Analyst Stickiness and Stock Return Predictability
研究发现分析师在更新预测时存在粘性，倾向于向先前预测靠拢，这种粘性在认知噪声较高的分析师中更明显，且粘性分析师的共识修正比传统共识修正对股票收益有更强的预测能力。
Abstract This study estimates analyst-level stickiness in forecast updating and investigates its underlying determinants. Consistent with recent experimental findings on belief updating under cognitive noise, analysts often compress their forecasts toward an intermediate default, such as prior forecasts, when uncertain about forecast precision, leading to forecast stickiness. This tendency is more evident among analysts with characteristics associated with higher cognitive noise, including lower forecast accuracy, limited experience, and complex portfolio coverage, and during periods of heightened macroeconomic uncertainty. A model incorporating sticky updating behavior shows that the consensus revision by sticky analysts exhibits stronger return predictability than the traditional consensus revision by all analysts, with this predictability increasing with the proportion of sticky analysts covering a stock. Empirical evidence supports these predictions. Additionally, the return predictability of sticky revisions is especially pronounced when forecast difficulty is elevated. Analyst-level stickiness provides more information about the cross-section of stock returns than firm-level stickiness.