Elsevier

Environmental Pollution

Volume 230, November 2017, Pages 291-301
Environmental Pollution

Statistical polarization in greenhouse gas emissions: Theory and evidence

https://doi.org/10.1016/j.envpol.2017.06.074Get rights and content

Highlights

  • We study the evolution of global polarization in GHG emissions.

  • We consider the four main GHGs: CO2, CH4, N2O and F-gases.

  • We use the multidimensional polarization indices (Gigliarano and Mosler, 2009).

  • We consider an endogenous grouping of countries (Aghevli and Mehran, 1981).

  • Most of the polarization indices showed a slightly increasing pattern.

Abstract

The current debate on climate change is over whether global warming can be limited in order to lessen its impacts. In this sense, evidence of a decrease in the statistical polarization in greenhouse gas (GHG) emissions could encourage countries to establish a stronger multilateral climate change agreement. Based on the interregional and intraregional components of the multivariate generalised entropy measures (Maasoumi, 1986), Gigliarano and Mosler (2009) proposed to study the statistical polarization concept from a multivariate view. In this paper, we apply this approach to study the evolution of such phenomenon in the global distribution of the main GHGs. The empirical analysis has been carried out for the time period 1990–2011, considering an endogenous grouping of countries (Aghevli and Mehran, 1981; Davies and Shorrocks, 1989). Most of the statistical polarization indices showed a slightly increasing pattern that was similar regardless of the number of groups considered. Finally, some policy implications are commented.

Introduction

Climate change is one of the most important challenges facing the international community nowadays. Given its possible far-reaching consequences for ecosystems and the quality of life of hundreds of millions of people, climate change is a political issue on the global agenda as it was firmly established in the Third and Fourth Assessment Reports of the Intergovernmental Panel on Climate Change (IPCC, 2001, IPCC, 2007).

The principal causes of global warming are the anthropogenic emissions of greenhouse gases (GHGs), especially carbon dioxide (CO2) from fossil fuel combustion. However, emissions of non-CO2 GHGs, such as methane (CH4), nitrous oxide (N2O) and fluorinated gases (F-gases), also alter significantly the climate. A recent Greenhouse Gas Bulletin (World Meteorological Organization, 2016) shows that the concentration of CO2, CH4, N2O has increased by 144, 256 and 121 percent since the year 1750, respectively. The increase in global CO2 concentration is basically due to the fossil fuel combustion. CH4 is emitted into the atmosphere from both natural (about 40 percent) and anthropogenic sources (approximately 60 percent). In the case of N2O, close to 60 percent is emitted into the atmosphere by natural sources and about 60 percent comes from human activities. Although F-gases are still low in abundance, they are potent GHGs which are increasing at relatively rapid rates given its anthropogenic origin.

Given that non-CO2 GHGs contribute more to global warming per unit mass than CO2 (U.S. Environmental Protection Agency, 2012) and to reduce them is a relatively cheap complement to the cost associated to CO2-only mitigation (U.S. Environmental Protection Agency, 2006), these gases have an important function in limiting global climate change. In addition, as these gases have much shorter lifetimes than CO2, reducing their emissions offers an extra opportunity to curb climate change (Montzka et al., 2011, Weyant et al., 2006, Rao and Riahi, 2006).

In recent years the concept of statistical polarization has emerged to capture the inherent conflict or instability of a distribution. While inequality measures study the dispersion of a distribution with respect to a reference value, the statistical examination of polarization consists in identifying the appearance of poles in that distribution, which is related to multimodal distributions (Esteban and Ray, 1994, Wolfson, 1994). According to a specific attribute, the notion of statistical polarization considers the population divided into different groups so that, the groups are internally homogeneous but different each other.

In the environmental field, global negotiations on reducing emissions are constructed through alliances of groups of countries with conflicting interests. Thus, developed and developing countries have polarized positions given their different environmental responsibilities and level of development. Some experts have suggested that climate change will intensify resource scarcity, population displacements and fuel conflicts, being these effects particularly serious in developing countries where infrastructure is missing (Salehyan, 2008).

Given that climate change may cause conflicts between the haves and the have-nots, increasing even global inequality, international statistical polarization using only one gas has already been analysed in various studies. Thus, Ezcurra (2007) analysed the convergence in per capita CO2 emissions using the EGR indices for the period 1960–1999. Meanwhile, Duro and Padilla (2008) used this same measure to investigate the same fact between 1971 and 2001. Duro (2010) examined the statistical polarization in per capita CO2 emissions with exogenous groups based on the ZK measure (Zhang and Kanbur, 2001), whose main differential advantage lies in its factor-decomposability. Duro and Padilla (2013) analysed the degree of statistical polarization in the international distribution of per capita CO2 emissions in the European Union, where the countries are grouped according to two criteria: their similarity in terms of emissions –endogenously– and their geographical location –exogenously. Finally, Duro and Teixidó-Figueras (2014) explored the distribution of per capita CO2 emissions for the period 1992–2010 comparing different statistical polarization measures.

The principal limitation of the previous studies is that they only consider the distribution of CO2 emissions, not giving a real picture of the international situation. In this sense, the extension of the preceding works to the analysis of the international distribution of the main GHG emissions is quite useful. On the one hand, it would give complete information about the possible political consequences of the emissions distribution, in terms of conflicts, and the probability of implementing international agreements. On the other hand, it would also give new insights of the Ecological Unequal Exchange framework.1

Using the multivariate inequality measures proposed by Maasoumi (1986), and considering their decomposition into the between- and within-group inequality components, it is possible to obtain statistical polarization indices from a multivariate perspective (Gigliarano and Mosler, 2009). The main aim of this article is therefore to apply these indices to study the international statistical polarization in the distribution of the principal GHG emissions: CO2, CH4, N2O and F-gases. Specifically, the empirical analysis is carried out for the time period 1990–2011 considering an endogenous grouping of countries (Aghevli and Mehran, 1981, Davies and Shorrocks, 1989). Moreover, this paper is an extension of the results recently obtained by Remuzgo et al. (2016) on the study of the evolution of global inequality in GHG emissions from 1990 to 2011.

To the best of our knowledge, this is the first attempt to use multivariate statistical polarization measures for analysing, in a joint manner, the global distribution of GHG emissions.2 In this sense, the use of quantitative methods for analysing the historical trend of global statistical polarization in GHG emissions is a significant step towards solving the problem of climate change. Moreover, modelling the social effects of global warming will facilitate the dialogue on this issue between national governments, international organizations, non-profit groups and multinational firms in order to design effective global polices.

The rest of the paper is organised as follows. Section 2 examines the concept of statistical polarization, including the principal measures proposed in the literature. Next, the multivariate statistical polarization indices used in this paper are detailed. The main results of the analysis are exposed in sections 4 Multivariate inequality analysis, 5 Multivariate statistical polarization analysis. Finally, with the conclusions of the chapter, some policy implications are discussed.

Section snippets

The concept of statistical polarization

Inequality measures quantify the dispersion of a distribution with respect to a reference value –usually the arithmetic mean. However, to study some social phenomena is interesting to use a measure of the degree to which population is clustered around a number of poles at a certain distance. The concept of statistical polarization (hereinafter referred to as polarization) is directly related to the emergence of social tensions caused by a general dissatisfaction (Esteban and Ray, 1994, Wolfson,

Methodology and data

The methodology applied in the multivariate polarization study is described in this section. In order to construct multivariate polarization indices, we proceed in two stages. In the first phase, the multivariate inequality indices based on the concept of generalised entropy are obtained, which can be expressed as the sum of the within and the between inequality components (Maasoumi, 1986, Maasoumi and Nickelsburg, 1988). Using the previous decomposition, the multivariate polarization indices

Multivariate inequality analysis

The multivariate indices used in this paper can be additively decomposed by population sub-groups. In this paper, the creation of groups has been made using the method proposed by Aghevli and Mehran (1981), technique which was later refined by Davies and Shorrocks (1989). This procedure involves minimizing disparities within each group considered. For this purpose, it is necessary to calculate the average emission between adjacent groups to find the border between them. This process converges

Multivariate statistical polarization analysis

Having analysed the evolution of inequality in GHG emissions from a multivariate perspective, the polarization in the distribution of the four most important gases –CO2, CH4, N2O and F-gases–, is studied in the same time period, 1990–2011, using the measures detailed in Section 3. Table 3 shows the evolution of the multivariate polarization –using the P1, P2 and P3 indices– for the four main pollutants over the period 1990–2011.

The γ parameter has been set to zero because it is the only

Conclusions and policy implications

Although the per capita emissions are frequently more abundant in the rich countries, there are considerable exceptions. For example, some middle-income countries have similar per capita emissions levels to those of the wealthier economies. Given the existence of a wide variety of countries with different features and similar per capita emissions patterns, one-size-fits-all strategies are not likely to be favourable to the implementation of a stronger international environmental agreement.

In

Acknowledgements

The authors thank to Ministerio de Economía y Competitividad (Project ECO2016-76203-C2-1-PE) for the partial support of this work. In addition, this work is part of the Research Project “New methods for the empirical analysis of financial markets” of the Santander Financial Institute (SANFI) of UCEIF Foundation resolved by the University of Cantabria and funded with sponsorship from Fundación Banco Santander (Project APIE 1/2015–17).

References (41)

  • CAIT

    Climate Data Explorer

    (2014)
  • J.Y. Duclos et al.

    Polarization

  • J.A. Duro

    Decomposing international polarization of per capita CO2 emissions

    Energy Policy

    (2010)
  • J.A. Duro et al.

    Cross-country polarization in CO2 emissions per capita in the European Union: changes and explanatory factors

    Environ. Resour. Econ.

    (2013)
  • J.A. Duro et al.

    World polarization in carbon emissions, potential conflict and groups: an updated revision

    Energy Policy

    (2014)
  • J.M. Esteban

    Desigualdad y polarización. Una aplicación a la distribución interprovincial de la renta en España

    Rev. Econ. Apl.

    (1996)
  • J.M. Esteban

    Polarización Económica en la Cuenca Mediterránea

    (2002)
  • J.M. Esteban et al.

    Extensions of a Measure of Polarization, with an Application to the Income Distribution of Five OECD Countries

    (1999)
  • J.M. Esteban et al.

    On the measurement of polarization

    Econometrica

    (1994)
  • J.M. Esteban et al.

    Comparing polarization measures

    Oxf. Handb. Econ. Peace Confl.

    (2012)
  • Cited by (0)

    This paper has been recommended for acceptance by Dr. Hageman Kimberly Jill.

    View full text