SHARING WATER: A HUMAN ECOLOGICAL ANALYSIS OF THE CAUSES OF CONFLICT AND COOPERATION BETWEEN NATIONS OVER FRESHWATER RESOURCES

DISSERTATION

Presented in Partial Fulfillment of the Requirements for

the Degree Doctor of Philosophy in the Graduate

School of The Ohio State University

By

Brian E. Green, M.A.

* * * * *

The Ohio State University 2002

Dissertation Committee: Approved by Professor Kazimierz M. Slomczynski, Adviser

Professor J. Craig Jenkins ______Adviser

Associate Professor Edward M. Crenshaw Department of Sociology ABSTRACT

The politics of fresh water in international contexts are becoming increasingly contentious. This study analyzes the effects of development, demographics and ecological factors on international disputes over water. From a human ecology approach,

I develop a model of water conflict that examines the extent to which population growth and density, urbanization, water scarcity and degradation, social organization, inequitable distribution of water, social inequality, and trade affect the likelihood of conflict over water. Using water event data from the Basins at Risk section of the Transboundary

Freshwater Dispute Database (Wolf 1998; Yoffe 2002) and ordinary least squares regression modeling, I tested hypotheses that specified predictors of international water conflict and cooperation. Field notes from a case study of the international dispute between Slovakia and Hungary were also analyzed.

The results of the analysis indicate that, of demographic predictors, population density has the clearest and most consistent association with international water conflict and cooperation. Countries with higher population densities have more frequent international water interactions of a more conflictual nature. Population growth and urbanization are also found to be associated with water conflict in various predicted ways.

Indicators of development tend to be associated with reduced levels of international water conflict, however, in the case of international inequality of development, water conflict is

ii more likely. Among environmental factors, several indicators of water degradation and depletion were associated with an increase in the level of international water conflict, however these findings were somewhat inconsistent. Inequality in terms of the amount of internally available water was consistently associated with higher levels of conflict. A surprising and counterintuitive finding is that countries that sign international water treaties continue to have water events of a conflictive nature after the treaty is signed. In the case study of Hungary and Slovakia, environmental degradation and depletion was found to increase environmental activism, which had the effect of destabilizing the national government of Hungary. After regime changes in both Hungary and Slovakia, international conflict increased due to the new political openness associated with democracy. Institutional mechanisms have since been established to reduce the intensity of the earlier dispute.

iii For my daughter, Hannah J. Green

iv ACKNOWLEDGMENTS

I would like to acknowledge and thank all of the people who helped me in big and small ways to successfully complete the Ph.D. program at Ohio State University. First and foremost I must thank my advisor, Dr. Maciek Slomczynski. Maciek worked with me and supported my studies for the five years during which I was in the program at Ohio

State. Maciek’s expertise in comparative research methods and social change, and his willingness to meet with me many times over many years, made this entire project possible. More than his specific knowledge of my research areas, Maciek’s overall mentoring and consistent interest in my academic and personal development helped me to want to strive for success at all levels.

I thank my dissertation committee members Drs. Craig Jenkins and Edward

Crenshaw. Through the classes I took with them and the many hours spent in offices and writing e-mail, I learned tremendously from both men. Craig and Ed provided crucial comments, letters of recommendation, and other support during my years at Ohio State and through the process of writing this dissertation. The quality of their scholarship is an inspiration to me.

There are many others who deserve recognition and thanks for their support in this project. Drs. Aaron Wolf and Shira Yoffe, both of Oregon State University, truly helped

v to make this entire project through academic support and especially by sharing the data from the Transboundary Freshwater Dispute Database. Both Aaron and Shira, along with their colleagues, have paved the way for the kind of quantitative analysis of freshwater disputes presented in this dissertation. Among others who helped directly with the completion of this study, Matt Moffitt of the Sociology Research Lab deserves the most thanks for helping me to write the programs that imported my data into a usable format.

Colin Odden, also of the SRL, deserves thanks for his help in that area.

Financial support for this dissertation study was provided by the Phyllis J. Krumm

Memorial Scholarship fund, the Graduate Student Alumni Research Award, and the

Department of Sociology at Ohio State. I thank all of these organizations for their assistance.

Tony Vigorito was a great friend during my time in graduate school. Together we went through the process of writing our dissertations and looking for academic jobs. I thank Tony for his companionship and emotional support during the past years.

Other friends who supported me both personally and professionally during work on this dissertation include Jack Selig and Dagmar Ruskova. Jack helped me through many difficult times and always inspired me to keep going and do my best. Dagmar wrote to me nearly every day and always reminded me “you can do it.”

Finally, my parents, Steve and Joyce Green, have always supported me in so many ways. Without them I could never have made this achievement.

vi VITA

April 9, 1970 ...... Born – Baltimore, Maryland USA

1992 ...... B.A. Sociology, Pennsylvania State University

1996–1997 ...... Graduate Teaching Assistant, Kent State University

1998 ...... M.A. Sociology, Kent State University

1997 – present ...... Graduate Teaching Associate, The Ohio State University

2000–2001 ...... Program for the Enhancement of Graduate Studies Dissertation Fellow

PUBLICATIONS

1. Green, Brian E. and Christian Ritter 2000. “Marijuana Use and Depression.” Journal of Health and Social Behavior 41 (March): 40-49.

2. Gregory, Stanford, Brian E. Green, Robert M. Carruthers, Kelly A. Dagan, and Stephen Webster. 2001. “Verifying the Primacy of Voice Fundamental Frequency in Social Status Accommodation.” Language and Communication 21(1): 37-60.

FIELDS OF STUDY

Major Field: Sociology Studies in: Comparative Social Change, Environmental Sociology

vii TABLE OF CONTENTS

Abstract ...... ii

Dedication ...... iv

Acknowledgments ...... v

Vita...... vii

List of Tables...... xi

List of Figures ...... xii

Chapters:

1. Introduction ...... 1

2. Theoretical Background ...... 5

The Social Science of Natural Resource Competition ...... 7 Classic Theoretical Perspectives on Natural Resource Distribution and Conflict...... 8 Contemporary Ecological Positions...... 11 The Roles of Population, Organization, Environment and Technology ...... 12 Population Dynamics ...... 13 Social Organization...... 14 The Environment...... 17 Technology ...... 18 Conclusion...... 20

3. Conflict and Cooperation Over International Freshwater as a Research Problem...... 21

Current Locations Where Water is Disputed...... 21

viii The Effects of Human Ecological Variables on International Water Conflict ...... 26 Population Dynamics ...... 26 Social Organization...... 27 The Environment...... 28 Technology ...... 29 Conclusion...... 30

4. The General Model and Hypotheses ...... 31

The Causes of International Freshwater Conflict...... 32 Demographic Impacts on International Water Conflict...... 33 Development and Economic Effects on International Water Conflict...... 34 Ecological Effects on International Water Conflict...... 36 Geo-political Effects on International Water Conflict ...... 36

5. Methods and Data ...... 40

Data Sources ...... 41 Freshwater Conflict...... 41 Independent Variables...... 43 Measures ...... 43 Water Conflict and Cooperation ...... 43 Demographic Indicators...... 47 Development and Political Variables ...... 48 Ecological Variables ...... 49 Final Data Set Construction...... 50

6. Results...... 61

The Link Between the Number of Interaction Events and the Level of Conflict/Cooperation ...... 62 The Frequency of International Water Conflict and Cooperation Events ...... 63 Development and Demographic Variables ...... 63 Export Trade ...... 66 Ecological Effects ...... 69 Treaties...... 78 Conflict and Cooperation ...... 79 Development and Demographic Variables ...... 80 Ecological Variables ...... 81

ix 7. The Role of Institutions and Organizations ...... 100

The Old World Order Vs. The New World Order ...... 101 The Slovakia–Hungary Case Study ...... 101 Gab íkovo–Nagymaros ...... 102 The Effects of Macro and Micro Organizational/Institutional Factors on Water Conflict in the Case of Hungary and Slovakia ...... 104

8. Conclusion ...... 106

Summary of Results ...... 106 Implications for Theory ...... 112 Implications for Policy ...... 114 Concluding Comments ...... 115

Appendices ...... 116

A. Syntax from Visual Basic Program Used to Import and Summarize Water Conflict and Cooperation Event Data from the Basins at Risk Data Set ...... 116 B. International River Basins ...... 121 C. Syntax from Visual Basic Program Used to Import Development, Demographic, and Ecological Variables into Working Data Set...... 139 D. Country Pairs in the Data Set ...... 144 E. Standard International Country Codes ...... 145

References ...... 149

x LIST OF TABLES

Table Page

5.1 Basins at Risk Event Database Example ...... 53

5.2 Modified COPDAB Scale Items ...... 54

5.3 Descriptive Statistics for All Variables in the Study ...... 57

5.4 The Number of Times Each Country Appears in the Final Data Set ...... 59

6.1 Regression of the Natural Log of the Number of International Water Events During Five Year Time Periods on Development and Demographic Variables ...... 86

6.2 Regression of the Natural Log of the Number of International Water Events During Five Year Time Periods on Development and Demographic Variables Including Export Trade Data ...... 87

6.3 Regression of the Natural Log of the Number of International Water Events During Five Year Time Periods on Ecological Variables, Controlling for Development and Demographic Variables ...... 88

6.4 Combined Regression Models of Logged Number of International Water Events During Five Year Time Periods on Ecological Variables and Development and Demographic Variables ...... 90

6.5 Regression of the Natural Log of the Number of International Water Events During Five Year Time Periods on Water Treaty Existence, Controlling for Development and Demographic Variables ...... 91

6.6 Regression of Average Level of Conflict or Cooperation Between Countries During Five Year Time Periods on Development and Demographic Variables ...... 92

xi 6.7 Regression of Average Level of Conflict or Cooperation Between Countries During Five Year Time Periods on Ecological Variables, Controlling for Development and Demographic Variables ...... 93

6.8 Combined Regression Models of Average Level of Conflict or Cooperation Over Water Between Countries During Five Year Time Periods on Ecological Variables and Development and Demographic Variables ...... 95

6.9 Regression of the Average Level of Conflict or Cooperation Over Water Between Countries During Five Year Time Periods on Water Treaty Existence, Controlling for Development and Demographic Variables ...... 96

xii LIST OF FIGURES

4.1 List of Hypotheses ...... 38

5.1 Distribution of Water Conflict and Cooperation Events ...... 60

6.1 Scatterplot of Average Level of Conflict and Cooperation by the Number of Water Events ...... 97

6.2 Graph of Average Conflict Score by Number of Events and Year ...... 98

6.3 Scatterplot of the Number of Water Events by Population Growth Rate ...... 99

6.4 Scatterplot of the Number of Dyads for Each Country by the Population Density ...... 101

xiii CHAPTER 1

INTRODUCTION

In the course of history, human societies have often challenged one another for the right to exploit natural resources. Indeed, the boundaries of most current nation states were defined based on either historical rights to the resources of certain regions or on the outcome of a war fought over possession of resource rich areas. During most of the twentieth century, after most state boundaries were established, wars and conflict tended to arise over ideological and political disagreements rather than scarce resources. The last thirty years, a period that included the end of the Cold War, have seen unprecedented levels of population growth, urbanization, and development worldwide, which have often contributed to the depletion and degradation of various natural resources. Because of the resultant social and ecological changes from these trends, and also in part due to a weak legal framework for addressing international resource disputes, many scholars and politicians predict that we will see an increasing number of disputes between countries over shared natural resources.

Fresh water, perhaps the most important resource to humans, is becoming increasingly scarce in many places around the world. Furthermore, water, which is

1 essential for agriculture, industry, and household use, is frequently shared across political borders, and this shared use is often the source of disagreement between countries.

Nations who share large bodies of fresh water, such as rivers, lakes, and underground aquifers, may increasingly come into conflict over water resources, as demands for water grow in the future. The goal of this study is to explore how and to what extent developmental, demographic, and ecological factors contribute to an increased likelihood of nations coming into conflict with one another over sources of fresh water. And, on the other hand, under what circumstances are we likely to see cooperation over water between nations who share it?

The results presented in this study, which attempt to illuminate the processes of international conflict and cooperation over natural resources, are from an analysis of events that were in regards to water and that occurred between pairs of countries during the period covering roughly from 1970 to 2000. Building from a theoretical model based on human ecology theory, quantified data was compiled on a number of variables measuring aspects of development, demographic, and ecological factors that were predicted to affect the level of conflict or cooperation between water sharing pairs of countries. Additionally, a case study based on field work conducted in Central Europe is also analyzed.

The analyses conducted in this study represent a break from previous investigations into trans-boundary water conflicts. Nearly all previous work has been in the form of case study analyses of one or several locations where water has been a cause of disputes. This study uses quantified data on a large number of country pairs to test

2 multivariate models that produce results that are generalizable at the macro-historical level. Two indicators of water interaction events are used throughout the study. One measure assesses the number of interactions that occurred between water sharing countries during specified time periods. This indicator provides us with an insight into what factors are associated with the occurrence of international water negotiations events.

The occurrence of water events reflects the salience of water issues to the international diplomacy between two countries. The second measure of water interaction events assesses the relative conflictive or cooperative nature of events that occurred between countries during five-year time intervals. Analysis of the factors associated with this measure show more specifically in what contexts we are likely to find conflict or cooperation between countries over fresh water.

The results of the study provide support for my hypothesis that the level of conflict or cooperation over fresh water in trans-boundary contexts, as well as the frequency of international water interaction, is affected by trends in development, demographics, and ecological change. More specifically, the results show that countries at higher levels of development tend to have more cooperative interaction over water, however countries with higher population density and higher levels of urbanization tend to have more conflictive interactions over water. The findings also show that water availability and water quality affect the level of conflict or cooperation between countries that share water. We observe events of more conflictive nature particularly in the case of inequality between water sharing countries in the amount of available water. Poor water quality is also found to be associated with conflict between countries that share water. A

3 surprising and counterintuitive finding of the study is that after countries sign a water treaty we do observe a decrease in the intensity of water interactions, however the interactions that do continue to occur are on average more conflictive than the interactions that occur between countries that do not have a water treaty.

These findings, and the analysis that lead to them, represent an important contribution to the broad interdisciplinary study of natural resource conflict. In addition to supporting the theoretical framework of the human ecology approach to resource conflicts, the multivariate analysis of a large cross-national data set represents a break from most previous studies of trans-boundary resource disputes and increases the overall generalizability of the conclusions.

This dissertation is organized in following way. The next chapter discusses the theoretical background for studies in natural resource conflicts. The third chapter discusses international cooperation and conflict over fresh water as a research problem.

The fourth and fifth chapters describe the hypotheses and the research model, and the methods and data used in the study. The sixth chapter describes the results of the quantitative analyses completed in the study. The seventh chapter describes the results of field work conducted on the conflict between Slovakia and Hungary regarding the

Gabcikovo-Nagymaros dam on the Danube River. Finally, the last chapter offers a discussion and some concluding remarks on the implications of the findings presented herein.

4 CHAPTER 2

THEORETICAL BACKGROUND

Recent international confrontations over oil reserves, freshwater supplies, and trans-boundary pollution indicate the potential for future disputes over natural resources

(Kamyar 2000; Kinzer 1999; Verhovek 1998). Supply and demand driven increases in the prices of gasoline and natural gas have also recently lead to political maneuvering as nations try to assure their continued supply of these valuable fuels (Ottoway 2000). This chapter explores the factors which may affect the likelihood of future conflicts over natural resources.

Hardin’s (1968) discussion of the “tragedy of the commons” points out the dilemma that human societies face as population and demand outstrip the supply of natural resources such as land, forests, and minerals. The tragedy is that the commons system works as long as human demands on the commons do not exceed the ability of nature to replenish itself; however, when demands do exceed the ability of the commons to replenish itself the system breaks down and will eventually collapse (Leal 1998).

Recognizing the potential for system collapse, human actors have the choice to voluntarily limit their use of the commons so that it will sustainably yield natural

5 resources. Since the choice to limit resource withdrawals is voluntary in a commons system, many individuals will choose not to limit their use of the commons so that they can gain a relative economic advantage compared to those who refrain from overuse.

Thus, in a commons system that experiences continuous growth of population and consumption levels, competition and conflict are likely outcomes.

The historical response to the tragedy of the commons has been privatization of land and resource ownership. Shiva (1991), however, points out that the transition from natural resource ownership held in common to the contemporary private property system is a process often filled with conflict. Furthermore, some natural resources are not easily privatized, such as rivers, oceans, the atmosphere, and wild animals. Weakly developed international law on the uses of oceans, rivers, the atmosphere and other resources has resulted in a system where nations use these resources in common, with few incentives to conserve or preserve them. Consequently, competition over these jointly held resources is increasing as countries attempt to ensure their shares of commonly held resources and gain relative economic advantages.

Social scientists have developed numerous theories which address the social, economic, demographic, and ecological causes of competition over natural resources.

Conflict over natural resources is defined broadly throughout the following discussion as disputes between individuals, groups, or nations over the right to utilize natural resources or over the responsibility of protecting and conserving the integrity of natural resources.

Humphrey and Buttel (1982) claimed that understanding the extent to which natural resource scarcity will cause struggles over access to limited resources is a critical issue in

6 environmental sociology. Furthermore, Namboodiri (1988) stated that the human ecological approach is a powerful perspective in the analysis of power relations and conflict processes. The human ecology approach is applied throughout this study to organize an attempt to determine how key variables affect the likelihood of conflicts over natural resources. In the next chapter a discussion of the factors which may cause international conflict and cooperation over freshwater resources is used to predict the effects of development, demographic, and ecological variables on natural resource conflict, and to review how the human ecology perspective can help explain such conflict.

The following section reviews literature which situates natural resource conflict studies within the social sciences generally, and within sociology specifically.

THE SOCIAL SCIENCE OF NATURAL RESOURCE COMPETITION

Several theoretical traditions in social science discuss the relationship between human societies and the environment. In this section I begin by briefly summarizing some classical theories which postulate on the demographic, economic, social, and ecological factors that affect the likelihood of conflict over natural resources. Then, I discuss some contemporary ecological positions which criticize and update the classic approaches and I discuss why an ecological approach to the study of natural resource conflicts is a perspective which provides a parsimonious understanding of this problem.

In the last part of this section I review a number of studies which present findings on the roles of ecological variables in causing or preventing conflicts over natural resources.

7 CLASSIC THEORETICAL PERSPECTIVES ON NATURAL RESOURCE

DISTRIBUTION AND CONFLICT

Traditional Malthusian theory suggests that—due to population growth—human consumption needs will eventually exceed the availability of natural resources, particularly food, causing a myriad of negative social outcomes like war, disease, and famine. Violence and war, from the Malthusian perspective, are “positive checks” that serve to reestablish the equilibrium that is disrupted by scarcity caused by population growth. According to Price (1998), Malthus’ theoretical statement was, simply, that population expands to the limits imposed on it by subsistence. The inevitable results when it reaches those limits are poverty and disaster. The traditional Malthusian perspective has been criticized for neglecting the role of technological innovation and other factors in increasing the carrying capacity of the world (for example, see Barnett

1974).

Classical economics theories on economic behavior and natural resources have emphasized the creation of markets as the key to balancing positive development and over-consumption. Adam Smith (1937 [1778]) suggested that a distributive system based on supply and demand could bring about a dynamic arrangement capable of effectively addressing scarcity. Simply put, as resources become more scarce their price goes up, thereby deterring over-consumption and spurring technological developments and substitutions. Critics of classical economics have pointed out that as natural resources become more scarce and their value goes up, the financial incentive to further exploit those resources also increases—to the point where the result is often extinction,

8 disappearance or devastation (Clark 1973; Davidson 1999). Trade in elephant ivory, leopard skins, and rhinoceros horns, and the destruction caused to the species who give us these things illustrate this point. Furthermore, many criticize the vast potential for divergent accumulation of wealth under the free-market system such that some gain control over much while the rest have little (Trainer 1998).

Following from that criticism, theories in the Marxist tradition have emphasized the conflicts of interest between groups with more or less control and ownership of natural resources. These approaches argue that free markets create such great disparities between the “haves” and the “have-nots” that social conflict is inevitable (Marx and

Engels 1962 [1848]). According to Dobkowski and Wallimann (1998: 4), the economic transformation of the last two centuries has “always been financed on the back(s) of peasants both in the nineteenth century, in the area that is now known as the center, and in the twentieth century, in the area now known as the periphery.” Contemporary scholars in the dependency and world-systems schools have further discussed how wealthy nations exploit other countries for their natural resources (Baran 1957; Chirot and Hall 1982; dos

Santos 1971; Wallerstein 1979). Critics of Marxist approaches, however, have included social Darwinists, who believe that social rewards are acquired based on merit or inherent ability (Spencer 1883), and functionalists, who feel that social hierarchy is a requisite aspect of modern society (Davis and Moore 1945).

In the classic sociological tradition, which emphasizes the effects of industrialization on human relations, Durkheim examined how macro-structural changes in human social organization affected social adaptability. According to Durkheim (1965

9 [1902]), population growth and its concomitant growth in competition for resources helped bring about the industrial era and the complex division of labor associated with it.

Furthermore, Durkheim contended that the division of labor in complex society increases social adaptability, thereby reducing conflicts between classes (Harper 1996; Humphrey and Buttel 1982). Marxists and critical theorists have criticized the Durkheimian view that organic solidarity is a stabilizing force strong enough to prevent conflict between the classes.

The classical social science approaches, with the notable exception of the

Malthusian tradition, have implicitly supported the notion that humans are in many ways exempt from the natural limits to growth prescribed by a finite resource base. For example, classic economic approaches emphasize the point that free and open markets, along with human ingenuity, can overcome natural resource limits by redefining the resource base. Marxist and conflict approaches generally claim that the scarcity of resources is not what limits development but rather it is the inequitable distribution of these resources among the members of society. Furthermore, most classic sociological approaches, particularly those in the functionalist tradition, emphasize the importance of higher order social structures and institutions as key to the adaptability of societies.

Social scientific approaches which de-emphasize the limits that finite natural resources may hold on society are commonly labeled “human exemptionalist” approaches (Dunlap and Catton 1979). Such perspectives are, in general, unlikely to claim that natural resource scarcity will directly result in social conflict and disorder in free-market,

10 industrialized societies. Rather, “human exempt” approaches predict that scarcity can be overcome and is generally not a primary cause of social disorder.

CONTEMPORARY ECOLOGICAL POSITIONS

Ecologists have been highly critical of the human exemptionalists’ belief that economic and technological development overcome our dependence upon natural resources. Shiva (1991: 13) describes the belief that technology reduces human dependence on natural resources as a myth that overlooks the “long and indirect chain of resource utilisation which leaves invisible the real material resource demands of the industrial processes.” The “new ecological” approach (Dunlap and Catton 1979), which has arisen in the last half century and gives more importance to the natural environment as a salient factor in social development, argues that there are limits to growth and that human societies cannot rely on market adaptations to overcome these limits.

Human ecologists and other environmental social scientists since the 1960’s have stressed the importance of the interactions between nature and society. Recently, some of these environmentalists have posited that the deterioration of natural systems like water, air and soil, could have negative effects on social, political, and ecological security

(Dabelko 1996; Gleick 1993; Homer-Dixon 1991,1994; Ophuls 1977; Sosland 1998;

Swart 1996). While there is a legitimate, ongoing debate over whether economic and technological development will solve the disputes that occur over natural resources, there are several reasons why the ecological approach is a useful theoretical approach to take when studying the causes of natural resource conflicts.

11 Human ecology is primarily interested in building a conceptual framework for studying the relationship between humans and the environment, with an emphasis on the interactions between populations, social organization, environment, and technology

(Duncan 1959; Namboodiri 1988). As such, human ecologists study systems of interdependence between humans and the natural environment. McKenzie conceptualized human ecology as a “study of the spatial and temporal relations of human beings,” as they are “affected by the selective, distributive, and accommodative forces of the environment (Hawley 1968: 329).” This conceptualization was furthered by Hawley

(1968: 329), who conceived of human ecology as the study of the “form and development of the human community,” with community “construed as a territorially localized system of relationships with functionally differentiated parts.” More specifically, Hawley stated that “human ecology ... is concerned with the general problem of organization ... as an attribute of a population (329).” By focusing on the interactions between population, organization, environment and technology, the human ecological perspective attempts to integrate knowledge of various aspects of the human and natural ecosystems into a comprehensive understanding of the processes which contribute to social stasis or change

(Duncan 1959).

THE ROLES OF POPULATION, ORGANIZATION, ENVIRONMENT AND

TECHNOLOGY

Park and Burgess, in an early discussion of human ecology, stated that competition and conflict are sources of bringing about equilibrium in human systems

12 (Humphrey and Buttel 1982). In this sense, competition and conflict can be thought of as

“functional” for the human ecological system. Nevertheless, we endeavor to find ways of preventing intolerable levels and types of conflict. Many scholars speculate on the relationships between human ecological variables and social conflict.

In this section I review and summarize literature which comments on the interactions of population dynamics, social organization, environmental factors, and technology and how these are related to conflict over natural resources. The human ecology perspective emphasizes the interactions of the POET variables. Therefore, I discuss the direct and indirect effects of each of the four main variables on natural resource conflict.

POPULATION DYNAMICS

Population growth influences social organization, the environment, and technology and in various ways may cause conflict over natural resources. Some have found that simple population growth is associated with higher levels of social conflict, net of other factors (Eliot 1972; Goldstone 1980). However, the greatest effects of population factors on natural resource conflict are most likely due to indirect effects on aspects of social organization, environmental impacts, and technology which further increase the chance of conflict.

Population growth interacts with aspects of social organization to increase conflict. For example, population growth in conjunction with the spread of industrial capitalism is associated with increased energy consumption per capita. This exacerbates environmental problems, because even improvements in efficiency and technology cannot

13 eliminate growth in the absolute values of pollution worldwide (Duchin 1996, Meyerson

1998). Rapid population growth in societies not equipped to deal with it is likely to lead to aggravated inequities, increased competition, and fighting over the resources that are within reach (Goldstone 1999).

Population growth also has negative impacts on the environment which may lead to conflict in several ways. Pimental et al. (1997), Pimentel and Giampietro (1994), and

Bongaarts (1996) point out that population growth and density lead to scarcity of land and food. Food scarcity can contribute to political unrest and inequality as well as further degrade the environment by exacerbating deforestation, soil salinization, water pollution, and biodiversity loss. Whitmeyer and Hopcroft (1996) cite the Zapatista rebellion in

Chiapas, Mexico as an example of population growth combining with land and food scarcity to cause violent confrontation. Similar examples abound and include the Kurds in the Middle East and the Miskitos in Central America (Gurr and Harff 1994).

Population growth and environmental degradation and depletion also contribute to the displacement or migration of peoples—who have been called “environmental refugees” (Myers 1997). Displacement and migration contribute to ethnic competition and conflict when ethnically distinct groups try to exploit the same ecological niche

(Calhoun 1993; Olzak 1992; Williams 1994). This is a serious problem, particularly in

Africa and the Middle East where herders and farmers are increasingly relying on the same land (Rahim et al. 1991).

14 SOCIAL ORGANIZATION

We must also look at the effects of social organization on natural resource conflict. Proponents of the expansion of global free trade discuss the “Liberal peace,” in which it is assumed that open and fair markets discourage conflict because it is unprofitable (Oneal et al. 1996). However, the current forms of social organization interact in various ways with population issues, the environment, and technology to influence the likelihood of conflict. For example, Arizpe and Velázquez (1994) argue that we cannot limit discussion of the effects of population on the environment to matters of population size, density, rate of increase, or migration, but rather, we must consider issues such as access to resources, livelihoods, and other social structural issues like gender and power. The global industrial-capitalist economic system advanced by the most powerful countries in the world today is characterized by several features, such as high social inequality, global trade patterns, and increased levels of consumption and resource extraction associated with wealth, that potentially increase the effects of population growth, density, and migration on the environment and conflict.

Some have argued that the transition from traditional to modern economies is a process filled with conflict. Furthermore, many see the current political world economy as a source of conflict within and between countries. Shiva (1991) discusses how the development of India, from colonization to independence, has caused conflict. For example, the radical changes from traditional commons systems to industrial and market capitalism caused numerous conflicts, as traditional modes of sustenance and resource

15 allocation were outlawed or outmoded. Water, forests, and land remain highly contentious in India.

On the other hand, some theorists argue that the world is now undergoing a process of ecological modernization in which the negative effects of industrialization on the environment are being reversed in the most economically and technologically developed countries. These theorists suggest that while detrimental impacts on the environment were magnified during the early and middle stages of industrialization, societies that are in the advanced stages of post–industrialism have been able to abate negative environmental damage through effective policy implementation and improved efficiency, among other reasons. Some evidence has been found to show that efficiency in energy consumption and air pollution improves when countries cross a threshold where major infrastructure projects are completed and the economy has shifted toward service sector predominance. Following this argument, we may speculate that the relationship between industrial–capitalist development and conflict over natural resources is curvilinear rather than linear: during initial and middle stages of development environmental conflict may increase, but at later stages of development countries are more able to resolve environmental problems which lead to conflict, thereby reducing the chance of conflict.

In spite of the ecological modernization perspective, the conflict oriented approach of Schnaiberg and Gould (1994) posits that industrial capitalism and the environment are at enduring odds. Conflicts occur between staunch environmentalists, who fight to protect the integrity of environmental systems, and staunch capitalists, who

16 advocate maintaining the status quo—consuming resources as long as it is profitable to do so. Furthermore, Schnaiberg and Gould argue, the capitalist system may cause inter- or intra-state confrontations over solutions to environmental degradation and responsibility for repairing ecological damage, as well as cause confrontations over the wide inequities that exist in resource distribution.

Other theorists also examine the effects of the spread of global capitalism on the environment and conflict. Smith and Sauer-Thompson (1998) believe that in focusing on

“growth” in the quantitative sense rather than “development” in the qualitative sense, the industrial-capitalist system undermines human security by harming the environment, by expanding social stratification, and by degrading the individual—treating him as a “cost factor” rather than as a human being. Homer-Dixon (1999) refers to inequitable access to valuable resources as “structural scarcity.” Structural scarcity makes certain regions of the world more insecure due to their heavy external debt and dependence on raw material exports. Malhotra (1998) discusses how downturns in the international economy contribute to extensive natural resource conflict in Southeast Asia as countries there try to export their way out of debt.

Barbosa (1996) points out a case where the international order actually helped to reduce natural resource conflict. Prior to the environmental consciousness of the late

1980’s and 1990’s, the Brazilian government ignored the rights of indigenous people and clear-cut the Amazon rain forest to the dismay of environmental and social activists.

Changes in the international economic and political systems which favored preserving natural resources and giving property rights to indigenous groups forced Brazil to change

17 its policies and take a more cooperative stance. Nevertheless, rapid deforestation is still a problem in the Amazon.

THE ENVIRONMENT

The previous discussions on population dynamics and social organization have touched on many of the ways in which environmental factors may contribute to conflict.

Specifically, a number of demographic, organizational and technological factors can cause the degradation and depletion of natural resources in various forms, such as deforestation, global warming, desertification, and pollution. Such degradation and depletion of resources can cause stress and conflict within human communities.

Two forms of conflict over natural resources are likely to occur as a result of environmental degradation and depletion. These are 1) competition over the right to exploit resources and 2) disagreements over the importance of protecting and cleaning up the environment. Examples of the first type of conflict include confrontations, sometimes violent, over the right to harvest old growth forests or to hunt endangered species. Other examples might include disputes between clans over watering holes, as wells dry up due to drought associated with global warming. Examples of the second type of conflict include the debate between the developed and the less developed worlds over who should be responsible for the burden of cleaning up the global environment (Porter and Brown

1996) and the debate between industrialists and staunch environmentalists over the types of ecological withdrawals and additions which are acceptable for maintaining a healthy environment (Schnaiberg and Gould 1994).

18 TECHNOLOGY

Although environmental damage is often caused by the introduction and use of new technologies, it is also often the source of new technologies, which are created to help alleviate some of the social problems associated with depletion and degradation.

The development of technologies are in many ways connected to changes in population dynamics and social organization (Boserup 1981). An increase in population creates environmental problems to be solved, but it also means more minds to ponder solutions to a problem. So population growth sometimes necessitates new technologies, but new technologies allow even more population growth. Increases in population, population density, and technology cause environmental and organization changes, as people consume more of their natural resources and need new social institutions to address new concerns.

Some theorists in the ecological modernization school speculate that an environmental Kuznet’s curve—that is the reduction of negative environmental impacts in later stages of development—is the result of increasing technology (Crenshaw and

Jenkins 1996; Ehrhardt-Martinez, Crenshaw, and Jenkins 2002; Sonnenfeld 2002). The argument is that while environmental damage was severe in the earlier stages of industrialization, technological developments, combined with changes in social organization associated with post-industrial society, have improved environmental efficiency and stabilized, or even reversed, the amount of harm caused to the natural world. There is some debate over the extent to which this is occurring and, historically,

19 developments in technology have both alleviated and created conflicts over natural resources (Roberts and Grimes 1997; Stern, Common, and Barbier 1996).

Two examples illustrate the potential for new technology to cause conflict. The current debates over genetically modified foods and nuclear energy show how attempts to address scarcity through technological change can lead to serious disputes. Food and energy scarcity has inspired scientific developments in the forms of new types of engineered food crops and nuclear power. Opponents of these technologies have demonstrated against the industries which provide the new products and the economic institutions which favor their application (Shaw 1999; Walsh 1985). This opposition culminated in the protests against the policies of the World Trade Organization and the

International Monetary Fund at Seattle in 1999. These examples show how population growth and social organizational changes lead to changes in environmental and technological circumstances, which create social conflict over natural resources.

CONCLUSION

In this chapter I have discussed the nature of contemporary conflicts over natural resources. I described several social scientific approaches to understanding natural resource conflict and I detailed how the human ecology model can be applied in these analyses. In the next chapter I discuss cases of international freshwater conflict and the relationships of development, demographic, and ecological variables with conflict and cooperation between nations over fresh water resources.

20 CHAPTER 3

COOPERATION AND CONFLICT OVER

INTERNATIONAL FRESHWATER AS A RESEARCH PROBLEM

A large body of literature in political science, geography, and development studies has discussed cases of current and potentially ongoing international freshwater conflicts

(see for example: Donahue and Johnston 1998; Gleick 1993; Haftendorn 2000; Postel

1996; Samson and Charrier 1997; Scheuman and Schiffler 1998). Because water is a valuable resource which frequently crosses political borders, the right to exploit water resources has often come into dispute. The literature on this subject has extensively speculated on the causes of international water disputes. In this section I describe a number of locations where internationally shared fresh water resources have been contentious, and then I describe the findings of research which illustrates the ways in which population dynamics, social organization, the environment, and technology can potentially interact to cause international conflict over water.

CURRENT LOCATIONS WHERE WATER IS DISPUTED

Numerous major water systems throughout the world are experiencing problems related to water usage that may lead to conflict. The Jordan River, the Tigris and

21 Euphrates system, the Ganges, the Nile, the Aral Sea region, the Danube, and the

Colorado river systems are all experiencing stress in terms of increasing demand for water and increasing tension over water rights. These systems may be considered “hot spots” and can provide a basis for a discussion about the factors that may lead to conflict over water.

The Jordan River region is perhaps one of the most volatile watersheds in the world and the difficulties associated with sharing the waters of the Jordan have been discussed extensively (Assaf et al. 1993; Biswas et al. 1997; Lowi 1993; Saliba 1968;

Scheumann and Schiffler 1998; Stevens 1965; Wolf 1995). The countries and political entities who share the Jordan River are Israel, Jordan, Syria, Lebanon and the West Bank.

According to the authors cited above, the politics of water rights in the region has been one of the dominant issues in the region’s politics, perhaps even being a cause of the Six

Day War of 1967 (Lowi 1993; Saliba 1968; Wolf 1995). Due to the limited supply of water in this region and the increasing need for irrigation and drinking water there, this area is particularly vulnerable to conflict.

The Tigris and Euphrates River systems are illustrative of one of two typical scenarios regarding water conflict. This is a system where a powerful upstream neighbor,

Turkey, controls the water sources and is likely to be unyielding to its downstream neighbors in the future as water resources get scarcer and scarcer. Another scenario which will be discussed below is when a powerful downstream neighbor is controlling a region’s water resources. The Tigris and Euphrates region is vulnerable to conflict for similar reasons as the Jordan region (Lowi 1993; Scheumann 1998). Turkey’s attempts to

22 control and utilize the headwaters of these two international rivers as part of the Grand

Anatolia Project have been sharply criticized by downstream neighbors Syria and Iraq.

Attempts to organize cooperative treaties between the countries have met with mixed success at best and throughout the past three decades tensions have been high between the three water sharing nations.

The Ganges River system is also exemplary of a situation where a powerful upstream neighbor, India, controls the region’s water to the detriment of its downstream neighbor, Bangladesh. Headwaters of the Ganges also originate in Nepal and Bhutan.

During the dry season much of the Ganges water is siphoned off upstream for irrigation and other projects as India has engaged in a number of controversial water diversion programs that have severely limited water supply to Bangladesh (Begum 1988; Crow,

Lindquist and Wilson 1995). During the wet season the Ganges floods the delta creating standing water on much of the Bangladeshi flood plain. Though the Bengladeshis are limited in the extent to which they can protest against their more powerful neighbor, after some demonstrations and violence these two countries have begun procedures that will lead to more cooperation over the water of the Ganges.

The Nile River system is also one of the most hotly disputed water systems in the world. Nine countries—Burundi, Rwanda, Tanzania, Kenya, Zaire, Uganda, Ethiopia,

Sudan and Egypt—share this water system (Lowi 1993). Much of this region has semi- arid or extremely arid weather. Egypt, the most powerful country in the region, receives nearly one hundred percent of its yearly water from the Nile (Postel 1996; Smith and Al-

Rawahy 1991). This water system is an example of a watershed where a powerful

23 downstream neighbor is working to prohibit or deter water utilization upstream. Several of the upstream countries in the Nile River system, particularly Ethiopia, are now poised to expand their economies and make use of the abundant fresh water supplies that they have. Egypt is doing everything it can to assure its own water supply, which would be significantly cut if Ethiopia began to dam the Nile tributaries (Schiffler 1998).

The Aral Sea region provides another example of a water system that is impoverished and also in a politically vulnerable area. The Amu Dar’ya and Syr Dar’ya

Rivers are the main sources of water for this sea of freshwater. The waters of the Aral

Sea are shared by five former Soviet republics: Kazakhstan, Kyrgyzstan, Tajikistan,

Turkmenistan, and Uzbekistan. The Aral Sea has been largely depleted by the diversion and pollution of its waters. Once the fourth largest freshwater lake in the world, the flow of water into the Sea is now seven hundred percent less than 1970 levels (Postel 1996).

This is also an area experiencing economic and population growth, which are likely to require even more water in the region. The depletion of this water source will both inhibit the viability of the region and threaten the people who make their living off of the

Sea.

The Danube River in Europe and the Colorado River in North America are two examples of a unique set of political and hydrological scenarios present in Europe and the

Americas. Countries in the Americas follow a traditional European approach to water rights (Fizmaurice 1996; Saliba 1968). According to Saliba,

(t)he most developed system of law which governs waterways of common interest is to be found in the provisions of water treaties

24 concluded among European states. This is not surprising since it was in Europe where the thrust for the harnessing of water for industrial and economic development began and where the need for development of general principles of law first arose (1968:49).

In the Americas and Europe, water disputes do arise, especially between environmentalists and politicians, but most major water systems are governed by existing international treaties. The existence of treaties does not overcome the fact that there is currently depletion and degradation of water sources throughout Europe and America.

There can still be international disputes in these regions, as pointed out by Ingram,

Delaney and Gillian (1995) and Reisner (1986). These authors point out that the U.S. and

Mexico are still in conflict over the uses of the Rio Grande and the Colorado River.

Cross–border pollution and excessive water withdrawals have been the main causes of disputes between the United States and Mexico over these rivers. In Europe, the waters of the Danube River appear to be unable to keep pace with the growth in the region of

Hungary, Slovakia and Austria. Disputes in the Danube region have occurred over the issues of pollution, water diversion projects, and environmental protection. Fitzmaurice

(1996) points out that the use of water systems such as the Colorado and the Danube will increasingly be disputed by capitalist interests, politicians, sovereign neighbors and environmentalist groups. Researchers have discussed the impacts changes in development, demographic, and ecological variables will have on the likelihood of international water conflict.

25 THE EFFECTS OF HUMAN ECOLOGICAL VARIABLES ON INTERNATIONAL

WATER CONFLICT

POPULATION DYNAMICS

Increases in population size, population density, and migration are all factors which are predicted to influence the likelihood of water conflict in the future. While many of the great sources of fresh water are already overextended, current population growth in certain regions of the world, particularly arid regions, will markedly reduce per capita water availability in the near future (Gardner-Outlaw and Engleman 1997). In the

Middle East, it is predicted that, due to population growth, water will be critically scarce, even if agricultural uses are dramatically reduced (Biswas et al. 1997). According to

Postel (1996), within 25 years over 1.3 billion people in Africa and the Middle East will live in countries that do not have enough internal water resources to supply each person with a healthy supply of potable water.

While population growth will contribute to water scarcity, urbanization and migration will also have effects which increase the chances of water conflict. Densely populated urban areas found in industrialized cities consume large amounts of water, require more elaborate water works, and generate tremendous amounts of waste water which pollutes rivers. Therefore, as the international trend towards urbanization continues, we can expect to find more water infrastructure needed and more pollution generated.

26 Migration caused by water scarcity is another important demographic factor which can contribute to international water conflict. Ethnic competition is occurring in a number of regions where people have been forced to move in search of more reliable water sources. Across Sub-Saharan Africa, nomadic and settled people are increasingly coming into conflict as groups travel farther and farther in search of water for human, animal, and agricultural needs (Rahim et al.1991). Ethnic strife has also occurred in the border region of India and Bangladesh over the use of the Ganges River (Crow, Lindquist and Wilson 1995; Crow and Sultana 2000; Swain 1996). Environmentalists claim that the excessive flooding in Bangladesh (caused in part by deforestation upstream in India and Nepal), along with pollution and water shortages in the dry season, is forcing many

Muslim Bangladeshis to migrate into predominately Hindu India and has caused localized violence and tension between the Indian and Bangladeshi governments in recent years.

SOCIAL ORGANIZATION

The current order of international social organization may in some ways contribute to conflict over water between nations. Industrial-capitalist norms which support large scale endeavors such as massive dams, diversion and navigation projects, and land reclamation may harm the balance of ecosystems, making it difficult for subsistence farmers and peasants to make a living. In a sociological/ecological analysis of the development of water scarcity in Sri Lanka, Starkloff (1998) finds that human conduct in ecological systems is dependent on social systems. Traditional norms for water use supported agricultural and resource extraction technologies that maintained the integrity of the hydrological cycle. Development brought a paradigm shift which

27 supported maximization of water extraction for agriculture and created an asymmetric power scenario in which lower class agrarian communities were marginalized. In this case, human land use patterns directly caused the degradation of the water cycle, leading to competition for water. Similar situations have occurred around the world. Wherever corporate and wealthy interests have taken precedence over the sustainability of local populations the potential for conflict increases (Samson and Charrier 1997). When one nation’s development occurs at the expense of another—as in the case of large scale dam and water diversion projects on international rivers—conflict between the countries is one potential outcome and has already occurred on the Nile, Tigris and Euphrates, Jordan, and

Colorado rivers, among others (Ingram, Laney and Gillian 1995; Scheumann and

Schiffler 1998; Sosland 1998; Wolf 1995).

THE ENVIRONMENT

Environmental factors can also contribute to conflict over water between nations.

As discussed above, environmental scarcity and degradation are typically the result of a number of interacting demographic, organizational, and technological factors. Population growth and changes in social organization and technology have interacted to cause the degradation of the world’s major water systems due to deforestation, mining, agricultural runoff, damming, sewage discharge, chemical discharge and oil spills, soil erosion, and channeling for navigation. The degradation of international watersheds can lead to conflict as nations attempt to address environmental and economic impacts of degradation.

28 One place where such factors have interacted to cause an ecological calamity is the Aral Sea region in Central Asia. Many years of poor watershed management and the drawing off of river water for cotton farming and industrial use has decimated the Aral

Sea. According to De Cordier (1996) and Lipovsky (1995), the ecological consequences of this could paralyze the regional economy and upset the political and social balance.

The sea, which is now nearly dead, can no longer provide food and a living for people on its shores and the necessary steps to restore the sea, which would require the cessation of major water diversion projects, would cripple cotton production—the region’s main agricultural commodity. The deterioration of the Aral Sea watershed has caused low intensity conflicts between the Central Asian Republics over the use of water and how to resolve the unfortunate decline of the sea.

TECHNOLOGY

Technology is also an important factor in water conflicts. Technology has in many ways contributed to population growth, social organizational changes, and environmental degradation. For instance, technology has interacted with population growth to bring about major social organizational and environmental changes, all of which affect competition for natural resources. Technology has an interesting association with water conflict in that it may both cause and reduce disputes.

Large scale hydroelectricity and irrigation projects have been the sources of many water disputes, while newer technologies like water desalination and drip irrigation technology may alleviate water demand issues that lead to conflict (Farinelli 1997).

According to Gleick (1993), the population displacements, ecosystem changes, inequality

29 of resource access, and economic dislocations associated with large scale water development projects may lead to cross-border disputes. The Gab íkovo-Nagymaros dam project on the Slovakia-Hungary border along the Danube River is the site of an international water dispute which arose when Hungary withdrew support for the project based on economic and environmental concerns (Murphy 1997). The Farakka barrage in

India and the Grand Anatolia project in Turkey have caused international disputes for similar reasons (Begum 1988; Scheumann and Schiffler 1998). On the other hand, some researchers and politicians are hopeful that technology can help reduce the sources of water conflict. Although prohibitively expensive in many places now, there is some hope that water desalination will be an effective option for reducing scarcity. Other technologies which could dramatically improve water efficiency, such as drip irrigation, flow restricters, and wastewater recycling, may also help.

CONCLUSION

In this chapter I described several locations where water shared in international contexts have been in contention. Further, I described how the POET variables from the human ecological model can influence the nature and intensity of water conflict. In the next chapter I articulate a number of testable hypotheses which speculate on the nature of the relationships of development, demographic, and ecological variables with conflict and cooperation between nations over fresh water resources.

30 CHAPTER 4

THE GENERAL MODEL AND HYPOTHESES

This chapter describes several testable hypotheses and a general model which specifies the proposed linkages between macro level demographic, development, and ecological variables and international conflict and cooperation over fresh water. The analyses that follow these hypotheses are limited in several ways by data availability.

These limitations reduce the extent to which we can hypothesize on and test relationships between the POET variables discussed in previous chapters and water conflict. While we can test a number of hypotheses on population, development, and environmental variables, we are limited in the extent to which we can test the effects of organization and technology on water conflict and cooperation. In spite of these limitations, the purpose of posing these hypotheses and testing them is to illuminate the effects of macro level variables on water conflict and cooperation using quantified data measured at the country level. Through this analysis we will gain a sense of which predictors are most important in causing water conflict, and the results will assist policy makers in their attempts to peacefully negotiate environmental disputes between countries. Additionally, the results

31 of the tests of these hypotheses will expand social scientific understanding of social change resultant from demographic, developmental, and ecological factors.

THE CAUSES OF INTERNATIONAL FRESHWATER CONFLICT

International freshwater conflict is nominally defined throughout this study as disputes or competitive interactions between countries over the right to utilize fresh water or over the responsibility of protecting and conserving the integrity of freshwater ecosystems. Conflict, though, is truly part of a continuum that ranges from war, on one extreme, to peaceful cooperation, on the other extreme. Wolf (1995) describes the conflict continuum as ranging from cooperation to competition to conflict. Cooperation is defined as there being “coordination of behaviour among entities to realize at least some common goals (87).” Competition is defined as “two or more entities, one or more of which perceives a goal as being blocked by another entity (87).” Finally, conflict is defined as the exertion of power to overcome the perceived blockage of the competition.

It is useful to define conflict as a continuum because explaining the factors which contribute to the absence of conflict or minimal levels of conflict is just as important as explaining the causes of intense conflict. In fact, most international freshwater disputes are at low or medium intensity—more competition than violent conflict.

In the second chapter I discussed research which has analyzed the relationships between human ecological variables and natural resource conflict. Building on the results of that previous research we may hypothesize on the causes of international conflict over fresh water. In this dissertation I focus on the effects of several predictors which are

32 derived from the key human ecological variables. I examine the effects of population growth, population density, urbanization, development level, trade, water quality, water availability, contiguous borders, and the existence of previous water treaties on water conflict between countries.

In the next chapter I describe the construction of the data which is used to test my hypotheses, however here it is necessary to discuss one particular methodological issue which pertains to the nature of analyzing interactions between countries. The dependent variable of concern in this study is conflict/coopoeration events between countries. These outcomes occur between pairs of actors. Therefore the unit of analysis in this study is the pair of countries. The strategy employed in this study is to assess the effects of the independent variables on the outcome variables by testing the effects of the mean levels and first difference levels of the predictors across pairs of water sharing countries on the level of conflict or cooperation between those pairs. The following paragraphs describe the hypotheses which I test in this study. Figure 4.1 lists the hypotheses that are tested in this study.

DEMOGRAPHIC IMPACTS ON INTERNATIONAL WATER CONFLICT

Rapid population growth exerts pressure on ecological systems and human infrastructure. The ability of nature to absorb the impacts of increasing rates of withdrawals from watersheds is limited, and it is also often difficult for engineers and hydrologists to increase the amount of water available to people when the amount of water available in nature remains essentially constant. Therefore, I hypothesize that the average rate of population growth of two countries is associated with an increased level

33 of conflict over water between those countries. Furthermore, I hypothesize that a large difference in population growth rates between two countries is associated with an increased level of conflict over water between those countries, due to increasing needs of the country with more rapid population growth, which will be likely to seek to extract a higher share of water which is shared with another country.

Population density and urbanization may also be associated with increased conflict over water between countries. Urban areas and densely populated areas require more extensive and more intensive water works projects to provide potable water for household, commercial and industrial uses. Water treatment is usually necessary but where it does not exist water pollution is increased. I hypothesize that higher average levels of population density and urbanization across countries contributes to a greater likelihood of conflict between those countries. Further, I predict that large differences in population density and urbanization between countries will contribute to more conflict over water.

DEVELOPMENT AND ECONOMIC EFFECTS ON INTERNATIONAL

WATER CONFLICT

The process of development from agriculturally oriented social organization to an industrial and market oriented society with a complex division of labor often generates disputes over the path of development and other issues that surround cultural change.

Changes in the patterns of water usage from small scale, localized consumption to large scale irrigation projects, dam projects, land reclamation, deep aquifer tapping, and water treatment can cause disagreements, particularly when historical rights are infringed or

34 when large groups are displaced from land or suffer reductions in water availability and quality. Therefore, I predict that, in general, international water conflict will increase with national development. Large differences between countries in terms of development level is also predicted to be associated with higher levels of water conflict due to the inequality and asymmetries of water consumption and power relationships between two countries at different stages of development.

The ecological modernization hypothesis discussed in Chapter 2 argues that the negative effects of development on the environment are reduced at the highest levels of development. Furthermore, proponents of the “Liberal peace” hypothesis argue that countries with highly developed market economies are less likely to come into conflict with one another because in general it is more socially and politically profitable to engage in peaceful trade compared to the costs of conflict, especially warfare. Thus, I hypothesize that the relationship between development and international conflict over fresh water is positive but curvilinear, with a decrease in conflict at the higher levels of development.

One aspect of the Liberal peace hypothesis is the importance of economic interdependence. The hypothesis argues that countries who are financially tied to one another have a mutual interest in maintaining stable and healthy trade regimes. Again, peaceful trade is more profitable than conflict. Based on this logic, I hypothesize that countries that are dominant trade partners with one another are less likely to come into conflict over fresh water.

35 ECOLOGICAL EFFECTS ON INTERNATIONAL FRESHWATER CONFLICT

When water is abundant and of high quality, disputes over water are less likely, but when water becomes scarce and polluted then the demands for potable water are harder to fulfill. Such scarcity and degradation of water may lead to a higher likelihood of conflict between nations competing for shared water resources. At the international level we can hypothesize that decreases in average per capita water supplies will contribute to increased levels of conflict. Furthermore, as average water quality decreases, conflicts between countries may arise over access to clean water or over responsibility for restoring water sources. We may also expect that inequality in water distribution and pollution across countries contributes to international water conflict.

Thus, I hypothesize that larger differences between countries in water availability and water quality will contribute to increased conflict.

GEO-POLITICAL EFFECTS ON INTERNATIONAL WATER CONFLICT

Research has shown that countries who are nearer to one another are more likely to have conflict (Oneal et al. 1996). That is, conflicts between neighboring countries tend to be more likely, more intense, and longer in duration. In terms of international water disputes, we observe that many watershed basins around the world are shared by numerous countries. For example, the Danube basin is shared by 13 countries, the Nile by 11 countries, the Mekong by 6 countries, and the Amazon by 8 countries. In many cases the countries who are sharing water in a shared watershed are not actually border countries. In such cases where water is shared by many countries, I hypothesize that

36 conflicts between countries over water will be more likely among countries who actually share a common border, rather than countries who share water but from a distance.

One type of international freshwater cooperation is the water treaty. When two or more countries sign a formal treaty regarding the rights and responsibilities of each country involved we may say that the countries have cooperated in a positive way. But what are the long term effects of signing a treaty? Is water conflict reduced because a treaty has been signed? Or is signing a treaty an indicator of a larger problem that is likely to continue into the future? I hypothesize that having a signed treaty between two countries is an indicator of larger problems regarding water, and therefore when two countries have a previously existing water treaty we may expect continued conflict in the future.

37 Demographic Effects 1. The mean rate of population growth of two countries has a curvilinear association with the number of water events between those countries and is associated with a lower average conflict intensity score.

2. A large difference in population growth rates between two countries is associated with an increased level of conflict over water between those countries.

3. Higher mean levels of population density across countries is associated with a greater likelihood of conflict between those countries.

4. Larger differences in population density between countries are associated with more international conflict over water.

5. Higher mean levels of urbanization across countries is associated with an increase in the number of water events between countries, and is associated with a greater likelihood of conflict between those countries.

6. Larger differences in urbanization rates between countries are associated with more international conflict over water.

Development Effects 7. International water conflict will increase with national development, however at the higher levels of development conflict intensity will decrease. Thus, I predict a positive curvilinear association between development and international water conflict.

8. Large differences between countries in development is associated with higher levels of water conflict.

9. Water sharing countries that are dominant trade partners with one another will have fewer water interaction events and the events that do occur will be more cooperative rather than conflictive.

continued.

Figure 4.1: List of Hypotheses.

38 Figure 4.1 Continued.

Ecological Effects 10. Lower levels of mean per capita water supplies will contribute to an increased frequency of water interaction events and those events will be more conflictive in nature.

11. Larger differences between countries in average per capita water supplies will be associated with higher levels of water conflict.

12. Lower levels of mean internal water per capita will contribute to an increased frequency of water interaction events and the events that do occur will be more conflictive rather than cooperative.

13. Larger differences between countries in average internal water per capita will be associated with higher levels of water conflict.

14. Lower levels of mean percent of access to potable water between countries will be associated with an increased frequency of water interaction events and will also be associated with events of more conflictive nature.

15. Larger differences between countries in percent access to potable water will be associated with higher levels of water conflict.

16. Higher levels of mean water pollution between water sharing countries will be associated with an increased frequency of water interaction events and will be associated with events of more conflictive nature.

17. Larger differences between countries in water pollution levels will be associated with higher levels of water conflict.

Geo-Political Effects 18. Water sharing countries that have a common border will have more water interaction events than those water sharing countries without a common border, and those with a common border will experience events of a more conflictive nature.

19. Countries with a signed water treaty at the beginning of a time period will experience more water interaction events during that time period, and those events that occur will be more, rather than less, conflictive in nature.

39 CHAPTER 5

METHODS AND DATA

This chapter describes the methods and data used to test my hypotheses on the causes of freshwater conflict between nations. The primary method of hypothesis testing used in this analysis is multi–variate linear regression of water conflict intensity between pairs of water sharing countries on quantified data for a number of predictor variables. In addition to this quantitative analysis, I report results from a case study based on field research and other sources of water conflict between Slovakia and Hungary. The population of interest is all pairs of water sharing countries that existed in the period from

1970 to 2000.

The analysis proceeds in several steps. First, I establish a link between the two indicators of water conflict intensity described in more detail below. The two measures are 1) the number of water interaction events between countries and 2) the mean level of cooperation or conflict of those events. In order to establish the link between these two indicators I show a scatterplot and a trend analysis illustrating the level of correlation between the two measures and how this relationship has changed over time. Then I run a series of ordinary least squares regression models that predict the frequency of water

40 interactions between countries. Finally, I run a second series of ordinary least squares regression models that predict the mean level of conflict or cooperation over water between countries. A comparison of these regression models allows for generalizations on the relative impact of the predictor variables on water conflict/cooperation event outcomes. After discussing these results I discuss the results of the case study on

Slovakia and Hungary.

DATA SOURCES

The data for the quantitative analysis of this study come from several sources.

The event data, or the outcome data, come from the Basins at Risk (BAR) section of the

Transboundary Freshwater Dispute Database (TFDD) (Wolf 1999; Yoffe and Larson

2001). The development, ecological and demographic data come from various sources including the World Bank (2001), the United Nations Human Development Report (1997,

1991, 2000), and The World's Water, 1998-1999: The Biennial Report on Freshwater

Resources (Gleick 1998).

FRESHWATER CONFLICT

The BAR data were compiled by Shira Yoffe of Oregon State University, under the project direction of Aaron Wolf. The present study is the second dissertation written which utilizes the BAR data. The BAR data are exceptional because, prior to their compilation, no comprehensive data set existed which allowed for the analysis of a large quantified data set specifically on international freshwater conflict. The stated goal of the overall BAR Project was “to identify historical indicators of international freshwater conflict and, from these indicators, create a framework with which international river

41 basins at potential risk for future freshwater conflict may be identified and further evaluated (Yoffe and Larson 2001: 8).” Furthermore, the specific goal of creating the

BAR Events Database was:

to identify all reported instances of conflict or cooperation over international freshwater resources for the entire world for the past fifty years, to classify those events by the international river basin in which they occurred, the countries involved in the event, the date, level of intensity of conflict or cooperation, and the main issue associated with each event (Yoffe and Larson 2001: 8).

As defined by the BAR project coordinators, water events are

instances of conflict and cooperation that occur within an international river basin, that involve the nations riparian to that basin, and that concern freshwater as a scarce or consumable resource (e.g., water quality, water quantity) or as a quantity to be managed (e.g., flooding or flood control, managing water levels for navigational purposes) (Yoffe and Larson 2001: 8-9).

Yoffe and her colleagues coded 3,278 water related conflict and cooperation events that occurred between water sharing countries around the world during the last fifty years. Each event was coded separately for every pair of countries that the event pertained to. In other words, if an event occurred that involved five countries, then that event was coded uniquely for every pair of countries that was party to it. From this event data I was able to import and code several relevant indicators of international water conflict and cooperation into a final data set which was used to complete the current study.

The BAR water conflict/cooperation event data was gathered by the Oregon State researchers from several sources, including the COPDAB (Azar 1993), GEDS (Davies

1998), FBIS (CIA 2000), WNC (CIA 2000), and Lexis–Nexis databases. Further information was collected from case studies and newspaper reports. For each event

42 included in BAR, relevant information was also coded on such factors as the date of the event, the individual countries involved, whether or not the event was initiated by one of the countries or was mutually initiated, and what issues were involved in the event. The events themselves were described in detail, and they were also coded on a numerical scale derived from the COPDAB Conflict Intensity Scale (complete description of this scale is below). Table 5.1 shows the basic layout of the original Basins at Risk data set.

INDEPENDENT VARIABLES

The majority of the development, ecological and demographic data were extracted from the World Development Indicators cd-rom published by the World Bank (2001).

This was the source for the following indicators: gross domestic product per capita, country population, population growth rate, percent of the population living in urban areas, water pollution, and access to potable water. The United Nations’ Human

Development Report was the source for the human development index (HDI) and the amount of internal water available per capita. The World's Water, 1998-1999 : The

Biennial Report on Freshwater Resources (Gleick 1998) is the source for data on the total amount of water available per country, per capita water use, the number of threatened fish species, volume of river flow into and out of each country, and the percent of total water withdrawn.

43 MEASURES

WATER CONFLICT AND COOPERATION

In order to assess the patterns of water conflict and cooperation between pairs of countries I summarized the occurrence of water conflict/cooperation events across five year time intervals. To do this, I tallied from the BAR data the number of events which occurred during the seven five year periods from 1970 to 2000 (-1970, 1970-1975, 1975-

1980, 1980-1985, 1985-1990, 1990-1995, 1995-2000). All events which occurred prior to 1970 were included in a single panel. It is useful to examine five year panels, compared to one year panels for example, because that approach allows us to obtain a composite estimate of patterns of water interaction not necessarily present in shorter intervals. I tallied the number of events in order to assess the magnitude, or the frequency, of interaction between each pair of countries. It is important to assess the frequency of occurrence of water conflict and cooperation events because it is an important indication of underlying issues. Although both cooperative and conflictive events are both included in the event frequency measure, it is reasonable to speculate that a high number of interaction events, even when many of those events are cooperative in nature, indicates at the very least some underlying disagreement about shared water issues.

From the events that occurred during the five year intervals for each pair of water sharing countries, I also coded a mean conflict/cooperation intensity score. The purpose of this measure is to indicate clearly the average level of conflictive or cooperative events taking place between two countries. This mean conflict/cooperation intensity score for

44 each country pair/time interval was calculated by summing the modified COPDAP conflict intensity score of all of the events during the time interval and dividing by the number of events in that interval:

Mean BAR Scale Score t = BAR Scale Score t ÷ Number of Scores t

(t = time interval)

The number of events during each time period and the average conflict intensity scores were computed from the BAR data using a program in Visual Basic. The program is included in Appendix A. This program read in the data from the original BAR data set and returned outcome data in a usable form.

The modified COPDAB conflict intensity score was coded in the following way.

First of all, the score ranges from –7 to +7, with a negative score indicating a conflictive or competitive interaction and a positive score indicating a cooperative interaction. The scale was modified from the original COPDAB scale to reflect water related events, rather than conflict events generally. Here I give a brief description of what each score indicates. I more detailed description of each score can be found in Table 5.2. A –7 score on the modified COPDAB scale indicates a formal declaration of war over water; a

–6 score indicates extensive war acts causing deaths, dislocation or high strategic costs over water; a –5 score indicates small scale military acts pertaining to water; a –4 score indicates political or military hostile actions over water; a –3 score indicates diplomatic or economic hostile actions such as unilateral construction of water projects against another country’s protests, reducing the flow of water to another country, or abrogation of

45 a water agreement; a –2 score indicates strong verbal expressions displaying hostility in interaction, official interactions only; a –1 score indicates mild verbal expressions displaying discord in interaction, both official and unofficial, including diplomatic notes of protest; a 0 score indicates a neutral or non–significant act for the inter–nation situation, a +1 score indicates a minor official exchange, or talks or policy expressions of mild verbal support; a +2 score indicates official verbal support of goals, values, or a regime; a +3 score indicates cultural or scientific agreement or support of setting up cooperative working groups; a +4 score indicates non–military economic, technological or industrial agreement, or legal cooperative actions between nations that are not treaties and cooperative projects for watershed management, irrigation, or poverty alleviation; a

+5 score indicates military, economic, or strategic support; a +6 score indicates a major strategic alliance or an international freshwater treaty over water; a +7 score indicates voluntary unification into one nation or the merging of states. Figure 5.1 shows the distribution of the original individual events coded in the BAR data set on this modified

COPDAB scale.

The information extracted from the BAR data for use in this study includes three things: 1) the country pairs that experienced water conflict/cooperation events as coded by Yoffe and colleagues, 2) the number of events that occurred between those pairs during each five year interval, and 3) the mean conflict/cooperation score of the events in each period. It is important to note that not all possible country pairs actually experienced conflict or cooperation events during each time interval. In the final data set on which the analyses are conducted, each eligible country pair is represented for every possible time

46 interval, and those pairs for which no events occurred during the time period in question were simply scored with a “0” on the frequency of interaction score. Each and every dyad did not, however, have a mean conflict/cooperation score for each time interval; only those country–pair time intervals during which at least one actual event occurred received a mean conflict/cooperation score. Table 5.3 shows the descriptive statistics for each of the two outcome variables, as well as for the predictor variables discussed below. From this table we can see that there are 807 pairs that received a mean BAR scale score. Of those pairs, the minimum score was -5.00 and the maximum was +6.00, and the mean of all mean scores was 2.98. Every case (3,434) received a score for the number of interaction events. The maximum score was 67 events in one five year time interval and the mean was 0.91. This measure is highly skewed and therefore in all analyses of this variable the natural log form is used to correct for this skewness.

DEMOGRAPHIC INDICATORS

I hypothesized that the dynamics of population have an effect on the occurrence of water conflict. I am particularly interested in population growth, population density, and urbanization. Indicators of these variables were extracted from the World Development

Indicators CD–Rom (2000). First I extracted each indicator for each individual country for the years 1965, 1970, 1975, 1980, 1985, 1990, and 1995. These variables were measured five years prior to the beginning of each time period in question to account for a lag effect. Population growth was measured as the annual percent increase or decrease in population for that year. Population density was measured as the number of inhabitants per square kilometer. And urbanization was measured as the percent of inhabitants living

47 in urban areas. Since I am interested in the effects of these variables on the behavior of pairs of countries, I created two variables to represent information about each demographic indicator for each pair of countries. The first variable is the mean across both countries, i.e. the mean population growth rate, and the second variable is the first difference between the countries, i.e. the absolute value of the difference in population growth rates between the two countries. The first difference measures used throughout this study refer to the absolute value of the differences between measures for pairs of countries.

DEVELOPMENT AND POLITICAL VARIABLES

I hypothesized that the social, political, and economic development levels of countries affect the occurrence of conflict over freshwater. I assessed the effects of several variables. From the World Development Indicators CD-ROM I extracted the gross domestic product (GDP) per capita of each country in my population for each year for which data was available. The GDP per capita was adjusted to current international dollars so the measure is comparable across time periods. Like the demographic variables, GDP per capita was measured five years prior to the beginning of each time period to account for a lag effect. I also calculated a mean and a first difference score on

GDP per capita for each pair of countries in the analysis. From the United Nations

Human Development Report (1991, 1997, 2000) I extracted the Human Development

Index (HDI) of each country in my population for each year for which data was available.

I calculated a mean and a first difference score on HDI for each pair of countries in the analysis.

48 I also hypothesized that the previous existence of a water treaty would affect the level of international conflict over water. Using data from the Transboundary

Freshwater Dispute Database (Wolf et al. 2002), I measured whether or not there was a bilateral or multilateral treaty on freshwater in existence between each pair of water sharing countries at the beginning of each five year time interval. This variable was coded as a dummy—0 indicates no treaty in existence and 1 indicates that a treaty is in existence.

ECOLOGICAL VARIABLES

Several ecological variables are included in the analysis in order to assess their effects on international water conflict. More specifically, various indicators of water depletion and degradation are measured. Water quality was measured by two variables:

1) an indicator of the volume of inorganic chemical pollutants and 2) an indicator of the number of threatened fish species in a country. The volume of inorganic chemical pollutants in water was extracted from the World Development Indicators CD-ROM

(World Bank 2001) for each country at the beginning of each five year interval starting with 1965. There was a lot of missing data on this measure. The number of threatened fish species was taken from The World’s Water 1998-1999 (Gleick 1998). There was also a lot of missing data on this measure. See Table 5.3 for descriptive statistics on these indicators. The effects of each of these variables were assessed by testing the impacts of the mean and first difference scores. Again, the goal was to look at the main, or overall, effect by testing the association between the mean level of water pollution, and also to look at the effect of a difference in water quality between the two countries.

49 Water depletion and availability were measured with several variables. These include 1) the percent of people with access to safe water, 2) the amount of water used per capita, 3) the total amount of water available per capita 4) the amount of internal water available per capita, and 5) the amount of water withdrawn as a percent of the total water available. These indicators were extracted from several sources, including Gleick

(1998) and WRI (1998), and in each case their effects on water conflict are assessed by looking at the mean score and the first difference scores of the water sharing country pairs.

FINAL DATA SET CONSTRUCTION

In this study the unit of analysis is the country pair. The population of interest is all pairs of water sharing countries that existed in the period from 1970 to 2000. Water sharing country dyads were considered to be any two countries that exist within the same watershed basins. Therefore, water sharing dyads may include countries that do not in fact border one another but only share water as part of a common drainage basin. The internationally shared water basins of the world have been identified by geographers and are listed in Wolf (2002). These watersheds, and the countries that share them, are listed in Appendix B.

The development, ecological, and demographic data used in the study were imported into the base data set with a similar algorithm to the one used to import the event data. This Visual Basic algorithm is shown in Appendix C. Once I obtained both the event data and the data on independent variables for the country dyads in the population, I merged these two data matrices into one set of data that included as its cases

50 all possible water sharing dyads between 1970 and 2000 and the corresponding data on outcome and predictor variables. There are 646 unique pairs of countries who shared water during the time period in question—1970 through 2000. Appendix D lists these pairs. The countries in each of the pairs are identified by their standard three letter country code abbreviations. Appendix E lists the codes with the corresponding conventional long country names. Each of these pairs is represented at multiple points in time, and taking into consideration countries which no longer exist as well as recently formed countries, there are 3,434 unique dyad-time interval pairs of water sharing countries during the period from 1970 to 2000.

Regression models that tested the effects of predictors on the number of water interaction events between country pairs are run on the full sample of 3,434 cases.

Models that tested the effects of predictors on the mean BAR conflict/cooperation scale are run on the 807 cases that actually had a score for that measure. This is due to the fact that country pair/intervals in which no water conflict or cooperation events occurred could not legitimately be coded using the current coding scheme. A separate set of analysis were run in which a “0” score was substituted in missing cases, however since the results were essentially the same these analyses are not included here.

There is perhaps some bias in the cases of the samples due to disproportionate representation of certain country pairs. As mentioned above, the population of interest is all country pairs who shared water during the period from 1970 to 2000. Each country in the world who shares water with at least one other country is included in the study and is paired separately with all of those countries with whom it shares water. The tremendous

51 variation in the sizes of countries, combined with the fact that some international watersheds are shared by many countries, results in a relative over-representation of some countries and a relative under-representation of other countries in the data. For example, small countries in Europe and Africa, like Burundi and Albania, share water with many countries while larger and politically more important countries in the Americas and Asia, like the USA and South Korea, only share water with one or two countries. Each country pair is represented in the population for all relevant time periods, and so when one small country shares water with many other countries this biases the cases in favor of that country because it will be represented proportionately more frequently in the population.

Table 5.4 shows the number of times each country is represented in the final data set.

The table shows that, for example, Burundi (BDI)–because of its size and location in the water basin–is paired with other countries in the population 107 times. Albania (ALB) is paired 101 times, while the United States (USA) is only paired 14 times in the entire population—seven times with Canada and seven times with Mexico. South Korea (KOR) is included in the population only seven times, all with North Korea. The problem of this representation issue is that we may expect a relatively higher level of conflict between

South and North Korea due to their political history, but their limited representation in the cases of the population reduces the effect on the outcomes of the equations. Similarly, the United States has a history of both conflict and cooperation over water with its neighbors, but with only 14 cases out of 3,434 the importance of these cases to the final analyses will be underestimated. On the other hand, we might expect, for a variety of reasons, a lower amount of water conflict in Burundi or Albania, but the relative over

52 representation of these countries among all cases will increase the importance of these cases in the equations. This problem of over-representation of certain countries is controlled for by including a dummy variable for whether or not the countries are bordering one another.

53 DATE BASIN COUNTRIES INVOLVED BAR SCALE EVENT SUMMARY SCORE 12/5/73 Laplata Argentina-Paraguay 4 Paraguay and Argentina agree to build dam, hydroelectric project 1/1/76 Ganges Bangladesh-India –2 Bangladesh lodges formal protest against India with United Nations, which adopts consensus statement encouraging parties to meet urgently, at level of minister, to arrive at settlement

7/3/78 Amazon Bolivia-Brazil-Colombia- 6 Treaty for Amazonian Cooperation Ecuador-Guyana-Peru- Suriname-Venezuela

4/7/95 Jordan Israel-Jordan 4 Pipeline from Israel storage at Beit Zera to Abdullah Canal (East Ghor Canal) begins delivering water stipulated in Treaty (20 mcm summer, 10 mcm winter). The 10 mcm replaces the 10 mcm of desalinated water stipulated Annex II, Article 2d until desalinization plant complete.

6/1/99 Senegal Mali-Mauritania –3 13 people died in communal clashes in 6/99 along border between Maur. & Mali; conflict started when herdsmen in Missira- Samoura village in W. Mali refused Maur. horseman use of watering hole; horseman returned w/ clansmen, attacking village on 6/20/99, causing 2 deaths; in following retaliation 11 more died. Table 5.1: Basins at Risk Event Database Example (Source Yoffe 2002).

54 ORIGINAL COPDAB BAR EVENT DESCRIPTION CODING SCALE 15 –7 Formal Declaration of War 14 –6 Extensive War Acts causing deaths, dislocation or high strategic costs; Use of nuclear weapons; full scale air, naval, or land battles; invasion of territory; occupation of territory; massive bombing of civilian areas; capturing of soldiers in battle; large scale bombing of military installations; chemical or biological warfare.

13 –5 Small scale military acts ; Limited air, sea, or border skirmishes; border police acts; annexing territory already occupied; seizing material of target country; imposing blockades; assassinating leaders of target country; material support of subversive activities against target country. Intermittent shelling or clashes; sporadic bombing of military or industrial areas; small scale interception or sinking of ships; mining of territorial waters. Official actions only. 12 –4 Political-military hostile actions; Inciting riots or rebellions (training or financial aid for rebellions); encouraging guerilla activities against target country; limited and sporadic terrorist actions; kidnapping or torturing foreign citizens or prisoners of war; giving sanctuary to terrorists; breaking diplomatic relations; attacking diplomats or embassies; expelling military advisors; executing alleged spies; nationalizing companies without compensation. Unofficial actions.

11 –3 Diplomatic-economic hostile actions Increasing troop mobilization; boycotts; imposing economic sanctions; hindering movement on land, waterways, or in the air; embargoing goods; refusing mutual trade rights; closing borders and blocking free communication; manipulating trade or currency to cause economic problems; halting aid; granting sanctuary to opposition leaders; mobilizing hostile demonstrations against target country; refusing to support foreign military allies; recalling ambassador for emergency consultations regarding target country; refusing visas to other nationals or restricting movement in country; expelling or arresting nationals or press; spying on foreign government officials; terminating major agreements. Unilateral construction of water projects against another country’s protests; reducing flow of water to another country, abrogation of a water agreement.

(continued)

Table 5.2: Modified COPDAB Scale Items. (Source Yoffe 2002).

55 Table 5.2: Continued.

ORIGINAL COPDAB BAR EVENT DESCRIPTION CODING SCALE 10 –2 Strong verbal expressions displaying hostility in interaction Warning retaliation for acts; making threatening demands and accusations; condemning strongly specific actions or policies; denouncing leaders, system, or ideology; postponing heads of state visits; refusing participation in meetings or summits; leveling strong propaganda attacks; denying support; blocking or vetoing policy or proposals in the UN or other international bodies. Official interactions only.

9 –1 Mild verbal expressions displaying discord in interaction Low key objection to policies or behavior; communicating dissatisfaction through third party; failing to reach an agreement; refusing protest note; denying accusations; objecting to explanation of goals, position, etc.; requesting change in policy. Both unofficial and official, including diplomatic notes of protest.

8 0 Neutral or non-significant acts for the inter-nation situation Rhetorical policy statements; non-consequential news items; non-governmental visitors; indifference statements; compensating for nationalized enterprises or private property; no comment statements.

7 1 Minor official exchanges, talks or policy expressions--mild verbal support Meeting of high officials; conferring on problems of mutual interest; visit by lower officials for talks; issuing joint communiqués; appointing ambassadors; announcing cease-fires; non-governmental exchanges; proposing talks; public non-governmental support of regime; exchanging prisoners of war; requesting support for policy; stating or explaining policy.

6 2 Official verbal support of goals, values, or regime Official support of policy; raising legation to embassy; reaffirming friendship; asking for help against third party; apologizing for unfavorable actions or statements; allowing entry of press correspondents; thanking or asking for aid; resuming broken diplomatic or other relations.

Continued.

56 Table 5.2: Continued.

ORIGINAL COPDAB BAR EVENT DESCRIPTION CODING SCALE 5 3 Cultural or scientific agreement or support (non-strategic) Starting diplomatic relations; establishing technological or scientific communication; proposing or offering economic or military aid; recognizing government; visit by head of state; opening borders; conducting or enacting friendship agreements; conducting cultural or academic agreements or exchanges. Agreements to set up cooperative working groups.

4 4 Non-military economic, technological or industrial agreement. Making economic loans, grants; agreeing to economic pacts; giving industrial, cultural, or educational assistance; conducting trade agreements or granting most favored nation status; establishing common transportation or communication networks; selling industrial- technological surplus supplies; providing technical expertise; ceasing economic restrictions; repaying debts; selling non-military goods; giving disaster relief. Legal, cooperative actions between nations that are not treaties; cooperative projects for watershed management, irrigation, poverty-alleviation.

3 5 Military economic or strategic support. Selling nuclear power plants or materials; providing air, naval, or land facilities for bases; giving technical or advisory military assistance; granting military aid; sharing highly advanced technology; intervening with military support at request of government; concluding military agreements; training military personnel; joint programs and plans to initiate and pursue disarmament.

2 6 International Freshwater Treaty; Major strategic alliance (regional or international). Fighting a war jointly; establishing a joint military command or alliance; conducting joint military maneuvers; establishing economic common market; joining or organizing international alliances; establishing joint program to raise the global quality of life.

1 7 Voluntary unification into one nation Merging voluntarily into one nation (state); forming one nation with one legally binding government.

57 Variable n Min. Max. Mean Std.Dev. Skewness Kurtosis

Mean BAR Scale Score 807 –5.00 6.00 2.98 2.18 –0.19 –0.27

Number of Events 3434 0.00 67.00 0.91 3.41 10.18 –2.00

Mean GDP/Cap Level 1680 237.48 25830.45 4011.80 4637.17 2.08 4.32

Difference in GDP/Cap 1680 0.81 23561.10 2403.76 3566.30 2.86 9.34

Mean HDI Level 1195 0.20 0.95 0.56 0.20 0.24 –1.31

Difference in HDI 1195 0.00 0.60 0.11 0.10 1.27 1.97

Mean Pop. Growth Level 3122 –3.50 5.85 2.08 1.16 –1.07 1.70

Difference in Pop. Growth 3122 0.00 9.20 0.82 0.93 3.51 18.33

Mean Pop. Density Level 2883 0.91 616.33 59.90 69.77 2.57 9.78

Difference in Pop. Density 2883 0.01 905.30 54.18 85.80 4.10 25.08

Mean Percent Urban Level 2741 2.80 93.01 39.31 20.34 0.34 –0.99

Difference in Percent Urban 2741 0.00 67.10 14.59 11.51 1.14 1.33

Shared Border 3434 0.00 1.00 0.49 0.50 0.06 –2.00

Export Partners 497 0.00 1.00 0.21 0.41 1.42 0.01

Treaty in Existence 790 0.00 1.00 0.85 0.36 –2.00 1.95

Mean Internal Water Level 3010 78.50 367015.50 18325.26 39428.16 4.79 27.52

Difference in Internal Water 3010 5.00 450876.00 23999.96 61030.74 4.91 26.60

Mean Threatened Fish Level 3115 0.00 106.00 5.84 8.67 5.10 46.36

Difference in Threatened Fish 3115 0.00 111.00 7.83 11.71 3.41 20.82

Mean Used Water Per Capita 2890 10.50 4398.50 455.64 692.00 3.38 13.29

Difference in Used Water 2890 0.00 7437.00 478.95 1071.62 4.86 25.85

Mean Water Pollution Level 845 492.13 5x106 234959.00 5x105 5.31 32.62

Difference in Water Pollution 845 3.77 8x106 337033.22 9x105 5.36 31.68

Continued.

Table 5.3. Descriptive Statistics for All Variables in the Study.

58 Table 5.3 Continued.

Mean Water Access Level 607 6.00 100.00 52.68 24.73 .20 –1.07

Difference in Water Access 607 0.00 87.00 18.58 15.81 1.38 2.37

Mean Water Withdrawal Level 2966 0.00 1366.50 63.99 194.74 4.39 19.19

Difference in Water Withdrawal 2966 0.00 2362.00 107.45 367.23 4.54 20.28

Mean Total Water Per Capita 3000 173.19 540094.01 35702.22 66030.44 3.50 13.56

Difference in Total Water 3000 0.46 747754.44 46776.62 104497.54 3.71 14.75

59 Country Frequency Country Frequency Country Frequency Country Frequency Code Code Code Code

AFG 41 FIN 21 MLI 98 TUN 7 AGO 98 FRA 49 MMR 62 TUR 69 ALB 101 GAB 77 MNG 22 TZA 135 ARG 35 GDR 46 MOZ 70 UGA 58 ARM 10 GEO 10 MRT 21 UKR 37 AUT 114 GFR 71 MWI 98 URY 28 AZE 10 GHA 35 MYS 21 USA 14 BDI 107 GIN 105 NAM 70 USR 95 BEL 42 GMB 14 NER 91 UZB 10 BEN 84 GNB 14 NGA 98 VEN 49 BFA 84 GNQ 21 NIC 14 VNM 35 BGD 30 GRC 29 NLD 42 YGF 54 BGR 104 GTM 28 NOR 21 YUG 36 BIH 32 GUF 7 NPL 34 ZAF 42 BLR 18 GUY 49 PAK 34 ZAR 135 BLZ 14 HND 21 PAN 14 ZMB 98 BOL 77 HRV 33 PER 56 ZWE 63 BRA 77 HTI 7 PLO 5 BRN 7 HUN 93 PNG 7 BTN 34 IDN 14 POL 101 BWA 70 IND 49 PRK 23 CAF 112 IRN 64 PRT 7 CAN 7 IRQ 40 PRY 28 CHE 103 ISR 31 ROM 93 CHL 21 ITA 103 RUS 38 CHN 107 JOR 58 RWA 107 CIV 91 KAS 1 SAU 35 CMR 161 KAZ 20 SDN 107 COG 77 KEN 65 SEN 35 COL 56 KGZ 14 SLE 77 CRI 14 KHM 35 SLV 14 CZE 34 KOR 7 SOM 21 CZS 64 LAO 35 SUR 49 DEU 50 LBN 35 SVK 34 DJI 14 LBR 21 SVN 33 DOM 7 LBY 49 SWE 14 DZA 105 LSO 21 SWZ 14 ECU 49 LTU 8 SYR 56 EGY 86 LVA 10 TCD 91 ERI 18 MAR 7 TGO 42 ESP 14 MDA 33 THA 42 EST 6 MEX 21 TJK 14 ETH 72 MKD 8 TKM 14

Table 5.4: The Number of Times Each Country Appears in the Final Data Set.

60 Figure 5.1. Distribution of Water Conflict and Cooperation Events (Source: Yoffe 2002).

61 CHAPTER 6

RESULTS

The testing of the hypotheses advanced in this study focused on the explanation of two outcomes: 1) the frequency of international water conflict and cooperation events and

2) the average level of conflict or cooperation that occurred in those events. This chapter presents the results of analyses conducted to assess the association between the various predictors and these two outcomes.

There are three parts to the analysis that are reported here. The first set of analyses describes the relationship between the two measures of the dependent variable.

This analysis shows that the two measures—the number of water interaction events and the mean level of conflict/cooperation—are closely related and that country pairs that had more interaction events experienced relatively more conflictive, or less cooperative, interaction in general. The second set of analyses present the results of models predicting the frequency of occurrence of water conflict and cooperation events, and the third set of analyses presents the results of models predicting the intensity of conflict and cooperation of those events.

62 THE LINK BETWEEN THE NUMBER OF INTERACTION EVENTS AND THE

LEVEL OF CONFLICT/COOPERATION

I argued above that the number of interaction events between two countries indicates an order of magnitude of the salience of water to the international diplomacy between the countries. Furthermore, I argued that a higher number of water interaction events indicates disagreement on water policy between the parties. Figure 6.1 shows, in scatterplot format, the bivariate relationship between the number of water interaction events between two countries and the mean level of conflict/cooperation of those events.

The scatterplot shows a moderately strong correlation. The Pearson r correlation coefficient is –.28, indicating that country pairs with more interaction events also have more conflictive interactions on average.

Figure 6.2 shows a trend analysis of the relationship between the number of water interaction events and mean level of conflict/cooperation over the time period under investigation. For each five year time period I calculated the mean level of conflict/cooperation between country pairs that had 1 event, 2 events, 3 or 4 events, five through 7 events, and 8 or more events. The graph shows that with a couple of exceptions the pairs that had fewer interaction events had more cooperative interactions

(i.e. higher average BAR scores), while pairs that had more interaction events had more conflictive interaction (i.e. lower average BAR scores). An F test of significance for comparison of means found that these differences are significant (results not shown).

The clear relationship between the number of interaction events and the level of conflict/cooperation between country pairs provides a justification for analyzing the

63 number of events as an outcome measure of international water conflict/cooperation.

While it could be argued that there is a reporting bias in events, such that conflictive events are more likely to be recorded and therefore coded into the dataset, the distribution of all events in the dataset shown in Figure 5.1 shows that the majority of events in the data are cooperative rather than conflictive. Therefore I believe that the number of interaction events is a valid measure of the order of magnitude of dispute between countries over water. The next section presents the results of regression models that assess factors predicted to be associated with the number of interaction events.

THE FREQUENCY OF INTERNATIONAL WATER CONFLICT AND

COOPERATION EVENTS

Tables 6.1 through 6.5 show the regression results from analyses of predictors of the frequency of international water conflict and cooperation events. It was impossible, due to data constraints—i.e. missing values—to test one equation model of this function.

Therefore, the results are presented in a series, such that the effects of demographic and development predictors are assessed first, followed by analyses of the effects of ecological variables and treaties.

DEVELOPMENT AND DEMOGRAPHIC VARIABLES

The data were well represented for the development and demographic indicators, with the exception of export trade status. It was therefore reasonable to include these variables in a single model predicting the frequency of water conflict and cooperation events across time periods. Table 6.1 below shows the results from two multi–variate

64 linear regression analyses of the association between development and demographic predictors and the natural log of the frequency of international water events. The number of water interaction events was logged to account for a known skewed distribution. Data on the existence of a shared border was universally available, so this variable is included in all models as a control. All of the cases in the data set were eligible for inclusion in these models, with the limiting factor being data availability. The measures of GDP per capita and the HDI are highly collinear and therefore the effects of those variables on the number of water interaction events are assessed in two separate models, shown as Model

One and Model Two in Table 6.1.

The results shown in Table 6.1 lend support for many of the hypotheses of this dissertation. In both models we see a consistent positive curvilinear association between population growth rate and the number of water interaction events, as was predicted.

That is, at lower levels population growth contributes to an increase in the number of water events. Countries with the lowest levels of population growth have on average fewer water interaction events. Increases in average population growth rates contribute to higher levels of water interaction events, however country pairs with the highest average population growth rates have fewer water interaction events. This effect is net of other indicators of development. Figure 6.3 shows the bivariate association between population growth rate and the number of interaction events during five year intervals.

The association appears curvilinear with an inflection point occurring around 2.5% annual increase in population. There is also a significant association between differences in population growth rates between water sharing countries and the number of water

65 interaction events. The results show that, on average, larger absolute differences between countries in their population growth rates are associated with a higher frequency of water interaction events.

In terms of population density, there is a significant positive association between the mean level of population density and the number of water events. On average, countries pairs with a higher mean level of population density will have more water interaction events. As far as differences between countries in population density, there is no statistically significant increase in the number of water interaction events due to differences between countries in population density.

Neither the mean level nor the difference in urbanization rates of water sharing countries was statistically significantly associated with an increase in the number of water interaction events, when controlling for other variables (both models). There were, however, associations between the measures of development, GDP per capita and HDI, and the number of water interaction events. For both mean level of GDP per capita and mean level of the HDI we find curvilinear relationships in the hypothesized directions.

GDP per capita and the HDI have a positive association with the number of water events at lower levels, however at the higher levels of development we see a decrease in the number of events, on average. The difference in GDP per capita between water sharing countries is significantly weakly associated with an increase in the number of interaction events, however the difference in the HDI of countries is not significantly associated with an increase in the number of water interaction events. These results provide mixed

66 support the hypothesis that water sharing countries at different levels of development will be more likely to experience water interaction events.

The above results are calculated controlling for whether or not the two countries in the dyad pair are bordering one another, or whether they are in the same water basin but not bordering one another. The results from Table 6.1 show a statistically significant association, such that, on average, countries that border one another will be likely to experience more water interaction events over a five year period.

None of the variables in these models explain the majority of the variation in the number of water interaction events between water sharing pairs of countries. However, all of the variables together do explain a meaningful proportion of the overall variation in the outcome measure. For Model One the adjusted R2 =.111 and for Model Two adjusted

R2 =.146. Given the nature of international politics and the range of factors that could contribute to international water interactions, combined with likely measurement error at various levels, these R2 values are not bad.

EXPORT TRADE

The data available for analyzing the effects of export trade dominance on the number of water interaction events were scarce. I wanted to test whether if either country relied on the other country in the pair as a dominant trade partner that would affect the frequency of water interactions. Data were gathered on the ranking of export trade partners for 1970 and 1980. Since this variable, like the other predictor variables, was assessed at the beginning of each five year time interval, this meant that the panels which could be analyzed were the 1970-1975 panel and the 1980-1985 panel. Analysis of only

67 these two panels reduced the sample size significantly. A further problem with this analysis was caused by the fact that GDP per capita and HDI data were not available (in constant units) prior to 1975. This limitation prevented the analysis of a full model on the

1970-1975 panel. So a two step approach was taken to assess the effect of export trade dominance on water interaction intensity. First, one equation was analyzed without GDP per capita data. This allowed for the inclusion of export trade data from 1970 and 1980 in the model and it also allowed for the maximum number of cases in a multi–variate model, minus GDP per capita. A second equation was analyzed using only data from the 1980-

1985 panel. This allowed for the inclusion of GDP per capita in the model, but it limited the number of cases. The results from these two models are shown in Table 6.2.

In the first model of Table 6.2, which included all available data on export trade dominance but which did not include data on GDP per capita, support is found for the hypothesis that export trade dominance is significantly associated with the number of water interaction events. This association is found when controlling for the other predictors in the model, including shared border status. The relationship is positive—that is, water sharing countries who are dominant export trade partners tend to have more water interactions, on average. There are three important caveats to this finding. First, these models are not assessing the qualitative nature of the interaction events, or to what extent the interactions are cooperative or conflictive. Second, even though GDP per capita was dropped from the equation to enlarge the sample size, there are still many missing values on the export data in this model. Since there are two panels in this model there are approximately 1,000 cases, or 2/7 of entire data set, which could be included,

68 however only 422 are in the model due to missing values. Third, the omission of a measure of development, i.e. GDP per capita, leaves the equation without an important component. Overlooking these caveats, initial support was found for the hypothesis that export trade dominance influences the frequency of water interaction events between water sharing countries. There are two major difference between Model One on Table

6.2 and the models from Table 6.1, which both include measures of development. First, we see a significant association between the urbanization level of water sharing countries and the frequency of water events. I suspect that this finding is an artifact of the absence of GDP per capita from the model. While urbanization is certainly conceptually different from GDP per capita, both are potential indicators of development due to their high correlation. Second, the statistically significant association between population density and the number of events is no longer significant in the expected direction. This result may indicate that export trade explains the association between population density and water events. However, since there are substantially fewer cases included in this model relative to analyses presented in Table 6.1, the results from Table 6.1 are given more weight.

The second model from Table 6.2 does include GDP per capita as a measure of development, but this inclusion restricts the sample to the 1980-85 panel only (n=157).

In this model there is still a significant association between export trade dominance status and the number of water interaction events. Some other effects in the model become non–significant, including the average level of GDP per capita, average population density, and shared border status. It is likely that these variables show up as

69 non–significantly related due to the dramatic reduction in sample size in this equation.

These models do have higher adjusted R2 values than the models shown in Table 6.1, which lends further support to the finding that export trade dominance is an important factor in predicting the frequency of water interaction events.

ECOLOGICAL EFFECTS

In order to test the effects of ecological variables, particularly variables related to water quality and availability, several regression models were analyzed. The results of these regression models are shown in Table 6.3 and Table 6.4. The generalizability of the results from these models are limited due to the very poor data coverage of many of the ecological predictor variables. Nevertheless, the models presented in Table 6.3 include the best data which are available on these measures, and therefore I will interpret the results of the models.

To test the effects of ecological variables relating to water quantity and quality I analyze the association between several variables and the number of water events. These variables are the amount of total water available per capita, the percentage of available water withdrawn, the amount of water actually used per capita, the amount of internal water available, the percentage of people with access to potable water, the level of water pollution, and the number of threatened fish species. The effects of each of these variables are assessed in a similar way as the demographic and development variables examined above. That is, for each pair of countries, i.e. each case, I calculated a mean score and a difference score to assess first the main effect of each variable and second the

70 effect of differences or inequality between countries on the number of water events.

The following strategy was employed to analyze the effects of ecological variables on the number of water events. First, I added each of the above variables to a base model the same as or similar to the one shown on Table 6.1, Model One. For these models I use

GDP per capita as an indicator of development because of better overall data coverage.

To test the effects of the amount of water actually used per capita, the percentage of water actually withdrawn, the amount of internal water available, and the number of threatened fish species I used only data from the 1995–2000 panel because those predictor variables were only measured in 1995.

Some meaningful results that provide support for my hypotheses are depicted in

Table 6.3. The first half of Table 6.3 shows the associations of various water quantity and quality variables with the number of water interaction events, controlling for demographic and development indicators shown on the bottom half of the table. Model

One tests the association between the total amount of water available per capita and the frequency of water interactions. Neither the mean level of total water available per capita nor the difference between countries on this indicator was significantly associated with the number of water interaction events in the predicted directions. This non–finding is not especially surprising. The total amount of water available per capita does not provide much insight into issues of actual water use, the distribution of water use, or whether the available water is from internal or external supplies.

Model Two from Table 6.3 tests the association between the amount of internal water available per capita. The goal of this test was to assess whether having internal

71 water resources as opposed to relying on external water decreases the level of international water events. According to the results of the equation, when country pairs have a higher mean level of internal water supplies they will, on average, experience fewer international water event interactions. The difference between countries in available internal water supplies was not, however, associated with an increase in the number of water events, as was expected.

Model Three from Table 6.3 assesses the association between the amount of water actually used per capita and the number of water interaction events. The results show that on average as the level of water use per capita increases the frequency of water interaction events increases. This indicates that countries with the highest levels of actual water use per capita are likely to experience international water events more often. The standardized coefficient for this predictor is large (.49), indicating that water use is indeed an important predictor of water interaction events. Larger absolute differences between countries in the amount of water used per capita are associated with a decrease in the number of water interaction events. This finding was not expected. A possible explanation may be that net of development when one country is using much less water per capita then there is less likely to be conflict. It is perhaps only when both countries are using the same amount of water, especially if it is a high consumption level, that we are likely to see interaction events.

Model Four from Table 6.3 tests the association between the percentage of total water withdrawn and water interaction events. The results for this variable are similar to those for the amount of water actually used per capita. Water sharing country pairs that

72 withdraw a higher percentage of their total available water are more likely to have water interaction events. Country pairs with a larger difference in the percentages of water actually withdrawn have on average fewer water interaction events. This effect is likely the result of one country using a much lower proportion of its water relative to another. If one country is using only a small portion of its total available water then interaction events are less likely.

Model Five from Table 6.3 examines the association between the percent of people with access to potable water and the number of water interaction events, controlling for other variables in the model. We see that the mean level of access to water is not significantly associated with the number of water interaction events, however there is a significant relationship for the difference between countries in access to water.

When there is a larger difference between two countries, the results of the analysis show that on average there will be more water interaction events. This effect is most likely due to the inequities in access to water in the one country relative to the other.

Model Six from Table 6.3 looks at the association between water pollution and the number of water interaction events. First of all, the results of this analysis, like some of the others presented in this study, have to be taken with a grain of salt due to the limited coverage of this particular variable. Over the entire study population only 313 cases had data on water pollution and the other variables in the model. This is an extremely low proportion of the entire population. Nevertheless, since this is one of the only globally available, national level indicators of water quality I will interpret the findings, acknowledging that they are tentative at best. It is found that the average level

73 of water pollution across two countries is not significantly related to the number of water interaction events, however the difference in the level of pollution is significantly related.

When there is a larger difference between countries in terms water pollution we may expect that, on average, there will be more water events between those two countries.

This finding is an indication that, when the water is more polluted in one country relative to its water sharing partner, there may be some attempts to lay blame for pollution in the form of official and unofficial discourse.

In the final model from Table 6.3, Model Seven, I examine the association between the number of threatened fish species and the number of water events. No significant associations were found. This could indicate several things. First, the number of threatened fish species is only a proxy measure of water quality and therefore a non–finding here does not prove that water quality is not associated with water interaction events. Second, since the number of threatened fish species measured at the national level is an aggregated indicator, there are problems here with measurement error. Of course, this non–finding could indicate that water quality is not an especially relevant predictor of the number of water interaction events, however I believe this to be unlikely.

When we look at the pattern of results from the bottom half of Table 6.3 we see some interesting things. In general, the results are consistent with the findings shown in

Table 6.1. There are only three predictors which have consistent, statistically significant effects across all models. These are the effects of average level of population density, the level of difference between countries in population growth rate, and having a shared border. All three predictors are positively associated with the number of water interaction

74 events. These consistent results, which come from a variety of different sub–samples, indicate meaningful relationships here. Two predictors, the difference in urbanization levels and the difference in population density levels, are not significantly associated with the number of water events in any of the models.

We do see some interesting patterns of findings on the bottom of Table 6.3, reflecting the different samples for each equation. First of all, there are two different base samples from which these equations are being drawn. Models One, Five and Six are drawn from all cases across all time periods in the study (with varying sample sizes, depending on variables included). Models Two, Three, Four, and Seven are drawn only from the 1995–2000 panel. While the curvilinear association between population growth level and the number of water events is observed in model one, which has the largest sample size of all the models presented, it is found in none of the models drawing on the

1995–2000 samples.

Similarly, we see that in Models One, Five and Six the difference in country’s

GDP per capita levels is significantly positively associated with the number of water interaction events, however for the 1995–2000 panel samples it is not. Again, this suggests some differences in terms of the processes causing water interaction between the population as a whole, which includes country pairs across three decades, and the most recent period. The effects of the mean level of GDP per capita for country dyads are inconsistent across the seven models in Table 6.3. We find the expected curvilinear association in models three, four, and six but not in the other models. There is no distinctively different pattern in the association between GDP per capita level and the

75 number of water events between the models drawn from the full sample and those drawn from the 1995–2000 sample only. Thus, there is mixed evidence in support of the hypothesized curvilinear association between average development level and water conflict events. In analyses of models which were identical to those on Table 6.3, but which included the HDI as the measure of development as opposed to GDP per capita, all models had a significant curvilinear association between the HDI and the number of water interaction events, when controlling for ecological variables. This supports the original hypothesis.

After I examined models which analyzed the unique associations between each of the ecological variables, I considered several models which included more than one of the ecological variables in order to control for certain effects. The results of those regression analyses are presented in Table 6.4. In Model One of Table 6.4 I include one measure for water quantity—total water available per capita—and one measure of water quality—level of water pollution, along with the other main variables in the model. This model includes data from all the panels in the study, however due to a large number of missing values for the water pollution measure many cases are lost from the analysis.

The results from the equation show that, when controlling for other factors, water quantity and quality are associated with the number of water interaction events.

Countries with higher average levels of water availability per capita tend to have more water events. Country pairs with a large difference in the total amount of water available tend to have fewer water events. The level of water pollution does not appear to have a significant association with the number of water interaction events, however larger

76 differences in the levels of water pollution between countries are associated with a greater number of water interaction events. This finding might indicate that when one country suffers from more pollution they are more likely to seek to confer with their water sharing partners to address the problem.

Model Two of Table 6.4 is an analysis of the effects of water availability and use on the number of water events among the 1995–2000 panel. The water availability and use indicators in this model were only available for the final time period in the study

(1995-2000). The results show that country pairs which have higher average levels of internal water supplies will experience fewer water interaction events on average. Large differences in the amount of internal water available are not, however, associated with an increased number of water events. The amount of water actually used per capita on average is positively associated with the number of water interaction events. That is, country pairs that consume higher amounts of water per capita are likely to have more water events over a five year period. Large differences in the amount of water consumed per capita are negatively associated with the number of water events, however. The results from Model Two did not show significant associations between the mean level of the percent of total water actually withdrawn, nor between the difference in the percent of water actually withdrawn, and the number of water events.

Model Three of Table 6.4 takes a closer look at the relationship between the amount of water available and the number of water events. While the measure of the total amount of water available per capita does give an indication of the overall availability of water it does not take into account the fact that in many cases much of that

77 total available water is not internal to any given country. Actually, the lack of internally available water, as opposed to water from all sources, may be a more important factor in predicting whether or not one country challenges another. Thus, Model Three assesses the effect of internal water availability on the number of water events, controlling for the total availability of water. In this model I apply the 1998 measure of internally available water per capita to all years and run the model on the full sample. This is perhaps a methodological stretch, however we know that in general water availability is consistent over time with some fluctuations. Any interpretations from this model have to consider that the results are from estimations, rather that true measures of internal water. We still see the same pattern as in Model One, that countries with more total water tend to have more water events and countries with large differences in total water availability have fewer events. However, there is a also a clear effect that the less water is available internally the more water interaction events will take place. And large differences in the amounts of internally available water are associated with an increase in the number of water events.

TREATIES

Table 6.5 shows two models which assess the association between water treaties and water interaction events between countries. The first model draws cases from all possible country pairs in the population. Country pairs which had a known bilateral or multi–lateral water related treaty were assigned a value of one on the treaty measure and country pairs which had no known water related treaty were assigned a score of zero. The

78 results show that, when controlling for other variables, countries that had a treaty at the beginning of a five year time period were likely to experience more water interaction events on average compared to countries that did not have treaties. This is not a surprising finding given that countries that have signed water treaties are more likely to have contentiously shared waters in the first place. Even with the existence of a water treaty, countries that have historically disputed water are likely to continue at least dialog over the shared water. But what are the effects of signing a water treaty? Or what happens when countries that formerly did not have a treaty then sign one? Model Two from Table 6.5 tries to answer this question.

Model Two from Table 6.5 includes only data from those country pairs that at some time signed a water treaty. Those cases received a score of one after the treaty was signed but received a score of zero at all time previous to the signing of a treaty. The results from the analysis of the effects of having a signed treaty on this sample show a different pattern. Among countries who had a signed treaty at some time, we see that the effect of having a signed treaty is a decrease in the average number of water interaction events over the next five year period. The effects of other predictors in Models One and

Two from Table 6.5 are consistent with the effects found in other models discussed above.

CONFLICT AND COOPERATION

The results discussed so far have described the relationships between the predictors and the frequency of water interaction events. The occurrence of water events

79 indicates that countries are in the process of negotiating various aspects of their water sharing. But what about the nature of those events? Are they mostly cooperative or mostly conflictive? In the following discussion I describe the results from several analyses that examined the relationships between the predictor variables and the average level of conflict or cooperation of events which occurred. The regression models described below test the effects of the predictors on the level of cooperation or conflict among country pairs which actually experienced water interaction events. More specifically, only country pairs which had at least one event occur during a given time period are eligible for inclusion into these models. The following models roughly parallel those presented in the previous section, with the main difference being that a different dependent variable is being analyzed.

There are 807 different country–pair/time dyads in which at least one international water interaction event occurred. Each of those 807 cases received an Average BAR

Scale Score for the events that occurred between those two countries during that time period. The average score (or average BAR score) is the sum of the modified COPDAB scores for each event divided by the number of events during that time period. Table 6.6 shows the results of regression analyses of the basic models predicting the average conflict/cooperation score.

DEMOGRAPHIC AND DEVELOPMENT VARIABLES

The regression models in Table 6.6 include the measures of population dynamics and development, along with a dummy indicator of whether or not the countries are bordering one another. There are two models in the table, with the distinction between

80 the models being two different measures of development. Model One includes GDP per capita as a measure of development and Model Two includes the HDI as the development measure. Square terms for population growth and development measures are not included in these models because the curvilinear associations predicted between those indicators and the number of water interaction events are not expected in the associations between those indicators and the average level of conflict/cooperation.

The results shown in Table 6.6 are largely consistent with the results found when examining the associations between the same predictors and the number of water interaction events. A major exception is that population growth does not appear to be significantly associated with the average level of conflict or cooperation over water.

Apparently, population growth is significantly associated with an increase in the number of water interaction events, however those events are not necessarily more conflictive in nature. Increases in population density do appear to be associated with more conflictive water interaction events, as do increases in urbanization.

The association between population density and international water conflict has consistently been found to be positive throughout these analyses. That is, country pairs with higher average population density have, on average, more water interaction events of a more conflictual nature. This association could be spurious however. Since population density is likely to be higher in regions with more rivers compared to more arid regions, the association between population density and water conflict may be due to the sheer higher number of water sharing pairs in densely populated areas. In order to rule out this explanation, a correlation analysis was used to determine the association between the

81 number of water sharing dyads each country is included in and the population density of each country. The Pearson r coefficient for this association is low, –0.06. Figure 6.4 shows this association in scatterplot format. The figure shows that countries involved in more water sharing pairs are not significantly more densely populated than other countries.

Countries at higher levels of development, as measured by both GDP per capita and the HDI, exhibit on average more cooperative water interaction events. The measure of the absolute difference between countries’ GDP per capita is significantly associated with more conflictive interaction events, however the measure of the difference between countries’ HDI is not significantly associated with more conflictive events. Thus, there is mixed evidence that larger differences in development between water sharing country pairs contributes to more conflictive interaction. Countries with a common border have on average more conflictive interactions over water, with nearly a one point difference on the average BAR Scale score.

In Model Two of Table 6.6, significant associations in the expected directions are found for the effects of differences between the countries in population growth rate and the level of urbanization. Although these findings do provide some support for my hypotheses, the associations are weak (as indicated by the standardized beta coefficients) and the findings are not consistent across both models.

ECOLOGICAL VARIABLES

Table 6.7 shows the results of a series of regression models which include measures of water quality and availability. There are seven models in the table with each

82 one including a different measure of water quality or availability. Contrary to expectations, there is little statistical support for the hypothesized associations between water quality and quantity measures and the intensity of conflict or cooperation between countries over water. There are, however, two significant findings. First, we see in

Model Two that large differences between countries in the amounts of internal water per capita are associated with more conflictive interaction events. Second, in Model Seven we find that country pairs that have a higher average number of threatened fish species are also likely to have more conflictive interaction events. While support for my original hypotheses is not found among the other equations, Models Two and Seven do lend some support to the hypotheses.

The bottom half of Table 6.7 shows the associations between the demographic and development indicators when controlling for the ecological effects assessed in the top part of the table. In general, these findings are consistent with what was found in previous analyses. The mean level of population growth does not appear to have a significant linear association with the level of conflict or cooperation. In several models, however, particularly those in which only data from the last panel was included, the difference in the level of population growth has a negative association with the BAR

Scale. That is, country pairs that have a higher difference in population growth rates may on average have more conflictive interaction events. The average level of population density among country pairs is found in all of the models to be associated with more conflictive water interaction events. The difference in the level of population density is not, however, significantly associated with the level of conflict or cooperation in any of

83 the models. Higher mean levels of urbanization are found to be associated with more conflictive interaction events, on average, in all of the models presented in Table 6.7, however differences in the level of urbanization is not found to have a significant relationship in any of the models.

The mean level of GDP per capita between country pairs is consistently found to have a significant association with the level of water conflict and cooperation. More specifically, country pairs with higher average GDP per capita tend have more cooperative interaction events. Country pairs with a large difference in the GDP per capita, however, tend to have more conflictive interaction events. In all of the models, countries that border one another have on average more conflictive interactions than country pairs not bordering one another.

Table 6.8 presents the results of several analyses that test the effects of water quality and quantity variables on the average BAR Scale score with different controls. In

Model One I test to see whether water quality and water availability have an effect on the level of water conflict when they are controlled in the same model. Consistent with the findings from Table 6.7, the results show that neither total water availability nor water pollution measures appear to be significantly associated with the average level of conflict or cooperation over water between country pairs.

In Model Two I examine the effects of three different measures of water availability and consumption from the 1995–2000 panel. The results show that measures of the mean and difference levels of water use per capita and the percentage of water withdrawn are not significantly associated with the average level of conflict or

84 cooperation over water between country pairs. Higher average levels of internal water per capita are not significantly associated with more cooperative water interactions on average, however there is a significant association between larger differences in countries’ levels of internal water per capita and more conflictive water interaction events. This finding is consistent with the results found in analyses presented in Table

6.7 and discussed above.

When the effects of the mean level and absolute difference level of internal water are assessed with a control for the total amount of water available, as shown in Model

Three from Table 6.8, there are no significant associations found. In general, the results from tables 6.7 and 6.8 indicate that water quantity and quality factors are not important predictors of the level of conflict or cooperation between countries over water. The exception to this general conclusion is that when countries have larger differences in their amount of internally available water there may be more conflictive interactions as a result.

The results shown in Table 6.8 regarding the effects of demographic and development predictors are consistent with the findings shown above: larger differences between countries in population growth are associated with more conflictive interactions, higher average population density levels are associated with more conflictive interactions, higher average levels of urbanization are associated with more conflictive interactions, higher average levels of GDP per capita are associated with more cooperative interactions, larger differences between countries in GDP per capita levels are associated

85 with more conflictive interactions, and countries with a common border are more likely to have conflictive interactions.

The final set of analyses are presented in Table 6.9. There are two models shown in the table. The first model compares all water sharing country dyads that did not have a known water treaty to those dyads that did have a treaty at the beginning of a particular time period. Surprisingly, the results show that countries with a water treaty have significantly more conflictive interactions than countries without a treaty. On average, the events which occur between countries with a water treaty are l.5 points lower on the

BAR Scale of conflict and cooperation.

The second model on Table 6.9 examines the association between treaty existence and average conflict/cooperation scores among only countries who have at some time signed a water treaty. This model is comparing the average scores on the conflict/cooperation scale of countries after they signed a treaty to the scores of those same countries before they signed a treaty. Surprisingly, we again find that countries who have signed a water treaty have more conflictive interactions after the treaty is signed compared to before. Countries who have a signed treaty have an average BAR Scale score of nearly three points lower (more conflictive) after the treaty is signed than before it is signed. This finding is counterintuitive and suggests that the resolution of water disputes may not be as simple as writing a treaty. Perhaps this finding reflects longstanding water tensions that are resolved only temporarily.

86 Natural Log of Number of Water Interaction Events Predictors Model One Model Two Population Growth Mean Level .172*** .197*** (.329) (.346) Mean Level Squared –.042*** –.069*** (–.269) (–.386) Difference .121*** .160*** (.160) (.139) Population Density Mean Level (÷ 100) .165*** .076† (.213) (.089) Difference –.020 .017 (–.025) (.025) Urbanization Mean Level –.002 –.003 (–.076) (–.087) Difference .000 –.002 (.003) (–.036) GDP per Capita Mean Level (÷ 1000) .033* ..... (.257) Mean Level Squared –.002** ..... (–.287) Difference .009 ..... (.051) Human Development Index Mean Level ..... 3.542*** (1.133) Mean Level Squared ..... –2.883*** (–1.089) Difference ...... 253 (.039) Shared Border1 .313*** .352*** (.266) (.283) R–Square .117 .154 Adjusted R–Square .111 .146 Standard Error of Estimate .555 .576 Number of Cases in Model 15851161 †p<.10; *p<.05; **p<.01; ***p<.001, one–tailed tests. 1Dummy variable: 1=shared border, 0=no common border.

Table 6.1. Regression of Number of International Water Events During Five Year Time Periods on Development and Demographic Variables. Unstandardized Coefficients Shown, with Standardized Coefficients in Parentheses.

87 Natural Log of Number of Water Interaction Events Predictors Model One Model Two (w/o GDP/capita) (1980-85 only)

Export Trade Dominance1 .141* .262** (.110) (.252) Population Growth Mean Level .274** .428* (.589) (1.155) Mean Level Squared –.078*** –.133*** (–.651) (–1.300) Difference .046 .129* (.055) (.161) Population Density Mean Level (÷ 100) –.191 –.035 (–.269) (–.066) Difference .101† .087 (.144) (.160) Urbanization Mean Level .007*** .001 (.309) (.073) Difference .003† .001 (.074) (.022) GDP per Capita Mean Level (÷ 1000) ...... 020 (.172) Mean Level Squared ..... –.003 (–.252) Difference ...... 046* (.256)

Shared Border2 .142** .053 (.148) (.064) R–Square .180 .291 Adjusted R–Square .162 .232 Standard Error of Estimate .441 .363 Number of Cases in Model 422157 †p<.10; *p<.05; **p<.01; ***p<.001, one–tailed tests. 1Dummy variable: 1=at least one out the two countries relies on the other as a dominant export trade partner—the partner is one of the top three countries to which goods are exported, 0=neither country relies on the other as a dominant export trade partner. 2Dummy variable: 1=shared border, 0=no common border.

Table 6.2. Regression of the Natural Log of the Number of International Water Events During Five Year Time Periods on Development and Demographic Variables Including Export Trade Data. Unstandardized Coefficients Shown with Standardized Coefficients in Parentheses.

88 Natural Log of Number of Water Interaction Events Model One Model Two Model Three Model Four Model Five Model Six Model Seven Predictors (2000 only) (2000 only) (2000 only) (2000 only) Total Water Available Per Capita Mean Level (÷ 10,000) .001 ...... (.010) Difference –.007 ...... (–.127) Amount of Internal Water Per Capita Mean Level (÷ 1,000) ..... –.004* ...... (–.241) Difference ...... 001 ...... (.107) Water Actually Used Per Capita Mean Level (÷ 100) ...... 048*** ...... (.492) Difference ...... –.028*** ...... (–.406) Percent of Total Water Withdrawn Mean Level ...... 003* ...... (.727) Difference ...... –.001† ...... (–.524) Percent With Access to Clean Water Mean Level ...... 003 ...... (.113) Difference ...... 009** ...... (.182) Water Pollution Mean Level (÷ 10,000) ...... –.001 ..... (–.125) Difference ...... 002* ..... (.326) Number of Threatened Fish Species Mean Level ...... 006 (.070) Difference ...... 001 (.010) Continued. Table 6.3. Regression of the Natural Log of the Number of International Water Events During Five Year Time Periods on Ecological Variables, Controlling for Development and Demographic Variables. Unstandardized Coefficients Shown with Standardized Coefficients in Parentheses.

89 Table 6.3. Continued. Population Growth Mean Level .143** .126† .070 .042 .185 .031 .064 (.263) (.216) (.110) (.071) (.285) (.057) (.112) Mean Level Squared –.041*** –.040 .007 –.011 –.030 .033† –.002 (–.257) (–.191) (.034) (–.050) (–.180) (.189) (–.011) Difference .138*** .194*** .163** .165** .032 .128*** .201*** (.182) (.147) (.124) (.125) (.033) (.193) (.153) Population Density Mean Level (÷ 100) .001** .002* .003*** .002** .002† .001* .003*** (.166) (.206) (.308) (.232) (.194) (.166) (.285) Difference –.000 –.000 –.001 –.000 .000 –.000 –.000 (–.001) (–.028) (.100) (–.037) (.006) (–.008) (–.052) Urbanization Mean Level –.002 –.013 –.015 –.016 .003 .001 –.012 (–.066) (–.368) (–.403) (–.433) (.096) (.039) (–.341) 90 Difference .000 –.000 .001 –.002 –.000 .002 –.000 (.004) (–.006) (.012) (–.035) (–.002) (.033) (–.011) GDP per Capita Mean Level (÷ 1000) .000 .000 .000† .000† –.000 .000† .000 (.151) (.368) (.525) (.453) (–.756) (.357) (.333) Mean Level Squared –.000* –.000† –.000* –.000† .000 –.000* –.000 (–.216) (–.329) (–.494) (–.372) (.617) (–.319) (–.300) Difference .000* –.000 –.000 –.000 .000*** .000* –.000 (.067) (–.052) (–.042) (–.057) (.292) (.106) (–.052)

Shared Border1 .302*** .447*** .399*** .456*** .294*** .322*** .422*** (.254) (.321) (.278) (.322) (.229) (.254) (.306) R–Square .127 .253 .270 .275 .192 .167 .237 Adjusted R–Square .120 .231 .246 .253 .157 .151 .216 Standard Error of Estimate .558 .610 .623 .612 .578 .585 .609 Number of Cases in Model 1,542 469 408 448 313 661 486 †p<.10; *p<.05; **p<.01; ***p<.001, one–tailed tests. 1Dummy variable: 1=shared border, 0=no common border.

90 Natural Log of Number of Water Interaction Events Predictors Model One Model Two Model Three

Total Water Available Per Capita Mean Level (÷ 10,000) .019† ...... 050** (.188) (.547) Difference –.015* ..... –.029*** (–.238) (–.496) Water Pollution Mean Level (÷ 10,000) –.002 ...... (–.138) Difference .002* ...... (.333) Amount of Internal Water Per Capita Mean Level (÷ 1,000) ..... –.004* –.007*** (–.204) (–.532) Difference ...... 001 .003** (.128) (.331) Water Actually Used Per Capita Mean Level (÷ 100) ...... 033** ..... (.339) Difference ..... –.020** ..... (–.289) Percent of Total Water Withdrawn Mean Level ...... 001 ..... (.259) Difference ..... –.000 ..... (–.109) Population Growth Mean Level .058 .083 .131** (.104) (.130) (.240) Mean Level Squared .022 –.012 –.042*** (.126) (–.054) (–.260) Difference .148*** .151** .146*** (.224) (.115) (.193) Population Density Mean Level (÷ 100) .126* .002** .001** (.157) (.228) (.148) Difference .001 –.000 .000 (.001) (–.052) (.011) Urbanization Mean Level .000 –.014 –.002 (.001) (–.389) (–.073) Difference .002 –.000 .000 (.027) (–.006) (.009) GDP per Capita Mean Level (÷ 1,000) .054* .044 .021 (.430) (.369) (.163) Mean Level Squared –.002* –.002† –.002* (–.369) (–.348) (–.222) Difference .018* –.004 .010† (.106) (–.023) (.057) Shared Border1 .299*** .433*** .297*** (.231) (.302) (.250) R–Square .172 .301 .136 Adjusted R–Square .152 .271 .127 Standard Error of Estimate .594 .613 .555 Number of Cases in Model 635 407 1543 †p<.10; *p<.05; **p<.01; ***p<.001, one–tailed tests. 1Dummy variable: 1=shared border, 0=no common border.

Table 6.4. Combined Regression Models of Logged Number of International Water Events During Five Year Time Periods on Ecological Variables and Development and Demographic Variables. Unstandardized Coefficients Shown with Standardized Coefficients in Parentheses.

91 Natural Log of Number of Water Interaction Events Predictors Model One Model Two (All Countries) (Before and After)

Treaty in Existence1 .120*** –.643*** (.089) (–.219) Population Growth Mean Level .150*** .358*** (.288) (.491) Mean Level Squared –.036** –.073** (–.231) (–.329) Difference .115*** .149*** (.152) (.228) Population Density Mean Level (÷ 100) .172*** –.043 (.222) (–.045) Difference –.021 .414*** (–.033) (.353) Urbanization Mean Level –.003 –.009 (–.104) (–.247) Difference –.000 –.003 (–.006) (–.048) GDP per Capita Mean Level (÷ 1,000) .037* .147** (.290) (.903) Mean Level Squared –.002** –.006*** (–.314) (–.723) Difference .010† .034* (.058) (.119)

Shared Border2 .298** .423*** (.253) (.267) R–Square .124 .305 Adjusted R–Square .117 .285 Standard Error of Estimate .553 .645 Number of Cases in Model 1585 426 †p<.10; *p<.05; **p<.01; ***p<.001, one–tailed tests. 1Dummy variable: the sample for Model One is drawn from all country pairs in the study. The sample for Model Two is only those country–pairs which signed a treaty at some time. Model Two attempts to examine the change in the number of water interaction events after countries sign a treaty. 1=countries have an international water treaty, 0=there is no known international water treaty between the two countries. 2Dummy variable: 1=shared border, 0=no common border.

Table 6.5. Regression of the Natural Log of the Number of International Water Events During Five Year Time Periods on Water Treaty Existence Controlling for Development and Demographic Variables. Unstandardized Coefficients Shown with Standardized Coefficients in Parentheses.

92 Average Level of Conflict or Cooperation Predictors Model One Model Two Population Growth Mean Level .029 .302 (.014) (.144) Difference .099 –.350* (.048) (–.108) Population Density Mean Level (÷ 100) –.882*** –.844** (–.389) (–.363) Difference .189 .329 (.108) (.189) Urbanization Mean Level –.020** –.019* (–.190) (–.087) Difference –.006 –.014† (–.035) (–.079) GDP per Capita Mean Level (÷ 1,000) .146*** ..... (.272) Difference –.151*** ..... (–.216) Human Development Index Mean Level ..... 3.190** (.279) Difference ...... 635 (.030) Shared Border1 –.900*** –.715** (–.193) (–.164) R–Square .156 .112 Adjusted R–Square .136 .087 Standard Error of Estimate 2.034 1.954 Number of Cases in Model 403331 †p<.10; *p<.05; **p<.01; ***p<.001, one–tailed tests. 1Dummy variable: 1=shared border, 0=no common border.

Table 6.6. Regression of Average Level of Conflict or Cooperation Between Countries During Five Year Time Periods on Development and Demographic Variables. Unstandardized Coefficients Shown with Standardized Coefficients in Parentheses.

93 Average Level of Conflict or Cooperation Over Water Model One Model Two Model Three Model Four Model Five Model Six Model Seven Predictors (2000 only) (2000 only) (2000 only) (2000 only) Total Water Available Per Capita Mean Level (÷10,000) –.019 ...... (–.044) Difference .055 ...... (.186) Amount of Internal Water Per Capita Mean Level (÷1,000) ...... 074 ...... (.316) Difference ..... –.075* ...... (–.541) Water Actually Used Per Capita Mean Level (÷100) ...... –.025 ...... (–.086)

Difference ...... –.050 ...... (–.151) Percent of Total Water Withdrawn Mean Level ...... –.003 ...... (–.357) Difference ...... 001 ...... (.188) Percent With Access to Clean Water Mean Level ...... 008 ...... (.061) Difference ...... –.004 ...... (–.028) Water Pollution Mean Level (÷10,000) ...... –.000 ..... (–.012) Difference ...... –.002 ..... (–.147) Number of Threatened Fish Species Mean Level ...... –.052** (–.291) Difference ...... 001 (.008) Continued. Table 6.7. Regression of the Average Conflict or Cooperation Over Water Level Between Countries During Five Year Time Periods on Ecological Variables, Controlling for Development and Demographic Variables. Unstandardized Coefficients Shown with Standardized Coefficients in Parentheses.

94 Table 6.7. Continued.

Population Growth Mean Level .031 .181 –.114 .113 .188 –.175 .030 (.015) (.096) (–.059) (.072) (.060) (–.086) (.016) Difference –.101 –.515* –.531* –.589* –.457 .004 –.604* (–.049) (–.139) (–.151) (–.164) (–.117) (.002) (–.163) Population Density Mean Level –.008** –.013*** –.011*** –.009** –.028** –.015** –.012*** (–.336) (–.616) (–.560) (–.449) (–.963) (–.628) (–.592) Difference .001 .006 .004 .003 .018 .007 .005 (.055) (.403) (.276) (.195) (.648) (.333) (.328) Urbanization Mean Level –.019** –.017† –.021* –.019† –.033† –.036** –.027* (–.178) (–.165) (–.202) (–.190) (–.280) (–.300) (–.265) Difference –.008 –.004 –.010 –.003 –.012 –.009 –.015 (–.043) (–.023) (–.059) (–.016) (–.070) (–.045) (–.081) GDP per Capita Mean Level (÷1,000) .136*** .196*** .122* .168*** .020 .117* .213*** (.255) (.495) (.270) (.435) (.040) (.231) (.538) Difference –.136*** –.163*** –.108 –.149 –.204** –.194* –.092* (–.195) (–.295) (–.193) (–.277) (.344) (–.310) (–.166)

Shared Border1 –.849*** –.452† –.777** –.846** –1.609* –1.151** –.294 (–.183) (–.109) (–.190) (–.203) (–.279) (–.210) (–.291) R–Square .166 .262 .307 .295 .494 .271 .264 Adjusted R–Square .142 .210 .252 .242 .405 .221 .212 Standard Error of Estimate 2.023 1.796 1.713 1.743 1.842 2.138 1.794 Number of Cases in Model 402 168 151 159 74 172 168 †p<.10; *p<.05; **p<.01; ***p<.001, one–tailed tests. 1Dummy variable: 1=shared border, 0=no common border.

95 Average Level of Conflict or Cooperation Over Water Predictors Model One Model Two Model Three Total Water Available Per Capita Mean Level (÷10,000) –.042 ...... 010 (–.118) (.023) Difference .098 ..... –.053 (.377) (–.181) Water Pollution Mean Level (÷10,000) –.001 ...... (–.043) Difference –.002 ...... (–.130) Amount of Internal Water Per Capita Mean Level (÷1,000) ...... 015 –.005 (.068) (–.096) Difference ..... –.048† .016 (–.369) (.452) Water Actually Used Per Capita Mean Level (÷100) ..... –.005 ..... (–.019) Difference ..... –.049 ..... (–.150) Percent of Total Water Withdrawn Mean Level ..... –.003 ..... (–.423) Difference ...... 001 ..... (.323) Population Growth Mean Level –.209 –.037 .061 (–.101) (–.019) (.030) Difference –.394* –.397† –.203† (–.218) (–.113) (–.098) Population Density Mean Level –.012* –.015*** –.008** (–.514) (–.757) (–.340) Difference .004 .007 .001 (.224) (.485) (.074) Urbanization Mean Level –.033** –.021* –.018* (–.273) (–.203) (–.167) Difference –.009 –.007 –.005 (–.046) (–.040) (–.026) GDP per Capita Mean Level (÷1,000) .087† .152** .132*** (.174) (.336) (.248) Difference –.167** –.133** –.136*** (–.270) (–.238) (–.195) Shared Border1 –1.032** –.760** –.853*** (–.189) (–.186) (–.184) R–Square .299 .384 .180 Adjusted R–Square .240 .315 .152 Standard Error of Estimate 2.105 1.6392.011 Number of Cases in Model 170 151402 †p<.10; *p<.05; **p<.01; ***p<.001, one–tailed tests. 1Dummy variable: 1=shared border, 0=no common border. Table 6.8. Combined Regression Models of Average Level of Conflict or Cooperation Over Water Between Countries During Five Year Time Periods on Ecological Variables and Development and Demographic Variables. Unstandardized Coefficients Shown with Standardized Coefficients in Parentheses.

96 Average Level of Conflict or Cooperation Over Water Predictors Model One Model Two (All Countries) (Before and After)

Treaty in Existence1 –1.475*** –2.987*** (–.300) (–.538) Population Growth Mean Level .174 .192 (.085) (.093) Difference .152 .103 (.073) (.069) Population Density Mean Level –.007** .003 (–.321) (.147) Difference .001 –.005 (.078) (–.212) Urbanization Mean Level –.012* .000 (–.115) (.002) Difference –.010 –.005 (–.054) (–.028) GDP per Capita Mean Level (÷1,000) .171*** .131** (.320) (.289) Difference –.163*** –.240*** (–.232) (.380)

Shared Border2 –.446* –.927*** (–.096) (–.142) R–Square .226 .528 Adjusted R–Square .206 .492 Standard Error of Estimate 1 .950 .1.630 Number of Cases in Model 404 140 †p<.10; *p<.05; **p<.01; ***p<.001, one–tailed tests. 1Dummy variable: the sample for Model One is drawn from all country pairs in the study. The sample for Model Two is only those country–pairs which signed a treaty at some time. Model Two attempts to examine the change in the average level of conflict or cooperation after countries sign a treaty. 1=countries have an international water treaty, 0=there is no apparent international water treaty between the two countries.

Table 6.9. Regression of the Average Level of Conflict or Cooperation Over Water Between Countries During Five Year Time Periods on Water Treaty Existence Controlling for Development and Demographic Variables. Unstandardized Coefficients Shown with Standardized Coefficients in Parentheses.

97 8

6

4

2

0

-2

-4

-6 Average BS Score during prev 5 yr

0 10 20 30 40 50 60 70

Number of Events during prev 5 yr

Figure 6.1: Scatterplot of the Relationship Between the Number of Water Interaction Events and the Average BAR Scale Score. Pearson’s r = –.28 (p<.001).

98 99 70

60

50

40

30

20

10

0

-10 Number of Events During 5 Year Interval 5 Year During Events of Number -4 -2 0 2 4 6

Mean Population Growth Rate Level

Figure 6.3. Scatterplot of the Number of Water Events by Population Growth Rate.

100 1000

800

600

400

200

0 Population Density (people per sq. km) Population 0 20 40 60 80 100 120 140 160 180

Number of Dyads for Each Country

Figure 6.4. Scatterplot of the Number of Dyads for Each Country by the Population Density.

101 CHAPTER 7

THE ROLE OF INSTITUTIONS AND ORGANIZATIONS

I hypothesized that the level of conflict and cooperation between nations over fresh water is affected by various aspects of social organization. By social organization I am referring to both macro and micro level aspects of social, political, and economic institutions. For example, at the macro level of social organization I am referring to the broad and varying types of social systems associated with historical changes in the size, technology, economics, and political structures of societies. At this level, I am interested in determining how the level of water conflict changes among and between societies with different types of social systems. In contrast, at the micro level of social organization I am referring to the particular political organizations and institutions of societies that are involved in negotiating water disputes. At this level, I am interested in determining how the level of water conflict between countries is affected by the behavior of particular organizations vis-a-vis other particular organizations. Unfortunately, assessing the effects of these aspects of social organization on international water conflict with quantitative models is very difficult due to data limitations. Therefore, I conducted a case study on

102 the international water dispute between Hungary and Slovakia in order to assess the effect of patterns of social organization on water conflict. This chapter reports these findings.

THE OLD WORLD ORDER VS. THE NEW WORLD ORDER

During the communist era it was common to compare capitalist countries to communist countries and discuss how results varied across societies of different political and economic organization. In the years since the fall of communism in Europe it has become unclear to what extent such comparisons are meaningful. For example, we often cannot precisely state when a country made the complete transition from communism to capitalism. Many countries made the transition gradually. While democratic reforms were in place, often the state still held monopolies on much of the economy. Thus, there is a blurriness in terms of assigning cases to the “Old World Order” or the “New World

Order.” The problem that this blurriness presents for analysis of the current study is that my data span both the communist and the post-communist eras. Because of this problem,

I focused my case study research on countries that were previously communist but then made a transition to social organization based on market capitalism and democracy.

Through my field research in Slovakia and Hungary I was able to assess the effects of regime change on water conflict, as well as investigate the roles particular organizations play in dealing with water conflict problems.

103 THE SLOVAKIA–HUNGARY CASE STUDY

In September and December of 2001 I spent approximately four weeks in

Slovakia and Hungary researching the recent history of disputes on the Danube river.

Through contacts in the United States and Europe, I met and interviewed a number of activists, politicians, and academics who have knowledge of the history of the

Gab íkovo-Nagymaros dam dispute between Hungary and Slovakia. First I will summarize the history of events in the water dispute between Slovakia and Hungary, and then I will discuss the implications of these events for our understanding of how social organization and institutions affect the level of international water conflict.

GAB ÍKOVO-NAGYMAROS

In 1977, after years of studies and planning, Hungary and Czechoslovakia signed a treaty plan to build a series of dams on the Danube River along the stretch forming their common border—at Gab íkovo on the Slovak side and Nagymaros on the Hungarian side. This agreement was negotiated under Soviet supervision and was intended as a signal of communist countries’ ability to plan and execute large scale development and engineering projects that would contribute to economic growth in the region. Virtually no environmental impact analysis was done at the time. According to Murphy (1997: 62),

“intensive industrial development and wasteful farming methods in the socialist countries during the pre-transition period led to increasing degradation of water and other resources.”

Starting in 1981, and continuing to the present, Hungarian environmentalists began criticizing the negative impacts of the dam complex on ecologically sensitive areas

104 of the inland delta region of Hungary. The protesters eventually coalesced into a group called the Danube Circle which began collecting signatures in opposition to the project.

The Danube Circle and other similar groups were officially banned by communist leaders.

By 1988, support for the resistance groups had increased and the stability of the ruling regime had decreased. In September of 1988 nearly 30,000 Hungarians marched on parliament demanding the stoppage of the dam project. Shortly after that, the Hungarian parliament officially suspended construction, to the dismay of the Slovaks across the border. After the September 1988 rally other disenfranchised and disillusioned

Hungarians began to organize, calling for a transition to democracy in Hungary.

In 1990 democrats won parliament in Hungary following the Czechoslovak Velvet

Revolution of 1989. Environmentalists in the new Hungarian government expected the

Gab íkovo–Nagymaros project to be stopped altogether, however nationalism on the side of the Slovaks, which ultimately culminated in Slovak independence, supported the project completely as a sign of their economic and technological strength. This disagreement began the last twelve years of debate over whether to complete the project or to dismantle it and repair the ecological damage done. The tensions over the fate of the Hungarian minority in Slovakia became entangled with the dam issue. When Slovaks finally diverted the river in 1992, the Hungarians protested heavily and there were rumors of military movement on both sides of the border.

In 1993 both sides agreed to submit the case to the International Court of Justice.

This was the first case of international environmental dispute before the court. Ultimately the court’s ruling in 1997 gave a “split decision.” The court ordered both sides to

105 continue negotiations. No reparations were ordered for either side. Hungary was faulted for breaking the original agreement and Slovakia was faulted for unilaterally diverting the river. Negotiations are ongoing to find a final solution to this problem. The Nagymaros dam on the Hungarian side has never been completed, while the Slovak dam at

Gab íkovo is completed.

THE EFFECTS OF MACRO AND MICRO ORGANIZATIONAL/INSTITUTIONAL

FACTORS ON WATER CONFLICT IN THE CASE OF HUNGARY AND SLOVAKIA

In the period before and after the Hungarian–Slovak treaty that initiated the joint dam project along the Danube on the common border, cooperation on environmental issues among Eastern European communist countries was enforced by the Soviets.

According to Murphy (1997: 64), “Soviet domination kept potential political, economic, and ethnic conflicts among the east European Danube countries under control.

Allegiance to Moscow by most, if not all, of the country leaders, to the Warsaw Treaty

Organization, and to economic ties ... contributed to regional cooperation.” But the inattention of the communist regimes to control environmental degradation and depletion contributed to the rise of environmental social movement organizations, which ultimately helped bring down the old governments.

This case study shows how industrial and technological development by communist regimes lead to water scarcity and degradation. For example, the navigation canal and the hydroelectricity plant on the Danube along the Hungarian–Slovak border siphoned water away from the natural bed of the river, lowering water tables and drying

106 up side channels of the river. Protest in Hungary about the environmental effects of the dam lead to social breakdown in Hungary in the form of a weakened state and receptivity to insurgency. Protesters seized the opportunity to support environmental policy which was in conflict with the agenda of Slovakia. The political opportunities associated with democratic reforms in Hungary caused previously existing discontent to come out. So with the fall of communism, international environmental disputes became front and center. Historical ethnic disputes between the two countries exacerbated the problem.

Ultimately, third party intervention by the European Community and the International

Court of Justice helped bring about a greater degree of cooperation.

Many of the environmental and social activists in Hungary who protested the development of the Danube have since been elected or appointed into positions in the

Hungarian government. In their new positions, these former activists are pursuing institutionalized avenues of negotiating with the Slovak diplomats. Because of this, relations between Hungary and Slovakia have normalized, although the tensions are not entirely resolved. In the final picture, we see the following pattern. Under the centralized rule of Soviet backed communism prior to 1990 conflict over water and other natural resources was minimal or non–existent between the East European states. As a result of the opening up of the political systems, and ultimately the fall of communism, in these states, the voices of discontent grew louder and this was followed by formal dispute over water policy between Slovakia and Hungary. Once the new political and economic reforms became entrenched, and previous activists became members of the official state structure, conflict declined as institutionalized procedures were pursued.

107 CHAPTER 8

CONCLUSION

The primary goal of this study was to explore how and to what extent developmental, demographic, and ecological factors contribute to an increased likelihood of nations coming into conflict with one another over sources of fresh water. Another goal of the study was to contribute to our understanding of the factors that lead to conflict over natural resources more generally. These goals were accomplished in three ways.

First, a thorough review of research on natural resource conflicts described potential linkages between factors presumed to affect such conflict. Second, I discussed how the human ecological approach can be effectively applied to the study of resource conflicts.

And last, quantitative and qualitative analyses of international water events provided an assessment of the association between key predictor variables and water conflict or cooperation.

SUMMARY OF RESULTS

The results of this quantitative analysis provided support for many of the proposed hypotheses. The analyses were structured to assess the factors associated with two

108 outcome measures: the frequency of water interaction events and the average level of conflict/cooperation among those events. The results from analyses on these two outcome measures revealed several interesting findings.

Across pairs of water sharing countries, the mean level of population growth has a statistically significant positive curvilinear association with the number of water interaction events. This relationship, however, appears to be mediated by environmental factors. When measures of water degradation and depletion are entered into the equations, the association between population growth and the number of water events becomes non–significant in six of seven models. This finding indicates that the effect of population growth on water conflict/cooperation is indirect: population growth causes water degradation and depletion, which causes water interaction events. The difference in population growth rates between countries is also significantly associated with an increase in the number of water interaction events. This indicates that countries who share water and are experiencing different rates of population growth will be more likely to have water interactions.

The relationship between population growth and the average level of conflict or cooperation is not so clear. Mean population growth is not significantly associated with the level of conflict/cooperation in any of the models, however the absolute difference in population growth is negative and significant in most of the equations. This finding indicates that, in terms of the conflictive or cooperative nature of water events, it is not population growth that matters, but inequality in population growth rates: when one of

109 two water sharing countries is experiencing a much higher rate of population growth then conflict will be more likely.

Population density has a clear, strong, and robust association with water conflict, measured both as the frequency of interaction events and as the average level of conflict/cooperation. In virtually every regression model, higher population density is associated with an increase in the number of international water interaction events and with a decrease in the average conflict/cooperation score (indicating more conflict). The effect is stronger on the average level of conflict, meaning that, controlling for other factors, actual conflict over water is more likely in situations where population density is high.

The effect of urbanization on water conflict is inconsistent. On the one hand, in most of the regression models the level of urbanization is not associated with a significant increase in the number of water interaction events. On the other hand, the effect of urbanization on the average level of conflict and cooperation is negative and significant in virtually every model. These findings indicate that urbanization does not increase the number of interactions, but the interactions that do occur are more conflictive in nature.

The development indicators GDP per capita and the Human Development Index have consistent associations with indicators of water conflict. Results from analysis of the factors associated with the frequency of water interaction events suggest that there is a curvilinear relationship between development and water conflict events. This finding lends support to the argument that there is an environmental Kuznet’s curve: countries at low levels of development experience relatively few water events; as development

110 proceeds environmental interaction increases, but then decreases at the higher levels of development. Higher development is associated with an increase in the level of cooperation over water, as measured by the average Basins at Risk Scale score. This association is robust, as it is observed in nearly all of the models.

The absolute difference between countries in the level of development also has a consistent association with water conflict. This supports the argument that inequality across countries can contribute to an increased likelihood of conflict over natural resources like water. If this finding is true, then we may expect to see more conflict in places where there is inequality in development across countries.

Several interesting findings are observed in analysis of ecological predictors of water conflict. Environmental factors of water degradation and depletion appear to affect the frequency of water interaction events, but do not strongly influence the level of conflict or cooperation of those events. Five of seven indicators of environmental degradation and depletion have associations with the number of interaction events in the predicted direction. For example, the higher the amount of internal water available the fewer events occurred. As water use and the share of available water withdrawn increase so does the number of water interaction events. Inequality in the share of people with access to water and the level of water pollution is associated with an increased number of water interactions. These findings on the effects of inequality in water access and water pollution indicate that the cross-national contexts of ecological injustice do have consequences for international political interaction.

111 The effects of ecological variables on the average level of conflict or cooperation over water are less clear. Five of the seven variables measured do not have statistically significant associations with the level of conflict or cooperation. Two indicators, however, do have significant relationships. First, the absolute difference in the level of internally available water per capita between two countries is associated with an increased level of conflict. This finding supports the hypothesis that unequal supplies of internal water can contribute to conflict. Second, the number of threatened fish species is associated with an increase in the level of conflict over water. This finding may perhaps indicate that fish resources are an important area of concern for countries, to the extent that when the sustainability of fish is threatened conflict may increase. In general, though, the environmental variables do not dramatically affect the level of conflict between countries over water. Development and demographic factors are on the whole more important factors associated with water conflict.

The results from the analysis of the effects of economic trade on water conflict are somewhat limited. There were not enough cases to do a legitimate analysis of the association between trade dominance and the average level of water conflict/cooperation.

Results from the analysis of the association between trade dominance and the frequency of water interaction events showed that country pairs that are dominant trade partners have more frequent water interaction events. This finding may lead to the conclusion that countries engaged in extensive economic interaction will also engage in more extensive water negotiation. Unfortunately, the data do not allow for an analysis of the

112 average level of conflict or cooperation of the interaction events that do occur between trade partners.

The effect of international water treaties on the level of water conflict is surprising. First of all, when comparing country pairs that have a water treaty to those that do not have a treaty, we find that, on average, pairs with a treaty have more interaction events. But, when looking only at country pairs that had a signed treaty, we see that the number of water interaction events declines after a treaty is signed. This finding makes sense, however the results from the analysis of the effect of treaties on the average level of water conflict/cooperation reveal a counterintuitive pattern. In comparing water sharing country pairs with a water treaty to those without a treaty, we find that pairs with a treaty have on average more conflictive interactions over water.

Furthermore, among countries with a signed treaty, we find that the water interaction events that do occur in the five year time period after the treaty is signed are significantly more conflictive compared to periods before the treaty was signed. This unexpected finding indicates that perhaps water treaties do not reduce the conflict over water between countries. Maybe this reflects longstanding water tensions that are resolved only temporarily through a signed treaty.

A case study analysis of the international dispute over water between Slovakia and

Hungary provided data on the impact of social organization and institutions on water conflict between nations. The findings showed that technology and industrialization in the communist countries of Eastern Europe contributed to environmental degradation and depletion in numerous ways. But the imposition of cooperation by the Soviets prevented

113 sources of conflict from turning into actual conflict. With the decline and fall of communism in Slovakia and Hungary, political opportunities increased with the weakening of various institutions of communist social control. Following this, organization at the micro level increased, pressuring the Hungarians to initiate policies that were in opposition to those supported by Slovakia. Once new democratic reforms became institutionalized, and once political voices found roles within the establishment, the intensity of international conflict subsided, although it has not disappeared.

IMPLICATIONS FOR THEORY

Human ecology theory has attempted to build a conceptual framework for studying the relationships between humans and the environment, focusing on the interactions between populations, organization, the environment, and technology. In this study I have utilized this approach to generate and test a series of hypotheses on the causes of natural resource conflict, as measured by international freshwater interaction events. The human ecological framework proved to be an effective and appropriate logical system for the generation of such hypotheses. The results of the research presented here illustrated this effectiveness in numerous ways.

First, the population and demographic variables proved to be the variables most consistently associated with water conflict indicators. Population density had the overall most robust and strongest relationship with water conflict. Second, the qualitative analysis of changes in institutions and the roles of governmental and non–governmental groups in Hungary and Slovakia illustrated how changes in social organization affect the

114 level of natural resource conflict. These results showed that centralized political, economic and social organization can suppress conflict over natural resources between nations, but in a decentralized system these suppressed conflicts can erupt as groups within and outside of government struggle to control policy. Third, this research clearly showed how environmental factors of natural resource degradation and depletion are associated with conflict, particularly in circumstances of unequal access to resources and unequal water pollution across countries. Still, environmental degradation is not as important of a predictor of conflict as demographic and developmental factors. Lastly, the results of this study demonstrate the effects of technology on water conflict.

Technology was not measured directly in this study, it was measured indirectly by measures of development. According to Lenski, technology is the driving force behind development. Thus, it is reasonable to estimate degrees of technology through development indicators. The results of this study showed how development is related to natural resource conflict. In general, countries at higher levels of development

(technology) are more cooperative with one another. In situations of international interaction, conflict is more likely when there is a high degree of inequality between natural resource sharing countries in terms of relative development levels.

While support was found for most of the hypotheses advanced in this study, support was not found for several of the hypotheses. No evidence was found to show that differences in the population densities or levels of urbanization of countries who share water are associated with more water conflict. Further, while I hypothesized that countries who were dominant trade partners would have fewer water interaction events, in

115 fact the opposite is true. Trade partners have more interaction events, net of other factors in the model. No association was found between the total amount of water available per capita and water conflict. The findings showed that it is not total water availability, but rather water access, internal water, and water pollution that seemed to influence the level of water conflict. While the mean level of water access across countries was not a significant predictor of water conflict, the results indicate that inequality in water access across countries is associated with at least a higher number of water interaction events, if not a greater level of conflict. Finally, partial support was found for the hypothesis that water pollution contributes to water conflict between nations. Neither the mean level of water pollution nor the mean number of threatened fish species were associated with an increase in water interaction events, but large differences in water pollution levels were associated with more interaction events. And while water pollution was not associated with a higher level of conflict over water, the number of threatened fish species did increase the average level of international conflict over water.

IMPLICATIONS FOR POLICY

Policy makers will be interested in how the results of this study can be used to address concerns of environmental security in cross–national contexts. The findings of the study show that policy makers should focus diplomacy efforts first of all in regions of high population density, where population growth has contributed to significant environmental degradation and depletion. Furthermore, actual conflict is most likely in middle and low development regions where there are high levels of inequality in internal

116 water resources, water access, and water pollution across countries. And one last important consideration is that the signing of an international water treaty is not necessarily something that will prevent further conflict over water between two countries.

It is likely that tangible programs that ensure steady and reliable access to clean and safe water are the most important policies that can be implemented in transboundary contexts of shared water.

CONCLUDING COMMENTS

History shows that human societies have competed for control of important natural resources like water for thousands of years. Most likely, competition over natural resources will continue in various ways. In the case of internationally shared fresh water, the possibility of more serious conflict appears to be increasing due to numerous demographic, developmental, and ecological factors. Because of this possibility, and because of the importance of water for human survival, it is important that we give much consideration to the processes that may increase stress on hydrological systems and increase human competition for water in the face of increasing population and demographic shifts. With support from careful research and skillful diplomacy, we will hopefully achieve effective international policy regimes that can be implemented to prevent water from being a cause of major conflict among the water sharing countries of the world.

117 APPENDIX A

Syntax from Visual Basic Program Used to Import and Summarize Water Conflict and Cooperation Event Data from the Basins at Risk Data Set.

‘COMMENT Label Subroutine Sub sum_dyad_evts()

'COMMENT Input Column Labels Const colDATE = "E" Const colDYAD_CODE = "F" Const colIRM = "L" Const colBARSCALE = "P" Const colCTY1 = "J" Const colCTY2 = "K" Const colGROUPID = "B"

'COMMENT Upper Bound for Arrays Const UB = 900 Const UB_UNIQ = 50

'COMMENT Dimension Incoming Data Dim rng As Range Dim dyads(UB) Dim d_counter(UB) Dim bs_max(UB) As Integer Dim bs_min(UB) As Integer Dim bs_sum(UB) As Integer Dim d_code As String Dim direc(UB) As String Dim i(UB) As Integer Dim R(UB) As Integer Dim M(UB) As Integer Dim cty1(UB) As String Dim cty2(UB) As String Dim groupid() As String

118 Dim uniq_group As Boolean

'COMMENT Dimension Outgoing Data Dim outsheet

'COMMENT Dimension Miscellaneous Variables Dim j, n

‘COMMENT Read in Data from Each Time Period For Each rng In Sheet1.UsedRange.Rows

If (rng.Columns(colDATE).Value < CDate("12/31/1970")) Then d_code = "70" ElseIf (rng.Columns(colDATE).Value < CDate("12/31/1975")) Then d_code = "75" ElseIf (rng.Columns(colDATE).Value < CDate("12/31/1980")) Then d_code = "80" ElseIf (rng.Columns(colDATE).Value < CDate("12/31/1985")) Then d_code = "85" ElseIf (rng.Columns(colDATE).Value < CDate("12/31/1990")) Then d_code = "90" ElseIf (rng.Columns(colDATE).Value < CDate("12/31/1995")) Then d_code = "95" Else d_code = "00" End If

' COMMENT Count Number of Events in Each Time Period for Each Dyad For n = 0 To UBound(dyads)

If (dyads(n) = rng.Columns(colDYAD_CODE).Value & d_code) Then uniq_group = True For j = 0 To UBound(groupid) If groupid(j) = rng.Columns(colGROUPID).Value Then uniq_group = False ElseIf groupid(j) = "" Then groupid(j) = rng.Columns(colGROUPID).Value Exit For End If Next j

If uniq_group = False Then Exit For

119

'COMMENT Match Count d_counter(n) = d_counter(n) + 1

'COMMENT Compare w/Existing Values If (bs_max(n) < rng.Columns(colBARSCALE).Value) Then bs_max(n) = rng.Columns(colBARSCALE).Value ElseIf (bs_min(n) > rng.Columns(colBARSCALE).Value) Then bs_min(n) = rng.Columns(colBARSCALE).Value End If

'COMMENT Maintain bscale Total bs_sum(n) = bs_sum(n) + CInt(rng.Columns(colBARSCALE).Value)

‘COMMENT Count Number of Times First Country in Pair Initiates Event

If rng.Columns(colIRM).Value = "I" Then i(n) = i(n) + 1 ElseIf rng.Columns(colIRM).Value = "R" Then R(n) = R(n) + 1 ElseIf rng.Columns(colIRM).Value = "M" Then M(n) = M(n) + 1 End If

Exit For

ElseIf dyads(n) = Empty Then 'we're on the 'new' element

dyads(n) = rng.Columns(colDYAD_CODE).Value & d_code

bs_max(n) = rng.Columns(colBARSCALE).Value bs_min(n) = rng.Columns(colBARSCALE).Value bs_sum(n) = CInt(rng.Columns(colBARSCALE).Value)

'COMMENT Scrap and Redeclare Unique Group Id Info ReDim groupid(UB_UNIQ) As String groupid(0) = rng.Columns(colGROUPID).Value

direc(n) = rng.Columns(colIRM).Value If direc(n) = "I" Then i(n) = 1 ElseIf direc(n) = "R" Then R(n) = 1

120 ElseIf direc(n) = "M" Then M(n) = 1 End If

d_counter(n) = 1

‘COMMENT Identify Countries in Pair cty1(n) = rng.Columns(colCTY1).Value cty2(n) = rng.Columns(colCTY2).Value

Exit For End If Next n Next rng

'COMMENT create new sheet for output Set outsheet = ThisWorkbook.Sheets.Add

'COMMENT Results outsheet.Cells(1, 1) = "DYAD_CODE" outsheet.Cells(1, 2) = "BSMIN" outsheet.Cells(1, 3) = "BSMAX" outsheet.Cells(1, 4) = "BSSUM" outsheet.Cells(1, 5) = "ENTRIES" outsheet.Cells(1, 6) = "AVG" outsheet.Cells(1, 7) = "FIRSTCASE" outsheet.Cells(1, 8) = "MODEDIREC" outsheet.Cells(1, 9) = "CTY1" outsheet.Cells(1, 10) = "CTY2" j = n = 0 For j = 0 To UBound(dyads) If IsEmpty(dyads(j)) Then Exit For 'we hit the end

outsheet.Cells((j + 2), 1) = dyads(j) outsheet.Cells((j + 2), 2) = bs_min(j) outsheet.Cells((j + 2), 3) = bs_max(j) outsheet.Cells((j + 2), 4) = bs_sum(j) outsheet.Cells((j + 2), 5) = d_counter(j) outsheet.Cells((j + 2), 6) = (bs_sum(j) / d_counter(j))

121 outsheet.Cells((j + 2), 7) = direc(j)

‘COMMENT Determine Modal Initiation, I=Cty1 Initiate R=Cty1 Recipient M=Mutual

If i(j) > R(j) And i(j) > M(j) Then outsheet.Cells((j + 2), 8) = "I" ElseIf R(j) > i(j) And R(j) > M(j) Then outsheet.Cells((j + 2), 8) = "R" ElseIf M(j) > i(j) And M(j) > R(j) Then outsheet.Cells((j + 2), 8) = "M" Else: outsheet.Cells((j + 2), 8) = "X" End If outsheet.Cells((j + 2), 9) = cty1(j) outsheet.Cells((j + 2), 10) = cty2(j) Next j

122 APPENDIX B

International River Basins (Source: Wolf 2002).

AFRICA BASIN NAME AREA OF BASIN (km2) COUNTRY AREA (km2) PERCENT Akpa Yafi 4,900 Cameroon 3,100 62.26 Nigeria 1,900 37.74 Atui 10,400 Mauritania 9,300 89.71 Western Sahara 1,100 10.29 Awash 155,300 Ethiopia 144,000 92.71 Djibouti 11,100 7.14 Somalia 240 0.15 Baraka 66,600 Eritrea 41,800 62.84 Sudan 24,700 37.16 Benito 12,600 Equatorial Guinea 11,100 88.57 Gabon 1,400 11.16 Bia 11,900 Ghana 6,900 57.83 Ivory Coast 4,800 40.01 Buzi 27,900 Mozambique 24,700 88.81 Zimbabwe 3,100 11.18 Cavally 30,600 Ivory Coast 16,600 54.31 Liberia 12,700 41.48 Guinea 1,300 4.21 Cestos 15,000 Liberia 12,700 84.54 Ivory Coast 2,300 15.32 Guinea 20 0.14 Chiloango 11,700 Congo, Democratic Republic of the (Kinshasa) 7,700 65.91 Angola 3,700 32.11 Congo, Republic of the (Brazzaville) 230 1.97 Congo/Zaire 3,699,100 Congo, Democratic Republic of the (Kinshasa) 2,307,800 62.39 Central African Republic 402,000 10.87 Angola 291,500 7.88 Congo, Republic of the (Brazzaville) 248,400 6.72 Zambia 176,600 4.77

123 Tanzania, United Republic of 166,800 4.51 Cameroon 85,300 2.31 Burundi 14,300 0.39 Rwanda 4,500 0.12 Gabon 460 0.01 Malawi 90 0.00 Corubal 24,100 Guinea 17,600 72.89 Guinea-Bissau 6,500 26.82 Cross 52,800 Nigeria 40,300 76.39 Cameroon 12,400 23.56 Daoura 34,600 Morocco 18,300 52.82 Algeria 16,300 47.18 Dra 54,900 Morocco 40,600 73.97 Algeria 14,300 26.03 Etosha-Cuvelai 167,600 Namibia 114,300 68.24 Angola 53,200 31.76 Gambia 70,000 Senegal 50,800 72.55 Guinea 13,300 18.95 Gambia, The 5,900 8.41 Gash 31,700 Eritrea 17,700 55.75 Sudan 8,500 26.85 Ethiopia 5,500 17.41 Geba 12,800 Guinea-Bissau 8,700 67.71 Senegal 4,100 31.86 Guinea 50 0.41 Great Scarcies 11,400 Guinea 9,000 79.30 Sierra Leone 2,300 20.53 Guir 79,100 Algeria 61,400 77.60 Morocco 17,700 22.40 Incomati 46,200 South Africa 29,200 63.19 Mozambique 14,300 30.97 Swaziland 2,700 5.84 Juba-Shibeli 805,100 Ethiopia 367,700 45.67 Somalia 221,500 27.52 Kenya 215,900 26.81 Komoe 78,500 Ivory Coast 58,500 74.52 Burkina Faso 17,100 21.74 Ghana 2,300 2.93 Mali 630 0.81 Kunene 110,300 Angola 95,500 86.57 Namibia 14,800 13.43 Lake Chad 2,394,200 Chad 1,082,000 45.19 Niger 675,700 28.22

124 Central African Republic 218,900 9.14 Nigeria 180,800 7.55 Algeria 90,000 3.76 Sudan 83,100 3.47 Cameroon 46,900 1.96 Chad, Claimed by Libya 12,300 0.51 Libya 4,600 0.19 Lake Natron 55,600 Tanzania, United Republic of 37,300 67.06 Kenya 18,300 32.94 Lake Turkana 207,600 Ethiopia 113,600 54.75 Kenya 89,900 43.28 Uganda 2,600 1.23 Sudan 1,500 0.70 Sudan, Administered by Kenya 70 0.03 Limpopo 415,500 South Africa 184,100 44.31 Mozambique 87,300 21.01 Botswana 81,500 19.61 Zimbabwe 62,600 15.06 Little Scarcies 19,300 Sierra Leone 13,300 68.86 Guinea 6,000 31.11 Loffa 11,400 Liberia 10,000 87.49 Guinea 1,400 12.43 Lotagipi Swamp 38,900 Kenya 20,500 52.52 Sudan 10,000 25.58 Sudan, Administered by Kenya 3,300 8.44 Ethiopia 3,200 8.30 Uganda 2,000 5.16 Mana-Morro 6,900 Liberia 5,800 83.67 Sierra Leone 1,100 16.31 Maputo 31,300 South Africa 18,600 59.43 Swaziland 11,000 35.02 Mozambique 1,700 5.55 Mbe 7,000 Gabon 6,500 92.65 Equatorial Guinea 500 7.18 Medjerda 23,100 Tunisia 15,600 67.28 Algeria 7,600 32.72 Moa 22,600 Sierra Leone 10,900 48.16 Guinea 8,700 38.58 Liberia 3,000 13.27 Mono 23,400 Togo 22,400 95.43 Benin 1,100 4.57 Niger 2,117,700 Nigeria 563,000 26.59 Mali 541,600 25.58

125 Niger 499,200 23.57 Algeria 161,500 7.63 Guinea 96,300 4.55 Cameroon 88,200 4.17 Burkina Faso 83,100 3.93 Benin 45,200 2.14 Ivory Coast 22,800 1.08 Chad 16,600 0.78 Sierra Leone 30 0.00 Nile 3,038,100 Sudan 1,931,300 63.57 Ethiopia 356,900 11.75 Egypt 273,100 8.99 Uganda 238,900 7.86 Tanzania, United Republic of 120,300 3.96 Kenya 50,900 1.68 Congo, Democratic Republic of the (Kinshasa) 21,700 0.71 Rwanda 20,800 0.69 Burundi 13,000 0.43 Egypt, Administered by Sudan 4,400 0.14 Eritrea 3,500 0.12 Sudan, Administered by Egypt 2,000 0.07 Ntem 35,000 Cameroon 20,400 58.26 Gabon 9,400 26.99 Equatorial Guinea 5,200 14.75 Nyanga 12,400 Gabon 11,500 93.30 Congo, Republic of the (Brazzaville) 830 6.70 Ogooue 223,400 Gabon 189,800 84.93 Congo, Republic of the (Brazzaville) 26,600 11.91 Cameroon 5,100 2.30 Equatorial Guinea 1,900 0.87 Okavango 708,600 Botswana 359,000 50.67 Namibia 176,800 24.95 Angola 150,100 21.18 Zimbabwe 22,700 3.20 Orange 947,700 South Africa 565,600 59.68 Namibia 240,600 25.39 Botswana 121,600 12.84 Lesotho 19,900 2.09 Oued Bon Naima 510 Morocco 350 68.87 Algeria 160 31.13 Oueme 59,500 Benin 49,500 83.24 Nigeria 9,500 16.04 Togo 430 0.72

126 Ruvuma 152,200 Mozambique 99,500 65.39 Tanzania, United Republic of 52,200 34.30 Malawi 470 0.31 Sabi 116,100 Zimbabwe 85,700 73.88 Mozambique 30,300 26.12 Sassandra 68,200 Ivory Coast 59,700 87.51 Guinea 8,500 12.49 Senegal 437,000 Mauritania 219,100 50.14 Mali 151,300 34.61 Senegal 35,700 8.16 Guinea 31,000 7.08 St. John 15,600 Liberia 13,000 83.55 Guinea 2,600 16.44 Ivory Coast 2 0.01 St. Paul 21,200 Liberia 11,800 55.47 Guinea 9,500 44.52 Tafna 9,500 Algeria 7,000 74.16 Morocco 2,400 25.84 Tano 14,300 Ghana 13,900 97.07 Ivory Coast 420 2.93 Umba 8,200 Tanzania, United Republic of 6,900 83.83 Kenya 1,300 16.17 Umbeluzi 5,400 Swaziland 3,100 57.17 Mozambique 2,300 41.63 South Africa 70 1.19 Utamboni 7,700 Gabon 4,600 59.81 Equatorial Guinea 3,000 39.31 Volta 414,000 Burkina Faso 174,200 42.07 Ghana 166,500 40.21 Togo 25,900 6.25 Mali 18,900 4.57 Benin 15,000 3.62 Ivory Coast 13,400 3.24 Zambezi 1,388,200 Zambia 577,900 41.63 Angola 255,500 18.40 Zimbabwe 215,800 15.55 Mozambique 163,800 11.80 Malawi 110,700 7.97 Tanzania, United Republic of 27,300 1.97 Botswana 19,100 1.38 Namibia 17,100 1.23 Congo, Democratic Republic of the (Kinshasa) 1,000 0.07 Total Area 18,682,410

127 ASIA

BASIN NAME AREA OF BASIN (km2) COUNTRY AREA (km2) PERCENT

Amur 1,884,000 Russia 1,005,300 53.36 China 849,900 45.11 Mongolia 28,700 1.52 Korea, Democratic People's Republic of 120 0.01 An Nahr Al Kabir 1,200 Syria 730 60.60 Lebanon 470 39.40 Aral Sea 1,319,900 Kazakhstan 923,500 69.97 Uzbekistan 236,700 17.93 Kyrgyzstan 138,000 10.46 Tajikistan 13,000 0.99 Turkmenistan 1,500 0.12 China 40 0.00 Asi/Orontes 18,200 Syria 10,100 55.63 Turkey 6,600 36.13 Lebanon 1,500 8.21 Astara Chay 560 Iran 450 81.33 Azerbaijan 100 18.67 Atrak 34,200 Iran 23,500 68.64 Turkmenistan 10,700 31.36 Bangau 430 Malaysia 230 53.04 Brunei 200 46.26 Beilun 960 China 700 73.61 Vietnam 250 26.39 Ca/Song-Koi 33,800 Vietnam 19,600 57.83 Laos, People’s Democratic Republic of 14,300 42.12 Coruh 20,700 Turkey 18,800 90.93 Georgia 1,900 9.07 Dasht 31,800 Pakistan 25,100 78.87 Iran 6,700 21.13 Fenney 2,800 India 1,800 65.88 Bangladesh 950 34.12

Fly 64,600 Papua New Guinea 60,400 93.42 Indonesia 4,300 6.58 Ganges-Brahmaputra- 1,664,700 India 974,300 58.52 Meghna China 320,400 19.25 Nepal 147,300 8.85 Bangladesh 112,400 6.75 India, Claimed by China 67,100 4.03 Bhutan 39,900 2.40 Myanmar 2,100 0.13 Indian Control, Claimed by China 1,200 0.07 Golok 1,800 Thailand 1,100 58.09 Malaysia 780 41.91 Han 35,300 Korea, Republic of 25,100 71.18 Korea, Democratic People's Republic of 10,100 28.63 Har Us Nur 197,800 Mongolia 195,400 98.77 Russia 2,300 1.17 China 120 0.06 Hari/Harirud 92,600 Afghanistan 41,100 44.33 Iran 35,400 38.17 Turkmenistan 16,200 17.50 Helmand 345,200 Afghanistan 283,800 82.23

128 Iran 48,300 13.99 Pakistan 13,100 3.78 Hsi/Bei Jiang 361,500 China 351,700 97.28 Vietnam 9,800 2.72 Ili/Kunes He 161,200 Kazakhstan 97,200 60.28 China 55,300 34.31 Kyrgyzstan 8,700 5.40

Indus 1,086,000 Pakistan 609,100 56.09 India 282,200 25.98 China 111,000 10.22 Afghanistan 72,500 6.68 Chinese Control, Claimed by India 9,600 0.89 Indian Control, Claimed by China 1,600 0.15

Irrawaddy 404,100 Myanmar 368,400 91.15 China 18,600 4.60 India 14,200 3.52 India, Claimed by China 1,200 0.29

Jordan (Dead Sea) 42,800 Jordan 20,600 48.44 Israel 9,100 21.35 Syria 4,900 11.54 West Bank 3,200 7.40 Egypt 2,700 6.39 Golan Heights 1,500 3.54 Lebanon 570 1.34

Kaladan 30,500 Myanmar 22,700 74.39 India 7,400 24.24

Karnafauli 15,000 Bangladesh 11,200 74.78 India 3,700 24.98

Kowl-E-Namaksar 40,100 Iran 26,600 66.50 Afghanistan 13,400 33.50

Lake Ubsa-Nur 74,800 Mongolia 52,700 70.48 Russia 22,100 29.52

Ma 24,600 Vietnam 16,800 68.20 Laos, People’s Democratic Republic of 7,800 31.80

Mekong 780,300 Laos, People’s Democratic Republic of 198,400 25.42 Thailand 194,100 24.87 China 168,400 21.58 Cambodia 157,000 20.11 Vietnam 35,000 4.49 Myanmar 27,500 3.53

Merauke 6,500 Indonesia 4,000 61.35 Papua New Guinea 2,500 38.65

Murgab 60,900 Afghanistan 36,400 59.78 Turkmenistan 24,500 40.22

Nahr El Kebir 2,000 Syria 1,700 81.77 Turkey 370 18.18

Ob 2,734,800 Russia 2,109,600 77.14 Kazakhstan 573,400 20.97 China 50,400 1.84

129 Oral (Ural) 260,400 Kazakhstan 142,400 54.69 Russia 118,000 45.31

Pakchan 2,700 Thailand 1,600 59.64 Myanmar 1,100 38.86

Pandaruan 810 Brunei 410 50.25 Malaysia 400 49.75

Pu-Lun-To 88,400 China 76,300 86.27 Mongolia 12,100 13.65 Russia 40 0.04 Kazakhstan 40 0.04

Red/Song Hong 164,600 China 84,400 51.28 Vietnam 79,000 47.98 Laos, People’s Democratic Republic of 1,200 0.75

Rudkhaneh-ye/BahuKalat 20,600 Iran 20,600 99.68 Pakistan 50 0.25

Saigon/Song Nha Be 29,400 Vietnam 25,000 98.93 Cambodia 230 0.92

Salween 244,100 China 128,000 52.43 Myanmar 107,000 43.85 Thailand 9,100 3.71

Sembakung 15,200 Indonesia 8,200 53.53 Malaysia 7,100 46.46

Sepik 73,400 Papua New Guinea 71,100 96.94 Indonesia 2,200 3.06

Song Vam Co Dong 15,300 Vietnam 7,700 50.22 Cambodia 7,600 49.72

Sujfun 16,800 China 9,800 58.39 Russia 7,000 41.61

Tami 89,900 Indonesia 87,600 97.54 Papua New Guinea 2,200 2.46

Tarim 950,200 China 901,700 94.90 Kyrgyzstan 23,900 2.51 Chinese Control, Claimed by India 21,500 2.27 Pakistan 1,900 0.20 Tajikistan 1,000 0.11 Kazakhstan 110 0.01 Afghanistan 20 0.00

Tigris-Euphrates/Shatt al Arab 793,600 Iraq 318,900 40.19 Turkey 197,100 24.84 Iran 155,600 19.60 Syria 119,400 15.04 Jordan 2,200 0.28 Saudi Arabia 240 0.03

Tumen 33,000 China 22,600 68.56 Korea, Democratic People's Republic of 10,200 30.90 Russia 180 0.54

130 Wadi Al Izziyah 580 Lebanon 380 65.91 Israel 190 33.74

Yalu 63,000 Korea, Democratic People's Republic of 31,700 50.38 China 31,200 49.59

Yenisey/Jenisej 2,497,600 Russia 2,169,800 86.88 Mongolia 327,600 13.12 Total Area 16,930,840

EUROPE

BASIN NAME AREA OF BASIN (km2) COUNTRY AREA (km2) PERCENT

Bann 5,600 United Kingdom 5,400 97.15 Ireland 160 2.85

Barta 1,800 Latvia 1,100 60.86 Lithuania 670 37.73

Bidasoa 530 Spain 470 89.52 France 60 10.48

Castletown 380 United Kingdom 290 76.12 Ireland 90 23.88

Danube 779,500 Romania 228,800 29.35 Hungary 92,800 11.90 Yugoslavia (Serbia and Montenegro) 81,000 10.40 Austria 80,300 10.30 Germany 52,100 6.68 Bulgaria 47,300 6.06 Slovakia 46,800 6.01 Bosnia and Herzegovina 37,800 4.85 Croatia 34,000 4.37 Ukraine 25,600 3.29 Czech Republic 21,300 2.74 Slovenia 16,400 2.10 Moldova 12,100 1.55 Switzerland 1,700 0.21 Italy 740 0.09 Poland 550 0.07 Albania 140 0.02

Daugava 79,600 Byelarus 28,300 35.55 Russia 27,100 34.02 Latvia 23,200 29.14 Lithuania 1,000 1.29

Dnieper 495,500 Ukraine 296,800 59.90 Byelarus 116,700 23.55 Russia 81,900 16.53

Dniester 72,200 Ukraine 52,900 73.37 Moldova 19,200 26.60 Poland 20 0.03

Don 425,600 Russia 371,200 87.21 Ukraine 54,400 12.78

131 Douro/Duero 96,200 Spain 77,900 81.01 Portugal 18,300 18.99

Drin 18,500 Yugoslavia (Serbia and Montenegro) 9,000 48.55 Albania 7,200 39.23 Macedonia 2,300 12.21

Ebro 85,100 Spain 84,200 98.96 France 470 0.55 Andorra 410 0.49

Elancik 1,400 Russia 940 68.19 Ukraine 440 31.81

Elbe 139,500 Germany 88,600 63.54 Czech Republic 49,600 35.60 Austria 1,100 0.77 Poland 140 0.10

Erne 3,500 Ireland 2,000 56.39 United Kingdom 1,500 43.59

Fane 200 Ireland 190 96.46 United Kingdom 10 3.54

Flurry 60 United Kingdom 50 73.77 Ireland 20 26.23

Foyle 2,900 United Kingdom 2,000 67.23 Ireland 960 32.77

Garonne 55,800 France 55,100 98.80 Spain 620 1.11 Andorra 40 0.07

Gauja 8,100 Latvia 6,900 85.87 Estonia 1,100 14.13

Guadiana 65,700 Spain 55,300 84.13 Portugal 10,400 15.87

Isonzo 3,000 Slovenia 1,800 59.55 Italy 1,200 40.05

Jacobs 440 Norway 300 68.55 Russia 140 31.45

Kemi 55,800 Finland 52,700 94.38 Russia 3,100 5.56 Norway 10 0.01

Klaralven 51,500 Sweden 43,400 84.15 Norway 8,200 15.84

Kogilnik 6,100 Moldova 3,600 57.85 Ukraine 2,600 42.15

132 Krka 1,300 Croatia 1,100 89.84 Bosnia and Herzegovina 110 8.96

Kura-Araks 193,800 Azerbaijan 59,800 30.86 Georgia 34,500 17.78 Iran 33,500 17.28 Armenia 29,900 15.42 Turkey 28,500 14.70 Russia 110 0.06

Lake Prespa 1,400 Macedonia 610 42.76 Albania 420 29.54 Greece 390 27.71

Lava-Pregel 8,800 Russia 6,400 72.54 Poland 2,200 25.36

Lielupe 27,200 Lithuania 19,000 70.03 Latvia 8,100 29.71

Lima 2,300 Spain 1,200 50.88 Portugal 1,100 49.04

Maritsa 52,800 Bulgaria 35,000 66.38 Turkey 14,300 27.16 Greece 3,400 6.46

Mino/Minho 16,600 Spain 16,000 96.42 Portugal 590 3.56

Mius 7,100 Ukraine 4,800 67.83 Russia 2,300 31.53

Naatamo 710 Norway 530 74.08 Finland 170 24.37

Narva 58,200 Russia 29,300 50.26 Estonia 16,800 28.84 Latvia 12,200 20.89 Byelarus 10 0.02

Neman 93,000 Byelarus 41,500 44.64 Lithuania 39,700 42.73 Poland 6,600 7.10 Russia 4,800 5.15 Latvia 330 0.35

Neretva 10,800 Bosnia and Herzegovina 9,900 91.99 Croatia 500 4.68 Yugoslavia (Serbia and Montenegro) 360 3.32

Nestos 12,000 Greece 8,500 70.88 Bulgaria 3,500 28.75

Oder/Odra 116,500 Poland 103,000 88.43 Czech Republic 7,400 6.35 Germany 6,100 5.22 Slovakia 10 0.00

Olanga 18,800 Russia 16,800 89.38

133 Finland 2,000 10.62

Oulu 24,800 Finland 23,600 95.25 Russia 1,200 4.75

Parnu 5,900 Estonia 5,900 99.78 Latvia 10 0.22

Pasvik 16,900 Finland 16,200 95.71 Norway 700 4.13 Russia 30 0.15

Po 87,100 Italy 82,600 94.83 Switzerland 4,100 4.71 France 380 0.44 Austria 30 0.03

Prohladnaja 620 Russia 480 77.06 Poland 140 22.94

Rezvaya 670 Turkey 500 74.66 Bulgaria 170 25.34

Rhine 195,000 Germany 106,800 54.74 Switzerland 34,700 17.78 France 25,400 13.00 Netherlands 11,900 6.11 Belgium 11,200 5.74 Luxembourg 2,500 1.29 Austria 2,300 1.19 Liechtenstein 160 0.08 Italy 140 0.07

Rhone 84,700 France 84,000 99.08 Switzerland 730 0.86 Italy 50 0.06

Roia 660 France 450 67.78 Italy 200 30.09

Salaca 4,000 Latvia 2,700 66.27 Estonia 1,400 33.71

Samur 6,800 Russia 6,300 92.74 Azerbaijan 430 6.38

Sarata 1,800 Ukraine 1,100 63.78 Moldova 640 36.16

Schelde 17,500 France 8,900 51.11 Belgium 8,400 48.22 Netherlands 80 0.46

Seine 86,100 France 84,200 97.78 Belgium 1,800 2.13 Luxembourg 80 0.09

Struma 16,800 Bulgaria 8,400 49.84 Greece 6,000 35.45 Macedonia 1,800 10.63 Yugoslavia (Serbia and Montenegro) 690 4.08

134 Sulak 14,800 Russia 13,800 92.02 Georgia 1,000 6.70 Azerbaijan 20 0.11

Tagus/Tejo 69,900 Spain 55,500 79.29 Portugal 14,500 20.71

Tana 16,100 Norway 9,400 58.34 Finland 6,700 41.60

Terek 43,800 Russia 41,800 95.43 Georgia 2,000 4.57

Torne/Tornealven 37,300 Sweden 25,300 67.86 Finland 10,600 28.50 Norway 1,400 3.64

Tuloma 26,100 Russia 23,400 89.91 Finland 2,600 10.04

Vardar 33,200 Macedonia 20,400 61.29 Yugoslavia (Serbia and Montenegro) 8,900 26.79 Greece 4,000 11.92

Velaka 1,100 Bulgaria 780 72.47 Turkey 300 27.53

Venta 7,700 Latvia 5,400 70.20 Lithuania 2,200 28.33

Vijose 9,000 Albania 6,500 72.00 Greece 2,500 27.60

Vistula/Wista 193,900 Poland 169,200 87.28 Ukraine 13,000 6.70 Byelarus 9,700 5.03 Slovakia 1,900 0.99 Czech Republic 10 0.00

Volga 1,553,900 Russia 1,849,800 99.74 Kazakhstan 2,400 0.15 Byelarus 1,600 0.10

Vuoksa 62,700 Finland 54,100 86.26 Russia 8,600 13.74

Yser 920 France 500 53.63 Belgium 430 46.37 Total Area 5,695,590

NORTH AMERICA

BASIN NAME AREA OF BASIN (km2) COUNTRY AREA (km2) PERCENT

Alesek 8,300 Canada 7,200 87.27 United States 1,000 12.42

Artibonite 8,800 Haiti 6,600 74.37

135 Dominican Republic 2,300 25.55

Belize 11,500 Belize 7,000 60.86 Guatemala 4,500 39.14

Candelaria 12,800 Mexico 11,300 88.24 Guatemala 1,500 11.74

Changuinola 3,200 Panama 2,900 91.29 Costa Rica 270 8.34

Chilkat 4,100 United States 2,400 58.01 Canada 1,700 41.99

Chiriqui 1,700 Panama 1,500 86.17 Costa Rica 240 13.83

Choluteca 7,400 Honduras 7,200 97.68 Nicaragua 170 2.32

Coatan Achute 2,000 Mexico 1,700 86.27 Guatemala 270 13.73

Coco/Segovia 25,400 Nicaragua 17,900 70.52 Honduras 7,500 29.48

Colorado 651,100 United States 640,700 98.40 Mexico 10,400 1.60

Columbia 668,400 United States 566,500 84.75 Canada 101,900 15.24

Firth 6,000 Canada 3,800 63.60 United States 2,200 36.40

Fraser 239,700 Canada 239,100 99.74 United States 620 0.26

Goascoran 2,800 Honduras 1,500 53.36 El Salvador 1,300 46.64

Grijalva 126,800 Mexico 78,900 62.26 Guatemala 47,800 37.73

Hondo 14,600 Mexico 8,900 61.14 Guatemala 4,200 28.50 Belize 1,500 10.36

Lempa 18,000 El Salvador 9,500 52.45 Honduras 5,800 32.01 Guatemala 2,800 15.54

Massacre 800 Haiti 500 62.03 Dominican Republic 290 35.96

Mississippi 3,226,300 United States 3,176,500 98.46 Canada 49,800 1.54

Motaqua 16,100 Guatemala 14,600 90.85

136 Honduras 1,500 9.11

Negro 2,500 Honduras 1,300 52.34 Nicaragua 1,200 47.66

Nelson-Saskatchewan 1,109,400 Canada 952,000 85.81 United States 157,400 14.19

Paz 2,200 Guatemala 1,400 64.47 El Salvador 770 35.53

Pedernales 360 Haiti 240 67.32 Dominican Republic 120 32.68

Rio Grande 548,800 United States 325,100 59.25 Mexico 223,600 40.75

San Juan 42,200 Nicaragua 30,400 72.02 Costa Rica 11,800 27.93

Sarstun 2,100 Guatemala 1,800 87.63 Belize 260 12.37

Sixaola 2,900 Costa Rica 2,500 88.68 Panama 290 9.96

St. Croix 4,600 United States 3,300 70.86 Canada 1,400 29.14

St. John 55,100 Canada 35,600 64.60 United States 19,400 35.25

St. Lawrence 1,055,200 Canada 559,000 52.98 United States 496,100 47.02

Stikine 50,900 Canada 50,000 98.32 United States 850 1.68

Suchiate 1,600 Guatemala 1,100 68.79 Mexico 490 31.21

Taku 18,000 Canada 17,600 98.20 United States 320 1.80

Tijuana 4,400 Mexico 3,100 70.57 United States 1,300 29.43

Whiting 2,600 Canada 2,000 80.06 United States 510 19.94

Yaqui 74,700 Mexico 70,100 93.87 United States 4,600 6.13

Yukon 829,700 United States 496,400 59.83 Canada 333,300 40.17 Total Area 8,863,060

SOUTH AMERICA

137 BASIN NAME AREA OF BASIN (km2) COUNTRY AREA (km2) PERCENT

Amacuro 4,000 Venezuela 3,400 85.15 Guyana 600 14.61

Amazon 5,866,100 Brazil 3,672,600 62.61 Peru 974,600 16.61 Bolivia 684,400 11.67 Colombia 353,000 6.02 Ecuador 137,800 2.35 Venezuela 38,500 0.66 Guyana 5,200 0.09 Suriname 20 0.00

Aviles 260 Argentina 230 88.72 Chile 30 11.28

Aysen 13,300 Chile 11,300 85.06 Argentina 2,000 14.94

Baker 30,800 Chile 21,000 68.29 Argentina 9,800 31.69

Barima 8,700 Guyana 7,700 87.86 Venezuela 1,000 11.79

Cancoso/Lauca 32,100 Bolivia 26,200 81.57 Chile 5,900 18.43

Catatumbo 26,100 Colombia 16,700 64.02 Venezuela 9,400 35.97

Chico/Carmen Silva 1,700 Argentina 1,000 59.70 Chile 680 40.30

Chira 16,700 Peru 9,200 55.33 Ecuador 7,500 44.67

Chuy 180 Brazil 110 64.57 Uruguay 60 32.57

Comau 920 Chile 840 90.91 Argentina 80 9.09

Courantyne/Corantijn 67,700 Suriname 36,900 54.46 Guyana 30,800 45.45

Cullen 590 Chile 490 83.00 Argentina 100 17.00

Essequibo 154,300 Guyana 115,400 74.79 Venezuela 38,800 25.12 Brazil 140 0.09

Gallegos-Chico 11,600 Argentina 7,000 60.15 Chile 4,600 39.85

Jurado 820 Colombia 580 70.52 Panama 240 28.75 La Plata 2,966,900 Brazil 1,366,700 46.06

138 Argentina 817,900 27.57 Paraguay 400,100 13.49 Bolivia 270,200 9.11 Uruguay 111,600 3.76

Lagoon Mirim 54,900 Uruguay 31,200 56.75 Brazil 23,700 43.18

Lake Fagnano 3,800 Argentina 2,800 74.95 Chile 950 25.05

Lake Titicaca-Poopo 116,500 Bolivia 61,700 52.99 Peru 53,600 45.96 Chile 1,200 1.05

Maroni 65,900 Suriname 37,000 56.20 French Guiana 28,100 42.66 Brazil 640 0.96

Mataje 730 Ecuador 540 73.98 Colombia 190 26.02

Mira 11,700 Colombia 7,100 61.00 Ecuador 4,600 38.99

Orinoco 958,500 Venezuela 607,400 63.37 Colombia 351,100 36.63

Oyupock/Oiapoque 27,100 Brazil 14,200 52.38 French Guiana 12,800 47.28

Palena 13,300 Chile 7,300 54.58 Argentina 6,100 45.42

Pascua 13,700 Chile 7,400 53.72 Argentina 6,300 46.22

Patia 21,300 Colombia 20,900 97.97 Ecuador 430 2.03

Puelo 8,200 Argentina 5,100 62.33 Chile 3,100 37.63

Rio Grande 7,900 Chile 4,000 50.86 Argentina 3,900 49.14

San Martin 640 Chile 580 90.22 Argentina 60 9.78

Seno Union/Serrano 6,500 Chile 5,700 87.93 Argentina 670 10.34

Tumbes-Poyango 5,000 Ecuador 3,500 71.04 Peru 1,400 28.96

Valdivia 11,400 Chile 11,300 99.09 Argentina 100 0.89

139 Yelcho 10,600 Argentina 6,900 65.06 Chile 3,700 34.88

Zapaleri 3,600 Chile 2,400 68.42 Bolivia 610 16.99 Argentina 520 14.59 Zarumilla 670 Ecuador 580 87.29 Peru 90 12.71 Total Area 10,544,710

140 APPENDIX C

Syntax from Visual Basic Program which Imports and Development, Demographic, and Ecological Variables into Working Data Set. Comment Lines Describe Procedures.

'COMMENT Demographic Variables and Water Event Data Input Programs

Sub merge_demvars()

Dim arrDemVars, arrNew Dim i, j Dim cty, cty_yr, cty1, cty2, offset Dim outsheet

'COMMENT Input Worksheets Set arrDemVars = Worksheets("NEWTEMPWKST").UsedRange Set arrNew = Worksheets("newworkingfile2").UsedRange 'create new sheet for output Set outsheet = ThisWorkbook.Sheets.Add

'COMMENT Give Column Names to Outsheet outsheet.Cells(1, 1) = "CTY" outsheet.Cells(1, 2) = "CTY1" outsheet.Cells(1, 3) = "GDP1" outsheet.Cells(1, 4) = "POP1" outsheet.Cells(1, 5) = "PDEN1" outsheet.Cells(1, 6) = "PGRO1" outsheet.Cells(1, 7) = "PURBAN1" outsheet.Cells(1, 8) = "WATPOL1" outsheet.Cells(1, 9) = "WATAXES1" outsheet.Cells(1, 10) = "HDI1" outsheet.Cells(1, 11) = "CTY2" outsheet.Cells(1, 12) = "GDP2" outsheet.Cells(1, 13) = "POP2" outsheet.Cells(1, 14) = "PDEN2"

141 outsheet.Cells(1, 15) = "PGRO2" outsheet.Cells(1, 16) = "PURBAN2" outsheet.Cells(1, 17) = "WATPOL2" outsheet.Cells(1, 18) = "WATAXES2" outsheet.Cells(1, 19) = "HDI2"

For i = 1 To arrNew.Rows.Count cty = Trim(arrNew.Cells(i, 1).Value) 'MsgBox cty cty_yr = Right(cty, 2) cty1 = Left(cty, 3) cty2 = Mid(cty, 5, 3)

'COMMENT Determine Offset from Other Vars in DemVariables Select Case cty_yr: Case Is < 70: offset = 2 Case Is < 75: offset = 3 Case Is < 80: offset = 4 Case Is < 85: offset = 5 Case Is < 90: offset = 6 Case Is < 95: offset = 7 Case Else: offset = 8 End Select

'COMMENT Search Through Demvariables for Cty1, Then Cty2 For j = 1 To arrDemVars.Rows.Count If Trim(arrDemVars.Cells(j, 1).Value) = cty1 Then 'MsgBox "For " & cty1 & vbCrLf & _ "GDP = " & arrDemVars.Cells(j, offset).Value & vbCrLf & _ "Population = " & arrDemVars.Cells(j, offset + 7).Value & vbCrLf & _ "PDEN = " & arrDemVars.Cells(j, offset + 14).Value outsheet.Cells(i + 1, 1) = "dummy" outsheet.Cells(i + 1, 1) = cty outsheet.Cells(i + 1, 2) = cty1 outsheet.Cells(i + 1, 3) = arrDemVars.Cells(j, offset).Value outsheet.Cells(i + 1, 4) = arrDemVars.Cells(j, offset + 7) outsheet.Cells(i + 1, 5) = arrDemVars.Cells(j, offset + 14) outsheet.Cells(i + 1, 6) = arrDemVars.Cells(j, offset + 21) outsheet.Cells(i + 1, 7) = arrDemVars.Cells(j, offset + 28) outsheet.Cells(i + 1, 8) = arrDemVars.Cells(j, offset + 35) outsheet.Cells(i + 1, 9) = arrDemVars.Cells(j, offset + 42) outsheet.Cells(i + 1, 10) = arrDemVars.Cells(j, offset + 49)

142

Exit For End If Next j

For j = 1 To arrDemVars.Rows.Count If Trim(arrDemVars.Cells(j, 1).Value) = cty2 Then 'MsgBox "For " & cty2 & vbCrLf & _ "GDP = " & arrDemVars.Cells(j, offset).Value & vbCrLf & _ "Population = " & arrDemVars.Cells(j, offset + 7).Value ^ vbCrLf & _ "PDEN = " & arrDemVars.Cells(j, offset + 14).Value

outsheet.Cells(i + 1, 11) = cty2 outsheet.Cells(i + 1, 12) = arrDemVars.Cells(j, offset).Value outsheet.Cells(i + 1, 13) = arrDemVars.Cells(j, offset + 7) outsheet.Cells(i + 1, 14) = arrDemVars.Cells(j, offset + 14) outsheet.Cells(i + 1, 15) = arrDemVars.Cells(j, offset + 21) outsheet.Cells(i + 1, 16) = arrDemVars.Cells(j, offset + 28) outsheet.Cells(i + 1, 17) = arrDemVars.Cells(j, offset + 35) outsheet.Cells(i + 1, 18) = arrDemVars.Cells(j, offset + 42) outsheet.Cells(i + 1, 19) = arrDemVars.Cells(j, offset + 49) Exit For End If Next j Next i

End Sub

‘COMMENT Import Water Variables

Sub merge_watervariables()

Dim arrWatVars, arrNew Dim i, j Dim cty, cty_yr, cty1, cty2, offset Dim outsheet

'COMMENT Input Worksheets Set arrWatVars = Worksheets("watervariables").UsedRange Set arrNew = Worksheets("newworkingfile").UsedRange

'COMMENT Create New Sheet for Output Set outsheet = ThisWorkbook.Sheets.Add

143

'COMMENT Give Column Names to Outsheet outsheet.Cells(1, 1) = "CTY" outsheet.Cells(1, 2) = "CTY1" outsheet.Cells(1, 3) = "TOTALH201" outsheet.Cells(1, 4) = "H20USED1" outsheet.Cells(1, 5) = "USEDPCAP1" outsheet.Cells(1, 6) = "INTWAT981" outsheet.Cells(1, 7) = "THRTFISH1" outsheet.Cells(1, 8) = "RIVFLOIN1" outsheet.Cells(1, 9) = "RIVFLOTO1" outsheet.Cells(1, 10) = "WIDRALPT1" outsheet.Cells(1, 11) = "CTY2" outsheet.Cells(1, 12) = "TOTALH202" outsheet.Cells(1, 13) = "H20USED2" outsheet.Cells(1, 14) = "USEDPCAP2" outsheet.Cells(1, 15) = "INTWAT982" outsheet.Cells(1, 16) = "THRTFISH2" outsheet.Cells(1, 17) = "RIVFLOIN2" outsheet.Cells(1, 18) = "RIVFLOTO2" outsheet.Cells(1, 19) = "WIDRALPT2"

For i = 1 To arrNew.Rows.Count cty = Trim(arrNew.Cells(i, 1).Value) 'MsgBox cty cty_yr = Right(cty, 2) cty1 = Left(cty, 3) cty2 = Mid(cty, 5, 3)

'COMMENT search through watervariables for cty1, then cty2 For j = 1 To arrWatVars.Rows.Count If Trim(arrWatVars.Cells(j, 1).Value) = cty1 Then 'MsgBox "For " & cty1 & vbCrLf & _

outsheet.Cells(i + 1, 1) = cty outsheet.Cells(i + 1, 2) = cty1 outsheet.Cells(i + 1, 3) = arrWatVars.Cells(j, 2).Value outsheet.Cells(i + 1, 4) = arrWatVars.Cells(j, 3) outsheet.Cells(i + 1, 5) = arrWatVars.Cells(j, 4) outsheet.Cells(i + 1, 6) = arrWatVars.Cells(j, 5) outsheet.Cells(i + 1, 7) = arrWatVars.Cells(j, 6) outsheet.Cells(i + 1, 8) = arrWatVars.Cells(j, 7) outsheet.Cells(i + 1, 9) = arrWatVars.Cells(j, 8)

144 outsheet.Cells(i + 1, 10) = arrWatVars.Cells(j, 9)

Exit For End If Next j

For j = 1 To arrWatVars.Rows.Count If Trim(arrWatVars.Cells(j, 1).Value) = cty2 Then outsheet.Cells(i + 1, 11) = cty2 outsheet.Cells(i + 1, 12) = arrWatVars.Cells(j, 2).Value outsheet.Cells(i + 1, 13) = arrWatVars.Cells(j, 3) outsheet.Cells(i + 1, 14) = arrWatVars.Cells(j, 4) outsheet.Cells(i + 1, 15) = arrWatVars.Cells(j, 5) outsheet.Cells(i + 1, 16) = arrWatVars.Cells(j, 6) outsheet.Cells(i + 1, 17) = arrWatVars.Cells(j, 7) outsheet.Cells(i + 1, 18) = arrWatVars.Cells(j, 8) outsheet.Cells(i + 1, 19) = arrWatVars.Cells(j, 9) Exit For End If Next j Next i

End Sub

145 APPENDIX D List of Pairs of Countries in Final Data Set (continued). PAIR PAIR PAIR PAIR PAIR PAIR PAIR PAIR AFG_CHN ARG_CHL BDI_ZMB BGR_ROM BRA_SUR CHE_YUG COG_GNQ DEU_YUG AFG_IND ARG_PRY BEL_CHE BGR_SVK BRA_URY CHL_PER COG_MWI DJI_ETH AFG_IRN ARG_URY BEL_DEU BGR_SVN BRA_VEN CHN_IND COG_RWA DJI_SOM AFG_KAZ ARM_RUS BEL_FRA BGR_TUR BRN_MYS CHN_KAZ COG_TZA DOM_HTI AFG_KGZ ARM_TUR BEL_GFR BGR_UKR BTN_CHN CHN_KGZ COG_ZAR DZA_GIN AFG_PAK ARM_AZE BEL_ITA BGR_USR BTN_IND CHN_KHM COG_ZMB DZA_LBY AFG_TJK ARM_GEO BEL_NLD BGR_YGF BTN_MMR CHN_LAO COL_ECU DZA_MAR AFG_TKM ARM_IRN BEN_BFA BGR_YUG BTN_NPL CHN_MMR COL_GUY DZA_MLI AFG_USR AUT_BGR BEN_CIV BIH_CHE BWA_LSO CHN_MNG COL_PAN DZA_NER AGO_BDI AUT_BIH BEN_CMR BIH_CZE BWA_MOZ CHN_NPL COL_PER DZA_NGA AGO_BWA AUT_CHE BEN_DZA BIH_DEU BWA_MWI CHN_PAK COL_SUR DZA_SDN AGO_CAF AUT_CZE BEN_GHA BIH_HRV BWA_NA CHN_PRK COL_VEN DZA_SLE AGO_CMR AUT_CZS BEN_GIN BIH_HUN M CHN_RUS CRI_NIC DZA_TCD AGO_COG AUT_DEU BEN_MLI BIH_ITA BWA_TZA CHN_THA CRI_PAN DZA_TUN AGO_GAB AUT_FRA BEN_NER BIH_MDA BWA_ZAF CHN_TJK CZE_DEU ECU_GUY AGO_MOZ AUT_GDR BEN_NGA BIH_POL BWA_ZAR CHN_TKM CZE_HRV ECU_PER AGO_MWI AUT_GFR BEN_SLE BIH_ROM BWA_ZMB CHN_USR CZE_HUN ECU_SUR AGO_NAM AUT_HRV BEN_TCD BIH_SVK BWA_ZWE CHN_UZB CZE_ITA ECU_VEN AGO_RWA AUT_HUN BEN_TGO BIH_SVN CAF_CMR CHN_VNM CZE_MDA EGY_ERI AGO_TZA AUT_ITA BFA_CIV BIH_UKR CAF_COG CIV_CMR CZE_POL EGY_ETH AGO_ZAR AUT_MDA BFA_CMR BIH_YUG CAF_DZA CIV_DZA CZE_ROM EGY_ISR AGO_ZMB AUT_NLD BFA_DZA BLR_CZE CAF_GAB CIV_GHA CZE_SVK EGY_JOR AGO_ZWE AUT_POL BFA_GHA BLR_EST CAF_LBY CIV_GIN CZE_SVN EGY_KEN ALB_AUT AUT_ROM BFA_GIN BLR_KAZ CAF_MWI CIV_LBR CZE_UKR EGY_LBN ALB_BGR AUT_SVK BFA_MLI BLR_LTU CAF_NER CIV_MLI CZE_YUG EGY_RWA ALB_BIH AUT_SVN BFA_NER BLR_LVA CAF_NGA CIV_NER CZS_DEU EGY_SDN ALB_CHE AUT_UKR BFA_NGA BLR_POL CAF_RWA CIV_NGA CZS_GDR EGY_SYR ALB_CZE AUT_USR BFA_SLE BLR_RUS CAF_SDN CIV_SLE CZS_GFR EGY_TZA ALB_CZS AUT_YGF BFA_TCD BLR_SVK CAF_TCD CIV_TCD CZS_HRV EGY_UGA ALB_DEU AUT_YUG BFA_TGO BLR_UKR CAF_TZA CIV_TGO CZS_HUN EGY_ZAR ALB_GDR AZE_GEO BGD_BTN BLZ_GTM CAF_ZAR CMR_COG CZS_MDA ERI_ETH ALB_GFR AZE_IRN BGD_CHN BLZ_MEX CAF_ZMB CMR_DZA CZS_POL ERI_KEN ALB_GRC AZE_RUS BGD_IND BOL_BRA CAN_USA CMR_GAB CZS_ROM ERI_RWA ALB_HRV AZE_TUR BGD_MMR BOL_CHL CHE_CZE CMR_GIN CZS_SVN ERI_SDN ALB_HUN BDI_CAF BGD_NPL BOL_COL CHE_CZS CMR_GNQ CZS_UKR ERI_TZA ALB_ITA BDI_CMR BGR_BIH BOL_ECU CHE_DEU CMR_LBY CZS_USR ERI_UGA ALB_MDA BDI_COG BGR_CHE BOL_GUY CHE_FRA CMR_MLI CZS_YGF ERI_ZAR ALB_MKD BDI_EGY BGR_CZE BOL_PER CHE_GFR CMR_MWI DEU_FRA ESP_FRA ALB_POL BDI_ERI BGR_CZS BOL_PRY CHE_HRV CMR_NER DEU_HRV ESP_PRT ALB_ROM BDI_ETH BGR_DEU BOL_SUR CHE_HUN CMR_NGA DEU_HUN EST_LVA ALB_SVK BDI_GAB BGR_GFR BOL_URY CHE_ITA CMR_RWA DEU_ITA EST_RUS ALB_SVN BDI_KEN BGR_GRC BOL_VEN CHE_MDA CMR_SDN DEU_MDA ETH_KEN ALB_UKR BDI_MWI BGR_HRV BRA_COL CHE_NLD CMR_SLE DEU_NLD ETH_RWA ALB_USR BDI_RWA BGR_HUN BRA_ECU CHE_POL CMR_TCD DEU_POL ETH_SDN ALB_YGF BDI_SDN BGR_ITA BRA_GUF CHE_ROM CMR_TZA DEU_ROM ETH_SOM ALB_YUG BDI_TZA BGR_MDA BRA_GUY CHE_SVK CMR_ZAR DEU_SVK ETH_TZA ARG_BOL BDI_UGA BGR_MKD BRA_PER CHE_SVN CMR_ZMB DEU_SVN ETH_UGA ARG_BRA BDI_ZAR BGR_POL BRA_PRY CHE_UKR COG_GAB DEU_UKR ETH_ZAR CHE_USR

146 APPENDIX E

Countries and Standard Country Codes.

Country Code Afghanistan AFG Albania ALB Algeria DZA Andorra ADO Angola AGO Argentina ARG Armenia ARM Austria AUT Azerbaidzhan AZE Bangladesh BGD Belarus BLR Belgium BEL Belize BLZ Benin BEN Bhutan BTN Bolivia BOL Bosnia and Herzegovina BIH Botswana BWA Brazil BRA Brunei BRN Bulgaria BGR Burkina Faso BFA Burundi BDI Cambodia KHM Cameroon CMR Canada CAN Central African Republic CAF Chad TCD Chile CHL China CHN Colombia COL Congo COG Congo Dem Republic ZAR

147 Costa Rica CRI Cote D'Ivoire CIV Croatia HRV Czech Republic CZE Djibouti DJI Dominican Republic DOM Ecuador ECU Egypt EGY El Salvador SLV Equatorial Guinea GNQ Eritrea ERI Estonia EST Ethiopia ETH Finland FIN France FRA French Guiana GUF Gabon GAB Gambia GMB Georgia GEO Germany DEU Ghana GHA Greece GRC Guatemala GTM Guinea-Bissau GNB Guinea GIN Guyana GUY Haiti HTI Honduras HND Hungary HUN India IND Indonesia IDN Iran IRN Iraq IRQ Ireland IRL Israel ISR Italy ITA Jordan JOR Kazakhstan KAZ Kenya KEN Korea, Demo People’s Rep. PRK Korea, Republic of KOR Kyrgyzstan KGZ Laos LAO

148 Latvia LVA Lebanon LBN Lesotho LSO Liberia LBR Libya LBY Liechtenstein LIE Lithuania LTU Luxembourg LUX Macedonia MKD Malawi MWI Malaysia MYS Mali MLI Mauritania MRT Mexico MEX Moldova MDA Mongolia MNG Morocco MAR Mozambique MOZ Myanmar MMR Namibia NAM Nepal NPL Netherlands NLD Nicaragua NIC Niger NER Nigeria NGA Norway NOR Pakistan PAK Panama PAN Papua New Guinea PNG Paraguay PRY Peru PER Poland POL Portugal PRT Romania ROM Russia RUS Rwanda RWA Saudi Arabia SAU Senegal SEN Sierra Leone SLE Slovakia SVK Slovenia SVN Somalia SOM South Africa ZAF

149 Spain ESP Sudan SDN Suriname SUR Swaziland SWZ Sweden SWE Switzerland CHE Syria SYR Tadjikistan TJK Tanzania TZA Thailand THA Togo TGO Tunisia TUN Turkey TUR Turkmenistan TKM Uganda UGA Ukraine UKR United Kingdom GBR United States USA Uruguay URY Uzbekistan UZB Venezuela VEN Vietnam VNM Western Sahara WSH Yugoslavia YUG Zambia ZMB Zimbabwe ZWE

150 REFERENCES

Arizpe, Lourdes and Margarita Velázquez. 1994. "The Social Dimensions of Population." Pp. 15-40 in Population and Environment: Rethinking the Debate, Lourdes Arizpe, Priscilla M. Stone, and David C. Major. Boulder, Colorado: Westview Press, Inc.

Assaf, Karen, Nader al Khatib, Elisha Kally and Hillel Shuval 1993. A Proposal for the Development of a Regional Water Master Plan. Israel/Palestine Center for Research and Information. Jerusalem.

Azar, Edward E. Conflict and Peace Data Bank (COPDAB), 1948-1978. [Computer file]. 3rd release. 1993. Center for International Development and Conflict Management. College Park, MD.

Baran, Paul A. 1957. The Political Economy of Growth. New York: Monthly Review.

Barbosa, Luiz C. 1996. "The People of the Forest Against International Capitalism: Systemic and Anti-Systemic Forces in the Battle for the Preservation of the Brazilian Amazon Rainforest." Sociological Perspectives 39(2):317-31.

Barnett, Harold J. 1974. "Economic Growth and Environmental Quality Are Compatible." Policy Sciences 5:137-47.

Begum, Khuyrshid. 1988. Tension Over the Farakka Barrage. Stuttgart: Steiner Verlag Wiesbaden GMBH.

Biswas, Asit K., John Kolars, Masahiro Murakami, and Aaron Wolf. 1997. Core and Periphery: A Comprehensive Approach to Middle Eastern Water. Oxford: Oxford University Press.

Bongaarts, John. 1996. "Population Pressure and the Food Supply System in the Developing World." Population and Development Review 22(3):483-503.

Boserup, Ester. 1996. "Development Theory: An Analytical Framework and Selected Applications." Population and Development Review 22(3):505-15.

151 ———. 1981. Population and Technological Change: A Study of Long Term Trends, Chicago: The University of Chicago Press.

Calhoun, Craig. 1993. "Nationalism and Ethnicity." Annual Review of Sociology 19:211- 39.

Chirot, Daniel and Thomas D. Hall. 1982. "World System Theory." Annual Review of Sociology 8:81-106.

Central Intelligence Agency. 2000. Foreign Broadcast Information Service.

______. 2000. World News Connection.

Clark, Colin W. 1973. "The Economics of Overexploitation." Science 181:630-634.

Crenshaw, Edward M. and J. C. Jenkins. 1996. "Social Structure and Global Climate Change: Sociological Propositions Concerning the Greenhouse Effect." Sociological Focus 29(4):341-58.

Crow, Ben, Alan Lindquist, and David Wilson. 1995. Sharing the Ganges: The Politics and Technology of River Development. New Delhi: Sage Publications.

Crow, Ben and Farhana Sultana. 2000. "Water, Power, and Gender: Pressing Questions and Overlooked Interests in a Poor and Overcrowded Delta." (Washington, D.C.)American Sociological Association.

Dabelko, Geoffrey D. 1996. The Environment and Conflict in the Third World: Examining Linkage, Context and Policy. Occasional Paper No. 12. Harrison Program on the Future Global Agenda.

Davidson, Debra J. 1999. "Implementing the Endangered Species Act on Private Land in the Coastal Redwoods: A Political-Sociological Analysis of Environmental Bureaucracies." University of Wisconsin, Madison.

Davies, John L. 1998. The Global Event-Data System: Coders’ Manual, Center for International Development and Conflict Management and Department of Government and Politics. College Park, MD.

Davis, Kingsley and Wilbert E. Moore. 1945. "Some Principles of Stratification." American Sociological Review 10:242-49.

152 De Cordier, Bruno. 1996. "Ethnic Conflict and Ecological Degradation in Central Asia: The Ferghana Valley and the North of Kazakhstan." Central Asian Survey 15(3- 4):399-411. dos Santos, . 1971. “The Structure of Dependence.”

Dobkowski, Michael N. and Isidor Wallimann. 1998. "The Coming Age of Scarcity: An Introduction." Pp. 1-20 in The Coming Age of Scarcity: Preventing Mass Death and Genocide in the Twenty-First Century, Michael N. Dobkowski and Isidor Wallimann. Syracuse, New York: Syracuse University Press.

Donahue, John M. and Barbara R. Johnston. 1998. Water, Culture, and Power: Local Struggles in a Global Context. Washington, D.C.: Island Press.

Duchin, Faye. 1996. "Population Change, Lifestyle, and Technology: How Much Difference Can They Make?" Population and Development Review 22(3):321-30.

Duncan, Otis D. 1959. "Human Ecology and Population Studies." Pp. 678-716 in The Study of Population: An Inventory and Appraisal, Philip M. Hauser and Otis D. Duncan. University of Chicago Press.

Dunlap, Riley and William R. Catton Jr. 1979. "Environmental Sociology." Annual Review of Sociology 5:243-73.

Durkheim, Émile. 1965 [1902]. The Division of Labor in Society. New York: Free Press.

Ehrhardt-Martinez, Karen, Edward M. Crenshaw, and J. Craig Jenkins. 2002. “Deforestation and the Environmental Kuznets Curve: A Cross-National Investigation of Intervening Mechanisms.” Social Science Quarterly. 83: 226- 243.

Elliot, Gil. 1972. Twentieth Century Book of the Dead. New York: Scribners.

Farinelli, Xavier H. 1997. Freshwater Conflicts in the Jordan River Basin. Geneva: Green Cross International.

Fitzmaurice, John 1996. Damming the Danube: Gabcikovo and Post-Communist Politics in Europe. Westview Press. Boulder, CO.

Gardner-Outlaw, Tom and Robert Engelman. 1997. Sustaining Water, Easing Scarcity: A Second Update. Revised Data for the Population Action International Report, Sustaining Water: Population and the Future of Renewable Water Supplies. Washington, D.C.: Population Action International.

153 Gleick, Peter H. 1993. "Water and Conflict: Fresh Water Resources and International Security." International Security 18(1):79-112.

______. 1998. The World’s Water: The Biennial Report on Freshwater Resources. Island Press. Washington, DC.

Goldstone, Jack. 1980. "Population and Revolution." American Sociological Association.

Goldstone, Jack A. 1999. "Imminent Political Conflicts Arising From China's Population Crisis." Pp. 247-66 in Contested Grounds: Security and Conflict in the New Environmental Politics, Daniel H. Deudney and Richard A. Matthew. Albany, NY: State University of New York Press.

Gurr, Ted R. and Barbara Harff. 1994. Ethnic Conflict in World Politics. Boulder, Colorado: Westview Press.

Haftendorn, Helga. 2000. "Water and International Conflict." Third World Quarterly 21(1):51-68.

Hardin, Garrett. 1968. "The Tragedy of the Commons." Science 162:1243-48.

Harper, Charles L. 1996. Environment and Society: Human Perspectives on Environmental Issues. Upper Saddle River, New Jersey: Prentice-Hall.

Hawley, Amos. 1968. "Human Ecology." International Encycopedia of Social Sciences 4:328-37.

Homer-Dixon, Thomas F. 1999. Environment, Scarcity and Violence. Princeton, New Jersey: Princeton University Press.

———. 1994. "Environmental Scarcities and Violent Conflict: Evidence From Cases." International Security 19(1):5-40.

———. 1991. "On the Threshold: Environmental Changes As Causes of Acute Conflict." International Security 16(2):76-116.

Humphrey, Craig R. and Frederick R. Buttel. 1982. Environment, Energy, and Society. Belmont, California: Wadsworth Publishing Company.

Ingram, Helen, Nancy K. Laney, and David M. Gillian. 1995. Divided Waters: Bridging the U.S. Mexico Border. Tucson: University of Arizona Press.

154 Kamyar, Mehdiyoun. 2000. "International Law and the Dispute Over Ownership of Oil and Gas Resources in the Caspain Sea." American Journal of International Law 94(1):179.

Kinzer, Stephen. 28 Feb 2000. "Where Kurds Seek a Land, Turks Want the Water." New York Times (New York).

Leal, Donald R. 1998. "Community-Run Fisheries: Avoiding the Tragedy of the Commons." Population and Environment: A Journal of Interdisciplinary Studies 19(3):225-45.

Lipovsky, Igor. 1995. "The Deterioration of the Ecological Situation in Central Asia: Causes and Possible Consequences." Europe-Asia Studies 47(7):1109-23.

Lowi, Miriam R. 1993. Water and Power: the Politics of a Scarce Resource in the Jordan River Basin. Cambridge University Press. Cambridge.

Malhotra, Kamal. 1998. "Economic Crisis Will Intensify Conflict Over Natural Resources." Promoting Dialogue on the Natural Resources of the Lower Mekong (Bangkok, Thailand)American Friends Service Committee.

Marx, Karl and Friedrich Engels. 1962. Manifesto of the Communist Party. New York: International Publishers.

Meyerson, Frederick A. B. 1998. "Population, Development, and Global Warming: Averting the Tragedy of the Climate Commons." Population and Environment: A Journal of Interdisciplinary Studies 19(5):443-63.

Murphy, Irene L. 1997. The Danube: A River Basin in Transition. Boston: Kluwer Academic Publishers.

Myers, Norman. 1997. "Environmental Refugees." Population and Environment: A Journal of Interdisciplinary Studies 19(2):167-82.

Namboodiri, Krishnan. 1988. "Ecological Demography: Its Place in Sociology." Annual Review of Sociology 53:619-33.

Olzak, Susan. 1992. Dynamics of Ethnic Competition and Conflict. Stanford, California: Stanford University Press.

Oneal, John R., Frances H. Oneal, Zeev Maoz, and Bruce Russett. 1996. "The Liberal Peace: Interdependence, Democracy, and International Conflict, 1950-85." Journal of Peace Research 33(1):11-28.

155 Ophuls, William. 1977. Ecology and the Politics of Scarcity: A Prologue to a Political Theory of the Steady State. San Francisco, CA: Freeman.

Ottoway, David B. 16 May 2000. "Vast Caspian Oil Field Found; Discovery May Spur U.S.-Russian Pipeline Rivalry." Washington Post (Washington, D.C.), p. A1.

Pimentel, David and Mario Giampietro. 1994. Food, Land, Population, and the U.S. Economy. Washington, D.C.: Carrying Capacity Network.

Pimentel, David, Xuewen Huang, Ana Cordova, and Marcia Pimentel. 1997. "Impact of Population Growth on Food Supplies and Environment." Population and Environment: A Journal of Interdisciplinary Studies 19(1):9-14.

Porter, Gareth and Janet W. Brown. 1996. Global Environmental Politics. Colorado: Westview Press.

Postel, Sandra. 1996. Dividing the Waters: Food Security, Ecosystem Health, and the New Politics of Scarcity. New York: Worldwatch Institute.

Price, David. 1998. "Of Population and False Hope: Malthus and His Legacy." Population and Environment: A Journal of Interdisciplinary Studies 19(3):205- 19.

Rahim, Nafissa A., Tafesse Hailu, Bernadette Ouédraogo, Babacar Touré, Mahamat Hissène, Berhane Woldegabriel, Zeremariam Fre, Nhial B. Aken, Boubakar Ba, Omar Mohammed, and Cheik K. Maïga. 1991. Greenwar: Environment and Conflict. London: Panos Publications Ltd.

Reisner, Marc. 1986. Cadillac Desert : the American West and its Disappearing Water. Viking. New York.

Roberts, J. T. and Peter E. Grimes. 1997. "Carbon Intensity and Economic Development 1962-91: A Brief Exploration of the Environmental Kuznets Curve." World Development 25(2):191-98.

Saliba, Samir N. 1968. The Jordan River Dispute. Martinus Nijhoff. The Hague.

Samson, Paul and Bertrand Charrier. 1997. International Freshwater Conflict: Issues and Prevention Strategies. Geneva: Green Cross International.

Scheumann, Waltina. 1998. "Conflicts on the Euphrates: An Analysis of Water and Non-water Issues." In Scheuman and Schiffler, Water in the Middle East: Potential for Conflicts and Prospects for Cooperation. Springer Press. Berlin.

156 Scheumann, Waltina and Manuel Schiffler. 1998. Water in the Middle East: Potential for Conflicts and Prospects for Cooperation. Berlin: Springer Press.

Schiffler, Manuel. 1998. “International Water Agreements: A Comparative View.” In Scheuman and Schiffler, Water in the Middle East: Potential for Conflicts and Prospects for Cooperation. Springer Press. Berlin.

Schnaiberg, Allan and Alan Gould. 1994. Environment and Society: The Enduring Conflict. New York: St. Martin's Press.

Shaw, Alison. 1999. "What Are 'They' Doing to Our Food?": Public Concerns About Food in the UK." Sociological Research Online 4(3).

Shiva, Vandana. 1991. Ecology and the Politics of Survival: Conflicts Over Natural Resources in India. New Delhi: Sage Publications and United Nations University Press.

Smith, Adam. 1937 [1778]. An Inquiry into the Nature and Causes of the Wealth of Nations. New York: The Modern Library.

Smith, Joseph W. and Gary Sauer-Thompson. 1998. "Civilizations' Wake: Ecology, Economics, and the Roots of Environmental Destruction and Neglect." Population and Environment: A Journal of Interdisciplinary Studies 19(6):541-75.

Sonnenfeld, David A. 2002. “Social Movements and Ecological Modernization: The Transformation of Pulp and Paper Manufacturing.” Development and Change. 33: 1-27.

Sosland, Jeffrey K. 1998. "Cooperating Rivals: The Politics of Water Scarcity, Protracted Conflict, and Complex Cooperation in the Jordan River Basin." Georgetown University, Washington, DC.

Spencer, Herbert. 1883. First Principles. New York: D. Appleton.

Starkloff, Ralf. 1998. "Water Scarcity in Kitulwatte: The Social Causes and Consequences of Environmental Degradation in a Highland Uva Village of Sri Lanka." World Development 26(6):913-32.

Stern, David I., Michael S. Common, and Edward B. Barbier. 1996. "Economic Growth and Environmental Degradation: The Environmental Kuznets Curve and Sustainable Development." World Development 24(7):1151-60.

157 Stevens, Georgiana G. 1965. Jordan River Partition. Hoover Institute. Stanford.

Swain, Ashok. 1996. "Displacing the Conflict: Environmental Destruction in Bangladesh and Ethnic Conflict in India." Journal of Peace Research 33(2):189-204.

Swart, Rob. 1996. "Security Risks of Global Environmental Changes." Global Environmental Change 6(3):187-92.

Trainer, Ted. 1998. "Our Unsustainable Society: Basic Causes, Interconnections, and Solutions." Pp. 83-100 in The Coming Age of Scarcity: Preventing Mass Death and Genocide in the Twenty-First Century, Michael N. Dobkowski and Isidor Wallimann. Syracuse, New York: Syracuse University Press.

United Nations Development Programme. 1991. Human Development Report 1999. Oxford University Press. New York.

United Nations Development Programme. 1997. Human Development Report 1999. Oxford University Press. New York.

United Nations Development Programme. 2000. Human Development Report 1999. Oxford University Press. New York.

Verhovek, Sam H. 4 Jul 1998. "Pollution Puts People in Peril on the Border With Mexico." New York Times (New York).

Wallerstein, Immanuel. 1979. The Capitalist World-Economy : Essays. Cambridge University Press. New York.

Walsh, Edward J. 1985. "Three Mile Island: The Battle of Unit One." The Bulletin of the Atomic Scientists 41(5):30-31.

Whitmeyer, Joseph M. and Rosemary L. Hopcroft. 1996. "Community, Capitalism, and Rebellion in Chiapas." Sociological Perspectives 39(4):517-38.

Williams Jr., Robin M. 1994. "The Sociology of Ethnic Conflicts: Comparative International Perspectives." Annual Review of Sociology 20:49-79.

Wolf, Aaron. 1995. Hydropolitics Along the Jordan River: Scarce Water and Its Impact on the Arab-Israeli Conflict. New York: United Nations University Press.

______. 1999. “The Transboundary Freshwater Dispute Database Project.” Water International 24: 160-163.

158 ______. 2002. Transboundary Freshwater Dispute Database. http://terra.geo.orst.edu/users/tfdd/

Wolf, Aaron, Jeffrey natharius, Jeffrey Danielson, Brian S. Ward, and Jan K. Pender. 1999. “International River Basins of the World.” International Journal of Water Resources Development. 15(4): 387-427.

World Bank. 2001. World Development Indicators CD-ROM.

World Resources Institute. 1998. World Resources: 1998-99. Oxford University Press. New York.

Yoffe, Shira. 2001. Basins At Risk: Conflict and Cooperation Over International Freshwater Resources. Dissertation at Oregon State University.

Yoffe, Shira and Kelli Larson. 2001. “Basins at Risk: Water Event Database Methodology.” Unpublished article.

159