Noncompetes often cover highly trained, high-paid workers but are also widespread in low-skill, low-pay service jobs. This paper asks where they hurt workers more. Using a dynamic monopsony job-ladder framework, we show that noncompetes depress wages by reducing competition, with potentially severe effects when adoption is widespread. The impact on wages is particularly adverse when they are used by firms with high productivity and high costs of training workers, as these are the firms with large rents. In contrast, the effects are more muted when low-rent employers use noncompetes.
Gottfries et al. (Fri,) studied this question.