What makes stocks sensitive to investor sentiment: An analysis based on Google Trends

We capture Google’s vast search volume through Google Trends to generate a weekly investor sentiment index (2018–2022) using the most popular keywords (extracted from Google Search) from a keywords collection of 92,000+ words found in business, finance, and common language dictionaries. The results show that Google Trends is an efficient measure of investor sentiment as reflected in relative trading volume. To check what makes stocks sensitive to investor sentiment, 500 randomly selected US firms from various industries are categorised by firm characteristics. We generate two sub-portfolios: large, old, profitable, and dividend-yielding firms versus small, young, unprofitable, and non-dividend-yielding firms—and find the relative trading volume of the latter to be more sensitive to investor sentiment. Our results remain robust when control and autoregressive variables are introduced, in addition to when an alternative measure of sentiment is used, thereby confirming our primary findings.