Correlated Forecast Errors
研究投资者在收集公司信息时，如何选择其私人信息误差项之间的相关性，以平衡交易成本与价格可预测性，解释了为何个体盈利预测误差存在相关性。
If investors can choose what information to gather about a firm, will their information-gathering activities overlap, thereby creating correlated errors in their forecasts? To address this question, we present a model in which informed investors choose the correlation between the error terms in their private information. Our analysis implies the chosen correlation will lie between the two extremes of complete independence and total correlation. The advantage of the former, which arises from independent information gathering, is that it reduces the cost of trading on private information in illiquid markets. In contrast, correlated information gathering enhances the ability to predict price movements that result from the future disclosure of privately gathered information. An interior choice of correlation results when the marginal benefit of lower trading costs equates to the marginal benefit of greater price predictability. Our analysis provides one explanation for why individual earnings forecast errors are correlated (O'Brien [1988] and Lys and Sohn [1990]). In addition, our analysis suggests how that correlation arises endogenously. For example, Abarbanell, Lanen, and Verrecchia [1995] show how price and volume behave in response to analysts' forecasts, when the correlation across these forecasts is exogenous. While incorporating our analysis into that discussion would be complex, at a minimum this paper outlines how one might do so. Most models of rational trade characterize an uncertain event as a normally distributed random variable, and information about the event