BEYOND EDGE CITY: OFFICE GEOGRAPHY IN THE NEW METROPOLIS1
Robert E. Lang2 Brookings Mountain West University of Nevada, Las Vegas Thomas W. Sanchez Department of City and Metropolitan Planning University of Utah Asli Ceylan Oner School of Urban and Regional Planning Florida Atlantic University
Abstract: Few studies focus on the distribution of office development across the United States as an element of urban form. Joel Garreau argues that all office space found outside central-city downtowns is grouped into large, mixed-use clusters he termed “edge cities.” The reality is that edge cities capture only a modest share of non-CBD office space, with a significant proportion located in what Lang referred to as “edgeless cities.” This study reexamines the distribution of edge and edgeless cities using a spatial analysis of more than 3 billion square feet of office space in 13 large U.S. markets. The office patterns revealed do not fit a downtown versus edge city development model, but instead show complex and regional variation. The findings confirm the “edgeless” nature of U.S. office development. Although primarily descriptive, this paper highlights issues of urban form from a holistic and novel perspective. [Key words: urban form, office space, suburban sprawl.]
INTRODUCTION This study extends the analysis of office geography completed by Lang (2003b), in which he examined the distribution of office development in the 13 largest markets in the United States. Observers such as Joel Garreau (1991), author of the influential book Edge City, speculated that all of the office space found outside central-city downtowns could be grouped into large, mixed-use clusters that he termed edge cities. The reality is that edge cities capture only a modest share of non-CBD office space. Lang characterized this pattern by what he termed edgeless cities, a dispersed urban form that accounted for almost double the office space located in edge cities. Lang also referred to this pattern as “office sprawl,” more in the sense of its spatial dispersion than its architectural esthetic (2000).
1
The authors are grateful to Eric Leinberger, University of British Columbia, for his help with the cartography. Correspondence concerning this article should be addressed to Robert E. Lang, Brookings Mountain West, University of Nevada, Las Vegas, Greenspun Hall, 4505 S. Maryland Parkway, Las Vegas, NV 89154-3007; telephone: 571-296-1033; email:
[email protected]
2
726 Urban Geography, 2009, 30, 7, pp. 726–755. DOI: 10.2747/0272-3638.30.7.726 Copyright © 2009 by Bellwether Publishing, Ltd. All rights reserved.
BEYOND EDGE CITY
727
Lacking specific form, edgeless cities exist in multiple shapes, sizes, and densities, and are situated in similarly diverse settings. Lang noted that edgeless cities occur throughout the metropolis—from an urban region’s edge to near its historic core, and in a variety of residential contexts. Perhaps more importantly, he also argued that edgeless cities were not edge cities waiting to evolve but were instead a less clustered form of commercial development that often competed with edge cities as well as other urban locations. This study seeks to better understand the scope and context of edgeless cities as well as operationalize a new geography of metropolitan office location through a spatial analysis of office building activity. CBDs (henceforth referred to as downtowns), edge cities, and edgeless cities are redefined using a method based on the composition and clustering of office development. This methodology, and the new geographic categories it produces, offers an interesting spatial depiction of this nonresidential land use in selected large U.S. metropolitan regions. This research offers a new perspective on intrametropolitan form because it not only focuses on nonresidential components, but also on the residual urban space after traditional commercial clusters are deducted. THE NEW SUBURBAN METROPOLIS The largest share of new office space is now located in the suburbs (Lang, 2000).3 As the suburban ring has acquired significant commercial development, it has undergone a change in character. In the United States, the suburbs now contain more than half of the population, or just over 150 million people as of 2005, and are increasingly diverse (Lang et al., 2005). They are also assuming a leading role in defining the metropolis. Suburbs now have business districts that rival those of central cities, as is the case of Southfield, Michigan, which today contains more office space than downtown Detroit (Lang, 2003b). To claim that there is a new suburban metropolis is not to assume that some 1950s-era housing subdivision suddenly dominates the region. Rather, as suburbs boomed in the last third of the 20th century, they underwent quasi-urbanization to become a new form of city—or more accurately a “new metropolis.” The suburbs are now far more urbanized and diverse than they were only a generation ago (Lang and LeFurgy, 2007). Suburbs now contain a larger share of the nation’s poverty (Swanstrom et al., 2004), their share of single-person-households continues to grow (Frey and Berube, 2003), and they increasingly house the elderly (Frey, 2003). Although many suburbs possess most of the elements that make a place urban, they often differ enough from traditional cities to sustain the popular depiction of suburbia as home to affluent, White-majority households (Lang and Blakely, 2006). Researchers have documented profound changes in suburban composition, but the manner in which suburbs have urbanized in a looser and more automobile-dependent manner than central cities means that it may take some time until their transformation registers in the popular
3
The U.S. Census Bureau provides no official definition for a “suburb.” Those who use the term have conventionally assumed that suburbs constitute the portions of metropolitan areas that lie outside central cities (Katz and Lang, 2003).
728
LANG ET AL.
culture. Realtors, developers, and planners need to recognize the new suburban reality because it represents new patterns of demand for goods and services. This study seeks to emphasize the changing role of the suburbs using office space as an indicator of urban restructuring. Since the beginning of the 1960s, there has been a shift in the U.S. economy toward the service sector and office employment. Pivo (1990) found that, between 1959 and 1990, the nation’s office inventory almost quadrupled and that office space is a prominent and defining element of contemporary metropolitan form. This produced a “new metropolis” that is urban in function but not in form (Fishman, 1990; Lang, 2003b). It has all the elements of traditional urban areas, but the elements are reassembled in a way that appears more spread out, characteristically suburban. A mixture of land uses exists, but almost never at a concentration level that looks and feels urban (Lang and LeFurgy, 2007). As the new metropolis evolves, integrating the mixeduse components in a more cohesive manner is one aim of the Smart Growth and New Urbanism movements (see Duany et al., 2000). Some places, including several highlighted below, coalesce and become more pedestrian-friendly. Nonetheless, the question remains whether these environments are mainly islands of urbanity in seas of sprawl, or indicative of the dominant suburban form. At the moment, the new metropolis exhibits mostly low to middle densities, automobile dependency, and considerable dispersion. It may possess large-scale edge cities (Garreau, 1991), but more often it contains smaller and more spread out edgeless cities (Pivo, 1990; Gordon and Richardson, 1996; Ding and Bingham, 2000; Lang, 2000, 2003b). Beltways, highway retail strips, single-family-housing subdivisions, garden apartments, and low-rise office structures typify this landscape. The new metropolis sprawls in so many complicated ways that, according to Hayden (2004), it needs a “Field Guide.” To Fishman (1987, p. 189), the new metropolis (or his analogous “technoburb”) “lacks any definable borders, a center or a periphery, or clear distinctions between residential, industrial, or commercial zones.” Fishman (1987, p. 203) further notes that technoburbs spill over vast space and that they “can best be measured in counties rather than city blocks.” At the same time, Pivo (1990, 1993) was among the first researchers to track suburban office sprawl to describe the new metropolis. According to Pivo (1990), the sprawl pattern of office space in the new metropolis can best be described as a “net-ofmixed beads.” Most office space is located outside of CBDs in clusters of various sizes or randomly scattered across the suburban ring. The largest and densest clusters are found at the most accessible locations, especially freeway junctions (Pivo, 1993). Thus, even as Garreau was publishing Edge City, academic observers such as Pivo were already examining an office landscape that defied Garreau’s simple polycentric model. Why Follow Office Space? Tracking office space trends provides a useful perspective for understanding metropolitan change because offices are where a large proportion of job growth occurs. In some metropolitan areas, nearly half of all newly hired employees go to work in office buildings (U.S. Bureau of Labor Statistics, 2003). Office buildings are also the last major element of central cities to suburbanize, following residential, retail, and other commercial activities (Muller, 1980; Leinberger, 1996).
BEYOND EDGE CITY
729
The rapid growth of the suburban office industry is so significant a trend that it sparked a series of books and articles in the late 1980s and early 1990s that sought to explain how it was reshaping the American metropolis (see, for example, Baerwald, 1982; Erickson, 1985; Baldassare, 1986; Cervero, 1986, 1989; Fulton, 1986; Hartshorn and Muller, 1986, 1989; Leinberger and Lockwood, 1986; Fishman, 1987, 1990; Romanos, et al., 1989; Pivo, 1990, 1993; Garreau, 1991; Bruegmann and Davis, 1992). Some identified this trend as early as the 1970s (Greenhouse, 1971; Herbers, 1971; Kneeland, 1971; Breckenfeld, 1972; Daniels, 1974; Vance, 1977). The location of office space is critical in a number of public policy areas. For example, the distribution of new office space can help determine the extent to which there is a jobs/ housing balance in a metropolis (Cervero, 1986; Orfield, 2002). It can also influence the spatial mismatch between economic opportunity and the concentration of minority households (Kain, 1992). Moreover, office location also impacts urban sprawl. If most new office space is constructed at the urban periphery, it extends commutersheds for many miles into semi-rural areas and thereby fuels sprawl (Ding and Bingham, 2000). Finally, the geography of office location figures prominently in urban transportation analysis: if most new office space is built in areas without public transit access, then reliance on automobiles will continue to grow (Cervero, 1989). Whereas office location data are important indicators of metropolitan change, they cannot convey the complete picture. This article reports on only a portion of the whitecollar employment (albeit a significant share) in 13 of the nation’s largest metropolitan areas. Although retail space is not specifically tracked, the presence or absence of large regional malls amid office development is noted. Unfortunately, we know little about the business being conducted in the office spaces we analyze, assuming that their building types and classifications imply significant agglomerations of professional activity and economic productivity. Mapping the distribution of office space helps identify these new submarkets. Most commercial real estate firms have delineated office markets; however, most are still unsystematically derived from traditional categories. The metropolitan submarkets that many real estate firms use in their research often reflect local convention and typically mix edge cities and edgeless cities together. Given that suburban property is now the target of most real estate investment, it makes sense to update categories to better gauge suburban investment opportunities. The only current market indicator that separates types of non-CBD office development is the Urban Land Institute’s “Suburban High-Rise and Low-Rise Office Rent Forecasts.” The Meaning of Downtowns, Edge Cities, and Edgeless Cities Monitoring the new metropolis is not simply a geographic exercise, although the spatial configuration of office space is significant. One longstanding concern is the loss of office space from the downtown CBD to the suburbs and the role this shift plays in promoting metropolitan decentralization (Hartshorn and Muller, 1986, 1989; Muller, 1980, 1981). Advocates and critics of edge cities have long debated what their emergence means for the future of the city (Garreau, 1991; Pierce, 1991; Barnett, 1992). More recently, even some critics of the suburbs accept a polycentric urban form, as evidenced
730
LANG ET AL.
by the theme of the 2005 Congress for New Urbanism (CNU) on the “polycentric city.”4 Lang (2003a, 2003b), however, showed that office development in many places can be more accurately described as “post-polycentric,” no longer falling into well-defined nodes.5 A key assumption made by both sides in the edge city debate is that size—meaning a large concentration of office and retail space—is a prerequisite for city building (Goldberger, 1987; Cervero, 1989; Garreau, 1991; Bruegmann and Davis, 1992). The minimum for edge cities, established by real estate analyst Chris Leinberger (1988, 1996, 1998), was 5,000,000 square feet (s.f.) of office space and 600,000 of retail in a mostly cohesive cluster. Garreau (1991) popularized these criteria, and this definition has taken hold in the literature. In fact, Lang (2003b) and again in this study uses a reduced criterion for an edge city based on office space alone. But the edge city requirements are hardly definitive, and more importantly they may be deceptive and lack empirical support. Only in a limited number of places outside a CBD do the elements of a full-fledged edge city actually come together. Lang (2003b) estimated that the amount of office space that lies in a true office/retail–defined edge city is well under five percent of all space nationwide. Yet that is not necessarily a reason for the CNU and others to despair. Commercial clusters in the suburbs may require much less office and retail space to actually function as urban places. In fact, the smaller scale may also help office and retail activity mix more successfully with residential development. This analysis considers the issue of making edgeless cities more traditional city-like environments, and will include data and case studies of the conditions that shape this outcome. THE NEW OFFICE GEOGRAPHY Lang (2003b) developed four basic categories of office location: downtowns, secondary downtowns, edge cities, and edgeless cities. Building on that framework, improved data resources and spatial analytical techniques have helped to identify several additional subtypes. Downtowns and secondary downtowns are supplemented with an “urbanenvelope” category that captures the zone of office development that does not quite fit into the CBD, but lies nearby and is often found in high-density, mixed-use centers. Edge cities also feature a subtype—the corridor—that captures office areas which form along freeways or other major transportation routes. Edgeless cities are divided into three subcategories—high-, medium-, and low-density, depending on the intensity of surrounding residential development. We now describe these geographies and include the underlying assumptions and justification for each.
4 The meeting was held in June 2005 in Pasadena, California. In the 2000 version of its “Charter,” the CNU described edge cities as “a term which [sic] implies urbanism but is in fact only a statistical agglomeration of housing, subdivisions, shopping centers, and office parks.” 5 Descriptions of post-polycentric urban form have an extensive literature. For instance, see Lessinger (1962), Lewis (1983, 1995), Lynch (1960), Pressman (1985), and Webber (1964).
BEYOND EDGE CITY
731
Downtowns Downtowns are found in the older, core areas of metropolitan regions. The “primary downtown” or CBD lies at the center of the region and is the original site of significant commercial development. In some cases, such as Manhattan, the CBD may be split between an original downtown and a nearby midtown that emerged in the 20th century. In such instances, the two downtowns are combined into one general primary downtown. Other examples include the downtown and midtown areas of Atlanta, and the Loop and Northside sections of Chicago. What are referred to here as “secondary downtowns” form the centers of major suburbs and uptowns within a central city that developed a relatively modest, though focused, commercial center early in the 20th century. Secondary downtowns are scaleddown, slightly less dense versions of primary downtowns. They have their origin in the streetcar and early automobile era and as such support some pedestrian presence. These areas often developed in close proximity to upscale housing, as managers shifted business locations closer to their homes to shorten commutes. Examples of secondary downtowns include the West Side of Los Angeles and Bethesda, MD just outside northwestern Washington, DC. Urban Envelopes Urban envelopes surround downtowns in cases where office density declines with distance from a major activity cluster. In some regions these envelopes form a tight band of space that only slightly extends the downtown office market. In other places the envelope forms a wide ring around the core that encompasses much of the central city as well as contains some inner suburbs. Several metropolitan areas exhibit no urban envelope, which is to say that office development fits into a downtown without spillover. Residential areas that lie within urban envelopes are densely populated—often with more than 5,000 people per square mile. The term “urban envelope” is now most commonly used in Canadian planning circles and typically refers to the residential areas that surround downtown. The term has its origins in the 1960s and early 1970s “factorial ecology” analyses (Berry and Horton, 1970), which created a typology for all urban neighborhoods using multivariate methods. Urban envelopes are specifically defined in this study as being contiguous to primary downtowns, extending outward to office space densities averaging 2,000,000 s.f. per square mile. Edge Cities Edge cities have a specific definition that appears in the literature on suburban office development. Whereas Joel Garreau coined a unique term, real estate consultant Christopher Leinberger (1984) had earlier used the same criteria to identify what he referred to as “urban villages.” Because Garreau used Leinberger’s criteria for his own research, this study relies on the office space elements of their descriptions to define edge cities.
732
LANG ET AL.
Garreau asserts that edge cities are dense relative to their suburban surroundings. Leinberger also emphasized density when describing urban villages. He depicted urban villages as: “business, retail and entertainment focal points amid a low-density cityscape,” adding that “each Urban Village has a core—a kind of new downtown—where the buildings are the tallest, the daytime population largest, and the traffic congestion the most severe” (Leinberger and Lockwood, 1986). The number of edge cities varies per metropolitan area, with larger office markets typically containing several. In this study, we define edge cities as non-CBD clusters with more than 5 million square feet (with some exceptions) of office space. Were we to include the rest of the criteria, especially the one concerning substantial retail space, the amount of office space found in edge cities would have dropped significantly. Note also that Garreau’s (1991, p. 5) criteria include the fact that edge cities “were nothing like a ‘city’ as recently as thirty years ago.” That would mean that dozens of edge cities would have to be excluded from his analysis, because places such as Midtown Atlanta, Jersey City, NJ, and even Arlington, VA—all edge cities according to Garreau— were all very much like a city in 1960 (or 30 years prior to when Garreau made this statement). Edge City Corridors Edge city corridors have long been noted in the literature (Berry, 1959; Baerwald, 1982; Sternlieb and Hughes, 1988). Not every metropolitan area exhibits these types of corridors. For the purposes of this study, we define corridors as linear clusters that typically follow major transportation routes (freeways or major arterial highways), averaging edge-city office densities, and anchored by one or more secondary downtowns or edge cities. Edgeless Cities Edgeless cities, along with edge cities, identify a major subset of non-CBD office space. As the term implies, edgeless cities lack a well-defined boundary or edge. Edgeless cities extend across tens, if not hundreds, of square miles of urban space. Individual components of edgeless cities often have an identity, but collectively these places seldom strike a casual observer as unified in any meaningful way. Thus, unlike edge cities, edgeless cities are not perceived as one place. In this study, edgeless cities are specifically defined as the balance of a metropolitan region’s office space outside of primary downtowns, secondary downtowns, urban envelopes, edge cities, and edge-city corridors. Table 1 summarizes the characteristics of all six categories of office location that are used in this study. METHODOLOGY The data for this study came from Black’s Guide to Office Leasing, a nationwide directory of office space. Black’s Guide lists buildings by address and those with 15,000 square feet or more that exist, are under construction, or are proposed (Black’s Guide, 2005). Black’s Guide surveys even the smallest suburban office markets, making it possible to compare data across metropolitan regions. Buildings are listed in the publication
733
BEYOND EDGE CITY
TABLE 1. METROPOLITAN FORM CHARACTERISTICS Category
Scale
Office density
Basic units
Boundary
Downtown
Concentrated
Very high to high
City blocks
Sharp, well defined
Secondary downtown
Concentrated
High to medium
City blocks
Mostly well defined, but some soft edge
Urban envelope
Buffer/ring
Low
Neighborhoods
Fuzzy, but with a recognizable edge
Edge city
Concentrated
Medium to low
Freeway interchanges
Fuzzy, but with a recognizable edge
Corridor
Linear
Low
Freeway frontage road
Fuzzy, but with a recognizable edge
Edgeless cities
Dispersed
Very low
Municipalities or counties
Indeterminate, very difficult to define
at no cost to owners or developers, and the guide is distributed free to companies and institutions involved in office leasing. The 13 Office Markets This analysis focuses on the spatial structure of office space in 13 major metropolitan markets distributed throughout the United States. Six are in the Northeast and Midwest and seven are in the South and West; collectively they contain nearly one-third of the U.S. population. The urban regions studied are: Atlanta, Boston, Chicago, Dallas, Denver, Detroit, Houston, Los Angeles, Miami, New York, Philadelphia, San Francisco, and Washington, DC. The office space in this study includes leased multi-tenanted buildings that exclude such other major employers as government offices, warehouses, flex space (offices combined with light manufacturing), hospitals, and universities. This analysis does not include small office buildings that are used for professional services such as dental clinics. Rather, offices found in Black’s Guide house producer service firms that serve regional, national, and global clients. These offices also contain many corporate headquarters and regional offices of multinational businesses. Mapping Methods Density maps were created within a geographic information system (GIS) to identify the clustering of office space in the 13 metropolitan areas. To do this, office building locations were first geocoded (address-matched) using street address information obtained from Black’s Guide. After geocoding the office building data (as point locations), density maps were generated for each metropolitan area. Density maps represent point data as continuous value surfaces that show the distribution of office space such that areas of high concentration appear as clusters relative to the surrounding metropolis. The density maps were then converted to contour maps with isolines of office density
734
LANG ET AL.
delineating office concentrations at predefined thresholds that characterize the 6 urban and suburban geographies previously described. Thresholds were identified as follows: • Primary downtowns are distinct clusters central to a metropolis that have office space densities averaging 7,500,000 s.f. per square mile. There were no clusters above this threshold in downtown Miami and Detroit, but because CBDs do indeed exist in these places, an exception was made and their central office clusters were counted as downtowns. In the case of New York, Black’s Guide did not provide complete data for Manhattan, so a raw number for office space was provided from Collier’s February 2005 estimates (Colliers International, 2005). • Secondary downtowns are office clusters geographically distinct from primary downtowns that have office space densities averaging 3,500,000 s.f. per square mile. • Edge cities are office clusters geographically distinct from primary or secondary downtowns averaging office space densities of 2,000,000 s.f. per square mile. • Urban envelopes are contiguous to primary downtowns, extending outward to edge-city office space densities averaging 2,000,000 s.f. per square mile. • Corridors are linear clusters that usually follow major highway routes, averaging edge-city office densities and anchored by one or more secondary downtowns or edge cities. Office corridors exist in Washington, DC (along I-270 and the Dulles Toll Road), Dallas (from the primary downtown to the secondary downtown to the north), Houston (along the Katy Freeway), and Los Angeles (along Wilshire Blvd). • Edgeless cities are the balance of metropolitan regions outside of primary downtowns, secondary downtowns, urban envelopes, edge cities, and corridors. Edgeless cities have office densities averaging approximately 20,000 s.f. per square mile. Residential Density Measure Residential density was calculated for all office clusters identified in this study. As expected, downtowns, secondary downtowns, and urban envelopes had the highest densities. The residential density in edge cities varied considerably among regions as well as between individual locations within the same metropolitan area. The residential density measure most importantly applies to edgeless cities and was developed to show the context for dispersed office development. Lang (2003b) noted that many regions that maintained a high share of edgeless city office space, such as Miami, also exhibited an overall densely built environment marked by high population density. Thus office space could be scattered yet fit into a relatively dense urban fabric. This suggests that office space could conceivably anchor mixed-use developments at a sub-edge city threshold. Calculations of population density for edgeless cities were performed using census tracts with estimated population data for 2004. The breakpoints in population density are based on a geographic method developed by Nelson and Sanchez (2005) to track land use
BEYOND EDGE CITY
735
change and settlement patterns in the nation’s 50 largest metropolitan areas. Nelson and Sanchez determined that 3,000 people per square mile—or three times the population density of the Census Bureau’s “urbanized area” measure—is sufficiently dense to support urban services, and is therefore “urban.” For example, light rail projects are viable at this threshold. Note that much of the area that qualifies as urban in this study is actually suburban (or, in census definition terms, in areas “outside primary cities”). Lang et al. (2005) showed that many of these areas essentially function as cities, despite being referred to as suburban. An example of an area that exceeds 3,000 is the eastern part of Fairfax County, Virginia along the Washington Beltway. Land market pressures in such built-up, and often built-out, older suburbs are changing the pattern of land development, resulting in a greater share of high-density multifamily housing. Such housing, mixed with office and retail activities, and especially when combined with light rail (as in the north side of Dallas), represents a nascent, small-scale form of urbanism. Edgeless city office space in residential areas that have from 1,000 to 3,000 people per square mile are labeled “suburban” in this study. This density range still represents a moderately intense development level, especially compared to still developing parts of the metropolitan fringe. However, most housing developments in this range consist of single-family detached dwellings. As noted above, the census uses the 1,000 person per square mile threshold as the basis for its “urbanized” measure. There are regions in this study, such as Los Angeles and Miami, in which virtually the entire population lives at densities greater than the urbanized minimum (Lang, 2002). Other metropolitan areas, especially Atlanta, have a significant share of people residing outside the urbanized zone. Edgeless city office development found within suburbs, as defined by this study, is not likely to anchor small urban islands as is possible at a higher residential density. Finally, edgeless city offices found in places with less than 1,000 residents per square mile fall in the “exurb” category. Fulton (1986) dubbed such developments “the office in the dell.” Edgeless cities on the suburban fringe sit amid large-lot subdivisions and farmlands. Not only is there no prospect for establishing an urban feel in these office environments: they may further propel decentralized development by extending the commutershed deep into rural areas. RESULTS Overall, the 13 markets in this study have over 3 billion square feet of office space (see Table 2 and Appendix 1). Downtowns contain close to a billion square feet of this space, with their urban envelopes containing more than 150 million additional square feet while secondary downtowns add another 131 million square feet. Edge cities have a total of nearly 424 million square feet of office space; corridors add another 117 million. Edgeless cities have a massive 1.2 billion square feet of office space, with approximately half at “urban” residential densities. As a result, about 600 million square feet of office space (or one-fifth) lies in low-density suburbs and exurbs. This finding is encouraging in that it dispels the notion that most edgeless space is found at the metropolitan fringe. In fact, half of all edgeless city space is embedded in a relatively dense metropolitan fabric. Also, note that this figure would have been higher had Black’s Guide surveyed the Bay Area’s Silicon Valley, which is densely built and essentially one large edgeless city.
736
LANG ET AL.
TABLE 2. TOTAL DISTRIBUTION OF OFFICE SPACE IN THIRTEEN METROPOLITAN AREAS Location
Square feet
Percent
Downtowns
981,167,601
32.7
Urban envelope
153,788,481
5.1
Secondary downtowns
131,159,297
4.4
Edge cities
423,975,978
14.1
Corridors Edgeless cities
116,832,371
3.9
1,196,557,572
39.8
Urban
617,202,152
51.6
Suburban
418,461,925
35.0
148,587,938
12.4
Exurban Total
3,003,481,300
TABLE 3. SUMMARY OF OFFICE SPACE IN THIRTEEN METROPOLITAN AREAS (PERCENT OF TOTAL) Downtown
Urban envelope
Atlanta
6.7
38.6
0
16.4
0
38.4
Boston
38.7
4.6
0
6.7
0
50.1
Metropolitan area
Secondary downtown
Edge city
Corridor
Edgeless
Chicago
49.2
0.8
0
10.7
0
39.4
Dallas–Ft. Worth
20.6
0.0
6.5
17.3
21.7
33.9
Denver
27.3
0.0
0
21.9
0.0
50.8
Detroit
17.2
1.6
0
27.1
0
54.1
Houston
22.8
3.3
0
33.9
5.4
34.7
Los Angeles
16.1
0.0
13.6
15.6
8.7
46.0
8.7
0.0
5.0
14.2
0
72.1
Miami New York
54.5
0.0
6.8
6.3
0
32.3
Philadelphia
29.6
2.1
4.0
10.1
0
54.3
San Francisco
39.2
0.3
6.7
14.4
0
39.4
Washington, DC
22.7
18.4
1.5
23.3
11.6
24.0
Lang’s study (2003b; based on 1999 data) showed that downtowns contained the largest share of office space with 38 percent of the total, and that edgeless cities claimed virtually the same at 37 percent. Edge cities contained about one-fifth of the space in the selected major office markets. Secondary downtowns comprised the smallest share of the region, with just six percent of office space. This study identified 13 primary downtowns, 23 secondary downtowns, 58 edge cities, with all other space counted as edgeless cities.
BEYOND EDGE CITY
737
Fig. 1. New York metropolitan region.
The following summarizes a number of interesting individual findings from the study regions. Of the 13, New York had the greatest proportion of office space located in downtown clusters. This is not surprising given the prominence of Greater New York as an international financial center and location of many corporate headquarters (Taylor and Lang, 2005). New York is also the largest in terms of total office space and exhibits a polycentric regional pattern totaling over 600 million square feet of office space (see Figure 1). The New York and Chicago metropolitan areas have by far the largest downtowns, and the leading share of their regional space is located in CBDs (see Table 3). The third largest downtown is in Los Angeles, a city often derided for lacking a center. The smallest
738
LANG ET AL.
Fig. 2. Los Angeles metropolitan region.
downtowns are found in metropolitan Detroit and Miami. Atlanta technically has the smallest downtown, but that is due in part to a statistical artifact of our method. Atlanta has one dense cluster at the heart of its CBD that our GIS analysis determined to be its “downtown.” However, a chain of quasi-downtown spaces extends across Atlanta’s urban envelope, reaching north to the upscale Buckhead district. By contrast, the downtowns in Detroit and Miami more accurately reflect the actual space in the CBD. Only four regions in this study have secondary downtowns. Lang (2003b) found more secondary downtowns, but in this analysis many of these places were determined to exist within downtown urban envelopes. In terms of secondary downtown office space, Los Angeles has a significantly greater proportion than the other regions due to substantial office space clusters in West Los Angeles and near Orange County’s John Wayne Airport (see Fig. 2). Among the study regions, Atlanta has the greatest proportion of office space concentrated in urban-envelope (UE) locations. This in large part stems from the development pattern that connects the downtown core, Midtown, Buckhead, and Perimeter office
BEYOND EDGE CITY
739
Fig. 3. Atlanta metropolitan region.
agglomerations. Nearly 40 percent of Atlanta’s office space is located in the UE shown in Figure 3. It is important to note that the UE does not include office space concentrated in the downtown, which happens to represent only 7 percent of the Atlanta region’s office building space. As noted above, urban envelopes vary greatly by region and account for the overwhelming proportion of the space found in this study. The Atlanta and Washington regions possess by far the largest urban envelopes, while five metropolitan areas lack a UE altogether. The Washington, DC urban envelope encompasses much of the District, and extends across the Potomac into Arlington and Alexandria, Virginia. Arlington’s Wilson Corridor—with nodes at Rosslyn, Clarendon, Court House, and Ballston—falls within this zone (see Fig. 4). Old Town Alexandria, which has seen a recent construction boom near the Eisenhower Avenue transit station on the Washington Metro, also forms part of the region’s urban envelope. Interestingly, all of the space in the DC urban envelope
740
LANG ET AL.
Fig. 4. Washington, DC metropolitan region.
lies within the original pre–Civil War borders of the District of Columbia, which formed a 10 mile × 10 mile diamond on land ceded by Maryland and Virginia. Edge Cities and Corridors Houston has the greatest proportion of office space in edge city densities and locations, with all of the office space located within Harris County (see Fig. 5). The other Texas metro area, Dallas–Ft. Worth, has nearly 22 percent of its office space clustered into corridors along the LBJ Freeway and to the north of downtown Dallas (see Fig. 6). Finally, Miami is the leader in edgeless city office space among the study regions (see Fig. 7). Whereas New York has the largest amount of edgeless city office space (over 222 million
BEYOND EDGE CITY
741
Fig. 5. Houston metropolitan region.
square feet), Miami’s dispersed office space represents more than 70 percent of its total due to the wide spread of its metropolitan region. Detroit and Philadelphia have the next largest proportions at 54 percent of their total office space. Edgeless Cities Although New York has the largest quantity of square footage in edgeless city space, the biggest regional share of edgeless city office space is found in Miami. Washington, DC has the smallest share, which indicates the clustered nature of office development in the region. The major differences among regions show that the spatial distribution of office development in the U.S. does not follow the simple model described by Garreau (1991). In fact, this complexity goes well beyond edge cities. As to why there is such variation, we
742
LANG ET AL.
Fig. 6. Dallas–Ft. Worth metropolitan region.
can only speculate. The reality is that U.S. metropolitan areas in general do not follow a typical pattern. As Galster et al. (2001) demonstrated by looking at six dimensions of urban sprawl in the same 13 regions as this study, there is no direct match between any two regions in terms of overall spatial form. Thus we should not expect that the distribution of office space would exhibit a different pattern. Rather, each region seems to fit office space into existing urban patterns, which our evidence clearly shows are idiosyncratic. POLICY AND MARKET IMPLICATIONS Reexamining the definition of edge cities and the amount of edgeless space in a region is helpful not only for understanding comparative urban form, but also for spotting potential trends and policy implications. In some regions, as edge cities age they may be molded into more traditional downtowns through the addition of more housing and public transportation. Edge cities may be losing out to edgeless cities and downtown, but the significant amount of infrastructure and investment that they represent will keep local governments interested in maintaining them. Edge cities may no longer be able to grow outward due to surrounding residential growth, but will instead have to grow vertically
BEYOND EDGE CITY
743
Fig. 7. Miami metropolitan region.
and prepare to become denser—especially if they wish to remain competitive with edgeless cities and downtowns. For example, Tyson’s Corner is planning to link up to the Metro rail system as part of a new line running out to Dulles International Airport. What about edgeless cities? What does the future hold for the spaces “in between”? The Lincoln Land Institute and the Urban Land Institute have written about ways to redesign edgeless cities. Unlike nodal edge cities, edgeless cities are diffuse, so how can they be “fixed”? Their density may increase over time and this may encourage greater mixing of uses and viable public transportation. Table 5 shows the amount of edgeless city office space that lies in neighborhoods with more than 8,000 residents per square mile. Edgeless city space at this density can be made accessible to rail-based transit due to the density of the surrounding built environment. The 8,000 people per square mile metric was developed by Pushkarev and Zupan (1977) in a study that looked at the urban densities needed to support public transportation in the
744
LANG ET AL.
TABLE 4. RANKING OF EDGELESS CITIES SPACE BY SQUARE FEET
Location
Square feet
Percent of regional office space
New York
222,052,423
32.3
Los Angeles
144,779,382
46.0
Chicago
107,776,326
39.4
Boston
101,051,915
50.1
Philadelphia
98,260,477
54.3
Washington, DC
86,113,714
24.0
Miami
77,293,650
72.1
San Francisco
68,790,444
39.4
Atlanta
64,775,200
38.4
Dallas–Ft. Worth
61,306,527
33.9
Detroit
58,351,987
54.1
Houston
56,679,390
34.7
Denver
49,326,137
50.8
Total
1,196,557,572
TABLE 5. TRANSIT POTENTIAL IN EDGELESS CITIESª Metropolitan area
Office space (s.f.)
Percent of edgeless
Atlanta
0
0.0
Boston
3,841,515
3.8
Chicago
1,991,294
1.8
Dallas–Ft. Worth
1,350,628
2.2
Denver
3,216,312
6.5
Detroit
3,959,908
6.8
Houston
3,669,733
6.5
Los Angeles
53,187,220
36.7
Miami
10,659,968
13.8
New York
40,231,441
18.1
Philadelphia
10,704,494
10.9
San Francisco
19,120,361
27.8
Washington, DC Total
8,117,282
9.4
160,050,156
13.4
ªThese are offices that lie in neighborhoods with more than 8,000 residents per square mile.
BEYOND EDGE CITY
745
New York region. They found that this threshold was the minimum requirement for transit. This figure represents a high intensity of residential development and requires a large share of multifamily dwellings. Neighborhoods at this density are often urban, but may include examples from inner suburbs, such as the Clarendon section of Arlington, Virginia. In total, more than 160 million square feet of edgeless city office space lies in “transitpossible” neighborhoods. This is an enormous quantity of space that exceeds Chicago’s downtown “Loop” in overall size. However, the distribution of this space varies tremendously by metropolitan area. Large markets, such as New York, Los Angeles, and San Francisco, account for a significant percentage of the transit-possible, edgeless city office space. In the case of Los Angeles, well over a third of its edgeless city office development fits into this category, followed by San Francisco at 28 percent and New York at 18 percent. By contrast, none of Atlanta’s nearly 65 million square feet of edgeless city office space lies in transit-possible neighborhoods. Likewise, only very small shares of such space exits in Boston, Chicago, and Dallas–Ft. Worth. Regions such as Miami and Philadelphia fall into the middle of this category. This is not to say that all places with the potential for rail-based transit will develop such infrastructure. Rather, the 8,000 people per square mile figure represents a potential for rail systems. Interestingly, a significant share of edgeless city space nationwide meets this criterion. Given the common assumption that scattered offices are mostly in “sprawling places,” the fairly common pairing of edgeless city space and dense neighborhoods was not an expected outcome. CONCLUSIONS The findings of this study present numerous opportunities and challenges to planners, realtors, and developers. A “new metropolis” has emerged with a tremendous diversity of commercial forms, and the market potential of most places outside downtowns is currently not well understood. Many larger edge cities are approaching buildout and have nowhere to go but up. Yet adding density to edge cities often produces congestion that discourages businesses from locating and investing in these places (Lang, 2003b). Edgeless cities play major roles in the new metropolis—both positive and negative. Some edgeless cities anchor moderately dense, mixed-use development that includes housing and retail. Others are scattered in the periphery of the region and are targets of sprawl critics. There is also a substantial dissimilarity in the location of office space by region. All metropolitan areas have some offices located in downtowns, edge cities, and edgeless cities, but distributions across these regions vary. Some contain just one edge city, while others have several. The downtowns of New York and Chicago still dominate their regions, whereas the CBDs in Detroit and Miami contain only a small share of their metropolitan areas’ office space. This article has provided a new perspective on the “new metropolis” beyond a simple urban and suburban dichotomy. Suburbs have become increasingly complex and yet remain understudied. The urban form categories (urban envelope, corridor, edgeless city) developed and refined in this study will hopefully enable a new way of looking at metropolitan regions. Our research here is primarily descriptive and focused on urban
746
LANG ET AL.
morphology. Future investigations need to examine more closely the activities that occur in the urban geographies we have identified. This should include the types of business and occupations within each and their spatial and economic relationships to one another. It will also be important to examine the real estate performance of these submarkets. For instance, how do rents and vacancy rate trends compare for edgeless cities, edge cities, and downtowns? There are also important lessons to be learned by reviewing the histories of these geographies in order to note the roles of land use and development policies (or lack thereof) in shaping these outcomes. With this type of information, comparative analyses can be pursued that may provide even more valuable insights into the growth and life cycles of metropolitan form and structure. REFERENCES Baerwald, T., 1982, Land use change in suburban clusters and corridors. Transportation Research Record, Vol. 861, 7–12. Baldassare, M., 1986, Trouble in Paradise: The Suburban Transformation of America. New York, NY: Columbia University Press. Barnett, J., 1992, Accidental cities: The deadly grip of outmoded zoning. Architectural Record, Vol. 180, 94–101. Berry, B. J. L., 1959, Ribbon developments in the urban business pattern. Annals of the Association of American Geographers, Vol. 49, 145–155. Berry, B. J. L. and Horton, F. E., 1970, Geographical Perspectives on Urban Systems. Englewood Cliffs, NJ: Prentice-Hall. Black’s Guide, 2005, Black’s Guide to Office Leasing: Office Space Market Series for Atlanta, Boston, Chicago, Connecticut/New York Suburbs, Dallas, Denver, Detroit, Houston, Los Angeles, New Jersey, Philadelphia, San Francisco, South Florida, Washington DC Metro Area. Gaithersburg, MD: Black’s Guide Publishing. Breckenfeld, G., 1972, “Downtown” has fled to the suburbs. Fortune, October, 80–87, 158, 162. Bruegmann, R. and Davis, T., 1992, New centers on the periphery. Center: A Journal for Architecture in America, Vol. 7, 25–43. Cervero, R., 1986, Suburban Gridlock. New Brunswick, NJ: Rutgers University, Center for Urban Policy Research. Cervero, R., 1989, America’s Suburban Centers: The Land Use-Transportation Link. Boston, MA: Unwin Hyman. Colliers International, 2005, Colliers’ February 2005 Estimates. Boston, MA: Colliers International. Daniels, P., 1974, New offices in the suburbs. In J. Johnson, editor, Suburban Growth: Geographical Processes at the Edge of the Western City. New York, NY: John Wiley and Sons, 177–200. Ding, C. and Bingham, R. D., 2000, Beyond edge cities: Job decentralization and urban sprawl. Urban Affairs Review, Vol. 35, 837–855. Duany, A., Plater-Zyberk, E. and Speck, J., 2000, Suburban Nation: The Rise of Sprawl and the Decline of the American Dream. New York, NY: North Point Press. Erickson, R. A., 1985, Multinucleation in metropolitan economies. Annals of the Association of American Geographers, Vol. 76, 331–346.
BEYOND EDGE CITY
747
Fishman, R., 1987, Bourgeois Utopias: The Rise and Fall of Suburbia. New York, NY: Basic Books. Fishman, R., 1990, America’s new city: Megalopolis unbound. Wilson Quarterly, Vol. 14, 24–45. Frey, W., 2003, Boomers and Seniors in the Suburbs: Aging Patterns in Census 2000. Washington, DC: The Brookings Institution, The Living Census Series. Frey, W. and Berube, A., 2003, City families and suburban singles: An emerging household story. In B. Katz and R. E. Lang, editors, Redefining Cities and Suburbs: Evidence from Census 2000. Volume I. Washington, DC: Brookings Institution Press, 257–288. Fulton, W. P., 1986, The once and future suburbs. Planning, Vol. 52, No. 7, 13–17. Galster, G., Hanson, R., Ratcliffe, M., Wolman, H., Coleman, S., and Frei Hage, J., 2001, Wrestling sprawl to the ground: Defining and measuring an elusive concept. Housing Policy Debate, Vol. 12, 681–717. Garreau, J., 1991, Edge City: Life on the New Frontier. New York, NY: Doubleday. Goldberger, P., 1987, When suburban sprawl meets upward mobility. New York Times, July 26, B30. Gordon, P. and Richardson, H. W., 1996, Beyond polycentricity: The dispersed metropolis, Los Angeles, 1970–1990. Journal of the American Planning Association, Vol. 62, 289–295. Greenhouse, L., 1971, The Outer City: Growth Turing Into a Menace. New York Times, June 3, A2, 28. Hartshorn, T. and Muller, P., 1986, Suburban Business Centers: Employment Implications. Washington, DC: U.S. Department of Commerce. Hartshorn, T. and Muller, P., 1989, Suburban downtowns and the transformation of metropolitan Atlanta’s business landscape. Urban Geography, Vol. 10, 375–379. Hayden, D., 2004, A Field Guide to Sprawl. New York, NY: WW Norton and Company. Herbers, J., 1971, The outer city: Uneasiness over the future. New York Times, June 2, A4, A20. Kain, J., 1992, The spatial mismatch hypothesis: three decades later. Housing Policy Debate, Vol. 3, 371–406. Katz, B. and Lang, R. E., 2003, Editors’ introduction. In B. Katz and R. E. Lang, editors, Redefining Cities and Suburbs: Evidence from Census 2000. Volume I. Washington, DC: Brookings Institution Press. Kneeland, D. E., 1971, The outer city: There is no firm stereotype. New York Times, May 31, A4. Lang, R. E., 2000, Office Sprawl: The Evolving Geography of Business. Washington, DC: Brookings Institution Center on Urban and Metropolitan Research, Brookings Survey Series. Lang, R. E., 2002, Open spaces, bounded places: Does the American West’s arid landscape yield dense metropolitan growth? Housing Policy Debate, Vol. 13, 775–778. Lang, R. E., 2003a, Beyond Edge Cities: Office Sprawl in South Florida. Washington, DC: Brookings Institution Center on Urban and Metropolitan Policy, Survey Series. Lang, R. E., 2003b, Edgeless Cities: Exploring the Elusive Metropolis. Washington, DC: Brookings Institution Press.
748
LANG ET AL.
Lang, R. E. and Blakely, E. J., 2006, In search of the “Real OC”: Exploring the state of the American suburbs. The Next American City, Vol. 10, 16–20. Lang, R. E., Blakely, E. J. and Gough, M., 2005, Keys to the new metropolis: America’s big, fast-growing suburban counties. Journal of the American Planning Association, Vol. 71, 381–391. Lang, R. E. and LeFurgy, J., 2007, Boomburbs: The Rise of America’s Accidental Cities. Washington, DC: Brookings Institution Press. Leinberger, C. B., 1984, New shape of cities will impact corporate locations. National Real Estate Investor, Vol. 26, No. 14, 40, 140. Leinberger, C. B., 1988, The six types of urban village cores. Urban Land, Vol. 47, No. 5, 24–27. Leinberger, C. B., 1996, Metropolitan development trends of the late 1990s: Social and environmental implications. In H. L. Diamond and P. F. Noonan, editors, Land Use in America. Cambridge MA: Lincoln Institute of Land Policy, 203–222. Leinberger, C. B., 1998, The market and metropolitanism. Brookings Review, Fall, 35– 37. Leinberger, C. B. and Lockwood, C., 1986, How business is reshaping America. The Atlantic, October, 43–52. Lessinger, J., 1962, The case for scatteration. Journal of the American Institute of Planners, Vol. 28, 59–169. Lewis, P. F., 1983, The Galactic Metropolis. In R. H. Pratt and G. Macinko, editors, Beyond the Urban Fringe. Minneapolis, MN: University of Minnesota Press, 94–111. Lewis, P. F., 1995, The urban invasion of rural America: The emergence of the galactic city. In E. N. Castle, editor, The Changing American Countryside: Rural People and Places. Lawrence, KS: University Press of Kansas, 39–62. Lynch, K. C., 1960, The Image of the City. Cambridge, MA: MIT Press. Muller, P. O., 1980, The Outer City: Geographical Consequences of the Urbanization of the Suburbs. Washington, DC: American Association of Geographers, Resource Paper No. 75-2. Muller, P. O., 1981, Contemporary Suburban America. Englewood Cliffs, NJ: Prentice Hall. Nelson, A. C. and Sanchez, T. W., 2005, The effectiveness of urban containment regimes in reducing exurban sprawl. NSL Network City and Landscape, Vol. 160, 42–47. Orfield, M., 2002, American Metropolitics: The New Suburban Reality. Washington, DC: Brookings Institution Press. Pierce, N., 1991, Edge city: The best America can do? Nation’s Cities Weekly, Vol. 14, No. 40, 5. Pivo, G., 1990, The net of mixed beads: Suburban office development in six regions. Journal of the American Planning Association, Vol. 56, 457–469. Pivo, G., 1993, A taxonomy of suburban office clusters : The case of Toronto. Urban Studies, Vol. 30, 31–49. Pressman, N., 1985, Forces for spatial change. In J. Brotchie, P. Newton, P. Hall, and P. Nijkamp, editors, The Future of Urban Form: The Impact of New Technology. London, UK: Croom Helm, 349–361. Pushkarev, B. and Zupan, J. M., 1977, Public Transportation and Land Use Policy. Bloomington, IN: Indiana University Press.
BEYOND EDGE CITY
749
Romanos, M. C., Chifos, C., and Fenner, T., 1989, Emergence of metrotowns in the American metropolitan scene: Definitition and evolution of the concept. In M. C. Romanos, editor, The City of the 21st Century. Cincinnati, OH: University of Cincinnati, 18–37. Sternlieb, G. and Hughes, J. W., 1988, The Suburban Growth Corridor. In G. Sternlieb and J. W. Hughes, editors, America’s New Market Geography. New Brunswick, NJ: Center for Urban Policy Research, 277–288. Swanstrom, T., Colleen, C., Flack, R. and Dreier, P. O., 2004, Pulling Apart: Economic Segregation among Suburbs and Central Cities in Major Metropolitan Areas. Washington DC: Brookings Institution Metropolitan Policy Program, Living Cities Census Series. Taylor, P. J. and Lang, R. E., 2005, U.S. Cities in the World City Network. Washington, DC: Brookings Institution Metropolitan Policy Program. U.S. Bureau of Labor Statistics, 2003, The Employment Situation: September 2003. Washington, DC: U.S. Department of Labor. Vance, J. E., 1977, This Scene of Man: Role and Structure of the City in the Geography of Western Civilization. New York, NY: Harper Press. Webber, M. M., 1964, The urban place and the non-place urban realm. In M. M. Webber, editor, Explorations into Urban Structure. Philadelphia, PA: University of Pennsylvania Press, 79–137.
750
LANG ET AL.
APPENDIX 1. INDIVIDUAL SUMMARY TABLES FOR OFFICE SPACE IN METROPOLITAN AREAS, 2005 Area, sq. mi.
Density, OS/sq. mi.
Residential density, per sq. mi.
Location
Square ft.
Percent
Downtowns
76,304,599
45.2
63.06
11,221,650
6.7
1.57
7,147,548
7,963
Urban envelope –Buckhead/Lenox
65,082,949
38.6
61.49
1,058,431
3,335
Edge cities
27,660,523
16.4
16.76
1,650,882
17,377,134
10.3
15.79
1,100,515
2,170
5,013,193
3.0
0.97
5,195,019
1,995 529
Atlanta Downtown
Cumberland-Galleria Duluth
1,210,032
North Fulton
5,270,196
3.1
1.88
2,803,296
Edgeless cities
64,775,200
38.4
8,399.02
7,712
Urban
21,021,253
32.5
3,000+
Suburban
37,629,923
58.1
1,000–3,000
6,124,024
9.5
Exurban Total
168,740,322
