Poverty, crime, poor educational performance, and loss of the middle class are problems typically associated with central cities, particularly large central cities.¹ But suburbs can experience these problems as well. Urban location theory asserts that households will arrange themselves around a centralized workplace, with the lowest income households living at the highest densities nearest the workplace center.² If central-city boundaries do not expand as metropolitan populations grow, one would expect the poverty rate to rise in central cities and for poverty to eventually spill over into near-in suburbs. With higher poverty levels, one might expect suburbs to see increased crime and other problems. Older suburbs -- those with buildings and infrastructures built for a less mobile and efficient economy -- could also be expected to undergo economic and social decline. Recent discussions of urban problems have noted that some suburbs face the same problems and share the same needs as central cities.

This article reports modest attempts by the U.S. Department of Housing and Urban Development (HUD) to determine the extent to which urban problems have spilled over to suburbs. The simple tests described here find clear examples of suburbs with poverty characteristics similar to those of central cities, but it is easier to find suburbs with high poverty rates than it is to find concentrated poverty in suburban areas.

Poverty Levels

Poverty statistics are not available between censuses for individual central cities or suburbs, but the U.S. Census Bureau does report annual poverty rates for all central cities combined and for all suburbs combined. In 1996, the most recent year for which data are available, the central-city poverty rate was more than twice that of suburbs. In central cities, 19.6 percent of persons lived in poverty, compared with 9.4 percent in suburban areas. Overall, 13.7 percent of U.S. persons were poor, and in nonmetropolitan areas, the poverty rate was 15.9 percent.

To look at poverty rates for individual suburbs, one has to use information from the 1990 census. Employing a file combining poverty data from the 1990 census with 1996 population estimates and organized by current definitions of metropolitan statistical areas, HUD examined poverty rates for all suburban "places." A "place" can be either an independent jurisdiction or an area recognized locally as a distinct community.³ For example, in Virginia, where local government functions are performed either by counties or by independent cities, the Census Bureau classifies as "places" both independent cities, such as Falls Church, and unincorporated areas with local names inside counties, such as Annandale in Fairfax County. Table 1 reports the results of this search for the 50 metropolitan areas with the largest suburban populations.

In Table 1, the first column contains the names of the metropolitan areas, which have been sorted by the size of their suburban populations in 1996.⁴ The second column provides the Census Bureau's estimate of the 1996 suburban population. The third column is the proportion of the suburban population, as of 1996, living in places where the poverty rate in 1989 exceeded the U.S. average poverty rate at that time, 13.1 percent. Note that the third column does not report the percentage of the total suburban population living in poverty or even the percentage of the population in the identified places that is poor. Instead, it tells what proportion of the suburban population lives in places where poverty is more pervasive than the national average. For convenience, places above the national average are termed "poor" places. Looking at the prevalence of poor places provides a good measure of the extent to which high poverty levels are an important problem in suburbs.

Table 1 shows substantial variation in the prevalence of poor places across the large metropolitan areas. For the 50 largest metropolitan areas, the proportion of the suburban population living in places with above-average poverty ranges from 0.0 percent in the suburban areas around Milwaukee-Waukesha to 33.0 percent in the suburbs of Los Angeles-Long Beach. Seven of the 50 metropolitan areas have 15 percent or more of their suburban population living in poor places. In addition to Los Angeles-Long Beach, these metropolitan areas are Riverside-San Bernardino (25.2 percent), Miami (21.4 percent), Pittsburgh (19.6 percent), Phoenix-Mesa (17.1 percent), Bergen-Passaic (15.8 percent), and New Orleans (15 percent). Eleven have 10 percent or more, but less than 15 percent, living in poor places, but 6 of these 11 are clustered at just above 10 percent. The percentage living in poor places is 5 percent or less in 19 of the 50 metropolitan areas.

Table 2 lists the 50 metropolitan areas with the highest proportion of their suburban populations living in poor places. Dothan, a small metropolitan area in the southeast corner of Alabama, leads the list with 68.2 percent of the 77,001 people living in its suburbs living in poor places. Three large metropolitan areas from Table 1 -- Los Angeles- Long Beach (20th), Riverside-San Bernardino (41st), and Miami (50th) -- are also on this table. Most of the rest of the metropolitan areas listed are southern or border areas. Because the poverty threshold is not adjusted for the cost of living, the prevalence of southern and border areas is not surprising. Many of these places have small suburban populations; 24 have less than 100,000 and 8 have less than 50,000.

It is important to note that in this analysis, suburban means noncentral city. Many metropolitan areas have more than one central city. For example, in the Washington Metropolitan Statistical Area, Arlington, Frederick, and Fredericksburg are central cities in addition to Washington, D.C. The poverty situation in these central cities does not show on these lists. To get a better sense of the spatial distribution of poor places and to see whether excluding noncore central cities affects the results, HUD mapped poor places, including central cities, for 24 of the 32 metropolitan areas with a suburban population of more than 1 million.

Maps 1 through 4 provide examples of where poor places are located relative to central cities. The maps distinguish between places with poverty rates above the 1989 national average and places with poverty rates above the 1989 average for central cities, 18.0 percent. Map 1, Los Angeles-Long Beach, shows a spillover pattern in which most of the poor suburbs border Los Angeles. Long Beach and Pasadena show up with an above-average poverty rate as well. Map 2, Atlanta, presents a mixture of spillover poverty, more distant poverty, and rural poverty, with some of the poorest suburbs located 20 miles or more from the Atlanta city limits. In Map 3, Cleveland-Lorain-Elyria, there are one spillover suburb, East Cleveland, around Cleveland and 8 poor places at some distance from that central city. Note that both of the other central cities, Lorain and Elyria, are classified as poor. They are near each other but over 10 miles from Cleveland. In Map 4, Minneapolis-St. Paul has one poor inner suburb southeast of St. Paul and a few remote poor places more than 20 miles from the central city.

In general the Maps indicate that ignoring the noncore central cities does not have much of an impact. For example, none of the three noncore central cities in the Washington, D.C. metropolitan areas had poverty rates above the national average in 1989. The Maps also show that sometimes the poor suburbs are actually at the rural edge of the metropolitan areas.

High-Poverty Tracts

Table 3 presents the results of this effort for the 50 metropolitan areas with the largest suburban populations.⁶ The first column contains the names of the metropolitan areas, which have been sorted by the size of the suburban population in 1996.⁷ The second column shows the percentage of the 1990 population in the largest central city that lived in very poor census tracts. Note that three metropolitan areas -- Nassau-Suffolk, Bergen-Passaic, and Middlesex-Somerset-Hunterdon -- have no central cities. The third column is the proportion of the population in the other central cities that lived in very poor tracts. Note that 35 of these metropolitan areas have more than 1 central city. The fourth column provides the same information for the suburban portion of these metropolitan areas. The final column contains the proportion of the entire metropolitan area population living in very poor tracts as of 1990.

For the 50 metropolitan areas combined, 11.4 percent of the population in the largest central cities lived in very poor census tracts in 1990. For the other central cities, the proportion fell to 4.4 percent. The suburban percentage is only 0.5 percent. In 11 of the 50 metropolitan areas, the percentage of the suburban population living in very poor census tracts in 1990 rounded out to 0.0 percent. In only five of these metropolitan areas did more than 2.0 percent of the suburban population live in very poor tracts, with the suburbs of Phoenix-Mesa having the highest percentage, 3.9 percent.

The standouts among second central cities are in metropolitan areas with large bordering central cities: Camden (Philadelphia) and Miami Beach (Miami). Other second central cities with high concentrations of poverty are East St. Louis (St. Louis, MO-IL); Muskegon and Holland (Grand Rapids, MI); Spartanburg and Anderson (Greenville, SC); and Dearborn, Pontiac, and Port Huron (Detroit).

Conclusions

At the same time, HUD found little evidence of concentrated poverty -- census tracts with poverty rates of 40 percent or more -- in the suburban parts of the metropolitan areas with the largest suburban populations. Except for a few exceptional cases, concentrated poverty was also less frequent an occurrence in secondary central cities, that is, central cities other than the largest central city in metropolitan areas with multiple central cities.

This analysis is exploratory only and is not intended to provide definitive answers on the extent to which urban problems have spread to suburbs. Different variables, more recent data, and more refined techniques need to be applied to this question.

Notes

1. See The State of the Cities, U.S. Department of Housing and Urban Development, June 1997, for a discussion of the problems of large urban areas.

2. Edwin S. Mills and Luan Sende Lubuele, "Inner Cities," in Journal of Economic Literature, June 1997.

3. "Places, for the reporting of decennial census data, include census designated places and incorporated places. . . . Census designated places (CDPs) are delineated for the decennial census as the statistical counterparts of incorporated places. CDPs comprise densely settled concentrations of population that are identifiable by name, but are not legally incorporated places." From pages A-9 and A-10 of Technical Documentation 1990 Census of Population and Housing.

4. In consolidated metropolitan statistical areas, HUD looked at the component metropolitan statistical areas separately.

5. Paul A. Jargowsky and Mary Jo Bane, "Ghetto Poverty: Basic Questions," in Inter-City Poverty in the United States, Laurence E. Lynn, Jr., and Michael G. H. McGeary, Editors, National Academy Press, Washington, DC, 1990; and John D. Kasarda, "Inter-City Concentrated Poverty and Neighborhood Distress: 1970 to 1990," in Housing Policy Debate, Vol. 4, No. 3 (1993).

6. The ranking is by suburban population in 1996 and, therefore, the ranking is identical to that in Table 1. However, the numerator and denominator for the ratios in the remaining columns are based on 1990 population counts because the Census Bureau does not provide intercensal population estimates at the census tract level.

7. Once again in consolidated metropolitan statistical areas, HUD looked at the component metropolitan statistical areas separately. In most cases where there are multiple central cities, the largest is the one listed first in the name of the metropolitan area.