Predicting Abnormal Stock Market Returns and Trading Volumes Using Google Search

Abstract: We test if investor attention, measured by online search volume in Google, can predict abnormal stock market returns and trading volumes. Specifically, we use online searches for ticker symbols and company names as a proxy for investor attention and hypothesize that an increase in investor attention leads to higher trading volumes and returns in the subsequent weeks. We examine all stocks in the S&P 500 and S&P Europe 350 indices during the period 2005-2014 and control our results for market risk, size, value and momentum factors in accordance with the Carhart (1997) four-factor model. Over a ten-year period, we find that search volume in Google can predict abnormal trading volumes of up to 20% for the following week and yearly abnormal returns of up to 12%. However, when adding trading costs and taxes, the observed abnormal returns will diminish and probably be eliminated in most cases. In addition, we find that the abnormal trading volumes and abnormal returns are considerably higher during the global financial crisis and almost non-existent in the post-crisis period, indicating that the observed abnormal returns are only apparent during exceptional market conditions. Our study contributes to the existing literature on investor attention by reinforcing the relevance of online searches and by providing new insights relating to bull and bear markets, geographical differences and alternative online search measures, that give reason to further scrutinize previous and forthcoming studies relating to investor attention and online search.