Content uploaded by Ryoji Hasegawa
Author content
All content in this area was uploaded by Ryoji Hasegawa on Feb 26, 2018
Content may be subject to copyright.
Available via license: CC BY 4.0
Content may be subject to copyright.
R E S E A R C H Open Access
Carbon footprint analysis through constructing
a multi-region input–output table: a case study
of Japan
Ryoji Hasegawa
1*
, Shigemi Kagawa
2
and Makiko Tsukui
3
* Correspondence: hasegawa@oiu.jp
1
Faculty of Global Business, Osaka
International University, 6-21-57
Tohdacho, 570-8555 Moriguchi,
Osaka, Japan
Full list of author information is
available at the end of the article
Abstract
In line with recent trends toward decentralization, prefectural and municipal
governments in Japan are becoming increasingly involved with managing global
warming in their regions. As a result, there is a new need to estimate the environmental
effects of regional economic activities, which can be used to establish effective energy
policies at the regional level. However, the details of these effects remain unclear due to
a lack of basic data. In this paper, we construct an original multi-region input–output
(MRIO) table based on interregional shipments among Japan’s47prefectures;thisis
done using the prefectures’single-region input–output (SRIO) tables and by applying
a non-survey technique. We use the constructed MRIO table, which we make freely
available online, to estimate the carbon footprint and carbon leakage of every region
and consider the structure of emissions at the regional level from the standpoints of
consumer and producer responsibility. The results reveal that production-based
emissions often differ significantly from consumption-based emissions. In addition,
the regional-level ratio of carbon leakagetocarbonfootprintis51.7%onaverage
and ranges from 34.8 to 79.8 %. Furthermore, the effects of economic activity in
and around Tokyo in terms of CO
2
emissions and leakage vary across regions.
JEL classification: Q54, R11, R15
Keywords: Carbon footprint; Multi-region input–output table; Carbon leakage;
Economic leakage
1. Background
In 1998, the Japanese government encouraged municipalities to voluntarily seek solu-
tions to global warming by enacting the Law Concerning the Promotion of Measures
to Cope with Global Warming. As a result, many governments at the prefecture and
municipality levels have become increasingly concerned about global warming issues
and the need for appropriate regional policies.
All prefectural governments in Japan estimate regional greenhouse gas (GHG) emis-
sions according to guidelines established by the Ministry of the Environment and pub-
lish concrete plans for GHG reduction. All prefectures also set up departments or
divisions focused on environmental problems. In contrast, many municipalities, espe-
cially smaller ones, do not have designated staff focused on environmental administra-
tion. Therefore, prefectures play a more important role than municipalities in
addressing global warming at the sub-national level.
© 2015 Hasegawa et al. This is an Open Access article distributed under the terms of the Creative Commons Attribution License
(http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided
the original work is properly credited.
Hasegawa et al. Journal of Economic Structures (2015) 4:5
DOI 10.1186/s40008-015-0015-6
As a prerequisite to framing appropriate environmental policies at the regional
level, it is necessary to quantify each region’s environmental burden. For instance,
Hasegawa [1] estimated CO
2
emissions among industries at the prefectural level in
Japan in 1995 and 2000 and clarified the differences in emissions structures among
regions and industries. Kudoh et al. [2] estimated CO
2
emissions from vehicles at
the municipal level in Japan and investigated policies for emissions reduction, taking
into consideration regional characteristics. Most studies, including these two, con-
fine their accounting to direct emissions in a region, excluding emissions related to
electricity.
When considering the scope of emissions for which a region is responsible, two types
of regional emissions can be identified. The first, “production-based”emissions, refers
to the CO
2
actually emitted by industries in a region as a result of production activities.
The other is “consumption-based”emissions: the CO
2
emissions that were required to
satisfy final demand for goods and services by a region’s population as well as those
emitted directly by households from private automobiles or for heating.
Production-basedemissionscangenerallybeidentifiedbysimplysummingdirect
emissions; however, accounting for consumption-based emissions requires consider-
ing not only direct but also indirect emissions. Consumption-based emissions, ex-
cluding direct emissions from households, are frequently referred to as a “carbon
footprint.”
A carbon footprint does not necessarily align with the direct emissions in a given region
due to regional characteristics such as the locations of industrial production and housing,
the pattern of interregional transactions, and the division of labor. In general, the smaller
the region examined, the larger the difference between its carbon footprint and its direct
emissions. The portion of a carbon footprint that is generated outside a region can be
interpreted as the region’s“carbon leakage¹.”These amounts have grown recently because
of the increasing division of labor and more frequent interregional transactions. Large
levels of carbon leakage make it difficult to estimate carbon footprints.
Accurate identification of carbon footprints by estimating carbon leakage among
regions and industries is necessary, however, to establish effective policies to counter
global warming at the regional level. An input–output (IO) model, especially a
multi-region IO (MRIO)² model, is one useful tool for carbon footprint analysis³.
Several studies have verified the usefulness of MRIO models in estimating compre-
hensive environmental loads. For instance, Vringer et al. [3] estimated land use and
GHG emissions induced by household consumption in the Netherlands by using
different models, including an MRIO model, and investigated the differences in em-
pirical results among models. The authors concluded that an MRIO model could
comprehensively estimate environmental loads induced by final demand and pro-
posed a hybrid multi-region (HMR) method that linked MRIO models with process
analysis.
Numerous studies have analyzed carbon footprints by using MRIO tables and focusing
on the environmental impact of international trade. For instance, Ackerman et al. [4] in-
vestigated how trade between Japan and the USA influences the volume of global CO
2
emissions by changing the trading structures in MRIO models. Su and Ang [5] calculated
exports and imports of CO
2
using the Asian International Input–Output Table compiled
by the Institute of Developing Economies Japan External Trade Organization (IDE-
Hasegawa et al. Journal of Economic Structures (2015) 4:5 Page 2 of 20
JETRO) in order to reveal the process through which emissions embodied in trade are
distributed within a country’s final demand through feedback effects.
Although carbon footprint analyses tend to adopt an international perspective,
several studies, such as those by Su and Ang [6] and Zhou and Imura [7], have used
MRIO models at the sub-national level to investigate comprehensive environmen-
tal impacts. Both of these studies focused on spatial variation and aggregation re-
lated to environmental impacts and conducted environmental analyses by using IO
tables compiled by IED-JETRO for eight regions of China. Notably, Su and Ang [6]
argued that it is essential to consider spatial aggregation in a country with large re-
gional variations in emissions structures, particularly for large countries such as
China.
Most studies, including the aforementioned, use officially compiled, existing MRIO
tables; as a result, there are few regions for which the proposed models are applicable.
It is more useful and flexible to construct an original MRIO table, using the well-
established method⁴, according to the objective of the study. This paper thus con-
structs an original MRIO table linking all of Japan’s 47 prefectures, using this process
to both highlight issues involved with constructing MRIO tables and propose a con-
struction method that is widely applicable in many settings. This paper also under-
takes a regional analysis of carbon footprints using the constructed MRIO table; the
results are shown to have important implications for climate policy.
The rest of the paper is organized as follows. In Section 2, we explain the method
used for constructing the MRIO table based on inter-prefecture shipments among all
prefectures in Japan. In Section 3, we develop a CO
2
emissions model based on the
constructed MRIO table. In Section 4, we estimate the carbon footprints and carbon
leakages for all prefectures in 2005 and analyze the results. Finally, we discuss prefec-
tural emissions responsibilities by considering both consumer and producer responsi-
bilities⁵; this allows us to identify significant outstanding issues in sub-national
environmental policy.
2. Construction of an MRIO table
2.1 Background
Regional IO tables can be classified into single-region input–output (SRIO) tables
and MRIO tables. While SRIO tables endogenously identify transactions among in-
dustrial sectors within a single region, MRIO tables endogenously consider the trans-
actions between multiple regions and reveal interrelationships among regions. MRIO
tables can therefore comprehensively calculate economic repercussions by consider-
ing the effects inside and outside a region as well as the rebound effects. SRIO tables
underestimate economic repercussions because they cannot account for all of the ef-
fects that are included in an MRIO table. It is thus desirable to use MRIO tables when
analyzing carbon footprints, as carbon leakage largely depends on economic
repercussions.
Although Japan’s Ministry of Economy, Trade, and Industry (METI) compiles an in-
terregional IO table for Japan that is segmented into nine regions, this segmentation is
not thorough enough for conducting a detailed analysis of the regional implications of
emissions policies. A district is not simply a large monolithic area but rather includes
Hasegawa et al. Journal of Economic Structures (2015) 4:5 Page 3 of 20
many local governments that are diverse in terms of emissions sources, as we later
show in our analysis.
It would be almost impossible to construct a complete interregional table from
scratch, as proposed by Isard [8], because this requires an enormous amount of data on
interregional trade that is unavailable. Most studies have instead constructed simplified
MRIO tables, typically of the Chenery–Moses type, using limited data in an operational
framework. Ishikawa and Miyagi [9], however, attempted to construct interregional IO
tables that covered all prefectures in Japan; in the process, the authors noted certain
issues concerning METI’s compilation of interregional IO tables. Ishikawa and Miyagi
focused on accurately estimating interregional trade coefficients using a large quantity
of interregional trade data as well as IO data and adjusted the estimated coefficients to
make the total output of all prefectures consistent with that of Japan as a whole; they
were thus able to propose a sophisticated method for constructing interregional IO
tables. In their resulting table, industrial sectors were aggregated into 45 sectors, pre-
sumably due to data limitations.
The 45-sector classification groups together industries with significantly different
emissions intensities as one sector. For example, “ceramic, stone, and clay products,”a
single sector in the classification, includes both cement and glass production, despite
the fact that cement’s emissions intensity is more than 20 times that of glass. The
aggregation level significantly affects the results of the IO calculations, particularly
when emissions intensities differ largely among the subsector’s industries; Jacobsen [10]
illustrated this by comparing results from 27-sector IO calculations to 117-sector re-
sults. A more detailed classification is thus required for accurate carbon footprint
analysis.
This paper constructs an original MRIO table with classification into over 45 sectors
using only the data available from SRIO tables; we are thus able to demonstrate a
method that requires less data and thus can be widely applied in different countries.
We do this by taking advantage of SRIO tables compiled by each prefectural office;
such tables are available for all prefectures in Japan. Using these, we construct an
MRIO table for 2005 consisting of every prefecture in Japan by applying a non-survey
technique.
Figure 1 shows the names and locations of Japan’s prefectures. As there are 47
prefectures in Japan, the constructed IO table consists of 47 regions. In the MRIO
tables presented in this paper, we have used the SRIO tables, which offer the most
detailed sector classification available, to compile industrial sectors with the highest
level of detail possible. As a result, we consider 80 industrial sectors, which corre-
sponds to the middle classification in the SRIO table compiled by METI, as shown
in Table 1.
2.2 Table construction methods
Figure 2 shows the framework of the MRIO table constructed in this paper. The grey
cells indicate transactions that can be identified from the prefectural SRIO tables, while
those in the other cells must be estimated. Data for the output vector (x), value-added
vector (v), final demand vector (f), and intermediate demand matrix (Z) in Fig. 2 is
taken directly from the prefectural SRIO tables, as is that for the foreign export vector
Hasegawa et al. Journal of Economic Structures (2015) 4:5 Page 4 of 20
(e), domestic export vector (d), foreign import vector (m), and domestic import vector
(n)⁶. Therefore, the balance equation is expressed using the available data as follows:
xr
i¼X
j
zr
ij þfr
iþer
iþdr
i−mr
i−nr
ið1Þ
In Eq. (1), xr
i,zr
ij ,fr
i,er
i,dr
i,mr
i,and nr
iare elements of x,z,f,e,d,m,andn,
respectively; superscripts and subscripts denote prefectures and industries,
respectively.
The prefectural SRIO tables are in the competitive foreign import (and domestic im-
port) form. This paper divides foreign import volumes into intermediate and final de-
mand by using foreign import coefficients; the same is done for domestic imports.
Fig. 1 Prefectures in Japan
Hasegawa et al. Journal of Economic Structures (2015) 4:5 Page 5 of 20
Foreign import coefficients Mr
i
and domestic coefficients Nr
i
are obtained using
Eqs. (2) and (3):
Mr
i¼mr
i
X
j
zr
ij þfr
i
ð2Þ
Table 1 Industrial sectors included in the MRIO table
(1) Agriculture,
forestry, and
fisheries
(21) Medicaments (41) Office machines and
machinery for service
industries
(61) Gas, steam, and hot
water supply
(2) Metal ores (22) Petroleum refinery
products
(42) Household electric and
electronic appliances
(62) Water supply and
other sanitary
services
(3) Nonmetal ores (23) Coal products (43) Electronic computing
equipment and
accessory equipment
(63) Trade
(4) Coal, crude
petroleum, and
natural gas
(24) Plastic products (44) Communication
equipment
(64) Financial service and
insurance
(5) Food and tobacco (25) Rubber products (45) Applied electronic
equipment and electric
measuring instruments
(65) Real estate agencies,
managers, and rent
(6) Drinks (26) Glass and glass
products
(46) Semiconductor devices
and integrated circuits
(66) House rent (imputed
house rent)
(7) Fabric (27) Cement and
cement products
(47) Electronic components (67) Transport
(8) Apparel and other
ready-made textile
products
(28) Pottery, china, and
earthenware
(48) Industrial heavy
electrical equipment
(68) Telecommunication
(9) Timber and
wooden products
(29) Miscellaneous
ceramic, stone and
clay products
(49) Other electrical
equipment
(69) Broadcasting
(10) Wooden furniture
and accessories
(30) Pig iron and crude
steel
(50) Motor vehicles (70) Information service
(11) Pulp and paper (31) Steel (51) Other motor vehicles (71) Internet services
(12) Converted paper
products
(32) Cast and forged
materials
(52) Steel ships and repair (72) Video and data
entry
(13) Publishing and
printing
(33) Other iron or steel
products
(53) Other transportation
equipment and repair
(73) Advertising services
(14) Chemical fertilizer (34) Nonferrous metals (54) Precision machinery (74) Public
administration
(15) Industrial inorganic
chemicals
(35) Nonferrous metal
products
(55) Miscellaneous
manufacturing products
(75) Education and
research institute
(16) Petroleum
chemical basic
products
(36) Metal products for
construction and
architecture
(56) Reuse and recycling (76) Medical service,
health, social
security, and nursing
care
(17) Organic chemical
products
(37) Other metal
products
(57) Construction and repair
of construction
(77) Goods renting/
leasing
(18) Resin (38) General industrial
machinery
(58) Public construction (78) Other business
services
(19) Chemical fiber (39) Special industrial
machinery
(59) Other civil engineering
and construction
(79) Personal service
(20) Final chemical
products
(40) Other general
machines and parts
(60) Electric power (80) Other
Hasegawa et al. Journal of Economic Structures (2015) 4:5 Page 6 of 20
Nr
i¼nr
i
X
j
zr
ij þfr
i
ð3Þ
By using Mand N, which are the diagonal matrices of Mr
iand Nr
i, respectively, we
can identify intraregional transactions (the elements along the diagonal), domestic in-
puts, and foreign imports within both intermediate and final demand, as shown in
Fig. 2.
zrr
ij ¼1−Mr
i−Nr
i
zr
ij ð4Þ
frr
ij ¼1−Mr
i−Nr
i
fr
ij ð5Þ
Equations (4) and (5) indicate intraregional transactions for intermediate demand and
final demand, respectively. Using these two equations, we can divide both intermediate
and final demand into internal supply, foreign imports, and domestic imports, as shown
in Eqs. (6) and (7).
zr
ij ¼zrr
ij þMr
izr
ij þNr
izr
ij ð6Þ
fr
i¼frr
iþMr
ifr
iþNr
ifr
ið7Þ
Figure 3 shows only the transactions that must be estimated, which are extracted
from Fig. 2; as such, it can be regarded as a matrix of domestic trade among regions.
The diagonal elements of the matrix are set to zero for both intermediate and final de-
mand because intraregional transactions are excluded. The matrix’s row and column
sums correspond to total domestic exports and imports in each region, as obtained
from the prefectural SRIO tables.
This paper applies the RAS method, using total domestic exports and imports in each
region as the control totals to estimate each element of the matrix shown in Fig. 3. In
other words, we simultaneously estimate trade flows to meet intermediate and final
Fig. 2 Framework for the constructed MRIO table
Hasegawa et al. Journal of Economic Structures (2015) 4:5 Page 7 of 20
demand by applying the RAS method to a rectangular matrix. Although the RAS
method is most commonly used to update or estimate a square input coefficient matrix,
it is frequently applied to other types of matrices. For example, Hasegawa [1] con-
structed energy flow matrices for Japan consisting of 33 sectors and 47 prefectures by
applying the RAS method modified for a rectangular matrix, and Lahr and Mesnard
[11] explained the application of the RAS procedure to a rectangular matrix.
In the RAS calculation, a matrix for approximation is required. We construct the ap-
proximation matrix by distributing total domestic imports (i.e., NZ and Nf ) for each
column across the 46 prefectures according to the ratio of the corresponding prefec-
ture’s monetary output to the total (the sum for all 46 prefectures). The elements of the
approximation matrix are expressed in Eqs. (8) and (9).
zrs
ij ¼xs
i
X46
sxs
i
Nr
izr
ij fors≠rð8Þ
frs
i¼xs
i
X46
sxs
i
Nr
ifr
ifors≠rð9Þ
The diagonal elements of the approximation matrix are set to zero⁷for consistency
with Fig. 3. As a result of the RAS calculation, the diagonal parts are estimated to be
zero and the other estimated elements are consistent in that the sums of the row and
column elements are equal to total domestic exports and imports, respectively.
The economic interpretation of the ordinal RAS procedure is that a matrix of inter-
mediate transactions is adjusted to consider substitution and fabrication effects⁸. This
arises because we use the RAS method on an interregional trade matrix including inter-
mediate and final demand; the RAS procedure thus involves not only the two effects
but also the change of the trade outlet and the substitution of demand between self-
supplied products and imports. It is assumed that the former occurs via adjustments to
the row quantities and the latter via adjustments to the column quantities.
The constructed table is available online as Additional files⁹1, 2, 3, 4, 5, 6, 7, 8, and 9.
Fig. 3 Matrix of domestic trade among regions
Hasegawa et al. Journal of Economic Structures (2015) 4:5 Page 8 of 20
3. Carbon footprint analysis method
In this section, we explain the method used for carbon footprint analysis based on the
MRIO table, which was constructed using the method explained in the previous sec-
tion. The output balance equation in the row direction is expressed via Eq. (10) because
domestic exports and imports are endogenous in the MRIO table.
x1
x2
⋮
xR
2
6
6
4
3
7
7
5
|fflfflffl{zfflfflffl}
X
¼
A11 A12 ⋯A1R
A21 A22 ⋯A2R
⋮⋮⋱⋮
AR1 AR2 ⋯ARR
2
6
6
4
3
7
7
5
|fflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl}
A
x1
x2
⋮
xR
2
6
6
4
3
7
7
5
|fflfflffl{zfflfflffl}
X
þ
f11
f21
⋮
fR1
2
6
6
4
3
7
7
5
|fflfflffl{zfflfflffl}
f1
þ⋯þ
f1R
f2R
⋮
fRR
2
6
6
4
3
7
7
5
|fflfflfflffl{zfflfflfflffl}
fR
þ
e1
e2
⋮
eR
2
6
6
4
3
7
7
5
|fflfflffl{zfflfflffl}
e
ð10Þ
In Equation (10), x,A,f
r
, and eare the output vector, input coefficient matrix, re-
gional final demand vector, and foreign export vector, respectively.
As shown in Fig. 2, the constructed IO table excludes foreign imports (m) in the en-
dogenous sector and the final demand sector by subtracting them from the column dir-
ection in a lump sum. Therefore, the foreign import vector is originally excluded from
Eq. (10). We develop Eq. (10) into Eq. (11), below.
x¼I−AðÞ
−1f1þ⋯þfRþe
ð11Þ
Next, we estimate both production-based emissions qr
P
and consumption-based
emissions qr
C
by linking the dataset of emissions coefficients in industries (c) and dir-
ect household emissions coefficients (h)¹⁰with Eq. (11).
qr
P¼Cxrð12Þ
qr
C¼CI−AðÞ
−1fr
|fflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflffl}
Carbon footprint
þHfr
|{z}
Direct emissions in households
ð13Þ
In Eqs. (12) and (13), Cand Hare diagonal matrixes with emissions coefficients c
and h, respectively. Equation (12) refers to production-based emissions—i.e., the CO
2
emitted to produce domestic final demand goods and exports. Equation (13) then con-
siders consumption-based emissions, including indirect emissions needed to satisfy re-
gional final demand, which is the first term on the right-hand side, and direct
emissions of households from private vehicles or heating, which is the second term on
the right-hand side.
This paper defines the first term on the right-hand side of Eq. (13) as a carbon foot-
print and investigates carbon leakage among industries and prefectures. Equation (13)
does not include the carbon footprints generated in foreign countries to produce Ja-
pan’s foreign imports: we focus on analyzing carbon footprints and leakage for domes-
tic emissions only.
4. Results
4.1 Total emissions at the prefectural level
First, we investigate total emissions at the prefectural level. Table 2 shows total and per
capita emissions for each prefecture, broken down into production and consumption
causes. In terms of total emissions, Chiba (12) has the largest volume of production-
Hasegawa et al. Journal of Economic Structures (2015) 4:5 Page 9 of 20
Table 2 Production- and consumption-based emissions in each prefecture
Emissions from production Emissions from consumption
Total Per capita Total Per capita
(Mt CO
2
)(tCO
2
) (Mt CO
2
)(tCO
2
)
(1) Hokkaido 44 (10) 7.8 (25) 48 (5) 8.5 (10)
(2) Aomori 10 (31) 7.1 (31) 11 (29) 7.5 (33)
(3) Iwate 7 (40) 5.3 (43) 10 (31) 7.6 (30)
(4) Miyagi 17 (21) 7.1 (30) 18 (15) 7.6 (26)
(5) Akita 8 (38) 7.2 (29) 9 (34) 8.1 (18)
(6) Yamagata 6 (43) 5.1 (46) 9 (35) 7.6 (29)
(7) Fukushima 49 (7) 23.4 (2) 16 (17) 7.7 (25)
(8) Ibaraki 51 (6) 17.0 (6) 25 (11) 8.5 (9)
(9) Tochigi 12 (29) 5.8 (40) 15 (20) 7.4 (36)
(10) Gumma 13 (27) 6.3 (35) 14 (22) 6.9 (43)
(11) Saitama 30 (15) 4.3 (47) 47 (6) 6.7 (48)
(12) Chiba 91 (1) 15.0 (8) 44 (7) 7.3 (38)
(13) Tokyo 87 (2) 6.9 (33) 157 (1) 12.4 (1)
(14) Kanagawa 65 (4) 7.4 (26) 73 (2) 8.3 (12)
(15) Niigata 33 (13) 13.5 (10) 20 (13) 8.4 (11)
(16) Toyama 12 (30) 10.4 (17) 8 (39) 7.4 (37)
(17) Ishikawa 9 (34) 8.0 (24) 11 (30) 9.0 (4)
(18) Fukui 20 (19) 24.4 (1) 7 (43) 8.0 (21)
(19) Yamanashi 5 (46) 5.3 (42) 7 (42) 7.9 (22)
(20) Nagano 16 (22) 7.3 (28) 16 (18) 7.2 (41)
(21) Gifu 15 (25) 7.3 (27) 14 (21) 6.8 (46)
(22) Shizuoka 32 (14) 8.4 (23) 31 (10) 8.3 (14)
(23) Aichi 79 (3) 10.9 (15) 66 (4) 9.0 (3)
(24) Mie 22 (18) 11.9 (11) 19 (14) 10.0 (2)
(25) Shiga 7 (41) 5.2 (45) 12 (26) 8.6 (8)
(26) Kyoto 15 (24) 5.8 (41) 18 (16) 6.7 (47)
(27) Osaka 57 (5) 6.4 (34) 72 (3) 8.1 (17)
(28) Hyogo 47 (8) 8.5 (22) 40 (8) 7.2 (39)
(29) Nara 5 (44) 3.6 (48) 10 (33) 6.8 (44)
(30) Wakayama 15 (23) 14.9 (9) 9 (38) 8.3 (15)
(31) Tottori 4 (47) 5.9 (39) 5 (47) 7.9 (23)
(32) Shimane 8 (39) 10.6 (16) 6 (45) 8.1 (19)
(33) Okayama 35 (12) 17.9 (4) 15 (19) 7.9 (24)
(34) Hiroshima 45 (9) 15.6 (7) 25 (12) 8.7 (6)
(35) Yamaguchi 26 (16) 17.1 (5) 11 (27) 7.5 (31)
(36) Tokushima 9 (36) 11.4 (13) 7 (41) 8.7 (7)
(37) Kagawa 9 (35) 9.2 (18) 8 (40) 7.6 (27)
(38) Ehime 17 (20) 11.6 (12) 13 (24) 9.0 (5)
(39) Koch 5 (45) 6.2 (37) 7 (44) 8.2 (16)
(40) Fukuoka 43 (11) 8.5 (21) 36 (9) 7.1 (42)
(41) Saga 10 (33) 11.1 (14) 6 (46) 6.8 (45)
(42) Nagasaki 13 (26) 8.9 (20) 11 (28) 7.6 (28)
Hasegawa et al. Journal of Economic Structures (2015) 4:5 Page 10 of 20
based emissions, followed by Tokyo (13), Aichi (23), Kanagawa (14), and Osaka (27).
On the other hand, Tokyo (13) has the largest volume of consumption-based emissions,
far higher than the next-highest prefectures of Kanagawa (14), Osaka (27), Aichi (23),
and Hokkaido (1). Focusing on per capita emissions, Fukui (18) has the largest volume
of production-based emissions, followed by Fukushima (7), Oita (44), Okayama (33),
and Yamaguchi (35); the ranking for consumption-based emissions is Tokyo (13), Mie
(24), Aichi (23), Ishikawa (17), and Ehime (38).
These results imply that emissions volume and the relative ranking for a given prefec-
ture differ significantly depending on the criteria for calculating emissions. In 24 pre-
fectures, production-based emissions are larger than consumption-based emissions,
whereas in the remaining 23 prefectures, the inverse is true. Figure 4 plots production-
and consumption-based emissions at the prefectural level based on the framework
presented in Table 2. In this figure, the horizontal axis represents production-based
emissions and the vertical axis, consumption-based emissions. The more distant a
marker is from the 45° line in the figure, the larger the imbalance in the two categories
of emissions. Consumption-based emissions are much larger than production-based
emissions in Tokyo (13), while Fukui (18) shows an opposite trend.
Figure 4 shows that there is less prefectural variation in consumption-based emis-
sions than in production-based emissions. The former varies mostly depending on per
capita total expenditures on final demand and the latter on industrial structure and per
Table 2 Production- and consumption-based emissions in each prefecture (Continued)
(43) Kumamoto 10 (32) 5.2 (44) 13 (23) 7.2 (40)
(44) Oita 24 (17) 19.5 (3) 9 (37) 7.5 (35)
(45) Miyazaki 7 (42) 6.1 (38) 9 (36) 8.0 (20)
(46) Kagoshima 12 (28) 6.9 (32) 13 (25) 7.5 (34)
(47) Okinawa 8 (37) 6.2 (36) 10 (32) 7.5 (32)
Total 1163 –9.1 –1061 –8.3 –
Note: Figures in parentheses indicate the ranking in terms of emissions volume
Fig. 4 Differences between production- and consumption-based emissions in each prefecture
Hasegawa et al. Journal of Economic Structures (2015) 4:5 Page 11 of 20
capita total industrial output. Emissions intensities show wide variations among indus-
tries, with an approximately 1500-fold difference between the largest and smallest in-
tensities in the 80-sector classification. Therefore, we expect that differences in
industrial structure are the largest contributor to the large variation in production-
based emissions. Figure 4 also shows large differences between the two emissions cat-
egories in many prefectures, indicating the importance of investigating carbon foot-
prints at the sub-national level.
Next, we address the prefectures’carbon footprints. As defined in Eq. (13) in Section 3,
a carbon footprint is regarded as consumption-based emissions, excluding direct house-
hold emissions. Figure 5 shows a breakdown of the carbon footprints generated by each
prefecture within its boundaries and throughout the rest of Japan. Carbon footprints gen-
erated in the rest of Japan are interpreted as carbon leakage. The ratio of carbon leakage
to total carbon footprint averages 51.7 % at the prefectural level, ranging from 34.8 %
(Okinawa (47)) to 79.8 % (Shiga (25)). The results reveal that carbon leakage is relatively
large and differs significantly across prefectures. Therefore, it is essential to identify car-
bon leakage more quantitatively to estimate the prefectures’carbon footprints.
4.2 Carbon leakage and economic leakage: the case of Tokyo
In this subsection, we focus on the carbon footprint derived from final demand in
Tokyo. Tokyo directly and indirectly induces production and emissions in not only
Tokyo but also other prefectures. It is clear that Tokyo’s influence on other prefectures
is significant, but the influence differs in terms of economic or emissions-related ef-
fects. These differences are not always clearly identified at the prefectural level. Table 3
shows the carbon footprint and induced production derived from final demand for sev-
eral industrial sectors in Tokyo for which the differences are large or notable.
Fig. 5 Composition of prefectures’carbon footprints
Hasegawa et al. Journal of Economic Structures (2015) 4:5 Page 12 of 20
Table 3 Tokyo’s carbon footprint and induced production (in %)
Agriculture, forestry
and fisheries (1)
Fabric (7) Metal products for construction
and architecture (36)
Carbon
footprint
Induced
production
Carbon
footprint
Induced
production
Carbon
footprint
Induced
production
(1) Hokkaido 8.85 9.34 0.50 0.47 2.51 1.47
(2) Aomori 2.45 2.61 0.10 0.09 0.68 0.34
(3) Iwate 1.79 1.96 0.10 0.13 0.26 0.70
(4) Miyagi 2.14 2.22 0.27 O.29 1.20 2.14
(5) Akita 1.55 1.52 0.26 0.14 0.25 0.40
(6) Yamagata 1.66 1.79 0.37 0.48 0.14 0.36
(7) Fukushima 3.57 2.26 2.51 0.60 1.85 1.09
(8) Jbaraki 3.97 3.52 2.70 2.05 8.04 5.91
(9) Tochigi 1.70 2.02 0.27 0.39 11.8` 2.76
(10) Gumma 1.79 1.94 1.55 1.91 0.62 1.75
(11) Saitama 1.06 1.33 1.20 1.82 1.01 3.22
(12) Chiba 5.66 4.54 3.40 1.90 13.51 8.32
(13) Tokyo 14.29 17.43 8.49 13.19 13.51 8.32
(14) Kanagawa 2.21 2.01 2.25 2.07 4.13 3.34
(15) Niigata 3.48 2.82 2.74 2.00 1.78 1.91
(16) Toyama 0.96 0.86 1.93 1.56 2.59 5.87
(17) Ishikawa 0.70 0.70 4.115 3.35 2.21 0.35
(18) Fukui 0.96 0.45 5.80 5.52 0.93 0.87
(19) Yamanashi 0.59 0.68 0.61 0.46 0.57 1.32
(20) Nagano 2.26 2.19 0.61 0.46 0.57 1.32
(21) Gifu 0.78 0.79 4.03 5.00 0.86 1.86
(22) Shizuoka 2.51 2.61 12.72 13.22 7.01 5.99
(23) Aichi 2.51 2.61 12.72 13.22 7.01 5.99
(24) Mie 1.52 1.43 1.93 1.63 0.92 2.25
(25) Shiga 0.58 0.76 3.26 4.49 0.47 2.42
(26) Kyoto 0.60 0.66 3.09 3.22 0.42 0.66
(27) Osaka 1.56 2.01 5.47 7.70 4.54 8.70
(28) Hyogo 1.51 1.53 3.30 3.08 6.75 5.33
(29) Nara 0.36 0.43 0.62 0.86 0.20 0.87
(30) Wakayama 1.60 1.57 3.95 5.20 2.71 1.55
(31) Tottori 0.48 0.52 0.12 0.13 0.05 0.12
(32) Shimane 0.78 0.69 0.51 0.38 0.32 0.35
(33) Okayama 1.38 1.14 2.90 2.73 6.67 2.34
(34) Hiroshima 1.07 0.89 2.04 1.85 9.44 4.11
(35) Yamaguchi 1.32 0.86 2.43 1.16 1.97 1.73
(36) Tokushima 1.11 1.00 1.01 0.87 0.40 0.35
(37) Kagawa 0.90 0.94 0.39 0.41 1.00 1.96
(38) Ehime 2.22 0.91 4.22 2.63 0.65 1.96
(39) Koch 0.94 1.01 2.21 2.27 0.11 0.97
(40) Fukuoka 2.19 2.29 0.59 0.67 5.25 3.20
(41) Saga 1.44 1.37 0.33 0.18 0.31 0.34
(42) Nagasaki 1.99 1.87 0.39 0.10 0.30 0.20
Hasegawa et al. Journal of Economic Structures (2015) 4:5 Page 13 of 20
Table 3 Tokyo’s carbon footprint and induced production (in %) (Continued)
(43) Kumamoto 2.03 2.04 0.41 0.45 0.46 1.39
(44) Oita 1.71 1.52 0.89 0.66 4.12 1.32
(45) Miyazaki 2.23 2.32 1.35 0.76 0.27 0.25
(46) Kagoshima 0.52 0.52 0.45 0.44 0.19 0.27
(47) Okinawa 0.58 0.53 0.08 0.08 0.17 0.19
Total (%) 100 100 100 100 100 100
Total (1000 t Co
2
or
billion yen)
1180 814 180 91 60 30
House rent (imputed house rent)
(66)
Total final demand
Carbon
footprint
Induced
production
Carbon
footprint
Induced
production
(1) Hokkaido 2.26 0.12 1.72 1.13
(2) Aomori 0.48 0.02 0.35 0.24
(3) Iwate 0.52 0.03 0.31 0.33
(4) Miyagi 0.91 0.07 0.72 0.63
(5) Akita 0.43 0.02 0.54 0.20
(6) Yamagata 0.21 0.02 0.54 0.20
(7) Fukushima 3.83 0.07 5.65 0.90
(8) Ibaraki 4.23 0.12 2.82 1.18
(9) Tochigi 0.96 0.05 0.57 0.73
(10) Gumma 0.83 0.06 0.58 0.83
(11) Saitama 2.73 0.25 1.55 1.81
(12) Chiba 5.41 0.19 5.51 1.65
(13) Tokyo 31.74 96.54 41.78 65.68
(14) Kanagawa 3.16 0.25 2.98 2.73
(15) Niigata 2.29 0.06 2.77 0.67
(16) Toyama 0.59 0.04 0.57 0.30
(17) Ishikawa 0.42 0.03 0.52 0.28
(18) Fukui 1.61 0.03 2.20 0.27
(19) Yamanashi 0.21 0.02 0.18 2.26
(20) Nagano 0.88 0.05 1.22 0.76
(21) Gifu 1.18 0.07 0.70 0.53
(22) Shizuoka 2.55 0.15 2.48 1.95
(23) Aichi 4.81 0.23 3.11 3.02
(24) Mie 1.64 0.08 1.24 0.90
(25) Shiga 0.82 0.05 0.39 0.59
(26) Kyoto 0.71 0.07 0.65 0.70
(27) Osaka 3.24 0.46 2.56 3.36
(28) Hyogo 2.72 0.14 1.86 1.54
(29) Nara 0.12 0.02 0.13 0.21
(30) Wakayama 0.96 0.03 0.72 0.26
(31) Tottori 0.10 0.01 0.11 0.13
(32) Shimane 0.59 0.02 0.74 0.16
(33) Okayama 1.95 0.07 1.45 0.65
(34) Hiroshima 2.41 0.10 1.75 0.81
Hasegawa et al. Journal of Economic Structures (2015) 4:5 Page 14 of 20
The CO
2
emitted in other prefectures is described as carbon leakage. Similarly, the
production induced in other prefectures can be also described as “economic leakage”
because Tokyo “leaks”economic activities into other prefectures through the economic
repercussions of satisfying its final demand. We investigated carbon leakage and eco-
nomic leakage induced by Tokyo’s final demand within sectors to identify the differ-
ences in carbon and economic leakage. Table 3 shows the prefecture-country ratio of
the carbon footprint and production induced by Tokyo’s final demand.
First, we focus on leakages based on Tokyo’s final demand from the agriculture,
forestry, and fisheries sector (1). The carbon footprint generated by Tokyo accounts for
14.3 % of the total amount in Japan. At the same time, 17.4 % of the induced produc-
tion is generated in Tokyo, which implies that its carbon leakage is greater than its
economic leakage. Focusing on the breakdown by prefecture, Tokyo’s carbon leakage is
larger than its economic leakage to Fukushima (7) and Chiba (12), while the inverse is
true for Hokkaido (1) and Osaka (27). This offers a concrete example of how Tokyo’s
influence on the economy and environment differs across prefectures.
In the fabric (7) sector, 8.5 % of the carbon footprint and 13.2 % of the monetary output
are generated in Tokyo. Tokyo’s final demand generates a greater carbon leakage than
economic leakage around its own prefecture. Table 3 also shows that Tokyo generates a
relatively large carbon leakage to western Japan and a large economic leakage around its
own prefecture when considering metal products for construction and architecture (36).
In the house rent (imputed house rent) (66) sector, 96.5 % of induced production is
generated in Tokyo, meaning that there is almost no economic leakage. In contrast, the
carbon footprint generated in Tokyo accounts for only 31.7 % of the national total, and
Tokyo leaks carbon to many prefectures. Therefore, there are large differences between
carbon and economic leakage in this sector.
Finally, we verify the differences between carbon and economic leakage based on
total final demand in Tokyo. These are also estimated using the METI-compiled inter-
regional IO table, which is segmented into nine regions. Table 4 compares the results
Table 3 Tokyo’s carbon footprint and induced production (in %) (Continued)
(35) Yamaguchi 2.20 0.05 1.39 0.47
(36) Tokushima 0.57 0.03 0.65 0.21
(37) Kagawa 0.58 0.04 0.44 0.27
(38) Ehime 1.03 0.06 0.94 0.43
(39) Koch 0.51 0.01 0.19 0.09
(40) Fukuoka 3.14 0.15 1.55 1.08
(41) Saga 0.53 0.01 0.67 0.19
(42) Nagasaki 0.64 0.01 1.00 0.21
(43) Kumamoto 0.44 0.03 0.30 0.30
(44) Oita 1.59 0.04 1.09 0.37
(45) Miyazaki 0.30 0.01 0.23 0.19
(46) Kagoshima 0.70 0.03 0.69 0.37
(47) Okinawa 0.29 0.02 0.23 0.11
Total (%) 100 100 100 100
Total (1000 t Co
2
or billion yen) 1011 7974 130439 117219
Hasegawa et al. Journal of Economic Structures (2015) 4:5 Page 15 of 20
obtained from our MRIO table to the METI table. Although not completely consistent
because aggregation levels of regions and sectors vary between the two tables, both sets of
figures show that Tokyo generates a larger induced output than carbon footprint in Kanto
while the remaining eight regions show the opposite trend. However, when we investigate
this for each of the prefectures included in Kanto, as shown in the right-hand side of
Table 3, we find that a larger carbon footprint than induced production is generated in
Ibaraki (8), Chiba (S12), Kanagawa (14), Niigata (15), Nagano (20), and Shizuoka (22).
These prefectures thus show the opposite tendency to Kanto as a whole.
Table 3 shows that the carbon footprint generated in Tokyo accounts for 41.8 % of
the national total, while 65.7 % of induced production is generated in Tokyo. Table 3
also confirms that Tokyo’s carbon leakage is larger than its economic leakage.
5. Conclusions
This paper, motivated by increasing concern on the part of sub-national governments over
global warming, analyzed prefectural carbon footprints in Japan. We constructed an
original MRIO table, which we made freely available online, using data from 2005 that
consisted of all prefectures in Japan and 80 industrial sectors; by applying a non-survey
technique, we determined the structure of emissions at the prefectural level. The
Table 4 Comparison of results between our original MRIO tables and the METI table (in %)
Estimated by the table
constructed in this paper
Estimated by the table
compiled by METI
Carbon
footprint
Induced
production
Carbon
footprint
Induced
production
Hokkaido 1.72 1.13 1.84 1.13
Including prefecture (1)
Tohoku 7.79 2.61 8.13 2.63
Including prefectures (2), (3), (4), (5), (6), and (7)
Kanto 62.42 78.25 64.90 80.26
Including prefectures (8), (9), (10), (11), (12),
(13), (14), (15), (19), (20), and (22)
Chubu 6.13 5.03 6.92 5.10
Including prefectures (16), (17), (21), (23), and (24)
Kinki 8.51 6.92 8.01 5.79
Including prefectures (18), (25), (26), (27), (28), (29), and (30)
Chugoku 5.44 2.23 4.85 2.17
Including prefectures (31), (32), (33), (34), and (35)
Shikoku 2.22 1.00 1.68 0.83
Including prefectures (36), (37), (38), and (39)
Kyushu 5.53 2.71 3.47 2.01
Including prefectures (40), (41), (42), (43), (44), (45), and (46)
Okinawa 0.23 0.11 0.20 0.09
Including prefecture (47)
Total (%) 100 100 100 100
Total (1000 t CO
2
or billion yen) 130,439 117,219 128,071 118,482
Note: The table shows carbon footprint and induced production derived from total final demand in Tokyo (13)
Hasegawa et al. Journal of Economic Structures (2015) 4:5 Page 16 of 20
emissions structure was compiled considering both consumer and producer responsibility.
We also investigated the carbon footprint generated by final demand in Tokyo and identi-
fied Tokyo’s carbon leakage and economic leakage to prefectures across Japan.
Our analyses revealed that in many prefectures, production-based emissions differ
significantly from consumption-based emissions. There were larger variations in
production-based emissions than in consumption-based emissions. We also found that
the ratio of carbon leakage to carbon footprint averages 51.7 % at the prefectural level
and ranges from 34.8 % (Okinawa) to 79.8 % (Shiga).
Although various activities in prefectures affect both types of emissions, production-
based emissions are strongly influenced by policies enacted to attract industry, the spatial
division of labor, and production technology at the prefectural level. Consumption-based
emissions, in contrast, are shaped by consumer behaviors, such as consumption patterns
and environmental consciousness, as well as the scale of final demand. Based on these re-
sults, we conclude that environmental policies within each prefecture should divide emis-
sions sources and address them by considering producer and consumer responsibility.
We also investigated the carbon footprint and production induced by Tokyo, Japan’s
largest metropolitan area. It is clear that Tokyo’s influence on other prefectures is sig-
nificant, but this influence differs between economy and emissions. We illustrated how
these differences are not accurately identified at the regional level by comparing Tables 3
and 4. Our analysis noted differences at the prefectural level and found certain prefectures
benefitting from or suffering a loss in terms of carbon footprint and induced production
due to final demand in Tokyo. The results indicate that Tokyo’s influence in terms of
carbon and economic leakage varies significantly from prefecture to prefecture and that,
as a whole, Tokyo has a larger carbon leakage than economic leakage.
The prefectural variation in production-based emissions results from the industrial
distribution promoted by each industry at the national level; prefectural govern-
ments’policies have not strongly influenced industrial activities in the past. This
suggests that prefectural governments are less responsible for production-based
emissions than is the national government and have difficulty in directly addressing
these.
To implement effective emissions reduction policies, prefectures should thus focus on
addressing consumption-based emissions from the viewpoint of consumer responsibility.
This is because Japan’s prefectures can exercise relatively more discretion when framing
environmental policies related to the residential sector. To reduce consumption-based
emissions by promoting environmentally friendly consumer behavior, it is important to
inform consumers of how regional characteristics affect carbon footprints. The method-
ology and results presented in this paper can help to do this.
Before closing, we will note some topics for future research. In constructing its MRIO
table, this paper adopted a method that requires limited data to disaggregate industrial
sectors and can be widely applied. As a result, however, it was necessary to sacrifice
some precision in the constructed table. Interregional trade was estimated using the
RAS method based on output shares; finding alternatives to this approach is an area in
need of future research. In order to facilitate more reliable analyses of carbon footprints
at the regional level, it is necessary to develop a method, as in Ishikawa and Miyagi [9],
for constructing an interregional IO table with more accurate estimations of interre-
gional trade at a detailed industry level.
Hasegawa et al. Journal of Economic Structures (2015) 4:5 Page 17 of 20
With respect to our carbon footprint analysis, there are some shortcomings that
should be noted. First, this paper did not consider the carbon footprint induced in for-
eign countries by Japanese final demand. Second, this paper did not alter emissions in-
tensities for a given industry depending on the prefecture and therefore did not address
regional differences in emissions intensities. Third, our analysis was confined to identi-
fying the current status of prefectural emissions and did not extend to analyzing prefec-
tural policies related to global warming.
Although these issues remain to be addressed, this paper has expanded the scope of
IO analysis of carbon footprints at the regional level; its approach can be applied to
undertake quantitative analysis of global warming policies considered by prefectures.
6. Endnotes
1
Generally speaking, carbon leakage refers to the phenomenon wherein overseas
emissions (especially those in countries with less strict environmental regulations) in-
crease because of emissions restrictions in a given country. However, this study con-
siders carbon leakage to refer to the more general case of economic activity in one
country (region) leading to induced emissions in another country (region) through the
division of labor and trade.
2
IO models linking multiple regions are classified into interregional and multiregional
IO models. The former consists of a complete set of intra- and interregional data and
is often labeled “Isard type.”The latter links single-region models using simplifications
and is often labeled “Chenery–Moses type”or “Leontief–Strout type”(see pp. 76–101
of Miller and Blair [12] for details of the two models). This paper mainly focuses on
MRIO as an IO model linking multiple regions, as compared with a single-region in-
put–output (SRIO) model.
3
In a special issue of Economic Systems Research devoted to carbon footprints, Minx
et al. [13] and Wiedmann [14] summarized the applications of the IO model to carbon
footprint analysis, including a brief description of the historical context.
4
Peter et al. [15] proposed a method to construct environmentally extended MRIO
tables using the database of the Global Trade Analysis Project (G-TAP) and used the
constructed table for analysis. The authors also proposed six key questions regarding
the construction of an MRIO table using the G-TAP database. Similarly, Muñoz and
Steininger [16] constructed MRIO tables from the G-TAP database in order to account
for Austria’sCO
2
responsibility due to consumption-based emissions.
5
Gallego and Lenzen [17] present a discussion related to producer and consumer re-
sponsibility for environmental burdens and attempt to construct a framework that uses
an IO model to assign producer and consumer responsibility.
6
Some prefectural SRIO tables do not distinguish between foreign and domestic ex-
ports. In such cases, we divided the given figures using data from SRIO tables for the
nine regions, as compiled by METI. The same procedure was followed for imports,
where necessary.
7
It has been reported that when the diagonal elements are set to zero, the RAS calcu-
lation is unable to converge or requires a large number of iterations to converge. How-
ever, our calculation easily converged, presumably because the diagonal elements of
our matrix account for only 1/47th of all elements.
Hasegawa et al. Journal of Economic Structures (2015) 4:5 Page 18 of 20
8
For an economic interpretation of the RAS procedure, see pp. 328–329 of Millar
and Blair [12] for a basic overview and Lahr and Mesnard [11] for more detailed
coverage.
9
Cite this paper when reporting analytical results or other studies that use the table
at a conference or in an article.
10
This paper uses the CO
2
emissions coefficients of Nansai and Moriguchi [18], who
calculate the emissions coefficients at the national level in Japan. Therefore, this paper
does not alter the emissions coefficient for a given sector across regions.
Additional files
Additional file 1: Constructed MRIO table (part1).
Additional file 2: Constructed MRIO table (part2).
Additional file 3: Constructed MRIO table (part3).
Additional file 4: Constructed MRIO table (part4).
Additional file 5: Constructed MRIO table (part5).
Additional file 6: Constructed MRIO table (part6).
Additional file 7: Constructed MRIO table (part7).
Additional file 8: Constructed MRIO table (part8).
Additional file 9: Constructed MRIO table (part9).
Competing interests
The authors declare that they have no competing interests.
Acknowledgments
The first version of this paper was presented at the 19th International Input–Output Conference, held in Alexandria,
USA, from 13th to 17th June 2011. The authors are grateful to anonymous referees for their useful comments and
suggestions for revising the paper.
Author details
1
Faculty of Global Business, Osaka International University, 6-21-57 Tohdacho, 570-8555 Moriguchi, Osaka, Japan.
2
Faculty of Economics, Kyushu University, 6-19-1 Hakozaki, 812-8581 Higashi-ku, Fukuoka, Japan.
3
Faculty of Business
and Commerce, Tokyo International University, 1-13-1 Matoba-kita, 350-1197 Kawagoe, Saitama, Japan.
Received: 16 June 2014 Accepted: 20 May 2015
References
1. Hasegawa R (2009) An examination of estimation methods employed for determining CO
2
emissions at the
prefectural level in Japan. Journal of Applied Input–Output Analysis 15:1–20
2. Kudoh Y, Kondo Y, Matsuhashi K, Kobayashi S, Moriguchi Y (2004) Current status of actual fuel-consumptions of
petrol-fuelled passenger vehicles in Japan. Appl Energy 79:291–308
3. Vringer K, Benders R, Wilting H, Brink C, Drissen E, Nijdam D, Hoogervorst N (2010) A hybrid multi-region method
(HMR) for assessing the environmental impact of private consumption. Ecol Econ 69(12):2510–2516
4. Ackerman F, Ishikawa M, Suga M (2007) The carbon content of Japan–US trade. Energy Policy 35(9):4455–4462
5. Su B, Ang BW (2011) Multi-region input–output analysis of CO
2
emissions embodied in trade: the feedback effects.
Ecol Econ 71:42–53
6. Su B, Ang BW (2010) Input–output analysis of CO
2
emissions embodied in trade: the effects of spatial
aggregation. Ecol Econ 70(1):10–18
7. Zhou X, Imura H (2011) How does consumer behavior influence regional ecological footprints? An empirical
analysis for Chinese regions based on the multi-region input–output model. Ecol Econ 71:171–179
8. Isard W (1951) Interregional and regional input–output analysis: a model of a space economy. Rev Econ Stat
33:318–328
9. Ishikawa Y, Miyagi T (2004) The construction of a 47-region inter-regional input–output table and inter-regional
interdependence analysis at prefecture level in Japan. ERSA conference papers, ersa04: 432, European Regional Science
Association, University of Porto, Porto, Portugal.
10. Jacobsen HK (2000) Energy demand, structural change and trade: a decomposition analysis of the Danish manufacturing
industry. Econ Syst Res 12(3):319–343
11. Lahr ML, Mesnard LD (2004) Biproportional techniques in input–output analysis: table updating and structural
analysis. Econ Syst Res 16(2):115–134
12. Miller RE, Blair PD (2009) Input–output analysis: foundations and extensions, 2nd edn. Cambridge University Press,
Cambridge
Hasegawa et al. Journal of Economic Structures (2015) 4:5 Page 19 of 20
13. Minx J, Wiedmann T, Wood R, Peters G, Lenzen M, Owen A, Scott K, Barret J, Hubacek K, Baiocchi G, Paul A,
Dawkins E, Briggs J, Guan D, Suh S, Ackerman F (2009) Input–output analysis and carbon foot-printing: an overview of
applications. Econ Syst Res 21(3):187–216
14. Wiedmann T (2009) Editorial: Carbon footprint and input–output analysis—an introduction. Econ Syst Res
21(3):175–186
15. Peters G, Andrew R, Lennox J (2011) Constructing an environmentally-extended multi-regional input–output table
using the GTAP database. Econ Syst Res 23(2):131–152
16. Muñoz P, Steininger KW (2010) Austria’sCO
2
responsibility and the carbon content of its international trade. Ecol
Econ 69(10):2003–2019
17. Gallego B, Lenzen M (2005) A consistent input–output formulation of shared producer and consumer
responsibility. Econ Syst Res 17(4):365–391
18. Nansai K, Moriguchi Y (2012) Embodied energy and emission intensity data for Japan using input–output tables
(3EID): For 2005 IO table (Beta+ version). CGER, National Institute for Environmental Studies, Japan.
http://www.cger.nies.go.jp/publications/report/d031/index.html. Accessed 18 March 2015.
Submit your manuscript to a
journal and benefi t from:
7 Convenient online submission
7 Rigorous peer review
7 Immediate publication on acceptance
7 Open access: articles freely available online
7 High visibility within the fi eld
7 Retaining the copyright to your article
Submit your next manuscript at 7 springeropen.com
Hasegawa et al. Journal of Economic Structures (2015) 4:5 Page 20 of 20
