The Fuse

How Urban Design Affects VMT, Part II

by Hart Schwartz | March 20, 2018

This article is the second in a series looking at how urban design affects vehicle-miles-traveled (VMT).

  • Some urban planners promote a return to the mixed-use, high-density cities of the 1920s, to reduce VMT and fuel consumption.
  • However, population densities (as a proxy for mixed-use) have fallen so far since the 1920s that this goal may be unachievable.
  • The most realistic trend-line is “blended density,” whereby high-density and low-density coexist and offset each other.
  • Recognition of mixed, blended land uses is crucial for forecasting, because fuel consumption will vary according to city district.

The first part of the series on urban design and VMT explained that compact development movements (new urbanism, transit-oriented development) aim to impact transportation by either reducing the number of vehicle trips, the average distance of vehicle trips, or changing the travel mode choice away from personal vehicles. A key feature of their programs, especially as embodied by new urbanism, is to place different destinations closer together in a mixed-use pattern of development that utilizes densely interconnected streets. Moreover, if transit stations are placed at the nexus of mixed-use districts, shared travel can be further encouraged, in a design style known as transit-oriented development.

Collectively, new urbanism and transit-oriented development are known as “compact development movements” because they aim to limit the expansion of developed land in American metropolitan areas. How can the extent of these movements’ impact be evaluated, especially with respect to fuel consumption? For fuel consumption, the implication is that any reduction in vehicle trips or trip length, or to other travel modes, would reduce VMT, which in turn would cut fuel consumption.

Any reduction in vehicle trips or trip length, or to other travel modes, would reduce VMT, which in turn would cut fuel consumption.

Because it’s difficult to measure the prevalence of mixed-use design, population density has often been used instead. Dramatic declines in population density statistics over the past century make it apparent that instituting compact development programs similar to the 1920s might not be fully achievable. It would be difficult to increase density to levels seen a century ago. Recognizing this difficulty, the compact development movements have adopted a more realistic program of “blended density,” whereby they accept that low-density, land-consuming design will continue to exist in the same broader metropolitan areas where compact, mixed-use districts are developed. In a complex offsetting process that resembles the so-called “whack-a-mole” game, the driving-friendly suburban sprawl patterns of the mid and late 20th century will continue to exist alongside the walking-friendly and transit-friendly resurgence of mixed-use, higher density design.

The bottom line is that the most realistic trend-line for urban design over the next few decades involves population density either holding steady or slightly increasing, without a national trend of everywhere reverting to the highly dense and walkable cities of the early 20th century. While certain cities may certainly either develop new pockets of intense walkability, or maintain existing walkable zones, these “city within a city” districts will likely be surrounded by lower density, driving-friendly areas. Other factors aimed at reducing fuel consumption, such as autonomous vehicles, ride-sharing, or other unforeseeable forces, will need to reckon with this “blended” environment in managing their roll-out and implementation. Recognizing the blend of uses is a crucial element of accurate forecasting.

Can mixed-use design of the 1920s ever be recaptured? Long way to go

Looking at density statistics—as a proxy for mixed-use walkability—reveals just how far of an uphill climb is faced by compact development movements in attempting to utilize mixed-use, compact, walkable design to reduce driving and fuel consumption. Recall that these movements, especially new urbanism, openly declare that they seek to recapture urban design popular in the 1920s and earlier. This begs the question, what were population densities back then? The central cities of metropolitan areas have the longest tradition, of any land-use type, of relatively consistent boundaries, thus making them a suitable basis of long-term comparison. Between 1950 and 1980, a monumental transformation occurred in population density of central cities, rendering many urban landscapes vastly different in kind, in 2018, from those of a century ago.


The impact of the above graph is hard to overstate. Is it virtually impossible for subtle elements of new urbanist design, such as front porches or placing stores and apartments in the same building, to increase population density from the 2,754 figure of 2010, back to the 7,990 of 1920 or 8,238 recorded in 1930? Such vastly re-increased density seems implausible, especially considering that most of those people who left central cities moved to suburbs which often have densities under 500 people per square mile Change has been all the more monumental if one contrasts plummeting central city density against plunging transit ridership and soaring VMT. Extraordinary changes in travel patterns have been strongly correlated with falling central city density, as shown in the following graphs. Do these changes represent sweeping infrastructural changes that will be very hard to undo?



These graphs may fairly be called an exclamation point. In just thirty years, between 1950 and 1980, central city densities plunged, VMT soared, and transit ridership plummeted. In terms of VMT inputs—number of vehicle trips, average trip distance, travel mode choice—number of trips and trip distance cannot be measured back to the 1920s, due to lack of data. But a historic transformation of mode choice, from transit to driving, can be seen in the following graph.


In the face of this expanded historical background, the objectives of new urbanism and transit-oriented development assume new significance. The low, stable central city density every decade since the 1980s, and the historically low transit levels against historically high driving levels, strongly suggest that recent efforts to spread mixed-use design during the last 35 years have had, at most, limited impact. While it is certainly possible that small districts in larger metropolitan areas have increased in density, there has not been a nationwide impact perceivable in statistics.

Whack-A-Mole: “Blended Density” as the new nirvana of urban design

The immensity of the density challenge, along with the intensity of land-use contests in most metropolitan areas, have led compact development movements to define a limited goal of “blended density.”

The immensity of the density challenge, along with the intensity of land-use contests in most metropolitan areas, have led compact development movements to define a limited goal of “blended density.” The so-called “whack-a-mole” game represents the reasoning behind urban planners’ thinking. So long as market demand exists to build low-density developments, the builders of compact, higher-density developments can only hope to displace lower-density developments to other parts of the same metro area. For instance, a developer who wins the right to create a high-density, mixed-use, retail/commercial/residential/office district in an abandoned area of the city center may simply induce a separate developer, who had been considering that very same city center space, to build a low-density strip mall somewhere in the suburban hinterlands. If one then measures the overall density of the metropolitan area, it may be unchanged, or even have decreased, despite the success of the high-density project. Low and high densities co-exist within the same metro area and overall density undergoes minimal change.

This vexing dilemma has not gone unnoticed by compact development movements. Recognizing that they cannot necessarily achieve overwhelming area-wide increases in density due to the “whack-a-mole” dynamic, they hope instead to incrementally raise existing densities by marginally increasing the ratio of medium- or high-density developments to low-density sprawl. For example, building one high-density development for every low-density development would be better than building one high-density for every five or ten low-density. Adjusting the ratio could perhaps slow the rate of expansion of developed land area (a.k.a. sprawl), even if it would not stop sprawl in its tracks. Therefore, compact development literature often refers to “blended density” as a realistic, achievable outcome.

Impact on fuel consumption

Many commentators have speculated whether renewed “urbanization” will lead to large reductions in vehicle ownership and VMT.

After a decade-long pause, VMT began to grow again in the mid-2010s. Many commentators have speculated whether renewed “urbanization” will lead to large reductions in vehicle ownership and VMT. Others have speculated on the potential for autonomous vehicles, ride-sharing, or other new technologies to reduce VMT, especially in supposedly re-densified cities. But the historical data reveal just the opposite: Density has fallen massively and stayed low, relative to what it used to be in the heyday of walkable, mixed-use cities, a century and longer ago.

Consequently, the historic and durable transformations in density and mode choice between 1950 and 1980 highlight the scope of the challenge in using urban design to reduce fuel consumption. Even the most fervent proponents of mixed-use, compact development must labor within an environment of massively reduced density, and have acknowledged the challenge of systematically implementing mixed-use within a “whack-a-mole” environment of contested land use. If even these movements have adjusted their expectations downward to aim for “blended density,” and if density is so strongly correlated with vehicle travel, then what does the future hold?

Going Forward: Recognition of blended land-use realities

America’s metropolitan areas, home to nearly 90 percent of Americans, contain a hybridblended system of land-use where many different land-use types co-exist side-by-side.

Those interested in this topic should understand that, now and going forward, America’s metropolitan areas, home to nearly 90 percent of Americans, contain a hybrid, blended system of land-use where many different land-use types co-exist side-by-side. Whether the task is selling vehicles, introducing autonomous driving, promoting ride-sharing, policymaking, or something else, strategists and investors should understand the mixed system of land use that exists. Ironically, while at a small scale, the new urbanists can only introduce their version of “mixed-use” within narrowly defined and targeted districts, when seen from a larger bird’s-eye perspective America’s metropolitan areas appear profoundly mixed-use, in the sense that “mixed-use” broadly understood includes both walking-oriented and driving-oriented districts and that land uses can be functionally mixed even when not walkable. That is, even if uses are physically far away from one another, they may in fact be close (i.e. mixed) in terms of time, if the dominant mode of transport for that district is driving..

Understood in this manner, different levels of density, different street patterns, and different styles of buildings all co-exist nationwide. Recognition of this co-existence ought to be a cornerstone, a starting point for any serious discussions, in any realistic analysis of future fuel consumption and vehicle market trends.