Economists increasingly refer to monopsony power to reconcile the absence of negative employment effects of minimum wages with theory. However, systematic evidence for the monopsony argument is scarce. In this paper, I perform a comprehensive test of this argument by using labor market concentration as a proxy for monopsony power. Labor market concentration turns out substantial in Germany. Absent wage floors, higher concentration reduces wages and employment, reflecting monopsonistic conduct of firms. Sectoral minimum wages lead to negative employment effects in slightly concentrated or more competitive labor markets. This effect weakens with increasing concentration and, ultimately, becomes positive in highly concentrated or monopsonistic markets. Overall, the results lend empirical support to the monopsony argument, implying that conventional minimum wage effects on employment conceal heterogeneity across market forms.