Survey evidence finds a wage premium for married men over single in most of the western world. Three key theories are put forward as an explanation: 1) marriage makes men more productive and therefore increases their wages; 2) men with higher labour-market productivity, and therefore higher wages, are more likely to be married; 3) employers simply favour married men over unmarried. We use a two-step analysis to test these three theories. In the first step, we analyse national panel data from Switzerland to pinpoint the part of the penalty due to either productivity or selection. We use entropy balancing to match never-married men to married on a set of pre-labour market covariates, thus isolating the selection effect, before we perform fixed effects regressions for productivity effects and to uncover any unexplained residual. We find a premium for married men of 5%, much of which is explained by selection. Next, we seek to uncover employer preferences by using a factorial survey experiment among HR managers (N = 714) in Switzerland. We ask the managers to assign wages to the CVs of fictional job candidates, who vary randomly on their civil status, amongst other characteristics. We can therefore identify employers’ preferences concerning married and unmarried men. We find that recruiters assign a small premium to married men, contingent on the job applied for. Overall, the premiums we find are lower than those previously reported in the literature.