Biodiversity risk and agricultural futures markets

Abstract

This study constructs a novel Biodiversity Risk (BR) index based on textual analysis and evaluates its performance in forecasting volatility in China's major agricultural futures markets. Using an autoregressive (AR) model framework, we conduct comparative analyses of the BR index's predictive ability through univariate models, various combination forecasting methods, and dimensionality reduction techniques (PCA, sPCA, and PLS). Empirical results indicate that the BR index possesses a certain degree of forecasting power for volatility in agricultural futures markets, with PLS-based models demonstrating relatively superior predictive performance across multiple markets. This conclusion is supported by multiple robustness tests, high-volatility periods, and medium- to long-term forecasts. The study emphasizes that incorporating biodiversity risk indicators can help improve risk management in agricultural futures and promote sustainable development.