ArticlePDF Available

Abstract and Figures

This paper surveys the literature on, and examples of current implementation of, carbon taxes and carbon emission permits. It sets out the theoretical basis for these instruments, with special reference to the revenue-recycling and tax interaction effects. This theoretical work concludes that instruments which raise revenue which can be recycled so as to reduce pre-existing distortionary taxes are significantly less costly than those which do not. The paper then reviews the sizable literature on the distributional effects of these instruments, especially with regard to industrial competitiveness and regressive effects on low-income groups, evaluating attempts to mitigate these where they are perceived as unacceptable. The paper concludes that such efforts at mitigation, while possible, can substantially reduce the efficiency benefits of the instruments. The projected costs of carbon taxes depend on a wide range of assumptions. This is still a contested area, but the paper concludes that, on a range of plausible assumptions, these costs need not be high. Finally the paper notes that early evaluations of the environmental effectiveness of carbon taxes have been generally positive. This suggests that, if concern about anthropogenic climate change continues to increase, more countries will introduce carbon taxes and emission permits, with the latter increasingly auctioned.
Content may be subject to copyright.
Paul Ekins
SPIRE, Keele University
Terry Barker
DAE, University of Cambridge
Abstract. This paper surveys the literature on, and examples of current
implementation of, carbon taxes and carbon emission permits. It sets out the
theoretical basis for these instruments, with special reference to the revenue-
recycling and tax interaction effects. This theoretical work concludes that
instruments which raise revenue which can be recycled so as to reduce pre-
existing distortionary taxes are significantly less costly than those which do
not. The paper then reviews the sizable literature on the distributional effects of
these instruments, especially with regard to industrial competitiveness and
regressive effects on low-income groups, evaluating attempts to mitigate these
where they are perceived as unacceptable. The paper concludes that such
efforts at mitigation, while possible, can substantially reduce the efficiency
benefits of the instruments. The projected costs of carbon taxes depend on a
wide range of assumptions. This is still a contested area, but the paper
concludes that, on a range of plausible assumptions, these costs need not be
high. Finally the paper notes that early evaluations of the environmental
effectiveness of carbon taxes have been generally positive. This suggests that, if
concern about anthropogenic climate change continues to increase, more
countries will introduce carbon taxes and emission permits, with the latter
increasingly auctioned.
Keywords. Carbon taxes; Carbon trading; Double dividend
1. Introduction and policy context
The Intergovernmental Panel on Climate Change (IPCC) concludes in its Second
Assessment Report that `the balance of evidence suggests that there is a
discernible human influence on global climate' (Houghton et al., 1996, p. 5), due
to the emission of a number of `greenhouse gases' into the atmosphere, the most
important of which is carbon dioxide (CO2). The IPCC Third Assessment Report
(2001) concludes that `There is new and stronger evidence that most of the
warming observed over the last 50 years is attributable to human activities'. While
the impacts of this influence are still uncertain, they include sea level rise, an
0950-0804/01/03 0325± 52 JOURNAL OF ECONOMIC SURVEYS Vol. 15, No. 3
#Blackwell Publishers Ltd. 2001, 108 Cowley Rd., Oxford OX4 1JF, UK and 350 Main St., Malden,
MA 02148, USA.
increase in precipitation intensity, more severe droughts and=or floods in some
places, and less severe droughts and=or floods in others, `a possibility for more
extreme rainfall events', and shifts in the competitive balance of species which may
include forest die-back (ibid., pp. 6±7). The UK Royal Commission on
Environmental Pollution has described climate change as an `enormous
challenge ... threatening generations to come' (RCEP, 2000, p. 9).
182 governments have now signed the UN Framework Convention on Climate
Change (UNFCCC), in order to `prevent dangerous anthropogenic interference
with the climate system' (UNFCCC, 1992, Article 2). In 1997 the Kyoto Protocol
to the Convention was signed, and mandated that industrial countries should,
overall, reduce emissions by about 5% from 1990 levels by the period 2008 ±2012.
Countries accepted different targets of emission reduction such that the overall
emission reduction would be reached.
While the Kyoto Protocol still has not been ratified by a sufficient number of
countries to enter into force, with the new US President recently declaring his
opposition to it, and while some of its key provisions and mechanisms still remain
to be adequately defined, many countries are preparing, or have prepared, climate
change programmes in order to meet their targets.
An important instrument in many of the programmes is the taxation of energy
or, in some cases, the taxation of energy according to its carbon content, which is
known as a carbon tax. In addition, a number of countries are preparing schemes
for the trading of carbon emission permits. The Kyoto Protocol envisages that
such a scheme will in due course be introduced at the international level in order
to help countries meet their carbon reduction commitments at least cost.
However, proposals to introduce carbon or energy taxes have met with
substantial resistance. As will be seen, carbon taxes have so far only been
introduced in a few European countries, at relatively low levels and with
exemptions for many energy-intensive industries. There have been major political
campaigns against such taxes, especially in Europe when the European
Commission was considering a carbon=energy tax in 1993 and 1994 and in the
US preceding the Kyoto meeting in December 1997. One effect of the
politicization of such taxes has been that governments and interest groups have
commissioned research into particular aspects of carbon taxation. Indeed, much
of the research reported in the literature has been funded by governments and by
the industries likely to be the most affected by the tax, for example the coal and
electricity industries. A result of this controversy, therefore, is that the possible
impacts of carbon taxation on carbon emissions (and on the economic activities
that are responsible for them) have been intensively studied for individual
economic sectors and for countries as a whole.
This article first briefly reviews the theories underpinning carbon taxes and
carbon emissions trading (Section 2). It then sets out the practical experience
with these instruments to date (Section 3). Section 4 discusses the environmental
impacts of these instruments which have been estimated or projected, and
Section 5 explores their economic and distributional impacts. Section 6
#Blackwell Publishers Ltd. 2001
2. Carbon taxes and carbon emissions trading: theoretical considerations
The objective of a carbon tax is to slow global warming, usually in the context of
agreed targets for limiting or reducing GHG emissions, taking into account the
scientific evidence, the risks of inaction and the possibility that some carbon
reduction may be achieved at no net cost (policies to achieve which are sometimes
referred to as `no-regrets' policies). Since emissions are expected to be on an
indefinite rising trend, any tax would also have to go on rising in order to achieve
long-term stabilization and reduction. In this context, a carbon tax appears to
have all the hallmarks of good taxation:
*It tackles an accepted economic problem (a damaging externality agreed as
such by governments) helping to bring the private costs of emitting CO2into
line with social costs of global warming.
*Its revenues can be expected to grow with income because energy demand
tends to rise with income, and it is not easy to substitute away from fossil
fuels in energy supply.
*It should be simple and cheap to administer through use of many existing tax
structures for excise duties.
*It is expected to stimulate energy saving, innovation and investment in clean
technology, and hence, possibly, economic growth.
*Any regressive side effects (those that disproportionately affect low-income or
otherwise disadvantaged groups) are likely to be small enough for compensa-
tion to be able to remedy them using a small fraction of the expected revenues.
Both a carbon tax and carbon emissions trading are market-based instruments
that depend fundamentally on the efficient working of the market system for their
success. This efficiency has many requirements and implications. First, the legal
and institutional structure needs to ensure that contracts are
*freely entered into by the relevant parties and
*enforceable under clear and widely accepted laws and rules.
Thus countries beset by bribery and corruption may not be able to use taxation
because the taxes will be evaded or become an excuse for further corruption.
Second, prices should reflect costs to some degree, so that the carbon tax will
increase the price of carbon-intensive production. Third, buyers and sellers should
be well informed as to the costs and availability of alternatives, such that future
outcomes (even if not known) should at least be considered. In some cases,
especially amongst some socially disadvantaged groups such as the elderly, there
may be an unwillingness to consider alternatives, so extra taxation may of itself
(and before possible compensation measures are taken into account) have very
inequitable effects.
These are some of the issues to be explored in later sections. First, however, it is
worth setting out the basic neo-classical theory of environmental, including
carbon, taxation.
#Blackwell Publishers Ltd. 2001
2.1. Basic theory of environmental taxes and emissions trading
Environmental taxes
The theoretical basis for environmental taxes in general, and carbon taxes in
particular, is well developed and needs only to be briefly rehearsed here. Thus it is
generally agreed among economists that, in a situation where the production or
consumption of some good results in a negative external effect (i.e. one that is not
reflected in the price of the good in question), then social welfare can be improved
by imposing a tax on the good.
There are two basic ways of identifying the level at which the tax should be set.
The neo-classical, optimization approach was conceived by Pigou (1932) and
formalized by Baumol (1972) (one of many formalizations). The approach seeks
to calculate a damage function for different rates of emission of the pollutant, and
then seeks to equate the marginal net private benefit (MNPB) of the activity
causing the pollution (P) with the marginal external cost (MEC) to which it gives
rise. The equality is achieved by imposing a tax equal to the difference between
them at the optimal emission level. Figure 2.1 (which is stylised in the sense that it
assumes that the curves can be measured and remain stable as prices and outputs
change, which may not in practice be the case) sets out this basic theoretical
position, in which Qis the no tax pollution level, and Q*is the optimal pollution
level. The optimal tax is then t*. The theory of environmental taxation has been
much developed to take into account market situations other than perfect
competition and other considerations (see, for example, Baumol and Oates, 1988),
but this is outside the scope of this paper.
While Baumol proved the theoretical validity of the Pigouvian optimising
approach to environmental taxation, he despaired of the prospects of making it
operational, both because of the difficulty of calculating the marginal damage
function and because the presence of multiple local maxima seemed to rule out an
iterative approach to the optimal position, concluding: `All in all we are left with
little confidence in the applicability of the Pigouvian approach ... We do not how
to calculate the required taxes and subsidies and we do not know how to
approximate them by trial and error.' (Baumol, 1972, p. 318).
With regard to carbon taxes, setting aside the issue of ancillary local
environmental benefits, Baumol's pessimism about multiple maxima is un-
Marginal cost/
benefit MNPB MEC
Q* Q Pollution
Figure 2.1 The optimising approach to environmental taxation.
#Blackwell Publishers Ltd. 2001
founded, because carbon dioxide is a global pollutant: the location of its emission
is irrelevant to its impact and there will be no local maxima. However, the
problem remains with the calculation of the damage costs related to global
warming, given the huge uncertainties about the impacts, and the contested nature
of some of the methodologies which are used to value them, especially the non-
market impacts such as effects on human health, risk of human mortality and
damage to ecosystems. The Policymakers' Summary of IPCC's Working Group
III's contribution to the Second Assessment Report, on the Economic and Social
Dimensions of Climate Change, admits of its attempts to assess the social costs of
climate change that `The literature on the subject is controversial ... There is no
consensus about how to value statistical lives or how to aggregate statistical lives
across countries. ... The estimates of non-market damages are highly speculative
and not comprehensive.' (Bruce et al., 1996, pp. 9 ±10).
It is to address such considerations that Baumol and Oates (1971) proposed an
alternative approach to implementing environmental taxes, which has come to be
called the standards and pricing approach after their article. The approach
involves choosing environmental standards on the basis of their desired effects on,
for example, human health, or on quality of life more generally, and then using
environmental taxes on an iterative basis to bring levels of environmental damage
down to the standards. Baumol and Oates showed that such environmental
taxation had the property that it would achieve the desired environmental
improvement at minimum cost to society at large. This has now become the
principal approach to and justification of environmental, including carbon, taxes.
Certainly all the carbon taxes that have been implemented to date (see Section 3)
have been put in place in order to contribute to defined programmes of CO2
emission reduction, rather than on the basis of any optimality calculations.
Emissions trading
As noted above, the property of carbon taxes that particularly commends itself
both to economists and policy makers on climate change is its ability, in theory, to
achieve a given level of emission reduction at least cost to society overall. This
comes about because the incentive effects of the tax act to equalise the marginal
abatement cost across all emitters.
Precisely the same effect can be achieved in principle through a carbon
emissions trading system, whereby emitters have access to emission permits which
they can trade among themselves. Experience with emissions trading has until now
been largely limited to the US, where sulphur dioxide emissions trading is
generally regarded as successful (see Sorrell and Skea, 1999). Yet emissions
trading should be even more suited to CO2, as a uniformly mixed pollutant. It is
therefore not surprising that a number of carbon emission trading schemes are
now being set up or proposed in different countries (see Section 3).
It has been shown, under a precise set of restrictive assumptions, that there is
broad equivalence between an emissions trading scheme, where emission permits
are auctioned by the government, and levying a carbon tax at the auction price
#Blackwell Publishers Ltd. 2001
(Pezzey, 1992a; Farrow, 1995). The main difference between taxation and trading
concerns price=quantity adjustment. With a carbon tax, it is the tax on and hence
an increase in the price of carbon that is fixed, and the quantity of carbon emitted
as CO2that adjusts. With emissions trading, it is the quantity of carbon emitted as
CO2that is fixed, and the price of the emission permits that adjusts. Weitzman
(1974) has shown that (1) it is preferable to fix the price when there is uncertainty
over the control cost function, and a possibility that it is very sensitive to greater
than optimal carbon emissions reduction, and (2) it is preferable to fix the
quantity when there is uncertainty about the damage function, and a possibility
that it may be very sensitive to greater than optimal emissions. Using this insight
Pizer (1999) has argued that it would be preferable to control carbon emissions
using a price, rather than a quantity, instrument, in contrast to the provisions of
the Kyoto Protocol, which are for quantity control. Pizer suggests that the
problems in negotiating the details of the Protocol derive from the potentially high
costs which carbon limits may entail. These costs could themselves be limited by
specifying a `trigger price' for extra emission permits, which would effectively set a
maximum cost of abatement (though obviously emissions could then increase).
The proposal well illustrates the relation and interaction between prices and
quantity limits.
A key decision in emissions trading schemes is how to allocate emission permits.
The two most discussed options are their auction by the government (when the
scheme resembles or is equivalent to a carbon tax set at the level of the auction
price, depending on the auction rules), and their free allocation on the basis of
some formula related to current emissions (an allocation mechanism called
`grandfathering'). The two methods of allocation provide the same incentives for
emissions reduction at the margin, and, of course, yield the same environmental
outcome (because the quantity of permits is unchanged). The difference between
them is distributional. Auctions (and carbon taxes) transfer resources from
emitters to the government and therefore yield government revenue. Grand-
fathering of emission permits appears to give assets in the form of tradable
property rights to polluters. In any scheme, a proportion of the permits can be
auctioned and the rest allocated free of charge: this flexibility gives permit schemes
an advantage over corresponding carbon taxation where, conventionally, all
revenues are received by governments.
Cramton and Kerr (1998) argue strongly that emission permits should be
auctioned by the government rather than grandfathered, on three main grounds:
governments can use the revenues to reduce distortionary taxes and therefore
increase economic efficiency (this argument is discussed in a wider context below);
auctions spread both the costs of carbon control, and the gain of the permit
allocation, more equitably through the economy; and auctions remove the need
for difficult and inevitably contested decisions over allocation, and give fair access
to emission permits to small producers and new entrants. Against these points is
the argument that grandfathering better reflects the implicit social contract with
current producers, on the basis of which they undertook their investments, that
their use of the atmosphere as a carbon sink would be free.
#Blackwell Publishers Ltd. 2001
Given the political power of current producers, it is the latter argument which
has prevailed in practice, and all proposed carbon emissions trading systems (see
Section 3) envisage the grandfathering of permits, initially at least. Over time it
may be that an increasing proportion will come to be auctioned.
It is sometimes argued that a permit system is to be preferred to a tax because
its outcome is more certain in terms of achieving an emission reduction target.
This extra certainty is dependent on the permit regime having clear, enforceable
and enforced rules and well-monitored emissions. However, there is no real
dichotomy between taxes and permits. An effective scheme to abate carbon
emissions would be likely to combine taxes, permits and regulation, with each
instrument supporting the others depending on the country, sector and institution
concerned. Since carbon abatement does not require a precise abatement by a
definite date, the aim of policy must be to achieve significant reductions over a
number of years, with policies adjusting to outcomes repeatedly.
2.2. The costs of carbon emissions reduction
2.2.1. The costs of introducing a carbon tax
Economists are generally agreed that, because of the efficiency properties of
carbon taxes (and systems of emissions trading using auctioned permits), they will
achieve any given level of carbon emissions reduction at a lower cost than
regulation. There have been a number of studies which calculate the efficiency
advantages of economic instruments over other forms of environmental
regulation. OECD 1997 (Table 2, p. 30) cites a range of studies which show that
the costs for regulatory instruments of air-pollution control compared to those
arising from least-cost instruments can vary by a factor of 1.07 to 22.00. In
addition, applying an economic instrument across countries can achieve a given
emission reduction at lower cost than applying separate economic instruments
within countries, because of the transnational equalisation of the costs of
abatement which the common instrument achieves. Thus Bohm (1999, p. 316)
finds that a jointly implemented carbon emissions trading system applied across
the four Scandinavian countries (Denmark, Finland, Norway, Sweden) saves 48%
of the costs of a given level of carbon reduction, even when each country operates
its own carbon trading system internally. The reduction in carbon tax rates
brought about by coordinating a tax across 11 EU countries has also been
studied. Conrad and Schmidt (1998), using a general equilibrium model, GEM-
E3, estimate that the coordinated tax rate falls by about 1.5% compared with
uncoordinated carbon taxes achieving a 10% reduction in CO2emission below
baseline levels; and Barker (1999) using a disequilibrium econometric model,
E3ME, estimates the fall as 5.2%.
One measure of the cost of reducing carbon emissions is the value of the
reduction in output to which it gives rise. In the first instance, this will depend on
the opportunities for and economics of abatement. In the extreme, where there are
no opportunities to reduce the emissions associated with the output of a product,
#Blackwell Publishers Ltd. 2001
then they can only be reduced by cutting the output of the product itself. This
situation gives rise to the largest cost of emissions reduction.
Although there are fewer opportunities for end-of-pipe reductions of carbon
emissions than for some other gases, there will typically be a number of ways of
cutting the emissions associated with production, which are cheaper than simply
cutting production, and which incur different costs for different levels of emission
reduction. Thus the cost of a given level of carbon reduction will, in the first
instance, be a combination of the costs of abatement technologies and the value of
foregone output. However, in a general equilibrium framework, several other
effects can be identified which make the overall cost of reduction much more
difficult to compute.
1. The reduction in carbon emissions may be associated with the reductions of
other polluting emissions (and will be if it involves reduced use of fossil fuels,
which are associated with a number of polluting emissions apart from
carbon). These reductions will yield what are called `secondary' or `ancillary'
benefits of carbon reduction, which should be taken into account when
computing its overall economic cost.
2. The output of pollution abatement companies will have increased, which will
stimulate the economy.
3. The price of the abating firm's output may have increased, with further
knock-on effects through the economy.
4. In response to any increase in the price of the abating firm's output,
substitution towards other products will take place.
5. It may be that the productivity of people or other firms will have increased
due to the reduction in pollution.
6. On the one hand, the rise in price of the use of the environmental resource
may hasten the depreciation of the capital equipment affected. On the other
hand, it may be that the shift in relative prices will stimulate research,
innovation and investment with a view to economising on the resource,
which may be economically beneficial.
7. The tax will yield revenue, which will allow other taxes to be reduced, for the
same level of government expenditure, with yet more knock-on effects
through the economy. Where the revenues generate economic benefits by
permitting distortionary taxes to be reduced, then this benefit, together with
the environmental benefit of the tax, is known as a `double dividend' from
the tax.1
A number of these points will be discussed further in subsequent sections, together
with the numerous results of modelling which have sought to estimate their
quantitative importance. The last point, however, merits further theoretical
As regards the double dividend debate, the first issue to be addressed is that of
the economic distortions introduced by taxation. It is a standard result from the
neo-classical optimal taxation literature (e.g. Diamond and Mirrlees, 1971; Stiglitz
and Dasgupta, 1971) that taxes on economic output or factor inputs are
#Blackwell Publishers Ltd. 2001
distortionary, thereby reducing output and welfare. According to the static perfect
competition model, distortions arise from the taxes changing the prices facing
both producers and consumers, so that factors no longer receive their marginal
product and the cost of output does not reflect its true economic cost. Some
analysts apply such insights to environmental taxes. Thus Gaskins and Weyant
(1993, p. 320) write of `the distortions to the economy caused by the imposition of
the carbon tax', and Jorgenson and Wilcoxen (1993a, p. 518) write of the
`introduction of distortions resulting from fossil-fuel taxes'. But the whole point of
an environmental tax (at a rate at or below the optimal level) is that it wholly or
partially corrects a distortion from a pre-existing environmental externality. The
adjustments to the new relative prices, and the resulting shift in resource
allocation, are not to be regarded as a `distortion' due to the tax. Adjustments in
the economy to a higher level of allocative efficiency are the effects of removing
distortions by the internalization of an environmental externality. As Pearce
(1991, p. 940) has emphasised: `While most taxes distort incentives, an
environmental tax corrects a distortion, namely the externalities arising from
the excessive use of environmental services'.
However, given that the environmental improvement can be envisaged as being
brought about through the diversion of economic resources from producing
marketed goods or services to the production of unmarketed environmental goods
or services, a loss of marketed output is to be expected from such improvement
(unless the instrument of improvement permits the reduction of other, pre-
existing, economic inefficiencies, as discussed below). Moreover, with regard
purely to marketed output (i.e. setting aside the environmental benefits, assuming
no correction of an externality), the introduction of a carbon tax will be
distortionary, as will any other tax on factor inputs or economic products. In the
modelling of an environmental tax, if the base run is considered non-distortionary
and at full equilibrium in all markets, then, as Boero et al. (1991, p. 34) note, `any
deviation from a ``no distortions'' base run necessarily involves economic costs'.
However, no economy is at a point of non-distortionary equilibrium. There are
distortions due to current taxation patterns, which bear most heavily on labour;
and there are distortions due to market or government failure, such as, perhaps, in
the market for energy efficiency, or as a result of inefficient government
regulation. In addition, economies are not likely to be in equilibrium in all
markets: there may be substantial involuntary unemployment in the labour
markets; there may be chronic deficits or surpluses in the balance of payments
implying disequilibrium in the foreign exchange market. Depending on the type of
model used (e.g. general equilibrium or otherwise), the macroeconomic effect of
an environmental tax will depend on whether its introduction, or other associated
policy, affects these distortions.
It has already been noted that carbon taxes will normally allow a given carbon
reduction to be achieved at a lower cost than pure reliance on regulations. It
remains to explore their implications for markets other than those of the directly
affected goods, for different income groups and for the wider tax system. These
implications are significant because the revenues from a carbon tax could be large.
#Blackwell Publishers Ltd. 2001
Thus the levels of the carbon taxes proposed as necessary to make substantial
reductions in carbon emissions are between $100 and $400 per ton of carbon (see
Table 5 from Boero et al., 1991, pp. 87 ± 89). A $250 tax is equivalent to $0.75 per
gallon on petrol or $30 per barrel on oil (Cline, 1992, p. 147). Cline (1992, p. 151)
also estimates that a $100 per ton global tax rate would raise on the order of $500
billion annually, and about $130 billion from the USA alone. Schelling considers
that `a carbon tax sufficient to make a big dent in the greenhouse problem would
have to be roughly equivalent to a dollar per gallon on motor fuel ... (which)
would currently yield close to half a trillion dollars a year in revenue.' (Schelling,
1992, p. 11). Barker and Rosendahl (2000, p. 433 and Table 5.3 below) estimate
that a carbon tax of 153 euros per ton of carbon will allow western Europe to meet
its Kyoto targets, with total revenues by 2010 of 170bn euros.
This revenue can be used in a number of ways:
1. To achieve further environmental benefits (for example, by subsidising
energy efficiency measures, or low-carbon technologies). This is discussed in
Section 4.
2. To achieve distributional objectives, either in response to the distributional
impacts of the carbon tax, or more generally. This is discussed in Section 5.
3. To reduce government borrowing or debt, thereby reducing the level of
taxation that will be required in the future.
4. To reduce other inefficiencies in the economy, by enabling, for a given level
of government expenditure, other distortionary taxes to be reduced. This has
been termed `revenue-recycling', and the `dividend' to which it may be able
to give rise is the subject of what follows.
2.2.2. The revenue-recycling dividend
A revenue-recycling dividend is here defined as an economic (and non-
environmental) benefit resulting from the revenue-neutral imposition of a tax
(i.e. all the revenue from the tax is returned to taxpayers by cuts in other taxes or
lump-sum rebates, rather than saved or spent by the government). Such a
dividend can arise if the revenue recycling improves economic distribution (see
Section 5), or reduces unemployment or otherwise increases economic efficiency
(thereby increasing output).
As noted above, the theoretical possibility for such a dividend depends on the
economy being in a non-optimal state to start with: the existing tax structure must
be non-optimal in some sense, for example because the tax base is related to
employment. Alternatively, there must be existing deficiencies in distribution and
market failures in the labour and other markets. Any perception or assumption
that the initial condition of the economy is characterized by no externalities,
perfectly competitive markets operating in equilibrium, and with taxes imposed
on a per capita basis, will a priori rule out the existence or possible achievement of
a revenue-recycling dividend which increases economic efficiency.
#Blackwell Publishers Ltd. 2001
However, with a less ideal initial economic configuration, the existence of such
a dividend cannot be ruled out. Figures from the US suggest that distortions from
taxation are substantial. Thus Ballard et al. calculate the marginal excess burden
(MEB) of taxation in the US to be in the range 17 to 56 cents per dollar of extra
revenue (Ballard et al., 1985, p. 128). Jorgenson and Yun (1990) find that the
MEB of the US tax system as a whole, even after the tax reform of 1986, which
was widely held to have reduced the excess burden, is 38 cents per dollar of
revenue raised. Some components of the tax system had far higher costs, for
example the MEB for individual capital taxes was 95c=$ (Jorgenson and Yun,
1990, p. 20). Jorgenson and Yun (1990, p. 6) acknowledge that their MEB
estimates `are considerably higher than previous estimates. This can be attributed
primarily to the greater precision we employ in representing the US tax structure'.
Nordhaus (1993, p. 316) notes that `some have estimated [the marginal
deadweight loss of taxes in the US] as high as $0.50 per $1.00 of revenue'.
While there are no comparable figures for Europe, EC 1994 (p. 145) makes the
Marginal costs of taxation increase more than proportionally with the level of
taxation. In view of the much higher share of tax revenues in the Community
than in the USA (the tax burden in the Community is nearly 50% higher than
in the USA and Japan) it would appear that the costs of fiscal systems in
terms of forgone GDP and hence employment might be particularly high in
the Community. Only if the structure of the Community's fiscal system were
much more efficient than in the USA, would this not hold true.
In the absence of grounds for believing European tax systems to be more efficient
than that of the US, it seems likely that the distortions from taxation in Europe
are at least as great as those in the USA.
The key question then becomes whether the substitution of a non-distortionary
environmental tax for a distortionary tax can reduce the distortions from the tax
system as a whole, leading to increased efficiency and output. The focus here will
be on substituting an environmental tax for a labour tax of two kinds: an income
tax on employees and a social security tax levied on employers (the latter
substitution is the one implemented in relation to the revenues from the UK's
Climate Change Levy, implemented in April 2001). Finally some brief
consideration will be given to the so-called erosion and interdependency (or tax
interaction) effects, which have been the object of some recent analysis in the
Substituting for a personal income tax
Two immediate first-round effects of substituting an environmental tax for a
personal income tax (here considering only labour income) will arise:
S1 The environmental tax will raise the price of the affected items; the
reduction in income tax will increase the disposable income of employees.
#Blackwell Publishers Ltd. 2001
Insofar as the higher prices are paid by non-employees (unemployed,
pensioners, other non-employed, foreigners), and income tax rebates are
received by employees, the substitution will raise the real wage. This will
increase the labour supply.
S2 The price increases will be concentrated in the goods or activities subject to
the tax. Where the tax falls on inputs, producers will tend to substitute away
from the taxed input. Where it falls on final demand, consumers will shift
demand away from the affected sectors to others that are relatively less
environment-intensive and so less affected by the tax. It is of course the
intention of the tax to bring about this substitution by both producers and
consumers, and it will occur irrespective of how the revenue from the tax is
Insofar as there is an inverse correlation between labour-intensity and
environment-intensity, the demand for labour-intensive goods and services
will increase. Barker (1994, pp. 20±21) has shown that such a correlation
exists for the production and industrial use of energy (energy-intensive
industries tend to use relatively little labour; labour-intensive industries
tend to use relatively little energy). A revenue-neutral energy tax would
therefore be expected to increase labour demand.
The shift in relative prices would also alter the productivity of the
affected capital stock and, perhaps, hasten its depreciation. This adjustment
effect could be minimised by introducing the taxation in a gradual, pre-
announced way, so that the new relative prices were anticipated and
allowed for in investment schedules.
The increase in labour demand as a result of S2 will either reduce unemployment
(if it exists) or put to work the increased labour supply induced as a result of S1.
Either way employment and output would increase, yielding both an employment
and an efficiency dividend. The increased employment and output would result in
second-round macroeconomic improvements (lower benefits, higher tax revenues
so lower tax rates etc.). The only way these positive benefits would fail to
materialize would be if S2's increase in labour demand was far stronger that S1's
increase in labour supply in the context of a tight labour market. The increased
labour demand would then engender wage inflation with generally negative effects
on employment and output.
Substituting for an employers' social security tax (SST)
Consider an average firm using energy E1and employing labour L1, which is
subject to a SST of tL. The firm pays tax of L1tL.
Now let a tax tEbe imposed on energy, with full compensation by way of
reduction in SST. Then, initially, the firm will pay energy tax of E1tE, with a
reduction in its SST rate to:
#Blackwell Publishers Ltd. 2001
Now, with energy relatively more expensive and labour relatively cheaper, the
firm changes its proportion of inputs, say to L2(>L1) and E2(<E1).
The tax paid on energy use is now tEE2. Let the new tax rate on labour, t0
set so that the total tax paid by the firm is unchanged, so:
The first term on the right hand side (RHS) of equation 2.1 can be thought of as
the result of the substitution effect. The firm's tax rate on labour will be reduced in
proportion to its increased employment. The second term is the revenue effect of
the energy tax, further reducing the effective tax rate on labour.
The reduction in the effective tax rate on labour will reduce the overall marginal
cost of labour to the firm, as long as the wage paid to labour is unchanged. If this
is so, then the greater is the fall from tLto t0
L, the greater is the fall in labour's
marginal cost to the firm. The reduced marginal cost of labour will increase labour
demand. The reduction in labour costs may also enable firms to cut the price of
their output, thereby compensating for any price increases due to the imposition
of the environmental tax.
As in the earlier case, if all the SST decrease is passed through to employees as
an increase in wages, then evidently the marginal cost of labour to employers will
not fall and so labour demand will not rise. Even so, because of the increased cost
of energy, a substitution effect still takes place, but less strongly. The situation
then becomes identical to where the recycling was achieved by reducing a personal
income tax, as in S1 and S2 above.
The effect S1 comes about as a result of a reduction of a distortion in the
labour market (employees' disposable income moves closer to their marginal
product as paid by their employers), through the simultaneous reduction or
removal of a distortion in the market of the taxed good (an externality has been
wholly or partially internalised). However, the possibility has been addressed in
a number of papers that, through interdependencies in the tax system, or the
erosion of the environmental tax base, the reduction of the environmental
distortion in the context of a labour market with pre-existing labour taxes could
lead to more labour market distortion rather than less. This possibility is
considered below.
This discussion of the possible employment effects of introducing a revenue-
neutral carbon tax can be summarized as follows. The key question determining
the extent to which an efficiency dividend would arise is whether or not the
reduction of social security taxes results in higher wages. If it does, then the
implications for employment are ambiguous: being negative because of inflation
and macroeconomic deterioration, and being positive through the substitution
effects in both production and consumption. If it does not, then employment
would unambiguously increase, through both the substitution effects and firms'
falling labour costs. Although the higher employment would probably mean that
#Blackwell Publishers Ltd. 2001
overall labour productivity in the economy was lower, it is also likely that output
would increase somewhat.
2.3. Distortions of, and tax interaction effects between, environmental and labour
It was noted above that, unlike other taxes which tend to introduce distortions,
environmental taxes at (or below) their optimal level (partially) correct an
economic distortion. As far as the environmental dimension is concerned this
argument is uncontroversial.
However, as far as marketed output alone is concerned, and leaving aside issues
of unemployment, carbon taxes will only increase output when substituting for
other taxes if they have less effect on output than the taxes they replace. Using a
general equilibrium model Goulder (1995b) concludes that, leaving aside its
environmental benefits, a carbon tax is substantially more distorting than personal
income, corporate or payroll taxes. He argues that the distortions it introduces
arise from its non-uniformity across energy carriers (precisely the characteristic
that makes it an efficient environmental tax), the narrowness of the tax base, and
the fact that it falls on intermediate inputs as well as final output.
This result has been challenged by INFRAS (1996), who point out that the size
of the distortionary effect of a tax depends not only on the tax rate (which, for a
given revenue, will tend to be higher and more distortionary the narrower the tax
base), but also on the elasticities of demand and supply of the taxed quantity. The
net effect of increasing energy and, for example, reducing labour taxes will
therefore depend on the relative elasticities of labour and energy. Using the
ECOPLAN computable general equilibrium model for Switzerland, INFRAS
(1996, p. 141) finds that the so-called distortionary effects of a carbon tax are
substantially less than the reduced distortions from lowering employers' labour
taxes, yielding a positive effect on output, even before any extra output from
reducing unemployment is considered.
In addition to the independent distortionary effects of different taxes, there is
the possibility that the taxes interact such that changes in one tax affect the
distortions caused by others. This is the effect which has been investigated in a
number of recent papers (e.g. Bovenberg and de Mooij, 1994; Goulder, 1995a,b,
Parry, 1995, 1997; Bovenberg and Goulder, 1996; Goulder et al., 1997; Parry et al.,
What is known as the tax interaction effect arises when environmental taxes are
levied in a context of pre-existing distortionary taxes (the above papers focus
predominantly on labour taxes). The interaction of the environmental tax with
these pre-existing taxes invalidates the normal result from environmental tax
analysis that the tax should be set at a rate equal to the marginal external
environmental damage cost at the optimal damage level. In general, the result of
the tax interaction effect is that the optimal tax rate is less than the optimal
(Pigovian) tax calculated in the absence of pre-existing taxes.
#Blackwell Publishers Ltd. 2001
The mathematical analysis which generates this result is formidably complex
and cannot be explored here. It is based on the usual assumptions underlying
exercises in utility maximisation in perfectly competitive markets (for example, a
single representative household, full market clearing in general equilibrium), and
this context needs to be borne in mind when interpreting the results or seeking to
apply them in practice.
Parry et al. (1999), which is representative of the recent literature in this area,
explores the implications of the tax interaction effect for two policy instruments: a
carbon tax, with revenues recycled by reducing the rate of a pre-existing labour
tax; and grandfathered carbon emission permits, which still constrain carbon
emissions, but do not generate government revenues.
The paper's analysis generates three main results. It shows the existence of the
tax interaction effect, which acts to reduce the optimal carbon tax, as noted above.
In the case of the carbon tax the tax interaction effect is offset to some extent by
the revenue-recycling effect of using the carbon tax revenues to reduce the labour
tax rate. However, the offset is only complete on the first increment of the carbon
tax when it is first introduced. At all higher levels the tax interaction effect more
than cancels out the gains from revenue recycling, thereby increasing the costs of
the carbon tax above what they would have been in the absence of the pre-existing
labour tax. In the case of the carbon permits, there is no offsetting revenue-
recycling effect, so the negative impact of the tax interaction effect on the costs of
carbon control is correspondingly greater.
In a numerical simulation Parry et al. (1999) quantify the impact of the tax
interaction effect. They find that it is significant. A pre-existing labour tax of 40%
increases the cost of the carbon tax by 22% relative to the zero labour tax case.
For the carbon permits the cost increase is more than double that of the carbon
tax case (for all emission reduction levels up to 25%), reflecting the absence of the
revenue-recycling effect (Parry et al., 1999, pp. 69 ± 70).
The general conclusions from this literature are, in line with the arguments
above, that the tax interaction effect in actually existing economies (i.e. with
significant pre-existing labour taxes) can substantially increase the costs of carbon
control, thereby reducing the optimal emission reduction, and that this effect is
substantially greater where freely issued permits are the instrument of control,
rather than a carbon tax with revenue recycling.
2.4. Conclusions on the costs of carbon control
The costs of carbon abatement depend on many factors, which are discussed in a
subsequent section. Here the focus has been on theoretical analysis as to whether
the costs may be to some extent offset by an instrument which raises government
revenue, and returns these to the economy by cutting pre-existing distortionary,
especially labour, taxes.
The conclusion of the analysis is unequivocal: revenue-raising instruments such
as carbon taxes and auctioned emission permits, with revenue recycling that
reduces other taxes, will reduce emissions at lower cost than both regulations and
#Blackwell Publishers Ltd. 2001
market-based instruments without revenue recycling (such as grandfathered
emission permits). The benefits of the revenue-recycling effect will be higher when
cuts in employers' payroll taxes increase labour demand, especially where there is
unemployment. However, as discussed above, they are also highly significant in
the stylised theoretical context that has been used in the analysis of the tax
interaction effect. Estimates of the quantitative size of the revenue-recycling effect
in other contexts are presented in Section 5 below.
3. Practical experience of carbon taxes and carbon emissions trading
3.1. Carbon taxes in practice
Section 2 showed that carbon taxes have a well-established theoretical basis,
which suggests that they could be a cost-effective policy instrument for the
reduction of CO2emissions. As a result there has been much policy interest in
carbon taxes, and numerous recommendations from policy analysts that they
should form part of any package of measures to address climate change (Pearce,
1991; RCEP, 2000, Recommendation 11, p. 200).
In contrast to this enthusiasm derived from theory, the practical implementa-
tion of carbon taxes has been cautious and modest, and is limited to six European
countries (see below).
Italy was the most recent country to implement a carbon tax, which was
introduced in 1999, and was expected to raise Euros. 1.13bn in that year (ENDS,
Table 3.1 gives some summary information about the carbon taxes which have
been implemented in the five other European countries, together with some
comparative information about some other countries' revenues from energy
Table 3.1 shows revenues from carbon and energy taxes for a number of
European countries. Of course, even for those countries without specific carbon
taxes, taxes on fossil fuels act as implicit carbon taxes, but they are not efficient, in
the sense of equalising the marginal cost of carbon abatement across the tax base,
if their purpose is to reduce CO2emissions.
Table 3.1 shows that carbon taxes were introduced in the 1990s. Carbon tax
revenues compared to energy tax revenues are most significant for Sweden, but
even then they comprise only 26%. This reflects the fact that energy taxes were
important revenue-raising taxes well before climate change, and therefore CO2
taxes, became a policy concern. Some of the five countries (for example,
Denmark, Netherlands, Sweden) introduced CO2taxes as part of a reform of
existing energy and other taxes, in order to give greater weight to environmental
considerations. Other countries now treat energy taxes de facto as environmental
taxes, even if they were introduced purely for revenue reasons and have no specific
environmental focus.
Table 3.1 also shows that, with the exception of Finland, the countries with
CO2taxes, compared to the countries without them, derive a relatively large
#Blackwell Publishers Ltd. 2001
proportion of their taxes from energy and environmental taxes. If this proportion
is an indicator of environmental policy concern, then it may be that other
countries will introduce carbon taxes if such concern continues to increase.
All the countries which have introduced a carbon tax have also introduced
special tax reductions, rebates, or exemptions, in order to address concerns about
the effect of the tax on industrial competitiveness (of course, this applies to many
countries' treatment of energy taxes as well), as discussed below. These
exemptions are complex and discussed in detail in Ekins and Speck (1999a).
Some of the principal features of these special arrangements for industrial energy
users are:
*In Sweden manufacturing industry only pays 50% of the CO2tax rate of
around Euros 40=tonne CO2.
*In Denmark the CO2tax rate varies according to whether the use is for space
heating, when it is around Euros 80=tonne CO2for households and businesses,
`light industrial processes', which pay around Euros 12=tonne CO2, and `heavy
industrial processes', which only pay around Euros 3=tonne CO2.Sucha
differential in tax rates has serious implications for the theoretical efficiency
advantages of using carbon taxes to reduce carbon emissions.
*In Norway the CO2tax rate varies with the fossil fuel, from about Euros
16=tonne CO2for heavy fuel oil, to around Euros 20=tonne CO2for coal, to
around Euros 44=tonne CO2for natural gas. In addition, reduced tax rates
apply for some industrial sectors. For example, the pulp and paper industry
and the fishmeal industry only pay 50% of the tax on heavy fuel oil.
*The 1992 proposal from the European Commission for a carbon=energy tax
provided for the exemption of the six most energy-intensive industrial sectors
Table 3.1 Carbon, energy and environmental taxes in some European countries.
Countries with
carbon taxes
(date of introduction)
from CO2
from energy
revenues in
energy tax
% of energy and
taxes in total
Denmark (1993) 45712905116 9 (1997)
Finland (1990) 43612519117 5 (1995)
Netherlands (1996) 82826990112 14 (1997)
Norway (1991) 32312429113 10.8 (1993)
Sweden (1991) 134415140326 10.1 (1998)
Other countries
France ... 21 652 ... 5.7 (1995)
Germany ... 31 588 ... 6.3 (1995)
Spain ... 8482 ... 8.1 (1995)
UK ... 26 223 ... 7.9 (1995)
Sources: Baranzini et al., 2000, Table 2, p. 399; EC 1999, various tables
1Figure for 1996
2Half the total revenue of two energy taxes which are calculated on a 50=50 energy=CO2basis
3Excludes revenue from an electricity tax
#Blackwell Publishers Ltd. 2001
(but even so failed to be implemented, at least partly because of strong
business opposition, see Ikwue and Skea, 1996).
Such differentials in tax rates have serious implications for the theoretical
efficiency advantages, and may be expected to increase the macroeconomic cost,
of using carbon taxes to reduce carbon emissions.
Of the five North European countries listed in Table 3.1, Finland, Norway and
Sweden allocate the revenues from their carbon taxes to the general government
budget. This means that, for a given level of government expenditure and fiscal
balance, other taxes are lower than they would otherwise be, but there is no
specific offsetting of other taxes against the carbon taxes, as is sometimes the case
in what is called `environmental tax reform'. In Denmark the carbon tax revenues
from industry are recycled back to industry through reduced social security
contributions and through investment incentives, and in the Netherlands the
revenues are recycled back to households and industry through personal and
corporate tax relief (EC, 1999).
3.2. Carbon emissions trading in practice
It was noted in Section 1 that it is envisaged that an international carbon
emissions trading regime will be put in place under the terms of the Kyoto
Protocol. However, by the end of 2000 the details of the scheme had still not been
fully worked out and agreed. At that time the only operational carbon emissions
trading scheme was the one introduced by Denmark's government for the Danish
power sector. This sector currently emits around 40% of Denmark's CO2
emissions, and the scheme is intended to make a contribution to the 21%
emissions reduction (by 2008 ± 2012 from 1990's level) to which Denmark is
committed under the Kyoto Protocol. The sector's allocated emission permits are
due to be reduced from 23 million tonnes (mt) CO2in 2000 to 20 mt in 2003, a cut
of 13%, when the scheme will end. The permits will be divided among about 15
companies, which may bank or trade them between themselves. Firms which emit
CO2in excess of the permits they possess will become liable to an additional
carbon tax of Euros 5.4 per tonne CO2(ENDS, 1999b; ENDS, 2000; the power
sector is currently exempt from Denmark's carbon tax). Analysis of such
interactions between carbon taxes and emissions trading schemes would seem to
present a useful avenue for future research.
A number of other countries are developing carbon (or greenhouse gas)
emissions trading schemes, including Australia, Canada, the US and a number of
European countries (including France, Germany, the Netherlands, Norway,
Sweden and the UK). In Europe the European Commission issued a Green Paper
on greenhouse gas emissions trading in March 2000 (EC, 2000), outlining a
scheme that is intended to be introduced in 2005. The scheme is initially intended
to cover only CO2emissions from `large fixed point sources' (EC, 2000, p. 10), but
could later be extended to other greenhouse gases and sources of emissions.
#Blackwell Publishers Ltd. 2001
4. Environmental effectiveness of carbon taxes
The purpose of carbon taxes is to reduce carbon emissions cost-effectively. It has
already been noted that they are widely regarded as being (with auctioned
tradable permits) the most efficient policy instruments for carbon reduction,
meaning that they will achieve a given carbon reduction at less cost than other
instruments. This section explores how carbon taxes bring about their reduction
effects, the responsiveness of carbon emissions to carbon taxation, and how their
environmental effectiveness may be increased.
The reduction of carbon emissions comes about through several possible
different effects.
4.1. Effects of carbon taxes on the demand for carbon-based fuels
Effects on producers
For producers, carbon dioxide may be related to output (O) through the following
decomposition into different ratios:
CO2(CO2=E) +(E=ES) +(ES=I) +(I=O) +O(4:1)
where CO2is carbon dioxide emissions, E is energy inputs, ES is energy services
(useful heat, light, power), I is inputs to production, and +denotes multiplication.
Where the tax falls on the carbon inputs to production, to the extent that
producers cannot pass the tax on to consumers it will tend to reduce their profits
and output (O), and therefore their CO2emissions. To counteract the effect on
profits and output, producers will seek to reduce their CO2emissions by reducing
the other terms in equation (4.1), where it is cost-effective to do so. Thus they will
tend to switch to less carbon-intensive fuels (reducing CO2=E), to use energy more
efficiently in current production processes (reducing E=ES), and to reduce their
demand for energy services relative to other production inputs (reducing ES=I).
They will also seek to develop new technologies for future production processes
which reduce all these ratios.
Effects on consumers
For consumers, carbon dioxide may be related to consumer expenditure (CE)
through the following decomposition into different ratios:
CO2(CO2=E) +(E=ES) +(ES=EIGS) +(EIGS=CE) +CE (4:2)
where the letters are as defined above and EIGS stands for energy-intensive goods
and services.
To the extent that producers can pass the tax on production inputs on to
consumers, or where the tax falls directly on consumers' consumption of carbon-
based fuels, they will seek to switch to less carbon-intensive fuels (reduce CO2=E),
use energy more efficiently (reduce E=ES), and reduce both the proportion of
#Blackwell Publishers Ltd. 2001
energy services in energy-intensive goods and services (reduce ES=EIGS) and the
share of EIGS (including ES) in consumer expenditure on all goods and services
(reduce EIGS=CE).
Estimates of the effectiveness of carbon taxes in reducing CO2emissions will
therefore depend on assumptions about:
*The ease of switching between more and less carbon-intensive fuels (affecting
*The opportunities for delivering energy services more efficiently (affecting
*The growth in demand for energy services, relative either to other production
inputs (affecting ES=I), or to other energy-intensive goods and services
(affecting ES=EIGS)
*The growth in demand for energy-intensive goods and services relative to
other goods and services (EIGS=CE).
The standard method of estimating energy (and thereby CO2) elasticities is
through the econometric use of time-series or cross-section data. Yet the principal
price changes in time-series data have been the sudden price increases in the 1970s,
due to OPEC production restrictions, and the equally sudden price reduction in
the mid-1980s, when those restrictions proved unsustainable.
Barker et al. (1995a, p. 11) note that there are a number of reasons why such
price changes are unlikely to reflect satisfactorily the likely responses to a carbon
tax that was set to escalate gradually over a number of years:
*Such a tax would be more likely to be perceived as permanent than cartel-
induced price increases, increasing the likely behavioural response and giving
producers a greater stimulus to develop low-carbon technologies.
*The gradual introduction of the tax would reduce the impacts of the price
increase on the economy by allowing the capital stock to be replaced at the
end of its useful life rather than inducing premature scrapping.
*Because the tax would be part of a long-term, well understood policy
commitment (to mitigate climate change), consumers could also anticipate it
in their behaviour, rather than being caught by surprise.
*The revenues from the tax would stay in the domestic economy, rather than
accruing to foreign oil producers, giving opportunities to recycle them by
reducing distortionary taxes. Even if this were not done, the macroeconomic
effects of such taxes would be very different to the OPEC-induced price rises.
All these factors would be likely to increase the price-elasticity of demand in
response to a carbon tax relative to those estimated from data from the 1970s and
1980s. In addition, both price and income elasticities will be affected by the degree
of market saturation for energy services and energy appliances. For example,
domestic warmth will not tend to grow with income, once a desired comfort level
has been reached.
The different chapters in Barker et al. (1995b) discuss many of these effects in
detail. Their overall implication is that many of the published estimates of CO2
#Blackwell Publishers Ltd. 2001
reduction from carbon taxes are likely to be too low (and the costs of a given CO2
reduction are therefore too high) because these factors are often not taken into
account. This also results in a very wide range of published estimates of the
effectiveness of a carbon tax, as illustrated by the Energy Modeling Forum (EMF)
at Stanford University, which specified standardised scenarios for fourteen widely
differing economic models of the US economy. In these models a carbon tax of
$80=tonne carbon brought about a change in CO2levels with respect to 1990 of
between 35% and 20% (Gaskins and Weyant, 1993, p. 319).
Similarly Repetto and Austin (1997, p. 13) have shown that the cost estimates of
different models of reducing carbon emissions by 60% from a projected baseline by
2020, reflecting different required levels of a carbon tax, varies from 71
of GDP, depending on the models' treatment of issues including those listed above.
The assumptions behind these cost estimates are discussed further in Section 5.3.
4.2. Effects of existing carbon taxes
The problems of estimating the environmental effectiveness of a carbon tax ex
ante (i.e. in advance of its introduction) are in no way reduced ex post (i.e. after
the event). This is because ex post evaluations have to compare the situation after
the event with a baseline of what would have happened without a carbon tax.
Constructing such a baseline faces identical problems to estimating the effects of
the tax ex ante. In addition, carbon (and energy) taxes are often introduced as part
of a carbon control policy package, rather than as separate measures. Estimating
their effectiveness then requires the impact of the different components of the
package to be separated out, which, as OECD 1997 (p. 122) notes, `can prove
completely impossible'.
Notwithstanding these difficulties of evaluation, most of the countries that have
introduced carbon taxes have sought to estimate their effectiveness (Ekins and
Speck, 1999b for a discussion). Thus the Dutch Green Tax Commission in the
Netherlands has calculated that each Euros. 450m raised from its fuel taxes leads
to a CO2reduction of 1± 1.5 mt. Larsen and Nesbakken (1997, p. 287) found that
the total effect of the Norwegian CO2tax on CO2emissions was 3 ± 4% for the
period 1991± 1993. The Danish Environmental Protection Agency (DEPA)
estimated that the Danish CO2tax would reduce CO2emissions by 1.6% (DEPA,
1999, p. 87). In Sweden a study commissioned by the Swedish Environmental
Protection Agency (SEPA) concluded that Swedish CO2emissions in 1994 `were
just under 5 mt lower than they would have been without the carbon dioxide tax'
(SEPA, 1997, pp. 48± 49).
4.3. Increasing CO2reduction from a carbon tax
It has already been noted that the evaluation of the effects of a carbon tax is
incomplete without consideration of the use to which its revenues are put. One
possible use of the revenues is the enhancement of the environmental effectiveness
#Blackwell Publishers Ltd. 2001
of the tax through subsidies, tax expenditures or tax incentives for carbon
Rajah and Smith (1993) review the issues raised by these kinds of measures.
They note that `first-best policy' would have no role for such measures, but that
where considerations of competitiveness or distribution (see Section 5) introduce
political constraints on implementing a carbon tax at the desired level, these
measures may provide an alternative means of emission reduction. Rajah and
Smith explore a number of implications of these measures, noting in particular
that they tend to contravene the Polluter Pays Principle, and that their use should
therefore be cautious and subject to international monitoring if it is not to risk
becoming a trade distortion. However, they conclude that investment subsidies, in
particular, may have a role to play as a complement to environmental taxes
(Rajah and Smith, 1993, p. 62).
Whatever the soundness of their theoretical basis, tax expenditures and
investment incentives for environmental protection are widespread. Rajah and
Smith (1993, p. 54) noted that a 1993 OECD study had identified their use in 14
OECD countries, and they play an especially important part, alongside a carbon
tax, in the carbon-reduction policy packages of Denmark and the Netherlands
(EC, 1999).
5. Economic and distributional impacts of carbon taxes
Implementing a carbon tax will increase the relative price of fossil energy
compared to other inputs, an effect which may be enhanced by reducing taxes on
these other inputs (e.g. labour, capital). But the change in relative prices could
affect economic development in a number of other ways. As discussed in
Section 2, one of the most important determinants of the overall impact of a
carbon tax will be the use that is made of the revenues. This issue emerges
repeatedly in many of the areas of discussion which follow.
5.1. Investment, efficiency and technical change
Increased scrapping
A change in relative prices caused by the imposition of a carbon tax might affect
economic development by making existing capital equipment uneconomic,
thereby bringing forward its scrapping date. This could be a major potential
source of adjustment costs related to the tax. One would expect that the least
disruptive imposition of a carbon tax would be one introduced initially at a low
level, with modest annual increases over a substantial, pre-announced period of
time. This would allow responses to the tax to be synchronised with normal
investment schedules.
Ingham and Ulph (1991a) have developed a vintage model of the UK
manufacturing sector, which allows firms to change their machines' energy-output
ratio, according to relative factor prices, both between different machine vintages
#Blackwell Publishers Ltd. 2001
and with existing machines. Technical change is thus at least partly endogenised.
The model clearly confirms the above intuition. There is a significant increase in
the tax rate required to meet, and the associated cost of meeting, a given target if
the target date is brought closer or if action is delayed, or if the target entails
cutting existing emissions rather than preventing future growth. As Ingham and
Ulph (1991a, p. 143) say: `It is much more expensive to undo the effects of
emissions-generating plant already installed, than it is to offset the effects of
emissions-generating plant yet to be installed.'
Moreover, the Ingham and Ulph model suggests that the increased scrapping
may lead to higher long term growth that outweighs the short term adjustment
costs. The authors find that `in the short run output falls, and this induces
considerable scrapping of equipment which leads to lower costs and prices, and
output being higher in the longer term than in the case where demand is
determined exogenously. In the extreme case, output growth rises from 2 to
4.4%.' (Ingham and Ulph, 1991b, pp. 198 ± 199)
Improvements in energy efficiency
Many analysts have argued that market failures are preventing the implementa-
tion of already cost-efficient energy-conservation measures (e.g. Lovins and
Lovins, 1991; Jackson and Jacobs, 1991; Jackson, 1995). Jackson (1991) provides
evidence that the energy market is far from perfect. He finds that out of 17
technological possibilities for the reduction of CO2emissions, eight could be
implemented at negative cost on the basis of current prices, saving a total of 165
million tonnes of CO2per year by 2005, or 24% of UK 1991 emissions. On this
analysis the UK could have exceeded the Toronto target for CO2emissions (20%
reduction from 1988 levels by 2005) and saved money. This is not an
unrepresentative result. After reviewing this issue, Cline (1992, p. 227) decides
that a reasonable estimate is that the first 22% of carbon emissions from base can
be cut back at zero cost. The IPCC survey of this literature (Bruce et al., 1996,
pp. 310, 318) finds that zero cost emission reductions by 2025=2030 estimated by
various studies ranged from >61± 82% (for the US) and from >45± 60% for other
OECD countries.
If it is true that many presently economic opportunities for energy-saving
remain unimplemented because of market failures (for example, lack of
information, excessive discount rates, landlord=tenant problems etc.), then it
might be expected that a continuously increasing energy price, by providing a
continuously increasing incentive to correct such failures, would result in
substantial investments in energy efficiency and further innovation in this area.
Private efforts in this field could be complemented by government initiatives to
encourage energy conservation and efficiency. If energy efficiency could thereby
be increased at no net cost at the same rate as the price of energy, then the
negative effect of the rising price on the competitiveness of even the most energy-
intensive sectors would be cancelled out.
#Blackwell Publishers Ltd. 2001
Influencing technical change
In addition to encouraging the adoption of existing energy-efficiency technologies,
rising energy prices would also give a positive stimulus to the development of new
energy-efficiency and non-fossil energy technologies. Such technical change would
be what Grubb (1995, p. 305) calls `induced technology development', which he
speculates may be a cause of asymmetrical elasticities of energy demand, for which
there is now substantial evidence (in general falls in energy prices have not
increased energy demand by as much as the preceding energy price increases
reduced it). Grubb concludes: `If price rises stimulate technical development and
in addition governments take further associated action to encourage energy
saving, long-term solutions may emerge at relatively lower cost as a result of the
accumulation of technical change in the direction of lower CO2-emitting
technologies, infrastructure and behaviour.' (ibid., p. 309).
Technologies do not emerge from on high. They evolve in response to pressures,
which may be the competitive forces of the market, or the demands of public
policy. Grubb et al. (1995, p. 420) point out that the no-regrets potential for
increased energy efficiency in the UK in 1980 was identified as about 20%. In
1990 it was again identified as about 20%, despite the fact that the earlier 20%
had largely been realized. They comment: `It seems a curious feature of energy
efficiency studies that they seem regularly to identify cost-effective potentials of
around 20± 30% of current demand, almost irrespective of the potential already
exploited. ... The persistence of such results suggests that investing in greater
energy efficiency helps itself to stimulate and identify options previously
5.2. Competitiveness of firms and industrial sectors
Section 3 noted the reductions in rates of carbon taxes that are levied on
manufacturing industries in general, or energy-intensive industries in particular,
reductions which are often also applied to other energy taxes. The reason for these
reductions is fears of the potential impact of such taxes on industrial
competitiveness. The special treatment of Swedish manufacturing industry with
respect to the taxation of energy is explicitly due to such fears. It was felt that the
tax system prior to the reform of 1993 ` ... had proved to constrain the
competitiveness of the Swedish industry' (TemaNord, 1994, p. 95).
There are no studies explicitly of the ex post impact of carbon taxes on
industrial competitiveness, partly at least because such taxes have only recently
been introduced and have, in any case, been explicitly designed not to have such
impacts through the exemptions and lower tax rates that have been applied.
However, there have been many studies exploring the impacts on competitiveness
of environmental policy in general, which may be considered relevant to the
introduction of carbon taxes. There have also been numerous studies which model
the potential impacts of carbon and energy taxes on the competitiveness of
particular economic sectors and on the economy as a whole. The general and
#Blackwell Publishers Ltd. 2001
sectoral studies are discussed next below, while those focused on the
macroeconomic impacts of carbon taxes are discussed in Section 5.3.
Considerations of competitiveness are important to environmental policy, and
carbon taxes, for both economic and environmental reasons:
1. Economic Ð if environmental policy produces negative impacts on
competitiveness it will be associated with corporate, sectoral or national
economic decline. This is both important in itself, and would make the
introduction of environmental policy politically difficult or impossible.
2. Environmental Ð if domestic environmentally or energy intensive industry
declines, to be replaced by a growth in such industries in other countries,
overall environmental impacts may not change. In respect of a global
pollutant such as CO2, this would mean that the loss of domestic industrial
competitiveness will have brought no environmental gain at all.
Competitiveness basically denotes the ability of a firm, economic sector or
national economy, or a productive sector, to sell its goods and services in domestic
and world markets. There are many possible indicators of competitiveness, some
of which become policy targets in their own right and even become taken for
competitiveness itself. These indicators include: income per head; balance of trade;
exchange rate movements; unit labour costs; generation of employment; labour
productivity; market share; profitability; firm growth; and the share of world
At the firm level the logic behind the fear of impacts on competitiveness from
environmental taxes is simple and persuasive: taxes on business inputs inevitably
add to business costs; where these taxes are imposed in one country only, these
extra costs will impair the international competitiveness of the business or sector
concerned. However, it may not always be the case that environmental policy
imposes costs on firms; even where it does the costs may not be substantial enough
to affect competitiveness; or the policy may generate benefits for the firm to set
against the costs. In principle, however, it is clear that environmental, including
carbon, taxes could affect industrial competitiveness.
It should be clear that effects on competitiveness will only arise if environmental
policy in different countries imposes different levels of costs on competing firms.
Thus, although the economic effects of environmental policy may be measured in
terms of reduced labour productivity, or reduced rates of economic growth, these
are only effects on competitiveness if they differentially affect some firms and not
their competitors. However, because in practice environmental policy and the
regulations to which it gives rise are not greatly harmonised between countries
(although such harmonisation is more apparent in groups of countries like the
European Union), such measures are often interpreted as indicating effects on
There is now a substantial literature on the effects of past environmental policy,
and the possible effects of future environmental policy, on the competitiveness of
businesses and countries. Here it is only possible to summarise the main
#Blackwell Publishers Ltd. 2001
arguments and results of this literature (for a fuller treatment see Ekins and Speck,
1998; other contributions in Barker and Ko
Èhler, 1998).
The conventional economic view is that the realisation of environmental
benefits through environmental policy is likely to entail economic costs.
One result of these possible cost increases resulting from environmental policy
is that affected industries will move to countries which have less stringent or no
environmental policies, the so-called `pollution haven hypothesis', an issue studied
by Lucas et al. (1992). They found that the growth rate of the toxic intensity of
manufacturing was both higher in the poorest countries and increased through the
1970s and 1980s. The authors consider that this is consistent with the hypothesis
that `stricter regulation of pollution-intensive production in the OECD countries
has led to significant locational displacement, with consequent acceleration of
industrial pollution intensity in developing countries.' (Lucas et al., 1992, p. 80).
This result showing the absolute and relative growth of pollution-intensive
industry in poor countries is confirmed by Low and Yeats (1992) through quite
different non-econometric analysis of trade statistics. However, while Low and
Yeats agree that the result is consistent with the Lucas displacement hypothesis,
they stress that there are a number of other possible explanations, including that
the strong growth of `dirty' industries is a normal occurrence at an early stage of
Other studies, surveyed by Dean (1992, pp. 16± 20), give conflicting results,
but overall do not suggest that the forces for displacement are very great. In the
same vein, Jenkins (2001) has found that there is no relationship between
pollution intensity and foreign ownership in the manufacturing sectors of
Malaysia and Indonesia, suggesting that foreign direct investment in such
developing countries is not disproportionately driven by the quest for `pollution
havens'. Moreover, if it is true that environmentally intensive sectors are also
capital-intensive, as suggested by Nordstro
Èm and Vaughan (1999, p. 31),
comparative advantage theory suggests that these industries would tend to be
located in capital-abundant (i.e. industrialised) countries. In fact, they present
World Bank data which shows that high-income countries do indeed export
more pollution-intensive goods than they import, while the reverse is true for
upper-middle-income, lower-middle-income and low-income countries (Nord-
Èm and Vaughan, 1999, p. 32).
To set against the reasons for possible increases in cost from environmental
policy, there are a number of possible benefits from environmental policy,
including cost reduction (from more efficient use of environmental inputs), first-
mover advantages and stimulation to innovation, which have become important
elements in the debate on this issue. Considerations of this sort have led Porter
(1990, pp. 647± 648) to hypothesise that environmental regulations may be good
for economic competitiveness. This `win-win' hypothesis of the economic, as well
as environmental, benefits of environmental regulation runs clearly counter to
economists' normal assumptions of efficient, competitive markets. It has been
attacked as being at best a marginal phenomenon with regard to the costs of
environmental regulation as whole. Palmer et al. (1995, pp. 127 ±128) estimate
#Blackwell Publishers Ltd. 2001
that Porter's `innovation offsets' amount to only a few percent of the total costs of
conforming to environmental regulations, which in the US have been estimated by
the EPA at $135 billion in 1992. They contend that the vast majority of these costs
conform to the standard economic trade-off model, whereby environmental
benefits are gained at the expense of growth and competitiveness.
Analysis by Jenkins (1998) gives marginal support to Palmer et al. over Porter.
Overall, however, Jenkins' review of the literature suggested that `there was no
strong universal relationship between environmental pressures and competitive
performance, either at the firm level or the industry level' (Jenkins, 1998, p. 38).
This is very much in line with the OECD's conclusions on the impacts of
environmental policy, namely that: `The trade and investment impacts which have
been measured empirically are almost negligible.' (OECD, 1996, p. 45).
However, the past, in terms of the effect of environmental policy on
competitiveness, may not be a good guide to the future. The environmental
policy measures applied so far have been relatively modest compared to those that
are sometimes considered necessary, especially in respect of carbon abatement,
and they have not predominantly been in the form of environmental or carbon
taxes. These may impose higher costs on seriously affected economic sectors
because the firms concerned will need to pay for abatement (up to the efficient
level) and for residual emissions. While this feature of environmental taxes results
in the advantage over regulation that they give an incentive for continual
environmental improvement at all levels of use, it also means that affected
businesses pay more under an environmental tax regime than under regulations.
Therefore the impacts of environmental, including carbon, taxes need to be
examined separately from an assessment of the impacts of an environmental
policy which has so far relied largely on regulation. Such examinations have been
carried out principally through modelling studies.
Modelling the effects of carbon taxes on sectoral competitiveness
The effect of a carbon tax on sectoral competitiveness will be determined by a
number of influences, including:
*the carbon intensity of the product.
*the trade intensity of the product (ratio of exports plus imports to production).
*the size of the carbon tax and the manner in which the revenues from it are
Because the first two of these influences differ between countries, even the
imposition of a carbon tax with an identical use of revenues in all countries would
differentially affect their competitiveness. The role of these factors in determining
the sectoral impact of a carbon tax was clearly shown by Pezzey (1991), who
simulated the introduction of a carbon tax of $100 per tonne carbon in the UK.
The results are worth quoting in some detail because they clearly illustrate the
essential considerations in evaluating the first order sectoral outcome of a carbon
tax with revenue recycling.
#Blackwell Publishers Ltd. 2001
Using data for 10 UK manufacturing sectors, Pezzey showed how:
*The first-round impact of a carbon tax resulted in increased costs for all the
sectors when the revenues were simply added to the general government budget;
*The tax resulted in increased costs for only the four most relatively carbon-
intensive sectors (iron & steel, chemicals, non-ferrous metals, non-metallic
minerals) when the revenues were returned to the sectors relative to their
output. Six sectors therefore experienced cost reductions due to the carbon tax;
*The cost effects on the four negatively affected sectors were reduced when
their trade intensity was taken into account. In particular, the low trade
intensity of iron & steel and non-metallic minerals (both sectors comprising
heavy, bulky goods including iron and cement) substantially reduced the
trade impacts that these sectors suffered from the carbon tax.
It may further be noted that, even if the revenues from the carbon tax are not
returned to the affected sectors, they will only experience negative cost impacts
from the tax to the extent that they do not reduce their carbon-intensity at a rate
equal to the tax being applied. If it is true that there are substantial unexploited
opportunities for cost-effective gains in energy efficiency, as discussed above, then
competitiveness impacts will be reduced to the extent that these opportunities are
Further analysis by Ekins (2000, pp. 267± 268) showed that the four negatively
affected sectors accounted for 16% of UK exports, while those that might be
expected to be relatively advantaged by such a carbon tax=revenue measure
accounted for 55%. On these figures it may well be that the UK's international
trading position would be improved by the tax-plus-rebate arrangement. It may also
be noted that 57% of UK exports in 1995 were to EU countries, so that if the carbon
tax was imposed on an EU-wide basis (as was the proposal from the European
Commission in 1992), the trade effects for all sectors would be much attenuated.
The conclusion that environmental taxes need not result in unacceptable effects
on industrial competitiveness would appear to be borne out by the experience of
Denmark, which has a small, open economy, and which has been a pioneer in the
area of environmental taxation. According to its Ministry of Economic Affairs:
`Danish experience through many years is that we have not damaged our
competitiveness because of green taxes. In addition, we have developed new
exports in the environmental area.' (Kristensen, 1996, p. 126). The study of the
Norwegian Green Tax Commission (1996, p. 90) has also endorsed this essential
conclusion: `Reduced competitiveness of an individual industry is not necessarily a
problem for the economy as a whole. ... It is hardly possible to avoid loss of
competitiveness and trade effects in individual sectors as a result of policy
measures if a country has a more ambitious environmental policy than other
countries or wishes to be an instigator in environmental policy. On the other
hand, competitiveness and profitability will improve in other industries as a result
of a revenue neutral tax reform'.
While the Pezzey simulation reported above only takes account of immediate,
first-round effects of the relative price-changes, rather than eventual adjustments
#Blackwell Publishers Ltd. 2001
to equilibrium, the main mechanisms through which imposing environmental
taxes influences sectoral competitiveness are clear, as is the difference between the
impacts from environmental taxes on sectoral and national competitiveness. The
cost increases in the four most affected sectors will impair their position in
domestic markets with respect to the products of other sectors. The six sectors
whose costs are decreased by the revenue-recycling will be particular beneficiaries
from the shift in relative prices.
To gain more detailed insights into the sectoral competitiveness effects of
environmental, including carbon, taxes, a disaggregated sectoral economic model
incorporating interaction and feedback effects is required. The estimation of the
price effects induced by the imposition of an environmental tax is often carried
out using a cost driven input-output price model. The impacts on competitiveness
are then analysed by the development of the sectoral prices following the
introduction of an environmental tax and the respective recycling measures of the
generated revenues. The price increase induced by, for example, an energy tax
affects not only the economic sectors producing energy products. The prices of all
economic sectors are increasing depending on how much energy is required,
directly and indirectly via intermediate goods, in the production of the goods. By
taking into account indirect and feedback effects from the carbon tax, this goes
further than the Pezzey analysis discussed earlier, which only analysed the carbon
tax's direct effects.
Using such an approach Barker (1995) has examined the issue of
competitiveness using the MDM-E3 model for the UK economy analysing
the implications for industrial costs of a $10 per barrel carbon=energy tax in
the UK, with compensating cuts in employers' National Insurance Contribu-
tions (NIC). The result shows again the importance of how the generated
revenues are redistributed: `If the taxes are not compensated, most industries'
prices rise as they face higher energy and labour unit costs. However if NIC
contributions are reduced to keep the PSBR ratios at base levels then all
industries' costs fall depending on their use of labour Ð and the most labour-
intensive industries will have the largest reduction in costs.' (Barker, 1995,
p. 19). For most sectors the effects from the reduction in labour costs more
than offset the effects from the increase in energy costs. A very similar result
emerged from the study of Germany by the German Institute for Economic
Research (DIW, 1994).
Barker (1998) develops the approach further in a more ambitious study of the
effects on competitiveness of carbon taxes for the EU Member States. Using the
E3ME model, the effects on trade volumes and prices for 11 Member States of a
coordinated carbon tax were compared with those for an uncoordinated tax
achieving the same 10% reductions in CO2emissions below the baseline by 2010.
The main conclusions from the study of the effects of carbon taxes imposed
unilaterally, with revenues recycled via reductions in employers' social security
payments, was that `the effects on industrial price competitiveness ... are small
and mixed ... Nearly all manufacturing sectors have an increase in output ... Any
carbon leakage ... is negligible' (p. 1097).
#Blackwell Publishers Ltd. 2001
Nevertheless it remains the case that fear of the competitiveness effects of
environmental taxes has resulted in most countries that have introduced carbon
and energy taxes for environmental reasons giving vulnerable firms or sectors tax
exemptions or concessions. Theory suggests that these reduce the economic
efficiency of the environmental tax and reduce the economic advantage to be
gained from clean production systems. They also slow down the process of
structural change in the economy, which would lead to energy- and environment-
intensive economic sectors becoming both less intensive, and less important
economically relative to less environment-intensive sectors.
The question arises as to how economically costly these exemptions might be. A
study by Oliveira-Martins, Burniaux and Martin (1992) showed that, for a given
emission-reduction target, the tax exemption of energy-intensive industries in the
EU fails to increase the output level of these industries. This outcome arises
because the exemptions result in higher tax rates for the rest of the economy, so
that the costs of the other sectors are higher and total output falls. A similar result
has been reported by Bo
Èhringer and Rutherford (1997) in their analysis of the
consequences of exempting energy-intensive sectors from a carbon tax. They find
that wage subsidies to export- and energy intensive sectors, rather than tax
exemptions, retain more jobs and are less costly. The study's general conclusions
are: `Welfare losses associated with exemptions can be substantial even when the
share of exempted sectors in overall economic activity and carbon emissions is
small. Holding emissions constant, exemptions for some sectors imply increased
tax rates for others and higher costs for the economy as a whole' (Bo
Èhringer and
Rutherford, 1997, p. 201). Rather than exempting energy-intensive sectors from a
carbon tax, it would seem preferable either to return the revenues from the tax to
the sectors, on some other basis than carbon, or to allow the tax payments to be
set against investments in energy efficiency. Both of these measures would cushion
the tax's effects while maintaining its incentive for carbon-reduction.
Goulder (2000), based on more detailed work by Bovenberg and Goulder
(2001), finds that the potential costs of compensating energy sectors for losses of
profits due to an upstream carbon tax or permit system need not be high.
Specifically, equity values in these industries can be maintained by grandfathering
10% of the emission permits (auctioning the rest) in the primary energy sectors,
and cutting corporate taxes in the downstream energy sectors. These results derive
from the fact that these sectors are able to pass on a large part of the costs of the
carbon taxes or emission permits to energy consumers. The corporate tax cuts
actually improve the efficiency of the economy overall (because corporate taxes
are more distorting than the carbon tax), while the grandfathering of the permits
only increases the cost of carbon reduction by 7%. Goulder considers that such
measures should increase the political feasibility of introducing carbon-control
measures of this sort.
Barker and Rosendahl (2000) look at the same issue of maintaining the profits
of the energy industries that might otherwise lose out under a carbon tax. They
compare 3 scenarios to achieve the Kyoto target of an 8% cut in GHG emissions
below 1990=95 levels by 2008± 12 for 17 west European countries. The scenarios
#Blackwell Publishers Ltd. 2001
are: (1) a carbon tax (2) a grandfathered CO2emission permit scheme and (3) a
mixed multilateral scheme with a permit scheme for the energy sector (energy-
intensive industries and electricity generation) combined with a carbon tax for the
rest of the EU economy. Permit prices are endogenously determined year by year
in the model by market demand and supply, and are the same across the regions.
The mixed scheme has an allocation of permits (between grandfathering and
auctioning) designed to maintain the profits of the energy industries at levels close
to the baseline. In the scheme, 70% of permits are allocated on a grandfathered
basis on 2000 emissions in 2001, 60% in 2002, 2003 and 2004, 55% in 2005 and
50% for all later years. Reductions for CO2emissions in terms of permits issued to
the year 2010 are calculated to be 25% below those of 1990 levels for the scenario
to achieve the Kyoto target. All implied values of grandfathered permits are
allowed to increase profits. A carbon tax at the rates in scenario 1 above is
introduced for all fuel users not covered by the permit scheme, including
transportation and households. All revenues from taxes and auctions are used to
reduce regional employers' contributions to social security. The study shows that
it may be possible for Europe to achieve Kyoto targets using market-based
policies, with negligible effects on GDP and inflation, and without damaging the
profitability of energy-producing and using industries (Barker and Rosendahl,
2000, p. 433).
However, other modelling studies have produced a wide range of macro-
economic impacts of carbon taxes, analysis of which is the subject of the next
5.3. Macroeconomic assessments of carbon taxes and auctioned permits
The objectives of the carbon tax studies
There is an extensive literature on assessing the macroeconomic effects of carbon
taxes. The studies are of particular tax proposals, such as the European
Commission's carbon=energy tax proposed in 1993, or they are designed for the
tax to achieve target reductions in CO2or CO2-equivalent emissions for particular
countries or groups of countries. The targets in these studies have changed as the
international community has moved towards more detailed and binding
*Studies of the early 1990s aimed for a 20% reduction in CO2emissions below
1988 levels by 2005 (the Toronto target);
*Studies in the mid 1990s aimed at the Rio non-binding target of a return to
1990 levels of CO2emissions by 2000;
*Studies in the late 1990s aimed at the Kyoto target for a basket of six GHGs,
with a range of targets, intended to become legally enforceable, for Annex B
(industrial) countries (US 7%, EU 8%, Japan 6% and all Annex B countries
about 5%) for average emissions 2008 ± 12 below 1990 and=or 1995 levels.
#Blackwell Publishers Ltd. 2001
The macroeconomic modelling of carbon tax effects
The macroeconomic modelling of the carbon tax has adopted 3 main approaches
to the problem (see Barker, 1998 and Zhang and Folmer, 1998).
*Dynamic optimisation. Optimising models (e.g. GLOBAL2100 (Manne and
Richels, 1992), POLES (Criqui et al., 2000)) show how energy demand can be
met at least cost, then impose an additional constraint of a restriction on the
CO2emissions to yield a shadow cost, which can be interpreted as the implied
carbon tax to achieve the reduction. The main problem with such models is
common to bottom-up engineering approaches to mitigation, viz. how to deal
with new technologies, not already in the process of development and diffusion.
This limits them to time horizons in which new, unknown technologies will not
substantially affect the outcome. There are other limitations: e.g. there is often
a very aggregate treatment of the effects on energy demand of factors other
than technology, namely the effects of relative price changes of energy carriers
or the effects of real income changes. The models show what the optimised
energy structure will be, given a set of technological options and their costs, but
have difficulty modelling the institutional introduction of the new technologies
and their diffusion over sectors and countries.
*Computable General Equilibrium (CGE). CGE models are organised around
the assumption that the economy is in a state of equilibrium in all markets,
with prices and wages adjusting to match supply and demand. The carbon tax
is introduced as an increase in costs and therefore prices, with revenues
available for distribution by the government sector (if it is included in the
model). The models may be solved with some markets out of equilibrium e.g.
the labour market, as in the GEM-E3 application (Conrad and Schmidt, 1998),
or to give time paths and adjustment to a series of equilibria, but these aspects
of the solution must be imposed by assumption. CGE models are usually
calibrated on consensus parameters, rather than estimated on relevant data.
*Macroeconometric simulation. Models following this approach are estimated
on time-series data, sometimes incorporating a fully specified long-run
solution as part of the model structure (Hargreaves, 1992). The carbon tax is
modelled alongside other indirect taxes as an increase in the price of energy
use according to carbon content. Since the models give short-term solutions
out of equilibrium, the economic responses tend to be smaller than those of
the CGE models, at least in the first few years following the introduction of a
carbon tax.
The effects of a carbon tax on GDP using the CGE and macroeconomic models
depends on their treatment of production. Usually an explicit production function
is included in the model with energy as a factor of production. In this case, the
functional form and the estimated or imposed parameters will determine how
output will change when energy prices rise as a result of a carbon tax or the cost of
buying carbon permits. The way this is done almost always implies that a carbon
tax will have the effect of reducing output and GDP. If the parameters are
#Blackwell Publishers Ltd. 2001
imposed, as they are with nearly all the models, then the extent of the GDP cost
will be a direct result of the size and sign of these parameters. Since there is a wide
range of plausible values for the parameters (see Burniaux et al., 1992 for the
ranges and a discussion) there is also a wide range of computed effects of the
carbon tax on output. In other words, the GDP costs given in the literature are
likely to be the direct result of assumptions rather than the outcome of empirical
research. The costs are significantly affected by the researchers' judgements of
what values to adopt for parameters in the wide range available.
These types of models are often combined together in an integrated assessment
framework and Weyant and Hill (1999) provide a helpful way of categorising the
models actually used for carbon tax studies. Table 5.1 adapts and extends this
categorisation. The models are grouped on 2 dimensions according to their
treatment of their energy and economy components. In both dimensions, the more
aggregate approaches are given first, so the models further across and down the
table are more detailed in their treatment of the energy system and the economy.
The role of assumptions in the modelling
In a comparative study of the results from US models, Repetto and Austin (1997)
from the World Resources Institute (WRI) used econometric regression
techniques to assess the role of assumptions in determining the projected GDP
costs of CO2mitigation. Most of the studies used a carbon tax explicitly or as an
implicit addition to the price of carbon needed to restrict its use. The WRI study
Table 5.1 Types of model used in the GHG mitigation studies.
Energy=carbon model
Fuel Supplies &
Demands by
General Equilibrium
Note: Most of the models (marked *) are global models described in the EMF-16 study (see Weyant
and Hill, 1999). (CPB, 1999) describes WorldScan; (Conrad and Schmidt, 1998) describes GEM-E3;
and (Barker, 1995 and 1998) describes MDM-E3 and E3ME.ENDSource: adapted from Weyant and
Hill (1999)
#Blackwell Publishers Ltd. 2001
covered 162 different predictions from 16 models. The study explains the %
change in US GDP in terms of the CO2reduction target, the number of years to
meet the target, the assumed use of carbon tax revenues (how the revenues are
`recycled' through the economy) and 7 model attributes. It estimates that in the
worst case combining these assumptions and attributes, a 30% reduction in US
baseline emissions by 2020 would cost about 3% of GDP. The corresponding best
case implies an increase of about 2.5% in GDP above the baseline. The total
difference of 5.5 percentage points (pp) of GDP in lower costs can be allocated to
the recycling assumption (1.2pp) and across the attributes:
*CGE models gave lower costs than macroeconometric models (1.7pp)
*the inclusion of averted non-climate change damages, e.g. air pollution
effects (1.1pp)
*whether Joint Implementation and=or international emission permit trading
is included (0.7pp)
*the availability of a constant cost backstop technology (0.5pp)
*the inclusion of averted climate change damages in the model (0.2pp)
*whether the model allows for product substitution (0.1pp)
*and how many primary fuel types are included, so as to allow for interfuel
substitution (0.0pp).
80% of the variation in GDP is explained by these factors, plus the extent of the
CO2target reductions. In summary, worst case results come from using a
macroeconometric model with lump-sum recycling of revenues, no emission
permit trading, no environmental benefits in the model and no backstop
The WRI study is convincing in showing how model approaches and
assumptions can and do influence the results. It reveals the influence of the
model methodology adopted and the importance of the assumption concerning
the recycling of tax revenues. If the published estimates of the macroeconomic
effects of carbon taxes are interpreted in the light of these findings, the results of
carbon taxes for the US and indeed for the implementation of the Kyoto Protocol
may not be as costly as at first sight.
Lump-sum recycling
In the context of CGE modelling, this is the most neutral means of recycling tax
revenues because in theory and by assumption it has no behavioural implications
in the model, although it can have substantial effects on the distribution of
income. The assumption, combined with the usual CGE treatment of the
production structure, has the inevitable outcome that any carbon tax will entail
GDP costs, and these are therefore to be expected in the empirical studies. The
interest is in the magnitude and distribution of these costs. The lump-sum
recycling of carbon tax receipts is not very likely and is certainly sub-optimal.
Jorgenson and Wilcoxen argue: `(Lump-sum recycling) is probably not the most
likely use of the revenue. ... Using the revenue to reduce a distortionary tax would
#Blackwell Publishers Ltd. 2001
lower the net cost of a carbon tax by removing inefficiency elsewhere in the
economy.' (Jorgenson and Wilcoxen, 1993b, p. 20). This is precisely the effect that
Jorgenson and Wilcoxen (1993b, Table 5 p. 22) obtain in their model when they
do in fact reduce distortionary taxes to offset a carbon tax, finding that a 1.7%
GDP loss under lump-sum redistribution is converted to a 0.69% loss and a 1.1%
gain by reducing labour and capital taxes respectively. However, the assumption
does provide a benchmark to compare effects for different countries and other
forms of recycling.
The effects of carbon taxes on GDP: the EMF-16 studies
A series of studies undertaken in a consistent framework of assumptions is
reported in an Energy Modeling Forum (EMF) exercise assessing the costs of
adopting elements of the Kyoto Protocol. Weyant and Hill (1999) summarise the
studies and the results. All the studies use carbon emission permits as the
instrument for mitigation and therefore yield implicit carbon tax rates to achieve
the targets; all assume lump-sum recycling of revenues; and all set aside the
environmental benefits. A consistent range of scenarios is considered by 13
modelling teams, with the emphasis given to how emission permit trading may
reduce costs.
Table 5.2 gives the range of results for the carbon tax rates and GDP changes
from the EMF-16 studies.
The most striking feature of the extremes presented in Table 5.2 is the wide
range of the carbon taxes estimated as needed to reach the Kyoto targets by
domestic policies using an efficient economic instrument. Although several of the
assumptions are consistent across the studies, there is an extra feature (excluded
from the WRI study) that gives rise to differences. The Kyoto target is an absolute
one in relation to a 1990 or 1995 base, whereas the WRI study considers CO2
Table 5.2 Energy Modeling Forum results for the carbon tax and GDP effects in 2010
Top of range for tax rate in
Bottom of range for tax rate in
per tC
% Model
per tC
% Model
USA 410 1.78 Oxford 76 0.42 G-Cubed
OECD-Europe 966 2.08 Oxford 159 0.55 RICE
Japan 1074 1.88 Oxford 97 0.57 G-Cubed
Australia, New
425 1.96 ABARE
145 0.96 RICE
Source: Weyant and Hill (1999, pp. xxxi± xxxiv)
#Blackwell Publishers Ltd. 2001
emissions relative to a base line over the projection period. This means that the
range in the results is partly due to the different regional targets and different rates
of growth of CO2emissions in the base projection.
The effectiveness of the carbon tax in achieving a given target in different
countries depends on the tax and energy systems in place. If energy is already
taxed, then a carbon tax will have to be that much higher in order to push up
prices by a given proportion. If the existing energy system is such that there are
substantial opportunities to switch from carbon-based energy, e.g. substitution
possibilities for a switch from relatively high to relatively low carbon fuels, such as
from coal to gas, then the tax will be lower. Carbon taxes tend to be estimated at
higher levels for Japan and Europe than for the USA, because of the latter's
relatively low energy tax rates.
The range of GDP costs is much narrower than that of the carbon tax rates, but
a high carbon tax does not necessarily imply high GDP costs. The costs in Japan
tend to be lower than in the other regions but there is no strong pattern. However
all the GDP effects are negative. It is very likely that this result comes from the
assumption in these studies that all the revenues from the carbon tax are returned
to the economy by means of lump-sum payments to consumers. This form of
recycling implies that the average consumer has a loss in real income from the
carbon tax compensated by a gain in wealth from the lump-sum repayment. In
most models the loss and the gain, even if they are the same monetary values, have
a net effect of reducing expenditure because spending is modelled as being more
responsive to a fall in income than to an equivalent rise in wealth.
The results from the Oxford model (Cooper et al., 1999) stand out from the rest
in showing very high costs of mitigation. The Oxford study considers the costs of
the US reaching its Kyoto target with a carbon tax (whose revenues are recycled
lump-sum) but without international permit trading, and holding emissions at
their 1990 levels after 2010. It estimates that US GDP is reduced by 4% below the
baseline by 2020, e.g. costs rising to 4% of GDP for the US by 2020. This is a
striking result especially since the Oxford model deals `with very important issues
not addressed elsewhere in the [Energy Journal] volume', e.g. `macroeconomic
adjustment costs' (Weyant and Hill, 1999, p. xlii). Indeed the study itself asserts
that `unlike CGE models, the Oxford model has been subject to statistical
verification and so is capable of explaining accurately the historical data.' (Cooper
at al. 2000, p. 338, italics in original).
The high costs in the Oxford results appear to be due to three features in the
*the choice of assumptions. The assumptions used in the Oxford study appear to
correspond exactly with those identified by Repetto and Austin (1997) as
leading to the most pessimistic outcome for GDP costs for the US (see above).
*the equation used to explain total factor productivity growth. This appears to
be derived from one given by Marion and Svensson (1986, p. 109, footnote
12). It appears to impose high costs of carbon taxation on potential output
growth simply by assuming a value for a substitution elasticity picked from
#Blackwell Publishers Ltd. 2001
the literature. The equation itself is suspect in terms of its derivation and
theoretical validity in this context (see Barker and Ekins, 2001).
*lump-sum recycling in a macroeconometric model. If a carbon tax is
introduced in such a model, then what happens to the revenues becomes
critical to the short-term outcome. This is not so important in CGE models
because the solution is one of long-run equilibrium. The huge revenues from
the carbon tax in the Oxford model are recycled lump-sum to consumers. If
these additions to consumers' income and wealth are largely saved, then the
economy will become progressively depressed as the carbon tax rises in order
to curb the growth in CO2emissions.
The effects of carbon taxes on GDP: other studies with lump-sum recycling
Individual studies for a number of countries considering carbon taxes with lump-
sum recycling support the EMF findings. The general results are that reductions
in CO2emissions of 15% to 25% by 2010 incur GDP or welfare costs of 0.1% to
The problem is that the frame of reference adopted in many of the studies is
inappropriate. They start from the position that the economy is in equilibrium at
maximum welfare and then measure the loss in welfare from introducing a carbon
tax. From this standpoint, a carbon tax is treated as a distortionary tax and is
being compared to an ideal instrument, lump-sum payments, which (in theory) are
non-distortionary, and is then inevitably judged as being costly. However, as
noted earlier, the carbon tax is in essence a means of correcting a distortion arising
out of an externality, i.e. the use of the atmosphere to dispose of waste emissions.
If the standpoint is changed to one in which a political process has determined a
GHG reduction target, as in the draft Kyoto Protocol, then other questions
become relevant. What are the least-cost means of achieving the target? If a
carbon tax is to be introduced, how should its revenues be recycled to improve
economic performance? How does use of economic instruments, such as a carbon
tax, compare with regulation to achieve the GHG reduction target?
The effects of carbon taxes on GDP: revenues recycled via reductions in other taxes
Many of the European studies of the effects of carbon taxes have adopted the
assumption that the revenues are used to reduce employment taxes. This follows
suggestions made by the European Commission for use of revenues in its
proposed carbon=energy tax. Many of the studies confirm the WRI study in
finding that under this assumption GDP rises above the baseline with a carbon tax
(Barker and Kohler, 1998; Capros et al., 1999; Barker, 1999; Bernard and Vielle,
1999; Hourcade et al., 1999). Several studies for Germany adopted the
government's 25% reduction target for 2010 and found that an environmental
tax reform with a carbon tax replacing some employment taxes led to increases in
GDP (Welsch and Hoster, 1995; Schmidt and Conrad, 1996; Boehringer, 1997).
#Blackwell Publishers Ltd. 2001
A study for Australia, (McDougall and Dixon, 1996) confirms the magnitude
and direction of these results. Studies on mitigation costs for China also suggest
that if the tax revenues are recycled, GDP may rise above base. Zhang (1998), with
a CGE model, suggests that if carbon taxes are used to reduce some indirect taxes,
welfare may improve. Garbaccio et al. (1999, 2000) with another dynamic CGE
model for China, but using carbon tax revenues to reduce all other taxes, find that
after the first year GDP is always above base.
One study (Ha
Êkonsen and Mathiesen, 1997) directly compares lump-sum and
employment-tax-reduction as methods of recycling revenues. For a 20% cut in
Norwegian CO2emissions, an index of welfare is reduced by 1% under lump-sum
recycling and 0.3% when social security contributions are used.
The European results differ from most US results when employment taxes are
reduced. In the US, this use of revenues does not generally appear to lead to
increases in GDP above base (Goulder, 1995b; Jorgenson et al., 1995;
Shackleton et al., 1996). An exception is Jorgenson et al. (1999), which finds
that if revenues from auctioned permits are recycled via reductions in personal
income tax, using a detailed, dynamic, general equilibrium model, then GDP is
0.6% above the baseline by 2020. The general result appears to be partly
because the US economy is closer to full employment in the base line and partly
because US employment taxes are lower than those in Europe (OECD Jobs
Study, 1994).
Another result from the Jorgenson model was that, if the revenues are used to
reduce corporate taxes on capital, US GDP is below base by 0.5%. Corporation
tax reductions therefore appear to be less beneficial for the US economy than
reductions in taxes or capital, which (as noted above, Jorgenson and Wilcoxen,
1993b) lend to increases in GDP.
Outside Europe and the US, Garbaccio et al. (1999, 2000), using a dynamic
CGE model for China, consider reductions in CO2emissions of 5%, 10%,
and 15% below baseline with carbon tax revenues recycled by reducing all other
taxes proportionally. In all the scenarios, a very small decline in GDP occurs in
the first year of the simulation, but GDP is above base in every year thereafter.
The magnitude of the double dividend for the European countries is lower in
studies that use CGEs compared with the results from studies that use
macroeconometric models. Barker (1999) directly compares sets of results from
GEM-E3 (a CGE model) and E3ME (an econometric model) and concludes that
the econometric model tends to have stronger effects for a very similar set of
assumptions. Linking this finding with the Repetto and Austin (1997) result
suggests that the key difference between the two modelling approaches in the
carbon tax literature is that the CGE models, without adjustment for short-term
interactions, tend to have smaller feedbacks affecting GDP when the tax is
introduced. Therefore the CGE responses are smaller both if GDP is projected to
fall and if it is projected to rise.
A recent comprehensive study of the effects of achieving the Kyoto targets for
the EU is that by Barker and Rosendahl (2000). The macroeconomic results of 3
scenarios are shown in Table 5.3.
#Blackwell Publishers Ltd. 2001
According to the E3ME model, the tax rates or permit prices in 2010 lie
between 135 and 154 euro (2000) per tonne carbon. Moreover, consistent with
other simulations of the Energy Modelling Forum (see e.g. Weyant, 1999), the
net impact on GDP is quite small, less than 1% from base in all scenarios.
Indeed, in two of the three scenarios the GDP effect is positive. The effects are
very small with inflation higher in all the scenarios, with the fully grandfathered
permit scheme implying the highest price rises. Two scenarios increase
employment by about 1% above base, whereas the second scenario, that is
grandfathered permits with higher profits, shows more or less no change in
employment. Introducing carbon taxes with revenue recycling seems to be
the best policy choice measured in GDP and employment effects. In contrast,
the permit scheme scenario with higher profits seems to be the least
The European studies also give examples of reductions projected in GDP for
small economies taking unilateral action (Proost and van Regemorter, 1995
(Belgium); Andersen et al., 1998 (Denmark); Barker, 1999 (the Netherlands)).
Table 5.3 Macrovariables in EURO-19 for 2010 in the three mitigation scenarios.
Tax rate euro(2000) per tonne carbon 0 153.1 0 153.1
Tax revenue (billion euro) 0 170.1 0 108.4
Permit price euro(2000) per tonne
0 0 135.2 147.8
Permit revenue billion euro 0 0 0 30.7
GDP %pa 2000±10 2.6 2.7 2.6 2.6
GDP% difference from 2010 base 0 0.8 0.3 0.5
GDP cost euro(2000) per tonne carbon
01008.5 355.7 698.6
Ancillary benefits difference as % of
0 0.11 0.10 0.10
Ancillary benefits euro(2000) per tonne
carbon equivalent
0 137.5 126.3 133.0
Employment 2010 million 162.2 163.9 162.1 163.5
Employment % difference from 2010
0 1.1 0.1 0.8
Consumer prices %pa 2000± 10 2.4 2.4 2.5 2.4
Consumer prices % difference from
2010 base
0 0.2 1.4 0.4
Trade balance Percentage point
difference from base
00.2 0 0.1
Government financial balance
percentage point difference from base
01.2 0.2 0.7
Energy profits billion (1990)euros
difference from base
019.2 20.1 0.8
Source: (Barker and Rosendahl, 2000)
#Blackwell Publishers Ltd. 2001
In summary, the literature points to the following factors giving rise to
differences in estimated GDP costs:
*the extent of the required reduction in GHG emissions
*the cost and availability of less carbon-intensive technologies and products
compared to the reference case
*the responsiveness of the production structure and final demand to increases
in real energy prices
*the time period available for adjustment to higher real prices
*the use of revenues from carbon taxes, with the opportunities for negative
costs, i.e. increases in GDP, depending on the existing tax structure
*the coverage of the emission permit trading scheme, with a wider scheme
covering more countries, leading to lower costs.
In addition, there are environmental benefits from reductions in emissions
associated with CO2emissions.
5.4. Distributional implications of carbon taxes
The different impact of carbon taxes, and environmental taxes in general, on the
competitiveness of different sectors is an example of the distributional effects of
the policy instruments. In this case, the differential impact is in accordance with
the polluter pays principle, such that relatively carbon-intensive sectors experience
a higher burden from the tax than sectors using relatively less carbon. However, a
carbon tax may have other distributional effects which are unintended, and may
be undesirable. The effects may be on different regions or income groups within a
country, or, in the case of a global tax regime, between different countries. These
effects are the focus of this section.
In principle it would be desirable to analyse the net distributional effects of the
full range of costs and benefits deriving from the carbon tax. Relevant
considerations include:
1. The increased costs to consumers of purchasing those goods to which the
carbon tax has been directly applied.
2. The increased costs to consumers of purchasing those goods and services
which were produced using a taxed input (both these costs being computed
after taking into account the responses of producers and consumers to the
incentive effects of the tax).
3. The benefits deriving from reduced carbon emissions, and from reductions in
other polluting emissions, as a result of the carbon tax.
4. The distributional situation following the allocation of the revenues from the
carbon tax (which may, of course, be made specifically to counteract the
distributional effects, as was the case in the preceding discussion on
There is also the issue of the baseline against which the costs and benefits should
be evaluated. In particular, should it be the status quo of no action on carbon
#Blackwell Publishers Ltd. 2001
emissions, or should it be relative to other actions on carbon emissions that yield
the same reduction? This point is particularly important because of carbon taxes'
efficiency property of being able to achieve a given carbon reduction at least
overall social cost.
In practice, this full distributional analysis has proved too complex, and the
majority of the studies do not address all these elements, as will be seen, and take
the `no action' situation as their baseline. The third point is largely outside the
scope of this paper altogether. However, it is important to bear all these points in
mind when overall conclusions about a carbon tax's distributional impacts are
being drawn.
The is little doubt with regard to the first of the cost categories above that
carbon taxes levied on domestic fuel tend to be regressive. In Poterba (1991,
Table 3.5, p. 79), household fuel accounted for 7.6% of total expenditure for the
lowest expenditure decile, but only 4.0% for the highest expenditure decile. For
Europe, in Smith (1992, p. 253), fuel expenditures relative to total expenditures
for the lowest expenditure quartile were 1.2% higher than for the highest quartile
in Italy, 1.9% in Spain, 2.7% in Netherlands, 3.4% in France, 3.5% in Germany,
and 5.8% in Ireland. In the UK the difference in expenditure share between the
highest and lowest decile was more than 10% (Smith, 1992, p. 252). In addition,
Johnstone and Alavalapati (1998, p. 9) calculated that in the UK the poorest
income quartile uses relatively more carbon-intensive coal, and less low-carbon
gas, than the richest quartile, so that they would be relatively hard hit by a carbon
tax for this reason as well.
The expenditure distribution of transport fuel is very different, accounting for
the lowest share of total expenditures in the lowest-expenditure quartile in the six
European countries in Smith (1992). In terms of the direct effects on households,
therefore, a carbon tax might be expected to be regressive in respect of domestic
fuel and progressive in respect of transport fuel.
In modelling studies where these effects of the tax on household and transport
fuels are combined with consideration of the second of the four considerations
listed above (the effects of the tax on the prices of other goods and services), the
review in Bruce et al. (1996, p. 420) shows the overall effect to be generally mildly
regressive for OECD countries, a finding also in line with that of Barker and
Èhler (1998) for the Member States of the EU. The regressivity is significantly
less if expenditure rather than income shares are used in the analysis (Poterba,
1991). Poterba (1991, p. 82) also notes that the effects of a carbon tax on asset
markets (especially the stocks of energy-intensive and fossil fuel industries)
would be likely to reduce the tax's regressivity. For developing countries, OECD
1995 (p. 25) reports a single study for Pakistan (Shah and Larsen, 1992), which
suggests that regressivity may be less important in such countries than in the
With regard to the inter-country distribution of costs from a global system of
carbon taxes, Whalley and Wigle (1991) show this, not surprisingly, to be very
dependent on the nature of the tax. A system of national production taxes would
cause revenues to flow from oil consumers to oil producers, benefiting oil
#Blackwell Publishers Ltd. 2001
exporters. A system of national consumption taxes would relatively benefit oil-
consuming countries at the expense of producers. A comparison of models
reported in Dean (1994), in which the same percentage emission reduction was
imposed in five world regions (US, Other OECD, China, ex-USSR, Rest of
World) found no clear pattern except that the Rest of World region generally
experiences the highest relative costs. This outcome should be considered in
conjunction with evidence (for example, Fankhauser, 1995, Table 3.16, p. 55),
that developing countries are expected to bear the highest relative costs from
climate change, and therefore have the most relatively to gain from carbon
The fourth important consideration in respect of the distributional impact of a
carbon tax is what is done with the revenues. The basic position may be simply
expressed. High-income groups or countries use more energy than low-income
groups or countries, and will therefore contribute more in absolute terms to
carbon tax revenues. It is therefore always possible in theory to remove any
regressive effects, if desired, by distributing the revenues such as to benefit
adversely affected groups.
Between countries this is clearly shown in Whalley and Wigle (1991, Table 7.6,
p. 250). Under a system of national production taxes, developing countries have
the highest relative cost, at 7.1% of GDP. With the redistribution of the revenues
from a global production tax on the basis of regional population, developing
countries receive a benefit of 1.8% of GDP, with all other regions experiencing a
net cost.
For households, one way of effecting such a redistribution is through the so-
called `eco-bonus', whereby tax revenues are distributed to households on an
equal per household basis. This would clearly make low-income households better
off than they were before, more than compensating them for any regressive effects
experienced from the tax itself. Such a redistribution is carried out in Switzerland
with the revenues from the taxes on light fuel oil, sulphur and volatile organic
compounds (VOCs), through a per capita reduction in medical insurance (EC,
1999). Equivalent in its distributional effect to an eco-bonus is the granting of a
tax-free allowance of energy use to all households, as practised by the Netherlands
(see below).
Alternatively governments may wish to target the redistribution of tax revenues
on adversely affected groups, through the welfare or benefit system or in some other
way (although it should be noted that it may be administratively difficult to reach
all those in these groups). For the UK Symons et al. (1994) show that, for the two
lowest decile expenditure groups, benefits reform can change an 18% and 15%
reduction in disposable expenditure from a carbon tax (with no revenue recycling)
into a 54% and 25% increase for the same groups. In practice governments are
likely to seek to distribute only part of the tax revenues in this way.
The most sophisticated rebate system aimed at avoiding regressive effects from
energy taxation is that associated with the small energy users' tax in the
Netherlands. This combines a tax-free allowance for electricity and natural gas
with personal income tax reductions for individuals, and reductions in social
#Blackwell Publishers Ltd. 2001
security contributions and other tax reductions for businesses (EC, 1999).
Calculations prior to the introduction of the tax showed that such measures were
able to offset potentially regressive effects completely and also to prevent large
transfers between tax-paying groups.
It is therefore clear that, while a carbon tax could have a regressive effect on
low-income households (or countries), it does not need to do so. The overall
desirability of a carbon tax, as well as its likely political and ethical acceptability,
depend on regressive effects being avoided.
5.5. The costs of climate stability
The objective of the UN Framework Convention on Climate Change is to prevent
dangerous anthropogenic interference with the global climate. Climate scientists
have not yet converted this objective into a maximum atmospheric concentration
of CO2(and other greenhouse gases), but have projected a number of different
carbon emissions scenarios which result in different atmospheric concentrations.
The pre-industrial concentration was about 280ppmv (parts per million by
volume), and in 1994 was 358ppmv. The lowest of the IPCC's projections involved
stabilisation of the carbon concentration at 450ppmv, concerning which the IPCC
estimated: `For stabilisation at 450ppmv (parts per million by volume) in the 1994
calculations fossil emissions had to be returned to about a third of today's levels
(i.e. to about 2GtC=year) by the year 2200.' (Houghton et al., 1996, p. 80) It is
therefore worth enquiring what the costs of such carbon reduction might be.
In the study cited above, Repetto and Austin (1997) find that, with their most
unfavourable assumptions, a 50% reduction in US baseline emissions by 2020 (the
kind of cut that would be necessary to get on the IPCC 450ppmv trajectory) was
projected to cost about 6% of GDP. With favourable assumptions there would be
a modest increase in GDP relative to the baseline.
For another estimate, the EMF modelling exercise discussed earlier (Gaskins
and Weyant, 1993) found that the average GDP loss from a 50% emissions cut by
2050 was about 3% of GDP. However, this does not take account of the factors
reviewed above which could substantially reduce this cost:
*the 35% to over 100% that can be offset by recycling the revenues in such a
way that distorting taxes are reduced;
*the >45% (minimum) reduction in emissions (by 2025=2030) by implement-
ing energy efficiency technologies.:
*in addition, 4± 18% of global CO2emissions could be cut with increases in
output by phasing out fossil fuel subsidies (Bruce et al., 1996, p. 73);
*and it is now well established that the ancillary benefits of CO2abatement
(the reduction in air pollutants apart from CO2) substantially reduce the
gross costs of such abatement; Ekins (1996) has calculated that, even if only
the ancillary benefits from SO2-reduction are considered, 25 ±50% of the
CO2abatement brought about by a $100=tC carbon tax would be achieved at
negative net cost.
#Blackwell Publishers Ltd. 2001
Combining these offsets and no-regrets options makes it seem unlikely that the
EMF 50% reduction would cost anything at all.
Looking beyond 2050, whether further reductions in emissions will incur more
costs depends most importantly on whether backstop energy technologies will
have been developed which are competitive with the by-then substantially
depleted fossil fuels. Nordhaus, in his DICE model, rather pessimistically assumes
that they will not, and projects several scenarios with rising costs of abatement
through to the next century. These suggest that, by 2100, the climate stabilisation
scenario (which restricts the global average temperature rise to 1.5 C above the
level in 1900) would cost about 8% of per capita consumption, reducing it from
about $9,500 in the no-controls case to $8,750. This scenario involves
substantially earlier stabilisation than the IPCC 450ppmv case (which was by
2200), and therefore much more ambitious emission reduction up to 2100.
Nordhaus (1994, Figure 5.2, p. 87) has CO2-equivalent emissions falling
practically to zero by 2035 and staying below about 1.5GtC (compared to
current levels of about 7GtC) per annum thereafter.
The cost estimate of 8% of global consumption by 2100, as with the EMF
estimates, does not contain the offsets from revenue-recycling, energy efficiency or
removing subsidies, nor any considerations of risk aversion or secondary benefits.
It is also pessimistic about the possibility of stimulating the development of a
backstop technology, such as solar power. Taking such considerations into
account would reduce the cost of climate stabilisation very substantially.
6. Conclusions
This paper has shown that there is now a very large literature on carbon taxes and
carbon emission permits, which this paper has, despite its length, only been able to
survey in general terms. Nevertheless a number of conclusions about these
instruments have now clearly emerged from the literature. These may be
summarised as follows.
First there is general agreement that market-based instruments of carbon
control will achieve a given level of emissions reduction at lower cost than
regulations. Among market-based instruments, those that raise revenue, and
recycle it through the economy by reducing pre-existing distortionary taxes, will
have lower costs than those that do not raise revenue. This conclusion clearly
favours carbon taxes and auctioned emission permits (with revenue recycling as
described) over grandfathered emission permits.
Notwithstanding the perceived advantages of carbon taxes and auctioned
emission permits, they have to date only been introduced in relatively few
countries, and then at relatively low levels. This is largely because the
distributional effects of the instruments have made them politically problematic.
These distributional effects have also received considerable attention in the
The most important of the distributional effects is that on industrial
competitiveness, such that energy-intensive sectors, which are often perceived as
#Blackwell Publishers Ltd. 2001
industrially important, emerge as significant losers. The result of this has been that
carbon taxes have been introduced with substantial exemptions for these sectors,
and grandfathered permits have been favoured over auctioned permits. While this
may have allayed concerns over the competitiveness of the relevant sectors, the
literature clearly shows that these choices have significant economic costs for
society as a whole.
Other potential distributional effects of concern, especially from a carbon tax,
are those on low-income groups. It is clear that revenues from the tax can be
devoted to removing regressive effects, although this may be institutionally
complex to achieve. It also means, of course, that revenues used for this purpose
cannot be used to gain the benefits of the revenue-recycling effect by reducing the
rates of distortionary taxes.
There is a very wide range of projections of the costs of imposing a carbon tax.
The paper has showed how these projections depend on a number of crucial
assumptions about the rate of development of non-carbon fuels, the flexibility of
the economy, the way the revenues are used and other factors. It is this paper's
conclusion that, with realistic assumptions on these matters, the costs of attaining
the Kyoto targets, in particular, need not be high. However, this is still a matter of
some controversy.
It is still early days to come to firm conclusions about the environmental
effectiveness of carbon taxes, but studies of their effects where they have been
introduced have broadly concluded that they have achieved the projected level of
carbon reduction. Their effectiveness can be enhanced by dedicating some of the
carbon tax revenues to stimulating further carbon control measures, though this
means that fewer revenues are then available to reduce other tax rates or
compensate for regressive distributional effects. Such trade offs are, of course,
common in complex areas of policy.
The overall conclusion of this paper is that, assuming that concern over
anthropogenic climate change continues to increase, the advantages of carbon
taxes and auctioned permits over other means of carbon abatement will lead them
to be progressively introduced for this purpose. However, the distributional
concerns will mean that their introduction is slower, and departs further from the
first-best means of their implementation, than would otherwise be the case.
1. There has been an extensive controversy on the nature and existence in theory and
practice of the double dividend, since the concept was first clearly set out by Pearce
(1991). Much of the theoretical literature is confusing and misleading in that it sets aside
the essential character of the carbon tax as an environmental tax. In particular the
distinction between the `weak' and the `strong' form of the double dividend brought
about by spending revenues from a carbon tax, introduced by Goulder (1995a), is not
helpful as explained by Bohm (1997, p. 114). More recently, Sanstad and Wolff (2000)
have elucidated some of the more refined theoretical niceties in the debate and have
#Blackwell Publishers Ltd. 2001
explored why one empirical model might yield a higher double dividend for the US than
Andersen, F. M., Jacobsen, H. K., Morthorst, P. E., Olsen, A., Rasmussen, M., Thomsen,
T. and Trier, P. (1998) EMMA: En energi- og miljùrelateret satellitmodel til ADAM
(EMMA an energy an environmental related satellite model to ADAM). Nationalù-
konomisk Tidsskrift, 136, 3, 333± 349.
Ballard, C. L., Shoven, J-B. and Whalley, J. (1985) General equilibrium computations of
the marginal welfare costs of taxes in the United States. American Economic Review,
75, 1, March, 128± 138.
Baranzini, A., Goldemberg, J. and Speck, S. (2000) A future for carbon taxes. Ecological
Economics, 32, 395± 412.
Barker, T. (1995) Taxing pollution instead of employment: greenhouse gas abatement
through fiscal policy in the UK. Energy and Environment, 6, 1, 1 ± 28.
Barker, T. (1998) The effects on competitiveness of coordinated versus unilateral fiscal
policies reducing GHG emissions in the EU: an assessment of a 10% reduction by
2010 using the E3ME model. Energy Policy, 26, 14, 1083 ± 1098.
Barker, T. (1999) Achieving a 10% cut in Europe's CO2emissions using additional excise
duties: coordinated, uncoordinated and unilateral action using the econometric model
E3ME. Economic Systems Research, 11, 4, 401± 421.
Barker, T. and Ekins, P. (2001) How high are the costs of Kyoto for the US Economy?,
mimeo, Department of Applied Economics, University of Cambridge, Cambridge.
Barker, T. and Rosendahl, K. E. (2000) Ancillary benefits of GHG mitigation in Europe:
SO2, NOx and PM10 reductions from policies to meet Kyoto targets using the E3ME
model and EXTERNE valuations, Ancillary Benefits and Costs of Greenhouse Gas
Mitigation, Proceedings of an IPCC Co-Sponsored Workshop, March 2000, OECD,
Barker, T., Ekins, P. and Johnstone, N. (1995a) Introduction. In Barker, T., Ekins, P. and
Johnstone, N. (eds), 1995b Global Warming and Energy Demand (pp. 1± 16).
Routledge, London=New York.
Barker, T., Ekins, P. and Johnstone, N. eds (1995b) Global Warming and Energy Demand.
Routledge, London=New York.
Barker, T. and Ko
Èhler, J. (1998) International Competitiveness and Environmental Policies.
Edward Elgar, Cheltenham.
Baumol, W. (1972) On taxation and the control of externalities. American Economic
Review, 62, 3, 307± 321.
Baumol, W. and Oates, W. (1971) The use of standards and prices for the protection of the
environment. Swedish Journal of Economics, March, 73, 42 ± 54.
Baumol, W. and Oates, W. (1988) (2nd edition) The Theory of Environmental Policy.
Cambridge University Press, Cambridge.
Bernard, A. L. and Vielle, M. (1999) Rapport a
Ála Mission Interministe
Ârielle sur l'Effet de
Serre relatif aux e
Âvaluation du Protocole de Kyoto effectue
Âes avec le mode
E3. Ministe
Áre de l'e
Âquipement des transports et du logement. Commissariat a
Boehringer, C. (1997) NEWAGE Ð Modellinstrumentarium zur gesamtwirtschaftlichen
Analyse von Energie- und Umweltpolitiken. In S. Molt and Fahl U. (eds), Energiemodelle
in der Bundesrepublik Deutschland Ð Stand der Entwicklung.Juelich,p.99±122.
Boero, G., Clarke, R. and Winters, L. (1991) The Macroeconomic Consequences of
Controlling Greenhouse Gases: a Survey. Department of the Environment, HMSO,
#Blackwell Publishers Ltd. 2001
Bohm, P. (1997) Environmental taxation and the double dividend: fact or fallacy? In Tim
O'Riordan (ed.), Ecotaxation (pp. 106± 124). Earthscan Publications, London.
Bohm, P. (1999) An emission quota trade experiment among four Nordic countries. In S.
Sorrell and J. Skea (eds), Pollution for Sale: Emissions Trading and Joint
Implementation (pp. 299±321). Edward Elgar, Cheltenham.
Èhringer, C. and Rutherford, T. F. (1997) Carbon taxes with exemptions in an open
economy: a general equilibrium analysis of the German tax initiative. Journal of
Environmental Economics and Management, 32, 189± 203.
Bovenberg, A. L. and de Mooij, R. A. (1994) Environmental levies and distortionary
taxation. American Economic Review, 94, 4, September, 1085 ± 1089.
Bovenberg, L. and Goulder, L. (1996) Optimal environmental taxation in the presence of
other taxes: general equilibrium analyses. American Economic Review, 86, 985 ± 1000.
Bovenberg, L. and Goulder, L. (2000) Neutralizing the adverse industry impacts of CO2
abatement policies: what does it cost? In C. Carraro and G. Metcalf (eds), Behavioral
and Distributional Impacts of Environmental Policies. University of Chicago Press,
Brendemoen, A. and Vennemo, H. (1994) A climate treaty and the norwegian economy: a
CGE assessment. The Energy Journal, 15, 1, 77 ± 93.
Bruce, J., Lee, H. and Haites, E. eds (1996) Climate Change 1995: Economic and Social
Dimensions of Climate Change. Contribution of Working Group III to the Second
Assessment Report of the Intergovernmental Panel on Climate Change (IPCC),
Cambridge University Press, Cambridge.
Burniaux, J. M., Martin, J. P., Nicoletti, G. and Oliveira-Martins, J. (1992) GREEN- A
Multi-Region Dynamic General Equilibrium Model for Quantifying the Costs of Curbing
CO2Emissions: A Technical Manual. Working Paper 116. Economics and Statistics
Department, OECD, Paris.
CPB Netherlands Bureau for Economic Policy Analysis (1999) WorldScan: the Core
version. CPB, The Netherlands.
Cline, W. R. (1992) The Economics of Global Warming. Institute for International
Economics, Washington DC.
Cooper, A., Livermore, S., Rossi, V., Walker, J. and Wilson A. (1999) Economic impacts of
reducing carbon emissions: The Oxford model. The Energy Journal, Special Issue,
Conrad, K. and Schmidt, F. N. (1998) Economic impacts of an uncoordinated versus a
coordinated carbon dioxide policy in the European Union: An applied general
equilibrium analysis. Economic Systems Research, 10, 2, 161± 182.
Cramton, P. and Kerr, S. (1998) Tradable permit auctions: how and why to auction not
grandfather. Discussion Paper 98± 34, Resources for the Future, Washington DC,
Criqui, P., Kouvaritakis, N. and Schrattenholzer, L. (2000) The impacts of carbon
constraints on power generation and renewable energy technologies. In L. Bernstein
and J. Pan, Sectoral Economic Costs and Benefits of GHG Mitigation: Proceedings of
an IPCC Expert Meeting, 14 ± 15 February 2000, Technical Support Unit, IPCC
Working Group III, RIVM, The Netherlands.
Dean, A. (1994) Costs of cutting CO2emissions: evidence from `top-down' models. In
OECD (Organisation for Economic Cooperation and Development) 1995 The
Economics of Climate Change: Proceedings of an OECD=IEA Conference, OECD,
Dean, J. (1992) Trade and the environment: a survey of the literature. In P. Low (ed.)
International Trade and the Environment (pp. 15± 28). World Bank Discussion Paper
159, World Bank, Washington D.C.
DEPA (Danish Environmental Protection Agency) (1999) Economic Instruments in
Environmental Protection in Denmark, DEPA=Ministry of Environment and Energy,
#Blackwell Publishers Ltd. 2001
Diamond, P. and Mirrlees, J. (1971) Optimal taxation and public production, I: production
efficiency and II: tax rules. American Economic Review, 61, 8 ± 27, 261 ±278.
DIW (1994) Wirtschaftliche Auswirkungen einer o
Èkologischen Steuerreform, Berlin.
EC (European Commission) (1994) Taxation, employment and environment: fiscal reform
for reducing unemployment. Study 3, European Economy, 56, 137±177, Directorate-
General for Economic and Financial Affairs, European Commission, Brussels.
EC (European Commission) (1999) Database on Environmental Taxes in the European
Union Member States, plus Norway and Switzerland, Evaluation of Environmental
Effects of Environmental Taxes, European Communities, Office for Official Publica-
tions of the European Communities, Luxembourg.
EC (European Commission) (2000) Green Paper on greenhouse gas emissions trading
within the European Union, EC, Brussels,
Ekins, P. (1996) How large a carbon tax is justified by the secondary benefits of CO2
abatement? Resource and Energy Economics, 18, 2, 161 ± 187.
Ekins, P. (2000) Economic Growth and Environmental Sustainability: the Prospects for Green
Growth. Routledge, London=New York.
Ekins, P. and Speck, S. (1998) The impacts of environmental policy on competitiveness:
theory and evidence. In T. Barker and J. Ko
Èhler (1998) International Competitiveness
and Environmental Policies (pp. 33± 69). Edward Elgar, Cheltenham.
Ekins, P. and Speck, S. (1999a) Competitiveness and exemptions from environmental taxes
in Europe. Environmental and Resource Economics, 13, 4, 369± 395.
Ekins, P. and Speck, S. (1999b) Evaluation of environmental effects of environmental taxes.
In EC 1999 Database on Environmental Taxes in the European Union Member States,
plus Norway and Switzerland, Evaluation of Environmental Effects of Environmental
Taxes (pp. 1±23). European Communities, Office for Official Publications of the
European Communities, Luxembourg.
ENDS (1999a) Carbon tax boosts Italian fossil fuel prices. ENDS Daily email information
service, January 19, ENDS, London.
ENDS (1999b) Go-ahead for Danish power sector CO2quotas. ENDS Daily email
information service, May 28, ENDS, London.
ENDS (2000) EU backs first Kyoto trading scheme. ENDS Daily email information
service, March 29, ENDS, London.
Fankhauser, S. (1995) Valuing Climate Change: the Economics of the Greenhouse.
Earthscan, London.
Farrow, S. (1995) The dual political economy of taxes and tradable permits. Economic
Letters, 49, 217±220.
Frandsen, S. E., Hansen, J. V. and Trier, P. (1995) GESMEC Ð En generel ligevúgtsmodel
for Danmark. Dokumentation og anvendelser (GESMEC Ð An applied general
equilibrium model for Denmark Ð Documentation and applications). The Secretariat of
the Economic Council, Copenhagen.
Garbaccio, R. F., Ho, M. S. and Jorgenson, D. W. (1999) Controlling carbon emissions in
China. Environment and Development Economics, 4, 4, 493± 518.
Garbaccio, R. F., Ho, M. S. and Jorgenson, D. W. (2000) The health benefits of controlling
carbon emissions in China. Ancillary Benefits and Costs of Greenhouse Gas Mitigation.
OECD, Paris.
Gaskins, D. W. and Weyant, J. P. (1993) Model comparisons of the costs of reducing CO2
emissions. American Economic Review (AEA Papers and Proceedings), 83, 2, May,
2-skatter, dobbelt-dividende og
konkurrence i energisektoren: Anvendelser af den danske AGL-model ECOSMEC
(CO2taxes, double dividend and competition in the energy sector: Applications of
the Danish CGE model ECOSMEC)', Arbejdpapir 1999:1, Danish Economic
#Blackwell Publishers Ltd. 2001
Goulder, L. H. (1995a) Environmental taxation and the double dividend: a reader's guide.
International Tax and Public Finance, 2, 157± 183.
Goulder, L. H. (1995b) Effects of carbon taxes in an economy with prior tax distortions: an
intertemporal general equilibrium analysis. Journal of Environmental Economics and
Management, 29, 271±297.
Goulder, L. (2000) Confronting the adverse industry impacts of CO2abatement policies:
what does it cost?. Climate Issues Brief No.23, September, Resources for the Future,
Washington DC,
Goulder, L., Parry, I. and Burtraw, D. (1997) Revenue-raising versus other approaches to
environmental protection: the critical significance of pre-existing tax distortions.
RAND Journal of Economics, 28, 4, Winter, 708 ± 731.
Grubb, M. (1995) Asymmetrical price elasticities of energy demand. In T. Barker, P. Ekins
N. Johnstone (eds), 1995 Global Warming and Energy Demand (pp. 305± 310).
Routledge, London=New York.
Grubb, M., Chapuis, T. and Ha Duong, M. (1995) The economics of changing course:
implications of adaptability and inertia for optimal climate policy. Energy Policy, 23,
4=5, 417±432.
Hargreaves, C. (1992) Macroeconomic Modelling of the Long Run. Edward Elgar,
Êkonsen, L. and Mathiesen, L. (1997) CO2-stabilization may be a `no-regrets' policy.
Environmental and Resource Economics, 9, 2, 171± 98.
Houghton, J., Meira Filho, L., Callander, B., Harris, N., Kattenburg, A. and Maskell, K.
eds (1996) Climate Change 1995: the Science of Climate Change, contribution of
Working Group I to the Second Assessment Report of the Intergovernmental Panel on
Climate Change (IPCC), Cambridge University Press, Cambridge.
Ikwue, A. and Skea, J. (1996) The energy sector response to European combustion emission
regulations. In F. Le
Âveà que 1996 Environmental Policy in Europe: Industry, Competition
and the Policy Process (pp. 75 ± 111). Edward Elgar, Cheltenham, UK.
INFRAS (1996) Economic Impact Analysis of Ecotax Proposals: Comparative Analysis of
Modelling Results, Final Report of a project conducted in co-operation with the 3rd
Framework Programme of DGXII of the European Commission, INFRAS, Zurich.
Ingham, A. and Ulph, A. (1991) Market-based instruments for reducing CO2emissions: the
case of UK manufacturing. Energy Policy, 19, 3, March, 138 ± 148.
Ingham, A., Maw, J. and Ulph, A. (1992) Energy conservation in UK manufacturing: a
vintage model approach. In D. Hawdon (ed.) 1992 Energy Demand: Evidence and
Expectations, Surrey University Press, Guildford, 115± 141.
Jackson, T. (1991) Least-cost greenhouse planning. Energy Policy, January=February,
Jackson, T. (1995) Price elasticity and market structure Ð overcoming obstacles to ensure
energy efficiency. In T. Barker, P. Ekins and N. Johnstone (eds), Global Warming and
Energy Elasticities (pp. 254± 266). Routledge, London.
Jackson, T. and Jacobs, M. (1991) Carbon taxes and the assumptions of environmental
economics. In T. Barker (ed.), Green Futures for Economic Growth (pp. 49± 67).
Cambridge Econometrics, Cambridge.
Jenkins, R. (1998) Environmental regulation and international competitiveness: a review of
literature and some european evidence. UNU=INTECH Discussion Paper #9801,
January, UNU=INTECH, Maastricht.
Jenkins, R. (2001) Trade, investment and industrial pollution: lessons from South-East
Asia. In N. Adger, M. Kelly and N. Hun Ninh (eds), 2001 Living with Environmental
Change. Routledge, London=New York.
Jensen, J. and Rasmussen, T. N. (1998) Allocation of CO2emission permits: a general
equilibrium analysis of policy instruments. The MobiDK Project, The Ministry of
Business and Industry, Copenhagen.
Jerkkola, J., Kinnunen, J. and Pohjola, J. (1993) A CGE model for Finnish environmental
#Blackwell Publishers Ltd. 2001
and energy policy analysis: effects of stabilizing CO2emissions. Discussion paper No. 5,
Helsinki School of Economics. Helsinki.
Johnsen, T. A., Larsen, B. M. and Mysen, H. T. (1996) Economic impacts of a CO2tax. In
K. H. Alfsen, T. Bye and E. Holmùy (eds), MSG-EE: An Applied General Equilibrium
Model for Energy and Environmental Analyses. Social and Economic Studies 96,
Statistics Norway.
Johnstone, N. and Alavalapati, J. (1998) The distributional effects of environmental tax
reform. Discussion Paper 98± 01, IIED (International Institute for Environment and
Development), London.
Jorgenson, D. W., Goettle, R. J., Gaynor, D., Wilcoxen, P. J. and Slesnick, D. T. (1995)
Social cost energy pricing, tax recycling and economic change. Final Report submitted
to the US EPA, EPA contract No. 68-W2-0018.
Jorgenson, D. W., Goettle, R. J., Slesnick, D. T. and Wilcoxen, P. J. (1999) Carbon
mitigation, permit trading and revenue recycling. Harvard-Japan Project on Energy
and the Environment 1998±99, March 1999.
Jorgenson, D. and Wilcoxen, P. (1993a) Reducing U.S. carbon dioxide emissions: an
assessment of different instruments. Journal of Policy Modeling, 15, 5&6, 491 ± 520.
Jorgenson, D. and Wilcoxen, P. (1993b) Reducing US carbon emissions: an econometric
general equilibrium assessment. Resource and Energy Economics, 15, 1, March, 7 ± 25.
Jorgenson, D. and Yun, K-Y. (1990) The excess burden of taxation in the US. Harvard
Institute of Economic Research Discussion Paper 1528, November, Harvard
University, Cambridge MA.
Kristensen, J. P. (1996) Environmental taxes, tax reform and the internal market Ð some
Danish experiences and possible community initiatives. In Environmental Taxes and
Charges: NATIONAL Experiences and Plans (pp. 121±136). European Foundation
for the Improvement of Living and Working Conditions, Dublin, and Office for
Official Publications of the European Communities, Luxembourg.
Larsen, B. N. and Nesbakken, R. (1997) Norwegian Emissions of CO2 1987 ±1994.
Environmental and Resource Economics, 9, 275± 290.
Lovins, A. B. and Lovins, H. L. (1991) Least cost climatic stabilization. Annual Review of
Energy and Environment, 16, 433± 531.
Low, P. and Yeats, A. (1992) Do `dirty' industries migrate? In P. Low (ed.) International
Trade and the Environment (pp. 89± 103). World Bank Discussion Paper No.159,
World Bank, Washington DC.
Lucas, R., Wheeler, D. and Hettige, H. (1992) Economic development, environmental
regulation and the international migration of toxic industrial pollution: 1960 ± 88. In P.
Low (ed.) 1992 International Trade and the Environment (pp. 67± 86). World Bank
Discussion Paper 159, World Bank, Washington DC.
Manne, A. (1993) International trade: the impacts of unilateral carbon emissions limits.
Paper presented at the OECD International Conference on the Economics of Climate
Change, Paris, June 14±16.
Manne, A. and Richels, R. (1992) Buying greenhouse insurance: the economic costss of CO2
emission limits. MIT Press, Cambridge, Mass.
Marion, N. and Svensson, L. (1986) The terms of trade between oil importers. Journal of
International Economics, 20, 99± 113.
McDougall, R. A. and Dixon, P. (1996) Analysing the economy-wide effects of an energy
tax: results for Australia from the ORANI-E model. In W. J. Bouma, G. I. Pearman
and M. R. Manning (ed.), Greenhouse: Coping with Climate Change (pp. 607 ± 619).
Wellington New Zealand.
Nilsson, C. (1999) Unilateral versus multilateral carbon dioxide tax implementations Ð a
numerical analysis with the European model Ð GEM-E3. Preliminary version,
National Institute of Economic Research, Stockholm.
Nordhaus, W. D. (1993) Optimal greenhouse gas reductions and tax policy in the `DICE'
model. American Economic Review (AEA Papers and Proceedings), 83, 2, May, 313±317.
#Blackwell Publishers Ltd. 2001
Nordhaus, W. D. (1994) Managing the Global Commons: the Economics of Climate Change.
MIT Press, Cambridge MA.
Èm, H. and Vaughan, S. (1999) Trade and Environment. Special Studies 4, World
Trade Organisation (WTO), Geneva.
Norwegian Green Tax Commission (1996) Policies For A Better Environment and High
Employment, 1996, Oslo.
OECD (Organisation for Economic Cooperation and Development) (1994) The OECD
Jobs Study Ð Evidence and Explanations. OECD, Paris.
OECD (Organisation for Economic Cooperation and Development) (1995) Climate
Change, Economic Instruments and Income Distribution. OECD, Paris.
OECD (Organisation for Economic Cooperation and Development) (1996) Implementation
Strategies for Environmental Taxes. OECD, Paris.
OECD (Organisation for Economic Cooperation and Development) (1997) Evaluating
Economic Instruments for Environmental Policy. OECD, Paris.
Oliveira-Martins, J., Burniaux, J.-M. and Martin J. P. (1992) Trade and the effectiveness of
unilateral CO2-abatement policies: evidence from GREEN. OECD Economic Studies,
19, 123±140, OECD, Paris.
Palmer, K., Oates, W. and Portney, P. (1995) Tightening environmental standards: the benefit-
cost or the no-cost paradigm? Journal of Economic Perspectives, 9, 4, Fall, 119± 132.
Parry, I. (1995) Pollution taxes and revenue recycling. Journal of Environmental Economics
and Management, 29, S64±S77.
Parry, I. (1997) Environmental taxes and quotas in the presence of distorting taxes in factor
markets. Resource and Energy Economics, 19, 203±220.
Parry, I., Williams, R. and Goulder, L. (1999) When can carbon abatement policies
increase welfare? The fundamental role of distorted factor markets. Journal of
Environmental Economics and Management, 37, 52± 84.
Pearce, D. (1991) The role of carbon taxes in adjusting to global warming. Economic
Journal, 101, 938±948.
Pezzey, J. (1992a) The symmetry between controlling pollution by price and controlling it
by quantity. Canadian Journal of Economics, 25, 4, 983 ± 991.
Pezzey, J. (1992b) Analysis of unilateral CO2control in the european community and
OECD. The Energy Journal, 13, 159± 171.
Pigou, A. C. (1932) The Economics of Welfare, 4th edition (1st edition 1920), Macmillan,
Pizer, W. (1999) Choosing price or quantity controls for greenhouse gases. Climate Issues
Brief No.17, Resources for the Future, Washington DC,
Porter, M. (1990) The Competitive Advantage of Nations. Free Press, New York.
Poterba, J. (1991) Tax policy to combat global warming. In R. Dornbusch and J. Poterba
(eds), 1991 Global Warming: Economic Policy Responses (pp. 71± 98). MIT Press,
Cambridge MA.
Proost, S. and Van Regemorter, D. (1995) The double dividend and the role of inequality
aversion and macroeconomic regimes. International Taxes and Public Finances,2,
217±219, Kluwer Academic Publishers.
RCEP (Royal Commission on Environmental Pollution) (2000) Energy Ð the Changing
nd Report of the Royal Commission of Environmental Pollution, The
Stationery Office, London.
Repetto, R. and Austin, D. (1997) The Costs of Climate Protection: a Guide for the
Perplexed, World Resources Institute, Washington D.C.
Rutherford, T. (1992) The welfare effects of fossil carbon restrictions: results from a
recursively dynamic trade model. OECD Dept. of Economic Statistics, Resource
Allocation Division, Working Paper No.112, OECD=GD (91), Paris.
Sanstad, A. and Wolff, G. (2000) Tax Shifting and the Likelihood of Double Dividends:
Theoretical and Computational Issues, Redefining Progress,
#Blackwell Publishers Ltd. 2001
Schelling, T. (1992) Some economics of global warming. American Economic Review, 82, 1,
March 1992, 1±14.
Schmidt, T. and Conrad, K. (1995) National economics impacts of an EU environmental
policy Ð an applied general equilibrium analysis. ZEW Discussion Paper No. 95± 22,
SEPA (Swedish Environmental Protection Agency) (1997) Environmental Taxes in Sweden
Ð Economic Instruments of Environmental Policy, Report 4745, Stockholm.
Shackleton, R., Shelby, M., Cristofaro, A., Brinner, R., Yanchar J., Goulder, L., Jorgenson
D., Wilcoxen, P. and Pauly, P. (1996) The efficiency value of carbon tax revenues. In
D. Gaskins and J. Weyant (eds), Reducing Global Carbon Dioxide Emissions: Costs and
Policy Options. Stanford University Energy Modeling Forum, Stanford, CA.
Shah, A. and Larsen, B. (1992) Carbon taxes, the greenhouse effect, and developing
countries. Background Paper No.6, for the 1992 World Development Report, World
Bank, Washington DC.
Smith, S. (1992) The distributional consequences of taxes on energy and the carbon content
of fuels. European Economy (pp. 241± 268). Special Edition on The Economics of
Limiting CO2Emissions, Commission of the European Communities.
Sorrell, S. and Skea, J. eds (1999) Pollution for Sale: Emissions Trading and Joint
Implementation, Edward Elgar, Cheltenham.
Stiglitz, J. and Dasgupta, P. (1971) Differential taxation, public goods, and economic
efficiency. Review of Economic Studies, 38, 2, 151± 174.
Symons, E. Proops, J. and Gay, P. (1994) Carbon taxes, consumer demand and carbon
dioxide emissions: a simulation analysis for the UK. Fiscal Studies, 15, 2, 19 ±43.
TemaNord (1994) The Use of Economic Instruments in Nordic Environmental Policy.
TemaNord 1994:561, Copenhagen.
UNFCCC (UN Framework Convention on Climate Change) (1992) the full text of the
Convention is available on, UNFCCC,
Weitzmann, M. (1974) Prices vs. Quantities. Review of Economic Studies, 41, 477 ± 491.
Welsch, H. and Hoster, F. (1995) A general equilibrium analysis of European
carbon=energy taxation Ð model structure and macroeconomic results. Zeitschrift
fuer Wirtschafts- und Sozialwissenschaften (ZWS), 115, 1995, 275± 303.
Weyant, J. P. and Hill J. N. (1999) Introduction and overview. pp. vi to 1xliv in The Costs
of the Kyoto Protocol: A Multi-Model Evaluation, Special Issue of the Energy Journal.
John Weyant (ed.).
Whalley, J. and Wigle, R. (1991) The international incidence of carbon taxes. In R.
Dornbusch and J. Poterba (eds), 1991 Global Warming: Economic Policy Responses
(pp. 233±263). MIT Press, Cambridge MA.
Zhang, Z.X (1998) Macro-economic and sectoral effects of carbon taxes: a general
equilibrium analysis for China. Economic Systems Research, 10 (2) 135 ± 160.
Zhang, Z.X and Folmer, H. (1998) Economic modelling approaches to cost estimates for
the control of carbon dioxide emissions. Energy Economics, 20, 1, 101± 120.
#Blackwell Publishers Ltd. 2001
... Focusing on Austria, Eisner et al. (2021) find a regressive effect for electricity and heating, but a disproportionate tax burden on middle-income households for transport fuels. Elkins and Baker (2001) explain these patterns by the different expenditure distributions of transport fuels and domestic fuels, i.e. heating fuels and electricity. Support for this finding is provided by Büchs and Schnepf (2013) and Callan et al. (2009). ...
... Berry, 2019;Martini, 2009). A broader picture of the distributional effects of carbon taxes on transport fuels is provided by literature surveys (Elkins and Baker, 2001;Pizer and Sexton, 2019;Speck, 1999;Wang et al., 2016;Zhang and Baranzini, 2004) and meta analyses (Alvarez, 2019;Ohlendorf et al., 2018). Their findings question the universal validity of a ''regressivity assumption''. ...
We introduce a new microsimulation model built on household transport data to study the distributional effects of carbon-based fuel taxation of private road transport in Germany. Our data includes car-specific annual mileage, fuel efficiency, and the distinction between fuel types, allowing for a very detailed analysis. The model enables focusing on different household types as well as identifying effect heterogeneity across the income distribution. We compare the recent fuel tax scheme with three policy reform scenarios to empirically test several hypotheses regarding distributional effects of carbon pricing. We find that the legal status quo of the fuel tax has overall regressive effects, with the tax on petrol being regressive and the tax on diesel being progressive. A transformation of the current tax into a revenue-neutral carbon-harmonised fuel tax yields a progressive distributional effect, while an introduction of a new carbon tax on transport fuels is neither clearly regressive nor progressive. Combining both tax schemes also has non-regressive effects. On aggregate, the distributional effects are modest. Subgroup analyses, however, reveal that the individual burden can be substantial. In sum, our results suggest that policy makers face various options for pricing road transport greenhouse gas emissions without causing an overall disproportionate tax burden on low-income households.
... It is imperative to investigate whether environmental tax policies result in CO 2 emissions reduction while exerting a minimal effect on economic growth (Li 2019;Dong et al. 2021). Elkins and Baker (2001) argued that environmental tax partially and wholly modifies environmental issues by enhancing incentives for clean energy consumption. The literature found different types of environmental tax such as carbon tax, fuel tax, and energy tax to achieve the target of reducing environmental pollution (Tamura et al. 1999). ...
Full-text available
Environmental stringency policy shocks and environmental tax have become fundamental policy tools for mitigating the degradation of the environment. The study explores the effects of environmental tax and environmental stringency policy shocks in the reduction of pollution emissions in China for the time 1993 to 2019. This study is a pioneer in assessing the simultaneous impact of these two policy instruments on pollution emissions in China. For empirical investigation, the study employed NARDL estimation techniques. The NARDL results show that positive shocks in environmental tax reduce N2O emissions by 0.03%, PM2.5 emissions by 0.13%, CO2 emissions by 0.18%, and GHGs emissions by 0.01%, however, negative shocks in environmental tax increase N2O emissions by 0.01%, PM2.5 emissions by 0.07%, CO2 emissions by 0.28%, GHGs emissions by 0.17% in the long run. The long-run results also show that positive shocks in environmental policy stringency reduced CO2 emissions by 0.94%, GHGs emissions by 0.77%, while negative shocks in environmental policy stringency increased N2O emissions by 0.17%, PM2.5 emissions by 0.50%, CO2 emissions by 0.63%. The findings suggest vigorous policy implications.
There is a growing interest in using carbon taxes to reduce greenhouse gas emissions, not only in industrialized economies but also in developing economies. Many countries have considered carbon pricing, including carbon taxes, as policy instruments to meet their emission reduction targets set under the Paris Climate Agreement. However, policy makers, particularly from developing countries, are seeking clarity on several issues—particularly the impacts of carbon taxes on the economy, the distribution of these impacts across households, carbon tax design architectures, the effects of carbon taxes on the competitiveness of carbon-intensive industries, and comparison of carbon taxes with other policy instruments for climate change mitigation. This paper aims to offer insights on these issues by synthesizing the literature available since the 1970s, when the concept of carbon tax was first introduced. This paper also identifies the areas where further investigations are needed. (JEL H23, Q35, Q38, Q54, Q58)
Full-text available
Penelitian ini bertujuan untuk mengetahui keunggulan dan kelemahan dari pajak karbon dan skema cap and trade dalam carbon pricing serta menggambarkan implementasinya di beberapa negara yang telah menerapkannya lebih dahulu, yaitu Kanada, Britania Raya, dan Australia. Selain itu, penelitian ini juga mendeskripsikan persiapan Indonesia dalam menerapkan carbon pricing serta memberikan alternatif faktor yang perlu dipertimbangkan oleh pemerintah. Penelitian ini menggunakan pendekatan kualitatif dengan studi analisis komparatif pada negara-negara tersebut. Pengumpulan data dilakukan dengan metode kepustakaan serta mewawancarai beberapa narasumber yang memiliki perhatian di bidang energi dan pajak karbon. Penelitian ini menunjukkan pajak karbon dan skema cap and trade memiliki keunggulan dan kelemahan bawaannya sehingga keputusan instrumen carbon pricing mana yang diterapkan, perlu disesuaikan dengan kebutuhannya masing-masing. Secara garis besar, kebijakan di ketiga negara tersebut memiliki tujuan yang sama, namun instrumen, mekanisme, dan tarif yang dikenakan berbeda. Pemerintah Indonesia telah melakukan persiapan implementasi carbon pricing sejak tahun 2021, namun hingga saat ini peraturan teknis terkait penerapannya masih belum ditetapkan. Masih terdapat beberapa alternatif faktor yaitu: regulasi penerapan dan regulasi kebijakan lingkungan lainnya, penentuan tarif yang optimal, redistribusi penerimaan, serta alternatif energi pengganti yang sesuai dengan kesiapan masyarakat untuk menjadi pertimbangan oleh pemerintah Indonesia dalam persiapan penerapan carbon pricing.
Against the background of the mixed carbon policy (carbon tax‑carbon trading) implemented in China, the authors analyze the complexity of a series of competition decisions between supply chains through constructing Cournot(Duopoly)game model between supply chains. The dynamics of the proposed game model and the time evolution of the supply chains are described by a nonlinear discrete dynamical system. The equilibrium point of the discrete dynamical system is obtained and its local stability is discussed. The stability of the equilibrium point is further discussed by numerical simulation, which is affected by the speed of adjusting parameters and some other cost parameters. Finally, we apply the delayed feedback control method to suppress and overcome the chaos of the discrete dynamical system. It can effectively promote the discrete dynamical system to return to the stable behavior. It is hoped that the current research can provide a theoretical basis for the competitive decision-making of supply chain node enterprises.
Full-text available
With the proposal for a Regulation of the European Parliament and of the Council establishing a Carbon Border Adjustment Mechanism (CBAM) adopted on July 14, 2021 by the European Commission, it begins to reach the heart of the global debate on the opportunity to adopt mechanisms to combat the phenomenon of ‘carbon leakage’, that is the attempt to evade existing carbon pricing systems through the delocalization of the most polluting activities. In this way, the European Union demonstrates its serious intention to strengthen actions against climate change and to identify customs taxation as an instrument to guide the choices of its trading partners and, therefore, of the major world economies. The precautions adopted in the application of the CBAM and the long transitional period before its entry into force, however, prevent the full appreciation of its potential repercussions (also) in terms of the new own resource of the European budget.
Full-text available
In order to solve the problems of improper order allocation and the lack of a carbon emission constraint system in the road freight transportation industry, this paper proposed an order allocation mechanism of network freight transportation with carbon tax constraints and established an order allocation optimization model with carbon tax constraints. Based on the basic characteristics of the problem, this paper redesigns the ant colony labor division expansion model, and designs a corresponding algorithm to solve the problem. By improving the update rules of the stimulus value and the threshold value, the matching difference between the order and the driver of the network freight platform is enlarged, and the matching relation-ship is dynamically adjusted, the order allocation scheme is optimized, and a more appropriate carbon tax rate range in this industry is explored. Furthermore, the problem is solved by a 0-1 integer programming algorithm, which is compared with the algorithm designed in this paper. Through multiple numerical simulation experiments, the effectiveness and feasibility of the algorithm are verified. The experimental results show that the order allocation arrangement of the online freight platform with carbon tax constraints is more economical and environmentally friendly.
Under neoclassical assumptions, this paper develops a model to illustrate the effects of government use of carbon allowances for carbon emission control. We find that control using the instrument of issuing long-term carbon allowances does not produce the same good results in the Decentralised Equilibrium as in the Planning Problem. Thus, while Coase's Theorem ensures that the Pareto optimum is maintained in an economy with multiple agents, it does not align the social development with the economic optimum of the planner. We believe this is due to the that the price of carbon allowances is determined by asset profiles of investors rather than externalities. The steady-state under the special pathway shows that consumption is determined by the rate of technological progress, the total amount of carbon dioxide at steady-state, the level of technology at which steady-state is reached and the total amount of carbon allowances remaining. The comparison with the optimal tax path reveals that the price of carbon allowances has increased too quickly, leading to excessive consumption of fossil fuels in the early stages.
Full-text available
Personal Carbon Budgets (PCBs) are a radical policy innovation that seek to reduce an individual’s carbon consumption. This review identifies three archetypes of PCBs in the current literature; Personal Carbon Trading, Carbon Tax and Carbon Labelling. We theorised that carbon trading could affect equity and allow quality of life and consumption to be driven by income rather than needs. We, therefore, developed a new model (Personal Carbon Allowance with no trading) to compare to existing archetypes. A PESTLE (Political, Economic, Social, Technological, Legal, Environmental) framework was applied to each archetype to analyse and compare their costs and benefits and to critically evaluate and identify which model may be the most appropriate to reduce emissions severely but equitably. We conclude that the only model that can achieve this is our proposed Personal Carbon Allowance (PCA) model with no trading. PCA has a hard cap on emissions allowing for controllable severe cuts to emissions, and the lack of trading would prohibit those with wealth from continuing high-consumption lifestyles at the expense of those with lower incomes.
Full-text available
Government policies to reduce environmental pollution and global warming are often criticised as damaging the economy, particularly by reducing international competitiveness. This book addresses this issue directly by examining some of the policies concerned and their effects on competitiveness. The Introduction summarises the arguments and covers: market-based instruments versus regulation, the definition of competitiveness, and its relationship with the exchange rate and with technology.
Full-text available
The purpose of this paper is to examine the long-run equilibrium between green growth and some environmental variables like deforestation, energy depletion and carbon dioxide (CO2) emissions in Nigeria from 1980 to 2015. To examine these long-run linkages, the study adopted the autoregressive distributive lag bound testing approach to cointegration. The bound testing approach shows an evidence of a negative long-run relationship between carbon dioxide emission and environmental depletion and renewable energy which stand as the proxy for green growth variable. Inversely, a positive long-run relationship exists between green growth variable and deforestation. Generally, this paper submits that efficient investment in green growth technologies and policies in Nigeria will assist to ensure a sustainable environment.
Fiscal changes designed to achieve the two objectives of reductions in greenhouse gas emissions and increases in employment are explored using a large-scale energy-environment-economy model of the UK, the Cambridge Multisectoral Dynamic Model (MDM). Two options are considered in detail: the road fuel duty escalator, introduced in the March 1993 Budget, and the CEC's proposed carbon/energy tax; revenues for each tax are recycled via reductions in employers' NIC calculated to keep the ratio of the PSBR to GDP at base scenario rates. The changes are introduced in 1996 and standardised so that each reduces CO2 emissions by some 21 mtC by 2005. The escalator is shown to raise far more revenue per mtC reduced and so NIC charges can be reduced much more than under the carbon/energy tax option. A detailed analysis is made of the effects on industrial costs of the carbon/energy tax option; the effects on competitiveness, measured as the effects on the balance of payments under a fixed exchange rate regime, are found to be negligible, even with a unilateral tax. Both options appear to provide substantial environmental net benefits as well as economic gains. However, each option has its strengths and weaknesses and a combination may allow the UK to reach the Toronto target of a 20% cut in CO2 emissions below 1988 levels by 2005.
A central concern of Danish energy and environmental policy is the fulfilment of international obligations to limit emissions of CO2, SO2 and NOx. The model presented aims at giving a frame of reference within which energy and environmental issues may be analysed in line with - and consistent with - macroeconomic analyses. The paper presents the theoretical and empirical foundations of the model as well as the results of applying it to calculating the impacts on emissions and the economy of a CO2-tax.