Inner-City Poverty in the United States (1990)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

6 Ghetto Poverty and Federal Policies and Programs MICHAEL G.H. MCGEARY Other chapters in this report have reviewed what is known about the geographic and socioeconomic dimensions of concentrated urban poverty in the United States and its effects on children and adults. This chapter reviews the evidence concerning several additional questions that must be addressed in a policy analysis of concentrated urban poverty: What effects, if any, have federal policies and programs had on the concentration of poverty in central-city ghettos of the United States? What role could they have in reducing concentration or alleviating its negative effects? The United States does not have a true urban policy in the sense of a set of comprehensive federal policies and programs that explicitly try to influence the size, location, or internal spatial structure of urban settlements as such (Mills, 1987~. As a result, private market forces play the predominant role in shaping urban areas. Some policies and programs, however, such as those for urban mass transportation, are explicitly intended to increase the mobility of poor people, and many other, nonurban policies have indirect or unintended spatial effects (Barro, 1978; Glickman, 1980; Mills, 1987; Wiley et al., 1979; Vaughan, 1977a, b; Vaughan and Vogel, 1979~. Mills (1987) reviewed the evidence and found that federal policies have probably had an insignificant impact on the overall level of urbanization in the United States, compared with the effects of the private market. Mills concluded, however, that in the aggregate, federal policies and programs have influenced the shape or pattern of urbanization by reinforcing and extending private market forces that encourage the suburbanization of employment growth and of better-off residents of metropolitan areas (p. 568~. Another review concluded that federal policies have favored the development of new and growing regions over older, settled regions and 223

224 INNER-CITY POVERTY IN THE UNITED STATES thereby caused older cities in the Northeast and Midwest to experience higher unemployment and poverty rates than they otherwise would have (Glickman and Wilson, 1986:22-25~. 1b the extent that they actually have accelerated or increased the suburbanization of employment and population within urban areas, and given the income and racial barriers facing poor people who try to move out of a central city, federal policies and programs have tended to reinforce the growing concentration of the poor and minorities in central cities relative to what would have occurred in their absence. Policies and laws against segregation have enabled many better-off minorities to move out of the inner city (Jaynes and Williams, 1989:Ch.4~. Well-designed education and training programs could increase the earnings of low-skilled residents of poverty areas, enabling them to move out also. Such programs, however, are limited in size. Few are targeted enough to reach many residents of high-poverty neighborhoods. FEDERAL POLICIES AND PROGRAMS THAT PROMOTE OR FAIL TO PREVENlr INCOME SEGREGATION The research on federal housing policies and programs, policies against housing discrimination, and transportation, economic development, and welfare programs is reviewed in this section for evidence of effects on the geographic concentration of poverty. Federal policies and programs have tended to promote concentration indirectly, for example, by encouraging the suburbanization of the better-off while having little or no direct effect in the opposite direction (e.g., through increasing low-income housing in the suburbs or commuting of the poor across city lines). Housing Because housing links all households to a specific location, federal housing policies and programs could have a direct effect on the cluster- ing of low-income households in particular areas of central cities. The deductibility of mortgage interest from federal taxes, for example, has promoted the suburbanization of higher income homeowners by making more expensive housing in the suburbs relatively cheaper for them (Mills, 1987:564-565), which in turn gives further impetus to the concentration of the poor in central cities. Historically, the low-rent public housing pro- gram has contributed directly to the concentration of poverty by locating high-rise projects in poorer and minority sections of central cities and, later, by lowering income ceilings. In addition, the early practices of the Federal Housing Administration, which adopted the racially discriminately practices of private real estate and mortgage lending institutions, helped to

GHETTO POVERTY AND FEDERAL POLICIES AND PROGRAMS 225 create predominantly white suburbs and increase the concentration of mi- norities in older, less expensive housing in central cities (Lief and Goering, 1987:228-231; Orfield, 1974~. lax Treatment of Housing The federal tax code and many state tax codes have permitted home- owners to deduct mortgage interest from their income before determining their tax liability. Because the federal tax schedule is progressive, housing- related deductions are worth more to high-income owners than low-income owners. Although suburbanization has occurred in the metropolitan areas of other advanced industrialized countries, it has proceeded further in the United States, led by high-income households (Mills, 1987:563~. Making housing relatively less costly for high-income than for low-income owners adds to the incentives for high-income homeowners to live in suburbs in which they can consume expensive housing where land is cheaper, receive expensive local government services, and at the same time, use zoning and other land-use controls to avoid subsidizing low-income taxpayers. ~ the extent that federal tax policies have increased the suburbanization of high-income households relative to poor households, they have reinforced the concentration of the poor in central cities. Federal Housing Subsidy Programs The 1.3 million units in low-rent public housing projects account for a third of all publicly subsidized housing units, and most of them- 72 percent are located in central cities (Rasmussen, 1980:261~. These projects probably account for many of the census tracts that had poverty rates of 60 percent or more in 1980, and, to the extent that public housing residents are less mobile than other residents of high-poverty tracts, con- tributed to the large increase in concentrated poverty between 1970 and 1980. For example, 59 percent of the households in the census tract con- taining the Cabrini-Green high-rise public housing project in Chicago had lived in the same housing unit between 1975 and 1980, compared with 40 percent in the surrounding poverty tracts without high-rise public housing. In addition, a series of federal policy changes beginning with the Brooke amendment of 1969 lowered income limits for households in public housing. This enabled people with very low incomes to afford public housing and increased the supply of housing for such households. It also populated family projects with predominantly low-income rather than mixed-income residents. The 1.3 million units built by more recent supply-side federal housing programs (e.g., Section 236, Section 202, and Section 8 New Construction

226 INNER-CTlY POVERTY IN THE UNITED STATES and Rehabilitation) are less concentrated in central cities. Nevertheless, more than half these units are in central cities, but because they tend to be in smaller projects and more scattered, they are less concentrated within central cities. Special attempts were made to locate Section 8 New Construction in the suburbs. An evaluation using 1979 data found that, while 88 percent of the residents of central-city projects had come from the central city, 40 percent of the residents of projects built in the inner suburbs had come from the central city, which resulted in a net shift of 10 percent of house- holds in the program from the central cities to the suburbs. Participating black households moved to census tracts in which the minority share of the population was 19 percentage points less, on average. Black families, however, benefited much less than single and elderly blacks. They de- creased the minority percentage of their census tracts by only 7 points on average, from 37 to 30 percent. A strong correlation remained between minority concentration in the Section 8 New Construction projects and mi- nority concentration in the census tracts in which the projects were located (Newburger, 1987:~ble 11, using data from Wallace et al., 1981~. In recent years, federal low-income housing policy has shifted to demand-side, or voucher-type, rent supplement programs, largely because of the high per-unit cost of supply-side programs but also in the hope that providing assistance directly to low-income families would encourage mobility and reduce income and racial segregation. These programs will be more important in the future, because new construction of low-income housing with federal funds has almost ceased. The number of commitments for new supply-side rental subsidies was less than 18,000 in 1986, down from more than 180,000 a year in the late l970s (Newburger, 1987:Table fib). The effects of demand-side programs on mobility have been limited, however. For example, analyses of the effects of the federally funded Housing Assistance Supply Experiment on mobility and desegregation by income and race found that about 40 percent of the renters participating in the program in St. Joseph County, Indiana, moved in an Month period, but that was less than the 54 percent that researchers estimated would have moved without the program (Lowry, 1983:212-218~. Nearly two-thirds of the 3,600 moves were within the same group of neighborhoods and averaged less than a mile. The rest of the moves were between neighborhoods that differed in degree of segregation, but the net shift toward majority or 95 percent white neighborhoods was just 98 renters. Those moves did not noticeably affect the racial or income composition of any of the neighborhood populations, even in the group of neighborhoods in which more than one-fifth of all households were enrolled and even though 12 percent of all the renters in the county received housing allowances. The Housing Assistance Supply Experiment, also supported by the

GHETTO POVERTY AND FEDERAL POLICIES AND PROGRAMS 227 Department of Housing and Urban Development, randomly assigned par- ticipants in Phoenix and Pittsburgh to treatment and control groups, which allows a direct comparison of the effects of housing allowances on mobility and desegregation by income and race. The control groups exhibited high mobility rates; 35 percent moved in Pittsburgh and 53 percent moved in Phoenix over two years. The provision of housing allowances increased mo- bility another 5 percentage points in Pittsburgh and 10 percentage points in Phoenix (Ross), 1981:168, using data from MacMillan, 1978:57~. Although participants, on average, moved into neighborhoods that had fewer poor and minorities and better amenities, including transportation, they were no more likely to do so than control households (Ross), 1981:170~. In the Section 8 Existing Housing program, an early evaluation found that half the black households that moved went to areas with lower percent- ages of black residents and 29 percent went to areas with higher percentages (Drury et al., 1978:74~. In the absence of controls, however, the possibility cannot be ruled out that low-income movers without Section 8 housing certificates were just as likely to go to less segregated areas as the control group members were in the housing-allowance experiment. Historically, federal housing supply programs for low-income house- holds have contributed to concentrated poverty in central cities because they were more likely to be located in neighborhoods that were already poor, although recent programs have tended to be more dispersed within central cities and within most metropolitan areas. Supply-side subsidies go with the unit rather than the tenant, and the number of units is small compared with the number of eligible households. There is some evidence that tenants of subsidized housing are less likely to move, probably because they would lose a significant benefit. There is little evidence that voucher- type, demand-side programs by themselves enable recipients to move to neighborhoods less segregated by income or race (compared with the poor in private housing). Mobilitv rates are orobablv higher than in subsidized projects, however. ~, A, Residential Desegregation Title VIII of the Civil Rights Act of 1968 prohibited racial and other discrimination in the sale or rental of most housing. In 1968 the Supreme Court decided in the Jones v. Mayer case (392 U.S. 409) that the Thirteenth Amendment of the Constitution banned racial discrimination in all housing. These laws, along with favorable changes in the racial attitudes of whites (Schuman et al., 1985:Table 3.3; 1bylor et al., 1978) and higher economic status among some minorities (R. Parley, 1984), enabled many minorities to move in the 1970s to higher income tracts, which were more likely to be suburban. This outward movement of minorities lowered the high levels

228 INNER-CITY POVERTY IN THE UNITED STATES of residential segregation by only a little, however, because suburban~zing minorities tended to move to neighborhoods that were already minority or were becoming minority. And minority suburbanization resulted in in- creased income segregation within minority groups, with the poorer among them remaining in central cities. Between 1970 and 1975 there was a net out-migration of 243,000 blacks from central cities; between 1975 and 1980 it was 439,000. There was net in-migration of blacks to the suburbs of 381,000 between 1970 and 1975 and of 556,000 between 1975 and 1980 (Bureau of the Census, 1981:Table F). Most of the increase in the number of black suburbanites resulted from migration from central cities (Spain and Long, 1981~. As a result, the black proportion of the suburban population increased from 4.8 to 6.1 percent (Long and DeAre, 1981:Table 1~. Only 20 percent of the black population lived in the suburbs in 1980, however, compared with 42 percent of the nonblack population. According to the index of dissimilarity, which measures how evenly a minority group Is distributed over census tracts within a metropolitan area on a scale from 0 to 100, spatial segregation of blacks from whites decreased. In a sample of the 50 largest MSAs (metropolitan statistical areas) plus 10 MSAs with large Hispanic populations, the index went from 79.2 in 1970 to 69.4 in 1980, a drop of 12.4 percent (Massey and Denton, 1987:1bble 3~. In the 161 MSAs that had populations that were at least 4 percent black, the index went from 74 to 68, a drop of 8 percent (Wilger, 1987:6~. The index went down in all regions and in all metropolitan areas whether grouped by size, rate of population growth, size of minority population, or rate of minority in-migration (R. Farley and Wilger, 1987:Table C; Massey and Denton, 1987:1bble 4~. The absolute level of black-white segregation remained very high, however. Nearly 70 percent of blacks would have had to move to achieve complete integration in 1980 in the 60 cities studied by Massey and Denton. The indices for black-white dissimilarity were especially high in the five northeastern and midwestern metropolitan areas whose central cities experienced large increases in concentrated poverty during the 1970s. On average (unweighted for size), the index went from 84.4 to 83.4. In Chicago it fell from 91.9 to 87.8, in Philadelphia from 79.5 to 78.8, and in Detroit from 88.4 to 86.7. In New York it increased from 81.0 to 82.0, and in Newark it increased from 81.4 to 81.6. Another dimension of residential segregation is exposure, or the like- lihood that minorities will encounter a white resident or another minority resident in their home census tract. Despite the higher proportion of minorities in metropolitan areas in 1980 compared with 1970, exposure indices increased slightly. The average probability of blacks interacting

GHETTO POVERTY AND FEDERAL POLICIES AND PROGRAMS 229 with non-Hispanic whites, for example, increased from .333 in 1970 to .376 in 1980 in the 60 large metropolitan areas studied by Massey and Denton (1987:~ble 1~. Isolation, or the probability of contact with other blacks, fell from .553 to .491. In MSAs with high and fast-growing levels of concentrated poverty in their central cities, the isolation of blacks was especially high, even in 1980. In the Chicago MSA, for example, the level of black isolation was .828 (it was .855 in 1970~. The probability of contact with whites was .125, a bare improvement over 1970, when it was .118 (Massey and Denton, 1987:14~. Although the level of isolation was lower in the New York, Detroit, Philadelphia, and Newark MSAs than in Chicago in 1980, it had actually increased (the unweighted average for the four metropolitan areas was .675 in 1970, .697 in 1980), and the probability of interaction with whites had fallen (from .245 to .211) (calculated from Massey and Denton, 1987:~ble 1~. Black suburbanization also did not have much impact on the level of residential segregation in the suburbs (Wilger, 1987:15~. Segregation of blacks from whites across suburbs was nearly as great in 1980 as in 1970 (Logan and Schneider, 1984:1kble 1~. Apparently, blacks were moving to suburban neighborhoods that were either already black or in the midst of racial transition (see J. Parley, 1982, and Lake, 1981, for studies of this process in St. Louis and New Jersey, respectively). The timing of the suburban movement of blacks indicates that federal fair housing laws were a factor (McKinney and Schnare, 1986:16-17; Wilger, 1987:18-19~. Black incomes grew rapidly in the 1960s, from 55 percent of white incomes, on average, to 64 percent, but segregation levels increased overall and in neighborhoods at each income level; the proportion of blacks living in lower income census tracts fell only from 85 percent to 80 percent (McKinney and Schnare, 1986:~bles 4, 5, 64. In the 1970s the ratio of median black income to median white income fell from 61 percent to 58 percent, yet the movement of blacks out of lower income census tracts and central cities increased greatly. The proportion of housing built in an MSA after 1969 might also contribute to lower levels of residential segregation. Most new housing is built in large tracts by a single owner, which makes it easier to enforce fair housing laws. New tracts are also less likely to be racially labeled and may not deter prospective white buyers fearful of racial transition or black buyers reluctant to be the first blacks in an all-white neighborhood (Wilger, 1987:13~. Indeed, the greater the proportion of housing built after 1969 in an MSA, the greater the decrease in the index of dissimilarity between 1970 and 1980. Where the proportion of new housing was 30 per- cent or more in 1980, the index fell more than 11 percent in the 1970s. If the

230 fNNER-CllY POVERTY IN THE UNITED STATES proportion of new housing was less than 15 percent, the index fell only 1 percent. If new housing amounted to between 15 and 30 percent of the stock, the index fell between 5 and 8 percent (Wilger, 1987:%ble 3~. Although the level of residential segregation did not decrease much despite black suburbanization, the income disparities between suburban blacks and central-city blacks increased sharply in the l970s. Average household income for central-city blacks fell from 92 percent of that of suburban blacks in 1970 to 78 percent in 1980 (among whites it went from 84 to 81 percent) (Manson and Schnare, 1985:1bble II-9~. This trend occurred at a time when inequality of income distribution among blacks was increasing (Jaynes and Williams, 1989:275~. It is consistent with the finding that increases in the concentration of poverty between 1970 and 1980 were caused in large part by out-migration of better-off minorities as well as whites. Federal civil rights laws and policies helped many better-off minorities move to the suburbs. Only a few moved to predominantly white suburbs, however. Segregation levels remained high in metropolitan areas, especially for blacks. At the same time, the movement of working- and middle-class blacks out of ghetto areas into higher income neighborhoods in central cities and the suburbs, especially in the Midwest and the Northeast, increased income segregation among blacks. One result was the further concentration of poor blacks and other minorities in central-city poverty neighborhoods. Reducing the income segregation of minorities presumably would require policies and programs that increase the supply of affordable housing as well as overcome racial discrimination in higher income neighborhoods in the central cities and suburbs. Transportation In 1980 federal assistance accounted for a major portion of govern- ment spending on urban transportation: 40 percent of capital expenditures on urban highways, 80 percent of capital assistance for urban mass tran- sit, and 30 percent of mass transit operating subsidies (Gomez-Ibanez, 1985:183~. Annual federal capital expenditures (in 1984 dollars) on high- ways, which were $11 billion in 1980, had peaked in 1965 at $16 billion and had declined steadily during the 1970s as the interstate highway system was completed (Congressional Budget Office, 1985:Fig. 7~. But half of the 1,200 uncompleted miles in the interstate system are in urban areas (Congressional Budget Office, 1985:14-15~. Federal capital and operating expenditures for mass transit grew rapidly from small beginnings in 1964, increasing 40 percent annually in the 1970s to more than $2.8 billion in 1980 (Congressional Budget Office, 1985:42~. Operating subsidies began

GHETTO POVERTY AND FEDERAL POLICIES AND PROGRAMS 231 in 1974, and new subway systems were built in Atlanta, Baltimore, Miami, San Francisco, Washington, D.C., and begun in Los Angeles and San Jose. Highways Federal assistance in the construction of highways has helped to reduce the cost of commuting by automobile (Vaughan and Vogel, 1979:94-95~. A1- though the main purpose of the interstate highway system is to improve the intercity transportation system, the urban segments of the system are used most by suburbanites to travel to work in central cities and in the suburbs (Mills, 1987:563~. Beltways and radial freeways (and radial transit systems) provide much higher travel-to-work cost savings to suburban households than to central-city households. This enables the growing number of white- collar jobs that require relatively high levels of education and skills to locate in central business districts without a commensurate increase in the number of nonpoor households in the central city (Small, 1985:212-213~. Although urban highway users as a group more than pay for the capital and operating expenses of those highways through fuel taxes, rush-hour commuters to the densely settled central cities of larger metropolitan areas probably pay less than the full costs because downtown highway segments are very expensive to build and maintain and their external social costs are high (Gomez-Ibanez, 1985:199~. This encourages excessive highway use and permits workers to live farther from their place of employment than if they paid the full user charge (Mills, 1987:563; Vaughan and Vogel, 1979:9~95~. Although the net impact of federal highway programs on the spatial extent of suburbanization has been large, the impact on the proportion of the population living in suburbs may have been much less. However, it may have promoted the suburbanization of high-income households, for whom lower transportation costs facilitate paying more for housing. Mass Transit The federal government became involved in the support of urban mass transportation lo help cities maintain their financially failing mass transit systems and discourage the use of private automobiles (Hilton, 1974:3~. Mass transit has also been expected to increase the mobility of the poor (as well as the elderly and the handicapped) and to improve land use by promoting more concentrated urban development (Gomez-Ibanez. 1985:206~. r ~ Despite the large increases in federal mass transit operating subsidies beginning in the 1970s, increases in ridership gains have been modest, and the impact on automobile use has been negligible. At the same time,

232 INNER-CITY POVERTY IN THE UNITED STATES federal operating subsidies have helped the poor only marginally by lowering transit fares. Since 1975 only 24 percent of additional government subsidy dollars have gone to reduce fares; the rest supported low-ridership service expansions to new suburban markets or went to higher factor prices (e.g., higher wage rates and fuel costs) and to compensate for lower productivity within the industry (Cervero, 1985; Pickrell, 1986~. The frequency of bus service in many central cities was cut back (Pickrell, 1983~. In addition, most of the federal capital subsidies have gone to aid subways and commuter rail projects rather than bus systems, which are used most by the poor (Pucher, 1981:Table 3). Only 23 percent of federal operating assistance went to low-income users in 1983; most long-distance, peak-hour commuters were individuals from upper income groups who used the more costly subway and commuter rail services. Only 8 percent of commuter rail users were poor, for exam- ple. Accordingly, travelers from low-income households received a federal subsidy of 12 cents per transit trip compared with 20 cents a transit trip for persons from households with incomes of $50,000 or more (Charles River Associates, 1986, quoted in Urban Mass Transportation Administration, 1987:29~. Moreover, new and extended subway systems appear to have the same impact as radial freeways, although to a lesser extent (Gomez-Ibanez, 1985:209~. They stimulate high-rise office development and, therefore, employment in the downtown area, while favoring the suburbanization of higher income households (Small, 1985:213~. In 1980 only 6.4 percent of all workers nationwide used public transit (down from 9 percent in 1970), but transit use was much higher in the Northeast, where more than 14 percent of workers used it (Fulton, 1983~. More than 45 percent of the workers in the New York MSA relied on public transit in 1980, and they accounted for 28 percent of all workers using transit in the United States. Workers in the Chicago area had the next highest reliance on public transit 18 percent. Only 10 other MSAs had more than 10 percent of their workers using public transit, most of them in the Northeast or the Midwest. Fewer workers rode mass transit in 1980 than in 1970, however. The net loss nationally of metropolitan workers using transit was 487,000, but metropolitan areas in the Northeast lost 596,000 and in the Midwest 211,000. Public transit in the New York MSA alone lost 355,000 workers (17 percent). In the Chicago MSA, public transit lost 82,000 (13 percent), in Philadelphia 108,000 (28 percent), and in Detroit 61,000 workers (49 percent). Another study of public transit ridership in 1970 and 1980 in the 25 largest urbanized areas found that there was a substantial gain in the number of workers using mass transit to commute from suburban homes to central-city jobs, but it was more than offset by the loss of centrality

GHEl7O POVERIY AND FEDERAL POLICIES AND PROGRAMS TABLE 6-1 Number of Workers Using Public Transit in the 25 Largest Urbanized Areas, by Origin and Destination, 1970 and 1980 233 Type of Trip 1970 1980 Percentage (000) (000) Difference Change Central-city home to central-city job3,230.22,727.0 -503.2 -15.6 Suburban home to central-city job686.51,014.0 +327.5 +47.7 Central-city home to suburban job343.0192.6 -150.4 -43.8 Suburban home to suburban job503.8357.4 -146.4 -29.1 NOTES: Public transit includes bus/streetcar, subway/elevated rail, commuter railroad, and taxi. Workers are ~1 persons aged 16 years or older living in the 25 largest urbanized areas who use mass transit to get to work. "Urbanized areas" include the central city and only that part of the surrounding suburban area with at least 1,000 persons per square mile MSA boundaries include enure counties and thus can include rural areas. This criterion excludes the thinly settled outer fringes of metropolitan areas. It should be noted that work-related trips account for less than half of all trips on public transit. SOURCE: Calculated from Joint Center for Political Studies (1985), Tables M-70 and M-80. workers using mass transit (Joint Center for Political Studies, 1985~. There were also losses in the number of "reverse commuters," those using public transit to get from central-city residences to suburban jobs, and in the number of workers using public transit to commute within the suburbs Amble 6-1~. The precise impact of these changes in transit use patterns on central- city poor people is unknown. The reduced number of work trips within central cities and between central cities and suburbs may be accounted for by workers who moved from the central cities to the suburbs during the 1970s, but the resulting cutback in bus schedules in the central cities would have reduced the mobility of those left in the central cities. This may have been a factor in the reduction in earnings among families remaining in high-poverty tracts. According to the Federal Highway Administration's 1977-1978 Nation- wide Personal Transportation Study, households with low incomes (less than $6,000 a year) constituted 12 percent of travelers using all forms of transit but 25 percent of the users of public transit (Pucher et al., 1981~. Poor people also accounted for 27 percent of all taxi trips, more than higher income groups. However, they relied on transit for less than 7 percent of all trips (8.3 percent of all work trips) and on taxis for 0.5 percent of all trips. Poor people were much more likely to travel by car (66 percent of all trips) or by foot (23 percent).

234 INNER-CITY POVERTY IN THE UNITED STATES Noting the low use of public transit by the poor, the high cost of subsi- dizing low fares for all regardless of income, and the limits of public transit systems in providing transportation for central-city workers to suburban job locations, analysts of federal transportation policies have recommended the use of user subsidies, in conjunction with paratransit alternatives, to meet the needs of poor, elderly, and handicapped people (Altshuler, 1979; Congressional Budget Office, 1979; Kirby, 1981; Kirby et al., 1975; Meyer and Gomez-Ibanez, 1981~. User subsidies permit designated groups of in- dividuals to buy rides from a provider at reduced fares by using a voucher, token, or scrip that the provider can redeem at full-fare value from the subsidizing agency. Paratransit modes include taxis, car and van pools, and dial-a-ride services. During the late 1970s, the Urban Mass Transportation Administration funded a number of demonstrations and case study evaluations of user subsidies, usually shared-ride taxi services for elderly and handicapped people. Evaluations of these federal demonstrations and case studies of local initiatives showed that such programs were feasible, reached the most transit-dependent and reduced their travel costs, and were more flexible than fixed-route transit (Spear, 1982~. These programs, however, were not true experiments and, in the absence of control groups, it is impossible to know what the real effects were. None of the demonstrations took place in a large urban area, and only one included poor people (Santa Fe). There was also an indication that, although the unit costs of subsidizing a dial-a-ride or shared taxi trip may be lower, the increase in demand from those previously unable to afford or use conventional fixed-route transit can result in large total program costs (Echols, 1985:85-87~. Costs can be controlled by limiting per-trip subsidies or the number of trips or by restricting eligibility, but such limitations reduce the effectiveness of the program or create inequities. In summary, the federally financed freeway system has enabled those able to afford an automobile to live at a distance from their jobs in the suburbs and provided incentives to the very poor to stay in central cities, where low-cost public transit, primarily buses and taxis, is more available. Federal subsidization of fixed-rail transit, justified in part as a means of helping the poor, has probably had the primary effect of enabling well- educated, highly skilled suburbanites to reach white-collar service jobs that are concentrating in central cities. Radial transit systems, whether subways or highways, are not well suited for reverse commuting from the central city to low-skilled employment in the suburbs. Meanwhile, federal operating subsidies have tended to go to expanding suburban bus service, while bus service within central cities has been reduced.

GHETTO POVERTY AND FEDERAL POLICIES AND PROGRAMS 235 Economic Developments Historically, federal infrastructure programs (e.g., highways; sewage and water treatment facilities) and tax policies (e.g., investment tax cred- its that favor construction of new plant and facilities over rehabilitation; tax exemption of industrial development bonds) have reinforced private market forces that have promoted the suburbanization of jobs and higher income people (Vaughan and Vogel, 1979~. At the same time, economic development programs aimed at revitalizing central business districts and blighted neighborhoods have not had much of a countervailing impact (Mills, 1987:565-566). Cost-effective public actions to foster economic activity in particular locations are difficult to develop and implement, and public intervention in the economic decisions of firms has uncertain effects and tends to be costly. Federal strategies usually focus on providing preferential tax treatment for firms that locate in specified places or hire workers with particular characteristics. Policies toward urban enterprise zones, for example, try to encourage businesses to locate in depressed areas and to hire unemployed or underemployed workers by providing tax subsidies to the employing organization. A voluminous literature on the economics of firm location, however, finds nearly universally that taxes are too small a component of firm costs or even of firm profits to have much impact on location decisions (Leonard, 1986:Ch. 5~. Other variables labor costs, site acquisition costs, transportation costs~tend to exhibit substantially more variation across sites than tax variables can (Advisory Commission on Intergovernmental Relations, 1981; Carlton, 1983; Kieschnick, 1981; Schmenner, 1982~. Tax relief has relatively little capacity to compensate for higher labor costs, higher construction or site acquisition costs, higher transport costs, or higher energy costs. Firms that can be influenced to locate in or near areas of concentrated poverty will tend to bring jobs that cannot break the cycle of concentrated poverty. Any firm whose location decisions are sensitive to modest dif- ferences in tax benefits is one that faces intense competition. Such firms tend to be unreliable because they can easily be influenced to change loca- tions again in response to foreign wage reductions, fluctuations in foreign exchange rates, and a host of other factors beyond the control of their host communities. Also, the jobs they can offer a community are often in low-skill and low-paying manufacturing activities that offer little training or prospect for advancement or growth and little in the way of benefits. Full-time, year-round employment in many of these jobs will not get a poor family above the poverty line. Willis section is based on material prepared by Herman B. Leonard.

236 INNER-CITY POYERIY IN THE UNITED STATES Even if firms can be induced to locate in areas of concentrated poverty, they may not hire the most disadvantaged residents. Even in areas of concentrated poverty, many of the residents are not below the poverty line. Many are employed, some in skilled occupations. Creating new opportunities for appropriate employment closer to where they live may result in shifts in their employment location rather than an increase in employment among those not currently employed. The jobs that are needed moderate-skill, light manufacturing in grow- ~ng firms with reasonably assured future prospects are the hardest to iden- tify and nurture. High-skill jobs are beyond the reach of most who live in areas of concentrated poverty; those who have such skills in occupations for which firms have open slots have, for the most part, found their way into good jobs and out of concentrated poverty. There are moderate-skill jobs (custom light manufacturing jobs, like printing, rebuilding car parts, and assembling scaffolding) that can be mastered relatively quickly but that require some skill, provide some training, and that can pay reasonably high wages. Firms doing that kind of work however, tend to be small, hard to identify, and difficult to assess. They have a high rate of business failure. Stying to specify programs whose subsidies could be targeted on that subset of these firms that can be expected to be successful and to provide long-term, stable employment would almost surely result in subsi- dies flowing to many firms that will come to depend on public support for their competitive edge and their continued existence, or alternatively, the subsidies may go to support activities that would have taken place anyway. Such substitution effects have been found in federal economic development programs (Congressional Budget Office, 1983:47~. In short, the prospects for using public intervention to create economic development in areas of concentrated poverty as a means of increasing the employment of the very poor are not promising, although the strategy may work in certain situations. Welfare Income maintenance transfers, which are mostly federally financed (including Aid to Families With Dependent Children [AFDC], Supplemen- tal Security Income [SSI], food stamps, local general assistance, energy assistance, refugee assistance, and earned-income tax credits) have been a growing component of income in the United States. Income transfers increased from .92 percent of total personal income in 1969 to 1.28 percent of total personal income in 1984 (Garnick, 1988:Tables 12, 14~. In the same period, in the central counties of MSAs, which contain the central cities, income maintenance transfers increased from 1.09 percent to 1.54 percent of personal income.

GHETTO POVERTY AND FEDERAL POLICIES AND PROGRAMS 237 Between 1970 and 1980 in the 50 largest central cities, the number of families with public assistance income who were living in high-poverty areas increased from 129,568 to 309,117. More important, the proportion of families with public assistance income who were living in high-poverty areas increased from 30.2 percent to 42.5 percent, and the proportion of families with earnings income fell from 74.4 percent to 62.6 percent, which indicates a greater reliance on welfare during the time period among families living in high-poverty areas (calculated from Bureau of the Census, 1973, 1985~. Welfare benefits for similar recipients differ from state to state, even though welfare programs are funded mostly with federal dollars, because federal welfare policies permit a high degree of local discretion in setting payment levels and eligibility requirements. In 1985, for example, the maximum AFDC benefit that a family of four could receive in California was $698 a month, nearly five times the maximum benefit the same family could receive in Mississippi. The addition of food stamps reduces the interstate variation by about half, because food stamp benefit levels are nationally uniform, but the combined benefits for a family of four in California are still double those in Mississippi (P. Peterson and Rom, 1988:27). The influence of these interstate differences in benefit levels on the migration patterns of the poor has been a much-studied subject. Using a variety of data sets, variables, and statistical techniques, most early studies did not find a significant effect of high benefit levels on in-migration to states or metropolitan areas (for a review of the leading studies, see P. Peterson and Rom, 1987:22-27). The few that did find migration effects had methodological shortcomings. Researchers also discovered that black interstate migrants to big cities, including migrants from the South, were less likely to be poor or on welfare than blacks born in the states in which the cities were located Tong, 1974, using 1970 census data). Several studies of New York found that most recipients who had migrated to New York City had not applied for welfare until several years after arrival (DeFerranti et al., 1974; Ostow and Dutka, 1975). Studies of decisions to migrate have found that the primary determi- nants of migration patterns are differences in employment opportunities and wage rates, although other factors, such as local spending and tax- ing policies, weather conditions, pollution levels, and differences in other amenities, are also significant. Differences in unemployment and wage rates, however, were larger factors in explaining in-migration than out- migration. In-migration, for example, is stimulated by low unemployment rates, but out-migration is insensitive to locally high unemployment rates (see review of research in Vaughan and Vogel, 1979:28-31), even though the unemployed move more than any other group (Bureau of the Census, 1981:3).

238 INNER-CITY POVERTY IN THE UNITED STATES In 1969 the Supreme Court ruled in Shapmo v. Thompson (394 U.S. 618) that states could not impose duration-of-residence requirements (typ- ically one year) on applicants for AFDC and that new state residents are, therefore, immediately eligible for AFDC (assuming they meet the other eligibility requirements). Recent studies that have focused on trends since the Supreme Court decision have found that differential welfare benefits have statistically significant migration effects (Blank, 1988; Gramlich and Laren, 1984; P. Peterson and Rom, 1987~. The interstate differences in benefits for recipients in similar circumstances can be quite large, even taking into account differential wage and tax rates (Blank, 1985~. The migration effect, while significant, is small but could result in substantial cumulative effects over time. Few AFDC beneficiaries move to another state over a five-year period, for example, but when they do move they are much more likely to go to a state with higher AFDC benefits (Gramlich and Laren, 1984:505-506~. The studies also found, however, that policymakers in high-benefit states reduce benefit levels in the face of in-migration by the poor (Gramlich and Laren, 1984; P. Peterson and Rom, 1987~. Assuming that differences in state welfare benefits have a significant long-term effect on the distribution of poor people, they could have in- creased the concentration of the poor in central cities if those moving in order to take advantage of higher benefits had a propensity to settle in central cities. There are some published data providing evidence that this occurs. According to a Census Bureau report on geographic mobility in the United States, for example, 9.1 percent of the 3.3 million households on welfare in 1979 had moved to a central city from somewhere else be- tween 1975 and 1980 (a few from other central cities), compared with 4.1 percent of the nonwelfare households (calculated from Bureau of the Cen- sus, 1981:Table 35~. (And, as discussed below, welfare households already living in central cities were much less likely to move out than nonwelfare households.) Unfortunately, the published data do not permit a similar comparison of poor and nonpoor households or of where households that were poor in 1975 ended up in 1980 in comparison with nonpoor house- holds. The published data also do not show the flows of public assistance recipients in and out of high-poverty tracts within central cities. The main factors in the growth of concentrated poverty during the 1970s were (1) out-migration of the nonpoor at a higher rate than the poor from tracts that were or became high-poverty tracts and (2) increasing poverty among existing residents of those tracts. It is, however, impossible to tell from the net migration data in the published reports on high-poverty areas how much the in-migration of poor people from outside the central city was replacing some of the poor people who left concentrated poverty tracts, thereby keeping concentration levels higher than they would have

GHETTO POVERTY AND FEDERAL POLICIES AND PROGRAAiS 239 been and contributing to the higher proportion of welfare recipients in high-poverty areas in 1980 than in 1970. The final possibility is that differences in welfare benefits and the availability of other programs benefiting poor people in central-city poverty areas slowed their out-migration from high-poverty areas. One of the early studies of welfare effects on migration, for example, found that receipt of public assistance reduced the tendency of poor families to leave a central city and that the low rate of out-migration contributed to the growing concentration of welfare recipients within central cities (Long and Heltman, 1976, cited in Vaughan and Vogel, 1979:32~. A recent analysis of data from the Census Bureau's Survey of Income and Program Participation (Clark, 1988) found that welfare recipients, especially AFDC recipients, are significantly less likely to move to another state. That study controlled for characteristics common to welfare recipients that also act to inhibit out- migration (racial minority status, having young children, less education, receiving public housing assistance, and not being in the labor force). It also controlled for characteristics that tend to increase mobility (youth and female-headed family structure). Census data on geographic mobility between 1975 and 1980 show that 65 percent of the households receiving public assistance in 1979 had moved, compared with only 51 percent of nonwelfare households. The central-city households receiving public assistance, however, were much more likely to move within the same central city (47 percent compared with 23 percent of nonwelfare households) and much less likely to move out of the metropolitan area. After five years, 82 percent of the public-assistance households remained in the same city, 5 percent moved to the central city of another MSA, and 13 percent moved outside the central city to a suburb or nonmetropolitan area. Among households not receiving public assistance in 1979, the comparable figures were 72 percent, 5 percent, and 24 percent (calculated from Bureau of the Census, 1981:Table 35~. Finally, a number of federally subsidized programs and services for poor people are relatively more available in central cities. The historical legacy of low-income, high-rise public housing and its role in concentrating poverty in poor, minority neighborhoods of central cities, and the tendency of recipients of housing allowances or vouchers to make only short moves to areas similar in racial composition and income, have already been discussed. Bus and taxi services, which are disproportionately used by poor people, are also most available in central cities. Other social and health services subsidized with federal funds are disproportionately located in central cities. During the 1970s most of the federal funding for community health centers went to centers in central cities, and about half the federal funding for maternal and child health (MCH) services was earmarked for special projects in urban areas for maternity and infant care and for

240 INNER-CITY POVERTY IN THE UNITED STATES child and youth services (replaced after 1981 by MCH block grants) (G. Peterson et al., 1986:Ch. 5~. Many community mental health centers are also located in central cities, as are many community action agencies that administer community senice and low-income energy assistance programs (G. Peterson et al., 1986:Ch. 3, 4~. State and local general assistance and medical assistance programs also vary in some states and tend to be both more available and more generous in central cities (Burke, 1987:63 66~. Services and programs for poor people are usually located in central- city poverty areas because that is where the people who need them are. But a number of studies using data from the 1960s found that high expenditures for local public services were associated with higher in-migration and lower out-migration, especially among poor and minority households. Wealthy white households tended to leave or avoid high-expenditure localities, ap- parently because of the high tax rates usually associated with high public expenditures (Cebula, 1974, 1979; Greenwood and Sweetland, 1972; Pack, 1973~. In summary, whether or not the greater availability of welfare, social, health, and other services for poor people in central cities relative to suburbs or rural areas played a role in attracting poor in-migrants during the 1970s, it may have played an inadvertent role in slowing the out-migration of the poor (see Kasarda, 1985, 1988~. This effect, if it existed at all, was probably very small, not least because most programs reached small proportions of those in poverty. Some of the services, such as education and training and other human-capital investment programs, which are reviewed in the next section, may have enabled more people to escape dependence and migrate out than the number of people that such programs may have "anchored" in central cities. FEDERAL PROGRAMS THAT COULD REDUCE POVERTY CONCENTRATION Some federal programs that are or could be targeted at central-city poverty areas have the potential of increasing the long-range ability of poor people to increase their incomes and move into nonpoverty areas. Such programs include education and employment training programs. Education Family income has a positive association with the academic achieve- ment of children and their subsequent earnings (Sewell and Hauser, 1975~. A recent study of the effects of parental characteristics on the educational achievement of children and their economic success when they grow up

GHETTO POVERTY AND FEDERAL POLICIES AND PROGRAMS 241 (using longitudinal data from the Panel Study of Income Dynamics con- ducted by the University of Michigan's Survey Research Center PSID) confirms the positive but modest effect of family income on achievement and income. In addition, a spell of family poverty has an independent and strong negative impact on the achievement, earnings, and income of sons (Corcoran et al., 1987:Table 4~. The same study, as well as an earlier one using PSID data (Datcher, 1982), also found that neighborhood character- istics, such as high unemployment rates, high welfare participation rates, and high proportion of female-headed families, as measured at the zip code level, have a negative effect on educational attainment and income (Chapter 4 in this volume discusses these studies). Given that the poor and minorities are concentrated in central cities, that public schools are locally controlled by each community in a metropoli- tan area, and that federal policies do not require the transportation of poor students to nonpoor schools across community boundaries, it is not surpris- ing that central-city schools have a disproportionate share of poor students. The federal government funds two major education programs aimed at counteracting the educational disadvantage of poor children and of low- achievers in poor schools: Head Start and Chapter 1 of the Education Consolidation and Improvement Act (formerly Title 1 of the Elementary and Secondary Education Act of 1965~. Head Start, which began with the Economic Opportunity Act of 1964, is a preschool program that provides year-round, comprehensive education and other child development services (medical, dental, nutritional, social) to children from poor families or families that would qualify for AFDC if they did not have child care assistance. The program costs about $1.2 billion a year in federal funds and serves about 460,000 three- to five-year-olds a year (but only a fifth of them are in full-day programs) (U.S. Congress, 1985:341~. Although more than 90 percent of the participants are below the poverty line, the program reaches just 20 percent of all poor preschool children in a given year. The program is operated by local school systems and nonprofit sponsors; it is not aimed at cities per se, and an unknown percentage of the funding goes to programs located in central cities. A number of research studies have found that preschool programs for disadvantaged children, some of them Head Start programs, can have immediate effects on cognitive test scores compared with control groups, but these effects disappear after several years. They also can achieve longer term gains on other indicators of school success (the Head Start studies are synthesized in Aitken et al., 1985; studies of other preschool programs in Lazar and Darlington, 1978; Lazar et al., 1982~. Head Start children are, for example, less likely to be placed in special education programs or to repeat a grade. A follow-up study of one of the best-known non-Head Start experiments, the Perry Preschool Project, found that poor children in the

242 INNER-CITY POVERTY IN THE UNITED STATES project had higher employment rates and lower rates of criminal behavior when they were 19 years old than individuals in the control group (but the sample sizes were very small) (Barrueta-Clement et al., 1984~. The Chapter 1 education program serves almost 5 million children in elementary and secondary schools, at a cost of $4.3 billion in 1988. Chapter 1 is intended to support special compensatory services for low-achieving students in schools with high concentrations of poor children. The focus of the program is consistent with research evidence that there is a strong association between family poverty and average achievement in a school and only a weak association with individual student achievement (Kennedy et al., 1986a:3-4, citing a review of studies of the relationship between socioeconomic status and academic achievement by White, 1982~. Chapter 1 funds are distributed in proportion to the number of poor children in a school district. The district, in turn, provides special services to children, poor or nonpoor, of low educational achievement in schools with above- average concentrations of poverty. The services provided are primarily basic skills instruction in reading and mathematics for elementary school students (Birman et al., 1987: Ch. 3~. Schools with as few as 10 low-income students are eligible for Chapter 1 services, and more than 90 percent of the school districts in the United States participate. Nevertheless, because funding is proportional to the number of poor students in a school district, the poorest quartile of districts contains 45 percent of the students receiving Chapter 1 services. Central- city districts, which have 26 percent of all students, have 37 percent of the Chapter 1 students. Still, 13 percent of schools with more than half their students in poverty offer no Chapter 1 services, and some schools with Chapter 1 services have very few poor students (Birman et al., 1987:Ch. 2~. In a single year, about 11 percent of school-aged children are served by Chapter 1, but, with a 40 percent turnover rate annually, an estimated 25 percent of all public school students receive services at some point (Kennedy et al., 1986b:7~. There are no recent data on the proportion of Chapter 1 recipients who are poor. Academic achievement, at least as measured by standardized tests, dropped substantially from the mid-1960s until the mid-to-late 1970s (Con- gressional Budget Office, 1986:Ch. 3~. In the subsequent upturn in test scores, minority students, students in schools with a high proportion of mi- norities, and students in disadvantaged urban areas made greater relative gains than other students (Congressional Budget Office, 1986:Ch. 4~. Some researchers have attributed the gains to federal funding because the pattern of increases in achievement scores paralleled patterns of increased federal financial support (LaPointe, 1984; Riddle, 1984~. Only one evaluation, using data from 1976-1979, followed a set of students over time, including comparable students who did not receive Chapter 1 services (there are

GHEITO POVERTY AND FEDERAL POLICIES AND PROGRAMS 243 no evaluations of the program's impact using a randomly selected control group). The study found that Chapter 1 students, especially in the early grades, gained relative to comparable students, but not enough to bring them close to the achievement levels of advantaged students. A more recent cross-sectional study without a comparison group found more substantial gains relative to advantaged students, but again the gains were not enough to bring Chapter 1 students close to the level of the average student (both studies are reanalyzed and their methodological shortcomings discussed in Kennedy et al., 1986b:Ch. 3~. A careful evaluation of the research evidence by the Congressional Budget Office (1987:94-98) concluded that Chapter 1 services could have contributed only a small amount to the upturn in test scores of minorities and of students in the early grades. In summary, federally subsidized education programs aimed at poor students and poor schools can improve the educational achievement and attainment and, therefore, the future employment and income prospects of poor children, including those in poor neighborhoods. Any impact on the concentration of poverty would be long term, however, and the magnitude of that impact would depend on how large and well targeted the programs were on poor families in areas of concentrated poverty and on how well designed and well delivered the programs were. As already noted, Head Start reaches about 20 percent of poor children aged three to five years; Chapter 1 reaches about 25 percent of all school-aged children at some point. The proportion of Chapter 1 students that are poor is unknown, but it is known that some schools with concentrations of poor children do not participate and that nonpoor children participate in other schools. The formulas for distributing program funds do not try to target the funding geographically but spread the program widely. The variation among individual Head Start and Chapter 1 projects in their effects is considerable, and their average impact would be increased appreciably if less effective programs performed as well as more effective programs (Kennedy et al., 1986a:Ch. 5~. Training Federal involvement in employment and training programs began in the Great Depression. The federal-state Employment Service has long existed to address the short-term, or frictional, unemployment of those temporarily between jobs. Many other federal employment and training programs were created to reduce cyclical unemployment caused by downturns in the business cycle. The most pertinent programs for residents of concentrated poverty areas, however, are those intended to reduce the long-term, or structural, unemployment of low-skilled workers (Bass) and Ashenfelter, 1986~.

244 INNER-CITY POVERTY IN THE UNITED STATES Despite the long history of federal involvement in training programs, few programs have been subjected to experimental analysis using randomly assigned control groups. Some of those that have, however, demonstrate significant positive and cost-effective results, including the National Sup- ported Work Demonstration, the state welfare-to-work experiments inspired by the supported work experiment, and the Job Corps. The National Supported Work Demonstration ran from 1975 through 1979 in 15 sites. It subsidized highly structured and closely supervised work experiences that gradually became more demanding until they approximated regular private employment. The program had the most positive impact on the long-term employment and earnings of enrollees from two target groups, long-term AFDC recipients and former drug addicts (there was no postprogram impact on the two other target groups, former offenders or young school dropouts). The AFDC recipients averaged about nine months in the program. Eighteen months later, 42 percent of the enrollees were employed, compared with 35 percent of the control group members. Enrollees worked 62 hours a month on average, compared with 46 hours among control group members, and they averaged $248 a month in earnings, about $81 more than the control group average (Hollister et al., 1984:~ble 4.6~. Similar results were obtained by the former drug users JIollister et al., 1984:Table 5.5~. The Supported Work Demonstration was also cost-effective for the AFDC recipients and former drug users. In the long run, the costs of the program were far outweighed by the higher income and reduced de- pendence on welfare payments of the AFDC recipients and by the higher income and reduced criminal activity of the former drug users (Hollister et al., 1984:Table 8.6~. The Manpower Demonstration Research Corporation, which was cre- ated to run and evaluate the Supported Work Demonstration, has sub- sequently conducted experimental evaluations of a number of state-level programs aimed at increasing the employment of AFDC recipients. Those programs, which rely on relatively inexpensive interventions (e.g., job clubs), have proven in most cases to have a significant and cost-effective impact, although the impact is typically modest (Gueron, 1987~. The Job Corps, which was established in 1964, is a residential program for school dropouts between ages 14 and 21 that provides them with basic education, vocational skills, and health care for an average of 30 months per participant. About 90 percent of the enrollees are from poor or welfare- dependent households, more than 75 percent are minorities, and 30 percent are female (Betsey et al., 1985:110-116, describe the program and assess evaluations of it). An evaluation in 1982 by Mathematica Policy Research found that participants in the Job Corps did better than comparison group members along several dimensions that can be attributed to the program

GHEl7O POVERTY AND FEDERAL POLICIES AND PROGRAMS 245 (Mallar et al., 1982~. After three and a half years, for example, Job Corps enrollees were employed an average of 26 weeks a year, compared with 23 weeks among nonparticipants; they earned an average of $2,592 (1977 dollars), compared with $2,025 among nonparticipants; and they had less income from welfare or unemployment. Moreover, they experienced better health and were arrested less often for serious crimes. After an initial six-month period after completing the program, when enrollees did worse than the comparison group members, these outcomes emerged and persisted during the rest of the four-year follow-up (Betsey et al., 1985:112, summarizing Mallar et al., 1982~. Like educational programs, carefully designed and targeted training programs have the potential of increasing the career earnings of low-skilled, disadvantaged residents of concentrated poverty areas and enabling them to move to nonpoor neighborhoods. The most successful of the programs (Job Corps, Supported Work), however, are very expensive and pay for themselves only in the long run. The state welfare demonstration programs are cheaper, but their impact is much more modest. CONCLUSION Historically, federal policies and programs have had five main if un- intentional effects on the spatial distribution of poverty. First, federal programs have favored the suburbanization of higher income people rela- tive to lower income people, thereby reinforcing the concentration of poor people, many of them minorities, in central cities. Second, federal actions have encouraged the development of new areas in the South and the West relative to the older, developed metropolitan areas in the Midwest and the Northeast. The economies of the latter regions have declined in relative and, in some cases, absolute terms; as a result, poverty has increased in the central cities of those regions and it has become even more difficult for poor people there to escape poverty through work. Third, some federal programs, such as high-rise public housing projects, have had the direct effect of concentrating poverty, especially after changes in the income eligi- bility rules drastically reduced the average income levels in public housing after 1969. Fourth, after 1968, fair housing laws helped nonpoor minorities to leave ghetto areas, which contributed to the dramatic increase in concen- trated poverty among central-city minorities during the 1970s. Fifth, some federal policies intended to increase the mobility of poor families, such as housing vouchers and mass transit subsidies, have not had the expected ef- fect of reducing residential or income segregation. Other programs, such as AFDC, may have the unintended effect of discouraging mobility. There is evidence, however, that programs that improve the education and training skills of the poor can increase the career earnings of low-skilled residents

246 INNER-C1IY POVERTY IN THE UNITED STATES of high-povert~r areas and thus enable them to move to nonpoor neigh- borhoods, although these programs are limited in size and are not well targeted geographically to high-povertr ghettos. REFERENCES Advisory Commission on Intergovernmental Relations 1981 Regional Growth: Interstate Tax Competition. Report No. A-76. Washington, D.C.: Advisory Commission on Intergovernmental Relations. Aitken, Sherne S., Ruth Hubell McKey, LarTy Condelli, Hamet Ganson, Barbara J. Barrett, Catherine McConkey, and Margaret C Plantz 19B5 The Impact of Head Start on Children, Families and Communities: Viral Report of the Head Start Evaluation, Synthesis and Utilization Project. Report prepared by CSR, Inc., for the Head Start Bureau. Washington, D.C: U.S. Department of Health and Human Services. Altshuler, Alan A., with James P. Womack and John R. Pucher 1979 The Urban Transportation System Politics awl Pony Innovation. Cambridge, Mass.: MIT Press. Barro, Stephen M. 1978 The Urban Impacts of Federal Policies Vot 3, Fiscal Conditions. Report No. R-2114-KF/HEW. Santa Monica, Calif.: Rand Corporation. Barrueta-Clement, John R., Lawrence H. Schweinhart, W. Steven Barnett, Ann S. Epstein, and David P. Weikart 1984 Changed r ices: The Effects of the Peny Preschool Program on Youths Through AN l9. Ypsilanti, Mich.: High/Scope Press. Bassi, Laurie J., and Orley Ashenfelter 1986 The effect of direct job creation and training programs on low-skilled workers. Pp. 133-151 in Fighting Poverty: Mat Works and What Doesn't, Sheldon H. Danziger and Daniel H. Weinberg, eds. Cambridge, Mass.: Harvard University Press. Betsey, Charles L^, Robinson G. Hollister, Jr., and Mary R. Papageorgiou, eds. 1985 Youth Employment and Training Programs: The YEDPA Years. Committee on Youth Employment Programs, National Research Council. Washington, D.C.: National Academy Press. Birman, Beatrice F., Martin E. Orland, Richard K Jung, Ronald J. Anson, Gilbert N. Garcia, Mary T. Moore, Janie E. Funkhouser, Donna Ruane Morrison, Brenda J. Ibmbull, and Elizabeth R. Reisner 1987 The Current Operas of the Chapter 1 Program. Washington, D.C.: U.S. Department of Education. Blank, Rebecca M. 1985 The impact of state economic differentials on household welfare and labor force behavior. Journal of Public Economics 28(Fall):25-58. The effect of welfare and wage levels on the location decisions of female- headed households. Journal of Urban Economics 24:186 211. Bureau of the Census 1973 Low-Income Areas in Large Clues. 1970 Census of Population, Subject Report PC(23-9B. Washington, D.C.: U.S. Department of Commerce. 1981 Geographical Mobility: March 1975 to March 1980. Current Population Reports, Series P-20, No. 368. Washington, D.C: U.S. Department of Commerce. 1988

GHElTO POVE~lY AND FEDERAL POLICIES AND PROGRAMS 247 1985 Poverty Areas in Large Cities. 1980 Census of Population, Subject Report PC80-2 8D. Washington, D.C.: U.S. Department of Commerce. Burke, Vee 1987 Cash and Noncash Benefits for Persons with Limited Incomes: Eligibility Rules, Recipient and Expenditure Data, FY 1984-86. Report No. 87-759 EPW. Washington, D.C.: Congressional Research Senace. Carlton, Dennis 1983 The location and employment choices of new firms. Review of Econorrucs and Statistics 65(August):440 449. Cebula, Richard J. 1974 Local government policies and migration: An analysis for SMSAs in the United States, 1965-1970. Public Choice l9(Fall):85-93. The Determinants of Human Migration. Lexington, Mass.: Lexington Books. 1979 Cervero, Robert 1985 Else anatomy of transit operating deficits. Urban Law and Policy 6(Janua~y): 477497. Charles River Associates 1986 Allocation of Federal Transit Operating Subsidies to Riders by Income Group. Draft research report submitted to the Urban Mass Transportation Administration. Clark, Rebecca 1988 Outmigration Among Welfare Recipients. Paper presented at conference on Individuals and Families in Transition: Understanding Change through Longitudinal Data, sponsored by the Social Science Research Council, Annapolis, Md., March 16-18. Congressional Budget Office 1979 Urban Transportation for Handicapped Persons: Alternative FederalApproaches. Washington, D.C.: U.S. Government Printing Office. The Federal Govemment in a Federal System: Current Intergovemmental Programs and Options for Change. Washington, D.C.: U.S. Government Printing Office. The Federal IRudget for Public Works Infrastructure. Washington, D.C.: U.S. Government Printing Office. Trends in Educational Achievement. Washington, D.C: U.S. Government Printing Office. Educational Achievement: Explanations and Implications of Recent Trends. Washington, D.C.: U.S. Government Printing Office. Corcoran, Mar,, Roger H. Gordon, Deborah Laren, and Gary Solon 1987 Intergenerational Transmission of Education, Income, and Earnings. Institute of Policy Studies, University of Michigan. 1983 1985 1986 1987 Datcher, Linda 1982 Edects of community and family background on achievement. 17'e Review of Economics and Statistics 64(Februaty):3241. DeFerranti, David M., et al. 1974 The Welfare and Nonwelfare Poor in New York City. Report No. R-1381-NYC. New York: New York City-Rand Institute. DruIy, M., O. Lee, M. Springer, and Lo Yap 1978 Nationwide Evaluation of He Exalting Housing Program. Washington, D.C.: U.S. Department of Housing and Urban Development.

248 INNER-CITY POVERTY IN THE UNITED STATES Echols, James C. 1985 Use of private companies to provide public transportation seances in Tidewater Virginia. Pp. 79-100 in Urban Transit: The Private Challenge to Public Transportation, Charles A. Lave, ed. San Francisco: Pacific Institute for Public Polipy Research. Farley, John E. 1982 Metropolitan housing segregation in 1980: The St. Louis case. Urban Affairs Quarterly 18(March):347-359. Farley, Reynolds 1984 Blacks and Whites: Narrowing the Gap? University Press. Cambridge, Mass.: Harvard Farley, Reynolds, and Robert J. W'lger 1987 Recent Changes in the Residential Segregation of Blacks From Whites: An Analysis of 203 Metropolises. Background report prepared for the Committee on the Status of Black Americans, National Research Council, Washington, D.C. Fulton, Philip N. 1983 Public transportation: Solving the commuting problem? Pp. 1-9 in Trans- p~tation Research Record 928. National Research Council. Washington, D.C.: Transportation Research Board. Garnick, Daniel H. 1988 Local area economic growth patterns: A comparison of the 198~; and previous decades. Pp. 199-254 in Urban Change and Poverty, Michael G. H. McGeary and Laurence E. Lynn, Jr., eds. Committee on National Urban Policy, National Research Council. Washington, D.C.: National Academy Press. Glickman, Norman J., ed. 1980 The Urban Impacts of Federal Policies. Baltimore: Johns Hopkins University Press. Glickman, Norman J., and Robert H. Wilson 1986 National contexts for urban economic development. Pp. 15-36 in Local Economies in Transition: Policy Realities and Development Potentials, Edward M. Bergman, ed. Chapel Hill: University of North Carolina Press. Gomez-Ibanez, Jose ~ 1985 1be federal role in urban transportation. Pp. 183-Z3 in American Domestic Priorities: An Economic Appraisal. John M. Quigley and Daniel L Rubinfeld, eds. Berkeley: University of California Press. Gramlich, Edward M., and Deborah S. Laren 1984 Migration and income redistribution responsibilities. Animal of Human Resources 19~0ctober):489-511. Greenwood, Michael J., and D. Sweetland 1972 The determinants of migration between standard metropolitan statistical areas. Demography 9(November):665~92. Gueron, Judith M. 1987 Refonn~g Welfare With Work. Occasional Paper 2, Ford Foundation Project on Social Welfare and the American Future. New York: Ford Foundation. Hilton, George W. 1~74 Federal Transit Subsidies: The Urban Mass Transportation Assistance Program. Washington, D.C.: American Enterprise Institute.

~9 Hollisler, an C., Peter Emma and ~ A. away, ads. 1~ ~s ~~' ~ ~ Ha. Radian: Un-miV of ~s ~~ PI J.-=, wild D., and Rabin a. Williams, a, eds. 1~9 ~ ~ _ Ha =d _ _ ~mmill~ on me Sulus of Blab Sedans, atonal away Hunch. ~shinglon, D.C.: National academy Pan. Joint caller ~ ~1 Studio 1~S Demographic Mange and -~ldp Mel Lands. ~1. II, Subsli=1 labl=. ~n prepaid for the Organ Baa Scion ~minislebon under Coma a~emeD1 DC-7. loins caller for halide Sludi=, bin~lon, D.C ~a, John D. 1~S Urban Range and minored op~=nili=. Pp. 3~7 in ~e ^- Paul E. Pele~n, ed. Dashingly, D.C: Beings Inshtubon. low, ml:mlion, and emerging urban misally. Pp. 1~-198 in - ~ ~ ~ H~ ~ ~ ~ ~ Em-= ~ -~ ~ eds. ~mmillee on rational Urban Polity National Mob Council. ~sbinglon, D.C.: Tonal Idea Pax. Penney, Baa hi., Rival ~ lung and Satin E. Hand 1~ ~ ~' ~d ~s ~~ ~- ~ S~. ~shingloni D.Cx U.S. Depanmen1 of Edumbon. Penney, Baa ~ Beald~ E Bidden, and Rang E. Romaine lamb ~e ~ ~ ~ amp. ~sbinglon, D.C.: U.S. Depanmenl of Mullion. Aft, ~ 1~1 ~ =d ~: ~ ~ =d ^~ a. ~sb ~g~n, D.C: ^~1 of Sale P~nn~g Andy. Ha, Ronald F 1~1 Meg monk e~cli=~: Offside lens~nalio; Untidies. ~' ~ ^~ ~nle0:~2. Ha, Ronald a, Can u. Bball, Snivel A. amp, Bong O. ~cOill~, and Satin Hi 1~3 ~-~ _ ~# -= Amp. `~inglon, D.C.: awn InsUlum. an, ~~ ~ 1~1 ~s _ S"~: ~= =d ~g ~ ~ Amp. Alley Un-=i9. Nag B~n~ck ha.: caller far Urban Policy fib. Spoil, Tie E. 1S4 He ~ nags afoul Sedan edu=[ion. ^' ~63Qune):~- . Irving, and Ritual B. Dading[on 1~8 Hag ~ Ha. Upon of 1be ~n=nium far ~ngiludinal Studio. ~shinglon, D.C: U.S. Depa~enl of H=kb, Edu~lion, and -1~. Irving, Richa~ Dadinglon, Hay Auk, ]a~uil1De a_, and An Sniper 1~2 ~sl1ng ewes of =~ edu~lion. ~ ~ ~e -- ~ ^~ air, ~1. a, As. 2-3, Seder ha. 193. ~ing10n, D.C. S=ieV at R=eamb in Child O~elopmen[. 1~

250 INNER-CITY POVERTY IN THE UNITED STATES Leonard, Herman B. 1986 Checks Unbalanced. The (quiet Side of Public Shindig. New York: Basic Books. Lief, Beth J., and Susan Goering 1987 The implementation of the federal mandate for fair housing. Pp. 227-267 in Divided Neighborhoods: Changing Pattems of Racial Seg~ega~n. Gary A. Tobin, ed. Vol. 32, Urban Affairs Annual Reviews. Newbu~y Park, Calif.: Sage Publications. Logan, John R., and Mark Schneider 1984 Racial segregation and racial change in American suburbs, 197~1980. American Joumal of Sociology 89:874-888. Long, Larry H. 1974 Poverty status and receipt of welfare among migrants and nonmigrants in large cities. American Sociological Review 39(Februa~y):4~56. Long, Larry H., and Diane DeAre 1981 The suburbanization of blacks. American Demographics 3 (September):17-21. Long, Larry H., and Lynne R. Heltman 1976 Do Welfare Payments Reduce Migration Potential? Paper presented at the Annual Meeting of the American Sociological Association, New York, August. Lowry, Ira S., ed. 1983 E~nenting with Hmsing Allowances: The Final Report of He Housing Assistance Supply E~irnent. Cambridge, Mass.: Oelgeschlager, Gunn & Hain. MacMillan, Jean 1978 Mobility in the Housing Allowance Demand Experiment. Draft report. Abt Associates, Cambridge, Mass. Mallar, C., S. Kerachsky, C. Thornton, and D. Long 1982 Evaluation of the Impact of the Job Corps Program, Third Follow-up Report. Mathematica Polipy Research, Princeton, NJ. Manson, Donald M., and Ann B. Schnare 1985 Change in the City-Suburb Income Gap, 197~1980. Project Report. Wash- ington, D.C: Urban Institute. Massey, Douglas S., and Nancy A. Denton 1987 1tends in residential segregation of blacks, Hispanics, and Asians: 197~1980. American Sociological Review 52(Dec3:802~25. McKinney, Scott, and Ann B. Schnare 1986 Trends in Residential Segregation by Race: 19601980. Project Report. Washington, D.C.: Urban Institute. Meyer, John R., and Jose A. Gomez-Ibanez 1981 Autos, Transit, and Cities. Cambridge, Mass.: Harvard University Press. Mills, Edwin S. 1987 Non-urban policies as urban policies. Urban Studies 24(December):561-569. Newburger, Harriet B. 1987 The Impact of Federal Housing Programs on Black Americans. Background report prepared for the Committee on the Status of Black Americans, National Research Council, Washington, D.C. Orfield, Gary 1974 Federal policy, local power and metropolitan segregation. Poetical Science Quarterly 89(Winter):777-785.

GHETTO POVERTY AND FEDERAL POLICIES AND PROGRAMS 251 Ostow, Miriam, and Anna B. Dutka 1975 Work and Welfare in New York City. Baltimore: Johns Hopkins University Press. Pack, Janet Rothenberg 1973 Determinants of migration to central cities. Joumal of Regional Science 13(August):249-260. Peterson, George E., Randall R. Bovbjerg, Barbara A. Davis, Walter G. Davis, Eugene C. Durman, and Theresa ~ Gullo 1986 The Reagan Block Grants: What Have We Leamed? Washington, D.C.: Urban Institute. Peterson, Paul E., and Mark C. Rom 1987 Federalism and Welfare Reform: The Determinants of Interstate Differences in Poverty Rates and Benefit Levels. Paper presented at the Annual Meeting of the American Political Science Association, Chicago, September 3~, The case for a national welfare standard. Brookin£s Review 6(W~nter1:2~32. 1988 Pickrell, Don H. 1983 The Causes of Rising Transit Operating Deficits. Report MA-11~037. John F. Kennedy School of Government, Harvard University, Cambridge, Mass. 1986 Federal operating assistance for urban mass transit Assessing a decade of experience. Pp. 1-10 in Transportation Research Board 1078. National Research Council. Washington, D.C.: lianportation Research Board. Pucher, John 1981 Equity in transit finance: Distribution of transit subsidy benefits and costs among income classes. Joumal of the American Planning Association 47:387 407. Pucher, John, Chris Hendrickson, and Sue McNeil 1981 Socioeconomic characteristics of transit riders: Some recent evidence. Tragic Quarterly 35(July):461-483. Rasmussen, David SU 1980 The urban impacts of the Section 8 Existing Housing assistance program. Pp. 24~263 in The Urban Impacts of Federal Policies, Norman J. Glickman, ed. Baltimore: Johns Hopkins University Press. Riddle, W. 1984 Achievement Score Trends and Federal Invah~ement in Elementary and Sec ondary Education: An Exploration of Their Relationships. Report No. 84~27 EPW. Washington, D.C.: Congressional Research Service. Rossi, Peter H. 1981 Residentialmobility. Pp.147-172in DoHousingABowances Work? Katharine Lo BradbuIy and Anthony Downs, eds. Washington, D.C.: Brookings Insti- tution. Sewell, William H., and Robert M. Hauser 1975 Education, Occupation and Faming. New York: Academic Press. Small, Kenneth A. 1985 Transportation and urban change. Pp. 197-223 in The New Urban Reality, Paul E. Peterson, ed. Washington, D.C: Brookings Institution. Schmenner, Roger W. 1982 MakingIRusu~ess Location Decisions. Englewood Cliffs, NJ.: Prentice-Hall. Schuman, Howard, Charlotte Steeh, and Lawrence Bobo 1985 Racial Attitudes nz America Trends and Interpretations. Cambridge, Mass.: Harvard University Press.

252 INNER-CTIY POVERTY IN THE UNITED STATES Spain, Daphne, and Lany H. Long 1981 Black Movers to the Suburbs: Are They Moving to Predominantly White Suburbs? Special Demographic Analysis. Washington, D.C.: Bureau of the Census. Spear, Bruce D. 1982 User-side subsidies: Delivering special-needs transportation through private provided. Pp. 13-18 in Transportation Research Record 850. National Research Council. Washington, D.C.: Transportation Research Board. Taylor, D. Garth, Paul B. Sheatsley, and Andrew M. Greeley 1978 Attitudes toward racial integration. Scientific American 238(June):42-49. Tolley, George, Philip E. Graves, and John L" Gardner 1979 Urban Growth Policy in a Market Economy. New York: Academic Press. Urban Mass Transportation Administration 1987 The Status of the Nation's Local Mass Transportation: Performance and Con d~tions. Report to Congress lay the Secretary of Transportation. Washington, D.C.: U.S. Department of Transportation. U.S. Congress 1985 Children in Poverty. Prepared for the Committee on Ways and Means by the Congressional Research Service and the Congressional Budget Office. Ways and Means Committee Print 99-8. 99th Cong., 1st sess. Washington, D.C.: U.S. Government Printing Office. Vaughan, Roger J. 1977a Me Urban Impacts of Federal Policies. Vol. 1, Overview. Report No. R-2206-KF/HEW. Santa Monica, Calif.: Rand Corporation. 1977b The Urban Impacts of Federal Policies. Vol. 2, Economic Development. Report No. R-2028-KF/RC. Santa Monica, Calif.: Rand Corporation. Vaughan, Roger J., and Mary E. Vogel 1979 The Urban Impacts of Federal Policies. Vol. 4, Population and Residen- tial Location. Report No. R-2205-KF/HEW. Santa Monica, Calif.: Rand Corporation. Wallace, James, Jr., S. Bloom, William Lo Holshouser, S. Mansfield, and Daniel H. Weinberg 1981 Participation and Benefits in the Urban Section 8 Program: New Construction and Existing Housing. Cambridge, Mass.: Abt Associates. White, K.R. 1982 The relationship between socioeconomic status and academic achievement. Psychological Bulletin 91(May):461~81. Wllger, Robert J. 1987 Black-White Residential Segregation in 1980: Have the Civil Rights Laws Made a Difference? Unpublished paper, Population Studies Center, Uni- versity of Michigan.