Search
Program Calendar
Browse By Day
Search Tips
Virtual Exhibit Hall
Personal Schedule
Sign In
We develop and test an empirical model of discretionary accruals that can con-
trol for the simultaneous correlation of multiple determinants with accruals. By
estimating logistic regressions of earnings management proxies on conventional
determinants of earnings manipulation, we ascertain the propensity to manage
earnings for each firm-year in our sample. In order to estimate the discretionary
component of accruals, firms are matched according to year, industry and its
propensity score. Comparisons with similar models in tests of statistical spec-
ification and power indicate that the propensity score matched model further
mitigates the probability of producing false positives in earnings management
hypothesis testing when the null hypothesis is no earnings management.