University of Southern California USC
Peter Gordon

Peter Gordon's Blog

A blog exploring the intersection of economic thinking and urban planning/real estate development and related big-think themes.

Recommended Books

Recommended Films

A Critique of New Urbanism

Peter Gordon and Harry W. Richardson
University of Southern California

Presented at the November, 1998 Meeting of the American Collegiate Schools of Planning (Pasadena, CA).

Table of Contents

Top Part I: The Failings of New Urbanism

Top Introduction
This paper deals with New Urbanism in the broad sense of prescribing public policies for the built environment as a whole rather than solely in the narrowest sense of examining the pros and cons of New Urbanist communities (although the latter interpretation is also considered). From this perspective, a useful starting point is some elements of the charter of the Congress for the New Urbanism (we admit that we are being selective here to highlight some of the issues that we would like to discuss; readers might like to consult the whole document). The second part of the paper discusses a much broader issue, agglomeration economies, that we believe sheds a much sharper light on the suburbanization (equals sprawl?) issue.

First, the diagnosis of problems provides the impetus for the Congress's actions and prescriptions. Its membership "views disinvestment in central cities, the spread of placeless sprawl, increasing separation by race and income, environmental deterioration, loss of agricultural lands and wilderness, and the erosion of society's built heritage as one interrelated community-building challenge." Regardless of the importance and legitimacy of these issues, our reaction is that it is simplistic to expect a dissolution of these problems by physical planning interventions. The CNU is somewhat ambivalent on this point by recognizing "that physical solutions by themselves will not solve social and economic problems but neither can economic vitality, community stability, and economic wealth be sustained without a coherent and supportive physical framework," although their actions appear to go far beyond this point. David Harvey denounces this position as "spatial determinism" (Harvey, 1997).

Among the key principles of public policy espoused by the New Urbanists are the following: the promotion of neighborhoods that are diverse in terms of use (e.g. mixed use developments) and populations (mixed in terms of age, race and income); designing communities with transportation alternatives (especially walking, cycling and public transit) to reduce automobile dependence, implying a strong emphasis on compactness; a strong preference for infill development rather than peripheral expansion; some priority to accessible public spaces, community institutions and a variety of parks and other open spaces to foster communitarian behavior; the provision of affordable housing distributed throughout the metropolitan region as part of a jobs-housing balance strategy; stressing the importance of farmland preservation and environmental conservation, combined with architectural and landscape design principles that pay attention to local history and cultural heritage, climate and ecology; and the recognition of the metropolitan region as the functional economic region coupled with revenue sharing among its municipalities to finance the alleviation of region-wide problems.

Problems What is wrong with this approach? Most important of all, it embraces pie-in-the-sky social engineering based on a false diagnosis of society's urban problems, an excessive faith in the ability to change the world, and the prescription of policies that are implementable only under very special circumstances. We will illustrate this claim with some examples.

Top i. Durability of Capital.
Even if the New Urbanists could capture both political and popular support for their physical planning prescriptions, the results would do little to change the metropolitan landscape (Downs, 1994). The reason is that the urban capital stock is already largely in place and changes very slowly. As for the residential capital stock, much of it has been built in the last forty years and the time of its physical obsolescence is far off. Hence, the practical consequences of New Urbanism continue to be a small number of relatively small communities accommodating a miniscule proportion of metropolitan population growth. Demonstration projects, the object of international study tours, a pleasant living environment for a few thousand households, well-paid lecture tours for a small clutch of somewhat immodest architects, the New Urbanist communities amount to little more.

Top ii. Residential Preferences.
Fannie Mae has been conducting surveys about housing preferences for years. The findings have changed little. Regardless of income, race or current tenure status, 75-80 percent of households would prefer to live in a single family home with a private yard. Whereas it may be possible via creative architectural and landscape design to produce high-density single family home developments in the suburbs that are compatible with these preferences, it is probably impossible at the close-in infill sites promoted by the New Urbanists. Developers are not stupid, large ones have extensive marketing expertise, and in general they produce the housing that buyers want that guarantees their profitability. If New Urbanist-type developments were in demand by consumers, they would be built. Obviously, we have no objection in principle to the idea that producers should offer consumers what they want, and we favor experiments by builders that provide a market test to see whether households are open to a change in residential lifestyles. An interesting question, especially with regard to infill projects, is whether these alternatives are acceptable to the community at large, as opposed to the prospective purchasers. There are many examples of broader community objections to high-density projects, usually on traffic generation grounds.

Top iii. Farmland Preservation.
A favorite argument of the New Urbanists and other anti-sprawl protagonists is that low-density suburban residential development is eating up prime agricultural land. We have addressed this question elsewhere (Gordon and Richardson, 1998). Agricultural land use in the United States peaked in the 1930s; nevertheless, agricultural productivity has risen sharply because of a shift to profitable land-intensive crops. Urban development still absorbs less than five percent of the landmass. Much hand wringing is associated with the gobbling-up of agricultural land contiguous to urban development without due consideration for the overall land-use allocation picture; the "highest and best use" argument has been too easily discarded because of the rejection of market principles in New Urbanist thinking. Two other points deserve a mention. First, the argument that U.S. urban development is adversely affecting the world food supply is nonsense. Starvation is a problem of distribution and inefficient food policies not of aggregate supply. Second, the environmental argument for preserving agricultural land is undercut by the fact that agriculture is, by far, the country's largest polluting sector, generating $173 billion of pollution damages in water pollution alone.

Top iv. Mixed Land Uses.
As implied by the CNU documents, New Urbanist communities are intended to be more than residential subdivisions. The plans are to have shops, a wide array of personal and consumer services, and workplace sites. Only by developing a broad mix of land uses can the goals, perhaps a dream, of walking to work and walking to shop be met. This is one of the plans for Kentlands, perhaps the most successful of the NU communities, yet commercial development there lags far behind. Apart from the pedestrian opportunities objective, however, there is no particular reason why these communities need to create an employment base. The idea of "selfcontainment" was one of the principles behind the creation of the British New Towns. Certainly, with the freestanding New Towns on green field sites (less clearly with the modified Expanding Town concept), it never worked well. Employment centers did emerge, but they did not cater to the local population. For skill mismatch and other reasons, the overwhelming tendency was for New Town residents to work elsewhere while the jobs in the New Towns were filled by commuters from outside. As a result, the strategy probably resulted in more commuting rather than less. This would be more true today than it was then because of ubiquitous accessibility by automobile. There is a stronger argument for having retail and other consumer services provided locally, but even in this case facilities have developed slowly as shoppers are attracted to major malls and other large-scale clusters.

Top v. Social Equity Issues.
CNU rhetoric gives substantial attention to promoting equity, fostering residential mixing, affordable housing provision, and reducing central city-suburb income differentials via middle-class infill development. Yet there is little evidence that NU communities have achieved these goals. Instead, they are turning out to be rather elistist settlements with average income levels much higher than in the surrounding areas. The Laguna West area, for example, has a household income two-thirds higher than Sacramento County, where it is located. At Seaside, the 1996 average sales price reached $503,500 (Garvin, 1998, p. 18). Offering variety in the housing stock does result some income mixing, but there are few signs of racial mixing, and supplying a range of housing products is typical of many standard residential subdivisions and is not restricted to NU communities. As for the idea that somehow New Urbanism can contribute to the stability, if not revival, of the central city, it remains just that - an idea. There is very little to show for it in practice. Despite the call for an integrated metropolitan unity, most NU communities are being built on greenfield sites some distance away from the central city, and infill development has been limited -- probably of necessity because of land scarcity - to tiny pockets. Hence, there is no identifiable relationship between NU communities and the fate of central cities and those who live there. If there is some consensus for tackling the social problems found in the central cities (and it is by no means clear that this consensus exists), it would be far better to deal with these problems via direct, tightly targetted measures rather than via land use controls and social experiments on the metropolitan fringe. As Harvey states: New Urbanism "builds an image of community and a rhetoric of place-based civic pride and consciousness for those who do not need it, while abandoning those that do to their 'underclass' fate" (Harvey, 1997, p. 69).

Top vi. Communitarianism.
Although the New Urbanists are regarded as very conservative, reactionary, and even stodgy, from an architectural point of view, they take a very old architectural tradition, that design affects social behavior, and radicalize it to the extent that they argue that incorporating specific design elements not only in buildings but in street layouts and neighborhood patterns can generate a communitarian spirit and dramatically increase social interaction. Although there are precedents for this view (in the writings of Jane Jacobs [1992], for example), and most people would accept that our behavior is sensitive to, and affected by, the surrounding physical environment, the New Urbanists take the argument to extremes. A major problem with their argument is that on the ground rather than in their polemics NU communities look little different than standard suburban areas, so that even if you accepted the communitarian argument it is difficult to believe that such subtle changes in the built environment could have more than miniscule social interaction effects. A more fundamental problem is that many New Urbanist projects are so influenced by the nostalgic longing for the archtypical small town of the past that they fall into the trap of believing that recreating its physical structure (at least to some degree) can simultaneously recreate its social and civic behavior. But society, culture and behavior have changed so much that this is a false dream. Harvey makes the point very well: "The New Urbanism assembles much of its rhetorical and political power through a nostalgic appeal to 'community' as a panacea for our social and economic as well as urban ills. . . . (H)arking back to a mythological past carries its own dangerous freight" (Harvey, 1997, 68-9). Duany himself argues that NU communities make American society and human behavior better in three ways: i. making life richer for children; ii. allowing one to age in place (not so much by creating nearby housing opportunities for empty nesters but by making pedestrian mobility possible); and iii. eliminating the need for more than one car (Harvard Design Magazine, Winter/Spring 1997, 53, 55). But the first two are attainable in a wide variety of urban and suburban residential environments, while the third has not been achieved in the sense that NU community residents have similar automobiles per household ratios to households elsewhere. The explanation of this last point is obvious: the accessibility and mobility needs of individuals cannot be satisfied by constraining them to inside the community, at least within walking distance.

Top vii. Tripmaking.
This brings us to a major claim of the New Urbanists is that their proposals will lead to major changes in travel behavior: reduced automobile dependence, more transit use, increased cycling, and pedestrian-friendly development. Unfortunately for them, there is little justification for these claims. A high proportion of trips is external to the community (for instance, almost all jobs are outside), and cars remain vitally necessary for mobility. No significant transit services have been developed to link NU communities with nearby centers, for example, the plans for a transit system to link Laguna West with Sacramento (less than 10 miles away) never materialized. The majority opinion is that the NU communities will never be dense enough or large enough to justify significant (i.e. frequent) transit service (see Downs' [1994] critique of Calthorpe [1993]). Duany himself admits that market preferences, heterogeneous housing demands and the open space provisions that drastically reduce gross compared with net residential densities result in relatively low densities compared with transit-oriented neighborhoods. Careful analysis of the tripmaking impacts (Crane, 1996) suggest that it is unclear whether higher density communities will result in more auto trips or less. The limited scope of retail and other consumer services in NU communities (typically, one shopping center at best) means that even within these communities most services are beyond the average American's tolerance for service-oriented walking, i.e. between one-quarter and one-half mile. The NU communities often lend themselves to comfortable cycling, but bicycles remain a niche travel mode, at least for Americans.

Top Visiting Laguna West
One of the authors recently visited Laguna West, the New Urbanist community designed by Peter Calthorpe a few miles south of Sacramento, California. It was not a research visit per se, but an opportunity to look at the physical characteristics of the site and to chat informally with some of the residents. The community is a poor reflection of New Urbanist principles. It contains many large houses on sizeable lots (perhaps a half acre), it is riddled with cul-de-sacs (a New Urbanist no-no), there are negligible traffic-calming measures except for tree planters on some of the side roads, and the main road through the community (Babson) is a race track. The main park (an approximation to a medieval village green) was deserted at 9 a.m. on a Saturday morning, although 30 minutes later the AYSO (American Youth Soccer Organization) teams began to turn up. The commercial center is pitiful with a small market, a couple of coffee shops, and a few other services (e.g. a dentist, a health store, a restaurant, a market, etc ). The office/industrial complex is very small, largely consisting of an Apple facility. There are a few bus shelters scattered through the community, but while we were there no bus was in sight. With respect to the vaunted social interaction and communitarian goals of New Urbanism, the residents we spoke to knew few people beyond their immediate neighbors, although there were some relatively well attended concerts in the park. On the positive side, some of the houses incorporate New Urbanist design elements, such as front porches and garages at the back. Even more important, Laguna West includes a variety of housing styles, including moderately priced duplex-type attached for-sale units (in the $120,000 range) and a senior rental complex (in the final stages of completion). In general, however, Laguna West loooks little different from the standard residential community, especially with respect to automobile ownership and use; five cars in the driveway (not on a party night) is a case in point.

Top Part II: The Role of Agglomeration Economies: Beyond the New Urbanists' Focus

Rather than tampering at the margins of urban problems (as we believe New Urbanists do), we prefer to focus our attention on a more significant issue from the perspective of future urban form: what is happening to agglomeration economies. A major rationale for this is that many of the benefits of urban settlement patterns such as pecuniary economies (higher wages and land values) or consumption externalities (e.g. more choice of restaurants, major ball clubs, concert halls, teaching hospitals) reflect agglomeration economies. The anti-sprawl advocates, among whom the New Urbanists are a vocal subgroup, have paid excessive attention to the costs side of alternative settlement patterns, a picture - by the way - that remains very blurred.

Agglomeration economies distinguish urban economics from the rest of economics. Even models of residential location begin by assuming an employment center. Yet, why are there such centers? The answer usually has to do with agglomeration economies. Vernon (1960) suggested that these economies are especially important for small start-up firms that are unable to benefit from scale economies. Such firms would, therefore, most likely locate in the cores of urbanized areas, trading off high land costs for external agglomeration economies. Such firms would, therefore, most likely locate in the cores of urbanized areas, trading off high land costs for external agglomeration benefits.

More recently, "mainstream" economists have discovered spatial economics. The Journal of Economic Perspectives (Spring, 1998) included a Symposium on urban agglomeration. In it, Krugman elaborated on the agglomeration idea, noting that now that theorists find increasing returns tractable, they are better able to apply their modeling skills to spatial questions. In the same Symposium, Quigley (1998) argues that the current interest in agglomeration matches the New Growth Theory's interest in the external economies provided by new ideas. Cities' "internal heteorgeneity and diversity" obviously play a role. The Symposium's third author, Glaeser, pays attention to the external costs as well as the external benefits to be found in cities but, we believe, would have developed a more powerful argument had he differentiated between what transpires in city centers vs their suburbs (see Correspondence by Gordon and Richardson, Journal of Economic Perspectives, forthcoming). In another Symposium, Bollinger et al (1998) report on a case study which demonstrates that access that allows face-to-face contacts is a significant determinant of office rents, suggesting that electronic access is not (yet) a good enough substitute (for enough people).

Models of city size emphasize the trade-off between the benefits of agglomeration (increasing with city size but at a decreasing rate) and congestion costs (also increasing with city size but at an increasing rate). This approach suggests an "optimal" city size, where marginal benefits of city size are equated with marginal costs. City size could be defined in many ways, e.g. area and population, but as long as the variables are highly correlated it does not matter much.

Dynamic treatments emphasize how both curves shift over time. Falling transportation costs shift the congestion curve outward, making it flatter and making the city larger. Falling communications costs may cause some of the benefits of face-to-face contact to become available to individuals that have electronic access to each other. Then this curve too becomes flatter. Both changes suggest a larger metropolis.

Dispersed employment and other opportunities flatten most price gradients. Growth, then, may not raise housing prices. Greater affluence prompts more consumption of space which suggests that spatial growth exceeds population or employment growth. This is part of the explanation of the accelerating spread of metropolitan areas.

Easterlin (1998) has discussed the importance of increasingly "footloose" industries. Whereas workers were once compelled to locate in the vicinity of industries that sought mineral-rich sites or places offering peculiar transportation advantages (ports, rail sidings, crossroads, etc.), footloose locators are now able to follow the labor force. With this new freedom, workers and their families choose suburban high-amenity low-rent sites. The historic pattern has been reversed. If so, does this mean that agglomeration economies are more ubiquitous or that they are less important? Of course, if they ever really become ubiquitous, they would no longer be important.

Aside from the obvious ("what are we measuring?"), the discussion of agglomeration economies prompts two questions:

    i. how important are they? and
    ii. where are they?

In a recent paper (Gordon, Richardson and Yu, 1998), we approached these questions by utilizing BEA county-level employment data (REIS file) for the period 1969-1994. Rather than patterns of "rural renaissance" (some writers' allusion to a seeming reversal of urbanization in the 1970s) vs. "urban revival" (one writer's demonstration that the 1970s reversal had itself reversed), we found that private sector job growth swings were really between exurban and rural dominance (1969-77 and 1988-94) and suburban dominance (1977-88). It is now possible to extend the discussion because two more years of REIS data are available. Non-metropolitan job growth was dominant through 1995 whereas metropolitan growth was greater for 1995-96 (Figures 1a, 1b and 2).

Tables 1a and 1b show these trends for the (roughly defined) "sunbelt" and "frostbelt" regions. The patterns across the two regions are similar. Among metropolitan counties, most growth occurred in the suburbs of the large metropolitan areas in the years 1977-89. In more recent years, there was high sunbelt growth in these counties as well as in exurban adjacent counties; the highest frostbelt growth took place in the non-adjacent exurban counties.

Can these findings elaborate Vernon's idea? Can the changing location of rapid job growth suggest anything about the location of agglomeration benefits? One way to approach these questions is to ask where growth takes place during periods of economic expansion when there are more business starts. Our previous results suggest that in the most recent years, these economies exist in the outer suburbs because, roughly speaking, periods of high employment growth are associated with growth in the outer suburbs whereas periods of lower employment growth are associated with faster rural and exurban growth. If these findings hold up, they support our doubts about the assault on suburbanization ("sprawl") by the New Urbanists and others, on the ground that it undermines the dominant characteristics of economic growth (by the way, many New Urbanists would regard this as a good thing).

Direct tests of the Vernon hypothesis would require the availability of data on business starts by location. The only business starts data known to us are by Dun & Bradstreet for the U.S. as a whole. These only cover the years 1985-1996 because the D&B pre-1985 data are not compatible with the newer series. Over the 1985-96 period, business starts per total number of businesses (corporations plus partnerships plus nonfarm proprietorships) peaked in 1985, reached a low in 1991, and rose again into the mid-1990s. Yet, recent business starts do not match the levels of the mid-1980s (Table 2). While these observations are from a series that is too short to generate robust conclusions, they do corroborate the hypothesis. Metropolitan-area private sector employment growth dominated when business formations were strong. And this metropolitan-area growth was strongest in the suburbs. These occurrences and the Vernon hypothesis lead us to conjecture that agglomeration economies were most attractive in the suburbs. If cities want to prosper then employment growth in their suburbs is to be welcomed. If suburban expansion is inevitably linked with sprawl, there are clearly serious risks in anti-sprawl actions.

Top Part III: Conclusions

What has this discussion of agglomeration economies to do with New Urbanism? First, the insignifance of NU or even quasi-NU communities in terms of population absorption implies that if they are regarded as a solution to metropolitan problems, all the cliches (shuffling the deckchairs on the Titanic, fiddling while Rome burns, etc.) come to mind. We can disagree about the scale and type of interventions needed to address central city economic and social problems, but certainly New Urbanism as judged by its implementation hitherto is irrelevant.

Second, the real problem for the central cities and the stability of its economic base is how to counter or accommodate the obvious decentralization of agglomeration economies to the suburbs and to edge cities (Garreau, 1991), or even the possibility that such economies are disappearing (Gordon and Richardson, 1996). New Urbanism has nothing to contribute to this discussion, in part because it believes, at least implicitly, that social problems are remediable by architectural and design prescriptions rather than by economic development. Third, while accepting the CNU argument that the metropolitan region is the appropriate unit for analysis, it is difficult to find any concrete details in the New Urbanist discussions as how they will influence the future metropolitan region. Agglomeration economies have much to contribute to such a discussion because what happens to them could affect the survival of the central city, the role of the suburbs, and the potential of exurban and rural locations. As Robert Campbell ponted out in the Harvard Design Magazine symposium, with some exaggeration, that we are "entering an era during which half the population will move to Montana and the other half will move downtown, because those are the two best places to live, and new communications technology lets you live anywhere. So what happens to the suburbs? Given this larger framework of settlement patterns, what happens to Kentlands?" (Harvard Design Magazine, 1997, 50). Duany's reply: "Kentlands will be resilient" (Ibid., 50). Yet what happens to metropolitan regions, and whether Campbell's prediction makes sense, depends on how the role of agglomeration economies is transformed by information technology and other developments. On these issues New Urbanism is silent.

Top References

Bartik, Timothy J. (1989), "Small Business Start-Ups in the United States: Estimates of the Effects of Characteristics of the States" Southern Economic Journal, 55:4, 1004-1018.

Bollinger, Christopher, Keith Ihlandfeldt and David Bowes, (1998) "Spatial Variation in Office Rents within the Atlanta Region" Urban Studies, 35:7.

Calthorpe, Peter (1993), The Next American Metropolis. New York: Princeton Architectural Press.

Crane, Randall (1996), Cars and Drivers in the New suburbs: Linking Access to Travel in Neotraditional Planning," Journal of the American Planning Association, 62(1), 51-65.

Downs, Anthony (1994), A New Vision for Metropolitan America. Washington, D.C.: Brookings Institution.

Easterlin, Richard A. (1998), "Twentieth Century American Population Growth" in S. Engerman and R.E. Gallman (eds.) The Cambridge Economic History of the United States, Vol III. The Twentieth Century. Cambridge: Cambridge University Press.

Garvin, Alexander (1998), "Is the New Urbanism Passe?" Lusk Review, 4(1), 12-32.

Glaeser, Edward L. (1998). "Are Cities Dying?" The Journal of Economic Perspectives, 12:2, 139-160.

Gordon, Peter and Harry W. Richardson (1998), "Farmland Preservation and Ecological Footprints: A Critique," Planning and Markets, 1, .

Gordon, Peter and Harry W. Richardson (1996), "Beyond Policentricity: Los Angeles, the Dispersed Metropolis" Journal of the American Planning Association, 62:3, 289-95.

Gordon, Peter and Harry W. Richardson and Gang Yu (1998), "Metropolitan and Non- metropolitan Employment Trends in the US: Recent Evidence and Implications" Urban Studies, 35:7, 1037-1057.

Harvard Design Magazine (1997), "New Urbanism: Urban or Suburban?" Winter/Spring, 47-63.

Harvey, David (1997), "The New Urbanism and the Communitarian Trap," Harvard Design Magazine, Winter/Spring, 68-9.

Jacobs, Jane (1992), The Death and Life of Great American Cities. Vintage Books.

Katz, Peter (1993), The New Urbanism: Toward an Architecture of Community. New York: McGraw-Hill.

Krugman, Paul (1998), "Space: The Final Frontier" The Journal of Economic Perspectives, 12:2, 161-174.

Quigley, John M. (1998), "Urban Diversity and Economic Growth" The Journal of Economic Perspectives, 12:2, 127-138.

Vernon, Raymond (1960), Metropolis 1985. Cambridge: Harvard University Press.

Top | Home