Sie sind auf Seite 1von 264

Grow Rich and Clean Up Later?

Joint Eects of
International Integration and Democracy on
Environmental Quality in Developing Countries
1
Gabriele Ruo
August 21, 2009
1
For detailed suggestions and comments, I thank Thomas Bernauer, Anna Kalb-
henn, Erik Gartzke, Simon Hug, Vally Koubi, Lena Schaer, Cornelia Schmidt and
Vera Troger. For their help concerning the formal model I thank Donja Darai, Har-
ris Dellas, Thorsten Ehlers and Christian Heppenstrick. This thesis was written
in the context of the Swiss National Research Program on Democracy in the 21st
Century.
Abstract
Many forms of environmental degradation rst increase, then level o and
decrease as national income grows. Does this mean that poorer countries
have to grow rich before they can improve their environmental performance?
This thesis argues that integration into the international system in combi-
nation with democratic forms of government mitigates this dilemma. This
argument is tested on panel data for 115 developing countries in 1970-2000.
The thesis examines whether enlarging the scope of international integra-
tion to encompass membership in international governmental organizations
(IGOs), international aid, trade openness and FDI inows adds to our un-
derstanding of developing countries environmental performance.
The empirical analysis shows that membership in IGOs is the only aspect
of globalization that is robustly associated to a reduction in air and water
pollution. Furthermore, whereas the type of the political system does not
appear to directly aect developing countries environmental performance, it
strongly mediates the eect of international integration. For IGO member-
ship we see that democracy seems to amplify the positive integration eect
with respect to SO
2
emissions but dampens the eect for CO
2
emissions and
water pollution. Concerning trade openness and foreign direct investment,
the results show that at any given level of trade openness or FDI, air pollu-
tion emissions are higher in autocratic developing countries than democratic
developing countries.
To better understand the mechanisms underlying the robust eect of
IGO membership, four case studies illustrate how the mandate and objec-
tives of IGOs can increase the environmental performance of developing
countries. The analysis of Cameroon, India, Laos and Lithuanias IGO
memberships points to three main mechanisms through which IGOs seem
to inuence a countrys environmental performance: issue linkage, technical
cooperation and know-how diusion, and pushing the topic of environmental
protection or sustainable development on the agenda of developing countries.
Relying on a specic-factors trade model, one of these mechanisms, the
transfer of technologies, is then taken up to build a simple formal theoreti-
cal model that links IGO membership to environmental pollution. Modeling
IGO membership as a trade-o between selling ones exports at better con-
ditions but using a more environmentally stringent technology, the results
show that for a reasonable setting of the model parameters, both democratic
and autocratic forms of government opt for at least some IGO membership,
thus making production more environmentally friendly.
Contents
1 Introduction 11
2 Literature Review 19
2.1 Literature on environmental quality: the Environmental Kuznets
Curve . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
2.2 Enlarging the Scope Step 1: Environmental Quality and the
Political System . . . . . . . . . . . . . . . . . . . . . . . . . . 22
2.3 Enlarging the Scope Step 2: Environmental Quality and Glob-
alization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
2.3.1 International Governmental Organizations . . . . . . . 26
2.3.2 International Aid . . . . . . . . . . . . . . . . . . . . . 30
2.3.3 Economic Integration: Trade and Foreign Direct In-
vestment . . . . . . . . . . . . . . . . . . . . . . . . . 33
2.3.3.1 Trade . . . . . . . . . . . . . . . . . . . . . . 35
2.3.3.2 Foreign Direct Investment . . . . . . . . . . . 37
2.3.3.2.1 Competition for FDI . . . . . . . . 37
2.3.3.2.2 Eects on the Environment . . . . . 39
2.4 The Political System and International Integration . . . . . . 41
2.4.1 The Political System and International Government
Organizations . . . . . . . . . . . . . . . . . . . . . . . 42
2.4.2 The Political System and International Aid . . . . . . 43
2.4.3 The Political System and Trade . . . . . . . . . . . . . 44
2.4.4 The Political System and Foreign Direct Investment . 44
2.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
3 Theory 48
3.1 Income and Environmental Quality . . . . . . . . . . . . . . . 49
3.2 No Independent Eect of Political System . . . . . . . . . . . 50
3.3 International Integration . . . . . . . . . . . . . . . . . . . . . 52
3.3.1 Membership in International Organizations . . . . . . 52
3.3.2 International Assistance . . . . . . . . . . . . . . . . . 54
3.3.3 Trade . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
3.3.4 FDI . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
1
CONTENTS
3.4 Interconnection with the Political System . . . . . . . . . . . 60
3.4.1 The Political System and the Response to Interna-
tional Integration . . . . . . . . . . . . . . . . . . . . . 61
3.4.2 The Political System and the Degree of International
Integration . . . . . . . . . . . . . . . . . . . . . . . . 63
4 Research Design 67
4.1 General Research Strategy . . . . . . . . . . . . . . . . . . . . 67
4.2 Sample Selection . . . . . . . . . . . . . . . . . . . . . . . . . 68
4.3 Estimation Method . . . . . . . . . . . . . . . . . . . . . . . . 69
4.4 Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
4.4.1 Dependent Variable . . . . . . . . . . . . . . . . . . . 73
4.4.1.1 SO
2
. . . . . . . . . . . . . . . . . . . . . . . 74
4.4.1.2 CO
2
. . . . . . . . . . . . . . . . . . . . . . . 75
4.4.1.3 BOD . . . . . . . . . . . . . . . . . . . . . . 75
4.4.2 Independent Variables . . . . . . . . . . . . . . . . . . 76
4.4.2.1 International Governmental Organizations . 76
4.4.2.2 International Aid . . . . . . . . . . . . . . . 79
4.4.2.3 Trade . . . . . . . . . . . . . . . . . . . . . . 79
4.4.2.4 Foreign Direct Investment . . . . . . . . . . . 80
4.4.2.5 Political System . . . . . . . . . . . . . . . . 80
4.4.2.6 National Income . . . . . . . . . . . . . . . . 81
4.4.3 Control Variables . . . . . . . . . . . . . . . . . . . . . 83
4.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
5 General Results 85
5.1 Results Standard Model: Independent Eect of International
Integration . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
5.2 Analyzing the Mediating Eect of the Political System . . . . 92
5.2.1 Indirect Eect of the Political System through Inter-
national Integration . . . . . . . . . . . . . . . . . . . 92
5.2.2 Interaction between IGO Membership and Political
System . . . . . . . . . . . . . . . . . . . . . . . . . . 94
5.2.3 Interaction between Aid and Political System . . . . . 98
5.2.4 Interaction between Trade and Political System . . . . 101
5.2.5 Interaction between FDI and Political System . . . . . 105
5.3 Robustness Checks . . . . . . . . . . . . . . . . . . . . . . . . 108
5.3.1 Evaluating the Eect of Democracy: Regime Change
and the Eect of Democracy . . . . . . . . . . . . . . 108
5.3.2 Robustness Checks with regard to Model Specication 111
5.3.3 Robustness Checks with regard to Estimation Method 116
5.4 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
2
CONTENTS
6 IGO Membership and Environmental Quality: Understand-
ing the Mechanisms 122
6.1 Evaluating the Robustness of IGO Membership . . . . . . . . 123
6.1.1 Accounting for the Stringency of IGO Membership . . 123
6.1.2 Accounting for the Trend in IGO Membership . . . . 125
6.2 Disaggregating the Eect of IGO Membership . . . . . . . . . 127
6.2.1 Disaggregating IGO Membership according to its Func-
tion . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
6.2.2 Disaggregating IGO Membership according to its De-
gree of Institutionalization . . . . . . . . . . . . . . . . 141
6.3 Endogeneity . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
6.4 Case Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . 150
6.4.1 State of the Environment . . . . . . . . . . . . . . . . 151
6.4.2 Common IGO Membership . . . . . . . . . . . . . . . 156
6.4.3 Cameroon . . . . . . . . . . . . . . . . . . . . . . . . . 168
6.4.4 India . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169
6.4.5 Laos . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171
6.4.6 Lithuania . . . . . . . . . . . . . . . . . . . . . . . . . 172
6.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173
7 Formal Model on IGO Membership and Environmental Qual-
ity 178
7.1 The Economy . . . . . . . . . . . . . . . . . . . . . . . . . . . 178
7.2 Government . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180
7.2.1 Government Favors Elite Member in the Export Sector 183
7.2.2 Government Favors Elite Member in the Domestic
Sector . . . . . . . . . . . . . . . . . . . . . . . . . . . 193
7.2.3 Government Favors Representative Voter (Worker) in
the Export Sector . . . . . . . . . . . . . . . . . . . . 201
7.2.4 Government Favors Representative Voter (Worker) in
the Domestic Sector . . . . . . . . . . . . . . . . . . . 208
7.3 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . 214
8 Conclusion 216
9 Appendix 225
9.1 Additional Information Chapter 4 . . . . . . . . . . . . . . 225
9.2 Additional Tables and Figures Chapter 5 . . . . . . . . . . 231
9.3 Additional Information Chapter 7 . . . . . . . . . . . . . . 236
3
List of Tables
4.1 Variables and their variation sample: all developing countries 73
4.2 Theoretical concepts and their operationalization . . . . . . . 82
5.1 Descriptive statistics sample: all developing countries . . . 86
5.2 Baseline model sample: all developing countries . . . . . . . 88
5.3 Baseline model: BOD only sample: low- and low-middle
income countries . . . . . . . . . . . . . . . . . . . . . . . . . 89
5.4 Baseline model: ACLP democracy variable sample: all de-
veloping countries . . . . . . . . . . . . . . . . . . . . . . . . 91
5.5 Eect of political system on international integration sam-
ple: all developing countries . . . . . . . . . . . . . . . . . . . 93
5.6 Interaction between IGO membership and political system
sample: all developing countries . . . . . . . . . . . . . . . . . 95
5.7 Interaction between aid and political system sample: all
developing countries . . . . . . . . . . . . . . . . . . . . . . . 99
5.8 Interaction between trade and political system sample: all
developing countries . . . . . . . . . . . . . . . . . . . . . . . 102
5.9 Interaction between FDI and political system sample: all
developing countries . . . . . . . . . . . . . . . . . . . . . . . 106
5.10 Regime change and the eect of democracy sample: all
developing countries . . . . . . . . . . . . . . . . . . . . . . . 110
5.11 Regions sample: all developing countries . . . . . . . . . . . 112
5.12 Urban population and agricultural land area sample: all
developing countries . . . . . . . . . . . . . . . . . . . . . . . 113
5.13 Number of cars and industry shares sample: all developing
countries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
5.14 Government share of GDP sample: all developing countries 115
5.15 Baseline results with a lagged dependent variable sample:
all developing countries . . . . . . . . . . . . . . . . . . . . . 117
5.16 Regression with both time and country xed eects sample:
all developing countries . . . . . . . . . . . . . . . . . . . . . 118
5.17 Regression with decade dummies and country xed eects
sample: all developing countries . . . . . . . . . . . . . . . . . 119
4
LIST OF TABLES
6.1 Varying the stringency of the IGO variable sample: all de-
veloping countries . . . . . . . . . . . . . . . . . . . . . . . . 124
6.2 IGO membership in deviations from the mean sample: all
developing countries . . . . . . . . . . . . . . . . . . . . . . . 126
6.3 IGO membership disaggregated according to IGO function
sample: all developing countries . . . . . . . . . . . . . . . . . 128
6.4 IGO membership in % of possible IGO membership sample:
all developing countries . . . . . . . . . . . . . . . . . . . . . 134
6.5 Military Organizations . . . . . . . . . . . . . . . . . . . . . . 138
6.6 Military Organizations . . . . . . . . . . . . . . . . . . . . . . 139
6.7 IGO membership disaggregated according to whether the IGO
has some environmental function sample: all developing
countries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
6.8 IGO membership disaggregated according to the degree of
institutionalization sample: all developing countries . . . . 143
6.9 Testing potential instrumental variables sample: all devel-
oping countries . . . . . . . . . . . . . . . . . . . . . . . . . . 145
6.10 Regression instrumenting IGO membership sample: all de-
veloping countries . . . . . . . . . . . . . . . . . . . . . . . . 148
6.11 Regression instrumenting environmental IGO membership
sample: all developing countries . . . . . . . . . . . . . . . . . 149
6.12 Membership in environmental IGOs . . . . . . . . . . . . . . 166
6.13 Membership in environmental treaties . . . . . . . . . . . . . 167
7.1 Starting values for the baseline simulation . . . . . . . . . . . 187
9.1 Time periods for countries SO
2
and CO
2
. . . . . . . . . . . 225
9.2 Time periods for countries SO
2
and CO
2
. . . . . . . . . . . 226
9.3 Time periods for countries SO
2
and CO
2
. . . . . . . . . . . 227
9.4 Additional countries and time periods CO
2
. . . . . . . . . 227
9.5 Time periods for countries BOD . . . . . . . . . . . . . . . 228
9.6 Time periods for countries BOD . . . . . . . . . . . . . . . 229
9.7 Time periods for countries BOD . . . . . . . . . . . . . . . 230
9.8 Interaction between IGO membership and political system
sample: low-income countries . . . . . . . . . . . . . . . . . . 231
9.9 Interaction between aid (divided by GDP) and political sys-
tem sample: all developing countries . . . . . . . . . . . . . 232
9.10 Interaction between log of export divided by GDP and polit-
ical system sample: all developing countries . . . . . . . . . 234
9.11 Gravity model including joint IGO membership . . . . . . . . 238
5
List of Figures
1.1 Relationship between GDP per capita and SO
2
emissions in
2000 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
3.1 Illustration of theoretical argument . . . . . . . . . . . . . . . 49
5.1 Marginal eect of IGO membership SO
2
. . . . . . . . . . . 96
5.2 Marginal eect of IGO membership CO
2
. . . . . . . . . . . 96
5.3 Marginal eect of IGO membership BOD . . . . . . . . . . 96
5.4 Marginal eect of IGO membership BOD . . . . . . . . . . 97
5.5 Marginal eect of aid SO
2
. . . . . . . . . . . . . . . . . . . 100
5.6 Marginal eect of aid CO
2
. . . . . . . . . . . . . . . . . . . 100
5.7 Marginal eect of aid BOD . . . . . . . . . . . . . . . . . . 100
5.8 Marginal eect of trade openness SO
2
. . . . . . . . . . . . 103
5.9 Marginal eect of trade openness CO
2
. . . . . . . . . . . . 103
5.10 Marginal eect of trade openness BOD . . . . . . . . . . . . 103
5.11 Marginal eect of trade openness SO
2
. . . . . . . . . . . . 104
5.12 Marginal eect of trade openness CO
2
. . . . . . . . . . . . 104
5.13 Marginal eect of FDI SO
2
. . . . . . . . . . . . . . . . . . 105
5.14 Marginal eect of FDI CO
2
. . . . . . . . . . . . . . . . . . 107
5.15 Marginal eect of FDI BOD . . . . . . . . . . . . . . . . . . 107
5.16 Marginal eect of FDI BOD . . . . . . . . . . . . . . . . . . 107
6.1 Marginal eect of IGO membership according to its function
SO
2
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
6.2 Marginal eect of IGO membership according to its function
CO
2
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
6.3 Marginal eect of IGO membership according to its function
BOD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
6.4 Marginal eect of % of IGO membership according to its func-
tion SO
2
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
6.5 Marginal eect of % of IGO membership according to its func-
tion CO
2
. . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
6.6 Marginal eect of % of IGO membership according to its func-
tion BOD . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
6
LIST OF FIGURES
6.7 State of the environment: Cameroon . . . . . . . . . . . . . . 152
6.8 State of the environment: India . . . . . . . . . . . . . . . . . 153
6.9 State of the environment: Laos . . . . . . . . . . . . . . . . . 154
6.10 State of the environment: Lithuania . . . . . . . . . . . . . . 155
6.11 IGO membership by country and over time . . . . . . . . . . 158
7.1 Optimal IGO membership at dierent levels of technology A 188
7.2 Optimal IGO membership at dierent levels of
K
y
K
x
. . . . . . 189
7.3 Optimal IGO membership with varying price of export good p
x
191
7.4 Optimal IGO membership with varying share of value added
that goes to workers . . . . . . . . . . . . . . . . . . . . . . 192
7.5 Optimal IGO membership at dierent levels of technology A 197
7.6 Optimal IGO membership at dierent levels of
K
x
K
y
. . . . . . 198
7.7 Optimal IGO membership with varying price of export good p
x
199
7.8 Optimal IGO membership with varying share of value added
that goes to workers . . . . . . . . . . . . . . . . . . . . . . 200
7.9 Optimal IGO membership at dierent levels of technology A 204
7.10 Optimal IGO membership at dierent levels of
K
y
K
x
. . . . . . 205
7.11 Optimal IGO membership with varying price of export good p
x
206
7.12 Optimal IGO membership with varying share of value added
that goes to workers . . . . . . . . . . . . . . . . . . . . . . 207
7.13 Optimal IGO membership at dierent levels of technology A 210
7.14 Optimal IGO membership at dierent levels of
K
y
K
x
. . . . . . 211
7.15 Optimal IGO membership with varying price of export good p
x
212
7.16 Optimal IGO membership with varying share of value added
that goes to workers . . . . . . . . . . . . . . . . . . . . . . 213
8.1 Environmental quality and GDP per capita (year 2000) . . . 217
8.2 The eect of GDP on SO
2
at various levels of IGO membership224
9.1 Marginal eect of aid scaled by GDP SO
2
. . . . . . . . . . 233
9.2 Marginal eect of aid scaled by GDP CO
2
. . . . . . . . . . 233
9.3 Marginal eect of aid scaled by GDP BOD . . . . . . . . . 233
9.4 Marginal eect of export divided by GDP SO
2
. . . . . . . 235
9.5 Marginal eect of export divided by GDP CO
2
. . . . . . . 235
7
List of Abbreviations
AARDO Afro-Asian Rural Development Organization
ASEAN Association of Southeast Asian Nations
ATO African Timber Organization
BC Baltic Council
BEF Baltic Environmental Forum
BIONET Global Network for Biosystematics
BOD Biological Oxygen Demand
CBD Convention on Biological Diversity
CENTO Central Treaty Organization
CFC Chlorouorocarbons
CO
2
Carbon Dioxide
COW Correlates of War
EAPC Euro-Atlantic Partnership Council
EBRD European Bank for Reconstruction and Development
ECCAS Economic Community of Central African States
EKC Environmental Kuznets Curve
EU European Union
FAO Food and Agriculture Organization
FDI Foreign Direct Investment
GATT General Agreement on Taris and Trade
GDP Gross Domestic Product
GEF Global Environment Facility
8
LIST OF FIGURES
IADefB Inter-American Defense Board
IAEA International Atomic Energy Agency
IBRD International Bank for Reconstruction and Development
ICivDO International Civil Defence Organization
ICSG International Copper Study Group
IDA International Development Association
IFAD International Fund for Agriculture and Development
IFC International Finance Corporation
IGO International Governmental Organization
IIR International Institute of Refrigeration
ILO International Labour Organization
IMF International Monetary Fund
IMO International Maritime Organization
IV Instrumental Variable
MEA Multilateral Environmental Agreement
MNC Multinational Corporation
NAM Non-Aligned Movement
NATO North Atlantic Treaty Organization
NGO Non-Governmental Organization
ODS Ozone-Depleting Substances
OECD Organisation for Economic Co-operation and Development
OLADE Latin American Energy Organization
OLS Ordinary Least Squares
OSCE Organization Security Cooperation Europe
PCSE Panel Correct Standard Errors
PPP Purchasing Power Parity
SAARC South Asian Association for Regional Cooperation
SEAMEO Southeast Asian Ministers of Education Organization
SEATO Southeast Asia Treaty Organization
9
LIST OF FIGURES
SO
2
Sulfur Dioxide
TSCS Time-Series Cross-Section
UN United Nations
UNCCD UN Convention to Combat Desertication
UNEP United Nations Environment Programme
UNESCO United Nations Educational, Scientic and Cultural
Organization
UNFCCC United Nations Framework Convention on Climate Change
UNIDO UN Industrial Development Organization
US United States
VASAB Vision and Strategies around the Baltic Sea
WHO World Health Organization
WMO World Meteorological Organization
WTO World Trade Organization
WWF World Wide Fund For Nature
10
Chapter 1
Introduction
Does integration into the international system help developing countries
provide public goods? The importance of this question lies in the fact that
poverty or low national income are often associated with low levels of public
goods provision, for example in areas such as health care, education or
environmental sustainability. Wealthy nations, in contrast, are characterized
by higher life expectancy, lower illiteracy rates
1
, superior environmental
performance
2
, and other conditions that the Millennium Assessment Process
(UN 2009a) and policy-makers worldwide regard as highly desirable.
This thesis concentrates on one important type of public good, namely
environmental quality. The theoretical and empirical literature on the re-
lationship between income and environmental degradation indicates that
many forms of pollution tend to rise monotonically with growing income
(Grossman & Krueger 1995). However, after a certain income threshold is
reached, some pollutants tend to level o and decline. Typical examples
include air and water pollution. This non-linear inverted U-shaped relation-
ship is known as the Environmental Kuznets Curve (EKC) (Grossman &
Krueger 1995; Selden & Song 1994). According to the logic of the EKC, an
increase in national income should be associated with better environmental
quality in industrialized countries whereas developing countries should be
confronted with deteriorating environmental quality when their GDP per
capita grows.
Does this regularity imply that poor countries need to become rich in
order to be able to oer high levels of environmental public goods to their
1
For a ranking of countries by life expectancy and illiteracy rates, see for example the
Human Development Index (UNDP 2009).
2
The Environmental Performance Index (Center for International Earth Science Infor-
mation Network (CIESIN) 2002) shows that the average sustainability score of high-income
countries is higher compared to those of middle- or low-income countries. This nding,
however, should not obscure the fact that there is strong variation within income brackets:
some high-income countries exhibit poor environmental performance and some low-income
countries high environmental performance.
11
CHAPTER 1. INTRODUCTION
population? As shown in Figure 1, the empirical pattern between environ-
mental degradation (shown here in terms of local air pollution in particular
sulfur-dioxide (SO
2
) emissions for all developing countries in the year 2000)
and GDP per capita suggests that this may not be the case. The wide vari-
ation in pollution levels around the regression line indicates that national
income is not as decisive for national environmental quality as the literature
on the EKC suggests.
Figure 1.1: Relationship between GDP per capita and SO
2
emissions in 2000
Although several countries seem to t the EKC pattern, like India or
Sri Lanka which both show increasing income together with increasing pol-
lution levels, many other countries do not. When we look at Albania or
Lithuania, for example, we observe, even long after the collapse of com-
munism, a strong rise in environmental quality. This is especially striking
because the national income of both countries has increased at the same
time. According to the EKC, this increase in income should have implied a
decline in environmental quality because both Albania and Lithuania, like
all developing countries, are still situated on the upward sloping part of the
Environmental Kuznets Curve for which increases in national income are
associated with decreases in environmental quality. In contrast, in the case
of Ecuador or Ivory Coast, environmental quality has worsened while at the
same time their national income has stagnated or even declined. Finally, we
also observe cases in which there is no obvious relationship between GDP
per capita and environmental quality at all, as for example in Cameroon or
the Philippines.
Positive or negative deviations from the EKC for any given number of
countries and environmental quality measures can in principle be due to a
variety of factors. The scholarly literature pays considerable attention to
the eects of democracy and economic openness to explain variations in
environmental quality at any given level of income (Antweiler et al. 2001;
Neumayer 2002b; Ward 2008; Baettig & Bernauer 2009). This thesis adds
12
CHAPTER 1. INTRODUCTION
to this existing body of research by developing and testing the argument
that integration into the international system, dened more broadly than
just economic openness, can help developing countries to cut through the
EKC (i.e. reach the turning point at lower levels of environmental degra-
dation, and/or earlier on in the stage of economic development). Moreover,
this thesis combines this argument with earlier analyses on the relationship
between the political system and the provision of environmental quality by
arguing that the eect of international integration is conditional on the type
of the political system (democratic versus autocratic forms of government)
of a country. This conditionality implies that I expect an interaction ef-
fect between a countrys political system on the one side and its level of
international integration on the other.
More precisely, I argue that international integration might improve the
environmental performance of developing countries because it can increase
both the capability and willingness of developing countries to provide better
environmental quality than could be expected given their level of national
income. Most of the literature on globalization
3
and the provision of public
goods focuses solely on the economic facets of international integration, such
as trade (Antweiler et al. 2001; Frankel & Rose 2005) and foreign direct in-
vestment (FDI) (Mani & Wheeler 1998; Busse 2004). Very few studies have
taken the political dimension of globalization into account. Ward (2006) is
a notable exception as he examines whether countries that are more central
to the network of international environmental regimes act more sustainably
at the national level. However, all of the existing studies so far are limited
to one particular facet of international integration (be it trade openness, for-
eign direct investment or membership in environmental regimes) capturing,
therefore, only a specic angle of globalization. This thesis therefore intends
to overcome this limitation by following a broader understanding of interna-
tional integration by analyzing how several dierent aspects of international
integration inuence the capability and willingness of developing countries
to provide environmental quality.
In particular, the analysis covers four dierent facets of international
integration, which are membership in international governmental organiza-
tions (IGOs), international aid, trade openness and foreign direct invest-
ment. Although the four dierent aspects of international integration are
assumed to aect the capability and willingness of developing countries to
provide environmental quality in dierent ways and to varying degrees, I
argue that the overall eect of all of them is positive and thus conducive to
the provision of environmental quality
4
.
3
For the remainder of the thesis, I use the terms globalization and international inte-
gration interchangeably.
4
At this point it is important to clarify that the interest of this thesis lies not with eval-
uating the eectiveness of membership in environmental organizations or with analyzing
the eectiveness of green development aid or investment but with assessing how interna-
13
CHAPTER 1. INTRODUCTION
These four aspects of international integration membership in IGOs,
international aid, trade openness and FDI were chosen for several reasons.
First, they cover the most important angles of globalization comprising the
whole spectrum from political towards economic international integration.
Second, all of these four aspects are well researched in the literature while
the relationship between these facets of international integration and envi-
ronmental quality has only been analyzed with regard to trade openness and
foreign direct investment. Third, these aspects of globalization take place
in a structured and institutionalized setting that is geared towards long-
term interaction. Hence these facets of international integration describe
repeated interactions, making it possible to analyze how they aect envi-
ronmental quality, a public good whose provision is often characterized by its
long-term nature
5
. Fourth, with respect to all four aspects of globalization,
governments are important driving forces behind the extent of a countrys
intensity of international integration by either making the decisions to join
a particular IGO, by setting appropriate regulations to attract foreign in-
vestment and trade ows or by being the executing agency for development
aid. Hence analyzing the interaction between a countrys political system
and any of the four forms of international integration, which constitutes the
second major contribution of this thesis, is in all of these four instances a
meaningful endeavor.
Concerning the interaction between international integration and the po-
litical system, I argue that the type of the political system should not have
an independent eect on environmental quality in developing countries, but
its interplay with international integration should be decisive for improved
environmental quality. More specically, I posit that the hypothesized pos-
itive eect of international integration is intensied in democratic political
systems, suggesting that at any given level of international integration envi-
ronmental quality will be better in democracies relative to autocracies. This
implies that the political system does not aect the environmental perfor-
mance of developing countries directly but only through the indirect link via
international integration.
I empirically test the propositions of the theoretical arguments with a
time-series cross-section analysis of 115 developing countries from 1970 to
2000. To proxy environmental quality I rely on three dierent measures of
environmental degradation, which are sulfur dioxide (SO
2
) and carbon diox-
ide (CO
2
) emissions measuring air quality, and biological oxygen demand
tional integration dened more broadly by including dierent facets of globalization
alters countries environmental performance.
5
This is one important reason why only membership in IGOs and not membership in
international treaties such as environmental treaties is considered in this analysis. Al-
though some of these treaties lead to repeated interaction even after ratication, some
treaties do not extend into member countries domestic actions later on at all, rendering
their analysis unsuitable for the setting at hand.
14
CHAPTER 1. INTRODUCTION
(BOD) measuring water quality. These three indicators cover important as-
pects of a countrys environmental performance, thus giving a broad picture
of environmental quality. After analyzing the eect of IGO membership,
international aid, trade openness and foreign investment and their interplay
with the political system on the environmental performance of developing
countries, this thesis focuses on one specic factor of international integra-
tion, which is membership in IGOs. Since this is the only factor of interna-
tional integration that is robustly associated to an increase in environmental
quality, I strive to understand more fully the mechanisms that are behind
these ndings. To this end further time-series cross-country analyses are
employed using a more rened coding of IGO membership. Moreover, this
detailed quantitative analysis of IGO membership on environmental quality
is further complemented by four illustrative case studies. These case studies
are conducted to assess in more detail which of the mechanisms proposed
in the theoretical section are indeed underlying the aggregate relationship
between IGO membership and environmental quality that come out of the
statistical analysis. Therefore, the main goal of these case studies is to eval-
uate and illustrate which of the theoretically proposed links are empirically
at play in the everyday business of IGOs in developing countries.
One mechanism, the diusion of technology, which is highlighted by both
the statistical as well as the case study analysis as being one driving force
behind the inuence of IGO membership, is then taken up in the last part
of the thesis to develop a simple formal theoretical model that links IGO
membership to environmental pollution. The aim of this model is to obtain
a more general theoretical framework for the study of IGO membership and
its links to environmental pollution. Using a specic-factors trade model,
this theoretical framework conceives of IGO membership as a trade-o for
each country between being able to sell ones exports at better conditions but
using a more environmentally stringent technology, thus making production
more costly. The results of the model show that for a reasonable setting of
the parameters in the model all types of government both democratic as
well as autocratic opt for at least some IGO membership, thus making
production more environmentally friendly.
The results that come out of this mixed methods approach show a di-
verse picture of environmental quality in developing countries. As pointed
out above, the only factor that is robustly associated to a decrease in all three
forms of environmental pollution is IGO membership. Trade openness only
signicantly inuences CO
2
emissions; however, it leads to an increase in
emissions and not to a decrease as was theoretically expected. International
aid is in none of the models signicantly associated to environmental pollu-
tion. Similarly, with the exception of water quality in low-income countries,
FDI is not signicantly associated with any of the three indicators of envi-
ronmental quality, as well. As hypothesized, political regime type does not
have an independent eect on environmental quality in developing countries;
15
CHAPTER 1. INTRODUCTION
however, it strongly mediates the eect of the four aspects of international
integration. With regard to IGO membership and in line with the theo-
retical arguments, democracy amplies the positive integration eect with
respect to SO
2
emissions, but contrary to theoretical predictions democ-
racy dampens the eect with respect to CO
2
emissions and water pollution.
Turning to trade openness and FDI, for both SO
2
and CO
2
emissions the
increasing eect on pollution levels is more pronounced in autocracies than
in democracies. This implies that at any given level of trade openness or
foreign investment, pollution emissions are higher in autocratic developing
countries compared to democratic developing countries. Concerning inter-
national aid, there is no systematic pattern of aid and the political system
on a countrys environmental performance.
The more detailed analysis of IGO membership that follows the aggre-
gate statistical analysis shows that the positive impact of IGO membership
on environmental quality is robust to various measurement specications.
Furthermore, when disaggregating IGO membership according to the func-
tion and the degree of institutionalization of the organization, we get some
very interesting ndings. First, we see that it is not membership in environ-
mental IGOs that is driving the results. This result arms our conclusion
that IGO membership is not endogenous to environmental quality and that
therefore the causality indeed runs in the expected direction
6
. In contrast, it
is membership in umbrella, nuclear, agricultural, economic standardization
and military organizations that is positively associated to an increase in en-
vironmental quality. Second, these ndings provide us with a rst indication
of the mechanisms behind this inuence of IGO membership. For example,
the inuence of umbrella organizations such as the European Union (EU)
or the Association of Southeast Asian Nations (ASEAN) is a rst indica-
tion that some IGOs seem to be very successful in connecting dierent issues.
Although environmental protection is usually not the main goal of these um-
brella organizations, countries joining these IGOs for economic or political
reasons are as a side eect also inuenced with regard to their environmen-
tal quality. In contrast, the inuence of agricultural, economic and nuclear
organizations rather seems to be grounded in the provision of technological
and nancial resources, thus increasing the capacity of developing countries
to provide environmental quality.
These ndings are clearly supported by the four illustrative case studies
on how IGOs inuence environmental quality in Cameroon, India, Laos and
Lithuania. When evaluating the mandate of the various organizations to
which these four countries belong, three main mechanisms become apparent
through which IGOs seem to inuence a countrys environmental perfor-
6
This conclusion is further supported by an instrumental variables regression showing
that even if we instrument IGO membership we obtain this positive and signicant eect
on environmental quality.
16
CHAPTER 1. INTRODUCTION
mance: issue linkage, technical cooperation and know-how diusion, and
pushing the topic of environmental protection or sustainable development
on the agenda of developing countries.
Concerning the rst issue we see that umbrella organizations such as the
EU or ASEAN can be very successful in linking topics such as environmental
or labor protection to trade or other economic issues. This seems to work
especially well in those cases in which the organization possesses enough
power to coerce its future member countries, as is the case with the EU or
ASEAN. Both Laos and Lithuania are examples of countries that adopted
extensive environmental regulations in order to become members of ASEAN
or the EU respectively.
With regard to the second issue, it becomes apparent from the case stud-
ies that many organizations such as the FAO, WHO, the World Tourism
Organization or UNESCO provide developing countries with knowledge and
technology needed to combat environmental pollution. This capacity build-
ing by transferring technology and diusing environmentally sound know-
how is even seen by organizations such as the Global Environment Facility
(GEF) as one of the most important strategies to deal with environmental
threats such as climate change (GEF 2008b).
Finally, the four cases show that an increasing number of IGOs have
taken up the course of sustainable development. Although not all of these
organizations do indeed provide developing countries with technologies and
resources necessary to achieve sustainable development, countries can hardly
avoid this topic any more. This trend is also becoming evident in the lending
behavior of the major development aid actors such as the World Bank. For
example, countries applying for loans with the International Development
Association (IDA), which is the part of the World Bank that provides loans
to the poorest developing countries, are obliged to provide a National En-
vironmental Action Plan, which outlines a countrys major environmental
problems and describes solutions to mitigate these problems (Gutner 2005).
These last two mechanisms further hint at a major dierence between
democracies and autocracies: both the promotion of sustainable develop-
ment as well as the diusion and transfer of technology appear to thrive
more intensively within democracies than within autocracies. This could
also be one explanation of why we see that most of the time the eect of
the four aspects of international integration are more conducive to environ-
mental quality in democracies than in autocracies. In contrast, the rst
mechanism, which refers to issue linkage, seems to work equally well for
democracies and autocracies. As long as the incentives are strong enough
to join these rather coercive organizations, even autocracies such as Laos
agree to implement environmental regulations although their major interest
in joining an organization such as ASEAN are clearly economic in nature.
The empirical and theoretical implications of this thesis for the study of
public goods provision are twofold: rst, contrary to the results of previous
17
CHAPTER 1. INTRODUCTION
studies, I do not nd a direct eect of the type of the political system
on developing countries environmental performance. The political system
appears to become important only in its interplay with the international
system in that it mediates the positive eect of international integration on
environmental quality.
Second, the ndings suggest that it is crucial not to limit international
integration to one aspect such as trade openness alone but to include vari-
ous aspects of the international system such that a more complete picture
of how globalization aects environmental quality arises. This is especially
important because, as the analysis shows, out of the four facets of interna-
tional integration analyzed in this thesis only IGO membership is always
associated to increases in environmental quality. These results also warn us
from drawing too optimistic a picture of foreign investment and trade ows:
whereas the study does not provide any evidence in favor of the pollution
haven argument, the positive perspective on trade and FDI as facilitators of
technology transfer does not seem to hold either. Similarly, with regard to
development aid the results suggest that the positive eects arising from the
promotion of sustainable development and the provision of environmentally
sound technology are outweighed by the aid induced economic growth.
Overall, however, the study also allows for some positive conclusions for
the study of public goods provision since IGO membership provides devel-
oping countries with a channel through which they can tackle their envi-
ronmental problems already at their current low national income. Hence
developing countries do not need to grow rich before they can combat envi-
ronmental pollution.
This thesis is structured as follows: The next chapter gives a detailed
overview of the various strands of literature important for the topics of inter-
national integration, the political system and the provision of environmental
quality. Chapter 3 develops the theoretical argument and outlines the hy-
potheses to be tested. Chapter 4 describes the data and research methods
used. Chapter 5 presents the results of the main empirical analysis. This is
followed in Chapter 6 by a detailed evaluation of the mechanisms underlying
IGO membership. Chapter 7 complements these ndings by a formal model
developed to generalize the argument of how IGO membership inuences en-
vironmental quality in developing countries. Finally, Chapter 8 summarizes
the ndings and discusses the theoretical and policy implications.
18
Chapter 2
Literature Review
As the introductory section has shown, this thesis ties in with several dif-
ferent strands of literature. A natural starting point for a literature review,
however, are the many studies dealing with national income and the provi-
sion of environmental quality. Building on these studies, I will then enlarge
the scope of the relevant literature by reviewing research that analyzes how
the political system inuences the provision of environmental quality. Sub-
sequently, I will further expand the scope by presenting studies that deal
with any of the four aspects of international integration analyzed in this
thesis IGO membership, international aid, trade and foreign direct invest-
ment. Finally, the literature is presented that connects these various forms
of international integration with a countrys political system.
2.1 Literature on environmental quality: the En-
vironmental Kuznets Curve
Starting with Shak & Bandyopadhyay (1992), Selden & Song (1994), Holtz-
Eakin & Selden (1995) and Grossman & Krueger (1995, 1993) a vibrant
literature has evolved that deals with the question of how national income
is related to a countrys environmental performance
1
. Most of these early
studies empirically found an inverted U-shaped relationship between growing
GDP per capita and several indicators of environmental pollution. Following
Kuznets (1955), who proposed an inverted U-shaped relationship between
the level of income and income inequality, this environmental version of the
curve was labeled the Environmental Kuznets Curve (EKC).
The logic underlying the EKC is that in the course of a countrys eco-
nomic development, rising national income increases the scale of economic
activity, which - all else being equal - leads to rising pollution levels (scale
eect). However, after a certain threshold of national income has been
1
If not indicated otherwise the studies under review follow a quantitative approach and
refer to all countries (developing and industrialized).
19
CHAPTER 2. LITERATURE REVIEW
reached, pollution is supposed to decline due to two eects: rst, the com-
position of the economy is likely to change from manufacturing to service
(composition eect). Second, with rising national income, technological
progress tends to lead to less environmental pollution (technology eect).
In addition, there is a political component to the EKC, which implies that
at early stages of economic development, environmental quality is generally
considered a luxury good and since states at this stage of economic devel-
opment have only limited resources available, environmental performance
usually ranks far behind the demand for better economic conditions, for
example. However, once people attain a certain level of living standards,
environmental quality turns into a normal public good and constituents de-
mand that their government take actions to reduce or avoid pollution, for
example by enacting appropriate environmental regulations (Selden & Song
1994).
The results of these early studies on the EKC somewhat qualied the
hitherto negative perspective on income growth and environmental quality:
Contrary to the alarmist cries of some environmental groups, we
nd no evidence that economic growth does unavoidable harm to
the natural habitat. Instead we nd that while increases in GDP
may be associated with worsening environmental conditions in
very poor countries, air and water quality appear to benet from
economic growth once some critical level of income has been
reached (Grossman & Krueger 1995, 370).
Relying on the development trajectory proposed by the EKC, some stud-
ies even went as far as to suggest that countries should simply grow rich and
then enact the appropriate policies to deal with environmental problems in-
stead of trying to counteract increasing pollution in the course of economic
development (Beckerman 1992). This recommendation, however, is not only
problematic as it absolves developing countries of their environmental re-
sponsibilities during early phases of economic development but also because
later and more sophisticated studies on the Environmental Kuznets Curve
qualify earlier results on both empirical and on theoretical grounds.
Starting with theoretical improvements and criticisms of the EKC, some
studies provide formal theoretical models to explain the particular shape
of the EKC. Lopez (1994) or Selden & Song (1994), for example, model
the relation between emissions and income using technology and consumer
preferences as the main parameters. In a more recent model by Andreoni
& Levinson (1998), the inverted U-shaped relationship between income and
pollution arises due to the technology that links consumption of a good to
its polluting byproduct.
In addition to these studies that are intended to improve the theoretical
underpinning of the EKC, some theoretical criticisms also exist (Stern 2004).
20
CHAPTER 2. LITERATURE REVIEW
First, the assumption underlying the EKC that economic growth will not be
aected by increasing environmental damage is challenged. Second, Stern
(2004) suggests that although some pollutants may decrease with income,
other pollutants will increase instead such that there will be a composition
change in pollution without a real overall reduction in emissions.
Turning to those studies that improve the research on the EKC empiri-
cally, Nordstroem & Vaughan (1999), for example, conclude that the EKC
only holds for a very specic set of pollutants such as local air pollutants
or some water pollutants but not for global air pollutants such as carbon
dioxide (CO
2
). In contrast, an early study by Holtz-Eakin & Selden (1995)
nds a diminishing marginal propensity of countries to emit CO
2
, which
would be well in accordance with an Environmental Kuznets Curve. This
is supported by Lamla (2006) who nds evidence using a Bayesian averag-
ing of classical estimates technique in favor of the EKC independent of the
pollution proxy.
As a further advancement on statistical grounds, Millimet et al. (2003)
propose a semi-parametric alternative to estimate the Environmental Kuznets
Curve. They nd that this semi-parametric specication yields even lower
turning points than the usual parametric approach, which consisted of us-
ing GDP per capita and its squared (and often its cubed) term. This would
imply that countries would reduce environmental pollution at lower levels
of national income than suggested previously. However, Stern (2004) argues
that none of Millimet et al. (2003)s estimates seems to t the data very
well, indicating a
fragility of the EKC rather than evidence for a low turning point
semi-parametric specication (Stern 2004, 1430).
This fragility of the EKC is further supported by Hettige et al. (2000):
using data on the water pollution intensity of dierent industry sectors, they
nd no evidence for an Environmental Kuznets Curve. Instead, they show
that pollution intensity increases until countries reach middle-income levels
whereupon pollution levels seem to stay constant. This is in line with Shen
(2006) and Plassmann & Khanna (2006) who also nd little evidence for the
inverted U-shaped relation between income and environmental degradation.
Finally, recent evidence from Perman & Stern (2003) suggest empirically
that the EKC does not exist and that if appropriate statistical techniques
and more inclusive samples are used, a probably more realistic picture of
the EKC arises: Although there may be an EKC for concentrations of pol-
lutants, emissions tend to rise monotonically with income (Stern 2004). This
even seems to be true for SO
2
emissions, a pollutant that has usually been
used as evidence in favor of the EKC. However, time-related factors that
are unrelated to income seem to reduce pollution in countries at all levels
of income (Stern 2004). This is in line with Dasgupta et al. (2002) who
21
CHAPTER 2. LITERATURE REVIEW
describe evidence that even in developing countries environmental problems
are addressed and remedied.
Taken together, all of these later studies on the EKC show that on empir-
ical as well as on theoretical grounds this inverted U-shape between income
and pollution seems to be highly questionable. It thus becomes evident from
the studies focusing merely on the question of how income inuences pollu-
tion that important facets of environmental quality provision are missing in
the literature on the Environmental Kuznets Curve, leading to these rather
questionable results. More recent studies therefore attempt to include fac-
tors that seem to be crucial for environmental performance but which have
been neglected in earlier studies.
2.2 Enlarging the Scope Step 1: Environmental
Quality and the Political System
A number of studies exist that are linked to the topic of democracy or
democratic structures and the environment. Most of the empirical liter-
ature shows that in general, democracies tend to be better providers of
environmental quality, although the eect size seems to vary strongly with
the particular indicator used to proxy environmental quality (Barrett 2000;
Bernauer & Koubi 2009; Deacon 1999, 2003; Torras & Boyce 1998).
The general logic underlying most of these studies is derived from the
literature dealing more broadly with regime type and the provision of public
goods. In this literature, it is argued that democracies provide more public
goods to their citizens than autocracies because in order to survive in of-
ce, democratic leaders need the support of the majority of their citizenry
(Bueno De Mesquita et al. 2003; Deacon 1999, 2003; McGuire & Olson 1996;
Midlarsky 1998; Olson 1993).
In this context, most authors rely on Bueno De Mesquita et al. (2003)s
selectorate theory to analyze the inuence of the political system on the
provision of public goods (Bernauer & Koubi 2009; Smith 2006). Following
the terminology of Bueno De Mesquita et al. (2003), we can dene the
selectorate as the group in a society that is eligible to select or elect the leader
of the country. The winning coalition is then dened as the subgroup of the
selectorate whose support is necessary for the leader to gain or stay in power.
In a democracy with a majoritarian electoral system, the winning coalition
would be at least 50% of the electorate, whereas in an autocracy this would
be a small elite group usually consisting of the military or party elites.
Applying Bueno De Mesquita et al. (2003)s terminology we can deduce that
leaders in autocracies need to satisfy a smaller winning coalition in order to
stay in power than leaders in democratic countries. Since leaders have only
limited resources at their disposal to attain political support, it follows that
autocratic leaders can concentrate these resources on their small winning
22
CHAPTER 2. LITERATURE REVIEW
coalition whereas leaders in democracies need to disperse their resources to a
larger group. To maximize political support given their resources, it is more
ecient for democratic leaders to provide public goods than to buy o their
winning coalition with targeted private goods. Consequently, democratic
leaders should provide more public goods to ensure the support of a larger
group, whereas in autocracies leaders should mostly rely on private goods
targeted to their small winning coalition.
In contrast to the reasoning of Bueno De Mesquita et al. (2003), it can
be argued that autocracies might supply more public goods as their leaders
usually have a longer planning horizon
2
. As most environmental problems
develop slowly and often only become apparent in the distant future, demo-
cratic leaders who can be seen as being more myopic due to more frequent
elections might not be interested in facing the short-term costs of provid-
ing long-term environmental quality (Congleton 1992; Midlarsky 1998). In
addition, some studies claim that in mature democracies public goods are
underprovided because special interest groups have gained a disproportion-
ate inuence on the government. Therefore, democratic governments often
provide private goods to these interest groups instead of providing public
goods to the whole population (Congleton 1992; Olson 1982).
However, as mentioned above, most of the empirical literature shows that
democracies tend to be better providers of environmental quality (Barrett
2000; Bernauer & Koubi 2009; Deacon 1999, 2003; Torras & Boyce 1998). In
his studies Deacon (1999, 2003), for example, shows that non-democracies
seem to under-provide public goods such as education, infrastructure and
environmental quality. These results are supported by Bernauer & Koubi
(2009) who nd that democratic structures are conducive to local air quality
as measured by SO
2
. Using a wide range of environmental indicators Li
& Reuveny (2006) nd that in general, democracy reduces environmental
degradation; however, the strength of the eect varies across indicators.
Gassebner et al. (2006) employ extreme bounds analysis to determine which
factors are robustly related to environmental quality (measured by SO
2
, CO
2
and BOD). Except for a dictatorship dummy, which is negatively related to
CO
2
emissions, none of the political variables exerts a consistent inuence.
Ward (2008) analyzes the inuence of democracy on sustainability (mea-
sured by a countrys national footprint and genuine savings, which is a mea-
sure of weak sustainability) while controlling for public opinion (world value
survey). He nds that stable core autocracies perform worse with regard
to strong sustainability compared to stable core democracies. Furthermore,
liberal democracies only seem to score higher on the indicator measuring
weak sustainability if the party system is not too open. Surprisingly, public
opinion on the importance of environmental issues does not seem to matter
2
As Bueno De Mesquita et al. (2003) note: once autocratic leaders survive the rst
year in oce they usually stay in oce for a long period.
23
CHAPTER 2. LITERATURE REVIEW
much.
Fredriksson & Wollscheid (2007) nd that democracies are associated
with stricter environmental policy measures than autocratic countries. How-
ever, the positive impact of democracies seems to be due to parliamentary
democracies. Presidential-congressional democracies, in contrast, do not
seem to set signicantly stricter environmental policies than autocracies
where environmental policies are measured by cross-sectional variables from
the Environmental Sustainability Index (Center for International Earth Sci-
ence Information Network (CIESIN) 2002) and prices of super and diesel
gasoline.
In their analysis of the Montreal Protocol, Murdoch & Sandler (1997)
nd that both democratic and more wealthy countries are characterized by
larger reductions in chlorouorocarbon (CFC) emissions, which is an ozone-
depleting substance. Finally, Torras & Boyce (1998) show that for certain
pollutants such as sulfur dioxide and smoke, higher levels of political rights
and civil liberties are associated with better environmental quality, especially
in low-income countries.
In contrast to the supporting evidence on the democracy and environ-
mental quality nexus, Baettig & Bernauer (2009) nd that in the context
of the Kyoto Protocol process, democracy has only a weak eect on green-
house gas emissions. Similarly, in their cross-sectional analysis, Fredriksson
& Svensson (2003) nd no eect of democracies displaying higher environ-
mental stringency.
In addition to distinguishing between democratic versus autocratic struc-
tures, some studies delve further into the specic aspects of democracies that
seem to be related to the provision of environmental quality. To this end,
these studies focus on very stable democracies, resulting in a sample that
is restricted to OECD countries only. By constructing a composite index of
air, water and soil pollution as well as waste management Jahn (1998) nds
that strong neo-corporatist arrangements and strong (oppositional) social
democratic parties are correlated with better environmental performance.
Furthermore, Jahn & Walti (2007) nd that federally and centrally gov-
erned OECD countries perform equally well with regard to environmental
performance because in federal systems corporate actors play a crucial role
compensating for the lack of strong central governing structures. Finally,
Jahn (2008) shows that government orientation regarding green values as
well as corporatism are associated with a reduction in CO
2
emissions.
Altogether, the literature on regime type and the provision of environ-
mental quality counts more studies in favor of a positive eect of democ-
racies. However, as the studies of Fredriksson & Svensson (2003); Baettig
& Bernauer (2009) and Ward (2008) show, the general result may not hold
for all specic settings. Thus, one needs to carefully adjust the theoretical
arguments underlying the connection between a countrys political system
and its environmental performance to the specic context of each study.
24
CHAPTER 2. LITERATURE REVIEW
Consequently, the theoretical section will apply the general argumentation
of the literature on regime type and public good provision to the specic
setting in developing countries, which lies at the heart of this study.
2.3 Enlarging the Scope Step 2: Environmental
Quality and Globalization
Not only do more recent studies incorporate various aspects of the polit-
ical system in their analysis, they also include non-domestic factors that
mostly fall under the label of globalization. The general inuence of glob-
alization or international integration on environmental quality is strongly
debated. On the one hand, several positive consequences of international
integration with regard to the provision of environmental quality have been
posed. Firstly, globalization implies a ow of information and knowledge
on environmental problems, their consequences and how to abate pollution.
Consequently, people could become aware of environmental quality as a pub-
lic good that needs to be protected (Ward 2006). Secondly, due to increasing
trade and foreign direct investment, recent technologies might spread even
to less developed countries. Additionally, some authors argue that trade
and FDI in general produce welfare enhancing gains, which might move
a country upward on the Environmental Kuznets Curve (Frankel & Rose
2005). Furthermore, international actors, like international organizations,
non-governmental organizations (NGOs) or individual countries, might pro-
vide low- and middle-income countries with knowledge, technologies and
development aid, which could enable those countries to provide better en-
vironmental quality. Finally, many international treaties and organizations
oblige their member countries to cooperate on a number of environmental
problems (Neumayer 2002b; Ward 2006).
On the other hand, globalization has been blamed for various negative
consequences with regard to environmental quality, most prominently in
the literature on pollution havens (Jae et al. 1995; Esty & Giradin 1998;
Staord 2000). The gist of the pollution haven argument is that indus-
trialized countries are only able to become cleaner by shifting their dirty
industries to less developed countries, which usually have less stringent en-
vironmental regulations. Some authors go even further and argue that for
fear of losing industries to countries with less stringent regulations, gov-
ernments will lower their environmental standards in order to prevent this
shift from taking place (Revesz 1992; Drezner 2001). This argumentation is
accordingly referred to as the race to the bottom hypothesis.
Taken together, it can be concluded that the overall impact of glob-
alization on environmental quality is strongly debated. The next sections
therefore review the relevant literature dealing consecutively with each of
the four aspects of international integration membership in international
25
CHAPTER 2. LITERATURE REVIEW
organizations, development aid, trade and FDI ows.
2.3.1 International Governmental Organizations
When refering to any kind of cooperation between states, international re-
lations theory usually uses the label regime. Perhaps the most prominent
denition of international regimes describes them as
sets of implicit or explicit principles, norms, rules, and decision-
making procedures around which actors expectations converge
in a given area of international relations (Krasner 1983, 3).
However, this denition causes several diculties for empirical research
as it covers a wide range of international cooperation including non-explicit
or non-formalized agreements, which are dicult to identify and measure
(Tooze 1990; Porter et al. 2000). Consequently, scholars in later research
have tried to narrow down the term regime as
institutions with explicit rules, agreed upon by governments,
that pertain to particular sets of issues in international relations
(Keohane 1989, 4)
or
social institutions consisting of agreed-upon principles, norms,
rules, procedures, and programs that govern the interactions of
actors in specic issue areas (Levy et al. 1995, 274).
Although these denitions are more concrete compared to the original de-
nition by Krasner, they still leave room for many dierent forms of interna-
tional cooperation such as bilateral versus multilateral treaties or short-term
agreements versus long-term international organizations. In order to be able
to analyze the eect of international institutions, it is therefore necessary to
distinguish between the dierent forms of cooperation and to specify what
type of cooperation one is analyzing.
In this thesis I therefore focus on international governmental organiza-
tions since they are the most formalized and long-term forms of international
political cooperation. These two features make IGOs a likely regime type to
inuence public good provision in developing countries as these organizations
possess at least some organizational capacity to interact with government
and society representatives in developing countries.
According to Pevehouse et al. (2004), IGOs are organizations that
1. consist of at least three members of the COW-dened state system
2. hold regular plenary sessions at least once every ten years
3. possess a permanent secretariat and corresponding headquarters
26
CHAPTER 2. LITERATURE REVIEW
4. and which are not emanations of already existing organizations.
The general purpose of international governmental organizations is to enable
or facilitate cooperation at the international level. Hence IGOs should allow
countries to realize benets from mutual cooperation by reducing transac-
tion costs, by limiting uncertainty, by widening the shadow of the future, by
increasing reputational costs and by allowing credible commitments (Keo-
hane 1984; Abbott & Snidal 2000).
More precisely, according to Keohane (1984) and Krasner (1983), IGOs
facilitate collective action by reducing transaction costs that would otherwise
deter cooperation eorts. Transaction costs may result from bargaining and
enforcement eorts that arise since incomplete information does not permit
states to get a clear picture of other states interest and actions (Coase 1960;
Milgrom & Roberts 1992; Williamson 1985). IGOs can solve these problems
by providing information, making rules explicit and therefore making devi-
ations recognizable
3
. One example of an institution that worked in this way
was the informal GATT (General Agreement on Taris and Trade) regime.
Although it had no enforcement powers before it was transformed into the
World Trade Organization (WTO) in 1995, GATT claried the expecta-
tions of the actors involved and thus facilitated international cooperation
(Reinhardt 2001; Stone et al. 2008).
Furthermore, IGOs enlarge the scope of cooperation by allowing for the
possibility of issue linkage (Martin 1993, 1995) and by increasing the reputa-
tion costs of non-compliance with agreements (Simmons 2000)
4
. Moreover,
Moravcsik (2000) points out that international organizations allow states
to make credible commitments, thereby enhancing cooperation. A further
mechanism by which regimes can inuence states is that they facilitate learn-
ing (Young & Levy 1999).
However, not all scholars in international relations accept these argu-
ments of how IGOs can inuence state behavior. Realists, in contrast, argue
that
3
Similarly, Keohane (1982) expresses these arguments in a framework of supply and
demand. In particular, he argues that one can interpret the theory of hegemonic stability
as reecting the supply side of international organizations, implying that if a hegemon is
present, the supply of an IGO is possible. In contrast, he sees his explanation as situated
on the demand side of international organizations. Since the international system often
displays what he calls market failures (uncertainty and transaction costs), states acting
unilaterally forgo benets that they could obtain by acting in a coordinated way. Regimes
are therefore a way to circumvent uncertainty and transaction costs and thus to benet
from mutual cooperation.
4
Simmons (2000) argues that this reputation eect is more pronounced for countries
with a strong tradition in respecting the rule of law. Similarly, Manseld et al. (2002)
argue that these reputation costs are higher for elected leaders, hence they play a greater
role in democracies.
27
CHAPTER 2. LITERATURE REVIEW
IGOs reect, rather than eect, world politics (Boehmer et al.
2004, 1).
According to this view, although countries may create international
regimes, these regimes do not have an independent eect on countries course
of action as these IGOs merely reect existing interests. Hence countries will
only comply with IGOs if it is in their direct short-term interest. This per-
spective is of course in strong contrast to how constructivists, functionalists
or liberal institutionalists conceive of IGOs as a means of allowing and fos-
tering international cooperation that would hardly take place without the
context of international organizations.
Today, most international relations scholars, however, agree that
global politics is increasingly organized around regimes and in-
stitutions that foster cooperation by providing information and
organization structure, promoting norms and common belief sys-
tems, and reducing transaction costs (Boehmer et al. 2004, 3)
5
.
Boehmer et al. (2004), for example, show that joint membership in struc-
tured IGOs, i.e. regimes that are characterized by well dened and struc-
tured organizations, signicantly reduces the likelihood of conict onset.
Using social network analysis, Dorussen & Ward (2008, 190) show in a sim-
ilar study that direct as well as indirect IGO membership ties reduce the
likelihood of interstate conict as IGO membership allows the transmission
of information on countries interests and intentions. They further argue,
as do Russett & Oneal (2001), that IGOs not only allow for information
exchange but also promote cooperation as country ocials might come to
appreciate other points of view.
Ingram et al. (2005) argue that IGOs reduce transaction costs because
they reduce uncertainty, allow states to commit to future actions, and pro-
mote rules such as fair exchange. Building on these arguments, they claim
that joint membership in IGOs increases bilateral trade. They make the ar-
gument that even organizations with only a social and cultural purpose such
as the World Health Organization enhance bilateral trade. The reasoning
behind this is that social and cultural IGOs
(1) [...] increase awareness, sympathy, empathy, and even trust
between the citizens of dierent countries: and (2) the resulting
5
However, this does not mean that realist objections are not embedded within newer
theories on the functions of international organizations. Neoliberal institutionalists, for
example, although they accept most of the functionalist perspective on IGOs, combine
these with the realist axioms of systemic anarchy, the importance of power, and the
preeminence of states interests. In this view, IGOs may mainly promote cooperation by
strengthening reciprocity through regularizing interactions (Boehmer et al. 2004; Mitchell
& Deane 2008).
28
CHAPTER 2. LITERATURE REVIEW
shift in cross-national relations and perceptions results in more
trade (Ingram et al. 2005, 830).
In addition, similar to Boehmer et al. (2004) Ingram et al. (2005) argue
that the eect on trade will be stronger for more structured IGOs because,
through their organized bureaucracy, structured IGOs are better able to
enforce rules, monitor states and allow for a more detailed exchange between
member states. Their results indeed show that IGOs in general, as well as
economic, social and cultural IGOs in particular, increase bilateral trade.
As expected, the eect is more pronounced for structured IGOs than for
minimally institutionalized organizations.
In contrast to these numerous studies showing a strong eect of IGOs,
Stein (2005) nds that international treaties have little constraining power.
She argues that those factors that lead states to enter into internation-
ally binding agreements are later decisive for their compliance, which would
suggest that treaties seem to have merely screening purposes rather than
constraining power.
Overall, however, this multiplicity of studies shows that international
governmental organizations do indeed appear to foster cooperation between
states in various issue areas such as trade and peace-building. In addition
to these studies, there exists a more recent strand of literature that analyzes
the domestic outcomes of IGO membership. Whereas the previous strand of
literature focuses on the many ways in which IGOs can foster international
cooperation, this new strand of literature highlights that IGO membership
can also inuence domestic outcomes.
Pevehouse (2002b) and Manseld & Pevehouse (2008), for example, show
that membership in international organizations is a way for governments to
signal their willingness to stick to a particular course of action. They nd
that countries undergoing a democratization process are more likely to join
standards-based (such as human rights or environmental organizations) and
economic IGOs because they allow governmental leaders to signal to their
domestic and international audience that they are willing to continue polit-
ical and economic reform. In contrast to political or universal organizations
such as the UN, standards-based and economic organizations demand that
countries implement specic policies or standards, making membership in
those organizations a more credible signal of democratization. Pevehouse
(2002a) goes even further by showing that membership in specic regional
IGOs signicantly increases the likelihood that a political regime will change
from autocratic to democratic.
The more recent literature on policy diusion
6
also shows that IGO mem-
bership aects domestic policy choices since connections in the network of
6
A commonly accepted denition of diusion comes from Strang (1991, 325), who states
that diusion happens when the prior adoption of a trait or practice in a population alters
the probability of adoption for remaining non-adopters.
29
CHAPTER 2. LITERATURE REVIEW
international organizations can provide countries with information that fos-
ters policy learning, thereby providing a socialization process among states
(Cao forthcoming; Hafner-Burton et al. 2008; Simmons & Elkins 2004; Frank
et al. 2000). This is underlined by Dobbin et al. (2007, 462) who state that
international organizations, either as agents or as sets of rules
that enhance transparency, appear to have had important eects
on information ows and policy transmission.
Finally, studies analyzing how IGOs aect a countrys environmental
performance are very rare. One notable exception is Ward (2006) who ex-
amines whether countries that are more central to the network of inter-
national environmental regimes act more sustainably at the national level.
Using social network analysis, he nds that indeed countries that are more
central also care more about domestic environmental quality. Although his
study concentrates on the network of international environmental treaties
and organizations he notes that
[. . . ] the Kantian view is that IGOs, economic interdependence
and democracy form a mutually supportive triangle that pro-
motes peace. The network of IGOs facilitates deterrence of bad
behaviour, mediation and problem-solving, sharing of informa-
tion and the generation of norms and trust (Russett & Oneal
2001). This raises the possibility that nations environmental
records may relate to their general position in the international
system, just as recent work suggests that joint membership of
non-trade related IGOs increases trade between pairs of nations
(Ingram et al. 2005) Ward (2006, 154).
Consequently, this argumentation indicates that there are good reasons
to analyze the impact of IGO membership in general on a countrys willing-
ness and capability to take care of its environment.
To summarize the literature on IGO membership, one can conclude that
most studies show that IGOs not only foster cooperation between states but
also inuence domestic outcomes such as regime changes or policy learning.
It is even suggested that IGOs can be an important actor in shaping the
capacity of their member countries to improve their environmental records
(Ward 2006). Consequently, this study argues that IGO membership can
play a crucial role in inuencing a countrys environmental performance
while taking both other forms of globalization as well as important domestic
factors such as the political system of a country and its national income into
account.
2.3.2 International Aid
Similar to the literature dealing with the eects of IGOs, there are many
studies analyzing the eectiveness of international aid as well as the rea-
30
CHAPTER 2. LITERATURE REVIEW
soning behind aid allocation decisions. As in the case of IGO membership
studies linking international aid to environmental quality are rare. This
section therefore begins with a short review of the literature dealing with
the eectiveness and the reasons behind aid allocation in general and then
discusses the literature that ties aid to environmental degradation.
One of the earlier studies dealing with the eectiveness of aid is Boone
(1996), in which he analyzes whether aid contributes to the alleviation of
poverty. His empirical results, however, show that aid does not seem to
benet the poor but that it is mostly proting a small political elite and
increasing the size of government. Moreover, he shows that the impact of
aid is not conditional on the political system of a country, implying that
democratic governments are not more eective than autocratic governments
in using aid ows. Based on these ndings, he conclude that aid might only
be eective if it is conditional on policy change such that the political elite
cannot easily use aid ows for its own enrichment.
One reason why Boone (1996) nds aid to be ineective in alleviating
poverty might be due to the motives of aid allocation. Many studies show
that in allocating bilateral aid, donors seem to pay more attention to their
own interests and strategic aims (such as promoting exports, supporting the
implantation of donor rms in recipient economies, and maintaining access
to essential imports) than to the characteristics of the recipient countries
themselves (for example the level of democracy, the character of economic
policies, and good governance practices) (Alesina & Dollar 2000; Alesina &
Weder 2002; Lundsgaarde et al. 2006; Maizels & Nissanke 1984; McKinlay
& Little 1977).
This nding that aid allocation is mostly driven by donor interests seems
to hold mainly for bilateral aid. Multilateral aid, in contrast, seems to
be driven more by recipient needs (Maizels & Nissanke 1984). Neumayer
(2003c), for example, shows that economic needs of recipient countries are
an important selection criterion for most multilateral donor institutions, es-
pecially the UN agencies. Furthermore, strategic concerns such as higher
military expenditures, which were important for bilateral aid allocation, do
not seem to be crucial for multilateral aid allocation. Looking at aggre-
gate multilateral aid Neumayer (2003c) shows, furthermore, that political
freedom is signicant in explaining aid allocation whereas personal integrity
rights are not.
The nding that political rights are only important for aid allocation if
multilateral in contrast to bilateral aid is considered is also reinforced by
a study by Zanger (2000) who shows that good governance criteria do not
play any signicant role for the allocation of European bilateral aid.
In a similar study, Neumayer (2003b) dierentiates between factors that
are crucial for aid eligibility and factors that explain the amount of aid
received by a country. His results show that at the aid eligibility stage,
which can be seen as a gate-keeping stage, respect for civil and political rights
31
CHAPTER 2. LITERATURE REVIEW
plays a statistically signicant role for most donors. However, looking at the
determinants of the level of aid, he shows that most donors provide more aid
to countries that have a poor record on either civil and political or personal
integrity rights. Based on these results, he concludes that contrary to their
commitment to human rights and good governance, donor countries do not
reward respect for these rights in their foreign aid allocation (Neumayer
2003b, 650).
Overall, these studies on bilateral and multilateral aid allocation seem
to suggest that the main driving forces behind development aid are donor
economic interests instead of recipient need interest, which would explain
to some extent why Boone (1996), for example does not nd that develop-
ment aid signicantly reduces poverty. However, some studies show that
international aid inuences certain domestic outcomes such as the adoption
of liberal economic reforms. Similar to the studies that investigate how
IGO membership can serve to diuse policy choices and policy outcomes
as described above, there are some studies which argue that policy change
can be brought about by the use of nancial channels such as trade ties or
conditional lending.
Henisz et al. (2004), for example, nd that multilateral lending coerces
states to adopt more liberal reforms in the telecommunications and the
electricity sector. The logic underlying the conditional lending argument
is similar to the reasoning of how the International Monetary Fund (IMF)
can inuence policy change in developing countries. In this context, Vree-
land (2003) argues that governments often use the conditionality imposed
on them by the IMF in order to implement policies that they are unable to
implement independently. Hence blaming an international lending institu-
tion such as the IMF or the World Bank allows governments to introduce
important policy changes without risking electoral punishment.
Studies connecting foreign aid with environmental performance in de-
veloping countries are rather rare. Gutner (2005), for example, argues that
the World Bank has undergone a so-called greening reform since the mid
1980s. Although environmental projects are still a minority of projects that
the World Bank nances, Nielson & Tierney (2003) nd evidence for the fact
that an environmental reform of the Banks lending decisions occurred in
the early 1990s with the establishment of the Global Environment Facility
(GEF).
Using a principal-agent framework, McLean (2006) investigates the strate-
gic interactions between the GEF
7
, as the aid donor, on the one side and
7
The Global Environment Facility was created in 1991 by Western governments and
was initially attached to the World Bank, later becoming nominally independent of the
Bank (Young 2002). Its purpose is to address global environmental issues while sup-
porting national sustainable development initiatives. The GEF is the designated nancial
mechanism for a number of multilateral environmental agreements (MEAs) or conventions
[...] such as the Convention on Biological Diversity (CBD), the United Nations Framework
32
CHAPTER 2. LITERATURE REVIEW
governments and NGOs, as aid recipients, on the other. The author shows
that the eectiveness of environmental aid is strongly dependent on the
strategic interaction between donor and recipient characterized by factors
such as
conicting donor and recipient motivations, the donors uncer-
tainty about its recipients commitment to project objectives, the
donors inability to make a credible commitment to stop aid dis-
bursements to inuential non-complying recipients, and tensions
between various policy goals that national governments pursue
(McLean 2006, 164).
The ndings of McLean (2006) are in line with the conclusions of an
older study by Keohane & Levy (1996) who claim that the prospects for
successful environmental aid are rather small. Factors that seem to make
eective use of environmental aid so dicult include the divergent interests
between donor and recipient countries and the diculty of building lasting
capacities in recipient countries such that environmental protection eorts
continue once international aid ceases to ow. In addition, it is argued that
most of the time, aid ows are seen as short-term inputs whereas most en-
vironmental problems are long-term in nature, which is another reason why
environmental aid is often rather ineective. However, Keohane & Levy
(1996) also underline positive impacts of environmental aid, for example
the possibility of attracting more attention towards environmental protec-
tion such that more eort is devoted to solving environmental problems in
recipient countries.
To sum up, the literature on the eects of aid does not give rise to
very optimistic conclusions. Rather, donor countries seem to be awarding
aid according to reasons that relate to their own economic and strategic
interests, resulting in fairly low eectiveness of development aid. Although
this conclusion must be qualied to some extent for multilateral aid since
recipient interests seem to be of greater importance for multilateral aid, the
eects of aid on domestic outcomes are still questionable as the studies of
McLean (2006) and Keohane & Levy (1996) show.
2.3.3 Economic Integration: Trade and Foreign Direct In-
vestment
The literature dealing with the relationship between economic globalization
(trade openness and foreign direct investment) and the provision of public
Convention on Climate Change (UNFCCC), the Stockholm Convention on Persistent Or-
ganic Pollutants (POPs), and the UN Convention to Combat Desertication (UNCCD)].
The GEF is also associated with many global and regional MEAs that deal with interna-
tional waters or transboundary water systems. The GEF is not a nancial mechanism for
the Montreal Protocol on Ozone-Depleting Substances; however, its activities complement
and enhance the work of the Montreal Protocol (GEF 2008a).
33
CHAPTER 2. LITERATURE REVIEW
goods revolves mostly around two issues: general welfare enhancing gains
and regulatory competition (Frankel & Rose 2005). As arguments are more
or less similar for both trade and foreign direct investment, I will provide a
short summary of the arguments before looking in more detail at the specic
literature on trade and FDI and their eect on environmental quality. This
similarity in the theoretical underpinning of the eects of both FDI and trade
on public goods provision is also an important reason why this thesis includes
both trade and FDI in its analysis of a countrys economic integration. Or
as Zeng & Eastin (2007, 991) argue,
[...] without measuring FDI and trade openness, the entire
eect of global economic integration cannot be captured.
8
Following the rst argument of how economic globalization aects the
provision of public goods, increasing economic integration is considered to
have, on average, a positive eect through the welfare enhancing gains of
economic globalization. In particular, it is argued that increasing economic
integration allows countries to specialize in those industries in which they
have a comparative advantage and hence allows for a more ecient alloca-
tion of resources. Due to the resulting welfare enhancing gains, a countrys
national income is supposed to increase (Frankel & Rose 2005). In the long
run, it can be argued that these increases in general welfare lead to a rise
in the demand for public good provision as the demand for most public
goods such as better environmental quality, better education or health care
is supposed to increase with growing income (Deacon 2003).
The second argument concerning economic integration and the provi-
sion of public goods focuses on regulatory mechanisms. It is argued that
due to increasing economic interdependencies, countries face strong regula-
tory competition. In order to attract foreign capital, companies or products,
countries are pushed towards less stringent taxation or environmental regu-
lation, for example. This should either result in a regulatory chill or even in
a regulatory race to the bottom, which in the end would lead to a decrease
in the provision of public goods (Zarsky 1999)
9
. Which of these two eects,
8
This is supported by Araya (2002) who argues that investment and trade are not
independent of each other and by Guerin & Manzocchi (2006) who nd that trade and
FDI are complements: countries that trade a lot, also share a lot of investment. This
point is reinforced by the study of B uthe & Milner (2008) who show that both WTO
membership and participation in preferential trade agreements imply a boost in foreign
investment.
9
These arguments underlying the regulatory race to the bottom assertion closely re-
semble the logic behind the eciency hypothesis with regard to welfare spending. This
eciency hypothesis states that in order to stay competitive in an economically more inte-
grated world, countries cut their public spending and thus provide a less generous welfare
state system (Garrett 2001; Jahn 2006). It is in strong contrast with what Garrett (2001)
or Rodrik (1998) call the compensation hypothesis. Here, it is argued that instead of
cutting back, governments actually increase their public spending in order to compensate
34
CHAPTER 2. LITERATURE REVIEW
general welfare enhancing gains or regulatory competition, will in the end
prevail is, according to Frankel (2003), mainly an empirical question.
2.3.3.1 Trade
Similar to the decomposition of income growth in the literature on the En-
vironmental Kuznets Curve, the eect of trade on pollution can be divided
into a scale, technology and composition eect (Antweiler et al. 2001; Bar-
rett 2000; Copeland & Taylor 1994; Frankel & Rose 2005; Frankel 2003).
The scale eect posits that due to expanded economic activity triggered by
an increase in trade, pollution will rise, all else being equal (Grossman &
Krueger 1995; Esty 2001; Cole 2004). However, trade also allows access to
better technologies and management practices such as catalytic converters
for cars or scrubbers on smokestacks, which would imply that this technology
eect yields a positive impact on environmental quality (Esty 2001).
Furthermore, growing trade is also supposed to lead to a more ecient
allocation of resources; this means that industries should settle where they
have a comparative advantage, which is labeled the composition eect of
trade. Whether this change in the allocation or composition of the industry
will bring about an increase or decrease in pollution depends on whether
a particular country has a comparative advantage in a more or less pollut-
ing industry (Cole 2004). Following standard trade theory, Hecksher-Ohlin,
countries should specialize in their abundant factor. For developing coun-
tries, this would imply a specialization in labor intensive industries. Since
polluting industries such as chemical, pulp and paper or oil rening indus-
tries tend to be more capital intensive, this would imply that they should
stay located in industrialized countries (Antweiler et al. 2001; Nordstroem
& Vaughan 1999). Consequently, standard trade theory would predict that
an increase in trade would lead to a reduction in pollution for developing
countries since their comparative advantage lies in labor intensive and thus
less pollution intensive production.
Against this argument, the proponents of the pollution haven hypoth-
esis argue that because of their usually rather low levels of environmental
regulation developing countries have a comparative advantage in pollution
intensive production. Therefore, due to the composition or allocation eect
of trade, dirty industries will settle in those countries which have the laxest
environmental regulations (Jae et al. 1995; Esty & Giradin 1998; Staord
2000). Consequently, it is argued that the eect of trade on pollution might
only be positive for countries with stricter regulations, which will most likely
be high-income countries. Some authors even claim that the threat of dirty
their people for the perceived risks arising from globalization. In addition, there is also
a convergence hypothesis stating that due to globalization, national policies will become
more alike, albeit without specifying whether this will be at high or low policy levels (Jahn
2006).
35
CHAPTER 2. LITERATURE REVIEW
industries settling in countries with lax environmental regulations will lead
to a competition between countries to oer those industries a better produc-
tion environment and therefore to a regulatory race to the bottom (Revesz
1992; Drezner 2001).
Empirical studies nd little evidence for the pollution haven hypothesis
and even less for the race to the bottom assertion (Nordstroem & Vaughan
1999; Frankel 2003)
10
. Rather, it seems to be the case that due to technology
diusion and the overall welfare gains of growing trade openness
11
, trade has
depending on the particular situation of the individual country a positive
eect on environmental quality on average (Naughton 2006).
As most studies on the pollution haven argument also focus on FDI, a
more detailed overview of the numerous studies on this topic is provided
in the following section dealing with foreign direct investment. Only those
studies focusing solely on the impact of trade on environmental quality are
discussed here.
In their study Botcheva-Andonova et al. (2006) analyze whether environ-
mental policies in post-Communist countries are inuenced by their trade
regimes. They measure environmental governance, their dependent variable,
by the national environmental funds spent each year divided by a countrys
population. Their evidence suggests that increasing trade openness is con-
nected with lower environmental regulation, lending support to a pollution
haven or race to the bottom argument. In contrast, Zeng & Eastin (2007) ar-
gue that in order to retain access to foreign markets in developed countries,
exporting rms in developing countries often fulll rather strict environ-
mental standards. In addition, fearing a potential dispute under the WTO,
those rms tend to comply ex-ante with strict environmental standards.
Concerning water quality Sigman (2004) shows that countries with higher
levels of bilateral trade are characterized by lower levels of river water pollu-
tion. This is in contrast to Bernauer & Kuhn (forthcoming), who nd that
trade has either a negative or insignicant eect on water pollution.
Eliste & Fredriksson (2002) theoretically show that pollution intensive
producers are often compensated for the introduction of pollution taxes by
governments mitigating the eect of the pollution tax on output. This also
has an impact on a countrys import and export composition since the intro-
duction of pollution taxes consequently does not reduce pollution intensive
exports as predicted by usual pollution haven arguments. They empirically
show that the introduction of more stringent environmental standards is
connected to an increase of transfers to the agricultural sector.
Finally, using social network analysis to measure trade competition, Cao
& Prakash (forthcoming) show that trade competition between countries
10
One exception is the study of Cole (2004) who nds some evidence in favor of trade
pollution havens; however, they appear to be rather small and not widespread.
11
Depending on a countrys location on the EKC an increase in trade, which theoretically
implies a rise in income, should result in a higher demand for environmental regulation.
36
CHAPTER 2. LITERATURE REVIEW
that export similar products to the same countries lead to downward races
with respect to air pollution as measured by SO
2
emissions. This nding
thus provides some of the rare evidence in favor of a race to the bottom with
regard to environmental indicators between countries that are competing for
the same export markets.
2.3.3.2 Foreign Direct Investment
Foreign direct investment ows have dramatically increased since the be-
ginning of the 1990s (UNCTAD 2009). This is mostly due to the fact that
in the early 1990s, many developing countries adopted neo-liberal policies
advocated by the IMF and the World Bank, which implied large-scale trade
and investment liberalizations (Zarsky 1999). This rise in foreign direct
investment has led to great debate about the environmental (and social)
eects of FDI. On the one side of this debate are mostly environmental-
ists arguing that the inow of foreign direct investment will lead to pollu-
tion havens in developing countries and might even bring about a race to
the bottom in world-wide environmental standards. On the other side are
mostly economists claiming that due to technological transfer FDI might
even induce so-called pollution halos.
In order to structure the arguments relating foreign direct investment
and environmental quality, it is useful to distinguish between arguments
and studies that center on the question of whether environmental regula-
tions are the driving force behind the competition for FDI, i.e. behind
the decision of rms location to developing countries, and arguments and
studies that focus on the question of what eects foreign ownership has on
environmental quality in developing countries (Zarsky 1999; Araya 2002).
Since the second strand of literature dealing with the eects of foreign own-
ership on environmental quality is more relevant for the arguments of this
thesis, I shall keep the literature review on the reasons behind rms relo-
cation decisions comparatively short.The presentation of the literature in
the subsequent paragraphs will follow this dierentiation between the com-
petition for FDI, i.e. rms relocation decisions, and the eects of FDI on
environmental quality.
2.3.3.2.1 Competition for FDI As introduced above, the gist of the
pollution haven assertion is that low environmental standards in developing
countries will attract dirty industries away from high regulated industrialized
countries towards low regulated developing countries. Furthermore, it is
argued that even if rms do not explicitly move for reasons connected to
the avoidance of environmental regulation, they will most likely perform
badly if left to operate uncontrolled in developing countries (Zarsky 1999).
Moreover, it is suggested that in order to remain competitive, industrialized
countries will in response also cut their environmental regulations, which will
37
CHAPTER 2. LITERATURE REVIEW
lead to a global race to the bottom with regard to environmental standards.
In contrast, most economist argue that along with foreign direct invest-
ment, newer and cleaner technologies are spread since rms from industrial-
ized countries possess better technology (Guerin & Manzocchi 2006; Jensen
& Rosas 2006) leading to a pollution halo instead of a pollution haven ef-
fect
12
. In addition, companies with their headquarters located in industri-
alized countries are said to have better environmental management systems
in many cases because of the more stringent regulations in industrialized
countries and because these rms are often more aware of the interests of
green consumers, which commonly characterize industrialized export mar-
kets (Zarsky 1999, 8). Hence in this view the best way to diuse best
practice production techniques (Zarsky 1999, 5) is via FDI.
Usually studies empirically test the pollution haven argument by evalu-
ating whether sectors of typically dirty industries receive a larger proportion
of foreign investment compared to non-dirty industry sectors. Most statisti-
cal studies do not nd any eect in favor of the pollution haven eect (Jae
et al. 1995; Esty & Giradin 1998; Mani & Wheeler 1998; Busse 2004; Dean
et al. 2005). This is also the case if data on rm level is used, as in the study
on the former Soviet countries by Javorcik & Wei (2004).
One argument explaining why no pollution havens are found is that
abatement costs are very small compared to other production costs such
as investments in infrastructure etc. (Jae et al. 1995). This would imply
that studies do not nd pollution havens because abatement costs are too
low in developing countries but because they are too low in industrialized
countries, implying that rms do not need to internalize the environmental
costs they are causing (Zarsky 1999, 13). A further reason seems to be that
rms are afraid of negative reputation eects resulting, for example, in a
loss of consumers in developed countries if they conict with environmental
norms in developing countries (Dasgupta, Laplante & Mamingi 1998). In
addition, it is argued that there are more important reasons to reallocate
plants than environmental regulations, such as cheap labor, natural resource
endowments, taxes or transportation costs (Grey & Brack 2002).
In contrast to these studies, which nd no evidence for pollution havens,
there are a number of case studies that nd some evidence in favor of the
pollution haven argument (Clapp 1998; Xing & Kolstad 2002; Homann
et al. 2004). These case studies show that heavy pollutant industries in
particular relocate their plants in response to tightening regulations in their
home country. Clapp (1998), for example, nds that large US chlorine com-
panies relocated their plants to developing countries in response to the US
government banning the land disposal of chlorine wastes.
12
In addition to FDI, Chudnovsky & Lopez (2003) list three other mechanisms through
which technology can be diused to developing countries via Multinational Corporations
(MNCs): joint ventures, purchase of technology by rms in developing countries in con-
tractual form (e.g. patents or licensing), and reverse engineering, imitation or copying.
38
CHAPTER 2. LITERATURE REVIEW
In the context of the race to the bottom hypothesis, there is no statisti-
cal evidence that countries lower their environmental standards in order to
stay competitive (Grey & Brack 2002). However, there is some anecdotal
evidence with regard to the regulation of energy and taxation that favors the
regulatory chill hypothesis, which states that countries abstain from tight-
ening their environmental standards in order to stay competitive (Neumayer
2001).
In contrast, there are some examples of countries in which the opposite
of a regulatory race to the bottom has taken place, namely a race to the
top. The race to the top eect is often called the California Eect after
the example of California, which introduced higher car emission controls
than specied by the U.S. Clean Development Act (Vogel 1997). In 1990
the Californian standard even became the national standard, thus serving as
the most prominent example of a race to the top. This argumentation is in
line with the so-called Porter hypothesis, which states that a country can
benet from tighter environmental policies since this might foster innovation
and eciency and therefore also competitiveness (Porter 1990). Vogel (2000)
reports another country example, namely Israel, which in order to be able
to export to the EU market has introduced EU pesticides standards. This
evidence in favor of the California eect is further supported by the studies of
Christmann & Taylor (2001), Prakash & Potoski (2006) and Zeng & Eastin
(2009). However, Vogel (2000) cautions against too much optimism since
the California eect may only take place in very narrow circumstances, for
example if a country is aiming to enter into a trade agreement with the
EU or other industrialized countries. In addition, this eect seems to hold
predominantly for products (e.g. cars that have a catalytic converter) and
not for production processes (Bernauer & Cadu 2004).
In summarizing the literature on pollution havens and pollution ha-
los, Zarsky (1999) concludes that rms location decisions are generally not
driven by dierences in environmental standards and/or abatement costs.
This implies that there is little evidence in favor of pollution havens or a
race to the bottom. Although there are some examples of pollution halos,
notably in the energy sector in which foreign rms usually have superior
technology or if there are close links to green consumer markets, in general
a broad pollution halo eect for FDI also does not appear to exist (Zarsky
1999, 2).
2.3.3.2.2 Eects on the Environment Turning to the literature that
deals with the eects of FDI on environmental quality, we see that, analo-
gous to trade, the inuence of FDI can be divided into a scale, composition
and technology eect (Homann et al. 2004). The arguments are similar
to the trade context with slight dierences concerning the technology and
composition eect.
39
CHAPTER 2. LITERATURE REVIEW
The arguments for the technology eect closely resemble the arguments
for pollution halos. It is argued that foreign investors have newer and thus
more environmental friendly technologies, which supposedly leads to tech-
nology development, diusion or transfer (Araya 2002). An example would
be Novo Nordisk, a Danish pharmaceutical company, that opened a joint
venture in China, which used the latest technology to treat the plants
wastewater (Eriksen & Hansen 1999). In addition to this rm-level tech-
nology transfer, it is argued that there is also a spill-over eect: In general,
technology spill-overs can happen via human capital upgrading and employ-
ment turnover, increased competition requiring local rms to become more
ecient, due to supply chain requirements or imitation of technology by
local rms (Chudnovsky & Lopez 2003; Birdshall & Wheeler 1993). More-
over, foreign rms that invest in developing countries are usually large and
have more money available for research and development as well as for en-
vironmental management systems. Local rms in developing countries are
mostly reliant on learning and copying from more advanced rms, which
implies that they might adopt the more stringent environmental practices
of those advanced rms (Zarsky 1999, 8). However, it is also likely that
foreign rms will adapt to the local context, implying that the environmen-
tal standards might ultimately be well below those in their home country.
Evidence on the diusion of technologies by MNCs has so far been mixed
(Chudnovsky & Lopez 2003).
Finally, the composition eect states that FDI is supposed to lead to a
more ecient allocation among economies, implying that the production of
goods will take place with lower capital per unit of output and lower input
than without liberalization, which would imply a positive eect of both FDI
and trade. However, whether this composition eect is indeed ultimately
positive for the environment depends also on the particular situation of the
country (for example whether the investment implies a change from the rst
(agrarian) to the second (manufacturing) or to the third (service) sector)
(Araya 2002).
Studies dealing with the eects of foreign ownership in developing coun-
tries on environmental quality usually follow two dierent approaches: on
the one hand, they investigate particular plants in the receiving countries
and test whether they are more pollution intensive than non-foreign plants.
On the other hand, they look at national or subnational foreign investment
and relate this to national environmental quality measures. Altogether,
these studies have found mixed results.
Concerning evidence on the plant level, many studies show that foreign-
owned plants are usually less pollution intensive and use more stringent tech-
nology than domestic plants in developing countries (Gentry 1998; Black-
man & Wu 1999; Li & Yeung 1999; Eskeland & Harrison 2003; Ivarsson &
Alvstam 2005). In contrast, using survey evidence from Mexico Dasgupta,
Hettige & Wheeler (1998) nd that foreign investment has no eect on envi-
40
CHAPTER 2. LITERATURE REVIEW
ronmental eort. They measure environmental eort by the adoption of ISO
14 000 type procedures for pollution management and by the use of plant
personnel for environmental inspection and control. Hettige et al. (1996)
provide a summary of three studies from Asia that also do not nd any
signicant inuence of foreign direct investment on rm-level environmental
performance after controlling for rm size and the age of the plant. This
means that larger and newer rms, whether they are foreign or not, usually
perform better. In addition, what these studies nd is that it is not owner-
ship structures that matter but rather regulation and community pressure.
Apparently, richer and more educated communities are able to bargain ef-
fectively with rms (Zarsky 1999, 14) to pressure them to introduce higher
environmental standards.
Studies dealing with the eects of FDI on environment quality on the na-
tional or sub-national level are numerous and provide mixed results. Wheeler
(2001) nds that in China, Brazil and Mexico, the three top FDI receiving
countries in the developing world, air quality in the major cities has im-
proved with increasing FDI. Similarly, Zeng & Eastin (2007) nd that both
trade openness and FDI are positively associated with environmental quality
in Chinas provinces. In contrast, Gassebner et al. (2006) show that both
trade and FDI are associated with higher CO
2
emissions. Dierentiating
between inward and outward foreign direct investment, Koubi & Bernauer
(2009) nd that inward FDI is associated with higher concentrations of SO
2
whereas outward FDI is associated with lower concentrations.
Altogether, the literature on trade and on foreign direct investment in-
dicates that both positive and negative eects on pollution seem to exist.
Although the positive eects seem to be somewhat more pronounced, it is
dicult to come to a clear conclusion as context factors (investment and
country characteristics) seem to be of great importance (Zarsky 1999). This
is one reason why this thesis argues in favor of conditioning the eects of
international integration both political and economic on a countrys po-
litical system. The next section therefore presents the body of literature that
attempts to explain how the political system and the international system
are related.
2.4 The Political System and International Inte-
gration
Section 2.2 presented the literature relating environmental quality to the
political system. However, not only is the political system potentially con-
nected to environmental quality in a direct way; it also becomes important
through the indirect link via international integration. There are many stud-
ies showing that democracies react dierently than their autocratic counter-
parts to the incentives arising from the international system. More precisely,
41
CHAPTER 2. LITERATURE REVIEW
the decision whether to join a particular IGO is among other things de-
pendent on the political system of a country, the allocation of aid is also
determined partly by the political system of a country, and the intensity of
a countrys trade and investment ties are also not independent of its polit-
ical system. There are therefore good reasons not only to take the direct
link of the political system and international integration on environmental
performance in developing countries into account but also to include the
indirect eects of international integration through a countrys political sys-
tem. Hence the following sections discuss the literature linking the type of
the political system to international integration, focusing on each of the four
aspects of international integration in consecutive order.
2.4.1 The Political System and International Government
Organizations
Concerning membership in international governmental organizations, vari-
ous studies show that on average, democracies choose to join more inter-
national agreements and international organizations than autocracies. The
reasons put forward to explain this nding are manifold. Abbott & Snidal
(2000) argue that strong domestic legal institutions and traditions should
increase international credibility. Hence democracies, as they usually have
stronger legal institutions, should more easily commit themselves and thus
can be expected to cooperate more often on the international level. Other
studies argue that democratic countries tend to be more transparent than
autocracies and therefore can more easily and credibly commit themselves
to international cooperation (Remmer 1998; Gaubatz 1996). The argument
that democracies are more transparent is empirically supported by Bueno
De Mesquita et al. (2003) who nd that leaders responsible to a larger win-
ning coalition are more likely to report data on tax revenue collection and
on per capita income of their countries. Similarly, Rosendor & Vreeland
(2006) show that democracies provide more information on ination and un-
employment rates, controlling for income and IMF membership, interpreting
this as evidence in favor of the argument that democracies are indeed more
transparent.
Moreover, the literature on the democratic peace states that as democ-
racies are used to solving their domestic conicts by cooperation, they are
assumed to project their cooperative behavior to the international realm as
well, resulting in the fact that democracies engage more frequently in in-
ternational cooperation (Russett & Oneal 2001). Russett & Oneal (2001)
empirically show that democracies do indeed tend to join more IGOs, es-
pecially if these are already characterized by a high membership rate of
democracies.
In a similar vein, Manseld & Pevehouse (2008) point out two addi-
tional reasons why democracies join more IGOs. First, membership in in-
42
CHAPTER 2. LITERATURE REVIEW
ternational organizations allows governments to achieve benets for their
constituents such as increased trade levels which they would be unable to
achieve independently. Second, constituents are often unable to distinguish
between poor performances caused by incompetent government and those
caused by international factors. Becoming a member of dierent interna-
tional organizations can therefore hedge democratic governments from elec-
toral punishment.
The nding that democracies tend to join more international organiza-
tions is supported by the literature dealing with membership in international
regimes more generally. In their study on states ratication behavior with
regard to 22 global environmental treaties, Roberts et al. (2004) nd that
next to the total number of NGOs and a countrys narrowness of its ex-
port base, the main driving force behind ratication decisions seems to be
the voice and accountability of citizens through their domestic institutions.
Beron et al. (2003) show that decisions whether to ratify the Montreal Pro-
tocol depend strongly on the type of political system since democracies are
more likely to ratify the Montreal Protocol. This nding is supported by
Fredriksson & Gaston (2000); Neumayer (2002a,b) and by Bernauer et al.
(2009) who nd that democracy has a signicantly positive although weak
eect on the likelihood of environmental treaty ratication.
However, democracies do not only seem to be more willing to join in-
ternational organizations. They also tend to be more compliant with their
international obligations. Stein (2004) argues in her study on human rights
agreements that domestic enforcement plays an important role for govern-
ments decisions to commit to a particular agreement. She therefore con-
cludes that
democracies only make promises they can keep, ensuring that
the necessary conditions that promote subsequent compliance
are in place before ratifying (Stein 2004, 2-3).
2.4.2 The Political System and International Aid
As already hinted above, democratic countries are not only more likely to
join and comply with the rules set forth by international organizations. They
are also supposed to receive more international assistance. Although a coun-
trys political system does not seem to play a signicant role for the allocation
of bilateral aid as shown in Section 2.3.2, the allocation of multilateral aid
seems to depend at least partially on a countrys regime type. Neumayer
(2003c), for example, shows that political freedom is a signicant predictor
of multilateral aid allocation. Similarly, Neumayer (2003b) nds that at the
aid eligibility stage, respect for civil and political rights plays a statistically
signicant role for most donors although the nal amount of aid allocated
to a particular country in the end is not driven by its political system.
43
CHAPTER 2. LITERATURE REVIEW
Furthermore, the political system is not only important for the allocation
of aid but also for its eectiveness. Using a game theoretic model following
Bueno De Mesquita et al. (2003)s selectorate politics theory, Smith (2006)
shows that democratic political leaders use foreign aid to provide more public
goods whereas in autocracies foreign aid is likely to be skimmed by the ruling
elite.
2.4.3 The Political System and Trade
Turning to the economic side of international integration, there are many
studies showing that in general, democracies seem to trade more intensively.
Manseld et al. (2002), for instance, show that democracies are more likely
to sign trade liberalization agreements. The reasoning behind this nding is
that, rst, by signing trade agreements political leaders can signal to their
voters that they are acting in the best interest of their country on economic
issues. Second, since liberalizing trade is seen as welfare enhancing, elected
governments have an incentive to liberalize trade and therefore increase their
chances of reelection. In addition, it is argued by Manseld et al. (2002) and
Kono (2006) that political opponents inform citizens if governments do not
adopt welfare enhancing trade policies, thereby providing governments with
an additional incentive to liberalize. Controlling for trade policies, Kono
(2006) shows that democracies do indeed trade more than autocracies.
Eichengreen & Leblang (2006) support this argument that democratic
countries have higher trade levels empirically using an instrumental variable
approach. They show that globalization, which the authors operationalize
using trade openness, has an eect on democracy and, conversely, democ-
racies seem to be more globalized, i.e. they are more open to trade. These
results are conrmed by Doces (2006) who shows that increasing exports to
the US seem to cause democracy and at the same time democracy seems
to cause higher exports to the US. Furthermore, Russett & Oneal (2001)
show that democracies tend to trade more with each other than with auto-
cratic countries. This claim is empirically supported by Ingram et al. (2005)
who nd, as do Bliss & Russett (1998), that bilateral trade is higher for
democracies.
2.4.4 The Political System and Foreign Direct Investment
Concerning the connection between foreign direct investment and political
regime type, the literature is well-developed. However, studies exist that
argue both that democratic political systems are more favorable for foreign
investment and that autocratic systems attract more investment. Empir-
ically, the literature has up to now produced inconclusive results: several
studies (Ahlquist 2006; Feng 2001; Jacobsen & De Soysa 2006; Jensen 2003;
Jensen & McGillivray 2005; Jensen 2006), have found that the likelihood of a
44
CHAPTER 2. LITERATURE REVIEW
multinational company investing in a democratic country is higher whereas
Li & Resnick (2003) found that large amounts of FDI ow into authoritar-
ian regimes. Concentrating on investment from industrialized to developing
countries only, Guerin & Manzocchi (2006) show that democracy, measured
in a binary way, is associated with increased FDI inows controlling for
trade and market-oriented reforms.
The argument underlying the ndings in favor of democracy attract-
ing more FDI is that in democracies, due to more veto players and higher
audience costs, more constraints are imposed on political leaders, which re-
sults in a higher policy credibility for MNCs (Jensen 2003). More precisely,
Jensen (2003) argues that rst, breaking a deal with a foreign investor im-
plies greater audience costs for a democratic leader than for an autocratic
leader since in democracies voters might punish their leader in the next elec-
tion. Second, he brings forward the argument that in countries with more
veto players, policy change is less likely to occur because a higher number of
veto players limits the discretion of democratic leaders to change investment
policies at short notice, implying greater security for investors and therefore
inducing higher FDI inows.
In contrast, according to Choi & Samy (2008, 85), the conventional
wisdom has long been that autocratic countries provide a more favorable
investment environment because investment deals are usually not subject
to demands from trade unions and pressures from civil society to increase
wages. Furthermore, Li & Resnick (2003) argue that democracies limit the
monopolistic or oligopolistic power of foreign rms and increase the ability
of local businesses to pursue protection against foreign capital, implying that
democracies are limited in their ability to oer generous scal and nancial
incentives to foreign investors (Choi & Samy 2008, 86).
To test which of the rival theories concerning democracy and FDI holds
true Choi & Samy (2008) propose using the number of veto players as a
more valid measure for the stability argument made by Jensen (2003). This
approach enables Choi & Samy (2008, 83, 98) to show that
democratic institutions are, at best, weakly associated with in-
creases in FDI inows (measured by FDI/GDP ratios). [. . . ] Put
dierently, FDI inows are correlated with politics of veto players
who are likely to have a direct inuence in thwarting investment
policy rersals against MNEs, but not connected with the politics
of audience costs that may not closely follow particular details
of FDI policy changes by national leaders.
This nding is supported by B uthe & Milner (2008) who show that the
number of veto players is signicantly associated with the net ow of inward
FDI, whereas dierent measures of electoral democracy, such as Polity or
Freedom House, do not reach standard signicance levels. Hence it appears
to be more important for foreign investors to choose countries which seem to
45
CHAPTER 2. LITERATURE REVIEW
oer a stable investment environment, whereas the question whether these
countries are democratic or autocratic is of secondary importance. How-
ever, as the number of veto players is usually higher in stable democracies,
democracies ultimately represent the more stable investment environments
for FDI.
In summary, one can conclude that in general, democracies seem to
be more willing to engage internationally and perhaps more importantly,
appear to be more eective in using their international ties for a better
provision of public goods. Although the degree to which this statement em-
pirically holds true seems to vary with the type of international integration
under consideration more for IGO membership and trade and less so for
aid and foreign investment the general tendency is clearly supported by
the literature. Consequently, this nding warrants including both the direct
eect of international integration on the provision of environmental perfor-
mance in developing countries as well as the indirect link of a countrys
integration through its political system.
2.5 Conclusion
Altogether, the dierent strands of literature important for this thesis sug-
gest several important conclusions for the study of the environmental perfor-
mance of developing countries. First, the diverse ndings of the numerous
studies on the Environmental Kuznets Curve show that focusing on a coun-
trys national income as the major explanatory factor of its environmental
performance leads to inconclusive results. Hence it is warranted to enlarge
the scope of the analysis and to include other important factors that can help
us better understand the variation in developing countries environmental
quality.
One such factor is the political system of a country. Here most stud-
ies suggest that democratic forms of government should be conducive to
a better provision of environmental quality. However, some studies only
show weak or mixed evidence in favor of democracies, which indicates to
more carefully align the arguments how the political system is to inuence
environmental quality to the specic research situation (Midlarsky 1998;
Fredriksson & Svensson 2003; Baettig & Bernauer 2009). Therefore, the
theoretical framework of this thesis adjusts the arguments on the political
system to the context of developing countries.
In addition to a countrys national income and its political system, the
capability to provide environmental quality can be inuenced by factors
emanating from outside a countrys borders. The existing literature has so
far focused mainly on trade and FDI to explain variations in environmen-
tal pollution. However, the various studies dealing with the eect of both
IGO membership and international aid on domestic variables such as policy
46
CHAPTER 2. LITERATURE REVIEW
changes and democratization processes, suggest that there are good reasons
to enlarge the scope of international integration to also include these two
aspects of international integration.
Furthermore, the literature on the inuence of the political system on in-
ternational integration points to the importance of taking the indirect eect
of the political system through a countrys international integration into
account. This literature does not only suggest that democratic countries
should join more IGOs, receive more international aid and FDI and have
higher trade levels but this literature hints at the possibility that democra-
cies are also more responsive to the inuences emanating from the interna-
tional system. Consequently, the theoretical framework established in the
next chapter accommodates both the direct and the indirect eect of the
political system on developing countries environmental performance.
Having presented the various strands of literature that are important in
examining the question of how a countrys international integration and its
political system inuence its environmental performance, I now turn to the
theoretical framework established to answer this question.
47
Chapter 3
Theory
The last chapter presented the various strands of literature that are impor-
tant to understand the eects of the political system and the international
system as well as their interaction eect on the environmental performance
of developing countries. Following closely the structure of the literature
review, the theoretical arguments developed to analyze the eect of inter-
national integration and the political system on the environmental quality
of developing countries evolve in four parts.
First, following the logic of the Environmental Kuznets Curve, I present
arguments showing why developing countries should theoretically face in-
creasing environmental pollution when experiencing economic growth. These
arguments highlight the fact that developing countries usually have low envi-
ronmental quality because they lack both the capability and the willingness
to decrease pollution due to their low national income levels. Second, I con-
clude from these arguments that the political system should not have an in-
dependent eect on the environmental performance of developing countries.
This implies that democratic developing countries per se should not be char-
acterized by better environmental performance relative to their autocratic
counterparts. Third, the eect of international integration on environmental
performance is analyzed. Considering consecutively each of the four aspects
of international integration IGO membership, international aid, trade and
foreign direct investment it is shown how international integration could
increase both the capability and the willingness of governments in devel-
oping countries to boost the provision of environmental quality. Finally,
an interaction eect between international integration and the political sys-
tem is introduced to capture the idea that although the political system
should not have an independent eect on environmental quality, it might
nevertheless mediate the eect of globalization on environmental quality in
developing countries. This implies that the presumed positive eect of the
four factors subsumed under the term international integration is expected
to be stronger in democratic countries compared to the eect in autocratic
48
CHAPTER 3. THEORY
countries. Figure 3.1 outlines the theoretical argument.
4FactorsofInterna.onal
Integra.on:
IGOMembership,Interna.onalAid,
TradeandFDI
EnvironmentalQualityin
DevelopingCountries
Poli.calSystem
Democraciesmaybe
be-erintegrated
interna2onally
Democraciesmaybemore
responsivetotheinuenceof
thevariousformsof
interna2onalintegra2on
+
Eectofinterna2onal
integra2onshouldbe
strongerindemocracies
Figure 3.1: Illustration of theoretical argument
3.1 Income and Environmental Quality
As described in detail in the literature review, many studies empirically
nd an inverted U-shaped relationship between GDP per capita and several
indicators of environmental pollution, which was labeled the Environmental
Kuznets Curve (EKC) (Grossman & Krueger 1995; Selden & Song 1994;
Shak & Bandyopadhyay 1992). To recapitulate briey, the logic underlying
the EKC is that in the course of a countrys economic development, rising
national income leads initially to an increase in environmental pollution
before pollution levels reach a turning point and eventually decrease. The
initial increase in pollution is due to the growth in scale of economic activity
associated with economic growth, which all else being equal results
in rising pollution levels (scale eect). However, when attaining a certain
threshold of national income, pollution may decline due to the composition
and the technology eect. The composition eect describes the possibility
that with rising income the composition of an economy is likely to change
from manufacturing to service, which would imply a reduction in pollution
emissions. The technology eect posits that with rising national income,
technological progress tends to lead to less environmental pollution.
In addition to these pure economic eects, there is also a political compo-
nent to the EKC: early on in the course of a countrys economic development,
environmental quality is usually considered a luxury good and since states
have only limited resources available, environmental quality at this stage of
economic development usually ranks, for example, far behind the demand for
better economic conditions. However, once people attain a certain level of
49
CHAPTER 3. THEORY
living standards, environmental quality turns into a normal public good and
constituents demand that their governments take action to reduce or avoid
pollution, for example by enacting appropriate environmental regulations
(Selden & Song 1994).
Applied to the context of developing countries, the EKC argument im-
plies that most if not all of these countries should be situated on the upward
sloping part of the curve. This, however, suggests that these countries will be
confronted with deteriorating environmental quality when experiencing eco-
nomic growth since at their present stage of economic development neither
the composition nor the technological eect should outweigh the scale eect.
Consequently, following the arguments underlying the EKC, we should ex-
pect that for developing countries an increase in national income should be
associated with a reduction in environmental quality because these countries
have not yet reached the turning point after which an increase in national in-
come is associated with an increase in environmental quality. This therefore
implies that
H1: An increase in income for developing countries is associated
with a decrease in environmental quality.
3.2 No Independent Eect of Political System
In addition to the fact that according to the EKC argument, developing
countries should theoretically face increasing environmental pollution, the
citizenry in these countries is also unlikely to demand that their govern-
ments spend scarce resources on improving environmental performance at
the expense of basic needs such as education and health. As already hinted
at in the previous section, the particular nature of environmental quality as
a luxury good
1
(Rogowski 2003) implies that environmental quality is not or
only very little demanded at low-income levels. Hence it follows that people
in developing countries should not demand that their governments take ac-
tion in order to enhance environmental quality. As Neumayer (2002a, 150)
notes:
This need not imply that poor countries care less about the
environment per se. Rather, because of their poverty, they might
prioritize issues other than the environment.
Consequently, one can argue that the political system should not have an
independent eect on the environmental performance of developing coun-
tries. Or put dierently, if there is no (or only very little) public demand
1
A luxury good is a good for which the demand increases with the income level. This
implies that at low-income levels the good is usually not demanded or only in low quantities
whereas its demand increases with growing income.
50
CHAPTER 3. THEORY
pressuring a government to increase environmental quality, democratic gov-
ernments have no electoral incentives to spend scarce resources on improving
their environmental performance. It therefore follows that without the nec-
essary electoral incentives, democratic governments should not care more
about their environment than autocratic governments. Hence
H2: Democratic developing countries per se are not character-
ized by higher environmental quality relative to their autocratic
counterparts.
Taken together, these two arguments suggest a very dark picture for the
environmental performance of developing countries. On the one hand, in-
creasing national income should theoretically lead to more environmental
degradation in developing countries. On the other, at this early stage of
economic development democracies should not be better caretakers of their
environment than their autocratic counterparts since public demand for bet-
ter environmental quality should not constitute a great enough electoral in-
centive for democratic leaders to spend scarce resources on the improvement
of environmental quality. Consequently, one could conclude that developing
countries need to grow rich before they are able to clean up. However, this
conclusion is not in line with the empirical evidence presented in the intro-
duction, in which we saw a wide variation in the environmental performance
of developing countries that is unexplained by their national income.
Moreover, this conclusion ignores the possibility that a countrys capa-
bility and willingness to reduce pollution might also be aected by factors
emanating from outside its national boundaries. Up to this point only do-
mestic factors have been presented as being important and decisive for the
provision of better environmental quality. However, states do not act in a
vacuum: they interact and are inuenced by other states and international
actors (such as IGOs), by economic pressures and dependencies. Conse-
quently, it has been argued that international integration also inuences the
capacity and interest of states to provide environmental quality (Antweiler
et al. 2001; Frankel & Rose 2005; Neumayer 2002b). However, the extent
to which international integration aects a countrys capability and willing-
ness to provide public goods is likely to depend on the political system of the
particular country. Consequently, I argue that although the political system
should not have an independent eect on environmental quality in develop-
ing countries, the type of the political system might nevertheless mediate
the eect of the international system on the environmental performance of
developing countries implying, thus, a dierence in the strength of the eect
of international integration in democracies versus autocracies.
In the next sections I therefore elaborate on the mechanisms by which
international integration and its interaction with the political system are
presumed to inuence environmental quality. To enhance the clarity of the
51
CHAPTER 3. THEORY
argument, I rstly analyze the impact of the four dierent manifestations
of international integration IGO membership, international aid, trade and
foreign direct investment on environmental quality and then incorporate
their interaction with the political system.
3.3 International Integration
3.3.1 Membership in International Organizations
The general purpose of international governmental organizations is to fa-
cilitate cooperation at the international level and thus enable countries to
solve problems which they would often be unable to solve independently, and
hence realize benets from mutual cooperation (Keohane 1984). IGOs pro-
mote cooperation by reducing transaction costs, by decreasing uncertainty,
by widening the shadow of the future, by increasing reputational costs and
by allowing credible commitments (Keohane 1984; Abbott & Snidal 2000).
Consequently, I argue that membership in IGOs may inuence the environ-
mental performance of developing countries for the following reasons:
First, since IGOs raise the reputational stakes for reneging on agree-
ments, members of a particular IGO should in principle obey the corre-
sponding rules (Keohane 1984). In addition, depending on the particular
type of IGO, non-compliance could even be sanctioned, meaning that the
responsible organization can demand that participant countries enforce the
corresponding rules. However, this will directly aect environmental quality
only in the case of IGOs that have some environmental purpose, such as the
Multilateral Fund for the Implementation of the Montreal Protocol or the
North Atlantic Salmon Conservation Organization
2
.
Second, IGOs create norms on good behavior or, rather, on what con-
stitutes bad conduct (Ward 2006; Young & Levy 1999). Although an IGO
might not have a direct link to environmental purposes, the general idea
of seeing environmental protection as important might be spread through
membership in IGOs in general. Abbott & Snidal (2000) subsume these
two mechanisms under techniques ranging from litigation and sanctions to
persuasion, normative appeals, and shaming. Hence international organi-
zations can constitute a forum in which countries or other actors such as
environmental non-governmental organizations (ENGOs) that have an in-
terest in environmental protection can promote this idea and also try to
inuence other states in this direction. This eect of IGO membership is
emphasized in the literature on policy diusion. Cao (forthcoming), for ex-
ample, shows that interaction at the international level through trade ties
or connections in the IGO network can serve as channels of communication
2
Out of the 415 IGOs that are considered in the empirical analysis, around 50 organi-
zations can be classied as environmental. For a more detailed discussion of the various
mandates of the IGOs considered in this analysis, see section 6.4 in chapter 6.
52
CHAPTER 3. THEORY
that allow countries to learn from other countries policy decisions and thus
provide for a socialization process among states.
Third, IGOs often deal with a variety of dierent issues (Vogler 1995;
Young & Levy 1999), and countries although they often join an organiza-
tion because of a certain issue only are also inuenced by other issues dealt
with by this organization. Hence even though countries become members
in an organization for some specic reasons, e.g. nancial assistance, they
are also exposed to the other purposes of this organization such as environ-
mental protection (an example would be the World Bank or the EU). Jahn
(2002), for example, posits that
[. . . ] the membership in an international organization such as
the EU may support the improvement of environmental stan-
dards. The rationale is that particular environmentally conscious
countries take the role of pioneers that push environmental is-
sues onto the agenda of countries that might be perceived as
environmental laggards.
3
The case of Laos corroborates this argument: by joining the Association of
Southeast Asian Nations (ASEAN) in 1997, Laos was required to implement
a number of agreements dealing with agro-economic development. The im-
plementation of these agreements aimed at making the agrarian sector more
sustainable had also, as a side eect, a positive impact on Laos environment
(UNEP 2001b).
These three arguments mentioned above all refer to mechanisms through
which IGOs might inuence a countrys willingness to enhance its environ-
mental performance. However, an additional argument suggests that mem-
bership in IGOs may also inuence the capability of developing countries to
improve their environmental performance by highlighting that IGOs also al-
low information and technology transfer. Porter et al. (2000) argue that due
to new information and knowledge on environmental pollution, awareness
of environmental quality as an important public good could be created and
knowledge and technology on abatement possibilities could be distributed.
This implies that a country that is better politically integrated in the in-
ternational system could receive more relevant information and could more
easily draw on technical knowledge and assistance from international ac-
tors. Examples of IGOs that diuse information and technologies that also
3
This idea is closely related to the concept of issue linkage (Keohane 1984): in order to
achieve cooperation on an environmental issue, benets in other areas like trade, nancial
or technological assistance etc., could be oered to countries that would not be willing to
cooperate on environmental issues in general. Similarly, in the words of Ward (2006, 151)
there is greater scope for sanctioning non-compliance in relation to one regime when
nations can withdraw their cooperation over other issues (Lohmann 1997); and there
is greater scope for getting cooperation through issue-linkage, supposing there is some
heterogeneity of preference (Martin 1995; Finus 2001). Thus the existence of multiple ties
between nations allows sanctions and bribes to be used.
53
CHAPTER 3. THEORY
benet the environment are the Food and Agriculture Organization (FAO)
and the World Health Organization (WHO). Although these organizations
mostly pursue goals that are not motivated by environmental protection per
se, such as access to safe drinking water and better sanitation facilities, their
actions often have positive spill-over eects on the environment as well.
Taken together, these arguments lead to the conjecture that membership
in IGOs is expected to positively inuence the environmental performance
of developing countries. Thus, the third hypothesis states:
H3: Developing countries that are members of more IGOs pro-
vide higher levels of environmental quality.
3.3.2 International Assistance
Similar to membership in IGOs, the provision of nancial and technological
assistance can inuence developing countries capability and willingness to
increase environmental performance in various ways. However, not all of
these eects necessarily lead to an increase in environmental quality.
First, international assistance often implies a transfer of technologies
from donor countries to developing countries. Similar to the technology ef-
fect of FDI described in section 2.3.3.2 in chapter 2, this technology transfer
usually implies that newer and better technologies are diused to devel-
oping countries since the technological standards of industrialized countries
are in general well above the standards of most developing countries
4
(Araya
2002). This technology transfer accompanying international aid can imply a
diusion of technologies that are directly related to reducing pollution, such
as abatement or sewage treatment technologies. In this case, a direct and
positive eect on pollution reduction can be expected. However, even tech-
nologies that are not necessarily intended to reduce pollution, such as more
ecient production technologies, may as a side eect improve environmental
quality. Hence the transfer of technology through international aid should
in general enhance the capability of developing countries to reduce pollution
since these technologies allow developing countries to produce their goods
in a cleaner way and tackle environmental degradation more eectively.
Secondly, as in the case of IGO membership, international assistance al-
lows information and knowledge transfer, thus increasing both the capability
and the willingness of developing countries to improve their environmental
performance (Porter et al. 2000). In contrast to the rst eect, which refers
directly to better technology being transferred to developing countries, this
argument refers to a softer diusion process in the form of knowledge and
information transfer. This increase in information and knowledge can raise
4
Although it is likely that it is not the newest technologies that are diused, their
technological standards should still be above those in place in most developing countries.
54
CHAPTER 3. THEORY
governments awareness of environmental protection and thus lead to a so-
cialization process towards sustainable development. To illustrate this argu-
ment, one can use the example of the World Bank, which is also described in
more detail in the case studies below. Since the beginning of the 1990s, the
World Bank has aligned its lending with the principle of sustainable devel-
opment, which aims to reconcile economic growth with the sustainable use
of a countrys national endowments (Gutner 2005; Nielson & Tierney 2003).
Therefore, by promoting sustainable development through its lending, the
World Bank provides important information and knowledge about environ-
mental protection and thus socializes developing countries in this direction
without necessarily providing better technologies.
Thirdly, international donors can assist states in enforcing environmental
regulations, thereby enhancing developing countries willingness to improve
their environmental performance. Through the principle of conditionality,
donors often demand that developing countries implement certain regula-
tions as a prerequisite for the provision of nancial benets. In general,
these conditions for international assistance are related to macroeconomic
policies or various forms of good governance. However, in recent decades
some donors have also included environmental regulations in their catalogue
of conditionalities. One example would again be the World Bank: since it
underwent its so-called greening reform in the early 1990s (Gutner 2005;
Nielson & Tierney 2003), environmental protection measures, along with
good governance and economic criteria, belong to the set of issues demanded
under the conditionality principle. In addition, governments might even use
this conditionality to implement regulations that they themselves consider
important, but which they have been unable to implement independently
(Vreeland 2003). Thus the pressure from international donors can be used
to introduce regulations that have been opposed domestically by actors such
as the opposition or major domestic corporations.
In addition to these three eects, international assistance provides gov-
ernments with additional funding. These additional resources could in prin-
ciple allow developing countries to invest more funds in the provision of
public goods, for example. Hence recipient countries would not need to
spend their own scarce resources on providing environmental quality but
could rely on this additional funding from international donor institutions.
However, whether the provision of additional funding through international
donors would indeed enhance the environmental performance of developing
countries is questionable since other types of public goods such as better
health care might be prioritized by governments in developing countries, as
argued earlier.
Contrary to these positive eects, international assistance could also
aect environmental quality in developing countries negatively. Since one
important goal of international assistance is to alleviate poverty by stimulat-
ing growth, most international assistance aims to boost economic activity
55
CHAPTER 3. THEORY
in developing countries (World Bank 2009a). Following the logic of the
scale eect underlying the Environmental Kuznets Curve, this aid induced
increase in economic activity would imply an increase in pollution levels,
all else being equal. Counteractive to this negative scale eect, economic
growth leading to an increase in national income might also trigger a higher
demand for environmental regulation and thus result in better environmen-
tal quality, depending of course on where on the Environmental Kuznets
Curve a country is located
5
.
Altogether, the provision of nancial and technological assistance would
lead to better environmental quality only if the positive eects of aid out-
weigh this scale eect arising from aid induced economic growth. On the
basis of the many studies showing that the technology transfer accompanying
foreign direct investment usually outweighs the scale eects of investment as
outlined in section 2.3.3.2 in chapter 2, it is reasonable to assume that the
technology eect together with the other positive eects of international as-
sistance should compensate for the increase in pollution induced by a boost
in economic growth. Hence I posit that international assistance increases
the capability and the willingness of developing countries to provide better
environmental quality.
H4: Developing countries that receive more international assis-
tance provide higher levels of environmental quality.
3.3.3 Trade
As presented in detail in the literature review, it is customary to analyze the
eect of both trade and foreign direct investment on environmental quality
in line with the scale, technology and composition eect (Antweiler et al.
2001; Frankel & Rose 2005). This decomposition of the eect of trade and
FDI allows all the dierent pathways to be taken into account along which
trade and FDI can in principle aect the provision of environmental quality.
Starting with the scale eect, it is argued that along with growing trade,
at least if the growth in trade is partly due to an increase in exports, an
expansion in economic activity is inevitable (Grossman & Krueger 1995; Esty
2001; Cole 2004). All else being equal, an expansion in economic activity
would imply a rise in pollution levels.
5
This potential increase in the demand for environmental protection induced by rising
national income might not be of great importance for the very poor developing countries.
In this case, an increase in national income would not move this country much further
upward on the Environmental Kuznets Curve. Hence environmental protection would
probably still not be considered highly important even after an increase in national income.
However, for those developing countries that are already further towards the potential
turning point of the EKC, an increase in national income might indeed imply a rise in the
demand for environmental protection.
56
CHAPTER 3. THEORY
The technology eect, which accounts for the possibility that trade might
be accompanied by a spread of more recent technologies between trading
partners, operates against this scale eect. Especially for developing coun-
tries, the technology eect could imply access to better and cleaner tech-
nologies and management practices such as catalytic converters for cars or
scrubbers on smokestacks (Esty 2001) and thus could lead to a decrease in
pollution.
The most debated eect of trade is the composition eect. As described
in detail in the literature review, one important consequence of trade is
that it allows each country to specialize in those activities in which it has a
comparative advantage, i.e. those activities that it performs relatively well
(Caves et al. 2002; Cole 2004). Following standard trade theory (Hecksher-
Ohlin) this would imply that developing countries should specialize in their
abundant factor which is labor intensive industries. However, since in gen-
eral polluting industries such as chemical industries tend to be more cap-
ital intensive, these polluting industries should stay located in industrial-
ized countries because the comparative advantage of industrialized coun-
tries lies in capital intensive production (Antweiler et al. 2001; Nordstroem
& Vaughan 1999).
In contrast, the pollution haven argument suggests that it is not most
important that developing countries have a comparative advantage in labor
intensive industries but that they have a comparative advantage in pollu-
tion intensive production. Accordingly, dirty industries should settle in those
countries which have the laxest environmental regulations, which are usu-
ally the low-income countries, implying that the composition eect of trade
leads to a decrease in environmental quality for developing countries (Jae
et al. 1995; Esty & Giradin 1998; Staord 2000). Some authors even claim
that the threat of dirty industries settling in countries with lax environmen-
tal regulations will lead to a competition between countries to oer those
industries a better production environment and therefore to a regulatory
race to the bottom (Revesz 1992; Drezner 2001).
In addition to the scale, technology and composition eect, trade can
aect environmental quality indirectly through its eect on income. As trade
allows for a more ecient allocation of resources, it is argued that trade can
therefore be seen as generally welfare enhancing (Caves et al. 2002). This
implies that increasing trade should lead to a growth in national income.
Depending on where a country is located on the Environmental Kuznets
Curve, this increase in national income may lead to a higher demand for
environmental regulation and thus to better environmental quality.
As described in the literature review, empirical studies nd little evi-
dence for the pollution haven hypothesis and even less for the race to the
bottom assertion (Cole 2004; Frankel 2003). Rather, it seems to be the case
that due to technology diusion and the overall welfare gains of growing
trade openness, trade has, overall, a positive eect on environmental quality
57
CHAPTER 3. THEORY
(Zarsky 1999).
In accordance with these earlier empirical studies, I posit that the eect
of trade on a countrys environmental performance is positive:
H5: Developing countries with higher levels of trade provide higher
levels of environmental quality.
3.3.4 FDI
Although, as we have seen already in the literature review, the logic of how
FDI can be linked to environmental quality is similar to the inuence of trade
on environmental quality, some important dierences also exist. Analogous
to trade and income growth, the inuence of FDI on environmental quality
is usually divided into a scale, composition and technology eect (Homann
et al. 2004). Whereas the scale and the composition eect of FDI work in a
similar way to the eects of trade, the technology eect is the major point
where the eects of trade and foreign direct investment dier theoretically.
In general, as a consequence of increasing FDI, economic activity is ex-
pected to grow. All else being equal, more FDI should therefore lead to more
pollution as the scale of production rises (Zarsky 1999). This scale eect
could be especially signicant since foreign rms usually invest in large-scale
plants.
However, whether production and thus pollution will necessarily grow
because of foreign direct investment depends heavily on the sector in which
it is invested. Investments in the manufacturing or agricultural sector will
result in higher production and hence, according to the scale eect, will
lead to higher pollution. Investments in the service sector, however, such as
call centers or software development rms, might have very little eect on
pollution levels. Consequently, the level of pollution depends on the sector
in which it is invested (composition eect). Furthermore, as FDI might have
positive spill-over eects, it is likely that the particular sector which receives
increasing foreign direct investment will also experience higher growth levels
in national rms as a consequence of FDI ows. This means that if, for
example, the service sector of a particular country receives increasing FDI
ows, the whole sector might experience growth and might thus attract more
capital and labor away from the more polluting manufacturing sector. In
this case, according to the composition eect, FDI would lead to an overall
positive eect on pollution levels. However, the exact opposite could also be
possible if FDI ows were mainly directed into the manufacturing sector.
This composition eect is, again, the main concern of the proponents
of the pollution haven argument. Assuming that developing countries have
a comparative advantage in pollution intensive production, the pollution
haven argument posits that it is those rms that are highly pollutant, in
particular, which will invest in developing countries. This would imply that
58
CHAPTER 3. THEORY
the composition eect of FDI translates to an increase in investment in
pollution intensive production and hence would lead to an increase in overall
pollution levels.
Finally, the eect of foreign direct investment on pollution depends heav-
ily on the technological level of the investment. Here the dierence between
trade and FDI becomes most apparent. Whereas increased trade leads to
more intensive economic contact between trading partners, which could po-
tentially lead to a technology transfer, an increase in FDI implies that a
foreign company investing in a plant brings along its own technology. Hence
with regard to foreign direct investment, a multinational corporation (MNC)
is not trading with a local partner that usually has a lower technological stan-
dard but is establishing a branch of its own company and therefore should
deploy newer technologies (Zarsky 1999). As MNCs are mostly located in
industrialized countries they are assumed to have higher technological stan-
dards than companies in developing countries. Therefore, along with foreign
investment, newer and cleaner technologies are supposedly diused to de-
veloping countries (Guerin & Manzocchi 2006; Jensen 2006).
In addition to this rm-level technology transfer, it is argued that there
is also a spill-over eect, which can happen via human capital upgrading
and employment turnover, increased competition due to which local rms
need to become more ecient, due to supply chain requirements or due to
imitation of technology by local rms (Chudnovsky & Lopez 2003; Birdshall
& Wheeler 1993). In addition, since foreign rms that invest in developing
countries are usually large, they are assumed to have more money avail-
able for research and development as well as for environmental management
systems (Zarsky 1999).
However, it is possible that foreign rms will adapt to the local context,
implying that the environmental standards nally in practice might be below
those in their home country (Chudnovsky & Lopez 2003). Nevertheless, one
of the reasons why the nal environmental standards of foreign companies
will still be higher than those in place in most developing countries is to
retain access to markets in industrialized countries (Zeng & Eastin 2007).
The access to industrialized markets is especially important because most
foreign trade occurs between multinational companies and their suppliers
(UNCTAD 1996), implying that a certain minimum technological standard
needs to be maintained. In addition, the access to foreign markets exposes
rms to the pressures of green consumers forcing foreign rms to comply
with higher technological standards (Zarsky 1999). Finally, for fear of a
potential dispute under WTO rules, many multinational companies tend to
comply ex-ante with strict environmental standards (Zeng & Eastin 2007).
Thus the overall inuence of FDI depends on which eect scale, com-
position or technology dominates. Based on the many studies that nd
a positive eect of FDI due to the technology eect, I tentatively posit a
positive eect of foreign direct investment:
59
CHAPTER 3. THEORY
H6: Developing countries which receive more foreign direct in-
vestment provide higher levels of environmental quality.
To summarize, I posit that the eect of each of the four aspects of inter-
national integration IGO membership, international aid, trade openness
and FDI on the environmental performance of developing countries is pos-
itive. Although there are both positive and negative arguments that link
aid, trade and FDI to environmental quality, I build on the ndings of early
studies to hypothesize an overall positive relationship between these forms
of international integration and environmental quality. As the next step I
turn to the discussion of how the political system acts as a mediating factor
between a countrys level of international integration and its environmental
performance.
3.4 Interconnection with the Political System
In section 3.2, I have established the argument that the political system
should not have an independent eect on environmental quality in develop-
ing countries. However, as shown in the literature review, there might be
an important indirect eect of the political system on environmental quality
as democracies seem to react dierently than their autocratic counterparts
to the incentives arising from the international system. Therefore, I argue
that the political system acts as a mediating factor between a countrys in-
ternational integration and its environmental performance. More precisely,
I posit that the positive eect of international integration is intensied in
democracies and mitigated in autocracies. That is, at any given level of
international integration, environmental performance should be higher in
democracies relative to autocratic countries because the presumably positive
eect of international integration is stronger in democracies. Consequently,
I introduce an interaction eect between the four dierent manifestations
of international integration IGO membership, international aid, trade and
foreign direct investment and a countrys political system to capture this
idea.
In particular, I argue that the reason why the positive inuence of inter-
national integration on environmental quality should be enhanced in democ-
racies is twofold: First, democracies might be more responsive to the inu-
ence of the various forms of international integration, thus using this in-
uence for a higher provision of public goods compared to their autocratic
counterparts. Second, democracies may in general be better integrated into
the international system, therefore indirectly intensifying the positive inu-
ence of international integration on environmental quality. The next two
sections consecutively discuss these two arguments.
60
CHAPTER 3. THEORY
3.4.1 The Political System and the Response to Interna-
tional Integration
The gist of the rst argument on how the political system indirectly inu-
ences environmental quality in developing countries is that the political sys-
tem should be important for a countrys responsiveness to the various forms
of international integration. Hence the political system should act as a me-
diator in relation to the eect of international integration on environmental
performance. The reasons why democracies could be more responsive to the
inuences of the international system are twofold. First, democracies might
use the resources provided by IGOs and international aid donors for a higher
provision of public goods. Second, environmental quality might be higher in
democratic developing countries because technological and knowledge trans-
fer induced by the various forms of international integration might happen
more easily in democracies. Both arguments are presented successively in
the following paragraphs.
As outlined above, both IGOs and international aid donors diuse re-
sources such as knowledge, technologies and funding to their member or
recipient countries. These resources may constitute a ready opportunity for
democratic leaders to provide public goods such as environmental quality
and thus enhance their chances of remaining in oce.
As presented in detail in section 2.2 in chapter 2, several authors argue
that democracies provide more public goods to their citizens than autocra-
cies (Deacon 2003, 1999; McGuire & Olson 1996; Bueno De Mesquita et al.
2003; Olson 1993). The reasoning behind this argument is that in order to
survive in oce, democratic leaders need the support of the majority of their
citizenry whereas in an autocracy leaders only need the support of the ruling
elite (Bueno De Mesquita et al. 2003). Hence in order to maximize politi-
cal support given their resources, it should be more ecient for democratic
leaders to provide public goods than to buy o their winning coalition with
targeted private goods. This should result in democracies, overall, providing
more public goods to their citizenry than autocracies.
Conversely, one can argue that due to their longer planning horizon, au-
tocratic leaders are more willing to invest in the provision of environmental
quality (Congleton 1992; Midlarsky 1998)
6
. Since most environmental prob-
lems are long-term in nature, democratic leaders facing frequent reelections
might be unwilling to bear the short-term costs of providing long-term envi-
ronmental quality. Furthermore, mature democracies might under-provide
public goods because special interest groups have gained a disproportion-
ate inuence on the government (Midlarsky 1998; Olson 1982). Therefore,
democratic governments tend to rely on the provision of targeted private
goods to these interest groups instead of providing public goods to the whole
6
As Bueno De Mesquita et al. (2003) note: once autocratic leaders survive the rst
year in oce they usually stay in oce for a long period.
61
CHAPTER 3. THEORY
population.
A substantial part of the empirical literature, however, shows that democ-
racies tend to be better providers of environmental quality (Torras & Boyce
1998; Barrett 2000; Bernauer & Koubi 2009). From these ndings it can
be concluded that by providing resources such as knowledge and technology
the inuence of IGO membership and the funding oered by international
aid donors, should constitute a ready opportunity for democratic leaders
to provide higher levels of public goods, in this case environmental quality,
consequently increasing their chances of reelection.
The argument with respect to international aid is further supported by
Smith (2006) who shows using a game theory model that democratic politi-
cal leaders should convert foreign aid into a higher provision of public goods
whereas the ruling elite in autocracies should put most of the foreign aid in
their own pocket (Tullock 1993). In a similar vein, the higher transparency
of democratic regimes, as introduced above, can be an additional reason
why democracies use their international assistance for the provision of pub-
lic goods. Due to the higher transparency of democracies, donors are better
able to control the use of their funds or the implementation of regulations.
As donors are also more likely to observe the misuse of funds, they can
easily punish a country for non-compliance. Fearing negative consequences,
the country concerned will thus avoid such misuse in the rst place. Conse-
quently, due to their higher transparency, democracies are expected to use
their technological and nancial assistance to provide higher levels of public
goods in general and therefore better environmental quality in particular.
Overall, this rst argument of why democracies should respond dif-
ferently to the international system than autocratic countries states that
democracies should use the resources provided by international organiza-
tions and aid donors to provide higher levels of public goods.
Turning to the second argument, we see that democratic developing
countries might provide better environmental performance because tech-
nological and knowledge transfer might take place to a greater extent. As
discussed in detail in sections 3.3.1 to 3.3.4, all of the four facets of interna-
tional integration provide developing countries with newer and supposedly
more environmental friendly technology and know-how. Thus the capacity
of developing countries to provide environmental quality increases. How-
ever, as mentioned above, the intensity of this technology and knowledge
eect depends to some extent on whether these better technologies and
practices and the know-how on how to deal with environmental problems
are diused within a particular country or whether they remain conned to
a particular company (in the case of FDI and trade) or development project
(in the case of aid or IGO activity). Since democracies are arguably more
transparent and more open for their civil society (Remmer 1998; Gaubatz
1996), it is a reasonable conjecture that the diusion and spill-over of these
new technologies and environmental practices should happen more easily
62
CHAPTER 3. THEORY
within democracies. This could therefore lead to a stronger positive eect
of the various forms of international integration in democracies compared to
autocracies.
Taken together, these two arguments imply that overall, democracies
should be more responsive to the positive inuences of IGOs, international
aid donors and foreign rms, leading to a better provision of environmental
quality given a countrys extent of international integration. The reason
why a democratic political system should intensify the positive eect of
international integration is, rst, that democracies should use the resources
available to them from their international ties for a higher provision of public
goods. Second, democratic political systems might be more conducive to
knowledge transfer and technological spill-over, thus increasing the positive
eect of the four aspects of international integration on the environmental
performance of developing countries.
3.4.2 The Political System and the Degree of International
Integration
In addition to the fact that democracies might be more responsive to the
incentives emanating from the international system, democratic developing
countries might in general be better integrated into the international system,
thus indirectly intensifying the positive inuence of international integration
on environmental quality. In the next paragraphs I will therefore analyze
each of the four factors subsumed in this thesis under international inte-
gration to show that we can theoretically expect democracies to join more
IGOs, receive more international assistance and have more intense trade
and foreign investment ties. This then implies that democratic developing
countries should indirectly perform better with regard to the provision of
environmental quality through their higher level of international integration.
Starting with membership in IGOs, there are several reasons why demo-
cratic countries might join more IGOs than autocratic countries. Relying
on the argument that democratic countries tend to be more transparent
than autocracies, this higher transparency should in turn allow democra-
cies to more easily and credibly commit themselves to international coop-
eration (Remmer 1998; Gaubatz 1996). Similarly, one can posit that the
strong domestic legal institutions and traditions usually found in democra-
cies should increase their international credibility (Abbott & Snidal 2000).
Moreover, according to the literature on the democratic peace, democra-
cies, because they are supposedly able to solve their domestic conicts by
cooperation, are assumed to project this cooperative behavior also to the
international realm (Russett & Oneal 2001). In addition, through member-
ship in international organizations, governments could achieve benets for
their constituents, such as gains from increases in trade, which they would
be unable to achieve independently (Manseld & Pevehouse 2008). Finally,
63
CHAPTER 3. THEORY
membership in dierent international organizations allows governments to
blame their own poor performance on these international actors, therefore
hedging democratic governments from electoral punishment (Manseld &
Pevehouse 2008).
Turning to international assistance, we see that democratic countries
should not only be more likely to join international organizations, they are
also supposed to receive more international assistance. The reasoning be-
hind this is twofold: on the one hand, international organizations such as
the World Bank often include good governance in the set of criteria required
under the conditionality principle. Under such criteria, democratic countries
are supposed to be preferred to autocracies with regard to the receipt of in-
ternational assistance. On the other hand, because of the higher levels of
transparency in democratic countries, international organizations can more
easily monitor the assistance provided to them. The fact that they can track
these funds more easily is one reason why they should give more assistance
to democratic countries in the rst place. Empirical evidence suggests that
although the political system of a country does not seem to play a signicant
role in the allocation of bilateral aid, as shown in section 2.3.2, the alloca-
tion of multilateral aid seems to depend at least partially on a countrys
regime type. Hence at least where multilateral aid is concerned, democratic
developing countries should receive more funds compared to their autocratic
counterparts (Neumayer 2003c).
With regard to international economic integration, one could argue that
democracies have more intensive trade and investment ties. Concerning
trade, elected governments have an incentive to liberalize trade because lib-
eralizing trade is seen as welfare enhancing. Hence in order to increase their
chances of reelection, it is reasonable for governments to boost trade and
thus enhance the welfare of their constituency. This argument is supported
by Manseld et al. (2002) who show that democracies are more likely to sign
trade liberalization agreements. The authors reasoning behind this nding
is that by signing trade agreements, political leaders can signal to their vot-
ers that they are acting in the best interest of their respective countries
with regard to economic issues. In addition to these arguments, one can
further argue that political opponents in democracies would inform citizens
if governments were not adopting welfare enhancing trade policies, thereby
generating information which provides governments with an additional in-
centive to liberalize (Manseld et al. 2002; Kono 2006). These arguments
are empirically supported by various studies (Eichengreen & Leblang 2006;
Russett & Oneal 2001; Ingram et al. 2005; Doces 2006).
Finally, concerning the connection between foreign direct investment and
political regime type there are various arguments claiming both a positive
and a negative eect of democracy on foreign direct investment. The argu-
ment in favor of democracy is that democracies constitute a more credible
policy environment for MNCs than autocracies (Jensen 2003). In particular,
64
CHAPTER 3. THEORY
because voters could punish their leaders in the next election, breaking a
deal with a foreign investor implies greater audience costs for democratic
governments than for autocratic governments. Furthermore, policy change
might be less likely to occur in democracies because the typically higher
number of veto players limits the discretion of democratic leaders to change
investment policies at short notice. According to these arguments, democra-
cies should provide investors with greater security and therefore democracies
should be characterized by higher FDI inows (Jensen 2003; Choi & Samy
2008).
In contrast, autocratic countries might attract more investment because
the power of local businesses, trade unions and civil society is limited in au-
tocracies (Choi & Samy 2008). Hence governments in autocracies have the
ability to oer generous scal and nancial incentives to foreign investors
(Choi & Samy 2008, 86) since local businesses in autocracies cannot eas-
ily pursue protection against foreign capital and since investment deals are
usually not subject to demands from trade unions and pressures from civil
society to increase wages.
Empirical studies show that it is indeed the number of veto players that is
associated with the net ow of inward FDI and consequently that democratic
systems, which are in general characterized by a higher number of veto
players, should attract more foreign investment than autocratic countries
with fewer veto players (Choi & Samy 2008; B uthe & Milner 2008).
Altogether, one can deduce from these arguments that in general, democ-
racies seem to be more willing to engage internationally. Although the degree
to which this statement empirically holds true might vary with the type of
international integration under consideration more for IGO membership
and trade and less for aid and foreign investment in general, it clearly
follows that democracies should be better integrated into the international
system. Furthermore, since we have argued in sections 3.3.1 to 3.3.4 that
international integration inuences both the willingness and the capabil-
ity of governments in developing countries to provide environmental quality
positively, it follows that indirectly through the link of international inte-
gration, democracies should achieve better environmental performance than
their autocratic counterparts.
In sum, the various arguments presented in this and the previous sec-
tion suggest that democratic political systems should amplify the positive
eect of international integration because democracies may a) be more re-
sponsive to the inuences of the international system by using the resources
available to them through their international ties for a higher provision of
public goods, and b) in general be more inclined to engage internationally.
I therefore expect an interaction eect between a countrys political system
and the eect of international integration on the provision of environmental
quality:
65
CHAPTER 3. THEORY
H7: A democratic political system enhances the positive eect of
international integration on environmental quality.
In summary, the theoretical arguments of this thesis postulate that all
four forms of international integration membership in international or-
ganizations, international assistance, trade and foreign direct investment
inuence the environmental performance of developing countries positively
with a countrys political system acting as a mediating factor. In partic-
ular, I argue that international integration improves environmental quality
in developing countries by providing a channel through which these coun-
tries may receive technologies and resources necessary to reduce pollution.
Moreover, I argue that although the political system should not have an in-
dependent eect on the environmental performance of developing countries,
the positive eect of international integration is intensied in democratic
countries, implying an interaction eect between the four facets of inter-
national integration on the one hand and a countrys political system on
the other hand. This interaction eect captures the idea that at any given
level of international integration, environmental quality will be higher in
democracies relative to autocratic countries. This should be the case be-
cause democracies may, rstly, be more responsive to the inuences of the
international system by using the resources available to them through their
international ties to provide better environmental quality relative to their
autocratic counterparts and, secondly, because democracies may in general
be more inclined to engage internationally.
The next chapters will test the various hypotheses established in this
theoretical section, employing dierent strategies. First, chapter 5 aims to
test hypothesis 1 to 7 using a quantitative analysis of a maximum of 115
countries over 30 years. Chapter 6 then goes into more detail concerning the
relationship between one facet of international integration IGO member-
ship and environmental quality and attempts to determine which of the
theoretically proposed mechanisms underlying IGO membership do indeed
inuence environmental performance in developing countries. Finally, chap-
ter 7 establishes a formal model to analyze the impact of IGO membership
on environmental quality within the framework of a specic-factors trade
model.
66
Chapter 4
Research Design
4.1 General Research Strategy
This chapter explains the empirical research strategy underlying this thesis.
In particular, I rst present the general method of investigation used to
assess the relationship between international integration, the political system
and environmental quality in developing countries. This is followed by a
discussion of the sample selection and the statistical estimation method
appropriate for analyzing the data at hand. The chapter ends with a detailed
description of the variables and data used in the statistical analysis.
The main component of my empirical analysis consists of a quantitative
time-series cross-country analysis. Using this method allows me to analyze
whether dierences in the level of international integration and dierences
in the political system lead to a higher or lower provision of environmental
quality in developing countries. Furthermore, the time component of my
analysis makes it possible to investigate how changes in the level of integra-
tion or changes in the political system of a particular country aect that
countrys environmental performance. Hence the time-series cross-sectional
structure of the data allows a more rigorous test of my theoretical arguments.
After presenting the main results of this time-series cross-country anal-
ysis in chapter 5, this thesis delves into one specic factor of international
integration in more detail: membership in international governmental orga-
nizations. Since this is the only factor of international integration that is
robustly associated to an increase in environmental quality, as will become
apparent in chapter 5, I aim to understand more fully the mechanisms that
are behind these ndings. To this end, chapter 6 presents further time-series
cross-country analyses using a more rened coding of IGO membership to
assess which of the mechanisms proposed in chapter 3 in section 3.3.1 are
indeed behind this positive impact of IGO membership on environmental
quality in developing countries.
This detailed quantitative analysis of IGO membership on environmen-
67
CHAPTER 4. RESEARCH DESIGN
tal quality is further complemented by four illustrative case studies. These
case studies serve to assess in more detail which of the mechanisms proposed
in the theoretical section and resulting from the statistical analysis are em-
pirically at play in the everyday business of IGOs in developing countries.
The case selection process and the exact method of analysis of these case
studies are described at the beginning of the case study analysis in section
6.4 in chapter6.
4.2 Sample Selection
Based on the theoretical arguments, the sample consists of all developing
countries of the world. Every year, the World Bank publishes a classication
of countries into low-income, lower-middle income, upper-middle income and
high-income (World Bank 2009b). Following this classication, I include all
countries in my sample that are characterized as non high-income. Since the
classication threshold varies each year, the highest threshold, which is the
one in the year 2000, is chosen. Hence all countries with a GNI per capita
level
1
of less than USD 9,266 are included in the sample, which results in a
total of 115 countries. However, in order to test the robustness of my results,
all statistical models are re-estimated using the samples that include only
all countries up to the lower-middle income (USD 2,995) and up to the low-
income (USD 755) classication thresholds respectively. In those cases in
which the results are sensitive to the sample specication, they are presented
below, otherwise only the results for the sample including all countries with
GNI per capita of less than USD 9,266 are shown.
Furthermore, all results presented in this thesis exclude Somalia, Ethiopia,
Cambodia, Lebanon and Bosnia-Herzegovina. The reason for this is simply
that the data coverage for these countries is very thin, leading to the prob-
lem that there are no common time periods between these ve countries and
all the other countries in the sample. Leaving the countries in the sample
would have implied that all regressions would have to be estimated using
pairwise instead of casewise inclusion
2
. Unfortunately, the data coverage
for BOD is even worse than compared to SO
2
and CO
2
emissions, which
implies that in the case of BOD there is no other option than to estimate
all regression models using pairwise instead of casewise inclusion
3
.
The time span of the analysis covers the years 1970 to 2000. Due to
data availability, the time span of the analysis is reduced to the years 1980
1
The World Bank calculates yearly GNI per capita levels using its Atlas method to
diminish the impact of exchange rate uctuations on the cross-country comparison of
national incomes.
2
The results are, however, not sensitive to the inclusion of these ve countries and
estimation of the regressions using pairwise regression.
3
For reasons of comparability, the ve countries remain excluded for the models using
BOD as the dependent variable.
68
CHAPTER 4. RESEARCH DESIGN
to 2000 in the case of water quality
4
.
In order to provide an overview of the countries in the sample and the
time periods for which data are available, tables 9.1 to 9.7 in the Appendix
provide this information. Data coverage is reported for the three dependent
variables SO
2
, CO
2
and BOD separately. In addition, tables 9.1 to 9.7
show whether a country belongs to the low-income, lower-middle income or
upper-middle income group.
4.3 Estimation Method
In order to empirically test the implications of my theoretical arguments I
use a xed-eects estimation procedure with panel correct standard errors
(PCSE) and a Prais-Winsten specication to control for autocorrelation.
The reasoning behind this choice is elaborated in detail below.
Due to the time component of time-series cross-section (TSCS) models
5
various diculties arise with respect to the model estimation compared to
the cross-section context of regression (Beck & Katz 1995, 1996; Pl umper
et al. 2005). I will discuss these diculties successively to derive, step by
step, the estimator used in the empirical analysis.
One ordinary least squares (OLS) regression assumption that is fre-
quently violated in the context of TSCS models is the assumption of in-
dependent and identically distributed errors as the errors mostly suer from
panel heteroskedasticity and serial correlation. Although OLS estimates are
still unbiased in the presence of autocorrelation or heteroskedasticity, they
are inecient, resulting in the case of positive autocorrelation, for example,
in overly optimistic standard errors (Wooldridge 2003).
In particular, panel heteroskedasticity implies that the error variance
changes from unit to unit (i.e. from panel to panel), whereas the variance
within each panel is constant (Beck 2001). One way to control for this
panel heteroskedasticity is to use panel correct standard errors (PCSE) as
4
The average time period for each country in the data is 24 years in the case of SO
2
and CO
2
and 13 years in the case of BOD.
5
The terms time-series cross-section data and panel data are usually used interchange-
ably in the literature, a convention I also adopt for this thesis. In general, the dierence
between panel and TSCS data is that for panel data the number of cross-sectional ob-
servations N is clearly larger than the time component T whereas for TSCS data the
time component is usually larger than the number of countries or individuals N (Beck
& Katz 1995). Hence in a panel context the asymptotic behavior of the estimators is
derived thinking of N becoming large. In contrast, for TSCS data the asymptotics lie in
T. Although my analysis consists of more countries N than time periods T, the number
of countries N in the sample at hand is xed to all developing countries. Thus we do not
sample our countries from a larger population but restrict ourselves to drawing inferences
about the units we observe (Beck 2001). Hence for the data at hand it is appropriate to
see the asymptotic behavior of the corresponding estimators lying in T. Consequently, it
would be more correct to refer to my analysis as a TSCS than as a panel data model.
69
CHAPTER 4. RESEARCH DESIGN
developed by Beck & Katz (1995). Using Monte Carlo simulations they
show that for T > 15 PCSE are very accurate and that even in the case
of no panel heteroskedasticity PCSE do not perform worse than plain OLS
standard errors. Consequently, there are good reasons to use panel correct
standard errors when analyzing time-series cross-section data to control for
potential panel heteroskedasticity.
Turning to autocorrelation
6
, there are two common approaches to deal
with serial correlation: OLS with PCSE and a lagged dependent variable
and the autoregressive, AR(1), model (Beck & Katz 1995, 1996; Pl umper
et al. 2005). Using Monte Carlo experiments Beck & Katz (1996) shows that
the approach using a lagged dependent variable instead of the AR(1) model
specication seems to be preferable with regard to eciency and accuracy
of the variability estimation.
Compared to this potential advantage of using a lagged dependent vari-
able, there are several serious disadvantages connected with this approach.
A rst severe problem that arises by introducing a lagged dependent variable
in the model is that in the case including unit dummies (country xed ef-
fects), the coecients are no longer consistent (Wooldridge 2001). Another
problematic feature of using the lagged dependent variable to control for
serial correlation is that including a lagged dependent variable is the same
as if we were to include a weighted average of the right-hand side variables
(Pl umper et al. 2005). Therefore, due to the inclusion of a lagged depen-
dent variable, the coecients of the theoretically interesting independent
variables can only be interpreted as the contemporaneous and not as the
general eect of the right-hand side variables on the dependent variable.
Furthermore, the inclusion of a lagged dependent variable usually captures
a large part of the variation in the dependent variable. This in turn leads to
an underestimation of the eect of the theoretically interesting independent
variables, resulting in often implausible small and insignicant coecients
(Achen 2000; Pl umper et al. 2005). Consequently, Pl umper et al. (2005)
propose using the AR(1) model instead since the AR(1) model is still con-
sistent in the presence of unit eects and it also corrects for autocorrelation
of the errors while leaving the interpretation of the coecients unchanged.
In addition:
In principle, AR1 error models tend to absorb less time-series
dynamics and may therefore be the method of choice for ap-
plied researchers seeking to explain not only cross-sectional vari-
ance and cross-sectional dierences in changes, but also average
changes in levels (Pl umper et al. 2005, 343).
Taking all these arguments into account, I decided to use the AR(1) model
as the main estimation specication since the approach of including a lagged
6
Using a Lagrange multiplier test shows that the models estimated without any cor-
rection do indeed suer from autocorrelation (Engle 1984; Beck & Katz 1996).
70
CHAPTER 4. RESEARCH DESIGN
dependent variable is aicted with too many problems and disadvantages
7
.
However, in addition to the AR(1) model, I also provide the results using
the lagged dependent variable approach in the section on robustness checks
in chapter 5.
Concerning the question of how to exactly estimate the AR(1) model,
Beck & Katz (1995, 1996) argue that it is better to assume that serial cor-
relation follows a common parameter for all panels instead of a serial cor-
relation parameter that varies by unit. The reasons behind this suggestion
are twofold: rst, they argue that if using time-series cross-section data we
assume that we can pool together our data and hence that all panels follow
a common set of the parameters of interest . They further argue that if
we assume that is common to all panels why should the serial correlation
parameters be the only parameters allowed to vary by units? Second, they
show using Monte Carlo simulation that
[...] the assumption of a common serial correlation process leads
to superior estimates of even when the data are generated with
diverse, unit-specic
i
(Beck & Katz 1996, 6).
In addition to panel heteroskedasticity and correlation of the errors, it is
important to deal with the question of how to model heterogeneity between
units. Since it is unlikely that my econometric model captures all unit
heterogeneity, there are good reasons to include unit eects in the statistical
analysis. The inclusion of unit eects is further indicated by a simple F-test,
which shows that the unit eects cannot assumed to be zero
8
. Leaving out
unit eects is especially problematic if the unit eects are correlated with the
other independent variables. In this case the estimation of the coecients on
those other variables can be aicted with both inaccurate point estimates
and inaccurate error estimates, implying that OLS is no longer unbiased
under these circumstances (Green et al. 2001; Kristensen & Wawro 2003;
Pl umper & Troger 2007).
One approach to model these unit eects is to use xed eects, which,
instead of assuming a common intercept for all units, allows each unit to have
its own intercept (Beck 2001; Hsiao 2003; Wilson & Butler 2007). Instead
of treating these unit eects as a xed eect, one could also conceive of
this unit-specic eect as a random variable. On the assumption that the
unit eects are not correlated with the explanatory variables, using random
eects will yield an unbiased and more ecient estimation than xed eects
(Hsiao 2003). However, if the assumption is violated, estimates will be
7
More precisely, I use a Prais-Winsten specication, which in contrast to a Cochrane-
Orcutt estimation procedure does not throw away the rst observations of each panel.
8
The value for the baseline model using SO
2
emissions as the dependent variable as
displayed in table 5.2 in chapter 5 is
2
= 4200000 which corresponds to a p-value of
0.00. The values in the case of CO
2
are
2
= 230000 and p = 0.00 and in the case of BOD

2
= 15785 and p = 0.00.
71
CHAPTER 4. RESEARCH DESIGN
biased. Using a Hausman test clearly indicates that random eects are not
appropriate in this context
910
.
Relying on xed eects estimation though comes at the price of not being
able to estimate the eect of time-invariant variables as they are perfectly
co-linear with the unit eects this implies, for example, that we cannot
include a dummy variable in the model to measure whether a country is
an oil exporter. In the case of slowly changing variables such as the polit-
ical system of a country, using xed eects leads in principle to unbiased
estimates. However, if the within variation is heavily dominated by the be-
tween variation, the standard errors for these variables are often rather high
as these variables are still strongly correlated with the unit eects (Troger
2008). However, in the case of the data used in this analyses, all of the
independent variables with the exception of GDP per capita and population
density have a within variation that is close to the between variation as
can be seen in table 4.1 and thus xed eects estimation is the appropriate
choice (Troger 2008).
In addition to including unit xed eects, one could also introduce time
xed eects. These time dummies are often incorporated in panel models to
control for those exogenous shocks that inuence all countries alike at one
specic point in time but for which we do not have a clear measurement.
However, introducing these time dummies usually absorbs a great deal of
the variance in the dependent variable without theoretically explaining these
interesting time variations (Pl umper et al. 2005) - this is essentially the
same problem as with the introduction of a lagged dependent variable as
described above. Nevertheless, since it is important to control for time in
a TSCS setting, I include a linear time trend in all of the models presented
below. Furthermore, in order to show that the results are not sensitive to
the inclusion of time dummies I present the estimates for my baseline model
9
With a
2
value of 14.91 (which corresponds to a p-value of 0.09) the Hausman
test rejects the use of random eects in the baseline model using SO
2
emissions as the
dependent variable as displayed in table 5.2 in chapter 5. The values in the case of CO
2
are
2
= 168.85 and p = 0.00 and in the case of BOD
2
= 294.97 and p = 0.00.
10
In the case of non-correlation between the unit eects and the explanatory variables,
using random eects is more ecient. However, in the case of correlation, random eects
are no longer unbiased. Hence the Hausman test essentially checks whether the coecients
of the random eects signicantly deviate from the coecients of the xed-eects model,
which are supposedly always unbiased. If the null hypothesis can be rejected, random
eects are biased and one should rely on xed-eects whereas if the null hypothesis cannot
be rejected random eects are unbiased and more ecient. However, as Tr oger (2008)
shows, there are cases in which the Hausman test rejects the null hypothesis not because
random eects are biased but because the point estimates of the xed eects model are
unreliable due to slowly changing explanatory variables. Since all of my variables, however,
display a large amount of within variation see table 4.1 I have some condence in the
results of the Hausman test. Furthermore, I used an autocorrelation and heteroskedastic
robust version of the Hausman test, which also indicates that the null hypothesis of the
appropriateness of random eects should be rejected (Hoechle 2007).
72
CHAPTER 4. RESEARCH DESIGN
Table 4.1: Variables and their variation sample: all developing countries
Overall Between Within
Variation Variation Variation
IGO membership 255.7 159.0 119.9
Aid per capita 7192.7 4062.8 3008.5
ln trade openness 0.7 0.5 0.3
FDI % GDP 15.1 2.6 13.3
Polity 48.0 31.7 18.9
GDP pc 7547162.8 7402807.1 744044.3
Population density 8772.2 7728.2 304.2
Growth 0.01 0.003 0.008
also including time xed eects - see section 5.3.3 in chapter 5.
Following from this detailed discussion of the various statistical estima-
tion methods designed to analyze panel data, we can conclude that a xed-
eects estimation procedure with panel correct standard errors (PCSE) and
a Prais-Winsten specication to control for autocorrelation is the most suit-
able estimation procedure for the data at hand. Furthermore, the section
on robustness checks in chapter 5 provides the results using other possible
estimation procedures to show that the results are indeed not sensitive to
the statistical estimation procedure.
4.4 Data
After having discussed the method of analysis, I now turn to the opera-
tionalization of the theoretical concepts underlying my arguments and to
the description of the data used in the analysis. Table 4.2 summarizes the
indicators and the data sources for all the variables.
4.4.1 Dependent Variable
In the literature, many dierent measures of the dependent variable, namely
environmental quality, are employed, such as indicators of air quality (Bernauer
& Koubi 2009; Gassebner et al. 2006; Neumayer 2003a; Grossman & Krueger
73
CHAPTER 4. RESEARCH DESIGN
1995), water quality (Gassebner et al. 2006; Sigman 2002), or indices which
combine several measures of environmental quality, such as the environmen-
tal sustainability index or the genuine savings index (Ward 2006; DeSoysa &
Neumayer 2005). To refer to a broader picture of environmental quality, this
analysis uses dierent measures covering the most important aspects of en-
vironmental quality. However, availability of data that exist as time-series
cross-section data
11
limits the choice to sulfur dioxide (SO
2
) and carbon
dioxide (CO
2
) emissions measuring air quality, and to biological oxygen de-
mand (BOD) measuring water quality
12
. Altogether, using these dierent
indicators allows us, rst, to refer to a broad picture of environmental qual-
ity and, second, it serves as a robustness check of the results as we can test
whether our theoretical arguments hold with regard to all three indicators,
which refer to dierent types of environmental pollution.
4.4.1.1 SO
2
SO
2
is a primary contributor to acid rain and is thus detrimental to both
human health and the well-being of the ecosystem. It mostly arises from
burning fossil fuels such as coal or oil and is mainly emitted by petroleum
reneries, metal processing facilities, locomotives and large ships.
Stern (2005) provides data on SO
2
emissions using an econometric emis-
sions frontier model to estimate the data using a countrys prole with regard
to energy use, economic structure and metal smelting. The obtained esti-
mates were compared to measured SO
2
emissions data for those countries
and years for which these data existed. The t between real and estimated
data is very good, indicating the precision of the estimated SO
2
data
13
.
11
Unfortunately, data availability is a key issue for environmental data. Several inter-
esting indicators such as the percentage of the population with access to sanitation and
safe drinking water or deforestation rates are only available as cross-sectional data and
are thus unsuitable for this analysis.
12
Although environmental performance indices exist, their use for time-series cross-
section analysis is limited. For example, the composition of the environmental sustain-
ability index changes from year to year, rendering it non-valuable for time-series analysis.
Moreover, the concept of genuine savings is intended to measure the true rate of savings in
an economy after taking into account investments in human capital, depletion of natural
resources and damage caused by pollution (Bolt et al. 2002). Although it can be seen as a
measure of weak sustainability (DeSoysa & Neumayer 2005), its environmental component
is very small, implying that genuine savings are not a valuable indicator of environmental
performance per se.
13
For several countries the SO
2
data needed to be adjusted. Stern (2005) provides
separate estimations for the Czech Republic and Slovakia from 1980 onwards. Since these
two countries have only been independent from each other since 1993, I have added Czech
and Slovak emissions to give a single gure for Czechoslovakias SO
2
. Similarly, I added
Macedonias, Croatias, Bosnia-Herzegovinas and Slovenias emissions to the emissions of
Yugoslavia prior to these countries independence. Since Stern (2005) provides separate
estimates for European Turkey and Asian Turkey as well as European Russia and Asian
Russia, I created a joint estimate for both Turkey and Russia. For Vietnam and Yemen,
74
CHAPTER 4. RESEARCH DESIGN
4.4.1.2 CO
2
CO
2
is largely a by-product of energy production and use. It is released from
the combustion of fossil fuels and cement manufacturing, for example (World
Bank 2006). CO
2
emissions are a main contributor to global warming and
climate change since it accounts for the major share of greenhouse gases
(World Bank 2006). The data on CO
2
emissions come from the World
Banks Development Indicators (World Bank 2006).
Both SO
2
and CO
2
emissions are measured in thousand metric tons per
capita in order to capture size eects (see Gassebner et al. 2006). I use the
natural log of both variables because of their long right tails.
4.4.1.3 BOD
Biological oxygen demand (BOD) is a proxy for organic water pollution
(World Bank 2006). In particular, BOD measures the amount of oxygen
consumed by water bacteria in order to break down waste such as bacteria,
toxic chemicals, metals and organic matter. The BOD data as provided by
the World Bank Development Indicators are constrained/conned to emis-
sions of organic pollutants arising from industrial activities, which constitute
a major cause of water quality degradation. More specically, the data are
based on a study on industrial emissions by Hettige et al. (2000) in which
the authors used actual measurement of water pollution at the plant level.
Combining these data with sector-level output and employment data, Het-
tige et al. (2000) then estimated BOD loads over time and across countries
in kilograms per day. According to the World Bank (2006, 153), the data
on BOD are quite reliable because a) many countries have pollution control
programs regulating water pollution and thus providing some data on water
pollution and b) because measuring water pollution is less expensive and
better understood than the measurement of other forms of pollution.
BOD is measured in kilograms per day per capita to capture size eects
and again, the natural log is used due to the long right tail distribution. All
indicators are measured in levels instead of growth rates for several reasons.
Since all my arguments refer to the state of environmental quality and not
to changes in environmental quality I consider levels to be more appropriate.
In addition, changes could be misleading in the sense that the growth rate
of pollution may be high, implying bad environmental quality, whereas the
overall state of the environment is still good. In contrast, growth rates
may be low and therefore hinting at very good environmental quality while
in reality the state of the environment is quite bad. Furthermore, since all
Stern (2005) only gives estimates for the single countries, although Yemen was divided for
the period between 1967 until 1990 and Vietnam between 1954 until 1975. Hence in the
case of Vietnam only the northern part is included in the analysis for the period in which
the country was divided. For Yemen, only the southern part is included in the analysis
during the period in which the country was divided.
75
CHAPTER 4. RESEARCH DESIGN
three variables are stationary
14
I think it is more appropriate to use the level
specication instead of dierencing the dependent variables. Furthermore,
using dierenced indicators implies that one can only model the short-term
eects (Beck & Katz 1996, 10), which is of only limited interest in this
analysis.
4.4.2 Independent Variables
The six independent variables of my analysis are membership in interna-
tional governmental organizations, international aid, trade openness, for-
eign direct investment inows, a countrys political system and its national
income. I will discuss the operationalization and measurement of each in-
dependent variable in consecutive order.
4.4.2.1 International Governmental Organizations
A countrys membership in international governmental organizations is op-
erationalized as the yearly cumulative count of each countrys IGO mem-
bership. The use of this cumulative measure, as opposed to the use of the
number of new memberships in a certain year, for example, is appropriate as
the cumulative measurement acknowledges that being a member of a certain
number of IGOs limits a countrys options for further membership.
The data come from the Correlates of War Project (COW) (Pevehouse
et al. 2004)
15
. More precisely, according to Pevehouse et al. (2004) this
includes all organizations that:
1. consist of at least three members of the COW-dened state system
2. hold regular plenary sessions at least once every ten years
3. possess a permanent secretariat and corresponding headquarters.
The Correlates of Wars IGO dataset oers three distinct variables of IGO
membership, which are dierentiated according to whether a state has full,
associated or observer membership. In general, I use the most inclusive (or
least stringent) variable, the one that includes full, associated and observer
membership. I further check the robustness of the results by using both
the most narrowly dened variable, which includes only full membership,
and the variable that includes both full and associated membership while
excluding observer status.
14
Using a Fisher test for unit roots in non-balanced panel data shows that all of the
dependent, independent and control variables are stationary with the exception of the
log of BOD per capita which is trend stationary. This implies that all variables can be
integrated into the model without the need for dierencing them.
15
This measure excludes emanations of IGOs.
76
CHAPTER 4. RESEARCH DESIGN
To better assess the theoretical mechanisms that link IGO membership to
the provision of environmental quality in developing countries, this general
IGO variable is disaggregated with regard to its degree of institutionalization
and with regard to its function. Concerning its degree of institutionalization,
the eect of IGO membership on environmental quality is further evaluated
in order to see whether more institutionalized IGOs may be more eective
than IGOs that only exist on paper. The data come from Boehmer et al.
(2004). In particular, Boehmer et al. (2004) code three degrees of institu-
tionalization: minimal, structured and interventionist IGOs. Minimal IGOs
are organizations that lack any bureaucratic, executive, or judicial organs
that possess any formalized power (Boehmer et al. 2004, 37). Structured
IGOs, by contrast, are organizations to which member states relinquish
minimal amounts of their sovereignty to support IGO projects and mis-
sions (Boehmer et al. 2004, 37) and interventionist IGOs are organizations
that possess clear mechanisms for coercing or inuencing state behavior
(Boehmer et al. 2004, 38).
Another way to disaggregate multiple IGO membership is according to
the specic function of the IGO. The aim is to assess which types of IGOs
are indeed associated with an increase in environmental quality. This eort
of aggregating membership according to the function of the organization
allows the mechanisms behind the inuence of IGO membership to be eval-
uated. Particular functions of IGOs are connected with specic theoretical
mechanisms, i.e. umbrella organizations such as the EU are potential or-
ganizations that allow for issue linkage whereas organizations such as the
Food and Agriculture Organization (FAO) are more inclined towards trans-
fer technology.
One existing study that disaggregates IGO membership along the func-
tion of the dierent organizations is Ingram et al. (2005). In particular,
Ingram et al. (2005) code the mandate or purpose of an IGO into four
broad categories:
1. general purpose (these are umbrella organizations characterized by
multiple functions and aiming at the communication between and the
administration of governments (Ingram et al. 2005, 854); examples
are the UN, EU, Organization of American States)
2. military/political organizations (these are military or regional political
alliances; an example is NATO)
3. economic organizations (these are all organizations with an economic
function; examples are the World Bank, IMF)
4. social organizations (these are all environmental, health, social welfare
and cultural organizations; examples are WHO, FAO).
77
CHAPTER 4. RESEARCH DESIGN
In a second step both the economic and social IGOs are further dierenti-
ated. The economic IGOs are coded into four categories:
1. monitoring, surety and general economic: organizations that deal with
one of the following issues: property rights, trade relations, monitoring
and enabling international transactions
2. standardization and harmonization: organizations that ensure and
harmonize international standards
3. cooperation and development: organizations that advocate develop-
ment
4. industry specic: organizations that are tailored to particular needs
of a specic industry.
The social IGOs are coded into three categories:
1. environmental organizations
2. general social organizations including organizations dealing with health,
social welfare, humanitarian or cultural issues
3. education and research organizations.
Unfortunately, Ingram et al. (2005)s coding only covers the time period
until 1992. Furthermore, they subsume nuclear and agrarian IGOs, which
are often characterized by strong environmental provisions, under social or
economic IGO mandates. I therefore decided to recode all IGOs according
to their function while using the coding instructions described in Ingram
et al. (2005) as a blueprint. Consequently, following the example and the
coding instructions of Ingram et al. (2005), I have coded all IGOs accord-
ing to their mandate into 10 dierent categories: military (e.g. NATO),
umbrella (e.g. ASEAN, EU), economic cooperation (e.g. World Bank), eco-
nomic general (e.g. WTO, IMF), economic standardization (e.g. Caribbean
Postal Union), economic industry specic (e.g. Cocoa Producers Alliance),
social (e.g. WHO, ILO, UNESCO), environmental (e.g. Global Environ-
ment Facility, GEF), nuclear (e.g. IAEA), and agricultural (e.g. FAO)
organizations.
In a similar attempt to Ingram et al. (2005), Gartzke et al. (2006) dis-
aggregate international organizations according to their mandate into po-
litical, economic, social and cultural organizations. Although the Gartzke
et al. (2006) data cover the years until 2000, I still decided to undertake
the additional coding eort because their data rely on only four categories
whereas Ingram et al. (2005)s coding originally consists of 9 categories, al-
lowing for a more detailed coding. However, I use Gartzke et al. (2006)s
data to cross-check the robustness of my own coding regarding the functional
disaggregation of IGOs.
78
CHAPTER 4. RESEARCH DESIGN
4.4.2.2 International Aid
Besides the industrialized countries in general, multilateral development or-
ganizations such as the World Bank are the main actors providing inter-
national assistance to developing countries. As most actors like the World
Bank or the EU nance dierent projects and loans in a given year, the
problem arises of how to measure the joint inuence of these single assis-
tances. The approach followed in this dissertation is to use the total amount
of money each country receives per year in the form of bilateral and multi-
lateral development aid.
The data stem from OECD (2008). This data set includes nancial
ows to developing countries that are provided by ocial agencies, including
state and local governments, or by their executive agencies (OECD 2008).
Furthermore, nancial ows are only characterized as development aid if
they are concessional in nature, convey a grant element of at least 25% and
if they are administered with the promotion of the economic development
and welfare of developing countries as the main objective (OECD 2008).
More precisely, the aid data constitute net ows, which equal total new
ows minus amounts received such as repayments of principal, osetting
entries for debt relief, repatriation of capital, and occasionally recoveries on
grants or grant-like ows (OECD 2008). All gures are expressed in million
constant USD (2000 prices).
The yearly aid gures are scaled by the population in each country to
capture size eects. Another possibility would be to scale international aid
not by a countrys size in terms of population but by its size in terms of
GDP as in Henisz et al. (2004, 26), for example. The rst approach aims
to measure how much development aid a country has at its disposal per
citizen, whereas the second approach measures how much development aid
a country has at its disposal compared to its economic performance. As the
theoretical arguments center around the provision of public goods, I consider
it more meaningful to scale aid by population to capture the volume of
additional resources a country has available to provide its population with
those public goods. However, the results using aid scaled by a countrys
GDP are additionally provided in the results section.
4.4.2.3 Trade
As is common in the literature, a countrys trade level is measured by its
yearly ratio of the sum of its exports and imports scaled by a countrys
GDP to account for the size of the economy (Gleditsch 2002; Frankel 2003).
The natural log is used due to the long right tail distribution. Since the
scale eect underlying the trade-environmental nexus refers to an increase
in economic activity induced by trade, it refers to exports only. Therefore,
I test the robustness of the results using exports divided by GDP instead of
79
CHAPTER 4. RESEARCH DESIGN
both exports and imports divided by GDP. The data come from Gleditsch
(2002) and are an updated version of the Penn World Tables trade gures.
This updated version is used as it provides estimates for the missing data
common in the original data set.
4.4.2.4 Foreign Direct Investment
Foreign direct investment is measured by FDI inows as a percentage of
GDP to capture market size (B uthe & Milner 2008; Choi & Samy 2008). The
data are taken from World Bank (2006)
16
. Since all theoretical arguments
concerning the eects of foreign direct investment refer to investment that
is provided by a foreign owner and not to the value of foreign capital in a
specic country, I use FDI inows instead of the stock of FDI. More precisely,
the variable measures the net inows to acquire a lasting management
interest in an enterprise operating in an economy other than that of the
investor. It is the sum of equity capital, reinvestment of earnings, and other
short- and long-term capital, as shown in the balance of payments (World
Bank 2006, 319).
4.4.2.5 Political System
The political system of a country is operationalized using the combined
Polity IV (Marshall & Jaggers 2002) score ranging from 1 institutional-
ized autocracy to 21 institutionalized democracy
17
. Polity IV captures
the competitiveness of political participation, the guarantee of openness and
competitiveness of executive recruitment, and the existence of institutional-
ized constraints on the exercise of executive power.
Since Polity IV has been criticized with regard to diculties concerning
the aggregation procedure (Munck & Verkuilen 2002), I use the dichotomous
index of democracy developed by Alvarez et al. (1996) (ACLP index) for a
robustness check. According to this index, democracy is dened if both the
executive and the legislature are chosen by contested election (Przeworski
et al. 2000). The original measure of Alvarez et al. (1996) covered the years
until 1990 only; however, Cheibub (2004) and Cheibub & Gandhi (2004)
have updated the data. I decided not to use the ACLP index as the main
measure of a countrys political system because being a dummy variable,
16
Another possible source of data on foreign direct investment is UNCTADs Handbook
of Statistics (B uthe & Milner 2008). However, the data from UNCTAD contain fewer
observations than the data from the World Bank. As the main results do not change
when using the UNCTAD FDI data I rely on the World Development Indicators to ensure
the most inclusive sample size.
17
The original Polity IV score ranges from -10 most autocratic to 10 most demo-
cratic. To facilitate interpretation of the interaction eects, the variable is transformed
to be positive throughout its complete range. This innocuous change does not aect the
results.
80
CHAPTER 4. RESEARCH DESIGN
it only distinguishes between democracies and autocracies per se and thus
masks the variation in regime type between these two extremes
18
.
In order to capture the interconnection between a countrys political sys-
tem and its international integration, the political system variable is inter-
acted with the four dierent variables measuring international integration.
4.4.2.6 National Income
To test for the eect of national income on pollution, the log of GDP per
capita is included in the model. Although the logic of the Environmen-
tal Kuznets Curve suggests that both GDP per capita and its square term
should be included in the regression model, doing so results in statistically
insignicant estimated coecients for both variables. This is due to the fact
that the analysis is restricted to a particular sample of countries. Looking
only at developing countries, we should observe a negative eect of income
on environmental quality as hypothesized in the theoretical section be-
cause none of the countries has yet reached the turning point after which
income is supposed to be associated with increasing environmental qual-
ity
19
. Therefore, using a quadratic specication of GDP per capita does not
capture the linear relationship between national income and environmental
performance for the countries included in this sample. Consequently, I only
include the linear term in all of the statistical models. The data on national
income come from Gleditsch (2002).
18
Of course, many more measures of democracy exist. However, most of them are either
limited to a specic time period or considered to be problematic measures of democracy.
For example, indicators such as the democracy measures of Bollen (1993); Coppedge &
Reinicke (1991); Hadenius (1992) are restricted to a particular year or to certain countries,
which makes them unsuitable for this analysis. Other indices like Freedom House (2006);
Gasiorowski (1996); Vanhanen (2000) are acknowledged to exhibit several problems with
regard to conceptualization, measurement, and aggregation (Munck & Verkuilen 2002,
28), which makes it advisable not to rely on those indicators. Another potential indicator
would be the Logic of Political Survival data (Bueno De Mesquita et al. 2003). I refrain
from using these data due to their slightly dierent focus compared to the other indicators.
This democracy indicator measures the size of the winning coalition, i.e. the number of
people needed to come or to stay in oce, relative to the size of the selectorate, i.e. those
people whose endowments include the qualities or characteristics institutionally required
to choose the governments leadership (Bueno De Mesquita et al. 2003, 42). Therefore,
this measure focuses on the inclusiveness of a political system while not reecting questions
such as the openness or competitiveness of executive recruitment.
19
Estimating a quadratic model of GDP per capita on environmental quality using all
countries, i.e. both high- and low-income countries, yields a turning point of USD 23,192
in the case of SO
2
and turning points that are out of sample in the case of CO
2
and BOD.
81
CHAPTER 4. RESEARCH DESIGN
T
a
b
l
e
4
.
2
:
T
h
e
o
r
e
t
i
c
a
l
c
o
n
c
e
p
t
s
a
n
d
t
h
e
i
r
o
p
e
r
a
t
i
o
n
a
l
i
z
a
t
i
o
n
C
o
n
c
e
p
t
I
n
d
i
c
a
t
o
r
D
a
t
a
S
o
u
r
c
e
E
n
v
i
r
o
n
m
e
n
t
a
l
Q
u
a
l
i
t
y
l
o
g
o
f
S
O
2
e
m
i
s
s
i
o
n
s
p
e
r
c
a
p
i
t
a
(
a
i
r
q
u
a
l
i
t
y
)
S
t
e
r
n
(
2
0
0
5
)
E
n
v
i
r
o
n
m
e
n
t
a
l
Q
u
a
l
i
t
y
l
o
g
o
f
C
O
2
e
m
i
s
s
i
o
n
s
p
e
r
c
a
p
i
t
a
(
a
i
r
q
u
a
l
i
t
y
)
W
o
r
l
d
B
a
n
k
(
2
0
0
6
)
E
n
v
i
r
o
n
m
e
n
t
a
l
Q
u
a
l
i
t
y
l
o
g
o
f
B
O
D
p
e
r
c
a
p
i
t
a
(
w
a
t
e
r
q
u
a
l
-
i
t
y
)
W
o
r
l
d
B
a
n
k
(
2
0
0
6
)
M
e
m
b
e
r
s
h
i
p
i
n
I
G
O
s
Y
e
a
r
l
y
c
o
u
n
t
o
f
I
G
O
m
e
m
b
e
r
s
h
i
p
P
e
v
e
h
o
u
s
e
e
t
a
l
.
(
2
0
0
4
)
M
e
m
b
e
r
s
h
i
p
i
n
I
G
O
s
Y
e
a
r
l
y
c
o
u
n
t
o
f
I
G
O
m
e
m
b
e
r
s
h
i
p
d
i
s
a
g
g
r
e
g
a
t
e
d
b
y
f
u
n
c
t
i
o
n
o
f
I
G
O
O
w
n
c
o
d
i
n
g
a
n
d
G
a
r
t
z
k
e
e
t
a
l
.
(
2
0
0
6
)
M
e
m
b
e
r
s
h
i
p
i
n
I
G
O
s
Y
e
a
r
l
y
c
o
u
n
t
o
f
I
G
O
m
e
m
b
e
r
s
h
i
p
d
i
s
a
g
g
r
e
g
a
t
e
d
b
y
i
n
s
t
i
t
u
t
i
o
n
a
l
i
z
a
-
t
i
o
n
o
f
I
G
O
B
o
e
h
m
e
r
e
t
a
l
.
(
2
0
0
4
)
I
n
t
e
r
n
a
t
i
o
n
a
l
A
s
s
i
s
t
a
n
c
e
Y
e
a
r
l
y
a
m
o
u
n
t
o
f
b
i
l
a
t
e
r
a
l
a
s
w
e
l
l
a
s
m
u
l
t
i
l
a
t
e
r
a
l
d
e
v
e
l
o
p
m
e
n
t
a
i
d
i
n
U
S
d
o
l
l
a
r
s
p
e
r
c
a
p
i
t
a
O
E
C
D
(
2
0
0
8
)
T
r
a
d
e
O
p
e
n
n
e
s
s
l
o
g
o
f
(
I
m
p
o
r
t
+
E
x
p
o
r
t
s
/
G
D
P
)
G
l
e
d
i
t
s
c
h
(
2
0
0
2
)
F
o
r
e
i
g
n
D
i
r
e
c
t
I
n
v
e
s
t
m
e
n
t
F
D
I
n
e
t
i
n

o
w
s
i
n
%
o
f
G
D
P
W
o
r
l
d
B
a
n
k
(
2
0
0
6
)
P
o
l
i
t
i
c
a
l
S
y
s
t
e
m
P
o
l
i
t
y
I
V
:
C
o
m
b
i
n
e
d
I
n
d
e
x
J
a
g
g
e
r
s
&
G
u
r
r
(
1
9
9
3
)
P
o
l
i
t
i
c
a
l
S
y
s
t
e
m
A
C
L
P
d
e
m
o
c
r
a
c
y
d
u
m
m
y
A
l
v
a
r
e
z
e
t
a
l
.
(
1
9
9
6
)
N
a
t
i
o
n
a
l
I
n
c
o
m
e
l
o
g
o
f
G
D
P
p
e
r
c
a
p
i
t
a
G
l
e
d
i
t
s
c
h
(
2
0
0
2
)
P
o
p
u
l
a
t
i
o
n
D
e
n
s
i
t
y
N
u
m
b
e
r
o
f
p
e
o
p
l
e
p
e
r
s
q
u
a
r
e
k
i
l
o
-
m
e
t
e
r
W
o
r
l
d
B
a
n
k
(
2
0
0
6
)
E
c
o
n
o
m
i
c
G
r
o
w
t
h
Y
e
a
r
l
y
g
r
o
w
t
h
o
f
G
D
P
P
e
n
n
W
o
r
l
d
T
a
b
l
e
s
82
CHAPTER 4. RESEARCH DESIGN
4.4.3 Control Variables
In addition to the six independent variables, a number of control variables
are included in the model. Only those control variables that are associated
with both environmental performance and any of the independent variables
are included. Hence, I strive to control for those factors that may have the
potential to bias the interaction eect between international integration and
the political system of a country.
I include economic growth to control for the fact that a growing economy
is often associated with environmental degradation. Furthermore, economic
growth plays an important role in the literature on democratization (Barro
1996) and therefore has the potential to bias the mediating eect of the
political system. Since a larger population usually consumes more natural
resources and thus produces greater environmental degradation (Gassebner
et al. 2006), population density (Correlates of War 2008; Singer et al. 1972)
is also included in the statistical model
20
. Finally, as mentioned above, a
time trend is added to the model in order to control for time eects.
4.5 Conclusion
The aim of this chapter was, rst, to present an overview of the research
strategy employed to test the theoretical hypotheses derived in chapter 3.
In particular, the empirical analysis of this thesis consists of a time-series
cross-country analysis that relates the four facets of international integra-
tion IGO membership, international aid, trade openness and foreign direct
investment and a countrys political system to three dierent forms of en-
vironmental quality SO
2
, CO
2
and BOD while controlling for important
country characteristics such as income or population density. This aggregate
statistical analysis is further complemented by four case studies aimed at
illustrating the theoretical mechanisms proposed in chapter 3.
Second, this chapter has provided a detailed discussion and justication
of the precise estimation method used in the statistical analysis, which in
particular is a xed-eects estimation procedure with panel correct stan-
dard errors and a Prais-Winsten specication to control for autocorrelation.
Finally, the sample selection has been discussed and the data used to test
the various hypotheses empirically have been introduced, an overview being
provided in table 4.2.
The next chapter tests the seven hypotheses specied in the theoretical
section, relying on various time-series cross-country analyses. This is then
followed in chapter 6 by further time-series cross-country analyses using a
more rened coding of IGO membership and by the case studies to assess
20
I do not include population per se in the model since all dependent variables are
already scaled by population.
83
CHAPTER 4. RESEARCH DESIGN
which of the mechanisms are indeed behind the impact of IGO membership
on environmental quality in developing countries.
84
Chapter 5
General Results
Does international integration in the form of membership in IGOs, interna-
tional aid, trade and investment ows lead to a better provision of environ-
mental quality in developing countries? This chapter presents the results of
various time-serious cross-section models set out to answer this question. In
section 5.1 the regression model testing hypotheses 1 to 6 is discussed. This
model constitutes the touchstone whether the four facets of international
integration are indeed associated with an increase in the three measures of
environmental quality SO
2
, CO
2
and BOD. Section 5.2 then evaluates how
the political system mediates the eect of the four aspects of international
integration on the environmental performance of developing countries. This
section begins with a discussion of the indirect eect of the political sys-
tem through a countrys international integration (section 5.2.1). Sections
5.2.2 to 5.2.5 then focus on the interplay between the four aspects of in-
ternational integration and a countrys political system. Considering each
of the four aspects consecutively, I examine the interaction eect between
the aspects of international integration and the political system respectively
to evaluate whether democracies do indeed react dierently to the incen-
tives emanating from the international system. Finally, various robustness
checks are presented to evaluate the sensitivity of the results to estimation
and measurement specications.
To get an overview of the data before delving into the details of the sta-
tistical analysis, in table 5.1 we see the descriptive statistics of all variables
employed in the standard models. An interesting observation that comes
out of this descriptive overview of the data is that average CO
2
per capita
emissions are much higher compared to SO
2
per capita emissions (this is
also true for the minimum and maximum values). Another striking obser-
vation that can be made from table 5.1 is that the average Polity score for
all country years is 9.8, which refers to an autocratic form of government.
If we consider the median Polity score instead of the mean this score is even
further decreased as 50% of all our observations have a Polity score of 5 or
85
CHAPTER 5. GENERAL RESULTS
below. Hence the democracies in this sample of developing countries are
well outnumbered.
Concerning the four aspects of international integration we see that there
is quite some variation in the data with regard to IGO membership and aid
ows but less so with regard to FDI and trade ows
1
. With 50 memberships,
average IGO membership is rather high compared to the maximum of 96
IGO memberships (out of 415 possible IGO memberships). In contrast,
average development aid is only USD 57 compared to the maximum of USD
1,446 per capita. Similarly, average FDI net ows in percentage of GDP
are only 1.38% compared to the maximum of 72.32%. For all countries in
the sample, trade ows (exports and imports) are very small compared to
their level of GDP, which leads to very small values of trade openness as
displayed in table 5.1.
Table 5.1: Descriptive statistics sample: all developing countries
Mean Std. Deviation Minimum Maximum
SO
2
per capita
1
0.01 0.02 0 0.19
CO
2
per capita
1
1847.11 2771.01 0 24188.81
BOD per capita 0.005 0.004 0 0.02
IGO membership 50 16 1 96
Aid per capita 57.21 84.81 -17.88 1446.21
ln trade openness -16.00 0.84 -19.03 -12.56
FDI % GDP 1.38 3.89 -82.81 72.32
Polity 9.81 6.93 1 21
GDP pc 3454.28 2747.21 281.26 16667.88
Population density 57.12 93.66 0 .017 844.18
Growth 0.01 0.10 -0.62 2.89
1
measured here in metric tons
1
Negative values for FDI net inows per capita imply that more capital is going out of
the country than coming into the country. Negative values for aid per capita imply that
the amount received by donors in the form of repayments of principal, osetting entries
for debt relief, repatriation of capital etc. exceed new aid ows.
86
CHAPTER 5. GENERAL RESULTS
Before we nally turn to the results testing whether these four aspects
of international integration are indeed associated with an increase in the
environmental performance of developing countries, some details of the sta-
tistical analysis need to be recapitulated. First, since international integra-
tion might not aect environmental quality contemporaneously (the eects
may need some time to trickle through the political system of a country),
all of the independent and control variables are incorporated into the model
with a one-year time lag
2
. Second, as mentioned in section 4.2, the sample
always includes all high-middle, low-middle and low-income countries. In
those cases in which the results are sensitive to the sample size they are
provided and discussed below.
5.1 Results Standard Model: Independent Eect
of International Integration
Table 5.2 shows the results testing hypothesis 1 to 6
3
. Therefore, table
5.2 displays all results testing the independent eect of a countrys income,
its political system and the four aspects of international integration on its
environmental performance while leaving aside for the moment the condi-
tional eect of international integration and the political system postulated
in hypothesis 7. This conditional eect between a countrys political system
and the four aspects of international integration is tested in the subsequent
sections.
Starting with hypothesis 1, we see that in all models, GDP per capita is
associated with an increase in pollution as predicted by the EKC argument
4
.
Hence an increase in national income is associated with a decrease in envi-
ronmental quality, conrming the theoretical assertion that the countries in
this sample have not yet reached the turning point on the EKC after which
an increase in GDP would lead to an increase in a countrys environmental
performance.
Turning to the results testing whether the postulated positive relation-
ship between the four dierent characteristics of international integration
and environmental quality in developing countries indeed holds, we see that
only one of the four manifestations of international integration is indeed
associated with better environmental quality, namely membership in inter-
national organizations. More precisely, the results show that membership
2
However, results do not change if contemporaneous variables are used. In contrast,
some results are even more signicant. Consequently, using a one-year time lag of all of
the independent and control variables seems to be a more conservative approach.
3
Since I use a Prais-Winsten specication to control for autocorrelation, the autore-
gressive parameter rho is reported for all the models. As we can see, in all of the models,
rho is smaller than 1, which indicates that the error process does not have a unit root.
4
Since the dependent variable is pollution emissions, a negative coecient sign implies
a reduction in emissions and therefore an increase in environmental quality.
87
CHAPTER 5. GENERAL RESULTS
in international organizations has a positive inuence on the environmental
performance of developing countries as measured by any of the three indi-
cators. Since the dependent variable is the log of SO
2
, CO
2
or BOD, the
coecients have a percentage interpretation (Wooldridge 2003). For exam-
ple, if a countrys IGO membership increases by one unit (i.e. one further
membership), the model predicts a reduction in this countrys SO
2
emissions
by 2% and a reduction in its CO
2
emissions and BOD by 1%.
Table 5.2: Baseline model sample: all developing countries
(1) (2) (3)
SO
2
CO
2
BOD
IGO membership -0.01*** -0.01*** -0.01***
(0.00) (0.00) (0.00)
Aid -0.00 0.00 -0.00
(0.00) (0.00) (0.00)
ln trade openness 0.04 0.10*** -0.02
(0.03) (0.02) (0.03)
FDI % GDP 0.00 -0.00 -0.00
(0.00) (0.00) (0.00)
Polity -0.00 -0.00 -0.00
(0.00) (0.00) (0.00)
ln GDP pc 0.42*** 0.54*** 0.55***
(0.05) (0.05) (0.08)
Population density -0.00 0.00*** 0.00***
(0.00) (0.00) (0.00)
Growth -0.01 -0.01 -0.10
(0.04) (0.04) (0.09)
Time trend 0.00 0.01*** 0.00
(0.01) (0.00) (0.00)
Constant 0.00 -6.16*** 0.00
(0.00) (0.55) (0.00)
Country xed eects yes yes yes
rho 0.77 0.72 0.66
Observations 2572 2663 1183
R
2
0.96 0.87 0.96
Number of countries 106 110 94
Standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1
In contrast, international integration as measured by aid or foreign di-
rect investment is not signicantly associated with any form of environmen-
tal degradation. Trade openness is even associated with an increase in CO
2
emissions while having no eect on SO
2
emissions or BOD. This implies
88
CHAPTER 5. GENERAL RESULTS
Table 5.3: Baseline model: BOD only sample: low- and low-middle income
countries
(1) (2)
BOD BOD
low + low-middle low income
IGO membership -0.01*** -0.01***
(0.00) (0.00)
Aid 0.00 -0.00
(0.00) (0.00)
ln trade openness 0.04 0.01
(0.03) (0.03)
FDI % GDP -0.01** -0.02***
(0.00) (0.01)
Polity -0.00 0.00
(0.00) (0.00)
ln GDP pc 0.53*** 0.61***
(0.08) (0.12)
Population density 0.00*** 0.01***
(0.00) (0.00)
Growth -0.19** -0.16
(0.08) (0.15)
Time trend -0.00 -0.02***
(0.00) (0.01)
Constant -12.60*** -11.38***
(0.97) (0.93)
Country xed eects yes yes
rho 0.01 0.70
Observations 817 370
R
2
1.00 1.00
Number of countries 72 40
Standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1
that countries which are more open to trade are characterized by higher
greenhouse gas emissions compared to those developing countries that are
less open to trade. Relating this nding to the theoretical arguments pro-
posed in section 3.3.3, we can deduce that the scale eect of trade seems
to dominate the technology and composition eect leading to an increase in
CO
2
emissions.
However, the results in table 5.2 need to be qualied to some extent
since the coecient of foreign direct investment on water quality is sensitive
to the sample size. If we only consider low and low-middle income coun-
tries see table 5.3 FDI seems to be positively related to water quality,
89
CHAPTER 5. GENERAL RESULTS
indicating that those countries that receive more foreign investment are char-
acterized by lower water pollution than those countries with less FDI. These
results would speak against the pollution haven argument, which postulates
that foreign investment would result in more environmental degradation in
developing countries as dirty industries would move from highly regulated
industrialized countries to low regulated developing countries.
To summarize the results with regard to international integration, only
hypothesis 3 receives full support since IGO membership seems to be the
only facet of international integration that is associated with an increase in
the environmental performance of developing countries. Hypothesis 4 and 5,
which postulate a positive relationship between aid and trade on the one side
and environmental quality on the other, are not conrmed by the statistical
analysis. Concerning FDI inows, hypothesis 6 does not seem to hold with
regard to air pollution. However, when considering water quality we see a
positive relationship between foreign investment and better water quality
for the sample of low and low-middle income countries.
Before discussing the results with regard to the political system of a
country, the results of the control variables warrant some discussion. Over-
all, none of the control variables exerts a robust inuence. Population den-
sity is associated with an increase in CO
2
emissions and BOD but is not
statistically signicant for SO
2
emissions. Interestingly, the time trend is
signicant only for CO
2
emissions, implying that CO
2
emissions are signif-
icantly increasing over time. Finally, a surprising nding is that economic
growth is not signicantly associated with SO
2
and CO
2
emissions and is
even connected to a decrease in water pollution.
Finally, we focus our discussion on the eect of the political system on the
environmental performance of developing countries. As argued in the the-
oretical part, democratic developing countries do not perform better with
regard to environmental quality relative to their autocratic counterparts.
The results in table 5.2 show no statistically signicant eect of the political
regime type on any of the indicators of environmental quality. This implies
that the coecient of the political system cannot be dierentiated from
zero, which would be the equivalent of no eect. Although it is of course
dicult to conrm that there is indeed no eect of political regime type
on environmental quality in developing countries, the rather large number
of observations, which renders the estimation more ecient, and the coe-
cient size, which is indeed zero, lend some support to the conclusion of no
independent eect of the political system.
Moreover, the results in table 5.2 are further supported when the Polity
IV measure of democracy is replaced with the dichotomous ACLP democ-
racy variable. As shown in table 5.4 the dichotomous democracy measure
is also not signicantly associated with the environmental performance of
developing countries as measured by both air and water quality.
These results are therefore compatible with the assertion that at the
90
CHAPTER 5. GENERAL RESULTS
Table 5.4: Baseline model: ACLP democracy variable sample: all devel-
oping countries
(1) (2) (3)
SO
2
CO
2
BOD
IGO membership -0.01*** -0.01*** -0.01***
(0.00) (0.00) (0.00)
Aid -0.00 0.00 0.00
(0.00) (0.00) (0.00)
ln trade openness 0.04 0.09*** 0.01
(0.03) (0.02) (0.02)
FDI % GDP 0.00 -0.00 -0.01**
(0.00) (0.00) (0.00)
ACLP democracy -0.03 -0.01 0.01
(0.02) (0.02) (0.02)
ln GDP pc 0.41*** 0.53*** 0.61***
(0.05) (0.05) (0.07)
Population density -0.00 0.00*** 0.00**
(0.00) (0.00) (0.00)
Growth -0.01 -0.00 -0.14**
(0.04) (0.04) (0.07)
Time trend 0.00 0.01*** 0.00
(0.00) (0.00) (0.00)
Constant -15.86*** -6.14*** -10.85***
(0.60) (0.55) (0.93)
Country xed eects yes yes yes
rho 0.78 0.72 0.54
Observations 2606 2735 1190
R
2
0.96 0.87 0.99
Number of countries 108 114 96
Standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1
stage of early economic development, environmental quality is perceived as
a luxury good, implying that constituents do not demand that their gov-
ernments spend scarce resources on enhancing environmental quality. Only
in mediating the eect of a countrys international integration should the
political system play a decisive role concerning environmental performance.
91
CHAPTER 5. GENERAL RESULTS
5.2 Analyzing the Mediating Eect of the Political
System
The reasoning behind the assertion that the political system mediates the in-
uence of the international system, as put forward in the theoretical section,
is twofold: First, democracies should be more responsive to the incentives
emanating from the international system in the sense that they should use
the resources available to them through their international ties for a better
provision of environmental quality. Second, democratic countries should in
general be more inclined to join IGOs, they should receive more development
aid and foreign investment and they are supposed to trade more. Overall,
then, democracies should be better integrated into the international sys-
tem. The rst reason suggests an interaction term between the four forms
of international integration and a countrys political system. The second
reason, however, refers to a more indirect eect that the political system
exerts, through the intensity of a countrys international integration, on its
environmental performance.
5.2.1 Indirect Eect of the Political System through Inter-
national Integration
In order to quantitatively assess the size of the indirect eect of the political
system through the intensity of a countrys international integration one can
rely on path analysis (Cohen et al. 2003; Garson 2008). The logic behind
this path analysis is that because the political system is supposed to aect
a countrys level of international integration, each of the four coecients of
international integration consists partly of this indirect eect of the political
system.
Using path analysis means, rstly, estimating the eect of the political
system on the four dierent forms of international integration (IGO mem-
bership, aid, trade openness and FDI). In a second step, it is necessary to
multiply the obtained coecient of the political system on the respective
form of international integration with the coecient of this particular form
of international integration (e.g. IGO membership) on environmental qual-
ity. This would then give the overall indirect eect of the political system
on environmental quality, which works through a countrys international
integration
5
.
5
To illustrate, the overall indirect eect of the political system on SO
2
emissions would
be:

DEMonIGO

IGOonSO2
+
DEMonAID

AIDonSO2
+
DEMonFDI

FDIonSO2
+
DEMonTrade

TradeonSO2
92
CHAPTER 5. GENERAL RESULTS
Table 5.5 shows the regression coecients for the political system on
each of the four aspects of international integration
6
, which we require to
calculate the overall indirect eect of the political system on the three forms
of environmental quality. The remaining coecients that are necessary to
determine the overall eect of democracy are displayed in table 5.2, which
shows the results testing the independent eects of international integration.
Table 5.5: Eect of political system on international integration sample:
all developing countries
(1)

DEMonIGO
-0.017
(0.020)

DEMonAID
0.369
(0.278)

DEMonTrade
-0.004
(0.002)

DEMonFDI
0.052**
(0.024)
Interestingly, according to table 5.5 only the coecient of the political
system on aid is statistically signicant and positive. In contrast to our
theoretical considerations, a countrys political system is not signicantly
associated to IGO membership, trade or FDI.
Using the coecients displayed in table 5.5 together with the coecients
of the four aspects of international integration as displayed in table 5.2 nally
allows us to calculate the overall indirect eect of the political system on
the three indicators of environmental quality. Surprisingly, doing so results
in an indirect eect of democracy on all three indicators of environmental
quality that cannot be dierentiated from zero. This is the case because
the dierent paths through IGO, aid, trade and FDI through which the po-
litical system could inuence environmental quality cancel each other out,
implying that the overall indirect eect of the political system through a
countrys international integration is zero. Hence in contrast to the theoret-
ical considerations, the political system does not positively aect a countrys
environmental performance through the indirect link of its international in-
tegration.
However, this indirect link through international integration is only one
of the two hypothesized channels through which the political system might
inuence a countrys environmental performance. The second channel through
6
These are obtained by using a xed eects regression while controlling for the time
context of the panel data only.
93
CHAPTER 5. GENERAL RESULTS
which the political system might become important for environmental qual-
ity is that it amplies the eect of international integration. In particular,
democracies should be more responsive to the incentives emanating from
the international system in the sense that they should use the resources
available to them through their international ties for a better provision of
environmental quality. Therefore, the next four sections analyze the inter-
action between each of the four aspects of international integration, IGO
membership, international aid, trade and foreign direct investment, and the
political system to fully test hypothesis 7, which posits that the eect of
any of the four aspects of international integration is more conducive to
environmental quality in democracies compared to autocracies.
5.2.2 Interaction between IGO Membership and Political
System
The results showing the interaction eect of IGO membership and a coun-
trys political system are presented in table 5.6. In order to facilitate inter-
pretation of the interaction terms, I calculated marginal eects and plotted
them in various graphs
7
. Figure 5.1 shows the marginal eect of IGO mem-
bership, i.e. the eect of one further membership, on SO
2
emissions over
the range of the democracy variable. As predicted by theory, IGO member-
ship reduces SO
2
emissions signicantly, no matter whether the country is a
democracy or an autocracy. However, the eect is much more pronounced in
democracies than in autocracies, which is supportive to hypothesis 7. Hence
a one-unit increase in IGO membership implies a reduction of approximately
1% in SO
2
emissions in an autocratic country whereas it implies a decrease
of around 1.5% in a full democracy.
These results provide initial conrmation of the theoretical arguments set
out above. A countrys number of IGO memberships has a positive inuence
on its SO
2
emissions, regardless of its political system. Furthermore, a
democratic political system increases this positive eect of integration into
the international system, suggesting that democracies do indeed appear to
be more responsive to IGOs than their autocratic counterparts.
Conversely, if we look at the eect of IGO membership on CO
2
emissions
and BOD, the picture is reversed (gures 5.2 and 5.3). Although IGO mem-
bership is again associated with a statistically signicant reduction in CO
2
emissions and BOD for all dierent regime types, the eect is now stronger
in autocracies and less pronounced in democracies. These results, although
in line with hypothesis 3, contrast with hypothesis 7. However, the dierence
in eect size between full autocracies and full democracies is not as distinct
as in the case of SO
2
. A one-unit increase in IGO membership implies an
approximately 1% reduction in CO
2
emissions in an autocratic country and
7
All gures showing marginal eects with interaction terms were produced using the
computer code developed by Brambor et al. (2006)
94
CHAPTER 5. GENERAL RESULTS
Table 5.6: Interaction between IGO membership and political system
sample: all developing countries
(1) (2) (3)
SO
2
CO
2
BOD
IGO membership -0.01*** -0.01*** -0.01***
(0.00) (0.00) (0.00)
Polity 0.01 -0.01 -0.01
(0.01) (0.01) (0.01)
Polity*IGO -0.00 0.00 0.00
(0.00) (0.00) (0.00)
Aid -0.00 0.00 -0.00
(0.00) (0.00) (0.00)
ln trade openness 0.04 0.10*** -0.02
(0.03) (0.02) (0.03)
FDI % GDP 0.00 -0.00 -0.00
(0.00) (0.00) (0.00)
ln GDP pc 0.41*** 0.53*** 0.55***
(0.05) (0.05) (0.08)
Population density -0.00 0.00*** 0.00***
(0.00) (0.00) (0.00)
Growth -0.00 -0.01 -0.10
(0.04) (0.04) (0.09)
Time trend 0.00 0.01*** 0.00
(0.01) (0.00) (0.00)
Constant -15.81*** -6.06*** -12.25***
(0.63) (0.56) (1.02)
Country xed eects yes yes yes
rho 0.78 0.72 0.66
Observations 2572 2663 1183
R
2
0.96 0.87 0.96
Number of countries 106 110 94
Standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1
is reduced to a 0.8% decrease in CO
2
in a full democracy. Turning to water
pollution (gure 5.3) we see that one further IGO membership reduces BOD
by 0.12% in an autocracy and by 0.11% in a democracy. Hence although
the eect of IGO membership on CO
2
emissions and water quality seems to
be greater in autocracies than in democracies, the overall dierence between
the eect sizes in autocracies and democracies is almost negligible.
Looking at dierent sample sizes, the results concerning the interaction
eect between IGO membership and the political system do not change with
95
CHAPTER 5. GENERAL RESULTS
-.03
-
.
0
3
-.03 -.02
-
.
0
2
-.02 -.01
-
.
0
1
-.01 0
0
0Marginal Effect of IGO
M
a
r
g
i
n
a
l

E
f
f
e
c
t

o
f

I
G
O
Marginal Effect of IGO 1
1
16
6
611
11
11 16
16
16 21
21
21 Polity
Polity
Polity Marginal Effect of IGO
Marginal Effect of IGO
Marginal Effect of IGO 95% Confidence Interval
95% Confidence Interval
95% Confidence Interval

Dependent Variable: ln SO2 pc, full sample
Dependent Variable: ln SO2 pc, full sample
Dependent Variable: ln SO2 pc, full sample

Marginal Effect of IGO membership as Democracy Changes
Marginal Effect of IGO membership as Democracy Changes
Marginal Effect of IGO membership as Democracy Changes
Figure 5.1: Marginal eect of IGO membership SO
2
-.03
-
.
0
3
-.03 -.02
-
.
0
2
-.02 -.01
-
.
0
1
-.01 0
0
0Marginal Effect of IGO
M
a
r
g
i
n
a
l

E
f
f
e
c
t

o
f

I
G
O
Marginal Effect of IGO 1
1
16
6
611
11
11 16
16
16 21
21
21 Polity
Polity
Polity Marginal Effect of IGO
Marginal Effect of IGO
Marginal Effect of IGO 95% Confidence Interval
95% Confidence Interval
95% Confidence Interval

Dependent Variable: ln CO2 pc, full sample
Dependent Variable: ln CO2 pc, full sample
Dependent Variable: ln CO2 pc, full sample

Marginal Effect of IGO membership as Democracy Changes
Marginal Effect of IGO membership as Democracy Changes
Marginal Effect of IGO membership as Democracy Changes
Figure 5.2: Marginal eect of IGO membership CO
2
-.03
-
.
0
3
-.03 -.02
-
.
0
2
-.02 -.01
-
.
0
1
-.01 0
0
0Marginal Effect of IGO
M
a
r
g
i
n
a
l

E
f
f
e
c
t

o
f

I
G
O
Marginal Effect of IGO 1
1
16
6
611
11
11 16
16
16 21
21
21 Polity
Polity
Polity Marginal Effect of IGO
Marginal Effect of IGO
Marginal Effect of IGO 95% Confidence Interval
95% Confidence Interval
95% Confidence Interval

Dependent Variable: ln BOD pc, full sample
Dependent Variable: ln BOD pc, full sample
Dependent Variable: ln BOD pc, full sample

Marginal Effect of IGO membership as Democracy Changes
Marginal Effect of IGO membership as Democracy Changes
Marginal Effect of IGO membership as Democracy Changes
Figure 5.3: Marginal eect of IGO membership BOD
96
CHAPTER 5. GENERAL RESULTS
the exception of water quality in low-income countries
8
. As gure 5.4 shows,
if we only consider countries with a GNI per capita level below USD 756,
which is the World Bank classication threshold for low-income countries,
we see that the eect of IGO membership is now stronger in democracies
compared to its eect in autocracies. This again is in line with hypothesis
7 indicating a reduction in water pollution independent of the regime type
with the eect being more pronounced in democracies than in autocracies.
-.03
-
.
0
3
-.03 -.02
-
.
0
2
-.02 -.01
-
.
0
1
-.01 0
0
0Marginal Effect of IGO
M
a
r
g
i
n
a
l

E
f
f
e
c
t

o
f

I
G
O
Marginal Effect of IGO 1
1
16
6
611
11
11 16
16
16 21
21
21 Polity
Polity
Polity Marginal Effect of IGO
Marginal Effect of IGO
Marginal Effect of IGO 95% Confidence Interval
95% Confidence Interval
95% Confidence Interval

Dependent Variable: ln BOD pc, low income sample
Dependent Variable: ln BOD pc, low income sample
Dependent Variable: ln BOD pc, low income sample

Marginal Effect of IGO membership as Democracy Changes
Marginal Effect of IGO membership as Democracy Changes
Marginal Effect of IGO membership as Democracy Changes
Figure 5.4: Marginal eect of IGO membership BOD
Altogether, the interaction eect between IGO membership and envi-
ronmental quality provides mixed support for hypothesis 7. Democracies do
indeed appear to be more responsive to the inuence of IGO membership
when it comes to SO
2
pollution (and water pollution if we consider the low
income sample only). They are, however, not more responsive with regard
to CO
2
pollution. One possible explanation for this nding could be due to
the dierent nature of the environmental problem: SO
2
in contrast to CO
2
is a local air pollutant whose adverse eects for the environment and human
health are felt locally compared to the harms arising from CO
2
pollution.
Hence elected governments should have a greater incentive to combat SO
2
pollution since its adverse eects are more closely felt by the citizenry.
Moreover, SO
2
is mainly a byproduct of industrial activity whereas CO
2
additionally arises to a large extent from car emissions. This implies that in
the case of SO
2
industrial activity needs to be regulated whereas in the case
of CO
2
individual behavior needs to be regulated as well. Following the logic
of Bueno De Mesquita et al. (2003) democratic countries need to convince
the median voter if they want to be re-elected whereas an autocracy needs
to be supported by the elite to stay in power. Since regulating industrial
activity hurts the elite more than the median voter but regulating individual
behavior hurts the median voter more than the elite, it seems logical that a
8
The regression results are displayed in table 9.8 in the Appendix.
97
CHAPTER 5. GENERAL RESULTS
democracy nds it easier to reduce SO
2
emissions than CO
2
emissions.
5.2.3 Interaction between Aid and Political System
Turning to the interaction between aid and the political system, we see
that in all of the gures 5.5 to 5.7, the condence interval indicated by
the two dashed lines encompasses the zero line. This implies that over
the whole range of the polity variable, the eect of international aid does
not reach statistical signicance (the results of the statistical model are
displayed in table 5.7). Hence international aid per capita does not seem to
aect the environmental performance of democratic or autocratic developing
countries.
Instead of scaling development aid by the population of a country, one
could scale it by a countrys GDP as in the case of FDI and trade openness.
However, using aid divided by GDP does not change the results either
9
. In
none of the models is international aid signicantly related to the provision
of environmental quality as measured in terms of air and water quality.
The result that international aid does not aect the environmental per-
formance of developing countries is in line with some of the earlier literature
which looks at the domestic outcomes of international aid. Boone (1996), for
example, empirically shows that development aid does not further poverty
reduction but rather prots the ruling elite. Furthermore, he shows that
democratic countries do not seem to be more eective in using development
aid for poverty reduction than autocratic countries. This is very much in line
with the ndings in this section since democratic developing countries do not
seem to use the additional resources provided by international donors for a
higher provision of environmental quality compared to autocratic developing
countries.
One potential reason for this non-nding could be that in contrast to
the theoretical expectations the positive eects of aid do not outweigh the
aid induced economic growth. Hence it seems that the positive eects of aid
such as technology and knowledge transfer and the promotion of sustainable
development can at least neutralize the negative eect of international aid
arising from a boost in economic activity, leading to neither a positive nor
a negative eect on environmental quality in developing countries.
9
The results using aid divided by GDP instead of aid divided by population can be
found in table 9.9 and gures 9.1 to 9.3 in the Appendix.
98
CHAPTER 5. GENERAL RESULTS
Table 5.7: Interaction between aid and political system sample: all devel-
oping countries
(1) (2) (3)
SO
2
CO
2
BOD
Aid 0.00 0.00 -0.00
(0.00) (0.00) (0.00)
Polity -0.00 -0.00 -0.00
(0.00) (0.00) (0.00)
Polity*Aid -0.00 -0.00 -0.00
(0.00) (0.00) (0.00)
IGO membership -0.01*** -0.01*** -0.01***
(0.00) (0.00) (0.00)
ln trade openness 0.04 0.10*** -0.02
(0.03) (0.02) (0.03)
FDI % GDP 0.00 -0.00 -0.00
(0.00) (0.00) (0.00)
ln GDP pc 0.42*** 0.53*** 0.55***
(0.05) (0.05) (0.09)
Population density -0.00 0.00*** 0.00***
(0.00) (0.00) (0.00)
Growth -0.01 -0.01 -0.10
(0.04) (0.04) (0.09)
Time trend 0.00 0.01*** 0.00
(0.01) (0.00) (0.00)
Constant 0.00 0.00 0.00
(0.00) (0.00) (0.00)
Country xed eects yes yes yes
rho 0.77 0.72 0.66
Observations 2572 2663 1183
R
2
0.96 0.87 0.96
Number of countries 106 110 94
Standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1
99
CHAPTER 5. GENERAL RESULTS
-.001
-
.
0
0
1
-.001 -.0005
-
.
0
0
0
5
-.0005 0
0
0.0005
.
0
0
0
5
.0005 .001
.
0
0
1
.001 Marginal Effect of Aid
M
a
r
g
i
n
a
l

E
f
f
e
c
t

o
f

A
i
d
Marginal Effect of Aid 1
1
16
6
611
11
11 16
16
16 21
21
21 Polity
Polity
Polity Marginal Effect of Aid
Marginal Effect of Aid
Marginal Effect of Aid 95% Confidence Interval
95% Confidence Interval
95% Confidence Interval

Dependent Variable: ln SO2 pc, full sample
Dependent Variable: ln SO2 pc, full sample
Dependent Variable: ln SO2 pc, full sample

Marginal Effect of Aid as Democracy Changes
Marginal Effect of Aid as Democracy Changes
Marginal Effect of Aid as Democracy Changes
Figure 5.5: Marginal eect of aid SO
2
-.001
-
.
0
0
1
-.001 -.0005
-
.
0
0
0
5
-.0005 0
0
0.0005
.
0
0
0
5
.0005 Marginal Effect of Aid
M
a
r
g
i
n
a
l

E
f
f
e
c
t

o
f

A
i
d
Marginal Effect of Aid 1
1
16
6
611
11
11 16
16
16 21
21
21 Polity
Polity
Polity Marginal Effect of Aid
Marginal Effect of Aid
Marginal Effect of Aid 95% Confidence Interval
95% Confidence Interval
95% Confidence Interval

Dependent Variable: ln CO2 pc, full sample
Dependent Variable: ln CO2 pc, full sample
Dependent Variable: ln CO2 pc, full sample

Marginal Effect of Aid as Democracy Changes
Marginal Effect of Aid as Democracy Changes
Marginal Effect of Aid as Democracy Changes
Figure 5.6: Marginal eect of aid CO
2
-.001
-
.
0
0
1
-.001 -.0005
-
.
0
0
0
5
-.0005 0
0
0.0005
.
0
0
0
5
.0005 Marginal Effect of Aid
M
a
r
g
i
n
a
l

E
f
f
e
c
t

o
f

A
i
d
Marginal Effect of Aid 1
1
16
6
611
11
11 16
16
16 21
21
21 Polity
Polity
Polity Marginal Effect of Aid
Marginal Effect of Aid
Marginal Effect of Aid 95% Confidence Interval
95% Confidence Interval
95% Confidence Interval

Dependent Variable: ln BOD pc, full sample
Dependent Variable: ln BOD pc, full sample
Dependent Variable: ln BOD pc, full sample

Marginal Effect of Aid as Democracy Changes
Marginal Effect of Aid as Democracy Changes
Marginal Effect of Aid as Democracy Changes
Figure 5.7: Marginal eect of aid BOD
100
CHAPTER 5. GENERAL RESULTS
5.2.4 Interaction between Trade and Political System
In contrast to our theoretical expectations, the eect of trade on environ-
mental quality is never positive whether we consider a democracy or an
autocracy (the regression results are displayed in table 5.8). More precisely,
we see in gures 5.8 to 5.12 that trade is either associated with an increase in
pollution levels or that the coecient does not reach statistical signicance.
However, all of the gures 5.8 to 5.12 indicate that the positive eect on
pollution levels is larger in autocracies and less pronounced in democracies.
Looking at gure 5.8, we see that trade is only associated with an increase
in SO
2
emissions if we consider an autocracy, whereas for the entire spectrum
of the polity variable that stands for democratic political systems (a value of
10 and above) the eect of trade is no longer signicant. However, when we
restrict the analysis to the sample that includes low-middle and low-income
countries only, the eect of trade is insignicant throughout the entire range
of the polity variable (see gure 5.11).
Considering CO
2
, the eect of trade is associated with an increase in
emissions throughout the whole range of the polity variable. However, as
indicated above, the eect is stronger in autocracies compared to democ-
racies. Hence a one-unit increase in the log of trade openness would imply
a 14% increase in CO
2
emissions whereas a one-unit increase in a democ-
racy would only imply an increase of 6%. These results are also sensitive
with regard to the sample size: when looking at low-middle and low-income
countries only, the eect of trade becomes insignicant for democracies (see
gure 5.12).
Finally, we see in gure 5.10 that the eect of trade on water quality is
insignicant throughout the entire range of the polity variable.
Altogether, these results show us that in contrast to the theoretical ex-
pectations the scale eect of trade clearly outweighs the composition and
the technology eect, leading regardless of the political system to either
a positive eect of trade on environmental pollution levels or to an insigni-
cant eect. However, the political system indeed mediates the eect of trade
on environmental pollution since democracies seem to dampen this pollution
increasing eect of trade. Although trade is not associated with an increase
in environmental quality in democracies, the eect of trade is at least not
as bad in democracies as it is in autocracies.
101
CHAPTER 5. GENERAL RESULTS
Table 5.8: Interaction between trade and political system sample: all
developing countries
(1) (2) (3)
SO
2
CO
2
BOD
ln trade openness 0.08*** 0.13*** 0.02
(0.03) (0.04) (0.06)
Polity -0.07* -0.06 -0.05
(0.04) (0.04) (0.04)
Polity*Openness -0.00* -0.00 -0.00
(0.00) (0.00) (0.00)
IGO membership -0.01*** -0.01*** -0.01***
(0.00) (0.00) (0.00)
Aid -0.00 0.00 -0.00
(0.00) (0.00) (0.00)
FDI % GDP 0.00 -0.00 -0.00
(0.00) (0.00) (0.00)
ln GDP pc 0.42*** 0.54*** 0.55***
(0.05) (0.05) (0.08)
Population density -0.00 0.00*** 0.00**
(0.00) (0.00) (0.00)
Growth -0.00 -0.00 -0.10
(0.04) (0.04) (0.09)
Time trend 0.00 0.01*** 0.00
(0.01) (0.00) (0.00)
Constant 0.00 -5.66*** 0.00
(0.00) (0.65) (0.00)
Country xed eects yes yes yes
rho 0.77 0.72 0.66
Observations 2572 2663 1183
R
2
0.96 0.87 0.96
Number of countries 106 110 94
Standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1
102
CHAPTER 5. GENERAL RESULTS
-.1
-
.
1
-.1 -.05
-
.
0
5
-.05 0
0
0.05
.
0
5
.05 .1
.
1
.1 .15
.
1
5
.15 Marginal Effect of Trade Openness
M
a
r
g
i
n
a
l

E
f
f
e
c
t

o
f

T
r
a
d
e

O
p
e
n
n
e
s
s
Marginal Effect of Trade Openness 1
1
16
6
611
11
11 16
16
16 21
21
21 Polity
Polity
Polity Marginal Effect of Trade openness
Marginal Effect of Trade openness
Marginal Effect of Trade openness 95% Confidence Interval
95% Confidence Interval
95% Confidence Interval

Dependent Variable: ln SO2 pc, full sample
Dependent Variable: ln SO2 pc, full sample
Dependent Variable: ln SO2 pc, full sample

Marginal Effect of Trade Openness as Democracy Changes
Marginal Effect of Trade Openness as Democracy Changes
Marginal Effect of Trade Openness as Democracy Changes
Figure 5.8: Marginal eect of trade openness SO
2
0
0
0.05
.
0
5
.05 .1
.
1
.1 .15
.
1
5
.15 .2
.
2
.2 Marginal Effect of Trade Openness
M
a
r
g
i
n
a
l

E
f
f
e
c
t

o
f

T
r
a
d
e

O
p
e
n
n
e
s
s
Marginal Effect of Trade Openness 1
1
16
6
611
11
11 16
16
16 21
21
21 Polity
Polity
Polity Marginal Effect of Trade openness
Marginal Effect of Trade openness
Marginal Effect of Trade openness 95% Confidence Interval
95% Confidence Interval
95% Confidence Interval

Dependent Variable: ln CO2 pc, full sample
Dependent Variable: ln CO2 pc, full sample
Dependent Variable: ln CO2 pc, full sample

Marginal Effect of Trade Openness as Democracy Changes
Marginal Effect of Trade Openness as Democracy Changes
Marginal Effect of Trade Openness as Democracy Changes
Figure 5.9: Marginal eect of trade openness CO
2
-.1
-
.
1
-.1 -.05
-
.
0
5
-.05 0
0
0.05
.
0
5
.05 .1
.
1
.1 .15
.
1
5
.15 Marginal Effect of Trade Openness
M
a
r
g
i
n
a
l

E
f
f
e
c
t

o
f

T
r
a
d
e

O
p
e
n
n
e
s
s
Marginal Effect of Trade Openness 1
1
16
6
611
11
11 16
16
16 21
21
21 Polity
Polity
Polity Marginal Effect of Trade openness
Marginal Effect of Trade openness
Marginal Effect of Trade openness 95% Confidence Interval
95% Confidence Interval
95% Confidence Interval

Dependent Variable: ln BOD pc, full sample
Dependent Variable: ln BOD pc, full sample
Dependent Variable: ln BOD pc, full sample

Marginal Effect of Trade Openness as Democracy Changes
Marginal Effect of Trade Openness as Democracy Changes
Marginal Effect of Trade Openness as Democracy Changes
Figure 5.10: Marginal eect of trade openness BOD
103
CHAPTER 5. GENERAL RESULTS
-.2
-
.
2
-.2 -.1
-
.
1
-.1 0
0
0.1
.
1
.1 .2
.
2
.2 Marginal Effect of Trade Openness
M
a
r
g
i
n
a
l

E
f
f
e
c
t

o
f

T
r
a
d
e

O
p
e
n
n
e
s
s
Marginal Effect of Trade Openness 1
1
16
6
611
11
11 16
16
16 21
21
21 Polity
Polity
Polity Marginal Effect of Trade openness
Marginal Effect of Trade openness
Marginal Effect of Trade openness 95% Confidence Interval
95% Confidence Interval
95% Confidence Interval

Dependent Variable: ln SO2 pc, low income sample
Dependent Variable: ln SO2 pc, low income sample
Dependent Variable: ln SO2 pc, low income sample

Marginal Effect of Trade Openness as Democracy Changes
Marginal Effect of Trade Openness as Democracy Changes
Marginal Effect of Trade Openness as Democracy Changes
Figure 5.11: Marginal eect of trade openness SO
2
-.1
-
.
1
-.1 0
0
0.1
.
1
.1 .2
.
2
.2 .3
.
3
.3 Marginal Effect of Trade Openness
M
a
r
g
i
n
a
l

E
f
f
e
c
t

o
f

T
r
a
d
e

O
p
e
n
n
e
s
s
Marginal Effect of Trade Openness 1
1
16
6
611
11
11 16
16
16 21
21
21 Polity
Polity
Polity Marginal Effect of Trade openness
Marginal Effect of Trade openness
Marginal Effect of Trade openness 95% Confidence Interval
95% Confidence Interval
95% Confidence Interval

Dependent Variable: ln CO2 pc, low income sample
Dependent Variable: ln CO2 pc, low income sample
Dependent Variable: ln CO2 pc, low income sample

Marginal Effect of Trade Openness as Democracy Changes
Marginal Effect of Trade Openness as Democracy Changes
Marginal Effect of Trade Openness as Democracy Changes
Figure 5.12: Marginal eect of trade openness CO
2
104
CHAPTER 5. GENERAL RESULTS
5.2.5 Interaction between FDI and Political System
Turning to the interaction eect between FDI and the political system, we
get some very interesting results (the regression results are displayed in table
5.9). Whereas the eect of FDI is always insignicant throughout the entire
range of the polity variable for CO
2
emissions (gure 5.14), the eect varies
strongly between democracies and autocracies in the case of SO
2
(gure
5.13). In the case of SO
2
emissions, we see that the marginal eect of
FDI is positive in autocracies but negative in democracies. Hence foreign
investment is associated with an increase in SO
2
emissions in autocracies
whereas it is associated with a decrease in emissions in democracies. This
result would imply that only autocracies seem to be considered as pollution
havens, whereas in democracies the positive eects of foreign investment
such as the transfer of more recent technology seem to prevail.
When looking at the eect of FDI on BOD the picture slightly changes.
Although the eect of FDI on BOD is always insignicant throughout the
entire polity spectrum if we consider the sample of all developing countries
(gure 5.15), the eect of FDI becomes statistically signicant negative in
autocracies if we consider the sample of low income countries only (gure
5.16). Interestingly, in both gures 5.15 and 5.16, the marginal eect line is
upward sloping, indicating that the eect of FDI on water quality is higher
in autocracies than in democracies.
-.02
-
.
0
2
-.02 -.01
-
.
0
1
-.01 0
0
0.01
.
0
1
.01 .02
.
0
2
.02 Marginal Effect of FDI
M
a
r
g
i
n
a
l

E
f
f
e
c
t

o
f

F
D
I
Marginal Effect of FDI 1
1
16
6
611
11
11 16
16
16 21
21
21 Polity
Polity
Polity Marginal Effect of FDI
Marginal Effect of FDI
Marginal Effect of FDI 95% Confidence Interval
95% Confidence Interval
95% Confidence Interval

Dependent Variable: ln SO2 pc, full sample
Dependent Variable: ln SO2 pc, full sample
Dependent Variable: ln SO2 pc, full sample

Marginal Effect of FDI net inflows as Democracy Changes
Marginal Effect of FDI net inflows as Democracy Changes
Marginal Effect of FDI net inflows as Democracy Changes
Figure 5.13: Marginal eect of FDI SO
2
105
CHAPTER 5. GENERAL RESULTS
Table 5.9: Interaction between FDI and political system sample: all de-
veloping countries
(1) (2) (3)
SO
2
CO
2
BOD
FDI % GDP 0.01*** 0.00 -0.01
(0.00) (0.00) (0.01)
Polity -0.00 -0.00 -0.00
(0.00) (0.00) (0.00)
Polity*FDI -0.00*** -0.00 0.00
(0.00) (0.00) (0.00)
IGO membership -0.01*** -0.01*** -0.01***
(0.00) (0.00) (0.00)
Aid -0.00 0.00 -0.00
(0.00) (0.00) (0.00)
ln trade openness 0.04 0.10*** -0.02
(0.03) (0.02) (0.03)
ln GDP pc 0.42*** 0.53*** 0.56***
(0.05) (0.05) (0.08)
Population density -0.00 0.00*** 0.00***
(0.00) (0.00) (0.00)
Growth -0.00 -0.00 -0.10
(0.04) (0.04) (0.09)
Time trend 0.00 0.01*** 0.00
(0.01) (0.00) (0.00)
Constant -15.61*** -6.13*** 0.00
(0.60) (0.56) (0.00)
Country xed eects yes yes yes
rho 0.77 0.72 0.65
Observations 2572 2663 1183
R
2
0.96 0.87 0.96
Number of countries 106 110 94
Standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1
106
CHAPTER 5. GENERAL RESULTS
-.01
-
.
0
1
-.01 -.005
-
.
0
0
5
-.005 0
0
0.005
.
0
0
5
.005 .01
.
0
1
.01 Marginal Effect of FDI
M
a
r
g
i
n
a
l

E
f
f
e
c
t

o
f

F
D
I
Marginal Effect of FDI 1
1
16
6
611
11
11 16
16
16 21
21
21 Polity
Polity
Polity Marginal Effect of FDI
Marginal Effect of FDI
Marginal Effect of FDI 95% Confidence Interval
95% Confidence Interval
95% Confidence Interval

Dependent Variable: ln CO2 pc, full sample
Dependent Variable: ln CO2 pc, full sample
Dependent Variable: ln CO2 pc, full sample

Marginal Effect of FDI net inflows as Democracy Changes
Marginal Effect of FDI net inflows as Democracy Changes
Marginal Effect of FDI net inflows as Democracy Changes
Figure 5.14: Marginal eect of FDI CO
2
-.02
-
.
0
2
-.02 -.01
-
.
0
1
-.01 0
0
0.01
.
0
1
.01 Marginal Effect of FDI
M
a
r
g
i
n
a
l

E
f
f
e
c
t

o
f

F
D
I
Marginal Effect of FDI 1
1
16
6
611
11
11 16
16
16 21
21
21 Polity
Polity
Polity Marginal Effect of FDI
Marginal Effect of FDI
Marginal Effect of FDI 95% Confidence Interval
95% Confidence Interval
95% Confidence Interval

Dependent Variable: ln BOD pc, full sample
Dependent Variable: ln BOD pc, full sample
Dependent Variable: ln BOD pc, full sample

Marginal Effect of FDI net inflows as Democracy Changes
Marginal Effect of FDI net inflows as Democracy Changes
Marginal Effect of FDI net inflows as Democracy Changes
Figure 5.15: Marginal eect of FDI BOD
-.04
-
.
0
4
-.04 -.02
-
.
0
2
-.02 0
0
0.02
.
0
2
.02 .04
.
0
4
.04 Marginal Effect of FDI
M
a
r
g
i
n
a
l

E
f
f
e
c
t

o
f

F
D
I
Marginal Effect of FDI 1
1
16
6
611
11
11 16
16
16 21
21
21 Polity
Polity
Polity Marginal Effect of FDI
Marginal Effect of FDI
Marginal Effect of FDI 95% Confidence Interval
95% Confidence Interval
95% Confidence Interval

Dependent Variable: ln BOD pc, low income sample
Dependent Variable: ln BOD pc, low income sample
Dependent Variable: ln BOD pc, low income sample

Marginal Effect of FDI net inflows as Democracy Changes
Marginal Effect of FDI net inflows as Democracy Changes
Marginal Effect of FDI net inflows as Democracy Changes
Figure 5.16: Marginal eect of FDI BOD
107
CHAPTER 5. GENERAL RESULTS
5.3 Robustness Checks
The last ve sections discussed the general results testing hypothesis 1 to
7. It was shown that out of the four aspects of international integration
considered in this thesis, only IGO membership hypothesis 3 is positively
associated with a higher provision of environmental quality in developing
countries. Furthermore, the analysis suggests that democratic developing
countries are in general no better caretakers of their environment than their
autocratic counterparts. However, as predicted by hypothesis 7, democracies
do indeed react dierently to the incentives emanating from the international
system. With the exception of international aid, the eect of international
integration leads either to a higher environmental quality in democracies
compared to autocracies (IGO membership and FDI) or to an increase in
pollution that is at least not as bad as in autocracies (trade).
This section reports various additional analyses undertaken to evaluate
the robustness of these general results with regard to model specication
and estimation methods. I thus aim to evaluate how condent we can be
concerning the various ndings presented earlier in this chapter.
5.3.1 Evaluating the Eect of Democracy: Regime Change
and the Eect of Democracy
In the following analysis I intend to further evaluate the eect of the polit-
ical system. A potential argument that one could put forward to challenge
the fact that we do not see a direct eect of the political system on environ-
mental quality in developing countries could be that the sample consists of
many unstable regimes. Since the provision of public goods might require
some regime stability to invest in these mostly long-term processes, political
regime change might interrupt the provision of public goods, blurring the
eect of the political system on environmental quality in developing coun-
tries. Hence I intend to show in this section that the results are not sensitive
with regard to changes in the political regime.
In the literature, dierent measures of regime change exist (Cederman
et al. forthcoming). Following Hegre et al. (2001), I rstly code regime
change with a 1 if there is at least a change of 2 units on the Polity Scale
since the previous year; otherwise, the variable remains at zero. In order
to distinguish between a regime change that results in a more democratic
system versus a change that leads to a more autocratic system, I have created
two variables: one measuring a democratization process and one measuring
an autocratization process
10
.
As can be seen in the rst three columns of table 5.10, controlling for
political regime change does not alter the results: neither the variable mea-
10
However, the results do not change if only one general measure of regime change is
used, which ignores the direction of the regime change.
108
CHAPTER 5. GENERAL RESULTS
suring the political system nor the two variables measuring democratization
and autocratization are signicantly dierent from zero. Hence even if we
take the (in)stability of the political system into account, democratic devel-
oping countries do not take better care of their environment compared to
their autocratic counterparts.
To evaluate the robustness of these results I use a measure of regime
change that relies on the dichotomous democracy measure developed by
Alvarez et al. (1996). If there was a change from democracy to autocracy
in the last year I coded the variable democratization 2 with a 1 and if
there was a change from autocracy to democracy in the last year I coded
the variable autocratization 2 with a 1; otherwise, the variables are zero.
As in the case of the regime change measure based on the polity index, we
do not see any signicant inuence on environmental quality of either the
democracy measure or the regime change measures see columns 4 to 6 in
table 5.10.
Finally, I use an even broader measure of regime change: relying on the
Polity IV index, I assign countries with a Polity score above 5 to the category
of democracies, those with a Polity score of below -5 to autocracies and all
other countries to the medium category of anocracies (Gleditsch & Ward
2000). Building on this classication of political regimes, I have constructed
a variable for regime change that is coded 1 if a country belonged to a
dierent category in the year before and zero otherwise. The last three
columns in table 5.10 show that this approach to measuring regime change
also does not alter the results. Even if we control for regime change the
political system does not have an independent eect on the provision of
public goods in developing countries
11
.
In addition to evaluating the robustness of my results, the inclusion of
regime change would have allowed some forms of non-linear relationship be-
tween democracy and the provision of environmental quality in developing
countries to be detected
12
. For example, it would have been possible that
a change in the political system does mean something dierent for an au-
tocracy than for a democracy. However, as this is not the case, we can be
fairly condent in the result that the political system has no independent
eect on the environmental performance of developing countries.
11
Including any of the measures of regime change also does not alter the results con-
cerning the interaction eects between the political system and any of the four dierent
characterizations of the international system analyzed in this thesis IGO membership,
aid, trade openness and foreign investment.
12
Including the square term of Polity IV does not change the results: democracy is not
signicantly related to any of the three indicators of environmental quality.
109
CHAPTER 5. GENERAL RESULTS
T
a
b
l
e
5
.
1
0
:
R
e
g
i
m
e
c
h
a
n
g
e
a
n
d
t
h
e
e

e
c
t
o
f
d
e
m
o
c
r
a
c
y

s
a
m
p
l
e
:
a
l
l
d
e
v
e
l
o
p
i
n
g
c
o
u
n
t
r
i
e
s
(
1
)
(
2
)
(
3
)
(
4
)
(
5
)
(
6
)
(
7
)
(
8
)
(
9
)
S
O
2
C
O
2
B
O
D
S
O
2
C
O
2
B
O
D
S
O
2
C
O
2
B
O
D
D
e
m
o
c
r
a
t
i
z
a
t
i
o
n
1
-
0
.
0
1
-
0
.
0
2
0
.
0
3
(
0
.
0
2
)
(
0
.
0
2
)
(
0
.
0
2
)
A
u
t
o
c
r
a
t
i
z
a
t
i
o
n
1
-
0
.
0
5
0
.
0
2
-
0
.
0
4
(
0
.
0
3
)
(
0
.
0
3
)
(
0
.
0
4
)
P
o
l
i
t
y
-
0
.
0
0
0
.
0
0
-
0
.
0
0
(
0
.
0
0
)
(
0
.
0
0
)
(
0
.
0
0
)
D
e
m
o
c
r
a
t
i
z
a
t
i
o
n
2
0
.
0
2
0
.
0
0
0
.
0
1
(
0
.
0
3
)
(
0
.
0
3
)
(
0
.
0
2
)
A
u
t
o
c
r
a
t
i
z
a
t
i
o
n
2
0
.
0
1
0
.
0
1
-
0
.
0
6
(
0
.
0
4
)
(
0
.
0
5
)
(
0
.
0
6
)
A
C
L
P
d
e
m
o
c
r
a
c
y
-
0
.
0
4
-
0
.
0
1
-
0
.
0
0
(
0
.
0
3
)
(
0
.
0
3
)
(
0
.
0
3
)
R
e
g
i
m
e
c
h
a
n
g
e
-
0
.
0
0
-
0
.
0
0
-
0
.
0
0
(
0
.
0
2
)
(
0
.
0
1
)
(
0
.
0
2
)
D
e
m
o
c
r
a
c
y
-
0
.
0
3
-
0
.
0
1
-
0
.
0
7
(
0
.
0
4
)
(
0
.
0
4
)
(
0
.
0
6
)
A
n
o
c
r
a
c
y
-
0
.
0
3
-
0
.
0
3
-
0
.
0
6
(
0
.
0
3
)
(
0
.
0
3
)
(
0
.
0
5
)
A
u
t
o
c
r
a
c
y
-
0
.
0
2
-
0
.
0
0
-
0
.
0
5
(
0
.
0
3
)
(
0
.
0
3
)
(
0
.
0
4
)
I
G
O
m
e
m
b
e
r
s
h
i
p
-
0
.
0
1
*
*
*
-
0
.
0
1
*
*
*
-
0
.
0
1
*
*
*
-
0
.
0
1
*
*
*
-
0
.
0
1
*
*
*
-
0
.
0
1
*
*
*
-
0
.
0
1
*
*
*
-
0
.
0
1
*
*
*
-
0
.
0
1
*
*
*
(
0
.
0
0
)
(
0
.
0
0
)
(
0
.
0
0
)
(
0
.
0
0
)
(
0
.
0
0
)
(
0
.
0
0
)
(
0
.
0
0
)
(
0
.
0
0
)
(
0
.
0
0
)
A
i
d
-
0
.
0
0
0
.
0
0
0
.
0
0
-
0
.
0
0
0
.
0
0
0
.
0
0
-
0
.
0
0
0
.
0
0
0
.
0
0
(
0
.
0
0
)
(
0
.
0
0
)
(
0
.
0
0
)
(
0
.
0
0
)
(
0
.
0
0
)
(
0
.
0
0
)
(
0
.
0
0
)
(
0
.
0
0
)
(
0
.
0
0
)
l
n
t
r
a
d
e
o
p
e
n
n
e
s
s
0
.
0
4
0
.
1
0
*
*
*
0
.
0
1
0
.
0
4
0
.
0
9
*
*
*
0
.
0
1
0
.
0
4
0
.
0
9
*
*
*
0
.
0
1
(
0
.
0
3
)
(
0
.
0
2
)
(
0
.
0
3
)
(
0
.
0
3
)
(
0
.
0
2
)
(
0
.
0
2
)
(
0
.
0
3
)
(
0
.
0
2
)
(
0
.
0
3
)
F
D
I
%
G
D
P
0
.
0
0
-
0
.
0
0
-
0
.
0
1
*
0
.
0
0
-
0
.
0
0
-
0
.
0
1
*
*
0
.
0
0
-
0
.
0
0
-
0
.
0
1
*
(
0
.
0
0
)
(
0
.
0
0
)
(
0
.
0
0
)
(
0
.
0
0
)
(
0
.
0
0
)
(
0
.
0
0
)
(
0
.
0
0
)
(
0
.
0
0
)
(
0
.
0
0
)
l
n
G
D
P
p
c
0
.
4
2
*
*
*
0
.
5
4
*
*
*
0
.
6
1
*
*
*
0
.
4
1
*
*
*
0
.
5
3
*
*
*
0
.
6
2
*
*
*
0
.
4
2
*
*
*
0
.
5
3
*
*
*
0
.
6
1
*
*
*
(
0
.
0
5
)
(
0
.
0
5
)
(
0
.
0
7
)
(
0
.
0
5
)
(
0
.
0
5
)
(
0
.
0
7
)
(
0
.
0
5
)
(
0
.
0
5
)
(
0
.
0
7
)
P
o
p
u
l
a
t
i
o
n
d
e
n
s
i
t
y
-
0
.
0
0
0
.
0
0
*
*
*
0
.
0
0
*
*
*
-
0
.
0
0
0
.
0
0
*
*
*
0
.
0
0
*
*
-
0
.
0
0
0
.
0
0
*
*
*
0
.
0
0
*
*
*
(
0
.
0
0
)
(
0
.
0
0
)
(
0
.
0
0
)
(
0
.
0
0
)
(
0
.
0
0
)
(
0
.
0
0
)
(
0
.
0
0
)
(
0
.
0
0
)
(
0
.
0
0
)
G
r
o
w
t
h
-
0
.
0
1
-
0
.
0
1
-
0
.
1
3
*
-
0
.
0
1
-
0
.
0
0
-
0
.
1
3
*
-
0
.
0
1
-
0
.
0
0
-
0
.
1
5
*
*
(
0
.
0
4
)
(
0
.
0
4
)
(
0
.
0
7
)
(
0
.
0
4
)
(
0
.
0
4
)
(
0
.
0
7
)
(
0
.
0
4
)
(
0
.
0
4
)
(
0
.
0
7
)
T
i
m
e
t
r
e
n
d
0
.
0
0
0
.
0
1
*
*
*
0
.
0
0
0
.
0
0
0
.
0
1
*
*
*
0
.
0
0
0
.
0
0
0
.
0
1
*
*
*
0
.
0
0
(
0
.
0
1
)
(
0
.
0
0
)
(
0
.
0
0
)
(
0
.
0
0
)
(
0
.
0
0
)
(
0
.
0
0
)
(
0
.
0
0
)
(
0
.
0
0
)
(
0
.
0
0
)
C
o
n
s
t
a
n
t
0
.
0
0
-
6
.
1
9
*
*
*
-
9
.
9
3
*
*
*
0
.
0
0
0
.
0
0
-
1
0
.
8
7
*
*
*
0
.
0
0
0
.
0
0
-
9
.
8
3
*
*
*
(
0
.
0
0
)
(
0
.
5
4
)
(
1
.
3
0
)
(
0
.
0
0
)
(
0
.
0
0
)
(
0
.
8
5
)
(
0
.
0
0
)
(
0
.
0
0
)
(
1
.
8
8
)
C
o
u
n
t
r
y

x
e
d
e

e
c
t
s
y
e
s
y
e
s
y
e
s
y
e
s
y
e
s
y
e
s
y
e
s
y
e
s
y
e
s
r
h
o
0
.
7
7
0
.
7
2
0
.
5
1
0
.
7
8
0
.
7
2
0
.
5
3
0
.
7
7
0
.
7
2
0
.
5
4
O
b
s
e
r
v
a
t
i
o
n
s
2
5
7
2
2
6
6
3
1
1
8
3
2
6
0
6
2
7
3
5
1
1
9
0
2
6
0
6
2
7
3
5
1
1
9
0
N
u
m
b
e
r
o
f
c
o
u
n
t
r
i
e
s
1
0
6
1
1
0
9
4
1
0
8
1
1
4
9
6
1
0
8
1
1
4
9
6
S
t
a
n
d
a
r
d
e
r
r
o
r
s
i
n
p
a
r
e
n
t
h
e
s
e
s
;
*
*
*
p
<
0
.
0
1
,
*
*
p
<
0
.
0
5
,
*
p
<
0
.
1
110
CHAPTER 5. GENERAL RESULTS
5.3.2 Robustness Checks with regard to Model Specication
This section provides various robustness checks of the results of the model
reported in table 5.2 in section 5.1. More precisely, the results are evaluated
with regard to the model specication. Here, the aim is to show that the
main results are not sensitive to the inclusion of additional control variables
that could theoretically bias the results when left out of the model.
Starting with the inclusion of regional dummies we see, as the results in
table 5.11 show, that including regional dummies in the model does not alter
the overall results. Hence even if we include a control that captures regional
dierences, IGO membership is still the only independent variable that is
associated with an increase in environmental quality. In general, it seems
that countries outside of Europe, which serves as the baseline category, are
characterized by lower pollution levels since most of the coecients of the
regional dummy variables are negative and statistically signicant.
Turning to important country characteristics that are often associated
with changes in environmental quality (Eliste & Fredriksson 2002; Gasseb-
ner et al. 2006; Bernauer & Koubi 2006), the next paragraphs discuss the
results including urban population, agricultural land area, the number of
cars, industrial and agricultural output and government spending
13
. As we
can see in tables 5.12 and 5.14, none of the additional characteristics alters
any of the main results. Hence the positive eect of IGO membership as
well as the non-eect of the political system are robust to the inclusion of
further country characteristics.
More precisely, as can be seen in the rst three columns of table 5.12, the
percentage of people living in urban areas is associated with an increase in all
three forms of environmental pollution. Thus the more people live in cities
compared to rural areas, the higher are the overall pollution levels of this
country. In contrast, the size of land used for agricultural purposes only has
an increasing eect on CO
2
emissions whereas it is signicant and negative
in the case of water pollution. This result is not very surprising since the
World Development Indicators only measure water pollution arising from
industrial sources.
In the rst two columns of table 5.13 we see that the results concern-
ing air pollution are robust to the inclusion of the number of cars in that
country measured as the number of passenger cars per 1,000 people. As
could have been expected, the more cars are in use in a particular country
the higher are the air pollution emissions. Finally, the last three columns
in table 5.13 show the results while additionally including industrial and
agricultural output. Industrial output is measured as a countrys net out-
put covering mining, manufacturing, construction, gas, electricity and water
whereas agricultural output covers the net output of the agricultural sector
13
Unless otherwise indicated, the data to measure these various control variables are
taken from the World Bank World Development Indicators (World Bank 2006).
111
CHAPTER 5. GENERAL RESULTS
Table 5.11: Regions sample: all developing countries
(1) (2) (3)
SO
2
CO
2
BOD
IGO membership -0.01*** -0.01*** -0.01***
(0.00) (0.00) (0.00)
Aid -0.00 0.00 0.00
(0.00) (0.00) (0.00)
ln trade openness 0.04 0.10*** 0.01
(0.03) (0.02) (0.03)
FDI % GDP 0.00 -0.00 -0.01*
(0.00) (0.00) (0.00)
Polity -0.00 -0.00 -0.00
(0.00) (0.00) (0.00)
ln GDP pc 0.42*** 0.54*** 0.61***
(0.05) (0.05) (0.07)
Population density -0.00 0.00*** 0.00***
(0.00) (0.00) (0.00)
Growth -0.01 -0.01 -0.14*
(0.04) (0.04) (0.07)
Time trend 0.00 0.01*** 0.00
(0.01) (0.00) (0.00)
Africa -14.49*** -3.54*** -0.62
(0.49) (0.19) (1.38)
Asia -16.66*** -4.30*** -1.30
(0.52) (0.24) (1.38)
Latin America -13.10*** -4.06*** 0.14
(0.59) (0.32) (1.40)
Middle East -13.89*** -1.66*** -0.75
(0.57) (0.10) (1.36)
Constant 0.00 -0.51 -9.92***
(0.00) (0.52) (1.52)
Country xed eects yes yes yes
rho 0.77 0.72 0.51
Observations 2572 2663 1183
R
2
0.96 0.87 0.99
Number of countries 106 110 94
Standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1
including sheries and forestry (World Bank 2006). Interestingly, when we
look at water quality neither the size of the industrial sector nor the size of
the agricultural sector is signicantly associated with pollution. In contrast,
concerning SO
2
emissions the size of the industrial sector is signicantly as-
112
CHAPTER 5. GENERAL RESULTS
Table 5.12: Urban population and agricultural land area sample: all de-
veloping countries
(1) (2) (3) (4) (5) (6)
SO
2
CO
2
BOD SO
2
CO
2
BOD
IGO membership -0.01*** -0.01*** -0.01*** -0.01** -0.01*** -0.01***
(0.00) (0.00) (0.00) (0.00) (0.00) (0.00)
Aid 0.00 0.00 0.00 -0.00 0.00 0.00
(0.00) (0.00) (0.00) (0.00) (0.00) (0.00)
ln trade openness 0.04 0.10*** 0.02 0.04 0.10*** 0.01
(0.03) (0.02) (0.02) (0.03) (0.02) (0.03)
FDI % GDP 0.00 -0.00 -0.00 0.00 -0.00 -0.00*
(0.00) (0.00) (0.00) (0.00) (0.00) (0.00)
Polity -0.00 -0.00 -0.00 -0.00 -0.00 -0.00
(0.00) (0.00) (0.00) (0.00) (0.00) (0.00)
ln GDP pc 0.40*** 0.51*** 0.59*** 0.44*** 0.55*** 0.64***
(0.06) (0.05) (0.08) (0.06) (0.05) (0.08)
Population density -0.00 0.00*** 0.00*** -0.00 0.00*** 0.00**
(0.00) (0.00) (0.00) (0.00) (0.00) (0.00)
Growth -0.00 -0.00 -0.10 -0.01 -0.01 -0.12*
(0.04) (0.04) (0.07) (0.04) (0.04) (0.07)
Time trend -0.00 0.00 -0.02*** -0.00 0.01* 0.00
(0.01) (0.00) (0.01) (0.01) (0.00) (0.00)
% urban population 0.01*** 0.01*** 0.04***
(0.00) (0.00) (0.01)
Agricultural land 0.01 0.01*** -0.01*
(0.00) (0.00) (0.00)
Constant 0.00 -6.28*** -10.69*** 0.00 0.00 -9.70***
(0.00) (0.51) (2.49) (0.00) (0.00) (1.71)
Country xed eects yes yes yes yes yes yes
rho 0.77 0.72 0.50 0.77 0.71 0.55
Observations 2572 2663 1183 2558 2650 1174
Number of countries 106 110 94 105 109 93
Standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1
sociated with an increase in pollution. Hence as could have been expected,
countries with a larger industrial sector are characterized by higher SO
2
emissions. Surprisingly, the industrial sector is not related to CO
2
emis-
sions while the size of the agricultural sector is associated with a decrease
in CO
2
emissions. So whereas a larger industrial sector does not necessarily
imply more CO
2
emissions, a larger agricultural sector implies lower CO
2
emissions.
113
CHAPTER 5. GENERAL RESULTS
Table 5.13: Number of cars and industry shares sample: all developing
countries
(1) (2) (3) (4) (5)
SO
2
CO
2
SO
2
CO
2
BOD
IGO membership -0.03*** -0.02*** -0.01*** -0.01*** -0.01***
(0.00) (0.00) (0.00) (0.00) (0.00)
Aid -0.00 -0.00 -0.00 0.00 0.00
(0.00) (0.00) (0.00) (0.00) (0.00)
ln trade openness -0.04 0.14*** 0.02 0.09*** 0.02
(0.07) (0.03) (0.03) (0.02) (0.03)
FDI % GDP 0.00 -0.00 -0.00 -0.00 -0.01*
(0.00) (0.00) (0.00) (0.00) (0.00)
Polity 0.00 0.01*** -0.00 -0.00 -0.00
(0.00) (0.00) (0.00) (0.00) (0.00)
ln GDP pc 0.33*** 0.51*** 0.41*** 0.51*** 0.62***
(0.10) (0.10) (0.06) (0.05) (0.07)
Population density 0.00** 0.00 -0.00 0.00*** 0.00**
(0.00) (0.00) (0.00) (0.00) (0.00)
Growth -0.04 0.03 -0.02 -0.02 -0.14**
(0.06) (0.09) (0.04) (0.04) (0.07)
Time trend 0.02** 0.02*** 0.01 0.01*** 0.00
(0.01) (0.00) (0.01) (0.00) (0.00)
Industry net output 0.01*** 0.00 -0.00
(0.00) (0.00) (0.00)
Agriculture net output -0.00 -0.01*** -0.00
(0.00) (0.00) (0.00)
Cars -0.00* -0.00***
(0.00) (0.00)
Constant 0.00 0.00 -16.30*** 0.00 -10.83***
(0.00) (0.00) (0.81) (0.00) (1.34)
Country xed eects yes yes yes yes yes
rho 0.50 0.38 0.76 0.71 0.87
Observations 779 800 2465 2528 1146
Number of countries 103 106 106 109 92
Standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1
Finally, table 5.14 shows the results including government expenditure
in the model to control for the fact that in order to provide higher levels of
public goods, more government resources might be necessary (Bernauer &
Koubi 2006). However, the inclusion of the government share of GDP
14
is
14
The data on government spending come from the Penn World Tables and measure
114
CHAPTER 5. GENERAL RESULTS
only signicantly related to the emission of CO
2
whereas a higher share of
government expenditure in GDP is associated with a decrease in emissions.
In contrast, government expenditure is not related to either BOD nor SO
2
emissions.
Table 5.14: Government share of GDP sample: all developing countries
(1) (2) (3)
SO
2
CO
2
BOD
IGO membership -0.01*** -0.01*** -0.01***
(0.00) (0.00) (0.00)
Aid 0.00 0.00 0.00
(0.00) (0.00) (0.00)
ln trade openness 0.05** 0.10*** 0.02
(0.02) (0.02) (0.03)
FDI % GDP 0.00 -0.00 -0.00
(0.00) (0.00) (0.00)
Polity -0.00 -0.00 -0.00
(0.00) (0.00) (0.00)
ln GDP pc 0.45*** 0.55*** 0.62***
(0.05) (0.05) (0.07)
Population density -0.00 0.00*** 0.00**
(0.00) (0.00) (0.00)
Growth -0.01 -0.00 -0.17**
(0.03) (0.04) (0.08)
Time trend -0.00 0.01*** 0.00
(0.00) (0.00) (0.00)
Government share of GDP 0.00 0.00** 0.00
(0.00) (0.00) (0.00)
Constant 0.00 0.00 -12.09***
(0.00) (0.00) (0.96)
Country xed eects yes yes yes
rho 0.76 0.72 0.61
Observations 2493 2584 1153
Number of countries 104 108 92
Standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1
the government share of real GDP (Heston et al. 2006).
115
CHAPTER 5. GENERAL RESULTS
5.3.3 Robustness Checks with regard to Estimation Method
The aim of this section is to evaluate the main results of this chapter with re-
gard to the dierent estimation procedures as discussed in chapter 4. Start-
ing with table 5.15 we see the results for the baseline model (reported in
table 5.2 in section 5.1) using a lagged dependent variable instead of an
AR(1) specication to deal with temporal dynamics. In the case of SO
2
and CO
2
we see no dierences between the model using a lagged dependent
variable and the baseline model with the AR(1) specication. Only if we
consider BOD as the dependent variable do we see slight changes. Whereas
the main nding of this analysis, namely that IGO membership reduces
emissions, is persistent throughout all model specications, the non-nding
for democracy disappears in the case of BOD. Democracy is signicantly
associated with an increase in water quality if we use a lagged dependent
variable to control for temporal dynamics.
As discussed in chapter 4, including a lagged dependent variable implies
that we can interpret the coecients of all other variables in the model only
as the short-term eect since the inclusion of a lagged dependent variable is
equivalent to including a weighted average of the coecients of all right-hand
side variables. Hence the nding in the case of BOD suggests that although
the long-term eect of the political system implies that democracies do not
provide better water quality than autocratic countries (since the coecient
of the political system variable is insignicant in the AR(1) model), the
short-term eect hints at a positive eect of democracy on water quality.
Another dicult question concerning panel data, as discussed in chapter
4, is the question of how to deal with time. In general, a linear time trend
is included in all of the models to capture time eects. Another possibility
is to include time xed eects. The downside of this approach is that these
time xed eects usually capture a great deal of the theoretically interesting
variation in the dependent variable without, however, explaining it. To show
that the results are nevertheless robust to the modeling of time, table 5.16
shows the results including time dummies, whereas table 5.17 shows the
results using decade dummies instead of a linear time trend. The 1970s are
taken as the reference category in the latter models
15
.
As both tables 5.16 and 5.17 show, the results do not change if we use
time dummies or decade dummies instead of a linear time trend. In all mod-
els, IGO membership is associated with an increase in environmental quality
whereas the political system does not signicantly aect environmental pol-
lution in developing countries.
15
For the models using BOD as the dependent variable, it was not possible to estimate
the model using time xed eects as the models did not converge. Therefore only the
results including decade dummies are included.
116
CHAPTER 5. GENERAL RESULTS
Table 5.15: Baseline results with a lagged dependent variable sample: all
developing countries
(1) (2) (3)
SO
2
CO
2
BOD
IGO membership -0.00** -0.00* -0.01***
(0.00) (0.00) (0.00)
Aid -0.00 0.00* -0.00
(0.00) (0.00) (0.00)
ln trade openness 0.01 0.03*** 0.02
(0.02) (0.01) (0.02)
FDI % GDP 0.00 -0.00 -0.00
(0.00) (0.00) (0.00)
Polity -0.00 0.00 -0.00*
(0.00) (0.00) (0.00)
ln GDP pc 0.10*** 0.10*** 0.09
(0.03) (0.03) (0.06)
Population density -0.00 0.00*** 0.00
(0.00) (0.00) (0.00)
Growth 0.05 0.05 -0.06
(0.05) (0.04) (0.10)
Time trend 0.00 0.00 0.01**
(0.00) (0.00) (0.00)
lagged SO
2
0.81***
(0.03)
lagged CO
2
0.81***
(0.02)
lagged BOD 0.76***
(0.07)
Constant -2.95*** -0.96*** -1.93**
(0.46) (0.23) (0.87)
Country xed eects yes yes yes
Observations 2572 2656 1055
R
2
0.98 0.99 0.96
Number of countries 106 110 91
Standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1
117
CHAPTER 5. GENERAL RESULTS
Table 5.16: Regression with both time and country xed eects sample:
all developing countries
(1) (2)
ln SO
2
pc ln CO
2
pc
IGO membership -0.01*** -0.01***
(0.00) (0.00)
Aid -0.00 0.00
(0.00) (0.00)
ln trade openness 0.03 0.10***
(0.03) (0.02)
FDI % GDP -0.00 -0.00
(0.00) (0.00)
Polity -0.00 -0.00
(0.00) (0.00)
ln GDP pc 0.39*** 0.54***
(0.06) (0.05)
Population density -0.00 0.00***
(0.00) (0.00)
Growth 0.01 -0.01
(0.04) (0.05)
Constant -15.72*** 0.00
(0.63) (0.00)
rho 0.77 0.71
Time dummies yes yes
Country dummies yes yes
Observations 2572 2663
R
2
0.96 0.88
Number of countries 106 110
Standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1
118
CHAPTER 5. GENERAL RESULTS
Table 5.17: Regression with decade dummies and country xed eects
sample: all developing countries
(1) (2) (3)
SO
2
CO
2
BOD
IGO membership -0.01*** -0.00* -0.01***
(0.00) (0.00) (0.00)
Aid -0.00 0.00 0.00
(0.00) (0.00) (0.00)
ln trade openness 0.05* 0.11*** 0.01
(0.03) (0.02) (0.02)
FDI % GDP -0.00 -0.00 -0.01**
(0.00) (0.00) (0.00)
Polity -0.00 0.00 -0.00
(0.00) (0.00) (0.00)
ln GDP pc 0.43*** 0.55*** 0.62***
(0.05) (0.05) (0.07)
Population density -0.00 0.00*** 0.00***
(0.00) (0.00) (0.00)
Growth -0.01 -0.01 -0.15**
(0.04) (0.04) (0.07)
Eighties -0.03 -0.02 -0.01
(0.03) (0.03) (0.02)
Nineties -0.02 -0.03 -0.02
(0.04) (0.04) (0.03)
Constant 0.00 0.00 -11.26***
(0.00) (0.00) (0.93)
rho 0.77 0.71 0.74
Time dummies no no no
Country dummies yes yes yes
Observations 2572 2663 1183
R
2
0.96 0.87 0.99
Number of countries 106 110 94
Standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1
119
CHAPTER 5. GENERAL RESULTS
5.4 Conclusion
The aim of this chapter was to test the dierent hypotheses set out in chapter
3 with the methods and data discussed in chapter 4. The results presented
in this chapter lend strong support to the assertion of hypothesis 1 that for
developing countries, an increase in national income is associated with a rise
in environmental pollution. The logic behind this nding is that developing
countries are still situated on the upward sloping part of the Environmental
Kuznets Curve and have not yet reached the turning point of the EKC
after which an increase in GDP would lead to an increase in a countrys
environmental performance.
The analysis of how international integration in the form of membership
in IGOs, international aid, trade openness and foreign direct investment
ows changes the capacity of democratic and autocratic countries to pro-
vide environmental quality has provided mixed results. Interestingly, only
membership in international organizations is associated to a decrease in both
air and water pollution. Trade openness does not aect SO
2
emissions and
water quality but signicantly increases greenhouse gas emissions. These
ndings suggests that more economic activity induced by higher trade lev-
els outweigh the technological transfer and the composition eect of trade,
resulting in a pollution increasing eect. International aid and foreign invest-
ment are not signicantly related to the environmental performance of de-
veloping countries with the exception of FDI in the case of water quality and
when excluding all high-middle income countries. This last nding speaks
to some extent against the pollution haven argument since low-income coun-
tries that receive more foreign investment seem to have better, not worse,
water quality.
As theoretically expected, the political system does not have an inde-
pendent eect on any of the three indicators of environmental quality. This
result underlines the hypothesis that at the stage of early economic develop-
ment, citizens do not demand that their governments spend scarce resources
on enhancing environmental quality, which implies that the political system
is not decisive for the provision of environmental quality when we consider
developing countries only. However, the political system plays an important
role as a mediator of the eect of international integration on environmen-
tal pollution. Although the indirect eect of the political system through
a countrys international integration is negligible, the various interaction
eects give a dierent picture: With the exception of international aid,
the eect of international integration either leads to a higher environmental
quality in democracies compared to autocracies (IGO membership and FDI)
or to an increase in pollution that is at least not as bad in democracies as it
is in autocracies (trade). Hence as expected from the theoretical arguments,
democracies do indeed react dierently to the incentives emanating from the
international system than their autocratic counterparts.
120
CHAPTER 5. GENERAL RESULTS
As these results have shown, the only factor that is associated with an
increase in the environmental performance of developing countries under all
settings and robustness checks is membership in IGOs. The theoretical ar-
guments relating IGO membership to environmental quality, as described in
section 3.3.1, suggest various mechanisms from issue linkage to the transfer
of technology and know-how that could enhance both the capability and
willingness of developing countries to provide better environmental quality.
However, the analyses presented in this chapter remain at the aggregate
level, making it dicult to assess which of the proposed mechanisms are at
play empirically. Consequently, the next chapter evaluates in more detail,
using a more rened coding of IGO membership, which of the mechanisms
proposed are behind this positive impact of IGO membership on environ-
mental quality in developing countries. Furthermore, this detailed quan-
titative analysis of IGO membership on environmental quality is further
complemented by 4 illustrative case studies. The main goal of these case
studies is to evaluate and illustrate which of the theoretically proposed links
is empirically at play in the everyday business of IGOs in developing coun-
tries. With this approach I expect to evaluate the aggregate quantitative
ndings more precisely and provide empirical illustrations concerning the
various mechanisms proposed to link IGO membership to a better provision
of environmental quality in developing countries.
121
Chapter 6
IGO Membership and
Environmental Quality:
Understanding the
Mechanisms
As the last chapter showed, IGO membership is the only factor that seems
to be robustly associated to environmental quality in developing countries.
In order to better understand the mechanisms underlying this eect, this
chapter evaluates the relationship between IGO membership and environ-
mental quality in greater detail. To this end, various dierent strategies are
pursued. First, the robustness of the results linking membership in IGOs
to an increase in environmental quality in developing countries is evaluated
using dierent measurement specications.
In a second step, IGO membership is disaggregated according to the
function and the degree of institutionalization of the respective IGO. Us-
ing this rened coding of IGO membership allows us to test which specic
types of IGOs are indeed associated with an increase in environmental qual-
ity. These results then permit us to better assess which of the mechanisms
proposed in the theoretical section are empirically driving the relationship
between IGO membership and environmental quality.
Third, using an instrumental variables approach, the potential endo-
geneity between IGO membership and environmental quality is evaluated.
In particular, two instruments whether a country votes in accordance with
the US in the UN General Assembly and the dierence between a countrys
per capita income measured at purchasing power parity and per capita in-
come measured at ocial exchange rate prices are employed to show that
IGO membership is leading to an increase in the environmental performance
of developing countries and that a countrys environmental record is not
the driving force behind its membership in IGOs (especially environmental
122
CHAPTER 6. IGO MEMBERSHIP AND ENVIRONMENTAL
QUALITY: UNDERSTANDING THE MECHANISMS
ones).
Finally, the analysis ends with four illustrative case studies, which are de-
signed to show which of the mechanisms suggested by the aggregate analyses
of IGO membership and environmental quality are at play in the every-day
business of IGOs in developing countries. I therefore evaluate the scope and
mandate of the particular IGOs of which the four countries chosen for this
study Cameroon, India, Laos and Lithuania are members. These cases
illustrate in some detail the various strategies by which IGOs help to make
environmental protection an important issue in developing countries, thus
underlining the plausibility of the aggregate statistical results.
6.1 Evaluating the Robustness of IGO Member-
ship
6.1.1 Accounting for the Stringency of IGO Membership
In a rst step, the robustness of the results is evaluated using the dierent
codings oered by the Correlates of War project (Pevehouse et al. 2004). As
described in section 4.4.2.1, the Correlates of War IGO dataset oers three
slightly dierent variables measuring IGO membership. These measures dif-
fer according to whether a state has full, associated or observer membership.
Up to now I have used the most inclusive (or least stringent) variable, the
one that includes full, associated and observer membership.
However, as table 6.1 shows, using either the most narrowly dened vari-
able, which includes full membership only, or the IGO variable that includes
both full and associated membership while excluding observer status instead
of the most inclusive IGO variable does not alter any of the earlier results.
IGO membership, independent of its measurement, is always associated with
a reduction in pollution emissions.
123
CHAPTER 6. IGO MEMBERSHIP AND ENVIRONMENTAL
QUALITY: UNDERSTANDING THE MECHANISMS
T
a
b
l
e
6
.
1
:
V
a
r
y
i
n
g
t
h
e
s
t
r
i
n
g
e
n
c
y
o
f
t
h
e
I
G
O
v
a
r
i
a
b
l
e

s
a
m
p
l
e
:
a
l
l
d
e
v
e
l
o
p
i
n
g
c
o
u
n
t
r
i
e
s
(
1
)
(
2
)
(
3
)
(
4
)
(
5
)
(
6
)
S
O
2
C
O
2
B
O
D
S
O
2
C
O
2
B
O
D
I
G
O
m
e
m
b
e
r
s
h
i
p
-
0
.
0
1
*
*
*
-
0
.
0
1
*
*
*
-
0
.
0
1
*
*
*
e
x
c
l
u
d
i
n
g
o
b
s
e
r
v
e
r
(
0
.
0
0
)
(
0
.
0
0
)
(
0
.
0
0
)
I
G
O
m
e
m
b
e
r
s
h
i
p
-
0
.
0
1
*
*
*
-
0
.
0
1
*
*
*
-
0
.
0
1
*
*
*
e
x
c
l
u
d
i
n
g
o
b
s
e
r
v
e
r
+
a
s
s
o
c
i
a
t
e
(
0
.
0
0
)
(
0
.
0
0
)
(
0
.
0
0
)
A
i
d
-
0
.
0
0
0
.
0
0
0
.
0
0
-
0
.
0
0
0
.
0
0
0
.
0
0
(
0
.
0
0
)
(
0
.
0
0
)
(
0
.
0
0
)
(
0
.
0
0
)
(
0
.
0
0
)
(
0
.
0
0
)
l
n
t
r
a
d
e
o
p
e
n
n
e
s
s
0
.
0
4
0
.
1
0
*
*
*
0
.
0
1
0
.
0
4
0
.
1
0
*
*
*
0
.
0
1
(
0
.
0
3
)
(
0
.
0
2
)
(
0
.
0
3
)
(
0
.
0
3
)
(
0
.
0
2
)
(
0
.
0
2
)
F
D
I
%
G
D
P
0
.
0
0
-
0
.
0
0
-
0
.
0
1
*
-
0
.
0
0
-
0
.
0
0
-
0
.
0
1
*
(
0
.
0
0
)
(
0
.
0
0
)
(
0
.
0
0
)
(
0
.
0
0
)
(
0
.
0
0
)
(
0
.
0
0
)
P
o
l
i
t
y
-
0
.
0
0
-
0
.
0
0
-
0
.
0
0
-
0
.
0
0
-
0
.
0
0
-
0
.
0
0
(
0
.
0
0
)
(
0
.
0
0
)
(
0
.
0
0
)
(
0
.
0
0
)
(
0
.
0
0
)
(
0
.
0
0
)
l
n
G
D
P
p
c
0
.
4
2
*
*
*
0
.
5
4
*
*
*
0
.
6
2
*
*
*
0
.
4
2
*
*
*
0
.
5
4
*
*
*
0
.
6
2
*
*
*
(
0
.
0
5
)
(
0
.
0
5
)
(
0
.
0
7
)
(
0
.
0
5
)
(
0
.
0
5
)
(
0
.
0
7
)
P
o
p
u
l
a
t
i
o
n
d
e
n
s
i
t
y
-
0
.
0
0
0
.
0
0
*
*
*
0
.
0
0
*
*
-
0
.
0
0
0
.
0
0
*
*
*
0
.
0
0
*
*
(
0
.
0
0
)
(
0
.
0
0
)
(
0
.
0
0
)
(
0
.
0
0
)
(
0
.
0
0
)
(
0
.
0
0
)
G
r
o
w
t
h
-
0
.
0
0
-
0
.
0
0
-
0
.
1
4
*
-
0
.
0
0
-
0
.
0
0
-
0
.
1
4
*
(
0
.
0
4
)
(
0
.
0
4
)
(
0
.
0
7
)
(
0
.
0
4
)
(
0
.
0
4
)
(
0
.
0
7
)
T
i
m
e
t
r
e
n
d
0
.
0
0
0
.
0
1
*
*
*
0
.
0
0
0
.
0
0
0
.
0
1
*
*
*
0
.
0
0
(
0
.
0
1
)
(
0
.
0
0
)
(
0
.
0
0
)
(
0
.
0
1
)
(
0
.
0
0
)
(
0
.
0
0
)
C
o
n
s
t
a
n
t
-
1
5
.
6
8
*
*
*
0
.
0
0
-
1
0
.
0
0
*
*
*
-
1
5
.
6
7
*
*
*
-
6
.
1
6
*
*
*
-
1
1
.
0
3
*
*
*
(
0
.
6
0
)
(
0
.
0
0
)
(
1
.
6
4
)
(
0
.
6
0
)
(
0
.
5
5
)
(
0
.
9
5
)
C
o
u
n
t
r
y

x
e
d
e

e
c
t
s
y
e
s
y
e
s
y
e
s
y
e
s
y
e
s
y
e
s
r
h
o
0
.
7
7
0
.
7
2
0
.
5
3
0
.
7
7
0
.
7
2
0
.
5
4
O
b
s
e
r
v
a
t
i
o
n
s
2
5
7
2
2
6
6
3
1
1
8
3
2
5
7
2
2
6
6
3
1
1
8
3
R
2
0
.
9
6
0
.
8
7
0
.
9
9
0
.
9
6
0
.
8
7
0
.
9
9
N
u
m
b
e
r
o
f
c
o
u
n
t
r
i
e
s
1
0
6
1
1
0
9
4
1
0
6
1
1
0
9
4
S
t
a
n
d
a
r
d
e
r
r
o
r
s
i
n
p
a
r
e
n
t
h
e
s
e
s
;
*
*
*
p
<
0
.
0
1
,
*
*
p
<
0
.
0
5
,
*
p
<
0
.
1
124
CHAPTER 6. IGO MEMBERSHIP AND ENVIRONMENTAL
QUALITY: UNDERSTANDING THE MECHANISMS
6.1.2 Accounting for the Trend in IGO Membership
Although the test statistic for non-stationarity of panel data shows that
all independent variables are stationary, each countrys IGO membership
nevertheless grows over time without usually decreasing. Therefore, it could
be possible that the eect of IGO membership is merely due to this time
trend and is not driven by IGO membership per se. To show that this is not
the case, I created a variable that measures the deviation of each countrys
IGO membership from the average IGO membership per year. Positive
deviations thus imply that this country is a member of more IGOs than
the average country in this year whereas negative deviations imply that this
country is a member of fewer IGOs than the average country.
The results in table 6.2 show the same results as in the case of the general
IGO variable: countries that are members of more than the average number
of IGOs are better caretakers of their environment. Accordingly, we can
exclude the possibility that the time trend underlying IGO membership is
causing the reduction in environmental pollution in developing countries.
125
CHAPTER 6. IGO MEMBERSHIP AND ENVIRONMENTAL
QUALITY: UNDERSTANDING THE MECHANISMS
Table 6.2: IGO membership in deviations from the mean sample: all
developing countries
(1) (2) (3)
SO
2
CO
2
BOD
Deviation of IGO membership -0.02*** -0.01*** -0.01***
(0.00) (0.00) (0.00)
Aid 0.00 -0.00 -0.00
(0.00) (0.00) (0.00)
ln trade openness 0.06** 0.09*** -0.01
(0.03) (0.01) (0.02)
FDI % GDP 0.00 0.00 0.00
(0.00) (0.00) (0.00)
Polity -0.00 -0.00** -0.00
(0.00) (0.00) (0.00)
ln GDP pc 0.53*** 0.54*** 0.61***
(0.05) (0.04) (0.07)
Population density -0.00 0.00*** 0.00***
(0.00) (0.00) (0.00)
Growth 0.01 -0.00 -0.10
(0.04) (0.04) (0.09)
Time trend 0.01 0.01*** 0.00
(0.00) (0.00) (0.00)
Constant -16.37*** 0.00 -11.95***
(0.61) (0.00) (0.79)
Country xed eects yes yes yes
rho 0.78 0.74 0.61
Observations 2677 2764 1200
Number of countries 106 110 94
R
2
0.96 0.86 0.99
Standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1
126
CHAPTER 6. IGO MEMBERSHIP AND ENVIRONMENTAL
QUALITY: UNDERSTANDING THE MECHANISMS
6.2 Disaggregating the Eect of IGO Membership
6.2.1 Disaggregating IGO Membership according to its Func-
tion
Measuring IGO membership using a count variable of a countrys IGO mem-
berships implies that we assume that all IGOs aect environmental quality
to the same extent (Boehmer et al. 2004). Although in general, the above-
mentioned theoretical reasons for the inuence of IGOs apply to all dierent
types of IGOs, some arguments are more important for specic IGOs than
for others. The possibility of issue linkage, for example, is more pronounced
in the umbrella type of IGOs since they deal with a great variety of issues.
Technology transfer, in contrast, is more likely to happen in more technically
oriented IGOs such as agrarian or environmental IGOs. Hence although the
above-mentioned theoretical arguments imply that all types of IGOs inu-
ence environmental performance positively, some IGOs may have a stronger
inuence than others.
Following closely the typology and the coding instructions of Ingram
et al. (2005) I coded, as described in section 4.4.2.1, all IGOs according to
their mandate into 10 dierent categories: military (e.g. NATO), umbrella
(e.g. ASEAN, EU), economic cooperation (e.g. World Bank), economic
general (e.g. WTO, IMF), economic standardization (e.g. Caribbean Postal
Union), economic industry specic (e.g. Cocoa Producers Alliance), so-
cial (e.g. WHO, ILO, UNESCO), environmental (e.g. Global Environment
Facility, GEF), nuclear (e.g. IAEA), and agricultural (e.g. FAO) organiza-
tions.
Table 6.3 shows the results using this disaggregated IGO measure instead
of the general count variable. Interestingly, only membership in umbrella
and military IGOs exerts a signicantly negative eect on pollution over
all three environmental indicators while controlling for membership in any
of the other IGO types. Membership in general economic IGOs is of great
importance when considering air quality but it has no eect in the case of
water pollution. In contrast, agricultural IGOs are only associated with
better water quality and not with air pollution. Social and nuclear IGOs
are only linked to lower SO
2
emissions whereas economic standardization
IGOs are only associated to lower CO
2
emissions. Surprisingly, environ-
mental organizations, with the exception of CO
2
, are not linked to better
environmental quality.
127
CHAPTER 6. IGO MEMBERSHIP AND ENVIRONMENTAL
QUALITY: UNDERSTANDING THE MECHANISMS
Table 6.3: IGO membership disaggregated according to IGO function
sample: all developing countries
(1) (2) (3)
SO
2
CO
2
BOD
Environmental IGOs 0.01 -0.01* -0.00
(0.01) (0.01) (0.01)
Umbrella IGOs -0.04*** -0.04*** -0.05**
(0.01) (0.01) (0.02)
Military IGOs -0.12*** -0.05*** -0.05***
(0.03) (0.02) (0.02)
Economic IGOs (development) -0.01 0.01 -0.02
(0.01) (0.01) (0.01)
Economic IGOs (general) -0.01** -0.01* 0.01
(0.01) (0.01) (0.01)
Economic IGOs (standardization) -0.02 -0.02** -0.01
(0.01) (0.01) (0.01)
Economic IGOs (industry specic) 0.01* 0.00 0.01
(0.01) (0.01) (0.01)
Social IGOs -0.00* -0.00 -0.00
(0.00) (0.00) (0.00)
Agricultural IGOs 0.01 0.02* -0.05***
(0.01) (0.01) (0.02)
Nuclear IGOs -0.14*** -0.02 -0.00
(0.04) (0.02) (0.02)
Aid 0.00 0.00 -0.00
(0.00) (0.00) (0.00)
ln trade openness 0.04 0.10*** -0.01
(0.03) (0.02) (0.03)
FDI % GDP 0.00 -0.00 -0.00
(0.00) (0.00) (0.00)
Polity -0.00 -0.00 -0.00
(0.00) (0.00) (0.00)
ln GDP pc 0.39*** 0.54*** 0.57***
(0.05) (0.05) (0.09)
Population density -0.00*** 0.00*** 0.00***
(0.00) (0.00) (0.00)
Growth 0.01 0.00 -0.09
(0.04) (0.04) (0.09)
Time trend -0.00 0.01** -0.00
(0.00) (0.00) (0.00)
Constant 0.00 -5.77*** 0.00
(0.00) (0.54) (0.00)
Country xed eects yes yes yes
rho 0.77 0.70 0.64
Observations 2572 2663 1183
R
2
0.96 0.89 0.96
Number of countries 106 110 94
Standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1
128
CHAPTER 6. IGO MEMBERSHIP AND ENVIRONMENTAL
QUALITY: UNDERSTANDING THE MECHANISMS
To better compare the eect sizes of the dierent IGO types, gures
6.1 - 6.3 show for each type of IGO membership its estimated eect on
environmental pollution and the condence interval around this estimate. If
the condence interval crosses the vertical line placed at zero, the coecient
of the respective variable is not statistically dierent from zero. In the
case of SO
2
emissions, we see that ve variables are to the left of the zero
line, indicating a signicantly negative eect on emissions. Both nuclear
and military organizations exert the strongest eect of all organizations,
followed by umbrella organizations
1
. The eect of umbrella organizations
is still greater than both the eect of general economic organizations and
the eect of social organizations, whereas the eect of the latter two is not
statistically dierent from each other
2
.
Environmental IGOs
Environmental IGOs
Environmental IGOs Umbrella IGOs
Umbrella IGOs
Umbrella IGOs Military IGOs
Military IGOs
Military IGOs Economic IGOs (development)
Economic IGOs (development)
Economic IGOs (development) Economic IGOs (general)
Economic IGOs (general)
Economic IGOs (general) Economic IGOs (standardization)
Economic IGOs (standardization)
Economic IGOs (standardization) Economic IGOs (industry)
Economic IGOs (industry)
Economic IGOs (industry) Social IGOs
Social IGOs
Social IGOs Agricultural IGOs
Agricultural IGOs
Agricultural IGOs Nuclear IGOs
Nuclear IGOs
Nuclear IGOs -.2
-.2
-.2 -.15
-.15
-.15 -.1
-.1
-.1 -.05
-.05
-.05 0
0
0.05
.05
.05 Marginal effect in percentage points
Marginal effect in percentage points
Marginal effect in percentage points Dependent Variable: SO2
Dependent Variable: SO2
Dependent Variable: SO2
Figure 6.1: Marginal eect of IGO membership according to its function
SO
2
Turning to CO
2
emissions we see that this time, the eect of military
and of umbrella organizations is greatest (both eects are not statistically
dierent from each other). The two organizations are followed by economic
organizations dealing with standardization processes. However, the eect of
this type of economic organization can only be signicantly dierentiated
from military organizations but not from umbrella organizations. Although
the eect of environmental IGOs seems to be smaller than the eect of
economic organizations as shown in gure 6.2, the dierence between the
coecients is not statistically dierent. In contrast, the eect of social
1
Using a t-test one can see that the eect of military or nuclear organizations is statis-
tically dierent from the eect of umbrella organizations, implying that the former eects
are indeed greater than the eect of umbrella organizations
2
If I characterize an eect as greater than another eect this implies that a one-sided
t-test has been conducted.
129
CHAPTER 6. IGO MEMBERSHIP AND ENVIRONMENTAL
QUALITY: UNDERSTANDING THE MECHANISMS
IGOs is indeed smaller than the eect of economic organizations but is not
distinguishable from the eect of environmental organizations.
Environmental IGOs
Environmental IGOs
Environmental IGOs Umbrella IGOs
Umbrella IGOs
Umbrella IGOs Military IGOs
Military IGOs
Military IGOs Economic IGOs (development)
Economic IGOs (development)
Economic IGOs (development) Economic IGOs (general)
Economic IGOs (general)
Economic IGOs (general) Economic IGOs (standardization)
Economic IGOs (standardization)
Economic IGOs (standardization) Economic IGOs (industry)
Economic IGOs (industry)
Economic IGOs (industry) Social IGOs
Social IGOs
Social IGOs Agricultural IGOs
Agricultural IGOs
Agricultural IGOs Nuclear IGOs
Nuclear IGOs
Nuclear IGOs -.1
-.1
-.1 -.05
-.05
-.05 0
0
0.05
.05
.05 Marginal effect in percentage points
Marginal effect in percentage points
Marginal effect in percentage points Dependent Variable: CO2
Dependent Variable: CO2
Dependent Variable: CO2
Figure 6.2: Marginal eect of IGO membership according to its function
CO
2
Finally, gure 6.3 shows the dierent eect sizes for water quality. Only
three IGO types, namely umbrella, military and agricultural organizations,
exert a statistically signicant and negative inuence on water pollution.
However, using a t-test shows that none of the three coecients is signi-
cantly greater than any of the other coecients, which does not allow us to
rank order these organizations according to their eect size.
Environmental IGOs
Environmental IGOs
Environmental IGOs Umbrella IGOs
Umbrella IGOs
Umbrella IGOs Military IGOs
Military IGOs
Military IGOs Economic IGOs (development)
Economic IGOs (development)
Economic IGOs (development) Economic IGOs (general)
Economic IGOs (general)
Economic IGOs (general) Economic IGOs (standardization)
Economic IGOs (standardization)
Economic IGOs (standardization) Economic IGOs (industry)
Economic IGOs (industry)
Economic IGOs (industry) Social IGOs
Social IGOs
Social IGOs Agricultural IGOs
Agricultural IGOs
Agricultural IGOs Nuclear IGOs
Nuclear IGOs
Nuclear IGOs -.1
-.1
-.1 -.05
-.05
-.05 0
0
0.05
.05
.05 Marginal effect in percentage points
Marginal effect in percentage points
Marginal effect in percentage points Dependent Variable: BOD
Dependent Variable: BOD
Dependent Variable: BOD
Figure 6.3: Marginal eect of IGO membership according to its function
BOD
130
CHAPTER 6. IGO MEMBERSHIP AND ENVIRONMENTAL
QUALITY: UNDERSTANDING THE MECHANISMS
In addition to being interesting in its own right, disaggregating IGO
membership can provide some insight into the underlying mechanisms of how
membership in IGOs may aect environmental performance in developing
countries. The strong eect of umbrella organizations such as the EU or
ASEAN, for example, is a rst indication that some IGOs seem to be very
successful in connecting dierent issues. Although environmental protection
is not usually the main goal of these umbrella organizations, countries joining
these IGOs for economic or political reasons are also inuenced with regard
to their environmental quality. Moreover, the strong eect of social IGOs in
the case of SO
2
emissions seems to hint at two dierent mechanisms. On the
one hand, these organizations may constitute a good forum for information
dissemination, and through the socialization process that takes place at
these fora they may also diuse environmental protection as proposed in the
literature on policy diusion (Cao forthcoming; Simmons & Elkins 2004).
On the other hand, many of these organizations such as the WHO and ILO or
research related organizations such as the International Commission for the
Scientic Exploration of the Mediterranean Sea might diuse knowledge and
technology helpful to the environment in general while aiming to improve
or preserve human health, labour conditions or biodiversity.
This last mentioned mechanism of technology and knowledge transfer
might also underlie the eect of agricultural organizations. Since one major
goal of these organizations is to promote sustainable agriculture, as in the
case of the FAO, they also improve environmental quality as a potential side
eect. Interestingly, technology transfer and capacity building are also two
of the tasks of some of the nuclear IGOs, as described in the four case stud-
ies below. Perhaps more importantly, however, nuclear IGOs such as the
International Atomic Energy Agency (IAEA) or the Latin American Energy
Organization (OLADE) mostly promote the use of nuclear energy produc-
tion. By relying on nuclear energy production instead of energy production
based on fossil fuels, countries that opt for nuclear energy production emit
less SO
2
and CO
2
3
.
In contrast, the negligible impact of environmental organizations seems
to suggest that enhancing environmental quality in developing countries by
establishing IGOs that pursue environmental interests is not very promising.
Although by joining environmental IGOs, countries commit to the idea of
taking care of their environment, they do not seem to act in accordance with
this idea at the national level. However, it is important to note that this
result does not necessarily imply that environmental regimes are ineective.
In the analysis at hand, only those regimes are included that are considered
to be an international organization, implying that they need to possess an
international secretariat. This is only true for a very limited number and for
3
Of course, this does not reect the fact that nuclear energy production produces
nuclear waste, which itself constitutes a major environmental problem.
131
CHAPTER 6. IGO MEMBERSHIP AND ENVIRONMENTAL
QUALITY: UNDERSTANDING THE MECHANISMS
only a very specic set of environmental regimes
4
. Many of these organiza-
tions, however, pursue goals such as combating locust pests in Africa that
are not related to any of the measures of environmental quality used in this
analysis. Hence even if these organizations were eective for the purpose
for which they were created, they would not aect air and water quality
because of their specialization. In contrast, other important environmental
regimes such as the Convention on Long-Range Transboundary Air Pollu-
tion are not included in this sample because they are not considered to be
an international organization. Hence the results of this analysis cannot be
interpreted as evidence against the ineectiveness of environmental regimes
in general.
However, this non-existing inuence of environmental IGOs on environ-
mental quality also has a positive side as it allows us to assess the problem
of endogeneity. One potential risk of looking at IGO membership and its
inuence on environmental performance is that, at least concerning environ-
mental IGOs, countries may self-select into these IGOs. This would imply
that, for example, only countries with a very good environmental record
would join environmental IGOs because these countries can be sure to meet
the obligations prescribed by the respective organization. In contrast, one
could also argue that countries with major environmental problems may
be more willing to join environmental organizations. On the one hand,
environmental organizations could provide these laggard countries with ad-
equate technology and know-how to ght their environmental problems and
may therefore provide these countries with an incentive to join these orga-
nizations. On the other hand, polluting countries may join environmental
organizations in order to signal to their citizenry that henceforth, they are
planning to take the environment more seriously.
Since environmental IGOs are not associated with environmental quality,
with the exception of CO
2
emissions, we can be rather condent that coun-
tries do not self-select into these organizations. In contrast, the generally
positive eect of IGO membership on environmental performance is, ac-
cording to the results in table 6.3, due above all to non-environmental IGOs
such as umbrella and social IGOs. Consequently, membership in IGOs does
indeed appear to constitute a channel through which developing countries
are urged to take better care of their environment by providing them with
technologies and resources necessary to reduce pollution.
Comparing the dierent types of IGOs, as in table 6.3, to some extent
4
For example, the set of environmental IGOs includes organizations such as the In-
ternational Commission of the Northwest Atlantic Fisheries, now called the Northwest
Atlantic Fisheries Organization, the International Whaling Commission, the Joint Anti-
Locust and Anti-Aviarian Organization, which aims to control the damage arising from
locust and aviarian pests, the International Seabed Authority, which regulates deep seabed
mining and aims to protect the marine environment, or Bionet, whose purpose it is to clas-
sify natural species.
132
CHAPTER 6. IGO MEMBERSHIP AND ENVIRONMENTAL
QUALITY: UNDERSTANDING THE MECHANISMS
masks the fact that an equal number of economic, social or environmental
organizations does not exist. Hence most countries are members of more
economic and social IGOs as compared to environmental or agricultural
IGOs, mainly because there are many more economic and social IGOs to
join in the rst place. One possibility to control for this is to measure IGO
membership as a percentage of all possible IGO memberships in a specic
category (economic, environmental etc.).
As we can see in table 6.4, the results change to some extent if IGO
membership is measured as a percentage of all possible IGO memberships.
Membership in general umbrella IGOs is no longer signicant for SO
2
emis-
sions and water quality whereas it is still signicantly associated with CO
2
emissions. Interestingly, economic organizations that promote cooperation
and development are connected with an increase in CO
2
emissions but with
a decrease in BOD. Similarly, agricultural organizations seem to be bad for
SO
2
whereas they have an emission reducing eect in the case of water qual-
ity. Hence if we control for the potential number of organizations in each
category, we see that there are organizations that seem to have a mixed
impact on environmental quality. Although membership in agricultural or
economic organizations furthering development are connected to improve-
ments in water quality, they have adverse eects on SO
2
or CO
2
emissions.
133
CHAPTER 6. IGO MEMBERSHIP AND ENVIRONMENTAL
QUALITY: UNDERSTANDING THE MECHANISMS
Table 6.4: IGO membership in % of possible IGO membership sample: all
developing countries
(1) (2) (3)
SO
2
CO
2
BOD
% environmental IGOs 0.25 -0.57** 0.15
(0.31) (0.26) (0.32)
% umbrella IGOs -0.30 -0.67*** -0.57
(0.23) (0.22) (0.35)
% military IGOs -1.43*** -0.79*** -0.38*
(0.40) (0.21) (0.21)
% economic IGOs (development) -0.40 0.45* -1.23**
(0.27) (0.27) (0.57)
% economic IGOs (general) -0.43 -0.55** 0.38
(0.27) (0.26) (0.35)
% economic IGOs (standardization) -1.02** -0.58** -0.51
(0.41) (0.27) (0.32)
% economic IGOs (industry) 0.12 0.45 -0.19
(0.32) (0.38) (0.41)
% social IGOs -0.80 -1.29 -2.16***
(0.70) (0.79) (0.71)
% agricultural IGOs 0.31** 0.17 -0.42*
(0.15) (0.14) (0.24)
% nuclear IGOs -0.57*** -0.01 -0.04
(0.16) (0.08) (0.09)
Aid -0.00 0.00 0.00
(0.00) (0.00) (0.00)
ln trade openness 0.04 0.10*** 0.01
(0.03) (0.02) (0.02)
FDI % GDP 0.00 -0.00 -0.01*
(0.00) (0.00) (0.00)
Polity -0.00 -0.00 -0.00
(0.00) (0.00) (0.00)
ln GDP pc 0.40*** 0.54*** 0.62***
(0.05) (0.05) (0.07)
Population density -0.00** 0.00*** 0.00**
(0.00) (0.00) (0.00)
Growth -0.00 -0.00 -0.15**
(0.04) (0.04) (0.07)
Time trend -0.01*** 0.00 -0.01**
(0.00) (0.00) (0.00)
Constant -15.01*** -5.92*** -9.68***
(0.58) (0.54) (3.07)
Country Fixed Eects yes yes yes
rho 0.77 0.70 0.63
Observations 2572 2663 1183
R
2
0.96 0.88 0.99
Number of countries 106 110 94
Standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1
134
CHAPTER 6. IGO MEMBERSHIP AND ENVIRONMENTAL
QUALITY: UNDERSTANDING THE MECHANISMS
As above, gures 6.4 - 6.6 provide graphical illustrations of the dierent
eect sizes of the various types of IGO memberships. In the case of SO
2
emissions, only military, economic general and nuclear organizations are
to the left of the zero line whereas umbrella and social IGOs, which were
signicant in the model before as shown in gure 6.1, cross the zero line
this time. Furthermore, as was the case in the model above, the eect of
military organizations is statistically greater than the eect of both economic
general and nuclear organizations, whereas the latter two are not statistically
dierent from each other. Hence after controlling for the number of IGOs
that exist per category, we see that the eect of nuclear organizations is now
smaller compared to the eect of military organizations whereas before, the
two eects were the same.
% Environmental IGOs
% Environmental IGOs
% Environmental IGOs % Umbrella IGOs
% Umbrella IGOs
% Umbrella IGOs % Military IGOs
% Military IGOs
% Military IGOs % Economic IGOs (development)
% Economic IGOs (development)
% Economic IGOs (development) % Economic IGOs (general)
% Economic IGOs (general)
% Economic IGOs (general) % Economic IGOs (standardization)
% Economic IGOs (standardization)
% Economic IGOs (standardization) % Economic IGOs (industry)
% Economic IGOs (industry)
% Economic IGOs (industry) % Social IGOs
% Social IGOs
% Social IGOs % Agricultural IGOs
% Agricultural IGOs
% Agricultural IGOs % Nuclear IGOs
% Nuclear IGOs
% Nuclear IGOs -2
-2
-2 -1
-1
-1 0
0
01
1
1Marginal effect in percentage points
Marginal effect in percentage points
Marginal effect in percentage points Dependent Variable: SO2
Dependent Variable: SO2
Dependent Variable: SO2
Figure 6.4: Marginal eect of % of IGO membership according to its function
SO
2
Concerning CO
2
emissions we see a less drastic change than for SO
2
emissions. Compared to the above model, agricultural organizations are no
longer associated with an increase in CO
2
emissions but economic devel-
opment organizations are. As could be expected by looking at gure 6.5,
all of the eect sizes are very similar, implying that we cannot dierentiate
between the eects of military, umbrella, environmental, general economic
and economic standardization organizations.
Finally, gure 6.6 shows the dierent eect sizes for water quality. Whereas
in the model above, in which we did not control for the potential number
of each organization per category, only three types of organizations were
inuential (umbrella, military and agricultural), now after controlling for
the number of IGOs four types of IGOs seem to play a role: military, agri-
cultural, social and economic organizations that promote development. So
whereas umbrella organizations do not seem to be inuential any more, so-
135
CHAPTER 6. IGO MEMBERSHIP AND ENVIRONMENTAL
QUALITY: UNDERSTANDING THE MECHANISMS
% Environmental IGOs
% Environmental IGOs
% Environmental IGOs % Umbrella IGOs
% Umbrella IGOs
% Umbrella IGOs % Military IGOs
% Military IGOs
% Military IGOs % Economic IGOs (development)
% Economic IGOs (development)
% Economic IGOs (development) % Economic IGOs (general)
% Economic IGOs (general)
% Economic IGOs (general) % Economic IGOs (standardization)
% Economic IGOs (standardization)
% Economic IGOs (standardization) % Economic IGOs (industry)
% Economic IGOs (industry)
% Economic IGOs (industry) % Social IGOs
% Social IGOs
% Social IGOs % Agricultural IGOs
% Agricultural IGOs
% Agricultural IGOs % Nuclear IGOs
% Nuclear IGOs
% Nuclear IGOs -3
-3
-3 -2
-2
-2 -1
-1
-1 0
0
01
1
1Marginal effect in percentage points
Marginal effect in percentage points
Marginal effect in percentage points Dependent Variable: CO2
Dependent Variable: CO2
Dependent Variable: CO2
Figure 6.5: Marginal eect of % of IGO membership according to its function
CO
2
cial and economic development organizations seem to matter. In addition,
social organizations are now in fact the most inuential IGOs, with their ef-
fect being greater than all of the other organizations. In contrast, the eect
of military, agricultural and economic development organizations cannot be
dierentiated.
% Environmental IGOs
% Environmental IGOs
% Environmental IGOs % Umbrella IGOs
% Umbrella IGOs
% Umbrella IGOs % Military IGOs
% Military IGOs
% Military IGOs % Economic IGOs (development)
% Economic IGOs (development)
% Economic IGOs (development) % Economic IGOs (general)
% Economic IGOs (general)
% Economic IGOs (general) % Economic IGOs (standardization)
% Economic IGOs (standardization)
% Economic IGOs (standardization) % Economic IGOs (industry)
% Economic IGOs (industry)
% Economic IGOs (industry) % Social IGOs
% Social IGOs
% Social IGOs % Agricultural IGOs
% Agricultural IGOs
% Agricultural IGOs % Nuclear IGOs
% Nuclear IGOs
% Nuclear IGOs -4
-4
-4 -3
-3
-3 -2
-2
-2 -1
-1
-1 0
0
01
1
1Marginal effect in percentage points
Marginal effect in percentage points
Marginal effect in percentage points Dependent Variable: BOD
Dependent Variable: BOD
Dependent Variable: BOD
Figure 6.6: Marginal eect of % of IGO membership according to its function
BOD
Following from this eort of dierentiating between the various types of
organizations, we can conclude that except for military organizations, which
seem to be always related to better environmental quality, very dierent or-
136
CHAPTER 6. IGO MEMBERSHIP AND ENVIRONMENTAL
QUALITY: UNDERSTANDING THE MECHANISMS
ganizations are inuential for the three indicators of environmental quality.
In most of these cases, the results help us to better understand the mech-
anisms behind the inuence of IGO membership in developing countries as
described above. For example, the inuence of umbrella organizations is
likely to be traced back to issue linkage, whereas the inuence of agricul-
tural, economic and social organizations may be due to the provision of
technological and nancial resources. Furthermore, social organizations in
particular may provide a good context for information dissemination and the
socialization of countries as proposed in the literature on policy diusion.
What is puzzling about these results is the robust inuence of military
organizations. Consequently, in order to nd out what the driving force
behind this eect is, I split up the military membership variable even further.
More precisely, I created a dummy variable for each military organization in
the sample, such that it is possible to trace back this eect of military IGOs
to particular organizations. Tables 6.5 and 6.6 provide an overview of the
specic organizations that have a signicant eect on environmental quality
in any of the three subsamples. The last column in table 6.6 shows the
number of countries in the sample that are members of this organization
5
. As
we can see, membership varies widely among these organizations, indicating
that we are dealing with both very specic and more broadly positioned
organizations.
In order to understand what might be behind the signicant inuence of
these eight organizations, the next paragraphs provide some more detailed
information on these IGOs. Starting with the Euro-Atlantic Partnership
Council (EAPC) we see that its main purpose is to bring together North
Atlantic Treaty Organization (NATO) member countries and some partner
countries to consult and exchange information on political and security re-
lated issues (NATO 2009). Since most of the partner countries are former
communist countries, the eect of the EAPC is most likely due to the fall
of the Soviet Union and the associated closure of many outdated industrial
plants.
In a similar vein, the signicant eect of the Organization for Security
and Co-operation in Europe (OSCE) probably also comes from its former
communist member countries that have joined the organization since the
early 1990s. The purpose of the organization is to serve as an instrument for
conict prevention, early warning, crisis management and post-conict re-
habilitation in Eastern and South-Eastern Europe and Central Asia (OSCE
2009). Finally, the same logic seems to hold true for the Baltic Council (BC)
with its member states Lithuania, Estonia and Latvia.
In contrast, the Inter-American Defense Board (IADefB), which consists
5
If an organization was only signicant in the sample comprising only low or low-
middle income countries, the number of countries per organization was calculated using
this specic sample only.
137
CHAPTER 6. IGO MEMBERSHIP AND ENVIRONMENTAL
QUALITY: UNDERSTANDING THE MECHANISMS
Table 6.5: Military Organizations
SO
2
CO
2
BOD
low, low-middle and high middle EAPC EAPC EAPC
low, low-middle and high middle OSCE OSCE OSCE
low, low-middle and high middle IADefB IADefB
low, low-middle and high middle CENTO CENTO
low, low-middle and high middle SEATO
low and low-middle EAPC EAPC EAPC
low and low-middle OSCE OSCE
low and low-middle IADefB
low and low-middle SEATO
low and low-middle BC
low and low-middle ICivDO
low income EAPC
low income OSCE OSCE
low income IADefB IADefB
low income ICivDO ICivDO
low income NAM
of countries of the Americas, is closely connected to the Organization of
American States to which it provides information on defense and security
issues concerning the Americas. Since the Organization of American States
has a department dealing with sustainable development, which includes the
promotion of environmental quality, it is likely that the signicant eect of
the Inter-American Defense Board can be attributed to the Organization
of American States more generally and its promotion of sustainable devel-
opment more specically (Organization of American States 2009). Another
potential explanation for the strong eect of the Inter-American Defense
Board might be that its membership correlates strongly with membership
in the Latin American Energy Organization (OLADE). As described above,
this organization diuses the latest energy production technologies and pro-
motes nuclear energy production, which is associated with a decrease in air
pollution (OLADE 2009).
Both the Southeast Asia Treaty Organization and the Central Treaty
Organization were regional defense organizations that were only active for
a short time period. The members of the Central Treaty Organization were
Iraq, Iran, Pakistan,Turkey, the UK and the US. The organization was dis-
solved in 1979. The Southeast Asia Treaty Organization encompassed Aus-
tralia, Bangladesh, France, New Zealand, Pakistan, the Philippines, Suri-
name, Thailand, the UK and the US and aimed to block communist ad-
vancement in Southeast Asia. The alliance was dissolved in 1977. For both
138
CHAPTER 6. IGO MEMBERSHIP AND ENVIRONMENTAL
QUALITY: UNDERSTANDING THE MECHANISMS
Table 6.6: Military Organizations
Abbreviation Name Number of countries
EAPC Euro-Atlantic Partnership Council 21
OSCE Organization Security Cooperation Europe 26
IADefB Inter-American Defense Board 23
CENTO Central Treaty Organization 3
SEATO Southeast Asia Treaty Organization 3
BC Baltic Council 2
ICivDO International Civil Defence Organization 40
NAM Non-Aligned Movement 50
organizations, the connection to environmental quality is not obvious at all.
The Non-Aligned Movement was founded in 1955 by a number of coun-
tries that did not consider themselves as part of any international military
alliance. These countries aimed to preserve their sovereignty and security
and opposed colonialism and imperialism (NAM 2009a). Since its establish-
ment the organization has also focused on other, non-security related issues,
such as human rights and economic development. More importantly, the
organization, in one of its resolutions, subscribes to the ideal of sustainable
development, which might be one important reason underlying its positive
eect on environmental quality (NAM 2009b).
Finally, the objective of the International Civil Defence Organization
(ICivDO) is
to contribute to the development by States of structures ensur-
ing the protection and assistance of populations and safeguard-
ing property and the environment from natural or man-made
disasters (ICivDO 2009a).
More precisely, the organization aims to gather and provide
information on organisation protection and operation princi-
ples with respect to the hazards threatening populations in case
of oods, earthquakes, avalanches, major res, hurricanes, dam
bursts and other kinds of destruction as well as from air and wa-
ter pollution or attacks with modern means of warfare (ICivDO
2009b).
Hence the International Civil Defence Organization provides its member
countries with both knowledge and technology that are directly related to
pollution control and can thus assist developing countries as well in tackling
environmental degradation.
139
CHAPTER 6. IGO MEMBERSHIP AND ENVIRONMENTAL
QUALITY: UNDERSTANDING THE MECHANISMS
Altogether, we can see that these eight organizations seem to be con-
nected to environmental quality for very dierent reasons. Some organi-
zations such as the OSCE or the Euro-Atlantic Partnership Council seem
to be mere proxies for the improvement in environmental quality that has
been achieved by the former communist countries since the fall of the Soviet
Union. For other organizations, especially the Central Treaty Organization
and the Southeast Asia Treaty Organization, a direct link to the environ-
mental performance of its member countries is hard to imagine. However, in
the two cases of the Non-Aligned Movement and the International Civil De-
fence Organization, it seems that the transfer of technology and know-how
as well as the promotion of sustainable development are driving the eect
of these military organizations.
This shows that some IGOs serve environmental purposes, as in the case
of the International Civil Defence Organization or the Non-Aligned Move-
ment, although we would not have expected it. This nding suggests that it
might be advisable to code each organization not only according to its major
function (economic, social, environmental) but according to whether the or-
ganization subscribes to some environmental mandate. Hence I recoded all
organizations according to whether they serve some environmental purpose.
All organizations that mention in their charter or their mission statement
that their aims include environmental protection, sustainable development
or clean technology transfer were dened as organizations with an environ-
mental function
6
.
However, as the results in table 6.7 show, only in relation to CO
2
emis-
sions and water quality in low-income countries is the eect of IGOs with an
environmental function indeed greater than the eect of organizations with-
out any environmental functions using a t-test. This result is very surprising
as it suggests that an organization does not necessarily need to pursue any
environmental function to be associated with an increase in environmental
quality in developing countries. This could imply several things: the social-
ization process towards environmentally sustainable development may well
take place even in organizations that are not directly linked to any environ-
mental purposes. Or some organizations promote environmental purposes
without making it obvious in their charter or mission statement, such that
they fall through the coding grid and are not labeled as organizations with
an environmental function. Finally and most likely, it could simply be the
case that due to the high correlation of membership in organizations with
and without any environmental function, the statistical estimation proce-
dure cannot distinguish the eect sizes very well.
6
For example, both the above-mentioned organizations, the Non-Aligned Movement
and the International Civil Defence Organization, would fall in this category of organiza-
tions with an environmental function. This is also true of the World Bank, which is an
important actor in the eld of environmental development aid.
140
CHAPTER 6. IGO MEMBERSHIP AND ENVIRONMENTAL
QUALITY: UNDERSTANDING THE MECHANISMS
Table 6.7: IGO membership disaggregated according to whether the IGO
has some environmental function sample: all developing countries
(1) (2) (3) (4)
SO
2
1
CO
2
1
BOD
1
BOD
2
IGO with env. func. -0.01** -0.01*** -0.02*** -0.01*
(0.01) (0.00) (0.00) (0.01)
IGO without env. func. -0.01*** -0.00 -0.01*** -0.00
(0.00) (0.00) (0.00) (0.00)
Aid -0.00 0.00 0.00 -0.00
(0.00) (0.00) (0.00) (0.00)
ln trade openness 0.04 0.10*** 0.01 0.01
(0.03) (0.02) (0.02) (0.03)
FDI % GDP 0.00 -0.00 -0.01** -0.02***
(0.00) (0.00) (0.00) (0.01)
Polity -0.00 -0.00 -0.00 0.00
(0.00) (0.00) (0.00) (0.00)
ln GDP pc 0.42*** 0.54*** 0.62*** 0.61***
(0.05) (0.05) (0.07) (0.12)
Population density -0.00 0.00*** 0.00** 0.01***
(0.00) (0.00) (0.00) (0.00)
Growth -0.01 -0.01 -0.14** -0.17
(0.04) (0.04) (0.07) (0.15)
Time trend 0.00 0.01*** 0.00 -0.02***
(0.01) (0.00) (0.00) (0.01)
Constant -15.68*** 0.00 -10.93*** -11.47***
(0.61) (0.00) (0.87) (0.93)
Country Fixed Eects yes yes yes yes
rho 0.77 0.72 0.51 0.62
Observations 2572 2663 1183 370
R
2
0.96 0.87 0.99 1.00
Number of countries 106 110 94 40
Standard errors in parentheses; *** p<0.01, ** p<0.05, * p<0.1
1
sample: all developing countries
2
sample: only low-incomce countries
6.2.2 Disaggregating IGO Membership according to its De-
gree of Institutionalization
The last section showed that taking into account the dierent functions of
international organizations can provide us with some more information con-
cerning the mechanisms behind the eect of IGOs on environmental qual-
ity in developing countries. However, IGOs do not only dier according
to their function; they also dier according to their degree of institution-
alization (Boehmer et al. 2004). The degree of institutionalization partly
determines to what extent an organization has the capabilities to inuence
its member countries. While minimalist organizations lack the capabilities
141
CHAPTER 6. IGO MEMBERSHIP AND ENVIRONMENTAL
QUALITY: UNDERSTANDING THE MECHANISMS
to exert power over their member countries, both structured and interven-
tionist IGOs could in principle limit member states in their sovereignty.
Hence we expect that structured and interventionist IGOs possess the in-
stitutional requirements to inuence environmental quality in developing
countries whereas minimal IGOs should not exert any signicant inuence.
Relying on the data of Boehmer et al. (2004)
7
it is possible to dieren-
tiate between three degrees of institutionalization: minimal, structured and
interventionist IGOs. Using these three measurements of IGO membership
in our model, we see that interestingly, it is not always the most institu-
tionalized IGOs that are signicantly associated with better environmental
quality. As table 6.8 shows, interventionist IGOs are the most important
organizations only in the case of SO
2
emissions. In the case of both CO
2
and
BOD emissions, structured IGOs, i.e. those IGOs with a medium degree of
institutionalization, seem to be decisive. As expected, we see that minimal
organizations are never associated with a signicant reduction in pollution
emissions.
We can conclude that some degree of formalization or institutionaliza-
tion of an organization is needed to aect environmental quality in devel-
oping countries. However, it is not necessarily the most institutionalized or
powerful organizations, in the sense that they can constrain their member
countries the most, that aect environmental quality in developing coun-
tries. This probably has to do with the fact that the main mechanisms
behind the inuence of IGOs as we saw in the paragraphs above are
technology and knowledge transfer, issue linkage and a broader socialization
process towards sustainable development and environmental protection. Or-
ganizations that fulll these purposes as long as they possess some degree
of institutionalization do not necessarily need to be fully constraining.
7
For a more detailed explanation of the coding concerning the institutionalization of
IGOs see the section on data measurement in the research design chapter.
142
CHAPTER 6. IGO MEMBERSHIP AND ENVIRONMENTAL
QUALITY: UNDERSTANDING THE MECHANISMS
Table 6.8: IGO membership disaggregated according to the degree of insti-
tutionalization sample: all developing countries
(1) (2) (3)
SO
2
CO
2
BOD
Minimal IGOs 0.00 0.00 0.01**
(0.00) (0.00) (0.00)
Structured IGOs -0.02 -0.03*** -0.03***
(0.01) (0.01) (0.01)
Interventionist IGOs -0.02* -0.00 -0.01
(0.01) (0.00) (0.01)
Aid -0.00 0.00 0.00
(0.00) (0.00) (0.00)
ln trade openness 0.03 0.10*** -0.01
(0.03) (0.02) (0.02)
FDI % GDP -0.00 -0.00 -0.01**
(0.00) (0.00) (0.00)
Polity -0.00 -0.00 -0.00
(0.00) (0.00) (0.00)
ln GDP pc 0.43*** 0.54*** 0.57***
(0.05) (0.05) (0.07)
Population density -0.00 0.00*** 0.00**
(0.00) (0.00) (0.00)
Growth -0.01 -0.01 -0.11
(0.04) (0.04) (0.07)
time trend -0.01*** 0.00 -0.01***
(0.00) (0.00) (0.00)
Constant -15.88*** -6.30*** 0.00
(0.60) (0.55) (0.00)
Country Fixed Eects yes yes yes
rho 0.76 0.71 0.80
Observations 2572 2663 1183
R
2
0.96 0.88 0.99
Number of cowcode 106 110 94
Standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1
143
CHAPTER 6. IGO MEMBERSHIP AND ENVIRONMENTAL
QUALITY: UNDERSTANDING THE MECHANISMS
6.3 Endogeneity
As described above, one potential risk of looking at IGO membership and
its inuence on environmental performance is that, at least concerning en-
vironmental IGOs, countries may self-select into these organizations. In
particular, this would mean that a countrys environmental record would
be decisive for its decision whether to join a particular IGO. It is likely, for
example, that countries with very few environmental problems might join
more environmental IGOs because these countries can easily comply with
the obligations of the respective organization. In contrast, it is also plausi-
ble that countries which have some environmental problems may be more
willing to join environmental organizations for two reasons. First, environ-
mental organizations often supply adequate technology and know-how to
ght environmental problems, thus providing an incentive for countries with
environmental problems to join these organizations. Second, joining an en-
vironmental organization can serve as a signal to a countrys citizenry that
henceforth, environmental protection will be on the agenda of governmental
action. Consequently, it is uncertain whether the causality runs from IGO
membership to environmental quality as hypothesized in the theoretical part
or whether the reverse is true.
As a rst step to tackle this problem of endogeneity, I disaggregated
the inuence of IGOs according to their function. Since the endogeneity
problem mainly arises for environmental organizations, it was advisable to
look at the inuence of each dierent type of IGO separately. As we could
see above, environmental IGOs are only signicantly associated with CO
2
emissions, indicating that this endogeneity problem does not seem to play
a huge role. However, in order to fully dispel any doubts concerning the
direction of causality, it is advisable to rely on an instrumental variables
approach.
The diculty with this approach comes with nding a suitable instru-
ment for IGO membership. This means that we need to nd one or more
variables that are closely linked to IGO membership without being related
to environmental quality (Wooldridge 2003). Following studies in the eld
of international relations, two dierent instruments would qualify for the
context at hand: the degree to which a country votes in accordance with
the US in the UN General Assembly and the dierence between a countrys
per capita income measured at purchasing power parity and per capita in-
come measured at ocial exchange rate prices (Barro & Lee 2005; Deger &
Smith 1983; Dreher & Jensen 2007).
The degree to which a country votes in accordance with the US in the
UN General Assembly is often used as an instrument for development aid
(Barro & Lee 2005; Dreher & Jensen 2007). These studies argue that UN
voting is strongly correlated with the ideological proximity of a country to
the US (and therefore also the proximity to the other major industrialized
144
CHAPTER 6. IGO MEMBERSHIP AND ENVIRONMENTAL
QUALITY: UNDERSTANDING THE MECHANISMS
countries). Therefore, UN voting is also correlated with the amount of aid
a country receives from international aid donors and whether a country
can obtain more favorable lending conditions from the IMF or the World
Bank. Following this logic, it can be argued that voting in accordance with
the US is not only a good indicator of whether a country receives more
favorable lending conditions but also whether a country is better politically
integrated into the international system in general. This would imply that
it is reasonable to assume that a countrys membership in international
organizations in general is also correlated with the degree to which a country
votes in accordance to the US. Indeed, if we look at a regression of UN voting
on IGO membership, we see a strong correlation indicated by a statistically
signicant coecient see table 6.9 rst column.
Table 6.9: Testing potential instrumental variables sample: all developing
countries
(1) (2)
IGO membership IGO membership
UN voting 3.71***
(1.21)
Current GDP - GDP PPP 1.72**
(0.75)
Time trend 1.24*** 1.16***
(0.04) (0.06)
Constant 29.93*** 31.01***
(1.42) (1.50)
Country Fixed Eects yes yes
rho 0.82 0.83
Observations 3552 3241
Number of countries 130 118
R
2
0.74 0.70
Standard errors in parentheses; *** p<0.01, ** p<0.05, * p<0.1
Furthermore, the studies using UN voting as an instrument of inter-
national aid argue that voting in accordance with or against the US in
the UN General Assembly does not feedback in any reasonable way to do-
mestic outcomes. More importantly, one cannot convincingly argue how
voting in accordance with the US should inuence a countrys environmen-
tal performance in any reasonable way. Consequently, voting in the UN
General Assembly should constitute a reasonable instrument for measuring
IGO membership and its inuence on environmental quality in developing
countries.
Relying on data from Voeten & Merdzanovic (2009), I have coded a
countrys vote in the UN General Assembly that is in accordance with the
145
CHAPTER 6. IGO MEMBERSHIP AND ENVIRONMENTAL
QUALITY: UNDERSTANDING THE MECHANISMS
US with a 1 and a vote not in accordance with the US with a 0. When a
country abstained from the vote or was absent, I followed the example of
Dreher et al. (2009) and coded this with a 0.5. Since there are many votes
in the UN General Assembly per year, I have summed the measures for each
vote and divided it by the number of UN General Assembly votes in this
year. Thus, the measure of voting in accordance with the US ranges from 0
to 1 with higher values indicating a greater accordance to the US voting in
the UN General Assembly.
Turning to the second potential instrument the dierence between
per capita income measured at ocial exchange rate prices (current GDP)
8
and per capita income measured at purchasing power parity (GDP PPP)
9
Deger & Smith (1983) argue that this dierence reects variations in a
countrys economic structure. More specically, the authors state that
as an economy becomes more open to world market forces and
the importance of trade grows, the dierence between the two
estimates shrinks. [...] This gap reects the degree of integration
into the world economy. (Deger & Smith 1983, 343-346).
Hence the dierence between current GDP and GDP PPP can be seen as a
measure of a countrys international economic integration
10
. Furthermore,
it is reasonable to expect that a country that is economically well integrated
into the international system should also be politically integrated into the
international system. As we can see in the second column of table 6.9,
the dierence between current GDP and GDP PPP is indeed signicantly
associated with IGO membership, making it a good potential instrument
for (measuring) IGO membership. Moreover, as in the case of UN voting,
it is not obvious how the dierence between current GDP and GDP PPP
should aect a countrys environmental performance, making it a second
valid instrument for the context at hand.
Table 6.10 shows the results of the IV regression using both UN voting
and the dierence between current GDP and GDP PPP as exogenous in-
struments for IGO membership
11
. As we can see, IGO membership is still
negatively associated with all three indicators of environmental pollution.
8
A countrys current GDP (or its GDP measured at ocal exchange rates) is the sum
value of all goods and services produced in this country divided by the bilateral average
US exchange rate in that year (Index Mundi 2009b).
9
A countrys GDP measured at purchasing power parity (PPP) is the sum value of all
goods and services produced in this country evaluated at prices prevailing in the United
States (Index Mundi 2009a).
10
For a similar argument see also Krugman & Obstfeld (2006, 381)
11
As we can see, the democracy variable becomes signicant in the IV regressions.
This is the case because using the IV regression does not allow for the Prais-Winsten
transformation to control for autocorrelation. As autocorrelation, which is denitely a
problem in these data, aects the estimates of the standard errors since we have positive
autocorrelation here, we underestimate the standard errors and hence the variables seem
to be more signicant than they are these standard errors should be interpreted with
146
CHAPTER 6. IGO MEMBERSHIP AND ENVIRONMENTAL
QUALITY: UNDERSTANDING THE MECHANISMS
However, the coecient loses statistical signicance in the case of water qual-
ity. Hence these results conrm that IGO membership does indeed bring
about a reduction in environmental degradation at least as regards SO
2
and
CO
2
emissions.
Using a Wald test to see whether IGO membership is indeed endogenous
to environmental quality in developing countries, we need to reject the null
hypothesis of exogeneity for all three environmental indicators (Wooldridge
2003, 507). This implies that IGO membership is at least to some extent
endogenous to environmental quality in developing countries. However, the
results in table 6.10 lend support to our earlier conclusions that IGO mem-
bership does indeed reduce air quality in developing countries. Concerning
water quality we need to be more cautious since the results of the IV regres-
sion are no longer statistically signicant.
To see how valuable our instruments of IGO memberships are, we can
compare the regressions with and without the instruments. This comparison
shows that the instruments explain about 0.02% of the variance. Although
this does not sound very much, when compared to the variance explained by
the other variables in the model, we see that the instruments explain more
variance than economic growth, trade openness, polity or international aid
respectively. Population density, in contrast, explains approximately the
same variance as the instruments, and GDP per capita is the only variable
to explain more of the variance in environmental quality.
In the introduction to this section on endogeneity, it was argued that the
endogeneity mainly arises because of environmental organizations. Hence to
further evaluate the robustness of the earlier results on the disaggregation
of IGO membership, I use the instruments proposed above to explain envi-
ronmental IGO membership only. As we can see in table 6.11, the results
of the instrumental variables regression indicates that after controlling for
the potential endogeneity of environmental IGOs, membership in these or-
ganizations does not seem to be statistically associated to air quality in
developing countries. Only in the case of water quality do environmental
IGOs seem to signicantly reduce pollution.
Again using a Wald test to see whether environmental IGO membership
is endogenous to our three indicators of environmental degradation, we need
to reject the null hypothesis of exogeneity for both SO
2
emissions and water
quality. In contrast, environmental IGO membership does not seem to be
endogenous to CO
2
emissions. As IV regressions always perform worse under
exogeneity than compared to least squares regression in the sense that IV
regression is less ecient this result indicates that we should only consider
the IV results in the case of SO
2
and BOD but rely on the conventional
caution. As the only goal of the IV regression is to evaluate the endogeneity of IGO
membership it is advisable not to rely on these results with respect to the democracy
variable.
147
CHAPTER 6. IGO MEMBERSHIP AND ENVIRONMENTAL
QUALITY: UNDERSTANDING THE MECHANISMS
Table 6.10: Regression instrumenting IGO membership sample: all devel-
oping countries
(1) (2) (3)
SO
2
CO
2
BOD
Instrumented IGO membership -0.03*** -0.02*** -0.02
(0.01) (0.01) (0.01)
Aid -0.00 -0.00 -0.00*
(0.00) (0.00) (0.00)
ln trade openness 0.13*** 0.21*** 0.04
(0.02) (0.02) (0.05)
FDI % GDP 0.00* -0.00 -0.00
(0.00) (0.00) (0.00)
Polity -0.01*** -0.00** -0.00*
(0.00) (0.00) (0.00)
ln GDP pc 0.75*** 0.70*** 0.69***
(0.04) (0.04) (0.07)
Population density -0.00*** 0.00*** 0.00
(0.00) (0.00) (0.00)
Growth -0.19 -0.17* -0.37**
(0.12) (0.10) (0.15)
Time trend 0.03*** 0.02** 0.01
(0.01) (0.01) (0.02)
Constant -14.60*** -1.90*** -10.12***
(0.59) (0.51) (1.43)
Country xed eects yes yes yes
Observations 2363 2434 1105
Number of countries 98 101 86
Standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1
Instrumented: IGO membership
Instruments: Aid, polity, ln trade openness, population density
FDI, time trend, ln GDP pc, economic growth
UN Voting in accordance with US, (ln GDP PPP - ln GDP current)
results for CO
2
emissions. As shown in table 6.3, the conventional estimates
for CO
2
emissions indicate that environmental IGOs are associated with an
increase in environmental quality. Altogether, this leads us to conclude that
environmental IGOs do not exert any systematic inuence in the case of SO
2
emissions only. However, if we consider CO
2
emissions or water quality, we
see that environmental IGO membership signicantly reduces environmental
degradation in developing countries even after controlling for the potential
endogeneity of this form of IGO membership.
148
CHAPTER 6. IGO MEMBERSHIP AND ENVIRONMENTAL
QUALITY: UNDERSTANDING THE MECHANISMS
Table 6.11: Regression instrumenting environmental IGO membership
sample: all developing countries
(1) (2) (3)
SO
2
CO
2
BOD
Instrumented environmental IGOs 0.02 0.01 -0.13**
(0.03) (0.03) (0.06)
Umbrella IGOs -0.06*** -0.05*** -0.11***
(0.01) (0.01) (0.02)
Military IGOs -0.17*** -0.07*** -0.00
(0.03) (0.03) (0.04)
Economic IGOs (development) 0.00 0.03*** -0.03*
(0.01) (0.01) (0.02)
Economic IGOs (general) -0.01** -0.02*** 0.03**
(0.01) (0.01) (0.01)
Economic IGOs (standardization) -0.03*** -0.03*** -0.01
(0.01) (0.01) (0.01)
Economic IGOs (industry) 0.00 -0.01 -0.03*
(0.01) (0.01) (0.02)
Social IGOs -0.01*** -0.00 -0.01
(0.00) (0.00) (0.01)
Nuclear IGOs -0.22*** 0.03 0.07
(0.03) (0.03) (0.04)
Agricultural IGOs 0.04*** 0.05*** -0.00
(0.01) (0.01) (0.03)
Aid -0.00 -0.00** -0.00
(0.00) (0.00) (0.00)
ln trade openness 0.11*** 0.19*** 0.10*
(0.02) (0.02) (0.05)
FDI % GDP 0.00* -0.00 -0.00
(0.00) (0.00) (0.00)
Polity -0.01*** -0.00 -0.00
(0.00) (0.00) (0.00)
ln GDP pc 0.65*** 0.62*** 0.71***
(0.05) (0.04) (0.08)
Population density -0.00*** 0.00*** 0.00**
(0.00) (0.00) (0.00)
Growth -0.16 -0.18* -0.24
(0.11) (0.10) (0.17)
Time trend -0.00 -0.00 0.03***
(0.00) (0.00) (0.01)
Constant -14.35*** -1.89*** -9.34***
(0.49) (0.44) (1.01)
Country xed eects yes yes yes
Observations 2363 2434 1105
Number of countries 98 101 86
Standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1
Instrumented: Environmental IGO membership
Instruments: Aid, polity, ln trade openness, FDI, ln GDP pc, time trend,
population density, ln economic growth, all other types of IGO membership
UN Voting in accordance with US, (ln GDP PPP - ln GDP current)
149
CHAPTER 6. IGO MEMBERSHIP AND ENVIRONMENTAL
QUALITY: UNDERSTANDING THE MECHANISMS
6.4 Case Studies
As a nal eort to understand how IGO membership might aect the en-
vironmental performance of developing countries, I decided to look at four
countries more closely. Following Lieberman (2005, 436) it is possible to
increase the condence in the central ndings of a study, e.g. by showing
that the hypothesized mechanisms were actually manifest in reality. More
precisely, these illustrative case studies should identify how specic IGOs
work in developing countries and how their actions can inuence the envi-
ronmental performance of their member countries. Although it would be
desirable to show that certain actions of a particular IGO have led, for ex-
ample, to a reduction in air pollution in the country under analysis, doing
so is practically impossible. Therefore, these case studies focus on analyz-
ing the mandate and objectives of the specic organizations to which the
four countries belong, in order to evaluate how these objectives of the IGO
in question can support a country in its environmental protection eorts.
Hence these case studies aim to show how the mandate and objectives of
the IGOs under analysis could in principle be linked to a better environmen-
tal performance of developing countries. Furthermore, the analysis examines
whether any systematic dierences exist with regard to the functioning of
the IGOs in democratic compared to autocratic developing countries. Al-
together, it is therefore important to note that these case studies are not
intended to test any hypotheses or assumptions; they are merely illustra-
tions of how IGO membership could in reality aect environmental quality
in developing countries. This should help to better evaluate which of the
mechanisms outlined in the theoretical section are indeed at work in reality
and which mechanisms do not seem to be of great importance.
In particular, I analyze two democracies and two autocracies, of which
one is a member of many IGOs and one is a member of only a few IGOs. The
analysis covers all major variation in the two main independent variables,
IGO membership and political system. Furthermore, I restrict the selection
to countries that have stable political regimes to exclude the possibility that
changes in the regime type aect the relationship between IGO membership
and environmental quality.
To specically decide which countries to analyze more closely, I follow
Lieberman (2005)s nested analysis approach. Since the general idea that
IGO membership is associated with better environmental performance is
conrmed by the analysis, the selection method proposed by Lieberman
(2005) suggests analyzing those cases that are not outliers to the statistical
analysis but fall into the pattern predicted by the theoretical arguments. By
selecting cases that are in line with the theoretical predictions, it is possible
to get in-depth insights into the specic mechanisms that are driving the
150
CHAPTER 6. IGO MEMBERSHIP AND ENVIRONMENTAL
QUALITY: UNDERSTANDING THE MECHANISMS
results in reality
12
.
Being constrained to select four countries that fall into the pattern pre-
dicted by the theoretical arguments, of which two are democratic and two
are autocratic with one of each category being a member of many IGOs and
one being a member of only a few IGOs while ensuring some regional vari-
ation, limits the analysis to the following four countries: Cameroon, India,
Laos and Lithuania. Cameroon and Laos are both autocratic regimes with
Cameroon being a member of many IGOs whereas Laos joined relatively
few IGOs until the year 2000. India and Lithuania were both democratic
countries for the time span in which they are included in the analysis. India
in contrast to Lithuania is a member of a great many international organi-
zations.
I begin the analysis of the four cases by briey describing the state of
the environment in each of the four countries. In particular, I present the
major environmental problems and describe which environmental authori-
ties, such as ministries of environment, exist in the country that could tackle
these problems. I then turn to the specic IGOs each country is a mem-
ber of. By analyzing the activities of each of these organizations I intend
to illustrate how these IGOs might aect the environmental performance
of their member countries. As IGO membership lies somewhere between 36
and 89 organizations per country I only describe the activities of those IGOs
that could help us understand how the positive inuence on environmental
quality might come about. It should thus be possible to illustrate which
inuence channels are most common and how many organizations do indeed
have a potential inuence on environmental quality in developing countries.
6.4.1 State of the Environment
Starting with the state of the environment in Cameroon, we see that Cameroons
economy is mainly agricultural with a focus on cocoa, coee, tobacco, cot-
ton, and banana plantation with the major export good being petroleum
(World Bank 2009c). Deforestation is Cameroons biggest environmental
problem since most activities related to Cameroons tropical forests are still
little regulated (WWF 2009). In addition, bushmeat hunting has strongly
12
Lieberman (2005) suggests only looking at outliers to the statistical analysis if the
theoretical arguments are not conrmed by the analysis. In this case, the goal of the
case study analysis should be to build a better theoretical model of the hypothesized
relationships. Therefore, it is necessary to obtain information about outlier cases to get
a better understanding of the relationship being analyzed in general. However, if the
statistical analysis conrms the theoretical argumentation, the goal of the case studies
should be to further test the theoretical arguments and to gain more insights into the
proposed theoretical relationship. Therefore, Lieberman (2005) argues that in this case
one should only look at cases that can provide the researcher with knowledge on the
proposed theoretical relationship, which are cases that are not outliers to the statistical
analysis.
151
CHAPTER 6. IGO MEMBERSHIP AND ENVIRONMENTAL
QUALITY: UNDERSTANDING THE MECHANISMS
increased over the last years, which threatens forest wildlife (WWF 2009).
When looking at the indicators of environmental quality used in the
statistical analysis we see, as shown in gures 6.7(a) and 6.7(b), an increase
in both SO
2
and CO
2
emissions until the mid 1980s with both pollutants
seeming to have decreased thereafter. Water pollution has declined more or
less steadily for the time period 1980 to 2000 see gure 6.7(c).
1000
1
0
0
0
1000 1200
1
2
0
0
1200 1400
1
4
0
0
1400 1600
1
6
0
0
1600 1800
1
8
0
0
1800 2000
2
0
0
0
2000 GDP pc
G
D
P

p
c
GDP pc .001
.
0
0
1
.001 .002
.
0
0
2
.002 .003
.
0
0
3
.003 .004
.
0
0
4
.004 SO2
S
O
2
SO2 1970
1970
1970 1980
1980
1980 1990
1990
1990 2000
2000
2000 year of observation
year of observation
year of observation SO2 per capita
SO2 per capita
SO2 per capita GDP per capita
GDP per capita
GDP per capita
(a) SO
2
emissions
1000
1
0
0
0
1000 1200
1
2
0
0
1200 1400
1
4
0
0
1400 1600
1
6
0
0
1600 1800
1
8
0
0
1800 2000
2
0
0
0
2000 GDP pc
G
D
P

p
c
GDP pc 0
0
0200
2
0
0
200 400
4
0
0
400 600
6
0
0
600 800
8
0
0
800 CO2
C
O
2
CO2 1970
1970
1970 1980
1980
1980 1990
1990
1990 2000
2000
2000 year of observation
year of observation
year of observation CO2 per capita
CO2 per capita
CO2 per capita GDP per capita
GDP per capita
GDP per capita
(b) CO
2
emissions
1000
1
0
0
0
1000 1200
1
2
0
0
1200 1400
1
4
0
0
1400 1600
1
6
0
0
1600 1800
1
8
0
0
1800 2000
2
0
0
0
2000 GDP pc
G
D
P

p
c
GDP pc .001
.
0
0
1
.001 .0012
.
0
0
1
2
.0012 .0014
.
0
0
1
4
.0014 .0016
.
0
0
1
6
.0016 .0018
.
0
0
1
8
.0018 .002
.
0
0
2
.002 BOD
B
O
D
BOD 1980
1980
1980 1985
1985
1985 1990
1990
1990 1995
1995
1995 2000
2000
2000 year of observation
year of observation
year of observation BOD per capita
BOD per capita
BOD per capita GDP per capita
GDP per capita
GDP per capita
(c) BOD
Figure 6.7: State of the environment: Cameroon
The highest political authority dealing with environmental regulation is
the Ministry of Environment and Nature Protection of Cameroon (UNEP
2009b). However, it is dicult to assess how active the Ministry of Environ-
ment is with regard to environmental protection. NGOs such as the World
Wide Fund For Nature (WWF) state that regulation in the eld of envi-
ronmental protection is very thin (WWF 2009). Since the government of
Cameroon and more specically the Ministry of Environment do not provide
much information on their environmental activities that is accessible on the
Internet, it is very dicult to evaluate the correctness of this statement.
Although mainly a rural economy, Indias industrial sector has been
characterized by immense growth over the last decades. This growth has led
to a strong increase in industrial pollution emissions, especially air pollution,
as well as an increase in hazardous waste (UNEP 2001a). Furthermore, since
most of Indias power is produced using coal, a major problem with regard
152
CHAPTER 6. IGO MEMBERSHIP AND ENVIRONMENTAL
QUALITY: UNDERSTANDING THE MECHANISMS
to air quality has emerged (UNEP 2001a).
This trend can also be seen in gures 6.8(a) and 6.8(b), which show that
both SO
2
and CO
2
emissions have increased along with GDP per capita.
Only water pollution seems to have decreased over the same period of time
see gure 6.8(c).
In addition to these problems, the State of the Environment Report in
India (UNEP 2001a) further refers to biodiversity loss and land degradation
as major threats to the environment. All of these problems are especially
pressing due to Indias immense population growth. Although India enacts
important environmental regulations, eorts to protect the environment are
often countervailed by the increase in population (UNEP 2001a).
1000
1
0
0
0
1000 1500
1
5
0
0
1500 2000
2
0
0
0
2000 2500
2
5
0
0
2500 GDP pc
G
D
P

p
c
GDP pc .0015
.
0
0
1
5
.0015 .002
.
0
0
2
.002 .0025
.
0
0
2
5
.0025 .003
.
0
0
3
.003 .0035
.
0
0
3
5
.0035 SO2
S
O
2
SO2 1970
1970
1970 1980
1980
1980 1990
1990
1990 2000
2000
2000 year of observation
year of observation
year of observation SO2 per capita
SO2 per capita
SO2 per capita GDP per capita
GDP per capita
GDP per capita
(a) SO
2
emissions
1000
1
0
0
0
1000 1500
1
5
0
0
1500 2000
2
0
0
0
2000 2500
2
5
0
0
2500 GDP pc
G
D
P

p
c
GDP pc 400
4
0
0
400 600
6
0
0
600 800
8
0
0
800 1000
1
0
0
0
1000 1200
1
2
0
0
1200 CO2
C
O
2
CO2 1970
1970
1970 1980
1980
1980 1990
1990
1990 2000
2000
2000 year of observation
year of observation
year of observation CO2 per capita
CO2 per capita
CO2 per capita GDP per capita
GDP per capita
GDP per capita
(b) CO
2
emissions
1000
1
0
0
0
1000 1500
1
5
0
0
1500 2000
2
0
0
0
2000 2500
2
5
0
0
2500 GDP pc
G
D
P

p
c
GDP pc .0016
.
0
0
1
6
.0016 .0017
.
0
0
1
7
.0017 .0018
.
0
0
1
8
.0018 .0019
.
0
0
1
9
.0019 .002
.
0
0
2
.002 .0021
.
0
0
2
1
.0021 BOD
B
O
D
BOD 1980
1980
1980 1985
1985
1985 1990
1990
1990 1995
1995
1995 2000
2000
2000 year of observation
year of observation
year of observation BOD per capita
BOD per capita
BOD per capita GDP per capita
GDP per capita
GDP per capita
(c) BOD
Figure 6.8: State of the environment: India
However, as early as 1974 the government in India set up a control
board to monitor environmental degradation all over the country and in
1984 the Ministry of Environment and Forests was established to enhance
environmental performance in India (Ministry of Environment and Forests
2009). The State of the Environment Report in India (2002) (UNEP 2001a)
refers to the 1972 United Nations Conference on the Human Environment
in Stockholm as the trigger event after which environmental protection was
integrated into nationwide policy-making. Since then the Indian govern-
ment has enacted many environmental regulations and experts conclude
153
CHAPTER 6. IGO MEMBERSHIP AND ENVIRONMENTAL
QUALITY: UNDERSTANDING THE MECHANISMS
that the state of the environment in India would be much worse without
its eorts with regard to environmental regulation and capacity building
(UNEP 2001a).
As in the case of Cameroon, one of Laos major environmental prob-
lems is deforestation, mainly because the government of Laos has not yet
enacted appropriate regulation concerning logging and limiting wood ex-
traction (UNEP 2001b). Deforestation is further problematic since Laos
rich biological diversity suers greatly from the loss of natural habitat and
because ground water levels are reduced (IMF 2009). Most people are de-
pendent on agriculture, primarily in the Mekong River basin, Laos main
river. Its water quality is generally assessed to be of good quality (UNEP
2001b). Laos industry consists of manufacturing (mainly the garment in-
dustry), mining, electricity and construction with manufacturing being by
far the largest part. However, most of the goods consumed in Laos are
imported, mainly from Thailand (UNEP 2001b).
When looking at air pollution in Laos
13
we see a very peculiar pattern
(gures 6.9(a) and 6.9(b)): both SO
2
and CO
2
emissions declined from 1970
to the beginning of the 1980s. Since then air pollution has steadily increased
despite the fact that Laos GDP per capita declined during the same period.
1400
1
4
0
0
1400 1500
1
5
0
0
1500 1600
1
6
0
0
1600 1700
1
7
0
0
1700 1800
1
8
0
0
1800 1900
1
9
0
0
1900 GDP pc
G
D
P

p
c
GDP pc .0004
.
0
0
0
4
.0004 .0006
.
0
0
0
6
.0006 .0008
.
0
0
0
8
.0008 .001
.
0
0
1
.001 .0012
.
0
0
1
2
.0012 SO2
S
O
2
SO2 1970
1970
1970 1980
1980
1980 1990
1990
1990 2000
2000
2000 year of observation
year of observation
year of observation SO2 per capita
SO2 per capita
SO2 per capita GDP per capita
GDP per capita
GDP per capita
(a) SO
2
emissions
1400
1
4
0
0
1400 1500
1
5
0
0
1500 1600
1
6
0
0
1600 1700
1
7
0
0
1700 1800
1
8
0
0
1800 1900
1
9
0
0
1900 GDP pc
G
D
P

p
c
GDP pc 50
5
0
50 100
1
0
0
100 150
1
5
0
150 200
2
0
0
200 CO2
C
O
2
CO2 1970
1970
1970 1980
1980
1980 1990
1990
1990 2000
2000
2000 year of observation
year of observation
year of observation CO2 per capita
CO2 per capita
CO2 per capita GDP per capita
GDP per capita
GDP per capita
(b) CO
2
emissions
Figure 6.9: State of the environment: Laos
The responsible agency for environmental management is the Science,
Technology and Environmental Agency in the Prime Ministers Oce (UNEP
2001b). Unfortunately, environmental protection does not seem to be of high
priority to Laos, which can be deduced from the fact that most environmen-
tal institutions and activities are under-resourced (UNEP 2001b). The State
of the Environment Report in Laos (UNEP 2001b, 22) even goes as far as
to state that Laos rich wildlife only persists because population density is
so low, not because any major conservation eorts are undertaken. This is
especially frustrating as Laos joined the UN Convention on Biological Diver-
sity in 1996 see table 6.13. Laos has thus committed itself to develop both
13
Data on water pollution are unfortunately not available for Laos.
154
CHAPTER 6. IGO MEMBERSHIP AND ENVIRONMENTAL
QUALITY: UNDERSTANDING THE MECHANISMS
a national strategy for conservation of the countrys biological diversity and
regulatory provisions for protecting threatened species and populations.
Finally, the case of environmental protection in Lithuania is very dierent
from the other three countries. Since the country regained its independence
in 1991, environmental quality has improved dramatically compared with
the Soviet era. This is especially apparent in the case of air quality as both
SO
2
and CO
2
emissions have decreased over the last decade (Ministry of
Environment of the Republic of Lithuania 2001). Figures 6.10(a) and 6.10(b)
further illustrate this point. Figure 6.10(c) shows that water pollution has
also decreased signicantly over the last decade.
6200
6
2
0
0
6200 6400
6
4
0
0
6400 6600
6
6
0
0
6600 6800
6
8
0
0
6800 7000
7
0
0
0
7000 7200
7
2
0
0
7200 GDP pc
G
D
P

p
c
GDP pc .005
.
0
0
5
.005 .01
.
0
1
.01 .015
.
0
1
5
.015 .02
.
0
2
.02 .025
.
0
2
5
.025 .03
.
0
3
.03 SO2
S
O
2
SO2 1990
1990
1990 1992
1992
1992 1994
1994
1994 1996
1996
1996 1998
1998
1998 2000
2000
2000 year of observation
year of observation
year of observation SO2 per capita
SO2 per capita
SO2 per capita GDP per capita
GDP per capita
GDP per capita
(a) SO
2
emissions
6200
6
2
0
0
6200 6400
6
4
0
0
6400 6600
6
6
0
0
6600 6800
6
8
0
0
6800 7000
7
0
0
0
7000 7200
7
2
0
0
7200 GDP pc
G
D
P

p
c
GDP pc 3500
3
5
0
0
3500 4000
4
0
0
0
4000 4500
4
5
0
0
4500 5000
5
0
0
0
5000 5500
5
5
0
0
5500 6000
6
0
0
0
6000 CO2
C
O
2
CO2 1990
1990
1990 1992
1992
1992 1994
1994
1994 1996
1996
1996 1998
1998
1998 2000
2000
2000 year of observation
year of observation
year of observation CO2 per capita
CO2 per capita
CO2 per capita GDP per capita
GDP per capita
GDP per capita
(b) CO
2
emissions
6200
6
2
0
0
6200 6400
6
4
0
0
6400 6600
6
6
0
0
6600 6800
6
8
0
0
6800 7000
7
0
0
0
7000 7200
7
2
0
0
7200 GDP pc
G
D
P

p
c
GDP pc .01
.
0
1
.01 .011
.
0
1
1
.011 .012
.
0
1
2
.012 .013
.
0
1
3
.013 .014
.
0
1
4
.014 BOD
B
O
D
BOD 1990
1990
1990 1992
1992
1992 1994
1994
1994 1996
1996
1996 1998
1998
1998 2000
2000
2000 year of observation
year of observation
year of observation BOD per capita
BOD per capita
BOD per capita GDP per capita
GDP per capita
GDP per capita
(c) BOD
Figure 6.10: State of the environment: Lithuania
However, some environmental problems persist in Lithuania, especially
with regard to nuclear energy production, leading to environmental pollution
with radionuclides (Ministry of Environment of the Republic of Lithuania
2001). Still, compared to the beginning of the 1990s, environmental and
human hazards arising from nuclear substances have also decreased.
The Ministry of Environment of the Republic of Lithuania is the re-
sponsible government agency for the protection of the environment. As
will become apparent in the discussion below, environmental regulation in
Lithuania is well developed, mainly because of Lithuanias accession to the
European Union in 2004 (BEF 2009).
155
CHAPTER 6. IGO MEMBERSHIP AND ENVIRONMENTAL
QUALITY: UNDERSTANDING THE MECHANISMS
To summarize briey the discussion on the state of the environment, the
four countries are characterized by very dierent environmental problems.
Whereas for both Cameroon and Laos, deforestation and biodiversity loss
are the major environmental threats, India is confronted with a strong in-
crease in industrial pollution due to its economic growth. Lithuanias state
of the environment, in contrast, has increased over the last two decades since
the country gained independence from the Soviet Union. Furthermore, the
discussion suggests that environmental regulation seems to be much more
developed in India and Lithuania compared to Cameroon and Laos. In addi-
tion, it is noticeable that in compiling the environmental country proles, it
was much more dicult to obtain valid information for the two autocracies,
Cameroon and Laos, than for the two democracies, India and Lithuania.
This was the case because a) government agencies in the two democratic
countries provide much more information on the Internet and b) because
NGOs such as the WWF and IGOs such as the United Nations Environ-
ment Programme (UNEP) seem to be more active in reporting about India
and Lithuania than about the two autocracies Cameroon and Laos
14
.
6.4.2 Common IGO Membership
After having described the state of the environment in each of the four
countries I now turn to their IGO membership. I begin by discussing the
inuence channels on environmental quality of those organizations to which
all four countries belong. These are mainly the major UN organizations such
as UNESCO and FAO as well as the World Bank and the IMF. Although
these organizations do not provide for any variance in the membership with
regard to the four countries, it is still worthwhile analyzing their objectives
and actions. As the aim of these case studies is to illustrate which specic
functions of an IGO can in reality lead to a better provision of environmen-
tal quality in developing countries, it is advisable to analyze all IGOs and
not only those IGOs that a particular country is a member of. Therefore,
by evaluating the actions of all IGOs, a more complete picture of all the po-
tential mechanisms emerges through which IGOs could aect their member
countries environmental performance.
The next paragraphs thus analyze the functions of the IGOs to which
all four countries belong. I then turn to each country separately and discuss
the various channels through which their membership in specic IGOs might
inuence their environmental performance.
In general, due to the case selection process, IGO membership varies
greatly between the four countries. Both Cameroon and India are members
of many IGOs with Cameroons IGO membership rising from 37 in 1970 to
14
Of course, it is important to bear in mind that Indias size and Lithuanias status as an
EU accession (now member) country could well be the reason for this better information
provision.
156
CHAPTER 6. IGO MEMBERSHIP AND ENVIRONMENTAL
QUALITY: UNDERSTANDING THE MECHANISMS
87 in 2000 and Indias IGO membership increasing from 50 to 87 over the
same period. In contrast, both Laos and Lithuania are members of only a
small number of IGOs as Laos IGO membership grew from only 26 to 36
from 1970 to 2000 and Lithuanias IGO membership increased from 13 in
1991 to 46 in the year 2000. Figures 6.11(a) to 6.11(d) show the variation in
membership in the specic IGO types for each country over time. In each
of the cases, membership in economic IGOs is the most common followed
by membership in social organizations
15
.
15
For most of the IGOs of which Cameroon is a member, the country joined either after
it gained independence in 1960 or immediately after the specic organization was estab-
lished. The only exceptions are the Maritime Organisation of West and Central Africa,
the International Oil Pollution Compensation Fund and the International Civil Defence
Organization, all of which Cameroon joined only in the late 1980s. Since its independence
Cameroon has had two presidents, both of whom belong to the Cameroonian Democratic
Union. In 1982 Paul Biya took over the presidency from Ahmadou Ahidjo and so all of
these later accessions to IGOs fall into the early years of the second presidents period in
oce (The World Factbook 2009). The only organization that India did not join shortly
after the organization was established is the International Whaling Commission. The In-
dian government did not join this organization until 1981, although it was established as
early as 1946. With the exception of ASEAN, the International Finance Corporation and
the Multilateral Fund for the Implementation of the Montreal Protocol, which Laos only
joined in the late 1990s, the country joined the various IGOs shortly after their establish-
ment. Lithuania joined most IGOs shortly after the country gained independence in 1991.
The major exception is the European Union, which the country joined in 2004.
157
CHAPTER 6. IGO MEMBERSHIP AND ENVIRONMENTAL
QUALITY: UNDERSTANDING THE MECHANISMS
0
0
02
0
2 0
2
0
4
0
4 0
4
0
6
0
6 0
6
0
1
9
7
0
1
9
7
0
1
9
7
0
1
9
8
0
1
9
8
0
1
9
8
0
1
9
9
0
1
9
9
0
1
9
9
0
2
0
0
0
2
0
0
0
2
0
0
0
y
e
a
r

o
f

o
b
s
e
r
v
a
t
i
o
n
y
e
a
r

o
f

o
b
s
e
r
v
a
t
i
o
n
y
e
a
r

o
f

o
b
s
e
r
v
a
t
i
o
n

E
n
v
.

I
G
O

E
n
v
.

I
G
O

E
n
v
.

I
G
O

U
m
b
.

I
G
O

U
m
b
.

I
G
O

U
m
b
.

I
G
O

M
i
l
.

I
G
O

M
i
l
.

I
G
O

M
i
l
.

I
G
O

S
o
c
.

I
G
O

S
o
c
.

I
G
O

S
o
c
.

I
G
O

E
c
o
n
.

I
G
O

E
c
o
n
.

I
G
O

E
c
o
n
.

I
G
O

A
g
r
i
.

I
G
O

A
g
r
i
.

I
G
O

A
g
r
i
.

I
G
O

N
u
c
.

I
G
O

N
u
c
.

I
G
O

N
u
c
.

I
G
O
(
a
)
C
a
m
e
r
o
o
n
0
0
02
0
2 0
2
0
4
0
4 0
4
0
6
0
6 0
6
0
1
9
7
0
1
9
7
0
1
9
7
0
1
9
8
0
1
9
8
0
1
9
8
0
1
9
9
0
1
9
9
0
1
9
9
0
2
0
0
0
2
0
0
0
2
0
0
0
y
e
a
r

o
f

o
b
s
e
r
v
a
t
i
o
n
y
e
a
r

o
f

o
b
s
e
r
v
a
t
i
o
n
y
e
a
r

o
f

o
b
s
e
r
v
a
t
i
o
n

E
n
v
.

I
G
O

E
n
v
.

I
G
O

E
n
v
.

I
G
O

U
m
b
.

I
G
O

U
m
b
.

I
G
O

U
m
b
.

I
G
O

M
i
l
.

I
G
O

M
i
l
.

I
G
O

M
i
l
.

I
G
O

S
o
c
.

I
G
O

S
o
c
.

I
G
O

S
o
c
.

I
G
O

E
c
o
n
.

I
G
O

E
c
o
n
.

I
G
O

E
c
o
n
.

I
G
O

A
g
r
i
.

I
G
O

A
g
r
i
.

I
G
O

A
g
r
i
.

I
G
O

N
u
c
.

I
G
O

N
u
c
.

I
G
O

N
u
c
.

I
G
O
(
b
)
I
n
d
i
a
0
0
05
5
51
0
1 0
1
0
1
5
1 5
1
5
2
0
2 0
2
0
2
5
2 5
2
5
1
9
7
0
1
9
7
0
1
9
7
0
1
9
8
0
1
9
8
0
1
9
8
0
1
9
9
0
1
9
9
0
1
9
9
0
2
0
0
0
2
0
0
0
2
0
0
0
y
e
a
r

o
f

o
b
s
e
r
v
a
t
i
o
n
y
e
a
r

o
f

o
b
s
e
r
v
a
t
i
o
n
y
e
a
r

o
f

o
b
s
e
r
v
a
t
i
o
n

E
n
v
.

I
G
O

E
n
v
.

I
G
O

E
n
v
.

I
G
O

U
m
b
.

I
G
O

U
m
b
.

I
G
O

U
m
b
.

I
G
O

M
i
l
.

I
G
O

M
i
l
.

I
G
O

M
i
l
.

I
G
O

S
o
c
.

I
G
O

S
o
c
.

I
G
O

S
o
c
.

I
G
O

E
c
o
n
.

I
G
O

E
c
o
n
.

I
G
O

E
c
o
n
.

I
G
O

A
g
r
i
.

I
G
O

A
g
r
i
.

I
G
O

A
g
r
i
.

I
G
O

N
u
c
.

I
G
O

N
u
c
.

I
G
O

N
u
c
.

I
G
O
(
c
)
L
a
o
s
0
0
05
5
51
0
1 0
1
0
1
5
1 5
1
5
2
0
2 0
2
0
2
5
2 5
2
5
1
9
9
0
1
9
9
0
1
9
9
0
1
9
9
2
1
9
9
2
1
9
9
2
1
9
9
4
1
9
9
4
1
9
9
4
1
9
9
6
1
9
9
6
1
9
9
6
1
9
9
8
1
9
9
8
1
9
9
8
2
0
0
0
2
0
0
0
2
0
0
0
y
e
a
r

o
f

o
b
s
e
r
v
a
t
i
o
n
y
e
a
r

o
f

o
b
s
e
r
v
a
t
i
o
n
y
e
a
r

o
f

o
b
s
e
r
v
a
t
i
o
n

E
n
v
.

I
G
O

E
n
v
.

I
G
O

E
n
v
.

I
G
O

U
m
b
.

I
G
O

U
m
b
.

I
G
O

U
m
b
.

I
G
O

M
i
l
.

I
G
O

M
i
l
.

I
G
O

M
i
l
.

I
G
O

S
o
c
.

I
G
O

S
o
c
.

I
G
O

S
o
c
.

I
G
O

E
c
o
n
.

I
G
O

E
c
o
n
.

I
G
O

E
c
o
n
.

I
G
O

A
g
r
i
.

I
G
O

A
g
r
i
.

I
G
O

A
g
r
i
.

I
G
O

N
u
c
.

I
G
O

N
u
c
.

I
G
O

N
u
c
.

I
G
O
(
d
)
L
i
t
h
u
a
n
i
a
F
i
g
u
r
e
6
.
1
1
:
I
G
O
m
e
m
b
e
r
s
h
i
p
b
y
c
o
u
n
t
r
y
a
n
d
o
v
e
r
t
i
m
e
158
CHAPTER 6. IGO MEMBERSHIP AND ENVIRONMENTAL
QUALITY: UNDERSTANDING THE MECHANISMS
Turning to the specic organizations, I begin the analysis with the World
Bank. Since the World Bank has become one of the most important actors
in nancing environmental development projects (Gutner 2005; Nielson &
Tierney 2003), its actions and objectives are discussed in greater detail than
the functions of the other organizations. This more detailed analysis of the
World Bank therefore allows us to better assess when and how the topic of
sustainable development became important for international organizations.
In general, the World Bank encompasses two institutions: the Interna-
tional Bank for Reconstruction and Development (IBRD) and the Interna-
tional Development Association (IDA) with the IBRD providing loans to
middle-income and creditworthy poor countries and the IDA focusing on
the poorest countries only (World Bank 2009a). As early as 1970 the World
Bank established an Oce of Environmental and Health Aairs in order
to mitigate some of the environmental problems accompanying large-scale
infrastructure projects (Gutner 2005). However, most eorts of the Oce
of Environmental Aairs to reconcile the large-scale projects with environ-
mental concerns were not successful (Wade 1997). In 1987, due to immense
pressure from the US and other major shareholders of the Bank, the Oce
became a separate environmental department and was given more inu-
ence on World Bank operations (Nielson & Tierney 2003). Two years later,
in 1989, the Bank introduced the environmental assessment procedure to
alleviate potential adverse eects World Bank projects may have on the en-
vironment. Furthermore, countries applying for loans with the International
Development Association are obliged to provide a National Environmental
Action Plan, which outlines a countrys major environmental problems and
describes solutions to mitigate these problems (Gutner 2005). Moreover,
in the early 1990s the World Bank became the implementing agency for
the Montreal Protocols Multilateral Fund and for the Global Environment
Facility (GEF), making the Bank one of the most important nancing in-
stitutions with respect to environmental issues (Gutner 2005).
When looking at the specic projects the World Bank nances in our
four countries, it becomes apparent that projects focusing on environmental
issues have become more important over the years. In Cameroon, for ex-
ample, the World Bank has carried out 109 projects to date (World Bank
2009e). In the 1970s, most of these projects focused either on building infras-
tructure such as roads or ports or on supporting certain types of industries
e.g. rice, oil palm, rubber or cocoa industry in order to make them
more competitive. At the beginning of the 1980s the Bank launched various
projects in Cameroon to improve water supply and give people access to safe
water and thus to improve public health and water quality. In 1995, the rst
environmental project was launched in Cameroon to conserve biodiversity.
However, it was not until 2000 that one of the major infrastructure projects,
the Petroleum Development and Pipeline Project, was accompanied by an
environmental impact assessment. Since then, many more environmental
159
CHAPTER 6. IGO MEMBERSHIP AND ENVIRONMENTAL
QUALITY: UNDERSTANDING THE MECHANISMS
capacity building projects have been launched, e.g. to promote sustainable
agriculture or to enhance environmental education
16
. However, this environ-
mental about-face has only happened within the last 5 to 10 years (World
Bank 2009e).
In addition to the World Bank, Cameroon receives funding and resources
from the African Development Bank, which in recent years has also shifted
its focus to promoting sustainable development. An example is a project to
ensure the sustainable use of Lake Chad (African Development Bank 2009).
In India the World Bank has carried out almost 600 projects to date,
with the picture in India being very similar to Cameroon: most projects
in the 1970s and 1980s were aimed at either building infrastructure or sup-
porting industries and the agricultural sector (World Bank 2009e). The only
exceptions were projects that were targeted at improving water and sanita-
tion facilities for people living in the urban areas of Calcutta or Bombay, for
example. In contrast to Cameroon, however, the rst pure environmental
projects started earlier. In 1990, for example, several projects were approved
that dealt with the degradation of Indias watersheds. Furthermore, in 1991
the World Bank launched a project to deal with industrial pollution con-
trol. From the mid 1990s onward, environmental projects were launched
more often, for example projects dealing with renewable energy, energy e-
ciency, eco-development, environmental management and capacity building
etc. (World Bank 2009e).
Only about 60 projects have been carried out by the World Bank in Laos
to date. Interestingly, most environmental projects focus on biodiversity and
nature conservation starting as early as 1994 (World Bank 2009e). Hence
World Bank lending does indeed appear to mirror the specic environmental
needs of the particular countries. In India, the World Bank concentrates
more on pollution control, whereas in Laos it focuses on one of the countrys
major environmental threats, which is biodiversity loss. Both India and
Laos receive resources from the Asian Development Bank in addition to
the funding from the World Bank. However, the number of environmental
projects nanced by the Asian Development Bank seems to be lower and
the number of infrastructure projects higher compared to the World Bank
nancing (Asian Development Bank 2009).
In Lithuania the World Bank has only begun to nance projects since
the end of the cold war. Moreover, most projects were launched under
the direction of the European Bank for Reconstruction and Development
16
Of course, when describing those areas of operations of an IGO that are related to
environmental protection it is important to bear in mind that other areas might contribute
to pollution at the same time. For example, although the World Bank devotes many
resources to promoting sustainable development, it still nances large-scale infrastructure
projects such as the construction of airports or roads, which necessarily contribute to
increased pollution levels. Similarly, increasing agriculture even if it is done ecologically
is inevitably connected with exploitation of natural resources such as water and land.
160
CHAPTER 6. IGO MEMBERSHIP AND ENVIRONMENTAL
QUALITY: UNDERSTANDING THE MECHANISMS
(EBRD), resulting in the fact that only 19 projects have been nanced by
the World Bank in Lithuania (World Bank 2009e). However, from the early
1990s onwards, several projects with an environmental focus were carried
out such as the Klaipeda Environment Project, which aimed to reduce the
discharge of partially treated and untreated waste water to the Baltic Sea
or a project targeted at improving energy eciency in residential areas.
Similar to the processes at the World Bank, projects nanced by the EBRD
must undergo environmental appraisal before they are approved in order
to meet the Banks goal of sustainable development (EBRD 2009a). More
specically, in addition to the 19 World Bank projects, the EBRD has, since
its establishment in 1990, nanced another 18 projects in Lithuania. Among
these projects were several environmental projects focusing mainly on waste
water treatment in urban areas (EBRD 2009b).
Altogether, the discussion of the World Banks actions shows that with
the exception of projects that aimed to improve sanitation facilities it took
until the 1990s before environmental management and capacity building
were promoted in the four countries. Since then, however, the World Bank
has strongly advocated sustainable development, among other things by
providing environmentally sound technologies and know-how to reconcile
economic growth with environmental protection.
The International Finance Corporation (IFC), which also belongs to the
World Bank Group, focuses on the private sector in developing countries by
nancing investment and providing advice on topics such as tax revenue, job
creation, corporate governance, nancial market stability etc. (IFC 2009b).
To ensure that the companies the IFC invests in fulll their environmental
and social responsibility, the IFC provides them with knowledge on best
practices regarding waste recycling, waste water treatment, energy use or
business travel (IFC 2009a). The IFC thus aims to reconcile their invest-
ment induced economic growth with the environmental responsibility of the
companies they invest in. Hence all major World Bank institutions today
seem to be committed to the goal of sustainable development by at least
aiming to minimize the adverse eects of their projects on environmental
quality in developing countries.
In the case of the International Monetary Fund (IMF), which together
with the World Bank constitutes the so-called Bretton Woods institutions,
the connection to environmental issues is less obvious than with the World
Bank. Although the IMF acknowledges that environmental degradation
might have adverse eects on macroeconomic indicators and that in princi-
ple all growth should be sustainable, the promotion of environmental pro-
tection is not very pronounced within the IMF (IMF 2009). One way in
which the IMF tries to promote environmental sustainability is to advise
those specialists who are responsible for implementing the macroeconomic
policies demanded by the IMF on potential risks to the environment and
possible solutions. Even so, the IMF mostly seems to rely on the expec-
161
CHAPTER 6. IGO MEMBERSHIP AND ENVIRONMENTAL
QUALITY: UNDERSTANDING THE MECHANISMS
tation that sound macroeconomic policies will, as a byproduct, also favor
the environment since these policies will make economic production more
ecient and increase the demand for environmental protection (IMF 2009).
Turning from the Bretton Woods institutions to the UN system, we see
that many UN institutions have also incorporated environmental issues into
their every-day business. Although the UNs main purpose is to maintain
international peace and security, environmental topics are becoming more
and more important (UN 2009b). For example, one of the eight Millennium
Development Goals adopted by the UN in 2000 is the goal of environmen-
tal sustainability, which encompasses actions to combat climate change and
biodiversity losses and to provide safe access to drinking water for people in
developing countries (UN 2009a). More important, however, are other UN
institutions such as the United Nations Environment Programme (UNEP)
which focus exclusively on combating environmental degradation in its many
forms (UNEP 2009a). UNIDO, the UN Industrial Development Organiza-
tion, which promotes industrial development mainly by providing technical
cooperation, also has a special thematic focus on environment and energy. In
this context, UNIDO provides technical advice on topics such as cleaner pro-
duction processes, renewable forms of energy and waste recycling (UNIDO
2009).
Furthermore, the UN Educational, Scientic and Cultural Organization
(UNESCO) and the World Meteorological Organization (WMO) also pro-
mote among other things environmental sustainability. UNESCO, for
example, has established several biosphere regions in which sustainable de-
velopment is practiced. Famous examples are the Serengeti National Park
in Tanzania or the central Amazon region in Brazil (UNESCO 2001). The
WMOs Atmospheric Research and Environment Programme is another ex-
ample of promoting environmental protection as this program provides re-
search and technical support to promote compliance with conventions deal-
ing with issues such as ozone depletion, climate change and long-range trans-
port of air pollution (WMO 2009).
The main goal of the Food and Agriculture Organization (FAO) is to
achieve food security by promoting sustainable agricultural and rural devel-
opment. To arrive at this goal the organization recognizes that it is vital
to manage natural resources responsibly and thus to advance environmen-
tal protection (FAO 2009b). Since 2007, the FAO has therefore shifted the
focus of its former sustainable development department to natural resource
management and environmental issues. However, even the old department
dealt with issues such as bioenergy
17
and sustainable land and water use,
thus promoting an ecologically friendly use of natural resources. In particu-
lar, the FAO provides technical assistance and research driven knowledge on
the best ways of achieving sustainable farming and the sustainable manage-
17
Bioenergy includes, for example, sustainable wood energy and biofuels such as ethanol.
162
CHAPTER 6. IGO MEMBERSHIP AND ENVIRONMENTAL
QUALITY: UNDERSTANDING THE MECHANISMS
ment of natural resources. Consequently, the organization directly transfers
technology and knowledge necessary to reduce environmental pollution to
developing countries.
Another major IGO that all four countries belong to is the International
Labour Organization (ILO). Although the ILO mainly focuses on issues con-
cerning labour standards, it also operates a program that deals with safety,
health and the environment, which is called SafeWork (ILO 2009). One
of the measures in the program is the provision of technical assistance to
promote the health and safety of workers worldwide. Among other things,
workers employed in hazardous industries should acquire knowledge on how
to handle dangerous chemical substances or how to avoid industrial acci-
dents. By promoting these measures, the SafeWork program also furthers
environmental protection.
Another rather surprising example of how environmental protection is
promoted by an organization that in principle does not have much to do
with environmental issues is the World Tourism Organization, of which
Cameroon, Laos and India are members. Since the early 1990s, the or-
ganization has trained its member countries on how to reconcile tourism
with the local culture, economy and the environment to achieve the goal of
sustainable tourism (World Tourism Organization 2009). In particular, the
organization provides information on best practices on topics such as waste,
water and energy.
Acknowledging the fact that 25% of all diseases worldwide arise due to
environmental hazards, the World Health Organization (WHO) carries out
research and gives policy advice on how to deal with various forms of envi-
ronmental hazards most eectively (WHO 2009). Among other things, the
WHO oers guidelines on how to dispose of hazardous waste, documents
the health risks associated with indoor air pollution, advocates measures
to decrease indoor air pollution and monitors the changes in pollution lev-
els in households around the world. These far-reaching measures make the
organization an important diuser of the latest research and knowledge on
the impacts of various forms of environmental degradation and on eective
countermeasures to tackle the dierent health problems associated with en-
vironmental pollution. Hence although it aims to promote better public
health, the WHO also promotes environmental quality in important ways.
Although liberalizing trade is often seen as a danger for the environment
(see race to the bottom arguments), the World Trade Organization (WTO)s
rules contain some environmental provisions
18
. For example, member coun-
tries are allowed to use trade-related measures to protect human, animal or
plant life and health, and natural resources as long as these measures are
not designed to introduce protectionism through the back door
19
(WTO
18
Only Cameroon and India joined the WTO during the period under review. Lithuania
joined the WTO in 2001 and Laos has observer status only.
19
One famous example is the shrimp-turtle dispute between the US on the one side and
163
CHAPTER 6. IGO MEMBERSHIP AND ENVIRONMENTAL
QUALITY: UNDERSTANDING THE MECHANISMS
2009a). Furthermore, other important trade-environment related issues such
as the relationship between WTO and Multilateral Environmental Agree-
ments are part of the Doha Round negotiations, which have not yet reached
a successful conclusion (UNEP 2009c).
Another very interesting example of an IGO that pursues some en-
vironmental goals is the International Atomic Energy Agency (IAEA)
Cameroon, India and Lithuania are all members of the IAEA. Besides pro-
viding information and technology on the safe use of nuclear energy and
treatment of nuclear waste, the organization supports developing countries
with technical cooperation (IAEA 2009b). For example, the IAEA has car-
ried out a project in Cameroon focusing on the use of nuclear techniques in
groundwater assessment (IAEA 2009a). In addition to this technical cooper-
ation, the inuence of the IAEA on environmental quality becomes apparent
in the case of energy production as well. Since nuclear energy production
does not produce SO
2
or CO
2
emissions, countries that rely more on nuclear
energy production with the help of IAEA will have less air pollution as a
byproduct
20
.
The only two environmental IGOs to which all four countries belong
are the Global Environment Facility and the Multilateral Fund for the Im-
plementation of the Montreal Protocol. The GEF was created in 1991 to
promote sustainability on the international level by providing funding to
developing countries for projects within each of the following six areas: bio-
diversity, climate change, international waters, land degradation, the ozone
layer, and persistent organic pollutants (GEF 2008c). The GEF is therefore
the largest provider of funding aimed at the improvement of environmental
quality. An example of a GEF project is the establishment of 20 National
Biodiversity Conservation Areas in Laos to protect the countrys biodiversity
(UNEP 2001b).
However, compared to development aid in general
21
the resources pro-
vided through the GEF are rather limited. To 2008, Cameroon received
USD 28 million of funding from the GEF compared to USD 535 million of
total development aid in the year 2000 alone. The GEF provided Laos with
13 million US dollars to 2008 (compared to USD 533 million of total devel-
India, Malaysia, Pakistan and Thailand on the other. In particular, the four developing
countries launched a complaint against the US because the US only allowed countries to
export shrimp to the US if their shrimp trawlers used turtle excluder devices in their nets
(WTO 2009b). The US nally lost the case because the Appellate Body decided that the
US practice discriminated between dierent member countries of the WTO. Although the
Appellate Body recognized that in principle the US was allowed to protect sea turtles
the US practice was discriminating since the US provided technical assistance to some
Caribbean states but not to other developing countries.
20
Of course, this does not reect the fact that nuclear energy production produces
nuclear waste, which itself constitutes a major environmental problem.
21
The funding provided by the GEF is included to 77% in the OECD aid data used in
the quantitative analysis of this thesis (OECD 2008).
164
CHAPTER 6. IGO MEMBERSHIP AND ENVIRONMENTAL
QUALITY: UNDERSTANDING THE MECHANISMS
opment aid in the year 2000), Lithuania with 27 million (the OECD does
not provide data on aid directed to Lithuania) and India with 235 million
(compared to USD 1,777.22 million of total development aid in the year
2000) to invest in environmental projects.
The Multilateral Fund for the Implementation of the Montreal Protocol
assists developing countries to comply with their obligations under the Pro-
tocol to phase out the use of ozone-depleting substances (ODS) by providing
them with nancial resources (Montreal Multilateral Fund 2009). In partic-
ular, the Funds experts estimate how much it would cost a country per year
to phase out ozone-depleting substances at a rate determined in the Mon-
treal Protocol. This allows the Fund to provide each country, after it has
achieved the demanded reductions in ODS, with funding to compensate for
the economic losses associated with ODS phase-out. The Fund thus ensures
that developing countries do indeed reduce emissions of substances such as
chlorouorocarbons and methyl bromide to decrease ozone depletion.
In addition to the GEF and the Montreal Fund, India, Cameroon and
Lithuania are members of the International Institute of Refrigeration. In
particular, the IIRs mission is to promote knowledge of refrigeration tech-
nology and all its applications in order to address todays major issues,
including food safety and protection of the environment (reduction of global
warming, prevention of ozone depletion), and the development of the least
developed countries (food, health). The IIR commits itself to improving
quality of life and promotes sustainable development (IIR 2009).
Furthermore, the three states are also members of the International Mar-
itime Organization (IMO), which aims among other things to reduce pol-
lution in maritime waters (IMO 2009). The International Convention for
the Prevention of Pollution from Ships, which is one of the major global
environmental treaties, falls within the auspices of the IMO (Roberts et al.
2004).
Finally, to get a more general overview of all the international environ-
mental commitments of our four countries, table 6.12 lists all environmental
IGOs of which any of the four countries is a member as well as the year
in which the particular country joined the organization. Table 6.13 comple-
ments this information by showing the major global environmental treaties
22
of which any of the four countries is a member. Interestingly, Cameroon cur-
rently has the highest number of environmental IGO memberships (12) fol-
lowed by India (8), Lithuania (6) and Laos (4). The picture slightly changes
if we look at membership in environmental treaties. Here India has the high-
est number of memberships (14) followed by Lithuania (10). Cameroon has
only joined 8 global environmental treaties to date and Laos only 5. One
reason for the high number of Cameroons environmental IGO memberships
22
I selected all environmental treaties which were included in the analysis of global
environmental treaty ratication by Roberts et al. (2004).
165
CHAPTER 6. IGO MEMBERSHIP AND ENVIRONMENTAL
QUALITY: UNDERSTANDING THE MECHANISMS
might be the fact that 4 out of the 12 IGOs have a clear regional focus
providing solutions to environmental problems that are directly relevant to
Cameroon such as the Joint Anti-Locust and Anti-Avarian Organization.
The discussion so far has shown that many of the major IGOs of which
all four countries are members include at least some environmental provi-
sions. Although the degree to which environmental topics play a role in a
specic organization varies considerably from organization to organization
for the World Bank or the WHO, environmental issues are clearly more
important than for the WTO or the IMF sustainable development has
become a crucial goal for most economic cooperation, agricultural and even
social organizations. To obtain more detailed information on how member-
ship varies among our four countries and how these countries are therefore
exposed to environmental exercises of inuence by IGOs, I now turn to the
specic IGO proles of each of the four countries.
Table 6.12: Membership in environmental IGOs
Environmental IGOS Cameroon India Laos Lithuania
Global Environment Facility (GEF) 1994 1994 1994 1994
Multilateral Fund for the Implementation
of the Montreal Protocol 1990 1991 1998 1995
International Institute of Refrigeration 1973 1971 1997
Mekong River Commission 1978
Baltic Environmental Forum 1995
International Commission
for the Northwest Atlantic Fisheries 1992
Conference of African Ministers Responsible
for Sustainable Development 1993
International African Migratory Locust Organization
a
1960
Joint Anti-Locust and Anti-Aviarian Organization 1965
International Seabed Authority 1994
Lake Chad Basin Commission 1964
Niger River Commission 1964
South Asia Cooperative Environment Programme 1987
International Whaling Commission 1981
International Oil Pollution Compensation Fund 1984 1987
Intergovernmental Oceanographic Commission 1971 1960
International Maritime Organization 1961 1959 1995
Global Network for Biosystematics (BIONET) 1998 1998 1998
International Plant Genetic Resources Institute 1994 1994
a
no longer in existence since 1985
166
CHAPTER 6. IGO MEMBERSHIP AND ENVIRONMENTAL
QUALITY: UNDERSTANDING THE MECHANISMS
T
a
b
l
e
6
.
1
3
:
M
e
m
b
e
r
s
h
i
p
i
n
e
n
v
i
r
o
n
m
e
n
t
a
l
t
r
e
a
t
i
e
s
E
n
v
i
r
o
n
m
e
n
t
a
l
I
G
O
S
C
a
m
e
r
o
o
n
I
n
d
i
a
L
a
o
s
L
i
t
h
u
a
n
i
a
I
n
t
e
r
n
a
t
i
o
n
a
l
C
o
n
v
e
n
t
i
o
n
O
n
C
i
v
i
l
L
i
a
b
i
l
i
t
y
F
o
r
O
i
l
P
o
l
l
u
t
i
o
n
D
a
m
a
g
e
1
9
8
4
1
9
9
7
I
n
t
e
r
n
a
t
i
o
n
a
l
C
o
n
v
e
n
t
i
o
n
f
o
r
t
h
e
R
e
g
u
l
a
t
i
o
n
o
f
W
h
a
l
i
n
g
1
9
8
1
I
n
t
e
r
n
a
t
i
o
n
a
l
C
o
n
v
e
n
t
i
o
n
f
o
r
t
h
e
P
r
e
v
e
n
t
i
o
n
o
f
P
o
l
l
u
t
i
o
n
o
f
t
h
e
S
e
a
b
y
O
i
l
1
9
6
2
a
n
d
1
9
6
9
1
9
7
4
C
o
n
v
e
n
t
i
o
n
o
n
I
n
t
e
r
n
a
t
i
o
n
a
l
T
r
a
d
e
i
n
E
n
d
a
n
g
e
r
e
d
S
p
e
c
i
e
s
o
f
W
i
l
d
F
a
u
n
a
a
n
d
F
l
o
r
a
1
9
8
1
1
9
7
6
C
o
n
v
e
n
t
i
o
n
o
n
t
h
e
C
o
n
s
e
r
v
a
t
i
o
n
o
f
M
i
g
r
a
t
o
r
y
S
p
e
c
i
e
s
o
f
W
i
l
d
A
n
i
m
a
l
s
1
9
8
2
C
o
n
v
e
n
t
i
o
n
o
n
L
o
n
g
-
R
a
n
g
e
T
r
a
n
s
b
o
u
n
d
a
r
y
A
i
r
P
o
l
l
u
t
i
o
n
1
9
9
4
C
o
n
v
e
n
t
i
o
n
c
o
n
c
e
r
n
i
n
g
t
h
e
P
r
o
t
e
c
t
i
o
n
o
f
t
h
e
W
o
r
l
d
C
u
l
t
u
r
a
l
a
n
d
N
a
t
u
r
a
l
H
e
r
i
t
a
g
e
1
9
8
2
1
9
7
7
1
9
8
7
1
9
9
2
C
o
n
v
e
n
t
i
o
n
o
n
W
e
t
l
a
n
d
s
o
f
I
n
t
e
r
n
a
t
i
o
n
a
l
I
m
p
o
r
t
a
n
c
e
e
s
p
e
c
i
a
l
l
y
a
s
W
a
t
e
r
f
o
w
l
H
a
b
i
t
a
t
1
9
8
1
1
9
9
3
U
n
i
t
e
d
N
a
t
i
o
n
s
C
o
n
v
e
n
t
i
o
n
o
n
t
h
e
L
a
w
o
f
t
h
e
S
e
a
1
9
8
5
1
9
9
5
C
o
n
v
e
n
t
i
o
n
f
o
r
t
h
e
P
r
o
t
e
c
t
i
o
n
o
f
t
h
e
O
z
o
n
e
L
a
y
e
r
1
9
9
1
1
9
9
5
P
r
o
t
o
c
o
l
o
n
S
u
b
s
t
a
n
c
e
s
t
h
a
t
D
e
p
l
e
t
e
t
h
e
O
z
o
n
e
L
a
y
e
r
1
9
8
9
1
9
9
2
1
9
9
8
1
9
9
5
C
o
n
v
e
n
t
i
o
n
o
n
E
n
v
i
r
o
n
m
e
n
t
a
l
I
m
p
a
c
t
A
s
s
e
s
s
m
e
n
t
i
n
a
T
r
a
n
s
b
o
u
n
d
a
r
y
C
o
n
t
e
x
t
2
0
0
4
I
n
t
e
r
n
a
t
i
o
n
a
l
C
o
n
v
e
n
t
i
o
n
o
n
O
i
l
P
o
l
l
u
t
i
o
n
P
r
e
p
a
r
e
d
n
e
s
s
R
e
s
p
o
n
s
e
a
n
d
C
o
-
o
p
e
r
a
t
i
o
n
1
9
9
7
C
o
n
v
e
n
t
i
o
n
o
n
T
r
a
n
s
b
o
u
n
d
a
r
y
E

e
c
t
s
o
f
I
n
d
u
s
t
r
i
a
l
A
c
c
i
d
e
n
t
s
2
0
0
4
U
n
i
t
e
d
N
a
t
i
o
n
s
F
r
a
m
e
w
o
r
k
C
o
n
v
e
n
t
i
o
n
o
n
C
l
i
m
a
t
e
C
h
a
n
g
e
1
9
9
4
1
9
9
3
1
9
9
5
1
9
9
5
C
o
n
v
e
n
t
i
o
n
o
n
B
i
o
l
o
g
i
c
a
l
D
i
v
e
r
s
i
t
y
1
9
9
4
1
9
9
4
1
9
9
6
1
9
9
6
I
n
t
e
r
n
a
t
i
o
n
a
l
C
o
n
v
e
n
t
i
o
n
t
o
C
o
m
b
a
t
D
e
s
e
r
t
i

c
a
t
i
o
n
i
n
t
h
o
s
e
C
o
u
n
t
r
i
e
s
E
x
p
e
r
i
e
n
c
i
n
g
S
e
r
i
o
u
s
D
r
o
u
g
h
t
a
n
d
o
r
D
e
s
e
r
t
i

c
a
t
i
o
n
1
9
9
7
1
9
9
6
1
9
9
6
2
0
0
4
167
CHAPTER 6. IGO MEMBERSHIP AND ENVIRONMENTAL
QUALITY: UNDERSTANDING THE MECHANISMS
6.4.3 Cameroon
As already mentioned above, one of the major environmental problems in
Cameroon is deforestation. One organization that is directly related to this
issue is the African Timber Organization (ATO) (FAO 2009a). Similar to
the International Tropical Timber Organization, of which both Cameroon
and India are members, the ATO aims to support its member countries
to achieve ecient and protable forest-based industries while sustaining
natural forests to ensure that biodiversity is conserved and to prevent un-
controlled depletion of this natural resource (International Tropical Timber
Organization 2009).
The African Union, formerly the Organisation of African Unity, is an
umbrella organization trying to promote greater unity and solidarity be-
tween African nations and to foster cooperation in dierent elds such as
economic issues, human rights etc. Among other institutional bodies, the
organization has a Committee on Industry, Science and Technology, Energy,
Natural Resources and Environment (African Union 2009a). The organiza-
tion also adopted a Declaration on the Conservation of Nature and Natural
Resources which underscores the intention of the African Union to promote
sustainable development (the declaration has not been signed or ratied by
Cameroon yet). However, it remains dicult to assess how seriously this
intention is adhered to in reality since the Union provides very little docu-
mentation on its every-day business. An exception is a project on biosafety
jointly undertaken by the African Union and the German technical coopera-
tion agency (Deutsche Gesellschaft f ur Technische Zusammenarbeit GmbH)
to provide the Unions member states with the necessary skills and technical
tools to implement the Cartagena Protocol
23
(African Union 2009b).
Similarly, the Economic Community of Central African States (ECCAS),
of which Cameroon is a member, promotes environmental protection along-
side its main goal of economic development and has also launched a state-
ment on environmental issues (ECCAS 2009). Again, as in the case of the
African Union, it is very dicult to assess how much of the statements
intention is carried over into every-day procedures due to the lack of infor-
mation provided by the organization.
In contrast, the Conference of the West and Central African Ministers
of Agriculture seems to be more application-oriented since the organization
aims to diuse technologies and best practices to its member countries to
help them realize a more ecient yet sustainable and environmental friendly
agricultural production (La Conference des Ministres de lAgriculture de
lAfrique de lOuest et du Centre 2009).
Africare, an international organization that promotes development in
23
The Cartagena Protocol seeks to protect biological diversity from the potential risks
posed by living modied organisms resulting from modern biotechnology (Convention on
Biological Diversity 2009).
168
CHAPTER 6. IGO MEMBERSHIP AND ENVIRONMENTAL
QUALITY: UNDERSTANDING THE MECHANISMS
Africa, also has a considerable focus on environmental issues such as defor-
estation, alternative energy sources, water protection etc. (Africare 2009).
However, compared to donors such as the World Bank, the organization
is rather small with an annual budget in 2006 that amounted to USD 50
million
24
(Africare 2006). Nevertheless, Africare seems to contribute to the
diusion of knowledge and technologies necessary to avoid environmental
degradation in many dierent ways as mentioned above.
The Lake Chad Basin Commission was created in 1964 by the four coun-
tries bordering Lake Chad Cameroon, Chad, Niger, and Nigeria and was
joined in 1994 by the Central African Republic. The aims of the Commission
are to regulate and control the use of water and other natural resources in
the basin and to initiate, promote, and coordinate natural resource develop-
ment projects and research (Lake Chad Basin Project 2009). Cameroon is
also a member of both the Niger Basin Authority and the Maritime Organi-
zation of West And Central Africa, which have adopted various provisions to
protect the natural environment of the Niger river and the coastal maritime
areas respectively.
Overall, in contrast to the more general IGOs discussed before of which
all of the four countries are members, it seems that for many of the Cameroon-
specic organizations the link to the environment is less obvious or seems
to exist merely on paper, as in the case of ECCAS. Hence environmental
capacity building induced by IGOs seems to come from the large organiza-
tions such as the FAO or the World Bank rather than from the other IGOs
of which Cameroon is a member.
6.4.4 India
One of the many IGOs of which India is a member is the Afro-Asian Rural
Development Organization (AARDO), which promotes rural development in
its 30 member countries. In its most recent triennium program AARDO has
laid the focus on issues such as water scarcity, environment, desertication
and sustainable agricultural and rural development (AARDO 2009a). The
organization intends to achieve these goals mainly by human development,
i.e. training rural workers in using the latest technologies and by dius-
ing new technologies through pilot projects to rural communities (AARDO
2009b).
Similarly, the Network of Aquaculture Centers in Asia-Pacic advocates
rural development through sustainable aquaculture, which they do mostly
through capacity building i.e. through education, training and research
(Network of Aquaculture Centres in Asia-Pacic 2009).
A pure environmental organization is the South Asia Co-operative En-
vironment Programme, which is an inter-governmental organization, estab-
24
Development aid administered by Africare is not included in the OECD data on
development aid (OECD 2008).
169
CHAPTER 6. IGO MEMBERSHIP AND ENVIRONMENTAL
QUALITY: UNDERSTANDING THE MECHANISMS
lished in 1982 by the governments of South Asia to promote and support
protection, management and enhancement of the environment in the region
(South Asia Co-operative Environment Programme 2009).
Another South Asian organization is the Association for Regional Co-
operation (SAARC) which aims to foster social and economic development
among its member countries in seven main areas. These issue areas include
agriculture and rural development and also environment and forestry. Under
this heading SAARC, for example, adopted an action plan on the environ-
ment that requires each member state to assess the national state of the
environment and to launch a national report (SAARC 2009). Although the
ultimate goal is to adopt joint environmental regulations to prevent envi-
ronmental degradation, this has yet to happen. However, the organization
still increases the awareness of its member countries with respect to envi-
ronmental issues, which is of course an important precondition for enacting
appropriate environmental regulations.
Although the Group of Fifteen was established as an organization to
foster south-south trade and to be recognized as a dialogue partner of the
G8, the group also conducts development and technical projects to increase
south-south cooperation (Group of Fifteen 2008). Some of these projects
such as the solar energy application project in India even have a clear envi-
ronmental focus, contributing therefore to the dissemination of technology
and knowledge needed to reduce environmental pollution.
India is also a member of CAB International, an organization that fo-
cuses on scientic publishing, research and communication in the elds of
agriculture and environment (CAB International 2009). An example of the
work of CAB International is the Methyl Bromide Critical Use Nomination
Assessment Project, which seeks to provide instructions to the participating
countries on how to meet the obligations concerning the phase-out of methyl
bromide an ozone-depleting substance under the Montreal Protocol.
Although one would not expect that the International Copper Study
Group (ICSG) serves any environmental purpose the group, besides provid-
ing information on copper production, consumption and trade, also consults
its member countries on various environmental and health issues related to
copper production
25
. An interesting example of the work of the ICSG in
this regard is the implementation of a project that aims to promote the
use of copper for energy ecient products in India and China (ICSG 2009).
However, it is still important to note that the main goal of most of these
industry specic organizations is to increase production and consumption of
the specic product that the organization deals with (copper, cocoa, sugar,
bauxite, to name only a few examples), which mostly conicts sharply with
25
The same rationale holds for the International Lead and Zinc Study Group, which
also provides knowledge on how to avoid adverse eects of lead and zinc on human health
and the natural environment (The International Lead and Zinc Study Group 2009).
170
CHAPTER 6. IGO MEMBERSHIP AND ENVIRONMENTAL
QUALITY: UNDERSTANDING THE MECHANISMS
environmental protection. This is also supported by the statistical results
described above, since in none of the models have economic industry spe-
cic organizations been associated with a signicant reduction in any of the
forms of environmental quality considered in this analysis
The International Fund for Agricultural Development (IFAD), of which
Cameroon, India and Laos are members, is another organization that pro-
vides nancial and technical assistance to rural communities. For commu-
nities in the semi-arid tropical areas of India, for example, IFAD tries to
improve living conditions by establishing better water management systems
(IFAD 2009). More generally, the organization aims to provide communities
with technology and know-how on the sustainable use of natural resources.
In contrast to the example of Cameroon, many of the IGOs of which India
is a member (in addition to the large organizations which all countries have
joined) seem to promote sustainable development and diuse technologies
and know-how related to environmental capacity building.
6.4.5 Laos
A very important organization for Laos is the Mekong River Commission,
previously known as the Interim Committee for Coordination of Investiga-
tions of the Lower Mekong Basin and before that the Mekong Committee.
This organization consisting of Laos, Cambodia, Thailand and Vietnam
intends to regulate the Mekong River Basin for both agriculture and hy-
dropower. The Commission also monitors pollution in the river (especially
arising due to salinity intrusion), thus striving for sustainable use of the
river basin (Jacobs 1996).
The Southeast Asian Ministers of Education Organization (SEAMEO)
is an inter-governmental organization consisting of 11 Southeast Asian na-
tions to foster cooperation in the realm of education and culture (SEAMEO
2009b). Through training programs the organization provides human re-
source development in the elds of learning, health and environment, agri-
culture and natural resources. In this way, SEAMEO aims to increase peo-
ples awareness and knowledge of important health and environmental haz-
ards. For example, SEAMEO developed a toolkit on water sanitation and
hygiene education to improve sanitation facilities in their member countries
(SEAMEO 2009a).
Both Laos and India are part of the Colombo Plan for cooperative eco-
nomic and social development in Asia and the Pacic, which is again an
organization committed to the promotion of its member countries economic
and social development. In recent years, the organization, at the initiative
of the Thai government, has also mobilised resources to launch a program
on the environment covering various topics such as climate change, air pol-
lution management and industrial ecology (The Colombo Plan 2009). How-
ever, this shift towards environmental topics seems to have occurred only
171
CHAPTER 6. IGO MEMBERSHIP AND ENVIRONMENTAL
QUALITY: UNDERSTANDING THE MECHANISMS
recently, which makes it unlikely that a potential positive eect on the envi-
ronmental performance of its member countries is reected in the statistical
analysis of this thesis.
Perhaps most importantly for Laos, the country became a full mem-
ber of the Association of Southeast Asian Nations in 1997. Membership in
ASEAN required Laos to ratify various agreements on agro-economic de-
velopment with signicant positive inuences on its environment (UNEP
2001b). Among these agreements Laos needed to ratify were, for example,
the Framework for Environmentally Sustainable Cities and ASEAN Agree-
ment on Transboundary Haze Pollution. In addition to obligating its mem-
ber countries to adopt certain environmental regulations, the environmental
ministers of the ASEAN member countries meet annually.
Altogether, it is striking that Laos has not only joined very few IGOs
but also that the country joined many organizations only in the late 1990s.
In a similar vein, the majority of the World Bank projects in Laos were
only carried out in the last decade. This suggests that the inuence of
international actors is generally fairly limited, although increasing in the
last few years. This impression is further supported by the fact that Laos
receives the smallest amount of environmental aid from the GEF and has
joined the lowest number of environmental treaties compared to any of the
other four countries.
6.4.6 Lithuania
In 1990, the three Baltic nations Estonia, Lithuania and Latvia founded the
Baltic Council, which serves as an inter-governmental organization to foster
cooperation on political, economic and security issues (Estonian Ministry of
Foreign Aairs 2009). The Baltic Council also established an Environmen-
tal Committee whose purpose is to mitigate the impacts of transboundary
pollution and to support the Baltic countries to meet the environmental
obligations they need to comply with under the European Union (Estonian
Ministry of Foreign Aairs 2009).
Established in 1995 as a technical assistance project to strengthen coop-
eration between the Baltic States in the eld of environmental policy, the
Baltic Environmental Forum (BEF) later assisted the Baltic states in the
EU accession process. In particular, the BEF has organized information
and knowledge transfer events with a focus on EU accession policy in the
environmental sector for national environmental administrations, industries,
experts, scientic institutions and municipalities of the three Baltic states
(BEF 2009).
Consequently, becoming a member in the European Union in 2004 also
had important implications for environmental quality in Lithuania. This is
the case because joining the EU requires a country not only to ratify all
EU treaties but also to adopt all common EU law, which includes its en-
172
CHAPTER 6. IGO MEMBERSHIP AND ENVIRONMENTAL
QUALITY: UNDERSTANDING THE MECHANISMS
vironmental legislation. One example of European environmental law that
Lithuania needed to adopt were regulations to limit industrial pollution
(Bundesministerium f ur Umwelt, Naturschutz und Reaktorsicherheit 2009).
Hence similar to Laos accession to ASEAN, membership in such an um-
brella organization can have important implications for the environmental
performance of its member countries. Although countries want to join these
umbrella organizations mostly for economic reasons, they also need to adopt
the organizations regulations in other realms such as labor or environmen-
tal policy. This seems to work especially well in those cases in which the
organization possesses enough power to coerce its future member countries,
as is the case with ASEAN or the EU.
A similar forum to the Baltic Council is the Council of the Baltic Sea
States, which encompasses not only Estonia, Lithuania and Latvia but all
countries with access to the Baltic Sea (The Council of the Baltic Sea States
2009). The Council describes itself as a political forum to intensify regional
inter-governmental cooperation. It does this in the areas of economic de-
velopment, education and culture, energy and environment. In particular,
in 2003 the Council adopted a declaration on sustainable development with
a special focus on water quality. In this declaration the member states of
the Council of the Baltic Sea States commit to cooperate on environmental
issues and transfer the EUs fairly strict environmental policies to member
states that are not part of the EU, such as Russia (Council of Baltic Sea
States 2003).
Finally, a further Baltic organization with something of an environmen-
tal mandate is the Vision and Strategies around the Baltic Sea (VASAB),
which is an organization that focuses on sustainable, balanced and harmo-
nious spatial development in the Baltic Sea Region (VASAB 2009a). To
achieve this goal VASAB disseminates best practices on spatial develop-
ment to ensure that urban planning is undertaken with a strong focus on
sustainable development (VASAB 2009b).
Overall, it seems that environmental capacity building was boosted in
Lithuania during the 1990s, mainly because of the upcoming EU accession.
So although the country has not joined many IGOs overall, Lithuanias
environmental performance seems to have beneted a great deal from the
few organizations it has joined, such as the Baltic Environmental Forum or
the European Bank for Reconstruction and Development.
6.5 Conclusion
The case studies have illustrated various examples of how very dierent
types of IGOs do indeed inuence their member countries with regard to
environmental topics. For some organizations such as the Global Environ-
ment Facility, the connection to environmental protection is quite obvious
173
CHAPTER 6. IGO MEMBERSHIP AND ENVIRONMENTAL
QUALITY: UNDERSTANDING THE MECHANISMS
whereas for other organizations such as the World Tourism Organization or
the International Copper Study Group the link to the environment is more
surprising. Altogether, the aim of these case studies was not to link specic
actions of IGOs to increases in the environmental performance of the four
countries but to show how the objectives and the every-day business of IGOs
could help developing countries to improve their environmental performance.
To obtain more far-reaching and general conclusions on the mechanisms by
which IGO membership can inuence environmental quality in developing
countries I provide a short summary of the above case studies. Furthermore,
I intend to relate the ndings of the case studies to the results of the various
statistical analyses at the beginning of this chapter, in order to highlight
the overall conclusions that we can draw from this in-depth analysis of IGO
membership and environmental quality in developing countries.
Altogether, three main mechanisms by which IGOs seem to inuence
a countrys environmental performance come out of the case studies: issue
linkage, technical cooperation and know-how diusion, and pushing the topic
of environmental protection or sustainable development on the agenda of
developing countries.
Concerning the rst issue we see, as has already been indicated by the
various statistical analyses, that umbrella organizations such as the EU or
ASEAN can be very successful in linking topics such as environmental or
labor protection to trade or other economic issues. This seems to work espe-
cially well in those cases in which the organization possesses enough power
to coerce its future member countries, as is the case for ASEAN or the EU,
whereas it does not seem to hold for organizations that are less coercive and
comprehensive such as the African Union. As described in more detail above,
examples of this mechanism are Laos and Lithuania, which adopted exten-
sive environmental regulations in order to become members of ASEAN and
the EU respectively, although they aimed to join these organizations mainly
for economic reasons. Interestingly, this mechanism of issue linkage seems to
work equally well for democracies and autocracies. As long as the incentives
are strong enough to join these comprehensive and restrictive organizations,
even autocracies such as Laos accept the requirement to implement envi-
ronmental regulation, although their own interest in environmental quality
seems to be rather limited as the discussion above has shown.
With regard to the second issue, it becomes apparent from the case stud-
ies that many organizations such as FAO, WHO, UNESCO or the many rural
development organizations provide developing countries with knowledge and
technology needed to combat environmental pollution. For example, due to
the World Bank Kerala Rural Water Supply and Environmental Sanita-
tion Project, the water sanitation coverage in Kerala, India, has increased
by 10% (World Bank 2009d). Interestingly, this diusion of environmentally
sound technology and know-how and the promotion of environmental capac-
ity building often happen as a side eect while these organizations promote
174
CHAPTER 6. IGO MEMBERSHIP AND ENVIRONMENTAL
QUALITY: UNDERSTANDING THE MECHANISMS
goals such as health care and food security. For example, the FAO, with
a focus on forestry management and shery, disseminates know-how on the
sustainable use of natural resources. This nding of the case studies is in line
with the statistical analysis that shows a signicant eect for social, agrar-
ian, economic cooperation and economic general organizations, which are
all organizations suited for the diusion of better technology and capacity
building.
Moreover, this capacity building by transferring technology and diusing
environmentally sound know-how is seen by organizations such as the GEF
as one of the most important strategies to deal with environmental threats
such as climate change (GEF 2008b). This is underlined by one of the major
ndings of the State of the Environment Report in Laos (UNEP 2001b) since
the report suggests that in order for Laos to deal with any environmental
problems the country needs to invest heavily in capacity building to raise
awareness of environmental issues. Furthermore, the report suggests that
educational and technical training for ocials is needed to help them in their
eorts to combat environmental pollution.
However, judging from the project descriptions and mission statements
of the individual organizations, technology and know-how diusion seem to
work best in larger organizations whose membership also includes industrial-
ized countries, such as the FAO, WHO and IAEA. In the case of Cameroon,
for example, it seems that many of the inter-African organizations that also
aim at capacity building are composed of statements of will and understand-
ing rather than being eective and strong organizations. This impression is
very much in line with the ndings concerning the degree of institutional-
ization of the organizations. As the statistical analysis has shown, in order
to inuence the provision of environmental quality in developing countries,
organizations need to possess some degree of formalization and institution-
alization.
Finally, the four cases show that an increasing number of IGOs have
taken up the course of sustainable development. Although not all of these
organizations do indeed provide developing countries with technologies and
resources necessary to achieve sustainable development, countries can hardly
avoid this topic any more. Sustainable development is mainly promoted by
the many development organizations and rst and foremost by the World
Bank. However, even organizations that on the face of it are not related
to environmental topics subscribe to the goal of sustainable development.
The World Tourism Organization is an interesting example in that it aims
to reconcile tourism with the local culture, economy and the environment
in order to achieve the goal of sustainable tourism (World Tourism Orga-
nization 2009). Furthermore, the eect of the economic organization with
general functions such as promoting trade shown in the statistical analysis
also seems to be due to the promotion of sustainable development. Both
the International Finance Corporation and UNIDO are examples of organi-
175
CHAPTER 6. IGO MEMBERSHIP AND ENVIRONMENTAL
QUALITY: UNDERSTANDING THE MECHANISMS
zations trying to reconcile economic growth with a countrys environmental
responsibility.
Furthermore, the case studies show that strong interlinkages exist be-
tween IGO membership, international aid and the environment. Most inter-
national organizations that provide international aid to developing countries
also follow this trend of sustainable development in their lending behavior.
For example, countries applying for loans with the International Develop-
ment Association are obliged to provide a National Environmental Action
Plan, which outlines a countrys major environmental problems and de-
scribes solutions to mitigate these problems (Gutner 2005). Furthermore,
all projects carried out by the World Bank need to undergo environmental
appraisal before they are approved (Gutner 2005).
Of course, when describing those areas of operations of an IGO that are
related to environmental protection, it is important to bear in mind that
the very same organization might also contribute to pollution. For example,
although the World Bank devotes many resources to promote sustainable
development, one of its major tasks is still to nance large-scale infrastruc-
ture projects such as the construction of airports or roads, which necessarily
contribute to increased pollution levels. This might be one reason why we
do not see a statistically signicant eect of international aid on environ-
mental quality in developing countries. Since international aid is, on the one
hand, characterized by the transfer of better technology and the promotion
of sustainable development but on the other hand, most international aid
aims to induce economic growth, the overall eect of aid seems to be neutral
with regard to its environmental implications.
The analysis of the dierent mechanisms behind the relationship between
IGO membership and environmental quality further hints at a major dier-
ence between democracies and autocracies: both the diusion and trans-
fer of technology and the promotion of sustainable development seem to
ourish more intensively within democracies than within autocracies. This
conclusion is supported by the fact that whereas it was no problem to get
information from the environmental ministries in India and Lithuania, the
governments of Cameroon and Laos do not provide valuable information
concerning environmental issues on the Internet. Furthermore, many more
non-governmental organizations such as the WWF seem to be active in In-
dia and Lithuania compared to Cameroon and Laos, thereby also fostering
the diusion and transfer of technologies and know-how necessary to re-
duce pollution (UNEP 2001b). This is very much in line with the State
of the Environment Report in Laos (UNEP 2001b), which identies Laos
centralized and closed system as a major obstacle to better environmental
performance since it seems to be very dicult to build up responsibilities
for the enforcement of environmental regulations in such a context. This
consequently serves as an indication that information and knowledge might
indeed spread more easily in democratic countries than in autocracies since
176
CHAPTER 6. IGO MEMBERSHIP AND ENVIRONMENTAL
QUALITY: UNDERSTANDING THE MECHANISMS
democracies provide many more access point for the development of envi-
ronmental capacities.
Furthermore, another striking dierence between the two democracies
and the two autocracies seems to be the degree to which these countries
enact environmental regulations: both India and Lithuania are character-
ized by a high degree of environmental regulation whereas both Laos and
Cameroon stand out due to the lack of appropriate regulations (UNEP
2001b,a; WWF 2009). This pattern also manifests itself in the implementa-
tion of international treaties. As described above, Laos became a member
of the Convention on Biological Diversity in 1996. However, appropriate
regulation to protect Laos immense wildlife is still lacking (UNEP 2001b).
Similarly, Cameroon has still not ratied the African Unions Declaration on
the Conservation of Nature and Natural Resources (African Union 2009a).
Additionally, both India and Lithuania have joined more global environmen-
tal treaties, as shown above, and the proportion of environmental projects
funded by the World Bank is also higher for both India and Lithuania com-
pared to Cameroon and Laos.
Hence all of these observations underline the nding that most of the time
we see that democratic developing countries seem to be more responsive to
the inuence of international actors than their autocratic counterparts (with
the exception of the interaction eect between IGO membership and the
political system on CO
2
pollution, which indicates that autocracies perform
better in decreasing CO
2
emissions). Thus in the majority of cases the
eect of international integration is indeed more conducive to environmental
quality in democracies than in autocracies, with IGO membership being
the one facet of international integration that is eectively and robustly
associated to an increase in the environmental performance of developing
countries.
177
Chapter 7
Formal Model on IGO
Membership and
Environmental Quality
In the last chapter, the relationship between IGO membership and environ-
mental quality in developing countries was analyzed in great detail empir-
ically. This analysis showed that technology and knowledge transfer is an
important mechanism through which IGOs bring about an increase in the
environmental performance of developing countries. To accommodate these
ndings theoretically as well, this chapter aims to build a simple formal the-
oretical model that links IGO membership with environmental pollution.
The aim of this model is to obtain a more general theoretical framework for
the study of IGO membership and its links to environmental pollution.
In order to theoretically model the eect of IGO membership on envi-
ronmental quality in developing countries, I rely on a specic-factors trade
model, which is a widely accepted modern treatment for analyzing trade im-
pacts (Rogowski 2006). Using this framework, it is possible to analyze how
membership in IGOs alters the incentive structure of a countrys economy
and thus inuences environmental quality in that country.
7.1 The Economy
In particular, I consider an economy that produces two goods X and Y
using the two inputs labor L and capital K. Production follows a standard
Cobb-Douglas production function:
X = A
x
L

x
K
1
x
(7.1)
Y = A
y
L

y
K
1
y
(7.2)
where (0, 1)
178
CHAPTER 7. FORMAL MODEL ON IGO MEMBERSHIP AND
ENVIRONMENTAL QUALITY
The parameter A captures the production technology and therefore the e-
ciency of production. A higher level of A implies that more output is reached
given the inputs of capital and labor. In the context of this model, A also
stands for the level of environmental stringency of the production where
high values of A imply low environmental stringency or regulation and low
values of A stand for more stringent environmental regulation. This is a rea-
sonable set-up to capture environmental stringency or regulation since envi-
ronmental regulations usually imply the use of more stringent and advanced
technology. The use of more stringent technology in turn usually renders
production more expensive, therefore implying that less output is obtained
given the same input levels. Modeling pollution abatement as costly for the
production process is in line with other models on environmental pollution
such as Antweiler et al. (2001) and Andreoni & Levinson (1998).
By assumption, good X is exported whereas good Y is only consumed do-
mestically.
1
A rm selling good X (good Y respectively) makes the following
prots:

x
= p
x
X wL
x
r
x
K
x
(7.3)

y
= Y wL
y
r
y
K
y
(7.4)
where w is the cost of labor and r
x
the cost of capital in sector X and r
y
the
cost of capital in sector Y. p
x
is the price of good X and good Y is taken to
be the numeraire hence p
y
= 1. We further assume a small open economy,
which is true for most developing countries in the world, and that therefore
the country itself cannot inuence p
x
. Substituting out for X in equation
(7.3) using equation (7.1) (respectively for Y in equation (7.4) using equation
(7.2)) yields:

x
= p
x
A
x
L

x
K
1
x
wL
x
r
x
K
x

y
= A
y
L

y
K
1
y
wL
y
r
y
K
y
Assuming that markets are competitive and that rms maximize prots we
obtain the following rst-order conditions for rms choice of L:
p
x
A
x
_
L
x
K
x
_
1
= w (7.5)
A
y
_
L
y
K
y
_
1
= w (7.6)
1
Furthermore, to facilitate later analysis we assume that citizens only consume good Y
and that all of good X is exported. One could think of this assumption as good Y being
food and good X being T-shirts, for example.
179
CHAPTER 7. FORMAL MODEL ON IGO MEMBERSHIP AND
ENVIRONMENTAL QUALITY
Since in this model labor is mobile and can thus move freely from sector
X to sector Y and vice versa, the marginal product of labor, i.e. the wage,
is the same in both sectors and hence we can set equation (7.5) equal to
equation (7.6):
p
x
A
x
_
L
x
K
x
_
1
= A
y
_
L
y
K
y
_
1
L
x
=
_
A
y
p
x
A
x
_ 1
1
K
x
K
y
L
y
Using the fact that aggregate labor is L = L
x
+L
y
we obtain:
L L
y
=
_
A
y
p
x
A
x
_ 1
1
K
x
K
y
L
y
L
y
=
_
1 +
_
A
y
p
x
A
x
_ 1
1
K
x
K
y
_
1
L
Similarly:
L
x
=
_
1 +
_
p
x
A
x
A
y
_ 1
1
K
y
K
x
_
1
L
This allows us to express production in the two sectors as a function of
aggregate labor endowments:
X = A
x
_
1 +
_
p
x
A
x
A
y
_ 1
1
K
y
K
x
_

K
1
x
(7.7)
Y = A
y
_
1 +
_
A
y
p
x
A
x
_ 1
1
K
x
K
y
_

K
1
y
(7.8)
7.2 Government
Now a government needs to decide whether it should join international gov-
ernmental organizations. Joining an IGO is treated in this set-up as if a
country were joining a club. This club allows the country to sell its exports
at better conditions, modeled as higher prices, but at the expense of using
a more environmentally stringent technology. This implies that the eect
of IGO membership m on the price p
x
is positive whereas the eect of m
on the technology A is negative. We further assume that the international
180
CHAPTER 7. FORMAL MODEL ON IGO MEMBERSHIP AND
ENVIRONMENTAL QUALITY
community only has inuence over the production of the exported good
2
.
Hence the prot of a rm in the export sector becomes:

x
= p
x
(m)A
x
(m)L

x
K
1
x
wL
x
r
x
K
x
where m indicates membership in IGOs (m can be understood as the in-
tensity with which a government decides to be integrated in the network of
international organizations, hence m is not a binary decision) and where
p
x
m
> 0 and
A
x
m
< 0
Hence joining an IGO poses a trade-o for each country: upon entering, it
can sell under better terms of trade (i.e. at higher prices) but at the same
time it needs to use more expensive technology. Inserting these specications
into the production equations (7.7) and (7.8) we obtain
X = A
x
(m)
_
1 +
_
p
x
(m)
A
x
(m)
A
y
_ 1
1
K
y
K
x
_

K
1
x
(7.9)
Y = A
y
_
1 +
_
A
y
p
x
(m)A
x
(m)
_ 1
1
K
x
K
y
_

K
1
y
(7.10)
To progress the analysis, we need to specify a functional form for both p
x
(m)
and A
x
(m). Concerning the price at which a country can sell its exports,
it is reasonable that the price does not depend linearly on a countrys IGO
membership but that joining the rst IGOs has a more pronounced eect
than joining later IGOs, implying that the eect of IGO membership m on
the price p
x
(m) is increasing but at a decreasing rate. One simple function
that captures this idea is
3
p
x
(m) = p
x
(1
1
m
) (7.11)
implying that
2
This assumption seems to be reasonable since other countries could decide not to buy
foreign goods that are produced under low standards but it is dicult to imagine how a
country could punish a foreign domestic sector for applying low environmental standards
(with the exception of the European Union, which is an IGO that prescribes production
regulation for its member countries). However, the simulation results are not sensitive
to this decision since the simulation shows similar results if we assume that both sectors
need to employ a more stringent technology.
3
Using this functional form requires, however, that IGO membership is positive, thus
that m = 0. Since in reality no country exists that is not a member of at least one IGO,
I do not consider this specication of the functional form for m as too restrictive.
181
CHAPTER 7. FORMAL MODEL ON IGO MEMBERSHIP AND
ENVIRONMENTAL QUALITY
p
x
m
= p
x
1
m
2
> 0
An empirical justication for modeling IGO membership in this way is de-
rived in section 9.3 in the Appendix.
After having specied the functional form showing how the price of the
export good p
x
depends on IGO membership m we turn to the second pa-
rameter that depends on m. Concerning the technology A
x
used in the
exporting sector, we employ an even more straightforward and simple spec-
ication by using a linear functional form:
A
x
(m) = A
x
mr (7.12)
and with
A
x
m
= r < 0
(7.13)
where r is the rate at which IGO membership makes the technology more
stringent and thus more environmental friendly
4
. The technology A
x
there-
fore decreases with IGO membership, implying that becoming a member in
an IGO induces countries to employ a more stringent production technol-
ogy. Employing a more stringent production technology means that on the
one hand, rms can produce less given the same amount of inputs and, on
the other hand, that production is more environmentally friendly due to the
greater stringency.
Some evidence that this assumption concerning the technology param-
eter is reasonable can be derived from the case study evidence in section
6.4. The analysis of how the dierent IGOs try to inuence their member
countries has shown that one major function of IGOs in developing coun-
tries is to diuse better technologies. Most of the IGOs analyzed in the
case study section either supply countries directly with new technology or
promote capacity building by transferring knowledge and information to im-
portant stakeholders in developing countries. Consequently, the assumption
that IGO membership has a positive eect on the stringency of the techno-
logical level in a country is therefore clearly supported by this case study
evidence.
4
Although this speccation means that we add an additional parameter to the model,
introducing r allows us to see how dierent rates at which the technology becomes more
stringent inuence the results. This makes the results even more generic since we can
consider both very coercive IGOs that induce countries to use a more stringent technology
and rather lax IGOs that cannot inuence the environmental stringency of its member
countries so much.
182
CHAPTER 7. FORMAL MODEL ON IGO MEMBERSHIP AND
ENVIRONMENTAL QUALITY
After having specied how IGO membership inuences both the price of
the export good and the level of technology, we turn to the question of how
the government makes its decision concerning the degree to which it should
become a member of international organizations m. Following Rogowski
(2003) I model the distinction between a democracy and an autocracy by
whether the government embraces the optimal strategy of the representative
voter, i.e. the average worker, or the optimal strategy of the representative
elite member, i.e. the representative capitalist in either sector X or Y. In
particular, I will consider four dierent scenarios: a government that adopts
the strategy of the capitalists in the exporting sector, the capitalists in the
domestic sector, the workers in the exporting sector and the workers in
the domestic sector. Therefore, we can think of the rst two scenarios as
reecting a non-democracy in which the decision process reects the wishes
of the representative elite member (i.e. the representative capitalist) whereas
the last two scenarios reect a democratic setting in which the decision
process matches the optimal strategy of the representative voter (i.e. the
average worker).
More precisely, again following Rogowski (2003), we assume that the rep-
resentative voter (elite member) cares only about her income. Consequently,
we model the governments choice of IGO membership by the government
maximizing the income of the respective representative voter or elite mem-
ber. We thereby assume that the representative voter (capitalist) earns the
marginal product of labor (of capital), which can be thought of as the aver-
age wage (average capital earning). This implies that we abstract from the
eect of taxes or government expenditures and that we treat the decision to
which degree a government should join international organizations as a one-
shot game. Hence we assume that if a government embraces the position of
a certain actor (worker, capitalist), this means that the government tries to
maximize the income of this actor.
Furthermore, since we have already assumed that citizens only consume
the domestic good Y, a change in the price of good X does not aect them as
consumers. This allows us to abstract from the fact that an increase in the
price of good X would aect consumers negatively. If we had not done so,
we would have needed to specify for each representative agent a preference
order concerning the consumption of the two goods. This would have im-
plied even more assumptions without yielding great additional benets since
considering the income eects IGO membership has for each representative
agent is most important.
7.2.1 Government Favors Elite Member in the Export Sector
The rst scenario we consider is the case in which we imagine a country
in which the government embraces the optimal strategy of the producers of
good X only. This would correspond to a non-democracy since the govern-
183
CHAPTER 7. FORMAL MODEL ON IGO MEMBERSHIP AND
ENVIRONMENTAL QUALITY
ment would only consider the optimum for the representative capitalist in
sector X. In this case, the government would maximize the return to capital
in sector X, which is what producers in sector X earn, with respect to m in
order to determine the optimal degree of IGO membership for this country.
The marginal rent on capital in sector X follows from maximizing equation
7.9 with respect to K:
MRC
x
=
X
K
x
= (1 )A
x
(m)
_
1 +
_
p
x
(m)
A
x
(m)
A
y
_ 1
1
K
y
K
x
_

_
L
K
x
_

+ A
x
(m)
_
1 +
_
p
x
(m)
A
x
(m)
A
y
_ 1
1
K
y
K
x
_
1
L

K
1
x
_
p
x
(m)
A
x
(m)
A
y
_ 1
1
K
y
K
2
x
Using the functional forms we specied for both price and technology
equations 7.11 and 7.12 we obtain
MRC
x
= (1 )(A
x
mr)
_
1 +
_
p
x
(1
1
m
)
(A
x
mr)
A
y
_ 1
1
K
y
K
x
_

_
L
K
x
_

+ A
x
(m)
_
1 +
_
p
x
(1
1
m
)
(A
x
mr)
A
y
_ 1
1
K
y
K
x
_
1
L

K
1
x
_
p
x
(1
1
m
)
(A
x
mr)
A
y
_ 1
1
K
y
K
2
x
Now the government makes its choice regarding the optimal level of IGO
membership:
184
CHAPTER 7. FORMAL MODEL ON IGO MEMBERSHIP AND
ENVIRONMENTAL QUALITY

M
R
C
x

m
=

A
x
(
m
)

m
_ _ _
(
1

)
_
1
+
_
p
x
(
1

1
m
)
(
A
x

m
r
)
A
y
_
1

1
K
y
K
x
_

_
L
K
x
_

_
1
+
_
p
x
(
1

1
m
)
(
A
x

m
r
)
A
y
_
1

1
K
y
K
x
_

1
L

K
1

x
_
p
x
(
1

1
m
)
(
A
x

m
r
)
A
y
_
1

1
K
y
K
2 x
_ _ _
+
(
A
x

m
r
)
_ _ _
(
1

)
_
L
K
x
_

)
_
1
+
_
p
x
(
1

1
m
)
(
A
x

m
r
)
A
y
_
1

1
K
y
K
x
_

1
1

1
_
p
x
(
1

1
m
)
(
A
x

m
r
)
A
y
_
1

1
_

p
x
(
m
)

m
(
A
x

m
r
)
A
y
+

A
x
(
m
)

m
p
x
(
1

1
m
)
A
y
_
K
y
K
x
+

K
1

x
K
y
K
2 x
_
p
x
(
1

1
m
)
(
A
x

m
r
)
A
y
_
1

1
(

1
)
_
1
+
_
p
x
(
1

1
m
)
(
A
x

m
r
)
A
y
_
1

1
K
y
K
x
_

2
K
y
K
x
1

1
_
p
x
(
1

1
m
)
(
A
x

m
r
)
A
y
_
1

1
_

p
x
(
m
)

m
(
A
x

m
r
)
A
y
+

A
x
(
m
)

m
p
x
(
1

1
m
)
A
y
_
+

K
1

x
K
y
K
2 x
_
1
+
_
p
x
(
1

1
m
)
(
A
x

m
r
)
A
y
_
1

1
K
y
K
x
_

1
1

1
_
p
x
(
1

1
m
)
(
A
x

m
r
)
A
y
_
1

1
_

p
x
(
m
)

m
(
A
x

m
r
)
A
y
+

A
x
(
m
)

m
p
x
(
1

1
m
)
A
y
_
_
!
=
0
185
CHAPTER 7. FORMAL MODEL ON IGO MEMBERSHIP AND
ENVIRONMENTAL QUALITY
After rearranging, this simplies to
r
_
(1 )
_
_
p
x
(1
1
m
)
(A
x
mr)
A
y
_
(
1
1
)
+
K
y
K
x
_
+
K
y
K
x
_
= (A
x
mr)

1
_
p
x
(1
1
m
)
(A
x
mr)
A
y
_
1
_
p
m
2
(A
x
mr)
A
y
r
p
x
(1
1
m
)
A
y
_
K
y
K
x
_
( 1) +
_
p
x
(1
1
m
)
(A
x
mr)
A
y
_ 1
1
( 1)
K
y
K
x
_
1 +
_
p
x
(1
1
m
)
(A
x
mr)
A
y
_ 1
1
K
y
K
x
_
1
+ 1
_
_
(7.14)
Since this equation does not have an analytical solution, we cannot simply
determine the optimal value for IGO membership m. Therefore, it is nec-
essary to use simulation to obtain optimal values for IGO membership m
given specic values of the other variables in the equation. In doing so, it
is also possible to vary the values of all the other variables in equation 7.14,
such as the ratio between K
x
and K
y
, to see how IGO membership changes
given the changing values of the respective variable
5
.
As a baseline setting for the simulation I use a situation in which the
exporting sector X is smaller than the domestic sector Y thus
K
y
K
x
> 1. In
particular, I set
K
y
K
x
= 1.1 in the baseline setting. Similarly, I set p
x
= 1.1
in this baseline setting, which implies that the price of the exported good is
higher than the price of the domestic good, which was normalized to one.
Furthermore, I assume that without IGO membership sector X and sector
Y have the same technology A thus A
x
= A
y
. For simplicity, A is set to 1 in
the baseline setting of the simulation. Regarding the share of value added
that goes to workers, , I rely on empirical studies that estimate this share
using mostly US data to determine a value for . These estimated values
for usually range from 0.5 to 0.9 (Rogowski 2003). I therefore set = 0.6
in the baseline setting. Table 7.1 summarizes the starting values for the
baseline simulation.
The nal variable that needs to be specied is the rate r at which IGO
membership renders the production technology more stringent. Since little
theoretical guidance exists on how to choose a starting value, I set r = 0
at the outset, which corresponds to the case in which the IGO would not
demand any adjustments in a countrys technology level. In this setting,
we therefore see no benets for environmental quality since the production
process does not become more environmentally friendly. I then increase
the value of r in steps of 0.001 increments until r reaches the value 0.1,
5
All simulations are obtained using MATLAB.
186
CHAPTER 7. FORMAL MODEL ON IGO MEMBERSHIP AND
ENVIRONMENTAL QUALITY
Table 7.1: Starting values for the baseline simulation
export sector
elite in domestic sector workers in domestic sector
A 1 1
p
x
1.1 1.1
0.6 0.6
r 0 r 0.1 0 r 0.1
K
y
K
x
1.1
K
x
K
y
1.1
which would correspond to a setting in which the IGO is rather coercive and
demands a large adjustment in a countrys technological level. This setting
therefore illustrates the case in which the production process becomes much
more environmentally friendly. Hence in all simulations below, r varies
between 0 r 0.1, reecting a steady increase in the environmental
stringency of the production process.
Figure 7.1(a) shows the value of m while varying r setting all other
variables to their baseline values. Interestingly, until r is approximately
0.015, IGO membership increases with r. Only after this peak does IGO
membership decreasewith r, which implies that after a certain threshold if
r increases IGO membership becomes too costly for the exporting sector,
resulting in a lower IGO membership. Hence in general, the higher the rate
r at which IGO membership renders the production technology more strin-
gent, the costlier IGO membership becomes for a country, leading to lower
optimal values of IGO membership. Figures 7.1(b) and 7.1(c) show the same
simulation as in gure 7.1(a), however with higher values for the technology
parameter A. Therefore, gures 7.1(b) and 7.1(c) illustrate countries with a
higher level of technology. In particular, we see that countries that are more
productive in general opt for higher IGO membership than countries with
a lower level of technology A. This is apparent as the values on the y-axis,
which reect the intensity of IGO membership, are higher in gures 7.1(b)
and 7.1(c) compared to gure 7.1(a) whereas the overall shape of the curve
is similar in all three gures.
187
CHAPTER 7. FORMAL MODEL ON IGO MEMBERSHIP AND
ENVIRONMENTAL QUALITY
(a) A=1
(b) A=2
(c) A=3
Figure 7.1: Optimal IGO membership at dierent levels of technology A
188
CHAPTER 7. FORMAL MODEL ON IGO MEMBERSHIP AND
ENVIRONMENTAL QUALITY
(
a
)
K
y
K
x
=
0
.
8
(
b
)
K
y
K
x
=
0
.
9
(
c
)
K
y
K
x
=
1
(
d
)
K
y
K
x
=
1
.
1
(
e
)
K
y
K
x
=
1
.
2
F
i
g
u
r
e
7
.
2
:
O
p
t
i
m
a
l
I
G
O
m
e
m
b
e
r
s
h
i
p
a
t
d
i

e
r
e
n
t
l
e
v
e
l
s
o
f
K
y
K
x
189
CHAPTER 7. FORMAL MODEL ON IGO MEMBERSHIP AND
ENVIRONMENTAL QUALITY
In the next setting, gures 7.2(a) to 7.2(e), we see optimal IGO mem-
bership at dierent ratios between the two sectors K
y
and K
x
while varying
r. In particular, the gures show that if the exporting sector is larger than
the domestic sector gures 7.2(a) and 7.2(b) optimal IGO membership is
smaller than if the domestic sector is larger than the exporting sector g-
ures 7.2(d) and 7.2(e). Thus a government embracing the optimal strategy
of the representative capitalist in the exporting sector (sector X) will opt
for a lower IGO membership if its own sector (the exporting sector) is larger
than the domestic sector whereas it opts for a higher IGO membership if its
own sector is smaller than the domestic sector.
The reverse picture arises if we consider optimal IGO membership at
dierent prices p
x
for the export good gures 7.3(a) to 7.3(e) while
varying r and xing all other variables at their baseline value. Here, we see
that optimal IGO membership is higher for all values of r if the price at
which a country can export good X increases. This implies that the higher
the price of the export good p
x
, the more benecial joining IGOs becomes,
which increases the optimal level of IGO membership. Hence higher values
of p
x
shift the curve between IGO membership and the rate r at which IGO
membership renders the production process more stringent upwards, leading
to generally higher values of IGO membership.
Finally, gures 7.4(a) to 7.4(e) show optimal IGO membership for vary-
ing degrees of r at dierent values for the share of value added that goes to
workers . We see that when the value added to production that is due to
labor increases ( becomes larger), optimal IGO membership is smaller for
all levels of r.
To summarize all of the ndings, we see that in general, optimal IGO
membership for the representative member of the elite in sector X is initially
increasing in r until r is approximately 0.015 whereupon IGO membership
is decreasing in r. Furthermore, we have seen that for more productive
countries (countries with a higher A), optimal IGO membership is generally
higher as is the case for higher values of the price p
x
at which good X is
exported. In contrast, IGO membership is lower if sector X is larger than
sector Y and if the share of value added that goes to workers increases.
As a next step, we look at exactly the same simulation settings for the other
three scenarios (representative member of the elite in sector Y, representa-
tive voter (worker) in sector X and Y) in order to compare these results with
those just obtained.
190
CHAPTER 7. FORMAL MODEL ON IGO MEMBERSHIP AND
ENVIRONMENTAL QUALITY
(
a
)
p
r
i
c
e
p
=
0
.
9
(
b
)
p
r
i
c
e
p
=
0
.
9
5
(
c
)
p
r
i
c
e
p
=
1
(
d
)
p
r
i
c
e
p
=
1
.
0
5
(
e
)
p
r
i
c
e
p
=
1
.
1
F
i
g
u
r
e
7
.
3
:
O
p
t
i
m
a
l
I
G
O
m
e
m
b
e
r
s
h
i
p
w
i
t
h
v
a
r
y
i
n
g
p
r
i
c
e
o
f
e
x
p
o
r
t
g
o
o
d
p
x
191
CHAPTER 7. FORMAL MODEL ON IGO MEMBERSHIP AND
ENVIRONMENTAL QUALITY
(
a
)

=
0
.
5
(
b
)

=
0
.
6
(
c
)

=
0
.
7
(
d
)

=
0
.
8
(
e
)

=
0
.
9
F
i
g
u
r
e
7
.
4
:
O
p
t
i
m
a
l
I
G
O
m
e
m
b
e
r
s
h
i
p
w
i
t
h
v
a
r
y
i
n
g
s
h
a
r
e
o
f
v
a
l
u
e
a
d
d
e
d
t
h
a
t
g
o
e
s
t
o
w
o
r
k
e
r
s

192
CHAPTER 7. FORMAL MODEL ON IGO MEMBERSHIP AND
ENVIRONMENTAL QUALITY
7.2.2 Government Favors Elite Member in the Domestic Sec-
tor
If we imagine a country in which the government embraces the position
of the representative elite member in sector Y only, the government would
maximize the return to capital in sector Y - this is what capitalists in sector
Y earn - with respect to m in order to determine the optimal degree of
IGO membership for this country. The marginal rent on capital in sector Y
follows from maximizing equation (7.10) with respect to K:
MRC
y
=
Y
K
y
= (1 )A
y
_
1 +
_
A
y
p
x
(m)A
x
(m)
_ 1
1
K
x
K
y
_

_
L
K
y
_

+ A
y
_
1 +
_
A
y
p
x
(m)A
x
(m)
_ 1
1
K
x
K
y
_
1
L

K
1
y
_
A
y
p
x
(m)A
x
(m)
_ 1
1
K
x
K
2
y
Using the functional forms we specied for both price and technology
equations 7.11 and 7.12 we obtain
MRC
y
= A
y
_
L
K
y
_

_
_
1 +
_
A
y
p
x
(1
1
m
)(A
x
mr)
_ 1
1
K
x
K
y
_
_

_
(1 ) +
K
x
K
y
_
_
1 +
_
A
y
p
x
(1
1
m
)(A
x
mr)
_ 1
1
K
x
K
y
_
_
1
_
A
y
p
x
(1
1
m
)(A
x
mr)
_ 1
1
_
_
Now the government makes its choice regarding the optimal level of IGO
membership:
193
CHAPTER 7. FORMAL MODEL ON IGO MEMBERSHIP AND
ENVIRONMENTAL QUALITY

M
R
C
y

m
=
A
y
_
L
K
y
_

_ _
(
1

)
+

K
x
K
y
_ _
1
+
_
A
y
p
x
(
1

1
m
)
(
A
x

m
r
)
_
1

1
K
x
K
y
_ _

1
_
A
y
p
x
(
1

1
m
)
(
A
x

m
r
)
_
1

1
_ _
(

)
_ _
1
+
_
A
y
p
x
(
1

1
m
)
(
A
x

m
r
)
_
1

1
K
x
K
y
_ _

1
K
x
K
y
1

1
_
A
y
p
x
(
1

1
m
)
(
A
x

m
r
)
_
1

1
_

p
x
(
m
)

m
(
p
x
(
1

1
m
)
)
2
A
y
(
A
x

m
r
)

A
x
(
m
)

m
(
A
x

m
r
)
2
A
y
p
x
(
1

1
m
)
_
+
A
y
_
L
K
y
_

_ _
1
+
_
A
y
p
x
(
1

1
m
)
(
A
x

m
r
)
_
1

1
K
x
K
y
_ _

_ _ _

K
x
K
y
_ _
1
+
_
A
y
p
x
(
1

1
m
)
(
A
x

m
r
)
_
1

1
K
x
K
y
_ _

2
K
x
K
y
1

1
_
A
y
p
x
(
1

1
m
)
(
A
x

m
r
)
_
1

1
_

p
x
(
m
)

m
(
p
x
(
1

1
m
)
)
2
A
y
(
A
x

m
r
)

A
x
(
m
)

m
(
A
x

m
r
)
2
A
y
p
x
(
1

1
m
)
_
_
A
y
p
x
(
1

1
m
)
(
A
x

m
r
)
_
1

1
+

K
x
K
y
_ _
1
+
_
A
y
p
x
(
1

1
m
)
(
A
x

m
r
)
_
1

1
K
x
K
y
_ _

1
1

1
_
A
y
p
x
(
1

1
m
)
(
A
x

m
r
)
_
(
1

1
)
_

p
x
(
m
)

m
(
p
x
(
1

1
m
)
)
2
A
y
(
A
x

m
r
)

A
x
(
m
)

m
(
A
x

m
r
)
2
A
y
p
x
(
1

1
m
)
_
_ _ _
!
=
0
194
CHAPTER 7. FORMAL MODEL ON IGO MEMBERSHIP AND
ENVIRONMENTAL QUALITY
After rearranging, this simplies to
K
x
K
y
_
A
y
p
x
(1
1
m
)(A
x
mr)
_ 1
1
+
_
A
y
p
x
(1
1
m
)(A
x
mr)
_ 1
1

K
x
K
y

_
_
1 +
_
A
y
p
x
(1
1
m
)(A
x
mr)
_ 1
1
K
x
K
y
_
_
= 0
Using the same set-up as in the previous section, I simulate optimal IGO
membership allowing the rate r at which IGO membership renders the pro-
duction technology more stringent to vary from 0 to 0.1 at various values of
the other variables in the model.
When looking at the baseline setting in gure 7.5(a), one major dier-
ence to the above setting becomes apparent: for all settings, optimal IGO
membership is much smaller if the government favors the elite member in
the domestic sector instead of the elite member in the export sector. For
example, in the specic case of A = 1 (as displayed in gure 7.5(a)), optimal
IGO membership, with m = 8, is highest at a rate of r = 0.02. In contrast,
in the scenario before when the government was composed of the elite mem-
ber in the exporting sector, optimal IGO membership, with m = 30 at a rate
of r = 0.02, was much higher compared to this setting. Hence optimal IGO
membership diers by sector with the exporting sector favoring generally
higher levels of IGO membership than the domestic sector.
Interestingly, if we vary the level of productivity A we see that the curve
becomes less steep. Unlike the setting before, though, increasing A does not
lead to a rise in the maximal demanded IGO membership (gures 7.5(a)
to 7.5(c)). However, as in the case above, if sector Y is larger than sec-
tor X optimal IGO membership increases (gures 7.6(a) to 7.6(e)). This
is a somewhat peculiar nding: in general, we observe that a government
embracing the position of the elite member in the domestic sector would
demand less IGO membership than a government embracing the position of
the elite member in the export sector. However, the larger this domestic
sector (sector Y) in contrast to the export sector (sector X), the higher the
optimal number of IGO memberships (independent of whether the govern-
ment favors the export or the import sector). Hence whereas a government
composed of the elite in the domestic sector prefers an overall lower level of
IGO membership than a government composed of the elite in the exporting
sector, a larger domestic sector leads to a higher optimal IGO membership.
In contrast to the setting before, if we look at changes in the price p
x
of
the export good (gures 7.7(a) to 7.7(e)) we see that this time, an increase in
the price p
x
leads to a lower optimal number of IGO membership demanded
by the government. Concerning the simulation results for various settings
of we again see the same picture as above (gures 7.8(a) to 7.8(e)). The
195
CHAPTER 7. FORMAL MODEL ON IGO MEMBERSHIP AND
ENVIRONMENTAL QUALITY
higher the share of value added that goes to workers the lower the opti-
mal number of IGO membership demanded by a government embracing the
position of the elite member in the export sector.
Comparing these two settings, which both correspond to a non-democratic
form of government since the government considers the position of the elite
only, we see similar results concerning variations in the ratio between the
export and the domestic sector and concerning variations in the value of .
In contrast, the results dier with respect to the level of technology A and
with respect to an increase in the price of the export good p
x
. Whereas
the level of technology A in the domestic sector does not aect the height
but only the steepness of the curve, which relates r to IGO membership, in
the exporting sector A increases the height of the curve but not its steep-
ness. Concerning the price of the export good, an increase in p
x
aects IGO
membership negatively in the domestic sector but positively in the export
sector.
The major dierence is, however, that optimal IGO membership is much
smaller if the representative decision-maker is an elite member in the domes-
tic sector than if the representative decision-maker is an elite member in the
export sector. Hence if we imagined a non-democratic country consisting
of the elite of both sectors, government would need to compromise between
the two positions, leading to less IGO membership than demanded by the
export sector but to more IGO membership than demanded by the domestic
sector. This in turn implies that environmental stringency, which is a pos-
itive function of IGO membership, is higher if a non-democracy favors the
position of the elite member in the export sector than if a non-democracy
favors the position of the elite member in the domestic sector.
196
CHAPTER 7. FORMAL MODEL ON IGO MEMBERSHIP AND
ENVIRONMENTAL QUALITY
0.02 0.04 0.06 0.08 0.1
3
3.5
4
4.5
5
5.5
6
6.5
7
7.5
8
I
G
O

m
e
m
b
e
r
s
h
i
p
rate r
(a) A=1
0.02 0.04 0.06 0.08 0.1
4
4.5
5
5.5
6
6.5
7
7.5
8
8.5
I
G
O

m
e
m
b
e
r
s
h
i
p
rate r
(b) A=2
0.02 0.04 0.06 0.08 0.1
4
4.5
5
5.5
6
6.5
7
7.5
8
8.5
I
G
O

m
e
m
b
e
r
s
h
i
p
rate r
(c) A=3
Figure 7.5: Optimal IGO membership at dierent levels of technology A
197
CHAPTER 7. FORMAL MODEL ON IGO MEMBERSHIP AND
ENVIRONMENTAL QUALITY
0
.
0
2
0
.
0
4
0
.
0
6
0
.
0
8
0
.
1
3
3
.
5 4
4
.
5 5
5
.
5 6
I G O m e m b e r s h i p
r
a
t
e

r
(
a
)
K
y
K
x
=
0
.
8
0
.
0
2
0
.
0
4
0
.
0
6
0
.
0
8
0
.
1
3
3
.
5 4
4
.
5 5
5
.
5 6
6
.
5
I G O m e m b e r s h i p
r
a
t
e

r
(
b
)
K
y
K
x
=
0
.
9
0
.
0
2
0
.
0
4
0
.
0
6
0
.
0
8
0
.
1
3
3
.
5 4
4
.
5 5
5
.
5 6
6
.
5 7
7
.
5
I G O m e m b e r s h i p
r
a
t
e

r
(
c
)
K
y
K
x
=
1
0
.
0
2
0
.
0
4
0
.
0
6
0
.
0
8
0
.
1
3
3
.
5 4
4
.
5 5
5
.
5 6
6
.
5 7
7
.
5 8
I G O m e m b e r s h i p
r
a
t
e

r
(
d
)
K
y
K
x
=
1
.
1
0
.
0
2
0
.
0
4
0
.
0
6
0
.
0
8
0
.
1
3 4 5 6 7 8 9
1
0
I G O m e m b e r s h i p
r
a
t
e

r
(
e
)
K
y
K
x
=
1
.
2
F
i
g
u
r
e
7
.
6
:
O
p
t
i
m
a
l
I
G
O
m
e
m
b
e
r
s
h
i
p
a
t
d
i

e
r
e
n
t
l
e
v
e
l
s
o
f
K
x
K
y
198
CHAPTER 7. FORMAL MODEL ON IGO MEMBERSHIP AND
ENVIRONMENTAL QUALITY
0
.
0
2
0
.
0
4
0
.
0
6
0
.
0
8
0
.
1
0 5
1
0
1
5
2
0
2
5
3
0
I G O m e m b e r s h i p
r
a
t
e

r
(
a
)
p
r
i
c
e
p
=
0
.
9
0
.
0
2
0
.
0
4
0
.
0
6
0
.
0
8
0
.
1
2 4 6 8
1
0
1
2
1
4
1
6
I G O m e m b e r s h i p
r
a
t
e

r
(
b
)
p
r
i
c
e
p
=
0
.
9
5
0
.
0
2
0
.
0
4
0
.
0
6
0
.
0
8
0
.
1
3 4 5 6 7 8 9
1
0
1
1
1
2
I G O m e m b e r s h i p
r
a
t
e

r
(
c
)
p
r
i
c
e
p
=
1
0
.
0
2
0
.
0
4
0
.
0
6
0
.
0
8
0
.
1
3 4 5 6 7 8 9
1
0
I G O m e m b e r s h i p
r
a
t
e

r
(
d
)
p
r
i
c
e
p
=
1
.
0
5
0
.
0
2
0
.
0
4
0
.
0
6
0
.
0
8
0
.
1
3
3
.
5 4
4
.
5 5
5
.
5 6
6
.
5 7
7
.
5 8
I G O m e m b e r s h i p
r
a
t
e

r
(
e
)
p
r
i
c
e
p
=
1
.
1
F
i
g
u
r
e
7
.
7
:
O
p
t
i
m
a
l
I
G
O
m
e
m
b
e
r
s
h
i
p
w
i
t
h
v
a
r
y
i
n
g
p
r
i
c
e
o
f
e
x
p
o
r
t
g
o
o
d
p
x
199
CHAPTER 7. FORMAL MODEL ON IGO MEMBERSHIP AND
ENVIRONMENTAL QUALITY
0
.
0
2
0
.
0
4
0
.
0
6
0
.
0
8
0
.
1
3
3
.
5 4
4
.
5 5
5
.
5
I G O m e m b e r s h i p
r
a
t
e

r
(
a
)

=
0
.
5
0
.
0
2
0
.
0
4
0
.
0
6
0
.
0
8
0
.
1
3
3
.
5 4
4
.
5 5
5
.
5 6
6
.
5 7
7
.
5 8
I G O m e m b e r s h i p
r
a
t
e

r
(
b
)

=
0
.
6
0
.
0
2
0
.
0
4
0
.
0
6
0
.
0
8
0
.
1
3 4 5 6 7 8 9
1
0
1
1
1
2
I G O m e m b e r s h i p
r
a
t
e

r
(
c
)

=
0
.
7
0
.
0
2
0
.
0
4
0
.
0
6
0
.
0
8
0
.
1
2 4 6 8
1
0
1
2
1
4
1
6
I G O m e m b e r s h i p
r
a
t
e

r
(
d
)

=
0
.
8
0
.
0
2
0
.
0
4
0
.
0
6
0
.
0
8
0
.
1
2 4 6 8
1
0
1
2
1
4
1
6
1
8
2
0
I G O m e m b e r s h i p
r
a
t
e

r
(
e
)

=
0
.
9
F
i
g
u
r
e
7
.
8
:
O
p
t
i
m
a
l
I
G
O
m
e
m
b
e
r
s
h
i
p
w
i
t
h
v
a
r
y
i
n
g
s
h
a
r
e
o
f
v
a
l
u
e
a
d
d
e
d
t
h
a
t
g
o
e
s
t
o
w
o
r
k
e
r
s

200
CHAPTER 7. FORMAL MODEL ON IGO MEMBERSHIP AND
ENVIRONMENTAL QUALITY
7.2.3 Government Favors Representative Voter (Worker) in
the Export Sector
After having described the optimal positions towards IGO membership and
thus towards environmental stringency when considering a non-democratic
form of government, we now turn to the two democratic settings. Although
labor is mobile across the two sectors and thus the wage is the same in both
sectors, IGO membership aects both sectors dierently. This implies that
before a government joins any IGO, workers in the two sectors might have
dierent positions towards the optimal degree of IGO membership, which
makes it necessary to look at both sectors separately.
First we imagine a country in which the government embraces the posi-
tion of the representative voter (worker) in the export sector only. In this
scenario, the government would maximize the average workers wage, which
corresponds to the marginal product of labor, with respect to m in order
to determine the optimal degree of IGO membership for this country. Since
we want to have the wage as an expression of total labor endowment in
the economy, we maximize equation (7.9) with respect to L to obtain the
average wage in sector X:
MPL
x
= w
x
=
X
L
= A
x
(m)
_
1 +
_
p
x
(m)
A
x
(m)
A
y
_ 1
1
K
y
K
x
_

_
K
x
L
_
1
Using the functional forms we specied for both price and technology
equations 7.11 and 7.12 we obtain
w
x
= (A
x
mr)
_
1 +
_
p
x
(1
1
m
)
(A
x
mr)
A
y
_ 1
1
K
y
K
x
_

_
K
x
L
_
1
Now the government makes its choice regarding the optimal level of IGO
membership:
MPL
x
m
=
_
K
x
L
_
1
A
x
(m)
m
_
1 +
_
p
x
(1
1
m
)
(A
x
mr)
A
y
_ 1
1
K
y
K
x
_

+
_
K
x
L
_
1
(A
x
mr)()
_
1 +
_
p
x
(1
1
m
)
(A
x
mr)
A
y
_ 1
1
K
y
K
x
_
1
1
1
K
y
K
x
_
p
x
m
(A
x
mr)
A
y
+
A
x
m
p
x
(1
1
m
)
A
y
_
_
p
x
(1
1
m
)
(A
x
mr)
A
y
_ 1
1
1
!
= 0
201
CHAPTER 7. FORMAL MODEL ON IGO MEMBERSHIP AND
ENVIRONMENTAL QUALITY
This simplies to
(A
x
mr)

1
K
y
K
x
_
p
x
(1
1
m
)
(A
x
mr)
A
y
_ 1
1
1
_
1 +
_
p
x
(1
1
m
)
(A
x
mr)
A
y
_ 1
1
K
y
K
x
_
1
_
p
x
m
2
(A
x
mr)
A
y

r
A
y
p
x
(1
1
m
)
_
= r (7.15)
Again, equation 7.15 cannot be solved analytically, which means that we use
the same simulation set-up as in the sections before to simulate optimal IGO
membership while varying the rate r at which IGO membership renders the
production technology more stringent from 0 to 0.1 at various values of the
other variables in the model.
Figure 7.9(a) shows the simulation for the baseline setting in which the
level of technology A is 1, the ratio between the domestic and the exporting
sector
K
y
K
x
is 1.1, the price of good X is p
x
= 1.1 and the share of value
added that goes to workers is 0.6. Similar to the setting in which the
government embraces the position of the representative elite member in the
export sector, we see that optimal IGO membership is rather high at low
levels of r and then decreases at an also decreasing rate. The only dierence
to the setting above is that the initial increase in IGO membership is missing
in the scenario where the government consists of the representative voter in
the export sector.
In addition, we also see that IGO membership is again higher in countries
that are more productive, i.e. that have a higher level of A, as can be seen in
gures 7.9(a) to 7.9(c). Furthermore, as before, IGO membership is higher if
the domestic sector Y is larger than the export sector X gures 7.10(a) to
7.10(e) and if the price p
x
of the export good increases gures 7.11(a) to
7.11(e). Hence the positions concerning optimal IGO membership are very
similar whether we consider a government representing the elite in the export
sector or a government consisting of the optimal worker in the export sector.
Counterintuitively, however, the optimal IGO membership demanded by the
representative elite in sector X is higher than the optimal IGO membership
demanded by the representative worker.
The only dierence to the non-democratic scenarios we considered before
is that optimal IGO membership is now increasing in (gures 7.12(a) to
7.12(e)). Thus the higher the share of value added that goes to workers
the higher is the optimal level of IGO membership demanded by the rep-
resentative worker. This is an interesting nding since higher values of
are equivalent to higher wages for the workers (this follows from equations
7.5 and 7.6, which describe the wage in the two sectors). Consequently, this
202
CHAPTER 7. FORMAL MODEL ON IGO MEMBERSHIP AND
ENVIRONMENTAL QUALITY
implies that countries in which more of the production arises from labor and
thus in which the workers earn higher wages would like to join more inter-
national organizations and in turn accept more environmentally stringent
technology.
203
CHAPTER 7. FORMAL MODEL ON IGO MEMBERSHIP AND
ENVIRONMENTAL QUALITY
0.02 0.04 0.06 0.08 0.1
0
5
10
15
20
25
I
G
O

m
e
m
b
e
r
s
h
i
p
rate r
(a) A=1
0.02 0.04 0.06 0.08 0.1
0
5
10
15
20
25
I
G
O

m
e
m
b
e
r
s
h
i
p
rate r
(b) A=2
0.02 0.04 0.06 0.08 0.1
0
5
10
15
20
25
30
I
G
O

m
e
m
b
e
r
s
h
i
p
rate r
(c) A=3
Figure 7.9: Optimal IGO membership at dierent levels of technology A
204
CHAPTER 7. FORMAL MODEL ON IGO MEMBERSHIP AND
ENVIRONMENTAL QUALITY
0
.
0
2
0
.
0
4
0
.
0
6
0
.
0
8
0
.
1
2 4 6 8
1
0
1
2
1
4
1
6
1
8
I G O m e m b e r s h i p
r
a
t
e

r
(
a
)
K
y
K
x
=
0
.
8
0
.
0
2
0
.
0
4
0
.
0
6
0
.
0
8
0
.
1
2 4 6 8
1
0
1
2
1
4
1
6
1
8
I G O m e m b e r s h i p
r
a
t
e

r
(
b
)
K
y
K
x
=
0
.
9
0
.
0
2
0
.
0
4
0
.
0
6
0
.
0
8
0
.
1
2 4 6 8
1
0
1
2
1
4
1
6
1
8
I G O m e m b e r s h i p
r
a
t
e

r
(
c
)
K
y
K
x
=
1
0
.
0
2
0
.
0
4
0
.
0
6
0
.
0
8
0
.
1
2 4 6 8
1
0
1
2
1
4
1
6
1
8
I G O m e m b e r s h i p
r
a
t
e

r
(
d
)
K
y
K
x
=
1
.
1
0
.
0
2
0
.
0
4
0
.
0
6
0
.
0
8
0
.
1
2 4 6 8
1
0
1
2
1
4
1
6
1
8
I G O m e m b e r s h i p
r
a
t
e

r
(
e
)
K
y
K
x
=
1
.
2
F
i
g
u
r
e
7
.
1
0
:
O
p
t
i
m
a
l
I
G
O
m
e
m
b
e
r
s
h
i
p
a
t
d
i

e
r
e
n
t
l
e
v
e
l
s
o
f
K
y
K
x
205
CHAPTER 7. FORMAL MODEL ON IGO MEMBERSHIP AND
ENVIRONMENTAL QUALITY
0
.
0
2
0
.
0
4
0
.
0
6
0
.
0
8
0
.
1
2 4 6 8
1
0
1
2
1
4
1
6
1
8
2
0
I G O m e m b e r s h i p
r
a
t
e

r
(
a
)
p
r
i
c
e
p
=
0
.
9
0
.
0
2
0
.
0
4
0
.
0
6
0
.
0
8
0
.
1
2 4 6 8
1
0
1
2
1
4
1
6
1
8
2
0
I G O m e m b e r s h i p
r
a
t
e

r
(
b
)
p
r
i
c
e
p
=
0
.
9
5
0
.
0
2
0
.
0
4
0
.
0
6
0
.
0
8
0
.
1
2 4 6 8
1
0
1
2
1
4
1
6
1
8
I G O m e m b e r s h i p
r
a
t
e

r
(
c
)
p
r
i
c
e
p
=
1
0
.
0
2
0
.
0
4
0
.
0
6
0
.
0
8
0
.
1
2 4 6 8
1
0
1
2
1
4
1
6
1
8
I G O m e m b e r s h i p
r
a
t
e

r
(
d
)
p
r
i
c
e
p
=
1
.
0
5
0
.
0
2
0
.
0
4
0
.
0
6
0
.
0
8
0
.
1
2 4 6 8
1
0
1
2
1
4
1
6
1
8
I G O m e m b e r s h i p
r
a
t
e

r
(
e
)
p
r
i
c
e
p
=
1
.
1
F
i
g
u
r
e
7
.
1
1
:
O
p
t
i
m
a
l
I
G
O
m
e
m
b
e
r
s
h
i
p
w
i
t
h
v
a
r
y
i
n
g
p
r
i
c
e
o
f
e
x
p
o
r
t
g
o
o
d
p
x
206
CHAPTER 7. FORMAL MODEL ON IGO MEMBERSHIP AND
ENVIRONMENTAL QUALITY
0
.
0
2
0
.
0
4
0
.
0
6
0
.
0
8
0
.
1
2 4 6 8
1
0
1
2
1
4
1
6
I G O m e m b e r s h i p
r
a
t
e

r
(
a
)

=
0
.
5
0
.
0
2
0
.
0
4
0
.
0
6
0
.
0
8
0
.
1
2 4 6 8
1
0
1
2
1
4
1
6
1
8
I G O m e m b e r s h i p
r
a
t
e

r
(
b
)

=
0
.
6
0
.
0
2
0
.
0
4
0
.
0
6
0
.
0
8
0
.
1
2 4 6 8
1
0
1
2
1
4
1
6
1
8
2
0
2
2
I G O m e m b e r s h i p
r
a
t
e

r
(
c
)

=
0
.
7
0
.
0
2
0
.
0
4
0
.
0
6
0
.
0
8
0
.
1
0 5
1
0
1
5
2
0
2
5
3
0
I G O m e m b e r s h i p
r
a
t
e

r
(
d
)

=
0
.
8
0
.
0
2
0
.
0
4
0
.
0
6
0
.
0
8
0
.
1
0 5
1
0
1
5
2
0
2
5
3
0
3
5
I G O m e m b e r s h i p
r
a
t
e

r
(
e
)

=
0
.
9
F
i
g
u
r
e
7
.
1
2
:
O
p
t
i
m
a
l
I
G
O
m
e
m
b
e
r
s
h
i
p
w
i
t
h
v
a
r
y
i
n
g
s
h
a
r
e
o
f
v
a
l
u
e
a
d
d
e
d
t
h
a
t
g
o
e
s
t
o
w
o
r
k
e
r
s

207
CHAPTER 7. FORMAL MODEL ON IGO MEMBERSHIP AND
ENVIRONMENTAL QUALITY
7.2.4 Government Favors Representative Voter (Worker) in
the Domestic Sector
The nal scenario we consider is the case when the government consists of
the representative voter (worker) in the domestic sector. As before, this kind
of government would maximize the average wage of the worker in sector Y
to obtain the optimal level of IGO membership. To derive the average wage
in sector Y as an expression of total labor endowment in the economy, we
maximize equation (7.10) with respect to L:
MPL
y
= w
y
=
Y
L
= A
y
_
1 +
_
A
y
A
x
p
x
_ 1
1
K
x
K
y
_

_
K
y
L
_
1
Using the functional forms we specied for both price and technology
equations 7.11 and 7.12 we obtain
w
y
= A
y
_
_
1 +
_
A
y
p
x
(1
1
m
)(A
x
mr)
_ 1
1
K
x
K
y
_
_

_
K
y
L
_
1
Now the government makes its choice regarding the optimal level of IGO
membership:
MPL
y
m
= A
y
_
_
1 +
_
A
y
p
x
(1
1
m
)(A
x
mr)
_ 1
1
K
x
K
y
_
_
1
_
K
y
L
_
1
()
1
1
_
A
y
p
x
(1
1
m
)(A
x
mr)
_ 1
1
1
K
x
K
y
_

p
x
(m)
m
(p
x
(1
1
m
))
2
A
y
(A
x
mr)

A
x
(m)
m
(A
x
mr)
2
A
y
p
x
(1
1
m
)
_
!
= 0
This simplies to
_
_
1 +
_
A
y
p
x
(1
1
m
)(A
x
mr)
_ 1
1
K
x
K
y
_
_
1
_
A
y
p
x
(1
1
m
)(A
x
mr)
_ 1
1
_

1
m
2
(1
1
m
)
+
r
(A
x
mr)
_
= 0
Using, again, the same simulation set-up as before, we obtain the optimal
level of IGO membership for various values of the other variables while vary-
ing the rate r at which IGO membership renders the production technology
208
CHAPTER 7. FORMAL MODEL ON IGO MEMBERSHIP AND
ENVIRONMENTAL QUALITY
more stringent from 0 to 0.1. The only dierence to the above setting is
that in this baseline simulation we set
K
x
K
y
= 1.1 instead of
K
y
K
x
= 1.1.
The picture we get from this simulation (gure 7.13(a)) is very similar
to the results for the workers in the export sector (gure 7.9(a)). Although
the curve is less smooth than for the workers in the export sector, opti-
mal IGO membership is also decreasing at a decreasing rate when the rate
r at which IGO membership renders the production technology more strin-
gent increases. Furthermore, although optimal IGO membership is generally
slightly lower for the workers in the domestic sector, the results do not dier
much from those for the workers in the export sector. Remarkably, the dif-
ference in optimal IGO membership is greater between the two autocratic
scenarios (at a rate r at which IGO membership renders the production
technology more stringent of 0.01 optimal IGO membership is about 25 for
the exporting sector compared to about 5 for the domestic sector) than be-
tween the two democratic scenarios (at r = 0.01 optimal IGO membership
is about 12 for the exporting sector compared to 10 for the domestic sector).
Interestingly, as in the case of the elite in the domestic sector, countries
that are more productive, i.e. that have a higher value of A, have a less steep
curve, which links IGO membership to r, without an increase in the overall
optimum of IGO membership see gures 7.13(a) to 7.13(b). Concerning
both the price p
x
of the export good and the ratio between the two sectors
we do not see noticeable changes in IGO membership while varying these
variables see gures 7.14(a) to 7.15(e). The same holds true for the share
of value added that goes to workers . As gures 7.16(a) to 7.16(d)
6
show,
the relationship between the rate r at which IGO membership renders the
production technology more stringent and optimal IGO membership does
not noticeably change for dierent values of .
Comparing the two scenarios that correspond to a democratic type of
government, we rst see that overall, workers in the domestic sector fa-
vor IGO membership (and with it less environmental stringent production)
slightly less than do workers in the export sector. Second, workers in the
domestic sector react less to changes in the economic conditions such as vari-
ations in the price of the export good than their counterparts in the export
sector. Thus if we imagine a government representing both workers from the
export and the import sector, IGO membership would be lower than if we
only considered the workers in the export sector but higher compared to the
scenario if we only considered the workers in the domestic sector. Overall,
however, the dierences between the two scenarios in which the government
represents the workers in the two sectors are not as large as between the
scenarios in which the government represents the representative elites in the
two sectors.
6
The simulation results for = 0.9 do not converge properly and are therefore not
shown here.
209
CHAPTER 7. FORMAL MODEL ON IGO MEMBERSHIP AND
ENVIRONMENTAL QUALITY
0.02 0.04 0.06 0.08 0.1
3
4
5
6
7
8
9
10
11
12
I
G
O

m
e
m
b
e
r
s
h
i
p
rate r
(a) A=1
0.02 0.04 0.06 0.08 0.1
4
5
6
7
8
9
10
11
12
I
G
O

m
e
m
b
e
r
s
h
i
p
rate r
(b) A=2
0.02 0.04 0.06 0.08 0.1
5
6
7
8
9
10
11
12
I
G
O

m
e
m
b
e
r
s
h
i
p
rate r
(c) A=3
Figure 7.13: Optimal IGO membership at dierent levels of technology A
210
CHAPTER 7. FORMAL MODEL ON IGO MEMBERSHIP AND
ENVIRONMENTAL QUALITY
0
.
0
2
0
.
0
4
0
.
0
6
0
.
0
8
0
.
1
3 4 5 6 7 8 9
1
0
1
1
1
2
1
3
I G O m e m b e r s h i p
r
a
t
e

r
(
a
)
K
x
K
y
=
0
.
8
0
.
0
2
0
.
0
4
0
.
0
6
0
.
0
8
0
.
1
3 4 5 6 7 8 9
1
0
1
1
1
2
1
3
I G O m e m b e r s h i p
r
a
t
e

r
(
b
)
K
x
K
y
=
0
.
9
0
.
0
2
0
.
0
4
0
.
0
6
0
.
0
8
0
.
1
3 4 5 6 7 8 9
1
0
1
1
1
2
1
3
I G O m e m b e r s h i p
r
a
t
e

r
(
c
)
K
x
K
y
=
1
0
.
0
2
0
.
0
4
0
.
0
6
0
.
0
8
0
.
1
3 4 5 6 7 8 9
1
0
1
1
1
2
I G O m e m b e r s h i p
r
a
t
e

r
(
d
)
K
x
K
y
=
1
.
1
0
.
0
2
0
.
0
4
0
.
0
6
0
.
0
8
0
.
1
3 4 5 6 7 8 9
1
0
1
1
1
2
I G O m e m b e r s h i p
r
a
t
e

r
(
e
)
K
x
K
y
=
1
.
2
F
i
g
u
r
e
7
.
1
4
:
O
p
t
i
m
a
l
I
G
O
m
e
m
b
e
r
s
h
i
p
a
t
d
i

e
r
e
n
t
l
e
v
e
l
s
o
f
K
y
K
x
211
CHAPTER 7. FORMAL MODEL ON IGO MEMBERSHIP AND
ENVIRONMENTAL QUALITY
0
.
0
2
0
.
0
4
0
.
0
6
0
.
0
8
0
.
1
3 4 5 6 7 8 9
1
0
1
1
1
2
1
3
I G O m e m b e r s h i p
r
a
t
e

r
(
a
)
p
r
i
c
e
p
=
0
.
9
0
.
0
2
0
.
0
4
0
.
0
6
0
.
0
8
0
.
1
3 4 5 6 7 8 9
1
0
1
1
1
2
I G O m e m b e r s h i p
r
a
t
e

r
(
b
)
p
r
i
c
e
p
=
0
.
9
5
0
.
0
2
0
.
0
4
0
.
0
6
0
.
0
8
0
.
1
3 4 5 6 7 8 9
1
0
1
1
1
2
I G O m e m b e r s h i p
r
a
t
e

r
(
c
)
p
r
i
c
e
p
=
1
0
.
0
2
0
.
0
4
0
.
0
6
0
.
0
8
0
.
1
3 4 5 6 7 8 9
1
0
1
1
1
2
I G O m e m b e r s h i p
r
a
t
e

r
(
d
)
p
r
i
c
e
p
=
1
.
0
5
0
.
0
2
0
.
0
4
0
.
0
6
0
.
0
8
0
.
1
3 4 5 6 7 8 9
1
0
1
1
1
2
I G O m e m b e r s h i p
r
a
t
e

r
(
e
)
p
r
i
c
e
p
=
1
.
1
F
i
g
u
r
e
7
.
1
5
:
O
p
t
i
m
a
l
I
G
O
m
e
m
b
e
r
s
h
i
p
w
i
t
h
v
a
r
y
i
n
g
p
r
i
c
e
o
f
e
x
p
o
r
t
g
o
o
d
p
x
212
CHAPTER 7. FORMAL MODEL ON IGO MEMBERSHIP AND
ENVIRONMENTAL QUALITY
0
.
0
2
0
.
0
4
0
.
0
6
0
.
0
8
0
.
1
3 4 5 6 7 8 9
1
0
1
1
1
2
1
3
I G O m e m b e r s h i p
r
a
t
e

r
(
a
)

=
0
.
5
0
.
0
2
0
.
0
4
0
.
0
6
0
.
0
8
0
.
1
3 4 5 6 7 8 9
1
0
1
1
1
2
I G O m e m b e r s h i p
r
a
t
e

r
(
b
)

=
0
.
6
0
.
0
2
0
.
0
4
0
.
0
6
0
.
0
8
0
.
1
3 4 5 6 7 8 9
1
0
1
1
1
2
1
3
I G O m e m b e r s h i p
r
a
t
e

r
(
c
)

=
0
.
7
0
.
0
2
0
.
0
4
0
.
0
6
0
.
0
8
0
.
1
3 4 5 6 7 8 9
1
0
1
1
1
2
I G O m e m b e r s h i p
r
a
t
e

r
(
d
)

=
0
.
8
F
i
g
u
r
e
7
.
1
6
:
O
p
t
i
m
a
l
I
G
O
m
e
m
b
e
r
s
h
i
p
w
i
t
h
v
a
r
y
i
n
g
s
h
a
r
e
o
f
v
a
l
u
e
a
d
d
e
d
t
h
a
t
g
o
e
s
t
o
w
o
r
k
e
r
s

213
CHAPTER 7. FORMAL MODEL ON IGO MEMBERSHIP AND
ENVIRONMENTAL QUALITY
7.3 Conclusion
Overall, the simulations have shown that regardless of the type of govern-
ment, as long as the rate r at which IGO membership renders the production
technology more stringent is not very high, governments always opt for some
IGO membership and therefore for a more environmentally stringent produc-
tion technology. Since I decided at the beginning to rely on a sector-specic
trade model we see that in fact, the preferences concerning the optimal level
of IGO membership separate along sector lines. Both the representative elite
member and worker in the export sector opt for relatively high levels of IGO
membership and thus for more environmentally stringent production tech-
nology than their counterparts in the domestic sector. Interestingly, elites
in the export sector favor even higher levels of IGO membership than the
average worker in this sector.
Another interesting result of the simulations is that the higher the price
of the export good p
x
, the more benecial joining IGOs becomes for both the
representative elite member and the representative worker in the exporting
sector. Thus the higher the price at which they can sell their product, the
larger is the optimal level of IGO membership the government will demand
and consequently the more environmentally friendly the production of the
export good becomes. The opposite scenario is true if the government repre-
sents the capitalists in the domestic sector. Then higher prices of the export
good are associated with a lower level of demanded IGO membership.
Finally, we observe from the simulations of the model that an increase in
the value added to production that is due to labor () implies a decrease in
optimal IGO membership if the government favors either the capitalists in
the domestic or the exporting sector but implies an increase in optimal IGO
membership if the government favors the workers in the exporting sector
7
.
Interestingly, higher values of , which characterize countries in which more
of the production arises from labor, are associated with higher wages for the
workers
8
. Hence the results of the simulation imply that countries in which
the workers earn higher wages would like to join more international organi-
zations and in turn accept more environmentally stringent technology if the
form of the government is democratic but vice versa if the form of the gov-
ernment is autocratic. Hence is the only parameter for which the results
of the simulation do not dier according to sectoral lines (domestic versus
exporting sector) but according to factor lines (workers versus capitalists).
Potentially, this model can be enhanced in various aspects. First, an
interesting extension of the model would be to include consumer eects. So
far, I have abstracted from any potential adverse consumption eects that
might arise because IGO membership increases the price of the export good
7
For workers in the domestic sector a change in does not aect the simulation out-
comes.
8
This is due to the Cobb-Douglas production function underlying this model.
214
CHAPTER 7. FORMAL MODEL ON IGO MEMBERSHIP AND
ENVIRONMENTAL QUALITY
X. Doing so would require specifying a preference ordering of the representa-
tive agents over the various goods produced in this economy. Furthermore,
opting for this route would make it possible to include a preference for
cleaner production technologies and thus for better environmental quality
in this utility function of the agents. This would imply that consumers, as
do the producers in the current model, face a trade-o between cheaper but
more pollution intensive consumption and more expensive but less pollution
intensive consumption.
As these suggestions show, the model in this chapter builds a rst formal
framework showing how the eects of IGO membership can be analyzed in
a stylized setting that still allows for interesting extensions.
215
Chapter 8
Conclusion
Do poor countries need to become rich in order to be able to oer high
levels of environmental public goods to their population? According to
the theoretical and empirical literature on the relationship between income
and environmental degradation, the answer to this question is supposedly
yes. Many studies that analyze how growing income inuences a countrys
environmental record have shown that many forms of pollution tend to follow
an inverted U-shape, which became known as the Environmental Kuznets
Curve: at the beginning of a countrys economic development, pollution
increases along with national income until a certain threshold is reached
after which some types of pollution such as air and water pollution tend
to level o and decline (Grossman & Krueger 1995; Selden & Song 1994).
Hence developing countries, which are all situated on the upward sloping
part of this curve, should be confronted with increases in pollution until
they grow rich enough to reach the turning point after which environmental
pollution supposedly levels o again.
However, empirically, as gures 8.1(a) to 8.1(c) show, we see a huge vari-
ation in the environmental performance of developing countries independent
of their national income, which puts a major question mark over this con-
clusion. Consequently, this thesis aims to evaluate which factors can help us
better understand this variation in environmental quality that we observe
empirically.
The existing literature focuses mainly on the type of the political system
(democratic versus autocratic forms of government) and economic openness
(trade openness and foreign direct investment) to explain positive or negative
deviations from the EKC at any given level of income (Antweiler et al. 2001;
Neumayer 2002b; Ward 2008; Baettig & Bernauer 2009). This thesis adds
to that body of research by evaluating how international integration and its
interplay with the political system inuence the environmental performance
of developing countries. In particular, I argue that international integra-
tion, which is dened in this thesis as encompassing four dierent aspects
216
CHAPTER 8. CONCLUSION
-16
-
1
6
-16 -14
-
1
4
-14 -12
-
1
2
-12 -10
-
1
0
-10 -8
-
8
-8 6
6
67
7
78
8
89
9
910
10
10 ln GDP pc
ln GDP pc
ln GDP pc 95% CI
95% CI
95% CI Fitted values
Fitted values
Fitted values log of SO2 per capita
log of SO2 per capita
log of SO2 per capita
(a) SO
2
emissions
-4
-
4
-4 -2
-
2
-2 0
0
02
2
24
4
46
6
67
7
78
8
89
9
910
10
10 ln GDP pc
ln GDP pc
ln GDP pc 95% CI
95% CI
95% CI Fitted values
Fitted values
Fitted values log of CO2 per capita
log of CO2 per capita
log of CO2 per capita
(b) CO
2
emissions
-9
-
9
-9 -8
-
8
-8 -7
-
7
-7 -6
-
6
-6 -5
-
5
-5 -4
-
4
-4 6
6
67
7
78
8
89
9
910
10
10 ln GDP pc
ln GDP pc
ln GDP pc 95% CI
95% CI
95% CI Fitted values
Fitted values
Fitted values log of BOD per capita
log of BOD per capita
log of BOD per capita
(c) BOD
Figure 8.1: Environmental quality and GDP per capita (year 2000)
membership in international government organizations, international aid,
trade openness and foreign direct investment can increase the capacity and
willingness of developing countries to provide higher environmental quality
than could be expected given their current low level of national income.
Furthermore, a countrys political system is hypothesized to mediate this
proposed positive eect of international integration on environmental qual-
ity. Although I argue that democratic developing countries should in general
not be characterized by better environmental quality than their autocratic
counterparts, democratic political forms of government should nevertheless
intensify the positive eect of international integration. Hence I propose
to analyze the interaction eect between international integration and the
political system on a countrys environmental performance.
In doing so, this analysis diers in three important respects from the ex-
isting literature. First, by empirically focusing on developing countries only,
this thesis allows us to evaluate which of the dierent aspects of international
integration considered in this thesis (membership in IGOs, international aid,
trade openness and FDI) might help countries to cut through the EKC,
i.e. to reach the turning point at lower levels of environmental degrada-
tion, and/or earlier on in the stage of economic development. This focus
217
CHAPTER 8. CONCLUSION
on developing countries is important since up to now, most studies dealing
with the environmental performance of developing countries paint a rather
gloomy picture of the relationship between economic growth and environ-
mental pollution (Arrow et al. 1995). Second, in contrast to older studies
which theoretically and empirically concentrate on the economic facet of in-
ternational integration only, this thesis follows a broader understanding of
international integration by looking at four dierent aspects of international
integration. Thus the analysis overcomes the shortcoming of previous stud-
ies that have captured only a specic angle of globalization, leaving aside
the other facets of international integration. Third, I theoretically adopt
the various arguments put forward to analyze the relationship between a
countrys political system and its environmental performance to the spe-
cic context of developing countries. This leads me to argue that the type
of the political system should not have an independent eect on environ-
mental quality in developing countries, but its interplay with international
integration should be decisive for improved environmental quality. More
specically, I posit that the presumably positive eect of international inte-
gration should be intensied in democratic political systems, implying that
at any given level of international integration, environmental quality will be
better in democracies relative to autocracies.
The seven hypotheses derived from the theoretical framework are empir-
ically tested using time-series cross-section analysis of 115 developing coun-
tries from 1970 to 2000. To present a broad picture of environmental quality
while covering important aspects of a countrys environmental performance,
I rely on three dierent measures of environmental pollution, which are sul-
fur dioxide (SO
2
) and carbon dioxide (CO
2
) emissions measuring air quality,
and biological oxygen demand (BOD) measuring water quality. The results
of this aggregate statistical analysis suggest, as discussed in more detail be-
low, that only one facet of international integration, namely membership in
IGOs, is robustly associated to an increase in the environmental performance
of developing countries. This is a remarkable nding in at least two respects:
rst, the relationship between IGO membership and environmental quality
in developing countries has not been analyzed in this form to the best of
my knowledge before, making this nding novel to the literature. Second,
the robust inuence of IGO membership suggests that this more political
aspect of international integration is more important for the environmental
performance of developing countries than the economic aspects of interna-
tional integration (trade openness and FDI), which overall are not robustly
associated to environmental pollution.
The empirical analysis then goes on to interpret the potential mecha-
nisms behind this nding more closely. Using a more rened coding of IGO
membership, further time-series cross-country analyses are presented that
allow for a rst evaluation concerning the theoretically proposed mecha-
nisms (issue linkage, technology and know-how transfer and the diusion of
218
CHAPTER 8. CONCLUSION
the ideal of sustainable development) underlying the aggregate relationship
between IGO membership and environmental quality. These results are fur-
ther evaluated in a second step using four illustrative case studies. The aim
of these case studies is to evaluate and illustrate which of these theoretical
links between IGO membership and environmental quality are empirically
at play in the every-day business of IGOs in developing countries.
Finally, one mechanism behind this inuence of IGO membership, the
diusion of technology, is then taken up to develop a simple formal theoreti-
cal model relating IGO membership to environmental pollution. The aim of
this model is to obtain a more general theoretical framework for the study of
IGO membership and its links to environmental pollution that can be used
in future analyses.
Using this mixed methods approach suggests that international integra-
tion and its interplay with the political system do not lead to one single-
edged conclusion. Starting with the aggregate statistical analysis presented
in chapter 5, we see that IGO membership is the only one out of the four
factors of international integration that seems to be robustly related to a
decrease in pollution with respect to all three indicators. Trade openness,
in contrast, does not aect SO
2
emissions and water quality but increases
greenhouse gas emissions in the form of CO
2
. This last nding suggests that
in contrast to what might be expected theoretically, the positive eects of
trade openness, i.e. the diusion of technology and the possibility of each
country to specialize according to its comparative advantage, are outweighed
by the trade induced increase in economic activity, labeled as the scale eect
of trade.
When considering foreign direct investment, it appears at rst glance
that FDI is not related to the environmental performance of developing
countries at all. However, in the particular case of water quality in low-
income countries, we see that FDI, as predicted theoretically, is indeed asso-
ciated with a decrease in water pollution. This result speaks to some extent
against the arguments of the pollution haven hypothesis since it suggest that
foreign investment does not bring about an increase in dirty industries but,
rather, leads to a transfer of better technologies.
Finally, although the theoretical arguments propose a positive relation-
ship between aid and environmental quality, we do not see any statistically
signicant eect of international aid on the three indicators of environmen-
tal quality. Nevertheless, this nding at least shows that the positive eects
of aid, such as the transfer of technology and the promotion of sustainable
development, are strong enough to cancel out the negative eect of pollution
arising through the aid induced economic growth.
Turning to the political system, the results are supportive of the theo-
retical arguments. The results show that at the stage of early economic de-
velopment, the political system does not seem to have an independent eect
on environmental quality. The reasoning behind this nding, as proposed in
219
CHAPTER 8. CONCLUSION
this thesis, is that at the stage of early economic development the citizenry
in developing countries will not demand that their governments spend scarce
resources on improving environmental quality at the expense of basic needs
such as education and health. Without these necessary electoral incentives,
however, democratic governments would not care more about their environ-
ment than autocratic governments, leading to the results that at the stage of
early economic development democratic developing countries are per se no
better caretakers of their environment than their autocratic counterparts.
Nevertheless, the political system plays an important role in that it me-
diates the eect of international integration on environmental pollution. As
the various interaction eects in sections 5.2.2 to 5.2.5 in chapter 5 show,
with the exception of international aid, the eect of international integration
is, in most of the cases, more conducive to the environment in democracies
than in autocracies. More precisely, with regard to foreign investment we
see that FDI is associated with an increase in air pollution in autocracies but
with a decrease in air pollution in democracies (however, the eect reaches
signicance in the case of SO
2
only and is reversed when considering water
quality).
When considering trade openness, we see a very interesting interaction
eect: although the eect of trade openness is not associated with a de-
crease in pollution for democracies, democratic political systems neverthe-
less soften the pollution increasing eect of trade. Hence we see that trade
is mostly associated with an increase in pollution that is at least not as bad
in democracies as it is in autocracies.
In contrast, IGO membership is always associated with a decrease in
pollution whether we consider a democracy or an autocracy. However, the
eect is more pronounced for democracies in the case of SO
2
emissions and
water quality when considering low-income countries but less pronounced
for democracies when looking at CO
2
emissions. Hence concerning SO
2
emissions, the inuence of IGO membership is more conducive to environ-
mental quality in democracies whereas with CO
2
emissions the inuence of
IGO membership is stronger in autocracies. This dierence with regard to
the environmental indicators is probably due to the specic nature of the
pollution indicator. Since the adverse eects of SO
2
for the environment
and human health are felt much more locally compared to the harms arising
from CO
2
pollution, elected governments might have a greater incentive to
combat SO
2
than CO
2
pollution. Furthermore, SO
2
is mainly a byproduct
of industrial activity whereas CO
2
additionally arises from car emissions.
Following the logic of Bueno De Mesquita et al. (2003), we can deduce that
in the case of SO
2
industrial activity needs to be regulated, hurting mainly
the elite, whereas in the case of CO
2
individual behavior also needs to be
regulated, hurting both the elite and the median voter. Consequently, a
democracy should nd it easier to reduce SO
2
emissions than CO
2
emis-
sions.
220
CHAPTER 8. CONCLUSION
Turning from these more general ndings to the detailed analysis of IGO
membership, which is conducted to shed more light on the specic mecha-
nisms underlying the eect of IGO membership, we can rst conclude that
the positive impact of IGO membership on environmental quality is robust to
various measurement specications. Furthermore, when disaggregating IGO
membership according to the function and the degree of institutionalization
of the organization, we observe some interesting results: the rst conclusion
we can draw from the disaggregation of IGO membership according to the
function of the specic IGO is that interestingly it is not membership in
environmental IGOs that is driving the results. This result, together with
an instrumental variables regression, arms our conclusion that IGO mem-
bership is not endogenous to environmental quality but that the direction
of the eect indeed runs from IGO membership to environmental quality.
The disaggregation of IGO membership rather suggests that it is mem-
bership in umbrella, nuclear, agricultural, economic, social and military or-
ganizations that is positively associated with an increase in environmental
quality. These ndings therefore serve as an initial indication of the mecha-
nisms behind this inuence of IGO membership. For example, due to their
large thematic scope, the inuence of umbrella organizations such as the EU
or ASEAN can be read such that some IGOs seem to be very successful in
connecting dierent issues. Although countries usually join these umbrella
IGOs for economic or political reasons, they are, as a side eect, also inu-
enced by the environmental functions of these organizations. Moreover, the
strong eect of social IGOs in the case of SO
2
emissions suggests two dierent
mechanisms. On the one hand, these organizations may constitute a good
forum for information dissemination and through the socialization process
that takes place at these fora they may also diuse environmental protection
as proposed in the literature on policy diusion (Cao forthcoming; Simmons
& Elkins 2004). On the other hand, many of these organizations such as the
WHO and UNESCO diuse knowledge and technology helpful to the envi-
ronment in general while aiming to protect or improve human health or the
worlds cultural heritage. This latter mechanism of information and know-
how diusion seems also to underlie the inuence of agricultural, economic
and nuclear organizations, as these IGOs often seem to be conducive to the
provision of technological and nancial resources, increasing the capacity of
developing countries to provide environmental quality.
As expected, when dierentiating IGO membership according to the de-
gree of institutionalization, we see that some degree of formalization or insti-
tutionalization of an organization is needed to aect environmental quality in
developing countries. More surprisingly, however, the most institutionalized
or powerful organizations, which are those organizations that can constrain
their member countries the most, are not necessarily the most eective ones.
With regard to both CO
2
emissions and water quality, the eect of medium
structured organizations is more pronounced than the eect of the most
221
CHAPTER 8. CONCLUSION
structured organizations. A possible explanation of this nding could be
that the main mechanisms behind the inuence of IGO membership, which
are technology and knowledge transfer, issue linkage, and a more broader
socialization process towards sustainable development and environmental
protection, might work well as long as the organization has some degree of
institutionalization without necessarily needing to be fully constraining.
These results are further reinforced by the four illustrative case studies.
When evaluating the mandate of the various organizations of which the four
countries, Cameroon, India, Laos and Lithuania are members, three main
mechanisms are underlined through which IGOs seem to inuence coun-
trys environmental performance: issue linkage, technical cooperation and
know-how diusion, and pushing the topic of environmental protection or
sustainable development on the agenda of developing countries. These nd-
ings are supported by Dasgupta et al. (2002) who conclude with respect to
the inuence of international actors on environmental quality in developing
countries that
the international community can play a valuable role in lower-
ing and attening the environmental Kuznets curve by nancing
appropriate training, policy, reforms, information gathering and
public environmental education (Dasgupta et al. 2002, 164).
Moreover, the ndings of the case studies provide us with further evi-
dence as to why democracies and autocracies dier in their reaction towards
IGOs: both the diusion and transfer of technology and the promotion of
sustainable development seem to ourish more intensively within democra-
cies than within autocracies. For example, whereas it was no problem to
get information from the environmental ministries in India and Lithuania,
the governments of Cameroon and Laos do not provide valuable information
about potential environmental problems and governmental activities on the
Internet. This is in line with the general observation that the two democratic
cases, India and Lithuania, are assessed to have more environmental regu-
lations (both domestic environmental regulation and ratication of interna-
tional environmental treaties) in place than the two autocratic countries,
Cameroon and Laos (UNEP 2001b,a; WWF 2009). In addition, the activity
of non-governmental organizations such as the World Wide Fund for Nature
seems to be more intense in India and Lithuania than compared to Cameroon
and Laos, also fostering the diusion and transfer of technologies and know-
how necessary to reduce pollution (UNEP 2001b). This is supported by the
State of the Environment Report in Laos (UNEP 2001b), which posits that
Laos centralized and closed system is a major obstacle to improvements in
the countrys environmental performance since it seems to be very dicult to
build up responsibilities for the enforcement of environmental regulations in
such a context. Altogether, these ndings are supportive of the conclusion
222
CHAPTER 8. CONCLUSION
that information and knowledge might indeed spread more easily in demo-
cratic countries than in autocracies since democracies provide many more
access points for the development of environmental capacities.
Taking up the mechanism of technology and knowledge transfer, the last
chapter of this thesis proposes a simple theoretical framework to formally
model the link between IGO membership and environmental pollution. Re-
lying on a specic-factors trade model, IGO membership is incorporated
in this set-up as a trade-o for each country between selling the countrys
exports at better prices but using a more environmentally stringent and
therefore more expensive technology. The results of the model, which are
obtained using simulations, show that under reasonable assumptions about
the rate at which IGO membership makes production more environmentally
friendly, both democratic and autocratic governments always opt for some
IGO membership and therefore for less environmental degradation.
Turning to the implications of this thesis for the study of public goods
provision, two major points are worth mentioning. First, in contrast to the
ndings of existing studies, the type of the political system per se does not
seem to play a signicant role for the environmental performance of devel-
oping countries. Only in its interplay with the international system does the
political system become important in that it mediates the positive eect of
international integration on environmental quality. This nding underlines
the importance of aligning the theoretical arguments that link the political
system to the provision of public goods to the specic research situation. In
the case at hand, the dierence to other studies that nd a positive eect
of democracy on environmental quality arises because this study focuses on
developing countries only, excluding all rich and consolidated democracies.
Second, this study shows that it is important not to reduce international
integration to one facet, such as trade openness alone, since other impor-
tant characteristics of the international system such as membership in IGOs
seem to aect environmental performance in important ways. Furthermore,
by including various aspects of the international system, a more complete
picture of how globalization aects environmental quality arises. This is
especially important because, as the analysis shows, the impact on envi-
ronmental quality is very dierent for the dierent facets of international
integration. Whereas international aid, trade openness and foreign invest-
ment are not systematically associated with a reduction in emissions, IGO
membership is robustly connected to a decrease in environmental pollution
for all indicators used in this study (SO
2
, CO
2
and BOD).
These general ndings also caution against an overly optimistic per-
spective on foreign investment and trade ows: although I do not nd any
evidence of pollution havens for the sample at hand, the positive perspective
on trade and FDI as catalysts of technology transfer does not seem to be
right either. The same seems to hold true for development aid: although the
case studies show that aid ows are focusing more and more on sustainable
223
CHAPTER 8. CONCLUSION
development and on the provision of environmentally sound technology, this
positive impact of international aid is far outweighed by the aid induced
economic growth, leading neither to a positive nor to a negative eect on
environmental quality in developing countries.
-14
-
1
4
-14 -12
-
1
2
-12 -10
-
1
0
-10 -8
-
8
-8 -14
-
1
4
-14 -12
-
1
2
-12 -10
-
1
0
-10 -8
-
8
-8 6
6
67
7
78
8
89
9
910
10
10 6
6
67
7
78
8
89
9
910
10
10 1
1
12
2
23
3
34
4
495% CI
95% CI
95% CI Fitted values
Fitted values
Fitted values ln SO2 pc
l
n

S
O
2

p
c
ln SO2 pc ln GDP pc
ln GDP pc
ln GDP pc
Figure 8.2: The eect of GDP on SO
2
at various levels of IGO membership
Overall, however, the study also oers some positive conclusions since
IGO membership oers a channel for developing countries to provide bet-
ter environmental quality given their current low level of national income.
This conclusion is underlined by gure 8.2, which shows the regression line
linking GDP per capita to SO
2
emissions at four dierent levels of IGO mem-
bership. Picture 1 shows the regression for those countries whose number
of IGO memberships belongs to the lowest 25 percentile whereas picture 4
shows the regression for those countries whose number of IGO memberships
belongs to the highest 25 percentile. These pictures allow us to see that the
regression line between GDP per capita and SO
2
emissions is steeper for
those countries with few IGO memberships than for those countries with a
larger number of IGO memberships. Therefore, the Environmental Kuznets
Curve is indeed attened for those countries that belong to more IGOs.
Hence developing countries do not need to grow rich before they can com-
bat environmental pollution as international actors have the possibility to
positively inuence the capability and willingness of developing countries to
enhance their environmental performance.
224
Chapter 9
Appendix
9.1 Additional Information Chapter 4
Table 9.1: Time periods for countries SO
2
and CO
2
Country Time Period Income Category
1
Haiti 1970 - 2000 L
Nicaragua 1970 - 2000 L
Moldova 1992 - 2000 L
Ukraine 1992 - 2000 L
Armenia 1991 - 2000 L
Georgia 1991 - 2000 L
Azerbaijan 1991 - 2000 L
Guinea-Bissau 1974 - 2000 L
Mali 1970 - 2000 L
Senegal 1970 - 2000 L
Benin 1970 - 2000 L
Mauritania 1970 - 2000 L
Niger 1970 - 2000 L
Ivory Coast 1970 - 2000 L
Guinea 1986 - 2000 L
Burkina Faso 1970 - 2000 L
Liberia 1970 - 2000 L
Sierra Leone 1970 - 2000 L
Ghana 1970 - 2000 L
Togo 1970 - 2000 L
Cameroon 1970 - 2000 L
Nigeria 1970 - 2000 L
Central African Republic 1970 - 2000 L
Chad 1970 - 2000 L
Republic of Congo 1970 - 2000 L
Democratic Republic of Congo 1970 - 2000 L
1
L: low-income country
LM: low-middle income country
UM: high-middle income country
225
CHAPTER 9. APPENDIX
Table 9.2: Time periods for countries SO
2
and CO
2
Country Time Period Income Category
1
Uganda 1970 - 2000 L
Kenya 1970 - 2000 L
Tanzania 1988 - 2000 L
Burundi 1970 - 2000 L
Rwanda 1970 - 2000 L
Eritrea 1993 - 2000 L
Angola 1985 - 2000 L
Mozambique 1970 - 2000 L
Zambia 1970 - 2000 L
Zimbabwe 1970 - 2000 L
Malawi 1970 - 2000 L
Madagascar 1970 - 2000 L
Sudan 1970 - 2000 L
Tajikistan 1992 - 2000 L
Uzbekistan 1992 - 2000 L
Mongolia 1993 - 2000 L
India 1970 - 2000 L
Bhutan 1980 - 2000 L
Pakistan 1972 - 2000 L
Bangladesh 1972 - 2000 L
Nepal 1970 - 2000 L
Laos 1984 - 2000 L
Indonesia 1970 - 2000 L
Jamaica 1970 - 2000 LM
Guatemala 1970 - 2000 LM
Honduras 1970 - 2000 LM
El Salvador 1970 - 2000 LM
Colombia 1970 - 2000 LM
Guyana 1970 - 2000 LM
Ecuador 1970 - 2000 LM
Peru 1970 - 1999 LM
Bolivia 1970 - 2000 LM
Paraguay 1970 - 2000 LM
Albania 1984 - 2000 LM
FYROM(Macedonia) 1993 - 2000 LM
Bulgaria 1980 - 2000 LM
Romania 1987 - 2000 LM
Russia 1992 - 2000 LM
Latvia 1992 - 2000 LM
Lithuania 1992 - 2000 LM
Belarus 1991 - 2000 LM
Djibouti 1985 - 2000 LM
Morocco 1970 - 2000 LM
Algeria 1970 - 2000 LM
Tunisia 1970 - 2000 LM
Swaziland 1970 - 2000 LM
Iran 1974 - 2000 LM
Egypt 1970 - 2000 LM
1
L: low-income country
LM: low-middle income country
UM: upper-middle income country
226
CHAPTER 9. APPENDIX
Table 9.3: Time periods for countries SO
2
and CO
2
Country Time Period Income Category
1
Syria 1970 - 2000 LM
Jordan 1970 - 2000 LM
Kazakhstan 1992 - 2000 LM
China 1970 - 2000 LM
Thailand 1970 - 2000 LM
Sri Lanka 1970 - 2000 LM
Philippines 1970 - 2000 LM
Papua New Guinea 1975 - 2000 LM
Fiji 1970 - 2000 LM
Trinidad and Tobago 1970 - 2000 UM
Mexico 1970 - 2000 UM
Panama 1970 - 2000 UM
Costa Rica 1970 - 2000 UM
Venezuela 1970 - 2000 UM
Brazil 1970 - 2000 UM
Chile 1970 - 2000 UM
Argentina 1970 - 2000 UM
Uruguay 1970 - 2000 UM
Poland 1985 - 2000 UM
Hungary 1970 - 2000 UM
Czech Republic 1990 - 2000 UM
Slovakia 1993 - 2000 UM
Croatia 1993 - 2000 UM
Estonia 1992 - 2000 UM
Gabon 1970 - 2000 UM
South Africa 1970 - 2000 UM
Botswana 1970 - 2000 UM
Mauritius 1980 - 2000 UM
Turkey 1970 - 2000 UM
Oman 1971 - 2000 UM
South Korea 1976 - 2000 UM
Malaysia 1970 - 2000 UM
1
L: low-income country
LM: low-middle income country
UM: upper-middle income country
Table 9.4: Additional countries and time periods CO
2
Country
1
Time Period Income Category
2
Gambia 1970 - 2000 L
Comoros 1980 - 2000 L
Solomon Islands 1980 - 1999 LM
Equatorial Guinea 1970/77 and 1985/2000 UM
1
Otherwise identical to SO
2
2
L: low-income country
LM: low-middle income country
UM: upper-middle income country
227
CHAPTER 9. APPENDIX
Table 9.5: Time periods for countries BOD
Country Time Period Income Category
1
Haiti 1980 - 1988 L
Nicaragua 1980 - 1985 L
Moldova 1992 - 2000 L
Ukraine 1992 - 2000 L
Armenia 1991 - 2000 L
Azerbaijan 1991 - 2000 L
Equatorial Guinea 1987 and 1990 L
Gambia 1980/82 and 1993 L
Senegal 1980 - 2000 L
Benin 1980 and 1981 L
Niger 1980 - 1982 L
Ivory Coast 1980/83 and 1994/97 L
Burkina Faso 1980/83 and 1998 L
Liberia 1984 - 1986 L
Sierra Leone 1981/82 and 1993 L
Ghana 1980/87 and 1993/95 L
Togo 1980 - 1984 L
Cameroon 1980/84 and 1989/99 L
Nigeria 1980/85 and 1991/96 L
Central African Republic 1980/83 and 1985/93 L
Republic of Congo 1981 - 1988 L
Uganda 1984 - 1989 L
Kenya 1980 - 2000 L
Tanzania 1988, 90/91 and 95/99 L
Burundi 1980, 83 and 86/91 L
Rwanda 1984 - 1986 L
Angola 1991 - 1993 L
Mozambique 1986/96 and 2000 L
Zambia 1980/82, 90 and 94 L
Zimbabwe 1980 - 1996 L
Malawi 1980 - 1998 L
Lesotho 1982/85, 90, 92/94, 96, 98 L
Madagascar 1980 - 1988 L
Sudan 1970 - 1999 L
Mongolia 1993 - 1995 L
India 1980 - 2000 L
Bhutan 1989 L
Pakistan 1980/91 and 1996 L
Bangladesh 1980/92, 95, 97 L
Nepal 1982, 86/91, 93/94, 96 L
Indonesia 1980 - 2000 L
1
L: low-income country
LM: low-middle income country
UM: upper-middle income country
228
CHAPTER 9. APPENDIX
Table 9.6: Time periods for countries BOD
Country Time Period Income Category
Jamaica 1980 - 1996 LM
Guatemala 1980 - 1998 LM
Honduras 1980 - 1995 LM
El Salvador 1980 - 1998 LM
Panama 1980 - 2000 LM
Colombia 1980 - 2000 LM
Guyana 1980 - 1982 LM
Ecuador 1980 - 2000 LM
Peru 1980 - 1996 LM
Bolivia 1980 - 2000 LM
Paraguay 1991 LM
Albania 1988 - 1997 LM
FYROM (Macedonia) 1993 - 1996 LM
Bulgaria 1980 - 2000 LM
Romania 1987 - 2000 LM
Russia 1993 - 2000 LM
Latvia 1992 - 2000 LM
Lithuania 1992 - 2000 LM
Swaziland 1980 - 1995 LM
Morocco 1980 - 2000 LM
Algeria 1980 and 1984/97 LM
Tunisia 1980/81 and 1993/2000 LM
Iran 1980/90 and 1993/2000 LM
Egypt 1980 - 1998 LM
Syria 1980 - 1998 LM
Jordan 1980 - 2000 LM
China 1980 - 2000 LM
Sri Lanka 1980 - 2000 LM
Thailand 1982, 84, 86, 88/91, 93/94 LM
Philippines 1980 - 1997 LM
Papua New Guinea 1980 - 1989 LM
Solomon Islands 1980 - 1984 LM
Fiji 1980/93 and 1996/98 LM
1
L: low-income country
LM: low-middle income country
UM: upper-middle income country
229
CHAPTER 9. APPENDIX
Table 9.7: Time periods for countries BOD
Country Time Period Income Category
1
Trinidad and Tobago 1980 - 2000 UM
Mexico 1980 - 2000 UM
Costa Rica 1984 - 2000 UM
Venezuela 1980 - 1998 UM
Brazil 1980/1985 and 1990/1995 UM
Chile 1980 - 2000 UM
Argentina 1980 - 1999 UM
Uruguay 1980 - 2000 UM
Poland 1985 - 2000 UM
Hungary 1980 - 2000 UM
Czech Republic 1990 - 1998 UM
Slovakia 1993 - 1999 UM
Croatia 1992 - 2000 UM
Gabon 1980/81 and 1991/95 UM
South Africa 1980 - 2000 UM
Botswana 1980 - 2000 UM
Mauritius 1980 - 1999 UM
Turkey 1980 - 2000 UM
Oman 1985 - 2000 UM
South Korea 1980 - 2000 UM
Malaysia 1980/97 and 1999/2000 UM
1
L: low-income country
LM: low-middle income country
UM: upper-middle income country
230
CHAPTER 9. APPENDIX
9.2 Additional Tables and Figures Chapter 5
Table 9.8: Interaction between IGO membership and political system
sample: low-income countries
(1) )
BOD
IGO membership -0.01
(0.00)
Polity 0.01
(0.01)
IGO*Polity -0.00
(0.00)
aid -0.00
(0.00)
ln trade openness 0.02
(0.03)
FDI % GDP -0.02***
(0.01)
ln GDP pc 0.61***
(0.13)
Population density 0.01***
(0.00)
Growth -0.18
(0.15)
Time trend -0.02***
(0.01)
Constant -11.35***
(0.93)
Country xed eects yes
rho 0.54
Observations 370
Number of countries 40
R
2
1.00
Standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1
231
CHAPTER 9. APPENDIX
Table 9.9: Interaction between aid (divided by GDP) and political system
sample: all developing countries
(1) (2) (3)
SO
2
CO
2
BOD
Aid/GDP 0.20 0.58 0.88
(0.29) (0.39) (0.54)
Polity -0.00 0.00 -0.00
(0.00) (0.00) (0.00)
Aid*Polity -0.00 -0.04 -0.04
(0.03) (0.04) (0.04)
IGO membership -0.01*** -0.01*** -0.01***
(0.00) (0.00) (0.00)
ln trade openness 0.04 0.10*** -0.00
(0.03) (0.03) (0.03)
FDI % GDP 0.00 -0.00 -0.01*
(0.00) (0.00) (0.00)
ln GDP pc 0.43*** 0.55*** 0.65***
(0.06) (0.06) (0.07)
Population density -0.00 0.00*** 0.00**
(0.00) (0.00) (0.00)
Growth 0.01 -0.01 -0.15**
(0.04) (0.03) (0.07)
Time trend 0.00 0.01*** 0.00
(0.01) (0.00) (0.00)
Constant -15.73*** -6.28*** -11.36***
(0.70) (0.56) (1.04)
Country xed eects yes yes yes
rho 0.77 0.72 0.54
Observations 2512 2600 1155
R
2
0.96 0.87 0.99
Number of countries 104 108 93
Standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1
232
CHAPTER 9. APPENDIX
-1
-
1
-1 -.5
-
.
5
-.5 0
0
0.5
.
5
.5 1
1
11.5
1
.
5
1.5 Marginal Effect of Aid
M
a
r
g
i
n
a
l

E
f
f
e
c
t

o
f

A
i
d
Marginal Effect of Aid 1
1
16
6
611
11
11 16
16
16 21
21
21 Polity
Polity
Polity Marginal Effect of Aid
Marginal Effect of Aid
Marginal Effect of Aid 95% Confidence Interval
95% Confidence Interval
95% Confidence Interval

Dependent Variable: ln SO2 pc, full sample
Dependent Variable: ln SO2 pc, full sample
Dependent Variable: ln SO2 pc, full sample

Marginal Effect of Aid as Democracy Changes
Marginal Effect of Aid as Democracy Changes
Marginal Effect of Aid as Democracy Changes
Figure 9.1: Marginal eect of aid scaled by GDP SO
2
-2
-
2
-2 -1
-
1
-1 0
0
01
1
12
2
2Marginal Effect of Aid
M
a
r
g
i
n
a
l

E
f
f
e
c
t

o
f

A
i
d
Marginal Effect of Aid 1
1
16
6
611
11
11 16
16
16 21
21
21 Polity
Polity
Polity Marginal Effect of Aid
Marginal Effect of Aid
Marginal Effect of Aid 95% Confidence Interval
95% Confidence Interval
95% Confidence Interval

Dependent Variable: ln CO2 pc, full sample
Dependent Variable: ln CO2 pc, full sample
Dependent Variable: ln CO2 pc, full sample

Marginal Effect of Aid as Democracy Changes
Marginal Effect of Aid as Democracy Changes
Marginal Effect of Aid as Democracy Changes
Figure 9.2: Marginal eect of aid scaled by GDP CO
2
-2
-
2
-2 -1
-
1
-1 0
0
01
1
12
2
2Marginal Effect of Aid
M
a
r
g
i
n
a
l

E
f
f
e
c
t

o
f

A
i
d
Marginal Effect of Aid 1
1
16
6
611
11
11 16
16
16 21
21
21 Polity
Polity
Polity Marginal Effect of Aid
Marginal Effect of Aid
Marginal Effect of Aid 95% Confidence Interval
95% Confidence Interval
95% Confidence Interval

Dependent Variable: ln BOD pc, full sample
Dependent Variable: ln BOD pc, full sample
Dependent Variable: ln BOD pc, full sample

Marginal Effect of Aid as Democracy Changes
Marginal Effect of Aid as Democracy Changes
Marginal Effect of Aid as Democracy Changes
Figure 9.3: Marginal eect of aid scaled by GDP BOD
233
CHAPTER 9. APPENDIX
Table 9.10: Interaction between log of export divided by GDP and political
system sample: all developing countries
(1) (2)
SO
2
CO
2
ln export/GDP 0.06** 0.12***
(0.02) (0.02)
Polity -0.08** -0.08***
(0.03) (0.03)
Polity*Export -0.00** -0.00***
(0.00) (0.00)
IGO membership -0.01*** -0.01***
(0.00) (0.00)
Aid -0.00 0.00
(0.00) (0.00)
FDI % GDP 0.00 -0.00
(0.00) (0.00)
ln GDP pc 0.42*** 0.54***
(0.06) (0.05)
Population density -0.00 0.00***
(0.00) (0.00)
Growth -0.00 -0.00
(0.04) (0.04)
Time trend 0.01 0.01***
(0.01) (0.00)
Constant -15.37*** 0.00
(0.67) (0.00)
Country xed eects yes yes
rho 0.77 0.72
Observations 2572 2663
R
2
0.96 0.87
Number of countries 106 110
Standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1
234
CHAPTER 9. APPENDIX
-.1
-
.
1
-.1 -.05
-
.
0
5
-.05 0
0
0.05
.
0
5
.05 .1
.
1
.1 Marginal Effect of Export/GDP
M
a
r
g
i
n
a
l

E
f
f
e
c
t

o
f

E
x
p
o
r
t
/
G
D
P
Marginal Effect of Export/GDP 1
1
16
6
611
11
11 16
16
16 21
21
21 Polity
Polity
Polity Marginal Effect of Trade openness
Marginal Effect of Trade openness
Marginal Effect of Trade openness 95% Confidence Interval
95% Confidence Interval
95% Confidence Interval

Dependent Variable: ln SO2 pc, full sample
Dependent Variable: ln SO2 pc, full sample
Dependent Variable: ln SO2 pc, full sample

Marginal Effect of Export/GDP as Democracy Changes
Marginal Effect of Export/GDP as Democracy Changes
Marginal Effect of Export/GDP as Democracy Changes
Figure 9.4: Marginal eect of export divided by GDP SO
2
-.05
-
.
0
5
-.05 0
0
0.05
.
0
5
.05 .1
.
1
.1 .15
.
1
5
.15 Marginal Effect of Export/GDP
M
a
r
g
i
n
a
l

E
f
f
e
c
t

o
f

E
x
p
o
r
t
/
G
D
P
Marginal Effect of Export/GDP 1
1
16
6
611
11
11 16
16
16 21
21
21 Polity
Polity
Polity Marginal Effect of Trade openness
Marginal Effect of Trade openness
Marginal Effect of Trade openness 95% Confidence Interval
95% Confidence Interval
95% Confidence Interval

Dependent Variable: ln CO2 pc, full sample
Dependent Variable: ln CO2 pc, full sample
Dependent Variable: ln CO2 pc, full sample

Marginal Effect of Export/GDP as Democracy Changes
Marginal Effect of Export/GDP as Democracy Changes
Marginal Effect of Export/GDP as Democracy Changes
Figure 9.5: Marginal eect of export divided by GDP CO
2
235
CHAPTER 9. APPENDIX
9.3 Additional Information Chapter 7
This section presents an empirical justication for the assumption of the the-
oretical model that IGO membership increases the price at which a country
can sell its exports. This empirical justication is derived using the logic of
the gravity model. The gravity model says that trade intensity between two
countries depends rst on their size in terms of the two countries GDPs with
countries having higher GDP trading more. Second, trade intensity between
two countries depends on the distance between them since the transporta-
tion costs are lower for countries that are closer together, implying that these
countries should trade more with each other (Tinbergen 1962; Anderson &
van Wincoop 2003; Santos Silva & Tenreyro 2006). Furthermore, trade in-
tensity also depends on trade distortions between these countries since any
obstacles to bilateral trade ows decrease the intensity of trade between the
two countries (Anderson & van Wincoop 2003).
According to our model, IGO membership serves to increase the price of
a countrys exports. This could also be thought of as if IGO membership re-
duced the trade barriers or transportation costs of trade, inducing countries
to trade more with each other. Hence if our assumption concerning IGO
membership is realistic we should see that those countries with more IGO
memberships should also trade more with each other since IGO membership
serves to reduce the trade barriers between them.
We can test this assumption using a standard gravity model set-up.
Following Santos Silva & Tenreyro (2006) I use a Poisson pseudo-maximum-
likelihood method to estimate the eect of IGO membership in the context
of a standard gravity model
1
. Using a Poisson model corrects for both the
heteroskedasticity of the error term in the standard gravity model and for
the fact that bilateral trade between many pairs of countries is zero. In
addition to the Poisson model, I provide the results using standard OLS
regression while dropping those pairs of countries with zero trade, which
used to be the standard approach followed by most empirical studies on the
gravity model (Santos Silva & Tenreyro 2006).
In particular, I use bilateral export data from Gleditsch (2002) as the
dependent variable, which encompasses country As export to country B
and vice versa for 172 countries. To measure IGO membership I rely on a
variable indicating the number of common IGO membership between each
country pair per year. The data is taken from the Correlates of War database
(Pevehouse et al. 2003). As further independent variables I include the usual
gravity variables: the log of GDP and GDP per capita for both the exporter
and the importer country as well as the geographical distance between the
two countries. The data on GDP also come from Gleditsch (2002) and the
1
Although trade data are no count data, using a Poisson model is the correct speci-
cation since theoretically deriving the gravity equation leads to a form corresponding to
the Poisson model (Santos Silva & Tenreyro 2006).
236
CHAPTER 9. APPENDIX
data on distance are from Gleditsch & Ward (2001).
In addition to these variables, I follow Santos Silva & Tenreyro (2006)
and control for several other variables. I include a dummy variable measur-
ing whether the two countries are contiguous and a dummy variable mea-
suring whether the two countries have the same language. Furthermore, I
incorporate two dummy variables to capture whether the exporter or the
importer is landlocked. Moreover, various dummy variables are included
that capture the extent to which the two nations have common origins. I
therefore control for whether the two countries have had a common colo-
nizer after 1945, whether they have had a colonial relationship after 1945,
whether they have ever had a colonial link and whether they are currently
in a colonial relationship. The data on the various dummy variables come
from Centre dEtudes Prospectives et dInformations Internationales (2005).
Moreover, I control for the level of democracy by including the Polity
score of both the exporter and the importer country. Finally, as is common
in the gravity literature, the model contains importer, exporter and time
xed eects. I estimate the model for the time period after the Cold War,
i.e. from 1990 to 2000, to exclude the possibility that military alliances
might bias the results.
As can be seen in table 9.11, introducing joint IGO membership in the
gravity model conrms our assumption that IGO membership is indeed as-
sociated with higher trade levels. In all of the four models we see that
joint IGO membership is positive and statistically signicant, implying that
IGO membership can indeed be interpreted as reducing the trade distortions
or transportation costs between trading partners as assumed in the formal
model. In model (1) we further see that the results for most independent
variables follow the results in the gravity literature. Countries with higher
GDP levels trade more as do countries that speak the same language, are
in a colonial relationship or are located next to each other. In contrast, the
further apart two countries are, the less they trade, and landlocked coun-
tries also trade less. Only the political system does not seem to matter since
being a democracy does not have a signicant inuence on bilateral trade
ties.
Overall, the results in table 9.11 conrm the assumption that IGO mem-
bership implies a positive eect on the conditions at which countries can sell
their exports since those countries that have more common IGO member-
ships are characterized by higher export levels.
237
CHAPTER 9. APPENDIX
Table 9.11: Gravity model including joint IGO membership
(1) (2)
1990-2000 1990-2000
Poisson OLS
ln exporter GDP 0.63*** 1.12***
(0.16) (0.12)
ln importer GDP 0.81*** 0.60***
(0.14) (0.12)
ln exporter GDP pc 0.42** -0.75***
(0.21) (0.12)
ln importer GDP pc 0.34* -0.13
(0.17) (0.12)
ln distance -0.65*** -0.94***
(0.01) (0.03)
Joint IGO membership 0.01*** 0.03***
(0.00) (0.00)
Importer Polity level 0.10 0.06**
(0.10) (0.03)
Exporter Polity level 0.18 0.03
(0.13) (0.03)
Current colonial relationship 0.18*** 0.53***
(0.05) (0.16)
Colonial relationship after 1945 0.40*** 0.96***
(0.07) (0.15)
Common colonizer -0.00 0.44***
(0.06) (0.07)
Ever in a colonial relationship -0.02 0.79***
(0.03) (0.12)
Common language 0.16*** 0.14***
(0.03) (0.05)
Contiguity 0.56*** 0.66***
(0.03) (0.10)
Exporter landlocked -1.50*** 1.71***
(0.58) (0.48)
Importer landlocked -1.33** -1.66***
(0.55) (0.56)
Constant -34.80*** -28.24***
(3.99) (3.83)
Exporter xed eects yes yes
Importer xed eects yes yes
Time xed eects yes yes
Observations 314644 179254
R-squared 0.94 0.71
Robust standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1
238
Bibliography
AARDO (2009a), Activities, http://www.aardo.org/activities1.htm
last checked: 01.07.2009.
AARDO (2009b), Development Pilot Projects, http://www.aardo.org/
developmentProject.htm last checked: 01.07.2009.
Abbott, K. W. & Snidal, D. (2000), Hard and Soft Law in International
Governance, International Organization 54(3), 421456.
Achen, C. (2000), Why Lagged Dependent Variables Can Surpress the Ex-
planatory Power of Other Independent Variables, Paper Presented at the
Annual Meeting of the Political Methodology Section of the American Po-
litical Science Association .
African Development Bank (2009), Programme de developpement
durable du bassin du Lac Tchad (PRODEBALT), http://www.
afdb.org/en/projects-operations/project-portfolio/project/
programme-de-developpement-durable-du-bassin-du-lac-tchad-prodebalt-604/
last checked: 01.07.2009.
African Union (2009a), African Union in a Nutshell, http://www.
africa-union.org/root/au/AboutAu/au_in_a_nutshell_en.htm last
checked: 01.07.2009.
African Union (2009b), Capacity Building for an Africa-Wide Biosafety
System, http://www.africa-union.org/root/au/AUC/Departments/
HRST/biosafety/AU_Biosafety_1b.htm last checked: 01.07.2009.
Africare (2006), Annual Report, http://www.africare.org/news/
annual/ last checked: 01.07.2009.
Africare (2009), About Us, http://www.africare.org/aboutus/index.
php last checked: 01.07.2009.
Ahlquist, J. S. (2006), Economic Policy, Institutions, and Capital Flows:
Portfolio and Direct Investment Flows in Developing Countries, Interna-
tional studies Quarterly 50(3), 681704.
239
BIBLIOGRAPHY
Alesina, A. & Dollar, D. (2000), Who gives Foreign Aid to Whom and
Why?, Journal of Economic Growth 5(1), 3363.
Alesina, A. & Weder, B. (2002), Do Corrupt Governments Receive Less
Foreign Aid?, The American Economic Review 92(4), 11261137.
Alvarez, M., Cheibub, J. A., Limongi, F. & Przeworski, A. (1996), Classify-
ing Political Regimes, Studies in Comparative International Development
31(2), 137.
Anderson, J. E. & van Wincoop, E. (2003), Gravity with gravitas: A solu-
tion to the border puzzle, American Economic Review 93(1), 170192.
Andreoni, J. & Levinson, A. (1998), The Simple Analytics of the Environ-
mental Kuznets Curve, NBER Working Paper No. w6739 .
Antweiler, W., Copeland, B. R. & Taylor, M. S. (2001), Is Free Trade Good
for the Environment?, The American Economic Review 91(4), 877908.
Araya, M. (2002), Environmental Benets of Foreign Direct Investment: A
Literature Review, OECD Report, Working Party on Global and Struc-
tural Policies .
Arrow, K., Bolin, B., Costanza, R., Dasgupta, P., Folke, C., Holling, C.,
Jansson, B.-O., Levin, S., Mueler, K.-G., Perrings, C. & Pimentel, D.
(1995), Economic Growth, Carrying Capacity, and the Environment,
Science 268(5210), 520521.
Asian Development Bank (2009), Projects, http://www.adb.org/
Projects/ last checked: 01.07.2009.
Baettig, M. & Bernauer, T. (2009), National Institutions and Global Public
Goods: Are Democracies More Cooperative in Climate Change Policy?,
International Organization 63, 281308.
Barrett, S. (2000), Trade and Environment: Local Versus Multilateral Re-
forms, Environment and Development Economics 5, 349359.
Barro, R. J. (1996), Democracy and Growth, Journal of Economic Growth
1, 127.
Barro, R. J. & Lee, J.-W. (2005), IMF-Programs: Who is Chosen and What
are the Eects?, Journal of Monetary Economics 52(7), 12451269.
Beck, N. (2001), Time-Series-Cross-Section Data: What Have We Learned
in the Past Few Years?, American Review of Political Science 4, 271293.
Beck, N. & Katz, J. N. (1995), What to Do or (Not to Do) with Time-Series
Cross-Section Data, American Political Science Review 89, 63447.
240
BIBLIOGRAPHY
Beck, N. & Katz, J. N. (1996), Nuisance vs. Substance: Specifying and
Estimating Time-Series-Cross-Section Models, Political Analysis 6(1), 1
36.
Beckerman, W. (1992), Economic Growth and the Environment: Whose
Growth? Whose Environment?, World Development 20(4), 481496.
BEF (2009), About BEF, http://www.bef.lv/10/ last checked:
01.07.2009.
Bernauer, T. & Cadu, L. (2004), In Whose Interest? Pressure Group
Politics, Economic Competition and Environmental Regulation, Journal
of Public Policy 24(1), 99126.
Bernauer, T., Kalbhenn, A., Koubi, V. & Ruo, G. (2009), A Comparison
of International and Domestic Sources of Global Governance Dynamics,
Paper Presented at the 50th Annual ISA Convention, February 15-18 New
York .
Bernauer, T. & Koubi, V. (2006), States as Providers of Public Goods:
How Does Government Size Aect Environmental Quality?, Center for
Comparative and International Studies (CIS) Working Paper .
Bernauer, T. & Koubi, V. (2009), Eects of Political Institutions on Air
Quality, Ecological Economics 68(5), 13551365.
Bernauer, T. & Kuhn, P. (forthcoming), Is There an Environmental Ver-
sion of the Kantian Peace? Insights From Water Pollution in Europe.
European Journal of International Relations, European Journal of Inter-
national Relations .
Beron, K. J., Murdoch, J. C. & Vijverberg, W. P. M. (2003), Why Coop-
erate? Public Goods, Economic Power, and the Montreal Protocol, The
Review of Economics and Statistics 85(2), 286297.
Birdshall, N. & Wheeler, D. (1993), Trade Policy and industrial pollution
in Latin America: Where are the Pollution Havens?, Journal of Environ-
ment & Development 2(1), 137147.
Blackman, A. & Wu, X. (1999), Foreign Direct Investment in Chinas Power
Sector: Trends, Benets, and Barriers, Energy Policy 27, 695711.
Bliss, H. & Russett, B. (1998), Democratic Trading Partners: The Liberal
Connection, 1962-1989, The Journal of Politics 60, 11261147.
Boehmer, C., Gartzke, E. & Nordstrom, T. (2004), Do Intergovernmental
Organizations Promote Peace, World Politics 57(1).
241
BIBLIOGRAPHY
Bollen, K. A. (1993), Liberal Democracy: Validity and Method Factors in
Cross-National Measures, American Journal of Political Science 37(4).
Bolt, K., Matete, M. & Clemens, M. (2002), Manual for Calculating Ad-
justed Net Savings, World Bank .
Boone, P. (1996), Politics and the Eectiveness of Foreign Aid, European
Economic Review 40(2), 289329.
Botcheva-Andonova, L., Manseld, E. D. & Milner, H. V. (2006), Interna-
tional Trade and Environmental Policy in the Post-Communist World,
Paper Presented at the International Political Economy Society Inaugural
Meeting, Princeton, November 2006 .
Brambor, T., Clark, W. & Golder, M. (2006), Understanding Interaction
Models: Improving Empirical Analyses, Political Analysis 14, 6382.
Bueno De Mesquita, B., Smith, A., Siverson, R. M. & Morrow, J. D. (2003),
The Logic of Political Survival, MIT Press, Cambridge.
Bundesministerium f ur Umwelt, Naturschutz und Reaktorsicherheit (2009),
Umweltschutz im Beitrittsverfahren, http://www.bmu.de/europa_und_
umwelt/eu-erweiterung/doc/2244.php last checked: 01.07.2009.
Busse, M. (2004), Trade, Environmental Regulations and the World Trade
Organization, World Bank Policy Research Working Paper 3361 .
B uthe, T. & Milner, H. V. (2008), The Politics of Foreign Direct Investment
into Developing Countries: Increasing FDI through International Trade
Agreements?, American Journal of Political Science 52(4), 741762.
CAB International (2009), About Us, http://www.cabi.org/datapage.
asp?iDocID=156 last checked: 01.07.2009.
Cao, X. (forthcoming), Networks of Intergovernmental Organizations and
Convergence in Domestic Economic Policies, International Studies Quar-
terly .
Cao, X. & Prakash, A. (forthcoming), Trade Competition and Domestic
Pollution: A Panel Study, 1980-2003, International Organization .
Caves, R. E., Frankel, J. A. & Jones, R. W. (2002), World Trade and Pay-
ments. An Introduction, Addison Wesley, Boston.
Cederman, L.-E., Hug, S. & Krebs, L. F. (forthcoming), Democratization
and Civil War: Empirical Evidence, Journal of Peace Research .
Center for International Earth Science Information Network (CIESIN)
(2002), Environmental Sustainability Index, Yale Center for Environ-
mental Law and Policy, Yale University .
242
BIBLIOGRAPHY
Centre dEtudes Prospectives et dInformations Internationales (2005),
Distances Database, http://www.cepii.fr/anglaisgraph/bdd/
distances.htm last checked: 01.07.2009.
Cheibub, J. A. (2004), Political Institutions, Working Paper Yale Univer-
sity .
Cheibub, J. A. & Gandhi, J. (2004), Classifying Political Regimes: A
Six-Fold Measure of Democracies and Dictatorships, Paper Presented at
the 2004 Annual Meeting of the American Political Science Association,
Chicago. .
Choi, S.-W. & Samy, Y. (2008), Reexamining the Eect of Democratic Insti-
tutions on Inows of Foreign Direct Investment in Developing Countries,
Foreign Policy Analysis 4(1), 83103.
Christmann, P. & Taylor, G. (2001), Globalization and the Environment:
Determinants of Firm Self Regulation in China, Journal of International
Business Studies 32(3), 439458.
Chudnovsky, D. & Lopez, A. (2003), Diusion of Environmentally Friendly
Technologies by Multinational Corporations in Developing Countries, In-
ternational Journal of Technology Management and Sustainable Develop-
ment 2(1), 518.
Clapp, J. (1998), Foreign Direct Investment in Hazardous Industries in
Developing Countries. Rethinking the Debate, Environmental Politics
7(4), 92113.
Coase, R. (1960), The Problem of Social Cost, Journal of Law and Eco-
nomics 3, 144.
Cohen, J., Cohen, P. & West, S. G. (2003), Applied Multiple Regres-
sion/Correlation Analysis for the Behavioral Sciences, Lawrence Erlbaum
Associates, Mahwah, New Jersey.
Cole, M. A. (2004), Trade, the Pollution Haven Hypothesis and the Environ-
mental Kuznets Curve: Examining the Linkages, Ecological Economics
48, 7181.
Congleton, R. (1992), Political Institutions and Pollution Control, The
Review of Economics and Statistics 74(3), 412421.
Convention on Biological Diversity (2009), Cartagena Protocol on
Biosafety, http://www.cbd.int/biosafety/background.shtml last
checked: 01.07.2009.
Copeland, B. R. & Taylor, M. S. (1994), North-South Trade and the Envi-
ronment, The Quarterly journal of Economics 109(3), 755787.
243
BIBLIOGRAPHY
Coppedge, M. & Reinicke, W. H. (1991), Measuring polyarchy, in A. Inkeles,
ed., On Measuring Democracy: Its Consequences and Concomitants,
Transaction, New Brunswick, NJ.
Correlates of War (2008), Data Set of National Military Capabilities. Ver-
sion 3.02.
Council of Baltic Sea States (2003), Baltic Sea States Declaration
on Environment and Sustainable Development, http://www.cbss.
org/Environment/creating-a-sustainable-baltic-sea-regionlast
checked: 01.07.2009.
Dasgupta, S., Hettige, H. & Wheeler, D. (1998), What Improves Environ-
mental Performance? Evidence from Mexican Industry, Policy Research
Working Paper 1877, World Bank Development Research Group .
Dasgupta, S., Laplante, B. & Mamingi, N. (1998), Capital Market Re-
sponses To Environmental Performance in Developing Countries, World
Bank, Development Research Group, Working Paper wps1909 .
Dasgupta, S., Laplante, B., Wang, H. & Wheeler, D. (2002), Confronting
the Environmental Kuznets Curve, Journal of Economic Perspectives
16(1), 147168.
Deacon, R. T. (1999), The Political Economy of Environment-Development
Relationships: A Preliminary Framework, Working Paper, University of
California, Santa Barbara .
Deacon, R. T. (2003), Dictatorship, Democracy, and the Provision of Public
Goods, Working Paper, University of California, Santa Barbara .
Dean, J. M., Lovely, M. E. & Wang, H. (2005), Are Foreign Investors At-
tracted to Weak Environmental Regulations? Evaluating the Evidene
From China, World Bank Policy Research Paper No. 3505. Washington,
DC: World Bank .
Deger, S. & Smith, R. (1983), Military Expenditure and Growth in Less
Developed Countries, Journal of Conict Resolution 27(2), 335353.
DeSoysa, I. & Neumayer, E. (2005), False Prophet, or Genuine Savior?
Assessing the Eects of Economic Openness on Sustainable Development,
1980-99, International Organization 59(3), 731772.
Dobbin, F., Simmons, B. A. & Garrett, G. (2007), The Global Diusion
of Public Policies: Social Construction, Coercion, Competition, or Learn-
ing?, Annual Revue of Sociology 33, 449472.
244
BIBLIOGRAPHY
Doces, J. A. (2006), Globalization and Democracy: An Empirical Analysis
Using A Simultaneous Equation Approach, Paper Presented at the Inter-
national Political Economy Society Inaugural Meeting, Princeton, Novem-
ber 2006 .
Dorussen, H. & Ward, H. (2008), Intergovernmental Organizations and the
Kantian Peace. A Network Perspective, Journal of Conict Resolution
52(2), 189212.
Dreher, A. & Jensen, N. M. (2007), Independent Actor or Agent? An
Empirical Analysis of the Impact of US Interests on IMF Conditions,
Journal of Law and Economics 50(1), 105124.
Dreher, A., Nunnenkamp, P.,

Ohler, H. & Weisser, J. (2009), Acting Au-
tonomously or Mimicking the State and Peers? A Panel Tobit Analysis of
Financial Dependence and Aid Allocation by Swiss NGOs, KOF Working
Papers No. 219 .
Drezner, D. W. (2001), Globalization and Policy Convergence, Interna-
tional Studies Review 3, 5378.
EBRD (2009a), Environmental Procedures, http://www.ebrd.com/
about/policies/enviro/procedur/index.htmlast checked: 01.07.2009.
EBRD (2009b), Project Summary Documents Lithuania, http://www.
ebrd.com/projects/psd/country/lithu.htmlast checked: 01.07.2009.
ECCAS (2009), Environnement, http://www.ceeac-eccas.org/index.
php?rubrique=domaine-intervention&id=13last checked: 01.07.2009.
Eichengreen, B. & Leblang, D. (2006), Democracy and Globalization,
NBER Working Paper No. W12450 .
Eliste, P. & Fredriksson, P. G. (2002), Environmental Regulations, Trans-
fers, and Trade: Theory and Evidence, Journal of Environmental Eco-
nomics and Management 43, 234250.
Engle, R. (1984), Wald, Likelihood Ratio and Lagrange Multiplier Tests
in Econometrics, in Z. Griliches & M. Intriligator, eds, Handbook of
Econometrics, Vol. 2, North-Holland, Amsterdam, pp. 775826.
Eriksen, J. & Hansen, M. (1999), Environmental Aspects of Danish Foreign
Direct Investment in Developing Countries. Managing the Environment
in an Open Economy, Occasional Paper. Cross Border Environmental
Management Project, CHP: Copenhagen Business School .
Eskeland, G. & Harrison, A. E. (2003), Moving to Greener Pastures? Multi-
nationals and The Pollution-Haven Hypothesis, Journal of Development
Economics 70, 123.
245
BIBLIOGRAPHY
Estonian Ministry of Foreign Aairs (2009), Baltic Cooperation: Estonia,
Latvia and Lithuania, http://www.vm.ee/eng/kat_202/840.html last
checked: 01.07.2009.
Esty, D. C. (2001), Bridging the Trade-Environment Divide, Journal of
Economic Perspectives 15(3), 113130.
Esty, D. C. & Giradin, D. (1998), Environmental Protection and Inter-
national Competitiveness: A Conceptual Framework, Journal of World
Trade 32(3), 546.
FAO (2009a), African Timber Organization (ATO), http://www.fao.org/
docrep/004/AC135E/ac135e03.htmlast checked: 01.07.2009.
FAO (2009b), The Natural Resources Management and Environment
Department (NR), http://www.fao.org/nr/dep/about_en.htmlast
checked: 01.07.2009.
Feng, Y. (2001), Political Freedom, Political Instability, and Policy Uncer-
tainty, International Studies Quarterly 45(2), 271294.
Finus, M. (2001), Game Theory and International Environmental Coopera-
tion, Edward Elgar, Cheltenham.
Frank, J. D., Hironaka, A. & Schofer, E. (2000), Environmentalisms as
a Global Institution. Reply to Buttel, American Sociological Review
65(1), 122127.
Frankel, J. A. (2003), The Environment and Globalization, NBER Working
Paper No. W10090 .
Frankel, J. A. & Rose, A. K. (2005), Is Trade Good or Bad for the En-
vironment? Sorting out the Causality, The Review of Economics and
Statistics 87(1), 8591.
Fredriksson, P. G. & Gaston, N. (2000), Ratication of the 1992 Climate
Change Convention: What Determines Legislative Delay?, Public Choice
104(3-4), 345368.
Fredriksson, P. G. & Svensson, J. (2003), Political, Instability, Corruption
and Policy Formation: The Case of Environmental Policy, Journal of
Public Economics 87, 13831405.
Fredriksson, P. G. & Wollscheid, J. R. (2007), Democratic Institutions
versus Autocratic Regimes: The Case of Environmental Policy, Public
Choice 130(3), 381393.
Freedom House (2006), Annual Survey of Freedom Country Scores, http:
//freedomhouse.org/ratings/index.htmlast checked: 01.07.2009.
246
BIBLIOGRAPHY
Garrett, G. (2001), Globalization and Government Spending Around the
World, Studies in Comparative International Development 35(4), 329.
Garson, G. D. (2008), Path Analysis, http://faculty.chass.ncsu.edu/
garson/PA765/path.htm#biblio last checked: 10.08.2009.
Gartzke, E., Boehmer, C., Nordstrom, T. & Hewitt, J. J. (2006), Disagre-
gating International Organizations in Time and Space, Paper Presented
at the 49th Annual ISA Convention, March 26-29 San Francisco .
Gasiorowski, M. J. (1996), An Overview of the Political Regime Change
Dataset, Comparative Political Studies 29(4), 469483.
Gassebner, M., Lamla, M. L. & Sturm, J. E. (2006), Economic, Demo-
graphic and Political Determinants of Pollution Reassessed: A Sensitivity
Analysis, CESifo Working Paper No. 1699 .
Gaubatz, K. T. (1996), Democratic States and Commitment in interna-
tional Relations, International Organization 50(1), 109139.
GEF (2008a), About the GEF, http://www.gefweb.org/interior.aspx?
id=50 last checked: 01.07.2009.
GEF (2008b), Transfer of Environmentally Sound Technologies: The GEF
Experience, http://www.gefweb.org/interior_right.aspx?id=21584
last checked: 01.07.2009.
GEF (2008c), What is the GEF, http://www.gefweb.org/interior_
right.aspx?id=50 last checked: 01.07.2009.
Gentry, B. S. (1998), Private Capital Flows and the Environment, Lessons
from Latin America, Edward Elgar Publishing, Cheltenham, UK.
Gleditsch, K. S. (2002), Expanded Trade and GDP Data, Journal of Con-
ict Resolution 46(5), 712724.
Gleditsch, K. S. & Ward, M. D. (2000), War and Peace in Space and Time:
The Role of Democratization, International Studies Quarterly 44(1), 1
29.
Gleditsch, K. S. & Ward, M. D. (2001), Measuring Space: A Minimum
Distance Database, Journal of Peace Research 38, 749768.
Green, D. P., Kim, S. Y. & Yoon, D. H. (2001), Dirty Pool, International
Organization 55(2), 441468.
Grey, K. & Brack, D. (2002), Environmental Issues in Policy-Based Com-
petition for Investment: A Literature Review, OECD Report, Working
Party on Global and Structural Policies .
247
BIBLIOGRAPHY
Grossman, G. M. & Krueger, A. B. (1993), Environmental Impacts of a
North American Free Trade Agreement, in P. Garber, ed., The U. S.-
Mexico Free Trade Agreement, MIT Press, Cambridge, MA.
Grossman, G. M. & Krueger, A. B. (1995), Economic Growth and the
Environment, The Quarterly Journal of Economics 110(2), 353377.
Group of Fifteen (2008), Current Projects of the Group of Fifteen, http:
//www.g15.org/gfprojects.html last checked: 01.07.2009.
Guerin, S. S. & Manzocchi, S. (2006), When FDI Flows from Rich to Poor
Countries: Do Democracy and Economic Reform Matter?, CEPS Work-
ing Document No. 251 .
Gutner, T. (2005), Explaining the Gaps between Mandate and Performance:
Agency Theory and World Bank Environmental Reform, Global Environ-
mental Politics 5(2), 1037.
Hadenius, A. (1992), Democracy and Development, Cambridge University
Press, Cambridge UK.
Hafner-Burton, E., Stein, J. V. & Gartzke, E. (2008), International Orga-
nizations Count, Journal of Conict Resolution 52(2), 175188.
Hegre, H., Ellingsen, T., Gates, S. & Gleditsch, N. P. (2001), Toward a
Democratic Civil Peace? Democracy, Political Change, and Civil War,
1816-1992, American Political Science Review 95(1), 3348.
Henisz, W. J., Zelner, B. A. & Guilen, M. F. (2004), International Co-
ercion, Emulation and Policy Diusion: Market-Oriented Infrastructure
Reforms, 1977-1999, http://ssrn.com/abstract=557140 last checked:
01.07.2009.
Heston, A., Summers, R. & Aten, B. (2006), Penn World Table Version 6.2,
Center for International Comparisons of Production, Income and Prices
at the University of Pennsylvania, September 2006 .
Hettige, H., Huq, M., Pargal, S. & Wheeler, D. (1996), Determinants of
Pollution Abatement in Developing Countries: Evidence from South and
Southeast Asia, World Development 24(12), 18911904.
Hettige, H., Mani, M. & Wheeler, D. (2000), Indutrial Pollution in Eco-
nomic Development: Kuznets Revisited, Journal of Development Eco-
nomics 62(2), 445476.
Hoechle, D. (2007), Robust Standard Errors for Panel Regressions with
Cross-sectional Dependence, The Stata Journal 7(3), 281 312.
248
BIBLIOGRAPHY
Homann, R., Ging, L. C., Ramasamy, B. & Yeung, M. (2004), FDI and
Pollution: A Granger Causality Test using Panel Data, CEABuR Work-
ing Papers 2: Centre for Europe-Asia Business Research .
Holtz-Eakin, D. & Selden, T. M. (1995), Stoking the Fires? CO
2
Emissions
and Economic Growth, Journal of Public Economics 57, 85101.
Hsiao, C. (2003), Analysis of Panel Data, Cambridge University Press, Cam-
bridge.
IAEA (2009a), National Projects, http://www-tc.iaea.org/tcweb/
tcprogramme/projectsbycountry/default.asp last checked:
01.07.2009.
IAEA (2009b), Pillars of Nuclear Cooperation, http://www.iaea.org/
OurWork/ST/index.html last checked: 01.07.2009.
ICivDO (2009a), About the International Civil Defence Organization,
http://www.icdo.org/index.html last checked: 01.07.2009.
ICivDO (2009b), Constitution of the International Civil Defence Or-
ganization, http://www.icdo.org/glance-icdo.html last checked:
01.07.2009.
ICSG (2009), Environmental and Economic Activities, http://www.icsg.
org/index.php?option=com_content&task=view&id=37&Itemid=50
last checked: 01.07.2009.
IFAD (2009), IFAD in India, http://operations.ifad.org/web/ifad/
operations/country/home/tags/india last checked: 01.07.2009.
IFC (2009a), Sustainability, http://www.ifc.org/sustainability last
checked: 01.07.2009.
IFC (2009b), What We Do, http://www.ifc.org/ifcext/about.nsf/
Content/WhatWeDo last checked: 01.07.2009.
IIR (2009), About the IIR, http://www.iifiir.org/en/about.php?rub=
1 last checked: 01.07.2009.
ILO (2009), SafeWork, http://www.ilo.org/public/english/
protection/safework/mandate.htm last checked: 01.07.2009.
IMF (2009), The IMF and the Environment, http://www.imf.org/
external/pubs/ft/exrp/environ/ last checked: 01.07.2009.
IMO (2009), About IMO, http://www.imo.org/ last checked: 01.07.2009.
249
BIBLIOGRAPHY
Index Mundi (2009a), Country Comparison: GDP (Purchasing Power
Parity, http://indexmundi.com/g/r.aspx?c=cv&v=65 last checked:
01.07.2009.
Index Mundi (2009b), GDP (Ocial Exchange Rate), http:
//indexmundi.com/world/gdp_(official_exchange_rate).html
last checked: 01.07.2009.
Ingram, P., Robinson, J. & Bush, M. (2005), The Intergovernmental Net-
work of World Trade: IGO Connectedness, Governance, and Embedded-
ness, The American Journal of Sociology 11(3), 824858.
International Tropical Timber Organization (2009), About ITTO, http:
//www.itto.int/en/about_itto/ last checked: 01.07.2009.
Ivarsson, I. & Alvstam, C. G. (2005), Technology Transfer From TNCs to
Local Suppliers in Developing Countries: A Study of AB Volvos Truck
and Bus Plants in Brazil, China, India and Mexico, World Development
33(8), 13251344.
Jacobs, J. W. (1996), Adjusting to Climate Change in the Lower Mekong,
Global Environmental Change 6(1), 722.
Jacobsen, J. & De Soysa, I. (2006), Do Foreign Investors Punish Democ-
racy? Theory and Empirics, Kyklos: Internationale Zeitschrift fr Sozial-
wissenschaften 59(3), 383410.
Jae, A., Peterson, S. R., Portney, P. R. & Stavins, R. N. (1995), Environ-
mental Regulation and the Competitiveness of U.S. Manufactoring: What
Does the Evidence Tell Us?, Journal of Economic Literature 33, 132163.
Jaggers, K. & Gurr, T. R. (1993), Tracking Democracys Third Wave with
the Polity III Data, Journal of Peace Research 32(4), 469482.
Jahn, D. (1998), Environmental Performance and Policy Regimes: Explain-
ing Variations in 18 OECD-Countries, Policy Science 31(2), 107131.
Jahn, D. (2002), The Impact of Globalization on Comparative Analysis, in
ISA Convention, New Orleans.
Jahn, D. (2006), Globalization as Galtons Problem: The Missing Link in
the Analysis of Diusion Patterns in Welfare State Development, Inter-
national Organization 60(2), 401431.
Jahn, D. (2008), The Politics of Climate Change, Paper Presented at the
49th Annual ISA Convention, March 26-29 San Francisco .
250
BIBLIOGRAPHY
Jahn, D. & Walti, S. (2007), Umweltpolitik und Foederalismus: Zur
Klaerung eines ambivalenten Zusammenhangs, in K. Jacob & F. Bier-
mann, eds, Politik und Umwelt, Vol. 262-282, VS Verlag, Wiesbaden.
Javorcik, B. S. & Wei, S.-J. (2004), Pollution Havens and Foreign direct
Investment: Dirty Secret or Popular Myth?, Contributions to Economic
Analysis & Policy 3(2), 132.
Jensen, N. M. (2003), Democratic Governance and Multinational Corpora-
tions: Political Regimes and Inows of Foreign Direct Investment, Inter-
national Organization 57(3), 587616.
Jensen, N. M. (2006), Nation-States and the Multinational Corporation:
A Political Economy of Foreign Direct Investment, Princeton Unviersity
Press, Princeton and Oxford.
Jensen, N. M. & McGillivray, F. (2005), Federal Institutions and Multina-
tional Investors: Federalism, Government Credibility, and Foreign Direct
Investment, International Interactions 31(4), 303325.
Jensen, N. M. & Rosas, G. (2006), Foreign Direct Investment and Income
Inequality in Mexico, 1990-2000, International Organization, Vol. 61, No.
3, 2007 61(3), 467487.
Keohane, R. (1982), The Demand for International Regimes, International
Organization 36(2), 325355.
Keohane, R. (1984), After Hegemony: Cooperation and Discord in the World
Political Economy, Princeton University Press, Princeton.
Keohane, R. (1989), International Institutions and State Power. Essays in
International Relations, Westview, Boulder.
Keohane, R. & Levy, M. A. (1996), Institutions for Environmental Aid:
Pitfalls and Promiss, MIT Press, Cambridge, MA.
Kono, D. Y. (2006), Optimal Obfuscation: Democracy and Trade Policy
Transparency, American Political Science Review 100(3), 369384.
Koubi, V. & Bernauer, T. (2009), Globalization and the Environment, CIS
Working Paper .
Krasner, S. (1983), International Regimes, Cornell University Press, Ithaca
(NY).
Kristensen, I. P. & Wawro, G. (2003), Lagging the Dog?: The Robustness
of Panel Corrected Standard Errors in the Presence of Serial Correlation
and Observation Specic Eects, Paper Presented at the 2003 Summer
Methods Conference .
251
BIBLIOGRAPHY
Krugman, P. R. & Obstfeld, M. (2006), International Economics: Theory
& Policy, Pearson International Edition, Boston.
Kuznets, S. (1955), Economic Growth and Income Inequality, American
Economic Review 49, 128.
La Conference des Ministres de lAgriculture de lAfrique de lOuest et
du Centre (2009), Presentation du PDDAA, http://cmaoc.org/index.
php?option=com_content&task=view&id=158&Itemid=184 last checked:
01.07.2009.
Lake Chad Basin Project (2009), About the Project, http://lakechad.
iwlearn.org/about/.
Lamla, M. L. (2006), Long-run Determinants of Pollution: A Robustness
Analysis, Paper Presented at the International Political Economy Society
Inaugural Meeting, Princeton, November 2006 .
Levy, M. A., Oran, Y. R. & Zuern, M. (1995), The Study of International
Regimes, European Journal of International Relations 1(3), 267330.
Li, Q. & Resnick, A. (2003), Reversal of Fortunes: Democratic Institutions
and Foreign Direct Investment Inows to Developing Countries, Interna-
tional Organization 57(1), 175211.
Li, Q. & Reuveny, R. (2006), Democracy and Environmental Degradation,
International studies Quarterly 50(4), 935956.
Li, X. & Yeung, Y. (1999), Inter-Firm Linkages and Regional Impact of
Transitional Corporations: Company Case Studies From Shanghai China,
Geograska Annaler Series B: Human Geography 81(2), 6172.
Lieberman, E. S. (2005), Nested Analysis as a Mixed-Method Strategy for
Comparative Research, American Political Science Review 99, 435452.
Lohmann, S. (1997), Linkage Politics, Journal of Conict Resolution
41(1), 3867.
Lopez, R. (1994), The Environment as a Factor of Production: the Eects
of Economic Growth and Trade Liberalization, Journal of Environmental
Economics and Management 27(2), 163184.
Lundsgaarde, E., Breunig, C. & Prakash, A. (2006), Instrumental Philan-
thropy: Trade and Geographical Distribution of Foreign Aid, Interna-
tional Political Economy Society Inaugural Conference November 17-18,
2006 .
Maizels, A. & Nissanke, M. K. (1984), Motivations for Aid to Developing
Countries, World Development 12(9), 879900.
252
BIBLIOGRAPHY
Mani, M. & Wheeler, D. (1998), In Search of Pollution Havens? Dirty
Industry in the World Economy, 1960-1995, Journal of Environment and
Development 7(3), 215247.
Manseld, E. D., Milner, H. V. & Rosendor, B. P. (2002), Why Democ-
racies Cooperate More: Electoral Control and International Trade Agree-
ments, International Organization 56(3), 477513.
Manseld, E. D. & Pevehouse, J. C. (2008), Democratization and the
Varieties of International Organizations, Journal of Conict Resolution
52(2), 269294.
Marshall, M. G. & Jaggers, K. (2002), Polity IV Project: Political Regime
Characteristics and Transitions, 1800-2002.
Martin, L. L. (1993), Credibility, Costs, and Institutions: Cooperation on
Economic Sanctions, World Politics 45(3), 406432.
Martin, L. L. (1995), Heterogeneity, Linkage and Commons Problems, in
R. Keohane & E. Ostrom, eds, Local Commons and Global Interdepen-
dence, Sage, London.
McGuire, M. C. & Olson, M. (1996), The Economics of Autocracy and
Majority Rule: The Invisible Hand and the Use of Force, Journal of
Economic Literature 34(1), 7296.
McKinlay, R. D. & Little, R. (1977), A Foreign Policy Model of U.S. Bilat-
eral Aid Allocation, World Politics 30(1), 5886.
McLean, E. (2006), International Organizations and Environmental Aid,
Doctoral Thesis, University of Rochester.
Midlarsky, M. I. (1998), Democracy and the Environment: An Empirical
Assessment, Journal of Peace Research 35(3), 341361.
Milgrom, P. & Roberts, J. (1992), Economics, Organization, and Manage-
ment, Prentice Hall, Englewood Clis, NJ.
Millimet, D. L., List, J. A. & Stengos, T. (2003), The Environmental
Kuznets Curve: Real Progress or Misspecied Models?, The Review of
Economics and Statistics 85(4), 10381047.
Ministry of Environment and Forests (2009), About: Ministry of Environ-
ment and Forests, http://envfor.nic.in/ last checked: 01.07.2009.
Ministry of Environment of the Republic of Lithuania (2001), State of
Environment 2001, http://www.am.lt/VI/en/VI/index.php#a/32 last
checked: 01.07.2009.
253
BIBLIOGRAPHY
Mitchell, R. B. & Deane, G. (2008), Comparing Institutional Inuence: The
Relative Eectiveness of Three Environmental Regimes, Paper Presented
at the 49th Annual ISA Convention, March 26-29 San Francisco .
Montreal Multilateral Fund (2009), About The Multilateral Fund, http://
www.multilateralfund.org/about_the_multilateral_fund.htm last
checked: 01.07.2009.
Moravcsik, A. (2000), The Origins of Human Rights Regimes: Democratic
Delegation in Postwar Europe, International Organization 54(2), 217
252.
Munck, G. L. & Verkuilen, J. (2002), Conceptualizing and Measuring
Democracy, Comparative Political Studies 35(1), 534.
Murdoch, J. C. & Sandler, T. (1997), The Voluntary Provision of a Pure
Public Good: The Case of Reduced CFC Emissions and the Montreal
Protocal, Journal of Public Economics 63(3), 331349.
NAM (2009a), About NAM, http://www.nam.gov.za/background/
background.htm last checked: 01.07.2009.
NAM (2009b), Resolution on the Right to Development, http://www.nam.
gov.za/resolutions/index.html last checked: 01.07.2009.
NATO (2009), The Euro-Atlantic Partnership Council, http://www.nato.
int/issues/eapc/index.html last checked: 01.07.2009.
Naughton, H. (2006), The Impact of Trade on the Environment, University
of Oregon Working Paper .
Network of Aquaculture Centres in Asia-Pacic (2009), About NACA,
http://www.enaca.org/modules/about/ last checked: 01.07.2009.
Neumayer, E. (2001), Do countries fail to raise environmental standards?
An evaluation of policy options addressing regulatory chill, Interna-
tional Journal of Sustainable Development 4(3), 231244.
Neumayer, E. (2002a), Do Democracies Exhibit Stronger International En-
vironmental Commitment? A Cross-Country Analysis, Journal of Peace
Research 39(2), 139164.
Neumayer, E. (2002b), Does Trade Openness Promote Multilateral Envi-
ronmental Cooperation?, World Economy 25(6), 815832.
Neumayer, E. (2003a), Are Left-wing Party Strength and Corporatism
Good for the Environment? Evidence from Panel Analysis of Air Pol-
lution in OECD Countries, Ecological Economics 45, 203220.
254
BIBLIOGRAPHY
Neumayer, E. (2003b), Do Human Rights Matter in Bilateral Aid Alloca-
tion? A Quantitative Analysis of 21 Donor Countries, Social Science
Quaterly 84(3), 650665.
Neumayer, E. (2003c), The Determinants of Aid Allocation by Regional
Multilateral Development Banks and United Nations Agencies, Interna-
tional Studies Quarterly 47(1), 101122.
Nielson, D. L. & Tierney, M. J. (2003), Delegation to International Or-
ganizations: Agency Theory and World Bank Environmental Reform,
International Organization 57(2), 241276.
Nordstroem, H. & Vaughan, S. (1999), Trade and Environment, Special
Studies 4. Geneva: World Trade Organization .
OECD (2008), International Development Statistics, http:
//oberon.sourceoecd.org/vl=479747/cl=41/nw=1/rpsv/ij/
oecdstats/16081110/v77n1/s4/p1 last checked: 01.07.2009.
OLADE (2009), Institutional: Mission and Vision, http://www.olade.
org/indexEn.html last checked: 01.07.2009.
Olson, M. (1982), The Rise and Decline of Nations, Yale University Press,
New Haven.
Olson, M. (1993), Dictatorship, democracy, and development, The Ameri-
can Political Science Review 87(3), 567576.
Organization of American States (2009), About OAS, http://www.oas.
org/main/english/ last checked: 01.07.2009.
OSCE (2009), About OSCE, http://www.osce.org/about/19298.html
last checked: 01.07.2009.
Perman, R. & Stern, D. I. (2003), Evidence from Panel Unit Root and
Cointegration Tests That the Environmental Kuznets Curve Does Not
Exist, The Australian Journal of Agricultural and Resource Economics
47(3), 325347.
Pevehouse, J. C. (2002a), Democracy from the Outside-In? Interna-
tional Organizations and Democratization, International Organization
56(3), 515549.
Pevehouse, J. C. (2002b), With a Little Help from My Friends? Regional
Organizations and the Consolidation of Democracy, American Journal
of Political Science 46(3), 611626.
255
BIBLIOGRAPHY
Pevehouse, J. C., Nordstrom, T. & Warnke, K. (2004), Intergovernmental
Organizations. 1815-2000: A New Correlates of War Data Set. Version
2.1.
Pevehouse, J., Nordstrom, T. & Warnke, K. (2003), Intergovernmental
Organizations, 1815-2000: A New Correlates of War Data Set, http:
//www.correlatesofwar.org/ last checked: 01.07.2009.
Plassmann, F. & Khanna, N. (2006), Household Income and Pollution.
Implications for the Debate About the Environmental Kuznets Curve,
The Journal of Environment and Development 15(1), 2241.
Pl umper, T. & Troger, V. E. (2007), Ecient Estimation of Time-Invariant
and Rarely Changing Variables in Finite Sample Panel Analyses with Unit
Fixed Eects, Political Analysis 15, 124139.
Pl umper, T., Troger, V. & Manow, P. (2005), Panel Data Analysis in Com-
parative Politis: Linking Method to Theory, European Journal of Politi-
cal Research 44, 327354.
Porter, G., Brown, J. W. & Chasek, P. S. (2000), Global Environmental
Politics, Westerview Press, Boulder.
Porter, M. E. (1990), The Competitive Advantage of Nation, Free Press,
New York.
Prakash, A. & Potoski, M. (2006), Racing to the Bottom? Trade, Envi-
ronmental Governance, and ISO 14001, American Journal of Political
Science 50(2), 350364.
Przeworski, A., Alvarez, M., Cheibub, J. A. & Limongu, F. (2000), Democ-
racy and Development: Political Regimes and Economic Well-being in the
World, 1950-1990, Cambridge University Press, New York.
Reinhardt, E. R. (2001), Adjudication without Enforcement in GATT Dis-
putes, Journal of Conict Resolution 45(2), 174195.
Remmer, K. L. (1998), Does Democracy Promote Interstate Coopera-
tion? Lessons from the Mercosur Region, International Studies Quarterly
42(1), 2551.
Revesz, R. L. (1992), Rehabilitating Interstate Competition: rethinking the
Race-to-the-Bottom Rationale for Federal Environmental Regulation,
New York Law Review 64, 12101254.
Roberts, J. T., Park, B. C. & Vasquez, A. A. (2004), Who Raties Envi-
ronmental Treaties and Why? Institutionalism, Structuralism and Par-
ticipation by 192 Nations in 22 Treaties, Global Environmental Politics
4(3), 2264.
256
BIBLIOGRAPHY
Rodrik, D. (1998), Why Do More Open Economies Have Bigger Govern-
ments?, The Journal of Political Economy 106(5), 9971032.
Rogowski, R. (2003), International Capital Mobility and National Policy
Divergence, in M. Kahler & D. A. Lake, eds, Governance in a Global
Economy: Political Authority in Transition, Princeton University Press,
Princeton, pp. 355374.
Rogowski, R. (2006), Trade, Immigration, and Cross-Border Investment, in
B. R. Weingast & D. A. Wittman, eds, The Oxford Handbook of Political
Economy, Oxford University Press, Oxford, pp. 814828.
Rosendor, B. P. & Vreeland, J. R. (2006), Democracy and Data Dissemi-
nation: The Eect of Political Regime on Transparency, Paper Presented
at the International Political Economy Society Inaugural Meeting, Prince-
ton, November 2006 .
Russett, B. & Oneal, J. (2001), Triangulating Peace: Democracy, Interde-
pendence and International Organizations, W. W. Norton & Company,
New York.
SAARC (2009), About Us, http://www.saarc-sec.org/?t=2.5 last
checked: 01.07.2009.
Santos Silva, J. a. M. C. & Tenreyro, S. (2006), The Log of Gravity, The
Review of Economics and Statistics 88(4), 641658.
SEAMEO (2009a), SEAMEO-UN HABITAT HVWSHE Project,
http://www.seameo.org/index.php?option=com_content&task=
view&id=312&Itemid=179 last checked: 01.07.2009.
SEAMEO (2009b), Training Programmes, http://www.seameo.org/
index.php?option=com_content&task=blogcategory&id=77&Itemid=
193 last checked: 01.07.2009.
Selden, T. M. & Song, D. (1994), Environmental Quality and Development:
Is There a Kuznets Curve for Air Pollution Emissions?, Journal of Envi-
ronmental Economics and Management 27, 147162.
Shak, N. & Bandyopadhyay, S. (1992), Economic Growth and Environ-
mental Quality, Working Paper: World Development Report .
Shen, J. (2006), A Simultaneous Estimation of Environmental Kuznets
Curve: Evidence from China, China Economic Review 17(4), 383394.
Sigman, H. (2002), International Spillovers and Water Quality in Rivers: Do
Countries Free Ride, The American Economic Review 92(4), 11521159.
257
BIBLIOGRAPHY
Sigman, H. (2004), Does Trade Promote Environmental Coordination?:
Pollution in International Rivers, Contributions to Economic Analysis
& Policy 3(2), Article 2.
Simmons, B. A. (2000), International Law and State Behavior: Commit-
ment and Compliance in International Monetary Aairs., American Po-
litical Science Review 94(4), 819835.
Simmons, B. A. & Elkins, Z. (2004), The Globalization of Liberalization:
Policy Diusion in the International Political Economy, American Jour-
nal of Political Science 98(1), 171188.
Singer, J. D., Bremer, S. & Stuckey, J. (1972), Capability Distribution,
Uncertainty, and Major Power War, 1820-1965, in B. Russett, ed., Peace,
War, and Numbers, Sage, Beverly Hills, pp. 1948.
Smith, A. (2006), Pernicious Foreign Aid? A Political Economy of Political
Institutions and the Eect of Foreign Aid, in November 17-18, 2006 In-
ternational Political Economy Society Inaugural Conference, Princeton.
South Asia Co-operative Environment Programme (2009), About
Us, http://www.sacep.org/html/about_overview.htm last checked:
01.07.2009.
Staord, S. L. (2000), The Impact of Environmental Regulations on the
Location of Firms in the Hazardous Wast Management Industry, Land
Economics 76, 569589.
Stein, J. V. (2004), Making Promises, Keeping Promises: Democracy, Rat-
ication, and Compliance in International Human Rights Law, Paper
Presented at the Annual Meeting of the American Political Science Asso-
ciation Chicago 2004 .
Stein, J. V. (2005), Do Treaties Constrain or Screen? Selection Bias and
Treaty Compliance, American Political Science Review 99(4), 611622.
Stern, D. I. (2004), The Rise and Fall of the Environmental Kuznets Curve,
World Development 32(8), 14191439.
Stern, D. I. (2005), Global Sulfur Emissions form 1850 to 2000, Chemo-
sphere 58, 163175.
Stone, R. W., Slantchev, B. L. & London, T. R. (2008), Choosing How to
Cooperate: A Repeated Public-Goods Model of International Relations,
International Studies Quarterly 52(2), 335362.
Strang, D. (1991), Adding Social Structure to Diusion Moddels: An Event
History Framework, Sociological Methods and Research 19(3), 324353.
258
BIBLIOGRAPHY
The Colombo Plan (2009), Programme for Public Administration & En-
vironment, http://www.colombo-plan.org/cp/images/cp_pr/cp_pr_
ap.asp?cm=30&cp_num=250 last checked: 01.07.2009.
The Council of the Baltic Sea States (2009), The Council, http://www.
cbss.org/CBSS-The-Council/the-council last checked: 01.07.2009.
The International Lead and Zinc Study Group (2009), Special Projects
- Common Fund Projects, http://www.ilzsg.org/static/projects.
aspx?from=3 last checked: 01.07.2009.
The World Factbook (2009), Cameroon, https://www.cia.gov/library/
publications/the-world-factbook/geos/cm.html last checked:
17.08.2009.
Tinbergen, J. (1962), The world economy: Suggestons for an international
economic policy.
Tooze, R. (1990), Regimes and International Co-operation, in A. J. R.
Groom & P. Taylor, eds, Frameworks for International Co-operation,
Palgrave Macmillan, New York.
Torras, M. & Boyce, J. K. (1998), Income, Inequality, and Pollution: A
Reassessment of the Environmental Kuznets Curve, Ecological Economics
25, 147160.
Troger, V. (2008), Problematic Choices: Testing for Correlated Unit Spe-
cic Eects in Panel Data, Paper prepared for the Political Methodology
Meeting in Ann Arbor 2008 .
Tullock, G. (1993), Rent-seeking, Edward Elgar, Brookeld, VT.
UN (2009a), Millenium Development Goals, http://www.un.org/
millenniumgoals/environ.shtml last checked: 01.07.2009.
UN (2009b), The UN in brief, http://www.un.org/Overview/uninbrief/
about.shtml last checked: 01.07.2009.
UNCTAD (1996), World Investment Report, UNCTAD, Geneva.
UNCTAD (2009), Foreign Direct Investment Statistics, http:
//www.unctad.org/Templates/Page.asp?intItemID=3199&lang=1
last checked: 01.07.2009.
UNDP (2009), Human Development Index, http://hdr.undp.org/en/
statistics/ last checked: 01.07.2009.
UNEP (2001a), India: State of the Environment, http://www.rrcap.
unep.org/reports/soe/ last checked: 01.07.2009.
259
BIBLIOGRAPHY
UNEP (2001b), Laos: State of the Environment, http://www.rrcap.
unep.org/reports/soe/ last checked: 01.07.2009.
UNEP (2009a), About UNEP, http://www.unep.org/ last checked:
01.07.2009.
UNEP (2009b), Ministries of Environment, http://www.unep.org/
resources/gov/MEnvironment.asp last checked: 01.07.2009.
UNEP (2009c), WTO: Environmental Mandate, http://www.unep.org/
un-env/default.asp?gegid=42 last checked: 01.07.2009.
UNESCO (2001), Die Liste der UNESCO-Biospharenreservate, http://
www.unesco.de/brliste.html?&L=0 last checked: 01.07.2009.
UNIDO (2009), Energy and Environment, http://www.unido.org/
index.php?id=905 last checked: 01.07.2009.
Vanhanen, T. (2000), A New Dataset for Measuring Democracy, 1810-1998,
Journal of Peace Research 37(2), 251265.
VASAB (2009a), Mission for VASAB Co-operation, http://www.vasab.
org/ last checked: 01.07.2009.
VASAB (2009b), VASAB 2010 Plus: II. Action During Recent
Years, http://www.vasab.org/documents.php?go=display&ID=39 last
checked: 01.07.2009.
Voeten, E. & Merdzanovic, A. (2009), United Nations General Assem-
bly Voting Data, hdl:1902.1/12379UNF:3:Hpf6qOkDdzzvXF9m66yLTg==
last checked: 01.07.2009.
Vogel, D. (1997), Trading up and governing across: transnational gover-
nance and environmental protection, Journal of European Public Policy
4(4), 556571.
Vogel, D. (2000), Environmental Regulation and Economic Integration,
Journal of International Economic Law 3(2), 265280.
Vogler, J. (1995), The Global Commons: A Regime Analysis, Wiley, New
York.
Vreeland, J. R. (2003), The IMF and Economic Development, Cambridge
University Press, New York.
Wade, R. H. (1997), Greening the Bank: The Struggle over the Environment,
1970-1995, in D. Kapur, J. P. Lewis & R. Webb, eds, In The World Bank:
Its First Half Century, Vol. 2, The Brookings Institution, Washington,
D.C., pp. 611736.
260
BIBLIOGRAPHY
Ward, H. (2006), International Linkages and Environmental Sustainability:
The Eectiveness of the Regime Network, Journal of Peace Research
43(2), 149166.
Ward, H. (2008), Liberal Democracy and Sustainability, Environmental
Politics 17(3), 386409.
Wheeler, D. (2001), Racing to the Bottom: Foreign Investment and Air
Pollution in Developing Countries, World Bank Policy Research Working
Paper No. 2524 .
WHO (2009), Public Health and Environment (PHE), http://www.who.
int/phe/news/en/index.html last checked: 01.07.2009.
Williamson, O. (1985), The Economic Institutions of Capitalism, Free Press,
New York.
Wilson, S. E. & Butler, D. M. (2007), A lot more to do: The Sensitivity of
Time-Series Cross-Section Analyses to Simple Alternative Specications,
Political Analysis 15(2), 101123.
WMO (2009), Programmes Summary, http://www.wmo.int/pages/
summary/progs_summary_en.html last checked: 01.07.2009.
Wooldridge, J. M. (2001), Econometric Analysis of Cross Section and Panel
Data, The MIT Press, Cambridge, Massachusetts.
Wooldridge, J. M. (2003), Introductory Econometrics: A Modern Approach,
Thomson South Western, Mason.
World Bank (2006), World Development Indicators.
World Bank (2009a), About Us, http://web.worldbank.org/WBSITE/
EXTERNAL/EXTABOUTUS/0,,pagePK:50004410
~
piPK:36602
~
theSitePK:
29708,00.html last checked: 01.07.2009.
World Bank (2009b), Country Classication, http://
web.worldbank.org/WBSITE/EXTERNAL/DATASTATISTICS/0,
,contentMDK:20420458
~
menuPK:64133156
~
pagePK:64133150
~
piPK:
64133175
~
theSitePK:239419,00.html last checked: 01.07.2009.
World Bank (2009c), Country Facts: Cameroon, http://web.worldbank.
org/WBSITE/EXTERNAL/COUNTRIES/AFRICAEXT/CAMEROONEXTN/
0,,menuPK:343823
~
pagePK:141132
~
piPK:141107
~
theSitePK:
343813,00.html last checked: 01.07.2009.
World Bank (2009d), India Kerala Rural Water Supply and Environmen-
tal Sanitation Project, http://www-wds.worldbank.org/external/
261
BIBLIOGRAPHY
default/main?pagePK=64193027&piPK=64187937&theSitePK=
523679&menuPK=64187510&searchMenuPK=64187283&siteName=
WDS&entityID=000333038_20090424015538 last checked: 17.08.2009.
World Bank (2009e), Projects & Operations, http://web.worldbank.
org/WBSITE/EXTERNAL/PROJECTS/0,,menuPK:115635
~
pagePK:
64020917
~
piPK:64021009
~
theSitePK:40941,00.html last checked:
01.07.2009.
World Tourism Organization (2009), Sustainable Development of Tourism,
http://www.unwto.org/sdt/fields/en/policy.php?op=2&subop=1
last checked: 01.07.2009.
WTO (2009a), About WTO, http://www.wto.org/ last checked:
01.07.2009.
WTO (2009b), United States Import Prohibition of Certain Shrimp and
Shrimp Products, http://www.wto.org/english/tratop_e/envir_e/
edis08_e.htm last checked: 01.07.2009.
WWF (2009), About Cameroon, http://www.panda.org/who_we_are/
wwf_offices/cameroon/about_cameroon/ last checked: 01.07.2009.
Xing, Y. & Kolstad, C. D. (2002), Do Lax Environmental Regulations
Attract Foreign Investment?, Environmental and Resource Economics
21(1), 122.
Young, O. R. & Levy, M. A. (1999), The Eectiveness of International
Environmental Regimes, in O. R. Young, ed., The Eectiveness of In-
ternational Environmental Regimes. Causal Connections and Behavioral
Mechanisms, The MIT Press, Cambridge, Massachusetts.
Young, Z. (2002), A New Green Order? The World Bank and the Politics
of the Global Environment Facility, Pluto Press, London.
Zanger, S. C. (2000), Good Governance and European Aid: The Impact of
Political Conditionality, European Union Politics 1(3), 293317.
Zarsky, L. (1999), Havens, Halos and Spaghetti: Untangling the Evidence
about Foreign Direct Investment and the Environment, OECD Confer-
ence on Foreign Direct Investment and the Environment, The Hague .
Zeng, K. & Eastin, J. (2007), International Economic Integration and En-
vironmental Protection: The Case of China, International Studies Quar-
terly 51(4), 971995.
Zeng, K. & Eastin, J. C. (2009), Export Destination, FDI Origin, and the
Environmental Performance of Chinese Provinces, Paper Presented at the
50th Annual ISA Convention, February 15-18 New York .
262

Das könnte Ihnen auch gefallen