Corporate responsibility and corporate misbehavior: are CSR reporting firms indeed responsible?

Reitmaier, Christine; Schultze, Wolfgang; Vollmer, Julia Article — Published Version Corporate responsibility and corporate misbehavior: are CSR reporting firms indeed responsible? Review of Accounting Studies Provided in Cooperation with: Springer Nature Suggested Citation: Reitmaier, Christine; Schultze, Wolfgang; Vollmer, Julia (2024) : Corporate responsibility and corporate misbehavior: are CSR reporting firms indeed responsible?, Review of Accounting Studies, ISSN 1573-7136, Springer US, New York, NY, Vol. 30, Iss. 2, pp. 1804-1872, https://doi.org/10.1007/s11142-024-09850-8 This Version is available at: https://hdl.handle.net/10419/330628 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by/4.0/ Vol:.(1234567890) Review of Accounting Studies (2025) 30:1804–1872 https://doi.org/10.1007/s11142-024-09850-8 Corporate responsibility andcorporate misbehavior: are CSR reporting firms indeed responsible? ChristineReitmaier1· WolfgangSchultze1 · JuliaVollmer1 Accepted: 11 July 2024 / Published online: 28 October 2024 © The Author(s) 2024 Abstract We investigate whether firms that proclaim a commitment to corporate social responsibility (CSR) by CSR reporting indeed internalize such a commitment and behave more responsibly. We analyze the association of the issuance and quality of voluntary CSR reports with the occurrence, number, and severity of corporate misbehaviors, both preceding and subsequent to CSR reporting. We find a significantly positive association of CSR reporting with our measures of prior and future misbehavior. The results are corroborated by a quasi-natural experiment around the Rana Plaza disaster where we find that the signatories of an accord for better working conditions have significantly higher prior and future misbehavior relative to nonsignatories and firms unaffected by the exogenous shock. Our results are in line with legitimacy theory implying that, on average, the firms’ proclaiming commitment to CSR is not a signal of internalized commitment but more likely serves greenwashing and impression management purposes. Keywords Corporate social responsibility (CSR)· Corporate misbehavior· CSR reporting· Real effects· Signaling theory· Legitimacy theory JEL Classification G18· G32· K38· K42· M41· M48· Q01 * Wolfgang Schultze W[email protected]g.de 1 Department ofAccounting andControl, University ofAugsburg, Universitaetsstr. 16, 86135Augsburg, Germany 1805 Corporate responsibility andcorporate misbehavior: are… 1 Introduction Whether firms that publicly proclaim commitment to corporate social responsibility (CSR) indeed internalize such commitments and act responsibly is an unresolved research question.1 Recent initiatives to regulate CSR reporting, such as the European Non-Financial Reporting Directive 2014/95/EU, consider CSR reporting a means to better align managerial decision-making with stakeholders’ interests, that is, to alter firm behavior toward more responsibility (e.g., GRI 2011; Dechow 2023). However, prior empirical evidence on whether CSR reporting is indeed associated with more responsible behavior is inconclusive (see, e.g., Christensen etal. (2021) for a review). While more and more firms voluntarily disclose CSR information, there is no indication that the number of cases of corporate irresponsibility is decreasing. Media reports reveal misbehavior such as human rights abuse or environmental damage daily. Disclosure theory suggests that firms that signal commitment to CSR should “walk their talk” and reduce misbehavior (Hermalin 2013; Hoi etal. 2013), in line with stewardship theory and the related principle of “what gets measured gets done” (Donaldson and Davis 1991; Davis etal. 1997). Socio-political legitimacy theory (Dowling and Pfeffer 1975), in contrast, argues that firms primarily proclaim a commitment to CSR to legitimize misbehavior and enhance their public reputation (Moneva etal. 2006). Which theory more accurately explains firm behavior is largely an empirical question, which we revisit from a novel perspective. Prior studies commonly measure CSR-related activities as an aggregate measure of CSR performance. These measures aggregate responsible and irresponsible behavior and thus capture whether one of the two predominates (Hughes etal. 2001; Cormier etal. 2004). This measurement and aggregation issue has been found to be one of the main reasons why prior studies find contradictory evidence about the relation between CSR reporting and corporate (mis)behavior (e.g., Patten 2002; Mattingly and Berman 2006; Strike etal. 2006; Semenova and Hassel 2015; Christensen etal. 2021). When firms do positive things in some areas to compensate for irresponsible behavior in others, aggregate measures of CSR performance may indicate improvements despite increases in misbehavior (Kotchen and Moon 2012; Adams and Abhayawansa 2022). Such aggregation is highly problematic because irresponsible activities such as corruption, fraud, and child labor are associated with high economic and social costs, whereas responsible activities such as CO2 reduction and development aid do not have benefits on the same scale and cannot compensate for the irresponsible behavior (Hail etal. 2018). Responsible and irresponsible activities are in fact considered distinct concepts that do not simply describe opposite reflections of CSR but are largely independent of each other and follow different purposes (e.g., McGuire etal. 2003; Strike etal. 2006). We 1 Practitioners and academics interchangeably refer to “corporate responsibility,” “corporate social responsibility (CSR),” “environmental, social, and governance (ESG) issues,” and “sustainability” to describe commitment to sustainable economic development and ethical standards (e.g., VanMarrewijk 2003; Christensen, Hail, and Leuz 2021). 1806 C.Reitmaier et al. therefore do not examine an aggregate measure of CSR performance, but focus on irresponsible behavior to analyze if firms that proclaim commitment to CSR subsequently avoid misbehavior, in line with Friedman’s (1970, 126) demand on firms to stay “within the rules of the game” as well as the EU Taxonomy principle to “do no significant harm” (EU 2019, 4).2 To date, to the best of our knowledge, only one study (Christensen 2016) has analyzed the relation of voluntary CSR reporting and a non-aggregated measure of misbehavior. It finds a decline in CSR-related lawsuits subsequent to issuing a CSR report. These results may imply that firms improve their CSR performance subsequent to CSR reporting as they internalize the proclaimed commitments, consistent with the intentions of those who emphasize the stewardship role of disclosure, like the Global Reporting Initiative (GRI) (Christensen 2016). An alternative explanation, however, is that CSR reporting provides a mechanism for successful image improvement, such that legal prosecution decreases while corporate behavior is not improved. Moreover, prosecuted cases of corporate misbehavior are rare (Mark-Ungericht and Weiskopf 2007), so many cases of misbehavior go unnoticed when using such measures, particularly those that occur in subsidiaries under foreign jurisdiction. In fact, legal prosecution is not the only form of sanctions for corporate irresponsibility, as firms face repercussions from customers and other stakeholders when their public image is harmed. The public media plays a vital role in revealing misbehavior and shaping stakeholders’ perceptions (Einwiller etal. 2010; Dube and Zhu 2021). We thus study a broader measure of misbehavior that captures the firms’ accountability to the public, based on critical media reports. Moreover, prior literature has only studied the dichotomous decision whether to voluntarily issue a CSR report. We extend this by also considering CSR disclosure quality based on Bloomberg’s ESG disclosure score to provide a deeper understanding of the relation between CSR reporting and misbehavior. We use the independent Corporate Critic Research Database (CCRD), which includes a broad range of summaries of public information on (responsible and irresponsible) corporate behavior (CCRD 2020a). The CCRD is part of the British Ethical Consumer Initiative, which provides ethical ratings based on its research database. It collects firm-specific CSR-related news from all over the world. The database covers all firms that sell products or services in Europe. We identify cases and evaluate the severity of misbehavior based on this information. We define misbehavior as violations of the principles of the United Nations Global Compact (UNGC), which may be considered a consensus on global ethical standards (e.g., Rasche and Waddock 2014).3 Our measures of corporate irresponsibility thus do not 2 “There is one and only one social responsibility of business—to use its resources and engage in activities designed to increase its profits so long as it stays within the rules of the game, which is to say, engages in open and free competition without deception or fraud” (Friedman 1970, 126). “Responsible businesses enact the same values and principles wherever they have a presence, and know that good practices in one area do not offset harm in another” (UNGC 2020a). 3 The UNGC is a worldwide voluntary compact between the United Nations, firms, and other parties (e.g., NGOs and business associations) (UNGC 2020b). The ten UNGC principles encompass human rights, labor, environment, and anti-corruption that reflect responsibility to people, the planet, and longterm corporate success (UNGC 2020a). 1807 Corporate responsibility andcorporate misbehavior: are… focus on only one dimension of CSR, but comprehensively include environmental, social, and governance (ESG) issues. For robustness, we also use the ESG controversy score from Refinitiv as an alternative measure of corporate irresponsibility. Another feature that distinguishes our study from prior research is that we consider lead-lag relations (Granger 1969), i.e., corporate irresponsibility before and after CSR reporting. One might expect that firms with (more severe) prior misbehavior continue to engage in (more severe) misbehavior in the future (Treviño 2005; Wu 2014). However, if CSR reporting is indeed a credible signal of increased commitment to CSR, it should foster change and break a direct relation between past and future misbehavior (e.g., Linnenluecke and Griffiths 2010). Our main sample comprises the firms listed in the Fortune Global 500 between 2003 and 2013. For robustness, we also analyze the period until 2018. We focus on voluntary stand-alone CSR reports based on the GRI guidelines, which have become the de facto standard of CSR reporting (DeVilliers and Alexander 2014). We also consider CSR reports that do not apply the GRI guidelines in robustness analyses. The year 2003 represents the onset of a period of intense CSR reporting (Freundlieb and Teuteberg 2013). Our main sample period ends in 2013 to avoid potential biases due to (1) early adoption of the EU directive 2014/95/EU, which was issued in 2014 and renders CSR reporting mandatory for large European firms as of January 2017; and (2) adoption of the International Integrated Reporting Council’s framework, published in December 2013, which commonly leads to CSR reports being combined with annual reports and other reports in a single integrated report that is not captured by the GRI measure. We collect data on misbehavior for the time before and after CSR reporting, covering the period 1999–2018. Compared to prior CSR studies, we use a larger dataset that allows for broader conclusions. To corroborate our findings, we analyze a second sample that comprises the firms affected by an exogenous shock to textile-related industries. In a quasi-natural experiment, we investigate a different form of proclaiming commitment to CSR. After the Rana Plaza disaster in Bangladesh in 2013, many firms voluntarily signed an accord for better working conditions. We compare the signatories to non-signatories within textile-related industries and to a control group of firms that were not affected by the disaster for the period 2000–2018. Signaling theory implies that firms that proclaim commitment to CSR are committed to reducing misbehavior. In contrast, legitimacy theory suggests that prior misbehavior increases the likelihood of firms subsequently proclaiming CSR commitment on the surface but not reducing misbehavior. In both samples, we find evidence consistent with legitimacy theory. We find a significantly positive relation between all measures of CSR reporting and prior as well as future misbehavior. Both the number and severity of prior misbehaviors are positively associated with the decision to disclose a CSR report as well as its quality as captured by Bloomberg’s ESG disclosure score. On the other hand, the number and severity of future misbehaviors increase subsequent to CSR reporting. This indicates that firms expend higher effort on CSR reporting when they have a history of past misbehavior and expect future misbehavior. These findings are robust to different methods of addressing endogeneity and to alternative variable measurements, model specifications, and sample compositions. Likewise, the quasi-natural experiment 1808 C.Reitmaier et al. indicates that the signatories have significantly higher misbehavior pre and post 2013, compared to both peer groups. Signatories show an increase in misbehavior after signing the accord, compared to non-signatories of the same industry and a control group of non-textile firms. Our contribution is twofold: First, we contribute to the literature on the consequences of CSR reporting and greenwashing (e.g., Clarkson etal. 2008; Heflin and Wallace 2017) by focusing on irresponsible behavior identified by critical media reports to resolve measurement issues in prior research. Our results are consistent with the explanation that the firms’ proclaiming their commitment to CSR by voluntary CSR reporting or signing an accord for better working conditions is not a signal of internalized responsibility but more likely serves greenwashing purposes. These results are important because in many parts of the world there is an ongoing controversial debate about mandating CSR reporting (e.g., Christensen etal. 2021). Our analyses show that voluntary CSR reporting is not sufficient to alter firm behavior towards more responsibility and is linked to an increase in the number and severity of future misbehaviors. However, whether mandatory CSR reporting does better and achieves the goal of indirectly regulating firm behavior through public pressure is up to future research. First supportive evidence is limited to specific contexts like hydraulic fracturing (Bonetti etal. 2023b), greenhouse gas emissions (e.g., Downar etal. 2021; Grewal etal. 2023; Tomar 2023), and climate change risk disclosures under the SEC 2010 rule (Kim etal. 2023), or to positive CSR activities (Fiechter etal. 2022). Second, we contribute to the literature on the determinants of CSR reporting (e.g., Cormier etal. 2005; Thorne etal. 2014) by identifying prior misbehavior across the broad range of CSR dimensions as an additional determinant of CSR reporting. Knowledge about what motivates voluntary CSR reporting is increasing but still limited (e.g., Hahn and Luelfs 2014; Huang and Watson 2015). Evidence on prior misbehavior as a precedent to CSR reporting is limited to a few case studies (e.g., Bonetti etal. 2023a). Other related studies either restrict misbehavior measurement to the environmental dimension or measure performance and reporting at the same point in time (e.g., Neu etal. 1998; Haddock 2005). 2 Related research andhypotheses development There is a large literature on the consequences of voluntary CSR reporting, including capital market effects and real effects on firm behavior (e.g., Dhaliwal etal. 2011; Matsumura etal. 2014; see Christensen etal. 2021 for a review). These prior studies largely rely on signaling theory or legitimacy theory.4 According to signaling theory, responsible firms will signal their quality to the market to differentiate themselves from competitors (Akerlof 1970; Spence 1973; Morris 1987). By providing additional information that is hard to mimic, firms with superior CSR performance can signal their concern about and compliance with stakeholders’ expectations (Clarkson 4 Frynas and Yamahaki (2016) provide an overview of theories used in prior studies to explain CSR. 1809 Corporate responsibility andcorporate misbehavior: are… etal. 2008; Plumlee etal. 2015). In line with the stewardship theory of disclosure, additional information improves stakeholders’ monitoring and hence the alignment of managerial decisions with stakeholders’ interests (Lambert etal. 2007). The use of measurement and control systems for CSR reporting increases firms’ awareness of misbehavior and promotes internal self-regulation and responsible corporate governance (Topping 2012; Wu 2014; Adams and Abhayawansa 2022). Standard setters consider disciplining mechanisms the main reason for mandating CSR reporting (e.g., EC 2021). Yet, it is unclear whether reporting can indeed achieve the aspired real effects. First supportive evidence on mandatory reporting is limited to specific cases of CSR. However, it is unclear whether voluntary CSR reporting would not also be sufficient to achieve the desired real effects (Christensen etal. 2021; Fiechter etal. 2022). Legitimacy theory, in contrast, argues that firms use voluntary disclosures to legitimize corporate activities (Patten 1991; Cho and Patten 2007). In particular, firms with poor prior CSR performance might face strong public pressure and, hence, report on CSR efforts to prevent or reduce reputational and legitimacy damages (Lindblom 1994). Managers may use CSR reporting opportunistically for impression management and greenwashing purposes despite the potential costs of getting caught (Merkl-Davies and Brennan 2007; Mishina etal. 2010). Numerous studies have shown that CSR reporting is often unbalanced and self-laudatory and neglects negative incidents (Wiseman 1982; Holder-Webb etal. 2009). Prior empirical studies are inconclusive about the relation between voluntary CSR reporting and CSR performance (Christensen etal. 2021). To measure CSR performance, prior studies focus on environmental aspects (e.g., Neu etal. 1998; Kim and Lyon 2011), aggregate responsible and irresponsible behavior (e.g., Schultze and Trommer 2012; Mahoney etal. 2013; Lee and Maxfield 2015), and measure variables of interest at the same point in time without considering lead-lag relations. The literature on the consequences of CSR reporting on misbehavior has focused on the issuance of a CSR report to capture CSR reporting, without considering the quality of the reporting (e.g., Christensen 2016). Other studies, however, have shown that the quality of CSR disclosures is informative in establishing a relation between CSR reporting and, for instance, firm value (Plumlee etal. 2015). To signal their CSR commitment or, alternatively, to legitimize their behavior, firms may not only decide to issue a voluntary CSR report but also expend higher effort for better CSR disclosure quality. We hence also consider CSR disclosure quality in our analyses. Some studies find support for signaling theory (e.g., Al-Tuwaijri et al. 2004; Mahoney etal. 2013); others support legitimacy theory (e.g., Kim and Lyon 2011; Cho et al. 2012). Even within single studies, the results are not conclusive. For instance, Clarkson etal. (2008) find a positive relation between CSR disclosures and CSR performance, which supports signaling theory, but they also find that legitimacy theory explains patterns in these disclosures. Explanations for the inconclusive results include inadequate sample selection, omitted or confounding variables measuring CSR performance (Patten 2002; Christensen etal. 2021), and, in particular, aggregating responsible and irresponsible behaviors (Mattingly and Berman 2006; Strike etal. 2006; Semenova and Hassel 2015). To measure responsibility in 1810 C.Reitmaier et al. the sense of avoiding misbehavior, it is important to distinguish irresponsible from responsible behavior (McGuire etal. 2003; Strike etal. 2006; Krüger 2015). To the best of our knowledge, only Christensen (2016) does not use an aggregate measure of CSR performance; he instead uses CSR-related lawsuits to capture misbehavior. He finds that voluntarily issuing a CSR report reduces the likelihood of CSR-related lawsuits in the following year, and concludes that after CSR reporting, firms improve their CSR performance due to an internalization of proclaimed commitments. However, the negative association of CSR reporting and CSR-related lawsuits is not an unambiguous indication that CSR reporting indeed improves the firms’ behavior. An alternative explanation, which is not ruled out in his study, is that the voluntary issuance of a CSR report provides a mechanism of successful image improvement, such that legal prosecution decreases without real improvements in corporate behavior. In a similar vein, Raghunandan and Rajgopal (2023) study the relation between firms’ signing the Business Roundtable’s “Statement on the Purpose of a Corporation” and violations of US federal laws. As in Christensen (2016), their dependent variable requires a legal investigation, but such investigations are rare (MarkUngericht and Weiskopf 2007) and may be biased by the firms’ political connections (McCarten etal. 2022). Most cases of corporate misbehavior will not result in a federally reported violation. In particular, cases of misbehavior that occur in foreign subsidiaries that are not subject to US laws will go unnoticed. In contrast, the public media plays a vital role in revealing corporate misbehavior and shaping stakeholders’ perceptions (Einwiller etal. 2010; Dube and Zhu 2021). In its desire to publish attention-provoking pieces, the media has the incentives, resources, and scrutiny to reveal misbehavior. Therefore, to capture firms’ overall accountability to the public, we measure corporate misbehavior based on critical media reports across the broad range of CSR dimensions. We consider misbehavior before and after CSR reporting to directly address lead-lag relations, which helps to identify causal relations (Granger 1969). Regarding the relation between CSR reporting and prior misbehavior, signaling theory implies that responsible firms benefit from CSR reporting by signaling that they have learned from past mistakes. Stakeholders are particularly vigilant in observing the activities of previously irresponsible firms, such that strategic costs increase (DeTienne and Lewis 2005). To reduce such costs, responsible firms have incentives to signal their type. Legitimacy theory argues that firms use CSR reporting after the revelation of misbehavior to respond to increased public pressure. To legitimize themselves, firms may expend higher effort to issue a CSR report or improve the report’s quality as captured by the public’s eye. Evidence of this is limited to a few case studies. Jantadej and Kent (1999) show that the mining firm Broken Hill published more environmental information after the disaster at Ok Tedi Mine. Others find similar effects after the Exxon Valdez oil spill (Patten 1992), the BP oil spill (Heflin and Wallace 2017), and the Fukushima nuclear disaster (Bonetti etal. 2023a, b).5 Both theories imply a positive relation between prior misbehavior and 5 Christensen (2016) includes prior misconduct, measured as publicized lawsuits, as a control variable in his analysis and finds insignificant results. 1811 Corporate responsibility andcorporate misbehavior: are… CSR reporting. However, irresponsible firms may not have sufficient favorable CSR information to be willing to issue a CSR report or improve their CSR disclosure quality, implying a negative relation. Based on signaling and legitimacy theory, we hypothesize: H1: Prior corporate misbehavior is positively associated with CSR reporting. Regarding the effects of CSR reporting on subsequent firm behavior, the arguments of signaling and legitimacy theory differ. Signaling theory implies that firms expect capital market benefits from improved CSR disclosure and better CSR information, which, in turn, improves monitoring and hence firm behavior. This is the rationale that standard setters are following (e.g., Christensen 2016; EC 2021) and suggests that firms that voluntarily issue CSR reports or spend higher efforts to improve CSR disclosure quality will internalize their CSR commitment and improve their behavior. However, firms have no strong incentives to change their behavior. Sanctions on euphemistic reporting and misbehavior are largely absent (Mark-Ungericht and Weiskopf 2007). Prior evidence suggests that investors only show very limited reactions to voluntary CSR reports (Yoon andSerafeim 2020; Burzillo etal. 2023). Firms may hence use CSR reporting to improve their legitimacy and reputation without changing their activities (Lindblom 1994). Even higher CSR disclosure quality, as measured by publicly available ESG scores, does not reflect internal improvements in the responsibility of operations but largely represents the fulfilment of information requirements by the rating agencies (Ahmed et al. 2023). If CSR reporting helps firms to successfully improve their public image, it may lead to less legal scrutiny and investor activism but not necessarily to less misbehavior. Based on legitimacy theory, we hypothesize: H2: CSR reporting is positively associated with future corporate misbehavior. 3 Measurement ofCSR reporting andcorporate misbehavior We measure CSR reporting (CSRRi,t) in two ways: Firstly, we use an indicator variable for issuing a voluntary stand-alone GRI-based CSR report (CSRR_Ii,t), as in prior research (e.g., Simnett etal. 2009; Dhaliwal etal. 2011; Christensen 2016).6 We collect data from the GRI Database, which covers GRI reports since 1999. We use CorporateRegister, the firms’ websites, and internet searches to verify the data 6 The GRI guidelines provide a CSR reporting framework of high quality (e.g., Christensen 2016) and are the most frequently used (e.g., KPMG 2022). The GRI does not focus solely on financial materiality but extends to corporate impacts on sustainable development. Such double materiality is central to current debates and developments involving mandatory non-financial reporting (Adams and Abhayawansa 2022). Despite recent attempts to integrate CSR topics into annual reports (e.g., the introduction of the Integrated Reporting Framework), stand-alone reports are still the most common form of CSR reporting (e.g., KPMG 2008, 2011, 2017, 2022). We differentiate neither between the GRI versions G2 to G4 nor between the different application levels (A + to C, undeclared, GRI-referenced). 1818 C.Reitmaier et al. Rusticus 2010). Several studies use firm age as a valid IV (Harjoto and Jo 2011; Christensen 2016). Others use firm age as a proxy for reputation (e.g. Datta etal. 1999). Firms with a higher reputation tend to disclose more to maintain reputation (Reitmaier and Schultze 2017). Older firms have more experience and lower reporting costs (Camfferman and Cooke 2002; Alsaeed 2006), suggesting a positive relation of AGEi,t to CSR reporting. In contrast, there is no relation to future misbehavior. Misbehavior depends on factors that develop independently of firm age, like corporate culture or competition. Literature on accounting fraud supports the irrelevance of firm age with respect to misbehavior since fraud firms do not differ in age compared to non-fraud firms (Farber 2005; Chen etal. 2006). Our second IV is FFLOATi,t. High free float comes with high agency costs (Fama and Jensen 1983; Prencipe 2004). To reduce these costs and to announce compliance, managers use voluntary disclosures, implying a positive relation of FFLOATi,t with CSR reporting (Brammer and Pavelin 2006). On the other hand, a direct relation between FFLOATi,t and corporate misbehavior is not intuitive. Chen et al. (2006) find that boardroom characteristics are more relevant in explaining fraud than ownership characteristics. Other studies find no significant relation between financial fraud and blockholders (Larcker etal. 2007; Johnson etal. 2009). We use two further IVs in additional analyses that have been used in prior research23: The dichotomous variable DJSIi,t-1 indicates DJSI listing in t-1, which positively relates to nonfinancial disclosures (Cho etal. 2012). Listed firms are more likely to issue a CSR report, as related disclosures are required for continued listing (Dhaliwal etal. 2012). In contrast, there is no evidence for a relation between DJSI listing in t-1 and corporate misbehavior in t + 1. The dichotomous variable CROSSi,t indicates cross-listing at three or more stock exchanges in t. Additional listings imply additional regulatory and reporting requirements, which increase voluntary disclosures. In contrast, cross-listing does not subject firms to the CSR regulations or laws of that country and hence does not affect misbehavior through legal bonding (Shi etal. 2018). 4.2.3 Second stage SR model The SR method does not require a complete set of IVs and fits nonlinear models.24 In comparison to the 2SLS model, it requires an additional regressor (V) in the second stage with the following characteristics (Lewbel 2014): V is exogenous and continuously distributed (Lewbel etal. 2012). V has a positive coefficient and 23 For robustness purposes, we repeat our analyses with all combinations of three or two of the four IVs and also analyze alternative IVs (Sect.7.2). 24 Lewbel (2014, 38) explains the SR method as follows: “For example, suppose an observed binary variable D satisfies D = I (V + W* ≥ 0), where V is the observed special regressor, and W* is an unobserved latent variable. […] Special regressor methods work by exploiting the fact that if V is independent of W* time. Likewise, we do not use industry aggregates since the endogenous part of the reporting decision varies not only across industries but also within industries. The use of ranked endogenous regressors would only transfer endogeneity from the original regressors to the ranked ones. Footnote 22 (continued) 1819 Corporate responsibility andcorporate misbehavior: are… is monotonous (Dong and Lewbel 2015).25 We include the average USD-EUR exchange rate (EXCHANGEt) as V in model (2). We use all different measures of future and prior misbehavior and CSR reporting, as outlined in Sect.326: USD and EUR are the most important currencies worldwide (e.g., Norrlof 2009). The US and Europe are the most strongly represented regions in our sample. The USD/EUR exchange rate is exogenous.27 While it does not influence CSR reporting, it might influence future misbehavior since exchange rates influence the competitive environment in which a firm operates. Exchange rate fluctuations induce additional risk for internationally operating firms (Kotabe and Murray 2004). The yearly average exchange rate is continuously distributed. 5 Sample Our main sample is based on all 814 firms listed in the Fortune Global 500 in the period from January 1, 2003, to December 31, 2013. The Fortune Global 500 lists the world’s largest firms in terms of total sales and is frequently used in CSR studies because of its global impact (e.g., Muller and Whiteman 2009; Lee and Maxfield 2015).28 The sample period starts in 2003, which marks the onset of intense voluntary CSR reporting due to rising public and governmental interest in CSR (Freundlieb and Teuteberg 2013). The sample period ends in 2013 for two reasons. First, the European Parliament and the Council issued EU directive 2014/95/EU in October 2014, which rendered CSR reporting mandatory as of January 2017 for large European firms. Early adopters might bias our analyses that explicitly focus (4) FutureCIR i,t+1 =𝛽 0 +𝛽 1 CSRR i,t +𝛽 2 UNGC i,t+1 +𝛽 3 PO i +𝛽 4 HO i +𝛽 5 GLOBAL i,t+ 1 +𝛽6PriorCIRi,t−3∕Σ3y+𝛽7SIZEi,t+1+𝛽8ROEi,t +𝛽 9 EXCHANGE t +IND +YEAR +𝜀 i,t+1 25 Dong and Lewbel (2015) relax the additional requirement of great support for V. 26 FutureCIRi,t+1 = CIR_Ii,t+1, CIR_Ni,t+1, or CIR_Si,t+1; PriorCIRi,t-3/∑3y = CIR_Ii,t-3/∑3y, CIR_Ni,t-3/∑3y, or CIR_Si,t-3/∑3y; and CSRRi,t = CSRR_Ii,t or CSRR_Qi,t. The first stage regression remains basically unchanged. Since it includes all determinants and control variables of the second stage regression, it also includes V in the SR approach. 27 IVs and the SR need to be exogenous in the respective model. That is, the IVs in model (3) need to correlate with CSRRi,t but not FutureCIRi,t+1, and the SR in model (4) needs to correlate with FutureCIRi,t+1 but not CSRRi,t. Polynomial regressions of future misbehavior (CIR_Ii,t+1, CIR_Ni,t+1, or CIR_Si,t+1) on the SR confirm a positive coefficient and monotonic graphical appearance (Dong and Lewbel 2015; Bontemps and Nauges 2016). 28 Analyzing only the largest firms reduces potential concerns about the disproportionate media coverage of larger firms and improves the likelihood that all cases of misbehavior are detected by public media and hence the CCRD. (either unconditionally or after conditioning on covariates) then variation in V changes the probability that D = 1 in a way that traces out the distribution of W* (either the unconditional distribution or the distribution conditional on covariates).”. Footnote 24 (continued) 1820 C.Reitmaier et al. on voluntary CSR reporting. Second, the International Integrated Reporting Council published the Integrated Reporting Framework in December 2013. Adopting firms no longer issue stand-alone CSR reports but generally integrate CSR reports, annual reports, and other reports into one report, which our GRI measure does not capture. We exclude 127 firms with missing data in Datastream or Eikon, 96 firms not covered by the CCRD, and 104 firms with missing data in other sources (UNGC homepage, DJSI or Interbrand ranking), resulting in a sample of 487 unique firms. We further exclude 858 firm-year observations due to bankruptcy or mergers and acquisitions. We do not exclude those firms completely to avoid survivorship biases. Our final main sample includes 4,499 firm-year observations. We collect data on misbehavior for the time before and after voluntary CSR reporting (2000–2014). As Bloomberg’s ESG disclosure score does not cover all firms over the entire period, the analyses refer to a reduced sample of 3,150 firm-year observations. Table 1 panel A shows the distribution of our main sample firms per ICB supersector and the proportion of firms with misbehavior (CIR_Ii,t) per supersector. CIR_Ii,t is highest in the supersectors Food & Beverage (ICB code 3500), Personal & Household Goods (3700), and Technology (9500), where more than 50 percent of the firms engage in misbehavior. At 17.65 percent, CIR_Ii,t is lowest in the supersector Construction & Materials (2300). Panel B presents the firms’ country of origin and shows that more than one-third of the firms are from the US (36.96 percent), followed by Japan (14.78), the UK (8.42), France (7.19), and Germany (5.13). The remaining 27.5 percent are in 22 different countries. The percentage of firm-year observations with misbehavior is highest in Turkey (100 percent, one firm with misbehavior in each year) and Taiwan (71.43 percent). It is lowest in Austria and Colombia (0 percent each, one/two firms without misbehavior). Table 2 presents the development of the number of firms (not) issuing a GRI report in a particular year (CSRR_Ii,t = 0/1) over our sample period. The third (fourth) column presents the corresponding number of firms with (without) misbehavior in the following year (CIR_Ii,t+1 = 0/1). The number of firms issuing a GRI report increased from 58 in 2003 to 262 in 2013. The number of firms with subsequent misbehavior does not follow a systematic trend; the levels are similar in the first and last sample year (145 in 2003 and 157 in 2013). This suggests that the increase in CSR reporting does not come with a systematic reduction in future misbehavior. Table 3 provides the descriptive statistics. For ease of interpretation, the values for the number (CIR_Ni,t-3) and severity of current (CIR_Si,t), prior (CIR_Ni,t-3, CIR_ Ni,∑3y, CIR_Si,t-3, or CIR_Si,∑3y), and future misbehaviors (CIR_Ni,t+1 or CIR_Si,t+1) are raw data before logtransformation.29 Forty-three percent of the observations are GRI reporters in t. Fifty-three percent issue some kind of CSR report in t. The average firm 29 To adjust for its strongly skewed distribution, we logtransform our measures of the number and severity of misbehaviors in our empirical analyses (see Sect.3 for details). 1821 Corporate responsibility andcorporate misbehavior: are… Table 1 Classification of sample firms Industry Supersector Number of Firms %of Firms per CIR_Ii,t Supersector Industry No Misbehavior Misbehavior % Panel A: Sample firms per ICB supersector 0001 Oil & Gas 0500 Oil & Gas 41 8.42 8.42 247 116 31.96 1000 Basic Materials 1300 Chemicals 12 2.46 6.37 76 31 28.97 1700 Basic Resources 19 3.90 116 52 30.95 2000 Industrials 2300 Construction & Materials 21 4.31 18.48 182 39 17.65 2700 Industrial Goods & Services 69 14.17 519 180 25.75 3000 Consumer Goods 3300 Automobiles & Parts 29 5.95 13.55 226 62 21.53 3500 Food & Beverage 15 3.08 67 76 53.15 3700 Personal & Household Goods 22 4.52 111 114 50.67 4000 Health Care 4500 Health Care 22 4.52 4.52 122 77 38.69 5000 Consumer Services 5300 Retail 43 8.83 13.14 255 130 33.77 5500 Media 9 1.85 61 18 22.78 5700 Travel & Leisure 12 2.46 70 30 30.00 6000 Telecommunications 6500 Telecommunications 15 3.08 3.08 91 32 26.02 7000 Utilities 7500 Utilities 38 7.80 7.80 247 86 25.83 8000 Financials 8300 Banks 48 9.86 18.48 326 107 24.71 8500 Insurance 31 6.37 215 67 23.76 8600 Real Estate 0 0 0 0 - 8700 Financial Services 11 2.26 46 26 36.11 9000 Technology 9500 Technology 30 6.16 6.16 136 143 51.25 Total 487 100 100 3.113 1.386 30.81 1822 C.Reitmaier et al. Table 1 (continued) Nation Number of Firms % of Firms CIR_Ii,t No Misbehavior Misbehavior % Panel B: Sample firms per country Australia 10 2.05 71 19 21.11 Austria 2 0.41 20 0 0.00 Brazil 3 0.62 17 4 19.05 Canada 18 3.70 128 19 12.93 Colombia 1 0.21 1 0 0.00 Denmark 2 0.41 15 1 6.25 Finland 5 1.03 38 11 22.45 France 35 7.19 236 109 31.59 Germany 25 5.13 160 102 38.93 Hong Kong 3 0.62 27 6 18.18 India 7 1.44 71 4 5.33 Ireland 3 0.62 26 5 16.13 Italy 11 2.26 82 12 12.77 Japan 72 14.78 532 142 21.07 Mexico 2 0.41 16 4 20.00 Netherlands 11 2.26 56 49 46.67 Portugal 2 0.41 6 1 14.29 Singapore 2 0.41 17 1 5.56 South Korea 17 3.49 87 32 26.89 Spain 12 2.46 106 21 16.54 Sweden 6 1.23 56 10 15.15 Switzerland 11 2.26 80 40 33.33 Taiwan 4 0.82 10 25 71.43 1823 Corporate responsibility andcorporate misbehavior: are… Table 1 (continued) Thailand 1 0.21 8 1 11.11 Turkey 1 0.21 0 6 100.00 United Kingdom 41 8.42 140 228 61.96 United States 180 36.96 1.107 534 32.54 Total 487 100 3.113 1.386 30.81 Presents the distribution of sample firms across ICB codes and geographic regions. Panel A shows the absolute and percental distribution of sample firms across ICB supersectors (SP) and the percental distribution of sample firms across ICB industries (IN). In addition, it shows the absolute number of firm-year observations per ICB supersector without (CIR_Ii,t = 0) and with misbehavior (CIR_Ii,t = 1) as well as the percentage of firm-year observations with misbehavior per ICB supersector (CIR_ Ii,t = 1). Panel B shows the absolute and percental distribution of sample firms across countries. In addition, it shows the absolute number of firm-year observations per country without (CIR_Ii,t = 0) and with misbehavior (CIR_Ii,t = 1) as well as the percentage of firm-year observations with misbehavior per country (CIR_Ii,t = 1) 1824 C.Reitmaier et al. has a Bloomberg ESG disclosure score of 40.32 percent. On average, almost one-third of the sample firms conduct misbehavior (CIR_Ii,t). The mean of 0.94 for CIR_Ii,∑3y indicates that firms engage in misbehavior in about one of the past three years on average. The average firm conducts 0.54 cases of misbehavior per year; the maximum number in t and t + 1 is 23 (CIR_Ni,t, CIR_Ni,t+1). The maximum number of misbehaviors in t-3 is 30 (CIR_Ni,t-3). The average firm conducts 1.71 cases of misbehavior over the prior three years; the maximum number is 44 (CIR_Ni,∑3y). The maximum severity score of misbehavior is 93.00; the mean is 2.37 in t (CIR_Si,t), 2.36 in t + 1 (CIR_Si,t+1), and 2.61 in t-3 (CIR_Si,t-3). The maximum sum over the prior three years is 226.00 (CIR_Si,∑3y). The severity measure based on the Refinitiv controversy score ranges from zero to one with a mean of 0.33 in t (CIR_Ri,t), 0.32 in t + 1 (CIR_Ri,t+1), and 0.35 in t-3 (CIR_Ri,t-3). Notably, around one-third of the observations show DJSI listing (mean of dichotomous variable DJSIi,t = 0.33) and UNGC membership in t (mean = 0.31). Table 4 presents the results of difference in means tests. Panel A presents the determinants of GRI reporting for reporting vs. non-reporting firms. The difference is significant for all variables except ROEi,t. The mean of all measures of prior misbehavior is significantly higher for GRI reporters than non-reporters. This indicates that CSR reporting increases after misbehavior, consistent with H1. Panel B shows the determinants of future misbehavior for firms with and without misbehavior in t + 1. The difference in means is significant for all variables. The mean of CSRR_ Ii,t and CSRR_Qi,t is significantly higher for firms with future misbehavior than for firms without future misbehavior. Consistent with H2, this indicates a positive relation between CSR reporting and future misbehavior. Table5 shows significant correlations between our measures of CSR reporting and prior (panel A) as well as future corporate misbehavior (panel B), consistent with H1 and H2. Table 2 Sample firms with/without GRI reports and future misbehavior per year Presents the distribution of sample firms with and without GRI reports per year (CSRR_Ii,t) and the distribution of firms with and without at least one case of misbehavior in the following year t + 1 (CIR_Ii,t+1) Year CSRR_Ii,t CIR_Ii,t+1 Reporter Non-reporter Misbehavior No misbehavior 2003 58 332 145 245 2004 88 306 184 210 2005 106 301 143 264 2006 119 283 132 270 2007 147 251 111 287 2008 195 216 100 311 2009 213 204 117 300 2010 228 187 130 285 2011 254 171 101 324 2012 263 160 100 323 2013 262 155 157 260 Total 1,933 2,566 1,420 3,079 1825 Corporate responsibility andcorporate misbehavior: are… Table 3 Descriptive statistics Variable N Mean SD Min Median Max CSR reporting CSRR_Ii,t 4,499 0.43 0.50 0.00 0.00 1.00 CSRR_I2i,t 4,499 0.53 0.50 0.00 1.00 1.00 CSRR_Qi,t 3,150 40.32 11.31 1.98 39.93 74.33 CSR reporting history PriorGRI_Ii,t 4,499 0.47 0.50 0.00 0.00 1.00 PriorGRI_Ni,t 4,499 1.84 2.68 0.00 0.00 14.00 CSRR_Ii,t-3 4,499 0.28 0.45 0.00 0.00 1.00 Corporate misbehavior CIR_Ii,t 4,499 0.31 0.46 0.00 0.00 1.00 CIR_Ni,t 4,499 0.54 1.25 0.00 0.00 23.00 CIR_Si,t 4,499 2.37 5.38 0.00 0.00 93.00 CIR_Ri,t 3,986 0.33 0.37 0.00 0.16 1.00 Future corporate misbehavior CIR_Ii,t+14,499 0.32 0.46 0.00 0.00 1.00 CIR_Ni,t+14,499 0.55 1.24 0.00 0.00 23.00 CIR_Si,t+14,499 2.36 5.30 0.00 0.00 93.00 CIR_Ri,t+13,982 0.32 0.36 0.00 0.16 1.00 Prior corporate misbehavior CIR_Ii,t-3 4,499 0.31 0.46 0.00 0.00 1.00 CIR_Ii,∑3y 4,499 0.94 1.14 0.00 0.00 3.00 CIR_Ni,t-3 4,499 0.58 1.41 0.00 0.00 30.00 CIR_Ni,∑3y 4,499 1.71 3.45 0.00 0.00 44.00 CIR_Si,t-3 4,499 2.61 5.97 0.00 0.00 93.00 CIR_Si,∑3y 4,499 7.68 15.84 0.00 2.00 226.00 CIR_Ri,t-3 3,194 0.35 0.40 0.00 0.16 1.00 Further variables AGEi,t 4,499 35.00 26.99 0.00 30.00 156.00 BRANDi,t 4,499 0.14 0.35 0.00 0.00 1.00 CROSSi,t 4,499 0.41 0.49 0.00 0.00 1.00 DJSIi,t 4,499 0.33 0.47 0.00 0.00 1.00 DJSIi,t-1 4,499 0.33 0.47 0.00 0.00 1.00 EXCHANGEt4,499 0.77 0.05 0.68 0.76 0.88 FFLOATi,t 4,499 80.01 21.71 0.00 88.00 100.00 GLOBALi,t 4,499 41.03 32.42 -73.47 39.48 689.70 GLOBALi,t+14,499 41.53 30.70 -73.47 41.13 215.62 HOi4,499 3.98 0.38 3.18 4.17 4.96 LEGALi4,499 0.46 0.50 0.00 0.00 1.00 LEVi,t 4,499 26.04 15.08 0.00 24.75 98.92 MSHAREi,t 4,499 -7.26 7.55 -59.00 -4.89 -0.00 POi4,499 4.32 0.25 3.58 4.36 4.94 RDi,t 4,487 712,730.20 1,583,548.00 0.00 1,029.00 1.36e + 07 1826 C.Reitmaier et al. 6 Empirical results 6.1 Prior corporate misbehavior andCSR reporting Table6presents the results for model (1), which tests H1 (that prior corporate misbehavior and voluntary CSR reporting are positively associated). Panel A shows the results of a logistic regression on CSRR_Ii,t based on our six measures of PriorCIRi,t-3/∑3y, displayed in column (A) trough (F). In all columns, the McFadden pseudo-R2 is about 0.26, which is similar to related studies (e.g., Dhaliwal etal. 2011; Christensen 2016) and indicates that the regression is well-specified. The goodness of fit (GOF) test is not significant, indicating that the model fits the data well (Kleinbaum and Klein 2010). The area under the ROC curve is higher than 82 percent. The model classifies about 75 percent of reporting and non-reporting firms correctly and is highly significant (p < 0.01). Overall, the quality of our model specifications is very similar to related studies (Brammer and Pavelin 2006; Legendre and Coderre 2013).30 All variance inflation factors (VIFs) are below the conservative threshold of 5. Robust standard errors control for heteroscedasticity. Industry and year fixed effects control for industry and time differences. All measures of prior misbehavior are significantly positively related to CSRR_ Ii,t. The odds ratios indicate that CSR reporting is between 6 and 22 percent more likely for firms with prior misbehavior than for their peers (CIR_Ii,t-3: 1.197 (p < 0.05), CIR_Ii,∑3y: 1.084 (p < 0.05), CIR_Ni,t-3: 1.223 (p < 0.05), CIR_Ni,∑3y: 1.165 (p < 0.01), CIR_Si,t-3: 1.096 (p < 0.05), CIR_Si,∑3y: 1.064 (p < 0.1)). The effect Table 3 (continued) Variable N Mean SD Min Median Max ROAi,t 4,488 5.12 5.64 -38.96 4.38 62.35 ROEi,t 4,499 15.24 64.94 -662.10 13.57 3,821.40 SIZEi,t 4,499 16.86 1.15 9.32 16.85 20.04 SIZEi,t+14,499 16.91 1.13 9.88 16.90 20.29 STATEi,t 4,499 2.24 10.48 0.00 0.00 93.50 UNGCi,t 4,499 0.31 0.46 0.00 0.00 1.00 UNGCi,t+14,499 0.33 0.47 0.00 0.00 1.00 VOLAi,t 4,499 26.06 8.14 6.65 25.01 73.02 Presents the descriptive statistics for all variables used in our main and robustness analyses for our main sample of Fortune Global 500 firms over the period 2003–2013. Values for the number and severity of current (CIR_Ni,t, CIR_Si,t), future (CIR_Ni,t+1, CIR_Si,t+1), and prior misbehaviors (CIR_Ni,t-3, CIR_ Ni,∑3y, CIR_Si,t-3, CIR_Si,∑3) are raw data before logtransformation. Since these data are strongly skewed, they are logtransformed in our empirical tests (Tables4, 5, 6, 7and 8). All variables are defined as in Appendix3 30 Some studies calculate other quality measures like the Nagelkerke R2 or the Cox & Snell R2. Values of about 0.41 and 0.30 for all our model specifications are in an uncritical range. 1827 Corporate responsibility andcorporate misbehavior: are… Table 4 Difference in means tests Presents the difference in means tests. Panel A shows the difference in means tests of the determinants of CSR reporting (CSRR_Ii,t). Panel B shows the difference in means tests of the determinants of future misbehavior (CIR_Ii,t+1). Values for CIR_Ni,t-3, CIR_Ni,∑3y, CIR_Si,t-3, and CIR_Si,∑3y are logtransformed as in all empirical tests. All variables are defined as in Appendix3 * , **, *** indicate that the estimated differences in means are statistically significant at the 10 percent, 5 percent, and 1 percent level, respectively, using a two-tailed test Panel A: Difference in means tests of the determinants of CSR reporting (CSRR_Ii,t = 0/1) All GRI reporter Non-reporter Variable Mean Mean Mean Diff. in means t CIR_Ii,t-3 0.3136 0.3678 0.2728 -0.0950 -6.8341*** CIR_Ii,∑3y 0.9438 1.0879 0.8352 -0.2528 -7.4151*** CIR_Ni,t-3 0.2932 0.3550 0.2466 -0.1083 -7.4254*** CIR_Ni,∑3y 0.6305 0.7344 0.5523 -0.1821 -7.9941*** CIR_Si,t-3 0.7107 0.8306 0.6203 -0.2103 -7.4962*** CIR_Si,∑3y 1.3064 1.4653 1.1867 -0.2786 -7.4268*** DJSIi,t 0.3350 0.4744 0.2300 -0.2445 -17.7888*** UNGCi,t 0.3121 0.4904 0.1777 -0.3127 -23.7712*** AGEi,t 34.9953 39.5820 31.5401 -8.0419 -10.0016*** CROSSi,t 0.4072 0.4775 0.3542 -0.1232 -8.3924*** BRANDi,t 0.1445 0.1945 0.1068 -0.0877 -8.3481*** MSHAREi,t -7.2645 -7.6841 -6.9484 0.7357 3.2397*** STATEi,t 2.2394 2.6453 1.9336 -0.7117 -2.2568** LEGALi0.4614 0.5732 0.3772 -0.1960 -13.3036*** FFLOATi,t 80.0113 83.3963 77.4614 -5.9349 -9.1604*** VOLAi,t 26.0618 25.4587 26.5162 1.0575 4.3194*** LEVi,t 26.0360 26.7908 25.4675 -1.3233 -2.9154*** SIZEi,t 16.8573 17.1238 16.6565 -0.4673 -13.8026*** ROEi,t 15.2401 14.0574 16.1310 2.0736 1.0603 Panel B: Difference in means test of the determinants of future corporate misbehavior (CIR_Ii,t+1 = 0/1) All Future misbehavior No future misbehavior Variable Mean Mean Mean Diff. in means t CSRR_Ii,t 0.4297 0.4908 0.4014 -0.0894 -5.6496*** CSRR_Qi,t 40.3241 13.1035 39.1184 -3.9851 -9.2032*** UNGCi,t+10.3343 0.3697 0.3180 -0.0518 -3.4240*** POi4.3190 4.3290 4.3144 -0.0146 -1.8455* HOi3.9799 3.9367 3.9998 0.0631 5.2324*** GLOBALi,t+141.5294 49.2299 37.9781 -11.2518 -11.5956*** CIR_Ii,t-3 0.3136 0.5648 0.1977 -0.3670 -26.5095*** CIR_Ii,∑3y 0.9438 1.8183 0.5404 -1.2779 -40.9976*** CIR_Ni,t-3 0.2932 0.5751 0.1631 -0.4120 -28.6581*** CIR_Ni,∑3y 0.6305 0.3637 1.2090 -0.8453 -40.3836*** CIR_Si,t-3 0.7107 1.2714 0.4521 -0.8193 -29.8303*** CIR_Si,∑3y 1.3064 2.2136 0.8880 -1.3255 -37.8636*** SIZEi,t+116.9086 17.4152 16.6750 -0.7402 -21.4503*** ROEi,t 15.2401 17.6670 14.1208 -3.5462 -1.7027* 1834 C.Reitmaier et al. Table 6 (continued) ROEi,t ? 0.001 1.45 0.001 1.25 0.001 1.37 Constant -18.260 -3.78*** -13.760 -2.87*** -13.583 -2.82*** Industry fixed effects Included Included Included Year fixed effects Included Included Included N 3,150 3,150 3,150 Adjusted R20.395 0.393 0.393 Highest VIF 2.45 2.48 2.52 Mean VIF 1.51 1.49 1.51 Presents the regression results for model (1). Panel A uses CSRR_Ii,t as the dependent variable and hence refers to the association between prior corporate misbehavior and the voluntary issuance of a stand-alone GRI report in t. It shows the odds ratios, coefficients, and z-statistics from pooled logistic regressions with industry and year fixed effects. Columns (A) to (F) differ in the measures of prior misbehavior: Column (A) includes CIR_Ii,t-3. Column (B) includes CIR_Ii,∑3y. Column (C) includes CIR_Ni,t-3. Column (D) includes CIR_Ni,∑3y. Column (E) includes CIR_Si,t-3. Column (F) includes CIR_Si,∑3y. Panel B uses CSRR_Qi,t as dependent variable and hence refers to the association between prior corporate misbehavior and CSR disclosure quality in t. It shows the coefficients and t-statistics from OLS regressions with industry and year fixed effects. We do not run logistic regressions here since, in contrast to CSRR_Ii,t, CSRR_Qi,t is not binary. As in panel (A), columns (A) to (F) differ in the measures of prior misbehavior. In both panels, values for CIR_Ni,t-3, CIR_Ni,∑3y, CIR_Si,t-3, and CIR_Si,∑3y are logtransformed as in all empirical tests. All variables are defined as in Appendix3 *, **, *** indicate that the estimated coefficients are statistically significant at the 10 percent, 5 percent, and 1 percent level, respectively, using a two-tailed test 1835 Corporate responsibility andcorporate misbehavior: are… Bloomberg.32 Overall, our findings support H1 and imply that firms use CSR reporting to signal that they have learned from past mistakes, or, alternatively, that firms greenwash past misbehavior to reduce reputational damages according to legitimacy theory. 6.2 CSR reporting andfuture corporate misbehavior Table 7 presents the results for model (2), which tests H2 on the relation of CSR reporting with future misbehavior.33 Panel A columns (A) through (F) show the results for the decision to issue a CSR report (CSRR_Ii,t) and our three measures of future misbehavior CIRi,t+1, with two alternative prior period measures of misbehavior as a control. The Durbin-Wu-Hausman test indicates endogeneity related to CSRR_Ii,t (p < 0.01/0.05) in all cases. The centered R2 is between 0.10 and 0.49. The models are highly significant (p < 0.01). AGEi,t and FFLOATi,t fulfill the criterion of relevance, as simple correlations (untabulated) and multivariate tests (Tables6and 7) indicate significant relations to CSRR_Ii,t.34 Regarding exogeneity, AGEi,t and FFLOATi,t show no economically meaningful correlations with any measure of FutureCIRi,t+1 (between 0.03 and 0.08).35 The Kleinbergen-Papp F statistic of the first stage is between 46 and 48, which is above the critical value of 19.93 for a model with one endogenous regressor and two IVs (Stock and Yogo 2002; Bascle 2008).36 We can reject weak identification and associated asymptotic biases based on these results (Bascle 2008) and confirm the relevance and suitability of the IVs. We find a significantly positive association of CSRR_Ii,t with all measures of FutureCIRi,t+1 (p < 0.01/0.05), consistent with H2 and legitimacy theory. In all columns, we document a positive association between past and future misbehavior. Column (G) analyzes the incremental effect of the severity of prior misconduct in more detail. We include the interaction of CIR_ Ii,∑3y and CIR_Si,∑3y in model (2) with future misconduct CIR_Si,t+1 as the dependent variable. We continue to find a significantly positive association of CSRR_Ii,t with CIR_Si,t+1 (0.263; p < 0.05). Moreover, while we find that the coefficient on CIR_Si,∑3y is positive (0.352; p < 0.01), the coefficient on CIR_Ii,∑3y is negative (-0.256; p < 0.01), which indicates that prior incidents of misbehavior in the absence of high severity are associated with less severe future misconduct. However, the interaction is positively significant (0.114; p < 0.01), implying that the severity of future misbehavior increases in the frequency of prior misbehavior when the incidents are more severe. 32 This is in line with literature on ESG rating coverage, such as by Bloomberg, that finds an increase in CSR disclosures for more-covered firms (Bikmetova and Pirinsky 2022). 33 We use heteroscedasticity-robust standard errors (r) since the Pagan-Hall test is significant with p < 0.01 and rejects homoscedasticity. 34 This also holds for the additional IVs DJSIi,t-1 and CROSSi,t. 36 We cannot use the Anderson-Rubin or Cragg-Donald tests (which also serve to validate relevance) since we use heteroscedasticity-robust standard errors (e.g., Baum etal.2007; Cheng etal.2014). 35 Correlations of DJSIi,t-1 and CROSSi,t are slightly higher (between 0.16 and 0.18), but the C statistic test of exogeneity for DJSIi,t-1 and CROSSi,t finds insignificant results in our additional analyses (p > 0.1, untabulated), supporting their exogeneity (Bascle 2008). Prior studies have validated the exogeneity of firm age (e.g., Harjoto and Jo 2011; Christensen 2016); hence, the prerequisite to perform the C statistic test of at least one exogenous instrument (Wooldridge 2013) may be considered fulfilled. 1836 C.Reitmaier et al. Table 7 CSR reporting and future corporate misbehavior (2SLS model) Panel A: Voluntary issuance of a stand-alone GRI report and future misbehavior Futu reCIRi,t+1 = β0+ β1CSRR_Ii,t + β2UNGCi,t+1 + β3POi+ β4HOi+ β5GLOBALi,t+1 + β6PriorCIRi,t-3/∑3y + β7SIZEi,t+1+ β8ROEi,t + IND + YEAR + εi,t+1 (2) Column (A)Column (B)Column (C)Column (D)Column (E)Column (F)Column (G) Dep. var . 2nd stage CIR_It+1 CIR_It+1 CIR_Nt+1 CIR_Nt+1 CIR_St+1 CIR_St+1 CIR_St+1 Coeff. zCoeff.z Coeff. zCoeff.z Coeff. zCoeff.z Coeff. z CS RR_Ii,t (instrument.) 0.4485.84 *** 0.2773.93 *** 0.324 4.56***0.170 2.57***0.509 3.99 ***0.251 2.17 ** 0.263 2.29** UN GCi,t+1 -0.116 -4.67*** -0.078 -3.52*** -0.083 -3.58*** -0.053 -2.52**-0.145-3.54 ***-0.101-2.75 ***-0.102 -2.85*** PO i0.1042.98*** 0.0571.84* 0.074 2.35** 0.0291.010.142 2.60***0.065 1.31 0.071 1.46 HO i-0.081 -3.78*** -0.049 -2.53**-0.065 -3.34*** -0.028 -1.58-0.100-2.83 ***-0.024-0.74 -0.041 -1.33 GLOBAL i,t+1 0.0001.450.000 0.81 0.001 1.97** 0.0001.260.000 0.06 -0.000 -0.83-0.000 -0.97 CIR_ Ii,t-30.25714.71*** CIR_ Ii,∑3y0.17726.71 ***-0.256 -9.17*** CIR_ Ni,t-30.336 18.40*** CIR_ Ni,∑3y0.30326.84*** CIR_ Si,t-30.44727.73*** CIR_ Si,∑3y0.42539.84***0.352 18.47*** CIR_ Ii,∑3y*CIR_Si,∑3y0.11411.39*** SIZE i,t+1 0.0698.65*** 0.0476.62*** 0.075 9.79***0.053 7.79***0.150 11.27*** 0.1089.14*** 0.095 8.24*** RO Ei,t 0.0000.410.000 0.65 0.000 0.29 0.0000.640.000 3.06***0.000 3.78***0.000 3.62*** Constant -1.162 -5.77*** -0.699 -3.89*** -1.230 -6.50*** -0.790 -4.61*** -2.491 -7.61*** -1.797 -6.14*** -1.468 -5.11*** Industry fixed effect sIncluded IncludedIncludedIncludedIncludedIncludedIncluded Ye ar fixed effectsIncluded IncludedIncludedIncludedIncludedIncludedIncluded N4 ,499 4,4994,499 4,4994,499 4,49 94 ,499 F stati stic 39.69 *** 75.68 ***34.18 *** 51.34 *** 78.77 ***116.50 *** 131.04 *** Centered R² 0.0950.27540.22890.35830.35920.475 10 .4936 Highest VIF9.739.879.7810.01 9.61 9.64 9.88 Mean VI F2.342.362.352.382.352.3 62 .76 Durbin-W u-Hausman Chi2 (endog eneity) 36.5096 ***14.988 *** 20.144 *** 5.917 ** 18.417 *** 5.083 ** 6.066 ** Kleinbergen -Paap Wald rk F statistic (weak identification) 47.11746.28246.94045.70247.99047.62 64 6.345 Hans en J statistic (overidentification ) 10.213 ** 7.973 ** 11.742 *** 8.843 ** 16.769 *** 11.174 ** 12.973 *** 1837 Corporate responsibility andcorporate misbehavior: are… Table 7 (continued) 1st stage results with dependent variable: CSRR_I i,t Coeff. tCoeff.t Coeff. tCoeff.t Coeff. tCoeff.t Coeff. t AGEi,t 0.002 6.29***0.002 6.30 ***0.002 6.28 ***0.002 6.27***0.002 6.37 ***0.002 6.37 ***0.002 6.31*** FFLOATi,t 0.001 2.42 ** 0.0012.40**0.001 2.42** 0.001 2.36** 0.001 2.42 ** 0.0012.42**0.001 2.44** UNGCi,t+1 0.21112.58***0.211 12.59*** 0.21012.55***0.210 12.55*** 0.21112.62***0.211 12.60*** 0.21112.59*** POi-0.191 -6.37*** -0.190 -6.35*** -0.192 -6.40*** -0.192 -6.40*** -0.190 -6.35*** -0.189 -6.33*** -0.189 -6.32*** HOi0.054 2.51** 0.0542.52**0.054 2.50 ** 0.055 2.57***0.054 2.52** 0.0542.50**0.052 2.41** GLOBALi,t+1 0.001 4.59***0.001 4.58 ***0.001 4.51 ***0.001 4.45***0.001 4.60 ***0.001 4.62 ***0.001 4.55 *** CIR_Ii,t-30.028 1.92 * CIR_Ii,∑3y0.0101.59-0.012 0.84 CIR_Ni,t-30.0322.24** CIR_Ni,∑3y0.022 2.25** CIR_Si,t-30.013 1.71 * CIR_Si,∑3y0.0061.05-0.008-0.86 CIR_Ii,∑3y*CIR_Si,∑3y0.0101.89* SIZEi,t+1 0.018 2.66***0.018 2.58 ***0.017 2.49 ** 0.016 2.31 ** 0.017 2.43 ** 0.0182.52**0.017 2.36** ROEi,t -0.000 -1.22-0.000-1.22 -0.000 -1.21-0.000 -1.19-0.000 -1.22-0.000-1.23 -0.000 -1.25 Constant 0.700 3.82***0.699 3.80 ***0.726 3.95 ***0.734 3.96***0.716 3.87 ***0.696 3.75 ***0.727 3.91*** Industry fixed effectsIncluded IncludedIncludedIncludedIncludedIncludedIncluded Year fixed effectsIncluded IncludedIncludedIncludedIncludedIncludedIncluded Panel B: CSR disclosu re quality and future misbehavior Futur eCIRi,t+1 = β0+ β1CSRR_Qi,t + β2UNGCi,t+1+ β3POi+ β4HOi+ β5GLOBALi,t+1 + β6PriorCIRi,t-3/∑3y + β7SIZEi,t+1+ β8ROEi,t + IND + YEAR + εi,t+1 (2) Column (A)Column (B)Column (C)Column (D)Column (E)Column (F)Column (G) Dep. var . 2nd stage CIR_It+1 CIR_It+1 CIR_Nt+1 CIR_Nt+1 CIR_St+1 CIR_St+1 CIR_St+1 Coeff. zCoeff.z Coeff. zCoeff.z Coeff. zCoeff.z Coeff. z CS RR_Qi,t (instrument.) 0.0214.11*** 0.0153.37*** 0.0142.97*** 0.0081.88* 0.0313.72*** 0.0192.65*** 0.0212.97*** UN GCi,t+1 -0.045 -2.01**-0.039 -1.97**-0.036-1.81 *-0.033-1.85 *-0.076-2.03 ** -0.078 -2.33**-0.075-2.29 ** PO i0.0932.28**0.079 2.22** 0.0411.170.025 0.82 0.1301.98**0.077 1.31 0.0841.46 HO i-0.014 -0.51-0.012 -0.49-0.017-0.73 -0.010 -0.470.018 0.41 0.0340.910.025 0.66 GLOBAL i,t+1 -0.000 -0.22-0.000 -0.450.000 0.20 0.0000.09-0.001 -1.76* -0.001 -2.06**-0.001 -2.08** CIR_ Ii,t-30.25812.10*** CIR_ Ii,∑3y0.17621.19***-0.256 -7.20*** CIR_ Ni,t-30.321 15.45 *** CIR_ Ni,∑3y0.28821.73*** CIR_ Si,t-30.417 21.79*** CIR_ Si,∑3y0.40832.03***0.316 14.19*** CIR_Ii,∑3y * CIR_Si,∑3y 0.1219.98*** 1838 C.Reitmaier et al. Table 7 (continued) SIZE i,t+1 0.0241.320.010 0.61 0.0422.48**0.031 2.04** 0.0893.03*** 0.0622.42**0.041 1.65* ROEi,t -0.000 -1.93* -0.000 -0.84-0.000 -1.57-0.000 -0.320.000 1.41 0.000 3.07***0.000 3.06*** Constant -1.240 -5.17*** -0.790 -3.68*** -1.085 -5.01*** -0.690 -3.53*** -2.851 -7.08*** -1.960 -5.48*** -1.623 -4.59*** Industry fixed effectsIncludedIncludedIncludedIncludedIncludedIncludedIncluded Year fixed effectsIncludedIncludedIncludedIncludedIncludedIncludedIncluded N3,150 3,150 3,150 3,150 3,1503,150 3,150 F statistic 332.84*** 263.54 *** 51.17 *** 62.42*** 98.48*** 125.63 ***138.93*** Centered R² 0.1016 0.2678 0.2490 0.3673 0.3213 0.4559 0.4711 Highest VIF6.356.326.276.256.316.2 78 .13 Mean VIF2.452.472.452.472.462.4 83 .46 Durbin-Wu-Hausman Chi2 (endogeneity) 18.385 *** 10.649 *** 8.728 *** 3.073 *15.108*** 7.437*** 9.333*** Kleinbergen-Paap Wald rk F statistic (weak identification) 35.172 34.399 34.87533.72235.46834.58334.112 1 st stage results with dependent variable: CSRR_Qi,t Coeff. tCoeff.t Coeff. tCoeff.t Coeff. tCoeff.t Coeff. t AG Ei,t -0.014 -2.14**-0.014 -2.15**-0.014 -2.16**-0.015-2.17** -0.014 -2.05**-0.014-2.08** -0.014 -2.12** FFLOA Ti,t 0.0848.36*** 0.0848.26*** 0.0848.32*** 0.0838.17*** 0.0848.41*** 0.0848.30*** 0.0838.23*** UN GCi,t+1 2.659 6.61***2.653 6.60***2.637 6.54***2.620 6.49***2.678 6.67***2.653 6.61***2.652 6.60*** PO i-3.260 -4.10*** -3.242 -4.08*** -3.297 -4.14*** -3.277 -4.12*** -3.267 -4.11*** -3.281 -4.13*** -3.265 -4.10*** HO i-2.740 -4.58*** -2.720 -4.54*** -2.744 -4.59*** -2.695 -4.50*** -2.746 -4.58*** -2.708 -4.51*** -2.710 -4.51*** GLOBAL i,t+1 0.0547.81*** 0.0537.72*** 0.0537.71*** 0.0537.61*** 0.0547.85*** 0.0547.76*** 0.0537.71*** CIR_ Ii,t-30.848 2.26** CIR_ Ii,∑3y0.354 2.09** 0.055 0.14 CIR_ Ni,t-31.0022.94*** CIR_ Ni,∑3y0.681 2.77*** CIR_ Si,t-30.3681.93* CIR_ Si,∑3y0.3292.10**0.121 0.52 CIR_ Ii,∑3y*CIR_Si,∑3y0.0750.56 SIZE i,t+1 2.745 14.30*** 2.719 13.70***2.705 13.89***2.672 13.32***2.720 13.58*** 2.683 12.99***2.673 12.93*** RO Ei,t 0.0031.460.003 1.48 0.0031.490.003 1.51 0.0031.480.003 1.51 0.0031.50 Constant 14.0322.78*** 14.2442.80*** 14.985 2.93***15.1742.95*** 14.5162.81*** 14.9522.88*** 15.086 2.89*** Industry fixed effect sIncludedIncludedIncludedIncludedIncludedIncludedIncluded Year fixed effects IncludedIncludedIncludedIncludedIncludedIncludedIncluded 1839 Corporate responsibility andcorporate misbehavior: are… Presents the first and second stage coefficients and t-/z-statistics from a 2SLS regression of model (2) with industry and year fixed effects. Panel A uses CSRR_Ii,t as CSR reporting variable and hence refers to the association between the voluntary issuance of a stand-alone GRI report in t and future corporate misbehavior in t + 1. Columns (A) to (G) differ in the measure of prior and future misbehavior: Column (A) includes CIR_Ii,t-3 and CIR_Ii,t1. Column (B) includes CIR_Ii,∑3y and CIR_Ii,t+1. Column (C) includes CIR_Ni,t-3 and CIR_Ni,t+1. Column (D) includes CIR_Ni,∑3y and CIR_Ni,t+1. Column (E) includes CIR_Si,t-3 and CIR_Si,t+1. Column (F) includes CIR_Si,∑3y and CIR_Si,t+1. Column (G) includes CIR_Ii,∑3y, CIR_Si,∑3y, their interaction, and CIR_Si,t+1. Panel B uses CSRR_Qi,t as CSR reporting variable and hence refers to the association between CSR disclosure quality in t and future corporate misbehavior in t + 1. As in panel A, columns (A) to (G) differ in the measures of prior and future misbehavior. In both panels, values for CIR_Ni,t-3, CIR_Ni,∑3y, CIR_Ni,t+1, CIR_Si,t-3, CIR_Si,∑3y, and CIR_Si,t+1 are logtransformed as in all empirical tests. All variables are defined as in Appendix3 * , **, *** indicate that the estimated coefficients are statistically significant at the 10 percent, 5 percent, and 1 percent level, respectively, using a two-tailed test Table 7 (continued) 1840 C.Reitmaier et al. Panel B shows the results for CSRR_Qi,t. The endogeneity results are similar to CSRR_Ii,t. In columns (A) through (G), we find a significantly positive association of CSRR_Qi,t with all measures of FutureCIRi,t+1 (p < 0.01/0.1), consistent with H2 and legitimacy theory. Voluntary CSR reporting does not seem to foster behavioral change and break the direct relation between prior and future misbehavior. The interaction results in column (G) are largely equivalent to panel A. Overall, the results imply that more extensive voluntary CSR reporting is not a signal of credible CSR commitment. Instead, more extensive voluntary CSR reporting is associated with an increase in the occurrence, number, and severity of future misbehaviors; hence, it likely serves as an instrument of impression management and greenwashing. Table8 presents the results for the SR model, where we trim outliers in all continuous variables at the 2.5th and 97.5th percentiles (panel A with CSRR_Ii,t: N = 4,278, panel B with CSRR_Qi,t: N = 2,996) and de-mean V to ensure that its mean is zero, consistent with the underlying methodology (Dong and Lewbel 2015; Bontemps and Nauges 2016). The general White (1980) test negates heteroscedasticity. We present marginal effects since regression coefficients in binary models have less explanatory power. The results are inferentially equivalent to the 2SLS results. All measures of FutureCIRi,t+1 are positively associated with CSRR_Ii,t (p < 0.01/0.05/0.1) and CSRR_Qi,t (p < 0.01/0.05). All SR model specifications show adequate test statistics. 7 Additional analyses androbustness In additional and robustness analyses, we address concerns that factors such as a firm’s CSR reporting history, alternative IVs, variable measurement, model specification, or sample composition may affect our results. All tests strongly support our main inferences. Table9 summarizes the results for models (1) and (2) for our measures of prior and future misbehavior based on CIR_Ii,t. Results with our other measures of misbehavior (based on CIR_Ni,t, CIR_Si,t) are inferentially equivalent. 7.1 Reporting history First, we address the concern that the firm’s history of CSR reporting may affect our results. Following Granger (1969), we include an additional variable in model (1) that accounts for prior CSR reporting. We use either a dichotomous variable that indicates whether firm i has issued a GRI report in any year prior to t (PriorGRI_Ii,t) or a categorical variable that counts the number of prior GRI reports before t (PriorGRI_Ni,t). For both variants, we find a significantly positive association with CSR reporting in t (CSRR_Ii,t or CSRR_Qi,t) (p < 0.01, untabulated). As in our main analyses, we find a significantly positive association between each measure of prior misbehavior (CIR_Ii,t-3, CIR_Ii,∑3y, CIR_Ni,t-3, CIR_Ni,∑3y, CIR_Si,t-3, or CIR_Si,∑3y) and CSRR_Qi,t (p < 0.01/0.05). However, there is no significant relation with CSRR_Ii,t. Subsequent to misbehavior, firms with a (more experienced) history of CSR reporting improve the quality of their reports rather than increase their likelihood of issuing another CSR report, which is not surprising since they have already been issuing 1841 Corporate responsibility andcorporate misbehavior: are… Table 8 CSR reporting and future corporate misbehavior (SR model) Panel A: Voluntary issuance of a stand-alone GRI report and future misbehavior FutureC IRi,t+1 = β0+ β1CSRR_Ii,t + β2UNGCi,t+1 + β3POi+ β4HOi+ β5GLOBALi,t+1+ β6PriorCIRi,t-3/∑3y+ β7SIZEi,t+1 + β8ROEi,t+ β9EXCHANGEt + IND + YEAR + εi,t+1 (4) Column (A)Column (B)Column (C)Column (D)Column (E)Column (F) Dep. variable 2nd stageCIR_It+1 CIR_It+1 CIR_Nt+1 CIR_Nt+1 CIR_St+1CIR_St+1 Coeff. zMarg. Effect zCoeff.z Marg. Effect zCoeff.z Marg. Effect zCoeff.z Marg. Effect zCoeff.Z Marg. Effect zCoeff.z Marg. Effect z CSR R_Ii,t(instrument.)0.018 3.81***0.104 3.85***0.015 3.21***0.092 2.39**0.017 3.27***0.108 3.86***0.019 4.28***0.118 7.41***0.026 3.32*** 0.2043.99*** 0.0151.92* 0.1311.80* UNGC i,t+1-0.004 -2.44**-0.021-3.31***-0.003-2.34** -0.021 -1.61-0.004-2.52** -0.026 -1.80* -0.005 -3.53*** -0.030 -3.15*** -0.010 -4.13*** -0.079 -4.83***-0.008-3.42***-0.072-3.02*** PO i0.008 3.83***0.048 4.90***0.006 2.60***0.034 2.15**0.008 3.34***0.050 3.50***0.006 3.08***0.039 2.23** 0.0030.830.023 0.64 0.0010.420.013 0.36 HO i-0.006 -4.16*** -0.033 -2.63*** -0.006 -4.39*** -0.036 -3.33*** -0.007 -4.38*** -0.042 -4.59*** -0.005 -3.87*** -0.031 -4.29*** -0.006 -2.49**-0.045 -2.56**-0.004-1.85*-0.036-1.78* GLOBAL i,t+10.000 1.71*0.000 1.38 0.0001.140.000 0.66 0.0000.960.000 1.15 0.0000.440.000 0.41 0.0000.080.000 0.08 -0.000 -0.33-0.000-0.27 CIR_ Ii,t-30.005 4.59***0.026 2.35** CIR_ Ii,∑3y0.0049.41***0.024 4.30*** CIR_ Ni,t-30.0087.43***0.049 5.19*** CIR_ Ni,∑3y 0.0069.87*** 0.0376.81*** CIR_ Si,t-30.01113.51***0.089 6.40*** CIR_ Si,∑3y 0.01218.20***0.102 9.71*** SIZE i,t+10.002 3.52***0.010 1.95*0.001 2.51**0.007 1.57 0.0023.84***0.013 2.65*** 0.0013.14*** 0.0093.21*** 0.0556.73*** 0.0444.05*** 0.0055.77***0.041 5.91*** RO Ei,t-0.000 -0.42-0.000-0.11 0.0000.440.000 0.08 0.0000.440.000 0.30 0.0000.640.000 0.13 0.0000.560.000 0.31 0.0000.750.000 0.43 Cons tant -0.055 -4.46*** -0.310 -3.77*** -0.031 -2.61*** -0.191 -1.83* -0.060 -4.48*** -0.375 -3.94*** -0.049 -4.28*** -0.303 -3.17*** -0.102 -5.17*** -0.808-4.55***-0.083-4.30***-0.737-6.13*** Sp ecial regressorIncludedIncludedIncludedIncludedIncludedIncludedIncludedIncludedIncludedIncluded IncludedIncluded Indust ry fixed effects IncludedIncludedIncludedIncludedIncludedIncludedIncludedIncludedIncludedIncludedIncludedIncluded Ye ar fixed effects IncludedIncludedIncludedIncludedIncludedIncludedIncludedIncludedIncludedIncludedIncludedIncluded 872,4872,4872,4872,4872,4872,4872,4872,4872,4872,4872,4872,4 N Ch i21,075.14 ***1,253.10 ***1,026.23 ***1,168.12 ***900.43 ***1,119.98 *** 340.0440.0520.0030.0720.0720.0ESM R 43.662.698.505.551.669.5amgi S Wh ite test (homoskedasticity)241.70 *** 220.54 *** 246.79 *** 222.65 ***237.41 ***230.21 *** 5.25.25.25.25.25.2levelgnimmir T 010101010101selpmas partstoo B Sargan statistic (overid. restrictions test)-346.252-354.041-329.438-270.521-487.05 4474.606 Basm ann statistic ( overid. restrictions test )-4,293.60-4,294.78 -4,290.86-4,278.61 4,309.19 -4,308.18 1842 C.Reitmaier et al. Table 8 (continued) Panel B: CSR disclosure quality and future misbehavior FutureC IRi,t+1 = β0+ β1CSRR_Qi,t + β2UNGCi,t+1 + β3POi+ β4HOi+ β5GLOBALi,t+1 + β6PriorCIRi,t-3/∑3y+ β7SIZEi,t+1 + β8ROEi,t + β9EXCHANGEt + IND + YEAR + εi,t+1 (4) Column (A)Column (B)Column (C)Column (D)Column (E)Column (F) Dep. variable 2nd stageCIR_It+1 CIR_It+1 CIR_Nt+1 CIR_Nt+1 CIR_St+1CIR_St+1 Coeff. zMarg. Effect zCoeff.z Marg. Effect zCoeff.z Marg. Effect zCoeff.z Marg. Effect zCoeff.Z Marg. Effect zCoeff.z Marg. Effect z CS RR_Qi,t(instrument.)0.001 6.73***0.005 9.79***0.001 6.09***0.005 8.37***0.002 6.26***0.005 5.30***0.002 6.31***0.005 4.26*** 0.0024.97*** 0.0135.67*** 0.0014.00***0.012 2.03** UNGC i,t+1-0.002 -2.01**-0.008-1.21 -0.003 -2.16**-0.010-2.47** -0.004 -2.87*** -0.012 -2.21**-0.005-3.76***-0.017-1.78*-0.007-3.83***-0.057-3.96*** -0.006 -3.31*** -0.056 -4.16*** PO i0.001 0.42 0.0030.270.004 4.52 0.0131.14-0.002-0.93 -0.007 -0.67-0.002-0.77 -0.006 -0.98-0.004-1.24 -0.033 -3.33*** -0.003 -1.13-0.034-0.70 HO i-0.001 -0.84-0.004-0.29 -0.002 -1.39-0.008-0.93 -0.000 -0.13-0.001-0.08 -0.001 -0.62-0.004-0.51 0.0010.370.007 0.26 0.0020.940.020 0.41 GLOBAL i,t+1-0.000 -4.31***-0.000-3.61***-0.000-3.73***-0.000-3.01***-0.000-4.62***-0.000-8.79*** -0.000 -5.19*** -0.000 -5.35*** 0.000-3.06***-0.001 -3.48*** 0.000-2.85***-0.001-1.99** CIR_ Ii,t-30.004 3.87***0.013 1.07 CIR_ Ii,∑3y0.0036.54*** 0.0122.75*** CIR_ Ni,t-30.0064.71***0.017 2.67*** CIR_ Ni,∑3y 0.0067.37*** 0.0203.66*** CIR_ Si,t-30.0089.20*** 0.0643.70*** CIR_ Si,∑3y 0.00812.98***0.081 2.44** SIZE i,t+1-0.002 -1.99**-0.005-0.67 -0.002 -2.35**-0.007-2.14** -0.002 -1.68* -0.005 -1.42-0.002-2.32** -0.007 -1.210.001 0.38 0.0040.330.000 0.32 0.004 0.11 RO Ei,t0.000 0.57 -0.000 -0.11-0.000-0.41 -0.000 -0.16-0.000-0.24 -0.000 -0.07-0.000-0.16 -0.000 -0.05-0.000-0.27 -0.000 -0.09-0.000-0.15 -0.000 -0.02 Cons tant -0.074 -3.18***-0.141-0.92 -0.044 -3.26***-0.169-1.84*-0.047-2.89***-0.137-1.48 -0.037 -2.31**-0.118-1.48 -0.077 -3.61*** -0.600 -3.30***-0.069-3.64***-0.683-1.59 Sp ecial regressorIncluded Included IncludedIncludedIncludedIncludedIncludedIncludedIncludedIncludedIncludedIncluded Indust ry fixed effectsIncludedIncludedIncludedIncludedIncludedIncludedIncludedIncludedIncludedIncludedIncludedIncluded Ye ar fixed effectsIncludedIncludedIncludedIncludedIncludedIncludedIncludedIncludedIncludedIncludedIncludedIncluded 699,2699,2699,2699,2699,2699,2699,2699,2699,2699,2699,2699,2 N Ch i2622.97 *** 715.05 *** 564.50 ***638.46 ***539.93 ***674.04 *** 430.0930.030.030.0520.0620.0ESM R 40.68.516.647.662.681.6amgi S Wh ite test (homoskedasticity)203.61 *** 210.28 *** 198.64 *** 211.76 *** 206.27 *** 207.70 *** 5.25.25.25.25.25.2levelgnimmir T 010101010101selpmaspartstoo B Sargan statis tic (overid. restrictions test)-231.035-202.186-173.809-153.478-375.77 4344.216 Basm ann statistic ( overid. restrictions test )-2,993.20 -2,984.61-2,980.32-2,973.48 -3,008.4 13,006.17 1843 Corporate responsibility andcorporate misbehavior: are… Presents the coefficients and marginal effects with z-statistics from a SR regression of model (4) with industry and year fixed effects. Panel A uses CSRR_Ii,t as CSR reporting variable and hence refers to the association between the voluntary issuance of a stand-alone GRI report in t and future corporate misbehavior in t + 1. Columns (A) to (F) differ in the measures of prior and future misbehavior: Column (A) includes CIR_Ii,t-3 and CIR_Ii,t+1. Column (B) includes CIR_Ii,∑3y and CIR_Ii,t+1. Column (C) includes CIR_Ni,t-3 and CIR_Ni,t+1. Column (D) includes CIR_Ni,∑3y and CIR_Ni,t+1. Column (E) includes CIR_Si,t-3 and CIR_Si,t+1. Column (F) includes CIR_Si,∑3y and CIR_Si,t+1. Panel B uses CSRR_Qi,t as CSR reporting variable and hence refers to the association between CSR disclosure quality in t and future corporate misbehavior in t + 1. As in panel (A), columns (A) to (F) differ in the measures of prior and future misbehavior. In both panels, values for CIR_Ni,t-3, CIR_Ni,∑3y, CIR_Ni,t+1, CIR_Si,t-3, CIR_Si,∑3y, and CIR_Si,t+1 are logtransformed as in all empirical tests. All continuous variables are trimmed at the 2.5th and 97.5th percentile, consistent with the underlying methodology (Dong and Lewbel 2015; Bontemps and Nauges 2016). Standard errors (untabulated) of the marginal effects at the mean are bootstrapped. All variables are defined as in Appendix3 * , **, *** indicate that the estimated coefficients are statistically significant at the 10 percent, 5 percent, and 1 percent level, respectively, using a two-tailed test Table 8 (continued) 1850 C.Reitmaier et al. higher public scrutiny on CSR and related (mis-)reporting due to the stronger regulations in Europe. Most specifications of model (2) confirm a significantly positive association of CSR reporting with future misbehavior (p < 0.01/0.05/0.1).44 7.6 Alternative sample composition: extended timeframe Finally, we analyze a broader sample of Fortune Global 500 firms until 2018.45 We compare our main sample period (2003–2013) to the periods 2014–2016 (after the issuance of the EU directive and the Integrated Reporting Framework) and 2017–2018 (after the effective date of the EU directive). For 2014–2016, we find a significantly negative association between prior misbehavior and CSRR_Ii,t in model (1) (p < 0.01/0.05/0.1). For 2017–2018, the association is insignificant for CIR_It-3, CIR_I∑3y, and CIR_N∑3y and significantly negative for CIR_Nt-3, CIR_St-3, and CIR_S∑3y (p < 0.05/0.1, untabulated). In model (2), the association of CSRR_Ii,t with future misbehavior is insignificant for CIR_Ii,t-3, CIR_Ii,∑3y, and CIR_Ni,t-3 and negatively significant for CIR_Ni,∑3y, CIR_Si,t-3, and CIR_Si,∑3y (p < 0.05/0.1, untabulated) for 2014–2016 and insignificant for 2017–2018. However, this extended time period is characterized by a transition from voluntary to mandatory CSR reporting. We observe in the data collection process that many firms have replaced GRI-based stand-alone CSR reports with integrated reports or mandatory reports which are not captured by our measure for CSRR_Ii,t. The CSRR_Ii,t measure is therefore not suitable for this time period, and the results cannot be interpreted as evidence of a decrease in misbehavior. For CSRR_Qi,t, we find no significant associations with prior or future misbehavior in either time period. Again, the interpretation of these results requires caution, as many of our sample firms are operating in jurisdictions that are affected by the European and other initiatives to mandate the disclosure of CSR-related information. The firms’ reporting and hence our measure for CSRR_Qi,t will therefore be affected by these regulations. Overall, our measures of CSRRi,t do not well reflect voluntary CSR reporting during the later time periods, which is why we excluded these periods from our main analyses. More future research is needed to analyze the effects of mandatory CSR reporting on misbehavior in a clean setting. 8 Quasi‑natural experiment To further validate our results, we use a quasi-natural experiment and analyze an exogenous shock to textile-related industries. In 2013, the Rana Plaza disaster in Bangladesh led many firms to sign an accord for better working conditions. We analyze the voluntary signature as an additional form of proclaiming CSR commitment, to shed more light on the relation between public CSR commitment and actual 45 After excluding firms not covered by the databases used, the sample includes 632 firms. 44 The association is insignificant when using the combinations CSRR_Ii,t and CIR_Ii,∑3y or CSRR_Qi,t and CIR_Si,∑3y. 1851 Corporate responsibility andcorporate misbehavior: are… behavior.46 The general expectation is that catastrophes like the Rana Plaza disaster lead to better working conditions in the related industries, since public scrutiny following the events is high and firms have incentives to fix their shortcomings. We compare the firms affected by the disaster to a control group of unaffected firms from other industries in a difference-in-differences design. In addition, we divide the group of affected firms into signatories and non-signatories of the accord. Signaling theory implies a particular reduction in misbehavior for signatories if the signature constitutes a credible signal of CSR commitment. However, the results of our analyses above imply that many firms proclaim CSR commitment for legitimization rather than signaling. Legitimacy theory implies that prior misbehavior increases the likelihood of firms signing the accord and that signatories do not reduce, and may even increase, their misbehavior afterwards. We thus expect that signatories behave less responsibly than non-signatories pre and post accord.47 We analyze 528 firms in five ICB subsectors: Clothing & Accessories (ICB code 3763), Footwear (3765), Food Retailers & Wholesalers (5337), Apparel Retailers (5371), and Broadline Retailers (5373). The 176 firms that signed the accord in 2013 (Bangladeshaccord 2013) are concentrated in these five ICB subsectors, except for three firms that we exclude since their main activities are in inadequately represented subsectors (Food Products (3577), Furnishing (3726), and Toys (3747)). We exclude 95 firms that signed the accord after 2013 or signed the follow-up accord that became effective in 2018. We further exclude 139 signatories and 171 non-signatories that are not covered by the databases used (Datastream, Eikon, and CCRD). The final sample includes 34 signatories and 89 non-signatories. As in Paik etal. (2017), the size of these two subsamples is relatively small but sufficient (Barnes 2017). We analyze misbehavior from 2000 to 2018. As a control sample, we analyze the Fortune Global 500 firms described in Sect. 7.6 (632 firms). Table10 shows the results. Panel A provides the means and standard deviations of CIR_Ii,t for the period pre (2000–2012) and post the disaster (2014–2018) for the three samples. Linear regressions test for differences (panel B). Figure1 illustrates the results. 46 In April 2013, the Rana Plaza building in Dhaka collapsed, killing more than 1,100 people and badly injuring thousands more (Bangladeshaccord 2019). This was an unexpected shock to the market and affected textile-related firms all over the world that commonly outsource their manufacturing in Bangladesh to save costs (Quinlan and Sheldon 2011; Barnes 2017). In response to an increase in external pressure after the disaster, outsourcing firms, trade unions, and witnesses initiated a safety agreement (the Accord on Fire and Building Safety in Bangladesh) effective in May 2013 – only one month after the disaster (Bangladeshaccord 2019). The aim was to gather money and resources to improve safety and working conditions and foster CSR over economic targets (Paik, Lee, and Krumwiede 2017). 47 Another, more recent incident occurred in Dhaka in 2017. An old heating boiler exploded in a factory despite inspections in accordance with the accord (Bangladeshaccord 2017). The factory was a supplier for several firms that had signed the accord in 2013. Akbar and Deegan (2021) highlight that a main reason for increased disclosures on workplace safety after the Rana Plaza disaster is to respond to outside pressure. Sinkovics etal. (2016) further find that outside pressure after the disaster has led firms to prioritize measurable standards implementation above the needs of their employees and society as a whole. In addition, firms seek greater efficiency in their processes to cope with the increasing cost of compliance, and neglect harmful side effects that destroy social value. Overall, this indicates that improvements after the disaster do not necessarily address all dimensions of CSR. 1852 C.Reitmaier et al. Table 10 Quasi-natural experiment: Bangladesh Accord Panel A: Means (standard deviation) of CIR_Ii,t by firm groups and periods FIRMS TOTAL F500 Signatories Non-signatories POST Pre shockevent 0.2807 0.3710 0.1780 0.2726 (0.4494) (0.4836) (0.3827) (0.4453) N = 8,216 N = 442 N = 1,157 N = 9,815 Post shock event 0.2066 0.3882 0.2449 0.2193 (0.4050) (0.4888) (0.4305) (0.4139) N = 3,160 N = 170 N = 445 N = 3,775 TOTAL 0.2601 0.3758 0.1966 0.2578 (0.4387) (0.4847) (0.3976) (0.4375) N = 11,376 N = 612 N = 1,602 N = 13,590 1853 Corporate responsibility andcorporate misbehavior: are… Table 10 (continued) Panel B: Linear regression results for the dependent variable CIR_Ii,t Without determinants and control variables (N = 13,590) With determinants and control variables (N = 8,966) Variable Coefficient t Coefficient t POST -0.0740 -8.13*** -0.0633 -6.37*** FIRMS  Signatories 0.0904 4.25*** 0.1015 3.26***  Non-signatories -0.1026 -7.51*** 0.0222 0.97 POST * FIRMS  F500 Signatories 0.0912 2.26** 0.0874 1.87*  F500 Non-signatories 0.1409 5.44*** 0.0358 1.37 UNGCi,t 0.0027 0.27 POi0.0495 2.93*** HOi-0.0620 -4.67*** GLOBALi,t 0.0006 4.07*** CIR_Ii,t-1 0.3847 39.57*** SIZEi,t 0.0593 16.92*** ROEi,t -0.0114 -1.79* Constant 0.2807 58.47*** -1.1928 -10.98*** Industry fixed effects Included Presents the results of a quasi-natural experiment surrounding the Rana Plaza disaster. Panel A shows means and standard deviations of CIR_Ii,t as well as the number of estimations for each firm group and time period. Panel B shows the results of linear regressions excluding and including the determinants and control variables of model (2). All variables are defined as in Appendix3 * , **, *** indicate that the estimated coefficients are statistically significant at the 10 percent, 5 percent, and 1 percent levels, respectively, using a two-tailed test 1854 C.Reitmaier et al. Overall, CIR_Ii,t has significantly declined for the average firm in the total sample from pre to post 2013. The coefficient of -0.0740 (p < 0.01) of POST indicates that the average unaffected Fortune Global 500 firm has reduced misbehavior by 0.0740 from pre to post 2013 (0.2807 vs. 0.2066). The mean of CIR_Ii,t is higher for signatories than for unaffected firms by 0.0904 pre 2013 (p < 0.01; 0.3710 vs. 0.2807) and by 0.1816 post 2013 (p < 0.01; 0.3882 vs. 0.2066). This difference has thus increased by (0.1816–0.0904 =) 0.0912 from pre to post 2013 (p < 0.05). CIR_Ii,t of non-signatories is lower than that of unaffected firms by 0.1026 pre 2013 (p < 0.01; 0.1780 vs. 0.2807). Post 2013, it is higher by 0.0383 (p < 0.1; 0.2449 vs. 0.2066). This difference has thus increased by (0.0383–0.1026 =) 0.1409 (p < 0.01). The main inferences hold after including the determinants and control variables of model (2) and using industry fixed effects.48 Overall, the results indicate that misbehavior is more likely for signatories compared to both non-signatories and unaffected firms pre and post 2013 and that signatories do not reduce their misbehavior after signing the accord. Firms seem to sign the accord to greenwash misbehavior rather than to signal true CSR commitment. Fig. 1 Mean corporate misbehavior pre and post 2013 by firm groups. Displays the mean corporate misbehavior (CIR_Ii,t) pre and post 2013 by firm groups (created in STATA). On average, corporate misbehavior is significantly lower for the overall sample of firms and for the control group of unaffected Fortune Global 500 firms in the period after the incident compared to the period before the incident. While for both non-signatories and signatories corporate misbehavior is significantly higher after the incident, the average corporate misbehavior remains significantly higher for signatories compared to non-signatories as well as to the unaffected Fortune Global 500 firms 48 The difference in the mean of CIR_Ii,t pre 2013 and the change of CIR_Ii,t from pre to post 2013 for non-signatories compared to unaffected firms is no longer significant. All other differences, particularly those of signatories, remain significant. POi, HOi, GLOBALi,t, CIR_Ii,t-1, and SIZEi,t are significant at p < 0.01; ROEi,t is significant at p < 0.1. UNGCi,t is not significant. We measure prior misbehavior (CIR_ Ii,t-1) in the year before the dependent variable CIR_Ii,t in order to control for the direct relation between past and future misbehavior. We measure all other variables at the same time as CIR_Ii,t, i.e., in t (comparable to model (2) where FutureCIRi,t+1 and further variables are measured at the same time, t + 1). We do not use year fixed effects since POST captures time differences. 1855 Corporate responsibility andcorporate misbehavior: are… 9 Conclusion We analyze corporate misbehavior pre and post voluntary CSR reporting to address the reciprocal relation of proclaimed and implemented CSR commitment over time. Prior literature on the relation between CSR reporting and CSR performance often aggregates responsible and irresponsible behaviors. To avoid measurement problems arising from this aggregation, we measure irresponsible behavior only, but across all CSR dimensions and preceding as well as subsequent to CSR reporting. We find significantly positive associations between CSR reporting (the voluntary issuance of a GRI report and CSR disclosure quality) and both prior and future misbehavior (occurrence, number, and severity). This implies that prior misbehavior increases the efforts expended for CSR reporting. However, our results imply that CSR reporting, even of better quality, is not an indication of improved operations but is related to increased future misbehavior, both in terms of number and severity. We substantiate our findings using the Rana Plaza disaster in Bangladesh in 2013 as a quasi-natural experiment. We analyze the voluntary decision to sign an accord for better working conditions as an additional form of proclaiming CSR commitment. We find that misbehavior is more likely for signatories before and after the signature, compared to non-signatories and firms from industries not affected by the exogenous shock. Signatories show no sign of improvement. Consistent with legitimacy theory, our results imply that voluntarily proclaiming CSR commitment is not a credible signal of the latter but serves impression management and greenwashing purposes. An inevitable limitation of our study is that we cannot entirely rule out endogeneity issues. Still, multiple approaches that address these issues support our inferences. Also, due to a lack of data, we had to exclude some firms from countries with apparent human rights violations like China. Our analyses focus on the period until 2013 when CSR reporting was generally voluntary and before some jurisdictions had begun to mandate it. Whereas we find an increase in misbehavior subsequent to voluntary CSR reporting during our main sample period, we find no such evidence for the extended time period of 2014–2018. However, this later time period is characterized by a transition to mandatory reporting during which our measures do not well reflect voluntary CSR reporting, which is why we excluded it from our main analyses. Whereas some jurisdictions, like Europe, have decided to mandate the disclosure of CSR information, others, like the US, remain undecided about whether and how to mandate CSR reporting. The results from our main sample period are unaffected by CSR reporting mandates and provide valuable insights for debates over what voluntary CSR reporting can and cannot achieve. Nevertheless, future research is needed to investigate whether mandatory CSR reporting is more effective in altering firm behavior toward more responsibility than voluntary CSR reporting. 1856 C.Reitmaier et al. Table 11 UNGC principles and corresponding CCRD (sub-)categories UNGC categories CCRD categories CCRD subcategories Examples of misbehavior per CCRD subcategory Human rights Principle 1: Businesses should support and respect the protection of internationally proclaimed human rights; and Principle 2: make sure that they are not complicit in human rights abuses People Human rights Racial discrimination, gender discrimination, dispossession, human rights violations, sexual harassment, intimidation, violence Labor Principle 3: Businesses should uphold the freedom of association and the effective recognition of the right to collective bargaining; Principle 4: the elimination of all forms of forced and compulsory labour; Principle 5: the effective abolition of child labour; and Principle 6: the elimination of discrimination in respect of employment and occupation Workers’ rights Unnecessary worker injuries, poverty wages, excessive overtime, forced labour, poor working conditions Supply chain management Poor supply chain policy, unsustainable sourcing, poor conditions in supplier companies Environment Principle 7:Businesses should support a precautionary approach to environmental challenges; Principle 8: undertake initiatives to promote greater environmental responsibility; and Principle 9: encourage the development and diffusion of environmentally friendly technologies Environment Climate change Destruction of rainforest/deforestation, release of greenhouse gas emissions Pollution & toxics Air pollution, water pollution, oil spills, widespread use of pesticides Habitats & resources Habitat destruction, threats to endangered species (biodiversity), illegal logging, displacement of local communities, water depletion Anti-corruption Principle 10:Businesses should work against corruption in all its forms, including extortion and bribery Politics Political activities Political donations (e.g., against environmental laws) Anti-social finance Bribery, tax avoidance, price-fixing Appendix1 1857 Corporate responsibility andcorporate misbehavior: are… Table 12 Misbehavior severity score Panel A: Scoring scheme Scoring Effects Range & Magnitude Criteria Description 1 point limited effects Number of categories Only one category affected Impact Little: only one person//living being affected Health severity Rarely a direct threat to life (no significant magnitude and consequences) Social severity Typically results in little inconveniences Environmental severity Rarely a direct threat to the environment 2 points minor effects Number of categories One or two categories affected Impact Moderate: impact on a small group (≤ 500) of people/living beings Health severity Temporary discomfort, rarely a direct threat to life Social severity Typically results in an inconvenience to daily life Environmental severity Rarely a direct threat to the environment 3 points moderate effects Number of categories More than one category affected Impact Great: large group of people/living beings is affected Health severity Temporary injuries, often threatening to life, some damage unavoidable Social severity Typically results in disruptions to daily life Environmental severity Often threatening to the environment, some damage unavoidable 4 points major effects Number of categories More than two categories affected Impact Great: large group of people/living beings is affected Health severity Permanent injuries, extensive damage to communities likely Social severity Results in major disruptions to daily life Environmental severity Extensive damage to the environment likely Appendix2 1858 C.Reitmaier et al. Table 12 (continued) 5 points severe effects Number of categories Several categories affected Impact Great: large group of people/living beings is affected Health severity Life saving actions needed, death occurs, extensive and widespread severe damage to communities Social severity Results in extreme disruptions to daily life Environmental severity Extensive and widespread severe damage to the environment Panel B: Examples Category Score Examples E 1 point Poor toxic chemicals policy 2 points Small gas leak; unauthorised release of dust; breach of license for using more water than permitted 3 points Illegal discharge of waste water, water contamination, river pollution; emitting illegal levels of air pollution 4 points Destruction of rainforest, clear-cut land 5 points Large-scale oil spill, extreme pollution S 1 point Unfair dismissal of a worker 2 points Different types of discrimination (gender, race, disability etc.) 3 points Wages below the subsistence level; excessive overtime 4 points ‘Slave-like conditions’; increase in landlessness 5 points Factory collapse killing workers G1 point Excessive directors’ pay 2 points Bribery scheme; corruption 3 points Tax evasion; money laundering schemes 4 points Profound infiltration of a national government 5 points (No cases identified) 1859 Corporate responsibility andcorporate misbehavior: are… Table 13 List of variables Variable Measurement Source CSR reporting CSRRi,t A placeholder for one of the three CSR reporting variables (below) in year t CSRR_Ii,t Dichotomous variable: 1 indicates the voluntary issuance of a stand-alone GRI report by firm i in year t; 0 indicates no stand-alone GRI report in year t GRI database, CorporateRegister, firm websites, internet CSRR_I2i,t Dichotomous variable: 1 indicates the voluntary issuance of any kind of stand-alone CSR report by firm i in year t; 0 indicates no stand-alone CSR report in year t CorporateRegister, firm websites, internet CSRR_Qi,t CSR disclosure quality measured by the Bloomberg’s ESG disclosure score for firm i in year t (in percent) Bloomberg CSR reporting history PriorGRI_Ii,t Dichotomous variable: 1 indicates the voluntary issuance of at least one stand-alone GRI report by firm i before t; 0 indicates no stand-alone GRI report before t GRI database, CorporateRegister, firm websites, internet PriorGRI_Ni,t Categorical variable: counts the number of GRI reports issued by firm i before t, i.e., it measures in how many years firm i had issued a GRI report until t-1 GRI database, CorporateRegister, firm websites, internet PriorCSRR_Qi,t CSR disclosure quality measured by the Bloomberg’s ESG disclosure score for firm i in the last available year prior to t-3, i.e. in t-4 or, if CSRR_Qi,t is not available in t-4, then in t-5 or, if CSRR_Qi,t is also not available in t-5, then in t-6, etc. (in percent) Bloomberg CSRR_Ii,t-3 Dichotomous variable: 1 indicates the voluntary issuance of a stand-alone GRI report in year t-3; 0 indicates no stand-alone GRI report in year t-3 GRI database, CorporateRegister, firm websites, internet Corporate misbehavior CIRi,t A placeholder for one of the four variables of corporate irresponsibility (below) in year t Appendix3 1866 C.Reitmaier et al. Christensen, H. B., L. Hail, and C. Leuz. 2021. Mandatory CSR and sustainability reporting: Economic analysis and literature review. Review of Accounting Studies 26(3): 1176–1248. Christodoulou, D., and S. McLeay. 2014. The double entry constraint, structural modeling and economic estimation. Contemporary Accounting Research 31(2): 609–628. Clarkson, P. M., Y. Li, G. D. Richardson, and F. P. Vasvari. 2008. Revisiting the relation between environmental performance and environmental disclosure: An empirical analysis. Accounting, Organizations and Society 33(4–5): 303–327. Cormier, D., I. M. Gordon, and M. Magnan. 2004. Corporate environmental disclosure: Contrasting management’s perceptions with reality. Journal of Business Ethics 49(2): 143–165. Cormier, D., M. Magnan, and B. VanVelthoven. 2005. Environmental disclosure quality in large German companies: Economic incentives, public pressures or institutional conditions? European Accounting Review 14(1): 3–39. Corporate Critic Research Database (CCRD). 2020b. Publications Referenced by Ethical Consumer. Retrieved March 4, 2020, from http:// www. corpo ratec ritic. org/ info/ rr/ publi catio ns. aspx. Corporate Critic Research Database (CCRD). 2020a. Research & Ratings. How is Corporate Critic compiled? Retrieved March 4, 2020, from http:// www. corpo ratec ritic. org/ info/ rr/ compi led. aspx Corporate Critic Research Database (CCRD). 2020c. Gulf of Mexico oil spill, habitats, pollution and GHG impacts. Retrieved January 5, 2020, from http:// www. corpo ratec ritic. org/ abstr acts. aspx?% 20IDs= 545454. Corporate Critic Research Database (CCRD). 2021a. Ethical Consumer: About Ethical Consumer. Retrieved September 27, 2021, from https:// www. ethic alcon sumer. org/ aboutus. Corporate Critic Research Database (CCRD). 2021b. Ethical Consumer Research & Consultancy: Corporate Research Database. Retrieved September 27, 2021, from https:// resea rch. ethic alcon sumer. org/ corpo rateresea rchdatab ase. Corporate Critic Research Database (CCRD). 2021c. Ethical Consumer Research & Consultancy: Our Ethical Ratings System. Retrieved September 27, 2021, from https:// resea rch. ethic alcon sumer. org/ corpo rateresea rchdatab ase/ ourethic alratin gssystem. Corporate Critic Research Database (CCRD). 2021d. Ethical Consumer Research & Consultancy: Partnerships. Retrieved September 27, 2021, from https:// resea rch. ethic alcon sumer. org/ resea rchconsu ltancy/ partn ershi ps. Corporate Critic Research Database (CCRD). 2021e. Ethical Consumer Research & Consultancy: Global Directory of Ethical Consumption Organisations. Retrieved September 27, 2021, from https:// resea rch. ethic alcon sumer. org/ resea rchhub/ globaldirec toryethic alconsu mptionorgan isati ons. Corporate Critic Research Database (CCRD). 2021f. Ethical Consumer: Our Ethical Ratings. Retrieved September 27, 2021, from https:// www. ethic alcon sumer. org/ aboutus/ ourethic alratin gs. Datta, S., M. Iskandar-Datta, and A. Patel. 1999. Bank monitoring and the pricing of corporate public debt. Journal of Financial Economics 51(3): 435–449. Davis, J. H., F. D. Schoorman, and L. Donaldson. 1997. Toward a stewardship theory of management. The Academy of Management Review 22(1): 20–47. Dechow, P. M. 2023. Understanding the sustainability reporting landscape and research opportunities in accounting. The Accounting Review 98(5): 481–493. DeTienne, K. B., and L. W. Lewis. 2005. The pragmatic and ethical barriers to corporate social responsibility disclosure. The Nike case. Journal of Business Ethics 60(4): 359–376. DeVilliers, C. J., and D. Alexander. 2014. The institutionalisation of corporate social responsibility reporting. The British Accounting Review 46(2): 198–212. Dhaliwal, D. S., O. Z. Li, A. Tsang, and Y. G. Yang. 2011. Voluntary nonfinancial disclosure and the cost of equity capital: The initiation of corporate social responsibility reporting. The Accounting Review 86(1): 59–100. Dhaliwal, D. S., S. Radhakrishnan, A. Tsang, and Y. G. Yang. 2012. Nonfinancial disclosure and analyst forecast accuracy: International evidence on corporate social responsibility disclosure. The Accounting Review 87(3): 723–759. Donaldson, L., and J. H. Davis. 1991. Stewardship theory or agency theory: CEO governance and shareholder returns. Australian Journal of Management 16(1): 49–64. Donaldson, T., and T. W. Dunfee. 1994. Towards a unified conception of business ethics: Integrative social contracts theory. The Academy of Management Review 19(2): 252–284. Donaldson, T., and T. W. Dunfee. 1999. Ties that bind: A social contracts approach to business ethics. Harvard Business School Press. 1867 Corporate responsibility andcorporate misbehavior: are… Dong, Y., and A. Lewbel. 2015. A simple estimator for binary choice models with endogenous regressors. Econometric Reviews 34(1–2): 82–105. Dowling, J., and J. Pfeffer. 1975. Organizational legitimacy: Social values and organizational behavior. The Pacific Sociological Review 18(1): 122–136. Downar, B., J. Ernstberger, S. Reichelstein, S. Schwenen, and A. Zaklan. 2021. The Impact of carbon disclosure mandates on emissions and financial operating performance. Review of Accounting Studies 26(3): 1137–1175. Dube, S., and C. Zhu. 2021. The disciplinary effect of social media: Evidence from firms’ responses to Glassdoor reviews. Journal of Accounting Research 59(5): 1783–1825. Einwiller, S. A., C. E. Carroll, and K. Korn. 2010. Under What Conditions Do the News Media Influence Corporate Reputation? The Roles of Media Dependency and Need for Orientation. Corporate Reputation Review 12(4): 299–315. European Commission (EC). 2021. Corporate sustainability reporting. Retrieved May 31, 2021, from https:// ec. europa. eu/ info/ busin essecono myeuro/ compa nyrepor tingandaudit ing/ compa nyrepor ting/ nonfinan cialrepor ting_ en. European Union (EU). 2019. Regulation (EU) 2019/2088 of the European Parliament and of the Council of 27 November 2019 on sustainability‐related disclosures in the financial services sector. Retrieved January 14, 2022, from https:// eurlex. europa. eu/ eli/ reg/ 2019/ 2088/ oj? msclk id= 55f05 e62b1 8111e c851a 1b441 213b6 c1& locale= en. Fama, E. F., and M. C. Jensen. 1983. Separation of ownership and control. The Journal of Law & Economics 26(2): 301–325. Farber, D. B. 2005. Restoring trust after fraud: Does corporate governance matter? The Accounting Review 80(2): 539–561. Fiechter, P., J. -M. Hitz, and N. Lehmann. 2022. Real effects of a widespread CSR reporting mandate: Evidence from the European Union’s CSR Directive. Journal of Accounting Research 60(4): 1499–1549. Flammer, C. 2013. Corporate social responsibility and shareholder reaction: The environmental awareness of investors. Academy of Management Journal 56(3): 758–781. Freundlieb, M., and F. Teuteberg. 2013. Corporate social responsibility reporting – a transnational analysis of online corporate social responsibility reports by market-listed companies: Contents and their evolution. International Journal of Innovation and Sustainable Development 7(1): 1–26. Friedman, M. 1970. The social responsibility of business is to increase its profits. The New York Times Magazine, 119(41): 122–126. Frynas, J. G., and C. Yamahaki. 2016. Corporate social responsibility: Review and roadmap of theoretical perspectives. Business Ethics: A European Review 25(3): 258–285. Gamerschlag, R., K. Moeller, and F. Verbeeten. 2011. Determinants of voluntary CSR disclosure: Empirical evidence from Germany. Review of Managerial Science 5(2–3): 233–262. Global Reporting Initiative (GRI). 2011. A new phase: The growth of sustainability reporting. GRI’s Year in Review 2010/11. Retrieved March 4, 2019, from https:// www. globa lrepo rting. org/ resou rceli brary/ GRIYearInReview20102011. pdf. Granger, C.W.J. 1969. Investigating causal relations by econometric models and cross-spectral methods. Econometrica 37 (3): 424–438. Grewal, J., G. D. Richardson, and J. Wang. 2023. Effects of Mandatory Carbon Reporting on Unrepresentative Environmental Disclosures. Retrieved August 16, 2023, from https:// ssrn. com/ abstr act= 41661 84. Hackston, D., and M. J. Milne. 1996. Some determinants of social and environmental disclosures in New Zealand companies. Accounting, Auditing & Accountability Journal 9(1): 77–108. Haddock, J. 2005. Consumer influence on internet-based corporate communication of environmental activities: The UK food sector. British Food Journal 107(10): 792–805. Hahn, R., and M. Kuehnen. 2013. Determinants of sustainability reporting: A review of results, trends, theory, and opportunities in an expanding field of research. Journal of Cleaner Production 59:5–21. Hahn, R., and R. Luelfs. 2014. Legitimizing negative aspects in GRI-oriented sustainability reporting: A qualitative analysis of corporate disclosure strategies. Journal of Business Ethics 123(3): 401–420. Hail, L., A. Tahoun, and C. Wang. 2018. Corporate scandals and regulation. Journal of Accounting Research 56(2): 617–671. Harjoto, M. A., and H. Jo. 2011. Corporate governance and CSR nexus. Journal of Business Ethics 100(1): 45–67. 1868 C.Reitmaier et al. Harjoto, M. A., and H. Jo. 2015. Legal vs. normative CSR: Differential impact on analyst dispersion, stock return volatility, cost of capital, and firm value. Journal of Business Ethics 128(1): 1–20. He, J. 2006. Pollution haven hypothesis and environmental impacts of foreign direct investment: The case of industrial emission of sulfur dioxide (SO2) in Chinese provinces. Ecological Economics 60(1): 228–245. Heflin, F., and D. Wallace. 2017. The BP oil spill: Shareholder wealth effects and environmental disclosures. Journal of Business Finance & Accounting 44(3–4): 337–374. Hermalin, B. E. 2013. Leadership and corporate culture. In The handbook of organizational economics, ed. R. Gibbons and J. Roberts, 432–478. Princeton University Press. Hoi, C. K., Q. Wu, and H. Zhang. 2013. Is corporate social responsibility (CSR) associated with tax avoidance? Evidence from irresponsible CSR activities. The Accounting Review 88(6): 2025–2059. Holder-Webb, L., J. R. Cohen, L. Nath, and D. Wood. 2009. The supply of corporate social responsibility disclosures among U.S. firms. Journal of Business Ethics 84(4): 497–527. House, R. J., P. J. Hanges, M. Javidan, P. W. Dorfman, and V. Gupta. 2004. Culture, leadership, and organizations. The GLOBE study of 62 societies. SAGE Publications. Huang, X., and L. Watson. 2015. Corporate social responsibility research in accounting. Journal of Accounting Literature 34:1–16. Huang, Q., Y. Li, M. Lin, and G. A. McBrayer. 2022. Natural disasters, risk salience, and corporate ESG disclosure. Journal of Corporate Finance 72:102152. Hughes, S. B., A. Anderson, and S. Golden. 2001. Corporate environmental disclosures: Are they useful in determining environmental performance? Journal of Accounting and Public Policy 20(3): 217–240. Interbrand. 2017. Best Global Brands. Retrieved May 30, 2017, from http:// inter brand. com/ bestbrands/ bestglobalbrands/ metho dology/. Ioannou, I., and G. Serafeim. (2017). The consequences of mandatory corporate sustainability reporting.The Oxford Handbook of Corporate Social Responsibility, Vol. 2, Oxford University Press. Jantadej, P., and P. Kent. 1999. Corporate environmental disclosures in response to public awareness of the Ok Tedi Copper Mine disaster: A legitimacy theory perspective. Accounting Research Journal 12(1): 72–88. Javidan, M. 2004. Performance orientation. In Culture, leadership, and organizations The GLOBE study of 62 societies, ed. R.J. House, P.J. Hanges, M. Javidan, P.W. Dorfman, and V. Gupta, 239–281. SAGE Publications. Johnson, S. A., H. E. Ryan Jr., and Y. S. Tian. 2009. Managerial incentives and corporate fraud: The sources of incentives matter. Review of Finance 13(1): 115–145. Kabasakal, H., and M. Bodur. 2004. Humane orientation in societies, organizations, and leader attributes. In Culture, leadership, and organizations. The GLOBE study of 62 societies, ed. R.J. House, P.J. Hanges, M. Javidan, P.W. Dorfman, and V. Gupta, 564–601. SAGE Publications. Kim, E. -H., and T. P. Lyon. 2011. Strategic environmental disclosure: Evidence from the DOE’s voluntary greenhouse gas registry. Journal of Environmental Economics and Management 61(3): 311–326. Kim, J. -B., C. Wang, and F. Wu. 2023. The real effects of risk disclosure: Evidence from climate change reporting in 10-Ks. Review of Accounting Studies 28(4): 2271–2318. Kleinbaum, D. G., and M. Klein. 2010. Logistic regression. A self-learning text, 3rd ed. Springer. Kolk, A., and P. Perego. 2010. Determinants of the adoption of sustainability assurance statements: An international investigation. Business Strategy and the Environment 19(3): 182–198. Kolk, A., S. Walhain, and S. VanDeWateringen. 2001. Environmental reporting by the Fortune Global 250: Exploring the influence of nationality and sector. Business Strategy and the Environment 10(1): 15–28. Kotabe, M., and J. Y. Murray. 2004. Global sourcing strategy and sustainable competitive advantage. Industrial Marketing Management 33 (1): 7–14. Kotchen, M., and J. J. Moon, 2012. Corporate social responsibility for irresponsibility. The B.E. Journal of Economic Analysis & Policy, 12(1), Article 55. KPMG. (1999). KPMG International Survey of Environmental Reporting 1999. Retrieved March 4, 2020, from https:// www. resea rchga te. net/ publi cation/ 25479 6996_ KPMG_ Inter natio nal_ survey_ of_ envir onmen tal_ repor ting_ 1999. 1869 Corporate responsibility andcorporate misbehavior: are… KPMG. 2002. KPMG International Survey of Corporate Sustainability Reporting 2002. Retrieved March 4, 2020, from https:// www. resea rchga te. net/ publi cation/ 25474 6739_ KPMG_ Inter natio nal_ Survey_ of_ Corpo rate_ Susta inabi lity_ Repor ting_ 2002. KPMG. 2008. KPMG International Survey of Corporate Responsibility Reporting 2008. Retrieved October 21, 2020, from https:// www. in. kpmg. com/ secur edata/ aci/ files/ sustc orpor atere spons ibili tyrep ortin gsurv ey2008. pdf. KPMG. 2011. KPMG International Survey of Corporate Responsibility Reporting 2011. Retrieved October 21, 2020, from https:// assets. kpmg/ conte nt/ dam/ kpmg/ pdf/ 2012/ 02/ Corpo raterespo nsibl ityrepor ting2012eng. pdf. KPMG. 2017. The KPMG Survey of Corporate Responsibility Reporting 2017. Retrieved October 21, 2020, from https:// assets. kpmg/ conte nt/ dam/ kpmg/ xx/ pdf/ 2017/ 10/ kpmgsurveyofcorpo raterespo nsibi lityrepor ting2017. pdf. KPMG 2022. KPMG Survey of Sustainability Reporting 2022. Retrieved October 10, 2022, from https:// assets. kpmg. com/ conte nt/ dam/ kpmg/ sg/ pdf/ 2022/ 10/ ssrsmallstepsbigshifts. pdf. Krüger, P. 2015. Corporate goodness and shareholder wealth. Journal of Financial Economics 115(2): 304–329. Lambert, R., C. Leuz, and R. E. Verrecchia. 2007. Accounting information, disclosure, and the cost of capital. Journal of Accounting Research 45(2): 385–420. Larcker, D. F., and T. O. Rusticus. 2010. On the use of instrumental variables in accounting research. Journal of Accounting and Economics 49(3): 186–205. Larcker, D. F., S. A. Richardson, and I. Tuna. 2007. Corporate governance, accounting outcomes, and organizational performance. The Accounting Review 82(4): 963–1008. Larcker, D. F., L. Pomorski, B. Tayan, and E. M. Watts, 2022. ESG ratings – a compass without direction.Rock Center for Corporate Governance at Stanford University Working Paper Forthcoming. Retrieved August 9, 2024 from https:// papers. ssrn. com/ sol3/ papers. cfm? abstr act_ id= 41796 47. Lee, J., and S. Maxfield. 2015. Doing Well by Reporting Good: Reporting Corporate Responsibility and Corporate Performance. Business and Society Review 120(4): 577–606. Legendre, S., and F. Coderre. 2013. Determinants of GRI G3 application levels: The case of the Fortune Global 500. Corporate Social Responsibility and Environmental Management 20(3): 182–192. Lewbel, A. 2014. An overview of the special regression method. In The Oxford handbook of applied nonparametric and semiparametric econometrics and statistics, ed. J.S. Racine, L. Su, and A. Ullah, 38–61. Oxford University Press. Lewbel, A., Y. Dong, and T. T. Yang. 2012. Comparing features of convenient estimators for binary choice models with endogenous regressors. Canadian Journal of Economics 45(3): 809–829. Li, X., Y. Lou, and L. Zhang, 2023. Do commercial ties influence ESG ratings? Evidence from Moody’s and S&P. https:// papers. ssrn. com/ sol3/ papers. cfm? abstr act_ id= 41902 04. Lindblom, C. K. 1994. The implications of organizational legitimacy for corporate social performance and disclosure. Conference Paper, Critical Perspectives on Accounting Conference, New York, NY. Linnenluecke, M. K., and A. Griffiths. 2010. Corporate sustainability and organizational culture. Journal of World Business 45(4): 357–366. Mahoney, L. S., L. Thorne, L. Cecil, and W. LaGore. 2013. A research note on standalone corporate social responsibility reports: Signaling or greenwashing? Critical Perspectives on Accounting 24(4–5): 350–359. Maignan, I., and D. A. Ralston. 2002. Corporate social responsibility in Europe and the U.S.: Insights from businesses’ self-presentations. Journal of International Business Studies 33(3): 497–514. Mark-Ungericht, B., and R. Weiskopf. 2007. Filling the empty shell. The public debate on CSR in Austria as a paradigmatic example of a political discourse. Journal of Business Ethics 70(3): 285–297. Matsumura, E. M., R. Prakash, and S. C. Vera-Munoz. 2014. Firm-value effects of carbon emissions and carbon disclosures. The Accounting Review 89(2): 695–724. Mattingly, J. E., and S. L. Berman. 2006. Measurement of corporate social action – discovering taxonomy in the KLD ratings data. Business & Society 45(1): 20–46. McCarten, M., I. Diaz-Rainey, H. Roberts, and E. K. M. Tan. 2022. Political connections, tacit power and corporate misconduct. Journal of Business Finance & Accounting 49(9–10): 1530–1552. McGuire, J., S. Dow, and K. Argheyd. 2003. CEO incentives and CSR. Journal of Business Ethics 45(4): 341–359. Meek, G. K., C. B. Roberts, and S. J. Gray. 1995. Factors influencing voluntary annual report disclosures by U.S., U.K. and Continental European multinational corporations. Journal of International Business Studies 26(3): 555–572. 1870 C.Reitmaier et al. Merkl-Davies, D. M., and N. M. Brennan. 2007. Discretionary disclosure strategies in corporate narratives: Incremental information or impression management? Journal of Accounting Literature 26:116–194. Mishina, Y., B. J. Dykes, E. S. Block, and T. G. Pollock. 2010. Why “good” firms do bad things: The effect of high aspiration, high expectations, and prominence on the incident of corporate illegality. The Academy of Management Journal 53(4): 701–722. Moneva, J. M., P. Archel, and C. Correa. 2006. GRI and the camouflaging of corporate unsustainability. Accounting Forum 30(2): 121–137. Morris, R. D. 1987. Signalling, agency theory and accounting policy choice. Accounting and Business Research 18(69): 47–56. Muller, A., and G. Whiteman. 2009. Exploring the geography of corporate philanthropic disaster response: A study of Fortune Global 500 firms. Journal of Business Ethics 84(4): 589–603. Muttakin, M. B., and A. Khan. 2014. Determinants of corporate social disclosure: Empirical evidence from Bangladesh. Advances in Accounting 30(1): 168–175. Neu, D., H. Warsame, and K. Pedwell. 1998. Managing public impressions: Environmental disclosures in annual reports. Accounting, Organizations and Society 23(3): 265–282. Norrlof, C. 2009. Key currency competition. The Euro versus the Dollar. Cooperation and Conflict 44(4): 420–442. Paik, G. H., B. Lee, and K. R. Krumwiede. 2017. Corporate social responsibility performance and outsourcing: The case of the Bangladesh tragedy. Journal of International Accounting Research 16(1): 59–79. Patten, D. M. 1991. Exposure, legitimacy, and social disclosure. Journal of Accounting and Public Policy 10(4): 297–308. Patten, D. M. 1992. Intra-industry environmental disclosures in response to the Alaskan oil spill: A note on legitimacy theory. Accounting, Organizations and Society 17(5): 471–475. Patten, D. M. 2002. The relation between environmental performance and environmental disclosure: A research note. Accounting, Organizations and Society 27(8): 763–773. Plumlee, M., D. Brown, R. M. Hayes, and R. S. Marshall. 2015. Voluntary environmental disclosure quality and firm value: Further evidence. Journal of Accounting and Public Policy 34(4): 336–361. Prencipe, A. 2004. Proprietary costs and determinants of voluntary segment disclosure: Evidence from Italian listed companies. European Accounting Review 13(2): 319–340. Qianqian, D., and S. Rui. 2018. Peer performance and earnings management. Journal of Banking & Finance 89:125–137. Quinlan, M., and P. Sheldon. 2011. The enforcement of minimum labor standards in an era of neo-liberal globalization: An overview. The Economic and Labor Relations Review 22(2): 5–31. Raghunandan, A., and S. Rajgopal, 2023. Do socially responsible firms walk the talk? Journal of Law and Economics, forthcoming. Raghunandan, A., and S. Rajgopal. 2022. Do ESG funds make stakeholder-friendly investments? Review of Accounting Studies 27(3): 822–863. Rasche, A., and S. Waddock. 2014. Global sustainability governance and the UN Global Compact: A rejoinder to critics. Journal of Business Ethics 122 (2): 209–216. Refinitiv. 2022. Environmental, social and Governance scores from Refinitiv. Retrieved September 29, 2022, from https:// www. refin itiv. com/ conte nt/ dam/ marke ting/ en_ us/ docum ents/ metho dology/ refin itivesgscoresmetho dology. pdf. Reimann, F., J. Rauer, and L. Kaufmann. 2015. MNE subsidiaries’ strategic commitment to CSR in emerging economies: The role of administrative distance, subsidiary size, and experience in the host country. Journal of Business Ethics 132(4): 845–857. Reitmaier, C., and W. Schultze. 2017. Enhanced business reporting: Value relevance and determinants of valuation-related disclosures. Journal of Intellectual Capital 18(4): 832–867. Roberts, R. W. 1992. Determinants of corporate social responsibility disclosure: An application of stakeholder theory. Accounting, Organizations and Society 17(6): 595–612. Schultze, W., and R. Trommer. 2012. The concept of environmental performance and its measurement in empirical studies. Journal of Management Control 22:375–412. Semenova, N., and L. G. Hassel. 2015. On the validity of environmental performance metrics. Journal of Business Ethics 132(2): 249–258. Shi, H., X. Zhang, and J. Zhou. 2018. Cross-listing and CSR performance: Evidence from AH shares. Frontiers of Business Research in China 12(11): 126–140. 1871 Corporate responsibility andcorporate misbehavior: are… Simnett, R., A. Vanstraelen, and W. Chua. 2009. Assurance on sustainability reports: An international comparison. The Accounting Review 84(3): 937–967. Sinkovics, N., S. Ferdous Hoque, and R. R. Sinkovics. 2016. Rana Plaza collapse aftermath: Are CSR compliance and auditing pressures effective? Accounting, Auditing & Accountability Journal 29(4): 617–649. Skinner, D. J. 1994. Why firms voluntarily disclose bad news. Journal of Accounting Research 32(1): 38–60. Skouloudis, A., N. Jones, C. Malesios, and K. Evangelinos. 2014. Trends and determinants of corporate non-financial disclosure in Greece. Journal of Cleaner Production 68:174–188. Spence, M. 1973. Job market signaling. The Quarterly Journal of Economics 87(3): 355–374. Stock, J. H., and M. Yogo, 2002. Testing for weak instruments in linear IV regression. National Bureau of Economic Research Technical Working Paper Series 284. Retrieved May 30, 2017, from http:// www. nber. org/ papers/ T0284. Stock, J. H., and M. W. Watson. 2012. Introduction to econometrics, 3rd ed. Pearson. Strike, V. M., J. Gao, and P. Bansal. 2006. Being good while being bad: Social responsibility and the international diversification of U.S firms. Journal of International Business Studies 37(6): 850–862. Taylor, M. S. 2005. Unbundling the pollution haven hypothesis. Advances in Economic Analysis & Policy 4(2): 1–28. Thorne, L., L. S. Mahoney, and G. Manetti. 2014. Motivations for issuing standalone CSR reports: A survey of Canadian firms. Accounting, Auditing & Accountability Journal 27(4): 686–714. Tomar, S. 2023. Greenhouse Gas Disclosure and Emissions Benchmarking. Journal of Accounting Research 61(2): 451–492. Topping, N. 2012. How does sustainability disclosure drive behavior change? Journal of Applied Corporate Finance 24(2): 45–48. Treviño, L. K. 2005. Out of touch – The CEO’s role in corporate misbehavior. Brooklyn Law Review 70(4): 1195–1212. Tschopp, D. J. 2005. Corporate social responsibility: A comparison between the United States and the European Union. Corporate Social Responsibility and Environmental Management 12(1): 55–59. United Nations Global Compact (UNGC). 2020a. United Nations Global Compact. The ten principles of the UN Global Compact. Retrieved October 21, 2020, from https:// www. unglo balco mpact. org/ whatisgc/ missi on/ princ iples. United Nations Global Compact (UNGC). 2020b. Who we are. Our Participants. Retrieved October 21, 2020, from https:// www. unglo balco mpact. org/ whatisgc/ parti cipan ts. United Nations Global Compact (UNGC). 2020c. Reporting on the SDGs. Retrieved October 21, 2020, from https:// www. unglo balco mpact. org/ takeaction/ action/ sdgrepor ting. VanMarrewijk, M. 2003. Concepts and definitions of CSR and corporate sustainability: Between agency and communion. Journal of Business Ethics 44(2–3): 95–105. Watson, A., P. Shrives, and C. Marston. 2002. Voluntary disclosure of accounting ratios in the UK. The British Accounting Review 34(4): 289–313. Welford, R. 2004. Corporate social responsibility in Europe and Asia. Journal of Corporate Citizenship 13:31–47. Welford, R. 2005. Corporate social responsibility in Europe, North America and Asia. Journal of Corporate Citizenship 17:33–52. White, H. 1980. A heteroskedasticity–consistent covariance matrix estimator and a direct test for heteroskedasticity. Econometrica 48(4): 817–838. Wickert, C., A. G. Scherer, and L. J. Spence. 2016. Walking and talking corporate social responsibility: Implications of firm size and organizational cost. Journal of Management Studies 53(7): 1169–1196. Williams, O. F. 2014. The United Nations Global Compact: What did it promise? Journal of Business Ethics 122(2): 241–251. Wiseman, J. 1982. An evaluation of environmental disclosures made in corporate annual reports. Accounting, Organizations and Society 7(1): 53–63. Wooldridge, J. M. 2003. Solutions manual and supplementary materials for Econometric analysis of cross section and panel data. MIT Press. Wooldridge, J. M. 2013. Introductory econometrics. A modern approach, 5th ed. Mason. Wu, J. 2014. The antecedents of corporate social and environmental irresponsibility. Corporate Social Responsibility and Environmental Management 21(5): 286–300. 1872 C.Reitmaier et al. Yoon, A. and Serafeim, G. 2020. Stock price reactions to ESG news: The role of ESG ratings and disagreement. https:// schol arspa ce. manoa. hawaii. edu/ server/ api/ core/ bitst reams/ b3b44 b92706b4474aab2a20a6 e51bd 2c/ conte nt.Accessed 1 Jul 2024. Young, S., and M. Marais. 2012. A multi-level perspective of CSR reporting: The implications of national institutions and industry risk characteristics. Corporate Governance: An International Review 20(5): 432–450. Zhao, M. 2012. CSR-based political legitimacy strategy: Managing the state by doing good in China and Russia. Journal of Business Ethics 111(4): 439–460. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.