Investor sentiment and the variance risk premium are well-established learning-based predictors of aggregate stock market returns. This study investigates whether the return predictability of investor sentiment and the variance risk premium is impacted by the quality of political signals. Our analysis shows that low-quality political signals substantially weaken return predictability via a prolonged mispricing correction associated with lower market participation. The explanatory power of predictive regression models is significantly improved when a proxy for the quality of political signals is included. Overall, our robust findings provide evidence that low-quality political signals have a negative impact on the functioning of financial markets.