Commodity Futures Report Text Sentiment and Returns

ABSTRACT This paper investigates how textual sentiment in analyst research reports influences commodity futures returns in China. Using the Natural Language Processing and Information Retrieval platform to analyze a comprehensive sample of reports from 2016 to 2021, we construct sentiment indices for 37 commodity futures across agriculture, metals, chemicals, and energy sectors. Our results show that text sentiment significantly explains both contemporaneous and future returns, with negative sentiment having a stronger impact. Increased optimism is associated with higher contemporaneous returns through investor sentiment and attention channels, especially for future contracts with low open interest growth, low volatility, a contango market (futures premium), and high basis‐momentum. We document heterogeneous results for subsequent returns and the three main risk premium factors. The findings remain robust after controlling for established pricing factors and macroeconomic variables. This study highlights the informational value of analyst sentiment, filling a gap in text sentiment research within the commodity futures market and offering insights for investors and policymakers.