And they call it homogamy

The educational homogamy meme has put a burr under my saddle. While I do think it is an important element in understanding inequality trends, what concerns me is the insinuation that this explains most of the rise in inequality, a position taken by Margaret Wente and William Watson to sound off that there is nothing we can do.

Guess what? Women’s liberation is to blame, so take that, you lefties.

Like all good spin, there is a grain of truth to the charge. Several decades ago, a lot fewer people went to university, and men were more likely to go than women. When they married their spouse was more likely to become the home-maker while they went to work. As universities expanded, a greater share of young people went, and the scales have evened out in terms of women and men attending. You don’t have to have attended a frat party to know that many relationships formed out of those years. And after graduation, professions that draw on people with university degrees also offer up another venue to meet a mate with similar credentials. Thus, more educated people tend to be marrying other more educated people.

So this explanation makes some sense, but as Andrew Jackson has pointed out, it is far from the whole story. There are actually two sides to the research behind the story. First, there is a study on educational homogamy, “The changing role of education in the marriage market: Assortative marriage in Canada and the United States since the 1970s” (available here) and second, there is how that aspect of family inequality plays into a bigger report by Andrew Heisz, “Income inequality and redistribution in Canada: 1976 to 2004” (available here).

It is worth noting that Heisz’s paper rather subtlely points his finger in this direction of family inequality one of several contextual explanations of rising inequality trends. He does not connect the dots in the way that Wente and Watson do, and in fact does not use the term “educational homogamy” at all. He states on page 14:

The family of the 2000s is very different from the family of the 1970s, and many of these differences would be expected to affect income inequality. Among the most relevant trends for the present study are the aging of the population, the rising share of persons in lone-parent families, and trends in marital earnings correlations. From 1970 to 2005, the share of the population aged 65 and over has steadily increased from 8% to slightly more than 13%. Seniors have lower average after-tax income than others and receive a larger share of their income from transfers, and less from market sources. Moreover, the senior population places some downward pressure on after-tax-income inequality. Thus, an increasing share of seniors in the population may influence both inequality and income redistribution. A rise in the share of lone-parent families will, all else equal, affect the bottom end of the distribution more as lone parents tend to have lower income: it will also affect trends in redistribution, as lone parents tend to receive more from transfers than others. Finally, rising correlations in income among spouses will tend to increase family-income dispersion, and greater increases in hours among wives of high-wage men than among others would also increase inequality. The net effect of these changes would be ambiguous, as some are expected to increase inequality and others decrease it.

And like a good researcher, he qualifies this explanation as needing further research. As for the study on homogamy, it posits that this is a cause of rising inequality but makes no attempt to assess what share of the total rise is due to homogamy. So a fair reading of the two studies would not lead one to jump to rash conclusions, like the two columnists I cite above.

A final gripe on terminology of the homogamy study: getting married is one of the most profound things one can do in life; calling it a “marriage market” trivializes that relationship, as if choosing your life partner was equivalent to selecting the right SUV. This is poor choice of metaphor, though perhaps just reflective of the times, as pop culture icons get married for a few years then move on to their next marriage. Which, now that I think about it, is kind of like the relationship we have with cars.


  • I am similarly skeptical of homogamy as the “cause” of inequality. At the risk of stating the obvious, homogamy can only amplify (among households) inequality that already exists (among individuals). Homogamy itself is not really the origin of any inequality.

    If individual incomes were equal, all marriages would be homogamous (in terms of income) and there would be no household inequality. Although this example is extreme and unrealistic, the policy implication is clear: reducing individual inequality reduces the household inequality that homogamy entails.

    In fact, I think that homogamy strengthens the “left-wing” view of inequality. If rich people were marrying poor people, then individual inequality might naturally even out at the household level. The rise of homogamy underscores the importance of addressing inequality among individuals.

  • I’d agree that homogamy is unlikely to have been the driving force behind the increase in inequality we’ve been seeing, but the long-run implications for social mobility are disquieting. Parents’ education levels are a very good predictor for PSE attainment, so education-based homogamy may further reduce social mobility.

  • andrew jackson

    John Myles kindly sent me some references. Homogamy in itself is not a huge driving force but patterns of family formation do loom large in some studies of family income inequality in Canada (the other part of family formation trends being the growth of single parent families.) The Fortin study in the Kesselamn book looks at this but was for an earlier period.

    The link from individual earnings inequality to family inequality is indeed tricky eg to soem extent the (modest) convergence of male and female individual earnings has – for the whole earnings distribution – masked increased inequality smong both men and women considered separately.
    Like Marc I found Heisz’s study very useful on close reading. One striking thing is that he concentrates on a 1989 to 2004 comparison to show that tax/transfers continue to offset growing market income inequality – but there is a real break in the numbers from the mid 1990s showing erosion of the equalizing role of transfers. Heisz is comparing cylcically similar years which is reasonable – but there’s still an important trend which would be unlikely to be reversed in a major way in a downturn given changes to EI and social assistance.

Leave a Reply

Your email address will not be published. Required fields are marked *