Comovements between multidimensional investor sentiment and returns on internet financial products

We establish multidimensional investor sentiments of Internet financial products (ISIFP) using a text mining method and data from WeChat subscriptions in China. Our comprehensive ISIFP measure is more affected by negative sentiment information. We investigate eight dimensions of ISIFP and find that seven sentiments (joy, anger, sadness, fear, calmness, disgust, and surprise) increase market risk and decrease expected returns, whereas the goodness sentiment has a leverage effect on returns on Internet financial products (IFPs). The dimensions of ISIFP have different impacts on returns on IFPs at different quantiles. •Establish the multidimensional investor sentiments of Internet financial products.•Comprehensive sentiment is more affected by negative sentiment information.•Seven sentiments increase market risk and decrease expected returns.•There is a leverage effect between multidimensional sentiments and returns.•The dimensions of investor sentiments have different impacts on returns at quantiles.