A Literature Review of CSR Disclosure Quality: Evidence From Restatements

In the preceding 20 years, CSR reporting has made significant strides. This study examines the quality of CSR disclosure from the perspective of CSR restatements after reviewing prior literature. Long-term improvement in enterprises‟ quality of CSR disclosure is believed to be made possible by more advanced reporting requirements, improved services from skilled auditors, and continuously evolving CSR reporting systems. These findings indicate that companies may proactively manage litigation exposure and risk of fines led by overstated or misstated CSR information, which implies that the level of legitimacy power is positively associated with CSR reports‟ reliability. This deduction is supported by previous research; Al-Shaer and Zaman (2016) state that companies tend to display a better commitment to implementation when regulators guide CSR practices. Abernathy et al. (2017) note that firms engage keenly to improve the reliability of CSR reports when supervised by regulators.


Theories for CSR Disclosures
A review of the mainstream theories may produce valuable insights into companies" engagement motives in CSR disclosure then contribute to a more comprehensive understanding of QCD. Legitimacy, stakeholder, and institutional theory, all derived from social and political theories, are widely employed to interpret companies" incentives of CSR disclosure (Deegan and Samkin, 2012) but are often used separately.
Legitimacy theory is the most employed theoretical perspective in CSR disclosure research, and the most prominent empirical evidence supports that. It underscores that organizations can only remain existing if the society on which they depend perceives an organization"s operating to a value system commensurate with the society"s value system (Gray et al., 2010). In line with legitimacy theory, CSR activities and reporting are regarded as a legitimization strategy (Fernando & Lawrence, 2014); thus, companies may engage in CSR activities and disclosures to obtain and preserve their legitimacy. Suchman (1995) finds that positive events (i.e.,environmentally-friendly behaviors, disclosing good news) enhance organizational legitimacy, vice versa. He defines pragmatic legitimacy as companies act to manipulate public perceptions of legitimacy based on their self-interested calculations.
Stakeholder theory highlights an organization"s accountability and its stakeholders" rights. Freeman (1984) defines a stakeholder as "any individual or group who affect or is affected by the achievement of a firm"s objectives." Based on stakeholder theory, companies involved in CSR disclosure attempt to demonstrate accountability to stakeholder groups (Fernando and Lawrence, 2014), primarily to manage the most potent groups (managerial perspective).
Institutional theory views organizations as operating within a social framework of norms and values about what constitutes acceptable or appropriate economic behavior (Carpenter and Feroz, 2001). Companies within the same field are likely to become homogeneous through following commonly institutional practices and conforming to generally accepted social values and norms (Fernando and Lawrence, 2014). Various forces, such as vital stakeholders" pressure and expectation, greatly influence companies" institutional practices (i.e., CSR disclosure), called coercive isomorphism (Deegan et al., 2009).
A theoretical foundation combining these three theories suggests more than one motive for CSR disclosure. Companies may consider CSR disclosure to legitimize their business, demonstrate accountability to stakeholders, and conform to societal beliefs and norms simultaneously. Powerfully external actors (i.e., government agencies, regulators), who can be seen as legitimacy sources, vital stakeholder groups, and dynamics in an institutional environment, may significantly influence firms" CSR disclosure. Besides, legitimacy theory, in especial the pragmatic perspective, implies that companies are likely to mask their CSR performance and seek "greenwashing" to manage legitimacy risk, which may negatively impact QCD.

Interpretation of CSR Restatements
Scholars deem a CSR report as being restated when a subsequent CSR report presents changes or corrected errors from the prior report (Ballou et al., 2018). Hence, there are two distinct types of CSR restatements, non-errors, and errors (Appendix 1). One-third of G250 companies had restated their CSR information from previous years in 2011 and 26% in 2013, whereas the percentage of N100 firms that issued CSR restatements slightly increased from 21% in 2011 to 25% in 2013 (KMPG, 2011;. Michelon et al. (2019) review 1200 stand-alone CSR reports issued by S&P 500 firms over 2010-2014 and find 177 that contain at least one restatement, 116 of which include restatements due to errors. The incidence of CSR restatements is alarmingly higher than that of financial report restatements (Note 1).
Prior literature documents that reasons for CSR restatements are substantially different from those for financial restatements. First, social and environmental performance, including qualitative and subjective components, is difficult to measure (Pinnuck et al., 2020). Second, without globally adopted reporting standards (Moroney and Trotman, 2016), managers have the discretion to report CSR performance and change measurement methodologies. Third, CSR reporting is still in its infancy, where definitions, scopes, and methodologies need continually improving, and companies" reporting systems are under-developed (Michelon et al., 2019). KMPG (2011; state that 79% of restatements belonged to non-errors (i.e., improved calculation methodologies, changes of applied definitions, and updates of scopes at 33%, 26%, and 20%, respectively). However, a significant proportion of restatements were due to errors or omissions. The percentage among G250 reporters was at 35% in 2011 and 22% in 2013.
CSR disclosures (Pinnuck et al. 2020). The pervasiveness of reporting errors erodes CSR disclosure"s informational value then weakens investors" confidence (Backof et al., 2020). However, a large volume of literature views CSR restatements positively. KMPG (2013) notes that restatements result from companies strengthening their internal reporting systems and improving data quality for decision making. Michelon et al. (2019) suggest that companies attempt to use restatements as a legitimacy-building tactic to improve disclosure quality. Ballou et al. (2018) consider such restatements as "good" and a positive step toward enhancing the effectiveness and quality of CSR disclosure.
No mandatory reporting standards, uncertainties and unquantifiable nature in social/environmental performance, and imperfect reporting systems pose challenges for companies to disclose accurate and reliable CSR information. However, a significant proportion of restatements due to errors or omissions, in combination with legitimacy theory, implies that self-serving management might intentionally overstate or misreport CSR performance, at least in the past leading to inaccurate and unbalanced disclosures. Overall, given that CSR restatements enable companies to demonstrate accountability and transparency to stakeholders and most restatements due to non-errors, consistent with previous research, restatements could be viewed as a positive sign and an indirect evidence for examining QCD.

Independent Assurance and QCD
In response to stakeholders" pressure for management to increase the credibility of disclosed CSR information (Maso et al., 2020), having CSR reports independently assured becomes a majority reporting practice, with 71% of G250 companies doing so, compared to 42% in 2010 (KPMG, 2020). There are two distinct assurance providers, accounting firms and consultants (GRI, 2020).
Prior literature provides empirical evidence on the association between independent CSR assurance and the incidence of restatements. Ballou et al. (2018) examine 2339 of the 7540 companies surveyed by KPMG between 2011-2013 and document that CSR assurance and use of accounting providers are significantly and positively associated with CSR restatements (both errors and non-errors). Further, they find that accounting firms improve reporting quality to a greater extent than non-accounting providers. Based on a sample of U.S. listed firms from 2010 to 2014, Michelon et al. (2019) conclude that voluntary assurance of CSR reports is associated with increased CSR restatements and more strongly related to restatements due to errors than to methodological updates. Focusing on CSR reports issued by G250 companies over 2006G250 companies over -2013G250 companies over , Pinnuck et al. (2020 find a positive association between having CSR reports assured and the frequency of restatements. They note that auditors are more likely to detect items that should be revised and encourage management to report revisions because they face liability and reputational costs. CSR assurance improves CSR reporting quality through identifying inaccuracies in prior reports and improvements to definition, scopes, and methodologies that require restatements for comparability (Ballou et al., 2018). The majority of previous publications (Pinnuck et al., 2020;GRI, 2013;Simnett et al., 2009) hold that CSR reports with external assurance indicate higher disclosure quality, and the quality will be greater when an auditing profession provides the service. Importantly, CSR assurance has been mandated in some jurisdictions (Maso et al., 2020) (Note 2).
Nonetheless, some criticize that CSR assurance is often unduly influenced by the management, thus failing to improve disclosure relevance or completeness (Simth et al., 2011), exhibiting high variation in scope, independence of provider, and the use of external criteria (Manetti and Becatti, 2009). Furthermore, Michelon et al. (2019) argue that providers only offer a limited level of assurance and may use restatements as a strategy to legitimize their services then expanding share in the assurance market.
Given its significant association with the occurrence of restatements, third-party CSR assurance, primarily provided by auditing firms (i.e., Big 4), plays an irreplaceable role in improving QCD. Whereas, being an emergent field, assurance services may need establishing essential guidelines and further regulating.

GRI Reporting Standards and QCD
GRI assists firms in making effective disclosures of CSR activities and information (Backof et al., 2020) through issuing non-binding reporting standards (Note 3). The GRI standards are acknowledged as the dominant global standard for CSR reporting (Mahoney et al., 2013), adopted by around two-thirds of the N100 and three-quarters of the G250 (KPMG, 2020).
Previous research provides limited evidence on the association between the use of GRI standards and CSR restatements. Ballou et al. (2018) find that CSR reports following GRI standards are only significantly associated with non-error restatements (i.e., definitions, scopes, and methodologies). Further, they suggest that although CSR reporting frameworks help enhance CSR disclosure quality through prompting non-error CSR restatements, these do not have the equivalent impact on error discovery and correction like assurance-related practices. Pinnuck et al. (2020) reveal that the adoption of GRI standards is only negatively related to the likelihood of restatements regarding overstatements of environmental information.
Scholars debate whether the adoption of GRI standards is the key to ensuring and improving CSR disclosure quality. Supporters argue that the GRI standards may improve the quality and comparability of disclosed CSR information across firms and time (Grewal et al., 2020). Pinnuck et al. (2020) suggest that GRI provides a benchmark against which reporting decisions can be judged; thus, firms following standards are less likely to report opportunistically. Since the standards" worldwide prominence, clarity and coverage, GRI followers can provide more reliable and relevant CSR information (Hahn and Lulfs, 2014).
However, opponents claim that GRI guidelines seem widely dispersed (Joseph, 2012) and fail to completely define and explain "materiality" (Michelon et al., 2019) (Note 4) and "completeness"; thus, companies may deliberately avoid disclosing the negative aspects of their CSR performance (Khan et al., 2020). Backof et al. (2020) suggest that due to their non-binding nature and relatively flexible characteristics, the GRI standards may be used in a biased way by companies to report on their social or environmental performance opportunistically. Under the pragmatic legitimacy perspective, they state, "companies tend to improve their apparent CSR performance simply through "ticking more GRI boxes," but this symbolic practice is unlikely to bring about any improvement within CSR disclosures." Michelon et al. (2015) find little evidence that CSR disclosures by firms adopting GRI guidelines are more balanced, comparable, and precise. They document that the use of GRI guidelines is not associated with the improved quality of CSR disclosure.
In light of the insufficient empirical evidence and the competing argument, whether adopting the GRI standards leads to higher QCD remains questionable. These standards provide a certain level of credentials and may enhance the comparability of CSR reports. Nevertheless, due to GRI standards" non-binding nature and vagueness in some fundamental concepts, linked to legitimacy theory, CSR reporters may follow these standards as a symbolic practice to portray an image of better commitment to QCD, thus managing stakeholders" perception.

Stakeholder Power and QCD
CSR reporting is frequently viewed as a kind of stakeholder-oriented behavior and responds to the stakeholders" information needs. The importance of meeting stakeholders" demand increases as the level of stakeholder power increases (Roberts, 1992).
Stakeholder power also refers to stakeholder orientation. Cheung et al. (2018) define stakeholder orientation as the extent to which corporate management"s vision of its roles and responsibilities includes the interests and claims of stakeholder groups and their legitimacy and power to influence corporate activities. Basing on the legal environment, laws related to CSR disclosure, and the level of public awareness for CSR practices, they rank stakeholder-orientated countries (Appendix 2). Prior literature notes that stakeholder orientation has a considerable influence on companies" CSR disclosure (Ramanna, 2013;Dhaliwal et al., 2014) and whether a country is stakeholder-oriented can affect the amount of companies" CSR information disclosure (Liang and Renneboog, 2017) Pinnuck et al. (2020) examine the frequency of CSR restatements by industry and country. They find that firms belonging to Electric utilities, Pharmaceuticals, and Oil, gas, and consumable fuel have a higher incidence of restatements (Appendix 3). Stakeholders demand these industries for more CSR disclosures. Companies residing in strong law countries (i.e., Netherlands, Denmark, Finland) are more likely to restate their CSR reports (Appendix 4). Simnett et al. (2009) provide some circumstantial evidence. Using a sample of 2113 companies from 31 countries, they find that companies operating within the mining, utilities, and finance are more likely to have their CSR reports assured. In stakeholder-orientated countries, firms tend to issue assured CSR reports. They further suggest that firms belonging to environmentally or socially sensitive industries have incentives to develop their CSR reporting systems and mechanisms to enhance the credibility of disclosed CSR information.
These findings indicate that companies may proactively manage litigation exposure and risk of fines led by overstated or misstated CSR information, which implies that the level of legitimacy power is positively associated with CSR reports" reliability. This deduction is supported by previous research; Al-Shaer and Zaman (2016) state that companies tend to display a better commitment to implementation when regulators guide CSR practices. Abernathy et al. (2017) note that firms engage keenly to improve the reliability of CSR reports when supervised by regulators. http://iref.julypress.com International Research in Economics and Finance Vol. 6, No. 3; By comparison, Michelon et al. (2020) observed a social movement over 2006-2012 launched by shareholders to demand CSR disclosure transparency. They conclude that companies did not substantially improve their CSR practice, except "producing" more information.
In line with stakeholder theory (managerial perspective), these results suggest that firms attempt primarily to signal accountability to powerful stakeholder groups and reduce litigation risk through enhancing the accuracy and reliability of disclosed CSR information. By contrast, they are less responsive to the claims of other groups (i.e., investors, environmentalists). Therefore, powerfully external actors (i.e., legislative bodies (Freeman, 1984), regulators, governments) have a more significant influence on firms" CSR reporting, and companies may increase QCD in response to their demand.

Conclusion
CSR reporting has achieved remarkable progress in the last two decades. A theoretical foundation suggests that firms have multiple motives to engage in CSR disclosures, including legitimacy, accountability, and isomorphic explanations. As stakeholders increasingly considering disclosed CSR information for decision-making, disclosure quality is vital.
During the past period, a high incidence of restatements led to concerns about the reliability of CSR report. Whereas most CSR restatements are due to non-errors used to signal accountability, consistent with prior literature, restatements could be viewed as a positive sign and used as a dimension to examine QCD. Nonetheless, a fair proportion of restatements due to errors, linked to implications from legitimacy theory, suggest that managers have incentives to report on CSR performance opportunistically, resulting in cosmetic disclosures and lower QCD.
Given their strong and positive association with CSR restatements, independent assurance services could bring about improvements in QCD, primarily provided by audit firms. In contrast, based on the ambiguous association and the competing argument, whether the adoption of GRI reporting standards leads to a substantial increase in QCD is still doubted. In light of the evidence that companies belonging to stakeholder-orientated countries or operating in specific industries are more likely to restate their CSR reports, stakeholder power may vary across countries and industries. Some companies proactively improve QCD, possibly because they need to reduce the litigation exposure and consequent risk of fine led by false CSR disclosure, rather than respond to demands from all stakeholder groups or fulfill public expectations. This inference highlights the role of legitimacy power.
In conclusion, more fine-tuned reporting standards, greater services from well-trained auditors, and fast-developing CSR reporting systems will build a solid foundation and undoubtedly enhance firms" QCD in the long term. However, corporate management might not make substantive efforts to improve QCD in the short term without legislative regulation and supervision. Policy-makers, regulators, and government agencies should have a higher responsibility for making companies more accountable for disclosed CSR information.
As a thriving area, CSR disclosure contributes ample directions to future research. Given its importance for social and environmental accounting, the quality of CSR disclosures could be examined through more relevant dimensions. Concretely, researchers can attempt to measure the disclosure quality from the perspectives of the institutional environment (i.e., the level of litigation) or stakeholder perception. Furthermore, future studies could focus on cross-country and cross-industry analyses because companies reside in different institutional environments. Given that the EU Commission is about to impose CSR assurance on large and media-size enterprises (Reuters, 2021), more direct evidence on the association between independent CSR assurance and disclosure quality could be expected. Considering that they remain continually evolving and improving, future research could investigate how and to what extent GRI reporting standards affect the disclosure quality.
CSR reporting develops over time. However, prior research generally samples CSR data for 2-6 years, and the significant lag between the sample period and the publication makes many data no longer referential. CSR information"s non-quantitative nature and many firms disclosing their information without standalone reports create difficulties for data collection. For example, data on restatements revealed by KPMG is hand-collected by accounting professions. Nowadays, emerging tools (i.e., Natural Language Processing, Python) enable researchers to capture first-hand data on CSR reports and conduct long-term studies. With enhanced data relevance and timeliness, the disclosure quality and its association with institutional dynamics or market reactions could be explored further and more timely.
Lastly, future research could pay closer attention to emerging economies where firms" initiatives, the level of quality, and stakeholder orientation for CSR disclosure remain under-explored.    Vol. 6, No. 3;