Lynchings, Labour and Cotton in the US South

I show that cotton price shocks predict lynchings of African Americans in the US South from 1882 to 1930. Specifically, a standard deviation decrease in the cotton price leads to a 0.095 to 0.16 standard deviation increase in lynchings, within a cotton-producing county. Lynchings also predict more black out-migration from 1920 to 1930. Using a simple model, I show that this is consistent with lynchings having labour market effects that benefitted whites: lynchings cause blacks to migrate away, lowering labour supply and increasing wages for white labourers. I run complier tests to show the mediating effects of railroads, the black-white farm worker ratio, and slavery. I then turn to the long-term effects of lynchings. I show that Mississippi counties with more violence during a 1964 Civil Rights campaign also had more lynchings during the earlier period. Finally, present-day outcomes show that lynchings predict higher black-white worker, family, and household income gaps; a standard deviation rise in past lynchings predicts a 0.08 to 0.15 standard deviation increase in black-white income gaps. These present-day results are robust to historical controls, to the use of California lynchings as a falsification, to the use of white-on-white lynchings as a placebo, and to the use of 1879 cotton acreage as an instrument; they also survive tests using Altonji, Elder, and Taber (2005) statistics.