Spillover effects of air pollution-induced sentiment across industries: Evidence from a LASSO machine learning technique

While existing research investigates the impact of investor sentiment on stock markets, it often underexplores the potential for sentiment closely related to specific industries to spill over into other, less related industries. Utilizing a Least Absolute Selection and Shrinkage Operator (LASSO) machine learning technique for robust variable selection and statistical inference, we identify significant cross-industry spillover effects of air pollution-induced investor sentiment, underscoring its important role in the whole cross-section of stocks. The spillover patterns remain consistent before and after the COVID-19 outbreak for most industries. We demonstrate that air pollution induces sentiment spillover in industries with minimal direct exposure, driven by investors’ environmental concerns and their subsequent investment decisions. Incorporating these spillover effects into portfolio strategies enhance risk reduction and hedging effectiveness. Our findings provide new insights into how environmental concerns influence market dynamics and stock premiums, offering implications for portfolio management and risk mitigation.