Renewable Energy Consumption-Economic Growth Nexus-Researchgate

Rabar, Danijela.
“Renewable Energy Consumption-Economic Growth Nexus: A Nonparametric

Evidence for Developing, Graduated Developing and Developed European Union Countries.”
Modelling Economic Growth: Issues and New Insights. Juraj Dobrila University of Pula, 2018.
RENEWABLE ENERGY CONSUMPTION-ECONOMIC

GROWTH NEXUS:
A NONPARAMETRIC EVIDENCE FOR DEVELOPING,
GRADUATED DEVELOPING AND DEVELOPED
EUROPEAN UNION COUNTRIES
ABSTRACT
This study empirically investigates the relationship between economic growth

and renewable energy consumption within the data envelopment analysis (DEA)
framework. The research spans the period from 1995 to 2012, and employs six energy
and non-energy indicators based on which the relative efficiency of 28 European Union
(EU) countries is assessed. The results imply a strong divergence both within and
among the three subgroups of countries (developing, graduated developing and
developed), also pointing to inefficiency sources and their magnitude. Renewable
energy consumption appears as a major contributor to inefficiency. The findings
should assist policy makers in their attempts to improve energy related decisions.
KEYWORDS:
Economic growth, renewable energy, (in)efficiency, European Union, data envelopment
analysis, window analysis
INTRODUCTION
Today’s world is confronted with two major concerns: an increase in energy-

consuming activities in the developed and emerging countries, and waste in rich
countries. These lead to the overexploitation of most easily accessible energy resources
and to the global warming problem. However, energy is essential for the sustainable
development of nations. As a solution to these energy and environmental challenges,
it is necessary to improve the efficiency of energy use, which is possible by increasing
1
Rabar, Danijela. “Renewable Energy Consumption-Economic Growth Nexus: A Nonparametric
the share of renewable energy in total final energy consumption and by the consequent
decrease of greenhouse gas (GHG) emissions.
Renewable energy is collected from solar, wind, rain, biomass and waste,
geothermal, waves and tides. Unlike the conventional energy, it is safe, clean and its
sources (also called renewables) are naturally replenished on a human timescale.
Therefore, its share in total energy consumption has been growing fast worldwide in
recent years and is expected to occupy a leading position. Renewable energy provided
an estimated 19.2 % of global final energy consumption in 2014. The contribution of
modern renewable technologies increased to approximately 10.3 %, while the
traditional use of biomass accounted for about 8.9 %. An unprecedented growth in
renewable capacity and generation occurred in 2015 despite tumbling global prices for
all fossil fuels, ongoing fossil fuel subsidies and other challenges faced by renewables,
including the integration of rising shares of renewable generation, policy and political
instability, regulatory barriers and fiscal constraints (REN21, 2016).
The necessity to reconcile energy demand and implications for the environment
with sustainable economic growth led to the understanding of the importance of
renewables and, consequently, to comprehensive administrative efforts in terms of
EU’s renewable energy policy. These endeavours have made the EU one of the leading
participants in the world in sustainability and renewable energy implementation, and
resulted in various official publications of different EU institutions (EU commission,
EU parliament, Council of ministers, etc.), including white papers, directives,
strategies, communications, reports and resolutions. The use of renewable energy in
the EU is projected to increase substantially by 2020. Although this horizon is still a
long way off, according to the latest reports significant progress has been made since
2005 and the EU is on track to reach its 2020 renewable energy targets.
A comprehensive review of the policy framework for renewable energy in the
EU and an analysis of the progress made by the use of renewable energy are given by
Scarlat et al. (2015) and European Parliament (2017). Although of indisputable
importance, these numerous documents, and the resource-specific issues they
consider, are not the focus of this research. Hence only a short chronological overview
2
of the main acts, along with set climate and energy goals and targets, is provided
(Table 1).
Table 1. A brief survey of EU policy framework for renewable energy

Year Document title Description Main EU’s climate and energy
(policy) goals and targets
Initial steps
1997 Energy for the future: White Paper for a - doubling the share of renewable
Renewable sources of Community Strategy and energy in the gross energy
energy Action Plan consumption from 6% to 12% by
2010
2001 Renewable Electricity Directive on the promotion - increasing the share of renewable
Directive 2001/77/EC of electricity produced from sources in the total electricity
renewable energy sources in production from 12.9% to 21% by
the internal electricity market 2010
2003 Biofuels Directive Directive on the promotion - increasing biofuels incorporation
2003/30/EC of the use of biofuels or other rate to 5.75% in the energy content
renewable fuels for transport of all fuels used for transport in
2010
Renewable Energy Road Map
2007 Renewable Energy Road A long-term strategy for - using renewable energy sources to
Map–Renewable energies renewable energy until 2020 meet 20% of final energy
in the 21st century: consumption needs and 10% of
building a more transport fuel consumption coming
sustainable future from biofuels by 2020
Renewable Energy Directive
2009 Renewable Energy Directive on the promotion - increasing the share of renewable
Directive (RED) of renewable energy sources energy to 20% of gross final energy
2009/28/EC (broken down into nationally consumption and to 10% in
binding sub-targets taking transport by 2020
into account different
starting points of the
member states)
Future steps
2011 Energy Roadmap 2050 A strategy paper that - decreasing primary energy
explores different scenarios demand from 16%-20% by 2030 and
on how the goal to reduce 32%-41% by 2050 (compared to
GHG emissions from 80%- maximum levels in 2005-2006);
95% below 1990 levels by contribution of carbon capture and
2050 can be achieved without storage with 19-32% in power
disrupting energy supplies generation; increasing the share of
and competitiveness renewable energy resources in 2050
up to 55-75% of gross final energy
consumption and 64-97% in
electricity consumption (compared
to 19.7% in EU-28 in 2010)
2012 Renewable Energy: a Communication from the - ensuring that renewable energy
major player in the commission on the further contributes to security and
European energy market development of renewable diversity of energy supply,
energy for a medium term competitiveness, and environment
perspective beyond 2020 and climate protection, and
supports economic growth,
employment creation, regional
development and innovation
3
2013 Delivering the internal Communication from the - completely overhauling the
electricity market and commission on the further subsidies that Member States are
making the most of public guidance on renewable allowed to offer the renewable
intervention energy support schemes and energy sector, preferring tendering,
on the use of cooperation feed-in premiums and quota
mechanisms to achieve obligations to commonly used feed-
renewable energy targets at a in tariffs
lower cost
2014 Guidelines on State aid Communication from the - establishing the support by a
for environmental commission on the further competitive bidding process for at
protection and energy- shaping of the new least 5% of the new renewable
related objectives for framework for renewable capacity (in a transitional phase
2014-2020 energy support schemes covering the years 2015 and 2016)
Conclusion
2016 Renewable energy A resolution adopted by the - presenting a more ambitious
progress report European Parliament climate and energy package by
2030 which increases the target for
renewable energy sources to at least
30% to be implemented by means
of individual national targets
Source: Compiled by the authors based on European Parliament (2017) and Scarlat et al. (2015).
This brief survey is not intended to completely cover this voluminous literature,
but to show researchers and practitioners the major EU policy initiatives concerning
renewable energy and its share in total energy consumption. It chronologically covers
all the period since 1997 when the European Commission set the basis for the EU policy
on renewable energy. At the same time, the scenarios contained in these documents
extend even to 2050.
The purpose of this study is to articulate the nexus between economic growth
and renewable energy consumption in the case of EU-28 for the period 1995-2012. As
a cross-country time-series nonparametric analytical tool, DEA is used assuming both
constant and variable returns to scale and input orientation.
The paper is organized as follows. A succinct overview of the literature on the
role of renewable energy in economic growth, using DEA approach, is provided in the
next section. Section 3 briefly describes the chosen socio-economic indicators along
with criteria for their selection, and the data collection, also presenting the DEA
approach setting up methodological framework used in this investigation. Section 4
shows the empirical results of the DEA-based efficiency analysis of EU-28 members.
The findings of the study and general remarks about their implications for
policymakers are summarized in the last section of this paper.
4
1. REVIEW OF LITERATURE ON THE ROLE OF

RENEWABLE ENERGY IN ECONOMIC GROWTH
USING DATA ENVELOPMENT ANALYSIS
In recent years, a number of empirical economic scientific studies have focused

on cross-country energy efficiency evaluation and comparison, attempting to identify
the countries that performed better, to determine major barriers and to overcome the
challenges by suggesting potential solutions. The following is a brief but fairly
complete synopsis of the relevant published literature dealing specifically with the
importance of renewable energy for economic growth, and using DEA-based
approaches.
Chien and Hu (2007) analysed the effects of renewable energy on the technical
efficiency of 45 economies across the world during the 2001–2002 period. Labour,
capital stock and energy consumption were the three inputs and real GDP was the
single output. The analysis revealed that, compared to non-OECD economies, OECD
economies had higher technical efficiency and a higher share of geothermal, solar, tide,
and wind fuels in renewable energy. However, non-OECD economies had a higher
share of renewable energy in their total energy supply than OECD economies.
Zhou et al. (2007) employed the ratio of total final energy consumption to GDP,
the percentage of renewable energy in total final energy consumption and the ratio of
CO2 emissions to GDP for modelling the sustainable energy development of eighteen
Asia-Pacific Economic Cooperation economies in 2002. The results indicated that the
sustainable energy development of these economies was not so good since most of
them had a very small sustainable energy index value.
Ceylan and Gunay (2010) analysed economy wide energy efficiency performance
and the energy saving potential of Turkey by means of cross-country comparison with
the EU countries for the period of 1995-2007. The model took capital, labour and total
R&D as non-energy inputs, oil, gas, solid fuels, nuclear energy and renewable energy
consumption as energy inputs, GDP as the desirable output and GHG emissions as the
undesirable output. The results indicated an improvement in energy efficiency over
the years, but showed that environmentally aware energy efficiency performance and
its improvement pace were much lower in all countries.
5
Halkos and Tzeremes (2013) examined the relationship between renewable

energy consumption and economic efficiency on a sample of 25 European countries
for the year 2010, using total labour force and capital stock as inputs, GDP as output
and primary energy consumption for renewables as external variable. The results
showed that renewable energy consumption had a positive effect on countries’
economic efficiency for lower consumption levels, while for higher levels the analysis
reveals mixed effects that are also subject to regional disparities. It appeared that this
effect depended also on countries’ specific regional characteristics as well as on the
environmental policies adopted.
Menegaki (2013) calculated inefficiencies in the growth of 31 European countries
within the period 1997-2010, using the variables typically used in the growth-energy
nexus literature such as GDP, fuel energy consumption, CO2 emissions, employment
and capital, but also with a particular focus on renewable energy sources consumption.
The analysis suggested that countries with remarkable renewable energy performance
had medium to low efficiency, while renewable energy laggards were among the most
technically efficient countries in Europe.
Using the number of employees, renewable and non-renewable energy
consumption and real productive capital stock as inputs, and real GDP as good and
CO2 emissions as bad output, Apergis et al. (2015) showed that efficiency levels of 20
selected OECD countries, in the period 1985-2011, were high and were reducing over
time. Higher efficiency levels were found in EU countries, followed by NAFTA and
G7. Moreover, capital-intensive countries were more energy efficient than labour-
intensive countries.
Based on labour, capital and renewable energy supply as inputs, and GDP as a
desirable and carbon emissions as an undesirable output, Woo et al. (2015) examined
the environmental efficiency of renewable energy from the static and the dynamic
perspective in 31 OECD countries from 2004 to 2011. The results demonstrated
geographical differences in environmental efficiency across the countries, and showed
that they implemented both catch-up and frontier shift strategies to improve
environmental efficiency.
6
Using energy environmental and non-radial Malmquist indices1, an empirical

study of 26 OECD and four BRIC countries was performed by Chen and Geng (2016)
for the period 2002-2011. The outcomes, based on renewable energy, fossil energy,
capital stock and labour force as inputs, and real GDP as a desirable and CO2 emission
as an undesirable output, are as follows. The 30 countries can be divided into four
categories with corresponding specific strategies for energy saving and emissions
reduction (ESER). ESER performance is underestimated for most countries, regardless
of renewables. Technological progress and efficiency improvement are generally out
of sync, mainly because of the difficulty to achieve the efficiency improvements.
Although the importance of renewables and their role in reducing climate change
and achieving a sustainable energy system is reasonably well understood, there is a
lack of literature on the nexus between renewable energy consumption and economic
growth. Therefore, the first aim of this paper is to help fill this void by conducting an
empirical study, with reference to EU members over an 18-year period at the turn of
the millennium.
2. MODEL SETUP – INDICATORS, DATA AND

METHODOLOGY
A different selection of input and output variables leads to different efficiency

results that are additionally influenced by model selection. Hence, the choice of
indicators, along with the model design, specifically regarding assumptions on model
orientation and returns to scale, are among the most significant and most delicate steps
in constructing an appropriate DEA model for evaluating the relative performance of
countries. Due attention should also be paid to data availability and accessibility.
Therefore, suitability of the model for a specific process application, in terms of
appropriate indicators and model selection, is the main postulate.
1Although Malmquist productivity index is defined based on the concept of distance function, it can
also be directly represented by DEA efficiency measures (Zhou et al., 2008).
7
2.1. Indicators
Global warming, caused by the rapidly expanding emissions of greenhouse gases

such as methane and carbon dioxide, along with the scarcity of energy resources and
their allocation, makes an omnipresent reminder of the crucial role that renewable
energy plays in the functioning of modern world economy. This phenomenon should
be addressed by a proper selection of indicators that will demonstrate the relationship
between renewable energy consumption and economic growth. After considering a
broad spectrum of relevant economic, energetic and environmental indicators, and
after ranking them by importance, unemployment, gross capital formation, renewable
energy consumption, energy use, GHG emissions and gross domestic product (GDP)
were selected for this research.
GDP, commonly used to determine the economic performance of a country, is
used as a proxy for economic growth, renewables are represented by renewable energy
consumption, while energy use (in kg of oil equivalent) and GHG emissions (in kt of
CO2 equivalent) are taken as indicators of energy and environmental impact.
Additional non-energy and non-environmental variables are labour and capital,
represented respectively by unemployment and gross capital formation.
Unemployment is used as an indicator of the state of the economy in terms of socio-
economic well-being, while capital formation is chosen as an indicator of investment
of the economy.
With the intention to make this analysis more reliable, some specific adjustments
of the indicators were required. To account exclusively for the rise in production and
to eliminate the effect of increasing prices, GDP is at constant 2010 US dollars. To
neutralise significant inter-country variance in absolute levels of indicators, in a part
resulting from large differences in population, GDP, energy use and GHG emissions
are taken per capita, while renewable energy consumption, unemployment and gross
capital formation are expressed as percentage of, respectively, total final energy
consumption, total labour force and GDP2. Such operations adjust the data to account
2 For simplicity, the names of the variables will not be used in their full forms which include the
aforementioned adjustments. For example, unemployment, total (% of total labour force) will be simply
referred to as unemployment.
8
for changes in prices and population size during the observed period, thus offering
more appropriate cross-country comparisons and simpler interpretations of the
results. These six adjusted variables will be merged into a single performance scale
which synthesizes separate indicators into one single statistic.
To build a model adequate for capturing the dynamic relationship between each
of the selected indicators and GDP as a standard measure of national economic
performance, with special emphasis on renewable energy-GDP nexus, only GDP is
considered an output while other five indicators are considered inputs.
2.2. Data Collection
The nature of the selected variables allows for comparisons to be made on an

annual basis. Data have been taken from the World Bank’s online World Development
Indicators (http://data.worldbank.org/indicator, accessed: 1 March 2017), and then
modified according to the aforementioned requirements. Since the data on some of
these indicators for some countries in transition were not available for earlier years,
1995 is chosen as the initial year for this research. Also, at the moment this study was
started, the latest available data for renewable energy consumption and GHG
emissions were from 2012, which is the reason for selecting this year as the final year
for the research. Thus, the selected period, spanning from 1995 to 2012 and divided
into yearly sub-periods, fulfils our original intention to include the longest period of
available data. For the purpose of this research, the EU countries were classified into
three subgroups: 15 developed (Austria, Belgium, Denmark, Finland, France,
Germany, Greece, Ireland, Italy, Luxembourg, Netherlands, Portugal, Spain, Sweden
and United Kingdom), eight graduated developing (Cyprus, Czech Republic, Estonia,
Latvia, Lithuania, Malta, the Slovak Republic and Slovenia) and five developing
countries (Bulgaria, Croatia, Hungary, Poland and Romania)3. The fragmentation of
3This division is based on the International Monetary Fund’s classification of countries into ‘emerging
market and developing’ economies, on one side, and ‘advanced’ economies, on the other. Since some of
the now advanced economies were considered developing until recently, they are classified into a
separate subgroup designated as ‘graduated developing’, regardless of the year in which they became
advanced. With the exception of Cyprus and Malta, these are all transition economies. The rest of the
advanced economies are labelled as ‘developed’.
9
Renewable Energy Sources policies across EU member states is not regarded as

relevant for the study.
Summary statistics for the adjusted data, aggregated over time and across
countries, are contained in Table 2 for each input and output used in the research.
Table 2. Data summary statistics, 1995-2012

Variable Countries Mean Median St. dev. Min. Max.
Developed 7.9 7.7 3.9 1.8 25.2
Graduated
9.4 7.9 4.2 3.3 19.3
Unemployment developing
Developing 10.6 10.0 3.9 5.5 20.5
EU-28 8.8 7.9 4.1 1.8 25.2
Developed 22.3 22.2 3.0 12.8 31.9
Gross Graduated
26.2 26.1 5.5 12.6 41.5
capital developing
formation Developing 23.4 23.3 5.0 0.3 36.8
EU-28 23.6 22.7 4.6 0.3 41.5
Developed 12.7 8.2 11.9 0.9 49.9
Renewable Graduated
13.1 10.4 10.8 0.1 40.4
Inputs energy developing
consumption Developing 11.3 9.8 5.4 3.1 24.1
EU-28 12.5 8.9 10.7 0.1 49.9
Developed 4,240.9 3,846.4 1,581.0 1,986.8 9,428.8
Graduated
2,933.4 3,012.6 845.3 1,618.5 4,491.5
Energy use developing
Developing 2,247.1 2,389.7 357.2 1,510.1 2,746.7
EU-28 3,511.3 3,246.9 1,493.2 1,510.1 9,428.8
Developed 12.3 11.4 4.5 6.7 28.9
Graduated
GHG 9.9 8.9 3.8 4.6 23.0
developing
emissions
Developing 8.0 7.5 1.7 5.3 11.9
EU-28 10.9 9.8 4.3 4.6 28.9
Output Developed 42,746.6 40,212.7 16,719.1 18,081.6 110,001.1
Graduated
16,154.7 15,579.8 6,732.0 5,134.6 32,651.9
GDP developing
Developing 9,060.5 8,893.0 3,237.0 3,781.9 14,782.7
EU-28 29,133.5 26,211.1 19,596.4 3,781.9 110,001.1
Source: Authors’ calculations based on data from the World Bank
2.3. Methodological framework
DEA is a non-parametric performance measurement technique commonly

utilized to assess the relative efficiency of a group of peer entities, i.e. of decision
making units (DMUs) that consume the same multiple inputs to create the same
multiple outputs. The calculation is made by incorporating empirical data on the
10
selected inputs and outputs of all analysed entities into a linear program that
represents the DEA model and provides a single relative performance efficiency index.
Because of its unique robust characteristics, DEA has been broadly practiced in a
variety of both profit and non-profit fields, and on both organizational and country
level, over nearly four decades. The DEA concept is based on the seminal work of
Farrell (1957). Since it was initially developed in 1978 by Charnes et al., and up to the
year 2014, about 6,500 DEA papers of both theoretical and practical interest have been
published in ISI Web of Science database (Liu et al., 2016). Thus, this approach is
proven to be considered as an internationally and academically recognized tool for
decision support.
The decisive factor in the selection of this method over traditional benchmarking
techniques for the present study was its possibility of dynamic efficiency
measurement, with inputs and outputs expressed in different and often incongruous
measurement units, and without the necessity of a priori determination of variable
weights or the exact knowledge of the functional form linking inputs and outputs. As
a matter of fact, the weights are set by the model itself, in a manner that maximizes the
efficiency rating for each evaluated entity, thus circumventing the subjectivity
connected with the estimation of each weight’s significance. The production possibility
frontier is constructed by simply enveloping observed inputs from below and outputs
from above. The DMUs on this ‘best practice’ frontier are classified as efficient (i.e.
benchmarks), compared to the rest of the analysed DMUs, and are all assigned the best
efficiency measure of one (or 100%). The non-frontier DMUs turn out to be inefficient
and are scored somewhere between zero and one, depending on their distance from
the frontier. This inefficiency is a result of input surpluses (i.e. using excessive inputs
at a given output level) and/or output shortfalls (i.e. producing poor output at a given
input level), and can be removed by reaching a model-calculated efficient projection
point of operation on the efficient boundary. Since empirically generated, this
boundary appears as an objectively achievable goal for each inefficient DMU and, at
the same time, serves as the basis for identifying its inefficiency sources and their
amounts, improvement directions and reference DMUs to most directly compare with.
11
The basic DEA models are characterized above all by the assumption on returns
to scale (constant or variable) and by the model orientation (to input minimization or
to output maximization). The numerous advanced models, built upon the basic ones,
are distinguished by miscellaneous extensions and relaxations.
Nevertheless, the first issue the analyst must face after the selection of indicators
is to determine the type of returns to scale in order to estimate the production frontier.
In this sense, two basic models most commonly exploited in DEA applications are CCR
(Charnes et al., 1978) and BCC (Banker et al., 1984), named after their authors’ initials.
The first model, under constant returns to scale (CRS) assumption, measures
inefficiencies due to the size of operations and the input-output configuration,
resulting in an overall efficiency score. One of its various theoretical extensions
represents the second model that, under the assumption of variable returns to scale
(VRS), estimates pure technical efficiency which purely reflects managerial
underperformance. The CCR score takes no account of scale effect and is therefore
called the (global) technical efficiency (TE). The BCC model, on the other side, separates
the scale effect thus expressing the (local) pure technical efficiency (PTE). Particular
previous studies and preliminary surveys on the process to be investigated and its
properties could provide a first indication on the type of returns to scale and confront
the validity of its assumption. Despite this, the characteristics of the production
frontiers sometimes, as in our application, cannot be determined with certainty, which
complicates the selection of model type. Namely, in these cases, it may be risky to rely
on only one particular model. A potential solution is to try some alternative methods
before arriving at a definitive conclusion, such as using statistical regressions and DEA
to cross check each other. Another possible solution is to try models under each of the
two assumptions, compare the similarity between their results and, based on the
magnitude of their differences and utilizing expert knowledge of the problem, find the
most suitable type of assumption for the analysis. In the thus conducted initial phase
of our research, the results received from the said models appeared to have significant
discrepancies. They may be ascribed to the return effect with respect to the range of
activities, which makes the BCC model more appropriate for describing the
investigated process, and consequently is the reason why it receives somewhat greater
12
representation in this paper. However, it is considered important to explore the extent

of inefficiency that can be attributed to scale economies and the one that resulted from
the waste of inputs independent of the scale of operations. For this reason, both
constant and variable returns to scale DEA models are considered.
The next issue the analyst must deal with is to select the model orientation.
Namely, depending on whether managerial choice is primarily to reduce the input
values or to increase the output amounts, DEA models are oriented to inputs or
outputs. This distinction in model orientation results in different projection trajectories
onto the efficient frontier and hence in different projection points of inefficient DMUs.
Since the distances from an inefficient DMU to its input and output projections
obviously differ, efficiency is not equally achievable by models with different
orientations. As we decided to treat GDP as the only output among the six selected
indicators, the input-oriented approach is chosen as more appropriate. This
orientation provides the opportunity to assess to what extent inputs would have to be
improved, without reducing current output levels, in order for a DMU to be
considered relatively efficient.
The following description of the models is based on Cooper et al. (2006, pp. 43-
47, 87-89, 140-142).
DEA decomposes the data set into two matrices – the matrix of inputs and the
matrix of outputs. They are denoted respectively as = ∈R and = ∈
R , where n is the number of entities under consideration, m is the number of
consumed inputs, and s is the number of created outputs. The model oriented on
inputs assesses the relative efficiency of , ∈ 1, 2, … , , by solving the following
linear program:
min
subject to − λ≥0 (1)
λ≥ (2)
λ≥0 (DEA-CCR) (3)
$λ = 1 (DEA-BCC) (4)
The objective function, together with the constraints (1) to (3), forms the CCR
model. In addition to these conditions, the BCC model contains the constraint (4). The
13
first three conditions consist of m, s and n constraints, respectively. In our study, m is

∗
5, s is 1 and n is 28. is the efficiency score of DMU , and can take any real value
between 0 and 1 (including 1). For an inefficient DMU , it also represents the input
reduction rate, while vector ) ∈ R corresponds to the proportions contributed by
efficient entities to its projection onto efficient frontier. e is a row vector with all
elements equal to 1.
∗
From the constraints (1) and (2), it is clear that, in case < 1, + λ, λ,
∗
outperforms + , ,. Consequently, the input excesses and the output shortfalls are
identified as “slack” vectors and defined, respectively, by
-. = − λ +- . ∈ R ,, -/ = λ − +- / ∈ R ,,
which are both non-negative for any feasible solution + , λ, of the above linear
program. To reveal the possible input surpluses and output shortages, the second
phase of the aforedescribed procedure should be performed. In that phase, the sum of
the mentioned surpluses and shortages is maximized while keeping equal to the
optimal objective value derived from the first phase + ∗ ,.
Definition 1 (Efficiency):
If an optimal solution + ∗ , λ∗ , - .∗ , - /∗ , obtained in this two-phase process satisfies
∗
= 1 and has no slack +- .∗ = 0, - /∗ = 0,, then the is called efficient, otherwise
it is inefficient.
The provided information can be used as a basis for goal setting for the DMU
being assessed as inefficient. A first step in determining its targets is to compare it with
the efficient DMUs that make up its reference set.
Definition 2 (Reference Set):

For an inefficient , its reference set 0 is defined based on an optimal solution λ∗
by
0 = 12 4 λ∗ > 06 +2 ∈ 1,2, … , ,.
The optimal solution is given with

∗
=∑ ∈9: λ∗ + - .∗,
=∑ ∈9: λ∗ − - /∗.
14
It can be noted, based on these relationships, that the efficiency of can be

∗
upgraded if the input amounts are decreased radially by the factor and the input
surpluses recorded in - .∗ are removed, and if the output amounts are increased by the
output shortages in - /∗ . These input and output improvements form the following
projection:
∗
; = − - .∗ ,
; = + - /∗.
The constraint (4) distinguishes the BCC from the CCR model, causing notable
differences in their production frontiers. Consequently, in practice, the projection of
any inefficient DMU on the CRS frontier is farther than its projection on the VRS
frontier. CCR efficiency is therefore more difficult to accomplish, resulting in lower
scores. Moreover, it may happen that a DMU is BCC efficient, but has a low CCR score,
which means that it is locally, but not globally efficient due to its scale size. It is
therefore recommended to evaluate the scale efficiency (SE) of a DMU as the ratio of the
CCR and BCC efficiency scores.
Definition 3 (Scale Efficiency):

∗ ∗
Let the CCR and BCC scores of a DMU be <<= and ><< , respectively. The scale
efficiency is defined by
@∗
?0 = @AAB
∗ .
CAA
Obviously, this number cannot exceed one and, for a CCR efficient DMU, it is
equal to one. This formula reveals an efficiency decomposition as
Technical Efficiency = Pure Technical Efficiency × Scale Efficiency,
proposing an answer whether the overall technical inefficiency (TE) is caused by

inefficient operation (PTE) or by disadvantageous conditions (SE) or by both.
Detecting trends in efficiency scores associated with economic growth appears
also very significant for economic policy makers. Within the DEA framework, it can
be provided by one of the numerous extensions to basic models – the window analysis.
15
In that case data for several periods are incorporated into analysis for each entity,
which is regarded as a different entity in each of the observed periods. Efficient frontier
of the window analysis model is built analogously to the one in the basic model with
the difference in sample size, also depending on the window length which can range
from one year to all available years. If in our case, for example, we choose a window
length of five years, there would be 14 windows (1995-1999, 1996-2000, ..., 2008-2012)
provided by dropping one year and adding another each time. In this way, each of the
years from 1999 to 2008 is tested ten times. On the other hand, the initial and final years
are tested only once. This makes the overview of the obtained results over time
somewhat unbalanced, which could be circumvented by equal treatment of all years
from the observed period. It is possible only in the cases with minimum and maximum
window length. In the first case there are 18 one-year windows, while in the second
one there is one 18-year window. Since the countries are mutually compared only
within and not across windows, the first case implicates 18 sets of 28 entities, while the
second implicates one set of 504 (18 x 28) entities4. The opportunity to compare each
country’s performance in a given sub-period not only to other countries’
performances, but also to its own performance in any of the other sub-periods is of
utmost importance. Therefore, the purpose of this research will be fulfilled by
examining the second of the above cases.
3. MODEL APPLICATION AND EMPIRICAL ANALYSIS
The relative efficiency assessment of 28 EU current member states was carried

out on the empirical data relating to six performance indicators over the period from
1995 to 2012. For this purpose, the input-oriented CCR and BCC models, extended by
the window analysis with one maximum size window of 18 years, are employed.
To provide a starting point for an appropriate decision making, a high-quality
analysis of inefficiency should always be accompanied by information on its sources
and their magnitudes, along with the improvement paths based on reference sets.
4Among these entities, 270 (18 x 15) correspond to developed, 144 (18 x 8) to graduated developing and
90 (18 x 5) to developing countries.
16
Since these valuable facts are available with the basic DEA models, but not with the
window analysis, the new combined models were built to compute them while leaving
the relative efficiency scores unaltered (Rabar, 2013). For brevity, these two models are
hereafter referred to as CCR-I-W1 and BCC-I-W1.
To increase the efficiency of an inefficient entity, the input-oriented DEA models
first tend to reduce its inputs and, if that is insufficient, to enlarge its outputs. Since in
this study gross capital formation and renewable energy consumption are specified as
inputs but, in practice, preferred to have larger amounts, in the model calculation their
reciprocals are taken. This at the same time preserves the relationship between inputs
and outputs and enables the assessment of countries’ performance regarding their
capability to minimize unemployment, energy use and GHG emissions, and to
maximize gross capital formation, renewable energy consumption and GDP.
However, recalculation of the obtained results and an additional caution in their
interpretation are necessitated. The evaluation of the relative performance is computed
by the software package DEA-Solver-Pro 7.0F from the Saitech Company, and further
processed by the authors.
The technical and pure technical efficiency results respectively generated by the
models CCR-I-W1 and BCC-I-W1, along with the calculated scale efficiency results, are
classified by subgroups of countries (developing, graduated developing and
developed). Due to space limitations, it would be far too extensive to analyse each
country in each year, so the focus will be placed on the average efficiency scores that
are provided in Table 3 and Figures 1 and 2 (for detailed results see the supplement).
Table 3. Summary statistics on input-based efficiency scores, 1995-2012

(Sub)group of countries
Result of
Efficiency type Graduated
analysis Developed Developing EU-28
developing
Efficiency scores TE 0.700 0.361 0.226 0.519
(averaged per PTE 0.863 0.850 0.848 0.856
country) SE 0.809 0.417 0.264 0.600
Efficiency scores TE 0.685 0.361 0.226 0.506
(averaged per PTE 0.847 0.812 0.831 0.835
inefficient country) SE 0.799 0.417 0.264 0.589
Number (%) TE 13 (4.8%) 0 (0%) 0 (0%) 13 (2.6%)
of efficient PTE 27 (10%) 29 (20.1%) 9 (10%) 65 (12.9%)
entities SE 13 (4.8%) 0 (0%) 0 (0%) 13 (2.6%)
On average TE Luxembourg Cyprus Croatia Luxembourg
17
most PTE Austria Latvia Romania Latvia

efficient country SE Denmark Cyprus Croatia Denmark
On average TE Portugal Lithuania Bulgaria Bulgaria
least PTE Germany Slovakia Poland Slovakia
efficient country SE Portugal Lithuania Bulgaria Bulgaria
On average TE 2008 2008 2012 2008
most PTE 2008 2007 2008 2008
efficient year SE 2011 2012 2012 2012
On average TE 1996 1995 1995 1996
least PTE 1996 1999 1996 1996
efficient year SE 1995 1995 1995 1995
Source: Authors’ work based on DEA-Solver-Pro calculations
Figure 1. Relative efficiency scores averaged by country
Figure 2. Relative efficiency scores averaged by year: (a) developed countries; (b)
graduated developing countries; (c) developing countries; (d) EU-28
(a)
(b)
18
(c)
(d)
Among other things, both similarities and differences between the results of the
models are revealed. The differences between efficiency scores are considerable and,
in average terms, most prominent in the case of developing countries. Average
countries’ efficiencies are, without exception, below one, which is indicating some
degree of inefficiency for each country. This means that none of the 28 countries was
efficient throughout the entire period. Moreover, 13 of them were continuously pure
technically inefficient, and even 23 exhibited permanent technical and scale
inefficiency. We note as a curiosity that 2003 is the only year with no pure technically
efficient country. Only 13 entities were technically efficient, which is every fifth pure
technically efficient entity. Interestingly, all technically efficient entities are from the
subgroup of developed countries. Moreover, the developed countries are on average
the most efficient, followed by the graduated developing countries, regardless of
efficiency type. The differences between the subgroups of countries are however much
more prominent in technical and scale than in pure technical terms, both in each year
(Figure 3) and averaged over all years (Table 3). According to both models, on average
the least and most efficient years were 1996 and 2008 respectively. This may seem
19
surprising, given that 2008 was the first full year of the Great Recession. It should
therefore be underlined that five out of six indicators upon which these results are
based had better than average values in that year.
Figure 3. Differences between the subgroups of countries based on average scores
The standard deviations in both models are significantly lower when considering
a particular country in different years than when considering different countries in a
particular year. This testifies to a well-balanced countries’ performance over time and,
at the same time, indicates large differences across countries. These differences are
especially pronounced among graduated developing countries, which comes as no
surprise given their heterogeneity. Namely, only this subgroup includes both
transition and non-transition countries, also with considerable variation in the dates
of their inclusion to the list of advanced economies.
The efficiency scores of inefficient entities allow their immediate ranking, while
efficient entities cannot be ranked directly because of their maximum efficiency. Since
the reference set of an inefficient entity is composed of efficient ones, one among many
approaches proposed by researchers for ranking an efficient entity is to sum the
frequency of its appearance in each reference set. Consequently, the higher the
frequency, the more robust the entity is. Table 4 shows the results of the BCC-I-W1
model in these terms. Because of the large sample size, the findings are not presented
at the entity level, but at the country level. In this sense, it should be noted that a
country may occur as a reference, even in the same reference set, through its
performance not only in one year but in any year in which it was efficient. At the same
time, any of these efficient performances may become a reference not only to a number
of different inefficient countries, but to the same country (including itself) in different
years. Nevertheless, 15 out of 28 countries have been selected as referential for the
20
proposed model. The total number of their appearances in the reference sets (1,723) is
divided into four groups – the one named ‘Self’ for the cases in which a country’s
performance in a particular year is a reference to its own performance in another year,
and one for each subgroup of countries (excluding self-comparison). Latvia is the
country with the highest average pure efficiency score and, although efficient in only
eight years, sets an exemplar by serving as a reference for the greatest number of
inefficient entities (370) that mostly represent developed countries (45.68 %). On the
other side, the Netherlands is a reference in only nine cases, out of which in even eight
to itself. By contrast, Lithuania is the only country that is no reference to itself, but only
to other graduated developing countries.
Table 4. Reference set membership in the model BCC-I-W1

Graduated % in
Self Developed Developing
EU-Country developing ∑ Total
No. % No. % No. % No. % No.
Austria 28 56.00 9 18.00 13 26.00 0 0.00 50 2.90
Denmark 30 11.58 209 80.69 12 4.63 8 3.09 259 15.03
Ireland 11 16.92 43 66.15 11 16.92 0 0.00 65 3.77
Developed Luxembourg 20 40.82 22 44.90 7 14.29 0 0.00 49 2.84
Netherlands 8 88.89 1 11.11 0 0.00 0 0.00 9 0.52
Portugal 17 32.08 13 24.53 17 32.08 6 11.32 53 3.08
Sweden 22 12.15 132 72.93 23 12.71 4 2.21 181 10.50
Cyprus 23 7.23 191 60.06 49 15.41 55 17.30 318 18.46
Czech Republic 9 60.00 5 33.33 1 6.67 0 0.00 15 0.87
Graduated Estonia 2 6.67 6 20.00 22 73.33 0 0.00 30 1.74
developing Latvia 24 6.49 169 45.68 111 30.00 66 17.84 370 21.47
Lithuania 0 0.00 0 0.00 13 100.00 0 0.00 13 0.75
Malta 19 15.97 46 38.66 20 16.81 34 28.57 119 6.91
Croatia 8 22.22 13 36.11 8 22.22 7 19.44 36 2.09
Developing
Romania 26 16.67 27 17.31 33 21.15 70 44.87 156 9.05
Total 247 14.34 886 51.42 340 19.73 250 14.51 1723 100.00
Source: Authors’ calculations using DEA-Solver-Pro
The differences between target and actual values provide the most direct insight
into the extent of each indicator’s contribution to inefficiency and dictate performance
improvements of the countries. The improvements needed according to the model
BCC-I-W1 are averaged over time and across countries, and expressed as percentages
in Table 5 for each indicator (for detailed results see the supplement). Thus represented
inefficiencies can be decomposed into radial and non-radial components and removed
by carrying out the aforedescribed two-phase procedure. The overall pure efficiency
score, averaged over time and across inefficient countries, is 0.835 (Table 5) which
21
suggests that on average the countries have 16.5 % inefficiency5. To perform efficiently,
∗
the first phase enjoins the countries to eliminate radial inefficiency by (0.835), i.e.
through the maximum possible equiproportional reduction (-16.5 %) of all five inputs6,
while preserving the output quantities at their actual levels and the proportions in
which the inputs are utilized. All the percentage differences in the last column of Table
5 indicate that the inefficiency cannot be completely eliminated by solely removing the
radial inefficiency. Hence, after radial reduction of all five inputs, the countries should
eliminate their non-radial inefficiency, through further decrease of inputs in the
second phase, which alters their proportions. The different percentages of
improvements needed to remove inefficiency signify that for all inputs, except for
renewables, the radial inefficiency is much more pronounced than non-radial for the
majority of inefficient countries in the observed year. Conversely, GDP, as the single
output, causes solely non-radial inefficiency.
Table 5. Summary statistics on inefficiency sources in the model BCC-I-W1, 1995-2012

Proposed input and output improvements per inefficient country (%)
(Sub)group of countries
Indicator Graduated
Developed Developing EU-28
developing
Unemployment -26.18 -26.39 -28.83 -26.76
Gross capital formation 18.58 26.86 22.01 21.23
Renewable energy consumption 51.56 47.33 41.85 48.63
Energy use -17.38 -24.82 -17.74 -19.41
GHG emissions -21.09 -26.81 -23.53 -23.04
GDP 0.27 10.63 32.07 8.86
Renewables obviously represent a major source of inefficiency during the whole

period, with the exception of the last investigated year when this was unemployment.
When this insufficient share of renewable energy in the total final energy consumption
is considered by subgroups of countries, it can be concluded – the more the country is
developed, the higher the required increase. This confirms the often advocated, but
516.5 = (1 – 0.835) * 100

6 As previously mentioned, gross capital formation and renewable energy consumption are entered as
reciprocals of actual values. Thus the initial calculation resulted in the respective average decreases in
reciprocal values of 17.51 % and 32.72 %, which corresponds to the average increases of 21.23 % and
48.63 % in their actual values respectively. Namely, if the initial calculation results in a decrease of X %
DEEF
in reciprocal value, the equivalent increase in actual value is %.
DEE.F
22
insufficiently practiced fact that the structure of the energy consumed and its
environmental consequences should indeed become a major issue for policy makers,
especially in developed nations. On the other side, GDP has the least impact on
efficiency. Consideration of GDP by subgroups of countries leads to quite the opposite
conclusion from the one concerning renewables, i.e. the more the country is developed,
the lower the required increase. Moreover, it causes no inefficiency at all in eleven
years for developed and in three years for graduated developing countries. More
specifically, the inefficient entities’ unemployment is 26.76 % above target, gross
capital formation is 21.23 % too low, renewable energy consumption is an astonishing
48.63 % below target, energy use is 19.41 % too high, GHG emissions are 23.04 %
higher, while GDP is 8.86 % lower than required, all in average terms. Viewing the
variables’ contribution to inefﬁciency in proportion, as given in Figure 4, renewables
have the largest (33 %) while GDP has the smallest (6 %) share. A stronger impact of
inputs rather than output on inefficiency is one of the predicted consequences of model
orientation selection.
Figure 4. Input-output contribution to inefficiency
Source: Authors’ calculations based on DEA-Solver-Pro results
As previously explained, all EU countries exhibit a certain type of inefficiency.

The question that comes up next is whether it is local or global. The answer lies in the
shares of pure technical and scale inefficiencies. The comparison of the scores
associated with the different types of efficiency (Figures 1 and 2 and the supplement)
shows considerable variety across countries. Thus their average values of 0.856 and
0.600 indicate that the overall inefficiency of EU countries can be mainly attributed to
23
scale inefficiency. To portray this problem, two contrary examples are briefly
introduced. On one side, pure technical scores are extremely higher than scale scores
in the case of Romania during the whole period under analysis. The difference is most
pronounced for the period 1997-2001, meaning that the major part of its overall
inefficiency can be attributed to scale inefficiency. The investigation also reveals
continuous increasing returns to scale, which means that Romania performs at sub-
optimal scale. To achieve the optimal scale, it obviously needs to increase the size of
operations. On the other side, the influence of scale inefficiency on overall inefficiency
is somewhat lower than the influence of pure technical inefficiency for Germany, also
throughout the whole analysed period. To remove the pure technical inefficiency
policy makers should embrace the strategies from benchmarking countries, of course
in the sense of here selected input and output variables. Nevertheless, regardless of
the type and extent of inefficiency, less than optimally demanded outcomes with
respect to the employed resources were produced. The efficiency assessment of
Germany may seem improper with regard to the overall socio-economic condition of
this country. It should therefore be noted that such result is to the greatest extent
related to renewable energy consumption, which also puts this issue in the context of
this research.
However, the following consistent positive trends are noted based on the data on
this indicator. The share of renewables on the EU level has increased from 10 % at the
beginning to 17.5 % at the end of the investigated period. Thanks to a greater relative
shift upward, from 2.2 % to 12.4 %, that Germany managed to make during the same
period, the unfavourable ratio of these shares on the EU and German level declined
from 4.7 to 1.4.
Relatively similar average pure technical efficiency scores and, at the same time,
mutually quite different average technical efficiency scores, on the subgroup level,
result in significant differences between the subgroups of countries regarding their
scale efficiencies. Thus the scale efficiency scores are 0.809, 0.417 and 0.264 for
developed, graduated developing and developing countries respectively. On average,
the highest scale efficiency scores in these three subgroups of countries are achieved
by Denmark, Cyprus and Croatia, respectively.
24
Given the dissimilarity of results associated with different types of efficiency, it

is difficult to point out the most or the least efficient country. Nevertheless, the only
country with all three average efficiency scores above 0.9 is Luxembourg, while
Bulgaria is the only one with both overall and scale efficiency under 0.2.
Based on all these numbers, it can generally be concluded that scale inefficiency
has much stronger effect on the overall inefficiency than pure technical inefficiency,
and the less the country is developed, the stronger this scale effect is. Since all
graduated developing and developing countries exhibit increasing returns to scale,
they need to expand their operations to reach the optimal scale. Is this feasible and, if
so, how? The answers to these two questions lie above all in the concept of empirical
relative efficiency, supported by the fundamental assumption behind the DEA
method. According to this concept, each DMU is assessed in relation to the actual
DMUs and, if inefficient, should take example by those efficient selected to constitute
its reference set. The experimentally supported assumption is that if a DMU can
produce a certain level of output utilizing specific input levels, other DMUs of equal
scale should be capable of doing the same.
The above assertions necessitate a more thorough examination of the reasons for
such outcomes and the responsibility to take adequate actions for enhancing countries’
efficiencies. This primarily implies a more detailed examination of renewable energy
consumption which generally represents the input with the largest contribution to
inefficiency.
4. CONCLUDING REMARKS
The dynamic relative efficiency assessment of the European Union countries was
carried out through the mutual performance comparison of all 28 current member
states classified into three subgroups, covering the period 1995-2012. The analysis was
conducted using input-oriented DEA window analysis models under the assumptions
of both constant and variable returns to scale. The research explores countries’
efficiency and its dynamics over an 18-year period, relying on six energy and non-
energy indicators – five inputs and one output. Although the empirical findings could
25
have not been discussed in detail due to space limitations, some challenging
implications for policy and practice can be drawn. The efficiency scores reveal
significant disparities, both within and among the subgroups of countries, with
different intensities depending on the type of efficiency. Nevertheless, the results
suggest that all countries have certain potential to raise the scores by improving their
inputs. Thus the overall technical inefficiency, averaged over time and across
countries, could be reduced by 48 %. This percentage can be ascribed to pure technical
inefficiencies and to suboptimal scale operations, with the impact of the latter being
more pronounced, particularly in the case of less developed countries. Renewable
energy consumption is the most influential input variable that contributes significantly
to the inefficiency of each country. These facts pose a major threat to sustainability and
an imperative to policy makers to perceive the problems comprehensively and to tailor
their decisions on energy strategies to address contemporary challenges.
To gain a more comprehensive and, at the same time, more detailed insight into
the energy-growth nexus, the analysis relying on the DEA method should be
broadened by covering more countries from Europe and other parts of the world, by
extending the time frame and by including more indicators that would reflect some
additional energy and environmental burdens. In this way, the limitations of the
present study can serve as opportunities for future research.
REFERENCES
1. Apergis, N., Aye, G.C., Barros, C.P., Gupta, R. and Wanke, P. (2015) ‘Energy
efficiency of selected OECD countries: A slacks based model with undesirable
outputs’, Energy Economics, 51, pp.45-53.
2. Banker, R.D., Charnes, A. and Cooper, W.W. (1984) ‘Some models for estimating
technical and scale inefficiencies in data envelopment analysis’, Management
science, 30(9), pp.1078-1092.
3. Ceylan, D. and Gunay, E.N.O. (2010) ‘Energy efficiency trends and policies:
Cross-country comparison in Europe’, paper presented in International
26
Conference of Economic Modelling (ECOMOD), July 7-10, Istanbul. Available at:

https://ecomod.net/sites/default/files/document-
conference/ecomod2010/1292.pdf (Accessed: 1 March 2017)
4. Charnes, A., Cooper, W.W. and Rhodes, E. (1978) ‘Measuring the efficiency of
decision making units’, European journal of operational research, 2(6), pp.429-
444.
5. Chen, W. and Geng, W. (2017) ‘Fossil energy saving and CO 2 emissions

reduction performance, and dynamic change in performance considering
renewable energy input’, Energy, 120, pp.283-292.
6. Chien, T. and Hu, J.L. (2007) ‘Renewable energy and macroeconomic efficiency
of OECD and non-OECD economies’, Energy Policy, 35(7), pp.3606-3615.
7. Cooper, W. W., Seiford, L. M. and Tone, K. (2006) Introduction to data

envelopment analysis and its uses: with DEA-solver software and references. 1st
edn. New York: Springer Science & Business Media.
8. EU Commission. (2013) ‘Delivering the internal electricity market and making

the most of public intervention’, C (2013) 7243 final, Brussels, 5th November 2013.
Available at: https://ec.europa.eu/energy/sites/ener/files/documents/
com_2013_public_ intervention_en_0.pdf (Accessed: 1 March 2017)
9. EU Commission. (2001) ‘Directive 2001/77/EC of the European Parliament and

of the Council of 27 September 2001 on the promotion of electricity produced
from renewable energy sources in the internal electricity market’, Official Journal
of the European Communities, L283, pp.33-40.

of the Council of 8 May 2003 on the promotion of the use of biofuels or other
renewable fuels for transport’, Official Journal of the European Union, L123,
pp.42-46.

of the Council of 23 April 2009 on the promotion of the use of energy from
renewable sources and amending and subsequently repealing Directives
27
2001/77/EC and 2003/30/EC’, Official Journal of the European Union, L140,

pp.16-47.
12. EU Commission. (1997) ‘Energy for the future: renewable sources of

energy’, White Paper for a Community strategy and action plan, COM (97) 599
final, Brussels, 26th November 1997. Available at: http://europa.eu/documents/
comm/white_papers/pdf/com97_599_en.pdf (Accessed: 1 March 2017)
13. EU Commission. (2011) ‘Energy Roadmap 2050’, COM (2011) 885 final, Brussels,
15th December 2011. Available at: http://eur-lex.europa.eu/legal-content/EN/
TXT/PDF/?uri=CELEX:52011DC0885&from=EN (Accessed: 1 March 2017)
14. EU Commission. (2014) ‘Guidelines on state aid for environmental protection

and energy-related objectives for 2014–2020’, Official Journal of the European
Union, C200, pp.2-44.
15. EU Commission. (2012) ‘Renewable Energy: a major player in the European

energy market’, COM (2012) 271 final, Brussels, 6th June 2012. Available at:
http://eur-lex.europa.eu/legal-content/EN/TXT/PDF/
?uri=CELEX:52012DC0271&from=EN (Accessed: 1 March 2017)
16. EU Commission. (2015) ‘Renewable energy progress report’, COM (2015) 293
final, Brussels, 15th June 2015. Available at: http://eur-lex.europa.eu/
resource.html?uri=cellar:4f8722ce-1347-11e5-8817-01aa75ed71a1.0001.02/
DOC_1&format=PDF (Accessed: 1 March 2017)
17. EU Commission. (2007) ‘Renewable Energy Road Map–Renewable energies in

the 21st century: building a more sustainable future’, COM (2006) 848 final,
Brussels, 10th January 2007. Available at: http://eur-lex.europa.eu/legal-
content/EN/TXT/PDF/?uri=CELEX:52006DC0848&from=EN (Accessed: 1
March 2017)
18. European Parliament. (2017) ‘Renewable Energy’, Fact Sheets on the European
Union – 2017, Brussels. Available at: http://www.europarl.europa.eu/ftu/pdf/
en/FTU_5.7.4.pdf (Accessed: 1 March 2017)
28
19. Farrell, M.J. (1957) ‘The measurement of productive efficiency’, Journal of the
Royal Statistical Society. Series A (General), 120(3), pp.253-290.
20. Halkos, G.E. and Tzeremes, N.G. (2013) ‘Renewable energy consumption and
economic efficiency: Evidence from European countries’, Journal of Renewable
and Sustainable Energy, 5(4), 041803.
21. Liu, J.S., Lu, L.Y. and Lu, W.M. (2016) ‘Research fronts in data envelopment
analysis’, Omega, 58, pp.33-45.
22. Menegaki, A.N. (2013) ‘Growth and renewable energy in Europe: Benchmarking
with data envelopment analysis’, Renewable Energy, 60, pp.363-369.
23. Rabar, D. (2013) ‘Assessment of regional efficiency in Croatia using data

envelopment analysis’, Croatian operational research review, 4(1), pp.76-88.
24. Renewable Energy Policy Network for the 21st century – REN21 (2016)
Renewables 2016: Global Status Report. Paris: REN21 Secretariat. Available at:
http://www.ren21.net/wp-content/uploads/2016/10/REN21_GSR2016_
FullReport_en_11.pdf (Accessed: 1 March 2017)
25. Scarlat, N., Dallemand, J.F., Monforti-Ferrario, F., Banja, M. and Motola, V. (2015)
‘Renewable energy policy framework and bioenergy contribution in the
European Union–An overview from National Renewable Energy Action Plans
and Progress Reports’, Renewable and Sustainable Energy Reviews, 51, pp.969-
985.
26. Woo, C., Chung, Y., Chun, D., Seo, H. and Hong, S. (2015) ‘The static and dynamic
environmental efficiency of renewable energy: A Malmquist index analysis of
OECD countries’, Renewable and Sustainable Energy Reviews, 47, pp.367-376.
27. Zhou, P., Ang, B.W. and Poh, K.L. (2007) ‘A mathematical programming
approach to constructing composite indicators’, Ecological economics, 62(2),
pp.291-297.
28. Zhou, P., Ang, B.W. and Poh, K.L. (2008) ‘A survey of data envelopment analysis
in energy and environmental studies’, European Journal of Operational
Research, 189(1), pp.1-18.
29
Rabar, Danijela. “Renewable Energy Consumption-Economic Growth Nexus: A Nonparametric Evidence for Developing, Graduated Developing and
Developed European Union Countries.” Modelling Economic Growth: Issues and New Insights. Juraj Dobrila University of Pula, 2018.
APPENDIX
CCR efficiency scores
Year St.
EU-Country Mean Min. Max.
1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 dev.
Austria 0.725 0.701 0.723 0.740 0.808 0.851 0.832 0.794 0.747 0.760 0.774 0.834 0.913 0.991 0.904 0.922 0.982 1 0.833 0.701 1 0.098
Belgium 0.483 0.475 0.504 0.509 0.537 0.565 0.563 0.567 0.559 0.601 0.633 0.656 0.691 0.691 0.646 0.652 0.681 0.686 0.594 0.475 0.691 0.075
Denmark 0.753 0.699 0.781 0.813 0.849 0.913 0.899 0.906 0.861 0.915 0.961 0.967 1 1 0.948 0.894 0.972 1 0.896 0.699 1 0.089
Finland 0.490 0.494 0.559 0.622 0.632 0.708 0.699 0.681 0.693 0.746 0.815 0.826 0.913 0.914 0.716 0.749 0.827 0.793 0.715 0.490 0.914 0.124
France 0.589 0.568 0.585 0.611 0.645 0.682 0.683 0.682 0.678 0.704 0.722 0.755 0.788 0.785 0.747 0.756 0.813 0.812 0.700 0.568 0.813 0.078
Germany 0.518 0.506 0.523 0.554 0.579 0.601 0.591 0.575 0.566 0.566 0.577 0.603 0.646 0.648 0.598 0.639 0.691 0.685 0.593 0.506 0.691 0.053
Greece 0.497 0.495 0.498 0.491 0.504 0.506 0.513 0.523 0.555 0.563 0.539 0.595 0.619 0.599 0.577 0.581 0.545 0.509 0.539 0.491 0.619 0.041
Ireland 0.521 0.541 0.586 0.617 0.661 0.703 0.723 0.753 0.805 0.857 0.938 1 0.973 0.884 0.824 0.823 0.900 0.890 0.778 0.521 1 0.148
Developed
Italy 0.628 0.637 0.640 0.632 0.633 0.648 0.658 0.665 0.645 0.655 0.657 0.681 0.704 0.702 0.714 0.711 0.726 0.710 0.669 0.628 0.726 0.033
Luxembourg 0.733 0.718 0.810 0.896 0.961 1 1 0.954 0.932 0.927 0.954 0.898 1 1 0.919 0.953 0.989 0.978 0.923 0.718 1 0.087
Netherlands 0.523 0.536 0.574 0.610 0.668 0.734 0.794 0.760 0.687 0.666 0.687 0.747 0.828 0.886 0.811 0.741 0.792 0.765 0.712 0.523 0.886 0.102
Portugal 0.481 0.503 0.499 0.489 0.484 0.503 0.508 0.482 0.481 0.476 0.468 0.500 0.509 0.535 0.515 0.551 0.549 0.556 0.505 0.468 0.556 0.027
Spain 0.500 0.521 0.509 0.523 0.542 0.561 0.581 0.579 0.590 0.598 0.613 0.650 0.661 0.682 0.644 0.656 0.653 0.644 0.595 0.500 0.682 0.059
Sweden 0.608 0.599 0.635 0.675 0.712 0.840 0.908 0.889 0.842 0.838 0.860 0.931 1 1 0.961 0.957 1 1 0.848 0.599 1 0.142
United Kingdom 0.456 0.454 0.474 0.497 0.529 0.558 0.580 0.602 0.620 0.645 0.664 0.668 0.698 0.703 0.672 0.658 0.718 0.706 0.606 0.454 0.718 0.091
Mean 0.567 0.563 0.593 0.619 0.650 0.692 0.702 0.694 0.684 0.701 0.724 0.754 0.796 0.801 0.746 0.749 0.789 0.782 0.700
Minimum 0.456 0.454 0.474 0.489 0.484 0.503 0.508 0.482 0.481 0.476 0.468 0.500 0.509 0.535 0.515 0.551 0.545 0.509 0.454
Maximum 0.753 0.718 0.810 0.896 0.961 1 1 0.954 0.932 0.927 0.961 1 1 1 0.961 0.957 1 1 1
Stand. deviation 0.101 0.088 0.106 0.121 0.136 0.152 0.154 0.145 0.130 0.136 0.155 0.150 0.162 0.161 0.142 0.133 0.155 0.162 0.155
Cyprus 0.666 0.609 0.655 0.626 0.652 0.654 0.705 0.737 0.689 0.762 0.762 0.780 0.783 0.767 0.736 0.747 0.762 0.781 0.715 0.609 0.783 0.058
Czech Republic 0.237 0.248 0.236 0.233 0.242 0.253 0.258 0.256 0.255 0.265 0.283 0.302 0.327 0.335 0.309 0.308 0.319 0.316 0.277 0.233 0.335 0.035
Estonia 0.127 0.145 0.187 0.188 0.166 0.187 0.205 0.236 0.268 0.283 0.298 0.361 0.401 0.326 0.237 0.241 0.281 0.325 0.248 0.127 0.401 0.076
Graduated developing
Latvia 0.149 0.155 0.172 0.194 0.213 0.235 0.237 0.246 0.261 0.279 0.314 0.356 0.394 0.364 0.302 0.283 0.318 0.324 0.266 0.149 0.394 0.072
Lithuania 0.119 0.116 0.133 0.151 0.168 0.188 0.196 0.199 0.219 0.227 0.239 0.260 0.297 0.286 0.243 0.282 0.289 0.297 0.217 0.116 0.297 0.062
Malta 0.420 0.443 0.415 0.495 0.476 0.566 0.489 0.541 0.492 0.492 0.481 0.518 0.512 0.649 0.680 0.676 0.696 0.734 0.543 0.415 0.734 0.100
Slovak Republic 0.163 0.193 0.204 0.215 0.201 0.203 0.220 0.228 0.227 0.247 0.268 0.290 0.323 0.337 0.297 0.313 0.332 0.328 0.255 0.163 0.337 0.056
Slovenia 0.289 0.287 0.293 0.310 0.336 0.351 0.343 0.345 0.362 0.387 0.397 0.424 0.469 0.479 0.406 0.397 0.399 0.372 0.369 0.287 0.479 0.057
Mean 0.271 0.274 0.287 0.301 0.307 0.330 0.332 0.348 0.347 0.368 0.380 0.412 0.438 0.443 0.401 0.406 0.424 0.435 0.361
Minimum 0.119 0.116 0.133 0.151 0.166 0.187 0.196 0.199 0.219 0.227 0.239 0.260 0.297 0.286 0.237 0.241 0.281 0.297 0.116
Maximum 0.666 0.609 0.655 0.626 0.652 0.654 0.705 0.737 0.689 0.762 0.762 0.780 0.783 0.767 0.736 0.747 0.762 0.781 0.783
Stand. deviation 0.189 0.171 0.172 0.170 0.174 0.182 0.180 0.191 0.165 0.182 0.173 0.170 0.158 0.176 0.197 0.195 0.192 0.201 0.179
Bulgaria 0.074 0.075 0.082 0.088 0.091 0.094 0.094 0.103 0.105 0.115 0.122 0.132 0.146 0.162 0.159 0.152 0.144 0.151 0.116 0.074 0.162 0.030
Croatia 0.301 0.310 0.307 0.303 0.303 0.322 0.326 0.331 0.327 0.340 0.351 0.366 0.372 0.393 0.373 0.373 0.378 0.394 0.343 0.301 0.394 0.033
Hungary 0.196 0.191 0.205 0.224 0.227 0.244 0.245 0.256 0.258 0.282 0.285 0.299 0.307 0.313 0.316 0.314 0.329 0.334 0.268 0.191 0.334 0.047
Developing
Poland 0.136 0.139 0.151 0.169 0.181 0.197 0.198 0.204 0.206 0.216 0.221 0.223 0.243 0.249 0.263 0.256 0.267 0.281 0.211 0.136 0.281 0.044
Romania 0.128 0.130 0.132 0.141 0.159 0.163 0.169 0.172 0.175 0.193 0.202 0.211 0.226 0.250 0.260 0.257 0.254 0.262 0.194 0.128 0.262 0.048
Mean 0.167 0.169 0.175 0.185 0.192 0.204 0.206 0.213 0.214 0.229 0.236 0.246 0.259 0.273 0.274 0.270 0.274 0.284 0.226
Minimum 0.074 0.075 0.082 0.088 0.091 0.094 0.094 0.103 0.105 0.115 0.122 0.132 0.146 0.162 0.159 0.152 0.144 0.151 0.074
Maximum 0.301 0.310 0.307 0.303 0.303 0.322 0.326 0.331 0.327 0.340 0.351 0.366 0.372 0.393 0.373 0.373 0.378 0.394 0.394
Stand. deviation 0.086 0.089 0.086 0.082 0.079 0.086 0.086 0.086 0.084 0.086 0.087 0.089 0.085 0.086 0.079 0.082 0.088 0.091 0.086
Mean 0.411 0.410 0.431 0.451 0.470 0.501 0.508 0.509 0.504 0.522 0.539 0.565 0.598 0.605 0.563 0.566 0.593 0.594 0.519
Minimum 0.074 0.075 0.082 0.088 0.091 0.094 0.094 0.103 0.105 0.115 0.122 0.132 0.146 0.162 0.159 0.152 0.144 0.151 0.074
EU
Maximum 0.753 0.718 0.810 0.896 0.961 1 1 0.954 0.932 0.927 0.961 1 1 1 0.961 0.957 1 1 1
Stand. deviation 0.214 0.205 0.218 0.227 0.242 0.259 0.262 0.254 0.241 0.245 0.255 0.257 0.268 0.269 0.252 0.250 0.268 0.265 0.252
30
BCC efficiency scores

Year St.
1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 dev.
Austria 0.999 0.950 0.945 0.946 0.989 1 0.973 0.942 0.908 0.888 0.887 0.919 0.977 1 0.964 0.951 0.995 1 0.957 0.887 1 0.038
Belgium 0.667 0.660 0.688 0.693 0.715 0.753 0.739 0.694 0.693 0.748 0.788 0.806 0.835 0.861 0.755 0.780 0.819 0.804 0.750 0.660 0.861 0.061
Denmark 0.817 0.755 0.834 0.861 0.882 0.938 0.919 0.926 0.882 0.929 0.966 0.967 1 1 0.961 0.898 0.972 1 0.917 0.755 1 0.069
Finland 0.691 0.685 0.737 0.783 0.774 0.839 0.820 0.788 0.797 0.838 0.906 0.899 0.971 0.965 0.792 0.816 0.891 0.889 0.827 0.685 0.971 0.082
France 0.776 0.743 0.757 0.773 0.802 0.831 0.825 0.823 0.817 0.837 0.853 0.884 0.915 0.913 0.883 0.890 0.947 0.943 0.845 0.743 0.947 0.063
Germany 0.715 0.691 0.702 0.730 0.746 0.766 0.737 0.714 0.706 0.701 0.709 0.727 0.767 0.765 0.737 0.759 0.814 0.812 0.739 0.691 0.814 0.036
Greece 0.907 0.904 0.880 0.882 0.881 0.872 0.857 0.854 0.871 0.859 0.825 0.876 0.893 0.868 0.825 0.845 0.823 0.812 0.863 0.812 0.907 0.028
Ireland 0.725 0.736 0.758 0.780 0.795 0.851 0.896 0.865 0.882 0.909 0.978 1 0.990 0.943 0.895 0.874 0.950 0.957 0.877 0.725 1 0.087
Developed
Italy 0.819 0.818 0.818 0.814 0.824 0.829 0.836 0.844 0.824 0.830 0.831 0.854 0.878 0.884 0.902 0.898 0.915 0.899 0.851 0.814 0.915 0.035
Luxembourg 0.844 0.781 0.916 0.957 1 1 1 0.962 0.955 0.960 0.973 0.905 1 1 0.959 0.973 1 1 0.955 0.781 1 0.060
Netherlands 0.683 0.717 0.760 0.808 0.874 0.941 1 0.942 0.835 0.775 0.786 0.849 0.935 1 0.923 0.814 0.860 0.819 0.851 0.683 1 0.092
Portugal 0.984 1 1 1 0.981 1 1 0.934 0.917 0.897 0.864 0.901 0.916 0.964 0.927 0.956 0.942 0.970 0.953 0.864 1 0.043
Spain 0.840 0.859 0.833 0.825 0.830 0.837 0.850 0.844 0.853 0.859 0.887 0.930 0.937 0.941 0.912 0.932 0.919 0.904 0.877 0.825 0.941 0.042
Sweden 0.787 0.748 0.797 0.823 0.838 0.970 1 0.975 0.930 0.901 0.892 0.942 1 1 1 0.972 1 1 0.921 0.748 1 0.086
United Kingdom 0.643 0.640 0.653 0.675 0.697 0.724 0.758 0.763 0.777 0.801 0.810 0.793 0.819 0.827 0.799 0.784 0.844 0.828 0.758 0.640 0.844 0.068
Mean 0.793 0.779 0.805 0.823 0.842 0.877 0.880 0.858 0.843 0.849 0.864 0.883 0.922 0.929 0.882 0.876 0.913 0.909 0.863
Minimum 0.643 0.640 0.653 0.675 0.697 0.724 0.737 0.694 0.693 0.701 0.709 0.727 0.767 0.765 0.737 0.759 0.814 0.804 0.640
Maximum 0.999 1 1 1 1 1 1 0.975 0.955 0.960 0.978 1 1 1 1 0.973 1 1 1
Stand. deviation 0.110 0.107 0.099 0.094 0.095 0.093 0.098 0.090 0.077 0.070 0.076 0.070 0.073 0.074 0.082 0.073 0.067 0.078 0.093
Cyprus 1 0.946 1 0.926 0.939 0.918 0.975 1 0.928 1 1 1 1 1 1 1 1 1 0.980 0.918 1 0.032
Czech Republic 0.973 1 0.894 0.780 0.707 0.765 0.771 0.746 0.720 0.725 0.727 0.763 0.880 0.933 0.733 0.717 0.741 0.720 0.794 0.707 1 0.096
Estonia 0.690 0.701 0.831 0.815 0.661 0.694 0.725 0.797 0.849 0.831 0.805 0.970 1 0.869 0.645 0.661 0.701 0.738 0.777 0.645 1 0.104
Latvia 0.958 0.946 0.918 0.977 1 1 1 1 0.997 0.999 1 1 1 0.998 0.987 0.921 0.956 1 0.981 0.918 1 0.028
Lithuania 0.761 0.702 0.764 0.781 0.821 0.815 0.805 0.829 0.838 0.836 0.842 0.948 1 0.873 0.719 0.769 0.790 0.770 0.815 0.702 1 0.073
Malta 0.999 0.971 0.872 0.964 0.891 1 0.886 0.970 0.871 0.873 0.847 0.893 0.864 1 1 1 1 1 0.939 0.847 1 0.061
Slovak Republic 0.650 0.865 0.867 0.845 0.705 0.687 0.755 0.740 0.678 0.702 0.718 0.721 0.729 0.728 0.693 0.688 0.714 0.709 0.733 0.650 0.867 0.063
Slovenia 0.739 0.725 0.743 0.720 0.750 0.771 0.803 0.756 0.752 0.805 0.792 0.846 0.967 0.997 0.767 0.718 0.703 0.702 0.781 0.702 0.997 0.083
Mean 0.846 0.857 0.861 0.851 0.809 0.831 0.840 0.855 0.829 0.847 0.841 0.893 0.930 0.925 0.818 0.809 0.826 0.830 0.850
Minimum 0.650 0.701 0.743 0.720 0.661 0.687 0.725 0.740 0.678 0.702 0.718 0.721 0.729 0.728 0.645 0.661 0.701 0.702 0.645
Maximum 1 1 1 0.977 1 1 1 1 0.997 1 1 1 1 1 1 1 1 1 1
Stand. deviation 0.150 0.128 0.083 0.095 0.124 0.127 0.103 0.116 0.107 0.110 0.109 0.107 0.099 0.097 0.151 0.142 0.136 0.142 0.117
Bulgaria 0.609 0.573 0.620 0.697 0.747 0.735 0.739 0.731 0.737 0.763 0.780 0.804 0.835 0.956 0.840 0.754 0.703 0.727 0.742 0.573 0.956 0.089
Croatia 1 1 1 0.937 0.945 0.940 0.944 0.961 0.947 0.950 0.954 0.973 0.946 0.985 0.945 0.920 0.902 0.938 0.955 0.902 1 0.027
Hungary 0.773 0.764 0.793 0.822 0.838 0.879 0.902 0.902 0.877 0.881 0.824 0.815 0.828 0.830 0.838 0.824 0.844 0.854 0.838 0.764 0.902 0.040
Developing
Poland 0.688 0.695 0.723 0.774 0.790 0.798 0.744 0.724 0.717 0.735 0.726 0.728 0.771 0.764 0.744 0.717 0.731 0.737 0.739 0.688 0.798 0.030
Romania 0.848 0.874 0.971 0.997 1 1 1 0.948 0.936 0.940 0.941 0.945 0.987 1 1 1 0.988 0.988 0.965 0.848 1 0.045
Mean 0.784 0.781 0.822 0.846 0.864 0.870 0.866 0.853 0.843 0.854 0.845 0.853 0.873 0.907 0.873 0.843 0.834 0.849 0.848
Minimum 0.609 0.573 0.620 0.697 0.747 0.735 0.739 0.724 0.717 0.735 0.726 0.728 0.771 0.764 0.744 0.717 0.703 0.727 0.573
Maximum 1 1 1 0.997 1 1 1 0.961 0.947 0.950 0.954 0.973 0.987 1 1 1 0.988 0.988 1
Stand. deviation 0.151 0.164 0.162 0.121 0.106 0.106 0.119 0.117 0.109 0.100 0.100 0.103 0.090 0.104 0.100 0.117 0.119 0.117 0.111
Mean 0.807 0.802 0.824 0.835 0.836 0.863 0.866 0.856 0.839 0.849 0.854 0.881 0.916 0.924 0.862 0.851 0.874 0.876 0.856
Minimum 0.609 0.573 0.620 0.675 0.661 0.687 0.725 0.694 0.678 0.701 0.709 0.721 0.729 0.728 0.645 0.661 0.701 0.702 0.573
EU
Maximum 1 1 1 1 1 1 1 1 0.997 1 1 1 1 1 1 1 1 1 1
Stand. deviation 0.127 0.124 0.106 0.096 0.103 0.104 0.100 0.098 0.089 0.085 0.087 0.085 0.083 0.083 0.108 0.104 0.105 0.109 0.104
31
Scale efficiency scores

Year St.
1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 dev.
Austria 0.726 0.738 0.766 0.782 0.818 0.851 0.856 0.843 0.823 0.856 0.872 0.907 0.934 0.991 0.937 0.969 0.987 1 0.870 0.726 1 0.087
Belgium 0.724 0.720 0.732 0.734 0.751 0.751 0.763 0.818 0.807 0.803 0.804 0.814 0.827 0.803 0.856 0.835 0.832 0.853 0.790 0.720 0.856 0.045
Denmark 0.922 0.925 0.936 0.944 0.962 0.974 0.978 0.979 0.976 0.985 0.995 0.999 1 1 0.986 0.995 0.999 1 0.975 0.922 1 0.027
Finland 0.709 0.722 0.758 0.794 0.817 0.844 0.853 0.863 0.869 0.891 0.899 0.918 0.940 0.947 0.904 0.918 0.928 0.891 0.859 0.709 0.947 0.073
France 0.759 0.764 0.773 0.790 0.805 0.821 0.828 0.829 0.830 0.841 0.846 0.854 0.861 0.859 0.846 0.850 0.858 0.861 0.826 0.759 0.861 0.034
Germany 0.725 0.733 0.746 0.759 0.776 0.784 0.802 0.805 0.801 0.808 0.814 0.829 0.843 0.847 0.811 0.841 0.849 0.844 0.801 0.725 0.849 0.040
Greece 0.548 0.548 0.566 0.557 0.572 0.580 0.598 0.612 0.637 0.656 0.654 0.679 0.694 0.689 0.699 0.687 0.662 0.627 0.626 0.548 0.699 0.054
Ireland 0.719 0.735 0.774 0.790 0.831 0.827 0.807 0.870 0.913 0.943 0.960 1 0.983 0.938 0.920 0.941 0.947 0.930 0.879 0.719 1 0.087
Developed
Italy 0.768 0.779 0.782 0.777 0.768 0.782 0.787 0.788 0.783 0.789 0.791 0.798 0.802 0.794 0.791 0.792 0.793 0.790 0.786 0.768 0.802 0.009
Luxembourg 0.869 0.920 0.884 0.936 0.961 1 1 0.991 0.976 0.965 0.980 0.992 1 1 0.959 0.979 0.989 0.978 0.966 0.869 1 0.039
Netherlands 0.765 0.748 0.755 0.756 0.764 0.780 0.794 0.807 0.822 0.860 0.875 0.880 0.886 0.886 0.878 0.910 0.921 0.935 0.834 0.748 0.935 0.064
Portugal 0.489 0.503 0.499 0.489 0.494 0.503 0.508 0.517 0.525 0.531 0.542 0.555 0.555 0.555 0.555 0.576 0.582 0.573 0.531 0.489 0.582 0.032
Spain 0.596 0.607 0.611 0.634 0.653 0.670 0.683 0.687 0.692 0.696 0.691 0.698 0.706 0.725 0.706 0.704 0.710 0.712 0.677 0.596 0.725 0.040
Sweden 0.773 0.800 0.796 0.821 0.850 0.866 0.908 0.912 0.905 0.930 0.964 0.989 1 1 0.961 0.985 1 1 0.915 0.773 1 0.079
United Kingdom 0.710 0.710 0.725 0.736 0.758 0.770 0.765 0.789 0.798 0.805 0.820 0.842 0.853 0.850 0.841 0.840 0.851 0.853 0.795 0.710 0.853 0.052
Mean 0.720 0.730 0.740 0.753 0.772 0.787 0.795 0.807 0.810 0.824 0.834 0.850 0.859 0.859 0.843 0.855 0.861 0.857 0.809
Minimum 0.489 0.503 0.499 0.489 0.494 0.503 0.508 0.517 0.525 0.531 0.542 0.555 0.555 0.555 0.555 0.576 0.582 0.573 0.489
Maximum 0.922 0.925 0.936 0.944 0.962 1 1 0.991 0.976 0.985 0.995 1 1 1 0.986 0.995 1 1 1
Stand. deviation 0.110 0.114 0.111 0.121 0.125 0.130 0.129 0.126 0.121 0.123 0.128 0.130 0.130 0.131 0.118 0.123 0.129 0.133 0.129
Cyprus 0.666 0.644 0.655 0.676 0.694 0.712 0.722 0.737 0.742 0.762 0.762 0.780 0.783 0.767 0.736 0.747 0.762 0.781 0.729 0.644 0.783 0.045
Czech Republic 0.243 0.248 0.264 0.299 0.342 0.331 0.335 0.343 0.354 0.366 0.390 0.397 0.371 0.359 0.422 0.430 0.431 0.439 0.353 0.243 0.439 0.061
Estonia 0.184 0.207 0.225 0.231 0.251 0.270 0.282 0.296 0.316 0.341 0.370 0.372 0.401 0.375 0.368 0.365 0.400 0.441 0.316 0.184 0.441 0.076
Latvia 0.155 0.164 0.187 0.198 0.213 0.235 0.237 0.246 0.262 0.279 0.314 0.356 0.394 0.365 0.306 0.307 0.332 0.324 0.271 0.155 0.394 0.071
Lithuania 0.156 0.165 0.174 0.193 0.205 0.231 0.243 0.240 0.262 0.272 0.284 0.275 0.297 0.328 0.338 0.367 0.366 0.386 0.266 0.156 0.386 0.072
Malta 0.420 0.456 0.476 0.514 0.534 0.566 0.553 0.558 0.565 0.564 0.568 0.580 0.592 0.649 0.680 0.676 0.696 0.734 0.577 0.420 0.734 0.085
Slovak Republic 0.251 0.223 0.236 0.254 0.286 0.296 0.291 0.309 0.334 0.351 0.374 0.402 0.443 0.463 0.428 0.455 0.465 0.463 0.351 0.223 0.465 0.087
Slovenia 0.392 0.396 0.394 0.430 0.449 0.455 0.427 0.456 0.482 0.480 0.501 0.502 0.485 0.481 0.529 0.553 0.567 0.529 0.473 0.392 0.567 0.053
Mean 0.308 0.313 0.326 0.349 0.372 0.387 0.386 0.398 0.415 0.427 0.445 0.458 0.471 0.473 0.476 0.487 0.502 0.512 0.417
Minimum 0.155 0.164 0.174 0.193 0.205 0.231 0.237 0.240 0.262 0.272 0.284 0.275 0.297 0.328 0.306 0.307 0.332 0.324 0.155
Maximum 0.666 0.644 0.655 0.676 0.694 0.712 0.722 0.737 0.742 0.762 0.762 0.780 0.783 0.767 0.736 0.747 0.762 0.781 0.783
Stand. deviation 0.176 0.171 0.169 0.175 0.174 0.175 0.172 0.175 0.169 0.168 0.158 0.160 0.153 0.157 0.159 0.157 0.158 0.163 0.169
Bulgaria 0.121 0.132 0.133 0.127 0.122 0.128 0.128 0.141 0.143 0.151 0.156 0.164 0.175 0.170 0.189 0.201 0.204 0.207 0.155 0.121 0.207 0.030
Croatia 0.301 0.310 0.307 0.323 0.321 0.342 0.346 0.345 0.345 0.358 0.368 0.376 0.393 0.399 0.395 0.406 0.419 0.420 0.360 0.301 0.420 0.039
Hungary 0.254 0.251 0.259 0.272 0.270 0.277 0.271 0.283 0.294 0.320 0.346 0.367 0.370 0.376 0.377 0.382 0.390 0.391 0.319 0.251 0.391 0.054
Developing
Poland 0.198 0.199 0.208 0.218 0.228 0.247 0.266 0.282 0.288 0.294 0.304 0.307 0.315 0.326 0.354 0.357 0.365 0.380 0.285 0.198 0.380 0.059
Romania 0.151 0.149 0.135 0.141 0.159 0.163 0.169 0.181 0.187 0.206 0.215 0.223 0.230 0.250 0.260 0.257 0.257 0.265 0.200 0.135 0.265 0.046
Mean 0.205 0.208 0.209 0.216 0.220 0.232 0.236 0.246 0.251 0.266 0.278 0.287 0.297 0.304 0.315 0.321 0.327 0.333 0.264
Minimum 0.121 0.132 0.133 0.127 0.122 0.128 0.128 0.141 0.143 0.151 0.156 0.164 0.175 0.170 0.189 0.201 0.204 0.207 0.121
Maximum 0.301 0.310 0.307 0.323 0.321 0.342 0.346 0.345 0.345 0.358 0.368 0.376 0.393 0.399 0.395 0.406 0.419 0.420 0.420
Stand. deviation 0.074 0.073 0.076 0.084 0.081 0.086 0.087 0.083 0.083 0.085 0.090 0.092 0.093 0.094 0.088 0.088 0.092 0.092 0.089
Mean 0.511 0.518 0.527 0.542 0.559 0.573 0.579 0.590 0.598 0.611 0.624 0.638 0.648 0.650 0.644 0.654 0.663 0.665 0.600
Minimum 0.121 0.132 0.133 0.127 0.122 0.128 0.128 0.141 0.143 0.151 0.156 0.164 0.175 0.170 0.189 0.201 0.204 0.207 0.121
EU
Maximum 0.922 0.925 0.936 0.944 0.962 1 1 0.991 0.976 0.985 0.995 1 1 1 0.986 0.995 1 1 1
Stand. deviation 0.263 0.265 0.266 0.268 0.272 0.274 0.276 0.276 0.271 0.271 0.269 0.272 0.270 0.269 0.256 0.258 0.258 0.255 0.267
32
Inefficiency sources and average amounts in the BCC model

Proposed input and output improvements per inefficient country (%)
Year 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012
Developed -35.5 -39.5 -36.2 -35.1 -29.8 -28.8 -26.6 -21.7 -24.1 -24.1 -20.8 -20.1 -17.7 -19.0 -20.9 -20.6 -20.6 -30.2
Graduated developing -29.8 -25.3 -28.3 -22.8 -29.0 -33.6 -28.6 -29.0 -23.9 -25.7 -27.8 -16.1 -18.3 -11.7 -26.5 -33.7 -33.3 -31.7
Unemployment
Developing -31.9 -32.4 -27.2 -19.8 -24.9 -33.4 -43.9 -33.5 -36.2 -36.9 -34.9 -30.4 -21.5 -18.6 -21.0 -24.7 -23.2 -24.4
EU -33.6 -34.4 -32.6 -28.6 -28.8 -30.9 -30.4 -25.7 -26.2 -26.9 -25.2 -21.0 -18.7 -16.7 -22.3 -24.4 -24.3 -29.2
Developed 26.1 31.0 26.4 23.4 20.6 18.6 19.5 17.6 19.2 18.3 16.0 14.3 10.8 12.0 17.5 15.2 12.8 19.0
Gross
Graduated developing 36.9 24.7 20.7 21.1 30.0 31.6 29.2 31.9 22.9 22.7 26.8 14.0 16.3 8.6 43.7 39.1 30.3 40.7
capital
Developing 37.1 70.1 32.1 18.3 20.5 19.3 20.2 17.2 18.7 17.1 18.3 17.2 14.5 13.1 18.8 24.4 19.9 17.8
formation
EU 30.2 34.0 25.6 21.7 23.1 22.1 22.5 20.6 20.1 19.2 18.8 14.8 12.7 12.2 23.3 21.7 18.3 23.2
Developed 55.9 70.7 61.4 55.8 57.2 81.6 88.1 60.5 63.7 65.0 58.7 55.7 43.8 54.4 24.6 26.2 23.9 23.2
Renewable
Inputs energy
Developing 153.3 99.9 76.0 52.9 71.4 63.2 36.6 36.7 27.9 31.1 29.8 38.2 30.6 18.0 23.2 27.5 22.7 17.8
consumption
EU 68.8 71.7 65.4 55.0 63.4 76.5 66.4 57.1 51.6 54.0 52.2 43.2 36.3 35.5 26.1 28.2 25.2 24.9
Developed -21.4 -25.2 -22.9 -21.1 -18.2 -18.8 -20.3 -15.7 -18.1 -18.5 -16.8 -15.3 -13.7 -13.5 -13.8 -14.2 -12.1 -13.4
Energy Graduated developing -22.3 -21.7 -23.2 -20.6 -25.3 -27.5 -26.6 -27.1 -25.1 -28.7 -32.2 -21.9 -19.3 -15.8 -25.5 -26.5 -26.2 -31.1
use Developing -27.1 -27.4 -22.3 -15.4 -17.0 -16.2 -18.1 -15.9 -16.9 -15.7 -16.2 -15.1 -13.1 -12.4 -16.9 -20.7 -17.4 -15.6
EU -22.5 -24.6 -22.9 -19.9 -20.0 -20.7 -21.9 -18.4 -19.9 -20.6 -20.3 -17.1 -14.6 -14.0 -17.2 -18.2 -16.7 -18.1
Developed -27.9 -30.9 -28.6 -24.9 -22.0 -21.7 -23.6 -19.1 -21.0 -19.9 -19.0 -16.7 -15.4 -17.3 -17.7 -17.8 -16.6 -19.4
GHG Graduated developing -31.4 -29.7 -31.3 -22.6 -27.7 -31.0 -27.0 -26.6 -23.1 -26.3 -29.3 -17.6 -14.0 -12.2 -27.9 -33.8 -33.3 -37.8
emissions Developing -34.6 -32.3 -27.8 -19.2 -23.8 -24.4 -23.2 -18.8 -19.8 -20.0 -20.3 -21.5 -21.2 -20.0 -22.4 -27.6 -23.8 -22.8
EU -29.8 -30.8 -29.2 -23.2 -23.9 -24.7 -24.6 -20.8 -21.4 -21.6 -21.7 -17.9 -16.5 -16.3 -21.0 -23.2 -22.3 -24.6
Developed 1.2 0.0 1.1 0.5 0.5 1.2 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.1 0.0 0.2
Output GDP
Developing 59.1 46.5 65.8 48.4 36.8 38.1 42.1 37.8 29.1 26.7 23.5 28.1 31.1 34.8 14.2 6.4 4.6 4.2
EU 21.8 16.8 17.9 15.3 10.9 10.7 10.1 8.8 5.8 5.2 4.6 6.6 7.4 10.1 2.6 1.9 1.0 1.1
33

Renewable Energy Consumption-Economic Growth Nexus-Researchgate

Hochgeladen von

Dokumentinformationen

Originaltitel

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

Renewable Energy Consumption-Economic Growth Nexus-Researchgate

Hochgeladen von

Copyright:

Verfügbare Formate

Rabar, Danijela.

“Renewable Energy Consumption-Economic Growth Nexus: A Nonparametric

RENEWABLE ENERGY CONSUMPTION-ECONOMIC

This study empirically investigates the relationship between economic growth

Today’s world is confronted with two major concerns: an increase in energy-

Table 1. A brief survey of EU policy framework for renewable energy

1. REVIEW OF LITERATURE ON THE ROLE OF

In recent years, a number of empirical economic scientific studies have focused

Halkos and Tzeremes (2013) examined the relationship between renewable

Using energy environmental and non-radial Malmquist indices1, an empirical

2. MODEL SETUP – INDICATORS, DATA AND

A different selection of input and output variables leads to different efficiency

Global warming, caused by the rapidly expanding emissions of greenhouse gases

2.2. Data Collection

The nature of the selected variables allows for comparisons to be made on an

Renewable Energy Sources policies across EU member states is not regarded as

Table 2. Data summary statistics, 1995-2012

Source: Authors’ calculations based on data from the World Bank

2.3. Methodological framework

DEA is a non-parametric performance measurement technique commonly

representation in this paper. However, it is considered important to explore the extent

first three conditions consist of m, s and n constraints, respectively. In our study, m is

Definition 2 (Reference Set):

The optimal solution is given with

It can be noted, based on these relationships, that the efficiency of can be

Definition 3 (Scale Efficiency):

Technical Efficiency = Pure Technical Efficiency × Scale Efficiency,

proposing an answer whether the overall technical inefficiency (TE) is caused by

3. MODEL APPLICATION AND EMPIRICAL ANALYSIS

The relative efficiency assessment of 28 EU current member states was carried

Table 3. Summary statistics on input-based efficiency scores, 1995-2012

most PTE Austria Latvia Romania Latvia

Figure 1. Relative efficiency scores averaged by country

Source: Authors’ work based on DEA-Solver-Pro calculations

Source: Authors’ work based on DEA-Solver-Pro calculations

Figure 3. Differences between the subgroups of countries based on average scores

Source: Authors’ work based on DEA-Solver-Pro calculations

Table 4. Reference set membership in the model BCC-I-W1

Source: Authors’ calculations using DEA-Solver-Pro

Table 5. Summary statistics on inefficiency sources in the model BCC-I-W1, 1995-2012

Renewables obviously represent a major source of inefficiency during the whole

516.5 = (1 – 0.835) * 100

Figure 4. Input-output contribution to inefficiency

Source: Authors’ calculations based on DEA-Solver-Pro results

As previously explained, all EU countries exhibit a certain type of inefficiency.

Given the dissimilarity of results associated with different types of efficiency, it

Conference of Economic Modelling (ECOMOD), July 7-10, Istanbul. Available at:

5. Chen, W. and Geng, W. (2017) ‘Fossil energy saving and CO 2 emissions

7. Cooper, W. W., Seiford, L. M. and Tone, K. (2006) Introduction to data

8. EU Commission. (2013) ‘Delivering the internal electricity market and making

9. EU Commission. (2001) ‘Directive 2001/77/EC of the European Parliament and

10. EU Commission. (2003) ‘Directive 2003/30/EC of the European Parliament and

11. EU Commission. (2009) ‘Directive 2009/28/EC of the European Parliament and

2001/77/EC and 2003/30/EC’, Official Journal of the European Union, L140,

12. EU Commission. (1997) ‘Energy for the future: renewable sources of

14. EU Commission. (2014) ‘Guidelines on state aid for environmental protection

15. EU Commission. (2012) ‘Renewable Energy: a major player in the European

17. EU Commission. (2007) ‘Renewable Energy Road Map–Renewable energies in

23. Rabar, D. (2013) ‘Assessment of regional efficiency in Croatia using data

BCC efficiency scores

Scale efficiency scores

Inefficiency sources and average amounts in the BCC model

Das könnte Ihnen auch gefallen