Minimum hourly wages were randomly imposed on firms posting job openings in an online labor market. A higher minimum wage raised the wages of hired workers substantially. However, there was some reduction in hiring and large reductions in hours-worked. Treated firms hired more productive workers, which can explain, in part, the reduction in hours-worked: with more productive workers, projects were completed in less time. At the conclusion of the experiment, the platform imposed a market-wide minimum wage. A difference-in-differences analysis shows that, in equilibrium, firms still substitute towards more productive workers, adversely affecting less productive workers.