Age and Corporate Social Responsibility
The effect of CEOs’ Age on CSR Performance and the
moderating role of their national culture
Master Thesis
MSc International Business and Management
Double Degree “Managing Multinationals”
Rijksuniversiteit Groningen and Uppsala University
Francesca Pangrazi
S3468070
f.pangrazi@student.rug.nl
Supervisor: Dr. O. Lindahl
Co-Assessor: Dr R.W. De Vries
Faculty of Economics and Business
University of Groningen
ABSTRACT
Table of Content
1. Introduction ... 1
2. Literature Review and Hypothesis Development ... 4
2.1 Corporate Social Responsibility ... 4
2.2 Upper Echelon Theory ... 5
2.3 CEO Age ... 7 2.4 CEO Nationality ... 11 2.4.1 Individualism ... 13 2.4.2 Masculinity ... 13 2.4.3 Uncertainty Avoidance ... 14 2.4.4 Power Distance ... 15 2.5 Conceptual Model ... 15 3. Methodology ... 16 3.1 Data Collection ... 16 3.2 Sample ... 16
3.3 Variables and measurement ... 18
3.3.1 Dependent variable: CSR ... 18
3.3.2 Independent Variable: CEOs’ AGE ... 20
3.3.3 Moderator: CEOs’ National Culture ... 20
3.3.4 Control variables ... 21
3.4 Empirical Data Analysis ... 22
4. Results ... 25 4.1 Descriptive Statistics ... 26 4.2 Correlations ... 27 4.3 Regression Results ... 30 5. Discussion ... 34 6. Conclusion ... 37 6.1 Theoretical Implication ... 37 6.2 Practical Implication ... 38
6.3 Limitations and Future Research ... 38
Appendices ... 48
Appendix A: Dimension Scores ... 48
Appendix B: Histogram & QQ Plot variables ... 49
Appendix B bis: Test for the normality of independent variables ... 50
Appendix C: Plot & Test for the normality of residuals ... 50
Appendix D: Breusch-Pagan results ... 51
Appendix E: Durbin-Watson results ... 51
Appendix F: VIF Test for Multicollinearity ... 51
List of Figures Figure 1: Age distribution ... 8
Figure 2: Conceptual model ... 15
Figure 3: Box-Plot graph ... 26
Figure 4: Simple Slopes graph ... 31
List of Tables Table 1: Descriptive Statistics ... 25
Table 2: Correlation Matrix ... 29
Table 3: Regression Analysis ... 33
List of Abbreviations
CEO Chief Executive Officer
CSR Corporate Social Responsibility CV Curriculum Vitae
1. INTRODUCTION
“There is one and only social responsibility of business: to use its resources and engage in activities designed to increase its profits" claimed Milton Friedman in a famous article (Friedman, 1970, p.6). Since then, many things changed in the business world, starting from the concept of Corporate Social Responsibility that today is considered “not as a cost center but an effective management tool with multi-dimensional benefits” (Gupta, 2009, p.396). Social and environmental problems are now trending topics in the media, the corporates' practices and the consumer preferences are shifting toward a more ethical dimension thus, as a response, CSR has become a priority.
Since the ability of being flexible to meet the changing conditions of the environment and survive (Smith & Grimm, 1987) represents a first motivation towards a sustainable approach, companies need to shift their corporate logic "in which caring for employees, communities, and environment is not only seen as morally correct but also the best approach for continuity and profitability” (Gupta, 2009, p.397). In this new scenario is necessary to consider the stakeholders, that are defined by Freeman (1984) as the group of people affected by the organizations' actions. O’Riordan and Fairbrass (2008) suggest establishing a dialogue with those stakeholders, in order to identify which are the appropriate business behaviors that they expect the firms to follow. Some recent studies confirm that those stakeholders are putting increasing pressure on companies to act responsibly, but sometimes it seems to be really complicated for the firms to understand and meet some conflicting expectations which differ among geographical regions and cultures (Daniels & Radebaugh, 2001; Castka et al., 2004).
Alongside the increasing phenomenon of social responsibility, among the changes that are taking place in our society, we can certainly not ignore the generational shift, where even if the new generations represent 35% of the actual workforce (BlackRock, Larry Fink’s Annual Letter to CEOs, 2018), only a small part of them covers leading roles in the companies, that are instead occupied mostly by older generations. For this reason, since we are talking about strategic decisions and choices, this research will be focused on the figure of the CEOs, seen as the leaders who can show their commitment for specific social issues, expressing which are their priorities and ideas while taking decisions within the boards of multinationals. This research sets out its foundations here, among the desks of those who decide and influence the operations of companies, trying to understand what influence younger people have on strategic decisions regarding CSR.
generations is how HR employee can attract younger workers, if the new generations can be considered ethical consumers or selfish individuals not caring for the social and environmental issues around them (e.g. Kanchanapibul et al., 2013), how the leadership and followership styles change among the older and younger generations (e.g. Shih Yung Chou, 2012). Instead, what has not been taken into deep analysis is if those younger generations can currently be considered as game players in the boards of the firms where they could actually make the difference and express their ideas and values. In fact, even if we know how much is important to implement CSR practices, and the evidence that firms are changing their mentality are many, the studies on what are the generational characteristics that influence the propensity toward social responsibility, if the new generations already have an influence or not, are topics that can still be considered as rare and not contemplated in the current literature. Starting from the idea stressed by Meister and Willyerd (2010) that the younger generations will bring changes and will play an essential role in CSR because they positively influence society towards a stakeholder-centered approach (Reavis, Tucci & St. Pierre, 2017), this thesis aims to fill the before mentioned gap. Thus, when we analyze which are the characteristics of the CEOs studied until now, it is rare to find differences in demographic data, and rather researchers focused on their skills, experiences, gender, or ethnic differences (Huang, 2013). This research instead aims to investigate how age difference is a crucial factor since the directors work together with people from different generations on a daily basis, and this mix of ages is also reflected in the decision-making process where various degrees of importance is given to the CSR. Additionally, a lot of consulting companies are recently encouraging to give more attention to age diversity. They are trying to push firms to introduce a more significant number of younger directors and managers in the top management teams explaining to them the importance of having diverse, non-traditional backgrounds, skills and points of view.
When we speak in generational terms, though, there is always a tendency to generalize. In fact, what is covered by most of the studies are common traits or characteristics of certain generations, for example the fact that younger generations are shaped by the forces of globalization, which is deemed to standardize the tastes, the attitudes, and making ancient national differences disappearing (Friedman & Miles, 2006). Although we might suspect a younger CEO's to be more interested in CSR, it is less clear whether this is equally the case across cultures. Thus, this research aims to analyze how the nationality of origin of younger CEOs influences their propensity towards sustainability issues. Using some of the dimensions created by Hofstede (Hofstede, 1984), this research aims to understand the relationship between the level of individualism, long-term orientation, and uncertainty avoidance of the national cultures of the CEOs involved and their attitudes. Thus, the second research question that will find a response in this paper is: do the CEO's national culture moderate the relationship between CEOs' age and CSR performance?
Thereby, this thesis will contribute to adding to the already existing research a better overview of how a demographic characteristic as the age of the CEOs can influence the CSR performance of different firms, and more in-depth if and how their nationality acts as a moderator for this relationship.
To answer the research questions, this study analyzed a sample based on the Global Fortune 500 firms that are the top worldwide corporations by revenue. The first question was answered relating the age of the CEOs with their firms' CSR scores provided by CSRHub, and the second question was answered testing the moderating of each Hofstede’s cultural dimension, provided by Hofstede Insight.
2. LITERATURE REVIEW AND HYPOTHESES DEVELOPMENT 2.1 CORPORATE SOCIAL RESPONSIBILITY
During the last few decades, the concept of Social Responsibility appears to be increasingly important when analyzing the ability of the firms to adapt to changing environments and needs. However, scholars still have not agreed on one single definition that exhaustively describes this concept; thus, multiple definitions were developed until now, together with the evolution of the CSR idea (Rahman, 2011).
Starting from the 1950s, when the awareness about the CSR began to increase (Carroll, 2008), through the years, also the definitions started to change and to underline different aspects and dimensions. In the 50s the CSR was considered as an obligation to the society; in the 60s definitions were more focused on the relationship between corporation and society; later in the 70s the stakeholders were included, promoting CSR as a philosophy that improves the quality of life and the social wellbeing. During the 80s, the highlight went on the fact that CSR is a voluntary and philanthropic way of acting that could give profits, social support and be both permitted by the law and ethically tolerated. Thus, while through the years the definitions were focusing on single aspects, finally in the 21st century it was proposed a more wide idea that integrated together all the concept mentioned before, considering the CSR as a combination of social and environmental concerns, aimed to enhance life quality and the respect of both human and labor rights, the safeguard of the environment, but also the fight against the corruption and it is also aimed to increase the transparency and accountability of firms’ actions (Rahman, 2011).
responsibility explains the desire of businesses to engage in social activities without being forced but instead in a voluntarily way, pursuing a philanthropic goal. Thus, according to the author, both ethical and discretionary responsibilities are reflected in the concept of CSR.
Additionally, two projects that have been able to achieve the objective of creating comprehensive frameworks are those of Dalhsrud (2008) and the one from Arguinis and Glavas (2012). While the first concentrated on the synthesis of the different definitions, the last two focused on the “understanding of underlying mechanisms and micro-foundations of CSR” (Arguinis & Glavas, 2012, p. 932).
Dalhsrud (2008) created a taxonomy of the 37 most influential and used definition of CSR. In this regard, according to linguists, the most frequently used definitions are considered to be more significant than the ones that are rarely used (Dalhsrud, 2008). For the purpose of this research, it will be taken the definition proposed by the Commission of the European Communities in the Green Paper 3157 of 2001. The frequency count associated with this definition is 286 (Dalhsrud, 2008), which is the highest number recognized in the taxonomy created by Dalhsrud, and additionally, it represents a complete description of the phenomenon which includes more of the dimensions mentioned before. Although the study made by Sheehy (2015) states that there is not a possible universal definition for CSR, the one that will be considered in this research is that one which describes the CSR as “a concept whereby companies integrate social and environmental concerns in their business operations and in their interaction with their stakeholders voluntarily" (EC- European Commission, GreenPaper 3157, 2001, p. 6).
2.2 UPPER ECHELON THEORY
Taking strategic decisions, like being involved in responsible activities, is a duty of the top management who influence the implementation of those practices with their values, as the upper echelon theory claims (Lee, Sun & Moon, 2018).
The upper echelon theory is built on the concept of “bounded rationality” (Simon, 1972), which explains the cognitive limitation of decision-makers. The uncertain situations are not analyzed and understood objectively, but rather, they are interpreted by the top executives (Hambrick, 2007). Accordingly, the managers' perceptions of the different situations they face must be combined with their values in order to generate a strategic decision. Thus, personal values are considered as something that can influence people’s perceptions (Scott & Mitchell, 1972).
It is generally useful to understand which are some of the background observable characteristics of the decision-makers that could predict their choices. Examples of characteristics that have been analyzed until now are education, gender, international experience, tenure and many others. Thus, the reason why organizations take different strategic choices depends on those aspects that differ among the managers, and the upper echelon theory explains this link. Therefore, it could be reasonable to state that the implementation of CSR practices in a company is linked to the individuality of its managers and its CEO. It is important to notice that, in order to be effective, the selection of proper CEOs and managers is needed to align their characteristics with the firm strategic choices. Research until now often connected CEO’s needs of attention or narcissism with CSR initiatives (e.g., Petrenko et al.,2016; Park et al., 2018; Al-Shammari et al., 2019), their religious beliefs (Mazereeuw-van der Duijn Schouten, 2014) or their level of confidence (McCarthy, Oliver & Song, 2017) and many other attributes, both in character and demographic terms.
Among the many characteristics of the CEOs, this study will focus on their age and their nationality as factors that influence their decisions and propensity toward sustainability choices. It is essential to focus on age because until now the studies about it were dedicated to how to manage an aging workforce (Ali et al., 2019), how to manage the age diversity in the teams to achieve better outcomes (Boehm & Dwertmann, 2015), how different CEOs characteristics change as they age (Graham, Harvey & Puri, 2013) or how CEOs take different decisions in various steps of their career, for example, according to how far is their retirement or how strong is their reputation in the firm. Studies on generational characteristics and their propensity toward specific issues like social responsibility are still rare, and this research aims to fill this gap.
to deepen the knowledge of how the national culture can be a moderator on the relationship between the CEO’s characteristics and the CSR performance.
2.3 CEO AGE
Nielsen (2010, p. 305) described the TMT as “the top executives who have a direct influence on the formulation of a firm’s strategy” and literature until now, generally includes in this group old executives, since they are associated with better knowledge of the firm and of the strategies. However, an increasing number of younger executives have an active part in the decision-making process, and studies like the ones made by Ng and Feldman (2008; 2013), analyzed many factors related to firms’ performance differing among workers of different ages. Some examples are the propensity of old employees to achieve extra-role performance (Ng & Feldman, 2008), or younger generations that seem to be more into innovative behaviors (Ng & Feldman, 2013).
Hambrick and Mason (1984) explain the importance of having members of different ages in the top management positions, since this kind of diversity includes a variety of values and perceptions.
The influential role of the CEO
The problem is that literature until now focused on the difference in age among the TMT members, and the CEO role was usually included in this group with no particular distinction. Nevertheless, the CEO position is unique, and it has become more visible in the press in the last decades (Park & Berger, 2004). The CEO image is considered an “intangible asset that is actively managed and promoted by an organization” (Ranft et al., 2006, p.279) since it influences the corporation’s reputation (Gaines-Ross, 2000), stock prices (Harrison et al., 2019) and investors’ behaviors (Quarrels, 2002). The CEO is that figure in the company that Ranft et al. (2006) describe as "an individual at the apex of an organization whose personal reputation can have immediate and long-lasting impact upon the organization” (Ranft et al., 2006, p.281), underlining how CEOs actions, decisions and ideas have a significant impact on firms' performance. They are considered "celebrities" (Ranft et al., 2006) since they attract large-scale public attention and because of their notoriety, the responsibility for the leadership role they cover is continuously under the spotlights.
organization is clear, it is relevant to consider also their involvement and interest in the CSR, since CEOs could result as active promoters of responsible strategies that a firm could implement.
Overview of the current ages present in the TMT
A fundamental fact to analyze is that for a firm is necessary to both adapt to environmental changes and shareholders’ needs in order to be lastingly successful. Lately, investors are encouraging boards to diversify as much as possible; for this reason, many companies have added younger directors to their boards. Additionally, there is a societal change in demographics where the new generations represent 35% of the actual workforce (BlackRock, Larry Fink’s Annual Letter to CEOs, 2018), and they will replace the old ones as the largest living adult generation (Fry, 2016).
Although, the reality is that the younger CEOs are still few, compared to the totality, and most of the directors among the S&P 500 boardrooms are around 75, making the ones aged 50 or under still a minority (BoardEx). PWC provided a younger directors census where their findings show that there are 315 younger directors in the S&P 500, that are aged from 29 to 50 and that, as shown in Figure 1, the distribution of those younger directors is “heavily concentrated at the older end of the range” (Pwc Census of directors 50 and under, April 2018, p.5). Additionally, this report state that fewer than half of those companies have at least one younger director (only the 43%), but also that those companies which have a younger CEO are more likely to have also younger directors. However, it is important to be able to predict and have an idea of who will be the leaders of the future and what their propensities are.
Figure 1: Age distribution
Definition of “age” differences
It is difficult, and it always leads to arbitrary decisions to define and divide people in subcategories by their age. Until now, literature identified some groups in order to be able to point out generational differences regarding consumption, leadership styles, and behaviors. A common division that has been done by researchers is dividing the current population into five categories, that, according to the analysis made by Jain (2016), are Traditionalists (born in the 1945 and before), Baby Boomers (1946 to 1964), Generation X (1965 to 1981), Millenials (1982 to 2000) and Centennials (from 2001 onwards). Some researchers suggest that groups of people can also be categorized by shared cultural events, i.e., living in a period of frequent terrorist attacks, or by radical societal progress, like the introduction of the Internet. The use of “strict” categories could lead to the exclusion of some individuals from the research, perhaps younger or older than some of his colleagues of only one year. For this reason, this thesis did not consider any category to identify the age but instead a comparison among younger and older individuals in the same group.
The role and importance of having younger CEOs
One aspect that seems to be relevant for the changes that new generations could bring is the importance given to the responsibility toward the society and the environment.
The relationship between age and CSR
Generational studies made until now found out that younger generations feel more directly responsible for their actions, and they think to have the chance of making a difference in the actual societies (Cone study, 2015). “Individuals of this civic-minded generation not only believe it is their responsibility to make the world a better place, a majority of them believe that companies have a responsibility to join them in this effort are prepared to reward or punish a company based on its commitment to social causes” (McGlone, Spain & McGlone, 2011, p. 196). This generation is expected to change the corporate culture of the firms where they cover top management positions toward a mentality that not only care for profits but also have a social conscience (Sharp, 2014). One reason why they are concerned about CSR is that in the last decades, training centers and universities have been trying to give them a perspective able to anticipate, prevent and resolve problems that societies will face in the future (Ashton et al., 2017). Suffrin (2017), indeed, as a demonstration, his study showed how MBA students are more likely to have a lower salary and work for a responsible firm than the opposite situation. Borghesi et al. (2014), too, found that younger managers are more prone to invest in CSR activities. Turner (2015) instead focused on the possible reasons, suggesting that something that could explain why some younger generations are more involved and aware of social issues is the historical period they lived in, in which there were two wars (Iraq War and the one in Afghanistan). Wiersema and Bantel (1992) also studied how age influences the decision-making process, finding that younger leaders are less risk averse but more likely to embrace changes.
Hence, all these considerations made above suggest a positive relationship between the CEOs from younger generations and the CSR performance of the firms where they exert their decision-making power. Nevertheless, as mentioned before, contracting results comes from previous studies, and next to the one which supports the younger, there are also the ones who reject the possible correlation of the age and the CSR performance, and some others which support the opposite view like Fabrizi et al. (2014), which state that younger CEOs invest less in CSR compared to their older peers. However, since the increasing pressure that younger generations are perceiving concerning the responsibility they have for the future of our societies and our planet, this thesis is expecting to see a positive relationship between the younger age of the CEOs and the CSR performance of their firms. Thus, the first hypothesis of this research is
2.4 CEO NATIONALITY
A problem that arises when dealing with age-related studies is that generalizations are made about groups of people under analysis. What has been said so far about the characteristics of the younger generations is undoubtedly relevant, but it is also important to realize that there are factors that can act as moderators in the hypothesized relationship.
One factor that certainly influences people's responsible behaviors and decision-making is their national culture (Ferrell & Gresham, 1985). To describe one's culture, the framework of six different dimensions proposed by Hosftede (1984) has been widely used in previous academic research.
A vast amount of studies has been done on trying to understand national differences and the way they shape individuals’ values, beliefs and behaviors (Hosftede, 2011), and consequently also the organizations where they work. Previous research applied it in different fields like economics, sociology, management or marketing, where, for example, the cultural influence was used as a tool to understand consumers’ behaviors (De Mooij & Hofstede, 2010). According to Hofstede (1984), each country scores differently on certain dimensions, and thus the most part of market research and analysis were conducted according to this national clustering. In management studies, culture was largely used since it helps in understanding when a particular strategy that is working in a specific place can result in being inappropriate in another country (Hofstede, 1984). Furthermore, it helps to understand the environment where a firm is operating or would like to operate in, determining the entry modes, the portfolio of countries, and the way to adopt to enter new markets, like the one theorized by the Uppsala model of internationalization (Forsgren, 2002). Some researchers already examined the relationship between CSR and national culture (Veser, 2004), but examining most of the time the nationality of the organizations, thus their country of origin, and rarely the nationality of the TMT.
then too far from what happens in reality, and that they fail to grasp the real needs and dynamics of the current situation.
Hofstede Framework
The framework developed by Hofstede is important in order to see how the cultural dimensions that characterize the national countries of the CEOs influence their points of view regarding the socially responsible practices of the firms. Hofstede defines culture as “the collective programming of the mind that distinguishes the members of one group or category of people from others” (Hofstede, 2011, p. 3) and to create the framework a study on IBM employeeswas conducted between the 1967 and 1973. From the worldwide survey Hofstede made, he theorized four dimensions that could describe the common traits of the individuals who were part of the same cultural block. In the initial work the dimensions were only four: Individualism, Masculinity, Power Distance and Uncertainty Avoidance. Later on, a fifth dimension called long-term orientation was added; it wanted to complement some missing aspects not covered in the original framework that were particularly evident in the Asian countries. In 2010 the last dimension was added; it describes indulgence that is occurring when a society permits gratification of the natural human desires explicated in the chance of enjoying their life and having fun (Hofstede, 2001).
Although more works related to the initial Hofstede’s one has been done, included some like the GLOBE, which extended it including a societal clustering of the 62 analyzed countries (Javidan & Dastmalchian, 2009), the initial work of Hofstede is still considered the most appropriate to frame cultural values on the national level. One reason that supports the choice of many researchers to keep using this framework is that even if Hofstede's dimensions have changed over the past decades, the results of the research of Beugelsdijk, Maseland and Van Hoorn (2015) demonstrate that they are still valid and applicable. In fact, while the modernization theory (Inglehart & Welzel, 2007)predict that cultural values through the societies change as a consequence of the economic development and the globalization, that research replicated the Hofstede's work using the World Value Survey data and different birth cohorts, and showed that cultural differences between countries are stable and change only relatively to the scores of the others. Some examples are that contemporary societies score higher in Individualism and lower on Power distance (Beugelsdijk, Maseland & Van Hoorn, 2015).
studied as the other four dimensions (Hofstede, 2001). In fact, the dataset for the dimensions is different; while the first four dimensions data are taken from the surveys made at IBM, the scores of the other two recent dimensions are estimated from data of the WVS (Beugelsdijk, 2018).
Thus, conforming with the previous literature of omitting the last two dimensions, this thesis tests the moderating role of power distance, uncertainty avoidance, individualism, and masculinity on the relationship between Age and CSR performance.
2.4.1. Individualism
The dimension of individualism/collectivism can be described as the preference between a social framework where “individuals are expected to take care of only themselves”, or another where “individuals can expect their relatives or members of a particular ingroup to look after them in exchange for unquestioning loyalty” (Hofstede, 1980, p. 45). People from more collectivistic societies are considered to be more eager to prioritize the goals of the collectivity and they are more concerned about public and shared goods like the environment and the planet (Sharma, 2010) because their principal aim is to protect those goods in order to allow their group to enjoy it (McCarty & Shrum, 1994). Several articles of research on CSR also predicted a negative relation between CSR and individualistic cultures, since being more sensitive and caring about the impact that business has on societies and to the stakeholders' interests is typical of collectivistic cultures (Ho, Wang & Vitell, 2012). Additionally, some researchers argued that the stakeholders' pressures are greater in more collectivist societies and that those strains usually affect CSR practices (Gallen & Peraita, 2018). Thus, CEOs that come from a more collectivistic country are expected to have a positive attitude towards responsible behaviors. Hence, the following has been hypothesized:
Hypothesis 2A: the effect of age of CEO on firm CSR performance is weaker when their national culture scores high in individualism
2.4.2. Masculinity
The concept of a masculine society refers to “the extent to which the dominant values in society are
‘masculine’- that is, assertiveness, the acquisition of money and things, and not caring for others, the quality of life, or people” (Hofstede, 1980, p. 46). This dimension is generally considered in
people from masculine societies are supposed to care more for their personal achievements and the improvement of their own lives, people from feminine societies are more interested in the welfare of the whole community. Thus, if we shift the point of view to the CSR practices, it seems natural to think that CEOs that come from masculine cultures will be less engaged in those responsible activities that support the community, the employees, the environment and, generally, what is external to their personal interests. Consequently, strong masculine cultures will affect the involvement of the CEOs in CSR activities negatively, leading their firms to poor CSR performance. Therefore, the hypothesis that is generated from this reasoning is:
Hypothesis 2B: the effect of age of CEO on firm CSR performance is weaker when their national culture scores high in masculinity (MAS)
2.4.3. Uncertainty Avoidance
Uncertainty avoidance is described as the propensity of a culture of being tolerant toward the unpredictability (McCornack & Ortiz, 2019). Some cultures are more comfortable than others in dealing with uncertainty and ambiguity, and this is reflected in the ability of an individual to face unstructured situations. Thus, cultures with high levels of uncertainty avoidance try to minimize unknown situations introducing strict laws, rules or safety measures (Hofstede, 1984).
Shreds of evidence show that individuals from low uncertainty avoiding cultures are more prone to take risky decisions (Hofstede 1984), and one example of risk is being involved in unethical actions or behaving irresponsibly. Therefore, CEOs from countries with high uncertainty avoidance could perceive the negative consequences of not acting in a responsible way (Vittell, Nwachukwu & Barnes, 1993). Deciding not to adopt CSR could increase the financial risk of a company since generally responsible activities contribute to having a long-term profitability keeping and well managing the relationships with stakeholders (Maggina & Tsaklanganos, 2012). Disli, Ng and Askari (2016) in their research, also supported by the study made by Thanetsunthorn (2015), state that cultures that try to escape uncertainty and unpredictability are more likely to preserve and improve environmental quality. Thus, it is reasonable to think that CEOs that come from countries with high UA would engage more in CSR activities. Consequently, the propose of our last hypothesis is as follows:
2.4.4. Power Distance
Power distance is described by Hofstede as “the extent to which a society accepts the fact that power in institutions and organizations is distributed unequally” (1985, p. 347). Societies with a high-power
distance have formal codes of conduct; in fact, the individuals are more likely to accept an unequal distribution of power between superiors and their subordinates, where the superiors have privileges, the loyalty and the dedication of their dependents (Hofstede, 1983;1984). Consequently, the societies with a strong power distance are usually also characterized by the acceptance of corruption (Park, 2003) and since individuals in those societies accept the lack of equality of the power, they tend to justify and consider legitimate those who have high-level public positions and take advantage of their roles to gain personal benefits through illegal ways such as bribery (Getz & Volkema, 2001).
Consequently, we expect that people from high power distance countries will be more likely to accept irresponsible, unfair or unethical behaviors, and consequently, they are also expected not to be involved in CSR practices. CEOs from those countries are not expected to show interest and be engaged in activities that positively influence CSR performance, but they are expected to care more for their personal reputation and achievements. Therefore, the hypothesis developed for this dimension is the following:
Hypothesis 2D: the effect of age of CEO on firm CSR performance is weaker when their national culture scores high in Power Distance (PDI)
2.5 CONCEPTUAL MODEL
3. METHODOLOGY
This section of the thesis provides a detailed explanation of the methodology used to carry out this research. There will be provided the description of the variables of the conceptual model and the relative measures used to collect the data. Lastly, the sampling method and the statistical analysis will also be provided.
3.1 DATA COLLECTION
In order to test the hypothesis mentioned in the previous section, this research used a quantitative approach. The data collected are secondary data that were public or made available from the University of Groningen through access to databases. The first hypothesis want to test the relationship between the CEOs' age and the firm CSR performance level; data for CSR which represent the dependent variable were obtained from CSRHub database, while CEOs' age data that is the independent variable, were collected via the database BoardEx, and information that was missing was collected through personal CV or CEOs' personal page on LinkedIn or Bloomberg as well as other websites like the official corporates' ones. The variables for the moderator, which is the national culture of the CEOs, are represented by four of Hofstede's dimensions and the score related to each country was obtained through the use of Hofstede Insight. Lastly, the first control variable is the Firm Size, which represents an influencing factor on firms' socially responsible practices for many reasons outlined in the below section of the methodology, and as also demonstrated in other past research where it has been used as a control variable (Manner, 2010). The second control variable is the Industry type that was selected since for many reasons explained in the variables section, it seems to play an active role in determining which industries are more likely to be responsible compared to others. This variable was considered as a dummy one and the firms were categorized in manufacturing or service ones. The sample of firms comes from Fortune Global 500 from 2019, which was selected among other popular indexes because it includes firms from the whole globe and different sectors, enabling this research to be generalizable and not industry or country related.
3.2 SAMPLE
companies from the US. The firms' data were taken from BoardEx and the results of four different areas were combined: the UK, USA, Europe and the Rest of the world.
One remark that could be done is that the Global Fortune 500 Index includes firms that are not only from different countries but also from different industries; thus, having an extensive group of different firms allows the research outcomes to be generalizable and not related to one single industry. Additionally, the sample is based on the worldwide 500 largest firms measured by revenue, that are influential in our research since, even if they represent a minority compared to the total amount of existing multinational corporations, they still represent “ 90% of the world’s stock of foreign direct investment” (Rugman & Verbeke, 2004, p.3 ). According to those authors, those largest firms are also the ones which conduct about half of the total global trade, thus becoming not only very influential but also, on the contrary, much more exposed to criticism, judgment, and pressure on their behaviors, that lead also to greater obligations of transparency (Cowen, Ferreri & Parker, 1987). Moreover, CSR is considered as a costly strategy (Russo & Perrini, 2009) and while SMEs are considered to be financially limited (Lepoutre & Heene, 2006), larger firms have the monetary resources to invest in CSR and thus, as the stakeholders know this, those firms are expected to undertake this strategy. Therefore, considering that the sample is composed of the largest firms, the likelihood that they are not involved in CSR activities should be low.
3.3 VARIABLES AND MEASUREMENT
3.3.1 Dependent variable: CSR
The dependent variable of this research is the CSR performance of the firms. Multiple indexes have been developed to measure CSR and some commonly used ones are the MSCI ESG Index (formerly named KLD) and Thomas Reuters Asset4 ESG database. However, in order to make quantitative research, only a few databases are available that also include firms outside the US in their datasets for CSR performance (Dragomir, 2018). Next to the indexes some third parties companies, provide a service of ranking the firms that score higher in CSR performance according to different variables that are used for the evaluation, one famous example is the Ranking provided by the Reputational Institute which lists the best 100 firms from all over the world each year. Another example is the service provided by CSRHub, that seems to be the more complete, since it merges together 556 data sources (CSRHub, 2019), including for example Asset4 index, the KLD index and numerous Fortune ratings, but giving a ranking to a large number of firms and not only listing the ones which score higher.
For the vast number of firms that are included in their analysis and for the comprehensive methodology utilized to give ratings to each firm, CSRHub was used for this research. This variable is considered as a continuous one, as the score given to each firm is from 1 to 100, where 1 refers to the lowest level of CSR performance and 100 to the maximum one.
The criterion used to assign scores to each company is based on the analysis of four categories for each of them, which are namely: community, employees, environment, governance. The overall score assigned to each firm is created on a weighted average of the four categories' scores.
The community category covers "the company's commitment and effectiveness within the local, national and global community in which it does business" (CSRHub Data Schema, 2019). It includes the analysis of the human rights, supply chain, the quality, safety and sustainability of products and services offered, but also the development of the community and the social impact that the company's production has.
The environment category embraces both the respect for the planet ecosystems and the natural resources used by the firms and also the respect of the regulations about environmental restrictions. The subcategories present in this class regard energy and climate change, resource management and policies and reporting standards and performance, such as the Global Reporting Initiative.
The last category is the Governance one; this one covers the examination of the board members, of their equal opportunities, of their diversity and effectiveness in providing proportionate and fair compensations, in respect of the work of the employees. It also includes leadership ethics, namely the firm's effectiveness in treating all the shareholders equally, and in promoting an ethical decision-making process and an integrated company image. Lastly, this category covers the transparency of the firm's activities and the adherence to reporting standards.
Thus, the consideration of all those categories makes the selection of this index coherent with the definition of CSR considered in this paper and described before, specifically the one proposed by the Commission of the European Communities in 2001 which describe the CSR as "a concept whereby companies integrate social and environmental concerns in their business operations and in their interaction with all their stakeholders on a voluntary basis" (EC- European Commission, GreenPaper3157, 2001, p. 6).
Time Lags
3.3.2 Independent Variable: CEOs’ AGE
The first independent variable is represented by a demographic characteristic of the CEOs, namely their age. That information is taken from BoardEx, the ones that are not available were integrated with the website pages of the CEOs' firms, and other reliable sources like Bloomberg and LinkedIn. The age could be measured in two different ways, as a continuous variable, which means that the age is studied as a number from 1 to 100, or as a category variable, where the ages are grouped together. For this thesis, the age was considered as a continuous variable, where the data are represented by a number from 1 to 100. The distribution of the age of the sample is broad; in fact, it is present at least one observation for each year from 35 to 89, which are specifically the youngest and the oldest observations of the sample.
3.3.3 Moderator: CEOs’ National Culture
3.3.4 Control Variables
Firm Size
Firm Size was selected as a control variable for this research as numerous previous ones already used it, since it is clear that it can be related to the level of responsible performance of firms (Manner, 2010). Prior studies state that there are many reasons why the firm size influences the chance of the companies to be active in social responsibility aspects. The dimension of the firms is considered as the physical and not financial one in this section, namely the size was measured by the employee number of each respective MNE and the data were collected from DataStream. The variable is considered as a continuous one, where the number of employees is expressed in numbers and not in categories. One reason why the size matters is that larger firms are exposed to a greater pressure from the stakeholders, since their actions are under a high level of public attention and exposure to media (Cowen, Ferreri & Parker, 1987), and consequently they could be more likely to engage in responsible activities to show a good company image and build a positive reputation (Hyun et al., 2016). Another reason is the fact that being involved in responsible activities is costly and a large number of resources are needed; for this reason, smaller firms that have no resources to finance those activities will be less likely to implement them (Hafsi & Turgut, 2013). By contrast, larger firms that have more resources to invest, consider CSR activities as a chance to gain future benefits such as stakeholders' satisfaction, customers' respect and greater financial incomes (Jackson & Apostolakou, 2010). Hence, for those explanations, this thesis assumes that firm size may affect CSR performance, thus firm size was selected as a control variable.
Industry Type
Previous research on CSR argue that the industry of the company could influence their attitude toward responsible activities. One reason is that each industry has specific regulations regarding its activities, and that could include limitations on production processes or the salaries and many other aspects (O'Connor & Shumate, 2010). Depending on the industry, the firms could be more or less in direct contact with the consumers, they could use environmental resources or with their activities, they could influence the communities around them, and ,consequently, those activities make some specific industries more likely to be involved in CSR than others (Dabic et al., 2016).
while communications, electric, finance, gas, insurance, real estate, retailing, sanitary services, trade, transportation, and wholesale firms in the service group.
3.4 EMPIRICAL DATA ANALYSIS
The data analysis was realized using the statistical software R-Studio (Version 1.1.463). The approach used was a deductive one. In fact, it has been tried to understand if certain hypotheses based exclusively on the theoretical aspects were also reflected in the empirical data that were observed. The goal of this analysis was, therefore, aimed exclusively at accepting or rejecting the previously theorized hypotheses.
To this purpose, a multiple linear regression based on the ordinary least squares (OLS) method has been carried out. Multiple linear regression analysis is commonly used to explain the relationship between a continuous dependent variable and one or more independent variables. It should be noted that we are dealing with multiple regression since the moderation variables and the control variables are in practice used as independent variables because they are also considered as regressors.
In addition to the multiple linear regression analysis, a further analysis was conducted in order to verify the existence of moderators, namely third variables that are involved in the relationship between two or more variables (in this specific context between the independent variable Age of the CEOs and the continuous dependent variable firms' CSR performance).
In other words, if the intensity of the effect of an independent variable on a dependent variable change as the level of a third variable change, the latter is defined as a moderation variable.
This further analysis has been performed by inserting in the multiple regression models also the effect of the interaction between two variables through the creation of a new variable given by the product of the independent variable and the moderator. The interaction between the two variables is given by the coefficient of the new variable. This coefficient is the term of interaction or rather it indicates how much the effect of independent variable changes with the variation of the other independent variable called a moderator. The interpretation of conditional regression models, namely the ones with interaction, could be complicated depending on the characteristics of the units of measurement of the independent variables. In order to guarantee an adequate interpretation of the data, the decision to operate on standardized variables has been taken, which means that they are mean-centered and with a constant variance equal to 1.
correlation, multicollinearity) had failed, the assumptions underlying the multiple linear regression based on the ordinary least squares method would not have been valid.
Shapiro-Wilk Test: normality of independent variables
Before carrying out the main regression analysis, the normality of the independent variables was tested both through the graphic analysis and through the Shapiro-Wilk Test. The Shapiro-Wilk Test is one of the most powerful tests for the verification of normality, which is tested by comparing two alternative variance estimators, a non-parametric one (at the numerator) and a parametric one, the sample variance, (at the denominator). From the graphic analysis, carried out through histogram and QQ Normal Plot, only the variable Age seems to be able to be distributed normally but to have confirmation of it, it seems appropriate to analyze the results of the Shapiro-Wilk Test. Since for all the variables, the statistic test is very large and the p-value is always lower than 0.05, it can be claimed that no variable is distributed normally. However, the results of this test do not compromise the feasibility of using the multiple linear regression model, as it is not a necessary condition that the independent variables, moderators and control variables are distributed normally. The graphs and tables with the results are shown in APPENDIX B.
Multiple linear regression models
A total of ten multiple linear regression models have been created, four of them with interaction. The results from the regressions are described in detail in the Results section. For all models, in order to assure a correct interpretation of the results, statistical tests were carried out to ensure preliminary requirements underlying the multiple linear regression theory.
These assumptions require the distribution of residuals to be normal, with zero mean and with constant variance (homoscedasticity), to the residuals to be correlated to two by two, and that there is no collinearity between the independent variables. Moreover, it is advisable to verify the presence or absence of outliers among the residuals.
In order to guarantee the normality of the residuals, it was necessary to transform the dependent variable as follows:
CSR Performance = CSR (row data) /100 – CSR (row data)
graph of all the models. In fact, unlike the test performed on independent variables, the Shapiro-Wilk test performed on residuals shows a high p-value (p≥0.05) for all models, so it is not possible to reject the null hypothesis of normality of residuals. Both the QQ plot and the table with the results of the Shapiro-Wilk Test are present in APPENDIX C.
Student t test: normality of residuals
After checking the normality of the residuals, it is necessary to verify that the average of the residuals is not significantly different from zero. For this purpose, the Student t test is carried out on the residuals of all the models created. From the results of this test, it emerges that all the residuals have an average of zero, therefore the hypothesis is accepted. Moreover, it is possible to demonstrate that there are no outliers among the residuals, which means that there are no values particularly distant from those predicted by the model. Indeed, all residuals have values in the range -3, +3.
Breusch-Pagan test: Heteroscedasticity
In order to verify the heteroscedasticity of the errors, on the other hand, the Breusch-Pagan test was conducted (results in APPENDIX D). From the test findings, it is clear that the errors have a constant variability in all models, in fact, the p-value is always greater than 0.05, therefore the hypothesis of homoscedasticity of errors is never rejected.
Durbin-Watson test: Independence of errors
The independence of the errors, however, is tested by the Durbin-Watson test. For all models, the Durbin-Watson coefficient was calculated, obtained from the ratio between the sum of the differences of two adjacent residuals raised to the square and the sum of the squares of the residuals.
Durbin-Watson statistics will always have a value between 0 and 4. A value of 2 means that no correlation between residuals has been found. From the results shown in APPENDIX E, it emerges that the values are always very close to 2 and in any case never go beyond the thresholds of acceptability 1.5-2.5, therefore it can be said that the residuals of all the models created are correlated to two by two.
Multicollinearity of the variables
a high correlation between the independent variables. If this were the case, the collinear variables would not provide additional information and it would be difficult to identify the effect that each of them has on the dependent variable. The coefficients of these variables could vary greatly depending on which regressors are included in the model and this would be an issue when interpreting the data. A method for measuring multicollinearity is based on the variance inflationary factor (VIF). If, on the contrary, the independent variables are highly correlated, the VIF is high and could exceed the 10. Some experts are in a more prudent position and suggest applying different regression methods than the ordinary least squares when the VIF is greater than 5. Concerning this analysis, from the results in APPENDIX F, it can be concluded that there is no evidence of multicollinearity between the independent variables since the values of the VIF are all very close to 1.
4. RESULTS
This chapter presents the results of the analysis. In the first part, the descriptive statistics and the correlations between the variables will be explained; in the second part, the results of the regression will be described in order to test the hypotheses.
4.1 DESCRIPTIVE STATISTICS
First, the attributes of the dependent variable should be analyzed. The CSR variable has a minimum score of 35 and a maximum score of 70 with an average of 56 and a standard deviation of 6.58. Transforming this variable into CSR Performance, as described above, results in a new data range with a significant reduction in variance (mean = 1.32; SD = 0.333).
In the sample analyzed, the CEO with the lowest age is 35 years old, while the oldest is 89 years old. The average age is however 58.6 years with a standard deviation of 7.48. Despite the fact that we have already shown that no variable is distributed as a normal one, among all the independent variables, the age is certainly the one that shows the greatest symmetry. For the description of the moderation variables, we use the following graph in which the box-plots of the moderators are compared.
The box-plot is a graph (Figure 3) obtained from 5 numbers of synthesis (minimum, 1° quartile, median, 3° quartile and maximum) that describes the most important characteristics of the distribution. The box represents the interquartile difference and its size can give a first approximation of the variability of the distribution itself. Moreover, comparing the heights of the whiskers and the lengths of the two rectangles that make up the box we can obtain some information about the skewness of the distribution: this is less skewness when the lengths of the whiskers are similar to each other and the heights of the two rectangles are similar to each other.
From the observation of this graph, it is possible to notice even the presence or not of outliers. In fact, the whiskers have a length equal to 1.5 times the height of the box.
From the graphic analysis of the moderation variables, it is possible to notice that all the moderators are very heterogeneous and have very different distributions among them. The Individualism variable, in indeed, has a distribution that goes from a minimum score of 18 to a maximum of 91, but it is very skewed, in fact, it has an average of 72.89 and a standard deviation of 21.41. It also has several outliers in the lower part of the distribution.
Masculinity, on the other hand, varies from a minimum of 5 to a maximum of 95 (mean = 58.7; SD = 15.64). However, it seems that most of the distribution is concentrated around the central values, in fact, it can be seen that the height of the box is lower than all the others. The size of the range is caused by the presence of outliers both in the upper part and in the lower part of the distribution.
The variables Uncertainty Avoidance and Power Distance are both skewed, in fact, the two rectangles inside the box have completely different heights. The first in particular has an interval ranging from 23 to 100 (mean = 58.87; SD=20.21) and there is no detected presence of any outlier. The second, instead, varies from 13 to 93 (mean= 47.27; SD = 14.91). In this one, it is possible to notice the presence of an outlier in the upper part of the distribution.
Finally, the average size of the sample companies, measured in terms of the number of employees, is 133038 employees with a standard deviation of 172725.7. The variability is very high. The smallest firm in the sample has 500 employees while the largest one has 2300000 employees.
4.2 CORRELATIONS
The correlation matrix is presented in Table 2 on the following pages. In the data set the variables Age, Masculinity and Power Distance are correlated with the dependent variable CSR Performance. The variable Age, in particular, is negatively correlated with the dependent variable (p≤0.1), which is a first positive indication of the validity of the first hypothesis. This aspect will be further investigated later, and the hypothesis will be tested with appropriate linear regression models.
the other hand, is not correlated to the independent variable Age. This information indicates that the Hypothesis 2C is unlikely to be accepted.
The Industry control variable is significantly correlated with Power Distance, Individualism and Uncertainty Avoidance, while the Number of Employees variable is not correlated with either the independent variable or the moderators.
4.3 REGRESSION RESULTS
Table 3 shows the results of the regression models created. The dependent variable, clearly, is CSR Performance for all models.
In model 1, in addition to the dependent variable, only the control variables Number of Employees and Industry are included. Both are not significant: Number of Employees (B=-0.47; p=0.426), and Industry (B=0.012: p=0.922). However, the control variables do not seem to create noise so that we can keep them in the following calculations.
In model 2, the independent variable Age is added separately to the control variables to verify hypothesis 1. Looking at the results we see that the variable Age is negatively correlated with the dependent variable CSR Performance (B=-0.14) at a significant level (p≤0.05).
The F test on the model's significance also seems to give us positive results. In fact, it has a value of 2.22 with a p-value ≤0.01. This allows us to ensure that at least one of the estimated coefficients is significantly different from zero.
It can, therefore, be said that hypothesis 1 can be accepted: the younger the CEO, the higher the CSR Performance. A further confirmation of the validity of hypothesis 1 consists in the fact that also in the following models, in which are introduced, one at a time, the modifiers and the interactions, in addition to the control variables, the coefficient of the independent variable Age is always negative at a significant level (p≤0.05). The only exception is model 5 in which the coefficient is significant at a level of p=0.08233.
Model 3 shows that the moderation variable Individualism rarely has a direct effect on the dependent variable. Adding the terms of interaction in model 4, it becomes obvious that the variable itself not only rarely has an effect (B=-.070; p=.252) but also rarely and insignificantly interacts with the variable independently of previous expectations due to correlations between variables. In fact, the interaction with the variable Age (B=-.001; p=.979) is, in any case, null and not significant. Therefore, the variable Individualism does not seem to moderate the effects of Age on CSR performance and therefore the hypothesis 2A is rejected.
p=.0005), we can notice how the Masculinity variable also interacts with the independent variable Age (B = .104; p=.061), as we expected also from the correlation between the two variables.
It is, therefore, possible to affirm that the intensity of the effect of the independent variable Age on the dependent variable CSR Performance changes with the variation of variable Masculinity values. Therefore, it is possible to define Masculinity as a moderator of the effect of age on CSR performance.
However, it is not possible to accept hypothesis 2C as the coefficient of interaction is positive. In fact, to affirm that the effect of the CEO's Age on CSR Performance is weaker when their national culture gets high scores in terms of masculinity, the coefficient of interaction should have been negative. In fact, even looking at the Simple Slopes graph (Figure 4) the hypothesis must be rejected.
Figure 4: Simple Slopes graph
Model 7 shows that the variable Uncertainty Avoidance (B=-0.085; p=.155) has no direct effect on the dependent variable CSR Performance. Moreover, from model 8, it can be seen that this variable does not interact with the independent variable Age (B=-0.054; p=.325). This result was predictable since the moderation variables Uncertainty Avoidance and Individualism are related to a significant level (p≤0.01). In fact, after rejecting hypothesis 2A, hypothesis 2C must also be rejected.
the terms of interaction in model 10 we notice that the variable itself does not interact with the independent variable Age (B=0.023; p=0.732) despite the results of the matrix of correlations we had expected a possible interaction. Additionally, from model 10 the direct effect of the moderation variable on the dependent variable is not significant (B=-0.86; p=0.1562). As a consequence of what has been said, it can be concluded that the Power Distance variable does not seem to moderate the effects of Age on CSR performance and therefore the hypothesis 2D is rejected.
5. DISCUSSION
The purpose of this research was to find out whether the age-related demographic characteristic of CEOs influenced their disposition towards CSR activities. Analyzing this relationship between the CEOs' characteristics and the CSR performance, this thesis further investigated the role that the nationality of the CEO has in moderating the relationship.
After conducting the analysis, we can finally examine the results and see if they are in line with what had been hypothesized in the theoretical section.
Before discussing the hypothesis, it is important to note that both the control variables found no significance even if firm size and industry generally have significant effects on the CSR of the firms. However, one thing that can be observed is that the CSR performance are higher for the Manufacturing industries. One possible reason could be that this kind of industry is generally considered responsible for "actively contributing to the pollution of air and water, as well as environmental damage and disruption” (Handayani, Wahyudi & Suharnomo, 2017, p.153), and consequently the attention given to avoid scandals and to avoid a bad reputation appears to be higher. CSR, in fact, can be used as a tool to both benefit society and help companies' activities and competitiveness (King & Lenox, 2002).
Regarding the main relationship existing between the CEOs' age and their attitude toward the CSR,
Hypothesis 1 stated that a younger age of the CEOs results in higher CSR scores for the firms they
same time, their analysis confirms that age is "the most important driver of CSR activities” (Fabrizi et al., 2014, p. 324), supporting the credibility of our findings.
Regarding the cultural dimensions, we need to do different reflections, since the first hypothesis was the only one out of five to have been accepted. Apparently, in fact, against what stated in theory and against the expectations, nationality does not seem to have a moderating role, but rather the cultural dimensions that describe each nationality could fit better as other independent or control variables.
In general, although many researchers, for example, Ringov and Zollo (2007) studied and confirmed the effect that different cultures have on the CSR performance of the firms, in this thesis the non-significance of three out of four hypotheses concerning nationality could be explained by the presence of many American CEOs, namely 108 out of 289, that make the sample not very heterogeneous.
Hypothesis 2A accurately predicted that the relationship between age and CSR would be weakened
if the CEO is from a culture that scores high in individualism. Since the result of the regression say that the variable itself does not have any effect on the relationship (B=-.070; p=.252), the hypothesis at-issue was rejected. Consequently, the theoretical predictions made in section 2.4.1 must be considered invalid since they are statically non-significant. Similar non-significant results regarding Individualism and CSR performance are also present in Thanetsunthorn (2014), who gave as a valid explanation the fact that the composition of the sample can become a critical factor when the number of observations is not a very extensive amount. Additionally, a possible explanation for this result is that this dimension comprise an opposing effect in which individualist CEOs could use responsible actions as a means to achieve economic and personal rewards and could implement CSR strategies through their influential role in the firm, while collectivistic CEOs, although more caring about the public welfare, would like to have a greater collective participation that could result in making the process of implementation of CSR strategies, slow and ineffective.
Hypothesis 2C regarding Uncertainty Avoidance was assuming how a high score of this dimension
However, even if in this study the outcomes are not significant, our theoretical assumptions were based on most of the existing research, whose findings argue that certainty avoiding countries implement responsible practices to avoid and prevent negative reputational consequences (Rallapalli et al.,1994). In this regard, a possible explanation for the non- significance is given by Katz, Swanson and Nelson (1999) who explain an internal contradiction of this dimension, in fact, consumer and community activism occur mainly in cultures with lower Uncertainty Avoidance; thus, cultures with an high level of UA with the absence of activism could have less pressure to implement CSR strategies.
Hypothesis 2D assumed that a high level of Power Distance results in a weakening of the effect of
the age on firm CSR performance. The results of the regression show that the dimension does not interact with the independent variable Age (B=0.023; p=0.732) and moderate on the dependent variable in a non-significant way (B=-0.86; p=0.1562). As a consequence, the hypothesis was rejected. Regarding this non-significant result, it seems essential to mention the fact that the power distance dimension is considered to have a "controversial nature”, since it resulted in inconsistent outcomes in many other studies (Miska et al., 2018). An argument that could explain this nature is given by Waldman et al. (2006), who explain that managers in countries with high power distance are not sensitive to social issues, but they could anyway feel obliged to take care of "greater social purposes" for the role they cover.
Hypothesis 2B instead assumed that the relationship between age and CSR performance would be
weakened if the CEO came from a culture with a high level of masculinity. The result of the regression for masculinity is the only statistically significant one among all the dimensions. In fact, while the other dimensions result in not moderating at all the relationship, masculinity has a direct influence on the dependent variable (B =-0.206; p=.0005) and also with the independent variable Age (B = .104; p=.061), and therefore, it is possible to define masculinity as a moderator of the effect of age on CSR performance. However, while the hypothesis suggested that this dimension should have weakened the relationship, masculinity strengthens it. Thus, the hypothesis was rejected since the coefficient of interaction is positive instead of being negative. Although a positive moderation is observable, interestingly, these findings go in contrast with the extensively documented previous literature where many researchers (i.e., Blodgett et al. 2001; Ringov & Zollo 2007; Peng et al. 2012; Thanetsunthorn, 2015) showed how masculinity interact negatively with CSR.
