ArticlePDF Available

Abstract and Figures

We document the falling socioeconomic status of American households without private vehicles and the continuing financial burden that cars present for low-income households that own them. We tie both these trends to the auto-orientation of America’s built environment, which forces people to either spend heavily on cars or risk being locked out of the economy. We first show that vehicle access remains difficult for low-income households and vehicle operating costs remain high and volatile. Using data from the Panel Study of Income Dynamics, Survey of Consumer Finances, and Census Public Use Microdata, we then show that in the last fifty years households without vehicles have lost income, both in absolute terms and relative to households with vehicles. We link these trends to the built environment by examining the fortunes of carless households in New York City, and particularly in Manhattan. Most of New York’s built environment did not change to accommodate cars, and in New York the fortunes of the carless did not fall. Our results suggest that planners should see vehicles, in most of the United States, as essential infrastructure, and work to close gaps in vehicle access.
Content may be subject to copyright.
Journal of Planning Education and Research
1 –18
© The Author(s) 2019
Article reuse guidelines:
DOI: 10.1177/0739456X18823252
Planning Research
Over the course of the twentieth century, the cost of shipping
goods fell while the money spent to move people rose. In
1890, moving a ton of goods cost more than 8 cents per mile;
by 2000, it was about 2 cents per mile (Glaeser and Kohlhase
2004). Over the same period, transportation’s share of total
household expenditures rose from essentially zero—people
mostly walked—to more than 17 percent. Automobility
explains both these changes. Internal combustion engines
made transportation overall faster and more convenient and
let shippers move goods across land instead of water, but it
also made household transportation more costly. Today the
car, in many ways, is private transportation spending. Almost
95 percent of household transportation expenditures go
toward purchasing, operating, and repairing personal auto-
mobiles (Glaeser and Kohlhase 2004; U.S. Bureau of Labor
Statistics [BLS], 2016a).
Widespread personal vehicle ownership has of course had
many consequences. This article focuses on the conse-
quences for people without vehicles. We start from the prem-
ise that personal vehicles have network externalities (Webber
1992). Vehicles become more valuable when more people
own them, because as vehicle ownership increases, so too do
the public and private investments that complement it. When
just one person owns a car, that car is more novelty than util-
ity, as society is not designed to make cars useful. Roads are
narrow, unpaved, and shared with pedestrians. Parking is
scarce. Destinations are close together. When most people
own cars, in contrast, auto ownership becomes an assump-
tion: tacitly or explicitly, people and institutions—employ-
ers, merchants, planners, and friends—believe that everyone
owns a vehicle and arrange private and public life around
that idea. In this world, roads are wider and reserved for cars,
parking is abundant, and daily schedules require travel to
places spread further apart. These changes make vehicles
more useful.
Network externalities cut two ways, however. As society
becomes more organized around vehicles, people without
vehicles risk being left out of society. This exclusion occurs
not just because people with cars can cover more ground
more quickly than people without them, but because changes
made to accommodate automobiles can affirmatively disad-
vantage other ways of moving around. The physical changes
that enable high-speed automobile travel, like low-density
development and expansive surface parking, also penalize
823252JPEXXX10.1177/0739456X18823252Journal of Planning Education and ResearchKing et al.
1Arizona State University, Tempe, AZ, USA
2Rutgers, The State University of New Jersey, New Brunswick, NJ, USA
3University of California, Los Angeles, Los Angeles, CA, USA
Initial submission, October 2017; revised submissions, May, September,
November 2018; final acceptance, December 2018
Corresponding Author:
David A. King, School of Geographical Sciences & Urban Planning, Arizona
State University, 975 Myrtle Avenue, Tempe, AZ 85281, USA.
The Poverty of the Carless: Toward
Universal Auto Access
David A. King1, Michael J. Smart2,
and Michael Manville3
We document the falling socioeconomic status of American households without private vehicles and the continuing financial
burden that cars present for low-income households that own them. We tie both these trends to the auto-orientation of
America’s built environment, which forces people to either spend heavily on cars or risk being locked out of the economy.
We first show that vehicle access remains difficult for low-income households and vehicle operating costs remain high
and volatile. Using data from the Panel Study of Income Dynamics, Survey of Consumer Finances, and Census Public Use
Microdata, we then show that in the last fifty years households without vehicles have lost income, both in absolute terms
and relative to households with vehicles. We link these trends to the built environment by examining the fortunes of carless
households in New York City, and particularly in Manhattan. Most of New York’s built environment did not change to
accommodate cars, and in New York the fortunes of the carless did not fall. Our results suggest that planners should see
vehicles, in most of the United States, as essential infrastructure, and work to close gaps in vehicle access.
transportation, urban form, urban history, transportation poverty, income inequality
2 Journal of Planning Education and Research 00(0)
low-speed modes by pushing destinations apart and making
roads less safe for people outside of vehicles.
This logic yields a prediction. Over time, as driving
becomes more necessary, anyone who can acquire a vehicle
will, even if doing so is financially burdensome. As a conse-
quence, the population without vehicles will become increas-
ingly disadvantaged, because only the most disadvantaged
people will be unable to afford cars. More poor people will
acquire automobiles, and more people without automobiles
will be poor. Owning a car will become an increasingly unre-
liable hallmark of affluence, while not owning one will
become an increasingly reliable sign of poverty.
Simple evidence from the U.S. Census seems to bear this
prediction out. Between 1960 and 2014, the U.S. poverty rate
fell from 24 percent to 14 percent. For households without
vehicles, however, the poverty rate slightly rose, from 42
percent to 44 percent. And within the population in poverty,
vehicle access increased sharply, from almost 60 percent to
just above 80 percent.1
It is this basic idea—that carlessness is increasingly asso-
ciated with poverty, even as more poor people have cars—
that motivates our analysis and motivates the recommendation
flowing from it. We proceed in the following steps. We
review the rise of American automobility, and in doing so we
quantitatively document the changes in household spending
that automobility triggered, and qualitatively describe the
network externality that resulted as roads, land uses, and pat-
terns of living changed, resulting in fewer places that accom-
modate nonauto travel. From there, we develop two important
stylized facts. First, as motor vehicles became more neces-
sary they did not, in many ways, become less expensive.
More low-income people acquired cars, but in doing so they
took on a large financial burden. Second, zero-vehicle house-
holds have become increasingly concentrated among people
with very low incomes. Households without vehicles are
falling further behind households with vehicles and are
poorer in absolute terms today than they were sixty years
ago. This falling socioeconomic status implies that—for
most people in most places—the high cost of owning a vehi-
cle is probably lower than the cost of living without one.
The article’s final section tests one implication of our
argument. If the growing socioeconomic difference between
households with and without vehicles owes in part to soci-
ety’s reorganization around automobiles, then these inter-
household differences should be smaller in places where that
reorganization was weakest. Our test of this idea is New
York, America’s largest city and the one that changed least to
accommodate automobiles. We compare New York with
both the United States as a whole and Los Angeles—
America’s second-largest city and one built primarily around
private vehicles.2 We show that in New York, the relationship
between vehicle ownership and socioeconomic status differs
substantially from the U.S. norm. Vehicle availability and
income remain highly correlated in New York, but the cor-
relation arises because households with cars are inordinately
affluent, not because households without vehicles are poor.
Unlike in Los Angeles or the United States overall, zero-
vehicle households in New York have gained ground abso-
lutely, and absence of a vehicle is not a reliable sign of
disadvantage. Our regression results suggest that in
Manhattan the probability of a vehicle-free household being
poor has actually fallen since 1960; in the United States,
overall it has risen.
Because most places are not Manhattan, and are unlikely
to resemble Manhattan anytime soon, our findings suggest
that planners should reconsider the role of personal vehicles,
particularly in debates about transportation and equity. A
great deal of planning around vehicles is focused, justifiably,
on vehicular externalities. Cars emit greenhouse gases, cre-
ate congestion, and pose real threats to public safety. But in
many if not most places, cars are also something close to a
necessity. We argue that planners might do well to treat them
like essential infrastructure, and work to guarantee basic
access to automobiles while also better regulating their use.
We briefly summarize this position, which is not original to
us but has arguably been a minority view in transportation
planning, in the next section.
Some caveats and methodological points before proceed-
ing: In what follows, we build on a sizable existing literature
that analyzes the relationship between auto access and eco-
nomic outcomes. This literature addresses the endogeneity
problem embedded in that relationship: acquiring a vehicle
can make poor households better off, but becoming better
off is what allows many poor people to acquire vehicles.
Our sense of this research is that even in the presence of
controls for two-way causality, vehicle access is associated
with less stress, more employment, and higher earnings
(Gurley and Bruce 2005; Pendall et al. 2015; Raphael et al.
In part because others have established the causal role of
automobiles in socioeconomic well-being, we do not dwell
on endogeneity problems here. Nor do we argue that vehicle
access alone explains America’s widening inequality.
Educational attainment, macroeconomic shifts, and tax poli-
cies favorable to the affluent all play larger roles in that trend.
We are concerned, again, with the narrower question of
whether households without vehicles are being left further
behind. As such, we emphasize the growing gap between
households that own vehicles and households that do not,
and in particular the role played by the built environment and
other institutions in mediating that gap. The endogeneity
problem that haunts research into travel and economic out-
comes arises because people with sufficient income select
into vehicle ownership. But they do so because automobiles
are themselves essential to income, and this entire relation-
ship occurs because society has been organized around auto-
mobiles. The automobile’s role in socioeconomic status—in
both directions—is substantial because society has made it
so. In places where society has not made it so, its role is
smaller and qualitatively different.
King et al. 3
Our focus on the mediating role of the built environment
justifies our use of New York City. We emphasize New York
for the same reason many transportation studies exclude it:
the city is sui generis, and its largely prewar built environ-
ment differs dramatically from the rest of the contemporary
United States. For our purposes, New York’s outlier status is
a feature, not a bug.
A final point is that in documenting the socioeconomic
trajectory of carless households, we rely on long-run time-
series data that sometimes have gaps or ambiguities. These
ambiguities can make interpreting the data difficult. To
address this concern, when we are able we validate the trends
we observe with multiple data sets. Although these data all
paint a similar broad picture, in their details they sometimes
differ, usually as a result of differences in the design of the
surveys. Fully explaining these differences would lengthen
the main article considerably, so most of this discussion is in
the appendix.
Motivation and Implication of the
The empirical work in this article is built around a particular
vision of transportation disadvantage: one that sees the
absence of vehicles as the source of disadvantage and the
provision of vehicles as an appropriate remedy. We outline
that vision here.
Cars were introduced as part of a larger wave of infra-
structure innovation—including indoor plumbing, house-
hold electrification, and refrigeration—that profoundly
transformed American life and to a great extent created the
modern United States. Like automobiles, these other innova-
tions were initially rare and reserved for the rich. In 1900, for
example, only 3 percent of households had electric lights
(Gordon 2016). Also like automobiles, these other innova-
tions create environmental externalities; burning fossil fuels
to generate electricity accounts for more U.S. carbon emis-
sions than does driving.3 Unlike automobiles, however,
access to these other innovations today is almost universal.
Virtually no households in 2018 lack electricity or indoor
plumbing—indeed, by 1970 every American household had
a mechanical refrigerator and electric lighting, and over 99
percent had flush toilets and running water (Gordon 2016).
Every year, however, almost 10 percent of households lack
an automobile. For some of these households, cars are genu-
inely unnecessary. For others, the absence of a car has high
economic and social costs.
Automobility differs in three important ways from other
essential infrastructure. But these differences strengthen
rather than weaken the case for universal access. First, basic
access to automobility is more difficult. Access to a home
generally implies access to water, electricity, and fuel.
Automobility, in contrast, demands the additional upfront
private investment of buying a vehicle. The typical American
home is connected to a road network, but full access to that
network requires the purchase of an expensive depreciating
Second, because of automobility’s network effects, lack
of access to a car carries a qualitatively different—and argu-
ably higher—price. Compared with its sister infrastructure
innovations, automobility has had more, and more conse-
quential, spillover impacts on land use and scheduling.
People without running water endure large hardships, but for
the most part those hardships do not arise because other peo-
ple do have running water. Some of the greatest costs of liv-
ing without a car, in contrast, arise because in most places
most people do have cars, and everyday activities thus
assume the presence of a vehicle. Lacking vehicle access
inhibits economic participation in ways that lacking plumb-
ing does not.
Third, despite the high penalty associated with lacking a
car, when people cannot afford automobility, governments
do not help them get it (Mattioli, Lucas, and Marsden 2017).
Americans who cannot afford electricity or heating fuel
receive assistance to help them buy it.4 Americans who can-
not afford private transportation are generally offered public
transportation—a very different and often inferior product
that leaves them with highly constrained mobility relative to
almost everyone else. Public transportation can and should
be improved, but if transportation disadvantage is largely a
function of vehicle access, the simplest and most direct way
to address that disadvantage might be providing auto access
to people without it. Universal access to electricity and clean
water were once policy priorities, integral components of
broader programs to increase inclusion and reduce social and
economic isolation (e.g., Wright 2010). Universal auto
access could advance similar goals.
Household Automobility and Its
Automobiles did not make Americans spend more time trav-
eling so much as they let Americans spend more money to
travel farther in the same amount of time. Scholarship related
to Marchetti’s Constant and similar research on the “univer-
sal travel time budget” suggests that for hundreds of years,
and possibly for millennia, people have devoted a roughly
equal share of each day to moving around (Crozet 2005;
Marchetti 1994; Zahavi and Talvitie 1980). For most of his-
tory, however, that movement was slow. It occurred either on
foot or through shared infrastructure (like horsecars or street-
cars) that could access only some places at some times.
Cars—privately owned infrastructure—allowed more speed,
more distance, and more flexibility in when and where peo-
ple went.
These advantages came, literally, at a price. The first cars
were artisanal goods for the rich. In 1900, only four thousand
vehicles were sold in the United States, and only eight thou-
sand had been registered. Per capita vehicle miles traveled
(VMT) in 1901 were just over one (Gordon 2016; Jakle and
4 Journal of Planning Education and Research 00(0)
Sculle 2008). But advances in vehicle manufacturing, along
with the advent of consumer credit, soon jumpstarted vehicle
ownership. In 1913, Ford by itself sold almost two hundred
thousand cars (Jakle and Sculle 2008). From 1906 to 1940,
the quality-adjusted price of automobiles fell 85 percent
(Raff and Trajtenberg 1996) and quality-adjusted prices as a
share of disposable income per capita fell even more (Gordon
2016). The result, somewhat paradoxically, was that house-
holds began spending more on transportation because auto-
mobiles became more affordable. Cars converted personal
transportation from something that principally cost time into
something that cost time and money.
Figure 1 shows one hundred years of household vehicle
ownership and transportation expenditures. Transportation’s
share of household spending rose sevenfold from 1917 to
2015, and more than doubled between 1934 and 1985, when
it peaked at 20 percent. Only housing, during this time, grew
faster as a share of household expenditures (U.S. Bureau of
Labor Statistics 2006, 2016b).
We take two other lessons from Figure 1. First is that rising
transportation expenditures were basically rising vehicle
expenditures. Transportation spending closely tracks vehicles
per capita, and the simple correlation between these trends
from 1935 to 2013 is 0.8. Data from the Consumer Expenditure
Survey (CES) reinforce this impression: in most years, more
than 90 percent and often more than 95 percent of household
transportation spending was devoted to cars (U.S. BLS 1970–
2017). Second, growth in vehicle ownership was widespread
rather than concentrated. In 1935, almost 60 percent of house-
holds did not own a vehicle but this share steadily plunged
and by 2000 was below 10 percent.
A central premise of our argument is that as private vehi-
cles proliferated, society changed in ways that rewarded their
ownership and penalized their absence. This idea is not new,
and others have documented it extensively elsewhere, so we
only summarize it here. In essence, vehicle ownership had
both a network effect and a feedback effect: mass auto owner-
ship transformed America’s landscape, and the landscape’s
transformation further encouraged mass auto ownership.
Both private and public institutions facilitated this transfor-
mation. Automobiles let private entrepreneurs develop busi-
nesses in ways and places that were previously unnecessary
or impossible. Filling stations and mechanic shops sprang up.
Advertising reinforced the status importance of owning auto-
mobiles. Roadside areas between cities and towns—once
devoid of economic activity—became places of commerce,
with businesses like motels, rest stops, and drive-in restau-
rants that were both designed for automobiles and largely
inaccessible to those without them (Jakle and Sculle 2008).
Governments, meanwhile, reacted to cars by changing the
quantity, dimensions, composition, and regulation of streets.
Governments provided coordination mechanisms: they put
Vehicles per capita
Transportation as % of household expenditure,
% of households without a vehicle
Transportation Expenditures
Left axis: Right axis:
Share HHs no vehicles Vehicles per Capita
1900 1920 1940 1960 1980 2000 2020
Figure 1. Household transportation expenditures and vehicle ownership, 1918–2015.
Source: Expenditure data are from U.S. BLS (2006) and U.S. Consumer Expenditure Surveys (2007–2016) (U.S. BLS 1970–2017). Vehicles per capita is
from Federal Highway Administration 1918–2015). Share of households without vehicles is assembled from multiple sources. The 1935 data are from the
U.S. BLS (2006). The 1950, 1972, and 1998 data are from 1955, 1972, and 1998 Survey of Consumer Finances. The 2002 data are from 2000 Decennial
Census, and data from 2006 to 2007 are from American Community Survey. HH = household.
Note: BLS = Bureau of Labor Statistics.
King et al. 5
white stripes in the middle of roads, created numbered routes
to identify major thoroughfares, and introduced stop signs
and traffic signals (Gordon 2016). They built new roads,
most visibly in the form of the Interstate system (DiMento
and Ellis 2011). Less visibly but arguably more important,
they paved existing roads. A 1901 Mercedes could drive 50
mph, but only on a smooth surface, and in 1901 more than 80
percent of U.S. roads were unpaved (Gordon 2016). Between
1900 and 2000, the U.S. road network doubled but the
amount of paved road grew eightfold (Forman et al. 2003;
National Research Council 2005).
State and local governments also widened streets, invented
jaywalking to keep pedestrians off streets now reserved for
vehicles, and redefined sidewalks as transportation corri-
dors—places to put the pedestrians no longer allowed on
streets (Loukaitou-Sideris and Ehrenfuecht 2010; Manville
2017; Norton 2011). Cities abandoned narrow grids with
short block patterns in favor of longer blocks with wider
curve radii, giving drivers longer sightlines that enabled con-
sistent fast driving (Barrington-Leigh and Millard-Ball
These interventions enabled speed. Speed let people live
outside dense urban centers without losing access to them,
and thus opened up cheaper outlying areas to residential
development. Average residential lots for new construction
grew threefold between the 1930s and 2008 (Hirt 2015) and
residential floor space per capita doubled from 1890 to 2010,
even as average household sizes fell (Moura, Smith, and
Belzer 2015). Purchasing speed, in short, let people purchase
space. More space, in turn, encouraged automobile travel, by
enabling and requiring higher speed.5
People who owned vehicles also needed places to store
them. Off-street parking transformed the landscape, arriving
partly in response to market demand and mostly in response
to parking requirements in zoning codes (Jakle and Sculle
2008; Shoup 2011). Low-density zoning with ample parking
pushed buildings apart from each other and back from the
street. Parking lots put asphalt and cars between pedestrians
and storefronts, and created curb cuts that let vehicles intrude
into sidewalks. Parcel-by-parcel, the built environment
transformed, and the new landscape that accommodated peo-
ple who drove disadvantaged people who did not. The same
changes that made vehicle travel more convenient made
walking, cycling, and mass transit less so.
As land uses dispersed density fell. Figure 2 shows the
population-weighted density of U.S. Census tracts from
1940 to 2013. In 1940 the median American household lived
in a neighborhood eight times as dense as the median house-
hold today. Today 25 percent of Americans live in neighbor-
hoods well below one thousand people per square mile. Only
5 percent of the contemporary U.S. population lives at 1940s
median density—and 40 percent of those people are in New
York City.
Low densities required driving and undermined other
forms of mobility. Figure 3 shows the strong correlation
between falling density, increased driving, falling carless-
ness, and falling transit use from 1960 to 2013 (1960 = 1.0).
Density declines in lockstep with carlessness. Transit use
falls sharply as more households get automobiles, then pla-
teaus. VMT per capita, meanwhile, rises steadily, nearly tri-
pling by the mid-2000s.
1930 1940 1950 1960 1970 1980 1990 2000 2010 2020
25th Percentile
75th Percentile
Figure 2. U.S. population-weighted residential density, persons
per square mile, 1940–2013.
Source: U.S. Bureau of the Census (1940–2010); American Community
Survey (2012–2016) five-year estimates.
1950 19601970 1980 1990 2000 2010 2020
Mean densityTransit trips per capita
VMT per capita Road Length per Capita
Percent with no car
Figure 3. Changes in density, transit use, driving, road capacity,
and carlessness, 1960–2013.
Source: U.S. Bureau of the Census, National Transit Database. Note: VMT
= vehicle miles traveled.
6 Journal of Planning Education and Research 00(0)
Driving increased in large part because places that could
be accessed without driving began to disappear. From 1969
to 2009, both the average number of household vehicles and
average household VMT grew 60 percent, even as average
household size fell. Over the same time, the average length
of household vehicle trips grew only modestly, by 9 percent.
The average number of household vehicle trips, however,
grew substantially, by 50 percent. Thus in the average house-
hold fewer people used more vehicles to drive more places,
even though the places they drove to were not on average
much farther away (Santos et al. 2011). Put another way, the
average distance between driving destinations did not sub-
stantially grow. What grew was the number of destinations
that required driving.
Voulgaris et al. (2016) examined thirty thousand U.S.
Census tracts and showed that by 2010 neighborhoods sup-
porting nonauto travel had all but disappeared. Neighborhoods
where solo driving accounted for less than half of all trips
accounted for less than 5 percent of all Census tracts. Half of
these tracts, moreover, were in the New York metropolitan
area. The vast majority of the country was built around the car.
The Burden of Automobility
A landscape that demands vehicles is a demanding landscape
for the poor, for the simple reason that driving is expensive
(Rice 2004; Smart and Klein 2018a; Thakuriah and Liao
2005). Driving has high fixed costs and volatile operating
costs; this section examines trends in both.
High Fixed Costs
Vehicles are expensive, and somewhat surprisingly in the last
fifty years vehicle sticker prices have not fallen much. The
quality-adjusted price of vehicles has fallen, but mostly
because quality has increased, not because prices have
declined. Gordon (2015) observes that a 1911 Model T was
priced at 11 percent of current-dollar household consumption
expenditures, while a 1950 Plymouth was 33 percent of cur-
rent-dollar expenditures, and a midrange 2014 automobile
was 21 percent—less than the 1950 vehicle but still almost
double the ratio of the Model T.6 The contemporary vehicle
is far superior to the Model T, but not more affordable.
Similarly, data from the U.S. Bureau of Economic
Analysis data show inflation-adjusted growth in average
expenditures per vehicle from just above $23,000 in 1967 to
$26,000 in 2014. Over the same time, real median household
incomes grew over from just below $51,000 to just above
$52,000 (U.S. Bureau of Economic Analysis, 1967–2014). If
anything, then, average vehicle expenditures have grown in
real terms and grown slightly as a share of median income.
Low-income households, of course, do not purchase aver-
age-priced vehicles, so their burden of automobile ownership
may be lower. Figures 4 and 5 use Survey of Consumer
Finances (SCF) data from 1955 to 2013 to examine this idea.
We first compute, for each household with vehicles, an aver-
age household vehicle value: the total value of household
vehicles divided by the total number of vehicles. We make
this calculation for the rich (the top income decile), the poor
(the bottom decile), and the middle class (the middle three
Across all vehicle-owning households, this computation
suggests that average household vehicle values grew mod-
estly, from just below $9,000 in 1955 to just more than
$11,000 in 2013 (all figures are expressed in 2013 dollars).
This small average change is deceptive, however, as it was
driven by substantial growth among the rich, whose average
vehicle values more than doubled. Vehicle values among the
poor and middle class, in contrast, moved much less. Vehicle
values among the bottom decile stayed largely constant and
rose slightly for the middle class.
1955 1963 1977 1983 1989 1998 2007 2013
Top income decile
Middle three deciles
Bottom decile
Figure 4. Average value of household vehicles, 1955–2013.
Source: Survey of Consumer Finances (all dollars constant).
1955 1963 1977 1983 1989 1998 2007 2013
Top income decile
Middle three deciles
Bottom decile
Figure 5. Average value of household vehicles as a percent of
household income.
Source: Survey of Consumer Finances.
King et al. 7
Figure 5 plots the average burden of vehicle ownership,
which we express as the ratio of household average vehicle
value to household income.7 The burden has fallen for all
groups, but in proportional terms it has fallen most for the
rich and least for the poor. The burden for the top decile
began small (at about 9% of income) and fell 77 percent, to
just 2 percent of income by 2013. Affluent households over
time bought more (and more valuable) vehicles, but their
incomes also grew sharply. The burden for middle-class
households fell 28 percent: these households added not only
more vehicles, and slightly more valuable vehicles, but also
more earners.
The poorest households stand out. In absolute terms, the
burden for this group has fallen the most (24 percentage
points), but in proportional terms it has fallen least (26%).
More important, the burden in 2013 remains extremely
high—more than 70 percent of income. Despite driving the
lowest-valued cars (and not seeing their average car values
rise over time), the ratio of vehicle value to income for
bottom-decile households was consistently four to five times
that of middle-income households. Furthermore, while the
burden for these households did fall from 1955 to 2013, more
households became subject to it, because car ownership in the
bottom decile more than doubled (from 26% to 55%). Thus
the prevalence of the burden grew as its severity declined.
Admittedly, vehicle values are not perfectly correlated
with purchase prices or payment burdens. For a vehicle of a
given value, individual customers can negotiate better prices
or use more advantageous financing instruments, making
that vehicle easier to own. Yet low-income people are less
able than others to secure advantageous prices and financing.
The poor often pay high interest rates (Barger 2003) and the
poor who are racial minorities may face the additional obsta-
cle of discrimination in financing terms (Cohen 2007; but
also see Charles, Hurst, and Stephens 2008).
The SCF shows that in 2013, 8.5 percent of U.S. house-
holds were unbanked (defined as lacking a checking account)
and 42 percent of these unbanked households did not have
vehicles, compared with just below 10 percent of households
with bank accounts. Some of this association is driven by low
incomes—30 percent of unbanked households were in the
bottom income decile, so low income might explain both the
lack of cars and the lack of bank accounts. Even within the
bottom decile, however, being unbanked is associated with
carlessness. Thirty-eight percent of bottom decile households
with checking accounts did not have a vehicle, compared with
59 percent of unbanked bottom decile households.8
Operating Costs
Measuring the burden of operating a vehicle is difficult,
because most estimates of driving “costs” (such as those
released annually by American Automobile Association
[AAA]) are actually assumptions about expenditures, and
expenditures are endogenous to income. Richer households
drive more expensive well-kept cars further distances, while
lower income households buy used vehicles, defer mainte-
nance on them, and drive them less (Rice 2004). The Panel
Study of Income Dynamics (PSID), for example, shows that
from 1999 to 2011, just 20 percent of cars owned by poor
families were new, compared with 48 percent of vehicles
owned by nonpoor families. And the National Household
Travel Surveys (NHTS) show that low-income people drive
less than high-income people (Santos et al. 2011). Precisely
because low-income households avoid costs by spending
less and traveling less, travel behavior across groups is an
inaccurate gauge of travel burdens.
We estimate relative travel costs by examining per-gallon
gasoline prices. Fuel is a relatively small part of total driving
costs (over the 27 years of AAA data, the fuel share averages
18%; AAA 1990–2017), but a larger share for the poor,
because it is a hard cost to avoid. The poor can spend less
than the rich to travel 100 miles if they drive cheaper vehi-
cles, drive without insurance, or drive vehicles with needed-
but-deferred maintenance. It is much harder, however, to
save money by buying cheaper gas. Most vehicles take the
same fuel, and the price difference between premium and
regular gasoline is relatively small (usually about 10%).
Drivers can save by consuming less gas, but doing so implies
less travel, not less costly travel. These facts, when coupled
with gas prices being easy to track, make gas prices a reason-
able proxy for the burden of driving, especially for the poor.9
Figure 6 shows, for 2000–2015 and for three different
points in the wage distribution (the top 10%, the median, and
the bottom 10%), the hours of work needed to buy enough
gas to drive 100 miles in a vehicle of average fuel economy.
The trends in this metric are determined by both gas price
volatility and wages’ ability to keep pace with that volatility.
2000 2005 2010 2015
Bottom decile
Top decile
Hours of work required to pay for
gas for 100 miles of driving
Figure 6. Labor needed to purchase 100 miles worth of
gasoline, at low, median, and high wages for a vehicle of average
fuel economy.
Source: Federal Highway Administration (1918–2015), U.S. Energy Information
Agency (2000–2015), U.S. Bureau of Labor Statistics (1970–2017).
Note: “Average fuel economy” is fleet mean miles-per-gallon for that year.
8 Journal of Planning Education and Research 00(0)
Because wages in the top decile kept pace with prices, from
2000 to 2015 top decile workers never needed more than
twenty-seven minutes of labor to buy enough gas to drive
100 miles.
Bottom-decile workers, in contrast, had lower and more
stagnant wages over these fifteen years, making them more
vulnerable to gas price volatility. Workers in the bottom
decile always needed at least one hour of labor to buy 100
miles’ worth of gas, and in many years they needed close to
two hours. Note that this calculation is for bottom decile
wages, meaning it could understate driving’s burden for bot-
tom decile households, since many poor households have no
regular wage-earners.
To summarize, since World War II, as the country reorga-
nized around the car, an increasingly large proportion of dis-
advantaged Americans have purchased automobiles. What
evidence we have suggests that these purchases were not
easy. Even low-value vehicles consume a large share of a
poor household’s budget, and even modest amounts of driv-
ing demand the equivalent of large amounts of low-wage
labor. These data reinforce other evidence (e.g., Blumenberg
and Agrawal 2014; Klein and Smart 2017) that low-income
households with vehicles constantly struggle to manage the
costs of their cars.
The Declining Fortunes of Zero-Vehicle
However difficult owning an automobile may be for lower-
income Americans, the cost of not owning one may be larger.
Most American carlessness appears to be involuntary: car-
less households often live in places where walking and tran-
sit use are difficult, which suggests that absence of a vehicle
is a constraint rather than a choice (Brown 2017). The value
of cars to low-income people is also evidenced by how
eagerly the poor acquire them. Low-income households
often convert even small increases in spending power—such
as increases in the minimum wage—into vehicle purchases.
They do so because automobiles can help people get and
keep employment and increase their earnings (e.g., Aaronson,
Agarwal, and French 2012; Cervero, Sandoval, and Landis
2002; Kawabata 2003; Lichtenwalter, Koeske, and Sales
2006; Pendall et al. 2015; Smart and Klein 2018b; Souleles
This association between car ownership and economic
outcomes appears to be strengthening over time. The PSID
shows that from 1969 to 1979, families with vehicles saw
their real incomes grow by an average of about 2 percent
year-over-year, while families without vehicles saw real
incomes fall about 2 percent.10 Between 2001 and 2015, this
gulf widened even as income growth among families with
cars slowed. The income increases of families with cars now
averaged only 1 percent every two years, but families without
cars saw average biennial declines of about 7 percent.
Figure 7 uses the PSID data to plot the falling real income
(in 2013 dollars) of families without vehicles and the widen-
ing gap between families with and without cars. From 1969
to 2013, the median income of families with cars rose 20
percent, from $55,000 to $62,000. Families without automo-
biles, meanwhile, saw their incomes fall 34 percent, from
$26,000 to $17,000. Car-owning families had incomes about
twice as high as the carless in 1969, but over three and a half
times as high in 2013. The real incomes of the carless fell
We use SCF data from 1955 to 2013 (not shown) to cor-
roborate this finding. In 1955, households with vehicles had
just over twice the median income of households without
vehicles ($39,000 to $19,000). By 2013, households with
vehicles had three times the median income of households
without ($53,000 to $17,000).11 In real income, households
without vehicles fell behind both relatively and absolutely.
Households without autos were poorer in 2013 than they
were in 1955.
The SCF also shows, perhaps more importantly, that car-
lessness has become increasingly concentrated at the bottom
of the income distribution (Figure 8). In 1955, more than 90
percent of households in the top income decile had automo-
biles, as did 75 percent of households in the middle class.
More than three-quarters of households in the bottom decile,
however, did not have automobiles. In the next sixty years,
the middle class converged with the top; carlessness among
middle-class households fell 80 percent, and vehicle owner-
ship became almost universal. Carlessness also fell in the
bottom decile, but by 2013, 45 percent of bottom decile
households remained vehicle-free.
1960 1970 1980 1990 2000 2010 2020
Median Income (No Car) Median Income (Has a Car)
Figure 7. Median income in 2013 dollars of families with and
without cars.
Source: Panel Study of Income Dynamics 1969–2013 (all dollars constant).
King et al. 9
Thus, where in 1955 the 75 percent of bottom-decile
households without an automobile accounted for 25 percent
of the nation’s carless households; by 2013, the 45 percent of
bottom decile households without an automobile—who were
less than 5 percent of U.S. households—accounted for 41
percent of American carlessness. Even as households in the
bottom decile became more likely to have cars, households
without cars became more likely to occupy the bottom decile.
Vehicles and Socioeconomic Status in a
Nonauto Built Environment
We have shown thus far that as the built environment changed,
more low-income people incurred a large burden to acquire
automobiles, while those unable to carry that burden fell
behind in both absolute and relative terms. If this falling
socioeconomic status of carless households does in fact result
from the transition to an auto-oriented built environment,
then in places where the built environment changed less, the
fortunes of carless households should not have declined as
much. The association between income and carlessness
should be different in places not organized around cars.
We test this idea using New York City. For an American
city, New York changed little to accommodate cars. The
average population-weighted density of a U.S. Census tract
in the United States fell by more than 50 percent between
1950 and 2013, and settled at about five thousand people per
square mile (see Figure 2). New York’s average tract density
in 2015 was more than ten times that amount, and its city-
wide population density grew from roughly twenty-five
thousand people per square mile in 1950 to more than twenty-
eight thousand in 2014, never falling below 23,500 in the
intervening years. The city’s streets remain mostly narrow
and organized primarily in a dense grid. Partly because of its
old housing (even today almost one-third of its housing stock
predates 1940) and partly because of its low minimum and
maximum parking regulations, parking in much of the city is
scarce and expensive. Only 30 percent of New York’s hous-
ing units include a parking space with the rent or purchase
price, compared with more than 90 percent of U.S. housing
units overall (Manville, Beata, and Shoup 2013). This rela-
tive absence of off-street parking gives the city fewer curb
cuts than other American places, which makes walking safer.
Manhattan, the densest of the city’s five boroughs, now has
only fifty gas stations (Nir 2016), despite having a daytime
population larger than twenty-seven states. New York, in
short, has long been a difficult place to own and operate a
vehicle. It should also, given our logic, be a place where
households without automobiles fare better, relative to both
the rest of the country and to households with automobiles.
Figure 9 uses IPUMS (Integrated Public Use Microdata
Series) Census microdata (Ruggles et al. 2017) to compare
households without vehicles in the United States, New York,
and Los Angeles. We include Los Angeles not only because
it is another large city but because it is in many ways New
York’s converse: an exemplar of auto-oriented urbanization.
Los Angeles has grown steadily denser since 1950, but its
zoning has ensured that even with that density, the automo-
bile has retained primacy (e.g., Chester et al. 2015; Eidlin
2005; Manville and Shoup 2005; Manville et al. 2013). We
1950 1960 1970 19801990200020102020
Top Income Decile Middle 3 Deciles Bottom Income Decile
7.5% of households
25% of carless
4.5% of households
41% of carless
Figure 8. Share of zero-vehicle households by income group, 1955–2013.
Source: Survey of Consumer Finances.
10 Journal of Planning Education and Research 00(0)
use data from 1960 to 2014, but we exclude 1970 because
the IPUMS database does not provide city-specific data for
that year.
The figure shows, first (in vertical bars), that since 1960
carlessness has been the dominant condition for New York
households. In every decade since 1960 over half of New
York households have been vehicle-free. In Los Angeles and
the United States overall, in contrast, carlessness has been
rare and getting rarer. The figure also shows (in horizontal
lines) that the economic fortunes of New York’s carless
households differ dramatically from those of carless house-
holds in the United States or Los Angeles. Since 1960, the
median household income of zero-vehicle households in Los
Angeles has fallen 14 percent in constant terms. In New
York, the median household income of the carless has risen,
and in absolute terms, it is more than double that of carless
households in Los Angeles ($36,600 to $15,400). The for-
tunes of the carless fell much less in the United States overall
than in Los Angeles—only about 1 percent—but this trend is
somewhat deceptive, since over this same time the carless,
and particularly the carless who are not poor, became
increasingly concentrated in New York. In 2014, New York
City had 3 percent of the nation’s households but 15 percent
of its carless households, and 18 percent of its nonpoor car-
less households. Removing New York makes the median
income of American households without cars fall 17 percent
between 1960 and 2014.
These data are consistent with the idea that carlessness
imposes a smaller economic penalty in places that are not
organized around automobiles. But to be clear, these findings
do not suggest that no income gap exists between New York
households with and without cars, or that vehicle ownership
in New York is uncorrelated with income. On the contrary,
such a gap exists, and the correlation is strong. In 2014, the
simple correlation between household income and carless-
ness in New York was –.18, equal to the correlation in Los
Angeles and larger than the correlation in the United States
overall (–.14). The correlation in New York, however, has a
different source: it arises because people with cars are more
likely to be rich, not because people without cars are more
likely to be poor. In Los Angeles, the average income of
households with vehicles is 10 percent higher than the
1960 1980 1990 2000 2010 2014
Median income of households without vehicles, $2014
% of households without a vehicle
USALos Angeles
% with no vehicle
median income, households with no vehicle
New York
Figure 9. Percent of households with no car and median income of households with no car, 1960–2014.
Source: American Community Survey and Decennial Census, Integrated Public Use Microdata Series.
King et al. 11
citywide average, while the average income of households
without vehicles is 67 percent below the citywide average.
The correlation between vehicles and income is thus driven
by the very low income of the carless. In New York, in con-
trast, the average income of households without vehicles is
20 percent below the citywide average, while the average
income of households with vehicles is 25 percent above it.
The income-vehicle correlation is thus driven by the inordi-
nately high incomes of households with cars.
The poor in New York, compared with those in Los
Angeles and the United States, are actually more likely to be
carless, because owning a car in New York is so expensive.
In 2014, 36 percent of Los Angeles’s bottom decile house-
holds had no vehicles, compared with 80 percent of New
York’s. But while the share of the poor who are carless is
higher in New York, the share of the carless who are poor is
lower, because many affluent households lack cars, and these
households actually account for most of the carless. In 2014,
41 percent of carless households in the United States were in
the country’s bottom income decile. In Los Angeles, this fig-
ure was 37 percent. In New York, it was only 18 percent.
Similarly, where only 1 percent of carless households in Los
Angeles (and 2% in the United States) were in the top income
decile, 7 percent of New York’s carless households were in
the top decile. And where only 1 percent of households in
Los Angeles’s top income decile do not have a car, 35 per-
cent of New York’s top decile households are zero-vehicle.
Carless households in New York, in sum, are seven times as
likely as those in Los Angeles to be in the top income decile,
and less than half as likely to be in the bottom. Relative to
Los Angeles and the United States overall, New Yorkers are
more likely to lack cars if they are disadvantaged, but less
likely to be disadvantaged if they lack cars.
Obviously one can object to these simple comparisons:
New York differs from Los Angeles and the United States in
ways beyond the auto-orientation of its built environment.
Our tabulations above do not control for New York’s cli-
mate, labor market, or governance, all of which might influ-
ence the median incomes of its residents without cars. We
can address this concern by making a comparison within
New York, using the borough of Staten Island. Staten Island
has the highest median income of all New York’s boroughs,12
and it shares a government, climate, and labor market with
the rest of the city. But it has a dramatically different (and
far more auto-oriented) built environment. The borough has
no subway connection, more off-street parking, and many
more single-family homes—Staten Island has only 5 per-
cent of New York’s housing units, but 20 percent of its
detached single-family homes. Indeed, as the three panels of
Figure 10 show, from 1980 to 2014, Staten Island—in the
level and trend of its density, car ownership, and housing
stock—has looked much more like Los Angeles than like
New York overall.
If our logic holds, the fortunes of Staten Island’s carless
should have declined more dramatically than those in New
York overall, even though Staten Island is more prosperous
than the city generally. As we have already documented
New York’s changes, Figure 11 compares Staten Island to
Manhattan, the borough closest to Staten Island in income
and least hospitable to automobiles. Recall that in New York
overall, the median income of carless households grew 12
percent from 1960 to 2014. In Manhattan, the median
income of carless households more than doubled. This
growth was slower than the income growth of Manhattan
households with automobiles (which more than tripled), but
zero-vehicle households nevertheless improved in absolute
terms, and since 1980 their income relative to that of vehicle-
owning households has been largely constant. Thus unlike
in the United States overall, the carless in Manhattan gained
ground absolutely and, in recent decades, did not lose
ground relatively.
The story is very different on Staten Island. Median
incomes grew 64 percent among Staten Island’s vehicle-
owning households between 1960 and 2014, but fell 41 per-
cent among its vehicle-free households. In 1960, Staten
Island households with vehicles had, on average, 1.5 times
the median income of households without. By 2014, they had
more than four times the median income of carless house-
holds. On Staten Island, unlike in Manhattan or New York
overall—but like the United States and Los Angeles—the
fortunes of the carless fell both relatively and absolutely.
A skeptic might still argue that we are reporting only uni-
variate descriptive trends. To address this concern, we use
the IPUMS data to estimate logit regressions that control for
an array of household-level characteristics. The dependent
variable in every equation is a binomial variable indicating if
the household is in poverty (1 = yes, 0 = no) and the inde-
pendent variable of interest is a dichotomous variable indi-
cating if the household lacks vehicle access. The regressions
also control for race, age, educational attainment, gender,
employment status, tenure status, type and age of housing
unit, household size, and year. We estimate regressions for
New York, Los Angeles, Manhattan, Staten Island, and the
United States.
We present the full regression results in the appendix.
Table 1 shows, for each place we analyze, the predicted prob-
ability of households with and without vehicles being in pov-
erty. We derive these predictions from the regressions: for
each place, we make a prediction for all years combined, for
1960 alone, and for 2014 alone. The predictions assume a
household that rents, does not live in a detached single-fam-
ily home, has one child, and that is in every other way “aver-
age”—that is, all other variables in the regression are held at
their means. The households are, as a result, largely artificial:
they are all mixed-race, multigenerational, and internally
heterogeneous in educational attainment. Our predictions are
thus not attempts to replicate actual poverty rates; our goal is
to simply hold confounding variables constant and compare
the independent association between poverty and vehicle
ownership across times and places.
12 Journal of Planning Education and Research 00(0)
1980 1990 2000 2010 2014
New York Los AngelesStaten Island
1980 1990200020102014
New York Los AngelesStaten Island
1980 1990 2000 2010 2014
New York Los AngelesStaten Island
(a) Population Density
(b) Share of Housing that is Single-Family Detached
(c) Share of Households without a Ve hicle
Figure 10. Density, housing, and vehicle ownership in New York, Los Angeles, and Staten Island.
King et al. 13
The results are largely consistent with the descriptive
findings above, although they do suggest that Manhattan
contributes a disproportionate share of what we observe in
New York overall. In each place, averaged across all years,
households without vehicles are about twice as likely to be
poor as households with vehicles (on Staten Island closer to
three times as likely: 21% to 8%). The absolute probability
that a zero-vehicle household will be poor, however, is sub-
stantially lower in New York City (20%), and especially in
Manhattan (16%), than it is in Los Angeles or the United
States (35% and 28%).
We also see that between 1960 and 2014, the probability
of a household without a vehicle being in poverty grew
everywhere, except Manhattan. On Manhattan, the probabil-
ity fell 20 percent, and the gap between households with and
without vehicles actually shrank. On Staten Island, in con-
trast, the gap between households with and without vehicles
widened considerably, and the probability that a household
1960 1980 1990 2000 2010 2014
households with vehicles
households without vehicles
Staten Island
Manhattan, with vehicle
230% growth
Staten Island, with vehicle
65% growth
Manhattan, without vehicle
105% growth
Staten Island, without vehicle
41% decline
Figure 11. Change in household income by vehicle availability, Manhattan, and Staten Island, 1960–2014.
Source: Integrated Public Use Microdata Series (1960–2014); 1970 is omitted for lack of city-level data.
Table 1. Probability of Household Poverty, by Vehicle Access.
All years 1960 2014
With vehicle No vehicle With vehicle No vehicle With vehicle No vehicle
Manhattan .08 .16 .08 .19 .10 .15
New York .09 .20 .05 .12 .13 .22
Staten Island .08 .21 .04 .08 .12 .29
Los Angeles .18 .35 .10 .25 .22 .42
The United States .15 .28 .18 .21 .20 .34
Note: Estimated from logit regressions using IPUMS samples from 1960, 1970, 1980, 1990, 2000, 2010, and 2014. Only regressions examining the United
States include 1970 data. Regression controls include number of adults, number of children, tenure, living in a detached single-family home, year fixed
effects, and the fraction of household members that are male, black, Hispanic, aged 65 years or above, aged 18 to 35 years, college-educated, and
employed. Predicted probabilities assume that the household rents, does not live in a detached single-family home, and has one child. All other variables
are held at their means. IPUMS = Integrated Public Use Microdata Series.
14 Journal of Planning Education and Research 00(0)
without a car would be poor more than tripled. Perhaps most
relevant to our thesis, in the United States overall the pre-
dicted probability that a household without a vehicle would
be in poverty grew dramatically between 1960 and 2014. In
1960, we see an 18 percent chance that a household with a
vehicle will be poor, compared with a 21 percent chance for
a household without a vehicle. By 2014, the probability of
poverty for a household with a vehicle had grown only
slightly (to 20%) while the probability for a household with-
out a vehicle had jumped to 34 percent.
These regressions are robust to a variety of perturbations,
and if we use linear regressions to examine household
income (as opposed to poverty), the results are consistent:
absence of a vehicle predicts large reductions in income.
Nevertheless, we do not wish to claim too much from these
regressions—certainly, the associations they capture are not
causal. But they broadly verify the trends we see descrip-
tively: outside the densest parts of New York, carlessness is
increasingly associated with economic distress.
The same changes to America’s society and landscape that
made driving easier have made not driving harder. In most of
the country, as a result, many disadvantaged people hover
just above carlessness, struggling (and periodically failing)
to maintain access to vehicles. Below these people on the
income ladder are households without automobiles, who
have lost ground in both absolute terms and in comparison to
households with automobiles. Both of these phenomena—
the burden of keeping a car and the income penalty associ-
ated with not doing so—appear to be artifacts of America’s
car-oriented built environment. The simplest evidence for
this proposition is that the socioeconomic penalty associated
with carlessness declines sharply in New York City.
Furthermore, the penalty declines most in that part of New
York City (Manhattan) that looks least like the rest of
America, and it declines least in the part of New York (Staten
Island) that looks most like the rest of America.
One obvious policy goal that flows from these findings is
to create more places like Manhattan, where automobiles are
not essential for economic success, and car-free living can
coincide with affluence. Places like Manhattan need not be
Manhattan, of course; they need only share Manhattan’s
attribute of being less oriented to automobility. This goal,
however, while undeniably important, is also indisputably
long-term, and pursuing it offers little help to transportation-
poor households today. For this reason, the long-range goal
of helping most nonpoor Americans drive less needs to be
paired with a shorter range goal of helping some poorer
Americans drive more.
Driving is America’s dominant form of transportation.
The United States has constructed a vast and comprehensive
public infrastructure that allows, and often requires, most
people to drive most places. Yet every year, close to one in
ten households (and sometimes more) cannot access that
public infrastructure, because they cannot afford the large
private investment it demands. Auto access is the starkest
transportation disparity in most of the United States. People
without automobiles cannot access employment, complete
errands, or generally move around in the same manner as the
vast majority of their fellow residents.
The simplest way to eliminate this disparity may be to
provide auto access. A policy to advance this goal could
involve buying people vehicles outright, but it need not go
that far and, in some cases, should not go that far. Some peo-
ple who lack vehicles are physically, cognitively, or legally
unable to drive—giving them cars would be unhelpful. The
goal instead should simply be to deliver the benefits of auto-
mobility to those who lack them. “Our central mission,” Mel
Webber (1992, 275) wrote over twenty-five years ago, “is to
redress the social inequities thrown up by widespread auto
use, and our central task is to invent ways of extending the
benefits of auto-like transportation to those who are pres-
ently carless.” The prospect of auto-like transportation for
the carless is perhaps now more real than it has ever been,
with the rise of Transportation Network Companies (TNCs)
like Uber and Lyft, as well improvements to traditional taxi
service. A program of universal access could include broad
subsidies to use such services.13
One objection to the goal of universal auto access is that
driving carries serious externalities, so planning and social
policy should promote less rather than more of it. By that
logic, however, society should also not help lower-income
households access electricity or heating fuel. Burning fossil
fuels to heat and power households generates greenhouse gas
emissions that rival or exceed those from cars. Yet we do not
deny heat and power to the poor in the name of sustainability.
Instead, we help the poor get basic access to electricity and
heating fuel, even as we work toward less overall consump-
tion of these resources by metering their use. Our current
approach to automobility is the exact reverse. We offer no
assistance with basic access, but heavily subsidize use for
those affluent enough to gain access on their own. As a result,
we have a small group of people who need vehicles and lack
them and a large group who have vehicles and use them
needlessly. A just and sustainable society would help the first
group drive more while encouraging the latter group to drive
less. Our status quo instead suppresses driving only by deny-
ing it to some of the people who need it most, even as it tac-
itly encourages low-value trips by the affluent. This approach
seems neither fair nor effective.
King et al. 15
Table A1. Associations with Household Poverty Status, 1960–2014.
New York
City Staten Island Los Angeles
The United
No HH vehicle 0.8101***
Fraction male −0.3058***
Fraction Black 0.3938***
Fraction Hispanic 0.5510***
Fraction aged 65+ years −0.9812***
Fraction with BA or Higher −0.8328***
Homeowner −0.8568***
Fraction employed −2.4845***
Detached single-family home 0.4458***
Fraction aged 18–34 years 0.0784
Total adults −0.3070***
Total children 0.2508***
Constant −0.4275***
N167,136 887,019 42,255 413,853 33,268,913
Pseudo-R2.249 .276 .333 .227 .235
Log Pseudo-likelihood 124,000 5,440,000 98,100 −7,206,329 20,200,000
Note: Standard errors in parentheses. All regressions include year fixed effects. Denominator for educational attainment is adults aged twenty-one years
or above. Denominator for employment is household members aged fifteen years or older. National regression includes 1970 Census data, others do not.
HH = household; BA = bachelor’s degree.
*p < .05. **p < .01. ***p < .001.
Data Sources
We draw on three primary data sources for our analysis: the
Panel Study of Income Dynamics (PSID), the Survey of
Consumer Finances (SCF), and the Census Integrated Public
Use Microdata Series (IPUMS). While at other points we
also draw on secondary data from the Bureau of Transportation
Statistics, the National Household Travel Surveys, and the
Consumer Expenditure Survey; it is these three data sources
that we manipulate the most; we briefly describe each one
Panel Study of Income Dynamics
The PSID is a long-running panel data set that began in 1968
with a sample of 18,000 individuals living in 5,000 thou-
sand families across the United States, and it has followed
the same individuals and their descendants ever since. The
sample has grown over time as descendants formed their
own PSID families, and also as Latinos and immigrants were
added to maintain the survey’s representativeness
(McGonagle et al. 2012). The PSID was originally designed
to track the effectiveness of poverty alleviation programs, so
it oversampled low-income and minority families, though
the PSID-provided weights correct for this oversample. Self-
identified family heads report data on all family members,
including children until they leave the home and become
financially independent, at which time they are recruited to
become stand-alone PSID family units. Data were collected
annually through 1997 and biennially since then. The PSID
focuses on sources of income and categories of expenditure,
including—for many waves of the data set—expenditures on
transportation including the components of automobile costs
such as auto finance, repair, fuel, and insurance. The surveys
16 Journal of Planning Education and Research 00(0)
from 1981 to 1997, however, did not include automobile-
related variables.
Survey of Consumer Finances
The SCF began in 1947 as a program of the University of
Michigan’s Economic Behavior Program and was ultimately
taken over by the Board of Governors of the Federal Reserve
and is now administered triennially by the National Opinion
Research Center on behalf of the Federal Reserve. Over this
time period, the survey’s questions, interview approach, and
weighting methodology have changed, but for our purposes
what matters is that in almost every year the survey tracks vehi-
cle ownership, vehicle value, and socioeconomic status. The
SCF, unlike the PSID, is not a panel and normally interviews
between six thousand and seven thousand households in each
wave. From 1989 forward, a multiple imputation algorithm
used to address missing data has made its summary extract data
appear to have five times more observations than actual inter-
views conducted: this imputation does not affect calculations
of summary statistics, which is what we use the SCF for.
Census Public Use Microdata (IPUMS)
The IPUMS is a much larger data set than either the PSID or
SCF: it is the microdata sample of the Decennial Census and
the American Community Survey. The IPUMS’s size (in most
years it is more than a million observations) allows more con-
fidence in its point estimates. More important, the greater size
and coverage permit us to make comparisons at the subna-
tional level (such as comparisons of Manhattan and Staten
Island) which are harder with the PSID and impossible with
the SCF. The disadvantage of the IPUMS is that it contains no
data about vehicle values or household expenditures on travel.
Also important is that the IPUMS tracks household access to
a vehicle, which need not indicate vehicle ownership (though
often it does). The PSID and SCF explicitly track a property
right in vehicles (through ownership or leasing) so estimates
of vehicle access vary across the surveys.
Board of Governors of the Federal Reserve System. n.d.
“The Survey of Consumer Finances.” Washington, DC.
McGonagle, Katherine A., Robert F. Schoeni, Narayan
Sastry, and Vicki A. Freedman. 2012 “The Panel Study of
Income Dynamics: Overview, Recent Innovations, and
Potential for Life Course Research.” Longitudinal and Life
Course Studies 3 (2): 1–21.
Daniel Kuhlmann and Sivaan Naaman provided excellent research
assistance. We thank anonymous reviewers for their helpful com-
ments. Any errors are the authors.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect
to the research, authorship, and/or publication of this article.
The author(s) disclosed receipt of the following financial support
for the research, authorship, and/or publication of this article: This
research was partially funded through the University Transportation
Research Center RF Grant No: 49198-36-27.
1. Calculated from IPUMS (Integrated Public Use Microdata
Series) U.S. Census data.
2. Los Angeles’s earliest development occurred around streetcar
lines (Wachs 1984). But the streetcars had declined by 1930,
and the region’s subsequent development was autocentric.
3. In 2015, electricity accounted for 29 percent of U.S. car-
bon emissions, compared with 27 percent for transportation.
Within the transportation sector, moreover, 17 percent of
emissions come from maritime, air, and rail transport (U.S.
Environmental Protection Agency n.d.).
4. For example, New Jersey offers up to $225 per year to help
residents pay for electricity and gas (State of New Jersey n.d.).
5. Fischel (2004) takes this argument further and contends that
automobiles actually spawned modern low-density zoning.
By liberating industrial uses and worker housing from prox-
imity to ports or rail lines, cars created footloose “undesir-
ables,” and prompted stricter land use controls. Minimum lot
sizes, in this telling, were both symptom and source of mass
6. Prior to the Model T, of course, vehicles were much more
expensive—often six to seven times the average income.
7. We also computed this burden as vehicle debt as a share of
income, with similar results (the main difference being that
households in lower income quintiles, because they often lack
credit, are less likely to carry vehicle debt).
8. This relationship is robust to multiple controls. Logit regres-
sions of bottom-decile households that predict carlessness
and control for income, age, the square of age, race, and sex,
for 2013 alone and all years in Survey of Consumer Finances
(SCF), show the odds of carlessness being 2.5 times higher for
unbanked households. These results are available upon request.
9. The Panel Study of Income Dynamics (PSID) shows that from
1999 to 2013 fuel was about 30 percent of total vehicle costs
for American families overall, but about 40 percent for poor
families. This disproportionate spending did not arise because
the poor spent more on fuel absolutely (they averaged $2,000
per year compared with $3,000 for the nonpoor) but because
they spent so much less than the nonpoor on other driving
costs ($3,000 compared with $7,000).
10. This category includes families who, between panel waves,
lost all earned income through unemployment.
11. The income difference between households with and without
vehicles actually peaked in 1989, when households with vehi-
cles had four times the income of households without them.
12. The median incomes are as follows: Manhattan $66,700;
Brooklyn $44,850; Queens $54,373; Bronx $38,900; and
Staten Island $70,295. Manhattan had a higher mean income
than Staten Island but a lower median.
13. Myers, in 1970, proposed an elaborate dial-a-ride system
for the poor that looks strikingly like modern Transportation
Network Companies (TNCs; Myers 1970). One obstacle to
using TNCs as a social service, however, is customers’ need
to have credit cards or (in some cases) debit cards, which
King et al. 17
makes these services less accessible to the unbanked (King
and Saldarriaga 2017). Nevertheless, Brown (2018) shows that
even poor neighborhoods in Los Angeles use ridehail services
frequently. Many cities are currently experimenting with pilot
projects to subsidize ridehailing trips, though as of this writing
the results of these projects have not been publicly released.
David A. King
Aaronson, Daniel, Sumit Agarwal, and Eric French. 2012. “The
Spending and Debt Response to Minimum Wage Hikes.”
American Economic Review 102 (7): 3111–39. doi:10.1257/
American Automobile Association. 1990–2017. “Your Driving
Costs.” Heathrow, FL: American Automobile Association.
Barger, Howard J. 2003. “No Deals on Wheels: Why the Poor Pay
More for Transportation.” Journal of Poverty 7 (4): 93–102.
Barrington-Leigh, Christopher, and Adam Millard-Ball. 2015. “A
Century of Sprawl in the United States.” Proceedings of the
National Academy of Sciences of the United States of America
112:8244–49. doi:10.1073/pnas.1504033112.
Blumenberg, Evelyn, and Asha W. Agrawal. 2014. “Getting around
When You’re Just Getting By: Transportation Survival Strategies
of the Poor.” Journal of Poverty 18 (4): 355–78.
Brown, Anne. 2017. “Car-Less or Car-Free? Socioeconomic and
Mobility Differences among Zero-Car Households.” Transport
Policy 60:152–59.
Brown, Anne. 2018. “Ridehail Revolution.” PhD diss., Department
of Urban Planning, University of California, Los Angeles.
Cervero, Robert, Onésimo Sandoval, and John Landis. 2002.
“Transportation as a Stimulus of Welfare-to-Work: Private
versus Public Mobility.” Journal of Planning Education and
Research 22:50–63. doi:10.1177/0739456X0202200105.
Charles, Kerwin K., Erik Hurst, and Melvin Stephens. 2008.
“Rates for Vehicle Loans: Race and Loan Source.” American
Economic Review 98 (2): 315–20.
Chester, Mikhail, Andrew Fraser, Juan Matute, Carolyn Flower,
and Ram Pendyala. 2015. “Parking Infrastructure: A Constraint
on or Opportunity for Urban Redevelopment? A Study of Los
Angeles County Parking Supply and Growth.” Journal of the
American Planning Association 81:268–86. doi:10.1080/0194
Cohen, Mark. 2007. “Imperfect Competition in Auto Lending.”
Vanderbilt Law School Law and Economics Working Paper 07-01.
Crozet, Yves. 2005. Time and Passenger Transport. Lyon, France:
Laboratory of Transportation Economics, University Lumiere
Lyon, Lyon.
DiMento, Joseph, and Cliff Ellis. 2011. Changing Lanes: Visions
and Histories of Urban Freeways. Cambridge: The MIT Press.
Eidlin, Eric. 2005. “The Worst of All Worlds: Los Angeles and the
Emerging Reality of Dense Sprawl.” Transportation Research
Record 1902:1–9.
Federal Highway Administration. 1918–2015. Highway Statistics.
Washington, DC: Department of Transportation.
Fischel, William A. 2004. “An Economic History of Zoning and a
Cure for Its Exclusionary Effects.” Urban Studies 41:317–40.
Forman, Richard T. T., Daniel Sperling, John A. Bissonette,
Anthony P. Clevenger, Carol D. Cutshall, Virginia H. Dale,
Lenore Fahrig, Robert L. France, Charles R. Goldman, Kevin
Heanue, Julia Jones, Frederick Swanson, Thomas Turrentine,
and Thomas C. Winter. 2003. Road Ecology: Science and
Solutions. Washington, DC: Island Press.
Glaeser, Edward L., and Janet E. Kohlhase. 2004. “Cities, Regions
and the Decline of Transport Costs.” Papers in Regional
Science 83:197–228. doi:10.1007/s10110-003-0183-x.
Gordon, Robert. 2016. The Rise and Fall of American Growth.
Princeton: Princeton University Press.
Gurley, Tami, and Donald Bruce. 2005. “The Effects of Car
Access on Employment Outcomes for Welfare Recipients.”
Journal of Urban Economics 58:250–72. doi:10.1016/j.
Hirt, Sonia A. 2015. Zoned in the USA: The Origins and
Implications of American Land-Use Regulation. Ithaca, NY:
Cornell University Press.
Jakle, John A., and Keith A. Sculle. 2008. Motoring: The Highway
Experience in America. Athens: University of Georgia Press.
Kawabata, Mizuki. 2003. “Job Access Employment among
Low-Skilled Autoless Workers in US Metropolitan Areas.”
Environment Planning A 35 (9): 1651–68.
King, David, and Juan Saldarriaga. 2017. “Access to Taxicabs for
Unbanked Households.” Journal of Public Transportation 20
(1): 1–19.
Klein, Nicholas J., and Michael J. Smart. 2017. “Car Today, Gone
Tomorrow: The Ephemeral Car in Low-Income, Immigrant and
Minority Families.” Transportation 44:495–510. doi:10.1007/
Lichtenwalter, Sara, Gary Koeske, and Esther Sales. 2006.
“Examining Transportation and Employment Outcomes:
Evidence for Moving beyond the Bus Pass.” Journal of Poverty
10:93–115. doi:10.1300/J134v10n01_05.
Loukaitou-Sideris, Anastasia, and Renia Ehrenfuecht. 2010. “Vibrant
Sidewalks in the United States.” Access 36 (Spring): 22–9.
Manville, Michael. 2017. “Automatic Street Widening.” Journal of
Transport and Land Use 10 (10): 375–93.
Manville, Michael, Alex Beata, and Donald Shoup. 2013. “Turning
Housing into Driving: Parking Requirements and Density in
Los Angeles and New York.” Housing Policy Debate 23:350–
75. doi:10.1080/10511482.2013.767851.
Manville, Michael, and Donald Shoup. 2005. “People, Parking and
Cities.” Journal of Urban Planning and Development 131 (4):
Marchetti, Cesare. 1994. Anthropological Invariants in Travel
Behavior. Laxenburg: International Institute for Applied
Systems Analysis.
Mattioli, Giulio, Karen Lucas, and Greg Marsden. 2017. “Transport
Poverty and Fuel Poverty in the UK: From Analogy to
Comparison.” Transport Policy 59:93–105.
Moura, Maria Cecilia P., Steven J. Smith, and David B. Belzer.
2015. “120 Years of U.S. Residential Housing Stock and
Floor Space.” PLoS ONE 10:e0134135. doi:10.1371/journal.
Myers, Sumner. 1970. “Personal Transportation for the Poor.”
Traffic Quarterly 24:191–206.
National Research Council. 2005. Assessing and Managing the
Ecological Impacts of Paved Roads. Washington, DC: National
Academies Press.
18 Journal of Planning Education and Research 00(0)
Nir, Sarah. 2016. “With Gas Station Closing, a Fuel Desert Expands
in Manhattan.” The New York Times, August 15. https://www
Norton, Peter D. 2011. Fighting Traffic: The Dawn of the Motor
Age in the American City. Cambridge: The MIT Press.
Pendall, Rolf, Christopher Hayes, Arthur (Taz) George, Casey Dawkins,
Jae Sik Jeon, Elijah Knaap, Evelyn Blumenberg, Gregory Pierce,
and Michael Smart. 2015. “Driving to Opportunities: Voucher
Users, Cars, and Movement to Sustainable Neighborhoods.”
Cityscape: A Journal of Policy Development and Research 17:57–87.
Raff, Daniel, and Manual Trajtenberg. 1996. “Quality-Adjusted Prices
for the American Automobile Industry.” In The Economics of
New Goods, edited by Timothy Bresnahan and Robert Gordon,
71–108. Cambridge: National Bureau of Economic Research.
Raphael, Steven, Michael A. Stoll, Kenneth A. Small, and Clifford
Winston. 2001. “Can Boosting Minority Car-Ownership Rates
Narrow Inter-Racial Employment Gaps? [with Comments].”
Brookings-Wharton Papers on Urban Affairs, 99–145.
Rice, Lorien. 2004. Transportation Spending by Low-Income
Households: Lessons for the San Francisco Bay Area. San
Francisco: Public Policy Institute of California.
Ruggles, Steven, Katie Genadek, Ronald Goeken, Josiah Grover,
Matthew Sobek. 2017. “Integrated Public Use Microdata
Series: Version 7.0 [Dataset].” University of Minnesota,
Minneapolis. doi:10.18128/D010.V7.0.
Santos, Adella, Nancy McGuckin, Hikari Y. Nakamoto, Danielle
Gray, and Susan Liss. 2019. Summary of Travel Trends: 2009
National Household Travel Survey. Oak Ridge National Laboratory.
Washington, DC: US Department of Transportation, Federal
Highway Administration.
Shoup, Donald. 2011. The High Cost of Free Parking. Updated ed.
Chicago: APA Planners Press.
Smart, Michael J., and Nicholas J. Klein. 2018a. “Complicating
the Story of Location Affordability.” Housing Policy Debate
28:393–410. doi:10.1080/10511482.2017.1371784.
Smart, Michael J., and Nicholas J. Klein. 2018b. “Disentangling
the Role of Cars and Transit in Employment and Earnings.”
Transportation. doi:10.1007/s11116-018-9959-3.
Souleles, Nicholas S. 1999. “The Response of Household
Consumption to Income Tax Refunds.” American Economic
Review 89 (4): 947–58. doi:10.1257/aer.89.4.947.
State of New Jersey. n.d. “Lifeline Utility Programs.” https://www
Thakuriah, Piyushimita, and Yihua Liao. 2005. “Analysis of
Variations in Vehicle Ownership Expenditures.” Transportation
Research Record: Journal of the Transportation Research
Board 1926:1–9. doi:10.3141/1926-01.
U.S. Bureau of Economic Analysis. 1967–2014. Table 7.2.5s: Auto
and Truck Sales. Washington, DC: Government Printing Office.
U.S. Bureau of Labor Statistics. 2016a. 100 Years of U.S. Consumer
Spending Data for the Nation, New York City, and Boston.
U.S. Bureau of Labor Statistics. 2016b. “Table 1110. Deciles of Income
before Taxes: Annual Expenditure Means, Shares, Standard Errors,
and Coefficients of Variation, Consumer Expenditure Survey,
U.S. Bureau of Labor Statistics. 2006. “Consumer Expenditure
U.S. Bureau of the Census. 1940–2010. American Community Survey,
Summary Files. Washington, DC: Government Printing Office.
U.S. Bureau of the Census. 1950–2010. U.S. Decennial Census
Summary Files. Washington, DC: Government Printing Office.
U.S. Energy Information Agency. 2000–2015. “Motor Gasoline Retail
Prices.” Washington, DC.
U.S. Environmental Protection Agency. n.d. “Fast Facts on
Transportation Greenhouse Gas Emissions.” https://www.epa
Voulgaris, Carole T., Brian D. Taylor, Evelyn Blumenberg, Anne
Brown, and Kelcie Ralph. 2016. “Synergistic Neighborhood
Relationships with Travel Behavior: An Analysis of Travel in
30,000 US Neighborhoods.” Journal of Transport and Land
Use 10:437–61. doi:10.5198/jtlu.2016.840.
Wachs, Martin. 1984. “Autos, Transit and the Sprawl of Los
Angeles.” Journal of the American Planning Association 50
(3): 279–310.
Webber, Melvin. 1992. “The Joys of Automobility.” In The Car
and the City, edited by Martin Wachs and Margaret Crawford,
274–84. Berkeley: University of California Press.
Wright, Gavin. 2010. “The New Deal and the Modernization of the
South.” Federal History 2:58–73.
Zahavi, Yacov, and Anti Talvitie. 1980. “Regularities in Travel
Time and Money Expenditures.” Transportation Research
Record 750:13–19.
Author Biographies
David A. King is an assistant professor in the School of
Geographical Science and Urban Planning at Arizona State
University. He studies transportation and land use policies.
Michael J. Smart is an assistant professor in the Edward J.
Bloustein School of Planning and Public Policy at Rutgers, The
State University of New Jersey. He studies transportation, demo-
graphic change, and social justice.
Michael Manville is an associate professor in the Luskin School of
Public Affairs at University of California, Los Angeles. He studies
public finance and transportation policy.
... In this article, we examine low-income households' transitions into and out of car ownership. We focus on car ownership because of the crucial role that access to a car plays in enabling participation in society and the hardships that those without cars face (King, Smart, and Manville 2019). In almost all neighborhoods in the United States, having access to a car enables greater accessibility to jobs, amenities, and social interactions compared to the accessibility of public transportation, walking, and cycling (Grengs 2010;Kawabata and Shen 2007;Liao et al. 2020). ...
... Car ownership is so common across the United States, even among low-income households, because social and economic mobility is tied to having a car (King, Smart, and Manville 2019). In almost all neighborhoods in the United States, access to a car enables drivers and passengers to access many more jobs and other opportunities within the same amount of time compared with traveling by transit, by bicycle, or on foot (Grengs 2010;Kawabata and Shen 2007;Liao et al. 2020). ...
... Losing access to a car typically results in decreased activity participation and lower quality of life, while regaining a car plays the opposite role. This should come as no surprise; much of the country's built environment, cultural norms, and labor market conditions are predicated on near-universal automobility (King, Smart, and Manville 2019). While recognizing that our data are pre-COVID, the pandemic has likely increased mobility gaps between the haves and have-nots (Sevtsuk et al. 2022). ...
Full-text available
We examine transitions into and out of car ownership among low-income households. We use a novel online survey of U.S. residents to investigate why households lose access to a car, how long they are without a car, why they regain a car, and how these transitions affect their quality of life. We find that car ownership transitions are primarily motivated by economic security and insecurity. The median length of a car-less episode is 1.7 years, while black and Hispanic respondents experienced longer car-less episodes. Despite their precarious grasp on car ownership, respondents felt that owning a car was worth it.
... Importantly, the economic divide between those with and without cars has grown over time. Between 1969 and2013, the median household income of carless households decreased by 35% in real terms ($26,492 vs. $17,237), while households with at least one car earned 13% more ($54,992 vs. $62,187) over the same period (King et al. 2019). ...
... Research shows that about half of families in poverty transitioned into and out of carlessness across a one decade time horizon, while less than one-third of other families did so (Klein and Smart 2017). The high levels of volatility in car transitions among low-income groups may be partially explained by that they often tend to have volatile incomes, purchase less reliable cars that increases their average car maintenance and repair costs, and pay high costs of loan interest rates and insurance on cars due to discriminations in these programs (Fletcher et al. 2010;King et al. 2019;Klein and Smart 2017;Ong and Stoll 2007). As incomes increase, low-income households are more likely to purchase a car than their middle-income counterparts (Blumenberg and Pierce 2012;Klein and Smart 2017), indicating that having a household car may impact mobility and quality of life of low-income households and residents to a larger degree compared to that of their peers (Klein 2020). ...
Full-text available
Car access is practically a necessity to get around in rural areas of the United States. Yet approximately 4% of rural residents or 4.3 million people do not have a car. Despite important rural–nonrural differences in the built environment and how people travel, research on rural mobility disparities by car access is limited. This paper addresses this gap by asking and answering the following questions: (1) What are the scope and scale of rural car access in the U.S. and what are the factors associated with not having a car? (2) How do rural carless residents get around relative to their nonrural peers? (3) What are the consequences of not having car access in rural areas? We used a dichotomous rural–nonrural classification to evaluate differences in car access and mobility outcomes between rural and nonrural areas using the U.S. Census Bureau Public Use Microdata Sample and the National Household Travel Survey. The results reveal new evidence on socioeconomic and mobility disparities by car access within and between rural and nonrural areas. Rural carless households earn 64% less and 40% less than their counterparts with full car access and their nonrural counterparts, respectively. Rural Native Americans, Black, and Asian populations have a 2.1 times, 1.3 times, and 1.3 times the odds of being carless, respectively, than their rural white peers, after controlling for other factors. Despite their lack of access to a car, a significantly smaller proportion of rural carless residents bike, walk, or take transit for work trips than their nonrural peers (38.9% vs. 74.6%), which may indicate infeasibility of sustainable modes in some rural areas. Consequently, rural carless residents are 2 times more likely to forsake trips than their nonrural peers due to a lack of transportation options. Our findings highlight rural mobility inequities that carless people face relative to their nonrural peers, which merit attention and tailored policy interventions to improve rural mobility, sustainability, accessibility, and equity.
... Experts classify AVs into five levels, from mere driver assistance to full autonomy (See Figure 1). (See Figure 4) Nov. 13, 2017;David King, Michael Smart, and Michael Manville, "The Poverty of the Carless" (UCLA ITS Working Paper, UCLA ITS, Los Angeles, CA, 2018), 4-5; and Kevin DeGood and Andrew Schwartz, "Can New Transportation Technologies Improve Equity and Access to Opportunity?," Center for American Progress, Apr. 27, 2016, 11. ...
Full-text available
"Self-driving cars won’t work until we change our roads," wrote Andrew Ng, chief of research at Chinese tech giant Baidu. Though Ng may exaggerate, autonomous vehicles (AVs) do benefit greatly from "smart infrastructure." Witnessing AVs’ rise, American governments have eagerly begun smart infrastructure projects, but few have developed stable methods of funding them. In this paper, I explore the funding mechanisms that exist today—the national gas tax, pooled and local financing, and AV mileage taxes—and evaluate their track record, future feasibility, and ability to make the largest beneficiaries pay a proportionate share. While the federal government has dedicated decreasing amounts of gas tax revenue to smart infrastructure, it also distributed gas tax dollars in the Smart City Challenge, an innovative contest that successfully if unsustainably drew in private contributions. Meanwhile, the Connected Vehicle Pooled Fund Study and Atlanta’s North Avenue Smart Corridor show how states and localities have self-funded intelligent transportation systems through general revenues. However, none of these methods impose costs on smart infrastructure’s largest users to the same degree as AV mileage fees, seriously considered in Tennessee and other states. No funding stream has simultaneously levied costs proportionately and dedicated its revenue back to smart infrastructure. The Smart City Challenge and especially AV taxes have come closer, but both involve political tradeoffs. In coming decades, autonomous vehicles and their supporting infrastructure offer a chance to rethink the fairness of transportation finance and the role of the public and private sectors in the city of tomorrow.
... The sprawling land use characteristics of the suburbs and limited alternate transportation options make travel by private vehicle a practical necessity for many residents. Researchers have found that car access allows low-income households to move to lower poverty neighborhoods, enables households to access employment, healthcare, shopping, childcare, and that lack of car access comes with a significant socioeconomic penalty (Jeon et al., 2018;King et al., 2022;Klein, 2020). Yet low-income and racial/ethnic minorities are more likely to be carless or exhibit inconsistent car ownership. ...
The budget allocation for poverty alleviation is getting bigger from time to time. However, the poverty rate did not change significantly. This study aims to determine the effectiveness of poverty programs that have been initiated by the government, through programs (KUBE, PANSIMAS, BUNDES, PNPM, and P2KP), which are believed to improve poverty alleviation efforts in Manyampa Village, Ujung Loe District, Bulukumba Regency. The research method used in this research is quantitative research, with the AHP (Analytic Hierarchy Process) analysis method. Data collection techniques were obtained from questionnaires or primary data sources, while secondary data were obtained from reports from BPS South Sulawesi, BPS Bulukumba, District Offices and related agencies in Bulukumba Regency. The novelty of this study is to explore further and evaluate what programs are quite effective in reducing poverty rates, with the help of expert choice with a consistency test. The results of this study indicate that in poverty alleviation efforts, the P2KP program is the most effective in reducing poverty with small business empowerment-based programs, while from community empowerment-based programs, the PANSIMAS program is more consistent than other programs in poverty alleviation efforts. P2KP and the PANSIMAS program each got a score of 0.02<0.1. It can be interpreted that the Inconsistency ratio <0.1 (the maximum limit) is also tested to be consistent, it's just that the achievement is lower than other programs. The implication of this research is that the local government needs to further evaluate ineffective poverty alleviation programs so that budget misallocations do not occur and are able to accelerate effective poverty alleviation programs that have been running. Keywords: Poverty, Program, Efectiveness
Full-text available
In the context of increasing urgency and interest in the need to reduce private car dependence, it is surprising that we know very little about people who chose to live without cars. This is particularly so for those living through life-stages, and in structural and cultural contexts, generally associated with private car use. Parenting children is one such life-stage. The common understanding is that the vulnerabilities and complexities associated with parenting are best attenuated by the autonomy, security and seclusion of the private car. Others, by choice or circumstance, parent by ‘altermobility’ – without the private car. Using data from in-depth interviews with car-free parents of young children, this paper records how parents come to live without private cars in a city that is dominated by structures, cultures and expectations of private car use. It proposes travel trajectories based on past and present events, experiences and inclinations. In doing so, the paper exposes the sheer complexity of influences of the past on present and future mobility practices, calling into question linear understandings of travel socialisation and mobility biography research. Auto-dependent childhoods, for example, do not necessarily produce car-dependent adults. And the onset of parenthood is not necessarily a time of increased private car attachment. In conclusion, the findings query assumptions about the impact of the early years of parenting on private car dependence, proposing that it is the child, not the infant, that cements the seemingly intractable bond between parenting and private car use.
Transportation affordability, the ability of a household to comfortably bear necessary transportation expenses, is a pressing challenge to the development of sustainable and equitable places. Transportation planners have historically identified transit user groups as choice riders, those with access to other modes or the ability to purchase access, and captive riders, transit-dependent populations who must use transit regardless of service quality. However, this characterization is incomplete, disregarding built-environment pressures that compel a much larger population of households to own and use vehicles regardless of their ability to afford the very high price of vehicle ownership. We propose a new theoretical construct, illuminated using the Bureau of Labor Statistics Consumer Expenditure Survey Public Use Microdata—the transportation choice/captivity matrix—which examines transportation user groups via vehicle ownership and income variables.
Full-text available
We examine the relationship between transportation access on the one hand and individuals’ employment and labor earnings on the other. We improve on existing studies by bringing a large national panel data set to bear on this question, attempting to disentangle the mechanisms by which individuals improve their economic standing and, finally, comparing the economic benefits to the direct costs of car ownership. To do this, we use nine waves from the Panel Study of Income Dynamics from 1999 to 2015. We find that access to a car is a strong predictor of future economic benefit for individuals, and that at very high levels of transit access, carless individuals can also fare equally well. Access to an automobile is strongly associated with employment, job retention, and earning more money over time. Though having a car is associated with economic benefits, owning and operating a car is expensive; yet, our findings suggest that the benefits may outweigh the costs for most people living outside neighborhoods with truly excellent transit service.
Full-text available
There is significant concern about the high cost of vehicle ownership, although several authors have acknowledged in the literature that owning and operating a vehicle have accessibility benefits. This paper uses Consumer Expenditure Survey data to explore vehicle ownership expenditures of American households. Differences in vehicle ownership expenditures among urban and rural households are estimated in different regions of the United States, controlling for socioeconomic, demographic, life cycle, and other household factors. First, patterns in vehicle ownership expenses are examined over a year. Three models of expenditures for vehicle ownership are estimated: the first model is estimated by ordinary least squares, the second is a Tobit model estimated to take into account zero expenditures on vehicles by a segment of households during the survey year, and the third model is a two-stage sample selection model that explicitly models first the household vehicle ownership decision and then, given that first decision, the expenditure intensity decision for the subsample of households that own vehicles. It was found that locational and regional effects were not constant but that urban–rural differences depended on region, and differences among regions depended on whether a household was located in an urban or a rural area.
Full-text available
In recent years, researchers and advocates have turned their attention to the trade-offs between housing affordability and transportation expenses. They argue that were families to move to more compact, transit-accessible, and walkable neighborhoods, they would reduce their driving and, possibly, forego the need for one or more cars, thus saving them money. We use the Panel Study of Income Dynamics to test this assumption with descriptive statistics and panel regression models, and we find little evidence to support it. We conclude that the location affordability literature may significantly overstate the promise of cost savings in transit-rich neighborhoods.
Full-text available
The notion of ‘fuel poverty’, referring to affordable warmth, underpins established research and policy agendas in the UK and has been extremely influential worldwide. In this context, British researchers, official policymaking bodies and NGOs have put forward the notion of ‘transport poverty’, building on an implicit analogy between (recognised) fuel poverty and (neglected) transport affordability issues. However, the conceptual similarities and differences between 'fuel' and 'transport' poverty remain largely unaddressed in the UK. This paper systematically compares and contrasts the two concepts, examining critically the assumption of a simple equivalence between them. We illustrate similarities and differences under four headings: (i) negative consequences of lack of warmth and lack of access; (ii) drivers of fuel and transport poverty; (iii) definition and measurement; (iv) policy interventions. Our review suggests that there are important conceptual and practical differences between transport and domestic energy consumption, with crucial consequences for how affordability problems amongst households are to be conceptualised and addressed. In a context where transport and energy exhibit two parallel policy worlds, the analysis in the paper and these conclusions reinforce how and why these differences matter. As we embark on an ever closer union between our domestic energy and transport energy systems the importance of these contradictions will become increasingly evident and problematic. This work contributes to the long-term debate about how best to manage these issues in a radical energy transition that properly pays attention to issues of equity and affordability.
Full-text available
A now substantial body of literature finds that land use and urban form have a statistically significant, albeit relatively modest, effect on travel behavior. Some scholars have suggested that various built-environment characteristics influence travel more in concert than when considered in isolation. Yet few previous studies have combined built-environment measures to create holistic descriptions of the overall character of neighborhoods, and fewer still have related these neighborhoods to residents’ travel decisions. To address this gap in the literature, we develop a typology of seven distinct neighborhood types by applying factor analysis and then cluster analysis to a set of 20 variables describing built-environment characteristics for most census tracts in the United States. We then include these neighborhood types in a set of multivariate regression models to estimate the effect of neighborhood type on the travel behavior of neighborhood residents, controlling for an array of personal and household characteristics. We find relatively little variation in the number of daily trips among neighborhood types, but there is substantial neighborhood variation in both person miles of travel and mode choice. Travel by residents of one particular neighborhood type is notably distinguished from all others by a very low number of miles traveled, little solo driving, and high transit use. However, this neighborhood type is found almost exclusively in just a few very large metropolitan areas, and its replicability is uncertain. © 2016 Carole Turley Voulgaris, Brian D. Taylor, Evelyn Blumenberg, Anne Brown, and Kelcie Ralph.
Full-text available
Los Angeles, California, is generally considered the archetypal sprawling metropolis. Yet traditional measures equate sprawl with low population density, and Los Angeles is among the densest and thereby the least sprawling cities in the United States. How can this apparent paradox be explained? This paper argues that the answer lies in the fact that Los Angeles exhibits a comparatively even distribution of population throughout its urbanized area. As a result, the city suffers from many consequences of high population density, including extreme traffic congestion, poor air quality, and high housing prices, while offering its residents few benefits that typically accompany this density, including fast and effective public transit, vibrant street life, and tightly knit urban neighborhoods. The city's unique combination of high average population density with little differentiation in the distribution of population might best be characterized as dense sprawl, a condition that embodies the worst of urban and suburban worlds. This paper uses Gini coefficients to illustrate variation in population density and then considers a number of indicators–-most relating either to the provision of transportation infrastructure or to travel behavior–-that demonstrate the effects of low-variation population distribution on the quality of urban life in Los Angeles. This approach offers researchers, practitioners, and policy makers in Los Angeles and in smaller cities that are evolving in similar ways a useful and user-friendly tool for identifying, explaining, measuring, and addressing the most problematic aspects of sprawl.
Transportation professionals have long identified the important division between choice and constraint in modal decision-making. However, while heterogeneity within some modal groups such as transit riders is well documented, intragroup differences in other groups have been largely ignored. In particular, significant heterogeneity exists among zero-car households, who may not own a car due to choice (car-free) or constraint (car-less). Recognition of intragroup heterogeneity among zero-car households yields policy implications as cities consider where to invest their ever more precious and more limited transportation resources. Using activity diary data from the 2012 California Household Travel Survey, I investigate two research questions: first, how do car-less households compare to car-free households in terms of both their socioeconomic characteristics and their relative share of zero-car households? Second, how do motivations behind not owning a car translate into mobility differences - including daily trip counts and miles traveled? I find that, contrary to media reports that the number of car-free households—that is, households that choose not to own a car—is “booming” 79 percent of zero-car households do not own a car because of economic or physical constraints. Car-less households in many ways mirror the captive transit population, in that they have significantly lower household incomes, lower educational attainments, and are disproportionately non-white compared to car-free households. Observed socioeconomic variations translate into mobility differences, with car-free travelers taking more trips and traveling more miles per day. Policy makers should consider extending access to carshare services, which are positively associated with more trips and miles traveled among both car-free and car-less households.