The Impact of Forced Migration on Modern Cities : Evidence from 1930s Crop Failures

We show that a surprisingly large portion of current city-level variation in unionization was set in place by exogenous events during the 1930s. Further, this exogenous unionization has real impacts on city-level economic outcomes through the present day. We first show that a primary factor behind city-level differences in unionization rates (within the same industry and occupation) is random - a result of substantially different rainfall levels during the Dust Bowl. We find that individuals in drought-ridden areas were significantly more likely to migrate to nearby cities. Workers in these cities - facing an influx of rural migrants - then became far more inclined to unionize than those facing less competition for their jobs. Using rainfall in surrounding counties to generate exogenous variation in city migration inflows, we show that random differences in 1930's drought conditions predict migration patterns, and variation in union formation rates that persist over 80 years later. These unionization shocks in turn predict modern-day city-level economic outcomes such as education levels, establishment growth, salary growth, and the presence of high-value industries. Moreover, the impacts are seen solely in those industries impacted by less-skilled farmer migration, and are absent in higher-skilled industries (e.g., medical fields) and fields that did not exist in the 1930s (e.g., telecommunications)