The contents of assurance statements for sustainability reports and information asymmetry

Abstract

This paper investigates how the assurance of sustainability reports enhances the credibility of such reports in the eyes of the investors and, thus, results in lower information asymmetries, as measured by bid-ask spreads. We measure the assurance of sustainability reports based on a content analysis of the assurance statements in which the assurance providers describe the design of the assurance process. For a matched sample of 442 STOXX 600 Europe companies with and without assured sustainability reports, our results indicate that a high-quality design of the assurance process reduces the level of information asymmetry. While an assurance process substantiating a high assurance level decreases information asymmetries, an assurance process that ensures only a moderate assurance level is insufficient. If an assurance provider performs tests of details of numerical data, this further reduces information asymmetries. For countries without regulations on sustainability reporting, we provide evidence that analytical tests of aggregated indicators, the description of the assurance provider’s competencies and the description of the sustainability assurance-specific work steps also contribute to a reduction of information asymmetries.

Keywords:

• voluntary assurance
• ISAE 3000
• AA1000AS
• information asymmetry
• GRI
• sustainability reporting

Acknowledgments

The authors appreciate the helpful comments and remarks of Michael Stich, Daniel W. Collins, and Edward J. Riedl and of participants of the 2013 Eighth Accounting Research Workshop at the University of Basle, Switzerland, and the 2013 4th WHU Summer Program in Accounting Research on previous versions of the paper. Furthermore, the authors would like to thank the joint editor, Juan Manuel Garcia Lara, and the associate editor, Chris van Staden, and two anonymous reviewers for very fruitful comments and recommendations which helped us to further improve our paper.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1 Simnett et al. (2009b) report that the share of assured sustainability reports is 6% in the US compared with 36% in other countries, and KPMG (2008) reports that 14% of US companies (compared with 30–73% of European companies) issue assured sustainability reports.

2 This paper is based on the GRI G3 guidelines because these guidelines were valid when the data were collected. There are no material changes regarding the assurance of GRI reports in the GRI G4 guidelines, which were launched in July 2013 (see GRI 2013)

3 Besides matching on the level of the decision to add an assurance statement, we could also match the companies based on their decision to publish a sustainability report. However, the characteristics of companies that provide a sustainability report do not have to be also determinants of the decision to add an assurance statement and to include particular assurance statement contents. Endogeneity is more likely to be an issue with regard to the content analytical variables that represent the contents of the assurance statements and the decision to add an assurance statement.

4 In addition to the analysis based on PSM, we conducted an analysis based on the procedure recommended by Chenhall and Moers (2007). The first step of this procedure is based on an auxiliary logit regression with the same explanatory variables as in Equation (1). In a second step, the residuals of that auxiliary regression are used as an additional independent variable in the OLS regression. Overall, this procedure results in similar results, which are available upon request.

5 Table 3 reports the results of difference of means tests for the comparison of the treatment and the control group. The results show that both groups are very similar with respect to almost all firm characteristics, except for the market value and the matching variable (i.e. the decision to add an assurance statement).

6 We also considered a broader definition of concomitant disclosures with a control variable based on the total number of press releases for each firm-year observation. This control variable is not significantly associated with the level of information asymmetry and does not affect the results of our content analytical variables.

7 Our subsample of Non-Big 4 auditors includes 68 assurance statements from sustainability specialists and only 3 from smaller auditing firms.

8 In an additional untabulated analysis based on model (2), we included an additional Big 4 dummy that measures possible interrelations between the bid-ask spread and the occurrence of a Big 4 assurance statement. However, the coefficients for both the assurance dummy (Assurance) and the Big 4 dummy (Big4) are not significant.