The influence of CEO characteristics and national institutions on Corporate Social Performance

(1)

The influence of CEO characteristics and national

institutions on Corporate Social Performance

Master Thesis Accountancy

University of Groningen, Faculty of Economics and Business P.O. Box 800, 9700 AV Groningen

The Netherlands Roderick J. Honig S1795406 Vechtplantsoen 14-3 3554 TB Utrecht Tel.: 06 30 13 84 63 E-mail: r.j.honig@student.rug.nl

Abstract: This study researches whether CEO age and CEO tenure have an influence on corporate social performance (CSP). Also, it looks at the moderating influences of two national institutions, namely political stability and environmental regulations. Theories used to explain the relationship between these previous variables arrive from upper-echelon, legacy, legitimacy and the institutions theory. Results show significant relationships between the four used variables on the one hand and CSP on the other hand. Also, this study finds a positive moderating effect of political stability on the relationship between CEO age and CSP. These results contribute to the different theories used in this research. Data in this research comes from Thomson Reuters’ ASSET4 ESG dataset, Yale’s EPI Index, the World Governance Indicator (WGI) Index and Fama and French’ Five Industries classification. This resulted in a sample of 7732 unique observations from over 40 countries in the years 2002 to 2010.

Supervisor Rijksuniversiteit Groningen: R.B.H. Hooghiemstra

Second Supervisor: S.A.L. Wakkerman

(2)

2

3.3 Independent variables 23 3.3.1 CEO age 23 3.3.2 CEO tenure 24 3.3.3 Environmental regulations 24 3.3.4 Political stability 24 3.4 Control variables 25 3.4.1 Firm size 25 3.4.2 Firm performance 25 3.4.4 Dummy variables 25 3.5 Analysis 26 4. Results 28 4.1 Descriptive statistics 28 4.2 Multicollinearity 32 4.3 Hypotheses 32

(3)

3

5. Conclusion and discussion 35

(4)

4

1. Introduction

In 2008, a survey by the Economist shows that 53.5% of firms that took part in the survey say that corporate social responsibility (CSR) “is a necessary cost of doing business”. Of the responding firms, 53.3% also says that “it gives us a distinctive position in the market”. According to Milyo, Primo & Groseclose (2000), in 1998, corporations gave over $17 billion to charity during the elections in the United States. This shows that firms try to improve their social image, by for example donating money to charities.

But not only firms pay growing attention to the social side of their actions. Since founded in 1969, the Council on Economic Priorities (CEP), an organization that is based in New York and London (with partners in over 10 countries), is committed to making information on CSR available to millions of consumers. It does so by evaluating the performances of companies on a wide scale of social dimensions. This results in guidelines to help consumers make purchasing decisions. Another example is the Social Performance Task Force (SPTF), who launched the Universal Standards for Social Performance Management in 2012. With over 1500 members, they help micro financial institutions to achieve their social goals 1.

Maybe the most important party in this matter is the public. In recent years, people seem to react to firms that perform bad on social and environmental issues. A survey by McKinsey shows that environmental, social and governance programs create shareholder value2.

One well known recent example is that of Apple, which made use of the so called ‘Double Irish with a Dutch Sandwich’ tax evasion. This results in them paying a mere 1.9% on taxes in 20123. As a reaction, people began boycotting products and even the American Senate put attention to the matter, which resulted in Tim Cook (Apple’s CEO since 2011) having to appear before the Senate4.

Although a lot of businesses embrace this new element of performance, according to Baron et al. (2011), the relations between the social, financial and social pressure elements within a firm are vague and there isn’t a clear direction of causation. They say that good corporate social performance (CSP) could affect good corporate financial performance (CFP) but in their turn, CFP could provide slack resources to be spent on CSP.

1 http://www.cgap.org/news/top-10-lessons-social-performance-management-microfinance 2 http://www.mckinsey.com/insights/corporate_finance/valuing_corporate_social_responsibility_mckinsey_global_survey_res ults 3 http://www.elsevier.nl/Economie/nieuws/2012/11/Apple-betaalt-19-procent-belasting-buiten-Amerika-ELSEVIER353787W/ 4 http://www.iphoneclub.nl/252248/tim-cook-moet-apples-belastingbeleid-uitleggen-aan-amerikaanse-senaat/

(5)

5

To explain why firms give more attention to, and invest more in CSP, numerous theories have been used. First of all, there is the agency theory (Jensen & Meckling, 1976; Jensen, 1986; Ghosh et al., 2007). This theory suggests that managers can divert resources at the expense of the firm. To maintain a good image, firms will try to have a good social performance. They can do this by, for example, hire mentally challenged people instead of people that function normally, which will cost them more of their resources. However, the manager can have incentives to use these resources for their personal gains. Another situation may arise when the manager will use these resources to invest in CSP improving elements, although this will not even improve the firm’s total performance (a so called pet project). For example, a manager may invest in disabled people, although this is not according the plan the firm has with these resources.

The stewardship theory (Davis et al., 1997) assumes the CEO and firm owner will work together to accomplish their goals, which may result in better alignment between principal and steward, resulting in higher (social) performance.

Another theory is the upper-echelon theory (Hambrick & Mason, 1984). This theory states that certain characteristics of top management teams (TMTs) and CEOs may influence their decisions and thus can give a reason why TMTs and CEOs will make certain decisions that can make CSP decline or grow. Within the upper-echelon theory, the legacy perspective (Ring, 1984; Cozzolino et al., 2004; Zacher et al., 2011) gives reasons why leaders change their behavior as they get older. CEOs may want to leave a so called legacy after their pension or death, and may thus start behaving more socially towards the public to leave a good image for when their place has been taken by a new CEO. On the other hand the resistance to change (RTC) view (Kunze et al, 2012) states that older CEOs are more conservative and hesitant for making big decisions, which gives implications regarding the explanation of results for firm and social performance.

Within this research, I will also have a look at national institutions. This will be done through the legitimacy and institutions theories.

The legitimacy theory can explain why managers will take decisions that are socially received well. Deegan (2002) states that firms can only exist because the public gives them legitimacy. Vanhonacker (2000) show that when a firm loses its legitimacy, social support will also be gone. This means the public will no longer give resources to this firm, which will make it harder for firms to continue their operations. Managers will change their actions and decisions to be socially more accepted, which hopefully results in more resources.

The institutions theory (DiMaggio & Powell, 1983), states that firms will take these institutions into account when developing its strategy. According to DiMaggio & Powell

(6)

6

(1983), different national institutions also give different results in different countries, which is why they can also influence the CSP of a firm.

1.1 Research plan

The main research question, which arises from previous research and the theories stated above is:

“What are the effects of certain CEO characteristics and national institutions on corporate social performance”

To answer the main research question, several sub questions are made: - “How does CEO age affect CSP”

- “In what way is CSP influenced by CEO tenure”

- “Do environmental regulations have an influence on CSP and how” - “What are the effects of political stability on CSP”

1.2 Research methods

To test whether the formulated hypotheses shown in the theoretical framework have significant effects, I make use of different variables. These variables come from four different databases, namely the Thomson Reuters’ ASSET4 ESG, Yale’s EPI Index, the World Governance Indicator (WGI) Index and Fama and French’ Five Industries classification. In this research, the first and second variables used, CEO age and tenure, originate from the upper-echelon theory (Hambrick & Mason, 1984). This states that certain CEO characteristics may have influences on performance. Firstly, the resistance to change relationship and legacy perspective give different arguments for both a negative and positive relationship between CEO age and CSP. Also, CEO tenure can be seen as having both a negative and positive association with CSP, due to the different perspectives, which will be explained in chapter two.

Also, in this research I will make use of two moderating variables, namely environmental regulations and political stability. Both of them are expected to have positive effects on CSP and the previously named CEO characteristics. This is due to the legitimacy theory (DiMaggio & Powell, 1983).

In the results section, three analyses are made, which are the descriptive statistics, Pearson correlations for testing multicollineairity, and lastly to test the hypotheses, linear regression analyses.

(7)

7

1.3 Scientific relevance

Earlier research acknowledges that CSP is an element that should be taken into account for managing a firm and making decisions that may result in better financial performance (Ioannou & Serafeim, 2012). This could imply that CSP is a subject that has been researched widely, but this isn’t the fact. There still is a large amount of unexplained variance across studies regarding financial firm performance and CSP (Guiral, 2012). Waddock & Graves (1997) and Orlitzky et al. (2003) do find a relationship between these two, but the direction of this relationship remains unclear. This means still a lot of research regarding these two important subjects can be done. However, the focus in this research is not on financial firm performance, but will only be used to make a link between other variables and CSP.

First of all, certain CEO characteristics will be explained via the upper-echelon theory. This theory (Hambrick & Mason, 1984) relates certain CEO characteristics to economic firm performance. In this paper, instead of looking at the influences on the financial part of firm performance, I look at CSP directly. Although Manner (2010) investigate the relationship between CEO characteristics and CSP, he looks at the educational background and gender of a CEO. I will focus on CEO age and tenure.

Most research done with regards to upper-echelons focuses on team and firm levels of analysis (Yamak et al., 2013). Instead, this research looks at upper-echelons on a national level. Also, it does not focus on one country, more than 40 countries are used in the analysis, and will also differentiate from earlier research through the use of national institutions. Thirdly, Hambrick & Mason (1984) state that executive characteristics will likely be different in various industries. According to Yamak et al. (2013), the understanding of these effects is still limited. This research may add more understanding in this matter, because a segregation will be made between five industries (following Fama & French, 1997) when using the CEO characteristics.

With respect to CEO characteristics, CEO age and tenure are the two variables that will be researched. These characteristics have been researched widely (Mudrack, 1989; Peterson et al., 2001; Sundaram and Yermack, 2007). For example, Sundaram & Yermack (2007) look at the part of debt of a firm that can be explained by CEO age. Peterson et al. (2001) state that younger females and older males will have a higher level of ethical beliefs. Instead of these firm related elements, this research focuses on the effects that CEO age and tenure have on CSP. Also, CEO characteristics through the RTC-relationship and legacy perspective are not linked to CSP yet. This means this research can give new insights of the behavior of CEOs regarding social performance.

(8)

8

The relationship between CEO tenure and job performance is a widely researched subject as well (Henderson et al., 2006; Simsek, 2007). Henderson et al. (2006) say that “in the stable food industry, firm level performance improved steadily with tenure”, while Simsek (2007) suggests CEO tenure only has an indirect influence on performance. Although this research also looks at CSP via job and firm level performance, the usage of national institutions in this matter has not been used. This means maybe new thoughts can arise from this research regarding the determinants of CSP.

Secondly, I look at national institutions. Van Essen et al. (2012) use the institution-based view to show the relationship between firm performance and CEO compensation and find empirical evidence which implies that, when looking at contracting environments, national institutions have a conditioning effect. Others focus on informational strategy issues like earnings quality (Kanagaretnam et al., 2011) and financial reporting (Hope et al., 2011) and find a positive relationship between higher institutional development and these two variables. This shows national institutions can have an influence on decisions regarding firm performance, which links these researches to CSP.

In this research, the national institutions are split in environmental regulations and political stability. Hypotheses for these two are explained through the legitimacy theory. Deegan (2002) says firms can only exist because of the public, which means it is an important theory that has to be taken into account, but he also says that this subject has not been investigated enough. This research might give a better understanding of implications for this theory.

1.4 Structure of research

In the next section, the theoretical framework will show the hypotheses used in the research. These are explained using the earlier named theories, namely the upper-echelon and legitimacy theory. Within these two theories, several perspectives will show how the hypotheses were made. Also, the conceptual model gives a brief and clear image of this research. After that, the used methods and variables will be briefly described. The results section shows whether the hypotheses will be accepted or rejected. These results will be discussed in the conclusion. Lastly, the discussion section talks about limitations to this research and will give suggestions regarding future research.

(9)

9

2. Theoretical Framework

2.1 Corporate social performance

More and more, firms are facing the pressure to maximize their corporate social performance (CSP) (Guiral, 2012; Hull & Rothenberg, 2008). CSP is a factor that has to be accounted for, which is acknowledged by Ioannou & Serafeim (2012), who say that “in recent years, growing attention has been paid to corporate social responsibilities”. These issues, like environmental pollution by emissions from production processes, are relevant to CSP and have become more prominent in recent years. In comparison with 1995, the investments in socially responsible investment funds (SRI funds) have increased with 2,634% to $316.1 billion in 2010 (U.S. Social Investment Forum). All of this shows that CSP is an element that should not be ignored.

CSP has been described as “a business organization’s configuration of principles of social responsibility, processes of social responsiveness, and policies, programs, and observable outcomes as they relate to the firm’s social relationships” (Wood, 1991: 693). Oikonomou et al. (2014) say CSP can be seen as the degree to which firms incorporate social and environmental concerns in their operations. This is almost the same as the definition for CSR as given by Homburg et al. (2013), which is “a firm’s voluntary consideration of stakeholder concerns both within and outside its business operations”. In this research, CSP and CSR will be treated as equivalents.

As is shown above, CSP can be seen as an important part of a firm. But what can influence the decisions that firms make regarding CSP? Some of these factors will be researched in this paper. We have a look at certain national institutions through CEO characteristics. Below, first of all I give a theoretical background via the upper-echelon and legitimacy theory, which will result in different hypotheses.

2.2 Upper-echelon theory

According to Peni (2014), the CEO is the highest ranking executive within a firm, and is responsible for managing the organization. He has a big role when it comes to decision making, (Fabrizi et al., 2013) which means CEOs can have a big influence on the way the firm performs.

Hambrick & Mason (1984) created the general upper-echelon perspective. This perspective focusses on top management (“top decision-makers”) in firms and brings the team level of analysis into the area of strategic management. For example, this theory states that performance can be partially predicted by looking at the background of a CEO. According to Rost & Osterloh (2010) this upper-echelon theory says that, when trying to explain the

(10)

10

choices that are being made by TMTs, you have to look at the mental models of TMTs. In another research they say that “the linkage between TMT characteristics and firm performance can be used as a valid indicator of real psychological and social processes that mediate between executives’ demography and their behavior” (Rost & Osterloh, 2009). Because of their experience and values, executives will process information regarding the firm itself, its environment, rivalry and all other aspects that have to be faced when running a firm. These attributes will affect the actions and decisions made by executives (Geletkanycz & Sanders, 2012). This means upper-echelon will have an influence on the strategy that a firm follows, which has implications regarding social issues. For example, Manner (2010) investigates the impact of the CEO having a bachelor’s degree in humanities, the width of his or her career experience and being female on CSP and found significant positive relationships, which implies CEO characteristics and CSP are directly linked. Kaplan et al. (2012) say there is a positive relation between subsequent performance and general abilities from the CEO. This paper will focus on the relationship between CEO age and tenure on the one hand and CSP on the other hand.

2.2.1 CEO age

The first CEO characteristic that will be investigated is the age of the CEO, which is measured in years.

For decades, CEO age is a subject that has been researched (e.g., Mudrack, 1989; Peterson et al., 2001; Sundaram and Yermack, 2007). Results from these researches show that individuals for instance become more conservative as they get older, which, according to Huang et al. (2012), will provide less aggressive earnings management by the CEO. Research on the relation between CEO age and compensation shows a weakening linkage over time (McKnight, 2000). Another finding from McKnight is the non-linear relationship between CEO age and the level of the bonus. In accordance with the research by Huang et al. (2012) this non-linear relationship may imply that because CEOs show less focus on aggressive earnings management and their own bonuses, the older a CEO gets, the less attention he may give to profits (financial performance). Instead of the focus on financial performance, a CEO may focus more on other aspects of performance, like CSP. This mechanism, the so called legacy perspective, will be explained later on in this paper.

Bertrand & Schoar (2003) suggest that, as said before, older CEOs are more conservative and thus may have an impact on firm performance, but this impact could be either positive or negative. Hirshleifer (1993) says that younger CEOs may want to focus on short-term goals because of the desire to let their reputation grow, which can be in contrary to older CEOs who

(11)

11

may be in the stage of maintaining their career (Kunze et al., 2012). This is referred to as the resistance to change relationship. It implies that older CEOs will be more hesitant in making decisions that will have big influences. Kunze et al. (2012) also say that older workers are assumed to be more cognitively rigid and are associated with lower potential for development, both resulting in higher levels of RTC. On the contrary, Kunze et al. (2012) also link RTC with individual performance outcomes and find a positive relationship. This may imply that the older a CEO is, the higher his RTC is and this will thus result in higher performance, of which CSP is one. Another result they find is that older employees seem to be better performers in terms of higher goal accomplishment, which also implies a positive relationship between age and (social) performance. Another reason for better performance by older CEOs may be that the older a CEO is, the more knowledge he will have about life and entrepreneurship. This may result in better understanding of, and reaction to certain problems that arise when taking decisions for the firm. These higher resources due to better firm performance can be used for investments that will heighten social performance.

Also, the RTC-CSP relationship can be positive when looking at the CEO age and CSP due to the higher knowledge about the firm and entrepreneurship (Kunze et al., 2012). The CEO may understand that investments in CSP will also be good for the firm because society will have a more positive image of the firm.

On the other hand, the RTC perspective can have a negative impact on firm and social performance, because older CEOs are in the stage of maintaining their career and have lower potential for development (Kunze et al., 2012). This means CEOs may not want to pay attention to the social performance of their firm. Also, CEOs can be more hesitant in making decisions that have big influences, like investments that will result in higher CSP.

As these earlier results show (Bertrand & Schoar, 2003; Hirsleifer, 1993; Kunze et al., 2012), the relationship between RTC, decision making and performance does not show one certain direction. Therefore, I conclude that although RTC and CSP are related, the direction in which the relationship goes is not clear.

Another way of linking CEO age to (different kinds of) performance is through the legacy perspective. Legacy, in this case, is described as that which a CEO will hand down to their predecessor after their pension (by choice or death) or as Zacher et al. (2011) call it: “individuals’ convictions about whether they and their actions will be remembered, have an enduring influence, and leave something behind after death”. For example, when a CEO is coming closer to his retirement, he may look at other factors than just the financial performance of his firm. Also, he may be aware that as he is getting older the chance of dying will rise. These two elements combined are called death reflection and describe the way

(12)

12

people review how others will look at them after their death (Cozzolino et al., 2004; Ring, 1984). Zacher et al. (2011) show that as leaders of firms get older, they more and more believe their actions will result in leaving a lasting and (hopefully positive) legacy in the future. This may also implicate that they will try to leave a positive image from their decisions while still in the function of CEO.

The legacy perspective can be linked easily to CSP . When a firm is already performing well, CEOs and managers may want to change their actions, resulting in supporting social instead of economic performance. Examples can be the funding of mentally or physically handicapped instead of normal functioning employees or using biological produced materials in their own production process. Although this will cost the firm more money, people may consider it as actions that will benefit the community.

Although the legacy perspective shows a positive relationship between CEO age and CSP, the impact of the RTC-relationship remains unclear. In view of the competing results from the RTC-relationship and legacy perspective, I arrive at the following non-directional hypothesis.

“H1: CEO age is associated with corporate social performance”

2.2.2 CEO tenure

Like CEO age, CEO tenure is a subject that has been widely researched (Simsek, 2007). His research shows a relationship between CEO tenure and job performance and also says the influence of tenure on risk taking can be seen as the lower half of an inverted-U relationship. According to Henderson, Miller & Hambrick (2006), after 10 to 15 years of a CEO’s tenure firm performance will start to suffer due to the saturation of the CEO’s learning curve. Miller (1991) also talks about a CEO being ‘stale in the saddle’, which means that the CEO is less open to change and will ignore the environment of the firm. Miller shows that CEO’s will eventually fail to match the firm’s strategy and structure with environment, resulting in loss of the organization’s performance. This in turn may lead to less investments in social performance by the firm. Also, according to Meyer (1975), CEOs with longer tenure will have more power to resist external pressures, like the pressure to invest in CSP. Meyer (1975) says this is because the capacity of leadership to ignore environmental demands increases over time, resulting in the CEO taking less note of external demands or pressure.

On the other hand, according to Peni (2014), executive experience and quality have a positive impact on firm performance. This result is consistent with Baysinger & Hoskisson’s (1990), who say that “as an executive’s tenure advances, the executive has more firm-specific knowledge and a better ability to monitor and provide valuable resources, which may improve

(13)

13

the firm’s performance”. All of the previous may imply a CEO will have more experience and knowledge regarding the firm as tenure gets higher. This will have a positive effect on his decision making, which can result in higher firm performance. In turn, this may lead to a growth of CSP.

Also, Dikolli et al. (2014) find that a firm’s attention for monitoring the CEO will decline over his tenure. This may result in the CEO having more control over the firm and his decisions may be accepted earlier. When a CEO has strong incentives to invest in CSP heightening elements, these will thus be accepted earlier. On the other hand, when a CEO is hesitant in improving social performance, the earlier stated mechanism works the other way around.

When a CEO has been working longer within the company (higher tenure), chances that he will have to leave are bigger. Consistent with the legacy perspective, a CEO then may want to think about the image he will leave when he has to resign. This will result in the CEO giving more attention and resources to investments in CSP heightening factors, trying to leave a positive image of himself.

As Baysinger & Hoskisson (1990) say, tenure is positively related to firm performance, of which CSP is a part. Like the legacy perspective, this mechanism contributes to the positive relationship between CEO tenure and CSP. On the other hand, the stale in the saddle perspective (Miller, 1991) shows that higher CEO tenure may result in less investments in CSP. Because of these contradicting arguments, and as research by Simsek (2007) talks about an inverted U-relationship, I conclude that the relationship between CEO tenure and CSP remains unclear. In accordance with the arguments above, I hypothesize:

“H2: CEO tenure is associated with corporate social performance”

2.3 Institutions and legitimacy theory

National institutions can be described as “the actions, rules, social structures, and practices that persist over time and are features of social aggregates that are larger than a single organization” (Murmann, 2003). Researchers continue to argue in which way these national institutions matter, although they do agree they have a certain influence (Doellgast et al., 2009). According to DiMaggio & Powell (1983), firms are influenced by national institutions and firms will adapt their strategy to these institutions. As said by DiMaggio & Powell (1983), several studies show that institutions will not give the same results on a national level in different countries. This so called institutions based view is also used by van Essen et al. (2012). They state there can be made a difference between formal and informal institutions.

(14)

14

Formal institutions are usually created and enforced by the state, which can provide incentives for managers and firms to “rationally pursue their interests and make choices” (Peng & Khoury, 2008). Informal institutions, rather than really created, emerge from reactions to encountered social or economic problems (van Essen et al., 2012) and are more created from the expectations that the public has about a firm. They are maintained through the continuance of behavior and give value because the effects of complying with the long-term results are bigger than the expected results from non-compliance in the short turn. These institutions can result in more pressure to firms. Examples are regulations on emission from production (formal) or loss of brand name capital through personal reputation damage from not working in accordance with the expectations that the public about the CEO or firm (informal). This in turn can be linked to another view on firms, the government and society, called the legitimacy theory.

According to Deegan (2002), the only reason why firms exist in this world is because society gives them legitimacy. A good legitimate status can be seen as a way of easy access to resources which will lead to long term survival and access to markets (Brown, 1998). In this research, I follow the definition by Arnold et al.: “the congruence of organizational results with institutional norms” (Arnold et al., 1996), and the lack of legitimacy is described as the perception that an organization is not acting according to regulations and social norms and values. It consists of two parts, namely the legislative and the non-legislative part. The legislative part consists of the rules that are shown clearly by the regulators. The non-legislative part has a more subjective look and consists of the way society views managers’ actions. According to Gray et al. (1996), these two can vary a lot. For example, a firm may produce weapons, which is not socially accepted. This will result in the public not granting it social contracts which may lead to increased difficulty in getting resources (non-legislative). A good example of the legislative consequences is the accounting firm Arthur Andersen. In 2002 they were found guilty of criminal charges in the United States, which led to the disappearance of their world-wide operations. Research by Vanhonacker (2000) finds that when a firm loses its legitimacy, social support disappears.

Also, legitimacy theory can be seen different from other rational theories. It can provide a framework for decision-making, in which the decisions people make are influenced by the assumption that those of (legitimated) organization’s or persons are correct and must be followed (Zelditch, 2001). This can be used to link legitimacy theory to CSP. To keep a constant flow of resources from society, firms will want to make decisions that are accepted by society. A loss of legitimacy may result in the loss of resources because the people around the organization will stop granting them these resources. To regain these resources, the firm

(15)

15

may want to act more in accordance with what the public wants, like increasing investments in factors that can increase CSP (and thus legitimacy). Deegan & Rankin (1996) researched the relationship between the reporting of favorable environmental information in periods of increased media attention and found a positive relationship. Hogner (1982) investigate whether legitimacy theory is an explanatory factor for CSR disclosure and finds that these disclosures are made in response to societal factors. Hogner’s research (1982) suggests that legitimacy and CSR are positively related and legitimacy can be a valid theory to explain why firms may want to invest more in CSP related factors.

To show how decisions by managers regarding social performance are influenced, we can both look at the institutions-based view and the legitimacy theory. First of all, formal institutions can be combined with the legislative part of the legitimacy view. Governments may maintain strict regulations which results in firms having (for example) less emissions. This heightens their environmental and thus corporate social performance. Through the non-legislative view and informal institutions, managers will feel pressure from society to socially perform well. When his firm is facing negative attention due to events that damages the firm, a manager may want to decide on taking actions that will change this attention in a more positive way. This results in higher CSP as well. These thoughts are in accordance with research by Ioannou & Serafeim (2012), who provide empirical evidence that nation-level institutions play a role in CSP variation and Hogner (1982), who says CSR disclosure is partially explained by societal institutional factors.

In the remainder of this research, the national institutions ‘environmental policy’ and ‘political stability’ will be looked at in relation to CSP.

(16)

16

2.3.1 Environmental regulations

According to Cormier et al. (2004) and Henriques & Sadorsky (1996, 1999), various stakeholders are concerned with corporate environmental issues. For this, several reasons can be given, like the fact that more and more natural resources (oil, gas) are being used in production processes, resulting in scarcity.

Earlier performed studies imply that firms take regard of the environmental influences from their operations, like Gouldson (2004) and Ambec & Lanoie (2008) who state that the success of businesses are increasingly influenced by elements like environmental risk and performance.

To control environmental problems like pollution, growing scarcity of natural resources and CO2 emission from production processes and the products themselves, governments may want to set regulations that firms will have to follow. On the one hand, researchers say that environmental regulations will lead to additional costs for firms (Palmer et al., 1995; Walley & Whitehead, 1994). This is based on the view that firms have decreasing marginal net benefits because of pollution decreasing and environmental improvements. On the other hand, Porter (1991) and Porter & van der Linde (1995a,b) give arguments that environmental regulation can improve performance. According to Porter (1991), both social welfare and private benefits will be created when environmental regulations arise. This implies firms will pay attention to what society thinks of their impact on the environment, which is in accordance with the informal institutional-based view (“emerge from reactions to encountered social or economic problems” (van Essen et al., 2012)). Porter & van der Linde (1995a,b) say that because of economic inefficiency due to pollution, environmental innovations can “partly or fully offset the cost of complying with environmental law”.

We can use all of this information to explain the link between environmental regulations and CSP. In accordance with the legitimacy theory, in countries where there are higher environmental regulations, firms will have to invest more in CSR to keep their legitimacy and thus their flow of resources from society. Environmental regulations can be seen as a part of governmental regulations, which are formal institutions. Informal institutions like, for example, activist groups that support a firms decisions, may result in the previously stated increased performance. Thus, I hypothesize:

(17)

17

2.3.2 Political stability

De Romilly (1977) states that people have long been fascinated by problems regarding political stability. According to Ahmed & Pulok (2013), in the process of development in any country, political stability generally plays an important role. In this research, we follow Campos & Karanos (2007) and use the term political stability to describe the state of a country. For example, Campos & Karanos (2007) use the height of informal political stability by describing the way people think of a country and how they act, like strikes. Formal political stability is described as the constitutional changes, legislations and regulations. In this research we make use of the “Political Stability and Absence of Violence” variable from the Worldwide Governance Indicators made by the Worldbank. This indicator can be described as the likelihood that the government of a country will be overthrown or altered by violent means, including terrorism and motivations due to other political disagreement like a coup. Kaufmann et al. (2010) review this methodology and conclude the indicators permit meaningful cross-country comparisons.

Aisen & Veiga (2013) perform an empirical investigation on the relationship between political instability and economic growth. Results show that higher degrees of political instability are associated with lower growth of gross domestic product (GDP) per capita. Abeyasinghe (2004) and Feng (2001) say that instability can have serious consequences on country performance due to the fact that investment and savings, which are two essential factors for economic growth, will be “affected adversely by political upheavals”.

Van Essen et al. (2012) say that a well-functioning legal system provides institutional investors and can result in higher incentives for CEOs to perform better in the long-term, because investors will react stronger to a well performing firm. This eventually results in higher country performance. Another reason can be that companies may have a better understanding of how to behave in a politically more stable country, due to the fact that there is a better understanding of regulations, and there is less chance that regulations will drastically change. All of this can imply formal institutions also help when it comes to improving performance and combined with the earlier stated researches suggest that a politically more stable country performs better.

To investigate whether political stability and CSP are linked and how, we can look at the legitimacy theory.

According to the legitimacy theory, when a firm lacks legitimacy it “is not acting according to regulations and social norms and values”. In this research, I look at the strictness of regulations in a country as the legislative part of the legitimacy theory. This means the government in the country will have a more strict policy regarding firms taking decisions with

(18)

18

regard to, for example, environment and human rights. When there is more political stability, companies will feel a stronger pressure to act according the way the government wants and it is less likely that there will be a strong change in regulations. Both the government and the public (non-legislative) want that firms have higher social performance, for example because it is better for the environment. In the case of higher political stability this will thus lead to a stronger connection between the influence of regulations on the behavior of the firm.

The non-legislative part consists of what society thinks of the actions managers take. Research by Ioannou & Serafeim (2012), who say that countries where values and beliefs are stronger will have higher CSP, support this mechanism. As with the pressure from governments, in a country where the community is more concerned with environmental elements, people tend to show their (dis)content to actions by firms earlier. This means firms may have stronger incentives to act the way society wants.

This means that in politically more stable countries, regulations are less likely to change, which results in society having more expectations with regards to social responsibility (which is the required legitimacy). To satisfy society and thus keeping legitimacy, firms will tend to invest more in CSP increasing elements. Also, managers’ actions will be more closely monitored, and the decisions they make are more influenced by society and the government. Also, firms in countries with higher political stability may have a better understanding of what the government expects of them (where they stand).

I state that political stability can be seen as a both informal and formal institution. The formal side consists of governmental regulations, while things like activists or media attention can be seen as informal institutions.

Following the legitimacy theory and earlier research, arguments are provided that show political stability improves social performance. I hypothesize:

“H4: More political stability will result in higher corporate social performance”

2.4 Moderating effects

In this research not only the direct effects of certain national institutions on CSP will be tested but also the moderating effects on the previous stated CEO characteristics with regard to CSP. In their paper, van Essen et al. (2012) also look at the moderating role of formal and informal institutions. They describe the relationship between corporate performance and executive compensation and find evidence that cross-country differences between this relationship can for a significant part be explained by formal and informal institutions.

(19)

19

This means national institutions play an important role in explaining why there could be differences between countries when looking at the CEO characteristics-CSP relationship.

As is shown by Porter (1991), both on social and private level, there will be benefits from higher environmental regulations. As the (informal) institutional-based view shows, society has an impact on firms’ actions regarding the environment. Firms that have production processes that have a big negative impact on the environment may be judged by the people and can receive negative (media) attention. According to the legitimacy theory, this means this firm receives less resources, which leads to worse (social) performance. This implies the decisions CEOs make will be influenced by society as well.

Like Zelditch (2001) says, the CEOs decisions will follow from influences by certain groups, like society (non-legislative) and governmental regulation (legislative). The stronger these legislative elements are, the higher monitoring on CEOs will be.

The legacy perspective shows that CEOs who are older and are in the function of CEO for a longer time (higher tenure) may want to change their decision taking from economic to social and environmental performance. On the other hand, the stale in the saddle perspective and RTC-relationship give arguments that may lead to a negative relationship between CEO tenure and age on the one hand and CSP on the other hand.

When this is combined with the statement that higher environmental regulations result in higher monitoring on CEOs, the legacy perspective implies that CEOs who are older and have higher tenure may even have more motivation to, hopefully, leave a good legacy by making decisions that are good (less worse) for the environment. These actions will be picked up by society and the government, which results in higher legitimacy. Also, when environmental regulations are higher, a CEO will have stronger incentives for acting socially responsible, thereby maintaining legitimacy.

On the other hand, when there will be more monitoring on the CEO through more environmental regulations, it means the CEO has less room for decision making, which weakens the relationship between environmental regulations on the one hand and the CEO characteristics – CSP relationship on the other hand. When looking at the RTC relationship, this implies that a CEO may not want to look at social performance because he is hesitant in making decisions that have a big influence. Also, when looking at tenure, a CEO who is longer in the company may be less open to change, and as Miller shows, the CEO will eventually fail to match the firm’s strategy.

(20)

20

Although I state that environmental regulations have a positive effect on CSP, it will not influence the unclear direction between CEO age and tenure on the one hand and CSP on the other hand.

Thus, I hypothesize:

“H5a: More environmental regulations strengthen the association between CEO age and CSP”

“H5b: More environmental regulations strengthen the association between CEO tenure and CSP”

As is shown earlier, in a country that is politically more stable, people tend to give more attention to legitimacy and thus to activities by firms that will improve social performance. The legacy perspective shows that the older a CEO gets and the higher his tenure will be, the less his focus may be on the economic results by the firm, and may give more attention to the image that society will have of him when he leaves the firm. In research by Baysinger & Hoskisson (1990), it is stated that the higher the tenure of the CEO is, the more firm-specific knowledge he has. This can result in better monitoring regarding valuable resources. In this case, it may imply the CEO uses these resources to invest in, for example, machines that produce less emissions. This leaves a better social and environmental image, resulting in a more positive legitimacy.

On the other hand, Hirshleifer (1993) shows that a younger CEO may want to have higher motivation to focus on firm improving goals, in contrast to an older CEO who is in the stage of maintaining his career. Also, Kunze et al. (2012) show that an older CEO can have lower career development.

As is shown by Ioannou & Serafeim(2012), in countries with higher political stability, the relationship between values and beliefs on the one hand and CSP on the other hand is stronger.

When combined, it shows that in a country where there is more focus on social performance due to higher political stability, a CEO may want to have even higher incentives to leave a good image because society attaches greater importance to this aspect. On the one hand, this means the CEO will want to have higher CSP to leave a more positive legacy. On the other hand, an older CEO who has been in the firm longer (higher tenure), can be more cognitively rigid.

(21)

21

I conclude that the influence of political stability has a strengthening effect on the relationship between CEO age and tenure and CSP, but like hypotheses 5a and 5b we do not now in which direction this relationship points.

Therefore, I hypothesize:

“H6a: Higher political stability strengthens the association between CEO age and CSP” “H6b: Higher political stability strengthens the association between CEO tenure and CSP”

2.4 Conceptual model CEO tenure CEO age Environmental regulations Political stability CSP

(22)

22

3. Methods

3.1 Databases and sample

In this research, several databases will be used. Most of the variables that are being used come from Thomson Reuters’ ASSET4 database. As described by Ioannou & Serafeim (2012), this database specializes in giving “objective, relevant, auditable and systematic information and investment analysis tools to professional investors”. From this information environmental, social and governance (ESG) metrics emerge that will be used in this research. According to Thomson Reuters’ themselves, the ESG database can be used for decision-making regarding issues like climate change, environmental risk, health & safety and executive remuneration. They are collected and standardized by over 120 analyst to ensure accuracy and compatibility and the database contains information on more than 4000 worldwide companies and over 750 data points which are used to cover every aspect of sustainability report. Also, over 280 key performance indicators (KPI’s) are featured, which include normalized scored and actual computed values to ensure ratings that are updated twice a week5. Some of these data points are: CO2 Equivalents Emmission Total (Environmental), Total Injury Rate (Social) and Anti-Takeover Devices (Corporate Governance).

To measure the CSP of firms in this research, and relate it to the CEO characteristics and control variables, this database will be used. In table 1 below, all variables that I will use in this research are shown, as is the regression model.

To measure the influences of the national institutions, two databases are used.

First of all, to measure the height of environmental regulations, the Environmental Performance Index (EPI) by Yale is used. Secondly I use one of the six Worldwide Government Indicators (WGI) made by the Worldbank, namely the Political Stability and Absence of Violence indicator, to test the influence of political stability.

The data in this research comes from the fiscal years 2002 to 2010. After deleting non usable data with regards to all variables used, a sample of 7732 unique observations was left.

5

(23)

23

3.2 Dependent variable

3.2.1 Corporate social performance

According to Ioannou & Serafeim (2012), there is a big challenge when it comes to the construction of a measurement method for CSP due to multidimensionality of theoretical construct and, when only a single aspect of CSP gets measured, this will lead to a limited perspective on the firm. Also, as Waddock & Graves say, there is “need for a multidimensional measure applied across a wide range of industries and larger samples of companies” (Waddock & Graves, 1997). In different researches, several CSP measurement elements have been used, like behavioural and perceptual measures (Wokutch & McKinney, 1991), content analysis of corporate documents (Wolfe & Aupperle, 1991) and measures from case study methodologies (Clarkson, 1991).

In this research, we follow Ioannou & Serafeim (2012) and use the global ESG dataset from Thomson Reuters’ ASSET4 to measure CSP.

As stated before, over 750 data points are used to identify more than 250 KPI’s which are then divided into economic, environmental, social and corporate governance performance. All of these four pillars form their own z-score which ultimately results in a score between 0 and 100%. This percentage can be used as a benchmark for comparing a firm’s performance with that of other firms (Ioannou & Serafeim, 2012).

To measure CSP in this paper, only the environmental and social performance pillars will be used. The first pillar, environmental performance, consists of the following categories: Resource Reduction, Emission Reduction and Product innovation. The social performance is consists of: Employment Quality, Health & Safety, Training & Development, Diversity & Opportunity, Community and Product Responsibility.

Both the environmental and social performance result in a score from 0 to 100%, which will then be equally weighted to form the CSP. This method is in accordance with the one used by Ioannou & Serafeim (2012).

3.3 Independent variables 3.3.1 CEO age

CEO age means how old the CEO is and will be measured in years. Research by Kunze et al. (2012), who also uses this variable, shows that there is a relationship between CEO age and performance, although the direction in which this relationship points remains unclear. The data from this variable comes from Thomas Reuters’ ASSET 4 dataset.

(24)

24

3.3.2 CEO tenure

CEO tenure consists of the years that the CEO has been in this function within the company and will be measured in years. Earlier research (Baysinger & Hoskisson, 1990; Miller, 1991) also use this variable, but analyses shows contradicting results between CEO tenure and performance. Like CEO age, the data in connection with CSP comes from Thomas Reuters’ ASSET 4 dataset

To measure the influence of environmental regulations on CSP, the Environmental Performance Index (EPI) by Yale will be used. The EPI ranks how good countries are performing on issues within the environment. It ranks over 120 countries (differing between years) with the use of over 20 or more indicators (also differing between years) that reflect national-level environment data. For example, the EPI In 2012 existed of 22 indicators that were put in 10 categories, which were: Environmental burden of disease, Air pollution (effects on humans), Water (effects on humans), Air Pollution (effects on ecosystem), Water (effects on ecosystem), Biodiversity & Habitat, Agriculture, Forestry, Fisheries and Climate Change. This results in a ranking in countries. These country rankings will be linked to the companies in the database. The country where the company is linked to belongs on the place where the headquarters is placed.

For measuring the height of political stability in a country, the Political Stability and Absence of Violence indicator will be used. This indicator is one of six, which all together lead to the Worldwide Governance Indicators (WGI). In total, the WGI report consists of six dimensions on governance for over 200 countries in the years 1996 to 2012. To get to these indicators, thirty one underlying data sources are used which get rescaled and combined. Kaufmann et al. (2010) summarize the methodology of the WGI and related analytical issues. They conclude that there are some margins of error, but in the end the WGI can be seen as a meaningful indicator system to make cross-country and over-time comparisons.

In this research, over 50 countries will be used, which will then be ranked from by using the political stability from the WGI. As with the environmental regulations, the link between the companies from the database and the country it is based in, depends on where the company’s headquarter is located.

(25)

25

3.4 Control variables 3.4.1 Firm size

The first control variable I will make use of in this research is firm size, which is measured in a firms total assets, as is done by Ioannou & Serafeim (2012) and Waddock & Graves (1997) as well. As shown by Ioannou & Serafeim (2012), firm size and CSP have a positive relationship. Earlier research by De Villiers et al. (2011) shows that the larger firms get, the more likely they are to identify environmental issues. Also, Waddock & Graves (1997) state that firm size is a relevant variable because a larger firm will exhibit more socially responsible behaviours. These elements show that there may be a relationship between firm size and CSP.

3.4.2 Firm performance

Another control variable that will be used is firm performance. Research by Ioannou & Serafeim (2012) shows better performing firms are more socially responsible. To measure firm performance, return on assets (ROA) (net income divided by average total assets) will be used, like Ioannou & Serafeim (2012) do as well. Waddock & Graves (1997) found a significant relationship between ROA and CSP, which implicates it can be seen as a useful control variable.

3.4.3 Board independence

The third control variable that will be used is the independence of the board. De Villiers et al. (2011) find that the more independent a board is, the higher environmental performance is, due to a higher level of effective monitoring. As independent board members have more influence on their reputation, they will take social elements more into account when making decisions (De Villiers et al., 2011). Board independence will be measured by looking at the number of independent board members divided by total board members, which comes from Thomas Reuters’ ASSET 4 dataset.

3.4.4 Dummy variables

To look whether CSP will differ in several industries and years, two dummy variables will be used. First of all, industries are divided into five sectors (Consumer, Manufacturing, Technology, Healthcare and Others) by using the Fama-French industry classification (Fama & French, 1997). This means they are given certain codes (SIC), to account for differences in CSP between industries. These differences may arise because firms in industries that are environmentally more sensitive will disclose more information about environment (Cho &

(26)

26

Patten, 2007). This method is also used by Waddock & Graves (1997) and Ioannou & Serafeim (2012).

Secondly, the probability exists that there are events which have an impact on CSP, although these events are not due to things within this research. For this, I use years as a dummy variable. The first year, 2002, is used as the reference group. If the observation was equal to the column year, a value of 1 is given and if it isn’t the same, a value of 0 is given.

3.5 Analysis

Instead of using the real value of the total assets, first of all I standardized them by taking the natural logarithm (LN) of its values in the database.

Secondly, I winsorized CEO age, CEO tenure, total assets and ROA, by giving values that are bigger or smaller than the average plus or minus three times the standard error the same value as the average plus or minus three times the standard error. I did not winsorize the CSP, environmental regulations, political stability and board independence because these are all scored on a scale from 1% to 100%.

By using the winsorized values instead of their real values, I try to reduce the effects of outliers.

(27)

27 Table 1. Variables Name Measurement CSP Corporate social performance ASSET4 ESG

CEOAGE CEO Age in years ASSET4 ESG

CEOTEN CEO Tenure in years ASSET4 ESG

ENVREG Environmental regulations EPI Yale Index

POLSTAB Political stability Political variable WGI

FIRMSIZE Firm size measured by total assets

ASSET4 ESG

FIRMPERF Firm performance measured by ROA

ASSET4 ESG

BOARDIND Percentage of independent board members

ASSET4 ESG

DUMSEC Industry classification based on SIC and Fama-French classification

ASSET4 ESG and Five industries Fama-French classification

DUMYR Year ASSET4 ESG

To test the hypotheses in this research, the following regression model will be used:

CSP = β0 + β1 * CEOAGE + β2 * CEOTEN + β3 * ENVREG + β4 * POLSTAB + β5 * FIRMSIZE + β6 * FIRMPERF + β7 * BOARDIND + β8 * DUMSEC + β9 * DUMYR + β10 * CEOAGE * ENVREG + β11 * CEOAGE * POLSTAB + β12 * CEOTEN * ENVREG + β13 * CEOTEN * POLSTAB + εi

β0 = constant βi = coefficients εi = distortion term

(28)

28

4. Results

4.1 Descriptive statistics

In table 2 the descriptive statistics with regard to the dependent variable, independent variables and control variables are shown.

CSP, which can have a score between 0 and 100, has an average of 49.21. The highest score is 97.82, while the lowest scoring country has a score of 6.33. Standard deviation is 29.608. These numbers show almost similar results as research by Ioaunnou & Serafeim (2012), who found a mean of 49, a maximum of 98, a minimum of 6 and a standard deviation of 29. The first independent variable, CEO age, is measured in years. The average CEO in this research is 54.4 years, whereas the youngest is 32.4 and the oldest 76.2 years, due to winsorizing. Standard deviation is 7.089 years.

The average of CEO tenure is 8.4 years, while the lowest tenure is 0 years, which means the CEO just got appointed. The longest seated CEO is in this function for 34 years, and standard deviation has a value of 8.305 years. As age, CEO tenure has been winsorized.

On average, countries have a score of 60.16 points on the Yale Index, with a standard deviation of 6.409. The highest score is 77.99 and the lowest is 34.21 points.

With regards to the political stability, countries have a score of 67.06 points on average. The minimum is 5.19, while the maximum is 100 out of 100 points. Standard deviation is 17.546.

Table 2. Descriptive statistics

Mean Minimum Maximum Standard Dev.

Dependent variable CSP 49.21 6.33 97.82 29.608 Independent variables CEOAGE 54.4 32.4 76.2 7.089 CEOTEN 8.4 0 34 8.305 ENVREG 60.16 34.21 77.99 6.409 POLSTAB 67.06 5.19 100 17.546 Control variables FIRMSIZE 15.787 10.231 21.375 1.784 FIRMPERF 6.309 -23 35 7.88 BOARDIND(%) 60.02 0 100 27.362 Note: (N = 7732).

(29)

29

Thirdly are the control variables. Firm size, which is measured by using the natural logarithm of total assets (in USD), has an average score of 15.787, a maximum of 21.375 and a minimum of 10.231, which is also due to winsorizing them. Standard deviation is 1.784. As with firm size, the firm performance has been winsorized as well. It is measured by using the return on assets (ROA). On average, companies have a ROA of 6.309. The best scoring company has a ROA of 35, while the minimum is -23 and standard deviation is 7.88.

On average, a board has 60.02% independent members. Both minimum (0%) and maximum (100%) score their lowest and highest possible scores, and standard deviation is 27.362.

(30)

Table 3. Pearson correlations for

multicollinearity CSP CEOAGE CEOTEN ENVREG POLSTAB FIRMSIZE FIRMPERF BOARDIND

CSP 1 CEOAGE 0.043** 1 CEOTEN 0.098** 0.295** 1 ENVREG 0.013 -0.009 0.026* 1 POLSTAB 0.016 0.004 0.05** 0.484** 1 FIRMSIZE 0.385** 0.140** -0.071** -0.01 -0.036** 1 FIRMPERF 0.006 -0.043** 0.070** 0.013 0.009 -0.153** 1 BOARDIND -0.003 0.063** -0.046** -0.023** -0.013 0.013 0.023** 1

(31)

Table 4. Multiple regression analysis (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) Constant -38.895*** (5.04) -48.438*** (5.847) -49.023*** (5.755) -47.409*** (6.469) -47.035*** (6.001) 52.587*** (0.289) 52.404*** (0.327) 52.547*** (0.282) 52.362*** (0.321) 53.016*** (0.338) FIRMSIZE 7.072*** (0.116) 8.033*** (0.143) 7.953*** (0.166) 7.378*** (0.142) 7.482*** (0.138) 14.992*** (0.31) 15.038*** (0.355) 15.125*** (0.301) 15.087*** (0.346) 15.787*** (0.372 FIRMPERF 0.243*** (0.025) 0.345*** (0.29) 0.349*** (0.032) 0.318*** (0.031) 0.318*** (0.03) 3.241*** (0.278) 3.465*** (0.309) 3.227*** (0.271) 3.392*** (0.301) 3.427*** (0.311) BOARDIND -0.52*** (0.007) -0.43*** (0.01) -0.94*** (0.012) -0.71*** (0.009) -0.07*** (0.009) -2.095*** (0.338) -3.430*** (0.383) -2.058*** (0.329) -3.296*** (0.374) -3.79*** (0.398) CEOAGE (H1) -- -0.054* (0.033) -- -- -- -1.049*** (0.276) -- -0.93*** (0.268) -- -1.81*** (0.327) CEOTEN H(2) -- -- 0.214*** (0.031) -- -- -- 1.665*** (0.301) -- 1.672*** (0.291) 1.839*** (0.322) ENVREG (H3) -- -- -- 0.121*** (0.038) -- 0.924*** (0.278) 0.764 (0.308) -- -- 0.318 (0.346) POLSTAB (H4) -- -- -- -- 0.073*** (0.013) -- -- 1.239*** (0.268) 0.936*** (0.298) 0.884*** (0.343) ENVREG*CEOAGE (H5a) -- -- -- -- -- 0.72 (0.287) -- -- -- 0.567 (0.366) ENVREG*CEOTEN (H5b) -- -- -- -- -- -- 0.309 (0.315) -- -- -0.31 (0.367) POLSTAB*CEOAGE (H6a) -- -- -- -- -- -- -- 0.773*** (0.273) -- 0,429 (0,359) POLSTAB*CEOTEN (H6b) -- -- -- -- -- -- -- -- 0.429 (0.291) 0.169 (0.347)

DUMYEAR YES YES YES YES YES YES YES YES YES YES

DUMSEC YES YES YES YES YES YES YES YES YES YES

R-squared 0.176 0.197 0.184 0.181 0.186 0.205 0.197 0.210 0.2 0.204

Adj. R-squared 0.175 0.196 0.183 0.18 0.185 0.203 0.196 0.209 0.199 0.201

F-value 255.801*** 205.436*** 153.723*** 177.295*** 191.619*** 136.885*** 108.436*** 147.536*** 115.138*** 86.02***

Maximal VIF 1.048 1.098 1.09 1.058 1.056 1.103 1.095 1.105 1.094 1.378

Notes: In all models, CSP score is the dependent variable. ***, ** and * are significant at respectively 1%, 5% and 10% (two-tailed). Standard error is shown between brackets. N = 7732.

(32)

4.2 Multicollinearity

To look whether there may be multicollinearity between the independent and control variables, table 3 shows Pearson correlations. When there are values that are ± 0.7, multicollinearity may be a problem. As is shown in table 3, the highest value is 0.484, which means multicollinearity will not be a problem. Another way of looking for multicollinearity is by using VIF values. If these are over 10, multicollinearity may arise. As can be seen in table 4, there are no VIF values over 1.378, so multicollinearity will not be a problem.

4.3 Hypotheses

To test the hypotheses in this research, multiple linear regressions are made, which are shown in table 4. This table consists of 10 models, where model 1 uses only control variables, models 2 to 9 test the hypotheses (including all control variables) and model 10 uses all the variables at once.

From the first model, using only control variables, it shows that firm size (b = 7.072) and performance (b = 0.243) have a significant positive effect on CSP (p < 0.01), which also stays the same in the other models. This is in accordance with the theory stated in the methodology section. Board independence has a significant negative effect (b = -0.52; p < 0.01). This is not in accordance with previously stated results from earlier research. The adjusted R-squared value of 0.175 shows that 17.5% of CSP can be explained by looking at the control variables.

Model 2 shows the relationship between CEO age and CSP, which also is the first hypothesis. A significant negative relationship is found (b = -0.054), although it is not that strong (p < 0.1). Based on this, we can state that CEO age has a negative effect on CSP. This means that hypothesis 1, CEO age has an effect on CSP, can be accepted, which is in line with research by Hirshleifer (1993) and Kunze et al. (2012), who state that there is a negative relationship between CEO age and performance due to the fact that older workers will be more cognitively rigid, have lower potential for development and that older CEOs are in a more career maintaining stage.

Hypothesis 2 comes from the third model, which looks at the relationship between CEO tenure and CSP. When looking at table 4, results show a positive significant effect (b = 0.214; p < 0.01). This means that CEO tenure has a positive effect on CSP, thereby we can accept hypothesis 2. This is in line with research by Baysinger & Hoskisson (1990), who find that tenure is positively related to firm performance due to the fact that a CEO with longer tenure has more experience regarding the firm. This will positively influence his decision making.

The influence of CEO characteristics and national institutions on Corporate Social Performance