• For many families, the way out of poverty is to move to a better neighborhood
  • A new study shows modest investments in information combined with supportive services can help them make that move.
  • We need to rethink attachment to neighborhood, and see moving as one way out of poverty for many low income households

The Seattle Experiment: Encouraging moves to opportunity

The arc of Raj Chetty’s work has been to convincingly illuminate, via big data, the critical role that neighborhood characteristics play in shaping our life chances, particularly for kids growing up in poverty. His project’s latest work, now under the banner of Opportunity Insights, is to test some of his findings in the real world with policy experiments. A new paper reports the results of a housing voucher experiment in Seattle, designed to figure out and overcome the barriers low income families face in getting housing in neighborhoods with better life chances.

The study’s approach and methodology are explained in some detail in excellent articles at Vox and CityLab, so we won’t reprise them here.  In most basic terms, the experiment consisted of randomly assigning families scheduled to receive housing vouchers to an experimental group–which received additional information about high opportunity neighborhoods along with customized counseling, and work with landlords–to a control group, which received just vouchers.

When provided with better information and some supporting services, voucher holders select and obtain housing in neighborhoods with higher levels of economic opportunity.  The results are stunningly large:  Those who got information and support services were 40 percentage points more likely to obtain housing in high opportunity neighborhoods than a control group, who just received vouchers; about 54 percent of “treated” participants moved into high opportunity neighborhoods, compared to just 14 percent in the control group.

To get an idea of what this means on the ground, consider this map of King County.  The area’s high opportunity neighborhoods are shaded in blue, and are principally located to the north and east of downtown Seattle. Most of the voucher holders in the control group (red pushpins) selected apartments in the Western half of the county, and especially in southern King County, where there are few high opportunity neighborhoods. (This mirrors a consistent pattern across the country in which most voucher holders are concentrated in lower income neighborhoods).  In contrast, a high fraction of the treatment group (the green push pins) selected housing in high opportunity areas (blue shading).

It’s too early, of course, to observe the lifetime effects of kids in the treatment and control groups as they grow up in these different neighborhoods (although the study will track families over time). But given Chetty, et al’s estimates of the difference in expected lifetime earnings between kids who’ve grown up in high opportunity v. high poverty neighborhoods in past decades, the author’s estimate kids in treatment families will earn an  additional $200,000 over a lifetime, with an equivalent net present value at birth of $85,000.

An earlier indicator of success, however, is qualitative analysis.  By a wide margin, those in the experimental program who moved to high opportunity neighborhoods are much more satisfied with their housing and their neighborhood, and report they are much more likely to stay than those in the control group.

. . . the treatment increased the share of families who report being “very satisfied” with their new neighborhoods by 33.6 percentage points (s.e. = 11.6, p < 0.01), from 28.6% in the control group to 62.2% in the treatment group

This kind of experiment is the gold standard of social science research methodologies:  a random-selection trial  that provides regular housing vouchers (with no added services) to a control group, and a set of information, counseling, preparation and modest financial support for move in costs (application fees, deposits, paying off outstanding utility bills, etc).

There are some important limits to the study; it’s scale is small, and ultimately, the effectiveness of the program hinges on there being enough affordable apartments in high opportunity neighborhoods.  The experiment, with just a few hundred households was hardly enough to strain the market, but the fact that even in a hot market like Seattle, low income households were able to find apartments was encouraging. Ultimately, though, the ability scale up the promising results of this program will hinge, yet again, on expanding the housing supply, particularly in high opportunity neighborhoods.

Housing search and barriers to moving to high opportunity neighborhoods

The fact that some places offer better living and better opportunities, while confirmed in dramatic detail by Chetty, et al’s work, isn’t exactly a secret.  But despite being aware that other places might offer a better shot out of poverty, many low income families seem rooted in low opportunity neighborhoods.  The big question is: “Why?”

The Seattle experiment illuminates how a series of small and seemingly prosaic barriers get in the way of making these moves. Time and the press of circumstance curtail the duration and scope of housing search. A significant fraction of program participants are dealing with medical and emotional challenges, and even domestic violence. The need to care for children, get to and from work, and deal with other challenges leaves them little time to explore the housing market. The supportive services help widen the areas voucher recipients consider and with help, they look at other neighborhoods, and much more frequently choose high opportunity places.

Its easy to imagine how people get trapped into considering only a limited number of choices.  If you’ve only lived in one part of town (a low opportunity part), your awareness of housing options in other places may be limited.
It’s likely that few of your friends, neighbors, co-workers or others have that familiarity as well.  Some housing opportunities in high opportunity neighborhoods may either be effectively invisible to low income families, or seem distant and unattainable.  It’s also clear that special efforts to work through the bureaucracy of rental applications, rectifying credit history issues, coming up with application fees and deposits, and not incidentally, overcoming discrimination are all important to getting different results.
Studies by Krysan and Crowder show that the choice set of neighborhoods is socially constrained, and affected by limited information, bias, and discrimination.  Other illuminating work on black migration shows that African-Americans who move to a different metropolitan area are much more likely to live in integrated neighborhoods than African-Americans who make moves within a metropolitan area.  One’s prior experience with a place constrains and informs the selection process. For many low income households, their only frame of reference is living in low opportunity neighborhoods, which contributes to the persistence of poverty.
What the Seattle experiment shows is that some modest and feasible efforts to provide information and services, and to lower the barriers between landlords and voucher holders reveals a dramatically different set of locational preferences on the part of low income households.

Bottom Line:  Most low income households don’t have a strong commitment to low opportunity neighborhoods. 

As the report concludes:

Low-income families tend to live in neighborhoods that offer limited prospects for upward income mobility, amplifying the persistence of poverty across generations. This paper has shown that this pattern of segregation is not driven by deep-rooted preferences among tenants or landlords. Rather, low-income families live in such areas primarily because of barriers that prevent them from moving to higher-opportunity neighborhoods – barriers that can be addressed through short-term assistance in the housing search process. These findings challenge canonical economic models of neighborhood choice in which residential sorting patterns are determined primarily by families’ preferences and call for greater modeling of the underlying structure of search costs.

The data suggest that while long term residence in an area might appear to be represent high attachment to a particular neighborhood, it is really in many cases a lack of knowledge, time and resources to become aware of other alternatives. As the authors explain, its hard to reconcile the widespread and consistent pattern of satisfaction reported by those  moving to high opportunity neighborhoods with strong attachment to their old neighborhoods.

The underlying assumption of many debates about neighborhoods and poverty is that people are tightly bound to place.  The reason we’re concerned about displacement is the sense that when people are “forced out” of a neighborhood that they have lived in and are attached to, that their networks of resources and connections are automatically and perhaps irreparably damaged.
A subsidiary argument could be made that low income households have, at least by dint of staying, chosen their neighborhoods, and despite some objectively negative effects, they feel that other factors (family, neighbors of similar ethnicity or income, or just familiarity) justify their choice.  In fact, as Chetty and his colleagues note, this assumption of “revealed preference”—that people have essentially voted with their feet and chosen the neighborhood that maximizes their well-being—underlies the canonical Rosen-Roback model of assessing urban amenities. The results of the experiment cast a great deal of doubt on that assumption.
A key takeaway from the Seattle experiment is that a very large fraction of low income residents would actually prefer not to live in low opportunity neighborhoods—provided they’re given the information and assistance to make that choice.

This has important implications for how we think about neighborhood change and migration. As we discussed last week in the context of two recent studies of gentrification, it’s often taken as a given that if a low income households moves out of a neighborhood it’s necessarily a bad thing: some equate movement with displacement. But this study (and more broadly the work of the Equality of Opportunity Project) suggests that for a significant fraction of households, moving to a better neighborhood is a path out of poverty.

A series of studies, including most recently reports  New York University, and from the Philadelphia Federal Reserve)  show that gentrification tends to produce very small increases in the rate at which low income households move out of gentrifying neighborhoods. That’s true in part because of the high rate of home-moving by low income households.  The average rental tenure in the United States is roughly two years; a wide majority  of all low income renters move over a decade or so, regardless of whether their neighborhood has gentrified or not.  Undoubtedly, as Matthew Desmond has demonstrated in Evicted, involuntary relocation is both an effect and a cause of persistent poverty. But that shouldn’t lead us to conclude that all moves represent displacement, or assume that reducing movement is always a positive in itself.

Its simply unreasonable–and unfair–to ask low income families to wait until their neighborhood gets better to find a better life for themselves and their children. For many, their best hope will be to find housing in a neighborhood that’s safer, has better schools and has less concentrated poverty and its associated ills. This experiment shows how we might move in that direction.


Peter Bergman, Raj Chetty, Stefanie DeLuca, Nathaniel Hendren, Lawrence Katz & Christopher Palmer, Creating Moves to Opportunity: Experimental Evidence on Barriers to Neighborhood Choice,  August 2019