Nationalism is widely viewed as a force for interstate violence, but does it really have an important effect on state aggression that cannot be explained by strategic concerns? I provide strong evidence that it does using regression discontinuity analysis. I take advantage of the fact that many countries experience a surge of nationalism when they go to the World Cup, and the World Cup qualification process from 1958-1998 produced a large number of countries that barely qualified or barely missed. I show that these countries are well-balanced across a wide range of factors, including past levels of aggression. However, the qualifiers experienced a significant spike in aggression during the World Cup year. I also replicate the analysis using the FIFA regional soccer championships and find similar results. In both cases, the estimated treatment effect is larger for authoritarian states than democracies, suggesting that democratic norms may help constrain nationalistic aggression.