Global Environmental Change 23 (2013) 1807–1818

Contents lists available at ScienceDirect

Global Environmental Change

journal homepage: www.elsevier.com/locate/gloenvcha

Better cars or older cars?: Assessing CO2 emission reduction potential

of passenger vehicle replacement programs
Shigemi Kagawa a,f,*, Klaus Hubacek b, Keisuke Nansai c, Minori Kataoka a,
Shunsuke Managi d, Sangwon Suh e, Yuki Kudoh f
a
Faculty of Economics, Kyushu University, Fukuoka, Japan
b
Department of Geographical Sciences, University of Maryland, College Park, MD, USA
c
Research Center for Material Cycles and Waste Management, National Institute for Environmental Studies, Ibaraki, Japan
d
Graduate School of Environmental Studies, Tohoku University, Sendai, Japan
e
Bren School of Environmental Science and Management, University of California, Santa Barbara, CA, USA
f
Research Institute of Science for Safety and Sustainability, National Institute of Advanced Industrial Science and Technology, Ibaraki, Japan

A R T I C L E I N F O A B S T R A C T

Article history: The primary motivation for the vehicle replacement schemes that were implemented in many countries
Received 29 November 2012 was to encourage the purchase of new cars. The basic assumption of these schemes was that these
Received in revised form 14 July 2013 acquisitions would benefit both the economy and the environment as older and less fuel-efficient cars
Accepted 21 July 2013
were scrapped and replaced with more fuel-efficient models. In this article, we present a new
environmental impact assessment method for assessing the effectiveness of scrappage schemes for
Keywords: reducing CO2 emissions taking into account the rebound effect, driving behavior for older versus new
CO2 emissions
cars and entire lifecycle emissions for during the manufacturing processes of new cars. The assessment of
Vehicle replacement schemes
the Japanese scrappage scheme shows that CO2 emissions would only decrease if users of the scheme
Life cycle assessment
Cost–benefit analysis retained their new gasoline passenger vehicles for at least 4.7 years. When vehicle replacements were
Climate change mitigation restricted to hybrid cars, the reduction in CO2 achieved by the scheme would be 6–8.5 times higher than
Rebound effect the emissions resulting from a scheme involving standard, gasoline passenger vehicles. Cost–benefit
analysis, based on the emission reduction potential, showed that the scheme was very costly. Sensitivity
analysis showed that the Japanese government failed to determine the optimum, or target, car age for
scrapping old cars in the scheme. Specifically, scrapping cars aged 13 years and over did not maximize
the environmental benefits of the scheme. Consequently, modifying this policy to include a reduction in
new car subsidies, focused funding for fuel-efficient cars, and modifying the target car age, would
increase environmental benefits.
ß 2013 Elsevier Ltd. All rights reserved.

1. Introduction emissions associated with car use, the Japanese government also
launched a vehicle scrappage scheme to promote the replacement
Vehicle replacement schemes, such as the ‘‘Car Allowance of old vehicles with new fuel-efficient vehicles. The Japanese
Rebate System program’’ in the USA, the ‘‘Scrappage Scheme’’ in scheme consisted of a lump-sum subsidy of 250,000 yen
the UK, and similar programs in countries such as Austria, Canada, (equivalent to approx. 3189 US dollars on August 6, 2012) being
China, France, Germany, Italy, Romania and Spain, have featured paid to the consumer of a new car for their old car if it had been
prominently in the economic stimulation packages that were registered for the first time at least 13 years previously and then
implemented in response to the global economic crisis (European scrapped (Ministry of Land, Infrastructure, Transport and Tourism,
Automobile Manufacturers’ Association, 2012; Japan Automobile Japan, 2012). What is not yet known is the nature of the impact of
Manufactures Association, 2012; Executive Office of the President these schemes on greenhouse gas and how many years the
of the United States, 2009; Institute for Energy and Environmental consumer who participated in the scheme should continue to own
Research Heidelberg, 2009; The World Bank, 2012). On April 10, his/her new car in order to maximize the environmental benefits of
2009, to stimulate the economy and reduce greenhouse gas the scheme. Answering these questions will clarify whether or not
these vehicle replacement schemes pass the cost–benefit test.
Recent studies (Abrams and Parsons, 2009; Sachs, 2009)
* Corresponding author. Tel.: +81-92-642-2489; fax: +81-92-642-2489. analyzed the gasoline-reduction impacts brought about by
E-mail address: [email protected] (S. Kagawa). replacing old vehicles (average fuel economy: 15.8 mi/gal) with

0959-3780/$ – see front matter ß 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.gloenvcha.2013.07.023
1808 S. Kagawa et al. / Global Environmental Change 23 (2013) 1807–1818

new vehicles (average fuel economy: 25.0 mi/gal) purchased greenhouse gas emissions. As in Abrams and Parsons (2009), we
through the Car Allowance Rebate System program in the USA. also conducted a cost–benefit analysis based on the environmental
The findings of these studies revealed that the Car Allowance Rebate benefits of the scrappage scheme and the governmental cost.
System program would decrease gasoline consumption by 280 gal Finally, we also suggest possible improvements to the scrappage
per car per year, which is equivalent to a reduction of 2.4 Mt of CO2 a scheme.
year. On the other hand, Knittel (2009) stressed the importance of The rest of this paper is organized as follows. Section 2 describes
considering the energy rebound effect in which new, more fuel- four policy scenarios for different car scrappage schemes, Section 3
efficient vehicles, are likely to be driven further than old vehicles formulates the analysis framework developed in this study and
(Greene, 1992; Jones, 1993; Goldberg, 1998; Greene et al., 1999; describes data sources and parameter settings, Section 4 describes
Greening et al., 2000; Small and Van Dender, 2007). In addition, life cycle CO2 emissions associated with the scrappage scheme,
Knittel (2009) also estimated the range in the implied cost of the Section 5 analyzes how changes in annual driving distance on CO2
reductions in CO2 and criteria pollutants (NOx, VOCs, PM10 and CO) emissions can be avoided through implementing the scrappage
under the Car Allowance Rebate System program by incorporating scheme, and Section 6 presents the discussion and policy
the rebound effect into a CO2 accounting formula. Although these implications.
previous studies (Abrams and Parsons, 2009; Sachs, 2009; Knittel,
2009) explicitly contributed to quantifying the environmental 2. Four policy scenarios
benefits of the US Car Allowance Rebate System program, they did
not consider CO2 generated by all of the supply chains associated According to a report by the Next Generation Vehicle Promotion
with the manufacture of new vehicles purchased through the Car Center, Japan (2010), 731,057 cars were scrapped and replaced as
Allowance Rebate System program, or emissions associated with the part of the vehicle scrappage scheme. We compared four
disposal of old vehicles (Lane, 2006). A vehicle life cycle perspective alternative policy scenarios (A–D). Under the baseline scenario,
is required in order to more accurately evaluate the environmental case A, in which no scrappage scheme was implemented, the
benefits resulting from implementation of the system. number of new cars would gradually increase and the number of
An important life cycle analysis study of the US Car Allowance old cars would naturally decrease over time according to a lifetime
Rebate System program was that of Lenski et al. (2010). They function (Kagawa et al., 2011) (blue line in Fig. 1). Eventually, the
proposed a life cycle assessment framework for analyzing the net entire stock of old cars would be replaced by new cars and the stock
environmental effect of the Car Allowance Rebate System program of old cars would be zero after the remaining lifetime of 12 years
on greenhouse gas emissions and found that, in the USA, (Fig. 1). In case A, it is assumed that consumers purchase ‘‘new
reductions of 3.7 million metric tons could be achieved over the gasoline passenger cars’’ (average fuel consumption: 15.8 km/
expected remaining lifetimes of the old vehicles retired through l = 37.2 mi/gal) immediately after relinquishing their old cars,
the program. The merits of their life cycle analysis approach which are older than 13 years (average fuel consumption: 11.8 km/
(Lenski et al., 2010) can be summarized as follows: (i) Life-cycle l = 27.8 mi/gal). As in case A, case B describes the situation without
greenhouse gas emissions associated with the production of a new a scrappage scheme; however, under this scenario, the newly
vehicle were estimated using the ‘‘Greenhouse gases, Regulated purchased cars are hybrid passenger cars (average fuel consump-
emissions, and Energy use in Transportation’’ models (Argonne tion: 35.5 km/l = 83.5 mi/gal).
National Laboratory, 2009), (ii) the annual driving distance of In case C, the vehicle scrappage scheme is implemented and
vehicles tends to decrease with vehicle age, (iii) the environmental the switch to ‘‘new gasoline passenger cars’’ occurs as the old cars
benefits associated with the US Car Allowance Rebate System are scrapped. In this case, all of the old cars (i.e., cars 13 years
program were reduced through the energy rebound effect, which and older) are replaced as part of the scheme in the first year after
meant that switching from old vehicles to new, more fuel efficient the new subsidy is implemented. Consequently, the cumulative
vehicles increases mileage. scrappage rate in the first year is 1, and this does not change over
Lenski et al. (2010) assumed that if the Car Allowance Rebate time (red line in Fig. 1). Case C also assumes that consumers
System program was not introduced, all of the participants would replace their old cars in the first year that the policy is
have retained their old vehicles for the remainder of the vehicles’ implemented and they simultaneously purchase ‘‘new gasoline
lifetime, and that all of the old vehicles would have been scrapped passenger cars’’. The life-cycle CO2 emissions for cases A and C can
after 2.52 years. However, this assumption is not plausible,
because vehicle lifetime is stochastic in reality. Furthermore, as
discussed in Suh et al. (2004), the bottom–up-type greenhouse
gases, regulated emissions, and energy use in Transportation
models used in Lenski et al. (2010) had a system boundary
problem in which the analysis framework did not cover
greenhouse gas emissions associated with all of the passenger
vehicle supply-chains. To address these issues, we developed an
analytical method by combining a top–down-type environmental
input–output life-cycle assessment framework (Moriguchi et al.,
1993; Lave et al., 1995; Hendrickson et al., 1998; Lenzen and Dey,
2002; MacLean and Lave, 2003a, 2003b; Hertwich, 2011),
providing a theoretical framework for a complete system
boundary, with vehicle lifetime distributions analysis (Müller,
2006; Müller et al., 2006; Kagawa et al., 2011; Pauliuk et al., 2012).
Although a previous study (Kagawa et al., 2011) found that
extending the lifetime of a vintage vehicle helps to reduce life-
cycle CO2 emissions throughout the supply chain, the study did not
analyze the environmental benefits of the scrappage scheme. The Fig. 1. Cumulative scrappage rate and four alternative policy scenarios (A–D). (For
modeling technique proposed in this article makes it possible to interpretation of the references to color in this figure legend, the reader is referred
precisely estimate the impacts of the scrappage scheme on to the web version of this article.)
 S. Kagawa et al. / Global Environmental Change 23 (2013) 1807–1818 1809

be estimated using Eqs. (8) and (10) shown in Section 3, the number of scrapped cars during a single year, y, was estimated
respectively. using Eq. (4).
In case D, the scrappage scheme is introduced (as in case C), but For the benchmark analysis, the age of all old cars scrapped as
the scrapped cars are replaced with ‘‘new hybrid passenger cars’’ part of the vehicle replacement scheme (henceforth, target car
(Fig. 1). The life-cycle CO2 emissions for cases B and D can be age) was assumed to be ‘‘13 years’’, and the focus was placed on
estimated using Eqs. (9) and (11) shown in Section 3, respectively. the trend in scrapping these old cars if the vehicle replacement
This study did not consider the aging of ‘‘new cars’’ under any of the scheme had not been implemented. In the absence of data on the
policy scenarios, and it was assumed that the car stock remained age composition of the scrapped cars, we considered a range of
constant at 731,057 cars. target car age scenarios in the sensitivity analysis, and present
the results in Section 5.First, the case in which 731,057 old cars
3. Methodology are scrapped in accordance with an appropriate lifetime
distribution shall be considered. In this case, by taking the
3.1. Estimation of the number of cars scrapped in situations with and number of cars remaining on the day when vehicle age reaches
without the vehicle replacement scheme 13 years (end of Year 13), Q ð13Þ, as 731,057, and substituting
this value into the right-hand side of Eq. (4), the number of cars
During the period when the vehicle replacement scheme was scrapped annually in Year 14 and thereafter can be estimated as
implemented – from April 10, 2009 to September 7, 2010 – the Dð yÞ ¼ 731; 057  ’ð1
yÞ’ð y1Þ
’ð13Þ ð y ¼ 14; 15; . . .Þ or alternatively
number of new, standard-size passenger cars that were exchanged Dðt Þ ¼ 731; 057  lðt Þ where
for cars at least 13 years old was 731,057 (Next Generation Vehicle
Promotion Center, Japan, 2010). In other words, if the vehicle ’ðt þ 13Þ  ’ðt þ 12Þ
lðtÞ ¼ ðt ¼ 1; 2; . . .Þ
replacement scheme had not been implemented, it is considered 1  ’ð13Þ
that some of the 731,057 old cars would have remained in use after
September 2010 and that these cars would eventually have been
3.2. Estimation of the life-cycle CO2 emissions from cars
scrapped at a natural pace in accordance with lifetime distribution.
The vehicle replacement scheme is thus considered to have
Next, we present a method for estimating life-cycle CO2
decreased the lifetime of the cars.
emissions from cars. If the average fuel consumption (km/l) of a
Based on the results of a recent statistical analysis (Kagawa et
car manufactured in Year s is represented by eðsÞ, and the average
al., 2011), the cumulative scrappage rate for a car that is purchased
annual mileage of a standard-size passenger car is represented by d
in Year 0 (new) and which is scrapped in Year y is known to follow
(km), the annual gasoline consumption of one standard-size
the generalized gamma function described by Eq. (1).
passenger car manufactured in Year s, EðsÞð1Þ, can be calculated
 b 
y using the expression EðsÞ ¼ eðdsÞ. Therefore, tank to wheel CO2
’ðy; a; r; bÞ ¼ P ;r ðy > 0Þ (1)
emissions derived from gasoline consumption generated during
a
car use can be estimated as shown below using specific emissions,
where a represents a scale parameter, and b and r represent shape fg, which represents CO2 emissions (t CO2-eq) directly produced by
parameters. P represents the incomplete gamma function, Pðz; aÞ ¼ the combustion of 1 l of gasoline.
R z t a1 R1
1
G ðaÞ 0 e t dt with G ðaÞ ¼ 0 et t a1 dt (Cohen and Whitten,
1988; Cohen, 1991). In this case, the number of cars purchased in d
Rg ðsÞ ¼ EðsÞ  f g ¼  f (5)
Year 0 is taken as C0 and the cumulative number of cars scrapped eðsÞ g
by the end of Year y becomes C 0 ’ð yÞ. Therefore, the number of cars
Also, if life-cycle CO2 emissions generated during gasoline
remaining at the end of Year 13 can be determined by subtracting
refining are taken as fc (t CO2-eq/l) (see Section 2.3 for the
the cumulative number of scrapped cars at the end of the 13th year,
estimation method), then CO2 emissions from gasoline refining can
C 0 ’ð13Þ, from the number of new cars purchased in Year 0, C0, as
be calculated as follows:
shown in the following equation:
d
Q ð13Þ ¼ C 0  C 0 ’ð13Þ ¼ C 0 ð1  ’ð13ÞÞ (2) Rc ðsÞ ¼ EðsÞ  f c ¼  f (6)
eðsÞ c
When y  13 is true for vehicle age, then y, from the cumulative
Therefore, the annual CO2 emissions from running one old
scrappage rate, ’ð yÞ and the single-year scrappage rate in Year y
standard-size passenger car manufactured in Year s can be
can be represented as ’ð yÞ  ’ð y  1Þ. Therefore, the total
obtained as follows:
scrappage rate for when vehicle age 13 years can be determined
by summing scrappage rates after Year 13, as follows: RðsÞ ¼ Rc ðsÞ þ Rg ðsÞ (7)
X
1
On the other hand, if a new car has been purchased, it is
ð’ð yÞ  ’ð y  1ÞÞ ¼ ’ð1Þ  ’ð13Þ (3)
necessary to consider CO2 emissions resulting from the manufac-
y¼14
ture of the new car and CO2 emissions resulting from scrapping
A characteristic of the generalized gamma function is that ’ð1Þ the old car. According to the Embodied Energy and Emission
converges to 1, so the remaining rate at the end of 13th year after Intensity Data for Japan using 2005 Environmental Input–Output
the day of registration can be represented as ’ð1Þ  ’ð13Þ ¼ Table published by the National Institute for Environmental
1  ’ð13Þ. Therefore, the number of cars scrapped in a single year, Studies of Japan (2012), the life-cycle CO2 emissions per million
y, when y  13, can be represented as follows: yen (producer’s price) spent on cars, from materials to parts,
assembly, distribution, and waste disposal, is 3.61 t CO2-eq (see
’ð yÞ  ’ð y  1Þ
Dð yÞ ¼ Q ð13Þ (4) Section 2.3). It should be noted that the waste disposal mentioned
1  ’ð13Þ
here refers to the disposal of industrial wastes as well as wastes
Substituting Eq. (2) into the right-hand side of Eq. (4) gives generated from the manufacture of the new car.
Dð yÞ ¼ C 0 ð’ð yÞ  ’ð y  1ÞÞ, which can be used to determine the The total number of new cars (gasoline passenger cars and
number of scrapped cars at time y. However, because the number hybrid cars) produced in 2005 in Japan was 4,748,391 (cars), with
of new cars purchased in Year 0, C0, is unknown, in this research, new hybrid cars occupying only 1.3% of the car market of Japan.
1810 S. Kagawa et al. / Global Environmental Change 23 (2013) 1807–1818

Accordingly, the average ‘‘passenger motor car’’ sector in the 2005 specific emissions at each life-cycle stage, the cumulative life-
Environmental Input–Output Table (401 sectors) can be regarded cycle emissions from the year that the eco-car subsidy was
as the ‘‘gasoline passenger car’’ sector. Therefore, the specific CO2 implemented (y = 13) to Year Y (Y > 13) can be estimated as
emissions attributed to the manufacture of one gasoline passenger shown in the equations below.
car, fa, was determined by multiplying this value (3.61) (t CO2-eq/ If the vehicle replacement scheme is not implemented and the
one million yen) by 1.78 million yen/vehicle, the average switch to new gasoline cars progresses in accordance with lifetime
producer’s price of a standard-size passenger car (2005 Input– distribution (case A):
Output Table published by the Ministry of Internal Affairs and
Communications of Japan). The CO2 emissions for the gasoline
passenger car are thus fa = 6426 kg CO2-eq/vehicle. X
Y X
Y
LA ðY Þ ¼ Dð yÞ f a þ 0:01  Dð yÞqs
A crucial problem is determining the CO2 emissions attributed y¼13 y¼13
to the manufacture of a single hybrid electric vehicle. Hybrid |fflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflffl} |fflfflfflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl}
electric vehicles have a battery, high capacity generator, electric Cumulative emission Cumulative emission
motor, and more complex power distribution electronics than associated with associated with
standard-size gasoline cars. Hybrid electric vehicles are also vehicle production scrapping end-of-life vehicles
typically constructed of lighter materials that the same ( )
X
Y XY
mechanical properties as those used in standard-size gasoline þ Dð yÞRð13Þ þ Q ð13Þ  Dð yÞ Rð0Þ (8)
cars, but these lighter materials require relatively more energy to y¼13 y¼13
|fflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflffl} |fflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl}
produce. Further, based on life cycle analysis results estimated by
Cumulative emission Cumulative emission
Toyota practitioners (Asakura, 2009; Ministry of Environment,
Japan, 2012a,b), the specific life-cycle CO2 emissions for a hybrid associated with associated with
car was u times that of a standard-size gasoline passenger car driving new vehicles driving vintage vehicles
(i.e., fh = u  fa), where u is the CO2 emission multiplier which
ranges between 1.0 and 1.5. A sensitivity analysis for which u ¼
If the vehicle replacement scheme is not implemented and the
1:0 ðlow caseÞ; 1:2 ðmedium caseÞ; 1:5 ðhigh caseÞ in a hy-
switch to new hybrid cars progresses in accordance with lifetime
brid car was conducted to estimate the influence of u on the
distribution (case B):
difference in life-cycle emissions between standard-size gasoline
passenger cars and hybrid cars.
If the vehicle replacement scheme was not implemented and X
Y X
Y
the 731,057 old cars aged 13 years or older were scrapped in LB ðY Þ ¼ Dð yÞ f h þ 0:01  Dð yÞqs
accordance with their lifetime distribution, then the number of y¼13
|fflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflffl}
y¼13
|fflfflfflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl}
old cars not scrapped and still being used after the eco-car Cumulative emission Cumulative emission
subsidy period (y  13) would be Q ð13Þ  Dð yÞ. The number of associated with associated with
old cars scrapped after the eco-car subsidy period (y  13) is Dð yÞ,
vehicle production scrapping end-of-life vehicles
and an equal number of new cars (standard-size gasoline
( )
passenger cars or hybrid cars) are purchased. Car owners who X
Y XY

relinquished their old cars were required to pay a disposal fee of þ Dð yÞRð13Þ þ Q ð13Þ  Dð yÞ Rð0Þ (9)
y¼13 y¼13
approximately 10,000 yen per vehicle. These fees are finally |fflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflffl} |fflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl}
channeled to waste disposal facilities to cover the costs Cumulative emission Cumulative emission
associated with scrapping the end-of-life vehicles (Ministry of associated with associated with
Economy, Trade and Industry of Japan, 2012). Accordingly, the driving new vehicles driving vintage vehicles
cost to dispose end-of-life vehicles can simply be calculated as
Dð yÞ  10; 000 yen (or Dð yÞ  0:01 million yen). Importantly, the
environmental input–output life-cycle assessment provides where Rð0Þ represents the CO2 emissions attributed to running one
information on the CO2 emissions associated with the production 13-year-old car that was registered as new in Year 0 and has been
of one unit (one million yen) in the waste disposal service driven for 13 years or more, and Rð13Þ represents the CO2
(sector) (Section 3.3 for the environmental input–output life- emissions attributed to running one new car that was registered as
cycle assessment method). The life-cycle emission intensity of new during the eco-car subsidy period. Since the fuel efficiency of
the waste disposal sector, qs, was estimated as 5.76 t CO2-eq per the new car is better than that of the old car, Rð13Þ < Rð0Þ. In
one million yen (Section 3.3). We therefore determined the CO2 addition, the specific emissions resulting from the manufacture of
emissions associated with scrapping the end-of-life vehicles by hybrid cars is higher than that of gasoline cars, and the emissions
multiplying the total disposal costs by the life-cycle emission associated with vehicle manufacture are u(=1.0, 1.2, 1.5) times
intensity for the waste disposal sector. Accordingly, the CO2 higher in case B than in case A (compare the first terms on the
emissions resulting from the disposal of passenger cars is right-hand sides of Eqs. (8) and (9)). We assumed that once car
58 kg CO2-eq/vehicle. Since CO2 emission associated with vehicle owners relinquish their old cars aged 13 years and older and buy
production is 6426 kg CO2-eq/vehicle, the CO2 emission associ- new cars, they continue using their new cars for an average car
ated with vehicle production and the disposal for passenger cars lifetime of 12 years (Kagawa et al., 2011).
is 6500 kg CO2-eq/vehicle in Japan. Using the ‘‘Greenhouse gases, Next, we shall consider the cases in which the vehicle
Regulated emissions, and Energy use in Transportation’’ models, replacement scheme was implemented. In these cases, a portion
Lenski et al. (2010) estimated that CO2 emissions associated with (Q ð13Þ) of old cars are simultaneously scrapped during the period
vehicle production and the disposal of passenger cars was of the vehicle replacement scheme and the same number of new
7800 kg CO2-eq/vehicle in the US. cars are purchased. The cumulative life-cycle CO2 emissions in
In this study, the life-cycle CO2 emissions attributed to cars such cases can be calculated as shown below.
after the eco-car subsidy period can be estimated by summing If the vehicle replacement scheme is implemented and the
the total emissions from running old cars, new cars, from the transition to new gasoline passenger cars occurs at the same time
manufacture of new cars, and from scrapping old cars. Using the as all of the old cars are scrapped (case C), then
 S. Kagawa et al. / Global Environmental Change 23 (2013) 1807–1818 1811

LC ðY Þ ¼ Q ð13Þ f a þ 0:01  Q ð13Þqs producer’s prices of a passenger car and gasoline are 1,778,551
|fflfflfflfflfflffl{zfflfflfflfflfflffl} |fflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflffl}
Emission Emission (yen/vehicle) and 115 (yen/l), respectively. Accordingly, multiply-
ing the life-cycle CO2 emissions (2005 Environmental Input–
associated with associated with
Output Table provided by the National Institute for Environmental
vehicle production scrapping end-of-life vehicles Studies of Japan) by the producer’s prices yields life-cycle CO2
X
Y
emissions of a passenger car and well-to-tank CO2 emissions of
þ Q ð13ÞRð13Þ (10)
gasoline as fa = 6.4258(t CO2-eq/vehicle)6426(kg CO2-eq/vehi-
y¼13
|fflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflffl} cle) and fc = 0.00063(t CO2-eq/l)0.6(kg CO2-eq/l), respectively.
Cumulative emission In particular, using the environmental input–output life-cycle
associated with assessment method, the CO2 emissions generated by upstream
driving new vehicles sectors in the automobile supply chains can easily be estimated as

If the vehicle replacement scheme is implemented and the qa ¼ diagðeÞðI  AÞ1 ya (13)
transition to new hybrid cars occurs as all of the old cars are
where ya is the final demand vector, for which the only element
scrapped (case D), then
associated with passenger car sector is one, while all other
elements are zero, diagðeÞ is the diagonal matrix in with the
LD ðY Þ ¼ Q ð13Þ f h þ 0:01  Q ð13Þqs
|fflfflfflfflfflffl{zfflfflfflfflfflffl} |fflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflffl} elements of e along the diagonal. The ith vector element of qa
Emission Emission represents the CO2 emissions associated with the production of the
associated with associated with commodity upstream from sector i that is required for one unit of
vehicle production scrapping end-of-life vehicles passenger car production. A complete solution for the life cycle CO2
emissions in the automobile supply chains (i.e., qa) is presented in
X
Y
þ Q ð13ÞRð13Þ (11) Table S1 of the supplementary information. The system boundary
y¼13 of our analysis is presented in Fig. S1.
|fflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflffl}
Cumulative emission 3.4. Adding the direct rebound effect into the environmental input–
associated with output life-cycle assessment framework
driving new vehicles
Based on energy rebound effect studies (Greene, 1992; Jones,
Comparing Eqs. (8) and (10), in Eq. (10) when consumers scrap 1993; Goldberg, 1998; Greene et al., 1999; Greening et al., 2000; Small
and simultaneously replace their old cars during the vehicle and Van Dender, 2007), the direct rebound effect can be defined as
replacement scheme, the emissions from running old cars aged 13
years and older (the part corresponding to the third term on the
Dd=d Dd=d
h¼ ¼   (14)
right-hand side of Eq. (8)) are zero.
D ps = ps D e1  pg = e1  pg

3.3. Environmental input–output life-cycle assessment method, data where d is the average annual travel distance, e1 is liters of gasoline
and supplementary results per kilometer (i.e., reciprocal of fuel mileage), pg is the gasoline price,
and ps = e1  pg represents the gasoline price per kilometer (i.e.,
Environmental input–output life-cycle assessment methods have energy service price). Eq. (14) denotes the gasoline cost-per-
thus far been used to estimate economy-wide emissions directly and kilometer elasticity of driving. As in Small and Van Dender (2007), it
indirectly induced by the production of one unit of a specific product can be understood that as the fuel mileage increases and the energy
(Lave et al., 1995; Hendrickson et al., 1998 for seminal contributions service price conversely decreases, car owners may want to drive
and Murray et al., 2012 for the practical importance of the further. Consequently, this implies that the gasoline cost-per-
environmental input–output life-cycle assessment methods). The kilometer elasticity of driving would be negative. Although other
essence of environmental input–output life-cycle assessment additional maintenance costs in other vehicles, especially hybrid
methods is to connect a highly-disaggregated input–output matrix cars (e.g., battery costs), may play a role in increasing the energy
with an emission intensity vector. More concretely, the CO2 service price, we assumed that the maintenance costs are negligible
emissions directly and indirectly induced by producing one unit in this study. In addition, the change in fuel mileage after the shift to
(one million yen) of a specific product k, can be estimated as follows: new cars may affect the cost-per-kilometer elasticity of driving and
the resultant change in distance may affect the car lifetime
qk ¼ eðI  AÞ1 yk (12) distribution. This study did not consider this complicated factor.
  It should be noted that eð0Þ represents the fuel mileage of one 13-
where, A ¼ ai j ði; j ¼ 1; . . . ; nÞ is a direct requirements matrix
year-old car that was registered as new in Year 0 and has been driven
whose (i, j) element represents the input of commodity i that is
for 13 years or more, and eð13Þ ( > eð0Þ) represents the fuel mileage of
directly required to produce one unit of output of commodity j, I is
one new car that was registered as new during the passenger vehicle
the identity matrix of order n, yk is the final demand vector for which
replacement period. If consumers exchange their old cars with a fuel
the only element associated
 with commodity k is one, while all other mileage of eð0Þ for new cars with the fuel
n mileage of eð13oÞ, the energy
elements are zero, e ¼ e j is the CO2 emission intensity vector
service price will decrease by D p̃s ¼ eð13Þ1  eð0Þ1  pg . Since
representing CO2 emissions per unit production of commodity j.
the energy service price before switching to new cars is
Using the 2005 Environmental Input–Output Table and CO2
p̃s ¼ eð0Þ1  pg , we obtain the following additional travel distance
emissions data provided by the National Institute for Environ-
induced by the decrease in the energy service price:
mental Studies of Japan), we estimated life-cycle CO2 emissions
directly and indirectly induced by producing one unit (one million D p̃s
Dd ¼ h d (15)
yen in producer price) of passenger cars, gasoline, and waste p̃s
treatment services as qa = 3.61 (t CO2-eq/one million yen), Consequently, the annual travel distance after switching to new
qg = 5.51 (t CO2-eq/one million yen), and qs = 5.76 (t CO2-eq/one cars can be represented as follows:
 
million yen), respectively. According to the producer’s price table D p̃s
d̃ðhÞ ¼ d þ Dd ¼ 1 þ h d (16)
in the 2005 Environmental Input–Output Table, the average p̃s
1812 S. Kagawa et al. / Global Environmental Change 23 (2013) 1807–1818

Considering the direct rebound effect in Eq. (16), the CO2 estimated as
emissions derived from gasoline consumption generated during
new car running can be estimated as shown below using specific 9224ð1  eÞy1
Rg ðs; yÞ ¼ Eðs; yÞ  f g ¼  fg (23)
emissions, fg, which represents CO2 emissions (t CO2-eq) directly eðsÞ
produced by the combustion of 1 l of gasoline. where fg represents CO2 emissions (t CO2-eq) that are directly
produced by the combustion of 1 l of gasoline (see Section 3.3).
d̃ðhÞ
R̃g ð13Þ ¼  f (17) Also, if life-cycle CO2 emissions generated during gasoline
eð13Þ g
refining are taken as fc (t CO2-eq/l) (see Section 3.3), then the
If life-cycle CO2 emissions generated during the gasoline CO2 emissions from gasoline refining can be calculated as
refining process are taken as fc (t CO2-eq/l) (see Section 3.3), then follows:
CO2 emissions from gasoline refining can be calculated as follows:
9224ð1  eÞy1
Rc ðs; yÞ ¼ Eðs; yÞ  f c ¼  fc (24)
d̃ðhÞ eðsÞ
R̃c ð13Þ ¼  f (18)
eð13Þ c
Therefore, the annual CO2 emissions associated with operating
Therefore, the annual CO2 emissions from running one new car one old standard-size passenger car manufactured in Year s can be
during the passenger vehicle replacement period can be obtained obtained as follows:
as follows:
Rðs; yÞ ¼ Rc ðs; yÞ þ Rg ðs; yÞ
R̃ð13Þ ¼ R̃c ð13Þ þ R̃g ð13Þ (19)
9224ð1  eÞy1 9224ð1  eÞy1
¼  fg þ  fc (25)
Finally, substituting Eq. (19) into the right-hand side of Eqs. (8)– eðsÞ eðsÞ
(11), we obtain the life-cycle emission accounting models with
It should be noted that if 0 < e < 1, then the annual CO2
direct rebound effects. Here, it should be noted that if h in Eq. (16)
emissions associated with running a vehicle will decrease with
is equal to zero, and d̃ð0Þ coincides with d, then life-cycle emission
vehicle age y. From Eq. (29), Rð0; yÞ represents the CO2 emissions
accounting models with direct rebound effects will be exactly
after y years (y  1) attributed to running one 13-year-old car that
same as Eqs. (8)–(11).
was registered as new in Year 0. Rð13; yÞ represents the CO2
In an econometric analysis for the five years 1997–2001, Small
emissions after y years attributed to running one new car that was
and Van Dender (2007) estimated the direct gasoline rebound
registered as new during the eco-car subsidy period. If the
effect in the USA to be h = 0.11. In this study, a sensitivity analysis
decreasing rate of annual driving distance, e, is zero, then Rð0; yÞ
in the range h = 0 to h = 0.4 was conducted.
and Rð13; yÞ will coincide with Rð0Þ and Rð13Þ, respectively (see
Eqs. (5)–(7)).
3.5. Considering the effect of decreasing annual driving distance with
Substituting Rð0; yÞ and Rð13; yÞ obtained from Eq. (25) into the
vehicle age into the environmental input–output life-cycle assessment
right-hand side of Eqs. (8)–(11), gives us the following life-cycle
framework
emission accounting models that consider the effect of decreasing
annual driving distance with vehicle age.
Although there is no empirical evidence in Japan for annual
driving distance decreasing with vehicle age, such an assumption
may be plausible (Lenski et al., 2010). In the benchmark analysis of X
Y X
Y
this study, we assume that all vehicles are driven the same annual LA ðY Þ ¼ Dð yÞ f a þ 0:01  Dð yÞqs
y¼13 y¼13
average travel distance, d = 9224(km), under all scenarios and that |fflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflffl} |fflfflfflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl}
the annual average travel distance does not change with vehicle age. Cumulative emission Cumulative emission
Here, without loss of generality, as in Lenski et al. (2010), we assume associated with associated with
that the annual average travel distance decreases by e  100% per
vehicle production scrapping end-of-life vehicles
year; the annual average travel distance is defined as an average ( )
X
Y XY
value of cross-sectional travel distances of passenger cars. þ Dð yÞRð13; yÞ þ Q ð13Þ  Dð yÞ Rð0; yÞ (25)
Based on this assumption, we can use the following equation to y¼13 y¼13
determine the annual driving distance of a vehicle with age y. |fflfflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflfflffl} |fflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl}
Cumulative emission Cumulative emission
dð yÞ ¼ ð1  eÞdð y  1Þ ð y  2Þ (20) associated with associated with
where dð yÞ is the annual driving distance of a vehicle with age y driving new vehicles driving vintage vehicles
and dð y  1Þ is the annual driving distance of a vehicle with age y-1.
If the initial annual driving distance in the first year (i.e., y = 1) is
denoted as dð1Þ ¼ 9224, then the annual driving distance after Y X
Y X
Y
years can be obtained using the following equation: LB ðY Þ ¼ Dð yÞ f h þ 0:01  Dð yÞqs
y¼13 y¼13
Y1 Y1 |fflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflffl} |fflfflfflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl}
dðY Þ ¼ dð1Þ  ð1  eÞ ¼ 9224ð1  eÞ (21)
Cumulative emission Cumulative emission
Under this situation, the annual gasoline consumption of one associated with associated with
standard-size passenger car manufactured in Year s will decrease vehicle production scrapping end-of-life vehicles
with vehicle age y as follows: ( )
X
Y XY

y1 þ Dð yÞRð13; yÞ þ Q ð13Þ  Dð yÞ Rð0; yÞ (26)

dð yÞ 9224ð1  eÞ y¼13 y¼13
Eðs; yÞ ¼ ¼ (22) |fflfflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflfflffl} |fflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl}
eðsÞ eðsÞ
Cumulative emission Cumulative emission
where eðsÞ is the average fuel consumption (km/l) of a car associated with associated with
manufactured in Year s. Therefore, the CO2 emissions derived from
driving new vehicles driving vintage vehicles
gasoline consumption generated during car operation can be
 S. Kagawa et al. / Global Environmental Change 23 (2013) 1807–1818 1813

LC ðY Þ ¼ Q ð13Þ f a þ 0:01  Q ð13Þqs from 1.0 to 1.5, the intersection point would range from 2.4 years
|fflfflfflfflfflffl{zfflfflfflfflfflffl} |fflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflffl}
Emission Emission to 4.0 years.
associated with associated with The curves for cases B and D that are shown in Fig. 3(c) intersect at
vehicle production scrapping end-of-life vehicles approximately 3.3 years when u = 1.2 (medium case), after which
the scheme using hybrid cars generates carbon savings that exceed
X
Y
þ Q ð13ÞRð13; yÞ (27) those of the baseline case without the scheme. In other words,
y¼13 compared to standard gasoline passenger cars (Fig. 3(a)), carbon
|fflfflfflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl}
savings are realized sooner when consumers switch to hybrid cars
Cumulative emission
(Fig. 3(c)). This implies that the overall emissions associated with
associated with driving are decreased. In summary, we found that the scheme would
driving new vehicles mitigate greenhouse gas emissions if the new cars remained in use
for at least 3.3 years in the case of hybrid cars, or for at least 4.7 years
LD ðY Þ ¼ Q ð13Þ f h þ 0:01  Q ð13Þqs in the case of standard gasoline cars. Fig. 3(c) also shows that if the
|fflfflfflfflfflffl{zfflfflfflfflfflffl} |fflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflffl}
CO2 multiplier for hybrid cars is changed from 1.0 to 1.5, then the
Emission Emission
intersection point, Y0, would range between 3.0 and 3.7.
associated with associated with We then calculated the cumulative CO2 reductions that could be
vehicle production scrapping end-of-life vehicles achieved over a car’s average lifetime by exchanging an old car
X
Y with a new car. Fig. 4 shows the cumulative life-cycle CO2
þ Q ð13ÞRð13; yÞ (28) emissions obtained after switching to standard-size passenger cars
y¼13
|fflfflfflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl} or hybrid cars under the scheme. When switching to standard-size,
Cumulative emission gasoline passenger cars, the difference in cumulative CO2
associated with emissions between cases A and C after the remaining car lifetime
driving new vehicles of 12 years is approximately one million tons CO2-eq (see the scope
of our analysis in Fig. 1). Such a reduction in CO2 is equivalent to
As in Lenski et al. (2010), a sensitivity analysis ranging from 0.09% of the CO2 emissions produced in Japan in 2009 (1.145 billion
e = 0 to e = 0.04 was conducted in this study. tons CO2-eq) (National Institute for Environmental Studies of
Japan, 2012). However, when switching to hybrid cars, the
3.6. Data sources and parameter settings difference in cumulative CO2 emissions between natural replace-
ment (case A) and replacement scenarios (case D with u = 1.2) after
Table 1 shows the data sources and parameter settings used in this 12 years is more than eightfold, i.e., approximately 8.5 million tons
study. Following Kagawa et al. (2011), the three parameters in Eq. (1), CO2-eq, which equivalent to 0.74% of the CO2 emissions in Japan in
a, b, and r, are set as 15.41, 4.96, and 0.55, respectively (Table 1). The 2009 (National Institute for Environmental Studies of Japan, 2012).
other variables necessary for the study are presented in Table 1. These findings imply that the scheme would mitigate greenhouse
gas emissions slightly in the medium term.
4. Life-cycle CO2 emissions associated with the scrappage It is important to note that the above benchmark analysis is based
scheme on the assumption that the target age of all of the old cars that are
scrapped as part of the vehicle replacement scheme is 13 years. We
As shown in Fig. 2, introduction of the scrappage scheme (case found that as the target car age decreases, the rate of scrapping old
C) increases the amount of life-cycle CO2 emissions in the first year cars decreases, and vice versa (Fig. S2). The benefit of slowly replacing
by about 3.5 million tons CO2-eq compared to the baseline case old cars with new cars is that CO2 emissions resulting from the
(case A) in which no such scheme was implemented (see Table S1 manufacture of new cars can be reduced; however, slow replacement
for details on the sectoral aggregation). Although the reduction in leads to an increase in CO2 emissions from driving cars, due to the
gasoline consumption achieved by switching from old cars to new, difference in the average fuel consumption of newer cars and older
more fuel-efficient, gasoline passenger cars led to the reduction in cars. As a result, one might expect for there to be an optimal target car
CO2 emissions (compare CO2 emissions for the driving phase in age in the sense that CO2 emissions avoided through the scrappage
cases A and C, large amounts of CO2 are emitted through the scheme are maximized during the remaining car lifetime of 12 years
premature scrapping of older cars and the subsequent production after switching to standard-size passenger cars or hybrid cars under
of the new cars to replace them. the scheme. Fig. 5 shows how changing the target car age for the
Fig. 3(a) shows a comparison between the cumulative life-cycle scrappage scheme affects cumulative CO2 emissions during the
CO2 emissions for cases A and C. The results show that the curves for remaining car lifetime of 12 years. A comparison between natural
cases A and C intersect at a point, Y0, approximately 4.7 years after replacement (case A) and a replacement scenario (case D with u = 1.2)
implementation of the scheme. After 4.7 years, the emissions in case A after 12 years shows that if the target car age for the scrappage
are higher than in case C, implying that the scheme only leads to scheme is 9 years, then CO2 emissions avoided by switching from old
reductions in emissions when the new cars are used for at least 4.7 cars to new petrol cars would be maximized, and would amount to
years. 8.8 million tons CO2-eq (Fig. 5). This also shows that if the Japanese
Fig. 3(b) shows a comparison between the cumulative life- government set the target car age as 9 years, an additional reduction
cycle CO2 emissions for cases A and D. In this comparison, the CO2 of 0.3 million tons CO2-eq would have been brought about in a
emission multiplier for hybrid cars, u, affects the intersection comparison with that achieved by the actual scrappage scheme
point, Y0. As the CO2 emission multiplier increases, the CO2 (target car age = 13 years) (Fig. S2).
attributed to the manufacture of hybrid cars also increases. A
higher multiplier consequently leads to higher cumulative life- 5. Impacts of changes in annual driving distance on CO2
cycle CO2 emissions and pushes the cumulative emission line emissions avoided through implementing the scrappage
upward as the scheme is implemented (red line in Fig. 3(b)). The scheme
results show that the scheme only leads to reductions in
emissions when the new hybrid cars are used for at least 3.6 The results in Fig. 4 show that switching from gasoline
years when u = 1.2 (medium case). If the multiplier is changed passenger cars aged 13 years or more, to new hybrid cars through
1814 S. Kagawa et al. / Global Environmental Change 23 (2013) 1807–1818

Table 1
Parameter settings and data sources.

Equation Variable and parameter settings Data source

Eq. (1) a = 15.41 Kagawa et al. (2011)

Eq. (1) b = 4.96 Kagawa et al. (2011)
Eq. (1) r = 0.55 Kagawa et al. (2011)
Eq. (4) Q ð13Þ ¼ 731; 057 Report on the Japanese vehicle replacement
scheme, Next Generation Vehicle Promotion Center, Japan
Eq. (5) d = 9224 (km) Annual Report of Road Transport Statistics, Ministry of Land, Infrastructure, Transport and
Tourism, Japan
Eq. (5) Vintage gasoline vehicle aged 1 year: 15.8 (km/l) The Energy Data and Modeling Center handbook of energy & economic statistics in Japan
Vintage gasoline vehicle aged 2 years: 14.4 (km/l)
Vintage gasoline vehicle aged 3 years: 13.9 (km/l)
Vintage gasoline vehicle aged 4 years: 14.1 (km/l)
Vintage gasoline vehicle aged 5 years: 13.9 (km/l)
Vintage gasoline vehicle aged 6 years: 13.9 (km/l)
Vintage gasoline vehicle aged 7 years: 13.9 (km/l)
Vintage gasoline vehicle aged 8 years: 13.3 (km/l)
Vintage gasoline vehicle aged 9 years: 12.6 (km/l)
Vintage gasoline vehicle aged 10 years: 12.3 (km/l)
Vintage gasoline vehicle aged 11 years: 12.1 (km/l)
Vintage gasoline vehicle aged 12 years: 11.9 (km/l)
Vintage gasoline vehicle aged 13 years: 11.8 (km/l)
Vintage gasoline vehicle aged 14 years: 11.6 (km/l)
Vintage gasoline vehicle aged 15 years: 11.8 (km/l)
Vintage gasoline vehicle aged 16 years: 11.7 (km/l)
Vintage gasoline vehicle aged 17 years: 11.6 (km/l)
Vintage gasoline vehicle aged 18 years: 12.1 (km/l)
Vintage gasoline vehicle aged 19 years: 12.4 (km/l)
Vintage gasoline vehicle aged 20 years: 12.7 (km/l)
Eq. (5) New gasoline vehicle: 15.8 (km/l) The Energy Data and Modeling Center handbook of energy & economic statistics in Japan
New hybrid vehicle: 35.5 (km/l)
Eq. (5) fg = 0.00231 (t CO2–eq/l) National Institute for Environmental Studies, Japan: Embodied Energy and Emission
Intensity Data for Japan Using 2005 Environmentally-extended Input–Output Table
Eq. (6) fc = 0.00063 (t CO2-eq/l) National Institute for Environmental Studies, Japan: Embodied Energy and Emission
Intensity Data for Japan Using 2005 Input–Output Table
Eq. (8) New gasoline vehicle: fa = 6.426 (t CO2–eq/car) National Institute for Environmental Studies, Japan: Embodied Energy and Emission
Intensity Data for Japan Using 2005 Input–Output Table

the scrappage scheme helps to reduce greenhouse gas emissions in increases CO2 emissions from running hybrid cars and decreases
the medium term. However, the analysis in the preceding section the overall emissions-reduction effect of the scrappage scheme.
does not consider the possibility that, as the cost of gasoline Therefore, this study estimated the impact of the rebound effect
(gasoline service price) required to travel one kilometer decreases on the scrappage scheme’s emissions reduction by incorporating
with improvements in fuel efficiency, overall mileage may increase. the energy rebound effect into the environmental input–output
Considering this energy rebound effect, the annual mileage of hybrid life-cycle assessment framework shown in Eqs. (8)–(11). According
car owners may thus be greater than when they owned old, more to recent studies, the elasticity of mileage h is approximately 0.1
fuel-inefficient gasoline cars. This rebound effect eventually for the US (Small and Van Dender, 2007). Therefore, in this study,

7,000

Scrapping end-of-life vehicles

6,008 Driving
6,000 42
Other sector
Life-cycle CO2 emissions (1000 tons CO2-eq)

1,256 Freight transport

5,000 Electricity
Assembly
Parts
4,000 Motor vehicle bodies
Internal combuson engines for motor vehicles
Materials
3,000
2,547
8
4,710
2,000
1,596

1,000

943
0
Case A Case C

Fig. 2. Comparison between life-cycle CO2 for case A (without the scrappage scheme) and case C (with the scrappage scheme for switching to a new standard-size gasoline
passenger car).
 S. Kagawa et al. / Global Environmental Change 23 (2013) 1807–1818 1815

emission intensity from the manufacture of hybrid cars is

improved by 20%, reducing the emission intensity to the same
level as gasoline cars, the rebound effect associated with switching
from a gasoline car to a hybrid car cancels out the emissions
reduction.
Interestingly, when the value of rebound elasticity h was 0.1,
the emissions reduction effect of the scrappage scheme was 7.7
million tons CO2-eq. However, assuming that the average annual
mileage of an old car aged 13 years or more and the average annual
mileage of a hybrid car that has been newly purchased through the
scheme decrease at an annual rate of 4% (in other words, e = 0.04),
as employed by Lenski et al. (2010), then the emissions reduction
effect of the scrappage scheme decreases by 1.7 million tons CO2-
eq (Fig. 6(b)). Thus, the results of this estimation show that,
ultimately, the environmental benefits associated with the
scrappage scheme is seriously undermined as a result of average
annual mileage decreasing with car age. The main reason is that a
decrease in mileage reduces the environmental benefits of the
scheme through the difference in the average fuel consumption of
newer cars and older cars.

6. Discussion and policy implications

In Japan, a lump-sum subsidy of 250,000 yen, which was worth

approximately 3188 US dollars on August 6, 2012, was paid to any
consumer buying a standard-size passenger car that met emissions
and fuel economy standards after scrapping their old car (provided
that their old car was newly registered at least 13 years earlier).
Importantly, such a policy does not incentivize consumers to
continue using their old cars and, in so doing, extend the lifetime of
their vehicles. For both of the schemes encouraging consumers to
switch to new gasoline or hybrid cars, i.e., cases C and D, we found
that greenhouse gas emissions would only decrease if the
participants in the scheme retained their new cars for at least
4.7 years after the replacement purchase (Fig. 3(a) and (b)).
Figs. 3 and 4 show that vehicle replacement schemes
contribute to a moderate reduction in emissions. The primary
reason for this moderate reduction is due to the considerable life-
cycle emissions associated with the manufacture of new cars. Thus,
in order to improve the efficacy of vehicle scrappage schemes,
reductions in life-cycle emissions throughout the car manufacture
supply chain are required in order to realize the benefits of
efficiency improvements. It should be noted that when u = 1, the
CO2 emission intensity of a gasoline passenger car is equivalent to
Fig. 3. (a) Comparison between cumulative lifecycle CO2 emissions attributed to that of a hybrid car. The comparison between case D (u = 1.0) and
cars under case C (scrappage scheme used to switch to new standard-size gasoline case D (u = 1.2) in Fig. 4 shows that the 20% reduction in the CO2
passenger car) and case A (no scrappage scheme) scenarios. (b) Comparison
between cumulative lifecycle CO2 emissions from cars under case D (scrappage
intensity of a hybrid car will bring about an additional emission
scheme used to switch to hybrid car) and case A (no scrappage scheme) scenarios. reduction of approximately one million tons CO2-eq over 12 years
(c) Comparison between cumulative lifecycle CO2 emissions from cars under case D (Fig. 4). In this study, it was found that the rebound effect by the
(scrappage scheme used to switch to hybrid car) and case B (no scrappage scheme) driver was sufficiently large to completely cancel out any
scenarios. (For interpretation of the references to color in this figure legend, the
additional emission reduction resulting from technical improve-
reader is referred to the web version of this article.)
ments by the car manufacturer (Fig. 6). It is probably unrealistic to
adopt a policy that aims to limit long-distance driving among
the value of h was varied between 0, 0.1, 0.2, 0.3, 0.4, and the drivers who have purchased a hybrid car because they want to
impact of this variation in h on the scrappage scheme’s emissions drive further and more cheaply than when they had a gasoline car.
reduction effect was estimated (Fig. 6(a)). It is thus important to educate consumers who purchase hybrid
It was found that when there is absolutely no rebound effect (in cars about how the rebound effect greatly reduces the emissions
other words, when h = 0), the scrappage scheme has a reduction reduction effect of the scrappage scheme.
effect of 8.5 million tons CO2-eq (see case when u = 1.2 in Fig. 6). We also found that the CO2 reduction effect achieved through
Importantly, the emissions reduction effect of the scrappage the introduction of the scheme differs by a factor of 8.5 depending
scheme decreases by 800,000 t when the value for rebound on whether consumers purchase a standard-size gasoline passen-
elasticity h is 0.1 (Fig. 6(a)). Meanwhile, if the emission intensity ger car or a hybrid vehicle. Current schemes are too broad and
associated with the manufacture of hybrid cars is 20% higher than allow consumers to purchase any car, even new cars with very poor
that of gasoline cars, then the emissions reduction effect of the fuel efficiency. Consequently, these schemes are not taking full
scrappage scheme decreases by approximately 900,000 t or 11% advantage of the fuel and carbon savings that are potentially
(compare the blue and red lines in Fig. 6(a)). Therefore, even if possible.
1816 S. Kagawa et al. / Global Environmental Change 23 (2013) 1807–1818

25

Scrapping end-of-life vehicles

Cumulave life-cycle CO2 emissions (million tons CO2-eq)

20.8 Driving

20 19.8 Other sector

Freight transport
Electricity
15.1 Assembly
15
Parts
12.4 Motor vehicle bodies
Internal combuson engines for motor vehicles
10 16.1 Materials
15.1
9.4
6.7

5
CO2 emission in the case of 1.5
4.7 5.7 5.7
4.7

0 CO 2 emission in the case of 1.0

Case A Case C Case B ( =1.2) Case D ( =1.2)
: CO 2 emission mulplier on the difference in life-cycle CO2 emission intensies
associated with the manufactures of petro cars and hybrid cars

Fig. 4. Comparison between cumulative lifecycle CO2 emissions generated during the remaining car lifetime of 12 years after switching to standard-size passenger cars or
hybrid cars under the scheme.

Although the vehicle scrappage scheme contributed positively yield monetary benefits amounting to 1.73 billion yen (22 million
toward reducing CO2 emissions, the policy also needs to pass the US dollars) or 78 million euros (97 million US dollars), respectively.
cost–benefit test. The total cost of this scheme (based on a lump This cost–benefit ratio of 0.001–0.04 is extremely small. In other
sum payment of 250,000 yen/car and a total of 731,057 cars) was words, this implies that the scheme was a relatively costly policy
approximately 183 billion yen (approx. 2.3 billion US dollars on for mitigating greenhouse gas emissions.
August 6, 2012). If the energy rebound elasticity and the decrease We conclude that the structure and implementation of the
in annual driving distance were 0.1 (Small and Van Dender, 2007) policy needs to be reassessed. Specifically, we propose that
and 4% (Lenski et al., 2010), respectively, the reduction in incentives for buying more fuel-efficient models, such as hybrid
cumulative CO2 emissions achieved by the scrappage scheme cars, needs to be encouraged in order to achieve both economic
when consumers switched to hybrid cars would be approximately stimulus and environmental benefits. The government should
6 million tons CO2-eq (Fig. 4(b)). Using the currently planned CO2 periodically review relevant policies and adjust these depending
tax rate of 289 yen/t CO2-eq (Ministry of Environment, Japan, on current levels of technical development. In addition, technical
2012a,b), or the European carbon permit price of 13 euros/t CO2-eq progress could also be a focus of government policy, i.e., providing
(Committee on Climate Change (CCC), UK, 2009), the policy would incentives to car producers to implement new technologies in

10000

8807
Cumulave life-cycle CO 2 emissions avoided through the

8000
scrappage scheme (thousand tons CO 2-eq)

6000
5882 Case A vs. Case C
Case A vs. Case D ( =1.2)
4000
Case B ( =1.2) vs. Case D ( =1.2)

2000
1385

0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

-2000

- 4000
: CO2 emission mulplier on the difference in lifecycle CO 2 emission intensies
associated with the manufactures of petro cars and hybrid cars

Fig. 5. Relationship between target car age of scrapped cars and cumulative lifecycle CO2 emissions generated during the remaining car lifetime of 12 years after switching to
standard-size passenger cars or hybrid cars under the scheme. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of
this article.)
 S. Kagawa et al. / Global Environmental Change 23 (2013) 1807–1818 1817

10.0
9.4 (a)
9.0

CO2 reudcon effects of the passenger vehicle

replacement program (million tons CO2-eq)
8.5 8.6 =1.0
8.0 =1.2
7.7 7.8

7.0 7.1 7.0 =1.5

6.9
6.3 6.2
6.0 6.1
5.5
5.3
5.0
4.6
4.0 3.8

3.0

2.0

1.0

0.0
0 0.1 0.2 0.3 0.4 0.5
Rebound elascity -

10.0
(b)
CO2 reudcon effects of the passenger vehicle

9.0
replacement program (million tons CO2-eq)

8.5
8.0 7.9
7.7
7.2 7.4
7.0 6.9 6.9
6.7
6.5 6.3 6.4
6.0 6.1 6.1 6.0
5.8 5.8
5.5 5.5
5.3 5.1 5.3
5.0 4.9 5.0
4.7 4.6
4.0
=0
3.0
=-0.1
2.0 =-0.2
=-0.3
1.0
=-0.4
0.0
0 0.01 0.02 0.03 0.04

Decreasing rate of annual driving distance

Fig. 6. CO2 reduction effects of decreasing annual driving distance with vehicle age. (For interpretation of the references to color in this figure legend, the reader is referred to
the web version of this article.)

vehicle design, fuel efficiency, and the production of smaller cars. (K122024) from the Japanese Ministry of Environment. The
Since the fuel efficiency of the average new gasoline car in Japan is authors accept full responsibility for the views expressed in this
considerably better than that of the average car in the USA, and paper. We thank three anonymous referees for their helpful
even better than some of the hybrid vehicles in the USA, this study comments on this manuscript.
could be used to justify the creation of subsidies for hybrids in the
USA and other countries. Appendix A. Supplementary data
Furthermore, we conclude that the government failed to
determine an optimal target car age for the scrappage scheme Supplementary data associated with this article can be found, in
in the sense that the environmental benefits of the scheme might the online version, at doi:10.1016/j.gloenvcha.2013.07.023.
not be maximized during the remainder of the cars’ lifetimes
(Fig. 5). In the event that such a scheme were reintroduced in the
future, more careful policy decisions regarding the age of cars References
earmarked for scrapping need to be taken.
Abrams, B.A., Parsons, G.R., 2009. Is CARS a clunker? The Economists’ Voice 6 ,
http://dx.doi.org/10.2202/1553-3832.1638.
Acknowledgments
Argonne National Laboratory, 2009. The Greenhouse Gases, Regulated Emissions,
and Energy Use in Transportation Model. hhttp://greet.es.anl.gov/i.
This research was supported by a Grant-in-Aid for research (No. Asakura, Y., 2009. Hybrid vehicles. MotorRing 28, 1–3.
Cohen, C., Whitten, B.J., 1988. Parameter Estimation in Reliability and Life Span
22360386) from the Ministry of Education, Culture, Sports, Science
Models. Marcel Dekker, Inc, New York, NY.
and Technology in Japan and partially supported by the Cohen, C., 1991. Truncated and Censored Samples: Theory and Applications. Marcel
Environment Research and Technology Development Fund Dekker, Inc, New York, NY.
1818 S. Kagawa et al. / Global Environmental Change 23 (2013) 1807–1818

Committee on Climate Change (CCC), UK, 2009. Meeting Carbon Budgets—The Need Lenzen, M., Dey, C.J., 2002. Economic, energy and greenhouse emissions impacts of
for a Step Change. hhttp://hmccc.s3.amazonaws.com/21667%20CCC%20Report some consumer choice, technology and government outlay options. Energy
%20AW%20WEB.pdfi. Economics 24, 377–403.
European Automobile Manufacturers’ Association, 2012. Vehicle Scrapping MacLean, H., Lave, L.B., 2003a. Life cycle assessment of automobile/fuel options.
Schemes in the European Union. hhttp://www.acea.be/images/uploads/files/ Environmental Science and Technology 37, 5445–5452.
20090406_Scrapping_schemes.pdfi. MacLean, H., Lave, L.B., 2003b. Evaluating automobile fuel/propulsion system
Executive Office of the President of the United States, 2009. Economic Analysis of technologies. Progress in Energy and Combustion Science 29, 1–69.
the Car Allowance Rebate System. hhttp://www.whitehouse.gov/assets/docu- Ministry of Land, Infrastructure, Transport and Tourism, Japan, 2012. hhttp://
ments/CEA_Cash_for_Clunkers_Report_FINAL.pdfi. www.mlit.go.jp/jidosha/jidosha_fr1_000008.htmli.
Greene, D.L., 1992. Vehicle use and fuel economy; how big is the rebound effect? Ministry of Environment, Japan, 2012a. hhttp://www.env.go.jp/council/02policy/
Energy Journal 13, 117–143. y025-01/ref_02-2.pdfi.
Greene, D.L., Kahn, J.R., Gibson, R.C., 1999. Fuel economy rebound effect for US Ministry of Environment, Japan, 2012b. hhttp://www.env.go.jp/policy/tax/plans/
households. Energy Journal 20, 1–31. 2011/about.pdfi.
Greening, L.A., Greene, D.L., Difiglio, C., 2000. Energy efficiency and consumption – Ministry of Economy, Trade and Industry, Japan, 2012. hhttp://www.meti.go.jp/
the rebound effect – a survey. Energy Policy 28, 389–401. policy/recycle/main/admin_info/law/08/i.
Goldberg, P.K., 1998. The effects of the corporate average fuel efficiency standards. Moriguchi, Y., Kondo, Y., Shimizu, H., 1993. Analyzing the life cycle impact of cars:
Journal of Industrial Economics 46, 1–33. the case of CO2. Industry and Environment 16, 42–45.
Hendrickson, C., Horvath, A., Joshi, S., Lave, L.B., 1998. Economic input–output Murray, J., Cawthorne, G., Dey, C., Andre, C. (Eds.), 2012. Enough for All Forever: A
models for environmental life-cycle assessment. Environmental Science and Handbook for Learning about Sustainability. Common Ground Publishing,
Technology 32, 184A–191A. Champaign, IL.
Hertwich, E.G., 2011. The life cycle environmental impacts of consumption. Eco- Müller, D.B., 2006. Stock dynamics for forecasting material flows—case study for
nomic Systems Research 23, 27–47. housing in The Netherlands. Ecological Economics 59, 142–156.
Institute for Energy and Environmental Research Heidelberg, 2009. Umweltprämie Müller, D.B., Wang, T., Duval, B., Graedel, T.E., 2006. Exploring the engine of
oder Abwrackprämie? hhttp://www.bmu.de/files/pdfs/allgemein/application/ anthropogenic iron cycles. Proceedings of the National Academy of Sciences
pdf/ifeu_abwrackpraemie_bf.pdfi. of the United States of America 103, 16111–16116.
Japan Automobile Manufactures Association, 2012. Japan’s Measures to Withstand National Institute for Environmental Studies, Japan, 2012. Embodied Energy and
Impact of Global Crisis on its Automotive Industry. hhttp://www.jama-eng- Emission Intensity Data for Japan Using Input–Output Tables. hhttp://
lish.jp/asia/news/2009/vol36/index.htmli. www.cger.nies.go.jp/publications/report/d031/eng/datafile/index.htmi.
Jones, C.T., 1993. Another look at US passenger vehicle use and the rebound effect Next Generation Vehicle Promotion Center, Japan, 2010. Report on the Japanese
from improved fuel efficiency. Energy Journal 14, 99–110. Vehicle Replacement Scheme.
Kagawa, S., Nansai, K., Kondo, Y., Hubacek, K., Suh, S., Minx, J., Kudoh, Y., Tasaki, T., Pauliuk, S., Dhaniati, N.M.A., Müller, D.B., 2012. Reconciling sectoral abatement
Nakamura, S., 2011. Role of motor vehicle lifetime extension in climate change strategies with global climate targets: the case of the Chinese passenger vehicle
policy. Environmental Science and Technology 45, 1184–1191. fleet. Environmental Science and Technology 46, 140–147.
Knittel, C.R., 2009. The Implied Cost of Carbon Dioxide Under the Cash for Clunkers Sachs, J.D., 2009. A clunker of a climate policy. Scientific American 301, 34.
Program. University of California, Davis, CA. Small, K.A., Van Dender, K., 2007. Fuel efficiency and motor vehicle travel: the
Lane, B., 2006. Life Cycle Assessment of Vehicle Fuels and Technologies. London declining rebound effect. Energy Journal 28, 25–51.
Borough of Camden, London, UK. Suh, S., Lenzen, M., Treloar, G., Hondo, H., Horvath, A., Huppes, G., Jolliet, O., Klann,
Lave, L.B., Cobas-Flores, E., McMichael, F.C., 1995. Using input–output analysis to U., Krewitt, W., Moriguchi, Y., Munksgaard, J., Norris, G., 2004. System boundary
estimate economy-wide discharge. Environmental Science and Technology 29, selection for life cycle inventories using hybrid approaches. Environmental
420–4269. Science and Technology 38, 657–664.
Lenski, S.M., Keoleian, G.A., Bolon, K., 2010. The impact of ‘Cash for Clunkers’ on The World Bank, 2012. Global Economic Prospects January 2012. hhttp://sitere-
greenhouse gas emissions: a life cycle perspective. Environmental Research sources.worldbank.org/INTPROSPECTS/Resources/334934-1322593305595/
Letters 5, 1–8. 8287139-1326374900917/ExecutiveSummary_GEPJan2012_Eng.pdfi.