The effect of corporate governance elements on corporate social responsibility reporting of listed companies in Vietnam

Abstract This paper examines the impact of various corporate governance characteristics on the disclosure of corporate social responsibility (CSR) information in the context of developing markets during the COVID-19 pandemic. We used data from the annual reports of the top 100 companies listed on the Ho Chi Minh Stock Exchange from 2019 to 2021 to investigate the relationship between board independence, board size, CEO duality, foreign ownership, government ownership, and the disclosure level of CSR reporting. We employed a content analysis to measure CSR disclosure level by using the CSR reporting index (CSRRI) with 17 items. The panel regressions including three techniques like the pooled-ordinary least squares (OLS), fixed effects model (FEM), and random effects model (REM), were used to analyze the data, and FEM) shows its best fit to the model. The results suggest that board independence and government ownership correlate positively and significantly with companies’ decisions to disclose CSR information. In contrast, board size, CEO duality, foreign ownership and was found to be insignificant. Our study extends the literature on CSR particularly in developing countries where governments play a significant role in promoting economic development. It also evaluates the CSR disclosure level of the top listed companies in Vietnam. Since then, the paper’s results provide several insights to the policymakers to identify the corporate governance characteristics that will promote CSR reporting in Vietnam’s listed companies.


PUBLIC INTEREST STATEMENT
In the context of international integration, CSR disclosure is always necessary and important to achieve sustainable development goals. The objective of the study is to examine corporate governance elements affect the disclosure level of CSR reporting of Vietnamese listed companies in the last 3 years (2019-2021). By using data from a sample of the top 100 listed enterprises on the Vietnam Stock Exchange, we used panel regressions to test the hypothesis. The results show that listed companies with higher proportion of board independence, higher state ownership rate will disclose more information in the CSR reporting. The findings from this study not only may fill the gap in research but also will be foundations for recommendations to improve the level of CSR disclosure in Vietnamese listed firm in the future.

Introduction
Disclosure of corporate social responsibility has been becoming a hot topic in academia and practice, in which business managers pay more attention to the social and environmental effects of business operations (Amel-Zadeh & Serafeim, 2018;Cohen et al., 2015). Enterprises that want to pursue sustainable development need to solve problems that influence the social community and create value for shareholders and other stakeholders (Christensen et al., 2021). In response to this demand, many companies worldwide started disclosing some sustainability information in their regular reports. Reporting on corporate social responsibility activities plays an increasingly important role for enterprises to show their commitment to environmental and social issues (Adams, 2004;Brammer & Pavelin, 2008).
Many organizations offer (voluntary) reporting standards for ESG activities that aim to improve or standardize reporting practices. For example, the Sustainability Accounting Standards Board (SASB) creates industry-specific disclosure standards across financial material, environmental, social, and governance topics that companies can use in their SEC filings. Similarly, the GRI (Global Reporting Initiative) is making global standards for sustainability reporting to help companies communicate their impact on critical sustainability issues. The IFRS Foundation is the latest group to get involved. It proposes a global approach to sustainability reporting to deal with many standards and the people who set them (IFRS, 2020). The introduction of the standard-setting mentioned above and regulatory initiatives has implied that sustainability reporting is considered one of the essential components to achieving climate and sustainability goals.
Due to the importance of CSR information disclosure, the academic community has been highly responsive to research to understand critical determinants of firms' (voluntary) CSR reporting decisions. Most of the previous studies on CSR reporting focused on developed countries (Guthrie & Parker, 1989;Deegan & Rankin, 1996;Kim et al., 2012) where the capital markets are mature, the legal framework on social responsibility has been completed, and stakeholder awareness of business accountability is high (Muttakin & Subramaniam, 2015). In contrast, the research on CSR reporting in developing countries is limited. The capital market in developing nations is still maturing, and regulations on corporate governance and disclosure of CSR information are generally weak. The demand for CSR reporting or sustainable development reporting is increasing significantly in developing markets as more and more companies want to join the regional stock market and attract foreign investment capital (Khan, 2010). Due to the far difference in sociocultural factors between the developed and developing economies (Blowfeld & Frynas, 2005;Jamali & Mirshak, 2006), further research on key driving factors of voluntary CSR reporting in a developing country needs to be further investigated.
Vietnam's economy is a market-based economy with a socialist orientation, characterized by the government's tight control and state-owned enterprises. The state-owned enterprises contribute to approximately 28.8% gross domestic product in Vietnam, compared to 10% of the average global (Bruton et al., 2014). Government intervention in the economy is decisive, and protection rights for creditors and shareholders are weak (Nguyen, 2021). In recent years, the Vietnamese government has paid attention to corporate governance for enterprises, especially listed companies in the stock market. The first introduction of the Law on Enterprises in 2005, the Law on Securities in 2006, and the Code of Corporate Governance for Listed Companies in 2007 laid the initial foundation for corporate governance in Vietnam. The Law on Enterprises is built based on the regulations of the legal principles of Anglo-American jurisdictions requiring greater board independence, separation of the CEO and chairman, provision of audit committees, etc. In 2019, the Vietnamese Ministry of Finance (MOF) and the Vietnamese State Securities Commission (SSC) partnered with the International Finance Corporation (IFC), the Swiss State Secretariat for Economic Affairs (SECO) released the "Vietnam Corporate governance code 2019" (State Securities Commission of Vietnam, 2019) consisting of some guidelines in regards to corporate governance practices by the listed companies. The Code of Conduct is also expected to support listed companies to ensure the board of directors' effectiveness in maintaining a high quality of disclosures, primarily environmental and social information. Therefore, the codes aim to achieve corporate accountability, which is in step with the new governance regulation model.
Several pieces of research have been conducted to analyze corporate governance's influence on CSR information disclosure in Vietnam. Vu and Buranatrakul (2018) implied several factors affecting the level of CSR disclosure, such as board independence, CEO duality, size, and profit. Besides, T. L. H. Nguyen et al. (2021b) reported a positive relationship between deputy CEO and state holding with CSR publication. In contrast, the proportion of independent directors, CEO duality, and CEO ownership was insignificant with CSR reporting information of 166 Vietnamese listed firms at the Hanoi Stock Exchange. However, very few studies focused on the Vietnam Corporate governance code 2019 and the role of this Code on CSR reporting information in Vietnam. Thus, our paper will fill this gap by examining the disclosure of CSR from the perspective of listed enterprises applying the New Code 2019. Our paper has several significant contributions. Firstly, it tests the impact of the Corporate governance factor on CSR disclosure under the New Code 2019 and Circular No 96/2020/TT-BTC (Minister of Finance, 2020) dated 16 November 2020 guiding information disclosure on securities market. Secondly, the paper identifies the critical determinant factors impacting the disclosure of CSR in developing countries where the characteristics of the market are different from that of developed markets. Thirdly, our study uses the data of the top 100 companies listed on the Ho Chi Minh Stock Exchange from 2019 to 2021 to provide a comprehensive picture of the CSR disclosure following Circular No 96/2020/TT-BTC. Several studies on CSR information disclosure in Vietnam used the data from 2019 backward (Hoang et al., 2018;T. L. H. Nguyen et al., 2021b;T.H. Nguyen et al., 2021a). Our study uses the latest data to investigate CSR disclosure in the context of the global pandemic.
The remainder of this paper is structured as follows. Section 2 discusses the background of CSR reporting in Vietnam. Section 3 introduces the theoretical framework, followed by the hypothesis development in section 4. Section 5 describes the data and methodology. Section 6 discusses the research findings, and the last section provides conclusions, implications, limitations, and suggestions for future research.

CSR reporting in Vietnam
Corporate social responsibility has become a key criterion for evaluating the firm performance of enterprises in Vietnam (Vu & Buranatrakul, 2018). Guidelines for CSR and Sustainability reporting in Vietnam (State Securities Commission of Vietnam, 2013) defined CSR as "the commitment of businesses to contribute to sustainable economic development by working with employees, their families, the local community, and society at large to improve their lives in ways that are good for business and for development." CSR reporting provides information to the public regarding companies' activities with the environment, community, employees, and consumers.
In Vietnam, CSR reporting is an integrated component of the annual report. Following the Law on Securities 2021 (Vietnam National Assembly, 2019) and Circular No 96/2020/TT-BTC, Vietnamese listed companies must prepare and publish CSR reporting. These companies must disclose 17-item CSR on CSR reports with 11 environmental indicators, 05 employee indicators, and 01 local community indicator. In addition, listed companies can prepare separate CSR reporting, namely Sustainability reports. The State Securities Commission of Vietnam has promulgated a document to guide enterprises to set up Sustainability reports according to GRI's standards (Global Reporting Initiative, 2017) with 3 Universal Standards (GRI 101, 102, 103) and three topicspecific standards (GRI 200-Economic, GRI 300-Environmental, GRI 400-Social). However, the number of listed companies making Stand-alone sustainability reporting is dramatically low as the disclosure is voluntary (Nguyen, 2021). Moreover, most enterprises have not paid adequate attention to CSR reporting information (T. L. H. Nguyen et al., 2021b). Bui (2010) indicated that Vietnamese listed firms were less interested in CSR disclosure due to three main reasons: the lack of understanding of disclosure impacts on society, financial resources, and the mandate legal framework. Thus, most listed firms in Vietnam showed low CSR disclosure quantity levels (Vu et al., 2011). There are almost no studies on corporate governance and CSR reporting of listed companies in Vietnam in the context of applying New Code 2019 and Circular No 96/2020/TT-BTC. Therefore, more studies are needed to clarify this relationship.

Theoretical framework
The most frequently cited theories in CSR reporting studies are the Legitimacy theory and Stakeholder theory (Adams et al., 1998;Gray et al., 1995;Moerman & Van Der Laan, 2005). The legitimacy theory provides a more comprehensive perspective on CSR reporting as it acknowledges that social contracts bind firms. In essence, the organization will receive support from the relevant parties and continue as its activities benefit or are not harmful to society (Buniamin et al., 2008). On the other hand, companies agree to take the various actions that the community desires to achieve their goals, ensuring their continued existence (Brown & Deegan, 1998;Deegan, 2002;Guthrie & Parker, 1989). The company will disclose more information on the environment and society to deal with these pressures from stakeholders, maintain its image as a legitimate company, and avoid adverse effects caused by the legitimacy crisis (De Villiers & Van Staden, 2006). Legitimacy theory, therefore, implies that an organization's top management is responsible for recognizing the legitimacy gap, carrying out necessary social responsibility activities, and disclosing to stakeholders through CSR reporting (Jizi et al., 2014). Thus corporate governance consisting of ownership structure and board composition (board size, CEO duality, female on board) plays a vital role in reducing the legitimacy gap through announcing CSR information.
Another theory to explain the CSR reporting practice in Vietnam is the Stakeholder theory. The stakeholder theory refers to the interest of many other objects in a company's CSR besides the traditional users of accounting information, such as shareholders and creditors (Moneva & Llena, 2000). The other objects demand information regarding the influence of a 'company's activities on the environment and society. When the company acknowledges the legitimate "stakeholders" interests, it will voluntarily report environmental and social information required by stakeholders (Monteiro & Aibar-Guzmán, 2010). Disclosure of social responsibility information builds a positive social image towards the public and satisfies their interests with stakeholders. Stakeholder theory also shows that corporate governance characteristics (company size, board independence, government ownership, and foreign ownership) influence CSR disclosure. Based on the stakeholder pressure, large enterprises are aware of the importance of CSR reporting more than smaller companies. Consequently, these enterprises tend to disclose more CSR information than smaller firms (Branco & Rodrigues, 2008;Moore, 2001). R. M. Haniffa and Cooke (2005) argued that independent boards that were supposed to represent the interests of other stakeholders would have more effect on CSR reporting. In contrast, Ghazali (2007) revealed that in compliance with stakeholder theory, Government-owned companies tended to be politically sensitive because their activities are more visible in the public eyes. Therefore, the government is likely to pressure companies to disclose more information because the government, as a body trusted by the public, will need to meet public expectations.

Board Independence
Boards are essential in monitoring and directing managers to satisfy stakeholders' interests. According to Article 153 of the Law on enterprises, the board of directors is the managerial body of the company. It has the right to make decisions on behalf of the company, perform its rights and obligations, and decide its company's medium-term development strategies and annual business plans (Vietnam National Assembly, 2020). Therefore, the presence of independent board members is considered a central corporate governance mechanism (A. Khan et al., 2013). Stakeholder theory argues that a high ratio of independent directors on the board could be an essential element that would help to resolve firm problems and advance the interests of other stakeholders, such as employees, customers, and local communities (Amran et al., 2009;Chen & Roberts, 2010).
The board, which comprises many independent members, has a greater monitoring and controlling ability over management and ensures that investors' interests are protected (Fama & Jensen, 1983). The State of "independence" is met when he or she is not working for the company or its parent company or a subsidiary company; did not work for the company or its parent company or subsidiary company within the last 03 years or longer; not directly or indirectly holding 1% of the 'company's voting shares or more (Vietnam National Assembly, 2020). It is expected that independent members of the Board of Directors are supposed to represent the interests of other stakeholders; they will have more influence on environmental and social reporting (R. M. Haniffa & Cooke, 2005). Most of the empirical studies that examined the relationship between corporate governance and information disclosure on CSR reporting found that board independence, measured in terms of the percentage of independent directors on the board, positively affects CSR disclosure (Barako & Brown, 2008;Harjoto & Jo, 2011;A. Khan et al., 2013;Shu & Chiang, 2020). Therefore, we propose the first hypothesis of this study: H1: There is a positive relationship between board independence and the degree of CSR reporting information.

Board size
From an efficiency and legitimacy perspective, corporate boards with a large number of members will strengthen supervision and management that could strictly comply norms and regulations of the enterprise (Alnabsha et al., 2018;Elamer et al., 2018;Ntim & Soobaroyen, 2013). According to the corporate governance mechanism, a larger board size may bring more members with experience (Xie et al., 2001) that may represent the stakeholders' interests (Halme & Huse, 1997). They concluded that larger boards were positioned to improve the company's governance. Consequently, the large boards were expected to be involved more in CSR activities and disclose more information through CSR reporting (Elmagrhi et al., 2016). M. Khan et al. (2020) indicated that board size was positively associated with information in CSR reporting-part of annual reports of Pakistan listed companies across a range of industries. In others developing countries, research by Said et al. (2009) in the 250 non-financial companies listed on the main board of Malaysia reported a positive relationship between board size and the level of CSR disclosures. However, Halme and Huse (1997) studied the relationship between the number of board members and information disclosure in the environment report of the 40 largest corporations from Scandinavian countries, including Finland, Norway, Spain, and Sweden. The findings showed that board size did not influence the extent of environmental reporting. Besides, Sufian and Zahan (2013) reported no relationship between board size and the disclosure of information in the CSR reporting in 70 Bangladeshi listed companies. To test whether board size influences CSR reporting information, we develop the following hypothesis: H2: There is a positive relationship between board size and the degree of CSR reporting information.

CEO duality
CEO duality occurs when the same person holds the organization's CEO and board chairman positions (Rechner & Dalton, 1989). The combination of CEO and chairman positions reflects leadership and governance issues (Said et al., 2009). R. Haniffa and Cooke (2002) point out the problem from two views. The first view supports the separation of the two top-level managers to controls and balances for the performance of management. If the CEO holds the chairman position, they will have greater power, which enables them to make decisions that do not maximize the shareholder's wealth. Powerful CEOs can use their influence to restrict voluntary disclosures, including CSR disclosures. Therefore, CEO duality may reduce the board of directors' effectiveness in monitoring and controlling the management (Buniamin et al., 2008). The second view argues that the separation is inconsequential because many companies have good performance results combined with the roles of CEO and chairman.
Furthermore, role duality increases the concentration of decision-making power and provides solid strength to the boards, enhancing the quality of CSR reporting (Al-Janadi et al., 2013). Previous studies by Forker (1992) found that CEO duality was associated with lower voluntary information in financial statements. Muttakin and Subramaniam (2015) found that CEO duality was negatively associated with levels of CSR reporting based on the data from the top 100 companies listed on the Bombay Stock Exchange. Shu and Chiang (2020) found similar results in their study, which analysed sample consists of 11,439 firm-year observations in the Taiwan Stock Exchange in the period of 2008-2015. The result showed a significant and negative relationship between CEO/chairman role duality and the levels of voluntary corporate disclosure. However, the studies of Said et al. (2009) in 150 Malaysian listed companies, Michelon and Parbonetti (2012) in 114 firms listed on the US and European stock markets, A. Khan et al. (2013) in 135 manufacturing companies listed in Bangladesh and Habbash (2016) in Saudi Arabian non-financial listed firms during 2007-2011, found no relationship between CEO duality and CSR disclosure. Hence, we propose the third hypothesis as below: H3: There is a negative relationship between CEO duality and the degree of CSR reporting information.

Foreign ownership
The ownership structure of the firm may lead to legitimacy gaps. CSR disclosure can function as a proactive legitimating strategy to obtain foreign investment further and satisfy ethical investors (R. M. Haniffa & Cooke, 2005). Foreign owners are also likely to be more aware and sensitive to society's growing expectations for CSR activities in the community. Bradbury (1991) stated that due to the geographical separation between management and owners, foreign shareholders had higher requirements for information disclosure when their ownership proportion increased. Thus, a company with foreign ownership is expected to disclose more social and environmental information to help them make decisions. Research by R. M. Haniffa and Cooke (2005), A. Khan et al. (2013), and Muttakin and Subramaniam (2015) provided some evidence support from an Asian context perspective for a positive relationship between foreign ownership and CSR disclosures. It showed that the involvement of foreign shareholders in Asian listed companies would enhance the extent of corporate social disclosure. However, in Malaysia-a Southeast Asian country, the study of Amran and Devi (2008) and Said et al. (2009) showed no association between foreign ownership and corporate social reporting. Though, we hypothesize that: H4: There is a positive relationship between foreign ownership and the degree of CSR reporting information.

Government ownership
Government ownership is another critical characteristic of large-scale listed companies in Vietnam.
Most of these state-owned enterprises have been equitized and listed on the stock market (Nguyen, 2021b). The State still holds a certain percentage of those enterprises depending on the importance of their business. Government involvement can pressure companies to disclose more information because the government is a trusted authority by the public (Said et al., 2009). Government shareholdings are expected to lead to greater corporate social responsibility disclosures in their annual reports, as government should promote transparency among publicly listed companies in Vietnam. Most of the previous studies, Nazli and Ghazali (2007), Amran and Devi (2008), Said et al. (2009), A. Khan et al. (2013), and Muttakin and Subramaniam (2015), and Gerged (2021) found that government ownership is statistically significant and positively linked to the extent of information disclosure in the CSR reporting. However, Dam and Scholtens (2012) discovered a negative relationship between government ownership and the extent of CSR disclosure. Since the majority of empirical studies still suggest a significant positive association, the study's fifth hypothesis is: H5: There is a positive relationship between Government ownership and the degree of CSR reporting information.

Control variables
The study considers size, profitability, and company age as the control variables. Previous studies have indicated a positive relationship between the degree of CSR reporting and company size, profitability, and age.
In the context of legitimacy theory, companies with larger sizes face pressure to disclose information about their compliance with state regulations and stakeholders (Patten, 1991). Several published studies provide evidence of the association between company size and corporate social disclosures on annual or CSR reporting (Kansal et al., 2014;M. Khan et al., 2020;Neu et al., 1998;Suwaidan et al., 2004). Besides, larger firms will realize better social responsibility practices, considering that social responsibility and disclosure are a way to enhance the 'company's reputation and image. This leads to our hypothesis: H6: There will be a positive relationship between company size and the degree of CSR reporting information.
R. M. Haniffa and Cooke (2005) discovered that profitable firms disclosed CSR information to show their role in society's well-being with the aim of validating their existence. The relationship between profitability and CSR reporting is also found in Patten (1991) and Roberts (1992) studies. However, some authors found a negative relationship as there existed the belief that disclosing CSR was a stakeholders' disadvantage because a company must use its resources only to maximize its profits (Preston & O'bannon, 1997;Simpson & Kohers, 2002). Thus, we proposed the hypotheses seventh as follows: H7: There is a positive relationship between profitability and the degree of CSR reporting information.
Stakeholders expect firms to continue carrying out and setting up CSR reporting to disclose information because long-term experience helps them use resources effectively and protect the 'business's reputation with social responsibility activities (Roberts, 1992). From the legitimacy theory perspective, the 'company's renown is built with age. Research by Hamid (2004) in Malaysia and Kansal et al. (2014) in India found a positive relationship between company age and the level of social responsibility disclosure. They indicated that the firms listed for a longer period would also be sensible towards more disclosure. In contrast, Rettab et al. (2009) reported a negative association between CSR disclosure and firm age. The eighth hypothesis is: H8: There is a positive relationship between company age and the degree of CSR reporting information.

Data
Our samples were listed companies on the Vietnam Stock Exchange (HSX). These enterprises are required to publish information on CSR following the regulations of the State Securities Commission of Vietnam. As of 31 December 2021, Vietnam had 1641 listed companies. Due to our funding and time, we are unable to collect data of all enterprises, so that, we chose Vietnam's top 100 listed companies (called VN100). The group of 100 listed companies includes 30 type-one enterprises (called VN30) and 70 type-two enterprises (called VNMidcap), with market capitalization accounting for about 85% and representing more than 80% of the transaction value of the whole market.
Social responsibility information is disclosed in the annual report. The previous year's annual report will be published from April to June of the following year. The annual reports can be easily downloaded from the company's website or the Ho Chi Minh Stock Exchange website. With the annual report collection in 3 years (2019-2021), the article produces 300 firm-year observations. The survey sample is classified into 11 sectors according to the Global Industry Classification Standard (GICS) developed by Morgan Stanley Capital International (MSCI) and Standard & Poors in 1999. The sample selection is reported in Table 1.

Dependent variable-CSR reporting
Although companies use different media for publishing social responsibility disclosures, we use annual reports to guarantee consistency in measuring the level of disclosure on the CSR reporting of listed companies. The dependent variable is the decision to report CSR information in the annual report. The dependent variable is measured by content analysis, a method of codifying the text (or content) of a piece of writing or categories depending on selected criteria (Weber, 1988).
We use the CSR reporting index (CSRRI) to measure CSR disclosure level by collecting information from Section 6, "Report the enterprise's impact on the environment and society," in the annual report. A "yes/no" or (1, 0) scoring methodology was employed (Branco & Rodrigues, 2008). "1" is used if firms present information on categories (items) in Section 6 of the report. In contrast, companies that do not publish any information are coded as 0. The maximum point is 17 when firms present all 17 categories in the checklist, including 11 environmental items, five employee items, and one society item (Table A1). The level of disclosure on the CSR reporting of each company is measured as the ratio of its disclosure score to the maximum points possible. For example, if a company reports no item (0) out of 17 items, the dependent variable score will be 0 percent. Likewise, if half of the total items are disclosed, the score for the dependent variable is 50 percent. Consequently, the formula to compute the CSRRI index is as below:

Independent variables
The study employed five independent variables: board independence, board size, CEO duality, foreign ownership, and government ownership. The construction of independent variables and their measurement technique is indicated in Table 2.

Control variables
The study used three control variables, namely company size measured by the natural logarithm of all assets at the end of the year (Andrikopoulos & Kriklani, 2013;Hussainey et al., 2011;Liu & Anbumozhi, 2009); company age, measured by the number of years from inception (Andrikopoulos & Kriklani, 2013;Rettab et al., 2009); and profitability measured by Return on assets (ROA), for three years from 2019 to 2021 (Andrian, 2020;Gray et al., 2001;Jizi et al., 2014).

Model specification
We used hierarchical regression analysis to test the relationship between the dependent variable and five independent variables (board independence, board size, CEO duality, foreign ownership, and government ownership), including three control variables. The regression equation is as follows: CSRRI ¼β 0 þβ 1 BOINDþβ 2 BOSIZEþβ 3 CEODUþβ 4 FOROWNþβ 5 GOVOWNþ β 6 LSIZEþβ 7 ROAþβ 8 AGE þ ε (1) Where CSRRI: CSR reporting index;   Vu and Buranatrakul (2018) and Ta et al. (2020), respectively. There is an improvement in CSR information disclosure, which may be linked to the application of the Vietnam Corporate governance in 2019 and the Law on Securities in 2021, Circular No 96/2020/TT-BTC, which regulated closer information disclosure in CSR reporting. In addition, the survey sample is large-scale companies in Vietnam's top 100 listed companies with abundant financial resources and more professional departments to collect, process and provide social responsibility information than other companies.

Descriptive statistics
The average board independence (BOIND) is 23.9%, while the average board size (BOSIZE) is 6.82. The statistics also show that only 3.7% of the CEOs in the research sample are the chairman  Table 4 presents the statistics for various information categories in CSR reporting. The highest rate of disclosed information belongs to the Social/local community information, followed by employee information (68.6%), and environmental data has the lowest rate of 31.1%. The result shows that listed companies have not paid adequate attention to environmental issues. It is consistent with the fact that companies avoid disclosing sensitive information related to environmental impact.
Social/local community information is published with the highest percentage. In the context of COVID-19 pandemic, listed companies always have an interest in performing social responsibility with the community and locality through charity activities, supporting doctors and nurses in hospitals, etc. However, social/local community information is calculated on a single indicator (SO1), so the representativeness is not high.
For employee information, the results indicate that number of employees (EM1) is important information, with 88.3% of firms presenting (Table 5). Following, labor policies to ensure health, safety and welfare of workers (EM3) is 81%. Listed companies disclose the lowest rate of employee training (EM4) with 45.3% Regarding environment information, although enterprises announce eleven environmental items, the level of disclosure of these indicators is low (<50%). EN1 is at the lowest level, with disclosure percentages amounting to 13%, followed by EN2 "Measures and initiatives to reduce GHG emission" and EN9 "Percentage and total volume of water recycled and reused" at 18% and 19.3%, respectively. In addition, all environmental indicators have rates below the average (< 50%). It shows that listed firms in Vietnam seem to have not put weight on environmental issues. Simultaneously with the disclosure of CSR information according to Circular No 96/2020/TT-BTC, listed companies have applied the GRI regulations to improve the quality of CSR information on reports. In 300 CSR reports of listed companies, we found that 50 reports were established under GRI. The number of CSR reports under GRI Standards in 3 years from 2019-2021 is 10, 18, and 22, respectively.

Analysis of the main results
According to Cooper and Schindler (2003), serious multicollinearity problems exist when the bivariate correlation score is 0.80 or greater. Before conducting a regression analysis, we undertook the correlation analysis to test the direction and degree of a linear relationship between the variables as it aids in identifying the potential existence of multicollinearity amongst the independent variables. Table 6 presents the result of the correlation analysis. Although there are significant relationships between Government ownership and company age, the correlation coefficient of .406 indicates that linear correlation at a weak level and multicollinearity is absent.
Based on estimation techniques, panel data regression models can be estimated using three estimation methods, including Pooled Ordinary Least Square (OLS), Fixed Effect Model (FEM), and Random Effect Model (REM). Of the three approaches will be chosen which approach is most suitable. A comparison between the OLS with the FEM was done by using F test, while the comparison of REM with the OLS was done by Breusch-Pagan's Lagrange Multiplier test (LM (LM Test). The Hausman test is used to choose between the FEM or REM models .
The results of F-test were significant at F = 27.02, with probability = 0.0000 (<0.05), so the conclusion is that the FEM test result is better than the OLS test result. After that, we also conducted LM test to know the suitability of using pooled OLS regression whereby the Chi-Square value = 19.51, p-value = 0.000, indicates that pooled OLS model is not appropriate for the data set. Moreover, we studied same sample of listed companies along different periods of time, OLS regression automatically become an inappropriate as suggested by Wooldridge (2010). The next stage is the Hausman test, which is performed to select better test results between FEM and REM. This results in a probability value of 0.0212, which is lower than 5%, so the conclusion is that that FEM test result is more appropriate than the REM test result.
After selecting the appropriate method to run the model (FEM model), we will detect the multicollinearity, heteroscedasticity and autocorrelation of the model. The multicollinearity test results of the model show that VIF < 3. Thus, there is no multicollinearity phenomenon in our study. To test the heteroscedasticity, the authors use the Wald test. The results show that probability = 0.0000, suggesting that we reject the null hypothesis of constant variance, which means that there is a variable variance phenomenon. Finally, we use Wooldridge test for autocorrelation in panel data. With prob value = 0.1691 we conclude that autocorrelation does not exist. In order to solve heteroscedasticity, the study uses GLS (Generalized Least Squares) estimation to overcome the variance in the model. After affirming the fit for the whole model, the summarized results presented in Table 7 are retrieved from OLS, FEM, REM methods.
The T-statistics are reported inparentheses. Table 7 presents the results of testing research hypothesis with random effects model (REM). The overall value of R2 for the model is 0.283; and significant at p-value = 0.000, that means 28.3% of the variations in the level of CSR disclosures can be explained by variations in explanatory variables.
We find a positive significant coefficient (β = 0.4318, p < 0.05) for our board independence (BOIND) variable, which supports H1. Independent directors can likely decrease agency conflicts between managers and owners by encouraging management to disclose more CSR activities. This result is also consistent with the research of Khan (2010) for private commercial banks of Bangladesh, Jizi et al. (2014) for US-listed banks and Shu and Chiang (2020) for listed firms in Taiwan. This can be explained by the fact that independent directors are more conscious of promoting their reputation and thus will pay attention to the firm's stakeholder interests when making board decisions. Therefore, independent directors are likely to support the disclosure of CSR activities to reduce information asymmetry between insiders and outsiders.
Contrasting with the previous literature, board size is insignificant for the effects on CSR reporting, with level sig. > 0.1, which does not support Hypothesis 2. We predicted that larger boards would be better able to direct management to engage in CSR activities and publish CSR reports to stakeholders. Although the results of the regression analysis are not as our expectations, the result is consistent with Halme and Huse (1997) in developed nations (including Finland, Norway, Spain and Sweden), Haji (2013) and Sufian and Zahan (2013) in developing countries (Malaysia and Bangladesh), which found no significant association between the number of board members and the tendency for companies to report on the CSR activities.
The relationship between CEO duality and the degree of CSR reporting information is negative but insignificant (p > 0.1). It implies that CEO duality does not influence the degree of CSR reporting information. The finding is inconsistent with Jizi et al. (2014), which found a positive relationship between CEO duality and CSR reporting, and Muttakin and Subramaniam (2015) with a negative relationship between CEO duality and levels of CSR disclosure. The result is consistent with previous research in developing countries such as Malaysia (R. Haniffa & Cooke, 2002;Said et al., 2009), Bangladesh (A. Khan et al., 2013), and Vietnam (Nguyen, 2021b). It can be explained that listed companies in Vietnam comply with regulations on corporate governance (Vietnam Corporate governance code in 2019), with 96% of the sample separated in the position of CEO and chairman of the board.
Hypothesis 4, which predicted that foreign ownership positively influences the degree of CSR reporting information, is empirically rejected (p > 0.1). From view of Stakeholder theory, a higher percentage of foreign ownership results in a greater extent of CSR disclosures. Managers of foreign firms tend to invest more in CSR disclosure in line with the expectations of their shareholders. However, the results of this study are contrary to those of Stakeholder theory and research of R. M. Haniffa and Cooke (2005) in Malaysia, A. Khan et al. (2013) in Bangladesh, and Muttakin and Subramaniam (2015) in India. Even though, we also find similarities in our study results with the study of Amran and Devi (2008) and Said et al. (2009) for Malaysia listed companies in 2002-2003 and 2006 respectively. Perhaps, this is due in part to companies with a higher percentage of foreign ownership use other alternative media rather than the annual reports for CSR purposes. They regularly provide information through the website, quarterly investors meeting to meet the requirements of domestic and foreign investors. In addition, these companies tend to make separate sustainability reports to disclose CSR information.
In Hypothesis 5, the coefficient for the path from GOVOWN to CSRRI is positive and significant (β = 0.1996, p < 0.01). Hypothesis H5 is accepted, which means the proportion of Government shareholders would significantly affect CSR reporting, indicating that companies with higher government ownership are more likely to disclose greater CSR information than other companies. This finding contradicts the results of Dam and Scholtens (2012) in 600 European firms from 16 countries. However, this result confirms the arguments of Devi (2008), Said et al. (2009), and Muttakin and Subramaniam (2015). Government-owned companies may be pressured to disclose their CSR information. It suggests that firms tend to be more politically sensitive; thus, there is a strong expectation for such companies to be aware of their public responsibility. They are more involved in socially responsible actions; they increase the publicity of social activities to legitimize their presence. In addition, government-owned companies in Vietnam are large companies operating in sensitive and essential sectors such as energy, chemicals, mining, etc. Therefore, they tend to disclose more CSR information to satisfy stakeholders.
Finally, regarding control variables, our overall findings suggest that larger firm size (LSIZE), older firms (AGE), and better profitability (ROA) are significantly related to the level of CSR reporting (p < 0.05). The results of our analysis concerning the control variables are consistent with the previous studies (Ghazali, 2007; A. Khan et al., 2013;R. M. Haniffa & Cooke, 2005;Roberts, 1992).

Conclusions and suggestions for further research
This study has examined the level of CSR reporting in the annual reports among Vietnam-listed companies and its association with corporate governance characteristics. Our results suggested that CSR reporting information level is higher than in previous research in Vietnamese listed companies. This shows that these enterprises have become more and more interested in disclosing information about CSR activities in their annual reports. Additionally, we found that board independence and Government ownership positively impacted the degree of CSR reporting information. BOIND has the most substantial influence on the dependent variable. However, we failed to find any significant effect of CEO duality, board size and foreign ownership on CSR reporting. We also found company size, company age, and profitability to be essential determinants of levels of CSR disclosures in Vietnamese CSR reporting. Therefore, CSR information in larger and older firms, higher profitability is more transparent than that of smaller-scale companies and has lower profitability.
The present study makes several significant contributions. First, it contributes to the literature by providing the current State of CSR reporting practices in the listed firms in Vietnam during the COVID-19 period. These results provide deeper insights into CSR reporting in developing countries such as Vietnam. In the context of the covid pandemic, listed companies have actively carried out social responsibility activities with the community and localities; fulfill responsibilities to employees but not pay attention to preserve and protect the environment. Second, our research is helpful to investors by providing an analysis of the relationship between the level of disclosure in CSR reporting and corporate governance characteristics of listed companies in developing countries. Investors should select largescale, older firms with higher percentage of State ownership and independent directors that provide more information about social responsibility and better performance of the business. Finally, the overall findings of our study provide empirical evidence, which adds literature to relationship between corporate governance attributes and CSR reporting. Beside, our result can also help managers and policymakers to promote CSR disclosure by applying the Vietnam Corporate governance code, increasing the percentage of state ownership, extending the number of independent board members.
This study suffers from several limitations. First, although the research focuses on the top 100 companies that accounted for more than 80% of the whole market's transaction value, it leaves the small listed companies unchecked. Using a comprehensive sample may add new insights on CSR reporting in Vietnam. Second, our analysis focused on only disclosures in CSR reporting-part of the annual report. The annual report is a fundamental channel for the dissemination of information for both non-financial and financial data to stakeholders. Future research may consider disclosures in other channels, such as the internet, newspapers, shareholder meetings, etc. Third, we assessed only three aspects of board composition (independence, size, and CEO duality), but prior studies suggest that board experience and expertise could lead to better governance (Gul & Leung, 2004;Khan, 2010;). Future studies may consider assessing the impact of board characteristics such as director expertise and women representation on CSR disclosure. Finally, a cross-country study should be undertaken to understand and compare the effect of corporate governance regulations, and disclosure requirements on preparing and publishing CSR reports in developing countries in Southeast Asia.