The effect of the interest coverage covenants on classification shifting of revenues

ABSTRACT

While prior studies focus on real/accrual-based earnings management and expense misclassification to investigate earnings manipulation in avoiding covenant violations, this paper extends such research in a new direction. In particular, it examines whether firms employ classification shifting of revenues when they are subject to interest coverage EBITDA-based covenants close to their threshold values or limits. This earnings management tool allows firms to increase reported EBITDA by misclassifying non-operating revenues as operating revenues to remain within covenant limits that include EBITDA. Using a sample of 559 UK listed firm-years for the period 2005–2014, it establishes that the use of classification shifting of revenues is high when interest coverage covenants are close to their limits. Further analysis suggests that firms also employ revenue shifting when all their loan covenants are EBITDA-related.

KEYWORDS:

• Classification shifting of revenues
• interest coverage covenant
• debt covenant violations

Acknowledgement

We are grateful to four anonymous referees for their helpful and constructive comments. These have helped us to improve the content and exposition of the paper.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1 Tight covenant slack is defined as situations where a company is close to its covenant threshold value.

2 Ertimur, Livnat, and Martikainen (2003) and Marquardt and Wiedman (2004) show that investors value a dollar of operating revenues surprises greater than a dollar decrease in operating expenses.

3 This is not the case in all countries. For example, the USA has a debtor-friendly bankruptcy code which means that banks in the USA cannot place firms directly into liquidation if the firms file for Chapter 11 bankruptcy protection (Acharya, Sundaram, and John 2011; Li, Lou, and Vasvari 2015). The latter gives firms 120 days to recapitalize and potentially find a solution to their financial difficulties.

4 Similar results are found by Kim, Lisic, and Pevzner (2010) who employ net worth covenants.

5 One exception for this is the study by Hsu and Kross (2011) who provide evidence that one motivation for showing non-operating revenues, transitory gains, as part of operating revenues is to meet/beat operating earnings benchmarks.

6 Dead firms are included across the test period to avoid survivorship bias.

7 Note, UK firms quoted on the Alternative Investment Market were required to follow IFRS from 2007. Therefore, for these firms, our sample period begins in 2007.

8 While annual reports highlight that these firms are subject to interest coverage covenants, they do not reveal their threshold values.

9 Our results do not change if we define tight covenant slack as firms that have interest coverage covenant slack within the lower quartile.

10 We estimate the model cross-sectionally for each industry-year using all UK firms included in Compustat Global. We employ the Global Industry Classification Scheme and require, following Athanasakou, Strong, and Walker (2009), at least 6 observations per industry-year to ensure that we have sufficient data for the estimation of classification shifting of revenues EM measure.

11 In other words, unexpected operating revenues are measured as the residuals from the regression model in equation (2).

12 The results do not change if we employ other measures of accruals EM such as total discretionary accruals (e.g. Peasnell, Pope, and Young 2000, 2005).

13 The following covenants are regarded as EBITDA-related covenants: (1) interest coverage, (2) debt to EBITDA, (3) fixed charge coverage, (4) debt service coverage, (5) EBITDA, and (6) EBITDA to net debt. In our sample, 842 firm-year observations out of 1,272 have only EBITDA-related covenants.

14 The results do not change if we use a continuous variable that is defined as the number of EBITDA-related covenants divided by the number of all financial covenants that a firm has in a specific year.

15 As can be seen in Table 1, the use of other covenants is not really common in the UK and therefore we do not consider them.

Additional information

Funding

Coakley gratefully acknowledges support from grant number ES/L011859/1, from the Business and Local Government Data Research Centre, funded by the Economic and Social Research Council.