Factors influencing corporate social responsibility disclosure (CSRD) by Libyan state-owned enterprises (SOEs)

Abstract This paper examines the relationship between selected firm-specific variables and the extent of corporate social responsibility (CSR) disclosures by Libyan state-owned enterprises (SOEs). The initial sample in this study consists of the 310 annual reports from 95 firms over a period of eight years from 2010 to 2018 inclusive. A content analysis approach of the firms’ annual reports is applied using a CSR disclosure index based on the Global Reporting Initiative guidelines. The factors in this study are analysed against the extent of CSR disclosures by SOEs using multiple regression. The main finding from the content analysis indicates that the extent of CSR disclosures made by SOEs in their annual reports is mostly descriptive, with charity and donation being the most disclosed items. The main findings from the correlation and regression analyses show that four of the eight proposed independent variables – firm sizes, firm age, type of sector, and CSR responsibilities—are statistically significant and positively related to the dependent variable CSR disclosure. With respect to the profile of SOE managers, those who are exercise responsibility in areas of CSR appear to have a significant influence on information disclosure in this respect, perhaps because their profession makes them more aware of such issues. The paper offers significant contributions to the extant literature by examining whether there are any effects of CSR disclosure practices by SOEs and expands the pool of CSR knowledge in the developing country context, especially in the public sector. This study provides fruitful implications for policy-makers and practitioners about state holdings, which may either hinder, or enhance corporate social performance.


PUBLIC INTEREST STATEMENT
Today, CSRD is considered a major topic in academic literature and an essential business practice along with significant debate remains about the relationship between CSRD and characteristics of the state-owned enterprises (SOEs). The results show that the extent of CSR disclosures made by SOEs in their annual reports is mostly descriptive, with charity and donation being the most disclosed items. The added value of this paper is to fill a research gap in the emerging field of the disclosure of quality information in the annual reports of statutory bodies by entities of the public sector, such as SOEs. The value of this study strengthens arguments which hold that the implementation of CSRD, even implemented by SOEs, functions more as corporate survival mechanism rather than for solving social and SOE managers appear to have a significant influence on information disclosure in this respect.

Introduction
Corporate social responsibility (CSR) is not a new concept, it means ethics, distinguishing right from wrong, and doing right. From a historical point of view, since the 1930s the social responsibility of business was not widely considered to be a significant problem from Adam Smith's time to the Great Depression during the 1930s and 1940s (Moon & Parc, 2019). As a result, in the late 1940s and during the 1950s, the primary developing body of knowledge on CSR was determined by the business organization, government policy, and society in general of social responsibility (Lafferty, 1996), as a means of doing good deeds for society (Carroll & Shabana, 2010) and is now an aspect of growing importance in many academic and research circles (Dentchev, 2005). CSR was defined in 1953 by Bowen as "the obligations of business to pursue those policies, to make those decisions or to follow those lines of action which are desirable in terms of the objectives and values of our society" (p. 6). Likewise, "CSR has been represented as an umbrella term covering a diverse range of issues which have grown steadily in importance for business performance at a global level" (Masoud, 2017, p. 1).
Though there are a large number of studies which have shown that participation in CSR activities can bring economic benefits to the firm, such returns are often of a long term and uncertain nature (Nikolaeva & Bicho, 2011). Given the special characteristics of Libyan firm's background, we are firstly interested in their CSR reports to release behavior Libyan firms. Moreover, CSR activities can be separated into CSR disclosure (CSRD). Thus, CSR disclosure is considered the main communication tool for stakeholders of firms regarding CSR activities (Belal & Cooper, 2011). However, most of the literature assumes that the CSR disclosure is full information and can reflect the actual CSR (Liao et al., 2018) while the actual credibility and integrity of CSRD are relatively low (Luo et al., 2019;She & Michelon, 2019).
A key feature of the issue of CSR in the field of public enterprises has been addressed in various studies (Córdoba Pachón et al., 2014;Roper & Schoenberger-Orgad, 2011) mostly published since 2007, although some date back to 1990. The primary function of SOEs is to fulfill the social objectives for which they were created (Cunningham, 2011;Heath & Norman, 2004) and not necessarily to obtain financial profits (Cunningham, 2011), although they must assure their own sustainability. However, with respect to SOEs, the objectives of CSR should go beyond considerations of financial indicators and utility.
Within the Libya context, the issue of CSR emerged in the early 2000s (Pratten & Mashat, 2009) due to pressures from stakeholders for information which may influence organizational performance for Libyan firms. This has led to the development of the concept of CSR disclosure. Business corporations in Libya, either owned by domestic and foreign investors, particularly the State-Owned Enterprises (SOEs), are legally liable to the provisions of CSR laws. Furthermore, in Libyan firms there is a lack of concern among regulatory bodies regarding CSR disclosure; this may be because they feel that CSR disclosure is not relevant for investors. Therefore, this paper aims to examine the practice of CSR disclosure in a developing country, particularly Libya.
The main contribution of our paper is to extend the knowledge about CSR disclosure to the field of SOEs in the context of an emerging economy (Libya) where state-owned enterprises still plays a significant role. This type of paper of literature revision allows to have a previous situation of the research, the theoretical framework that has been used, and to specify the issues studied. It was documented in prior studies most have focussed on private sector firms, with particular emphasis on business-related matters. However, little empirical research has been conducted into CSR disclosure by public-sector firms, particularly as regards public enterprises. For these explanations, the present study examines the situations of public enterprise disclosure in the Libyan context. The study highlights the methodologies and the main profiles of the principal researchers to be of significance to CSR. In addition, emerging trends, gaps, and areas of future research could be identified. This study also tends to provide an important implication for the corporate sector that, after the shift in SOEs, firms need to look at strategies to maintain the status of CSR disclosure. For that purpose, firms shave to adopt the policies of SOEs regarding CSR disclosure because they may have a different working environment.
The remainder of the paper is structured as follows. The next section presents the institutional background of the study. Section 3 explains the theoretical framework, followed by a fourth section aimed at explaining the variables and hypotheses. Section 5 presents the sample, data collection, and methods. Next, we report the results followed by the discussion section and the limitations and suggestions for future research.

Institutional background on CSR disclosure in Libya
Libya, as a developing country, was chosen by the researcher as it has an important standing in the world economy and unique changes over a short period of time in terms of economic, environmental, and social changes. Changes in the regulatory environment may have an impact on firms in terms of their disclosures. Since 1999, the Libyan government has gradually been implementing measures to reform and following an open-door policy. The efforts are aimed at changing the country's state-controlled economy to a more market-oriented one in order to integrate it into the global economy, restructuring the economic mechanism from central administrative planning to mostly market-driven pricing and exposing the economy to the world through trade (expert oil, and imports) and foreign investment (Masoud, 2013). Since 2000 the level of CSR disclosure in Libya has increased (Pratten & Mashat, 2009) due to pressures from stakeholders for information which may influence organizational performance for Libyan firms. This has led to the development of the concept of CSRD. There is the idea that governments represent public interests and SOE managers should be made accountable to identify and cater to them (Matten & Crane, 2005). Stakeholders can also have different degrees of influence in organizational action (Mitchell et al., 1997) and therefore they play a key part in supporting or undermining CSR initiatives (Costa & Menicini, 2013;Calabrese & Lancioni, 2008;Maclagan, 2008). Firm disclosure practices, within the Arab world context, in which Libya constitutes an important part, there is still a paucity of empirical studies on CSR practices to various stakeholder groups (e.g., Al-khater & Naser, 2003;Jahamani, 2003). Thus, CSR was understood as a philanthropic work, a noninstitutionalized phenomenon based on traditions presented by the Arab norms, and religious beliefs, a strong culture of giving, Islam recommends social co-operation (Masoud & Halaseh, 2016).

Theoretical framework
CSR disclosure is interpreted in many ways by different authors and theories that attempt to explain what it is are just as diverse, the most widely accepted, and are the ones this research will focus on; these are, stakeholder theory and legitimacy theory (De Villiers and Van Staden, 2006;O'Dwyer, 2003;Milne, 2002;Gray et al., 1995). The theoretical framework adopted incorporates both influences, by adopting institutional theory perspectives, specifically stakeholder theory focuses upon the expectations of particular interest groups, and legitimacy theory focuses upon the expectations of society in general. Over the last decades, firms have come under growing pressure to actively pursue CSR efforts from a variety of stakeholders (Brammer et al., 2007). In the context of a firm, society can be grouped as shareholders, employees, customers, creditors, and suppliers, who may be interested in the firm's social and environmental activities. Freeman (1984) identified these groups as "stakeholders." Carroll (1991, p. 43) states that there is a natural fit between the idea of CSR and an organization's stakeholders. From a stakeholder theory perspective (Cahaya et al., 2015;Hamidu et al., 2015;Freeman et al., 2007;Freeman et al., 2010;Freeman, 1984), which suggests that CSR includes managing multiple stakeholder ties concurrently, but other stakeholders will benefit from firm's activities as well, "[. . .] the goal of any firm is or should be the flourishing of the firm and all its principal stakeholders" (Werhane & Freeman, 1999, p. 8). In addition, the stakeholder theory attempts to explain how a firm identifies the powerful stakeholder groups within the social dynamics who may affect or be affected by, the firm's CSR disclosure practices, and how the firm responds to their expectations thus winning their support and ensuring the survival of the firm (Khlif et al., 2015;Gray et al., 1996).
The institutional perspective of Legitimacy theory has become one of the most cited theories within the CSRD and environmental accounting area (see, for instance, Patten & Crampton, 2004;Deegan, 2002;Neu et al., 1998). Suchman (1995, p. 574) defines legitimacy as " [A] generalized perception or assumption that the actions of an entity are desirable, proper, or appropriate within some socially constructed system of norms, values, beliefs, and definitions", or "is a social judgment of appropriateness, acceptance, and desirability" (Zimmerman & Zeitz, 2002, p. 418). It would suggest that legitimacy theory attempts to explain why a firm makes CSR disclosure, and therefore becomes a resource that a firm can create, influence, or manipulate through various disclosure-related strategies for acquiring and managing legitimacy (Deegan, 2002;Woodward et al., 2001;Lindblom, 1994) and as a means of protecting an organization's reputation and identity (Panwar et al., 2014;Cho & Patten, 2007;Campbell, 2003;Hooghiemstra, 2000). These strategies vary from changing goals, outputs, and methods to changing perceptions about the firm's goals, outputs, and methods. Harte and Owen (1992), and Deegan & Rankin (1996) also figure out the downside of legitimacy theory is that it may encourage business managers to put out positive information (i.e., only information that will guarantee stakeholder approval). Hybels (1995, p. 243) argues that good models in legitimacy theory must examine the relevant stakeholders, and how "Each influences the flow of resources crucial to the organizations' establishment, growth, and survival, either through direct control or by the communication of goodwill." In line with the above discussion, empirical tests in this study consider the influence of the dissemination of CSR information, including the organization's size, type of sector, institutional ownership, and age of operation (Jenkins & Yakovleva, 2006;Haniffa & Cooke, 2005;Maignan & Ralston, 2002;Adams et al., 1998;Hackston & Milne, 1996;Gray et al., 1995). The personality characteristic of the managers responsible was also related to social responsibility issues. In this regard, Kelly et al. (2001) observed how variables related to age, gender, education level, and CSR responsibilities were directly related to ethical issues in business. However, several research studies conducted to ascertain the specific characteristics of managers have led to divergent findings. Different authors continue to maintain different viewpoints. These are: (a) personal characteristics including age and gender of individual managers, are irrelevant; (b) individual effort or creativity, and are a matter of choice, can exert a direct influence on perceptions of CSR; and (c) other structures include a person's work experience, religious beliefs, education, and training (Quazi, 2003).
Following this theoretical framework, it is clear that legitimacy requires a reputation that must be retained. It requires a firm to convince its relevant publics that its activities are congruent with their values and that of perspective. Thus, these perspectives can be explored by using stakeholder theory insights. On the other hand, organizational legitimacy and organizational reputation are inextricably linked, and in this study the distinction between the two will not be explored further.

Empirical literature review and hypotheses development
In what follows, explanations for CSR disclosure based on the theoretical framework presented in the previous section are developed by selecting the most relevant factors influencing CSRD.
Variables are chosen to represent particular aspects of social visibility, and in each case, an expectation regarding its relationship to CSR disclosure is stated based on prior literature.

Firm size
As an issue of major public concern, the legitimacy theory points out, organizations should take into account their public responsibilities and observe the social contract (Reverte, 2009). However, the idea that larger firms have a greater social reputation and the agency cost is higher (Jensen & Meckling, 1972) because, as firms grow, they attract more attention from stakeholders and need to respond more openly to their demands (Hillman & Keim, 2001). Particularly, from an empirical standpoint, a number of studies over the past decades tests the influence of firm size on the level of disclosure. Most researchers have found a positive relationship between firm size and the disclosure of CSR information in both developing and developed countries (Reverte, 2009;Oyelere et al., 2003;Cooke, 1992;Singhvi & Desai, 1971). Similarly, a number of empirical studies found that there is a positive association between the size of the firm and corporate social disclosure (Hossain, 2008;Bozzolan et al., 2003).  concluded that firm size is positively associated with a significant change in the amount of CSR disclosure.
According to legitimacy theory, the larger the organization, the greater the volume of information it publishes, in order to enhance its image and reputation. Small firms may not be able to afford such costs from their resource base. This study measured by the logarithm of employees as the firm size variable, because it is an important variable that has often been positively related to social performance (McWilliams & Siegel, 2000;Waddock & Graves, 1997). The following hypothesis can thus be presented based on this idea: H 1 : There is a positive significant relationship between SOE size and the level of CSR information disclosed.

Firm age
This variable refers to the length of time a firm has been in business. The legitimacy of an organization is increased when it reflects acceptable and desirable norms, standards, and values (Suchman, 1995). According to the stakeholder theory, as firms grow older, they are expected to establish relationships with a more widespread group of stakeholders. With this increasing spread of stakeholders, the need for social responsibility disclosure as a means of communicating with stakeholders' increases can participate in a variety of ways, then would provide more or better resources and returns for the organization's successful functioning and survival (Parsa & Kouhy, 2008;Donaldson & Preston, 1995). Delaney and Huselid (1996) found a positive relation between CSR information and firm age. Though older firms have had more time and gained more experience in the development of CSR policies and in an information disclosure to stakeholders, which reflect the public for longer. However, some empirical studies found that there is a negative association between the amount of CSRD and firm age. For instance, Rettab et al. (2009) found a negative relationship between CSR and firm age, while Liu & Anbumozhi (2009) also found a negative relationship between environmental disclosure and firm age. In this study, it can be argued that older firm's SOEs may have improved their disclosure practices over time as they have been subject to the demands of stakeholders over a longer period of time. This study measured firm age by the logarithm of the number of years the SOE since founding, because the maturity of a firm may affect its social responsibility activities (Moore, 2001;Roberts, 1992). Taking into account these considerations, we propose the following hypothesis: There is a positive significant relationship between old firm's SOEs disclose more that in the level of CSR information than younger SOEs.

Type of sector
Another commonly used proxy for social visibility is a type of sector. According to legitimacy theory, corporations from environmental sensitive industries with high public visibility and a significant potential for environmental impact tend to disclose more socially responsible information (Reverte, 2009;Brammer & Pavelin, 2008;Patten, 2002;Adams et al., 1998;Hackston & Milne, 1996. Deegan & Jeffry (2006), and Deegan (2009) argue that if some firms in the industry are disclosing CSR information; then, other firms operating in the same industry would disclose information about their CSR activities in order to look similar to other firms. Previous results of empirical studies indicate that level and type of disclosure are significantly different when firms are from different industries. For instance, several empirical studies have found a positive and significant relationship between CSRD and type of industry (Wanderley et al., 2008;Castelo & Lima, 2008;Newson & Deegan, 2002;Gray, 2002;Cowen Linda & Scott, 1987). The following hypothesis can thus be put forward based on previous ideas: There is a positive significant relationship between SOE sector and the level of CSR information disclosed.

Institutional ownership
According to the theory of legitimacy, the actions of SOEs should respond to the expectations of their stakeholders. The ownership structure or degree of dependence of an organization, can be a determinant factor of a few large investors or dispersed among many can affect disclosure policy Cormier et al., 2005;Ullmann, 1985;Roberts, 1992), as firms conducting CSR are explained to be more attractive in the eyes of investors and especially institutional investors (Coffey & Fryxell, 1991). Thus, ownership concentration and institutional shareholding are used to measure the ownership effect on CSR disclosure. Previous studies have stated a significant association between firms' voluntary disclosure of information and their ownership structure (Eng & Mak, 2003;Hossain et al., 1994). Although a negative relationship between block-ownership and disclosure is reported in the previous research (Kelton & Yang, 2008;Schadewitz & Blevins, 1998;Mitchell et al., 1997). It can be argued that higher degree of public ownership would be expected to encourage management to react positively to social pressure, by increasing the level of CSD to acquire owners' satisfaction, and consequently the negative association between CSD and blockownership. The following hypothesis can be presented based on these assertions: There is a positive significant relationship between institutional ownership of SOEs and the level of CSR information disclosed.

SOE age
According to legitimacy theory, in public organizations managers tend to have longer tenure, seniority is respected and more experienced public administrators have a greater insight into the process of performance improvement. Based on theories of identity, age is a variable that is associated with the personal characteristics of individuals' ethics and social characteristics (Erickson, 1968). In principle, "advancing age could be commensurate with a broader perspective on the issues and entities surrounding decision-making, including multiple stakeholders and thus stronger CSR values" (Waldman et al., 2006, p. 831) because older managers may be more conservative in their ethical attitudes. Moreover, age provides managers with greater experience and thus an accumulation of skills and abilities (Herrmann & Datta, 2005). However, younger managers are less averse to risk and the opportunities offered by new changes in technologies, products, and business and are often more strongly oriented toward issues of social responsibility within the firm (Post et al., 2011). The following hypothesis can thus be put forward based on previous ideas: There is a positive significant relationship between old firm's mangers in SOEs disclose more that in the level of CSR information than younger ones.

Gender
According to stakeholder theory, firms involving more females on board and senior management levels reflect protecting the interests of various stakeholders. Thus, the relationship between gender and business ethics receives the most attention and is widely researched. Among many empirical researches indicate females are more ethical than males (Atakan et al., 2008), there is a strong distinction in terms of leadership styles and behavior (Hooijberg & DiTomaso, 1996). Gill (2010) indicates female business students are more ethically predisposed than their male counterparts. It is further observed that mixed results on the effects of having women directors on CSR disclosure. Jia & Zhang (2012), Giannarakis et al. (2014) and Wu et al. (2015) found a positive relationship between women directors and CSR disclosure. Other studies have found a negative effect of having women directors on firms' CSR disclosure (see, for instance, Muttakin et al., 2015;Shamil et al., 2014). On the other hand, Bear et al. (2010), Fernandez-Feijoo et al. (2014, and Sundarasen et al. (2016) found that having women directors enforce a positive influence on firms' CSR disclosure only when there is a higher representation of women directors on boards. However, empirical studies have reported conflicting and often confusing results on CSR disclosure (Glass et al., 2015;Giannarakis, 2014;Kahreh et al., 2014;Bowrin, 2013). The following hypothesis can be presented based on these assertions: H 6 : There is a positive significant relationship between percentage of female of SOE managers and the level of CSR information disclosed.

Education
Prior studies suggest that managers with higher educational degree affects their ability to deal with the external environment, communication, and coordination, information processing, innovation, tolerance, and knowledge (Bantel & Jackson, 1989;Hambrick & Mason, 1984) therefore, they are more capable of making decisions under complex environments (Usdiken, 1992;Kimberly & Evanisko, 1981).  stated that top managers with higher education have more opportunities to become involved with and manage external activities in comparison with managers with lower education. Furthermore, managers with a high level of education with relevant training will have better information processing and integration capability to improve CSR performance, in the view that promoting innovation and devising new ideas and solutions requires both knowledge and experience (Damanpour & Schneider, 2009;Lee et al., 2005;Quazi, 2003) because appropriate training for managers can equip them with the skills and knowledge necessary to deal with the social, ethical, and environmental impacts of business activities (Brampton & Maclagan, 2005). However, a lack of knowledge and awareness of issues related to the area of social responsibility can result in scant or poor disclosure of CSR information (Belal, 2001) due to the training of future managers in questions of social responsibility is of crucial importance for the appropriate disclosure of such information (Lungu et al., 2009;Al-Khater & Naser, 2003). For this reason, it is hypothesized that: There is a positive significant relationship between ethical business education of SOE managers and the level of CSR information disclosed.

CSR responsibilities
According to the modern theory of CSR, ethical and socially responsible initiatives constitute a legitimate and sustainable way to generalized perception or assumption that the firm's actions are desirable since these actions help to strengthen the relations with all stakeholders and contribute to improve conditions within the firm or in the business environment (Bhattacharya et al., 2008;Barnett, 2007). If the relationship with stakeholders is properly managed, costs and risks may decrease (Jensen, 2002) and competitive advantages may therefore arise (Porter & Kramer, 2006). In addition, doing what is right and responsible use of the firms' power increases their legitimacy (Sethi, 1979), prevents social sanctions, develops positive attitudes towards the firm, and its products (Sen et al., 2006), CSR would remain at a merely theoretical level within the firm (Rodrìguez Bolìvar et al., 2015;Córdoba Pachón et al., 2014;Pedersen & Neergaard, 2009;Caldwell et al., 2008). For this reason, these managers will perceive that ethics and social responsibility are essential with regard to the overall effectiveness of the firm. In this regard, we propose the following hypothesis: There is a positive significant relationship between CSR responsibilities of SOE managers and the level of CSR information disclosed.

Sample and data collection
The initial sample in this study consists of the 310 annual reports were analysed, consists of the 95 listed firms' annual reports over a period of eight years from 2010 to 2018 inclusive, using multiple regression analysis. The time of duration is selected for the following reasons: this period is the recovery period from the level of awareness and attitude among Libyan managers towards CSR practices that has just appeared, and therefore, it is the period of firms' involvements in CSRD. During this period, firms also started to addresses the interests and demands of its stakeholder's environmental concerns through appropriate CSR activities. Firms with no available complete annual reports or which were delisted during the sample period were excluded as well as firms that experienced significant mergers or reconstructions. This paper adopts a qualitative content analysis and quantitative correlation analysis to explain factors influencing CSR disclosure levels by Libyan state-owned enterprises (SOEs). The content analysis was conducted as a systematic method of categorizing and analysing the content of text from the annual reports of sample firms. Content analysis is one of the most used techniques for analysing and understanding the collections of text in qualitative research (Bayoud et al., 2012;Erlingsson & Brysiewicz, 2017;Gray et al., 1995;Mayring, 2014;Menassa, 2010). It is a method of codifying written text into various groups or categories based on selected criteria (Weber, 1988). Prior researchers suggest that content analysis provides valid results for CSR disclosures in both developed countries (Ali et al., 2017;Gamerschlag et al., 2011;Deegan, 2008) and developing countries (Ali & Frynas, 2018;Belal & Cooper, 2011;Hassan et al., 2006). Consequently, the data was coded first based on the framework (Table 1), and then with a refined category system formed by using a data-driven (abductive) approach (Flick, 2009). The entire process was performed iteratively. In other words, the texts of the CSR reports were analysed with NVivo 11.0 software, and when the category system was created, a process of manually going back-and-forth between the text and the framework was implemented.

Dependent variable
The disclosure index developed in this study will measure the extent of disclosure of quality information in the annual reports of statutory bodies. In other words, the study identifies the presence of the disclosure items in the annual reports. This means once an instance is found for a particular checklist item, the search would immediately stop for that item. The presence or existence, instead of frequency of occurrence, of data is applied. Secondly, the index adopts the dichotomous or binary method of scoring each disclosure item. This method is chosen instead of the polychotomous or qualitative scale method (Coy & Dixon, 2004;Hall & Steiner, 2020;Robbins & Austin, 1986) as this study is interested in determining the presence of information and extent of disclosure. The binary method is also consistent with many prior studies (Gore, 2004;S. Ismail & Abu Bakar, 2011;. Thirdly, the study adopted the unweighted approach for scoring each of the disclosure items therefore assuming that all items are equally important. This method was used in many previous studies (S. Ismail & Abu Bakar, 2011;Wei et al., 2008). For the purposes of this study, the level of the quality of information is defined asDIS i shown in Table 1. Several empirical studies in the area were of great utility in developing the CSRD index used (Azizul Islam & Deegan, 2010;Massa et al., 2015). CSRD refers in this study to nine categories reflecting the socially responsible information disclosed by SOEs. As in the study by Prado-Lorenzo et al. (2008), we considered each item through dummy variables, in which a value of 1 was given if the corresponding initiative was present in the firm and a value 0 if not. Under-unweighted approach, the disclosure score indexes for each category are constructed to take into account these considerations: where CSRD refers to the CSRD index expresses the level of disclosure for a firm i, DIS i related to "0" if the item does not disclose any information. It is assigned a value of 1 if a measure the (quantity or quality) of disclosure in the phrase j of the annual report of firm i. In other words, when the disclosure score index is equal to 0, it indicates that firm i does not disclose any item. Furthermore, index values equal to i = 1, . . . , id i = number of considered in the annual report of firm i. In order to assess the level of the quality of CSR reports, nine criteria were identified in the category of the relevance of information disclosed by SOEs. By adopting this framework, it is possible to calculate an index of disclosure quality. The quality assessment CSR disclosure is shown in Table 1, which combines different types of information which Bouten et al. (2011) argue can be used as a valuable tool for assessing the overall quality of a firm's reporting practice. It is assigned a value of 1 if there is the information of subcategories (items), whereas a score of 0 will be awarded if no information subcategory is disclosed. In other words, when the disclosure score index is equal to 0, it indicates that firm i does not disclose any item. Furthermore, index values equal to i = 1, . . . , m means that a level of disclosure is provided. In order to assess the level of the quality of CSR reports, nine criteria were identified in the category of the relevance of information using (0/1) scoring system, to avoid any absence or presence of each item (Jones et al., 1998). In the quality assessment CSR disclosure, nine criteria has been identified as shown in Table 1, which combines different types of information. Bouten et al. (2011) argue that it can be used as a valuable tool for assessing the overall quality of a firm's reporting practice.

Independent variables
This section concerns the independent variables employed in this study. Existing literature provides models and scales to explain the variation of CSRD across various factors including; the size of the SOE, measured by the number of employees, as the value of 0 if it was 0-299 employees, and the value of 1 if it was 300-600 employees or more because it is an important variable which has often been positively related to social performance (Castelo & Lima, 2008;McWilliams & Siegel, 2000), as firms grow, they attract more attention from stakeholders and need to respond more openly to their demands (Hillman & Keim, 2001). Firm age (CAGE), measured by the number of years, as 0 if it was 25 years or less, 1 if it was between 25 and 50 years old or more, because the maturity of the firm may affect its social responsibility activities (Gallego et al., 2011;Castelo & Lima, 2008;Moore, 2001). SECTOR refers to the type of industry that was measured by a dummy variable that takes the value of 1 if the SOE operates in a sensitive environmental sector such as manufacturing, oil and gas, transport, construction, tourism, food, and the value of 0 if otherwise (Castelo & Lima, 2008;Elsayed & Hoque, 2010;Patten, 1991). Institutional ownership (OWN) is the proportion of public ownership of the firm, as a higher degree of public-ownership would be expected to be more committed to CSR (Ghazali, 2007;. Another variable concerns the types of SOE profile firm managers were analysed as the average age of the SOE manager (SOEAGE), measured as 1 if it was 35 years or less, 2 if it was between 35 and 50, and 3 if the SOE was 50 years old or more, because some argue that managers tend to give more priority to personal growth than wealth or advancement (Hall, 1976) and that they become more ethical attitudes (Quazi, 2003;Singhapakdi et al., 1999;Terpstra et al., 1993) as they grow older which can influence the disclosure of CSR information. Gender, measured as a dummy variable which takes the value 1 if the SOE manager is male, and the value of 0 otherwise (Safari Kahreh et al., 2014;Smith et al., 2010). Education (EDU) that was related ethical business education (degree) of the SOE manager with CSR subjects (Lungu et al., 2009;Maclagan, 2005), measured as dummy variable which takes the value 1 if the ethical business education of SOE manager, and the value of 0 otherwise, and CSR responsibilities (CSRR), measured as whose work involves CSR concerns that tend to be more affected by the public interest (Bernardi & Threadgill, 2010;Brunton & Eweje, 2010;Peterson, 2006;Quazi, 2003). The regression model is presented to analyses the relationship between the infusing factor and the CSR disclosure of the characteristics of the SOEs. Following previous studies, we test the relationships between the dependent variable, CSRD, and the independent variables used multiple regression analysis (Muttakin & Khan, 2014). Since the implementation of CSR by SOEs normally utilizes public budget allocations, as such public disclosure and public audit are judicially consequential. This study further applied robustness tests to address major endogeneity issues. Specifically, we use the following regression model based on the hypotheses proposed Figure 1 : where for the firm i: CSRD represents the dependent variables refer to the CSRD index (DIS1 . . . DIS9), while independent variable represents, such as its SIZE: firm size, CAGE: firm age, SECTOR: type of sector, OWN: institutional ownership, SOEAGE: age, GENDER: Gender, EDU: education, CSRR: CSR responsibilities refer to the workplace.β 0 is a constant whereas β 1 . . . β 8 are co-efficiency of each calculation of the independent variables, ε represents the error term and i indicates the firm number.

Descriptive statistics
In the context of this study, the areas that were least often disclosed concerned CSR management indicators (DIS1) with a mean value of 0.32, responsibilities for environments (DIS4) at 0.47, responsibilities for products and consumers (DIS6) at 0.38, and responsibilities for governance (DIS7) at 0.46 shown in Table 2. Similarly, the low levels of disclosure on environmental issues seem to be consistent with the findings of many studies and particularly those conducted in developing countries (Andrew et al., 1989;Belal, 2001;Imam, 2000;Rizk et al., 2008). However, content analysis results show that the amount of CSR disclosure is satisfactory in the developing country context. The results are inconsistent with those who found in their study that the level of environment disclosure was developed in Libya (N.S.M. Ahmad & Mousa, 2010). Our results also suggest that, according to the study by  as one of the major reasons for low social disclosure in developing countries, including those whose economies are in transition, may therefore be the lack of mandatory (i.e., statutory) disclosure requirements. Haniffa & Cooke (2002, pp. 237-238) agree that, if there is no public pressure function (which is the case in Libya) or laws and regulations governing the firms, firms are unlikely to provide high-quality information without this mandatory requirement. Table 3 provides descriptive statistics for the variables used in the study. The mean value is the most commonly used to measure central tendency. Results reveal that the mean value for the CSR disclosure variable is 4.1350, indicating that the average CSRD of the sample firms is about 44% (4/9). This outcome average is higher than the averages results done in developing countries for two samples of Saudi Arabia firms of 16% and 14.61% found by Macarulla and Talalweh (2012) and Al-Janadi et al. (2013), during 2008-2007. This signifies an improvement in disclose of CSR information which may be linked to the application of the public image. The mean value of SIZE is 0.6955, indicating that on average; SOEs are 70% suggesting that large firms are more likely to disclose CSR information. There is an increasing number of  firms that disclose employee relation information in their annual reports. The mean value of the firm age variable is 0.6686, implying that, on average; tend to be old SOEs disclose more CSR information, with 69% having been operating for more than 25 years. The mean value of the type of sector is 0.1267, indicating that, on average, only 13% operate in environmentally sensitive sectors. This implies that the type of sectors disclosing their environment is limited, while the mean value of institutional ownership of SOEs is 41%. As shown in Table 3, the mean value for SOE profile firm managers, their average age is in the range 35-50 years, although this is subject to considerable variability. For managers' agenda, the descriptive statistics show that most top managers for sample firms are male 59%, and have an average of 29% ethical business education possesses a bachelor's degree or above. Besides, the results show that, on average, in positions directly related to CSR responsibilities are 23%.
Reliability is measured by Cronbach's alpha (α) coefficients of the variables tested as illustrated in Table 3, were all above 0.80, which was considered satisfactory (Fornell & Larcker, 1981;Hair et al., 2010). According to Sekaran (2003), alpha coefficients less than 0.60 is poor, those in the 0.70 range as acceptable, and those over 0.80 is good. The results are consistent with that of Abdul and Ibrahim (2002), and Brønn & Vidaver-Cohen's (2008) studies on CSR, which yielded results of 0.61 and 0.98, respectively. Therefore, this shows that there is inter-item consistency reliability among the variables analysed. In addition, we checked the condition for discriminant validity among constructs, we used the average variance extracted (AVE). As suggested by Barclay et al. (1995), and Fornell and Larcker (1981) values of AVE higher than 0.5 indicate that the set of items has an appropriate convergence in measuring the concern construct, making the validity of indicators and the validity of the entire construct questionable. Based on the results of Table 3, the values of AVE of all the constructs were ranged between 0.60 and 0.65. Thus, the measures used have an appropriate level of convergent validity. Table 4 presents the Pearson correlation between the set of independent variables. The correlation shows no indication of multicollinearity as the correlations between independent variables do not exceed 0.8 (Guajarati, 1995;Judge et al., 1985). There is a high correlation between the dependent variable CSR disclosure (CSRD) with three of the eight proposed independent variables -firm size, type of sector, and CSR responsibilities-are positively and statistically significant at the 95% significance level (p <0.05) (0.498, 0.350, 0.382, respectively). These results mean that they are significantly positively correlated indicating that in this sample, as firm size, type of industry, and CSR responsibilities increases, the level of CSR disclosure also increases. An additional justification for this result could be that this correlation does not represent a serious multicollinearity problem between the independent variables because is lower than 50%. For instance, Hossain and Hammami (2009) do not consider a correlation of 52% to be a serious multicollinearity problem. Furthermore, Bryman and Cramer (2001) argue that the correlation between independent variables represents a serious problem only if exceeds 80%. Specifically, our results indicate that the examination of variance inflation factors (VIFs) indicates no evidence of multicollinearity between variables (Table 5). Table 5 highlights the results of multiple regression analysis are used for multivariate testing of the hypotheses in this research. The analysis of the factors that influences the disclosure of CSR information shows that four of the eight proposed independent variables -firm size, firm age, type of sector, and CSR responsibilities-are statistically significant and positively related to the dependent variable CSRD (β 1 =0.351, β 2 =0.822, β 3 =0.269, β 8 =0.247, respectively). Regarding the results of regressing the characteristics of the SOEs, the coefficient of the firm size is statistically significant at the 99% significance level (p <0.01) (β 1 =0.351, t-statistic =4.200), with regard to the characteristics of the SOEs, the influence of firm size on CSR disclosure has been studied extensively (Da Silva Monteiro & Aibar Guzmán, 2010). In line with earlier findings, this result is consistent with agency theory that larger firms need to disclose more information in order to reduce the larger information asymmetry and agency costs (Elzahar & Hussainey, 2012). Another possible explanation may be that larger firms are playing a major role in society, more likely to disclose higher levels of CSR information, and thus, our results show that the size of an SOE is positively related to the level of CSR information disclosed, which supports for hypothesis H 1 . Accordingly, Z. Ismail and Koh (1999) have found a negative significant relationship between the firm size and the extent of disclosure but most of the researchers such as Alam and Deb (2010), Akhtaruddin (2005), and Ahmed and Nicholls (1994), have found a significant positive relationship between the firm size and the extent of disclosure.

Empirical results and discussion
In the second hypothesis, the coefficient for the firm age is positive and statistically significant at the 90% significance level (p <0.10) (β 2 =0.822, t-statistic =2.670), indicating that older firms tend to disclose higher levels of social information than newer ones (Gallego et al., 2011;Murias et al., 2008). In line with earlier findings, this result is consistent with Delaney and Huselid (1996) who indicate that firm age has a positive relationship with CSR practices and CSRD. In another study, Alam and Deb (2010) reveal a positive association between the level of disclosure and firm age while the counter result has been got by Akhtaruddin (2005). Thus, these results confirmed the possible explanation could be that once these practices are implemented, stakeholder expectations increase, and the firm is forced to meet them and even reinforce them (Moore, 2001;Roberts, 1992), and they, therefore, provided strong support for H 2 .
Additionally, the results also indicate a positive and significant influence of the type of sector in which the SOE at the 95% significance level (p <0.05) (β 3 =0.269, t-statistic =2.382) is operating corroborates hypothesis H 3 . However, our result is inconsistent with previous studies that have rejected this hypothesis (Da Silva Monteiro & Aibar Guzmán, 2010). Therefore, the empirical results are consistent with the view that the level and type of disclosure are significantly different when firms are from different industries. This difference refers to stakeholder's pressure (Patten, 1991) & regulation imposed on some industries (Dierkes & Preston, 1977), in our model, the fact that a firm operates in an environmentally sensitive sector seems to be a differentiating attribute in its level of disclosure. The results also show a positive but statistically insignificant correlation between the institutional ownership of SOEs and CSR disclosure (β 4 =0.067, t-statistic =0.530), which implies that institutional ownership could not be deemed a determinant of CSR disclosure. This result is consistent with those of Elzahar and Hussainey (2012), and Eng and Mak (2003) who find an insignificant correlation between the two variables. Accordingly, this study rejects the relationship between institutional ownership of SOEs and the level of CSR disclosed; hypothesis H 4 .
As may be seen from Table 5 providing details of the second set of variables, regarding public manager's characteristics, the analysis shows that only responsibilities for CSR could be determinant of CSR disclosure, while age and gender, as well as the education received are not. Given the interesting finding in previous studies, the evidence of younger subjects may have more interest in the wider social responsibility arena (Arlow, 1991). This means that younger managers are more receptive to new ideas (Mellahi & Guermat, 2004) and the relatively new stakeholder approach (Ramasamy et al., 2007). Although some authors have argued on the effect of age are contradictory. Further, it appears that managers tend to give more priority to personal growth than wealth or advancement (Hall, 1976) and that they become more ethical (Singhapakdi et al., 1999;Terpstra et al., 1993) as they grow older. Thus, the age of firm manager variable was found to be reflected negatively on the CSR disclosure (β 5 =−0.058, tstatistic =−0.640); hypothesis H 5 .
The result also shows that a positive but statistically insignificant correlation between female SOE managers and the level of CSR disclosed (β 6 =0.039, t-statistic =0.413). Thus, H 6 is not supported. The insignificant effect of female managers on CSR disclosure is similar to those found by Glass et al. (2015), Giannarakis (2014), Kahreh et al. (2014), and Bowrin (2013. The insignificant effect could be due to a number of reasons. First, the low representation of female managers may require these women directors to align their leadership style with that of men, thus affecting their influence on CSR disclosure's decision. In Libya, the focus to include (more) female managers is still new. Newly appointed female managers may have less experience related to CSR activities and disclosure, thus affecting their leadership style and influence on the decision-making process. However, the results are inconsistent with those who found a positive relationship between female managers and CSR disclosure (Liao et al., 2015;Giannarakis et al., 2014;Jia & Zhang, 2013). According to some, there are no significant differences in gender (Atakan et al., 2008;Gill, 2010), but many studies have suggested that females pay more attention to ethical responsibilities than their male counterparts (Marz et al., 2003;Smith et al., 2001).
In the seventh hypothesis, H 7 , the result shows that a positive but statistically insignificant correlation between ethical business education of SOE managers and the level of CSR disclosed (β 7 =1.574, t-statistic =1.562) in our model. Some authors have stated that managers with higher education have a greater ability for information processing and innovation and devising new ideas; therefore, they are more capable of making decisions under complex situations (Damanpour & Schneider, 2009;Lee et al., 2005;Bantel & Jackson, 1989). Our results suggest that, according to the study by Belal (2001), a lack of knowledge and awareness of issues related to the field of social responsibility can result in scant or poor disclosure of CSR information.
Due to the fact that CSR practices may be conditioned by the characteristics of a firm and its managers (Hillman & Keim, 2001;McWilliams & Siegel, 2000;Waddock & Graves, 1997), our result shows the variable CSR responsibilities of SOE managers are positively and significantly associated with CSR disclosure at the 95% significance level (p <0.05) (β 8 =0.247, t-statistic =5.502); hypothesis H 8 . This indicates that SOE managers seek not only to further the firm's interests, as is the case of social entrepreneurship but also to help achieve social goals (Heath & Norman, 2004). This is the case because, as public management research has shown that managers' activities and strategies have an evident impact on business performance and information disclosure (Rodríguez Bolívar et al., 2015;Hicklin & Godwin, 2009). Our results also suggest that according to the earlier studies the relationship between responsibilities for CSR issues and disclosure is weak or negative, the later studies, having taken into account Patten's (2002) suggestions, tend to find a positive relationship. Patten (2002) suggested a number of reasons for this, mainly that the samples were too small and that the studies did not control for extraneous variables that could influence the relationship.

Robustness analysis
We conduct a series of robustness estimations to examine the extent to which our main findings are robust or sensitive to the use of standard OLS regression which will provide inconsistent parameters due to the correlated-omitted variables' problem. The common econometric solution to endogeneity matter is the use of instrumental variables' specification procedure. The instrumental variable approach should be associated with endogenous regressors but unrelated to the error term in the structural equation models (Habib & Azim, 2008). Larcker & Rusticus (2010) showed that OLS estimates provide better parameter estimates than the two-stage least square approach if the chosen instrumental variables do not conform to the standard definition of instrumental variables. The way of addressing the potential endogeneity for our study, we follow the Ntim (2016), Ntim andSoobaroyen (2013a, 2013b), Larcker & Rusticus (2010), Frank (2000), and Frank et al. (2008) alternative approach. Their method involves assessing how large the endogeneity issue (unmodeled variable) has to be to change the OLS coefficient estimates and, in particular, how large it has to be to make the coefficients statistically insignificant.
In our analysis, we first examined the presence of multicollinearity in our dataset, which is considered to exist when the variables correlate highly. Even though the rule of thumb is that VIFs figures of more than 10 depict the presence of multicollinearity, Freund and Littell (2000) suggests that VIFs figures below 5 indicate the absence of multicollinearity. The VIFs figures in our regression output (Table 5) confirm the absence of multicollinearity between variables in this study. Second, to check for potential endogeneity that may occur as a result of joint determination between independent variables and dependent variable CSR disclosure, the Durbin-Wu test was applied. In a statistical model, the F-values = 4.469 for the model is significant at a confidence level of 99% (p <0.01) level. This suggests that the independent variables considered, when taken together, explain the dependent variable CSR disclosure and its categories took individually. However, this does not mean that each of the independent variables contributes to the explanation of the dependent variable CSR information disclosure by SOEs. Furthermore, the explanatory power of the regression model explains about 0.33% of the total variation of CSR disclosure scores between the firms can be explained by the independent variables included in the regression models.

Conclusions and implications
The first objective of this study is to establish the CSR information disclosure of Libyan public listed firms. The data analysis for the listed firms' annual reports during the period 2010-2018 reveals that the involvement and disclosures of CSR information are generally low levels of disclosure. The findings of this research are consistent with those of other studies and with those in other developing countries generally and Libya in particular (N.S. Ahmad & Gao, 2005;N.S.M. Ahmad & Mousa, 2010;Pratten & Mashat, 2009). Most public listed firms in Libyan disclose their CSR information in a general statement where content is limited, namely, that provide a new understanding of policies that should be applied by public firms in the continuous improvement of their accountability, with specific regard to CSR information.
The second objective of this study is to examine whether there are any factors that influence CSR information disclosure by SOEs in Libyan, taking into account stakeholder theory and legitimacy theory. Additionally, the low levels of CSR disclosure in Libya are connected to the levels of demands and needs of stakeholders to improve the firm's image. An additional justification for this result could be that in developed countries, where the economy is well-established stakeholders in the public sector can afford to be more aware of, and receptive to, environmental and social impacts, whereas in developing countries worry more about meeting basic material needs. Similarly, the low level of disclosure on environmental disclosure seems to be consistent with the findings of many studies and particularly those conducted in developing countries (Andrew et al., 1989;Belal, 2001;Imam, 2000;Rizk et al., 2008).
The results of this study however show that among the variables such as firm size, age, and type of sector are determinant factors in CSR disclosure. This finding is consistent with the findings of prior studies that indicate CSR activity and its disclosure are important and systematically determined by firm characteristics that affect the relative levels of CSR information disclosure and its performance (Patten, 2002;Wang et al., 2013). Hence, firms can improve their advantage in social responsibility concerns and respond to society's demands by disclosing CSR information; thus, recruitment of managers who are concerned with environmental protection. Although previous research has observed that in countries where state ownership dominates, such as Libya, only a few users other than the government are expected to be provided with accounting disclosure and standards (Arpan & Radebaugh, 1985), our own findings suggest that the fact of government ownership is not significant with respect to CSR information disclosure. The reason for this result can be found in the firms in Libya, which are predominantly state-owned, pay more attention to commercial information (i.e., selling prices) rather than financial information, perhaps because most such firms tend to lack rich market conditions and the weakness of the Libyan stock market.
Quantitative results also suggest that, according to the characteristics associated with SOE managers, those that are inherent, such as age and gender, as well as education are not significant variables with respect to CSR information disclosure. We also found other interesting relations referring to the SOE managers' variable (female): shows that a positive but statistically insignificant correlation between female of SOE managers and the level of CSR disclosed is similar to those found by Glass et al. (2015), Giannarakis (2014), andKahreh et al. (2014). In our model show that the variable CSR responsibilities of SOE managers are positively and significantly associated with CSR disclosure. This seems to indicate that the characteristics and attitudes of SOE managers seek not only to further the firm's interests, as is the case of social entrepreneurship, but also to help achieve social goals and strategies (Heath & Norman, 2004;Pedersen, 2011).
This study presents a comprehensive overview of the factors that may affect the disclosure of CSR, therefore, has obtained significant findings that could help characterize a business and the profile of its SOE managers. The consciousness of the significance and influence of these characteristics can help researchers better understand CSR information disclosure by SOEs in Libyan firms. Firm factors such as its size, age, type of sector, and CSR responsibilities in which it operates were found to exert a positive influence on the disclosure of CSR information. With respect to the profile of SOE managers, those who are exercise responsibility in areas of CSR appear to have a significant influence on information disclosure in this respect, perhaps because their profession makes them more aware of such issues. The diversity of stakeholders and notions associated with CSR as perceived by SOE managers can become an opportunity rather than a constraint, and therefore it is essential to have a good understanding of their characteristics and attitudes.
Although our results are important, there are some limitations in terms of generalization. During the eight-years period examined, CSR remained largely unresponsive to significant changes in the political scene that occurred over a much longer period of time. Given the relevance of this field, it should be noted that view previous research has been conducted at the organizational issues related to the social responsibility of SOEs (Rodríguez Bolívar et al., 2015;Córdoba Pachón et al., 2014) and that the approach presented in this paper will provide a better understanding of the issues involved in order to assist decision-making practices by firm managers and politicians, and opening up interesting lines for future research. It would also be interesting to know what policies and criteria are currently applying CSR public managers in their firms as well as the causes and improved factors that could make public firms more socially responsible.

Funding
The authors received no direct funding for this research.