Comments on the Spatial Dimensions of Urban Inequality

A fundamental question is whether the spatial arrangements of the different dimensions of urban inequality merely reflect larger non-spatial causal forces or are in and of themselves a driving force in creating or reinforcing those inequalities.

Certainly, the ways in which urban inequalities show up in spatial arrangements do reflect (or are a spatial expression of) fundamental inequalities. For example, given that household incomes, assets, and disposable incomes are unequally distributed, and given that housing is largely distributed through the private market in cities, and finally given that people with similar tastes or wants are likely to cluster together, then it follows that there will be a fair degree of segregation across income classes and that neighborhoods will be ranked in a market “pecking order,” with those with the most income to spend on housing dominating the most desirable neighborhoods, ranging down to those with the least income clustering in the least desirable areas.

This basic market sorting may be reinforced by practices of incumbent owners, other participants in the residential real estate market (brokers, bankers, etc.), and even legal mechanisms (zoning, covenants, etc.) that keep lower income or otherwise less appreciated households out of more exclusive areas.

The spatial patterns of this form of income (and also related clustering by race, ethnicity, lifestyle, stage in the life cycle, and so on) vary a lot across urban settings. For most of the last 60 years, suburban neighborhoods rapidly eclipsed central city neighborhoods in attracting the top earning households in the U.S., but in the last 20 or so, central city locations have increased their leverage. In Europe, central city neighborhoods more often retained their top-tier positions and social housing was placed in inner suburbs, such as the banlieux of Paris or the so-called Western garden suburbs of Amsterdam. Similarly, the array of high and low income neighborhoods within metropolitan areas varies across U.S. metropolitan areas, often not following a “concentric circles” pattern. Some have many of their high income areas on the periphery and low income neighborhoods in the center (think Detroit), others retain high income areas in the center (San Francisco or Manhattan), and others are more mixed.

In other words, the spatial patternings do not only reflect a rent or housing cost gradient, but the accumulated or embedded results of various other decisions about the location of industrial areas, railways, highways, noxious facilities, institutions of various sorts, and social housing, many of which are political or governmental as opposed to strictly economic. Land tenure tends to be “sticky” and has perhaps become more difficult to alter whole scale through urban renewal or eminent domain, though market change can have an acidic upward or downward influence on prevailing forms of tenure.

But what about the feedback loop? Once in place, what influences do specific forms of spatial inequality exert on the life quality and life chances of the people who live in different parts of the unequal spaces?

A lot of social scientific water has flowed under the “neighborhood effects” bridge. The best studies suggest that place effects account for one-tenth or so of the main effects flowing from individual and family characteristics. Critics of those studies, however, such as Robert Sampson, argue that each of the individual controls contain place effects, so partialling them leads to their under-estimation.

To apply the Brooklyn Cab Driver test, it is clear that people spend a lot of time trying to locate themselves in the “right” place and to defend that place against threats to its relative attraction. Much of this is tied to the spatial fixity of primary education and the desire of better off and more educated families to get their kids into the “best” primary schools. (It is said that being in the PS321 school catchment area in Park Slope is worth $200,000 on a $1.5 million property.) They also want proximity to all kinds of amenities, such as convenient travel to work, shopping, restaurants, cultural activities, and sometimes religious institutions. And they organize to prevent or keep out land uses they think are inconsistent with their neighborhood quality of life. So in their behavior people seem to suggest that place effects or context are quite important to them and their family members. (At the other end of the scale, people clearly try to leave dangerous low income neighborhoods.)

Even if we lean toward believing that “place matters” (as I clearly do), we still have to answer the questions “when, why, and how”? Do all spatial patterns of inequality have the same positive and negative effects? If those effects vary across the different spatial patterns of inequality (for example, splotchy patterns of poverty might be less harmful than concentrated patterns of poverty), what are the mechanisms generating and transmitting the effects? And what are the magnitudes of variation along the different parts of the inequality distribution (i.e. are the effects more extreme at the ends than in the middle)? (For thinking about the possible mechanisms by which these effects could be transmitted, see Joseph, Chaskin, and Webber 2006, Galster 2010, and Sampson 2012).

The idea that income segregation has negative effects on the bottom and unfairly concentrates advantage on the top is so compelling that many nations are experimenting with various forms of “mixed income housing” development. This can come about by tearing down some social housing and building less deeply subsidized or even market rate housing in its place (Hope VI in the U.S. or the mixed income programs in the Netherlands), by creating inclusionary housing (requiring developers to create some affordable units through zoning or other incentives), by building “scattered site” social housing in high income areas, by providing people with vouchers to move there, and many other ways.

At present, it does not seem like these forms of deliberate mixing yield large positive effects on members of poor families, for example low income families moving out of high poverty areas (Moving to Opportunity) or that lower income people living in the affordable units in market rate developments interact much with the higher income residents. At the same time, they doubtless benefit by being able to exit bad situations, especially if they can avoid going into another bad situation. But none of these experiments have been designed with rigorous control groups (even Moving to Opportunity, which had this aim, did not always get people to move away from bad neighborhoods and stay away from them).

Given that we are not likely to have new rigorous experimental situations to tease out the variety of effects that might flow from different types of income desegregation, it might be worthwhile to compare otherwise similar urban fabrics where the past history of social housing has led to a spatially less concentrated pattern of neighborhood inequality (such as Amsterdam) with those where the market has more effectively sorted people (such as Newark).   This would be especially valuable if we could take repeated cross sections of individual and family outcomes and neighborhood contexts over time and analyze their relationships; better yet would be place-conscious panel studies, but they do not exist.

How should we move toward a better understanding of how different spatial forms of inequality affect the quality of urban life for individuals, families, neighborhoods, and cities? A first step might be to develop a rigorous inventory of the different spatial forms and shapes or patterns of income inequality, keeping an eye on scale – how much is within (or across) blocks, within neighborhoods, or within districts of the metropolitan regions? A second step might be to inventory the institutions, processes, and policy decisions that seem to have framed those patterns, from the ways in which residential real estate markets work to governmental decisions about basic infrastructure, land use, and housing tenure, not excluding the cultural dimensions of how and why certain groups tend to cluster together. Final steps might then be to tease out causal patterns and consequences for individuals, groups, and cities.


Galster, George C. (2010). “The Mechanism(s) of Neighborhood Effects: Theory, Evidence, and Policy Implications,” (Paper for presentation at the ESRC Seminar “Neighbourhood Effects: Theory & Evidence” St. Andrews University, Scotland, UK 4-5 February).

Joseph, Mark, R. Chaskin and H Webber (2006). “The theoretical basis for addressing poverty through mixed income development, Urban Affairs Review 42 (3):369-409.

Sampson, Robert J. (2012) Great American City: Chicago and the Enduring Neighborhood Effect (Chicago: University of Chicago Press).

About the Author
John Mollenkopf is a Distinguished Professor at the Graduate Center of the City University of New York.
Posted on July 23rd, 2014.