Do people in poverty create bad neighborhoods, or do bad neighborhoods cause people to be poorer? This chicken-or-the-egg question is known as “selection bias” among social scientists and has been debated for decades. But a new University of Michigan study on housing vouchers provides some clarity.
The study, by researcher Eric Chyn, compared people who moved from demolished public housing with people whose public housing was untouched. Chyn found that children who moved to lower-poverty areas had better educational and employment outcomes as adults. But are the positive results because of the neighborhood change or something else?
How the study worked
Chyn’s study used a “natural experiment” of public housing demolitions in Chicago. All the residents of demolished buildings received vouchers to move elsewhere in the city. His study brings to mind the Moving to Opportunity (MTO) experiment, which gave vouchers to families in five different cities to move to lower-poverty neighborhoods, and then followed them to see how they fared.
While there are several important differences between MTO and Chyn’s study, the biggest one is agency. Families in MTO could decide whether to use their vouchers to move, but the ones in Chyn’s study had no choice because of the demolition. This kind of unilateral action gets around a common criticism of MTO and other housing and school voucher experiments: that the families who are willing to apply for and use the vouchers are more committed to improving the wealth and health of their children to begin with. Chyn’s study essentially eliminates selection bias to show that leaving a particularly bad area as children can lead to more education and more pay as adults, regardless of parental characteristics or motivations.
For instance, approximately half of the families who were assigned vouchers in MTO actually moved to different housing. The families that Chyn observed were all residents of public housing in poor neighborhoods in Chicago. Ultimately, he shows that there wasn’t any notable difference between the families aside from the fact that some families had to move because of the demolition and others could stay put. So when he finds that children from the demolished housing were 9 percent more likely to be employed and earned 16 percent more in income as adults, he can confidently claim that something about just moving—beyond any family motivation—caused more upward mobility.
But what exactly is behind the improvement?
Chyn and others who discussed his findings argue that these improved outcomes are because of the neighborhood change that the demolition and subsequent move caused. The study clearly demonstrates that the displaced families found themselves in areas with lower poverty, unemployment, and crime rates than nondisplaced families in the years immediately following demolition. Families that were displaced lived in neighborhoods with poverty rates nearly 28 percentage points lower than residents who were not displaced. However, that difference decreased by half after three years and disappeared completely after 10 years, mostly because nondisplaced residents moved out of public housing over time.
The average neighborhood poverty rate for residents in Chyn’s study was 78 percent before demolition—ranking it among the poorest locales in the country. Having more than 40 percent of residents in a neighborhood live in poverty is a well-documented marker of concentrated poverty, a line where the negative factors associated with living in poverty compound, making circumstances harder for all the residents of that neighborhood. Put simply, the displaced families moved from very bad areas to somewhat less bad areas, but the new neighborhoods still look very different from the ones where most Americans live.
Though the displaced families moved to places with lower poverty rates, they did not move to places with low poverty rates, meaning these families never escaped the trap of concentrated disadvantage. Compare this with the families in the MTO study who received special “experimental” vouchers that required them to move for at least one year to neighborhoods where less than 10 percent of the residents live in poverty. That difference in neighborhood composition appears to translate to bigger differences in opportunity gains. The children in Chyn’s study who were displaced went on to earn about $602 more a year as adults than those who weren’t displaced, while young children in MTO who received the experimental vouchers to move to low-poverty neighborhoods went on to earn $3,477 more a year than those who didn’t receive vouchers.
If these positive outcomes weren’t because of moving to a different kind of neighborhood, what caused them? While there is no definitive answer, I would be willing to guess that just leaving these public housing projects played a large role. The high-rise public housing projects of 1980s and 1990s Chicago were notorious in their dysfunction and the subject of many news reports, books, and documentaries.
MTO shows us that not all residents of bad neighborhoods are willing and readily able to move to better ones, but when they do, they will likely see better health, more safety, and a chance for their kids to move out of poverty. Chyn’s study shows us that if policymakers can get even more of our most disadvantaged residents into better housing and neighborhoods, whether by providing them opportunities to move or by improving the neighborhoods in which they live, these residents would see positive results. Even if earnings are more modest, more people were helped, and that seems like a trade-off worth exploring.