We use linked employer-employee data to investigate the job satisfaction effect of unionisation in Britain. We depart from previous studies by developing a model that simultaneously controls for the endogeneity of union membership and union recognition. We show that a negative association between membership and satisfaction only emerges where there is a union recognised for bargaining, and that such an effect vanishes when the simultaneous selection into membership and recognition is taken into account. We also show that ignoring endogenous recognition would lead to conclude that membership has a positive effect on satisfaction. Our estimates indicate that the unobserved factors that lead to sorting across workplaces are negatively related to the ones determining membership and positively related with those generating satisfaction, a result that we interpret as being consistent with the existence of queues for union jobs.