The quality assessment of stakeholder engagement disclosure in the EU mandatory non-financial reporting framework

Purpose – This article investigates the quality of stakeholder engagement (SE) process disclosure in the context of non-financial reporting (NFR) introduced by Directive 2014/95/EU (NFRD). SE implies the involvementof the subjects interested in the organization ’ s activity, according to the principle of inclusiveness and the key concepts of the stakeholder theory (ST). Design/methodology/approach – Theauthorsconductedacontentanalysison75non-financialstatements (NFSs) published by companies listed on the Italian Stock Exchange in 2018 and 2021 to evaluate the evolutionary profiles of SE quality through the years. Findings – The average level of SE is not significantly high. The research showed an overall poor quality of disclosureconcerningstakeholders ’ keyexpectationsandissuestobeaddressedandanswered.Furthermore,a certain variability emerged in the quality of the disclosure between the various reports, and no significant improvements in SE quality were noted from 2018 to 2021. Research limitations/implications – The conclusions provide a replicable method for the analysis of SE quality in NFSs and the development of new standpoints in the ongoing debate on the implications of mandatorylegislativeframeworksforNFR.Contentanalysesintrinsicallypresentmarginsofsubjectivity.The samplewaslimitedtoasubsetofNFSfromItaly;hence,theresultscouldbecountryspecific. Practical implications – This work suggests some possible ways of improvement of SE practices by companies. Originality/value – Original assessment model based on eight variables identified from the academic literature and the most common international sustainability reporting standards. These variables were stakeholder identification, stakeholder selection process, degree of involvement, SE approach, dialogue channels, SE results, different points of view and integration of the SE process.


Introduction
Non-financial reporting (NFR) represents a topic of significant interest in academic studies and professional practice (Haller et al., 2017;Stolowy and Paugam, 2018), which has recently received further impetus from legislative changes at the international level.This represents a sign of the widespread and increased need for greater disclosure required to companies, which is not limited to purely economic-financial aspects, but also considers issues relating to environmental, social and governance (ESG) topics, with specific emphasis on some aspects, such as respect for human rights, diversity, and the fight against corruption (Alsayegh et al., 2020;Cardoni et al., 2020;Manita et al., 2018).
Stakeholder engagement (SE) (Greenwood, 2007;Noland and Phillips, 2010) is one of the most characteristic and complex activities that have emerged in preparing nonfinancial reports.SE draws its main theoretical reference from stakeholder theory (ST) (Donaldson and Preston, 1995;Freeman, 1984).The underlying and common purpose of ST and SE is substantiated by the need to first understand the requests and expectations on information of multiple categories of stakeholders (Freeman et al., 2010).Therefore, SE implies the consideration of stakeholders' visions and points of view to evaluate the significant information to be communicated in relation to company performance (Lopatta et al., 2017).
In this sense, SE can be considered an instrument that can limit information asymmetries between companies and investors (Garc ıa-S anchez and Noguera-G amez, 2017).Moreover, the strategic role of SE is rising, for example, for the development of innovation management and entrepreneurial development (Leonidou et al., 2020).Given that SE is important in the context of NFR processes (Zarzycka et al., 2021), a reflection on the ways in which it is conducted is essential.In particular, conducting SE that aims more at seeking stakeholders' consensus rather than their real involvement misrepresents the purposes and potential positive effects of this practice (Manetti, 2011).
For these reasons, the issue of the quality of the SE process appears essential.Some preliminary studies have shown that many companies fail to provide complete disclosure of SE modalities (Moratis and Brandt, 2017), and the general quality level of SE processes appears to be quite low (Venturelli et al., 2018).Nonetheless, in light of the relevance of the subject, the number of academic contributions on SE quality seems to remain limited, and this work aims to deepen knowledge on this topic.
Academic research on SE represents a developing line of investigation because of both the scientific and professional growing interest in issues of sustainability reporting (Camilleri, 2015;Stocker et al., 2020) and to one of the most important innovations recently introduced by the EU legislation, namely, the obligation to publish a document reporting the sustainability performance, called "non-financial statement" (NFS) and introduced by the Directive 2014/95/ EU on NFR (NFRD).It contains the obligation to communicate information of a non-financial nature for public interest entities that exceed certain dimensional parameters.These companies have been obliged to publish the NFS, a report with relevant information of a nonfinancial nature relating to environment, social, personnel, compliance with human rights and fight against active and passive corruption (La Torre et al., 2018).
Nevertheless, the NFRD represents only the first step of a possible path of institutionalization (Pizzi et al., 2023) of sustainability reporting in the European context.The EU has recently approved a new regulation, Directive 2022/2464/EU (CSRD).It aims to fill the gaps in the current NFRD framework on the communication of non-financial information, which is considered of insufficient quality for investors.The declared objective of the CSRD is twofold: on the one hand, to promote the publication by a wider audience of companies than that envisaged by the first directive of sustainability reporting, an expression that will replace that of NFR (Baum€ uller and Sopp, 2022); on the other hand, to favor an even more effective disclosure about the ESG communication profiles toward external stakeholders.
The first studies on the application of the NFRD seem to suggest that this directive has pushed improvements in sustainability reporting quality and quantity (Ottenstein et al., 2022).However, these issues require the development of further study (Venturelli et al., 2022), which will also have to be conducted in light of the new regulatory role promoted by the application of the CSRD.In particular, several authors have argued that progress is needed with respect to the NFRD on the need to introduce a greater focus on sustainability and more integration between non-financial and financial aspects and of new perspectives about the introduction of a new materiality concept (Fiandrino et al., 2022;Raith, 2023).
The role of legal frameworks, like that proposed by the EU, to favor more effective forms of reporting connected with ESG topics is constantly growing (Dumitru et al., 2017;Venturelli et al., 2019).The relevance of this topic is incentivized by the increasing centrality of the Sustainable Development Goals promoted by the United Nations for the coherent economic development of the competitive environment in which companies operate (Pizzi et al., 2022;Pizzi et al., 2021b).In this sense, the present study aims at extending the existing literature, which analyzes the role that mandatory non-financial disclosure could play in terms of improving NFR quality (Carungu et al., 2020;Mart ınez-Ferrero et al., 2015), with specific reference to the SE process, within an interpretative framework consistent with ST.
For this research, the analysis focused on the NFSs published according to the NFR framework by the companies listed in the FTSE MIB index of the Italian Stock Exchange, which represents the most significant Italian stock index in terms of size and liquidity of the companies belonging to it.We used a research sample represented by the main listed companies in Italy, homogeneous from a dimensional point of view, thus favoring a comparative analysis on the topics under investigation.Therefore, this work aimed to investigate the quality of the disclosure of the SE process for the companies included in the research sample.To this end, an interpretative theoretical framework was proposed to carry out an empirical analysis based on an assessment model that provides for examination through a manually performed content analysis (Beck et al., 2017;Krippendorff, 2018) of eight research variables.These variables were identified on the basis of an academic literature review and the indications reported in the main international reporting standards applicable.
Therefore, this work seeks to analyze the effectiveness of NFRD in improving SE practices (Korca and Costa, 2020), similar to other studies on related topics (Zarzycka and Krasodomska, 2022).Hence, this research focuses on the importance of performing effective SE in approaching NFR practices-a challenging topic in the academic debate concerning the revision of NFRD (Venturelli et al., 2022).The analysis of these research topics appears to be particularly relevant at this specific moment when the transition from the regulatory framework of the NFRD to that of the CSRD raises important questions about the role of these tools in favoring more effective sustainability reporting.This is also relevant in view of the future conversion of CSRD at the level of individual EU members and of its future application.In this sense, this work aims to provide a contribution on the level of the effectiveness of the regulatory initiatives already undertaken in the context of the ongoing debate on possible future developments of the same.
The article has the following structure: the second section develops a literature review, with an analysis of the main topics related to the subject of this work, i.e. the relationship between SE and ST, the emerging gaps and controversies in literature about SE quality and the possible role of a regulatory framework in pursuing more effective SE processes.The third section describes the research method, with reference to the methodological approach, the research sample and the elements of evaluation and scoring adopted for the empirical analysis.The fourth section provides the results, and the final section contains the discussion and the conclusions of the work.

Literature review
2.1 Stakeholder engagement in the realm of stakeholder theory SE can be defined as the process used by an organization to involve its stakeholders to understand their expectations and interests and the resulting information needs (GSSB, 2016).In other words, SE is the commitment to embrace the principle of "inclusiveness," which means recognizing the right to be heard by stakeholders and accepting the commitment to report on the choices made for their businesses (Anggraeni et al., 2019;Herremans et al., 2016).

The stakeholder engagement disclosure
Therefore, SE is recognized as a fundamental accountability mechanism, as it obliges an organization to involve stakeholders in identifying, understanding and responding to sustainability issues and concerns and to report, explain and respond to their decisions, actions and performance (Zarzycka et al., 2021).SE, if properly and effectively used, allows us to gain an understanding of the key topics, interests and concerns of stakeholders and to explain how the company has reacted and responded to those aspects (Freeman et al., 2017;Greenwood, 2007;Lopatta et al., 2017).The principle of "inclusiveness," explicitly indicated and described by the GRI standards (GSSB, 2016) and the AccountAbility Principles (AccountAbility, 2018), summarizes the essential concepts of ST (Donaldson and Preston, 1995;Freeman, 1984) and hence of the SE process.
Therefore, SE finds its theoretical foundation in ST, as also noted by one of the authors, who contributed most to the development of this theory (Freeman et al., 2017).Nonetheless, Greenwood (2007) pointed out that the ideas underlying the SE process can also be drawn from other managerialist theories and approaches, such as accountability and responsibility theories (Gray et al., 2014), legitimacy theory (Deegan, 2002) and social control and construction theories (Livesey and Kearins, 2002;Power, 2004).Consistent with this vision, SE has been considered a corporate social responsibility tool (Lopatta et al., 2017) that is useful for reducing the information asymmetries between firms and investors (Garc ıa-S anchez and Noguera-G amez, 2017) and introducing good corporate governance practices (Devinney et al., 2013;Fassin and Van Rossem, 2009;Klein et al., 2019).Moreover, the involvement of stakeholders belonging to a company's corporate governance system, such as the board members or the audit committees (Velte, 2023), in the NFR process appears to give positive results in pursuing sustainability aims aligned with the Sustainable Development Goals (Cosma et al., 2022).
In the SE process, bi-directional communication between the company and its stakeholders is essential (Bellucci et al., 2019;Unerman, 2007) and takes place as a part of a relationship of trust and interdependence.Business managers should be aware of the need for ongoing and proactive SE (Cooper and Owen, 2007;Morsing and Schultz, 2006).Freeman et al. (2010) showed that the involvement of multiple categories of stakeholders, a characteristic feature of ST, is necessary for the processes of value creation because, in this way, the problems deriving from the involvement of only the dominant stakeholders are overcome.In this sense, Boesso and Kumar (2016) examined the association between stakeholder culture, the relevance of stakeholder groups and the SE process and highlighted the importance for managers to shift their attention from corporate reputation to activities that create value for a wide range of stakeholders.

Gaps and controversies about SE quality
Although this topic is attracting growing interest among scholars, there is still a limited amount of academic research on the quality of the SE process.Among the few contributions, some surveys have been conducted through questionnaires presented to company managers to evaluate the SE process (Ayuso et al., 2014;Boesso and Kumar, 2009;Cummings, 2001).Herremans et al. (2016) suggested that it is necessary to deepen knowledge on capabilities and internal processes that drive SE processes in sustainability reporting practices; the resource needs of the various categories of stakeholders lead to the development of various stakeholder relationships and, consequently, adequate resources within the company to implement SE strategies.
Among the authors investigating the quality of SE in the sustainability report, Manetti (2011) conducted content analysis on 174 sustainability reports and obtained evidence of how SE had been used as a means of legitimizing and managing stakeholders (in compliance with the legitimacy theory), rather than as a real tool for involving them (as required by the ST).
Other examples of studies on content analysis are those by Haddock-Fraser (2012), who examined the SE process in the context of news media, and Manetti and Bellucci (2016), who investigated the use of social media in the SE process.In this work, the authors show that there are still few organizations that use social media in SE processes and that, where present, these types of interactions do not bring positive contributions, leading to the emergence of divergent socio-political opinions from a competitive perspective.Furthermore, Pizzi et al. (2021a, b) suggested that a dialogue with stakeholders using social media channels cannot replace an appropriate and complete SE process.
Among the research contributions on the quality of disclosure on the SE process, Moratis and Brandt (2017) examined 55 sustainability reports issued by European companies according to GRI reporting standards.They showed that many companies fail to provide complete disclosure of SE modalities and discussed how the SE process helps define the content of the report.Venturelli et al. (2018) developed a model for assessing the quality of disclosure in the SE process.They proposed a score for assessing the quality of the SE process based on nine variables taken from an examination of international standards and literature on the SE.The score was applied to a sample of 313 European banks through content analysis.The results revealed the low quality of SE processes.However, Civera et al. (2019) showed how actions of empowerment and the involvement of vulnerable and less powerful stakeholders contribute to value creation processes, confirming the ST.Moreover, recent studies have emphasized the central relevance of SE practices for pursuing strategies of innovation management and entrepreneurial development (Leonidou et al., 2020), also within the perspective of value "co-creation" between companies and stakeholders (Loureiro et al., 2020).

Role of a regulatory framework in pursuing more effective SE processes
Academic research on the SE process and its quality has a relevant space worldwide but with a particular rate of interest in the EU context because of the innovation introduced by the NFRD.It has established the compulsory disclosure of information of a non-financial nature for public interest entities that exceed certain dimensional parameters, with the aims of improving the comparability of information and enhancing corporate accountability (La Torre et al., 2018).The first studies on the application of NFRD revealed that it has produced some positive results in terms of sustainability disclosure quality and quantity (Mion and Adaui, 2019;Ottenstein et al., 2022), although some exceptions have been presented (Carungu et al., 2020;Lippai-Makra et al., 2022;Posadas et al., 2023).Moreover, several scholars have claimed the need to introduce more focus on sustainability, more integration between non-financial and financial aspects (Fiandrino et al., 2022), and a new approach to the concept of materiality (Raith, 2023), with the introduction of the "double materiality" perspective (Adams et al., 2021;Baum€ uller and Sopp, 2022).
Furthermore, to consider the observations that emerged both from the academic world and from the realities involved in the NFR processes, the EU developed a debate that led to the recent approval of the CSRD.Therefore, the academic community is questioning what the evolution of sustainability reporting could be in the European context, also in light of this significant evolution of the regulatory framework (Pizzi et al., 2022;Venturelli et al., 2022).
The role of the legal frameworks in pushing to more effective forms of sustainability reporting is a topic of interest not only in the European context (Cupertino et al., 2022;Dumitru et al., 2017;Mion and Adaui, 2019;Venturelli et al., 2019;Zanellato and Tiron-Tudor, 2022) but also in other countries or territorial contexts.With reference to the worldwide situation, Krueger et al. (2023) showed that a compulsory ESG disclosure regulation can contribute to the improvement of corporate information, with general benefits on the capital market, particularly by helping reduce differences and therefore protecting the interests of The stakeholder engagement disclosure investors (Rossignoli et al., 2022).Other studies involving extra-European countries (Syed and Butt, 2017;Tran et al., 2021) confirm that a regulatory framework that seeks to consider the interests of stakeholders, particularly investors, can constitute a positive opportunity for the development of more complete and adequate practices of sustainability reporting (Alshbili and Elamer, 2020;Sadou et al., 2017).
Starting from this analysis of the existing literature, the present work was intended to extend it by investigating the role that a mandatory disclosure of a non-financial nature could play in terms of improving the quality of information on the SE process.This role was assessed according to the interpretative framework of the ST.In addition to assessing the quality of SE processes in the context of corporate reporting activities and in compliance with the principle of inclusiveness, this study aimed to shed light on the effectiveness of NFRD in favoring improvement paths in SE practices (Korca and Costa, 2020), with an aim similar to that of Zarzycka land Krasodomska (2022) about KPIs.
Consequently, the fundamental research question of this work is as follows.
RQ1.What is the level of compliance with the principle of inclusiveness and with the fundamental concepts of ST during the SE process by companies that prepared the NFS according to the Italian application of the NFRD?
Assessing the effective compliance of SE process with ST principles in a mandatory framework provides useful indications in relation to the objectives of this study and, in particular, to evaluate the effectiveness of NFRD in improving SE practices, especially in light of the importance of this topic also for the forthcoming application of the CSRD in Europe.

Research method 3.1 Methodological approach and research sample
In light of this theoretical framework, this work was intended to contribute to the development of the qualitative analysis of the disclosure of the SE process, proposing an evaluation model that provides an examination of specific research variables through content analysis (Krippendorff, 2018).The research variables were identified on the basis of an academic literature review and the disclosure requirements or indications reported in the main international reporting standards applicable, following an approach similar to that of Venturelli et al. (2018).This analysis concerns the disclosure of the SE process and related results by Italian-listed companies obliged to prepare the NFS in compliance with Legislative Decree 254/2016, the Italian legislation that transposed the NFRD.The NFRD identifies the recipients in large companies that are public-interest entities.Through the scoring system and the developed evaluation model, this study aimed to verify effective compliance with the principle of inclusiveness in the sample.This analysis was favored by the application of the scoring system to two years of analysis: 2018 (the second year of application of the NFRD in Italy) and 2021 (the last year available).
The empirical application of the principle of inclusiveness can lead to various methods and levels of engagement (from passive methods to forms of bidirectional and multidirectional dialogue, up to the integration of stakeholders in decision-making, strategic and corporate governance processes), with related effects on the quality of the process itself and on its effectiveness in pursuing the aims of ST.To answer the research question, we should verify compliance with some conditions derived from the analysis of the literature on the SE process.
(1) Condition n. 1 (C1): The SE process should cover all identified stakeholders in order to pay attention to the legitimate interests of all appropriate stakeholder groups; (2) Condition n. 2 (C2): Different approaches to the SE process, in terms of the methods of involvement (including the consideration of aspects, such as timing and frequency of involvement) and the number and type of dialogue channels used, influence the effectiveness of the same and related disclosure; (3) Condition n. 3 (C3): The SE process should identify key topics for stakeholders, and the company should explain how it has responded to their interests and expectations.
A good SE process should identify direct inputs from stakeholders or their legitimate social expectations.
The verification of compliance with these conditions makes it possible to assess the effectiveness of the SE process in the companies examined, in relation to both the objectives of this work, thus obtaining information on the susceptibility of a mandatory regulatory context to improve the SE approach and related disclosure, and to the research question posed, in particular, by verifying compliance with the principle of inclusiveness envisaged by the ST.The methodological approach adopted was qualitative because it was considered suitable for investigating the SE process in the context of NFS, given the elements of novelty and complexity (De Villiers et al., 2019).Furthermore, qualitative research is suitable for investigating the strategic choices of management (Parker, 2018) that underlie reporting decisions.Among the types of qualitative research, content analysis was used to examine the quality of the SE process disclosure within the NFS, referring to the financial years 2018 and 2021 for the sample of companies examined.Content analysis is a technique that lends itself to exploratory empirical investigation aimed at producing valid, reliable and replicable inferences (Krippendorff, 2018).This method has been widely used in previous research in the field of environmental, social and sustainability disclosure investigations (Milne and Adler, 1999;De Villiers et al., 2019).It has also been applied to assess the quality of non-financial information, as it is considered an objective and reproducible method that allows comparisons, both in terms of time and space (Guthrie and Abeysekera, 2006;Guthrie et al., 2004).
In this case, content analysis was performed manually, without the use of software or computerized assistance techniques, considering the need to interpret the information contained in the NFS.Both data and elements of evaluation defined in paragraph 3.2 were cross-checked by the authors and by two different and independent scholars, as described in par.3.2.
The research sample consisted of NFS referring to 2018 and 2021, published by the companies whose shares were listed on the Italian Stock Exchange and included in the FTSE MIB index.The FTSE MIB index measures the performance of the top 40 Italian stocks by size and liquidity and represents the most significant stock index of the Italian Stock Exchange in terms of capitalization, free float and liquidity.
Through this research, we aimed to investigate whether and how the assessments on the quality of the information from the research perspective described above had changed in 2021 reporting compared to that in 2018.The choice of the years 2018 and 2021 was motivated as follows: for almost all the entire sample, 2018 represented the second year of the obligation to prepare the NFS, with the consequent expectation of a better and effective understanding of the reporting requirements introduced by the NFRD and by the Legislative Decree 254/2016, whereas 2021 was the last year available.Therefore, the choice of these years was justified with the intention of understanding the actual evolution of the quality of disclosure by comparing the first real "stable" or comparable year of reporting in the NFS with the last one available.The choice to focus on NFS prepared by the listed companies in Italy was influenced by the fact that NFRD had a very detailed application in Italy through Legislative Decree 254/2016.

The stakeholder engagement disclosure
Starting from the 40 companies listed on the FTSE MIB index, we verified which of them had prepared the NFS.This is because, in some cases, an exclusion from this obligation could be envisaged according to the EU and Italian regulation [1].The final sample consisted of 36 companies for 2018 and of 39 companies for 2021.Furthermore, 34 companies were present for both 2018 and 2021 reporting, thus allowing for specific comparative assessments over time.See Table 1 for the list of companies and related industries whose NFS have been examined.

Elements of evaluation and scoring
To assess the quality of the disclosure, we selected eight variables on the basis of the literature review and the indications contained in the main international reporting standards (GRI Standards, AccountAbility and IR Framework) on the quality of the SE process.For each examined NFS, a score was assigned, which increased to the extent that the disclosure provided for each variable more closely reflected the requirements envisaged for each specific element being assessed.
Different scoring schemes have been used because the nature of the investigated variables is different.In particular, where the variable to be investigated did not allow elements of gradation (two out of eight variables), we used a binary "yes" or "no" system.Gradation was performed by introducing elements of evaluation that were reasonably measurable or assessable as much as possible and related to the nature and type of the variable.The scoring scheme differences were standardized and resolved in the calculation of the final evaluation index, as specified below.
See Table 2 for the variables investigated, mainly expressed in the form of questions and the scoring scheme outlined.
Variable V1 reflects a necessary operation for the ST.The need to identify and describe stakeholders is expressly provided for in the three main reporting references considered (GRI, IR Framework, AccountAbility).In particular, GRI disclosure 102-40 (GSSB, 2016) requires an organization to include the list of stakeholder groups involved in the report.
Variable V2 relates to the description of the process followed by the company to intercept and define the stakeholder groups with which to relate consistently with the ST.Both the GRI Standards and the AccountAbility Principles require this to be described in the report.In particular, GRI disclosure 102-42 requires an organization to disclose the reasoning followed to identify its stakeholder groups and determine which ones to involve and which ones not to involve.In the evaluation of this disclosure, a score of zero was assumed in the event of the absence of this information in the NFS, and an increasing score from 1 to 3 was used in relation to both the width and depth of the disclosure provided and the underlying logical reasoning.In particular, the maximum score was assigned in the presence of extensive and complete information, which also highlighted the use of innovative analysis and selection methodologies (e.g.databases to periodically map stakeholders, use of big data or of digital platforms).Note that for this variable, the subjectivity of the assessment can influence the assignment of the score, especially in some cases in the attribution of scores that range between the values 1 and 2.
Variable V3 was inspired by Manetti (2011), who, among the research elements relating to the degree of stakeholder representation, proposed verifying whether all the groups of stakeholders identified had been involved.This aspect concerns the principle of inclusiveness of the ST and, in particular, the broader interpretation of the same illustrated by authors such as Clarkson (1995) and Garriga and Mel e (2004).This contrasts with the more selective approach of other scholars, typically the identification and salience theory proposed by Mitchell et al. (1997).Evidently companies are not obliged to involve all the stakeholders identified to comply with the reporting standards; however, we preferred to assign a reward score for those who did so in adherence with the broader meaning of the principle of inclusiveness.
Variable V4 was intended to investigate the quality of the information provided in NFS about the approach to SE, including the description of specific aspects, such as (1) methods of engagement, (2) frequency of engagement by type and stakeholder group, (3) timing (Manetti, 2011) and (4) disclosure about the existence of a specific process.The evaluation of such aspects is useful for understanding how the SE process took place.The need to provide such disclosure is referenced in the reporting standards GRI 102-43 and AA1000SES.Furthermore, we refer to the academic contributions of Manetti (2011), Moratis and Brandt (2017).

JAAR
With regard to the scoring scale, in addition to establishing a score of zero in the event of absence of information, the grade from 1 to 3 varies according to the completeness of the disclosure with reference to the elements of analysis listed above (a-b-c-d).In particular, a score of 3 was assigned only in the presence of full information on everything, whereas the progression between scores 1 and 2 varied depending on whether the information covered only one aspect of those indicated (score 1) or more than one aspect (score 2), giving greater weight to the information referred to in letters b) and c) above, which, from the first screening on the NFS observed, appeared more fluctuating as a presence in the disclosure.It follows that in the presence of adequate information on the frequency of involvement and timing of the SE process, the minimum score assigned was equal to 2.
Variable V5 focuses on the dialogue channels and the levels of involvement chosen by companies in their SE process.Organizations can choose various ways and levels of SE, ranging from more passive methods (e.g.monitoring information or complaints) to gradually increasing forms of dialogue (bidirectional or multidirectional dialogue) up to the integration of stakeholders into governance (Cooper and Owen, 2007) in decision-making and strategic processes and in company operations (AccountAbility, 2015).Given its importance for the purposes of the principle of inclusiveness, this aspect is the specific subject of disclosure by the AccountAbility Principles (AA1000SES) and GRI standards (disclosure 102-43) and has also been investigated by academic researchers (Venturelli et al., 2018;Manetti, 2011).
Variable V6 intends to address the quality of the disclosure on key topics, expectations, critical issues and concerns raised by stakeholders and on the company's answers or followup to them.All major reporting standards require the company to describe the results of the SE process; however, as highlighted by Moratis and Brandt (2017), this disclosure frequently appears to be lacking in practice.In the assignment of the score, a value of zero was established in the event of absence of information, whereas the progression from 1 to 3 varied according to the completeness and depth of the information reported, especially with reference to the responses by the company to the most relevant issues and interests highlighted by stakeholders.This aspect was the most deficient, as emerged from the article by Moratis and Brandt (2017).In assessing the quality of the information, it was also verified whether the NFS mentioned the direct inputs received from stakeholders.The scoring scheme for variable V6 was as follows: no information 5 0; little information, absence, or scarcity of disclosure on the responses by the company 5 1; appropriate information 5 2; comprehensive information regarding the answers from the company and evidence of direct inputs from stakeholders 5 3.
For the attribution of the maximum score, we verified that the information reported in the NFS really brought out the key topics, concerns and issues of interest to the stakeholders.However, this variable lends itself to margins of subjectivity in the evaluation, especially in the extreme case of assigning a score oscillating between the values 1 and 2.
Variable V7 relates to any disclosure of conflicting views or expectations among stakeholders.In such cases, a company is expected to be able to explain how it balances them when making decisions about its reporting (GRI 101).Note that this disclosure is not mandatory for companies.Nonetheless, it appears particularly interesting, both as it fully expresses the principle of inclusiveness and because it provides indications of the quality and consistency of the SE process.For these reasons, it was considered useful to evaluate this information element by assigning a reward score to this disclosure.
Finally, variable V8 was intended to verify whether the NFS gives evidence of the integration of the SE process into the governance, strategies and operations of the company, an issue considered important by the AccountAbility Principles and fully in line with the ST.With reference to scoring, in addition to establishing a score of zero in the event of the absence of information, the gradation from 1 to 3 varies according to the completeness of the disclosure with reference to the three elements of integration (governance, strategies and The stakeholder engagement disclosure company operations).In particular, a score of 3 was assigned only in the presence of exhaustive disclosure showing the integration of all three of the aforementioned elements, whereas the progression between the scores 1 and 2 varied depending on whether the information covered only one aspect than those indicated, or more than one aspect.An evaluation schedule was compiled for each of the NFS examined.After assigning a score to each variable, we proceeded to correlate the score for each variable [var (i)] to the highest value attributed [max var (i)], multiplying the result obtained by 100.The following is the explanatory formula of the calculations performed: The score can range from 0 to 100, with the latter value representing the highest score if each evaluation element reaches its maximum relative value.
To verify the degree of reliability of the scores obtained and of the underlying evaluations of the variables described above, two scholars independently repeated the content analysis.The control protocol adopted was as follows: the independent researchers carried out the analysis separately based on the scoring scheme reported above and of the coding indications commented on the previous pages.The reliability test was performed on 10 NFSs, the same for both observers and the NFFs were selected to capture the extreme and median values resulting from the final score.In particular, the following NFSs were selected: the first three classified, the last three classified, the three with values equal to the median and one with a random criterion.The reliability control protocol was performed according to Krippendorff's indications.The Krippendorff-α (nominal) reliability coefficient obtained was 0.831 for the 2018 NFS and 0.840 for the 2021 NFS; such results can be considered positive, being higher than the suggested reliability threshold of 0.800 (Krippendorff, 2018).

Results
Following the evaluations carried out according to the research methodology described above, the following scores were obtained, (see Table 3), with related descriptive statistics (see Table 4).
The evaluation scores obtained show an overall quality to be considered barely sufficient, both in 2021 and 2018, with regard to the disclosure of the SE process by the top Italian-listed companies.
The mean and median values do not show appreciable variations in the comparison between the analysis carried out for 2021 and that for 2018.There was a slight improvement in the maximum score and a more consistent one in the observed minimum value.We noted an improvement in terms of less variance of scores and related standard deviation, thus suggesting a slight qualitative realignment of the NFSs.Therefore, the descriptive statistics do not show a sounding overall qualitative improvement of the SE disclosure in the comparison between 2021 and 2018, if not a lower general variability and qualitative difference between one NFS and another.
The final scores assigned to the companies for which it was possible to make a comparison between 2021 and 2018 (i.e. for 34 companies and related annual NFSs) show an improvement in the score for 17 NFSs, a worsening for 16 NFSs and an equal result for 1 NFS.This confirms that there has not been a significant overall improvement in the quality of the information.
In the analysis of the eight research variables, the results and related descriptive statistics are summarized, respectively, in Table 5 and Table 6.Regarding variable V1 (stakeholder identification), the examination of the NFS showed that almost all the companies in the sample provided information on the stakeholders involved in the SE process.Both in 2018 and in 2021, there was only one case in which this disclosure was not identified.
For variable V2 (stakeholder selection process), the research highlights some variability in the results obtained; however, there was a qualitative improvement in the comparison between 2018 and 2021, as evidenced by the increase in cases with an "Enough" or "A lot" rating.In 2021, the NFS that obtained a positive evaluation was 74% (29 out of 39), which was only 50% (18 out of 36) in 2018.This shows an improvement in the disclosure provided in the process followed by the company to identify its stakeholder groups and determine which ones to involve and which do not involve.Such improvement seems related to a greater internalization of SE practices following the introduction of the NFS reporting obligation, which appears to contribute to reducing the degree of information asymmetry and, thus, mitigating agency problems, as highlighted by Garc ıa-S anchez and Noguera-G amez (2017).With reference to variable V3 (degree of involvement), the verification of whether all the stakeholder groups identified were involved gave an overall positive result (75% for 2018 and 95% for 2021).Such results are fairly in line with the evidence collected by Manetti (2011), who found an engagement rate of all identified stakeholders equal to 86%, albeit referring to a different sample of analysis.In general, the empirical evidence that emerged was considered positive, considering the broader meaning of the inclusiveness principle of the ST, theorized by scholars such as Clarkson (1995) and Garriga and Mel e (2004).This is especially because there is no specific obligation in this regard in the 2016 GRI standards, which are the reporting documents applied by all the companies examined.
With regard to variable V4 (the SE approach), the results showed a wide variability of scores between one NFS and another.However, in the comparison between 2018 and 2021, there was some general improvement in the quality of the information provided in the NFS regarding the approach to SE.In both years, we noticed some gaps in the indications of the execution times (timing) of the SE process.This lack confirms the low quality of disclosure on frequency and timing of SE observed by Moratis and Brandt (2017).
In variable V5 (dialogue channels), the overall quality of the disclosure was good, both in 2018 (mean value of 82.41) and 2021 (mean value of 85.47).The results obtained were better than those shown by Venturelli et al. (2018) and more in line with the evidence gathered by Moratis and Brandt (2017).Nevertheless, it should be considered that there was a certain variability in the score from one NFS to another.
All the companies examined held meetings with their stakeholders.Moreover, among the most frequently used dialogue channels were the administration of a survey (86% in 2018, 97% in 2021), the creation of focus groups (64% in 2018, 77% in 2021) and the organization of workshops (44% in 2018, 51% in 2021).All the engagement channels mentioned represent an intermediate form of involvement, classified as a "Consult" level by AccountAbility (2015).Such a kind of limited two-way engagement was also observed by Bellucci et al. (2019), who noted that fully developed frameworks for dialogic accounting are rare.Therefore, this research confirmed observations in prior literature findings on such elements of SE and highlighted the need for the top Italian-listed companies to implement methods of engagement with their stakeholders that provide for greater collaboration and empowerment, especially through the integration of stakeholders into governance.
With regard to variable V6 (SE results), the examination of the descriptions given in the NFS regarding the results of SE activities and particularly the analysis of the disclosure on key topics, interests, concerns, crucial issues raised by stakeholders and the company's responses to those aspects was overall unsatisfactory for both years and was characterized by a high variability.Adding the assessments with answers "No" and "Little," we obtained a poor quality of disclosure for 56% of the sample analyzed in 2021 and 46% in 2021.The improvement observed in 2021 reporting was still insufficient, given the importance of such disclosure and the research sample made up of the main listed companies in Italy.
The evidence collected on such a research variable fully confirmed the criticality identified by Moratis and Brandt (2017) regarding the lack of information provided in the report with reference to the company's responses to the most relevant issues, concerns and interests highlighted by stakeholders.This is despite the disclosure 102-44 of the GRI standards expressly requiring it, as well as the AccountAbility Principles and the IR Framework.This is undoubtedly a deficiency that limits the principle of inclusiveness and reduces the quality of the disclosure, casting the shadow of reducing the NFS to mere storytelling on such an important aspect, where it is required to provide a clear and thorough explanation of the company's responses to stakeholders' needs and expectations.Moreover, it is interesting to observe that although the analysis of the results of variable V5 showed an improvement in The stakeholder engagement disclosure the dialogue channels with stakeholders, as described in the NFS from 2018 to 2021, this aspect was not reflected in a tangible improvement in the overall quality of the disclosure of the results of the SE process (variable V6), placing some doubts on the real effectiveness of the SE and dialogue methods implemented.
For variable V7 (different points of view), the results showed an almost total absence of this disclosure in the research sample: as many as 92% of the sample in 2018 and 90% in 2021 did not report any information on different or conflicting points of view between stakeholders.Although this disclosure is not mandatory, and the existence of discordant or conflicting points of view between the stakeholders can be, abstractly, only possible, it is quite natural that different groups of stakeholders can express different, if not conflicting, needs and expectations.This is assuming that an entity wants to intercept and monitor this aspect, which is certainly less easy to grasp, by implementing specific activities or procedures as part of the SE process.Therefore, the verification of such disclosure included in GRI 101-1.1 (GSSB, 2016), although not mandatory, provides useful information on the depth of the SE process and on the full understanding of the principle of inclusiveness and the foundations of the ST by companies (as theorized by Freeman, 1984;Donaldson and Preston, 1995).Hence, these aspects should be improved in light of the results of the search variable V7. by the assessment of variables V4 and V5, showed how an appropriate engagement approach and different methods of dialogue and levels of stakeholder involvement influence the effectiveness of the process, with consequent effects on the information reported in the NFS.However, we observed wide variability in the quality of the disclosure on the engagement approach.Among the aspects to be improved was the disclosure of the possible integration of stakeholder groups in governance, strategic processes, or company operations.
However, the empirical investigation did not confirm the expectations regarding C3.The SE process should identify the key topics, expectations and interests of stakeholders to be answered, whereas the results obtained clearly highlight a significant gap with regard to this important disclosure.This showed how this aspect constitutes the main criticality observed.In this sense, the research results confirmed what was observed by Venturelli et al. (2018) and by Moratis and Brandt (2017).It is also important to underline that this criticality was not significantly reduced in the comparative analysis of the variable V6 between the 2018 and 2021 NFS, a sign of a certain inability of the top Italian-listed companies to address one of the key aspects of the principle of inclusiveness.Reading the results of variable V6 with those of variable V7, we infer a scarce depth of the SE process, especially in reaching the core of the principle of inclusiveness, according to the ST.

Discussion and conclusions
The research question was intended to ascertain whether the companies examined had respected the principle of inclusiveness and the fundamental concepts of the ST during the SE process, aimed at preparing the NFS pursuant to Legislative Decree n. 254/2016, which implemented the NFRD in Italy.
In the analysis of the results, the empirical investigation did not show a convincing positive response to the research question, which also confirmed for 2021 what emerged from the examination of the NFS prepared in 2018 (i.e. at the beginning of the obligation to prepare such reports by Italian companies).The results showed qualitative differences between the various reports, even if this aspect improved over time.Above all, the expectation that disclosure of the SE process would identify and describe stakeholders' key topics, expectations and interests was not verified for many of the companies in the research sample.Moreover, disclosure of the answers to such issues was poor.
From a theoretical implication point of view, such results highlight how this aspect constitutes the main criticality observed regarding compliance with the principle of inclusiveness by companies, confirming in some ways what was found in the literature.In addition, the empirical part shows a lack of attention by the top Italian-listed companies in intercepting different points of view among stakeholders and in providing an explanation of their balancing and management.
These criticalities show a scarce depth of execution of the SE process, inducing us to also hypothesize its use as a legitimacy tool by companies and/or its managers (i.e. the fulfillment of a sort of disclosure "obligation" envisaged by the reference standards or a corporate social responsibility practice deemed necessary, rather than as a real activity of involvement of stakeholders, aimed at effectively implementing the principle of inclusiveness of the ST).Therefore, this research highlights the need for a deeper and more comprehensive involvement of stakeholders in preparing non-financial reports.
Consistently with these emerging theoretical issues, this work adds its contribution to the debate on the role of mandatory frameworks for NFR, with specific reference to the SE initiatives.In particular, the emerging picture leads to state that the mandatory framework did not have a very significant effect on the improvement of SE activities reporting.At the same time, the research showed a reduction in the differences, from 2018 to 2021, in the reporting of SE practices, possibly due to mimetic mechanisms resulting from the increasing diffusion of NFR.
This research also provides some positive aspects that are useful for theoretical developments.First, the appreciable results in terms of the involvement of all groups of identified stakeholders reveal how this qualification of the principle of inclusiveness appears to be largely assimilated by the Italian-listed entities of the sample.Second, this research confirms the importance of dialogue channels and levels of engagement used, which both significantly influence the effectiveness of the process itself, with consequent impacts on the quality of the information reported in the NFS.Another possible theoretical development of this work is that it provides an assessment framework for the quality of SE in the NFR.This framework, based on the literature review and the most common international sustainability reporting standards, appears to be useful for other analyses applied to other geographical contexts or to different time basis.
From a practical implication point of view, the work indicates some possible ways of improvement and enrichment for the development of SE initiatives by companies that must undertake paths of SE.The analysis of the proposed variables for the definition of the quality of an SE process represents a possible benchmark for preparers of sustainability reports in considering SE initiatives.The article also provides some suggestions on the most recurring challenges and the most common difficulties in implementing SE activities.

The stakeholder engagement disclosure
From a point of view of the policy and societal implications, this work contributes to enriching the debate on the role of mandatory frameworks for the development of more effective sustainability reporting practices, with specific reference to the quality of SE processes behind them.The results show that the presence of a regulatory framework leads to internalization by operators of the importance or necessity-even if sometimes only for the purposes of regulatory compliance-of SE practices.However, this framework is not sufficient to contribute to the development of quality SE processes, for which it appears necessary to make greater efforts to disseminate best practices and the positive effects they can have, not only on SE processes, but also on the overall performance of the companies.In this sense, in the context of the future application of the CSRD, as an evolution of the NFRD, a more relevant role should be reserved, not only for the methods of implementation of the SE processes, but also for how the various stakeholders' categories can exert the right pressure on the companies to be properly involved.
With reference to its possible future developments, this research could hopefully represent a starting point for further investigations aimed both at extending the investigation to all NFSs published in Italy by examining any additional or different examination profiles and at checking the quality of disclosure in other countries that have implemented the NFRD and will implement the CSRD in the next years.Furthermore, among the possible future perspectives of this research, there is insight, through surveys or interviews with prepares, on some observed behaviors that were found to be discordant with respect to the theoretical framework of ST.Examples are the deficiencies found in the identification of the key topics for stakeholders and related responses to those issues and the disclosure of conflicting points of view among stakeholders.
Finally, the main limitations of this work are as follows.
(1) The methodology (content analysis) intrinsically presents margins of subjectivity (Guthrie and Abeysekera, 2006).Some evaluation elements, particularly variables V2, V4, V6 and V8, are more suitable for different choices or scores, according to the sensitivity and background of the researcher, with consequent difficulties in validating the results achieved.
(2) The sample analyzed was limited to a subset, albeit significant and important, and substantially homogeneous from a dimensional point of view of the Italian public interest entities required to draw up the NFS, with the effect that the conclusions reached may not be representative and extendable to the entire population of the NFS prepared in the years 2018 and 2021.Furthermore, the research concerned only one country, Italy.Therefore, the results obtained could be country specific.

Table 3 .
Final scores