Investor Sentiment Indicators and the Prediction of European Equity Index Returns: A Machine Learning Approach

This paper investigates the predictive power of various sentiment indicators on European equity index returns, focusing on the EUROSTOXX600 and EUROSTOXX50 over the period 2010–2022. Using a set of sentiment indicators ranging from options-based measures (Call/Put ratios, Skew), economic and political uncertainty (EPU, RAX), and sector-based sentiment proxies (EASS indices), we estimate a suite of models from classical linear regression to regularized models (Ridge, Lasso) and machine learning techniques (Random Forest, Gradient Boosting). The aim is to assess both the statistical significance and predictive importance of these sentiment measures. Our findings show that sentiment indicators carry significant explanatory power for European equity returns, with certain indicators consistently emerging as relevant across methodologies. The results have implications for asset managers and policymakers interested in incorporating behavioral factors into asset pricing models.