Earnings Management: A Methodological Review of the Distribution of Reported Earnings Approach

Abstract

Hayn (1995) first introduced the concept of the pooled, cross-sectional distribution of reported earnings approach to assess whether there is any evidence of earnings management. This approach was further developed by Burgstahler and Dichev (1997) and since then, a substantial volume of new research has applied this methodology to alternative earnings thresholds and in different operational settings. This paper evaluates some aspects of the methodology and its applicability to different settings and contexts. This paper shows that the choice of interval width is a critical consideration in the distribution of reported earnings approach and, therefore, results should include a variety of interval widths to illustrate the robustness of findings. It also demonstrates that when the peak of the distribution falls adjacent to a threshold, the distribution of reported earnings approach will not provide statistically reliable and robust results. Finally, it concludes that the assumption of symmetry used by Burgstahler and Dichev (1997) to test for the prevalence of earnings management can only be justified where there is a known symmetrical distribution for the data in question.

Keywords: Earnings management, earnings benchmarks, distribution of reported earnings

JEL Classification: M41, M43, L14, C89