Disentangling market drivers and macro uncertainty risks in crude oil futures pricing: a multi-scale quantile regression and causal forest approach

In the accelerating global energy transition, crude oil futures price formation has become increasingly complex. This study employs multi-scale decomposition, multi-scale quantile regression, and causal forest models to investigate the volatility mechanism of China's INE crude oil futures from multiple dimensions, including supply and demand, monetary factors, and macro uncertainty. Using a "decomposition-reconstruction" approach, we extract multi-frequency dynamic patterns of crude oil futures prices. Multi-scale quantile regression reveals complex relationships between prices and influencing factors across different time scales, while the causal forest model quantifies the causal effects of economic policy uncertainty, climate policy uncertainty, and geopolitical risks on price volatility. We find that Brent crude oil futures prices have the strongest linkage with China's market in the medium to long term, while the Nanhua Industrial Index significantly impacts short-term fluctuations. Climate policy uncertainty significantly affects the market across all time scales, while economic policy uncertainty has limited impact. Geopolitical risks indirectly influence price volatility via supply-demand relationships. This study provides new perspectives and practical insights for investors, policymakers, and regulators to better manage and respond to the risks of crude oil price volatility in an increasingly complex global market.