Dynamic Attention Behavior Under Return Predictability
研究了投资者在回报可预测性下如何动态分配注意力以最大化终身效用，发现最优注意力与预测因子呈U型关系，并得到实证支持。
We investigate the dynamic problem of how much attention an investor should pay to news in order to learn about stock-return predictability and maximize expected lifetime utility. We show that the optimal amount of attention is U-shaped in the return predictor, increasing with both uncertainty and the magnitude of the predictive coefficient and decreasing with stock-return volatility. The optimal risky asset position exhibits a negative hedging demand that is hump shaped in the return predictor. Its magnitude is larger when uncertainty increases but smaller when stock-return volatility increases. We test and find empirical support for these theoretical predictions. This paper was accepted by Gustavo Manso, finance.