0
STRUCTURAL
DECOMPOSITION OF
RENEWABLE ENERGY USE:
1995-2009
Master Thesis
Filippo Capurro
S2964953
filippocapurromu@gmail.com
Supervisor: prof. dr. H.W.A. E. Dietzenbacher
Co-assessor: prof. dr. mr. C.J. Jepma
Faculty of Economics and Business
University of Groningen
1
STRUCTURAL DECOMPOSITION
OF RENEWABLE ENERGY USE:
1995-2009
ABSTRACT
Anthropogenic activity is on the break of causing irreversible and harmful changes in planet’s climate. As acknowledged during the Paris Climate Change Conference in 2015, global warming is also, if not primarily, an energy problem. Thus renewable sources of energy, as an alternative for fossil fuel combustion, represent an opportunity towards the pursuit of a more sustainable socio-economic development. This study aims to investigate the global renewable energy (RE) use evolution between 1995 and 2009. In a globalized and extremely interconnected world it is interesting, also for policy purposes, to identify the main drivers behind RE use trend and its geographical distribution. To this extent, through a structural decomposition analysis, it was possible to isolate the contributions of six key determinants. Global RE use rose from 52.3 Exa Joules (EJ) in 1995, to 68.7EJ in 2009. This 16.4EJ was the net result of overall energy efficiency improvements (-18.4EJ), consumption per capita (+26.3EJ) and population (+10.3EJ) changes, plus the marginal net contribution (+1.6EJ) of the three remaining factors: shifts towards RE, commodity mix and trade structure. The BRIIC’s and EU’s national RE use changes rose global RE use by 5.5EJ and 3.0EJ respectively. US (+0.7EJ) and EU (+0.26EJ) were the main net importers of RE use, while the biggest net exporter (+1.2EJ) was the aggregate Rest of the World (RoW), containing mainly developing countries, and its contribution to global RE use change was +7.0EJ. Although the impact of trade on global RE use was negligible, it played an important role in terms of delocalization of RE use across countries.
2
Table of Contents
ABSTRACT ... 1
I. INRODUCTION ... 3
II. LITERATURE REVIEW ... 8
III. DATA AND METHODOLOGY ... 13
3.1 Input-Output analysis ... 13
3.2 Multiregional Input-Output framework ... 15
3.3 Structural decomposition analysis ... 16
3.4 Deflation issue ... 23
3.5 Data ... 24
IV. RESULTS ... 27
4.1 Global trends of renewable energy use ... 27
4.2 Drivers of renewable energy use at global level ... 29
4.3 Determinants by country ... 31
4.4 Countries’ footprints and trade balance ... 33
4.5 Discussion ... 34
V. CONCLUSIONS... 40
REFERENCES... 43
3
I. INRODUCTION
In the last few decades environment has been one of the striking concerns and discussion cornerstones among countries all over the world: from the early United Nations Framework Convention on Climate Change (UNFCCC) signed in Rio de Janeiro (1992), passing through the 3rd Conference of Parties (COP) which led to the famous Kyoto Protocol (1997), to the very last United Nations Climate Change Conference (2015) held in Paris, interposed by many other international summits1. Such urgency and commitment stemmed from the undergoing drastic
climate changes, indicating an environment under pressure due to the increasing and intensive anthropogenic activity.
These global summits were primarily focused on global warming and coped with problems such as the reduction or annulment of anthropogenic greenhouse-gas (GHG) emissions, seen as the main culprits of climate change. Nevertheless, if compared to the other environmental issues, global warming has the peculiarity of being primarily an energy problem. The latest 2015 report by International Energy Agency (IEA) and the 2015 European Commission report highlight that the energy production and use are responsible for approximately two-thirds of the world’s GHG emissions. The relation between GHG emissions and energy consumption stems from the fact that the world's energy is largely supplied through the combustion of fossil fuels2. This process releases CO2 in the atmosphere which in
turn is the principal component of the GHGs causing global warming. Thus energy use and climate change are two sides of the same coin. In EDGAR 4.23 database the
1 The United Nations Framework Convention on Climate Change (UNFCCC) is a framework
designed to rule how international agreements, also called protocols, should be negotiated among the parties, how should be set the binding limits on emissions, hence the UNFCCC itself does not contain neither enforcement mechanism neither emission targets. From the 1995, within the UNFCCC system, member states meet in yearly events, the COP, to discuss environmental issues and assess the improvements in dealing with climate change. The 1st United Nation Climate Change Conference was held in Berlin during the 1995. The Kyoto Protocol is the first agreement that contains legal obligations for participants.
4
total amount of CO2 emissions in 2014 was 35.7Gt4. In the same year around
29.7Gt of CO2 emissions were only due to fossil fuel combustion which, according
to Renewables 2015 Global Status Report projections (REN21), still has a share of 78% of global final energy consumption in 2014. The remaining 22% is divided between renewable (19%) and nuclear energy (3%). These figures are consistent also with the Tracking Clean Energy Progress Report (2015) by IEA. Although nuclear energy is a non-fossil fuel source and actually for its characteristics it is possible to label it as a renewable energy (from now on RE), both in REN21 and IEA reports it is not included in the renewable category. The main reason is that nuclear energy exerts a striking environment pressure as well, through e.g. radioactive waste.
One of the possible answer to global warming, widely discussed in the global conventions, outlined also in the Fifth Assessment Report (2014) of the Intergovernmental Panel on Climate Change (IPCC), is the implementation of country-specific policies promoting the shift from non-renewable to renewable sources of energy. It should be specified that RE is neither 100% GHG free nor completely free from any other type of environmental stress. It is, however, unquestionable that energy coming from such sources is much less GHGs intensive, which justifies the political effort in speeding up the transition from fossil fuel to renewable sources, as pointed in Tilman et al. (2009). This is crucial if the pledges and commitments, made in Paris few months ago – stating in the Article 2 the goal of “Holding the increase in the global average temperature to well below 2 °C above pre-industrial levels and to pursue efforts to limit the temperature increase to 1.5 °C above pre-industrial levels, recognizing that this would significantly reduce the risks and impacts of climate change” – are willing to have a real impact on the planet.
The measures aiming to cope with or prevent climate change or any other environmental issues may collide with other development goals, pursued by the same actors that display an environmental vow. An emblematic example would be
4
5
the Millennium Development Goals (MDGs), contained in the United Nations Millennium Declaration. Goal 1 refers to the eradication of extreme poverty while Goal 7 highlight the need of environmental sustainable development. These two scopes appear to be, at least partially, contradicting, in fact improving life standards or rising the income per capita in developing countries is likely to increase also the energy use, foster population growth, thus magnifying even more the energy demand of those regions and their environmental pressure. To what extent this rise in energy demand, even if completely satisfied by RE, would not boost global emissions and consequently frustrate the effort of stabilizing temperature increase below 2°, remains unclear. However it challenges the apocalyptic threshold of 2°C, outlined by the IPCC report (2014), that in case of not being achieved before 2050 will probably lead to severe, pervasive and irreversible impacts for people and ecosystems, and long-lasting changes in all components of the climate system.
From an economic perspective the last decades have been also very eventful. The globalization wave has eased world integration and this is reflected in the sharp trade volume growth5 and in the production fragmentation phenomena,
both very interconnected. Concurrently technological innovations are frequent, but, more strikingly, they spread worldwide at a very fast pace causing important structural changes in terms e.g. of production efficiency.
Within such context a thoroughly analysis of global RE use seems appropriate. This study aims to shed light on global RE use changes and the main drivers behind such evolutions. It will be quantified the countries’ territorial contribution to global RE use change (producer responsibility) and how much global RE use is embodied in their final demand (consumer responsibility). This will make possible to unveil nations’ energy trade balance, hence the geographical distribution RE use. Finally the quantitative analysis will enable not only the
5 According to Arto and Dietzenbacher (2014), between 1995 and 2008, world trade tripled from
6
evaluation of past policies and structural changes, but it will also allow to mark the path for future interventions.
By using the World Input-Output Database (WIOD) and a structural decomposition analysis (SDA), the global6 RE use change during the period
1995-2009 will be uncovered. SDA allows the break-down of a variable, RE use in this case, in a multiplicity of underling determinants. Six main RE use drivers were identified: population growth, trade structure, consumption per capita, composition of final demand, technological changes and shifts towards RE. The Input-Output (IO) analysis will also allow to compute countries’ footprints and disclose the geographical distribution of RE use. The last two aspects are the key in determining each country’s responsibility, necessary for reaching equitable international agreements in the future given that a nation’s energy footprint is all the energy globally needed for the final demand of such country.
To the best of my knowledge, no such a comprehensive study on RE use and its determinants has been conducted. Although the literature on IO and structural decomposition analysis applied to environmental issues is vast, the research has mainly focused on other parallel topics. There are many papers on GHG emissions that either run a SDA at global level as in Arto and Dietzenbacher (2014); or compute emission responsibility for one country (Spain) as in Serrano & Dietzenbacher (2010). Steen-Olsen et al. (2012) conducted a wide study on land, water and carbon footprints and displacements, yet only at European level. Dietzenbacher et al. (2012) investigated carbon emissions in China, with a specific focus on trade, consumer and producer responsibility. The constituents of the US CO2 emissions, between 1997 and 2013, were explored by Feng et al. (2015).
Although all these studies are to a certain extent similar to the analysis that will be carried out in this work, they do not tackle energy use.
6 In the database the world is represented by 41 regions: the 27 EU members states (except for
7
There are papers addressing energy use, for example Lin and Polenske (1995) decomposed Chinese energy use in the 80s exploiting the availability of IO tables and the SDA, their focus was at sectorial level as well. Lim et al. (2008) conducted a SDA of CO2 emissions due to energy use in South Korea during the
1990-2003 period. Whilst Wachsmann et al. (2008) paper shed light on Brazil energy use and its drivers from 1970 to 1996, however they used a different structural decomposition method. In another paper Alcàntara and Duarte (2004), by means of a SDA, investigated the energy intensity in European member states and sectors. US energy use, between 1997 and 2002, was decomposed with both IDA and SDA methods in Weber (2009). Many of these studies are limited in terms period considered, almost all of them focus only on one country and foremost no one dedicates specific attention to RE use.
8
II. LITERATURE REVIEW
The decomposition analysis that will be conducted in the next section is based on the Input-Output model, theorized in Leontief (1953). Depending on the data availability, IO tables may have numerous uses. The fundamental information, the National Input-Output (NIO) tables contains, is: the inter-industry transactions, the industry output and the industry sales to specific final markets7. Such data
allows to quantify, e.g., how much agriculture production is required in manufacturing production, how much capital is embodied (directly and indirectly) in the national final demand of services. NIO tables are also meaningful from a policy perspective, in fact they can help developing projections on the effects of government spending on a specific sector.
By using Multiregional Input-Output tables (MRIO) similar points can be explored, but from a global perspective. In addition to transactions within industries also inter-country trade is captured. Hence it is possible to uncover complex relationships such as how much manufacturing production in country is embodied in country final demand for country service industry. As for production fragmentation and trade relationship Weidmann et al. (2007) underlined that inter-industry demand of inputs is satisfied by both domestic and foreign markets. At the same time the produce of an industry may be just an intermediate embodied in the production of an international supply chain. This because in such a globalized world, countries import from a multiplicity of nations, each one with different production technologies, and concurrently they export goods to satisfy foreign consumption: all these facets cannot be properly captured through NIO tables.
Moreover IO analysis can be extended also to the assessment of environmental impacts, as Leontief showed in another of his seminal papers
7 Final demand typical categories are: export, public investment and consumption, private
9
(1970). By combining NIO and MRIO tables with environmental satellites, energy accounts in this study, it is feasible to uncover different issues. At national level it is possible to answer not only which industry is more energy intensive and how much energy it uses, but also how much energy use in an industry is due to production in another industry or embodied in the final demand for another product.
At a global level, questions such as how much energy is produced worldwide can be answered. On the other hand it becomes feasible to track geographical distribution of energy use, highlight “trade” patters, compute energy footprints and determine how much energy use in a certain country is due to final demand by another state. Steem-Olsen et al. (2012) paper is an emblematic example of comprehensive footprint and displacement assessment within the EU.
Footprints are indicators based on the consumer responsibility approach that account for the total direct and indirect effects of consumption activity. In other words a country’s energy footprint indicates all the energy globally embodied in the final demand of such economy. Instead by producer responsibility it is meant how much energy is involved in national production (irrespective of where final consumption happens), hence territorial or national energy use. From this perspective country’s energy use is the sum of all energy embodied in local production destined to both local and foreign consumers. While Lenzen et al. work (2007) sheds light on these two ways of assessing country’s environmental pressure, in Serrano and Dietzenbacher (2010) paper is demonstrated that evaluating a country’s emission international responsibility with the trade emission balance (emissions embodied in exports minus emissions embodied in imports) and the responsibility emission balance (territorial emissions minus emissions footprint) is the same.
10
country, on intermediate deliveries across its industries and the foreign ones, on industries total output, final consumption data and eventually environmental accounts. In this sense WIOD is a very complete database.
As mentioned in the introduction, the world economy is becoming more and more complex, trade is increasing at vertiginous rates, production fragmentation has revolutionized the traditional way of thinking about sectorial clusters. Globalization has brought economic development and consumption levels to peaks never seen before albeit unevenly distributed. Innovative technology is often introduced and it spillovers very quickly due to the worldwide progressive interconnectedness. On top of that, population is growing and new economies are emerging. In such chaotic context, determining and quantifying what drives a certain phenomenon has important implications also from a policy perspective.
A tool widely used by researchers to cope with such need is the Structural Decomposition Analysis (SDA). It is a recognized methodology to quantify the relative impact of constituents that determine together the actual change in a certain variable of interest. By applying a simple version of SDA to energy use, it is possible e.g. to pinpoint three main determinants of energy use variation: changes in energy use coefficients, technology and final demand changes. Lin et al. (1995) conducted this type of decomposition analysis for energy use in China. Arto and Dietzenbacher (2014) focused on global emissions change and decomposed it in five drivers: technological, final demand per capita, commodity mix, population and trade structure changes. Lim et al. (2008) conducted a national SDA of CO2
emissions in Korea, while Wachsmann et al. (2008) did it for Brazilian energy use, both papers split the variable of interest in eight drivers.
11
are explored, this will lead to 720 equivalent exact8 decomposition equations of
which the average9 should be taken. More generally, if there are determinants,
equivalent exact decomposition equations can be derived. One way to bypass such cumbersome computations is to use two polar equations as demonstrated by Dietzenbacher and Los (1998), from now on D&L. The authors agree that such approximation is a reasonable compromise between computation complexity and loss of accuracy10. For these reasons the D&L method is adopted and thoroughly
developed in the methodology section.
Another approach capable of digging into multiple aspects of an economic phenomena is the index decomposition analysis (IDA). Although it is not the aim of this work to explain the differences between these two methodologies, a briefing on their pros and cons is needed in order to justify why, within this study, it was opted for a SDA. As underlined in Hoekstra and van der Bergh (2003) one first advantage of IDA is its lower data requirement. Nevertheless lower data requirement constitutes also a limitation because it implies that IDA is capable of less detailed decompositions compared to SDA11. Furthermore SDA can isolate a
wider range of technological and final demand effects. In spite of such differences and different theoretical background the two approaches lead to qualitatively similar conclusions as emerges in Weber (2009), where the U.S. energy use change, between 1997 and 2002, is decomposed with both strategies12.
A final concern that needs to be addressed is the deflation procedure. Generally speaking IO tables are available in money terms, specifically in current prices. This means that variation of prices is likely to affect the numeric results of IO analysis and SDA. Therefore to avoid this price distortion it is necessary to apply
8 It is a decomposition for which the sum of the changes from each driver equal the overall change
in the disentangled variable.
9 Taking the average of each decomposition means giving the same weight to each contribution. 10 D&L conclude, after empirical analysis that the discrepancy of sectorial effect estimates, given in
terms of the ratio of the average of two polar decomposition to that of full decomposition, ranges from 0.9% to 7.7%.
11 This is true because SDA exploits the IO framework whereas IDA uses aggregate sectorial data. 12 Hoekstra and van der Bergh (2003) suggested as well that the two methodologies can to a certain
12
13
III. DATA AND METHODOLOGY
3.1 Input-Output analysis
Before specifically presenting the decomposition analysis, a brief overview of IO analysis is necessary. To become familiar with IO tables, a stylized NIO table in money terms and a satellite account13 are presented. Together they summarize the
most important elements that are going to be used in the study.
Industries Final Demands Total
Industries
Primary inputs
Total
Energy Industries
Renewable
Given that is the number of industries in the economy, then is a matrix displaying the intermediate deliveries across industries, the element gives the money value of the deliveries from industry to industry ; the subscripts and indicate the row and column coordinates respectively. is the final demand vector14. The element denotes the money value of the delivery from
industry to final demand. is the primary input matrix, where refers to the primary input taxonomies (i.e. wages, indirect taxes, imports, etc.), the element
represents how much primary input category is used in industry . Similarly
13 Satellite accounts can include data on emission, employment, water and land as well. Moreover
each dimension can be divided in subgroups depending on the information contained in the database. Within WIOD, as far as energy, it is possible to distinguish between renewable and nonrenewable energy and eventually reach a further level of categorization (coal, gas, nuclear energy, hydro, wind, geothermal and solar energy, etc).
14 Note that in WIOD’s NIO tables the final demand is a matrix where represents the
14
is the column vector (whereas is the transposed vector) of gross industry output; thus identifies the total output in industry .
Regarding the satellite accounts, they are accounts that contain additional data on multiple issues, within which the environmental and social ones are the most commonly used. The former display information, e.g., on emissions, water consumption and land use, whereas the latter provide information on employment, educational attainment and capital stocks amongst other things. The vector composed by elements, in this study, gives the RE use in each industry. The advantage of satellite accounts, combined with NIO or MRIO tables, is allowing to quantify energy use footprints, national or industrial energy intensity and energy use “trade balance”.
Another important item for the decomposition analysis is the matrix , called the input coefficient matrix or the matrix of domestic technical coefficients; its element tells how much production in industry is required for an additional unit of output in industry . Matrix is obtained:
(1)
Where is the inverse diagonal matrix of output vector .
Consequently following the seminal work of Leontief (1953) the total output of an economy can be expressed as:
(2)
Then identity (2) can be written as:
(3)
Where is the identity matrix and is known to be the Leontief
inverse matrix of dimensions, from now on indicated as . The elements
15
In a similar fashion, in order to proceed in the analysis, also energy coefficients must be derived. These coefficients are obtained by dividing the energy vector with the output vector . Subsequently, the element of the energy coefficient vector displays how much energy is involved in one unit of output of industry . In algebra notation:
(4)
3.2 Multiregional Input-Output framework
In terms of structure, MRIO tables are similar to NIO tables: the main difference comes from the interpretation of its elements. Note that the number of industries is still while the number of countries is indicated by . Therefore the element of the intermediate deliveries matrix represents the money
value of how much inputs from industry in country are involved in the production of industry in country . Hence the component 15 in the
final demand matrix gives the delivery in money terms from industry in country to country final demand. The primary input matrix , in this case, contains the constituent which expresses how much primary input from
country is used in industry in country . Similarly the element belonging to the output vector shows the total output for industry in country . Now the component in the k matrix of technical coefficients indicates how
much production from industry , located in country , is needed for an additional unit of output in country industry . Finally from the Leontief inverse
matrix displays how much output in industry country is required for an extra unit of final demand of industry in country . When considering energy use satellite the element from renewable energy vector identifies the amount of energy use by industry in country . Thus, in a similar fashion, the component in vector shows the amount of energy use per unit of output in country industry .
16
3.3 Structural decomposition analysis
As mentioned before, in order to investigate the global and regional trends of RE use and to measure the contributions of each one of the six factors, the SDA analytical tool will be used.
To perform a decomposition of global energy use, two more vectors are needed apart from the elements presented until now. These vectors will enable the assessment of how population and consumption changes impact national and regional energy use. They are: and , the total direct energy use by household and the population size vectors, respectively16. Hence the element would
be the total direct energy use by household in country , while is the population in country .
It is immediate that the vector of country’s total energy use is given by the sum of nation’s household consumption and sectorial energy use. This is expressed formally as:
(5)
Note that in identity (5) is the vector of country’s total energy use in production, however its dimension is now . is obtained by (pre)multiplying the previous vector with a matrix with the following structure:
Where is a vector of ones.
Remembering the accounting equation (3) and (4), then it is possible to reformulate identity (5) in the following way:
(6)
17
being the energy coefficients matrix with the following structure:
While is the final demand vector resulting from:
With being a vector of ones. Note that the element of the vector of energy coefficients can be decomposed in two other factors: one capturing the efficiency in overall energy use (renewable and non-renewable), the other capturing the share of renewable energy use on overall energy use. Formally these elements can be expressed as follows:
Where element belongs to the output vector and shows the total output for industry in country . The component from the RE use vector identifies how much RE use happens in industry in country . represents the total energy use by industry in country , hence both RE and non-RE use.
Consequently the term captures how much energy (non-RE and RE) is involved
in a unit of output of industry in country , it is in other words a proxy for production efficiency in terms of energy use. On the other hand the term is able to unveil the potential shift from RE to non-RE use or vice versa, undergone in industry in country . In matrix notation the above identity would be:
(7)
18
Also the matrix single element can be decomposed further:
(8)
Where is the fraction of the total final demand in country for good coming
from country ; when this amount is imported from country , whereas if the fraction of final demand is produced domestically, thus in country ; is the population in country ; is the share of commodity in country final demand; and is the final demand per capita in country . In matrix notation:
(9)
Where is a stacked matrix structured in the following way:
is the diagonal matrix of final demand per capita vector and is the population vector .
Likewise the household energy use element , from vector , can be expressed in an alternative way such that:
(10)
Where is the total direct energy use (RE and non-RE) by households. Hence
19
RE to non-RE. Instead reflects the total direct energy use by household per unit of final demand in country . Whilst is the population in country and is the final demand per capita in country . In matrix notation equation (10) can be written:
(11)
Here and are the diagonal matrices of the vectors and
respectively.
Now, substituting the expressions (7), (9) and (11) in equation (6) gives the following identity:
(12)
Since in this study the D&L method is adopted, the SDA of the variable of interest is conducted by taking the average of two polar equations. The analysis target is to assess RE use trends from 1995 to 2009, thus it is necessary to calculate the variations along this time span of both RE use and its constituents. The formal steps to obtain the expressions capable of computing the changes between two random points in time, initial time 0 and final time 1, are explained below17.
Starting from the variation of RE use between period 0 and period 1:
(13)
The first polar decomposition (indicated by the subscript ) is obtained by substituting expression (12) in equation (13), which gives:
17 Note that the implementation of this procedure to a broader time range it is only a matter of
20 (14)
In similar fashion the second polar decomposition (indicated by the subscript ) can be derived, therefore obtaining:
(15)
Then it is necessary to take the average of the two polar decompositions:
(16)
The next step consists in decomposing the Leontief inverse matrix. It would be of particular interest breaking the input coefficient effect in two additional contributions: variations in technology coefficients and changes in trade coefficients. This can be formalized, starting from the third term of equations (14) and (15), in the following way:
21
In order to isolate the two effects attributable to: changes in the production technology and variations in the intermediate coefficients matrix due to trade it is necessary to define the technology coefficients matrix and the trade coefficients matrix . The elements of the matrix are given by:
The elements of the matrix are obtained by divinding the input coefficient elements with
element:
With being the share of each intermediate good bought by sector in country
that is produced domestically if or imported from country if . Consequently, given that , can be expressed as follows:
(18)
22 (23) (24) (25) (26) (27) (28) (29) (30)
In the introduction it was advocated the split of global RE use change in six determinants: the technological change (or overall energy efficiency change); the trade structure change; the variation in the commodity mix of the final demand; the change in final demand per capita; the population growth; and finally the shift from non-RE to RE use. The first contribution is captured by terms number (19), (21) and (27); the second by (22) and (23); the third by (24); the fourth by (25) and (29); the fifth by terms number (26) and (30); and finally the last one by (20) and (28).
23
It is worth noting that the RE use vector can be expressed also as a matrix . In this case equation (12) should be rewritten in the following way:
(31)
Where is the diagonal matrix obtained from the population vector .
The element of matrix shows how much RE use in country is embodied in
country final demand. Similarly by substituting with in equations (19) to (30) the matrix change is obtained instead of the vector change .
3.4 Deflation issue
24
MRIO tables for each year18 in current prices ( ) and previous year prices .
Following Arto and Dietzenbacher (2014) strategy, the deflation procedure is reduced to a chain in which energy use change between two years and is computed using data in CP for the former and PYP for the latter19. So that the
change is expressed in constant prices and therefore the price effect is eliminated. Just for explanatory reasons, a brief sketch of the deflation chain is presented. Assuming that is a factor affecting the energy use change, year by year; in order to compute between time and , the data in and the data in are needed. In formulas this can be expressed as:
Similarly for time and the deflation is accounted as follows:
Therefore the deflation chain for the time period to becomes:
After this deflation process all the yearly variations are in constant prices and can be summed in order to unveil the long run RE use change. In the same fashion the deflation chain can be extended to the whole time period considered in this study 1995-2009 and to all the factors included in the final decomposition expression (19-30).
3.5 Data
The analysis was conducted by means of the World Input-Output Database20
(WIOD). WIOD was selected for three main reasons: first of all because it is an open source database; secondly because it provides data on a wide number of
18 For 1995 there is only the MRIO table in current prices.
19 In expressions (19-30), the subscripts and refer to elements expressed in and in
respectively.
25
economic and environmental indicators. The environmental accounts, which include energy use, are of specific relevance within this study. Finally because it contains MRIO tables in current and previous year prices, a unique and essential characteristic for the deflation method selected.
WIOD covers data on 40 countries plus an aggregate called Rest of the World (RoW)21, 35 industries and 59 products. Besides the database provides National
and World Input-Output (WIO) tables22 , both types are available from the
1995-2011 in current US dollars. World IO tables in previous year prices are accessible, but they cover only the 1996-2009 period; environmental accounts instead contain data from 1995 to 2009. For further details on WIOD Database it is possible to consult Timmer et al. (2015); and for satellite accounts, Genty et al (2012); all documents downloadable at WIOD’s webpage.
The gross energy use accounts encompass data on energy use in TJ23 by
country, industry and energy commodity. In these satellite accounts, 26 different energy sources are recorded, nevertheless within this study the attention is on RE sources only. According to IEA reports (2015) and the REN21 report (2015), the energy commodity taxonomies, included in WIOD, that can be considered coming from renewable sources are: biogas, biogasoline, biodiesel, geothermal heat, waste combustion, solar, wind and hydroelectric energy24. The database comprises also a
category named “other renewables”. WIOD energy taxonomies are more aggregated than the IEA energy classification, Tables 3.5a and 3.5b in the appendix give a correspondence overview.
For sake of simplicity, the results concerning the country level, are presented with regional aggregates. EU2725 refers to the European member states; Brazil,
21 It is worth noting that this aggregate contains mainly developing countries.
22 National IO tables are , whereas world IO tables are . Basically countries
and industries.
23 Tera Joules. .
24 Energy use categories will be abbreviated in the following way: biogas (biogas), biogasoline
(biogasol), biodiesel (biodiesel), geothermal heat (geo), waste combustion (waste), solar (Sol), wind (wind) and hydroelectric energy (hydro).
26
Indonesia, India, Russia and China can be considered both singularly or as the aggregate BRIIC; the East Asia label includes Japan, South Korea and Taiwan; RoW represents the rest of the world, as already mentioned. US, Canada, Turkey, Australia and Mexico complete the sample26.
Population data, needed for computing the population and final demand per capita determinants, was gathered from World Bank27 (WB) database. Since
population time series on Taiwan were not available in WB dataset, they were retrieved directly from the Taiwanese National Statistic Agency28.
26 Countries’ names will be abbreviated in the following way: AUS (Australia), BRA (Brazil), CAN
(Canada), China (CHN), Indonesia (IDN), India (IND), Mexico (MEX), Russia (RUS), TUR (Turkey), United States (US), European Union (EU27), rest of the world (RoW) and East Asia (East Asia).
27
IV. RESULTS
4.1 Global trends of renewable energy use
Between 1995 and 2009 the use of RE increased by 16.5 EJ29, from 52.3EJ to
68.7EJ. Although the average growth percentage rate in the time span considered is almost 2% per year, between 2000 and 2001 there was a negative growth and global RE use decreased by 0.2%. In 1995 the share of RE use attributable to households was 56.2%, while in 2009 it dropped to 51.9%; the RE use share due to production, on the contrary increased from 43.8% to 48.1%. All these figures are included in Table 4.1a (green part). If we break the period in two segments, 1995-2002 and 2003-2009, some features emerge: total, household and production RE use all grew on average faster in the second period. This was particularly marked for RE use due to production, for which the average growth pace of the first period (1.5%) was less than half of the second one (3.9%). Finally in 2008 the growth rates of both household and production RE use slowed, probably because of the global financial crisis.
When focusing on RE source a higher grade of fluctuations emerge. Overall, each energy type30 has seen its use augmented as shown in Table 4.1a (red part).
Nevertheless biodiesel, solar and wind energy displayed average growth rates in the period of 42.4%, 23.2% and 31.9% respectively. Biodiesel use rocketed from a 4.7PJ in 1995 to 398.8PJ in 2009. Although solar energy use had a less impressive average growth rate, its use in production process moved from 23.6PJ to 430PJ during the period analyzed. Wind energy use level in 1995 was around 23PJ and in 2009 it reached 983.2PJ. Nevertheless, despite this amazing upsurge, hydroelectric and other renewables energy use remained dominant along the fourteen years.
29 EJ refers to Exa Joules, joule being the international unit of measure of energy, and Exa being 1018,
it may be used as well PJ (Peta Joules 1015) and TJ (Tera Joules 1012).
30 Biodiesel, biogas, biogasoline, geothermal, hydro, solar, waste and wind energy and other
28
Table 4.1a: Overview of global RE use change, taxonomies, quantities, growth rates and shares (1995-2009)
Notes: HH is household consumption; RE_P is renewable energy use due to production; TOT_RE is total renewable energy use; TOT_E is total energy use; 09-95_c is the period change; %c is percentage change; %c_a is percentage change average; Tre_share is the share in total renewable energy use; T_share is the share in total energy use; REP_share is the share in renewable energy use due to production; period_a is the percentage change average of 2002-1995 and 2009-2003 respectively.
29
In 1995 the share of the former in global renewable energy use was 39.1%, whereas the share of the latter was 48.5%. Both shares decreased, but remained around their initial levels, indeed in 2009 they were 35.4% and 43.3% respectively (% of renewable energy use only due to production). The third most used energy source is the geothermal one, whose average global share has been only 7.4%31.
Such figures are qualitatively similar to the RE category shares pinpointed in REN21 (2015)32, besides from the same report it is understandable why there was
found a boom in solar and wind energy use. In fact only between 2004-2009 on average 35.5 and 50.8 Billion USD per year were allocated to new investments respectively in solar and wind technologies, which may reflect to a certain extent the political and policy attention given in the last decades to ecological issues.
In terms of household consumption, three main features emerge. Firstly, in the period considered, on average 97.3% of household energy use came from hydroelectric sources. Secondly, wind energy, according to WIOD, was not consumed at all by households, in any year. Finally, on one hand biodiesel household consumption began in 2000, while on the other hand energy coming from “other renewable resources” was not consumed anymore from 2002 onwards. For more detailed data on household consumption see Table 4.1b in the appendix.
4.2 Drivers of renewable energy use at global level
One of the main research questions of this work is to shed light on the drivers behind such RE use variations. To this end six key determinants are identified: technological changes (overall energy use efficiency), changes in trade structure, variations in the commodity mix of final demand, changes in consumption per
31 These results are consistent with REN21 (2015) report and in the Tracking Clean Energy
Progress Report (2015) by IEA.
32 They differ quantitatively because the aggregates are formed with partially different RE
30
capita, population growth and trade-off between non-RE and RE use. Figure 4.2a tracks the yearly cumulative change of the six constituents at global level.
Between 1995 and 2009 the technological factor caused a sharp reduction in RE use of around -18.4 EJ. This reduction in energy intensities is likely to reflect the effort of policy-makers in promoting more efficient means of production. The shifts towards non-RE use caused only a marginal drop (-2.4EJ) of global RE use. Despite the effect increased in absolute terms during the time span its evolution is quite steady. An interesting target from a policy perspective would be to achieve a positive value of this factor. This would mean that governments really found a way to substitute non-RE use with energy use coming from RE sources, hence promoting a more sustainable socio-economic growth path.
Figure 4.2a: Cumulative change by renewable energy use driver
Source: WIOD
31
of production, the more human beings are welcome on the planet the more energy will be used. This intuition is clearly reflected in the population and final consumption effects. The impact of commodity mix is almost negligible, in fact it reduced RE use only by -0.3EJ. Also changes in trade structure were almost irrelevant (+1.1EJ), which is quite surprising. As noted in Arto and Dietzenbacher (2014), between 1995 and 2008, world trade tripled from $6.3 trillion (21% of world GDP) to $19.5 trillion (32% of world GDP), however trade itself had a little impact on global RE use changes. The intuition behind the small effect is that at global level imports and exports cancel out. In other words exports substitute RE use that otherwise would have taken place abroad; and similarly imports substitute RE use that otherwise would have taken place locally. Trade in energy did happen as will emerge more clearly in section 4.5, but its more interesting implications have to do with the delocalization issue. In fact international trade is more a tool through which countries displace RE use to other locations than an important driver of RE use changes. The evidence on trade as a negligible determinant of global RE use in this analysis, coupled with the same conclusion found by Arto and Dietzenbacher (2014) for CO2 emissions suggest that all the
attention given by policy-makers and environmental organizations to trade as one of the main environmental pressure culprit seem exaggerated.
4.3 Determinants by country
32
variations due to global trade and commodity mix change seem again only marginal.
Population growth did affect the RE use levels, but still not as much as consumption per capita and technology. China’s territorial net effect represented 11.7% (+1.9EJ) of the global energy use change within the fifteen years, strikingly global changes in final demand per capita rose Chinese RE use by 69.5% (+11.4EJ). However this figure was cancelled out by a seemingly sharp global efficiency improvement of 55.2% (-9.1EJ). Overall RoW’s national net RE use change contributed to global RE use by 42.4% (+7.0EJ). Global changes in technology, population and consumption per capita led to changes in RoW territorial RE use equal to -26.2% (-4.3EJ), +31.5% (+5.2EJ) and +29.7% (+4.9EJ) respectively. Surprisingly RoW, which represents a huge portion of emerging countries, was the region that more has contributed to RE use growth. Worldwide population changes caused only a modest impact on RE use in China (+6%) and India (+9.3%).
Figure 4.3a: % change in global RE use by driving country and driving force
Source: WIOD
33
increase. Its contribution was marginally induced by global population growth (1.7%) and global final demand per capita effect in the “old continent” was low as well (6.9%). Net worldwide changes in the six determinants caused an increase of 3.2% in United States’ territorial RE use (0.5EJ). Interestingly EU27 and US are the only countries in which, along the period considered, global changes induced a shift from non-RE to RE use. Such evidence is likely to highlight the commitment of these economies to reduce their environmental pressure by promoting production processes relying more on RE, this may be true because in the global changes also the changes in such regions are included. The territorial RE use variations in China, Brazil, Indonesia and India contributed to global RE use change by 1.9EJ, 1.7EJ, 1.3EJ and 0.7EJ respectively.
4.4 Countries’ footprints and trade balance
Until now the focus has been on territorial energy use. Therefore those figures do not tell for how much RE use each nation is actually responsible for, in terms of consumer responsibility, and at the same time no specific delocalization pattern can be unveiled.
An economy’s energy use footprint is coincident with all the energy globally necessary to satisfy the final demand of such country. Thus, to compute a country RE footprint it is necessary to add together the RE use in national production destined to national consumption, imports (foreign production destined to national consumption) and household consumption. This approach follows the consumer responsibility as depicted in Dietzenbacher et al. (2012) and in Serrano and Dietzenbacher (2010).
34
(0.09EJ), East Asia (0.08) and China (0.003EJ). Whereas the biggest net importer is US with 0.7EJ, followed by India (0.28EJ), EU27 (0.26EJ) and Russia (0.15EJ).
As anticipated in section 4.3 this evidence unveils other interesting insights on trade role. Although globalization boosted international trade at levels never seen before, changes in the trade structure contributed only marginally to changes in global RE use. The main international trade effect has to do with the displacement of RE use across regions, thus causing a redistribution of RE use from producers to consumers. Therefore, uncovering the trade balance really sheds light on who produces RE for whom.
Figure 4.5a: Country’s territorial and footprint cumulative changes in RE use and trade balance
(1995-2009)
Source: WIOD
4.5 Discussion
35
China, as remarked in Arto and Dietzenbacher (2014), was the largest emitter of CO2 in 2009 and still is today, according to REN21 (2015). REN21 (2015)
report indicates that air pollution is a striking issue especially in Chinese metropolis. It is causing millions of premature deaths annually and more in general it has a negative impact on human health. Hence, for China, RE represents a strong answer to pollution and health related matters, but, at the same time, it also shields the country against energy supply risk. In fact its massive and fast economic growth in the last decades has increased also its dependency on imported fossil fuels. Chinese investments on renewable sources and fuels amounted to 0.3 Billion USD in 2004, reached 39.5 Billion USD in 2009 and rocketed to an impressive 83.3 Billion USD in 2014, as remarked in REN21 (2015). This may reflect China’s proactive investments environment towards RE sources. In the present study it was found that China was the biggest single country also in terms of territorial contribution (+1.9EJ) to global RE use change between 1995-2009. It is worth noting that Chinese territorial RE use changes are driven by worldwide changes in the six determinants, however in these global changes there are included also all the changes undergone in China itself which directly relate to Chinese investments33.
Other Asian nations such as Japan, South Korea, Taiwan, Indonesia, India, Malaysia, Philippines, Singapore, Thailand and Vietnam have to cope with similar pollution and imported fossil fuel dependency concerns. Governments of all these countries in the last decade have promoted policies34, as highlighted in IEA (2015)
and REN21 (2015) reports, enhancing both the private and public investments on RE sources. These commitments probably explains why for example Indonesia resulted to be a RE use net exporter.
Note that Malaysia, Philippines, Singapore, Thailand, and Vietnam are included in RoW’s aggregate. From the REN21 (2015) and IEA (2015) reports, it
33 This is true also for all the other country’s territorial RE use variations.
34 According to REN21 (2015) they involve primarily: capital subsidy, grant or rebate; investment
or production tax credits; reduction in sales, VAT, energy and CO2 or other related taxes; energy
36
pops up that many other emerging and developing countries included in the RoW aggregate are investing or have invested important funds on RE sources. E.g. in 2014 the top five countries in terms of investments in renewable power and fuels per unit of GDP were Burundi, Kenya, Honduras, Jordan and Uruguay, according to REN21 (2015) report. At the same time Kenya ranked first and Philippines fourth when looking at net geothermal capacity additions in 2014, in particular Kenya alone installed more than half of the planet’s new geothermal capacity. In the same year South Africa led its continent in new wind installations. Rwanda e.g. was able to significantly enhance its total generating capacity with the addition of new hydropower and solar energy capacity. In 2014 Chile has been acknowledged to have improved its wind and solar photovoltaic (PV) capacity, whereas Uruguay was, globally speaking, the first economy in terms of wind capacity per capita additions. Moreover within the countries with the largest amounts of geothermal electric generating capacity, it is possible to find: New Zeland, Philippines, Iceland and Kenya. Finally REN21 (2015) report shows that the share of countries that has implemented renewable energy policies by income group increased for all categories between 2004 and 2014. However this is surprisingly true for upper and lower middle and low income countries. In 2004 only 28% of upper middle income, 13% of lower middle income and 3% low income countries had approved pro-RE policies, but in 2014 the shares rocketed to 82%, 67% and 62% respectively. All these combined efforts are mirrored in the quantitative results and help to explain why RoW was the largest contributor of global RE use changes 7.0EJ (42.4%) and simultaneously the biggest RE use net exporter.
37
biodiesel to EU and that its investments on renewables sources and fuels went from 0.8 in 2004 to 7.6 Billions of USD in 2014. Despite this is not as high as the Chinese investments, it is still a massive allocation of public and private resources to the RE sector; especially if it is considered that the Americas’ investments, excluding US and Brazil, amounted to 1.7 in 2004 and 14.8 Billion USD in 2014.
Despite, according to REN21 (2015), EU and US had in 2014 a renewable global electric capacity second only to China, they are far from being self-independent. From the results in section 4.4, US and EU were, although their national RE use change contributed to global RE use change by 3.0EJ and 0.5EJ respectively, net importers of RE use between 1995 and 2009. At the same time it emerged that they were the only regions for which global changes eased a shift from non-RE to RE use, which is consistent with the environmental commitments publicly expressed by the governments of these economies. As for US investment on renewable power and fuels, they went from 5.4 Billion USD in 2004 to 24.3 Billion USD in 2009, reaching in 2014 the 38.3 Billion USD. EU and Russia saw a steady rise in their investment from 2004 to 2009 when their value reached 81.2 Billion USD. However, probably due to the crisis and adverse policy conditions, the investments on renewable sources and fuels declined, and in 2014 they amounted to 57.5 Billion USD. This is also reflected, according to REN21 (2015), in the EU renewable employment drop, recorded in the 2010-2013 period.
38
seem to have properly eased the shift from non-RE to RE, indeed the only fiscal incentive or public funding measure currently in effect is capital subsidy for infrastructure. The US Energy Information Administration (EIA) released a report (2015) in which Russia is acknowledge to be the largest producer of crude oil and the second largest producer of dry natural gas. Besides it has the largest reserve of gas in the world, the second largest reserve of coal and the eighth largest reserve of oil. This incredibly abundant availability of non-RE sources is likely to have constituted the main bottleneck against the expansion of the RE sector. At the same time the trade relationship with EU may have hindered the incentive to exploit RE sources. In fact according to EIA report (2015) only in 2014, more than 70% of Russian crude exports and almost 90% of Russian natural gas exports had Europe as destination. If Europe will succeed in shifting to RE sources, thus reducing its non-RE dependency, there is also the likelihood that Russia will become more proactive towards the development of the RE sector. According to this study, Russia was the only country for which both a negative territorial RE use change and a net import position was registered during the 1995-2009 period.
A final remark would be that, despite RE use has increased steadily between 1995 and 2009, its share on total energy use remained almost constant (table 4.1a). Similarly, along the same period, the substitution of RE to non-RE use has been almost stabilized at global level, as from figure 4.2a. Nevertheless its sign is negative, which means that every year there has been a shift from RE to non-RE use. This is an essential insight for future interventions, if governments all around the world are truly willing to stabilize temperature increase below 2° before 2050. Immediate efforts should be addressed to reduce to zero the contribution of this RE use determinant, thus stopping a shift from RE to non-RE use. On the medium term instead measures should focus on incentive and regulatory mechanisms capable of easing the opposite process, hence a shift from non-RE to RE use.
39
40
V. CONCLUSIONS
In the last decades the growing concern on environmental issues, reflected on the numerous and lively international meetings and debates, justifies the attention that this study dedicates to RE use. Energy use due to anthropogenic activities and global warming are two strictly connected elements, thus a shift towards more ecological means of production implies unavoidably policies addressing energy use. RE is one path that can be followed in the pursuit of a more green and sustainable socio-economic development.
To both assess the validity of past efforts and supply guidelines for future reforms on RE use, a thoroughly understanding of its evolution and main constituents is necessary. In the absence of a comprehensive study on RE use within the literature, this work aimed to give a broader overview on RE either in time frame and geographical terms. Six determinants of RE use change have been identified, they are: population growth, shift from non-RE to RE use, technological (overall energy efficiency), trade structure, commodity mix and consumption per capita chances. To this extent a SDA based on IO tables, for the 1995-2009 period, was conducted.
41
As for the six drivers between 1995 and 2009, improvements in efficiency caused a sharp reduction in global RE intensity (-18.4EJ). This effect was completely compensated by the consumption per capita change (+26.3EJ). The third biggest determinant was population growth (+10.3EJ). The contribution of these last two factors implies that independently of the effort to improve energy efficiency or to consume less energy intensive products and services, the more people in the planet the more energy will be used. Instead the global changes in RE use were only marginally affected by variations in commodity mix (-0.3EJ), trade structure (+1.1EJ) and shifts towards RE use (-2.3EJ). Especially the low contribution of trade was unexpected, considering how much its volume and value has risen in the last decades also due to the globalization phenomena. From the analysis emerged that trade itself is not an important driver of RE changes: this casts doubts on the excessive attention policy-making and research environments has addressed to it. However international trade does play a key role in explaining RE use displacements across countries, and RE use trade balances are useful indicators to uncover “who produces RE for whom”. Furthermore the negative sign of the shift determinant is a warning and suggests that immediate measures should be designed and implemented by governments to prevent further shifts from RE to non-RE use. Concurrently the process of moving towards RE sources, hence reducing the dependency on the more polluting fossil fuel and nuclear energy sources, should be accelerated.
42
combined efforts have been great, but they still seem to be insufficient to match the 2° target highlighted in the Paris agreement. The trade balance analysis uncovered that between 1995 and 2009 the main net exporters were RoW (1.2EJ), Brazil (0.3EJ) and Indonesia (0,09EJ) whereas the biggest net importers were US (0.7EJ), India (0.3EJ) and EU27 (0.3EJ). Finally EU and US were the only regions in which a positive shift towards RE use was recorded, but, as shown by their trade balances, they are still far from being self-sufficient from an RE use point of view.
In spite of its uniqueness, the study can be complemented on many aspects that therefore constitute starting points for future research. First of all the present study can be expanded also to embrace a sectorial level assessment. WIOD offers IO tables with 35 industries, consequently industry energy intensity and intra-industry trade could be investigated as well. Another improvement would be to consider more regions and a longer time span. While the advantages of a broader time period are quite straightforward, the advantages of using more single countries may be unclear at first sight. As indicated in the methodology section, the matrix contains the industry specific technical coefficients for each country. Although the aggregate RoW comprises many countries, their technologies are all proxied with the same coefficients. In other words it is assumed that e.g. Petroleum refinery sector (industry 23 in WIOD) technology is the same in Venezuela and Bhutan. This is very unlikely, knowing that the first is one of the main exporters of crude and refined petrol, whereas the second is mainly an importer35.
Consequently the former economy is likely to have an advanced Petroleum refinery industry, while the latter may not have such sector at all. One way to account for countries’ technological heterogeneity is including them separately. Including more single countries will also help in tracking more precise geographical RE use displacements via international trade. To this extent other existing databases might offer a solution.
35 According to the Observatory of Economic Complexity (OEC) database, developed by the MIT
43
REFERENCES
Alcantara, V., & Duarte, R. (2004), “Comparison of energy intensities in European Union countries. Results of a structural decomposition analysis”, Energy policy, 32(2), 177-189.
Arto, I. and E. Dietzenbacher (2014), “Drivers of the growth in global greenhouse gas emissions”, Environmental Science and Technology, 48, 5388-5394, (plus supplement).
Ciais P., Crisp D., Denier van der Gon H.A.C., Engelen R., Heimann M., Janssens-aenhout G., Rayner P., Scholze M., reviewed by Bergamaschi P., Dowell M., Broquet G., Chevallier F., Pinty B., Zunker H. (2015), “Towards a European Operational Observing System to Monitor Fossil CO2 emissions”, European
Commission: Joint Research Centre. Retrieved:
http://edgar.jrc.ec.europa.eu/news_docs/CO2_report_22-10-2015.pdf
Dietzenbacher, E., & Los, B. (1998). “Structural decomposition techniques: sense and sensitivity”. Economic Systems Research, 10(4), 307-324.
Dietzenbacher, E., Hoen, A. R. (1998), “Deflation of input‐output tables from the user's point of view: a heuristic approach”, Review of Income and Wealth, 44(1), 111-122.
Dietzenbacher, E., & Hoen, A. R. (1999), “Double deflation and aggregation”, Environment and Planning A, 31(9), 1695-1704.
Dietzenbacher, E., Pei, J., & Yang, C. (2012). “Trade, production fragmentation, and China's carbon dioxide emissions”. Journal of Environmental Economics and Management, 64(1), 88-101.
Energy Information Administration (EIA), US, (2015), “Russia”. Online reports. Retrieved: https://www.eia.gov/beta/international/analysis.cfm?iso=RUS European Commission, Joint Research Centre (JRC)/PBL Netherlands
Environmental Assessment Agency. Emission Database for Global
Atmospheric Research (EDGAR), release version 4.2.
http://edgar.jrc.ec.europe.eu, 2015.
European Commision (2015), “Towards a European Operational Observing System to Monitor Fossil”, Copernicus Europe eye’s on Earth. Retrieved:
44
Feng, K., Davis, S. J., Sun, L., & Hubacek, K. (2015). Drivers of the US CO2 emissions 1997-2013. Nature communications, 6.
Genty, A., I. Arto and F. Neuwahl (2012), “Final Database of Environmental Satellite Accounts: Technical Report on their Compilation”, WIOD. Retrieved:
http://www.wiod.org/publications/source_docs/Environmental_Sources.pdf
Hoekstra, R., & Van den Bergh, J. C. (2003), “Comparing structural decomposition analysis and index”, Energy economics, 25(1), 39-64.
International Energy Agency (2015), “Special Report on Energy and Climate
Change”, Energy and Climate Change. Retrieved:
https://www.iea.org/publications/freepublications/publication/WEO2015S pecialReportonEnergyandClimateChange.pdf
International Energy Agency (2015), “Tracking Clean Energy Progress”, Energy
Technology Perspectives. Retrieved:
http://www.iea.org/publications/freepublications/publication/Tracking_Cle an_Energy_Progress_2015.pdf
International Energy Agency (2003), “Renewables in Russia, from opportunity to
reality”. OECD/IEA publications service. Retrieved:
http://www.iea.org/publications/freepublications/publication/renewrus_2 003.pdf
IPCC (2014), “Climate Change 2014: Synthesis Report”. Contribution of Working Groups I, II and III to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change [Core Writing Team, R.K. Pachauri and L.A. Meyer (eds.)]. IPCC, Geneva, Switzerland, 151 pp.
Lenzen, M., Murray, J., Sack, F., & Wiedmann, T. (2007). “Shared producer and consumer responsibility—Theory and practice”. Ecological Economics, 61(1), 27-42.
Leontief W. (1953), "Domestic production and foreign trade; the American capital position re-examined." Proceedings of the American philosophical Society 97.4 : 332-349.
Leontief W. (1970), "Environmental repercussions and the economic structure: an input-output approach." The review of economics and statistics: 262-271. Lim, H. J., Yoo, S. H., & Kwak, S. J. (2009). “Industrial CO 2 emissions from energy
use in Korea: a structural decomposition analysis”. Energy Policy, 37(2), 686-698.
