Banks’ Sustainability Reporting in Brazil

Abstract

The purpose of this study is to evaluate the quality of sustainability reporting from banks operating in Brazil from the perspective of the GRI reporting principles and the coverage of reported content and its correlation with the SDGs. We also examine whether there is some association between the quality of such reporting and some characteristics of the banks (e.g., national vs. foreign, publicly traded vs. privately held, government vs. privately controlled). We used two different methodologies proposed in existing literature to assess quality based on economic, social, and environmental contents and developed our own methodology, based on previous studies and official documents, to analyze compliance with GRI principles. The results indicate that the reports are generally of low quality, although there are some cases of medium or very low quality, demonstrating the need for improvement in their preparation. The best evaluations were for reports produced by private banks that publicly trade, are based in Brazil, and use integrated reporting.

1. Introduction

The business sector has an important role in contributing to the Sustainable Development Goals (SDGs) due to the knowledge it has of resource management and control (Yiu & Saner, 2017). Dressler and Bucher (2018) examined the complexity of the SDGs in the business world, identifying a link between the implementation of the SDGs and possible innovations. Topple et al. (2017) suggested that the use of international sustainability standards and guidelines, such as the SDGs and the Global Reporting Initiative’s (GRI) standards, can help companies achieve SDGs more successfully. Thanks to their position as intermediaries, banks play an essential role in directing financial resources towards sustainable activities (Cerqueti et al., 2023), both through credit policies and investment strategies (Xia et al., 2023). For Avrampou et al. (2019), the financial sector is essential in the transition to a sustainable economy, as large amounts of capital are needed (from 5 to 7 billion dollars/year) and banks are responsible for directing funds to promote sustainable growth.

Over the last 20 years, the global banking system has adopted various “pacts” in favor of sustainability, including the Equator Principles (managing environmental and social risks), PCAF—Partnership for Carbon Accounting Financials (measuring and disclosing carbon emissions from financed activities), and UN Principles for responsible banking (aligning strategy with the UN Sustainable Development Goals and international agreements). The Equator Principles (EPs) are an important risk management framework and set of standards created by the financial sector to address the environmental and social impacts of banks’ loan portfolios for large projects (Light & Skinner, 2021). The Global GHG Accounting and Reporting Standard, launched by the PCAF, was the first standard aimed at financial institutions measuring and reporting greenhouse gas (GHG) emissions in their loan and investment portfolios (Light & Skinner, 2021). These actions by banks aim to reduce emissions and promote environmentally positive actions throughout their value chain.

When financing a project, institutions should adopt the “Triple Bottom Analysis”, evaluating the environmental and social impacts and performances and not just the financial ones of the borrowers (Singh, 2015), or refuse financing to companies that negatively affect vulnerable people and violate human rights (Brightwell & Gardener, 2018).

Krasodomska (2015) argues that banking is certainly not as damaging to the environment as the chemical, pharmaceutical, or extractive industries. Notwithstanding, as emphasized by Simpson and Kohers (2002, p. 108), although not facing “the same social responsibility challenges in terms of pollution, product and worker safety”, banks “do have a social and legal responsibility because they grant loans to companies that pollute, produce unsafe products, etc.”.

At the same time as there is concern about financed emissions and the decarbonization of credit portfolios, the 60 largest banks in the world, including some with activities in Brazil (which are the subject of this study), have invested more than USD 5.5 trillion from 2016 to 2022 in the fossil fuel industry according to a report on financing the fossil fuel sector issued in 2023 (Banking on Climate Chaos, 2023, p. 3).

Forests and Finance (2023, p. 3) showed that, from January 2016 to September 2023, Brazilian financial institutions invested at least USD 127 billion in organizations with a history of socio-environmental violations in the production of beef, soy, palm oil, pulp and paper, and rubber and timber, representing 41% of total global loans.

In 2020, the Central Bank of Brazil included the “Sustainability” pillar in the BC# Agenda, following “an international trend motivated by the identification that climate change is not only a source of social and environmental concern, but also a risk to financial stability” (Banco Central do Brasil, n.d.).

The aim of this work is to assess the quality of the sustainability reports of the main national or foreign banks operating in Brazil, and which use the GRI Standards in their sustainability reporting. To achieve the objectives of the study, the following research questions were developed:

Q1: Are there any differences in the quality of the reports considering control (public vs. private), capital (public vs. private), headquarters (Brazilian vs. foreign), or scope of activity (multiple vs. development)?

Q2: Are there any differences in quality depending on the type of report (integrated versus sustainability/ESG)?

Q3: Has the quality of the reports improved?

Q4: Which banks stood out positively and negatively?

In terms of methodology, the reports were assessed in terms of their coverage of the 17 Sustainable Development Goals using a scoring system in which they are classified as having very low, low, medium, or high quality. They were also assessed for adherence to the Quality Principles defined by the GRI.

Among the main results, it stands out that the highest-quality reports are in the form of integrated reports from private, publicly traded banks based in Brazil. We hope that this study will contribute by helping managers who are responsible for preparing reports at financial institutions, banking regulators, and auditing, consulting, certification, or assurance firms, as well as the academic community interested in researching the reports issued by banks.

Section 2 presents a review of the relevant literature. Section 3 identifies the methodology used. Section 4 presents the results and their discussion. Finally, the last section presents the main conclusions, identifies the main limitations of this study, and suggests avenues for future research.

2. Review of the Relevant Literature

The disclosure of non-financial information has become a relevant issue for preparers, users, and regulators. Recent developments, such as the creation of the International Sustainability Standards Board (ISSB) by the International Financial Reporting Standards (IFRS) Foundation to develop this organization’s sustainability-related disclosure standards and the development of the European Sustainability Reporting Standards (ESRS) (Hummel & Jobst, 2024), highlight the need to make sustainability information comparable, reliable, and relevant.

The reports used to disclose sustainability information can be divided into two main groups: integrated and sustainability. While sustainability reports serve a wide range of stakeholders and communicate the organization’s impacts on the economy, environment, and society, integrated reports explain to investors how it creates value in the short, medium, and long term. According to a KPMG study on sustainability reporting, 86% of the Brazilian companies examined used the integrated reporting structure and 93% reported using the GRI methodology (KPMG, 2023).

The GRI Standards present practices for reporting different economic, environmental and social impacts. Reports based on these standards provide information on an organization’s positive or negative impacts on sustainable development. The GRI Standards have three series that underpin the reporting process: the Universal Standards, which apply to all organizations, the GRI Sector Standards, applicable to specific sectors, and the Thematic Standards, presenting relevant content on economic (200 series), environmental (300 series), and social (400 series) impacts.

Considering the seriousness and likelihood of their impacts, their size, worldwide distribution and the number of financial organizations using the GRI Standards, the Global Sustainability Standards Board (GSSB) has started to develop specific standards for the sector, which will be included in a Sector Standard (Global Reporting Initiative, 2023). The Sector Standards currently cover the Oil and Gas, Coal, Mining and Agriculture, Aquaculture and Fisheries sectors.

The document “From chaos to order” presents the 10 types of change that are taking place with sustainability reports, including moving from voluntary to mandatory, becoming audited, being combined with financial reports and disclosed simultaneously, and being guided by the needs of stakeholders and not just investors (Fincomm Services, 2024). The better the reports are, the easier it will be for stakeholders to assess the actions and impacts of banking activities on the environment and society and, thus, make more informed decisions about their relationship with financial institutions.

Daub (2007) analyzed the GRI reports of 103 large Swiss companies, using a system of weights and scores in four performance groups (economic, social, environmental, and integrated) and found that the majority of large Swiss companies still have a long way to go to meet the requirement to disclose their contribution to the sustainable development of the economy, society, and the environment in their reports. On average and across all sectors, companies achieved a third (33.33%) of the total possible points.

Other studies used different ways of assessing the sustainability reports’ quality. Analyzing the 2011 reports of 60 Brazilian publicly traded companies, Ching et al. (2013) assigned a score of 0 or 1 to the reporting of environmental, social, and economic aspects, concluding that those listed on the stock exchange’s Corporate Sustainability Indicator (ISE) produced better-quality reports. Hąbek and Wolniak (2016) used 17 evaluation criteria grouped into the categories “relevance” and “credibility” of the information in 507 reports from six EU countries with a “low” quality score. Yadava and Sinha (2016) analyzed 104 reports from major Indian companies and assigned scores from 0 to 3 for each of the 84 economic, social, and environmental performance indicators proposed by the GRI, resulting in only 20 that were considered “comprehensive”. Stranieri et al. (2019) assessed 69 GRI reports from European food industries, but graded them from 0 to 4 for elements of the report other than just social, economic, and environmental indicators, and concluded that listed companies differ from privately held companies in that they need to comply with specific procedures related to the disclosure of financial, social, and environmental information in order to attract investors.

Ismail et al. (2021) analyzed reports for the years 2012 to 2017 from 90 companies listed on the Pakistani stock exchange (including 11 banks) and assigned quality scores based on the evaluation of the reporting principles recommended by the GRI. They concluded that the principles with the best ratings were Timeliness (73%) and Clarity (72%), while the worst were Balance (38%), Accuracy (24%), and Comparability (21%). They also found that most companies publicize positive or favorable events to improve their reputation and legitimacy, while hiding negative impacts and aspects.

Research focused on the sustainability reporting of financial institutions has also been produced. Sethi et al. (2017) analyzed 104 reports from the world’s largest financial institutions in 2012; 92% of them were from Western Europe, North America, or East Asia, resulting in an average score of 42.26 out of a possible 100. Kumar and Prakash (2019) analyzed 2015 and 2016 reports from 42 Indian banks and concluded that there is no significant difference in the disclosure of social and environmental indicators between public and private banks. Tsalis et al. (2020) developed a methodology to assess the level of compliance of corporate sustainability reporting practices within the scope of the 17 SDGs. They analyzed 48 reports from 20 Greek companies published between 2014 and 2016, and the authors found that, in general, the quality of the reports was very low and that the companies provided more information on the reduction and use of renewable energy (SDG 7), investments in infrastructure and innovation (SDG 9), and actions to combat climate change and reduce GHG emissions (SDG 13). Specifically, for the financial sector, they obtained a quality score of 103, which is considered “very low” quality, worse than the scores given to the aviation, energy, metallurgy, and telecommunications sectors.

Chagas et al. (2022) used the same methodology but focused on comparing the main Brazilian banks and only for the reports for 2020. The authors selected reports from 11 banks that were among the largest in assets and used the GRI standard, including private, public, and development banks, as well as national and foreign institutions, and ranked the reports, concluding that they have a low level of quality. They also concluded that the Brazilian banking sector produces higher-quality reports than the Greek banking sector since, in the analysis of their reports, the Brazilian average was higher. These researchers also identified a direct and positive relationship between the number of material contents reported and the DQPI score, which led to the conclusion that the amount of content disclosed can lead to an increase in the quality of reports. They also found that the SDG with the highest score was the one on decent work and economic growth (SDG 8), that the three banks that reported the least on SDGs scored the lowest, and that the highest-quality reports were from private banks.

These researchers propose, as an interesting avenue for further research, the use of different methodologies to assess sustainability reporting with reference to the same sample and contrast the results of their application (Chagas et al., 2022, p. 1984). This is what we propose to do in our study by applying two different methodologies to assess the quality of the sustainability reports of major banks operating in Brazil.

3. Research Design

3.1. Sample

This study follows up on Chagas et al.’s (2022) analysis and includes in the sample the same 11 banks examined by these authors. We analyzed the content of the reports of 11 banks that use the GRI Standards and are among the 20 largest banks operating in Brazil, by total assets, for the years 2021 and 2022. We added to these 11 banks the bank Inter, due to being a digital and native bank, totaling 12 banks and 25 documents, since, for bank Inter, we will also analyze the Report for the year 2020.

Table 1. Sample.

Firm	Ownership	Listed	Headquarters	Scope
Banco do Brasil (BB)	State	Yes	Brazil	Multiple
Banco do Nordeste (BNB)	State	Yes	Brazil	Development
BNDES	State	No	Brazil	Development
Bradesco	Private	Yes	Brazil	Multiple
BTG Pactual	Private	Yes	Brazil	Multiple
Caixa Econômica Federal (CEF)	State	No	Brazil	Multiple
Citibank	Private	Yes	USA	Multiple
Inter	Private	Yes	Brazil	Multiple
Itaú	Private	Yes	Brazil	Multiple
J.P. Morgan	Private	Yes	USA	Multiple
Santander	Private	Yes	Spain	Multiple
Sicredi	Private	No	Brazil	Multiple

provides information on the banks included in the sample. Of the 12 banks, 4 are state-owned and 8 are privately owned. Only three are not headquartered in Brazil (two of these have their headquarters in the USA; the other is from Spain). Two banks are development banks; all the others are multiple.

3.2. Data Collection

The analysis focused on a single document per year from each bank whose characteristics are presented in

Table 2. Report characteristics.

Bank	Year	Type	Pages	Material Topics
BB	2021	Integrated	225	11
	2022	Integrated	216	11
BNB	2021	Sustainability	101	9
	2022	Sustainability	94	9
BNDES	2021	Integrated	103	5
	2022	Integrated	157	5
Bradesco	2021	Integrated	299	12
	2022	Integrated	275	7
BTG Pactual	2021	Integrated	147	6
	2022	Integrated	161	6
CEF	2021	Integrated	305	13
	2022	Sustainability	150	13
Citibank	2021	ESG	183	22
	2022	ESG	108	22
Inter	2020	Integrated	106	10
	2021	Integrated	191	8
	2022	Integrated	113	8
Itaú	2021	ESG	178	10
	2022	ESG	168	10
J.P. Morgan	2021	ESG	78	18
	2022	ESG	74	23
Santander	2021	ESG	98	6
	2022	Integrated	64	11
Sicredi	2021	Sustainability	76	12
	2022	Sustainability	105	12

. All the reports were obtained from the banks’ websites. Forms, links, web pages, etc. referenced in the documents listed below were not accessed, as these were the only source of information for the results of this work. Of the 24 reports analyzed, 13 were integrated reports, 4 were sustainability reports, and 7 were ESG reports; 5 banks published integrated reports regarding the entire period examined, while 2 banks published sustainability reports and 2 published ESG reports. Regarding Santander, the 2021 report was an ESG report and the 2022 report was an integrated report. The first report from CEF was an integrated report, and the second was a sustainability report. At least in the case of the banks we examined, “ESG report” and “sustainability report” are two different names that banks used to refer to similar documents.

3.3. Reporting Quality Assessment

In the first part of the study, we followed the research suggestion of Chagas et al. (2022), who analyzed the 2020 reports, to carry out a time series evaluation to reflect the evolution over time of the quality of the non-financial reports of the main banks operating in Brazil.

The comparison and quality assessment of the reports analyzed in this study took place using the benchmarking scoring technique and consisted of two processes: Sustainability Disclosure Matrix (SDM) and the Measurement System (MS) (Tsalis et al., 2020). To prepare the SDM, we applied the linkage drawn up by Tsalis et al. (2020) of the contents of the Thematic Standards proposed by the GRI, which cover the economic (200 series), environmental (300 series), and social (400 series) impacts of corporate operations, with the 17 SDGs, as shown in

Table A1. SDG x GRI linkage based on Tsalis et al. (2020).

SDG	GRI	Total
1	201-1, 201-3, 202-1, 202-2, 203-2, 413-2	6
2	201-1, 203-1, 203-2, 206-1, 411-1, 413-2, 416-1, 416-2	8
3	203-2, 305-1, 305-2, 305-3, 305-6, 305-7, 306-1, 306-2, 306-3, 306-4, 401-2, 403-2, 403-3	13
4	205-2, 404-1, 404-2, 404-3, 410-1, 412-2	6
5	201-1, 202-1, 203-1, 401-1, 401-2, 401-3, 404-1, 404-3, 405-1, 405-2, 406-1, 414-1, 414-2	13
6	303-1, 303-2, 303-3, 304-1, 304-2, 304-3, 304-4, 306-1, 306-2, 306-3, 306-5	11
7	201-1, 203-1, 302-1, 302-2, 302-3, 302-4, 302-5	7
8	201-1, 201-3, 202-1, 202-2, 203-2, 204-1, 301-1, 301-2, 301-3, 302-1, 302-2, 302-3, 302-4, 302-5, 303-3, 401-1, 401-2, 401-3, 402-1, 403-1, 403-2, 403-3, 403-4, 404-1, 404-2, 404-3, 405-1, 405-2, 406-1, 407-1, 408-1, 409-1, 414-1, 414-2	34
9	201-1, 203-1	2
10	201-1, 202-1, 203-1, 203-2, 204-1, 205-1, 205-3, 401-1, 404-1, 404-3, 405-2, 406-1, 412-3	13
11	203-1, 413-1, 413-2	3
12	204-1, 301-1, 301-2, 301-3, 302-1, 302-2, 302-3, 302-4, 302-5, 303-3, 305-1, 305-2, 305-3, 305-4, 305-6, 305-7, 306-1, 306-2, 306-3, 306-4, 308-1, 308-2, 417-1	23
13	201-2, 302-1, 302-2, 302-3, 302-4, 302-5, 305-1, 305-2, 305-3, 305-4, 305-5, 305-6, 305-7	13
14	304-1, 304-2, 304-3, 304-4, 305-1, 305-2, 305-3, 305-4, 305-5, 305-7, 306-1, 306-3, 306-5	13
15	303-1, 303-2, 304-1, 304-2, 304-3, 304-4, 305-1, 305-2, 305-3, 305-4, 305-5, 305-7, 306-1, 306-2, 306-3, 306-5	16
16	205-1, 205-2, 205-3, 206-1, 307-1, 406-1, 408-1, 410-1, 411-1, 412-1, 412-2, 412-3, 414-1, 414-2, 415-1, 416-2, 417-1, 417-2, 417-3, 418-1, 419-1	21
17	201-1, 203-1, 203-2, 413-1, 413-2	5
Total		207

Source: Tsalis et al. (2020).

(Appendix A).

The next step was to use the Measurement System, which provides the means to objectively quantify the analysis of sustainability reports in relation to the quality of the information reported for each SDM content, using three indicators:

• Accountability Indicator (AI), with scores assigned as follows:
  0 points = no information for thematic content;
  1 point = there is qualitative information;
  2 points = there is quantitative information.
• Total Accountability Indicator (TAI), which is the sum of the IA scores for each SDG.
• Disclosure Quality Performance Indicator (DQPI), calculated by adding up the TAI score for each SDG.

The DQPI score ranges from 0 to 414. A report with a score from 307 to 414 is considered to be of high quality, from 207 to 306 medium quality, from 107 to 206 low quality, and from 0 to 106 very low quality.

In order to bring the study more up to date, the results are also presented with the link to the SDGs updated in 2022 by the GRI (Global Reporting Initiative, 2022b), as shown in

Table 3. SDG x GRI linkage based on Global Reporting Initiative (2022b).

SDG	GRI	Total
1	202-1, 203-2, 207-1, 207-2, 207-3, 207-4, 413-2	7
2	411-1, 413-2	2
3	203-2, 305-1, 305-2, 305-3, 305-6, 305-7, 306-1, 306-2, 306-3, 306-4, 306-5, 401-2, 403-6, 403-9, 403-10	15
4	404-1	1
5	202-1, 203-1, 401-1, 401-2, 401-3, 404-1, 404-3, 405-1, 405-2, 406-1, 408-1, 409-1, 414-1, 414-2	14
6	303-1, 303-2, 303-3, 303-4, 303-5, 304-1, 304-2, 304-3, 304-4, 306-1, 306-2, 306-3, 306-5	13
7	302-1, 302-2, 302-3, 302-4, 302-5	5
8	201-1, 202-1, 202-2, 203-2, 204-1, 301-1, 301-2, 301-3, 302-1, 302-2, 302-3, 302-4, 302-5, 306-2, 401-1, 401-2, 401-3, 402-1, 403-1, 403-2, 403-3, 403-4, 403-5, 403-7, 403-8, 403-9, 403-10, 404-1, 404-2, 404-3, 405-1, 405-2, 406-1, 407-1, 408-1, 409-1, 414-1, 414-2	38
9	201-1, 203-1	2
10	207-1, 207-2, 207-3, 207-4, 401-1, 404-1, 404-3, 405-2	8
11	203-1, 306-1, 306-2, 306-3, 306-4, 306-5	6
12	301-1, 301-2, 301-3, 302-1, 302-2, 302-3, 302-4, 302-5, 303-1, 305-1, 305-2, 305-3, 305-6, 305-7, 306-1, 306-2, 306-3, 306-4, 306-5, 417-1	20
13	201-2, 302-1, 302-2, 302-3, 302-4, 302-5, 305-1, 305-2, 305-3, 305-4, 305-5	11
14	304-1, 304-2, 304-3, 304-4, 305-1, 305-2, 305-3, 305-4, 305-5, 305-7	10
15	304-1, 304-2, 304-3, 304-4, 305-1, 305-2, 305-3, 305-4, 305-5, 305-7, 306-3, 306-5	12
16	205-1, 205-2, 205-3, 206-1, 307-1, 403-4, 403-9, 403-10, 408-1, 410-1, 414-1, 414-2, 415-1, 416-2, 417-2, 417-3, 418-1, 419-1	18
17	207-1, 207-2, 207-3, 207-4	4
Total		186

Source: our own elaboration based on Global Reporting Initiative (2022b).

which shows a total of 186 contents related to the 17 SDGs, unlike Tsalis et al. (2020), who presented 207.

The results are also presented without linking them to the SDGs, evaluating the dissemination of the content set out in the Thematic Standards according to the methodology developed by Tsalis et al. (2020).

In the second part of this work, the quality of the reports is assessed from the perspective of the GRI reporting principles, which must be observed to guarantee the quality and adequate presentation of the information reported, allowing information users to make informed assessments and decisions about the organization’s impacts and its contribution to sustainable development (Global Reporting Initiative, 2022a).

The evaluation is based on an analysis of the content of the reports and the assignment of scores for nine quality questions according to

Table 4. Quality assessment.

GRI Principle	Questions	Score
1. Balance	Q1.1: To what extent does the organization highlight both positive and negative events/impacts?	0 = No mention of negative events/impacts 1 = Negative events/impacts mentioned with less prominence and/or depth 2 = Balance between positive and negative impacts/events
	Q1.2: Does the organization report the sector’s main challenge related to sustainability, including examples and data?	0 = There is no mention of the main challenge of the sector 1 = There is mention, but without examples and data 2 = There is mention, examples, and data
2. Comparability	Q2.1: Does the organization adopt guidelines, standards or frameworks that allow information to be comparable to that of other organizations?	0 = Does not adopt any standard or framework 1 = Adopts one standard or framework 2 = Adopts two or more standards or frameworks
	Q2.2: Does the report compare information year on year?	0 = No comparison 1 = Comparison mostly with previous year 2 = Comparison mostly with two or more previous years
3. Accuracy	Q3.1: Are the measurement techniques and bases for calculating the information described?	0 = Not described 1 = Described in the least number of cases 2 = Described in the greatest number of cases
4. Timeliness	Q4.1: Are reports issued regularly and on time?	0 = No regularity or timeliness 1 = Regularity or timeliness 2 = Regularity and timeliness
5. Clarity	Q5.1: Are resources used to make the reports accessible and understandable?	0 = No resource utilization 1 = Minority resource utilization 2 = Majority resource utilization
6. Verifiability	Q6.1: Is the report externally assured?	0 = No external assurance 1 = Limited external assurance or restricted scope 2 = Reasonable external assurance and broad scope
	Q6.2: Is the organization’s contact information provided and responsive?	0 = No contact information 1 = Non-responsive information 2 = Responsive information

, prepared by the authors based on Ismail et al. (2021), the document GRI 1: Foundation 2021 (Global Reporting Initiative, 2022a), and the methodology presented in the publication of the results of Conselho Empresarial Brasileiro para o Desenvolvimento Sustentável (2024). Please note that the main reference for developing this methodology was Ismail et al. (2021), who used the distinction that GRI made between “quality” and “content” principles. Since the objective of this work was to assess quality, only the six principles were utilized, as in the reference work. Currently, GRI lists eight principles without separating them into “content” or “quality” (the two additional principles are completeness and sustainability context).

In addition to reading all the reports in the sample, highlighting the information needed to answer the questions relating to each principle, we carried out the procedures described below.

• Balance: “The organization shall report information in an unbiased way and provide a fair representation of the organization’s negative and positive impacts.” (Global Reporting Initiative, 2022a, p. 21)

To answer the first question (Q1), we carried out a search on Gemini, Google’s Artificial Intelligence (AI) platform, in which we submitted a question, for example, “What percentage of positive vs. negative mentions in Itaú bank’s 2021 sustainability report?”.

We also used the AI tool Machine Learning ESG Analyst (Malena) developed by ESG experts at the International Finance Corporation (IFC), linked to the World Bank, available at https://malena.ifc.org/#/ 9 June 2025. Malena analyzes documents uploaded by users and can read relevant ESG texts, such as sustainability reports, environmental and social impact assessments, and news articles.

To answer the second question (Q2), we analyzed what banks disclose about financing/investments granted to companies that emit large amounts of carbon (carbon majors), since the role of financial intermediation highlights the importance of banks in terms of the emissions they finance.

• Comparability: “The organization shall select, compile, and report information consistently to enable an analysis of changes in the organization’s impacts over time and an analysis of these impacts relative to those of other organizations.” (Global Reporting Initiative, 2022a, p. 21)

According to the GRI (Global Reporting Initiative, 2022a), information reported in a comparable way allows for the assessment of the organization’s current impacts in relation to its past impacts and its objectives and targets. It also allows external parties to assess and compare the organization’s impacts with those of other organizations.

To answer (Q1), we assessed how many frameworks/guidelines/standards were adopted by the organizations when preparing their reports. The diverse number of reports makes comparisons between companies extremely challenging and, for this reason, it is important to use the available standards so that readers can compare their information with that of their competitors, for example.

For Q2, we analyzed how many previous years the information provided by the banks was compared with, giving maximum points to those that present information from at least two previous periods, as indicated by the GRI (Global Reporting Initiative, 2022a).

• Accuracy: “The organization shall report information that is correct and sufficiently detailed to allow an assessment of the organization’s impacts.” (Global Reporting Initiative, 2022a, p. 20)

To answer Q1, we mainly analyzed the indicator data in graphs and tables to see whether they contain explanations of the ways in which they were obtained and the bases for calculation.

• Timeliness: “The organization shall report information on a regular schedule and make it available in time for information users to make decisions.” (Global Reporting Initiative, 2022a, p. 23)

To answer Q1, since the GRI does not set a maximum deadline for disclosure but does indicate that it should take place in time for readers to make decisions about their relationship with the organization, we used data from the “Reporting Matters Brasil” report (Conselho Empresarial Brasileiro para o Desenvolvimento Sustentável, 2024), which shows that most organizations publish their reports in the first half of the following year.

In order to obtain a maximum score in this category, banks had to publish their reports in the first half of the following year and maintain regularity of publication, i.e., the report must have been published in the same month or up to one month before or after the previous year’s report.

When it was not possible to identify the date of publication of the report in the document itself or on the website of the institutions, we considered the lowest between the date of creation of the report in the PDF version and that of the assurance letter.

• Clarity: “The organization shall present information in a way that is accessible and understandable.” (Global Reporting Initiative, 2022a, p. 21)

To answer Q1, we followed the GRI guidelines (Global Reporting Initiative, 2022a) and checked whether the reports used visual resources such as graphs, tables, and images to make the information accessible and understandable. In addition, we also assessed the existence of lists of acronyms and abbreviations, a glossary, a table of contents, and an index to enable navigation with less effort and greater ease of understanding.

Considering that 84% of respondents to a study carried out by the Central Bank of Brazil have a bank account (Banco Central do Brasil, 2022), but only 5% of the Brazilian population is fluent in English (British Council, 2014), we believe that the report of a bank with a strong presence in the country should have a Portuguese version. Therefore, even though they present the other items in the reports, reports without a Portuguese version were given a maximum score of 1. With regard to file size, we considered a size of up to 15.5 Megabytes (Mb) to be adequate, which was the average size of the reports in the sample. We consider this item to be relevant in the evaluation since the majority of the Internet user population (58%) declared that they only use their cell phone to access it and mobile access is characterized by important limitations, since 60% of people who own a cell phone use a prepaid plan, which can mean a limited data package, which can be insufficient for the activities carried out in a month (Núcleo de Informação e Coordenação do Ponto BR, 2024).

• Verifiability: “The organization shall gather, record, compile, and analyze information in such a way that the information can be examined to establish its quality.” (Global Reporting Initiative, 2022a, p. 23)

To answer the first question Q1, we considered the importance that the GRI attributes to external assurance, listing it as a factor in increasing the credibility of reports and considering that it is used to rank companies (Acionista, 2021; Global Reporting Initiative, 2022a). We, therefore, checked the assurance letters in the reports, which are generally signed by the Big Four, to see what the scope and breadth of the checks carried out were.

For Q2, we used the criterion adopted by Ismail et al. (2021) that it is important to have contact information for the organization in the report so that readers can clarify any doubts about the data contained therein. We assessed the existence of contact information and its responsiveness, i.e., whether, after asking the questions below and sending them to the email address provided in the reports, the organization responded by answering the questions raised.

Is feedback given to stakeholders to find out what they think of the Reports after they have been published? If so, which stakeholders are heard and what methodology is applied? What was the outcome of the SR/RI for 2021, 2022, and 2023? Has the result been published and/or can it be made available?

4. Results1

4.1. Quality Analysis

In

Table 5. Quality analysis according to the DQPI.

Banks	DQPI Tsalis et al. (2020)			DQPI GRI
	DQPI 2020	DQPI 2021	DQPI 2022	DQPI 2021	DQPI 2022
BB	165	177	154	152	134
BNB	119	133	202	109	166
BNDES	121	166	218	156	204
Bradesco	241	245	257	210	248
BTG Pactual	93	182	206	155	191
CEF	70	53	85	54	73
Citibank	207	180	164	148	141
Inter	213	194	170	150	134
Itaú	184	168	187	169	194
J.P. Morgan	145	66	65	46	47
Santander	136	145	120	149	102
Sicredi	142	148	154	108	106
Average	153	154.8	165.2	133.8	145

, we present the main results concerning the quality of the reports according to the DQPI for each bank and for each year of analysis. The data presented in this table are the basis for some of the analyses that follow. The institutions with the best overall performance concerning the DQPI were Bradesco and BNDES and the worst were CEF and J. P. Morgan.

4.1.1. Quality Analysis According to Tsalis et al. (2020) Link

In

Table 6. Quality analysis according to Tsalis et al. (2020) link.

Years	Average DQPI	Quality	Average Amount of Content	SDG with Greatest Disclosure	No. of Banks Communicating All SDG
2020	153	2 very low; 3 average; 7 low	31.9	SDG 8 for all banks	4
2021	154.8	2 very low; 1 average; 9 low	33.3	SDG 8 for 11 banks; SDG 5 and 8 for one bank	6
2022	165.2	2 very low; 4 average; 6 low	36.2	SDG 8 for all banks	7

, we present the main results of our own analysis for the years 2021 and 2022, and also the results obtained by Chagas et al. (2022) regarding 2020, to which we added the results pertaining to bank Inter.

Chagas et al. (2022) found that, for the 2020 reports analyzed, the average number of reported contents was 31, while the average DQPI was 147.5. With the inclusion of the 2020 bank Inter report, we arrived at an average of 31.9 reported topics and 153 DQPI.

Comparing the average scores and the average amount of content, we can see that the quality of the reports evolved in 2021 and 2022, although most banks present “low” quality. The result corroborates the perception of Chagas et al. (2022) that no report was of “high” quality and, therefore, that banks still need to advance in the quality of the disclosure of their sustainable practices.

SDG 8 was the one that obtained the highest score in all the reports, which is also in line with the findings of Chagas et al. (2022) and confirms other studies that highlight the prevalence of this SDG in non-financial reports (Avrampou et al., 2019; KPMG, 2020). The number of banks communicating all SDGs has risen from four in 2020 to seven in 2022.

Table 7. Average DQPI comparison.

Characteristics	Average DQPI 2021	Average DQPI 2022
Private	166.0	165.4
State	132.3	164.8
Listed	165.6	169.4
Not listed	122.3	152.3
Headquarters in Brazil	162.9	181.4
Headquarters abroad	130.3	116.3
Development	149.5	210.0
Multiple	155.8	156.2
Integrated report	169.5	187.5
Sustainability/ESG report	140.0	142.8

presents the average DQPI results according to control, capital, headquarters, activity, and type of report.

The results above show that the highest-quality reports are those produced by private, publicly traded banks with headquarters in Brazil and in the “Integrated” category. Even though, in 2021, the difference was considerable, in 2022, the state-owned banks came closer to the score of the private ones. Although they had a lower average DQPI in 2021, development banks made a leap in quality and, in 2022, had a considerably higher average than multiple banks.

4.1.2. Quality Analysis According to GRI Link

In May 2022, the GRI updated the document “Linking the SDGs and the GRI Standards” (Global Reporting Initiative, 2022b), bringing its vision of the correlation between the contents and the SDGs, the result of which we reproduce in

Table 8. Quality analysis according to GRI link.

Years	Average DQPI	Quality	Average Amount of Content	SDG with Greatest Disclosure	No. of Banks Communicating All SDG
2021	133.8	2 very low; 9 low; 1 average	33.3	SDG 8 for all banks	0
2022	145.0	2 very low; 6 low; 4 average	36.2	SDG 8 for all banks	1

. We, therefore, applied the same methodology to the 2021 and 2022 reports to compare the results.

The DQPI score, in this case, ranged from 0 to 372, with a report with a score from 275 to 372 considered to be of high quality, from 185 to 274 of medium quality, from 95 to 184 of low quality, and from 0 to 94 of very low quality.

Comparing the DQPI averages, we can see that there has been an improvement in the quality of the 2022 reports, although all the average scores indicate “low” report quality. The result reinforces that banks must improve the quality of their non-financial reports.

In the two years analyzed, Bradesco obtained the best DQPI score and the quality of its reports was “average”. At the other extreme, J.P. Morgan and Caixa’s reports were rated “very low”.

In terms of evolution, BTG Pactual, BNDES, and Itaú stand out, having obtained “low” quality with DQPI 155, 156, and 169 points, respectively, in 2021, rising to “medium” quality with DQPI 191, 204, and 194.

SDG 8—decent work and economic growth—was, again, the highest scoring of all the reports, but the one with the worst score was SDG 2—ending hunger—unlike Chagas et al. (2022), who identified SDG 11—sustainable cities and communities—as the worst performer.

Table 9. Average DQPI by group and year.

Characteristics	Average DQPI 2021	Average DQPI 2022
Private	141.9	145.4
State	117.8	144.3
Listed	143.1	150.8
Not listed	106.0	127.7
Headquarters in Brazil	140.3	161.1
Headquarters abroad	114.3	96.7
Development	132.5	185.0
Multiple	134.1	137.0
Integrated report	146.2	168.8
Sustainability/ESG report	121.5	121.2

presents a comparison of the average DQPI results according to control, capital, headquarters, activity, and type of report.

The results above reaffirm that the highest-quality reports are those produced by private, publicly traded banks headquartered in Brazil and in the “Integrated” modality. It can also be seen here that, although, in 2021, the public banks scored well below the private banks, in 2022, they scored very close to the private banks. Once again, the average DQPI for development banks was slightly lower in 2021, but made a qualitative leap in 2022, coming in well above the average for the other banks.

4.1.3. Quality Analysis Without Links to SDGs

Table 10. Average AI comparison.

Characteristics	Average AI 2021	Average AI 2022
Private	60.3	61.0
State	52.3	61.0
Listed	60.4	61.8
Not listed	49.0	58.7
Headquarters in Brazil	61.7	68,4
Headquarters abroad	45.3	38.7
Development	56.5	71.5
Multiple	57.8	58.9
Integrated report	65.2	71.2
Sustainability/ESG report	50.0	50.8

shows the average results of the AIs according to control, capital, headquarters, operations, and type of reporting.

The results confirm that the highest-quality reports are those produced by publicly traded banks headquartered in Brazil and in the “Integrated” category. It can also be seen that public banks obtained the same score as private banks in 2022, although, in the sum of the two years analyzed, they still lag behind. The average AI for development banks was slightly lower in 2021, but they were well above the average for multiple banks.

4.2. Evaluation of Reporting Principles

In order to present the consolidated result, we divided the results of the principles that had two linked questions, so that they had the same weight as the others. In this way, the total possible score was 12 points, as shown in

Table 11. Reporting principles analysis.

Bank	Year	Balance	Comparability	Accuracy	Timeliness	Clarity	Verifiability	Total
BB	2021	1	2	2	2	2	1.5	10.5
	2022	1	2	2	2	2	1.5	10.5
BNB	2021	1	1.5	2	1	1	1	7.5
	2022	1	1.5	2	0	1	1	6.5
BNDES	2021	1	1.5	2	2	2	1.5	10
	2022	1	2	2	2	2	1.5	10.5
Bradesco	2021	1.5	2	2	1	2	1.5	10
	2022	1.5	2	2	2	2	1.5	11
BTG Pactual	2021	1	1	2	2	2	1	9
	2022	1	2	2	2	2	1	10
CEF	2021	0.5	1.5	2	2	1	1	8
	2022	0.5	1.5	2	0	2	1.5	7.5
Citibank	2021	1.5	1	2	2	1	1	8.5
	2022	1.5	1.5	2	2	1	1	9
Inter	2021	0.5	1	2	2	2	1	8.5
	2022	0.5	1.5	2	2	2	1	9
Itaú	2021	1.5	2	2	2	2	1.5	11
	2022	1.5	2	2	2	2	1.5	11
J.P. Morgan	2021	1	1	2	2	1	0.5	7.5
	2022	1	1	2	2	1	0.5	7.5
Santander	2021	1	2	2	1	2	1	9
	2022	1.5	2	2	2	2	1	10.5
Sicredi	2021	0.5	2	2	2	2	1.5	10
	2022	1	2	2	2	2	1.5	10.5
Average		1.04 (52%)	1.65 (83%)	2 (100%)	1.71 (86%)	1.71 (86%)	1.19 (60%)

.

The institutions with the best overall performance concerning the principles were Itaú, BB, and Bradesco and the worst were CEF, BNB, and J.P. Morgan.

Comparing the performance of 2021 with 2022, there was an increase in the performance of BNDES, Bradesco, BTG, Citibank, Inter, Santander, and Sicredi, a decrease in BNB and Caixa, and a maintenance in Banco do Brasil, Itaú, and J.P. Morgan. In the final tally, the institutions’ reports advanced in terms of adherence to the principles since, in 2021, they totaled 109.5 compared to 113.5 in 2022.

The worst results were for the principles of Balance (52%) and Verifiability (60%), while the best were Accuracy (100%), Clarity (86%), and Timeliness (86%). The results are partially in line with the findings of Ismail et al. (2021), who also highlighted Timeliness and Clarity as the principles with the best scores and Balance as one of the worst. However, there is some distance between the results for the Accuracy principle, which, in the current study, scored the highest and, in the previous one, was one of the worst.

The results corroborate the perception of 98% of Brazilian investors who believe that corporate reports contain greenwashing (PricewaterhouseCoopers, 2024), which would be mitigated if the reports were more balanced and had external verification of greater scope and assurance.

Table 12. Average reporting principles score by group.

Characteristics	2021	2022
Private	9.2	9.8
State	9.0	8.8
Listed	9.1	9.4
Not listed	9.3	9.5
Headquarters in Brazil	9.4	9.6
Headquarters abroad	8.3	9.0
Development	8.8	8.5
Multiple	9.2	9.7
Integrated report	9.3	10.3
Sustainability/ESG report	8.9	8.7

presents a comparison of the average reporting principles score according to control, capital, headquarters, activity, and type of report.

The results indicate that the highest-quality reports in terms of adherence to the GRI principles are those drawn up by private banks headquartered in Brazil and in the “Integrated” category, in line with the other results of this study. This time, the unlisted banks were slightly ahead of the listed ones, driven by the good scores received by BNDES and Sicredi, while the development banks were below the multiples.

4.3. Overall Discussion

In general, the banks presented poor quality reports, similar to the findings of Chagas et al. (2022), who analyzed the 2020 reports of the same banks (except for the bank Inter), and Avrampou et al. (2019), who evaluated the reports of five large European banks and for whom the disclosure of the SDGs was considerably low, with an average score of 0.78 on a scale from 0 to 4. They are, however, a little better than the conclusions of Tsalis et al. (2020), who showed the reports of the Greek financial sector with an average DQPI of 103 points, i.e., very low quality.

The SDGs with the highest scores were those related to decent work and economic growth (SDG 8) and gender equality (SDG 5), and those with the worst scores were SDG 2—ending hunger, SDG 9—innovation and infrastructure, SDG 11—sustainable cities and communities, and SDG 17—partnerships for development. This is in line with Chagas et al. (2022), who identified SDG 8 as the highest scoring and SDG 11 as the worst.

The contents with the best results (ratio of points possible to points achieved) were economic with 37%, followed by social with 36% and environmental with 29%.

Using the linkage with the SDGs proposed by Tsalis et al. (2020), the reports considered to be of medium quality were those of Bradesco (2021 and 2022) and BNDES (2022), a result different from that of Chagas et al. (2022), who found this to be Bradesco and Citibank. For the very-low-quality reports, Chagas et al. (2022) identified Caixa and BTG Pactual, while the present study had Caixa and J.P. Morgan in both years analyzed.

As for the principles, the overall average was 9.3 points, equivalent to 77.5% of the maximum possible score (12 points), demonstrating opportunities for improvement in the preparation of the reports. In agreement with the findings of Ismail et al. (2021), most companies disclose clear easy-to-find information and contact details so that stakeholders can verify the information, but lack balanced information on positive vs. negative events, highlighting only favorable data to preserve or improve their image and reputation.

Q1: Are there any differences in the quality of the reports considering control (public vs. private), capital (public vs. private), headquarters (Brazilian vs. foreign), or performance (multiple vs. development)?

Yes. The highest-quality reports were those produced by privately controlled publicly traded banks headquartered in Brazil, although in the evaluation of the GRI principles, privately held banks obtained a slightly higher average score. With regard to performance, in the 2021 reports and in the evaluations of the GRI principles, multiple banks were better evaluated; but, in 2022, development banks obtained better average scores. With regard to the analysis of the economic, environmental, and social content disclosed, the privately controlled banks performed better than the public ones, although the difference narrowed in 2022, and there was even a tie in the assessment of content not linked to the SDGs.

The conclusion is different from that obtained by Kumar and Prakash (2019), who, based on the analysis of reports from 42 Indian banks, concluded that there is no significant difference between public and private banks, but is similar to that of Chagas et al. (2022), for whom the set of private banks had an average of 164 points while the public ones had 119.

The banks headquartered in Brazil obtained a better average score, which was the largest difference among the characteristics evaluated, a result different from that found by Chagas et al. (2022), who found a higher average score for banks with foreign control than those headquartered in Brazil.

Q2: Are there any differences in quality depending on the type of report (integrated or sustainability/ESG)?

Yes. The highest-quality reports were those that used the integrated model. Most of the banks used integrated reporting (52%), a figure similar to that observed by Chagas et al. (2022), which was 56%. Of the four banks with the highest DQPI scores and content without a connection to the SDGs, three (75%) were reported via integrated reporting. Regarding the evaluation of the GRI principles, of the five that scored 10 in both years evaluated, three (60%) used the integrated model.

With regard to the DQPI quality indicator, we found that the averages in both 2021 and 2022 point to better results for reports using the “Integrated” type. In 2021, the average for “Integrated” reports was 157.9 points compared to 130.8 for the “Sustainability/ESG” type. In 2022, the difference increased to an average DQPI of 178.2 for “Integrated” reports and 132.0 for “Sustainability/ESG”. Also, in the evaluation of the content reported without relating it to the SDGs, we found that the banks that used the integrated model scored better than the others.

In the qualitative dimension related to the GRI principles, of the eight reports that stood out with the highest scores (10.5 or 11), five are of the “Integrated” type and three “Sustainability/ESG”.

The findings corroborate a previous study which stated that the adoption of integrated reporting generated a steady increase in the quantity and quality of sustainability disclosures (Montecalvo et al., 2018).

Of the six with the lowest scores (from 6.5 to 8), five are “Sustainability/ESG” and one “Integrated”. The average score for integrated reports was 9.8 and sustainability/ESG was 8.8, making it possible to conclude that integrated reports comply better with the GRI principles.

Q3: Has the quality of the reports improved?

Yes. Regarding the evaluations of the disclosure of economic, environmental, and social content, whether correlated with the SDGs or not, it was found that, in all scenarios, there was progress. Using the Tsalis et al. (2020) link, the average DQPI went from 147.5 in 2020 to 154.8 in 2021 and 165.2 in 2022. According to the GRI definition, it went from 133.8 in 2021 to 145 in 2022 and, according to the AI assessment, it went from 57.6 to 61 points.

In the qualitative dimension related to the GRI principles, the average score in 2021 was 9.1 and in 2022 9.5 out of a maximum of 12 points, which is equivalent to a 3.4% increase. The principle with the biggest increase was Comparability with a +10.5% average score, and the only principle with a decrease was Timeliness with −4%.

Q4: Which banks stood out positively and negatively?

The banks with the best reports were Bradesco, BNDES, and Itaú, while the worst performers were J.P. Morgan and CEF. Bradesco ranked first in the DQPI quality indicator for both years, both in the content x SDG correlation of Tsalis et al. (2020) (Table A1) and in the one defined by the GRI (Table 5), always with average quality. In the assessment of compliance with the GRI principles, it came second in the sum of the years, only behind Itaú. Itaú is the best bank in terms of adherence to the GRI principles and obtained average quality for the 2022 report in the content x SDG correlation defined by the GRI (Table 5).

J.P. Morgan had very low quality in all assessments related to the SDGs, occupying the worst position in the DQPI in 2021 in the Tsalis et al. (2020) linkage (Table A1) and in 2022 in both forms of linkage (Table 5 and Table A1). It achieved the penultimate place in relation to adherence to the GRI principles (Table 11). CEF also had very low quality in all the evaluations related to the SDGs (Table 5 and Table A1), having also ranked third-worst in the analysis of adherence to the GRI principles (Table 11).

5. Conclusions

The analyses carried out on 25 reports from 12 large banks operating in Brazil lead to the conclusion that, compared to state-owned banks, private banks produce better quality sustainability reports in the period 2021–2022. The same can be said regarding listed banks compared to their non-listed counterparts as well as concerning banks with head offices in Brazil compared to their foreign counterparts. Regarding integrated reports versus autonomous (ESG/sustainability) reports, the banks whose reports used the “Integrated” model seem to be better-quality sustainability reports in 2021 and 2022. It is also possible to conclude that the quality of the reports has evolved, considering both the economic, environmental, and social content and its correlation with the SDGs, as well as adherence to the GRI reporting principles.

This study contributes to the discussion on the preparation of sustainability reports by banks, helping them in their approaches to making the reports an effective communication tool that improves sustainability, including the possibility of making comparisons with their competitors (Tsalis et al., 2020). It also contributes to possible improvement in the application of GRI 1 principles in the sense that it allows us to identify those to which more attention should be given and how to improve their application. The finding that Balance and Verifiability are the principles to which banks seem to pay less attention supports the perception of 98% of Brazilian investors who believe that corporate reports contain greenwashing (PricewaterhouseCoopers, 2024). This could be mitigated if the reports were more balanced, focused on transparency rather than marketing, and included more comprehensive external verification and assurance. Banks’ managers would be well advised to incorporate this finding into their decision making concerning sustainability reporting.

Another contribution, in line with the conclusion of Chagas et al. (2022), is the direct and positive relationship between the number of contents reported and the Disclosure Quality Performance Indicator (DQPI), i.e., the more contents, the higher the DQPI. This study responds to Chagas et al.’s (2022) call for research applying different methodologies to assess sustainability reporting of the same set of firms and contrasting the results obtained. Some limitations identified during the research were the fact that only the 2021 and 2022 reports were analyzed from a specific economic sector and from a sample of 12 organizations. Future research could explore a longer period and compare different sectors of the Brazilian economy, as well as using other frameworks or standards in addition to the GRI. Although all the banks in the sample use sustainability reports to make their actions transparent, it is true that they also use other resources such as forms, websites, repositories, etc., and, therefore, using only the reports as a source of information was also a limiting factor. Another limitation encountered was the fact that, during the course of the investigation, the banks released new versions of their reports without this being explicitly stated in the document produced, and so the analyses carried out may have been in versions other than those currently available. Lastly, the GRI summaries did not always indicate the topic dealt with in that content in isolation, but rather aggregated several others, making it difficult to separate the subjects in order to later assign a score. Other issues that could generate future research would be to explore the relationship between the number of frameworks used and comparability between companies and to identify the methodologies applied in the selection and engagement of stakeholders to define material topics and how this influences the result, as well as analyzing the quality of the reports based on feedback from these same stakeholders. One can also look for explanations for the better quality of reports from publicly traded and private banks. Pressure from shareholders for transparency? The fact that they have two regulators, the banking regulator and the capital market regulator? Could it be related to management continuity, since the top level of public institutions is changed at least every time the government changes? Could it be related to the ease with which private banks can hire people or consultancy firms on the market and/or prepare reports with more expertise in contrast to the need for public tenders and bids by public institutions?