The inability of poor kids to get ahead in the US is a growing topic of concern, both because of the unfairness of having one's life outcome determined at birth, and because denying children opportunities on the basis of income risks making society as a whole worse off. "Poor kids, through no fault of their own, are less prepared by their families, their schools, and their communities to develop their God-given talents as fully as rich kids," Harvard political scientist Robert Putnam writes in his new book, Our Kids. "For economic productivity and growth, our country needs as much talent as we can find, and we certainly can't afford to waste it."
New research from a group of economists at Harvard, the Treasury Department, and the London School of Economics provides a particularly vivid illustration of how disadvantage can harm the economy at large. The researchers — Harvard's Alex Bell, Raj Chetty, and Xavier Jaravel, Treasury's Neviana Petkova, and LSE's John Van Reenan — match up data on patents in the United States with tax returns of the people getting the patents — and the returns of their parents. That lets them see not just how much inventors earn as adults, but how much their families made when they were children. That means they can compare how often children born into poor or middle-class families get patents relative to children of privilege.
Kids from rich families are much likelier to get patents later in life. That's not surprising, but the sheer magnitude of the gap is. Patent rates grow exponentially with childhood income:
It's not just that the poor are at a disadvantage; the rich-but-not-super-rich are, too. Patent rates of the 90th percentile are roughly half that of the 99th percentile.
Worse, Bell and his coauthors also get the public-school test scores for the New York City residents included in their study. Children from poor families were less likely to get patents even when they got extremely high test scores:
So the issue isn't just that rich kids are inherently more talented. The difference seems to be education. The researchers suggest that about 90 percent of the relationship between childhood income and patent rates is explainable by children of privilege being more likely to attend high-quality colleges; they note that 23 percent of patents (weighted by the number of times they're cited, so higher-quality patents count more) come from students from just 10 colleges. Those colleges account for only 2.7 percent of US college enrollment.
The research isn't bulletproof. Patents aren't a perfect measure of innovation; plenty of important inventions go unpatented, like the World Wide Web or friction matches, and plenty of patents that get approved are low quality. A person's ability to patent inventions may have more to do with access to and money for patent lawyers and familiarity with the process than genuine skills (which could actually strengthen the researchers' point about children of poor and middle-class homes being shut out from patents). That said, Bell et al. are hardly alone here, and it's standard within economics to use patents as a proxy for innovation.
Further, the researchers were only able to reliably identify the parents of children born after 1980, and since they looked at patents up to 2012, they effectively only looked at patents for people under 32. That's not a tiny sample, but it might not be representative of patent recipients as a whole. And test scores are, while useful, an imperfect measure of aptitude.
So don't take Bell and his coauthor's findings as gospel. But despite the research's limitations, it's one of the more compelling analyses of the costs of social mobility, not just to children of poor families but to innovation and the overall health of the economy.