Pre-print, Working paper: In this paper, I study the wage premium associated with firm-level union recognition in France and show that this premium is due to a rent-extraction phenomenon. Using a large matched employer-employee dataset from a 2002 survey in France, I first estimate a series of wage determination models that control for individual and firm-level characteristics. I find that union recognition is associated with a 2-3% wage premium. To show that this premium results from a non-competitive phenomenon, I construct a bargaining model and estimate it empirically using a smaller but very detailed matched employer-employee dataset for 2004. The model predicts in particular that the wage premium obtained by unions should increase both with their bargaining power and with the amount of quasi-rents per worker available in the firms they organize. These predictions are validated empirically when I use the firms' market share as a proxy for their quasi-rents and the percentage of unionized as a proxy for the unions' bargaining power. All the results remain valid when I control for the firm-level workers' average productivity.