Journal of Policy Modeling
25 (2003) 795–810
Environmental taxation and economic effects:
a computable general equilibrium
analysis for Turkey
Gürkan Selçuk Kumbaroğlu∗
Department of Industrial Engineering, Bogaziçi University, 34 342 Bebek, Istanbul, Turkey
Abstract
This study explores economic effects of environmental taxation using an energy–economy–environment computable general equilibrium model of the Turkish economy. The
model disaggregates the Turkish economy into seven sectors and describes production
within a nested CES representation. Results, which are obtained under a Business-As-Usual
as well as various environmental tax scenarios, provide insight into energy–economy–
environment interactions in Turkey and indicate opportunities for an ecologically and economically sustainable development of the country. Besides general policy implications, it is
found that a second dividend of environmental taxation, that is economic benefits in addition
to environmental improvements, is possible when imported fuels are the primary source of
pollutant emissions. This result has been obtained under tax revenue recycling that assumes
public consumption of tax revenues instead of the common practice of using tax revenues
for reducing existing tax distortions in order to obtain a second dividend.
© 2003 Society for Policy Modeling. Published by Elsevier Inc. All rights reserved.
JEL classification: C68; O21; O52
Keywords: Environmental policy modeling; CGE analysis; Double dividend; Emission taxation;
Energy–economy–environment interactions
∗
Tel.: +90-212-3581540/2079; fax: +90-212-2651800.
E-mail address: [email protected] (G.S. Kumbaroğlu).
0161-8938/$ – see front matter © 2003 Society for Policy Modeling.
doi:10.1016/S0161-8938(03)00076-0
796
G.S. Kumbaroğlu / Journal of Policy Modeling 25 (2003) 795–810
1. Introduction
Parallel to ongoing industrialization, Turkey’s primary energy needs have increased rapidly at an average annual growth rate of 5.1% over the last 50 years. The
per capita energy use still ranks below the world’s average and only 35% of the primary energy consumption in 2001 (78 MTOE) is provided from domestic sources.
The energy imports amount to US$ 8.3 billion, some 20% of Turkey’s total imports
in 2001. The bill can be expected to increase significantly in the coming years as
the Ministry of Energy and Natural Resources forecasts a doubling of the primary
energy consumption by 2010, with the domestic share falling to 27%. Increasing
energy imports might become an essential burden to the Turkish economy as the
country has a persistent current account deficit and relies on foreign exchange inflows to finance her outstanding external debt (which has reached some US$ 114
billion — nearly 78% of GDP — in 2001). Naturally, such implications are vitally
important for the success of the IMF-supported structural reform package.1
The Turkish energy sector is undergoing severe structural changes in a restructuring process to establish competitive energy markets, for which the legal and
regulatory framework has been established in late 2001. The market opening underway is triggered by the hope to achieve efficiency gains and improve welfare.
The prospect for a rapidly developing energy market is attracting private investors,
who are naturally primarily concerned about maximizing profits rather than reducing harmful environmental impacts. Clean energy technologies and emission
reduction measures are therefore to be promoted by policy-makers if ecological
sustainability is desired. This study explores the economic impacts of environmental taxation as a policy instrument to steer the country into a path of sustainable
development.
Sustainability is likely to be an issue on Turkey’s roadway to the EU where more
stringent emission standards are present. The EU has ratified the Kyoto Protocol
on Climate Change committing member countries to reduce their greenhouse gas
emissions. Turkey’s compliance with the Kyoto requirements will most probably
be on the agenda of her accession talks which can be expected to begin by 2005.2
This puts additional emphasis on the importance of the present study highlighting
energy–economy–environment interactions in Turkey.
1 Since 1961, there have been 18 loan arrangements with the IMF to support the Turkish economy. Turkey’s latest stand-by arrangement, a comprehensive program of macroeconomic and fiscal
adjustment, covers the period 2002–2004 with a total Fund financing of some US$ 17 billion.
2 The EU has agreed to decide upon Turkey’s progress in fulfilling the Copenhagen political criteria
in December 2004, and then begin accession talks without delay should the country pass the review.
It should be noted that there is a wide consensus and strong political will in Turkey to undertake
the necessary reforms. On the other hand, Turkey’s joining the EU might yield significant economic
benefits for the Union as indicated by Goldberg and Levi (2000). But, more importantly, Turkey’s
evolution as a secular modern democratic state would be confirmed under a EU membership (that’s
why Turkey’s aspirations to EU accession are also being strongly supported by the US).
G.S. Kumbaroğlu / Journal of Policy Modeling 25 (2003) 795–810
797
Computable general equilibrium (CGE) modeling is the primary analytical
tool available for conducting economic analyses of energy and climate policies.
Commercial software packages such as GEMPACK (Harrison & Pearson, 1996)
or MPSGE (Rutherford, 1995) have taken advantage of recent algorithmic advances triggering the development of large-scale dynamic CGE type of models
(Shoven & Whalley, 1992; among others). Such modeling applications in energy
and environmental policy planning are various. The models GTAP-E (Burniaux &
Truong, 2002), RICE (Nordhaus & Yang, 1996), DREAM (Vennemo, 1997),
WARM (Carraro & Galeotti, 1997) and SCREEN (Frei, Haldi, & Sarlos, 2003;
Kumbaroğlu & Madlener, 2003) can be mentioned as some outstanding examples.
Each model exhibits different originalities, e.g., the GTAP-E model includes a specific treatment of carbon emission trading, WARM decomposes the capital stock
into environment-friendly and polluting parts, DREAM defines environment–
economy feedback links via a health-induced productivity index as a function
of energy consumption, RICE incorporates an environmental damage function
and SCREEN integrates a process-oriented energy–environment linear activity
analysis framework into a CGE setting.
This paper describes the environment–energy–economy model ENVEEM, a
dynamic CGE model with particular reference to energy–environment interactions, presents results from the empirical analysis for Turkey and discusses policy
implications. ENVEEM has been designed to study long-term consequences of
environmental policy decisions on the adoption of energy technologies and on
socio-economic indicators. It is calibrated using Turkish data to reveal opportunities for an ecologically and economically sustainable development of the country.
The following section summarizes the main features of ENVEEM. Section 3
provides an overview on energy use and pollutant emissions in Turkey, defines
different energy and emission tax scenarios considered in the empirical application, presents the results obtained from the scenario analysis, and discusses policy
implications. The final section summarizes the relevant findings for policy-making
and concludes the study.
2. Main features of the model
The model ENVEEM is structurally similar to Goulder’s (1994) energy–
economy CGE model. It is developed as a multisectoral version of the model
MEEET (Arıkan, Güven, & Kumbaroğlu, 1997) and calibrated under the same
benchmark data assumptions. Agents accounted for in the model are producers,
consumers and the government. Consumers maximise utility by consuming commodities subject to endowments of primary factors. Producers maximise profits
subject to available production technologies, which transform factors and intermediates into commodities. The government receives the revenue from indirect and
environmental taxes and spends them on purchasing goods and services at market
prices. The interactions between the involved agents are illustrated in Fig. 1 which
798
G.S. Kumbaroğlu / Journal of Policy Modeling 25 (2003) 795–810
I/O matrix
Intermediate
goods
Energy
Technology matrix
Imported goods
Supply of capital, labor
Producers
Consumers
Supplyofcom modities
Imposition of
indirect and
environmental taxes
Supply of
commodities
Government
Fig. 1. General structure of ENVEEM.
represents the general model structure. A revenue is associated with each arrow
connecting agents in the figure. That is, the government imposes a tax and receives
the tax revenue, producers supply commodities and receive the revenue obtained
from selling the commodities at market prices, consumers supply their labour and
capital endowments to be used in production and, in return, receive a revenue from
the producers.
The model represents the economy as a competitive Arrow–Debreu equilibrium.
Agents’ supplies and demands depend only on relative prices. An equally-proportioned change in all prices neither changes the agents’ budget sets nor their
behaviours. Agents’ behaviours are as follows. On the production side, firms minimise their cost of using labour, capital, energy and intermediate inputs subject
to technological constraints. On the demand side, a representative consumer allocates his income from factor endowments to maximise total utility over the
planning horizon. The government spends all of its tax revenue on the purchase of
domestically produced goods and services.
The model includes seven producing sectors, three of which are the energy
industries aggregated as electricity, oil and gas and solids. The remaining four
sectors disaggregate the Turkish economy into the four sectors of transport, manufacturing and basic industries, services, and others. For each sector, a nested
constant-elasticity-of substitution CES function describes the technological substitution possibilities in domestic production between the input factors of capital,
labour, electricity, oil and gas, and solids. The nesting structure of production is
presented in Fig. 2.
Labour, determined exogenously in the model, and capital are supplied by
households. The exogenously specified growth of labour, together with exogenously specified technical progress, drives the growth of the model. That is, all
quantities grow as a result of the exogenous labour and productivity growth rate
while all prices decay at the interest rate. This gives the model a template for
projecting the initial equilibrium along a dynamic path.
G.S. Kumbaroğlu / Journal of Policy Modeling 25 (2003) 795–810
799
Gross Output
CES
Capital-Labor
Composite
Energy
Composite
Cobb-Douglas
Capital
Intermediate
Inputs
Cobb-Dougl as
Labor
Electricity
Oil & gas
Solid fuels
Fig. 2. Nesting structure of production in ENVEEM.
3. Empirical analysis and results
The subject of the empirical analysis is Turkey. An overview of energy use and
pollutant emissions is provided in the next subsection, followed by an elaboration of model results scrutinizing energy–economy–environment interactions and
discussing policy implications.
3.1. Energy use and pollutant emissions in Turkey
Turkey’s primary energy consumption ranks with 1.11 TOE per capita much
below the International Energy Agency (IEA) average of 5.1 TOE per capita in
the year 1998. The fast increase in primary energy consumption (PEC) and compositionary changes since 1980 are shown in Table 1.
As evident from Table 1, the last decade of PEC is characterized by a spectacular
growth of natural gas, which has been intensively used in electricity generation.3
The main energy source is still oil, with a share of 39.6% of total PEC in year
2000.
Pollutant emissions due to energy use have amounted to 187.5 million tons
CO2 (2.9 tons per capita; 0.41 tons/$1000 GDP), 1.93 million tons SOx (29.8 kg
per capita; 5.2 kg/$1000 GDP), and 940 million tons NOx (14.5 kg per capita;
2.5 kg/$1000 GDP) in 1998, according to IEA-statistics (IEA, 2001). Per capita
emissions are lower than OECD averages (CO2 : 10.9 tons per capita; SOx : 39.2 kg
per capita; NOx : 40.6 kg per capita), and only the SOx emission intensities of GDP
3 64.8% of natural gas consumption in 2000 has been used for generating electricity. The share of
electricity generated from natural gas-fired power plants in total electricity production has risen from
1% in 1995 to 37% in 2000. A detailed analysis on the technological changes in the electricity sector
can be found in Ediger (2003a, 2003b).
800
G.S. Kumbaroğlu / Journal of Policy Modeling 25 (2003) 795–810
Table 1
Primary energy consumption in Turkey (MTOE)
Source
Hardcoal
Lignite
Oil
Natural gas
Hydropower
Wood
Waste
Other
Total
1975
1980
1985
1990
1995
2000a
2.883
1.732
7.958
–
0.261
3.845
2.128
0.042
3.025
2.692
14.178
–
0.508
4.369
2.414
0.195
2.824
3.970
16.074
0.021
0.976
4.730
2.953
0.365
3.775
7.933
18.134
0.062
1.036
5.210
2.539
0.478
6.150
9.765
23.901
3.110
1.991
5.361
1.847
0.507
5.905
10.605
29.324
6.313
3.057
5.512
1.556
0.943
9.983
13.219
32.595
13.327
2.656
5.081
1.376
3.989
18.849
27.381
31.913
39.167
52.632
63.215
82.226
1970
Source: WECTNC (1999).
a Data taken from the WECTNC website (http://www.dektmk.org.tr/turkish/Rapor/2000istatistik/
2000denge1.htm).
are higher than OECD averages (CO2 : 0.61 tons/$1000 GDP; SOx : 2.3 kg/$1000
GDP; NOx : 2.4 kg/$1000 GDP).
Air quality standards in Turkey are presently less stringent than those recommended by the World Health Organization (WHO). The maximum daily averages
(short-term standards) for SO2 and NO2 are, respectively, 400 and 300 ␮g/m3 in
Turkey as opposed to the 125 and 150 ␮g/m3 recommended by the WHO. A similar gap exists in the long-term standards for SO2 and NO2 , as well as for other
pollutants. A more detailed comparison of air quality standards can be found in
IEA (2001).
Environmental impacts of energy use in Turkey have been studied by various
authors (e.g., Demirbaş, 2003; Kumbaroğlu, 1997; Plinke, Haasis, & Rentz, 1990;
Taşdemiroğlu, 1992). These studies all agree on a rapid growth of pollutant emissions, with projected growth rates for total SO2 , NOx and CO2 emissions ranging
from 3.2 to 5.1% p.a. Official forecasts of pollutant emissions from electricity generation project a similar increase as documented in various studies (e.g., Sakaryalı,
Sevgör, Erdoğ, & Yıldıran, 2000; TEAŞ, 1996). These projections do not assume
a diffusion of renewable energy technologies (which is indeed a quite realistic
assumption for the next 10 years, given the current technological and economical
restrictions). But it should be noted that a considerable renewable potential exists
so that a stepwise shift might be a serious alternative for emission mitigation in
the long term (see Ediger & Kentel, 1999 and Kaygusuz & Sarı, 2003 for a more
detailed discussion). Official projections of energy-related pollutant emissions are
based on the IMPACTS module of energy and power evaluation program (ENPEP), as described in Sakaryalı et al. (2000). It should be noted, however, that
the ENPEP methodology used in national energy policy making fails to represent
some crucial energy–economy–environment interactions as discussed in Arıkan
and Kumbaroğlu (2000) and has provided quite inaccurate energy forecasts as
revealed in Ediger and Tatlıdil (2002).
G.S. Kumbaroğlu / Journal of Policy Modeling 25 (2003) 795–810
801
Applied modeling studies exploring the implications of economic policy measures to reduce pollutant emissions, like environmental taxation, are quite rare for
the case of Turkey. In a general equilibrium type optimization format (aggregate
economic equilibrium), Arıkan and Kumbaroğlu (2001) analyze the economic impacts of emission taxation. However, the analytical framework provides a highly
aggregated representation of the nonenergy economy and fails to simulate intersectoral substitution effects. ENVEEM overcomes this deficiency by featuring a
sectoral disaggregation within a CGE representation.
3.2. Empirical results
The model has been calibrated using 1991 data for Turkey. This year has been
chosen as the base year since it is an ordinary year in which no extraordinary political, social or economic events have taken place. The integrated programming environment GAMS/MPSGE has been utilised to formulate the model. Equilibrium
solutions are obtained with the solver PATH (Ferris & Munson, 1999) at intervals of
5-year time periods. A reference run which includes “Business-As-Usual” (BAU)
assumptions has been taken first. The BAU scenario includes no environmental
taxes. The model results for the first period (1995) turned out to approximate actual
realisations fairly closely. Fluctuations from actual values were within a tolerance
limit of +/− 1%. In addition to the BAU scenario, various SO2 emission and sulphur tax scenarios have been defined. The definition of environmental scenarios
is as follows:
Scenario
Definition
S1
S3
N1
N3
SUL5
100 $/ton tax on SO2 emissions
300 $/ton tax on SO2 emissions
100 $/ton tax on NOx emissions
300 $/ton tax on NOx emissions
500 $/ton tax on sulphur contained in fuels
In the definition of the scenarios, the environmental taxes are first applied in year
2005. The amount of tax is set so that high-tax scenarios reduce emissions to satisfy
European regulations which allow maximum annual emissions of 1.8 million tons
of NOx and 2.6 million tons of SO2 (Plinke, 1992).
Results of main economic indicators are summarised in Table 2. The highest GDP losses caused by emission taxes are in the order of 1.5% of BAU-GDP
(caused by scenario S3 in 2020). GDP losses induced by sulphur taxation, on the
other hand, are much higher: scenario SUL5 causes a reduction in BAU-GDP lying
in the range of 3.3–6.9%. Obviously, there is a significant difference in the economic impact that the two tax variants induce. That is, if policy-makers were to
introduce a sulphur instead of emission tax for reducing SO2 emissions, the loss in
GDP to be expected would roughly be four times higher. This finding emphasizes
the significant impact of energy and environmental policy decisions on economic
802
G.S. Kumbaroğlu / Journal of Policy Modeling 25 (2003) 795–810
Table 2
Results of ENVEEM: main economic indicators (×109 $)
GDP
Investment
Exports
Imports
2000 BAU
204.75
57.26
32.28
54.93
2010 BAU
S1
S3
N1
N3
SUL5
319.06
316.24
314.69
318.96
319.03
306.95
73.40
72.80
72.10
73.36
73.32
69.05
68.61
68.17
67.80
68.62
68.66
67.57
105.15
103.94
102.63
105.04
104.92
103.87
2020 BAU
S1
S3
N1
N3
SUL5
475.86
471.92
468.84
475.77
475.99
442.55
124.00
123.56
122.97
123.96
123.93
116.35
129.95
129.07
128.17
129.98
130.09
127.08
164.54
162.40
159.92
164.35
164.15
162.88
2030 BAU
S1
S3
N1
N3
SUL5
759.61
757.83
754.96
759.82
760.47
734.20
290.55
292.09
294.25
290.65
290.79
287.66
227.19
225.79
224.21
227.24
227.41
224.16
240.61
237.36
233.42
240.33
240.01
239.89
growth. A more detailed analysis on the specific costs of the two distinct types of
taxes follows at the end of this section. It is further observed from Table 2 that
there is a slight increase in GDP as a result of NOx emission taxation. This is a
surprising result that deserves more attention as a reduction in GDP would normally be expected in response to a new tax. The unexpected gain occurs because
NOx emissions result, to a large extent, from the consumption of imported fuels
— essentially petroleum in the transport sector. The NOx emission tax scenarios,
therefore, decrease petroleum imports and lead to a shift away from the transport
sector towards other sectors. The rise in the production of other sectors increases
simultaneously those sectors’ exports and leads, together with the fall in fuel imports, to a decline in the foreign trade gap and an increase in GDP. Thus, it is
possible to achieve an economic gain by taxing emissions that originate mainly
from the use of imported energy carriers. This finding can be interpreted as an
indicator of the significant impact of fuel imports on the Turkish economy.4 The
essential discovery from environmental policy scenarios N1 and N3 that should be
underlined, however, is the given opportunity to achieve environmental improvements and economic benefits simultaneously under the described circumstances.
Results in Table 2 further show that the emission tax scenarios reduce investments
4 Using a multivariative multi-attribute hierarchial dynamic technology adoption model, Madlener,
Kumbaroğlu, and Ediger (2003) reveal similar findings for electricity generation investments in Turkey.
G.S. Kumbaroğlu / Journal of Policy Modeling 25 (2003) 795–810
%
40
%
45
35
40
30
35
803
30
25
25
20
20
15
15
10
10
5
5
0
2000
2010
2020
2030 Year
0
2000
BAU Electricity Composition
2010
2020
2030 Year
S3 Electricity Composition
%
40
%
50
hy dro
hardcoal
fu el-oil
45
40
natgas
lignite
nuclear
35
35
30
25
30
20
25
15
20
15
10
10
5
5
0
0
2000
2010
2020
N3 Electricity Composition
2030 Year
2000
2010
2020
2030 Year
SUL5 Electricity Composition
Fig. 3. Electricity composition of selected scenarios (%).
in the short and medium term (up to 20 years) whereas they cause a slight increase in the long term. The short-term decline in investments can be explained by
a general shrinkage of the economy. The long-term increase, on the other hand,
is a result of costly abatement investments as well as more expensive energy investments in cleaner technologies. In the sulphur tax scenarios, investments also
decline in the long term since those scenarios do not encourage abatement technology installation. It is also observed that SO2 emission and sulphur tax scenarios
cause a decrease in foreign trade activities. This is again explained by the general
shrinkage of the economy.
Fig. 3 presents the composition of electric energy in scenarios BAU, S3, N3 and
SUL5. The scenarios S3 and N3 have been selected from the emission tax scenarios
since these are the “high tax” cases of SO2 and NOx taxation and implications are
easier observed.
It can be seen from Fig. 3 that the share of hydroelectric energy declines in all
scenarios. This is because all hydroelectric potential is depleted in all scenarios
by 2025. It is observed that the share of natural gas fired power plants does not
change significantly in response to an environmental tax. This is because natural
804
G.S. Kumbaroğlu / Journal of Policy Modeling 25 (2003) 795–810
Table 3
Results of ENVEEM: energy aggregates (electricity: Twh; oil and gas, solids: MTOE)
BAU
S1
S3
N1
N3
S1N1
S3N3
SUL5
Oil and gas
2010
88.5
2020
133.3
2030
174.3
87.7
131.9
172.4
86.5
130.0
169.6
88.04
132.5
173.3
87.2
131.2
171.5
87.2
131.2
171.4
85.3
128.1
167.1
86.9
130.4
169.0
Solids
2010
2020
2030
52.6
112.4
182.3
49.8
109.2
179.4
58.7
119.6
186.0
58.5
119.3
185.8
55.5
112.1
182.1
49.6
108.7
178.9
48.8
86.9
147.4
58.9
119.9
186.3
Oil and gas/solids
2010
1.50
2020
1.11
2030
0.94
Electricity
2010
283.5
2020
555.7
2030
775.5
1.67
1.17
0.95
278.2
547.3
769.8
1.74
1.19
0.95
276.7
541.8
764.9
1.50
1.11
0.93
283.1
555.1
775.0
1.49
1.10
0.92
282.8
554.6
774.6
1.66
1.17
0.94
278.0
546.9
769.5
1.72
1.18
0.93
276.2
540.9
764.2
1.78
1.50
1.15
266.4
499.0
698.8
gas imports in the BAU scenario are very close to their upper bounds. Similarly, the
maximum limit of nuclear power installation is reached in the BAU scenario and
therefore does not change significantly in response to an environmental tax — but it
still follows an increasing trend in all scenarios. A comparative investigation of the
scenarios’ lignite shares indicates that emission and sulphur taxes both reduce the
share of lignite-based electric energy significantly from its BAU level. A significant
reduction is observed in initial years which slightly recovers later in the emission
tax scenarios. There is no recovery in the sulphur tax scenario since there is no
encouragement for abatement technology installation (i.e., abatement technology
installation does not reduce the tax to be paid by producers). The recovery in the
emission tax scenarios is coupled with abatement technology installation in new
power plants using lignite. However, even in the recovered situation the share of
lignite-based electricity is less in environmental scenarios than it is in the BAU
case.
Table 3 presents energy aggregates; it can be seen that environmental tax scenarios except NOx emission taxes increase the oil and gas/solids ratio. This is
an expected result since most of the SO2 emissions are due to solids (essentially
low-quality domestic lignites). The change in the oil and gas/solids ratio is more
pronounced in early periods when there is greater possibility of substitution. The
scarcity of hydro reserves and upper bounds on natural gas imports inevitably lead
to the utilisation of domestic solid fuel reserves in the long term. It is also observed
that taxation reduces electricity consumption in all scenarios.
Table 4 presents the emission intensity EMI (EMI = total emissions/total primary energy consumed) and energy intensity ENI (ENI = total primary energy
G.S. Kumbaroğlu / Journal of Policy Modeling 25 (2003) 795–810
805
Table 4
Results of ENVEEM: emission and energy intensities (EMI: ton/TOE; ENI: TOE/100 $)
BAU
S1
S3
N1
N3
S1N1
S3N3
SUL5
2000
EMI
ENI
0.048
0.040
0.043
0.040
0.043
0.039
0.048
0.040
0.048
0.040
0.043
0.040
0.043
0.040
0.048
0.040
2010
EMI
ENI
0.046
0.046
0.032
0.044
0.031
0.043
0.046
0.046
0.046
0.046
0.032
0.044
0.031
0.043
0.037
0.044
2020
EMI
ENI
0.055
0.053
0.031
0.052
0.030
0.051
0.055
0.053
0.055
0.053
0.031
0.052
0.030
0.050
0.032
0.049
2030
EMI
ENI
0.058
0.047
0.034
0.047
0.033
0.046
0.058
0.047
0.058
0.047
0.034
0.047
0.033
0.046
0.037
0.043
consumed/GDP) results. It can be seen from Table 4 that in the BAU scenario, EMI
and ENI, respectively, grow at an average of 0.6 and 0.5% per year between 2000
and 2030. It is observed that SO2 emission taxes and the sulphur tax significantly
decrease EMI reflecting more intensive use of lower pollutant-emitting energy
sources. There is no significant effect of NOx emission taxation on EMI. This is
because taxation implies only a minor reduction in NOx emissions, which does
not significantly affect EMI. The energy intensity ENI, on the other hand, is not
significantly affected by taxation. However, a slight reduction of ENI is observed
70
60
Reduction (%)
50
40
30
S1
20
S3
SUL5
10
0
2005
2010
2015
2020
Year
Fig. 4. Reduction in SO2 emissions.
2025
806
G.S. Kumbaroğlu / Journal of Policy Modeling 25 (2003) 795–810
8
S1
7
S3
SUL5
6
Loss (%)
5
4
3
2
1
0
2005
2010
2015
2020
2025
Year
Fig. 5. GDP losses.
in scenarios S3, S3N3 and SUL5 indicating that more extreme tax scenarios induce less energy use. In the present model this is reflected by a shift from energy
towards other inputs; it would reflect energy efficiency improvements if they were
determined endogenously.
The percentage reduction in SO2 emissions and GDP losses are plotted in Figs. 4
and 5, respectively.
It is observed from Figs. 4 and 5 that initially both emission tax scenarios
achieve more emission reduction with less GDP loss than the sulphur tax scenario. However, in the long-term, more emission reduction is achieved with the
sulphur tax scenario although GDP losses of emission taxes remain still below
the loss caused by the sulphur tax. The sulphur tax scenario reduces more emissions in the long term since it achieves emission reduction through a change in
the technological structure of the energy system rather than through an installation
of abatement technologies as is the case in the emission tax scenarios. A change
in the technological structure occurs naturally more slowly due to the putty-clay
structure of the model which necessitates to wait for the retirement of existing
technologies in order to replace them. A glance at the above figures indicates
that in the short term the emission taxes are more effective in reducing emissions since they induce more emission reduction with less GDP loss. However,
due to the change of the picture in the long term, it is worth taking a look at the
marginal abatement costs in order to draw conclusions about the effectiveness of the
taxes.
Fig. 6 presents the marginal abatement costs defined as the ratio of GDP
loss/reduced emissions. It is observed from the figure that there are significant
differences between the marginal abatement costs (MACs) of the sulphur and
G.S. Kumbaroğlu / Journal of Policy Modeling 25 (2003) 795–810
807
9000
8000
S1
7000
S3
SUL5
Cost ($/Ton)
6000
5000
4000
3000
2000
1000
0
2010
2015
2020
2025
2030
Year
Fig. 6. Marginal abatement costs (GDP loss/reduced emissions).
emission tax scenarios. In 2010, the MAC of scenario SUL5 exceeds that of scenario S1 by a factor of five. The gap is reduced in the long term, however the MACs
of the sulphur tax scenario are persistently higher than those of the emission tax
scenarios. Accordingly, it is concluded that a tax on emissions is much more effective in reducing SO2 emissions than a sulphur tax. It should be emphasized that
this effectiveness has economic consequences which allow to save up to 10 billion
dollars in 2010.
4. Conclusions
The use of the dynamic CGE model ENVEEM has been discussed in this paper,
studying the economic implications of environmental tax reforms in Turkey. Interactions of the energy sector with the rest of the economy are elaborated with respect
to various energy and emission tax scenarios. Model results yield several policy
implications, which are essential for an environmentally and economically sustainable development of the country. The main findings from the scenario analyses
highlight the importance of (1) preferring emission taxation as a policy instrument to achieve environmental improvement, (2) substituting oil and gas instead
of hardcoal and lignite to reduce pollutant emissions, and (3) reducing energy
imports to speed-up economic development. It has been found that the economic
burden of not following these policy suggestions might reach considerable levels
amounting to nearly 6% of GDP. Model results further indicate that the above policy suggestions inherit a potential to simultaneously improve both environmental
808
G.S. Kumbaroğlu / Journal of Policy Modeling 25 (2003) 795–810
quality and economic performance in Turkey.
As Turkey possesses considerable coal reserves and negligible oil and gas
reserves,5 the above findings (2) and (3) seem to face policy makers with a trade-off
between reducing energy imports and substituting oil and gas for solid fuels. The
model, however, combines both policies through a reduction in energy use. The
consumption of all energy carriers is reduced (along with a general economic
shrinkage), with solids declining faster than oil and gas. Hence, the oil and gas/
solids ratio increases while imports of oil and gas are declining. Naturally, policymakers should aim to achieve such a result without curtailing economic growth.
These findings emphasize the significance of energy conservation. It is concluded
that policy-makers should support policies to improve energy efficiencies, to increase energy savings, and to reduce the demand for energy-intensive activities.
The prospects for a double dividend of environmental taxation, i.e., improvement in environmental quality and at the same time better economic performance,
are indicated by the NOx emission tax scenarios. This is because NOx emissions originate to a large extent from the consumption of imported fuels, mainly
petroleum products. It has been found that NOx emission taxation decreases the
oil imports of NOx -intensive sectors thereby increasing the gross production of
nonintensive ones as a result of intersectoral substitution — hence GDP increases.
These observations underline the important impact of energy imports, essentially
petroleum, on the Turkish economy. It is concluded that policy-makers should
target a higher domestic share in primary energy consumption.6
The achievement of environmental benefits together with economic ones brings
a new dimension to the discussion of necessary circumstances for obtaining a double dividend, which mainly centers on the use of environmental tax revenues to
correct pre-existing inefficiencies of non-environmental taxes (e.g., Böhringer,
Pahlke, & Rutherford, 1997; Bovenberg & Goulder, 1997; Parry & Bento, 2000).
The results of this research indicate that a second dividend of environmental taxation yielding economic benefits is possible even when the tax revenues are not
recycled to reduce existing tax distortions — provided that imported fuels are the
primary source of pollutant emissions.
As a final remark, I would like to underline that — as opposed to most EU
countries — Turkey is yet in the stage of industrialization with rapidly increasing
energy needs of crucial importance for her economic growth. The development of
sustainable energy and environmental policies is therefore a challenging task, for
5 Proven reserves are some 423 Mton hardcoal, 7339 Mton lignite, 43.7 Mton oil, and 8.8 Bcm
natural gas (WECTNC, 1999).
6 It should be noted that this conclusion poses a dilemma due to limited indigenous resources.
Renewable energy technologies are yet too expensive and technologically not matured for a large-scale
adoption. Domestic lignites on the other hand are of low quality (low calorific values, high emission
factors) and don’t seem to be suitable for use in future advanced coal technologies. Still it seems
inevitable for the Turkish economy, especially after 2010 (as suggested by all scenarios of the study),
to exploit the indigenous lignite reserves unless an equally cheap domestic alternative can be introduced.
G.S. Kumbaroğlu / Journal of Policy Modeling 25 (2003) 795–810
809
which policy-makers need to take assistance from empirical research.
Further research is needed towards exploring the double dividend hypothesis
under different tax revenue recycling assumptions. A more detailed representation
of future energy technologies is required for model refinement. Research is also
further necessary for an explicit modeling of uncertainty in energy demand and
prices (essentially due to energy market liberalization).
References
Arıkan, Y., Güven, Ç., & Kumbaroğlu, G. (1997). Energy–economy–environmental interactions in a
general equilibrium framework: The case of Turkey. In D. W. Bunn & E. R. Larsen (Eds.), Systems
modelling for energy policy (pp. 117–139). New York: Wiley.
Arıkan, Y., & Kumbaroğlu, G. (2000, May). Geçmişten günümüze enerji modelleri (Energy models
from past to the future). Proceedings of the 8th Turkish energy congress: Energy and technology
for sustainable development in the 21st century (pp. 9–20). WEC Turkish National Committee.
Arıkan, Y., & Kumbaroğlu, G. (2001). Endogenising emission taxes: A general equilibrium type optimisation model applied for Turkey. Energy Policy, 29(12), 1045–1056.
Böhringer, C., Pahlke, A., & Rutherford, T. F. (1997). Environmental tax reforms and the
prospects for a double dividend: An intertemporal general equilibrium analysis for Germany.
http://nashcolorado.edu/tomruth/germany/ecotax.htm.
Bovenberg, A. L., & Goulder, L. H. (1997). Costs of environmentally motivated taxes in the presence
of other taxes: General equilibrium analyses. National Tax Journal, 50(1), 59–88.
Burniaux, J.-M., & Truong, T. P. (2002, January). GTAP-E: An energy–environmental version of the
GTAP model. GTAP Technical Paper No. 16. http://www.agecon.purdue.edu/gtap/.
Carraro, C., & Galeotti, M. (1997). Economic growth, international competitiveness and environmental
protection: R&D and innovation strategies with the WARM model. Energy Economics, 19(1), 2–28.
Demirbaş, A. (2003). Energy and environmental issues relating to greenhouse gas emissions in Turkey.
Energy Conversion and Management, 44(1), 203–213.
Ediger, V.Ş. (2003a, January). Elektrik sektöründeki tarihsel yakıt rekabetinin sistem dinamikleri-1
(System dynamics of historical fuel competition in the electricity sector-1). PetroGas, 36, 60–64.
Ediger, V.Ş. (2003b, March). Elektrik sektöründeki tarihsel yakıt rekabetinin sistem dinamikleri-2
(System dynamics of historical fuel competition in the electricity sector-2). PetroGas, 37, 62–67.
Ediger, V. S., & Kentel, E. (1999). Renewable energy potential as an alternative to fossil fuels in Turkey.
Energy Conversion and Management, 40, 743–755.
Ediger, V. S., & Tatlıdil, H. (2002). Forecasting the primary energy demand in Turkey and analysis of
cyclic patterns. Energy Conversion and Management, 43(4), 473–487.
Ferris, M. C., & Munson, T. S. (1999). GAMS/PATH user guide. http://www.gams.com/solvers/
path/main.htm.
Frei, C. F., Haldi, P.-A., & Sarlos, G. (2003). Dynamic formulation of a top-down and bottom-up
merging energy policy model. Energy Policy, 31(10), 1017–1031.
Goldberg, M. A., & Levi, M. D. (2000). The European Union as a country portfolio. European Journal
of Political Economy, 16, 411–427.
Goulder, L. H. (1994). The Goulder model: Structure and simulation results. Energy Modeling Forum
EMF Report 12 (Chapter 8). Stanford University, California.
Harrison, W. J., & Pearson, K. R. (1996). Computing solutions for large general equilibrium models
using GEMPACK. Computational Economics, 9, 83–127.
IEA. (2001). Energy policy of IEA countries. Turkey 2001 review. Paris: OECD/IEA.
Kaygusuz, K., & Sarı, A. (2003). Renewable energy potential and utilization in Turkey. Energy Conversion and Management, 44(3), 459–478.
810
G.S. Kumbaroğlu / Journal of Policy Modeling 25 (2003) 795–810
Kumbaroğlu, G. S. (1997). A model for long-term global air quality and development of efficient
control strategies in Turkey. European Journal of Operations Research, 102, 380–392.
Kumbaroğlu, G., & Madlener, R. (2003). Energy and climate policy analysis with the hybrid bottom-up
computable general equilibrium model SCREEN: The case of the Swiss CO2 Act. Annals of Operations Research, 121(1–4), 181–203.
Madlener, R., Kumbaroğlu, G., & Ediger, V. Ş. (2003, July). Technology adoption modelling in situations of irreversible investments under uncertainty: The case of the Turkish Electricity Supply
Industry. Proceedings of the international conference on policy modeling EcoMod 2003. Istanbul.
Nordhaus, W. D., & Yang, Z. (1996). A regional dynamic general-equilibrium model of alternative
climate-change strategies. The American Economic Review, 86(4), 741–765.
Parry, I. W. H., & Bento, A. M. (2000). Tax deductions, environmental policy, and the double dividend
hypothesis. Journal of Environmental Economics and Management, 39(1), 67–96.
Plinke, E. (1992). Energy and emission control strategies for Turkey. Ph.D. Thesis in the Department
of Economic Sciences, University Fredericiana Karlsruhe, Germany.
Plinke, E., Haasis, H.-D., Rentz, O., & Sivrioğlu, M. (1990). Analysis of energy and environmental
problems in Turkey by using a decision support model. AMBIO, 19(2), 75–81.
Rutherford, T. (1995). Applied general equilibrium modeling with MPSGE as a GAMS subsystem: An
overview of the modeling framework and syntax. http://www.gams.com/solvers/mpsge/syntax.htm.
Sakaryalı, Z., Sevgör, S., Erdoğ, F., & Yıldıran, M. (2000, May). IMPACTS Modeli ile Uzun Dönem
Enerji Kaynakları Arz-Talep Planlama Çalışmalarında Çevre Faktörünün ıncelenmesi (Analyzing
the environmental factor in long-term energy resource supply-demand planning with the IMPACTS
model). Proceedings of the 8th Turkish energy congress: Energy and technology for sustainable
development in the 21st century (pp. 39–53). WEC Turkish National Committee.
Shoven, J. B., & Whalley, J. (1992). Applying general equilibrium. Cambridge University Press, Cambridge.
Taşdemiroğlu, E. (1992). Air pollutant emissions due to energy utilization in Turkey. Energy, 17(1),
95–97.
TEAŞ. (1996). Elektrik Enerjisi Üretiminde Çevre Faktörünün İncelenmesi (Analysis of environmental factors in electric energy production). Turkish Electricity Generation and Transmission Co.
Rep. Nr.: TEAŞ–Ç 96-03-01.
Vennemo, H. (1997). A dynamic applied general equilibrium model with environmental feedbacks.
Economic Modelling, 14(1), 99–154.
WECTNC. (1999). Energy report 1998. World Energy Council Turkish National Committee. Ankara:
Nurol Publishers.