Unraveling asymmetrical spillover effects originating from China's green finance markets: Insights from asymmetric TVP-VAR and interpretable machine learning.

Abstract

This study combines an asymmetric TVP-VAR model with interpretable machine learning algorithms to confirm the presence of asymmetries in spillover effects within China's green finance market and to identify the macroeconomic drivers behind these effects. The key findings are as follows: First, China's green finance market has become a prominent transmitter of energy risk spillovers, with a significant asymmetry in its external effects-negative return spillovers exceed positive ones. This asymmetry is especially evident during extreme events like the 2014 oil price crash and the COVID-19 pandemic, indicating that investors in this market are more responsive to negative news. Second, using interpretable machine learning models, we identify the macroeconomic factors that significantly impact spillover effects in the green finance market, with Economic Policy Uncertainty and the U.S. Energy Price Index standing out as particularly influential. Third, while the drivers of positive and negative spillovers differ, their directional impact is consistent across both. These insights are crucial for investors aiming to diversify portfolios and for policymakers managing risks in asymmetric market conditions.