economic growth, energy demand and the environment ... · transportation have lead to an unexpected...

Economic Growth, Energy Demand and the Environment –

Empirical Insights Using Time Series and Decomposition Analysis

Dirk C. Böhm

http://www.weissensee-verlag.de/autoren/boehm.htm

Bibliografische Information Der Deutschen BibliothekDie Deutsche Bibliothek verzeichnet diese Publikation in der Deutschen Nationalbibliografie; detaillierte bibliografische Daten sind im Internet über www.dnb.de abrufbar.

Als Inaugural-Dissertation zur Erlangung des Grades eines Doktors der Wirtschaftswissenschaften (Dr. oec.) an der Universität Hohenheim 2010, Fakultät Wirtschafts- und Sozialwissenschaften, vorgelegt von Dirk C. Böhm. D 100

Erstgutachter: Prof. Dr. Ansgar BelkeZweitgutachter: Prof. Dr. Michael AhlheimPrüfungsvorsitzender: Prof. Dr. Werner F. SchulzDatum der mündlichen Prüfung: 07. 02. 2011

Gedruckt auf holz- und säurefreiem Papier, 100 % chlorfrei gebleicht.

©Weißensee Verlag, Berlin 2011www.weissensee-verlag.de

Alle Rechte vorbehalten

Umschlagbild: Peter Elvidge (pixmac.com)

Printed in Germany

ISBN 978-3-89998-193-3

�

Contents

Abbreviations ................................................................................................................��

List of Figures ..............................................................................................................�X��

List of Tables ................................................................................................................��X��

1.� Preface ....................................................................................................................... 1�2.� Energy and the Economy .......................................................................................... 7�

2.1� The Current Energy System ............................................................................... 7�2.2� Resources and Reserves of Fossil Fuels: Limits to Growth? ........................... 11�2.3� Growth and the Environment: The Environmental Kuznets Curve ................ 18�

3.� The Relationship between Energy Prices, Energy Consumption and Growth ........ 23�3.1� Granger Causality ............................................................................................ 23�3.2� Cointegration and Vector Error Correction Models ........................................ 27�

3.2.1� Individual Unit Root Test ......................................................................... 28�3.2.2� Testing for Cointegration ......................................................................... 28�3.2.3� Panel Unit Root and Cointegration Tests ................................................. 29�3.2.4� Causality Tests .......................................................................................... 31�

3.3� Empirical Evidence of Causal Relationships ................................................... 37�3.3.1� Causality between Energy Consumption and Economic Growth ............ 37�3.3.2� The Role of Energy Prices ........................................................................ 44�

3.4� Mutual Causality and Heterogeneity ............................................................... 50�4.� The Impact of Energy Use on the Environment ...................................................... 53�

4.1� Introduction: The Kaya Identity and the Concept of Decomposition .............. 53�4.2� The Decomposition of CO2 Emissions............................................................. 57�4.3� Which are the Factors behind Emission Changes? .......................................... 62�4.4� How to Achieve Emission Reductions: An Aggregated View ........................ 72�

5.� Industrial Energy Demand and Emissions .............................................................. 75�5.1� Introduction: The Role of the Industry Sector ................................................. 75�5.2� Index Decomposition Analysis for the Industrial Sector ................................. 77�5.3� Energy and CO2 Efficiency in the European Manufacturing Sector ............... 81�

5.3.1� Decomposition of Changes in Energy Consumption ............................... 81�5.3.2� Decomposition of Changes in CO2 Emissions ......................................... 86�

5.4� Energy Productivity Gains in the Manufacturing Sector ................................. 92�6.� The Transport Sector ............................................................................................... 95�

6.1� The Need for Mobility and the Dependence on Oil ......................................... 95�6.2� A Short Decomposition Analysis of CO2 Emissions in the Transport Sector 100�6.3� Drivers of Road Transport Energy Demand and CO2 Emissions .................. 105�6.4� The Future of Transport Energy Use ............................................................. 109�

7.� Conclusions ........................................................................................................... 113�7.1� Summary ........................................................................................................ 113�7.2� Outlook .......................................................................................................... 116�

References .................................................................................................................... 119�

Abbreviations

�II�

Abbreviations

ADF test Augmented Dickey Fuller test

AIC Akaike Information Criterion

APEC Asia-Pacific Economic Cooperation

ARDL Autoregressive Distributed Lag

AUS Australia

AUT Austria

BEL Belgium

BGD Bangladesh

BGR Bulgaria

BGR Bundesanstalt für Geowissenschaften und Rohstoffe

(Federal Institute for Geosciences and Natural Resources)

bn Billion

BRA Brazil

BTU British Thermal Units

C CO2 Emission Coefficient

CAN Canada

CE Cointegrating Equation

Ceff Emission Coefficient Effect

CHE Chemical and Petrochemical Industry

CHN People's Republic of China

CO Carbon Oxide

CO2 Carbon Dioxide

CYP Republic of Cyprus

CZE Czech Republic

DEU Germany

DNK Denmark

e Natural logarithm of per capita total final energy demand

E Total Energy Consumption in the Industry sector

Eact Activity Effect (Energy Consumption)

ECM Error Correction Model

ECMT European Conference of Ministers of Transport

ECT Error Correction Term

Eint Intensity Effect (Energy Consumption)

Abbreviations

V��

EKC Environmental Kuznets Curve

EM CO2 Emissions from the Consumption of Fossil Fuels

EMact Activity Effect (CO2 Emissions)

EMemf Emission Factor Effect

EMint Intensity Effect (CO2 Emissions)

EMmix Fuel Mix Effect

EMstr Structure Effect (CO2 Emissions)

EOR Enhanced Oil Recovery

ESP Spain

Estr Structure Effect (Energy Consumption)

ET Total CO2 Emissions from the Transport Sector

EU European Union

EUR19 Aggregate of 19 European Union and OECD members

(Germany, France, Great Britain, Italy, Spain, the Netherlands,

Poland, Belgium, Sweden, Austria, the Czech Republic, Finland,

Greece, Portugal, Hungary, Denmark, Ireland, Slovakia,

Luxemburg)

EUR23 Aggregate of 23 European Union members

(Germany, France, Great Britain, Italy, Spain, the Netherlands,

Poland, Belgium, Sweden, Austria, the Czech Republic, Romania,

Finland, Greece, Portugal, Hungary, Denmark, Ireland, Slovakia,

Bulgaria, Luxemburg, Cyprus, Malta)

FIN Finland

FM Fuel Mix Variable

FMeff Fuel Mix Effect

FOD Food and Tobacco Industry

FPES Primary Energy Supply of Fossil Fuels

FRA France

FSU Former Soviet Union

FSUREG Region of the Former Soviet Union (= FSU)

G-11 Group of Eleven (Jordan, Croatia, Ecuador, Georgia, El Salvador,

Honduras, Indonesia, Morocco, Pakistan, Paraguay, Sri Lanka)

G-7 Group of Seven

(The United States, Japan, Great Britain, Germany, Canada,

France and Italy)

GBR Great Britain

GCC Gulf Cooperation Council

Abbreviations

��

GDP Gross Domestic Product

GHG Greenhouse Gas

GRC Greece

HUN Hungary

I Energy Intensity

IBRD International Bank for Reconstruction and Development

IDA Index Decomposition Analysis

IDN Indonesia

IEA International Energy Agency

Ieff Energy Intensity Effect

IND India

IPCC Intergovernmental Panel on Climate Change

IPS test Im, Pesaran and Shin panel unit root test

IRL Ireland

IRN Islamic Republic of Iran

IRS Iron and Steel Industry

ITA Italy

JAMA Japan Automobile Manufacturers Association

JPN Japan

KOR Republic of Korea (South Korea)

LLC test Levin, Lin and Chu panel unit root test

LMDI Log Mean Divisia Index

LNG Liquefied Natural Gas

LPG Liquefied Petroleum Gas

LR Long Run

LUX Luxemburg

MAC Machinery Industry

MAIC Modified Akaike Information Criterion

MET Non-Ferrous Metals Industry

MEX Mexico

M Fuel Mix Variable

MLT Malta

MS Modal Split

Abbreviations

�

MSeff Modal Split Effect

Mt Million Tons

MWI Malawi

MYS Malaysia

NLD The Netherlands

NMM Non-Metallic Minerals Industry

NOx Nitrogen Oxide

NZL New Zealand

OECD Organisation for Economic Co-operation and Development

OPEC Organization of Petroleum Exporting Countries

p Natural logarithm of the index of real energy prices for industry

and households

PAK Pakistan

PAP Paper, Pulp and Printing Industry

Peff Population Effect

PHL Philippines

pkm Passenger kilometres

POL Poland

POP Population

PP test Phillips-Perron unit root test

PPP Purchasing Power Parity

PRT Portugal

Q Activity Level

R/P ratio Reserves-to-Production ratio

RAR Reasonably Assured Resources

ROU Romania

S Share of Fossil Fuels; Activity Share

SAU Saudi Arabia

Seff Substitution Effect

SGP Singapore

SIC Schwarz Information Criterion

SO2 Sulfur Dioxide

SPM Suspended Particulate Matter

��

SR Short Run

STAN Structural Analysis Database (OECD)

SVK Slovak Republic

SWE Sweden

t Tons

TEX Textile and Leather Industry

TFC Total Final Consumption

THA Thailand

tkm Thousand Kilometres

TPES Total Primary Energy Supply

TRA Transport Equipment Industry

TUR Turkey

TWN Taiwan

U.S. United States

UK United Kingdom

UNFC United Nations Framework Classification for Fossil Energy

and Mineral Resources

US United States

USA United States of America

USD U.S. Dollars

VA Value Added

VAR Vector Autoregression

VECM Vector Error Correction Model

vkm Vehicle Kilometres

VMT Vehicle Miles Travelled

WBCSD World Business Council for Sustainable Development

WEO World Energy Outlook (IEA Publication)

WOD Wood and Wood Products Industry

y Natural logarithm of real per capita GDP

Yeff Income Effect

ZAF South Africa

Preface

1

1. Preface

Industrialization and increasing wealth in emerging markets – especially in China

and India – as well as intensifying globalization and the associated boost in

transportation have lead to an unexpected rise of global energy demand in the last

decade. According to the Reference Scenario in the 2007 issue of the IEA World

Energy Outlook, the global primary energy demand is projected to grow by 55%

between 2005 and 2030. The developing countries with fast growing populations

contribute 74% of this increase, China and India alone 45%.1 The bulk of global

energy supply is coming from fossil fuels. Although renewable energy sources

are promoted heavily in industrialized countries, they will neither be able to

replace fossil fuels, nor fill the gap between growing demand and the current

level of supply of fossil resources in the near future. This has important

implications for the world’s energy security. We will see a growing dependence

of consuming countries on oil and gas imports as well as a reduction of

geographic supply diversity with an increasing market dominance of the Middle

East and Russia. As a geopolitical consequence, energy will greatly determine

foreign relations in the future as the uninterrupted flow of energy will mainly

depend on the political and economic stability of the producer regions.

Furthermore there will be more competition among consumer countries for

energy supplies following this growing energy import dependency. Another

consequence of the soaring demand in combination with the existing global

energy mix is accelerating climate change due to increasing greenhouse gas

emissions. According to the Intergovernmental Panel on Climate Change Fourth

Assessment report2, most of the observed rise in global average temperatures

since the mid-20th century is very likely3 related to the increase in anthropogenic

greenhouse gas (GHG) concentrations. The combustion of fossil fuels is the

largest contributor to GHG emissions and carbon dioxide (CO2) is responsible for

about 95% of the energy-related emissions. Thus fossil fuel combustion is the

single largest human influence on climate. While emissions have doubled in the

period between 1971 and 2005, real gross domestic product (GDP) has reached

three times the value of the base year (see Figure 1). But although declining

global CO2 emissions per unit of GDP could be observed in many industrialized

countries, strong economic growth in emerging markets has led to a worrying 1 See IEA (2007b), p.73. 2 IPCC (2007) 3 In the IPCC terminology “very likely” means a likelihood of occurrence of > 90%.

Preface

2

rise in global CO2 emissions in the last years. Similar to the supply security

problem GHG emissions can be tackled by either increasing the share of

renewable energy sources or by the implementation of strict energy conservation

measures as well as efficiency technologies. During the negotiation process for a

new post Kyoto climate regime energy saving measures play a vital role, as

energy saving is far easier in the short term than restructuring the primary energy

mix. However a lot of countries, mostly developing countries, fear that such

policy measures will harm their economic development.

Figure 1: World CO2 emissions from fuel combustion and real GDP4

Since the oil crisis of the 1970’s the relationship between energy consumption,

environmental pollution and economic growth has been analyzed and discussed

in many different ways. Especially between 2004 and 2008 energy became an

issue once again as global economic growth led to an enormous increase in oil

prices.

The IEA World Energy Outlook 2006 stated: “The world is facing twin energy-

related threats: that of not having adequate and secure supplies of energy at

affordable prices and that of environmental harm caused by consuming too much

of it. …” and “…the need to curb the growth in fossil-energy demand, to increase

geographic and fuel-supply diversity and to mitigate climate-destabilising

emissions is more urgent than ever.”5 In the very same year Sir Nicholas Stern

published his widely acknowledged and discussed “Stern Review on the

Economics of Climate Change”6, presenting scientific evidence for the economic

4World GDP in 2000-USD using Purchasing Power Parities (PPP). 5 See IEA (2006), p.37. 6 Stern (2006).

�

��

��

��

��

��

��

��

��

Data Source: IEA; author’s own illustration.

Preface

3

impacts of climate change and necessary policy responses. The subsequent IEA

World Energy Outlook 2007 focused on the role of China and India as the

“emerging giants of the world economy and international energy markets.”7 It

became clear that the growth rates of these countries and the impact on global

energy demand would pose a big challenge to the present energy system and

existing efforts to combat climate change. Thus a transition to a more secure,

lower-carbon energy system seemed inevitable, but only without the risk of

undermining economic and social development. Furthermore, soaring energy

prices fueled the fear of a recurrence of the oil crisis of the 1970s and reports

about continuous resource depletion led to a revival of the peak oil theory

initially developed by Marion King Hubbert in the 1950s.8

These developments and discussions were the starting point for this text. The

idea was to dig deeper into the mechanisms of energy use and economic

development and to assess which are the main factors behind energy savings and

a reduction of energy related carbon dioxide emissions. As discussed in detail in

the following chapters, the existing empirical literature only gives a limited view

on these issues as to the methods used, to the geographical and sectoral coverage,

as well as the time periods observed. This thesis thus aims at empirically giving

insights about the relationship between energy consumption, economic growth

and CO2 emissions on a global scale. The analysis is carried out for the top

energy consumers and CO2 emitters worldwide with a special emphasis on the

European Union and some focus countries for the detailed investigation of the

industry and transport sector. In addition, recent panel cointegration and

decomposition methods are used to analyze detailed data of the last 25-35 years.

The statements or assertions below are to be tested within this context:

- There is a mutual interrelationship between economic growth and energy

demand.

- Developing and emerging economies are more energy-dependent than the

economies of highly developed countries.

7 See IEA (2007b), p.41. 8 See Hubbert (1956).

Preface

4

- Rising energy productivity in combination with changes in the primary

energy mix with less carbon content is the best strategy to mitigate energy

related emissions.

- The reduction of CO2 emissions within the industrial sector in Western

European countries was only possible due to a relocation of energy

intensive subsectors to developing and newly industrialized countries.

The text is structured as follows. Chapter 2 gives a description of the relationship

between energy, economic growth and the environment. The current energy

system in terms of availability and use of energy sources is illustrated. The

question is discussed, whether pollution is rising inevitably with economic

growth, or if emissions fall again at a certain level of per capita income

(Environmental Kuznets Curve).

In Chapter 3 the causality between energy prices, energy consumption and

economic growth is analyzed empirically. This empirical investigation applies

cointegration and error correction techniques and consists of two parts. In the

first part the bivariate relationship between energy and GDP is examined for the

15 biggest global energy consumers between 1978 and 2005. In the second part,

energy prices are added as a third variable.

Chapter 4 uses decomposition analysis of the change in carbon dioxide emissions

in order to analyze the relationship between emission growth and changes in

underlying factors using the Log Mean Divisia Index (LMDI) method. It covers

the biggest carbon dioxide emitting countries and regions that together account

for over 80% of total emissions worldwide in the period from 1971 to 2005.

The industry sector is one of the largest consumers of energy and also one of the

largest energy-related CO2 emitters. Chapter 5 gives insights into the

mechanisms of change in industrial energy consumption as well as changes in

CO2 emissions for ten manufacturing industries in five European countries by

using the same decomposition technique.

The transport sector is one of the largest and fastest growing sources of

greenhouse gas emissions. As a consequence Chapter 6 provides an overview of

energy consumption and emission drivers in the transport sector, with empirical

examples for the United States, Japan and Germany. The second part of the

Preface

5

chapter focuses explicitly on road transport, as this sub-sector has the lion’s share

of transport energy use and emissions.

Finally Chapter 7 summarizes the empirical results and presents policy

implications as well as further need for research beyond the scope of this text.

This thesis uses recent panel cointegration and error correction models to test for

causality between energy consumption and economic growth. It also assesses the

role of energy prices in this context using real indices of energy prices for

industry and households, while most other studies on this issue use consumer

price indices as a proxy. In chapters four to six a perfect index decomposition

method LMDI1 is applied to analyze the underlying factors of energy

consumption patterns and CO2 emissions. Time series decomposition is used

because the decomposed results given by this approach can better explain the

underlying mechanisms of change in energy use. Most other decomposition

studies of energy consumption and CO2 emission use period wise decomposition,

which is based on the data of two benchmark years, and the data for the

intervening years are discarded. These period wise decomposition results are less

informative and hence may not result in superior representation to the real

situation. Apart from rising incomes and population, these models also account

for fuel mix in primary and final energy consumption as well as structural

effects. The analysis is not only conducted on an aggregate level, but also for the

industry and transport sector.

Energy and the Economy

2. Energy and the Economy

2.1 The Current Energy System

Energy is involved in everything that happens on earth. Vast quantities of energy

sources (fossil, nuclear and thermal energy) are still accessible by either

conventional or enhanced technologies; but the consequences of location,

development and utilization of these energy sources is shaping the global

political and economic landscape and directly affect living standards and

prosperity. Therefore any discussion of employment, technology,

competitiveness and economic growth must consider the strategic role of energy.

In a historical perspective, the human utilization of energy commenced about

half a million years ago with the burning of wood for lighting, heating and

cooking. By the eighteenth and nineteenth century, coal was used instead of

firewood and formed the basis for industrialization through the production of iron

and the invention of the steam engine. These coal-fired steam engines later gave

way to oil and gas supplied systems. Since the 1950s nuclear energy is used in

parts of the world to generate electricity. Figure 2 presents the development of

U.S. energy consumption by source over the last centuries.

Figure 2: History of energy consumption in the United States

Within the current energy system, oil is still the largest primary energy source

consumed globally, followed by coal, natural gas and hydroelectricity. Coal is

the fastest-growing fuel, whereas oil has lost market share in the last years. On a

regional level, oil is the dominant fuel in the Americas, the Middle East and

Africa. Gas is the dominant fuel in Europe and Eurasia and as coal meets 70% of

PetroleumHydroelectricCoalWoodNatural GasNuclear

quadrillion Btu

0

40

35

30

25

20

15

10

5

45

1775 19751950192519001875185018251800 2000Source: Energy Information Administration (EIA).


8

China’s energy needs it is the most important primary energy source in Asia-

Pacific. In 2006 12.7% of Total Primary Energy Supply (TPES) was produced

from Renewables (see Figure 3).

Figure 3: Total primary energy supply by fuel – World 2006

While in high-income countries the majority of renewable energy sources like

wind and solar energy are used to generate electricity, on a global basis a big part

of renewables is used in the residential sector in the form of biomass. On a

global basis solid biomass is by far the largest renewable energy source due to its

extensive non-commercial use in developing countries. Together with other

renewable combustibles and waste it represents 9.9% of world TPES and 78.1%

of renewable supply. With 17.5% and 3.1% of global renewable supply the

second and third largest renewable sources are hydro power and geothermal

energy. More interesting, solar, wind and tidal energy only contribute 1.2% of

renewable supply, which is approximately 0.5% of TPES (see Figure 4).

Figure 4: Renewable energy supply by product – World 2006

��

��

��

�� !��

"!�!#�$�! ��

��%"!�!#��&� �!��

"!�!#�$�!��$� ��$�! ��'�&� �!

�(��

)��*�+�'!��

,-'��

&��'��

�!��.!��

Source: IEA online database; author’s own illustration.



9

Following coal and natural gas, renewables are the third largest contributor to

worldwide electricity production. According to the 2008 edition of the IEA

Renewables Information they accounted for 18.1% of global generation in 2006

(see Figure 5)9.

Figure 5: Global electricity production 2006 – fuel shares

Driven by their use in transport, oil products are the most important final energy

commodity with a global share of 43% in 2006. Electricity accounts for about

17% of total final consumption (see Figure 6) and is still growing.

Figure 6: Total final energy consumption by fuel – World 2006

The main reasons for this rising electricity consumption are the increased use of

electrical appliances (air conditioning, lighting, IT equipment, etc.) and the

advent of new electrical devices in the service sector. In the residential sector,

rising incomes, higher living standards and the trend towards smaller households 9 See IEA (2008c), p.5.

��

��(�

��

�� !��(�

��%"!�!#��&� �! ��

"!�!#�$�! �(��

��'� ��/��'� �

��

��

��0� ��

��$� ��$�!��!�!#�$�! ��'�#� �!

��

�! �� -��

,!��




10

lead to more and larger dwellings and a growing demand for electrical appliances

such as refrigerators, freezers, washing machines, dishwashers, TVs and dryers

as well as a growing number of smaller appliances such as videos and computers.

Around 17% of total final energy consumption is attributed to natural gas. In the

industry sector it is used to produce steel, glass, paper, clothing, brick, and as an

essential raw material for many other products such as paints, fertilizers, plastics,

medicines and explosives. Households use natural gas as a major heating fuel and

to fuel stoves, water heaters, clothes dryers, and other household appliances.

Three sectors are responsible for approximately 80% of global energy

consumption, namely the transport, the industry and the residential sector (see

Figure 7). The transport sector is almost solely relying on oil and is growing fast;

particularly in developing countries (see also Chapter 6 for a detailed description

of the transport sector). The residential sector consumes electricity for all kinds

of appliances as well as gas and oil for heating purposes, whereas industrial

processes use all secondary energy sources (see also Chapter 5 for a detailed

analysis of the industrial sector).

Figure 7: Total final energy consumption by sector– World 2006

1�'� ��-� ! ��

+�� /�� ! ��(�

"! �'!��

��!� !��'�/�$�� !�2� !

(�

30�� !�45��! ��-�6�7� .��0��

"! ��



11

2.2 Resources and Reserves of Fossil Fuels: Limits to Growth?

Fossil fuels can be considered as mineral resources that mostly result from burial

and transformation of biomass during the last 200 million years. A resource in

this sense can be defined as “a concentration of naturally occurring solid, liquid

or gaseous material in or on the earth’s crust in such form and amount that

economic extraction of a commodity from the concentration is currently or

potentially feasible”.10 Reserves however refer to material contained in

established and exploitable deposits. At a given time, reserves of a given

commodity include only that part of its resources which is available for

exploitation. Resources of fossil fuels can be categorized by two parameters: (1)

degree of certainty under the present state of knowledge, and (2) feasibility of

economic recovery under the present conditions and available technologies.11

This implies that the amount of reserves depends on current prices of the energy

carrier as well as on technological progress.

Figure 8: Reserves and resources for fossil fuels (excl. Uranium)

Reserves and resources of oil, gas and coal are dynamic quantities as new

technologies are developed and new discoveries are made. The estimation of

potential resources of a commodity is difficult. Due to the uncertainty of price,

market and technological development it is hard to assess what can be considered

as potentially available and recoverable in the future. According to the United

Nations Framework Classification for Fossil Energy and Mineral Resources

10 See Brown (2002), p. 25. 11 See Brown (2002), p. 25.

Source: BGR (2007), p.35.


12

(UNFC)12, initial resources of an energy commodity can be described in terms of

a) produced quantities (or cumulative production in Figure 8), b) remaining

recoverable quantities and c) additional quantities remaining in-place. Produced

quantities are the sum of sales and non-sales quantities between the time of the

first recorded production and the time of the evaluation. The remaining

recoverable quantities are the sum of (sales and non-sales) quantities estimated to

be produced from the evaluation time onward. Additional quantities are

estimated to be in-place at the time of evaluation but not yet produced or

recoverable. The UNFC categorizes the remaining resources using the following

recoverability criteria:

• Economic and commercial viability

• Field project status and feasibility

• Geological knowledge

These criteria are also implicitly used in the other existing resource

classifications. The fact that the future availability of reserves depends on the

level of energy prices and technological progress has led to a rising figure for

crude oil and natural gas reserves in the last three decades (see Figure 9 and

Figure 10).

Figure 9: Proved oil reserves (1980-2007)

12 See UN (2010), p.16.

�

��

��

��

(��

��

��

��

��(� ��(� ��(� ��(� �� ( ��

+.�� '��$��!� Source: BP (2008); author’s own illustration.


13

Figure 10: Proved gas reserves (1980-2007)

An indicator often used and as often misinterpreted in this context is the

Reserves-to-Production (R/P) ratio. The R/P ratio is expressed in years and

interpreted as the remaining amount of a depleting natural resource, mostly

applied to fossil fuels. These ratios may be calculated for individual countries or

globally for specific resources, the latter being used as an indicator of the time

remaining before the resource is completely exhausted.

Figure 11: Global Reserves-to-Production ratio for oil (1980-2007)

As Figure 11 and Figure 12 exhibit, the R/P ratios for crude oil and natural gas

remained almost stable in the last 25 years despite rising energy demand and

resulting growth in production, which is also due to the increase in proved

reserves (see Figure 9 and Figure 10).

�

��

��

��

(�

��

��

��

��

�(�

��

��(� ��(� ��(� ��(� �� ( ��

+�� $� ��!�!�

�

�

��

��

��

��

��

��

��

��

��

��(� ��(� ��(� ��(� �� ( ��

-!��

Source: BP (2008); author’s own illustration.

Source: BP (2008); author’s own illustration.


14

Figure 12: Global Reserves-to-Production ratio for gas (1980-2007)

Today, the global R/P ratio for crude oil is slightly above 40 years, proved

natural gas reserves would be depleted in about 60 years at the current rate of

extraction. According to data of the World Energy Council13 the R/P ratio for

coal was 133 years in 2007, so that coal is the fossil energy carrier with the

longest range.

As fossil fuels are exhaustible resources there have always been discussions

about when the world runs out of these resources and what consequences the

ongoing depletion has on energy prices and energy supply security. The so called

“Peak Oil” theory is the most prominent base for controversial discussions of

experts and politicians. This theory was originally developed in 1956 by the

geophysicist Marion King Hubbert, who used extraction and depletion rates of

oil in order to determine the peak of global oil production.14 The advocates of

this theory argue that despite growing statistical reserves the gap between annual

global consumption and new global discoveries has widened within the same

time period. Others argue that there is no physical problem, rather an investment

problem on the transition path from oil – mainly needed for transport – to

alternative fuels.15 As mentioned before, the future availability of fossil energy

resources is a function of their prices and technological progress. This is

expressed graphically by Figure 13 which shows the various oil prices at which

the exploitation of different resources becomes an economical option, taken into

13 See WEC (2007). 14 See Hubbert (1956). 15 See Mabro (2006).

�

��

��

��

��

��

��

��

(�

��(� ��(� ��(� ��(� �� ( ��

-!�� Source: BP (2008); author’s own illustration.


15

account the cost of capture and storage of CO2 produced during the extraction of

non-conventional oils.

For oil the extraction represents the dominant cost, whereas for natural gas the

economics are dominated by the cost of transportation. Future gas supply is

therefore determined by the development of liquefied natural gas and other

transportation technologies. 16

Figure 13: Oil cost curve

A more profound concern is the global allocation of fossil energy resources.

More than 60% of global crude oil reserves are located in the Middle East, 12%

in Europe and Eurasia – mostly within the territory of the Former Soviet Union –

and 9% in Central and South America (here the bulk of oil reserves lies in

Venezuela) as well as Africa. Only small reserves can be found in the top

consuming regions North America, Western Europe, Japan and China (see Figure

14).

16 See IEA (2005), pp.111 ff..


16

Figure 14: Oil reserves 2007 by region

Natural gas reserves are also allocated unevenly. As shown in Figure 15, most of

the reserves lie in the Middle East as well as Europe and Eurasia. Once again it is

the Russian Federation that holds 75% of the gas reserves within this region.

Figure 15: Gas reserves 2007 by region

Europe’s and North America’s dependency on imported energies will increase

even further in the future, as domestic reserves will be depleted first. This has a

major geopolitical impact, as the political and economic power of the oil and gas

exporting countries will continue to rise. An uninterrupted flow of energy

depends on the political and economic stability of the producer regions. At least

for the Middle East this stability is questionable and Russia is already using its

position strategically in foreign relations. Many oil producing countries have

��.�3�!��

)��6��!��3�!��

��

��/!�6��

8�''�!��

35��

3 �� 5� ��

��.�3�!��

)��6��!��3�!��

��

��/!�6��

8�''�!��

35�� (�

3 �� 5� (�




17

experienced the so called “Dutch disease”, i.e. negative effects of producing a

singular product for the world market and an uneven internal distribution of oil

revenues can lead to political instability17. However, growing energy import

dependency also leads to sharper competition among consumer countries. As

global demand for energy continues to rise, the major players like the United

States, European Union, Japan and China are facing a race to secure long-term

energy supplies. Due to its booming economy, China is intent on getting the

resources needed to sustain its rapid growth and has already turned to Africa as a

major source for oil often overlooked by the Europeans and Americans.

Consequently Chinese energy companies are committing large amounts of

funding and labour for exploration and development rights in resource-rich

countries.

Figure 16: Coal reserves 2007 by region

Figure 16 shows the regional distribution of coal reserves. Due to its large

reserves and availability within the major energy consuming regions, coal will

remain the dominant fuel for global electricity generation. Amongst the major

energy sources, coal is the most rapidly growing fuel as it is abundant and

broadly distributed around the world. Economically recoverable reserves of coal

are available in each major world region.18 The high specific carbon dioxide

content however makes it problematic on ecological grounds.

17 See CIEP (2004), p.19. 18 See WEC (2007), p. 1.

��.�3�!��

)��6��!��3�!��

��

��/!�6��

8�''�!�� 6�35��

3 �� 5� ��



18

The role of nuclear power for electricity generation depends on the availability of

uranium. The biggest “Reasonably Assured Resources” (RAR) can be found in

Australia, Canada, Kazakhstan, Niger and Brazil. These countries account for

over 80% of global uranium resources19. As global electricity consumption is

expected to continue growing over the next several decades to meet the needs of

an increasing population and further economic growth, nuclear energy will

continue to play an important role in generating the required electricity, although

the magnitude of that role remains uncertain. This is mostly depended on the

political debate about advantages and risks of nuclear energy. Some argue that

the use of nuclear power in electricity generation reduces carbon emissions and

increases energy security by decreasing dependence on oil and gas imports.

Opponents point to the problems associated with processing, transport and

storage of radioactive nuclear waste, or the risk of nuclear accidents.

In summary it can be said that the world will not run out of fossil fuels within the

next decades taking technological progress into account. Fossil fuels thus will

continue to dominate energy supply in the future. Nevertheless the geographic

distribution of resources spurs concerns about supply security, and energy prices

will probably not fall back to the level of the 1990s again. Renewable energy

sources – which until now played a minor role in the energy mix – can help to

level out supply disruptions and to reduce the dependency on energy imports

from the Middle East and Russia.

2.3 Growth and the Environment: The Environmental Kuznets Curve

The environmental Kuznets curve (EKC) hypothesis asserts that pollution

follows an inverted-U path with respect to economic growth, i.e. pollution

increases (environmental quality declines) with economic growth up to a certain

income level, after which it declines (improvement of environmental quality).

This hypothesis is adapted from Kunznets’ (1955) original study on the

influences of economic development on income inequality.20 Equally the 1992

World Development Report (IBRD, 1992) argued: “The view that greater

economic activity inevitably hurts the environment is based on static assumptions

19 See OECD (2006b), p.13ff.. 20 See Cavigilia-Harris et al (2009).


19

about technology, tastes and environmental investments” and also that “as

incomes rise, the demand for improvements in environmental quality will

increase, as will the resources available for investment”.21

Pure growth in the scale of an economy would result in a proportional growth in

pollution and other environmental impacts if there were no changes in the

structure or technology of this economy. The view that economic development

and environmental quality are conflicting goals reflects this scale effect alone.22

An inverted-U shape for environmental quality can thus be explained by several

structural factors. One of the arguments is, that production and consumption

change with income. If the industry structure changes from relatively clean

agricultural economies to polluting industrial economies and then to a bigger

share of information-based industries and services, an inverted-U path for the

relationship between economic growth and environmental quality is plausible

due to the changing output mix. If we look at CO2 emissions, increasing income

also leads to changes in the fuel mix, e.g. from coal to natural gas. Another

reason for this relationship could be that environmental awareness tends to grow

with income, i.e. there is positive income elasticity for environmental quality.

This increased demand for environmental quality then forces an active

environmental policy and the creation of institutions to internalize negative

effects caused by pollution. Technological progress can also explain the

appearance of an EKC, e.g. in the form of increasing production efficiency or

changes in the production process that are emission specific.

Early empirical research on the EKC was conducted by Grossman and Krueger

(1991), Shafik and Bandyopadhyay (1992) and Selden and Song (1994). The

working paper by Grossman and Krueger (1991) estimated Environmental

Kuznets Curves for SO2, fine smoke and suspended particles as part of a study on

the potential environmental impacts of NAFTA. The regressions involve a cubic

function in levels of GDP per capita and various site-related variables, a time

trend, and a trade intensity variable. The turning points for SO2 and fine smoke

are at around USD 4,000 to 5,000 while the concentration of suspended particles

appeared to decline even at low income levels. The results of the paper of Shafik

and Bandyopadhyay (1992) were used in the 1992 World Development Report 21 See IBRD (1992), p.38. 22 See Stern (2003), p. 4.


20

cited above. They estimated EKCs by using three different functional forms (log-

linear, log-quadratic and a logarithmic cubic polynomial in PPP GDP per capita

as well as a time trend and site related variables) for ten different indicators (lack

of clean water, lack of urban sanitation, ambient levels of suspended particulate

matter, ambient sulfur oxides, change in forest area, annual observations of

deforestation, dissolved oxygen in rivers, fecal coliform in rivers, municipal

waste per capita, and carbon emissions per capita). Only the two air pollutants

conform to the EKC hypothesis. The turning points for both were found for

income levels of between USD 3,000 and 4,000. Selden and Song (1994)

estimated EKCs for SO2, NOx, SPM, and CO emissions primarily from

developed countries. The estimated turning points are very high compared to the

studies by Grossman and Krueger (1991) and Shafik and Bandyopadhyay (1992),

varying from USD 7,114 for CO to 13,383 for NOx. According to their analysis

the turning point for emissions is likely to be higher than that for ambient

concentrations. Following these initial papers several studies have been focusing

on air pollution, water pollution, deforestation, hazardous waste and toxins,

carbon dioxide and other pollution indicators.23

Looking at the results of these studies it becomes clear that the EKC only holds

for some of the measures of environmental quality and that there is no simple and

predictable relationship between economic growth and pollution. The results do

not only vary depending on the pollutant measurement chosen (e.g. emissions or

ambient concentrations), but also the choice of countries observed, trade effects,

functional form and methodology.

Amongst others Arrow et al. (1995) criticize that most EKC models assume there

is no feedback from environmental damage to economic activity as income is

assumed to be an exogenous variable, i.e. there is an assumption that the

economy is sustainable. The levels per unit of output in specific processes have

declined for many pollutants in developed countries over time due to strict

environmental regulations and technical innovations. Nevertheless for other

pollutants this relation is still rising, as the mix of waste changes over time so

that in sum per capita waste may not have declined. Economic activity is

inevitably environmentally disruptive in some way (simultaneity). Arrow et al.

(1995) admit that while some empirical findings “indicate that economic growth

23 See for example Huang et al. (2008), p. 240 f. for an extensive literature review.


21

may be associated with improvements in some environmental indicators, they

imply neither that economic growth is sufficient to induce environmental

improvement in general, nor that the environmental effects of growth may be

ignored, nor, indeed, that the Earth's resource base is capable of supporting

indefinite economic growth”.24

If an EKC type relationship is found it might also be a result of the effects of

trade on the distribution of polluting industries. Trade theory suggests that, under

free trade, developing countries specialize in the production of goods that are

intensive in the more abundant factors labour and natural resources. The

developed countries in turn specialize in human capital and manufactured capital

intensive activities. This specialization leads to reduction in environmental

degradation levels in developed countries and an increase in environmental

degradation in middle and low income countries. This effect is even aggravated

by higher environmental regulation standards in developed countries. So even if

an EKC exists locally, global pollution may increase, since the fastest economic

growth occurs in developing countries with the highest population growth.

Lucas et al. (1992) found evidence that stricter environmental regulation in

OECD countries has led to a relocation of polluting industries towards poorer

developing countries. According to Dasgupta et al. (2002) some critics even

argue that globalization promotes a “race to the bottom”, i.e. the curve will rise to

a horizontal line at maximum pollution levels. This is the case when

industrialized countries relax their environmental standards in order to win back

capital outflows to low-income countries. But it remains to be proven that

developed countries are really reducing their environmental standards. Others

argue that even if some pollutants are reduced with rising incomes our economies

continuously create new and unregulated pollutants. According to Dasgupta et al.

(2002) “such concerns raise the possibility that economic development will

always be accompanied by environmental risks that are either newly discovered

or generated by the use of new materials and technologies”.25

The environmental Kuznets curve therefore implies that less developed countries

– also represented by large and fast growing economies like China and India –

will experience increasing pollution levels until a large rise in per capita income

24 See Arrow et al. (1995), p. 520. 25 Dasgupta et al. (2002), p. 162.


22

has been achieved. Nevertheless the goal of environmental regulation must be a

shift to the left and flattening of the EKC (“Revised EKC” in Figure 17).

Figure 17: Alternative views on the Environmental Kuznets Curve

The results from numerous empirical papers on the EKC show that there is little

evidence for a common inverted U-shaped pathway which countries follow with

rising income. It is therefore questionable whether the EKC is a complete model

of emissions. Stern (2003) explicitly points to the importance of decomposition

analysis to study the impact of economic growth on pollutants, especially carbon

dioxide emissions. “The research challenge now is to revisit some of the issues

addressed earlier in the EKC literature using the new decomposition models and

rigorous panel data and time series statistics”.26

In the following chapters advanced time series and decomposition methods will

be used to analyse whether economic growth inevitably increases pollution,

taking into account energy efficiency improvements, changes in the energy mix

and structural effects.

26 See Stern (2003), p. 20.

New Toxics

Race to the Bottom

The EKC

Revised EKC

Economic Growth (GDP per capita)

Pollution

Source: Dasgupta et al. (2002).

The Relationship between Energy Prices, Energy Consumption and Growth

3. The Relationship between Energy Prices, Energy Consumption and

Growth

3.1 Granger Causality

One possibility to investigate the relationship between energy consumption and

GDP is the use of Granger causality tests in time series analysis. The directions

and policy implications for the causal relationship between energy consumption

and economic growth can be categorized as follows. If it is found that

unidirectional causality runs from energy consumption to economic growth, then

restrictions on the use of energy could lead to a reduction in GDP growth. Many

countries worry about this negative effect on income caused by the restricted use

of energy, as the pressure to mitigate CO2 emissions in order to slow down the

rate of climate change grows. On the other hand, if unidirectional causality runs

from GDP to energy consumption, then energy conservation measures may be

implemented with little or no adverse impacts on growth. A bi-directional causal

relationship implies that energy use and economic growth are jointly determined

and affected at the same time. If no causal relationship between the two variables

is found, then the “neutrality hypothesis” holds. Adding energy prices as a third

variable allows for an additional channel of causality and helps to investigate

whether energy prices have a significant impact on energy consumption or even a

direct effect on GDP growth.

In recent years there has been a revival of literature on the causal relationship

between energy use and economic growth. This is not only due to the fact that

supply security, climate change and rising energy prices have been on the

political agenda once again. Apart from that, there are also new developments in

econometrics, most of all the use of panel cointegration techniques that make it

possible to solve the problem of small samples which is still prominent in the

energy literature, as there is only limited availability of time-series for energy

use. In general the causality literature differs in the methodology applied, the

variables analyzed, the countries or regions studied and the time period observed

(see overview in Table 1).

The Relationship between Energy Prices, Energy Consumption and Growth

24

Stu

dy

Me

tho

dC

ou

ntr

ies

Pe

rio

db

ivar

iate

de

man

dsu

pp

lye

->y

y->

e-

1A

kinl

o (2

008)

Aut

oreg

ress

ive

Dis

trib

uted

Lag

(A

RD

L)11

Sub

-Sah

ara

Afr

ican

cou

ntrie

s19

80-2

003

1�

��

2A

l-Iria

ni (

2006

)P

anel

Coi

nte g

ratio

n an

d V

ecto

r E

rror

Cor

rect

ion

6 G

CC

cou

ntrie

s19

71-2

002

1�

3A

ltina

y an

d K

arag

ol (

2004

)G

rang

er c

ausa

lity,

Hsi

aoT

UR

1950

-200

01

X4

Asa

fu-A

djay

e (2

000)

Coi

nteg

ratio

n an

d V

ecto

r E

rror

Cor

rect

ion

IND

, ID

N, T

HA

, PH

L19

71/7

3-19

951

�

5C

hen,

Kuo

and

Che

n (2

007)

Pan

el C

oint

egra

tion

and

Vec

tor

Err

or C

orre

ctio

n10

asi

an c

ount

ries

1971

-200

11

��

6C

hen g

and

Lai

(19

97)

Coi

nteg

ratio

n an

d V

ecto

r E

rror

Cor

rect

ion

TW

N19

55-1

993

1�

7C

hont

anaw

at e

t al (

2006

)G

ran g

er c

ausa

lity,

Hsi

ao, E

CM

> 1

00 c

ount

ries

1960

-200

0 (O

EC

D);

1

XX

X8

Cho

ntan

awat

et a

l (20

08)

Gra

nger

cau

salit

y, H

siao

, EC

M>

100

cou

ntrie

s19

60-2

000

(OE

CD

);

1X

XX

9D

inda

and

Coo

ndoo

(20

06)

Pan

el C

oint

egra

tion

and

Vec

tor

Err

or C

orre

ctio

n88

cou

ntrie

s gr

oupe

d in

to 1

2 r

egio

ns19

60-1

990

1X

X10

Ero

l and

Yu

(198

7)S

ims

& G

rang

erU

SA

1973

-198

41

�

11F

atai

et a

l (20

04)

Coi

nte g

ratio

n an

d V

ecto

r E

rror

Cor

rect

ion

(Eng

le G

rang

er,

NZL

, AU

S, I

ND

, ID

N, P

HL,

TH

A19

60-1

999

1X

X12

Gho

sh (

2002

)C

oint

egra

tion

and

Vec

tor

Err

or C

orre

ctio

nIN

D19

50-1

997

1X

13G

lasu

re a

nd L

ee (

1997

)C

oint

egra

tion

and

Vec

tor

Err

or C

orre

ctio

nK

OR

, SG

P19

61-1

990

1�

�

14H

ondr

oyia

nnis

et a

l (20

02)

Coi

nteg

ratio

n an

d V

ecto

r E

rror

Cor

rect

ion

GR

C19

60-1

996

1X

X15

Hu

and

Lin

(200

8)T

hres

hold

Coi

nteg

ratio

n an

d V

ecto

r E

rror

Cor

rect

ion

TW

N19

82-2

006

(qua

terly

) 1

XX

16Ji

nke

et a

l (20

08)

Coi

nteg

ratio

n an

d V

ecto

r E

rror

Cor

rect

ion

US

A, J

PN

, KO

R, C

HN

, IN

D, Z

AF

1980

-200

51

XX

17Ju

mbe

(20

04)

Coi

nteg

ratio

n an

d V

ecto

r E

rror

Cor

rect

ion

MW

I19

70-1

999

1X

X18

Kra

ft an

d K

raft

(197

8)S

ims

& G

rang

erU

SA

1947

-197

41

�

19Le

e (2

005)

Pan

el C

oint

e gra

tion

and

Vec

tor

Err

or C

orre

ctio

n18

dev

elop

ing

coun

trie

s19

75-2

001

1�

20Le

e (2

006)

Gra

nger

cau

salit

y, T

oda-

Yam

amot

oG

-11

1960

-200

1 (D

EU

: 197

1-20

01; C

AN

: 196

5-20

01)

1X

XX

21Le

e an

d C

hang

(20

07)

Pan

el C

oint

egra

tion

and

Vec

tor

Err

or C

orre

ctio

n22

dev

elop

ed a

nd 1

8 de

velo

ping

cou

ntrie

s19

65(1

972)

-200

21

��

22Le

e an

d C

hang

(20

08)

Pan

el C

oint

egra

tion

and

Vec

tor

Err

or C

orre

ctio

n16

Asi

an c

ount

ries

1971

-200

21

�

23Le

e an

d Le

e (2

010)

Pan

el C

oint

egra

tion

and

Vec

tor

Err

or C

orre

ctio

n25

OE

CD

cou

ntrie

s19

78-2

004

1X

X24

Lee

et a

l (20

08)

Pan

el C

oint

egra

tion

and

Vec

tor

Err

or C

orre

ctio

n22

OE

CD

cou

ntrie

s19

60-2

001

1X

X25

Mah

adev

an a

nd A

safu

-Adj

aye

(200

7)P

anel

Coi

nteg

ratio

n an

d V

ecto

r E

rror

Cor

rect

ion

20 e

nerg

y im

port

ers

and

expo

rter

s19

71-2

002

1�

�

26M

asih

and

Mas

ih (

1996

)C

oint

egra

tion

and

Vec

tor

Err

or C

orre

ctio

nIN

D, P

AK

, MY

S, S

GP

, ID

N, P

HL

1955

-199

0, 1

955-

1991

, 19

60-1

990

1X

XX

27M

asih

and

Mas

ih (

1997

)C

oint

egra

tion

and

Vec

tor

Err

or C

orre

ctio

nK

OR

, TW

N19

55-1

991

& 1

952-

1992

1�

�

28M

ehra

ra (

2007

)P

anel

Coi

nteg

ratio

n an

d V

ecto

r E

rror

Cor

rect

ion

11 o

il ex

port

ing

coun

trie

s19

71-2

002

1�

29M

ozum

der

and

Mar

athe

(20

07)

Coi

nteg

ratio

n an

d V

ecto

r E

rror

Cor

rect

ion

BG

D19

71-1

999

1X

30N

acha

ne e

t al (

1988

)C

oint

egra

tion

and

Vec

tor

Err

or C

orre

ctio

n16

cou

ntrie

s (1

1 LD

Cs,

5 D

Cs)

1950

-198

51

��

X

31N

aray

an a

nd P

rasa

d (2

008)

boot

stra

pped

Gra

nger

cau

salit

y te

st30

OE

CD

cou

ntrie

s19

60-2

002,

196

5-20

02,

1970

-200

2, 1

971-

2002

1X

XX

32N

aray

an a

nd S

myt

h (2

008)

Pan

el C

oint

egra

tion

and

Vec

tor

Err

or C

orre

ctio

nG

719

72-2

002

1X

33O

h an

d Le

e (2

004)

Coi

nteg

ratio

n an

d V

ecto

r E

rror

Cor

rect

ion

KO

R19

70-1

999

1�

�

34S

ari e

t al (

2008

)A

RD

LU

SA

2001

:1-2

005:

6 (m

)1

XX

35S

hiu

and

Lam

(20

04)

Coi

nteg

ratio

n an

d V

ecto

r E

rror

Cor

rect

ion

CH

N19

71-2

000

1X

36S

oyta

s an

d S

ari (

2003

)C

oint

egra

tion

and

Vec

tor

Err

or C

orre

ctio

nG

7 an

d to

p 10

em

ergi

ng c

ount

ries

(exc

l. C

HN

)

1950

-199

2 (A

RG

: 195

0-19

90; I

DN

: 196

0-19

92;

KO

R: 1

953-

1991

; PO

L:

1965

-199

4)1

XX

X

37S

tern

(19

93)

VA

R a

nd G

rang

er c

ausa

lity

US

A19

47-1

990

1X

X

38S

tern

(20

00)

Coi

nteg

ratio

n an

d V

ecto

r E

rror

Cor

rect

ion

US

A19

48-1

994

1X

39Y

ang

(200

0)C

oint

egra

tion

and

Vec

tor

Err

or C

orre

ctio

nT

WN

1954

-199

71

��

40Y

u an

d H

wan

g (1

984)

Sim

s &

Gra

nger

US

A19

47-1

979

1�

41Za

char

iadi

s (2

007)

3 m

odel

s: V

EC

, AR

DL,

Tod

a-Y

amam

oto

G-7

1960

-200

41

XX

X

42Zo

u an

d C

hau

(200

6)C

oint

egra

tion

and

Gra

nger

cau

salit

yC

HN

1953

-200

21

X(X

)

mu

ltiv

aria

teD

ire

ctio

n o

f ca

usa

lity

Table 1: Literature review

Source: author’s own illustration.

economic growth, energy demand and the environment ... · transportation have lead to an unexpected...

Documents