Elsevier

Applied Energy

Volume 261, 1 March 2020, 114371
Applied Energy

Inequality and convergence in energy intensity in the European Union

https://doi.org/10.1016/j.apenergy.2019.114371Get rights and content

Highlights

  • Energy intensity convergence is examined by using inequality decomposition.

  • Energy intensity convergence in the European Union is explored over the 2003–2014 period.

  • Results show that convergence mainly occurs in the first years of the period considered.

Abstract

Disparities in energy intensity across the European Union member countries are large, especially after the enlargement that took place in 2004 when Central and Eastern European countries with high energy intensities accessed the European Union. All member countries have committed themselves to the target of reducing energy intensity, however the rate of change in energy intensity differs across countries. In this context, monitoring the convergence process among member countries is crucial to assess the progress in achieving the energy saving targets stated by the European Union. In this article, the convergence process in energy intensity is examined by using an approach based on inequality decomposition. The change in inequality is broken down into two components measuring β-convergence in energy intensity and the re-ranking of countries within the energy intensity distribution, respectively. Since the change in inequality measures the relative variation of energy intensity dispersion, the inequality change in itself is a measure of σ-convergence. Moreover, the inequality change and its components are further decomposed to detect the spatial components of convergence and re-ranking. The convergence in energy intensity in the European Union from 2003 to 2014 is investigated. Results show that convergence mainly occurs in the first years of the period considered, whereas there is a slowdown of the convergence process in the following years. In this second phase, spatial effects on convergence and re-ranking are more evident.

Introduction

Reduction of energy consumption is a priority for the European Union (hereafter, EU) member countries to ensure a sustainable economic growth in the long run [1]. Energy intensity is one of the main indicators used to monitor the progress in achieving energy efficiency targets [2]. The overall trend of energy intensity in the EU is decreasing, however the rate of change in energy intensity differs across countries [3]. In this context, examining the convergence process in energy intensity can explain how relative disparities among countries have changed over time, indicating whether energy intensities are converging or diverging. Such an analysis is more informative when explaining not only whether convergence (or divergence) occurs but also where it occurs. This is particularly relevant to the EU policy-makers after the largest enlargement of the EU that took place in 2004, when new member states of Central and Eastern Europe joined the EU, as their energy intensities were higher than those of old member states of Western Europe.

Convergence in energy intensity among the EU member countries may be examined from the perspective of β-convergence or σ-convergence. However, the two types of convergence are usually examined separately, without clarifying the nexus between them. A first objective of this article is establishing a link between β-convergence and σ-convergence in energy intensity. We show that such a link can be established by using a nonparametric approach based on the decomposition of the change in inequality, which was previously used to analyze income convergence [4]. The second objective of the article is separating the spatial components of the measures of β-convergence, σ-convergence and re-ranking.

β-convergence and σ-convergence are the main two types of convergence originally proposed in the economic growth literature [5], [6], [7] and subsequently studied in other contexts.1 β-convergence in energy intensity occurs when more energy-intensive countries are catching up with less energy-intensive countries. σ-convergence in energy intensity means a reduction of the dispersion of the energy intensity distribution among countries. Mulder [10] examined σ-convergence and β-convergence in manufacturing energy intensity for a subset of OECD countries, explaining the effects of changes in sector structure and energy efficiency on β-convergence. Markandya et al. [11] analyzed β-convergence in energy intensity among the EU countries from 1992 to 2002, finding evidence of convergence in both energy intensity and per capita income. Liddle [12] examined σ-convergence and β-convergence in energy intensity for a subset of OECD countries from 1960 to 2006. In addition, the author explored the re-ranking of countries in energy intensity distribution by using the re-ranking measure suggested by Boyle and McCarthy [13]. Ezcurra [14] also argued that the re-ranking of countries should be taken into account when analyzing convergence, since re-ranking is part of the energy distribution dynamics.2

In the studies cited above, re-ranking, β-convergence and σ-convergence are measured separately, without explicitly linking the respective measures. O’Neill and Van Kerm [4] suggested a nonparametric approach to income convergence analysis that explains the link between re-ranking, β-convergence and σ-convergence. This approach is based on the Jenkins and Van Kerm decomposition of the change in inequality between two points in time [15], in which the change in the Gini index is broken down into a component measuring the re-ranking effect and a component measuring the effect of progressive income growth [4]. Progressive income growth means that the growth rates of low-income countries tend to be higher than those of high-income countries, reducing inequality between initially poorer and richer countries over time. Since the progressive income growth component measures the reduction in inequality due to the fact that initially lower incomes have grown faster than higher incomes, this component can be interpreted as a “distributive measure of β-convergence” [4]. Since the change in the Gini index is a measure of the change in income dispersion, a reduction in inequality indicates σ-convergence among countries. In the O’Neill and Van Kerm approach, σ-convergence is the outcome of the interaction between β-convergence and re-ranking.

In this paper, the O’Neill and Van Kerm approach is applied to the analysis of energy intensity convergence. For this purpose, the method for decomposing the change in the Gini index recently suggested by Mussini and Grossi [16] is used. Mussini and Grossi [16] decomposed the change in carbon dioxide emission inequality in Europe by comparing all pairs of countries.3 Using the Mussini and Grossi decomposition, the change in energy intensity inequality, measuring σ-convergence, is split into two components. The first component measures the re-ranking effect on inequality change. The second component measures the effect of changes in relative disparities between initially more and less energy-intensive countries, which indicates whether β-convergence has occurred. The fact that the decomposition is based on pairwise comparisons between countries enables us to combine the decomposition of inequality change with the spatial decomposition of the Gini index recently suggested by Rey and Smith [22]. The Rey and Smith decomposition separates inequality between neighboring countries from inequality between non-neighboring countries. We obtain a decomposition separating the spatial components of σ-convergence, β-convergence and re-ranking. This decomposition is used to examine convergence in energy intensity among the EU member states from 2003 to 2014. Our findings indicate that convergence mainly occurs in the first years of the period considered, whereas there is a slowdown of the convergence process in the following years. In this second phase, spatial effects on convergence and re-ranking are more evident.

The rest of the article is organized as follows: Section 2 shows how the decomposition of the change in energy intensity inequality can be used to examine energy intensity convergence. Section 3 develops the spatial decomposition of inequality change and its components. Section 4 analyzes convergence in energy intensity among the EU-28 member states. Section 5 discusses results and policy implications. Section 6 concludes.

Section snippets

Energy intensity convergence and inequality change

The concept of convergence was originally developed in the economic growth literature by focusing on convergence between per capita incomes of different countries. Two types of economic convergence were defined. The first type of convergence occurs when initially poorer countries grow faster than richer countries, implying the reduction of disparities between poorer and richer countries. This type of convergence is known as β-convergence [5], [6] and is tested by regressing per capita income

A spatial decomposition of inequality change

ΔG, R and D are measures quantifying different distributional changes in the energy intensity distribution, but these summary measures do not tell us whether changes involve countries that are geographical neighbors. The fact that countries are geographically referenced is not taken into account by the measures of convergence and re-ranking, since they would remain the same if countries exchanged their positions over the map. In other words, ΔG, R and D are locationally invariant, borrowing the

Analysis of energy intensity convergence in the EU-28

The inequality decomposition is used to examine the convergence in energy intensity among the EU-28 member states from 2003 to 2014. The EU is committed to achieving its 20 per cent energy savings target by 2020 and, more broadly, an European Energy Union ensuring a secure and sustainable economic growth [31]. Improvements in energy efficiency play a key role in achieving these goals, and reductions in energy intensity are seen as a proxy of energy efficiency improvements [2]. In such a

Discussion and policy implications

The decomposition results indicate two distinct phases for convergence in energy intensity in the EU-28. In the first phase, coinciding with the sub-period 2003–2007, more energy-intensive countries narrowed the gap with less energy-intensive countries, resulting in a substantial reduction of energy intensity inequality and a limited tendency for non-neighboring countries to re-rank. This convergence process slowed in the second phase, from 2007 to 2014, showing more evident neighborhood

Conclusions

In this article, energy intensity convergence among the EU-28 member states is examined by using a nonparametric approach based on the decomposition of inequality change. This approach, originally developed to explore income convergence [4], is useful to investigate convergence in energy intensity as different distributional changes in the energy intensity distribution can be analyzed together. In the decomposition, σ-convergence in energy intensity is linked with β-convergence and the

Author contribution

I am the sole author of the paper.

Declaration of Competing Interest

The author declares that he has no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

I would like to thank three anonymous reviewers and the editor for insightful comments on earlier versions of this manuscript.

References (36)

  • B. Liddle

    Revisiting world energy intensity convergence for regional differences

    Appl Energy

    (2010)
  • D. Štreimikienė

    Review of financial support from EU Structural Funds to sustainable energy in Baltic States

    Renew Sustain Energy Rev

    (2016)
  • The European Commission. EUROPE 2020. A strategy for smart, sustainable and inclusive growth. The European Commission;...
  • L. Grossi et al.

    Inequality in energy intensity in the EU-28: evidence from a new decomposition method

    Energy J

    (2017)
  • D. O’Neill et al.

    An integrated framework for analysing income convergence

    Manchester School

    (2008)
  • R. Barro et al.

    Convergence

    J Polit Econ

    (1992)
  • X. Sala-i-Martin

    The classical approach to convergence analysis

    Econ J

    (1996)
  • D.T. Quah

    Twin peaks: growth and convergence in models of distribution dynamics

    Econ J

    (1996)
  • Cited by (27)

    • Examining the heterogeneity of financial development in the energy-environment nexus in the era of climate change: Novel evidence around the world

      2022, Energy Economics
      Citation Excerpt :

      In Nigeria, Adom (2015) presents that energy prices, trade openness, and industry structure are asymmetric effects of energy intensity's determinants. Besides the determinants of energy intensity, energy efficiency/intensity is also associated with a strand of other issues such as government expenditure (Yuxiang and Chen, 2010), international inequality (Duro et al., 2010; Mussini, 2020), foreign direct investment (Elliott et al., 2013; Jiang et al., 2014), exports (Zheng et al., 2011), changes in trade structure (Adom and Kwakwa, 2014), and among others (Bilgili et al., 2017; Bu et al., 2019; Chen et al., 2019; Dargahi and Khameneh, 2019; Herrerias, 2012; Huang et al., 2020; Liu and Xie, 2013; Tan and Lin, 2018). Besides the relation between energy resources, energy intensity, and economic growth, the role of financial development and its impacts on energy issues have been presented in the literature.

    • Club convergence in energy efficiency of Belt and Road Initiative countries: The role of China's outward foreign direct investment

      2022, Energy Policy
      Citation Excerpt :

      A range of studies has focused on whether the energy intensity/energy efficiency is converged or diverged, for the various economies, countries or sectors samples using parameter and non-parameter methods. Energy efficiency convergence at different sample levels has been studied, such as at the world level (Liu and Lee, 2020), Latin American countries (Zhang et al., 2021), European Union (Mussini, 2020), China (Cheng et al., 2020b), OECD countries (Mulder and de Groot, 2012), Asian countries (Mishra and Smyth, 2014) etc. Also, its convergence at the level of economic sectors is widely explored, covering the manufacturing sectors (Miketa and Mulder, 2005), energy sectors (Moutinho et al., 2014), residential and transport sectors (Ivanovski et al., 2018), and service sectors (Mulder and de Groot, 2012).

    • Will China's provincial per capita energy consumption converge to a common level over 1990–2017? Evidence from a club convergence approach

      2022, Energy
      Citation Excerpt :

      At the country level, Liddle [27] revisited the convergence of energy intensity across nations by developing two new large data sets, and both revealed the convergence of world energy consumption over 1990–2006. With a focus on the European Union countries, Mussini [28] examined their convergence of national energy intensity over 2003–2014, and the results showed the convergence of energy intensity throughout 2003–2007 and a slowdown of convergence speed in 2007–2014. Furthermore, Huang et al. [29] also found the contribution of globalization level to energy intensity convergence.

    View all citing articles on Scopus
    View full text