Decomposition and decoupling effects of carbon dioxide emission

ARTICLE IN PRESS
Energy Policy 35 (2007) 3226–3235
www.elsevier.com/locate/enpol
Decomposition and decoupling effects of carbon dioxide emission from
highway transportation in Taiwan, Germany, Japan and South Korea
I.J. Lua, Sue J. Linb,, Charles Lewisc
a
Department of Environmental Engineering, National Cheng Kung University, Tainan 701, Taiwan, ROC
Department of Environmental Engineering, Sustainable Environmental Research Center (SERC), National Cheng Kung University, Tainan 701, Taiwan,
ROC
c
Department of Resources Engineering, National Cheng Kung University, Tainan 701, Taiwan, ROC
b
Received 4 July 2006; accepted 1 November 2006
Available online 16 January 2007
Abstract
We adopted the Divisia index approach to explore the impacts of five factors on the total carbon dioxide emissions from highway
vehicles in Germany, Japan, South Korea and Taiwan during 1990–2002. CO2 emission was decomposed into emission coefficient,
vehicle fuel intensity, vehicle ownership, population intensity and economic growth. In addition, the decoupling effects among economic
growth, transport energy demand and CO2 emission were analyzed to better understand the fuel performance and CO2 mitigation
strategies for each country. From our results, we suggest that the rapid growths of economy and vehicle ownership were the most
important factors for the increased CO2 emissions, whereas population intensity contributed significantly to emission decrease. Energy
conservation performance and CO2 mitigation in each country are strongly correlated with environmental pressure and economic driving
force, except for Germany in 1993 and Taiwan during 1992–1996. To decouple the economic growth and environmental pressure,
proponents of sustainable transport policy in Taiwan should focus on improving the operation and energy use of its highway
transportation system by implementing an intelligent transportation system (ITS) with demand management, constructing an integrated
feeder system, and encouraging the use of green transport modes.
r 2006 Elsevier Ltd. All rights reserved.
Keywords: Divisia index; Road transportation; Decoupling
1. Introduction
A well-established highway transportation system is not
only an important indication of a country’s progress, but it
is also essential for the economic development of a nation.
Concomitant with the rapid growth of Taiwan’s economy,
socio-economic levels and industrial development, is a rise
in transportation demand. By the end of 2004, the energy
consumption and CO2 emission of the transportation
sector reached 15.7 million tons of oil equivalent and
42.5 million tons, respectively, accounting for 22.7% of
total energy consumption and 17.5% of CO2 emission in
Taiwan. According to International Energy Agency (IEA)
statistics, the aggregate energy demand of the global
Corresponding author. Fax: +886 6 2364488.
E-mail address: [email protected] (S.J. Lin).
0301-4215/$ - see front matter r 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.enpol.2006.11.003
transport system increased by 25% from 1990 to 2000. In
the coming three decades, the growth rates of energy
consumption and CO2 emission from highway transportation are expected to be 90% and 87%, respectively, because
of the high demand for transportation. If unchecked, this
situation reveals that the development of a well-established
transportation system could accelerate resource exhaustion
and result in serious environmental pollution, even though
it can boost regional economic growth and living
standards.
Because sustainable development has become an international priority for the 21st century, the energy consumption and CO2 emission from the highway transportation
sector will play a crucial role in economic development,
ecological environment and energy technology. As a result
of the Kyoto Protocol that entered into effect on February
16, 2005, many countries have initiated plans to reduce
ARTICLE IN PRESS
I.J. Lu et al. / Energy Policy 35 (2007) 3226–3235
greenhouse gas emissions and are especially interested in
CO2 mitigation. According to the Key World Energy
Statistics from the IEA (2005), the CO2 released from
Taiwan was 245.21 Mtons, accounting for 1% of world
emissions and ranking Taiwan as 22nd. Although Taiwan
is neither a member of the UNFCCC nor a non-Annex I
country, it held a national energy conference in June, 2005
to consider measures to reduce the indirect impacts of the
post Kyoto Protocol period on industrial development,
economic cooperation and international trade. Our study
tries to identify the major factors affecting the highway
transportation—related CO2 variations in Taiwan and to
assess the linkage effect between environmental pressure
and economic driving force to better understand the
development patterns between transportation energy demand, gross domestic product (GDP) and CO2 emissions.
Furthermore, comparisons with Germany, Japan and
South Korea are made to gain insight for planning the
national highway transportation policy and CO2 mitigation strategies in Taiwan.
Decomposition analysis is a popular and widely used
approach to identify the direct and intricate factors
affecting a system’s emission changes or energy consumption. The application of decomposition analysis has been
conducted by a variety of studies. For example, Howarth et
al. (1991) utilized Laspeyres index and Divisia index to
characterize energy consumption in the manufacturing
sector in eight OECD countries during 1973–1987 and
compared the advantages and disadvantages of these two
methods. The influences of transport activity, the mix of
travel modes, energy intensity, CO2 intensity and fuel mix
on the increase of CO2 emission in nine OECD countries
were explored by Lynn et al. (1996), who discovered that
travel-related activity was the major cause of emission
increase. Lin and Chang (1996) used the Divisia index
approach to examine emissions of CO2, NOx and SO2 from
major economic sectors in Taiwan during 1980–1992. They
found that economic growth had the greatest impact on the
variation of emission intensities during this period, while
the influence of fuel mix was limited. Shrestha and
Timilsina (1996) utilized the Divisia decomposition approach to examine the effects of fuel mix, fuel quality and
generation efficiency from thermal power plants on CO2
intensity in 12 selected Asian countries during 1980–1990.
Greening et al. (1999) adopted the adaptive weighted
Divisia index (with a rolling base year) to analyze energy
consumption and carbon intensity of the freight sector of
10 OECD countries. Key factors affecting the variation of
CO2 emission from the transport sector in Italy in the
period 1980–1995 were identified by Mazzarino (2000). He
decomposed CO2 emission into five components: including
fuel mix, energy intensity, modal structure, transportation
intensity and economic growth and found that the growth
of GDP was the main cause of the increase in CO2. Zhang
(2000) analyzed the relationships of fuel mix, energy
saving, economic productivity and population expansion
to the increase of China’s CO2 emissions during 1980–1997.
3227
Li and Lee (2001) combined structure decomposition,
input–output modeling and integrative index decomposition to decompose CO2 emissions from Taiwan petrochemical industries into eight factors. Gonzalez and Suarez
(2003) decomposed changes in aggregate electric energy
intensity in Spain and found that a considerable reduction
in intensity was primarily due to structural and energy
intensity effects. Paul et al. (2004) selected pollution
coefficient, energy intensity, structural changes and economic activity as primary input to identify the major
factors affecting the energy-related CO2 emissions from the
major economic sectors in India from 1980 to 1996. Their
results showed that economic growth was the most
important component of CO2 emissions. The Laspeyres
index method was adopted by Steenhof (2006) to decompose electricity demand in the industrial sector into
industrial activity, structure share and energy intensity.
He found that industrial activity and fuel shift were the key
factors for the increase of electricity demand in China. Lin
et al. (2006) adopted the simple average Divisia index
approach with a rolling base year to identify key factors
and strategies for industrial CO2 reduction in Taiwan.
In addition, the studies of methodological issues were
explored in recent years. Ang and Lee (1994) analyzed five
methods and found that the adaptive weighting and the
simple average Divisia index methods tended to yield
smaller residuals in decomposition. In order to get perfect
decomposition, handle the zero values in the dataset and
study the decomposition of a differential change, a refined
Divisia index method, logarithmic mean Divisia index
(LMDI) approach, was introduced by Ang et al., 1998.
Ang and Liu (2001) presented a new decomposition
method, log-mean Divisia method I (LMDI I), which had
the desirable characteristics of perfect decomposition and
consistency in aggregation. Divisia index approach was
applied by Choi and Ang (2002) to decompose the
conventional thermal efficiency index in the Korean power
sector. They concluded that thermal efficiency was
improved by 1.1% per year during 1970–1998, which was
higher than the 0.9% improvement per year based on the
conventional method. Ang (2004) compared the decomposition methodologies and concluded that the multiplicative and additive logarithmic mean Divisia index
approach is the best method for theoretical foundation,
adaptability, ease of operation and result interpretation.
Ang (2005) proposed a practical guide for the general
formulation of the logarithmic mean Divisia index (LMDI)
method and used industrial energy consumption and CO2
emissions as examples for realizing the applications and
advantages of the LMDI approach. Ang (2006) examined
and illustrated how the technique of index decomposition
analysis (IDA) provides a bottom-up framework for
economy-wide composite energy efficiency index and
national energy efficiency trend monitoring.
This study aims to assess the relative contributions of
emission coefficient, vehicle fuel intensity, vehicle ownership, population expansion and economic productivity to
ARTICLE IN PRESS
I.J. Lu et al. / Energy Policy 35 (2007) 3226–3235
3228
the increase of highway transportation CO2 emissions in
different countries using the Divisia index method.
Furthermore, the linkage between economic growth and
environmental pressure is explored in order to better
understand the relative variance between GDP vs. transport energy demand and GDP vs. CO2 emission.
2. Methodology
Decomposition methodologies, such as Laspeyres index
and Divisia index, have been used to analyze energy use,
energy intensity and pollution emission. These methods
decompose an object into a multiplicative product of
several components to explore the principal factors
affecting its variation. As Ang and Lee (1994) illustrated
by analysis of two Laspeyres-based methods, two simple
average Divisia methods and the adaptive weighting
method, the adaptive weighting and the simple average
Divisia index methods tend to yield smaller residuals in
decomposition.
Here, we use the simple average Divisia index approach
with a rolling base year to analyze factors that affect CO2
emission from road vehicles:
Qt ¼
Qt E t V t Pt
Gt ,
E t V t Pt G t
(1)
Qt, highway transportation CO2 emission from energy use
for year t (million metric tons); Et, highway transportation
energy consumption for year t (kilotons of oil equivalent);
Vt, the number of motor vehicles for year t (10,000 unit
vehicles); Pt, human population for year t (1000 people);
Gt, the GDP for year t (2000 billion US dollars).
Eq. (1) can be rewritten as
Qt ¼ C t E t V t P t G t ,
(2)
where Ct ¼ Qt/Et, the emission coefficient for year t;
Et ¼ Et/Vt, the vehicle fuel intensity for year t; Vt ¼ Vt/
Pt, the vehicle ownership for year t; Pt ¼ Pt/Gt, the
population growth per unit GDP for year t.
Eq. (2) can be further expressed as the decomposition of
emission growth rate into the sum of the growth rates for
each component; we obtain the following equation by
differentiating both sides of Eq. (2) with respect to time t:
dQt
dC t Qt dE t Qt dV t
¼
þ
þ
dt
Ct
dt
Et
dt
Vt
Qt dPt Qt dG t Qt
þ
þ
.
dt
Pt
dt
Gt
dt
¼ DC þ DE þ DV þ DP þ DG þ RD.
ð5Þ
DC, DE, DV, DP and DG in Eq. (5) represent the Divisia
indices for effects due to changes in emission coefficient,
vehicle fuel intensity, vehicle ownership, population
intensity and economic growth, respectively; RD is the
residual term.
3. Data
The CO2 emission and energy consumption of the road
transportation sector, GDP and population for each
country is adopted from the OECD statistics database,
whereas registered population in Taiwan is from the
Ministry of Interior in Taiwan. The number of motor
vehicles is based on the database of ‘‘world road statistics’’
The number of motor vehicles from 1963 to 1989
was adopted from the dataset of ‘‘International Road
Federation—World Road Statistics (January 8, 2003)’’,
whereas the data during 1990–2004 was observed from the
yearbook
of
‘‘International
Road
Federation—
World Road Statistics.’’ through the help of Ministry
of Transportation and Communications, ROC. Fuel
consumption of the highway transportation sector in
Taiwan is taken from ‘‘Energy Balances in Taiwan
Republic of China 2004’’ (Bureau of Energy Ministry of
Economic Affairs, 2005).
4. Results
4.1. Energy consumption in Germany, Japan, South Korea
and Taiwan
ð3Þ
Integrating both sides of Eq. (3) from year 0 to year t
yields
Z t
Z t
DQt ¼
dlnðC t Þ Qt þ
dlnðE t Þ Qt
0
0
Z t
Z t
þ
dlnðV t Þ Qit þ
dlnðPt Þ Qt
0
0
Z t
þ
dlnðG t Þ Qt .
ð4Þ
0
From the simple average parametric Divisia method, the
integral of Eq. (4) can be estimated by the mean of the
beginning points and end points over a short period of time
because the data in our study is discrete.
Ct
Qt þ Qo
Et
þ ln
DQt ¼ ln
Co
2
Eo
Qt þ Qo
Vt
Qt þ Qo
þ ln
2
Vo
2
Pt
Qt þ Qo
Gt
þ ln
þ ln
Po
2
Go
Qt þ Qo
þ RD
2
According to OECD statistics (2003), the energy consumption in Japan was much higher than other countries,
followed by Germany, South Korea and Taiwan. Energy
consumption of Japan’s highway transport system rose
from 63.1 million tons of oil equivalent (Mtoe) in 1990 to
76.7 Mtoe in 2003, for an annual growth rate of 1.52%
(Fig. 1). Energy requirements by the road transportation
network in Germany also experienced a slight increase. By
the end of 2003, the energy requirement in Germany was
54.0 Mtoe, which was higher than that of in 1990 by 105%,
ARTICLE IN PRESS
I.J. Lu et al. / Energy Policy 35 (2007) 3226–3235
1990 by 19.2%. In Taiwan, the growth of economic
productivity was higher than the demand for traffic fuel.
Thus, the energy intensity was improved by 7.8% from
1990 to 2003.
Energy consumption (kilo-tonnes of oil
equivalent)
although the energy consumption decreased from 1999 to
2003. Despite the fuel consumption in South Korea
being much lower than that of Japan and Germany, the
growth rate of road transportation energy demand in
South Korea was extraordinary. For instance, the aggregate energy consumption increased 15.3 Mtoe during
1990–2003, with an annual growth rate of 7.04%.
Furthermore, the consumption of vehicle fuels in Taiwan
reached 11.3 Mtoe in 2003, which was higher than that of
1990 by 180%, and it continued to increase at a rate of
4.62% per year.
Energy consumption per unit GDP (energy intensity)
from 1990 to 2003 is given in Fig. 2. The performance of
traffic energy consumption in Germany suggested an
improvement by 13.9% from 1990 to 2003. The energy
intensity in Japan’s road transportation sector was fairly
steady even though its energy performance was better than
the other countries. Besides, the fuel demand per unit GDP
in South Korea increased by 1.36% over the 13-year
period. In 2003, the energy intensity of Korea’s road
transportation was 29.6 kiloton of oil equivalent (ktoe) per
billion US dollars (year 2000), which was more than that of
90,000
80,000
Germany
South Korea
4.2. CO2 emission in Germany, Japan, South Korea and
Taiwan
The growth of highway transportation-related carbon
dioxide emission (Fig. 3) in each country had similar
patterns to its energy consumption. The CO2 emission of
road transportation in Japan contributed 39.4 million
metric tons of emission increase through 1990–2003 and
continued to rise at an annual growth rate of 1.49%
because of the strong transportation demand in Japan.
Along with the increase of fuel demand, the aggregate
contribution of CO2 emission from Germany’s road sector
rose from 151.3 million metric tons in 1990 to 158.7 million
metric tons in 2003, with an annual growth rate of 0.37%.
The CO2 released from motor vehicles in South Korea had
a significant influence on global CO2 emission, for the
average annual growth rate was 6.96% during 1990–2003,
Japan
Taiwan
70,000
60,000
50,000
40,000
30,000
20,000
10,000
0
1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003
Year
Fig. 1. Road energy consumption of various countries.
Energy intensity (kilo-tonnes of oil
equivalent/billion US$)
40
35
30
25
20
15
10
5
Germany
South Korea
3229
Japan
Taiwan
0
1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003
Year
Fig. 2. Energy intensity of various countries.
ARTICLE IN PRESS
I.J. Lu et al. / Energy Policy 35 (2007) 3226–3235
3230
even though the emission from South Korea was less
than that of Germany and Japan. Similar to South Korea,
road transport was responsible for more than 80%
of the CO2 emission from Taiwan’s system (Fig. 4).
Because of the heavy reliance on oil consumption by
motor vehicles, emission changes from the road sector
increased 14.7 million tons from 1990 to 2003 at a rate of
4.60% per year.
Similarly, the measure of CO2 emission per unit GDP in
Germany and Taiwan indicates steady improvement, while
trends of CO2 intensity in Japan and South Korea were
fairly steady (Fig. 5). The environmental pressure of
vehicular CO2 emission in Taiwan was much ‘‘heavier’’
than its economic development in comparison to other
countries, but it decreased at a rate of 2.96%. The CO2
intensity in South Korea increased by 18.0% from 1990 to
2003, whereas it declined by 14.0% in Germany for this
period. Furthermore, the growth of CO2 intensity in Japan
was minimal with an annual growth rate of 0.16% over this
period.
4.3. Decomposition of CO2 emission in Germany, Japan,
South Korea and Taiwan
4.3.1. Germany
The net CO2 emission from highway transportation in
Germany increased by approximately 14.0 million metric
tons from 1990 to 2002 (Table 1). Car manufacturing is an
important industry in Germany, a country where the high
ownership rate and the frequent use of motor vehicles are
the major factors affecting CO2 emission; overall, these
contributed to 59.7 million metric tons of emission
increase. Realizing the serious consequences of global
warming, German car manufacturers generally agreed to
fuel price hikes, vehicle taxation, and promotion of the use
of sulfur-free fuel, etc.––all of which were implemented in
the 1990s (Federal Ministry for the Environment, Nature
Conservation and Nuclear Safety, 2002). This resulted in a
reduction of energy consumption per 10,000 unit vehicles
up to 52.1 million metric tons. Not to be overlooked is the
energy demand from economic growth related to an
CO2 emission (million metric tons)
250
200
150
Germany
South Korea
100
Japan
Taiwan
50
0
1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003
Year
Fig. 3. Transport CO2 emission of various countries.
45,000,000
CO2 emission (tons)
40,000,000
35,000,000
Trasnport sector
Aviation
Road transportation
Railway
Water transport
30,000,000
25,000,000
20,000,000
15,000,000
10,000,000
5,000,000
0
1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004
Year
Fig. 4. CO2 emission by tranport mode in Taiwan.
ARTICLE IN PRESS
I.J. Lu et al. / Energy Policy 35 (2007) 3226–3235
3231
CO2 emission intensity (million metric
ton/billion US$)
0.20
0.16
Germany
South Korea
0.12
Japan
Taiwan
0.08
0.04
0.00
1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003
Year
Fig. 5. CO2 emission intensity of various countries.
Table 1
Decomposition of CO2 emission changes for Germany, Japan, South Korea and Taiwan
Country
Year
Index (emission unit: 106 metric tons)a
DC
DE
DV
DP
DG
EC
RD (%)
Germany
1990–1993
1993–1996
1996–1999
1999–2002
1990–2002
0.40
0.17
0.18
0.83
0.04
24.48
14.16
1.83
16.58
52.14
31.90
13.89
7.38
7.71
59.68
6.01
6.40
8.64
5.51
25.48
9.56
7.83
9.03
6.32
31.57
11.36
1.34
9.78
8.89
13.59
0.04
0.00
0.03
0.02
0.06
Japan
1990–1993
1993–1996
1996–1999
1999–2002
1990–2002
0.12
0.16
0.32
0.58
0.54
6.47
10.57
2.14
8.22
11.63
8.73
9.49
3.67
2.72
24.22
7.00
11.59
0.25
5.27
23.59
8.77
13.63
1.74
6.63
30.00
17.08
22.24
6.98
4.72
41.58
0.06
0.09
0.01
0.00
0.34
South Korea
1990–1993
1993–1996
1996–1999
1999–2002
1990–2002
0.06
0.08
0.43
0.07
0.58
6.04
2.12
15.54
7.23
17.56
20.13
19.71
7.77
6.55
57.45
6.48
11.44
2.59
11.04
30.47
7.68
13.12
4.07
12.45
36.04
15.04
19.00
6.71
15.06
42.39
1.24
0.96
0.10
0.43
5.88
Taiwan
1990–1993
1993–1996
1996–1999
1999–2002
1990–2002
0.02
0.06
0.00
0.03
0.05
2.44
1.51
0.74
1.02
0.03
2.94
4.17
3.17
2.35
12.29
3.96
4.30
3.99
1.63
13.78
4.58
4.95
4.75
2.24
16.32
5.94
3.25
3.19
1.97
14.35
0.65
0.13
0.10
0.03
2.73
a
DC, emission factor; DE, energy intensity; DV, vehicle ownership; DP, population intensity; DG, economic growth; EC, emission changes; RD,
residuals.
increase of 31.6 million metric tons of additional CO2
emissions. The population intensity index indicates a
reduction of 25.5 million metric tons in emission because
the population grew slower than the GDP. The emission
factor was extremely small for 1990–2002.
4.3.2. Japan
The aggregate change in emission from the highway
transportation sector in Japan was 41.6 million metric tons
for the period 1991–2002 (Table 1). Emissions related to
economic growth totalled 30.0 million metric tons for this
period. Additionally, the continuous increase in motor
vehicles was another important factor for the rise of CO2
emission. This caused a substantial increase in CO2 of 24.2
million metric tons. Based on the ‘‘Revised Law Concerning the Rational Use of Energy in 1998’’, Japan established
a series of measures to promote motor vehicle fuel
efficiency (The Government of Japan, 2002). Therefore,
the energy required per 10,000 unit vehicles has had a
positive effect on the reduction of CO2 emission since 1999.
The aggregate emission increase based on energy intensity
was 11.6 million metric tons for the past 12 years.
Population growth per unit GDP was the most important
factor for the emission decrease; the total decrease was 23.6
ARTICLE IN PRESS
I.J. Lu et al. / Energy Policy 35 (2007) 3226–3235
3232
million metric tons. The decrease in the emission coefficient
by 0.54 million metric tons indicates the improvement from
fuel switching and the promotion of clean vehicles, even
though motor gasoline and diesel oil are still the most
important fuels used in highway vehicles (Table 2).
4.3.3. South Korea
In comparison to the other countries in our study, South
Korea had the largest increase in CO2 emission by
contributing 42.4 million metric tons of aggregate emission
from 1990 to 2002 (Table 1). However, the CO2 emission
was not constant as evidenced by a 6.7 million metric tons
decrease from 1996 to 1999. The decline in transportation
CO2 emission was due to a major economic recession in
1999 and to a more efficient use of energy during this
period. Being a developing country, the number of motor
vehicles in South Korea grew at a rate of 10.4% per year.
Thus, the aggregate emission produced from the rapid
growth of vehicle ownership was 57.5 million metric tons.
Besides, the economic driving force was an indispensable
factor that affected the variation of CO2 emission. Hence,
the net increase in emission due to economic growth was up
to 36.0 million metric tons. Regarding the relative effect of
population growth on GDP and the resulting CO2
emission, the reduction reached 30.5 million metric tons.
Another important reason for the emission decrease was
the improvement in energy conservation. In total, it caused
a decrease in emission of 17.6 million metric tons from
1990 to 2002. The influence of emission coefficient on
Table 2
Fuel mix of road transportation energy consumption in Taiwan
Year
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
Motor
gasoline
Diesel
oil
1631
1777
1956
2138
2247
2451
2824
3449
4041
4490
4855
5534
5995
6415
6816
7160
7310
7670
7974
8079
8128
8282
8584
8802
1447
1456
1559
1643
1745
1902
2034
2175
2274
2377
2467
2871
3085
3172
3176
3088
3177
3295
3460
3569
3513
3927
3752
3992
Unit: kilotons of oil equivalent
LPG
Others
Electricity
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
8.25
19.61
18.19
8.77
14.94
18.94
17.82
19.50
29.08
0.71
0.72
0.50
0.70
1.94
2.04
2.46
1.61
1.87
1.57
2.91
3.28
0.89
0.59
0.62
11.97
0.78
0.76
1.24
2.35
2.28
3.01
3.23
4.00
0.37
0.48
0.57
0.62
0.68
0.68
0.67
0.72
0.78
0.83
0.96
1.00
1.17
1.24
1.31
1.62
1.72
2.06
2.41
2.94
3.02
3.23
3.33
3.51
carbon dioxide variation was much less obvious compared
to the others; the contribution of aggregate emission
reduction was 0.6 million metric tons.
4.3.4. Taiwan
The aggregate emission change of CO2 in Taiwan
increased 14.4 million metric tons during 1990–2002 (Table
1). The major factor for the increase of CO2 was economic
growth, with an aggregate increase of 16.3 million metric
tons. The number of registered motor vehicles in Taiwan
increased from 10.1 million vehicles in 1990 to 17.9 million
vehicles in 2002. Accompanying the growth of motor
vehicles, the index of per capita vehicle ownership
contributed 12.3 million metric tons of emission increase.
The emission change of population growth per unit GDP
was the main reason for the emission decrease and caused a
reduction of 13.8 million metric tons. Different from the
other countries in this study, the negative contribution of
fuel consumption per 10,000 unit vehicles in Taiwan was
insignificant compared to the other factors, with a total
decrease of 0.03 million metric tons. Furthermore, the
emission coefficient, which contributed 0.05 million metric
tons of emission decrease, had relatively little effect on CO2
reduction. Since economic development is an unavoidable
driving force for the increase of CO2, the government of
Taiwan should strive to improve fuel mix and energy
efficiency by eco-driving, scrapping of old vehicles,
adjusting speed limits, reducing idle vehicles and labeling
fuel efficiency, etc.
4.4. Decoupling index in Germany, Japan, South Korea and
Taiwan
4.4.1. Decoupling of energy consumption from GDP
According to the OECD report (2002), the term
‘‘decoupling’’ means breaking the connection between
environmental pressure and economic performance. In
other words, the decoupling indicator explores the relative
growth rate of various environmental factors and economic
driving force over a given period. Consider the relationship
between GDP vs. energy demand in Germany’s road
transportation (Fig. 6): the decoupling factor varied
between 0.01 and 0.14 from 1990 to 2003. That is, the
value of the decoupling index represents a relative
decoupling effect, which occurred when the growth rate
of energy consumption was lower than the increase of
economic benefit. The growth of transportation energy
consumption in Japan was positive and grew at a higher
rate than its economic productivity. Thus, we classified the
performance of Japan’s road transportation sector as
coupling for the decoupling factor ranged from 0.11 to
0.01 during 1990–2003. Besides, the variation in decoupling factor also implies a decline in energy intensity after
1999 because the requirement of fuel demand tended to
moderate as the GDP grew. During the same period, the
strong demand in fuel consumption caused coupling in
South Korea because the increase in traffic energy
ARTICLE IN PRESS
I.J. Lu et al. / Energy Policy 35 (2007) 3226–3235
consumption was higher than economic growth. As for the
development pattern, the promotion of vehicle fuel
consumption during 1996–1999 was the main reason for
the decrease of the negative value of the decoupling index.
In Taiwan, the variations of energy use and economic
productivity had similar tendencies, but the relative change
of energy consumption was much higher than GDP for the
period 1992–1996. Thus, the index showed a coupling effect
during 1992–1996 and a relative decoupling after 1996.
4.4.2. Decoupling of CO2 emission from GDP
Similar to the decoupling of fuel demand from GDP
growth, relative decoupling was observed in Germany from
1990 to 2003, except for the coupling index in 1993 (Fig. 7).
In fact, the increase of decoupling factor from 1999 to 2003
suggests that the economic development increased over the
period, but the aggregate carbon dioxide by the road
transportation declined steadily. In Japan, it experienced
0.20
Decoupling index*
0.10
0.00
Germany
South Korea
1990 1991
3233
coupling; the decoupling factor was negative. The comparison of the growth rate of traffic CO2 emission with respect
to GDP in Japan increased steadily from 1990 to 1999 and
then decreased for an improvement of CO2 intensity. The
variation of decoupling effect in South Korea suggests a
strong connection between CO2 emission and GDP over
the period. Additionally, the linkage effect of environmental pressure of CO2 and GDP in Taiwan’s road
transportation exhibited relative decoupling, except for
1992–1996.
5. Discussion
Several studies, as mentioned above, have found that
high-energy demand and CO2 emission are inevitable
results of economic development. In this study, the factors
such as fuel mix, vehicle fuel intensity and vehicle ownership, etc. are adopted to explore the variation in CO2
Japan
Taiwan
1992 1993 1994
1995 1996 1997 1998 1999 2000
2001 2002
2003
-0.10
-0.20
-0.30
-0.40
Fig. 6. Decoupling index of road energy demand.
ðEP=DF Þ end of period
,
ðEP=DF Þ start of period
where EP, environmental pressure and DF, driving force (OECD, 2002).
Decoupling factor ¼ 1 0.20
Germany
South Korea
Japan
Taiwan
Decoupling index*
0.10
0.00
1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003
-0.10
-0.20
-0.30
-0.40
Fig. 7. Decoupling index of road CO2 emission.
ðEP=DF Þ end of period
,
ðEP=DF Þ start of period
where EP, environmental pessure and DF, driving force (OECD, 2002).
Decoupling factor ¼ 1 ARTICLE IN PRESS
3234
I.J. Lu et al. / Energy Policy 35 (2007) 3226–3235
emission, but economic growth is still the key factor for the
rapid increase of CO2 during 1990–2002. Being a developing country, the GDP in Taiwan experienced an increased
growth rate of 5.28% per year over the past 13 years. This
effect also is reflected in the growth of transport fuel
consumption and CO2 emission of 80% and 79%,
respectively, from 1990 to 2003. Since sustainable development has become an important topic in the 21st century,
the national transport policy should not only focus on
pursuing economical efficiency, but should also enhance
energy conservation and environmental quality. In addition, road transportation is not only the major north–south
thoroughfare but also the primary transport mode for most
people in Taiwan. Consequently, the strategic measures of
sustainable transport policy should aim at (1) carrying out
high occupancy vehicle control, ridesharing and right of
way controls to restrain the growth of transport volume;
(2) implementing an intelligent transportation system (ITS)
and transportation demand management; (3) restraining
the growth of private vehicles by fuel taxation; (4)
encouraging the use of green transport mode, maximizing
energy efficiency and strengthening the emission standards
of new vehicles; and (5) integrating the land-use planning
and transportation construction projects.
6. Conclusion
In this study, we found that the transportation energy
use and CO2 emission for all compared countries increased
significantly from 1990 to 2003. Taiwan’s economic growth
and vehicle ownership were two important elements
causing the steady increase in CO2 emission. On the other
hand, the growth of population per unit GDP was the key
factor for the emission decrease. Comparing the four
countries in this study, economic development and motor
vehicles growth were the major factors for the rise of CO2
emission, while population intensity had a significant
positive contribution to emission decrease. Additionally,
the vehicle fuel intensity factor contributed to reduction of
CO2 in Germany and South Korea. Despite the improvement of energy intensity, the aggregate emission increase of
vehicle energy use in Japan’s road sector was up to 11.6
million metric tons during 1990–2002.
The decoupling index of transportation energy consumption had consistent growth patterns for CO2 emission
in each country. Regarding the coupling in Germany, the
conservation in fuel consumption and the control of CO2
emission resulted in a relative decoupling effect from 1990
to 2003, except for the coupling in 1993. During the same
period, a strong connection between energy consumption
vs. GDP and CO2 emission vs. GDP were observed in
Japan and South Korea. In Taiwan, environmental
pressure and economic growth were decoupled, except for
1992–1996.
The construction of well-established highway transportation system will improve industrial development, economic productivity and living standards, but it will also
bring serious environment problems, like enormous
energy consumption and increased CO2 emission by the
transportation system in the post Kyoto Protocol
period. Though not yet a member of the UNFCCC, the
CO2 released from Taiwan during the past 10 years
increased substantially, ranking Taiwan as 22nd in the
world. Realizing the importance of global warming
and the seriousness of CO2 emission mitigation, the
government of Taiwan is striving to fulfill its responsibility and obey relevant international environmental
agreements to reduce the indirect impacts of the post
Kyoto Protocol period on industrial development, economic cooperation and international trade. Consequently,
the emphasis for national transport policy in the 21st
century is pursuing sustainable development through the
harmonization of energy consumption, economic development and environmental protection. Since motor gasoline
and diesel oil are still the most important fuels in road
transportation and the decreased effect of clean fuel is
limited, an integrated and sustainable transportation policy
in the future should not only focus on the substitution of
fossil fuels but also should focus on the improvement in
vehicle fuel efficiency, fuel efficiency labeling and adjustment of speed limits, thereby developing inter-modal
transportation, well-developed feeder systems, etc. The
results of this study can be a helpful reference for setting
priority strategies of transportation-related CO2 mitigation
in Taiwan.
Acknowledgments
We wish to thank the National Science Council (Project
NSC 94-2211-E-006-031) for financial support. The
authors appreciate the editor and anonymous referees for
their valuable suggestions regarding the manuscript.
References
Ang, B.W., 2004. Decomposition analysis for policymaking in energy:
which is the preferred method? Energy Policy 32, 1131–1139.
Ang, B.W., 2005. The LMDI approach to decomposition analysis: a
practical guide. Energy Policy 33, 867–871.
Ang, B.W., 2006. Monitoring changes in economy-wide energy efficiency:
from energy–GDP ratio to composite efficiency index. Energy Policy
34, 574–582.
Ang, B.W., Lee, S.Y., 1994. Decomposition of industrial energy
consumption: some methodological and application issues. Energy
Economics 16 (2), 83–92.
Ang, B.W., Liu, F.L., 2001. A new decomposition method:
perfect in decomposition and consistent in aggregation. Energy 26,
537–548.
Ang, B.W., Zhang, F.Q., Choi, K.-H., 1998. Factorizing changes in energy
and environmental indicators through decomposition. Energy 23,
489–495.
Bureau of Energy (Ministry of Economic Affairs), 2005. Energy balances
in Taiwan, Republic of China, 2004.
Choi, K.H., Ang, B.W., 2002. Measuring thermal efficiency improvement
in power generation: the Divisia decomposition approach. Energy 27,
447–455.
ARTICLE IN PRESS
I.J. Lu et al. / Energy Policy 35 (2007) 3226–3235
Federal Ministry for the Environment, Nature Conservation and Nuclear
Safety, 2002. Third report by the government of the Federal Republic
of Germany in accordance with the Framework Convention of the
United Nations.
Gonzalez, F.P., Suarez, P.R., 2003. Decomposing the variation of
aggregate electricity intensity in Spanish industry. Energy 28, 171–184.
Greening, L.A., Ting, M., Davis, W.B., 1999. Decomposition of aggregate
carbon intensity for freight: ends from 10 OECD countries for the
period 1971–1993. Energy Economics 21, 331–361.
Howarth, R.B., Schipper, L., et al., 1991. Manufacturing energy use in
eight OECD countries: decomposing the impacts of changes in output,
industry structure and energy intensity. Energy Economics 13,
135–142.
International Energy Agency, 2005. Key World Energy Statistics 2005.
International Road Federation, 2003. World Road Statistics.
Lee, C.F., Lin, S.J., 2001. Structural decomposition of CO2
emissions from Taiwan’s petrochemical industries. Energy Policy 29,
237–244.
Lin, S.J., Chang, T.C., 1996. Decomposition of SO2, NOx and CO2
emissions from energy use of major economic sectors in Taiwan. The
Energy Journal 17, 1–17.
3235
Lin, S.J., Lu, I.J., Lewis, C., 2006. Identifying key factors and strategies
for reducing industrial CO2 emissions from a non-Kyoto protocol
member’s (Taiwan) perspective. Energy Policy 34, 1499–1507.
Lynn, S., Lee, S., Nancy, K., 1996. CO2 emissions from passenger
transport. Energy Policy 24, 17–30.
Mazzarino, M., 2000. The economics of the greenhouse effect: evaluating
the climate change impact due to the transport sector in Italy. Energy
Policy 28, 957–966.
OECD, 2002. Indicators to measure decoupling of environmental pressure
from economic growth.
Shrestha, R.M., Timilsina, G.R., 1996. Factors affecting CO2 intensity of
power sector in Asia: a Divisia decomposition analysis. Energy
Economics 18, 283–293.
Steenhof, P.A., 2006. Decomposition of electricity demand in China’s
industrial sector. Energy Economics 28, 370–384.
The Government of Japan, 2002. Japan’s Third National Communication
under the United Nations Framework Convention on Climate
Change.
Zhang, Z., 2000. Decoupling China’s carbon emissions increase from
economic growth: an economic analysis and policy implications.
World Development 28 (4), 739–752.