Wage equations using cross-sectional data typically find an earnings premium in excess of 10 percent for married men. One leading hypothesis for the premium is that marriage facilitates specialization that enables married men to become more productive than single men. Another is that the premium is attributable to an unobserved fixed effect, married men possessing qualities that are valued in the labor market as well as the marriage market. This paper suggests that the premium is attributable to an unobserved time-distributed fixed effect that emerges and grows with the approach of marriage and continues to grow for some years after marriage. A similar distributed fixed effect is found in the case of women, but it is smaller and declines after a few years of marriage. The results appear to cast doubt on the specialization hypothesis.