The Handbook of News Analytics in Finance

7.6 ASSET PRICING: THEORY AND EVIDENCE

In the appendix of the version of this chapter published in The Review of Financial Studies, we develop a model of price formation for a market in which attention-driven noise traders tend to buy stocks that catch their attention. Our theoretical model has the testable prediction that the underperformance of the stocks bought relative to stocks sold by these noise traders will be greatest following periods of high attention. The intuition underlying this result is straightforward; noise traders create uniform buying pressure in attention-grabbing stocks, and market-makers respond to this buying pressure by increasing prices, leading to lower subsequent returns—the more intense the attention-driven buying, the lower the subsequent returns. In this section, we test this return prediction.

There are two significant challenges in testing our model. Our first challenge is that the model does not specify the period of time over which attention-driven buying affects returns. Our evidence that investors do buy stocks that catch their attention is based upon 1-day sorts. It is likely, though, that investors' attention often spans more than a single day. How the effects on returns of attention-driven buying persist is likely to vary with the intensity and duration of the attention-grabbing event. In the following analysis, we form portfolios on the basis of daily attention sorts and evaluate performance over a 1-month horizon. We obtain similar results ...