Predictive power of timely measured attention on stock returns: 21st century evidence from the U.S.

Abstract

In this paper, I examine Google search frequency (Search Volume Index (SVI)) of S&P 1500 -index stocks in relation to raw stock returns in the U.S. between the years 2004 and 2018. I cross-sectionally regress stock returns on abnormal search frequency, also controlling the market capitalization and industry, in order to study price predictive effect. In line with existing literature, my results suggest a significant positive increase in returns in a week after abnormally high search frequency. Accordingly to the price pressure hypothesis of Barber and Odean (2008), this increase is then reversed in the following two weeks. In an attempt to make a contribution to the existing literature, I examine the price pressure effect between different industries. My results are unable to prove significant differences in this effect between industries.