This paper examines the influence of local (state-specific) and global Economic Policy Uncertainty (EPU) on the volatility of US state-level equity returns. We employ a GARCH-MIDAS approach that incorporates multiple EPU indices as low-frequency predictors of daily stock return volatility. To address the challenge of selecting the most relevant EPU indices, we utilize an Elastic Net (EN) shrinkage method to combine forecasts from different models. Our results reveal that the combined model, which leverages information from both local and global EPU indices, generally outperforms single specifications. Further, a cluster analysis based on the volatility forecasts uncovers distinct geographical patterns, suggesting that state-level volatility is influenced by both state-specific and nationwide policy uncertainties. These findings highlight the importance of considering both local and global economic policy uncertainty in understanding and predicting the volatility dynamics at the regional level.