The “Fading American Dream” paper isn’t the first time economics professor Raj Chetty and his team have used data to upend previous research.

Nearly 25 years ago, the U.S. Department of Housing and Urban Development (HUD) began a monumental social experiment to test whether moving low-income families to lower-poverty neighborhoods would improve the lives of the parents and children involved.

More than 4,600 families were enrolled and assigned to one of three randomized groups: a group that received vouchers to move to areas with poverty rates below 10 percent, as well as counseling on the potential relocation and help in leasing a new unit; a group that received vouchers that could be used anywhere; and a group that was eligible only for preexisting programs.

The results of the experiment were largely disappointing. Although the recipients of the lower-poverty vouchers became healthier, the adults did not experience better outcomes in employment or income.

Then, Chetty and Harvard professor Nathaniel Hendren came along. In 2015, the pair combined forces with Harvard professor Lawrence Katz to reanalyze the data. Because the children in the experiment were older than they had been in previous evaluations, Chetty and Hendren were able to test their employment outcomes.

The updated results? The experimental vouchers increased the future annual earnings of young children who moved to lower-poverty neighborhoods by 31 percent, an amount the researchers estimated would increase total lifetime earnings by about $302,000.

When Gregory P. Russ, then the executive director of the Cambridge Housing Authority, first started reading reports about the new research, he sent Chetty an email, and the pair made a commitment to study the issue further. More than a dozen public housing authorities around the country agreed to participate, including the Seattle Public Housing Authority, where Andria Lazaga is director of policy.

Lazaga said the new analysis of the HUD experiment was a “game changer.”

“We had all written it off,” she says of the experiment. “When the revised data came out, it was like, in the long run, there really is a payoff if we do this right.”

The United States spends approximately $20 billion each year on vouchers, but 80 percent of them are used in moderate- or high-poverty areas. Chetty and Hendren estimate that the eventual additional tax revenue from children who move to lower-poverty neighborhoods would offset the cost of the vouchers used to move them there.

Today, public housing agencies around the country have partnered with government officials, housing advocates, Chetty and other scholars in an effort to maximize low-income families’ chances for success.

“Here’s a case where the data has shaped thinking and now the data is about to shape practice,” says Russ, who became executive director of the Minneapolis Public Housing Authority in February and promptly signed it up to participate in Chetty’s research. “The real power is not just changing your mind that a method could work; the real power is in designing a program based on what really amounts to a scientific assessment of its impact. That’s rarely done.”


Rebecca Beyer is a freelance writer in New York City.