We examine how large and localized resource discoveries affect long-run population growth in the United States, and examine how these shocks interact with pre-existing geographic properties of the discovery site. Using a dynamic event study analysis and developing novel, geographically delineated measures of both amenity value and geographic isolation, we find that resource discoveries cause population to grow both in the short and long-run (e.g., fifty years). However, this effect is largely driven by discoveries in unfavorable locations that might struggle to grow in the absence of a resource discovery. More generally, this paper highlights the importance of considering heterogeneous effects of resource shocks and yields insights into the observed spatial distribution of people in the United States.
