social performance, consistent with the notion that CSR-related incentives lead to more CSR activities (Hong et al., 2016). Considering this, the following hypothesis is formulated below:
H1a: CSR-related performance targets in CEO compensation are positively related to firm CSR performance.
CEOs are expected to be rewarded for having achieved the predetermined CSR targets, resulting in a higher compensation (arrow „b‟ in figure 1). Higher compensation, based on incentives like CSR target setting, will predominantly be executed using variable compensation structures like bonuses (Mahoney & Thorn, 2006). In summary, the existence of CSR-related targets in the performance evaluation of CEOs is expected to lead to improved firm CSR performance, which in turn is expected to have a positive effect on the amount of compensation received by CEOs:
H1b: There is a positive indirect relationship between the inclusion of CSR-related performance targets of CEOs and the amount of CEO compensation acting through CSR performance.
8 CSR Performance
a b
c
CSR Targets CEO Compensation
Figure 1: Conceptual model.
A different, competing notion regarding the effect of the inclusion of CSR targets in the performance evaluation of CEOs on the amount of compensation received by the CEOs can be formed when looking at stakeholder theory (Huang & Watson, 2015). Stakeholders may impose pressure upon firms regarding CSR activities, influencing firms‟ environmental strategy and management control systems (Rodrigue et al., 2013; Pondeville et al., 2013; Arjaliès & Mundy, 2013). Adhering to the wishes- and taking care of the interests of stakeholders will be important, as ignoring them can cause disruptions in the relationship between the firms and its stakeholders, negatively affecting firm performance. The definitions of stakeholder theory and CSR closely resemble each other as well. Stakeholder theory attaches importance to all organizations‟ constituents, including suppliers, customers, employees, shareholders and the communities in which they operate (Benson & Davidson, 2010), which appears similar to the generally accepted description of CSR: giving attention to the broader social and environmental issues and contexts firms are part of. For this reason, prior research in the field of stakeholder theory can give insights which are of interest when researching the effect of CSR targets on CEO compensation amounts.
Management of stakeholders is the main independent variable in Coombs & Gilley‟s (2005) study. They investigated whether stakeholder management can serve as a predictor of CEO compensation. Building on stakeholder-agency theory (Hill & Jones, 1992), they propose that CEOs are compensated for maximizing value for stakeholders because it enhances the firms‟ effectiveness. This indicates recognition of the board of directors of the value of effective stakeholder management (Coombs & Gilley, 2005). However, they did not find a positive relationship, with the main effects being negative rather than positive. They conclude that CEOs are not being incentivized to pursue initiatives which are stakeholder related (Coombs & Gilley, 2005). In a related study, Benson & Davidson (2010) find that managers are not compensated for having good relationships with stakeholders, even though stakeholder management is positively related to firm value. These previous findings, however, have not studied the phenomenon of stakeholder theory with explicit attention to CSR. Stakeholder
9 theory and CSP are combined by Madsen & Bingham (2014), who find executives initial, first-year compensation to be positively related to strong CSP. They also find this relationship to be different for varying executive roles, with CEOs showing a stronger link between strong CSP and initial compensation than other executive roles because of the greater level of (general and firm-specific) human capital it requires to deal with many different stakeholders and to align their interests. Human capital concern the knowledge, skills, abilities and capabilities of people, allowing them to generate economic value and are built through experience accumulated over time (Madsen & Bingham, 2014). Human capital can be divided into general- and firm-specific human capital. General human capital can be transferred across firms, whereas firm-specific human capitals are unique to a specific firm and not transferrable across firms (Becker, 1964; Madsen & Bingham, 2014). Stakeholder-management responsibilities are thus a factor in the setting of executive‟s payment amounts (Madsen & Bingham, 2014). Following these findings, achieving strong CSR performance will require firms‟ cooperation with- and considerations of all of the firms‟ stakeholders. This results in a demand for strong stakeholder-managing capabilities of CEOs. Assuming firms are looking at CSR activities from a value-creating view, managing their stakeholders will be of importance and subsequently firms can be expected to reflect this in higher CEO payment amounts (illustrated by dotted arrow „c‟ in figure 1).
H2: There is a positive direct relationship between the inclusion of CSR-related performance targets of CEOs and the amount of CEO compensation because of higher
stakeholder-managing demands caused by CSR-related targets.
3. Methodology
My study is of quantitative nature, using archival data of the 2014 fiscal years of British FTSE100 firms. Since the study is focused on discovering effects on a broad scale of firms and tries to make inferences based on a sample, quantitative research based on archival data is deemed most appropriate. A cross-sectional picture is drawn of British FTSE100 index constituents, using a number of different data sources and collection methods. An overview of the variables used, their abbreviations/codes a short description and their sources can be found in table 1.
10 3.1 Sample selection
A sample of British firms was chosen for several reasons. First, since Anglo-American business culture (of which British firms are part) is primarily focused on shareholder considerations (García-Castro et al., 2008), applying CSR-related research can give interesting insights into the use of the stakeholder-related considerations demanded by CSR (Ferrell et al., 2016). Besides, if CSR activities are truly considered to be regarded as a valuable and value-creating opportunity, this should be especially reflected in actions of firms taking shareholder considerations as their top priority. While naturally all of the worlds‟ listed firms will devote considerable attention to shareholder considerations, Anglo-American firms are still regarded as taking this notion one step further compared to firms originating from other countries than the United Kingdom and the United States: Shleifer and Vishny (1997) (as cited by Farrell et al., 2016) describe Anglo-American corporate governance practices to be about “how investors get the managers to give them back their money” (p.738). The United Kingdom has been picked over the United States because of data-availability considerations and because British firms appear to be studied less often than American firms, which makes for a larger scientific void to fill using British firms. FTSE100 firms were chosen since these are the 100 largest listed firms of the country. Considering the large impact these big firms are making, it is interesting to understand them. Besides, using the biggest firms of the country also enhances data availability. The fiscal year 2014 was selected because at the time of this study data from 2014 provided the most complete recently available information. Still, because of data availability issues 9 firms had to be excluded from the analysis, resulting in a final sample size of 91 firms. This size of the sample should still prove to be sufficiently large, as only 9% of the original sample had to be excluded.
3.2 Independent variable
The independent variable concerns the existence of CSR-related CEO performance targets in the determination of CEO compensation amounts (CSR_Targets). These data have been hand-collected by studying corporate annual reports of the fiscal year 2014. Specifically, those parts of the annual reports called „Remuneration report‟ (or something similar) as part of the „Governance‟ section of the annual report have been subject to study. Hong et al. (2016) previously determined a number of descriptions which identified CSR-related targets (Hong et al., p. 210) which were used as guidance for coding CSR-related targets in this research as well. These targets focused on non-financial performance concerning topics such as (among others) safety, employee engagement, -well-being, environmental aspects, sustainability,
11 diversity and ethical compliance. A dummy variable has been created to specify the independent variable. Firms with CSR-related targets were assigned a „1‟, while firms who had not incorporated CSR-related targets received a dummy equaling „0.‟ More detailed information about the coding of CSR-related targets can be found in Appendix A.
3.3 Dependent variable
The dependent variable is total CEO compensation (Total_comp). Data concerning CEO compensation amounts have been provided by the BoardEx database. Specifically, the „Task 1878 UK Compensation 2014‟ Excel sheet output has been used. This sheet provides compensation information about all British listed firms‟ directors, both executive as well as non-executive directors. Naturally, only the information concerning CEO compensation was deemed useful for this research. The numbers in the „Total compensation‟ column in the „Summary‟ tab were used for providing information of CEO compensation amounts. These „Total compensation‟ numbers are the sum of two separate numbers defined as „Direct‟ and „Equity linked‟ compensation. „Direct‟ compensation constitutes of salary and bonuses, while „Equity linked‟ provides information regarding compensation linked to equity aspects like Long-Term Incentive Plans (LTIP), shares and options. It should be noted that BoardEx data does not consider (taxable) benefits, defined contribution plans (pensions) and „other‟ compensation as part of the „Total compensation‟ numbers. The exclusion of these numbers – which were commonly disclosed in annual reports – is because these numbers cannot be considered „true‟ compensation since they are a direct byproduct of- and related to primary compensation drivers like salary, or incidental expenses not related to performance. As such, they do not provide any valuable information about compensation amounts for this research. A number of firms in the sample are using fiscal book years which deviate from the calendar year. E.g., some firms‟ fiscal year 2014 ended on the 31st
of March 2015 or another date in 2015. Correspondingly, for 24 firms, compensation data had to be extracted from the „Task 1878 UK Compensation 2015‟ file. For fiscal years ending between the 30th
of June 2014 and the 31st of December 2014, the 2014 BoardEx data remained the source of compensation data. Reliability checks comparing the BoardEx data to the compensation information in annual reports (salary & bonuses) provided assurance of the use of correct compensation numbers. 3.4 Mediating variable
CSR performance is used as the mediating variable between CSR-related targets and CEO compensation amounts. CSR performance is constructed using environmental-, social- and
12 governance (ESG) scores from the Thomson Reuters Asset4-database, incorporated in Datastream. These ESG scores have been measured on the first month of the new fiscal year (2015), because these (yearly) updated scores will be based on the ESG performance of the previous fiscal year. Depending on the specific fiscal book years used by the sample firms, this meant that these measures took place at different points in time, but all immediately after the end of the respective firms‟ fiscal book years. The Asset4 ESG data is used by professional investors to define a wide range of responsible investment opportunities and by corporate executives to benchmark their firms‟ performance against industry peers (Thomson Reuters, 2015). ESG performance is measured in percentages based on 220 separate criteria related to environmental-, social- and governance practices with 100% being the maximal performance. In this study, I averaged these 3 ESG scores for every firm, resulting in 1 overall ESG score for every firm. Governance-, social-, environmental- and overall CSR performance are abbreviated to Gov_perf, Soc_perf, Env_perf and CSR_perf respectively.
3.5 Control variables
A number of control variables have been used in order to increase the confidence of the findings. Firm size has shown to have a significant effect on CEO compensation (Tosi et al., 2000), so this has been controlled for using 2014 net sales numbers. Besides, industry effects also have an influence on compensation strategies (Stroh et al., 1996). Industry has been controlled for by creating dummy variables of the Standard Industry Classification (SIC) codes‟ first two digits. Since CEOs main performance targets are related to financial performance, earning per share (EPS) has been used to control for the effect of financial performance on CEO compensation. Specifically, EPS excluding extraordinary items has been used as excluding extraordinary items better captures „core‟ financial performance, giving a less distorted indication of firms‟ financials. Board characteristics also have an effect on CEO compensation, since the boards of directors are responsible for the determination of CEO compensation amounts and structures (Van Essen et al., 2016). Boards‟ compensation committee independence (Comp_com_ind) levels thus have also been added as control variable, as compensation committees are specifically responsible for CEO compensation policies of the boards. Firm size (sales), - industry (SIC code) and –financial performance (EPS) have been composed from the Compustat Global database, while compensation committee independence data originates from Datastream.
13
Variables Abbreviation/code Description Source
CSR-related performance targets
CSR_Targets Dummy variable equaling 1 if
performance targets are related to CSR in the determination of CEO
compensation amounts.
Annual reports
CSR performance CSR_perf Scores relating to firms' performance regarding environmental, social and governance issues.
Asset4 / Datastream
CEO compensation amount Total_comp Total compensation received by CEOs, consisting of salary, bonuses and equity linked compensation.
BoardEx
Size Sales Firm size based on net sales. Compustat
Global Industry Industry Firms' industry based on SIC codes'
first two digits.
Compustat Global Financial performance EPS Firms' financial performance of the
year expressed in earning per share, excluding extraordinary items.
Compustat Global Compensation committee
independence
Comp_com_ind Percentage of independent board members on firms' compensation committee.
Datastream
Table 1: Overview of variables.
3.6 Treatment of outliers
In order to prevent skewed results, extreme outliers have been excluded from the analyses. Extreme outliers were identified using IBM‟s SPSS software, drawing boxplot graphs of each variable, including control variables. Outliers which were then marked with an asterisk by SPSS were deleted from the input sheets. In the case that one or more of the ESG (CSR) scores were excluded, the corresponding CSR performance score was excluded as well, since the CSR scores constitute of the average of the three separate ESG scores. Similarly, if one of the salary, bonus and LTIP variables were identified as being extreme outliers and subsequently excluded from the analysis, the corresponding „total compensation‟ number was excluded as well. Without excluding outliers, a significant T-statistic would have been found on Salary. All other statistical findings did not show any differences.
14
4. Results
Examining the results of this study, descriptive results will be presented and discussed first. After that, individual hypotheses are tested and presented using linear regression techniques. 4.1 Descriptive analysis
My descriptive analysis shows general observations (Table 2) and differences between firms with CSR-related targets and firms without, using statistical T-tests (Table 3). As can be seen from Table 3, firms with CSR-related targets on average appear to have higher governance-, environmental- and overall CSR performance. Firms with CSR-related targets also tend to provide higher CEO salaries, long-term incentive compensation and total compensation, as well as enjoy higher earnings per share. However, only environmental performance shows to be statistically significant (p = 0.021), whilst salary shows significance at the 10% level (p = 0.092). These tests have been conducted on a one-tailed basis in line with the hypotheses, except for the three control variables which have been tested two-tailed since no prior expectation was formed regarding the effect of CSR-related targets on these control variables. These preliminary results do not support any of the hypotheses. However, interesting signs appear with regards to the significant difference in environmental performance and salary amounts between firms with- and without CSR-related targets, which could indicate partial support for H1a and H2 respectively.
Variable N Mean Median St. deviation Minimum Maximum
CSR_Targets 91 0,51 1 0,50 0 1 Gov_perf 91 79,50% 84,70% 16,21% 21,33% 97,64% Soc_perf 90 79,99% 85,05% 14,65% 42,00% 96,44% Env_perf 91 78,99% 84,42% 16,95% 16,97% 95,04% CSR_perf 90 79,94% 85,09% 12,17% 40,91% 93,71% Salary (x1000) 89 £836,09 £831,00 £230,40 £188,00 £1.465,00 Bonus (x1000) 89 £795,79 £800,00 £656,77 £0,00 £3.590,00 LTIP (x1000) 86 £2.467,73 £2.213,50 £1.898,44 £0,00 £8.287,00 Total_comp (x1000) 83 £4.085,66 £3.790,00 £2.248,15 £393,00 £10.294,00 Sales (x1000) 87 £15.853.455,53 £6.982.000,00 £27.705.976,27 £207.000,00 £221.073.000,00 EPS 89 £75,76 £56,10 £81,71 -£195,72 £353,10 Comp_com_ind 89 94,49% 100% 11,23% 33,33% 100%
Table 2: Descriptive analysis.1
1 Sample sizes (N) vary because of the exclusion of (extreme) outliers in the data. See also ‘Treatment of
15
No CSR Targets CSR Targets
Variable Mean Mean T statistic Sig.
Gov_perf 78,67% 80,31% 0,480 0,316 Soc_perf 80,10% 79,88% -0,072 0,471 Env_perf 75,35% 82,55% 2,048 0,021 ** CSR_perf 78,92% 80,91% 0,770 0,220 Salary (x1000) £803,96 £868,95 1,337 0,092 * Bonus (x1000) £826,16 £767,39 -0,420 0,338 LTIP (x1000) £2.287,49 £2.631,96 0,839 0,202 Total_comp (x1000) £3.877,45 £4.279,35 0,812 0,210 Sales (x1000) £17.424.234,30 £14.318.376,27 -0,521 0,604 2-tailed EPS £71,35 £79,88 0,490 0,620 2-tailed Comp_com_ind 95,77% 93,30% -1,041 0,301 2-tailed Number of firms 45 46 * shows statistical significance at 10% (one-tailed) level.
** shows statistical significance at 5% (one-tailed) & 10% (two-tailed) level.
Table 3: T-test results.
Table 4 provides a Pearson-correlation coefficient matrix consisting of all variables. Environmental performance shows to be significantly correlated to CSR-related targets, which is consistent with the results of the T-tests in Table 3. Interestingly, salary significantly correlates with governance-, social-, environmental- and total CSR performance. Long term incentive plans (LTIP) also shows significant correlation with social-, environmental- and total CSR performance, while total compensation significantly correlates with social- and environmental performance but not with total CSR performance. CSR performance consists of the average of environmental, social and governance performance (ESG), so all correlations between ESG variables and total CSR performance suffer from multicollinearity. Multicollinearity also applies to the correlations between salary, bonus, LTIP and total compensation, since total compensation is the sum of salary, bonus and LTIP.
16
CSR_Targets Gov_perf Soc_perf Env_perf CSR_perf Salary Bonus LTIP Total_comp Sales EPS Comp_com_ind
CSR_Targets 1 Gov_perf 0,051 1 Soc_perf -0,008 0,564** 1 Env_perf 0,213* 0,360** 0,601** 1 CSR_perf 0,082 0,747* 0,893** 0,775** 1 Salary 0,142 0,189* 0,479** 0,370** 0,408** 1 Bonus -0,045 0,079 0,003 -0,023 0,029 0,229* 1 LTIP 0,091 0,130 0,245* 0,309** 0,257** 0,605** 0,279** 1 Total_comp 0,090 0,161 0,270** 0,280** 0,270 0,644** 0,494** 0,964** 1 Sales -0,056 0,104 0,166 0,087 0,133 0,346** -0,081 0,065 0,052 1 EPS 0,052 -0,028 -0,139 0,059 -0,068 -0,045 0,074 0,017 0,054 -0,071 1 Comp_com_ind -0,109 0,090 0,093 -0,016 0,090 0,029 0,094 -0,101 -0,059 0,070 -0,040 1
* shows significant correlation at 5% (one-tailed) level.
** shows significant correlation at 1% (one-tailed) level.
Table 4: Pearson-correlation coefficient matrix.
4.2 Hypothesis 1a
Table 5 provides linear regression results for hypothesis 1a. The effect of CSR-related targets has been tested on environmental-, social-, governance- and total CSR performance. Model 1 regresses all control variables on CSR performance and its components. Model 2 adds the independent variable CSR_Targets into the regression analysis. The results do not show support for H1a, as CSR_Targets does not show a significant effect on CSR_Perf. As such, CSR-related performance targets in CEO compensation does not lead to higher firm CSR performance. However, like in the previous T-tests and correlation results, a positive relationship between CSR_Targets and Env_Perf was found (p = 0.072, b = 6.425). Therefore, it can be concluded that CSR-related performance targets lead to higher firm environmental performance.
17
Dependent variable Env_perf Soc_perf Gov_perf CSR_perf
Model 1 2 1 2 1 2 1 2 Intercept 82,692 *** 76,522 *** 74,019 *** 74,583 *** 68,124 *** 66,220 *** 73,528 *** 71,767 *** (15,274) (15,431) (12,258) (12,639) (15,412) (15,866) (10,864) (11,169) Independent variable CSR_Targets 6,425 * -0,587 1,983 1,834 (3,520) (2,884) (3,620) (2,548) Control variables Sales 0,000 0,000 0,000 0,000 0,000 0,000 0,000 0,000 (0,000) (0,000) (0,000) (0,000) (0,000) (0,000) (0,000) (0,000) EPS 0,003 0,002 -0,034 -0,034 -0,006 -0,006 -0,016 -0,016 (0,022) (0,022) (0,018) (0,018) (0,022) (0,022) (0,016) (0,016) Industry -0,517 ** -0,493 ** -0,590 *** -0,593 *** -0,114 -0,106 -0,372 ** -0,365 ** (0,162) (0,161) (0,130) (0,131) (0,164) (0,165) (0,115) (0,116) Comp_com_ind 0,058 0,084 0,202 0,199 0,140 0,148 0,150 0,157 (0,161) (1,160) (0,130) (0,131) (0,163) (0,164) (0,115) (0,116) R² 0,123 0,159 0,244 0,244 0,024 0,028 0,139 0,145 ΔR² 0,036 0,000 0,004 0,006 F-statistic 2,814 ** 2,983 ** 6,450 *** 5,106 *** 0,489 0,448 3,237 ** 2,677 ** N 91 90 91 90
* shows significance at 10% level (two-tailed) ** shows significance at 5% level (two-tailed)
*** shows significance at 1% level (two-tailed)
Table 5: Linear regression results of CSR targets on CSR performance and its components of environmental-, social-, and governance performance.
4.3 Hypothesis 1b
Linear regression analysis results for hypothesis 1b are provided in Table 6. Hypothesis 1b predicted a positive indirect relationship between CSR-related performance targets and the amount of CEO compensation, acting through CSR performance. The mediated regression analysis consisted of three steps. First, CSR_Targets was regressed on CEO compensation (Total_comp), investigating a direct relationship between CSR-related targets and CEO compensation. No significant effect was found. The second step regressed CSR_Targets on CSR performance (CSR_perf), which also did not result in a significant effect. Lastly, CSR_Targets and CSR_perf were regressed on Total_comp. The models labelled „model 1‟ regressed all control variables on the dependent variables, while „model 2‟ added the independent variables. No support for H1b was found, as CSR_Targets shows no significant relationship with neither total CEO compensation (Total_comp) nor CSR performance (CSR_perf). CSR performance does show a significant relationship with total CEO compensation while controlling for CSR targets (p = 0.066, b = 41.614), indicating support for the „b-path‟ of figure 1 but no support for a mediating effect of CSR performance between
18 CSR targets and total CEO compensation has been found, even though coefficient of CSR_Targets on Total_comp did decrease from 301,477 to 225,172.
Dependent variable Total_comp CSR_perf Total_comp
Model 1 2 1 2 1 2 Intercept 5291,882 ** 5002,360 ** 73,528 *** 71,767 *** 5291,882 ** 2015,836 (2182,620) (2246,926) (10,864) (11,169) (2182,620) (2426,431) Independent variable CSR_Targets 301,477 1,834 225,172 (512,630) (2,548) (505,735) CSR_perf 41,614 * (22,258) Control variables Sales 0,000 0,000 0,000 0,000 0,000 0,000 (0,000) (0,000) (0,000) (0,000) (0,000) (0,000) EPS 0,616 0,555 -0,016 -0,016 0,616 1,233 (3,149) (3,165) (0,016) (0,016) (3,149) (3,133) Industry -45,267 * -44,150 * -0,372 ** -0,365 ** -45,267 * -28,948 (23,194) (23,374) (0,115) (0,116) (23,194) (24,380) Comp_com_ind -4,539 -3,316 0,150 0,157 -4,539 -9,860 (23,063) (23,374) (0,115) (0,116) (23,063) (23,138) R² 0,058 0,063 0,139 0,145 0,058 0,106 ΔR² 0,005 0,006 0,048 F-statistic 1,143 0,975 3,237 ** 2,677 ** 1,143 1,423 N 83 90 83
* shows significance at 10% level (two-tailed) ** shows significance at 5% level (two-tailed)
*** shows significance at 1% level (two-tailed)
Table 6: Linear regression results of the mediating effect of CSR performance on the relationship between CSR targets and CEO compensation.
Since Table 5 showed evidence of a relationship between CSR targets and environmental performance, a similar three-step mediated regression analysis to the one shown in Table 6 has been performed using environmental performance as a potential mediator, substituting CSR performance (Env_perf instead of CSR_perf in Table 6). Results of this regression are presented in Table 7. Interestingly, these results provide evidence for both a relationship between CSR targets and environmental performance (the adjusted „a-path‟ in figure 1, replacing CSR performance with environmental performance as previously mentioned in the results of H1a), as well as a relationship between environmental performance and total CEO compensation while controlling for CSR targets (the adjusted „b-path‟ in figure 1) (p = 0.075,
b = 29.119). In other words, there appears to be existence of indirect full mediation, creating
an indirect relationship between CSR targets and total CEO compensation through environmental performance: no direct relationship between the dependent and independent variable is found, but significant relationships exist between the independent and mediating
19 variable, as well as the mediating and dependent variable while controlling for the independent variable. Besides, the coefficient of CSR_Targets on Total_comp decreases when Env_perf is included as a mediator (from 301,477 to 114,385), which is another precondition for mediation.
Dependent variable Total_comp Env_perf Total_comp
Model 1 2 1 2 1 2 Intercept 5291,882 ** 5002,360 ** 82,692 *** 76,522 *** 5291,882 ** 2774,091 (2182,620) (2246,926) (15,274) (15,431) (2182,620) (2534,124) Independent variable CSR_Targets 301,477 6,425 * 114,385 (512,630) (3,520) (515,418) Env_perf 29,119 * (16,135) Control variables Sales 0,000 0,000 0,000 0,000 0,000 0,000 (0,000) (0,000) (0,000) (0,000) (0,000) (0,000) EPS 0,616 0,555 0,003 0,002 0,616 0,509 (3,149) (3,165) (0,022) (0,022) (3,149) (3,117) Industry -45,267 * -44,150 * -0,517 ** -0,493 ** -45,267 * -29,786 (23,194) (23,374) (0,162) (0,161) (23,194) (24,358) Comp_com_ind -4,539 -3,316 0,058 0,084 -4,539 -5,761 (23,063) (23,374) (0,161) (1,160) (23,063) (22,948) R² 0,058 0,063 0,123 0,159 0,058 0,103 ΔR² 0,005 0,036 0,045 F-statistic 1,143 0,975 2,814 ** 2,983 ** 1,143 1,381 N 83 91 83
* shows significance at 10% level (two-tailed) ** shows significance at 5% level (two-tailed)
*** shows significance at 1% level (two-tailed)
Table 7: Linear regression results of the mediating effect of environmental performance on the relationship between CSR targets and CEO compensation.
In order to advance the understanding of the effect of CSR-related targets on CEO compensation, more detailed results can be found in Table 8, which regresses environmental performance on the three components of total compensation while controlling for CSR targets. Environmental performance was found to also fully mediate the relationship between CSR targets and salary (p = 0.024, b = 3.21) and LTIP (p = 0.026, b = 30.181), but no significant effect on bonus was found.
20
Dependent variable Salary Bonus LTIP Total_comp
Independent variable CSR_Targets 42,432 -52,807 70,094 114,385 (44,537) (150,804) (423,877) (515,418) Env_perf 3,210 ** 0,206 30,181 ** 29,119 * (1,394) (4,721) (13,270) (16,135) * shows significance at 10% level (two-tailed)
** shows significance at 5% level (two-tailed)
A-path: CSR_targets on Env_perf p = 0.072, b = 6.425.
Table 8: Linear regression results of multiple adjusted „b-paths,‟ regressing environmental performance on compensation components.
4.4 Hypothesis 2
Hypothesis 2 predicted a positive direct relationship between CSR-related performance targets and the amount of CEO compensation because of higher stakeholder-managing demands caused by CSR-related targets. Linear regression results for H2 are presented in Table 9. Again, the models labelled „model 1‟ regress the control variables on the dependent variables. „Model 2‟ adds the independent variable in the regression analysis. Table 9 shows that CSR_Targets is not significantly associated with Total_comp, nor its components of salary, bonus and LTIP. Therefore, H2 is not supported.
Dependent variable Salary Bonus LTIP Total_comp
Model 1 2 1 2 1 2 1 2 Intercept 826,538 *** 765,984 *** 199,343 248,788 4215,541 ** 3962,007 ** 5291,882 ** 5002,360 ** (192,758) (196,357) (624,304) (643,386) (1816,504) (1869,241) (2182,620) (2246,926) Independent variable CSR_Targets 63,054 -51,487 264,004 5002,360 (44,798) (146,786) (426,461) (2246,926) Control variables Sales 0,000 * 0,000 * 0,000 0,000 0,000 0,000 0,000 0,000 (0,000) (0,000) (0,000) (0,000) (0,000) (0,000) (0,000) (0,000) EPS -0,209 -0,222 0,638 0,648 -0,213 -0,267 0,616 0,555 (0,278) (0,277) (0,901) (0,906) (2,621) (2,633) (3,149) (3,165) Industry -7,410 ** -7,177 ** 2,775 2,585 -31,187 -30,208 -45,267 * -44,150 * (2,048) (2,043) (6,634) (6,693) (19,303) (19,445) (23,194) (23,374) Comp_com_ind 1,349 1,605 5,548 5,339 -12,604 -11,534 -4,539 -3,316 (2,037) (2,033) (6,597) (6,660) (19,195) (19,349) (23,063) (23,259) R² 0,244 0,262 0,024 0,025 0,048 0,053 0,058 0,063 ΔR² 0,018 0,001 0,005 0,005 F-statistic 6,451 *** 5,621 *** 0,490 0,413 0,972 0,848 1,143 0,975 N 89 89 86 83
* shows significance at 10% level (two-tailed)
** shows significance at 5% level (two-tailed)
*** shows significance at 1% level (two-tailed)
21
5. Discussion & conclusions
The objective of this study was to improve the understanding of the effects of the use of CSR in management accounting and –control (systems). Since previous CSR research has primarily focused on CSR reporting (Cormier & Magnan, 2003; Clarkson et al., 2008; Cahan et al., 2016) and the implementation of CSR in performance measurement systems (Gond et al., 2012; Arjaliès & Mundy, 2013), the effects of the implementation and use of CSR-related performance measurements has remained largely unexplored. This research has tried to increase knowledge of this unexplored field of research, by investigating a relationship between CSR-related performance targets and CEO compensation amounts. Two competing notions have been adopted regarding this relationship. Drawing from agency theory, an indirect positive relationship between CSR-related targets and CEO compensation amounts was expected through higher CSR performance. Neglecting CSR performance and taking a stakeholder-theory view, the second notion expected a direct positive relationship between CSR-related targets and CEO compensation amounts because of higher stakeholder-management demands caused by CSR-related targets. British FTSE100 firms were studied using their 2014 fiscal years. These firms‟ annual reports were studied, after which a dummy variable was created, equaling „1‟ if CSR-related targets were included in the performance measurement of the firms‟ CEO. Data was then analyzed using linear regression techniques. The findings show no support for the hypotheses. CSR-related targets do not lead to higher CSR performance (H1a), thereby immediately making it impossible for CSR performance to be a mediator in the relationship between CSR-related targets and CEO compensation amounts (H1b). Besides, no direct relationship between CSR-related targets and CEO compensation amounts was found (H2). One reason why the findings show no support for the hypotheses could be because of the limited influence CSR-targets were found to have in the performance measurement system influencing CEO compensation amounts. For example, in most of the cases in which CSR-related targets were found to be present, these targets were part of some kind of „Personal objectives‟ set by the compensation committee. These „Personal objectives‟ generally accounted for a maximum of 25% of the total compensation objectives, with financial measures being the primary domain of performance target setting. CSR-related targets further decrease in importance since „Personal objectives‟ encompass a variety of targets, with CSR-related targets being one of a number of „Personal objectives.‟ Other variations of target setting incorporating CSR-related targets were all found to give limited attention to CSR-related targets. When CSR-related targets are only given a relatively
22 small share of the total targets upon which compensation is determined, the potential effect of CSR targets on CEO compensation amounts will automatically be limited. But why are CSR-related targets given such little attention in the performance measurements systems determining CEO compensation? Given that a large number of CSR studies in accounting have focused on CSR reporting, a number of critics have also emerged regarding CSR reporting by firms. One such critical issue views CSR reporting by firms as a tool helping to improve their legitimacy (Milne & Gray, 2007; Cho et al., 2012). In such cases, CSR reporting merely functions as a way to influence the way firms are perceived in society (Lindblom, 1993; Suchman, 1995). Ongoing activities (acting towards CSR) are limited and decoupled from the formal structures of CSR reporting (Meyer & Rowan, 1977). The fact that generally only relatively little attention is paid towards CSR-related targets in the performance evaluation of CEOs could be an extension of the legitimacy-seeking notion of CSR: some attention to CSR-related targets is paid, increasing firms‟ legitimacy and showing they take CSR seriously by embedding CSR in executive actions – but is only limited in order to prevent executives from being distracted on focusing on (non-CSR) performance targets which the firm feels are truly important. The lack of support for the hypotheses could also be explained by a firms‟ possible general lack of importance given to CSR, while still believing CSR adds value. The inclusion of CSR-related targets not yielding higher CEO compensation could simply be explained by firms not believing higher compensation amount are justified, i.e. CSR adding value, but not yielding high enough competitive or financial advantages to compensate their CEOs for CSR, nor believing higher stakeholder-managing demands caused by CSR actions justify higher compensation.
However, while CSR-related targets did not lead to higher CSR performance nor higher CEO compensation, further analysis of the results has shown CSR-targets to lead to higher environmental performance. Further, higher environmental performance has shown to lead to higher CEO compensation, showing evidence of indirect full mediation by environmental performance on the relationship between CSR-related targets and CEO compensation amounts. Had the hypotheses formed in this study focused on environmental performance instead of CSR performance, support would have been found for hypotheses 1a and 1b. This appears remarkable, because it contradicts all previous findings of this study which have focused on total CSR performance. A potential cause of this can be that if CSR-related targets are implemented in the performance measurements determining CEO compensation, these targets mainly focus on environmental targets, putting less emphasis on social- and
23 environmental targets. The central research question of this study was as follows: “What is
the effect of CSR-related performance targets of CEOs on the amount of compensation received by CEOs?” Answering this question in its most straightforward way, we can
conclude that CSR-related performance targets of CEOs do not have an effect on the amount of compensation received by CEOs – neither a direct relationship nor an indirect relationship mediated by CSR performance. A potential direct effect was examined from a stakeholder-managing perspective caused by CSR-related targets. The lack of a direct relationship is consistent with previous findings in literature, which found that CEOs are not being incentivized to pursue stakeholder-related initiatives (Coombs & Gilley, 2005) and managers are not compensated for having good relationships with stakeholders (Benson & Davidson, 2010). On the other hand, these findings contradict Madsen & Bingham‟s (2014) findings, who found that stakeholder-managing responsibilities are an important factor in determining CEOs first-year compensation amounts. Since my study did not focus on first-year compensation but compensation for all FTSE100 CEOs in a particular year, this incentive effect of compensation could be different to the sorting effect of compensation as studied by Madsen & Bingham (2014). For example, firms may only take the stakeholder-managing responsibilities into consideration when they are looking for a new CEO.
However, environmental performance functions as a mediator in the relationship examined in the central research question, creating an indirect effect between CSR-related targets and CEO compensation amounts: CSR-related targets result in superior environmental performance, with environmental performance resulting in higher CEO compensation amounts. Interestingly, superior environmental performance was not found to have a significant relationship with CEO bonuses, but LTIP and salary instead. This indicates that compensation rewarded to CEOs due to superior environmental performance is more or less fixed (salary) and long-term oriented (LTIP) and not based on short-term bonus structures. This finding contradicts Mahoney & Thorn (2006), who state that incentives like CSR would mainly be executed using variable compensation structures, most notably using bonuses. No higher social performance was found to be resulting from CSR-related targets. These findings contradict Hong et al. (2016) and Mahoney & Thorne (2005), who found CSR-related targets to lead to higher social performance.
5.1 Contribution and implications
Overall, this study has tried to identify some of the effects of the inclusion of CSR-related targets in the performance measurement systems determining CEO compensation. Previous
24 research into CSR in accounting literature has mainly focused on CSR reporting and the way CSR is integrated into performance measurement- and management control systems. This study has taken a more general, larger look on CSR integration in performance measurement systems and found a number of observations which so far had not been found in previous studies, focusing on the effects of the integration of CSR-related targets in performance measurement- and management control systems. The agency theory notion of CSR – specifically environmental performance – seems to prevail: CSR-targets lead to higher environmental performance, for which CEOs get rewarded.
A number of practical implications have been found by this study. First, linking CSR-related targets to compensation does not automatically result in superior CSR performance to firms which have not linked CSR targets to compensation. More incentives appear to be needed in order to yield superior CSR performance than merely linking CSR-related targets to compensation. On the other hand, linking CSR-related targets to compensation does result in superior environmental performance. When CSR targets are linked to compensation, these targets generally lack attention and form a relatively small part of the total (non-CSR) targets upon which CEO compensation is determined. Still, making predictions regarding CEO compensation amounts when firms are guiding their compensation systems towards CSR-related targets need to be performed carefully, since this study does not give a conclusive answer on the compensation effects of CSR-related targets. The topic remains very open to further studies and debate.
5.2 Limitations
A number of limitations are present in this study. Annual reports were screened looking for targets related to CSR. However, the presence of these targets does not say anything about the realization of these targets by the CEOs. Compensation will likely only be awarded when targets are met. Since the mere presence of CSR-related targets does not say anything about the fact whether these targets were achieved, drawing conclusions about the relationship between CSR-related targets and CEO compensation can become tricky. Second, targets relating to CSR were not divided into different sub-targets in this study, as this idea only became clear to me after completing the content analysis. Performing content analysis again would have been necessary, but would also cause time constrains, so I elected to refrain from splitting CSR-related targets into different sub-targets. Though, accuracy could have been improved if targets were split into the three different aspects of CSR (ESG) performance: environmental, social and governance aspects. Had this been done, more detailed results
25 would have been produced. For example, this study found CSR targets to lead to higher environmental performance. However, it is unknown whether this was because of targets relating to the environment specifically. All we know is that CSR-related targets are related to environmental performance, which is more general than knowing which specific aspect of CSR-related targets formed the relationship with environmental performance. Next, some of the regression results yielded low, insignificant F-statistic values. While this does not cause immediate concern, higher F-statistics would have improved the validity of the findings. Since CEO compensation can be expected to be based on the culmination of specific targets, concerns regarding reverse causality appear low: CSR-related targets being implemented because of higher compensation amounts appear illogical, however, reverse causality concerns cannot be fully mitigated. Lastly, a number of variables which relate to CEO compensation and CSR performance may have been missing in my study. Although all (control) variables have been selected carefully, the risk of missing variables is still present. Looking back, specifically CEO duality could have been included as a control variable, as CEO duality has been found to be an important determinant in CEO compensation (Shin, 2016; Van Essen et al., 2015).
5.3 Future research
To further improve our understanding of the effects of CSR-related targets for CEOs, a number of avenues of future research are identified. First, to confirm the findings of my study, similar studies can be performed which take a different measurement of CSR performance. While this study uses Asset4 ESG scores, an alternative measure of CSR performance is the KLD Research & Analytics MSCI KLD Scores, which dissects five key ESG scores: environment, community & society, employee & supply chain, customers and governance & ethics (Network for Business Sustainability, 2013). Next, a larger sample could be used to further investigate the relationship between CSR-related targets and CEO compensation, in different settings and contexts other than the British setting of this study. Next to larger samples, more in-depth analysis is needed as well. Case studies could be used to explore firms‟ motives why CSR-related targets remain relatively small in the total package of targets upon which CEO compensation is based. Improved measurement of CSR-related targets can also be achieved by in-depth case studies, instead of relying on public information provided in annual reports and using a proxy for CSR-related targets like in my study. As outlined in the introduction section of this study, another area of potential interesting results is the trade-off between CSR and CEO compensation. With CSR being considered positive and important to
26 society, (excessive) CEO compensation is a topic creating tension between firms and society, especially the rising pay-dispersion between CEOs and average employees (Connelly et al., 2016). Perhaps CSR-related targets are not being translated into higher CEO compensation for this reason: positive effects of CSR actions by firms risk being mitigated by the negative effects (reputational harm) of increased CEO compensation caused by CSR-related targets and high CSR performance. My hope is that new studies into the use- and effects of CSR-related performance targets in performance measurement systems will emerge to further develop our knowledge into an area of research we do not know much about yet.
27 References
Arjaliès, D.L., Mundy, J. (2013). The use of management control systems to manage CSR strategy: A levers of control perspective. Management Accounting Research, 24(4), 284-300. Baker, G.P., Jensen, M.C., Murphy, K.J. (1988). Compensation and incentives: Practice vs. Theory. The Journal of Finance, 43(3), 593-616.
Becker, G.S. (1962). Investment in human capital: A theoretical analysis. Journal of Political
Economy, 70(5), 9-49.
Benson, B.W., Davidson, W.N. (2010). The relation between stakeholder management, firm value, and CEO compensation: A test of enlightened value maximization. Financial
Management, autumn 2010, 929-963.
Bonner, S.E., Sprinkle, G.B. (2002). The effects of monetary incentives on effort and task performance: theories, evidence and a framework for research. Accounting, Organizations
and Society, 27, 303-345.
Buysse, K., Verbeke, A. (2003). Proactive environmental strategies: A stakeholder management perspective. Strategic Management Journal, 24, 453-470.
Cadsby, C.B., Song, F., Tapon, F. (2007). Sorting and incentive effects of pay for performance: An experimental investigation. Academy of Management Journal, 50(2), 387-405.
Cahan, S.F., De Villiers, C., Jeter, D.C., Naiker, V., Van Staden, C.J. (2015). Are CSR disclosures value relevant? Cross-country evidence. European Accounting Review, 25(3), 579-611.
Cho, C.H., Michelon, G., Patten, D. (2012). Impression management in sustainability reports: An empirical investigation of the use of graphs. Accounting and the Public Interest, 12, 16-37.
Clarkson, P.M., Li, Y., Richardson, G.D., Vasvari, F.P. (2008). Revisiting the relation between environmental performance and environmental disclosure: An empirical analysis.
Accounting, Organizations and Society, 33(4/5), 303-327.
Connelly, B.L., Haynes, K.T., Tihanyi, L., Gamache, D.L., Devers, C.E. (2016). Minding the gap: Antecedents and consequences of top management-to-worker pay dispersion. Journal of
