Case study 12.

3: Fuel taxes and optimality

How much petrol should be taxed?
The tax on petrol varies widely around the developed world. Americas gasoline tax is
currently about 40 cents an American gallon, euivalent to ! pence a litre. "any Americans
are calling #or it to be cut, as the summer increase in prices begins to ma$e itsel# #elt, and
re#lecting a more general alarm about the countrys %energy crisis. &n 'anada the tax is hal#
as big again as in America( in Australia it is more than double. &n )apan and most o# *urope,
the speci#ic tax on petrol is around #ive times higher than in America, standing at the
euivalent o# some +, pence a litre. At the upper extreme is -ritain, where #uel duty .paid in
addition to value/added tax0 has risen in recent years to a punitive rate o# 1ust under ,0 pence
a litre, seven times the American levy.
2ou would expect well/designed petrol taxes to vary #rom country to country, according to
national circumstances 3 but not, on the #ace o# it, by a #actor o# seven. &n America it is ta$en
#or granted that *uropes petrol taxes, let alone -ritains, are insanely high, and presumably
something to do with socialism. &n -ritain, on the other hand, it is ta$en #or granted that
Americas gas tax is insanely low, part o# a broader scheme to wrec$ the planet. 4rotests in
-ritain last year showed that petrol tax had #inally been raised all the way up to its political
ceiling 3 but nobody expects or even calls #or the tax to be cut to the American level.
America and -ritain may both be wrong about the gas tax, but it seems unli$ely that they can
both be right. 5o how heavily should petrol be taxed? A paper by &an 4arry o# 6esources #or
the 7uture, an environmental thin$/tan$ in 8ashington, 9', loo$s at the arguments.
The most plausible 1usti#ication #or taxing petrol more highly than other goods is that using
the stu## harms the environment and adds to the costs o# tra##ic congestion. This is indeed
how -ritains government de#ends its policy. -ut the #act that burning petrol creates these
%negative externalities does not imply, as many seem to thin$, that no tax on petrol could
ever be too high. *conomics is precise about the tax that should, in principle, be set to deal
with negative externalities: the tax on a litre o# #uel should be eual to the harm caused by
using a litre o# #uel. &# the tax is more than that, its costs .which include the inconvenience
in#licted on people who would rather have used their cars0 will exceed its bene#its .including
any reduction in congestion and pollution0.
The pollution costs o# using petrol are o# two main $inds: damage to health #rom breathing in
emissions such as carbon monoxide and assorted particulates, and broader damage to the
environment through the contribution that burning petrol ma$es to global warming.
6eviewing the literature, "r 4arry notes that most recent studies estimate the health costs o#
burning petrol at around ;0 pence a litre or less. The harm caused by petrols contribution to
global warming is, #or the time being, much more speculative. 6ecent high/damage scenarios,
however, put an upper limit on the cost at about <;00 per ton o# carbon, euivalent to , pence
1
a litre o# petrol. Adding these together, you come to an optimal petrol tax o# no more than ;,
pence a litre.
High petrol taxes also help to reduce tra##ic congestion. However, they are badly designed #or
that purpose. 'urbing the number o# car 1ourneys is only one way to reduce congestion.
=thers include persuading people either to drive outside pea$ hours or to use routes that carry
less tra##ic. High petrol taxes #ail to exploit those additional channels. As a result, "r 4arry
#inds, the net bene#its o# a road/speci#ic pea$/period #ee .the gain o# less congestion minus
the cost o# disrupted travel0 would be about three times bigger than a petrol/tax increase
calculated to curb congestion by the same amount. 5till, i# politics or technology rules out
congestion/based roadpricing, a second/best case can be made #or raising the petrol tax
instead. According to "r 4arry, congestion costs in -ritain might then 1usti#y an additional
;0 pence a litre in tax.
This brings you to a total petrol tax o# around >, pence a litre. The pre/tax price o# petrol is
currently about >0 pence a litre, so this upper/bound estimate o# the optimal tax represents a
tax rate o# well over ;00? 3 a %high tax, to be sure. 2et -ritains current rate is roughly
double this. =n the same basis, o# course, Americas rate is #ar too low .even a lower bound
#or the optimal rate would be a lot higher than ! pence a litre0.
-ritains rate, 1udged according to the environmental and congestion arguments, loo$s way
too high 3 but plainly the -ritish government has another reason #or taxing petrol so heavily.
&t needs the money to #inance its plans #or public spending. 4olitically, raising money through
the tax on petrol, protests notwithstanding, has proved #ar easier than it would have been to
collect the cash through increases in income tax or in the broadly based value added tax 3 or,
#or that matter, through congestion based road/pricing .always dismissed as %politically
&mpossible0.
This seems odd. 5upposing that actual and pro1ected public spending 1usti#ied higher
taxation, "r 4arrys analysis strongly suggests that the country would have been better o##
paying #or it through income taxes than through a punitive petrol tax. And the petrol tax is not
only waste#ul in economic terms, i# "r 4arry is right( it is also regressive in its distributional
e##ects, increasing the cost o# living #or poor car/owning households much more than #or their
richer counterparts.
At last, -ritain has #ound the political ceiling #or the petrol tax. 8hat is remar$able is 1ust
how high it proved to be.
Summary:
The tax on petrol varies widely around the developed world. Americas gasoline tax is
currently about 40 cents an American gallon, euivalent to ! pence a litre. &n 'anada the tax
is hal# as big again as in America( in Australia it is more than double. &n )apan and most o#
*urope, the speci#ic tax on petrol is around #ive times higher than in America, standing at the
2
euivalent o# some +, pence a litre. At the upper extreme is -ritain, where #uel duty .paid in
addition to value/added tax0 has risen in recent years to a punitive rate o# 1ust under ,0 pence
a litre, seven times the American levy. &n America it is ta$en #or granted that *uropes petrol
taxes, let alone -ritains, are insanely high, and presumably something to do with socialism.
&n -ritain, on the other hand, it is ta$en #or granted that Americas gas tax is insanely low,
part o# a broader scheme to wrec$ the planet. A paper by &an 4arry o# 6esources #or the
7uture, an environmental thin$/tan$ in 8ashington, 9', loo$s at the arguments. The most
plausible 1usti#ication #or taxing petrol more highly than other goods is that using the stu##
harms the environment and adds to the costs o# tra##ic congestion. -ut the #act that burning
petrol creates these %negative externalities does not imply, as many seem to thin$, that no tax
on petrol could ever be too high. *conomics is precise about the tax that should, in principle,
be set to deal with negative externalities: the tax on a litre o# #uel should be eual to the harm
caused by using a litre o# #uel. &# the tax is more than that, its costs .which include the
inconvenience in#licted on people who would rather have used their cars0 will exceed its
bene#its .including any reduction in congestion and pollution0. 6ecent studies estimate the
health costs o# burning petrol at around ;0 pence a litre or less. The harm caused by petrols
contribution to global warming is, #or the time being, much more speculative. 6ecent high/
damage scenarios, however, put an upper limit on the cost at about <;00 per ton o# carbon,
euivalent to , pence a litre o# petrol. Adding these together, we come to an optimal petrol
tax o# no more than ;, pence a litre. "r 4arry #inds, the net bene#its o# a road/speci#ic pea$/
period #ee .the gain o# less congestion minus the cost o# disrupted travel0 would be about
three times bigger than a petrol/tax increase calculated to curb congestion by the same
amount. The total petrol tax should be around >, pence a litre. The pre/tax price o# petrol is
currently about >0 pence a litre, so this upper/bound estimate o# the optimal tax represents a
tax rate o# well over ;00? 3 a %high tax, to be sure. 2et -ritains current rate is roughly
double this. =n the same basis, o# course, Americas rate is #ar too low. &t needs the money to
#inance its plans #or public spending. 4olitically, raising money through the tax on petrol,
protests notwithstanding, has proved #ar easier than it would have been to collect the cash
through increases in income tax or in the broadly based value added tax. "r 4arrys analysis
strongly suggests that the country would have been better o## paying #or it through income
taxes than through a punitive petrol tax. The petrol tax is not only waste#ul in economic
terms, i# "r 4arry is right( it is also regressive in its distributional e##ects, increasing the cost
o# living #or poor car/owning households much more than #or their richer counterparts.
Question 1: What are the economic reasons for fuel taxes being different in different
countries
3
!ns"er: The economic reasons #or which #uel taxes are di##erent in di##erent countries are:
;. 5ubsidies
>. *xchange rate
+. Availability
4. 4rices o# substitutes, li$e gas price
,. To #inance public spending and construction pro1ects
@. 'arbon tax
!. Tax o# state and central government
Question 2: What additional factors are rele#ant in explaining "hy fuel taxes in the $%
are se#en times the le#el in the $S!
!ns"er: The additional relevant #actors which are relevant #or AB are:
The most plausible 1usti#ication #or taxing petrol more highly than other goods is that
using the stu## harms the environment and adds to the costs o# tra##ic congestion. This
is indeed how -ritains government de#ends its policy.
The pollution costs o# using petrol are o# two main $inds: damage to health #rom
breathing in emissions such as carbon monoxide and assorted particulates, and
broader damage to the environment through the contribution that burning petrol
ma$es to global warming.
High petrol taxes also help to reduce tra##ic congestion.
Covernment needs the money to #inance its plans #or public spending.
Question 3: Why are fuel&taxes an inefficient "ay of reducing traffic congestion
!ns"er: 'urbing the number o# car 1ourneys is only one way to reduce congestion. =thers
include persuading people either to drive outside pea$ hours or to use routes that carry less
tra##ic. High petrol taxes #ail to exploit those additional channels. The net bene#its o# a road/
speci#ic pea$/period #ee .the gain o# less congestion minus the cost o# disrupted travel0 would
be about three times bigger than a petrol/tax increase calculated to curb congestion by the
same amount.
A higher #uel taxes or subsidiDing transit #ares as a short/run measure to mitigate congestion,
though it is recogniDed that these pricing schemes are ine##icient than a congestion #ee. 7uel
taxes raise the cost o# all driving and there#ore do not induce the e##icient substitution o##
congested roads onto other transport modes, or travel at o##/pea$ periods. 5imilarly, transit
4
#are subsidies only a##ect the price o# driving on congested roads relative to public transit, but
not the 4rices uali#ied to driving on other roads or at o##/pea$ hours.
Question: '. (i#en that fuel taxes are higher in the $% than the rest of )urope* "hat
implications does this ha#e for $% firms competing "ith )uropean +nes
!ns"er: The implications do this have #or AB #irms competing with *uropean ones are
given below:
;. Creater uality. AB 7irms are encouraged not 1ust to compete on price but also in terms o#
uality. They give uality products compare with *uropean ones.
>. "ore choice: AB providers o##er di##erent types o# service #or their customer. As a result
their customer will satis#y.
+. AB government imposed tax on *uropean products. As a result, *uropean product price is
higher in AB mar$et and AB #irms competing with *uropean =nes.
5