Automobile Externalities and Policies

Automobile Externalities and Policies
Author(s): Ian W. H. Parry, Margaret Walls and Winston Harrington

Source: Journal of Economic Literature , Jun., 2007, Vol. 45, No. 2 (Jun., 2007), pp. 373-
399
Published by: American Economic Association
Stable URL: https://www.jstor.org/stable/27646797
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms
is collaborating with JSTOR to digitize, preserve and extend access to Journal of Economic
Literature
This content downloaded from

182.255.0.242 on Wed, 10 Nov 2021 04:05:08 UTC
All use subject to https://about.jstor.org/terms
Journal of Economie Literature
Vol. XLV (June 2007), pp. 373-399
Automobile Externalities and Policies
Ian W. H. Parry, Margaret Walls, and Winston Harrington
This paper discusses the nature, and magnitude, of externalities associated with
automobile use, including local and gfobal pollution, oil dependence, traffic conges
tion and traffic accidents. It then discusses current federal policies affecting these
externalities, including fuel taxes, fuel economy and emissions standards, and alter
native fuel policies, summarizing, insofar as possible, the welfare effects of those poli
cies. Finally, we discuss emerging pricing policies, including congestion tolls, and
insurance reform, and summarize the appropriate combination of policies to address
automobile externalities.
1. Introduction maintenance and new construction (U.S.

Department of Transportation, Bureau of
Transportation Statistics 2005, table 3-29a).
Of allmore
consumer products,
heavily or regulated more few
exten are taxed And automobile use also has many undesir
sively than automobiles. In the United States, able side effects. About 40,000 people die on
ownership taxes over the life of a vehicle highways each year (U.S. Department of
average about 18 percent of the sales price, Transportation, National Highway Traffic
and combined state and federal gasoline Safety Administration 2002, table 3); urban
taxes average 40 cents per gallon, or about 20 road congestion causes 3.7 billion hours of
percent of pretax fuel prices (Winston delay a year (David Schrank and Timothy
Harrington and Virginia D. McConnell Lomax 2005); automobiles are a leading
2003). New vehicles also are subject to regu source of greenhouse gases and local air pol
lations governing local air pollution, safety, lutants; and gasoline accounts for nearly half
and fuel economy, and in many states, of the nation s dependence on oil.
mandatory purchase of liability insurance. For these reasons, it is not surprising that
At the same time, few consumer products automobiles and the fuels they use have
require such a gigantic public infrastructure attracted attention from government. What
in order to be useful, one that costs more we may wonder is whether the current col
than $100 billion per year in road and bridge lection of policies makes sense or whether it
could be improved upon, or even over
hauled completely. Several recent trends
* Parry: Resources for the Future. Walls: Resources for make it a particularly good time for
the Future. Harrington: Resources for the Future. reassessing federal automobile policies.
373

182.255.0.242 on Wed, 10 Nov 2021 04:05:08 UTC
374 Journal of Economic Literature, Vol. XLV (June 2007)
First, there are heightened concerns about appears to be the appropriate combination
energy security with the recent tripling of of policies.1
world oil prices and alarming instability in
the Middle East. Second, there is growing 2. Automobile Externalities
pressure on the federal government to curb
A number of other studies provide a gen
greenhouse gases, with solidifying scientific eral discussion of automobile externalities
evidence on global warming, various
including Bruno de De Borger and Stef
regional abatement initiatives, and the birth
Proost (2001); Mark A. Delucchi (2000);
of carbon trading in Europe. Third,
David L. Greene, Donald W. Jones, and
because of rising urban land costs and
Delucchi (1997); Douglas B. Lee (1993);
intense siting opposition, road capacity
Todd Litman (2006); David M. Newbery
enhancements lag far behind relentlessly
(2005a); Ian W. H. Parry and Kenneth A.
expanding vehicle use, with increasing grid
lock the inevitable result. Fourth, due to Small (2005); John Peirson, Ian Skinner, and
Roger Vickerman (1995); Richard C. Porter
declining real fuel tax revenue per vehicle (1999); Emile Quinet (2004); Werner
mile, there is a growing transportation
Rothengatter (2000); and U.S. Department
funding gap, increasingly met at the region
of Transportation, Federal Highway Admin
al level by referenda tying various tax istration (1997). Here we distill some of the
increases to specific highway projects.
main findings.
Finally, with advances in electronic meter
ing technology, it is now, for the first time, 2.1 Local Air Pollution
feasible to charge motorists on a per mile Gasoline vehicles emit carbon monoxide
basis according to the marginal external
(CO), nitrogen oxides (NOx), and hydrocar
costs of their driving.
bons (HC) (sometimes called volatile organ
From an efficiency perspective, the first
ic compounds). CO reduces oxygen in the
issue in analyzing automobile policies is to
bloodstream causing breathing difficulty
identify the specific externalities that they and cardiovascular effects, while HC and
are designed to address. Some of these
NOx react in sunlight to form ozone (the
externalities, such as those from green
main component of smog), which affects
house gases, vary with fuel combustion,
pulmonary function in children and asth
while others, such as congestion, vary with
matics, and reduces visibility. More impor
the extent, location, and timing of travel. In tant, NOx and HC also react to form
section 2 below, we describe each of the
particulate matter. Fine particles (PM2.5)
major externalities, explain what margins of
are small enough to reach lung tissue and
behavior they operate along, and summa studies have documented a causal relation
rize evidence on their magnitude. The rest
between particulate exposure and mortality
of the paper focuses on the implications of
these externalities for the appropriate
design of federal policy, with emphasis on 1 A number of broader issues are beyond our scope,
including automobile policies outside the United States,
some practical difficulties in policy reform. policies for heavy-duty trucks, how to improve highway
Section 3 discusses the major federal poli spending mechanisms to ensure that individual projects
cies currently affecting these externalities are economically efficient, and the interface between
automobiles and urban development.
including fuel taxes, fuel economy stan Gasoline vehicles account for about 95 percent of
dards, new-vehicle emissions standards, and new passenger vehicle sales; this includes the 4 percent of
alternative fuel policies. Section 4 describes flexible-fuel vehicles that could operate on ethanol but
typically use gasoline (U.S. Department of Energy,
emerging pricing policies that target con Energy Information Administration 2006b). Diesel vehi
gestion and accident externalities more cles are far more common in Europe, where they often
receive favorable tax treatment.
directly. Section 5 briefly summarizes what

182.255.0.242 on Wed, 10 Nov 2021 04:05:08 UTC
Parry, Walls, and Harrington: Automobile Externalities and Policies 375
(Douglas W. Dockery et al. 1993; Joel declined; for example, almost half of HC
Schwartz 1994). emissions in 1970 were from mobile sources
The generation of these pollutants can be compared with only 28 percent in 2003 (U.S.
reduced by curbing vehicle miles traveled Environmental Protection Agency 2004).
(VMT), improving average vehicle fuel Estimates of local pollution damages
economy or through other technologies to attributed to automobiles are dominated by
lower exhaust emissions per gallon of fuel mortality effects, especially due to particu
combustion. A gasoline tax alone is inade lares. Studies translate vehicle emissions into
quate, as it does not encourage the last of changes in ambient concentrations of pri
these behavioral responses. Therefore, since mary and secondary pollutants based on cli
the 1970 Clean Air Act, new passenger vehi mate, topography, and wind patterns, etc.,
cles have been subject to grams-per-mile health effects for exposed population groups
standards for CO, NOx, and HC. Initially are then inferred from epidemiological evi
these standards were slightly more stringent dence; finally, monetary damage estimates
for cars than for light-trucks (sport utility are obtained from assumptions about peo
vehicles or SUVs, pickup trucks and mini ple s willingness to pay to avoid health risks,
vans), though standards have been harmo using evidence from revealed and stated
nized since the mid-1990s so it will no longer preference studies.
be the case that light-trucks produce more Small and Camilla Kazimi (1995) project
emissions per mile than cars. ed pollution costs of 2.3 cents per mile for
The gram-per-mile standards allow manu the Los Angeles region for the year 2000
facturers flexibility in meeting them through (updating to 2005$), with a range of 1-8
abatement or fuel-saving technologies, and cents per mile; NOx and HC emissions con
ensure that all new vehicles emit at the same tribute about equally to these costs, mainly
per-mile rate, regardless of fuel economy. In through particulate formation, while CO
the past, as vehicles aged, their emission effects are ignored as their outdoor concen
control systems tended to deteriorate, so trations are too low to have noticeable
that in older vehicles the natural relationship health effects. Although meteorological
between fuel economy and emissions rates conditions in Los Angeles are especially
per mile reasserted itself (Winston favorable for pollution formation, Small and
Harrington 1997). However, given the dura Kazimis estimates are broadly consistent
bility of state-of-the-art abatement technolo with estimates for other urban areas
gies currently being used, emissions over the (Donald R. McCubbin and Delucchi 1999;
life of a vehicle have become largely decou U.S. Department of Transportation,
pled from initial fuel economy (Carolyn Federal Highway Administration 2000).
Fischer, Harrington, and Parry forthcom Nonetheless, local automobile emissions
ing); this means that tailpipe emissions now will continue to decline substantially in
vary primarily with VMT rather than total upcoming years as far more stringent regu
fuel consumption. lations are phased in and the vehicle fleet
Despite high growth in VMT over the past turns over (see below); therefore, in terms
thirty years, nationwide vehicle emissions of of future policy reform, other externalities
all local pollutants have fallen dramatically. are of greater concern.
This drop can be attributed to progressively
more stringent new-vehicle emissions stan 3 An alternative approach that avoids the last two steps
dards over time, along with retirement of the is to measure people s willingness to pay for cleaner air by
comparing differences in property values across clean and
oldest, most polluting vehicles (figure 1). dirty regions controlling for other factors (e.g., V. Kerry
The share of vehicle emissions in total sta Smith and Ju-Chin Huang 1995; Kenneth Y. Chay and
tionary and mobile emissions also has Michael Greenstone 2005).

182.255.0.242 on Wed, 10 Nov 2021 04:05:08 UTC
376 Journal of Economie Literature, Vol. XLV (June 2007)
Figure 1. Total Vehicle Emissions Relative to Base Year
Source: From EPA's MOBILE model (U.S. Environmental Protection Agency 2004). This model simulates
fleetwide emissions by classifying existing on-road vehicles by type and vintage, and applying emissions
coefficients to those vehicles.
2.2 Global Air Pollution and Joseph Boyer (2000) put the (popula
tion-weighted) expected global costs of a
Light-duty vehicles account for a fifth of 2.5?C warming in 2100 at almost 2.0 percent
nationwide emissions of carbon dioxide, of world GDP. Half of this is from the risk of
which is the leading greenhouse gas (U.S. catastrophic or abrupt climate change,
Department of Energy, Energy Information which they estimated based on subjective
Administration 2006b). A fuel tax essentially expert judgment about the likelihood of
is equivalent to a tax on vehicle carbon emis major disruptions to GDP. Another signifi
sions as, unlike for local pollutants, there are cant damage component is from the possible
no viable technologies for reducing carbon spread of tropical disease, especially in
emissions per gallon of fuel combustion. Africa, which is inferred from data on the
Economists have attempted to estimate incidence of various diseases across different
future damages from global warming. One climatic regions, and disability adjusted life
well-known study by William D. Nordhaus years lost per disease. Agricultural effects,

182.255.0.242 on Wed, 10 Nov 2021 04:05:08 UTC
allowing for farm-level adaptation to climate when warming rises beyond 5?C under busi
change, account for less than 10 percent of ness as usual assumptions. The appropriate
damages, though some countries, such as discount rate for such long-range effects
India, are predicted to suffer more than oth remains contentious (Paul R. Portney and
ers. Sea level rise accounts for a further 6 John P. Weyant 1999). Conventional discount
percent of damages; costs here reflect the rates currently warrant fairly modest actions
value of inundated land and infrastructure to slow climate change for future generations,
necessary to protect valuable coastal regions. which some object to on ethical grounds;
And damages to immobile settlements however, much lower rates are inconsistent
(Venice, Bangladesh), or to ecosystems, as with observed behavior, lead to perverse
measured by willingness to pay to preserve results in other contexts (e.g., drastic reduc
these assets, account for another 6 percent tions in current consumption), and imply that
of costs. the most speculative, distant effects have a
Overall, Nordhaus and Boyer (2000), large influence on current policy.
updated in Nordhaus (forthcoming), put the A gallon of gasoline contains 0.0024 tons of
discounted cost of current carbon dioxide carbon (National Research Council 2002);
emissions, over their expected 100-year therefore damages of $20, $50 and $300 per
atmospheric life, at equivalent to $20 per ton ton of carbon translate into 5,12, and 72 cents
of carbon (in 2005$). However, there is a per gallon of gasoline, respectively.
ramp-up effect over time with growth in
2.3 Oil Dependency
potentially affected world output, improved
mitigation technology, and as marginal dam The United States consumes 21 million
ages from additional warming rise with the barrels of oil a day, of which almost 60 per
temperature level; the projected cost per ton cent is imported (up from 27 percent in
of carbon rises to $84 by 2050, and $270 by 1985); gasoline is the single most important
2100. Most other economic assessments are source of oil use, accounting for 45 percent of
in the same ballpark (e.g., David Pearce 2005; petroleum products (U.S. Department of
Robert Mendelsohn et al. 2000); a meta Energy, Energy Information Administration
analysis by Richard S. J. Toi (2005) suggests a 2006a). Although U.S. Department of Energy,
current upper bound cost of $50 per ton. Energy Information Administration (2006a)
A strikingly different conclusion was projects oil consumption to increase to 26 mil
reached in the recent Stern Review lion barrels per day by 2025 (with the import
(Nicholas Stern 2006), which puts total dam share staying roughly constant), they predict
ages from future warming at 5-20 percent of that oil use relative to GDP will fall by around
world GDP in perpetuity and recommends a 30 percent, due to continued improvements in
current social cost equivalent of $311 per ton energy efficiency and growth in the overall
of carbon. This much larger estimate partly is economy outstripping growth in transportation
explained by assumptions about more rapid fuel demand. Dependence on oil and foreign
warming, greater disruptions from extreme imports exposes the economy to energy price
weather events, and a more limited scope for volatility, price manipulation (though the
adaptation to climate change. But according
to Nordhaus (2006), most of the difference is 4 The social discount rate per capita consists of the pure
rate of time preference for discounting utility (assumed to
from their assumption that the social discount be about zero in Stern 2006) and the growth in per capita
rate on future consumption is around 1 per consumption times the elasticity of marginal utility of con
cent rather than 3?5 percent as assumed in sumption. Uncertainty over the future discount rate
increases the expected value of the discount-factor applied
most previous studies; the lower rate greatly to future damages; this can significantly increase expected
magnifies the present value of damages marginal damages from todays emissions (Richard G.
occurring one or two centuries from now, Newell and William A. Pizer 2003).

182.255.0.242 on Wed, 10 Nov 2021 04:05:08 UTC
United States itself has some market power), private sector internalizes 25-100 percent of
and may compromise national security and for the risk of price shocks.
eign policy interests; however, the extent to 2.3.2 Market Power
which the market fails in all these regards often
is murky. Although noncompetitive pricing and
investment behavior by OPEC countries
2.3.1 Vulnerability to Oil Price Volatility
may reduce global welfare, this fact itself
Projecting future oil prices is especially does not drive any price wedge between the
hazardous as they are sensitive to so many domestic U.S. demand curve and the oil
uncertainties: vehicle growth in China, import supply curve; if the United States
OPEC behavior, supply from conventional were a price taker in the world oil market,
and non-conventional sources (e.g., oil shale there would be no efficiency basis for an oil
and tar sands), policy change in Saudi Arabia, tax on market power grounds. However,
etc. One recent assessment put the likeli studies have shown that the United States
hood of a temporary $15-$50 per barrel currently has a limited degree of monopsony
price shock in the next ten years at about 50 power in the world oil market; in principle
percent (Phillip C. Beccue and Hillard G. this justifies an optimum tariff, as is familiar
Huntington 2005). Simulation models and from trade theory, but only if welfare is
regression analyses suggest that a price shock viewed from a domestic rather than global
of this magnitude would lower U.S. GDP by perspective (which seems inconsistent with
anything from 0.5 to 6.0 percent by raising measuring climate damages on a global
energy costs and deflating demand via a basis). Again, simulations by Leiby et al.
transfer of purchasing power to foreign sup (1997) suggest this tariff would amount to
pliers (Hillard Huntington 2005). The shock around $3-10 per barrel, or 7-24 cents per
would be more disruptive the greater the gallon of gasoline (updated to 2005$).5
extent of market frictions, if the economy is
2.3.3 Military and Geo-Political Costs
in a slump at the time of the shock, or if there
are constraints on monetary and fiscal policy. In principle, some portion of U.S. Middle
However, the extent to which these macro East military expenditures may constitute
economic disruption risks constitute a mar part of the total external cost of oil depend
ket failure is questionable. U.S. consumption ency Delucchi and James Murphy (2004)
of oil itself adds little to the risk of a world allocate part of the defense budget to
wide price shock, and many analysts believe Middle East operations, add to this an esti
that firms and households adequately mate of the annualized costs to the United
account for oil price volatility in their capital States of conflicts in the region (prior to the
investment decisions, use of futures markets, protracted war in Iraq), and then allocate a
inventory strategies, and so forth (Douglas R. portion of these costs to protecting oil sup
Bohi and Michael A. Toman 1996). Others plies for domestic consumption (as opposed
argue that disruption costs partly reflect mar to other objectives, such as regional stabili
ket imperfections (e.g., price and wage ridig ty). They put the military burden attributed
ities, underinvestment in fuel-efficient to automobiles at $0.8-$8.5 billion a year, or
technologies) and are therefore not fully 0.5-6 cents per gallon of gasoline. However
internalized. The most widely cited study, by An alternative approach has been to estimate optimal
Paul N. Leiby et al. (1997), puts the uninter oil tariffs in a dynamic setting, accounting for scarcity
nalized disruption risks from oil price shocks rents to exhaustible resources and the dependency of cur
rent production decisions on the expected path of future
at around $0-8.50 per barrel of U.S. oil con tariffs. Summarizing this literature, Newbery (2005b) puts
sumption, or 0-20 cents per gallon of gaso the optimal oil tariff for the United States at $3.8-$15.6
line (updated to 2005$), assuming that the per barrel.

182.255.0.242 on Wed, 10 Nov 2021 04:05:08 UTC
the marginal cost of military spending will against an increase in lane-mile capacity of
only equal this average cost if military spend only 51 percent (U.S. Department of
ing falls in proportion to (modest) reductions Transportation, Bureau of Transportation
in gasoline demand; some analysts view mil Statistics 2005, tables 1-6 and 1-33); annual
itary spending as more of a fixed than a vari urban congestion delays increased from 16 to
able cost and, therefore, assume the 47 hours per driver, while the national cost of
marginal cost of military spending is zero. wasted time from congestion increased from
Dependence on oil also may constrain U.S. $12.5 to $63 billion (Schrank and Lomax
foreign policy if the government believes that 2005). Clearly a fuel tax, which raises driving
oil-producing nations would disrupt the oil costs for all regions at all times of day, is a
market in response to U.S. policy pressures. very blunt instrument for alleviating traffic
However, the overriding concern in the congestion, which is highly specific to rush
United States has become the flow of petrole hour periods in urban areas; the ideal instru
um dollars to governments, such as Iran, or ment is a road-specific congestion toll that
other groups, such as terrorists or insurgents varies with time of day.
in Iraq, that threaten regional or U.S. nation
2.4.1 Theory of Congestion Externalities
al security. These broader costs are exception
ally difficult to quantify: although the United The standard way economists think about
States acting unilaterally has very little influ traffic congestion is to plot demand, average
ence on these revenue flows, oil-conserving cost, and marginal cost as a function of vehi
technologies developed at home might still cle flow, that is, the number of vehicles pass
have a substantial longer-term impact if they ing a point on a highway in an hour (Alan A.
were ultimately deployed in other large, Walters 1961; William S. Vickrey 1963;
industrializing countries such as China. Kenneth Button 2004; Robin Lindsey and
Summing up, as with global warming, there Erik T Verhoef 2001). The average cost is
is room for legitimate debate about the extent vehicle operating costs per mile (mainly
and magnitude of externalities or broader fuel), plus the product of time per mile?the
market failures from oil dependence. Our reciprocal of speed?and the value of travel
discussion suggests a corrective tax might be time, usually assumed to be around 50 per
anything from roughly 8-50 cents (excluding cent of the wage (Small 1992a). At low vehi
geo-political costs); a recent panel of experts cle flows the average cost curve is flat, but it
(National Research Council 2002) recom starts to rise as more vehicles on the road
mended a value of 12 cents per gallon. eventually force the motorist to slow down to
Whether the ideal corrective tax should be maintain a comfortable time-separation with
on consumption of gasoline (and other oil the vehicle just ahead. In this analysis, the
products) or just oil imports depends on marginal cost curve lies above the average
whether the objective is to reduce the oil cost curve as it also reflects the additional
intensity of GDP or dependence on foreign travel time costs to all motorists due to
oil, though an import tariff is likely precluded added congestion from an extra vehicle mile
by international trade agreements. per unit of time. Since motorists only care
about average costs to themselves, efficiency
2.4 Traffic Congestion
requires a Pigouvian tax equal to the gap
Between 1980 and 2003, urban VMT in between the marginal cost to all drivers and
the United States increased by 111 percent, average cost to the individual, at the point
where the marginal cost intersects the
6 For example, a 10 percent reduction in U.S. oil demand curve.
imports would reduce long-term world oil prices by per
haps 0.5-2 percent, which is small when set against the In the 1960s, however, engineers and
recent tripling of oil prices. economists realized that the flow congestion

182.255.0.242 on Wed, 10 Nov 2021 04:05:08 UTC
model was often not a very good descrip costs as they also include the costs to those
tion of rush-hour, which is characterized by who adjust their behavior to avoid conges
very high densities, stop and go traffic, and tion by rescheduling trips or using mass
transit (Arnott, de Palma, and Lindsay
"hypercongestion," where travel speeds are
so low that total traffic flow actually Lindsey 1994; Small and Xuehao Chu
declines?often to considerably less than 2003).
half of road capacity. The existence of
hypercongestion seemed to imply that the 2.4.2 Nationwide Marginal Congestion
average cost curve was backward bending. Costs
Beginning with Vickrey (1969), models
began to appear analyzing "bottleneck" A number of studies estimate congestion
congestion, which results when traffic flow costs for individual roads or cities, which is
temporarily exceeds capacity at a point, needed for local congestion policies, but
either because of a spike in demand or a few attempt to compute an average over a
sudden reduction in road capacity, such as nation, which is needed to evaluate the fed
an intersection with obstructing cross traf eral fuel tax (at least until congestion is fully
fic or an accident (Richard Arnott, Andr? internalized through peak period pricing).
de Palma, and Robin Lindsaey 1993, 1994; One exception, based on speed-flow curves,
Arnott and Marvin Kraus 1995). is U.S. Department of Transportation,
Models combining bottleneck and flow Federal Highway Administration (1997,
congestion appear to have resolved the 2000), who weighted marginal external
problem of the backward-bending cost costs for representative urban and rural
curve. For example, Se-il Mun (1994) roads at different times of day by the
developed a dynamic model of travel respective VMT shares; they put "averaged"
between two distant points with a queue in marginal external costs at 5 cents per pas
the middle that forms and eventually clears senger mile, equivalent to $1.05 per gallon
during peak period; the travel cost is deter at current on-road fuel economy of 21 miles
mined by the standard speed-flow relation per gallon.
on either side of the bottleneck, but also For assessing the congestion effects of
includes the waiting cost in the queue. The fuel taxes, however, this cost should be
average travel cost over the entire trip is adjusted to account for the much weaker
always increasing in the travel flow, and does sensitivity of peak-period driving (which is
not bend backwards, as in the traditional dominated by commuting) to fuel prices
model. compared with off-peak or rural driving.
Recognition of the important dynamics of That is, higher fuel taxes will have a dis
congestion does not change the definition of proportionately large effect on roads with
the congestion externality, but it does sug minimal congestion and a disproportion
gest that estimating it is far more difficult ately small impact on congested roads;
than previously anticipated. Not only is Parry and Small (2005) assumed marginal
roadway congestion highly variable over congestion costs of 3.5 cents per mile in
time and space, but now delay on a road is their fuel tax analysis. A later study, by
understood to be not usually the result of Fischer, Harrington, and Parry (forthcom
inadequate capacity on that road, but the ing), estimates that national average mar
result of a bottleneck elsewhere in the net ginal congestion costs (in response to lower
work. A further difference compared with fuel costs) are equivalent to 6.5 cents per
the traditional formulation is that bottle mile; this is based on a spatially disaggre
neck models showed that the costs of con gated model of the Washington, D.C., met
gestion are broader than just the pure delay ropolitan area transport network, with results

182.255.0.242 on Wed, 10 Nov 2021 04:05:08 UTC
extrapolated to all other U.S. metropolitan Highway Traffic Safety Administration

areas.7 (2002), tables 3 and A-l (see Ted R. Miller
1993 for an earlier estimate). Costs encom
2.5 Traffic Accidents
pass quality-adjusted life years lost from
Annual fatalities on American roads have injuries, property damages to automobiles,
hovered around 40,000 since 1960. travel delays, medical costs, lost productivity
However, there has been a dramatic decline in the workplace and at home, and insur
in fatality rates, from 5.1 per 100 million ance/legal expenses. Total costs are quite
VMT in 1960 to 1.5 per 100 million VMT in substantial at $433 billion (4.3 percent of
2003 (U.S. Department of Transportation, GDP), equivalent to an average of 15.8 cents
Bureau of Transportation Statistics 2005, per VMT.
table 2-17), reflecting greater seatbelt use,
2.5.2 Marginal Accident Externality
improved vehicle technology, reduced drunk
driving, and reduced pedestrian deaths as A common assumption is that injury risks in
fewer people are inclined to walk. Again, single-vehicle crashes to the driver and other
the theoretically ideal tax to factor accident vehicle occupants are internalized, though
risks into the costs of driving would not be a whether driving by one individual raises
fuel tax but a tax on VMT reflecting differ injury risks for other drivers is unclear. All
ences in marginal external costs across driv else being the same, the presence of one extra
ers, vehicles, and regions. Although there is vehicle on the road raises the likelihood that
some consensus on how the total (internal other vehicles will crash, but if people com
and external) costs of traffic accidents can be pensate by driving more slowly or more care
estimated, separating out the external costs fully in heavier traffic this will lower both the
and putting them on a marginal basis is number and average severity of collisions.
much more challenging. The severity-adjusted risk for other drivers
may even fall with more traffic, though this
2.5.1 Social Costs of Traffic Accidents
may not imply a positive externality, as the
Table 1 compiles an estimate of the total compensating behavior is itself costly.
social costs of traffic accidents (for all motor Unfortunately, empirical evidence on this
vehicles) for the year 2000, based on U.S. issue is limited; recent studies often assume
Department of Transportation, National no effects on injury risks to other drivers (e.g.,
Inge Mayeres, Sara Oschelen, and Stef Proost
1996; U.S. Department of Transportation,
Both of the nationwide estimates of congestion costs
account for bottlenecks in a crude way in the calibration
Federal Highway Administration 1997).
of the speed-flow relations, and the second estimate Aside from the interdriver injury issue,
accounts for the costs of behavior to avoid congestion; nei studies typically include pedestrian and
ther estimate accounts for nonrecurrent congestion delays
caused, for example, by weather and road works (accident
cyclist injuries in computing marginal exter
delays are discussed below). nal costs (these account for about 13 percent
Fatalities attributed to nonuse of seatbelts fell from of fatalities attributed to passenger vehicles).
13,301 in 1975 to 9,238 in 2000, airbags and child
restraints saved 2,488 and 446 lives respectively in 2003,
alcohol-involved fatalities fell from 23,167 in 1985 to 9 Using panel data on state-average insurance premi
17,013 in 2003, and pedestrian deaths per VMT fell 84 ums and claims, Aaron S. Edlin and Pinar Karaca-Mandic
percent between 1960 and 2003 (U.S. Department of (2006) find that an additional driver can substantially
Transportation, National Highway Traffic Safety increase insurance costs for other drivers in urban areas.
Administration 2002, table 21; U.S. Department of However, insurance costs are far from a comprehensive
Transportation, Bureau of Transportation Statistics 2005, measure as they mainly reflect property damages and
tables 2-25 and 2-30). Data on nonfatal injuries is avail these account for only 14 percent of social costs in table 1;
able from 1990 and also shows a declining trend (U.S. Edlin and Karaca-Mandic find mixed evidence on
Department of Transportation, Bureau of Transportation whether fatality rates increase or decrease with heavier
Statistics 2005, table 2-17). traffic.

182.255.0.242 on Wed, 10 Nov 2021 04:05:08 UTC
382 Journal of Economie Literature, Vol XLV (June 2007)
TABLE 1
Social Costs of Traffic Accidents in the United States, 2000
Property Minor Serious Fatalities All injuries

damage only injury injury
Number of injuries 23,631,696 7,208,043 607,882 41,821 31,489,442
Total cost per injury, $ 2,532 10,401 259,718 3,366,387 13,766
quality adjusted life years 0 2,880 135,275 2,389,179 6,444
property damage 1,484 2,845 4,982 10,273 1,875
travel delay 803 776 1,004 9,148 812
medical & emergency services 31 1,539 33,899 22,928 1,106
market and household productivity 98 70 291 795,601 3,058
insurance & legal 116 1,131 45,965 139,258 617
Total injury costs, $billion 60 75 158 141 433
Note: Minor injuries are those with quality-adjusted life years below $4,500.
Source: U.S. National Highway Traffic Safety Administration (2002).
less safe for their occupants (as less of the

Property damages in single- and multivehicle
energy
crashes mostly are treated as an external in a crash is absorbed by the vehicle
cost
and more
given that premiums primarily are levied on isa transferred to its occupants) but
safer
lump sum rather than variable (per for other road users, though again
mile)
basis (though premiums do rise temporarily compensating behavior by drivers of lighter
following a claim); medical costs primarilyvehicles are
may weaken the effect. And all else
external, again because they largely are being the same, light trucks do more damage
borne
by third parties. Productivity effects, to the
netoccupants
of of other vehicles than cars
taxes, are internal for own-driver injury do, risks,
as trucks are heavier, have suffer frames
though the revenue loss to the government (and therefore
is transfer more energy to other
external. vehicles), and have bumper heights that are
Recent studies using this general not compatible with cars.
approach put the marginal external costs at Most empirical literature on this issue has
around 2-7 cents per mile, or 13 to 44 per focused on the relationship between vehicle
cent of average accident costs (e.g., U.S. size/weight and total highway fatalities or
Department of Transportation, Federal injuries, and this literature has mixed results
Highway Administration 1997; Gunnar (e.g., Robert W. Crandall and John D.
Lindberg 2001; Mayeres, Oschelen, and Graham 1989; J. Daniel Khazzoom 1997;
Proost 1996; Miller et al. 1998; Parry 2004). Charles J. Kahane 1997; Douglas Coate and
James VanderHoff 2001; Robert B. Noland
2.5.3 Safety across Vehicle Types
2004). For our purposes, we are interested
A further unsettled issue is the relation only in how (marginal) external costs differ
between vehicle size/weight and safety, across vehicle types; external costs are quite
which matters for policies that encourage different from total injuries as they exclude
vehicle downsizing or affect fleet composi own-driver injury risks, but include factors
tion. We might expect lighter vehicles to be such as property damage and medical costs.

182.255.0.242 on Wed, 10 Nov 2021 04:05:08 UTC
Studies that use accident data to attribute damage is attributed to heavy-duty trucks
injuries to different vehicles involved in the (e.g., Small, Clifford Winston, and Carol A.
crash find only modest differences in exter Evans 1989; Newbery 1988). U.S.
nal costs per mile between cars and light Department of Transportation, Federal
trucks (U.S. Department of Transportation, Highway Administration (1997), table V-9,
Federal Highway Administration 1997; put external costs at only 0.06-0.08 cents per
Miller et al. 1998; Parry 2004). However, mile for passenger vehicles.
econometric studies by Michelle J. White
(2004) and Ted Gayer (2004), which are able 2.6.3 Urban Sprawl
to control for a wide range of nonvehicle Although the low cost of auto travel may
characteristics such as driver age, gender, contribute to urban fringe development,
region, speed, negligence, road class, weath there is little consensus on the magnitude of
er, seatbelt use, etc., reach a different con this effect on external costs such as traffic
clusion. For example, White (2004) finds congestion, loss of habitat, and open space
that the probability of a vehicle occupant amenities (McConnell and Margaret A.
being killed in a two-vehicle crash is 61 per Walls 2005). Moreover, if sprawl is excessive,
cent higher if the other vehicle is a light this primarily is due to tax preferences for
truck than if it is a car.
housing and the failure of development fees
2.6 Other Externalities and zoning restrictions to fully account for
the external and infrastructure costs of new
Other highway externalities are small,
apply primarily to heavy trucks, or result from development.
other policy failures rather than suboptimal
2.6.4 Parking Subsidies
automobile and fuel policies.
2.6.1 Noise Many individuals park for free when they
Noise costs (from engine acceleration, work or shop; Litman (2006) puts the costs
tire/road contact, braking, etc.) have been from these parking subsidies at 3-10 cents
inferred from hedonic property value models per VMT (after dividing subsidies by average
that include distance to local roads and traf distance traveled). Again though, there
fic volumes as explanatory variables (though remains dispute over whether free parking
should be attributed as an external cost of
it is difficult to control for noise mitigation
barriers, such as hills, sound-proof walls, automobile use as it results from other pric
double glazed windows). Delucchi and Hsu ing distortions, including employee cost
Shi-Lang Ling Hsu (1998) estimate costs of sharing and the treatment of parking as a
0-0.4 cents per mile for passenger vehicles, tax-preferred fringe benefit.
while U.S. Department of Transportation, 2.6.5 Other Environmental Externalities
Federal Highway Administration (1997),
table V-22, put average external costs at 0.06 Improper disposal of vehicles and parts
cents per mile. (e.g., tires, batteries, oil) can result in envi
ronmental and health hazards; however Lee
2.6.2 Highway Maintenance Costs
(1993) puts these costs at only 0.0015 cents
Analysts have estimated the effect of axle per VMT, and they probably have declined
loads and traffic volumes on pavement dam with more stringent regulations governing
age for different vehicle classes, controlling disposal and recycling. Damages from
for factors such as pavement age and climate. upstream emissions leakage from the petro
The key finding is that a vehicle causes road leum industry also are relatively small,
wear at a rate that is a sharply increasing func around 2 cents per gallon according to the
tion of the weight per axle, so that virtually all National Research Council (2002).

182.255.0.242 on Wed, 10 Nov 2021 04:05:08 UTC
TABLE 2
Summary of External Costs
cent/gal.a cents/milea
Central values for marginal external costs
Fuel-related costs
Greenhouse warming 6 0.3
Oil dependency 12 0.6
sum 18 0.9
Mileage-related co
Local pollution 42 2.0
Congestion, cents/mile 105 5.0
Accidents 63 3.0
sum 210 10.0
Note: aCosts converted as
2.7 Summary 3.1 Fuel Taxes
Table 2 summarizes, albeit very tenta Although gasoline taxes currently average
tively, our best assessment of major auto 40 cents per gallon (18.4 cents at the feder
mobile external costs omitting components, al level and on average 22 cents at the state
such as the geo-political costs of oil level), they are low not only by internation
dependence, that have not been quanti al standards (see figure 2) but also by his
fied. Given the popular focus on the need torical standards. When expressed on a
to reduce U.S. gasoline consumption per-mile basis, fuel taxes have declined in
because of energy security and climate real terms by 40 percent since 1960 due to
change, it is striking that these fuel-related the failure of nominal rates to keep pace
externalities add to only 18 cents per gallon with inflation and improved fuel econo
while mileage-related externalities (con my. This latter trend will only be com
gestion, accidents, and pollution) are pounded in the future with the market
equivalent to $2.10 per gallon. Naturally, penetration of hybrid and alternative fuel
these figures need to be updated over time; vehicles. We first discuss how effective fuel
for example, marginal local pollution and taxes are in reducing fuel consumption
accident costs will likely fall in future with and driving and then evaluate the existing
improved vehicle technology, while mar level of fuel taxes vis-?-vis fuel-related
ginal congestion and carbon costs will rise externalities.
over time.
3.1.1 Behavioral Responses
3. Traditional Policies The own-price elasticity of gasoline
demand has been estimated from time
This section discusses what historically
series models (often with a lag structure
have been the two most important fuel con
servation policies, namely fuel taxes and fuel
economy standards, as well as emissions per 10 From U.S. Department of Commerce, Census Bureau
mile standards and alternative fuel policies. (2003), table 730, and www.vtpi.org/tdm/fueltrends.xls.

182.255.0.242 on Wed, 10 Nov 2021 04:05:08 UTC
3.50
3.00
2.50
2.00
tit
1.504
III Mill
Mill Ml
1.00
0.504
0.00 llllllllllllll
*jr f.J> j?
<//// *ssss N*
* &
I unleaded gasoline I diesel for noncommercial use
Figure 2. Fuel Taxes for Selected Countries, 2004

Source: Organization for Economic Co-operation and Development (2005).
Joyce
imposed) and cross-section data (which areDargay, and Mark Hanly (2004) a
better able to control for household charac
Daniel J. Graham and Stephen Glaist
(2002)
teristics). A decade ago, reviews pointed to a put the elasticity at -0.7 and -
long run gasoline demand elasticity of
respectively. Estimated VMT/fuel pr
around -0.7 to -1.0 (Carol Dahl and
elasticities typically are around -0.1 to -
Thomas Sterner 1991, table 2; Phil B.
(Graham and Glaister 2002; Goodwin 199
Goodwin 1992, table 1; Molly Espey 1996,
Goodwin, Dargay, and Hanly 2004; Green
table 4). Later U.S. studies that better conR. Kahn, and Robert C. Gibson 1999
James
Olof
trol for fuel economy regulations and Johansson and Lee Schipper 1997; Pau
corre
lation among vehicle age, use, and Schimek
fuel 1996; Small and Kurt Van Dend
2006). Therefore, around 20-60 percent
economy, suggest a less elastic response;
another factor may be the decline of thefuel
gasoline demand elasticity appears
costs relative to the value of travelreflect
time. changes in VMT, while the oth
U.S. Department of Energy (1996) propercent reflects long-run changes
40-80
average
posed a long run fuel price elasticity of fleet fuel economy, as manufact
-0.38, though other reviews by Goodwin,
ers incorporate fuel-saving technologies in

182.255.0.242 on Wed, 10 Nov 2021 04:05:08 UTC
new vehicles and consumers buy smaller In fact this is a timely issue; given looming
vehicles. pressures on the entitlement system, there
3.1.2 Second-Best Fuel Taxes have been calls for higher fuel taxes and
other environmental taxes as part of deficit
Assuming, for now, that more finely tuned reduction packages (to reduce the need for
pricing policies are unavailable, there are two future increases in other taxes).
alternative conceptual approaches to gauging Leaving aside distributional issues, and col
the efficient level of fuel taxes (Newbery lapsing the broader fiscal system into a single
2005a). One is to estimate the Pigouvian tax, tax on labor income, the optimal fuel tax can
and then consider whether additional taxa
be decomposed into an externality correcting
tion might be warranted on broader fiscal and a fiscal or Ramsey tax component, where
grounds; the other is to have the fuel tax be the latter is positive if the taxed commodity is
chosen in part to guide highway spending a relatively weak substitute for leisure. Using
toward efficient levels.
this approach, Parry and Small (2005) put the
Pigouvian Tax. The Pigouvian gasoline tax, optimal fuel tax at 17 cents per gallon more
denoted by tG, is given by (Parry and Small than the pure Pigouvian tax when marginal
2005): fuel tax revenues finance a proportionate
reduction in labor income taxes. Sarah E.
(1) t?=EF+EMf?,
West and Roberten C. Williams (2007) sug
where EF is the cost per gallon of carbon and gest a somewhat larger upward adjustment to
oil dependence and EM is the cost per mile of the optimal tax, based on their econometric
local pollution, congestion, and accidents, ? estimates of the gasoline/leisure cross-price
is the fraction of the gasoline demand elas elasticity. The upward adjustment would be
ticity due to reduced mileage; the smaller larger still if, following the recycling of fuel
the tax-induced reduction in fuel use that tax revenues in income tax cuts, account were
comes from reduced driving, the smaller the taken of efficiency gains in the tax-distorted
reduction in mileage-related externalities capital market, and efficiency gains from
per gallon of fuel conservation and, corre reducing distortionary tax preferences for
spondingly, the smaller the Pigouvian tax. housing, employer medical insurance, and
Finally, / is average on-road vehicle fuel other fringe benefits (A. Lans Bovenberg and
economy, which converts mileage externali Lawrence H. Goulder 1997; Parry and
ties into costs per gallon, though it will vary Antonio M. Bento 2000).n Accounting for
with the fuel tax. Let fuel economy be broader fiscal considerations therefore can
f=f0{(pG+tG)/(pG+t%)Y1-P)T> , where pG is strengthen the efficiency case for raising fuel
the pretax price of gasoline, r?GG is the (con taxes above their current levels, though the
stant) own-price gasoline demand elasticity, overall optimal fuel tax is difficult to pin
and 0 denotes a current value. down, given uncertainty over behavioral
Suppose we use the figures in table 2 responses to broader tax adjustments.
along with /3=0.4, ?7GG=-0.55, pG=$1.60, Using Fuel Taxes to Promote Efficient
tG=40 cents, and/0 = 21. Then, using (1), Highway Spending. Taking a somewhat dif
and the fuel economy/fuel price relation, the ferent perspective, Newbery (2005a) argues
Pigouvian tax would be $1.11 per gallon,
nearly three times the current tax. 11 In contrast, Louis Kaplow (2004) finds that the opti
Broader Fiscal Rationale for Fuel Taxes. mal environmental tax equals the Pigouvian tax for "distri
Additional fuel taxation might be justified bution neutral" tax shifts in heterogeneous agent models;
when the efficient balance between fuel this result stems from his assumptions that the taxed com
modity is an average leisure substitute, and that all exter
taxes and labor income taxes in financing the nal costs reduce the marginal value of work relative to the
governments overall budget is considered. marginal value of leisure.

182.255.0.242 on Wed, 10 Nov 2021 04:05:08 UTC
for dividing fuel taxes into a road-user charge 1990; Erik Caspersen and Gilbert E. Metcalf
set to cover the maintenance and capital costs 1994), though some have argued against
of the road network and an additional com using lifetime income due to liquidity and
ponent to account for externalities other than other constraints on the ability to smooth
congestion and road damage (with excess consumption over the life cycle (Thomas A.
revenue going to the general government). Barthold 1993). A more comprehensive dis
The rationale for the road-user charge is tributional analysis also would take into
that if, on average, motorists are taxed for account who benefits from the recycling of
their contribution to congestion and pave fuel tax revenues; moreover, low-income
ment damage, then earmarking revenues will groups may benefit more (relative to their
lead to the efficient level of spending on income) from the mitigation of congestion
highway maintenance and expansion, assum and other highway externalities, as VMT and
ing constant returns to spending and that the value of travel time decline relative to
income for wealthier households (Mark
projects are determined by a cost/benefit cri
terion (Newbery 1988; Button 2004). Any Wardman 2001).
change in the composition of road user Even if any distributional concerns could
charges would be revenue-neutral; thus, the be fully addressed through adjustments to
introduction of congestion pricing in urban the broader fiscal and benefit system, there
areas (and pavement damage pricing for has been immense political opposition to
heavy trucks) automatically would result in higher fuel taxes in the United States. Most
compensation to the average motorist likely this opposition is due in part to politi
(though perhaps not to owners of commercial cally powerful auto manufacturers and oil
vehicles), through a rearrangement of nation companies, as well as to the greater vulnera
wide fuel taxes. For the United Kingdom, bility of households to fuel prices in the
Newbery (2005a) puts the road user compo United States, where annual gasoline con
nent at about half of the current fuel tax; pol sumption per capita is around 470 gallons
lution and other externalities bring the compared with only 90 gallons in Western
optimal tax up to about 70 percent of the Europe.1 Just because gasoline tax increas
U.K. fuel tax, or more than $2 per gallon. es have been politically unpalatable in the
past, however, does not mean that they will
3.1.3 Obstacles to Raising Fuel Taxes
be infeasible ten or twenty years down the
Whichever of the two conceptual road; pressure for higher fuel (or other auto
approaches outlined above is taken, the cur mobile) taxes from those concerned about
rent level of fuel taxation in the United global warming, energy dependence, under
States appears to be too low, leaving aside funded highways, and widening federal
the issue of better pricing instruments. deficits, is only likely to intensify.
Nonetheless, skeptics of higher fuel taxes Some observers have questioned whether
maintain that such taxes are unfair, are polit revenues from any fuel tax increase would be
ically untenable, or that governments may used wisely. Without any change in existing
end up wasting the extra revenue. legislation, most of the extra revenue would
As regards equity, most incidence studies be returned to state governments for high
find that gasoline taxes are regressive way projects, roughly in proportion to their
because lower income groups have larger
budget shares for gasoline. However, regres
sivity is mild when a measure of lifetime or 12 From www.eia.doe.gov/emeu/international/contents.
html. Consistent with this last point, Henrik Hammar, Asa
permanent income is used in place of annu L?fgren, and Sterner (2004) find that high gasoline con
al income (e.g., James M. Poterba 1989, sumption causes lower gasoline prices across OECD
nations.
1991; U.S. Congressional Budget Office

182.255.0.242 on Wed, 10 Nov 2021 04:05:08 UTC
share of total lane mile capacity. The federal fuel economy of new passenger vehicles is still
cost share in these projects, which ranges up below its peak in 1987 (see figure 3).
Higher CAFE standards have been ration
to 90 percent, is difficult to justify by appeal
ing to the benefits accruing to citizens of alized on the grounds that they reduce car
other states (Wallace E. Oates 2006). It is bon emissions and oil dependence, as well as
easy to see how, in such a funding environ possibly addressing a market failure from
ment, the selection of projects to go forward consumer undervaluation of fuel economy.
would be less than optimal. Despite this We take up both of these issues.
argument however, estimates of the social
3.2.1 Do Externalities Warrant Higher
rate of return to highway spending are typi CAFE Standards?
cally within a range of around 0-30 percent,
compared with a social discount rate of Raising CAFE standards actually may
around 5 percent (Transportation Research increase the (net) costs of highway externali
Board 2006). This suggests that at the margin ties (Andrew N. Kleit 2004; Fischer, Har
many highway projects would still increase rington, and Parry forthcoming), even though
efficiency, though not all analysts agree on higher fuel taxes would decrease them.
this (Chad Shirley and Winston 2004). For This paradox can be explained by decom
more substantial gasoline tax increases, legis posing the change in external costs, net of
lation might specify that revenues in excess of existing taxes, following an incremental adjust
the desired level of highway spending accrue ment to either policy into two components.
to the U.S. Treasury, though whether that First is the change in gasoline demand, times
would ultimately lead to a reduction in other external costs per gallon from fuel-related
taxes, or higher general public spending is externalities net of the currendy prevailing
unclear. Ideally perhaps, any excess revenues fuel tax, which incorporates some of the socie
would be earmarked for other socially desir tal costs of driving into the fuel price; second
able spending, such as basic R&D into clean is the change in VMT times marginal costs
fuel vehicles, or be included in a broader from externalities that vary with distance trav
package of environmental taxes that raises eled rather than fuel use. Increasing the fuel
enough revenue to cut income taxes by an tax above its current level actually reduces effi
amount that can be clearly specified in the ciency in the gasoline market if external costs
accompanying legislation. from carbon emissions and oil dependence are
below 40 cents per gallon (as they are in table
3.2 Fuel Economy Standards
2). However, this effect is offset easily by
The Corporate Average Fuel Economy externality benefits from reducing congestion,
(CAFE) program, established in the wake of accidents, and local pollution, as VMT falls;
the 1973 oil crisis, requires automobile manu hence the Pigouvian fuel tax is (much) greater
facturers to meet standards for the average than the current tax. In contrast, higher fuel
fuel economy of their passenger vehicle fleets. economy standards actually increase, rather
The light-truck standard will be increased to than reduce, VMT as they lower (rather than
22.2 miles per gallon by 2007, while the stan increase) fuel costs per mile driven, though
dard for cars, 27.5 miles per gallon, has been the increase in VMT is modest. Nonetheless,
fixed since 1985. Although the CAFE pro the impact of tightening CAFE could be to
gram significantly boosted fuel economy dur
ing the 1980s, it is unclear whether the 13 Manufacturers must pay a penalty of $55 per vehicle
standards presently are binding or not (Small for every one mpg that their fleet average falls below the
and Van Dender 2006). Moreover, due to the relevant standard. A lower standard for light trucks origi
nally was permitted to limit the burden on industrial
rising share of light-duty trucks, which now interests, though this argument lost its relevance with the
account for half of new vehicle sales, average rapid growth in use of light trucks for passenger vehicles.

182.255.0.242 on Wed, 10 Nov 2021 04:05:08 UTC
Figure 3. Certified Fuel Economy of New Passenger Vehicles

Source: National Research Council (2002).
increase external costs overall (at least external exceed the upfront installation costs. Are
costs that have been quantified).14 there reasons why such seemingly cost-effec
tive technologies may not be adopted by the
3.2.2 Information Market Failures
market, without more stringent regulation?
Even if the externality rationale for higher One possibility is that consumers under
fuel economy standards may be open to value new vehicle fuel economy because
question, might they be justified by an they have short planning horizons, high dis
another, informational market failure? As count rates, care more about other vehicle
documented by National Research Council attributes, or do not expect savings to be
(2002), there is a wide range of existing and reflected in used vehicle prices (Greene
emerging technologies for increasing new 1998). Unfortunately, there is little in the
vehicle fuel economy for which the dis way of solid empirical (as opposed to anec
counted, lifetime fuel savings appear to dotal) evidence on this hotly contested
issue.1 Another possibility is that engineer
14 Small and van Van Dender (2006) project that only ing studies have ignored alternative uses for
around 10 percent of fuel savings from improved fuel
economy will be offset by extra driving. The effect of
tighter fuel economy regulation on mileage is not sym Though there is an empirical literature finding that
metrical to the effect of higher fuel taxes; in each case the consumer discount rates exceed market rates for many
demand for vehicles falls (either because vehicle prices or energy saving technologies, we are aware of only one
fuel costs increase), playing a counteracting effect in the study, by Mark K. Dreyfus and W. Kip Viscusi (1995),
first case, and a reinforcing effect in the second. The applied specifically to automobiles. They estimated
above discussion assumes the value of a dollar of gasoline implicit consumer discount rates for future fuel savings of
tax revenue is a dollar. As suggested above, the social value 11-17 percent, though average interest rates on car loans
may exceed a dollar, given that revenues are earmarked were 13-15 percent during their sample period, suggest
for highway spending. In this case, the efficiency loss from ing that high discount rates may reflect credit constraints
the reduction in gasoline tax revenues is slightly larger, as rather than myopia. A big data problem has been control
it crowds out socially desirable public spending (Fischer, ling for all other relevant vehicle characteristics when
Harrington, and Parry forthcoming). comparing prices of vehicles with different fuel economy.

182.255.0.242 on Wed, 10 Nov 2021 04:05:08 UTC
new technologies; the economic cost of a that: (a) total VMT changes; (b) the share of
technology used to improve fuel economy is cars versus light trucks in the vehicle fleet
the larger of its installation cost and its changes; and (c) manufacturers produce
potential value if used instead to enhance smaller or lighter vehicles. The first effect
other vehicle attributes such as horsepower probably is fairly minor, while the second
(David Austin and Terry Dinan 2005). And a could be important if the lower standard for
final possibility, particularly given current light trucks were removed, for example, by
fuel prices, is that emerging technologies allowing trading of CAFE credits across cars
will penetrate the market over time in and light trucks. There is little solid evidence
response to consumer demand for better on how vehicle downweighting would affect
fuel economy, in which case higher mileage a comprehensive measure of accident exter
standards may have little effect. nalities.
Whether higher fuel economy standards 3.3 Emissions Standards and Related
would increase or reduce efficiency, or have Policies
little effect, remains unsettled. Kleit (2004)
and Austin and Dinan (2005) find that costs The principal instrument targeted at
from binding increases in standards of 3-4 local emission rates is the set of new-vehi
miles per gallon would be around $3-4 billion cle grams-per-mile exhaust standards for
or more, assuming market adoption of all pri HC, NOx, and CO introduced in the 1970
vately cost-effective technologies. Higher Clean Air Act (CAA), and subsequently
fuel economy standards significantly increase tightened over time. In fact, once the "Tier
efficiency only if carbon and oil dependence Two" standards (which apply to all cars and
externalities greatly exceed the mainstream light trucks) are fully phased in, new vehi
estimates in table 2, or if consumers perceive cle emission rates will be just 0.8-5.0 per
only about a third of the actual fuel economy cent of pre-1970 rates (table 3); these
benefits (Fischer, Harrington, and Parry reductions have been achieved through
forthcoming). Economic analyses also strong technology improvements such as the
ly support higher fuel taxes over higher fuelthree-way catalyst that adjusts the air/fuel
economy standards (Kleit 2004; Austin and ratio using computer-controlled sensors.
Dinan 2005; West and Williams 2005). The 1977 CAA Amendments also required
But what if the only immediate choice is programs subjecting vehicles to periodic
to do nothing or to gradually increase CAFE emissions inspections and, if necessary,
standards? Although it can be argued both repairs to meet the standard for regions
ways, we ourselves would probably lean with air quality problems.
toward the latter, given the possible geo Automotive fuels also are subject to vari
political benefits of reduced oil dependence ous regulations. According to Newell and
that are not included in the above studies. At Kristian Rogers (forthcoming), the phase-out
the very least, the CAFE program could be of leaded gasoline produced benefits (e.g.,
reformed in ways that make economic sense; health improvements from reduced inci
for example, allowing manufacturers to trade dence of high blood pressure) that exceeded
fuel economy credits across their car and the control costs, and also facilitated the use
light-truck fleets would significantly reduce of catalytic converters to control other emis
compliance costs (Austin and Dinan 2005; sions. The 1990 CAA Amendments required
Jonathan Rubin, Leiby, and Greene 2006).
16 Due to its unique air quality problems, California can
3.2.3 Accidents and CAFE set its own standards, and these have generally been
stricter than the federal standards. Other states may adopt
Tighter fuel economy regulations may the California standards but not set their own independent
affect external accident costs to the extent standards.

182.255.0.242 on Wed, 10 Nov 2021 04:05:08 UTC
TABLE 3
New Car Emissions Standards, Grams Per Mile
Model year HC CO NOx

Pre-control vehicle 10.6 84.0 4.1
1970-71 4.1 34.0
1972 3.4 39.0
1973-74 3.4 39.0 3.0
1975-76 1.5 15.0 3.1
1977-79 1.5 15.0 2.0
1980 0.41 7.0 2.0
1981-93 0.41 3.4 1.0
1994-03 0.25 3.4 0.4
2004-07 0.09 4.2 0.07
Source: National Academy of Sciences (2001).
that oxygenated fuels be used in certain

In theory, the ideal instrument to control
urban areas during winter, a requirement
local tailpipe emissions would be an emis
sions tax levied on each vehicle that varied
that since has been replaced by an ethanol
mandate (see below). And, to reduce with
evap local population exposure and time of
orative HC emissions, the reformulated
year. However, the existing combination of
gasoline program specifies a maximum
emissions standards, inspection programs,
allowable fuel vapor pressure during fuel
sumtaxes, and fuel regulations does not rely
on the measurement of in-use emissions,
mer in areas out of compliance with national
ozone standards. and thus may be simpler to administer.
Moreover, there may not be much differ
EPA conducted a retrospective cost/bene
encet
fit analysis of the CAA over 1970-1990 and a on cost-effectiveness grounds as
prospective analysis for 1990-2010 (U.S.together these instruments exploit all the
Environmental Protection Agency 1997, potential margins for emissions reductions.
1999). The retrospective analysis put the
3.4 Alternative Fuel Policies
costs of auto emissions controls at $5.5 bil
Government policies also encourage pro
lion in 1990, while the prospective study,
duction of and research into a variety of
which includes the costs of tighter emissions
standards, inspection and maintenance emerging
pro and potential alternatives to
gasoline.
grams, and the reformulated and oxygenated
Ethanol receives a federal tax credit of 51
fuel programs, estimates costs at $9.1 billion
cents per gallon, and the 2005 Energy
and $12.3 billion for 2000 and 2010 respec
PolicyofAct imposes minimum ethanol pur
tively (in 1990$). Although the benefits
reduced automobile emissions arechase
notrequirements on refineries. However,
explicitly decomposed from those attributa
due to opposition from agricultural and pro
duceritinterests, ethanol imports have high
ble to stationary sources in these analyses,
seems highly likely that they would exceed
tariffs. Ethanol is blended with gasoline to
the estimated control costs. make E10 (10 percent ethanol, 90 percent

182.255.0.242 on Wed, 10 Nov 2021 04:05:08 UTC
gasoline) though, at modest extra cost, flexi costs of reducing gasoline dependence
ble fuel vehicles can be produced to run on through subsidizing the adoption of alter
either gasoline or an 85 percent ethanol native fuel vehicles more generally rise
blend (these vehicles are further encouraged sharply, as they are more expensive and
through CAFE credits). Ethanol currently is have shorter driving ranges than their gaso
produced from corn however, once the pro line counterparts (e.g., Alan J. Krupnick
duction technology has evolved, the hope is and Walls 1992; Walls 1996; Kazimi 1997;
to produce cellulosic ethanol from switch Leiby and Jonathan Rubin 2001). Although
grass and other fibrous plants; cellulosic it makes sense to invest in a diverse portfo
ethanol would not consume valuable farm lio of basic R&D prospects to reduce long
land and, accounting for energy used in fuel term oil dependence, tax credits for
production, the carbon savings over a gallon alternative fuels or alternative fuel vehicles
of gasoline are around 80 percent, compared are far less efficient instruments than fuel
with 20 percent for corn ethanol (Newell taxes, as they do not exploit the entire
2006). Although U.S. Department of range of fuel conserving options, which
Energy, Energy Information Administration include reduced use, and improved fuel
(2006b) projects ethanol production will rise economy, of conventional vehicles.
to 10-14 billion gallons by 2030, this is still
less than a tenth of projected gasoline 4. Emerging Pricing Policies
demand.
The government also is subsidizing an We now discuss more finely tuned instru
effort to develop vehicles that run on elec ments for addressing congestion and accidents
tricity produced from a reaction between that ideally would complement fuel taxes and
that have become feasible with advances in
hydrogen and oxygen, but the technological electronics and telecommunications.
challenges are formidable (National
Academy of Sciences 2004). These include 4.1 Congestion Tolls
the high cost of hydrogen fuel cells, the
problem of safe on-board storage (even Although congestion pricing, which
compressed hydrogen takes up far more involves charging motorists the difference
space than other fuels); the current high between marginal costs to all drivers and
costs of producing hydrogen from fossil average travel cost to the individual at a
fuels or water; and the cost of transitioning point in time, largely has been ignored by
to a fuel distribution network based on transportation policymakers to date, the
hydrogen. A possible intermediate technol confluence of several factors makes this an
ogy that could become competitive, at least especially favorable time for its serious
after substantial advances in battery tech consideration.1 One is the increasing diffi
nology, is the plug-in hybrid, which com culty of building new roads, due to rising
bines an internal combustion engine, urban property values and opposition from
regenerative braking, and on-board battery neighborhood and environmental groups,
storage that can be recharged from a power along with the realization that road build
socket. ing is partly self-defeating as it encourages
The new ethanol mandate (if binding)
seems likely to reduce efficiency, given the Many roads have tolls for revenue collection but
large preexisting tax preference for ethanol these vary little by time of day. Until recently, the only
(not to mention agricultural subsidies for other notable examples of road pricing were the cordon
tolling implemented in Singapore in the mid-1970s, which
corn production), and the relatively modest was effective in limiting downtown traffic, and the less
per-gallon externalities from gasoline itself. effective cordon toll rings implemented in several
And studies suggest that the incremental Norwegian cities in the 1980s (Georgina Santos 2004).

182.255.0.242 on Wed, 10 Nov 2021 04:05:08 UTC
more driving.1 Another is the growing be second-best optimal if they are set by
need for new fiscal instruments to pay for private, revenue-maximizing operators.
highways, given the steady erosion of real On the political side, electronic toll collec
fuel-tax revenue per VMT and the recogni tion raises privacy concerns, even though sys
tion that road pricing would aid in indicat tems can be designed to avoid the central
ing where capacity additions would have collection of travel information; in fact mar
greatest benefit (Transportation Research ket penetration of transponders is surprising
Board 2006; Todd Goldman and Martin ly low so far, less than 50 percent in most U.S.
Wachs 2003). Furthermore, congestion toll road systems. Another barrier is that,
fees now can be deducted electronically by prior to the distribution of congestion toll
in-vehicle transponders, thereby avoiding revenues, many motorists are worse off in
bottlenecks at worker-operated tollbooths, terms of paying tolls that exceed their value
or through direct billing with on-board of travel time savings; this makes some
global positioning systems. But there observers pessimistic that congestion pricing
remain formidable practical and political will ever be widely implemented (e.g.,
obstacles to widespread implementation of Genevi?ve Giuliano 1992).
ideal congestion pricing. One possibility for overcoming political
On the practical side, it may be computa opposition is judicious use of congestion toll
tionally infeasible to estimate marginal con revenues to create a broader coalition of
gestion costs on every single link and winners from the policy change. For exam
intersection in an urban road network, par ple, Small (1992b) and Goodwin (1995) rec
ticularly given that pricing at one point ommend a mix of spending on transportation
diverts traffic elsewhere within the network, alternatives, road improvements, and reduc
and that stop-and-go queues at bottlenecks tions in other taxes, while a stated prefer
are subject to rapid and substantial change. ence study by Harrington, Krupnick and
Even if such a pricing system could be reli Anna Alberini (2001) finds a discernable
ably simulated, it would impose substantial increase in support for congestion pricing
information processing costs on drivers; when toll revenues are recycled in the form
moreover their ability to respond to charges of other local tax reductions.19 Another pos
that vary in real time may be limited once sibility is to begin with pricing reforms for
the trip has begun. Thus far, most actual or which there is the least opposition, such as
proposed congestion pricing schemes vary converting high occupancy vehicle lanes to
with time of day (rather than with real-time high occupancy/toll lanes that single-occu
traffic conditions), and are confined to major pant vehicles can pay to use, and then build
urban expressways, or cordon tolls; this more ing new toll-lane capacity to create a
piecemeal approach limits efficiency gains, network of premium lanes covering an
partly because of worse congestion resulting entire urban area (Robert W. Poole and C.
elsewhere in the network (e.g., Small and Jia Kenneth Orski 2003; Transportation
Yan 2001; Erik Verhoef 2005; Newbery and Research Board 2006).
Santos 2002). Moreover, toll levels may not In Europe, congestion pricing may be on
the verge of breaking out into much more
widespread use. Following the successful cor
18 Someday, freeway capacity might be dramatically don toll introduced in central London in 2003,
increased without more pavement; advanced sensing
technologies, which are already being incorporated into the national U.K. government is considering
high-end vehicles, together with embedded detectors in
highways, may permit "platoons" of vehicles to travel
together at high speed under computer control with min Productive use of revenues is also important to
imal headway between vehicles (Roberto Horowitz and ensure that congestion tolls are efficiency improving over
Pravin Varaiya 2000). all (Parry and Bento 2001).

182.255.0.242 on Wed, 10 Nov 2021 04:05:08 UTC
replacing its fuel taxes with a system of per 2007). PAYD would provide incentives to
mile charges that would vary sharply with limit driving trips, unlike the current insur
region and time of day (Jonathan Leape ance system where annual premiums
2006; Santos 2004). Stockholm introduced a depend only weakly on mileage, and these
time-varying cordon toll in 2006 and heavy incentives would be greatest for the drivers
truck tolls have been introduced in Germany with highest risk factors. Drivers with below
and Switzerland. Though there is growing average mileage (who are the majority given
interest in the United States, congestion pric that the mileage distribution across drivers is
ing has been slower to get off the ground, skewed to the left) have an incentive to opt
and applications have been limited to the for PAYD as it would lower their annual pay
construction of high occupancy/toll lane ments; premiums would rise for those
capacity (e.g., SR91 in Orange and Riverside remaining under the present system, which
Counties in Southern California, 1-15 in San in turn would encourage additional switch
Diego and I-10 in Houston). ing at the margin. Edlin (2003) estimates
substantial efficiency gains from switching to
4.2 Charging for Accident Risk
PAYD, particularly when the reduction in
In this last section, we focus on innovative the entire spectrum of auto externalities is
pricing policies to internalize accident risks taken into account.
into the choice of how much to drive.20 As
already mentioned, the ideal policy would 5. Conclusion
be to charge drivers according to the mar
It could be argued that the externality
ginal external accident cost per mile, though
rationale for higher fuel taxes has come and
measuring how this varies across individu
gone. Electronic road pricing offers the only
als, vehicles, and regions is problematic.
Still, even a uniform VMT tax is more cost real hope of addressing relentlessly increas
ing urban gridlock, while encouraging a
effective at reducing accidents than fuel
transition to mileage-based insurance would
taxes as all, rather than a portion, of the
improve highway safety more effectively.
behavioral response to it comes from
And local tailpipe emissions are rapidly
reduced driving.
declining with improved technology to meet
In fact a form of charging by the mile for
progressively more stringent new-vehicle
accident risk may emerge on its own through emissions standards.
the market. This is pay-as-you-drive (PAYD)
insurance under which a person s insurance
That leaves climate change and oil
dependence, but even for these problems
payment would vary in direct proportion to
other policies are more appropriate. Climate
annual VMT, scaled by the drivers (and pos
change requires an economywide approach,
sibly the vehicles) relative risk factor. This
particularly given that most of the low-cost
risk factor would be determined by insur
options for emissions abatement are in the
ance companies, and it has been estimated
power sector. Economists typically favor a
that the charge for the average motorist
carbon tax imposed upstream on fossil fuels,
would be around 6 cents per mile (Litman
moderately scaled at first but rising steadily
20 Pricing and various nonpecuniary penalties are also

used to deter drunk drivers, though the expected penalty 21 PAYD schemes are emerging at the state level. In
per trip is well below the external cost per trip, due to the Oregon, insurance companies have been offered a state tax
very small likelihood of arrest (Donald S. Kenkel 1993; credit of $100 per motorist for the first 10,000 motorists
Steven D. Levitt and Jack Porter 2001). A promising new who sign up for PAYD. The Texas Legislature has passed
technology for deterring recidivism is the alcohol inter legislation authorizing auto insurance companies to offer
lock ignition technology, which is increasingly required for per-mile insurance, and state governments in Maryland
DUI offenders in New Mexico. and Connecticut are considering similar measures.

182.255.0.242 on Wed, 10 Nov 2021 04:05:08 UTC
over time, with credits for downstream car with those from other environmental taxes,
bon capture and sequestration. Similarly, such as carbon taxes (or auctioned carbon
taxing all oil products, including aviation permits), and used to lower personal income
fuel, diesel fuel, home heating oil, and petro taxes or pay down the federal deficit.
chemicals, would be more cost-effective References
than taxing gasoline alone. Moreover, higher
energy taxes need to be buttressed with Arnott, Richard, Andre de Palma, and Robin Lindsey.
1993. "A Structural Model of Peak-Period
technology policies if a radical transition Congestion: A Traffic Bottleneck with Elastic
away from fossil fuels is to be achieved over Demand." American Economic Review, 83(1):
161-79.
the next generation or two, though econom
Arnott, Richard, Andre de Palma, and Robin Lindsey.
ic analysis is needed on how prevailing R&D 1994. "The Welfare Effects of Congestion Tolls with
spending, and its composition across differ Heterogeneous Commuters." Journal of Transport
Economics and Policy, 28(2): 139-61.
ent technological possibilities, could be Arnott, Richard, and Marvin Kraus. 1995. "Financing
improved. Capacity in the Bottleneck Model. "Journal of Urban
Nonetheless, given that the widespread Economics, 38(3): 272-90.
Austin, David, and Terry Dinan. 2005. "Clearing the
adoption of the ideal set of externality taxes is
Air: The Costs and Consequences of Higher CAFE
unlikely in the near term, a progressive Standards and Increased Gasoline Taxes." Journal of
increase in the federal gasoline tax seems to Environmental Economics and Management, 50(3):
562-82.
make sense after all, given that it is adminis Barthold, Thomas A. 1993. "How Should We Measure
tratively simple, and unlike fuel economy Distribution?" National Tax Journal, 46(3): 291-99.
standards, it favorably affects all the external Beccue, Phillip C, and Hillard G. Huntington. 2005.
ities and raises revenue. The model here "An Assessment of Oil Market Disruption Risks."
Energy Modeling Forum Working Group Report,
might be the "fuel tax escalator" in place in no. EMF SR 8.
the United Kingdom during the 1990s under Bohi, Douglas R., and Michael A. Toman. 1996. The
Economics of Energy Security. Boston; Dordrecht
which the nominal tax rate automatically and London: Kluwer Academic.
increased each year; a four cent-per-gallon Bovenberg, A. Lans, and Lawrence H. Goulder. 1997.
increase each year would double (federal and "Costs of Environmentally Motivated Taxes in the
state) fuel taxes within a decade. Legislation Presence of other Taxes: General Equilibrium
Analyses." National Tax Journal, 50(1): 59-87.
could be introduced to require periodic Button, Kenneth. 2004. "The Rationale for Road
review and reauthorization of the fuel tax Pricing: Standard Theory and Latest Advances." In
Road Pricing: Theory and Evidence, ed. G. Santos.
escalator, and possibly require its suspension Research in Transportation Economics, vol. 9. San
in the event of an oil price shock, or if rev Diego and Amsterdam: Elsevier, JAI, 3-25.
enues from other road-user fees become Caspersen, Erik, and Gilbert E. Metcalf. 1994. "Is a
more significant. The thorniest issue would Value Added Tax Regressive? Annual versus
Lifetime Incidence Measures." National Tax
be designing legislation to ensure the produc Journal, 47(4): 731-46.
tive use of additional revenue flows. Initially, Chay, Kenneth Y., and Michael Greenstone. 2005.
given the pressure on local transportation "Does Air Quality Matter? Evidence from the
Housing Market." Journal of Political Economy,
budgets, revenues might remain earmarked 113(2): 376-424.
for highways. Down the road, legislation Chernick, Howard, and Andrew Reschovsky. 1997.
"Who Pays the Gasoline Tax?" National Tax Journal,
might require that revenues be combined 50(2): 233-59.
Coate, Douglas, and James VanderHoff. 2001. "The
Taxes have a number of advantages over a cap-and Truth about Light Trucks." Regulation, 24(1): 22-27.
trade permit approach (Nordhaus forthcoming). For Crandall, Robert W, and John D. Graham. 1989. "The
example, freely allocated permits do not raise government Effect of Fuel Economy Standards on Automobile
revenues, they transfer large rents from energy consumers Safety." Journal of Law and Economics, 32(1):
to (wealthy) stockholders in energy companies, volatility 97-118.
in carbon prices can deter large fuel-saving capital invest Dahl, Carol, and Thomas Sterner. 1991. "Analysing
ments, and rancorous negotiations over country-level quo Gasoline Demand Elasticities: A Survey." Energy
tas would be avoided if agreement were only needed on a Economics, 13(3): 203-10.
harmonized carbon tax. De Borger, Bruno, and Stef Proost, eds. 2001.

182.255.0.242 on Wed, 10 Nov 2021 04:05:08 UTC
Reforming Transport Pricing in the European Union: Goodwin, Phil B., Joyce Dargay, and Mark Hanly.
A Modelling Approach. Transport Economics, 2004. "Elasticities of Road Traffic and Fuel
Management and Policy series. Cheltenham, U.K. Consumption with Respect to Price and Income: A
and Northampton, Mass.: Elgar. Review." Transport Reviews, 24(3): 275?92.
Delucchi, Mark A. 2000. "Environmental Externalities Graham, Daniel J., and Stephen Glaister. 2002. "The
of Motor-Vehicle Use in the US." Journal of Demand for Automobile Fuel: A Survey of
Transport Economics and Policy, 34(2): 135-68. Elasticities." Journal of Transport Economics and
Delucchi, Mark A., and Shi-Ling Hsu. 1998. "The Policy, 36(1): 1-26.
External Damage Cost of Noise Emitted from Motor Greene, David L. 1998. "Why CAFE Worked." Energy
Vehicles." journal of Transportation and Statistics, Policy, 26(8): 595-613.
1(3): 1-24. Greene, David L., Donald W. Jones, and Mark A.
Delucchi, Mark A., and James Murphy. 2004. "U.S. Delucchi, eds. 1997. The Full Costs and Renefits of
Military Expenditures to Protect the Use of Persian Transportation: Contributions to Theory, Method
Gulf Oil for Motor Vehicles." University of California and Measurement. Heidelberg and New York:
Davis Research Report, no. UCD-ITS-RR-96-03(15) Springer.
rev2. Greene, David L., James R. Kahn, and Robert C.
Dockery, Douglas W., C. Arden Pope, Xiping Xu, John Gibson. 1999. "Fuel Economy Rebound Effect for
D. Spengler, James H. Ware, Martha E. Fay, U.S. Household Vehicles." Energy Journal, 20(3):
Benjamin G. Ferris, and Frank E. Speizer. 1993. "An 1-31.
Association between Air Pollution and Mortality in Hall, Fred L., and Lisa M. Hall. 1990. "Capacity and
Six U.S. Cities." New England Journal of Medicine, Speed Flow Analysis of the QEW in Ontario."
329(24): 1753-59. Transportation Research Record, 1287: 108-18.
Dreyfus, Mark K., and W. Kip Viscusi. 1995. "Rates of Ham mar, Henrik, Asa L?fgren, and Thomas Sterner.
Time Preference and Consumer Valuations of 2004. "Political Economy Obstacles to Fuel
Automobile Safety and Fuel Efficiency." Journal of Taxation." Energy Journal, 25(3): 1-17.
Law and Economics, 38(1): 79-105. Harrington, Winston. 1997. "Fuel Economy and Motor
Edlin, Aaron S. 2003. "Per-Mile Premiums for Auto Vehicle Emissions." Journal of Environmental
Insurance." In Economics for an Imperfect World: Economies and Management, 33(3): 240-52.
Essays in Honor of Joseph E Stiglitz, ed. R. Arnott et Harrington, Winston, Alan J. Krupnick, and Anna
al. Cambridge and London: MIT Press, 53-82. Alberini. 2001. "Overcoming Public Aversion to
Edlin, Aaron S., and Pinar Karaca-Mandic. 2006. "The Congestion Pricing." Transportation Research: Part
Accident Externality from Driving." Journal of A: Policy and Practice, 35(2): 87-105.
Political Economy, 114(5): 931-55. Harrington, Winston, and Virginia McConnell. 2003.
Espey, Molly. 1996. "Explaining the Variation in "Motor Vehicles and the Environment." In The
Elasticity Estimates of Gasoline Demand in the International Yearbook of Environmental and
United States: A M eta-analysis." Energy Journal, Resource Economics 2003/2004: A Survey of Current
17(3): 49-60. Issues, ed. H. Folmer and T. Tietenberg. New
Fischer, Carolyn, Winston Harrington, and Ian W II. Horizons in Environmental Economics. Cheltenham,
Parry. Forthcoming. "Should Automobile Fuel U.K. and Northampton, Mass.: Elgar, 190-268.
Economy Standards Be Tightened?" Energy Journal. Hoerner, Andrew J., and Beno?t Bosquet. 2001.
Gayer, Ted. 2004. "The Fatality Risks of Sport-Utility Environmental Tax Reform: The European
Vehicles, Vans, and Pickups Relative to Cars." Experience. Washington, D.C.: Center for a
Journal of Risk and Uncertainty, 28(2): 103-33. Sustainable Economy.
Gerard, David, and Lester B. Lave. 2004. "Why Gas Horowitz, Roberto, and Pravin Varaiya. 2000. "Control
Taxes and Fuel Economy Standards Are Too Low: Design of an Automated Highway System."
The Economics of the CAFE Reconsidered." Proceedings of the IEEE, 88(7): 913-25.
Unpublished. Huntington, Hillard G. 2005. "The Economic
Giuliano, Genevi?ve. 1992. "An Assessment of the Consequences of Higher Crude Oil Prices." Energy
Political Acceptability of Congestion Pricing." Modeling Forum Working Group Report, no. EMF
Transportation, 19: 335-58. SR 9.
Goldman, Todd, and Martin Wachs. 2003. "A Quiet Johansson, Olof, and Lee Schipper. 1997. "Measuring
Revolution in Transportation Finance: The Rise of the Long-Rim Fuel Demand of Cars: Separate
Local Option Transportation Taxes." Transportation Estimations of Vehicle Stock, Mean Fuel Intensity,
Quarterly, 57(1): 19-32. and Mean Annual Driving Distance." Journal of
Goodwin, Phil B. 1992. "A Review of New Demand Transport Economics and Policy, 31(3): 277-92.
Elasticities with Special Reference to Short and Kahane, Charles J. 1997. "Relationships between
Long Run Effects of Price Changes." Journal of Vehicle Size and Fatality Risk in Model Year
Transport Economics and Policy, 26(2): 155-69. 1985-93 Passenger Cars and Light Trucks." National
Goodwin, Phil B. 1995. "Road Pricing or Transport Highway Traffic Safety Administration Technical
Planning?" In Road Pricing: Theory, Empirical Report, no. DOT HS 808 570.
Assessment and Policy, ed. B. Johansson and L.-G. Kaplow, Louis. 2004. "On the (Ir)Relevance of
Mattsson. Transportation Research, Economics and Distribution and Labor Supply Distortion to Govern
Policy, vol. 3. Boston and Dordrecht: Kluwer ment Policy." Journal of Economic Perspectives,
Academic, 143-58. 18(4): 159-75.

182.255.0.242 on Wed, 10 Nov 2021 04:05:08 UTC
Kazimi, Camilla. 1997. "Valuing Alternative-Fuel Policy, 33(3): 253-86.

Vehicles in Southern California." American Economic Mendelsohn, Robert, Wendy Morrison, Michael E.
Review, 87(2): 265-71. Schlesinger, and Natalia G. Andronova. 2000.
Kenkel, Donald S. 1993. "Do Drunk Drivers Pay Their "Country-Specific Market Impacts of Climate
Way? A Note on Optimal Penalties for Drunk Change." Climatic Change, 45(3-4): 553-69.
Driving." Journal of Health Economics, 12(2): Miller, Ted R. 1993. "Costs and Functional
137-49. Consequences of U.S. Roadway Crashes." Accident
Khazzoom, J. Daniel. 1997. "Impact of Pay-at-the Analysis and Prevention, 25(5): 593-607.
Pump on Safety through Enhanced Vehicle Fuel Miller, Ted R., D. T. Levy, R. S. Spicer, and D. C.
Efficiency." Energy Journal, 18(3): 103-33. Leuna. 1998. "Allocating the Costs of Motor Vehicle
Kleit, Andrew N. 2004. "Impacts of Long-Range Crashes between Vehicle Types." Transportation
Increases in the Fuel Economy (CAFE) Standard." Research Record, 1635: 81-87.
Economic Inquiry, 42(2): 279-94. Mun, Se-il. 1994. "Traffic Jams and the Congestion
Krupnick, Alan J., and Margaret A. Walls. 1992. "The Toll." Transportation Research: Part B:
Cost-Effectiveness of Methanol for Reducing Motor Methodological, 28(5): 365-75.
Vehicle Emissions and Urban Ozone." Journal of National Academy of Sciences. 2004. The Hydrogen
Policy Analysis and Management, 11(3): 373-96. Economy: Opportunities, Costs, Barriers, and Rl?D
Leape, Jonathan. 2006. "The London Congestion Needs. Washington, D.C.: National Academy Press.
Charge." Journal of Economic Perspectives, 20(4): National Research Council. 2002. Effectiveness and
157-76. Impact of Corporate Average Fuel Economy (CAFE)
Lee, Douglas B. 1993. Full Cost Pricing of Highways. Standards. Washington, D.C.: National Academy
Washington, D.C.: U.S. Department of Transpor Press.
tation. Newbery, David M. 1988. "Road Damage Externalities
Leiby, Paul N., Donald W Jones, T Randall Curlee, and Road User Charges." Econometrica, 56(2):
and Russell Lee. 1997. "Oil Imports: An Assessment 295-316.
of Benefits and Costs." Oak Ridge National Newbery, David M. 2005a. "Road User and Congestion
Laboratory Report, no. ORNL-6851. Charges." In Theory and Practice of Excise Taxation:
Leiby, Paul N., and Jonathan Rubin. 2001. Smoking, Drinking, Gambling, Polluting, and
"Effectiveness and Efficiency of Policies to Promote Driving, ed. S. Cnossen. Oxford and New York:
Alternative Fuel Vehicles." Transportation Research Oxford University Press, 193-229.
Record, 1750: 84-91. Newbery, David M. 2005b. "Why Tax Energy? Towards
Levitt, Steven D., and Jack Porter. 2001. "How a More Rational Policy." Energy Journal, 26(3):
Dangerous Are Drinking Drivers?" Journal of 1-39.
Political Economy, 109(6): 1198-1237. Newbery, David M., and Georgina Santos. 2002.
Lindberg, Gunnar. 2001. "Traffic Insurance and "Estimating Urban Road Congestion Charges."
Accident Externality Charges." Journal of Transport CEPR Discussion Paper, no. 3176.
Economics and Policy, 35(3): 399-416. Newell, Richard G. 2006. "What's the Big Deal about
Lindsey, Robin, and Erik T Verhoef. 2001. "Traffic Oil? How We Can Get Oil Policy Right." Resources,
Congestion and Congestion Pricing." In Handbook 163: 6-10.
of Transport Systems and Traffic Control, ed. K. J. Newell, Richard G., and William A. Pizer. 2003.
Button and D. A. Hensher. Handbooks in Transport, "Discounting the Distant Future: How Much Do
vol. 3. Amsterdam; London and New York: Elsevier Uncertain Rates Increase Valuations?" Journal of
Science, Pergamon, 77-105. Environmental Economics and Management, 46(1):
Litman, Todd. 2006. Transportation Cost and Renefit 52-71.
Analysis: Techniques, Estimates and Implications. Newell, Richard G., and Kristian Rogers. Forthcoming.
Victoria, Can.: Victoria Transport Policy Institute. "The Market-Based Lead Phasedown." In Carrots,
www.vtpi.org/tca. Sticks, and Sledgehammers: Are Market Instruments
Litman, Todd. 2007. Pay-as-You-Drive Vehicle Really Better for Environmental Protection, ed. C.
Insurance: Converting Vehicle Insurance Premiums Kolstad and J. Freeman. Oxford: Oxford University
into Use-Rased Charges. Victoria, Can.: Victoria Press.
Transport Policy Institute, www.vtpi.org/tdm/ Noland, Robert B. 2004. "Motor Vehicle Fuel
tdm79.htm. Efficiency and Traffic Fatalities." Energy Journal,
Mayeres, Inge, Sara Ochelen, and Stef Proost. 1996. 25(4): 1-22.
"The Marginal External Costs of Urban Transport." Nordhaus, William D. 2006. "The Stern Review on the
Transportation Research: Part D: Transport and Economics of Climate Change." nordhaus.econ.yale.
Environment, 1(2): 111-30. edu/SternReviewD2.pdf.
McConell, Virginia D., and Margaret A. Walls. 2005. Nordhaus, William D. Forthcoming. "To Tax or Not to
The Value of Open Space: Evidence from Studies of Tax: Alternative Approaches to Slowing Global
Nonmarket Benefits. Washington, D.C.: Resources Warming." Review of Environmental Economics and
for the Future. www.rff.org/Documents/RFF Policy.
REPORT-Open20Spaces.pdf. Nordhaus, William D., and Joseph Boyer. 2000.
McCubbin, Donald R., and Mark A. Delucchi. 1999. Warming the World: Economic Models of Global
"The Health Costs of Motor-Vehicle-Related Air Warming. Cambridge and London: MIT Press.
Pollution." Journal of Transport Economics and Oates, Wallace E. 2006. "On the Theory and Practice

182.255.0.242 on Wed, 10 Nov 2021 04:05:08 UTC
of Fiscal Decentralization." Institute for Federalism Density Matter?" Transportation Research Record,
and Intergovernmental Relations Working Paper, no. 1552: 120-25.
2006-05. Shirley, Chad, and Clifford Winston. 2004. "Firm
Organisation for Economic Co-operation and Inventory Behavior and the Returns from Highway
Development. 2005. Energy Prices and Taxes: Infrastructure Investments." Journal of Urban
Quarterly Statistics: First Quarter: 2005. Paris and Economics, 55(2): 398-415.
Washington, D.C.: Organisation for Economic Co Schrank, David, and Timothy Lomax. 2005. The 2005
operation and Development. Urban Mobility Report. College Station: Texas
Parry, Ian W H. 2004. "Comparing Alternative Policies Transportation Institute, Texas A&M University.
to Reduce Traffic Accidents." Journal of Urban Schwartz, Joel. 1994. "Air Pollution and Daily
Economics, 56(2): 346-68. Mortality: A Review and Meta Analysis."
Parry, Ian W H., and Antonio M. Bento. 2000. "Tax Environmental Research, 64(1): 36-52.
Deductions, Environmental Policy, and the 'Double Small, Kenneth A. 1992a. Urban Transportation
Dividend' Hypothesis." Journal of Environmental Economics. Fundamentals of Pure and Applied
Economics ana Management, 39(1): 67-96. Economics series. Regional and Urban Economics
Parry, Ian W H., and Antonio M. Bento. 2001. section. Philadelphia and Reading, U.K.: Harwood
"Revenue Recycling and the Welfare Effects of Road Academic.
Pricing." Scandinavian Journal of Economics, Small, Kenneth A. 1992b. "Using the Revenues from
103(4): 645-71. Congestion Pricing." Transportation, 19(4): 359-81.
Parry, Ian W H., and Kenneth A. Small. 2005. "Does Small, Kenneth A., and Xuehao Chu. 2003.
Britain or the United States Have the Right Gasoline "Hypercongestion." Journal of Transport Economics
Tax?" American Economic Review, 95(4): 1276-89. and Policy, 37(3): 319-52.
Pearce, David. 2005. "The Social Cost of Carbon." In Small, Kenneth A., and Camilla Kazimi. 1995. "On the
Climate-Change Policy, ed. D. Helm. Oxford and Costs of Air Pollution from Motor Vehicles." Journal
New York: Oxford University Press, 99-133. of Transport Economics and Policy, 29(1): 7-32.
Peirson, John, Ian Skinner, and Roger Vickerman. Small, Kenneth A., and Kurt Van Dender. 2006. "Fuel
1995. "Estimating the External Costs of UK Efficiency and Motor Vehicle Travel: The Declining
Passenger Transport: The First Step towards an Rebound Effect." Energy Journal, 28(1): 25-52.
Efficient Transport Market." Environment and Small, Kenneth A., Clifford Winston, and Carol A.
Planning A, 27(12): 1977-93. Evans. 1989. Road Work: A New Highway Pricing
Poole, Robert W, Jr., and C. Kenneth OrskL 2003. and Investment Policy. Washington, D.C.: Brookings
"HOT Networks: A New Plan for Congestion Relief Institution.
and Better Transit." Reason Public Policy Institute Small, Kenneth A., and Jia Yan. 2001. "The Value of
Policy Study, no. 305. 'Value Pricing' of Roads: Second-Best Pricing and
Porter, Richard C. 1999. Economics at the Wheel: The Product Differentiation." Journal of Urban
Costs of Cars and Drivers. San Diego; London and Economics, 49(2): 310-36.
Toronto: Academic Press. Smith, V Kerry, and Ju-Chin Huang. 1995. "Can
Portney, Paul R., and John P. Weyant, eds. 1999. Markets Value Air Quality? A Meta-snalysis of
Discounting and Inter generational Equity. Hedonic Property Value Models." Journal of Political
Washington, D.C.: Resources for the Future. Economy, 103(1): 209-27.
Poterba, James M. 1989. "Lifetime Incidence and the Stern, Nicholas. 2006. Stern Review: The Economics of
Distributional Burden of Excise Taxes." American Climate Change. Cambridge: Cambridge University
Economic Review, 79(2): 325-30. Press.
Poterba, James. M. 1991. "Is the Gasoline Tax Toi, Richard S. J. 2005. "The Marginal Damage Costs
Regressive?" In Tax Policy and the Economy, Volume of Carbon Dioxide Emissions: An Assessment of the
5, ed. D. Bradford. Cambridge: MIT Press, 145-64. Uncertainties." Energy Policy, 33(16): 2064-74.
Quinet, Emile. 2004. "A Meta-analysis of Western Transportation Research Board. 2006. The Fuel Tax
European External Costs Estimates." Transportation and Alternatives for Transportation Funding.
Research: Part D: Transport and Environment, 9(6): Washington, D.C.: Transportation Research Board.
465-76. U.S. Congressional Budget Office. 1990. Federal
Rothengatter, Werner. 2000. "External Effects of Taxation of Tobacco, Alcoholic Beverages, and Motor
Transport." In Analytical Transport Economics: An Fuels. Washington, D.C.: U.S. Congressional Budget
International Perspective, ed. J. B. Polak and A. Office.
Heertje. Transport Economics, Management and U.S. Department of Commerce, Census Bureau. 2003.
Policy series. Cheltenham, U.K. and Northampton, Statistical Abstract of the United States: 2003.
Mass.: Elgar, 79-116. Washington, D.C.: U.S. Department of Commerce.
Rubin, Jonathan, Paul N. Leiby, and David L. Greene. U.S. Department of Energy. 1996. "Policies and
2006. "Tradable Fuel Economy Credits: Competition Measures for Reducing Energy Related Greenhouse
and Oligopoly." Unpublished. Gas Emissions: Lessons from Recent Literature."
Santos, Georgina, ed. 2004. Road Pricing: Theory and U.S. Department of Energy, Office of Policy and
Evidence. Research in Transportation Economics, vol. International Affairs Report, no. DOE/PO-0047.
9. Oxford; Amsterdam and San Diego: Elsevier, JAI. U.S. Department of Energy, Energy Information
Schimek, Paul. 1996. "Household Motor Vehicle Administration. 2006a. Annual Energy Outlook
Ownership and Use: How Much Does Residential 2006: With Projections to 2030. Washington, D.C.:

182.255.0.242 on Wed, 10 Nov 2021 04:05:08 UTC
U.S. Department of Energy. Pricing Schemes in the Monocentric City." Journal

U.S. Department of Energy, Energy Information of Urban Economics, 58(3): 367-88.
Administration. 2006b. Annual Energy Review 2005. Vickrey, William S. 1963. "Pricing in Urban and
Washington, D.C.: U.S. Department of Energy. Suburban Transport." American Economic Review,
U.S. Department of Transportation, Bureau of 53(2): 452-65.
Transportation Statistics. 2005. National Vickrey, William S. 1969. "Congestion Theory and
Transportation Statistics 2004. Washington, D.C.: Transport Investment." American Economic Review,
U.S. Government Printing Office. 59(2): 251-60.
U.S. Department of Transportation, Federal Highway Walls, Margaret A. 1996. "Valuing the Characteristics
Administration. 1997. 1997 Federal Highway Cost of Natural Gas Vehicles: An Implicit Markets
Allocation Study: Final Report. Washington, D.C.: Approach." Review of Economics and Statistics,
U.S. Department of Transportation. 78(2): 26f>-76.
U.S. Department of Transportation, Federal Highway Walters, Alan A. 1961. "The Theory and Measurement
Administration. 2000. Addendum to the 1997 of Private and Social Costs of Highway Congestion."
Federal Highway Cost Allocation Study: Final Econometrica, 29(4): 676-99.
Report. Washington, D.C.: U.S. Department of Wardman, Mark. 2001. "A Review of British Evidence
Transportation. on Time and Service Quality Valuations."
U.S. Department of Transportation, National Highway Transportation Research: Part E: Logistics and
Traffic Safety Administration. 2002. The Economic Transportation Review, 37(2-3): 107-28.
Impact of Motor Vehicle Crashes 2000. Washington, West, Sarah E., and Roberton C. Williams III. 2005.
D.C.: U.S. Department of Transportation. "The Cost of Reducing Gasoline Consumption."
U.S. Environmental Protection Agency. 1997. The American Economic Review, 95(2): 294-99.
Renefits and Costs of the Clean Air Act, 1970 to West, Sarah E., and Roberton C. Williams III. 2007.
1990. Washington, D.C.: U.S. Environmental "Optimal Taxation and Cross-Price Effects on Labor
Protection Agency. Supply: Estimates of the Optimal Gas Tax." Journal
U.S. Environmental Protection Agency. 1999. The of Public Economics, 91(3-4): 593-617.
Renefits and Costs of the Clean Air Act, 1990 to White, Michelle J. 2004. "The 'Arms Race' on
2010. Washington, D.C.: U.S. Environmental American Roads: The Effect of Sport Utility Vehicles
Protection Agency. and Pickup Trucks on Traffic Safety." Journal of Law
U.S. Environmental Protection Agency. 2004. Air and Economics, 47(2): 333-55.
Emissions Trends?Continued Progress through 2004. Yang, Hai, and Hai-Jun Huang. 1998. "Principles of
Washington, D.C.: U.S. Environmental Protection Marginal-Cost Pricing: How Does It Work in a
Agency. General Road Network?" Transport Research: Part
Verhoef, Erik T. 2005. "Second-Best Congestion A: Policy and Practice, 32(1): 45-54.

182.255.0.242 on Wed, 10 Nov 2021 04:05:08 UTC

Automobile Externalities and Policies

Uploaded by

Copyright:

Available Formats

Automobile Externalities and Policies

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Automobile Externalities and Policies

Uploaded by

Copyright:

Available Formats

Automobile Externalities and Policies

Author(s): Ian W. H. Parry, Margaret Walls and Winston Harrington

Stable URL: https://www.jstor.org/stable/27646797

This content downloaded from