Segregation and Neighborhood Growth

A common misconception about segregation in America is that segregated neighborhoods are dense hubs where large numbers of people are clustered -- the overcrowded inner cities of the popular imagination.

The reality of segregation is very different. Segregated areas often do not teem with people, and are rarely becoming more crowded. Instead, most are emptying out. This process, which can be traced through decades of neighborhood change, has left terrible scars on the social and economic fabric of cities and their suburbs.

For a poor or moderate-income neighborhood, there are few long-run trends more alarming than a decline in density. Falling density and population typically results in declining tax base, lowered school attendance, and underfunded municipal government. In places that are poor, out-movers tend to be wealthier than those that stay, concentrating poverty.

And falling density in a neighborhood spells trouble for the surrounding areas, too. Because the overall population is growing in most regions, reduced density generally results in sprawl elsewhere.

In the city centers where segregation is often the most pronounced, population declines disrupt the tendency of a region's economy and infrastructure to cluster around a geographic center. The result is the displacement of jobs and development to the urban fringe, which not only taxes financial and environmental resources but places economic opportunity out of reach for many. City centers are economic engines for entire regions; when the centers begin to decline, the engine begins to rust or break down entirely.

Segregation helps drive an unsustainable feedback loop of neighborhood decline. When segregated neighborhoods empty out, schools and public services take a hit. The community becomes less and less attractive to existing residents and potential newcomers, leading to further population loss. The feedback loop is particularly acute in suburban cities: when a disproportionately large share of city revenues are drawn from residential property tax, the negative impacts of residential flight are amplified.

Integrated neighborhoods, by contrast, not only break this cycle, but may be able to reverse it. Integrated neighborhoods are the fastest growing of all -- in the aggregate, faster-growing than even predominantly white areas, which are often a region's most affluent. (This is illustrated in the table at the bottom of this post.) By contrast, most integrated places are neither wealthy nor impoverished -- they are, instead, middle-class boom towns where the dream of economic advancement for the working class seems to still be alive.

Longitudinal census data lets us see the deleterious effects of segregation on a large scale, across many different places. Areas that were segregated in the past have consistently become less densely populated. In this regard, they differ noticeably from more integrated places.

This is illustrated in the graph below, which displays data from approximately 33,000 census tracts in the 50 largest American cities. The horizontal axis shows a tract's racial concentration in 1980 while the vertical axis shows change in density in that tract over the next twenty years. The dividing line is at the 100 percent mark -- any tract above this point grew more dense, while any below emptied out.

Despite the huge number of data points, a racial pattern is very clear: at high levels of segregation, declines in density become far more common. In neighborhoods where segregation nears 100 percent, very sharp declines in density become much more likely than not.  Meanwhile, very few neighborhoods in this range experience any sort of rapid growth.

This trend, suggested by the graph, is confirmed by the numbers. In the least white 10 percent of tracts, population fell 11 percent and two-thirds of tracts shrank. By contrast, in the whitest 90 percent of tracts, population increased 34 percent, with two-thirds of tracts growing. This is illustrated below.

Correlation, of course, does not imply causation, and there are a variety of confounding factors at play. Most notably, segregation is highly correlated with poverty -- at times so much so that economic and racial concentration are nearly impossible to disentangle.

However, the sharp downtick in neighborhood outcomes at the most segregated end of the spectrum makes the effect of race hard to ignore. It is unlikely that this cluster of census tracts experiencing sharp decline can be explained by reference to poverty alone.

The relationship between racial composition and density is further explored in the table below, which uses the same data. Of "hypersegregated" tracts where more than 90 percent of the population was nonwhite in 1980, only 5.4 percent saw rapid growth. Even in less severely segregated tracts where between 10 and 40 percent of the population was white, 12.1 percent saw rapid growth -- and over three times as many saw a decline in density.

Although predominantly white areas did not suffer the rapid decline, they still faced a degree of stagnation. Nearly one in three became less dense between 1980 and 2000, with fewer grew rapidly.

But there was a marked increase in population density in one kind of neighborhood: integrated places where 20 to 60 percent of the population was nonwhite in 1980. Over a third of these places grew rapidly -- in fact, rapid growth was more likely than any degree of decline.

For many cities and neighborhoods, segregation has been the first step on along a road to ruin. But these places can be restored, and others protected, by the proactive and thoughtful embrace of housing integration.