CONSTRUCTION OF ECONOMIC UNCERTAINTY INDICATORS AND CARBON PRICE PREDICTION BASED ON PARTIAL LEAST SQUARES METHOD

This study examines the ability of economic uncertainty measures to predict carbon prices. It introduces a novel economic uncertainty index using the partial least squares (PLS) method, which includes traditional economic uncertainty indicators for comparison. Our research indicates that the economic uncertainty index based on the PLS method (EUPLS) outperforms several existing economic uncertainty indicators. In-sample results and out-of-sample tests reveal that EUPLS possesses superior predictive capability compared with other measures and demonstrates significant predictive ability in multi-period forecasting (1, 3, 6, 9, and 12 months). The research offers a practical approach to enhancing market participants' predictive abilities in carbon markets.