Does the change of accounting regulation on employee share options give rise to greater scope for earnings management?

Purpose – The purpose of this paper is to address the concern about the impact of accounting regulatory change pertaining to employee share options (ESOs) on earnings management. Following Australia's adoption of International Financial Reporting Standards (IFRS) in 2005, companies are required to recognise the fair value of ESOs as expenses. Due to inherent imprecision in the estimate of ESO's fair value, the regulatory change from disclosure to recognition was widely claimed to potentially give rise to an alternative mechanism to manage earnings. This study provides empirical evidence on whether the regulatory change leads to earnings management problems. Design/methodology/approach – This study uses the regulatory change in accounting for ESOs to provide a direct test of earnings management between disclosed versus recognised regimes for the same sample of firms. The sample consists of Australian firms from S&P/ASX300 for the period from 2003 to 2006. Findings – The results show that, although the accounting regulatory change from disclosure to recognition may provide an alternative earnings management vehicle, there is no evidence of this occurring. There could be several reasons for this finding. First, the statistical tests lack power. Second, there are stricter audit tests on recognised amounts than on disclosed amounts. Third, given the concern of excessive pay and the close scrutiny of compensation, managers may have already understated ESO values in the disclosure regime. Finally, managers have limited time and resources and the effort involved in the adoption of IFRS in 2005 could have restricted the time available to manage earnings via the ESO reporting channel. Originality/value – This study adds to the limited research on whether a change in accounting regulation for employee share options from disclosure to recognition gives rise to greater scope for earnings management. One reason for the lack of empirical evidence in the research is due to the problem of designing a test. Bernard and Schipper suggest that within‐firm studies have limitations for comparing the effects of recognition versus disclosure when the change is driven by an estimate becoming more reliable. A cross‐sectional study is also problematic due to self‐selection bias if firms can choose between disclosure versus recognition. This study circumvents potential design problems raised by Bernard and Schipper by setting a test using regulatory change which allows the test to be compared directly using the same company.