Cross-Sectional Relations among Accruals, Cash Flows and Profitability: A Bayesian Model for Analyzing Earnings Management


  • 2021
English It is an important stylized fact that cross-sectional accruals are positively related to earnings and negatively related to operating cash flows. In this paper, we reconsider the question of why these associations occur and how to deal with them in earnings management studies. We propose a simple Bayesian model in which unobservable true profitability is related to cash flows by normal accruals and to reported earnings by discretionary accruals. The overlay of uncertainty concerning true performance and normal and discretionary accruals gives rise to the observed cross-sectional relations, even if earnings are not managed and true performance does not affect accruals. In our model, earnings management appears in the Bayesian estimates of both discretionary and normal accruals. The most efficient way to detect earnings management depends on the sampling of suspect firms. For random sampling, the test should be based on estimated discretionary accruals; for sampling with respect to reported earnings, estimated normal accruals are more efficient; and for sampling with respect to the Bayesian estimate of true profitability, total accruals are most appropriate. We propose graphic diagnostics and a simple regression approach to identify the appropriate performance adjustment setting.
Faculté des sciences économiques et sociales et du management
Département des sciences du Management
  • English
Other electronic version

Publisher's version

License undefined
Persistent URL

Document views: 1 File downloads:
  • 2021_Wallmeier_Cross-Sectional.pdf: 5