Studying the Spatial Dimensions of Mobility
As previous speakers discussed, place matters when it comes to economic mobility. This session of the workshop featured presentations on place-based investments, and issues and challenges in studying mobility in rural and urban areas.
RESEARCH PRIORITIES FOR PLACE-BASED INVESTMENTS
Rates of economic mobility vary drastically across different regions of the United States, said Matthew Staiger (Harvard University, Opportunity Insights). Staiger shared a map from the Opportunity Atlas, which shows the income of adults who were born into low-income families (see Figure 4-1). The red- and orange-shaded areas are areas of low economic mobility, whereas blue and green represent areas of high economic mobility. For example, he said, a child who grows up in a low-income family in Charlotte, North Carolina, has an expected annual income of $26,000 at age 35. A similar child growing up in Dubuque, Iowa, is expected to make almost twice that, earning $46,000 per year at age 35. Staiger said that a better understanding of why these disparities exist may enable improved outcomes for some children.
These disparities also exist at the hyperlocal level, he said. For example, in Brooklyn, New York, Black children from low-income families who were born on the north side of Dumont Avenue have an expected income of about $17,000 at age 35, while Black children born on the south side of the street are expected to make almost $10,000 more. The key takeaway from these data, said Staiger, is that there may be scope for policies to operate at a neighborhood level, rather than a city or regional level. Based on this evidence, there are two broad strategies for increasing economic opportunity, said Staiger. The first is to reduce segregation and reduce barriers that prevent low-income families from accessing and moving to high-opportunity neighborhoods. The second is to use place-based investments to increase upward mobility in low-opportunity neighborhoods; that is, bring economic opportunity to where people are already living. Staiger focused the remainder of his presentation on place-based investments, considering three central questions:
- What are the mechanisms through which neighborhoods shape economic opportunity?
- What are the causal impacts of historical place-based investments?
- Can researchers predict long-run impacts of current interventions more quickly?
The current state of the literature has shown convincingly that where a child grows up matters greatly for long-run outcomes, said Staiger, but much less is known about why that is the case. To better understand the mechanisms that shape mobility by place, Staiger and his colleagues are gathering data on additional characteristics of neighborhoods. One project is using social media data in order to construct neighborhood-level measures of social networks; this will allow researchers to relate measures of
social network to measures of economic mobility. The second project is using data from the Social Security Administration to create measures of life expectancy at the neighborhood level, with subgroups defined by sex, race, and income.
Staiger offered his perspective on where future research in this area should go. There is a great deal to be learned, he said, by constructing additional measures of both the characteristics of neighborhoods and the outcomes of its residents. For example, area-based measures could include crime, policing, discrimination, pollution, access to credit, and substance abuse. This is an area with enormous potential for using big data from private companies alongside some of the more traditional data sources, such as administrative sources or survey data. With these new measures in hand, said Staiger, there are two approaches for better understanding the mechanisms of mobility. The first is to compare the characteristics between high- and low-opportunity neighborhoods, and the second is to look across time in a given neighborhood. For example, if the crime rates decline in a neighborhood, is this decline associated with better life prospects for the children growing up there? One way to study changes across time is to look at past policies or programs that were targeted at neighborhoods; there is “great value” in fully evaluating the impact of historical place-based investments in order to understand who was affected and how.
Current understanding of the impact of place-based policies is incomplete, Staiger said, in large part because of significant data hurdles. He shared an example of his work in order to illuminate the issues with data. A federal program called HOPE VI provided almost $7 billion in order to improve the living conditions of distressed public housing projects in the 1990s and 2000s. One specific project, he said, was a $20 million grant to the Dixie Homes Projects in Memphis, Tennessee. The neighborhood was characterized by very high poverty rates and very high crime rates, and it was isolated from the rest of the city by the layout of buildings and roads. The grant money was used to demolish old buildings and construct new buildings that would house both public housing residents and people paying market prices, with the intent of transforming the neighborhood into a mixed-income community. The program had a massive impact on the characteristics of the neighborhood, said Staiger, with the poverty rate dropping from 80 percent in 2000 to 20 percent in 2017. However, he said, what is not clear is who benefited from the program and how. To the extent that the program simply displaced existing residents into other high-poverty neighborhoods, those residents were unlikely to have benefitted.
In order to understand the impact of these types of programs at an individual level, two kinds of data are needed: big data to observe most of the people in the geographic area, and data with longitudinal linkages to follow
people if they move out of the area. In the current project, said Staiger, these data hurdles have been addressed by using a combination of tax records, administrative records, Census survey records, and private company data. Another project that is aimed at building data infrastructure is working to link historical Census data to tax records. These datasets, said Staiger, will ideally provide both the large sample sizes and longitudinal linkages that are critical for evaluating place-based policies.
In addition to evaluating historical place-based programs, future research should prioritize tracking and monitoring the impact of current policies, said Staiger. Many place-based policies are motivated in part by a desire to improve the life prospects of children growing up in the targeted area, he said, and it would be useful to be able to evaluate the efficacy of the programs “without having to wait several decades for the children to grow up.” Short-run outcomes could be identified that would allow researchers to forecast long-run outcomes. An example of this kind of work is a project that looks at older research on charter schools in order to link short-run outcomes, such as test scores, with long-run outcomes. If successful, this approach will allow researchers to use the methods developed in order to evaluate and anticipate the long-run outcomes of current policies. Building a broader set of these types of intermediate or surrogate outcomes to predict long-run outcomes, said Staiger, will be very useful for evaluating policies and programs in real time.
STUDYING MOBILITY IN URBAN AREAS
It is clear that space is directly and causally related to prospects for economic and social mobility, said Patrick Sharkey (Princeton University). He offered three statements that drill down into this overarching idea and that have implications for both understanding mobility and interventions directed at mobility:
- Spatial advantages and disadvantages are long-term multigenerational processes;
- Spatial inequality is generated by active interventions into space; and
- Mobility happens through processes that can be captured by ethnography: interactions, local social processes, turning points, shifts in mindset, and sense of identity.
Long-Term and Multigenerational Processes
Most of the literature on neighborhood effects—that is, the relationship between where young people spend their childhoods and where they end
up—examines the neighborhood environment for only one or two years. However, said Sharkey, this does not capture the full experience of spatial advantage or disadvantage. He shared a figure that shows that about half of Black Americans in the United States have lived in the poorest quarter of U.S. neighborhoods for multiple, consecutive generations (see Figure 4-2). This experience, he said, is extremely rare for White families in the United States, and is not explained by economic characteristics or anything observable in the data. Instead, the combination of social policies and the persistence of segregation has led to these severe racial inequalities in neighborhood environments. The persistence of spatial inequality by race is directly linked with prospects for economic mobility, said Sharkey. For example, among families who are exposed to disadvantaged environments, downward mobility is much more common among Black Americans than among White Americans. Sharkey identified some of the implications of the fact that mobility is linked to long-term and multigenerational processes. First, he said, places—including institutions, exposures, and opportunities—are central to the mobility process. Second, understanding the histories of places and people is essential to understanding social and economic mobility. Third, to understand inequality today, research must be prioritized that links areas, people, and families across long periods of time and across regions and countries. For more information in this area,
Sharkey recommended that interested workshop participants read historical work by Rothstein1 and Derenoncourt.2
Conversations about policy are often focused on programs that have the potential to reduce inequality and enhance mobility, said Sharkey. Often not discussed, however, are policy interventions that play a role in creating or reinforcing spatial inequality and reducing mobility. There are two dominant approaches to improving mobility for people in poor neighborhoods: helping them move to other neighborhoods or investing in disadvantaged areas. However, a third approach, said Sharkey, is to end the interventions that have been implemented over time to create and reinforce inequality. For example, redlining and yellow-lining have been found to have long-term consequences for economic mobility. Sharkey shared an example from Atlanta, where “urban renewal” policies resulted in a reshaped city, with interstate highways connecting the suburbs to the central city, suburbs zoned for single-family housing, and highways built in ways that reinforced the boundaries between Black and White neighborhoods. “It’s not an accident,” said Sharkey, that children growing up in some Atlanta suburbs will grow up to make significantly more than children growing up in neighborhoods on the other side of the river. This inequality “is the result of active intervention into urban space, with the goal of dividing space, in order to hoard resources, in order to exclude disadvantaged populations, in order to reinforce spatial inequality.”
When researchers acknowledge that spatial inequalities are generated by active interventions in the space, said Sharkey, it becomes clear that inequality is relational; the decisions made by actors on one side of a boundary affect the outcomes on the other side of the boundary. Neighborhoods are not islands, but are part of social systems of advantage and disadvantage. There is a need for policy research that focuses on active interventions to divide space and reinforce spatial stratification, such as land use regulations, occupational licensing requirements, and local housing decisions, said Sharkey. “We don’t just need programs to help people move across boundaries [but need to] know why those boundaries are there to begin with.” To develop interventions that will have a long-term capacity to reduce spatial inequality and enhance mobility, he said, there is a need to demolish the boundaries drawn by current and past interventions.
1 Rothstein, R. 2017. The Color of Law: A Forgotten History of How Our Government Segregated America. New York: Liveright.
2 Derenoncourt, E. 2022. Can you move to opportunity? Evidence from the great migration. American Economic Review, 112(2), 369-408.
The Role of Ethnography
Economic and social mobility happen through processes that occur on the ground and in real time, said Sharkey, including interactions, local social processes, turning points in lives, shifts in mindset, identity formation, and networks. These are processes that are difficult to capture in large-scale big data, he said, but can be examined through ethnography. For example, he said, research has demonstrated that community violence is a central factor in the process of economic and social mobility. In 2020, violence rose in nearly all big cities and skyrocketed in some. Sharkey said that he was frequently asked to comment on what was happening, and he said he “didn’t have a great answer.” There was a need for ethnographic studies to examine, in real-time, factors such as changes to the local social order, the way police interact with residents, and the mindset and choices of young people. Ethnography must be at the center of the effort to understand social and economic mobility, said Sharkey; he pointed to the American Voices project as an example of the type of continuous fieldwork that needs to become more common.
STUDYING MOBILITY IN RURAL AREAS
As other speakers have noted, said Daniel Lichter (Cornell University), where you live matters—and this is true in rural areas as well. Unfortunately, he said, there has not been a great deal of attention on mobility in rural areas. An urban-centric view tends to dominate, with a lot of attention on poor children in big cities. However, systemic and institutional racism is pervasive and obvious in the “invisible places” in rural America—for example, the South’s rural Black Belt, the lower Rio Grande valley, or on Indian reservations. Rural segregation and geographic isolation in these areas are a direct result of the historical legacy of slavery, racial oppression, land grabs, conflict, and violence, and the effects persist to this day. Lichter emphasized that rural and urban areas and populations are not opposite or isolated from one another, but instead are interconnected and interface in multiple ways. In order to understand mobility in urban America, he said, you have to understand what is happening in rural America and vice versa.
It is critical to include rural America to ensure a spatially inclusive approach to studying social mobility, said Lichter. Fortunately, recent signs suggest that rural areas are on America’s political and legislative policy agenda; for example, a recent Brookings Institution report on reimagining rural policy3 and an Urban Institute report on defining rural areas for hu-
man services delivery.4 He identified a number of reasons for this growing attention on rural issues. First, there has been an increased interest in understanding the “disaffected, resentful, forgotten rural voters” who voted for Trump. Second, there is growing rural-urban inequality, with growing differences in concentrated poverty, intergenerational poverty, employment, family patterns, health outcomes, and life expectancy. Lichter gave several examples of these changes in rural areas, including increases in marital and family instability, reduction in access to abortion and contraception, and growth of unemployment and underemployment. The third reason for increased attention to rural areas, said Lichter, is the “3 Ds” of depopulation, death, and diversity: the population of rural America has declined over the past decade; “deaths of despair” are increasingly common; and a majority of the population growth over the last decade is from groups other than non-Hispanic White people.
The research community, said Lichter, “must work toward mainstreaming rural-oriented work as opposed to relegating it to the backwater as unimportant or, worse, falling prey to conventional stereotypes about rural people and places.” Research on rural America needs to acknowledge and examine the heterogeneity among communities, he added, sharing an old saying in the field of rural sociology: “When you’ve seen one rural community, you’ve seen one rural community.” In addition, the boundaries that separate rural and urban America are “fluid and ambiguous,” and research should examine the ways that urban and rural areas are interconnected in the world. Lichter noted that in universities, people who study rural areas are often isolated from other researchers, and there is a difference in status and resources allocated. He emphasized the need for researchers to bridge this divide and to study the rural-urban interface.5
Despite the growing need for research in rural areas, said Lichter, there are a number of data challenges. The first overarching issue, he said, is imprecise estimates due to a lack of sufficient data. This issue is caused by a variety of factors, including small numbers of participants in research, data suppression in Census data (especially for minority populations), rural heterogeneity, difficulty with tracking and retention as respondents move, and a lack of personnel and infrastructure to deal with big data. The second major challenge, said Lichter, is rural measurement; in particular, defining the term rural. In the 1960s, the Economic Research Service developed the rural-urban continuum code to categorize areas where people lived. Today, many of these categories no longer apply or no longer predict the types
5 Lichter, D.T., and Brown, D.L. 2014. The new rural-urban interface: Lessons for higher education. Choices, Quarter 1.
of outcomes that are common in particularly rural areas. The government uses core-based units to define non-metropolitan areas, which includes small towns and the open countryside. However, the majority of rural people in the United States actually live in metropolitan areas—these rural metropolitan people create “lots of conceptual confusion.” Furthermore, the administrative boundaries that define rural and urban places change over time; this can create significant problems in longitudinal research. Researchers need to consider, said Lichter, how to define and categorize rural areas, including whether to divide and count by rural region, commuting zone, or rural neighborhood.
Lichter closed by saying that the current preoccupation with urban neighborhoods “has distracted attention from the larger question of how different dimensions of the residential context, which operate at multiple geographic and social scales, become salient in the lives of individuals and families.” It is critical, he said, to break down the rural-urban divide and consider the multiple aspects and factors that impact individual’s environment and outcomes.
Following the presentations, Mario Luis Small (Columbia University) led a discussion with speakers and workshop participants.
Small’s first question was directed at Lichter. He noted that understanding the great migration during the 20th century is pivotal to understanding today’s rural South and the urban East and Midwest. He asked Lichter to talk about the current trends in rural and urban areas that will improve understanding of what is happening in the 21st century. Lichter responded that there are several important things impacting the flow of people across the country. First, policy decisions that impact immigration and ports of entry can affect where people settle; for example, there has been a great migration to the Midwest for workers in meat packing, dairy farms, and hospitality. Second, the cost of living in many metropolitan areas is turning rural areas into a “collecting ground for America’s poor.” Both suburban and metropolitan people are moving in increasing proportions into rural areas because of lower cost of living and housing, he said. Furthermore, people who do have resources are moving into rural areas for second homes or retirement. This results in a process of rural gentrification, as the movement of urban people into an area makes it more difficult for the original rural population to stay.
Challenges for Rural Research
A second question for Lichter involved the challenges involved in studying rural areas. Small noted that researchers working in isolated, resource-deprived communities need a lot of “fortitude and commitment,” particularly for longer-term ethnographic work. He asked Lichter to discuss potential ways to improve the quality of the data collected in rural areas, whether ethnographic, survey, or other types of data. Lichter responded that while there has been a lot of leadership in this area from the university community, one major challenge is that today’s students and researchers largely have “no experience, no connection to rural America whatsoever.” This lack of first-hand experience makes it difficult for researchers to gain entrée into some rural communities. For example, said Lichter, there are very, very few sociologists or economists who have spent time in the Delta, the colonias along the Rio Grande valley, or on American Indian reservations. “These are places you don’t go if you’re not on your way to anywhere else; they are forgotten places,” he said.
Limitations of Tax Data
Small asked Staiger to comment on the limitations of using tax data to study mobility, as the Opportunity Atlas does. Small noted that research has found that tax records at the bottom of the distribution may not accurately reflect reality; since these are the populations “we most care about,” how appropriate are these data to use? Staiger responded that while using tax data has many strengths, it is not a flawless approach. One way to make progress in both understanding the limitations to the data and taking steps to improve the data, he said, is to link tax data with other data sources. For example, tax data could be linked with Census data in order to determine to what extent tax data are missing some low-income families who do not file taxes. “When you collect data on an individual from multiple data sources, you can triangulate to determine a more accurate measure, as well as the strengths and weaknesses of different data sources,” he said.
A workshop participant asked Staiger about the potential role for leveraging cell phone GPS data or other private data to understand processes related to mobility. Staiger responded that there is “enormous potential” in the data owned by private companies, particularly in terms of understanding the role of neighborhoods in shaping economic mobility. For example, one strength is the large sample size that could allow researchers to create
very local measures of characteristics of places. In addition, data such as those from Google or Facebook can be used to measure things that can typically not be captured in Census surveys or administrative records. Staiger gave one example of a way that cell phone data specifically could be used; he said that they could be used to track where people are moving in order to measure the extent to which roads and other neighborhood features are isolating communities from one another.
Mobility research is valuable for explaining how policies and trends in the past have resulted in inequalities today, said Small. However, looking prospectively is more challenging. He asked Sharkey to opine on what current trends will end up being meaningful for the mobility of today’s youth. “It is a mistake to predict the future,” responded Sharkey. However, he said, intermediate outcomes can improve understanding of what is changing and how these changes might impact mobility. For example, research has focused on making precise causal estimates about how exposure to violence affects the economic outcomes of a young person. When violence levels change (such as the rise in crime in 2020), these estimates can be used to think about how the change might impact youth’s prospects for mobility. Interventions could then be developed based on the inference that the rise in crime will have long-term effects. These interventions, he said, could be targeted directly at youth and families, or at the environment that will also be impacted by the rise in violence (e.g., businesses, schools).
Family and Spatial Processes
Another workshop participant observed a potential relationship between family processes and spatial disadvantage in mobility. She said that most individuals live near parents and adult children, but that this spatial clustering is more common among individuals who are disadvantaged. Thus, she said, the characteristics associated with lack of mobility are likely to be shared among these family members (e.g., local labor markets). Small asked Sharkey to comment on how family processes and spatial processes may intersect and reinforce one another. Sharkey agreed that there are mechanisms that can contribute to families sharing networks that can support people during difficult times and on which they can rely to stay in school or purchase a house, and so on. However, he said, there is a broader shift of rising inequality across regions, and the chances for upward mobility are now based on where a person spends the early part of their life. The implication, he said, is that variation across regions is much more important than it was a generation ago. Sharkey said that individuals from
lower-mobility areas “have to balance the advantages and the supports that come from having networks rooted in a place” with the benefits that arise from moving to places that are economically more dynamic, with rising opportunities and wages. At this moment, he said, the “balance is tilted toward the benefits of making long-distance moves to areas of opportunity.”
Why Do Neighborhoods Matter?
Although many speakers made the point that neighborhoods matter, said Small, he asked them to elaborate on why neighborhoods matter—what are the specific mechanisms that create advantage or disadvantage in mobility? Sharkey listed a number of neighborhood factors related to mobility, including violence, pollutants in water and air, and exposure to lead. However, instead of considering these factors on their own as mechanisms that lead to economic outcomes, Sharkey said he is trying to push toward a model that considers how neighborhoods relate to other neighborhoods in the community; for example, how policies and choices impact a poor neighborhood, a gated neighborhood, a neighborhood zoned for single-family residence, or a neighborhood with a highway surrounding it. Sharkey suggested that if someone doesn’t think that neighborhoods matter, they should “immediately move to the most disadvantaged, most violent community” they can find because this move should not affect their lives in a negative way and will save them a great deal of money. This tongue-in-cheek suggestion, he said, makes the point that people have a natural intuition that dimensions of one’s environment affect their later life outcomes.
The fact that neighborhoods matter is well established, said Staiger. For example, research shows that children who move to higher-opportunity neighborhoods at younger ages earn more in adulthood than children who move at older ages. While some mechanisms are established, there is ongoing, extensive research into the other mechanisms that lead to this relationship. For example, said Staiger, researchers are looking into richer measures that may impact mobility, such as social networks or friendships across class lines.
Small qualified his opening question by noting that while it is important to think about and be able to explain why neighborhoods matter, there has been a long history of work in this area. He shared two review papers6,7
6 Sharkey, P., and Faber, J.W. 2014. Where, when, why, and for whom do residential contexts matter? Moving away from the dichotomous understanding of neighborhood effects. Annual Review of Sociology, 40(1), 559-579.
7 Small, M., and Newman, K. 2001. Urban poverty after the truly disadvantaged: The rediscovery of the family, the neighborhood, and culture. Annual Review of Sociology, 27, 23-45.
that discuss this work, and noted that it is important to keep the long-term evidentiary basis in mind as new data are collected.
Recalling Lichter’s comments about the “3 Ds,” Small asked Lichter to comment further on the specific reasons why depopulation has an impact on mobility. Lichter responded that there is chronic outmigration of younger people from rural areas due to a lack of job opportunities. Often, he said, this is because the single industry in a town (e.g., timber, manufacturing) shuts down, so younger people move to find new opportunities. This leaves older people aging in place who are reluctant to spend money on schools or other youth-centered organizations. In addition, depopulation leads to community institutions, such as hospitals and supermarkets, closing down, furthering the existing disadvantages. However, Lichter noted, it is difficult to speak broadly about all rural communities; there are significant differences in the history, discrimination and oppression issues, and isolation among rural communities that range from American Indian reservations to Appalachia.
Data for Relational Studies
A workshop participant asked Sharkey to elaborate on what type of data are needed to measure and understand the relational aspects of neighborhoods, particularly contemporaneously. Sharkey responded by giving an example of his current research and explaining what kind of data are used. He and his colleagues have been working to build a dataset to examine how space is divided and how outcomes differ across these boundaries. They use data that includes information on city and town boundaries, school districts, gated communities, land zoning, and other public and private data. They have used these data to examine the issue of police violence by looking at the resources of local police departments across urban area commuting zones. Sharkey explained that communities with few resources and high levels of violence may struggle more with police violence than neighborhoods with a higher property tax base and more resources. This is one example, he said, of how data can be pulled together to document how inequality lays out across space and to identify the boundaries that form the spatial structure of inequality.
Major Approaches for Research
Small asked each speaker to briefly identify approaches for improving and expanding the type of research that is most needed in this area. Responses included:
- Rural research needs to be mainstreamed in the university. (Lichter)
- More resources need to be allocated to build datasets that allow researchers to drill into large data sets to observe how interventions impact people differently in different areas of the country. (Staiger)
- Students need rigorous training in areas including big data, data science, and ethnographic research. (Sharkey)
This page intentionally left blank.