We provide evidence that political instability deteriorates economic growth. We establish this result based on panel difference-in-differences strategies and dynamic panel data models using a large sample of 180 countries, a novel geocoded dataset for 2,660 regions, and micro data for about 250,000 households. We exploit coups d'état as a source of exogenous variation in political instability, as they are difficult to anticipate, mirror the political zeitgeist, and reduce measurement error. We use spatial variations and synthetic control methods for identification and find that periods of instability reduce growth by 2-3 percentage points, increase unemployment, and impair health and life satisfaction. The adverse effects are stronger for women than for men.