Environmental disclosure in corporate websites: a study in Brazil and USA companies

Purpose – The internet allows much corporate information to be instantly accessed from anywhere, at any time. To better inform the more diverse stakeholders, companies have used their websites as another tool for disclosure. The purpose of this paper is to contribute to the area of environmental accounting, as it investigates whether the companies located in different countries, from different sectors, in different stages of development and regulatory environments present different levels of environmental disclosure and to explain the environmental disclosure extension on corporate websites of companies in Brazil and the USA through corporate characteristics. Design/methodology/approach – To achieve such purpose, an environmental disclosure index (EDI) was created and a model was used to investigate whether the variables environmental performance, size, profitability, debt, sector and country explain the disclosure on the website. Findings – It was pointed that US companies stood out compared to Brazilian companies throughout the EDI. On the one hand, the statistical model suggests that the variables, namely, organization size, sector and country of origin of the company, explain the environmental disclosure in corporate website, whereas the profitability and debt variables were not significant in the model. On the other hand, the environmental performance variable proved to be significant; however, it was contrary to what was expected from the theory of legitimacy, once a negative relation between environmental disclosure and environmental performance is expected. Originality/value – It is considered that transnational studies on corporate environmental responsibility can improve the understanding and eventually explain the difference of this disclosure.


Introduction
The environmental responsibility in the corporate field has arisen mainly in the last decades, thus initiating companies' global actions in search for a promotion of acts towards the environment and they have aimed to explain such actions to the external public. According to Deegan (2017) currently one of the main debates in environmental accounting is "accountability" that seeks to understand the level of responsibility of organizations with other stakeholders, since companies must provide information to users so that they make a decision whether do or do not support an organization.
Companies can use different media to promote an environmentally correct image as well as manage stakeholder¨s perceptions, one of which is from corporate websites (Patten & Crampton, 2004). Managers promote disclosure in different media in order to serve all the information users, eventually disclosing voluntary information addressed to investors, while some disclosures are for the benefit of customers, employees, the press, the general public and other interested people (De Villiers & Van Staden, 2011).
Companies therefore use resources from the internet and corporate websites to design a socially acceptable approach to environmental management for the group of people who are interested in . Thereby, organizations identify potential benefits in disclosing environmental information on a website, not only based on print management but also on how management can try to transmit shareholders value throughout environmental responsibility (Cormier, Leboux, & Magnan, 2009).
The information posted on a website has the ability to provide timely and continuous information, rather than periodic information such as printed reports, having benefits ranging from reduced costs to the ability to serve a wider range of stakeholders (Lodhia, 2010). The choice of investigating websites is therefore, due to the fact that it is a quick access tool, where information can be disclosed to a wide range of stakeholders, as well as it is a channel that companies use to disclose a diversity of environmental information and strategies (Aerts, Cormier, & Magnan, 2008;Andrikopoulos & Kriklani, 2013;Tagesson, Blank, Broberg, & Collin, 2009).
Formerly studies on environmental accounting focused only on developed countries, however, companies located in emerging countries have also seek for better environmental accountability towards international markets and even the domestic market. Currently, there is a need for transnational studies on environmental responsibility because previously studies focused only on developed nations, resulting in the need to understand how social and environmental responsibility is addressed in countries within different economic and social contexts (Xiao, Gao, Heravi, & Cheung, 2005).
The option of using two groups of companies is justified by the possibility of investigating whether organizations located in different countries, at different stages of development and regulatory environments have different levels of environmental disclosure, as well as if companies in different contexts identify economic opportunities in voluntary environmental disclosure through websites (Bagnoli, Wang, & Watts, 2014;Barbu Dumontier, Feleag a, & Feleag a, 2014;Gamble, Hsu, Jackson, & Tollerson, 1996;Xiao et al., 2005).
The choice of investigating companies based in the USA and Brazil is backed by the fact that environmental disclosure is strongly promoted in the USA, as well as it is concerned to a mature securities market operating in a highly regulated environment (Bagnoli, Wang, & Watts, 2014). However, it is important to highlight that while there are social problems in emerging countries such as Brazil, companies must still play an important social role in providing well-being to society and addressing their concerns to natural resources (Wanderley, Lucian, Farache, & De Souza Filho, 2008). Besides, an excellent way to evaluate Brazilian environmental disclosure is to compare it to more economically developed countries, with efficient markets and those ones with greater informational demands (Ribeiro, Nascimento, & Van Bellen, 2009).

RAUSP 55,3
As for the theoretical contribution, the research supports the legitimacy theory, which postulates that companies with lower environmental performances will take over a greater volume of environmental disclosure in the search of legitimacy for stakeholders (Gomez-Gutierrez & Cormier, 2019;Patten, 2002). As environmental disclosure through the internet and websites demonstrates a concern for legitimacy by companies that have poor environmental performance, rather than necessarily increasing corporate responsibility .
In this context, the general objective of the study is to explain the extent of environmental disclosure on corporate websites of companies located in Brazil and the USA through corporate characteristics. As for the specific objectives of the study, they are as follows: identify through an Environmental Disclosure Index (EDI) the extension of environmental disclosure on the investigated companies' websites; point out whether there is a difference in the extension of environmental disclosure between companies in Brazil and the USA; and to verify which variables explain the environmental disclosure in corporate website.

Theoretical framework 2.1 Environmental accounting and disclosure in websites
The current literature on environmental accounting can be classified into three stages, the first study group examines the relevance of corporate environmental information to users seeking to assess environmental liabilities, the second study line examines what factors promote disclosure and at last, another group of studies has sought to understand the relation between disclosure and environmental performance (Clarkson, Li, Richardson, & Vasvari, 2008).
According to De Villiers and Van Staden (2011), companies are commonly sharing environmental information through annual reports, corporate websites and social and environmental reports, but nevertheless, websites and annual reports have distinct users looking for a differentiated range of information. Therefore, online platforms are opportunities for companies to bring out more environmentally responsible information to light and provide public relations opportunities with stakeholders .
Previously, corporate information was traditionally evidencedthrough annual reports or print mediahowever, this communication could not reach all stakeholders and it was necessary to rethink communication strategies, using Web disclosure . One reason companies choose to disclose more environmental information on a website is because the cost of adding a dedicated environmental information space on the corporate website is lower than adding information in the annual report. Patten and Crampton (2004).
The alternative of providing information through the Web can offer advantages over the traditional print format, such as the ability to disclose more information to a wider range of stakeholders at a lower cost and at more regular posts than traditional reports (Williams & Pei, 1999). As well as the communication of environmental and social information through corporate websites have the ability to address multiple targets, besides serving a wide range of stakeholders, this disclosure has global range and can reach people anywhere in the world (Lodhia, 2010) In her study, Lipe (2018) states that companies have been looking for an appropriate "information package" to be disclosed and the elements that set up this package are essential to form the external image, examples of elements of this package are the interaction with the user, the language, personal image, as well as the mode in which the nformation will be disclosed.

Environmental disclosure
2.2 Legitimacy theory in the environmental disclosure context Among the theories commonly related to corporate evidence are economic theory, stakeholder theory, disclosure theory, contingency theory and legitimacy theory (Gray, Kouhy, & Lavers, 1995). In general, these theories say that companies that experience greater social and political pressures show more social and environmental information, whether they are companies with good environmental performance or not (Patten & Crampton, 2004). The theory of legitimacy is based on the idea that a company will use corporate disclosure in a way that influences society's perception, which will identify its behavior as acceptable (Adams, Hill, & Roberts, 1998;Campbell, 2003;Cormier & Magnan, 2013;Deegan & Rankin, 1997;O'Donovan, 2002;Patten, 1992).
The theory of legitimacy therefore highlights that the greater the adverse changes in society's perception of the company, the more convenient it will be to try to manage such social impressions, so the theory is based on how management can try to achieve compliance or social expectations and values through disclosure (O'Donovan, 1999). Therefore, the theory of legitimacy is supported by the social contract between company and society meaning that the company's ability to benefit interested parties in an economic, social and political way (Magness, 2006). Aerts and Cormier (2009) state that legitimacy is mainly about perception and how it is managed for relevant audiences, so the company evaluates the content, quality and process of environmental disclosure so that disclosure becomes convenient and desirable. Meaning that, companies have been disclosing socio-environmental information to fill the gap between society's expectations and corporate reputation (Campbell, 2003).
A company management will define its environmental strategy to meet financial and non-financial expectations, meaning, the company will seek to legitimize its actions for both financial analysts and other non-financial stakeholders, which can be difficult, because in several situations, company's environmental performance can validate or undermine efforts to legitimize through environmental disclosure (Cormier & Magnan, 2013).

Sample selection
The companies in the sample of this study should do as follows: attend the "Melhores & Maiores Exame" or the "Fortune 500"; in case of Brazilian companies, be listed on Ibovespa; not be part of the financial services and insurance industry; and have an available website.
It is necessary to standardize the sectors, as it makes easy for further analysis. Therefore, for the purpose of this work, companies were classified by the North American Industry Classification System. The companies that are part of the final sample are divided into 12 sectors. Of the 117 companies, 57 are from Brazil and 60 from the USA.

Environmental evidence contentenvironmental disclosure index
The environmental disclosure content of the present work was built from previous studies in this area to investigate the environmental disclosure attributes on a website. For the study purpose only, the EDI has 10 categories and 40 subcategories Table 1. RAUSP 55,3

Collecting data procedure
To investigate the attributes of environmental disclosure, the research reported to official corporate websites at predetermined dates. Website disclosure is able to provide sustainable sources of information targeted to specific stakeholders, such as hyperlinks and menus, which provide personalized Environmental disclosure information (Lodhia, 2010). At this line,  developed a framework for the investigation of environmental disclosure on websites that guided the diagram elaboration that enables the structuring of a corporate website investigation process focused on environmental information (Figure 1). Regarding environmental responsibility reports or sustainability reports published on corporate websites, only those made available from the years 2014 to 2016 were investigated.

Studies hypotheses
From previous studies, the study hypotheses that follow were formulated: H1. A negative relation between environmental performance and the extent of environmental disclosure on a corporate website is expected.

H2.
A positive relation between the size and extent of environmental disclosure on a corporate website is expected.

H3.
A positive relationship between profitability and the extent of environmental disclosure on a corporate website is expected.

H4.
A negative relationship between debt and the extent of environmental disclosure on a corporate website is expected.

H5.
A positive relationship is expected between companies operating in environmentally sensitive sectors and the extent of environmental disclosure on the corporate website is expected.
H6. It is expected that companies in the USA disclosure a greater extent of environmental information on corporate websites than companies in Brazil.]

Data analysis procedure
The current research is based on studies conducted by Patten (2002) (2014) and Singhania and Gandhi (2015) to investigate the validity of the hypotheses through the multiple regression analysis.
where EDI is given as: EDI is given as: To do so, 40 represents the subcategories of environmental disclosure attributes stablished in the environmental disclosure content. Therefore, if the information is disclosed it is assigned 1, otherwise it is assigned 0. Thus, measuring the extent of environmental disclosure on websites, this approach assumes that every item of environmental disclosure content is equal in importance in the model (Singhania & Gandhi, 2015).

Environmental disclosure
The explanatory variables used in this research were based on previous studies ( Figure 2). It is emphasized here that different researches may use different measures, such as the size variable, those ones chosen for this study to satisfy the research objectives.
The "environmental performance" variable is part of a range of corporate indicators developed by CSRHUBV R Sustainability management tools. For the purposes of this research, only the "Environment" indicator was used, which evaluates corporate environmental performance.
For the "sector" variable, it was determined that companies in the "Manufacturing Industry", "Mining Oil & Gas Extraction" and "Electricity, Gas and Water Companies" sectors are part of companies from environmentally sensitive sectors Patten & Crampton, 2004). Therefore, for the regression model were considered 1 for companies from environmentally sensitive sectors and 0 for those companies that are not part of the environmentally sensitive sectors. The same procedure was performed for the country variable, where it was considered 1 for US companies and 0 for Brazilian companies.
Regarding the methodological limitations of the present study, they are: the website investigation, since the visitation of corporate pages occurred at a single moment in time and companies can update, remove and add information without prior notice. Another methodological limitation is the measurement of the EDI, as binary assignment was used.

Results description and analysis 4.1 Descriptive analysis
Among the 57 Brazilian corporate websites investigated, 48 presented a section on environmental responsibility, representing therefore 84% of the Brazilian corporate websites studied. Among the 60 US corporate websites, only 1 company has no section on environmental responsibility. However, it was found that Brazilian companies most commonly present this section already on their homepage, 53% of the cases. While 59% of US companies submitted the environmental responsibility section from the drop-down menu, framework or other page than the homepage, requiring two or three clicks to access. Another step in the investigation of corporate websites was to identify and examine sustainability reports or environmental responsibility report. In this sense, it was found that the sample companies from the USA released more updated environmental reports, with reports mostly from 2015 or related to 2016. Regarding the investigation of the environmental disclosure content for the construction of the EDI, the main considerations between the categories and subcategories are that in the categories "impact of products and processes," "goals and targets" and the category "other environmental disclosures"; all subcategories were percentage more evidenced by US companies. Concerning the "statement of environmental policy," "wastes and odors," "efficient use or water reuse," "environmental accounting" and "soil contamination and reserve for environmental protection" no disparities were found between Brazilian and US companies. In the "energy" category, "conservation and energy saving" is part of the concerns expressed in almost all of the companies in the sample. However, the "development/exploration of new sources" was more evident in the US companies; this is also repeated for the subcategory "future energy saving." Concerning "spills and airs emissions," there was a greater disclosure about information on atmospheric emissions toward the information about spills, for both Brazilian and US companies.

Statistical analysis
In this section, the results on the suggested model about methodological aspects will be presented. The software used for the statistical analysis was the STATAV R 11. Table 2 presents the descriptive statistics of the EDI variables and environmental performance. It is noted that the average of EDIs is higher for US companies, with 21.13, while the average of Brazilian companies is 15.37. Among the US companies is the company with the largest extent of disclosure, 33 EDI. However, the environmental performance of the Brazilian companies investigated seems to have better average performance, 60, while the US companies averaged 58.23. Another point to be noticed is that among Brazilian companies is the company with the highest environmental performance indicator, 77, while among the US companies investigated the highest performance was 72. The minimum values also show good environmental performance by the Brazilian companies, while these ones had a minimum performance of 38, among the US companies the lowest performance was 33.
The results of the proposed multiple linear regression model are described in Table 3. R 2 and the adjusted R 2 show good fit of the data to the model. As well as the small difference between R 2 and adjusted R 2 demonstrates the adequacy of the number of explanatory variables considered. The adjusted R 2 value indicates that 47.76% of the variation of the dependent variable EDI can be explained by the explanatory variables included in the model, Table 4.
The model proved to be significant, since the probability of F-Statistics is less than 0.05, rejecting the hypothesis that the estimated parameters are simultaneously equal to zero. Regarding the regression assumptions, auxiliary tests were performed to verify the absence of problems in the tested model. The normality of the model residuals was verified, and the data were normal. Regarding the questions involving multicollinearity, the VIF statistics identified that all regressors of the model have values below 2, with the variable country with the highest VIF (1.75) and an average of 1.3 among the explanatory variables. Therefore, it can be affirmed that there are no multicollinearity problems. On the other hand, the Breusch-Pagan/Cook-Weisberg test result indicated the absence of heteroscedasticity in the model (prob> x 2 = 0.3172). Therefore, the variable environmental performance (H1) was significant for the model (p-value = 0.000). However, the hypothesis is rejected since it proved to be opposite to the expected, as the estimated parameter indicates a positive differential of 0.42 percentage points in the EDI for each environmental performance variance.
The variable size (H2) was significant at a significance level of 5% (p-value = 0.026), indicating that there is a positive relationship between the size and extent of environmental disclosure on a corporate website. This result resembles those from studies made by Gray, Javad, Power, and Sinclair (2001), Brammer and Pavelin (2006), Liu andAnbumozhi (2009), Cho et al. (2010), da Monteiro and Aibar-Guzmán (2010), Rover, Tomazzia, Mucia, and Borba (2012), Thompson andKe (2012), D'Amico et al. (2014), mainly to the findings of research made by Patten and Crampton (2004) and Andrikopoulos and Kriklani (2013), since the object of study was also the corporate websites.
The explanatory variable profitability (H3) was not statistically significant at the 5% confidence level (p-value = 0.173), it is rejected the hypothesis that there is a positive relationship between profitability and the extent of environmental disclosure on a corporate website. The studies by Thompson and Ke (2012) and  also applied ROA as a measure of profitability. However, unlike the present research, they had results that indicated a positive relationship between profitability and environmental disclosure. Another hypothesis is debt (H4), since it was not significant within the model (p-value = 0.457), it is rejected the hypothesis that there is a negative relationship between debt and the extent of environmental disclosure on a corporate website. Unlike the results of studies made by Brammer and Pavelin (2006), Andrikopoulos andKriklani (2013), D'Amico et al. (2014), Singhania and Gandhi (2015), resembling only the study made by Liu and Anbumozhi (2009). Environmental disclosure However, the explanatory variable sector (H5) was significant at the 1% confidence level (p-value = 0.001). This result resembles to the research made by Gray et al. (2001), Patten and Crampton (2004), Wanderley et al. (2008), Liu and Anbumozhi (2009), Suttipun andStanton (2012), D'Amico et al. (2014), Singhania and Gandhi (2015), as they conclude that environmentally sensitive companies provide more information than companies without high environmental impact. The estimated parameter in the model still indicates an increase ratio of 4.1 percentage points when the company is classified as environmentally sensitive.
The variable country (H6) was also significant at a significance level of 1% (p-value = 0.004), indicating that the company's origin explains the extent of environmental disclosure on the corporate website. Therefore, the estimated parameter points to a positive differential of 5.6 percentage in the EDI, when the company is located in the USA. This result resembles to research results made by Xiao et al. (2005), Wanderley et al. (2008), Liu and Anbumozhi (2009) and Singhania and Gandhi (2015) who conclude that the extent of environmental disclosure differs between developed and developing countries.
Finally, models were run separately for Brazilian and US samples. However, there were no distinct results than the results found in the complete model, so it was decided not to discuss them among the research results.

Discussion
This paper aimed to explain the extent of environmental disclosure on corporate websites of companies located in Brazil and the USA through corporate characteristics. One of the main methodological limitations of this study was investigating corporate websites, however it was found that the proposed diagram for this investigation of environmental disclosure, which was elaborated from the study by , was extremely important to guide the research, although there are large disparities between the websites investigated, such as layouts and distinct interfaces.
Overall, US company websites were more inviting, with graphics, animations, and layouts in the sections about environmental responsibility. The built environmental disclosure content also was proved to be adequate when investigating corporate websites, and there was no need during the search for new adaptations or totally non-existent subcategories, as well as being able to properly capture the environmental disclosure of companies from two different countries.
The first specific objective was to build EDIs for the investigated companies, and among the main conclusions are that on average, US companies have higher EDIs (21.13) than Brazilian companies (15.37). The second specific objective was to point out whether there is a difference in the extent of environmental disclosure between companies located in Brazil and the USA. From the analysis of the filling of the environmental disclosure content and, consequently, the EDI construction, it can be stated that there was a greater extension of disclosure by US companies than the disclosure on Brazilian corporate websites. Although in some subcategories the disclosure is similar between companies in the USA and Brazil, companies in the USA stood out in most of the subcategories of EDI. Except the subcategory "areas protected", where there was a greater disclosure by Brazilian companies.
After investigating websites, the third specific objective was to verify which variables explain environmental disclosure on a corporate website using a multiple regression model. Table 4 summarizes the research results regarding the examination of the previously formulated hypotheses.
Overall, the research results suggest that the variables organization size, the business sector and the country of origin of the company explain the extension of environmental RAUSP 55,3 disclosure on a corporate website. While the profitability and debt variables were not significant in the model. Yet, the variable environmental performance was significant, however, opposite to what was expected in the theory of legitimacy, since a negative relationship between environmental disclosure and environmental performance was expected. Although there is strong evidence that companies use environmental disclosure as a tool of legitimation, we identified the inverse through the variables used for this research, better performing companies have higher EDIs rates.

Final considerations and conclusions
The environmental performance measure used in this research may have its limitations and there was no control over such variable, as they were data from CSRHUBV R Environment and, therefore, may or may not correctly capture the environmental performance of the investigated companies. Although it was unable to reach conclusions on the theory of legitimacy, the study sought to obtain answers through statistical tests regarding the relation environmental performance versus environmental disclosure.
The results of the regression model, however, suggest a greater extension of environmental disclosure when the company is located in the USA, confirming the research by Xiao et al. (2005), Wanderley et al. (2008), Liu and Anbumozhi (2009) and da Monteiro and Aibar-Guzmán (2010).
Despite its methodological limitations, this study contributes to research on environmental accounting, as it examines non-traditional means of accounting as well as it improves understanding of environmental disclosure by companies in Brazil and the USA. In this sense, the attempt to understand the characteristics of environmental information disclosed by companies through electronic means or websites is still incipient in international literature.  (2019) Environmental disclosure