Disclosure on the Sustainable Development Goals – Evidence from Europe

Abstract

This study examines disclosure on the sustainable development goals (SDGs) in firms’ annual reports. For a sample of European firms listed in the STOXX Europe-600 index and a reporting period of four years, we use textual analysis to assess both firms’ explicit reference to the SDGs in their annual reports as well as the implicit prevalence of SDG topics. In addition, we use content analysis to manually assess the quality of firms’ disclosure on the SDGs based on eleven reporting items. The results show a substantial increase in SDG reporting quality over time but a distinct lack of disclosure of quantitative and forward-looking information. Further analyses reveal the relevance of both financial and non-financial stakeholders. Specifically, SDG disclosure is particularly associated with a high relevance of socially responsible investors, customers or environment-related public pressure, while financial analysts, employees and the media are not associated with SDG disclosure.

Keywords:

• sustainable development goals
• disclosure
• stakeholders
• textual analysis
• content analysis

Disclosure Statement

No potential conflict of interest was reported by the author(s).

Notes

1 Following Guenther et al. (2016, p. 365), stakeholder relevance is defined as the ‘measurable influence of a specific stakeholder group on the decisions of the firm.’

2 The targets are further delineated by the ‘Global indicator framework for the Sustainable Development Goals and targets of the 2030 Agenda for Sustainable Development’ provided by the IAEG-SDG (A/RES/71/313).

3 In addition, there is an online inventory that lists business tools and indicators (https://sdgcompass.org/business-tools/ and https://sdgcompass.org/business-indicators/).

4 Regarding the term ‘non-financial information’, Haller et al. (2017) ascertain the existence of various different definitions and interpretations of the term. In the context of the recent EU Directive 2014/95/EU, the authors therefore call for the issuance of regulatory guidelines to clearly define the term. Similarly, Carroll (2015) also notes the existence of various ‘competing and complementary concepts’ and terms for corporate social responsibility (CSR) and sustainability. He concludes that ‘none of these frameworks […] appear to be sufficiently distinct’ and thus refers to the traditional and accepted term of CSR (p. 93).

5 The underlying sustainability performance assessment is provided by RobecoSAM, an asset management and rating agency. Lai et al. (2018) have also documented the importance of rating agencies for the preparers of integrated reports.

6 These findings are moderated by Lai et al. (2016), who do not find evidence that the adoption of integrated reporting is related to industry membership.

7 Files that cannot be processed in the textual analysis are PDF files with copy protection.

8 First, we extract the text from the PDF files and generate txt files. Then, we eliminate all line breaks and tabulators, all Unicode-wide characters and all blanks that occur several times in sequence and divide the text into single words (tokens). Next, we eliminate words that consist of only one character and stop words, as provided by McDonald (2017). In addition, we exclude the names of the firms that are included in our sample. Finally, we apply a stemming algorithm that collapses words down to their word stem (e.g. ‘caring’, ‘cares’ and ‘cared’ to the root word ‘care’) (Porter, 1980).

9 # denotes the numbers 1-17.

10 This manual assessment revealed that all reports with SDG_BINARY=1 refer explicitly to the SDGs as defined by our search query. Thus, the results from textual analysis appear to be valid.

11 To further improve our ability to capture disclosure on the broad topics the SDGs encompass without requiring the firms to explicitly refer to the SDG framework, we also include the firms’ annual reports from 2014 for the construction of the vocabulary.

12 As is true for all CSR rating providers, Thomson Reuters ASSET4 relies on firms’ self-reported data to construct the overall CSR performance score. Thus, in the absence of a reporting mandate, there is a self-selection problem in the sense that firms choose their CSR activities and the subsequent disclosure on these activities based on cost-benefit considerations (Christensen et al., 2019). We argue that this problem is less severe in our setting since we focus on a very specific type of CSR disclosure, namely, disclosure on the achievements of the SDGs. Moreover, Thomson Reuters ASSET4 relies on a broad set of disclosure channels when assessing a firm’s CSR performance, whereas our SDG disclosure only accounts for disclosure in firms’ annual reports.

13 Specifically, the adoption of Directive 2014/95/EU in 2014 resulting in a mandatory non-financial reporting requirement in European countries from 2017 onwards might affect our dependent variable. For instance, Aureli et al. (2019) document significant heterogeneity in a country’s legal changes resulting from the implementation of the Directive. For a discussion of the transposition of the Directive in German law see (Kajüter, 2017).

14 Note that the results do not substantially change if we run logit regression for SDG_BINARY and if we run the regressions for SDG_QUAL for the reduced sample of SDG reporters, i.e., n=652.

15 Beginning with reporting year 2017, such a reporting mandate applies for all sample firms that are located in European Union member countries due to the non-financial reporting directive 2014/95/EU.

16 The scaling accounts for a general increase in news articles over the sample period.

17 For the list of employers, see https://www.forbes.com/lists/worlds-best-employers/#24a28bdb1e0c.