SOME EVIDENCE ON THE NEWS CONTENT OF PRELIMINARY EARNINGS ESTIMATES

This article provides evidence on the news content of managements' pre liminary earnings estimates, which we define as projections of earning s conveyed in expectational language after the end of the reporting pe riod but before the release of final earnings numbers. We examine stoc k price changes to assess whether a preponderance of these disclosures are interpreted as ''good news'' by investors, and the extent to whic h good news releases are disclosed earlier than bad news. Associated w ith preliminary earnings estimates are disclosure and timing issues. W hile previous theoretical work suggests managers have incentives to su ppress or delay disclosure of adverse information (Verrecchia 1983; Dy e 1985), studies examining the disclosure issue using management forec asts produced inconsistent results, and results on the timing of corpo rate earnings announcements are ambiguous. Although there is evidence on the information content of preliminary earnings estimates (Foster 1 973), no study has used these data in investigating the relation betwe en corporate disclosure and news content. Preliminary estimates are im portant because they embody aspects of disclosure choice similar to ot her voluntary disclosures such as forecasts, and the time lags between preliminary estimates and earnings releases are short, thereby assuri ng a strong timing aspect to their release. We document significant ne gative mean abnormal returns associated with the disclosure of prelimi nary estimates. The median is negative, but not significant at convent ional levels. Tests on the timing issue indicate an ambiguous relation between disclosure timing and news content. Preliminary estimates of quarters 1-3 earnings are more likely to be bad news compared to estim ates of quarter 4 and annual earnings, but within quarters there is no strong relation between news content and disclosure timing.