Board demographic diversity and human rights reporting in Western Europe

Purpose – This paper aims to examine the relationship between board demographic diversity and human rights reporting fora sample of largeWestern Europeancompanies. Design/methodology/approach – Grounded on resource dependence theory, the authors hypothesize that greater gender, age and nationality diversities will translate into enhanced levels of human rights reporting. The authors use ordinal logistic regression analysis to analyze the association between these types of board diversity andsuch reporting. Findings – The ﬁ ndings suggest that the companies in the sample attribute little importance to the reporting of information pertaining to the issue of human rights. They also suggest that only the diversity of nations represented intheboard of directors is signi ﬁ cant inexplainingthis type of reporting. Research limitations/implications – The sample includes only large companies from Western Europe andtheanalysis covers only oneyear. Originality/value – To the best of the authors ’ knowledge, this study provides the ﬁ rst empirical analysis of factors in ﬂ uencing human rights reporting conducted on a multiple-country setting. It is also the ﬁ rst investigatingtheassociation between boards of directors ’ demographic diversity andsuch reporting.


Introduction
Although the human rights (HR) issue is one of the well-established issues of corporate social responsibility (CSR), the study of HR reporting is a relatively new trend in academic literature. Numerous articles on the HR issue have already been published in international journals (Buhmann et al., 2019;Wettstein, 2012;Ramasastry, 2015). The same cannot be said regarding the HR reporting literature.
When compared to studies examining other specific aspects of CSR reporting, such as the ones concerning the environment, the number of published studies specifically on HR reporting or examining separately this type of reporting is negligible. Regarding the studies specifically on HR reporting, they are relatively recent and have been conducted predominantly in a single-country setting. The scarcity of research is especially obvious in the case of studies examining the factors that influence this type of reporting. As far as we are aware, only Probohudono et al. (2015) and Sahari and Hamzah (2014) used multivariate analysis to investigate some factors influencing such reporting.
Our motivation is thus to contribute to the examination of factors influencing HR reporting. To the best of the authors' knowledge, our study is the first examining such factors conducted on a multiple-country setting. Moreover, it is also the first investigating the association between boards of directors' demographic diversity and HR reporting. Our focus is on the "three main factors" (Jain and Jamali, 2016, p. 264) of such diversity: directors' age, gender and nationality/ethnicity.
We have defined a very specific purpose for our study that has been motivated by the findings of Kaspereit et al. (2016). Based on a sample of listed companies from the USA, these researchers examined the association between gender diversity and a number of CSR dimensions. The only dimensions regarding which these researchers found no statistically significant associations were human rights and product-related CSR. Given that we already have the evidence for the case of the USA, we focus on Western European countries with the aim at extending the insights offered by Kaspereit et al.'s study to the case of these countries, while focusing on HR reporting and examining also the associations concerning other diversity-related variables. By focusing on Western Europe, we also ensure a level of homogeneity of the sample, in view of the adoption by the European Union of a Directive (2014/95/EU) making non-financial reporting mandatory for the largest European firms (Monciardini et al., 2020). Our purpose is to examine whether such a lack of association between diversity on the board of directors and human rights related CSR practices (in this case, reporting practices) is also detected in the case of Western Europe.
Based on a sample of large Western European companies, ordinal logistic regression analysis is used to analyze the association between these factors and HR reporting. Grounded on resource dependence theory, we hypothesize that greater gender, age and nationality diversities will translate into enhanced levels of HR reporting. Our findings suggest that the companies in the sample attribute little importance to the reporting of information pertaining to the issue of HR. They also suggest that only the diversity of nations represented in the board of directors is significant in explaining this type of reporting.
This study contributes to the literature on the determinants of HR reporting as well as to the literature on the determinants of the different dimensions of CSR. We extend the findings of Kaspereit et al. (2016) by providing evidence of a statistically non-significant association between gender diversity and the HR CSR dimension also in the case of large Western European companies. Based on our results, it seems that Kaspereit et al. (2016, p. 50) claim "that board gender diversity matters, but not with respect to all board decisions" holds also for the case of Western Europe, and that HR is one of the CSR dimensions regarding which gender diversity is not a relevant determinant.

Prior literature 2.1 Human rights reporting
What is common to most studies examining HR reporting is the finding that such reporting is not highly developed. Most of the studies conducted in a developed country setting (Chiu and Wang, 2015;Islam and Jain, 2013;Islam et al., 2017;Morhardt, 2009;Parsa et al., 2018;Preuss and Brown, 2012), as well as the studies examining these practices in emerging countries (Hoang et al., 2018;Probohudono et al., 2015;Sahari and Hamzah, 2014;Salawati et al., 2012;Sobhani et al., 2012), reached the same conclusion. Only Islam and McPhail (2011), who examined the disclosure of information on the adoption of the International PRR Labour Organisation's HR standards by 18 major multinational clothing and retail companies, report the existence of a reasonable level of development of such reporting. Morhardt (2009) examined compliance with the HR performance indicators proposed in the G3 version of the Global Reporting Initiative Sustainability Reporting Guidelines in 100 companies, and concluded that "the lack of compliance is striking" (p. 151). In an examination of the attention given to HR issues by 98 of the FTSE 100 constituent companies, Preuss and Brown (2012) found that 42.8% of the companies do not seem to address such issues at all. Regarding those that address these issues, Preuss and Brown (2012, p. 297) found that the content of the HR policies was "rather shallow". Using a sample of the 201 largest international companies that prepared their reports in accordance with the GRI's G4 Sustainability Reporting Guidelines in 2015, Cubilla-Montilla et al. (2019) found that "75% of the indicators on human rights are disclosed only by 30%À40% of the largest companies worldwide" (p. 949). Parsa et al. (2018) examined the case of a sample of transnational companies from Forbes 250, and consider that the hesitance of the companies examined "to report on procedures used to identify breaches of human rights issues" provides evidence of "how little attention they are paying to systematically approaching human rights aspects of their external workforces" (p. 56).
Based on an examination of workplace HR reporting of 18 Australian retail and garment manufacturing companies, Islam and Jain (2013, p. 102) depict it as "poor". In their examination of the quality of social disclosure by Taiwanese listed companies, Chiu and Wang (2015) report an unsatisfactory overall disclosure quality. Using a sample of Australian mineral companies, Islam et al. (2017) compared HR reporting by those operating in high HR risk countries with those operating in low-risk countries. These researchers found that although the former provide more HR disclosures than the latter, "the average number of items disclosed by mineral companies is not very impressive" (p. 45). Hoang et al. (2018, p. 844) found that of 133 Vietnamese firms in their sample, only 13.5% disclosed the HR aspect. Sobhani et al. (2012, p. 81) report evidence of HR issues having "received poor attention" by the listed banks from Bangladesh in their sustainability disclosure in annual reports and websites. Probohudono et al. (2015), Sahari and Hamzah (2014) and Salawati et al. (2012) also examined HR reporting in developing countries. While Probohudono et al. (2015) studied the case of listed companies in Indonesia, Sahari and Hamzah (2014) and Salawati et al. (2012) investigated the case of construction companies listed on Bursa Malaysia. Probohudono et al. (2015) examined the relationship between some factors (institutional ownership, public ownership, foreign ownership and managerial ownership), and found that only foreign ownership was significantly (positively) associated with HR reporting. Of the variables they examined (leverage, international operations, government ownership, size and profitability), Sahari and Hamzah (2014) found that only size was significantly (positively) associated with HR reporting. We intend to contribute to this literature by examining the association between HR reporting and the three main factors of boards of directors' demographic diversity: directors' age, gender and nationality/ethnicity.

Board diversity and corporate social responsibility reporting
There is a wealth of empirical literature on the relation between board diversity and CSR reporting. Table 1  The primary board diversity variable that has been found consistently positively associated with CSR reporting is gender diversity. Of the 18 papers reviewed, only 7 found  (Ahmad et al., 2018;Giannarakis, 2014aGiannarakis, , 2014bManita et al., 2018;Ntim and Soobaroyen, 2013) or a negative relationship (Cucari et al., 2018;Muttakin et al., 2015). Regarding the other two board diversity variables examined in this study (nationality and age), the results are mixed. Of the six studies examining the relationship between CSR reporting and nationality diversity, only three found a positive relationship (Khan et al., 2019;Muttakin et al., 2015;Ntim and Soobaroyen, 2013). Barako and Brown (2008) found no relationship, and Katmon et al. (2019) and Liao et al. (2018) found a negative relationship. Of the six studies analyzing the association between age diversity and CSR reporting (Cucari et al., 2018;Giannarakis, 2014aGiannarakis, , 2014bKatmon et al., 2019;Khan et al., 2019;Ntim and Soobaroyen, 2013), only one found a negative association between these variables (Katmon et al., 2019). Ntim and Soobaroyen (2013) found a positive association. The other studies found no association between CSR reporting and age.

Development of hypotheses
This study is grounded on the resource dependence theory. According to the proponents of this theoretical framework, directors are seen as providing to the company or securing for the company essential resources "through linkages to the external environment" (Hillman et al., 2002, p. 748). Such resources include expertise in a number of areas, diverse perspectives, ties to stakeholders and legitimacy (Hillman et al., 2002). According to this theoretical lens, one should expect directors to be "chosen based on a desire to widen the resource base provided by the board" (p. 749). Hoang et al. (2018, p. 836) suggest that board demographic diversity is "a necessary resource for a firm to realize and respond to the impact of its activities on the environment". Such diversity in the board of directors can facilitate several of the important functions it serves (Ali et al., 2014). Ali et al. (2014, p. 499) also emphasize the reduction of "uncertainties and dependencies" that organizations are likely to achieve "if they capitalize on the full range of connections delivered by a diverse board".
More diverse boards are "more likely to represent diverse stakeholders, which should lead to better CSR" (Cucari et al., 2018, p. 261). Directors' diversity tends to support the external legitimacy of an organization and to improve relationships between the organization and stakeholders, and therefore, diversity in boards is likely to have a positive impact on the disclosure of social responsibility information (Hoang et al., 2018). Ali et al. (2014, p. 499) refer to gender diversity as an important attribute of the board in view of the "different skills, knowledge, and perspectives" possessed by men and women. Rao and Tilt (2016a, p. 334) mention that most studies examining the issue of gender differences point to the existence of important differences in "values, perceptions and beliefs between men and women in general". Moreover, there is strong evidence suggesting that "female directors may perceive community or stakeholders' interests, particularly CSR issues, differently than male directors" (ibid.). According to Bear et al. (2010, p. 217), the strengths that increased participation of women can bring to boards of directors include "increased sensitivity" and "participative decision-making styles". These researchers further argue that these aspects are likely to lead to CSR strength ratings:

Gender diversity
H1. The proportion of women on the board is positively associated with human resources reporting.
Board demographic diversity and human rights

Age diversity
The coexistence of different generations on corporate boards is likely to bring diversity in terms of worldviews, values, cultural norms and perspectives (Cucari et al., 2018;Katmon et al., 2019). Age diversity is likely to be an important attribute of the board in view of the complementary attributes of younger and older directors (Ali et al., 2014). While "older directors bring to the board valuable experience that they accumulated in the industry", their younger counterparts tend to present higher familiarity with new technologies as well as to be "highly educated" (p. 498). Katmon et al. (2019, p. 455) argue that "for a successful board a mixture of different ages of directors is desirable to disseminate knowledge and experience from the senior group to the younger group of directors that could contribute to robust decision making": H2. Age diversity is positively associated with human rights reporting. Muttakin et al. (2015) found that the presence of foreigners on the board has a significant positive influence on levels of disclosure of social responsibility in Bangladesh. They interpret these results as suggesting that foreign directors, because of their international knowledge and exposure, are more committed to disclosure transparency and reducing information asymmetry. Khan et al. (2019Khan et al. ( , p. 1373 suggest that multinational directorship has a positive impact on accountability, because of their "global exposure, experience, skills, and knowledge":

Nationality diversity
H3. Nationality diversity on the board is positively associated with human rights reporting.

Research design 4.1 Sample
Our original sample consists of the 111 companies from Western Europe included in the 2018 edition of the CSR-Sustainability Monitor (CSR-S Monitor) (Weissman Center for International Business, 2018). The CSR-S Monitor is a scoring framework developed by researchers from the Weissman Center for International Business, at the City University of New York's Baruch College. This scoring framework has already been used in a number of recent studies (Demir and Min, 2019;Ferns et al., 2008;Sethi et al., 2017aSethi et al., , 2017b. It is presented in some detail in Sethi et al. (2017c). The final sample includes 100 of the 111 companies for which we were able to obtain all the necessary data. Hence, our sample consists of 100 observations (one year; 100 companies). Germany is the country in which a larger percentage of companies in the sample have the headquarters located (23%) ( Table 2). It is followed by the UK and Ireland with 21% (19 companies from the UK, and 2 from Ireland) and France with 16%.

Variables
Our dependent variable is based on the CSR-S Monitor Human Rights Contextual element as provided by Weissman Center for International Business (2018). This contextual element is defined as measuring "the quality of information provided by the company about their management commitment and effectiveness regarding their impacts on local communities and the rights of indigenous peoples, support for any controversial regimes, and their commitment to protecting freedom of expression and preventing censorship" (Sethi et al., 2017b, p. 78). It is important to note that "human rights topics tied to labor issues are covered under the Labor Relations and/or Supply Chain Contextual Elements" (p. 78). Examples of "illustrative subelements" are (p. 78) community impact; indigenous peoples' relations; support for controversial regimes; freedom of expression/censorship; and discussion of publicized cases of human rights violations.
A total of 23 companies presented a score of 0. In addition, there was a high level of concentration in a few other scores (one score with a frequency of 12; one with a frequency of 10; and two other scores with frequencies of 8). Following Branco et al. (2019), because of these characteristics of the original, continuous, dependent variable, we considered appropriate to transform it into an ordinal variable. This transformation makes it possible to conduct a multivariate ordinal logistic regression analysis to test the hypotheses developed in Section 2. Logistic regression does not rely on the assumptions of normality of distribution of independent variables, linearity of association between them or equality of variances within each group (Mertler and Reinhart, 2016). We have used SPSS's Visual Binning instrument to conduct the transformation. The transformed dependent variable has the four categories presented in Panel A of Table 3.
Independent variables are presented in Panel B of Table 3. The main independent variables used in the model are gender diversity (DIVGEN), measured as the percentage of women in the board (Cucari et al., 2018;Muttakin et al., 2015); the diversity of nationalities in the board (DIVNATION), measured as the number of foreign directors on the board relative to total board size (Khan et al., 2019;Muttakin et al., 2015); and the diversity of ages on the board (DIVAGE), measured as the standard deviation of board members' ages divided by their mean age (Ali et al., 2014;Katmon et al., 2019).
Several control variables to capture other factors that are likely to influence HR reporting are also considered: industry, size, profitability, the country of headquarters and having branches in countries with high HR risk. Regarding industry, following Sethi et al. (2017b, p. 65), we use a dummy variable that takes the value 1 for companies that operate in industries considered to have a high "exposure to environmental and social risks and impacts" (financial, manufacturing, oil and gas and utilities industries), and 0 otherwise. Islam et al. (2017) found that Australian mineral companies with operations in high HR risk countries disclose more HR performance information than their counterparts operating in low-risk countries. Based on these findings, we include as a control variable the company having operations in high-risk countries.

Model
To examine the hypotheses developed above, we used an ordinal regression model. The model used is as follows:   Table 4. In total, 23% of the companies in the sample have a score of 0 in terms of the HR contextual element of the CSR-S Monitor. Only 9% of the companies present a score of over 81% of the HR contextual element; 71% of the companies operate in industries with a high exposure to environmental and social risks and impacts; and 47% of the companies have at least one branch in high HR risk countries. The average percentage of women on the board is 25%, and the average percentage of foreign directors on the board is 44%.

Bivariate analysis
Pearson correlations are provided in Table 5. At the bivariate analysis level, none of the independent continuous variables seem to be significantly associated with HR reporting. It PRR is important to note the significant associations between DIVNT and SIZE, and between SIZE and PROFIT. The data in Table 5 indicate that the correlation coefficients between each pair of the independent variables are not high, suggesting that multicollinearity is unlikely to be a problem.

Multivariate analysis
Ordinal regression analysis is used to test the hypotheses presented above. The results of the regression analysis are presented in Table 6. The regression model provides 28.2% predictive power using the Nagelkerke R 2 approximation. The parallelism test for the parameter estimates for the ordinal regression model used was not significant (x 2 = 34.251; sig. 0.193). This result indicates that the assumption of parallelism is plausible for the model. With the exception of IND, SIZE and DIVAGE, the coefficients of all variables are in the hypothesized directions. Of these, none is statistically significant. The variables DIVNAT and country are statistically significant at the 0.05 level. Companies that have higher diversity of nationalities in the board seem to present higher levels of HR reporting than their counterparts. H2 is supported. When compared to companies that have their headquarters in the UK or Ireland, companies from Germany, Italy, Spain and Switzerland present higher scores in the HR contextual element. Contrary to our expectation, companies Board demographic diversity and human rights with higher proportion of women on the board and with higher levels of age diversity in the board do not seem to present higher scores in the HR contextual element. We do not find support for H1 and H3. It is, however, important to note that while the positive sign of the variable DIVGEN is in accordance with the expectation, the negative sign of the variable DIVAGE is not.

Additional analysis
Given that our study focuses on the relationships between gender, nationality and age diversities with HR reporting, we conducted additional analyses related to the measurement of these three independent variables. We used alternative measures of the three types of diversity considered. In the case of nationality diversity, we used as alternative measure the number of foreign directors on the board. We have rerun the model using this latter variable instead of the percentage of foreign directors on the board. The results of this analysis are qualitatively consistent with the results reported above.
Regarding gender diversity, we used three alternative measures taking into account the presence of women on the board. The first of these variables is the number of women on the board (Al-Shaer and Zaman, 2016;Bear et al., 2010). The second is a dummy variable capturing the presence of women on the board (1 in case at least one woman is present, and 0 otherwise) (Ben-Amar et al., 2017;Liao et al., 2018;Manita et al., 2018). A total of 89% of the companies in the sample have at least one woman on the board. The third alternative measure is also a dummy variable that distinguishes between companies that have three or more women on the board from those that have less than three (1 in case at least three women are present, and 0 . A total of 60% of the companies in the sample have three or more women on the board. This latter variable is associated with the so-called "critical mass theory" (Fernandez-Feijoo et al., 2014;Jain and Jamali, 2016;Post et al., 2011;Torchia et al., 2011). According to this theory, the mere presence of women on the board is not enough to exert any influence, but "when the minority group reaches critical mass, a qualitative change will take place in the nature of group interactions" (Torchia et al., 2011, p. 302). A number of studies provide evidence of this critical mass to translate into a number of three women in the board (Ahmad et al., 2018;Post et al., 2011;Torchia et al., 2011). As above, we have rerun the model with these variables instead of the proportion of women on the board, and once again, the results are qualitatively consistent with the results reported above.
Regarding age diversity, we considered as alternative measure the average age of the board as a measure of age diversity (Arayssi et al., 2016;Giannarakis, 2014aGiannarakis, , 2014bCucari et al., 2018). We have rerun the model with this variable instead of the standard deviation of the ages of the board, and once again the results are qualitatively consistent with the results reported above.

Discussion and concluding remarks
The first overall conclusion of this study is that human rights reporting practices of the companies included in the sample are not highly developed. This study points to the conclusion that large Western European companies attribute little importance to the reporting of information pertaining to the issue of human rights. A total of 23% of the companies in the sample presented a HR information score of 0, and only 9% presented scores above 81%. The levels of this type of reporting presented by the companies in our sample are very low, with the majority of the companies with scores above 0 presenting scores below 50%. These results are consistent with the existing empirical literature on HR reporting reviewed in Section 2.
We found the companies with higher levels of nationality diversity on the board are more likely to present higher scores than their counterparts. These findings are not consistent with those of Barako and Brown (2008), who examined CSR reporting in the Kenyan banking sector, and found no significant relationship between the presence of foreign directors on the board and CSR reporting. They are also not consistent with the findings of Katmon et al. (2019), who, based on a sample of 200 listed firms in Bursa Malaysia, found a negative relationship between the quality of CSR reporting and nationality diversity. While the lack of meaningful relationship may be a feature of the banking industry, Barako and Brown (2008, p. 320) suggest that foreign directors often represent "the interests of foreign owners" and therefore their presence on the board "may act as a substitute for increased disclosure". Katmon et al. (2019, p. 464), for their part, suggest that the negative relationship they found may be related to the fact that the presence of foreigners on the board of directors may play a role more related with protecting the interests of shareholders and acting in the direction of reducing the importance of CSR reporting. They also argue that nationality diversity can be problematic from the point of view of communication because of the "language barrier" and the "cultural factor" (p. 464). In this regard, they refer to the fact that Malaysia is an Asian country characterized by values very different from those of Western countries (p. 464).
However, our findings are consistent with those of Khan et al. (2019), who found a positive relation between the presence of foreign directors on the board and CSR reporting in the case of Pakistani nonfinancial firms. They are also consistent with the findings of Muttakin et al. (2015), who used a sample of listed Bangladeshi non-financial companies, and found that a higher percentage of foreign directors on the board is associated with increased Board demographic diversity and human rights CSR disclosures. They are also consistent with the findings of Ntim and Soobaroyen (2013), who found that corporations with a higher board nationality diversity presented a higher level of black empowerment reporting. In any of these cases, in particular those of Pakistan and Bangladesh, the arguments presented above as justifying the absence of a relationship or a negative relationship could apply. This diversity of results points to the need, especially in the case of emerging countries, for the cultural context to be analyzed in some detail.
Given that we used a sample of large multinational companies from Western European developed countries, the similarity between the settings in which our study and the studies mentioned above were developed is low. However, our findings regarding the levels of HR reporting from countries with headquarters in different countries (companies from Anglo-Saxon countries present lower levels of reporting than companies from Germany, Italy, Spain, and Switzerland) also point to the need to examine thoroughly the relation between HR reporting and national culture. For example, Haniffa and Cooke (2005, p. 394) underline the need to place legitimacy "in its environmental context to avoid the problem of ethnocentricity which can occur when a country such as Malaysia is evaluated by the norms prevalent in an Anglo-Saxon culture". Referring to the case of CSR reporting, these researchers argue that "in an Anglo-Saxon society it may well be that demand for legitimacy comes from groups such as consumers whereas in Malaysia that demand comes from government and certain elite groups" (p. 394).
Contrary to our expectations, the findings concerning gender diversity are not consistent with the wealth of existing empirical studies that found CSR reporting to be positively influenced by this type of diversity, both in terms of quantity and quality (Al-Shaer and Zaman, 2016;Arayssi et al., 2016;Barako and Brown, 2008;Ben-Amar et al., 2017;García-S anchez et al., 2019;Hoang et al., 2018;Katmon et al., 2019;Khan et al., 2019;Rao and Tilt, 2016b;Rao et al., 2012). None of the alternative measures we used as proxies for gender diversity is significantly associated with HR reporting.
There are, however, some studies reporting findings similar to ours. Examining black economic empowerment reporting in South Africa, Ntim and Soobaroyen (2013) found a non-significant relationship between gender diversity and such reporting. In his examination of CSR reporting from samples of, respectively, 366 and 100 companies from the Fortune 500 list, Giannarakis (2014aGiannarakis ( , 2014b find no significant relationship between the percentage of women on the board and such reporting. Ahmad et al. (2018) also found the proportion of women on the boards of Malaysian listed companies not to be associated with CSR reporting. Although not testing the hypothesis, these researchers considered that such a finding could "potentially be associated with the critical mass theory" (p. 101). Based on a sample of 379 constituents of the Standard and Poor's 500 Index, Manita et al. (2018) found non-significant relations between ESG disclosure and several measures of gender diversity (proportion of women on the board, at least one woman on the board, at least two women on the board and at least three women on the board).
It is worthy of note that there are a few studies presenting evidence of a negative relationship between CSR reporting and gender diversity. In their examination of the relationship between environmental, social and governance disclosure and diversity of the board of directors in Italian listed companies, Cucari et al. (2018) found a negative association in the case of gender diversity, but only at the 0.1 level. Muttakin et al. (2015) found a negative association between gender diversity and CSR disclosure in Bangladesh.
Although we do not find support for a significant positive relationship between gender diversity and HR reporting, our findings lead us to a conclusion that gender diversity does not discourage such reporting. This is in accordance with Jain and Jamali's (2016, p. 264) contention that "gender diverse boards do not discourage CSR". Notwithstanding, our PRR findings do not allow us to argue that our result could be "explained by invoking the critical mass theory", as Jain and Jamali argue regarding neutral results concerning the relationship between gender diversity and CSR. We cannot interpret our findings concerning gender diversity as Manita et al. (2018) interpreted the results of their study. These researchers concluded that their findings confirmed in some way the hypothesis of the existence of a critical mass effect partly because their sample included very few companies with boards, including three or more women. That is not the case with our sample.
There is at least one study examining board gender diversity and CSR that, albeit not examining CSR reporting, is relevant to the discussion of our findings, least of all because we motivated our purpose on the findings of this study. In their examination of the association between board gender diversity and a number of CSR dimensions, Kaspereit et al. (2016) found no significant relationship concerning the human rights and productrelated CSR dimensions. They found a positive association concerning all the other CSR dimensions they examined (community, diversity, employees and environment). They offer as a possible interpretation for these findings "the fact that larger firms are more international with operations in countries where human rights are a critical issue. A large textile producer that manufactures in Asia, for instance, is more likely to encounter human rights-and product-related controversies than a smaller local manufacturer" (Kaspereit et al., 2016, p. 58). It may be the case that, as Blanc et al. (2019) argue regarding the case of anti-corruption reporting, the scarcity of disclosure on HR issues may be related to the lack of interest companies have in discussing them for a fear that addressing these sensitive issues could create suspicions that problems concerning them do exist or draw attention to the existing problems.
Our findings concerning age diversity are consistent with Giannarakis (2014a, 2014b), Cucari et al. (2018) and Khan et al. (2019), who found a non-significant relationship between the average age of the board and CSR reporting. Rao and Tilt (2016a, p. 333) depict age diversity as an "emerging diversity characteristic which is gaining attention in the CSR literature", albeit "no solid argument favouring one age group when it comes to board processes" exists.
However, they are not consistent with the results of Katmon et al. (2019), who found a negative relationship between age diversity and the quality of social responsibility reporting. However, these authors were not surprised by this finding, stating that "traditional Malaysian culture, the older tend to undermine or less appreciate the opinion of the people with younger age" (p. 462). In Malaysia, the ideas/opinions/suggestions/recommendations of younger generations are generally viewed by older generations as less mature and less valuable because of the lack of experience of younger generations. Hafsi and Turgut (2013, p. 466) also found a significant negative relationship between board age diversity and US corporate social performance. Hafsi and Turgut (2013, p. 473) point out as a possible explanation the possibility of age diversity leading to "polarization, a sort of generation conflict". They also added that "some evidence confirms that there is a level of diversity beyond which it may become unwieldy" (p. 473). They adduce "the possibility of age diversity being low everywhere and the few firms that have higher diversity go through difficult reconciliation across age groups, which leaves little space for the contentious issues of corporate social responsibility" (p. 474).
We used a sample of large multinational companies from Western European developed countries presenting cultural characteristics very different from those of Malaysia, and, albeit in a lower degree, from those of the USA. It may be the case that the difficulty of reconciliation across age groups to which Hafsi and Turgut (2013) refer, and which is implicit in Katmon et al.'s (2019) discussion, is not as problematic in Western European Board demographic diversity and human rights countries as it is in Malaysia or possibly in the USA. These results also point to HR rights reporting.
It is interesting to note the similarity between the study of Ntim and Soobaroyen (2013) and our own in that both examined a very specific and sensitive type of reporting. Similar to Ntim and Soobaroyen (2013), who examined black empowerment reporting in South Africa, our study reveals a positive association between the type of reporting examined and nationality diversity and age, and a non-significant association between such reporting and gender. It may be the case that such well-established relationships, as is the case of that between gender diversity and CSR reporting, do not hold when one is analyzing such sensitive issues as black empowerment reporting and HR reporting because companies fear that addressing them could make stakeholders suspicious of the existence of problems concerning them, as suggested above.
Concerning black empowerment reporting in South Africa, it may be the case that foreign directors would be more aware of the cruciality of the problem than their national counterparts, and be more open to disclosure of information on the issue. In the case of HR reporting, it may be the case that the presence of directors from a diverse range of nationalities has a similar influence on HR reporting.
The contributions of this study to the literature are twofold. First, it contributes to the literature on the determinants of HR reporting by examining the association between HR reporting and the three main factors of boards of directors' demographic diversity: age, gender and nationality/ethnicity. Second, it contributes to the literature on the determinants of the different dimensions of CSR. In particular, it extends to Western Europe Kaspereit et al.'s (2016) finding of a statistically non-significant association between gender diversity and the HR CSR dimension in the USA. This study also has some practical implications. Its findings suggest that when the HR issues are a critical one to a particular company, the best option to improve how it deals with this issue may not be by way of an increase in the proportion of women in the board. Rather, it may be preferable to consider the inclusion of additional foreigners on the board.
This study presents a number of limitations that could be addressed in future research. Further studies could examine two separate periods sufficiently far apart to encompass more than one board mandate. It would also be interesting to analyze the evolution of the disclosure of human rights information for the same companies over a period and to analyze the possible relationship of this evolution with changes in the composition and diversity of the boards. In addition, future studies could also examine the relationship between cultural aspects and HR reporting. At the same time, we also think it is important to extend this type of study to larger samples, including non-European companies, and to check whether the introduction of companies based in other geographical areas, notably the USA, changes the results or corroborates the findings.