We analyze the long-run evolution of populism and explore the role of globalization in shaping such evolution. We use an imbalanced panel of 628 national elections in 55 countries over 60 years. A rst novelty is our reliance on both standard (e.g., the "volume margin", or vote share of populist parties) and new (e.g., the "mean margin", a continuous vote-weighted average of populism scores of all parties) measures of the extent of populism. We show that levels of populism in the world have strongly fluctuated since the 1960s, peaking after each major economic crisis and reaching an all-time high – especially for right-wing populism in Europe – after the great recession of 2007-10. The second novelty is that when we investigate the "global" determinants of populism, we look at trade and immigration jointly and consider their size as well as their skill-structure. Using OLS, PPML and IV regressions, our results consistently suggest that populism responds to globalization shocks in a way which is closely linked to the skill structure of these shocks. Imports of low-skill labor intensive goods increase both total and right-wing populism at the volume and mean margins, and more so in times of de-industrialization and of internet expansion. Low-skill immigration, on the other hand, tends to induce a transfer of votes from left-wing to right-wing populist parties, apparently without aecting the total. Finally, imports of high-skill labor intensive goods, as well as high-skill immigration, tend to reduce the volume of populism.