What is the Effect of Firms’ home National Culture on its Corporate Social Performance? And How does the Level of Industry Competition moderates this relationship?

(1)

1

What is the Effect of Firms’ home National Culture on its

Corporate Social Performance? And How does the Level of

Industry Competition moderates this relationship?

By

Ignasi Sol – S4222148

MSc. International Business & Management

Faculty of Economics & Business

University of Groningen

(2)

2

ABSTRACT

Built on institutional theory this study aims to examine the effect of national culture on firms' sustainability performance and the moderating effect of the level of industry competition. Drawing on Hofstede’s framework I outline cultural dimensions which may predict CSR engagement: individualism, masculinity, uncertainty avoidance, and power distance. Using a sample of 1152 global companies from 30 nations, I acquired the CSP scores data from Thomson Reuters ASSET4 and tested the hypothesis using OLS. The findings suggest that companies’ home country masculinity and power distance influence positively, and uncertainty avoidance negatively, its CSP scores.

(3)

3

INTRODUCTION

In the past, Corporate Social Responsibility (CSR) was considered an extra cost for shareholders, and the only business obligation was to maximize profit (Friedman, 1970). However, in today’s world, enterprises are expected to take responsibility for social and environmental issues (Pedersen & Andersen, 2006).

Currently, with the increase of globalization, worldwide more companies are expanding to foreign countries and, with this expansion, their corporate responsibility is surpassing the borders as well (Ioannou and Serafeim, 2012, Roy and Goll, 2014, Gallego-Álvarez and Ortas, 2017). This phenomenon is increasing the pressure that stakeholders exert on companies for them to improve their performance in social and environmental issues (Ioannou and Serafeim, 2012). This pressure is especially strong since most multinational firms from developed countries have extended their global supply chain to emerging countries. These countries have the necessity of foreign capital investment, so their institutions and regulations tolerate poor environmental and social performance. (Husted, 2005). Hence, given the huge impact that multinational enterprises have on sustainability, corporate social responsibility must be a priority in their agenda. However, as the United Nations Global Compact —Accenture Strategy CEO Study on Sustainability (2019) pointed out, only 21% feel that their business is currently playing a critical role in contributing to the Global Sustainability Goals.

(5)

5 they influence the social and political development of the stakeholders’ interest. This has been confirmed by a small number of existing studies which have demonstrated consistent differences in the implementation and understanding of CSR across countries (i.e., Husted 2015, Jackson and Apostolakou, 2010, Ho et al., 2012, Roy and Goll, 2014, Cai et al., 2016, Gallego-Álvarez and Ortas, 2017, Miska et al., 2018). For instance, Maignan (2001) studied consumers from Germany, France, and the United States and found that depending on the different national cultural, consumers’ support of socially responsible organizations differed.

Institutions are divided into formal or informal. Most of the extant researches that have analysed antecedents of CSR have emphasized the effects of formal institutions, such as law, on CSR (i.e., Campbell, 2007; Chih et al., 2010; Moon, 2005) and have paid little attention to informal institutions such as culture (i.e., Maignan, 2001; Ringov & Zollo, 2007; Waldman et al., 2006, Ho et. al, 2011, Ioannou & Serafeim, 2012, Peng et al. 2014, Miska et. al. 2018, Armien et al. 2018). Various studies have identified culture as an essential explanatory factor (i.e., Waldman et al., 2006, Ringov & Zollo, 2007, Beekun, Hamdy, Westerman, & Hassab Elnaby, 2008; Haxhi & van Ees, 2010; Parboteeah, Addae, & Cullen, 2012). For example, Ioannou and Serafeim found that cultural characteristics play an important role in the explanation of CSR differences across companies (2012).

According to Miska et al., empirical discoveries in this field have revealed significantly mixed-effects (2018). For example, power distance Hofstede’s dimension was discovered to have negative (i.e., Ringov & Zollo, 2007) and positive (i.e., Ho et al., 2012, Ioannou & Serafeim, 2012) impact on sustainability. Moreover, inconsistencies have been found regarding the effect of masculinity and uncertainty avoidance Hofstede’s dimension on sustainability (i.e., Arnold, Bernardi, Neidermeyer, & Schmee, 2007, Scholtens & Dam,2007, Haxhi & van Ees, 2010, Cai et al., 2016).

(6)

6 interconnections between formal and informal institutions, can have several types of impacts on sustainability (Miska, et. al. 2018).

In addition, the industry-level circumstances in which companies engage in CSR, to a lesser or larger extent, have been overlooked (Aguilera, Rupp, Williams, & Ganapathi, 2007; Campbell, 2007). Our understanding of the environment for CSR will continue to be limited without looking at the role of competition (Campbell, 2007; Flammer, 2012). To the best of my knowledge, despite the fact that Ioannou and Serafeim (2012) pointed out that the 10% total variance in companys' CSR engagement is explained by industry characteristics, there has not been any research which has studied the moderating influence of the level of industry competition on the relationship between national culture and sustainability.

Hence, considering the existing discrepancies regarding culture as a predictor of CSP and the lack of research analysing the moderation role of the industry level of competition, this paper aims to address the gap of how firms’ home national culture affects its corporate social performance and how this relationship is moderated by the level of industry competition. In order to fill up this gap, my study is based on the institutional theory (e.i., Campbell, 2007; Campbell et al., 1991; North, 1990) and followed the Ioannou and Serafeim study (2012) with the objective of examining the CSP scores heterogeneities across companies and countries in accordance with the national-level institutions' variation.

This study contributes to the current emerging literature on CSR from two different perspectives: First, it brings light into the existing discrepancies found within the literature researching the influence of national culture on CSP and provides a plausible explanation for the observed CSP heterogeneities across firms and nations. Second, addressing the gap of how this influence is moderated by the level of industry competition I provide new useful information about how the joint effect of national culture and the level of industry competition affect firms’ CSP, which may provide additional insights to better understand the CSP heterogeneities across industries and countries.

(7)

7 Therefore, my research further investigates the relationship between national culture and CSP by posting the following question:

-What are the effects of firms’ home national culture on their Corporate Social Performance (CSP) variation across different countries and how is the relationship moderated by the level of industry competition?

(8)

8

LITERATURE REVIEW AND HYPOTHESES DEVELOPMENT

Definition of CSP

Generally, scholars failed to scientifically integrate the conceptual developments of CSP (Wood, 1991). The first study attempting to combine multiple conflicting perspectives on CSP was conducted by Caroll (1979). As stated by Caroll the social responsibility of a company “embodies the economic, legal, ethical and discretionary expectations that society has of organizations at a given point in time” (1979). Despite the diverging views in the literature on the methodology and definition of CSP, Wood (1991) study contains the most fully integrated and generally accepted definition. Therefore, this definition will be used in the present paper: “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”. Additionally, most scholars agree that CSP is a key element to the proper functioning of a company. Therefore, firms need to understand how to address social and sustainability issues. Notwithstanding, scholars were unsuccessful to recognize precisely which social and sustainability issues come within the bounds of CSP (Ho et al., 2011). In this study, the paper of Ioannou and Serafeim (2012) was followed and the matters relating to the environmental, social, and corporate governance were considered.

(9)

9 Institutional theory and CSP

Given that stakeholder identities and interests change from country to country, and considering that the objective of this thesis is capturing the variations of CSP heterogeneities across countries, the institutional theory is a useful instrument in order to understand these variations (Matten & Moon, 2008). The institutional theory assumes that businesses are embedded in formal and informal institutions (North, 1990). Formal institutions are rules and organized structures, such as laws and regulations (North 1990). Informal institutions are related to socially common rules which are outside of the officially sanctioned channel. (North, 1997). Across the international business and management literature, culture is often considered an informal institutional element. (Peng et al., 2008). Therefore, to precisely describe the effect of national culture on firms' sustainability performance my research is based upon the institutional theory. Hence, I followed the definition of Huntington (1969): “institutions are stable, valued, recurring patterns of behaviour,” specified by their “adaptability, complexity, autonomy, and coherence” (1969: 12).

Importantly, as recognized by Whitley (1999), institutions diverge across countries and they jointly constitute different “types” of national business systems (NBS) (Ioannou & Serafeim, 2012). NBS is defined by Whitley as “distinctive patterns of economic organization that vary in their degree and mode of authoritative coordination of economic activities, and in the organization of, and interconnections between, owners, managers, experts, and other employees” (1999:33). My study draws on Whitley’s NBS framework, as Matten & Moon (2008) and Ioannou & Serafeim (2012) did. The reason behind this election is that this framework focuses on stable and long-established characteristics of the business system, distinctly encompasses the main institutional categories that influence CSP, and, additionally, includes the key role of the culture system (Matten & Moon, 2008, Ioannou & Serafeim, 2012).

The literature on the relationship between institutions and CSP

(10)

10 and Neville culture is the condition, shaping the implementation of sustainability practices (2012). Similarly, Ioannou and Serafeim pointed out that cultural characteristics are an important factor in explaining differences in corporate social performance across firms (2012). Moreover, numerous studies have observed a link between corporate social responsibility and culture across countries (i.e.Szőcs, Schlegelmilch, Rusch, & Shamma, 2016, Williams & Zinkin, 2008, Gallego-Álvarez & Ortas, 2017, Peng et al. 2014).

However, despite the several studies emphasizing the influence of culture on corporate social responsibility aspects, little consistency has been seen across these studies. For example, some researchers observed a positive effect of Hofstede’s power distance on sustainability (Ho et al., 2012, Ioannou & Serafeim, 2012) whereas others have a negative influence (Ringov & Zollo, 2007). Or the power distance Hofstede dimension which has been found to have positive (Ho et al., 2012, Williams & Zinkin, 2008) and negative (Gallego-Álvarez & Ortas, 2017, Husted, 2005, Perk et al., 2007, Peng et al., 2014, Ringov & Zollo, 2007) effects on firms sustainability performance.

One of the reasons for these inconsistent findings is that different methods for measuring culture have been applied: some have relied on Hofstede’s national culture framework (Ringov & Zollo, 2007, Ho et al., 2011, Ioannou & Serafeim, 2012) whereas others on Global Leadership and Organizational Behaviour Effectiveness (GLOBE) (Waldman et al., 2006, Milsk, 2018, Witt & Stahl, 2016, Waldman et al., 2006, Parboteeah et al., 2012). Despite the similarities both models share, the different research designs with which they work may cause different results (Shi & Wang, 2011). First, GLOBE project distinct between cultural values and cultural practices and found a negative correlation between them. Contrarily, in Hofstede’s framework, cultural values and practices are positively related (Javidan, House, Dorfman, Hanges & Luque, 2006). Second, both studies diverge substantially in the way they analyse and aggregate the scales to the societal level (Peterson & Castro, 2006).

(11)

(12)

12

HYPOTHESES

Individualism

Hofstede defines individualism-collectivism cultural dimension as “the degree to which individuals are integrated into groups” (1994, p. 2). In higher individualistic cultures ties between individuals are loose, therefore, everybody looks after her/himself and her/his family. Individuals within individualistic societies tend to assign greater importance to personal interest and achievement rather than collective accomplishment (Ho et al., 2011). On the other side, people from collectivist cultures tend to be integrated into solid, cohesive groups and they place a higher value on the group welfare rather than on their own self (Hofstede, 1994). Consequently, in these cultures, decisions are made more consultatively and participatory (Ioannou & Serafeim, 2012) and the importance is given to future concern collectivist interest (Waldman et al., 2006). Hence, decision-makers individuals within these societies have a tendency to implement strategies that benefit the collective. In this sense, activities with the objective of protecting the sustainability of the environment and behaving in a socially responsible manner are more logical with collectivist values (Parboteeah et al., 2012). Furthermore, Akaah found that companies based in individualistic cultures have lower levels of ethics (1990), which may probably result in a negative influence on their CSP scores. Therefore, a negative correlation between companies based in high individualistic countries and Corporate Social Performance is predicted.

H1: Companies based in higher individualistic countries will have a lower CSP score.

Masculinity

(13)

13 Hofstede contends that a strong emphasis on personal gains, competition, and economic growth hinders any display of sustainability behaviour (2001). In line with this view, Cox et al. (2011) acknowledge that the pursuit of economic growth, which is the main characteristic of masculine societies, may cause enterprises to be hesitant to assume the additional costs deriving CSR-related activities. Hence, these companies would have a tendency to lower their CSR investments, resulting in lower CSP scores. Therefore, a negative correlation between companies based on high masculinity countries and their Corporate Social Performance scores are expected.

H2: Companies based in higher masculine countries will have a lower CSP score.

Uncertainty avoidance

According to Hofstede uncertainty avoidance is defined as follows: “the extent to which members of a culture feel either uncomfortable or comfortable by uncertainty or unknown situations” (1994, p. 5). Central to this idea is cultural awareness about the future, describing the degree of risk acceptance and ambiguity by a society (Husted, 2005). Societies that score low on uncertainty avoidance inspire differences of opinions, show an inclination for a few regulations and loose laws, therefore, being more prone to take risk (Hofstede, 1984). On the other end, high uncertainty avoidance societies desire organized environments, well-defined hierarchies, laws, and social rules, which enable them to reduce uncertainty. The relationship of companies with their stakeholders may be more sustainable with higher CSR investment. For this reason, CSR engagement may constitute an effective corporate strategy to decrease the level of uncertainty. Hence, companies based in higher uncertainty avoidance countries are expected to have higher CSR investments, which would increase their CSP scores.

H3: Companies based in greater uncertainty avoidance countries will have a higher CSP score.

Power distance

(14)

14 power in decision-making, hence, facing a stronger normative restriction (Ioannou & Serafeim, 2012). On the other hand, managers in countries with high power distance managers are perceived as “well-meaning autocrats” that offer favours to their subordinates in exchange for their loyalty (Hofstede, 2001). These managers put in front of their individual interests and are not interested in establishing meaningful relationships with key stakeholders (Ioannou & Serafeim, 2012). Therefore, as Husted pointed out, in these societies, there are fewer initiatives to debate and to question the ruling (2005). For this reason, people in high power distance societies tend to accept unsustainable business behaviour easier, at the same time as managers that support these conducts are held to account less than they would in a low power distance cultures (Cohen, Pant & Sharp, 1996). Hence, it is expected a negative correlation between companies based on greater power distance countries and its Corporate Social Performance scores.

H4: Companies based in greater power distance countries will have a lower CSP score.

Level of industry competition

Most findings up-to-date recognized that industry characteristics substantially influence CSR activities. For instance, Hendry (2006) contended that activists choose the industry before targeting a particular company and that changes in CSR by leading enterprises are expected to spread throughout the industry with other companies imitating such behaviour. Similarly, O’Connor and Shumate (2010) found that companies within the same industry tend to communicate CSR in the same way in order to influence similar stakeholders. Therefore, as seen in CSP ranking, firms’ sustainability scores vary across different types of Industries.

(15)

15 Prior strategic CSR literature has studied industry competition and CSR in various tangential ways. For my study, the most relevant way is the one which interest lies in understanding the impact of industry competitive forces on companies’ CSR practices or scores (i.e., Siegel & Vitaliano, 2007, Campbell, 2007, Flammer, 2012, Hawn 2014, Sheikh, 2018).

Within this strategic CSR literature, various studies have found a positive relationship between the level of industry competition and CSP scores. For example, within the strategic view of CSR, Fernandez-Kranz and Santalo (2010) and Declerck and M’Zali (2012), reported that firms in more competitive environments have higher social ratings due to the competitive advantage that CSR represents in this environments. Similarly, Flammer et al (2012) presented a big tariff reduction as an experiment for rises in competition, resulting in a CSP increase after the increase in competition, providing empirical support to the argument that to remain competitive, companies need to distribute their resources as efficiently as possible in all its areas. Therefore, considering these findings, it is expected that the level of industry competition would moderate the impact that each of Hofstede’s national culture dimensions (individualism, power distance, uncertainty avoidance, and masculinity) has on the CSP scores. However, given the different nature of each national culture dimension, it is expected a different moderation effect on each Hofstede’s national culture dimension. Hence, I argue as follows:

Individualism:

(16)

16 Nevertheless, a low level of competition implies little to no choice for suppliers and customers within the market(Campbell, 2007, Flammer 2012, Sheikh, 2018). Then, corporate reputation and brand loyalty do not determine sales, profitability, or survival (Campbell, 2007, Flammer 2012). Hence, since individuals within individualistic societies tend to assign greater importance to personal interest and achievement rather than collective accomplishment (Ho et al., 2011), firms based on individualistic countries have fewer incentives to engage in CSR, which may lead to low CSP scores. However, in collectivist societies, since individuals place a higher value on the group welfare rather than on their own self, companies in low competitive industries may tend to act more responsibly in order to increase the common welfare. Hence, I suggest:

H1a: The level of industry competition moderates the negative relationship between home country individualism and CSP in the following way:

- The relationship between home country individualism and CSP is weaker for companies within industries with a high level of competition.

- The relationship between home country individualism and CSP is stronger for companies within industries with low competition

Masculinity

(17)

17 working together leads to an inclination for cooperative behaviour, still in industries with a high level of competition. Nonetheless, provided that members of the alliance networks are benefited by supplying them with potential sources of competitive advantage in CSR, competitors' actions to neutralize these potential advantages are expected. This, in turn, may lead to an increase in CSR efforts of the whole industry, resulting in higher CSP scores.

On the other hand, within industries with a low level of competition, the little to no choice for suppliers and customers within the market makes that sales are not determined by reputation and brand loyalty (Campbell 2007, Flammer, 2012, Sheikh, 2018). Masculine societies are characterized as being focused on success in terms of material rewards rather than in welfare terms (Hofstede’s, 2001). Moreover, as Steensma, Marino, Weaver, & Dickson pointed out, in masculine societies the general well-being is not considered a priority since there is a tendency to see the world from a winner-loser perspective (2000). Therefore, given the lack of consideration for the general welfare and the fact that in a low level of competition industries profitability is not determined by brand reputation, companies from masculine countries may see CSR as an unnecessary cost which keep them away from the goal of maximum profitability. Contrarily, in a feminine society, there is a tendency to care for the weak and try to increase the general quality of life (Hofstede, 1994). Thus, even if the reputation does not influence profitability, they may tend to invest in CSR activities in order to keep a certain level of the general welfare. Hence, I suggest:

H2a: The level of industry competition moderates the negative relationship between home country masculinity and CSP in the following way:

- The relationship between home country masculinity and CSP is weaker for companies within industries with a high level of competition.

- The relationship between home country masculinity and CSP is stronger for companies within industries with low level of competition.

Uncertainty avoidance:

(18)

18 approach to doing so will differ (Flammer, 2012). On one hand, due to the higher competitive risk in low concentrated industries and the intolerance of high uncertainty avoidance societies (Millán, 2015), it is expected that in highly competitive environments, companies based in high uncertainty avoidance countries will attempt to maximize their profit by established models that have proofed reliability and follow the rules on the market. Therefore, it is expected that companies based in high uncertainty avoidance countries will increase their CSR efforts in low concentrated industries because of the fact that it is the general rule followed by companies within the industry and it is the action which involves less risk. Nevertheless, when these companies would be in a low competitive environment they may have a tendency not to act responsibly due to the low risk that the brand reputation has for financial performance.

On the other hand, low uncertainty avoidance societies inspire differences of opinions, show an inclination for few regulations, and are more prone to take risk (Hofstede, 1984). This mindset would lead them not to follow established models that have proofed reliability and try to find different business strategies on the market (Millán, 2015, Sheikh, 2018). This is why it is expected that companies based in low uncertainty avoidance countries would not increase their CSR investment when been present in low concentrated industries. Similarly, in these countries there is a tendency to fewer regulations and loose laws (Hofstede, 1994). For this reason, and given that in highly concentrated industries CSR efforts are more driven by institutional forces than by industry forces (Hawn, 2012), companies based on lower uncertainty avoidance are expected to have lesser CSP scores when facing low levels of industry competition. Therefore, I suggest:

H3a: The level of industry competition moderates the positive relationship between home country uncertainty avoidance and CSP in the following way:

- The relationship between home country uncertainty avoidance and CSP is stronger for companies within industries with a high level of competition.

- The relationship between home country uncertainty avoidance and CSP is weaker for companies within industries with low level of competition.

(19)

19 As commented above, in high levels of industry competition, CSR is not perceived as a cost but rather a resource that allows companies to i.e., attract new customers, differentiate themselves, improve their productive efficiency, and ultimately enhance their competitiveness (Cambpell 2007, Flammer, 2012, Hawn, 2012). Provided that managers operating in high power distance countries tend to put in front their individual interests (Ioannou & Serafeim, 2012), they may tend to increase the investment in CSR as valuable business strategy in order to increase its competitive advantage. For the same reason of investing in CSR as a valuable business strategy, companies based on low power distance countries may also have a tendency to enhance their CSR effort in industries with high level of competition.

Notwithstanding, as mentioned above, in industries with low level of competition corporate reputation and brand loyalty do not determine sales, profitability, or survival. Hence, managers operating in high power distance countries, which put in front their individual interests (Ioannou & Serafeim, 2012), may be inclined to reduce CSR investments in order to increase profitability. Since they are not interested in establishing meaningful relationships with key stakeholders (Brammer & Millington, 2008) and sales is not determined by corporate reputation, they do not have any incentive for expending in sustainability related activities. Therefore, CSP scores of companies based on high power distance countries may be lower in high concentrated industries. However, in low power distance countries, managers have less power in decision-making, hence, facing a stronger normative restriction (Ioannou & Serafeim, 2012). Thus, in these countries, the fear of government action (Baron, 2001), or high market power visibility, and public exposure (Easley & Lenox, 2006), may lead to an increase of social responsibility activities in highly concentrated industries. This in turn, would increase CSR investments of the whole industry, resulting in higher CSP scores of the companies within the industry. Hence, I suggest:

H4a: The level of industry competition moderates the negative relationship between home country power distance and CSP in the following way:

- The relationship between home country power distance and CSP is weaker for companies within industries with a high level of competition.

(20)

20

METHODOLOGY

The purpose of this study is to evaluate how companies’ home national culture effect its CSP scores and assess how the level of industry competition moderates this expected impact. In order to test the hypothesis, a series of variables were gathered from multiple sources. The chosen variables and the reason for selecting them are explained in the next paragraph and in Appendix 1 you can find a table summarising this section.

Variables:

Dependent variables: CSP

As Ioannou and Serafeim pointed out, the multi-domain nature of the theoretical constructs is the main challenge of corporate social performance measures (2012). For this reason, it is necessary a multidimensional measure applied across a large range of industries and a bigger sample of organizations (Ioannou & Serafim, 2012). Considering this requirement, I used the governance, social, and environmental metrics from the ASSET4 database. ASSET4 provides objective, relevant, and systematic environmental, social, governance, and economic data that is based on more than 750 crucial performance indicators (KPIs) per each firm and more than 250 individual data points together with their original data sources (Ioannou & Serafeim, 2012).

(21)

21 Independent variables: National culture dimensions

The independent variables are four of Hofstede’s dimensions of national culture, Power distance, Masculinity, Uncertainty Avoidance, and Individualism. These variables are continuous variables computed using Hofstede’s indices (1985). I decided to use Hofstede’s indices in my study instead of other used frameworks in the literature such as GLOBE for the two following reasons: First, throughout studies relaying on Hofstede’s indices, little coherence and uniformity can be observed (Ringov & Zollo, 2007, Waldman et al., 2006, Ho et al., 2011, Ioannou & Serafeim, 2012). Second, this framework makes it possible for a better comparison with earlier literature on the field since it has been more widely used over time than the GLOBE framework.

Moderator variable: Level of industry competition

The Herfindahl index of concentration within industry, HHI, was used to analyse the moderative effect of the level of industry competition. Ranging from 0 to 1, the higher the index, the less competitive the industry. The Herfindahl index was calculated in formula 1.

Where 𝑠𝑛 is the market share percentage of firm n. The HHI was calculated for each ICB sector.

In Appendix 3 it can be seen the HHI per sector.

Formula 1: HHI

=s

n

+s

n

+s

n

+…s

n..

Control variables:

I additionally included a series of control variables which are commonly used in the assessment of sustainability. The control variables are at the firm, industry, and country level.

(22)

22 performance (Ioannou & Serafeim, 2012, Arminen et al., 2018). Finally, by leverage, measured by the percentage of the total debt over the total assets. This variable will be utilized as a proxy for the riskiness behaviour of a company (Waddock Graves, 1997: Ho et al., 2012; Arminen et al., 2018). Arminen et al. (2018) pointed out that leverage has a negative impact on CSP, in line with the results from previous studies (Waddock Graves, 1997, Ho et al., 2012). Particularly, Campbell (2007) and Ioannou & Serafeim (2012) indicated that CSP is expected to decrease when company risk is especially high.

At the industry level, I controlled by the type of industry a company belongs to, using a dummy variable. The companies were classified by industries according to the Industry Classification Benchmark (ICB) taxonomy. The ICB is an industry classification taxonomy which the classification system allocates companies into industries, supersectors, sectors, and subsectors within the macroeconomy. For the industry level control variables only the industry category, which consists of 10 types of industries, was taken into account. As demonstrated by Gray et al. (1995), depending upon the product industry, firms are more willing to engage in CSP activities (Gallego-Alvarez & Ortas, 2017).

(23)

(24)

24

DATA

In order to test the hypothesis and answer the research question, a quantitative analysis was conducted. Due to the time constraints of the cultural Hofstede’s dimensions, the nature of this study was cross-sectional, using data from 2018. Because national culture dimensions and CSP are both publicly accessible and qualified measures, a secondary analysis of data using electronic databases was done. First, a sample of 1200 companies was chosen from the S&P Global 1200 index, which is a stock market index of global equities from Standard & Poor’s. It is composed of seven regional indices, from the United States, Asia, Australia, Europe, Latin America, Japan, and Canada. It covers 31 countries and approximately 70% of the global stock market, including companies in all ICB sectors and industries. Provided that it is based upon regional indices which include the largest companies listed in each respective stock market, it enabled me to have a sample of comparable firms and, therefore, a better and more valid analysis. Second, company-level data was retrieved using Eikon-x, which is a financial economic data source that combines data from several important sources. Specifically, data on CSP indices were obtained from ASSET4 database, while industry and accounting information was derived from Thomson Reuters DataStream. In addition, sector-level data on the level of industry competition was calculated with the Herfindahl index of concentration within the sector (HHI). Finally, the data of the four cultural dimensions was retrieved from Hofstede’s website.

Sample characteristics

The dataset utilized for the current research covers a total of 1.188 firms from 30 nations. As it can be observed in Table 1, the most common countries in the sample are the United States of America (41,23%), followed by Japan (12,50%) and the United Kingdom (7,03%).

(25)

25

Table 1: Country sample distribution

Country Freq. Percent Cum.

(26)

26

Table 2: Industry sample distribution

Industry Freq. Percent Cum.

Basic Materials 81 6.82 6.82 Consumer Discretionary 194 16.33 23.15 Consumer Staples 83 6.99 30.13 Energy 58 4.88 35.02 Financials 184 15.49 50.51 Health Care 107 9.01 59.51 Industrials 214 18.01 77.53 Real Estate 59 4.97 82.49 Technology 101 8.50 90.99 Telecommunications 40 3.37 94.36 Utilities 67 5.64 100.00 Total 1188 100.00 Descriptive Statistics

The descriptive statistics are shown in Tables 3, 4, and 5. The first table shows the average scores, maximum, minimum, and standard deviation for all the variables gathered in the sample, while the last two display the mean scores, and standard deviation of the dependent variable (CSP score).

In general, table 3 indicates that, within this sample, the CSP scores range from a minimum of 0,68 to a maximum of 97,77 on a scale of 0-100, with a mean of 62,24 and a SD of 21,27. Table 4 shows the CSP mean and standard deviation per country. It can be observed that companies in Portugal, Norway, and France present high CSP scores (Means > 79), while companies in New Zealand, China, Belgium, and Singapore present the lowest scores (Mean < 56). The significant difference in CSP index scores across nations (Table 4) shows that the level of economic development may have an impact on the companies’ CSP scores. In fact, CSP is comparatively higher in developed countries than in developing ones.

(27)

27

Table 3: Descriptive Statistics

Variable Obs Mean Std. Dev. Min Max

GDPpercapita 1188 51433.767 16178.284 572.427 82818.108

PopulationGrowth 1188 .524 .438 -.203 1.536

GovernmentEffectiv~s 1188 1.522 .38 -.447 2.231

PoliticalStability 1188 .589 .391 -.787 1.54

HHI 1188 .311 .18 .052 .987

TotalAssets 1188 2.922e+09 2.396e+10 197914 4.368e+11

ROA 1188 6.774 7.786 -61.78 128.42 Leverage 1188 26.895 21.357 0 378.15 PowerDistance 1188 44.12 11.65 11 81 Individualsim 1188 74.583 21.262 13 91 Masculinity 1188 61.066 19.111 5 95 UncertaintyAvoidance 1188 56.381 20.042 8 104 CSPScore 1188 62.247 21.272 .68 97.415

Table 4: Summary of CSP Score per Country

Country Mean Std. Dev.

(28)

28

Table 5: Summary of CSP Score per Industry

Industry Mean Std. Dev.

(29)

29

RESULTS

Correlation

In order to summarize and see patterns within the data a correlation matrix was runed (Appendix 4). The Correlation matrix is a table displaying the correlation coefficients among the variables. As it can be seen, there is no substantial correlation between the dependent variable (CSP Score) and the independent variables (Hofstede’s cultural dimensions). Furthermore, the independent variables (CSP Score) is not significantly correlated with any variable, being the most positively correlated Individualism (0,212) and the most negatively correlated Uncertainty Avoidance (-0,231).

The correlation matrix did not demonstrate significant and conclusive results and the correlation coefficients do not necessarily predict the potential impacts the selected variables will have on the dependent variable (CSP Score). Hence, a regression analysis needs to be conducted.

Results of the Regression Analysis

(30)

30 variables which express values from the year 2017. In addition, dummy variables were created for the industry variables in order to control for their influence on the analysis.

In order to detect and minimize the potential impact of large outliers, a scatter matrix plot with all the variables was performed (Appendix 6). As can be seen, the graphs of the firm-level control variables (Total Assets, ROA, and Leverage) show some potential outliers. Therefore, Cook’s distance and leverage were calculated. As it can be observed in Appendix 6, 36 observations were identified as outliers (Cook’s distance > 4/n or leverage > 0.1)

and were deleted from the sample, due to the anomalous representation of the company’s values. Appendix 6 (scatterplot with (y axes) Cook’s distance and (x-axes) leverage) gives a graphical picture of the outliers identification analysis.

In order to perform a regression analysis, the data needs to show homoscedasticity, defined as the homogeneity of variance of the residuals. Hence, to verify the presence of heteroscedasticity, a scatter plot with the residuals (y-axes) and the fitted values (x-axes) was performed (Appendix 7). As shown, there is no pattern to the residuals against the fitted values, therefore, suggesting that our estimation would not be affected by heteroscedasticity. Notwithstanding, the Breusch-Pagan test was performed. The results of the Breusch-Pagan test (Appendix 8) did allow us to accept the null hypothesis (Prob > chi2 =

0.1741), therefore, implying that no heteroscedasticity is present within the sample analysis. Additionally, a collinearity analysis was conducted, to check for the presence of multicollinearity. The term collinearity implies that two variables are near perfect linear combinations of one another. It needs to be analysed because in high degrees of multicollinearity, the coefficient estimates become unstable and the standard errors for the coefficients can get inflated. As it can be observed in Appendix 9, the results show low Variance Inflator Factor (VIF) values (VIF < 6 ) for all variables taken into account, which implies that there is no presence of multicollinearity within the sample.

(31)

31 HHI control variables were transformed to the natural logarithm. The first four variables were transformed following this formula: ln(constant+variable), being the constant the minimum value of the variable. The reason behind adding a constant is to prevent the logarithm transformation of the negatives values which would have reduced significantly the number of observations in the analysis due to the generation of missings when transforming the variable.

With the purpose of testing the hypothesis, I used different models with sqrt_CSP_neg as the dependent variable and a gradual introduction of the independent and control variables. The analysis was conducted with the software for statistics STATA. The regression results are presented in the regression table below (Table 6). Column (1) of the table displays the results of the first model, which was run only with the control variables (GDP per capita, const_ln_PopulationGrowth, const_ln_GovernmentEffectiveness, const_ln_PoliticalStability, ln_TotalAssets, const_ln_ROA, Leverage). Column (2) represents the results of the second model which explains the main regression results of our analysis. For this model, the independent variables were included (cultural dimensions: Power Distance, Individualism, Masculinity, and Uncertainty Avoidance). In addition, in column (3) it can be seen the results of the third model in which the moderator variable (ln_HHI) was incorporated. Finally, the fourth model (column 4) was performed adding the interaction between the independent variables (cultural dimensions) and the moderator variable (ln_HHI).

In models 1,2 and 3 the Energy dummy variable was automatically dropped out of the analysis by the software because of collinearity. For the same reason, the software automatically omitted the Utilities dummy variable in the 4rth model. Considering that the main objective of the research is analyse how companies’ home national culture dimensions affect the CSP score, the omission of one dummy variable, included as a control variable, did not impact the final outcomes.

Model 1 displays an R-squared of 0,0477 and an Adjusted R-squared of 0.0334. In this model, some firm-level and country-level control variables showed to be significant. At the country level, GDP per capita and Population Growth resulted to be significant. At the firm level, Total Assets resulted significantly at the **p < 0.01 level.

(32)

32 of all the independent variables resulted to be significant (p-value < 0.05), corroborating the impact of the companies home national culture in its sustainability scores. On one hand, the cultural dimension with the highest impact on the model was Uncertainty Avoidance which resulted to be significant at the *** p < 0.001 level. On the other hand, Individualism resulted not to be significant in the model with a p-value higher than 0.05. Regarding the control variables, all the variables that were significant in model 1 resulted to be significant in model 2, except for GDP per capita.

In model 3 it can be seen that after adding the moderator variable the model didn’t improve its explainability, showing the same R-squared of 0.0938. Furthermore, the moderator variable itself (ln_HHI) resulted not to be significant at all with an extremely high p-value. Due to the insignificance of the variable added, all the other variables remained practically the same as in the second model.

After including the interaction term of the independent variable and the moderator variable, model 4 presents a slightly higher R-squared (0.0962) and adjusted R-squared 0,0753. The tiny increase of the R-squared (0.0024) indicates an interaction effect on the model, although this effect was not very significant. Furthermore, any of the interactions resulted to be non-significant (all of them with a p-value > 0.05). In addition, in this model, Uncertainty Avoidance remained the only significant Hofstede’s cultural dimension (* p < 0.05). These results corroborated the fact that Uncertainty Avoidance was the cultural dimensions with the highest influence in the analysis.

(33)

33 Regarding the hypothesis, only Hypothesis 3, involving Uncertainty Avoidance, resulted to be significant in all the models (p-value always lower than 0.01) and presented negative coefficients. This contradicts Hypothesis 3 and indicates that companies based in greater uncertainty avoidance countries tend to have lower CSP scores. Similarly, Hypothesis 4, regarding Power Distance, was rejected since in models 2 and 3 it showed to be significant at the **p<0.01 level and it presented a positive coefficient, indicating that companies based in greater power distance countries show higher CSP scores. In the same way, Hypothesis 2, involving masculinity, resulted to be significant in model 2 and 3 and showed positive coefficients. This indicates that companies based in higher masculine countries have a tendency to have higher CSP scores. These results undermine the prediction of H2, being this rejected. Nevertheless, Hypothesis 1, considering Individualism, resulted not to be significant in any model, indicating that it cannot be demonstrated that there is any effect of companies’ home country Individualism on its CSP scores. All the independent variables turned out to have the opposite effect as I predicted and, therefore, none of the hypotheses could be confirmed.

Furthermore, the four hypotheses predicting that the level of industry competition moderates the relationship between the independent variable and the dependent were rejected. As commented above, any of the interactions (Power Distance#ln_HH / Individualism#ln_HHI / Masculinity#ln_HHI / Uncertainty Avoidance#ln_HHI) resulted to be significant, implying that there is no moderation effect of the level of industry competition within this sample analysis.

Considering firm-level control variables, in all the models' Total Assets was significant (p-value always lower than 0.01) and presented negative coefficients. Contrary to what was expected, these results indicate a negative influence of the company size on the CSP scores. Surprisingly, the other two firm-level control variables, Return On Assets, and Leverage, did not result to be significant in any model.

(34)

34 At the industry-level, any of the industry dummy variables resulted to be significant in any model.

In addition, in order to check the reliability of the results and have a better picture of whether the findings are conclusive and reliable or if they need to be evaluated cautiously, a robustness check was performed (Table 7). This robustness check consisted of running an additional analysis with four Schwartz cultural dimensions as independent variables, instead of the four Hofstede cultural dimensions. For the robustness check to be valid the four chosen Schwartz cultural dimensions should be correlated with each Hofstede cultural dimension used (Individualism, Masculinity, Uncertainty Avoidance, and Power Distance). According to Schwartz (1994), the following dimensions of both cultural frameworks are correlated: Hofstede’s individualism score is positively correlated with his intellectual autonomy dimension (0,53); Hofstede’s power distance score is positively correlated with his hierarchy dimension(0,45); Hofstede’s uncertainty avoidance score is positively correlated with his harmony (0,43) and Hofstede’s masculinity score is positively related with his mastery dimension (0,56). Hence, intellectual autonomy, hierarchy, harmony, and mastery Schwartz dimensions were used for the robustness check analysis.

(35)

35

Table 6: Regression analysis coefficients

(1) (2) (3) (4)

(36)

(37)

37

Table 7: Robustness check analysis coefficients

(38)

38

DISCUSSION

In this research, I have analysed the role of national cultures in determining Corporate Social Performance variation across countries and how is this relationship moderated by the level of industry competition. The aim of this paper was to make a contribution to the CSR literature, giving support to the idea that institutions are a determining factor for CSR. Cooperatively, this study underlines an often under-researched informal institutional component, called culture, in the corporate sustainability environment. First, drawing on institutional theory and the mechanisms that companies tend to engage in sustainability-related activities in a manner that is consistent with cultural traits, this study seeks to indicate a set of Hofstede’s cultural dimensions that are relevant to sustainability. Secondly, this research also aims to address the gap of which is the level of competition circumstances in which companies engage in CSR. In order to so, I analysed how the relationship between these cultural dimensions and sustainability performance is moderated by the industry level of competition.

Companies' home country Masculinity, Individualism, Power Distance, and Uncertainty Avoidance scores were expected to have an impact on its CSP scores, explaining CSP scores heterogeneity across countries. For the present research, I relayed on CSP scores, by drawing on Thomson Reuters ASSET4, gathering scores of 1200 firms from 30 nations for the year 2018, and all the hypotheses were tested using an ordinary least square regression (OLS). As will be explained below, the results indicate that some companies' home country cultural characteristics may well have a key role in the explanation of CSP heterogeneity across countries.

(39)

39 Contrary to my expectations, in all models Individualism dimension resulted to be non-significant and presented negative estimated coefficients, therefore, rejecting Hypothesis 1. When considering the non-significance of this variable, we should take into consideration the divisive nature of the individualism Hofstede dimension. It is a dimension which has been linked with insignificant findings also in other analysis (Hartmann & Uhlenbruck, 2015, Haxhi & van Ees, 2010, Ioannou & Serafim, 2012, Perk et al, 2007, Ringov & Zollo 2007, Scholtens & Dam, 2007), or with results showing the opposite relationship (Husted, 2005, Peng et al, 2014, Williams & Sinkin, 2008).

Despite that practices related to the sustainability of the environment and society are more coherent within decision-makers from collectivist societies, some researchers argued the contrary (Husted, 2005, Ioannou & Serafim, 2012), that higher levels of individualism involve greater CSR investments. Husted stated that in individualistic cultures there is a higher presence of interest groups instead of centralized associations (2005). Therefore, given the greater presence of environmental groups, individualistic societies should present institutional capacity to respond favourably to sustainability issues (Husted, 2005). In line with this, Ioannou and Serafim (2012) provided another possible explanation. They suggested that decision-makers from companies placed in high individualistic countries are more willing to take explicit decisions, which leads to higher performance in general, increasing the CSP scores. Generally speaking, it is possible that there is a trade-off between individualist and collectivist societies when it comes to invest and engage in CSR related practices, which, in turn, leads to no significant and conclusive differences in CSP scores. Furthermore, it is worth pointing out that the insignificance of the results may be given due to the characteristics of the sample or for outside control factors that I did not control.

(40)

40 However, contrary to my expectations, when it comes to analyse which is the direction of this effect, the coefficients estimated resulted to be positive. Therefore, according to these results companies based in higher masculine countries present higher CSP scores, which contradicts the predictions of Hypothesis 2. Interestingly, my rejected hypothesis was in line with most of the prior studies in the field (Gallego Alvarez & Ortas, 2017, Husted, 2012, Park et al., 2007, Peng et al., 2014, Scholters & Dam, 2007, Vachon, 2010, Winterich & Zhang, 2014), 2008, Peng et al., 2014), which found a negative relation between masculinity and sustainability. A possible explanation for this difference may be a consequence of whether endogeneity issues of the culture dimensions are taken into consideration and the manner in which each analysis utilizes and control variables. However, this paper is not the only one which findings suggested a positive relation between masculinity and sustainability (Williams & Zinik, 2008, Ho et al. 2012), suggesting that it may be a plausible relationship depending on the sample the study analyses. Therefore, these contradictions call for further analysis in future research.

Hypothesis 3 predicted that companies based in greater uncertainty avoidance countries will have higher CSP scores. Uncertainty Avoidance is the only cultural dimension in the analysis which a positive relationship is expected. The argument behind this positive relationship is that companies placed in high uncertainty avoidance countries may have a tendency to have a more sustainable relationship with their stakeholders in order to reduce the level of uncertainty. Higher investments in CSR is may be one important way to make these relationships more sustainable and strong and for this reason, CSR engagement may constitute an effective corporate strategy for companies within greater uncertainty avoidance societies.

(41)

41 gave a plausible explanation for the negative relationship between uncertainty avoidance and sustainability. He argued that uncertainty avoidance is negatively related to innovation, which is a core part of incrementing the efficiency of the sustainability related practices, especially being a foundational element of contamination prevention. Hence, the lower innovation of the companies placed in high uncertainty avoidance countries may have a negative in its CSP scores. Therefore, the explanation than in uncertainty avoidance societies, business is less willing to implement CSR related practices seem to be reasonable and it may well be the case within this sample.

Hypothesis 4 predicted that companies based in greater power distance countries will have lower CSP scores. The negative correlation between power distance and CSP is based upon the argument that in high power distance societies people tend to accept unsustainable behaviour easier, at the same time as managers that support these conducts are held to account less than they would in lower power distance cultures (Cohen, Pant & Sharp, 1996).

Power Distance resulted to be significant in model 2 and 3 at the ** p < 0.01 level, being less significant than Uncertainty Avoidance but more than Masculinity. Contrary to what I expected, the coefficient estimate in model 2 and 3 was positive. Therefore, these results refute Hypothesis 4 according to which companies based in greater power distance countries have lower CSP scores. My hypothesis was consistent with most of the earlier studies (Walmann et al., 2006, Husted, 2005, Ioannou & Serafeim, 2012, Hartman & Uhlenbruck, 2015, Cai et al, 2016, Parboteeah et al., 2015). Nevertheless, it is worth to point out the controversial nature of this dimension, since it has also been associated with a negative relation with sustainability. (Alas, 2006, Ho et al., 2012, Witt & Stahl, 2016). The idea behind this is the suggestion that sustainability procedures are applied to achieve consensus between stakeholders with regard to their economic behaviour. Moreover, stability, social order, and harmony are deemed peculiarities in power distance societies and, in general, these societies have more respect for companies. Therefore, as stated by Miska (2018), the proposal that, in high power distance societies, firms may be more inclined to increment their CSR engagement appears to be consistent.

(42)

42 variables. For the robustness check to be valid the four chosen Schwartz cultural dimensions were correlated with each of the Hofstede cultural dimensions used in the analysis (Individualism, Masculinity, Uncertainty Avoidance, and Power Distance). The four Schwartz cultural dimensions which have the strongest correlation are Intellectual Autonomy, Hierarchy, Harmony, and Mastery, hence, these were picked for the robustness check analysis.

In the robustness check analysis, only two cultural dimensions resulted to be significant (Intellectual Autonomy and Mastery). Surprisingly, both cultural dimensions are correlated with the cultural dimensions which were less significant in the analysis, Individualism, and Masculinity respectively. Intellectual Autonomy, which is a dimension strongly correlated with Individualism resulted to be significant at the *** p < 0.001 level, whereas Individualism was the only dimension that did not show any significance in model 2. Furthermore, Hierarchy and Harmony, which are correlated with Power Distance and Uncertainty Avoidance respectively, resulted no to be significant, while Power Distance and Uncertainty Avoidance were the dimensions which showed to be more significant in the regression analysis. Regarding the coefficients, it is surprising to see that all the dimensions show a completely opposite sign in the robustness check analysis. Therefore, these results question the robustness of the results and suggest that the findings need to be evaluated cautiously.

Apart from this, there is a point which I think is worth to mention from the robustness check analysis. It is interesting to point out that the squared was 0.0697 points higher than the R-squared in model 2 of the main regression analysis. Therefore, this means that within the selected sample, the Schwartz cultural dimensions have more explanatory power than the Hofstede cultural dimensions.

(43)

(44)

44

CONCLUSION

The main goals of this paper were to investigate how is the impact of the companies' home country national culture on its CSP scores, as well as to examine whether the strengthening of this relationship changes within different levels of industry competition. Therefore, the main goal of this study was to provide new facts on the importance of institutional factors such as cultural characteristics as a determining factor for heterogeneity of CSP scores across countries. The results of the analysis show that companies' home country uncertainty avoidance had negative effects, whereas companies' home country masculinity and power distance impact positively on their CSP scores. These findings were in the opposite direction of the predictions I made, which corroborate the challenge of doing empirical research in this field, leading to significantly mixed-effects throughout the literature. Taking all of these into account, the present research has some theoretical and managerial implications.

Theoretical Implications

he first objective of this study was to make a contribution to the theoretical and empirical literature about corporate social responsibility and, therefore, brings light into the existing discrepancies found within the literature researching the influence of national culture on CSP Most of the researches within this literature has used Hofstede’s framework measurement to evaluate the cultural dimensions in their empirical analysis (). Furthermore, within these studies, little consistency and uniformity had been observed. In order to have a better comparison with previous literature on the field, I decided to perform my research with Hofstede’s cultural dimensions framework.

(45)

45 Nevertheless, the direction of this effect has shown to be contrary to what was expected. As pointed out above, any of the three significant cultural dimensions directional effect is consistent with most of the previous studies in the playing area, thus adding inconsistency and uniformity to the literature in the field.

The second goal of this study contributed to the theoretical and empirical literature about corporate social responsibility by addressing the gap of how this influence is moderated by the level of industry competition and, thus, provide new useful information about how the joint effect of national culture and the level of industry competition effect the firms’ CSP.

As shown, contrary to my expectations, the level of industry competition resulted not to have any significant moderating effect on the impact of national culture on sustainability. Therefore, it can be concluded that the level of industry competition does not moderate the effects of firms’ home country cultural traits on Corporate Social Performance (CSP) variation across different countries. Nevertheless, this insignificance may be given because of a problem of low statistical power of the present sample, so these findings are not conclusive and thus calls for further analyses in future research.

To conclude, even though three out of four independent variables resulted to be significant, the mixed findings within the literature in the field, the non-significance of the individualism dimension, and the moderator variable calls for more research and detailed analysis in future research.

Managerial Implications

(46)

46 (2015), managers could construct CSR strategies in a manner that is consistent with the cultural traits of the area of interest.

Secondly, as stated by Hartmann & Uhlenbruck (2015), a big problem for multinationals is that they are usually carefully examined concerning CSR activities in the nation in which they locate their business operations and frequently are considered responsible for dubious practices. That's why, this study might help managers from multinationals with an international interrelated supply chain, providing knowledge about potential explanations for suppliers' sustainability performance. Hence, when making decisions regarding the relocation to a different country or the partner’s selection, decision-makers should keep in mind that there are some cultures with less inclination to engage in sustainability-related practices and, thus, may represent a risk to the reputation of the company.

In conclusion, given that decision-makers' decisions shape companies' sustainability performance, it is crucial that they take into account which factors may have an influence on their company CSP, such as culture and the level of industry competition.

Limitations and Future Research

This study presents some limitations, which ought to be taken into account while interpreting the results.

The first limitation is related to the sample used for the sample. While prior studies used a large sample of more than 10.000 companies, the current research is limited to 1.152 observations only. This may have caused a slight problem of statistical power which may have limited the reliability of the results of the moderating effect. Additionally, US firms are represented in the sample, thus future studies may take into account collecting more uniform representations of nations. In addition, future research may consider a longer time horizon period in order to confirm the CSP variation throughout time.

(47)

47 of the international business literature have underlined the necessity of a more precise and modern measurement of national culture variation. (i.e. Drogendijk & Slangen, 2006; Shenkar, 2001). Furthermore, in this study, we only used four of the seven Hofstede cultural dimensions. Hence, it may be interesting for future studies to use other dimensions to study for their effect of CSP differences across countries.

Thirdly, there might be possible biases in the cultural impacts on the firms included within the sample. Multinationals which are heavily internationally oriented might be affected and influenced by informal institutions belonging to the states where they are operating. Hence, as Graafland and Noorderhaven (2018) pointed out, it would be interesting for future studies to concentrate on the relevance of companies' internationalization on the relation between national culture and sustainability performance scores. Ho et al. (2012) for example, indicated that domestic firms may have features, like the more important of local stakeholders, which might influence positively firms engagement in sustainability practices. One more exciting opportunity for the future may take into account corporate culture’s role in the study of the correlation between national culture and corporate social responsibility.

(48)

48

REFERENCES

Aguilera, R., Jackson G. (2003) The cross-national diversity of corporate governance dimensions: dimensions and determinants. Academy of Management Review, 28: 447-465. Aguilera, R., Rupp, D., Williams, Ganapathi, J. (2007). Putting the S Back in Corporate Social Responsibility: A Multilevel Theory of Social Change in Organizations. The Academy of Management Review, 32: 836-863.

Akkermans, D., Harzing, A., Witteloostuijn, A. (2010). Cultural Accomodation and Lenguague Priming. Management International Review 50(5):559-583.

Alas, R. (2006). Ethics in countries with different cultural dimensions. Journal of Business Ethics, 69(3): 237–247.

Arminen, H., Puumalainen, K., Patari, S., Fellnhofer, K. (2018). Corporate social

performance: Inter-industy and international differences. Journal of Cleaner Production, 177: 426-437.

Arnold, D., Bernardi, Neidermeyer P., Schemee, J. (2007). The Effect of Country and Culture on Perceptions of Appropriate Ethical Actions Prescribed by Codes of Conduct: A Western European Perspective Among Accountants. ournal of Business Ethics 70(4):327-340.

Akaah, I. P. (1990). Attitudes of marketing professionals toward ethics in marketing research: A cross-national comparison. Journal of Business Ethics, 9(1), 45-53.

Bagnoli, M., Watts, S. (2004). Selling to Socially Responsible Consumers: Competition and The Private Provision of Public Goods. Journal of Economics & Management Strategy, 12(3).

Baron, D. (2001). Private Politics, Corporate Social Responsibility, and Integrated Strategy. Journal of Economics & Management Strategy, 10.

Beekun, R.I., Hamdy, R., Westerman, J.W. et al. (2008). An Exploration of Ethical Decision-making Processes in the United States and Egypt. J Bus Ethics, 82: 587–605.

Brammer, S. (2008). Does It Pay to Be Different? An Analysis of the Relationship Between Corporate Social and Financial Performance. Strategic Management Journal 29(12):1325 – 1343.

What is the Effect of Firms’ home National Culture on its Corporate Social Performance? And How does the Level of Industry Competition moderates this relationship?