We formulate and test the hypothesis that nonearnings disclosures of small, but not large, firms generally are good news. Nonearnings disclosures are defined as disclosures by managers and outsiders about news other than earnings (e.g., stock splits, takeovers, new orders). Good news is defined as a positive stock pi-ice reaction at the time of the information disclosure. Our hypothesis is motivated by two lines of prior research. First, managers have incentives to disclose their private information voluntarily when they expect the effects of the information on firm value to exceed the disclosure costs (Verrecchia 19831. Second, the firm-size differential information hypothesis, advanced by Atiase (1980, 1985) and the corroborating empirical evidence of Atiase (1985, 1987), Freeman (1987), and Bhushan (1989) suggest that incentives for information production and dissemination by outsiders are an increasing function of firm size. Thus, assuming that nonearnings disclosures concerning small firms are initiated primarily by managers, whereas those of large firms are not, small (but not large) firms' nonearnings disclosures are more likely to be good rather than bad news. Using firm-specific nonearnings disclosures, identified from the Dow Jones News Retrieval Service data base over the 1982 to 1987 period, we show that small firms' nonearnings disclosures, on average, are associated with significant stock price increases, whereas large firms' nonearnings disclosures, on average, are valuation-neutral. Given these results and the evidence that nonearnings disclosures are often made around the time of earnings announcements (Hoskin et al.1986; Thompson et al. 1987), we reexamine the puzzling result of Chari et al. (1988) that on-time earnings announcements of small, but not large, firms are associated with positive abnormal returns, unconditional upon the nature of the earnings news. We hypothesize that this phenomenon is attributable to nonearnings disclosures of good news around the time of small firms' earnings announcements. We show that small and large firms' pure on-time earnings announcements are not associated with positive abnormal returns, and that small (but not large) firms' contaminated on-time earnings announcements are associated with positive abnormal returns. We conclude that the Chari et al. (1988) results do not pertain to small firms' on-time earnings announcements per se, but to those that are accompanied by nonearnings news
