2.1. Schools of Thought
According to Chicago school thinkers, the outcome of both economic and social processes is an observable and quantifiable spatial concentration of population groups within cities. Accordingly, land values and housing density, which help dictate who lives where, also exhibit a clear spatial pattern. Competition for land and the availability of new housing construction, coupled with transportation infrastructure (which helps determine both land prices and the propensity for land to be developed), give rise to a rather ordered spatial separation of social groups and subsequent urban form. Burgess’ [1
] famous concentric zone model identified four zones situated in rings around a city center core descending linearly in socioeconomic status toward a catchall ‘commuter zone’ on the outskirts of the city. Harris and Lewis [4
] argue that this model likely oversimplifies even the realities of early 20th-century Chicago, as the suburbs likely contained far more diversity than the commuter zone implied. Over time, the model arguably helped to exacerbate the perception of a stark city center /suburban divide whereas in reality, suburbs have long been economically, physically and socially diverse places [4
]. Nonetheless, supporters of Chicago school and research that follows in its tradition, defend its evidence-based approach of seeking to identify and explain regularities across urban areas [7
The Great Inversion hypothesis posited by Ehrenhalt [10
] retains the ordered spatial patterns described by the Chicago school, but reverses the sorting pattern to reflect more recent back-to-the-city movements and gentrification trends reshaping American cities. According to this view, the city center will be home to the more affluent social classes, while the disadvantaged will increasingly live at a distance from this core. Similar to this hypothesis is the idea of a New York school proposed by Halle [11
]. According to this view, cities conform to a pattern prototyped by New York City featuring a strong urban core, home to the middle class and rich as opposed to the working class and new immigrants described by original Chicago school thinkers.
On the far opposite end of the spectrum of these cities described by strong spatial patterns, the Los Angeles school, rooted in postmodernism, embraces the polycentric and periphery-driven landscape of Los Angles as emblematic of the future of US cities [6
]. According to the Los Angeles school, Chicago is the prototypical modernist city featuring a strong core which dictates an organized sorting of neighborhoods and social groupings away from this center. Los Angeles, by contrast, is polycentric in nature, with cores dispersed around the metropolitan area including in the periphery. As acknowledged by Beveridge [12
], though the notion of a Los Angeles school did not emerge until the 1980s, Fogelson [13
] described the evolution of Los Angeles from its agricultural beginnings until its ‘fragmented par excellence, the archetype for better or for worse of the contemporary American metropolis’ (p.2). Debates over whether Los Angeles is archetypical of newer American cities or exemplary in its decentralization and fragmentation have persisted.
Of these predominant schools of thought regarding the sociospatial ordering of US cities, Chicago is suggestive of clustered, spatially cohesive groupings of neighborhood types of varying geometric patterns. New York flips the ordering so that the city center features a concentration of advantage rather than disadvantage, while Los Angeles rejects a strong central core theory in favor of polycentricity and a high degree of fragmentation.
2.2. Empirical Evidence
A small but growing body of literature has sought to empirically test the relevance of these epistemologies in describing the current spatial structure of cities. For example, Hackworth [14
] evaluated landscape complexity in the 10 largest metro areas from 1970–2000 according to four variables: population density, rent, average home value, and per-capita income. He analyzed the distribution of these variables with respect to distance from their city centers, and found very regular patterns with respect to space. Ultimately, he concluded that the chaotic urban form proposed by the Los Angeles school did not describe the spatial structure of the ten largest US cities. While landscape complexity did increase marginally, overall, he argued that these cities witnessed inner-city revitalization, inner-ring suburban decline, and suburbanization in a largely similar fashion. Shearmur and Charron [15
] provided supporting evidence for the presence of regular patterns and processes with respect to the income distribution in Montreal, arguing in favor of Chicago school-inspired quantitative analyses that seek to uncover these regularities, and Meyer and Esposito [16
] found the income distribution in Los Angeles to be significantly related to distance from the central business district, even more so than for Chicago. This latter finding may capture some of the ‘Great Inversion’ patterns present in Chicago, as wealth returns to strengthen its urban core.
In a more recent analysis, Florida and Adler [17
] examined the 2012 spatial patterns of three classes of workers: advantaged, blue collar and service in 12 of the largest US metropolitan areas. They categorized these cities into three groups: core-oriented, where the advantaged class occupies a large share of the core of the city; class-bloc, where the three classes were largely split into large, separated spatial blocs; and fractal, where class segmentation was less organized. In contrast to Hackworth [14
], Florida and Adler [17
] argue that the spatial organization of social classes in modern cities has been reshaped into a ‘patchwork metropolis in which class divides cut across city and suburb alike’ (p. 14). The spatial patterns of neighborhood socioeconomic dynamics in Chicago and Los Angeles from 1970–2010 were mapped by Delmelle [18
], who noted that the resulting landscapes of Chicago largely reminisced the patterns foretold by early theorists with clear concentric zones and multiple nuclei. A spatially contiguous revival of downtown waterfront properties also aligned with the Great Inversion Hypothesis. On the other hand, the pattern of Los Angeles was much less ordered, particularly as it moved away from a largely impoverished core. Other descriptive analyses on the spatial patterns of various neighborhood types in various metropolitan areas have commented on the more structured spatial patterns of older cities compared to more dispersed spatial patterns in newer, more rapidly growing urban areas [19
] examined patterns of population growth in New York, Chicago and Los Angeles, and found that early growth followed a more structured pattern in Chicago and Los Angeles, but had no obvious pattern in New York. More recent growth has mimicked the unstructured picture of New York in all three cities. In all three cases, population growth could be found in both downtown and outlying areas, but there was no specific pattern to this growth. He argued that once an urban area has been settled initially, subsequent population growth is a function of other urban features such as transportation networks, land use types, and other factors that are not distributed in the same way across all urban areas. Florida and Adler [17
] reiterated this suggestion by proposing that the most advantaged class has located in the most geographically desirable locations, leaving the remaining spaces for those less well-off. These desirable locations may be in the center of cities, but they may also be in the suburbs, moving the importance away from a strictly spatial perspective and instead emphasizing the importance of urban and natural in determining where growth occurs and for whom. Foote and Walter [19
] noted the importance of highways and arterials in describing the spatial patterns of growth in rapidly growing cities, harking back to Hoyt’s sectorial model, and Meyer and Esposito [16
] acknowledged the role of natural amenities in Chicago models including distance from the city center, elevation and waterfront in determining the income distribution of cities. The evidence is therefore mixed on the status of traditional models in explaining current cities.
In this article, I offer a complementary perspective on the evolution of urban landscape complexity in 50 US metropolitan areas. I take neighborhoods classified into nine similar groupings according to their social, economic and housing characteristics, and analyze how the spatial configuration of these neighborhoods has changed through time (from 1990–2010). In the first part, I investigate the fragmentation of neighborhood types at the metropolitan scale to determine the extent to which the spatial arrangement of neighborhoods has become more intermixed and to compare this fragmentation among cities. This analysis addresses the quandary raised by Harris [21
] as to whether Los Angeles is exceptional or exemplary in depicting contemporary urban spatial structures. In the second part, I analyze the spatial clustering of each neighborhood type in each city. This follows the analysis of Delmelle [22
] who analyzed the spatial clustering of these neighborhood types for all 50 MSAs grouped together. In this analysis, I segment out these results by MSA to determine which social groups are most spatially compact and how these patterns compare across cities and through time. What is found is a trend toward spatial fragmentation across the majority of the 50 MSAs, supporting the idea that modern metropolitan development does not conform to the patterns predicted by the established, geometric traditional schools of thought. Rather, metropolitan sorting patterns and development occur in a manner that follows attributes dispersed throughout the landscape, both in the city center and peripheral space alike.