10-K Disclosure of Corporate Social Responsibility and Firms’ Competitive Advantages

Abstract

In this paper, we offer evidence that disclosure of corporate social responsibility (CSR) in the 10-K provides information about firms’ competitive advantages. We create a textual measure of CSR disclosure that aggregates CSR keywords found in 10-Ks. We measure firms’ competitive advantages by using levels of and persistence of industry-adjusted gross margin; selling, general, and administrative (SG&A) margin; and operating margin. We first show that 10-K CSR disclosure intensity is associated with lower levels of gross margin but higher SG&A margin. We then observe that firms with more CSR keywords in their 10-Ks maintain more persistent above-industry-median gross and operating margins. Upon sorting our full set of 10-K CSR keywords into three subcategories (philanthropy, business practice, and product), we find that the intensities of disclosure in these CSR subcategories have associations with varying competitive advantages.

Keywords:

• Corporate social responsibility
• Non-financial disclosure
• Competitive advantage
• Operating performance

JEL Classifications:

• M14
• M41
• L29

Acknowledgement

We thank the editors, Reuven Lehavy and Florin Vasvari, two anonymous reviewers, Anna Brown, Brian Knox, Sue Ravenscroft, Yan Sun, Kimberly Swanson, Ke Wang, Amanda Wilford, seminar participants at Iowa State University, the 2015 Conference on Convergence of Financial and Managerial Accounting Research, the 2015 BYU Accounting Research Symposium, the 2015 AAA Midwest Region Meeting, the 2016 AAA Annual Meeting, and the 2017 Congress of the European Accounting Association, for their valuable comments and suggestions. We also thank Sen Wang for his excellent research assistance.

Data Availability

All data are publicly available from the sources identified in the text.

Disclosure statement

No potential conflict of interest was reported by the authors.

Supplemental Data and Research Materials

Supplemental data for this article can be accessed on the Taylor & Francis website, doi:10.1080/09638180.2019.1670223

Appendix S1. Select examples of CSR disclosure in 10-K filings

Appendix S2. CSR disclosure keyword vocabulary

Appendix S3. Summary of Bill McDonald’s detailed textual data parsing procedures

Appendix S4. Univariate analysis

ORCID

James N. Cannon http://orcid.org/0000-0003-0099-3236

Zhejia Ling http://orcid.org/0000-0002-0977-9263

Qian Wang http://orcid.org/0000-0002-5756-4400

Olena V. Watanabe http://orcid.org/0000-0002-1719-9023

Notes

1 All industry groups are defined based on the two-digit Standard Industry Classification (SIC) codes.

2 Harris (1998) argues that a significantly positive persistence of above-industry earnings suggests the inability of rival firms in the industry to drive down a firm’s profitability to a normal rate of return.

3 We acknowledge that there is always a potential for self-selection bias when managers have discretion over voluntary disclosure and, to some degree, discretion over their disclosure response to mandatory reporting standards. It is possible that managers choose to disclose information about CSR activities when they expect performance to reflect a competitive advantage unrelated to CSR. Alternatively, managers may engage in CSR activities when they anticipate a competitive advantage that is unrelated to CSR. However, in both cases, we would not have reason to expect performance to vary across disclosures of different types of CSR as we find in our empirical analyses. However, we draw no causal inferences but instead document associations between CSR disclosure and firm competitive advantage.

4 We note that measuring the quality of CSR performance to gauge the truthfulness of disclosure claims is not the focus of our study. Rather, we argue that 10-K CSR disclosure represents managers’ desire to convey information about competitive advantage.

5 These and other comment letters are available at https://www.sec.gov/comments/s7-06-16/s70616.htm.

6 In our observation, the two sections in 10-K filings that are most likely to include CSR disclosure are ‘Part I: Business’ and ‘Management Discussion and Analysis.’ For illustration, in supplementary online Appendix S1, we present extracts of text with related CSR disclosures from 10-Ks for four company-years.

7 It is also possible that firms use CSR disclosure to legitimize their subpar CSR or financial performance. That is, firm managers may use CSR disclosure to enhance readers’ perception of the firm’s purported CSR performance (Cho, Guidry, Hageman, & Patten, 2012; Clarkson, Li, Richardson, & Vasvari, 2008). Or, as an extension, firm managers may use CSR disclosure to attribute poor operating performance to their engagement in CSR activities (e.g., Hummel & Schlick, 2016). In the former case, we would expect no association between CSR disclosure and competitive advantage, while in the latter case, we would expect a negative association between CSR disclosure and competitive advantage.

8 We note that CSR activities impact not only business-to-consumer transactions but also business-to-business dealings. For example, Drumwright (1994) finds that social responsibility factors into corporations’ buying decisions in business-to-business transactional settings.

9 In a parallel line of thought, Lev (1983) argues that monopolistic firms would exhibit more persistent earnings than would competing firms. By extension, we expect that a monopolist’s market power is similar to a competitive advantage for a particular firm in a competitive industry.

10 Supporting this view, Clarkson, Li, Richardson, and Vasvari (2011) show that firms that dedicate significant resources to environmental strategies experience subsequent increases in cash flows and performance. Similarly, Radhakrishnan et al. (2018) cite numerous cases in which firms invested significant resources to achieve CSR-related competitive advantages.

11 As discussed later, in section 3, we measure ‘operating performance’ using industry-median-adjusted gross margin, SG&A margin, and operating margin to capture various types of competitive advantage.

12 See Auger et al. (2003); Choi and La (2013); Elliott and Freeman (2004); Folkes and Kamins (1999); and Mandhachitara and Poolthong (2011) for examples.

13 The distinction between product and business practice CSR activities lies in whether the CSR attribute is a feature intrinsic to the product (e.g., organic coffee) or is part of the process through which the product was created (e.g., fair-trade coffee) (Peloza & Shang, 2011).

14 We recognize that some may consider the entire gambling and alcohol industries’ operations to be socially irresponsible, due to risks of consumer addiction. However, in this study, we omitted these terms in favor of capturing within-industry variation in disclosed CSR activities.

15 Three of the co-authors independently assigned the keywords to Peloza and Shang’s (2011) categories. The fourth co-author reconciled the differences between the three codings and finalized the keyword classifications.

16 Following a paper by Bostwick et al. (2016), who find that significant discrepancies may exist between the Compustat cost-of-goods-sold data item and cost of goods sold as reported in the 10-Ks, we compute cost of goods sold using a Compustat adjustment procedure to reduce the discrepancy. Our inferences are not qualitatively different when we do not perform this adjustment procedure.

17 Less persistent negative industry-adjusted performance would reflect a quicker erosion of competitive disadvantage in firms that disclose CSR in their 10-Ks.

18 The repository is available online at https://sraf.nd.edu/data/. The detailed parsing procedures are described in our online Appendix S3.

19 We calculate the industry-median operating performance benchmark using the entire Compustat population. The median values of the operating performance measures (adjGM, adjSGAM, adjOM, adjSales, adjCOGS, and adjSGA) differ from zero as a result of the sample reduction criteria described in subsection 4.1. Specifically, we find that the reduced sample is skewed toward firms with better-than-industry-median operating performance.

20 We thank the referees for suggesting that we separately examine components of margins.

21 Given that adjSales is the dependent variable in our analysis of sales, we omit it as a control variable in that model specification.

22 We procured information on firms that publish stand-alone CSR reports from the Corporate Register (http://www.corporateregister.com/).

23 KLD uses a number of positive indicators (strengths) and negative indicators (concerns) to assess social performance along six dimensions: community, diversity, employee relations, environment, human rights, and product. Following prior literature, we calculate KLD score as total strengths minus total concerns in KLD’s six social rating categories, reported by the MSCI (e.g., Kim et al., 2012).

24 We thank the anonymous reviewer for bringing this issue to our attention.