Industry competition and non-GAAP disclosures

Abstract

We examine the role of industry-level product market competition on non-GAAP disclosure decisions. We consider traditional measures of industry competition (concentration, price-cost margin, and set up costs), and large reductions in import tariff rates that identify an exogenous increase in competition. We find that competition intensity influences the likelihood of non-GAAP disclosure and the magnitude of non-GAAP exclusions. Our evidence suggests that strong competition encourages managers to disclose higher non-GAAP earnings. However, when competition is strong, firms with low performance relatively to the industry exclude smaller amounts. We also find that in competitive environments, managers are more likely to provide reconciliations and are less likely to exclude recurring items that are commonly excluded by other firms in the industry. These findings indicate that industry competition has a positive influence on the transparency of non-GAAP disclosures.

Keywords:

• proforma earnings
• alternative performance measures
• non-IFRS earnings
• proprietary costs
• financial performance
• industry concentration

JEL Classification:

• M21
• M41

Acknowledgements

The authors greatly acknowledge the helpful comments of Giovanna Michelon and seminar participants at the Auckland University of Technology, Cass Business School, University of St. Gallen, and Munich School of Management. We also recognise the helpful comments received at the Lisbon Accounting conference, the Accounting and Finance Research Forum of University of Western Australia, the European Accounting Association Annual Congress, and the mid-year meeting of the International Accounting Section of the American Accounting Association. This work was supported by the Foundation for Science and Technology (grant PTDC/IIM-GES/2686/2014 and UID/GES/00315/2019). We are grateful for the excellent research assistance of Filipa Nunes.

Disclosure statement

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

Notes

1 In response to specific congressional directions contained in the Sarbanes-Oxley Act, the SEC issued Regulation G in 2003, establishing strict rules on non-GAAP disclosure. The SEC has also designated a taskforce to scrutinize potentially misleading non-GAAP disclosures, and in 2016 issued Compliance and Disclosure Interpretations on non-GAAP reporting. In addition, the SEC has increased monitoring of non-GAAP metrics resulting in more comments letters sent to companies regarding their non-GAAP reporting practices. In contrast, European regulators have only issued recommendations during our sample period. The Committee of European Securities Regulators issued non-GAAP guidelines in 2005 (CESR 2005), but a follow-up study by the European Financial Reporting Advisory Group, the entity providing advice to the European Commission on reporting matters, concluded that most companies do not followed the guidelines (EFRAG 2009). More recently, the European Securities and Markets Authority (ESMA), CESR’s successor, issued guidelines for the transparent disclosure of non-GAAP measures (ESMA 2015).

2 A striking example is: in early 2005 eBay reported that it had missed fourth-quarter 2004 consensus estimate by just one penny and saw its share price plunge 22% (McKinsey & Company 2013).

3 See also Boot and Thakor (2001) and Verrecchia (2001) on voluntary full disclosure incentives.

4 Differently from Brown et al. (2018), we consider performance relative to the industry to be poor if ROA is in the lowest industry decile.

5 These negative effects critically depend on whether investors and others stakeholders can verify non-GAAP exclusions. Verifiability is possible if the firm provides a complete reconciliation, and the exclusions correspond to the total value of the line items mentioned. When firms do not provide a reconciliation (as it is common in the European setting), investors and competitors can partially verify the recurring nature of non-GAAP exclusions ex-post, against GAAP and non-GAAP measures reported in future periods. As non-GAAP disclosure is relatively sticky, investors will also have prior beliefs about the firm’s non-GAAP disclosure behavior.

6 We select only non-GAAP measures that portray earnings, to ensure comparability with the GAAP earnings numbers.

7 When firms report non-GAAP earnings in levels rather than on a per share basis we divide non-GAAP earnings by the number of shares outstanding. To obtain a relative measure of non-GAAP exclusions and to account for scale effects, we express the variable as a proportion of stock price at the beginning of the period.

8 A large number of firms indicates more competition, whereas a higher concentration of sales indicates less competition. To facilitate interpretation of the results, we multiply the HH and four-firm sales by minus one so that all variables represent a high level of competition (i.e. low concentration). We then extract one principal component, as only one component has an eigenvalue higher than one.

9 In cases of no disclosure in the prior year, we code this variable as zero.

10 We collect measures of non-GAAP earnings per share, non-GAAP net income, and adjusted versions of EBITDA and EBIT.

11 The small negative mean value of NG exclusions (-0.03) is a result of 16% of our sample having non-GAAP earnings that are less than GAAP earnings. However, only 0.1% of these cases report negative NG exclusions when GAAP earnings miss analysts’ expectations. To examine whether these percentages reflect outliers that could affect our results we repeat the regression analysis using a winsorized dependent variable at the 1% and 2% top and bottom of the distribution. Additionally, we estimate a robust regression analysis following the suggestion of Leone et al. (2019) that robust regression is an effective method to deal with outliers. Our results do not change. We also performed a Cook’s distance analysis and do not find any cases where Cook’s exceed the usual threshold of one.

12 An alternative explanation is that a special item happened, that caused both effects.

13 Collinearity is considered high if the variance inflation factors exceed 10 (Belsley et al. 1980).

14 In our study, we model the selection decision. Alternatively, we can avoid the selection concern by estimating the magnitude of non-GAAP exclusions for the subsample of disclosing firms (Doyle et al. 2003, Brown et al. 2012, Lennox et al. 2012). In Appendix 2, we present the results for this alternative method. Our conclusions do not change.

15 Available at http://faculty.som.yale.edu/peterschott/.

16 We could also explore the top performers in the industry, but we prefer to test the bottom industry performers because our hypothesis development is focused on relatively lower performance.

Additional information

Funding

This work was supported by Foundation for Science and Technology [grant number PTDC/IIM-GES/2686/2014, UID/GES/00315/2019].