Exploiting a unique setting in United States military history, I study the origins of and persistence in the spatial distribution of the US population. I build a new dataset of the 19th century Army's frontier forts to document their place-based effects. Forts predict initial increases in population and population density, indicative of their role as man-made factors in explaining the origins of local population patterns. Increased population and density persist, on average, over a century after fort abandonment, indicative of path dependence at frontier fort locations. Results using new causal random forest estimation provide consistent evidence of frontier forts' treatment effects both initially and in the long run. Persistence at fort locations is not driven by variation in fort size or length of time in operation, providing evidence that initial, temporary place-based policies may have long run consequences. Separately, demographics of forts locations reflect those of the US Army itself, but only in the short run, with increased shares of men and foreign-born. Using the spread of the transcontinental railroad, I show that early rail access was critical for continued growth at fort locations and that forts are necessary, but not sufficient, in determining spatial patterns.
The Road to the Urban Interstates: A Case Study from Detroit (revise and resubmit, 2nd round, The Journal of Economic History)
I use the city of Detroit to examine the political economy behind site selection for urban segments of the Interstate Highway System and to measure effects of construction on neighborhoods. Using variation at the census-tract level, I show that interstates were routed through neighborhoods with low property value to minimize land acquisition costs and future losses to the city's tax base. Following construction, interstates led to short-run declines in property values, population density and the percentage of black residents. In the long run, property values continued to decline in tracts with an interstate, and tracts closest to construction remained less densely populated and had lower shares of black residents relative to those further away.
Work in Progress
Demand uncertainties create major obstacles for financing technological innovation, as well as creativity in the arts. This paper uses detailed book-level data on Romantic Period English literature to investigate crowdfunding as a mechanism to finance innovation in the presence of significant demand uncertainties. A simple model yields conditions under which authors choose alternative financing, and specifically, crowdfunding. We show that new authors, female authors, and authors in new genres face substantially greater demand uncertainty than established authors, men, and authors working in established genres. Detailed book-level contract data reveal that entrants, women, and authors in new genres are more likely to crowdfund their works. We find that crowdfunded works have lower payoffs on average but are substantially more likely to become a major hit. Exploring variation across genres, we show that crowdfunded novels for women, and by women, were more likely to be published in multiple editions, suggesting that realized demand exceeded expected demand in this emerging genre. Crowdfunded women’s novels are also more likely to be translated and have an increased probability of long-run success.
Forts and the Origins of Industrial Composition