The aim of this paper is to investigate the sensitivity of stock markets to election uncertainty and election shock. The analysis employs both fixed effect modelling approach and event study methodology, and utilizes a unique dataset of polling results measuring political preferences over 91 elections in EU countries. We show that election uncertainty induced by changes in political support significantly affects the volatility of stock markets in the pre-election period. Stock volatility also increases in post-election periods. We find that the difference between the outcome of the election and the expected one contributes to the magnitude of election shock, which influences stock markets. These suggest that the accuracy of pre-election polls can be used by market participants and academics as a proxy for market expectations. Our findings have also important implications for optimal investing strategies around elections and are of interest to fiscal policy makers and regulators of pollsters.