Effects of Activist Shareholding on Corporate Social Responsibility Reporting Practices: An Empirical Study in Spain

New business practices are mainly characteristic of large firms, especially those quoted on the stock market. Listed companies show a higher commitment to corporate social responsibility (CSR) practices because capital markets allow activists to become a firm’s socially oriented shareholders. These actors, although small in number, have a significant influence over other larger block-holders. Recent decades have witnessed a significant increase in societal pressure to control the behavior of companies owing to the risks deriving from the economic, social and environmental effects of their business activity. The aim of this work is to test the effect that CSR activist shareholders have on the decision to disclose corporate social responsibility information in the Spanish context, controlling for the rest of the dimensions in Ullmann’s theoretical framework.


INTRODUCTION
Situations such as growing pollution, the dumping of toxic waste or the use of child labour, among others, have brought about a process of social awareness that has increased the pressure felt by firms regarding certain theoretical ethical limits that should not be breached when doing business. This acceptable ethical threshold has meant that firms not only comply with legal requirements, but also voluntarily tend to adopt environmental improvements in their manufacturing processes, implement environmental management systems and make it easier for their employees to conciliate their work with their personal lives.
In this sense, companies need to perform well and undertake socially desirable actions, including the distribution of economic, social or political benefits to the groups from which they derive their power (Shocker & Sethi, 1973;Alcabés, 2005). Furthermore, as businesses recognize their stakeholders' social expectations, the role of corporate social reporting takes on increasing importance as a mechanism through which such duties of accountability may be discharged (Gray, Owen & Adams, 1996;Tran, 2009).
Several authors, such as Liu and Anbumozhi (2009), argue that external stakeholders, particularly the primary ones -shareholders, government and creditors -are the principal actors in the process of corporate social responsibility (CSR) transparency. In this line, a fewer number of papers such as Roberts (1992), Liu and Anbumozhi (2009) and Prado et al. (2009a), using Ullmann's theoretical framework, have analyzed the role of these stakeholders in the process, showing that there is a limited impact of shareholder power on it.
These papers assume that dispersed ownership is directly associated with firms' accountability practices, although they do not empirically confirm this relationship. Nevertheless, the last paper mentioned provides evidence for the Spanish setting that larger dominant shareholders are more likely to adopt the emission of CSR reports drawn up according to an international standards guide. Prado et al. (2009a) argue that this effect is due to the strong relationship between the reputation of the firm and that of the shareholders.
On the other hand, other authors such as Lee (2009) and Solomon (2006) indicate that firms' commitment to CSR reporting is justified by the fact that managers and larger shareholders could be influenced by activist stockholders. In this sense, the aim of this paper is to test the effect that activist shareholders have on the decision to disclose corporate social responsibility information, controlling for the rest of the dimensions proposed by Ullmann. The result obtained shows that activist shareholders have a significant effect on the process of transparency in relation to the triple bottom line of firm behavior.

SHAREHOLDER ACTIVISTS AND SOCIAL DISCLOSURES: RESEARCH HYPOTHESIS
Stakeholder Theory asserts that the reasons behind social information disclosure is that an organization's survival depends on the support of their stakeholders, understanding them as ''a person or group that can affect or is affected by the achievement of the organization's objectives'' (Freeman, 1984, p. 46). In this sense, CSR reporting is the mechanism used by corporations to show a firm's social performance to the stakeholders (Roberts, 1992).
Resource dependency theory suggests that power accrues to those parties that control the resources required by the organization, thus creating power differentials among stakeholders (Pfeffer, 1981). In this sense, the power of stakeholders is a function of the resources they control that are essential to the corporation (Ullmann, 1985).
At the most fundamental level, ownership stakes controlled by different stakeholder groups accrue power to these groups vis-à-vis the firm and heighten the urgency that the demands of these groups be met ( Van der Laan Smith et al., 2005).
Generally, authors have considered that the less the influence of the top shareholders, equivalent to more dispersed ownership, the greater the likelihood that firms will disclose more information (Keim, 1978;Ullmann, 1985;Craswell & Taylor, 1992;Christopher & Hassan, 1996;Frost, 1999); however, this effect has not been empirically tested for CSR information (Roberts, 1992;Liu & Anbumozhi, 2009;Prado et al., 2009a). In contrast, the positive effect of larger dominant shareholders has been confirmed for the Spanish setting (Prado et al., 2009a).
This last effect could be explained by the fact that the shares of listed firms are traded on the public stock exchanges so the door is open to anyone who can afford to purchase a share to become a shareholder and try to influence the firm. This door could be considered the unique access to corporate management for social and environmental activists in order to advance their socially oriented agendas (Lee, 2009).
The potential influence of this type of shareholder, according to Lee (2009), is higher, even though they are small in number, because (i) the voice of activist has a significant potential to influence the opinions of other large shareholders and, moreover, (ii) corporate managers could consider the most vocal group of shareholders in order to avoid losing legitimacy in the eyes of the majority stockholders. In fact, activist shareholders have expanded their demands from the circumscribed realm of shareholder rights to issues of how successors to the CEO are chosen, how much executives are paid, etc. (Davis & Thompson, 1994).
Taking into account these considerations, the following alternative hypothesis is posed: h 1 : Céteris páribus, the contents and quality of the corporate social responsibility report are positively affected by the presence of social activists in the firm's ownership structure.

Sample
The target population in the study corresponds to the 116 non-financial Spanish firms quoted on the Spanish continuous market. It should be pointed out that the population was selected taking into consideration the criteria of size and stock market listing used in previous studies (Guthrie & Parker, 1990;Hackston & Milne, 1996;Collet & Hrasky, 2005, Gallego, 2006, García-Sánchez, 2008, as well as the fact that they are obliged to deposit information on corporate governance before the Comisión Nacional del Mercado de Valores (National Stock Market Commission). Consultation of their database provided us with reports from 99 firms during 2009, the sample used in the analysis.

Dependent variable: Practices in corporate social reporting
In this study we have adopted the proposal for CSR reports made by Prado et al. (2009a). Thus, firms' CSR disclosures have been classified into five issues. They allow us to define five dependent dummy variables that take the value of one to identify the content and quality of CSR Reporting, and 0 otherwise: 1. The firms disclose several items of economic, social and environmental aspects. 2. The CSR report presents an informal format in accordance with stakeholder demands. 3. The CSR information report is adapted to a standard of the most widespread international model, the Global Reporting Initiative (GRI). 4. The information has been verified or audited by an independent entity which guarantees its accuracy and credibility. 5. The report's compliance with the demands of the GRI has been certified on the part of the organization responsible.
These five original variables were summarized and a principal components analysis was estimated to make it possible to simplify the dependent variables previously considered into components that reflect the underlying common dimensions.
Prior to estimation of the principal components analysis, the Kaiser-Meyer-Olkin measurement of sampling adequacy and Bartlett's sphericity test were run. The results obtained show an adequate basis for the empirical examination of factor analysis sufficiency. Table 1 shows how the sufficiency measurement of the general sampling falls within the range of acceptance, as well as the significance of Bartlett's sphericity test. Subsequently, principal components analysis (VARIMAX rotation) was run, and the results are given in Table 2.
By analyzing the loadings, it can be seen that Component 1, ASSURANCE, represents CSR reporting where the information has been verified by an independent entity and the contents and format certified as In Accordance by the GRI.
Component 2, STAKEHOLDER DEMANDS, identifies those firms that give out information on economic and social, and/or environmental matters, in an informal format according to their stakeholder demands.
Finally, the last component, GRI FORMAT, includes CSR reporting in which the contents and format meet the requirements and demands of the GRI, but have not been certified. The disclosing of information in accordance with this model would entail an increase in the contents, quality and objectivity of the information.

Independent Variable: Active Shareholders
In order to represent active shareholders, we have used the ratio number of directors that represent active shareholders' interests divided by the directors on the board.

Control Variables.
The control variables proposed in this paper are based on the theoretical foundations proposed by Ullmann (1985), which have been tested by Roberts (1992) and Prado et al. (2009a).
Ullmann defined a conceptual framework for the factors explaining social disclosure based on three dimensions: Stakeholder power, Strategic posture and Economic performance.
-Stakeholder Power reflects the theoretical basis of the cited framework since the firm, when stake- holders control resources critical to the organization, is likely to respond in a way that satisfies the demands of the stakeholders.
-Strategic Posture describes the mode of response of an organization's key decision-makers toward demands. An active posture implies a position in which managers seek to influence their organization's relationship with important stakeholders in order to achieve optimal levels of interdependence.
-Economic Performance is important in two ways: (i) it determines the relative weight of social demand and the attention it receives; for instance, in periods of high profitability, social demand receives more attention; and (ii) it influences the financial capability to undertake costly programs related to social demands.
The independent and control variables are summarized in Table 3.

Analysis Model
Testing hypotheses H 1 entails analyzing the effect of the power of the active stockholders on the characteristics and contents of the CSR report, which have been synthesized in three components through factor analysis. Dependence models or multiple linear regression models are used for this purpose.
With that goal in mind, we propose the following model [1], in which the CSR report characteristics are a function of activist shareholders and several variables that represent the dimensions of Ullmann's framework. In which: ActShaPower i is the independent variable that identifies the power of active shareholders on company i, measured by the number of directors that represent socially active stockholders.
OtherStkPower i are control variables that identify the power of other stakeholders on company i, measured by the variables: INDEPENDENT, SIZE, INDUSTRY and LEVERAGE.
StratePosture i are control variables that identify the environmental and social strategic posture of company i, measured by the variables: ISO14001 and OHSAS18001.
EcoPerfor i is a control variable that identifies the economic performance of company i, measured by ROA.

RESULTS OF THE ANALYSIS
The results obtained by the estimation of all of the dependence models proposed are synthesized in Table 4. Model 1, which analyzed the effect of active shareholder power on CSR report assurance, has an explanatory power of 41.30% for a confidence level of 99% (p-value < 0.01).
On analyzing the variables individually, it can be seen that the independent variable proposed, which represents the pressures exerted by active shareholders, does not affect the external assurance of the CSR report.
Regarding the control variables, those relating to the firm's strategic posture on presence/absence of ISO14001 and OHSAS18001 certification have a positive effect at a confidence level of 99% in the model relating to Component 1 -ASSURANCE.
The variables SIZE, INDUSTRY and ROA have a positive effect on assurance, but this impact is statistically insignificant, whereas CREDITORS PRESSURE has a negative insignificant effect on the dependent variable.
Model 2, which analyzed the effect of active shareholder power on CSR report format according to a firm's stakeholder demands, has an explanatory power of

Minority Shareholder powers
Several authors suggested that the less the influence of the top shareholders, equivalent to more dispersed ownership, the greater the likelihood that firms will disclose more information (Craswell & Taylor, 1992;Christopher & Hassan, 1996;Frost, 1999). These shareholders do not have enough power to influence the firm's decision and their motivations stem from personal preferences. But, in this respect, independent directors play a special role in ensuring observance of the law and defending minority shareholders' interests (Fama & Jensen, 1983). Prado et al. (2009b) NORITY SHAREHOLDER Numerical variable measuring the percentage of independent board members who represent the interests of the minority shareholders

Government power
Political Costs theory affirms that the information which a company provides is used to draft the government regulations affecting them. As a consequence, the executives of that company will attempt to choose the disclosure policies which most contribute to eliminating political interference and to producing a decrease in costs such as taxes, fees and regulated charges, among others. Trotman & Bradley (1981) Belkaoui & Karpik (1989) Patten (1991) Archel (2003) Adams et al. (1995, 1998) Hackston & Milne (1996 Deegan & Gordon (1996) Liu & Anbumozhi (2009) INDUSTRY Dummy, takes the value of 1 if the firm's activity belongs to sensitive industries, and 0 otherwise

Creditors' power
Creditors control access to financial resources that may be necessary for the continued operation of a corporation, so the greater the degree to which a corporation relies on debt financing to fund capital projects, the greater the degree to which corporate management would be expected to respond to creditor expectations concerning a corporation's role in socially responsible activities. Roberts (1992) Prado et al. Strategic posture refers to how a company may respond to social demands. An active strategic posture towards social demands is expected to result in greater social responsibility activities, outstanding among which are the drawing up and disclosure of CSR reports (Ullmann, 1985).

Economic Performance
Owing to the substantial costs involved in becoming socially responsible, economic performance is an important factor in determining whether social responsibility issues will be on the priority list. Cowen et al. (1987) Belkaoui & Karpik (1989 Hackston & Milne (1996) ROA Numerical variable that represents the return on assets.
19.60%, for a confidence level of 95% (0.01 < p-value < 0.05). The pressure exerted by active shareholders has a positive effect on this CSR report typology for a confidence level of 95%.
Regarding the control variables, the presence/absence of ISO14001 and the pressure exerted by minority shareholders have a positive effect at confidence levels of 95% and 90% (0.05 < p-value < 0.1), respectively, in the model relating to Component 2.
The variables INDUSTRY, CREDITORS PRES-SURE and ROA have a positive effect on assurance, but this impact is statistically insignificant. On the other side, SIZE and OHSAS18001 have a negative insignificant effect on the dependent variable.
Model 3, which analyzed the effect of active shareholder power on CSR reporting according to the GRI format, has an explanatory power of 29.10% for a confidence level of 99%.
The independent variable proposed has a negative effect on Component 3-GRI FORMAT for a confidence level of 95%. In contrast, two control variables, CREDI-TORS PRESSURES and ISO14001, have a positive effect on it for a confidence level of 99%.
The variables SIZE, INDUSTRY, ROA, OHSAS 18001 and MINORITY SHAREHOLDER PRESSURES have an insignificant positive effect on Component 3, except the two last variables, for which the effect is negative.
The overall analysis of the results obtained allows us to partially accept the hypothesis proposed since activist shareholder pressure has a contradictory effect on CSR disclosure practices. On the one hand, these stakeholders encourage the process of drawing up a report that satisfies the firm's stakeholder demands in the form of a GRI format that is highly general. However, on the other hand, these actors are not interested in the external verification of the information disclosed.

DISCUSSION OF RESULTS
The empirical evidence of the present paper indicates that activist shareholders play an important role in a firm's decision to disclose CSR information. Specifically, we have observed that the directors who represent this type of shareholder are very interested in having firms make available to their stakeholders the information they consider most suitable to their needs. In certain aspects, these demands for information are negatively related to the adoption of internationally accepted guidelines, such as those proposed by the GRI. In this sense, Logsdon and Van Buren (2008) have shown that the majority of activists' actions deal with specific issues that affect specific groups. Moreover, through direct comparison with the results obtained by García-Sánchez, Gallego-Álvarez and Prado-Lorenzo (2008) and Prado et al. (2009a), our results show that stakeholder influence, particularly government power, is not the driver of CSR disclosure. This contrast effect allows us to state that firms disclose information orientated to satisfying their stakeholders' demands and not in order to reduce political costs.
On the other hand, and in line with the results obtained by Prado et al. (2009a), the adoption of this international standard and the verification of the information disclosed is strongly linked to the strategic posture a firm has adopted regarding the social and environmental aspects of its behavior, particularly with respect to the latter. This relationship cannot be extended to the economic dimension of the corporations activity. We have therefore concluded that firms' disclosure practices are a trade-off between activist demands and the social and environmental strategic plan designed by managers.

CONCLUSIONS AND LIMITATIONS
In this study we posited a dependence model in an attempt to analyze the effect that activist shareholders have on a firm's decision to disclose information on corporate social responsibility in the Spanish context, controlled by the dimensions proposed by Ullmann. In the model posed, the dependent variables are grouped into three components: Component 1, ASSURANCE, which represents the guarantee of the CSR report when the information has been verified by an independent body and the contents and format are certified as In Accordance with the GRI. Component 2, STAKE-HOLDER DEMANDS, identifies the firms that provide information on economic, social and/or environmental aspects in an informal format in accordance with the demands of their stakeholders. Component 3, GRI FORMAT, includes CSR reports in which the contents and format comply with GRI requirements and demands, but have not been verified.
The independent variable is represented by the activist shareholders, and is obtained by the ratio of the number of directors representing the interests of the activist shareholders divided by the number of directors on the Board. The control variables in this research are based on the theoretical framework proposed by Ullmann (1985), and are the factors that explain the disclosure of social information on three levels: power of the stakeholders, strategic posture of the firm and economic performance.
The results obtained show that activist shareholders have a significant effect on the process of transparency related to the triple bottom line (economic, social and environmental information) of firm behavior. These results should be compared with those obtained for firms of other countries, with a different economic environment and in a different time frame, as possibilities for future research.