Audit Committee Features and CSR Disclosure: Additional Evidence From an Emerging Market

This study aims to measure the level of corporate social responsibility disclosure (CSRDL) in an emerging market, Jordan. It also aims to investigate the effect of audit committee (AC) features on CSRDL. The study’s data, obtained from annual reports, comprises 576 company-year observations from industrial and services companies listed on the Amman stock exchange for the period 2011 to 2016. Content analysis is used to measure CSRDL, and multiple regression techniques to test the proposed relationships. The results show that CSRDL among Jordanian listed companies is low, suggesting that companies have less incentive to disclose CSR practices. Additionally, AC independence and AC ownership positively influence CSRDL. The findings from this research have limited generalisability due to the specific characteristics of the sample, which is from a developing country. However, they have implications and value for policymakers and different stakeholders interested in improving CSR initiatives. The findings are also significant in reducing agency costs and enhancing corporate transparency. Within the limits of the authors’ knowledge, this research is among the first to provide preliminary evidence on the effect of AC features on CSR disclosure in an emerging market context, specifically in Jordan.


Introduction
There has been an upsurge in financial crises, cases of corruption and environmental hazards in recent years. These have a significant effect from the economic, social, and environmental perspectives. Because of these contemporary challenges, corporate social responsibility (CSR) reporting is being emphasised by government authorities, capital providers, stakeholders, academics, professionals, and society as a whole, as these parties are pressuring companies to accept greater responsibility for sustainable development (Aguinis & Glavas, 2012;Garcia et al., 2020). Well-known cases of fraud and corporate scandals like those of Enron, WorldCom, American insurance group, Lehman Brothers and Meadow have shown that contemporary corporate governance should dwell more on ethics, accountability and social responsibility and the need to tackle agency conflicts (Elkington, 2006). Consequently, corporate stakeholders now demand more understanding of how business decisions and strategies incorporate considerations of social, environmental and economic issues (Amran & Ooi, 2014). CSR activities are considered as one of the important strategies that allow companies to attract institutional investment, as these investors have no concern with companies that actively engage in community contribution practices (Daas & Alaraj, 2019).
The audit committee (AC) constitutes a corporate governance mechanism that enhances the integrity of financial and non-financial reporting. It is the most strategic of the committees in the financial reporting system (Ali & Kamardin, 2018). It is important to emphasise the quality of ACs because they are directly involved in providing information to the public. This information should be credible and contain more disclosure to meet stakeholders' expectations and to mitigate the asymmetry of information (Li et al., 2012). Studies have extensively examined the relationship between AC characteristics and CSR disclosure, and tend to agree on the positive role of the existence and characteristics of the AC on enhancing CSRDL (Appuhami & Tashakor, 2017;Garas & ElMassah, 2018;Khan et al., 2013;Said et al., 2009). However, the effect of AC features on CSRDL in Jordan is not yet clear. To the best of the

CSR Theoretical Background
Over the last two decades, several studies on CSR disclosure have contributed to further theoretical development and their practical application in the business world. This research is based on two theories: legitimacy and agency theories (Garcia et al., 2020). From the legitimacy perspective, companies' CSR disclosure practices constitute a tool for establishing and maintaining their legitimacy because CSR affects several important stakeholders (Tilt & Symes, 1999). Hence, corporate CSR disclosures may be utilised to show the extent to which the organisation conforms to the expectations of the wider society (Deegan et al., 2002). Companies are expected to disclose a sufficient level of financial and non-financial information to serve as evidence that they comply with societal expectations and conduct their operations within the limit considered acceptable by stakeholders (Wilmshurst & Frost, 2000). Numerous studies suggest that legitimacy theory is appropriate in understanding CSR disclosures as a way of improving companies' image and legitimising their actions (Barakat et al., 2015;Dyduch & Krasodomska, 2017;Garas & ElMassah, 2018;Sadou et al., 2017). The second theory employed here is agency theory, under which the AC is considered to be an internal monitoring mechanism in the financial and non-financial reporting process that attempts to mitigate the conflict problem (Jensen & Meckling, 1976;Madi et al., 2014); this may lead to improved performance and higher quality of CSR disclosure. The literature suggests that the presence of an effective AC will improve the disclosure quality and internal control system while helping to mitigate agency costs (Appuhami & Tashakor, 2017;Madi et al., 2014;Setiany et al., 2017).
Recently, many studies have indicated growing interest in CSR disclosure (e.g. Alazzani et al., 2019;Garcia et al., 2020). Researchers argue that it performs a vital function in enhancing the quality of communication among stakeholders (De Klerk et al., 2015). Additionally, research in emerging and advanced contexts has focused on the relationship between corporate governance proxies and CSR disclosure, providing inconsistent evidence regarding the manner in which corporate governance influences CSR disclosure (e.g. Garas & ElMassah, 2018;Khan et al., 2013;Majumder et al., 2017;Wu et al., 2012). Furthermore, a study conducted in Australia (Appuhami & Tashakor, 2017) focused on determining how AC characteristics influence CSR disclosure.
In Jordan, there has been limited research on CSR disclosure. The little available work concentrated on the factors that influence CSR disclosure in companies (Abu-Baker & Naser, 2000;Abu Farha & Alkhalaileh, 2016;Abu Qa'dan & Suwaidan, 2019;Al-Hamadeen & Badran, 2014;Alshannag et al., 2016;Barakat et al., 2015;Ismail & Ibrahim, 2008;Suwaidan et al., 2004). Most of these studies investigated the CSRDL in Jordan utilising content analysis. Their findings show that CSRDL is very low, especially when compared to developed countries. To cite one instance, Suwaidan et al. (2004) find that a company only reported around 13 per cent of the items included in the index, while only three companies had CSR disclosure scores above 30 per cent. In the work of Ismail and Ibrahim (2008), 15 per cent of the sampled businesses did not provide a report regarding CSR activities. The results of the study by Al-Hamadeen and Badran (2014) reveal that the average CSR disclosure score for the fiscal year 2011 was about 19 per cent, whereas Alshannag et al. (2016) find that the average level of CSR disclosure in the services and industrial sectors from 2011 to 2014 was about 34.1 per cent. Similarly, Abu Qa'dan and Suwaidan (2019) report an average of CSR disclosure level of 30.80 per cent. These findings point to a serious need for more CSR disclosure among companies in Jordan if they are to be compared with more developed economies.
Research on how corporate governance affects CSR information disclosure in Jordan is limited and provides inconsistent results (Abu Qa'dan & Suwaidan, 2019;Ibrahim & Hanefah, 2016). Ibrahim and Hanefah (2016) looked at board diversity features and how they affect CSRDL. Using a sample of 117 companies registered on the ASE from 2007 to 2011, they found that directors' independence and having women, foreign and young members on the board all significantly and positively affect CSR disclosure. However, it was revealed that profitability did not play a significant role in the decision to disclose CSR activities. Abu Qa'dan and Suwaidan (2019) found that board size is significantly and positively associated with CSRDL. However, the proportion of independent directors, the duality of CEO, board ownership, director's age, and institutional ownership have no impact on CSRDL. The effect of the AC on CSR, to the best of the researchers' knowledge, has not yet been addressed. Thus, the current study fills this gap in the literature.

Audit Committee Features and CSR Disclosure
The AC is an important link between the board of directors, management, internal and external auditors (Ahmed Haji & Anifowose, 2016). Thus, the presence of an AC is expected to improve communication between owners and management (Barako et al., 2006), reduce information asymmetry, enhance disclosure quality and improve market performance (Said et al., 2009). Additionally, the presence of the AC in public corporations reduces agency cost (Forker, 1992), which could improve corporate performance. Consistent with agency theory, the AC can act as a watchdog that promotes corporate disclosure and enhances transparency. This assures stakeholders and assists management in securing finance. Previous studies on the AC suggest that its effectiveness depends on certain features (Akhtaruddin & Haron, 2010;Dhaliwal et al., 2010;Li et al., 2012;Persons, 2009). Therefore, a suitable mix of experiences and skills is vital in enabling the AC to perform effectively. In view of this argument, some hypotheses are proposed and investigated in the present study to determine the influence of AC characteristics (size, independence, financial expertise, and share ownership) on CSRDL.

Audit Committee Size (ACSIZE) and CSR Disclosure
Larger ACs have the capacity, diverse expertise and perspectives to ensure appropriate oversight (Bé dard et al., 2004) of CSR disclosure. They may reduce the likelihood of information asymmetry and the agency problem (Chen & Jaggi, 2000). Mangena and Pike (2005) assert that a larger AC results in better oversight because it is more likely to be made up of the necessary expertise and skills. Previous empirical research yielded mixed results regarding the role of AC size. For example, some authors reported that AC size positively influences the extent of voluntary disclosure (Madi et al., 2014;Persons, 2009;Setiany et al., 2017). Similarly, Appuhami and Tashakor (2017) found that AC size positively influences the extent of CSR disclosure while other studies found no significant effect of AC size on voluntary ethics disclosure (e.g. Othman et al., 2014). In line with the proposition of agency theory, it is expected that larger ACs will possess a greater capacity for board monitoring to improve the level of CSR disclosure. Therefore, the following hypothesis is formulated: H1: The AC size positively influences the level of CSR disclosure.

Audit Committee Independence (ACIND) and CSR Disclosure
For ACs to be effective, they need to be free from top management influence (Lin et al., 2008). In line with agency theory, independent directors are more inclined to oversee management's activities and to increase disclosure. This reduces information asymmetry and agency conflicts to the minimum (Fama & Jensen, 1983). As noted by Fama (1980), including independent directors from outside the company decreases the chance of collusion among the management team and prevents expropriation of the assets of shareholders. Some past studies related to the subject found a positive effect of AC independence on the voluntary financial disclosure level (Akhtaruddin & Haron, 2010;Setiany et al., 2017). Garas and ElMassah (2018) found a positive effect of AC independence on CSRDL. In contrast, Othman et al. (2014) reported an insignificant positive effect of AC independence on voluntary ethics disclosure. Therefore, the current research investigates the influence of AC independence on CSR disclosure. In pursuance of this objective, the following hypothesis is proposed: H2: AC independence positively affects the level of CSR disclosure.

Audit Committee Financial Expertise (ACEXP) and CSR Disclosure
Based on the agency theory, the corporate board, particularly the AC, needs critical expertise and skills to supervise management and to monitor the preparation of financial reports effectively (Fama & Jensen, 1983). Essentially, members of the AC should possess a certain level of financial literacy to enable them to understand and interpret financial reports (Dhaliwal et al., 2010). According to Bé dard and Gendron (2010), expertise in the area of finance enables AC members to probe deeper and ask vital questions which challenge management and auditors. This will lead to higher quality financial statements. It will also improve the transparency of financial reports and keep agency conflict to the barest minimum. Persons (2009) asserts that independent AC members who also possess accounting skills have a higher probability of detecting irregularities in financial statements, and other inappropriate transactions. Such members are more diligent because they need to protect their professional image and code of ethics.
Empirical studies that examined the influence of AC financial expertise on CSRDL are few and provide mixed results. For instance, Madi et al. (2014) found that AC financial expertise did not significantly influence the voluntary disclosure level. Likewise, Appuhami and Tashakor (2017) reported an insignificant association between AC financial expertise and CSRDL. However, Mangena and Pike (2005) and Felo and Solieri (2009) reported a positive and significant link between these two variables. Therefore, the following hypothesis is formulated: H3: AC financial expertise has a positive effect on the level of CSR disclosure.

Audit Committee Share Ownership (ACOWN) and CSR Disclosure
There are two schools of thoughts regarding the influence of share ownership of the AC members on the company's disclosure level. From the point of view of the agency theory, the first argues that the interests of directors who hold a high proportion of company shares will be more closely aligned to those of other shareholders. It is also expected that they may be more zealous in overseeing management activities, which could improve the financial reporting system and encourage a higher quality of disclosure to the capital market (Karamanou & Vafeas, 2005). However, a contrary argument is that higher AC share ownership may cause weak oversight of the members because they would tend to act in their own interest at the expense of other shareholders. This may have a negative impact on their effectiveness (Mangena & Pike, 2005).
Previous researchers did not directly investigate the link between CSR disclosure and AC share ownership. However, some authors (Li et al., 2012;Mangena & Pike, 2005) provide indirect evidence of a significant negative effect of AC share ownership on the extent of voluntary disclosure. Therefore, the following hypothesis is established: H4: AC share ownership has an effect on the level of CSR disclosure.

Sampling and Data
The sample of the current study includes all non-financial Jordanian companies listed in the ASE between 2011 and 2016. Companies operating in the finance sector (e.g., banks, insurance, and diversified financial services) were excluded because they have special regulatory provisions and are subject to peculiar corporate governance laws which differ from other registered companies (Rainsbury et al., 2009). A company with missing data is also excluded. The final sample is 96 companies (576 company-year observations) in two sectors: industrial and services. These two sectors have more impact on society and the surrounding environment (Ismail & Ibrahim, 2008). The sample is summarised in Table 1.
Annual reports of the Jordanian companies are the main source of data, supplemented by board of directors and ownership disclosures submitted separately to the ASE, and formal contact with companies to gather the AC missing data. Content analysis was used in extracting CSR data from companies' annual reports. Content analysis is a technique for gathering data and codifying both qualitative and quantitative information into pre-determined categories to assist in the presentation and communication of vital information (Krippendorff, 2018). The study utilised the CSR checklist relevant to Jordanian companies used by Abu Qa'dan and Suwaidan (2019). This checklist consists of information in four categories of activity (community involvement, environmental, products/services to customers and human resources) covering an inventory of 42 items. A dichotomous scoring technique was utilised such that a company is scored 1 if a checklist item is disclosed and 0 otherwise. Therefore, the CSR disclosure index was arrived at by computing the ratio of actual scores to the maximum obtainable (42) by that company. As in Abu Qa'dan and Suwaidan' (2019) procedure, the CSR disclosure index was calculated as:

∑
Where: d = 1 if the item i is disclosed and 0 otherwise; and t = the time (year). So that, 0 per cent ≤ CSR disclosure score ≤ 100 per cent.

Independent and Control Variables Measurement
Independent variables in our research model are AC features (AC size, AC financial expertise, AC share ownership and AC independence). Previous studies suggest that it is important to control for company characteristics when examining CSR disclosure (Kansal et al., 2014;Ting, 2020;Wachira et al., 2019). Several variables that may affect CSR disclosure are therefore controlled for in the model: ownership concentration (OWC) (Crifo et al., 2016); audit fees (AFEE) (Chen et al., 2016); return on assets (ROA); market to book ratio (MTB); company size (CSIZE) (Ghaleb et al., 2020b); and debt ratio (DEBT). The regressions also include sector type and year indicator variables to further control for any industry and year effects. Definitions and measures of the variables investigated in this study are summarised in Table 2.

Model Specification
To determine the influence of AC features on CSRDL in the Jordanian market, we estimate the following ordinary least squares (OLS) regression model with robust standard errors ( Year is the period of study; and є = Error term (see Table 2 for variables definitions).

Descriptive Statistics
Descriptive analysis results are reported in Table 3. The mean CSRDL for the study period 2011-2016 is 0.293, indicating that, on average, Jordanian listed companies disclose 29.3 per cent of the items contained in the CSR checklist in their yearly statements, that is a low CSRDL. However, this result is similar to that of Abu Farha and Alkhalaileh (2016) and Abu Qa'dan and Suwaidan (2019). Importantly, CSRDL for each year (untabulated) is almost invariable, suggesting that companies did not increase their percentage of CSR item disclosure over the period of study. The statistics in Table 3 also show that the number of directors on the AC (ACSIZE) ranges from two to five with a mean of approximately 3.106. The mean proportion of companies with independent AC members (ACIND) is around 78.6 per cent. This shows that most companies have ACs with independent directors. On average, 58.8 per cent of AC members have financial expertise (ACEXP). This encompasses academic and professional qualifications/experience in finance or accounting. In terms of AC ownership (ACOWN), the average is approximately 3.7 per cent over time. For the control variables, Table 3 shows that companies in Jordan have highly concentrated ownership by families or government (Abu Qa'dan & Suwaidan, 2019). The mean ownership concentration (OWC) is around 61.148 per cent over the period of the sample. Moreover, the mean of audit fees (AFEE) is 9.402, where the average of the return on assets (ROA) is 2.30 per cent. The average market to book value ratio (MTB) is 1.34. Further, the statistics show that the mean company size (CSIZE) as measured by the natural log of total assets is 17.112, and the mean debt ratio (DEBT) is 32.4 per cent.
The correlation matrix between variables is illustrated in Table 4. The results show a positive relationship between AC size, AC independence, AC financial expertise and CSRDL. However, there is a negative correlation between AC ownership and CSRDL. The results show a high correlation between audit fees and firm size, which is not surprising given that the larger companies invest more in external audit. Importantly, Table 4 shows that multicollinearity is not a serious problem among the variables, as the coefficients of the relationships are below the critical level of 0.80 (Greene, 2003). Furthermore, variance inflation factors (VIF) were calculated to test for potential multicollinearity. According to Greene (2003), multicollinearity exists when the VIF value is more than 10. In this study, VIFs range from 1.08 to 2.66, so multicollinearity was not detected (see Table 4). Moreover, this study applied the Breusch-Pagan/Cook-Weisberg tests to detect any linear form of heteroscedasticity and the Durbin-Watson test to investigate autocorrelation of the error terms. The results show that the data suffer from autocorrelation. Thus, OLS regression with a robust function was employed (Alazzani et al., 2019;Appuhami & Tashakor, 2017).  Note: *** p <0.01, ** p <0.05, * p <0.10. Variables definitions are presented in Table 2. Table 5 depicts the empirical results of the multiple linear regression between the AC characteristics and CSRDL to investigate our hypotheses. The regression model tests whether the features of AC increase the level of overall CSR disclosure. The model is significant at the 1 per cent level (R-squared = 0.50, p-value <0.01). This shows that the regression model can account for about 50 per cent of the variance in CSRDL, and this is explained by the independent variables in the model.

Multivariate Analysis
Based on the results presented in Table 5, the coefficient of AC size is negative but not significantly associated with CSRDL (t-value = -0.05, p>0.10). Thus, H1 is rejected, suggesting that ACSIZE has no effect on CSRDL in the Jordanian market. However, this result is in contrast to previous studies (Appuhami & Tashakor, 2017) that report a significant positive association between AC size and CSRDL.
Moving on to the second hypothesis, the results show that AC independence (ACIND) has a significant and positive association with CSRDL at 1 per cent (t-value = 4.31, p<0.01). Thus, H2 is supported. This is consistent with agency theory, that emphasises the importance of directors' independence in enhancing financial reporting quality. It is also in line with previous empirical studies' findings (e.g. Appuhami & Tashakor, 2017;Setiany et al., 2017;Garas & ElMassah, 2018) which report a positive influence of AC independence on the level of corporate disclosure. This result affirms the argument that having independent directors on the AC strengthens its performance. Therefore, it can be argued that an independent AC provides more effective oversight of the information provided in companies' annual reports.
The results in Table 5 reveal that AC financial expertise (ACEXP) has no significant association with CSRDL (t-value = -0.02, p>0.10). Thus, H3 is rejected. This is consistent with Appuhami and Tashakor (2017) and Othman et al. (2014). A possible reason is that AC members with financial expertise pay more attention to the financial processes and transactions than to CSR activities and disclosure (Appuhami & Tashakor, 2017). Regarding AC ownership (ACOWN), the coefficient is strongly and positively related to CSRDL (t = 3.95, p< 0.01). Thus, H4 is supported. This can be interpreted to mean that AC ownership is vital in improving CSR disclosure and reducing the information asymmetry associated with agency problems between shareholders and managers. The results support the positive influence of directors' and managers' incentives predicted by agency theory (alignment hypothesis). However, the result contradicts other studies that show the relationship to be negative (Li et al., 2012).
Regarding control variables, the results reported in Table 5 show that ownership concentration is negatively and significantly associated with CSRDL. This result is in line with the claim that larger shareholders have interests other than CSR activities (Crifo et al., 2016). Thus, companies with high ownership concentration consider CSR activities as a cost rather than an investment that may contribute to the company value. Similar to the results of another study (Chen et al., 2016), our results show that audit fees are positively and significantly associated with CSRDL. Company characteristics ROA, MTB, CSIZE, and DEBT were found to be positively associated with CSRDL, suggesting that these characteristics have a significant role in enhancing CSRDL.  -statistic (16, 559) 46.72*** Prob. > F-statistic 0.000 Note: *** p <0.01, ** p <0.05, * p <0.1. Variables definitions are presented in Table 2.

Additional Tests
Recent research emphasises that AC measured through a composite score of its characteristics can better capture the AC's effectiveness (ACE) in monitoring financial reporting, as these characteristics complement each other (e.g., Lisic, 2014;Zaman et al., 2011). Therefore, we measure the AC features using binary variables (dummies). ACSIZEdum equals 1 if the AC has more than three members, and 0 otherwise. ACINDdum equals 1 if the AC is fully independent (all members are independent), and 0 otherwise. ACEXPdum equals 1 if the AC has at least one member with financial expertise, and 0 otherwise. ACOWNdum equals 1 if any AC member owns shares in the company and 0 otherwise. Finally, we score all dummies to measure the ACE and examine its influence on CSRDL. The results reported in Table 6 show that ACE has no significant influence on CSRDL. Overall, the results suggest that the role of AC in the Jordanian market is still inadequate, although some individual characteristics show a positive association with CSRDL.  -statistic (16, 559) 55.88*** Prob. > F-statistic 0.000 Note: *** p <0.01, ** p <0.05, * p <0.10. ACE is AC effectiveness measured as a composite score of the dummies of four AC features (ACSIZEdum, ACINDdum, ACEXPdum, and ACOWNdum). Definitions of other variables are presented in Table 2.

Conclusions
The research was conducted to investigate the effect of AC characteristics on CSRDL among Jordanian listed companies over the six years 2011 to 2016. Specifically, making use of agency theory, it tests the effect of several features (size, independence, financial expertise, and ownership) as determinants of CSR disclosure. The results reveal that the presence of independent AC members and AC ownership has a positive impact on CSR disclosure. This indicates how these elements are key factors in promoting information transparency for the benefit of shareholders and stakeholders. However, no link was found between AC size or AC financial expertise and CSRDL.
This study has significant managerial implications from a practical perspective. Jordanian policymakers and regulators (such as ASE and Securities Depository Center) could benefit from the study's insights and learn how to reform and formulate policies to encourage compliance by companies that are required to disclose CSR. This will make them more responsive to social interests because CSR disclosure has a very important role in meeting the interests of stakeholders and the company's long-term needs. Furthermore, findings from the study suggest that the ASE should review current listing requirements by putting more emphasis on CSR disclosure. It should improve the current structural and functional requirements of ACs to make them work more effectively.
The study has a few limitations. First, it used corporate annual reports as the medium through which to quantify CSR, although there are other mediums used by listed companies to communicate with their stakeholders, such as websites and corporate letters. Second, our results are based on a sample of non-financial companies, so the generalisability of the results requires including financial companies. Third, the study selected only one proxy for corporate governance, AC, although there are other proxies are available to represent board composition and ownership structure. Despite these limitations, the evidence provided in this study makes an interesting contribution regarding the role of AC members in the practice of CSR reporting. Additionally, these insights provide fresh opportunities for further research, for example, longitudinal studies of financial companies. Finally, future research might analyse the impact of other characteristics of the AC, such as the gender, age and educational level of the directors, multiple directorships, tenure, and remuneration.