نظریه افشای داوطلبانه و متغیرهای کنترل مالی:ارزیابی پژهش های اخیر افشای محیطی

A growing number of environmental disclosure studies are using financial control variables based on arguments from the voluntary disclosure theory (VDT). The VDT justifications for these controls are based on assumptions that disclosure is used as a tool for reducing information asymmetry between managers and investors. Given the findings reported in a broad sample of legitimacy-based environmental disclosure studies, we question whether the disclosures are primarily aimed at the market, and as such attempt to assess evidence to date on the relation between VDT financial control variables and differences in environmental disclosure. Based on a review of thirteen recent environmental disclosure studies including VDT financial control variables in their analyses, we fail to find, with the exception of firm size, evidence suggesting any systemic associations. Further, we assess whether including VDT financial control variables changes the inferences on the relation between environmental performance and environmental disclosure in one recent legitimacy-based study (Cho & Patten, 2007) and find that even with the controls, a negative association between performance and disclosure still exists. Overall, we question the need for VDT financial control variables in environmental disclosure research, but encourage further exploration of the relations using more consistent measures and media of disclosure.

A considerable body of research over the past two decades focuses on corporate environmental disclosure and what drives differences in the information provision across firms, industries, or time (e.g., Deegan and Gordon, 1996, Gao et al., 2005, Gray et al., 1995 and Guthrie et al., 2008). Many of the investigations rely upon statistical models to determine the significance of various factors posited to influence the disclosure (see, e.g., Aerts and Cormier, 2009, Brown and Deegan, 1998, Patten, 2002a, Patten, 2002b and Wilmshurst and Frost, 2000, although also see Cho, 2009, O’Donovan, 2002 and Laine, 2009 for examples of other approaches). However, as noted by Kadera and Mitchell (2005, p. 273), “model specification is a ubiquitous challenge in the social sciences” and has led to, among other things, concerns with the use of control variables in empirical analyses. Within the social and environmental accounting domain, a growing number of environmental disclosure studies adopt arguments from the economics-based voluntary disclosure theory (VDT) literature as justification for the inclusion of financial control variables in the explanatory models used (e.g., Bewley and Li, 2000, Clarkson et al., 2008, Cormier and Magnan, 1999 and Magness, 2006). Starr (2005, p. 360) argues that the inclusion of control variables in empirical models should be based on good theoretical reasons and only after “fairly extensive preliminary data analysis reveals… the form of the relationship.” In spite of this, we are aware of no attempts to date to assess either the theoretical justifications for the application of VDT arguments to the use of financial control variables in environmental disclosure research or the strength of the proposed relations. A careful review of the articles forming the foundation for the VDT arguments (Dye, 1985, Lang and Lundholm, 1993 and Verrecchia, 1983) indicates that all specifically address disclosure as a tool of communication to market participants and none specifically addresses the provision of environmental information. As such, if corporate environmental disclosure is made primarily to reduce information asymmetries between managers and investors, the application of VDT to this practice would be warranted. However, social–political theories of disclosure (see Gray et al., 1995) argue that instead of informing shareholders, corporate social and environmental disclosure is used more as a tool of impression management to reduce the exposures companies face owing to social and political pressures (Patten, 1991 and Walden and Schwartz, 1997). As noted by Deegan, 2002 and Deegan, 2007 and others, there is considerable empirical support for these social–political arguments. As such, we question the applicability of VDT as justification for the use of financial control variables in corporate environmental disclosure research. In order to assess evidence to date on the relation between VDT financial control variables and corporate environmental disclosure we review the findings from thirteen recent environmental disclosure studies that include VDT variables in their models. With the exception of firm size, we find no consistent patterns of a significant relation between the financial control variables and environmental disclosure. We next investigate whether omission of VDT financial control variables in one specific legitimacy-based research study, Cho and Patten (2007), may have led to erroneous inferences regarding the relation between environmental performance and environmental disclosure. Using data from Cho and Patten (2007) supplemented with VDT financial control variables used by Clarkson et al. (2008), we find that, consistent with the results originally presented by Cho and Patten (2007), environmental performance continues to exhibit a significant negative relation to disclosure. Thus, the differences in relation between environmental performance and environmental disclosure reported by Cho and Patten (2007) and Clarkson et al. (2008) do not appear to be due to VDT financial control variables. Overall, we fail to find meaningful evidence supporting the need for VDT financial control variables in environmental disclosure research. It is important to note that we are not dismissing VDT as a basis for trying to understand corporate environmental disclosure. It is plausible that some firms with superior but publicly unobservable environmental performance may wish to signal this to their stakeholders. The issue we are raising is that VDT-based models seem to have been adapted from the financial disclosure literature without careful consideration of whether the control variables relevant to that body of work are equally as relevant in explaining disclosure targeted at a different stakeholder group. Indeed, our results suggest these variables may not be relevant. However, we concede that the body of work to date, particularly due to differences in environmental disclosure measures used and the media of disclosure examined, as well as differences across country and time in the samples investigated, may not be sufficient to uncover existing systemic relations. As such, additional research focusing on consistency between these factors should be encouraged. We begin our examination with a review of the accounting articles that principally support the VDT arguments.

A growing number of environmental disclosure studies have begun using VDT-based financial control variables in their analyses. And while careful consideration of potentially omitted variables is important in empirical research, Starr (2005) notes that it is also important to assure that measures included for control purposes be theoretically justified and that the form of the proposed relations be well specified. The concern we have identified in this examination is that VDT justification for financial control variables appears to hinge on the argument that environmental disclosure is used by corporations as a tool for reducing information asymmetries between managers and investors, and based on evidence from the legitimacy-based research, that is, at best, a questionable premise. We thus question the existence of a sound theoretical argument for inclusion of these control variables. Further, our review of 13 environmental disclosure studies using VDT-based financial control variables fails to find, with the exception of firm size, a consistent association between the control measures and differences in corporate environmental disclosure. The lack of a systemic relation holds when examining disclosures deemed more likely to be targeted at market participants (financial report environmental disclosures and items of financial environmental information). Finally, our extension of Cho and Patten's (2007) legitimacy-based examination of environmental disclosure shows that the negative relation between the disclosure and firm environmental performance reported in the original continues to hold when financial control variables are included in the analysis. In sum, evidence to date does not support the need for including financial control variables in environmental disclosure research. It is important to note that our investigation centers only on the use of financial control variables in environmental disclosure research and is not meant to be a critique of the potential role that VDT might play in explaining at least some corporate environmental disclosure. Firms with unobservable superior environmental performance would clearly have incentives for signaling this information to its relevant stakeholders.8 Our concern is that the use of financial control variables in the VDT-based environmental disclosure work seems to be based on arguments that, while relevant for examinations of financial disclosure quality (because it is targeted at market participants), may not be valid for explaining differences in environmental disclosure. Rather than focusing on controls for observable financial performance, VDT-based environmental disclosure research should perhaps, instead, attempt to control for differences in observable environmental performance in order to isolate any potential signaling due to unobservable superior environmental performance by disclosing firms. Although our analysis fails to find evidence supporting the use of financial control variables in environmental disclosure research, we concede the body of work in this area to date is limited. The lack of comparability across disclosure measures, media of disclosure, and sample attributes all may be leading to difficulties in uncovering systemic relations. As such, we encourage future researchers exploring VDT-based arguments for environmental disclosure to focus more specifically on the use of consistent measures and media of disclosure. This may be particularly relevant given the movement of environmental reporting from financial reports to standalone sustainability reporting documents.