Do popular party leaders increase the vote share of their political party? The individual-level relationship between the popularity of party leaders and reported vote choice has been well established by existing survey research on parliamentary elections. Yet, whether a greater popularity of party leaders actually translates into a higher vote share for political parties has received no systematic attention. I argue that polarizing party leaders have a greater positive effect on their party's vote share compared to universally popular leaders, because of their greater mobilization potential. This paper examines the relationship between leader popularity and changes in party vote share using cross-national survey data and administrative election statistics. The results enhance our understanding of polarization and the personalization of politics by identifying potential micro ecological fallacies in existing research. They further have important implications for leadership selection within political parties.