A central question in economic history is whether the nature of production of certain crops, such as cotton, necessitates the coercion of labor. Using a unique data source, newly digitized samples of Egypt’s individual-level population censuses of 1848 and 1868, I document that the cotton boom in 1861-1865 increased the demand for imported slave labor in cotton-favorable districts, but not in non-cotton-favorable districts (that produced only cereals and beans). Demand for slave labor peaked among village headmen, who were medium landholders, but was negligible among both smaller landholders and large estates; the latter had preferential access to local agricultural labor via state coercion tools. Overall, the findings suggest that the crop-slavery association is not necessary but conditional on the magnitude of the price shock, the relative supplies of local and (imported) slave labor, and landholders’ access to coercive tools of local and foreign labor.