Relationship between Differential Amounts of Prior Information and Security Return Variability
发现，在年报发布时，分析师预测数量越多或公司规模越大，股票收益波动越小，说明这两者衡量了先验信息的不同维度，且对小公司或分析师覆盖少的公司，检验统计量更容易显著。
This study investigates the relation between stock return variability at annual earnings announcements and the amount of prior information available about a firm, which is measured by both the number of security analysts' annual earnings forecasts and firm size (market value of common equity). We find that each of these information measures is inversely related to return variability at annual earnings announcements, after controlling for differences in the other. This result suggests that number of security analysts' earnings forecasts and firm size measure dimensions of the amount of prior information available about a firm that are not subsumed by one another. The study also provides evidence, using empirically generated distributions, that test statistics derived from squared, standardized market model prediction errors have larger means and variances for firms with few analysts' forecasts or for small firms than for firms with many analysts' forecasts or for large firms. This result implies that the probability of obtaining large values of those test statistics is greater for firms with few analysts' forecasts or for small firms; therefore, the probability of concluding there is increased return variability at annual earnings announcements is greater for those firms. Consistent with Marais [1984],