Cultural Values and International Differences in Business Ethics Bert Scholtens Lammertjan Dam ABSTRACT. We analyze ethical policies of firms in industrialized countries and try to find out whether culture is a factor that plays a significant role in explaining country differences. We look into the firm’s human rights policy, its governance of bribery and corruption, and the comprehensiveness, implementation and communication of its codes of ethics. We use a dataset on ethical policies of almost 2,700 firms in 24 countries. We find that there are significant differences among ethical policies of firms headquartered in different countries. When we associate these ethical policies with Hofstede’s cultural indicators, we find that individualism and uncertainty avoidance are positively associated with a firm’s ethical policies, whereas masculinity and power distance are negatively related to these policies. KEYWORDS: business ethics, codes of ethics, cultural values JEL: G300, L210, M140 Introduction Are there differences with respect to the ethical policies of firms that are headquartered in different countries? And are there differences among firms that belong to different industries? Chryssides and Kaler (1996), Ferrell et al. (2000), and Crane and Matten (2004) discuss that the conduct of business emerges and evolves in response to religious, philosophical, societal, economical, and institutional concepts and notions. They also point out that ethical theories can help to clarify the different moral presuppositions of the various parties involved in a decision or action (e.g. Chapter 3 in Crane and Matten, 2004). As such, ethical theories are being applied to business ethics (see also De George, 1999; Ferrell et al., 2000). Then, we find that business ethics, as part of culture, does not happen in vacuum or isolation. It takes place in a social and cultural environment that is being governed by a complex set of laws, rules and regulations, formal values and norms, codes of conduct, policies, and various organizations (see Hofstede, 1991; Scott, 2001; Trompenaars, 1993). Ethical theories can be used to analyze the (changes in) ethics and ethical policies of business in time and among countries and industries. Berkert (1995) contends that corporations differ from individual agents with respect to their susceptibility for moral responsibilities. In his view, it is a special set of values, principles and ideas which regulates behavior in business. As ethical conduct of individuals and organizations is part of and very much intertwined with culture and society, it is quite common to assume that the ethics of firm behavior too will be subject to change (see also McInnes, 1996). While various explanations have been offered to explain these societal differences, an ever-growing body of literature argues that cultural differences between countries are one of the main drivers of a nation’s level of economic and entrepreneurial conduct (McGrath et al., 1992; Thomas and Mueller, 2000). Recognizing the critical role that culture plays in determining corporate behavior, several scholars have called for future research Bert Scholtens received his Ph.D. at the Universtiy of Amsterdam. Since 1999 he has been working at the Department of Finance of the University of Groningen, the Netherlands. His research particularly looks into the interaction between financial institutions and corporate social responsibility. He has published in, among others, Ecological Economics, Journal of Banking and Finance, Finance Letters, Journal of Investing, Sustainable Development, and Journal of Business Ethics. Lammertjan Dam is a Ph.D. student at the Universtiy of Groningen. He expects to defend his thesis about the integration of corporate social responsibility in economic valuation in Summer 2007. Journal of Business Ethics (2007) 75:273–284 Springer 2007 DOI 10.1007/s10551-006-9252-9 addressing the impact of national culture on corporate activity. For example, Sethi and Sama (1998) argue that in order to investigate ethical business conduct, both corporate and industry structure has to be considered (see also Zahra et al., 1999). They assess industry sectors on the basis of their structural and institutional opportunities towards exploitation. However, they do not test their framework. Thus, it is not clear how culture is related to the ethical conduct of firms in practice. Fortunately, much empirical research in this direction already has been undertaken. For example, in an empirical study after the adoption of voluntary codes of conduct, Bondy et al. (2004) find that there are significant differences between the UK, Germany, and Canada. Sanyal (2005) finds bribery differs significantly among countries and that it is both economic and cultural factors that are important explanatory factors of bribery. Many studies focus on particular aspects of ethical codes or on the use of codes in specific industries. For example, Koehn (2005) treats integrity of the firm as an important business asset (see also Pearson (1995) for a similar approach). Diller (1999) focuses on the improvement of customer relationships. As customer interaction differs per industry, this might be a determinant of the differences among industries. King and Lenox (2000) analyze the role of peer pressure in the chemical industry. Boatright (1999) goes into the role of ethics in finance (see also Statman, 2004) and Van Tulder and Kolk (2001) analyze the sporting goods industry. O’Higgins and Kelleher (2005) analyze the ethical orientations of human resources, marketing and finance managers, whereas Stevens et al. (2005) investigate the impact of ethics codes on financial executives’ decisions. The approach taken in our study is in line with a tradition that started with Langlois and Schlegelmilch (1990). These authors investigated codes of conduct for a large number of companies from different countries. They analyze 189 companies from the UK, (Western) Germany, and France and compare them with 174 firms from the US. Langlois and Schlegelmilch focus on large, predominantly industrial companies. They find that US firms have more codes of ethics than firms from Europe. When going into the content of the codes, Langlois and Schlegelmilch find various significant differences between the US firms and those from France and Germany and sometimes also the UK. This study was complemented by Schlegelmilch and Robertson (1995) who went into the ethical perceptions of senior executives in the US, the UK, Germany, and Austria. Their study also showed that the country has a significant impact. Kaptein (2004) investigates the content of the codes of conduct of 200 multinationals in 17 countries. He reports what elements are included in these codes and what stakeholder principles are addressed. Kaptein (2004) concludes that the companies specifically differ in what they include and exclude from their codes and inthe wording that is used. There is much research that finds that country origin is an issue in the content and design of ethical codes. For example, Wood (2000) for the US, Canada, and Australia, Hood and Logsdon (2002) for the US, Canada, and Mexico, Maignan and Ralston (2002) for the US, the UK, France, and the Netherlands, Reich (2005) for Germany, Japan, and the US, Lindfelt (2004) for Finland, Singh et al. (2005) for Australia, Canada, and Sweden, and Mele´ et al. (2006) for Argentina, Brazil, and Spain. We will try to bring this line of research one step further by analyzing the key attributes of ethics in different countries and industries. Our purpose is to come up with an assessment of the business ethics of a large number of firms in the tradition of Langlois and Schlegelmilch (1990). To this extent, we will use data from EIRIS to find out whether there are significant differences in the assessment of ethical policies of firms in different countries and industries. We use data for almost 2,700 firms from 24 countries and 35 industries. In this respect, our paper differs from other quantitatively oriented approaches as that of – among others – Sanyal (2005) who focuses on macro (country) data. Furthermore, we investigate how culture is to be associated with ethical conduct in different countries. To this extent, we use the Hofstede (1980, 1991) data to find out whether and how culture matters in this respect. The Hofstede database gives us detailed information about key dimensions of culture. As such, we analyze firms’ ethical policies on an international level from a micro perspective. We look into the different attributes of the firm’s relation with ethics and investigate whether and how they differ between firms operating in different countries. Hood and Logsdon (2002) and Singh et al. (2005) included Hofstede’s dimensions in their analyses and found 274 Bert Scholtens and Lammertjan Dam them relevant. However, they did not try to estimate the extent of the impact of cultural values on business ethics. As such, to our knowledge, this paper is the first to engage in a quantitative analysis of the association between international differences in business ethics and cultural values. We build on the findings of Langlois and Schlegelmilch (1990), Hood and Logsdon (2002), Kaptein (2004), and Singh et al. (2005). But there are some important differences. First is that we do not use a questionnaire but we base our data on an investigation that also uses other sources about the ethical codes of the firm. Second is that the quality of the codes is taken into consideration. Third is that we include more firms and more countries in our analysis. Fourth is that our firms are evenly spread across the whole spectrum of the economy. A fifth difference is that we relate ethical codes to cultural values on the basis of a quantitative model. The contribution of this paper is that it not only establishes the existence of important differences in the ethical conduct of firms in a large group of countries and industries, but it also aims at advancing the theoretical discussion of the character and direction of cultural differences in business ethics. The structure of the remainder of this paper is as follows. We first come up with a description of our dataset. Then, in Countries, we analyze firms’ ethical policies at the country level. In Culture and ethical conduct, we relate ethical policies at the country level to Hofstede’s measures of culture. The conclusion is in last section. Data and methodology This section introduces the data about codes of ethics and cultural values that are subject to our analysis. The data about codes of ethics are derived from Ethical Investment Research Service (EIRIS). EIRIS is a charity set up in the UK in 1983. EIRIS covers over 40 different areas including animal testing, military, environmental performance and human rights. It gathers the data on the basis of a questionnaire and a survey of the firms in six different areas: Environment, governance, human rights, positive products and services, stakeholder issues, and ethical concerns. The philosophical background of EIRIS is not very clear; it argues that ‘‘we do not promote on particular view on ethical issues’’, but ‘‘companies are judged fairly against common standards and meaningful comparisons can be made between them’’ (see http://www.eiris.org). The survey was conducted in late 2004 and EIRIS analyzes independent sources of information on companies, including regulatory authories’ databases. For some research areas, where external sources are not available, they rely on company responses to their questionnaires. Given the nature of this paper, we focus on ethics. This is compatible with the approach proposed by Krajnc and Glavic?(2005) who suggest a procedure for assessing companies on different aspects of sustainability. We find that ethics is one of these aspects. As such, we look into the firm’s governance of bribery and corruption, human rights and the systems or comprehensiveness, communication, and implementation of their ethical codes. EIRIS assigns grades on specific attributes in the different areas. This procedure implies that some subjectivity is involved in assessing the ethics of the firms. However, given the ways in which the topics and questions are framed (see also below), we are convinced that the research by EIRIS results in valid measures. Furthermore, we are very well aware of the fact that firms’ ethical policies may differ from their performance in this respect. An ethics code itself does not guarantee ethical behavior (Kitson and Campbell, 1996; see also Svensson and Woods, 2005). However, to our knowledge, there is no database that assesses the ethical performance of a large number of firms in different industries and countries. Therefore, we stick to the information about ethical policies and will refrain from deriving conclusions about their ethical behavior. To assess the firms, EIRIS has a scoring table which consists of six scales or grades. EIRIS does not provide an overall assessment or rating of the companies. Therefore, we give a score of three to the high positive grade, 2 to med positive, 1 to low positive, )1 to low negative, )2 to med negative, and )3 to high negative. With respect to the five key items, EIRIS answers the following questions: 1. Governance of bribery and corruption: Does the company have policies and procedures on bribery and corruption? Here, the firm can either have a clear policy and procedures, it has adopted or it has no policy disclosed. Cultural Values and International Differences 275 2. Systems of the codes of ethics: The first question about the firm’s code of ethics is whether the company does have a code of ethics and, if so, how comprehensive is it. The answer is either no, limited, basic, intermediate or advanced. 3. Implementation of the codes of ethics: The second question is whether the company does have a system for implementing a code of ethics and, if so, how comprehensive is it. The answer is either no, limited, basic, intermediate or advanced. 4. Communication of the codes of ethics: The third question is whether the company has adopted a code of ethics or business principles by which it communicates to all employees. The answer is either no evidence of, has adopted, or clearly communicates. 5. Human rights policy: What is the extent of policy addressing human rights issues? The answer is either no evidence of, has adopted, or clearly communicates. In our sample, we have that most of the firms are from the US and the UK (about 25% each). Japan ranks third with about one fifth of all the firms. The other 21 countries harbor the remaining 30% of the firms. Half of them are represented by less than 1% of the total number of firms. Luxembourg has only 3 firms in the sample and Portugal 8 (see Appendix 1). Firms based in Luxembourg were not assessed with respect to their human rights policy. Industries that are very well represented are the banks, media and entertainment, and support services (see Appendix 2). These three each have more than 5% of all the firms. However, it appears that our sample is quite well spread across the business sectors. There are two industries with less than 1% of all the firms: tobacco and water. Data for cultural values are derived from the Hofstede (1980, 1991) studies. His work consists of survey data about the values of people working in local subsidiaries of IBM in more than 50 countries. The actual surveys used in Hofstede (1980) date back to the 1970s. Updates and extensions have re-affirmed its main conclusions (see Hofstede, 1991). These data are used a lot in social and economic research (for example, see Garretsen et al., 2004; Licht et al., 2003; McGrath et al., 1992; Thomas and Mueller, 2000). The fact that the data are more than 30 years old is not a main concern under the assumption that culture changes very slowly over time. Another reason to use these data is that they pertain to general features of culture for the countries in the sample. This suits our research objective since we want to emphasize the role of cultural values that are general and not specific to certain markets or transactions. Hofstede (1980) defines the following societal or cultural indicators: PDI: Power distance is defined as the extent to which the less powerful members of institutions and organizations within a country expect and accept that power is distributed unequally. As such, it measures societal inequality. IDV: Individualism pertains to societies in which the ties between individuals are loose: everyone is expected to look after himself. Collectivism pertains to societies in which people from birth onwards are integrated into groups, which throughout their lives continue to protect them in exchange for unquestioning loyalty. MAS: Masculinity; this property shows the desirability for assertive behavior against the desirability of modest behavior. It appears that in some societies there are strong differences in answers given by men or women. In the modest countries the differences in gender are weak, but in assertive countries differences are strong. UAI: Uncertainty avoidance is defined as the extent to which the members of a culture feel threatened by uncertain or unknown situations. It is more general than risk avoidance, which is defined with respect to a certain object. Countries In this section, we analyze whether the firms differ from one each other with respect to human rights policy, governance of bribery and corruption, and the comprehensiveness (i.e. the actual systems in place), implementation, and communication of their codes of ethics in case the firms are clustered by country. As such, we try to find out whether there are significant differences in ethical policies along different countries. First, we discuss the scores of the firms in the different countries. 276 Bert Scholtens and Lammertjan Dam Table 1 reveals that the average EIRIS-score on the governance of bribery and corruption is 1.97. In this respect, firms from the US and Norway perform best. Companies from Australia, Italy, the Netherlands, and Finland also perform well. Firms from Luxembourg, Singapore, Hong Kong, Spain, Portugal, and Ireland perform weak on their governance of bribery and corruption. The average firm score on the extent and quality of the systems of the codes of ethics is 0.25. As to these systems, US, Australian, and Dutch firms perform best. Here, firms from Luxembourg, Singapore, and Hong Kong perform worst. With respect to the communication of the codes of ethics, the average firm score is 2.38. Here, firms from the US, Australia, and New Zealand top the ranking. Those from Luxembourg, Singapore, Hong Kong and Ireland rank lowest. As to the implementation of the codes of ethics, it is again US firms that receive the highest ratings from EIRIS. Firms from Luxembourg, Singapore, and Hong Kong perform worst. The average firm score on human rights policies is 0.31. Here, the 3 companies from Luxembourg were not given a score. Firms from Finland, Norway, and Sweden got on average the highest score on their human rights policies. Firms from Ireland, New Zealand, Portugal, and Singapore scored lowest. In all, it appears that firms based in the US and Scandinavia, and – excluding human rights policies – those from Australia and New Zealand did receive the highest scores on the five attributes of business ethics. Firms from Luxembourg, Singapore, Hong Kong, Ireland, and Portugal show the poorest results. In Culture and TABLE 1 Mean score of firms in the 24 countries on the five attributes of business ethics Number of firms Governance of bribery and corruption Codes of ethics systems Communication of codes of ethics Implementation of codes of ethics Human rights policies Australia 115 2.30 0.97 2.86 1.97 )0.11 Austria 13 1.69 0.00 2.00 0.15 1.00 Belgium 15 1.87 0.53 2.27 0.93 0.00 Canada 85 2.11 0.65 2.55 1.28 0.50 Denmark 15 1.80 0.13 2.33 0.67 1.50 Finland 16 2.25 )0.06 2.44 1.44 1.88 France 79 2.09 0.22 2.39 0.91 1.54 Germany 89 1.87 )0.39 2.17 0.30 0.72 Greece 15 1.60 )0.67 2.07 )0.07 0.50 Hong Kong 106 1.26 )1.36 1.54 )0.98 )0.85 Ireland 16 1.50 )0.81 1.81 )0.25 )1.00 Italy 54 2.30 0.17 2.41 1.37 0.40 Japan 487 1.64 0.40 2.21 0.28 )0.19 Luxembourg 3 1.00 )2.00 1.00 )2.00 Netherlands 38 2.26 0.84 2.61 1.68 1.32 New Zealand 23 2.17 0.43 2.70 1.52 )1.00 Norway 13 2.46 0.77 2.54 1.54 1.80 Portugal 8 1.50 0.38 2.63 1.25 )1.00 Singapore 49 1.10 )1.76 1.55 )1.10 )1.00 Spain 48 1.42 )0.90 2.27 )0.08 0.45 Sweden 42 1.88 )0.24 2.29 0.81 1.65 Switzerland 45 2.09 0.04 2.38 0.98 0.81 UK 656 1.82 )0.18 2.10 0.33 0.92 USA 651 2.49 1.04 2.93 2.17 0.32 All 2681 1.97 0.25 2.38 0.88 0.31 Cultural Values and International Differences 277 ethical conduct, we will try to find out whether these international performance differences can be related to differences in cultural values. To find out whether there are significant differences in ethical policies in the different countries, we perform an ANOVA (see Pindyck and Rubinfeld, 1985). The null hypothesis with the ANOVA is that the population means are identical. Rejection of H0 tells us that not all population means are equal. The issue in this section is whether the ethical policies of the firms with respect to human rights policies, the governance of bribery and corruption, and the systems, implementation, and communication of their codes of ethics does significantly differ among the firms in 24 countries. This indeed is the case for all five key variables; as the probability of the F-statistic in all instances points out that the firms within the various countries perform significantly different from the population’s average at the 1% confidence level and we may reject the H0 that the populations are equal. To investigate how different the ethical policies are among our 24 countries, Table 2 gives the number of indicators that are at least two standard deviations away from the mean score on each indicator of all firms (i.e. confidence >95%). For example, Finnish and French firms show a significantly higher score than the average firm on their human rights policy. Table 2 shows that most TABLE 2 Differences in ethical policies of firms with respect to countries (2 standard deviations above the mean = +1; 2 standard deviations below the mean = )1) Governance of bribary and corruption Code of ethics - systems Communication of code of ethics Implementation of code of ethics Human rights policy Total Australia +1 +1 +1 +1 0 +4 Austria 0 0 0 0 0 0 Belgium 0 0 0 0 0 0 Canada 0 +1 +1 0 0 +2 Denmark 0 0 0 0 0 0 Finland 0 0 0 0 +1 +1 France 0 0 0 0 +1 +1 Germany 0 )1 )1 00 )2 Greece 0 0 0 0 0 0 Hong Kong )1 )1 )1 )1 )1 )5 Ireland )1 0 )1 0 )1 )3 Italy +1 0 0 0 0 +1 Japan )1 +1 )1 )1 )1 )4 Luxembourg )1 )1 )1 )1 )4 Netherlands 0 +1 +1 0 0 +2 New Zealand 0 0 +1 0 +1 +2 Norway +1 0 0 0 0 +1 Portugal 0 0 0 0 )1 )1 Singapore )1 )1 )1 )1 )1 )5 Spain )1 )1 0 )1 0 )3 Sweden 0 0 0 0 +1 +1 Switzerland 0 0 0 0 0 0 UK )1 )1 )1 )1 +1 )4 USA +1 +1 +1 +1 0 +4 Total number of differences 2 standard deviations above / below mean 11 11 12 8 10 52 278 Bert Scholtens and Lammertjan Dam countries consistently either outperform or underperform the average firm in the sample. Only Japanese and British firms score significantly above the average firm on some items whereas they score significantly below average on other items. From Table 2, we conclude that there are substantial differences indeed. Firms from Australia and the US are significantly outperforming the other firms in most respects. Firms from Hong Kong, Singapore, Luxembourg, the UK and Japan perform worse than most other firms. Firms from Austria, Belgium, Denmark, Greece and Switzerland do not significantly differ from the average firm in the sample. So, we find that there are significant differences in the characteristics of ethical policies of firms located in different countries. This finding is in line with results found elsewhere in the literature (see e.g. Bondy et al., 2004; Hood and Logsdon, 2002; Kaptein, 2004; Langlois and Schlegelmilch, 1990; Lindfelt, 2005; Maignan and Ralston, 2002; Mele´ et al., 2006; O’Higgins and Kelleher, 2005; Reich, 2005; Singh et al., 2005; Stevens et al., 2005; Wood, 2000). But we established our conclusion on a much larger number of countries and industries and on the basis of much more firms. Therefore, we have succeeded in generalizing the existing observations. Note, however, that it may be the case that because of differences in the industrial structure of countries, the industry results are driven by the country differences. A simple Chi-square test of independence rejects the hypothesis that industry and home country are independent variables (Chi-square test statistic for independence of industry and country is equal to 209, dF = 120, p-value = 1.) Thus, indeed, there is significant dependence between country and industry. Therefore, in the remainder of this paper, we will focus on the interaction between culture and country differences as to firms’ ethical conduct. This also has a very practical reason, namely the fact that our data about culture are on a country basis and, unfortunately, not available on an industry basis. Culture and ethical conduct In this section, we investigate how culture affects firms’ ethical conduct. First, we will go into the ideas about the association between the two and then we will perform a simple test. Culture is a multifaceted concept. Literally, it means to build on, to cultivate, or to foster. But many authors have given their own interpretation and various schools of thought concerning the concept culture have emerged (see Bodley, 2005, for an overview). For example, there are the concepts of mass culture and popular culture, where it relates to taste and values. Alternatively, theories evolved that regard culture as values shared among different social groups and classes. Others view culture as a set of values and characteristics of a given group, the relation of an individual to culture, and his/her acquisition of those values and characteristics (see Soley and Pandya, 2003). Hofstede (1980) refers to this vision as the collective programming of the mind. Bodley (2005) argues that a crucial feature of culture is that people learn it. A lot of aspects of life are transmitted genetically, such as the desire for food. A person’s specific desire for milk and cereal or for a croissant and coffee in the morning, on the other hand, cannot be explained genetically. Culture, as a body of learned behaviors common to a given human society, has a predictable form and content and shapes behavior and consciousness within society from generation to generation. Then, according to Bodley (2005), culture resides in learned behavior as well as in some shaping consciousness prior to behavior. Language, organization, and technology are probably the most important elements of culture. Cultural differences manifest themselves in various ways. The deepest manifestation of culture is the set of values. Values are broad tendencies to prefer certain states of affairs over others. Norms are the standards for values that exist within a group or category of people. More superficial differences in culture can be found in symbols and rituals. Values are at the core of economic behavior and could help explain differences in the conduct of firms (Bodley, 2005). For example, Zaheer and Zaheer (2006) use cultural values to investigate international collaboration of business households, especially trust. Different cultures have their own mores of what is acceptable and unacceptable conduct. And each culture has methods for dealing with the violation of social norms (Svensson and Wood, 2003). Values are affected by the environment, by the cultural context. In this respect, Hofstede (1980) defines his four cultural values: uncertainty avoidance, power distance, Cultural Values and International Differences 279 individualism versus collectivism, and masculinity versus feminism (see Data and methodology). Hood and Logsdon (2002) as well as Singh et al. (2005) use the Hofstede dimensions to assess the international differences in business ethics. However, both studies only investigate three countries and do not use the exact scores on the Hofstede indicators in their analysis. Now, we try to relate the Hofstede (1980, 1991) data to the firms’ scores with respect to ethical conduct. Given the discussion above and the description of the data in Data and methodology, we expect that Hofstede’s indicators are significantly related to the various attributes of firms’ ethical policies. We expect that power distance and masculinity are negatively related to firms paying a lot of attention to ethical issues. This is because power distance measures societal inequality. We assume that countries that are characterized by relatively more inequality will also be characterized by relatively little attention for ethics. As to masculinity, we expect that firms in countries that are more assertive will regard their ethical policies of little importance and that they have a lower score in this respect. On the other hand, we expect that individualism and uncertainty avoidance are positively related to ethical conduct. Individualism puts an agent’s own responsibility on the foreground and, therefore, we expect that in countries with a relative high score on this indicator, firms will pay more attention to their ethical policies. As to uncertainty avoidance, we expect a positive association because firms in countries that feel relatively more threatened by uncertain and unknown situations will want to have the systems in place to deal with such situations which will, in our opinion, result in more attention being paid to codes of conduct and ethical policies. In order to test for these hypotheses, we use a simple linear model of the following general form: ETHICSi ¼ ai þ biCULTUREi þ ei: Where ETHICSi is the dependent variable reflecting the score of the average firm in a country on one of the indicators of ethical values (human rights, codes of ethics systems, codes of ethics communication, codes of ethics implementation, stance on corruption), ai and bi are parameters, and CULTUREi is a vector of the explanatory variables. For this vector, we take as independent variables the ones suggested by the Hofstede study (uncertainty avoidance, individuality, power distance, masculinity), and ei is an error term. Please note that this approach is a very simple and rough one in which we implicitly make a lot of assumptions about the dataset. Many of them will not hold. However, the estimations are undertaken to arrive at least at some preliminary insights into the association between the ETHICS and CULTURE variables. Table 3 gives the estimation results from our regressions of this model. All estimations have a reasonable explanatory power and the F-test shows that the models appear to be adequate descriptors. However, given the small number of observations, we have to be careful with drawing conclusions from these results. It appears that power distance and masculinity do have a negative association with the culture variables but in most circumstances, except TABLE 3 Estimation results (17 countries) Human rights Ethics systems Ethics communication Ethics implementation Corruption Coefficient p-value Coefficient p-value Coefficient p-value Coefficient p-value Coefficient p-value constant 1.8903 0.02 )1.9534 0.02 1.3768 0.00 )1.5629 0.07 0.9765 0.02 UAI 0.0151 0.02 0.0111 0.07 0.0029 0.31 0.0084 0.20 0.0041 0.20 IDV 0.0021 0.76 0.0266 0.00 0.0133 0.00 0.0348 0.00 0.0146 0.00 PDI )0.0249 0.01 )0.0063 0.43 0.0002 0.95 )0.0014 0.87 )0.0018 0.66 MAS )0.0229 0.00 )0.0021 0.68 )0.0025 0.32 )0.0080 0.18 )0.0036 0.21 adj. R2 0.6974 0.5811 0.5347 0.6126 0.5555 F-sign. 0.0008 0.0049 0.0089 0.0032 0.0069 280 Bert Scholtens and Lammertjan Dam for human rights policies, this relation is insignificant. Uncertainty avoidance and individuality are positively associated with the ethical conduct variables. In the majority of the cases this is a significant relationship. Individuality is highly significant with the ethical variables, except with human rights policies. Uncertainty avoidance only is significantly positive associated with ethical policies in the case of human rights policies and the codes of ethics systems. These results in part confirm our hypotheses. The ‘strongest’ finding is for the positive association between individuality and ethical conduct, but not with human rights policies. Uncertainty avoidance has the expected positive sign and is significant in two of the five cases. Masculinity also has the expected negative sign but is significant in one case only. Power distance has the expected negative sign in four of the five cases but is significantly negative in only one case. Power distance is positive but insignificantly associated with ethics communication. The results are in line with those found elsewhere in the literature. Especially, they confirm the findings of, among others, Langlois and Schlegelmilch (1990) about the US, the UK, France, and Western-Germany for a much larger sample of countries and firms. More specifically, our findings extend and generalize the observation by others such as Langlois and Schlegelmilch (1990) and Bondy et al. (2004) that there are significant differences in the codes of ethics to the observation that there also are significant differences with respect to the quality of these codes as assessed by an external independent rating agency. Furthermore, our association between cultural values and different attributes of codes of ethics substantiates the ideas put forward by Seth and Samal (1998). The results also complement the conclusions derived from sectoral, country, and functional studies by, among others, Van Tulder and Kolk (2001), Kaptein (2004), Lindfelt (2005), and Stevens et al. (2005). Conclusion On the basis of our analysis, we find for our sample of almost 2,700 firms in 24 countries that the location where the firm is headquartered appears to be a significant factor when it comes to the assessment of the firm’s communication, implementation and the systems of the code of ethics (comprehensiveness), its governance of bribery and corruption, and its human rights policies. We find that there are significant differences between these attributes in the 24 countries and among the 35 industries investigated. For example, firms from the US, Australia and Scandinavia perform significantly better than the average firm in the sample, whereas those from Luxembourg, Singapore and Hong Kong perform relatively poor. We can not detect a clear relation between economic development and firm’s ethical policies. For example, when we associate the ranks of the 24 countries on ethical policies with the countries’ ranks on per capita GDP, we have a correlation coefficient of only 0.24. Please keep in mind that we look into firms’ ethical policies, that are their human rights policies, the governance of bribery and corruption, and the systems, implementation and communication of their codes of ethics. On the basis of our dataset, it is not possible to assess the ethical performance or the ‘true’ ethical behavior of the firms. Our findings suggest that firms’ non-financial conduct is shaped by a combination of firm specific, industry specific, country specific and global factors. Furthermore, each firm’s unique set of characteristics is seen to shape the responses of the firm to specific challenges. We also undertook a very preliminary investigation into how the ethical conduct of firms might be associated with Hofstede’s societal norms and cultural values. This analysis was undertaken in a simple but novel manner. In many cases, we find that specific cultural values can be significantly associated with ethical policies of firms in the countries under investigation. Especially, individualism and uncertainty avoidance are positively associated with firms’ ethics, whereas masculinity and power distance tend to be negatively associated. These observations are in line with those found elsewhere in the literature (see Gnyawali, 1996; McGrath et al., 1992; Sanyal, 2005; Thomas and Mueller, 2000). For companies, our research implies that they should be well aware of the differences in business ethics in different countries and industries. This especially seems relevant if they want to export or invest abroad. Incongruence may lead to smaller chances of acceptance of the firms’ products and services and/or to higher costs with respect to acquiring human or financial resources. Cultural Values and International Differences 281 The major weakness of our study is that, so far, we lack a clear-cut theory about the exact interaction between ethics and culture. Therefore, we are unable to actually put hypotheses to the test. Rather, our research results in preliminary findings about the associations between the two. Furthermore, the quality and timeliness of the dataset (especially the culture variables) is a matter of concern. To conclude, we have established that there are significant international differences in ethical policies. Cultural values are an important determinant in this respect. As such, our analysis of variance has confirmed and generalized notions that have existed for long in the literature. The preliminary regression analysis suggests how different cultural values are to be associated with firms’ ethical conduct. Our research also gives rise to new questions. For example, a very interesting and logical question is whether firms’ attitude towards ethical issues is related to ethical performance. A major challenge we face is to come up with a theory of how ethics and culture interact. Also, we would love to have access to better data about societal norms and cultural values in a much larger number of countries and, especially, industries. Further research will have to shed light on these matters. APPENDIX 1 Composition of the data sample with respect to countries Number of firms % of total Australia 115 4.3 Austria 13 0.5 Belgium 15 0.6 Canada 85 3.2 Denmark 15 0.6 Finland 16 0.6 France 79 2.9 Germany 89 3.3 Greece 15 0.6 Hong Kong 107 4.0 Ireland 16 0.6 Italy 54 2.0 Japan 487 18.2 APPENDIX 1 Continued Number of firms % of total Luxembourg 3 0.1 Netherlands 38 1.4 New Zealand 23 0.9 Norway 13 0.5 Portugal 8 0.3 Singapore 50 1.8 Spain 48 1.8 Sweden 42 1.6 Switzerland 45 1.7 UK 656 24.5 USA 651 24.3 Total 2,683 100.0 APPENDIX 2 Composition of the data sample with respect to industries Sector Number of firms % of total Aerospace & Defence 26 1.0 Automobiles & Parts 65 2.4 Banks 184 6.9 Beverages 33 1.2 Chemicals 90 3.4 Construction & Building Materials 117 4.4 Diversified Industrials 42 1.6 Electricity 62 2.3 Electronic & Electrical Equipment 96 3.6 Engineering & Machinery 98 3.7 Food & Drug Retailers 36 1.3 Food Producers & Processors 79 2.9 Forestry & Paper 27 1.0 General Retailers 124 4.6 Health 84 3.1 Household Goods & Textiles 76 2.8 Information Technology Hardware 115 4.3 Insurance 75 2.8 Leisure & Hotels 80 3.0 Life Assurance 30 1.1 Media & Entertainment 138 5.1 Mining 31 1.2 282 Bert Scholtens and Lammertjan Dam Acknowledgements We wish to thank EIRIS for sharing their data with us. We are very grateful to IMUG – Hannover (Germany) for discussing the EIRIS database. We especially want to thank Silke Riedel and Kirein Franck. EIRIS nor IMUG can be held responsible for the analysis we conduct on the basis of their data. Furthermore, we thank several participants of the international workshop on CSR at London Metropolitan University, October 2005, and of on Business Ethics at the Freie Universita¨t Berlin, May 2006, as well as the participants in research seminars at the University of Groningen. Especially, we acknowledge the help and useful suggestions of Johanna Fernholm, Lise-Lotte Lindfelt, and Elmer Sterken. However, only the authors are responsible for all interpretations and conclusions as well as for all remaining errors and omissions.