Corporate Characteristics and Sustainability Disclosure in Bangladesh: An Analytical Study Based on Global Reporting Initiative Guideline 4

Purpose – The purpose of the research is to examine the level of compliance of sustainability disclosure as per GRI G4 framework and the influence of corporate characteristics on sustainability disclosure of listed companies in Bangladesh. Design/methodology/approach – A sample of 175 listed companies drawn from the Dhaka Stock Exchange (DSE) listed companies using purposive sampling from different categories and data collected from the annual reports of 2018 through content analysis. The hypotheses were tested using ordinary least square regression (OLS). Findings – Statistical results witness that the sustainability disclosure index (SDI) is inferior (12.19) with a high deviation (SD 9.61). This study also documents that industry membership, ISO certification, multi-nationality, and board size are positively associated with sustainability disclosure at a five per cent level. Still, the company category is negatively associated with sustainability disclosure at a six per cent level. This paper also found that sustainability disclosure is not likely to be significantly influenced by age, profitability, and leverage. Research limitations/implications - The research used content analysis to measure quantity ignoring the quality of sustainability disclosure based on GRI G4 guidelines from the company's only one-year published annual report. Practical implications - The research adds value to the sustainability disclosure literature and provides a message to the policy planners and practising authorities. Originality/value – This is one of the pioneer studies using GRI G4 guidelines to measure the extent of sustainability disclosure and examine the influence of corporate characteristics on sustainability reporting in Bangladesh, considered a developing nation with an emerging economy.

Sustainability disclosure has become a progressively relevant issue in business and academia at the end of the 1990s (Hahn & Kuhnen, 2013). Knowledge and technological development can encourage sustainable economic growth and defuse the crisis threatening society, the environment, and the world economy (Wang, 2017). Companies are trying to increase transparency, augment brand value, reputation and legitimacy, enable benchmarking against competitors, signal competitiveness, encourage employees, and care about corporate information and control procedures through relating sustainability information (Herzig & Schaltegger, 2006, cited in Hahn & Kuhnen, 2013. Every company should conduct a wideranging materiality assessment, integrating stakeholder engagement in recognising and reporting unique sustainability impact (Puroila & Mäkelä, 2019). Despite these benefits, it seems that some companies are still unenthusiastic to adopt this practice, as it involves added responsibilities and challenges for companies (Bhatia & Tuli, 2017a). However, there is a positive association between sustainability disclosure and earnings informativeness (Swarnapali, 2019). For instance, ecologically polluting businesses disclose significantly higher sustainability information than non-polluting industries in India (Kumar et al., 2021). Corporate governance has an influence on sustainability reporting in Bangladesh (Tasnim & Khan, 2022). Unlike obligatory financial disclosures, sustainability disclosure is voluntary and measured as an act of transparency (de Villiers & Marques, 2016, cited in Semuel, Hatane, Fransisca, Tarigan, & Dautrey, 2019. A standardized method of sustainability reporting would help to reduce the limitations of voluntary initiatives and recover the overall voluntary reporting mechanisms (Coulmont, Berthelot, & Gagne, 2022). However, consciousness and empathy about sustainability reports in Indonesia are still deficient while the government has made sustainability reports obligatory (Wahyuningtyas, Susesti, & Murtadho, 2022). Global Reporting Initiative (GRI) is dedicated to developing and providing universal guidance for sustainability disclosure (Wachira, Berndt, & Romero, 2019;Islam, 2020). GRI guidelines are developed in connotation with specialists from all stakeholder groups (Christofi et al., 2012) so that the framework should use unless a more comprehensive and better framework developed due to it is the only framework that facilitates disclosure of the complete scenario of the sustainability performance of the firm (Laskar & Maji, 2016).
From the above standpoint, the matter is imperative to find out whether there is any relationship between the volume of sustainability disclosure based on GRI G4 and corporate characteristics of DSE listed companies in Bangladesh. The study is expected to enrich the literature on sustainability reporting and provide valuable guidelines to the policymakers and practising authorities to implement sustainable business so that the world will be comfortable for the future generation.

Review of Literature
The following literature on the relevant field from home and abroad was reviewed to find the research gap and formulate hypotheses. In the study, the independent variables have been taken with interpretation from previous studies by other researchers. Each of the company traits deliberated in turn and established hypotheses based on the relation with the extent of sustainability disclosure proposed below.

Company Categories
A few numbers of researchers used industry type as an explanatory variable in the corporate sector. Rao and Tilt (2016) found that industry type has some influence on CSR disclosure. Dissanayake, Tilt, and Qian (2019) ;Shamil, Shaikh, Ho, and Krishnan (2014) found no association between industry type and sustainability disclosure. Mudiyanselage (2018) found that there is an insignificant relationship between industry type and CSR. Bhatia and Tuli (2017a); Girón et al. (2021) found a positive relationship between company nature and sustainability disclosure. Sarkar (2021) recognised that sustainability reporting is meaningfully linked to the company category.
Based on this background, company categories are expected to influence sustainability disclosure and the first research hypothesis formulated regarding company categories.
H1: There is a relationship between company categories and the extent of sustainability disclosure. Branco, Delgado, Gomes, and Eugenio (2014) and Sarkar (2021) found that industrial affiliation influences sustainability disclosure, but there is no sufficient study under the review, which considered industrial membership as a corporate characteristic. The survey attempt to assess the relationship between the industrial membership of the company and the extent of sustainability disclosure. Therefore, the second hypothesis was formulated regarding the industrial membership of the company.

Industry Membership
H2: There is a relationship between the industry membership of the company and the extent of sustainability disclosure.

ISO Certification
Many theoretical and empirical research has examined the relationship between company recognition and the length of environmental exposure. Ezhilarasi and Kabra (2017) ;Sarkar, Ahmed and Islam (2020); Yusoff, Othman, and Yatim (2013) showed a significant influence of ecological certification on the environmental disclosure of companies. Sarkar (2021) identified that sustainability reporting is significantly related to the ISO certification of the company. Therefore, the third hypothesis was formulated regarding the ISO certification of the company.
H3: There is a significant relationship between the ISO certification of the company and the extent of sustainability disclosure. Bhatia and Tuli (2017a); Bae, Masud, and Kim (2018) found a positive relationship between the multi-nationality of the company and sustainability disclosure, whereas Wang (2017); Tasnim and Khan (2022) explore the foreign shareholders' holdings positively related to the disclosure of sustainability reporting. Anazonwu, Egbunike, and Gunardi (2018) found a significant favourable influence of board members' nationality on sustainability disclosure. Discipline, Khulna University, Volume: 17, Number: 1, January-December 2022, pp. 103-126, DOI: 10.35649/KUBR.2022.17.1.6 (Print ISSN: 1811Online ISSN: 2664-3502) Sarkar (2021) identified that sustainability reporting is significantly related to the multinationality of the company. Majeed, Aziz, and Saleem (2015) found contrary relationships between foreign directors' representation in the board and CSR reporting.

Business Review-A Journal of Business Administration
Based on these different possibilities from empirical studies, a positive relationship between the multi-nationality of the company and the extent of sustainability disclosure is expected. Therefore, the fourth hypothesis was formulated regarding the multi-nationality of the company.
H4: There is a positive relationship between the multi-nationality of the company and the extent of sustainability disclosure.

Age of the Company
Several studies found a significant positive association between the age of the company and the extent of sustainability disclosure (Bhatia & Tuli, 2017a;Buallay & Al-Ajmi, 2019;Kumar et ., 2021), whereas Mudiyanselage (2018) found that there is a significant negative relationship between the age of the company and sustainability disclosure. Some of the studies found no statistically significant relationship between the age of the company and sustainability disclosure (Dienes, Sassen & Fischer, 2016). Shamil et al. (2014) found that younger firms are likely to adopt sustainability reporting. Sarkar (2021) found that sustainability reporting is significantly related to company age.
Based on these contradictory suppositions from empirical studies, a positive or negative relationship between the age of the company and the extent of sustainability disclosure is expected. Therefore, the fifth hypothesis was formulated regarding the age of the company.
H5: There is a relationship between company age and the extent of sustainability disclosure.

Board Size
A few researchers used board size as an explanatory variable for differences in disclosure level in the corporate sector. Most of the studies (Bae et al., 2018;Giannarakis, 2015;Hu & Loh, 2018;Mahmood, Kouser, Ali, Ahmad, & Salman, 2018;Majeed et al., 2015;Mudiyanselage, 2018;Olayinka, 2021;Shamil et al., 2014;Wang, 2017) showed a significant positive relationship between board size and CSR/ sustainability disclosure. However, some studies recognise no statistically significant relationship between board size and sustainability/ CSR disclosure (Rao & Tilt, 2016;Sarkar, 2022). Handajani, Subroto, Sutrisno, and Saraswati (2014) found that board size significantly affects corporate social disclosure. Akhtaruddin, Hossain, Hossain, and Yao (2009) showed a significant positive relationship between board size and voluntary disclosure. Tjahjadi, Soewarno, and Mustikaningtiyas (2021) found size has a positive consequence on economic, an adverse effect on communal and no effect on ecological sustainability performance. Tasnim and Khan (2022) found a negative impact of board size on sustainability reporting. Based on these empirical studies, a positive relationship between the board size of the company and the extent of sustainability disclosure is expected. Therefore, the sixth hypothesis was formulated regarding the board size of the company. Discipline, Khulna University, Volume: 17, Number: 1, January-December 2022, pp. 103-126, DOI: 10.35649/KUBR.2022.17.1.6 (Print ISSN: 1811Online ISSN: 2664-3502) H6: There is a positive association between board size and the extent of sustainability disclosure.

Profitability
Some of the studies show a significant positive relationship between sustainability disclosure and the profitability of the company (Branco et al., 2014;Dilling, 2010). On the other hand, some researchers recognise no statistically significant relationship between profitability and sustainability/ CSR disclosure (Dienes et al., 2016;Rao & Tilt, 2016;Shamil et al., 2014). Atan, Razali, Said, and Zainun (2016) found no association between environmental social and governance (ESG) disclosure level and the firm's financial performance. Similarly Wahyuningtyas, Susesti, and Murtadho (2022) found no substantial consequence of sustainability reporting on financial performance of companies. Tarmuji, Maelah, and Tarmuji (2016) found the influence of ESG practices on economic performance. Kee, Li, Sidik, Seng, and Suppiah (2020) observed ROE has a significant influence on ESG scores. Giannarakis (2014) found a positive relationship between profitability and CSR disclosure. Argento, Grossi, Persson, and Vingren (2019); Mudiyanselage (2018) found a positive correlation between profitability and sustainability disclosure. Orazalin and Mahmood (2019) found that firm profitability substantially influence the extent, nature and quality of sustainability-reporting practices. However, Bhatia and Tuli (2017a); Kumar et al. (2021) found a significant negative relationship between sustainability disclosure and the company's profitability. Sarkar (2021) found that sustainability reporting is significantly related to the profitability of the company. Pham, Do, Doan, Nguyen, and Pham, (2021) observed a positive relationship between corporate sustainability and financial performance.
Based on this contradictory conclusion from empirical studies, a significant positive, significant negative, or no meaningful relationship between profitability and the extent of sustainability disclosure is expected. Therefore, the seventh hypothesis was formulated regarding the profitability of the company.
H7: There is a relationship between company profitability and the extent of sustainability disclosure. Shamil et al. (2014) identified a significant positive relationship of sustainability disclosure with leverage, whereas some studies identified a significant negative association between sustainability/ CSR disclosure and leverage (Bhatia & Tuli, 2017a;Branco et al., 2014;Giannarakis, 2014;Habbash, 2016;Kumar et al., 2021). Mudiyanselage (2018) ;Hongming, Ahmed, Hussain, Rehman, Ullah, and Khan (2020) found that there is an insignificant relationship between leverage and CSR/ sustainability reporting. On the other hand, Sarkar (2021) identified that sustainability reporting is significantly related to the company's leverage.

Leverage
Several studies have found different results, either positive, negative or insignificant, between leverage and sustainability/ CSR disclosure. Therefore the eighth hypothesis has developed regarding the leverage of the company: Business Review-A Journal of Business Administration Discipline, Khulna University, Volume: 17, Number: 1, January-December 2022, pp. 103-126, DOI: 10.35649/KUBR.2022.17.1.6 (Print ISSN: 1811Online ISSN: 2664-3502) H8: There is a relationship between leverage and the extent of the sustainability disclosure.

The Methodology of the Study
The empirical research is based on secondary data sources collected through content analysis of the company's annual report listed in DSE, Bangladesh, based on GRI G4 guidelines.

Population and Sample
There were 316 companies listed in 2018 with Dhaka Stock Exchange (DSE), Bangladesh. The sample of the study was selected based on Krejcie and Morgan's table (1970 cited in KENPRO). The table suggested 175 companies as a sample (175 samples for the population size of 320 in the case of the finite population). A purposive and judgmental basis sample is used to select the companies from 18 categories based on DSE classification (Panel C of Table-2).

Measurement Procedure
Earlier researchers used different measurement procedures to assess the level of sustainability disclosure practices. Akter, Akter, and Akhter (2018) Nur et al. (2016) used the UN Global Compact framework to measure sustainability disclosure. Content analysis is a commonly used means in social science research for mining information in a numeric arrangement from the published report (Laskar & Maji, 2016). The study used a content examination of the annual report 2018 to develop a sustainability disclosure index (SDI) based on the GRI G4 guidelines. The annual reports are painstaking as a source of data because it is obligatory as required by legislation. Moreover, all listed companies recurrently produce it and are quite easy to compare (Tilt, 2001, cited in Akbas, 2014. Discipline, Khulna University, Volume: 17, Number: 1, January-December 2022, pp. 103-126, DOI: 10.35649/KUBR.2022.17.1.6 (Print ISSN: 1811Online ISSN: 2664-3502)

Information Items Included in the Index
Ninety-three items of GRI G4 guidelines are used to develop an appropriate compliance index on sustainability disclosure practices.

Scoring in the SDI
The dependent variable is indomitable as sustainability disclosure score (SDS) of each company as follows: SDS Where, d= 1 if the company disclosed the item di d= 0 if the company does not disclose the item di n= number of items SDI of each company is computed by using the following formula:

Data Analysis Techniques
The research used descriptive statistics to measure the level of sustainability disclosure in the corporate sector in Bangladesh. In addition, the ordinary least square (OLS) regression model uses to investigate the relationships between corporate characteristics and the level of sustainability disclosure.

Dependent and Independent Variables
The SDI has been considered the dependent variable for each company studied, seeing collected data based on GRI G4. The explanatory variables cast off in the study have taken into account previous studies undertaken by other researchers.

Figure 1: Independent and Dependent Variable
Source: Self Constructed.
Description of independent variables, their labels, expected signs, and relationships are presented in Table-1.  Volume: 17, Number: 1, January-December 2022, pp. 103-126, DOI: 10.35649/KUBR.2022.17.1.6 (Print ISSN: 1811Online ISSN: 2664-3502)  LEV has a relationship with the levels of sustainability disclosure H8 +/-Eight corporate attributes considered are company categories (proxied by DSE classification based on business categories), industrial membership (proxied by either the company is member of an industrial association or not), ISO certification (proxied by whether the company ISO certified or not), multi-nationality (proxied by whether the company multinational or not), age (proxied by year of establishment), board size (proxied by the number of directors in the board), the profitability of companies (proxied by EPS in 2018), and leverage (proxied by debt to assets ratio).

Multiple Regression Models
Multiple linear regression models developed for the study problem as- MULT: multi-nationality of the company i (categorical variable, it takes 1 for multi-national and 0 for non-multi-national companies) AGE: year of establishment of the company i as of 2018.

Results and Discussion
The section result of the study presented and discussed dividing three parts. In the first part, descriptive statistics of the independent and dependent variables are presented in the table and brief descriptions. The second part presents the Pearson correlation matrix to show the relationship among variables with a brief clarification. In the last part, an ordinary least square regression model was developed. Panel C testimony that only about one-fifth of the company have industry membership and companies from different categories selected with a justifiable share of the population.  Khulna University, Volume: 17, Number: 1, January-December 2022, pp. 103-126, DOI: 10.35649/KUBR.2022.17.1.6 (Print ISSN: 1811Online ISSN: 2664-3502)  Business Review-A Journal of Business Administration Discipline, Khulna University, Volume: 17, Number: 1, January-December 2022, pp. 103-126, DOI: 10.35649/KUBR.2022.17.1.6 (Print ISSN: 1811Online ISSN: 2664-3502)

Correlation Matrix
Table-3 presents the Pearson correlations matrix between the dependent and independent variables. I did this to check whether any multicoliniarity exits among the variables included in the regression model. The result of the correlations analysis indicates that the highest correlation coefficient between independent variables is -0.381 for the company category and industry membership of the company. Thus, there is no unacceptable level of multicollinearity between the independent variables. According to Farrar and Glauber (1967) the correlation between independent variables should not be considered harmful until the correlation coefficients reach 0.8 or 0.9 (Akbas, 2014). Source: Analysis of data.

Regression Results
In table-4 the estimated value for company category is -.228 and its t-value is -1.896 with pvalue 0.060, the estimated value for industry membership is 6.788 and its t-value is 8.204 with p-value 0.000, the estimated value for ISO certification is 3.286 and its t-value is 2.714 with p-value 0.007, the estimated value for multi-nationality of company is 4.174 and its tvalue is 2.588 with p-value 0.011, the estimated value for age of the company is -0.053 and its t-value is -1.397 with p-value 0.164, the estimated value for board size is 0.439 and its tvalue is 2.952 with p-value 0.004, the estimated value for profitability is 0.110 and its t-value is 1.211 with p-value 0.228, the estimated value for leverage is 0.332 and its t-value is 0.945 with p-value 0.346. Statistical results indicate that the company category has an insignificant relationship at 5 per cent but a significant negative relationship with sustainability disclosure at a 6 per cent level. In contrast, industry membership, ISO certification, multi-nationality Business Review-A Journal of Business Administration Discipline, Khulna University, Volume: 17, Number: 1, January-December 2022, pp. 103-126, DOI: 10.35649/KUBR.2022.17.1.6 (Print ISSN: 1811Online ISSN: 2664-3502) and the board size have a positive and statistically significant relationship with sustainability disclosure at a 5 per cent level of significance. But the company's age, profitability, and leverage have no statistically significant relationship with sustainability disclosure. The nonexistence of multicollinearity in the data because the variance inflation factor (VIF) values for all eight independent variables are less than two. Result of the correlation matrix testimony that there is no variable with a higher correlation in the data set. The Durban Watson test statistics value is 1.109, in the normal range of 1.0 to 2.5. Field (2009) suggests that values under one or more than 3 are a definite cause for concern. So, the result indicates that there is no autocorrelation. The value for this model is 0.490, and the Adj value is 0.465, which implies that the predictor variables can explain about 49.0 per cent of total variation by and about 46.5 per cent of total variation by Adj .

Conclusion
The objective of this study was to investigate the relationship between the extent of sustainability disclosure (SDI) based on GRI G4 guidelines and corporate characteristics using secondary data collected through content analysis of the annual report 2018 of 175 companies listed in DSE, Bangladesh. The disclosure items of G4 were used as a measure of the extent of the sustainability disclosure index. In addition, eight corporate characteristics are considered as independent variables based on the previous literature. Descriptive statistics and inferential statistics were used to analyse the data through SPSS (Statistical Packages for Social Science) version 20.
Descriptive statistics indicate that the mean SDI is 12.19 with a high deviation (standard deviation 9.61 and range 42.39). The mean age of the selected companies is 26.94 from 4 to 124 years old in 2018. Selected companies include both profitable and losing companies. Only about one-third of the companies have ISO certificates, one-fifth of the companies have industry membership, and about one-sixth of the companies are multi-national. The correlation matrix indicates no unacceptable level of multicollinearity in the independent variables because the highest correlation coefficient between independent variables is -0.381 between industry membership and company category. The regression result indicates that four out of eight hypotheses supported at 5 per cent level of significance and another one significant at 6 per cent significance level. The empirical result means that the company category has an insignificant negative relationship with the extent of sustainability disclosure at 5 per celt significance level but significant at 6 per cent Business Review-A Journal of Business Administration Discipline, Khulna University, Volume: 17, Number: 1, January-December 2022, pp. 103-126, DOI: 10.35649/KUBR.2022.17.1.6 (Print ISSN: 1811Online ISSN: 2664-3502) level, contrary to the expectation (hypothesis 1). Thus, the result of sustainability disclosure and the company category is inconsistent with the previous studies if we considered 5 per cent level, but it is consistent at 6 per cent level. On the other hand, industry membership (hypothesis 2) provide supporting evidence that there is a significant positive relationship between industry membership and the extent of sustainability disclosure. Furthermore, there is a meaningful positive relationship between the ISO certification of the company and the extent of sustainability disclosure (hypothesis 3), multi-nationality of the company and the extent of sustainability disclosure (hypothesis 4), and between the board size of the company and the extent of sustainability disclosure (hypothesis 6). This result of hypothesis 2, consistent with previous research, indicates that companies with industry membership disclosed more sustainability information in the annual report than those with no affiliation. In addition, the result of hypothesis 3 demonstrates that companies with ISO certification disclose more sustainability information than those without ISO certification. Finally, the result of hypothesis 4 shows that multi-national companies disclose more sustainability information than others.
The study results of hypothesis 6 indicate a positive relationship between board size and the degree of sustainability disclosure, which is consistent with some previous studies. The result implies that large board size companies disclosed greater sustainability information than those with small board size. The result of the OLS regression analysis does not provide statistical support for the remaining three hypotheses relating to age (hypothesis 5), profitability (hypothesis 7) and leverage (hypothesis 8). The coefficient for age is negative and statistically insignificant at 5 per cent but significant at 16.4 per cent level of significance. The result has similarities with the findings of Dienes et al., 2016. On the other hand, the coefficient for profitability is positive and statistically insignificant at 5 per cent but significant at 22.8 per cent level of significance. Thus, the result has consistency with Dienes et al. (2016); Rao and Tilt (2016); Shamil et al. (2014) found no statistically significant relationship between profitability and sustainability/ CSR disclosure, and Atan et al. (2016) found no association between ESG disclosure level and the firm's financial performance. Similarly, the coefficient for leverage is positive and statistically insignificant at 5 per cent but significant at 34.6 per cent level of significance. The result has consistency only with Mudiyanselage (2018), who tried to show the relationship between leverage and CSR. This study is not free from limitations. Mentionable limitations are that the study's period covers only one year, only considered annual reports of companies as the source of data on sustainability disclosure, and think only the volume but not the quality of disclosure. This study used data from the annual report 2018 which was latest at the time when the study was conducted. Following this, two more annual report were published and analyse the most recent one would be great and should be considered in future studies. However, this study findings are still important and guide the future research. The model explained 46.5 per cent proportion of SDI by the 8 variables included in the model as independent variables. To explain the unexplained 53.5 per cent, I need to add other possible independent variables. However, I could not do it because of the lack of data.