This study examines the dynamic association between Climate Policy, CO2 Emissions and European stock market returns. While the existing literature offers mixed evidence regarding the pricing of climate risk, we investigate whether these associations are regime-dependent and differ across levels of climate exposure.
Using an annual panel of 28 European countries spanning 2007–2024, we estimate dynamic conditional associations using a local projections framework (Jordà, 2005, 2023) to assess the dynamic responses of stock returns to changes in climate-related variables. We then combine the threshold model of Seo and Shin (2016) with the local projection framework by estimating the benchmark model within each regime, to capture endogenous regime shifts in climate risk exposure. Finally, we conduct robustness checks on the baseline specification adding alternative macroeconomic controls and sample restrictions.
The threshold estimation provides strong evidence of regime dependence. In particular, the pricing of CO2 emissions in stock returns depends critically on the level of climate policy stringency. Emissions are positively associated with stock returns in low-policy environments but are negatively associated in high-policy regimes. In contrast, changes in climate policy are generally associated with positive stock market responses.
The reliance on aggregate market indices limits our ability to assess how firm-level characteristics shape exposure to transition risk. A more granular analysis based on individual stock data could yield deeper insights, particularly with respect to portfolio diversification.
The findings highlight that the pricing of climate risk is shaped by policy credibility. Investors and policymakers should account for the interaction between emission intensity and regulatory stringency when assessing financial stability and transition risks.
This study provides novel evidence of regime-dependent and dynamic associations between climate risks and European stock markets by integrating local projections with an endogenous dynamic panel threshold framework.
