UNIVERSITAT
ROVIRA I VIRGILI
DEPARTAMENT D’ECONOMIA 

WORKING PAPERS
Col·lecció “DOCUMENTS DE TREBALL DEL
DEPARTAMENT D’ECONOMIA - CREIP”

Changes in Energy Output in a Regional Economy:
A Structural Decomposition Analysis

Maria Llop Llop

Document de treball n.6- 2016

DEPARTAMENT D’ECONOMIA – CREIP
Facultat d’Economia i Empresa

UNIVERSITAT
ROVIRA I VIRGILI
DEPARTAMENT D’ECONOMIA 

Edita:
Departament d’Economia
www.fcee.urv.es/departaments/economia/publi
c_html/index.html
Universitat Rovira i Virgili
Facultat d’Economia i Empresa
Av. de la Universitat, 1
43204 Reus
Tel.: +34 977 759 811
Fax: +34 977 758 907
Email: sde@urv.cat

CREIP
www.urv.cat/creip
Universitat Rovira i Virgili
Departament d’Economia
Av. de la Universitat, 1
43204 Reus
Tel.: +34 977 758 936
Email: creip@urv.cat

Adreçar comentaris al Departament d’Economia / CREIP

ISSN edició en paper: 1576 - 3382
ISSN edició electrònica: 1988 - 0820

DEPARTAMENT D’ECONOMIA – CREIP
Facultat d’Economia i Empresa

Changes in Energy Output in a Regional Economy:
A Structural Decomposition Analysis
Maria Llop
(Universitat Rovira i Virgili and CREIP)
Avinguda Universitat nº 1, 43204 Reus (Spain)
maria.llop@urv.cat

Abstract
In the input-output literature, structural decomposition analysis (SDA) has largely been used to
disentangle changes in key variables over time. This paper uses the demand-driven input-output
model and proposes a simple method to decompose the changes in energy gross output into
different determinants. Specifically, the total changes in energy output are divided into two
elements: technological changes, showing the effects of changing the technological coefficients
of the input-output model, and structural changes, showing the effects of changing the final
demand components. The empirical application, which is for the Catalan economy, uses two
input-output tables that cover an entire decade (2001 and 2011). The results show a positive
contribution of technology to increasing energy output, while the contribution of the final
demand for energy is negative. In addition, the various energy activities exert a different
repercussion on energy gross output changes. This highlights the importance of using detailed
methods in the study of energy issues.

Keywords
Energy Output, Input-Output Analysis, Structural Decomposition Analysis, Technological
Coefficients, Final Demand.
1

1. Introduction
Energy has become a strategic issue in the world economy in the twenty-first century. As
production activities require energy for use in the transformation processes, energy availability is
crucial for ensuring the production of goods and services that consumers demand. Moreover,
private welfare is directly related to energy uses and the modern standards of living require
increasing energy use.
Energy will face global challenges in the near future, which will require worldwide strategies.
First, it will be necessary to fill the gap between energy production and growing energy demand.
Second, the development of new forms of energy to ensure reliability, environmental
sustainability and inexpensiveness will be crucial. To achieve all these goals, major efforts must
be made towards developing alternative energy sources, reducing the environmental impacts of
energy production and energy consumption, promoting energy savings, increasing energy
efficiency and obtaining new transportation systems.
In recent decades, the input-output model has proved to be a helpful framework for analysing
energy issues. For instance, Proops (1988) used the input-output method to define indicators of
both direct and indirect energy consumption. Hawdon and Pearson (1995) applied the inputoutput model to the United Kingdom and showed the interrelations among energy issues, the
environment and the economy. Alcántara and Roca (1995) proposed an input-output method for
measuring the demand for energy and carbon dioxide emissions in Spain. Lenzen (2001)
constructed a generalized input-output model in which investment and imports were separated
from final demand and internalized into intermediate demand, and presented an empirical
application of energy multipliers for Australia. Manresa and Sancho (2004) estimated sectoral

2

energy intensities and CO2 emissions for the Catalan economy, using an extension of the inputoutput model: the SAM (social accounting matrix) model. In addition, Llop and Pié (2008)
defined a price version of the input-output model for Catalonia to analyse the economic impact
of alternative measures that could be implemented in the energy sector.
The input-output model has also largely been used to unmask the factors that underlie the
changes in an extensive set of variables over time, by defining intertemporal approaches. Thanks
to the usefulness of the input-output SDA methods, a large set of contributions has been
published based on different economies at different periods of time. In particular, the research
efforts undertaken to date have provided comprehensive and precise knowledge of the properties
of the SDA methods and their usefulness in both economic and environmental research. In this
field, Rose and Casler (1996) reviewed the literature produced so far. Dietzenbacher and Los
(1998) discussed some of the problems inherent in the non-uniqueness of the computational SDA
methods. Dietzenbacher and Los (2000) discussed the possible dependency between decomposed
determinants that may lead to incorrect conclusions when the co-related determinants are treated
independently. Dietzenbacher and Stage (2006) showed that the hybrid input-output model that
combines monetary and physical units may provide results that depend on the choice of units
rather than on changes in economic determinants. For environmental SDA methods, Su and Ang
(2012) examined the new methodological improvements dealing with energy and pollutant
emissions which have been developed in the last decade, and provided guidelines on method
selection.
As structural decomposition analysis has proved to be very helpful in unravelling the long-term
drivers of economic indicators, a wide range of issues have been analysed in this type of
literature. These issues include technological changes, changes in output, international trade
3

issues and changes in private consumption. The SDA has also proved to be extremely useful for
disentangling the factors that explain the changes observed in energy, such as the drivers behind
energy usages and energy efficiency.
In the contributions on SDA applied to energy, Gowdy and Miller (1987) examined the changes
in the US energy uses during the period 1963-1977, taking into account both technological and
demand changes. Casler and Hannon (1989) used the energy input-output model to explore the
reasons behind the changes in the US energy intensities, distinguishing between the direct and
indirect energy substitutions. Han and Lakshmanan (1994) analysed the modifications in
Japanese energy intensities between 1975 and 1985, and explicitly defined the contribution of
energy imports. Wier (1998) explored the changes in Danish gas emissions during 1966 and
1988, identifying changes in final demand, changes in input-output coefficients, changes in
energy intensities and finally changes in emissions from the household sector. Kagawa and
Inamura (2001) evaluated the sources of changes in Japanese energy demand between 1985 and
1990. Llop (2007) used an environmental input-output model for Spain to separate the effects of
changing the emission coefficients and the effects of changing the input-output structural
coefficients within the total changes in emission multipliers between 1995 and 2000. In another
study on the Spanish economy, Guerra and Sancho (2011) analysed the effects on energy uses
and emission levels caused by efficiency gains, changes in final demand and changes in input
requirements. More recently, Lan et al. (2016) compare the global energy footprint by
identifying various drivers and its distribution across different countries. This study, which
provides unique insights into the energy changes in all countries, is applied to 186 economies
and covers the period between 1990 and 2010.

4

To the best of my knowledge, all the SDA contributions applied to energy focus on explaining
the changes in (physical) energy uses and (physical) energy efficiency. Nevertheless, it would be
equally interesting to study the drivers behind the (monetary) gross output changes of energy. By
taking into account that energy is strategic for the economy and the society, a thorough
knowledge about the different forms of energy and its determinants over time is essential for
meeting both the economic and social challenges regarding energy.
The purpose of this paper is to identify the underlying factors that contribute to changing energy
gross output in an economy, with a particular focus on the various activities that make up the
energy sector. In particular, as energy is used by both production activities and final consumers
(i.e. households, government and the foreign sector), different economic agents can be
responsible for the alteration of energy production. Bearing this in mind, it is interesting to
analyse the patterns that explain the changes in energy output, by disentangling the effects
caused by the various agents involved (i.e. production sectors and final consumers) and by the
two possible sources of demand (i.e. intermediate demand and final demand). This analysis is
particularly appropriate for understanding the driving factors in the energy (economic) variables.
The empirical application is for the Catalan economy and uses two regional input-output tables,
covering the period 2001-2011. The results show that the decomposed effects behaved
differently: final demand contributed negatively to energy output while technological changes
contributed positively. As a result of these two (opposed) impacts, energy output experienced a
slight increase in real terms during the decade studied. In addition, the various energy activities
are very heterogeneous in relation to the individual temporal drivers observed in the regional
economy.

5

The rest of the paper is organised as follows. The next section shows the structural
decomposition analysis used to decompose the total changes in energy gross output. Section 3
describes the regional input-output tables used for the Catalan economy, and Section 4 contains
the empirical application. Finally, the last section of the paper concludes.

2. Changes in Sectoral Output and Decomposition
The standard representation of the demand-driven input-output model, in matrix notation, is:
x = Ax + y = (I – A)-1y = Ly,

(1)

where x is the vector of sectoral gross output, y is the vector of final demand, A is a n  n matrix
of technological coefficients (calculated by dividing the industry-by-industry direct requirements
by the sectoral output), and I is the identity matrix. In expression (1), (I – A)-1 = L is the n  n
matrix of input-output multipliers, or the Leontief inverse, and shows the overall effects (direct
and indirect) on sectoral output caused by unitary and exogenous increases in the final demand
of sectors.
The study of changes in gross output involves a comparison of variables for two different periods
of time. These periods are hereinafter denoted using the subscripts 0 (first period) and 1 (last
period). Moreover, as the model covers various years, there is the need to remove the distorting
effects of price changes on the relevant variables. This means that in order to avoid the influence
of prices, all the information in the input-output tables has to be expressed in terms of the prices
of a single year.
The so-called double deflation method enables price effects to be eliminated by revealing the
changes in variables in real terms. Assuming that each sector produces a single homogenous
good, then the sectoral gross output, intermediate demand and final demand are deflated by the
6

corresponding price index for this sector. Let pi be the ratio of the current price level and the
base-year price level for sector i. We denote by P the diagonal matrix containing the elements pi
on its main diagonal and zeros elsewhere. Expression (1) can therefore be defined in constant
prices as follows:
xd = Ad xd + yd = (I – Ad )-1yd ,

(2)

where xd = P-1x is the vector of deflated gross production, yd = P-1y is the vector of deflated final
demand and finally, Ad = P-1AP is the matrix of deflated input-output coefficients.
Let us now assume that all the components in the input-output model are in constant prices and
let us also assume an identical sectoral disaggregation in the two periods analysed.1 By taking the
first difference in expression (1), it follows that:2
x = L1y1 - L0y0
= L1(y0 + y) - (L1 - L) y0 = L1y + Ly0

(3.a)

= (L0 + L)y1 - L0 (y1 - y) = Ly1 + L0y,

(3.b)

where x, y and L contain the first differences of the elements in vector x, vector y, and
matrix L, respectively. Note that although expressions (3.a) and (3.b) are equivalent, the results
provided are necessarily different because the weights used to calculate the contribution of the
two increased components are different. Dietzenbacher and Los (1998) showed that the average
of the two polar decompositions yields a good approximation of the average of all the possible
decompositions. In particular, the average of expression (3) is equal to:
x = L(y0 + y1) + (L0 + L1)y.
1
2

In what follows, the superscript d is avoided for the sake of simplicity.
See Miller and Blair (2009) for a detailed presentation of SDA decompositions of sectoral output.
7

(4)

Next, we can go further in the decomposition of expression (4) by dividing the changes in the
matrix of multipliers (L) into the changes in the matrix of technical input-output coefficients
(A). If both sides of L0 = [I – A0]-1 are post-multiplied by [I – A0] and the derivative respect to
time is taken, it follows that:
L - L0A - LA0 - LA = 0;
L - L0A - LA0 - (L1 - L0)A = 0;
L = L1AL0.

(5.a)

Alternatively, changing the matrix and the period of post-multiplication gives rise to the
following equivalent expression:
L = L0AL1.

(5.b)

By substituting (5.a) and (5.b) into (4):
x = L(y0 + y1) + (L0 + L1)y
= (L1AL0y0 + L0AL1y1) + (L0 + L1)y
= (L1Ax0 + L0Ax1) + (L0 + L1)y.

(6)

Following this expression, the changes in sectoral gross output (x) have been calculated as the
addition of two different components. The first component reflects how sectoral output is
modified by changes in the technical coefficients of the input-output model, and the second
component reflects how changes in final demand modify sectoral gross output. In expression (6),
the total changes in production have therefore been disentangled as the aggregation of two
components with different economic interpretations: technological changes ( (L1Ax0 +
L0Ax1)) and structural changes ( (L0 + L1)y).
8

In order to individually quantify the effects of changing all the input-output coefficients used in
the model, matrix A can be decomposed by showing all the possible modifications that have
taken place. In specific terms, the individual effects of technical coefficients can be examined by
writing matrix A as the following sum:
A = A11 + A12 + A13 + … + Ann = ΣiΣjΔAij,

(7)

where each n  n matrix is made up of the single non-zero entry showing the change in the ijth
element, aij, and zeros elsewhere.
Additionally, the changes in the vector of final demand can be written as:
y = y1 + y2 + … + yn = ΣiΔyi,
where each decomposed vector is made up of the single non-zero entry showing the change in
the final demand of sector i. Changes in final demand can be further decomposed. If there are k
categories of demand instead of a single total value, then each element in vector y responds to yi1
+ yi2 + … + yik,, with yik being the amount of category k in the final demand from sector i. We can
therefore disentangle the changes in final demand by adding vectors containing the single nonzero entry showing the change in the final demand of the ikth element and zeros elsewhere, as
follows:
y = y11 + y12 + … + y1k + …+ yn1 + yn2 + … + ynk = ΣiΣkΔyik.

(8)

Using the divisions in (7) and (8), expression (6) can be re-written as:
x = [L1Ax0 + L0Ax1] + [L0 + L1]y
= [L1(ΣiΣjΔAij)x0 + L0(ΣiΣjΔAij)x1] + [L0 + L1](ΣiΣkΔyik). (9)

9

Expression (9) measures the effects of separately changing all the components in the model and
maintaining the changes occurred in all other components constant, as it shows the contribution
of every single change to total output changes. Note that this decomposition facilitates the
interpretation of results because among the huge set of elements within an input-output model,
the calculation in (9) makes it possible to individually illustrate the contribution of each one to
output modifications over time.

3. Databases
The empirical application is based on the latest two input-output tables available for Catalonia
(one for 2001 and one for 2011), published by the regional statistics office (IDESCAT). These
two databases provide statistical coverage of the regional economy for an entire decade.3
Before the empirical calculations, the original tables underwent a transformation process. The
official 2001 input-output table is limited to an extended use matrix at basic prices. This means
that neither the make matrix nor the symmetric matrix is available for 2001. To avoid this
statistical gap, the symmetric input-output table was indirectly calculated in a process involving
the estimation of the 2001 make table for Catalonia. This estimation took into consideration the
structure of the make matrix for the Spanish economy4 and some other indirect information
concerning sectoral production in the regional economy provided by the IDESCAT.
Subsequently, by using the extended use table and the previously estimated make table, a sectorby-sector symmetric table for 2001 was obtained.5 As the regional statistics office directly

3

IDESCAT (2015a, 2015b).
INE (2015a).
5
Miller and Blair (2009) describe the calculations required to obtain a symmetric input-output table by using the use
and the make matrices.
4

10

provides a symmetric input-output table, the only transformation required for the year 2011 was
the application of an identical sectoral aggregation as in the 2001 database.
The resulting two input-output tables have the same structure, consisting of twenty-eight
differentiated activities, five of which represent the energy sectors (Table 1): extraction of
minerals (sector 4), coke, petroleum and fuels (sector 5), electricity (sector 6), gas (sector 7) and
water (sector 8).
[PLACE TABLE 1 HERE]
To avoid price distortions, the 2001 input-output table was rescaled to the 2011 price levels
(expression (2)). In specific terms, vector x of gross output, vector y of final demand and matrix
A of structural coefficients were valuated at the 2011 price levels. For the service activities, the
elements in the diagonal matrix P were obtained from the consumption price index for the
regional economy (INE, 2015b). For the industrial activities, the elements in P were obtained
from the production prices index (INE, 2015c).

4. Empirical Application for the Catalan Economy
The analytical method defined in Section 2 shows the decomposition of gross output in all
sectors of production. However, as the interest in this study lies in analysing energy production
and its driving temporal determinants, the results in the next tables are limited to show the
accounts for the five energy activities in the Catalan input-output databases (j = 4, 5, 6, 7, 8).
[PLACE TABLE 2 HERE]
Table 2 shows the main economic indicators for the energy activities. In this table, the 2001
values have been rescaled to the 2011 price levels. For the aggregated energy sector, the

11

intermediate inputs increased by around 27% (from 8,780.8 million euros in 2001 to 11,137.8
million euros in 2011). The intermediate demand rose by 8.8% (with values of 18,812.4 and
20,470.6, respectively) and the final demand for energy decreased in real terms by a percentage
of around -14.8%. This decrease in demand is mainly explained by the significant reduction in
the energy exports, which is quantified by -46.1% (5,127.7 million euros in 2001 and 2,762.9
million euros in 2011). Finally, Table 2 shows that the total energy output was almost stable
during the period analysed, and shows an increase of 136.2 million euros, which represents a
variation of around 0.5%.
A closer look at Table 2 suggests that the different energy activities evolved asymmetrically
during the period studied. In particular, minerals and electricity reduced the intermediate inputs,
while coke, petroleum and fuel, gas and water increased the intermediate costs. On the other
hand, the final demand for electricity and gas decreased (by -52.5% and -27.5% respectively)
and increased in the other subsectors. Finally, the last row in Table 2 shows that the gross output
of electricity remained quite stable, the output of gas, water and minerals increased (by 28.3%,
50.1% and 10.6% respectively) and the output of coke, petroleum and fuel decreased by -14.6%.
In the light of these results, there is a heterogeneity in the intertemporal changes in the various
energy accounts, and no general trends can be traced from the figures in Table 2.
Table 3 contains the summarised results of the SDA decomposition. In specific terms, this table
divides the changes in the energy gross output into the contribution of the technological changes
(i.e. changes in the input-output technical coefficients) and the structural changes (i.e. changes in
the final demand for energy).
[PLACE TABLE 3 HERE]

12

Two main results can be seen from Table 3. The first is based on the fact that the two
decomposed elements have a different sign. While the contribution of technological changes to
energy production was positive (318.6 million euros), the contribution of structural changes was
negative (-182.47 million euros). At the individual level, changes in the technological relations
contributed positively in all subsectors except in coke, petroleum and fuel (-2,474.8 million
euros). However, changes in final demand behaved negatively in electricity and gas (-1,393.7
and -103.0 million euros respectively) and positively in the other energy activities.
The second result in Table 3 is based on the fact that the different energy subsectors behaved
differently in terms of their output changes during the period analysed. More specifically,
extraction of minerals, gas and water increased the real output (638.9, 665.3 and 567.4 million
euros, respectively), while the other two energy subsectors (coke, petroleum and fuel and
electricity) reduced the real output by -1,549.9 and -185.6 million euros. The aggregation of
these individual (and opposed) output changes results in a small overall increase of 136.1 million
euros. In addition, the energy activities show different patterns in the two isolated effects. While
in minerals and water both the technological and the structural component are positive, in
electricity and gas a positive technological change is combined with a negative structural change
and in coke, petroleum and fuel a (large) negative technological change is combined with a
positive structural change.
We can further analyse the factors underlying the intertemporal drivers of energy by showing by
how much the changes in matrix A of input-output coefficients contributed to energy output
modifications. Table 4 summarises the results for the coefficients of the five energy accounts
reflected in the model. Specifically, this table groups the coefficients’ changes into three
different categories, which have a different economic meaning. The changes in the column
13

coefficients reflect the modification in the cost structure, or the production effect, and show how
the energy activities have changed their intermediate purchases of products. The changes in the
row coefficients reflect the modification in the selling structure, or the substitution effect, and
show how the energy accounts have modified their intermediate output relations. The last two
columns in Table 4 show the joint modifications in rows and columns, or the mixed effect, and
reflect the changes in the total intermediate relations (i.e. purchases and sales) of the energy
activities. In Table 4, the values in parenthesis are the percentage of variation from the
corresponding initial value.
[PLACE TABLE 4 HERE]
In the coefficients column, there is an increase in the structure of the intermediate purchases in
coke, petroleum and fuel (35.6%), gas (75.8%) and water (29.8%). These activities therefore
show an increase in the proportion of their sectoral output that goes to buying inputs of
production to the rest of activities. However, the column coefficients decreased in extraction of
minerals (-36.8%) and to a lesser extent in electricity (-2.8%), meaning that these two sectors
reduced the fraction of their output spent on intermediate production from the other sectors.
In general, the row coefficients evolved positively during the period, illustrating an increase in
the proportion of energy output sold to the rest of the production system. The increase is
especially significant in water (157.9%) and gas (51.8%). The exception to this general increase
is coke, petroleum and fuel, where intermediate output coefficients fell by -38.6%. This may
show that the Catalan production system has substituted the intermediate uses of fuels by other
forms of energy.

14

The last two columns in Table 4 contain the joint row and column changes in coefficients, on
average terms and for each of the energy accounts. These figures reproduce the same sign as the
row changes, which predominate over the column changes in the total technological effects. As a
result, the technological component is positive in all energy activities except in coke, petroleum
and fuel. An interesting outcome is therefore that the proposed decomposition method shows a
different temporal adaptation of the structure of intermediate costs and intermediate output in the
various energy activities. Another result from the last column of Table 4 is that the contribution
of the technological component to output change is the combination of different drivers within
the technology of production, which are different in magnitude and sign in terms of the cost
structure (column changes) and the selling structure (row changes) of energy sectors.
The impact of the coefficient changes on energy output helps us to understand the effects within
the production system. Hereinafter, the analysis turns to the changes in the final uses of energy.
In specific terms, Table 5 disentangles the various elements that compound the final demand for
energy.
[PLACE TABLE 5 HERE]
Once again, there are asymmetrical effects in terms of both the different sectors and the different
elements within the final demand. By components, the reduction in energy exports (-2,364.8
million euros) is significant, which generated a reduction in output of -3,832.2 million euros.
This is mainly explained by the reduction in the exports of electricity, which show a negative
contribution on output of -2,714.6 million euros, and in the exports of gas, which show a
negative contribution of -1,121.1 million euros.

15

The other demand components in Table 5 evolved positively, especially final consumption,
which with a rise of 716.4 million euros, positively contributed to increasing energy gross output
(1,198.6 million euros). The increase in final consumption is significant in coke, petroleum and
fuel (670.4 million euros) and to a lesser extent in gas (310.4 million euros). On the contrary, the
rest of energy activities show a reduction in final consumption during the period analysed.
In relation to the various energy accounts, there is a significant reduction in the final demand for
electricity, explained by a reduction in electricity exports and to a lesser extent in private
consumption, which is quantified at -1,970.8 million euros. This negative contribution to the
energy output amounted to -1,393.7 million euros, a reduction of -4.7% compared to the 2001
figures for total energy output. Another interesting result in Table 5 is that the negative effect of
the electricity demand predominates over the other activities and the resulting final structural
component is slightly negative (-182.4 million euros).
The structural decomposition shows the importance of breaking down the overall changes in
energy output into their driving determinants. The changes in output are the result of a
combination of a huge amount of individual effects, which are extremely difficult to identify by
limiting the analysis to the final aggregated impacts. The results in these tables clarify which is
the individual contribution of hidden effects, that cannot be identified in general and aggregated
studies of the energy sector.

5. Conclusion and Policy Implications
This paper has analysed the changes in Catalan energy gross output during the period 2001-2011,
by dividing total changes in energy production into two different components: technological
changes and structural changes. The technological changes show the effects of changing the

16

intermediate coefficients in the input-output model. The structural changes show the effects of
changing the final uses for energy. In addition, the method proposed in the paper enables
individual quantification of the contribution of all the elements in the input-output model (i.e. the
changes in all the sector-by-sector coefficients and in all the sector-by-component elements of
final demand).
The results showed a moderate increase in the energy gross output during the period, which was
the result of combining a negative contribution of the final demand and a positive contribution of
the intermediate coefficients. In terms of the different energy activities, there is a different
sectoral repercussion on the changes in energy output. The technological changes positively
contributed in all sectors, except in coke, petroleum and fuel. On the other hand, the changes in
the demand for electricity and to a lesser extent gas contributed negatively to energy output
changes.
One important finding is therefore that intertemporal changes in the regional energy output are
mainly explained by changes in exports of electricity and by technological changes (i.e.
intermediate sales and purchases) of coke, petroleum and fuel. In the light of these results,
special attention should be paid to these specific areas of the energy system in order to define
successful energy policies in the regional economy.
The analytical framework proposed in this paper clarifies some aspects within the complex
process of changes in gross output. In particular, the input-output SDA analysis is an appropriate
method for disentangling the driving forces of temporal modifications in energy variables. This
information is obviously extremely useful for gaining further knowledge in order to be able to
attribute environmental responsibilities, as it enables the different agents involved in energy

17

usages and their changes over time to be identified. The outcomes in the paper therefore suggest
a need for using disaggregated and detailed methods to gain further insights into the driving
forces of changes in energy.
In the coming decades, the challenges in energy issues can only be met if important research
efforts are made to help the energy policy-making process. The need for new energy sources,
less energy-intensive production systems and environmental-friendly forms of energy will be the
focus of interest for academics, politicians and regulators. The message of this paper is that it is
necessary to focus on the underlying factors affecting energy production and energy
consumption as a means to clarify the complex aspects involved in energy issues.

Acknowledgements
The author gratefully acknowledges the financial support of the Chair of Energy and
Development (Universitat Rovira i Virgili, MADE and Diputació de Tarragona), the Spanish
Ministry of Culture (grant ECO2013-41917-P) and the Government of Catalonia (grant
SGR2014-299 and “Xarxa de Referència d’R+D+I en Economia Aplicada”).

References
Alcántara, V., Roca, J. 1995. Energy and CO2 emissions in Spain. Energy Econ. 17 (3), 221-230.
Casler, S. D., Hannon, B. 1989. Readjustment potentials in industrial energy efficiency and
structure. J. Environ. Econ. Management 17, 93-108.
Dietzenbacher, E., Los, B. 1998. Structural decomposition techniques: sense and sensitivity.
Econ. Syst. Res. 10, 307-323.

18

Dietzenbacher, E., Los, B. 2000. Structural decomposition with dependent determinants. Econ.
Syst. Res.12, 497-514.
Dietzenbacher, E., Stage, J. 2006. Mixing oil and water? Using hybrid input–output tables in a
structural decomposition analysis. Econ. Sysy. Res. 18, 85-95.
Guerra, A. I., Sancho, F. 2011. An analysis of primary energy requirements and emission levels
using the structural decomposition approach: The Spanish case, in: Llop, M. (Ed.), Air
pollution: Economic modelling and control policies. Bentham E-Books Series.
Gowdy, J. M., Miller, J. L. 1987. Technological and demand change in energy use: an inputoutput analysis. Environ. Plan. A 19, 1387-1398.
Han, X., Lakshmanan, T. K. 1994. Structural changes and energy consumption in the Japanese
economy 1975-1985: an input-output analysis. Energy J. 15, 165-188.
Hawdon, D., Pearson, P. 1995. Input-output simulations of energy, environment, economy
interactions in the UK. Energy Econ. 17 (1), 73-86.

IDESCAT, 2015a. Taula Input-Output de Catalunya 2001. Institut d’Estadística de Catalunya,
Barcelona.
IDESCAT, 2015b. Taula Input-Output de Catalunya 2011. Institut d’Estadística de Catalunya,
Barcelona.
INE, 2015a. Tabla de Origen 2001. Instituto Nacional de Estadística, Madrid.
INE, 2015b. Índice de Precios al Consumo. Instituto Nacional de Estadística, Madrid.
INE, 2015c. Índice de Precios de Producción. Instituto Nacional de Estadística, Madrid.

19

Kagawa, S., Inamura, H. 2001. A structural decomposition of energy consumption based on a
hybrid rectangular input-output framework: Japan’s case. Econ. Syst. Res. 13, 339-363.
Lan, J., Malik, A., Lenzen, M., McBain, D., Kanemoto, K. 2016. A structural decomposition
analysis of global energy footprints. Appl. Energy 163, 436-451.
Lenzen, M. 2001. A Generalized Input-Output Multiplier Calculus for Australia. Econ. Syst.
Res. 13, 65-92.
Llop, M. 2007. Economic structure and pollution intensity within the environmental input-output
framework. Energy Policy 35, 3410-3417.
Llop, M., Pié, L. 2008. Input-output analysis of alternative policies implemented on the energy
activities: An application for Catalonia. Energy Policy 36, 1642-1648.
Manresa, A., Sancho, F. 2004. Energy intensities and CO2 emissions in Catalonia: a SAM
analysis. J. Environ. Work. Employ. 1 (1), 91-106.
Miller, R. E., Blair, P. D. 2009. Input-Output Analysis: Foundations and Extensions, Cambridge
University Press, New York.
Proops, J. L. R. 1988. Energy intensities, input-output analysis and economic development. In:
Ciaschini, M. (ed.), Input-Output Analysis: Current Developments. Chapman and Hall,
London, 201-215.
Rose, A., Casler, S. 1996. Input-output structural decomposition analysis: A critical appraisal.
Econ. Syst. Res. 8, 33-62.
Su, B., Ang, B. W. 2012. Structural decomposition analysis applied to energy and emissions:
Some methodological developments. Energy Econ. 34, 177-188.

20

Wier, M. 1998. Sources of changes in emissions from energy: A structural decomposition
analysis. Econ. Syst. Res. 10, 99-112.

Table 1. Sectors of Production
1

Agriculture

15 Automobiles and transport material

2

Livestock

16 Other industries

3

Fishing

17 Construction

4

Extraction of minerals

18 Commerce

5

Coke, petroleum and fuel

19 Railway transport

6

Electric energy

20 Land transport

7

Extraction and distribution of gas

21 Maritime transport

8

Distribution of water

22 Air transport

9

Food production

23 Services linked to transport activities

10 Textiles

24 Finance

11 Manufactures of wood

25 Education

12 Paper

26 Medical assistance and social services

13 Chemistry

27 Public administration

14 Electric equipment and machinery 28

21

Other private services

Table 2. Indicators of the Energy Sector, Catalonia. Million Euros
4. Extraction of
5. Coke,
6. Electricity
7. Gas
8. Water
minerals
petroleum, fuel
2001
2011
2001
2011
2001
2011
2001
2011
2001
2011
Intermediate
inputs

379.2

270.4

3,663.9

Value added

286.8

332.2

548.2

Intermediate
outputs
Private
consumption

5,831.3 6,419.0

4,295.2 3,096.0 2,975.5 1,172.6 2,675.3
322.9

1,942.1 3,143.1

2011

921.4

8,780.8

11,137.8

607.0

781.5

4,159.3

4,723.0

3,724.0 5,205.2 6,990.4 1,348.9 2,289.4

505.5

1,047.8

2,909.8 1,924.6 1,744.1

18,812.4 20,470.6

413.9

724.4

627.5

547.6

5,213.1

5,929.5

0.0

0.0

0.0

0.0

77.1

0.0

77.1

0.0

2.6

0.0

-0.8

0.0

-51.5

-2.2

41.3

583.0

0.0

0.0

27.2

5,127.7

2,762.9

999.6

724.4

626.7

651.9

10,289.3

8,767.3

3.6

2,239.3

0.0

0.0

0.0

0.0

0.0

Investment

-11.8

5.0

-41.6

-7.2

0.0

Exports

198.2

237.0

2,514.8

2,457.4 1,831.7

Total demand

194.3

245.6

4,712.5

5,360.0 3,756.2 1,785.4

Total output

143.3

2001

469.1

5,921.5

7.8

Public consumption

775.2

Total Energy

6,025.6 6,664.5 10,634.0 9,084.0 8,961.4 8,775.8 2,348.5 3,013.9 1,132.2

22

1,699.7

29,101.7 29,237.9

Table 3. Changes in Energy Output, Catalonia (2001-2011). Million Euros
Technological
Structural
Total
Changes
Changes
Changes
4. Extraction of minerals
361.56
277.40
638.96
5. Coke, petroleum and fuel
-2,474.82
924.83
-1,549.99
6. Electricity
1,208.14
-1,393.77
-185.63
7. Gas
768.42
-103.04
665.37
8. Water
455.33
112.12
567.45
Total Energy
318.63
-182.47
136.16

23

4. Extraction of minerals

Table 4. Technological Changes in Energy Output, Catalonia (2001-2011)
Column Changes
Row Changes
Total Changes
Production
Production
Production
Coefficients
Coefficients
Coefficientsa
(Million Euros)
(Million Euros)
(Million Euros)
-0.02369
-241.61
0.33070
727.97
0.005482
361.56
(-36.89%)
(-0.83%)
(44.33%)
(2.50%)
(40.32%)
(1.24%)

5. Coke, petroleum and fuel

0.12436
(35.69%)

1,134.84
(3.90%)

-0.24124
(-38.63%)

-3,518.50
(-12.09%)

-0.002087
(-15.35%)

-2,474.82
(-8.50%)

6. Electricity

-0.01003
(-2.87%)

-191.14
(-0.66%)

0.10429
(25.68%)

1,563.46
(5.37%)

0.001683
(12.38%)

1,208.14
(4.15%)

7. Gas

0.38279
(75.83%)

1,258.45
(4.32%)

0.05769
(51.85%)

1,561.92
(5.36%)

0.007866
(57.86%)

768.42
(2.64%)

8. Water

0.12457
(29.84%)

265.50
(0.91%)

0.09711
(157.91%)

756.54
(2.60%)

0.003959
(29.12%)

455.33
(1.56%)

a Average individual changes.

24

Table 5. Structural Changes in Energy Output, Catalonia (2001-2011)
Private Consumption
Production
Value
(Million
Euros)

Public Consumption
Production
Value
(Million
Euros)

Investment
Production
Value
(Million
Euros)

Exports
Production
Value
(Million
Euros)

Total Demand
Production
Value
(Million
Euros)

4. Extraction of
minerals

-4.18
(-53.74%)

-4.54
(-0.02%)

0.00
(0.00%)

0.00
(0.00%)

16.76
(142.53%)

18.18
(0.06%)

38.76
(19,55%)

42.06
(0.14%)

51.34
(26.43%)

277.40
(0.95%)

5. Coke, petroleum and
fuel

670.46
(29.94%)

1,024.44
(3.52%)

0.00
(0.00%)

0.00
(0.00%)

34.38
(82.68%)

52.53
(0.18%)

-57.36
(-2.28%)

-87.64
(-0.30%)

647.49
(13.74%)

924.83
(3.18%)

6. Electricity

-180.45
(-9.38%)

-273.62
(-0.94%)

0.00
(0.00%)

0.00
(0.00%)

0.00
(0.00%)

0.00
(0.00%)

-1,790.35
(-97.75%)

-2,714.69
(-9.33%)

-1,970.81
(-52.47%)

-1,393.77
(-4.79%)

7. Gas

310.45
(74.99%)

596.99
(2.05%)

0.00
(0.00%)

0.00
(0.00%)

-2,61
(-100.00%)

-5.02
(-0.02%)

-583.05
(-100.00%)

-1,121.19
(-3.85%)

-275.21
(-27.53%)

-103.04
(-0.35%)

8. Water

-79.87
(-12.73%)

-144.64
(-0.50%)

77.10
(100%)

139.63
(0.48%)

0.77
(100.00%)

1.39
(0.00%)

27.20
(100.00%)

49.26
(0.17%)

25.19
(4.02%)

112.12
(0.38%)

Total Energy

716.41
(13.74%)

1,198.63
(4.12%)

77.10
(100.00%)

139.63
(0.48%)

49.3
(95.73%)

67.08
(0.23%)

-2,364.8
(-46.12%)

-3,832.20
(-13.17%)

-1,522.0
(-14.79%)

-182.47
(-0.63%)

25