Investor sentiment and stock market returns: A comparative analysis of mood, word, and trade

Abstract We examine and compare the return predictability of mood‐, word‐, and trade‐based sentiment measures across 18 international stock markets. Empirical results reveal that the trade‐based measure performs strongly across many settings; the word‐based measure contributes important complementary information, including in cases where the trade‐based measure is less informative; and the mood‐based measure provides useful signals at shorter horizons and in certain markets. Overall, no single measure dominates universally, and our findings highlight the significance of contextual factors such as test types, forecast horizons, and stock markets in determining the predictive efficacy of sentiment measures.