Uncertainty, Disagreement and Sentiment effect_data

This study proposes a sentiment asset pricing model featuring parameter uncertainty and disagreement within a framework of Bayesian learning. In the model, the stock-level sentiment is constructed as the cross-sectional wealth-weighted average of investors’ sentiments, and disagreement is constructed as their variance. We document that stock price is positively correlated with parameter uncertainty and disagreement when sentiment is optimistic but negatively when sentiment is pessimistic. Accordingly, parameter uncertainty and disagreement amplify sentiment-induced mispricing. Moreover, empirical evidence aligns with the model predictions, and further analysis confirms that the amplifying effects of parameter uncertainty and disagreement cannot be explained by existing explanations.

本研究提出了一个嵌入贝叶斯学习（Bayesian learning）框架、兼具参数不确定性与投资者意见分歧的情绪资产定价模型。在该模型中，个股层面情绪被构建为投资者情绪的截面财富加权平均值，而意见分歧则以投资者情绪的方差进行测度。本文研究发现，当投资者情绪乐观时，个股价格与参数不确定性、意见分歧呈正相关关系；而当情绪悲观时，二者则呈负相关关系。据此，参数不确定性与意见分歧会放大情绪引致的定价偏差。此外，实证证据与该模型的预测结果相一致，进一步分析证实，参数不确定性与意见分歧的放大效应无法通过现有解释体系予以说明。