Elsevier

Energy

Volume 148, 1 April 2018, Pages 687-700
Energy

How far is Colombia from decoupling? Two-level decomposition analysis of energy consumption changes

https://doi.org/10.1016/j.energy.2018.01.141Get rights and content

Highlights

  • Tapio's index values and the decoupling index values are not so dissimilar.

  • Strong decoupling subperiods between 2000 and 2007 are identified but not after 2008.

  • LMDI Population and activity effects drive energy consumption in Colombia.

  • The Transport and Industrial sectors are key drivers of energy consumption changes.

  • Current decoupling measures oriented are right but more efforts should be made.

Abstract

A decoupling elasticity analysis and a two-level decomposition analysis of energy consumption in Colombia from 2000 to 2015 are developed. Firstly, the decoupling elasticity approach is used to analyse the importance of energy consumption changes in relation to the GDP changes. Then, a Logarithmic Mean Divisia Index analysis is carried out, decomposing the changes in energy consumption into four effects: Population, Activity, Structural and Intensity. Secondly, a decoupling index determines the main drivers of the inhibiting effect on energy consumption. The results show that the Population and Activity effects contribute to the country's increase of energy consumption, while the Intensity effect and, to a lesser extent, the Structure effect help to decrease it. From a sectoral perspective, variations in the energy consumption are mainly caused by the Transport and Industrial sectors. In the light of the results obtained, current decoupling-oriented measures are steps in the right direction, but more efforts should be made because until now they have not been effective. New policy recommendations are provided.

Introduction

Colombia has the second largest hydropower potential in Latin America, after Brazil [1]. Brazil, Russia, Canada, the United States, Indonesia, China and Colombia are the top seven countries with water availability exceeding 2000 km3 [2]. In 2015, Colombia had a total installed electricity generation capacity of 16.4 GW, with a share of 70% of hydropower (high, mid and small plants together). As a consequence, Colombia's electricity sector is highly vulnerable to sufficient water availability [3]. During 2015 and 2016, the droughts suffered in Colombia as a result of “El Niño" were close to causing a blackout due to the drop in the electricity generation of the hydroelectric plants. Gutiérrez and Dracup [4] provided evidence of the relationship between “El Niño” and hydrology. Arias-Gaviria et al. [3] analyse the impact of “El Niño” on Colombian's power generation. Low levels of rain caused by “El Niño” provokes more extreme and longer dry seasons than usual, reduces the country's total water reservoirs and has led to an energy crisis in Colombia. A recent state of the art of “El Niño” is available in Trenberth [5]. Even more recently, Smith and Ubilava [6] analyse a causal relationship between “El Niño” and economic growth in Colombia.

The phenomenon “El Niño" occurs with some regularity. Evidence for the period 1979–2009 with data collected from 341 data stations is provided by Córdoba-Machado et al. [7] and by Smith and Ubilava [6] for the period 1961–2015. The possibility of building predictive indicators was analysed by Gutiérrez and Dracup [3]. Colombia needs to take measures to guarantee electricity supply in particular and energy in general. These measures can come from the Generation System [8], or measures oriented to the management of demand, particularly those aimed at acting on the drivers of energy consumption. This paper focuses on the latter. A pool of policy measures is required. In this sense, the possibility of Colombia moving properly towards decoupling between energy consumption and economic growth is a desirable outcome. If limiting the global average temperature growth below 2° C is our aim, delinking between economic growth and energy consumption is a necessary part of the right roadmap regarding current fossil fuels dependence.

Colombia's economic fundamentals, including macroeconomic stability and openness to global trade and finance, have remained relatively strong in recent years. The economy has expanded by an average of around 5% annually over the past five years. Its population has risen to 47.7 million people with US $ 6056.1 per capita. In absolute terms, the Colombian economy is the 4th largest in the Latin American Area and the 5th in terms of per capita GDP (Purchasing Power Parity - PPP). Between 2000 and 2014, the ratio of total primary energy sources to population measured in toe/capita increased by 10.9% (2.6% for non-OECD American countries). For the same period, total energy production measured in Mtoe increased by 75.29%. Focusing only on electricity consumption measured in TWh, this increased by 83.9% and, expressed in per capita terms, electricity consumption also increased significantly, around 55.6% [9]. It should be considered that Colombia is among the countries with the greatest production of electricity [10].

The paper addresses the following questions. How far is the Colombian economy along the road to decoupling energy consumption from economic growth? Which factors have determined the changes in Colombia's energy consumption in the period examined? Was the success of past energy policies in Colombia decoupling oriented? Together with answering these questions, policy recommendations at the sectoral level are provided.

To answer all these questions three methodologies are combined in a novel way of providing information items that supplement each other. In a first step, the decoupling status between energy consumption and GDP growth is analysed with a decoupling elasticity index following Tapio [11] approach. In a second step, an additive decomposition analysis index (LMDI) is applied in order to identify the driving factors of the energy consumption changes in Colombia at the sectorial level. Four effects in the decomposition were considered: Population, Activity, Structure and Intensity for the period 2000–2015. The Agriculture, Mining, Industrial, Electricity, Gas and Water, Construction, Transport and Commercial and Public sectors were analysed. Finally, in a third step, considering the four effects from the LMDI analysis, a second level of decomposition was conducted to explore the efforts made to meet decoupling between energy consumption and economic growth.

The decomposition analysis is one of the most widely applied tools for analysing the mechanisms influencing energy consumption and its environmental side-effects. The literature offers two methodologic approaches for this analysis, developed exhaustively in environmental topics: Structural Decomposition Analysis (SDA), and Index Decomposition Analysis (IDA). Both methods decompose the variation experienced by a variable between their determining factors. These techniques have usually been applied in isolation to analyse the energy consumption and CO2 emissions changes. Some of these papers are: Achão and Schaeffer [12]; Zhang et al. [13]; Ang and Su [10] and Cansino et al. [14] for the SDA method and Hoekstra and van den Bergh [15]; Hatzigeorgiou et al. [16]; Ma and Stern [17]; Andreoni and Galmarini [18] and Colinet and Román [19] for the IDA method. The latest comparisons between IDA and SDA have been shown in Su and Ang [20]. Comparatively, IDA has certain advantages over SDA. IDA enables decompositions for any aggregate (value, ratio or elasticity). Also, IDA requires less data than decomposition methods based on input-output (IO) analysis, and it is useful when decomposing changes in environmental variables between their various components [21]. It could be said that one of the main advantages of IDA methods over their competitors based on IO matrices is the abundant availability of data.

Within the IDA methods, there are two important types: the Laspeyres method and the Divisia method. Ang et al. [22]; Ang and Zhang [24]; Ang [25]; Fernández and Fernández [82]; among others, have compared these two types of analysis. Specifically the Laspeyres IDA method is proportionally less used than the Divisia IDA, as considerable residuals arise in its decomposition, although the calculation of the results can be simple and understandable as shown in Zhang et al. [27].

The LMDI method seems to offer the most advantages within these various IDA methods [22,23,[27], [28], [29], [30], [31],[33], [34], [35], [36], [37]]. Ang et al. [38] conclude that the LMDI-I and LMDI-II methods satisfy most of the index number tests that are considered relevant for the IDA family, except that the LMDI-I fails the proportionality test, whereas the LMDI-II fails the aggregation test. Ang [28] assessed the various decomposition methods and concluded that LMDI-I is a more recommendable method due to both its theoretical base and its set of properties, which are satisfactory in the case of index decomposition. Additionally, LMDI-I provides a simple and direct association between the additive and the multiplicative decomposition form [39].

The literature offers evidence of a relationship between Colombian's energy consumption and economic growth using different approaches. Destek and Aslan [40] analysed this relationship for Colombia during the period 1980 to 2012 applying a bootstrap panel Granger causality test. A similar approach was used by Narayan and Doytch [41]; who also studied the link between economic growth and energy consumption in Colombia over the period 1971 to 2011. However, drivers of the decoupling or coupling processes between energy consumption and economic growth were not included in the above literature. The literature also offers results from LMDI decomposition analysis including Colombia in the sample of countries analysed. These are the cases of Timilsina and Shrestha [36] and Sheinbaum et al. [42]. Malpede [43] developed a Multi-Regional analysis with diverse IDA methods focused on CO2 emissions, but not on energy consumption. Ang [10] included Colombia in his global analysis of the changes in the aggregate carbon intensity for electricity, relating CO2 emissions to electricity production in each country. More recently, Román et al. [44] conducted a LMDI decomposition analysis for the specific case of Colombia for the period 1990 to 2012. Nevertheless, all of these papers focused on decomposing CO2 emissions changes but not on energy consumption changes.

We take advantage of the previous decomposition analysis and try to analyse not only the driving factors of energy consumption changes during one period but also the effort needed to achieve decoupling between energy consumption and economic growth. This latter information will show us the success of previous decoupling policies carried out in Colombia. Some papers have applied a decoupling index for analysing CO2 emissions [[45], [46], [47]]. Our proposal is to apply as a novelty a decoupling index for energy consumption that, as far as we are aware, is being carried out for the first time in Colombia.

The contribution of this paper to the specialised literature is based on the following points: i) The combination of three methodologies - the elasticity index, the LMDI analysis and the decoupling index - is applied for the first time, as far as we are aware, to analyse the energy consumption changes in Colombia over the period 2000–2015, ii) a more comprehensive updated dataset analysis dissagregated at the sectoral level that contributes to better understanding energy consumption changes in Colombia, iii) the analysing of past decoupling-oriented policy measures and the proposal of energy policy recommendations.

The results are interesting not only for researchers but also for policy-makers. In fact, this paper speaks directly to the authorities of Colombia and the policy agenda regarding several issues, mainly energy saving.

The paper is structured as follows. After the introduction, the methodological approaches are described in Section 2. Section 3 describes the database used. The results are presented in Section 4. In the light of the results, past policies are commented upon in Section 5. The main conclusions and policy recommendations appear in Section 6.

Section snippets

Decoupling analysis. The elasticity index

The decoupling index proposed by the Tapio model has been widely used to analyse the relationship between economic growth and energy consumption. Recent papers based on Tapio [48] and [11] include those by Diakoulaki and Mandaraka [45,47]; Aiwen and Dong [49]; Gray et al. [50]; Li et al. [51] and, more recently, Jiang et al. [46] and Wang et al. [52]. The later papers linked the Tapio [11] decoupling index with the LMDI approach described in the next Subsection.

Technically, the Tapio decoupling

Database

Three sources were mainly used to support this research. Firstly, the energy consumption is taken from the balances offered by the Mining and Energy Planning Unit (UPME in its Spanish acronym), which detail the necessary supply of final energy consumed by the main economic activities of the country. This dataset has been updated near to the end of 2016 offering detailed information until 2015 inluded in the Second revision of Colombian's energy balances 1975–2015 [65]. Secondly, the gross

Decoupling elasticity in Colombia

In order to explore the relationship between energy consumption and economic growth in Colombia between 2000 and 2015, equation (1) is used to calculate decoupling elasticity by using the economic growth measured by GDP at 2005 constant prices. The results of the decoupling elasticity index are shown in Table 2.

Only four subperiods clearly show a strong decoupling and two subperiods inform of weak decoupling. Considering that most of them are in the subperiod 2000–2007, it can be concluded that

Discussion

The activity and the population effects are the main drivers of Colombia's energy consumption. The results reported are consistent also across the scientific literature when this is focused on CO2-eq emissions from fossil fuel combustion [44]. Malpede [43] as well found that the population effect acted as a driver for the Colombian case. For other countries and regions one can see Hatzigeorgiou et al. [16] for the case of Greece, Donglan et al. [72] for China's residential sector, Colinet and

Conclusions and policy implications

Decomposition analysis, in addition to being a powerful explanatory tool, offers valuable assistance in assessing and analysing the progress in decoupling energy consumption from economic growth, supplemented by the decoupling index and elasticiy.

For the whole period analysed, 2000–2015, the results let us conclude that efforts made to achieve decoupling energy consumption from economic growth were not at all effective in Colombia. The activity (or affluence) and population effects acted as the

Acknowledgements

The first and second authors acknowledge the funding received from the SEJ 132 project of the Andalusian Regional Government, and the Cátedra de Economía de la Energía y del Medio Ambiente. The first and second authors also acknowledge the funding provided by the Universidad Autónoma de Chile (Chile) and from the project Nº 018/FONDECYT/16 of Chile's Department of Education. The standard disclaimer applies.

References (83)

  • J.M. Cansino et al.

    Main drivers of changes in CO2 emissions in the Spanish economy. A structural decomposition analysis

    Energy Pol

    (2016)
  • R. Hoekstra et al.

    Comparing structural and index decomposition analysis

    Energy Econ

    (2003)
  • E. Hatzigeorgiou et al.

    CO2 emissions in Greece for 1990–2002: a decomposition analysis and comparison of results using the arithmetic mean Divisia index and logarithmic mean Divisia index techniques

    Energy

    (2008)
  • C. Ma et al.

    China's changing energy intensity trend: a decomposition analysis

    Energy Econ

    (2008)
  • V. Andreoni et al.

    Drivers in CO2 emissions variation: a decomposition analysis for 33 world countries

    Energy

    (2016)
  • B. Su et al.

    Structural decomposition analysis applied to energy and emissions: some methodological developments

    Energy Econ

    (2012)
  • J.M. Cansino et al.

    Driving forces of Spain׳s CO2 emissions: a LMDI decomposition approach

    Renew Sustain Energy Rev

    (2015)
  • B.W. Ang et al.

    Factorizing changes in energy and environmental indicators through decomposition

    Energy

    (1998)
  • L. Chen et al.

    A spatio-temporal decomposition analysis of energy-related CO2 emission growth in China

    J Clean Prod

    (2015)
  • B.W. Ang

    The LMDI approach to decomposition analysis: a practical guide

    Energy Pol

    (2005)
  • V. Alcántara et al.

    Structural analysis of electricity consumption by productive sectors. The Spanish case

    Energy

    (2010)
  • M. Zhang et al.

    Decomposition of energy-related CO2 emission over 1991-2006 in China

    Ecol Econ

    (2009)
  • B.W. Ang

    Decomposition analysis for policymaking in energy: which is the preferred method?

    Energy Pol

    (2004)
  • B.W. Ang

    LMDI decomposition approach: a guide for implementation

    Energy Pol

    (2015)
  • B.W. Ang et al.

    A new energy decomposition method: perfect in decomposition and consistent in aggregation

    Energy

    (2001)
  • V. Moutinho et al.

    The driving forces of change in energy-related CO2 emissions in Eastern, Western, Northern and Southern Europe: the LMDI approach to decomposition analysis

    Renew Sustain Energy Rev

    (2015)
  • M. Shahiduzzaman et al.

    Decomposition analysis to examine Australia's 2030 GHGs emissions target: how hard will it be to achieve?

    Econ Anal Pol

    (2015)
  • W. Zhang et al.

    Decomposition of intensity of energy-related CO2 emission in Chinese provinces using the LMDI method

    Energy Pol

    (2016)
  • B.W. Ang et al.

    Properties and linkages of some index decomposition analysis methods

    Energy Pol

    (2009)
  • B.W. Ang et al.

    Energy decomposition analysis: IEA model versus other methods

    Energy Pol

    (2007)
  • M.A. Destek et al.

    Renewable and non-renewable energy consumption and economic growth in emerging economies: evidence from bootstrap panel causality

    Renew Energy

    (2017)
  • S. Narayan et al.

    An investigation of renewable and non-renewable energy consumption and economic growth nexus using industrial and residential energy consumption

    Energy Econ

    (2017)
  • C. Sheinbaum et al.

    Energy consumption and related CO2 emissions in five Latin American countries: changes from 1990 to 2006 and perspectives

    Energy

    (2011)
  • R. Román et al.

    Analysis of the main drivers of CO2 emissions changes in Colombia (1990–2012) and its political implications

    Renew Energy

    (2018)
  • D. Diakoulaki et al.

    Decomposition analysis for assessing the progress in decoupling industrial growth from CO2 emissions in the EU manufacturing sector

    Energy Econ

    (2007)
  • J. Vehmas et al.

    Linking analyses and environmental Kuznets curves for material flows in the European Union 1980–2000

    J Clean Prod

    (2007)
  • P. Tapio

    Climate and traffic: prospects for Finland

    Global Environ Change

    (2002)
  • J. Brizga et al.

    Drivers of CO2 emissions in the former Soviet Union: a country level IPAT analysis from 1990 to 2010

    Energy

    (2013)
  • B.W. Ang et al.

    Decomposition of energy-induced CO2 emissions in manufacturing

    Energy Econ

    (1997)
  • J. Sun

    Changes in energy consumption and energy intensity: a complete decomposition model

    Energy Econ

    (1998)
  • E. Akbostancı et al.

    CO2 emissions of Turkish manufacturing industry: a decomposition analysis

    Appl Energy

    (2011)
  • Cited by (89)

    View all citing articles on Scopus
    View full text