¹ÏÌïºÚÁÏ

We characterize the factors that determine who becomes an inventor in America by using de-identied data on 1.2 million inventors from patent records linked to tax records. We establish three sets of results. First, children from high-income (top 1%) families are ten times as likely to become inventors as those from below-median income families. There are similarly large gaps by race and gender. Differences in innate ability, as measured by test scores in early childhood, explain relatively little of these gaps. Second, exposure to innovation during childhood has signicant causal effects on children's propensities to become inventors. Growing up in a neighborhood or family with a high innovation rate in a specic technology class leads to a higher probability of patenting in exactly the same technology class. These exposure effects are gender-specfic: girls are more likely to become inventors in a particular technology class if they grow up in an area with more female inventors in that technology class. Third, the financial returns to inventions are extremely skewed and highly correlated with their scientic impact, as measured by citations. Consistent with the importance of exposure effects and contrary to standard models of career selection, women and disadvantaged youth are as under-represented among high-impact inventors as they are among inventors as a whole. We develop a simple model of inventors' careers that matches these empirical results. The model implies that increasing exposure to innovation in childhood may have larger impacts on innovation than increasing the financial incentives to innovate, for instance by reducing tax rates. In particular, there are many lost Einsteins" - individuals who would have had highly impactful inventions had they been exposed to innovation.