Pré-publication, Document de travail: Labor scarcity is the main hypothesized determinant of labor coercion (Domar, 1970), however, its effects are theoretically ambiguous and remain empirically untested. This paper provides the first causal estimate of the effect of labor scarcity on labor coercion. I obtain quasi-exogenous variation in labor scarcity from immense spatial dispersion in deaths from three plagues in the Baltics (1605-6, 1657, 1710-2), which I show is uncorrelated to a host of local, pre-plague characteristics. To measure the intensity of labor coercion, I hand-collect thousands of serf labor contracts in Estonia, which capture the work obligations of serfs. I find that labor scarcity substantially increases coercion à la Domar (1970). Investigating mechanisms, I find that this effect is enhanced by the lack of outside options and increased labor monopsony power, in line with theoretical models. Investigating the consequences of (labor-scarcity instrumented) coercion, I find negative effects on education and increased migration. Taken together, these findings highlight the conditions under which labor scarcity raises coercion and provide suggestive evidence of why it does not in other cases (e.g., in Western Europe following the Black Death).