Energy Policy 30 (2002) 511523

Direct and indirect energy requirements of households in India

Shonali Pachauri*, Daniel Spreng
Centre for Energy Policy and Economics, Swiss Federal Institutes of Technology, ETH Zentrum WEC, CH-8092 Zuerich, Switzerland
Abstract
This study is based on the 115 sector classication inputoutput tables for India for the years 198384, 198990 and 199394.
Calculated total primary energy intensities along with private nal consumption expenditures are used as a basis for determining the
indirect energy requirements of Indian households. Results reveal that total household energy consumption is about evenly divided
between direct and indirect energy and together comprises 75% of the total energy consumption of India. Most of household energy
consumed directly is still non-commercial and the consumption of food is responsible for about half the indirect energy
consumption. Household energy requirements have increased signicantly, both in total and per capita terms over this time period.
The commercial component of direct household energy consumption and the indirect energy requirements have increased
continuously. The main drivers of this increase have been (1) the growing expenditures per capita, (2) population and (3) increasing
energy intensity in the food and agricultural sectors. r 2002 Elsevier Science Ltd. All rights reserved.
Keywords: Household energy requirements; Household expenditures; Total primary energy intensities
1. Introduction
Energy services make up a sizeable share of total
household expenditures in developing countries. In
India energy consumption patterns vary widely across
dierent areas and groups of households. Average per
capita energy consumption is low compared to devel-
oped countries and even world average gures. How-
ever, the demand for energy using services in the
household sector has been growing at an increasing rate
since the early 1980s and is likely to expand rapidly in
coming years too. Given this scenario and the growing
share of India in global energy use and CO
2
emissions, it
is important to analyse the responsible factors. This
paper aims to do so by quantitatively analysing total, i.e.
direct and indirect energy requirements of households in
India over the years 198384 to 199394.
It is dicult to dene a typical Indian household,
however, the average, albeit not necessarily typical
Indian household in 199394 comprised 4.5 members
including 1.5 children. More than 70% of Indians
resided in rural areas and engaged in agriculture, more
than 80% owned their own dwelling, and about 66%
relied on rewood as their main source of direct energy
The per capita expenditure of the average Indian was
about Rs. 6226 (US$ 198.5). Direct commercial plus
non-commercial energy consumption of households in
199394 was on average half the value Goldemberg et al.
(1988) postulated to be required for a decent standard of
living in a warm climate. This latter value is about
10 GJ/capita or 0.3 kW/capita. Ocial statistics state
that virtually all villages in India are now electried;
however, several studies show that still barely 40% of
rural households have electricity connections. Erratic
power supply and low incomes result in very low
penetration rates for most electric ttings and electronic
appliances.
This scenario, however, is changing rapidly. One
study shows that while the penetration of consumer
durables increased at an average annual rate of 4%
between 198990 and 199293, it was nearly double this
between 199293 and 199596. Additionally, while the
penetration rate of electrical goods in urban areas was
3573 per thousand households as compared to only 883
in rural areas in 199596, if all rural households were
assumed to be electried the gure for rural areas would
jump to 2574 per thousand households (Natarajan,
1998).
Changes in technology, income levels and lifestyles
are bringing about important changes in both direct and
*Corresponding author. Tel.: +41-1-632-4188; fax: +41-1-632-
1050.
E-mail address: shonali.pachauri@cepe.mavt.ethz.ch (S. Pachauri).
0301-4215/02/$ - see front matter r 2002 Elsevier Science Ltd. All rights reserved.
PII: S 0 3 0 1 - 4 2 1 5 ( 0 1 ) 0 0 1 1 9 - 7
indirect energy requirements of household in the
country. Incomes and spending habits clearly have a
strong bearing on energy consumption. In addition, the
embodied energy requirements of the goods and services
consumed are aected by changes in the energy
intensities of the producing sectors. By using data on
household consumption expenditures along with infor-
mation on the energy intensities of the consumed goods
and services, it is possible to nd out more about the use
of primary energy in a country. In fact, the results of this
study indicate that total (direct and indirect) primary
energy use of Indian households amounts to more than
75% of total primary energy use in the country. The
remainder comprises government use and the energy
content of investments and net imports. While net
imports of energy are included in the analysis, net
imports of other goods and services are neglected, as
they are still relatively small for India. (See Fig. 1 for an
overview of aspects of energy requirements included in
the analysis.)
Two main features of the approach adopted in this
paper are worth highlighting:
*
Firstly, it includes non-commercial (biomass) energy
use, which is crucial from a household and develop-
ment perspective, especially for a large developing
country like India. Including non-commercial energy
is also an attempt to incorporate the non-formal or
unorganised sector of the economy into our analysis.
Energy inputs associated with animal and human
labour, which probably comprise an equally impor-
tant part of the unorganised or non-monetised
sector could not, however, be included due to a lack
of data.
*
Second, it focuses not only on direct energy but also
indirect energy requirements of households and so
allows one to account for technological change, to
some extent, as this is reected in changes in energy
intensities of the various sectors and sub-sectors.
The paper thus presents the total, direct and indirect
energy requirements per main consumption categories
for the period from 198384 to 199394 for Indian
households. The indirect energy requirements are
estimated in two stages. In the rst stage, we apply the
energy inputoutput analysis methodology to calculate
the direct and total primary energy intensities of
producing sectors making use of the 115 115 sector
inputoutput data for India. In the next step, we
multiply the private nal consumption expenditures
for dierent goods and services by the appropriate
sectoral total primary energy intensities to determine the
total energy requirements of household consumption.
Finally, we calculate the relative importance of the main
drivers of the observed changes over the study period by
disaggregating total household energy requirements for
each of the main consumption categories, by structure,
intensity, activity and population eects.
Gross i nvestment
Energy
embodied
in final
demand
Public consumption
Consumption of
private households
Direct
energy
requirements
Indirect
energy
require-
ments
Non-commercial
energy carriers
Indiginous
commercial energy
Imported
commercial energy
Primary
energy
input
Energy
requirements of
energy sectors
Agricultural, industrial
and service sectors
Primary energy input
Direct energy requirements of agriculture, industry and the service
sector, including Indian and foreign energy sectors
Direct energy requirements of private & public consumption
Indirect energy requirements of private & public consumption
Fig. 1. Direct and indirect energy requirements.
S. Pachauri, D. Spreng / Energy Policy 30 (2002) 511523 512
2. Methodology for calculating total energy intensities
The use of the engineering oriented process analysis,
the economic inputoutput analysis, or a hybrid of the
two is well established in the energy analysis literature.
While the seminal work in inputoutput analysis was
done by Leontief as early as the 1930s and 1940s
(Leontief, 1994), it was Bullard and Herendeen (1975)
and their colleagues at the University of Illinois who
pioneered techniques in energy inputoutput analysis in
the 1970s. The oil price shocks of the 1970s prompted a
spurt in energy analysis research in an eort to nd ways
to increase the eciency of energy use. (See Spreng
(1988) for a discussion of much of this early work and
also an illustration of a process analysis approach to
calculating total household energy requirements.)
Early research focused on the determination of total
energy intensities of goods and services or other
processes by inputoutput analysis, in OECD countries,
and include those for Germany by Denton (1975), for
the USA by Bullard and Herendeen (1975) and Wright
(1974), for Canada by Bush (1981), for the UK by Pick
and Becker (1975), and for New Zealand by Peet (1993).
For India, the rst published literature related to this
eld is a study done in the early 1980s to calculate total
energy intensities for 15 broad sector groupings based
on the 197879 inputoutput tables by Rao et al. (1981).
A more recent study by Tiwari (2000) has calculated
direct and total energy intensities for 38 aggregated
sectors based on the 198384 and 198990 inputoutput
tables.
For the purposes of this paper, we calculated total
and direct primary energy intensities at the disaggre-
gated 100 sector level for India, using the 115 115
sector inputoutput transactions tables, sectoral price
indices, and energy ow data, both commercial and
non-commercial, for the study years. Inputoutput
transactions tables for India for 198384, 198990 and
199394, prepared by the CSO (Central Statistical
Organisation, 1990, 1997, 2000) were used for the
analysis.
Actual quantity data for each energy type was relied
on to obtain the physical ows of energy across sectors.
As no single comprehensive published source for such
data is available, various ocial sources were used such
as the ocial publicationsFCoal Statistics, Petroleum
and Natural Gas Statistics, Power Statistics, various
issues on Energy by CMIE, various issues of
Electricity Statistics by CEA, and the TERI online
Energy Database (Central Electricity Authority, various
issues; CMIE, 1995; Department of Coal, various issues;
Ministry of Petroleum and Natural Gas, various issues;
Ministry of Power, various issues; Tata Energy Research
Institute, 1999). For non-commercial energy ows, data
estimates from the Regional Wood Energy Develop-
ment Programme in Asia of the FAO (1997) were relied
upon primarily but crosschecked with other published
estimates as well. Where the physical ows of energy
required further sectoral desegregation than was avail-
able in the published sources, monetary ow values in
the inputoutput tables along with energy price data
were used to make proportional allocations (see
Appendix A for a detailed description of the method
used for allocating energy ows).
Energy intensities by main energy carriers for dierent
sectors were calculated using the national inputoutput
transactions tables by applying economic inputoutput
energy analysis. Since the published inputoutput tables
included only the monetary value of transactions for
commercial energy sectors alone, two modications to
the matrices were carried out for the purpose of these
calculations as is mentioned in Steps 1 and 2 of the
description of the analysis below. In the rst stage of the
analysis, the 115 sector inputoutput tables were
aggregated to a 99 99 table. The reason why such an
aggregation was carried out was two-fold. On the one
hand, energy data was available only at a very aggregate
level and even though for reasons of accuracy it made
sense to keep the analysis at as disaggregate a level as
possible, for certain industries, namely mining, iron and
steel, heavy engineering and machinery, and other
transport equipment, early calculations done at the
115 sector level revealed implausible uctuations in
energy intensities over the studied decade. This was due
to the fact that the procedure for allocating energy from
a more aggregate sector (based on the classication of
sectors in the energy statistics) to sub-sectors of the
sector (following the classication in the inputoutput
tables) only provides accurate results under the assump-
tion that the sub-sectors that comprise the sector have
similar energy intensities. However, for the sub-sectors
that comprised the industries mentioned above direct
energy use are widely varying as the processes involved
in these industries are quite disparate and therefore the
assumptions used to allocate energy ows to these
sectors resulted in skewed results.
The second reason for aggregating the sectors of these
industries was, that since none of the products produced
by these industries were consumed as nal demand by the
household sector, energy intensities of these sectors did
not have a direct bearing on the calculation of total
household energy requirements and would not make
much of a dierence to the next stage of the analysis.
Therefore, it was decided to aggregate the sectors that
comprised the above-mentioned industries, thus eectively
transforming the 115115 matrix to a 99 99 matrix.
Briey, the inputoutput analysis consisted of the
following steps.
Step 1: The 115 sector inputoutput tables were
aggregated to 99 sectors, by aggregating all the mining
sub-sectors, the iron and steel sub-sectors and the heavy
engineering sub-sectors into single sectors. A new row
S. Pachauri, D. Spreng / Energy Policy 30 (2002) 511523 513
and column were added to each years table to represent
the non-commercial energy sector, thus eectively
making the classication for each year 100 100. Of
these 100 sectors, ve were energy sectorsFcoal and
lignite, crude petroleum and natural gas, petroleum
products, electricity, and non-commercial energy.
Step 2: Values in the energy rows were changed from
monetary to physical ows using the sources of data
mentioned above. The ratio of monetary ows to
physical ows of energy were also determined and cross
checked with published data on energy prices for
dierent sectors, wherever available, and appropriate
corrections made. For a detailed description of how
energy ows were allocated refer to Appendix A.
Step 3: Direct energy intensities for each energy
carrier for all sectors were calculated as the physical
energy ows into the sector per Rupee output.
Step 4: Total energy intensities for each energy carrier
were calculated with the aid of the Leontief inverse
matrices whose elements represent total direct and
indirect requirements per Rupee output for each sector.
Step 5: The energy intensities were deated to
constant 199394 prices using sectoral wholesale price
indices in order to make the values for dierent years
comparable. For the service sectors the overall con-
sumer price index for urban non-manual employees was
used.
For details of the computational procedures used for
calculating the various energy intensities and require-
ments please refer to Appendix 61, in Miller and Blair
(1985) on the mathematics of energy inputoutput
analysis.
3. Results
3.1. Energy intensities of goods and services
The ten most energy intensive non-energy sectors in
India (i.e. other than the energy sectors) in terms of their
total primary energy intensity, which we dene as the
sum of total coal and lignite intensity, total oil and
natural gas intensity, the ratio electricity from primary
sources (i.e. hydro and nuclear) times total electricity
intensity and total non-commercial energy intensity, are
presented in Table 1. The most energy intensive non-
energy sectors in 198384 as can be seen from the table
were Cement, Structural Clay Products, Fertilisers, Coal
Tar Products and Iron and Steel, whereas in 199394 the
most energy intensive sectors were Fertilisers, Paper,
Paper Products and Newsprint, Structural Clay Pro-
ducts, Cement and Hydrogenated Oil. Looking at the
total energy intensities by dierent energy carriers the
ranking of dierent sectors can be seen to be quite
dierent. Table 2 shows the top ten non-energy sectors
in terms of total coal and lignite intensity, total oil and
natural gas intensity, total petroleum products intensity,
total electricity intensity and total non-commercial
energy intensity in 199394. Detailed data on energy
intensities of dierent sectors for the dierent years are
Table 1
Total primary energy intensity of most intensive non-energy sectors
a
198384 198990 199394
Sector Total primary
energy intensity
(kgoe/Rs.
9394
)
Sector Total primary
energy intensity
(kgoe/Rs.
9394
)
Sector Total primary
energy intensity
(kgoe/Rs.
9394
)
Cement 0.23400 (0.9798) Structural clay
products
0.16142 (0.6759) Fertilisers 0.1860 (0.7788)
Structural clay products 0.22176 (0.9285) Fertilisers 0.14556 (0.6095) Paper, paper
products and
newsprint
0.1701 (0.7123)
Fertilisers 0.21386 (0.8954) Paper, paper products
and newsprint
0.13356 (0.5592) Structural clay
products
0.1666 (0.6975)
Coal tar products 0.20432 (0.8555) Cement 0.11798 (0.4940) Cement 0.1130 (0.4730)
Iron and steel 0.13574 (0.5684) Iron and steel 0.10846 (0.4541) Hydrogenated oil 0.0896 (0.3750)
Paper, paper products
and newsprint
0.13172 (0.5515) Inorganic heavy
chemicals
0.08728 (0.3654) Iron and steel 0.0894 (0.3744)
Other non-metallic
minerals
0.11972 (0.5013) Soaps, cosmetics and
glycerine
0.06779 (0.2838) Wheat 0.0704 (0.2947)
Organic heavy chemicals 0.09331 (0.3907) Other non-metallic
minerals
0.06700 (0.2805) Water supply 0.0672 (0.2814)
Hotels and restaurants 0.08304 (0.3477) Woolen textiles 0.06691 (0.2802) Printing, publishing
and allied activities
0.0653 (0.2733)
Railway transport service 0.08112 (0.3396) Coal tar products 0.06691 (0.2802) Woolen textiles 0.0598 (0.2503)
a
Note: Figures in brackets are total primary energy intensities in MJ/100 Rs.
S. Pachauri, D. Spreng / Energy Policy 30 (2002) 511523 514
Table 2
Most intensive non-energy sectors by energy carrierF199394
Sector Total coal
and lignite
intensity
(kgoe/Rs.
9394
)
Sector Total oil
and n. gas
intensity
(kgoe/Rs.
9394
)
Sector Total petro.
products
intensity
(kgoe/Rs.
9394
)
Sector Total
electricity
intensity
(kgoe/Rs.
9394
)
Sector Total non-
commercial
energy intensity
(kgoe/Rs.
9394
)
Cement 0.0967 Fertilisers 0.1246 Fertilisers 0.0610 Water supply 0.0188 Paper, paper products
and newsprint
0.1247
Iron and steel 0.0744 Other transport
services
0.0310 Other transport
services
0.0335 Wheat 0.0140 Structural clay
products
0.1245
Fertilisers 0.0560 Wheat 0.0228 Other non-metallic
mineral
0.0151 Cement 0.0090 Printing publishing
and allied activities
0.0379
Water supply 0.0532 Pesticides 0.0162 Pesticides 0.0138 Hydrogenated oil 0.0088 Hydrogenated oil 0.0334
Inorganic heavy
chemical
0.0502 Paddy 0.0151 Jute, hemp,
mesta textiles
0.0129 Fertilisers 0.0085 Hotels and restaurant 0.0313
Wheat 0.0429 Hydrogenated oil 0.0146 Wheat 0.0126 Non-ferrous
basic metals
0.0072 Woolen textile 0.0235
Misc. metal
products
0.0396 Other non-metallic
minerals
0.0144 Iron and steel 0.0123 Pesticides 0.0068 Tea and coee
processing
0.0199
Hydrogenated oil 0.0392 Jute, hemp,
mesta textiles
0.0130 Railway transport
service
0.0118 Soaps, cosmetics
and glycerine
0.0067 Rubber products 0.0191
Other non-metallic
minerals
0.0372 Iron and steel 0.0123 Hydrogenated oil 0.0117 Organic heavy
chemicals
0.0063 Jute, hemp, mesta
textiles
0.0191
Paper, paper
products and
newsprint
0.0366 Synthetic bres
and resin
0.0122 Cement 0.0110 Paints, varnishes
and lacquers
0.0063 Other chemicals 0.0171
S
.
P
a
c
h
a
u
r
i
,
D
.
S
p
r
e
n
g
/
E
n
e
r
g
y
P
o
l
i
c
y
3
0
(
2
0
0
2
)
5
1
1

5
2
3
5
1
5
not included in here but can be obtained from the
authors.
The contribution of direct energy intensity to the total
diers widely across sectors. Except in the case of
petroleum products, coal tar products, structural clay
products, cement, and electricity sectors the contribu-
tion of direct energy inputs to the total energy intensity
was o80% for each of the studied years. In fact, for
most machine/ engineering industries and some manu-
facturing sectors, the percentage of direct to total energy
intensity was very low, mostly o8% for all years.
Clearly, for these sectors the contribution of indirect
energy inputs to total energy intensity was relatively
high, often as large as ten times the direct inputs. From
an energy conservation perspective, this highlights the
importance of determining total energy intensities as
lowering direct energy use alone may not be the most
eective strategy for reducing total energy use. Reduc-
tions in indirect energy use through the substitution of
production factors and reducing materials use is
especially important for sectors near the end of
production chains.
Table 1 also reects some changes in energy intensities
over the decade from 198384 to 199394. The changes
in energy intensities over time reect the changing
technologies of production within the country as well as
changing patterns of demand for dierent forms of
energy. In general, energy intensities increased for many
sectors during the period 198384 to 198990 but then
experienced a decrease during 198990 to 199394. In
all, of the 100 sectors for which intensities were
determined, 56 sectors experienced a decline in energy
intensity over the decade, 2 sectors showed no change,
and the remaining 42 sectors had an increase in their
total primary energy intensities over this period. The
total primary energy intensity of total production, which
equals the average of the primary energy intensities over
all production sectors weighted by the production of
each sector, declined marginally from 0.029 kgoe/
Rs.
9394
to 0.025 kgoe/Rs.
9394
between 198384 and
198990 and then declined further to 0.022 kgoe/
Rs.
9394
by 199394. For most of the food and
agricultural sectors, however, the energy intensities
increased during the entire time period. This is explained
in part by the fast rate of growth of electricity use in the
agricultural sector during this time period. In fact,
annual average electricity growth rate in the agricultural
sector was 14% between 1976 and 1996, the highest
among all sectors in India. It is dicult to specify
exactly the degree of accuracy of the reported energy
data used in our analysis; however, it is possible that
some of the observed growth in electricity use in the
agricultural and household (direct) sectors may be a
result of better reporting and accounting. It is a known
fact that transmission and distribution losses in elec-
tricity supply are very large in India. The Planning
Commission estimates that these losses are as high as
20% or more of total electricity available and a
signicant proportion of these losses are attributed to
pilferage, especially by the household and agricultural
sectors where proper metering is often lacking (Planning
Commission, 1999). In addition, the rapidly rising
electricity use in the agricultural sector is also likely, in
part, to be a result of the very low, subsidised electricity
taris to agricultural users.
In all, 15 sectors experienced an increase in energy
intensity of more than 50% during this decade, largely
those in the food and agricultural sectors but also the
Printing, Publishing and Allied Activities sector, the
Pesticides sector, and the Wood Products sector. There
were also seven sectors including Cement, the Miscella-
neous Manufacturing sector and Rail Transport Ser-
vices sector, which experienced a decrease of greater
than 50% in their primary energy intensities. For many
sectors, the general shift from coal and lignite and non-
commercial energy use to greater oil and electricity use
during this period are responsible, in part, for the
observed decreases in primary energy intensities. Parti-
cularly, the substitution of non-commercial with com-
mercial energy may lead to signicant decreases in
energy intensity.
3.1.1. Comparison with energy intensity values from
other countries
It is problematic to compare energy intensities of
goods and services in one country with the equivalent
energy intensities in another country. Both dierences in
prices of goods and services and in the product mix of
the sector-outputs limit the usefulness of the compar-
ison. We therefore do not attempt to compare our
results for the energy intensities of goods and services in
India with results from other studies but cite a
comparison of energy requirements for the production
of three commodities for the recent UNDP/WEC-study
(Jochem et al., 2000). These specic energy requirements
are computed for well-dened products per weight
(Table 3).
It is clear from the table that despite improvements in
the energy intensity of certain sectors in India during the
studied period, there is a signicant potential for further
Table 3
Specic energy use in selected basic products industries and countries
Mid-1990
a
Country Steel
(GJ/tonne)
Cement
(GJ/tonne)
Pulp and paper
(GJ/tonne)
India 39.7 8.4 46.6
USA 25.4 4.0 40.6
Japan 17.5 5.0 F
Sweden 21.0 5.9 31.6
a
Source: UNDP/WEC 2000.
S. Pachauri, D. Spreng / Energy Policy 30 (2002) 511523 516
improvements in energy eciency by adopting more
ecient technologies and production processes.
3.2. Direct energy consumption of households
Total direct nal energy use in Indian households
increased by 14.3% during the decade between 198384
and 199394. However, if one considers total direct
primary energy use, which we dene as direct nal
energy use multiplied by the energy intensity of the
energy sectors in kgoe
in
=kgoe
out
or energy requirements
of energy, during the same period, then one observes
a slightly greater increase of 18.7%. In fact, there
was an overall increase in direct energy requirements
for all energy carriers other than coal and lignite
during this period, with electricity use increasing the
fastest.
Looking more closely at the trends in per capita direct
energy requirements, we observe that while non-
commercial energy use still dominates household direct
energy requirements, its use, in per capita terms,
declined continuously over the study period. In contrast,
both electricity and petroleum products use increased
rapidly, while direct consumption of coal in private
households remained small and decreasing (see Fig. 2).
Of all important energy carriers, the use of electricity
grew the fastest between 198384 and 199394 at the
household level. The large increase in electricity use is
primarily a result of greater accessibility for households
to electricity grids, and the faster penetration of
electrical and electronic equipment in households. These
shifts in the energy carriers also had an impact on the
total direct primary energy requirements as the energy
contents and eciencies of dierent fuels vary by a large
range. Thus, while direct commercial energy use
increased at an average rate of 4.3% per annum during
the studied decade this was oset by a decrease of about
1% per annum in per capita non-commercial energy
consumption. These trends are in conformance with a
progression up the energy ladder, a phenomenon well
documented by other researchers (Reddy and Reddy,
1994) wherein households move to more sophisticated
fuels and energy applications with increases in incomes
and urbanisation.
3.3. Total (direct plus indirect) energy consumption of
households
Studies examining specically the total (direct and
indirect) energy requirements of households using
inputoutput analysis have been done for the USA by
Herendeen et al. (1981), New Zealand (Peet et al., 1985)
and Switzerland (Ospelt et al., 1996). Studies extending
the analysis to calculate total carbon emissions include
those for the UK by Gay and Proops (1995) and for
Australia (Common and Salma, 1992), (Lenzen, 1998).
A series of publications by researchers at the University
of Groningen and University of Utrecht look at
household metabolic ows (Noorman et al., 1998),
(Vringer and Blok, 1995a, b; Vringer and Blok, 2000). A
recent publication in this eld that links consumption to
environmental pollution reported results from the
European study for Germany, France and Netherlands
on Consumer Lifestyles and Pollutant Emissions
Weber and Perrels (2000).
Studies on developing countries are more dicult to
nd. Researchers at the Indira Gandhi Institute of
Developmental Research have used inputoutput ana-
lysis to calculate the CO
2
emissions from energy
consumption for dierent groups of households for the
year 198990 (Murthy et al., 1997; Parikh et al., 1997).
However, to the best of the authors knowledge, the
present study is the rst attempt at examining changes in
total energy requirements, including non-commercial
energy, of dierent household consumption categories
in a developing country using inputoutput analysis at
the disaggregated 100 sector level.
In Fig. 3, total household energy requirements,
broken down by direct commercial, direct non-commer-
cial and indirect energy requirements for each of the
study years are presented. As in conventional energy
statistics, one observes a very large but declining
proportion of direct non-commercial to direct commer-
cial energy use. However, in addition, it is also clear that
the ratio of indirect energy use to the total rose over this
period. This conforms to the pattern observed in most
industrialized countries.
The average per capita total energy requirement of
private consumption increased from 266 kgoe (11.2 GJ) Fig. 2. Direct household energy requirements per capita.
S. Pachauri, D. Spreng / Energy Policy 30 (2002) 511523 517
in 198384 to 288 kgoe (12.1 GJ) in 199394, exhibiting
an annual average rate of increase of about 0.8%.
During the same period, the level of direct energy
requirements per capita decreased marginally at an
average rate of 0.3% per annum, but indirect per capita
energy requirements grew at a rate of 2.2% per year.
Indirect energy requirements made up 41% of total
household energy requirements in 198384 but this
percentage increased to 47% by 199394. The per capita
indirect energy requirements per main consumption
categories for the study years show an increasing trend
for all the main consumption categories except clothing
and footwear (see Fig. 4). The largest proportion of
indirect energy requirements was for the food, bev-
erages and tobacco category (more than 40%), which is
also the one that experienced the largest absolute
increase in total energy requirements. This can be
explained in part by the fast rate of growth of direct
energy use in the agricultural sector during this period.
Energy requirements of the transport and communica-
tions group also showed a large increase during this
period. For details regarding which sectors are included
in each consumption category refer to Appendix B. The
following section will discuss in greater detail the
changes in total and indirect energy requirements for
dierent consumption categories over the studied
decade.
3.3.1. Comparison with other studies
As this is the rst such study carried out for India,
and is also the rst to include non-commercial energy
use, it is dicult to compare the results obtained from
our analysis with other studies. Total per capita house-
hold energy requirements in most industrialized nations
for comparable time periods are a factor 710 times
higher than the value for India determined in this study.
Dierences in energy requirements arise from dierences
in direct and indirect energy use, both of which are
larger in industrialized countries.
Rather than comparing actual numerical values of
energy requirements for Indian households with those of
other countries, it is interesting to compare some of the
trends in household energy requirements between
countries. Studies carried out for Netherlands show
indirect energy requirements to comprise more than
50% of total household energy requirements (Vringer
et al., 1995a, b). In India, this share is still slightly less
than half of the total but has increased rapidly over the
studied decade and will most likely exceed 50% in
the near future. Within the indirect energy category, the
share of basic commodities like food, clothing, and
housing, was lower in the Netherlands in 1990 (Vringer
et al., 1995a) as compared to the share determined in
this study of these basic commodities. This seems
reasonable given the vast dierences in living standards
between the two countries. The only anomalous trend is
the rising share of indirect primary energy requirements
for food, beverages and tobacco in India. However, the
trends for other basic commodities show that the share
of these in total household energy requirements, both
for India and the Netherlands, is declining. Similarly,
the share of transport and communication is seen to be
rising for both countries.
3.4. Changes in household energy requirements
To understand the evolution in total and indirect
energy requirements of households during the period
198384 to 199394, we related the changes that have
occurred to four factors: (1) change in the composition
of the household shopping basket or structure of
consumption; (2) change in the productive energy
intensities of the goods and services consumed; (3)
0
30
60
90
120
150
180
210
240
270
300
1983-84 1989-90 1993-94
E
n
e
r
g
y

i
n

k
g
o
e
Indirect Direct Commercial Direct Non-Commercial Total
Fig. 3. Total household primary energy requirements per capita.
Fig. 4. Indirect primary energy requirements per capita for the main
consumption categories.
S. Pachauri, D. Spreng / Energy Policy 30 (2002) 511523 518
change in per capita levels of energy-using activities or
the per capita expenditures on each item; and (4)
changes in total population size. For the total and for
each indirect consumption category, we measured the
changes in activity, structure, intensity and population
and calculated the change in energy use that would have
occurred in response to each factor if the other three had
remained constant at the base year (198384) values. In
this way, we disaggregated the contributions to energy
use made by changes in the shares of household
expenditure (structure), the primary energy intensity of
each item (intensity), per capita household expenditures
(activity), and population in order to understand the
relative importance of these dierent factors.
The relation used in order to disaggregate the eects
of changes in structure, intensity, activity and popula-
tion on energy requirements was
THER
X
i
PFCE
i
P
i
PFCE
i
TPEI
i

P
i
PFCE
i
POP
POP;
where THER refers to total household energy require-
ments, PFCE
i
=
P
i
PFCE
i
is the share of expenditure on
each item as proportion of the total, TPEI
i
is total
primary energy intensity of each sector,
P
i
PFCE
i
=POP
is per capita private nal consumption expenditure,
POP refers to total population, and PFCE refers to
private nal consumption expenditure.
At the rst level, the above analysis was carried out to
disaggregate changes in the overall total household
energy requirements. Results indicate that during the
studied decade, structural changes were responsible for a
25% increase in total household energy requirements
(see Fig. 5). Changes in productive intensities on the
other hand, decreased total household energy require-
ments by 16%. However, changes in activity levels as
measured by real per capita household expenditures,
were responsible for a 55% increase in total household
energy requirements of Indian households during this
period. Similarly, population changes led to a 29%
increase in total energy use. It is clear from these gures
that increases in activity levels in Indian households
were the largest contributor to the increase in energy
requirements during the study period.
A similar disaggregation was also carried out on
changes in indirect energy requirements for each of the
main consumption categories. The results of that
analysis can be seen in Fig. 6. In general, for most of
the categories, other than clothing and footwear,
structural changes were responsible for a decrease in
indirect energy requirements. Changes in energy inten-
sity also led to a decrease in energy use for all categories
except food, beverages and tobacco, where it led to a
34% increase, and the medical and hygiene category
where it led to a smaller increase. Changes in activity
levels and population exerted an upward pull on energy
use for each category.
While it is also dicult to directly compare results of
this study with those of other Indian studies, as most
tend to exclude non-commercial energy use, a recent
paper using decomposition analysis for total commercial
primary and nal energy consumption and carbon
emissions for India between 1970 and 1995 reached a
similar conclusion regarding the importance of changes
in activity levels in driving the increase in energy use and
carbon emissions during this period (Nag & Parikh,
2000). Nag and Parikh conclude that India is likely to
become more and more important as a global energy
consumer and therefore autonomous technical change
will not be enough to attain the desired level of eciency
and emissions reduction. Mukhopadhyay and Chak-
raborty (1999) have carried out a structural decomposi-
tion analysis of energy consumption changes in India
between 197374 and 199192. Their analysis also
concludes that shifts in nal demand levels were the
most important driver of increase in total commercial
energy consumption during the period between 197374
and 199394.
4. Discussion and conclusions
The paper presents direct and indirect energy require-
ments of Indian households during the decade from
198384 to 199394 and disaggregates the changes in
total energy use during this period by structure,
intensity, activity and population eects. Results in-
dicate that rising real per capita expenditures, which we
refer to as the activity eect, has been the largest
contributor to increasing energy requirements during
this period. An overall 43% increase in real per capita
0
20
40
60
80
100
120
140
160
180
200
1983-84 1989-90 1993-94
Structural change Intensity change
Activity change Population change
Actual change
Fig. 5. Index values for the decomposition of changes in total
household energy requirements.
S. Pachauri, D. Spreng / Energy Policy 30 (2002) 511523 519
household expenditures over the study period translated
into an 8% increase in per capita total energy
requirements, implying an expenditure elasticity of
about 0.19 over this period. Future increases in
expenditure levels will lead to further increases in
household energy requirements.
Our ndings also indicate that the intensity eects led
to a slight decrease in total energy use during this
period but the structure, activity and population eects
led to a large increase in total energy requirements.
Findings from the disaggregation of indirect energy use
for each of the consumption categories are similar to
those for total energy use, with the exception of the
category food, beverages and tobacco and clothing
and footwear. The food category comprised almost
half of the indirect energy requirements of Indian
households and also experienced the largest increase
due to the fact that energy intensity for most of the
agricultural and food sectors also increased during this
period. Thus, the combined eect of increasing expen-
ditures on food and rising energy intensities in this
sector led to a large increase in energy requirements for
this consumption category. In fact, the combination of
direct energy needs for cooking and the indirect energy
requirements of the food, beverages and tobacco
consumption category make up about 6065% of total
household energy needs. While, food remains a basic
necessity whose consumption must increase given
increasing populations levels in India and widespread
poverty, the results of this study indicate that large
energy savings can be realised by increasing the
eciency of energy use both in the food producing
and manufacturing sectors, as well as in food prepara-
tion. This requires putting in place appropriate energy
pricing policies that encourage energy conservation,
especially for agricultural users; adopting newer more
energy ecient technologies in upstream sectors that
provide inputs to the agricultural sectors (for e.g.
fertilisers and pesticides); and encouraging the use of
more sophisticated direct fuels and appliances for
cooking.
Clothing and footwear was the second category
that exhibited changes contrary to the average during
this decade. This was the only category for which per
capita indirect energy requirements decreased. More
interestingly, this was the only category for which real
per capita expenditure levels remained almost constant
during the studied decade. In order to better understand
this trend, we looked at some of the physical textile
consumption data for India during this decade. Data
from the World Apparel Fibre Consumption Survey of
the FAO (FAO, 1996) shows that for the studied decade,
per capita availability of bres increased only margin-
ally. In addition, a structural change within this sector
took place during this decade, with larger increase in the
consumption of natural bres during the late 1980s, but
a decrease in the same during the early to mid 1990s and
a larger increase in the consumption in man-made bres
during this period. Greater durability of man-made
bres might, in part, be responsible for the almost
Fig. 6. Impacts of changes in structure, intensity, activity and population on dierent categories of indirect household energy requirements.
S. Pachauri, D. Spreng / Energy Policy 30 (2002) 511523 520
constant per capita expenditures for this category during
the studied decade.
The results presented here are for the total of all
households in India or for the average per capita
household energy requirement. However, given the high
degree of variation in living conditions and consumption
levels across dierent groups of households and urban/
rural areas in India, an important area for future
research will be to carry out a similar analysis separately
for urban and rural households and households belong-
ing to dierent income classes or geographical regions.
In addition, the analysis can be extended to calculate the
environmental consequences (insult) resulting from total
energy consumption of households. These environmen-
tal emissions gain increasing signicance at the house-
hold scale (especially for the large majority of the rural
population that still use non-commercial energy) be-
cause of their serious health impacts. This paper makes
a rst step towards understanding the causes of changes
in total household energy requirements. Future research
will provide more insights into the drivers of total
energy consumption across dierent groups of house-
holds by looking at household consumption from a
more micro perspective.
Acknowledgements
The authors would like to acknowledge the Alliance
for Global Sustainability, a university research partner-
ship between the Massachusetts Institute of Technology
(MIT), the Swiss Federal Institutes of Technology
(ETH), and the University of Tokyo (UT), for partial
funding of the work reported in this paper. We would
also like to thank Eberhard Jochem for valuable
comments made on earlier drafts.
Appendix A. Technical note on inputoutput
exercise: conversion of sectoral energy transactions from
value units to physical units
This note describes the various steps taken to
calculate direct and total energy intensities for India
using inputoutput data by directly incorporating the
energy usage patterns for each activity/sector in physical
terms into the inputoutput tables of the economy. The
exercise was performed on the 115-sector inputoutput
transactions matrix of the Indian economy, published by
the Central Statistical Organisation, Department of
Statistics, Ministry of Planning, Government of India
for the years 198384, 198990 and 199394. The
exercise used sectoral energy transactions (in value
terms) reported in the matrices as the starting point,
and then converted these values into physical quantities
of energy used. These physical quantities replaced the
original values of energy transactions in the 115-sector
matrix, and the modied transactions matrix was used to
compute a technology coecient matrix, which was used
as the basis of the inputoutput energy analysis. The follo-
wing describes in detail the transformation of sectoral
energy value transactions into physical quantities.
The original matrices provided information on fossil-
fuel usage by each of the sectors in terms of the Rupee
values (current prices) of coal and lignite (Sector 23),
crude petroleum and natural gas (Sector 24), petroleum
products (Sector 52), electricity (Sector 84) and gas
(Sector 85) used by each sector during each of the three
accounting years 198384, 198990 and 199394. The
coal classication included all varieties of coal, including
coking coal. The petroleum products classication is,
again, an aggregated one, inclusive of all renery
products.
(1) Coal and lignite. Data on the sectoral consumption
of various types of coal and lignite was limited. The data
for some sectors was available in the CMIE and CSO
publications and TEDDY. Since data on sectoral
consumption of various dierent types of coal for all
sectors was not available, to start with the physical
quantities of coal was allotted to those sectors for which
physical data was available. These related to the largest
consumers of coal such as the electricity sector, iron and
steel, etc. In the next step, and average price of coal was
estimated for each of the years based on the available
physical data and monetary values in the inputoutput
tables. Finally, monetary values were divided by the
calculated average price to arrive at the physical ows.
Lignite production data provided information on the
total lignite produced and consumed within the country.
In addition, the CEA General Reviews provided data on
the physical quantity of lignite consumed by the
electricity sector, which was approximately 65% of the
total. The remaining quantity of lignite was assigned to
the manufacturing sector as a share of total coal
consumption of each sector. The fractions used were
0.035 for 198384, 0.05 for 198990 and 0.075 for 1993
94.
(2) Crude petroleum and natural gas. Crude petroleum
and natural gas ows were almost entirely shown as an
input into the petroleum products activity, as was to be
expected, with some small values feeding into several
other sectors. These ows were interpreted as the
amount of natural gas being directly consumed by these
sectors. For some sectors (large consumers such as
fertilisers and electricity sector) data on the physical
quantities of natural gas consumed for the dierent
years was available. For the other sectors for which
physical ows were not available physical quantities of
natural gas use were allotted in proportion to the
monetary or value ows from the relevant sectors.
(3) Petroleum products. The conversion exercise for
petroleum products involved an additional step. A
S. Pachauri, D. Spreng / Energy Policy 30 (2002) 511523 521
distinction had to be made between categories of
distillates (light, middle and heavy). The demand for
fuels by the production activities (intermediate demand)
in the system was taken to consist of furnace oil (FO),
LSHS (both heavy distillates) and HSD and LDO
(middle distillates) and LPG and Naphtha (light
distillates). The nal demand component for petroleum
products was assumed to consist of LDO (middle),
motor spirit and kerosene (light distillates).
For all production sectors the following procedure
was used. The document Indian Petroleum and Natural
Gas Statistics (P&NGS) (198990 and 199394) pro-
vided us some idea of the usage patterns of renery
products by some of the production sectors. In addition,
the CMIE and CSO publications also provided some
data on the physical consumption of selected petroleum
products by dierent sectors. The prices of various
distillates were also obtained from the same sources. The
value of petroleum product usage provided by the
inputoutput matrix was decomposed into HSD and FO
components, using the value distributions obtained from
the P&NG Statistics. Since this value decomposition
was not available for all sectors separately, some
judgements about usage patterns were necessary. The
nal result was, therefore, a combination of P&NGS
distributions applied to as many sectors as possible and
own judgements for the rest of the sectors. The rupee
value for dierent petroleum products thus obtained
were converted to physical units (tonnes) using the price
information described above.
For the nal demand components of petroleum
products, as well as for imports, the value shares
provided in the P&NG statistics was used to convert
the matrix information on aggregate values into
product-wise values, and then to physical quantities,
using price information from the same source.
(4) Electricity. Data on electricity consumption by
dierent end-use sectors was culled out from the CMIE
and CSO publications, CEA publications and TEDDY.
As data for electricity use for each of the 115 sectors was
not available, price data was also made use of to convert
the value data to physical quantities. In general,
quantity data was used wherever available, and for
sectors where the disaggregate physical consumption
data was not available, assumptions were made such
that price of electricity for like industries/sectors (HT/
LT) was taken to be the same. Wherever, possible
crosschecks were also carried out between calculated
prices (monetary value/physical value) and published
data on prices for electricity for dierent sectors.
(5) Gas. GasFrefers only to Gobar Gas/Biogas
(LPG is included in the Petroleum, Products sector and
Natural Gas in the Crude Petroleum and Natural Gas
sector) and is very small in value, hence values in
this sector were not converted from monetary to physical
terms.
A.1. Note on creating a new non-commercial energy use
sector and allocating non-commercial energy to the
dierent end-use sectors
For data on the non-commercial energy use by broad
sector groupings (industry, services and households) for
dierent years we used the RWEDP database of the
FAO. The data was crosschecked with other sources of
published data on NCE use in India. Further desegrega-
tions to end sectors were done in proportion to total
Rupee value of expense on coal and lignite of each
sector. In addition, assumptions regarding, which
sectors are the prime users of non-commercial energy
were made based on data on the consumption pattern of
wood fuels in industries from the Wood Energy News
(newsletter) of the RWEDP in Asia of the FAO.
Appendix B. Sectoral composition of consumption
categories
1. Food, beverages and tobacco: Paddy, Wheat, Jower,
Bajra, Maize, Gram, Pulses, Sugarcane, Groundnut,
Jute, Cotton, Tea, Coee, Rubber, Coconut, Tobac-
co, Other Crops, Milk And Milk Products, Animal
Services (Agriculture), Other Livestock Products,
Forestry and Logging, Fishing, Sugar, Khandsari
and Boora, Hydrogenated Oil, Edible Oil other than
Vanaspati, Tea and Coee Processing, Miscellaneous
Food Products, Beverages, Tobacco Products
2. Clothing and footwear: Khadi, Cotton Textile in
Handlooms, Cotton Textiles, Woolen Textiles, Silk
Textiles, Art Silk, Synthetic Fibre Textiles, Jute,
Hemp, Mesta Textiles, Carpet Weaving, Ready Made
Garments, Miscellaneous Textile Products, Leather
Footwear
3. Housing and household eects: Furniture and
Fixtures, Wood and Wood Products Except Furni-
ture, Leather and Leather Products Except Footwear,
Rubber Products, Plastic Products, Other Chemicals,
Structural Clay Products, Other Non-Metallic Miner-
al, Miscellaneous Metal Products, Oce Computing
and Accounting, Other-Non-Electrical Machinery,
Batteries, Electrical Appliances, Other Electrical
Machinery, Watches and Clocks, Miscellaneous
Manufacturing, Ownership Of Dwellings
4. Medical care and hygiene: Soaps, Cosmetics and
Glycerine, Water Supply, Medical and Health Ser-
vices, Other Services
5. Transport and communication: Communication
Equipment, Ships and Boats, Rail Equipment, Motor
Vehicles, Motor Cycle and Scooter, Bicycles, Cycle-
Rikshaw, Other Transport Equipment, Railway Trans-
port Service, Other Transport Service, Communication
6. Education and recreation: Paper, Paper Products and
Newspapers, Printing Publishing and Allied Activ-
S. Pachauri, D. Spreng / Energy Policy 30 (2002) 511523 522
ities, Electronic Equipment Incl. Television, Hotels
and Restaurants, Education and Research
7. Other services: Trade, Banking, Insurance
References
Bullard, C.W., Herendeen, R., 1975. The energy costs of goods and
services. Energy Policy 3(4), 268278.
Bush, M.J., 1981. The energy intensities of commodities produced in
Canada. Energy 6, 503517.
Central Electricity Authority, various issues 198485, 198990, 1990
91, 199394, 199495. Public electricity supply all India statistics.
Government of India, New Delhi.
Central Statistical Organisation, 1990. Inputoutput transaction table
for 8384. Government of India, New Delhi.
Central Statistical Organisation, 1997. Inputoutput transaction table
for 198990. Government of India, New Delhi.
Central Statistical Organisation, 2000. Inputoutput transaction table
for 199394. Government of India, New Delhi.
CMIE (Centre for Monitoring the Indian Economy), 1995. (1996),
(1999) Indias energy sector. Mumbai, India.
Common, M.S., Salma, U., 1992. Accounting for changes in
Australian carbon dioxide emissions. Energy Economics 14(3),
217225.
Denton, R.V., 1975. The energy costs of goods and services in the
Federal Republic of Germany. Energy Policy 3(4), 279284.
Department of Coal, various issues 199091, 199394. Annual Report.
Government of India, New Delhi.
FAO, 1997. Regional wood energy development programme in Asia
(RWEDP). Wood Energy Database.
FAO 1996. World apparel bre consumption survey. FAO, Rome,
Italy.
Gay, P.W., Proops, J.L.R., 1995. Carbon-dioxide production by the
UK economy: an inputoutput assessment. Applied Energy 44,
113130.
Goldemberg, J., Johansson, T.B., Reddy, A.K.N., Williams, R.H.,
1988. Energy for a Sustainable World. Wiley Eastern Limited, New
Delhi.
Herendeen, R., Ford, C., Hannon, B., 1981. Energy cost of living,
197273. Energy 6 (12), 14331450.
Jochem, E., 2000. End-use eciency. In: UNDP, UNDESA, WEC
(Ed.), World Energy Assessment: Energy and the Challenge of
Sustainability. New York (Chapter 6).
Lenzen, M., 1998. Primary energy and greenhouse gases embodied in
Australian nal consumption: an inputoutput analysis. Energy
Policy 26 (6), 495506.
Leontief, W.W., 1994. The Structure Of American Society 19191929.
Oxford University Press, New York.
Miller, R.E., Blair, P.D., 1985. InputOutput Analysis: Foundations
and Extensions. Englewood Clis, Prentice-Hall, NJ.
Ministry of Petroleum and Natural Gas, various issues 198990, 1993
94. Annual Report. Government of India, New Delhi.
Ministry of Power, various issues 199091, 199394. Annual Report.
Government of India, New Delhi.
Mukhopadhyay, K., Chakraborty, D., 1999. Indias energy consump-
tion changes during 1973/74 to 1991/92. Economic Systems
Research 11 (4), 423438.
Murthy, N.S, Panda, M.K., Parikh, J.K., 1997. Economic develop-
ment, poverty reduction and carbon emissions in India. Energy
Economics 19, 327354.
Nag, B., Parikh, J.K., 2000. Indicators of carbon emissions intensity
from commercial energy use in India. Energy Economics 22, 441
461.
Natarajan, I., 1998. India market demographics report (1998) NCAER
(National Council for Applied Economic Research), New Delhi.
Noorman, K.J., Uiterkamp, T.S. (Eds.) 1998. Green Households?
Domestic Consumers, Environment and Sustainability. Earthscan
Publications Ltd., UK.
Ospelt, C., Knoepfel, I., Spreng, D., 1996. Direkter und indirekter
Energieverbrauch der Haushalte Bulletin SEV/VSE, 7, June, 1996.
Parikh, J.K., Panda, M.K., Murthy, N.S., 1997. Consumption patterns
by income groups and carbon-dioxide implications for India: 1990
2010. International Journal of Global Energy Issues 9 (46), 237
255.
Peet, N.J., 1993. Inputoutput methods of energy analysis. Interna-
tional Journal of Global Energy Issues. Special Issue on Energy
Analysis 5 (1), 1018.
Peet, N.J., Carter, A.J., Baines, J.T., 1985. Energy in the New Zealand
household, 19741980. Energy 10 (11), 11971208.
Pick, H.J., Becker, P.E., 1975. Direct and indirect use of energy and
materials in engineering and construction. Applied Energy 1 (1),
3151.
Planning Commission, 1999. Annual report on the working of state
electricity boards and electricity department, Government of India,
New Delhi.
Rao, S.S., Raizada, S.I., Ramanathan, A., 1981. Determination of
energy costs and intensities of and goods and services in the Indian
economyFan inputoutput approach. In: Chatterji, M. (Ed.),
Energy and Environment in Developing Countries. Wiley, New
York.
Reddy, A., Reddy, B.S., 1994. Substitution of energy carriers for
cooking in Bangalore. Energy 19 (5), 561571.
RWEDP (Regional Wood Energy Development Programme), 1996.
Wood Energy News 11(2), 16.
Spreng, D., 1988. Net Energy Analysis and the Energy Requirements
of Energy Systems. Praeger Publishers, New York.
Tata Energy Research Institute, 1999. TERI online energy database.
TERI, New Delhi.
Tiwari, P., 2000. An analysis of sectoral energy intensity in India.
Energy Policy 28, 771778.
Vringer, K., Blok, K., 1995a. The direct and indirect energy
requirements of households in the Netherlands. Energy Policy 23
(10), 893910.
Vringer, K., Blok, K., 1995b. Consumption and energy-requirement: a
time series for households in the Netherlands from 1948 to 1992.
NW & Sonderzoeksrapport no. 95016, University of Utrecht,
Utrecht.
Vringer, K., Blok, K., 2000. Long-term trends in direct and indirect
household energy intensities: a factor in dematerialisation? Energy
Policy 28, 713727.
Weber, C., Perrels, A., 2000. Modelling lifestyle eects on energy
demand and related emissions. Energy Policy 28, 549566.
Wright, D.J., 1974. Goods and services: an inputoutput analysis.
Energy Policy 2(4), 307315.
S. Pachauri, D. Spreng / Energy Policy 30 (2002) 511523 523