| This paper proposes a measure of earnings management flexibility and examines its impact on firms' reporting strategies. Most prior earnings management studies assume that earnings management is free and managers have unlimited ability to manage earnings. In practice, though, with the limitations imposed by GAAP and the monitoring by auditors and the SEC, earnings management with accruals beyond the allowable set can be prohibitively costly. I propose that the manager's discretion in a quarter depends on the limits of the allowable set of accruals, the level of prior discretionary accruals, and the reversal rates of these accruals. Specifically, analysis illustrates that prior discretionary accruals have three effects—reversal, constraint, and benchmark effects—on current period's flexibility. While the magnitude of prior discretionary accruals determines the degree of impact on current quarter flexibility, the reversal rate of accruals determines the number of prior quarters whose accruals have an impact on current quarter's flexibility. Based on this analysis, I construct a flexibility measure that allows me to test the impact of flexibility empirically. The results show that firms with low earnings management flexibility (proxied by both my measure and Barton and Simko's (2002) measure) are more likely to miss analysts' forecasts. Results based on Kasznik's (1999) measure are not significant. Moreover, I show that my flexibility construct provides incremental information to that captured by the flexibility measures proposed in Barton and Simko (2002). I also show that after adjusting the flexibility measures by the industry mean, results using my flexibility measures continue to hold while those using Barton and Simko's flexibility measure are no longer significant. |
