Does the Directive of the European Union on disclosure of non-financial and diversity information improve reporting?

1. Introduction

Over recent decades, the importance of reporting that is not limited to financial performance but also considers companies’ ecological and social performance has risen sharply among all types of companies, particularly among large listed ones (KPMG, 2022). While non-listed companies often report their non-financial performance on a voluntary basis, in many countries large listed companies (LC) are now subject to stringent regulations that govern the non-financial reporting (NFR) requirements (GRI and University of Stellenbosch Business School, 2020).

Several years ago, the European Commission announced the need to raise the transparency of non-financial aspects to the same level in all member states to achieve harmonization of NFR. To facilitate this, Directive 2014/95/EU (ED) was passed in 2014. The directive requires large companies of public interest, including listed ones, to publish non-financial information on different topics. The member states should have transposed the directive into national law by 6 December 2016 (Directive 2014/95/EU).

The question of whether an obligation regarding NFR leads to an improvement in reporting has been raised several times in academic discourse. Several studies have investigated this question in different institutional settings. However, the research findings do not suggest a clear direction, as prior empirical studies show conflicting results (Criado-Jiménez et al., 2008; Asmaranti and Lindrianasari, 2014; Chelli et al., 2014; Fatima et al., 2015; Arrive and Feng, 2018; Issarawornrawanich and Wuttichindanon, 2019).

Moreover, several studies have already examined the effects of ED (Caputo et al., 2019; Mion and Loza Adaui, 2019; Nicolò et al., 2020; Agostini et al., 2021; Korca et al., 2021; Matuszak and Rózanska, 2021; Posadas and Tarquinio, 2021; Carmo and Ribeiro, 2022; Lippai-Makra et al., 2022). However, all such studies analyze the effects of the directive only in selected countries or sectors, or only consider a specific aspect of the directive. Therefore, there is a need to investigate the effect of the directive on all the information required by ED across the European Union (EU) and different sectors. Against this background, our study pursues the following research question:

To answer this research question, we conducted a content analysis of the published non-financial information of the EURO STOXX 50 companies on two different dates – 2016 and 2018, before and after the directive came into force. This approach allows us to examine the impact of the directive in close temporal proximity to its entry into force. Because we aim to evaluate the impact of the directive on reporting, the information required by the directive was the starting point for the development of an evaluation tool for our investigation.

The paper is structured as follows: Section 2 Background and hypotheses development reviews the literature, suggests a theoretical perspective and uses it as a basis for deriving the hypotheses for our study. Section 3 Methodology and sample describes the research approach and the sample in detail. Section 4 Results presents the findings of our analysis. Finally, Section 5 Discussion and conclusion discusses key insights, draws a conclusion, explains the main limitations of the study and shows the need for further research.

Our investigation contributes to the debate on the effects of mandatory NFR in several ways. First, it analyzes the effects of ED on companies from different EU countries and sectors in its entirety and in a longitudinal study design. As far as we know, such an investigation has not yet been carried out in this field and thus provides novel evidence. Existing studies in this context have either examined only selected countries (Mion and Loza Adaui, 2019; Caputo et al., 2019; Agostini et al., 2021; Korca et al., 2021; Matuszak and Rózanska, 2021; Posadas and Tarquinio, 2021; Carmo and Ribeiro, 2022; Lippai-Makra et al., 2022), sectors or industries (Nicolò et al., 2020; Korca et al., 2021) or specific areas of the directive (Agostini et al., 2021) and thus do not allow for a conclusive answer. Furthermore, it uses the Global Reporting Initiative (GRI) framework, the most widely used NFR framework worldwide (KPMG, 2022), as a basis for evaluating the improvement by assigning the individual GRI indicators to the information required by the directive. Showing how the various GRI indicators can be assigned to the information required enables report readers to obtain an overview of the company’s compliance with the requirements of ED in a reasonable amount of time, by checking which indicators are reported. Finally, from a theoretical perspective, it suggests an institutional view that allows for a better understanding of the changes discovered here.

Overall, our study provides novel findings on the effects of ED that are not only relevant for academic discourse but can also serve as a basis for political and managerial decision-making, especially against the backdrop of the tightening of reporting obligations by the Corporate Sustainability Reporting Directive (Directive (EU) 2022/2464) which not only significantly expands the group of affected companies, but also expands or specifies the required reporting content.

2. Background and hypotheses development

3. Methodology and sample

4. Results

In the category of business model, substantial differences were identified. While indicator GRI 102–2 or G4-4 – activities, brands, products and services – is disclosed by almost 90% of the companies in both reporting years, indicator GRI 202–1 or G4-EC5 – ratios of standard entry-level wage by gender compared to local minimum wage – is reported by well under a quarter with a downward trend. In general, the comparison of the two years shows similar reporting patterns. The CL of the category rises from 60.6% to 62.0%. The greatest differences are found in indicators GRI 102–4 or G4-6, which require information about countries where the organizations operate, and GRI 102–8 or G4-10, which demand information on employees and other workers, with a increase of over 8 percentage points (pp) in each case (see Figure 1) (GRI, 2016, 2017).

In the category policies and due diligence, the variation between the indicators is comparatively lower. The best-rated indicator or management approach in the second year under review is GRI 103–1 or G4-21 regarding the environment – explanation of the material topic and its boundary (85.4%), whereas GRI 103–1 concerning anti-corruption and bribery scored the fewest points in both years (32.3%, 2016 and 56.3%, 2018). The category CL increased from 57.3% to 70.8%, the largest increase in the category comparison of the entire investigation. At the individual indicator level, the greatest differences with a change of more than 20 pp are shown for GRI 103–2 regarding human rights and GRI 103–1 concerning anti-corruption and bribery (see Figure 1) (GRI, 2016, 2017).

The variation of the category principal risk is at a similar level as the category policies and due diligence. The most extensively reported indicator is GRI 103–3 or G4-DMA-c concerning environment – evaluation of the management approach –, which is rated with almost two-thirds in both years. However, the least reported indicators are GRI 103–3 regarding anti-corruption and bribery in 2016 (9.4%) and GRI 308–1 or G4-EN32 – new suppliers that were screened using environmental criteria – in 2018 (19.8%). The CL of the category rises from 31.3% to 37.3%, with the largest change by far in indicator GRI 103–3 concerning anti-corruption and bribery and human rights (each 17.7 pp) (see Figure 1) (GRI, 2016, 2017).

The variation in the environment category is greater than in the two categories described above. In the year after the directive entered into force, indicator GRI 305–2 or G4-EN16, which requires information about the energy indirect GHG emissions, is rated with almost 80%, whereas indicator GRI 306–4, which demands information on the transport of hazardous waste, is the least considered in both years (15.6%, 2016 and 13.5%, 2018). A comparison of the two years of investigation shows that the CL of the category decreases from 42.8% to 42.3%. Within the entire investigation, this category is the only one with a decreasing CL, albeit only marginally. The largest difference with almost 20 pp is shown in indicator GRI 305–3 or G4- EN17, which demands information on the other indirect GHG emissions (see Figure 1) (GRI, 2016, 2017).

In the category social, the variation is lower than in the other categories. Indicator GRI 413–1 or G4-SO1 – operations with local community engagement, impact assessments and development programs – is the best rated in both years under review (27.1%, 2016 and 30.2%, 2018), whereas indicator GRI 417–3 or G4-PR7 – incidents of non-compliance concerning marketing communications – is the least considered in the second year (15.6%). The comparison of the CL for the two years shows only a marginal change from 20.5% to 20.8%, with indicator GRI 413–2 or G4-SO2 – operations with significant actual and potential negative impacts on local communities – showing the greatest change with over 6 pp (see Figure 2) (GRI, 2016, 2017).

Compared to the other categories, variation in the employee matters category is higher. The most extensively published indicator is GRI 404–2 or G4-LA10, which demands information on the programs for upgrading employee skills and transition assistance programs (69.8%, 2016 and 71.9%, 2018). In contrast, GRI 401–2 or G4-LA2, which require information regarding the benefits provided to full-time employees that are not provided to temporary or part-time employees, is the least considered (7.3%). The CL of the category only rises slightly from 33.2% to 34.8%. The greatest differences can be identified with more than 8 pp – one positive and one negative – for indicator GRI 401–2 and indicator GRI 407–1 or G4-HR4 – operations and suppliers – in which the right to freedom of association and collective bargaining may be at risk (see Figure 2) (GRI, 2016, 2017).

In the human rights category, the variation is comparatively lower. In both years of the study, the best-rated indicator is GRI 412–2 or G4-HR2, which demands information on employee training on human rights policies or procedures (19.8%, 2016 and 26.0%, 2018). In the second year of the investigation, information on indicator GRI 410–1 or G4-HR7 – security personnel trained in human rights policies or procedures – is the least considered. The CL increased from 13.7% to 17.1%. The greatest difference with over 7 pp is shown in indicator GRI 412–1 or G4- HR9 – operations that have been subject to human rights reviews or impact assessments (see Figure 2) (GRI, 2016, 2017).

Variation in the anti-corruption and bribery category is also at a low level. The indicator GRI 205–2 or G4-SO3 – communication and training about anti-corruption policies and procedures – is rated with almost a third, whereas indicator GRI 205–3 or G4-SO5 – confirmed incidents of corruption and actions taken – is considered roughly a quarter of the times and thus has the lowest value within the category. The CL remained at the same level (26.6%). The greatest difference with a decrease of over 7 pp is shown for indicator GRI 415–1 or G4-SO6 – political contributions (see Figure 2) (GRI, 2016, 2017).

In the category of non-financial key performance indicators, the variation is comparatively higher. The best-rated indicators relate to environmental issues (65.6%, 2016 and 74.0%, 2018). Human rights indicators are least likely to be considered in the second year of the investigation (19.8%). The CL of the category rises from 38.8% to 45.6%, with the greatest differences by far, each with over 8 pp, shown in the indicators relating to environmental as well as anti-corruption and bribery issues (see Figure 3).

Figure 3 provides an overview of the results for the individual categories in comparison and illustrates the large differences between categories. In addition, it shows that in comparison of the two reporting years, the greatest change with almost 14 pp occurs in the policies and due diligence category. It also gives an overview of changes at the company level and shows that in some cases considerable differences can be identified between the TCL of the two years under observation (see Figure 3).

To assess whether the overall changes obtained are statistically significant (H1), t-tests were performed. It was shown that the difference in reporting considering all indicators between 2016 (Mean = 39.2%, SD = 21.605) and 2018 (Mean = 42.7%, SD 23.404) is statistically significant (t(80) = −4.587, p < 0.001), where the effect size suggests a medium to strong effect (r = 0.456) (Cohen, 1988; Gignac and Szodorai, 2016). The difference in the TCL, i.e. at company level, between 2016 (Mean = 37.4%, SD = 16.142) and 2018 (Mean = 41.2%, SD = 15.152) also proved to be statistically significant (t(47) = −2.937, p = 0.005). The effect size again indicates a medium to strong effect (r = 0.394).

To evaluate in which categories the changes in reporting behavior are significant (H2), t-tests were performed for the individual categories at company level. This showed that the difference between reporting in 2016 and 2018 in the categories of policies and due diligence, principal risk and non-financial key performance indicators can be classified as significant, with the effect size indicating a medium to strong effect in each case (Table 2). In the other categories, the changes could not be evaluated as significant (Table 3).

5. Discussion and conclusion

The aim of our investigation was to determine whether ED improved the reporting of non-financial and diversity information. Our study revealed a statistically significant improvement in reporting, which confirms not only our hypothesis H1, but also the first empirical findings regarding the effect of ED on particular EU member states, sectors or specific disclosure requirements and the trend toward improvement in disclosure identified therein (Caputo et al., 2019; Mion and Loza Adaui, 2019; Agostini et al., 2021; Korca et al., 2021; Matuszak and Rózanska, 2021; Posadas and Tarquinio, 2021; Carmo and Ribeiro, 2022; Lippai-Makra et al., 2022).

However, in line with our hypothesis H2, the improvement is not evident in all information categories. While there is a significant improvement in the categories policies and due diligence, principal risk and non-financial key performance indicators, this is not the case in the other categories– i.e. business model, environment, social, employee matters, human rights and anti-corruption and bribery.

Institutional theory suggests that the coercive pressure triggered by the directive’s requirements leads to an adjustment in reporting behavior, as noncompliance entails legitimacy problems and sanctions. Thus, from a theoretical perspective, the overall improvement in reporting that we revealed is expected. That coercive pressure is not exerted by all information categories of ED, however, was also to be assumed. On the one hand, the information requirements in some categories are not a novelty for companies, as they already have many years of experience in some subject areas; on the other hand, the pressure to disclose may not even be felt, as companies already publish the information voluntarily anyway.

However, that coercive pressure occurs in the categories of policies and due diligence, principal risk and non-financial key performance indicators leading to a change in reporting patterns seems obvious upon closer examination of the information required in these categories. The first two categories are information on the strategic concepts or management approaches and the risk management process with the inclusion or differentiated presentation of environmental, social, employee, human rights and anti-corruption and bribery issues. Reporting this information causes considerably more effort than, for example, simply disclosing actual performance in a particular area. Strategies or concepts for the required areas must be developed and implemented, which also entails an adaptation of existing corporate structures and processes. In addition, by disclosing this information, the company provides insights that enable an external assessment of the approach of the management to these issues, including issues in categories such as human rights or anti-corruption and bribery that are considered sensitive. When reporting was voluntary, companies could choose whether to disclose this information as well. Even if they followed the GRI guidelines, there was basically no obligation to disclose. Given the effort involved and the sensitivity of the information, it might have been opportune from a strategic perspective for certain companies not to report, at least in the short term.

The significant improvement in these categories immediately after the directive came into force can be seen as evidence that coercive pressure triggered by the directive’s requirements had an effect and led to an adaptation of reporting. Companies have suddenly started reporting on GRI indicators that address these requirements, even though they were already part of the GRI guidelines before the directive was implemented. Considering that NFR is a well-institutionalized practice among EURO STOXX 50 companies, especially in line with the GRI guidelines, it is unlikely that there will be a significant change in reporting behavior immediately after the adoption of the directive, which should have nothing to do with it. At the descriptive level, however, the improvements in the two categories of policies and due diligence and principal risk, are not found equally in all areas, but the topics of human rights and anti-corruption and bribery show the greatest improvements. This indicates that reporting obligations that also standardize the content or the topics to be disclosed can lead to information being reported that would not be disclosed by companies with reporting experience in a voluntary setting. This is even though these companies would have particularly high incentives for comprehensive reporting due to their capital market orientation.

In this context, however, it must be noted that the starting levels already varied greatly between the individual categories and issues before the introduction of ED. For example, while the information required in the category business model was reported particularly extensively before the directive entered into force, the reporting level was considerably lower in other categories, such as social or anti-corruption and bribery. It must also be noted that information requirements or availability of necessary information can vary significantly. In the business model category, much of the information required can be easily obtained or is already available in any case, whereas the information for other categories, such as human rights, is often not yet available or at least not as comprehensively.

The third category where significant improvement was found – non-financial key performance indicators – is novel in that it requires company-specific indicators that are not included in the common reporting standards or guidelines used by companies. Companies that want to report are looking for professional support in the preparation of their reports and therefore resort to established guidelines, such as those of the GRI. Content that is not covered by a reporting framework, such as the category non-financial key performance indicators, represents a novelty at the beginning and therefore likewise exerts coercive pressure. However, the identification and disclosure of these indicators means that company- or industry-specific conditions are considered and can thus be conducive to the principle of materiality in reporting.

In the theoretical construct of institutionalism, however, homogenization is a dynamic process. Hence, it can be assumed that further changes in reporting behavior will occur over time. In contrast to the initial changes or reactions to the directive, these are often characterized by normative and mimetic processes (DiMaggio and Powell, 1983; Shabana et al., 2017). This theoretical suggestion could help explain why we did not find significant differences in the other categories. According to the phase model of isomorphism in reporting, as proposed by Shabana et al. (2017), the initial phase, which is characterized by a dominance of coercive mechanisms, is followed by a normative and finally a mimetic one. In the coercive phase, companies try to maintain their legitimacy by adapting or fulfilling the requirements of the directive or a well-established reporting standard to avoid potentially negative public perception at a time of increased awareness. Subsequently, reporting is driven forward through professionalization and networks before imitation within a field occurs.

From this point of view, it could be concluded that the directive does not exert coercive pressure in the other categories, for example in the environmental category. It is not uncommon for many companies to have a long tradition in reporting on environmental topics. LC in particular have had a long tradition in environmental reporting. Hence, reporting in the environmental category may be driven by other mechanisms. There is therefore an alignment of the scope or the intensity of disclosure within a field, or, in other words, an informal standard of how companies report on environmental issues has emerged. This explanation can also be applied to the other categories in which we did not find significant improvements, especially considering that the pressure to disclose non-financial information has existed for a long time among these LC.

In conclusion, the results of our study clearly show that obligations can lead to an improvement in reporting. As a result, information that provides insights and would likely not be available without obligations is disclosed. This is even though the companies studied here have had many years of experience in NFR and the application of the GRI guidelines. In addition, the incentives for capital market-oriented companies for comprehensive disclosure are much higher compared to privately held companies with concentrated ownership, which in turn suggests comprehensive disclosure on a voluntary basis for strategic reasons. Although our findings revealed an improvement in reporting, there are also many critical voices in the literature related to reporting obligations. The negative effects identified in such critiques range from a restriction in flexibility to a reduction in reporting to a minimum standard (Schaltegger, 1997; Pedersen et al., 2013). Moreover, excessively strict or detailed standardization can undermine the principle of substance over form (Frost, 2007). On the other hand, too much leeway in reporting can sometimes lead to cherry-picking or greenwashing and thus to a dilution of the principle of materiality. Legal requirements that lead to standardization or harmonization of the structure of reporting might increase the clarity and comparability of reporting. The NFR of companies often includes a large amount of information. This often makes it difficult to obtain an overview or achieve comparability, especially in combination with the different focal points of the report.

Therefore, legislators must always weigh these two lines of argument against each other. The introduction of new or revision of existing reporting obligations should always be made with care, as adjustments may be required even for companies with experience in reporting. As theory and several empirical studies suggest, including our own, companies affected by regulation respond primarily by adjusting their existing reporting to maintain legitimacy and achieve compliance with regulatory requirements.

The main limitations of our study concern the sample and method. On the one hand, the study focuses exclusively on EURO STOXX 50 companies and therefore does not consider companies that are listed on another stock exchange or not listed at all. The reason for this focus is that the EURO STOXX 50 companies are guaranteed to be affected by the directive. In addition, they already have much experience in NFR, because they are exposed to more public attention and, consequently, to greater pressure to legitimize their actions. This, in turn, suggests that the observed changes can be attributed to the effect of the directive with more confidence.

The method used by the study also has a limiting effect, especially regarding the operationalization of the reporting improvement. Improvement in reporting was assessed using a three-stage scale in almost all categories. In doing so, the assessment aimed at checking the completeness of the information required for each GRI indicator. An investigation or evaluation of other forms of improvement was not carried out.

Further research should include companies other than the EURO STOXX 50, especially those that have not had such extensive experience with reporting non-financial information but are also affected by reporting obligations. In addition, it should be analyzed whether there will be further changes in reporting behavior, as suggested by institutional theory. Overall, there is a need for further research on the impact of obligations, especially against the background of the Corporate Sustainability Reporting Directive (Directive [EU] 2022/2464), which not only significantly expands the group of companies affected but also expands or specifies the reporting content.