Say you don’t need no diamond ring

There aren’t many articles I’ve read that have really changed the way I see the world, but “Gender Identity and Relative Income Within Households,” by Marianne Bertrand, Emir Kamenica, and Jessica Pan is one of them. It’s a remarkable exercise of Occam’s Razor, in that it shows that a remarkable number of different phenomena- in particular changes in propensity to marry both across and within races, economic strata, and geographic areas- can, to a significant degree, be summarized with a single number, the relative income of men and women who could potentially marry one another.

We examine causes and consequences of relative income within households. We show that the distribution of the share of income earned by the wife exhibits a sharp drop to the right of 1, where the wife’s income exceeds the husband’s income. We argue that this pattern is best explained by gender identity norms, which induce an aversion to a situation where the wife earns more than her husband. We present evidence that this aversion also impacts marriage formation, the wife’s labor force participation, the wife’s income conditional on working, marriage satisfaction, likelihood of divorce, and the division of home production. Within marriage markets, when a randomly chosen woman becomes more likely to earn more than a randomly chosen man, marriage rates decline. In couples where the wife’s potential income is likely to exceed the husband’s, the wife is less likely to be in the labor force and earns less than her potential if she does work. In couples where the wife earns more than the husband, the wife spends more time on household chores; moreover, those couples are less satisfied with their marriage and are more likely to divorce. These patterns hold both cross-sectionally and within couples over time.

The nice thing about this hypothesis is that you can pretty easily reproduce the basic pattern with publicly available data, in this case from the Census’s 2014 Current Population Survey. For example, most people probably know that there is a strong relationship between income and marriage:


But it may be less well known how much larger the difference between men and women’s earnings is at higher household incomes than at lower.


That graph is not conditioning on marriage, it is just showing the ratio of the income that women report earning themselves to the income men report earning, for each level of household income.

This ratio (of reported female to male income) corresponds closely to the fraction of respondents who are married:


This relationship is largely mediated through marriage (not just from women being shut out of high-earning careers, that is.) For example, here is the fraction of total household income women report making:


And here is the same relationship if you only look at married female respondents: the relationship is mostly attenuated.


It’s always been true that richer people are more likely to marry, but while well-off people’s marriage rates have slightly declined, middle and lower-income people’s marriage rates have declined more.  So here are 1980 numbers (adjusted for inflation, though more about that later) versus 2014 numbers:


This corresponds well to Bertrand and her colleagues hypothesis that groups with relative incomes closer to 1 (parity between male and female workers) would have larger drops in marriage rates.

I adjusted the 1980 income numbers in the previous graph by the Consumer Price Index (CPI), which tends to overstate inflation; another inflation index would tend to move the blue (1980) dots to the left, making the drop in marriage since 1980 larger for most incomes rather than smaller.

Most of these graphs are for middle aged (40-55 year old) respondents; relative incomes for men and women in 2016 are already much closer to 1 for younger workers, with some surveys suggesting that young women are already exceeding young men in earnings. If the pattern Bertrand and her colleagues observe persists, this would suggest that marriage rates may decline much faster than previously anticipated.



16 thoughts on “Say you don’t need no diamond ring

  1. What is left unstated by you and the authors of the article is the societal implications:
    1) Marriage is one the greatest stabilizing forces in society, nationally and locally. The trends described and predicted by the data implies grave consequences for society.
    2) Marriage remains the best place to raise the next generations of Americans.
    3) Everything described by Charles Murray in Coming Apart,


  2. Another related question which is not addressed is why do some classes tend to choose to have children out of wedlock at a younger age vs. other classes who choose to delay having children until they are married or have a career.

    I think that Ruby Payne is on to something here, even with all of the numerous flaws in her theories.


  3. Is it possible that the study has the cause and effect reversed? You quote “…gender identity norms, which induce an aversion to a situation where the wife earns more than her husband…” I take that to mean–in plain, non-Vulcan English–that men don’t want to marry women who out-earn them.

    But surely that works the other way at least as much…surely it is obvious that high-earning women are generally averse to low-earning men?

    Another factor not discussed is “assortative mating”, the contemporary phenomenon where people tend to marry mostly within their post-educational, in-career demographic. Business executives don’t marry their secretaries anymore: they marry their fellow executives who are graduates of the same or similar schools.

    Given that significant pay disparities within an organization’s workforce are generally associated with differences in power and authority, necessarily such encounters would be less likely, given the current hair-trigger environment with respect to sexual harassment: one would mostly have such encounters with people at one’s own level.

    It is of course not clear whether assortative mating is a cause or an effect.

    Another thing that is not entirely clear is, are the actual earnings adjusted for buying power? Here in the DC area, two mid-level civil servants, each at the GS-12 or GS-13 level, making a minimum of $62k or $74k respectively for a combined household income of at least $124k would be struggling to get by: obviously they would be much better off in a lower-cost area.


    1. Bertrand notes that within marriages, when the wife earns more than the husband, the wife expresses less satisfaction with the marriage; my reading is that she believes that the gender norms are held as strongly by women as by men. I don’t believe her data sources control for cost of living or buying power; there’s a separate argument to be had about whether two-earner households are sometimes monetizing services (eg, childcare) without increasing quality of life, which might be a potential confound.


  4. In other words, wage is often a signal for accomplishment. Women who marry men who earn less then they do might be disappointed in “marrying down”, independent of the cost of living question.

    If possible, I would like to run this data for women married to artists, academics, clergy, etc. (which theoretically could be done with Chetty’s data, but i don’t know about the other sources) to see if patterns change.


  5. Pingback: spottedtoad

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s