THE ROLE OF HUMAN AND SOCIAL BOARD CAPITAL IN DRIVING CSR

The objective of this paper is to analyse the effect of professional, technical and relational background (human and social capital) of outside directors on promoting firm CSR disclosure. Following the Hillman et al. (2000) taxonomy of board members, we classify outside directors as business experts, support specialists and community influentials, and examine whether business and technical expertise or political ties in the boardroom affect CSR disclosure. This study confirms that not all outside directors are equally effective in improving CSR disclosure and that only certain kinds of outside directors, those classified as support specialists, help increase it. On the other hand, our findings also show that directors with previous experience as politicians affect CSR disclosure negatively, probably due to their interests in safeguarding their reputation within the company against public scrutiny and in protecting their political connections. In addition, our set of analysis with interaction effects reveals that powerful CEOs have the incentive to promote CSR-related strategies and to press business expert and support specialist directors to enhance profitable sustainability strategies and transparency in its disclosure. Nevertheless, powerful CEO effect is not enough to compensate the negative role of political directors on CSR reporting. Therefore, this paper supports the theories that appeal for analysing the multiple configurations of corporate governance mechanisms by adopting a “holistic approach” and the need to combine them in order to analyse their impact on CSR behaviour.


INTRODUCTION
Through Corporate Social Responsibility (CSR) disclosure companies improve corporate transparency by reporting not only financial information, but also their social and environmental performances to stakeholders and society (e.g., Aribi and Gao, 2010).
CSR disclosure may allow firms to develop and enhance their corporate image and to provide useful information for investment and non-investment decisions (e.g., Deegan and Blomquist, 2006).In this way, companies whose boards support CSR reporting will show a more social orientation and a more engagement with stakeholders and society by satisfying their needs and demands of receiving social and environmental information.
The analysis of the association between the professional background of board members and Corporate Social Responsibility (CSR) disclosure is highly relevant in a world of continual governance scandals, failures, opacity, and social and environmental excesses (Jain and Jamali, 2016).According to Michelon and Parbonetti (2012), because disclosure policies emanate from boards of directors, sustainability disclosure can be conditioned by board attributes.Therefore, previous literature has evaluated the effect of board composition on CSR ratings, studying the influence of board structure characteristics such as independence, interlocking or gender diversity (Cuadrado-Ballesteros, Martínez-Ferrero and García-Sánchez, 2017;Shaukat et al., 2016;Bear et al., 2010).A significant limitation of this literature is that these papers adopt a traditional and short-sighted perspective of board composition and assume homogeneity in skills and abilities in the boardroom, which leads to an incomplete analysis of the level of expertise that actually exists on the board and its consequences (Gray and Nowland, 2013).This paper advances this stream of literature by analysing the effect of professional, technical and relational background on promoting firm CSR disclosure.These resources represent the board capital (human and social) as they refer to knowledge, skills, networks and ties (Hillman and Dalziel, 2003).We use the board director classification of Hillman et al. (2000), who develop board capital and classify board outside directors as business experts, support specialists, and community influentials.These dimensions, previously used by other authors (Markarian and Parbonetti, 2007;Jones, Makri and Gomez-Mejia, 2008;Bear, Rahman and Post, 2010;Haynes and Hillman, 2010), allow us to examine a comprehensive and broad range of board characteristics.We exploit this literature by arguing that boards with a higher proportion of business experts and support specialists are expected to be more effective in encouraging CSR reporting.Into the broad Hillman category of community influential, we differentiate directors with political connections and expect a negative influence on CSR.As an analysis extension, we study whether the effect of board composition on CSR reporting is moderated by powerful CEOs.In this vein, we follow Jain and Jamali (2016), who state that scholars are required to adopt a holistic approach where CEO power and board interact to form bundles that in turn influence CSR outcomes.
We examine a sample of Spanish listed firms during the 2008-2014 period.Spain is an interesting country to analyse board composition due to several reasons.First, the recent Good Governance Code of Spanish Listed Companies (2015) recommends that firms should have a diverse board in skills and background and states that the director selection policy should look for a balance of knowledge and experience in boardrooms.
Second, in contrast to the Anglo-Saxon capital markets, the board of directors in Spain is the prevalent mechanism of control (García-Meca et al., 2015).Finally, in Spanish boardrooms, it is noticeable the high number of directors with political connections, mainly explained by the number of privatizations made in Spain during the last decades and the high ownership concentration (Bona-Sánchez et al., 2014).Previous papers have examined the influence of boards in CSR disclosure of Spanish firms (e.g.García-Meca and Pucheta, 2017;Cabeza-García et al., 2017), but none of them has used the Hillman taxonomy to study how the human and social capital of directors can contribute to CSR disclosure.
The paper contributes to the literature focused on directors from the perspective of the resource dependence theory, by supporting the assumptions that boards provide critical resources to the firm, including professional (business experts), technical (support specialists) and relational (community influential) background as a way to influence firm CSR disclosure.Therefore, we contribute to the emerging stream of behavioural governance by noting that individual differences in personal styles, skills, ties and business knowledge between business experts, support specialists and community influentials can lead them to make different choices that affect corporate decisions (Hambrick and Mason, 1984).This paper is also the first empirical paper which tests the "board human and social capital" effect on promoting CSR disclosure, providing an answer to the recent call for further research on examining the knowledge and experiential diversity of boards which can significantly impact CSR outcomes (Jain and Jamali, 2016).
According to the recent revision of governance studies published by Filatotchev and Wright (2017), studies that analyse board composition are usually too simplistic and fail to examine the human and social capital of boards that is important for both monitoring and adding value.Finally, this is the first paper to test the connections between different board member categories, CSR reporting and CEO power.Our results are consistent with the trend in literature that suggests the interdependency among corporate governance mechanisms.

Theoretical framework
Drawing on resource dependence perspective, prior research (Pfeffer and Salancik, 1978;Hillman et al., 2000;Hillman and Dalziel, 2003) explores the effect of several board attributes on CSR reporting.A complete revision of past literature examining the association between board characteristics and CSR activities is done by Jain and Jamali (2016).They note that among boardroom attributes, human and social capital is crucial to improve CSR engagement (e.g., Wincent et al., 2010;Tian et al., 2011).According to Hillman and Dalziel (2003) and Haynes and Hillman (2010), board members may become providers of these two kinds of capital.Thus, when board directors provide skills and knowledge to boards, they will be allocating human capital, while that board directors will allocate social capital when providing abilities to obtain resources and maintain relationships with external environments or organizations.In this regard, Russo and Fouts (1997) also highlight "the importance of nurturing and building resources through sustained actions for creating and maintaining a pro-environmental internal capabilities and external reputation" and Hart (1995) states that for companies willing to maintain their competitive advantages concerning social and environmental matters, "it is important to consistently build upon their internal human and organizational competencies and resources, as these may otherwise erode over time as competitors catch up".
Resource dependence approach also supports the view that boards become a relevant mechanism for providing outside resources and for maintaining external relations with influential's organizations.In this regard, board of directors is an important resource for handling firm external needs like environmental and social challenges.Board members, according to resource dependence theory, become more effective when providing resources to companies like abilities, knowledge, experience, advice, legitimacy, reputation, assistance in achieving commitment and external links between firms and significant stakeholders or other relevant organizations (Pfeffer and Salancik, 1978), which may affect the role played by directors to monitor managers, to advise firms and to make decisions.Boards with this human and social capital are more likely to develop their tasks and duties in a more efficient way.Under this perspective, a board made up by directors with diversity of links, experience, abilities and knowledge and with a wider stakeholder perspective may affect positively CSR strategic decisions such as the reporting of CSR matters (e.g., Wang and Dwhirst, 1992).According to Dass et al. (2014), resource dependence approach also suggests the relevant advice role that boards might perform when their members possess reputation, expertise in different areas and influences.Thus, board structure will be shaped according to firm's needs for relevant external ties (e.g., Pfeffer and Salancik, 1978).Individual board members provide boards individual experience, different contributions and other firms expertise (Baysinger and Butler, 1985).Furthermore, outsiders, apart from supervising managers, also bring experience and resources for handling matters coming from outside such as ties with external firms, helpful experience and advise, legitimacy and support or commitment from external agents (e.g., Pfeffer and Salancik, 1978).
In order to survive and gain competitive advantages toward their competitors, firms should appoint board members who are able to satisfy demands from influent and relevant externals organizations.Diverse directorships have benefits for companies in terms of enriching experience and reinforcing knowledge and business links with outsiders (e.g., Harris and Shimizu, 2004).Terjesen et al. (2016) also claim that human capital (individual expertise, abilities or background) on boards might improve the decision-making process of firms because there will be a higher flow of information between board directors.Consequently, the connections with outside organizations and stakeholders can be strengthened, encouraging the reporting of CSR matters.

Hypotheses development
Prior expertise and background of board of directors become equal or more relevant than board independence or non-CEO duality (Gul and Leung, 2004).In this respect, Hillman et al. (2000) give a step and classify external directors into three groups: business experts, support specialists and community influentials.
According to Hillman et al. (2000), business experts are those board members who have knowledge based on their previous experience as executives of other firms.They provide abilities, skills and knowledge acquired previously in other organisations as insiders or executives.This past expertise in several firm's areas when these directors were developing executives tasks in other companies will have positive effects on firms, particularly on the decision-making process.Additionally, business experts can contribute with new perspectives coming from other environments concerning internal matters and markets as well as, according to Zahra and Pearce (1989) and Bear et al. (2010), being able to provide useful resources for firm activities and obtaining relevant connections with external agents such as boards and other significant stakeholders from crucial companies.Therefore, business experts will be able to bring human and social capital to boards, useful for all board members when non-financial and non-business decisions, such as CSR reporting, have to be made.This human and social capital comes from the education, knowledge and expertise acquired by these directors outside of the company (e.g., Wincent et al. 2010).Board members with these qualities will have a higher ability for processing information, for being involved with companies' strategies, for creating networks, for being more receptive to innovation and for supporting effective decision-making (Wiersema and Bantel, 1992;Goll et al., 2007;Wincent et al., 2009).
Thus, board directors with these characteristics will be more likely to develop their roles of service, control and provision of resources with positive effects on companies.
The appointment of business experts on boards may be positively appreciated by shareholders and several stakeholders, given their prior executive experience in similar sectors.Then, they may contribute to make relevant strategic decisions for their betterquality skills, because may detect threats and opportunities and because the background acquired in linked industries might bring information of quality to share with other board members (Jones et al., 2008;Dass et al. 2014;Faleye et al., 2014).This will improve board effectiveness since board members will be able to play an enhanced supervising role.
According to resource dependence theory, board members with diverse background and executive experience show a higher level of social and human capital, as well as having a positive effect on the relationships between board members and managers, which in turn may encourage CSR activities (e.g., Shropshire, 2010;Westphal, 1999).Ben Barka and Dardour (2015) show a positive association between board members with experience and skills and corporate social/environmental performance.Finkelstein et al. (2009) also argue that the educational background of executives will define their social orientation.In this regard, executives who have gained degrees relative to social and human sciences, for instance, will be more likely to engage with stakeholders' needs and, consequently, they will be more proactive in disclosing CSR issues.Business experts may also contribute towards making relevant strategic decisions due to their higher level of skills, because they may better detect threats and opportunities, and because their background acquired in related industries might bring quality information which they can share with other board members (Dass et al., 2014;Faleye et al., 2014;Jones et al., 2008).Therefore, the human and social capital supplied by business experts on boards may define the board's CSR orientation.
Prior empirical literature shows that investment in innovation (Faleye et al., 2014) or corporate value (e.g.Drobetz et al., 2014) is positively affected by business experts on boards.Furthermore, Thomas and Simerly (1995) provide evidence that managers with prior expertise and knowledge acquired in other organisations show more sensitivity towards stakeholders' needs and those CSR matters which firms may face.Gray and Nowland (2017) demonstrated that firms whose boards are composed of directors with prior business expertise in other companies and who have worked as directors on boards of different firms had a positive effect on market reaction.Additionally, Mallin and Michelon (2011), focusing on resource dependence theory, propose that external directors with a diversity of capabilities, external links, skills and advice on stakeholders' expectations can be considered as providers of social and human capital and, thereby, may improve their credibility and CSR reporting, strengthening ties with companies' stakeholders.
Thus, based on the above arguments, we posit the following hypothesis: H1: Business expert directors on boards have a positive impact on CSR reporting.
In comparison to business experts, who are regarded as decision monitors by Jones et al. (2008), among others, support specialist directors are deemed as decision supporters who bring specific expertise and knowledge to areas such as insurance, law, technology, industries and capital markets (e.g., Shaukat et al., 2016), amongst other things, which may have a positive effect on strategic activities and decisions.In this regard, Hillman et al. (2000) describe support specialists as directors serving on boards who provide expertise acquired as officers or executives in commercial and investment banks, accounting, auditing and consulting firms, law companies and advertising firms.This specialized know-how may help managers in particular matters with which they are not familiar, because support specialists will be able to advise them and contribute with specific abilities, skills and experience when making strategic decisions (Hillman et al., 2000), which might have a positive effect on the whole business deals of companies.
However, support specialists on boards do not tend to possess as much general experience in management matters (Hillman et al., 2000;Baysinger and Zardkoohi, 1986) as business experts do.Support specialists also become significant suppliers of social and human capital to boards, due to their particular knowledge and expertise in different areas of the firm and due to their external connections, which aid companies in obtaining support from outside organisations (Pfeffer and Salancik, 1978).This view is supported by Bear et al. (2010), who state that support specialists maintain ties with clients' networks, their main companies and professional organisations.The different external links and networks may result in a higher support from external agents, in a higher knowledge of the context where the firm operates and in a higher response to the demands and needs of stakeholders.As Beckman and Haunschild (2002) argue, all the networks might supply experience, counsel and relations that promote alliance and mutual aid with crucial stakeholders.
Firms might enhance the relations with their stakeholders if they voluntarily disclose CSR information and may be perceived as more socially responsible.Thus, directors (support specialists) with attributes such as particular background and connections with outside organizations that may allocate expertise, advice and support will be more likely to support decisions involving social and environmental issues in order to satisfy shareholders' and stakeholders' needs by engaging with the reporting of CSR matters.This idea is supported by Nikolova and Arsi (2017), who pose that "CSR cannot exist if individuals do not possess enough maturity and competence to act responsibly.It is up to companies to train and for society to socialize the individuals towards the development of such necessary competencies".
Authors such as Kroll et al. (2008) and Kor and Sundaramurthy (2009), among others, report that directors who possess specific expertise related to the industry where the company operates, will bring to the board solid information on the competitive environment of the company and on the way that the industry works and, consequently, their knowledge and advice will benefit managers.Furthermore, this particular and technical knowledge of the industry will allow directors to identify new industry's opportunities, to assess growth's schemes proposed by managers and to access to vital resources for their external connections, providing social and human capital to boards.
The capital of supporter specialists may play a relevant role on boards due to the wide benefits that firms might gain from it.Galbreath (2016Galbreath ( , 2009) also supports the idea that support specialists possess expertise in formulating strategies which are not related to business and markets, and Helfaya and Moussa (2017) noted that audit committee directors with financial expertise had a positive impact on environmental and sustainability disclosure of UK firms.Given that CSR activities are more associated with environmental and social matters than financial, board members who provide a unique knowledge, skills and expertise acquired in other organizations might bring improved social and human capital to boards to address CSR issues.These individuals have highly developed their human skills and, therefore, they will be more conscious of their own behaviours, beliefs and perceptions on other groups and individuals.In line with Ewert and Baker (2001), apart from formal education, academic major or specialization of board directors is involved with their general motivations, beliefs, perspectives and values.
According to these authors, the academic major affects the several levels of a person's environmental worries and attitude toward environmental issues.In this way, these directors might accept that others have different assumptions, opinions, perspectives and beliefs.Accordingly, they may be more sensitive to the motivations and the needs of others and might be willing to behave in line with taking into account the perceptions, needs and demands of others.As a consequence, they may increase their feeling in the direction of society and stakeholders' interest, which may result in a higher engagement and commitment with CSR disclosure.Thus, directors with specific abilities and backgrounds also tend to be more involved with the needs of others and to be more sensitive towards social and environmental demands and, therefore, these board members will make more efficient CSR decisions, such as CSR reporting (Knonrad et al., 2006;Rosener, 1995).
To sum up, support specialists become significant board members, especially from the beginning of the worldwide financial crisis, because shareholders, stakeholders and society, in general, claim that most of board members have scant specific background, knowledge and qualifications for serving as directors.As a result, the majority of these agents blames to the lack or poor specific technical knowledge of directors of the financial frauds, bankruptcy of firms or loss of credibility of capital markets and investors.
According to the above arguments, we predict that the specific qualifications, abilities and technical knowledge of support specialists will have a positive impact on CSR reporting.Hence, we propose the following hypothesis: H2: Support specialists on boards have a positive effect on CSR reporting.
Community influentials or community leaders play a significant role when serving as board directors because they provide support to companies through networking and also bring reputation and credibility.Most of the strategic activities and choices of companies have an effect on the community or non-business organisations, which have ties with community influentials who also possess expertise, skills, abilities and influence (Hillman et al., 2000).Among these communities or organisations, officers of communities or social organisations, or representatives of universities or other institutions like political parties can be emphasised.In this vein, many ex-politicians hold a directorship on firms.According to Li et al. (2008), the majority of community influentials are directors who have previously been politicians or renowned members in relevant communities and are therefore admired and respected and have power in nonprofit contexts.Braiotta and Sommer (1987) argue that the presence of community influentials' directors on boards tend to be high because they do not need to have technical qualifications or financial, accounting and business experience, as it would be expected for members who compose specialized board subcommittees like financial or audit committees.This lack of specific expertise, technical knowledge and background may justify the great appointment of community influentials' directors on boards.This idea is also supported by Peterson and Philpot (2007), who drawn on resource dependence theory, show that the representativeness of community influentials' directors on boards is high.These ideas suggest that community influentials are not appointed by providing resources as their background, experience or specific knowledge on certain matters, for instance, but by allocating other resources such as external links with relevant communities and organisations, which are also important for having competitive advantages, for getting significant resources and for surviving.Thus, community influentials, when holding a directorship on boards, are providers of legitimacy, reputation and credibility and have significant external connections for the significant organisations that they represent or have represented.
These outside links will allow firms to enhance the opportunities to have contacts with powerful individuals in the institutional context, which might lead to access to financial support, relevant information or other important resources.Furthermore, the external connections of community influentials might be beneficial for firms in order to survive and have success since their operations will be better accepted by external communities or organizations (Hillman et al., 2000;Hillman and Dalziel, 2003).According to Westphal (1999), community influentials on boards can give advice and guide to management team, due to their field of expertise as community leaders.
Consistent with Hillman et al. (2000) and Hillman and Keim (2001), Yekini et al. (2015) argue that companies which have to face uncertain environments will tend to appoint community influentials' directors on boards as a strategic move.In this way, nonprofit or non-business opinions and experiences will be shared among board members and boards might use the influence of these directors on several communities.Michelon and Parbonetti (2012) claim that community influential directors have a higher orientation toward social issues, due to most of them are top executives from non-profit or the military organizations, well-known academics and scientists and politicians and, as a result, they are more likely to affect positively CSR disclosure.Mallin et al. (2013) support this idea since they demonstrate the positive impact of community influential on all types of disclosure analysed, among them, CSR reporting.Hillman and Kein (2001) also show that community influentials affect positively firm value.Thus, the social and human capital, particularly social, supplied by community influentials might play a more social role, orienting boards toward social issues such as the reporting of CSR information.
On the other hand, most community leaders, particularly in the Spanish context, are characterised by maintaining political connections or by being ex-politicians.
Therefore, this might negatively influence the opinion of shareholders and stakeholders with regards to the firm, since politicians are sometimes reviled and stigmatized by society in general.These negative opinions on politicians or ex-politicians are due, in part, to corruption scandals, particularly in periods of economic crisis, and because they benefit or have benefited themselves to the detriment of citizens' interests and needs.Chen et al. (2011) claim that firms operating in institutional contexts where there have been a large number of corruption scandals and which have weak legal systems tend to appoint more community influential directors with political connections.Accordingly, community influential directors with political ties may be interested in achieving their own aims by using political resources to the detriment of shareholders' and stakeholders' needs.This leads community leaders to disclose, as Bona-Sanchez et al. (2014) suggest, less information to third parties in order to safeguard their reputation as politicians or ex-politicians within the company against public scrutiny, to protect their political connections, and also to avoid reporting competitive advantages to their rivals.As a result, the presence of community influentials on boards may have a negative effect on CSR disclosure; that is, directors with political ties will be less likely to report CSR matters in order to prevent rivals from discovering the companies' strategies (information effect) and to avoid that reported information may transmit a negative opinion about them to stakeholders and society (reputational effect).In this last case, whether firms voluntarily report CSR information, stakeholders and society will be more conscious on all their CSR practices.Stakeholders with high ethical standards, for instance, may be potential readers of this CSR information.Accordingly, when they read this information, they might perceive it negatively if its reporting has been encouraged by directors with political ties o ex-politician serving on boards, particularly if some information stresses the ethical attitude of companies, given that the ethical behaviour of these directors is questioned.Similar examples are provided by Halek and Eisenhauer (2001) and Griffin and Sun (2018), who explore the association between religious affiliations and CSR disclosure.Past research (Chaney et al., 2011;Boubakri et al., 2008) finds a negative association between political ties and corporate value.Additionally, Bona-Sanchez et al.
(2014) also show that Spanish companies with politicians on their boards reduce earnings informativeness.
Therefore, taking into account above premises, we hypothesise that the group of community influential' directors is not homogeneous concerning its impact on CSR disclosure.In this regard, we expect that CSR reporting will be affected differently by political directors and the remainder of community influentials' directors such as influential members in non-profit firms or cleric communities.So, we predict the directors with political connections will be less likely to report CSR information, while the rest of community influentials' directors will have a positive effect.Thus, we differentiate among the group of community influentials directors between directors who have political links or are ex-politicians and the remainder of community influentials directors.
To the best of our knowledge, in the Spanish context there is no past research focused on examining the association between the presence of community influentials' directors and CSR reporting, specifically splitting up the community influentials collective between political directors and the rest of community influentials.In view of that, we propose the following two hypotheses: H3a: Directors on boards who were politician or have currently political ties have a negative effect on CSR reporting.

H3b: The remainder of community influential directors on boards have a positive
effect on CSR reporting.

The moderating role of CEO power
The holistic approach highlights the need to consider the multiple relationships that exist among the different firm governance dimensions.In this sense, corporate governance scholars are required to take a holistic approach where CEO and board interact to form bundles that in turn influence CSR disclosure (Jain and Jamali, 2016).
The moderating role of CEO power is supported by Chen ( 2014), who suggests that research which aims to examine the resource dependence role of board capital in influencing R&D investment should consider the potential moderating role of CEO power.In the same vein, Haynes and Hillman (2010) also show that CEO power is an important moderator of the relationship between board capital and strategic change.This moderating effect is even more important in contexts like Spain, where there is a high power distance culture, what means that hierarchy is highly important and most of the decisions are centralized in the CEO.
Literature regarding the CEO power effect on firm outcomes is not homogeneous (Rhoades et al., 2001).The agency approach suggests that CEO power increases the concentration of managerial power, which can weaken the monitoring and advisory role of directors and, thus, reduce CSR investments and the transparency of firm information.
Therefore, powerful CEOs can make decisions that do not take into account the greater interests of stakeholders, reducing their boards' involvement in social and environmental activities as well as the disclosure of these activities (Michelon and Parbonetti, 2012).
On the other hand, another stream of literature holds that powerful CEOs have the incentive to promote CSR-related strategies and to press directors to enhance profitable sustainability strategies and transparency in its disclosure.According to this literature, the CSR engagement and promotion of CSR strategies by directors can be conditioned by the CEO incentives to perform these tasks (Hillman and Dalziel, 2003).Therefore, board capital can be more likely to promote CSR activities when there is a powerful CEO in the firm.Hence, non-monetary CEO incentives related to career concerns, reputation, entrenchment and power may have a positive influence on board attitude towards CSR decisions (Fabrizi et al., 2014;Jiraporn et al., 2015).Accordingly, powerful CEOs can reinforce the role of human and social board capital and push for more CSR disclosure, which aims to raise long-term profit and performance, as well as to reinforce their legitimacy in the eyes of the shareholders.According to Chen et al (2014), powerful CEO can also affect the nomination process and, therefore, they condition the loyalty and social obligation that directors feel when they are appointed by the current CEO.In addition, the implementation of entrenchment activities by CEOs can also increase the involvement of CSR activities to compensate for the damage suffered by shareholders as a result of self-entrenchment strategies (Prior et al., 2008).
The moderating role of powerful CEO on the association between boards and firm outcomes was already shown by Chen (2014), who found that directors with human and social capital tend to support R&D investments when powerful CEOs are present.In Spain, Godos-Díez et al. (2014) evidenced that the influence of large shareholders on CSR activities was moderated by the CEO profile, noting that Spanish large owners interested in CSR were more likely to achieve their objectives if the CEO is closer to the steward model and he/she behaved collectively in the interests of all the stakeholders.According to the above arguments, we suggest that powerful CEO and board members interact to form bundles which in turn affect CSR reporting.Therefore, we posit the following hypothesis: H4: The effect of business experts, support specialists and community influentials on CSR reporting is moderated by powerful CEOs.

METHODOLOGY
The study sample is comprised of an unbalanced panel of 152 non-financial firms listed on the Spanish stock market during the 2008-2014 period.The Global Reporting Initiative (GRI) is the most trustworthy and comprehensive reporting tool used by companies for disclosing sustainability information (Brown et al., 2009).Companies voluntarily disclose their annual GRI reports on the GRI website, so the information about CSR is collected from here as well as from the companies' websites.Financial and accounting information comes from the SABI1 database, which is comprised of general information and data from the financial statements included in the Spanish Companies Registration Office.Board characteristics are hand-collected not only from the website of the National Securities Market Commission (CNMV), but also from the corporate websites.The final sample consists of an unbalanced panel with a total of 763 observations2 .

Dependent variable: Measure of Corporate Social Responsibility
In order to find an appropriate measure for CSR disclosure, we have created an index based on the content analysis technique on the CSR reports to quantify the amount of CSR information in the reports.This technique is a method of grouping the text of the reports into several categories (Kuo et al., 2012;Prado-Lorenzo et al., 2009).Therefore, the CSR disclosure index is the aggregation of the six following items measured as dummy variables (García-Meca and Pucheta-Martínez, 2017;Prado-Lorenzo et al., 2009): (1) CSR information disclosure, which shows if companies disclose any social, environmental and economic information; (2) informal preparation of CSR report, where firms, besides using the GRI format, give information in other ways; (3) GRI format, where companies disclose the CSR information using the GRI format; (4) GRI certification, where the CSR report is certified by the GRI; (5) audited CSR report, which informs us of whether the CSR information is audited by an external and independent entity and (6) assurance of CSR report, which shows if the assurance scope refers to the entire CSR report.
When facing to binary information, as it is our case, indicators can only measure the quantity of disclosure, but they ignore the quality of the information (Leitoniene and Sapkauskiene, 2015).However, some items that we have included to create the index, such as if the CSR information is audited by an external and independent entity, allows companies to enhance the quality and reliability of the information provided and enhance the stakeholders' engagement process.

Independent and control variables
As independent variables, we use the proportion of outsiders classified as Business Expert (BE), Support Specialist (SS) and Community Influential (CI) over the board size.BE are active or retired executives from other firms, with experience in strategic decisionmaking in different companies, their main goal being to provide not only their expertise, but also their knowledge, advice and alternative viewpoints about internal concerns (Markarian and Parbonetti, 2007;Hillman et al., 2000).Outsiders classified as SS are professionals specialized in diverse fields (law, capital markets, insurance, public relations and industrial knowledge, among others) and who provide companies with their expertise and knowledge in their individual specialized field (Hillman et al., 2000).The CI group is made up of non-executive directors such as politicians, members of the clergy, and leaders of social organizations that provide networking and reputation opportunities to the company (Hillman, Cannella and Paetzold, 2000).Since it is rather common to find former politicians on the boards of the largest firms in Spain (Goldman et al., 2009, Faccio, 2006), we have divided the CI members into two groups.The CI_pol includes the proportion of politicians classified as community influential, while the CI_others is made up of the remaining CI, that means, those who are not classified as politicians.The main difference between the groups is that, unlike the BE, the members of the SS and CI groups lack general management expertise (Markarian and Parbonetti, 2007).
To avoid biased results and based largely on previous empirical literature, we include the following control variables.
Since boards are responsible for encouraging CSR disclosure (Li et al., 2010), a larger board size is more desirable because it allows board members with more opportunities to connect with external knowledge, skills and networks (Rhee and Lee, 2008), encouraging the communication of CSR information (Jizi, 2017).
Firm size (SIZE) is calculated as the logarithm of the total assets (Sotorrío and Sánchez, 2010;Martínez-Ferrero, Garcia-Sanchez and Cuadrado-Ballesteros, 2015).The company size is expected to have a significant impact on CSR reporting, due to the fact that larger companies are required to disclose more social, economic and environmental information than the smaller ones in order to satisfy stakeholders' needs (Cooke, 1991).
Moreover, since the preparation and disclosure of this information is costly, larger companies may have more resources and expertise to produce and disclose CSR information than the medium and small firms, which improve their image and reputation (Monteiro and Aibar-Guzmán, 2010).
Ownership concentration (OWNERSHIP) is measured by the percentage of shares controlled by the largest and second largest shareholders.Previous literature finds a negative relationship between the power of the largest shareholders and CSR disclosure (López-Iturriaga and López-de-Foronda, 2011; Prencipe, 2004).Those companies with a more disperse ownership suffer a higher pressure for voluntary disclosure (Cullen and Christopher, 2002) and, consequently, they are more prone to suffer an opportunistic behaviour and conflict of interests between the agent and the principal.
Ceo power (CEO_power) is represented as the sum of the following standardized variables: (1) CEO duality, (2) the ratio of directors appointed after the CEO began his tenure to the total number of directors, (3) the ratio of shares held by the CEO to director ownership (Haynes and Hillman, 2010;Chen, 2014).Previous literature suggests that CEO power can affect CSR disclosure in a positive (Michelon and Parbonetti, 2012) or negative (Hillman and Dalziel, 2003) way.While Michelon and Parbonetti (2012) consider that CEO power provokes a strong power base, which can reduce the board´s ability to exercise effective control and reduce their involvement in CSR disclosure, Hillman and Dalziel (2003) point out that powerful CEOs have incentives to promote CSR-related strategies and to press directors to enhance profitable sustainability strategies and transparency in its disclosure, stating a positive relationship.
Profitability is measured by using the return on assets (ROA) and it is calculated as the earnings before tax divided by total assets.In spite of the fact that some studies have found a negative relationship between profitability and the extent of their CRS disclosure (García-Ayuso and Larrinaga, 2003;Brammer and Pavelin, 2008;Stanny and Ely, 2008), most of previous literature confirms a positive relationship between them (Wallace et al., 1994;Wallace and Naser, 1995;Raffournier, 1995).
Leverage (LEV) is measured as total debt over assets (Michelon and Parbonetti, 2012).Previous empirical evidence is not conclusive.On the one hand, some studies find a positive relationship between the reporting of CSR matters and highly leveraged firms (Xiao et al., 2004;García-Meca and Pucheta-Martínez, 2017;Clarkson et al. 2008), because these companies are more prone to disclose voluntary information in order to reduce their agency costs and, therefore, their cost of capital.On the other hand, some studies show a negative relationship between both variables by arguing that more leveraged companies have fewer opportunities to allocate funds for CSR activities (Jizi, 2017).Other studies, have not found a statistically significant relation between the previous variables (Gul and Leung, 2004).
In order to measure the level or quality of firms' governance, we calculate an aggregate index that incorporates several characteristics of the functioning and structure of the board of directors (Lara, Osma and Penalva, 2007).This index is the sum of the following standardized variables: (1) number of board meetings, since it is perceived that more effective boards are those which are more monitored and controlled their directors; (2) existence of an audit committee that involves increased control and monitoring exerted by directors, which can entail into increased quality and transparency of financial statements (Klein, 2002;Xie et al., 2003).It is measured as a dummy variable that takes the value 1 if the company has an audit committee and 0, otherwise; (3) existence of a nomination committee, which guarantees the efficiency of independent board members and act as CEO monitors.It is measured as a dummy variable that takes the value 1 if the company has a nomination committee and 0, otherwise; (4) lack of an executive committee, that although at first sight may appear to be related to better corporate governance, the Olivencia Report (1998) points out that this committee assumes a key role in decision making, lowering the importance of having independent directors, which is a signal of bad governance.This variable takes the value of 1 if the firm does not have an executive committee (reflecting good governance) and 0, otherwise; finally, (5) since the Olivencia Report notes that the ideal number of directors is between 5 and 15 directors and following García-Lara et al. (2007), the last variable is measured as a dummy variable that takes the value 1 (good governance) if there are less than 16 directors on the board and 0, otherwise.Higher values of the index are related to stronger governance.After creating the index, we construct a dummy variable (STRONG_WEAK_BOARD) that takes the value 1 if the aggregate index is greater than or equal to the median value, indicating strong corporate governance, and 0 otherwise, indicating weak corporate governance.According to Gompers et al. (2003), Davila and Penalva (2006) and García-Lara et al. (2007), the use of this kind of indexes, which aggregate several governance measures, provides a way of classifying firms considering the strength of their governance more successfully.It shows the importance about the links between a company's CSR and the related board attributes.According to Helfaya and Moussa (2017), it is expected that stronger boards are more likely to disclose CSR information.
The IBEX-35 is the benchmark stock market index of the Spanish capital markets, which includes the largest 35 Spanish firms quoted in the stock market (Sierra, Zorio and García-Benau, 2013).It is included as a dummy variable that equals 1 if the company is listed on the IBEX-35 index and 0, otherwise (Gallego-Álvarez, García-Sánchez, and Rodríguez-Dominguez, 2010;García-Sánchez, 2008).Due to the IBEX 35 Spanish listed companies are large companies (Ortiz and Marín, 2014), it is expected a positive relationship between the inclusion of the company in the Ibex 35 and CRS disclosure.
Finally, in order to control for the effect of macroeconomic variables on companies' behaviour, our models include industry and temporal dummies (Miguel, Pindado and Torre, 2005).
Table 1 presents the description of the variables.

3.2.Model and analytical technique
The relationships proposed in previous hypotheses between the independent variables and CSR reporting take the following form: Since we are working with panel data, it is necessary to consider the individual and the time point, represented by i and t, respectively.Furthermore, the error term is decomposed into the stochastic error term varying cross-time and cross-section combined effect (µ it ,), and the unobserved time-invariant, firm-specific effect (ƞ i ), which varies among individuals, but is constant over time.
The dependent variable is a count data variable, so standard regression models are not suitable due to the fact that they do not take the discrete nature of the variable characteristic into account (Cameron and Trivedi, 1998).As a solution, and in order to control for endogeneity, we estimate the models by using the generalized method of moments (GMM) estimator proposed by Arellano and Bond (1991).Despite the endogeneity problems can also be addressed by using simultaneous equation estimators, the choice is based on consistency concerns (Miguel, Pindado and Torre, 2005).GMM is more consistent and efficient than others because it also controls for the unobservable heterogeneity.It arises because CSR disclosure decision is taken by specific individuals within a firm, thus generating a particular behaviour pattern (Martínez-Ferrero and García-Sánchez, 2014;Martínez-Ferrero, Ruiz-Cano, and García-Sánchez, 2015).In order to test the lack of second-order serial correlation, we apply a serial correlation test of order i by using residuals in first differences (m 2 ) (Miguel, Pindado and Torre, 2005;Martínez-Ferrero and García-Sánchez, 2014).As shown in Tables 4 and 5, this hypothesis is always rejected for all our models.Additionally, the Hansen test confirms the suitability of the instruments that have been used in the estimation.Wald test is also provided, represented as z, which corroborates the joint significance of the coefficients and of the time dummies.

Descriptive Statistics
Table 2 presents the mean value, the standard error, and the 25th, 50th, and 75th percentiles of the variables.The results confirm that CSR disclosure of the firms, on average, is 0.956.Regarding board composition, it is shown that, on average, the largest represented group is that of support specialists (SS) (13.956%), followed by business experts (BE) (12.0%) and community influentials (CI) (6.1%).Table 3 reports the Pearson correlation coefficients.Due to the fact that none of the correlation coefficients is high enough (>0.80), we conclude that multicollinearity is not a concern in our analysis.
As a supplement to the information presented, the Variance Inflation Factor (VIF) test is provided.The highest VIF value is 2.03, which is below the tolerance value of 10, indicating that the results are not biased due to multicollinearity (Kutner et al., 2005).
Table 2: Main Descriptive Statistics Mean, standard deviation, and 25, 50 and 75 percentile values of the main variables.Panel A and B show the continuous and dummy variables, respectively.CSR is the dependent variable, measured as the sum of a maximum of 6 items provided by the company; BE is the proportion of board members classified as business experts; SS is the proportion of board members classified as support specialists; CI_pol is the proportion of politics classified as community influential; CI_others is the proportion of outsiders with no political background classified as community influential; ROA is the return on assets measured as the proportion of operate income before taxes divided by the total assets; IBEX 35 is a dummy variable equal to 1 if the company is listed on the IBEX-35 index, or 0 otherwise; LEV is the leverage ratio measured as the ratio of total debt to total assets; CEO_power is the sum of standardized (1) CEO duality, (2) ratio of directors appointed after the CEO began his tenure to the total number of directors, (3) ratio of shares held by the CEO to director ownership; STRONG_WEAK_BOARD is a dummy variable, equal to 1 if the company has a strong corporate governance and 0 if it has a weak one; OWNERSHIP is the percentage of shares controlled by the largest and second largest shareholders; SIZE measures the company´s size as the logarithm of total assets; BDSIZE is the total number of board directors *** Significant at 1%, ** at 5% and * at 10% Pearson's correlation matrix.CSR is the dependent variable, measured as the sum of a maximum of 6 items provided by the company; BE is the proportion of board members classified as business experts; SS is the proportion of board members classified as support specialists; CI_pol is the proportion of politics classified as community influential; CI_others is the proportion of outsiders with no political background classified as community influential; ROA is the return on assets measured as the proportion of operate income before taxes divided by the total assets; IBEX 35 is a dummy variable equal to 1 if the company is listed on the IBEX-35 index, or 0 otherwise; LEV is the leverage ratio measured as the ratio of total debt to total assets; CEO_power is the sum of standardized (1) CEO duality, (2) ratio of directors appointed after the CEO began his tenure to the total number of directors, (3) ratio of shares held by the CEO to director ownership; STRONG_WEAK_BOARD is a dummy variable, equal to 1 if the company has a strong corporate governance and 0 if it has a weak one; OWNERSHIP is the percentage of shares controlled by the largest and second largest shareholders; SIZE measures the company´s size as the logarithm of total assets; BDSIZE is the total number of board directors.Variance Inflation Factor (VIF).*** Significant at 1%, ** at 5% and * at 10%.

Regression results
In Table 4, the estimations for testing our research hypotheses are displayed.To test the first hypothesis, Model 1 is provided in column 1, showing that the results disagree with our predictions, presenting a negative sign and, as a result, those active or retired directors who come from other companies (BE) negatively affect CSR disclosure.
Therefore, the hypothesis H1 is rejected.
As shown in the second column of the table, where the results of Model 2 are provided, there is a significant positive relationship (p < 0.01) between SS and CSR disclosure, indicating that the higher the number of SS on the board, the higher the probability of disclosing CSR information.It confirms our second hypothesis and supports the theory that, since SS are directors who provide support in areas that demand specialized expertise such as law, marketing, and finance, their personal image and social reputation can be strongly associated with the evolution of the company.
Regarding our hypothesis H3a, the results offered in Model 3, where community influentials with political ties are explored, are not statistically significant.Nevertheless, we note that the influence of those CI with political influence in their background is positively associated with CSR disclosure on the overall model.Therefore, Model 5, where all the variables are jointly examined, evidences that a greater of directors on boards with political connections has a negative effect on CSR disclosure, supporting H3a.
According to the results in Model 4, we reject H3b, which predicts that the community influential directors without political connections have a positive effect on CSR disclosure.
With regard to control variables, those of IBEX35, return on assets (ROA), leverage (LEV), board size (BDSIZE), firm size (SIZE) and strong or weak board (STRONG_WEAK_BOARD) present a significant positive sign in all the models, as predicted.These results find that large leveraged companies, with big and strong boards, as well as firms with high return on assets, have a positive effect on CSR disclosure.On the contrary, the remaining control variables have a negative and significant effect upon CSR reporting.

Extension Analysis
In this section, we use a moderator analysis to determine whether the relationship between CSR disclosure and the type of director is moderated by the CEO_power index (hypothesis 4).To perform this analysis, the variables corresponding to the multiplicative term between the independent variables (BE, SS, CI_pol and CI_others) and the moderator variable (DCEO_power) are created.DCEO_power is a dummy variable equals to 1 if CEO_power takes a value higher than its mean value and 0, otherwise.
The results of the GMM regression analyses are presented in Table 5.In Model 1, we report the positive effect of the interaction of DCEO_power and BE on CSR disclosure.Consistent with Table 4, while the proportion of BE has a negative effect on CSR disclosure, the interaction with DCEO_power has a positive and significant effect, noting that the negative relationship between the proportion of BE on board and CSR disclosure is weaker for firms with powerful CEOs.The results are in line with those reported by Gul and Leung (2004), who found a negative influence of the expertise of non-executive directors on voluntary disclosure in Hong Kong, as well as a positive moderating role of CEO duality.
The results in Model 2 (Table 5) also confirm that SS directors with specific abilities and backgrounds tend to be more sensitive towards social and environmental demands and that this effect is even higher under powerful CEOs.With respect to CI_pol and CI_others directors, Model 3 (Table 5) shows the interaction of these variables with CEO_power, suggesting that the negative effect of community influentials on CSR disclosure is so strong that CEOs are not sufficiently powered to compensate their negative role on CSR reporting.and Corporate Governance rank, 0 otherwise.The results, not reported for the sake of brevity, show that our main results are qualitatively the same, confirming that the negative impact of BE and politicians, the positive influence of SS directors and the neutral impact of other community influential on CSR, are independent of the CSR measure.

DISCUSSION
Overall, the empirical evidence of this study supports the view that firms should highlight the unique capabilities of directors in order to better understand how board composition impacts on CSR reporting.In particular, our results allow us to confirm the significant role of directors categorized as support specialists as well as the hypothesis that these directors have specific qualifications, abilities and technical knowledge that increase their feeling in the direction of society and stakeholders' interest, which results in a higher engagement and commitment with CSR disclosure.Our results are in line with previous literature (e.g.Francis et al., 2015;Gray and Nowland, 2017) that has noted that specialized directors on boards provide benefits and positive outcomes to firms.
In addition, the findings evidence a negative effect of political board connections on CSR reporting, what supports the view that community influential directors with political ties may be interested in achieving their own aims by using political resources to the detriment of shareholders' and stakeholders' needs.These results are in line with previous literature (e.g.Bona-Sánchez et al., 2014;Chaney et al., 2011), which has suggested that these directors may have a detrimental effect on firm transparency due to their interests on safeguarding their reputation within the company against public scrutiny and protecting their political connections.In summary, our results suggest that when firms are interested in increasing CSR disclosure as a potential firm strategy, they should consider the industry-specific knowledge and technical skills more than the professional experience or political networks of potential new board members.
This paper also supports the theories that appeal for analysing the multiple configurations of corporate governance mechanisms by adopting a "holistic approach" and the need to combine them in order to analyse their impact on CSR behaviour.In particular, our findings note that support specialists directors tend to be more sensitive towards social and environmental demands under powerful CEOs.In addition, the expected positive effect of business experts on CSR only happens under powerful CEOs.
Overall, these findings confirm that the non-monetary CEO incentives related to career concerns, reputation, entrenchment and power only have a positive influence on the attitude towards CSR commitment with business expert and support specialist directors.
On the other hand, the results also show that the negative effect of political directors on CSR disclosure does not change under firms with powerful CEOs.
As practical implications, these results confirm the recent calls for appointing more board members with specific knowledge and technical qualifications.These requests have been more popular after the recent governance scandals and criticisms over opacity and lack of transparency in social and environmental activities.Our results also have relevant implications for policy makers, since the recent international recommendations call for a more professional boardroom of directors to ensure the board understanding of the firm's financial aims, the difficulties of global markets, and the consequences of the business on different stakeholders (European Commission, 2011).In addition, our findings provide relevant implications for countries where politically connected boards are prevalent.In these countries, regulatory authorities interested in increasing transparency should recommend the need to report board political connections in governance reports as this information can condition CSR reporting.

CONCLUDING REMARKS
The objective of this paper has been to study the influence of human and social board capital on CSR disclosure of Spanish companies.Following the Hillman et al. (2000) taxonomy of board members, we have classified outside directors as business experts, support specialists, political directors and other community influentials, and have examined whether business, technical expertise or political ties in the boardroom affect CSR disclosure.For a sample comprising 152 non-financial listed firms on the Spanish stock market during the 2008-2014 period, and after applying several regressions for panel data, this paper provides the following evidence.First, outside directors classified as support specialists increase transparency in CSR reporting.Their specific skills and technical knowledge could justify their positive impact on strategic firm decisions related to CSR reporting, as well as their major qualifications to advise and manage environmental risks related to environmental fines or litigation costs.Second, the findings show that those directors with political connections negatively impact on CSR reporting.
These results are in line with previous literature which has hold that these directors may be interested in achieving their own aims to the detriment of stakeholders' needs and reduce firm transparency in order to safeguard their reputation and to protect their political connections.Finally, our results confirm the effect of powerful CEOs in affecting the association between board composition and CSR disclosure.Nevertheless, the findings show that the negative effect of political directors on CSR disclosure remains unchanged under firms with powerful CEOs.
As main limitations of the research, we must remark the use of our proxy of CSR.
Although we believe this measure to be reliable -following previous studies, e.g., García-Meca and Pucheta (2017) -we are cautious about the possible bias included in it, because the final assessment is conditioned by the person who processes the information.
Nevertheless, we must say that our results are robust to alternative proxies of CSR based on Merco and Dow Jones Sustainability rankings.Another limitation is the classification of outside directors among business experts, support specialists and community influentials, which has been based on the biographical information available.As future research, more evidence examining the implications, motivations and effects of these categories of directors on firm outcomes (e.g.innovation policies or risk strategy, etc) should be useful to better understand the role of these directors in the governance and reporting strategies of companies.In addition, we recommend researchers to explore our evidence in international samples as well as with non-listed and small and medium firms.

Table 1 :
Variable Definition

Table 4 :
Results of the Generalized Method of Moments of the baseline model Estimated coefficients (std.error).CSR (t-1) is the first lag of the dependent variable, measured as the sum of a maximum of 6 items provided by the company; BE is the proportion of board members classified as business experts; SS is the proportion of board members classified as support specialists; CI_pol is the proportion of politics classified as community influential; CI_others is the proportion of outsiders with no political background classified as community influential; CEO_power is the sum of standardized (1) CEO duality, (2) ratio of directors appointed after the CEO began his tenure to the total number of directors, (3) ratio of shares held by the CEO to director ownership; STRONG_WEAK_BOARD is a dummy variable, equal to 1 if the company has a strong corporate governance and 0 if it has a weak one OWNERSHIP is the percentage of shares controlled by the largest and second largest shareholders; SIZE measures the company´s size as the logarithm of total assets; BDSIZE is the total number of directors on the board; ROA is the return on assets measured as the proportion of operate income before taxes divided by the total assets; LEV is the leverage ratio measured as the ratio of total debt to total assets; IBEX 35 is a dummy variable equal to 1 if the company is listed on the IBEX-35 index, or 0 otherwise; Year and industry effect are included to control for possible effects on the results; z is a Wald test of the joint significance of the reported coefficients, under the null hypothesis of no relationship; m 1 m 2 are serial correlation tests using residuals in first differences, under the null hypothesis of no serial correlation; Hansen is a test of over-identifying restrictions, under the null hypothesis of non-correlation between the instruments and the error term; degrees of freedom in parentheses.*** Significant at 1%, ** at 5% and * at 10%.

Table 5 :
(Gallego-Álvarez and Quina- Custodio, 2016)s of the extension modelEstimated coefficients (std.error).CSR (t-1) is the first lag of the dependent variable, measured as the sum of a maximum of 6 items provided by the company; BE is the proportion of board members classified as business experts; SS is the proportion of board members classified as support specialists; CI_pol is the proportion of politics classified as community influential; CI_others is the proportion of outsiders with no political background classified as community influential; ROA is the return on assets measured as the proportion of operate income before taxes divided by the total assets; IBEX 35 is a dummy variable equal to 1 if the company is listed on the IBEX-35 index, or 0 otherwise; LEV is the leverage ratio measured as the ratio of total debt to total assets; DCEO_power is a dummy variable equal to 1 if CEO_power takes a value higher than its mean value and 0 otherwise; STRONG_WEAK_BOARD is a dummy variable, equal to 1 if the company has a strong corporate governance and 0 if it has a weak one; OWNERSHIP is the percentage of shares controlled by the largest and second largest shareholders; SIZE measures the company´s size as the logarithm of total assets; BDSIZE is the total number of board directors; Year and industry effect are included to control for possible effects on the results; z is a Wald test of the joint significance of the reported coefficients, under the null hypothesis of no relationship; m 1 m 2 are serial correlation tests using residuals in first differences, under the null hypothesis of no serial correlation; Hansen is a test of over-identifying restrictions, under the null hypothesis of non-correlation between the instruments and the error term; degrees of freedom in parentheses.***Significantat1%, ** at 5% and * at 10%.The robustness of the models has been checked by examining if the prior results are sensitive to alternate CSR measurements.To start with, it is important to mention the two indexes in which we are based on.On the one hand, we rely on Merco-Responsibility and Corporate Governance Index, which has been working since 2000 to measure the reputation of the 100 top responsible Spanish companies.On the other hand, the Dow Jones Sustainability Index (DJSI) is also used as a proxy variable of CSR.It considers the environmental, financial and social behavior of the companies that are committed to sustainability and voluntary disclosure good practices(Gallego-Álvarez and Quina- Custodio, 2016).Using both indexes we have created a new dummy variable assigned a value of 1 if the company pertains to the DJSI rank and/or to the Merco-Responsibility