ABSTRACT
Economic policy uncertainty (EPU) is an important driver of the correlation in the oil–stock nexus. However, whether the effect of EPU on oil–stock correlations across different market conditions is heterogeneous remains unclear. To fill this gap, we combine a dynamic conditional correlation with the mixed data sampling (DCC-MIDAS) model and the Markov regime-switching model to explore the market-state-dependent effects of EPU on oil–stock correlations under different regimes. Empirical results indicate that the impacts of EPU on oil–stock correlations are regime-dependent both at the aggregate and industry levels, with stronger effects in high-correlation regimes, and these effects are more significant in times of economic turmoil. Moreover, the impact of EPU on oil–stock correlations is larger during the COVID-19 pandemic than it was during the Global Financial Crisis. These findings highlight the need to consider the nonlinear impact of EPU under different market conditions.