What makes an industrial policy successful? This paper finds that the effect of an industrial policy changes tremendously with the implementing bureaucrat. We study Korean bureaucrats who promote exports on appointments to 87 countries between 1965, when Korea was one of the world’s poorest countries, and 2001. We exploit the rotation of bureaucrats between countries to show that individual bureaucrats matter greatly in boosting exports. Moving from a bureaucrat at the 20th percentile to the median is associated with a 40% increase in exports. This effect is comparable to that of opening an office, implying that this industrial policy has no effect under a 20th percentile bureaucrat. We exploit differential import demand growth to study a mechanism via which better bureaucrats increase exports - transmitting information about market conditions. Under better bureaucrats Korean exports increase more with a product’s import demand. Finally, we investigate whether experience can bridge the gaps between bureaucrats. We isolate quasi-random variation in experience exploiting a product’s import demand growth during the bureaucrat’s first appointment. In subsequent appointments exports increase in products with greater bureaucrat experience. This highlights learning-by-doing as a channel to build bureaucratic capacity. However, the differences between bureaucrats are larger than the effect of experience, suggesting selecting good bureaucrats may be more important than training them.