Why Are Black-White Marriage Rates So Low?

Pick two Americans who married this year and trace them back a decade, before they had ever met, and the odds are good they were already living close to one another. Not the same street, necessarily. Not even the same block. But near. Benjamin Goldman and his colleagues went hunting for this pattern across the tax records and census files of some 31 million people, and there it sat, stubborn and plain: most spouses were neighbours long before they were anything else.

That one fact carries a lot of weight. If we marry the people we live among, and we live among people who tend to look and earn much as we do, then residential segregation is not only sorting our schools and our streets; it’s arranging our marriages too.

So Goldman, now at Cornell, working with Jamie Gracie at Harvard and Sonya Porter at the US Census Bureau, asked the next question. What would happen if you mixed the neighbourhoods up? Would more couples form across the lines of class and race that the country has drawn so firmly? The answer, when it arrived, split cleanly down the middle, and the two halves do not match. More marriages across class lines, yes. Across the colour line, almost none.

Here is the snag any honest researcher hits first. People who live in mixed neighbourhoods are already unusual.

The white family on a mostly-Black block, the kid from a poor household growing up in a wealthy suburb, may have ended up there precisely because they were already open to crossing those lines. Compare them with everyone else and of course you will see more intermarriage, but you will never know whether the neighbourhood did the work or whether these were simply different sorts of people to begin with. Exposure and openness come hopelessly tangled together.

A lottery hidden in the birth records

What the team needed was something that scrambled the neighbourhood without scrambling the people in it. They found it in the oldest lottery going: whether a baby arrives as a boy or a girl.

Think it through. The sex of any given newborn is, near enough, a coin flip, and it has nothing to do with what sort of place the parents chose to live in. Suppose a boy happens to be born somewhere where, purely by chance, the nearby high-income families had rather more daughters than sons in the years around his birth. Because Americans mostly settle near where they grew up, that boy will come of age in a marriage market subtly tilted toward high-income women, an imbalance nobody engineered and the dice alone produced. The researchers call this “market tightness,” and it handed them a way to nudge a person’s pool of plausible partners without touching a thing about who that person actually was. That such a faint signal survives at all in the roar of 31 million lives is, frankly, a little astonishing.

It survives. And the result is sharp.

Tilt the market across class lines and people respond: a 10-percentage-point rise in cross-class tightness lifted the chance of marrying across class lines by about 1.3 points. Modest-sounding, until you notice it closes more than a third of the gap between how often poor and rich Americans actually marry each other and how often they would if class were ignored altogether.

The lever that wouldn’t engage

Now run the same machinery across race. The effect on Black and white marriage came out at roughly 0.001 of a percentage point, so close to nothing that the statistics can rule out anything more than a whisper. Same method, same vast dataset, same country. The lever that shifted class simply would not engage with race.

“At the beginning of the research, it wasn’t obvious that these two types of marriages should react differently to exposure,” says Gracie. The point of the paper, he is careful to add, is “to document that difference, as opposed to necessarily trying to understand why.” Which is the kind of restraint you learn to trust.

Documenting it took one more step, mind you, because a single nudge in one neighbourhood is not the same thing as redrawing the whole map. Marriages spill. If a woman in one tract finds a partner she would not otherwise have met, then somebody else, somewhere else, does not. To follow those ripples across an entire city, the team built a model of the marriage market, fed it their real-world numbers, and ran the experiment no government would ever sign off on: they wiped out segregation completely and watched who married whom. In a United States with no residential segregation at all, interclass marriage among people from low-income families climbs from 3.1 to 4.6 per cent, closing about 35 per cent of the homophily gap. Black-white marriage, over the same sweeping change, inches from 2.1 to 3.0 per cent, closing barely four per cent of its own.

There is a wrinkle here worth sitting with, because it very nearly contradicts itself. Shift only a handful of families out of the most segregated blocks, as real programmes such as Chicago’s Gautreaux Project once did, and you do get new interracial marriages, around a one-point rise for a ten-point jump in exposure, closely matching what actually befell those families decades ago. So small doses work. It is the wholesale version that fizzles, and the model offers a reason: in any population there is a thin slice of people open to marrying across racial lines, and a little integration is enough to let those matches find each other. Once they have, the well runs dry.

And that is the uncomfortable shape of the thing. For class, proximity is much of the story; bring people into contact and the marriages follow, with everything that implies for the resources the next generation of children will inherit. For race, proximity turns out to be almost beside the point. People can share a postcode and still keep their social worlds, their churches and friendships, sorted by colour. Why the two come so completely apart, the data cannot say. It might be sharper preferences; it might be that living near someone of another race in America still so rarely becomes knowing them.

Gracie puts the open question simply: “Does growing up in a more mixed neighborhood shape your attitudes toward people of different groups? We don’t have an answer to that.” The children of those 31 million are only now reaching adulthood. When they marry, the records will be waiting.

Frequently Asked Questions