“There’s a dispiriting fact that when a family gets a voucher, they typically don’t move to substantially better neighborhoods,” says Rob Collinson, a PhD candidate in public policy at NYU. Specifically, he’s talking about the HUD housing vouchers distributed to help lower-income Americans with monthly rent. While the federal agency said last year its research shows that the vouchers are effective when it comes to reducing homelessness, other studies have shown that vouchers alone might not be enough to get a recipient to leave a high-poverty neighborhood she knows for an unfamiliar higher-opportunity neighborhood.
Collinson and his research partner, Peter Ganong, a PhD candidate in economics at Harvard, set out to explore why vouchers haven’t been a stronger government tool when it comes to social mobility in the U.S.
A few popular explanations reign: lack of affordable housing in high-opportunity neighborhoods, resistance to leaving neighborhoods because of existing social ties, discrimination and vouchers not covering enough of the rent. Collinson and Ganong wanted to target and test that last theory, and in 2011, a unique occurrence in Dallas was beginning to provide just the right data. The pair seized the opportunity to develop some brand-new models, based on their background in economics, as well as more elusive factors such as “neighborhood quality.”
That just-right data became available thanks to a lawsuit that a fair housing nonprofit in Dallas brought against HUD regarding the agency’s method of determining fair market rent (FMR). By setting voucher amounts across a whole metro area, the nonprofit argued, HUD kept higher-opportunity areas with higher-quality housing out of reach for many recipients. The case outcome: HUD agreed to a pilot program in Dallas in which it would set amounts by ZIP code, a much smaller area than a metro. (Collinson was working at HUD at the time. He remains an unpaid HUD employee and asked that I note his views here are strictly his own.)
Around the same time, Peter Ganong was working at the White House’s Council of Economic Advisers while collaborating on government-sponsored enterprise reform research. His and Collinson’s mutual interest in housing affordability led them to collaborate, and ultimately produce the working paper “The Incidence of Housing Voucher Generosity.” In it they reveal their findings about metro-wide vouchers vs. those based on ZIP codes, as well as vouchers and their costs and benefits to tenants, landlords and the government.
This is worthy policy research, of course, but the most innovative aspect of their study is the model they developed, considering economics and human behavior, to analyze voucher recipients’ actions.
“It’s an original model primarily building on a literature of economic search models. There’s a large literature on how people search for jobs, and this adapts some search theoretic work and looks at a housing application,” explains Collinson. “But we worked on the model to tailor it to dynamics particular to the voucher program.”
The model tries to get at the tradeoff that voucher holders must make in finding a unit that’s both of acceptable quality and also comes with a landlord willing to accept a voucher. Unfortunately, some landlords of higher-quality units shun voucher holders. (There are states with laws against this, with varying degrees of success with enforcement.)
Collinson and Ganong also developed a neighborhood quality index that they used to determine whether or not, after receiving more “generous” or higher-ceiling vouchers, recipients moved to “better” neighborhoods. (They used established factors like child development, outcomes later in life, intergenerational mobility, violent crime rates, test scores, poverty rates and more.)
After measuring how many voucher holders had moved to better neighborhoods, they compared the data with new leases signed in Fort Worth, Texas, after HUD changed its Dallas policy. Fort Worth was a kind of control group. They found the quality of the neighborhoods that voucher recipients moved to in Dallas rose significantly. And interestingly, most movers had not received the mobility counseling made available through the same court settlement.
“Voucher holders have to trade off neighborhood quality with both the odds that they find a landlord that will lease to them and the structural quality of that unit. [With metro-wide vouchers], they might choose to search in the same poorer neighborhoods that they always did because they’ll have higher odds of successfully leasing up, or because they want better physical structure quality,” Collinson explains. But when HUD drilled down to a ZIP code snapshot, he says, “the old neighborhoods they searched in now have a lower allowable subsidy, so they are less likely to find a landlord in those neighborhoods than they previously were.”
Overall, the pair’s research points to ZIP-code-based vouchers being more helpful to recipients. From their paper:
Raising the generosity of the rent ceiling everywhere appears to primarily benefit landlords, who receive higher rents with very little evidence of medium-run quality improvements. Setting ZIP code-level rent ceilings causes rent increases in expensive neighborhoods and decreases in low-cost neighborhoods, with little change in aggregate rents. The ZIP code policy improves neighborhood quality as much as other, far more costly, voucher interventions.
HUD has now piloted the more targeted program in five other cities and wants to replace across-the-board FMRs altogether. But, Collinson asserts, that would merely represent “a reasonable first step.”
“It’s an important objective to get recipients of vouchers into better quality neighborhoods, but there’s no one-size-fits-all solution,” he says. “This policy might not work well for the disabled or elderly. And some areas have excess market and some are extremely tight. Policy has to be designed with differences in mind.”
The Science of Cities column is made possible with the support of the John D. and Catherine T. MacArthur Foundation.