This paper studies the impact of railroads on town-level growth in Sweden over 150 years. Our analysis builds on the fact that railroads historically were extended quasi-randomly across towns. Towns that gained access to a rail connection grew larger relative to other towns, with large negative spillovers on unconnected nearby towns. Over the 20th century, we find little adjustment to the initial shock in town populations, despite a sharp reversal in relative connectivity. Evidence on historical investments and present-day factors is consistent with this temporary shock giving rise to path dependence in the location of economic activity.
展开▼
