On the same boat: Three essays on the causes of pre-WWI Italian international migration

This thesis analyzes the causes of Italian mass international migration before the outbreak of WWI. While each chapter is a stand-alone paper that explores different motives and dynamics, there is a common quantitative core based on Ellis Island administrative records (1892-1912) and official Italian out-migration statistics (1876-1912). Chapter I is centered on the reconstruction of US-bound out-migration flows from Italian municipalities between 1892 and 1912. The main findings can be summarized as follows. First, international migration followed a process of path-dependency and progressive spatial diffusion at the local level. Second, the further away from a port a given municipality of origin was, the more passengers tended to travel together therefore creating ex-ante networks. Third, cyclical migration was a sizable and widespread phenomenon. Chapter II - joint with Prof Rowena Gray and Prof Gaia Narciso - investigates the local migration response to international agricultural commodity price shocks during the first globalization era. The paper revolves around the idea that price shocks affect different areas according to their agricultural production structure. We conduct a Fixed-Effects analysis at the province level and find that higher agricultural commodity prices released liquidity constraints thus promoting out-migration. On the other hand, higher volatility translated itself into higher incentives to out-migrate. Results are statistically significant across a battery of robustness checks. Chapter III - joint with Prof Yannay Spitzer and Ariell Zimran - studies the local migration response to the 1908 Messina and Reggio Calabria earthquake, one of the deadliest and most disruptive natural disasters in European history. The paper is based on a Difference-in-Difference approach coupled with a series of event studies that allow to test the pre-trends hypothesis. We find that the quake only produced a short-lived decline of out-migration rates from affected municipalities - as opposed to non-affected ones - in Calabria. We test two mechanisms: migration networks and damage accumulation. In the former case, we find that there is no statistically significant effect of networks; in the latter, we show how migration decreased more in municipality which had been affected by another earthquake in the recent past