Review on Energy Resilience

Energy Conversion and Management 92 (2015) 459–468

Contents lists available at ScienceDirect

Energy Conversion and Management

j o u r n a l h o m e p a g e : w w w . e l s e v i e r . c o m / l o c a t e / e n c o n m a n

Analyzing driving forces behind changes in energy vulnerability of Spanish electricity generation through a Divisia index-based method

http://dx.doi.org/10.1016/j.enconman.2014.12.083 0196-8904/� 2015 Elsevier Ltd. All rights reserved.

⇑ Corresponding author at: Department of Applied Economics, Faculty of Econ- omy and Business, University of Oviedo, Avda. del Cristo s/n, 33006 Oviedo, Spain. Tel.: +34 985103747; fax: +34 985105050.

E-mail addresses: [email protected] (P. Fernández González), morenob@uniovi. es (B. Moreno).

P. Fernández González a,⇑, B. Moreno b a Faculty of Economics and Business, Department of Applied Economics, University of Oviedo, Spain b Regional Economics Laboratory – REGIOlab, Department of Applied Economics, University of Oviedo, Spain

a r t i c l e i n f o a b s t r a c t

Article history: Received 31 October 2014 Accepted 27 December 2014 Available online 16 January 2015

Keywords: Electricity generation Energy fossil-fuel dependence Vulnerability Decomposition analysis Spain Divisia index

High energy dependence on fossil raises vulnerability concerns about security of supply and energy cost. This research examines the impact of high dependence of imported fuels for power generation in Spain through the quantification and analysis of the driving forces behind the change in its electricity bill. Fol- lowing logarithmic mean Divisia indexes approach, we present and perform a new method that enables a complete decomposition of changes in electricity vulnerability into contributions from several drivers. In fact, we identify five predefined factors behind the variations in vulnerability in Spain during the 1998– 2011 period: fuel price, average heat rate, fuel dependence, degree of electricity importance and energy intensity.

The application of this approach reveals a significant increase in Spanish vulnerability in the last two decades, promoted by increments in fuel price and importance of electricity over the primary energy con- sumption, but especially by increasing fuel dependence (particularly gas dependence). Therefore, findings mainly advocate for those strategies aimed at reducing Spanish energy dependence. Also those improving thermal efficiency and energy intensity are indicated.

� 2015 Elsevier Ltd. All rights reserved.

1. Introduction

It is well-known that energy is a key contributor to economic prosperity and military strength. It is a truly strategic commodity, leading governments to significant efforts to ensure its availability. Insufficient domestic energy supply, specific technical features of some resources or inadequate financial support may expose coun- tries to energy supply insecurity.

The case of Spain is emphasized for its high energy dependence which is close to 80% [1]. The Spanish low domestic energy produc- tion is based almost exclusively on nuclear generation, renewable energy resources, and a small contribution from domestic coal.

This greater dependence implies added risk for production pro- cesses related to ensuring energy supply. In that sense, Spanish electricity generation is one of the economic activities more exposed to the fuel dependence.

According to the International Energy Agency [2], half of gross electricity generated in Spain in 2011 came from power stations using fossil fuels (such as natural gas, coal and oil) becoming nat- ural gas the most important fuel type in the Spanish electricity generation mix as a consequence of the construction of numerous combined cycle power plants. Moreover, real fossil fuel prices have risen to record levels with the consequent increment of the vulner- ability of the Spanish electricity system.

This research examines the exposure of Spanish power genera- tion to the imported fuels between 1995 and 2011. For this pur- pose, we use a vulnerability indicator defined by Nakawiro and Bhattacharyya [3] and Bhattacharyya [4] to capture the fuel bill (for using fossil fuels) for electricity generation per unit of gross domestic product (GDP).

The fuel bill can be thought as a product of relevant driving fac- tors such as price of fuel price, fuel heat rate (thermal efficiency) and energy intensity, among others. Thus, we also analyze the driv- ing forces behind the change in Spanish electricity bill. We deter- mine these drivers by using an index decomposition analysis (IDA). In fact, we propose the use of the most popular approach within IDA, the so-called logarithmic mean Divisia index (LMDI) method. Researchers often apply this method to factorize changes

460 P. Fernández González, B. Moreno / Energy Conversion and Management 92 (2015) 459–468

in energy and environmental aggregates like energy consumption or intensity, or greenhouse gas emissions.

Usually, most of works implement widely accepted methodolo- gies in order to study changes in vulnerability. Jansen et al. [5], Constantini et al. [6] or Schaepers et al. [7], following Dutch Energy Research Centre-ECN and Clingendael International Energy Pro- gramme (CIEP), used the Shannon–Wiener diversity index as basic indicator of vulnerability. However, consideration of subjective weights to derive the composite index may be criticized. Other works like Gupta [8] or Gnansounou [9] relied on principal compo- nents and cluster analysis, respectively. In this case, the use of eigenvalues for estimating weights of principal components may also be questioned since the weight of each component should be the result of either objective measures or subjective perceptions of decision makers. Others like Yokoyama et al. [10] based their study on minimax regret criterion and Månsson et al. [11] discussed several methodologies to analyze vulnerability.

Although numerous index based methods of decomposition have been reported within the energy and environmental fields such as Inglesi-Lotz and Pouris [12], Choi and Oh [13], Fernández González et al. [14], they are uncommon in the energy exposure analysis. Only some authors as Bhattacharyya [4] use Laspeyres index based methodology to analyze changes in the vulnerability of electricity generation. This paper goes one step further by pro- posing the use of the most accurate method within IDA [15] that uses logarithmic mean weight functions: the logarithmic mean Divisia index (LMDI). It is an exhaustive method that uses logarith- mic mean weight functions instead of Laspeyres. These specific weight functions are based on Törnquist [16] and Sato [17] propos- als, and were firstly introduced in Ang and Choi [18] in order to analyze variations in aggregate energy intensity. Later, it has been widely studied and used in the fields of energy and environment such as energy intensity or consumption and gas emissions inten- sity. Moreover, it has been adopted by Canada [19] and Australia [20] to study trends of aggregates such as energy efficiency/ intensity.

In this paper, we present and derive its use in order to decom- pose changes in vulnerability. Unlike other methodologies, weights are not fixed nor subjectively stated by the researcher but they are adaptable and formulated in terms of logarithmic mean changes of the relevant variables. Furthermore, they satisfied most of the tests of index numbers which are considered to be relevant to IDA [21]: factor-reversal, time-reversal, proportionality and aggregation tests). On the other hand, it leaves no residual (factorization is exhaustive and complete), it can effectively handle value zero in the data set, it is consistent in aggregation and there exists a simple relationship between additive and multiplicative decomposition.1

For all these above (adaptability, theoretical foundation, but also from easy of use and result interpretation), we propose this method- ology to the study of the vulnerability of Spanish power generation.

Few studies of the exposure of the electricity sector on fuels have been carried out. Nakawiro and Bhattacharyya [3] analyzed the vulnerability of gas dependence of Thai electricity sector, Balat [27] studied security of energy supply in Turkey, Augutis et al. [28] explored energy mix optimization from an energy security per- spective, Ren and Sovacool [29] determined some of the factors to achieve China’s energy security and Grubb et al. [30] considered the diversity of electricity system in the UK. To the best of our knowledge, there is only one study that has considered security of electricity supply in Spain. This is the work of Bhattacharyya [4] which decomposed changes in vulnerability in the United

1 Demonstrations are collected in Ang et al. [22], Ang and Zhang [23], Ang and Liu [24], and Ang [25,26].

Kingdom, the Netherlands, Italy, Germany and Spain from 1996 to 2006 by using a Laspeyres index based methodology.

The objective of this paper is twofold: Firstly, it aims to add to the small body of empirical literature in Spain on this matter by taking the logarithmic mean Divisia index (LMDI) method to fac- torize driving forces underlining changes in the Spanish electricity bill; secondly, it aims to provide Spanish policy makers important findings on determinant factors driving the electricity vulnerability that may help them to design effective energy and environmental policies.

The organization of the paper is as follows. Section 2 reports the methodology. It presents the indicator used to measure electricity exposure and the adaptation of the LMDI decomposition to factor- ize driving forces underlining changes in vulnerability. In Section 3, we display the results, analyzing in detail the evolution of Spanish electricity vulnerability and its exposure to changes in the eco- nomic environment. In order to understand variations on it, we firstly define some factors that are commonly considered signifi- cant driving forces. Those are fuel price, fuel heat rate (thermal effi- ciency), fuel dependence, importance of electricity on gross energy consumption and energy intensity. Once determinant factors are clearly stated, we perform a preliminary but illuminating graphic analysis. Then, we implement the LMDI method to carry out a rig- orous study of vulnerability decomposition. Findings may help authorities to design effective energy and environmental policies. In that sense, in Section 4 we present discussion of the results, analyzing the Spanish regulatory framework influence on them. Finally, in Section 5, some concluding comments and policy recommendations are presented.

2. Method: adapting LMDI to vulnerability analysis

There are several studies such as Percebois [31]; Gnansounou [9]; Gupta [8]; Gnansounou and Dong [32], Ang et al. [33], Smith-Stegen and Palovic [34] and Jansen and Seebregts [35] sug- gesting a number of indicators and dimensions that could be used to measure weakness in energy supply. These include for example energy dependence, import concentration, energy intensity and net energy import bill, among others.

In this paper, we use a vulnerability indicator defined by Nak- awiro and Bhattacharyya [3] and Bhattacharyya [4]. This indicator seeks to display the fuel bill for electricity generation per unit of gross domestic product (GDP). Thus, one would expect that an increase in the price of fuel lead to changes in the costs of produc- ing electricity and thereby, in electricity prices, even if there is some transmission lag of the signal to the consumer market. As reasonable, an increment of the fuel bill as a share of GDP implies further weakness in the country as it is more subjected to any possible price shocks.

Mathematically, the vulnerability indicator (V) can be expressed as:

V ¼ Fuel bill for electricity power generation

GDP ¼ Xk j¼1

pj CEj

GDP ð1Þ

where, fuel bill for power generation is the total costs of using fuels for power generation; where Vj is the vulnerability of fuel j, Pj rep- resents price of fuel j, CEj is consumption of fuel j to produce elec- tricity and GDP is the country’s gross domestic product.

The numerator term ‘‘Fuel Bill for Power Generation’’ may be understood as a product of outstanding driving forces such as price of fuel, heat rate or efficiency of conversion of the fuel in electricity and dependence on the fuel for power generation.

Thus, we propose the use of a new approach within the index decomposition analysis IDA, the so-called logarithmic mean Divisia

P. Fernández González, B. Moreno / Energy Conversion and Management 92 (2015) 459–468 461

index (LMDI) method to factorize driving forces underlining changes in vulnerability.

4 From a historical perspective, the first proposals were based on indexes such as Laspeyres [36–38] and Marshall–Edgeworth [39–42]. Later, authors like Liu et al. [43], Ang and Lee [44], Ang [45], Ang and Choi [18], Sun [46] and Albrecht et al. [47] improved this methodology, proposing weight functions that are adaptable to changes in the magnitudes and lead to complete decompositions. Interesting

2.1. A logarithmic mean Divisia index (LMDI) decomposition

In this section, we present and derive the LMDI method use in order to decompose changes in vulnerability.

Since we describe the formulation in a multiplicative form,2

changes in energy exposure will be measured as a ratio, and this ratio will be equal to product of several determinant factors.

Assuming that the decomposition is carried out for m regions (or countries) and k types of fuels, we may express energy vulner- ability (V) in the following form:

V ¼ Xk j¼1

Xm r¼1

V jr ¼ Xk j¼1

Xm r¼1

Pjr CEjr

GDPr ¼ Xk j¼1

Xm r¼1

Pjr CEjr Ejr

Ejr Er

Er PCr

PCr GDPr

¼ Xk j¼1

Xm r¼1

Pjr Hjr Djr W r Ir ð2Þ

where Vjr is the vulnerability of fuel j in region r, Pj represents price of fuel j, CEjr is consumption of fuel j to produce electricity in region r, Ejr represents electricity generation from fuel j in region r, Er denotes gross electricity generation in region r, PCr is primary energy consumption in region r and GDPr denotes gross domestic production region r, Hj = CErj/Ejr denotes the average heat rate

3 by fuel j (energy efficiency or thermal performance for a fuel j in an average power plant), Djr = Ejr/Er is dependence of fuel j in region r, Wr = Er/PCr represents the importance of electricity in primary energy consumption and Ir = PCr/GDPr denotes energy intensity (defined as total primary energy consumption of region r over its gross domestic product).

Taking logarithmic derivatives with respect to time in (2) leads to:

d ln V dt

¼ Xk j¼1

Xm r¼1

Pjr Hjr Djr W r Ir V

d ln PjrðtÞ dt

þ d ln HjrðtÞ

dt þ

d ln DjrðtÞ dt

�

þ d ln W rðtÞ

dt þ

d ln IrðtÞ dt

� ð3Þ

Integrating (3) and applying the exponential function – the fol- lowing standard formula for the logarithmic change between 0 and T period gives:

V T V 0 ¼ exp

Xk j¼1

Xm r¼1

R T 0 w

� jrðtÞln

Pj;T Pj;0

� � dt

! exp

Xk j¼1

Xm r¼1

R T 0 w

� jrðtÞln

Hjr;T Hjr;0

� � dt

!

exp Xk j¼1

Xm r¼1

R T 0 w

� jrðtÞln

Djr;T Djr;0

� � dt

! exp

Xk j¼1

Xm r¼1

R T 0 w

� jrðtÞln

W r;T W r;0

� � dt

!

exp Xk j¼1

Xm r¼1

R T 0 w

� jrðtÞln

Ir;T Ir;0

� � dt

!

ð4Þ

Being wjr(t ⁄) the weight function evaluated at the point t⁄ [0, T].

Since different weight functions may be considered, different spe- cific methods may be derived. The earliest ones refer to those for- mulated on the basis of conventional indexes such as Laspeyres,

2 LMDI method verifies the additive property [26]. This means that a simple equivalence relationship between multiplicative and additive decompositions exists. Therefore, one of them is technically redundant.

3 Heat rate is defined as the ratio between the energy supplied to the power plant for a period and its electricity output in the studied period.

Paasche or Marshall–Edgeworth weights.4 Later, more refined methods like the Sun [46], the Path Based [49] or the so-called LMDI [18] methods were developed. Following the latter, we propose the following weight functions:

LðV jr;0; V jr;TÞ¼ ðV jr;T � V jr;0Þ lnðV jr;T=V jr;0Þ

ð5Þ

After its normalization, (5) may be transformed in:

w�jr ¼ LðV jr;0; V jr;TÞXk

i¼1 LðV jr;0; V jr;TÞ

; i ¼ 1; 2; . . . ; k: ð6Þ

being w�jr the normalized weight of fuel j (j = 1, 2, . . . , k) in region r (r = 1, 2, . . . , m).

Therefore, contributions from determinant factors (Rfactor) may be expressed as:

Rp ¼ exp Xk j¼1

Xm r¼1

Z T 0

w�jrðtÞln Pj;T Pj;0

� � dt

! ð7Þ

Rh ¼ exp Xk j¼1

Xm r¼1

Z T 0

w�jrðtÞln Hjr;T Hjr;0

� � dt

! ð8Þ

Rd ¼ exp Xk j¼1

Xm r¼1

Z T 0

w�jrðtÞln Djr;T Djr;0

� � dt

! ð9Þ

Rw ¼ exp Xk j¼1

Xm r¼1

Z T 0

w�jrðtÞln W r;T W r;0

� � dt

! ð10Þ

Ri ¼ exp Xk j¼1

Xm r¼1

Z T 0

w�jrðtÞln Ir;T Ir;0

� � dt

! ð11Þ

Finally, total change in V may be expressed as

Rtot ¼ Rp RhRdRw Ri Rr ð12Þ

where Rtot = VT/V0 is the total change in V between periods 0 and T (total effect), Rp denotes the impact of changes in fuels price (price effect), Rh reflects the influence of variations in average heat rate (heat rate effect), Rd quantifies the contribution on energy depen- dence in electricity generation (dependence effect), Rw reflects the influence of changes on the importance of electricity in primary energy consumption (weight effect), Rint denotes the impact of vari- ations in energy intensity (intensity effect), and Rr shows the residual term.5

2.2. Time series decomposition

Following this methodology, intermediate periods between benchmark years 0 and T are also considered. If the cumulative change in energy vulnerability (total effect) from 0 to T is denoted as (Ctot)0,T, the estimated cumulative price effect as (Cp)0,T, the esti- mated cumulative thermal efficiency effect by (Ch)0,T, the estimated cumulative energy dependence effect by (Cd)0,T, the estimated cumulative electricity weight effect by (Cw)0,T, the estimated cumu- lative intensity effect by (Ci)0,T, and the estimated cumulative

statistical and mathematical properties were attributed to these methods in Sun and Ang [48].

5 By definition, exhaustive decomposition methods ensure that there is no deviation from the target value, so the residual term is always unity (resp., null) in the multiplicative (resp., additive) case. In the following sections, residual term is computed in order to check correctness of our calculations.

462 P. Fernández González, B. Moreno / Energy Conversion and Management 92 (2015) 459–468

residual term6 is (Cr)0,T, then the estimated cumulative effects, from 0 to T may be multiplicatively decomposed as follows:

ðCtotÞ0;T ¼ðRtotÞ0;1ðRtotÞ1;2 . . .ðRtotÞT�1;T ð13Þ ðCpÞ0;T ¼ðRpÞ0;1ðRpÞ1;2 . . .ðRpÞT�1;T ð14Þ ðChÞ0;T ¼ðRhÞ0;1ðRhÞ1;2 . . .ðRhÞT�1;T ð15Þ ðCdÞ0;T ¼ðRdÞ0;1ðRdÞ1;2 . . .ðRdÞT�1;T ð16Þ ðCwÞ0;T ¼ðRwÞ0;1ðRwÞ1;2 . . .ðRwÞT�1;T ð17Þ ðCiÞ0;T ¼ðRiÞ0;1ðRiÞ1;2 . . .ðRiÞT�1;T ð18Þ ðCrÞ0;T ¼ðRrÞ0;1ðRrÞ1;2 . . .ðRrÞT�1;T ð19Þ

Results provided by time series decomposition make possible the detection of different phases or time patterns in the effects and enables the identification of potential structural breaks in intermediate periods. Actually, this procedure may be expected to be more accurate than periodwise decompositions in which only initial (0) and final (T) periods are taken into account.

3. Results

In this section, we analyze in detail the evolution of Spanish electricity power vulnerability from 1995 to 2011. First, we per- form a preliminary but illuminating graphic analysis (Section 3.1) that helps us to better understand the specific Spanish situation. Then, we go deeper in the study and we carry out a rigorous decomposition analysis (Section 3.2). For this analysis, we imple- ment the LMDI method presented and developed in the previous section.

In order to carry out the empirical analysis, the following vari- ables are used:

– Consumption from fuel – coal, oil and gas-in electricity generation (thousand tons of oil equivalent-1000 TOE): This variable mea- sure the consumption of each fossil fuel for Electrical energy generated and it is obtained from the International Energy Agency [2].

– Primary energy consumption of fossil fuels – coal, oil and gas (thousand tons of oil equivalent -1000 TOE): This variable mea- sures the primary energy consumption of each fossil fuel by the Spanish Economy. The information is obtained from the Inter- national Energy Agency [2].

– Gross electricity generation by fossil fuel – coal, oil and gas – and total (Gigawatt hour-GWh): This variable measures total gross electricity generated in Spain and the gross electricity gener- ated by each fossil fuel. They are obtained from the Interna- tional Energy Agency [2].

– Price of fossil fuels: The Price of gas (€/million of British Thermal Units) measures the price of imported gas in Spain, the Price of Coal (€/tonne) measures the Northwest Europe marker price of coal and the Price of Oil (€/barril) is the Brent crude petroleum price by barrel. The information is obtained from the BP Statis- tical Review of World Energy [50].

– Gross domestic product (Purchasing Power Parity – Millions of €). This variable measures the Spanish Economic Activity and it is obtained from Eurostat [1].

Table 1 gives some details on the variables used in the analysis.

3.1. Graphical analysis

We began the preliminary study by observing the need of differ- ent fuels to produce electricity (Fig. 1). Coal and peat was the

6 Since LMDI method is exhaustive and complete, (Cr)0,T = 1.

leading fuel contributor to electricity generation in the first part of the studied period, while gas becomes the major fuel for power generation in the second part of the period.

As happened in the rest of Europe, the demand for gas has sig- nificantly increased as a replacement for more expensive, less envi- ronmental friendly resources. The choice of gas to produce electricity raised dependence and vulnerability on this energy resource since the present European supply is dependent on too few sources and venues [51]. It is noticed that the Spanish share of electricity generated by natural gas has increased from 2.2% to 29% during the considered period 1995–2011. Natural gas has experienced a great growth in Spain, partly for the ample invest- ment in gas combined cycle production technologies. In that sense, the increasing large number of combined cycle installations have entered into service is basically due to their large advantages over those coal-based technologies: higher thermal efficiency, lesser volume of specific investment, lower requirement of personnel and shorter time of setting them up [52].

As the U.S. Energy Information Administration [53] points out, up until the 2008 financial crisis, Spain was one of countries with the fastest-growing natural gas markets in Europe. Further, accord- ing to PFC Energy, in 2011 Spain was the third-largest importer country of liquefied natural gas (LNG) after South Korea and Japan.

Fig. 2 reports evolution of both gas price and vulnerability in the Spanish economy. Findings exhibit a very similar behavior pattern between both variables particularly in the second part of the ana- lyzed time interval. Therefore gas vulnerability becomes an impor- tant challenge in this economy.

An increasing gas price positively influences the electricity bill and therefore, vulnerability of electricity consumers. As fuel prices have changed substantially in Spain, the price effect is a significant driver of changes in electricity exposure.

In that sense, changes in the costs of natural gas, coal and oil – can directly affect retail electricity prices, since generation costs are likely to be transmitted through the wholesale electricity market.

In the present context of a liberalized Spanish electricity market, the price is influenced by the short-term marginal energy cost of the different electricity generation technologies. Although the wholesale electricity prices are mainly sensitive to variations on crude oil, coal and natural gas prices – as Pereira Freitas and Pereira Da Silva [54] have shown in the context of the Iberian electricity market (MIBEL) – the price increment in the used fuel of the marginal generator is finally transmitted to the final consumer bill.

We would like to point out that electricity prices for industrial consumers are particularly important for industry competitive- ness, as electricity usually represents a significant proportion of total energy cost for industry. According to the last Spanish Energy Consumption Survey [55] the main energy product used by indus- trial companies was electricity – 51.7% of the total energy con- sumption – becoming Rubber and plastic products, Electrical, electronic and optical material and equipment and Transport material those sectors with consumption of electricity representing more than 70% of they total energy consumption. Competitiveness of Spanish industrial companies has suffered a significant negative impact since electricity prices have been going up in the last years. Variability of imported fuel prices obviously impacts competitive- ness of manufacturing and other sectors since it involves important effects on the level and stability of the final electricity prices paid by industrial consumers [56].

Attending to the heat rate, it exhibits an improvement along the period 1995–2011, particularly in oil and gas in the first years (Fig. 3).

An improvement in energy efficiency for power plant reduced Spanish exposure to fuel fluctuations.

Table 1 Definition of variables and descriptive statistics (1995–2011).

Variable Min Max Mean Standard deviation Units Source

Consumption from fuel in electricity generation Coal 6090 18,926 14,763.4 3746 1000 TOE IEA Oil 2646 5863 4175.9 915.3 Gas 752 18,036 7799.2 5984.9

Primary energy consumption of fossil fuels Coal 7938 21,602 17,440.9 4017.2 1000 TOE IEA Oil 55,800 62,232 59,272.8 1976.7 Gas 7720 34,900 21,675.6 9363.8

Gross electricity generation by fossil fuel and total Total 167,085 313,758 251,862.0 50,331.8 GWh IEA Coal 26,323 82,457 64,508.5 16,637.2 Oil 13,920 28,593 20,386.2 4434.3 Gas 3750 120,798 53,353.1 39,548.8

Price of fossil fuels Coal 27.0 100.4 49.5 20.9 €/Tonne BP Oil 11.4 79.9 36.1 20.4 €/barril Gas 2 9.1 4.5 2.2 €/MMBTU

Gross domestic product 528,405.6 1,180,970.1 882,318.1 228,068.2 MM of PPS Eurostat

Fig. 1. Energy consumption (in GWh) for electricity generation in the Spanish economy in 1995–2011 period. Source: IEA.

Fig. 2. Trends of total effect (Ctot) and gas price (Pgas, in euro per GWh) in 1995– 2011 period. Sources: British Petroleum and own elaboration.

Fig. 3. Trend of heat rate per fuel (Hj, in GWh/GWh) in the Spanish economy, 1995– 2011 period. Source: Own elaboration from IEA data.

Fig. 4. Fuel dependence (Dj) trend in the Spanish economy in 1995–2011 period. Source: IEA.

P. Fernández González, B. Moreno / Energy Conversion and Management 92 (2015) 459–468 463

However, heat rate in coal and peat power plants does not show improvements in thermal efficiency. Therefore, changes in this fuel continue influencing Spanish vulnerability, particularly in the first part of the study when coal and peat is the major fuel for power generation. This partly counteracts the reducing electricity exposure promoted by the rest of the fuels.

Focusing on the fuel dependence factor, significant changes in fuel dependence for power generation affect vulnerability index. Fig. 4 shows fuel dependence evolution in the Spanish economy.

Particularly oil but also gas dependence shows a slight decrease in the first years of study, therefore contributing to exposure

Fig. 6. Evolution of energy intensity in the Spanish economy from 1995 to 2011. Source: Eurostat.

Table 2 Decomposition results through multiplicative LMDI method (Base year = previous year).

Years Rtot Rp Rh Rd Rw Ri

1995 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1996 1.21870 1.07743 0.73649 1.84403 0.91864 0.90705 1997 2.67696 1.07524 1.06472 1.97174 1.17797 1.01289 1998 0.63941 0.88573 0.85933 0.89564 0.96475 0.97217 1999 1.23487 0.94201 1.18068 0.96373 1.19248 0.96629 2000 1.72170 2.06308 0.86362 1.01319 1.02072 0.93444 2001 1.06508 1.07711 0.91069 1.18651 0.98073 0.93311 2002 1.05632 0.72142 1.14671 1.14674 1.14016 0.97661 2003 1.20507 1.06755 0.97417 1.23426 0.95957 0.97847 2004 1.35149 1.00468 1.02511 1.21664 1.07168 1.00646 2005 1.91921 1.47000 0.98552 1.22464 1.10612 0.97841 2006 1.53489 1.32453 1.11146 1.15437 0.99255 0.91010 2007 0.85913 0.92692 0.95531 1.01625 1.02404 0.93229

464 P. Fernández González, B. Moreno / Energy Conversion and Management 92 (2015) 459–468

reductions. However, coal and peat dependence remains stable. Since this is the major fuel for power generation in these first peri- ods, impacts of energy dependence are very limited. Moreover, in the last part of the analyzed period, gas dependence increases, and being this the most widely used in power generation, it leads to a more vulnerable Spanish economy.

Furthermore, such vulnerability to imported fuels is aggravated because these imports are from countries in which supplies, eco- nomic policies or legal security are not totally reliable. In 2011, the imported petroleum came mainly from Russia (15.3%), fol- lowed by Saudi Arabia (14.7%), Iran (14.4%) and Nigeria (13.3%) while the imported natural gas came mainly from Algeria (37%), followed by Nigeria (20%) and Qatar (13%) (Spanish Ministry of Industry, Energy and Tourism [57]).

Meanwhile, the importance of electricity in the overall energy demand also influences the electricity sector exposure to changes.

Fig. 5 reports progress in electricity influence in the Spanish economy from 1995 to 2011.

Despite of the slight reduction showed by the importance of electricity sector in the Spanish energy consumption, Fig. 5 reveals an increasing trend in this variable. Therefore, consequences involve a more exposure of Spanish economy with an increment in its vulnerability.

Finally, we analyze the evolution of the last predetermined fac- tor: energy intensity which is reported in Fig. 6.

Spanish energy intensity has mostly dropped during this period, experiencing some growth, particularly in the final year of study. Since lower energy intensity means lower energy demand for an output level, this contributed to reduce vulnerability in the most part of the studied period.

Findings of this introductory graphical analysis are later evi- denced in next section in which each individual driver is quantified through a rigorous decomposition technique.

2008 1.89266 1.45928 1.02794 1.26013 1.02412 0.97784 2009 0.65805 0.68531 1.01199 1.02783 0.96144 0.96017 2010 0.89361 1.00162 1.01943 1.04723 0.86580 0.96522 2011 1.05069 1.21071 0.97413 0.83621 1.04797 1.01661

Table 3 Decomposition results through multiplicative LMDI method (Base year = 1995).

Year Ctot Cp Ch Cd Cw Ci

1995 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1996 1.21870 1.07743 0.73649 1.84403 0.91864 0.90705 1997 3.26241 1.15850 0.78416 3.63595 1.08212 0.91874 1998 2.08602 1.02612 0.67385 3.25650 1.04398 0.89317 1999 2.57596 0.96661 0.79560 3.13840 1.24492 0.86305 2000 4.43504 1.99420 0.68710 3.17979 1.27072 0.80647

3.2. LMDI decomposition results

In this section, we implement the LMDI approach presented in Section 2, decomposing changes in the Spanish vulnerability of electricity generation. Although this methodology makes possible an aggregate study of r economic regions, our willingness to ana- lyze a sufficiently large temporal interval and the difficulty of achieving homogeneous regional data sets, lead us to limit our study to a single country: Spain.

Results of periodwise and time series estimated effects are col- lected in Tables 2 and 3, respectively.

According to Table 2, Spain shows an increasing vulnerability path along the studied period, only altered in 1998, 2007, 2009 and 2010. In these particular years, Spain vulnerability

Fig. 5. Evolution in electricity importance (W, electricity generation over primary energy consumption) in Spain, period 1995–2011. Source: IEA.

2001 4.72368 2.14797 0.62573 3.77287 1.24623 0.75253 2002 4.98971 1.54960 0.71753 4.32648 1.42090 0.73493 2003 6.01296 1.65427 0.69900 5.34002 1.36346 0.71911 2004 8.12644 1.66200 0.71655 6.49691 1.46119 0.72375 2005 15.59637 2.44314 0.70617 7.95636 1.61626 0.70812 2006 23.93864 3.23602 0.78488 9.18459 1.60422 0.64446 2007 20.56646 2.99955 0.74980 9.33386 1.64279 0.60082 2008 38.92526 4.37717 0.77075 11.76184 1.68241 0.58751 2009 25.61494 2.99974 0.77999 12.08918 1.61753 0.56411 2010 22.88978 3.00461 0.79514 12.66010 1.40046 0.54449 2011 24.05003 3.63772 0.77457 10.58649 1.46764 0.55353

improvement basically relies in negative contribution of the price and weight effects. It is likely that the numerous measures taken by the Spanish government to meet the euro area access objectives and the oil price increases in the late 90s, and the consequences of the recent economic crisis in the late 2000 had influenced the aggregate in this sense.

0

2

4

6

8

10

12

14

16

19 90

19 92

19 94

19 96

19 98

20 00

20 02

20 04

20 06

20 08

20 10

20 12

$/ pe

r m ill

io n

B tu

0

20

40

60

80

100

120

$/ pe

r b ar

re l

Natural gas ($/per million Btu) Petroleum Brent ($/per barrel)

Fig. 8. Evolution of imported gas and petroleum prices. Source: British Petroleum, 2013

P. Fernández González, B. Moreno / Energy Conversion and Management 92 (2015) 459–468 465

Attending to Table 3, it shows a worrying increase in the energy vulnerability of the Spanish economy around 24.05% between 1995 and 2011. This increment is driven by factors such as fossil fuels price, energy dependence and importance of electricity in primary energy consumption. Particularly, increases in these factors increased vulnerability of the Spanish economy in about 363.77%, 1058.64% and 46.76%, respectively. Meanwhile, improvements in factors such as energy intensity and thermal efficiency contributed to reduce energy vulnerability in about a 22.55% and 44.65%, respectively. Results strongly show the greater importance of those determinants upward influencing than those downward influenc- ing electricity vulnerability. Therefore, total effect reveals a more vulnerable Spanish economy.

Fig. 7 shows graphic analysis of cumulative effects progress from 1995 to 2011. This chart facilitates a detailed analysis, reveal- ing some outstanding movements in effects development.

Then, we attend to each individual effect evolution. First, we realize that the energy dependence effect is the most influential factor in any period, growing its impact every year, except in 1998, 1999 and 2011 periods, when it diminishes its influence over previous years. Thus, action lines addressing energy dependence reduction are keys in order to control Spanish vulnerability. In this sense, measures looking for changes in the fuel mix such as fuel diversification and public promotion of renewable are strongly recommended.

Second, evolution of the fuel price effect shows a changeable but positive influence, having small impact on the first part of the period, inverting its impact in 1999 and significantly contribut- ing to increased exposure in the second part of the analyzed period. Strategies aiming at keeping stable fuel prices – intermediation in international conflicts, international cooperation, etc. – are recom- mended in order to reduce vulnerability.

In fact, attending to Fig. 8 representing evolution of imported gas and petroleum prices, fossil fuel prices have risen to record levels over the past decade with the consequent exposure of the Spanish Economy.

Third, the electricity importance effect exhibits a slight but increasing influence on changes in the aggregate. The only excep- tions are 1998, 2001, 2003 and 2010 years. Therefore, measures such as promotion of green attitudes among final consumers, proper regulation for the construction of clean buildings, financial support for renewing the vehicles fleet and appliances will contribute to limit this effect.

Fig. 7. Estimated cumulative effects from LMDI decomposition method (Base year = 1995).

Fourth, energy intensity effect proves to be the most leading one when reducing energy vulnerability in the Spanish economy. Particularly, in 2000 and 2006 when its impact exhibits a signifi- cant change. Energy use within industry sector includes highly intensive subsectors like steel, aluminum and petrochemicals in which Spain is a traditional producing European country. There- fore, those energy strategies pursuing an improvement on energy efficiency or looking for structural change toward less energy intensive sectors will help on energy exposure reduction in the Spanish economy. Furthermore, building material sector (including concrete and bricks) is also a significant energy intensive sector in Spain and its prospects are intimately linked to future activity. Consequently, economic boom or crisis periods play an important role in this sector. Action lines aiming at balance those periods may also contribute to get vulnerability under control.

Finally, heat rate effect shows a changeable contribution to energy vulnerability reduction, being its influence larger in the first part of the period than in the second one. Investment, innovation, R&D, adaptation to more efficient technologies in power sector and any public and financial support to achieve better thermal effi- ciency becomes interesting strategies in order to reduce exposure.

Findings on this work are endorsed by similar energy security studies. Bhattacharyya [4] decomposed changes in vulnerability in the United Kingdom, the Netherlands, Italy, Germany and Spain from 1996 to 2006. In order to perform this analysis, he relied on a Laspeyres index based decomposition method. Its implementation to the Spanish economy led to analogous results to those high- lighted in this paper: gas dependence in electricity generation is revealed as a crucial driver on vulnerability.

4. Discussion of the results

The long-term security of energy supply is a priority issue on energy policy agenda worldwide. The growing dependence on energy imports from insecure regions, the volatility of energy prices, the growing energy demand in emerging economies, the perspective of oil and gas reserves depletion and the concern of global climate change make public and political attention to energy supply security issues.

The European Union (EU) is one of the present international actors who are actively involved in environmental and energy challenges. Since Spain is a member state of the EU and it must abide and transpose current European legislation is also concerned about these issues, playing also an important role.

The EU has not always played a key role in energy security and environmental fields. Although energy was included in the European agenda from the first moment as reflected in Treaty of

466 P. Fernández González, B. Moreno / Energy Conversion and Management 92 (2015) 459–468

Paris [58] or Treaties of Rome [59], it did not develop a truly com- mon energy policy until the signing of Kyoto Protocol [60]. Finally, in 2000, the European Commission [61] presented the Action Plan on Energy Efficiency 2000–2006 which aims at protecting the envi- ronment, strengthening the security of energy supply, creating a more sustainable energy policy in order to reduce energy con- sumption. Then, it deepened reporting several regulatory proposals in EC [62]. Among them, the Directive 2001/77/EC [63] encourages electricity generation from renewable sources; the Directive 2002/ 91/EC [64] ensuring a minimum energy performance requirements in buildings; the Communication 2002/0488 [65] tackles internal energy market, coordinating measures on the security of energy supply, the Directive 2005/89/EC [66] sets a proper framework for ensuring adequacy between electricity demand and supply, and stimulating an appropriate interconnection capacity between EU Members; and the Directive 2006/67/EC [67] establishes mini- mum stocks and reserves of crude oil and/or petroleum products within EU.

A second Green Paper on European Strategy for Sustainability, Competitiveness and Secure Energy [68] goes beyond and it includes measures for actively combating energy vulnerability by: (i) promoting renewable energy sources and energy efficiency, (ii) improving the efficiency of the European energy grid by creat- ing a truly competitive internal energy market and (iii) coordinat- ing the EU’s supply of and demand for energy within an international context.

At last, a real energy policy for Europe is set up in Communica- tion 2007/0001 [69]. It seeks to maintain a careful balance between security of supply, competitiveness and environmental sustainabil- ity. Moreover, Communication 2008/0772 [70] considers several energy efficiency measures to achieve the so-called 20-20-20 goal, and Directive 2009/28/EC [71] deeps in this goal proposing strate- gies to increase the percentage of renewables use from 20 to 30%.

In recent years, European Communities continue facing energy risks, deepening on improvements in energy efficiency as it is shown in Directive 2012/27/EU [72] and Communication 2013/ 0762 [73].

Minimizing the EU and particularly the Spain vulnerability con- cerning imports, shortfalls in supply, possible energy crises and uncertainty with respect to future supply may be faced proceeding in different ways:

(a) Establishment the internal energy market by: (i) setting common rules for the internal market in electricity looking for integrated, competitive interconnected and single EU electricity market, (ii) establishing minimum rules and obli- gations to ensure continuous operation of the transmission network operators.

(b) Safeguard security of electricity supply by: (i) setting up a clear policy in place to maintain the balance between supply and demand, (ii) encouraging the establishment of whole- sale markets, (iii) ensuring an appropriate level of genera- tion reserve capacity, (iv) facilitating the development of new generation capacity and encouraging energy conserva- tion and technology for demand management in real time, (v) laying down a framework facilitating investment for net- work operators, (vi) stimulating SMEs to undergo energy audits and disseminate best practices, (vii) auditing energy consumption of large companies to help them on identifica- tion of their potential energy consumption reduction, (viii) monitoring efficiency levels of new energy generation capacities, (ix) ensuring solidarity between Member States, and (x) building up strategic oil stocks.

(c) Commitment to renewable resources by: (i) diversifying energy mix, (ii) ruling public incentives for investment in new energy options (e.g., synthetic fuels from coal), and

(iii) promoting the use of renewable energies (wind power, solar and photovoltaic energy, biomass and bio-fuels, geo- thermal energy and heat-pump systems).

(d) Improvement of energy efficiency by: (i) supporting invest- ment in generation capacity, (ii) developing more energy savings technologies, (iii) diversifying energy mix, and (iv) researching for new materials.

(e) Enhance a European strategy/implement a common interna- tional energy policy by: (i) enhancing EU relations with con- sumer countries (such as the United States, India, Brazil or China), producer countries (Russia, Norway, OPEC countries and Algeria, for example) and countries of transit (such as the Ukraine), (ii) providing a stronger mechanism to secure new import routes for increasing amounts of oil and gas, (iii) diversifying geographical sources of supply, (iv) extend- ing Community powers with regard to establish a complete strategy for security of supply, and (v) enhancing external relations (encouragement of geopolitical and economic sta- bility in supplier and transit countries, closer cooperation with international suppliers and predictability in the pro- ducer–consumer relationship).

(f) Actions in favor of a demand policy by: (i) bringing about a real change in consumer behavior (e.g., using taxation), (ii) enhancing the use of more energy saving vehicles and appli- ances, (iii) promoting the use of bio-fuel transport (diversifi- cation of transport fuel mix), (iv) assessing for co-generation, tri-generation and district heating potential and (v) promot- ing consumers ‘‘green attitudes’’ and rational use of energy.

We would like to point out that a reduction of the energy inten- sity has been observed in Spain due to technological improvements and energy efficiency policies, mainly driven by the Energy Saving and Efficiency Action Plan 2011–2020 [74]. This Plan includes final and primary energy savings, promotion objectives of renewable energies and more efficient transformation technologies. In fact, in the Energy Transformation sector, the proposed aims are ‘‘the installation of 3751 MW of new co-generation capacity till 2020, and the renewal of up to 3925 MW of co-generation capacity over 15 years old’’. Thus, it is expected specific government support to boost household co-generation capacity and regulatory develop- ments for a suitable incorporation of the micro-grids in the elec- tricity system [75–77].

To sum up, results of this work recommend actions such as the enhancement of a common European strategy, safeguard security of electricity supply and commitment to renewable energies. In gen- eral, all those actions particularly aiming at reducing electricity dependence and fuel price volatility since they are both the biggest contributors to vulnerability increase in the Spanish economy in the last years.

5. Conclusions

Nowadays, energy supply is considered an essential input for any economic activity of developed and developing countries. Therefore, ensuring an affordable and reliable supply is one of the important challenges for policy makers. Several indicators sys- tems were proposed in the literature for assessing energy security or vulnerability. Evaluation of these concepts in a particular national context reveals delicate due to its multidimensional char- acter and the presence of numerous interconnections. In this paper, we aim at studying changes in Spanish exposure in the electricity generation, quantifying and evaluating drivers under those changes. In order to achieve this goal, we propose the use of an interesting index-based decomposition approach. The so-called LMDI method is derived in order to decompose changes in vulnerability into contributions from factors like volatility of fuel

P. Fernández González, B. Moreno / Energy Conversion and Management 92 (2015) 459–468 467

prices, energy dependence in electricity generation, energy inten- sity, the importance of electricity over primary energy consump- tion and average heat rate. Findings show a significant increase in exposure (particularly in gas), especially driven by energy dependence and increases in fuel prices. Thus, those strategies seeking for control these factors will contribute to substantially increase energy security in the Spanish economy. In this context, a transition toward a sustainable, resource-efficient economy is instrumental for reducing Spanish economy vulnerability. In that sense, energy efficiency can reduce the impact of Spanish economy of volatile energy prices on the world market.

Efficiency improvement of energy use may be mainly achieved from technological improvement and innovation. However, the decision about the investment energy-efficient technologies is shaped by green-culture of consumers and their perception of energy efficiency level [78]. Thus, it could be advisable to audit energy consumption to help consumers to identify their energy saving potential. In that sense, technological progress in advanced metering infrastructures (AMI) are successful instruments to reduce energy consumption by aiming to change the amount or timing consumer’s electricity demand. Moreover, government pro- grams such as educational workshops, advertising, and product labeling [79] may be also a way to promote ‘‘green attitudes’’ in consumers and a rational use of energy.

To sum up, Spanish economy must speed up its transition to an energy-efficiency economy and reduce its energy dependence, achieving a cost reduction and increase in the level of self-supply. Moreover, the triple goal of the ‘20-20-20’ initiative for 2020, which means a saving of 20% of the Union’s primary energy con- sumption has characterized the new frame of reference for devis- ing energy policy.

References

[1] European Commission. European economic statistics (Eurostat). Luxembourg: Publications Office of the European Union; 2013. <http://epp. eurotat.ec.europa.eu/portal/page/portal/statistics/search_database>.

[2] International Energy Agency; 2013. <http://www.iea.org/statistics/>. [3] Nakawiro T, Bhattacharyya SC. High gas dependence for power generation in

Thailand: the vulnerability analysis. Energy Policy 2007;35:3335–46. [4] Bhattacharyya SC. Fossil-fuel dependence and vulnerability of electricity

generation: case of selected European countries. Energy Policy 2009;37:2411–20.

[5] Jansen JC, van Arkel WG, Boot MG. Designing indicators of long term energy supply security. ECN-C-04-007; 2004.

[6] Constantini V, Gracceva F, Markandya A, Vicini G. Security of energy supply. Comparing scenarios from European perspective. Energy Policy 2007;35:220–6.

[7] Schaepers M, Seebregts A, de Jong J, Matters H. EU standards for energy security of supply. ECN-E-07-004/CIEP; 2007.

[8] Gupta E. Oil vulnerability index of oil-importing countries. Energy Policy 2008;36:1195–211.

[9] Gnansounou E. Assessing the vulnerability: case of industrialized countries. Energy Policy 2008;36:3734–44.

[10] Yokoyama R, Fujiwara K, Ohkura M, Wakui. A revised method for robust optimal design of energy supply systems based on minimax regret criterion. Energy Convers Manage 2014;84:196–208.

[11] Månsson A, Johansson B, Nilsson LJ. Assessing energy security: an overview of commonly used methodologies. Energy 2014;73:1–14.

[12] Inglesi-Lotz R, Pouris A. Energy efficiency in South Africa: a decomposition exercise. Energy 2012;42:113–20.

[13] Choi KH, Oh W. Extended Divisia index decomposition of changes in energy intensity: a case of Korean manufacturing industry. Energy Policy 2014;65: 275–83.

[14] Fernández González P, Landajo M, Presno MJ. The driving forces behind changes in CO2 emission levels in the EU27. Differences between Member States. Environ Sci Policy 2014;38:11–6.

[15] Ang BW, Huang HC, Mu AR. Properties and linkages of some index decomposition analysis methods. Energy Policy 2009;37:4624–32.

[16] Törnquist L. A memorandum concerning the calculation of Bank of Finland consumption price index. Bank of Finland; 1935.

[17] Sato K. The ideal log-change index number. Rev Econ Statist 1976;58:223–8. [18] Ang BW, Choi KH. Decomposition of aggregate energy and gas emission

intensities for industry: a refined Divisia index method. Energy J 1997;18(3): 59–73.

[19] Natural Resources Canada. Energy efficiency in Canada, 1994–2004. Office of Energy Efficiency. Ottawa; 2006.

[20] Sandu S, Syed A. Trends in energy intensity in Australian industry. ABARE 2008, Canberra (Report 08.15).

[21] Ang BW, Liu FL, Chung HS. Index numbers and the Fisher ideal index approach in energy decomposition analysis. Research Report 38/2002; Department of Industrial and Systems Engineering, National University of Singapore, 10 Kent Ridge Crescent, Singapore 119260.

[22] Ang BW, Zhang FQ, Choi KH. Factorizing changes in energy and environmental indicators through decomposition. Energy 1998;23:489–95.

[23] Ang BW, Zhang FQ. A survey of index decomposition analysis in energy and environmental studies. Energy 2000;25:1149–76.

[24] Ang BW, Liu FL. A new energy decomposition method: perfect in decomposition and consistent in aggregation. Energy 2001;26:537–48.

[25] Ang BW. Decomposition analysis for policymaking in energy: which is the preferred method? Energy Policy 2004;32:1131–9.

[26] Ang BW. The LMDI approach to decomposition analysis: a practical guide. Energy Policy 2005;33:867–71.

[27] Balat M. Security of energy supply in Turkey: challenges and solutions. Energy Convers Manage 2010;51(10):1998–2011.

[28] Augutis J, Martišauskas L, Krikštolaitis R. Energy mix optimization from an energy security perspective. Energy Convers Manage 2015;90:300–14.

[29] Ren J, Sovacool BK. Enhancing China’s energy security: determining influential factors and effective strategic measures. Energy Convers Manage 2014;88:589–97.

[30] Grubb M, Butler L, Twomey P. Diversity and security in UK electricity generation: the influence of low-carbon objectives. Energy Policy 2006;34(18):4050–62.

[31] Percebois J. Energy vulnerability and its management. Int J Energy Sector Manage 2007;1(1):51–62.

[32] Gnansounou E, Dong J. Vulnerability of the economy to the potential disturbances of energy supply: a logic-based model with application to the case of China. Energy Policy 2010;38:2846–57.

[33] Ang BW, Choong WL, Ng TS. Energy security: definitions, dimensions and indexes. Renew Sustain Energy Rev 2015;42:1077–93.

[34] Smith Stegen K, Palovic M. Decision-making for supplying energy projects: a four-dimensional model. Energy Convers Manage 2014;86:644–52.

[35] Jansen JC, Seebregts AdJ. Long term energy services security: what is it and how can it be measured and valued? Energy Policy 2010;38:1654–64.

[36] Bossanyi E. UK primary energy consumption and the changing structure of final demand. Energy Policy 1979;7:253–8.

[37] Hankinson GA, Rhys JMW. Electricity consumption, electricity intensity and industrial structure. Energy Econ 1983;5:146–52.

[38] Park SH. Decomposition of industrial energy consumption – an alternative method. Energy Econ 1992;4(4):265–70.

[39] Reitler W, Rudolph M, Schaefer H. Analysis of the factors influencing energy consumption in industry – a revised method. Energy Econ 1987;9:145–8.

[40] Boyd GA, McDonald JF, Ross M, Hanson DA. Separating the changing composition of US manufacturing production from energy efficiency improvements: a Divisia index approach. Energy J 1987;8(2):77–96.

[41] Hulten CR. Divisia index. In: Eatwell et al., editors. The new palgrave dictionary of economics, Palgrave Macmillan; 1987. doi:10.1057/9780230226203.2399. <http://www.dictionaryofeconomics.com/article?id=pde1987_X000612>. Palgrave Macmillan. [14.09.12].

[42] Doblin CP. Declining energy intensity in the US manufacturing sector. Energy J 1988;9(2):109–35.

[43] Liu XQ, Ang BW, Ong HL. The application of the Divisia index to the decomposition of changes in industrial energy consumption. Energy J 1992;13(4):161–77.

[44] Ang BW, Lee SY. Decomposition of industrial energy consumption: some methodological and application issues. Energy Econ 1994;16(2):83–92.

[45] Ang BW. Decomposition methodology in industrial energy demand analysis. Energy 1995;20(11):1081–95.

[46] Sun JW. Changes in energy consumption and energy intensity: a complete decomposition model. Energy Econ 1998;20(19):85–100.

[47] Albrecht J, Francois D, Schoors K. A Shapley decomposition of carbon emissions without residuals. Energy Policy 2002;30(9):727–36.

[48] Sun JW, Ang BW. Some properties of an exact decomposition model. Energy 2000;25:1177–88.

[49] Fernández-Vázquez E, Fernández González P. An extension to Suńs decomposition methodology: the path based approach. Energy Econ 2008;30(3):1020–36.

[50] Bristish Petroleum. Statistical review of world energy; 2013. <http://www.bp. com/content/dam/bp/pdf/statisticalreview/statistical_review_of_word_energy_ 2013.pdf>.

[51] Weiser H. The security of gas supply – a critical issue for Europe? Energy Policy 2007;35(1):1–5.

[52] Fernández E, Xiberta J. Restructuring and generation of electrical energy in the Iberian Peninsula. Energy Policy 2007;35:5117–29.

[53] U.S. Energy Information Administration. Country analysis note-Spain; 2013. <http://www.eia.gov/countries/country-data.cfm?fips=SP#pet>.

[54] Pereira Freitas CJ, Pereira Da Silva P. Evaluation of dynamic pass through of carbon prices into electricity prices – a cointegrated VECM analysis. Int J Pub Policy 2013;9(1/2):65–85.

[55] Statistical National Institute (INE). Spanish Energy consumption survey 2011. INE: Madrid, Spain; 2013.

468 P. Fernández González, B. Moreno / Energy Conversion and Management 92 (2015) 459–468

[56] Moreno B, García-Álvarez MT, Ramos C, Fernández-Vázquez E. A General Maximum Entropy Econometric approach to model industrial electricity prices in Spain: a challenge for the competitiveness. Appl Energy 2014;135(5):815–24.

[57] Spanish Ministry of Industry, Energy and Tourism, 2011. Boletines trimestrales de coyuntura. Madrid; 2011.

[58] European Union. Treaty establishing the European Coal and Steel Community (ECSC). Treaty of Paris; 1951. <http://europa.eu/legislation_summaries/ institutional_affairs/treaties/treaties_ecsc_en.htm>.

[59] European Union. Treaty establishing the European Atomic Energy Community (Euratom). Treaties of Rome; 1957. <http://ec.europa.eu/energy/nuclear/ euratom/euratom_en.htm>.

[60] Unite Nations. Kyoto Protocol. United Nations Framework Convention on Climate Change; 1997. <http://unfccc.int/essential_background/kyoto_ protocol/items/1678.php>.

[61] European Commission. Communication from the Commission to the Council, the European Parliament, the Economic and Social Committee and the Committee of the Regions, of 26 April 2000, entitled: ‘‘Action Plan to improve energy efficiency in the European Community’’ [COM(2000) 247].

[62] European Commission. Commission Green Paper of 29 November 2000. Towards a European strategy for the security of energy supply [COM/2000/ 0769]; 2000.

[63] European Commission. Directive 2001/77/EC of the European Parliament and the Council of 27 September 2001 on the promotion of electricity produced from renewable energy sources in the internal electricity market; 2001. <http://eur-lex.europa.eu/LexUriServ/LexUriServ.do?uri=CELEX:32001L0077: ES:NOT>.

[64] European Commission. Directive 2002/91/EC of the European Parliament and of the Council of 16 December 2002 on the energy performance of buildings; 2002. <http://europa.eu/legislation_summaries/other/l27042_es.htm>.

[65] European Commission. Communication from the Commission of 11 September 2002 – the internal market in energy: coordinated measures on the security of energy supply [COM/2002/0488]; 2002. <http://eur-lex.europa.eu/LexUriServ/ LexUriServ.do?uri=CELEX:52002DC0488:EN:HTML>.

[66] European Commission. Directive 2005/89/EC of the European Parliament and of the Council of 18 January 2006 concerning measures to safeguard security of electricity supply and infrastructure investment; 2005.

[67] European Commission. Directive 2006/67/EC of 24 July 2006 imposing an obligation on Member States to maintain minimum stocks of crude oil and/or petroleum products; 2006. <http://eur-lex.europa.eu/LexUriServ/LexUriServ. do?uri=CELEX:32006L0067:EN:NOT>.

[68] European Commission. Commission Green Paper of 8 March 2006: A European strategy for sustainable, competitive and secure energy [COM/2006/0105]; 2006. <http://ec.europa.eu/energy/strategies/2006/2006_03_green_paper_ energy_en.htm>.

[69] European Commission. Communication from the Commission to the European Council and the European Parliament of 10 January 2007 – an energy policy for Europe [COM/2007/0001]; 2007. <http://eur-lex.europa.eu/LexUriServ/ LexUriServ.do?uri=CELEX:52007DC0001:EN:NOT>.

[70] European Commission. Communication 2008/0772 set from the Commission of 13 November 2008 – energy efficiency: delivering the 20% target [COM/ 2008/0772]; 2008. <http://eur-lex.europa.eu/LexUriServ/LexUriServ.do?uri= CELEX:52008DC0772:EN:NOT>.

[71] European Commission. Directive 2009/28/EC of the European Parliament and of the Council of 23 April 2009 on the promotion of the use of energy from renewable sources and amending and subsequently repealing Directives 2001/77/EC and 2003/30/EC; 2009. <http://eur-lex.europa.eu/LexUriServ/ LexUriServ.do?uri=CELEX:32009L0028:EN:NOT>.

[72] European Commission. Directive 2012/27/EU of the European Parliament and of the Council of 25 October 2012 on energy efficiency, amending Directives 2009/125/EC and 2010/30/EU and repealing Directives 2004/8/EC and 2006/ 32/EC; 2012. <http://ec.europa.eu/energy/efficiency/eed/eed_en.htm>.

[73] European Commission. Communication from the Commission to the European Parliament and the Council. Implementing the Energy Efficiency Directive – Commission Guidance. [COM/2013/0762]; 2013. <http://eurlex.europa.eu/ LexUriServ/LexUriServ.do?uri=CELEX:52013DC0762:EN:HTML:NOT>.

[74] Ministry of Industry, Tourism and Commerce (Government of Spain). Energy saving and efficiency action plan 2011–2012. Madrid: Spanish Ministry of Industry, Tourism and Commerce; 2011.

[75] Díaz G. A note on the multivariate Archimedean dependence structure in small wind generation sites. Wind Energy 2014;17(8):1287–95.

[76] Díaz G, González-Morán C. Fischer-Burmeister-based method for calculating equilibrium points of droop-regulated microgrids. IEEE Trans Power Syst 2012;27(2):959–67.

[77] Valenzuela J, Thimmapuram PR, Kim J. Modelling and simulation of consumer response to dynamic pricing with enabled technologies. Appl Energy 2012;96:122–32.

[78] Mallett A, Nye S, Sorrell S. Policy options to overcome barriers to industrial energy efficiency in developing countries: development policy, statistics and research branch. Working Paper of United Nations Industrial Development Organization; 2012. p. 1–45.

[79] Anderson ST, Newell RG. Information programs for technology adoption: the case of energy-efficiency audits. Resour Energy Econ 2004;26:27–50.