At most 10% of the estimated marriage premium is due to selection

Well this is interesting. In the past couple of years I have dug up the data to support something my father told me years ago; your chance of being in poverty is miniscule if you do just three things - graduate high school, get married, get a job, any job. I don't know where he came across this information but I have been able to validate it and in fact, if you have these three characteristics, you have a less than 2% chance of being in poverty.

All along though, it has been obvious that these are just markers for something else going on. Just the existential fact of being married isn't causative of income. Clearly, some of the human capital skills, or non-cognitive skills as they are sometimes called (empathy, self-control, tolerance, etc.) are at play - being empathetic probably increases both the probability of your being a good spouse and an effective employee. But is there an operational component as well? Is there something about the institution of marriage that also contributes to greater employment effectiveness? This study, Is the male marriage premium due to selection? The effect of shotgun weddings on the return to marriage by Donna K. Ginther and Madeline Zavodny answers in the affirmative.

In standard cross-sectional wage regressions, married men appear to earn 10 to 20% more than comparable never married men. One proposed explanation for this male marriage premium is that men may be selected into marriage on the basis of characteristics valued by employers as well as by spouses or because they earn high wages. This paper examines the selection hypothesis by focusing on shotgun weddings, which may make marital status uncorrelated with earnings ability. We compare the estimated marriage premium between white men whose first marriages are soon followed by a birth and other married white men in the United States. The return to marriage differs little for married men with a premarital conception and other married men, and the results suggest that at most 10% of the estimated marriage premium is due to selection.

So indeed, being a good spouse means you have some of the same attributes that make you desirable as an employee (selection) but there is some larger effect of being married that also makes you a more effective employee.