We use longitudinal data based on administrative archives from 1985 to 2002 to estimate the relationship between wages and firm size in Italy. By controlling for individual fixed effects, we find that larger firms pay significantly higher wages, although the individual unmeasured ability component accounts for about one half of the uncovered size–wage premium. To reduce potential self-selection problems arising from endogenous job changes, we focus on a sample of workers displaced by firm closures. By using this sample, we confirm that larger firms pay higher wages, in part for unmeasured workers' abilities and in part for firm size effects.