The relationships between total, electricity and biofuels residential energy consumption and income in Latin America and the Caribbean Countries

1 THE RELATIONSHIPS BETWEEN TOTAL, ELECTRICITY AND BIOFUELS RESIDENTIAL ENERGY CONSUMPTION AND INCOME IN LATIN AMERICA AND THE CARIBBEAN COUNTRIES Authors and e-mail of them: María P. Pablo-Romero [email protected] Antonio Sánchez-Braza [email protected] Robert Mejia robert.Mejí[email protected] Department: Departamento de Análisis Económico y Economía Política / Department of Economic Analysis and Political Economy University: Universidad de Sevilla / University of Seville Subject area: 07. Energy and territory Abstract: Controlling residential energy consumption in Latino America and the Caribbean countries is crucial to reduce CO2 emissions, as it has an important energy-saving potential, and its environmental controls are difficult to displace offshore. The aim of this study is to analyze the relationships between residential energy consumption and income for 22 Latin America and the Caribbean countries in the period 1990-2013. For this purpose, residential energy environmental Kuznets curves (EKC) are estimated by taking into account the heterogeneity among the countries by including two control variables: one representing the possible effect of urbanization on residential energy use and the second representing the possible effect of petrol production. 2 The EKC are estimated for total residential energy consumption, for residential electricity consumption and for biofuels and waste energy consumption. The elasticities of total, electricity and biofuels residential energy consumption with respect to income are calculated for each year and country, analyzing the different behavior between countries. Obtained results show that the EKC hypothesis is confirmed for the residential sector when the biofuels energy consumption is considered. Moreover, the results also show that the turning point has been reached in some countries. Nevertheless, the EKC is not confirmed when electricity or total residential energy consumption is considered. Thus, for total residential energy consumption, the elasticity is always positive, growing also as the income does. For electricity energy consumption, the elasticity is also always positive, since although the elasticity decreases until a threshold, from an per capita income value it begins to grow. Keywords: Latin America and the Caribbean Countries, Environmental Kuznets Curve, Residential energy consumption, Energy consumption-income elasticities JEL codes: C23, O52, Q43, Q52 3 THE RELATIONSHIPS BETWEEN TOTAL, ELECTRICITY AND BIOFUELS RESIDENTIAL ENERGY CONSUMPTION AND INCOME IN LATIN AMERICA AND THE CARIBBEAN COUNTRIES 1. Introduction Energy consumption is the main cause of CO2 emissions, so controlling its growth is going to be crucial (Soytas et al. (2007). However, while emissions growth could be controlled by reducing energy consumption, this reduction could also have negative effects on economic growth (Lotfalipour et al., 2010). The relationship between energy consumption and economic growth has been widely studied in recent literature, receiving a new important attention from the environmental perspective (Pablo-Romero and Sánchez-Braza, 2017a). This dynamic relationship between economic growth, energy consumption and environmental pressure has been analyzed for different regions and countries (Chang and Carballo, 2011; Wang et al., 2011; Arouri et al., 2012; Hamit- Haggar, 2012; Omri, 2013; Apergis and Payne, 2014; Cowan et al., 2014; Kasman and Duman, 2015; Pablo-Romero and De Jesús, 2016). Thus, an in-depth knowledge of this relationship is considered to be extremely important for the development of effective energy and environmental policies to promote sustainable development. Moreover, energy use in residential buildings is one of the largest sources of direct and indirect CO2 emissions (U.S. Energy Information Administration, 2013; Estiri, 2015). So, energy consumption in the residential sector is an area with great potential for implementing energy saving policies, which could be achieved not only through technical measures, but also by improving consumer behavior (Ouyang and Hokao, 2009). Likewise, the applied energy policies may be more globally-effective in this sector than in others, thus, the analysis of the evolution of residential energy use becomes interesting (Pablo-Romero and Sánchez-Braza, 2017b). Anyway, it is necessary to thoroughly understand the relationship between economic growth and energy consumption in general, and in the residential sector in particular, to be able to 4 develop environmental and energy efficiency policies, as increasing energy demand is at the centre of the environmental problem (Canadell et al., 2007). Otherwise, in the case of Latino America and the Caribbean countries controlling becomes crucial to reduce CO2 emissions, as it has an important energy-saving potential, and its environmental controls are difficult to displace offshore. The aim of this study is to analyze the relationships between residential energy consumption and income for 22 Latin America and the Caribbean countries in the period 1990-2013. For this purpose, residential energy environmental Kuznets curves (EKC) are estimated by taking into account the heterogeneity among the countries by including two control variables: one representing the possible effect of urbanization on residential energy use and the second representing the possible effect of petrol production. The EKC are estimated for total residential energy consumption, for residential electricity consumption and for biofuels and waste (biomass) energy consumption. From the estimate results, the elasticities of total, electricity and biofuels residential energy consumption with respect to income are calculated for each year and country according to Pablo-Romero and Sánchez-Braza (2015), analyzing the different behavior between countries. Thus, this paper is organized as follows: Section 2 analyzes the methodology used; Section 3 describes the database and the evolution of residential energy consumption and economic growth. Results and discussions are contained in Section 4. Finally, Section 5 concludes. 2. Methodology The general specification model for testing the EKC is expressed as follows: itititititit eYYYAE  3 3 2 21  (1) Where: E is a measure of environmental pressure, being in this study the energy final consumption per capita by the residential sector in natural logarithms, and considering 5 total residential energy (RECpc variable), electricity (RElCpc variable) and (RBCpc variable) biomass consumption; Y is the independent variable of income per capita expressed in logarithms, in this case GDP (GDPpc variable); A represents the sum of time and country effects;  1,  2 and  3 are the parameters of the function to be estimated; i and t denote countries and years, respectively, being i equal to 1, 2..., 22 countries of the sample, and t from 1990 to 2013; finally, e is a random error term.  coefficients values informs about the form of relationship between the variables of residential energy consumption per capita and GDPpc. The Equation [1] may be estimated including and excluding the cubic term of the variable Y, since both specifications are estimated in previous studies. However, according to Luzzati and Orsini (2009), estimations with a cubic term give a greater flexibility to the model. So, this is the option estimated in this study. In order to take into account the heterogeneity of the sample, two control variables have been included in the model (Piaggio and Padilla, 2012). On one side, a control variable representing the percentage of population living in urban areas has been included (Urb variable). This control variable measures the possible effect of represents the possible effect of urbanization in the residential energy consumption. In this sense, previous studies consider state that urbanization may affect the energy use in the residential sector, with it being less energy-intensive in rural areas (Wiedenhofer et al., 2013; Heinonen and Junnila, 2014). Due to the importance of urbanization in economic development, the relationship between urbanization and energy consumption or energyrelated CO2 emissions has been extensively studied, in terms of cross-countries, timeseries or panel scopes (Poumanyvong and Kaneko, 2010; Al-mulali et al. 2012, 2013; Zhang et al., 2015; Shahbaz et al., 2016; Wang et al., 2016). However, most of these studies investigate energy consumption as a gross term, especially ignore the difference between residential sector and other production sectors. On the other hand, a second control variable has been included in order to consider the possible effect of petrol production on residential energy consumption. So, oil 6 production (Oil variable) has been considered expressed in percentage of total production. Therefore, the equation to be estimated may be reformulated as: itititititititit eOilUrbYYYAE  21 3 3 2 21  (2) Several econometric problems have been observed in previous studies when estimating the EKCs. At first, some authors such as Narayan and Narayan (2016) and Narayan et al. (2016) have perceived multicollinearity problems among the explanatory variables. In this paper, the severity of multicollinearity among the explanatory variables was first quantified by using the values of the variance inflation factors (VIFs) (Sánchez- Braza and Pablo-Romero, 2014). Once this problem was shown, the data were converted to deviations from the geometric mean of the sample to mitigate it. This transformation avoids the multicollinearity among the variables, which is tested again by using the VIF values. These VIF values are reported in Table 1, including values obtained for the variables, both with and without being converted to deviations from the geometric mean of the sample. As observed, the VIF values do not exceed the value of 5 for any converted explanatory variable, ruling out possible problems of multicollinearity when considering the variables expressed in terms of deviations with respect to its geometric mean. Table 1 Variance inflation factors. Variables VIF (variables) VIF (deviations from the geometric mean) Y 1399.68 4.07 Y2 6410.57 2.63 Y3 1905.09 4.74 Urb 1.61 1.61 Oil 1.15 1.15 Mean VIF 1120.22 2.89 Source: Own elaboration. 7 Using a topline over variables to indicate these deviations, it is possible to rewrite (2), as follows: it itit ititititit eOilUrbYYYAE  21 3 3 2 21  (3) Additionally, in order to avoid spurious estimates (Stern, 2014), the stochastic nature and properties of the variables were examined: firstly, cross-section dependence in the data was tested by using the Pesaran (2004) CD test; secondly, the Pesaran (2007) second generation panel unit roots tests (CIPS tests) were used to investigate the presence of unit roots; finally, the error correction based on panel cointegration tests proposed by Westerlund (2007) was implemented to test the existence of a structural long-run relationship among the variables. Taking into account the previous tests results, the data were also transformed into first differences, therefore being similar to expressing the EKC in terms of long-run growth rates (Anjum et at., 2014). Using Δ to indicate first differences, it is possible to rewrite (3) as follows (where it A= δt ): it itit ititit t it eCCYYYE   3 3 2 21 (4) Once the Equation (4) has been estimated, the  coefficients obtained may inform about the relationships between the E and Y variables. If all  coefficients are positive, then an increasing relationship exists between E (residential energy consumption) and Y (GDPpc). However, the EKC hypothesis is verified and presents the classic inverted U shape if  1>0,  2<0 and  30. Likewise, if  1>0,  2>0 and  3<0, the curve may present the U shape from a certain Y value, which depends on the  coefficients values. Alternatively, if  1>0,  2<0 y  3>0, then the curve presents a N shape (Dinda, 2004). In the case that the EKC exists, a threshold or a turning point may be calculated, making the elasticity of E with respect to Y equal to zero. Then, from the estimate results, the elasticities of total, electricity and biofuels residential energy consumption with respect to income are calculated for each year and country according to Pablo-Romero and Sánchez-Braza (2015), analyzing the different behavior between countries. This elasticity may be calculated for each year and country as follows: 8 2 321 32 itit it YYelas   (5) These elasticities measure the residential energy (total, electricity or biomass) consumption sensitivity with respect to a change in the GDPpc, for each year and country. Thus, it is a measure of the responsiveness of E to an increase in Y. Therefore, these elasticities allow the possibility of analyzing different behavior between countries. 3. Data. 3.1. Data sources. This study uses a panel data of 22 countries over the period 1990-2013. The countries included in the analysis are the followings: Argentina, Bolivia, Brazil, Chile, Colombia, Costa Rica, Cuba, Dominican Republic, Ecuador, El Salvador, Guatemala, Haiti, Honduras, Jamaica, Mexico, Nicaragua, Panama, Paraguay, Peru, Trinidad, Uruguay and Venezuela. Residential energy consumption data proceed from the International Energy Agency (IEA, 2017) which offers energy data for its member countries as well as for wide range of non-member countries. Energy final consumption per capita by the residential sector is analyzed, considering total residential energy (RECpc variable), electricity (RElCpc variable) and (RBCpc variable) biomass consumption. Data are expressed in natural logarithm of kilograms of oil equivalent consumption by the residential sector per inhabitants. The rest of the data (GDP, total population, urban population and oil production) come from the World Bank Development Indicators (World Bank, 2017). Total population is used to convert energy consumption and income expressed in absolute terms in per capita terms. Y variable is expressed in natural logarithm of GDP in constant 2005 U.S. dollars per inhabitants. Urban population variable (Urb variable) is expressed in percentage of total population. Finally, Oil production (Oil variable) variable is expressed in percentage of total production. 9 3.2. Descriptive analysis. Table 2 presents the main descriptive statistics of the variables used to estimate the models. The between statistics refer to the average data values of each individual country, while the within statistics refer to intra groups values over time for each country and to the variation from each individual country’s average. Table 2 shows that the standard deviation of the data across countries is higher than across time for all variables during the considered period. Table 2 Descriptive statistics (1990-2013). Variable Mean Std. Dev. Min. Max. Observations Residential energy consumption pc (in logs) (RECpc) overall 3.119 0.801 -0.820 4.120 N = 528 between 0.796 -0.149 3.895 n = 22 within 0.186 2.448 3.717 T = 24 Residential electricity consumption pc (in logs) (RElCpc) overall 1.449 0.935 -2.180 3.280 N = 528 between 0.926 -1.749 2.717 n = 22 within 0.235 0.761 2.078 T = 24 Residential biomass consumption pc (in logs) (RBCpc) overall 2.353 1.009 0.050 4.030 N = 528 between 0.988 0.364 3.785 n = 22 within 0.290 1.159 3.259 T = 24 GDPpc (Y) (in logs) overall 3.143 0.575 1.470 4.500 N = 528 between 0.559 1.585 4.067 n = 22 within 0.179 2.615 3.696 T = 24 % of urban population (Urb) overall 63.928 19.113 8.534 94.983 N = 528 between 19.148 9.614 92.262 n = 22 within 3.828 52.050 79.708 T = 24 % of oil production (Oil) (in logs) overall 5.077 4.852 0.000 12.190 N = 528 between 4.956 0.000 12.025 n = 22 within 0.234 3.775 5.785 T = 24 Source: Own elaboration. 16 First, Graph (a) in Figure 2 shows the evolution of these calculated elasticity values by countries, when total residential energy consumption is considered. The results obtained from estimates in Column (c) in Table 3 were used; therefore the final equation used is the following: 2 036.0*3054.0*2149.0)( itit it YYtotalela  (6) The estimated elasticity values are not constant over period analyzed. The black line represents the trend of the elasticity median spline that shows a stable trend during the first tranct of the period considered but an important increasing trend from 2004. Anyway, notable differences among the countries RECpc elasticities values are registered, although the great majority of countries show a clear increasing trend in the last years of the period. All estimated elasticity values are positive and are in the range of values between 0.1 and 0.3, with de exception of Trinidad that holds a significant growth, going from the value 0.2 registered in 1990 to almost reaching the value of 0.5 at the end of the period. Among the countries with the highest elasticity values, Panama, Uruguay, Chile, Venezuela and Argentina finalized the period reaching values above 0.3. On the other hand, Paraguay, Jamaica, El Salvador, Guatemala, Bolivia and Nicaragua represent a below average level GDPpc, showing an stable evolution along the considered period around 0.12 value. Secondly, Graph (b) in Figure 2 shows the evolution elasticity values by countries, when residential electricity consumption is considered. The results obtained from estimates in Column (b) in Table 4 were used; therefore the final equation used is the following: 2 254.0*3246.0*2343.0)( itit it YYelectela  (7) Noticeable differences between countries elasticity values may be observed. The black line representing the median shows a stable trend with similar values at the initial and final years of the considered period. All values obtained are positive, and in the great 17 majority of cases they are in the range of variation between 0.3 and 0.7. Anyway, some groups of countries with different evolution of their elasticity may be highlighted. Haiti presents again an exceptional elasticity values with constant with an increasing trend. Nicaragua, Honduras y Bolivia register at the beginning of the period values above the unit, but after a marked downward trend finalized the period with values clearly lower than unity. Conversely, Trinidad is distinguished by a significant upward trend, especially in recent years, starting with an around 0.3 value and finalizing the period with values close to the unity. El Salvador, Dominican Republic, Peru, Guatemala y Peru show values above the average, but with a marked downward trend throughout the whole period. Finally, other countries as Brazil, Mexico, Costa Rica, Colombia and Ecuador have values slightly below the average, showing a constant trend evolution along time. Thirdly, Graph (c) in Figure 2 shows the evolution elasticity values by countries, when residential biomass consumption is considered. The results obtained from estimates in Column (c) in Table 5 were used; therefore the final equation used is the following: 2 043.0*3088.0*2195.0)( itit it YYbiomela  (8) In general terms, a moderate decreasing trend is registered at the beginning of the considered period that becomes in an important decreasing trend from 2004-2005. Nicaragua, Bolivia, Guatemala, Jamaica and Paraguay stand out as countries with values above the average, and also maintaining a constant trend throughout the period, without registering the general decreasing trend. Costa Rica, Brazil and Mexico show values slightly below the average. Argentina, Venezuela, Chile, Panama and Uruguay also register elasticity values below the average but with a much sharper decreasing trend at the end of the period, reaching negative values in recent years. Finally, it may be highlighted the evolution of Trinidad, with a more strong decreasing trend and registering negative important values from 1998. 18 4.3. Evolution of residential energy consumption elasticities by GDPpc levels. Finally, Figure 3 displays the elasticity values of residential energy consumption (total, electricity and biomass) per capita with respect to GDPpc for each GDPpc level. Figure 3 Estimated elasticities of residential energy consumption (total, electricity and biomass) per capita with respect to GDPpc by GDPpc level. Source: Own elaboration. Considering firstly the total residential energy consumption, the elasticity shows a U shape. Initially, as GDPpc increases the elasticity slightly decreases, although always is positive and being the minimum value reached by this elasticity around 0.12. Anyway, elasticity start to increase as GDPpc does from a GDPpc value close to 2.65 (in logs). From this value, GDPpc increases origin a positive and growing trend in total energy consumption. Therefore, as mentioned above, it is shows that the EKC hypothesis is not supported for total residential energy consumption. 19 In the case of the residential electricity consumption, although results in Table 7 show that the EKC hypothesis could be supported with a N shape, Figure 3 shows that it does not happen. Initially the elasticity decrease as GDPpc increase, with a strong downward trend. Then, this trend becomes positive from a GDPpc value close to 3.46 (in logs) registering a minimum value for this elasticity around 0.26. So, as in the case of total energy consumption, GDPpc increases origin a positive and growing trend in residential energy consumption. Nevertheless, the elasticity values show that the ECK hypothesis is supported for the case of the residential biomass consumption, reaching the ECK turning point for a GDPpc value around 3.86 (in logs), when the elasticity values become negatives. Below this value, elasticities are positive, rising until a GDP per capita level close to 2.46. From this value, being the maximum value reached by this elasticity around 0.26, elasticity values start to decrease. Therefore, for values above these levels for GDPpc increases origin a negative and degressive trend in residential biomass consumption. 5. Discussion and conclusions. Energy consumption is the main cause of CO2 emissions, so controlling its growth is going to be crucial. However, while emissions growth could be controlled by reducing energy consumption, this reduction could also have negative effects on economic growth. Then, the study and in-depth knowledge of this relationship is considered to be extremely important for the development of effective energy and environmental policies to promote sustainable development. On the other hand, energy consumption in the residential sector is an area with great potential for implementing energy saving policies, in the sense that the applied energy policies may be more globally-effective in this sector than in others, thus, the analysis of the evolution of residential energy use becomes interesting. In this study it has been analyzed the relationships between residential energy consumption and income for 22 Latin America and the Caribbean countries in the 20 period 1990-2013. Residential energy environmental Kuznets curves (EKC) have been estimated by taking into account the heterogeneity among the countries by including two control variables: one representing the possible effect of urbanization on residential energy use and the second representing the possible effect of petrol production. The EKC are estimated for total residential energy consumption, for residential electricity consumption and for biofuels and waste (biomass) energy consumption, and the corresponding elasticities of energy consumption with respect to income have been also calculated for each year and country. Obtained results show that the EKC hypothesis is confirmed for the residential sector when the biofuels energy consumption is considered. Moreover, the results also show that the turning point has been reached in some countries. Nevertheless, the EKC is not confirmed when electricity or total residential energy consumption is considered. Thus, for total residential energy consumption, the elasticity is always positive, growing also as the income does. For electricity energy consumption, the elasticity is also always positive, since although the elasticity decreases until a threshold, from an per capita income value it begins to grow. References Al-mulali, U., Binti Che Sab, C.N., Fereidouni, H.G., 2012. Exploring the bi-directional long run relationship between urbanization, energy consumption, and carbon dioxide emission. Energy 46, 156–167. Al-mulali, U., Fereidouni, H.G., Lee, J.Y.M., Sab, C.N.B.C, 2013. Exploring the relationship between urbanization, energy consumption, and CO2 emission in MENA countries. Renewable and Sustainable Energy Reviews 23, 107–112. Anjum, Z., Burke, P., Gerlagh, R., Stern, D., 2014. Modeling the emissions-income relationship using long-run growth rates. CCEP Working Paper 1403. Canberra: Centre for Climate Economic & Policy, Crawford School of Public Policy, The Australian National University. 21 Apergis, N., Payne, J.E., 2014. The causal dynamics between renewable energy, real GDP, emissions and oil prices: Evidence from OECD countries. Applied Economics 46(36), 4519–4525. Arouri, M.E.H., Youssef, A.B., M’henni, H., Rault, C., 2012. Energy consumption, economic growth and CO2 emissions in Middle East and North African countries. Energy Policy 45, 342–349. Canadell, J.G., Le Quéré, C., Raupach, M.R., Field, C.B., Buitenhuis, E.T., Ciais, P., Marland, G., 2007. Contributions to accelerating atmospheric CO2 growth from economic activity, carbon intensity, and efficiency of natural sinks. Proceedings of the National Academy of Sciences 104(47), 18866–18870. Chang, C., Carballo, C.F.S., 2011. Energy conservation and sustainable economic growth: The case of Latin America and the Caribbean. Energy Policy 39(7), 4215– 4221. Cowan, W.N., Chang, T., Inglesi-Lotz, R., Gupta, R., 2014. The nexus of electricity con-sumption, economic growth and CO2 emissions in the BRICS countries. Energy Policy 66, 359–368. Dinda, S., 2004. Environmental Kuznets Curve hypothesis: A survey. Ecological Economics 49, 431–455. Estiri, H., 2015. The indirect role of households in shaping US residential energy demand patterns. Energy Policy 86, 585–594. Greene, W., 2000. Econometric Analysis, fourth edition. Upper Saddle River, New Jersey: Prentice Hall. Hamit-Haggar, M., 2012. Greenhouse gas emissions, energy consumption and economic growth: a panel cointegration analysis from Canadian industrial sector perspective. Energy Economics 34, 358–364. Heinonen, J., Junnila, S., 2014. Residential energy consumption patterns and the overall housing energy requirements of urban and rural households in Finland. Energy and Buildings, 76, 295–303. IEA, 2017. International Energy Agency. Statistics report, energy balances. Paris: OECD/IEA. Kasman, A., Duman, Y.S., 2015. CO2 emissions, economic growth, energy consumption, trade and urbanization in new EU member and candidate countries: A panel data analysis. Economic Modelling 44, 97–103. 22 Lotfalipour, M.R., Falahi, M.A., Ashena, M., 2010. Economic growth, CO2 emissions, and fossil fuels consumption in Iran. Energy 35(12), 5115–5120. Luzzati, T., Orsini, M., 2009. Investigating the energy-environmental Kuznets curve. Energy 34(3), 291–300. Narayan, P.K., Narayan, S., 2010. Carbon dioxide emissions and economic growth: Panel data evidence from developing countries. Energy Policy 38(1), 661–666. Narayan, P.K., Saboori, B., Soleymani, A., 2016. Economic growth and carbon emissions. Economic Modelling 53, 388–397. Omri, A., 2013. CO2 emissions, energy consumption and economic growth nexus in MENA countries: Evidence from simultaneous equations models. Energy Economics 40, 657–664. Ouyang, J., Hokao, K., 2009. Energy-saving potential by improving occupants’ behavior in urban residential sector in Hangzhou City, China. Energy and Buildings 41(7), 711–720. Pablo-Romero, M.P., De Jesús, J., 2016. Economic growth and energy consumption: The Energy-Environmental Kuznets Curve for Latin America and the Caribbean. Renewable and Sustainable Energy Reviews 60, 1343–1350. Pablo-Romero, M.P., Sánchez-Braza, A., 2015. Productive Energy Use and Economic Growth: Energy, Physical and Human Capital Relationships. Energy Economics 49, 420–429. Pablo-Romero, M.P., Sánchez-Braza, 2017a. The changing of the relationships between carbon footprints and final demand: Panel data evidence for 40 major countries. Energy Economics 61, 8–20. Pablo-Romero, M.P., Sánchez-Braza, 2017b. Residential energy environmental Kuznets curve in the EU-28. Energy 125, 44–24. Pesaran, M., 2004. General Diagnostic Tests for Cross Section Dependence in Panels. Cambridge Working Papers WP0435. Cambridge: Faculty of Economics, University of Cambridge. Pesaran, M.H., 2007. A simple panel unit root test in the presence of cross section dependence. Journal of Applied Econometrics 22(2), 265–312. Piaggio, M., Padilla, E., 2012. CO2 emissions and economic activity: Heterogeneity across countries and non-stationary series. Energy Policy 46, 370–381. 23 Poumanyvong P, Kaneko S, 2010. Does urbanization lead to less energy use and lower CO2 emissions? A cross-country analysis. Ecological Economics 70, 434–444. Sánchez-Braza, A., Pablo-Romero, M.P., 2014. Evaluation of property tax bonus to promote solar thermal systems in Andalusia (Spain). Energy Policy 67, 832–843. Shahbaz, M., Loganathan, N., Muzaffar, A.T., Khali, A., Jabran, M.A., 2016. How urbanization affects CO2, emissions in Malaysia? The application of STIRPAT model. Renewable and Sustainable Energy Reviews 57, 83–93. Stern, D., 2014. The environmental Kuznets curve: A primer. CCEP Working Paper 1404. Canberra: Centre for Climate Economic & Policy, Crawford School of Public Policy, The Australian National University. Soytas, U., Sari, R., Ewing, B.T., 2007. Energy consumption, income, and carbon emissions in the United States. Ecological Economics 62(3), 482–489. U.S. Energy Information Administration, 2013. International energy outlook 2013 International Energy Outlook 2013 with projections to 2040. Washington, D. C.: U.S. Energy Information Administration, Office of Energy Analysis, U.S. Department of Energy. Wang, Q., Wu, S.D., Zeng, Y.E., Wu, B., 2016. Exploring the relationship between urbanization, energy consumption, and CO2, emissions in different provinces of China. Renewable and Sustainable Energy Reviews 54, 1563–1579. Wang, S.S., Zhou, D.Q., Zhou, P., Wang, Q.W., 2011. CO2 emissions, energy consumption and economic growth in China: A panel data analysis. Energy Policy 39(9), 4870–4875. Westerlund, J., 2007.Testing for error correction in panel data. Oxford Bulletin of Economics and Statistics 69(6), 709–748. Wiedenhofer, D., Lenzen, M., Steinberger, J.K., 2013. Energy requirements of consumption: Urban form, climatic and socio-economic factors, rebounds and their policy implications. Energy Policy 63, 696–707. Wooldridge, J., 2002. Econometric Analysis of Cross Section and Panel Data. Cambridge, Massachusetts: MIT Press. World Bank, 2017. World development indicators. Washington DC: The World Bank. Zhang, Y.J., Yi, W.C., Li, B.W., 2015. The impact of urbanization on carbon emission: empirical evidence in Beijing. Energy Procedia 75, 2963–2968.