Measuring Sprawl

Abstract

Sprawl is a popular subject in the urban literature, yet conceptualization and measurement have proven elusive. Projects which focus either on empirical advances in the quantification of urban form or related phenomena like travel behavior are rarely conversant, leading to a fundamental disconnect between operationalizing the concept and modeling its effects. Here, I build on previous work in developing a new index of sprawl and examine changes in urban morphology at the metropolitan level in the United States from 2000 to 2010. I then illustrate face validity by outlining suggestive relationships between the index and associated environmental and housing outcomes, while comparing it with other commonly used measures. I find that sprawl continues into the twenty-first century, and that this proposed measure demonstrates initial face validity with respect to key environmental and housing outcomes. I conclude with a discussion of the results and suggestions for future research.

Keywords

sprawl environment polycentricity density housing

Sprawl has long been characterized as urban pathology, a signifier of unchecked development which consumes an excess of resources through land speculation and low-density dispersion. Research has convincingly linked it to such diverse phenomena as automobile-reliant travel behavior (Bento et al. 2005; Ewing and Cervero 2010), lower public transit ridership (Taylor et al. 2009), higher public infrastructure costs (Speir and Stephenson 2002), pronounced electrical energy use (Ewing and Rong 2008), increased greenhouse gas emissions (Clark 2013; Grazi, van den Bergh, and van Ommeren 2008; Hankey and Marshall 2010), localized particulate pollution and air quality (Stone 2008; Stone et al. 2007), elevated obesity rates (Zhao and Kaestner 2010), and the spatial mismatch between poor populations and employment opportunities (Covington 2009), among others (but see Brueckner 2001; Glaeser and Kahn 2004, for more positive assessments). Yet defining exactly what we should consider to be sprawl—often alternately described using the “urban” or “suburban” modifier—has proven to be more elusive than illustrating its general relationship (however measured) with these outcomes of interest.

Although there is a rich literature that on one hand connects sprawl to socially and environmentally significant effects, and on the other an effort to more accurately quantify it, these research agendas seldom overlap. That is, although econometrically complex work often illustrates links between urban form and outcomes like transportation behavior, it frequently employs rather simplistic indicators of the former—for example, the overall density of urbanized areas, or shares of population within zones defined by central business districts (CBDs) or seats of government (e.g., Bento et al. 2005; Su 2011; Taylor et al. 2009). Similarly, work that strives to better empirically capture sprawl often does so in isolation, largely—with rare exceptions (Ewing and Hamidi 2014)—leaving unexplored how well new measures relate to purported effects in establishing face validity.¹ Moreover, extant studies examining the causes and consequences of sprawl do not consider whether results are sensitive to different measures of the concept.²

I make a modest attempt at bridging these gaps by (1) proposing a new sprawl index, and using it to examine recent changes in U.S. Metropolitan Statistical Areas (MSAs); (2) comparing it with other indices that are widely cited and commonly used in research; and (3) modeling associated outcomes to demonstrate face validity and to illustrate differences between this approach and other alternatives. I focus on residential sprawl and develop a new index based on the work of Lopez and Hynes (2003) using data from the U.S. Census, a method that is parsimonious, practical, and easily replicable. First, I examine current measures of sprawl and highlight their strengths and weaknesses. I then explain the new index and use it to describe the morphology and recent growth trends of large metropolitan areas in the United States. I compare this measure with existing ones by examining their correlations with environmental and housing outcomes, and statistically model these relationships using both standard multivariate and first-difference regressions, offering provisional but suggestive evidence of the predictive power of the index outlined here. Finally, I conclude with a discussion of the findings and suggestions for future research.

Measuring a Nebulous Concept

While sprawl may be an intuitively simple concept in the abstract, quantification and empirical investigation are complicated by a lack of concrete parameters and the litany of different ways to gauge where, when, how, and to what relative extent it occurs. As the subject became popularized in academic discourse, researchers began conceptualizing and measuring sprawl using multidimensional frameworks which reflected the growing complexity of its theorization. Geographic Information Systems (GIS) eventually provided researchers the technological resources to incorporate spatial data into empirical measurements by offering mathematical renderings of urban morphology (e.g., “centeredness,” contiguity, etc.).

Galster et al. (2001) helped pioneer this multifaceted spatial approach, initially proposing six dimensions of sprawl: density (residential units per square mile of developable land), concentration (whether housing is distributed evenly over the urban area), clustering (the degree to which development is distributed evenly within subareas), centrality (how close development is in relation to the CBD), nuclearity (whether the urban area is mono- or polycentric—that is, the number of nodes constituting robust development), and proximity (the degree to which predominantly residential or nonresidential square-mile grids are geographically close to one another). Cutsinger et al. (2005) updated and expanded the analytical approach by adding a mixed-use development metric—related but distinct from proximity, as it is a measure of the jobs to housing ratio within subareas, as opposed to the distances between them—and examined the sprawl profiles for 50 metropolitan areas in the United States. The authors found that urban areas often ranked highly in some dimensions while scoring low on others, suggesting a complex portrait of metropolitan morphology and growth, and argued that whether an area can be described as sprawling depends on which factors are being considered (e.g., overall density, CBD-proximate development, etc.).

Ewing, Pendall, and Chen (2003) took a similar but more streamlined approach, calculating a composite measure which included subindices representing density, centrality, mixed land use, and street accessibility (i.e., shorter or larger city blocks). The most recent incarnation (Ewing and Hamidi 2014) adds employment and walkability data in constructing the subindices. Similar to Cutsinger et al. (2005), Ewing and Hamidi incorporate numerous variables related to spatial morphology into their measure, among them centralized development (a measure of compact monocentric growth), density gradients (how fast density declines with distance from the CBD), street accessibility (average city block size), and “centering” measures (the proportion of population and employment within CBDs and subcenters). Using standardized scores to neutralize the influence of overall population size, the authors calculate sprawl indices for counties, metropolitan areas, and urbanized areas in the United States. They also demonstrate the face validity of their measure by regressing a number of outcome variables (e.g., housing affordability, obesity rates, etc.) on the composite measure and its subindices, a rarity in the literature.

These approaches have many strengths, first and foremost the complex way they statistically render many distinct aspects of urban form. Researchers who grapple with specific empirical questions can use these indices to parse out which morphological factors most precisely associate with their outcomes of interest both theoretically and empirically, rather than necessarily having to use composite or reductive measures. The index offered by Ewing and colleagues in particular has been brought to bear on numerous research projects exploring public health and energy use outcomes, establishing a track record in the literature (e.g., Ewing and Hamidi 2014; Ewing, Pendall, and Chen 2003; Ewing and Rong 2008). Yet multidimensional spatial approaches also have their weaknesses, ranging from the relatively minor (e.g., incorporating proprietary data, like walking scores in Ewing and Hamidi’s metro land-use mix subindex) to the more significant (relying on numerous measures which quantify urban centrality or concentration, which can provide misleading results).

Incorporating measures of centrality in particular into these composite measures tacitly positions monocentric morphology (i.e., urban areas that radiate outward from a dominant CBD) as less sprawling than polycentric forms (i.e., employment and development that are multinodal rather than concentrated at one or few points within a given region). Yet there is a paucity of empirical evidence that polycentric development is, in practice, less desirable in terms of ecological, economic, or general welfare outcomes. As researchers have pointed out with respect to commuting specifically, the general balance between housing and jobs in a given subarea is a more important influence on transportation behavior than proximity to a given CBD (Buliung and Kanaroglou 2006; Modarres 2011). As Gaigné, Riou, and Thisse (2012) illustrated, there are scenarios where compact monocentric growth patterns can lead to higher total travel-related emissions than polycentric ones because of the relocation of firms and residents between and within cities, which can lead to longer and more energy-intensive trips to a single CBD rather than localized commutes to a given subcenter. These multidimensional approaches largely avoid accounting for these possibilities.

For instance, Cutsinger et al. (2005) build housing and job centrality and monocentricity into their index rather straightforwardly, all of which are conditioned on proximity to a central city hall location. Ewing and Hamidi (2014) improved upon their previous measure of centrality (see Ewing, Pendall, and Chen 2003)—a simple measure of employment shares within concentric rings around a given CBD—by accounting for employment subcenters using a nonparametric procedure developed by McMillen (2001). This ostensibly accounts for urban areas which are polycentric by identifying development nodes other than the CBD. Specifically, they measure population and employment within subareas that meet the criteria (census block groups with significant positive residuals estimated from an exponential employment density function conditioned on distance from a given CBD).

Yet this only accounts for population and employment proportions within these rather geographically limited subareas. This could in practice position urban areas with concentrated nodes of population and employment surrounded by, say, detached single-family housing, as less sprawling than regions with a greater mix of employment and residential construction and overall higher-density development. Ewing and Hamidi (2014) also included the coefficient of variation of population and employment density as part of their centering subindex, arguing that greater variation among subareas around the mean within a given metro is indicative of less sprawled development. The weakness of this measure can be best illustrated by the example of Atlanta and Los Angeles—the former at or near the bottom of every sprawl ranking, including Ewing and Hamidi’s, and the latter the general opposite. Using the mean and standard deviation of population density among developed census blocks in 2010 to calculate coefficients of variation, Atlanta has a significantly higher value than Los Angeles (1.149 and 0.831, respectively). This is not because Atlanta is less sprawled but because Los Angeles is more uniformly dense and, thus, has a smaller standard deviation relative to its (much higher) mean.

In a unidimensional spatial approach, Lee et al. (forthcoming) incorporated mass transit access into their Compact City Index (CCI) and used it to examine midsize Japanese cities. The CCI is a function of the population and amenity densities surrounding mass transit stops, along with the proximity of those localized zones to the CBD (as determined by land values). Although this can be a valuable tool for assessing transit access and pedestrian-oriented urban growth, it relies on a single CBD for deriving its calculations and, thus, does account for polycentric form. Moreover, because transportation access is not deterministically driven by densities—even if it is made feasible by them—but complicated by local politics, revenues, and a number of other factors, it is less a true sprawl index than a metric of robust public transit and the composition of surrounding areas. Finally, though the CCI was designed to be more or less universally applicable, places with no centralized accounting of municipal public transportation—like the United States—can be problematic with respect to measurement.

Other more purely spatial methods for measuring sprawl rely on satellite imagery of land use and do not incorporate demography. Burchfield et al. (2006) used National Land Cover Database (NLCD) data—which assign land-use categories to 30m × 30 meter grid cells that cover the United States—to construct a sprawl index based on the percentage of undeveloped land in the square kilometer surrounding a given cell. While these data can be very useful for some applications, such as describing the total area of urban development over a bounded geography, calculating sprawl with this methodology cannot precisely account for what kinds of construction characterize a cell.³ Are developed grid cells that are geographically distant from a central city comprised of low-density single-family housing that leapfrog along a transportation corridor, or are they characterized by multifamily construction that are separated by land which is difficult, costly, or even impossible to build on? The findings cast further doubt on the methodology, which rank Memphis to be the second least sprawling metropolitan area with a population over one million in the second stage measurement (land use in 1992), with Dallas close behind. While counterintuitive results should not be dismissed out of hand, these results are likely an artifact of the research design (see Irwin and Bockstael 2007 for a more thorough critique).

Bereitschaft and Debbage (2014) likewise used NLCD data to quantify fragmented development in large urban areas in the United States. While their approach is more sophisticated, using nine different indicators of continuity and shape complexity (e.g., contiguity, landscape shape, etc.), the same weaknesses apply because the researchers do not account for the underlying composition of the urbanized patches of land. Using this measure, New Orleans and Buffalo are less sprawled than Boston or New York merely because they are more uniform in shape. Jaeger and Schwick (2014) offered a similarly sophisticated rendering of urban form and used it to construct a sprawl measure they termed Weighted Urban Proliferation (WUP), using it to explore sprawl in Switzerland. Unlike other approaches that rely on satellite data, WUP does account for total population, which offers a clearer and more comprehensive illustration of development than spatial morphology of urbanized patches of land alone. Still, the index relies on a specification of a bounded area around the land patches which the authors term the “horizon of perception,” and it is unclear how such aerial units would be determined for a larger-scale intermetro comparison, which includes places of vastly different size and underlying topography. Moreover, the incorporation of total regional population into the calculation does not address the granular composition of subareas which constitute a given urban area.

Recent work by Paulsen (2014) and Tsai (forthcoming) focused on changes in regional sprawl patterns over time. The former describes changes in housing density using four variables: Overall change in housing unit density, marginal land consumption of each new housing unit, the density of housing in newly urbanized areas, and the percentage of net new housing construction in places already urbanized. Tsai develops a sprawl index which expresses the proportion of metro population in low- and high-density subareas (i.e., the percentage of population in the top and bottom quintiles, based on subarea density distributions computed for each metro). Tsai’s measure must not necessarily be expressed dynamically, as unlike Paulsen’s land consumption approach, it is based on discrete sprawl scores calculated at different time points. Nevertheless, it pegs thresholds to regional percentile scores rather than establishing universal cut points, making it more suitable for examining changes over time within individual urban areas as opposed to illustrating the differences between them.⁴ Although both methods offer valuable tools for analyzing the changing nature of sprawl and urban development, they are less useful for deciphering these cross-sectional interurban differences.

Lopez and Hynes (2003) offered a simple density-based approach to measuring sprawl. The authors calculate the percentage of a given metropolitan area which is sorted among low-density census tracts above a rural threshold (over 200 but under 3,500 persons per square mile). This method has many advantages, from its ease of calculation and straightforward data requirements to its broad applicability and replicability. Although it leaves out other theorized dimensions of sprawl, it consequently avoids the weaknesses of spatial approaches, successfully zeroes-out rural or undeveloped land through a baseline cut point, and operates independently of regional size, municipal boundaries, and physical geography. Yet for all its strengths, it relies on one relatively modest threshold for high-density development (areas characterized by detached single-family housing can meet the 3,500 persons/m² standard with relative ease) and uses census tracts as its subarea unit of analysis, which can be large enough to skew density estimates. I build on Lopez and Hynes’ measure in attempting to retain the strengths of the approach while minimizing these weaknesses.

A New Sprawl Index

One key issue with any sprawl measure that relies on calculations involving subareas (i.e., the percentage of population in high- or low-density tracts, block groups, etc.) is what unit of measurement to use. Because fine-grained demographic data are only practically available from the U.S. Census, the choices are confined to blocks, block groups, and tracts, in ascending order of size. The modifiable aerial unit problem becomes an important consideration, as results can be highly sensitive to which subunits are used in calculations. For instance, census tracts can often include substantial amounts of land devoted to infrastructure or that are otherwise undevelopable, which can easily skew density calculations downward (see Figure 1 for an example).

Figure 1.

Census tract geography.

Using larger aerial units becomes particularly problematic when higher cut points for density are used. For instance, among the top 150 MSAs in the United States, more than two-thirds have no census tracts with a population density over 20,000 persons/m². Using blocks, the smallest unit level available using census data, every one of these MSAs has at least some subareas that exceed the same threshold. Using tracts can demonstrably bias estimates downward because for many urban areas they simply fail to capture fine-grained pockets of residential density, however rare they may be (see Figure 2 for a graphical illustration using the example of Dallas county of how using various aerial units can render significantly different density profiles). To minimize the influence of biasing scale effects, I calculate sprawl values using blocks, the smallest aerial unit available, while also offering tract-based scores to illustrate their differences in modeling the outcomes explored later in this article.⁵

Figure 2.

Block, block group, and tract densities.

Another important consideration is which cut points should be used; what density thresholds correspond with robust multiunit housing as opposed to detached single-family construction? Because these standards will often vary according to regional context and the unit levels at which density is calculated, researchers have had to rely on somewhat arbitrary baselines for what constitutes urban (or suburban) development. Lopez and Hynes zero-out tracts with less than 200 persons/m² as rural and designate those with over 3,500 persons/m² as dense in their calculation. In their density subindex, Ewing and Hamidi zero-out tracts with less than 100 persons/m² and designate those with less than 1,500 and over 12,500 persons/m² as low- and high-density, respectively, for their calculations. Paulsen (2014) used 500 persons/m² as the threshold and drop areas from the housing growth measures which both fail to meet the standard and are noncontiguous past a defined aerial scope. Here, I use multiple thresholds for calculating a new sprawl index. I retain the 200 persons/m² baseline because it roughly corresponds to the kind of growth fostered by large-lot zoning rules found in heavily regulated suburban municipalities and is generally consistent with established measures.⁶ I also retain the 3,500 cut point, which roughly signifies the transition from low- to moderate-density single-family housing development (Pendall, Puentes, and Martin 2006).

Because areas which only modestly exceed this threshold can still be characterized by relatively low-density development, however, I add cut points of 8,500 and 20,000 persons/m². I base these partly on research by Guerra and Cervero (2011), who analyzed what density levels were systematically conducive to cost-effective mass transit rail operation. The 8,500 and 20,000 cut points approximate the densities required for economically viable heavy rail construction with capital costs of $100 to $200 million per square mile in large cities and light rail with costs of $25 to $50 million (U.S. Dollars) per mile in medium-sized cities, respectively (Guerra and Cervero 2011). They also roughly correspond to the transition away from detached single-family to multiunit apartment building construction (Pendall, Puentes, and Martin 2006). While any threshold is arbitrary to some extent, I offer real-world examples of blocks that slightly exceed the cut points used here in Figure 3, which illustrate that they meaningfully distinguish between different forms of housing construction—ranging from large-lot exurban single-family homes to midrise apartment buildings with common courtyards.

Figure 3.

Example of census blocks at density cut points.

I calculate the total percentages of MSA population residing in census blocks below each of these cut points (3,500, 8,500, and 20,000 persons/m², while zeroing-out blocks under the 200 baseline; see Figure 4 for a graphic illustration of the extent of metropolitan development above 200 persons/m² in the study MSAs within the contiguous United States). I then average these three values, such that higher overall index scores indicate more population below each of these cut points and consequently more sprawl. While this equally weighs each threshold, I can provide no theoretical or empirical basis for doing otherwise. The result is an index which ranges from 0 to 100. First, I calculate the sprawl index for the largest 150 MSAs in the United States and show changes from 2000 to 2010. Although the extent of urbanized areas and what kind of development constitutes them is the subject of some debate, I use MSA boundaries in this analysis. While some MSAs include counties which lie far beyond the dense central cities that anchor them, the U.S. Census has developed objective criteria for its delineations.

Figure 4.

Extent of MSAs under study: Blocks with population density >200 persons/m².

For counties to be included in a given MSA, they must have substantial economic ties to the surrounding environment—that is, at least 25% commuting interchange between outlying counties and central urban areas. Of more practical importance, because MSAs are composed of counties, which have very stable historical boundaries, time trend calculations are straightforward, and indices can be easily coupled with a diverse array of outcomes of interest to researchers in a simpler way than they can be with, say, census-defined urbanized areas. I begin with the latest MSA delineations in tracking changes over time (Office of Management and Budget [OMB] 2013). While counties are added to and drop from MSA delineations because of changes in commuting patterns, I use the most recent boundaries to gauge how density levels have changed in places which eventually became coupled or decoupled with metropolitan regions as they are currently defined. Zeroing-out blocks with less than 200 persons/m² also addresses the “underbounding” problem and excludes outlying rural development that is qualitatively and conceptually distinct from urban sprawl. Of course, because this method does not require that blocks be contiguous, more distant nodes (as long as they meet the baseline threshold) would be counted as part of the index. Because as part of the MSA delineation they are nevertheless economically tied to the larger region, and because of the practical import of eventually modeling effects which include aggregate figures based on these populations, I deliberately leave them in the analysis without imposing a contiguity requirement on their inclusion.

Sprawl in the United States: 2000–2010

The total population living in blocks with greater than 200 persons/m² within the top 150 MSAs was 212.4m in 2010—about 68.8% of the national population and an 11.6% increase from 2000.⁷ This constituted the lion’s share of total growth in the United States during the decade; remaining areas (smaller MSAs, micropolitan statistical areas, rural locales, and blocks within larger MSAs with less than 200 persons/m²) grew 5.7% over the same time frame. While population change was positive at every threshold, the lowest-density areas grew the most and the highest-density areas least: The national index increased from 57.9 in 2000 to 59.4 in 2010 (see Table 1). The densest region in the United States in 2010 was the West, followed by the Northeast, Midwest, and the South; this order was unchanged from 2000 to 2010 (see Table 2). With respect to thresholds, however, the South’s densest blocks grew the fastest of any region, signaling substantial multifamily residential development, though not enough to offset the growth of detached single-family housing. The only region that experienced a decline in population among its densest areas was the Midwest, which lost 8.7% of its population among blocks with 20,000 persons per square mile or more. Across the United States, every region grew more sprawling over the decade.

Table 1.

National Changes in Population Among Blocks Above Population Density Thresholds.

	200–3,500 Persons/m²	3,500+	8,500+	20,000+	Sprawl Index
Total population 2000	56.56m	133.77m	75.30m	31.25m	57.91
Total population 2010	64.50m	147.83m	78.90m	32.23m	59.37
% change	14.22%	10.51%	4.79%	3.13%	2.52%

Table 2.

Population Above Density Thresholds in 2010 and Changes 2000–2010 by Region.

Total Population in Blocks Above Threshold	West	Northeast	Midwest	South
200+/total population (Δ2000–2010)	55.25m (14.17%)	44.96m (3.77%)	39.59m (5.38%)	72.63m (18.99%)
200–3,499 persons/m² (Δ2000–2010)	9.09m (9.10%)	13.31m (7.63%)	12.95m (12.69%)	29.25m (20.05%)
3,500 + (Δ2000–2010)	46.17m (15.22%)	31.65m (2.22%)	26.63m (2.16%)	43.38m (18.28%)
8,500 + (Δ2000–2010)	28.33m (12.34%)	21.75m (0.98%)	10.92m (−11.88%)	17.91m (10.85%)
20,000 + (Δ2000–2010)	9.53m (5.24%)	13.47m (2.19%)	3.57m (−8.70%)	5.66m (10.87%)

Note. Regional definitions are consistent with census delineations.

Although Los Angeles is often popularly associated with sprawl because of its pollution and traffic, its sheer lack of very low-density development places it atop all U.S. metro areas, with New York and San Francisco close behind (see Table 3 for top- and bottom-ranked MSAs and Figure 5 for a graphic illustration of sprawl in 2010 in the contiguous United States). While there is a moderately strong correlation between the total counted population (i.e., the sum of blocks which have population densities higher than 200 persons/m²) and the index in 2010 (r = −.515), there are notable defiers. For example, Santa Barbara ranks as the 8th least sprawled metropolitan area in the United States, yet is outside the top 100 in total population. Although the geography of coastal California is unforgiving, Santa Barbara has a long and unique history of land-use controls (Warner and Molotch 1995), offering another possible reason for its ability to channel growth into already densely built areas. Similarly, large metros do not necessarily achieve the kinds of dense development which often characterizes these regions. Atlanta is the 9th largest MSA in the United States, yet ranks 122nd on the sprawl index (81.14), similar to Birmingham and Mobile, Alabama. Consistent with general regional trends, the bottom-ranked metros are exclusively in the South. See Table 4 for descriptive statistics for 2000 and 2010 index values and 2000–2010 percentage changes.

Table 3.

2010 Sprawl Index: Top- and Bottom-Ranked MSAs.

Rank	MSA	2010 Index (Δ% 2000–2010)	Rank	MSA	2010 Index (Δ% 2000–2010)
1	Los Angeles, CA	32.59 (<−0.01)	141	Huntsville, AL	85.49 (−0.46)
2	New York, NY	33.20 (1.19)	142	Chattanooga, TN	87.77 (1.09)
3	San Francisco, CA	37.61 (−1.09)	143	Greenville, SC	87.79 (1.68)
4	Honolulu, HI	38.00 (−7.01)	144	Knoxville, TN	88.14 (0.36)
5	San Jose, CA	40.20 (−0.94)	145	Winston–Salem, NC	88.28 (0.68)
6	Salinas, CA	47.17 (5.22)	146	Ocala, FL	90.72 (2.05)
7	Chicago, IL	47.18 (6.85)	147	Asheville, NC	91.27 (−0.66)
8	Santa Barbara, CA	47.81 (−5.26)	148	Kingsport, TN	92.01 (0.19)
9	San Diego, CA	48.02 (−2.52)	149	Hickory, NC	95.22 (0.77)
10	Miami, FL	48.04 (0.16)	150	Columbia, SC	96.18 (0.76)

Figure 5.

2010 sprawl index among the top 150 MSAs.

Table 4.

Descriptive Statistics for 2000/2010 Sprawl Indices.

	2000 Index	2010 Index	Δ% 2000–2010
M	68.518	69.502	1.422
SD	12.134	12.404	2.874
Median	69.710	70.993	1.359

In terms of 2000–2010 trends, only 45 of the 150 MSAs surveyed here grew more dense (see Figure 6). The leaders were mostly in the West, with the notable exceptions of McAllen, Texas, Tallahassee, Florida, and Raleigh, North Carolina (see Table 5)—the latter two which include prominent universities and the former a center of border trade and industrial production between the United States and Mexico. The metropolitan areas which were most sprawling over the same period were mostly postindustrial rustbelt regions which experienced losses in employment and population. Of the MSAs surveyed, 10 declined in population from 2000 to 2010 and 6 of those rank as most sprawling over the same time frame. The overall correlation between MSA population growth among blocks with density greater than 200 persons/m² and changes in the sprawl index from 2000 to 2010 is significant, negative, and moderately strong, illustrating that population decreases correlate with more sprawl over time (r = −.379).

Figure 6.

Change in the sprawl index, 2000–2010.

Table 5.

Percentage Change in the Sprawl Index 2000–2010: Top- and Bottom-Ranked MSAs.

Rank	MSA	Δ% 2000–2010 (2010 Index)	Rank	MSA	Δ% 2000–2010 (2010 Index)
1	Honolulu, HI	−7.01 (38.00)	141	Ann Arbor, MI	5.36 (67.04)
2	Salem, OR	−5.49 (61.90)	142	Reno, NV	5.50 (61.52)
3	Santa Barbara, CA	−5.26 (47.81)	143	Buffalo, NY	5.73 (65.14)
4	Portland, OR	−4.37 (59.49)	144	Savannah, GA	5.79 (76.75)
5	Seattle, WA	−3.89 (62.27)	145	Toledo, OH	6.69 (69.06)
6	Tallahassee, FL	−3.84 (77.15)	146	Chicago, IL	6.85 (47.18)
7	Colorado Springs, CO	−2.98 (62.81)	147	Cleveland, OH	7.43 (68.47)
8	Riverside, CA	−2.83 (58.82)	148	Flint, MI	7.92 (82.08)
9	McAllen, TX	−2.69 (70.12)	149	Detroit, MI	8.43 (67.78)
10	Raleigh-Cary, NC	−2.53 (78.04)	150	New Orleans, LA	14.80 (59.32)

Population losses do not necessarily mean places must become less dense as a matter of course (i.e., if their lowest-density regions were the ones experiencing depopulation), but in the absence of carefully targeted decline, it is obviously difficult for contracting regions to preserve their overall morphological profile in practice. Some MSAs like Chicago were marked by increases in sprawl despite gaining population, likely driven by losses in the central city proper and surrounding environs. Yet the correlation between percentage change in the sprawl index and corresponding population change in the largest central cities anchoring them was virtually zero (r = −.005), consistent with Downs (1999), who also found no significant general relationship between core decline and regional sprawl.

Among the MSAs growing densest from 2000 to 2010, the presence of robust growth controls is conspicuous. As the first state to introduce metropolitan growth boundaries in the 1970s, Oregon pioneered comprehensive, regional land-use planning. Its MSAs have grown similarly to other areas which also have robust controls—for example, Honolulu and Santa Barbara—but with comparatively negligible physical constraint. In Washington state, as Anthony (2004) described, state policy targets restrictions to the fastest-growing regions and couples them with incentives for compact development, which may similarly contribute to the marked improvement of their MSAs. Other large MSAs which grew notably denser were San Diego (2.52% reduction in the index), El Paso (2.32%), Albuquerque (2.05%), Charlotte (1.94%), and Houston (1.42%), the latter two defying regional trends.

Extant scholarship on growth control regimes paints a rather inconsistent picture of their effect on compact development. Yet the most recent research on state and regional growth controls, which attempts to rectify the weaknesses of earlier studies, seems to generally support the trends illustrated from 2000 to 2010. Specifically, researchers have used panel data to more convincingly demonstrate causal effects and have gauged the relative strength of such growth controls rather than using simple dummy indicators for the presence or absence of statewide or regional policy. Using this approach, Howell-Moroney (2007) found that only strong growth management programs (like those found in Oregon) were able to significantly reduce urban land conversion and increase density levels. Similarly, Paulsen (2013) found that strong controls reduced marginal land consumption in urban areas, and that moderate programs counterintuitively produced even more peripheral land conversion than unregulated places. This points to the weaknesses of growth policies which lack robust regional coordination and implementation, and would help explain why some places which have state-mandated controls still contain MSAs which continue sprawling despite them (e.g., Tennessee and Arizona).

The plight of declining metropolitan regions—which sprawled the most from 2000 to 2010—highlights the difficulty in preserving compact communities in places suffering from significant losses in population and employment. As controversial as the imposition of growth controls has been, targeted decline raises even more vexing questions as to how to preserve relatively healthy areas amid widespread deprivation. Recently Detroit has taken the approach of triage, as outlined by the land-use policies in the “Detroit Future City” plan, calling for targeted investment in low-vacancy areas while gradually siphoning funds away from neighborhoods with the most severe problems (Detroit Future City 2014). While targeting resources to healthy areas may prove effective in producing more efficient service provision and public infrastructure expenditure, along with densifying places which remain relatively vibrant, there are obvious issues of logistics and equity concerns with respect to what happens to those who stay behind. Moreover, evidence on the imposition of growth controls and the lack of association between central city decline and metropolitan form suggest that municipal-level solutions may be found wanting in the face of an absence of regional frameworks for dealing with such issues.

Although low-density development continued from 2000 to 2010 in the United States, with single-family detached housing outpacing the growth of multiunit construction, there are recent signs that this may be changing. While this analysis includes the start of the global recession and the foreclosure crisis—which effectively halted housing construction and increased rentership, both of which would likely slow sprawl patterns—population growth in large central cities only began outpacing that of their suburbs in 2010. This trend has continued up to the present (Frey 2014), suggesting that updated figures may reflect a slowdown in sprawl, if not a total inversion of the dynamics which have characterized decades of urban growth in the United States. Whether this trend is a reflection of a “new normal” whereby the cultural predilections for suburban living are waning or a function of economic malaise (or a combination of these and other factors) remains to be seen, as does its durability, its relationship with sprawl patterns in the new decade, and whether the negative effects of sprawl may, thus, be attenuating as time passes.

Sprawl, Housing, and the Environment

Perhaps the most compelling examples of the negative effects of sprawl involve environmental outcomes and energy use. For instance, Clark (2013) found significant associations between vehicle miles traveled (VMT), carbon emissions, and core densities in 57 urbanized areas in the United States, yet argued that efficiency gains may be more effective policy-wise than imposing wholesale changes in urban morphology. Hankey and Marshall (2010) took the unique approach of modeling future urban growth scenarios, finding that continuing historical urban growth trends could wash out future advances in efficiency and technology, and argued that addressing the form of physical development and fleet emissions and providing new energy sources in tandem have the potential for profound reductions in carbon emissions that would be meaningful on a global scale.

Even in the face of technological advances which have dramatically reduced particulate emissions—particularly those that contribute to smog formation—research has also connected sprawl to these more conventional sources of pollution. Stone et al. (2007) build a model forecasting future pollution rates under various growth scenarios, finding meaningful elasticities between changes in the built environment and vehicular emissions. In a similar study, Stone (2008) examined data on annual ozone exceedance in 45 large metropolitan areas in the United States, finding sprawl to have significant effects on ambient air quality net other factors, which suggests that spatial form further influences pollution indirectly through urban heat island phenomena. Indeed, in subsequent research, Stone, Hess, and Frumkin (2010) found that the rates of increase in extreme heat events (EHEs) in the most sprawling metropolitan regions in their sample were double that of compact ones. This suggests that in addition to worsening climate change through vehicular emissions, sprawled regions may also be significantly more vulnerable to its localized effects.

Although much of the research on sprawl focuses on its negative implications, some researchers point to the potential positive trade-off of greater housing affordability and homeownership. In a simple rendering of the Alonso–Mills–Muth theory (see Alonso 1964; Muth 1968), a stylized monocentric model of urban growth would hold housing price at a given periphery as a function of commuting cost and distance to employment and would fall as these distances increase (see also Brueckner 2001, who argued for the model’s validity to polycentric urban forms). Consumers may willingly incur the expenses of commuting because land rent is sufficiently low for them to make this trade-off. In practice, cheap energy and ubiquitous automobility often make this desirable. While market interventions—chiefly in the form of land-use governance as it exists in the United States—surely exist, ceteris paribus, more peripheral development that is often low density as a matter of course would result in greater regional housing affordability under this theory. As Glaeser and Kahn (2004) illustrated, suburban homes are often cheaper on a price per square foot basis than those in corresponding central cities even if absolute costs are generally higher. Among individual metropolitan areas, Clark (2013) found both renter and owner affordability to be negatively related to the core densities of the larger urbanized areas. Focusing specifically on infill development in central cities, Steinacker (2003) found it to be correlated with lower general housing affordability, also broadly consistent with the theory.

To demonstrate the face validity of the sprawl measure devised here, I focus on how it relates to these selected environmental and housing outcomes. I use 2011 data from the National Emissions Inventory in exploring links between the sprawl index and conventional pollution, and 2002 data from the Vulcan Project (Gurney et al. 2009) to model its association with carbon dioxide emissions. Both of these are per capita figures and are derived from “onroad” sources (i.e., vehicular emissions), which helps isolate the influence of urban form and avoids confounding variation in energy generation or industrial capacity with systematically different individual-level behavior. I also use a new measure of housing affordability developed by the Department of Housing and Urban Development (HUD)—the Location Affordability Index—which is a calculation of housing cost as a percentage of income for both median renter and owner households, signifying the average housing cost as a percentage of income for households at the median income level and the average household size for a given region.

First, I present a simple correlation matrix to illustrate bivariate associations between these outcomes and a range of commonly available sprawl measures, including the one devised here. I then use standard multivariate regression to model the associations between the sprawl index and carbon emissions, hazardous pollution, and housing affordability. Finally, because these variables are only available cross-sectionally, I include a fixed-effects (first-difference) model which explores the relationship between the changes in owner-occupied home values and the change in the sprawl index from 2000 to 2010 (see Table 6 for a detailed explanation of dependent variables and sprawl measures). For the bivariate correlation matrix and the hazardous pollution/affordability ordinary least squares (OLS) models, I recalculate the index to conform to 2010 MSA delineations so it corresponds with other sprawl measures and the outcomes of interest. For the CO₂ model, because the outcome is measured at 2002, all the independent variables are from 2000 rather than 2010. Because the index is significantly correlated with total population, I also took the approach of Ewing and Hamidi (2014) and developed a normalized measure derived from regressing the sprawl scores on the natural logarithm of MSA population. Using the standardized residuals rather than the raw index results in no substantive differences with respect to any of the results.

Table 6.

Main Measures, Sources, and Details.

Sprawl Measures (Abbreviation)	Source; Unit Level; Year	Details
New Sprawl Index (IN10)	Census Summary File 1 (SF1); Blocks; 2010	Shares of MSA population under 3,500, 8,500, and 20,000 persons per square mile, averaged. Calculated at the block level
New Sprawl Index (INCT)	Census SF1; Tracts; 2010	Same as above at census tract level
Population-Weighted Density (CPB)	Census SF1; Blocks; 2010	D = Σ(Pi_di) / ΣPi, where D is the population-weighted density of a metro or micro area, and Pi and di are the population and density of the ith census tract (or block), respectively
Population-Weighted Density (CPT)	Census SF1; Tracts; 2010	Same as above at the tract level
Lopez and Hynes’ Sprawl Index (LH)	Lopez and Hynes (2003)—Data publicly available (Lopez 2014); Tracts; 2010	SI_i = (((S%_i − D%_i) / 100) + 1) × 50, where SI_i is the sprawl index, D%_i is the percentage of total population in high-density tracts, and S%_i is the percentage in low-density tracts
Lopez and Hynes’ Sprawl Index (LHB)	Census SF1; Blocks; 2010 (recalculated from above)	Same as above at the census block level
Ewing and Hamidi’s Composite Sprawl Index, and Density, Mixed-Use, Centrality, and Street Network Density Subindices (ECS, ED, EM, ECN, and EST)	Smart Growth America: Measuring Sprawl (see Ewing and Hamidi 2014)—Data publicly available; Various; 2010	Composite index of sprawl; Constitutes four dimensions—density, mixed use, centrality, and street block size/road network density. The data for several larger MSAs were originally at the metropolitan division level, which were then reaggregated using population weights
Tsai’s Sprawl Index (first-difference model only)	Calculated from Census 2000 and 2010 SF1 data	SI_i = 50 × {[S_i – D_i / 100] + 1}, where SI_i is the sprawl index, and S_i and D_i are the percentage of the total MSA population in the bottom and top 20% dense subareas, respectively. Sprawl scores were calculated for 2000 and 2010
Marginal Land Consumption (first-difference model only), based on Paulsen (forthcoming)	Calculated using Census 2000 and 2010 SF1 data	Land area per housing unit, in square meters; calculated for 2000 and 2010
Particulate Pollution Emissions per Capita, lbs. (PW)	National Emissions Inventory (2014); County data aggregated to MSA level; 2011	Total per capita emissions in pounds of hazardous pollutants, the sum of volatile organic compounds, NO_x, sulfur dioxide, carbon monoxide, ammonia, and small and large particulate matter (particles less than 2.5 and 10 µm in diameter, respectively). County data aggregated to MSA level
CO₂ Emissions per Capita from Onroad Sources, kg. (CO₂)	The Vulcan Project (see Parshall et al. 2010)—Data publicly available (Vulcan Project 2014); County data aggregated to MSA level; 2002	Total per capita onroad carbon dioxide emissions in kilograms from highway sources
Location Affordability Index—Housing Cost as Percentage of Income for Homeowners (HDO)	Local Affordability Index (Department of Housing and Urban Development [HUD] 2014); MSA level; 2006–2010 American Community Survey (ACS) data	Housing cost as a percentage of income for average household size and median income level for region (i.e., typical household affordability), among owners
Location Affordability Index—Housing Cost as Percentage of Income for Renters (HDR)	Local Affordability Index (HUD 2014); MSA level; 2006–2010 ACS data	Housing cost as a percentage of income for average household size and median income level for region (i.e., typical household affordability), among renters

Note. All data used in this project is available at http://goo.gl/N0TZUD.

For the bivariate correlations, I include Ewing’s composite index and the subindices, which separately measure density, centrality, mixed-use development, and street connectivity. With respect to Lopez and Hynes’ index, I recalculated figures at the block level in addition to using their original calculations, which are at the tract unit level. I also include population-weighted density at both the tract and block level, the preferred measure of sprawl used by the census and also employed in other research (e.g., Rappaport 2008). As the results show (see Table 7 for a pairwise correlation matrix), the index devised here at the block unit level correlates most strongly with each outcome of interest. With respect to scale effects, measuring the index at the block rather than the tract level results in modest but noticeable differences in the strength of correlations. Also notable are the very high correlations between the index devised here and Ewing’s mixed-use subindex (r = −.858; the coefficient is negative because for Ewing’s measure, higher values signify less sprawl, while the opposite is true of the index outlined here).

Table 7.

Pairwise Correlation Matrix.

	IN10	INCT	CPB	CPT	LH	LHB	ECS	ED	EM	ECN	EST	PW	CO₂	HDO	HDR
IN10	1.000	—	—	—	—	—	—	—	—	—	—	—	—	—	—
INCT	.963	1.000	—	—	—	—	—	—	—	—	—	—	—	—	—
CPB	−.812	−.847	1.000	—	—	—	—	—	—	—	—	—	—	—	—
CPT	−.766	−.849	.937	1.000	—	—	—	—	—	—	—	—	—	—	—
LH	.911	.897	−.662	−.636	1.000	—	—	—	—	—	—	—	—	—	—
LHB	.928	.840	−.635	−.577	.897	1.000	—	—	—	—	—	—	—	—	—
ECS	−.694	−.686	.616	.595	−.607	−.633	1.000	—	—	—	—	—	—	—	—
ED	−.723	−.811	.900	.970	−.609	−.560	.550	1.000	—	—	—	—	—	—	—
EM	−.858	−.915	.797	.808	−.778	−.733	.608	.841	1.000	—	—	—	—	—	—
ECN	−.262	−.340	.407	.396	−.189	−.123	.543	.386	.339	1.000	—	—	—	—	—
EST	−.630	−.632	.630	.598	−.540	−.570	.650	.593	.677	.196	1.000	—	—	—	—
PW	.683	.663	−.562	−.526	.600	.591	−.416	−.520	−.622	−.106	−.428	1.000	—	—	—
CO₂	.565	.519	−.465	−.431	.504	.536	−.340	−.414	−.486	−.022	−.422	.486	1.000	—	—
HDO	−.632	−.632	.527	.536	−.543	−.500	.447	.461	.530	.086	.440	−.596	−.372	1.000	—
HDR	−.334	−.320	.301	.273	−.271	−.265	.141	.212	.277	−.086	.302	−.359	−.210	.775	1.000

Note. Insignificant correlations at the p < .05 level are in bold type. See Table 6 for abbreviations.

In the multivariate models (see Table 8), sprawl is strongly associated with carbon emissions, hazardous pollution, and housing affordability in the expected directions, net control variables. While these models are merely suggestive and do not illustrate causation, the elasticities between sprawl and each outcome described here are both significant and robust in magnitude: For every 10% increase in sprawl, there is an approximately 5.7% increase in per capita carbon emissions, a 9.6% increase in per capita hazardous pollution, and a 4.1% and 2.9% reduction in the owner and renter housing affordability index, respectively. For all models, including the sprawl index devised here resulted in a higher amount of explained variance than using other sprawl indicators (detailed results not shown). In the hazardous pollution and affordability models, Ewing and Hamidi’s centrality subindex was insignificant (which was not true of either their other subindices or the composite measure), which suggests that statistical renderings of morphology may be less desirable for modeling certain effects and could contribute to underestimating associations between form and outcomes when included in a multidimensional measure.

Table 8.

OLS Regressions.

	Carbon Emissions	Conventional Pollution	Homeowner Costs, % Income	Renter Costs, % Income
	B (SE)	B (SE)	B (SE)	B (SE)
Sprawl index	0.566 (0.114)***	0.963 (0.145)***	−0.405 (0.055)***	−0.292 (0.060)***
Median household income	0.307 (0.198)	−0.962 (0.330)**	0.554 (0.119)***	0.384 (0.140)**
% college educated	0.177 (0.073)*	0.167 (0.138)	−0.063 (0.040)	−0.048 (0.040)
Poverty rate	0.118 (0.083)	−0.203 (0.169)	0.242 (0.053)***	0.227 (0.074)**
Household size	0.092 (0.150)	−0.427 (0.243)^†	−0.337 (0.080)***	−0.323 (0.105)**
% in labor force	−0.366 (0.166)*	1.358 (0.428)**	−0.862 (0.134)***	−0.906 (0.140)***
Total land area	−0.021 (0.015)	−0.056 (0.022)*	0.002 (0.009)	0.003 (0.009)
Commuting by car (%)	0.555 (0.362)	−0.310 (0.364)	—	—
Gas prices	−0.676 (0.267)*	0.587 (0.363)	—	—
Housing growth 2000–2010	—	—	0.023 (0.010)*	0.064 (0.008)***
Constant	−0.276 (2.897)	0.976 (2.546)	6.461 (0.769)***	6.196 (0.925)***
F	16.43***	48.07***	33.15***	27.35***
R ²	.473	.655	.678	.622
n	150	150	149	149

Note. Gas prices are determined at the state level; MSAs which lie on borders are assigned values based on the location of their central city (e.g., gas prices for St. Louis are assigned values for Missouri). Housing growth is the percentage change in housing units from 2000 to 2010 (2006–2010 five-year averages are used for the latter). Control variables for carbon and particulate pollution models are from 2000 SF3 and 2008–2012 five-year ACS averages, respectively. For housing outcomes, variables are based on 2006–2010 ACS averages to provide consistency with HUD’s calculations. All dependent and independent variables are logged so as to indicate elasticities. All models are reported with robust standard errors. Housing models have 149 cases because New Orleans lost housing units; other models including it offered substantively similar results. Robust regressions, which attenuate the influence of outliers, also delivered substantively similar results. OLS = ordinary least squares; HUD = Department of Housing and Urban Development.

†

p < .1. *p < .05. **p < .01. ***p < .001.

Fixed-effects statistical approaches are more robust to latent confounding factors because they effectively control for time-invariant individual metropolitan characteristics which are difficult (and often impossible) to model. I focus on owner-occupied home values from 2000 to 2010 (the latter based on 2008–2012 five-year ACS averages), adjusted for inflation, for the first-difference models presented here. I separately use the newly developed index, Lopez and Hynes’ sprawl measure, and weighted population density, all calculated at both the tract and block level, as predictors. I also include two additional sprawl measures described earlier in the paper based on the work of Tsai (forthcoming) and Paulsen (2014) which quantify changes over time. The former expresses the change in proportion of metro population in low- and high-density subareas over the time period, whereas the latter is similar to Paulsen’s marginal land consumption indicator, and is a measure of the change in land consumption per housing unit. I do not include Ewing and Hamidi’s sprawl measure because the methodology changed from 2000 to 2010, so for all practical purposes, they are not strictly comparable across years.⁸ The results illustrate that density-based measures are significantly related to changes in housing price in the expected direction. The index devised here was the only density-based measure to achieve significance using both the tract and block unit level (see Table 9 for results).

Table 9.

First-Difference Models: Housing Values (Owner-Occupied Units), 2000–2010.

	B (SE)	B (SE)	B (SE)	B (SE)	B (SE)	B (SE)	B (SE)	B (SE)
Median household income	1.100 (0.237)***	1.118 (0.210)****	1.225 (0.240)***	1.146 (0.220)***	1.251 (0.228)***	1.199 (0.241)***	1.239 (0.232)***	1.091 (0.220)***
College educated	0.513 (0.042)***	0.508 (0.044)***	0.489 (0.043)***	0.498 (0.043)***	0.489 (0.043)***	0.486 (0.043)***	0.488 (0.043)***	0.511 (0.044)***
Poverty rate	−0.185 (0.103)^†	−0.186 (0.094)^†	−0.162 (0.101)	−0.188 (0.095)*	−0.168 (0.101)^†	−0.180 (0.104)^†	−0.165 (0.101)	−0.191 (0.098)^†
Housing units	0.239 (0.154)	0.298 (0.142)*	0.394 (0.170)*	0.333 (0.138)*	0.525 (0.157)**	0.446 (0.144)**	0.462 (0.150)**	0.285 (0.144)*
Household size	−0.072 (0.150)	−0.059 (0.150)	−0.045 (0.151)	−0.079 (0.153)	−0.031 (0.146)	−0.030 (0.145)	−0.027 (0.145)	−0.079 (0.132)
% in labor force	0.210 (0.119)^†	0.183 (0.124)	0.201 (0.124)	0.226 (0.122)^†	0.187 (0.121)	0.202 (0.121)^†	0.181 (0.122)	0.207 (0.118)^†
Total land area	−0.054 (0.110)	−0.108 (0.094)	−0.186 (0.124)	−0.113 (0.097)	−0.319 (0.118)**	−0.234 (0.101)*	−0.238 (0.111)*	−0.120 (0.087)
Sprawl index—Blocks	−0.907 (0.404)*	—	—	—	—	—	—	—
Sprawl index—Tracts	—	−1.409 (0.464)**	—	—	—	—	—	—
Lopez and Hynes—Blocks	—	—	−0.092 (0.145)	—	—	—	—	—
Lopez and Hynes—Tracts	—	—	—	−0.346 (0.104)**	—	—	—	—
Tsai	—	—	—	—	0.200 (0.122)	—	—	—
Marginal land	—	—	—	—	—	−0.097 (0.069)	—	—
Weighted density—Blocks	—	—	—	—	—	—	0.017 (0.033)	—
Weighted density—Tracts	—	—	—	—	—	—	—	0.398 (0.124)**
Constant	−0.806 (4.268)	1.107 (3.676)	−6.825 (3.602)^†	−4.575 (2.902)	−9.022 (2.956)**	−6.471 (3.257)*	−7.939 (2.908)**	−7.841 (2.702)**
F	70.21***	66.12***	59.74***	63.74***	62.11***	60.20***	59.78***	64.04***
n	150	150	150	150	150	150	150	150

Note. All 2010 figures are based on 2008–2012 five-year ACS averages. Demographic data were collected at the county level and reaggregated using population weights to conform to 2013 OMB metropolitan statistical area boundaries. All models utilize robust standard errors and logged variables. OMB = Office of Management and Budget.

†

p < .1. *p < .05. **p < .01. ***p < .001.

For changes in home values, indices measured at the tract unit level achieved higher levels of significance in general. This could be because “exurban”-type blocks either drop out of the calculation after failing to meet lower-bound thresholds or are counted as higher-density development, while the larger tracts they constitute are counted as low-density development, the growth of which more closely corresponds with the recent price shocks in regional home values. Yet using foreclosure data from HUD which measures rates of delinquency in the 18 months preceding June 2008, the density-based measures generally produced similar results when restricted to highly distressed (or relatively stable) housing market subsamples.

I deliberately refrain from discussing policy implications related to these outcome measures as the results are merely attempts at illustrating initial face validity. Nevertheless, future research which traces the links between urban form and its effects would clearly benefit from exploring the range of options currently available and by performing formal sensitivity analyses where feasible which will help determine which measures most precisely model specific outcomes. This extends to exploring how different units of analysis affect results, with the possibility that different aerial units may be more or less suitable for specific research questions. Because of the sheer number of distinct social, economic, and welfare outcomes that are currently associated with sprawl, it is highly unlikely that one index will ever be sufficient to explaining every phenomenon. Rather, each measure has its own unique set of advantages and disadvantages which should be carefully weighed in empirical research. Research on the measurement of sprawl could likewise benefit from becoming better integrated in a general sense, whereby new measures are compared with existing ones and used to predict common outcomes in establishing initial face validity so as to demonstrate their value for further research applications.

Conclusion

In this article, I calculated a new density-based index of sprawl and examined recent changes in the top 150 MSAs in the United States. Although metropolitan regions in the United States continued to grow more sprawling over this time frame—that is, construction of detached single-family housing significantly outpacing that of multiunit developments—this trend could be changing in light of recent trends in suburban growth, and further research would do well to update these results as newer data become available. I also compared this index with other popular measures in drawing associations between sprawl and selected outcomes in carbon emissions, hazardous pollution, housing affordability, and home values. Far from arguing that this or any sprawl measure is definitive across a range of applications, I call for a greater integration of work on urban form and the effects of the built environment so as to more comprehensively address the basis for which various measures are used and how distinct approaches may deliver different results. Specifically, research on measurement could integrate established outcomes of sprawl into their work, and when possible, compare new calculations with established indicators. Likewise, econometric work linking urban form to its effects could examine how alternative indicators perform differently as part of a sensitivity analysis, ensuring that models are not underestimating the relationships between sprawl and outcomes in travel behavior, pollution, or public health. Doing so would build confidence both in our conceptualization and operationalization of sprawl, and the research linking it with the litany of relevant outcomes concerning researchers and policy makers.

Footnotes

Acknowledgements

Thanks to Eman Abdelhadi, Delia Baldassarri, Julia Behrman, Ned Crowley, Sara Duvisac, Noah Ennis, Jon Gordon, Jen Jennings, and Rob Riggs for comments and suggestions. Much gratitude to Reid Ewing and Shima Hamidi for supplying their data in advance of online publication.

Author’s Note

This article benefited tremendously from the incisive criticism of the anonymous reviewers, along with the Editors’ guiding hands. All remaining errors are, of course, mine alone. All data used in this project is available at .

Declaration of Conflicting Interests

The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author received no financial support for the research, authorship, and/or publication of this article.

Notes

Author Biography

Thomas Laidley is a PhD Candidate at New York University. His current work focuses on the social, economic, and environmental effects of urban form and the spatial morphology of cities.

References

Alonso

1964. Location and Land Use. Cambridge: Harvard Univ. Press.

Anthony

2004. “Do State Growth Management Regulations Reduce Sprawl?” Urban Affairs Review 39 (3): 376–97.

Bento

A. M.

Cropper

M. L.

Mobarak

A. M.

Vinha

2005. “The Effects of Urban Spatial Structure on Travel Demand in the United States.” The Review of Economics and Statistics 87 (3): 466–78.

Bereitschaft

Debbage

2014. “Regional Variations in Urban Fragmentation Among U.S. Metropolitan and Megapolitan Areas.” Applied Spatial Analysis and Policy 7 (2): 119–47.

Brueckner

J. K.

2001. “Urban Sprawl: Lessons from Urban Economics.” Brookings-Wharton Papers on Urban Affairs 2001:65–97.

Buliung

R. N.

Kanaroglou

P. S.

2006. “Urban Form and Household Activity-Travel Behavior.” Growth and Change 37 (2): 172–99.

Burchfield

Overman

H. G.

Puga

Turner

M. A.

2006. “Causes of Sprawl: A Portrait from Space.” The Quarterly Journal of Economics 121 (2): 587–633.

Clark

T. A.

2013. “Metropolitan Density, Energy Efficiency and Carbon Emissions: Multi-attribute Tradeoffs and Their Policy Implications.” Energy Policy 53:413–28.

Covington

K. L.

2009. “Spatial Mismatch of the Poor: An Explanation of Recent Declines in Job Isolation.” Journal of Urban Affairs 31 (5): 559–87.

10.

Cutsinger

Galster

Wolman

Hanson

Towns

2005. “Verifying the Multi-dimensional Nature of Metropolitan Land Use: Advancing the Understanding and Measurement of Sprawl.” Journal of Urban Affairs 27 (3): 235–59.

11.

Department of Housing and Urban Development. 2014. “Location Affordability Portal: Understanding the Combined Cost of Housing and Transportation.” http://www.locationaffordability.info/lai.aspx (accessed January 14th, 2015).

12.

Detroit Future City. 2014. “The Land Use Element: The Image of the City.” http://detroitfuturecity.com/wp-content/uploads/2014/02/DFC_LandUse_2ndEd.pdf (accessed June 15, 2014).

13.

Downs

1999. “Some Realities About Sprawl and Urban Decline.” Housing Policy Debate 10 (4): 955–74.

14.

Ewing

Cervero

2010. “Travel and the Built Environment: A Meta-analysis.” Journal of the American Planning Association 76 (3): 265–94.

15.

Ewing

Hamidi

2014. “Smart Growth America: Measuring Sprawl 2014.” http://gis.cancer.gov/tools/urban-sprawl/ (accessed April 25, 2014).

16.

Ewing

Pendall

Chen

2003. “Measuring Sprawl and Its Transportation Impacts.” Transportation Research Record 1831:175–83.

17.

Ewing

Rong

2008. “The Impact of Urban Form on U.S. Residential Energy Use.” Housing Policy Debate 19 (1): 1–30.

18.

Frey

W. H.

2014. “Brookings Institute: Will This Be the Decade of Big City Growth? ” http://www.brookings.edu/research/opinions/2014/05/23-decade-of-big-city-growth-frey (accessed June 15, 2014).

19.

Gaigné

Riou

Thisse

J. F.

2012. “Are Compact Cities Environmentally Friendly?” Journal of Urban Economics 72 (2): 123–36.

20.

Galster

Hanson

Ratcliffe

M. R.

Wolman

Coleman

Freihage

2001. “Wrestling Sprawl to the Ground: Defining and Measuring an Elusive Concept.” Housing Policy Debate 12 (4): 681–717.

21.

Glaeser

E. L.

Kahn

M. E.

2004. “Sprawl and Urban Growth.” Handbook of Regional and Urban Economics 4:2481–527.

22.

Glaeser

E. L.

Schuetz

Ward

2006. “Rappaport Institute of Greater Boston: Regulation and the Rise of Housing Prices in Greater Boston.” http://www.hks.harvard.edu/var/ezp_site/storage/fckeditor/file/pdfs/centersprograms/centers/rappaport/policybriefs/brief_housingregs.pdf (accessed April 29, 2014).

23.

Grazi

van den Bergh

J. C. J. M.

van Ommeren

J. N.

2008. “An Empirical Analysis of Urban Form, Transport, and Global Warming.” The Energy Journal 29 (4): 97–122.

24.

Guerra

Cervero

2011. “Cost of a Ride: The Effects of Densities on Fixed-Guideway Transit Ridership and Costs.” Journal of the American Planning Association 77 (3): 267–90.

25.

Gurney

K. R.

Mendoza

D. L.

Zhou

Fischer

M. L.

Miller

C. C.

Geethakumar

de la Rue du Can

2009. “High Resolution Fossil Fuel Combustion: CO₂ Emission Fluxes for the United States.” Environmental Science & Technology 43 (14): 5535–41.

26.

Hankey

Marshall

J. D.

2010. “Impacts of Urban Form on Future US Passenger-Vehicle Greenhouse Gas Emissions.” Energy Policy 38:4880–87.

27.

Howell-Moroney

2007. “Studying the Effects of the Intensity of US State Growth Management Approaches on Land Development Outcomes.” Urban Studies 44 (11): 2163–78.

28.

Irwin

E. G.

Bockstael

N. E.

2007. “The Evolution of Urban Sprawl: Evidence of Spatial Heterogeneity and Increasing Land Fragmentation.” Proceedings of the National Academy of Sciences 104 (52): 20672–77.

29.

Jaeger

J. A. G.

Schwick

2014. “Improving the Measurement of Urban Sprawl: Weighted Urban Proliferation (WUP) and Its Application to Switzerland.” Ecological Indicators 38:294–308.

30.

Jaret

Ghadge

Reid

L. W.

Adelman

R. M.

2009. “The Measurement of Suburban Sprawl: An Evaluation.” City & Community 8 (1): 65–84.

31.

Lee

Kurisu

Hanaki

Forthcoming. “Development of the Compact City Index and Its Applications to Japanese Cities.” Urban Studies.

32.

Lopez

2014. “Urban Sprawl 1970-2010 Data.” http://www.russlopez.com/data (accessed January 14th, 2015)

33.

Lopez

Hynes

H. P.

2003. “Sprawl in the 1990s: Measurement, Distribution, and Trends.” Urban Affairs Review 38 (3): 325–55.

34.

McMillen

D. P.

2001. “Nonparametric Employment Subcenter Identification.” Journal of Urban Economics 50 (3): 448–73.

35.

Modarres

2011. “Polycentricity, Commuting Pattern, Urban Form: The Case of Southern California.” International Journal of Urban and Regional Research 35 (6): 1193–211.

36.

Muth

1968. Cities and Housing. Chicago: Univ. of Chicago Press.

37.

National Emissions Inventory. 2014. “The 2011 National Emissions Inventory.” http://www.epa.gov/ttn/chief/net/2011inventory.html (accessed January 14th, 2015).

38.

Office of Management and Budget. 2013. “Revised Delineations of Metropolitan Statistical Areas, Micropolitan Statistical Areas, and Combined Statistical Areas, and Guidance on Uses of the Delineations of These Areas.” OMB Bulletin No. 13-01. Washington, DC.

39.

Parshall

Gurney

Hammer

S.A.

Mendoza

Zhou

Geethakumar

2010. “Modeling Energy Consumption and CO2 Emissions at the Urban Scale: Methodological Challenges and Insights from the United States.” Energy Policy 38: 4765-4782.

40.

Paulsen

2013. “The Effects of Growth Management on the Spatial Extent of Urban Development, Revisited.” Land Economics 89 (2): 193–210.

41.

Paulsen

2014. “Geography, Policy or Market? New Evidence on the Measurement and Causes of Sprawl (and Infill) in US Metropolitan Regions.” Urban Studies 51(12): 2629-2645.

42.

Pendall

Puentes

Martin

2006. “The Brookings Institution Policy Briefs—From Traditional to Reformed: A Review of the Land Use Regulations in the Nation’s 50 Largest Metropolitan Areas.” http://www.brookings.edu/~/media/research/files/reports/2006/8/metropolitanpolicy%20pendall/20060802_pendall.pdf (accessed April 25, 2014).

43.

Rappaport

2008. “A Productivity Model of City Crowdedness.” Journal of Urban Economics 63 (2): 715–22.

44.

Rudel

T. K.

O’Neill

Gottlieb

McDermott

Hatfield

2011. “From Middle to Upper Class Sprawl? Land Use Controls and Changing Patterns of Real Estate Development in Northern New Jersey.” Annals of the Association of American Geographers 101 (3): 609–34.

45.

Speir

Stephenson

2002. “Does Sprawl Cost Us All? Isolating the Effects of Housing Patterns on Public Water and Sewer Costs.” Journal of the American Planning Association 68 (1): 56–70.

46.

Steinacker

2003. “Infill Development and Affordable Housing: Patterns from 1996 to 2000.” Urban Affairs Review 38 (4): 492–509.

47.

Stone

2008. “Urban Sprawl and Air Quality in Large US Cities.” Journal of Environmental Management 86 (4): 688–98.

48.

Stone

Hess

J. J.

Frumkin

2010. “Urban Form and Extreme Heat Events: Are Sprawling Cities More Vulnerable to Climate Change than Compact Cities?” Environmental Health Perspectives 118 (10): 1425–28.

49.

Stone

Jr. Mednick

A. C.

Holloway

Spak

S. N.

2007. “Is Compact Growth Good for Air Quality?” Journal of the American Planning Association 73 (4): 404–18.

50.

2011. “The Effect of Population Density, Road Network Density, and Congestion on Household Gasoline Consumption in U.S. Urban Areas.” Energy Economics 33: 445-452.

51.

Taylor

B. D.

Miller

Iseki

Fink

2009. “Nature and/or Nurture? Analyzing the Determinants of Transit Ridership Across US Urbanized Areas.” Transportation Research Part A 43 (1): 60–77.

52.

Tsai

Forthcoming. “Housing Demand Forces and Land Use Towards Urban Compactness: A Push-Accessibility-Pull Analysis Framework.” Urban Studies.

53.

Vulcan Project. 2014. “Complete U.S.” Vulcan Fossil Fuel CO₂ Data. http://vulcan.project.asu.edu/research.php (accessed June 10, 2014).

54.

Warner

Molotch

1995. “Power to Build: How Development Persists Despite Local Controls.” Urban Affairs Review 30 (3): 378–406.

55.

Zhao

Kaestner

2010. “Effects of Urban Sprawl on Obesity.” Journal of Health Economics 29 (6): 779–87.