Who Competes for Whom? Monopsony in Ability-Segregated Labor Markets

I develop a general-equilibrium oligopsony model in which firms differ in their demand for worker ability, generating worker-specific monopsony power. Taking the model to matched employer–employee data for Italy and Germany reveals ability segregation that localizes competition: firms compete most intensely with similar firms targeting the same ability segment. In the calibrated model, monopsony power and welfare losses are largest for low- and high- ability workers, who face fewer effective employers. Output losses are modest relative to standard quantitative benchmarks. Labor-market power amplifies wage inequality because wider between-market wage dispersion outweighs compressed assortative matching and top wages.