Master thesis
Corporate Social Responsibility
The effect of short- and long- term executive remuneration on CSR and the
change during the financial crisis: an empirical study
E.T. Majoor
Master in Business Economics, Finance 5745594
July, 2014 M. Dijkstra
University of Amsterdam
Abstract
There has been an increase in socially responsible programs included in overall business strategies (Margolis et al., 2009). An indication of this increase is the fact that over 50% of Fortune 1000 companies now issue CSR reports on a yearly basis and companies are being checked by stakeholders if they include CSR related actions. This thesis examines if the board of directors can direct an executive towards more socially responsible actions by using short- or long-term incentives as a motivation. The effect of executive remuneration on CSR is researched by running linear regressions on companies in the US during and after the financial crisis that started in 2007. Executive
remuneration can be seen as the motivation for executives to work hard and to achieve the goals that are set by the directors (Kruger, 2009). In this retrospect executive compensation can be regarded as a steering tool towards the wishes of the board (McGuire et al, 2003; Mahoney and Thorn, 2006). Salary has a positive effect on net CSR strength. When salary rises with 1 million UDS, net CSR strengths rises with 0.27. However, the short-and long-term incentives have no significant effect on CSR. To examine the effect of the recent financial crisis three dummy variables were created for the interactions between salary, bonus and stock options with the crisis. Both interactions are not statistically significant and do not have a significant effect on the independent variable net CSR strengths. The conclusions are robust for the type of model (linear, ordered logistic regression), the specification of the time variable, and for net CSR strength excluding negative economic impact from the KLD scale.
Table of Contents
1. Introduction... 5
2. Literature Review ... 8
2.1 Theoretical motivation ... 8
2.1.1 Corporate Social Responsibility ...8
2.1.2 CSR and executive compensation ...11
2.2 Empirical motivation...12
2.3 CSR and the financial crisis ...15
3.
Hypothesis development and methodology ... 17
4. Data ... 18
4.1 Sample selection ...18
4.2 Variable description ...20
4.2.1 Dependent variable: net CSR strengths...21
4.2.2 Explanatory variables compensation...25
4.2.3 Control variables: Firm Size, ROE and Debt to Equity ...26
4.3 Industry level controls...27
4.4 Time dependent variables ...28
5. Results ...
285.1 Univariate analysis...28
5.2 Regression Analysis...29
5.2.1 Effect of variable compensation on CSR ...29
5.2.2 CSR before and during the financial crisis...30
5.2.3 The crisis and compensation structure ...32
5.3 Robustness check ...32
5.3.2 Robustness check for the effect of time on CSR ...34
5.3.3 Robustness check for negative economic impact in the KLD scale ...37
6. Conclusion
...397. Discussion... 39
Bibliography ...
40Appendix...
421. Introduction
The concept of Corporate Social Responsibility (CSR) began to culminate starting in the 1970’s (Wirl et al., 2013). Shareholders but also governments, entire communities and employees started to hold companies accountable for their actions and the social and environmental impact they have on society by mandating companies to issue CSR reports (Porter and Kramer, 2006). In 2008 more than 50% of the Fortune 1000 companies issue CSR reports on a regular basis, and 10% of the investments made in the US are checked if they encompass CSR related issues (El Goul et al., 2011). For twenty years KPMG has conducted a survey of Corporate Responsibility Reporting and in 2013 they published the eighth edition. In 1993 the first edition that KPMG published covered only 10 companies and the significant increase to 4100 companies across 41 countries covered in the last survey indicates the heightened attention that is given to CSR (KPMG Corporate Responsibility Reporting Survey, 2013). Compensation matters for CSR because it is the reward for the work that an executive has performed in the past year. Executive remuneration can be seen as the motivation for executives to work hard and to achieve the goals that are set by the directors (Kruger, 2009). In this retrospect executive compensation can be regarded as a steering tool towards the wishes of the board. Socially responsible business strategies are increasingly on the minds of the directors (Porter and Kramer, 2006). The board of directors can use compensation to steer an executive more towards socially responsible strategies. This thesis will examine if the variable portion of executive remuneration, bonus as a short-term incentive and stock options as a long-short-term incentive, can be used to steer executives towards more socially responsible actions.
This research will answer the following research question: What is the effect of short- and long -term executive remuneration on CSR? And has this changed during the recent financial crisis?
According to the KPMG Corporate Responsibility Survey (2013) 10 percent of the firms mention a relationship between CSR and executive compensation. And 5 percent clearly report the effect that environmental catastrophes and social risks can have on the performance of a firm.
According to Porter and Kramer (2006) companies were forced by the public to be more socially responsible because environmental and social issues are on the top of the minds of consumers and they are looking for companies that have ‘doing good’ in their strategy. Examples of this forced socially responsible awareness are the pressure from activist groups and journalists that led to a boycott of Nike, in the US in the 1990’s, due to abusive labour practices, forced pharmaceutical companies from all over the world to help fight the AIDS epidemic (also in the 1990’s) and made fast-food companies accountable for the increasing obesity in countries in 2000 (Porter and Kramer, 2006).
When a firm is engaged in CSR it means that the company takes both the shareholder’s and stakeholder’s interests into account (Wirl et al., 2013). A manager is said to be socially responsible when he not only takes into account his own comfort but also the wishes and demands of the entire firm, consumers, employees, the environment and the community (Carrol, 1999). So instead of only paying attention to short-term profits, managers keep in mind the long-term effects of the firm’s actions and act in the best interest of all stakeholders of a firm (Mahoney and Thorn, 2006).
Investments in CSR will payoff in the long-term according to Kruger (2009). According to Chatterji et al. (2008) the motivation for firms to engage in CSR is that the company will have better financial performance. The most common motivation for firm to engage in CSR is that it will enhance their relationship with shareholders because of lower costs of capital and less risk due to increased
transparency. Other reasons are that CSR attracts socially responsible consumers, it attracts motivated and highly educated employees and that engaging in more CSR will decrease the threat of activists and NGO’s that put pressure on firms to be more socially responsible (Chatterji et al. 2008). Investors realize that there is a correlation between firm performance and society (Carrol, 1999). Porter and Kramer (2006) also link performance and society by stating that a healthy firm needs a healthy society and vice versa. CSR is a tool to create shared value, both for the firm as for society, and to sustain the shared value the points of intersection between a company and society have to be found (Porter and Kramer, 2006). Preston and O’Bannan (1997) find that financial performance often precedes or runs simultaneous with social impacts. This indicates that social performance is boosted by money that is available or by positive coactions (Preston and O’Bannan, 1997). Sprinkle and Maines (2010)) states that firms engage in CSR to reduce the risk that is associated with all
management efforts. For example by improving working conditions or reducing gas emissions firms circumvent possible regulations or legal implications. Because shareholders can influence the corporate progress it is in the best interest of the firm to pursue non-financial goals that are aligned with the wishes of the shareholders and stakeholders (Kruger, 2009).
The variable portion of executive remuneration, stock options and bonus, are used to align the interests of the executive with that of the community and environment. Haigh and Jones (2006) state that to make the accounting models of the firm take CSR into account, and not only superficial mention CSR in their yearly business reports, the remuneration package has to force managers to keep in mind the negative influences that they can generate.
This aim of this thesis is to get a comprehensive look at whether CSR is influenced by either long- or short-term compensation in the form of stock options and bonus. The focus is laid on executive remuneration because a CEO is the one who is responsible for the performance of the firm and who has to report to the board of directors. It is interesting for the board of directors to know if they can direct the CEO towards more social goals by changing the amount of the bonus or stock options that
the executive receives at year-end. Mahoney and Thorn (2006) find that there is a positive relationship between bonus and CSR and between stock options and CSR. In their study they extend a study of McGuire et al. (2003) in which a regression is performed to find the effect of executive compensation on CSR strengths and CSR concerns in Canada.
This thesis is an extension on the researches of Mahoney and Thorn 92006) and McGuire et al. (2003) where the data for CSR is retrieved from the KLD (Kinder, Lydenberg and Domini) Research and Analytics Environmental and social performance database and the dataset that is used is from 2001 until 2012. The KLD database gives companies a rating for CSR strengths and concerns based on internal assessments. The dependent variable net CSR strengths is calculated by subtracting total CSR concerns from total CSR strengths. The data for executive remuneration is retrieved from Compustat and Execucomp also present in WRDS. The effect of the recent financial crisis on CSR will be examined by creating a dummy variable for the crisis years and by including interaction effects of the crisis with the salary, bonus and stock options. A regression will be run to see if there is a correlation between the different elements of executive remuneration and the amount of CSR a company exerts. As an addition on the effect of executive remuneration on CSR, the change in CSR during the recent financial crisis is also examined. First the link between CSR and executive remuneration will be elaborated on and then the effect of the most recent financial crisis, which started in 2007, on CSR will be studied.
In the first chapter the theoretical and empirical motivation of the effect of executive remuneration on CSR will be introduced in the literature review. In this section empirical evidence will be discussed that are inline with the existing theories. In chapter 2 the hypothesis will be stated and the
methodology will be provided. This section will include the type of data that will be used and what economic model will be used. Chapter four will include the data and descriptive statistics and chapter five will be the results that follow from the regression that is run, it will also include several
robustness checks that are performed. Chapter 6 will include the conclusion and the final chapter of this thesis includes the discussion.
2 Literature Review
In this chapter first the theoretical motivation for the link between executive remuneration and CSR is given and secondly the empirical relationship is elaborated on. The final section of this chapter includes a theoretical and empirical approach on the effect of the recent financial crisis on CSR.
2.1 Theoretical motivation
2.1.1 Corporate Social Responsibility
Carroll (1999) mentions that the start of the formation of the concept of CSR originated in the 1950’s with a publication of economist Howard Bowen who asked what the social responsibilities of businessman were for society. The notion of CSR started without the corporate connotation because corporations were not yet dominant in the business sector. In 2012 71 % of the listed companies worldwide published a corporate social responsibility report (KPMG, 2013). According to Margolis et al. (2009) since 1972 214 articles, books and working papers have been written about CSR or CSR related effects. Bowen’s (1950) definition of CSR was built around the concept that businessman should try to satisfy society by designing strategies that are in line with the wishes of society (Carroll, 1999). The concept of CSR moved from an awareness of businesses to recognize the relationship between firms and society towards a theoretic concept now widely used and applied in corporations throughout the world (Carroll, 1999). The core of the definitions of CSR is that it entails decisions of which the effect is beyond only economic benefits and includes responsibilities towards society (Carroll, 1999). The European Commission (2010) states that CSR is: “a concept whereby companies integrate environmental and social concerns in their business strategies and in their interaction with their stakeholders on a voluntary basis”.
Most definitions of CSR define the role a business plays in society (Giannarakis and Theotokas, 2011).There are several reasons why companies may engage in CSR. Porter and Kramer (2006) see CSR as a marketing tool, because the ranking of firms based on CSR is receiving considerable publicity, used by companies to attract new customers. Another reason may be because firms just think that it is the right thing to do (Sprinkle and Maines, 2010). Sprinkle and Maines (2010) also state that increasing CSR may be a reaction to satisfy various stakeholders. A third reason for incorporating CSR may be the contracting benefits such as increased motivation for employees (Porter and Kramer, 2006). Also the risks of a company can be reduced by engaging in CSR because of a reduction in incidents due to better working conditions or the reduction of hazardous emissions (Sprinkle and Maines, 2010). But there are also costs associated with incorporating CSR in business strategies (Porter and Kramer, 2006). These costs are associated with the activities that are undertaken for CSR (donations, additional personnel, contribution of employees time, costs of using sustainable
products) and the activities that were not undertaken because of a focus on CSR (Sprinkle and Maines, 2010).
According to Burke and Logsdon (1996) companies should not focus on the direct link between CSR and short-term profits, but should think of CSR as a tool to create strategic benefits. According to them CSR should be seen as a strategic tool that brings benefits to the firm when it supports the core activities of the firm and thus improving the effectiveness of the firm. Burke and Logsdon (1996) find five dimensions of corporate strategy that are important to establish success and create value while installing CSR policies and strategies. One of these dimensions is the importance of the fit of the CSR strategy with the mission of the firm. A second important dimension that Burke and Logsdon (1996) identify is that firms have to be able to capture the private benefits form increasing CSR and not only establishing public goods. An example of this is investments in research and development that leads to new products on which a patent can be given (Burke and Logsdon, 1996).
After the publication of Bowen’s book on CSR in 1950, people started to see the benefits of CSR and firms started to associate CSR with increased profitability and higher valuation in the future. McGuire et al. (2006), Porter and Kramer (2006) and Souto (2009) focus on the wishes of stakeholders, corporations and society as a reason for the incorporation of CSR (Gond and Moon, 2011). These changes were caused by pressure from stakeholders to hold companies accountable for their social impact (Porter and Kramer, 2006). According to Porter and Kramer (2006) difficulties arise because firms are aware of the risk of not acting on behalf of all stakeholders but they often try to solve it by using media campaigns instead of strategic or operational solutions. In 2006 64 % of the 250 largest multinationals published CSR reports that show the social and environmental actions the firm has undertaken (Porter and Kramer, 2006). These reports do not state the specific strategies used or outcomes established by the firm. These reports are an indication that companies don’t actually incorporate more social responsible practices, but often focus on the perception of increased CSR. An indication of this is that these reports often omit the actual impact or results that are booked, they only mention the number of hours spent on the project or document the results only for a division of the company but not the entire company (Porter and Kramer, 2006). Resulting in media coverage and publicity campaigns instead of actually making a difference in the business strategy. According to Besley and Ghatak (2007) CSR firms address both the wishes of the stakeholders and shareholders but with differences in the allocation when compared to non- CSR firms. This is often referred to as the stakeholder view (Freeman, 1984). This means that firms try to satisfy the preferences of their customers, employees and their community regarding social, environmental performance and the actual impact they have on society.
A company can benefit from CSR by building the reputation of a responsible business, attracting more reliable and more concentrated suppliers, customers are more dedicated based on the social capabilities of the company, the relationship with the community is improved by helping out with
community issues which includes increasing employee satisfaction (Souto, 2009). According to Giannarakis and Theotokas (2011) a firm also mitigates social risks such as ethnic conflicts, pollution and unemployment by including more social responsible projects, focusing on those aspects, in the business strategy. The mitigation of the social risk comes from projects ranging from raising awareness regarding ethnicity discrimination to reducing gas emissions of factories. An increase in CSR can increase the profitability in the future by using CSR as a positive signaling tool towards society (Kane, 2002). The motivation behind this statement is that CSR attracts expanded investment for the company, improves the image of the company and raises the value of the stock. This will increase the value of the company when the cash outflow of the company will be smaller than the cash inflow, otherwise the investment will have a negative Net Present Value, which will decrease the value of the company. According to Porter and Kramer (2006) the best way to increase profitability in the future is when the socially responsible strategies are in line with the economic and regulatory interest of the company. An example of this is a change in materials used at McDonald’s that led to a reduction of 30% in waste (Porter and Kramer, 2006).
One way to improve the competitive advantage is when a firm invests more in CSR according to Porter and Kramer (2006), Burke and Logsdon (1996). According to them both companies and consumers profit from establishing socially responsible programs. An example that increased CSR can be beneficial is that exploitation between firms can be circumvented by installing strong regulatory standards that protect companies and consumers (Porter and Kramer, 2006). Companies such as The Dow Jones Sustainability Index and the FTSE4GOOD Index but also non-profit organizations have begun to rank firms based on their social performances and gain substantial publicity for it (Porter and Kramer, 2006). The competitive advantage that can be created is one of the reasons to give directors incentives to pay more attention to the social and environmental consequences of their business practices (Burke and Logsdon, 1996). These incentives can be in the form of a bonus or as stock options, linked to social performance of the company in that year, that a director receives at the end of a calendar year.
Another important implication of the CSR rating (for example in the Dow Jones Sustainability Index, The KLD or the FTSE4GOOD) a company gets is the consequence that it can have on the future proliferation and profitability of the company. An example of this was when in 2006 the KLD database removed Coca Cola from their social listing due to distress regarding their environmental and social standards in the developing countries (Chatterji et al. 2007). This led to a divestment of 1.25 million shares of Coca Cola.
According to Porter and Kramer (2006) there are three big motivations for firms to engage in CSR. The first is that CSR can create value for a company since it stimulates a sustainable business model; firms should strive for economic stability and stay way from short-term strategies that will cause harm to the environment or be socially damaging.
Creating a sustainable business model can promote competitiveness and innovation inside the company to come up with new strategies that can be implemented (Haigh and Jones, 2006). By improving social performance of a company, new business strategies have to be formed. This sparks creativity and innovation in a company. This in turn develops and encourages higher human capital since human capital rises with creativity, competencies and a steep learning curve (Porter and Kramer, 2006). A second motivation for more CSR is that it can be seen as risk management (Porter and Kramer, 2006). CSR can reduce financial risk that arises when the actions of the firm are being regarded as unacceptable. An increase in CSR will also enhance external relationships with suppliers, investors and people in the community through the socially responsible projects that are installed in the firm. The third reason for an increase in CSR is that it is a form of corporate philanthropy that can enhance the reputation of a company. So instead of donating money to many small charities the management of the company incorporates responsible strategies and objectives in its business plan (Porter and Kramer, 2006). It is not as productive for the firm if they would focus on the outcome and the effect of enhanced social responsible projects. It would be more effective to adapt a CSR strategy that is not generic (a strategy that focuses on gaining a competitive advantage through lower costs, product differentiation and focus) but that fits the business strategy and core competencies of the firm (Giannarakis and Theotokas, 2011; Haigh and Jones, 1996).
With an enhanced socially responsible program a company can satisfy the wishes of shareholders, the community and other stakeholders (Porter and Kramer, 2006). The link between business and society is that they depend on each other; a successful company needs the society to be healthy, educated and productivity, safe working conditions will reduce accidents on the work floor, and productivity is enhanced with efficiently handling of land, energy and gas. Governments are obliging companies to file Corporate Social Responsibility reports and even legislation, such as the Sarbanes-Oxley act, the Working Time amendment (2001) and the Race Relations act (2001) in the US are formed that will mandate companies to include reports on their social, ethical and environmental performances (Porter and Kramer, 2006). The obligation to file CSR reports and the need for a healthy society indicates that incorporating CSR in firm strategy is on top of mind for the board of directors of companies.
2.1.2 CSR and executive compensation
According to the agency theory (Ross and Mitnick, 1972) the executive of a firm acts as an agent for the board of directors. By giving a director a reward in the form of a bonus for future behavior in line with the expectations of the board of directors, it can act as an inducement for better performance in the future (Mahoney and Thorn, 2006). The board of directors can use compensation to encourage the CEO to invest in projects that have a financial or social and environmental outcome (Mahoney and Thorn, 2006; McGuire et al. 2003). McGuire et al. (2003) state that the board of directors can control executive remuneration to prize executives for fulfilling specific strategies and objectives. This study
builds on previous research of McGuire et al. (2003) and Mahoney and Thorn (2006) by investigating if corporate governance, in specific executive remuneration, has an effect on the social performance of a firm. According to McGuire et al. (2003) executive remuneration targets not only financial performance objectives but also certain social objectives. McGuire et al. (2003) therefore indicate that it is important to investigate which form of executive remuneration has an effect on certain performance objectives. Executive compensation consists of three major elements: a fixed compensation salary, a bonus that can be applied to both short-term and long-term goals, and the third element is a term incentive such as the stock options. The difference between a bonus and long-term incentives is that a bonus offers an executive a direct cash compensation while stock options can be exercised in the future and change in value. Stock options might also encourage myopic behavior when the vesting period, the period in which options can be exercised, is short. So for stimulating long-term objectives it may be a logic step to increase vesting periods to align the time periodsof the stock options with the project. This may direct an executive towards building up longer-term relationships with the stakeholders instead of focusing on the short-term profits (Mahoney and Thorn, 2006). According to McGuire et al. (2003) when managerial performance standards are defined, compensation for executives can steer the executive towards specific outcomes. When CSR is defined as one of these performance standards, executive compensation can direct the executive towards those social goals. A reward for profit, a bonus, may be as effective as rewarding an executive for CSR since CSR can have a positive effect on firm performance (Porter and Kramer, 2001). But CSR is not only important for firm performance but also for the entire perception of the company. The perception of the firm translates into higher profits due to an increase in customers, declining costs because of for example reduced energy usage and increased innovation. For managers the decision to be more socially responsible can be seen as a trade-off between short-term profit and profitability in the future (Souto, 2009).
2.2 Empirical motivation
As mentioned in the previous section many researches have been looking for the link between CSR and profitability in the future. Ghoul et al. (2011) study the effect of CSR on firm value and on the cost of equity capital. They measure this by using an accounting-based approach, where they regress the cost of equity on several CSR proxies and control variables, to examine the relationship between CSR and the cost of equity. They find that a one standard deviation increase in CSR leads a decrease of 0,10 points in the cost of equity. Since the cost of equity is the required return rate based on the riskiness that the market attaches to the firm; the lower firm risks can be a reason for increased investment. Because investors are risk averse, they are inclined to invest in companies that have low risk of going bankrupt. So investing in CSR decreases riskiness, which attracts risk-averse investors
(Ghoul et al., 2011). This is in indication that CSR is linked to higher profitability, lower risk and hence a better valuation in the future (Ghoul et al., 2011 and Godfrey and Hatch, 2010).
Margolis et al. (2009) perform a meta-analysis on 214 manuscripts of companies, which all include a measure of CSR, a measure of financial performance and a measure of the effect of CSR on financial performance in their company manuscripts. They find a relatively small CSR- financial performance mean effect size (r= 0.13). Margolis et al. (2009) also report that 29% of the studies find a positive relationship between CSR and financial performance.
Wirl et al. (2013) investigate the market dynamics of CSR activities by remarking that if increasing CSR leads to higher profitability in one company, this must also hold for other companies in the same sector. Their starting point is that firms will increase CSR activities when it maximizes Net Present Value of profits. They find multiple equilibriums for firms in the same market trying to improve the improvements that can be gathered from increasing CSR. Wirl et al. (2013) try to explain CSR-waves, differences in CSR across countries and differences across time. They state that when a firm has a certain initial level of CSR, the return on CSR in the future will be higher and a CSR strategy will survive.
Since higher profitability is one of the goals for an executive or board of directors, increasing CSR can be a strategy to achieve this. Since the board of directors of a company is accountable for the organization’s performance, the will to increase CSR in the light of higher profitability is a goal (Godfrey and Hatch, 2010).
The link that will be investigated in this thesis is the effect that short- and long-term compensation of executives can have on socially responsible goals such as the environment, employee relations and health (McGuire et al, 2003). Mahoney and Thorn (2006) emphasize that the traditional role of the board of directors is to enhance firm profits by making sure that the objectives of the CEO are in line with the expectations of the shareholders.
A tool for the board of directors to steer an executive towards the goals that are set is executive compensation (McGuire et al., 2003). This is an indication that compensation can be used to direct executives towards more socially responsible strategies. Executive compensation can be divided in three components: a fixed amount salary, a bonus that is a short-term incentive and stock options that are long-term incentives. Assumed in this thesis is that the proportion of long- and short-term incentives that an executive receives influence executives to focus on either the long- or short- term. As mentioned before CSR increases profitability reduces risk and raises the value of the company in the long run (Wirl, et al., 2013 and Godfrey and Hatch, 2010 and Porter and Kramer, 2006). This indicates that the long-term portion of executive remuneration will direct an executive to increase CSR. However, McGuire et al. (2003) mention that stock options do not result in an actual loss for the executive. So they expect a negative relationship between stock options and CSR. Their point is that each component of executive remuneration has a different effect on the way an executive leads the
company and the business plan he formulates. This in turn affects the society as a whole. But Mahoney and Thorn (2006) state that stock options are dependent on the value of the stock in the future. This indicates that CEO’s that receive stock options will focus on goals set in the future (Mahoney and Thorn, 2006). If the board of directors and stakeholders want an increase in CSR in the future, stock options can be used to direct the executive in that way (Mahoney and Thorn, 2006).
McGuire et al. (2006) use salary measured as annual cash salary that an executive receives, bonus is measured as the percentage of additional payments an executive receives and the long-term incentives are measured as the percentage of stock options and other long-term incentives, measured by the value of the options relative to total salary of the executive (McGuire et al., 2003). They use a sample of one-year data from 1999 where they extract CSR data from the KLD Database. The KLD database rates companies based on 90 strengths and concerns on which a company can get a rating of 0 or 1. These strengths and concerns indicate that a firm can be strong and weak at the same time for certain social performance indicators and hence they are measures on two different scales. The dependent variable in their research is total CSR and the independent variables are the three different compensation variables (salary, bonus, and long-term incentives). Executive incentives act as an encouragement for positive or negative strategies for social responsibility (McGuire et al. 2003). McGuire et al (2003) find that corporate governance has a significant relationship with weak CSR. They also find that there is a negative relationship between long-term incentives and CSR and between salary and CSR. In the regression of executive compensation and firm-level control variables on the number of CSR concerns they report a beta coefficient of 0.27 on the variable salary and a coefficient of 0.16 on the variable long-term incentives.
McGuire et al. (2003) state that CSR is largely motivated by managerial conviction towards social performance and not by executive compensation. They state that the negative relationship that they found between executive compensation and CSR indicates that stimulating CSR by means of remuneration has no point because CSR is determined by managerial beliefs.
Mahoney and Thorn (2006) extend the four-component measure of CSR (including community, employee relations, environment, product and business practices) used by McGuire et al. (2003) with a seven-component measure of CSR (adding diversity, international and other). Mahoney and Thorn (2006) use data from Canada. They used the year 2001 for their research and used the Canadian database for CSR to construct the dimensions.
They find that there is a strong relationship between stock options and bonuses and CSR strengths. In the regression of executive compensation and firm-level control variables on the number of CSR strengths they report a beta coefficient of 0.41 on the variable stock option percentage and a coefficient of 0.28 on the variable bonus. They also find a strong relationship between salary and CSR concerns. In the regression of executive compensation on total CSR concerns they report a beta coefficient of 0.31 on the variable salary.
2.3 CSR and the financial crisis
Companies that start socially responsible projects or include CSR principles in their business strategy before or during the crisis are less affected by the financial crisis because trust in the company is enhanced due to stakeholders that are satisfied that their wish for increased CSR are accomplished (Souto, 2009).
Giannarakis and Theotokas, 2011 state that in crisis times businesses engage in new operations that may be socially responsible and circumventing the effects that the crisis can have on the business. But managers may also stay away form undertaking new socially responsible goals, since it will cost valuable time and effort that a company may not be willing to spend during a crisis (Souto, 2009). If a firm’s main objective is to make a short-term profit, it will not be likely that it will enhance its CSR projects because a profit from a socially responsible project will not be immediate (Souto, 2009). On the other hand socially responsible actions are more in demand by directors and stakeholders in times of crisis since it can have a turn-around effect, attract investors and customers who find CSR of great importance, and improve business operations (Karaibrahimoglu, 2010). Souto (2009) states that CSR can be a threat during the crisis because it can be a waste of crucial money that may be needed as a buffer for a potential bankruptcy. But he also states that CSR can be transformed from a threat to an opportunity in a crisis. More firms are including business practices that have a social goal that can be seen in the yearly reports that have been issued. Reasons for this are that it attracts quality employees, the name of the company is better and they are a more wanted as an employer. The recent financial crisis has caused executives and managers to shift from an outlook focused on profit to an outlook that incorporates the social footprint that they leave behind and the effects on society and the environment (Souto, 2009). Companies are redefining the core objectives of the firm, trying to focus less on strictly economic profit but also looking at the long-term effect they have on society by including socially responsible strategies to the business goals. Porter and Kramer (2006) state that CSR mainly has an impact on the intrinsic values in a company (employee satisfaction, work atmosphere, motivation) but less on short-term financial performance. Another thing that has to be pointed out is that engaging in social responsible projects generate costs; something firms try to avoid especially during a crisis when a buffer is needed to secure the future of the firm.
When comparing the requests for CSR and the desires in crisis periods, Souto (2009) finds some significant similarities. The first similarity is innovation, which is brought to a company when it starts focusing more on CSR and which may be a mechanism to get out of a crisis (Souto, 2009). The reason for an increase in innovation is that employees get motivated and challenged when installing more socially responsible projects. Also when employee/working conditions improve employees tend to become more innovative due to more responsibility or respect (Giannarakis and Theotokas, 2011).
When implementing CSR revising business strategies is often a following consequence. According to Souto (2009) this promotes increased transparency, investor confidence, positive market attitude and a deep internal reflection of the company, which is also a necessity for recovering from a crisis. When CSR is considered as an effective business strategy, it will have a positive effect on a firm’s position in the market, especially when they make an effort towards communication and transparency (Karaibrahimoglu, 2010). Because of the transparency a firm creates more confidence for the investors and stakeholders, especially when companies increase their communication with investors (Souto, 2009). Since most investors are risk-averse they are more inclined to invest in these
companies than very risky companies. Another reason for investors to invest in socially responsible companies is that CSR can offer a guarantee to investors (Souto, 2009).
Empirical researches on the link between a financial crisis and CSR are performed by Giannarakis and Theotokas (2011) and Karaibrahimoglu (2010). According to Giannarakis and Theotokas (2011) firms increase socially responsible performance during the financial crisis that started in 2007. Giannarakis and Theotokas (2011) study the effect of the financial crisis on CSR. The conduct a research on 112 companies that applied the GRI (Global Report Initiatives) in the time period of 2007-2010, to see if the financial downturn that started in 2007 has an effect on CSR. The Global Report Initiative is an organization that constructed a general sustainability-reporting framework so companies all over the world so that companies can be easily compared to each other based on sustainability and social responsibility (Giannarakis and Theotokas, 2011). Companies that operate conform the Global Report Initiatives are used over a time period containing the financial crisis that started in 2007. Giannarakis and Theotokas (2011) find that investment in CSR increased during the crisis rising from a mean of 4.133 in 2007 to a mean of 5.07 in 2010. The reason that Giannarakis and Theotokas (2011) provide is that firms invest in CSR in times of a crisis to salvage the lost trust in the firm. They state that a financial crisis can offer the opportunity to make a drastic positive change in either a firm’s brand, employee satisfaction, productivity and perception.
However, Karaibrahimoglu (2010) states that a decrease in the attention to CSR has been noticed after the financial downturn and uncertain business climate since 2008. He finds that companies engage less in socially responsible projects than is expected from literature. In his research he reviews 100 companies by extracting data from the non-financial reports of these companies and evaluates
companies based on information on their webpage and annual reports. Karaibrahimoglu (2010) finds a decrease of 0.357 in CSR projects from 2007 to 2008, indicating that the recent financial crisis that started in 2007 has a negative effect on CSR.
3 Hypothesis development and methodology
This chapter introduces and explains the hypothesis and the regression model used to study the relationship between CSR of firms and executive remuneration of the firm executives. The purpose of this thesis is to find the effect of variable remuneration, namely the short- and long- term incentives, received by directors on CSR. In addition the effect of the crisis on the relationship between CSR and executive remuneration is studied by comparing the effect during and after the most recent financial crisis that started in 2007.
The regression model below is used to introduce the three hypotheses of this thesis. This is also the regression that is performed in this thesis.
€
CSRi,t =
α
+β
1Salaryi,t+β
2Bonusi,t +β
3StockOptionsi,t+β
4Crisist* Bonusi,t+β
5Crisis * StockOptionsi,t+
β
6Crisisi,t+β
7Yeari,t+γ
jIndustryi,t+β
8ROEi,t +β
9Sizei,t +β
10DebttoEquityi,t+ε
i,t j =112
∑
Null hypothesis 1a: A bonus that is given to the executive of a company has no effect on CSR. From past literature (McGuire et al, 2003; Mahoney and Thorn, 2006) it is expected that the variable portion, which captures bonus, of executive remuneration can have a positive effect on CSR. By giving a director a reward in the form of a bonus for future profits, it can act as an inducement for better performance in the future (Mahoney and Thorn, 2006). One way to improve future performance in the form of competitive advantage is when a firm invests more in CSR according to Porter (2006). Therefore it is expected that the coefficient of
€
ˆ
β
2to be > 0, as can be seen in table 1.Null hypothesis 1b: Stock options as the variable part of executive remuneration are expected to have no effect on CSR. The motivation is that when a director receives compensation in the future for increased profits, he or she will pay more attention to actions that will have a positive influence on profits and overall firm stability in the future (Kruger, 2009). So stock options act as an incentive to increase profits. One of these actions can be heightened attention to CSR.
Therefore it is expected that the coefficient
€
ˆ
β
3to be > 0.Null hypothesis 2: The financial crisis has no on CSR (Giannarakis and Theotokas, 2011). Crisis is measured as a dummy variable that is 1 during the crisis years (2007, 2008) and zero before and after the crisis. Companies invest more in CSR during a crisis to regain the lost trust in the companies.
Based on the research of Giannarakis and Theotokas (2011), Souto (2009), Karaibrahimoglu (2010) and the Global Report Initiatives it is expected that
€
ˆ
β
6 will be positive.Null hypothesis 3: Executive remuneration will have no effect on CSR during the crisis. The effects of the short- and long-term incentives on CSR are measured with two interaction dummies. The effect of executive remuneration on CSR has increased during the crisis. As stated above it is expected that
€
ˆ
β
4 and€
ˆ
β
5 will be positive.The control variables that are included are Return on Equity, Size, Debt to Equity and dummy variables for industry are added to control for the industry the firm belongs to.
CSR is the dependent variable. The independent variables in the regression are salary, the percentage of bonuses and the percentage of stock options. The regression also controls for return on equity, firm size, debt to equity and industry. Manufacturing has been used as the baseline industry since it is most frequent industry. These variables are included to reduce omitted variable bias. And finally a dummy variable is added to account for the crisis.
4. Data
In this chapter firstly the construction of the dataset that is used for this thesis is elaborated on. Then an explanation for the origin of the dependent variable and the explanatory variables are given. Finally the control variables and the dummy variables are discussed.
The data is collected over time starting in 2001 until 2012. This time-span includes the financial crisis, which started in 2008. A problem occurs when not every firm is present in the dataset, the entry and exit of the firms in the panel are correlated with CSR and this disrupts the analysis. So the data that is used includes only firms that are present in all twelve years of the dataset.
4.1 Sample selection
For this research a sample of US publicly traded firms is used. The sample consists of the years 2001 until 2012. The KLD (Kinder, Lydenberg and Domini) database provides a rating of companies for CSR based on 90 strengths and concerns and was founded in 1988. They use SOCRATES, which is the social rating monitor that includes profiles of more than 4000 US firms. The ratings are creates by a group of analysts that use internal assessments to measure the CSR objectives. The internal
assessments include questionnaires sent to investor relations, interviews, visits of the site and annual reports reports. Government agencies, non-governmental organizations and news reports are also
being used to define the ratings for all the companies included in SOCRATES. The KLD database give a rating for both past socially responsible outcomes, such as environmental and social improvements, and current managerial actions undertaken that may be an indication for future outcomes (Chatterji et al., 2008). The CSR objectives are divided in two areas: qualitative issues and controversial practices. Some examples of the qualitative objectives that are being rated by the KLD analysts are community relations, diversity, environment and human rights. Examples of the
controversial practices include tobacco, firearms and alcohol. For each of these objectives KLD created a set of strengths and concerns that are rated with a 0 or 1. A concrete example of a CSR strength is generous giving. ‘A company gets a 1 if it has consistently given over 1.5% of trailing three-year net earning before taxes to charity, or has otherwise been notably generous in its giving’ (KLD ratings criteria, p.3, 2003).
Data on salary, bonus and stock options of executives are retrieved from Execucomp. For salary, bonus and stock options the absolute values in millions of USD are used. Since a director is defined as a board member, firms often have more than one director, even at one moment in time. Because the dataset runs over time a firm will likely have more than one director during this time period. To create one value for every firm for salary, bonus and stock options, the director variable had to be aggregated to the firm. The consequence is that for this research the average of the salaries, bonus and stock options of the board members of the company are used in the dataset.
To control for the interferences that the variables can have and for reducing the omitted variable bias, firm- and industry specific control variables are used. The firm specific control variables are: Firm Size, ROE and Debt To Equity, industry level dummies are created using SIC codes. These variables are retrieved from Compustat. Because data in Compustat has a monthly frequency, for this research the end of the last month of the year was taken as the amount for the year. By using the cusip codes, which is a nine-digit code that is assigned to all issues that can be shared, the Execucomp and KLD datasets are merged. After this first merge 15,293 observations were present and after the final merge 10,334 observations remained.
Since panel data was used in this thesis a check for the sequence of the companies had to be done. From figure 1 it can be concluded that most firms are present for seven, ten and twelve years. Because of missing values 3107 firm-years are present for all twelve years of the dataset, which is equal to 259 observations for every year. To control for firm level variation the regression uses the cross-section of 259 observations for every year. This elimination led to a loss of 69%. The number of items in the KLD database varies over time. When the number of items in the KLD database rises, new companies automatically will have a higher score than older companies.
Figure 1. Number of years that firms are present in data. The data is collected over time starting in 2001 until 2012. The firms are a sample of US Publicly traded firms that have been rated by the KLD database on CSR strengths and concerns. Most firms are present for 7, 10 and 12 years of the dataset. For this thesis the firms that are present in all 12 years are used.
4.2 Variable description
In this part of the thesis the variables that are used in the regression are described and it is exemplified how they are gathered from the databases. First the dependent variable (net CSR strengths) will be discussed, followed by the explanatory variables, the control variables, the industry level controls and finally the year dummies.
From table 1 it can be seen that a total of 3107 observations are used, missing values caused a loss of 97 observations. The average salary in this sample of an executive is 0.5 million USD. The variable compensation is on average higher but also has more variation. On average a firm has one CSR strength, which is comparable with McGuire et al. (2003) who report a mean of 0.73 CSR strength per firm. And for CSR concerns a firm in this sample has on average 0.8 CSR concern while McGuire et al. (2003) find a mean of 0.64.
0 1000 2000 3000 N u mb e r o f firms 1 2 3 4 5 6 7 8 9 10 11 12 Number of years
Table 1. Descriptive statistics (N=3107, CSR strengths is the number of reported strengths in the KLD database. CSR concerns is the number of reported concerns in the KLD database. Salary is mean executive salary in millions of USD. Bonus is mean executive bonus compensation in millions USD. Stock options is the mean of executive stock option compensation. Stock option compensation is sum of the income of the exercised stock options and the market value of unexercised in-the-money stock options. Stock options are measured in millions. Return on equity is net income divided by equity. Debt to equity is total debt divided by shareholders equity. Size is the logarithm of total assets. Source: CSR variables are retrieved from the KLD database, compensation variables are retrieved form Execucomp and firm-level variables are retrieved from Compustat
Variable Observations Mean Standard deviation Minimum Maximum CSR strengths 3107 1.015 1.327 0 8 CSR concerns 3107 0.801 1.208 0 5 Salary 3107 0.556 0.326 0.073 3.474 Bonus 3107 0.347 0.656 0 10.483 Option awards 3107 1.489 3.919 0 75.349 Return on Equity 3107 0.066 0.189 -‐7.366 1.419 Debt to Equity 3107 0.747 1.123 0.002 11.583 Size 3107 8.493 1.369 2.961 12.980
4.2.1 Dependent variable: net CSR strengths
The variable for CSR is gathered from the KLD (Kinder, Lydenberg and Domini) Research and Analytics Environmental and social performance database that can be found in WRDS. This database contains data on CSR values for the 4000 largest publicly traded firms for the US. The KLD database rates the firms based on 13 different areas.1 As mentioned in section 4.1 for every area that is rated the company gets a rating for the strengths and weaknesses in that specific area. These scores are
measured per category. A firm can get a rating of 0 or 1 for an objective and then they are added for each category.
The variable net CSR strengths is obtained by subtracting CSR concerns from CSR strengths. As can be seen from figure 2 the distribution resembles a normally distributed variable and therefore an OLS regression can be used. However net CSR strengths is a discrete variable. Strictly speaking an ordered logistic regression model is most suitable for this type of data. To check whether the results are
1
Community (14 items in rating), corporate governance (20 items in rating), product (12 items in rating), environment (24 items in rating), diversity (16 items in rating), human rights (16 items in rating), employee relations (21 items in rating), alcohol consumption (3 items in rating), firearms (2 items in rating), military (5 items in rating), tobacco (3 items in rating), gambling (3 items in rating) and nuclear power (5 items in rating).
qualitatively affected by the choice of regression (linear versus ordered logistic regression) a robustness check is performed in the form of an ordered logistic regression and is included in section 5.3.1.
McGuire et al. (2003) and Mahoney and Thorn (2006) used the sum of the CSR strengths and concerns as the dependent variable in their regressions. They find a strongly significant effect of the size of a company in assets on CSR strengths and concerns. In this thesis net CSR is used as a
standard for CSR. Net CSR is the difference between CSR strengths and CSR concerns. This measure exemplifies the direct actions that are undertaken by the firm towards CSR. Using net CSR strengths corrects the fact that some firms have both CSR strengths and concerns and may use the CSR strengths to compensate the concerns.
Figure 2 compares the distribution of net CSR and the sum of CSR that Mahoney and Thorn (2006) and McGuire et al. (2006) use in the regressions. Figure 2 shows that net CSR strengths has a roughly symmetrical distribution between -5 and 5. For the sum of CSR strengths the distribution is
concentrated at 0.5. Net CSR strengths shows a clearer distinction between high- and low social performers compared to the sum of CSR strengths. Only CSR items are included that are present in all twelve years of the sample so that data can be compared over time. 2
Figure 2. Distribution of net CSR strengths, calculated by subtracting CSR concerns from CSR strengths, and the sum of CSR strengths. Net CSR strengths is symmetrically distributed while sum CSR is concentrated to the left. Net CSR has a wider range than the sum of CSR strengths. (N=3107, Source: KLD Database)
2
Items included in all twelve years of the sample: Investment Controversies, Negative Economic Impact, Union Relations, Health and Safety Concern, Regulatory Problems, Substantial Emissions, Climate Change (from 1999), Transparency Strength, Charitable Giving (from 1991 through 2011), Innovative Giving, Union Relations, Cash Profit Sharing, Employee Involvement, Health and Safety Strength, Beneficial Products and Services, Pollution Prevention, Recycling, Clean Energy, Management Systems Strength
Both CSR strength and concerns variables are constructed from four CSR categories that are all present for all years of the dataset; corporate governance, employees, environment and community. From these four categories a total of 11 CSR strengths and 7 CSR concerns are used to compute the net CSR strengths variable. Figures A1 and A2 in the appendix show the distribution of the CSR items can be seen. The concerns are investment controversies, union relations, regulatory problems, climate change, negative economic impact, health and safety and substantial emissions. If a company does not have a strength or concern in that issue it is indicated with a 0 and otherwise it gets a 1. The concern that is most frequent is health and safety with an average of 22%. The strengths that are included in the regression are reporting quality, donations for innovation, profit sharing, employee health, recycling, management systems, charity donations, union relations, employee involvement, benefit product and clean energy. A management system, which indicates a firms commitment to setting up system standards such as ISO certification, and clean energy are the most frequent strengths with an average of 20% and recycling the least frequent present in the sample with an average of 3%. These items indicate a firm’s environmental performance.
The amount of net CSR is presented in figure 3. From this graph a downward trend can be observed starting simultaneous with the financial crisis and a significant upward trend beginning in 2010 can be observed as the years progress. This indicates that time controls have to be added. A linear time trend is chosen in the analysis. As a robustness check time is controlled for with a quadratic and a cubic time trend and with time fixed effects. In figure 3 a linear trend and a quadratic effect is fitted. From
0 500 1000 1500 F re q u e n cy -5 0 5 10 Net CSR strenghts 0 500 1000 1500 F re q u e n cy 0 2 4 6 8 Sum of CSR strenghts
the quadratic trend it can be seen that there is a non-linear effect of time. This will be further investigated in the robustness check section.
There is a possibility that financial crisis related items in the KLD scale drive the downward trend of CSR during the financial crisis. Especially the item negative economic impact will likely be strongly correlated with the financial crisis. As robustness check the analysis will be performed without the negative economic impact item. These results are included in section 5.3.3
Figure 3. Average Net CSR strengths per year. Net CSR strengths are stable until 2007, turn negative between 2007 and 2010 (the financial crisis) and increase strongly afterwards. Included are linear and quadratic trends. The quadratic trend indicates a non-linear effect of time. A (N=3107, source: KLD database)
Figures A3 and A4 in the appendix show the mean CSR per firm per year. Graph three indicates that the number of concerns have risen since the start of the crisis in 2007. An example is the accounting concern that has been added in 2006 and rates if a company has weak reporting qualities. The rise of concerns during the crisis is also caused by the domestic market weaknesses and uncontrolled growth that were the foundation of the crisis (Souto, 2009). The domestic market weaknesses lead to financial distress for firms and caused firms to get the rating 1 for concerns during the crisis. This was one of the reasons for an increase in concerns.
From figure 4 it can be seen that the number of strengths has increased significantly starting in 2010. A reason for this can be that firms changed their policy toward CSR during or after the crisis and the
-. 2 0 .2 .4 .6 .8 N e t C SR st re n g h ts 2000 2005 2010 2015 Year
(mean) net CSR strengths Linear trend Quadratic trend
result could be that firms started focusing more on the strengths within CSR (Souto, 2009). The crisis acted as a stimulant for companies to change their corporate policies or visions (Kruger, 2009).
Figure 4. Evolution of CSR strengths and concerns over time. Data on CSR is retrieved from the KLD database starting from 2001 until 2012, including the financial crisis that started in 2007.
4.2.2 Explanatory variables compensation
Since the effect of executive remuneration on CSR is studied, measurements of the explanatory variables compensation in the regression are needed.
Data on bonus, stock option and salary, are retrieved from the Execucomp database that can be found in WRDS. The salary that is used is total cash salary in millions of dollars that is paid to an executive in a calendar year. As mentioned before the salaries of all directors of a firm are aggregated so that there is one average salary per firm.
For the variable stock options in executive remuneration the value realized on exercised options in that year are added to the estimated value of the in-the-money unexercised options (the difference between the exercise price and the close price of the stock of the company at year end). The average of the stock options for directors is 1.49 million USD (table 1). For the bonus variable the absolute amount of bonuses received at the end of the calendar year is used. The average bonus is 0.34 million USD (table 1). 0 .5 1 1.5 2 Me a n C SR st re n g h ts 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 0 .5 1 1.5 2 Me a n C SR co n ce rn s 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012
When the average statistics of this thesis are compared to the descriptive statistics of McGuire et al. (2003) the average of the stock options appears to be on the lower side. Mahoney and Thorn (2006) use the discounted present value of the options to attach a nonnegative value to it.3
The bonus variable was also retrieved from Execucomp. It replicates the short-term incentives a director gets as an addition to induce future performance of the company. The absolute value of bonuses an executive receives at year-end is used. As can be seen from table 1 the average of bonuses for the directors is 0.34 million USD. Director’s bonus is included in this regression because
according to McGuire et al. (2003) and Mahoney and Thorn (2006) a bonus for the director can act as catalyst for them to invest more in CSR.
4.2.3 Control variables: Firm Size, ROE and Debt to Equity
The control variable size is included because it has been shown that it is positively related to social performance as well as executive remuneration (Mahoney and Thorn 2006 and McGuire et al. 2003). Often it is true that the bigger the firm, the higher the salaries (Mahoney and Thorn, 2006). To control for the variable size, total assets (in million of dollars) are retrieved from the Compustat database. In the regression analysis the logarithm of total assets is used. This adaptation of total assets does not affect the interpretation of the results since a firm with large assets will also have large logarithm assets4. Table 1 shows that the mean firm has total assets of approximately $3 billion.
To control for the profitability of a firm, Return on Equity is used as a control variable. To create the variable for ROE net income (Earnings before income and taxes) and shareholder equity are retrieved from Compustat. Equity is calculated as total assets minus total liabilities. Consequently ROE is calculated as follows:
€
ROE = NetIncome / Equity
To calculate total debt; total long-term debt is used and added is Accounts Payable which corresponds to the short-term debt. Equity is retrieved from Compustat and calculated by subtracting total debt from total assets. Leverage has to be controlled for because firms with a high amount of leverage are
3 McGuire et al. (2003) measure the long-term incentive payments by adding other long-term incentives such as cash and vacation increase to the total amount of stock options. The difference between this thesis and other researches is that stock options are restricted to only option awards.
4
The logarithm of assets is used because from a statistic point of view this performs better in the regression
expected to invest less in CSR since they will have extra constraints due to debt covenants, unless the debt covenants specify CSR (Kruger, 2009). According to Kruger (2009) firm with high amounts of debt are more likely to reward their directors less with salary, bonuses and stock options.
4.3 Industry level controls
This regression also includes dummy variables to control for the industry a company is in just as McGuire et al. (2003) and Mahoney and Thorn (2006). A control for industry has to be added in the regression because it has been proved that there are certain industries that restrain from socially responsible actions because they are unable to satisfy the socially responsible needs of their customers (Kruger, 2009). Based on the first digits of the SIC codes firms were classified in their specific industry groups as in McGuire et al. (2003) and Mahoney and Thorn (2006).
As can be seen from table 2 the distribution of industries varies a lot in the sample, from -0.25 to 0.65 mean CSR. In the oil and gas industry, 240 oil and gas observations are present which corresponds to 20 oil and gas firms are present in the dataset.
Table 2. Industries in sample based on firstdigitsic-code. Firms are classified in an industry based on the first digits of their SIC code. (N=3107, source=KLD database, Compustat, SIC)
Industries Frequency Percent Cumulative Mean CSR
Oil and gas 238 7.66 7.66 -0.52
Food, apparel and chemicals 516 16.61 24.27 0.35
Construction 780 25.10 49.37 0.20
Transportation, communication, electric
539 17.35 66.72 -0.55
Wholesale and retail trade 132 4.25 70.97 -0.13
Financial services 560 18.02 88.99 0.16
Services 282 9.08 98.07 0.62
Health care 48 1.54 99.61 -0.25
Public administration 12 0.39 100.00 0.57
Based on the variation of CSR per industry that can be seen from figures A5 and A6, the inclusion of the industry dummies test the effect of industry on CSR.
4.4 Time dependent variables
It is possible that over time the relationship between CSR and executive remuneration changes. The financial crisis is included as a dummy variable that equals 1 in 2007, 2008 and 2009. In times of crisis firms have the inclination to turn around the strategy of the firm. It can be the start of increased attention given to CSR. It is assumed that during the crisis consumers are more inclined to purchase goods from a firm that is socially responsible (Karaibrahimoglu, 2010). As can be seen from figure 4, CSR strengths have risen considerably after the crisis. A stronger reputation for consumers, creating shareholder confidence, increasing competitiveness and enhancing the working climate are all advantages that can be associated to CSR (Souto, 2009).
Next to the effect of the financial crisis on CSR, a time-trend effect is included. This is done by including a control for the effect of time (year). The trend variable is calendar year 2001. So for 2001 the trend variable is 0 and for calendar year 2002 it is 1 and so on. Due to heightened attention from shareholders to CSR and more transparency during the sample period it is expected that CSR has increased during those years (Souto, 2009). By controlling for the time trend the autonomous increase in CSR in firms is recorded for. In section 5.3 a robustness check is performed with a quadratic and cubic time trend.
5. Results
In this chapter an answer to the research question in the introduction will ne given; what is the effect of variable executive remuneration on CSR? And has this changed during the recent financial crisis? The results of the regression analyses will be discussed according to the predicted effect of the three main hypotheses in this thesis.
5.1 Univariate analysis
McGuire at al. (2003) and Mahoney and Thorn (2006) have shown that there is a relationship between compensation and CSR. The correlation table (table 3) shows that there is a significant correlation between the compensation variables and net CSR strengths. Table 3 shows that salary has a significant relationship (p-value<0.05) with net CSR strengths.
