We propose new ways to measure populism, using the Manifesto Project Database (1960-2019) as main source of data. We characterize the evolution of populism over 60 years and show empirically that it is significantly impacted by the skill-content of globalization. Specifically, imports of goods which are intensive in low-skill labor generate more right-wing populism, and low-skill immigration shifts the distribution of votes to the right, with more votes for right-wing populist parties and less for left-wing populist parties. In contrast, imports of high-skill labor intensive goods, as well as high-skill immigration flows, tend to reduce the volume of populism.