How far can we trust earnings numbers? What research tells us about earnings management

Abstract

The article reviews the recent academic research literature on earnings management (EM) with a view to identifying research themes and results of interest to users and preparers of financial statements, accounting standard setters, and others with responsibility for ensuring that companies provide financial information to shareholders that can be relied upon. Hopefully students of accounting with an interest in exploring the EM literature will find that the article provides a useful framework. The literature on this topic is vast, and it is not possible to cover every article in detail. I provide an impressionistic survey that highlights examples of specific research themes and methods that regularly appear in the literature. Most of the examples are chosen from the literature published since 2000, although I do also highlight a few methodological contributions that appeared earlier. It is inevitable that the selection of articles reflects to some extent my personal interests and biases (intentional or otherwise). Thus, I wish to acknowledge that I owe a very substantial intellectual debt to the insights and contributions of the many uncited authors of a literature that spans over 40 years in over 20 accounting and finance journals.

Keywords:

• capital market effects
• capital market motives
• contractual motives
• earnings management
• discretionary accruals
• real earnings management
• third-party motives

Acknowledgements

This article was written as the basis for the 2012 P.D. Leake lecture. I am grateful to the ICAEW and the P.D. Leake trust for encouraging me to undertake this task, to Robert Hodgkinson and Brian Singleton-Green for guidance during the writing of this lecture, and to Garen Markarian for helpful comments. Pauline Weetman, the editor of ABR, also gave many useful comments and suggestions. The comments of an anonymous reviewer are acknowledged with thanks. All remaining errors and omissions are my responsibility.

Notes

I have not covered any of the pre-1970 literature, but the origins of work on EM date back at least to the late nineteenth century.

See Levitt (1998), regarding SEC concerns of what he called ‘The Numbers Game’. Similarly Bob Herz and Mary Keegan wrote disparagingly about what they called ‘The Earnings Game’ (see Eccles et al. 2001).

By ‘free cash flows are given’, we mean that operating cash flows and investment cash flows are determined by management in advance, before any EM decisions are made.

Jorissen and Otley (2010) provide an interesting case study of two financially interdependent airlines that went bankrupt.

Myerson (1985) provides an authoritative overview of how the revelation principle was discovered.

I discuss this point further below in relation to capital market motives for earnings management.

Other useful surveys can be found in Mulford and Comiskey (2002), and Nelson et al. (2003).

In my public presentation of this article, I referred to the empirical archival literature as the ‘Big Muddy’ of accounting research. This was inspired by the song ‘Waist deep in the big muddy’, by Pete Seeger.

I define ‘free cash flows’ as cash flow from operations minus net cash investment in operations. Thus accruals in Equation (1) are defined very broadly to include all operating income accruals and the recognition of investments.

The original Jones model was developed as a time series model but, following Dechow et al. (1995), most researchers now estimate normal accrual models using cross section data, typically by industry-year.

The number of major international audit firms changes over time ranging from big eight to, more recently, big four.

Atieh and Hussain (2012) report the UK evidence on this phenomenon.

This methodology involves testing for a discontinuity in the distribution of a variable around a threshold. For example, in the case of loss avoidance one can construct a histogram of earnings surprises around zero, and compare the number of observations just above zero with the number of observations just below zero.

As book to market relies on an accounting measure of book value this could be influenced by accounting choice.

Myers et al. (2003) also present evidence on auditor tenure and earnings quality.

Following the SOX act of 2002, Regulation G was introduced by the SEC to regulate the reporting of non-GAAP financial measures.

‘Balance sheet bloat’ refers to the accumulation of earnings over a number of years that is much greater than the accumulation of free cash flows over the same period.