The Misinterpretation of Pigouvian Taxes

Kategori	Artikler
Underkategori(er)	n/a
År	2013
Forfatter(e)	Annegrete Bruvoll
	https://vista-analyse.no/site/assets/files/5754/2013_the_misinterpretation_of_pigouvian_taxes.pdf
Last ned	file_download (308.1 kB)
Les i nettleser	find_in_page

Journal of Environmental Protection, 2013, 4, 154-160 
doi:10.4236/jep.2013.48A1017 Published Online Augus t 2013 (http://www.scirp.org/journal/jep) 
The Misinterpretation of Pigouvian Taxes 
Annegrete Bruvoll 	 	
Research Department, Statistics Norway, Oslo, Norway. 
Email: annegrete.br uvoll@vista-analyse.no 
 
Received April 26	th, 2013; revised June 1	st, 2013; accepted July 15	th, 2013 	
 
Copyright © 2013 Annegrete Bruvoll. This is an open access article distributed under the Creative Commons Attribution License, 
which permits unrestricted use, distributio n, and reproduction in any medium, provided the original work is properly cited. 
ABSTRACT 
The tax revenues from Pigouvian taxes are difficult to calculate. The optimal pollution taxes are weighted averages of 
Ramsey taxes and Pigouvian taxes, and the entanglement of environmental, fiscal and other taxes complicate the meas- 
urements. The present international environmental tax statis tics do not rely on a theoretical basis, and include several 
fiscal and resource taxes. St arting with the additivity theorem by Sandmo [1], we propose theoretically consistent 
guidelines for environmental tax statistics. Calculations base d on this framework prove that the present international 
official statistics (Eurostat, IEA, OECD) severely overestim ate environmental taxes. The lack of theoretical consistency 
forms arbitrary results and the statistics used as measures of environmental taxes may cause major flaws in research and 
policy-making based on such statistics. 
 
Keywords: Environmental Taxes; Fiscal Taxes; Pigouvian Taxes; Ramsey Taxes 
1. Introduction 
Taxes and tradable permits are cost efficient instruments 
in internalizing negative externalities. Statistics on the 
use of such taxes can be usable in several contexts. In 
international policy comparisons, environmental taxes 
are frequently used as indicators of environmental con- 
cerns. Environmental tax statistics are potentially valu- 
able inputs in environmental instruments efficiency ana- 
lyzes, and the environmental economics literature advo- 
cates recirculation of emission tax and permit revenues to 
reduce efficiency losses, c.f the double dividend litera- 
ture (Bovenberg [2], Goulder and Parry [3]). Great effort is devoted to measuring environmental 
taxes and tax revenues in international statistical agencies. 
The OECD, Eurostat and the IEA have set in place ex- 
plicit routines for the definition, collection and publish- 
ing of environmental tax data across OECD countries 
(Eurostat [4,5]. OECD [6,7]), hereafter referred to as the 
“ international statistics ”. In their ongoing work on a 
manual for environmental eco nomic accounting, the UN 
seems to follow the same approach (United Nations [8]). 
In this article, we show that the international statistics 
are misleading in that they depart from the theoretical 
definition of environment taxes. According to economic theory, environmental taxes 
should correct the marginal e ffects of environmental ex- ternalities, and the optimal environmental tax rate equals 
the marginal damage cost (Pigou [9]). The tax revenue 
becomes the product of the environmental tax and the tax 
base. When moving from theory to practice, measuring 
the extent of environmental tax rates and revenues is far 
from straightforward for two main reasons. First, the 
presence of both environmental and fiscal taxes influ- 
ences the optimal tax structure (Sandmo [1]). Only part 
of the tax on the externality should count as correction 
for the externality. Second, the entanglement of different 
forms of taxes complicates the statistical definitions. Part 
of the taxes on polluting goods relates to externalities, 
while other parts may be politically motivated in fiscal 
concerns, income distribution or as corrections for other 
market failures. Although theory advocates one instru- 
ment per target, in practical policy, each political target 
involves use of several type
s of instruments, and each 
instrument is adjusted to achieve several targets (Bye and 
Bruvoll [10]). This complicat es the calculations of how 
large parts the tax should be ascribed to the environ- 
mental elements. Taxes may be described and incorrectly 
labeled as environmental, even if there is no or only a 
small direct link with the environmental externalities. The statistics developed by the OECD, Eurostat and 
the IEA cover all tax bases that have a “… particular 
environmental relevance”, i.e., taxes related to energy , 
transport and pollution , as well as taxes levied on re-

sources (Eurostat [4], OECD [6,7])	1. The statistics then 
clearly include additional elements, such as fiscal taxes 
and resource rents, and can per definition not be inter- 
preted as environmental taxes as defined by economic 
theory. 
Data failures may have important implications for sci- 
entific analyses, and for the comparisons of environmen- 
tally friendly policies. Sterner and Köhlin [13], for in- 
stance, use the international statistics and conclude that 
European countries have consistently higher levels of 
environmental tax revenue than the USA, and Ekins [14] 
states that the environmental taxes comprise an increas- 
ing proportion of total tax burden in European countries 
up to 1994. Eurostat [5] concludes that there has been a 
steady fall in environmental taxes from 2003 to 2008, 
after which the level has remained constant. These con- 
clusions are not reliable, since the data include large 
shares of other taxes than environmental taxes. In an 
econometric analysis, Morley [15] rejects the double 
dividend theory	
2 applying the Eurostat data set, and con- 
cludes that environmental taxes have had a negative ef- 
fect on economic growth. In fact, what this study shows 
is that environmental taxes and fiscal taxes reduce 
growth, which does not enlighten the double dividend 
theory. Failures in the statistical foundation may lead 
scientists and politicians to draw misleading conclusions 
on the causes and effects of environmental policy. In the 
end, such information failures can bring along wrong 
advice to the decision makers and a less efficient envi- 
ronmental policy. This paper discusses the separation of environmental 
taxes from other taxes in light of tax theory and suggests 
solutions to the classification problems disturbing the 
official international statistics. Following Pigou [9], the 
proper tax base is the environmental externality, or the 
most closely related tax base. We take the additivity 
theorem by Sandmo [1,16] as a starting point, and pro- 
pose theoretically consistent guidelines on how to disen- 
tangle and consistently calculate the environmental tax 
elements. The aim is to improve the data quality and the 
scientific basis for analyses of environmental instru- 
ments. We illustrate the consequences of the measurement 
problems empirically by calculating the revenues from 
the environmental taxes based on our guidelines and 
compare our results with the tax revenues defined and published as official international statistics, taking Nor- 
wegian taxes as an example. Our conclusion is that the 
present 
international statistics tend to significantly over- 
estimate the taxes and tax revenues; in our example, by 
more than five times the actual environmental taxes. This 
gives an incorrect picture of the environmental policy in 
and between countries and the potential efficiency im- 
provement by tax revenue recycling. Using the interna- 
tional statistics as environmental tax estimates, analytical 
research on efficiency, revenue recycling, double divi- 
dend, and environmental impacts based may be severely 
flawed. 	
2. Theoretical Framework 
Principally, Pigouvian taxes internalize environmental 
externalities into markets (Pigou [9], Sandmo [1]). Ac- 
cording to economic theory, an optimal environmental 
tax should be levied directly at the externality and equal 
the marginal damage cost ( MDC) of the pollution. In 
contrast, the purpose of fiscal taxes is to raise revenue for 
publicly provided goods. Fiscal taxes should be levied 
where they are least likely to distort economic activity to 
avoid any deadweight loss. Under simplifying assump- 
tions, e.g., disregarding extern alities, the fiscal tax rate 
on a good should then be inversely proportional to the 
corresponding own price elasticity of demand (Ramsey 
[17], Diamond and Mirrlees [18]). For polluting goods	
3, total taxes should vary according 
to the externality rule and the elasticity of demand by 
taking into account the additivity theorem in Sandmo 
[1,16]. Importantly, the optimal taxes will not simply be 
the sum of the fiscal tax and the MDC, rather a weighted 
average of the tax computed under the Ramsey inverse 
elasticity rule and the Pigouvi an marginal social damage. 
More formally, let the weights be a and (1 − a ), and T	
R 
the inverse elasticity of demand. The optimal tax rate for 
the good then becomes: 	
()1
01,	R	TaT aMDC a
=+−
<< (1) 
where the first element reflects the fiscal element, and 
the second element reflects the Pigou element. The pa- 
rameter a then reflects the tightness of the government’s 
budget constraint or the marginal cost of public funds 
(Sandmo [16]). Consistent guidelines for environmental tax calcula- 
tions should take this additivity theorem into account. 
We continue by developing a framework including this 	
1Indeed, the international framewor k uses the term environmentally related taxes in their original definition. However, the meaning is 
clearly to illuminate the trends in environmental policy, and the term 
“environmental taxes” is normally used in reports of the international 
statistics. E.g. Eurostat [5] uses this concept when reporting the latest 
version of the international data base, other examples are European 
Environment Agency [11] and OECD [12]. 
2The double dividend refers to an increase in environmental taxes re-
ducing levels of emissions as we ll as increasing economic growth. 	
3A polluting good should be defined as closely to the externality as possible, i.e. in most cases in terms of environmentally damaging emis-
sions. It may also be in terms of the quantities of goods causing the 
emission. E.g, since the CO	
2emissions and the use of gasoline are 	proportional, carbon taxes on petrol can be levied at the energy use, 
rather than the emissions.

element. When decomposing the total revenue into its 
respective fiscal and environmental elements, each as- 
cribed part ( a T	
R and (1 − a ) MDC ) will be equal to or 
lower than the respective optimal fiscal and Pigouvian 
taxes ( T	
R and MDC ). As such, the higher the govern- 
ment’s budget requirement, the lower the weight of the 
environmental element. Hence, the environmental tax 
rate counts for only a share of marginal social damage: 	
()	 1
Environmental tax a MDC MDC =− ≤	. (2) 
The theoretical framework refers to an optimal situa- 
tion where the value of a is a simple function of the 
shadow prices on the public and private budget con- 
straints. Thus, a is the same for all goods and all tax rates. 
In practice, the formulation of the tax system generally 
deviates from the theoretical framework because of mar- 
ket failures, conflicting political stands, pressure groups 
and considerations other than pure efficiency concerns. 
Consequently, a varies over goods. In practice, goods- 
and tax-specific estimates do not exist, and so it is not 
clear how one should decompose an observed tax rate 
into a fiscal contribution and the share ascribed to MDC. 
This is not an argument against using the general princi- 
ples underlying the optimal tax structure as guidelines 
when evaluating the actual tax structure. It is difficult to 
see how one can avoid arbitr ariness in the evaluations 
when using a theoretically opt imal structure as a refer- 
ence for the evaluation. We should regard the weighted 
average in (1) as one such principle. Moreover, as we 
demonstrate, one can derive policy-relevant conclusions 
concerning the actual tax structure without knowledge of 
the weights. We provide two separate cases, when the applied tax 
rate T is set higher and lower than T*, T* being the opti- 
mal tax rate where T* = MDC = MAC . 	
Figure 1	 illustrates the case the tax is higher than the 
optimal tax rate, under the general assumptions of in- 
creasing marginal abatement costs and increasing mar- 
ginal damage costs. In this case, the tax is also higher 
than the marginal damage at the actual emission level, T 
= T	
h > T* >MDC' . The revenue from the tax equals both 
shaded areas. This revenue in cludes more than the opti- 
mal Pigouvian tax and the marginal social damage. Con- 
trary to what may be perceived as standard intuition, part 
of the total revenue, i.e., corresponding to at least T	
h − 
MDC' , is then classified as fiscal and should be sub-
tracted from the revenue. The question is how much. As 
long as a in practice is unknown, we do not obtain any 
assistance from theory. To define the entire MDC' as the 
environmental tax is rather radical in that it implies that 
the marginal cost of funds ( a) is zero, cf. (2). An ap- 
proximation is to identify MDC' times the emission as an 
estimate of the maximal environmental tax revenue. 
Discrimination of e.g. carbon taxes are common out of 	
 	
Figure 1. Tax higher than marginal damage cost. 	
 
competitiveness concerns (Babiker et al. [19]). Typically, 
taxes are relatively high for emission sources with low 
price elasticites, such as tran sport and residential heating 
(Bruvoll and Larsen [20]). Clearly, it is debatable what is 
the “correct” level of MDC estimate for greenhouse gas 
emissions and other externalities. This must rely on dif- 
ferent estimates in the literature. Given that the MDC 
estimates are agreed upon, the taxes higher than MDC 
estimates have no foundation in environmental taxation, 
but rather involve efficiency losses, since the marginal 
abatement costs are larger th an the marginal benefits. 
Instead, these should be counted as non-environmental, 
or fiscal, taxes. Also, the fact that the Ramsey elasticity 
rule closely corresponds with the differentiation of the 
taxes, indicates that fiscal considerations heavily influ- 
ence the politicians’ determinations of carbon tax levels. We find many examples of taxes set equal to the esti- 
mated marginal costs. In ot her instances, taxes are dif- 
ferentiated and certain industries face lower tax rates 
than the assumed MDCs out of competitiveness concerns. 
This represents the cases where T = T	
l ≤ T* , the tax is 
equal to or lower than the marginal damage, T	l ≤ MDC'' 
at the actual emission level, see 	
Figure 2	. Typically, sec- 
tors with high price elasticites are exempt from the tax, 
or face lower rates (Bruvoll and Larsen [20]). Even if the 
tax is lower than the MDC, the environmental tax is prin- 
cipally lower than the app lied tax level, given that a > 0. 
Again, defining the entire revenue as environmental tax 
revenue implies an assumption that the marginal cost of 
funds, and hence the Ramsey tax, is equal to zero. As 
above, given that a is unknown, the revenue from the tax 
T	
l can be defined as an estimate for the maximal envi- 
ronmental tax revenue. This implies the following rules for the calculation of 
environmental taxes and their respective revenues: 
• 	Rule 1: If the tax rates on pollution are higher than the 
MDC , the revenues according to the estimates of the 
MDC equal the estimates of the maximum environ- 
mental tax revenues .

Figure 2. Tax lower than marginal damage cost. 	
 
• 	Rule 2: If the tax rates on pollution are equal to or 
lower than the MDC, the full revenues equal the esti- 
mates of the maximum environmental tax revenues . 	
• 	Rule 3: If a is known, the maximum environmental 
tax should be corrected accordingly by a rate (1 − a ) 
to obtain a closer estimate of the environmental taxes 
and the accordant environmental tax revenues. 	
3. Use of the Framework and Comparison with the Official International Statistics 
To investigate the impact of using a theoretically based 
framework, we calculate the related revenues and com- 
pare them with the numbers reported in the international 
statistics (Eurostat [5]). 
As the MDC values cannot be observed in the market, 
our calculations are based on MDC estimates	
4. Accord- 
ing to our Rule 1, we subtract the difference between the 
tax rates and the MDC estimates when the tax rates are 
higher than the MDC estimates to obtain the maximum 
environmental tax revenues. Total taxes on greenhouse 
gases contribute to the largest share of the total income 
from taxes on emissions in Norway, amounting to 1081 
million euros, see 	
Table 1	. Due to the global, long-term 
and uncertain nature of the effects of greenhouse gases 
and varying effects across regions, it is particularly 
problematic to choose MDC estimates for these emis- 
sions. As a point of departure, we used the highest price 
of emission permits in the EU ETS 2007 (25 euros per 
tonne CO	
2). This price level is al so in line with the aver- 
age Norwegian CO	
2 tax, which may be a proxy of the 
willingness to pay for CO	
2 reductions in Norway. The 
tax on CO	
2 emissions from petroleum activities on the 
continental shelf and on gasoline were about 50 - 80 per- 	
Table 1. Real revenues from Norwegian environmental 
taxes, and revenues according to the OECD, Eurostat and 
IEA definition in millions of euros, 2007. 	
 Revenues from 
environmental 
taxes OECD, 
Eurostat, 
IEA statistics	1	
Taxes on greenhouse gases ( Rule 1) 715 1081 	
Taxes on sulfur ( Rule 1) 
Taxes on NOx, waste incineration, 
beverage containers, pesticides, 
trichloroethane and tetrachloroethane (Rule 2) 246 248 
Petrol and diesel taxes ( Rule 2) 655 1816 
Other taxes according to the OECD, Eurostat, IEA guidelines	
2 5083 
Total 1616 8228 	
1Data reported to Eurostat/OECD (Næss and Smith 2009, Eurostat 2011). 2Motor vehicle registration tax, reregistration tax on motor vehicles, annual 
motor vehicle tax, annual weight-based tax on motor vehicles, electricity 
consumption tax, tax on mineral oils, tax on lubricating oil, base tax on 
disposable beverage packaging. 
 
cent higher and taxes on greenhouse gases from landfills 
four times higher than the MDC estimate. Rule 1 also 
applies for the sulfur tax on mineral products. This tax 
addresses environmental externalities associated with 
acid rain. Relative to the target set in the Gothenburg 
Protocol, the tax is 13 percent higher than the corre- 
sponding marginal abatement cost (Ministry of Finance 
[22]). We reduced the revenue accordingly. 
The remaining environmental taxes in 	
Table 1	 are 
emission taxes set equal to or lower than according MDC 
estimates, c.f. Rule 2. These taxes include taxes on NO	
x, 
emissions from waste incineration, environmental costs 
associated with beverage containers, pesticides, tri- 
chloroethane and tetrachloroet hane, and the share of the 
petrol and diesel taxes including road transport related 
emissions and noise	
5. 
According to Sandmo and Rule 3, the fiscal element 
should also be separated from the taxes. Principally, cor- 
rect adjustment requires a specific a for each taxed good, 
but such estimates are not available. Estimates on the 
average marginal cost of funds in the Norwegian econ- 
omy vary between 20 and 50 percent (Brendemoen and 	
5The taxes on NO	x emissions are designed to meet agreed targets under 	the Gothenburg Protocol. The taxes are lower than the estimated opti-
mal tax to reach the emission target (Ministry of Finance [22]), T < 
MDC, cf. Figure 2. The Norwegian tax on waste incineration is ad-	justed to correspond to the estimated MDC. The taxes on pesticides,
trichloroethylene and tetrachloroethylene are considered to have ful-
filled their intentions with respect to emission reductions (Ministry o	
f 
Finance [22]), cf. a reducti on to the optimal level at MDC = MAC in the 
figures above. The taxes on beverage containers assumedly reflect the 	
MDC (Raadal et al. [23]). The MDCs related to local road emission and 
noise are estimated to constitute about a quarter of total taxes on petrol 
and diesel fuels (ECON [24]). 	
4Environmental economics theory offe rs several methods for estimating the values of environmental externalities as approximations of the true 
shadow prices, the most important being contingent valuation, hedonic 
pricing, travel costs, choice experiments and market-	based methods 
(see, e.g., Pearce and Turner [21]).

Vennemo [25], Holmøy and Strøm [26]). The rule of 
thumb in the Norwegian public sector allows for an av- 
erage marginal cost of 20 percent (Ministry of Finance 
[27]). Using this as an estimate, the environmental tax 
revenues should be further reduced. If we use 20 percent 
as an approximate, we estimate the total environmental 
taxes in Norway to about 1300 euros. 
The International Statistics 
The Norwegian tax revenues according to the definition 
for the international statistics , published in the OECD/ 
Eurostat/IEA database, amount to 8200 million euros 
(Næss and Smith [28], Eurostat [5]), cf. 	
Table 1	. This is 
more than five times the environmental tax revenues ac- 
cording to our economic theory-based definition. The 
difference is partly due to the reduction of the surplus 
fiscal elements according to Rule 1 outlined above. The 
main reason however, is the inclusion of a range of 
purely fiscal taxes in the international statistics. As 
pointed out also by OECD [7], creating any definition of 
environmentally related taxes is problematic. Taxes have 
been implemented for a number of reasons, and some 
taxes have likely been implemented without stringent 
assessment of the costs and damages of the pollution. 
This contributes to explain why the international statis- 
tics rely on a broad definition. The EU, OECD and IEA 
refer to an “environmenta lly related tax” as “A tax whose 
tax base is a physical unit (or a proxy of it ) of something 
that has a proven , specific negative impact on the envi- 
ronment . It was decided to include all taxes on energy 
and transport in the definition of environmental taxes. 
Value added type taxes ar e excluded from the definition 
(Eurostat [4])	
6. As all manufacturing requires factor in- 
puts with a negative environmental impact, and all taxes 
and charges affect pollution through the equilibrium ef- 
fects, this definition in principle covers all taxes and 
economic activities. The framework excludes some large 
tax bases, such as VAT and taxes on oil and gas, but 
without any principal justification	
7. 
The lack of theoretical f oundation creates disturbing 
adjustments that increase the gap between statistics and a 
consistent and interpretable framework. For example, the 
statistics include the full revenues from petrol and diesel taxes, 
i.e., the costs related to road usage and accidents, 
and all other fiscally motivated transport-related taxes 
except VAT. E.g., the numbers include the motor vehi- 
cles registration tax , whose original purpose was to cre- 
ate revenue for the state (Min istry of Finance [22]), see 
line 4 in 	
Table 1	. Indeed, the tax has a differentiated 
element according to CO	
2 emissions, but defining this as 
environmental would in any case imply double counting, 
as CO	
2 and the other environmental externalities from 
road traffic are already intern alized in user-dependent tax 
bases. This also applies for the annual tax on motor vehi- 
cles that is higher for vehicles without particle filters. 
The purpose of other transport-related taxes (the reregis- 
tration tax, annual tax and the weight-based tax on mo- 
tor vehicles) is mainly to fund infrastructure. The international statistics also include the entire 
electricity consumption tax on electric power . Consump- 
tion of electricity has no known adverse environmental 
effects; the externalities origin from the production of 
power. The electricity good is based on a mixture of 
production technologies, ranging from zero emissions 
(e.g. wind, nuclear and hydro power) to emission inten- 
sive technologies (fossil fuels), and consumption taxes 
are only consistent to theory if they constantly vary with 
the current energy mix. According to theory, any pollu- 
tion related to production should be regulated at the 
emission sources. Given taxes on power production, 
consumption taxes should not be included as it would 
imply double counting. Finally, 100 percent of Norwe- 
gian electricity production is based on renewable energy 
sources, and hence emission free. Hence, our judgment is 
that the Norwegian electricity consumption tax does not 
include any Pigouvian tax elements. The purpose of the 
tax on lubricating oils is to fund a collection system and 
the responsible handling of oil waste. Consequently, it 
should be considered a charge covering the costs of 
abated emissions, not an environmental tax covering the 
costs of the remaining emissions. For the remaining taxes 
included by Eurostat and other (the taxes on mineral oils 
and disposable beverage packaging ), we base our judg- 
ment on the Ministry of Finance [22], and conclude that 
these taxes are pure fiscal taxes. Environmental tax and revenue calculations require 
careful examination of the se t of special duties in the 
light of tax theory for each country, evaluations of the 
marginal damage costs and the marginal rate of fund. All 
environmental tax calculations are thus encumbered with 
uncertainty. For example, the marginal costs of green- 
house gas emissions are particularly uncertain and sub- 
ject to normative judgments. It is clearly debatable 
whether the EU ETS price, as we used, is the most rele- 
vant estimate. The MDC may be defined as the necessary 
global carbon price to stab ilize the atmospheric concen- 
tration at a widely agreed level. Globally efficient solutions 	
6A specified list of tax bases is incl uded in the guidelines (Eurostat [4]).7Neither VAT nor resource taxes s hould be included according to the-
ory, as neither are directed toward externalities. VAT is fiscal taxes, 
and resource taxes are normally imposed to tax economic rents on the 
extraction of natural resources (including Ricardian, Hotelling and 
monopoly rents). However, the reas oning for excluding them is not 
based on theoretical principals. VAT are excluded on the erroneous 
argument that they do not influence relative prices in the same way as 
taxes on environmentally related tax bases and Eurostat justifies the 
exclusion of resource taxes by invoking the wide variation in revenue 
from taxing oil and gas and across nati onal tax systems. Hence, data is 
adjusted to avoid variation in the use of instruments.

implies the lowest possible price, while the price may be 
higher than any current CO	
2 tax if only a group of coun- 
tries should contribute to the necessary emission reduc- 
tions. However, these questions are out of scope of this 
paper, and do not influence the principal theoretical 
problems in the production of environmental tax statistics. 
Our aim in the empirical part of the paper is to illustrate 
the potential gains from basing the statistics on a theo- 
retically consistent framework. Even with a wide range 
of uncertainty, our comparisons strongly indicate that the 
current international statistics present estimates that are 
too imprecise to indicate the extent of environmental 
taxes. 	
4. Discussion 
International environmental tax statistics could be valu- 
able inputs in research and political decisions, and as 
indicators for the use of economic instruments in the en- 
vironmental policy and the environmental protection 
over time and across countries. Creating any definition of 
environmentally related taxes is inherently problematic. 
The entanglement with fiscal and other taxes complicates 
the calculation of these statistic s. The problem is twofold. 
First, part of any tax on polluting emissions includes a 
fiscal element. Contrary to standard approximation, the 
environmental tax element is generally even lower than 
the marginal damage cost in the presence of fiscal taxes. 
This should be corrected wh en calculating environmental 
tax revenues. Second, the pres ent official international 
statistics incorrectly include pure fiscal taxes and other 
non-environmental tax bases. Due to such conceptual 
challenges, the definition underlying the official interna- 
tional environmental tax statistics include a too wide 
range of tax bases. This paper suggests theoretically consistent guidelines 
for the calculations of environmental taxes, starting with 
the theoretical definition of Pigou taxes and the additivity 
theorem (Sandmo [1]), and illustrates the potential meas- 
urement failure empirically. Using Norwegian taxes as an example, we find that the 
tax revenues as published by official international statis- 
tics (OECD, Eurostat and IEA) are several times higher 
than the environmental tax revenues following our theory 
based definition. This is a serious warning that scientists 
using the international data can draw misleading conclu- 
sions on the causes and effects of environmental policy 
(see, e.g., Ekins [14], Sterner and Köhlin [13], Morley 
[15], Eurostat [5]). 
It is fundamental for the reliability that the underlying 
principles for the official da ta are based on robust and 
consistent theory. The framew ork presented in this paper 
represents principal economic theory and significantly 
improves the statistics on environmental tax revenues, 
and the statistics’ relevance to research and international comparisons of Pigouvian taxes. 	
5. Acknowledgements 
The author is grateful to Torstein Bye and Erling Holmøy 
for valuable comments on an earlier version of this pa- 
per. 	
REFERENCES 	
[1] 
A. Sandmo, “Optimal Taxation in the Presence of Exter- 
nalities,” Swedish Journal of Economics , Vol. 77, No. 1, 
1975, pp. 86-98. doi:10.2307/3439329	
 
[2] A. L. Bovenberg, “Green Tax Reforms and the Double 
Dividend: An Updated Reader’s Guide,” International Tax 
and Public Finance , Vol. 6, No. 3, 1999, pp. 421-449. 
doi:10.1023/A:1008715920337	
 
[3] L. H. Goulder and I. W. H. Parry, “Green Tax Reform 
and the Double Dividend,” AERE Newsletter, Vol. 20, No. 
1, 2000, pp. 9-13. 
[4] Eurostat, “Environmental Taxes—A Statistical Guide,” 
Office for Official Publications of the European Commu- 
nities, Luxembourg, 2001. 
[5] Eurostat, “Taxation Trends in the European Union. Data 
for the EU Member States, Iceland and Norway,” Euro- 
stat Statistical Books , Publications Office of the European 
Union, Luxembourg, 2011. 
[6] Organisation for Economic Co -operation and Develop-
ment (OECD), “Database on Environmental Policy In-
struments,” 2010. www.oecd.o rg/env/policies/database 
[7] Organisation for Economic Co -operation and Develop- 
ment (OECD), “Taxation, Innovation and the Environ- 
ment, OECD Green Growth Strategy,” OECD, Paris, 2010. 
[8] United Nations, “Outcome Paper for Global Consulta-
tion,” Issue 6: Environmenta lly Related Taxes, Depart-
ment of Economic and Social Affairs, Statistics Division, 
New York, 2011. 
[9] A. C. Pigou, “The Economic s of Welfare,” Macmillian, 
London, 1920. 
[10] T. Bye and A. Bruvoll, “Multiple Instruments to Change 
Energy Behaviour: The Emperor’s New Clothes?” Energy 
Efficiency , Vol. 1, No. 4, 2008, pp. 373-386. 
doi:10.1007/s12053-008-9023-9
 
[11] European Environment Agency, “Market-Based Instru- 
ments for Environmental Policy in Europe,” EEA Tech- 
nical Report No. 8, European Environment Agency, Co- 
penhagen, 2005. 
[12] OECD, “The Political Economy of Environmentally Re- 
lated Taxes,” OECD, Paris, 2006. 
[13] T. Sterner and G. Köhlin , “Environmental Taxes in 
Europe,” Public Finance and Management , Vol. 3, No. 1, 
2003, pp. 117-142. 
[14] P. Ekins, “European Envir onmental Taxes and Charges: 
Recent Experience, Issues and Trends,” Ecological Eco- 
nomics , Vol. 31, No. 1, 1999, pp. 39-62. 
doi:10.1016/S0921-8009(99)00051-8	
 
[15] B. Morley, “Empirical Evidence on the Effectiveness of

The Misinterpretation of Pigouvian Taxes 
Copyright © 2013 SciRes. JEP 
160 
environmental Taxes,” Bath Economic Research Papers, 
No. 2, University of Bath, Bath, 2010. 
[16] A. Sandmo, “The Public Econo mics of the Environment,” 
Oxford University Press, Oxford, 2000. 
doi:10.1093/019829798X.001.0001	
 
[17] F. Ramsey, “A Contribution to the Theory of Taxation,” 
Economic Journal , Vol. 37, No. 145, 1927, pp. 47-61. 
doi:10.2307/2222721	
 
[18] P. Diamond and J. Mirrlees, “Optimal Taxation and Pub- 
lic Production I: Production Efficiency,” American Eco- 
nomic Review , Vol. 61, No. 1, 1971, pp. 8-27. 
[19] M. H. Babiker, P. Criqui, A. D. Ellerman, J. M. Reilly 
and L. L. Viguier, “Assessing the Impact of Carbon Tax 
Differentiation in the European Union,” Environmental 
Modeling and Assessment , Vol. 8, No. 3, 2003, pp. 187- 
197. doi:10.1023/A:1025543207122	
 
[20] A. Bruvoll and B. M. Larsen, “Greenhouse Gas Emis- 
sions in Norway: Do Carbon Taxes Work?” In: T. Sterner 
and A. Muller, Eds., Environmental Taxation in Practice , 
Ashgate Publishing Limited,	
 Aldershot, 2006. 
[21] D. W. Pearce and R. K. Tu rner, “Economics of Natural 
Resources and the Environment,” Harverster Wheatsheaf, 
Ashford Colour Press Ltd., Gosport, Hampshire, 1990. [22] 
Ministry of Finance, “An Ev aluation of Excise Taxes,”	
 	
Ministry of Finance (Norway), Oslo, 2007. (in Norwe- 
gian) 
[23] H. L. Raadal C. A. Nyland , I. S. Modahl and O. J. 
Hanssen, “Environmental Valu ation of Recyclable PET 
Drinking Bottles in Norway,” Stiftelsen Østfoldforskning, 
Fredrikstad, 2003. (in Norwegian) 
[24] ECON, “Marginal External Transport Costs,” Rapport 
2003-054, ECON Analyse, Oslo, 2003. (in Norwegian) 
[25] A. Brendemoen and H. Vennem o, “The Marginal Cost of 
Funds in the Presence of En vironmental Externalities,” 
Scandinavian Journal of Economics , Vol. 98, No. 3, 1996, 
pp. 405-422. doi:10.2307/3440734	
 
[26] E. Holmøy and B. Strøm, “T he Social Cost of Govern- 
ment Spending in an Economy with Large Tax Distor- 
tions: A CGE Decomposition for Norway,” Discussion 
Papers 396, Statistics Norway, Oslo, 2004. 
[27] Ministry of Finance, “Benef it-Cost Analyses,” Ministry 
of Finance (Norway), Oslo, 1998. (in Norwegian) 
[28] E. M. Næss and T. Smith, “Environmentally Related 
Taxes in Norway. Totals and Divided by Industry,” 
Documents 5, Statistics Norway, Oslo, 2009.

Parse pdf-doc by clicking the button below. Parsed pdf-documents will be searchable.

Metadata Found
Author
CreationDate
2013-08-09T17:17:42+08:00
ModDate
2013-08-26T08:13:32+02:00
Producer
Acrobat Distiller 9.0.0 (Windows)
Pages
7