Employment protection legislation often varies with firm or establishment size and age, and many evaluations of EPL reforms exploit these thresholds. This paper develops a search-and-matching model with multi-establishment firms to disentangle direct effects from general-equilibrium spillovers. We combine the model with quasi-experimental evidence from a 2009 Portuguese reform that restricted fixed-term contracts in young establishments of large firms. Spillovers barely distort the reduced-form estimate, yet they overturn the aggregate conclusion: extrapolating the reduced-form estimate implies a 1.51% employment decline, compared with a 0.11% increase in general equilibrium. Counterfactuals show that conditioning EPL on size versus age can yield opposite welfare effects, even when employment effects are similar, and that fixed-term contract and firing restrictions must be evaluated jointly.