We use a novel instrument based on social norms towards marriage to present a new finding: being married has a positive causal effect on the wages of both men and women. Despite the striking changes in the labor market and the composition of families over the past decades, the premium for men is sizable and has only slightly diminished over time. In contrast, we document a stark change for women: the effect of being married has evolved from a penalty into a premium. We present new evidence to uncover the mechanisms behind these causal patterns. The premium for men can be mainly rationalized as a byproduct of household specialization à la Becker (1985, 1993): within married couples, men tend to concentrate on market work, which leads them to supply more market hours than their single counterparts and earn higher wages. In the past, the penalty for women was also a consequence of household specialization, with married women working significantly fewer market hours than their single counterparts. The degree of specialization between spouses has decreased over time, allowing married women to work virtually as many hours as singles and leading to the disappearance of the penalty. The marriage wage premium for women in recent years can be explained by one of the channels in Pilossoph and Wee (2021): because of income pooling with their husbands, married women have a higher reservation wage, which allows them to select better-paid job opportunities.