Mandatory sustainability reporting and the disclosure-performance gap: insights from the EU directive Open Access

In alignment with the classic objective of corporate disclosures, the purpose of sustainability disclosures is to increase corporate transparency by mitigating information asymmetry between the management and stakeholders (Hess, 2014). The key difference, nevertheless, lies in the extension of the disclosure scope from being shareholder-oriented to stakeholder-oriented by integrating information across the Environmental, Social and Governance (E, S and G) pillars of sustainability ^[1] (Hörisch et al., 2020). Correspondingly, several government policies have surfaced globally to ensure responsible corporate behavior. Mandating sustainability reporting, commonly considered soft regulation, aims to enhance corporate accountability, thereby enabling stakeholders, including investors, analysts, consumers and employees, to reward or penalize companies through their market activities (Jackson et al., 2020). A prominent regulation in this respect is the European Union’s Non-Financial Reporting Directive 2014/95/EU (NFRD) (the EU Directive hereafter), implemented in 2017, requiring large publicly listed companies with over 500 employees in the EU countries to disclose certain sustainability information (Kinderman, 2020; European Union, 2014; Hummel and Jobst, 2024).

The introduction and consistent development of mandatory sustainability reporting has sparked significant academic interest in evaluating its effectiveness from various perspectives. Studies have analyzed its influence on corporate transparency, focusing on the quantity (e.g. Ioannou and Serafeim, 2017; Fiechter et al., 2022) and quality (e.g. Dumitru et al., 2017) of the disclosures. Additionally, companies’ sustainability performance has also remained the focus (e.g. Jackson et al., 2020; Fiechter et al., 2022). While these approaches have merit, a crucial aspect often overlooked is the consistency and/or the gap between the disclosures and sustainability performance (the disclosure-performance gap, hereafter). We consider assessing this gap to avoid falling into the “greenwashing trap,” where the authenticity and credibility of sustainability disclosures are at risk of being merely symbolic (Yu et al., 2020; Ruiz-Blanco et al., 2022). Even though the purpose of sustainability disclosures is to provide a true and fair picture of corporate performance across E, S and G, these disclosures have often been criticized for being rhetoric in nature, issued solely for legitimacy purposes and unsubstantiated by actual performance (Blome et al., 2017; Pizzetti et al., 2021). Moreover, mandatory sustainability reporting is considered a hybrid or a soft regulation, which legally binds companies to issue disclosures but lends significant discretion over actual activities. Such regulations unfold their steering potential only in combination with other market factors and the company’s policies (Steurer, 2013). In a landscape of stringency regarding the issuance of disclosures but flexibility in content and type of information included, the companies might engage in box-ticking practices (Jackson et al., 2020). Therefore, we argue that without focusing on the disclosures and performance simultaneously, studies risk biased outcomes, undermining the primary goal of such reporting, which is to enhance transparency by aligning companies’ stated actions with their actual practices.

Our research framework identifies two contrasting outcomes of mandatory sustainability reporting in terms of corporate transparency: substantive and symbolic transparency. Substantive transparency, aligning with the actual goal of regulation, arises when it leads to high-quality disclosures aligning with sustainability performance ^[2]. This outcome is grounded in signaling theory, which suggests that companies use disclosures to convey their strengths and build stakeholder trust (Verrecchia, 1983; Karaman et al., 2021; Uyar et al., 2020). Symbolic transparency, on the other hand, occurs when companies resort to greenwashing, which refers to producing superficial or rhetorical disclosures without meaningful organizational changes (Ruiz-Blanco et al., 2022). Legitimacy theory underpins this phenomenon, positing that regulations may sometimes unintentionally lead to symbolic transparency, where poorly performing companies use disclosures as a tactic to maintain stakeholder approval (Deegan et al., 2002; O’Donovan, 2002).

Moreover, existing research on mandatory sustainability reporting often suffers from temporal, regional and thematic limitations (Korca and Costa, 2021; Dinh et al., 2023). Studies so far have focused on a narrow time frame, assessing only the short-term effects of the regulation (Dumitru et al., 2017; Cosma et al., 2022). Additionally, investigations are typically limited to specific countries, for instance, Italy (Korca et al., 2021), Italy and Spain (Posadas and Tarquinio, 2021), Poland (Matuszak and Różańska, 2017), Germany and Italy (Mion and Loza Adaui, 2019) and Poland and Romania (Dumitru et al., 2017). These studies neither offer a comprehensive view of the regulation’s large-scale impact nor insights into country-level variations. Another prevalent limitation is focusing on a single industry or exclusively one ESG pillar. Analysis of mandatory carbon disclosures (Grewal et al., 2022), mine-safety disclosures (Christensen et al., 2017) or the social dimension only (Jackson et al., 2020) does not provide holistic opinions about the regulation that holds for the three sustainability pillars and applies to all industries. Another important aspect overlooked in literature is the heterogeneity arising from contextual elements i.e. country, industry and culture. Korca and Costa (2021) highlighted the need for considering the effect of context while studying mandatory sustainability reporting to capture more meaningful results.

To address these gaps, this study applies a multitheoretical approach to assess the effect of the EU Directive on corporate transparency. Specifically, we examine whether mandatory sustainability reporting leads to substantive or symbolic transparency, where transparency refers to the alignment between a company’s sustainability performance and disclosures (Hummel and Schlick, 2016). We further analyze the moderating role of institutional factors and stakeholder pressure by examining country-level and industry-level heterogeneity. Applying institutional theory, we predict that prevailing market dynamics in a country, particularly in terms of stakeholder- and shareholder-orientation, can shape a company’s compliance with sustainability-related regulations (aligning with Dinh et al., 2023). Similarly, drawing on stakeholder theory, we propose that companies belonging to the industries with higher stakeholder scrutiny may show a comparatively higher degree of compliance than the others.

We empirically tested our theoretical predictions using the difference-in-differences estimation on a sample of large publicly listed companies in 28 OECD (Organization for Economic Co-operation and Development) countries over a ten-year period (2013–2022). The treatment group was derived from the 2017 implementation of the EU Directive, while the control group comprised non-EU OECD companies. The regression results showed an overall shift toward symbolic transparency for the EU companies in the postimplementation period. However, when divided into civil-law and common-law categories, the EU companies exhibited a significant heterogeneity in results. Civil-law countries showed a shift toward substantive transparency, while common-law countries tended to move toward symbolic transparency. Analysis of industry-level heterogeneity showed surprising results, as both environmentally-sensitive and high-customer proximity categories exhibited symbolic transparency in the postimplementation period. Therefore, building on a multitheoretical approach, our empirical results delineate the conclusion that companies issue sustainability disclosures primarily to appease stakeholders, even in a mandatory setting.

This study yields theoretical and practical implications. First, it contributes to the literature by examining the efficacy of mandatory sustainability reporting, distinguishing between substantive and symbolic transparency. Second, we examine the influence of the EU Directive over a ten-year period with a six-year postimplementation phase. This longitudinal approach enhances the robustness of the findings by capturing cumulative effects. Third, we highlight the significant country-level and industry-level heterogeneity.

Regarding practical implications for policymakers, the wider disclosure-performance gap highlights the lack of reliability of sustainability disclosure. These findings align with and reinforce the rationale behind revising the NFRD and implementing the upcoming Corporate Sustainability Reporting Directive (CSRD), particularly its introduction of mandatory assurance and the European Sustainability Reporting Standards (ESRS).

The rest of the paper is structured as follows: Section 2 offers background and literature, and Section 3 formulates hypotheses. Section 4 accounts for methodology, Section 5 presents empirical results, Section 6 includes robustness tests and Section 7 concludes with implications, limitations and potential areas for future studies.

The beginning of mandatory sustainability reporting in the EU dates to 2014, when the EU parliament and the Council adopted the current EU Directive (NFRD), which requires to include information on five key areas: environmental protection, social responsibility, labor rights protection, anticorruption and bribery and board diversity (Hummel and Jobst, 2024). The directive was implemented, across all large publicly listed EU companies having over 500 employees, with dual objectives. First, as articulated in the mandate (Directive 2014/95, Recital 1), the directive aims to enhance the transparency of sustainability-related disclosures. Second, as inferred from the directive’s underlying objectives, the regulators intend it to be a strategic measure to improve companies’ sustainability performance (Fiechter et al., 2022). This inference is derived from the statements “disclosure of non-financial information plays a crucial role in steering change toward a sustainable global economy” (Directive 2014/95, Recital 3) and “revealing non-financial information aids in the evaluation, oversight, and management of enterprises’ operations and their societal impact” (Directive 2014/95, Recital 3).

Since it is the first transnational sustainability reporting regulation, its implementation has stimulated empirical research (e.g. Dumitru et al., 2017; Venturelli et al., 2017; Stolowy and Paugam, 2018; Posadas and Tarquinio, 2021; Cosma et al., 2022; Ottenstein et al., 2022). Typically considered as soft regulation, mandatory sustainability reporting is expected to generate positive spillover effects on firms’ sustainability performance by embedding more integrated changes and reducing externalities (Ioannou and Serafeim, 2017). However, achieving the intended outcomes remains uncertain, as concerns persist regarding insufficient structure and ambiguity in effectively designing disclosure models, potentially resulting in disclosures of poor quality (Doni et al., 2020). Moreover, these outcomes may fall short if market incentives remain unchanged, increasing the risk of greenwashing (Lashitew, 2021). Similarly, the effectiveness of the EU Directive has frequently been questioned due to its inherent limitations, specifically, no mandatory assurance, lack of standardized reporting framework and “comply or explain” option (La Torre et al., 2018; Hummel and Jobst, 2024).

We acknowledge that the current EU Directive is deemed to be replaced by the Corporate Sustainability Reporting Directive (CSRD) in 2025 (Hummel and Jobst, 2024). However, in alignment with La Torre et al. (2018), we believe that it serves as a pivotal milestone in integrating sustainability and nonfinancial factors into corporate reports and will remain a reference for the current and upcoming regulatory framework. Therefore, the researchers and practitioners can gain actionable insights from its implementation, especially considering its substantial duration in effect. Moreover, the fundamental rationale underpinning the upcoming regulations, that is, to mandate corporate sustainability reporting, aligns with that of the current directive. Therefore, the scope of this study remains well-justified by focusing on the existing directive.

This section presents the theoretical framework to predict the influence of mandatory sustainability reporting on corporate transparency and the moderating role of institutional environment and stakeholder pressure. Figure 1 illustrates the research framework, which bridges mandatory sustainability reporting to corporate transparency, one leading to substantive transparency and the other to symbolic transparency. The following subsections detail the development of the corresponding hypotheses.

Stakeholders, including investors, analysts, consumers and employees, are interested in assessing companies’ sustainability performance, rewarding superior performers while penalizing the inferiors (Dhaliwal et al., 2012; Jackson et al., 2020). To address stakeholders’ concerns, governments implement mandatory sustainability reporting, primarily aimed at enhancing transparency by mitigating information asymmetry between the management and stakeholders (Hess, 2014). Such regulatory interventions often drive companies to increase sustainability-related disclosures, creating an initial surge in disclosure quantity (Ioannou and Serafeim, 2017; Ottenstein et al., 2022). Companies use these disclosures as legitimization tools to signal compliance and satisfy stakeholders (Cho and Patten, 2007; Cho et al., 2018). However, we argue that transparency encompasses more than the quantity of disclosures; it requires alignment between a company’s reported information and its actual behavior, ensuring a credible representation of its environmental, social and governance performance (Hummel and Schlick, 2016). In this context, the effectiveness of mandatory sustainability reporting, being soft regulation (Steurer, 2013), is widely debated (Aureli et al., 2019), with two contradictory outcomes: substantive transparency and symbolic transparency. The key distinction between both lies in the extent to which the disclosures represent a company’s actual performance (Truong et al., 2021).

On the one hand, proponents of mandatory sustainability reporting argue that such regulations prompt substantive transparency; that is, companies issue high-quality disclosures that align with the actual performance (Grewal et al., 2022). According to this view, mandatory sustainability reporting drives organizational as well as broader market changes as corporate disclosure and performance are mutually correlated, with one influencing the other (Leuz and Wysocki, 2016; Roychowdhury et al., 2019). Although mandatory sustainability reporting uses regulatory instruments to enhance transparency, it nevertheless stimulates companies’ self-regulation, making them more responsible (Lepoutre et al., 2007; Steurer, 2013). This premise is grounded in signaling theory (Verrecchia, 1983), which suggests that companies outperforming in sustainability will accelerate the release of disclosures with an incentive to get rewarded by the stakeholders (Hummel and Schlick, 2016; Uyar et al., 2020; López-Santamaría et al., 2021). Such companies strategically use mandatory reporting requirements to highlight their superior sustainability practices, thereby gaining financial and intangible benefits such as enhanced value, customer loyalty, brand reputation and competitive advantage (Porter, 1991; Cao and Rees, 2020; Downar et al., 2021). Simultaneously, increased disclosures often stimulate organizational changes by enabling managers to systematically measure, monitor and analyze a company’s environmental and social impacts. This systematic evaluation not only fosters improvements within the organization but also influences broader market dynamics. Institutional theory suggests that transparency driven by mandatory reporting motivates companies to imitate peers, leading to standardized and homogenous market-wide practices (Jackson et al., 2020). While this imitation can sometimes result in rigid strategies, it also raises the overall quality of sustainability practices across industries, setting higher benchmarks and promoting competition (Chatterji et al., 2016; Russo-Spena et al., 2018). Hence, mandatory sustainability reporting drives a change toward substantive transparency where companies’ disclosures are verifiable by the actual performance, narrowing the gap between disclosure and performance.

On the other hand, critics argue that companies may indulge in box-ticking practices while complying with mandatory sustainability reporting, leading to symbolic transparency. In this context, the emphasis is primarily placed on disclosure quantity without corresponding improvements in actual performance (Caputo et al., 2021). Legitimacy theory supports symbolic transparency, positing that the primary aim of corporate disclosures is to satisfy stakeholders (Deegan et al., 2002; O’Donovan, 2002). Under regulatory pressure, companies with poor environmental and social performance may feel compelled to increase disclosure quantity to signal compliance with regulation and commitment to stakeholders. Mandatory reporting frameworks, therefore, may create an environment where poor sustainability performers might strategically exploit disclosure requirements. Since the intention of sustainability reporting is to satisfy stakeholder demands, companies’ communication may exceed the actual performance by distorting the reality (Ruiz-Blanco et al., 2022). Companies resort to this approach to avoid being punished by stakeholders (Seele and Gatti, 2017).

Several factors may lead to symbolic transparency, specifically under mandatory sustainability reporting (Steurer, 2013). First, the lack of a standardized framework results in heterogeneity of disclosures, complicating users’ ability to interpret and assess the information or benchmark companies’ underlying performance (Christensen et al., 2021). Second, the “comply or explain” provision in current regulations and the lack of hard enforcement mechanisms, such as penalties or fines, may reduce companies’ incentives to alter existing practices. The issued disclosures may include explanations for noncompliance, highlight goal misalignments or report negative news (Ioannou and Serafeim, 2017; Hummel and Jobst, 2024), contributing to a broader disclosure-performance gap, as the quantity of reported information increases without underlying performance improvements. Third, compliance with mandatory sustainability reporting induces an economic burden on companies in the form of administrative, proprietary and political costs (Christensen et al., 2021). Given these factors, mandatory sustainability reporting is often linked to a mere disclosure quantity increase, while qualitative improvements remain limited (Caputo et al., 2019).

In summary, signaling theory anticipates a narrow disclosure-performance gap following regulatory interventions, whereas legitimacy theory suggests a wider gap. This theoretical ambiguity implies that both substantive and symbolic transparency are plausible outcomes of mandatory sustainability reporting, establishing a need for examination. Hence, we propose two alternative hypotheses:

The concept of sustainability varies widely across national and industry contexts (Ruiz-Blanco et al., 2022), potentially leading to heterogeneity in compliance with mandatory sustainability reporting. Moreover, the implementation of regulations is often context-dependent, leading to variability across companies (Banghøy et al., 2023). The following subsections delve deeper into these concepts.

The effect of regulations is rarely uniform; it depends significantly on the interaction between regulations and prevailing institutional factors in a country (Dumitru et al., 2017; Jackson et al., 2020). Institutional theory supports this premise, positing that there is a different macrolevel setting in each country that potentially affects companies’ sustainability performance (Kolk and Perego, 2010). Therefore, compliance with mandatory sustainability reporting is often steered by market norms (Steurer, 2013). Dinh et al. (2023) highlighted the need to analyze the variations in sustainability reporting between rather stakeholder- and shareholder-oriented countries. In a similar context, Liang and Renneboog (2017) evidenced that companies operating in civil-law countries exhibit a more substantial commitment to sustainability than common-law economies, with companies from Scandinavian countries demonstrating the highest levels. Civil-law countries are characterized by markets with low shareholder control, high stakeholder orientation and managerial discretion (Porta et al., 1998). These institutional factors allow companies to shift focus from solely financial goals to broader nonfinancial ones, such as fostering relationships with society (Dhaliwal et al., 2012). Consequently, companies operating in these markets are more likely to integrate sustainability goals into their strategies and operations, driving substantial organizational changes. In such environments, mandatory sustainability reporting can stimulate substantive transparency by aligning sustainability disclosures and actual sustainability performance, creating an equilibrium between both. Thus, we propose the following hypothesis:

Common-law countries, contrarily, represent markets with high shareholder control, low stakeholder rights and less managerial discretion (Porta et al., 1998). These institutional factors potentially constrain a company’s commitment to sustainability to minimum levels, as the primary focus is on profit-maximization goals (Demirbag et al., 2017). Therefore, the general market is characterized by a “race to the bottom” in terms of sustainability (Mellahi and Wood, 2004), where companies are less incentivized to pursue nonfinancial goals. In this context, companies may engage in box-ticking practices to satisfy regulatory requirements, leading to symbolic transparency. Therefore, we formulate the following hypothesis:

Stakeholder theory provides a suitable framework for understanding the industry-level heterogeneity arising in corporate transparency in response to regulations. This theory underscores the reciprocal relationship between companies and stakeholders (Freeman and Phillips, 2002). On the one hand, stakeholders’ expectations shape companies’ strategy, behavior and communication on sustainability (Fernandez-Feijoo et al., 2014). On the other hand, companies issue sustainability disclosures to modify stakeholder perceptions, aiming to enhance their reputation, demonstrate accountability and align with societal expectations (Ruiz-Blanco et al., 2022). The extent of stakeholders’ ability to assess the alignment between sustainability disclosures and performance is the key element in a company’s decision-making between substantive and symbolic transparency (Bebbington et al., 2008). Thus, companies with higher visibility (Delmas and Montes‐Sancho, 2010) and stakeholder proximity (Schons and Steinmeier, 2016) face higher reputational risks and fewer opportunities for issuing greenwashed information. In this case, mandatory sustainability reporting is expected to stimulate substantive transparency more efficiently. Therefore, we propose the following hypothesis:

Environmentally-sensitive industries, such as those associated with hazardous emissions or pollution, are subject to heightened stakeholder scrutiny (Ruiz-Blanco et al., 2022). Companies in these industries face greater pressure to align their sustainability practices with stakeholder expectations. Marquis et al. (2016) found that companies in environmentally-sensitive industries are less likely to engage in greenwashing, with those under regulatory pressure demonstrating even greater authenticity in their disclosures. Walker and Wan (2012) reported that greenwashing has significant negative financial implications for visible polluting companies, further incentivizing transparency.

Given these dynamics, we propose that companies in environmentally-sensitive industries are more likely to adopt substantive transparency in response to mandatory sustainability reporting, as the risks of reputational damage and financial penalties outweigh the benefits of symbolic compliance. Therefore, we propose the following hypothesis:

The unit of analysis in this study is a publicly listed company meeting the implementation criteria for the EU Directive. The final sample was constructed after merging data from multiple sources. The sustainability-related data were retrieved from Bloomberg and Refinitiv Asset4, and the financial data was obtained from Worldscope. The final sample comprises large companies with over 500 employees, publicly listed on the respective stock exchanges of 28 OECD countries from 2013 to 2022. The sample was split into a treatment group comprising 17 EU countries subject to the EU Directive and a control group containing 11 non-EU OECD countries not exposed to any comparable regulation during the studied period. The non-EU OCED countries serve as a reliable counterpart to EU countries for two reasons:

1. These are advanced economies with stable capital markets, where the companies are characterized by high standards of corporate governance, accounting practices and transparency and have readily available financial and operational data (OECD, 2021).

2. Their role as a control group against EU countries has been established in the literature on mandatory financial and sustainability reporting (e.g. Kim et al., 2012; Jackson et al., 2020; Ottenstein et al., 2022).

The analysis spans two phases centered around the implementation of the EU Directive in 2017: the preimplementation phase (2013–2016) and the postimplementation phase (2017–2022). The final sample consists of 1,511 companies (15,110 company-year observations). The treatment group comprises 440 companies (4,400 company-year observations), while the control group includes 1,071 companies (10,710 company-year observations). Table 1 presents the sample distribution across years, industries and countries.

The dependent variable of this study is Dis_perf_gap, defined as the extent of the gap between a company’s sustainability disclosures and performance (Yu et al., 2020; Ruiz-Blanco et al., 2022). It is quantified using equation (1); the higher (lower) the value, the larger (smaller) the gap:

The sustainability disclosure score was retrieved from Bloomberg, renowned for its comprehensive repository of data encompassing stock market, financial and ESG metrics, in alignment with prior studies (Ioannou and Serafeim, 2017; Yu et al., 2020). Bloomberg’s sustainability disclosure score encompasses the extent of all positive, negative, qualitative and quantitative ESG disclosures issued by companies. It varies from 0.1, the lowest level of disclosure, to 100, the highest, calculated through Bloomberg’s proprietary calculation methodology covering over 900 indicators, including metrics like direct greenhouse gas emissions, energy usage, water consumption, waste management, workforce diversity, safety incidents, governance practices and political contributions. This score primarily reflects the quantity of sustainability information disclosed and does not assess sustainability performance.

The sustainability performance score was proxied with Asset4 three pillar score on the E, S and G dimensions of sustainability, in tandem with the literature (Cheng et al., 2014; Ioannou and Serafeim, 2017; Yu et al., 2020). The Asset4 performance score is a peer-relative score, ranging between 0 and 100, for a company’s achievements in key metrics for E, S and G pillars, for example, from emission reduction to labor quality.

The hypotheses-specific independent variables are discussed below:

H1a and H1b: EU, an indicator variable taking value 1 if the company resides in the treatment group and 0 otherwise; Post is an indicator variable taking value 1 for the postimplementation period and 0 otherwise.

H2a: Civ_law, an indicator variable taking value 1 if the company resides in France, Sweden, Finland and Norway and zero for the rest of EU countries (Porta et al., 1998; Spamann, 2010).

H2b: Comm_law, an indicator variable taking value 1 if the company resides in the UK and zero for rest of the EU countries (Porta et al., 1998; Spamann, 2010).

H3a: Env_sens, an indicator variable taking value 1 for industries pharmaceutical, chemical, mining, metals, papers, transportation, petroleum and utilities, and zero for the rest of industries (cf., Brammer and Millington, 2005; Michelon et al., 2015).

H3b: Cust_prox, an indicator variable taking value 1 for energy utilities, financial services, food and beverages, health care, household and personal products, retailers, telecommunications, textiles and apparel, waste management, water utilities, commercial services, consumer durables, media and tobacco, and zero for the rest of industries (cf., Branco and Rodrigues, 2008; Fernandez-Feijoo et al., 2014).

To minimize the omitted variable bias, several control variables are added to account for alternative explanations and extraneous factors influencing the main analysis. Firm size, measured as the log of total assets, has been established as an essential variable in socio-economic studies (Orlitzky, 2001; Wang and Qian, 2011). ROA, Tobin’s q and Leverage are added to account for financial aspects. Since companies’ slack resources are considered to improve sustainability performance, the net cash flow from financing, investing, and operating activities scaled by total assets is included to mitigate the noise from firm size (Seifert et al., 2004). Companies’ sustainability performance across the E, S and G pillars separately are used, proxied by their respective Refinitiv Asset4 annual scores. To account for corporate governance factors likely to affect the company’s sustainability performance and disclosure strategies, we include board size measured by the number of directors on the company’s board and board independence measured by the proportion of independent directors (cf. Chams and García-Blandón, 2019). We also account for ownership structure by including the log of the percentage of shares in free float (cf. Ren et al., 2023).

To test our hypotheses, difference-in-differences analysis is conducted, estimating equation (2) using OLS regression:

$Dis_perf_gapijt$ accounts for the disclosure-performance gap of a company i, from industry j at time t. $Yijt$ is the hypothesized interaction term replaced in each respective analysis by the hypotheses-specific interaction terms, defined in Table 2.

Before estimating our dependent variable Dis_perf_gap, we explore the correlation between country-wise mean sustainability disclosure and performance scores for the treatment and control groups in Figures 2 and 3, respectively. The scatterplots exhibited a positive correlation between the two variables, consistent with the empirical evidence prevalent in the literature (Herbohn et al., 2014; Yu et al., 2020).

Subsequently, we create a boxplot for sustainability disclosure and performance scores, for all the companies included in our sample over the period of 2013–2022. As seen in Figure 4, the distribution for both variables varied significantly, evident from the difference in their respective means. The mean (median) for the sustainability disclosure score is 47 (46.9), while that for the sustainability performance score is 56 (58.5), hence the need for normalization.

Both the scores are min-max normalized, using equations (3) and (4).

where $Sus_dis_scrijt$ (⁠$Sus_perf_scrijt$⁠) represents the Bloomberg sustainability disclosure score (Asset4 sustainability performance score) of a company i belonging to industry j at time t, and Min (Max) represents the minimum (maximum) values for the respective scores. The normalization process transforms the data range for each score to a scale of 0–10, making them more comparable.

Table 3 presents the descriptive statistics summary for all variables used in regression analyses. The mean value for the variable Dis_perf_gap is −0.09, with a standard deviation of 0.16, indicating a moderate spread around the average.

To evaluate multicollinearity among variables, we estimate the variance inflation factors (VIFs). The average VIF value for all variables is 1.41 (maximum = 2.08 and minimum = 1.00), significantly below the commonly accepted threshold of 10, suggesting low risk of multicollinearity. To further assess the risk of multicollinearity Type 1 error (as outlined by Kalnins, 2018), we analyze bivariate correlations by generating the Pearson correlation matrix, see Table 4.

A fundamental premise of our research framework is that mandatory sustainability reporting leads to an increase in disclosures. However, the extent to which these disclosures reflect actual performance remains uncertain. Therefore, prior to the main analysis, we test the validity of this premise by testing the influence of the EU Directive on disclosure quantity. The difference-in-differences results in Table 5, using sustainability_disclosure_score as a dependent variable, confirm a statistically and economically significant increase in disclosure quantity for the company-year observations subject to the EU Directive.

Subsequently, we switch to the main analysis, which addresses our primary research question of whether mandatory sustainability reporting stimulates substantive or symbolic transparency. Table 6 presents the results for estimating the difference-in-differences model for our main outcome variable, the Dis-perf_gap, using multiple specifications. Model 1 only includes the main variable of interest, EU×Post, providing a baseline assessment of its influence. Model 2 incorporates the control variables only, while Model 3 includes the full set of variables. Model 4 addresses potential concerns regarding Type I errors by excluding the control variables, which positively correlated with EU×Post in Table 6 but demonstrated conflicting signs in the regression results. This step ensures the robustness and interpretability of the findings. The results across all models are consistent in terms of coefficient and signs.

In Model 3, accounting for the full set of variables, the regression results yield a statistically significant and positive coefficient of 0.0722 (t: 27.067) for the interaction term, EU×Post, indicating a shift toward symbolic transparency in the EU companies after the EU Directive implementation. These findings reject the substantive transparency H1a and support the symbolic transparency H1b. Beyond statistical significance, these results reflect a meaningful, practical impact of the EU Directive on the disclosure-performance gap. Given the mean of −0.09 and a standard deviation of 0.16 for the complete sample, this increase represents approximately 80% of the mean and 45% of the standard deviation. Such a relative change indicates that the Directive has significantly influenced companies’ reporting practices, incentivizing symbolic transparency rather than substantive improvements in sustainability performance. Our results align with the literature, which raises skepticism about soft regulations for sustainability and stresses that such requirements stimulate mere box-ticking practices (e.g. Caputo et al., 2021; Breijer and Orij, 2022). Supporting the legitimacy theory perspective, these results reiterate that companies’ primary motivation for issuing sustainability disclosures seems to be to demonstrate compliance and satisfy stakeholders rather than implementing genuine organizational change.

We examined the country-level heterogeneity within the EU by introducing two interaction terms, Civ_law×Post and Comm_law×Post, to assess the existence of symbolic versus substantive transparency in civil law and common law countries, respectively. First, we augmented Civ_law×Post in equation (2), and the regression results in Table 7 showed a negative coefficient (coefficient: −0.019, t = −3.876) for it in Model 3, indicating a lower disclosure-performance gap for civil-law countries in the postimplementation period compared to other EU countries, supporting H2a. While the coefficient for Comm_law×Post in Table 8, Model 3, takes a positive value (coefficient: 0.040, t = 8.102), affirming H2b.

Therefore, civil law countries, characterized by greater stakeholder orientation and managerial discretion, demonstrate a stronger commitment to sustainability by aligning more effectively with regulatory requirements, leading to substantive changes in core business practices. In contrast, common law countries tend to adopt compliance measures primarily for symbolic purposes. These findings provide empirical support for both institutional and stakeholder theories, aligning with the literature’s perspective that shareholder-centric markets tend to prioritize financial objectives over environmental and social considerations (Mellahi and Wood, 2004; Demirbag et al., 2017), while stakeholder-oriented markets exhibit the opposite trend (Dhaliwal et al., 2012). However, we extend this understanding by demonstrating that these dynamics also hold true in mandatory settings, underscoring the significant role institutional environments play in shaping the effectiveness and influence of regulations.

To investigate the industry-level heterogeneity, we introduced two interaction terms, Env_sens×Post and Cust_prox×Post. Regression results in Tables 9 and 10 show a statistically significant increase in the disclosure-performance gap for both variables, leading to the rejection H3a and H3b. Industries under heightened stakeholders’ scrutiny, specifically environmentally-sensitive and those with higher customer-proximity, are more likely to engage in symbolic transparency in the postimplementation period. These findings contradict studies on voluntary sustainability reporting, which observed lower levels of symbolic practices or greenwashing in such industries (e.g. Ruiz-Blanco et al., 2022). This discrepancy may suggest that regulatory enforcement gives stakeholders a sense of confidence, potentially reducing their vigilance over companies. Consequently, this leniency may lead companies to shift away from substantive actions, engaging more frequently in symbolic practices.

A typical challenge in regulation-focused studies is to empirically disentangle the impact of regulation itself from other confounding factors (Ioannou and Serafeim, 2017; Banghøy et al., 2023). We assess the sensitivity of our results to this issue by reestimating the main model using propensity score-matched groups. Specifically, we match the year before the EU Directive implementation, 2016, on the company characteristics Industry, Tobin’s q, Firm size, Leverage and ROA, using equation (5):

The treatment variable takes the value 1 (0), if the company i belongs to the EU (non-EU OECD) and Industry is an indicator variable that equals 1 (0) for industry j (for all others). The summary statistics pertaining to the matching algorithm are presented in Table 11 in terms of effect size, measured by Cohen’s d, for each covariate before and after matching. Cohen’s d is a measure used for comparing two groups, usually expressed as a standardized difference between their means (Cohen, 2013).

According to Figure 5, the matching procedure is well-executed. Cohen’s d value decreases after matching, which signifies that the mean differences for all covariates are practically significant across the treatment and unmatched control group (before propensity score matching) but practically insignificant between the treatment and matched control group (after propensity score matching). The insignificance of mean differences indicates that the groups have become closer and more similar in terms of the considered covariates. The final sample, after matching, comprises 374 closely similar companies in each of the counterpart groups.

Regression results in Table 12 for estimating equation (2) with the propensity score-matched groups are robust to initial findings. The coefficient for EU×Post is positive and statistically significant, signifying a higher disclosure-performance gap for EU companies in the postimplementation period compared to the non-EU OECD companies.

A critical assumption for our methodological approach is that the treatment and control groups are valid counterfactuals to each other (cf. Atanasov and Black, 2016), and that the variable Dis_perf_gap would have followed its respective trend, given the EU companies were not exposed to the EU Directive. To test this assumption, we estimated yearly treatment effects using 2014 as a benchmark year.

Figures 6 and 7 illustrate the result by plotting point-estimates with two-tailed 95% confidence intervals for all four yearly treatment effects on sustainability disclosure score and sustainability performance score, respectively. Our research framework is built on the premise that regulatory interventions lead to an increase in disclosure quantity, which results in either symbolic or substantive transparency, depending on the corresponding increase in sustainability performance. Therefore, analyzing both components individually provides deeper insights into the dynamics of disclosure quantity and actual performance, whereas focusing solely on the disclosure-performance gap would have revealed the extent of the gap without accounting for the underlying reason.

Graphical inspection of Figures 6 and 7 reveals two insights. First, the treatment effects in the preimplementation period (2015 and 2016) are insignificant for both sustainability disclosure and performance score. It means we find no evidence of different trends between the treatment and control groups with respect to the two variables. Second, the treatment effects of the EU Directive are significant for disclosure quantity in the postimplementation period (2017 and 2018) in Figure 5, as the confidence interval excludes zero, suggesting a meaningful divergence in outcomes between the groups. However, the treatment effects remain insignificant for sustainability performance postimplementation. These findings provide confidence in our results of symbolic transparency, revealing that the EU Directive influenced disclosure quantity only.

Since sustainability practices are not randomly assigned but instead arise from management’s strategic decision-making, endogeneity can be a threat in sustainability-focused studies, leading to biased conclusions (Du et al., 2023). Accounting for this issue, we thoroughly diagnosed and identified the potential causes of endogeneity, prescribed and explained the cause-specific solutions and reported the results for primary and supplementary analysis (cf., Hill et al., 2021). The main endogeneity threat in this study arises from the nonrandom or “selected” nature of the treatment, the EU Directive implementation, leading to “selection of treatment” bias (Hill et al., 2021). Many companies adopted sustainability disclosures and practices voluntarily before the directive’s implementation. Moreover, given the soft nature of the EU Directive, there is a likelihood of systematic variation in companies’ commitment to sustainability even in the postimplementation period. These issues can lead to a bias in inferring its true effect. Therefore, we resorted to estimating the average treatment effect technique (Angrist and Imbens, 1995; Wooldridge, 1997), using difference-in-differences analysis (Athey and Imbens, 2006).

Difference-in-differences accounts for temporal and cross-sectional effects of the treatment, thereby mitigating endogeneity (Hill et al., 2021). The risk of multicollinearity is addressed by testing multiple specifications of the main model. Parallel trend analysis is conducted to ensure the comparability of the treatment and control groups (see Section 6.2). Finally, for robustness, we repeat the analysis using propensity-score matched groups (see Section 6.1). To mitigate the omitted variable bias, we incorporate theoretically and empirically grounded control variables (see Section 4.4).

This study investigates the influence of mandatory sustainability reporting on corporate transparency, specifically examining whether such regulation fosters substantive transparency, where disclosures align with actual performance or merely symbolic transparency, where disclosures serve primarily as legitimacy tools. Leveraging the implementation of the EU Non-Financial Reporting Directive (NFRD) as a quasi-natural experiment, we apply a difference-in-differences design to assess changes in the disclosure-performance gap among EU companies relative to a control group of non-EU OECD companies.

The empirical results show that following the implementation of the Directive, EU companies exhibit a significantly wider disclosure-performance gap, indicating a systemic shift toward symbolic rather than substantive transparency. This finding is robust to multiple model specifications, parallel trend analysis, propensity score matching and comprehensive control variables. Moreover, we find compelling country- and industry-level heterogeneity in companies’ responses. At the country level, firms in civil-law countries display greater alignment between sustainability disclosures and performance, while firms in common-law countries exhibit more symbolic disclosure patterns. This divergence underscores the moderating role of national legal and institutional environments in shaping regulatory effectiveness.

Contrary to theoretical expectations, our industry-level analysis reveals that industries traditionally subject to high stakeholder pressure, such as environmentally-sensitive or high customer-proximity, are not more likely to adopt substantive transparency. Instead, these industries demonstrate a greater tendency toward symbolic transparency, suggesting that visibility and reputational risk may incentivize firms to report strategically rather than substantively.

Collectively, these findings indicate that while the EU Directive successfully increased the quantity of sustainability disclosures, it did not uniformly improve their substance. Instead, regulatory mandates may unintentionally encourage symbolic compliance, particularly in environments with weaker institutional enforcement or high legitimacy pressures. Our study thus contributes to the growing literature questioning the transformative capacity of soft regulation. In doing so, this study offers theoretical and practical implications, as discussed in the subsequent sections.

The study advances the debate on mandatory sustainability reporting by introducing and empirically validating the disclosure-performance gap as a comprehensive and meaningful construct for assessing transparency. Using this construct, we emphasize distinguishing between substantive and symbolic transparency, thereby underscoring a move beyond quantity or presence as proxies for transparency. This aligns with the dual objective of the EU Directive in specific and the rationale of mandatory sustainability reporting in general, i.e. to drive organizational change. While the core mechanism of the EU Directive is to increase disclosures/reporting, the broader ambition, as articulated in Recital 3 of the Directive, is “to stimulate a change toward a sustainable global economy” (European Union, 2014; Hummel and Jobst, 2024). In this sense, reporting serves as the means to an end, where the end goal is to improve companies’ sustainability performance (Fiechter et al., 2022). Our findings, of a widened disclosure-performance gap postimplementation, are interesting in this respect as they indicate that although disclosure quantity has risen, company-level sustainability performance has not improved proportionally.

The study also highlights the importance of incorporating contextual factors when evaluating regulatory effectiveness. Our results reveal significant heterogeneity across legal settings (civil vs common law) and industry characteristics (e.g. environmental sensitivity and stakeholder exposure), supporting the contingency view of regulation, i.e. its effectiveness is not uniform but context-dependent (Aragón-Correa et al., 2016).

Moreover, it addresses a gap in the literature by adopting a longitudinal design. Given the time lag between policy implementation and potential behavioral or performance changes, short-term studies may fail to capture the full effects of regulation (Korca and Costa, 2021). The multiyear, cross-country design illustrates these temporal dynamics and contributes to an increased knowledge of the pathways through which regulation may or may not achieve its intended outcomes.

The transition from the current version of the EU Directive, the NFRD, to the upcoming CSRD, set to fully apply in 2025, reflects a regulatory response to growing concerns about the reliability and credibility of sustainability disclosures (La Torre et al., 2018). These concerns are mirrored in our empirical findings, which reveal an increased gap between reported disclosures and actual sustainability performance.

The observed increase can be attributed to three main factors: box-ticking practices, greenwashing and disclosure of negative sustainability performance. While the latter may represent transparent reporting, the former two signal strategic compliance without substantive change, raising doubts about the material relevance of disclosure. Even if the wider gap stems from reporting negative performance, which, while transparent, does not fulfill the broader transformational aim of the directive: to drive organizations toward more sustainable behavior. Particularly, given the six-year postimplementation period, we expected to observe improvement in sustainability performance corresponding to reporting.

We conjecture that this gap may be addressed through more rigorous oversight and the implementation of structured, standardized guidelines for sustainability performance indicators. In this respect, the CSRD represents a significant advancement by mandating assurance of sustainability disclosure (European Union, 2022; Hummel and Jobst, 2024). Our findings suggest that these requirements are both timely and appropriate; however, the initial reliance on “limited assurance” may compromise their overall effectiveness. Following the implementation of NFRD, the number of companies seeking limited assurance had increased (Ottenstein et al., 2022), but such assurance is often limited to the existence of the report rather than verifying its content (Christensen et al., 2021). Moreover, the practical challenges arising from the current shortage of qualified sustainability auditors may hinder efforts to enhance the reliability of sustainability reporting (Krasodomska et al., 2021). Thus, policymakers must anticipate and mitigate these constraints by ensuring that assurance procedures are proportionate and meaningful for both companies and stakeholders.

Moreover, CSRD introduces a set of prescriptive guidelines, the ESRS, designed to enhance disclosure transparency. The ESRS is expected to narrow the disclosure-performance gap by promoting standardized and transparent sustainability reporting, thereby encouraging more robust long-term planning and accountability. Nonetheless, its full effectiveness may be influenced by several ongoing challenges, including capacity constraints among SMEs, variability in data quality, and uneven adoption across sectors and member states (Operato et al., 2025). Addressing these areas through continued support, harmonized implementation, and iterative refinement of the standards is essential in shaping effective reporting practices and ensuring a broader sustainability transition.

While our study provides valuable insights, it has limitations that future research could address. Notably, we did not account for variations in reporting requirements stemming from the transposition of the EU Directive into national legislation. Although this was beyond the scope of this study, which focuses on evaluating the broader impact of regulatory shifts, we suggest, in line with Dinh et al. (2023), that future research could explore the domestic implementation of the Directive. Additionally, our reliance on external evaluations of sustainability disclosure and performance could be expanded by incorporating innovative content analysis techniques, as proposed by Du et al. (2023), to generate more revealing insights. Moreover, after CSRD is implemented and has been in effect for a significant time, future studies can evaluate its effects on the disclosure-performance gap and compare it to the current study and the effects of Non-Financial Reporting Directive 2014/95/EU.