Environmental Externalities, Market Distortions and the Economics of Renewable Energy Technologies[Dagger]
Jul 20 - Energy Journal, The
This paper reviews life cycle analyses of alternative energy technologies in terms of both their private and societal costs (that is, inclusive of externalities and net of taxes and subsidies). The economic viability of renewable energy technologies is shown to be heavily dependent upon the removal of market distortions. In other words, the removal of subsidies to fossil fuel-based technologies and the appropriate pricing of these fuels to reflect the environmental damage (local, regional, and global) created by their combustion are essential policy strategies for stimulating the development of renewable energy technologies in the stationary power sector. Policy options designed to "internalize" these externalities are briefly addressed.
The twentieth century witnessed historically unprecedented rates of growth in
energy systems, supported by the widespread availability of fossil fuel
resources. During the second half of the century, however, concerns associated
with high levels of fossil fuel dependence began to surface. Two issues were of
particular significance: the impact of modern energy systems on the environment
and security issues associated with fuel supply lines.
Environmental concerns had been evident in more localised areas for many
hundreds of years. Ancient Rome burned wood and Emperor Nero's tutor, Seneca,
complained of the bad effect that smoke had on his health and of smoke damage to
temples, whilst anecdotal evidence indicates that air pollution had been a
concern in England as early as 1352 when a ban was introduced on coal burning in
London. Today, local pollution from energy systems remains a threat to the
health of the living environment. However, in the latter decades of the
twentieth century, pollution resulting from combustion of fossil fuels became a
global concern, with the publication of credible scientific evidence that the
planet's climate was changing as the result of a build up of so-called
greenhouse gases in the atmosphere.
Historically, regulatory instruments have been the basic mechanism for
enacting environmental policy throughout the industrialised world. Environmental
quality has been seen as a public good that the state must secure by preventing
private agents from damaging it. Direct regulation involves the imposition of
standards (or even bans) regarding emissions and discharges, product or process
characteristics, etc., through licensing and monitoring. Legislation usually
forms the basis for this form of control, and compliance is generally mandatory
with sanctions for non- compliance.
The proposal to impose taxes on pollution, whilst more recent, is also far
from new, having been advanced at the turn of the last century by the famous
British economist Arthur Cecil Pigou as a means of reducing London's famous fogs
(or smogs). Pigou observed that pollution imposed uncovered costs on third
parties that were not included in ordinary market transactions. His proposal was
to tax pollution by means of a so-called externality tax1 in order to
internalise within ordinary market transactions the damages caused by pollution.
At the time Pigou's proposal was regarded as an academic curiosity, but several
generations later it was rejuvenated as the core of the "polluter pays
principle."
Contemporary energy policy issues are dominated, directly and indirectly, by
major concerns at both local and global levels of environmental degradation
arising from combustion of fossil fuels. Even countries with relatively modest
fossil fuel requirements, such as the poorer nations of Africa, Asia, and the
South Pacific, could experience significant adverse consequences if the world's
requirement for energy from fossil fuels does not abate within a relatively
short time frame. Consequently, the economics of renewable energy technologies
has a core position in energy policy formulation over the foreseeable future.
However, a number of non-quantifiable policy objectives are also of
significance in the planning of future energy technology options. Currently, the
most important of these would appear to be the security of supply of energy
resources and their associated transmission and distribution systems. To the
extent that governments bear the security costs associated with ensuring that
uninterrupted supplies of fuels reach the relevant markets, then these fuels are
being subsidised and hence there exists an inefficient allocation of resources.
The price to the ultimate consumer would be too low, and consequently demand
(and pollution) levels would be higher than in the absence of the subsidy.
This paper commences with a summary of the economics of environmental
externalities. An overview of the methodology of life cycle analysis and its
application to the energy sector to derive estimates of environmental
externalities is then given. The implicit costs of externalities attributable to
power generation (from both commercial and potentially commercial technologies)
are then compared with the private costs that are generally passed on to the
consumer.
2. EXTERNALITIES2
2.1 Definition
Externalities are defined as benefits or costs generated as an unintended
by-product of an economic activity that do not accrue to the parties involved in
the activity and where no compensation takes place. Environmental externalities
are benefits or costs that manifest themselves through changes in the
physicalbiological environment.
Pollution emitted by road vehicles and by fossil fuel fired power plants
during power generation is known to result in harm to both people and the
environment. In addition upstream and downstream externalities, associated with
securing fuel and waste disposal respectively, are generally not included in
power or fuel costs. To the extent that the ultimate consumer of these products
does not pay these environmental costs, nor compensates people for harm done to
them, they do not face the full cost of the services they purchase (i.e.,
implicitly their energy use is being subsidised) and thus energy resources will
not be allocated efficiently.
The origin of an externality is typically the absence of fully defined and
enforceable property rights. However, rectifying this situation through
establishing such rights is not always an easy task. In such circumstances, at
least in theory, the appropriate corrective device is a Pigouvian tax equal to
marginal social damage levied on the generator of the externality. If the tax is
subsequently used to compensate the sufferer(s), then the externality is said to
have been "internalised."
2.2 Externality Adders
In the context of energy markets, an "externality adder" is simply
the unit externality cost added to the standard resource cost of energy to
reflect the social cost of its use. For power generation, the externality adder
would generally be specified in terms of milli-dollars (1000th of a dollar) per
kWh (m$/kWh) or / kWh. For the transport sector the corresponding units would be
m$/ vkm (i.e., one-thousandth of a $ per vehicle kilometre) for passenger
vehicles and m$/tkm (i.e., one-thousandth of a $ per tonne kilometre) for goods
vehicles, or the equivalent in cents.
Pearce (2002) lists five uses for externality adders:
i. For public or quasi-public ownership of sources of electric power
generation, the full social cost of alternative technologies could be used to
plan future capacity with preference being given to that with the lowest social
cost. Where electric power generation is privately owned, then regulators could
use the full social cost to influence new investment, perhaps through an
effective environmental tax.
ii. Environmental adders can be used to estimate the appropriate level of
environmental taxes. Although estimates of environmental adders have been
derived for a number of applications, examples of their actual implementation
are few.
iii. Environmental adders could be used to adjust national accounts data to
reflect depreciation of natural resources and damage to the environment arising
from economic activity, yielding so-called "green" national accounts.
iv. Environmental adders could be used for "awareness raising";
i.e., to inform the public of the degree to which alternative energy sources
have externalities that give rise to economically inefficient allocation of
resources.
v. Environmental adders might assist in determining environmental policy
priorities.
The task of estimating the value of an externality adder involves a
substantial commitment of resources and expertise in order to ensure credible
information for policy purposes. In the context of the energy sector, a life
cycle approach must be adopted in order to identify and quantify environmental
adders associated with the provision of energy services. The approach also
provides a conceptual framework for a detailed and comprehensive comparative
evaluation of energy supply options (based upon both conventional and renewable
sources). The methodology employed is the subject of the next section.
3. LIFE CYCLE ANALYSIS
When comparing the environmental footprints of alternative energy
technologies, it is important that the power generation or combustion stage of
the technology not be isolated from other stages of the "cycle." For
example, fuel cells emit virtually no greenhouse gases (GHGs) in their
operation. However production of their "fuel" (hydrogen) from fo\ssil
fuels may involve increases in GHG emissions in excess of those that would arise
from using current commercial fossil fuel technologies to meet the same level of
energy requirements. To avoid such distortions, the concept of life cycle
analysis has been developed.
Life cycle analysis (LCA) is based upon a comprehensive accounting of all
energy and material flows, from "cradle to grave," associated with a
system or process. The approach has typically been used to compare the
environmental impacts associated with different products that perform similar
functions, such as plastic and glass bottles. In the context of an energy
product, process, or service, a LCA would analyse the site-specific
environmental impact of fuel extraction, transportation and preparation of fuels
and other inputs, plant construction, plant operation/fuel combustion, waste
disposal, and plant decommissioning. Thus it encompasses all segments including
upstream and downstream processes and consequently permits an overall comparison
(in a cost benefit analysis framework) of short- and long-term environmental
implications of alternative energy technologies. Central to this assessment is
the valuation of environmental externalities of current and prospective fuel and
energy technology cycles. It should be noted, however, that only material and
energy flows are assessed in an LCA, thus ignoring some externalities (such as
supply security) and technology reliability and flexibility.
For the purpose of this paper, life cycle analysis will involve the following
methodological steps:3
* Definition of the product cycle's geographical, temporal, and technical
boundaries;
* Identification of the environmental emissions and their resulting physical
impacts on receptor areas; and
* Quantifying these physical impacts in terms of monetary values.
Traditionally, LCA has omitted the third of these steps and the final
analysis has therefore been expressed in terms of just the biophysical impacts
that can be quantified. The extension to include costing of these impacts is
generally known as the "impact pathway" methodology. Essentially,
however, it can be considered as a specific application of LCA. This methodology
formed the theoretical basis for the European Commission's ExternE4 study, which
was the first comprehensive attempt to use a consistent "bottom-up"
methodology to evaluate the external costs associated with a range of different
fuel cycles. The main steps are illustrated in Figure 1.
Figure 1. The Impact Pathway Methodology
3.1 Definition of the Product Cycle's Boundaries
The first task is to identify, both in terms of activities and geographic
locations, the various stages of the fuel/technology cycle. Each energy form is
viewed as a product, and impacts are included for the actual pathway. The
precise list of stages is clearly dependent on the fuel chain in question, but
would include activities linked to the manufacture of materials for plant
construction, demolition and site restoration as well as power generation. Other
stages may also be appropriate, such as exploration, extraction, processing and
transport of fuel, and the generation of wastes and by-products, and their
treatment prior to disposal.
The extent to which the boundaries must encompass indirect impacts is
determined by the order of magnitude of their resulting emissions. For example,
in theory externalities associated with the construction of plants to make the
steel that is used to make coal wagons to transport the coal to the power plants
should be included in the power plant's LCA. In reality, however, such
externalities are likely to have a relatively insignificant impact. In addition,
externalities that pass into another product's boundaries must be excised from
the analysis to avoid double counting. For example, the ultimate environmental
externality of by-products of power generation that are fully utilised in
another industry fall within the latter's life cycle as soon as product transfer
occurs.
For each fuel/technology cycle, boundaries are likely to vary, particularly
in relation to upstream impacts, and consequently derivation of a
"generic" LCA for each technology may be unrealistic. For example,
identical coal-fired power plants located in different areas of the same country
may use coal from different sources (perhaps one uses imported coal, the other
domestic), there may be variations in fuel quality or variations in atmospheric
dispersion, or there may be differences in the sensitivity of the human and
natural environment upon which fuel chain burdens impact. When different
generations of coal-fired plants enter the analysis, use of a generic approach
may lead to a further drop in precision. However, the increased precision
achieved by deriving a site specific LCA for all projects may well be offset by
the cost of such exercises. In reality, indicative or generic estimates may be
unavoidable.
The system boundary will also have spatial and temporal dimensions. These
will have major implications for the analysis of the effects of air pollution in
particular. For many air pollutants, such as ozone and SO2, the analysis may
need to focus on a regional, rather than local, scale in order to determine
their total impact. For emissions of GHGs, the appropriate range is clearly
global. Impacts must also be assessed over the full term of their effect, a
period that may extend over many decades or even centuries in the case of
emissions of GHGs and long-term storage of some nuclear waste products. This
introduces a significant degree of uncertainty into the analysis, as it requires
projections to be made of a number of variables that will form the basis of
future society. Among these would be the size of the global population, the
level of economic growth, technological developments, the sustainability of
fossil fuel consumption, and the sensitivity of the climate system to
anthropogenic emissions.
A generic "chain" for coal-fired electricity generation is
illustrated in Figure 2. Even from this simplified illustration, however, it is
clear that the data requirements to undertake a LCA are formidable, particularly
where sources in other countries have to be accessed. Data limitations and cost
constraints will obviously combine to prevent a complete enumeration of the
emissions of a given process. It is essential, therefore, that when this
situation is reached the proportion left unaccounted should be clearly
specified.
3.2 Identification of the Environmental Emissions and Their Resulting
Physical Impacts on Receptor Areas
Environmental emissions (or burdens) from the energy sector that are capable
of causing some form of impact can be identified in the following broad
categories:
* Solid wastes;
* Liquid wastes;
* Gaseous and particulate air pollutants;
* Risk of accidents;
* Occupational exposure to hazardous substances;
* Noise; and
* Others (e.g., exposure to electro-magnetic fields, emissions of heat).
All potential physical impacts of the identified burdens for all fuel chains
must be analysed comprehensively. However, it is possible to produce several
hundred burdens and impacts for the various fuel chains. Thus, for practical
reasons, the analysis must concentrate on those that are considered to be
non-negligible in terms of their externalities.
Some impact pathways may be relatively simple. For example, the construction
of a wind farm will affect the appearance of a landscape, leading to a change in
visual amenity. In other cases, the link between the burden, physical impact,
and monetary cost is far more complex. In reality, much of the required data is
either incomplete or simply does not exist. Thus any analysis is, of necessity,
only partial.
Figure 2. Generic Coal-Based Electricity Fuel Cycle Chain
Comparisons of alternative power generation technologies utilising LCA are
generally standardised as emissions per unit of energy produced (kWh) in order
to allow for different plant sizes and capacity factors. However, the data used
to quantify burdens is, to varying degrees, technology specific. For example,
emissions of carbon dioxide (CO2) in power generation depend only on the
efficiency of the equipment and the carbon/hydrogen ratio of the fuel;
uncertainty is negligible. Whereas emissions of SO^sub 2^ can vary by an order
of magnitude depending on the grade of oil or coal and the extent to which
emission abatement technologies have been adopted. As a general rule, one should
adopt the most efficient technology currently in use in the country of
implementation in order to compare environmental pollutants across different
technologies.
Quantifying the physical impacts of emissions of pollutants requires an
environmental assessment that ranges over a vast area, extending over the entire
planet in the case of CO2 emissions. Thus the dispersion of pollutants emitted
from fuel chains must be modelled and their resulting impact on the environment
measured by means of a dose-response function. Generally, for damages to humans,
such functions are derived from studies that are epidemiological; assessing the
effects of exposure to pollutants in real life situations.
3.3 Cost of Damage or Cost of Abatement?
The two principal methods generally used for assessing the value of
externalities are calculation of damage costs and calculation of control (or
abatement) costs. Although control costs are often (mistakenly) seen as
estimates of damage costs, conceptually they are very different. Damage costs
are a measure of society's loss of wellbeing resulting from the damage arising
from a specific adverse environmental impact. Control costs are what it costs
society to achieve a given standard that restricts the extent of the impact to
an acceptable level, and are thus likely to be only tenuously related to damage.
Control costs are often used as a surrogate for damage costs as they are a
relatively straightforward concept, are re\latively easy to derive, and can be
applied to most environmental impacts. Essentially, control costs can be
calculated simply by dividing the cost of mandated controls by the emissions
reduction achieved by the controls. In general, however, control costs must be
viewed as a poor substitute for estimating damage costs, since the methodology
is subject to inherent flaws. The implicit assumption in control costing is that
society controls pollution until the benefits of additional controls would be
outweighed by the costs of their imposition. But using the cost of regulation to
estimate the benefits is rather a meaningless, circular, procedure, given that a
cost benefit ratio of unity will always be achieved. A further flaw is that use
of control costs to value externalities implies that legislators are able to
make optimal decisions when imposing policy instruments to modify polluting
behaviour to achieve such an "optimal" outcome. However, in practice,
epidemiological studies of cost per life saved (for example) have indicated
large variations in the values implied by the costs and benefits of different
regulations.
Estimation of damage costs has economic theory as its basis. It focuses
directly on explicitly expressed preferences as revealed by willingness to pay
to avoid environmental damage or by stated preferences in either real or
simulated markets. In addition, it can be combined with financial assessment of
investment options in order to provide a societal estimate for the impacts of an
investment in a common numeraire. This methodology is fundamental to the
attribution of financial values to environmental impacts identified in LCA. The
last of the four stages in the environmental "impact pathway" involves
calculation of the economic value of the biophysical effects in terms of
willingness to pay to avoid damage arising from the emission of pollutants.
Clearly, however, a major disadvantage is the scale of the data requirements for
deriving estimates of these damage costs.
There is no reason why the two concepts should be of comparable dimension. In
fact, rationally, control costs should always be less than the estimated level
of damages.
3.4 Quantifying Physical Damage in Terms of Monetary Values
The many receptors that may be affected by fuel chain activities are valued
in a number of different ways. For example, forests are valued not just for the
timber that they produce, but also for providing recreational resources,
habitats for wildlife, their interaction (both direct and indirect) with
climate, the hydrological cycle, protection from soil erosion, etc. All such
aspects have to be valued in an externality analysis.
Figure 3 illustrates the major valuation methodologies that have been
developed in order to attribute a monetary value to the biophysical impacts of
environmental externalities. Commercial, observable, markets exist for some
goods, e.g., crops, timber, buildings, etc., and consequently valuation data are
relatively easy to derive. However, conventional markets do not exist for
assessing damage from many other impacts, such as human health, ecological
systems, and non-timber benefits of forests. A number of techniques have been
developed for assessing willingness to pay (WTP) for such "goods," and
these are set out in the figure.5
The temporal valuation of the cost of damage resulting from energy sector
emissions also raises the controversial issue of the appropriate rate for
discounting over future generations.6
Figure 3. Externality Valuation Methodologies
4. THE COSTS OF ELECTRICITY GENERATING TECHNOLOGIES
Power plants are most frequently compared on the basis of their levelized
electricity cost (LEC), which relates the discounted capital cost of the plant,
its annual operating and maintenance costs and fuel prices to the annual
production of electricity to yield a value in cents per kWh.7
Renewable energy technologies that are, by their very nature, intermittent
would incur fuel costs to the extent that backup capacity was used in order to
maintain the desired supply of peaking power to the grid. At low levels of
renewables penetration additional system costs would be negligible compared with
generation costs, since variability would still be within normal tolerance
levels for the system as a whole. Thereafter, higher levels of penetration will
involve additional cost, since additional generation or electricity storage
capacity would be required to meet peak demand if, for example, wind were
unavailable.8 As a consequence, at a purely financial level, the value of
intermittent generation should be less than that of conventional generation by
approximately these additional costs.9
Table 1 gives (indicative) levelized electricity costs in euro- cents per
kilowatt hour (euro/kWh) for electricity generation by the major renewable and
non-renewable technologies. Both coal and gas exhibit a clear absolute cost
advantage over the bulk of renewable technologies, although electricity
generated by "best performance" wind power has recently approached
similar cost levels. Back-up generation costs associated with the intermittency
of renewables to ensure reliability of supply are not included. Thus on purely
financial grounds (inclusive of all forms of subsidy), renewable technologies
would, in general, appear to be non-competitive. The cost "gap" has
been narrowed significantly over the past two decades, a process that is
expected to continue as reflected in projected cost levels for 2020 (Table 1).
However, it is clear that significant policy actions to increase investment in
research and development and to stimulate economies of scale in production and
dissemination of renewables are required to meet environmental commitments on
global climate change in any major way.
Table 1. Cost of Traditional and Renewable Energy Technologies Current and
Expected Trends
The cost data presented in Table 1, however, give a misleading indication of
the extent of the cost disadvantage of renewables.
* Unlike fossil fuel technologies, the efficiency of renewable technologies
is generally very site-specific. For example, it would be expected that
photovoltaics in the UK would incur a higher cost per kWh than countries located
at lower latitudes. In contrast, coal and (to a lesser extent) gas fired power
plants use a fuel that is internationally traded and therefore of similar cost
(net of transport charges) throughout the world. Thus, comparisons should be
made on the basis of "optimal conditions" costs, rather than the full
range that may incorporate old technologies and inappropriate siting decisions.
* Photovoltaics is generally "delivered" as distributed
electricity. Thus its cost should be compared with "delivered" (i.e.,
inclusive of transmission and distribution costs) electricity from other
sources, both renewable and fossil fuel. In Table 1, cost ranges for delivered
electricity are also given. Outside of rural electrification in developing
countries the cost difference still favours fossil fuel technologies, but the
divergence is considerably smaller than when delivery is ignored.
5. ASSESSING THE EXTERNALITIES OF POWER GENERATION
Environmental externalities of energy production/consumption (whether based
upon fossil fuel combustion, nuclear power or renewable technologies) can be
divided into two broad (net) cost categories that distinguish emissions of
pollutants with local and/ or regional impacts from those with global impacts:
* costs of the damage caused to health and the environment by emissions of
pollutants other than those associated with climate change; and
* costs resulting from the impact of climate change attributable to emissions
of greenhouse gases.
The distinction is important, since the scale of damages arising from the
former is highly dependent upon the geographic location of source and receptor
points. The geographic source is irrelevant for damages arising from emissions
of greenhouse gases.
Costs borne by governments, including direct subsidies, tax concessions,
indirect energy industry subsidies (e.g., the cost of fuel supply security), and
support of research and development costs, are not externalities. They do,
however, distort markets in a similar way to negative externalities, leading to
increased consumption and hence increased environmental degradation.
In order to address effectively these environmental matters, together with
energy supply security concerns, radical changes in power generation, automotive
engine, and fuel technologies will probably be required. Such changes must offer
the potential for achieving negligible emissions of air pollutants and
greenhouse gases, and must diversify the energy sector away from its present
heavy reliance on fossil fuels (and particularly gasoline in the transportation
sector). A number of technologies, including those that are solar or
hydrogen-based, offer the long term potential for an energy system that meets
these criteria.
However, a number of policy objectives that are more difficult to quantify
are also of significance in the planning of future technology options.
Currently, the most important of these would appear to be the security of supply
of energy resources and their associated transmission and distribution
systems.10
5.1 Pollution Damage From Emissions Other Than CO2
This category refers to costs arising from emissions that cause damage to the
environment or to people. These include a wide variety of effects, including
damage from acid rain and health damage from oxides of sulphur and nitrogen from
fossil fuel power plants. Other costs in this category include such factors as
power industry accidents (whether they occur in coal mines, on offshore oil or
gas rigs, in nuclear plant, on wind farms, or at hydro plants), visual
pollution, and noise.
Among the major external impacts attributed to electricity generation are
those caused by atmospheric emissions of pollutants, suchas particulates,
sulphur dioxide (SO^sub 2^) and nitrogen oxide (NO^sub x^), and their impacts on
public health, materials and crops. The impact of these atmospheric pollutants
on forests, fisheries and unmanaged ecosystems are also important but have not
yet been quantified. Emissions of SO^sub 2^ and NO^sub x^ have long range
transboundary effects, which makes calculation of damages an imprecise exercise.
Such calculations require measurement to be based upon the unique link between
fuel composition, characteristics of the power unit, and features of the
receptor areas. Thus estimated damage costs may vary widely across continents,
and even within individual countries.
Estimated damages per tonne of pollutant for SO^sub 2^, NO^sub x^, and
particulates vary greatly because of a number of factors. Briefly these are:
* Vintage of combustion technologies and presence of associated
emission-reducing devices such as flue gas desulphurisation or low NO^sub x^
burners;
* Population density in receptor areas for airborne pollutants;
* Fuel quality (particularly coal); and
* Mining and fuel transportation externalities (particularly accidents).
The major source of pollution is at the power generation stage for fossil
fuels, whereas for renewables it tends to be during equipment manufacturing
stages.
However, damage estimates are dominated by costs arising from human health
effects, which are largely determined by the population affected. Estimation of
health impacts is generally based upon exposure-response epidemiological studies
and methodologies for placing a valuation on human life remain controversial.11
As might be expected, countries that are sparsely populated, or populated in
largely non-receptor areas, tend to have relatively low health damage costs.
The ExternE study has produced estimates of human health damages and other
non-climate change pollution damages for the coal fuel cycle that range from 0.2
euro/kWh to 4.0 euro/kWh.12 For the gas fuel cycle, where SO^sub 2^ emissions
are negligible, combined cycle gas turbine technology produces damages that are
considerably lower per kWh than for coal, particularly for combined heat and
power plants. Again, the largest damages occur where the plants are located
close to high population density areas. Even then, damages do not exceed 1.0
euro/kWh, and are generally considerably lower than this figure. While power
generation damages arising from the oil fuel cycle are, on average, marginally
lower than those associated with coal, they too exhibit significant variation
between plants.
It is evident from these damage values that the country-specific nature of
these estimates does not permit an "average" global damage figure to
be derived, and thus country (or regional) specific policies would be required
in order to reduce existing damage levels. This could occur automatically if
investment in new plant derived benefits from utilising technological
developments that further reduced pollutants, whilst existing plants could be
retrofitted with improved technology as it became available.
However, Rabl and Spadaro (2000) have estimated "typical"
quantifiable,13 average European conditions, non-CO2 damages to be 4.54 euro/kWh
for the coal fuel cycle, with a comparable estimate for gas of 1.12 euro/kWh.
The discrepancy between these estimates and those of the ExternE study quoted
earlier (that produced separate damage costs estimates for each pollutant in
each country in the EU) is due to the higher damage costs attributed to the
pollutants by Rabl and Spadaro (2000).
5.2 The External Damage Costs of Emissions of Carbon Dioxide
This category refers to external costs arising from greenhouse gas emissions
from electricity generating facilities that lead to climate change with all its
associated effects. This is a very contentious area, and the range of estimates
for the possible economic ramifications of global climate change is vast. Costs
associated with climate change, such as damage from flooding, changes in
agriculture patterns and other effects, all need to be taken into account.
However, there is a lot of uncertainty about the magnitude of such costs, since
the ultimate physical impact of climate change has yet to be determined with
precision. Thus, deriving monetary values on this basis of limited knowledge is,
at present, an imprecise exercise.
Table 2. CO2 Emissions from Different Electricity Generation Technologies
Table 2 sets out typical life-cycle CO2 emissions (in tonnes per GWh) of the
major forms of electric power generation. From this table it can be noted that
CO2 emissions from coal and oil-based technologies far exceed those of the
"renewables" and are about twice those of gas. In terms of damage
costs from CO2 alone, based upon an updated ExternE estimate14 of 29 euro/tonne
CO2, (or 8 euro/ tonne C) a "typical," average European conditions,
coal fuel cycle would cause damage equivalent to 2.8 euro/kWh. The comparable
damage cost for gas would be 1.4 euro/kWh.15
5.3 External Damage Costs for Electricity Production
Table 3 gives cost ranges (euro/kWh) for quantifiable external costs
associated with the range of electricity generation technologies for countries
within the European Union. The ranges are often relatively large, reflecting
variations in generation technology (and hence emission levels per kWh) and
geographic location (and hence damage costs per kWh).
Based upon the Rabl and Spadaro estimates, a typical, average European
conditions, new baseload plant, would have total quantifiable damage costs of
7.27 euro/kWh for a coal fuel cycle, and 2.37 euro/kWh for gas. Both of these
estimates fall within their respective "EU range" in Table 3, despite
the relatively high assumed damage costs relative to the ExternE study.
These "typical" estimates indicate that total damage costs
associated with the coal cycle are (approximately) three times those of gas and
a very large multiple of those for renewable energy technologies.16 If these
typical "externality adders" are combined with the lower bounds of the
"current" cost data given in Table 1, the gas fuel cycle would exhibit
a marked societal cost advantage over all other modes of generation with the
exception of wind and hydro.
If the "environmental adders" were to be imposed upon expected
future costs, then it is clear that by 2020, under the best operating
conditions, many other renewables will become less costly than either gas or
coal on the basis of the societal cost of electricity production. Such a
comparison is fraught with problems, however, as the external costs per kWh
associated with both emissions of pollutants and climate change in 2020 are
likely to differ significantly from those given in Table 3.17 To a large extent
differences will depend upon the success or otherwise of GHG abatement programs
over the same period. A decline in damage costs arising from emissions of non-GHGs
can also be expected to occur as a consequence of continuing improvements in
emission-reduction technology and retirement of older plant.
Table 3. External Costs for Electricity Production in the EU
(range: euro/kWh)
5.4 Uncertainty and the Marginal Costs of CO2 Emissions18
Tol (2003) has reviewed 88 estimates, from 22 published studies, of the
marginal cost of carbon dioxide emissions and combined them to form a
probability density function. He found that the function is strongly skewed to
the right, with a mode of $5/tonne of carbon (tC), a mean of $104/tC, and a 95th
percentile of $446/tC. Including only peer-reviewed studies in the analysis,
gave corresponding estimates of $5, $57, and $307 respectively. Thus not only is
the mean estimate substantially reduced, but so is the degree of uncertainty.
Equity weighting19 and changing discount rates were also shown to have
significant effects on these estimates. Overall, Tol concluded that, for all
practical purposes, it is unlikely that the marginal costs of CO2 emissions
would exceed $50/tC or (40 euro/ tC) and are likely to be substantially lower.
Based upon a constant discount rate and without equity weighting, Pearce
(2003) quotes a range of $4-9/tC. Equity weighting, using a marginal utility of
income elasticity of unity, changes the range to $3.6-$22.5 (2.9-18.0 euro)/tC.
A time varying discount rate raised this range to $6.5-$40.5 (5.2-32.4 euro)/tC.
All estimates, therefore, are well below Tol's upper bound of $50/tC (40 euro/tC),
and the $8/tC quoted earlier falls towards the lower bound of Pearce's final
range.
6. INTERNALISING THE EXTERNALITIES OF ELECTRICITY PRODUCTION
6.1 Internalising Externalities
At least in theory, the most efficient process for imposing the
"polluter pays principle" would be to internalise as many of the
externalities of power generation as possible. Using the marketplace would
permit energy producers and consumers to respond to such price signals in the
most efficient and cost-effective way.
However, it should be emphasised that only external damage costs associated
with emissions from fossil fuel combustion have been considered explicitly in
these calculations. Those associated with other forms of power generation, in
addition to security of supply considerations and energy subsidies must also be
incorporated into the analysis in order to achieve a reasonable balance across
the range of power generating technologies, both conventional and renewable. For
example, without such action nuclear power, with its negligible level of CO2
emissions per kWh but significant subsidies and radioactive waste management
costs, would possess an apparent marked competitive advantage over all other
technologies (with the exception of some hydro systems), both renewable and
non-renewable. However, as noted earlier, costs associated with emission of
pollutants other than CO2 can be very variable and tend to be site- specific.
Once monetary values have been derived to reflect the external costs of d\iffering
technologies, the next step is to devise a mechanism for "internalising"
them into market prices. In theory, an energy tax would represent a relatively
straightforward solution, although the practicalities of its imposition would be
fairly complicated. The tax would be required to be imposed at differential
rates, depending upon the total estimated damages resulting from the fuel in
question. A simple carbon tax alone, for example, would not impose any cost on
the nuclear power industry. The tax would also have to be imposed by all
nations, to ensure that the competitiveness of their industries in global
markets was not compromised. The resulting tax revenue would also have to be
distributed in such a way that implicit energy subsidies were not (re-)
introduced. Finally, the worst of any social impact of energy taxes on poorer
sections of society would have to be offset to ensure that the tax burden was
not disproportionate in its incidence.
An alternative approach to the problem of reflecting external costs, and one
that would possibly cause less economic disturbance, would be to introduce
"environmental credits" for the uptake of renewable energy
technologies. Examples are currently commonplace. However, such credits do not
"internalise" the social costs of energy production but rather
subsidise renewables. In addition, the taxpayer pays the subsidy and not the
electricity consumer, thus rejecting the "polluter pays principle."
Their attractiveness to governments is that they can be justified as a carbon
offsetting initiative that is far more politically palatable than a carbon tax.
As noted earlier, leading renewable energy technologies are characterised by
relatively high initial capital costs per MW of installed capacity, but very low
running costs. This characteristic can make renewable technologies financially
unattractive compared with traditional fossil fuel derived power using
traditional project evaluation techniques based upon the anticipated life of the
electricity generating facility (say, 30 years). However, in terms of an
economic/environmental evaluation, the relevant time frame should be set by the
date at which all of the consequences attributable to the project had ceased to
exist. In the context of CO2 emissions from fossil fuel power stations this
period could exceed 100 years, and in the case of spent-fuel storage for nuclear
plants many hundreds of years. Further, it is likely that the value of emission
reduction will continue to rise into the future given projected world population
growth, economic growth, and the subsequent difficulties in meeting global
climate change agreements. In this context, the rate of discount is crucial in
assessing the relative cost and benefit streams of alternative energy
technologies in the context of intergenerational equity (Philibert, 1999; Newell
and Pizer, 2003; and Weitzman, 2001).
6.2 Policy Options for "Internalising" Externalities
Estimated damage costs associated with externalities of fossil fuel
combustion tend to lack precision,20 which would make the imposition of
environmental "adders" a very controversial policy option. Further, it
should be remembered that valuation of externalities is predicated on the
discipline of welfare economics, where economic (or allocative) efficiency is
the guiding principle. Distributional assumptions are, at least at that level,
ignored. In addition, most actions will be based upon control or abatement costs
and therefore their relationship with the precise cost of damage arising from
the externality may be very tenuous.21 However, a number of second-best options
are available that could, at least partially, approximate the desired outcome.
Direct Government Actions
Governments generally exercise effective control over many parts of western
economies, including buildings, employees, vehicle fleets, infrastructure,
government corporations, joint ventures, land and resource management, and the
allocation of research and development budgets. Because externalities are a form
of market failure, Government intervention is justified in order to minimise
their impacts on the community. Where taxing polluters is deemed to be
politically unacceptable, then environmentally benign technology could be
encouraged through grants and subsidies.
Voluntary Actions
Governments may try to influence the actions of households and firms by
voluntary means, such as information campaigns, advertising, environmental
product labelling, demonstration projects, and facilitating voluntary
environmental initiatives.
Economic Instruments
In principal, this would involve imposing an emissions tax on consumption of
the commodity in question, reflecting the damage incurred by society. In
practice, this is more likely to involve taxation at a level that would control
emissions to an acceptable standard (i.e., a control cost). Alternatively,
tradeable permits could be introduced to restrict emissions to the required
standard. In theory the two instruments are equivalent for meeting a given
standard, although in practice they can differ significantly in their impacts.22
Although the implementation of carbon taxes at the international level has
been discussed extensively, politically it has never been acceptable to a wide
range of countries. Both the negotiation of a carbon tax rate at the
international level and the implementation of a carbon tax regime have turned
out to be too complex. Difficulties lie in deciding on a level of tax and on how
the resulting revenue should be used or redistributed.
One of the first proposals for a carbon tax was US President Clinton's 'BTU'
tax, which was discarded in 1994. In 1992, the European Commission (EC) put
forward a proposal for a European Union- wide tax on all energy products, except
renewable energy sources. Half of the tax would have been based on the energy
content, and half on the carbon content of fuels. After the EC proposal had been
faced by severe opposition by the British government it was eventually abandoned
at the end of the nineties. The EC subsequently encouraged its member states to
adopt carbon taxes at the national level.
Carbon taxes have been implemented in Denmark, Finland, Germany, the
Netherlands, Norway, Sweden, and the United Kingdom. Details are given in Table
4. Although these taxes have been named carbon taxes, they don't usually have a
common tax base. For example, carbon taxes in Denmark and the United Kingdom are
imposed on a per kilowatt hour basis on the consumption of electricity, whilst
carbon taxes on natural gas in Denmark, Norway, Sweden and the United Kingdom
are imposed on cubic metres (m3) of natural gas consumed.23
In addition, there are many countries that have adopted taxes on energy
consumption that act implicitly as a carbon tax without, however, being called a
carbon tax. Moreover, the impact of these carbon taxes not only hinges on the
size of the tax rate but also on the modalities and rules for the recycling of
the revenue of these taxes. These are commonly very complex, as they are the
result of negotiations of all stakeholders, especially those firms who will be
affected by the tax.
Unlike carbon taxes, the first carbon emissions trading regime to emerge was
at the international level. In fact, the agreement on the Kyoto Protocol
negotiations in 1997 could only be achieved by adopting provisions for trading
greenhouse gas emissions internationally. The regime under the Kyoto Protocol is
a cap-and- trade regime. The most important driving factor was the concern of
the USA that they would not be able to implement sufficiently strong domestic
policies to meet their 7% emissions reduction target, and that they needed a
cost-effective means of meeting their emissions reductions. The trading
mechanisms adopted under the Kyoto Protocol are commonly referred to as
'flexibility mechanisms'.
Table 4. Taxes in OECD Member Countries Levied on Electricity Consumption
As part of countries' efforts to comply with their obligations under the
Kyoto Protocol, and also to be able to fully participate in international
emissions trading, a number of national and industry systems have emerged. These
include one regional scheme: the trading regime of the European Union.
Among the existing domestic regimes are Denmark, the United Kingdom, ERU-PT -
a Dutch programme, and the US state of Oregon. Of these, only the Danish trading
regime is a pure cap-and-trade regime. Among the industry schemes are the
internal trading programmes of Shell and British Petroleum (BP), and the
Canadian Pilot Emission Reduction Trading (PERT). Existing and emerging domestic
trading regimes are given in Table 5.
A European-wide scheme was adopted by the European Parliament in 2002. The
scheme provides for the introduction of legally binding, absolute emission caps
from 2005 for around 4000-5000 power stations and industrial plants with high
levels of energy consumption. The European trading scheme covers plants
midstream rather than in a purely up- or downstream fashion. Thus, the following
industries have been included: Power and heat generation (in plants with a
thermal input capacity exceeding 20 MW), mineral oil processing; coke ovens;
metal processing; cement and lime production, other building material and
ceramics, glass and glass fibre, and paper and cellulose. Minimum sizes apply,
and initially only CO2 emissions will be covered.
Regulation
This involves placing mandatory thresholds on the adoption of low emission
technologies or practices by power utilities and car manufacturers, energy use
in buildings, and land and other resource management codes. Renewables
obligations are being increasingly adopted by governments around the world.
Known as Portfolio Standards in the US, Renewables Obligation in the UK, and as
the Mandatory Renewable Energy Target in Australia, such legislation obliges
electric utilities to use renewable energy sources to me\et a specified target
percentage of their supply. The aim is to bring "green" energy online
quicker than would otherwise happen by providing incentives for renewables
generation. The targets are mandatory, with financial penalties for those who
fail to meet them.
Property Rights
By setting minimum standards for public exposure to pollutants, governments
give property rights to individuals or groups of individuals that would enable
them to take civil action against polluters who exceed mandated standards.
Table 5. Existing and Emerging Domestic Trading Regimes
7. CONCLUSIONS
This paper has considered the economics of renewable energy technologies
through the quantification in financial terms of the major environmental
externalities of electric power generation, for a range of alternative
commercial and almost-commercial technologies.
It has been shown that estimates of damage costs resulting from combustion of
fossil fuels, if internalised into the price of the resulting output of
electricity, could clearly lead to a number of renewable technologies
(specifically wind and some applications of biomass) being financially
competitive with generation from coal plants. However, combined cycle natural
gas technology would have a significant financial advantage over both coal and
renewables under current technology options and market conditions. Over the next
few decades, the costs of renewable technologies (particularly those that are
"directly" solar-based) are likely to decline markedly as technical
progress and economies of scale combine to reduce unit costs. On the basis of
cost projections made under the assumption of mature technologies and the
existence of economies of scale, renewable technologies would possess a
significant social cost advantage if the externalities of power production were
to be "internalised." Incorporating environmental externalities
explicitly into the electricity tariff today would serve to hasten this process
of transition.
Justification of energy subsidies to developing technologies may be based
upon the desire of a government to achieve certain environmental goals (e.g.,
enhanced market penetration of low GHG emissions technology). However, in
general, case specific direct action is likely to give a more efficient outcome.
Thus penalising high GHG (or other pollutant) emitting technologies not only
creates incentives for "new" technologies, but it also encourages the
adoption of energy efficiency measures with existing technologies and
consequently lower GHG emissions and other pollutants per unit of output. In
addition, if the existence of market failures is restricting the diffusion of
renewable energy technologies, then addressing those failures directly may again
provide an efficient outcome.
The principle of internalising the environmental externalities of CO2
emissions (and other pollutants) resulting from fossil fuel combustion is of
global validity. Whether this is achieved directly through imposition of a
universal carbon tax and emission charges, or indirectly as a result of ensuring
compliance with Kyoto targets and other environmental standards, a similar
result is likely to be achieved. Specifically, a rise in the cost of power
generation based upon fossil fuel combustion and a relative improvement in the
competitive position of an increasing range of renewable energy technologies. In
other words, the removal of both direct and indirect subsidies to power
generation technologies and the appropriate pricing of fossil (and nuclear)
fuels to reflect the environmental damage (local, regional, and global) created
by their combustion are essential policy strategies for stimulating the
development of renewable energy technologies.
The Energy Journal, Vol. 25, No. 3. Copyright 2004 by the IAEE. All rights
reserved.
[dagger] Editor's note: IAEE presidents have an option to publish a paper in
our Journal. Tony Owen, president in 2004, has submitted this paper.
1. Also known as a "Pigouvian" tax.
2. In this paper, the term "externality" will be used only in the
context of "environmental externalities." Non-environmental
externalities in the energy sector, with the exceptions of mining deaths and
traffic accidents, are likely to be relatively minor and site-specific.
3. These steps describe a "bottom up," as distinct from a "top
down," methodology for life cycle analysis. Top-down studies use highly
aggregated data to estimate the external costs of pollution. They are typically
undertaken at the national or regional level using estimates of total quantities
of emissions and estimates of resulting total damage. The proportion of such
damage attributable to certain activities (e.g., the transport sector) is then
determined, and a resulting monetary cost derived. The exercise is generic in
character, and does not take into account impacts that are site specific.
However, its data requirements are relatively minor compared with the
"bottom up" approach. The latter involves analysis of the impact of
emissions from a single source along an impact pathway. Thus all technology data
are project specific. When this is combined with emission dispersion models,
receptor point data, and dose-response functions, monetised values of the
impacts of specific externalities can be derived. Data requirements are
relatively large compared with the "top down" methodology, and
therefore omissions may be significant.
4. The European Commission (EC) launched the project in collaboration with
the US Department of Energy in 1991. The EC and US teams jointly developed the
conceptual approach and the methodology and shared scientific information for
its application to a range of fuel cycles. The main objectives were to apply the
methodology to a wide range of different fossil, nuclear and renewable fuel
cycles for power generation and energy conservation options. Although the US
withdrew from the project, a series of National Implementation Programmes to
realise the methodology for reference sites throughout Europe was completed. The
methodology was extended to address the evaluation of externalities associated
with the use of energy in the transport and domestic sectors, and a number of
non-environmental externalities such as those associated with security of
supply. Krewitt (2002) has provided a critique of the evolution of the
methodologies used in the ExternE analyses.
5. A detailed explanation of these techniques, with practical examples, is
given in Part III of OECD (1994).
6. See Pearce (2003) for a summary of these issues.
7. See, for example, Sorensen (2000) for a more detailed definition of
levelized electricity costs.
8. A high rate of penetration by intermittent renewables without electric
storage could be facilitated by emphasis on advanced gas turbine power
generating systems. Such power generating systems (characterised by low capital
cost, high thermodynamic efficiency, and the flexibility to vary the electrical
output quickly in response to changes in the output of intermittent power
generating systems) would make it possible to back up the intermittent
renewables at low cost, with little need for electrical storage.
9. The costs to the system of coping with unpredictable intermittency in the
UK have been explored by Milborrow (2001).
10. In the case of oil, this issue is covered in greater detail in Owen
(2004).
11. See Aunan (1996) for a survey of exposure-response epidemiological
studies. Rabl and Spadaro (2000) discuss the ExternE methodology used to derive
monetary estimates of health impacts, whilst Pearce (2002) raises questions
regarding the ExternE methodology.
12. European Commission (1998).
13. A number of impacts were ignored either due to their being of a very
minor nature or where insufficient knowledge is available to derive credible
estimates.
14. Rabl and Spadaro (2000).
15. For new baseload plants these damages are likely to be a little lower,
reflecting higher levels of efficiency in power generation. In this context,
Rabl and Spadaro (2000) quote estimates of 2.73 euro/kWh and 1.25 euro/kWh for
coal and gas respectively.
16. The exception being some biomass technologies.
17. In addition, the implicit assumption that the real price of fossil fuels
will remain constant may not be valid.
18. Original data were quoted in US$. At the time of writing, US$1.0 was
equivalent to approximately 1.25 euro.
19. Equity weighting gives a higher weight to damages that occur in poor
countries relative to the same cost of damage in a rich country. It requires the
specification of a social welfare function in order derive the weights. Pearce
(2003) illustrates the effects of equity weighting on damages arising from
climate change.
20. See Sundqvist (2004) for an analysis of the causes of the disparity of
electricity externality estimates.
21. See Section 9 for an extended discussion of the distinction between
control and damage costs.
22. See Missfeldt and Hauff (2004) for elaboration of this point.
23. In the case of the Climate Change Levy in the UK, Pearce (2003) has
calculated implicit carbon tax rates to be 16/tC for coal, 30/tC for gas and
31/tC fore electricity. For a genuine carbon text, of course, these rates should
be identical. Further, the UK government has adopted 70/tC (under review) as its
measure of marginal damage resulting from climate change. So the long-term
carbon tax is a long way from reflecting a true Pigovian tax rate. In contrast,
Pearce notes that the rate of a carbon tax implicit in UK fuel excise duty far
exceeds (by a factor of 5) this 70 figure (which in itself appears to be
unrealistically high).
REFERENCES
Aunan, K. (1996). "Exposure-Response Functions for Health Effects of Air
Pollutants Based on Epidemiological Findings." Risk Analysis 16(5):
693-709.
Energy Information Administration (EIA) (1995). Electricity Generation and
Environmental Externalities: Case Studies. Washington DC: US Department of
Energy.
Eu\ropean Commission (2003). External Costs: Research Results on
Socio-Environmental Damages Due to Electricity and Transport. Brussels.
European Commission (1998). ExternE - Externalities of Energy. Brussels.
Gagnon, L., C. Belanger and Y. Uchiyama (2002). "Life-Cycle Assessment
of Electricity Generation Options: The Status of Research in Year 2001."
Energy Policy 30(14): 1267-1278.
Haites, E. and F. Mullins (2001). Linking domestic and industry greenhouse
gas trading systems. Report prepared for EPRI, IEA and IETA.
Imperial College Centre for Energy Policy and Technology (ICCEPT) (2002).
Assessment of Technological Options to Address Climate Change. A Report to the
Prime Minister's Strategy Unit. December.
International Energy Agency (IEA) (1989). Environmental Emissions from Energy
Technology Systems: the Total Fuel Cycle. Proceedings of IEA/OECD Expert
Seminar, Paris, 12-14 April 1989.
International Energy Agency (IEA) (1999). World Energy Outlook, Looking at
Energy Subsidies: Getting the Prices Right. Paris: OECD/ IEA.
International Energy Agency (IEA) (1999). World Energy Outlook 2002. Paris:
OECD/IEA.
Krewitt, W. (2002). "External Costs of Energy - do the Answers Match the
Questions? Looking Back at 10 Years of ExternE." Energy Policy 30(10):
839-848.
Milborrow P. (2001). PIU Working Paper on Penalties for Intermittent Sources
of Energy. London: The Cabinet Office.
Missfeldt, F. and J. Hauff (2004). "The role of economic
instruments." in A.D. Owen and N. Hanley (eds.) The Economics of Climate
Change. London: Routledge.
Newell, R. and W. Pizer (2003). "Discounting the Future: How Much do
Uncertain Rates Increase Valuations?" Journal of Environmental Economics
and Management 46(1): 52-71.
Organisation for Economic Cooperation and Development (OECD) (2003).
Environmentally Related Taxes Database, http:// www.oecd.org.
Owen, A.D. (2004). "Oil supply insecurity: control versus damage
costs." Energy Policy 32(16): 1879-1882.
Pearce, D. (2003). "The social cost of carbon and its policy
implications." Oxford Review of Economic Policy 19(3): 362-384.
Pearce, D. (2002). "Energy Policy and Externalities: An Overview."
Paper presented at an IEA workshop on Energy Policy and Externalities: The Life
Cycle Analysis Approach, Paris: OECD/IEA.
Philibert, C. (1999). "The Economics of Climate Change and the Theory of
Discounting." Energy Policy 27(15): 913-927.
Rabl, A. and J. V. Spadaro (2000). "Public Health Impact of Air
Pollution and Implications for the Energy System." Annual Review of Energy
and Environment 25: 601-627.
Sorensen, B. (2000). Renewable Energy: Its Physics, Engineering,
Environmental Impacts, Economics & Planning. Second Edition. London:
Academic Press.
Sundqvist, T. and P. Soderholm (2002). "Valuing the Environmental
Impacts of Electricity Generation: A Critical Survey." The Journal of
Energy Literature VIII(2): 3-41.
Sundqvist, T. (2004). "What causes the disparity of electricity
externality estimates." Energy Policy 32(15): 1753-1766.
Tol, R.S.J. (2003). The Marginal Costs of Carbon Dioxide Emissions: An
Assessment of the Uncertainties, Unpublished manuscript.
Weitzman, M.L. (2001). "Gamma discounting." The American Economic
Review 91(1): 260-271.
Anthony D. Owen*
Work undertaken whilst the author was located at the International Energy
Agency, Paris, during the first six months of 2003. The author gratefully
acknowledges the hospitality of the IEA, but bears sole responsibility for the
contents of this paper. Detailed comments and suggestions from Campbell Watkins
were much appreciated.
* School of Economics, The University of New South Wales, Sydney, NSW 2052,
Australia. E-mail: a.owen@unsw.edu.au .
Copyright International Association for Energy Economics
2004