This study uses a historical setting in which expected litigation costs were low (i.e., Australia, from 1993 to 1996) to investigate whether companies with good news were more likely to preempt annual earnings than their counterparts with bad news. Empirical tests compare the probability of preemption conditional on having good news with the probability of preemption conditional on having bad news. The models control for other potential determinants of disclosure policy that have been documented in the literature. The results do not support the research hypothesis that companies with good news were more likely to preempt annual earnings than companies with bad news. This finding suggests that there may be other factors driving disclosure of bad news, in addition to those acknowledged in the extant literature. The evidence also indicates that in Australia during the investigation period, the probability of preemption was positively associated with firm size and analyst following and differed as a function of industry membership.