Non-Financial Disclosure and Intra-Industry Comparability: A Macro, Meso and Micro Analysis

Abstract

The inadequacy of financial reports for meeting the information needs of stakeholders has prompted companies to adopt non-financial communication systems (also called non-financial disclosure—NFD). Comparability of NFD is an issue as it allows making sense of the information. Nevertheless, while some argue that comparability is particularly difficult to pursue in practice on a general level, it can be achieved among companies belonging to the same industry. This study aims to understand whether, at the empirical level, the comparability of NFD is achieved and to what extent (macro, meso or micro). To achieve this aim, a text analysis of the NFD was performed. The object of analysis is represented by the NFD published by the listed companies belonging to the energy and banking industries, and that is part of the Dow Jones sustainability index. The main results are the following. First, there is a de facto comparability in terms of adopted standards, but not in terms of how the standards are applied. Second, the phenomenon of label creativity represents a relevant barrier. Third, although content standardization is lacking, common information dimensions in the reports seem to emerge. This appears to be an invitation to policymakers to transpose virtuous behavior and to implement the desired harmonization of jure.

1. Introduction

Literature and practice have highlighted a growing need for companies to legitimize their activities and respond more appropriately to the growing information needs of stakeholders [1,2,3]. The inadequacy of the traditional financial reporting tools to meet these needs has pushed companies towards the disclosure, on a compulsory or voluntary basis, of non-financial information [4,5,6]. Due to this, several national and international organizations have proposed different frameworks useful to support companies in the disclosure of non-financial information (e.g., social, environmental, governance, etc.) and in the implementation of non-financial disclosure (NFD) practice. These frameworks differ from one another in purposes, focus (on all ESG dimensions or only on some of them), approach, content, structure, dimensions considered, etc. [7,8,9]. Despite this, all these frameworks have at least one point in common: they all mention comparability as a fundamental principle of NFD.

Comparability of NFD is generally considered as a relevant issue not only among the producers of NFD but also among the users, i.e., all the stakeholders [6,10,11,12,13], as it allows evaluating the overall organizational performance and being “able to make peer-to-peer comparisons” [12] (p. 71). In addition, recent surveys and empirical analysis have highlighted the existence of a public interest in having comparable NFD [11] (p. 127) [14,15] (p. 8) [16] (p. 4). Moreover, the relevance of this principle is also witnessed by the existence of several government initiatives that aim to improve the comparability of NFD. This kind of initiatives have taken place in Europe [17,18], USA [19,20], Australia [21] and Asia [22]. The IFRS published, in September 2020, the “Consultation Paper on Sustainability Reporting” [23] to define a single NFD framework and, therefore, promoting the standardization of non-financial information. Even if the comparability is a desirable principle for standard setters, academics, practitioners, and regulatory bodies as it allows to have efficient financial markets, various empirical analyses that have investigated the practices adopted by companies, indicate that comparability is, in general, not concretely achieved [6,7,24,25,26,27].

The main reasons for the lack of inter-firm comparability are mainly related to the presence of a high number of frameworks [28], the different degree of sensitivity to the social and environmental issues of companies [29,30,31], the wide margins of discretion granted to the preparers [32,33] as well as “greenwashing” practices and opportunistic behaviors [34,35,36,37,38,39,40,41]. It must be pointed out that while comparability between firms belonging to different industries appears difficult to achieve, some argue that it could be achieved between companies operating in the same field. This can happen because the exogenous contextual variables (normative, cultural, macro-economic, etc.) tend to impact less on the NFD processes, and it seems reasonable to expect emulation behaviors between companies and convergence towards common practices as well as practices to prove to be the “best in class” [42,43]. In addition, some scholars have highlighted that mandatory NDF “increases the comparability of reports among firms as well as over time” [44] (p. 3473) and “promotes better comparability but may also help prevent some companies from “greenwashing” or deliberately manipulating stakeholders’ perceptions through discretionary reporting” [45] (p. 20). Other scholars [27,46] argue that sustainability ratings act to deter greenwashing and that they also stimulate the adoption of comparable non-financial information.

Moving from the abovementioned reflections—and also considering the call for research on “the differences between de jure (rules) and de facto (practices), as understanding their interactions can contribute to understanding whether “good quality reporting needs, good quality regulation or whether good reporting can develop without good regulation” [47] (p. 150)” [7] (p. 603)—this paper aims to investigate inter-firm comparability of NFD at a macro (framework), meso (structure) and/or micro (indicators) level. The analysis does not only aim to demonstrate the existence of the noncomparability issue, as done by previous studies but also (and mainly) to understand whether comparability can be identified at a macro, meso and/or micro level and the detailed differences between the examined NFD.

To achieve this aim, a manual text analysis [48,49] of NFD will be performed. The sampled companies are included in the Dow Jones sustainability index (DJSI) that tracks the stock performance of the world′s leading companies in terms of economic, environmental and social criteria. In particular, the analysis is focused on the companies operating in the banking and energy industries.

In comparison to the extant studies, this one does not focus only on indicators among a specific NFD framework [26] but offers a multilevel analysis (macro, meso, micro). Furthermore, this study tries to propose a comparison between comparability at a conceptual level (i.e., what is declared by the standard setters or what NFD comparability should be) [7,50] and what happens in practice (i.e., what companies do). Moreover, this is not based on quantitative analyses [6] but on in-depth textual analysis to identify processes of convergence at an analytical content [51,52]. In addition, the study tries to investigate comparability at an intraindustry level and not exclusively at an interindustry one [6,53]. Finally, this study considers a very specific sample of companies, i.e., the ones listed in the DJSI operating in the energy and banking industries.

The structure of the paper is the following. First, a literature review that systematizes the main contributions on the NFD comparability issue is presented. Then, the design of the study and the results of the research are provided. In the final sections, the authors attempt to make sense out of the case findings, present the theoretical arguments of the study and lastly, propose future research opportunities.

2. Literature Review

In recent years, an increasing number of scholars and practitioners have discussed the social and managerial implications related to the NFD on a mandatory or voluntary basis.

This growing interest can be due to several factors such as the decreasing relevance of financial reports and the consequent need and call for new accounting practices useful for disclosing the organizational value creation process and/or the value realized by a company more comprehensively, the worldwide introduction of mandatory reporting practices such as King Code III in South Africa and the Directive 95/2014/EU in Europe as well as the pressure of corporate governance bodies and consulting and auditing companies [6,34,54,55,56,57].

To support companies in the production of NFD, several national and international organizations have proposed different frameworks, none of which can be considered as generally accepted [6,7,58]. These frameworks differ from one another in several aspects, among which it is possible to mention the following [7,8,9]. First, some frameworks provide detailed guidelines and standards (close to a rule-based approach) (e.g., GRI, SASB, EFFAS, etc.), whereas others tend to adopt a principle-based approach, which leaves more freedom to the NFD preparers (e.g., IIRC, AA1000, etc.). Second, some frameworks are oriented towards a range of stakeholder perspectives (e.g., GRI, AA1000 and ISO) while others mainly focus on financial stakeholders (e.g., IIRC, SASB and EFFAS). Third, some tend to be centered on the business model and the value creation process (e.g., IIRC) while others on the value realized and how it is distributed among different stakeholders (e.g., GRI, etc.). Forth, some frameworks aim to disclose the different ESG dimensions (e.g., IIRC, GRI, etc.) while others only some of them, mainly the environmental ones (e.g., TCFD, UN Global Compact, GHG, CDP, Eco-Management Audit Scheme and ISO14000, etc.). To give an idea of the magnitude of this “new institutional global infrastructure” [7] (p. 606), Brown et al. [28] perform an analysis of the modes that companies can adopt for preparing NFD and identify 30 international frameworks. Similarly, Fiandrino recalls the existence of about 255 standards, codes of conduct and audit protocols worldwide to address sustainability information [59] (p. 47). In all, a plethora of NFD frameworks can be identified.

To help organizations in the selection of an NFD framework, the Federation of European Accountants [60] has reviewed the different models and officially recognized them as NFD eligible frameworks emanating from Global Reporting Initiative (GRI), International Integrated Reporting Council (IIRC), Sustainability Accounting Standards Board (SASB), AccountAbility1000 (AA1000), United Nations Global Compact (UNGC), Organization for Economic Cooperation and Development (OECD), European Federation of Financial Analysts Societies (EFFAS) and International Standards Organization (specifically ISO26000). Moreover, to this, also models for reporting specific environmental or social aspects should be mentioned: Carbon Disclosure Project, Greenhouse Gas Protocol Corporate Standard, tripartite declaration of principles concerning multinational enterprises and social policy of the International Labor Organization, ISO9000 and SA8000. Among all these frameworks, the GRI Guidelines and IIRC Framework are the most widely used [5,56,61,62]. Even if, as mentioned, the different standards present several differences, they all share a few aspects. One of these is comparability.

While in the financial reporting domain, the issue of comparability has been deeply investigated and has now reached a maturity level of conceptualization and application [63,64,65,66], in the NFD, it is an argument still debated. Even if with different nuances, comparability tends to be defined by the different standards as a reporting quality that allows performing temporal and/or inter-firm comparisons. In this study, the focus is on inter-firm comparability. All the standards consider inter-firm comparability as a fundamental principle of NFD, a “promise” that all the standards formulate to their adopters and supporters. In this respect, the IIRC, for example, envisages among its basic principles of information comparability “The framework... enabling a sufficient degree of comparability across organizations to meet relevant information needs.” (IIRC Framework, 2013, p. 7). Similarly, the GRI states, “The GRI Standards create a common language for organizations and stakeholders, with which the economic, environmental, and social impacts of organizations can be communicated and understood. The Standards are designed to enhance the global comparability and quality of information on these impacts, thereby enabling greater transparency and accountability of organizations.” (GRI 2016: GRI 101). Regard this [67] (p. 268) note that “GRI claims to provide the basis of worldwide standardized, comparable, reporting on the sustainability of (particularly business) organizations.”. Finally, several studies are based on the “grand theory” that the NFD should be comparable from both a temporal and spatial perspective [6,42,53] In fact, some scholars have recognized that mandatory NDF “increases the comparability of reports among firms as well as over time” [44] (p. 3473).

The relevance of this principle is also witnessed by several facts.

For example, the EU Directive 2014/95/EU calls explicitly for inter-firm NFD comparability as it aims to “enhance the consistency and comparability of non-financial information disclosed” [33,68]. Moreover, the European Commission (EU) has recently launched an international consultation regarding NFD: the majority of respondents (71%) believe that the non-financial information reported by companies is deficient in terms of comparability [17]. Drawing upon these pieces of evidence, the EU is working on “future standards on non-financial reporting” and declared the intention to “require some or all companies under the scope of the legislation to use a non-financial reporting standard” [18]. A few years before, the Sustainability Accounting Standards Board (SASB), concerning the U.S. NFD regulation, has highlighted the request of investors for more comparability of sustainability information [19]. The U.S. Council of Institutional Investors, more recently, calls “clearer and more comparable information about key ESG” and thinks that “adoption by investors and issuers of common ESG disclosure standards would be a highly significant market improvement” [20]. In Australia, the Minister recommended consideration of an “agreed reporting framework (...) to allow for greater comparability” [21]. In addition, in the Asian continent, there is a strong propensity to report non-financial information using uniform guidelines to allow international comparability [22].

Perhaps, on the stimulus provided by the institutional bodies and investors, something is changing. Recent news concerning this is testified to the intention of the International Integrated Reporting Council (IIRC) and the SASB to merge into a unified organization, the Value Reporting Foundation, providing a comprehensive corporate reporting framework across the full range of enterprise value drivers and standards to drive global sustainability performance. The IFRS published, in September 2020, the “Consultation Paper on Sustainability Reporting” [23] to define a single NFD framework and, therefore, promoting the standardization of non-financial information.

Comparability of NFD is considered as a relevant issue not only among the standard setters and regulators but also between the different NFD users [6,10,11,12,13]. Comparability allows stakeholders to evaluate the overall performance of a company and make them “able to make peer-to-peer comparisons” [12] (p. 71). Recent surveys highlight that professional investors, in particular, prefers “non-financial information that is concise, comprehensive, comparable, and credible” [11] (p. 127). The same relevant aspect is underlined by the Big Four. KPMG states that “The public interest requires the same for non-financial reporting so that information is relevant, comparable and reliable for decision-making” [14] while PWC argues that “Investors are interested in comparability against peers and within the company itself over time” [15] (p. 8). The ACCA affirms, “in order for non-financial information to be useful to investors, it must be comparable across companies” [16] (p. 4). In all, comparability allows users to make sense of the non-financial information disclosed, to enhance the usefulness of accounting information, to identify benchmarks and best practice as well as classifying and ranking organizational performances [42,69].

In summary, from a conceptual perspective, it seems that comparability is something desirable.

Despite this, NFD comparability appears to be still problematic in practice for several reasons [6,7,24,25,26,69]. First, there is a large number of frameworks [28], and companies are completely free in the choice of which framework to adopt [7,26]. Second, the different frameworks, even the most detailed ones, allow preparers to exercise their discretion over the choice of the indicators to report, the calculation methods and the related narratives to provide [30,50]. Third, the comparability of the NFD can be influenced by specific characteristics of the companies, such as size [70], type of operating sector [25], corporate and CSR governance [71,72,73], the degree of sensitivity to the social and environmental issues of companies [25,29,31,53] and the wide margins of discretion granted to the preparers [32]. Forth, specificities of the context (e.g., country, social, economic, political, cultural background, regulatory issues, stakeholder influence, etc.) are determinant in defining the quality of NFD and its comparability [26,57,74].

Another phenomenon should also be considered as a barrier against NFD comparability: “greenwashing”, an opportunistic behavior [36,37] which deters the credibility of the information disclosed and, consequently, the usefulness of NFD for investors [34,35,38]. Recently, the NFD has received criticism as it sometimes lacks “authentic effort” and failing to meet users’ expectations [35,39,40,41]. Despite this, some argue that mandatory NFD “promotes better comparability, but may also help prevent some companies from “greenwashing” or deliberately manipulating stakeholders’ perceptions through discretionary reporting” [45] (p. 20). In addition, some scholars also highlight that the sustainability rating act to deter greenwashing can reduce the risk that greenwashing affects the value of a firm [46] and promote comparable non-financial information [27].

Although cross-sectoral comparability appears to be problematic [12], some authors highlight the possibility of comparing the NFD of companies operating in the same sector with better results, especially if they adopt the same reporting framework. The idea is that companies operating in the same sector have similarities in terms of business models, environmental impact, governance, stakeholders, etc., and, thus, several of the abovementioned barriers should be overcome, and a large part of their NFD should be “naturally” comparable [26,75,76].

It should also be considered that some studies have shown that among companies operating in the same sector, there is a tendency towards convergence. This situation may be induced either by the willingness of companies to meet the needs of users to have comparable reports [42] or by procedures adopted by consultants who, adopting an approach guided by efficacy and effectiveness, i.e., strongly standardized, tend to replicate tested and consolidated models, limiting innovation and customization [77]. Still, at an intraindustry level, it seems possible that companies tend to adopt emulation and convergence processes towards common practices as well as practices to prove to be the “best in class” [42,43].

Moreover, it must be considered that recent studies have also highlighted that companies tend to adopt two main frameworks: the GRI Guidelines and the IIRC Framework. Consequently, the comparability problems related to the existence of different NFD frameworks should be limited. In all, while interindustry comparability seems still to be challenging, several arguments suggest that intraindustry comparability can be achieved, maybe not at a microlevel (indicator), but, at least, on a meso or macro level, especially if companies adopt the same reporting framework. This is the idea that this paper aims to explore.

3. Design of the study

The object of the analysis is represented by the English version of the NFD reports published by companies on their institutional websites. The web dissemination channel of information of a purely accounting nature is justified by the potential to be a communication tool capable of ensuring transparency, as can be seen from previous studies on the subject [31,78].

The investigation considers the NFD of the firms indexed to DJSI. These firms were identified as follows: the list of companies indexed to DJSI was downloaded from the database available on the SAM website, updated to 24 September 2018, and includes the reference to the country and industry for each company. The companies included in the DJSI are 317, belonging to 24 different industries. Of the 317 listed companies, 46% originates from Europe, 26% from Asia, 22% from America, 5% from Australia and 1% from Africa.

The choice to focus on DJSI firms is due to three sets of reasons. First, DJSI reports the performance of the world’s leading companies in terms of economic, environmental and social criteria and implicitly includes companies geared towards a strong commitment to corporate social responsibility (CSR) [79]. Second, since 1999, DJSI has been offering coverage in multiple markets, and this allows investigating differences between sectors. Finally, the existence of sustainability information directly accessible on the company’s website is one of the assessment parameters in the DJSI rating (see for more details what is defined in the revisions of “SAM Corporate Sustainability Assessment (CSA)” available at www.robecosam.com). In particular, the “2019 SAM Corporate Sustainability Assessment—Annual Scoring and Methodology Review” stated that “Assessed companies are expected to name and describe the living wage methodology used, and provide supporting evidence and indicate if the information is available in their public reporting or corporate website.”. The same assessment methodology recognizes that “Reporting and collecting high-quality sustainability information is the critical first step towards ensuring that sustainability information is more widely accepted and used by the investment community” and that the “data needs to be precise and comparable” [80] (p. 5) to provide investors with meaningful sustainability data to enable better-informed investment decisions. Consequently, several of the abovementioned factors that generate non-comparability should be mitigated for the companies of the sample.

The choice of the mentioned sample should also reduce the risks related to greenwashing and the consequent non-comparability related to this phenomenon [35,36,37,38]. The sustainability rating process related to the DJSI allows excluding from the group companies that adopt “greenwashing” practices [81]. In detail, the evaluation and review process employed in the development of the DJSI index is based on the application of criteria built into the corporate sustainability assessment, which quantifies the sustainability performance of a company by assigning a corporate sustainability performance score [82]. Once a company is included in the index, it is monitored with respect to the “critical environmental, economic, and social issues or crisis situations that can have a highly damaging effect on its reputation” [83]. Thus, these firms cannot declare but not adopt sustainable practices in practice.

As the aim of the paper is to analyze intraindustry comparability, two specific industries were selected: the energy and the banking ones. The choice is because, in both sectors, sustainability issues are particularly relevant and pushed forward by stakeholders as well as by law and policymakers. In particular, both in the energy industry and the banking industry, the labor, the environment and the social dimensions are particularly important, even if with different weight and perspective. Even if the two sectors appear to be strongly different, they have in common a relevant impact on sustainable issues. Previous research demonstrated that the energy sector is characterized by a high commitment to and development of sustainability practices [84,85]. At the same time, the bank sector presents the increased relevance of non-financial information [86,87,88]. Additionally, both sectors are under pressure by regulatory bodies and investors and, thus, operate in highly regulated markets and have a social and environmental, other than economic, obligation to meet the stakeholders’ expectations [84,86]. Drawing upon these considerations, we intentionally selected these two sectors to understand similarities or differences in terms of NFD comparability.

From the selection carried out, there are 45 companies in the sample, 16 of which are in the energy sector and 26 from the banking sector.

The NFD reports considered for the analysis are all related to the financial year 2019 except for four companies (two from the energy industry and two from the banking one), for which the last report available is for the financial year 2018.

To analyze NFD comparability, a manual text analysis of the companies′ reports was performed as it allows understanding how ideas and concepts (i.e., framework) are perceived, interpreted and translated into reality (i.e., NFD) [48,49]. The choice not to use a software (such as, for example, whose are typical for content analysis, as Wordstat 7, Nvivo, TLab) derives from the fact that much non-financial information is presented with the use of image, icons, graphics that cannot always be processed by these tools.

The research protocol is as follows:

The first step involves a structural analysis that includes the identification of the reference framework (GRI Guidelines, IIRC Framework, etc.) used for the preparation of the report, the total number of pages of the report and the analysis of the report summary. As far as the counting of the total number of pages of the report is concerned, only sections containing sustainability information were considered; it follows that any additions (e.g., financial statements) were not considered. The minimum, average and maximum number of pages of the NFD reports were then calculated to have some synthetical parameters.

The second step involves the analysis of the sections contained in each report. It was decided to focus attention on human capital and the environmental parts. This choice relies on the fact that several studies emphasize the importance of investing in human capital and the environment as it allows companies to create a sustainable competitive advantage and because these are two dimensions that surely all companies report [29,89,90,91]. Some further “context” elements have also been considered, such as the number of pages devoted to the two information dimensions: human capital and environmental capital. The counting of the number of pages, in this case, was carried out starting from what can be extracted from the index; it follows that any in-depth analysis in sections not properly focused on the two dimensions investigated has not been considered.

The third and last step provides for the highest degree of deepening and enters the content of the reports. Regarding human capital, the indicators were identified through a text analysis searching for words related to this dimension. In particular, a semi-objective approach was adopted as it is used by a large body of literature in the domain [24]. The team has initially defined a set of keywords related to human capital moving for the extant studies. Then, this list has served as a guide for identifying the most frequently used words. Thus, the keywords used for the investigation were defined as follows: “human”, “employ”, “personnel”, “workers”. The identified indicators were then traced back to the macro-categories of GRI indicators, version 2016, relating to human capital and available in the fact sheets. Each GRI indicator was then matched with the different “labels” it assumes in practice.

To ensure the validity and reliability of the work, the data collection and analysis was carried out independently by the researchers involved, who, at a later stage, were confronted to identify and understand the reasons for possible differences to bring them back to units (i.e., triangulation). The generalizability of the results was considered in terms of theoretical generalizability, i.e., as a possibility to trace the results achieved in this work to those of similar previous studies.

4. Results

Once the reports were downloaded, a first classification was made according to the name of each of them (see

Table 1. Names of reports.

	Energy	Bank	Total
Annual Report	4	6	10
Integrated report (including similar reports: integrated report, integrated management report, integrated annual report)	3	5	8
Sustainability report (including similar reports: sustainability performance report)	6	5	11
CSR report (including similar: corporate social responsibility, social responsibility)	1	4	5
Other	2	7	9
Non-financial data and engagement		1	1
Management report		1	1
Consolidated non-financial statement		1	1
Socioeconomic group impact		1	1
Environmental, social and governance report		2	2
Non-financial declaration		1	1
Global Stewardship report	1		1
Sustainable development report	1		1
Total	16	27	43

).

In the energy industry, the most widely used name is “Sustainability report”, while for the banking industry, it is curious to point out that 25% of companies (7 out of 27) adopt a different name, not homogeneous with any of the other companies examined.

About the length of the reports, it is evident that it is not homogeneous (see

Table 2. Number of pages in each report.

	Energy			Bank
	Minimum	Maximum	Average	Minimum	Maximum	Average
Annual report	70	217	138	40	246	115
Integrated report (including similar reports: integrated report, Integrated management report, Integrated annual report)	42	166	112	34	120	80
Sustainability report	82	169	107.14	10	182	111.2
CSR report (including similar: corporate social responsibility, social responsibility)	72	72	72	74	148	120.25
Other	48	88	68	28	224	118.14

). There are very short reports (e.g., 10 pages) and very long ones (of 246 pages).

As mentioned, the table of contents of the reports was also analyzed. The overall analysis shows that it is not possible to trace all reports to a single structure. In addition, there are similar names used to report the same dimensions (e.g., the section dedicated to human capital can be called “our people”, “employees”, etc.). For most of the content elements, however, the structure used is not easily traceable to a predefined model.

Another aspect related to the first step of analysis is related to the used framework. As already stated in the literature, the most widely used frameworks for NFD are the GRI Guidelines. This is valid for both the energy and banking sectors (

Table 3. Standards applied.

	Energy	Bank
Reporting standards for economic, social and environmental information
GRI Guidelines	15	21
IIRC Framework	4	7
AA1000 Principles	4	2
Principles of socially responsible conduct
UN Global Compact	6	19
Reporting standards for specific dimensions (environmental)
Recommendations of the Task Force on Climate-Related Disclosures (TCFD)	10	22
Greenhouse gas emissions (GHG)	14	15
Carbon Disclosure Project	1	8
Reporting standards and principles of conduct for specific sectors
Guide for the preparation of management reports of listed companies of the CNMV	1	1
EFQM model criteria	1	0
ISAE 3000	0	6
AA1000AS	1	2
IPIECA reporting guidelines (for the oil and gas industry)	2	0
UNEP FI Principles for responsible banking	0	8
Principals for Responsible Investments (PRIs)	0	2
The Equator Principles (for financial institutions)	0	9
ISO 26000	0	3
Sustainability Accounting Standards Board—SASB (sectoral standard for United States companies)	0	6

). One result to be highlighted is that four energy companies and six companies in the banking sector use both the GRI Guidelines and the IIRC Framework: while the IIRC Framework is mainly used at the structural level, the GRI Guidelines are used at an indicators’ level. This suggests that comparability should exist between companies that adopt the same framework. Regarding the reporting standards on specific dimensions, the recommendations of the Task Force on Climate-Related Disclosures (TCFD) and greenhouse gas emissions (GHG) have a high percentage of applications.

The second step of analysis concerns the sections included in each report. The focus is on two particular sections: human capital and environment (see

Table 4. Number of sections dedicated to the environment and human capital.

	Energy	Bank
Environment Section
Number of companies with a section dedicated to the environment	10	16
Number of companies with two sections dedicated to the environment	2	4
Number of companies with three sections dedicated to the environment	1	1
Number of companies with no section dedicated to the environment	3	6
Human capital section
Number of holdings with a section dedicated to human capital	8	16
Number of holdings with two sections dedicated to human capital	4	2
Number of holdings with three sections dedicated to human capital		3
Number of holdings with no section dedicated to human capital	4	6

).

As for the environmental section, 81% of companies in the energy industry and 78% of the banks have at least one specific section dedicated to the environment. Companies in the energy industry devote on average 11 pages to this topic while the ones in the banking industry about 12 pages. Thus, the average length is more or less similar. Even if the average is similar, both industries present a high variability: in the banking industry, the section moves from a minimum of one page to a maximum of 54 pages. For human capital, on the other hand, 75% of energy companies devote at least one section to human capital and 78% of the banking sector. The average number of pages per section is 12 pages for the energy industry and 15 pages for the banking one. In this case, the range varies between two pages and 52 pages.

The third level of analysis concerns only human capital (see

Table 5. Number of human capital indicators.

	Energy			Bank
Report GRI Compliant	15 Out of 16 Sample Reports			21 Out of 26 Sample Reports
	Minimum	Maximum	Average	Minimum	Maximum	Average
Total number of indicators	18	143	68.53	22	156	67.33
Number of quantitative indicators	7	130	47.27	16	89	50.19
Number of narrative indicators	2	60	21.33	1	79	17.14
Number of sections concerned	2	8	4.33	2	13	5.57
Report NOT-GRI Compliant	1 out of 16 Sample Reports			5 Out of 26 Sample Reports
	Minimum	Maximum	Average	Minimum	Maximum	Average
Total number of indicators	136	136	136	5	26	12.60
Number of quantitative indicators	92	92	92	3	21	10.80
Number of narrative indicators	44	44	44	0	5	1.80
Number of sections concerned	5	5	5	1	5	3.40

).

First of all, it is noted that the number of quantitative and qualitative indicators in each report for both the banking and energy industries go from a minimum of three indicators to a maximum of 143.

Table 6. GRI indicators coupling human capital—name used.

Designation of GRI Indicators		Number of Companies Reporting It		Number of Names Used
		Energy	Bank	Energy	Bank
Information on employees and other workers.	(a) Total number of employees by employment contract (permanent and temporary), by gender. (b) Total number of employees by employment contract (permanent and temporary), by region.	9	8	18	32
	(c) Total number of employees by employment type (full-time and part-time), by gender.	5	11	10	36
	(d) Whether a significant portion of the organization’s activities are performed by workers who are not employees.	2	0	5	0
Ratios of standard entry-level wage by gender compared to local minimum wage.	(a) When a significant proportion of employees are compensated based on wages subject to minimum wage rules, report the relevant ratio of the entry-level wage by gender at significant locations of operation to the minimum wage.	2	3	5	3
New employee hires and employee turnover.	(a) Total number and rate of new employee hires during the reporting period, by age group, gender and region.	9	10	29	28
	(b) Total number and rate of employee turnover during the reporting period, by age group, gender and region.	7	15	25	32
Benefits provided to full-time employees who are not provided to temporary or part-time employees.	(a) Benefits, which are standard for full-time employees of the organization but are not provided to temporary or part-time employees by significant locations of operation.	7	15	12	27
Parental leave.	(a)Total number of employees that were entitled to parental leave by gender. (b) Total number of employees that took parental leave, by gender. (c) Total number of employees that returned to work in the reporting period after parental leave ended, by gender. (d) Total number of employees that returned to work after parental leave ended that were still employed 12 months after their return to work, by gender. (e) Return to work and retention rates of employees that took parental leave by gender.	5	9	15	40
Minimum notice periods regarding operational changes.	(b) For organizations with collective bargaining agreements, report whether the notice period and provisions for consultation and negotiation are specified in collective agreements.	6	7	10	11
Types of injury and rates of injury, occupational diseases, lost days, and absenteeism, and number of work-related fatalities.	(a) Types of injury, injury rate (IR), occupational disease rate (ODR), lost day rate (LDR), absentee rate (AR), and work-related fatalities, for all employees, with a breakdown by (1) region; (2) gender.	13	13	96	61
Health and safety topics covered in formal agreements with trade unions.	(a) Whether formal agreements (either local or global) with trade unions cover health and safety. (b) If so, the extent, as a percentage, to which various health and safety topics are covered by these agreements.	1	1	2	1
Average hours of training per year for employees.	(a)Average hours of training that the organization’s employees have undertaken during the reporting period, by (1) gender; (2) employee category.	12	19	52	63
Programs for upgrading employee skills and transition assistance programs.	(a) Type and scope of programs implemented and assistance provided to upgrade employee skills.	9	14	26	41
	(b) Transition assistance programs provided to facilitate continuous employability and the management of career endings resulting from retirement or termination of employment.	4	6	5	11
Percentage of employees receiving regular performance and career development reviews.	(a) Percentage of total employees by gender and by employee category who received a regular performance and career development review during the reporting period.	8	10	22	21
Diversity of governance bodies and employees.	(b) Percentage of employees per employee category in each of the following diversity categories: (1) gender; (2) age group: under 30 years old, 30–50 years old, over 50 years old; (3) other indicators of diversity where relevant (such as minority or vulnerable groups).	10	21	21	62
Ratio of basic salary and remuneration of women to men.	(a) Ratio of the basic salary and remuneration of women to men for each employee category, by significant locations of operation.	3	2	6	2

shows the reclassification of the selected indicators in the macro-categories of the GRI, the number of companies reporting it and the number of labels it assumes. Companies that declare that to comply with the GRI (and even less those that do not apply it) use a large number of different names to monitor the same aspect.

This analysis shows that the names used by the different firms are not the same, even if the differences are limited. An example is the indicators on the types of injuries and rates of injuries, occupational diseases, lost days, absenteeism and the number of fatal accidents related to work. In both the energy and banking industries, there are 96 and 61 denominations, respectively, but the comparability of the data for the reader appears to be guaranteed. The different names used are: “absentee rate”, “absenteeism”, “absenteeism rate”, etc. Another example is the indicator for average annual training hours per employee, for which there are 52 different names in the energy sector and 63 in the banking sector. Even between these, the differences are easily overcome by the reader: in fact, it is named “average education hour per person”, “average hours of training per employee”, or “hours of training received by professional”. Nevertheless, there are a few instances where the names used for the same indicator differ completely. An example is an indicator for programs for upgrading employees’ skills and transition assistance programs. The names used are: “description of employee development program”, “employee assistance program utilization (%)”, and “investment in the quality of life programs”. Concerning the value of the indicator, the calculation method is not indicated. The different names used, and the non-explanatory method of calculation make comparability difficult.

Another analyzed aspect was the section of the document where the indicator is included. The same indicator is included by companies in different sections that are often not even easily traceable. For example, the same indicator is presented by one company in the “social employee” section and by another in the “key indicators” section. From

Table 7. GRI indicators coupling human capital—sections.

	Energy					Bank
	Number of Companies Reporting It	Number of Companies Where the Indicator is Contained in:				Number of Companies Reporting It	Number of Companies Where the Indicator is Contained in:
		One Section	Two Sections	Three Sections	More Than Three Sections		One Section	Two Sections	Three Sections	More Than Three Sections
(a) Total number of employees by employment contract (permanent and temporary), by gender. (b) Total number of employees by employment contract (permanent and temporary), by region.	9	8	0	1	0	8	3	4	1	0
(c) Total number of employees by employment type (full-time and part-time), by gender.	5	4	1	0	0	11	6	4	0	1
(d) Whether a significant portion of the organization’s activities are performed by workers who are not employees.	2	2	0	0		0	0	0	0	0
(a) When a significant proportion of employees are compensated based on wages subject to minimum wage rules, report the relevant ratio of the entry-level wage by gender at significant locations of operation to the minimum wage	2	1	1	0	0	3	3	0	0	0
(a) Total number and rate of new employee hires during the reporting period, by age group, gender and region.	9	6	3	0	0	10	6	3	1	0
(b) Total number and rate of employee turnover during the reporting period, by age group, gender and region.	7	5	2	0	0	15	13	2	0	0
(a) Benefits, which are standard for full-time employees of the organization but are not provided to temporary or part-time employees by significant locations of operation.	7	6	0	1	0	15	10	2	1	2
(a) Total number of employees that were entitled to parental leave by gender. (b) Total number of employees that took parental leave, by gender. (c) Total number of employees that returned to work in the reporting period after parental leave ended, by gender. (d) Total number of employees that returned to work after parental leave ended that were still employed 12 months after their return to work, by gender. (e) Return to work and retention rates of employees that took parental leave by gender.	5	4	1	0	0	9	7	1	0	1
(b) For organizations with collective bargaining agreements, report whether the notice period and provisions for consultation and negotiation are specified in collective agreements.	6	5	1	0	0	7	5	2	0	0
(a) Types of injury, injury rate (IR), occupational disease rate (ODR), lost day rate (LDR), absentee rate (AR), and work-related fatalities, for all employees, with a breakdown by (1) region; (2) gender.	13	3	6	0	4	13	5	7	0	1
(a) Whether formal agreements (either local or global) with trade unions cover health and safety. (b) If so, the extent, as a percentage, to which various health and safety topics are covered by these agreements.	1	0	2	0	0	1	1	0	0	0
(a) Average hours of training that the organization’s employees have undertaken during the reporting period, by (1) gender; (2) employee category.	12	6	5	1	0	19	9	6	3	1
(a) Type and scope of programs implemented, and assistance provided to upgrade employee skills.	9	4	3	2	0	14	10	2	0	2
(b) Transition assistance programs provided to facilitate continued employability and the management of career endings resulting from retirement or termination of employment.	4	4	0	0	0	6	5	1	0	0
(a) Percentage of total employees by gender and by employee category who received a regular performance and career development review during the reporting period.	8	3	3	0	2	10	6	4	0	0
(b) Percentage of employees per employee category in each of the following diversity categories: i. Gender; ii. Age group: under 30 years old, 30–50 years old, over 50 years old; iii. Other indicators of diversity where relevant (such as minority or vulnerable groups).	10	6	2	0	2	21	16	4	0	1
(a) Ratio of the basic salary and remuneration of women to men for each employee category, by significant locations of operation.	3	3	0	0	0	2	2	0	0	0

, it is evident that for the majority of companies, the indicators are found in only one section, but there are several cases in which the same indicator is reported in even more than three sections. This leads to the fact that all companies in the do not always report the same indicator in the same section: indicators are located in the sections considered as “the most appropriate” by each preparer. In other words, the use of different report structures involves the inclusion of indicators in different sections, which makes the reports not directly comparable.

5. Discussion

This study aimed to investigate inter-firm comparability of NFD at a macro (framework), meso (structure) and/or micro (indicators) level. To achieve this aim, a text analysis [48,49] of NFD was performed focusing on the companies operating in the banking and energy industry and indexed in the DJSI.

The performed investigations suggest the following reflections.

From a macro perspective, the first aspect that emerges is the variety of names used to label the NFD. Besides the “institutional” ones (e.g., Integrated report, Sustainability report, etc.), there are names that appear to be self-made to catch the attention of the users (e.g., Global Stewardship report, ESG report, etc.). This approach partly recalls the idea of some practices of window-dressing [92] and can, partly, act as a barrier for users that search a “sustainability report” of a company and make users believe, at first sight, that what they are looking at is not a “sustainability report”, but something different.

Concerning the frameworks, it emerges that the investigated companies adopt a plethora of different frameworks even if they operated in the same industry. On a first stance, it seems possible to argue that comparability is not possible. Nevertheless, some additional aspects must be considered. First, the study shows that the GRI Guidelines are the most widely used for NFD, followed by the IIRC Framework. This result, as expected, confirms the results of previous studies: the companies, regardless of their sector, mainly use the GRI Guidelines and less the framework of the IIRC, thus prioritizing the drafting of sustainability reports over the integrated report [5,56,62]. This prevalence of GRI-compliant reports should facilitate NFD comparability. In addition, the analysis also shows that although the GRI and IIRC models are conceptually different and proposed as alternatives, there is mutual contamination in practice: companies that adopt the IIRC Framework tend to identify the indicators related to each of the six capitals (one of the content elements of the IIRC Framework) by referring to the GRI Guidelines. In this way, companies can exploit the potential of the two models, namely flexibility, conciseness and focus on the value creation of the IIRC model as well as completeness, support and level of detail offered by the GRI. In summary, a spontaneous contamination process among different frameworks can be identified.

Regarding the length of the NFD, the analysis shows two main aspects. First, integrated reports and sustainability reports tend to be as long as the annual reports: thus, the idea of conciseness seems that it has not been translated into practice and/or it is difficult to balance the idea of completeness with the one of conciseness to have effective communication that. Second, there is a large variability of the reports in terms of the number of pages. This is the first hint of the existence of problems related to comparability issues.

Moving to a meso-level of analysis, it emerges that the structure of the reports is not comparable but appears to be freely defined by each company, despite operating in the same industry, having similar characteristics and adopting the same framework. This conclusion seems to support the argument put forward by Cardoni et al. [26] and by Boiral [53] that NFD cannot be comparable, even at an intraindustry level. The presence of highly differentiated structures limits direct comparability; in theory, it can be achieved through a reclassification of the different sections according to a specific model. In all, comparability can be achieved in part, but it requires some re-elaboration of the reports. Moreover, also the different labels adopted by companies to identify the same sections can represent a barrier for immediate comparability. In other words, the phenomenon of “label creativity”, i.e., similar elements labeled with different names [93], can be found at a structural level: similar sections of the NFD labeled differently.

At a micro- (indicator) level, the analysis has shown different comparability barriers related to the quantity, quality and position of the different indicators. Regarding the number of indicators, it is possible to identify a high variability of the indicators used to describe an organizational human capital (from 18 to 156). This means, of course, that there is a core of indicators that are adopted by most of the firms, but also other indicators that can be adopted in a few situations. Regarding the quality of indicators, the first aspect that emerges is the phenomenon of label creativity is particularly important when it comes to this level of analysis: indicators that seem to represent the same issue are often labeled differently. Sometimes labels are very similar, but in other cases, they are very different. Another qualitative aspect relates to the fact that the calculation method of the indicators is often not disclosed. Thus, even if the user can identify the same indicator in two reports, he cannot be sure that the measurement process is the same and, consequently, that the data are comparable. Regarding the position of the indicators in the report, the study shows that the same (or similar type of) indicator is not always presented in the same section. The users have then to search for it, collect the data and then compare. This leads to the idea that to achieve comparability of NFD; detailed rule-based standards should be provided instead of the principle-based ones now in force: the mentioned spontaneous forces (e.g., emulation processes, etc.) appears to be too weak to transform the idea of comparability into reality. Nevertheless, the problems related to a “one-size-fits-all” approach for NFDs should also be considered given the presence of narratives and other peculiarities that characterize such reporting that can be difficult to standardize or harmonize [94].

6. Conclusions

The comparability of NFD is a much debate topic from standard-setters (IASB, IFRS, IIRC), professional societies (KPMG, PWC, ACCA), academics [6,24,26], regulatory bodies and by companies’ stakeholders [6,10,11,12,13]. From the literature, it emerges the need to study “the differences between de jure (rules) and de facto (practices)” [7] (p. 603) on the NFD comparability. Thus, this paper investigates inter-firm comparability of NFD at a macro (framework), meso (structure) and/or micro (indicators) level. To achieve this aim, a manual content analysis was performed, focusing on the NFD published by companies operating in the energy and banking industries and included in the DJSI.

The main conclusions are as follows: It is overall possible to identify the potential of comparability on a macro and meso level. There are a phenomenon and contamination between the two main frameworks, i.e., the IR and the GRI. Moreover, even if some companies adopt self-made labels for their reports, there is a tendency to name them as IR or sustainability reports. Finally, even if the structure of the NFD is not homogeneous, it is possible to re-classify the information provided to have a comparable structure. With regard to the microlevel of analysis, the study shows that there is a core of indicators that are potentially comparable, but also a large number of indicators that are used by a few companies and, consequently, they are difficult to compare. In addition, the phenomenon of label creativity and the lack of information about how each indicator is calculated in practice can act as barriers against comparability. In summary, while there is, in practice, a substantial homogeneity of frameworks and dimensions of analysis that can allow comparability, even if often only after a re-elaboration of the report, i.e., comparability at a macro and meso-level, comparability of indicators (microlevel) is not achieved in practice. The lack of comparability at a micro level appears to be mainly related to the level of freedom that the preparers have and to the principle-based approach adopted by the standard-setters.

This study contributes to research into NFD and NFD comparability from an empirical point of view, offering different levels of analysis. In addition, the study shows that a convergence-contamination process among different frameworks can be identified in practice. Finally, this study contributes to the search question if “good quality reporting needs good quality regulation or whether good reporting can develop without good regulation” [47] (p. 150).

In terms of policymaking, the achieved results can support the discussions regarding the need for a tight and unique regulatory framework for NFD. Today several governments and standard setters are discussing the opportunity of creating a common (and unique) set of rules for NFD. The matter is not only the development of this new standard but also the decision about the approach it should adopt (principle vs. rule-based), and about the degree of freedom it should leave to the prepares to make NFD useful. In fact, between complete freedom and a one-size-fits-all approach, it is possible to consider intermediate positions like the definition of a mandatory structure (meso level) with a minimum set of mandatory indicators.

From the managerial point of view, this study contributes to the discussions about to what extent preparers should provide users with comparable NFD. At the moment, the preparators have great freedom in the drafting of NFDs regardless of the standard used, which is why our study directs preparers to standardize the drafting of reports, also using similar names and not fall into the phenomenon of label creativity.

The limitations of this work are mainly attributable to the analysis sample and to the method used to capture the detailed information. Only two areas were considered in the study: although the analysis is therefore not exhaustive, it shows that there is no comparability in two strongly different sectors. The second limitation concerns the application of the text analysis method solely to collect human capital indicators. It also must be considered that if it is true that the analyzed sample should minimize the risk of a phenomenon of “greenwashing” [35,36,37], which certainly reduces the comparability of the reports published by companies, it is not strictly excluded that some of the results achieved are due to this phenomenon.

Research perspectives are represented by extending the analysis to other industries and deepening to further dimensions of sustainability. Other interesting future research could concern the analysis of the comparability of NFDs from different countries with different regulations [33]. The purpose would be to check whether the legal system, besides being a determining factor in its level of regulatory compliance [33,95] it is a determining factor of comparability of NFD. Finally, another interesting research could verify the possible presence of a greenwashing phenomenon that would affect the comparability of the NFDs of companies, either in mandatory or voluntary context, to support findings of previous research [27,44,45,46].