- U.S. transit ridership has declined steadily over the past three years.
- Every size of urbanized area experienced a drop in transit.
- Absolute decline was far higher in largest metro areas.
- The advent of ride-sharing and car-sharing is likely partly behind falling transit use.
Transit ridership in the United States has been declining steadily over the past three years. One might conclude that this would impact fuel consumption, if the trend continues. But in fact, it is likely that fuel consumption may not drastically change, due to the enormous proportion of VMT already dedicated to personal vehicles as opposed to transit vehicles. This article will review the recent decline in transit ridership and present historical data, dating back to the dawn of streetcar railways in 1890 and coming all the way to the present, which places the current data in perspective. “Deep historical” data on transit is frequently overlooked, yet represents one of the most vital forces for clear understanding of whether and how transit trends can have a long-run impact on petroleum dependence.
One question that may arise concerns what modes lost transit ridership would switch to. Whether the lost transit ridership ends up being absorbed by any of several other modes—biking, walking, ride-sharing, or even just staying home—the fact remains that a transit vehicle will typically run its route (and consume fuel) no matter how many riders there are, so losing riders doesn’t necessarily mean losing transit vehicle trips, transit VMT, or transit fuel consumption. If a transit rider switches to a non-fuel-consuming mode, such as walking or biking, then there is no net change in fuel consumption, whereas a switch to private vehicle or ride-sharing means that fuel consumption can likely only increase, since all private vehicle and ride-sharing trips are optional and thus add fuel consumption at the margin. Notwithstanding this “switching analysis,” for several decades transit has constituted only about one percent or less of U.S. fossil fuel consumption for transportation, and only two percent of person-trips, so losing transit ridership nationwide is unlikely to have any more than a minimal impact on fuel consumption trends.
Falling ridership (2014 through 2017)
Annual ridership fell each year between 2014 and 2017 (see below), from nearly 11 billion rides in 2014 to just over 10 billion rides in 2017.
Ridership declined in an increasing fashion: 1.4 percent from 2014 to 2015; 1.9 percent from 2015 to 2016; and 2.9 percent from 2016 to 2017. Because U.S. population continued growing during these years, per-capita ridership declined as well, from 33.7 annual rides per person in 2014 to 31.0 in 2017.
This steady reduction was not confined to any one particular place, but rather affected cities of all sizes.
This steady reduction was not confined to any one particular place, but rather affected cities of all sizes: For the most recent years, 2016 to 2017, the National Transit Database of the Federal Transit Administration reveals that out of 290 total urbanized areas with available data, 229 experienced a decline in transit ridership (2016 to 2017), including 48 of the 50 most populated U.S. urbanized areas. On a percentage basis, every size of urbanized area experienced a drop in transit, although the absolute decline was far higher in the largest areas, since they have the greatest population and thus the highest number of rides. The following charts reveal the reduction in ridership across urbanized areas of varying population:
Mode share—the relative proportion of transit trips dedicated to bus, rail, or other modes—exhibits one of the clearest differentiators. While America’s national transit ridership has been split relatively evenly for many years between bus and rail, 90 percent of trips were lost from 2016 to 2017 due to changes in bus ridership, while only about 9 percent of rail trips were lost. Of the bus trip losses, nearly 90 percent occurred in urbanized areas of more than one million people. The following charts illustrate these joint trends.
While America’s national transit ridership has been split relatively evenly for many years between bus and rail, 90 percent of trips were lost from 2016 to 2017 due to changes in bus ridership.
In viewing these numbers, the critical question is—what’s next? Does this represent a new equilibrium, or will these changes continue to occur indefinitely, leading to a slow collapse of America’s transit system? Moreover, what are the likely factors behind the decline?
To address the first questions, regarding the equilibrium or “natural” level of transit ridership, it helps to look at historical data. The United States has experienced remarkable changes over time in both ridership levels and mode share, as seen in the following graphs.
From 1890 through 1950, transit was a predominant mode of travel, with very high levels of per-capita ridership as well as high ridership in absolute number of annual trips. In recent years, from roughly 1995 until the present, there has been a moderate, limited revival in ridership in raw numbers, but per-capita there has only been a stabilization at a much lower level than earlier in the 20th century.
In terms of mode of transit travel, initially nearly all transit occurred by rail, whereas over time motor buses grew in importance. The following graph shows mode-share over the entire recorded history of transit, from 1890 through 2017.
Indeed, from a rail-only, streetcar-dominated system in the “early days” of the late 1890s and early 1900s, motor buses steadily caught up, becoming the dominant mode of transit by the 1960s. Nevertheless, and pertinent to the discussion of what has happened from 2014 through 2017, rail has spent much of the most recent half-century playing “catch-up.” Many urbanized areas such as Denver, Austin, Portland, Los Angeles, Houston, and others, traditionally seen as automobile-dominated, have quietly installed rail systems to connect their far-flung suburbs to the city center, or to connect various neighborhoods and districts in the denser city center.
The remarkable changes in mode-share over time can be seen by reviewing data for Denver, Colorado, an urbanized area that has participated in every era. In the early 1900s, Denver was a compact city with an extensive network of streetcars connecting dense inner suburbs to a bustling urban core; yet by the 1950s, nearly all of Denver’s streetcar rail lines had been removed, in favor of an extremely sprawling, Interstate-based, automobile-oriented metropolitan area. Nevertheless, since the early 2000s, Denver has made a major investment in rail lines connecting the by now far-flung suburbs.
Rare and valuable data on Denver mode share from over a century ago is available from a 1915 survey commissioned by the Denver Tramway Company (cited on the urban planning website Denver Urbanism), on the chosen travel mode for people entering and exiting the city center. This rare peek into the past can be contrasted with data from 2016, 101 years later, on how Denverites journey to and from work.
The data reveal a remarkable transformation from 89 percent of Denverites either using the streetcar or walking, in 1915, to 85 percent driving in 2016. This city-level shift has occurred at the national level as well, and has had profound implications for fuel consumption. In fact, during the first two decades of the twentieth century, when the number of automobiles sold passed the one million mark nationwide, and when VMT accelerated exponentially, America was already becoming an “automobile nation” even at the apex of its streetcar and trolley systems. The 13 percent automobile share for Denver, in 1915, masks a massive revolution in travel— “the great VMT transformation”—that occurred between 1902 and 1922.
The great VMT transformation
In the “great VMT transformation,” the vast majority of mileage traveled by forms of mechanized transport became transferred from streetcars to automobiles, at a time when few Americans actually owned automobiles. This occurred in parallel with several rounds of decentralization and suburbanization of American metropolitan areas, creating intense pressure towards ever-more auto-mobility. As far back as the 1910s, 1920s, and 1930s, automobiles were covering enormous amounts of annual mileage in comparison to transit. This spurred the far-flung, sprawling development of many metropolitan areas, especially western ones such as Los Angeles, which were at that time far younger with far more developable land than their counterparts in the eastern and central regions of the United States, such as Chicago or New York.
By as early as 1930, 98.7 percent of all VMT were from automobiles, rising gradually over time.
The groundwork had been laid for nationwide, long-term impact. By as early as 1930, 98.7 percent of all VMT were from automobiles, rising gradually over time even from this exceptionally high proportion, to reach the even higher proportion of 99.86 percent by 1980. The following graph reveals the long-term trend, every two or three decades from 1902 through 2010.
Explaining transit ridership declines, and understanding fuel consumption impacts
Transit agencies with aging fleets may be resisting replacement of their fleets because they perceive an inability to compete with Uber, Zipcar, or Lyft.
Regarding the decline in annual U.S. transit ridership, from 2014 through 2017, many possible reasons exist, such as the advent of ride-sharing and car-sharing; aging infrastructure; and low fare revenue proportion of total operating costs. In a detailed analysis, scholars Boris Pushkarev and Jeffrey Zupan listed several major “prices” that people take into account when deciding what travel mode to use, in their 1977 classic book, Public Transportation and Land Use Policy:
- Price in money
- Price in travel time
- Price in access time and effort
- Price in discomfort and disamenity
In all likelihood, all of these factors have been playing out negatively for public transit at once in the past few years. For instance, the advent of point-to-point, on-demand ride-sharing has changed the calculus of travel time, access time and effort, and comfort/amenity. This could work hand-in-hand with traditional factors that have discouraged transit usage. Transit agencies with aging fleets may be resisting replacement of their fleets because they perceive an inability to compete with Uber, Zipcar, or Lyft, which in turn could increase the “price in discomfort and disamenity,” reduce ridership, and create a spiralling cycle to even lower ridership. Similarly, given that most transit agencies have long received only partial revenue from fares, typically 30-60 percent, and the rest from state or municipal disbursements, long-running agency budget shortfalls or fiscal pressures on states and municipalities may make it difficult to even maintain existing levels of service, let alone introduce new services. This would, in turn, increase the passenger “price in travel time” and “price in access time and effort,” reduce ridership, reduce already low fare revenue, and place even more financial pressure on transit agencies.
Implications for fuel consumption
Nevertheless, even such insight into pervasive forces behind transit ridership declines cannot avoid the fact that changes in transit ridership likely have little to no effect on U.S. fuel consumption. Indeed, when fuel consumption is heavily dependent on VMT, yet transit has been only a tiny proportion of all VMT since as far back as 1930, it’s not clear that any foreseeable change at all, in transit ridership, could have any non-trivial impact upon fuel consumption.
In 2015, 99.13 percent of all U.S. motor fuel was consumed by private vehicles (rail transit mostly consumes electric power, so hasn’t been considered here) to go along with more than 99 percent of VMT from privately operated vehicles. Any type of purported effect on fuel consumption due to switching away from transit to other modes (biking, walking, driving, ride-sharing, etc.) is negated by the overwhelming concentration of VMT and fuel consumption in privately operated vehicles. Even on a level of individual, trip-by-trip switching analysis, lost transit ridership will not usually decrease the amount of transit fuel consumption, because transit vehicles will normally run anyways (but with fewer riders per vehicle), so that any transit trips switched to private vehicles will leave transit fuel consumption the same while increasing private vehicle fuel consumption. The above chart illustrates these points very clearly.
While transit vehicles have become substantially more fuel-efficient in recent years, declining ridership makes it very hard to take full advantage of these gains in fuel efficiency.
The history shown in this article argues, then, that the well-known shift of American transportation to personal automotive travel has been so profound and far-reaching that meaningful reduction in petroleum dependence via public transit usage appears implausible. While transit vehicles have become substantially more fuel-efficient in recent years, declining ridership makes it very hard to take full advantage of these gains in fuel efficiency, since transit consumes so small a proportion of overall fuel consumption, and moreover since the vast majority of personal trips are taken by private vehicle (on the order of several hundred billion private vehicle trips per year, versus ten billion transit trips). Non-trivial future impact on fuel consumption is more likely to be found elsewhere, maybe in dramatic increases in automotive fuel efficiency or through introduction of alternative drivetrains, or maybe in other, as-yet-unforeseen forces.
Nonetheless, it should not be overlooked that transportation revolutions do occur, as evidenced by the “great VMT transformation” between 1902 and 1922. What will it take for another such revolution to occur? Are the early signs at hand, even if imperceptible, or will the transportation system continue in its current configuration for the foreseeable future?