ECONOMIC EFFICIENCY, ENERGY CONSUMPTION AND SUSTAINABLE DEVELOPMENT

This paper analyzes structural indicators of economic effi ciency and energy intensity consumption as determinants of sustainable economic development for the selected 33 European countries. The correlation, regression and multivariate factor analyses are applied to test the associations between the selected structural variables of energy intensity consumption, economic effi ciency, and the main driving forces behind these developments. Economic effi ciency is positively associated with expenditures on research and development (R&D) and a greater technological intensity of exports, while at the same time the economic effi ciency of R&D expenditures and technological intensity of exports reduce the energy intensity consumption of the economy. The results suggest that management strategies and policies directed towards R&D expenditures, human capital investments, and technologically intensive export oriented products are improving economic effi ciency performance and contributing to energy saving sustainable economic development. The technological intensity of products reduces energy consumption, which is related to restructuring of energy intensive industries into more advanced and energy saving ones with higher value added per unit of product, but with lower energy consumption per unit of product.


Introduction
The issue of sustainable development and sustainability has been analyzed in theory and application from different perspectives (Pearce, Warford 1993;Dasgupta 2007;Tvaronavicius, Tvaronaviciene 2008;Sobotka, Rolak 2009). Zavadska and Antucheviciene (2006) defi ned a set of indicators in the multicriteria analysis for a rational redevelopment of derelict immovable property from the perspective of sustainable development that includes environmental, social and economic aspects of sustainability. Wegscheider and Sabolovic (2006) underlined the importance of a bio-based economy with nonfood bio-based products. Kryk (2009) evaluated the implementation of the sustainable development concept and effectiveness of environmental protection policy during the economic transformation, European Union (EU) membership and globalization process of the Polish economy. Ighodaro (2010) found the existence of a long-run relationship between energy consumption and economic growth using the Johansen co-integration technique, but the causality depends on the variables used. Electricity consumption and gas utilization are found to determine economic growth, while economic growth determines domestic crude oil production. Chen (2009) investigated causalities between price competition, investment in clean production technologies in the presence of environmental concerns, and consumers' willingness to pay an extra premium for green eco-labelling products and systems in the market to reduce environmental impacts of consumption due to the environmental attributes of green products.
Economic effi ciency, effi cient energy consumption, and sustainable economic development are objectives which can be in a collision of different interests. The business interests of energy suppliers can be in a confl ict with effi cient energy consumption and with development of an alternative energy production and use, particularly from renewable sources of energy, which have impacts on the environment and competitiveness (Nordhaus 1994;Filbeck, Gorman 2004;Stern 2007;Wagner et al. 2007). Burinskienė and Rudzkienė (2007) provide a literature review dealing with economic, ecological and social components of sustainable development with analysis focusing on the aggregated indicators on air pollution variation, income, energy consumption and selected social indicators of national residents. In the literature there exists a recognition of the need for environmental management and sustainable development (Roome 2001;Schaltegger, Synnestvedt 2002;Li et al. 2009), which is beyond the narrow boundaries of an enterprise (Sinding 2000), by considering the sustainable component in economic growth (Priemus 1994;Ginevičius et al. 2008). In this development process there are important positive effects from technological changes and the development of sustainable technologies (Weaver et al. 2000), strategies and management of economic-ecological sustainable development of industries with negative impacts on the environment and sustainable development (Frosch, Gallopoulos 1989).
Moreover, the implications of the global warming and climate change have become one of the most crucial and challenging questions for sustainable economic and environmental development. Due to this, these subject areas have become important for research and policy questions in different sciences (e.g. European Commission 2003Eydeland, Wolyniec 2003). The economics and management of climate change (e.g. Nordhaus 1994;Stern 2007;Wagner et al. 2007) and sustainable economic development have become a constituent part of different documents, global, regional and national policy agendas. These subject areas of the global warming and climate change and their different implications are also causing changes in energy consumption, changing economic effi ciency and sustainability, which motivated our research. We focus on the analysis of the causalities between the intensity in energy consumption and economic effi ciency and their implications for long-term sustainable economic development.
The global warming and climate change have several implications for the economic developments with associated signifi cant implications for energy demands, effi cient supply and consumption of energy, and for sustainable economic development. Energy demand is increasing, which is determined by growth of incomes and by more extreme weather conditions (e.g. Papler, Bojnec 2007). With the global warming and climate changes, and changing economic structures of the developed and some emerging developing economies there are also changes in the seasonal consumption patterns of energy. In the continental northern hemisphere the consumption energy peak is no longer concentrated only on the colder winter season, but also on the hotter summer season.
One of the priorities of sustainable economic development is reduction of the impact of major economic activities on the environment. Burinskienė and Rudzkienė (2007) have explained the association between increase in the economic effi ciency and decrease in the environmental impact. One of the key indicators that reveal economic effi ciency is the amount of energy consumed for production. The previous studies confi rmed the causality between energy consumption and changes in socio-economic structures (Berndt 1978;GiamPietro, Pimentel 1991;Beckerman 1992;Suri, Chapman 1998;Schategger, Synnestvedt 2002;Rutkauskas 2008). Hall et al. (1986) with cross-country analysis confi rmed the strong correlation between the gross domestic product (GDP) and the fuel consumed. This correlation association could vary by different countries and by different periods. However, the signifi cant positive association between energy consumption and economic growth has important implications for further development of the economy's effi ciency and energy consumption in its close connection to problems of sustainable development (GiamPietro, Pimentel 1991;Spangenberg 2004;Blok 2005).
The object and goal of our research is the analyses of the causalities and relations between economic effi ciency and energy intensity consumption in the 33 European countries: EU-27 countries 1 , four European Free Trade Agreement (EFTA-4) countries (Iceland, Liechtenstein, Norway, and Switzerland), and two EU candidate countries (Croatia and Turkey). We analyze structural indicators of economic effi ciency and energy intensity consumption as determinants of sustainable economic development. Restructuring and transformation of the economies from energy intensive industries towards more technologically advanced products and services might lead to higher value added per unit of product, thus higher labour productivity, and energy saving sectors with lower energy consumption per unit of output. This might improve economic performance and lead to higher technological intensity of products, but might at the same time reduce energy intensity consumption and also reduce negative environmental pressures as an important factor of sustainable economic development considering possible (nonrenewable) resources needed for energy production and environmental implications.
The method applied is the statistical multivariate analysis (Kachigan 1991;Hair et al. 1995). We test the hypothesis that economic effi ciency and energy intensity consumption are associated with expenditures on research and development (R&D) and with the share of technologically intensive products in exports by underlying the importance of new challenges in investments in R&D resources and industrial experiences in energy consumption and economic effi ciency. We analyze structural indicators of economic effi ciency and energy intensity consumption as determinants of sustainable economic development for the selected 33 European countries. Economic effi ciency and energy intensity consumption in the selected 33 European countries are investigated in order to establish associations between the level of economic effi ciency on the one hand, and the expenditures on R&D and the share of technologically intensive products in exports on the other. With the cross-country correlation, regression and multivariate factor analyses, we identify signs and intensities of the associations of energy intensity consumption with the human capital and technologically intensive products in exports and the economic effi ciency of investments in R&D, which provide important policy implications for economic effi ciency, energy intensity consumption and sustainable economic development.
The paper is organized as follows. In the next, the second section, we present the methodology and data used. In the third section, the empirical results are presented and explained. The fi nal, fourth section derives main conclusions in order to increase economic effi ciency and to rationalize energy intensity consumption, which are important for sustainable long-term economic development.

Methodology and data used
Different methodological approaches have been used to investigate the relationship between economic effi ciency, energy consumption and sustainable economic development. Ighodaro (2010) employed the Johansen co-integration technique and causality relationship between different proxies of energy consumption, government activities, monetary policy, and economic growth using time-series data for Nigeria. We apply correlation analysis, regression analysis and multivariate factor analysis (e.g. Kachigan 1991;Hair et al. 1995) on the 33 European cross-country datasets to test the sign and statistical signifi cance of the associations between selected structural indicators' variables of energy intensity consumption and the economic effi ciency, and the main driving forces behind these developments across the selected 33 European countries. Sustainable economic development considers both effi cient economy and effi cient energy consumption in order to assure the quality of life also for future generations.
Effi cient economic development across the selected 33 European countries is measured by the labour productivity per person employed relative to the average for the EU-25 (LAB_P_E), where the EU-25 = 100. The energy intensity consumption of the economy (EN_INT) is defi ned as the gross inland consumption of energy divided by GDP at constant prices (1995 = 100) or as kilogram of oil equivalent (kgoe) per 1000 Euro.
The set of associations is tested by two hypotheses with two pertaining regression equations. Each of the regression equations is explained by two explanatory variables: (1) gross domestic expenditures on research and development activities (R&D) and (2) with the share of technologically intensive products in total exports (HTECH): and EN_INT = f(R&D, HTECH).
From (1) and (2) we specify the following empirical cross-section regression models: and where  0 and  0 are regression constants,  1,  2,  1 , and  2 are regression coeffi cients pertaining to the explanatory variables, and u and v are the stochastic error terms. We expect positive associations of the LAB_P_E with the R&D and the HTECH, respectively, but negative associations between the EN_INT with the R&D and the HTECH, respectively.
In addition to correlation and regression analysis, we employ multivariate factor analysis to test the reliability of our regression results with a greater number of included explanatory variables. In the multivariate factor analysis, in addition to the LAB_P_E, EN_INT, R&D and HTECH variables, we include also the additional four explanatory variables: the GDP per capita in purchasing power parity (GDP_PPP), the emissions of CO 2 (GHGEMISS), the share of renewable sources of energy (RE_SH), and the number of graduates in the fi eld of science and technology (H_ED). With the multivariate factor analysis we aim to identify a smaller number of common factors with the highest weights of variables inside them.
The data source for the selected 33 European cross-country analysis for the years 2003 and 2005 focusing on the factors of economic effi ciency, effi ciency in energy consumption, and determinants of sustainable economic development, is the Statistical Offi ce of the European Communities (Eurostat 2006 and.

Summary statistics for the selected European countries
We fi rst provide the summary statistics on the analyzed structural indicators' variables. Prior to interpreting the empirical results, it is worth mentioning that some missing data for some analyzed variables by the analyzed countries are found (e.g. in 2005 for Liechtenstein and Switzerland in the group of EFTA-4 countries) as a reason that the total number of observations (N) is not matched with the number of the analyzed selected 33 European countries. As can be seen from Tables from 1 to 6, the sample includes those European countries which on average are somewhat less developed than the average for the EU-25 countries. Among the least developed countries outside the EU-25 are the candidate countries for EU membership, particularly Turkey. This fi nding is even more considerable and clearly confi rmed in the case of labour productivity. Similarly, as in the case of GDP per capita and labour productivity, there is considerable variation across the analyzed 33 European countries for the gross domestic expenditure on R&D activities as percentage of GDP. This is confi rmed by a large gap between the minimum and maximum values across the analyzed countries and by the standard deviation. The total number of graduates in science and technology per 1000 of population aged 20-29 years increased from 10 to 11.4 graduates, but again with considerable differential across the analyzed 33 European countries and over time. The index of greenhouse gas emissions on average for the analyzed 33 European countries increased over time, but remained below its level in the base year in 1990. As an important fi nding, the economic growth for the analyzed countries between 2003 and 2005 was achieved by the reduced energy intensity consumption in the analyzed 33 European economies. Moreover, the share of electricity from renewable sources energy to gross electricity consumption increased, which can be again considered as a positive outcome for sustainable economic development. Again, there are large variations across the analyzed 33 European countries, as a fi nding which has important policy implications by the countries and for regional European development, with implications for effi ciency in energy intensity consumption and economic effi ciency and sustainable economic development.
Due to large variations in the analyzed variables across the analyzed 33 European countries, we explain the descriptive summary statistics by the more homogeneous groups of the analyzed European countries to exclude possible outliers' biases. Tables from 2 to 4 present the results for the groups of EU countries (for old EU-15, new EU-12, and total EU-27 jointly for the old EU-15 and the new EU-12), while Tables 5 and 6 present similar results for the EFTA-4 countries, and for the candidate EU countries Croatia and Turkey in the years 2003 and 2005, respectively (Table 1). Note: GDP_PPP = GDP per capita in Purchasing Power Standards (PPS), (EU-25 = 100); LAB_P_E = labour productivity per person employed, expressed as GDP in PPP standards per labour active population, relative to EU-25 (EU-25 = 100); R&D = gross domestic expenditure for research and development (R&D) activities as percentage of GDP; H_ED = total number of tertiary graduates in science and technology per 1000 of population aged 20-29; HTECH = exports of high technology products as a share of total exports; GHGEMISS = index of greenhouse gas emissions; percentage change since 1990 = 100, based on CO 2 equivalents and Kyoto Targets in CO 2 equivalents (actual base year = 100); EN_INT = energy intensity consumption of the economy defi ned as gross inland (domestic) consumption of energy divided by GDP (at constant prices, 1995 = 100) in kilogram of oil equivalent per 1000 Euro in constant prices, 1995 = 100; and RE_SH = share of electricity from renewable energy to gross electricity consumption. Source: Authors' calculations from Eurostat (2006 and.   In comparison with the EU-27 countries, the EFTA-4 countries experience similar tendencies, except having the increase in the energy intensity use. On average the EFTA-4 countries vis-à-vis the EU-27 countries have higher GDP per capita, labour productivity, the gross domestic expenditure for R&D activities as percentage of GDP, the index of greenhouse gas emissions, the energy intensity consumption of the economy, and the share of electricity from renewable energy to gross electricity consumption, and vice versa the EFTA-4 countries have a lower total number of tertiary graduates in science and technology per 1000 of population aged 20-29 and lower exports of high technology products as a share of total exports than the EU-27 countries. The EU candidate countries Croatia and Turkey in comparison with the EU-27 countries experience lower GDP per capita, labour productivity, the gross domestic expenditure for R&D activities as a percentage of GDP, total number of tertiary graduates in science and technology per 1000 of population aged 20-29, exports of high technology products as a share of total exports, and the energy intensity consumption of the economy, but higher is the index of greenhouse gas emissions and the share of electricity from renewable energy to gross electricity consumption. Croatia and Turkey, except for the total number of tertiary graduates in science and technology per 1000 of population aged 20-29, and the energy intensity use, have experienced an increase in the analyzed indicators of economic effi ciency, energy consumption and sustainable development as a positive sign for future sustainable economic development.

Correlation analysis
The correlation analysis is used to establish the signs, intensity of associations, and statistical signifi cance of the associations between the pairs of the variables that are used later in the regression and in the multivariate factor analyses. The correlation matrix between the analyzed variables for the 33 European countries indicates positive correlations between the analyzed variables: labour productivity (LAB_P_E) measured as GDP in PPP per labour active person, gross domestic expenditures for research and development activities (R&D), and the share of technologically intensive products in total exports (HTECH) ( Table 7). The intensity of the associations is found stronger in 2003 than in 2005. This could be explained by the EU enlargement in 2004, which seems -in combination with developments in energy markets -to have broadened the scope for the economic effi ciency, energy consumption and sustainable economic development. The estimated Pearson correlation coeffi cient between the R&D and the HTECH is relatively low (0.113 in 2003 and negative in 2005) and modest to low between the LAB_P_E and the R&D (0.589 in 2003 and 0.252 in 2005), and between the LAB_P_E and the HTECH (0.474 in 2003 and0.439 in 2005). On the other hand, less clear associations are found between the energy intensity consumption (EN_INT), the R&D, and the HTECH, respectively: negative in 2003 and vice versa positive in 2005. These correlation results suggest shifts from energy saving towards energy using technologies, which can be explained by the growth in world oil and energy real prices, which has caused the increase of the share of gross domestic consumption of energy in the unit of product or in GDP.

Regression analysis
The regression analysis is used to identify the signs, intensity of the associations, and statistical signifi cance for the associations of economic effi ciency and energy intensity consumption, respectively, as the dependent variables, with the expenditures on R&D and technologically intensive products in exports as the explanatory variables. These two explanatory variables are used as proxy determinants for sustainable economic development.
Both cross-section regression equations for economic effi ciency and energy consumption, respectively, are estimated for the years 2003 and 2005, respectively. As can be seen from Table 8, the regression analysis confi rmed: fi rst, positive and signifi cant associations of labour productivity (LAB_P_E) with the expenditures on R&D and with the share of technologically intensive products in exports (HTECH), respectively. Second, energy intensity is negatively and signifi cantly associated with expenditures on R&D and with the share of technologically intensive products in exports (HTECH), respectively. Third, the comparison between the cross-section regressions for the years 2003 and 2005, respectively, shows an increase in the regression constant for autonomous labour productivity, and vice versa a decline of autonomous energy intensity consumption. These fi ndings for the regression constants are consistent with the theoretical expectations and objectives of sustainable economic development to assure higher labour productivity with lower energy intensity consumption. The regression coeffi cients pertaining to the expenditures in R&D decline, implying a slight deterioration of the transmission of the expenditures in R&D on labour productivity on the one hand, but its effi ciency improvements in energy intensity consumption on the other. The regression coeffi cients pertaining to the high-tech exports are of the theoretically expected sign and are signifi cant: the greater share of the high-tech exports increases labour productivity on the one hand, but decreases energy intensity of consumption on the other.
These regression results suggest that management strategies and economic policies directed towards investments in R&D, and particularly in technologically intensive exports oriented products, are signifi cant for the macro-economic effi ciency performances and for the energy saving sustainable economic developments in the analyzed 33 Eu-ropean countries. The higher technological intensity of exported products is associated negatively with the energy consumption per unit of a product, which is related to the restructuring of production processes from the energy intensive industries towards industries with the higher value added per unit of the product with lower energy consumption per unit of the product.

Multivariate factor analysis
The multivariate factor analysis is used to investigate common factors and main weights of variables in associations between various analyzed variables, in order to fi nd a smaller number of joint variables that represent common factors of the analyzed variables explaining economic effi ciency, energy intensity consumption, and sustainable economic development. The following eight variables are included into the multivariate factor analysis: gross domestic product per capita in purchasing power parity (GDP_PPP), labour productivity (LAB_P_E) measured as GDP in PPP per labour active person, gross domestic expenditures on research and development activities (R&D), graduates in the fi eld of science and technology (H_ED), the share of technologically intensive products in total exports (HTECH), energy intensity consumption of the economy (EN_INT), emissions of CO 2 (GHGEMISS), and the share of renewable resources of energy (RE_SH).
The Kaiser-Meyer-Olkin (KMO) measure of sampling adequacy and Barlett's test of sphericity are used as measures of appropriateness of factor analysis. Bartlett's test of  The extraction method of the maximum likelihood with the rotation method of Oblimin with Kaiser Normalization does not change communalities considerably, but has improved the obtained results on the common factors. The fi rst common factor of the economic effi ciency and expenditures in R&D underlined the importance of ecological factors, whereas the second common factor of the human capital and knowledge with expenditures in R&D underlined the importance of energy intensity consumption. The highest weights in the fi rst common factor are found for variables GHGEMIS, EN_INT (with negative sign), GDP_PPP, and LAB_P_E. In the second common factor the highest weights for variables are found for GDP_PPP, LAB_P_E, R&D, H_ED, and in 2005 to a lesser extent also for GHGEMIS (with negative sign).
The extraction method of the maximum likelihood with the rotation method of Varimax with Kaiser Normalization confi rmed also the two common and most signifi cant factors of economic effi ciency and expenditures in R&D on the one hand, and the human GDP per capita in purchasing power parity (GDP_PPP), labour productivity (LAB_P_E) measured as GDP in PPP per labour active person, gross domestic expenditures for R&D activities, graduates in the fi eld of science and technology (H_ED), the share of technologically intensive products in total exports (HTECH), energy intensity consumption of the economy (EN_INT), emissions of CO 2 (GHGEMISS), and the share of renewable resources of energy (RE_SH).
capital and knowledge with expenditures in R&D on the other. In the fi rst common factor, the highest weights are found for variables GHGEMIS, GDP_PPP, LAB_P_E, and EN_INT. In the second common factor, human capital and knowledge with expenditures in R&D underlined the impact on the number of graduates in the areas of science and technology, where the highest weights are confi rmed for variables R&D, GDP_PPP, LAB_P_E, and for H_ED. The economic effi ciency in the analyzed European countries depends on the expenditures in R&D, energy intensity consumption, and human capital and knowledge investments in R&D.

Conclusion
Energy consumption in the European economies has increased, but greater efforts have been made towards reducing energy intensity consumption of the economy production. In order to achieve stabilities and effi ciencies in the energy markets, both effi cient energy supply and effi cient energy use are important. The energy market stabilities with rational energy consumption can contribute to energy friendly sustainable economic development with the aim of achieving a higher level of living standard of the population. The promotion of economic effi ciency in sustainable economic development with competitive energy supply and effi cient energy use could be an effective strategy providing benefi ts to energy producers, energy consumers, and society's environmental concerns to treat the environment in a sustainable way. According to our empirical results, it is less likely that conservation policy regarding energy consumption would harm economic growth. On the contrary, our results clearly confi rm that sustainable economic development can be achieved by a combination of higher economic effi ciency with at the same time more effi cient energy consumption.
Our results also confi rm that there are signifi cant differences in economic effi ciency, effi ciency in energy consumption and in sustainable economic development between the analyzed 33 European countries. This fi nding has been a reason for presenting and explaining the summary statistics results between more homogenous groups of countries in order to derive similarities and differences between them. The strengthening of the importance of economic effi ciency and sustainable energy projects is an objective of the EU policies and one of the possibilities for using EU structural funds to assist in developing sustainable energy projects to ensure environmental safety and effi cient usage of energy resources towards sustainability (Grundey 2008). Energy consumption can also be biased towards extra energy losses (Oke, Oyedokun 2007) associated with production facilities that occur in energy transfer as a result of ineffi ciencies in equipment and operations. Therefore, more in-depth analysis by countries is an issue for future research.
The EU strategy for smart, sustainable and inclusive growth "Europe 2020" (e.g. Balkytė, Tvaronavičienė 2010; Balkytė, Peleckis 2010) underlines the deeper relationship between sustainable development and competitiveness suggesting different concepts, models of competitiveness, evaluation criteria, challenges and opportunities in the context of international globalisation, economic growth, sustainable competitiveness and sustainable development. Consistently with changing policy context, growing role of sustainable development and the transition to a green economy, our present analysis has been geared towards achieving economic effi ciency, effective and effi cient energy consumption, and sustainable development by researching causalities between structural indicators of economic effi ciency and energy intensity consumption as determinants of sustainable economic development.
The correlation, regression and multivariate factor analyses results consistently show the feasibility of the applied procedure and the contribution of the results in analyzing structural indicators of economic effi ciency and energy intensity consumption as determinants of sustainable long-term economic development. With the cross-country correlation, regression and multivariate factor analysis of the economic effi ciency and energy intensity use variables, we have found a signifi cant association between the energy intensity consumption with two groups of factors: the economic effi ciency of expenditures in R&D and the intensity in energy consumption on the one hand, and the human capital and knowledge investments in R&D and energy intensity consumption on the other.
These results and fi ndings suggest that management strategies and policies directed towards economic effi ciency of expenditures in R&D with human capital knowledge and investments into technologically intensive export-oriented products are signifi cant for the economic effi ciency performance and for energy intensity saving technologies as the important determinants for long-term sustainable economic development. Restructuring and transformation of the European economies from energy-intensive industries towards energy-saving service and more technologically intensive and advanced industries with export-oriented products leads to higher economic effi ciency with higher value added per unit of product and to higher effi ciency in reducing energy consumption by energy saving -technologies with a lower energy use per unit of product. These restructuring and transformation processes towards a higher technological intensity, with higher value added per unit of product and with lower energy consumption per unit of product and their export orientation, are the potential to improve economic performance by reducing energy intensity consumption with implications for reduction in environmental pressures and with greater sustainability in long-term economic development in the direction of smart, sustainable and inclusive growth.