Strategies for Reducing Disparities
This chapter reviews a number of strategies to address the unequal spatial opportunity structure of metropolitan areas and the disparities that result, at least in part, from it. It sets forth a range of options that have been suggested and, in many cases, tried in at least a few areas. It discusses the rationale for the strategies and relevant arguments on both sides of the issue. If the strategy has been put into effect, we present available evidence on its effects.
Changing the Spatial Distribution of Population
Since large numbers of municipal governments have already been established in most U.S. metropolitan areas, policy options for reducing the sorting process that facilitates stratification by race and income have focused primarily on land use controls. One set of strategies seeks to break down exclusionary zoning practices that help to increase the cost of housing beyond the means of low- and moderate-income households. Another set of strategies is inclusionary, seeking to require local governments or developers to include housing for low-and moderate-income households in the community. Some approaches contain elements of both strategies.
Ending Exclusionary Practices
Ending exclusionary practices inevitably means placing limits on local government's control of decision making about land use. Existing local government control over land use could be eliminated, reduced, or circumscribed in a
variety of ways. These include statewide control of local land use decisions and statewide review of local land use decisions to determine their consistency with state guidelines; metropolitan-area-wide control of land use decisions within metropolitan-area boundaries or metropolitan review of local land use decisions; and state regulation of local land use, in particular the prohibition of specific exclusionary zoning practices. Another possibility would give developers the right to sue local jurisdictions that restrict their ability to provide low- and moderate-income housing, as in the Massachusetts Anti-Snob Zoning Law.
Zoning as a mechanism supports the system of small and fragmented local government that facilitates income segregation by jurisdiction (Mills and Oates, 1975). To the extent that such a system is associated with allocative efficiency by permitting the various local governments to respond to the relatively homogeneous preferences of their residents, as Tiebout (1956) and his followers argue, then reducing exclusionary zoning practices will adversely affect efficiency by promoting heterogeneity of preferences among the citizenry. The possibility of a trade-off between achieving optimal levels of allocative efficiency and reducing inequality of opportunity in metropolitan areas must be acknowledged.
State Land Use Plans
Several states—Oregon, Florida, New Jersey, California, and Vermont, among others—have adopted state land use planning laws. Oregon's structure is probably the most elaborate and strongest in terms of the limits it places on traditional local government control of land use. The state has established 19 statewide goals and guidelines, 2 of which are particularly relevant. Goal 10 requires local government plans to provide for land uses that make housing available that meets the needs of households at all income levels. Goal 14 requires all urban areas in the state to designate urban growth boundaries outside of which land cannot be converted to urban use. In effect, all urban growth is to occur within these areas.
Local government plans, which are reviewed by the state land use authority, must conform to these goals. One study found that the state authority does indeed carefully review the plans and require revisions (Knapp, 1990). Indeed, Knapp notes that, of the 53 urban jurisdictions with populations over 5,000, only one plan satisfied the requirements for goals 10 and 14 on its first review. He concludes that Oregon is exercising a consistent policy favoring high-density development and requiring local governments to zone for their fair share of multiple-family and high-density housing and that, furthermore, this policy is not one that local governments would have adopted on their own.
As to the question of whether these developments have actually reduced the cost of housing and opened up the suburbs to low- and moderate-income households, Knapp observed in 1990 that there had been little research on this question. He contended that, although there had been a multifamily housing boom, "there
is no evidence that the newly constructed housing is more affordable to the poor" (1990:43).
Two other researchers note the post-1990 rapid rise in housing prices in the Portland area (Abbott, 1997:34-35; Fischel, 1997). This suggests the possibility that the impetus for lower housing costs brought about through higher-density development may have been swamped by the rising cost of land caused by the pressures of recent rapid population growth on the restricted supply of land within the urban growth boundary. Additional research is clearly needed on the Portland experience.
Metropolitan Land Use Plans
Fischel (this volume) argues that lodging land use decisions at the metropolitan level would include those who would otherwise be excluded, and, as a consequence, political pressures would reduce the incidence of exclusionary zoning. He cites research suggesting that big cities tend to be less exclusionary in their zoning than do their suburbs.
Downs (1997) argues that the most logical way of reducing fragmented local government control over land use decisions is to adopt a single metropolitan-area-wide government. However, Downs then succinctly observes, "that tactic has absolutely no political support from either suburban or city residents or officials; hence it will never happen in most metropolitan areas" (1997:47).
A close approximation to metropolitan land use control, at least in its structural form, exists in the Minneapolis-St. Paul region. The Twin Cities Metropolitan Council, under state legislative authorization, established a Development Framework Plan for the region in 1976. All the region's municipalities and townships are required to submit their own comprehensive plans to the metropolitan council for approval. If the local plan is inconsistent with the growth projections of the Development Framework Plan, the council can hold up the plan's approval. Unlike the Oregon state land use plan, there is no specific provision for low- and moderate-income housing required of local governments by the Development Framework Plan. Also unlike the Oregon experience, the process appears to have had very limited impact. Orfield observes (1997:123): "the Met Council has narrowly construed its authority. Under a system of self-imposed restraint, the council will require a plan amendment only when the local comprehensive plan imposes a burden on a metropolitan system that 'threatens its capacity'—a fairly cataclysmic event."
Mills and Lubuele (1997) caution that there are inevitable limits on the ability to move low-income households to the suburbs by ending exclusionary practices. They note that constant-quality housing prices are lower in cities than in suburbs and that households seeking low-cost housing will inevitably find central-city housing cheaper (1997:754): "Poor-quality housing can be built in some suburbs, but old housing cannot, and higher-quality schools and lower
crime rates keep land values high in suburbs. The result is that the cheapest housing in the metropolitan area is often found in inner-city slums, even if land use controls impose almost no limits on the supply of low-income housing in the suburbs."
Inclusionary zoning specifically includes or requires a range of housing types and densities so that housing opportunities are available to low- and moderate-income households. Calavita and colleagues (1997:110) use the broader term "inclusionary housing" to refer to "a wide variety of techniques that link construction of low- and moderate-income housing to construction of housing for the marketplace, generally by including lower-income units in an otherwise market-driven development. The principal objective . . . is not only to increase the supply of affordable housing, but to do so in a manner that fosters greater economic and racial residential integration."
Inclusionary zoning can involve either state-imposed "fair-share" requirements on local communities (as in Oregon, New Jersey, and California) or publicly imposed requirements that developers, in return for gaining permission for housing development, will agree to provide a share of the units for low- and moderate-income households. In California and New Jersey (but not Oregon) these fair-share requirements are backed up by incentives to developers for the construction of low-income housing. Such incentives include density increases, land write-downs, impact fee waivers, and various other regulatory concessions. In Montgomery County, Maryland, outside Washington, D.C., new subdivisions of more than 50 units are required to include at least 15 percent of housing units at a price suitable for moderate-income families, with the county public housing agency purchasing up to one-third of these units. This program has created more than 9,000 units of moderate- and low-income housing over the past 25 years (Kleit, 1997).
Like Oregon, California has a state statute requiring localities to "make adequate provision for the existing and projected needs of all segments of the community" (Calavita et al., 1997), but, unlike Oregon, the state lacks power to mandate changes in local government plans. Nor does the state have the ability to require local governments to actually construct affordable housing. As Calavita et al. observe (1997:117-118): "Among the 527 cities and counties in the state required to adopt housing elements, the compliance rate at the end of 1992 was only 19 percent. Starting in 1993, HCD [the Department of Housing and Community Development] redoubled its efforts, raising the compliance level to 58 percent by December 31, 1995. . . . It is questionable, however, whether this increase represents a corresponding increase in housing opportunities or is instead little more than the creation of paper documents with little significance for implementation."
In contrast, New Jersey's inclusionary zoning is court imposed. In its 1975 Mr. Laurel decision, the state supreme court ruled that local governments were using a state police power, zoning, to exclude low- and moderate-income households in violation of the state constitution. The original remedy for exclusion was the creation of an obligation for communities to allow "least-cost" housing to be developed. However, as Fischel notes (this volume), newly developed least-cost housing still turned out to be too expensive for low- and moderate-income households. As a consequence, the court imposed a requirement that suburbs must accept a minimum amount of newly constructed low-income housing for any new market-rate housing constructed. The low-income housing is to be paid for by charging developers of the market-rate housing for rezonings.
Fischel (this volume) found little evidence that the inclusionary zoning process brought about by the Mr. Laurel decision has changed the general pattern of suburban housing development. Calavita et al. (1997:130-131) note that, although New Jersey developers were willing to build some units of low- and moderate-income housing as a cost of doing business during the housing boom in the 1980s (particularly as a part of large-scale town house and condominium developments), they have become much less willing to do so during the 1990s. They point to the increasing use of a provision in the law that permits developers to make per-unit cash payments to the municipality in lieu of constructing low- or moderate-income housing; the municipality, generally a suburban jurisdiction, is then permitted to give the cash contribution to another jurisdiction, usually a city, for the construction of affordable housing units that would be credited to the suburban municipality's fair-share account. Calavita et al. contend that the Mt. Laurel decision has had some impact at increasing low-and moderate-income housing units in the state, and they remain supportive of the inclusionary housing approach. However, they acknowledge the limits of the accomplishments as a means of reducing racial segregation, observing that "few projects have substantial minority—particularly African-American—populations and that the minority population is nominal or even nonexistent in many" (1997:129). As evidence they cite a study that sampled new suburban inclusionary housing developments occupied between 1988 and 1996 and found that only 12 percent of the occupants had previously resided in cities, and of these, less than 25 percent were black (Wish and Eisdorfer, 1996).
In the Minneapolis-St. Paul area, the state's Livable Communities Act, enacted in 1995, has created what is in effect a voluntary approach to inclusionary housing. Through a fund administered by the Twin Cities Metropolitan Council, resources to support affordable housing projects (as well as for tax base revitalization and economic and community development) are made available to communities that voluntarily agree to participate in the Housing Incentive Program created by the act. The program requires housing financed under its provisions to include housing at all income levels. In the first year of the program, eight grants
totaling $1 million were given for affordable housing projects in seven communities (Metropolitan Council of the Twin Cities, 1997).
Orfield, however, argues (1997) that the Livable Communities Act was a weak compromise that essentially reaffirmed powers the Metropolitan Council already had and, since it was voluntary, would result in little substantive accomplishment. Calling its passage ''symbolic,'' he nonetheless called it "a platform on which a stronger and more enforceable act could be built" (p. 152).
Efforts to Limit or Direct Growth
Another approach is aimed at redirecting growth toward denser, more developed urban and suburban areas. Although containing sprawl is the explicitly stated objective, achieving higher densities should (other things being equal) reduce the land component of housing prices and reduce the price of housing in the suburbs, but higher-density housing could increase construction costs for housing above two or three stories.
Efforts in Portland and Minneapolis-St. Paul have already been discussed. Another initiative is Maryland's Smart Growth Program, designed to "halt suburban sprawl and address its impacts" (Maryland Office of Planning, 1997). The legislation prohibits the provision of state funds for highways, sewer and water construction, and housing and economic development assistance to growth-related projects or developments not located within a "priority funding area," which is an area designated for growth and development. The initial legislation designated the city of Baltimore and suburbs within the Baltimore and Washington beltways (essentially inner and middle-core suburbs) as priority funding areas as well as other major urban areas in the state.
Changes in economic incentives may serve to redirect growth into more compact development as well. It has been argued that requiring automobile users to pay the full, unsubsidized cost of automobile use through higher gasoline taxes will make low-density development on the metropolitan fringe less likely. Similarly, some have argued that, in areas where new suburban developments do not pay the full marginal cost of infrastructure development, they should be required to do so.
Restoring Property Rights and Requiring Suburbs to Pay the Cost of Exclusion
The restoration of the development rights of owners of undeveloped land and land ripe for development would accomplish the same objective as metropolitan government with respect to reducing the economic stratification of suburban communities and opening up the suburbs to the poor, according to Fischel (this volume). He reasons that owners of such land "are representatives of people not resident in the community who would constitute net additions to the population. . . . Suburban zoning laws regulate the market in part because the market would
give existing suburbs more poor people than current residents would like. The owner of undeveloped land, however, does not care whether the high bidder for her 10-acre tract proposes to put 2 mansions or 40 bungalows. In most places, the developer of the higher-intensity use, which will usually serve lower-income people, can outbid others."
Fischel would not simply eliminate land use controls and give landowners and developers the unlimited tight to develop what they wanted. Instead, he would give them the tight to develop at normal suburban densities, "determined by comparisons to existing well-planned and socially diverse developments at comparable locations within the metropolitan area" (this volume). Communities would be able to purchase lower density for any tract by paying the difference between the land's value at the normal density and its value at the community's desired lower density. Fischel envisions beneficial results: "the existing residents can thus raise the suburban drawbridge if they are collectively willing to pay for it, and I do not doubt that some would. But I also have no doubt that the impassioned pleas by most neighbors who oppose development would be greatly muted if, to get their desires, they would have to pay more in property taxes. Requiting the local governments that want to preserve open space to buy that space (or the development tights) would make existing voters pay attention to the opportunity cost of exclusion and, for the most part, deter inefficient and inequitable low-density zoning."
Subsidies or Cost Reduction
The exclusionary effects of local land use regulation work by increasing the cost of housing beyond the means of low- and moderate-income households. Rather than focusing on the land use regulations themselves, one possible policy option is to attempt to lower the cost of suburban housing through other means or to increase the ability of low- and moderate-income households to pay for such housing through subsidies. Lowering housing costs could be achieved through devices such as improvements in building technology or lowering standards currently incorporated in building and housing codes. Increasing the effective demand of low- and moderate-income households to pay for housing could be achieved through substantially expanding the amount of federal housing subsidies such as Section 8 certificates, or by expanding them into an entitlement program. This approach is discussed in more detail in the section below on household mobility strategies.
Reducing Outcome Disparities
Five major categories of policy alternatives are directed toward reducing unequal opportunity and disparities in outcomes between minorities and whites and city and suburban residents within metropolitan areas:1
Place-based initiatives, which attempt to bring new resources and opportunities to distressed areas.
Worker mobility strategies, which attempt to link residents in cities to jobs in the suburbs, without changing either job or residential locations.
Household mobility strategies, which attempt to open up housing opportunities for residents of distressed areas in other areas where housing conditions, employment, and other opportunities are presumably better.
Human capital strategies, which attempt to improve the labor market skills of minorities and poor people.
Antidiscrimination strategies directed at reducing discrimination in labor markets.
Place-based strategies attempt to address unequal opportunity through economic and community development strategies designed to increase employment opportunities in distressed areas and improve them as residential and commercial environments. Targeted approaches aimed at specific geographic sections of a country's regions have a long history in many European countries and are not unknown in the United States (the programs of the Appalachian Regional Commission are the best known). Place-based economic development programs targeted at much smaller, more discrete areas on the sub-city level are more recent, originating with the British enterprise zones in the early 1980s.
Enterprise zones essentially designate a small distressed area as a zone in which various incentives are made available for firms that locate there. Many states have adopted variations of enterprise zone programs; the precise incentives and structure of such programs vary enormously. A federal empowerment and enterprise zone program, embodying some of the original enterprise zone concepts but also differing in important ways, was passed in 1994.
The federal program retains several components of traditional enterprise zone programs, in particular the tax credit for employers in the zone who hire zone residents, but it departs from tradition significantly by including large sums to address a variety of community social issues and issues related to job readiness ($100 million for each empowerment zone and $3 million for each enterprise zone). It thus combines a prior generation of poverty-directed, place-based activities, such as the Model Cities program, with the employer-directed incentives of enterprise zone programs.
The research suggests that enterprise zone programs have been successful in bringing about physical development in distressed areas. However, there is little evidence that they have resulted in additional employment for area residents. Ihlanfeldt (this volume) reviewed the evidence that the number of jobs increases in enterprise zones and found it mixed. However, even if jobs do increase, it appears doubtful that the result will be a net increase in jobs or that the area's
residents will be the beneficiaries. After reviewing the literature, Ladd (1994c:208) concluded, "Pure place strategies of the type represented by the English enterprise zone program are not an effective approach to pockets of urban distress. The main effect of the tax and regulatory relief provisions is simply to relocate firms to the zones from nearby locations." Vidal (1995:188) comes to the same conclusion with respect to state enterprise zones: "the information available about the design of state enterprise programs and about performance to date indicates that . . . these programs do not appear to represent a promising strategy for addressing the employment problems of poor inner-city neighborhoods."
Ihlanfeldt (this volume) contends that the fact that jobs are simply relocated from one area to another may not be a bad thing on fairness grounds if they are relocated from suburbs to central city. But this argument is not persuasive if the jobs are simply relocated from other nearby areas of the city, if area residents do not capture the jobs in any case, or, if they do so, but only at the expense of equally poor residents in nearby distressed areas.
More broadly, research by Galster (1997:62-63) casts doubt on the efficacy of place-based initiatives in general as a means of improving the economic condition of central-city residents. On the basis of results from his econometric model, he concludes that "the number of jobs located in the central city . . . did not have a statistically significant path directly to local labor market supply. . . . [T]he empirical estimates here suggest that even if a policy were instantaneously to create 10,000 . . . additional jobs in the typical sample central city, per capita incomes of city residents would rise by only $0.49. . . . The current cross-sectional evidence suggests that merely having more jobs located in the central city (relative to the working-age population) is insufficient to generate significantly more employment for residents."
The proximity of jobs to the local neighborhood (within a two-mile radius) also appears to have a very small effect on neighborhood employment rates in city neighborhoods, casting further doubt on the efficacy of small-area development strategies as a means of employing local residents (Immergluck, 1998). However, if the jobs are of similar occupational categories and skill levels as those of neighborhood residents, the effects are larger.
Another set of place-based strategies advocates building from within rather than attempting to attract jobs to distressed areas through incentives. Community development corporations and community development financial institutions are two such approaches. Community development corporations (called CDCs) are, according to Vidal (1995:204), "nonprofit, community-based organizations whose mission is to make the low- and moderate-income communities in which they work better places to live." Community development financial institutions, Vidal observes (1995:188-89), "have their genesis in the perception that certain types of communities and credit needs are not adequately served by mainstream financial institutions. Hence, despite their diversity, they have in common a commu-
nity development mission, a focus on servicing low-income communities or otherwise disadvantaged persons or distressed areas, and direct involvement in lending activities, often accompanied by other services." Peirce (1993:25;303) makes the case for such community-based efforts:
From Washington's Anacostia to Newark's Central Ward to Chicago's South Side to Miami's Liberty City, CDCs have been the fulcrum of amazing recovery stories. . . . more than 2000 CDCs are operating today across the country. In the last several years, they have built or refurbished 320,000 homes and apartments for low- and moderate-income households and have developed 17.4 million square feet of commercial and industrial space. Through an array of programs for business development and entrepreneurship, they also claim credit for the creation of some 90,000 permanent jobs. . . . Pessimists typically put down promising CDCs and neighborhood revitalization efforts as isolated achievements of charismatic individuals, unlikely to be replicated or spread. But the point is not to replicate precisely; it is to transfer the core ideas and some of the skills of human and economic revitalization under the toughest of circumstances.
Ihlanfeldt (this volume) reviewed the assessment of community development corporations and community development financial institutions by two social scientists, Vidal and Harrison, which, like the more prolific writings of community activists and advocates, is favorable, if more balanced. However, he notes that the evidence reviewed and provided by Vidal and by Harrison and his colleagues is entirely anecdotal and that "there is no statistical evidence on the effectiveness of CDCs as job generators."
Finally, it is possible to devise place-based strategies that target the residents of such places as poor inner-city neighborhoods for assistance, rather than, or in addition to, the jobs or employers that are located there. Thus, Lehman (1994) suggests targeting wage subsidies on the residents of low-income areas, regardless of where they work, a proposal that was originally considered as part of the federal empowerment zone legislation but ultimately rejected. The empowerment zone program does provide, however, for wage subsidies and other services for zone residents who also work within the zone. Another possibility is to target public service employment jobs on low-income neighborhoods and their residents.
Worker Mobility Strategies
Mobility strategies involve linking residents in distressed parts of the metropolitan area such as central cities to opportunities in the suburbs through improved transportation or through improved job information networks. Reverse commuting programs have been tried in many metropolitan areas, but, as Ihlanfeldt notes (this volume), there has been little systematic effort to evaluate their effectiveness. One study, which interviewed the staff members of 20 pro-
grams in 15 cities, concluded that there was little evidence that providing transportation had an impact on inner-city unemployment (Drachman Institute, 1992).
Bridges to Work, a demonstration program with a carefully constructed evaluation component by Public/Private Ventures, promises to provide better evidence on the effectiveness of reverse commuting strategies on worker mobility. Funded jointly by the federal government and a variety of philanthropic organizations, it is a four-year project providing reverse commuting services to central-city poor workers in five large metropolitan areas. The project, which will end in December 2000, includes both treatment and control samples, the latter consisting of poor central-city residents who will continue to receive their communities' usual employment-related services but not reverse commuting. Ihlanfeldt (this volume) observes that "this research promises to provide the first reliable evidence on the effectiveness of transportation programs as an anti-poverty strategy."
Improving job information networks and linking socially isolated residents of poor neighborhoods more effectively to them is another means of linking workers to jobs. Ihlanfeldt notes that computerized job opportunity networks are a frequent recommendation, but that "programs that are specifically designed to provide information on suburban jobs to central-city workers are rare and little is known regarding their effectiveness." Another approach, community-based job development, is designed to recreate the various components of informal job networks, including information on job availability, as well as counseling, job readiness, training, and contacts with employers.
Some critics of mobility strategies suggest that they are unlikely to succeed because physical access is only part of the problem; suburban employers will continue to discriminate against central-city residents, particularly black central-city residents, either through racism or through a belief that a central-city address is a proxy for poor education and work readiness. In addition, as Downs notes (1994), a worker mobility strategy, even if successful, would not break up segregated neighborhoods of concentrated poverty in inner cities or the adverse consequences that result from them.
Household Mobility Strategies
Household mobility strategies envision opening up residential opportunities in the suburbs, particularly middle-income suburbs, to poor and minority groups. The expected results would be moving these groups closer to job opportunities, thereby mitigating the spatial mismatch problem; reducing the adverse consequences of residing in areas of high poverty (assuming they are not replicated in the suburbs); and breaking the link between low-income and poor public services as low-income households move to jurisdictions with higher tax capacities.
One set of strategies, described above, involves breaking down existing regulatory barriers and exclusionary zoning devices that currently inhibit low-
income households from moving to predominantly middle-class suburbs by increasing the price of housing in these areas. Another set of strategies involves providing housing subsidies to low- and moderate-income households so that they can afford the higher cost of suburban housing. Such strategies might include scattered-site subsidized housing production throughout the metropolitan area or metropolitan-area-wide housing vouchers for low- and moderate-income households that would give them a housing subsidy wherever they found a suitable unit within the area.
An existing federal program, Section 8 certificates, has, since 1989, permitted recipients to search for housing eligible for Section 8 assistance throughout the metropolitan area. However, the potential impact of the program has been muted due to the lack of effective counseling for Section 8 recipients; a disinclination on the part of city public housing agencies who administer the program to lose the federal administrative payment that would result if their client found a house in the suburbs; and, in some cases, fair-market rental ceilings that effectively limit the choice of households searching for Section 8 housing in many suburbs. Prospective tenants also have difficulty finding units in the suburbs as a result of housing discrimination against minorities. An effective program might require metropolitan administration of federal housing programs or, at a minimum, more effective metropolitan coordination.
Such a strategy builds on the successes of the Gautreaux program, discussed in Chapter 3, but also suggests that counseling assistance as well as housing subsidies will be required. Although the results of the Gautreaux program appear promising (see Rosenbaum, 1995), studies in several other locales (Cincinnati, Hartford, Dallas, and Memphis) that are reviewed by Peterson and Williams (1995) appear somewhat more mixed; additional studies in other locales are under way but have not yet been completed.
Other complementary strategies aim at eliminating illegal discrimination in housing markets through more vigorous enforcement of existing laws and random audit-based testing in every metropolitan area (see below). Household mobility strategies do not necessarily imply either stable racially or economically integrated neighborhoods, although they do imply racially and economically mixed jurisdictions. Ihlanfeldt (this volume) discusses the literature on the extent of racially integrated neighborhoods and various policies to achieve them, including the debate over whether integration should be managed—and, if so, how—to foster racially integrated neighborhoods.
Critics of the household mobility approach argue that, if successful, it will further empty out city neighborhoods, exacerbating the problems of concentrated poverty. This will particularly be the case if the low- and moderate-income families who move out are self-selected on the basis of their motivation and are more likely to succeed. Household mobility strategies have political problems as well. At the receiving end, one such effort, the Moving to Opportunity program in Baltimore County, was brought to a halt because of the protests of suburban
residents who objected to low-income black households moving into their area. At the sending end, city politicians often do not see losing their constituents to the suburbs as a mark of success; black elected politicians in particular may fear the loss of political power and influence this implies.
In addition, the cost of a full-blown household mobility effort in terms of federal housing assistance could be prohibitive given federal budget fiscal constraints, and the politics of bringing a program such as Gautreaux to scale make it unlikely (Hughes, 1995:285): "simply consider that for African Americans to be represented in the Chicago CMSA suburbs in proportion to their presence in the metropolitan area (which is 27 percent black as a whole), the size of the suburban black population would have to quadruple, from 250,000 (in 1990) to I million. Likewise, for poor people to be represented in the Chicago CMSA suburbs in proportion to their presence in the metropolitan area (which is 11 percent poor as a whole), the size of the suburban poor population would have to increase by more than 300,000, from 190,000 (in 1990) to 500,000." Downs argues, to the contrary, that the scale necessary is not implausible (1994:110):
A simple quantitative analysis of conditions in the Chicago and Atlanta metropolitan areas shows that greater use of household mobility would be feasible. . . . [Assume] that inner-city poverty areas in the two cities contained 15 percent of their 1980 population. . . . If only a small percentage of new housing built during the 1980s in the suburbs of these cities had consisted of subsidized units occupied by inner-city households, the out-movement would have made a notable dent in reducing inner-city populations. In the greater Chicago area, if 5 percent of the suburban housing added from 1980 to 1990 had been subsidized and occupied by inner-city residents, 7.5 percent of Chicago inner-city households could have moved out of the inner-city. In the Atlanta metropolitan area, if 5 percent of added suburban housing had been similarly subsidized and occupied by inner-city residents, 82 percent of these people would have moved to the suburbs. This impact is much greater because Atlanta has a small population compared with its suburbs, and rapid suburban growth added a lot of new housing from 1980 to 1990.
Human Capital Strategies
There are a vast range of potential strategies that might be addressed to improve the human capital of poor and less-skilled people as a means of reducing disparities in outcomes. Taken broadly, these include policies related to early childhood development and preschool education, elementary and secondary education, higher education, job training and school-to-work programs, public health and nutrition, and others. Since the committee's prime concern is with metropolitan governance and its impact on disparities, we have focused attention on human capital policies related to an individual's location in the metropolitan area and, in particular, education.
As discussed, the quality of education and the skills one derives from it are
directly related to where one lives in a metropolitan area. Low-income areas and areas of concentrated poverty have, on average, poor-quality schools for a variety of reasons: if the schools are located in low-income school districts, resources are likely to be low relative to need; better-trained and higher-quality teachers are likely to choose to teach in "more pleasant" middle-income districts; students in low-income schools are likely to be attending classes with peers whose education deficits are as large as theirs, so that the potential for learning from peers is reduced.
One approach to this problem is to attempt to improve the quality of existing schools through school-based reforms, efforts to attract better teachers into low-income schools, reducing student-teacher ratios, and reorganizing school management systems (site-based management, decentralization, recentralization). A variety of approaches has been suggested, including "accelerated schools" by Henry Levin, "essential schools" by Ted Sizer, ''success for all" by Robert Slavin, and school development projects by James Comer (see Barnett, 1996, for a review and assessment of these programs). Linking district funding to student performance is another type of reform that has been tried in Dallas, South Carolina, and Kentucky (see Clotfelter and Ladd, 1996).
Another approach is to break the link between residential location and the location of schooling through "choice/mobility" strategies, such as metropolitan-area-wide desegregation or a metropolitan education vouchering system that would permit students to attend any public school in the area (some proposals include private schools as well) so long as space is available. Less comprehensive approaches include charter schools (publicly supported institutions chartered by the state but usually run by private or nonprofit institutions rather than the existing school system) open to students without regard to geographic location and, within existing school districts, magnet schools and school choice among any public school in the district. Since 1990, the city of Milwaukee has provided vouchers to low-income students to attend nonsectarian private schools; more recently, voucher programs have been undertaken in a small number of other jurisdictions.
Providing a market proxy mechanism that allows parents to choose the school their child will attend, with the money following the child, should provide a strong incentive for schools to produce outcomes that parents are seeking. However, a choice program, if it is to be practicable in terms of having a large impact, would have to do more than simply allow parents to place their children in more effective schools outside the traditional public school; it would also have to engage the public school system to respond by creating more effective schools within the traditional system.
The research evidence on the effectiveness of efforts to improve the quality of existing schools is to date somewhat inconclusive. However, there is at least some evidence that all of these potential approaches can have positive effects on the performance of poor inner-city students. For instance, Clotfelter and Ladd
(1996) show that performance can rise in districts in which funding is linked to student performance. Barnett (1996) shows that the Levin, Slavin, and Comer approaches have generated positive outcomes for students in particular settings, although questions remain about the extent to which these results can be generalized.
The evidence on the effectiveness of choice/mobility strategies, particularly vouchering strategies, is highly contested. Initial evidence on the performance of inner-city students in the Milwaukee voucher program is conflicting: Witte et al. (1995) and Witte (1997) found no significant effects on test scores, but Greene et al. (1997) found significant effects on both reading and math. Rouse (1997) found that the program had positive effects on math but not on reading. The program is very small (it involves only 1.5 percent of students in the Milwaukee school system), and the results are still unclear. Hoxby (1996b) also found positive effects from school competition on student outcomes.
There is also evidence that inner-city minorities have substantially higher test scores and educational attainment in private parochial schools than in public schools (see Greene et al., 1997). However there have been questions about whether this superior performance merely reflects selection bias: that is, are the results attributable to the effects of the schools or to the characteristics of the students who choose to attend them? Recent evidence from studies by Hoxby (1996a) and Neal (1997) that control for selection bias through more sophisticated instrumental variable modeling techniques (although both are still open to criticism) suggest that the parochial schools do succeed in raising performance.
Two recent reviews of the literature come to quite different conclusions. Levin (1998:378) concludes that "My own reading of the body of studies comparing student achievement in public and private schools is that differences are small." Henig (1998), in contrast, accepts that the existing body of research supports the finding that private schools produce superior outcomes, yet notes that it is still possible that the differences resulted from unmeasured selection biases.
Evidence on the effect of choice or voucher programs on the performance of the existing public school system is scant. A few studies are indicative and encouraging, though hardly definitive. These infer effects from cross-sectional work in which the degree of competition for public schools is proxied through creative means. Hoxby (1997) modeled student achievement as a function of the number of school districts in a metropolitan area and the evenness with which enrollment is spread over these districts. Thus, a district with a large number of school districts and a relatively small share of the area's student population in each district would be one in which a substantial amount of "choice" existed, whereas, at the extreme, a metropolitan area with only one school district or with two districts, one of which enrolled 98 percent of the area's students, would be one with low choice. Controlling for other factors, she found that increased choice results in a slight improvement in the reading and math test scores of
public school students in the area and a rather substantial reduction in per pupil expenditure. She also found that the greater the number of private schools providing competition to public schools in a metropolitan area, the better the performance of public school students, implying a response of public school systems to competition.
Couch and colleagues (1993) reported that, controlling for a variety of other factors, the performance of public school students on statewide algebra tests in North Carolina was higher in counties in which a larger percentage of county students were enrolled in private schools. This also suggests that public schools are responding to private school competition.
School-to-work programs can also play an important role in improving the skills of inner-city students and linking them to suburban labor markets (Bailey, 1995). At least potentially, the various career academies, tech prep schools, and other programs could serve to raise the incentives of students to improve their basic cognitive skills while in school, if they perceive clearer links between school performance and work opportunities with specific employers. Similarly, training programs for adults like the Center for Employment Training Program in San Jose, California, seem to be effective in providing inner-city minorities with skills that are directly linked to specific jobs and employers located in suburban areas. Although there are some promising models on school-to-work programs and training, it is not yet known how successfully these can be implemented on a wider scale. (For a review of the literature on job training, see Lalonde, 1995.)
Once again, there are also spatial considerations. Not only is school quality highly related to geographic location in the metropolitan area but also, as discussed earlier, the acquisition of human capital relates both to formal schooling and to the family and neighborhood environment. As Galster (1993:1440) observes, ''Children first may learn communication and interpersonal skills from family members and neighbors. Schools give children a second chance to learn these skills, however, because students interact with schoolmates from other families and neighborhoods." Thus, policies to reduce concentrations of poverty and achieve at least jurisdictional integration so that poor and middle-class and blacks and whites share the same schools and classrooms, even if not necessarily the same neighborhood, should also be seen as contributing to human capital.
Antidiscrimination in Employment Policies
Some of the differences in labor market outcomes between blacks and whites are due directly to illegal discrimination in the labor market. Stronger enforcement of existing laws is clearly called for. Matched testing employment audits on a random basis in all metropolitan areas might not only result in better enforcement but also act as a deterrent to discrimination if employers perceive that the qualified applicant they are interviewing may actually be testing their compliance with the law.
Special efforts could be made to raise the effectiveness of antidiscrimination policies in the suburbs. To date, these policies have been ineffective and, indeed, more so than they might otherwise seem, since discrimination is frequently measured by the fraction of black employees at firms relative to the fraction of blacks in the local population (rather than in the relevant labor market). Discriminatory companies thus may relocate to the outlying suburbs precisely to get away from black applicants and protect themselves from lawsuits (Bloch, 1994).
It should be noted that, to some extent, worker mobility policies and antidiscrimination employment policies are complements to each other in the suburbs: the former is designed to increase the supply of central-city labor in suburban labor markets, and the latter is designed to increase the demand for central-city minority labor.
Reducing Tax/Service Disparities
The existing distribution of population across space and jurisdictions in metropolitan areas results in serious tax/service disparities among local governments (see Chapter 3). These tax/service disparities in turn contribute to unequal spatial opportunity in metropolitan areas and to the disparities between residents of central cities and suburbs because they affect the nature of the public services residents receive (particularly education) and the price they pay for them. In this section, we explore alternative strategies for reducing these tax/service disparities or their effects.
There is a distinction between what has been called political stretching strategies and fiscal stretching strategies as means of addressing the tax/service disparity problem (Pagano, this volume). The principal objective of these approaches is either to extend the political boundaries of the city or to stretch its financial reach so that it can encompass the tax base of the broader metropolitan area (or beyond) and thereby lower average expenditure needs.
Political Stretching Strategies
Political stretching strategies attempt, in effect, to reduce or eliminate tax/ service disparities by the simple step of eliminating or greatly reducing the multitude of local governments in a metropolitan area and moving to a single unit within which taxes are levied and services are provided as though it were one jurisdiction. Options include aggressive annexation by the central city, reorganization to bring about broader units of government such as single-tier metropolitan government, two-tier federalized metropolitan government, city-county consolidation, and creation of metropolitan-wide, single-purpose special districts.
Annexation expands the boundaries of the central city and permits inclusion of the growing suburban population within central-city boundaries, usually adding households with higher incomes and lower service needs. However, annexation is possible (1) only when the central city is not already surrounded by incorporated municipalities, as is already the case in many parts of the country, particularly the older cities of the Northeast and the Midwest, and (2) when state annexation laws are permissive (see Rusk, 1993:20-22).
Creation of a single general-purpose metropolitan government, a two-tier federalized metropolitan government, or a city-county consolidation have all been suggested and, in some cases, put into place as political stretching reforms (discussed below). Of these, only a single general-purpose metropolitan-area government would in principle completely eliminate tax/service disparities. Two-tier metropolitan government envisions a metropolitan government responsible for some functions, but with an array of local governments, each with its own taxing and spending authority, responsible for other functions. As a consequence, tax/service disparities might be reduced but not eliminated. City-county consolidation would be equivalent to a single metropolitan government in those metropolitan areas that consisted of only one county, but in multicounty metropolitan areas it would simply reduce the number of local governments by eliminating all those previously within the central county, improving the relative fiscal situation for those who previously were city residents.
Within the units created by these reforms, it would seem that differences in tax burdens on residents due to differences in revenue-raising capacity of their previous local governments should be eliminated, assuming fair property assessment practices and administration of tax laws. However, this is not necessarily the case. In city-county consolidations, for example, the county is frequently divided into zones, with outlying areas receiving lower levels of service and tax rates and central-city areas receiving higher levels of services and tax rates (Harrigan, 1993:352). Two-tier metropolitan governments would also retain differences in tax rates and burdens as a result of local taxes imposed by the lower tier. Indeed, Hawkins and colleagues (1991) observe that the term "consolidation" is a misnomer, since in nearly all cases some independent governments with their own taxing authority continue to exist.
Whether differences in service quality would be eliminated, or the extent to which they would be reduced, is even more open to question. It is clear that simply creating a single jurisdiction does not mean that all residents within the component jurisdictions (or, more meaningfully, all neighborhoods) would receive equal services. Pagano (this volume), reviewing the literature on the degree
and nature of inequality of service delivery among neighborhoods in the same city, notes that, although studies have consistently shown that levels of service vary among neighborhoods, early studies concluded that these variations were largely due to professional norms and standards applied by service bureaucracies and did not systematically disadvantage low-income or minority neighborhoods. More recent, carefully designed studies by Bolotin and Cingranelli (1983) and Meier et al. (1991) found that levels of service to neighborhoods are correlated with both race and class; minority neighborhoods and low-income neighborhoods systematically receive poorer services than other neighborhoods in the same city.
Metropolitan consolidation efforts would seem to provide the potential for reducing tax/service disparities among local governments in metropolitan areas. Do they actually do so? Existing research provides only scant guidance, and most of the findings are little more than anecdotal. In the most systematic study, Gustely (1977) found the net fiscal impact of the Dade County two-tier consolidation was, although modest, in favor of poorer jurisdictions, with the city of Miami and its residents being the biggest winner. Blomquist and Parks (1995) argue that, contrary to expectations, the adoption of Unigov in Indianapolis has not redistributed the area tax base in favor of the central city. This results from the structure of the reform, which retains a variety of municipal functions for the city that must be financed solely from the city's tax base.
In terms of service distribution, Horan and Taylor (1977) found that the Nashville consolidation reduced service disparities between urban and rural areas; Swanson (1996), however, found that consolidation in Jacksonville did little to improve the flow of services to the poor. In a general review of a large number of consolidation efforts, Erie and colleagues (1972:32) conclude that "reformed metropolitan institutions often do not provide services any more equitably than unreformed institutions in the same area."
Using a different approach, Bradford and Oates (1974) tried to estimate the redistributive effects of moving from the present fragmented system of local government in the metropolitan areas of northeastern New Jersey to a unified system with a single area-wide property tax rate and equal per capita expenditures throughout the area. They concluded that, in the long run, a "somewhat more egalitarian" distribution of income would result, but that there would also be some adverse efficiency effects (see also Greene and Parliament, 1980).
In the same vein, Sacher (1993) attempted to model the incidence of local taxes and expenditures and the distribution of income in metropolitan Washington, D.C., under the current fragmented fiscal system and under a hypothetical metropolitan-area-wide system. He concluded that the overall redistributive effects of moving to a metropolitan system would be very modest. A metropolitan system would increase the burden of the suburban well-off relative to the central-city well-off (because the tax burden currently falls heavily on the well-off in central cities), but it would also increase the burden of the central-city poor (who
currently receive disproportionate levels of public services) relative to the suburban poor. There would be virtually no change in net incidence between income classes.
The political difficulties involved in bringing about these various forms of greater regionalism through boundary change are well known and are discussed in a later section, along with a variety of considerations concerning their desirability.
Metropolitan-Area-Wide, Single-Purpose Special Districts
Metropolitan-area-wide special districts with their own taxing and spending powers are, in effect, the equivalent of metropolitan governments for a specific purpose. Foster (1997:108) divides regional special districts into two categories: (1) "regionalizing-collectivizing districts" that spread costs across the metropolitan area through region-wide financing mechanisms for services enjoyed throughout the area such as zoos, convention centers, parks, and environmental protection and (2) "regionalizing-privatizing districts" financed directly by users of the service through fees and charges, such as bridge tolls, water consumption charges, and airport fees.
The question of service equalization across areas notwithstanding, creation of single-purpose regional special districts to replace or augment taxing and spending by large numbers of local governments should serve to reduce tax/ expenditure disparities for that purpose in metropolitan areas. Unlike other forms of boundary reorganizations in metropolitan areas, special districts, particularly for systems maintenance issues, are not always outside the realm of political feasibility. However, there is almost no empirical evidence on the actual distribution of service for various subregional areas within special districts or on the redistribution of tax burdens within the region as a consequence of moving to regional special-purpose districts. One exception is a study by Hawkins and Hendrick (1997), who found fiscal redistribution in favor of city residents as a consequence of the pattern of taxes and benefits from two special districts in the Milwaukee metropolitan area.
Fiscal Stretching Strategies
Fiscal stretching policies permit existing units of local government suffering from fiscal disparities to gain access to fiscal resources from outside their borders. Such options include metropolitan tax base sharing; county or state assumption of local service responsibilities; the ability to export taxes, particularly through devices such as taxes on income earned within the jurisdiction; and assistance from higher (federal and state) levels of government.
Metropolitan Tax Base Sharing
Tax base sharing involves a redistribution of revenues among local governments in a metropolitan area, with local governments contributing some portion of their tax base or tax base growth into a metropolitan fund, which is then reallocated among the local governments according to need-related criteria. The rationale is both to reduce fiscal disparities among local governments in the metropolitan area and to reduce the incentive for local governments to engage in nonproductive competition with each other to attract property-tax-paying economic activity within their borders. The extent to which fiscal disparities are reduced is dependent both on how much of the area's local tax base is set aside for the pool and the extent to which the distribution from the pool takes local fiscal capacity into account.
The best known example is the tax base sharing system in the Minneapolis-St. Paul area, which was imposed by the state legislature in 1971. All local governments in the seven-county metropolitan area are required to pay into an area-wide pool an amount equal to 40 percent of the growth in their commercial and industrial tax base since 1971. This amount includes growth from any source, including new construction, inflation, revaluation, and appreciation (Bell, 1994:160). The tax base is redistributed from the pool back to local governments on the basis of a fiscal capacity measure,2 so that jurisdictions with lower fiscal capacity receive relatively higher amounts of per capita tax base. The redistributed portion of the tax base is then taxed at the average area-wide rate. Two other examples of tax-base sharing are the New Jersey Meadowlands area and the Charlottesville, Virginia (Albemarle County), metropolitan area (Bell, 1994:154-55). Pagano (this volume) describes a voluntary tax base sharing program in the Dayton (Montgomery County), Ohio area.
The Twin Cities plan has reduced fiscal disparities among local governments in the metropolitan area from a range of 50:1 to a range of 12:1, according to Orfield (1997). A study for the Minnesota legislature cited by Bell (1994), however, reported that disparities in millage rates have been reduced but the effect in terms of reducing tax burdens has been much less.
Experience has shown that tax base sharing schemes do not inevitably work to the advantage of central cities. Chernick and Reschovsky (1995) report that St. Paul has been a net beneficiary of the process since its inception, but between 1984 and 1994 Minneapolis actually was a net contributor each year. They attribute this to a boom in commercial real estate in Minneapolis during that time period and to the fact that the distribution formula does not take into account the higher costs of providing services in central cities. Nunn and Rosentraub (1996:96) report that Dayton was a net contributor to the Montgomery County (Ohio) tax sharing scheme in 1992, despite the fact that it had a higher poverty rate than many other participating jurisdictions. This results from the distribution formula that allocates money to participating jurisdictions on the basis of popula-
tion (Pagano, this volume). The extent to which tax base sharing schemes assist central cities is thus critically dependent on the formula by which the pooled funds are distributed.
The relatively small number of such schemes suggests the political difficulty of bringing them into existence. Orfield, however, argues (1997) that a tax sharing program can be structured so that the winners far outnumber the losers, and political alliances can be created within the legislature so that representatives from winning districts outnumber those from losing districts. Even so, Orfield's account suggests how difficult it is to overcome ideological opposition and partisan political position taking, even when measures of objective interests would seem to point to easy victory.
The difficulty of financing local services that are either redistributive (and thus cannot be easily financed by a distressed city without further encouraging the out-migration of its tax base; see Peterson, 1981) or region-wide with respect to benefits has led, over time, to a state or county's assuming the responsibility for either financing or both financing and delivering particular services in many areas. For example, there are now only a few states (New York is one) in which city governments are responsible for financing a share of federal welfare or Medicaid services; most city universities have been taken over by the state; and in many places city public hospitals are now the responsibility of the county government. (In some states, there has been a movement for state government to take over failing school systems, although this usually does not imply a complete financial takeover.)
The problem with these approaches is that the incentive for service assumption by county or state government is lacking, except under situations of real stringency. Assuming financial responsibility will affect the fiscal status of these governments, leading perhaps even to unpopular tax increases and, in many cases, overlying county governments are nearly as fiscally distressed as are large city governments. Furthermore, the devolution of federal responsibility for a variety of programs to the states may have the effect of pushing responsibility downward toward local governments rather than upward to the county or state. However, where valued services are near collapse, county or state assumption may still be possible. In many cases, however, the most obvious candidates for assumption have already taken place.
All local governments are able to export some portion of their tax burden to nonresidents. The most obvious example is the property tax on commercial and industrial property owned by nonresidents, which in addition is frequently taxed
at a higher rate than residential property. Although there is substantial variation among states and even at times among localities within states, Chernick and Reschovsky (1995:153) note that the effective property tax rate on nonresidential property during the 1970s was typically about 25 percent higher than the rate on housing. Differential taxation of nonresidential property sometimes occurs through formal classification schemes and sometimes through de facto differences in assessment rates. However, some states have ruled that such differentials are constitutionally impermissible. In addition, there has been a growth in tax concessions given to businesses by local governments as a means of attracting them into an area or retaining them there, which has probably served to reduce the differential in favor of residential property.
Earnings or commuter taxes are the other major area with potential for tax exporting. For cities that are major tourist attractions, hotel, rental car, and entertainment taxes are also frequently resorted to. Local governments must receive permission from the state government in order to impose an earnings tax; as noted earlier, only 11 states provide local governments with such authority. Chernick and Reschovsky (1995:153) note that 6 of the nation's 24 largest cities make use of an earnings tax, and in 1989 the tax accounted for about 8 percent of total expenditures in these cities. Moreover, Yinger (1996:37) observes that five of the six central cities with the poorest "standardized fiscal health" in 1992 had access to a commuter tax, which served to improve their actual fiscal health.
As the above discussion suggests, the political problems of taxing the earnings of those who work in the city but live in the suburbs, while difficult, appear not to be insurmountable. But there is a question about the economic consequences of such taxes. Chernick and Reschovsky (1995:12) observe: "As with the property tax, the problem with a local earnings or income tax is that while it may be effective at exporting tax burdens and raising revenue in the short run, in the longer run it may hurt the competitive position of the city. In Philadelphia, nonresident earnings are taxed at almost the same rate as resident earnings, while there is no local income tax in the suburbs. Research suggests that the employment impacts of this tax have been quite negative."
To meet this concern, they suggest (1995:165) that suburban residents who work in the city be permitted to take a dollar-for-dollar credit against their residential property tax bill for any wage tax payments they make to the jurisdiction in which they are employed, with the state compensating localities for a portion of the foregone property tax revenues.
Equalization Aid from State and Federal Governments
Equalization aid is designed explicitly to reduce fiscal disparities among local governments by providing disproportionately greater aid to those with lower fiscal capacity and higher needs. Most countries explicitly provide general grants to their subnational governments for this purpose (see Wolman, 1985). However,
at the federal level, general revenue sharing was the only program explicitly designed for this purpose, and its distribution formula was poorly designed to do so, resulting in only a very mild equalizing effect. Other intergovernmental aid programs, although not specifically designed for equalization purposes, nonetheless can produce equalizing results. Fiscal assistance from state and federal government plays an important role in moderating the degree of fiscal disparities among local governments. Ladd and Yinger (1991) report that, between 1972 and 1982, much of the decline in the standardized fiscal health of large cities was made up for through grants from the state and federal governments. Nonetheless, in the aggregate, U.S. federal grant programs have a very modest impact on reducing disparities among local governments when compared with those of other countries (Wolman and Page, 1987:88-89).
In addition, federal aid to local governments has fallen precipitously since the late 1970s, from 16 percent of municipal general revenue in 1978 to 5 percent in 1992. State aid as a percentage of local government revenue remained nearly constant over that period, but there was substantial variation among individual states. State aid to local governments has also generally tracked the economic cycle, rising in growth periods and falling during downturns. Between 1978 and 1992, federal and state aid in the aggregate as a percentage of municipal government general revenue fell from 40 to 28 percent. Reductions in intergovernmental aid to local governments have made the equalizing potential of aid more difficult.
State governments provide both general grants for local government operations (about 25 percent of total state aid to local governments in 1992) and categorical grants for specific purposes. In some cases, these grants are explicitly equalizing in intent. Most of the debate about equalization grants has occurred in the context of categorical grants for education purposes, as state court decisions have ruled that gross disparities in fiscal capacity across school districts are impermissible. Although some of these efforts aim explicitly at reducing the actual differences in per capita spending among school districts, the basic principle is that of locational neutrality, i.e., the taxes that individuals bear to support a given level of educational services should not depend on where they reside (Downes and Pogue, 1994:55).
States have responded by designing explicit education equalization grant programs, either through structuring more generous foundation programs or through power equalizing. Foundation programs provide school districts with the difference between a state-set per student amount (the foundation grant) and what they would raise from levying a state-specified tax rate on their property tax base. Power equalizing permits each school district to levy the tax rate it wishes against its property tax base and then provides the district with the difference between a standard yield set by the state for each rate and what the district actually is able to raise at that rate. Districts that raise more from their own base than the state
standard receive no state aid (see Pagano, this volume; Downes and Pogue, 1994; Reschovsky and Wiseman, 1994).
Education equalization grants have been criticized from two perspectives: on one hand, because they have not, in fact, resulted in equalization, particularly if the goal is to equalize per pupil spending across school districts. Reschovsky contends (1994:191) that ''the limited available evidence suggests that, largely for political reasons, achieving school finance equalization . . . is by and large an elusive goal."
On the other hand, it has been argued that equalization programs may achieve equalization, but at the expense of lower overall spending on education (see Reschovsky and Wiseman, 1994; Hoxby, 1996b). This could result if education financing is transferred in large measure from the local to the state tax base, thus severing the link between local willingness to tax oneself for education and the understanding that the taxpayer will benefit from undertaking the burden with additional education spending in his or her district. By contrast, at the state level, taxpayers may be less willing to tax themselves for education, since it is not obvious that their district will gain the benefit from the tax. Hoxby (1996b) also argues that, although equalization schemes like that of Serrano in California, level down, achieving equalization by preventing wealthier districts from spending as much as they wish, and achieving redistribution from property-rich to property-poor districts, at the same time, overall spending on education may be reduced. She concludes that "students from poor households end up experiencing lower school spending under very stringent school finance equalization" (1996b:18).
Others argue that a well-designed equalization program does not inevitably have to reduce expenditure; indeed, if it is so desired, it can be structured to encourage expenditure. Both Chernick and Reschovsky (1995) and Ladd and Yinger (1994) argue for a foundation program that requires local districts to impose a minimum tax rate and then have the state make up the difference between what that rate yields and a generous foundation level (i.e., a minimum per pupil expenditure level for all districts), providing that level accounts for differences in costs. Assuming local districts are not prevented from spending above the foundation level, strict equalization may not result, but disparities will be reduced and the receipt of an adequate standard of education will no longer be dependent on the wealth of the community in which an individual resides.
Ladd (1994b:50) also argues that states could use a program designed to narrow the gap between fiscal capacity and expenditure need, as she and Yinger have defined it, as the criterion for distributing funds to local governments; this has been done in Massachusetts. The problem, as Pagano notes, is more one of political feasibility rather than inability to design a workable program.
Breaking the Link Between Location and Service Quality
In addition to political stretching and fiscal stretching strategies, a third option would be to break the link between residential location and service delivery, so that residents of jurisdictions with high fiscal disparities would be able to receive services throughout the metropolitan area or beyond. Voucher programs or metropolitan or statewide school choice programs (as in Minnesota) are examples, as is the METCO program in the Boston area. METCO enables more than 3,000 minority children in the city of Boston to attend suburban public schools at state expense.
Improving Metropolitan Governance
This section examines changes in metropolitan governance that might serve to improve the opportunity structure of metropolitan areas. These alternatives can be considered as counterfactuals to the existing metropolitan governance system. The section examines what we know about the potential of these various alternatives for positively affecting the metropolitan opportunity structure and reducing inequality of opportunity, as well as their political feasibility. It also examines the possible adverse consequences of metropolitan reform strategies for other important values, such as efficiency, choice, and local autonomy, as well as the possibility that there is a trade-off between these values and achieving greater equality of opportunity.
Reducing political fragmentation by creating a more encompassing unit or units of government within metropolitan areas has been the subject of long debate. In the context of the committee's concerns, the object of such reform proposals would be to even out the spatial opportunity structure of metropolitan areas.
The argument for reform of metropolitan government structure in this context is that, in a unitary rather than a fragmented system, the incentive to engage in exclusionary zoning practices for fiscal ends would be eliminated, as would the incentive for households to sort themselves based on their ability to pay for high-quality public services. Sorting by race and income might continue to appear at the neighborhood level, but differences in tax burdens among individuals would be reduced and the quality of public services, including services of fundamental importance to equal opportunity, such as education, would no longer be dependent on where in the metropolitan area an individual resided.
The debate about reform, however, has taken place on a much broader canvas. Proponents of metropolitan government reform in the academic and research literature have focused their arguments on efficiency and administrative considerations, such as the need to achieve economies of scale and improved coordina-
tion and cooperation, although equity concerns related to fiscal disparities have also been present. They also argue that a metropolitan government structure will allow for expression of interests and development of policies shared by the entire region, such as greater economic growth and development.
Opponents of such centralizing reforms argue that large numbers of governments actually are desirable, both for reasons of efficiency and for those of democratic political participation and accountability. The starting point of these ''public choice" arguments is the desire for a government structure that promotes efficiency, as defined by the allocation of resources that best accords with citizen preferences, in metropolitan areas (see Tiebout, 1956; Ostrom et al., 1961).
To that end, public choice proponents propose a market analog: local governments produce and sell public services to consumers (taxpayers) who purchase them. Taxpayers differ among themselves in their taste and demand for public services and in their willingness and ability to pay for them, just as they differ among themselves with respect to other goods. In choosing residential locations, they attempt to match their preferences for tax/service packages with those offered by the large number of local governments in a metropolitan area. In short, they "vote with their feet." The large number of governments providing services in the area promotes competition, thus both ensuring that prices (i.e., the tax price of public services) are driven down and that many specialized niche needs can be satisfied. As a consequence of the large number of relatively small local governments, the tax/service preferences of citizens in any one community are less likely to diverge from the actual tax/service package provided than would be the case if the government covered a larger and more heterogeneous area. This is efficient in the sense of optimal allocation of society's resources.
The public choice argument suggests that the sorting process is actually desirable. Efficiency is conditioned on many small local governments with homogeneous populations in terms of tax/service preferences. If such preferences are also correlated with other characteristics, such as income, class, and race, it also implies homogeneity with respect to these characteristics.
In political terms, the public choice argument is analogous to the proposition that government authority should be exercised at a very local level, where elected officials are close to the people and understand their needs and concerns and where individuals can engage in meaningful political participation and hold their elected officials accountable.
The above discussion suggests the very real possibility that there may be a trade-off between the values associated with equity (in particular, the reduction of unequal opportunity) and values that have undergirded the traditional American system of local government, such as efficiency, choice, and local autonomy. Certainly such a trade-off is perceived by many of the opponents of various proposals for metropolitan reform.
Many of the propositions in this debate have been subjected to the test of
empirical research. The literature on the efficiency consequences of consolidated versus fragmented systems of local government (or large versus small units) is particularly extensive (for reviews, see Hirsch, 1968; Parks and Oakerson, 1989; Dolan, 1990; Boyne, 1992a, 1992b; Dowding et al., 1994; Durning, 1995; Foster, 1997; Oakerson, 1998). The preponderance of the evidence indicates that small local governments (and thus metropolitan areas characterized by fragmentation) are more efficient for labor-intensive services, whereas larger units are more efficient for capital-intensive services (because of economies of scale) and for certain overhead functions.
Evaluation of the effects of metropolitan consolidation reforms that have occurred in the post World War II era is disappointing. Most of the studies of these reforms have focused on the politics that brought them about rather than on the consequences of reform for policy (Carl Vinson Institute of Government, 1989). Effects are inherently difficult to assess, since the number of governments that meet the reform criteria are small. City-county consolidations were not uncommon in the nineteenth century (Harrigan, 1993), but only 22 have been approved since 1921 (Pagano, this volume), and, of these, only 3 have involved populations of more than 250,000 people: Nashville, Jacksonville, and Indianapolis. Miami-Dade County is a rare example of two-tier metropolitan government, although incomplete consolidation renders many of the city-county consolidation efforts (such as Indianapolis-Marion County) into a form of two-tier government as well. There is no real example of unitary, general-purpose metropolitan government in the United States.
Evidence on the consequences of comprehensive reform is therefore sparse and impressionistic. Pagano (this volume) observes that "although the evidence on fairness or equity of service delivery tends to be anecdotal, studies on tax equity and service delivery in consolidated counties are nonexistent, focusing instead on residents' satisfaction of services."
There is general agreement that consolidation has not reduced costs (as predicted by some reform advocates) and, in fact, may have even increased total local expenditures (consistent with the literature described above). Gustely (1977) found that expenditures rose after the Dade county consolidation. Benton and Gamble (1984) came to the same conclusions in their study of Jacksonville. Erie and colleagues (1972:30) reviewed a range of consolidation efforts and concluded that "the net effect of restructuring is a per capita increase in service costs," which they attribute to an increase in average service levels.
What are the effects of metropolitan consolidation on the disparities in metropolitan areas between blacks and whites and cities and suburbs with which the committee is concerned? Again, there are no systematic empirical studies. The evidence that does exist, however, suggests that these efforts have had no significant impact on redistributing income or on addressing the problems of the poor or racial minorities (Erie et al., 1972; Canter, 1973; Blomquist and Parks, 1995; Swanson, 1996). Harrigan (1993) concludes that areas that have undergone
metropolitan reform have fared no better than multicentered metropolitan areas in dealing with problems that place poor people and minorities at a disadvantage.3 As has already been discussed, there is little evidence that metropolitan consolidation reforms have actually resulted in reduction in tax/service disparities in metropolitan areas, although they have the potential to do so. And although there appears to be some scope for consolidation schemes to achieve at least a modest degree of income redistribution, there may be a cost to pay in terms of lost efficiency for doing so (Bradford and Oates, 1974; Greene and Parliament, 1980).
How metropolitan governments affect spatial opportunity structures by reducing the ability of localities in the area to control access through local land use decisions is also questionable. Harrigan (1993) notes that in all four of the major metropolitan reformed governments (Indianapolis, Nashville, Jacksonville, and Miami), existing suburban municipalities were allowed to retain their authority over zoning and land use decisions unimpaired. He concludes (1993:361): "Metropolitan governments are uniformly much more successful in dealing with the physical questions such as sewers, water supply, or parks and recreation than they are in dealing with social issues such as fiscal disparities, race relations, open housing, and the location of public, low-income housing in the suburbs. . . . [T]hese governments have not eliminated the biases of the multicentered metropolis on social access issues of zoning, schools, and housing."
The apparently disappointing evaluations of metropolitan reform efforts do not mean that it is impossible to design metropolitan structures that would have the desired impact on spatial opportunity structures, but only that existing ones appear to have failed to do so What is perhaps more relevant, the politics of bringing into being even the weakened forms of metropolitan governments appear nearly insurmountable. With respect to city-county consolidations, Pagano (this volume) notes that less than 20 percent of proposed consolidations have been approved since 1921. Furthermore, most proposed consolidations in recent years have been in smaller metropolitan areas, mostly in the South, and the primary rationale has been efficiency gains and cost savings (Durning, 1995; Pagano, this volume).
Why have these metropolitan forms of government been so difficult to achieve? When the effort is made to bring about these reforms through voter referenda rather than through an act of the state legislature (as was the case in Indianapolis), it typically requires majority approval in both central city and suburbs separately. In their review of voting patterns on 28 consolidation efforts, Marando and Whitley (1972) conclude that black voters oppose consolidation efforts more that whites and suburban residents more than city residents. Suburban residents usually vote against approval, as Harrigan observes (1993:361), because, "suburban voters interpret metropolitan reform as an attempt by central-city officials to 'grab' their tax base" and to increase their taxes. However, recent votes have seen central-city minorities increasingly oppose consolidation as well,
fearing a dilution of political influence and of their ability to control political offices. Pagano adds (this volume) that, in states requiring preclearance of electoral changes under the Voting Rights Act, metropolitan government restructuring efforts must first be submitted to the U.S. Department of Justice to demonstrate that minority voting influence is not being diluted.
In general, campaigns for metropolitan reforms are generated at the elite level in response to a specific problem of governance; as Marando and Whitley (1972:200) observe, "consolidation is not basically a grass-roots movement." Furthermore, metropolitan reform campaigns are conducted in an environment of public apathy, ignorance, and misinformation (Marando and Whitley, 1972; Greer, 1963). As Greet (1963:199) observes, "transmitting to the electorate the complex issues of structural change in government is a thankless and near-impossible task." Indeed, there appears to be a tendency for informed voters to be more supportive of consolidation proposals (as Hawkins, 1966, found in his survey of Nashville voters).
Analysis of why some metropolitan reform efforts succeed and others fail tend to focus heavily on idiosyncratic characteristics of each effort. However, it appears that success is much more likely when elite groups are not involved in organized opposition. Greet (1963) compared unsuccessful efforts in St. Louis and Cleveland with the successful Miami referendum. He notes that, in all three areas, reform had at least the nominal support of business groups and the wholehearted support of area newspapers. However, he points to the active and vocal opposition of central-city mayors in the unsuccessful efforts in contrast to Dade County, where no major elite opposition formed (see also Banfield, 1957). Sofen (1963:214) expands on this: "In other metropolitan areas there were many centers of power or independent pyramids like labor, racial groups, political parties. . . . In Miami, aside from officeholders and municipal and county employees, there were no other strongly organized power groups to oppose the loose coalition of central-city businessmen, area-wide newspapers and 'do-good' organizations that supported Metro. Similar coalitions have, of course, been found in every organized attempt to create metropolitan government. But elsewhere in the nation they invariably met with defeat because of the entrenched position of the countervailing groups who did not want to jettison the status quo for an unproven product."
Most political scientists who have studied the politics of metropolitan consolidation are quite pessimistic about the political feasibility of adoption through the referendum process (Greer, 1963; Horan and Taylor, 1977; Harrigan, 1993). As Greet (1963:199) observes: "When the structure of government is to be decided by a public opinion poll (and this is what a referendum turns out to be) the outlook . . . would appear bleak. The change cannot move far beyond the understanding and normative commitment of the median in the distribution of voters. That, in turn, would seem to be a basically conservative, anti-government
position. The alternatives are (1) to manipulate the election through redefining (or misdefining) the issue or (2) to bring about change through fiat."
By fiat, Greer means through the actions of the state government, superimposing a new structure on the metropolitan area, as was the case in Indianapolis and with the less comprehensive metropolitan reforms in the Twin Cities and Portland. Horan and Taylor (1977:200) agree, concluding that "an active state role in the reorganization of local government and the absence of a referendum appear to be the two ingredients that will facilitate, but certainly not guarantee, metropolitan reform."
In short, the evidence of the effect of metropolitan government on reducing disparities and on changing spatial opportunity structures in desirable ways is not compelling, and the political feasibility of achieving comprehensive metropolitan reform of this nature in very many metropolitan areas seems slight.
Comprehensive metropolitanization is not the only alternative for restructuring metropolitan institutions. As discussed, most metropolitan areas engage in various, though limited, single-purpose forms of regional activity. Three of them—Minneapolis-St. Paul, Seattle, and Portland—have developed metropolitan institutions that, since they are multipurpose in nature, represent a form of partial metropolitan government.
In the Minneapolis-St. Paul region, the state legislature established the Twin Cities Metropolitan Council in 1967. The council, whose members are appointed by the governor, covers a seven-county region and is responsible for a wide range of functions, including land use, housing, transit, sewage, parks and recreation, and several other services. However, throughout most of its history, the Metropolitan Council has functioned as a planning and policy-making body rather than as an operating agency. Harrigan (1996) calls it a bifurcated model, in which the Metropolitan Council sets policy and reviews, but the policies are implemented by other local governments, with the exception (since 1994) of transit operations and waste control. He argues that the bifurcated model worked well in some areas (e.g., sewerage), but resulted in a relatively limited impact on important land use decisions: "The more important the project was to development interests, the less the Council seemed to affect the decisions. The result is that the siting of most of these projects resulted from the traditional politics of land use rather than from a guided land use policy directed by the Development Framework" (1996:218).
In 1971, the Minnesota state legislature passed a Metropolitan Council proposal to set up the regional tax base sharing system (described above), which the council administers. In 1975, the Metropolitan Council passed a Development Framework Plan, setting up a regional land use and development plan designed to contain sprawl within a metropolitan urban services area. The state legislature
then gave the Metropolitan Council the authority to review the plans of individual local governments in the area and to hold up the plan's approval if it was inconsistent with the growth projected for the area by the Development Framework Plan, thus providing it with considerable power over local land use decisions (Harrigan, 1996).
In Portland, Oregon, the Metropolitan Service District (called Metro), unlike the Twin Cities Metropolitan Council, is an elected body. Metro was authorized by the state legislature in 1977 and created in 1978 by a referendum in the three counties it serves. It is governed by a seven-person board elected on a district basis and an executive officer elected at large. Metro provides services, including the regional zoo, solid waste disposal, and regional tourist development to 3 counties and 24 cities in the Portland area. It is also responsible for coordinating growth management, land use, and transportation planning throughout the region.
Metro actually has a role in delivering only a small number of services (Nelson, 1996:263). Its importance lies with its ability to review local land use plans and its role as a coordinator and convenor in addressing regional issues. In this capacity, it has played a key role in such important issues as arranging the distribution of fair-share housing allocations, creating the urban growth boundary, adjusting the appropriation of transportation funds, and helping to create a consensus for new regional initiatives. Nelson argues that it is the fact that Metro is an elected body that serves to legitimize these functions.
It is difficult to relate the experience of the Twin Cities and Portland to an impact on reducing disparities in outcomes. Harrigan (1996) notes that the Twin Cities Metropolitan Council was highly successful in its first two decades, but has been less successful recently. He attributes this to two factors that are of direct concern to our inquiry (1996:223): "The nature of the most pressing issues had changed from physical development in the 1960s to the much more intractable issues of social deterioration in the 1990s. Economic and demographic change worsened the central-city/suburban disparities." Recently, efforts to restructure the metropolitan council to allow it to better cope with these issues have been undertaken but have yet to be enacted. These include proposals for an elective council, the mandating of low- and moderate-income housing goals for each suburb, with the metropolitan council having the authority to enforce these through denying sewer and highway extensions to noncomplying suburbs, and regional housing reinvestment and poverty reduction efforts.
There is also some debate over how applicable the experience of the Twin Cities and Portland are to other metropolitan areas. Both areas have relatively small minority populations and political traditions and cultures that place a relatively high emphasis on rational discussion and negotiation. As Abbott notes (1997:43), "Portlanders share a political culture that considers policy alliances and team building to be the normal way of doing public business. Nurtured in nonpartisan political institutions for local government, the Portland style prefers protracted discussion and negotiation to ideological battles and electoral confron-
tation." Despite this, he considers that the Portland experience is, at least to a certain extent, generalizable (1997:39): "Portland in the aggregate is not a unique metropolitan area. Many aspects of its economic base, social geography, and demography certainly set it apart from the typical city of the South or Northeast. However, it is not sui generis, despite its 'whiteness' and its muted class divisions. In particular, Portland bears many similarities to a number of 'middle American' cities, including Indianapolis, Des Moines, Minneapolis-St. Paul, Omaha, Denver, Salt Lake City, Sacramento, and Seattle."
Expansion of Existing Metropolitan Institutions
Most metropolitan areas already have one or more regional institutions for specific purposes. What are the prospects for the evolution of existing regional institutions into multipurpose regional government entities?
Metropolitan planning agencies, most of which exist largely as a result of federal assistance and prodding, frequently provide planning across several functional areas. The most prevalent such entities are councils of governments, which are voluntary associations of local governments that join together for the purpose of engaging in area-wide planning, coordination of activities, review of local applications for state or federal assistance, research, and, in some cases, allocation of federal funds. Most of the metropolitan planning organizations, charged with the responsibility of transportation planning and fund allocation under the Intermodal Surface Transportation Efficiency Act, are councils of governments. Councils of governments have obvious limitations: they lack operating authority in most areas; because they are voluntary they find it difficult to engage in controversial action; and, in most metropolitan areas representation is on a one-jurisdiction/one-vote basis, which implies severe underrepresentation for cities.
Nonetheless, councils of governments and, in some areas, other metropolitan-area-wide planning institutions exist and are accepted. As Harrigan (1993:384) suggests, organic growth of these institutions carries with it the possibility of "metropolitan policy making with teeth." However, he also recognizes that, even if such steps were to occur, it would probably involve metropolitan decision making in systems maintenance rather than life-style issues.
Informal Cooperation, Interlocal Agreements, and Nested Government
Despite the lack of formal metropolitan government structures, interaction does occur among local governments in metropolitan areas. Public choice theorists argue that, when joint action across a fragmented local government system is required, it will occur through activity by overlapping governmental units (such as counties), interlocal agreements, privatization of services across local boundaries, or the creation of special districts. They particularly stress the ability to separate the provision of public services (a public function) from the production
of services, which may be undertaken by public or private entities and provides the possibility for delivering services effectively across jurisdictional boundaries (see Tiebout, 1956; Ostrom et al., 1961, 1988; Bish, 1971; Parks and Oakerson, 1989; Oakerson, 1998).
The large number of special districts, including those covering more than one jurisdiction, is well known. The extent of privatization and interlocal arrangements is less well known and understood. Indeed, the number of such arrangements in a metropolitan area can be staggering. Thompson (1997) surveyed municipal governments in the Detroit metropolitan area and found more than 4,000 privatization or interlocal arrangements, an average of 30 for each local government. Of the 2,967 interlocal arrangements municipalities reported, 47 percent involved relationships with other municipalities or authorities that were the joint creations of municipalities.
However, Thompson found that most of the interlocal arrangements were pursued for reasons of efficiency and effectiveness and involved municipalities that had common borders and shared similar populations, particularly with respect to education and race. There were "no obviously redistributive or tax base sharing arrangements between the central city and suburbs that were identified by either side" (1997:15).
Special districts are another means of providing links across a fragmented metropolitan area, at least for a specific service. Indeed, special districts can be tailored to encompass the geographic area most appropriate to the service in terms of relevant economies of scale on the production side and homogeneity of preference on the provision side. As of 1992, there were more than 13,000 special districts (exclusive of school districts) in metropolitan areas, with an average of 43 per metropolitan area. Approximately 25 percent of these are regional, providing services at the county or larger scale (Foster, 1996).
But, like interlocal agreements, special districts are likely to be put in place for reasons of efficiency, and it is not clear what the impact, if any, will be on the spatial opportunity structure with which we are concerned.
Pagano (this volume) notes that, on the ground, interlocal agreements and special districts appear to be the political action of choice as the means of overcoming fragmentation. He observes, however, that most of the services covered through these arrangements "are either 'household' functions, such as garbage collection, fire protection, policing, tax collection activities, payroll, and planning, or infrastructure activities, such as transportation improvement districts, municipal utility districts, metropolitan planning organizations, and the like. They tend not to be social services."
A "nested" arrangement, with larger governmental units overlaying smaller ones, is the most efficient manner of overcoming fragmentation in metropolitan areas, according to Parks and Oakerson (1989:22-23). They observe that "those issues that cannot be effectively addressed by governments of a large number of relatively small, adjacent municipalities can be assigned to overlying special-
purpose or multipurpose jurisdictions.'' Moreover, they contend that such a fragmented and nested system of local government is highly functional, arguing (1989:23) that "local governments tend not to be 'balkanized' in multi-jurisdictional metropolitan areas; rather they are linked organizationally by webs of interlocal agreements and overlaid by larger-scale arrangements for specific purposes."
Within such a nested arrangement, distributional equity concerns can be pursued at an overlying level, without the need of creating a cumbersome and overarching metropolitan government. In this way, Parks and Oakerson suggest an efficient way in which redistribution might be achieved and the spatial opportunity structure altered.4 They fail, however, to identify the political incentives or conditions that might bring about such redistribution (Keating, 1995). Unfortunately, we have come full circle; under the present system, redistribution appears not to occur or to occur insufficiently to make a difference.
Metropolitan Governance: The Third Wave
There has been a recent resurgence of interest in regionalism, which Wallis (1994) terms the "third wave" of regional governance. He notes that the shift in terminology from metropolitan or regional government to governance is self-conscious (1994:292): "The change in terminology reflects a shift in focus from formal structural arrangements to informal structures and processes for setting policy and mobilizing action. . . . Attempts to achieve regional governance are being led largely by cross-sectoral coalitions or alliances whose interests tend to fall largely within a specific strategic arena. Private corporations, for example, are primarily concerned with economic development, but their interest clearly spills over into concerns over infrastructure, affordable housing, public education, and issues in other areas. Likewise, nonprofit agencies engaged in human service delivery advocacy tend to focus on social equity."
These collaborative efforts reflect an approach that focuses less on authority and more on consensus building as a means of bringing about coordinated activity in a variety of social and political settings (see Innes, 1992, 1996). With respect to regional activity, efforts have tended to focus on specific projects, but there is frequently (Wallis, 1994) substantial stability over time for specific projects and even institutionalization. Peirce calls this cooperation at the regional level "citistate governance." He examines six metropolitan regions in depth and observes, on the basis of his investigations (1993) that "what struck us most was the potential of virtually every citistate to alter its environment to position itself for the new economy. The right civic forces have the potential to coalesce to achieve some form of shared governance" (1993:36).
Nonetheless, Peirce is not convinced that this kind of regional collaboration is sufficient. He notes the impressive degree of informal collaboration among organizations in the Cleveland area to come to grips with the problems of the area
economy (1993). Participants include representatives of the city Department of Economic Development, Cleveland Tomorrow, the Greater Cleveland Round Table, Cuyahoga County, local universities, and two community foundations, the George Gund Foundation and the Cleveland Foundation. He concludes, however, that "moving from informal consultations to such tangible regional steps as tax-base sharing, cross-border school enrollment, or formation of a work force preparedness plan for the Cleveland region in the 21st century, is very difficult" (1993:34). Instead, Peirce concludes that "a region simply must have some form of umbrella regional governance structure. At a minimum, such an organization needs the power to resolve disputes between individual governments of the region. At a maximum, it would assume direct control of, and coordinate, the major cross-regional functions . . . now performed by independent special authorities" (1993:319).
As Peirce's discussion suggests, there is a debate among proponents of metropolitan governance between those who put great emphasis on consensus building and those who believe that progress requires the building of majority coalitions. There is little disagreement that third-wave metropolitanism is highly desirable. The question is how much an essentially consensus-based approach can do to move toward more equality of opportunity and a reduction in outcome disparities between city and suburbs and whites and minorities. Orfield (1997), for example, lauds the efforts of third-wave regional governance advocates to create coalitions on behalf of a regional perspective or regional interest; however, he is skeptical that such a regional interest will emerge or drive changes in the city's favor.
Federal Actions to Encourage Metropolitan Governance
The federal government has played an important role in encouraging metropolitan activity by providing assistance to metropolitan planning organizations such as councils of governments, and by requiring metropolitan planning and, in some cases, resource allocation within specific programs such as transportation. Federal general aid to agencies such as councils of governments reached its peak in the 1960s and 1970s and has now largely disappeared, but there are actions the federal government could take to encourage greater metropolitan cooperation and activity. It could, for example, make clear that all of its grant programs to local governments are also available to consortia of local governments and to metropolitan-area-wide organizations. It could further encourage regional applications by indicating that it would give a degree of priority to such applications or additional funds for local governments that engage in such regional applications. It could set aside a portion of grant funds for programs such as the community development block grant to be used only for metropolitan-area-wide projects. It could create a small program of grants for innovative metropolitan-area-wide initiatives, with an emphasis on those that deal with problems related to unequal opportunity in metropolitan areas.
More radical steps could also be taken. Downs (1994) proposes, for example, that the federal government allocate all federal assistance for local government in metropolitan areas through a regional institution created by area residents or by the state government. This regional allocation agency would have the power to ''allocate the federal funds within each program area either to local governments or directly to households, service delivery agencies, or other recipients" (1994:76-177). The Intermodal Surface Transportation Efficiency Act already does something similar to this for federal transportation funding.
State Actions to Encourage Metropolitan Governance
Governance arrangements in metropolitan areas are ultimately a product of state government action. Local governments operate within a framework of state laws and regulations; they are, legally and constitutionally, creatures of the state. State governments could thus bring about forms of metropolitan governance. Indeed, both the Twin Cities Metropolitan Council and the Indianapolis Unigov were the product of state government activity rather than local choice. States may set up special districts within metropolitan areas; they may structure state laws to make municipal incorporation more difficult and to make annexation by cities easier. They can also reduce the power of local governments to engage in exclusionary local zoning behavior and can engage in redistributive tax and expenditure policy that favor more distressed jurisdictions within metropolitan areas.
State governments can also encourage cooperative metropolitan activity among local governments. Recent Virginia legislation, for example, provides state funding for the creation of regional partnerships among local governments and private-sector actors in pursuit of economic competitiveness (Richman and Oliver, 1997). Indeed, state governments could engage in all of the activities suggested in the previous section for the federal government.
We do not consider income transfer programs, since these do not address the causes of unequal opportunities and disparities.
Fiscal capacity is defined as the per capita equalized market value of all real property in a jurisdiction.
Despite these findings, evidence from citizen satisfaction surveys has frequently indicated strong black support for the results of metropolitan reform (see, for example, Stowers, 1996, in reference to Dade County and Lyons et al., 1992, who find greater satisfaction with services among blacks living in a neighborhood in the consolidated government of Lexington than among blacks living in a socially and economically similar black suburb in the fragmented Louisville metropolitan area).
Only fragmentary evidence exists on whether such redistribution does indeed occur. One study (Hawkins and Hendrick, 1994) found fiscal redistribution between suburban areas and central cities through the tax and expenditure behavior of the overlapping county government in the Milwaukee area; another (Banovetz, 1965) found mixed evidence in the Twin Cities area.