Economic Consequences of Coercive Institutions: Evidence from the US Convict Labor System
Abstract I explore the consequences of convict labor in modern history. I argue that prison-produced goods competed with low-skill labor (particularly female) intensive products, thus pushing business owners to adopt new technologies to automate low-skilled jobs. Using newly-collected panel data on U.S. prisons and convict labor camps from 1886 to 1940, I show that the wages of free laborers were lower in the counties that forced prisoners to work. At the same time, those counties had higher rates of technology adoption: a greater number of registered patents and higher capital-labor ratio. To alleviate endogeneity concerns, I use state-level variation in the introduction of convict labor laws interacted with historical state-level prison capacity to construct instrument for the prevalence of convict labor. I find that one standard deviation in exposure to convict labor decreases wages in manufacturing by 38%, and increase capital-labor ratio and number of registered patents by 15.1% and 22.1% of its standard deviation respectively. I document that reallocation of welfare from wage-earners to capital-owners caused by convict labor had a long-lasting effect on the inequality of opportunities. Intergenerational mobility of the bottom income quintile gets worse, while outcomes of the richest improved. The result also sheds light on the evaluation of possible impact of contemporary prison industry on the U.S. economy.