There is a number of competing theories have been advanced to explain CSED. All these theories seem to provide plausible explanations of corporate social and environmental behaviour. Moreover, most CSED studies concentrate on examining these theories separately, maybe because of the complexity of testing more than one theory at a time as these theories are different in terms of ontological, epistemological and methodological positions.
This study aims to test and discriminate between two  of the CSED explanatory theories. Nevertheless, the discrimination between any two or more theories to explain CSED is somehow problematic because it may prove difficult to judge what encourage CSED and to what extent. For example, a company may make disclosures because it has a stock market quotation. In other words, it responds to stock market demands (as explained by agency theory). Alternatively, the disclosure can be explained by the public profile of the company and lobbying activity. In this case, the requirement to create public understanding is regarding to the acceptability of its activities (legitimacy theory explanation). Moreover, some of theses theories may share the same explanatory factors such as composition of non-executive directors, chairman with multiple directorships, size and profitability. This makes it difficult to recognise which of these theories is supported by which variable.
This sub-section introduces an overview of the two competing CSED theories legitimacy theory as well as agency theory. Moreover, the discrimination between the two theories will be discussed based on three areas; the level of CSED provision, themes of CSED and quality and sensitivity of CSED.
3.1.1 Legitimacy theory (LT)
Legitimacy theory can be defined as ''a generalised perception or assumption that the actions of any entity are desirable, proper, or appropriate within some socially constructed system of norms, values, beliefs and definitions'' (Suchman, 1995, p. 574). Actions by corporate management to convince wider society that the firm is socially responsible are treated as a part of the legitimation process (Gray et al., 1995a). Moreover, legitimacy theory is based on the premise that firms signal their legitimacy by disclosing certain information. Legitimacy theory is centred on the notion of a contract between a firm and its constituents (Shocker & Sethi, 1974). Since voluntarily revealing certain information, directors can communicate with stakeholders (such as environmentalists, other investors, Government agencies etc.) who as a result will feel more confidence about the performance (Cook & Haniffa 2005).
Lindblom (1994) suggests four broad legitimation strategies that firm may take: informing stakeholders about improvements in performance; aims to change stakeholder's perceptions of a particular enent; distracting attention away from an issue; and changing external expectations about its performance. As these strategies include gaining and repairing legitimacy (Suchman, 1995), Dowling and Pfeffer (1975) suggest three modes of action that firms can take to enhance legitimacy: adapt outputs, goals and methods of operation to conform to prevailing definitions of legitimacy; altering the definition of social legitimacy through communication, output and values; and finally, attempt through communication to become identified with symbols, values or institutions which have a strong base for social legitimacy. Gray et al. (1995a) link the actions proposed by Dowling and Pfeffer (1975) and the strategies suggested by Lindblom (1994) within the framework of legitimacy theory. This relationship is illustrated in Figure (2).
There are many of empirical studies have employed a number of proxies to test aspects of legitimacy theory except content analysis. Some of these proxies include: ownership structure, industry sensitivity (Deegan and Gordon, 1996); major corporate events (Guthrie and Parker, 1989; Patten, 1992); and type of users of annual reports (Deegan and Rankin, 1997). Some of the recent studies have focused on the motivations of managers for CSED. ODonovan (2002) supports legitimacy theory and provides insight into management disclosure behaviour based on scenarios that have different impacts.
Where a legitimacy gap exists, the company has to consider a response. One response is to ignore the gap, and assume there is no adverse consequences will arise (Cook & Haniffa 2005). Alternatively, company can disclose information that helps to reduce the legitimacy gap, and this is the focus of this study.
3.1.2 Agency Theory (AT)
Positive accounting theory that proposed by Watts & Zimmerman (1978) encompasses agency theory. This theory can be used to explain the incidence of social and environmental disclosure in firm annual reports. According to Choudhury (1989), agency theory can be defined as:
"Agency theory fires the imagination as much ac it raises hackles, it is not a subject that can be ignored" (p. 105).
Agency theory dictates that managers will disclose social information only if it increases their welfare, that is, when the benefits from disclosure outweigh the associated costs. Agency costs are incurred when the manager (agent), by acting in his/her self-interest, acts to the financial detriment of the principal (shareholder). An overview of agency theory is represented in figure (2).
Jensen and Meckling's (1976) positive agency theory provides a framework linking disclosure behaviour to corporate governance. Corporate governance mechanisms are introduced to control the agency problem and ensure that managers act in the interests of shareholders. In theory, the impact of internal governance mechanisms on corporate disclosures may be complementary or substitutive. If it is complementary, agency theory predicts that a greater extent of CSED is expected since the adoption of more governance mechanisms will strengthen the internal control of companies and provide an "intensive monitoring package" for a company in order to reduce opportunistic behaviours and information asymmetry (Leftwich, Watts and Zimmerman 1981; Welker 1995). Managers are not likely to withhold information for their own benefits under such an intensive-monitoring environment, which lead to improvement in disclosure comprehensiveness and quality of financial statements.
These apparently conflicting viewpoints on the impact of corporate governance have not been totally resolved. In spite of this theoretical ambiguity, Hill (1999) argues that no one single mechanism is a governance panacea and suggests that "it is desirable to have a system of overlapping checks and balances." Therefore, the hypotheses about the effect of internal governance mechanisms in the current study are predictions of a positive association.
The disclosure of ratios in firm accounts may provide users of financial statements with new information not calculable elsewhere. A point of note is that several empirical studies (e.g., Hossain et al., 1994; Cooke, 1989b), make no distinction as to whether disclosures provide old or new information. For example, both Cooke (1989) and Hossain et al. (1994) include in their disclosure indices `old' information such as ratios and financial history disclosures, as well as projections and segmental information (examples of `new' information) disclosures. The inclusion of old items can be justified if its inclusion aids the users' understanding, or economises on their time, or reduces the cost of obtaining the information elsewhere. That is, such disclosures may improve the overall quality of an annual report. If by improving disclosure quality, managers reduce agency costs, then agency theory can explain why managers seek to disclose accounting ratios.
3.1.3 Discriminating between the two Theories of CSED: LT and AT
The prior literatures have found difficulties arising when examining more than one of the competing CSED theories especially when theses theories come from different epistemology, ontological and methodological positions. In fact, these competing theories provide competing hypotheses and in order to discriminate between these competing hypotheses there is a need to an investigation (Cormier et al 2005). This study attempts to statistically discriminate between LT and AT by investigating the explanatory power of these theories on three areas; the level of CSED provision, themes of CSED and quality and sensitivity of CSED.
22.214.171.124 In explaining the level of CSED
The Hypotheses development in both sections 3.2.1 and 3.2.2 suggest nine variables to examine legitimacy theory LT: composition of non-executive directors, chairman with multiple directorships, ownership structure by foreign shareholders, the existence of an audit committee, size, profitability, leverage and auditor reputation. In contrast, they suggest the same number of variables to test agency theory AT, these are: composition of non-executive directors, chairman with multiple directorships, managerial ownership, block holder ownership, size, profitability, liquidity and leverage. Accordingly, the two theories from which hypotheses are drawn are linked regarding to some of their predictions. However, they both argue an association between CSED and five variables which are: composition of non-executive directors, chairman with multiple directorships, size, profitability and leverage.
While the discriminating between LT and AT with regards to composition of non-executive directors, chairman with multiple directorships, size, profitability and leverage is not an easy task because of the complexity of these hypotheses that yield the same empirical predictions, the discrimination between the two theories based on the hypotheses driven to examine each of the theories individually is not difficult. Further more, discriminating between the two theories can be applied by observing the pattern of hypotheses confirmation and also the outcomes of each hypothesis test for first, convergent evidence. In other words, confirmation of the hypotheses testing a given theory. Second, the test will be for discriminate evidence (what pattern of rejecting the null hypotheses permits one to favour one theory over the other).
According the evidence from reviewing CSED literature, it can be noted that, the majority of the prior studies have adopted legitimacy theory. In other words, these studies provide results and argument to support this theory (see Deegan 2000). In contrast, studies that are based on agency theory are very rare (Massoud 2009). Moreover, there are no serious attempt exists in the prior studies in order to discriminate between these theories. This study does not suggest the superiority of any particular theory in explaining CSED, as there is no clear theoretical justification to select one theory to be more favourable than other (Massoud 2009). In addition, this study suggests that, both LT and AT have equal explanatory power in explaining CSED phenomenon.
126.96.36.199 In explaining the volume of CSED in different themes
In order to analyse CSED, prior studies use a number of sub-areas (theme) of CSED, such as ; environmental information, community involvement, disclosure related to employees, fair business practices, and so on, see for example Deegan et al (2000), Gray et al (2001) and Deegan et al (2002). According to Gray et al (2001), the relationship between CSED and its determinants varies depending on the type of theme of CSED. Consistently, there are some studies provide evidence that is only no more than one applicable to specific disclosure theme (usually it refers to environmental disclosure) (Massoad 2009). Guthrie and Parker (1989) analysed CSED from a legitimacy theory point of view, and they identify a relationship between CSED and environmental disclosure. Moreover, this study was not able to confirm the relationship between LT and energy as well as community involvement. Another example here is Cormier and Gordon (2001). They examine disclosure of social and environmental information determinants using legitimacy theory. The findings of this study indicate that, information cost and benefits are depending on just environmental disclosure whereas, this relationship is not applied for social disclosure. Regarding to this, there are many studies are based on analysing the relationship between CSED and environmental disclosure (see for example, Pattern 1992; Deegan and Gordon 1996; Wilmshurst and Frost 2000).
Legitimacy theory suggests CSED as a channel to communicate legitimacy whereas agency theory explains CSED as a method/tool to reduce agency costs. In addition, there are no theoretical basis emphases that the explanatory power of theses theories is limited just to a particular theme of CSED (Massuad 2009). Therefore, this study suggests and examines the argument that both theories even they are driven from different perspectives explain all CSED themes. Accordingly, the hypotheses would be applicable to the extent of disclosure in each thematic group.
188.8.131.52. In explaining the volume of sensitive and quality CSED
In a theory, companies should disclose all the information they have even they might be bad news (Clatworthy & Jones 2003). In fact, firms tend not to disclose their information if it is bad (Suijs 2005) as these disclosures result usually in undesirable market effects. Cormier et al (2005) found that, environmental bad news is disclosed by third parties such as environmental agencies and this cause negative implication for a company's stock price as it reveals costly environmental obligations (Cormier & Magnan 1999; Hughes 2000). Moreover, market reaction may occur if investor thought that mangers do not disclose information when it is negative or bad news (Clatworthy & Jones 2003). According to the AT argument made earlier, the information withheld by management may result in agency costs in the future in the form of regulation. Therefore, regulators may impose new laws to force managers to provide a piece of information about the bad news. As a result, managers will try hardly to select carefully bad news through voluntary disclosure in order to avoid agency costs of regulations.
According to agency theory, mangers are expected to provide disclosures which can be understood as they are providing good performance and further more they seem to be acting in the interests of shareholders. They do this in order to reduce agency costs (Watts & Zimmeman 1978). Consistent with this, managers and according to legitimacy theory use CSED to represent companies' activities legitimation and compliance within the norms of their respective societies (Deegan & Rankin 1997). In another case, mangers try to reduce effects of bad news that might be unfavourable to the image of their companies (Deegan et al 2000).
In terms of the provision of quality (monetary) information, the both LT and AT are expected to have some explanatory power. In legitimacy theory point of view, firms provide some quality social and environmental information when the political-economic status maintains a good environmental for accounting communications (Massoud 2009). According to agency theory, firms with higher agency costs disclose more monetary social and environmental information. Although quality (monetary) information is linked to higher proportion costs, quality disclosure can in an indirect way reduce the agency costs by reducing information asymmetries (Massoud 2009).
3.2 Developments of Hypotheses
An objective of this study is to determine how corporate governance variables affect a firm's disclosure behaviours. Four corporate governance variables are examined in the present study. These are composition of non-executive directors, chairman with multiple directorships, ownership structure by foreign shareholders, managerial ownership, block holder ownership and the existence of an audit committee. Whereas, the five company characteristics are; size, profitability, liquidity, leverage, and audit reputation.
The hypotheses about the effect of governance mechanisms and company-specific characteristics in the current study are mainly predictions of a positive association as the two positive theories agency theory and legitimacy theory are considered.
3.2.1 Corporate Governance variables:
Corporate governance mechanisms are introduced to control the agency problem and ensure that managers act regarding to the interests of shareholders. These apparently conflicting viewpoints on the impact of corporate governance have not been totally resolved. Hill (1999) emphasises that no one single mechanism is a governance panacea and suggests that "it is desirable to have a system of overlapping checks and balances." Therefore, the hypotheses about the effect of governance mechanisms in the current study are mainly predictions of a positive association.
184.108.40.206 The proportion of independent nonexecutive directors on board
One major role of boards is its control functions (Pound 1995). Outside (independent) nonexecutive directors are perceived as a tool for monitoring management behaviour (Rosenstein and Wyatt 1990), resulting in disclosing more information to outside investors.
Both Lefwich et al. (1981) and Fama and Jensen (1983) argued that the larger the proportion of independent non-executive directors on the board, the more effective it will be in monitoring managerial opportunism. Therefore, there will be more voluntary disclosures. Forker 1992) found that a higher percentage of this variable enhanced the monitoring of the financial disclosure quality (Cook & Haniffa 2005). In other words, the benefits of withholding information are reduced. In fact, non-executive directors can put pressure on firms to disclose more on CSED in terms of ensuring congruence between organizational legitimacy and its actions. Therefore, and as a legitimacy theory explained, there is a positive relationship between outside directors and CSED (Cook & Haniffa 2005). Moreover, Chen and Jaggi (2000) found a positive relationship between proportion of independent directors and CSED. It is hypothesized that, the proportion of non-executive directors is positively associated with the level of CSED.
H1: Companies with a higher proportion of independent nonexecutive directors are more likely to have a higher extent of corporate social and environmental disclosure.
Although there are some studies as Wong and Ho (2001) and Stock Exchange of Hong Kong (SEHK) requirements emphasize the number rather than the proportion of independent non-executive directors to total directors, the use of a proportion was used in the this study. This is because independent non-executive directors might not apply a sufficient monitoring power especially if their numbers only account for a small percentage of board membership.
220.127.116.11 The chairman of the board
Companies that have one individual who is working as both chairman and chief executive officer/ managing director are considered to disclose more social and environmental information (Molz 1988). The person who occupies both roles would tend to withhold favourable information to outsiders. According to Fama and Jensen (1983), any adverse consequences could be eliminated by market discipline. But Forker (1992) stated that a dominant personality in both roles poses a threat to monitoring quality in terms of the quality of CSED. Another argument in favour is that members of other board may offer comparisons derived from personal knowledge of other organisations (Dahya et al., 1996). Therefore, decisions at one board can be part of the raw material to decisions-making at other boards. Further more, there was evidence that board members who have participated in various structural changes on other boards brings those beliefs to advocate changes in another board (see: Mizruchi, 1992; Haunschild, 1993). There is a positive relationship between chairman with multiple directorship and CSED drawing from legitimacy theory (Cooke and Haniffa 2003; Cooke and Haniffa 2005, Stulze et al 2007). Cooke and Haniffa (2005) found a significant positive relationship between corporate social disclosure and chairman with multiple directorships in Malaysia. Hence, it is hypothesized that:
H2: Companies with a chairman having multiple directorships are more likely to have a higher extent of corporate social and environmental disclosure.
18.104.22.168 The existence of an audit committee
The role of an audit committee is to ensure the quality of financial accounting as well as control system (Collier 1993). In fact, an audit committee consists mainly of nonexecutive directors. Therefore, it has influence to reduce the amount of information withheld. It is predicted by agency theory that, the audit committees are used as a means of attenuating agency costs (Shun & Ho 2001). Forker (1992) emphasized that; the existence of audit committees can improve internal control. Resulting from this, this factor can be used as an effective monitoring device for improving disclosure quality. He found a positive relationship between the disclosure of the audit committee and the quality of share-option disclosure for UK companies. This positive relationship is also supported by McMullen (1996) as he found an association between the presence of audit committees and more reliable financial reporting. Shun and Ho (2001) indicate that, the existence of an audit committee is positively related to the extent of voluntary disclosure. Therefore, it is hypothesized that:
H3: Companies that have an audit committee are more likely to have a higher extent of corporate social and environmental disclosure.
22.214.171.124 Ownership by foreign shareholders
The legitimacy gaps can be raised by ownership structure of the firm. In fact, different shareholders demand different disclosures and the demand is greater when the shareholders are foreigners (Craswell and Taylor, 1992). Since a substantial fund in the UK capital market comes from foreign investors, higher disclosure of information, including social and environmental information may be expected (Cooke and Haniffa 2005). According to legitimacy theory, there is a positive relationship between the CSED and foreign shareholders. This theory gives an explanation which is that, foreign shareholders have a greater demand than other kind of shareholders. This is due to the separation between management and owner geographically which holds a high proportion of shares. Thus, it is hypothesized that:
H4: Companies that dominated by foreign shareholders are more likely to have a higher extent of corporate social and environmental disclosure.
126.96.36.199 Managerial ownership
Managerial ownership can be defined as a percentage of ordinary shares held by the CEO and executive directors, and contains their deemed interests. In fact, there is a greater agency problem when managerial ownership is low. In other words, the manager has greater incentives to consume perks and reduced incentives to increase and maximize company's performance. Therefore, outside shareholders will increase monitoring of manager's behaviour in order to reduce the agency problem (Jensen and Meckling, 1976). According to this, it can be understood that, monitoring by outside shareholders maximizes costs of the company. However, monitoring by outside shareholders could be reduced if managers provided voluntary disclosure. This variable is explained by agency theory when there is an agency problem released. Ruland et al (1990) found that, managerial ownership has a negative relationship with voluntary disclosure. Mak and Eng (2003) found that, lower managerial ownership is associated with increased disclosure. Hence, it is expected that there is a negative association between managerial ownership and the level of CSED.
H5: Companies with a higher managerial ownership are more likely to have a lower extent of corporate social and environmental disclosure.
188.8.131.52 Block holders ownership
Block holder ownership is defined as the percentage of ordinary shares held by substantial shareholders (that is, shareholdings of 5% or more). When share ownership is diffused, more monitoring is required. This variable is viewed by an explanation from agency theory. According to prior studies, there is an empirical evidence indicates a negative association between block holder ownership and disclosure (McKinnon and Dalimunthe, 1993; Schadewitz and Blevins, 1998). The recent study of Mak and Eng (2003) also emphasised the negative association between this variable and CSED. It is therefore expected that there is a negative association between block holder ownership and voluntary disclosure.
H6: Companies with a higher block holder ownership are more likely to have a lower extent of corporate social and environmental disclosure.
As it can be seen from figure (3), there is a list of independent variables related to corporate governance. This figure represents these variables, their lables and expected relationship and sign in the regression.
3.2.2 Control variables (company-specific characteristics)
Several studies attempted to address the degree of disclosures in financial annual reports and the impact of various corporate characteristics on annual report disclosures. According to these studies, company characteristics have an impact on the level and quality of CSED. This impact can be either positive or negative. This study will examine the impact of five company characteristics which are; size, profitability, liquidity, leverage, and audit reputation on CSED.
184.108.40.206 Firm size
There are several studies have found a significant positive association between firm size and CSED in the corporate annual reports in both; developed and undeveloped countries (Andrew, J et al 2006). One explanation for the association is that large companies undertake more activities and have more operations. Therefore, the large companies have a greater impact on society (see: Trotman and Bradley, 1981; Teoh and Thong, 1984). Larger firms are subject to greater scrutiny by different groups in society and therefore would be under greater pressure which has encouraged these firms to disclose their social and environmental activities to legitimise their business (Cooke & Haniffa 2005). According to agency theory, social and environmental responsibility disclosure in larger companies can be used to reduce political costs which reduce wealth of a company. Since the magnitude of political costs is highly dependent on company size, the relationship between firm size and CSED is significantly positive (Andrew, J et al 2006). Thus, it is hypothesized that:
H7: Larger companies are more likely to have a higher extent of corporate social and environmental disclosure.
Profitability was used by several studies as an explanatory variable for difference in disclosure level. In addition, there are additional costs associated with CSED, and the profitability of the reporting firms is depressed (Andrew et al 2006). A possible explanation for positive association is that management has enough flexibility to undertake more extensive social responsibility activities to shareholders. Profitable firms tend to disclose social and environmental information in order to demonstrate their contribution to society (see; Cooke & Haniffa 2003, and 2005). In other words, profitable companies legitimise their existence.
Agency theory also suggests a possible relationship between CSED and profitability. Inchausti (1997) argues that managers of very profitable companies disclose more information to support the continuance of their positions and compensation arrangements. Further, (Ng & Koh, 1994) emphasized that more profitable companies will apply self-regulation mechanisms in order to avoid external regulation. In other words, financial ratios can be one form of such disclosures. (Maeston et al 2002). Hence, the hypothesis is stated as following:
H8: Profitable companies are more likely to have a higher extent of corporate social and environmental disclosure.
Leverage variable is measured as total liabilities divided by total assets. When Naser and Wallace (1995) tested the determinants of CSED in Hong Kong during 1991, the leverage ratios were low. Higher financial leverage is expected to reduce disclosure because leverage helps control the free cash flow problem. In fact, the agency costs of debt are controlled through restrictive debt covenants in debt agreements rather than increased disclosure in annual reports (Jensen, 1986). Agency theory would predict a negative relationship between financial leverage and disclosure although not necessarily disclosure of gearing ratios. When a company borrows, because of the divergence in interest between creditors and management. Resulting from this, further agency costs are incurred (Marston et al 2002).
As it has been explained by agency theory, there is a negative relationship between the value of leverage ratio with corporate social and environmental disclosure (Mak & Eng 2003). Thus, the hypothesis of this variable is stated as:
H9: Companies with a higher value of financial leverage are more likely to have a lower extent of corporate social and environmental disclosure.
Liquidity ratio assesses how well placed a firm is to meet its short term financial obligations. In other words, when the liquidity ratio is high, the better placed is the firm to meet its short term financial commitments. (Altman 1968). It is therefore possible that those companies in a secure financial position will be willing to disclose this to outsiders.
However, it has also been suggested (Wallace et al., 1994) that companies with weak liquidity may disclose more social and environmental information in order to provide more details of its "weak" performance to re-assure investors. Agency theory predicts that agency costs will be higher when the proportion of debt is greater (which could be reflected in liquidity ratio). Therefore, the higher the liquidity, the fewer ratios will be disclosed (Marston et al 2002).
Belkaoui & Kahl (1978) found a positive relationship between liquidity and disclosure. Further more, Williamson (1984) found significant differences between the current ratio value of reporting and non-reporting firms. However, Wallace et al (1994) found a negative relationship between company liquidity and disclosure. They suggest that, firms viewing low liquidity as "bad news" which gives greater details as part of their accountability to the users of an annual report. On the basis of these results, the tenth null hypothesis is:
H10: Companies with a higher value of liquidity are more likely to have a lower extent of corporate social and environmental disclosure
220.127.116.11 Auditor reputation
There is a number of studies have examined the relation between the characteristics of the audit company in terms of its size (whether the audit firm is one of the Beg-Four or not) and the extent of CSED. It was found that, there is a positive association between the audit firm size and the level of CSED. In fact, the important responsibility of auditors is to recommend their client firms and encourage them to practice the social responsibility of accounting practices/activities (Choi 1998). It is expected that, firms which have contract with larger audit firms (one the Big-four) are more likely to disclose more social and environmental information. This variable is used in the regression as "1" a company is audited by one of the Big-four and a "0" otherwise (Andrew et al 2006). The following hypothesis has been tested the auditor reputation variable:
H11: Companies that are audited by one of the Big-Four are more likely to have a higher extent of corporate social and environmental disclosure.
To be clear, figure (4) represents a list of independent variables related to control variables (company-specific characteristics). This figure represents these variables, their labels and expected relationship and sign in the regression.
Table (1) summarizes the linkage between theories and expected relationships between independent variables and CSED. This table includes both; corporate governance variable as well as control variable (company-specific characteristics).