Peace by Repatriation Concepts, Cases, and Conditions Patrik Johansson P

Peace by Repatriation Concepts, Cases, and Conditions Patrik Johansson P
Patrik Johansson
Peace by Repatriation: Concepts, Cases, and Conditions
Peace by Repatriation
Concepts, Cases, and Conditions
Patrik Johansson
ISSN 0349-0831
ISBN 978-91-7459-107-1
UMINF 08.01
Umeå University 2010
Department of Political Science
Umeå University
SE-901 87 Umeå, Sweden
Department of Political Science
Umeå University
Peace by Repatriation
Concepts, Cases, and Conditions
Patrik Johansson
Department of Political Science
901 87 Umeå
Umeå 2010
PEACE BY REPATRIATION
Copyright©Patrik Johansson
ISBN: 978-91-7459-107-1
ISSN: 0349-0831
Department of Political Science, Research Report, 2010:3
Cover photo by Nadine Walicki: IDP settlement Zubchuk, near Fizuli, Azerbaijan,
with Karabakh mountains in the background.
Printed by: Print & Media, Umeå University, 2008835
Umeå, Sweden, 2010
To Paul Johansson
1933-2009
PEACE BY REPATRIATION
PEACE BY REPATRIATION
Abstract
The focus of this study is the assumption that the return of refugees is a
necessary condition for the establishment of sustainable peace after armed
conflict. This assumption is often made in the peacebuilding literature as
well as by policy makers, but it has rarely been the object of systematic
analysis. The purpose of the study, therefore, is to test this assumption,
which I label the “peace-by-repatriation thesis.”
I adopt a two-step approach to analyzing the peace-by-repatriation
thesis. The first step is to formulate an analytical framework. The second
step is to use the framework to test the peace-by-repatriation thesis on a
medium number of cases. The formulation of the analytical framework starts
with an examination of previous research. I trace the theoretical foundations
of the peace-by-repatriation thesis in research on peacebuilding, forced
migration, and partition. The analytical framework is further informed by
case studies of Bosnia-Herzegovina and Nagorno-Karabakh, two cases that
represent opposing perceptions of repatriation as a condition for peace.
I adopt a set-theoretic approach to test the peace-by-repatriation thesis.
I describe in some detail how the key concepts of the analytical framework
are operationalized. I select and code forty-three cases—terminated conflicts
that caused large-scale displacement—and use fuzzy-set analysis to test the
peace-by-repatriation thesis. The analysis shows that repatriation is not a
necessary condition for sustainable peace. Instead, ending displacement—
irrespective of how this is done—turns out to be an important condition for
peace. This result is consistent across tests of different combinations of cases
and tests using alternative operationalizations of key concepts.
Taken together, the fuzzy-set analysis and the case studies suggest that
the relationship between repatriation and peace will vary from case to case
and that pre-war interethnic relations is one of the circumstances that affect
that relationship.
Keywords
Refugees, displaced persons, repatriation, sustainable peace, necessary
conditions, fuzzy sets, Bosnia-Herzegovina, Nagorno-Karabakh.
i
PEACE BY REPATRIATION
ii
PEACE BY REPATRIATION
Table of Contents
Abstract
Table of Contents
Tables, Figures, and Maps
Abbreviations
Preface
i
iii
v
vii
ix
1
1.1
1.2
1.3
1.4
Introduction
The Peace-by-Repatriation Thesis
The Purpose of the Study
Context and Relevance
Outline of the Study
1
1
3
4
11
2
2.1
2.2
2.3
Theorizing Peace by Repatriation
Introduction
Previous Research
An Analytical Framework of Peace by Repatriation
13
13
16
27
3
Case Studies of Peace by Repatriation
3.1 The Purpose of the Case Studies
3.2 The Structure and Focus of the Case Studies
Bosnia-Herzegovina and the Primacy
of Minority Return
4.1 Introduction
4.2 Implementing Peace in Bosnia-Herzegovina
4.3 Peace by Repatriation in Bosnia-Herzegovina
30
30
32
4
37
37
42
53
5
Nagorno-Karabakh and the Search
for Pragmatic Solutions
5.1 Introduction
5.2 Negotiating Peace in Nagorno-Karabakh
5.3 Peace by Repatriation in Nagorno-Karabakh
57
57
62
72
6
6.1
6.2
6.3
78
78
78
84
Case Study Analysis
Summary Results of the Case Studies
Implications for the Peace-by-Repatriation Framework
Conclusions
iii
PEACE BY REPATRIATION
7
7.1
7.2
7.3
A Comparative Methodology
A Set-Theoretic Argument
The Methodology of Necessary Conditions
Fuzzy Sets
8
8.1
8.2
8.3
8.4
8.5
8.6
8.7
From Theoretical Concepts to Fuzzy Sets
Concept Structures
Sustainable Peace
Repatriation
Ended Displacement Situation
Partition
Fractionalization
Group Concentration
85
85
88
96
109
109
114
120
125
128
130
132
9 Case Selection and Coding
9.1 Selection of Cases
9.2 The Coding Procedure
135
135
140
10 Systematic Comparative Analysis
10.1 Testing the Peace-by-Repatriation Thesis
10.2 All Conflicts
10.3 Conflicts over Territory
10.4 Conflicts over Government
10.5 Findings
10.6 Modifying the Analysis
153
153
155
163
164
166
168
11 Summary and Conclusions
11.1 Summary
11.2 Conclusions
176
176
177
References
183
Appendix 1:
Appendix 2:
Appendix 3:
Appendix 4:
Appendix 5:
iv
Interview Guide
Fuzzy Membership Scores
Coding Sustainable Peace
Complex versus Parsimonious Solutions
Truth Table Solutions
211
212
214
216
218
PEACE BY REPATRIATION
Tables, Figures, and Maps
Tables
Table 7.1
Table 7.2
Table 7.3
Table 7.4
Table 7.5
Table 7.6
Table 7.7
Table 7.8
Table 7.9
Table 7.10
A basic test of necessity
Twins always resemble each other
Hypothetical dataset on ethnic rebellion:
dichotomous data
Hypothetical truth table based on dichotomous data
Crisp versus fuzzy sets
Hypothetical dataset on ethnic rebellion: fuzzy data
Formulas for computing consistency and coverage
Hypothetical truth table based on fuzzy data
Truth table solution on ethnic rebellion
Membership in configurations
92
93
97
99
103
105
105
107
Table 8.1
Table 8.2
Table 8.3
Table 8.4
Table 8.5
Table 8.6
Table 8.7
Table 8.8
Table 8.9
The logic of concept structures
Definitions of peace
The fuzzy set of “sustainable peace”
The fuzzy set of “repatriation”
The fuzzy set of “ended displacement situation”
Outcomes of ethnic civil wars
The fuzzy set of “partition”
The fuzzy set of “fractionalization”
The fuzzy set of “group concentration”
111
115
120
124
128
129
130
132
133
Table 9.1
Table 9.2
A preliminary list of cases
Refugees and asylum seekers from Tajikistan,
1994–2002
Coding of selected cases
138
Table 9.3
Table 10.1
Table 10.2
Table 10.3
Table 10.4
Table 10.5
Table 10.6
Table 10.7
Table 10.8
Number of cases in different tests
Abbreviations of concepts used in the analysis
Consistent solutions—all conflicts
Consistent solutions—conflicts over territory
Consistent solutions—conflicts over government
Consistent solutions—alternative
operationalization of “sustainable peace”
Average membership scores
Alternative operationalization of the fuzzy set
of “ended displacement situation”
89
90
148
151
154
155
162
164
166
170
170
171
v
PEACE BY REPATRIATION
Table 10.9
Table 10.10
Consistent solutions—alternative
operationalization of “ended displacement situation”
Consistent solutions—alternative
operationalization of “fractionalization”
Figures
Figure 7.1
Figure 7.2
Figure 7.3
Figure 10.1
Figure 10.2
Figure 10.3
Fuzzy-set analysis of necessity, XY plot
Consistency and coverage, Venn diagrams
Measuring consistency and coverage, XY plots
Fuzzy-set analysis of peace by repatriation, XY plot
XY plot of membership in PEACE/REPALL
XY plot of membership in PEACE/ENDALL
Maps
Map 4.1
Map 5.1
Map 5.2
Bosnia-Herzegovina
Azerbaijan
Nagorno-Karabakh, conflict area
vi
172
172
99
101
103
156
157
167
38
58
59
PEACE BY REPATRIATION
Abbreviations
CIRI
CRPC
cs/QCA
DPA
fs/QCA
ICFY
ICG
IDP
MAR
NATO
NGO
NKAO
NKR
OHR
OSCE
PIC
PLIP
QCA
SFOR
SFRY
SSR
UCDP
UN
UNHCR
UNMIBH
UNMIK
USCRI
USSR
WRS
Cingranelli-Richards (Human Rights Index)
Commission for Real Property Claims of Displaced Persons
and Refugees
crisp-set Qualitative Comparative Analysis
Dayton Peace Agreement
fuzzy-set Qualitative Comparative Analysis
International Conference on the Former Yugoslavia
International Crisis Group
Internally Displaced Person
Minorities at Risk
North Atlantic Treaty Organization
Non-Governmental Organization
Nagorno-Karabakh Autonomous Oblast
Nagorno-Karabakh Republic
Office of the High Representative
Organization for Security and Cooperation in Europe
Peace Implementation Council
Property Law Implementation Plan
Qualitative Comparative Analysis
Stabilisation Force
Socialist Federal Republic of Yugoslavia
Soviet Socialist Republic
Uppsala Conflict Data Program
United Nations
United Nations High Commissioner for Refugees
United Nations Mission in Bosnia and Herzegovina
United Nations Interim Administration Mission in Kosovo
United States Committee for Refugees and Immigrants
Union of Soviet Socialist Republics
World Refugee Survey
vii
PEACE BY REPATRIATION
viii
PEACE BY REPATRIATION
Preface
The idea to this dissertation originally developed in reaction to the
international community’s different approaches to addressing displacement
in the Balkans and the Middle East. In the case of Bosnia-Herzegovina the
right of every displaced person to return home was a central feature of the
Dayton Peace Agreement of 1995, not because this is what the parties to the
negotiations necessarily wanted, but because of the insistence on such a right
by the international community. Only two years earlier, in the Oslo Accords
of 1993, the question of the Palestinian refugees was one of the so called final
status issues, explicitly left for Israel and the PLO to negotiate and reach
agreement on, with the international community neither then nor later
expressing a firm position on what such an agreement should entail, other
than repeating that it should be “just.”
Why these different approaches, I wondered. Because those displaced
from Bosnia ended up in Europe rather than in Arab countries? Because of
the special relationship between Israel and the United States? Because the
passage of time since the creation of the Palestinian refugee problem sixty
years ago has rendered return impracticable? These answers appeared either
cynical, biased, or likely to inspire those opposed to return to remain
recalcitrant, as time is on their side (none of which qualities necessarily
makes them wrong). Or could the different approaches to displacement be
adequately explained by other differences between the two cases?
As these thoughts developed into a broader research problem regarding
the relationship between repatriation and peace I ended up excluding the
Israeli-Palestinian conflict from the analysis (because, as explained in
chapter 9, I analyze cases that have been coded by the Uppsala Conflict Data
Program as terminated at one time or other, and the Israeli-Palestinian
conflict has not been so coded). However, I remain deeply sympathetic to the
plight of the Palestinian refugees, and their right, in principle, to return.
Writing this dissertation has been an enjoyable experience, which I have had
the fortune to share with a lot of people; I would like to mention a few.
Sincere thanks, first of all, to my supervisors: to Ramses Amer for
encouraging me to pursue my original plan as far as possible during the first
several years of the Ph.D. program, as well as to take on other challenges,
Cindy Kite for helping me focus on the issue at hand during the final year,
and for proof reading the manuscript, and Jan Engberg for always being
sensible and practical about my ambitions with the project. Together you
have provided an effective combination of advice and support.
As work on the study progressed, I have explored different ways of
addressing the issue, efforts that resulted in draft chapters and conference
papers of (hopefully) increasing quality and cohesion. Regardless of whether
I ultimately found it worthwhile to use the various options I tried out, I am
ix
PEACE BY REPATRIATION
grateful for the many comments I have received from colleages in Umeå, in
other parts of Sweden, and abroad, among them Nicholas Aylott, Ronald
Holzhacker, Lisa Hultman, Anna Jarstad, Emilia Korkea-aho, Kersti
Larsdotter, Jannie Lilja, Emelie Lilliefeldt, James Milner, Ann-Sofi
Rönnbäck, Eric Selbin, Isak Svensson, and Malin Eklund Wimelius.
Special thanks to Gino Sahovic for sharing many insights about Bosnia
as well as advice about the research process, and for commenting on several
drafts of the manuscript, Susanne Alldén for continually acting as a sounding
board (and, yes, I’ve enjoyed traveling with you, too), Johanna Johansson for
thoughtful comments on a late version of the manuscript, Johan Hellström
for many inspiring discussions about methodology, and Magnus Blomgren
for the rare ability to make constructive comments about any topic.
I am alone responsible for any remaining flaws and mistakes.
In relation to the field work I want to thank Stina Magnusson and Vida
Korén Holm at Kvinna till kvinna for sharing useful contacts, Nara Galstyan
for being my personal tour leader in Nagorno-Karabakh, and Suzana
Moćević for emergency interpretation in Sarajevo. I am deeply indebted to
all interviewees—especially those of you who shared personal experiences
with me. I am also grateful to those of your assistants who aided with
interpretation.
Thank you as well Nadine Walicki for the cover photo, Daniel Karlsson
for improving my figures, and Christina Boström for assistance with layout
and for keeping track of practical details during the final months.
I gratefully acknowledge the financial assistance provided by the
Wallenberg Foundation and Ambassador Lyndon Olson through the
“Scandinavian Policy and Politics Project,” as well as by the J. C. Kempe
Memorial Fund, the Knut and Alice Wallenberg Foundation, and
Forskraftstiftelsen Theodor Adelswärds Minne.
Three groups of people who had nothing to do with the dissertation also
deserve mentioning, because in different ways they helped me to regularly
put work aside and focus on something else entirely: the choir Kammarkören
Sångkraft, the floorball squad, and my five brothers from back in high
school. These breaks, I dare say, have helped me remain sane thoughout
most of the process.
Finally, I want to thank my wife, Elisabeth, for ceaseless encouragement
and support and for keeping our home up and running when my focus has
been elsewhere, and my son, Henning, for always expecting my full
attention.
Umeå, October 2010
x
PEACE BY REPATRIATION
xi
PEACE BY REPATRIATION
xii
PEACE BY REPATRIATION
Emperors and kings, dukes and marquises, counts, knights, and
townsfolk, and all people who wish to know the various races of
men and the peculiarities of the various regions of the world, take
this book and have it read to you. Here you will find all the great
wonders and curiosities of Greater Armenia and Persia, of the
Tartars and of India, and of many other territories. Our book will
relate them to you plainly in due order, as they were related by
Messer Marco Polo, a wise and noble citizen of Venice, who has
seen them with his own eyes. There is also much here that he has
not seen but has heard from men of credit and veracity. We will
set down things seen as seen, things heard as heard, so that our
book may be an accurate record, free from any sort of fabrication.
And all who read the book or hear it may do so with full
confidence, because it contains nothing but the truth.
Marco Polo, The Travels
xiii
PEACE BY REPATRIATION
xiv
INTRODUCTION
1 |Introduction
The conventional wisdom of most refugee experts holds that there is a necessary
connection between forging and implementing a peace agreement and ensuring the
successful return of refugees. Peace depends on refugee repatriation, and every peace
agreement must provide for it, or so it is widely believed.
Howard Adelman1
1.1 The Peace-by-Repatriation Thesis
The sun was barely visible through the cold and hazy air as Dr. van Heuven
Goedhart made his way to the Nobel Institute in Oslo on 12 December,
1955.2 The Office of the United Nations High Commissioner for Refugees
(UNHCR) had been awarded the Nobel Peace Prize, and High Commissioner
Dr. van Heuven Goedhart was in Oslo to accept the prize and deliver his
Nobel Lecture.3
This was neither the first nor the last time that the link between refugee
protection and peace has been highlighted through the Peace Prize. In 1922,
Norwegian explorer/diplomat Fridtjof Nansen received the prize after,
among other things, as High Commissioner for Refugees of the League of
Nations, having taken on the tasks of finding solutions for some 1.5 million
Russian refugees around Europe, and of handling the displacement situation
resulting from the Greco-Turkish War. Shortly after Nansen’s death in 1930,
the League of Nations established the Nansen International Office for
Refugees to continue his work in the field of refugee protection, and the
organization was awarded the Peace Prize in 1938.
Also, after the UNHCR was awarded the Peace Prize in 1955, in 1958 it
was given to Father Georges Pire for his work with assisting old and
unhealthy refugees to leave refugee camps around Europe ten years after the
end of World War II. In 1981 the UNHCR was awarded the Peace Prize a
second time, for its continued work on the issue of refugee protection,
including its adhesrence to “the fundamental principle that no refugee must
be repatriated against his or her will, or under any form of coercion
whatever.”
In his Nobel lecture in 1955 Dr. van Heuven Goedhart explained that
Refugee problems can only be solved in three different ways—through voluntary
repatriation, through resettlement overseas, and through integration either in the
country of present residence or in combination with intra-European migration. Of
1 Adelman (2002, 273).
2 Oslo weather observations for 12 December, 1955, kindly provided by the Climate Division of the Norwegian
Meteorological Institute.
3 The UNHCR received the 1954 Peace Prize, which was “reserved” in 1954, and awarded in 1955.
1
PEACE BY REPATRIATION
these solutions, voluntary repatriation is no longer of great importance (UNHCR
1955).
In fact, he went on to predict that “voluntary repatriation will in the years to
come account for not more than one percent of the solutions to refugee
problems still to be solved” (UNHCR 1955).
At this time the UNHCR’s mandate was limited to caring for persons
who had become refugees (a concept discussed further below) as a result of
“events occurring in Europe before 1 January 1951.” More specifically, the
1951 Convention Relating to the Status of Refugees (the Refugee Convention)
left it up to each Contracting State to choose whether the geographic
restriction would apply to its particular obligations or not; the Statute of the
UNHCR, on the other hand, contained no geographic reference (Barnett &
Finnemore 2004, 84–86; Orchard 2008, 278–282). Both documents were
limited to displacement occurring before 1951. In 1967 a Protocol Relating to
the Status of Refugees was adopted, which removed the time and geographic
restrictions, thereby also eliminating the inconsistency between the Statute
and the Convention. Earlier, the Statute had made the High Commissioner
competent to deal with an increasing number of refugees who nevertheless
were not covered by the Convention (cf. Orchard 2008, 301).
In 1955 Dr. van Heuven Goedhart’s remarks on repatriation no longer
being of great importance concerned some one million refugees still
displaced as a result of events occurring in Europe before 1951. Sixty years
later, the number of refugees around the world is ten million, with another
more than twenty million being internally displaced, and the international
approach to solutions has shifted 180 degrees to a strong emphasis on
repatriation as the solution in today’s refugee discourse. Chimni (2003, 195),
concludes that in the early twenty-first century local integration and third
country resettlement are applicable to around one percent of the world’s
refugees.
According to Barnett and Finnemore this shift began in the late 1970s
and developed into a “repatriation culture [. . .] a bureaucratic structure and
discourse coupled with formal and informal rules that made repatriation
UNHCR’s preferred solution and, in fact, nearly synonymous with
protection” (Barnett & Finnemore 2004, 93–105). In time, repatriation has
evolved from a solution to the refugee’s problem of being displaced to a
solution to the problem of the fact that people are displaced, and to some
extent even further into being instrumental in resolving other problems than
those related to displacement. For example, in 1990, then High
Commissioner Stoltenberg stated his ambition that the
2
INTRODUCTION
UNHCR should be prepared to seize all the possibilities for voluntary repatriation,
which is the best solution for refugees, the most productive use of resources, and a
concrete contribution to peace and stability (quoted in Barnett & Finnemore 2004,
97–98).
Voluntary repatriation is a concrete contribution to peace and stability. High
Commissioner Stoltenberg is not alone in making that claim, which I will
refer to in this study as the peace-by-repatriation thesis.4 Indeed, this is the
conventional wisdom of most refugee experts, as noted by Adelman in the
epigraph to this chapter. Obviously, this conventional wisdom has
implications for policy making. Since the late 1980s most major peace
agreement have contained specific provisions regarding the return of
displaced persons (UNHCR 1997, 159–160; cf. Harbom, Högbladh &
Wallensteen 2006; UCDP 2009b). Similarly, about half of all post–Cold War
multidimensional peacekeeping operations have had refugee repatriation as
part of their mandates (Howard 2008, 347–352). According to Black (2006,
26), the 1997 issue of the UNHCR’s State of the World’s Refugees “stressed
how return is symbiotically related to peace-building in conflict-torn
countries.”
However, the Adelman quote sensibly ends with the words “or so it is
widely believed” rather than “and this has been confirmed by social
scientists.” In fact, despite the popularity of the peace-by-repatriation thesis,
very little research has been done on the assumption that refugee
repatriation is crucial to achieving and maintaining peace. Consequently, in
its most rudimentary form, the problem addressed in the present study is
whether this assumption holds true.
Obviously, the peace-by-repatriation thesis needs to be both specified
and clarified before it can be tested. It should be noted at the outset that
refining the peace-by-repatriation thesis and making it testable is an
important facet of this enquiry, and one which goes beyond the definition of
key theoretical concepts. Rather, it is continually done throughout most of
the study.
1.2 The Purpose of the Study
In one of the few academic studies expressly addressing the peace-byrepatriation thesis, Adelman (2002) sets out to analyze the relationship
between peace and repatriation. He does this on the basis of his own
previous research as well as several case studies of peace implementation in
the same edited volume, of which his chapter is a part (Stedman, Rothchild
& Cousens 2002).
4 Throughout this study, ”the thesis” will refer to the peace-by-repatriation thesis; the dissertation will be
referred to as ”the study.”
3
PEACE BY REPATRIATION
Briefly, he questions “any causal, necessary, signifying, or even inherent
type of relationship of refugee repatriation to peace agreements” (Adelman
2002, 274). Importantly, his argument is not that refugee repatriation and
peace are totally unrelated phenomena, but rather that no one relationship is
valid across all cases; depending on the circumstances, repatriation may
contribute to peace, it may be irrelevant to peace, or it may be unfavorable to
peace. He concludes, inter alia, that factors such as the type of war and who
wins affect the link between repatriation and peace (Adelman 2002, 295).
What makes Adelman’s study unusual is that it draws on a number of
cases, and initiates a systematic discussion of why the role of repatriation in
the establishment of peace differs between cases. Most other studies that
consider the relationship between repatriation and peace do not have that
relationship as their main focus, and, unsurprisingly, they often reach
different conclusions regarding the role of repatriation in achieving and
maintaining peace. Hence, there is clearly a need for further systematic
analysis of the common assumption that repatriation is a necessary
condition for peace—the peace-by-repatriation thesis—but there is also a
need to develop the theoretical foundations of that assumption.
Therefore, the purpose of this study is to refine and analyze the peaceby-repatriation thesis. An important rationale for this study is that the
assertion that repatriation is necessary for peace is more common than
explicit accounts of the theoretical arguments on which the assumption
rests. Such accounts, in turn, are more common than even rough
descriptions of key concepts (for example, how much repatriation is required
for what kind of peace?).
This study therefore begins by turning to previous research in an
attempt to identify the theoretical foundations of the peace-by-repatriation
thesis. This results in a preliminary theoretical framework comprising a few
concepts relevant to an analysis of peace by repatriation. This framework is
further refined through an initial application of it in two empirical cases of
large-scale conflict-induced displacement. Finally, the peace-by-repatriation
thesis is tested through a fuzzy-set analysis of around forty cases. A more
detailed outline of the study follows at the end of this chapter. Before that,
however, I will discuss a number of key concepts and briefly describe the
context in which repatriation has become the default solution to
displacement.
1.3 Context and Relevance
Overview of Key Concepts
While concepts such as peace and repatriation are discussed further below,
the presentation will benefit from a brief account of a few other important
concepts already at this stage.
4
INTRODUCTION
The term refugee is frequently used to describe anyone who has left his
or her home for any one of a number of reasons and seeks protection in
another country. “Development refugees” are forced from their homes by
major dam projects and the like, “climate refugees” flee from the effects of
climate change, such as rising sea levels, “economic refugees” connotes
people who leave their homes voluntarily to seek a better life in a rich
country, et cetera. People who are displaced within their own country are
occasionally referred to as “internal refugees,” though the term internally
displaced person (IDP) is far more common (see below).
However, the term refugee is also a legal term with a very specific
meaning. A refugee is defined by the Refugee Convention, in Article 1. A (2),
as a person who
owing to well-founded fear of being persecuted for reasons of race, religion,
nationality, membership of a particular social group or political opinion, is outside the
country of his nationality and is unable or, owing to such fear, is unwilling to avail
himself of the protection of that country.
To emphasize that a displaced person is a refugee according to this
definition, terms such as “Convention refugee” or “statutory refugee” are
used. Clearly, IDPs are not statutory refugees. The Convention definition
also limits the legitimate reasons for flight to violations of civil and political
rights (see below). This means that under the Refugee Convention climate
refugees, development refugees, and others are not refugees.
In displacement statistics an overall number is often provided for
“refugees and asylum seekers.” The term refugees refers to those who have
been recognized by the UNHCR or a host state as refugees according to the
Refugee Convention definition; asylum seekers are those who seek
protection from persecution, but who have not yet been granted such
protection, or asylum. There is no right to asylum in international law, only
a right to seek asylum (see e.g. Goodwin-Gill & McAdam 2007, 358–365). It
remains a discretionary act of the host state to grant it. The most important
right awarded a refugee is the right not to be returned to a country where he
or she is likely to face persecution or torture; this is known as the principle of
non-refoulement (see e.g. Chimni 2000, 85; Clark 2004, 587; Goodwin-Gill
& McAdam 2007, 201–284).
Like all basic international refugee instruments, the Refugee Convention
contains a “cessation clause” (Art. 1.C), which specifies when protection
ends. This can be the case for a number of reasons. Most important in the
present context is the so-called “ceased circumstances clause” of Art. 1.C (5),
which specifies that the Convention shall cease to apply to a person if
5
PEACE BY REPATRIATION
he can no longer, because the circumstances in connexion with which he has been
recognised as a refugee have ceased to exist, continue to refuse to avail himself of the
protection of the country of his nationality.
Examples of ceased circumstances are the end of civil conflict, or regime
change from authoritarian to democratic government (see Bonoan 2001).
The peace-by-repatriation thesis most often calls for “refugee”
repatriation, but the theoretical arguments that inspire the thesis make clear
that it does not exclusively concern persons who meet the definition of the
Refugee Convention. Rather, the peace-by-repatriation thesis refers more
generally to conflict-induced displacement, or, in practice, to persons who
fall within the mandate of the UNHCR. These include
not only those who can, on a case-by-case basis, be determined to have a well-founded
fear of persecution on certain grounds (so-called “statutory refugees”); but also other
often large groups of persons who can be determined or presumed to be without the
protection of the government of their State of origin (Goodwin-Gill & McAdam 2007,
49).
Goodwin-Gill and McAdam (2007, 49) go on to argue that in order for
people in the second group, those who are not statutory refugees, to fall
under the protection mandate of the UNHCR the reasons for flight need to
be “traceable to conflicts, human rights violations, breaches of international
humanitarian law, or other serious harm resulting from radical political,
social, or economic changes” in the state of origin. They should also have
crossed an international border. This broader meaning of the term refugee is
the one used in this study.
Displaced persons who do not cross an international border become
internally displaced persons (IDPs). IDPs enjoy significantly weaker
international protection than refugees, but their situation has been given
increased attention during the past decade. IDPs do not constitute a legal
category in the same way that refugees do. The standard definition of IDPs is
as persons or groups of persons
who have been forced or obliged to flee or to leave their homes or places of habitual
residence, in particular, as a result of, or in order to avoid the effects of, armed
conflict, situations of generalized violence, violations of human rights or natural or
human-made disasters, and who have not crossed an internationally recognized state
border (Cohen & Deng 1998a, 18).
Since 1998 a set of Guiding Principles on Internal Displacement identify the
rights and guarantees relevant to the protection of IDPs (reproduced in e.g.
Chimni 2000, 423–433; Cohen & Deng 1998a, 305–316). The Guiding
Principles are not legally binding like the Refugee Convention. However,
they reflect and restate many existing norms of international human rights
law and international humanitarian law that are legally binding, but not
6
INTRODUCTION
specifically concerned with displaced persons (Goodwin-Gill & McAdam
2007, 484–485; Mooney 2003, 161). While the problem of internal
displacement is well recognized today, the rationale for treating IDPs as a
separate category of people in need is a matter of some dispute (see e.g.,
Hathaway 2007, and responses)5; the UNHCR’s role in assisting them is also
a matter of controversy (Goodwin-Gill & McAdam 2007, 485–488).
There are three so-called durable solutions to displacement, and they are
the same as in 1955 when Dr. van Heuven Goedhart gave his lecture in Oslo:
(1) return/repatriation and reintegration in the home society, (2) local
integration in the host country, and (3) resettlement to a third country
(Chimni 2000, 330–389; Goodwin-Gill & McAdam 2007, 489–501). These
solutions are a product of Cold War thinking, which interpreted everything
including displacement as a result of ideological differences. Earlier,
however, during the interwar years, refugee flows were primarily attributed
to ethnic conflict, and the three durable solutions proposed at the time were
(1) redrawing borders, (2) exchanging populations, and (3) securing
international guarantees for minority rights (Adelman 1992, 6).6
For more on the general development of the international protection
regime, see for example Orchard (2008) and UNHCR (2000, Chapter 1).
The Making of a Default Solution
The 1951 Refugee Convention is a child of its time, and highly influenced by
the emerging post–World War II ideological conflict between East and West.
Due to the voting strength of the Western alliance at the United Nations at
the time (Hathaway, quoted by Chimni 2000, 13), the Refugee Convention
sets out to protect “those whose (Western inspired) civil and political rights
are jeopardised, without at the same time protecting persons whose (socialist
inspired) socio-economic rights are at risk” (Hathaway, quoted by Chimni
2000, 14). Accordingly, during most of the Cold War, “from the point of view
of the Western governments which received them, refugees from the Eastern
Bloc were viewed as welcome evidence of the failure of the socialist system”
(Harrell-Bond 1989, 46). Repatriation was seen as an ideal solution, which
could not be implemented in practice. As expressed by Goodwin-Gill, the
Refugee Convention “was drafted at a time when voluntary repatriation was
effectively obsolete” (quoted by Chimni 2004, 61).
5 “Debate” section in the Journal of Refugee Studies, 20:3, 2007.
6 In fact, a population exchange was Nansen’s solution to the refugee problem created by the Greco-Turkish
war. The problem consisted of more than 1 million Greeks fleeing from Turkey to Greece, which was unable to
provide for them. Nansen believed that “to unmix the populations of the Near East will … secure the true
pacification of the Near East” (quoted by Huntford 1997, 529), and this was indeed agreed in the 1923
Convention of Lausanne. The population exchange was completed by the end of 1924. This, according to
Nansen’s biographer (Huntford 1997, 530), “was arguably his greatest achievement as an international
statesman.”
7
PEACE BY REPATRIATION
By the mid 1980s, however, repatriation started to gain acceptance as
the ideal solution also in practice, for example through the adoption of a
Conclusion on Voluntary Repatriation by the UNHCR Executive Committee
(1985). The Conclusion stated that
from the outset of a refugee situation, the High Commissioner should at all times keep
the possibility of voluntary repatriation for all or for part of a group under active
review and the High Commissioner, whenever he deems that the prevailing
circumstances are appropriate, should actively pursue the promotion of this solution.
Scholars pointed out that return is a poorly understood social phenomenon,
and regretted that the desire of all refugees to go home was presented as a
statement of fact rather than as a hypothesis to be tested (Chimni 2004, 58–
59). A few years later High Commissioner Ogata declared the 1990s the
“decade of repatriation.” “Needless to add,” comments Chimni, “it was not
the sudden availability of scholarly studies which emboldened the
organization to make such an announcement” (Chimni 2004, 59).
A milestone in the establishment of repatriation as the default solution
to displacement situations is the concept of temporary protection. In
response to the refugee flows from the former Yugoslavia in the early 1990s,
the UNHCR in 1992 requested host governments to give temporary
protection to displaced persons instead of going through the time-consuming
process of evaluating individual asylum applications (Bell 2000, 238).
However, in addition to not requiring individual refugee determinations,
temporary protection also allowed return “before conditions which would
safely satisfy the cessation clause are met” (Bayefsky & Doyle 1999, 8). The
Refugee Convention does not specify how or to what extent circumstances
need to change to be considered “ceased,” but Chimni argues that such
decisions have over time come to rely more on “objective” circumstances as
determined by the host state, and less on the “subjective” circumstances as
perceived by displaced persons themselves (Chimni 2004, 61). In light of the
strong preference for repatriation in today’s refugee discourse, such shifts
are significant.
One illustration of host state preferences for repatriation begins with an
August 2004 UNHCR position paper (in fact, one of a series of papers with a
similar content presented over several years), which explicitly argued against
the return to Kosovo of rejected asylum seekers, because the situation in
Kosovo was considered unsafe. In paragraph 6 the UNHCR (2004)
maintains its position that members of the Serb, Roma, Ashkaelia and Egyptian
communities should continue to benefit from international protection in countries of
asylum. Return of these minorities should take place on a strictly voluntary basis,
deriving from fully informed individual decisions. Along with Kosovo Albanians
originating from areas where they now are in a minority situation, they should not be
forced or compelled to return to Kosovo.
8
INTRODUCTION
Responding to UNHCR’s recommendation the Danish, Icelandic,
Norwegian, and Swedish governments, in a letter to the UNHCR and the
United Nations Interim Administration Mission in Kosovo (UNMIK),
claimed that
this ban sets aside the principle of individual status determination of the 1951
Convention, and we are concerned that it has the unintended effect of cleansing
Kosovo of ethnic minorities. [. . .]
The UNHCR position paper furthermore bases itself on an interpretation of the term
“homes in Kosovo” in Security Council Resolution 1244 of 1999 which in our view does
not appropriately take into account the generally accepted concept of internal flight
alternative (Foreign Ministry 2005).
As further discussed in later chapters, the need to reverse ethnic cleansing is
one of the main arguments underlying the peace-by-repatriation thesis, and
one made in relation to the former Yugoslavia more than anywhere else. In
effect, then, the Scandinavian governments used the peace-by-repatriation
thesis to argue in favor of returning displaced persons to areas considered by
the UNHCR to be unsafe for returnees. If the assumption that repatriation is
necessary for peace is used in this way by host state governments, then there
is every reason to refine and analyze that thesis and to find out whether it is
indeed more than just an assumption.
Fortress Europe and the Securitization of Migration
In early June 2006, seventeen members of a purported Islamist terrorist cell
were arrested in a series of counter-terrorism raids in the greater Toronto
area in Canada.7 Two weeks later, the International Association for the Study
of Forced Migration (IASFM) gathered in Toronto for its tenth Biennial
Conference. During one of the plenary sessions it was noted that
no self-respecting terrorist would try to enter a country as an asylum seeker,
considering all the interviews, finger printing, and often humiliation, which that
entails—not to mention how time consuming it is. He would arrive with a phony
Swedish passport as a business man. Yet, as soon as the seventeen were arrested, calls
were made for stricter asylum policies.8
This is a reference to an ever more common view of refugees and asylum
seekers as security threats—migration has become “securitized” (Abiri 2000;
Huysmans 2002). In order to deal with these threats states have employed
various devices “to keep asylum seekers from the procedural door”
(Goodwin-Gill & McAdam 2007, 390), in effect, preventing them from
seeking asylum in a country of their choice. Among these so-called “non7 See “2006 Toronto Terrorism Case” on Wikipedia for more details.
8 Donald Galloway; adapted from the present author’s conference notations.
9
PEACE BY REPATRIATION
admission policies” are the concepts of first country of asylum (FCA), safe
third country (STC), and safe country of origin (SCO) which are part of the
2005 EU Council Directive on minimum standards on procedures in
Member States for granting and withdrawing refugee status, aka the Asylum
Procedures Directive (Council of the European Union 2005; see also Costello
2005, 396–403; Goodwin-Gill & McAdam 2007).
In the EU context, the first country of asylum concept means that
Member States do not need to consider asylum applications if an asylum
seeker has already been granted “refugee status or otherwise sufficient
protection” in another country—the first country of asylum—to which the
asylum seeker can then be returned. The safe third country concept means
that the same principle can be applied if the asylum seeker could have found
protection in another country en route to the Member State in question, but
did not do so; he or she may still be returned to that third country. According
to Costello (2005, 40) “the practice of returning asylum-seekers to STCs is a
European invention, with scant foundation in international law.”
If an asylum seeker comes from a country which has been designated a
safe country of origin, Member States can presume that the asylum
application is unfounded because the country of origin is safe for the
applicant. While these rules are not absolute, they are clearly indicative of
the proliferation of non-admission policies in the Global North, as are the
proposals for establishing “transit processing centers,” or reception camps,
on the major transit routes into the EU (see e.g., Crawley 2005; cf. Fekete
2005; HRW 2009). As part of this development
roadblocks to asylum have become the rule rather than the exception in developed
countries beginning in the 1980s. Visa requirements, carrier sanctions, safe country of
origin and safe third country rules, expedited processing and removal, filing deadlines,
detention, and pre-inspection discourage or bar asylum seekers from receiving
protection in developed countries (Martin & Schoenholtz 2006, 422–423).
As a final example, it is symptomatic that the European Commission places
the issue of asylum under the Home Affairs portfolio, along with crime
prevention and law enforcement, rather than under International Cooperation, Humanitarian Aid, and Crisis Response, or under Employment,
Social Affairs, and Inclusion.
This study is not about the securitization of migration or stricter asylum
policies. However, to some extent at least, the peace-by-repatriation thesis is
a part of the same context. This, by itself, does not mean that the peace-byrepatriation thesis is necessarily unfounded, but it does mean that there is an
apparent risk that states will be tempted to refer to the peace-by-repatriation
thesis in their pursuit of other goals—such as those reflected in the Asylum
10
INTRODUCTION
Procedures Directive.9 Seen from this perspective, there is an urgent need to
start treating the peace-by-repatriation thesis as a hypothesis to be tested
rather than as an assumption. Therefore, it is my hope that this study can
make a small contribution to clarifying the relationship between repatriation
and peace, whatever it may be.
1.4 Outline of the Study
Having now introduced the purpose and the context of the study, I continue
to argue in chapter 2 that explicit theoretical arguments for the peace-byrepatriation thesis are rarely made in the literature. Like the assumption
mentioned above, that all refugees wish to return, the peace-by-repatriation
thesis is usually presented as an assumption rather than as a hypothesis to
be tested. In order to identify the theoretical foundations for the peace-byrepatriation thesis, I turn primarily to the peacebuilding literature, but also
to research on forced migration and on the ethnic security dilemma. The
chapter results in a preliminary theoretical framework, or pre-theory, based
on five key concepts: sustainable peace, repatriation, ended displacement
situation, partition, and ethnic structure.
Next, in chapters 3–6, I develop this framework with the help of two case
studies. First, the case studies are described in chapter 3 as being
“instrumental,” meaning that the cases were selected because they are
particularly useful for the purpose of developing the peace-by-repatriation
framework. Case studies on Bosnia-Herzegovina and Nagorno-Karabakh are
presented in chapters 4 and 5. Bosnia is the case where the argument about
the necessity of repatriation has been most consistently made, whereas in the
case of Nagorno-Karabakh repatriation has played a marginal role in
negotiations. On the basis of the case studies chapter 6 summarizes the
peace-by-repatriation framework used in the chapters that follow.
A systematic analysis of the peace-by-repatriation thesis is conducted in
chapters 7–10. Chapter 7 presents the methodological approach of this
analysis, namely set theory and necessary conditions. I stress that the peaceby-repatriation thesis is not a correlational argument about a dependent and
an independent variable, but a set-theoretic argument about an outcome and
a condition. Consequently, it should be tested as a set-theoretic argument. I
introduce the concept of fuzzy sets, and argue that fuzzy-set analysis is
suitable for a test of the peace-by-repatriation thesis.
Next, in chapter 8, the various concepts of the framework developed in
chapters 2–6 are operationalized as fuzzy sets. The operationalization of
9 It is interesting to note that when displacement is discussed in a peacebuilding context, humanitarian
considerations are usually seen as favoring repatriation. However, in an international legal context, such as in
relation to a common European asylum policy, the past twenty years’ increased focus on return as the solution
to displacement is seen as very problematic from a humanitarian point of view (see e.g. Costello 2005; Noll
1999; Thielemann 2004).
11
PEACE BY REPATRIATION
theoretical concepts is always a delicate matter, irrespective of the choice of
method. It is not necessarily more difficult to operationalize a concept as a
fuzzy set than as a continuous variable, but the difficulties inherent in the
process are often made more explicit. The operationalizations used in this
study will be open to debate, as would any alternative operationalizations.
Therefore, in addition to describing how the concepts are operationalized, I
make an effort to explain why I believe these operationalizations make
sense.10
In chapter 9 I select the cases for the fuzzy-set analysis. The cases are
conflict terminations and the question asked in the analysis is whether the
conflicts remained terminated or restarted. From the Uppsala Conflict Data
Program (UCDP) I select forty-three conflict terminations in thirty-seven
different conflicts. Each case (conflict termination) is then coded—given a
membership score in each of the five fuzzy sets.
The fuzzy-set analysis conducted in chapter 10 clearly shows that
repatriation is not necessary for peace. There is some support for the
alternative thesis that displacement situations have to be resolved in order
for sustainable peace to be achieved, but not that repatriation is preferable to
any other durable solution.
The study is summarized in chapter 11, which also draws some
conclusions and proposes important areas of further research.
10 This effort is also related to one of the premises of this study, namely that the observation of social
phenomena is not as straightforward an exercise as implied by the Marco Polo quotation in the beginning of
the study; describing “things seen as seen, things heard as heard,” however sincerely, does not make a
statement “true.”
12
CONCEPTS
2 |Theorizing Peace by Repatriation
Everything was simple, and direct. Cause and effect were good friends back then; thesis
and reality hugged each other as if it were the most natural thing in the world. And my
guess is that the sixties were the last time that’ll ever happen.
Haruki Murakami11
2.1 Introduction
So far few attempts have been made to systematize the assumed relationship
between repatriation and peace. Usually, it is simply one of a number of
activities described variously as important tasks of peacebuilding or as
conditions for peacebuilding success, without further specification either
how it supports peacebuilding or whether it does so only under certain
circumstances. As a rule, the positive effects of repatriation for peacebuilding
are taken for granted, and, at least implicitly, assumed to hold across all
cases of peacebuilding. In order to analyze the claim that repatriation is
important for peace the relationship needs to be more explicitly specified.
Hence, the purpose of this chapter is to elicit and concretize the theoretical
foundations of the peace-by-repatriation thesis. I rely most heavily, although
not exclusively, on the peacebuilding literature.
Two earlier attempts to describe the relationship between repatriation
and peace, or perceptions of it, constitute the starting point for the
development of the theoretical framework presented in this chapter. These
are Adelman’s four positions regarding the nature of the relationship
between repatriation and peace, as extracted from peacebuilding research
(Adelman 2002), and UNHCR’s four explanations12 of how repatriation
contributes to peace (UNHCR 1997). After introducing them I take a brief
look at previous research on displacement and peacebuilding to try to
establish what theoretical concepts should be part of the peace-byrepatriation framework.
I want to stress that the word “framework” is used loosely to refer to a set
of factors, or concepts, that would appear to be relevant to an analysis of the
peace-by-repatriation thesis. The present chapter identifies an initial set of
factors by looking at previous research on repatriation and peace. The
following chapters develop this framework through an initial application of it
to two empirical cases. While the case studies enhance understanding of the
cases, more importantly they contribute to evaluating and refining the peace11 Murakami (2006, 77).
12 For lack of a better term to refer collectively to the four “ways” in which repatriation can contribute to
peace, as described by the UNHCR, I use ”explanations.”
13
PEACE BY REPATRIATION
by-repatriation framework. In other words, the development of the
framework in chapters 2–6 is reminiscent of Rosenau’s (1966) “pre-theory”
of foreign policy, with theory and observation being mutually reinforcing, as
well as of Brecher’s (1999) description of “theory from the bottom up,”
according to which “theory is the end-point of an intellectual process, not the
starting-point.”
From literature on refugee repatriation and peacebuilding, specifically
several case studies in Stedman, Rothchild and Cousens (2002), Adelman
(2002, 274–275) extracts four positions, representing different views of the
relationship between refugee repatriation and peace implementation13. First,
he identifies two soft positions, one (Soft I) that considers any resolution of a
refugee problem as a sign that peace is in place, and one (Soft II) that
specifically sees repatriation as such a sign. Common to the two soft
positions is that they perceive the resolution of the refugee issue as “a
manifestation of peace.”
Adelman’s soft positions come dangerously close to using repatriation as
an indicator of peace, as Ali and Matthews (2004, 405–6) do:
Even in Mozambique, South Africa, and Zimbabwe, where the international
community can fairly claim some responsibility for a successful transition from war to
peace, measured by such criteria as effective economic recovery, free and fair
elections, the repatriation of refugees, the decommissioning of former combatants
and their integration into society, external intervention has also had a negative impact
[emphasis added].
Using repatriation as an indicator of peace means that repatriation cannot be
fruitfully analyzed as an empirically necessary condition for peace; it means
determining a priori that it is a logically necessary condition (or a
“constitutive” condition, cf. chapter 7). Adelman’s soft positions are not
necessarily of this character, but the distinction between “being a sign of”
and “being an indicator of” is not made explicit in his study.
Adelman also presents two hard positions, one (Hard I) that considers
peace and refugee repatriation as necessary conditions of one another, and
one (Hard II) that claims mutual causation between peace and refugee
repatriation.14 Additional versions of the hard positions may be conceived of
in the form of unidirectional conditionality and causation (in either
direction). While there are several examples of the soft positions in the
13 Peace implementation, ”the process of carrying out a specific peace agreement” (Stedman 2002, 2).
14 To argue that refugee repatriation and peace are necessary conditions of one another is methodologically
problematic. As explained in greater detail in chapter 7, a relation of necessity means that the outcome
constitutes a subset of the necessary condition. A relation of mutual necessity would mean that the two
relevant sets (here “cases of peace” and “cases of refugee repatriation”) are subsets of each another, or,
expressed more succinctly, that the two sets completely overlap. However, this means that they are also
supersets of each another, making them not only mutually necessary but mutually sufficient, which I think is
not what Adelman intends to imply with his Hard I position (cf. Johansson 2007, 97). It makes more sense to
see refugee repatriation and peace as mutually reinforcing.
14
CONCEPTS
literature, according to Adelman (2002, 274) “the hard view remains the
overwhelming conviction and presupposition of those who connect refugee
repatriation to the implementation of peace agreements.”
The most common and, in my opinion the most crucial position is the
one claiming unidirectional conditionality—that repatriation is a necessary
condition for peace. This is the one referred to in the epigraph to chapter 1
above. This position is common, and often explicitly articulated by the policy
community (cf. chapter 4).
But, “how exactly can repatriation and reintegration contribute to the
peacebuilding process in war-torn societies?” asks the UNHCR in the 1997
issue of The State of the World’s Refugees (UNHCR 1997, 160). Four
answers (“explanations”) are given.15
The first corresponds to Adelman’s Soft II position—repatriation can be
seen as a sign that peace is in place. Quoting a World Bank discussion paper,
the UNHCR argues that “as long as significant portions of a society’s
population are displaced, the conflict has not ended” (UNHCR 1997, 161).
Refugee movements and other forms of forced displacement are the result of
the state’s inability to protect its own citizens. Consequently, voluntary
repatriation and reintegration is “an important manifestation of the process
whereby national protection is restored and human security reinforced.” In
this way, voluntary return has “an important impact on public confidence in
the peacebuilding process” (UNHCR 1997, 162).
A second and related way in which repatriation contributes to peace is by
validating the post-conflict political order, not least in the form of refugees
returning to participate in elections. “When they choose voluntarily to go
back to their homeland, refugees are, quite literally, voting with their feet
and expressing confidence in the future of their country” (UNHCR 1997,
162). Clearly, voluntariness is the sine qua non of this logic.
The third way in which repatriation is said to contribute to peace is that
“the return and reintegration of an exiled population may be a precondition
for peace in situations where refugees are politically and militarily active.”
Examples include Cambodia, Namibia, Nicaragua, and Rwanda. In each of
these cases, it is argued, “the return of the refugees and their separation from
the military represented an important step in the transition from war to
peace” (UNHCR 1997, 162–163).
Finally, returning populations can make important contributions to the
economic recovery of war-torn states. “Indeed, repatriation may even be a
prerequisite for that objective to be achieved” (UNHCR 1997, 163). People
with education and other resources may have a greater chance of escaping
generalized violence, and their return is then likely to be important for
15 More or less the same arguments have been repeated by such prominent authorities as Walter Kälin
(2008), Representative of the Secretary General on the Human Rights of Internally Displaced Persons, and
Erika Feller (2009), Assistant High Commissioner (Protection) at the UNHCR.
15
PEACE BY REPATRIATION
rebuilding war-torn societies. Refugees can also get various forms of training
during displacement, which may prove valuable to the peacebuilding process
once they return.
Refugees voting with their feet and the economic contribution of
returnees are clearly related to the particular solution of repatriation. The
question of restoring protection, however, is addressed as well by the
solutions of local integration and third-country resettlement, and does not
necessarily require repatriation. And to a great extent, the issue of refugee
warriors is a question of disarmament, demobilization and reintegration
(DDR) rather than a question of repatriation.
This suggests that not only the Soft II position can have hard versions,
which it does in Adelman’s presentation. The Soft I position, too, can be
turned into a hard version, namely that any resolution of a refugee problem
(that is, repatriation, local integration, or third-country resettlement) is a
necessary condition for peace.
2.2 Previous Research
The theoretical foundations of the peace-by-repatriation thesis can be traced
in several fields of research, though thesis and reality do not exactly hug each
other in any of them. In this section I discuss three fields of research—on
peacebuilding, forced migration, and partition. The idea is to keep a clear
focus on the peace-by-repatriation thesis, and I make no claim to giving a
comprehensive presentation of any of these fields of research.
I want to point out that the starting point of my interest in these issues,
and the field of research where I primarily aim to make a contribution, is
peacebuilding. Although I have had to acquaint myself with several other
fields of research (certainly one of many rewarding aspects of working with
this project), their use in this study is limited to what they can contribute to
the understanding of the relationship between repatriation and peace.
Peacebuilding
Until 1989 peacekeeping missions were deployed with the express consent of
the (state) parties to the conflict in question, their mandates were usually
limited to supervising truces or disengagement agreements with minimum
involvement in the internal affairs of states, and they were almost exclusively
carried out by the United Nations. The end of the Cold War opened new
possibilities for the international community to engage in conflict resolution
around the world, and post-1989 peacebuilding missions differ in all these
respects. They are frequently deployed under Chapter VII of the UN Charter,
thus not requiring the consent of the (now often non-state) parties, they
cover a broad range of activities (civil as well as military), often issues at the
heart of national sovereignty, and they are carried out by numerous actors of
different character (see e.g., Paris 2004, 13–39).
16
CONCEPTS
The concept of peacebuilding is most often related sequentially to other
concepts such as preventive diplomacy, peacemaking, and peacekeeping.
Peacebuilding is then seen as a post-agreement phenomenon, simultaneous
with implementation and reconstruction, and referred to as post-conflict
peacebuilding. An Agenda for Peace (UN 1992) takes this approach, and
consistently refers to “post-conflict peace-building.” For example, paragraph
21 reads,
The present report in addition will address the critically related concept of postconflict peace-building—action to identify and support structures which will tend to
strengthen and solidify peace in order to avoid a relapse into conflict. Preventive
diplomacy seeks to resolve disputes before violence breaks out; peacemaking and
peace-keeping are required to halt conflicts and preserve peace once it is attained. If
successful, they strengthen the opportunity for post-conflict peace-building, which can
prevent the recurrence of violence among nations and peoples.
The Report of the Panel on United Nations Peace Operations—the so-called
“Brahimi Report” (UN 2000)—defines peacebuilding as “activities
undertaken on the far side of conflict to reassemble the foundations of peace
and provide the tools for building on those foundations something that is
more than just the absence of war.” Similarly, Paris (2004, 38) defines
peacebuilding as “action undertaken at the end of a civil conflict to
consolidate peace and prevent a recurrence of fighting.” Jarstad and Sisk
(2008, 17, note 1) define peacebuilding as “efforts to implement and
consolidate peace agreements,” which presupposes the existence of such
agreements. The aim of peacebuilding is to prevent a recurrence of fighting,
a conceptualization that presupposes that fighting has occurred and largely
ended.
An alternative approach to peacebuilding is to identify it with reference
to its objectives rather than by its sequencing in a peace process. Cockell
(1998, 207; cf. Darby & Mac Ginty 2003, 195) argues that
While the transition phase from a peace accord to stable peace is perhaps the conflict
situation most commonly associated with a broad range of peacebuilding activities,
and may be the most frequent type of operational environment for such applied
peacebuilding, the unique requirements of this transitional environment should not be
confused with the conceptualisation of peacebuilding itself.
In An Agenda for Peace (under the heading “Post-conflict peace-building”)
UN Secretary-General Boutros-Ghali described a number of activities that
would “consolidate peace and advance a sense of confidence and well-being
among people,” explaining that
these may include disarming the previously warring parties and the restoration of
order, the custody and possible destruction of weapons, repatriating refugees, advisory
and training support for security personnel, monitoring elections, advancing efforts to
17
PEACE BY REPATRIATION
protect human rights, reforming or strengthening governmental institutions and
promoting formal and informal processes of political participation (UN 1992,
paragraph 55).
Obviously, many such activities can be initiated before a formal peace
agreement has been reached. The point of the alternative approach is that it
is not the timing of a certain activity that determines whether it constitutes
peacebuilding or not.
Research on peacebuilding, like the trade itself, covers a broad range of
activities. Examples include (re)building democratic institutions, including
arrangements for power-sharing, as well as decentralization and post-war
elections (Hartzell & Hoddie 2003; Jarstad 2009; Lyons 2002; Paris 2004;
Roeder & Rothchild 2005); economic reconstruction (del Castillo 2008;
Woodward 2002); transitional justice, including questions of human rights,
reconciliation, and the rule of law (Brounéus 2008; Fletcher & Weinstein
2002; Hannum 2006; Kaufman 2006; Kostić 2007; Kritz 2001; Lederach
1997; Meernik 2005; Mertus & Helsing 2006; Putnam 2002; Roht-Arriaza &
Mariezcurrena 2006; Sannerholm 2007); security issues, such as
disarmament, demobilization, and reintegration (DDR) and security sector
reform (SSR) (Knight & Özerdem 2004; Peterson 2010; Schnabel & Ehrhart
2005; Spear 2002); and the return of displaced persons (Adelman 2002;
Chimni 2003; Johansson 2007).
However, it should be noted that not everyone agrees that the concept of
peacebuilding should be so all-encompassing, or that all of these activities
are always supportive of peace. For example, Cousens (2001, 13) argues that
peacebuilding should not be equated to the entire basket of postwar needs, as multiple
and complex as they are. Rather, it should be seen as a strategic focus on conflict
resolution and opening of political space, to which these other needs may or may not
contribute. What are frequently conceived as peacebuilding activities, then—
demobilization, economic reconstruction, refugee repatriation, human rights
monitoring, elections, community reconciliation—are not inherently equivalent to
peacebuilding unless they design themselves to be [emphasis in original].
In effect, the various activities that peacebuilders engage in, “however
commendable,” risk diverting attention from the original focus of
peacebuilding, which remains “consolidating whatever degree of peace has
been achieved in the short term and, in the longer term, increasing the
likelihood that future conflict can be managed without resort to violence”
(Cousens 2001, 4). This is a crucial point. Many of the activities that the
international community deals with in the aftermath of conflict are positive
values in their own right, but whether (and when) they contribute to
consolidating peace and avoiding future conflict depends on the
circumstances, which vary from case to case.
18
CONCEPTS
Consider, for example, democratization. As Jarstad and Sisk (cf. Baker
2008; cf. 2001) point out, international efforts to promote democracy may
undermine efforts to consolidate peace, and vice versa. Jarstad and Sisk and
the various authors in their edited volume demonstrate convincingly that
efforts to promote sustainable peace and efforts to promote democracy do
not necessarily reinforce, but instead often undermine each other. They
argue that peacebuilders face four dilemmas. Briefly, the horizontal dilemma
refers to what parties should be included in negotiations and in the post-war
political order; it is a dilemma between inclusion and exclusion at elite level.
Next, the vertical dilemma concerns the trade-off between efficacy and
legitimacy, in terms of elite processes versus public openness and
involvement. Third, the tension between international and local ownership
of the processes of peacebuilding and democracy are represented by the
systemic dilemma. Finally, the temporal dilemma involves the question of
short-term versus long-term priorities and effects of various activities. For a
more elaborate description of the dilemmas, and what causes them, see
Jarstad (2008).
A related dilemma is presented by Sahovic (2007), in his analysis of the
Office of the High Representative (OHR) in Bosnia-Herzegovina from 1996–
2006. Describing what he labels the socio-cultural viability dilemma he
finds that the OHR during its first ten years was continually torn between, on
the one hand, its own cultural bias in favor of implementing peace and
democracy by peaceful and democratic means and, on the other hand, the
Bosnian socio-cultural reality which required authoritarian decision making
in order to be effective—including, for example, the dismissal of dozens of
elected politicians who did not agree with the vision outlined in Dayton.
Assisting the development of democracy is, to echo Cousens, a commendable
exercise, which might contribute to peacebuilding, but there is no guarantee
that it will, particularly in the short term.
The premise of the present study is that the same thing is true for
refugee repatriation. Refugee repatriation is often one of the many activities
included in the theory and practice of peacebuilding. It is quite likely that in
most instances people displaced by violent conflict will want to return home
once the conflict is over, and the international community clearly has an
important role to play in assisting such return. Often, or occasionally, the
return of displaced persons can contribute to peacebuilding, but in this study
I question the assumption that that this is always and necessarily the case.
A few examples can illustrate how the peace-by-repatriation thesis is
articulated in the peacebuilding literature. They show that the perception of
the relationship between repatriation and peace is not straightforward.
Consider first del Castillo (2008, 40), who argues that “critical activities” in
war-to-peace transitions include
19
PEACE BY REPATRIATION
the delivery of emergency aid to former conflict zones (many of which may not yet be
under government control); the disarmament, demobilization, and reintegration
(DDR) of former combatants; the return of refugees and internally displaced groups;
the reform of the armed forces and the creation of a national civilian police.
In a similar vein she goes on to suggest that “refugee repatriation and the
relocation of displaced populations are critical in peace consolidation. Any
strategy for post-conflict reconstruction needs to include plans and sources
of finance for this process” (del Castillo 2008, 257).
Referring to refugee return as a critical activity of war-to-peace transitions
and in peace consolidation can reasonably be interpreted as examples of the
(unidirectional) Hard position I—repatriation is a necessary condition for
peace.
However, in the case study of economic reconstruction in Afghanistan in
the same volume, del Castillo (2008, 169) notes that “an estimated two
million refugees had returned to the country in 2002—a sign of confidence
and hope in the political transition,” a conclusion clearly more reminiscent
of the Soft II position, which makes no claim of necessity but instead sees
repatriation as a sign that peace is in place.
Another example of the Soft II position is Wallensteen’s (2007, 150),
argument that the “issue of the return of refugees is important as it signifies,
more than many other actions, that the extreme conditions which gave rise
to their flight have been remedied to some extent.” Similarly, according to
Stein (1997, 160), “repatriation of refugees to their homeland is a sign that
safety and control over one’s own life has the possibility of being restored,
but it does not necessarily mean that the bond of trust and loyalty has been
restored.”
An example of the Hard I position is Kumar’s claim that the “return and
resettlement of the refugees and IDPs are necessary for social peace and
economic growth” (Kumar 1997, 15).16 Less deterministic is Cohen and
Deng’s claim that “peace and reconstruction in war-torn societies depend in
part on the effective reintegration of displaced persons” (Cohen & Deng
1998a, 5). The Hard I position can be found within the policy community as
well. For example, UN Secretary-General Annan, in an address to the
UNHCR Executive Committee, argued that
the return of refugees and internally displaced persons is a major part of any postconflict scenario. And it is far more than just a logistical operation. Indeed, it is often a
critical factor in sustaining a peace process and in revitalizing economic activity (UN
2005).
16 Kumar uses the term “resettlement” to refer to the reintegration of returnees into their societies of origin,
not as the alternative durable solution of third-country resettlement (cf. p 17 where he discusses “the
resettlement of returnees”).
20
CONCEPTS
As noted above, one can also conceive of a hard version of Adelman’s Soft I
position, which would claim that solving the displacement situation is a
necessary condition for achieving sustainable peace, but without requiring a
particular solution to the displacement situation. For example, High
Commissioner Ogata (1997, vii) has argued that “peace-building requires just
solutions for refugees and displaced persons. In UNHCR’s experience, such
solutions are indispensable for lasting peace and true stability.” Similarly,
the Esquipulas II Declaration of 1987 proclaims that “there can be no lasting
peace without initiatives to resolve the problem of refugees, returnees and
displaced persons” (Stein 1997, 158).
While these examples can rather easily be related to Adelman’s positions
regarding the character of the relationship between repatriation and peace, it
is more difficult to link them to one or more of UNHCR’s explanations of
how repatriation contributes to peace—or, importantly, to find alternative
explanations. This is true for most of the peacebuilding literature. A rare
exception is Adelman (2002, 276–277), who notes that in some cases
repatriation may not be a goal in itself, but intimately tied to the success of
an election or a referendum (which, however, as argued above, does not
necessarily contribute to peace); examples are Cambodia, Namibia, and
Western Sahara.
In addition, it is possible to find examples of UNHCR’s explanations in
peace agreements, such as the Agreement on a Comprehensive Political
Settlement of the Cambodia Conflict (Paris Agreement 1991), which states
that
The task of rebuilding the Cambodian nation will require the harnessing of all its
human and natural resources. To this end, the return to the place of their choice of
Cambodians from their temporary refuge and elsewhere outside their country of origin
will make a major contribution.
A variant of this involves emphasizing the contribution to rebuilding that can
be made by people with special education and training, such as engineers,
teachers, physicians, et cetera, and to focus particularly on the return of such
individuals.17 The International Organization for Migration (IOM), for
example, carried out such a program in Bosnia-Herzegovina (cf. chapter 4).
Forced Migration
Another body of research is forced migration research, which like
peacebuilding is a field that comprises several sub-fields, including refugee
studies (Haddad 2008; Hathaway 2007; Milner 2009) as well as research on
various other forms of displacement, such as development induced
17 This is related to what Naudé (2009, 258) refers to as “entrepreneurial migration,” the fact that people who
emigrate are often relatively skilled individuals.
21
PEACE BY REPATRIATION
displacement, climate or environmental refugees, and internal displacement
(Biermann & Boas 2010; Cohen & Deng 1998a; 1998b; Global IDP Project &
Norwegian Refugee Council 2002; McNamara 2007; Mooney 2003; Weiss &
Korn 2006). There is also research on UNHCR and asylum, including the
securitization of migration and the development of “fortress Europe” (Abiri
2000; Alexseev 2006; Byrne, Noll & Vedsted-Hansen 2004; Koser, Walsh &
Black 1998; Loescher, Betts & Milner 2008; Noll 2003; Thielemann 2004;
van Selm 2003), protracted refugee situations, warehousing, and trafficking
(Koser 2003; Loescher & Milner 2005; Smith 2004), exile and diasporas,
land and property rights, and the search for durable solutions (Black & Koser
1999; Chimni 2004; Harrell-Bond 1989; Leckie 2007; Swain 2007). Further,
various social science themes are applied to the specific situation of forced
migrants, for example, gender studies, health issues, and sustainable
livelihoods (Hyndman 2000; 2010; Mertus 2003).
By way of simplicity one could perhaps say that if peacebuilding is about
consolidating peace in the short term and avoiding a recurrence of violence
in the long term, forced migration studies are about protection during
displacement in the short term (which is not always so short), and durable
solutions for displaced persons in the long term. In the peacebuilding
literature, the sustainability of return is usually not an issue. The important
thing is that displaced persons return, but the need for sustainable return is
at best implicit. Occasionally (as for example in the case of the Scandinavian
letter to UNMIK and UNHCR), it is explicitly made clear that the peace-byrepatriation thesis does not require return to be sustainable—the internal
flight alternative is good enough (cf. the discussion of minority return versus
relocation in chapter 4).
Conversely, in forced migration research the sustainability of return is
essential, but only rarely is it explicitly related to peace. When it is, the
argument is often that return should not be used as a means to achieve
another end, even if that other end is peace. Sustainable return is an end in
itself and the fate of displaced persons should not be reduced to being an
instrument for other purposes—not primarily because this might undermine
the peace process, but because it might put returnees in danger (see Bhatia
2003; Black 2006; Chimni 2003; Naqvi 2004; Petrin 2002; Phuong 2000;
Zolberg, Suhrke & Aguayo 1989). For example, Bhatia (2003, 797) asks how
the voluntariness of return is affected by a broader peace settlement that
puts repatriation on a list of tasks to be completed according to a timetable.
Focusing specifically on the situation of the Sahrawi refugees, he emphasizes
that repatriation is
related to cessation [of refugee status] by the fact that the end-point of return is an
individual’s renewed acquisition of national protection. Yet, what of repatriation prior
to cessation, as will be the case in the Western Sahara? (Bhatia 2003, 794)
22
CONCEPTS
If repatriation is to coincide with the renewed acquisition of national
protection, then a more comprehensive peace than the mere absence of
armed conflict needs to be in place prior to return. From this perspective it is
difficult, if not impossible, to conceive of repatriation as a condition for
peace. A similar argument is presented by Naqvi (2004, 83):
Given the link between massive human rights violations and refugee status, it would
seem logical that repatriation of refugees should only be promoted and encouraged
after an objective assessment of the human rights situation in the country of origin has
been carried out and where it has been ascertained that repatriates will be guaranteed
these rights if they choose to or are required to return [emphasis in original].
Lischer (2005) focuses on the issue of refugee related violence (the third of
UNHCR’s explanations). She finds four socioeconomic explanations for
refugee-related violence in the literature: (1) that large refugee camps
become breeding grounds for militant and criminal organizations because
they are harder to control, (2) that camps located near the border of the
sending state facilitate attacks either by refugee militias or by the sending
state itself, (3) that large numbers of young men among the refugees will
lead to greater violence, and (4) that poor living conditions encourage
discontent, which leads to militancy (Lischer 2005, 9). Finding these
explanations wanting, Lischer proposes that refugee-related violence can be
better understood by considering the political context of the crisis,
specifically three attributes of this context, namely (1) the origin of the
refugee crisis, (2) the policy of the receiving state, and (3) the influence of
external state and non-state actors (Lischer 2005, 18–43). Most importantly,
the origins of the refugee crisis are often disregarded in the routine
socioeconomic explanations of refugee-related violence.
Lischer categorizes refugees into three groups according to the cause of
their flight: First, there are situational refugees, who flee the general
conditions of civil war; they usually wish to return as soon as hostilities end
and are unlikely to engage in military activity. Second, persecuted refugees
escape ethnic cleansing or oppressive policies targeting them on the basis of
ethnic, religious, linguistic, or political identity; they are likely to organize
and be involved in cross-border violence, and they will not return home
unless the can be protected from further persecution. Third, state-in-exile
refugees include political and military leaders; these groups will only return
home in victory. They are very likely to organize and have a high propensity
for political violence.
In sum, the militarization of refugees—however likely or unlikely it was
in the individual case—can make the need for a solution to the displacement
situation more acute. However, peace is arguably more dependent on
disarming and demobilizing the militarized refugees and finding durable
solutions for them, than on resolving the displacement situation particularly
23
PEACE BY REPATRIATION
through repatriation. The advantage of considering the origin of the refugee
crisis is that specific characteristics of the outcome of the conflict can be used
to assess the likelihood of voluntary repatriation among different groups of
refugees once the conflict has ended and a certain level of national protection
has been restored. Again, peace comes before repatriation.
In this context, I should point out that the role of diasporas in
peacebuilding is not addressed in this study, even though the broad concept
of diaspora can comprise conflict induced displacement. Research on the
role of diasporas in sustaining or mitigating conflict in their countries of
origin focuses on established diaspora communities that display some form
of permanency of relocation, such as the Irish-American diaspora, the Sri
Lankan diaspora in the West, and the Ethiopian diaspora in North America
(see various contributions in Swain 2007). Conversely, the peace-byrepatriation thesis relates to asylum seekers, refugees, IDPs, et cetera—
people displaced by armed conflict and awaiting durable solutions.
Ethnicity and Partition
Widening the scope of potentially relevant literature reveals that, in addition
to the UNHCR’s four explanations of how repatriation contributes to peace, a
fifth explanation is related to ethnicity. The role of ethnicity in an analysis of
repatriation and peace is stressed by proponents as well as opponents of
repatriation.
Ethnicity is an inherently contested concept. Discussions about what
ethnicity “is” were more common in the 1980s and early 1990s than they are
today. Various authors have asked whether ethnicity is primordial or
acquired, static or contextual. They have compared and contrasted ethnicity
to related concepts such as race, nationality, and minority, and debated
alternative bases for membership in an ethnic group ranging from kinship
and shared history to religion and language (Hettne 1993; Horowitz 1985;
Lindholm 1993; Riggs 1991a; 1991b). Analyzing the link between ethnicity
and conflict, Gurr, one of the world’s leading authorities on ethnic conflict,
describes “groups whose ethnic identity is the basis for either differential
treatment or collective political action, or both” as “ethnopolitical groups”
(Gurr 2000, 6). Ethnopolitical groups are divided into several subgroups,
most importantly “national peoples,” who seek separation from or autonomy
within states, and “minority peoples,” who seek rights, access, and control
(Gurr 2000, 16).
In the context of repatriation and peacebuilding the concept of ethnicity
is not well explored or defined. It is basically accepted that people belong to
various ethnic groups and assumed that there is a need to remix them after
ethnic cleansing—or, according to the opponents of repatriation, to keep
them apart. This study does not discuss the concept of ethnicity at any
length, but considers the question of mixing or unmixing.
24
CONCEPTS
A brief overview of the academic debate about ethnic mixing and
unmixing has a natural starting point in the work of Kaufmann, who argues
that “the international community must abandon attempts to restore wartorn multi-ethnic states. Instead it must facilitate and protect population
movements to create true national homelands” (Kaufmann 1996b, 137).
Kaufmann’s argument is premised on the notion of the security
dilemma. The security dilemma was originally devised as a description of a
predicament states face under international anarchy, namely (1) that it is
difficult to determine whether other states have defensive or offensive
intentions, and (2) that offensive action is often (perceived as) superior to
defensive action. A state that feels just a little bit threatened will increase its
defensive capabilities. This is seen as an offensive move by another state,
which in turn increases its defensive capabilities, thus confirming to the first
state that it is indeed under threat. When both states believe that the key to
prevailing in a conflict is to strike first, a worst-case scenario is very likely
(cf. Herz 1950).
The security dilemma as an analytical tool was introduced to the study of
ethnic conflict by Posen (1993). He argued that the security dilemma was
well suited to explain the breakup of, for example, Yugoslavia or the Soviet
Union, where various ethnic, religious, and cultural groups were left to their
own devices as the “sovereign” faded and disappeared. There was “anarchy
within,” as Melander (1999) aptly described it.
Ethnic heterogeneity is central to Kaufmann’s perception of the security
dilemma in intrastate conflict:
The severity of ethnic security dilemmas is greatest when demography is most
intermixed, weakest when community settlements are most separate. The more mixed
the opposing groups, the stronger the offense in relation to the defense; the more
separated they are, the stronger the defense in relation to offense (Kaufmann 1996b,
148).
Because ethnic heterogeneity is seen as reinforcing the security dilemma,
Kaufmann calls for the separation of opposing groups into defensible
enclaves.
Another central feature of his argument is that ethnic fears and hatreds
are hardened by war. This, in turn, makes the security dilemma dynamic
even more compelling the next time a group feels threatened. Even after
ethnic wars, where the security dilemma was not initially an important
contributing cause, it will be the second time around if attempts are made to
rebuild a multi-ethnic state, for the simple reason that there has been a war.
Therefore, “solutions to ethnic wars do not depend on their causes”
(Kaufmann 1996b, 137, cf. 154).
Testing his argument that separation is a necessary (but not sufficient;
Kaufmann 1996b, 150) condition for the resolution of ethnic civil wars,
25
PEACE BY REPATRIATION
Kaufmann uses Minorities at Risk (MAR) data to analyzes the twenty-seven
ethnic civil wars that were resolved between 1944 and 1994 (1996b, 160).18
Twelve cases ended by victory, five by partition (de facto or de jure), and two
were suppressed by third party military occupation; these are in line with
Kaufmann’s argument. The remaining eight cases ended by agreement that
did not partition the country, which at first sight contradicts his argument.
However,
every case in which the state was preserved by agreement involved a regionally
concentrated minority, and in every case the solution reinforced the ethnic role in
politics by allowing the regional minority group to control its own destiny through
regional autonomy for the areas where it forms a majority of the population. There is
not a single case where non-ethnic civil politics were created or restored by
reconstruction of ethnic identities, power-sharing coalitions, or state-building
(Kaufmann 1996b, 161).
This convinces Kaufmann that it is the regionally concentrated minority that
is the decisive factor in preventing a recurrence of war: “The critical variable
is demography, not sovereignty” (Kaufmann 1996b, 161). However, all eight
cases in which the state was preserved by agreement were characterized by
both a regionally concentrated minority and regional autonomy. It is not
clear on what basis Kaufmann concludes that the regionally concentrated
minority is more important than the regional autonomy for preventing a
recurrence of war.
Ten years on Licklider and Bloom collected some of the responses to
Kaufmann’s ideas in an edited volume (Licklider & Bloom 2006). For
example, Laitin (2006)19 finds Kaufmann’s argument wanting in precision in
several respects, and sets out to refine it in order to allow for more elaborate
analysis of its viability. In particular, he focuses on the aspect of territorial
concentration of ethnic groups. Using MAR data like Kaufmann, Laitin
concludes, in contradiction to Kaufmann’s argument, that “the concentration
of groups within a region of a country, other things being equal, makes these
groups better candidates for insurgencies leveled against states” (Laitin
2006, 212).
Laitin also points out that Kaufmann does not specify whether his theory
requires a minimum level of regional autonomy or ethnic separation, or
whether what is required is greater autonomy and more unmixing than
before the war. This is problematic since many situations are characterized
by group concentration and regional autonomy already before the conflict. 20
18 Kaufmann considers these cases to constitute all ended conflicts 1944–1994, a claim, which is open to
question (cf. Kuperman 2004).
19 Previously published in Security Studies, 13(4), 350–365.
20 On a similar note, but without making explicit inferences about ethnic mixing or unmixing, Cornell (2002,
247) argues that rather than bringing peace and stability, the introduction of territorial autonomy “may foster
ethnic mobilization, increased secessionism, and even armed conflict.” He compares nine compactly settled
26
CONCEPTS
The fact that “Kaufmann at times combines partition and unmixing as
having the same properties, but at times separates them analytically”
complicates things even further (Laitin 2006, 217).
In another chapter in the same volume, Sambanis (2006, 74)21 finds that
“nonethnic partitions are more stable and peaceful than ethnic partitions
(although these results are driven by the very few cases of nonethnic
partition).” This is noteworthy since partition theorists tend to focus on
ethnic wars (cf. Kaufmann 1996a). Sambanis concludes (2006, 86) that “if
border redefinition is a viable option [. . .] then ethnic integration rather
than ethnic partition may be a winning strategy.”
Partition theory does not usually explicitly address the question of
refugee repatriation. However, because the question of ethnic mixing and
unmixing plays an important role for proponents as well as critics of
partition as a solution to ethnic civil war, their conclusions have implications
for how displacement ought to be handled. Kaufmann is explicit about how
to deal with displacement:
After an ethnic war, repatriating any substantial number of refugees back to territory
held by the other group risks making control of that territory once again uncertain,
thus re-creating the same security dilemma that helped escalate the conflict in the first
place (Kaufmann 1998, 156).
And when Sambanis argues that ethnic integration rather than ethnic
partition may be a winning strategy, this implies that repatriation is
preferable to local integration and third country resettlement as a durable
solution for displaced persons.
To sum up, the partition literature presents several concepts—ethnic
heterogeneity versus homogeneity, mixing and unmixing of ethnic groups,
and political and/or territorial autonomy and partition—that are all
potentially relevant to an analysis of the peace-by-repatriation thesis.
2.3 An Analytical Framework of Peace by Repatriation
In this chapter, I have traced the theoretical foundations of the peace-byrepatriation thesis in three broad fields of research, namely the fields of
peacebuilding, forced migration, and partition. The final task of the chapter
is to extract the relevant factors from the discussion and bring them together
minorities in the South Caucasian republics, Armenia, Azerbaijan, and Georgia, at the end of the Soviet era.
Four of these minorities enjoyed regional autonomy (the Armenians of Nagorno-Karabakh in Azerbaijan, and
the Abkhaz, the Ajars, and the South Ossetians in Georgia), and five did not (the Azeris of Armenia, the Azeris
and the Armenians of Georgia, and the Lezgins and the Talysh of Azerbaijan). Cornell finds that none of the
five non-autonomous minorities formed any broad-based, well-organized, or credible separatist movements.
Conversely, “all four autonomous minorities displayed high levels of separatism, with all but one case (Ajaria)
ending in armed conflict” (Cornell 2002, 258).
21 Previously published in World Politics, 52(4), 437–483.
27
PEACE BY REPATRIATION
into a preliminary analytical framework, or pre-theory, which will be further
refined in chapters 3–6 with the help of two case studies.
Obviously, the relationship of principal interest is the one between
repatriation and peace. These two concepts are central to the analysis, but
they need to be made more precise. As discussed, the relationship between
these concepts can be perceived in different ways, but the focus of attention
in the present study is repatriation as a necessary condition for peace—the
assumption that there will not be peace unless there is repatriation. A few
sources of this assumption are found in the literature: displaced persons vote
with their feet to express confidence in the post-conflict political order, they
participate in the physical rebuilding and the economic recovery of their
societies, and their return reverses what can be reversed of ethnic cleansing.
Clearly, the peace-by-repatriation thesis refers to the post-conflict period
(a concept which is, admittedly, not as straightforward as it might first
appear). What this means is that proponents of the peace-by-repatriation
thesis do not suggest that displaced persons be returned to areas of ongoing
armed conflict in order to bring about an end to the shooting; the idea is that
once the active armed conflict has come to an end—that is, once there is
negative peace22—the return of displaced persons is required, either in order
to prevent armed conflict from recurring, or in order to add some positive
qualities to that negative peace—this latter distinction is not always clear.23
A few additional assumptions underlying the peace-by-repatriation
thesis require an end to displacement, but not necessarily repatriation. They
correspond to the hard version of Adelman’s Soft I position (see p 16), where
any solution to a displacement situation is perceived as a necessary condition
for peace. For example, the suggestion that “as long as significant portions of
a society’s population are displaced, the conflict has not ended” means that
durable solutions have to be found for the displaced persons. This does not
privilege repatriation over local integration or third-country resettlement in
principle (though, in official rhetoric, it does).
Similarly, the need to address refugee-related violence is often essential,
but as a condition for peace it does not require the specific solution of
repatriation. The key is that refugee-related violence is brought to an end,
but whether displacement ends through repatriation, local integration, or
third-country resettlement is in principle a secondary issue. A new
government or a military victory may be necessary requirements for the
voluntary return of state-in-exile refugees, but this should not be confused
with the necessary requirements for peace in the country of origin. In other
words, an analysis of the peace-by-repatriation thesis should not focus
22 Negative peace is peace defined as the absence of war, cf. chapter 8.
23 Reversing ethnic cleansing—thereby sending the message that it will not be accepted—could have the
additional effect of discouraging potential ethnic cleansers in different locations from initiating conflict in the
first place. However, this dimension of the thesis is not analyzed in the present study.
28
CONCEPTS
exclusively on the relationship between repatriation and peace, but on the
somewhat broader relationship between ended displacement situations and
peace.
Further, ethnicity and partition need to be considered. These and
related concepts—such as heterogeneity and homogeneity, mixing and
unmixing, and various forms of autonomy—are used both by proponents and
opponents of the peace-by-repatriation thesis. Kaufmann is one of the few
who explicate the causal connection between repatriation and peace—or, in
his view, between repatriation and the risk of renewed conflict. In contrast,
the policy community often stresses the need for power-sharing within preconflict borders as part of the overall vision of peace, to which repatriation
also belongs. An example would be the Framework Agreement in Macedonia,
2001,24 negotiated under the intense auspices of the international
community, and which states that “there are no territorial solutions to ethnic
problems.”
As far as these opposing theories go, there is an assumed negative
relationship of sorts between repatriation and partition; the pro-repatriation
camp is anti-partition, and vice versa. However, any combination of the
absence or presence of repatriation and partition is possible.
In conclusion, then, a preliminary analytical framework for studying the
peace-by-repatriation thesis would inquire how peace after armed conflict is
related to (1) the repatriation of displaced persons, (2) the end of the
displacement situation, irrespective of how this is achieved, (3) the ethnic
structure of the post-conflict society, and/or (4) the partition of the country.
In the following chapters, I attempt to refine this theoretically derived
framework with the help of two real-world post-conflict situations—BosniaHerzegovina and Nagorno-Karabakh.
The previous chapter presented a preliminary framework based on previous
research. This and the following chapters develop this framework further
with the help of case studies of Bosnia-Herzegovina and Nagorno-Karabakh.
The purpose of the case studies is not to conduct an empirical test of the
thesis, but to gain better insights into how the relationship between
repatriation and peace is understood in real-world cases.
This chapter begins by defining the purpose of the case studies,
discussing what type of case study is most suitable to fulfill that purpose, and
motivating the selection of cases. Next, the method of structured, focused
comparison is described. The chapter concludes with an overview of how the
field research was conducted.
24 See http://faq.macedonia.org/politics/framework_agreement.pdf.
29
PEACE BY REPATRIATION
Studies of Peace by
3 |Case
Repatriation
Without an adequate understanding of the ways in which we apply concepts, appreciation
of the reasons for our conceptual choices, and recognition of the strengths and weaknesses
that our use of key concepts entail, we run the risk of conducting empirical studies that we
cannot justify or that amount to nothing more than aimless fact-finding.
Milja Kurki25
3.1 The Purpose of the Case Studies
As Gerring (2007, 17) observes, “researchers have many things in mind when
they talk about case study research.” He suggests that many of the terms
used to describe case studies, such as qualitative, ethnographic, holistic,
naturalistic research, process-tracing, and triangulation, are “better
understood as describing certain kinds of case studies, not the topic at large”
(Gerring 2007, 18).
Similar arguments have been made regarding comparative research.
Moses and Knudsen state that “one of the most confusing aspects of the
comparative method is the many names given to it,” such as different
systems/similar systems, comparable case strategies, focused comparison,
case-oriented comparisons, or the method of systematic comparative
illustration (Moses & Knutsen 2007, 96). This is not surprising, since
comparisons are used in all scientific methods, but it is makes the conception
of comparative research ambiguous.
These variations are mentioned in order to illustrate that conducting
case studies or comparative research—or, indeed, comparative case-study
research—requires further precision than these general labels entail. I would
argue that the starting point for achieving that precision should be the
purpose of the study.
In this regard, Stake (1995) distinguishes between intrinsic and
instrumental case studies. In the intrinsic study, the case is given. “We are
interested in it, not because by studying it we learn about other cases or
about some general problem, but because we need to learn about that
particular case” (Stake 1995, 3). Conversely, in instrumental case studies,
“we have a research question, a puzzlement, a need for general
understanding, and feel that we may get insight into the question by
studying a particular case,” and the first criterion for case selection is “to
maximize what we can learn” (Stake 1995, 3–4). It is worth noting that
selecting cases that maximize what we can learn means that case selection
25 Kurki (2008, 8–9).
30
CASES
for case study research does not chiefly aim at representativeness; “case
study research is not sampling research” (Stake 1995, 4).
The case studies in the following chapters are instrumental, and the two
cases are selected because they are believed to be particularly fruitful for a
further refinement of the analytical framework. While chapter 2 traced the
theoretical underpinnings of the peace-by-repatriation thesis in previous
research, the case studies aim at a better understanding of how that thesis is
perceived in relation to particular empirical situations rather than as a
general assumption. To be clear, the purpose of the case studies is not to test
the peace-by-repatriation thesis, or to explain a particular outcome; instead,
it is to further develop and synthesize the peace-by-repatriation framework.
Is the framework a useful tool for analyzing peace by repatriation? How can
it be improved?
Selecting the Cases
Two cases of large-scale, conflict-induced displacement are analyzed in the
following chapters, namely Bosnia-Herzegovina and Nagorno-Karabakh.
There are many similarities between the cases: both conflicts occurred in the
first half of the 1990s, both played out in relation to the dissolution of multiethnic states—Yugoslavia and the Soviet Union, respectively—both pitted
ethnic groups against each other, and both resulted in massive ethnic
cleansing in the form of deaths and displacement.
The most important difference between the cases—in the context of the
present study—concerns the perceived importance of repatriation for the
establishment of sustainable peace. In Bosnia, the international community
has consistently emphasized repatriation as the only acceptable solution to
displacement, explicitly arguing that repatriation is necessary for peace. In
Armenia and Azerbaijan, the accepted view is that repatriation is impossible
in the absence of a peace agreement, and the international community
consequently argues that the best solution for displaced persons is local
integration. In this way, Bosnia and Nagorno-Karabakh represent what
Seawright and Gerring (2008) refer to as “diverse cases.”
A related difference between the cases is that there is a peace agreement
in Bosnia-Herzegovina (the Dayton Peace Agreement, DPA, of 1995), but not
in Nagorno-Karabakh. This is undoubtedly part of the explanation for the
different views on how to resolve displacement in the two cases, but it is not
the whole explanation. And, more to the point, it does not explain the
different views on repatriation as a necessary condition for peace. As
illustrated below, repatriation was the agreed solution to displacement from
and within Bosnia-Herzegovina several years before the signing of the DPA,
and in Armenia and Azerbaijan both official and unofficial peace efforts aim
at resolving the political-territorial conflict with the issue of displacement
playing a rather marginal role in negotiations.
31
PEACE BY REPATRIATION
The focus of the case studies is the perception of the relationship
between repatriation and peace, and the analysis covers the period from the
early 1990s until the time of writing (2010). Developments over time inform
the overall analysis, but the idea is not to explain the most recent positions
with the help of these developments.
3.2 The Structure and Focus of the Case Studies
Yin presents a twofold, technical definition of case studies (2009, 18):
A case study is an empirical inquiry that

investigates a contemporary phenomenon in depth and within its real-life
context, especially when

the boundaries between phenomenon and context are not clearly evident.
The case study inquiry

copes with the technically distinctive situation in which there will be many more
variables of interest than data points, and as one result

relies on multiple sources of evidence, with data needing to converge in a
triangulating fashion, and as another result

benefits from the prior development of theoretical propositions to guide data
collection and analysis.
I believe the case studies of Bosnia-Herzegovina and Nagorno-Karabakh in
the following chapters fit this definition well. Both cases are contemporary;
the conflicts and the displacement situations are comparatively recent, and
the implementation and negotiation of peace, respectively, are still on the
international agenda. Also, refining the peace-by-repatriation framework—
the purpose of the case studies—involves trying to clarify what factors are
best seen as the causal (or, rather, permitting) conditions that make up the
peace-by-repatriation thesis, and what factors constitute the context, or
scope conditions, “the circumstances under which a pattern or relationship
holds” (Ragin 2000, 61). In order to address several questions in the two
cases I use data from previous research and other documents as well as
interviews (see below). The case studies are guided by the analytical
framework outlined in chapter 2.
More specifically, I employ the method of structured, focused
comparison (George & Bennett 2004, 67–72). Outlining the method, George
and Bennett stress the importance of identifying the universe of cases, of
which the cases to be studied are instances—each case must be an instance of
the same phenomenon, otherwise the comparative component of the
analysis will be severely undermined. In the present study, the universe of
cases consists of “armed conflicts that have caused substantial displacement,
32
CASES
and that have ended at one time or other, and may or may not have
restarted.”26
Further, the selection of cases from within the universe of cases should
be made on the basis of the research objective and an appropriate strategy to
achieve that objective. “Cases should not be chosen simply because they are
‘interesting’ or because ample data exist for studying them” (George &
Bennett 2004, 69). In other words, to make the most of structured, focused
comparison, case studies should be instrumental, not intrinsic; as argued,
the present case studies meet that requirement.
George and Bennett (2004, 69) also stipulate that case studies should be
conducted with the purpose of explaining, and to provide policy makers with
leverage to influence outcomes. As observed above, the purpose of the casestudy analysis in the present study is not to test the peace-by-repatriation
thesis, or to explain a particular outcome. They are not, in themselves,
concerned with explanation. However, the case studies are intended to
contribute to a better understanding of the peace-by-repatriation thesis.
Together with added insights from the fuzzy-set analysis in chapters 7–10,
the study as a whole can increase awareness of how repatriation is linked to
sustainable peace. This has at least the potential to enhance policymakers’
appreciation of what “effects” various policies in the area of conflict-induced
displacement are likely to have in different circumstances.
According to George and Bennett, once these preliminaries are
accomplished, the first step in conducting the actual structured, focused
comparison involves devising a set of standardized, general questions, which
are asked of each case covered by the analysis. The objective with this set of
questions is to ensure that comparable data are collected from each case.
This is what makes the method structured. Second, the analysis should not
attempt to provide a comprehensive picture of all interesting aspects of an
event. Instead, it should concentrate on those aspects that are relevant to the
research objective at hand. This is what makes the method focused.27
The problem addressed in the case studies is the same as in chapter 2,
namely how the link between repatriation and peace is perceived. However,
while chapter 2 answered that question with the help of previous research,
the following chapters use two empirical cases. Three broad research
questions asked in the case studies are:
26 The definition of this universe of cases is discussed in further detail in chapter 9 in relation to the selection
of cases for the fuzzy-set analysis. There would have been no methodological problems with selecting cases for
the case-study analysis from a different—broader or narrower—universe of cases than the one used for the
fuzzy-set analysis. However, neither are there any methodological problems with using the same universe of
cases.
27 George and Bennett go on to describe in further detail how to employ the method of structured, focused
comparison in three stages: designing case-study research, carrying out the case studies, and drawing
implications for theory (2004, 73–124). I will not repeat those instructions here, as they are rather closely
related to the purpose of explanation, which, as already noted, is not the purpose of the case studies in the
chapters that follow.
33
PEACE BY REPATRIATION
1.
Is repatriation perceived by various actors as contributing to peace,
and, if so, how?
2. How well does the analytical framework developed in chapter 2
capture the relationship between repatriation and peace as perceived
in the two cases?
3. What other factors are important in the context of peace by
repatriation in the two cases?
Data Collection
The case studies are based on primary and secondary written material, as
well as interviews. In the case of Bosnia-Herzegovina there is an abundance
of written material about displacement and return and their relationship to
sustainable peace. Consequently, the case study of Bosnia-Herzegovina is
heavily influenced by such material, particularly policy documents and
various analyses, among them reports by the International Crisis Group
(ICG).28
In the case of Nagorno-Karabakh there is considerably less written
material, both primary and secondary, regarding displacement and return,
and particularly regarding the relationship between repatriation and peace.
Consequently, there was a comparatively greater need to supplement the
written material with interviews in the case of Nagorno-Karabakh than in
Bosnia-Herzegovina.
The different availability of written material in the two cases mirrors the
uneven attention paid to them by the international community. This
imbalance is perhaps even more obvious in terms of the number of UN
Security Council Resolutions adopted regarding the two conflicts. Since 1991,
the Security Council has adopted more than eighty Chapter VII-resolutions
dealing with the situation in the Former Yugoslavia (including Kosovo).29
Regarding Nagorno-Karabakh, the Council has adopted four resolutions, all
in 1993, and none of which invoked Chapter VII (cf. Berdal 2004; Johansson
2009a).
Field research was carried out in the fall of 2008. I spent four weeks in
Armenia and Azerbaijan, including Nagorno-Karabakh, during August and
September, and two weeks in Bosnia-Herzegovina in October. Interviewees
28 The International Crisis Group (ICG) was established in 1995, largely as a reaction to the inability of the
international community to respond effectively to the war in Bosnia. Fifteen years later the organization has
become a leading and well-respected independent source of analysis of conflict and of international efforts to
resolve conflict, and has produced hundreds of reports on dozens of conflicts worldwide. The return of
displaced persons to and within Bosnia-Herzegovina has been a recurring feature of numerous reports. The
ICG has also produced several reports on Nagorno-Karabakh, though far fewer than on Bosnia-Herzegovina,
and with less focus on displacement.
29 In a 2004 study, Wallensteen and Johansson point out that more than 50 percent of all Chapter VII
resolutions adopted from 1946 to 2002 concerned the post–Cold War conflicts in Iraq and Yugoslavia.
“Whether this is a reasonable proportion of Council attention or not, it is testimony to the international
significance of these two situations for the post–Cold War order” (Wallensteen & Johansson 2004, 19).
34
CASES
were selected from three types of actors. My first priority was to speak to
representatives of the international community (IC), most importantly the
UNHCR, but also other organizations. Second, I wanted to meet
representatives of local civil society, first and foremost non-governmental
organizations (NGOs) dealing explicitly with displaced persons. Third, I
wanted comments from government representatives.
In distinguishing these three types of actors, I agree with Pouligny
(2005, 501) that
there is much more proximity between northern or international NGOs (which are
often the same, as there are very few actual transnational NGOs) and
intergovernmental organizations (IGOs), than between the northern and southern
NGOs.
Consequently, organizations such as the International Committee of the Red
Cross (ICRC) or the Norwegian Refugee Council (NRC) would be considered
part of the international community. The term NGO is reserved for local
NGOs (including local branches of international umbrella organizations).
With a few exceptions, I succeeded in meeting the people I wanted to
talk to. In the case of Bosnia-Herzegovina, the OHR turned down my
requests for an interview, arguing that formal responsibility for the return
process has been handed over to local authorities and that the OHR is no
longer involved, and that the turnover of OHR staff means that there is no
institutional memory regarding the return process. In the case of NagornoKarabakh, I was informed that the representatives of the Minsk Group of the
Organization for Security and Cooperation in Europe (OSCE, cf. chapter 5)
are not authorized to discuss the peace negotiations with researchers. In
addition, with reference to the Minsk Group being the forum for
negotiations, the regular OSCE office in Baku declined to meet me.
A majority of the interviews were conducted in English, but in a few
instances I had to rely on interpreters. With one exception, interpreters were
provided by the interviewees. Most interviews were recorded and
transcribed, but a few interviewees did not want to be recorded, and on a few
occasions my recording turned out to be inaudible. In those cases I have
relied on my interview notes. The duration of the interviews varied from 30
to 90 minutes.
The interviews were semi-structured: They combined “a sequence of
themes to be covered, as well as some suggested questions” with an
“openness to changes of sequence and forms of questions in order to follow
up the specific answers given and the stories told by the subjects” (Kvale
2007, 51). I usually began by briefly introducing the subject I was interested
in and then letting the interviewee develop his or her views on this, if
possible initially without much input on my part. In this way, I hoped to
avoid getting the answers the interviewee thought I wanted. As the interview
35
PEACE BY REPATRIATION
progressed I relied more and more on the set of standardized questions that
I wanted to address, and tried to steer the conversation towards those
questions that had not yet been covered. The interview guide is reproduced
in Appendix 1.
Most interviewees agreed to being identified by both name and
organization. However, there were exceptions, and since I believe it is
enough for my purposes to identify which of the three types of actors an
interviewee belongs to—the international community, a local NGO, or the
government—I decided not to reveal the identity of any interviewee, nor the
particular organization they represent (because in several cases this would
de facto have revealed their identity).
The case studies are presented in chapters 4 (Bosnia-Herzegovina) and 5
(Nagorno-Karabakh). They focus on the link between repatriation and peace,
but also discuss other solutions to displacement situations, as well as the
issues of partition and ethnic relations. Chapter 6 compares and analyzes the
cases, and concludes with a refined framework of peace by repatriation,
which is then applied to a larger number of cases in the fuzzy-set analysis in
chapters 7–10.
36
CASES
and the
4 |Bosnia-Herzegovina
Primacy of Minority Return
One of the mantras of the international engagement in BiH since Dayton has been that
the return of refugees and displaced persons to their pre-war homes is a necessary
precondition for the establishment of a sustainable (and self-sustaining) peace.
International Crisis Group30
Nowhere has the argument about repatriation as a necessary condition for
peace been more explicitly and emphatically pronounced than in the case of
Bosnia-Herzegovina (Bosnia). Further, not only return to Bosnia but also
return within Bosnia has been strongly emphasized by the international
community. The recreation of a multiethnic country has been a central
principle of the peacebuilding process, accompanied by numerous plans and
supported with large resources.
4.1 Introduction
While the underlying causes of World War I were abundant, the triggering
event was the assassination of Archduke Franz Ferdinand in the Bosnian
capital, Sarajevo, in June, 1914. As a result of the war, the Austro-Hungarian
Empire crumbled, leaving Bosnia and some of its other Balkan domains free
to join formerly independent Serbia to form the Kingdom of Serbs, Croats
and Slovenes in 1918—in essence, the first Yugoslav state.31
The interwar years were characterized by pan-Yugoslav ambitions losing
ground to de facto centralization under Serb domination. During World War
II, Croat Ustaša fought with the Nazis against Serb Četniks, who, in turn,
later joined the Germans against the Communist Partisans. “Muslims” writes
Malcolm (2002, 192) “had fought on all sides—Ustaša, German, Četnik,
Partisan—and had been killed by all sides.”
The pan-Yugoslav, but above all Communist, Partisans emerged
victorious, and in 1943 proclaimed the second country called Yugoslavia—
more specifically (after a few name changes) the Socialist Federal Republic of
Yugoslavia (SFRY), consisting of six republics: Bosnia-Herzegovina, Croatia,
Macedonia, Montenegro, Serbia, and Slovenia. This country survived for
almost fifty years, until June 1991, when Slovenia and Croatia declared their
independence. Macedonia followed suit in September, and Bosnia in April
1992. The last country to bear the name Yugoslavia was the Federal Republic
30 ICG (2002, 40). The Bosnian name of Bosnia-Herzegovina is Bosna i Hercegovina, and the abbreviation
BiH is often used in English as well.
31 In 1929, the Kingdom’s name was changed to Yugoslavia. Jug means ”south” and slavija means ”land of the
Slavs”; thus, Yugoslavia means the ”land of the south Slavs.” The name Bosnia means either ”running water”
or ”border”; Herzegovina refers to the ”land of the Herzog” (duke in German).
37
PEACE BY REPATRIATION
Map 4.1 Bosnia-Herzegovina. Source: UN Cartographic Section, Map No. 3729, Revised 6 March
2007. Reprinted with permission.
of Yugoslavia (FRY) established in 1992 by the two remaining republics,
Montenegro and Serbia. In 2003, it abolished the name Yugoslavia and
became the State Union of Serbia and Montenegro, and in 2006 both
Montenegro and Serbia declared independence, thereby definitively ending
the Yugoslav era. Most recently, in 2008, the province of Kosovo, having
been under international administration since 1999, seceded from Serbia,
and soon received widespread recognition.
38
CASES
Conflict
While the case of Kosovo is beyond the scope of the present study, it provides
a convenient bridge to a brief summation of the Bosnian war of 1992–1995,
because Kosovo is often considered the place where the break-up of
Yugoslavia started. It was there that Slobodan Milošević, in April 1987,
exclaimed to a Serb crowd that “no one should dare to beat you” (Silber &
Little 1996, 37). Two years later, as part of a Serbian “pounce for control of
all of Yugoslavia” (Rogel 2003, 172), he managed to revoke the provincial
autonomy that Kosovo (and Vojvodina) had received in 1974. The result was
nationalist protest among Kosovo’s Albanian majority, and all over
Yugoslavia ethnicity began gaining lost ground as a basis for organization
and loyalty.
When Croatia and Slovenia declared their independence in June 1991,
Belgrade responded with war. Slovenia secured its independence in a brief
conflict with relatively few casualties. The Croatian conflict was decidedly
more serious, but was, in turn, overshadowed by the war in Bosnia. The
republic of Bosnia-Herzegovina had the most complex ethnic structure of the
six republics with 44 percent Bosnian Muslims (Bosniaks), 31 percent Serbs,
and 17 percent Croats. That is, no ethnic group constituted a majority of the
population. In each of the other five republics, the principal ethnic group
represented at least 60 percent of the population.32
In February 1992, after the European Community (EC) had extended
recognition to Croatia and Slovenia, a referendum on independence was
arranged in Bosnia. Bosniaks and Croats came out overwhelmingly in favor
of independence, and Bosnia-Herzegovina declared its independence in early
March, and gained international recognition in April. However, the Bosnian
Serbs, who had largely boycotted the referendum, were adamantly opposed
to Bosnian independence, and ready to prevent it by force.
The ensuing war was fought primarily between the Bosniaks and the
Bosnian Croats on the one side, and the Bosnian Serbs, backed by Belgrade,
on the other. But nothing was holy. Already in 1991, Belgrade and Zagreb
conspired to divide Bosnia between them; by 1992, similar talks were held
between Bosnian Croats and Bosnian Serbs. In 1993, the Croat-Bosniak
alliance broke down and conflict intensified in central Bosnia between
government forces and the self-proclaimed Croatian Republic of HerzegBosna. One is reminded of Vladimir Gligorov’s dictum about the ethnic issue
in the Balkans: “Why should we be a minority in your state, when you can be
a minority in our state?” (cited in ICG 2004, 4). Whether or not there is any
truth to such a statement, it should not lend credit to attempts to explain the
Bosnian war with reference to stereotypical ancient hatreds among different
32 There were 62 percent Montenegrins in Montenegro, 66 percent Serbs in Serbia, 67 percent Macedonians
in Macedonia, 78 percent Croats in Croatia, and 91 percent Slovenes in Slovenia.
39
PEACE BY REPATRIATION
ethnic groups. As Kaufman (2001, 166) concludes, “Yugoslavia’s ethnic
conflicts were particularly intractable, but mostly because its leaders were
determined to make them so.” As discussed below, several interviewees
argued that ethnicity is still used by Bosnian leaders for political purposes,
and that this severely undermines the return process, and, therefore, the
peace process.
The international community was unprepared for and repulsed by the
brutality of the war and the massive ethnic cleansing that took place in the
form of mass murder and large-scale displacement. Civilians were directly
targeted by the warring parties, whose aim was to create and control
ethnically homogeneous territories.
The international community did not recognize the nature of the
conflict. At a European Community Conference in Geneva in the summer of
1992, the British Minister of Overseas Development, Baroness Chalker,
emphasized that “the world had a humanitarian, not a political, crisis on its
hands, and it called for a humanitarian, not a political response” (Silber &
Little 1996, 247). In a similar vein, Malcolm (2002, 242) describes how the
Western governments only saw the symptoms of the war, but not its causes.
As a result, the world used “humanitarian assistance as a substitute for
protection” in Bosnia (Frelick 1997, 25).
The war ended with the Dayton Peace Agreement (DPA) in November
1995, one of the most ambitious peace agreements to date. The DPA divided
Bosnia into two entities under the federal government level, the BosniakCroat Federation of Bosnia and Herzegovina (the Federation), which was
itself further divided into ten cantons, and the Serb-dominated Republika
Srpska.33 The DPA also confirmed the right of all displaced persons to return
home.
Displacement
The war in Bosnia resulted in the largest displacement crisis in Europe since
World War II. More than half of Bosnia’s 4.4 million pre-war population was
uprooted by the war. More than nine hundred thousand fled the country,
one-half ending up in other countries of the former Yugoslavia, the other half
mainly in the rest of Europe. An additional 1.3 million became internally
displaced (USCRI 1996, 6–7, 129).
As will be further developed below, it was decided very early that the
displacement situation resulting from the war in Bosnia should be resolved
through repatriation. This focus on repatriation had effects on the protection
regime. At a 1992 emergency conference, the UNHCR called on governments
to provide “temporary protection” to displaced persons from the former
33 I follow conventional practice and refer to the entities as Republika Srpska in Bosnian and to the
Federation in English.
40
CASES
Yugoslavia (UNHCR 1992). The concept existed earlier. “It is in the context
of former Yugoslavia, however, that the notion of temporary protection has
been developed most systematically as a means of resolving refugee
problems” (UNHCR 1995b, 85).
In 1996, in one of the most in-depth accounts of the collapse of the
former Yugoslavia, Silber and Little (1996, 247) argued that “to assume that
the refugees would be able to return after the fighting had ended was to miss
the whole point of the war, which was being waged deliberately to ensure
that they would never return.” Pape explicitly brings Kaufmann’s thesis to
bear on the case of Bosnia when he argues that
partition is in Bosnia’s future and no Western policy can avoid it. Rather than allow
ethnic boundaries to be written in blood after SFOR [the NATO-led Stabilization
Force] leaves, the West should help to manage a peaceful partition while it still has
troops on the ground (Pape 1997, 27).
However, the international community began working for an end to
temporary protection very soon after the end of the war (USCRI 1996, 134).
Under pressure from host governments, the UNHCR drafted a very
ambitious plan for return, initially aiming for most return to happen during
1996 (USCRI 1996, 134). In the end, it took several of years before return
became substantial, and another few years before minority return (see
below) really took off.
While hundreds of thousands of Bosnians have returned from
displacement since 1995, it is not yet possible to fully account for the
demographic changes that resulted from the war, or to what extent they have
been reversed. The most recent census was conducted in 1991, prior to the
war. It is estimated that Bosnia’s present population is around 4.5 million,
with Bosniaks constituting 48 percent, Serbs 37 percent, and Croats 14
percent (CIA 2010). However, it is difficult to find reliable information on
the ethnic composition of the two entities—the Federation and Republika
Srpska.34
A new census, the first since 1991, is being prepared for 2011, but is
politically sensitive. According to the Bosnian constitution (Art. IX. 3),
Officials appointed to positions in the institutions of Bosnia and Herzegovina shall be
generally representative of the peoples of Bosnia and Herzegovina.
34 However see Marko (2000). Using figures from the International Monitoring Group (IMG) on the basis of
the 1991 census and UNHCR estimates for 1997, he compares the post-war ethnic structures of the entities
with the demographics of the corresponding territories before the war. He finds that the share of Bosniaks
and Croats in the territory of the Federation increased from 74 percent in 1991 to 95 percent by 1997, and that
the share of Serbs decreased from 18 percent to 2 percent. In the territory of Republika Srpska, the share of
Serbs increased from 54 percent in 1991 to 97 percent by 1997, while the number of Bosniaks and Croats
decreased from 38 percent to 3 percent. Substantial return has taken place since 1997, but this has taken the
form of both majority and minority return (cf. below) and it is not known to what extent these returns have
affected the ethnic structure of the two entities.
41
PEACE BY REPATRIATION
This means that the (non-elected) staff of public institutions should broadly
represent the ethnic structure of the state, entity, canton, or municipality
they serve. Until now, such representation has been based on the 1991
census, first, because those have been the only available figures, but second,
and equally important, in order to avoid legitimizing the demographic effects
of ethnic cleansing. A new census would produce reliable up-to-date figures,
effectively negating the first motive for relying on the 1991 census. Despite
significant return, the new figures would largely reflect the effects of ethnic
cleansing, and this has already led to disagreement about which census to
use henceforth as the basis for representation in public institutions. To
circumvent this problem, it has been suggested that the 2011 census
disregard ethnicity and religion, as those should not be relevant for planning
purposes anyway. At the time of writing much remains unclear about the
2011 census.
4.2 Implementing Peace in Bosnia-Herzegovina
Several peace proposals were presented during the conflict. Already in
September 1991, as conflict broke out after Croatia’s and Slovenia’s
declarations of independence from Yugoslavia, the international community
initiated peace talks (Keesing’s September 1991, 38420). These talks, though
chaired by then NATO Secretary-General Lord Carrington, became known as
the European Community Conference on Yugoslavia.
The first peace plan for Bosnia was presented in March 1992, the
Carrington-Cutileiro Plan (or the Lisbon Agreement), but it did not prevent
the conflict from escalating. In August 1992, several months after the war
had spread to Bosnia, the International Conference on the Former
Yugoslavia (ICFY) was established at a meeting in London, building on the
work done by the European Community Conference. Under the framework
of the ICFY several additional peace plans were presented over the following
years: the Vance-Owen Plan of January 1993, the Stoltenberg-Owen Plan (or
the Invincible Package) of August 1993, the European Union Action Plan of
December 1993 and the Contact Group Plan developed during 1994 (see
Ramcharan 1997, 24–28, 249–341).
Finally, in November 1995, the so called proximity peace talks held at
Wright-Patterson air force base in Dayton, Ohio (famously described by
Holbrooke 1998), resulted in the General Framework Agreement for Peace
(GFAP), more commonly known as the Dayton Peace Agreement (DPA), or
simply the Dayton Agreement. The DPA was signed in Paris on December 14,
and since then the international community has devoted much effort to
support its implementation.
After agreement was reached in Dayton, but before it was signed in
Paris, a conference aimed at mobilizing support for the peace process was
arranged in London on December 8–9, 1995. One important result of the
42
CASES
conference was the establishment of the Peace Implementation Council
(PIC), consisting of 55 countries and agencies that support the peacebuilding
process in Bosnia in various ways. The PIC defines the goals of the peace
implementation in Bosnia and still meets on a regular basis. The High
Representative for Bosnia and Herzegovina is, strictly speaking, a
representative of the PIC.
The Right to Return
The division of Bosnia into two entities is one defining feature of the DPA.
The entities are a manifestation of the de facto division of the country which
resulted from the war. Despite this, another defining feature of the
agreement reflects the international community’s unwavering determination
not to accept the consequences of ethnic cleansing. This is the right of all
displaced persons to return to their pre-war homes. Annex 7 deals with the
issue of return, and its first Article reads:
All refugees and displaced persons have the right freely to return to their homes of
origin. They shall have the right to have restored to them property of which they were
deprived in the course of hostilities since 1991 and to be compensated for any property
that cannot be restored to them. The early return of refugees and displaced persons is
an important objective of the settlement of the conflict in Bosnia and Herzegovina.
The Parties confirm that they will accept the return of such persons who have left their
territory, including those who have been accorded temporary protection by third
countries.
The return of displaced persons is arguably the most important element of
the DPA. As expressed by the ICG (2000, 1):
The key to DPA implementation is refugee return, which is guaranteed in Annex 7 of
DPA and the Bosnian Constitution. All other DPA annexes either depend on refugee
return, or were created to assist in implementing refugee return.
The strong emphasis on return led Paddy Ashdown, High Representative
2002–2005, to proclaim that “we’ve invented a new human right here, the
right to return after a war” (quoted in ICG 2002, 39). The perception of
return as being a right of displaced persons is highly significant, for two
reasons. First, in the particular context of the return process in Bosnia,
several political agreements initially failed to produce any substantial return.
Prettitore (2006, 201) argues that
only when the international community encouraged a system for return and
repossession grounded in the rule of law, and not subject to political agreements, did
serious progress ensue. The DPA does provide individual rights, and the rule of law
approach is the best to ensure they are enforced.
43
PEACE BY REPATRIATION
Insofar as the return process in Bosnia is a success, then, this can largely be
attributed to the approach of treating return as a non-negotiable right, one
not subject to political agreements, but strictly a question of the rule of law.
In this way, it was possible to defuse demands by local politicians for
“reciprocal return,” that is, the idea of people more or less switching places
and moving into “each other’s” apartments. Theoretically, reciprocal return
is perceived as a way to break the logjam and set in motion a chain reaction
of return. In practice it meant that the entire return process became hostage
to the weakest link in the chain and that nothing happened. In the words of
the ICG (1999, 19), “the principle of reciprocity in negotiating returns ought
to be unacceptable, since it implies that individual human rights are
tradeable.” The rule-of-law approach was also important in the Property Law
Implementation Plan (PLIP) discussed below. The basic idea is that the right
to return should apply equally to every displaced person, irrespective of
ethnicity, place of origin, or place of displacement.
The second reason why treating return as a right is highly significant
becomes obvious when this principle is considered in a broader context than
Bosnia. The clearest example of the opposite approach is the Palestinian
refugee problem. Within the framework of the Oslo peace process, the
question of refugees was one of the so-called “final status issues” that was to
be negotiated during the five-year interim period (which formally ended in
May 1999). During the past fifteen years, while return to Bosnia has been a
non-negotiable right and a prerequisite for peace, the international
community has largely limited its efforts to address the Palestinian refugee
problem to calling for a just and comprehensive solution. For all practical
purposes that solution is decidedly and exclusively subject to political
agreement between the parties to the conflict.
Minority Return
During the first several years after the end of the war returning refugees
tended to settle in areas where their own ethnic group was in the numerical
majority, irrespective of whether they actually came from that area or not.
This was not considered enough to secure sustainable peace in Bosnia.
Instead, a central feature of the return process has been the need for
“minority return,” that is, return to areas now demographically dominated
by another ethnic group than one’s own.35
Minority return was a clear objective very early in the peace process. In
November 1995, even before the official signing of the DPA, a pilot project
35 According to the Bosnian Constitution (=Annex 4 of the DPA), Bosniaks, Croats and Serbs are the
“constituent peoples” of Bosnia-Herzegovina. In Bosnian law, the term “minorities” refers to people who are
not constituent or state-forming peoples, and is unrelated to the demographic situation in any given area.
Nevertheless, in the context of return to and within Bosnia, “minority return” is the accepted term used to
signify return to an area where one’s own ethnic group is now in the numerical minority (cf. ICG 2002, 4).
44
CASES
was planned for central Bosnia involving Croat return to Bugojno and
Travnik and Bosniak return to Jajce and Stolac (USCRI 1996, 130). By the
end of 1997, these pilot returns were completed in Bugojno, Travnik and
Jajce, but not Stolac, with just over sixty returning Bosniak families (ICG
1998c, 8; USCRI 1998, 166). Also, in the beginning of the return process,
people who ran for office in their pre-war municipality were prioritized in
the property repossession process in order to encourage minority
populations to engage in politics (Interview 28 IC BiH)36.
Two years after Dayton around two hundred thousand refugees and two
hundred twenty-three thousand IDPs had returned. Most of these were
“easy” cases who returned home to majority areas. Most of the others
originated from areas where they would now be in the minority if they
returned (USCRI 1998, 163). During 1996 and 1997 some forty-five thousand
minority returns took place, but they were offset by some eighty-thousand
newly displaced minorities as displacement continued after the end of open
warfare. This mainly involved ethnic Serbs leaving Sarajevo suburbs as these
were transferred to Federation control in 1996 (USCRI 1998, 164; cf. USCRI
1997, 174).
The year 1998 was declared the year of minority returns (ICG 1998b, i),
but it took until 2000 before minority returns became substantial. More
than sixty-five thousand minority returns (refugees and IDPs together) took
place during 2000, a 64 percent increase from 1999 (USCRI 2001, 207). In
2001, there were more than ninety thousand minority returns, seventeen
thousand refugees and seventy-five thousand IDPs. This was more than 90
percent of all returns—refugee as well as IDP returns—during 2001 (USCRI
2002, 197–198).
The need to achieve minority return is discussed in some detail in the
ICG report Minority Return or Mass Relocation (ICG 1998b). As indicated
by the title, the report emphasizes the difference between minority return,
that is, return to pre-war homes that happen to be situated in an area where
one’s own ethnic group is in the minority and relocation to another area
within Bosnia. Relocation refers to refugees returning to parts of the country
other than their places of origin as well as IDPs finding solutions through
local integration in majority areas.
A common problem in the first few years after the end of the war was the
return of refugees from abroad to other parts of Bosnia than where they
originally came from—that is, relocation. A 1997 report by the Commission
for Real Property Claims of Displaced Persons and Refugees (CRPC) and the
UNHCR argued that such refugee return “contributed significantly to the
36 I refer to interviews by their number in the list on pp 208–209, specifying (1) whether the interviewee is a
representative of the international community (IC), of a local NGO, or of the government, and (2) in which
country or area the interview was conducted (the latter being relevant for the case study on NagornoKarabakh).
45
PEACE BY REPATRIATION
displaced persons problem within Bosnia” in 1997 (quoted in ICG 1998b, 8).
The UNHCR accepted relocation as a durable solution, but underlined that it
had to be voluntary (which under the prevailing circumstances at the time
would have been difficult to ascertain). Bell (2000, 256; cf. Prettitore 2006,
198) argues that the “early enforced return by host states of refugees to parts
of the country which are safe, even though original home areas are not, in
effect consolidates the ethnic carve-up of BiH rather than reversing it.” The
ICG notes that the main proponents of relocation were Croat and Serb hardliners, who preferred the ethnic divisions created by the war (ICG 1998b, 9).
Overall, the strong emphasis on ethnic remixing through minority return
shows how skeptical the international community was at the time to
relocation, seeing it as increasing the risk of cementing the consequences of
ethnic cleansing.
Parenthetically, in light of these sentiments, which were clearly
expressed by the international community in the late 1990s and later, the
Scandinavian position regarding return to Kosovo (mentioned in chapter 1)
is all the more surprising. On the one hand, the Scandinavian governments
argue that they want to send minority populations back to Kosovo in order to
reverse ethnic cleansing. On the other hand, they argue that if return to
actual pre-war homes is not possible, the “internal flight alternative”—that
is, relocation—should be tried. In terms of the need to reverse ethnic
cleansing, therefore, the two arguments are plainly contradictory (and
possibly motivated by other concerns).
A similar stand was taken by other countries regarding Bosnia:
Germany actively promoted repatriation for minorities originating in Republika
Srpska who would not be able to return to their places of origin. An estimated 90,000
of the 1998 repatriates from Germany were minorities originating in Republika Srpska
who relocated to areas of the Federation (USCRI 1999, 178–179).
Even today (or in 2008, when I conducted the field research), a Bosnian
Government Ministry pointed out that simply returning is not enough, we
have to make sure that return does not turn into local integration, that is,
into de facto relocation (Interview 25 Gov’t BiH).
During the late 1990s the international community employed a number
of different strategies to promote minority return in Bosnia (ICG 1998b, 15–
34). For a few years, the Open Cities Initiative was the backbone of UNHCR
policy on minority returns (ICG 1998b, ii). It was launched in March 1997
(USCRI 1998, 167), and it was premised on linking increased international
assistance to local communities showing a “genuine commitment” to
minority return. Seven open cities were recognized during 1997: Bihać,
Busovača, Goražde, Kakanj, Konjic, Vogošća, and Mrkonić Grad (USCRI
1998, 167). By the end of 1999 there were fifteen open cities (USCRI 2000,
221). The initiative was not very successful, mostly because there was not
46
CASES
enough monitoring of the Open Cities (Prettitore 2006, 186). The initiative
was also undermined by poor coordination among the international
community. Many donors had their own funding priorities, which meant
that municipalities that were declared Open Cities did not necessarily receive
the extra funding promised. Additionally, “some European countries
preferred to allocate funding to areas that were the pre-war homes of
refugees they were currently hosting” (Prettitore 2006, 186). The top priority
of several host governments was obviously to assist or encourage people to
leave the host country, but what happened to them or to the country of origin
once they returned was often of secondary importance as best.
During spring of 1998, following the 1997 PIC meeting in Bonn, a
number of cantonal and municipal return plans were drafted and new
mechanisms put in place to promote minority return. The Central Bosnia
Cantonal return plan of November 1997 served as a model in this process. It
outlined three phases of return: first to unoccupied homes, second to
occupied private homes, and third to occupied socially-owned housing (ICG
1998b, 21–24). However, donor response was slow, and some municipalities
soon used the plans to oppose returns outside their particular frameworks.
In order to support a grass-roots movement for return, in October 1996
the OHR initiated the Coalition of Return, which soon developed into an
umbrella organization for more than one hundred refugee and IDP
associations. A central figure in the Coalition was Mile Marčeta, a Bosnian
Serb displaced from Drvar and head of the association of displaced persons
from Western Bosanska Krajina. Among other things, as described by the
ICG (1998a), he was instrumental in the return of displaced Bosnian Serbs to
Drvar, Bosansko Grahovo and Glamoč. When I conducted my field research
in Bosnia ten years later, international community representatives still
referred to Mr. Marčeta as an example of the important role displaced
persons themselves can play in the return process (Interview 30 IC BiH).
However, while their role may be important in specific instances, such
instances were too few and too limited to have a significant impact on the
overall process.
In 1998, hopes were high that the Sarajevo Declaration and the Banja
Luka Regional Return Conference would be important starting points for
minority return. Sarajevo, being the capital of Bosnia and of the Federation,
and Banja Luka, being the capital of Republika Srpska, should serve as
models for the rest of the country, making early minority return particularly
important there. In the end not much came out of either project.
Commenting on these initiatives, Prettitore argues that
it became apparent that political agreements on their own would not lead to
considerable returns. Such agreements made the ability to return subject to political
whims and objectives. Not only was this an inefficient way to support returns, but it
47
PEACE BY REPATRIATION
was also contrary to the guarantee of individual rights of refugees and displaced
persons enshrined in the DPA and the B&H Constitution (Prettitore 2006, 187).
Finally, internationally regulated and monitored return was most relevant
to the case of Brčko, a town and surrounding district in northeastern Bosnia.
The dispute over Brčko was not settled by the DPA. Instead, the parties
agreed to refer the issue to binding arbitration (DPA, Annex 2, Art. 5). In the
final award, delivered in March 1999, the Arbitral Tribunal decided on the
establishment of a self-governing “neutral” district to be part of both entities,
the Federation and Republika Srpska, but under the sovereignty of the state,
meaning that neither entity would exercise any authority in the new district
(for more on the Brčko arbitration, see ICG 2003; Schreuer 1998a; 1998b;
1999). The imperative of the return of displaced persons—described in the
Final Award as “the most important of all of Dayton’s objectives”—is a
central and recurring explanation for this decision (see Arbitral Tribunal for
Dispute Over Inter-Entity Boundary in Brčko Area 1999).
The success of these strategies varied (as did their aims), but none of
them had a major effect on the overall return to and within Bosnia. However,
a strategy that did have an impact was one that addressed the related
problem of property.
Property Repossession
By the end of the war there were considerably fewer empty houses than
displaced persons (Interview 25 Gov’t BiH). In part this was because many
houses had been destroyed, but also because attempts to create ethnically
homogeneous areas had been largely successful. A lot of the internal
displacement that took place was in the form of exchange migration;
Bosniaks displaced from a Serb to a Bosniak area took over the houses of the
Serbs displaced from the Bosniak area, et cetera. However, as noted above,
the idea of reversing exchange migration according to the principle of
reciprocal return was not acceptable to the international community—and it
was also not very efficient. But something had to be done, and the approach
chosen was to treat property repossession, like the issue of return, as a
matter of the rule of law to be implemented as an individual right. In
addition, the physical reconstruction of houses required a massive effort.
According to the International Monoitoring Group (IMG), during the period
1996–1999 housing represented the largest single category of international
expenditure, several times larger than, for example, education, landmine
clearance, peace implementation activities or government institution
building (Black 2001, 183–184).
For several years after Dayton the various Bosnian property laws—one
on the state level and two on the entity level—were major impediments to
minority return. Laws on the use of abandoned property were employed by
48
CASES
the wartime regimes trying to make the separation of the populations
irreversible. However, sustained efforts by the international community to
make return and property restitution part of entity law changed that, thus
making return possible (OHR 2000, 2; cf. Leckie 2007). The Federation
adopted new property legislation in April 1998, and Republika Srpska
followed suit in December. Over the next several years the High
Representative introduced additional amendments to these laws (Prettitore
2006, 189–190).
A milestone of these efforts was the Property Law Implementation Plan
(PLIP, referred to here as OHR 2000), launched in October 2000 by the lead
organizations in property and return issues, OHR, UNHCR, OSCE,
UNMIBH, and CRPC. The working approach of the PLIP was
standardization—that the same pressures, demands and expectations should
apply to all officials and municipalities in Bosnia (OHR 2000). Until 2000,
the focus of return efforts had been on selected return locations or certain
modalities of return. This had been necessary at the time, to initiate the
return process. Conversely, the PLIP was premised on neutral application of
the law. This had two concrete benefits: de-politicization of the property
issue, and institutionalization of the property return process.
The objective of the PLIP is to ensure that all outstanding claims by refugees and
displaced persons to repossess their properties are resolved. It aims to do this by
building domestic legal processes which apply the laws neutrally, processing property
claims as efficiently as possible until all claimants are able to exercise their rights
under Annex 7. By treating repossession of property as a question of rule of law, the
PLIP promotes respect for civil rights over political interests and opens enormous
possibilities for the overall return of DPs and refugees (OHR 2000).
The document confidently states that “the full and complete resolution of
property claims will prove to be the cornerstone of a sustainable and lasting
peace in Bosnia and Herzegovina.” (OHR 2000)
The introduction of the PLIP was followed by a marked increase in
minority returns already during 2000, with 67,445 minority returns
(refugees and IDPs) that year—a 64 percent increase from 1999 (USCRI
2001, 207). During the years 1996–1999, there was an average of 31,000
minority returns; in 2000–2003 the average increased to 26,000 (Ambroso
2006, 3).
In 2003 the Bosnian (state-level) Ministry for Human Rights and
Refugees (MHRR) took over responsibility for implementation of Annex 7.
By the end of 2006 the OSCE, along with UNHCR and OHR, declared that
property law implementation had been accomplished throughout Bosnia
(OSCE Undated). The PLIP had been completed in 129 municipalities across
the country, and nearly 95 percent of over two hundred thousand claims for
property repossession had been successfully resolved. Some interviewees
49
PEACE BY REPATRIATION
echoed the common criticism that many displaced persons reclaimed their
property only to rent or sell it instead of actually returning, and that property
repossession statistics and return statistics are two quite different things
(Interviews 27 NGO BiH, 30 IC BiH). If so, successful property law
implementation would not necessarily reflect a successful return process.
However, a 2006 survey carried out by the Nadel Institute in Sarajevo
showed that “only 5% of returnees have actually sold their houses and 10%
are considering doing so in the future” (see Ambroso 2006, 4). This
illustrates the overall success of the PLIP process also as a means of
furthering the return of displaced persons.
International organizations agree that while the overall return process is
not completed, what clearly have been successful are the property
repossessions (Interview 30 IC BiH). This promotes peace because it
addresses grievances—people get back what they lost (Interview 30 IC BiH).
Similarly, Prettitore (2006, 199) argues that
regardless of whether refugees or displaced persons actually did return in the end, an
extremely important factor in the post-conflict reconciliation in B&H is that they had
the right to choose whether or not to return and repossess property.
Even if many reclaimed their property only in order to sell it, rather than
actually returning, they usually used that money to settle somewhere else.
Thus irrespective of whether they actually returned or not, those who got
their property back are generally no longer displaced, no longer IDPs
(Interview 30 IC BiH). It is also worth noting that rural properties were more
often destroyed, while urban properties were more often occupied. This
means that people were often able to repossess their urban properties (and
possibly sell them at a good price), but that rural ones no longer existed and
could not be reclaimed (Interview 30 IC BiH). On the one hand, it is positive
that for those who reclaimed their property and sold, this process probably
provided them with a durable solution. On the other hand, in these cases the
solution has not contributed to minority return, which was one of the main
motivations for the PLIP.
Despite significant successes, there are more than one hundred fifteen
thousand IDPs still in Bosnia (UNHCR 2010, Europe section, 22). As late as
November 2009, the United Nations Security Council, in resolution 1895
(2009) on the situation in Bosnia, reiterated that “a comprehensive and
coordinated return of refugees and displaced persons throughout the region
continues to be crucial to lasting peace.”
A Revised Strategy for Annex 7 Implementation
Substantial return is completed in Bosnia (Interview 30 IC BiH). Security
incidents occur, but lack of security is no longer the main reason why people
50
CASES
do not return. Instead, the decision to return or not depends on access to
employment, education, health services, and other such factors (Interviews
28 IC BiH, 30 IC BiH). For example, there is better health care in Sarajevo
than in Zenica or Foča. Therefore, many displaced persons in Sarajevo do
not want to give up their status as “displaced person” and the social benefits
linked to it. Thus they do not deregister, even if they actually return
(Interview 28 IC BiH). Also, many are reluctant to return if it means their
children will have to go to school in a “minority area.” Some do return, but
leave their children behind in a majority area or send their children to school
in a majority area even if this means they must go by bus a long way every
day (Interview 30 IC BiH). There is also a lot of peer pressure among
displaced persons and returnees in this regard (Interview 30 IC BiH).
In light of this, discussions about declaring Annex 7 concluded have been
going on for several years. Already in 2002 the ICG warned against “the urge
to redefine the continuing challenge out of existence” (ICG 2002, 40).
According to interviewees, there have always been two attitudes to return:
the Republika Srpska wants everybody to stay where they are, while the
Federation would prefer that everybody returns (Interview 28 IC BiH). This
is reflected in views on the idea of closing Annex 7. The Serb side will say
that Annex 7 is already closed, and the Bosniak side will say that it will never
close (Interview 30 IC BiH).
Discussing the possibility of closing Annex 7, one international
community representative emphasizes that
the rights that are set out in Annex 7 exist independently of Annex 7. So the right to
return and, theoretically, the right to compensation, those rights exist regardless of
Annex 7. And people… Even if somebody, say the High Representative, came out and
said, Annex 7 is now officially implemented and closed, that doesn’t mean people don’t
still have the rights that exist within Annex 7, meaning the right to return, and to be
facilitated in that (Interview 30 IC BiH).
Another international community representative argued that there is a need
to define a limit to at least some of the rights associated with Annex 7,
namely those related to the status of “displaced person.” The time has come
for people to decide whether they are going to return or not. Those who
decide to return need to do so, and those who decide not to return should no
longer be considered potential returnees (Interview 29 IC BiH).
At this time it is not clear whether there is ever going to be an official
declaration regarding the completion of Annex 7 (Interview 30 IC BiH). Such
a declaration was made regarding the property repossession process, but
that was a process involving a smaller—and very specific—number of claims,
which could be clearly determined as settled or not. The sustainable return
of displaced persons (or compensation or other durable solutions) is not as
easy to measure.
51
PEACE BY REPATRIATION
A representative of a local NGO explains that Bosnian organizations
working with displaced persons do not want to close Annex 7. Apart from the
process not being completed, an important reason for this position is that
international assistance for the work that remains to be done would most
likely decrease if there was an official declaration that the job was already
done (Interview 26 NGO BiH).
Presently, a revised strategy for Annex 7 implementation is in the
process of being adopted. The revised strategy, while still emphasizing
return,
recognises the need to compensate people for lost property (instead of a sole focus on
restitution) and to assist the most vulnerable who cannot or do not want to return,
thereby providing de facto support to local integration (IDMC 2008).
UNHCR (2009, 2) describes the new strategy as follows:
A reinvigorated approach to the durable solutions is thus now required on behalf of
competent authorities, supported by the international community, primarily in the
following areas:
1.
Continuing to support the right to voluntary return, but with increased emphasis
on creating conditions for sustainable return and reintegration (primarily in
terms of access to rights and economic sustainability);
2.
Promoting and supporting durable solutions in the place of displacement for
extremely vulnerable IDPs who are not in a position to return.
The UNHCR expressly advocates “that solutions in the place of displacement
be made available, as an urgent priority for extremely vulnerable IDPs”
(UNHCR 2009, 3). In the same document, the UNHCR repeatedly calls for
“pragmatic durable solutions” for vulnerable IDPs who are still not able to
return after more than fifteen years in displacement.
The revised strategy will place less emphasis on return (Interview 25
Gov’t BiH), but at the same time it will not explicitly change its focus to
another durable solution (Interviews 28 IC BiH, 30 IC BiH). Rather,
according to the new strategy local integration will be accepted as a durable
solution for the most vulnerable (Interview 30 IC BiH). Also, the question of
compensation for destroyed property is getting attention for the first time in
many years (Interview 30 IC BiH).
The revised strategy addresses issues such as health, education, the
closure of collective centers—actually the whole social protection system.
Unfortunately, while all this is badly needed, the political will to actually
implement it is lacking (Interview 28 IC BiH). Others agree. According to
one representative of an international organization, when it comes to
addressing the social protection systems in general, it is largely just wishful
thinking (Interview 30 IC BiH).
52
CASES
Several interviewees described the revised strategy as involving a shift of
focus (or perhaps, part of a shift of focus), from finding solutions for
displaced persons in particular to addressing the needs of the Bosnian
population at large, including displaced persons. Improvements in
education, health, and social security are badly needed and would benefit
displaced persons and returnees as well as the rest of the population.
Before, it was return, return, return, return, reconstruction, return, reconstruction,
return. Now after all these years, people have come to the conclusion that you can’t
resolve everything that way, that you need to give the people the freedom to choose
(Interview 28 IC BiH).
The revised strategy does not impose a deadline for the implementation of
Annex 7. The strategy is supposed to cover four years, but, as pointed out by
a representative of an international organization, the important thing is not
to have a deadline. The timing of implementation can always be adjusted.
What is important is the content of the strategy (Interview 28 IC BiH).
The revised strategy has been adopted by the Council of Ministers and
the lower house of the Parliamentary Assembly, but its adoption by the
upper house has been blocked. On several occasions the PIC Steering Board
has expressed its dissatisfaction with the failure of the upper house to adopt
the revised strategy and reiterated the importance of doing so (see UNHCR
2009, 4).
4.3 Peace by Repatriation in Bosnia-Herzegovina
The general perception of the relationship between peace and repatriation in
the case of Bosnia is that repatriation is a necessary condition for peace. The
focus on return has been strong throughout the process, to the extent that
support for alternative solutions actually provided for in the DPA has not
materialized. As cited above, according to Annex 7 of the DPA, all refugees
and displaced persons “shall have the right to have restored to them property
of which they were deprived in the course of hostilities since 1991 and to be
compensated for any property that cannot be restored to them.” Article 14 of
Annex 7 provides for the establishment of a Refugees and Displaced Persons
Fund, to be used in cases of compensation, but as Prettitore (2006, 196–197)
notes, the right of claimants to express a preference for compensation has
been used only for statistical purposes. No compensation has ever been paid;
the Fund was not even established.
There are also those who do not want to return, for example many Serbs
who left Sarajevo in the immediate aftermath of Dayton, which has
implications for the reintegration of Sarajevo (Interview 30 IC BiH).
Displaced persons who do not want to return are organized and work, among
other things, for the right to compensation (Interview 26 NGO BiH). So far
this has been very difficult to achieve.
53
PEACE BY REPATRIATION
In terms of how repatriation contributes to peacebuilding in Bosnia, the
most important aspect has been to undo as far as possible the effects of the
war. More specifically, the goal has been to reverse ethnic cleansing, both in
terms of assisting and encouraging displaced persons to return to their
places of origin, and in terms of strongly promoting their right to reclaim
property lost during the conflict.
An exception to this picture is the Return of Qualified Nationals (RQN)
program37 led by the International Organization for Migration (IOM). Under
this program, “skilled medical, academic and technical experts in sectors
relevant to the reconstruction of the country” (IOM Undated) received
assistance for return to Bosnia. By mid-2000, more than eight hundred
qualified nationals and one thousand eight hundred dependants had
returned under the program. This represents a quarter of a percent of the
total returns, but it has generally been perceived as positive (USCRI 2000,
219; cf. however Black 2001, 187–188).
Any discussion of how return to and within Bosnia contributes to peace
must take into consideration the fact that Bosnia constitutes “an extreme
instance of the politicization of displacement” (Weiss & Pasic 1998, 208).
This means that any decision to return or not to return has political
implications, even when the displaced person making the decision has no
political motives. On the one hand, according to one NGO representative,
Bosnians have not returned in order to express their support for the postconflict political order. They have returned when they feel safe enough to
return, because they have a house, employment et cetera (Interview 26 NGO
BiH).
On the other hand, displaced persons are being used for political
purposes. One international community representative describes how
displaced persons from Srebrenica were convinced in 2007 to set up a
protest camp in Sarajevo and demand that the district of Srebrenica be
transferred to the Federation. This was promoted not because it is in the
interest of the displaced persons, but because it is what some politicians
want. In this particular instance, “there’s a political interest on the Bosniak
side to keep those people as displaced people” (Interview 30 IC BiH).
These problems not only stem from the Bosniak side, and not only from
politicians. There is also a problem of peer pressure among people who have
not yet returned, particularly in Republika Srpska. When people have been
displaced to and remain in an area where they are in the majority, they are
often pressured to stay there in order to help retaining that majority
(Interview 26 NGO BiH). At the same time there are examples of unfortunate
pressure on vulnerable people to return, not least on female heads of
37 It is an exception in the sense that it had another primary rationale than reversing ethnic cleansing; it was
obviously not based on acceptance of ethnic cleansing.
54
CASES
household who, like everybody else, are expected to return to their places of
origin when objective conditions for safe and dignified return are satisfied,
despite the risk of additional trauma, such as the likelihood of meeting
perpetrators of rape and torture (Dzumhur 2010).
Overall, the central argument for return has been the need to reverse
ethnic cleansing and to protect the right of return, which, as noted, was not
among the explanations presented by the UNHCR in 1997 (cf. p 13).
However, there is possibly more to this argument than what is found in the
numerous policy documents on return to Bosnia. Several interviewees
argued in various ways that the return of displaced persons and the remixing
of ethnic groups is important for peace because it contributes to confidence
building and reconciliation. One NGO representative argues that multiethnicity is stabilizing. For example, the recreation of a multiethnic Sarajevo
is the most important argument for return (to Sarajevo)—sharing a city
makes people learn from each other and respect each other (Interview 27
NGO BiH). Similarly, according to an international community
representative, as people return neighbors begin to help neighbors, people
begin to feel safer, and eventually it becomes irrelevant whether you are part
of the majority or not (Interviews 28 IC BiH).
In this context, it is interesting to recall Kaufmann’s theory. He argues
that ethnic conflict makes it more difficult for different ethnic groups to live
together irrespective of whether relations were good or bad prior to the
conflict. Therefore, in order to avoid repeating the security dilemma logic,
different ethnic groups need to be kept apart. One international community
representative argued that while this may be true in some parts of Bosnia it
cannot be generalized to the whole country (Interview 30 IC BiH). Similarly,
another international community representative commented:
Yes, but now they are kept apart. [. . .] You have various areas that are predominantly
Bosniak, predominantly Croat, predominantly Serb, but people are still afraid of each
other! (Interview 28 IC BiH)
Keeping ethnic groups apart after ethnic conflict is possibly a way to avoid
repeating the security dilemma, but it also preserves the fear and suspicion
that the various ethnic groups have developed towards each other, fears that
are in turn nurtured and exploited by nationalist politicians. A politician in a
predominantly Bosniak area can gain support from continually reminding
people about what the Serbs did during the war, and what they would do
again if they had the chance (Interview 28 IC BiH). Return of displaced
persons and ethnic remixing is the most reasonable way to overcome this,
because
55
PEACE BY REPATRIATION
if you let people to live as they used to live before, you know, the fear will disappear.
You know, because you live together, you build relationships, you start after certain
time to trust each other. And then the politicians…, this whole rhetoric would be
defeated (Interview 28 IC BiH).
The final corollary of this line of reasoning is that the idea of temporary
protection makes good sense—although it was supposed to be accompanied
by open borders, which was not always the case (UNHCR 1995b, 89).
Displaced persons will look for durable solutions. If return is out of the
question for the foreseeable future, alternative durable solutions will be
chosen if available. The implication is that if durable solutions had been
available in various host countries in the mid 1990s many refugees would
probably have opted for such solutions. As a result, fewer displaced persons
would have been interested in returning once return became a real
possibility around the year 2000. This would have contributed to cementing
the separation of the ethnic groups in Bosnia, and made mutual suspicion
and fear all the more difficult to overcome. According to this line of
reasoning, the insistence that alternative solutions were temporary, in
anticipation of the day when return would be possible, probably meant that
many more displaced persons returned than would otherwise have been the
case.
56
CASES
and the Search
5 |Nagorno-Karabakh
for Pragmatic Solutions
Although 2004 marked the 10th anniversary of the cease-fire agreement still holding today,
prospects in the future to achieve a peaceful solution to the conflict are considered slight.
Against this background, local integration remains the best and most realistic durable
solution available to refugees from Azerbaijan.
UNHCR38
The war in Nagorno-Karabakh in the early 1990s resulted in large-scale
displacement both within and across the borders of Armenia and Azerbaijan.
In fact, Azerbaijan has the largest per capita IDP population in the world,
with nearly 10 percent of its entire population being internally displaced. In
terms of durable solutions to displacement, the international community
argues that because peace is a necessary condition for return, and because
there is no peace, local integration is the most realistic solution to
displacement.
5.1 Introduction
Nagorno-Karabakh, the mountainous black garden39 of the South Caucasus,
is an old issue of contention between Armenia and Azerbaijan, though not as
ancient as is sometimes believed. After a brief interlude of independence
following the fall of the Russian Empire, the three countries of the South
Caucasus (or Transcaucasus, as viewed from Russia)—Armenia, Azerbaijan,
and Georgia—were conquered by the Red Army during 1920 and 1921, and
became part of the Soviet Union (the Union of Soviet Socialist Republics,
USSR) at its establishment in 1922.
Already at this time there was disagreement about where certain
Caucasian regions should belong, most importantly Nakhchivan and
Karabakh. After a few years of contradictory messages and decisions from
Moscow both regions became part of the Azerbaijan Soviet Socialist Republic
(SSR). Nakhchivan was made an Autonomous Soviet Socialist Republic
(ASSR), the second highest status after the Union Republics, the SSRs, and
Nagorno-Karabakh became an Autonomous Region—the Nagorno-Karabakh
Autonomous Oblast (NKAO)—with lower status than the ASSRs.
Armenian SSR and the Armenian majority population of NagornoKarabakh never accepted the incorporation of Karabakh into Azerbaijan
38 UNHCR (2006).
39 Karabakh means “black garden” and Nagorno means “mountainous.” With reference to the introduction to
this study, it might be noted that there is a Nansen Street in Yerevan (in honor of his devotion to the plight of
the Armenian people) and a Nobel Avenue in Baku (named after Alfred’s brothers, who were instrumental in
developing the Caspian oil industry at the turn of the nineteenth century).
57
PEACE BY REPATRIATION
Map 5.1 Azerbaijan. Source: UN Cartographic Section, Map No. 3761, Revised 7 February 2008.
Reprinted with permission.
SSR, but the dispute remained largely dormant during most of the Soviet
era. However, as the political climate became more open toward the end of
the 1980s, tensions grew, and in 1988 they escalated into violence.
Most of the Soviet republics declared independence in the fall of 1991. In
September, in response to Azerbaijan’s declaration of independence August
30, authorities in the NKAO and Shaumian proclaimed the independence of
the Nagorno-Karabakh Republic (NKR). In December a referendum was
held, and 99 percent of those who voted supported independence. The
Karabakh Azerbaijanis boycotted the referendum (Cornell 1999, 27).
Independence was re-declared in January 1992. No state, including Armenia,
has recognized the independence of the NKR.
Conflict
Following a week of mass demonstrations in Yerevan, on February 20, 1988,
the Regional Soviet of the NKAO appealed to the Supreme Soviets of
Armenian SSR, Azerbaijan SSR and the USSR to be transferred to Armenian
SSR (de Waal 2003, 10).40 This appeal was followed by a series of decisions,
40 The conflict is usually considered as having begun in February 1988. However, de Waal (2003, 18–19)
relates how low scale violence and displacement occurred already in the fall of 1987, a prelude to what was
about to unfold, “as if both sides were picking up a high-frequency radio signal.” See also Kaufman (2001).
58
CASES
Map 5.2 Nagorno-Karabakh, conflict area.
59
PEACE BY REPATRIATION
votes, and statements by the political bodies mentioned. The NKAO and
Armenian SSR spoke in favor of unification with Armenia, and Azerbaijan
SSR and the Supreme Soviet of the USSR rejected it. For a more detailed
description of these developments, see Cornell (1999, 13–30). Only a week
after the original appeal by the Regional Soviet of the NKAO that it be
allowed to join with Armenia, there were riots in Sumgait, an industrial town
just north of Baku on the Abseron peninsula. More than thirty people were
killed, including twenty-six Armenians and six Azerbaijanis.
In early 1989 Soviet troops were deployed to Nagorno-Karabakh, and the
region was placed under the direct rule of Moscow. From 1988 through 1991
the main protagonists were the central Soviet Authorities and the Armenian
SSR. After the dissolution of the Soviet Union in late December 1991 the
primary parties became instead the Republic of Azerbaijan and the selfdeclared NKR, backed by Armenia.
The war ended with a ceasefire, the Bishkek protocol, brokered by Russia
and signed on May 5, 1994. The ceasefire line is referred to as “the line of
contact.” By the end of the war and until today, Nagorno-Karabakh is de
facto connected to Armenia, but remains de jure part of Azerbaijan.
The territorial consequences of the war have been succinctly
summarized by the ICG (2005a, 1–2) (cf. map p 58). The NKAO covered
4,388 sq. km (5 percent of the territory of Azerbaijan). Since 1994 the de
facto NKR controls 4,061 sq. km (92 percent) of the former NKAO, or
Nagorno-Karabakh proper. In addition, the NKR occupies more than 7,400
sq. km outside Nagorno-Karabakh proper, including the entire districts of
Kelbajar, Lachin, Kubatly, Jebrail and Zangelan, as well as parts of Agdam
and Fizuli. In total, the NKR, supported by Armenia, occupies 11,722 sq. km,
or 13 percent, of Azerbaijan’s territory. Below, the term “district” will be used
to refer to individual districts, and the term “occupied territories” will be
used when all seven are considered.
In addition, the NKR claims another 1,000 sq. km of territory still under
Azerbaijani control (or, according to the NKR, under Azerbaijani
“occupation”). This includes 327 sq. km of the districts of Martuni and
Mardakert, both of which were part of the NKAO, plus 701 sq. km outside the
NKAO, namely the district of Shaumian, which also declared independence
from Azerbaijan in 1991, but remains under Azerbaijani control.
Displacement
The conflict resulted in large-scale displacement, both within and between
the two countries. During the first phase of the conflict, between Armenia
and the Soviet Union, some two hundred thousand ethnic Azerbaijanis left
Armenian SSR mainly for Azerbaijan SSR, and at least three hundred
thousand ethnic Armenians left Azerbaijan SSR primarily for Armenian SSR
and Russian SSR. Strictly speaking, with the USSR still in place at this time,
60
CASES
these people were internally displaced. However, since the dissolution of the
USSR in 1991 they have been considered refugees. 41
By the end of the conflict there were some two hundred ninety thousand
Armenian refugees from Azerbaijan residing in Armenia (including eighteen
thousand from Nagorno-Karabakh), and another hundred thousand in
Russia. There were also seventy thousand Armenian IDPs. In Azerbaijan
there were around one hundred eighty-five thousand Azerbaijani refugees
from Armenia, and six hundred seventy thousand IDPs (USCRI 1995;
1996).42 The vast majority of the Azerbaijani IDPs originated in the occupied
territories; some 10 percent came from Nagorno-Karabakh proper
(Interviews 1 IC Azerbaijan, 12 NGO Azerbaijan).
Because the de facto authorities in Stepanakert, the capital of NagornoKarabakh, considers the NKR to be an independent state, ethnic Armenians
displaced to Nagorno-Karabakh from other areas of Azerbaijan are
considered refugees, and only those displaced within the territory claimed by
the NKR are seen as IDPs. The international community, however, does not
recognize the NKR, and considers all those displaced within the borders of
Azerbaijan—including Nagorno-Karabakh and the occupied territories, and
irrespective of ethnicity—as IDPs. The number of Armenian IDPs in
Azerbaijan—that is, ethnic Armenians displaced within the borders of
Azerbaijan, and de facto residing in Nagorno-Karabakh and the occupied
territories—is estimated at between fifty and sixty thousand. These are more
or less equally divided between what the de facto authorities refer to as
refugees and IDPs (Interview 24 NGO Nagorno-Karabakh).
One result of the displacement is that Armenia and Azerbaijan, as well as
Nagorno-Karabakh, have all become more homogeneous than they were
before the conflict. According to the last Soviet census, conducted in 1989,
Armenia had a total population of just under 3.3 million, which was 93
percent Armenian, and Azerbaijan’s 6.8 million population was 85 percent
Azerbaijani (Library of Congress 2009). By contrast, in the early twenty-first
century, Armenia is 98 percent Armenian, and Azerbaijan is 91 percent
Azerbaijani (ICG 2005a, 2). In the pre-war NKAO, 77 percent of the one
hundred eighty-eight thousand population was Armenian, and in the postwar de facto NKR, the one hundred twenty-five thousand population is 99
percent Armenian (Rowland 2008, 104, 108). One NGO representative in
41 Displacement statistics for Armenia and Azerbaijan in the 1990s also contain large groups of displaced
persons not related to the conflict over Nagorno-Karabakh, and not considered in the present analysis. Most
importantly, Azerbaijan is the host of some fifty thousand Meskhetian Turks (deported from Georgia to
Central Asia in the 1940s) who fled ethnic violence in the Fergana Valley region of Uzbekistan in 1988. There
were also some ten thousand Chechens, who arrived after the conflict in Chechnya resumed in 1999. In
Armenia, several hundred thousand people remained displaced after an earthquake in December 1988.
42 Fifteen years on, the number of IDPs in Azerbaijan varies, and numbers from the same source do not
always add up. The Internal Displacement Monitoring Centre sets the number at 573,000–603,000 (IDMC
Undated).
61
PEACE BY REPATRIATION
Nagorno-Karabakh could think of no more than five mixed (ArmenianAzerbaijani) couples in Stepanakert at this time (Interview 24 NGO
Nagorno-Karabakh).
The population of Nagorno-Karabakh is markedly smaller today than
twenty years ago. In the area of the former NKAO, the population has
decreased by 35 percent from around one hundred ninety thousand to one
hundred twenty-five thousand, and in the total area of the NKR, the
population has decreased by 60 percent, from three hundred forty thousand
to one hundred thirty-five thousand (Rowland 2008, 104). These numbers
show that the occupied territories are very sparsely populated (cf. ICG
2005b, 25).
Azerbaijanis who left Armenia were primarily rural, as were most of the
IDPs, with the exception of people from Agdam (Interviews 1 IC Azerbaijan,
9 NGO Azerbaijan). Today, most of them live in Baku, Sumgait, or other
cities. Armenians who left Azerbaijan, however, were largely urban and
many used to work in the oil industry. Today they live in rural areas in
Armenia (Interviews 9 NGO Azerbaijan, 16 Gov’t Armenia). One NGO
representative, a displaced person from Baku, guided me around Shusha,
and as we passed a small puddle of oil outside the entrance to his garden he
inhaled and said that the smell made him nostalgic.
Most people displaced by the conflict have been unable to return home.
The exception is the Armenian IDPs and some of the Karabakh Armenians.
The ICG explains that some Azerbaijani IDPs returned in the late 1990s (ICG
2005b, 26), but according to a well-positioned interviewee in Azerbaijan
there has been no return to Nagorno-Karabakh since the cease-fire. During
the conflict control of certain territories went back and forth, and people
went with it; some refer to that as return. Further, in the last couple of years,
many Azerbaijani IDPs have moved to new settlements near the occupied
territories (cf. below), but this does not mean that they have returned home
(Interview 1 IC Azerbaijan).
5.2 Negotiating Peace in Nagorno-Karabakh
The OSCE Minsk Process
Since 1992, peace negotiations between Armenia and Azerbaijan have been
conducted under the auspices of the OSCE.43 The original plan to arrange a
peace conference in Minsk, the Belarusian capital, has not yet materialized
but lives on in the “Minsk Group,” which has responsibility for the
negotiation process. Since 1997 the Minsk Group has been co-chaired by
France, Russia and the USA. Also since 1997, Ambassador Andrzej Kasprzyk
43 Until 1995 called the Conference on Security and Cooperation in Europe, CSCE.
62
CASES
has been the Personal Representative of the Chairman-in-Office on the
Conflict Dealt with by the OSCE Minsk Conference.
The de facto authorities in the NKR are not party to the negotiations,
only Armenia and Azerbaijan are. However, relations between Armenia and
the NKR are very intimate. For example, the first de facto president of the
NKR, Robert Kocharyan, 1994–1997, went on to become prime minister of
Armenia in 1997 and then president from 1998 to 2008. The current
president of Armenia, Serzh Sargsyan (who also served as prime minster for
a short period), served as chairman of the Nagorno-Karabakh “Self-Defense
Forces Committee” during most of the war. Both of them were born in
Stepanakert. According to Cornell (1999, 27) “Karabakh can be said to
control Armenia to a much higher extent than the opposite.”
Armenia and Azerbaijan have been members of the OSCE since 1992.
Both pledge their support for the organization’s ten fundamental “principles
guiding relations between participating states,” but each has its own favorite
principle. Armenia emphasizes the equal rights and self-determination of
peoples outlined in principle VIII, while Azerbaijan hails the territorial
integrity of states guaranteed by principle IV.44 This was also noted by some
of the interviewees (Interviews 7 IC Azerbaijan, 18 IC Armenia).
Progress has been slow, though from 2004 negotiations have been more
intensive within the framework of the Prague Process (still within the Minsk
framework), involving direct meetings between the Armenian and
Azerbaijani presidents and foreign ministers. For example, Armenian
President Kocharyan and Azerbaijani President Ilham Aliyev met three times
during 2006. Still, no agreement was reached.
Parallel to these negotiations both governments present hard-line
positions for public consumption at home. An important consequence of this
is that there is very little public support for a peace agreement. In addition,
those who try to work for reconciliation and coexistence are often considered
traitors and risk harassment. The need to prepare the populations for
necessary compromises, and the complete failure to do so up to today, was
noted by several interviewees (Interviews 7 IC Azerbaijan, 9 NGO
Azerbaijan, 17 NGO Armenia, 18 IC Armenia); see also Dehdashti (1997,
478), ICG (2005a, 24), and Kaldor (2007, 165) for similar points.
44 The international community has not come out clearly in favor of either principle (which are obviously
relevant beyond the OSCE). An illustrative example of this vacillation is the European Neighborhood Policy
Action Plans for Armenia, Azerbaijan, and Georgia (which has its own breakaway regions, most importantly
Abkhazia and South Ossetia). These plans were all adopted in November 2006, but they emphasize different
principles. The EU/Azerbaijan and the EU/Georgia Action Plans both stress respect for sovereignty and
territorial integrity, and neither contains the term self-determination. Conversely, the EU/Armenia Action
Plan makes no mention of territorial integrity, but instead maintains that the Nagorno-Karabakh conflict
should be resolved “on the basis of international norms and principles, including the principle of selfdetermination of peoples” (Johansson 2009b, 126).
63
PEACE BY REPATRIATION
In the summer of 2006 the co-chairs of the Minsk Group presented a
number of basic principles for a settlement of the conflict: withdrawal of
Armenian forces from and demilitarization of the occupied territories,
deployment of international peacekeepers, demining, reconstruction,
resettlement of IDPs, and some interim status for Nagorno-Karabakh, with
final status to be determined in a popular referendum (OSCE 2006; cf. de
Waal 2009, 3; ICG 2007, i). The plan thus revealed was not a big surprise
(see e.g., ICG 2005b, 18–29; Lynch 2004, 116–119), but the decision to make
it official was highly significant in light of the hard-line positions presented
by both governments at home. Going public with what was on the agenda
was intended to prevent the two governments from misleading their publics
about what they were discussing.
Armenia has long favored a package approach to avoid making any
commitments until satisfied with the question of the final status of NagornoKarabakh, while Azerbaijan has preferred a more gradual approach. The
Prague process has been described as an attempt to combine agreement on a
package with step-by-step implementation of it (ICG 2005b, 11). For
example, on the particular issue of the future status of Nagorno-Karabakh,
the aim of the negotiations has shifted from reaching agreement on status to
reaching agreement on how to determine status.
The future status of Nagorno-Karabakh is the most contentious issue for
both sides. Both Armenians and Azerbaijanis are adamant that ultimate
sovereignty over Nagorno-Karabakh must rest with them. Some form of farreaching autonomy for Nagorno-Karabakh within Azerbaijan is a likely
arrangement, though presently not acceptable to the Armenian side. Several
Azerbaijani interviewees referred to President Aliyev’s proclaimed readiness
to offer “the highest degree of autonomy that exists in the world” (cf. ICG
2005b, 13) to Nagorno-Karabakh. Tatarstan and the Åland islands45 were
mentioned as models (Interviews 9 NGO Azerbaijan, 10 NGO Azerbaijan, 11
NGO Azerbaijan).
Various options for permanent status have been discussed over the
years, such as some confederal arrangement between Azerbaijan and
Nagorno-Karabakh (see e.g., ICG 2005b). There have been talks (at least by
observers) about a land swap, possibly involving the creation of a land link
between Nakhchivan and Azerbaijan proper. However, as noted, the present
strategy seems to be to postpone the status issue and eventually settle it
through an internationally sanctioned referendum. Many obstacles remain
to be overcome if a referendum is to be successful, or even possible.
Interviewees did not place a lot of hope in that solution. Importantly, the
Constitution of Azerbaijan requires a nation-wide referendum for any
45 Tatarstan is an autonomous republic within the Russian Federation. The Åland islands are an autonomous,
Swedish-speaking region of Finland.
64
CASES
change in the borders of the country. This precludes any regional
referendum limited to Nagorno-Karabakh. Additionally, the Constitution can
only be amended by way of a nation-wide referendum.
In the unlikely case that the Azerbaijani constitution is indeed amended
to make a regional referendum possible, the outcome will depend on who is
eligible to vote as well as on what territory the referendum would actually
concern. According to one Azerbaijani NGO representative there was a clear
65 percent Armenian majority in the former NKAO before the conflict, but in
the territory now controlled by the NKR (that is, the NKAO plus the occupied
territories) there were 70–80 percent Azerbaijanis before the war (Interview
11 NGO Azerbaijan). While the accuracy of these figures may be questioned
(cf. above), it is clear that the pre-war Armenian population in the area was
concentrated to Nagorno-Karabakh proper, and was markedly smaller in the
surrounding districts.
It is not known how far negotiations have progressed on such issues as
eligibility to vote and what geographic area the referendum would actually
concern, but the issues are clearly problematic (cf. Johansson 2009b). Most
Karabakh Armenians are against a referendum. They note that there have
already been two referenda on the issue, both confirming the independence
of the NKR, and argue that there is no need for a third one (Interviews 21 de
facto Gov’t Nagorno-Karabakh, 24 NGO Nagorno-Karabakh).
The security aspects of a solution are clearer. The general consensus
seems to be an Armenian withdrawal from five to six-and-a-half of the
occupied districts, with some special arrangements for Lachin, which
provides the land corridor between Armenia and Nagorno-Karabakh, and
probably for Kelbajar, which is also situated between NKAO and Armenia.
One option is to handle Lachin and Kelbajar in the same way that Brčko was
handled in Bosnia-Herzegovina (cf. chapter 4), that is, through international
arbitration (Interview 1 IC Azerbaijan).
In Azerbaijan the distinction between the former NKAO and the
occupied territories is very clear. Both local actors and the international
community in Azerbaijan basically take it for granted that the occupied
territories will be returned to Azerbaijani control in any resolution of the
conflict. They maintain that the real conflict is about Nagorno-Karabakh
proper (Interview 1 IC Azerbaijan, 13 Gov’t Azerbaijan). Conversely, the
distinction is not very relevant to Armenians—and it is becoming less so
(Interview 18 IC Armenia). In Armenia and Nagorno-Karabakh I saw several
maps—including in the offices of international organizations in Yerevan—
that did not distinguish between different parts of the NKR. A representative
of the de facto NKR authorities argued that while compromise will be needed
to reach a solution to the conflict, the NKR cannot make compromises about
territories that are either strategically important or “that were historically
ours” (Interview 21 de facto Gov’t Nagorno-Karabakh).
65
PEACE BY REPATRIATION
Deployment of a peacekeeping force has also been on the agenda during
the entire Minsk process. Dehdashti (2000, 263–268) describes a number of
plans presented in the mid 1990s, proposals for stationing from several
hundred up to five thousand peacekeepers and military observers in the
occupied territories. de Waal (2009, 3) describes plans to deploy up to ten
thousand peacekeepers, primarily in the Kelbajar region. So far, however,
the ceasefire has been basically self-regulated, merely overseen by
Ambassador Kasprzyk’s field assistants. They regularly visit the line of
contact, but are required to announce their intention to do so in advance.
Incidents do occur, and in 2005 the UCDP coded the conflict as active at
minor level. A few years later, the UCDP noted that according to some
accounts clashes in the spring of 2008 were the most serious since the 1994
ceasefire (UCDP 2009a).
The Minsk process is one of the most closed and confidential peace
processes in the world. Very few people are involved, and there is no
spokesperson or press secretary of the Minsk Group co-chairs (de Waal
2009, 3). The elite character of the Minsk Group negotiations is one of the
problems of the peace processes, because any peace initiatives must be
approved by either Baku or Yerevan, and everybody is extremely suspicious
(Interview 7 IC Azerbaijan).
At the same time, there seems to be a consensus that reaching a solution
to the conflict is the responsibility of the governments. NGO activists in both
countries explained that civil society organizations do important work on
humanitarian issues and reconciliation initiatives, but that it is up to the
governments and the presidents to resolve the political conflict (Interviews 6
NGO Azerbaijan, 10 NGO Azerbaijan, 12 NGO Azerbaijan, 20 NGO
Armenia). Even a high level ministry official in Yerevan, when asked about
the repercussions of the conflict in Georgia on the Minsk Group and Russia’s
role in the negotiations, declined to comment on the issue, because “our
department does not deal with problems of peace. [. . .] Only the minister of
the exterior is authorized to deal with these issues” (Interview 16 Gov’t
Armenia). While several interviewees regretted the closed character of the
Minsk process, some noted that the problem is not whether the Minsk Group
is open or closed, public or confidential—the problem is that they do not
produce any results (Interviews 10 NGO Azerbaijan, 11 NGO Azerbaijan).
Leaving conflict resolution to the governments, civil society concentrates
on the task of preparing their respective populations for compromise,
reconciliation, and once again living together in peace. The governments
make no efforts whatever in that regard, rather the opposite. Since there are
no official relations between Armenia and Azerbaijan, there can be no official
initiatives for reconciliation (Interview 13 Gov’t Azerbaijan). NGOs working
with these issues are often ridiculed or even accused of treason (Interviews 6
NGO Azerbaijan, 9 NGO Azerbaijan, 15 NGO Armenia, 17 NGO Armenia).
66
CASES
Nevertheless, a few organizations are able to engage Karabakh Armenians in
dialogue (Interviews 9 NGO Azerbaijan, 20 NGO Armenia, 24 NGO
Nagorno-Karabakh). However, as the head of an Armenian NGO points out,
the international community, which wants dialogue, does not fund projects in
Nagorno-Karabakh, and the Armenian diaspora, which funds projects in NagornoKarabakh, is not interested in promoting dialogue” (Interview 17 NGO Armenia).
In light of this, it is hard to see that NGO communication will make any real
difference (Interviews 8 IC Azerbaijan, 17 NGO Armenia).
A few interviewees noted, in the context of a possible future referendum,
that the Azerbaijani government claims it wants to convince Karabakh
Armenians that they are better off as part of Azerbaijan than as independent,
particularly in economic terms. However, interviewees also noted that it is
difficult to present a convincing argument to someone you refuse to talk to
(Interviews 12 NGO Azerbaijan, 14 NGO Azerbaijan). Azerbaijan also
considers Karabakh Armenians as Azerbaijani citizens. For example, they
were encouraged to participate in the 2005 elections in Azerbaijan, though
for obvious reasons there were no polling stations in Karabakh (Interview 13
Gov’t Azerbaijan).
The Regional Context
During my visit to the Caucasus in August and September 2008, the South
Ossetia conflict was high on the agenda. A ceasefire agreement was reached
only a week before I arrived in Baku, and Russian withdrawal continued for
another several weeks. A number of interviewees argued that the Minsk
formula would have to be reconsidered after the conflict, which had
demonstrated how strong Russia’s strategic interests in the South Caucasus
are (Interviews 6 NGO Azerbaijan, 7 IC Azerbaijan, 9 NGO Azerbaijan, 14
NGO Azerbaijan, 15 NGO Armenia, 17 NGO Armenia, 18 IC Armenia). On
the other hand, the difference between Russia’s role as party to the conflict in
South Ossetia and intermediary in the Nagorno-Karabakh conflict was also
noted (Interview 13 Gov’t Azerbaijan).
Some interpreted the South Ossetia conflict as a Russian message to all
three South Caucasian republics not to get too friendly with the West, or
alternatively as a message particularly directed at Azerbaijan that Russia
would react strongly to any Azerbaijani attempt to retake Nagorno-Karabakh
by force (Interviews 1 IC Azerbaijan, 6 NGO Azerbaijan). Several
interviewees contended that the three South Caucasian republics need to
agree on a common path, east or west, Russia or NATO, and that regional
integration is the key to long-term peace and development (Interviews 6
NGO Azerbaijan, 9 NGO Azerbaijan, 13 Gov’t Azerbaijan, 15 NGO Armenia).
67
PEACE BY REPATRIATION
One interviewee explained that after Russia entered South Ossetia many
Armenians greeted Russians in Nagorno-Karabakh and thanked them
(Interview 23 NGO Nagorno-Karabakh). Another expressed his hopes that
Russian recognition of Abkhazia and South Ossetia would “strengthen the
precedent of Kosovo” (Interview 24 NGO Nagorno-Karabakh). However, the
decision by the de facto authorities of the NKR to recognize Abkhazia and
South Ossetia will not make it easier to use Georgia as a neutral venue for
NGO meetings (Interview 17 NGO Armenia). As far as recognition is
concerned, one Karabaghian NGO representative hoped that the de facto
NKR government’s strategy was to be recognized by both sides, not only by
Armenia. In fact, he said, “I want to believe that the first country that
recognizes us is Azerbaijan” (Interview 24 NGO Nagorno-Karabakh).
Conversely, Armenian recognition of Nagorno-Karabakh is the one thing
that could lead to renewed conflict, according to one interviewee. It would
simply be too much for the Azerbaijani government (Interview 12 NGO
Azerbaijan).
Some interviewees argued that it is largely up to Russia to decide when
the Nagorno-Karabakh conflict can be settled (Interviews 1 IC Azerbaijan, 6
NGO Azerbaijan), while others believed that if Armenia and Azerbaijan were
to reach a peace agreement, outside powers could not prevent it (Interviews
13 Gov’t Azerbaijan, 14 NGO Azerbaijan). An Armenian NGO representative
noted that Russia has tried to handle the Nagorno-Karabakh conflict on its
own from the beginning, side-tracking the Minsk Group (Interview 15 NGO
Armenia). The Bishkek protocol (that is, the 1994 ceasefire agreement) was
mentioned as an example of that.
In recent years there have been warnings that Azerbaijan might attempt
to re-establish control over Nagorno-Karabakh by force. This perceived
threat is largely a result of Baku’s being emboldened by its oil wealth (ICG
2005b). Having been the world’s second largest producer of oil in the early
1900s, Azerbaijan saw its oil production decline toward the end of the
century. However, since 1998 it is once again advancing (Kaldor 2007, 167)—
and fast. Increasing production in combination with rising oil prices made
Azerbaijan the fastest growing economy in the world in 2006, with GDP up
34.5 percent (ICG 2007, 8). Substantial amounts of money have gone into
military spending, which increased by 51 percent from 2004 to 2005, and
again by 82 percent in 2006. In 2007, President Aliyev pledged to make the
Azerbaijani military budget as large as Armenia’s entire state budget (ICG
2007, 12).
Some argue that as the oil revenues begin to decline—as they will by
2012 according to most forecasts—the Azerbaijani authorities might want to
divert attention from declining wealth by pursuing renewed efforts at
retaking Nagorno-Karabakh (ICG 2007, 8). However, interviewees did not
believe that the threat is real, with several arguing that any resort to force is
68
CASES
bound to fail miserably and that the Azerbaijani government is fully aware of
this (Interviews 1 IC Azerbaijan, 7 IC Azerbaijan, 12 NGO Azerbaijan) In
addition, if oil revenues are bound to start declining sooner or later, a new
war would be instantly devastating to them (Interview 14 NGO Azerbaijan).
Azerbaijani IDPs are generally more positive to the military option than the
rest of the population. The ICG cites a survey showing that 13 percent of all
Azerbaijanis favored a military solution; the corresponding figure for IDPs
was 84 percent (ICG 2005a, 28). According to one Azerbaijani NGO
representative, most IDPs favor a peaceful solution to the conflict, but as
time passes there is increasing support for the military option (Interview 10
NGO Azerbaijan).
Oil and gas from the Caspian Sea are central to political developments in
the region. The Baku-Tbilisi-Ceyhan (BTC) pipeline, which takes Azerbaijani
oil to the Mediterranean through Georgia and Turkey, painstakingly avoids
both Russian and Armenian territory and competes with pipelines through
Russia. Suggesting that Russia is the key to resolving the Nagorno-Karabakh
conflict, one interviewee suspected that Russian inaction regarding the
conflict is a way of punishing Azerbaijan for the BTC pipeline (Interview 1 IC
Azerbaijan).
Addressing Displacement
As described above, Armenian refugees from Azerbaijan and Azerbaijani
refugees from Armenia are not expected to return when the conflict is
resolved (Interviews 7 IC Azerbaijan, 15 NGO Armenia, 16 Gov’t Armenia, 21
de facto Gov’t Nagorno-Karabakh). The solution to their displacement is
overwhelmingly local integration. According to an Armenian government
representative, of the three durable solutions to displacement return is
impossible and third country resettlement is improbable. Therefore, local
integration is the only reasonable solution for the Armenian refugees
(Interview 16 Gov’t Armenia). The final stage of local integration is
citizenship, and some eighty-five thousand Armenian refugees have already
received citizenship. Another two hundred fifty to three hundred thousand
still have refugee status (Interview 16 Gov’t Armenia). Local integration is
also believed to be in line with the wishes of the refugees—both Armenian
and Azerbaijani. While many are hoping for some kind of compensation for
lost property, very few would be interested in actually returning (Interviews
7 IC Azerbaijan, 15 NGO Armenia, 17 NGO Armenia).
However, the promotion of local integration as a durable solution is not
entirely uncontroversial. One NGO representative argues that, on the one
hand, the UNHCR puts pressure on Armenia to provide citizenship for
refugees from Azerbaijan, while promising to continue to assist those in need
of assistance. On the other hand, when new citizens are no longer considered
displaced, the UNHCR decreases its budget for Armenia (Interview 20 NGO
69
PEACE BY REPATRIATION
Armenia). Further, out of seventy thousand Armenian IDPs from the
districts close to the border with Azerbaijan, around sixty thousand have
been able to return or settle elsewhere. Between eight thousand and ten
thousand are still in need of durable solutions (IDMC 2010, 53).
The category of displaced persons for which repatriation remains an
issue, therefore, is the Azerbaijani IDPs, numbering more than half a million,
and they are most often seen as firmly committed to return (ICG 2005a, 16).
This picture was partly confirmed by several interviewees. As far as first
generation IDPs are concerned, return is decidedly perceived as the only
possible durable solution. Second generation IDPs do not necessarily
consider themselves Karabaghians but would nevertheless want to establish
ties with their ancestral homes, to be able to visit the areas, and to bring their
children there during holidays (Interviews 1 IC Azerbaijan, 10 NGO
Azerbaijan, 11 NGO Azerbaijan, 13 Gov’t Azerbaijan). One interviewee
pointed out that many IDPs compare their poor living conditions in
displacement with their normal and good living conditions before
displacement and argue that they want to return as soon as possible to get
away from the poor living conditions of displacement (Interview 10 NGO
Azerbaijan). The fact that the situation in the places of origin are not the
same as before the conflict is easily overlooked.
There is no alternative, no alternative to returning to Nagorno-Karabakh, because the
conditions here are bad, but conditions were okay in Karabakh. IDPs are now living in
dormitories, in awful conditions; they don’t see any alternative to returning (Interview
10 NGO Azerbaijan).
Similar sentiments have been expressed in relation to other conflicts, for
example among Georgian IDPs from Abkhazia (cf. Kabachnik, Regulska &
Mitchneck 2010).
The international community has been unequivocal about protecting the
right of return of Azerbaijani IDPs (ICG 2007, 5) as well as the other groups
of displaced Armenians and Azerbaijanis. However, return has to happen
within the framework of a broader political process, not least because there
is an enormous need for reconstruction in the occupied territories
(Interviews 1 IC Azerbaijan, 7 IC Azerbaijan).
The displacement situation is clearly not only a humanitarian issue. It is
part of the highly political demographic dimension of the conflict. An
example of this dimension is Azerbaijan’s 2004 complaint to the UN General
Assembly about non-Azerbaijani citizens moving from Armenia to NagornoKarabakh and the occupied territories in order to change the demographic
composition of the area (ICG 2005a). An OSCE Fact Finding Mission (FFM)
was subsequently deployed for the purpose of establishing whether such
migration was the result of a deliberate policy on the part of Armenia. The
FFM concluded that this was not the case.
70
CASES
When it comes to the return of displaced persons, most plans presented
during the peace process are weak on IDP return to Nagorno-Karabakh (ICG
2005b, 24–25). Regarding return, however, there is, again, an important
difference between the occupied territories and Nagorno-Karabakh proper.
As Lynch (2004, 118, 129) observes, while the great majority of Azerbaijani
IDPs could return to their former homes in the occupied territories as soon
as these were returned to Azerbaijani control, return of IDPs to their former
homes inside Nagorno-Karabakh “would not be immediate, nor would it be
pressed by the international community in order to avoid the security
dilemmas that may arise from the mingling of the two populations.”
Several interviewees agreed, emphasizing the fundamental distinction
between the former NKAO and the occupied territories also regarding the
prospects for return (Interviews 1 IC Azerbaijan, 8 IC Azerbaijan, 12 NGO
Azerbaijan). As noted, the share of Azerbaijani IDPs from NKAO is believed
to be around 10 percent, meaning that the vast majority of IDPs would be
able to return if the occupied territories, or even just five districts, were
returned to Azerbaijani control (Interview 1 IC Azerbaijan). Not everyone
believes that the Azerbaijani IDPs can return to Nagorno-Karabakh,
however, because the Karabakh Armenians do not trust them. “[The
Karabakh Armenians] argue that they were lucky to survive the first time,
but that they may not be so lucky the second time” (Interview 17 NGO
Armenia).
Another aspect of the politicization of demography is the claim that
Azerbaijan has a non-integration policy for IDPs, supposedly to demonstrate
that the conflict is unresolved and that displacement is temporary (see e.g.
ICG 2007, 16). Throughout the 1990s the Azerbaijani authorities
concentrated on resettling refugees from Armenia. The IDPs, on the other
hand, have always been expected to return, which means that any
arrangements to care for them during displacement are by their very nature
temporary. Summarizing these sentiments, one international community
representative in Baku observed that “the IDPs are the physical indicator of
Azerbaijan’s claim to Nagorno-Karabakh” (Interview 1 IC Azerbaijan).
Since the early 2000s, partly as a result of increasing oil revenues, the
IDPs have received better care, but the notion of temporality remains central
(ICG 2005a, 16). The same international community interviewee explains
that
while the government finally wants to give the IDPs better living conditions it is not
ready to accept local integration as a de facto durable solution. The conception is that
accepting local integration as a durable solution for the IDP is equivalent to giving up
Azerbaijan’s claims to Nagorno-Karabakh. In practice, however, local integration is
taking place on a large scale—we just can’t refer to it as a durable solution (Interview 1
IC Azerbaijan).
71
PEACE BY REPATRIATION
Despite the Azerbaijani unwillingness to accept local integration as a durable
solution, most interviewees (Azerbaijanis as well as international community
representatives) did not subscribe to the notion that the government is
deliberately trying to keep the IDP population isolated from the general
population as a reminder to the world of its claims to Nagorno-Karabakh
(Interviews 1 IC Azerbaijan, 7 IC Azerbaijan, 10 NGO Azerbaijan, 11 NGO
Azerbaijan). On the contrary, the international community is generally
positive to how the Azerbaijani government handles the displacement
situation (Interview 1 IC Azerbaijan). Interviewees emphasized that there is
no discrimination of IDPs; in some respects they are actually favored. For
example, IDPs cannot be evicted from their housing (Interviews 1 IC
Azerbaijan, 9 NGO Azerbaijan). Some, however, expressed doubt that the
government actually wants the IDPs to return; “if they did, we would already
be back” (Interview 11 NGO Azerbaijan).
Since 2004 the Azerbaijani government is pursuing a “Framework Plan
on the Return of Displaced Persons” (or Great Return Plan) in cooperation
with several international organizations (ICG 2005a, 16–17). It contains
three points: (1) to institutionalize a common view of what constitutes
voluntary repatriation, (2) to agree on necessary conditions for voluntary
repatriation, and (3) to define the government’s sectoral responsibilities. The
third point has thus far been the least successful (Interview 1 IC Azerbaijan).
Part of this plan is the construction of new settlements46 for IDPs near the
occupied territories, to which around ninety thousand IDPs had moved by
early 2008 (NRC 2008, 4).
5.3 Peace by Repatriation in Nagorno-Karabakh
While the return of displaced persons is not a high priority issue in the
various proposals—official and unofficial—produced by the Minsk Group
over the years, it is nevertheless included. However, what cannot be found in
the proposals is a perception of return as being important for peace. In
contrast to the Bosnian case, where peace cannot be achieved without return,
in the case of Nagorno-Karabakh, security and stability must come first—
peace is a necessary condition for return.
There are different views on the relationship between repatriation and
peace in the case of Nagorno-Karabakh. Refugees are not expected to return
at all for a long time, whether a political solution to the conflict is found or
not. This includes Armenian refugees from Azerbaijan as well as Azerbaijani
refugees from Armenia. As noted, the de facto authorities (as well as the
inhabitants) of the NKR also consider Armenians who fled to the NKR from
other parts of Azerbaijan as refugees; they are also not likely to return. The
group of displaced persons for whom return is an issue is the Azerbaijani
46 Such as the one depicted on the cover of this study.
72
CASES
IDPs, and as far as they are concerned the general consensus seems to be
that they may be able to return once a political settlement of the conflict has
been reached. With reference to Adelman’s positions, their eventual return
will be a sign that peace is in place.
An Armenian government representative argued that peace and return
are certainly linked, but emphasized that the order must be peace first and
return second (Interview 16 Gov’t Armenia). Others agree, and outline the
sequence of a settlement as first a political agreement that resolves the
conflict, then withdrawal of all military forces, then return. Because
displacement was a result of war, return will be the result of peace (Interview
11 NGO Azerbaijan).
Different actors draw different conclusions from this, and would
complete the following sentence in different ways:



Azerbaijanis: Return of Azerbaijani IDPs to Nagorno-Karabakh and
the occupied territories will be a sign that peace is in place, and that
is why we need a political solution to the conflict soon, so return—
the only conceivable durable solution to displacement—can begin.
Armenians: Return of Azerbaijani IDPs to Nagorno-Karabakh and
the occupied territories will be a sign that peace is in place, because
only with peace in place can we possibly accept the return of
Azerbaijanis.
International community: Return of Azerbaijani IDPs to NagornoKarabakh and the occupied territories will be a sign that peace is in
place, but because peace is not in place we must look to other
durable solutions to displacement.
Not surprisingly, then, the return of Azerbaijani IDPs to Nagorno-Karabakh
and the occupied territories is not expected to make an important
contribution to peace. Peace will make return possible, and return will help
resolving the problem of displacement, but return is not perceived as
something that will make peace any more likely or sustainable. In terms of
how return would contribute to peace, a few interviewees admitted that
returnees would contribute to the physical reconstruction of the war-torn
areas and to the development of the post-conflict economy. These comments
came only after I had suggested how return and returnees can contribute to
peace (Interviews 10 NGO Azerbaijan, 11 NGO Azerbaijan). Alternatively,
returnees can contribute to validating the post-conflict political order, in this
case not primarily through the act of returning, but by participating in an
eventual referendum, should it become possible to arrange one (Interviews 1
IC Azerbaijan, 7 IC Azerbaijan).
Regarding pre-war inter-ethnic relations, there was limited interaction
between Armenians and Azerbaijanis before the conflict (Interview 3 IC
73
PEACE BY REPATRIATION
Azerbaijan). One international community representative with previous
working experience from the Balkans compared relations between
Armenians and Azerbaijanis to those between Serbs and Albanians. These
latter groups displayed stronger feelings of hatred towards each other than
any of the other ethnic groups in the former Yugoslavia (Interview 1 IC
Azerbaijan, cf. Interview 29 IC BiH). Such lack of positive pre-war relations
means that there is less confidence to rebuild once the conflict is over. Thus,
even if the conflict is not the result of age old hatreds, the enmity created
during the conflict cannot easily be overcome with reference to the good
relations that existed before the conflict. For some, this means that young
people are more skeptical to reconciliation than older generations who have
still memories of coexistence (Interview 22 de facto Gov’t NagornoKarabakh). Others believe that it would be easier for young people to
reconcile since they do not remember the conflict as vividly as their parents
(Interview 3 IC Azerbaijan).47
An interesting experience was a meeting I arranged at the State
Committee for Refugees and IDPs.48 I had asked for the opportunity to
conduct an interview about the problem of displacement in Azerbaijan and
how the resolution of that problem would be related to the resolution of the
political conflict regarding Nagorno-Karabakh. What I got was a history
lesson, the main point of which was that the Armenians are latecomers to the
South Caucasus and therefore have less right to be there (Interview 5 Gov’t
Azerbaijan).49 While it is not surprising to find such sentiments among
people in conflict areas, it is disheartening that a high level representative of
the State Committee for Refugees and IDPs—in a meeting with a researcher
interested in the issue of displacement—chooses to engage in political
demagoguery rather than take the opportunity to discuss the situation of
displaced persons in Azerbaijan.
Still, the issue of territorial control turned out to be critically important
not only to this particular government representative, but—somewhat
surprising to me—also to civil society activists. As described briefly above,
there are NGOs that work for dialogue and cooperation, arrange workshops
about conflict resolution, and visit each other in Armenia and Azerbaijan.
Often these activities have a regional development perspective and include
Georgian NGOs and use Georgia as a neutral venue. Some of my
47 My personal impression is that the former view is more probable. Older generations may have both good
and bad memories, but young people have little else than stereotypes and nationalist rhetoric.
48 Also known as the Department for Problems of Refugees, IDPs, Migration and Affairs with International
Organizations.
49 Incidentally, the view that who was here first has the better right, which was the basis for this entire
speech, stands in sharp contrast to the argument that “even an unambiguous answer to the question of who
first entered the now disputed territory could not seriously affect its legal status,” which was the basis for the
Azerbaijani motion to the UN regarding Armenian attempts to change the demographic composition of the
territories, referred to above (UN 2004).
74
CASES
interviewees told me about friends on the other side and expressed sincere
appreciation and respect for them. They favored return of displaced persons
and coexistence among Armenians and Azerbaijanis, and they were more
than willing to offer far-reaching autonomy to the other group. Despite all
this, their view was that ultimate sovereignty over Nagorno-Karabakh must
rest with “us,” this is absolutely critical.
One Azerbaijani NGO representative expressed what appears to be a
common opinion in Azerbaijan when she explained that Azerbaijani IDPs do
not have a problem with even the highest possible level of autonomy for
Nagorno-Karabakh. They would still be ready to return and live in peace
with the Armenians. But autonomy means autonomy and not independence.
It is absolutely essential that Nagorno-Karabakh remain Azerbaijani territory
in any solution (Interview 10 NGO Azerbaijan). Azerbaijan is ready to offer
the Karabakh Armenians their own money, president, whatever, but
Nagorno-Karabakh is Azerbaijani territory. It belongs to Azerbaijan
(Interview 6 NGO Azerbaijan). An Armenian and an Azerbaijani NGO
representative among my interviewees work closely and have a close and
respectful relation in their work on reconciliation and coexistence—but as far
as the future status of Nagorno-Karabakh is concerned, neither is willing to
compromise.
According to Lynch, the importance of the status issue is a legacy of the
Soviet era, where “group rights had to be territorialized to mean anything.”
De jure, the republics/regions/areas had autonomy, but real power resided elsewhere,
with the union republics and Moscow. The legacy from this was paradoxical.
Structures of autonomy supported the territorialization of ethnicity. However, the
experience of autonomy was often negative for members of the titular nationality in
the autonomous structure, who were well aware that power lay elsewhere. It was also
negative for the titular nationality in the union republic in which the autonomy was
embedded, who saw it as a means of Soviet/Russian “divide and rule” (Lynch 2004,
23–24).
These sentiments are stronger on the Azerbaijani side, largely because
Armenia won the war. It is easier to defend the prevailing situation, as
Armenians and particularly Karabakh Armenians do, than to try to change it,
as Azerbaijanis do (Interview 17 NGO Arm). Several interviewees explained
that the Karabakh issue is not very high on the political agenda in Armenia
today (Interviews 18 IC Armenia, 19 IC Armenia, 20 NGO Armenia).
Particularly after the 2008 elections the dispute between the government
and the opposition are the focus of attention for Armenians (Interview 15
NGO Arm). Also, the Armenian refugees from Azerbaijan are being
integrated into Armenian society, their situation is being resolved, but the
Azerbaijani IDPs remain displaced (Interview 15 NGO Arm). “In Armenia,”
explained one interviewee, “the question of Karabakh is not in your face”
(Interview 17 NGO Arm). This might be an indication that to a certain extent
75
PEACE BY REPATRIATION
peace depends on the displacement situation being resolved, though not
necessarily through repatriation—what I referred to as the hard version of
Adelman’s Soft I position (p 16). As far as the refugees and the Armenian
IDPs are concerned, this seems to be the case.
Representatives of civil society, in both Armenia and Azerbaijan, made it
clear that civil society performs important work in preparing for
compromise, reconciliation and coexistence, but that peacemaking—
negotiating a solution to the political conflict over Nagorno-Karabakh—is the
prerogative of the government, of the president or of the foreign minister.
Somewhat ironically, however, those same interviewees made it very clear
that as far as the final status of Nagorno-Karabakh is concerned there is no
possibility for compromise—”it is our territory.”
Perceptions of peace and displacement in relation to the conflict over
Nagorno-Karabakh differ markedly from those regarding the conflict in
Bosnia. Local integration is considered the best solution for the five hundred
thousand refugees in Armenia and Azerbaijan. How their displacement
situation is addressed is not perceived as having an impact on the prospects
for peace, or on the sustainability of any future peace. Resolving the situation
of Armenian IDPs is also seen as unconnected to the peace efforts. The
perception of the Azerbaijani IDPs is basically the same—how and whether
their situation is resolved does not constitute a condition for a resolution of
the political conflict over Nagorno-Karabakh.
What sets the Azerbaijani IDPs apart from the other categories of
displaced persons is that, far more than for the other displaced persons, the
solution to their situation depends on peace. The Azerbaijani government
has always emphasized that return is the only possible durable solution for
the IDPs from Nagorno-Karabakh and the occupied territories. According to
this view, accepting local integration as a durable solution means giving up
the claim to the territories.
The international community also supports the right of return for
Azerbaijani IDPs, but because return is impossible in the absence of a peace
agreement the international community is more open to alternative durable
solutions, including local integration. Several interviewees confirmed this.
On the one hand, while local integration is clearly seen as a durable solution
by many in the international community, this cannot be stated officially. The
sensitivity of the issue means that any talk of durable solutions at this time is
impossible. On the other hand, return remains the ideal solution also to the
international community (Interviews 1 IC Azerbaijan, 2 IC Azerbaijan).
On the possibility of living together, the time perspective is probably
important. It is a bit more common on the Armenian side to say that
reconciliation and coexistence is impossible at this time and for the
foreseeable future, while Azerbaijanis more often argue that once we find a
solution to the political conflict we will be able to coexist once again. The de
76
CASES
facto authorities in the NKR explained that Azerbaijanis are welcome to
return to the NKR. They will become citizens, with the same rights and
obligations as other citizens; they have nothing to fear. However, they can
live in areas where there were many Azerbaijanis before the war, such as
Fizuli and Agdam, “but we don’t want them in Stepanakert or Shoushi”
(Interview 21 de facto Gov’t Nagorno-Karabakh).
77
PEACE BY REPATRIATION
6 |Case Study Analysis
True, sustainable peace and recovery are necessary to allow refugee returns, but refugee
returns are every bit as essential to sustained peace and recovery.
Antonio Guterres50
6.1 Summary Results of the Case Studies
Most policy documents on refugee repatriation to and within BosniaHerzegovina are based on the assumption that repatriation is a necessary
condition for peace (as reflected in the epigraph to chapter 4). Further, the
main motive for emphasizing return is the perceived need to reverse what
can be reversed of the ethnic cleansing perpetrated during the war.
Interviewees qualified this general impression in some ways. One
government employee claimed that “return is not the key to peace; it is a part
of peace” (Interview 25 Gov’t BiH). An NGO representative argued that the
relationship between peace and return is reciprocal, that they affect each
other (Interview 26 NGO BiH), and a representative of the international
community suggested that the theory about repatriation as a condition for
peace be reformulated into a claim that durable solutions are important for
peace: “It’s hard to imagine how peace can happen without durable solutions
for the people, but not necessarily return so long after” (Interview 30 IC
BiH). Nevertheless, the predominant view has been—for more than fifteen
years—that without the return of displaced persons there cannot be peace in
Bosnia, because peace depends on the reversal of ethnic cleansing and the
recreation of a multiethnic society.
In the case of Nagorno-Karabakh, in principle the international
community supports the right of return for all Azerbaijani IDPs (as well as
for all other groups of displaced persons in Armenia and Azerbaijan). In
practice other solutions are encouraged because return is not a realistic
option at this time. More to the point, return is not considered a factor that
would contribute to peace even in the sense of improving the “quality” of
peace once it is achieved. The general perception is that return to NagornoKarabakh and the occupied territories will be a sign that peace is in place,
and it will put an end to displacement for many people.
6.2 Implications for the Peace-by-Repatriation Framework
As mentioned in chapter 3, there are both differences and similarities
between the two cases, which may affect the perceptions of repatriation as a
condition for peace. One important difference already noted is that a peace
50 UN (2006, 4).
78
CASES
agreement ended the conflict in Bosnia only a few years after people were
displaced, whereas there is still only a ceasefire between Armenia and
Azerbaijan more than fifteen years after displacement occurred. Clearly, the
Dayton Peace Agreement (DPA)—and the international presence backing it
up—has created conditions in Bosnia that support perceptions of
repatriation, peace, and the relationship between them, that remain
inconceivable in the case of Nagorno-Karabakh. In light of this dissimilarity,
it is not surprising that the peace-by-repatriation thesis has been more
explicitly pronounced in Bosnia than in Armenia and Azerbaijan.
I want to repeat that the purpose of the case studies was not to explain
the outcome of the conflicts, how the two displacement situations have been
addressed, or even to explain different perceptions of the relationship
between repatriation and peace in Bosnia and Nagorno-Karabakh. The
purpose was to refine the preliminary analytical framework developed in
chapter 2, the “peace-by-repatriation framework.” Accordingly, the rest of
this chapter asks whether and how that framework can be informed by the
case studies.
Sustainable Peace
Chapter 2 concluded that the peace-by-repatriation thesis refers to the postconflict period. Repatriation is not seen as a condition for the warring parties
to stop shooting at each other, but rather as a condition for successful
peacebuilding (primarily) in the aftermath of conflict. The case studies
present no reason to alter that conclusion. Indeed, a government
representative in Azerbaijan argued that if return was to take place before a
political settlement was at least agreed, IDPs would become a force for
conflict, because they would be unprepared and confronted with
underdevelopment; “you have to finish one process before you start the next”
(Interview 13 Gov’t Azerbaijan).
The implication of this is that an analysis of the peace-by-repatriation
thesis should not be concerned primarily with peace as the absence of war.
Instead, the concept of peace should be sensitive to various ambitions of
peacebuilding, still without deviating too much from how the concept has
been used is previous research. This is discussed in further detail in
chapter 8.
Repatriation
In terms of Adelman’s positions on the relationship between repatriation
and peace, Bosnia is the epitome of the unidirectional Hard I position, which
portrays repatriation as a necessary condition for peace. In NagornoKarabakh, on the other hand, the general perception is that the repatriation
of Azerbaijani IDPs will be a sign that peace is in place. This is the Soft II
position.
79
PEACE BY REPATRIATION
Regarding the explanations for how repatriation contributes to peace,
the need to reverse ethnic cleansing tops the list in Bosnia, while in the case
of Nagorno-Karabakh repatriation will more likely be a manifestation and a
legitimization of the post-conflict political order, once there is such an order.
Some of the literature surveyed in chapter 2 is clear about the need to
remix ethnic groups separated by conflict (cf. below), but it mostly refers
only to refugee repatriation without further specification. The case study of
Bosnia highlighted the difference between minority return and relocation,
the former referring to the return of displaced persons to their former homes
in areas where they are now in the numerical minority, the latter signifying
displaced persons—including refugees returning from abroad—finding
solutions not in their places of origin, but more often in areas where they
belong to the ethnic majority. The significance of this difference is
contingent on what motivates the need for repatriation. If the main reason
for promoting return is that the conflict will not be seen as ended as long as a
significant portion of a society’s population is displaced, durable solutions
need to be found, whether in the place of origin or elsewhere. In such cases,
relocation is quite sufficient. Conversely, if peace is seen as being dependent
on the recreation of a multiethnic society with an intermixed population,
minority return is required.
It is also worth underlining that the literature surveyed in chapter 2
nowhere implies that the importance of repatriation for peace diminishes
over time. Nor do the opponents of repatriation argue that the dangers of
repatriation are only temporary. Neither, however, does the literature
explicitly preclude that peace can eventually be possible after a long time
even in the absence of repatriation. Previous research simply does not
specify whether there is an expiration date on the claims made by the peaceby-repatriation thesis. Nevertheless, the fact that repatriation is still
considered essential to peace in Bosnia fifteen years after the end of armed
conflict means that the claims are supposed to hold at least that long. While
this does not imply that the peace-by-repatriation thesis should be presumed
to be valid indefinitely, it shows that the duration of the Azerbaijani IDP
situation does not by itself undermine the applicability of the thesis to the
case of Nagorno-Karabakh.
Equally important, the peace-by-repatriation thesis is not limited to the
post-agreement phase of a conflict. Consider again the case of Bosnia.
Repatriation was more or less explicitly declared the solution to
displacement from and within Bosnia before the end of active warfare, as
indicated inter alia by the UNHCR’s 1992 call for temporary protection and
the 1995 pilot project on minority return, which was initiated prior to the
signing of the DPA. In fact, at the London Conference in August 1992, where
the ICFY was established, acting U.S. Secretary of State Lawrence
Eagleburger made the following statement:
80
CASES
Let me say, parenthetically, that we are aware of the risk that humanitarian assistance
could, if we are not careful, help consolidate the land-grab in Bosnia and the political
cantonization which the United States categorically opposes. Therefore, we believe it is
not too soon for the international community to begin addressing the issue of how we
will assist refugees to return to—and to rebuild—what is left of their homes and
villages. This is an issue which will have to be part of any political settlement of the
present crisis which obtains the support of the United States, and it will require
another substantial infusion of the international assistance (Ramcharan 1997, 118).
The (at least informal) decision that the displacement situation resulting
from the war in Bosnia should be resolved as far as possible through
repatriation was clearly taken long before the proximity talks in Ohio.
Ended Displacement Situation
In the case of Bosnia, ending displacement was not sufficient. Displacement
should end through repatriation. Fifteen years after the DPA, a revised
strategy for Annex 7 implementation calls for de facto support to local
integration as a solution for the most vulnerable, but this is an exception to
the time-honored rule.
In the case of Nagorno-Karabakh, most refugees have found, or are in
the process of finding, durable solutions in the form of local integration. For
the Azerbaijani IDPs, repatriation is the preferred solution in theory, but in
the absence of a peace agreement local integration is becoming a de facto
durable solution, a process encouraged off the record by the international
community in Azerbaijan. However, as argued above, finding durable
solutions for the Azerbaijani IDPs may end their displacement, but it is not
seen as contributing to peace between Armenia and Azerbaijan.
Partition
Neither of the two cases resulted in de jure partition and the establishment
of a new, independent state (barring the establishment of Bosnia itself as
part of the dissolution of SFRY). However, there are degrees of de facto
partition in both cases—and of de jure partition in the case of the Bosnian
entities. (The idea of degrees of partition is discussed further in chapter 8.)
The case studies are relevant to the partition component in the peace-byrepatriation framework as it relates to the component of ethnic structure
(rather than in terms of its relationship to peace).
In Bosnia, the establishment by the DPA of the entities—the Federation
and Republika Srpska—confirmed in principle the territorial division of
Bosnia generated by the war-time efforts to create and control ethnically
homogeneous areas. Those same war-time efforts were categorically rejected
by Annex 7 and the prominence given to return of displaced persons since
the end of armed conflict. In this way, the Bosnian case undermines, at least
in theory, the possibility of combining partition and unmixing as having the
81
PEACE BY REPATRIATION
same properties (which Laitin rightly criticizes Kaufmann for doing, cf.
chapter 2). Rather, the two concepts should be analyzed separately.
The lessons from the case of Nagorno-Karabakh are more equivocal. On
the one hand, the present situation of de facto partition manifestly prevents
the return of Azerbaijani IDPs, meaning that territorial partition equals
ethnic unmixing. On the other hand, Armenians and Azerbaijanis were not
intermixed before the war in what is today the de facto NKR, and even a
reintegration of the NKAO and the occupied territories into Azerbaijan
followed by large-scale return of displaced persons is unlikely to result in any
substantial degree of ethnic intermixing. Territorial unity, in other words,
does not equal ethnic mixing. In the end, both case studies caution against
using partition as a proxy for ethnic unmixing, and vice versa.
Ethnic Structure
The principal argument for peace by repatriation in Bosnia concerns the
need to undo the effects of the war, most importantly to reverse ethnic
cleansing. To a great extent this argument is backward-looking, focusing on a
need to redress past wrongs, which is clearly an important part of the
establishment of sustainable peace (the term backward-looking is not
intended to diminish the importance of redressing past wrongs). What is
usually not part of this argument in policy documents, but which was
emphasized by several interlocutors, is how repatriation is a prerequisite for
encounters, and hence for reconciliation. Reconciliation requires that people
meet. In Lederach’s words,
reconciliation must be proactive in seeking to create an encounter where people can
focus on their relationship and share their perceptions, feelings, and experiences with
one another, with the goal of creating new perceptions and a new shared experience
(Lederach 1997, 30).
Kaufmann’s argument that ethnic groups be kept apart after ethnic conflict
in order to avoid a repetition of the ethnic security dilemma was challenged
by interviewees. They argued that after ethnic conflict and ethnic cleansing
people are afraid and suspicious of each other irrespective of whether they
are kept apart of not. While separation can conceivably reduce the risk of
pre-emptive action initiating a new spiral of violence, it also effectively
prevents reconciliation.
Such observations were more common in Bosnia than in Armenia and
Azerbaijan, and this adds to the equation a factor which is not part of the
theoretical underpinnings of the peace-by-repatriation thesis, namely prewar interethnic relations. In the case of Nagorno-Karabakh, there was
limited interaction between Armenians and Azerbaijanis before the war.
There was a firm Armenian majority in the NKAO, and an overwhelming
Azerbaijani dominance in what are now the occupied territories. Armenians
82
CASES
and Azerbaijanis lived side by side rather than together. In Bosnia, on the
other hand, there was a lot of interaction among the ethnic groups (as far as
ethnicity was even relevant before the war), many identified themselves as
Yugoslavs rather than Bosniaks, Croats, or Serbs and some still do.
Intermarriages were common. The point is that where pre-war interethnic
relations were good, such as in Bosnia, there is something to build on, a
starting point for reconciliation. Conversely, where pre-war relations were
bad, or perhaps, rather, weak, as in Nagorno-Karabakh, there is less of a
foundation for post-conflict reconciliation between ethnic groups.
According to the Kaufmann thesis, the experience of ethnic war should
by itself be enough to render future coexistence next to impossible
irrespective of the quality of pre-war interethnic relations. However, though
Bosnians have far from overcome the experience of the war, they have
demonstrated that coexistence is possible. Importantly, according to several
interlocutors this has been possible because people remember what things
were like before the war. This line of reasoning suggests that pre-war
interethnic relations are not at all irrelevant, as Kaufmann would have it; on
the contrary, they could be decisive.
Obviously, this has implications for the peace-by-repatriation framework. If repatriation is necessary for reconciliation through encounters, and
if the prospects for successful reconciliation depends in part on pre-war
interethnic relations, then pre-war interethnic relations will affect the
usefulness of repatriation for the purposes of reconciliation. Neither the
strong focus on repatriation in the Bosnian case nor the absence of such a
focus in Nagorno-Karabakh has been motivated with reference to this
argument, but the difference between the cases is in line with it. Assuming
that pre-war interethnic relations are relevant, return has a far more
important role to play in the establishment of sustainable peace in Bosnia
than it does in Nagorno-Karabakh.
Importantly, there are at least two dimensions to the concept of ethnic
structure. One concerns the heterogeneity or fractionalization of a country. It
usually reflects the number and relative size of different ethnic groups in a
country, but is insensitive to the geographical dispersion of these groups.
The other is the particular distribution of ethnic groups, the extent to which
they live separately or intermixed. While the second one is more to the point
in relation to the peace-by-repatriation thesis, the first one is the more
common in social research. Both will be employed in the systematic
comparative analysis in chapter 10, and they are discussed in further detail
in relation to the operationalization of key concepts in chapter 8.
83
PEACE BY REPATRIATION
6.3 Conclusions
We are now halfway through the study, and a preliminary analytical
framework has been developed on the basis of previous research and refined
through case studies of Bosnia-Herzegovina and Nagorno-Karabakh.
The framework suggests that an analysis of the peace-by-repatriation
thesis should examine how peace after armed conflict is related to (1) the
repatriation of displaced persons, (2) the end of the displacement situation,
irrespective of how this is achieved, (3) the ethnic fractionalization of the
country concerned, (4) the geographic concentration of the ethnic group
concerned, and/or (5) the partition of the country. This is the framework
applied in the test of the of the peace-by-repatriation thesis presented in the
following chapters.
In chapter 7 I describe the methodological approach. I emphasize that
the peace-by-repatriation thesis makes a set-theoretic claim and that this has
important implications for how the thesis can be tested. In chapter 8 I
operationalize the concepts of the analytical framework (the outcome and
the five conditions) as fuzzy sets. I describe in some detail how I approached
this task. Cases are selected and coded in chapter 9. The actual fuzzy-set
analysis is conducted in chapter 10. Chapter 11 summarizes the study and
presents some conclusions.
84
CONDITIONS
7 |A Comparative Methodology
Perhaps every science must start with metaphor and end with algebra; and perhaps
without the metaphor there would never have been any algebra.
Max Black51
7.1 A Set-Theoretic Argument
An important premise of the analysis presented in the next few chapters, and
a rationale for the current chapter, is that the claim made by the peace-byrepatriation thesis is a set-theoretic argument, and therefore should be
tested as such. More specifically, I use fuzzy-set analysis to conduct this test.
Because most readers are probably not very familiar with fuzzy-set analysis,
this chapter presents a few basic features of set theory, necessary conditions,
and fuzzy sets.
The first part of the chapter describes a few relevant differences between
qualitative, comparative, and quantitative approaches to social science. The
purpose of this description is to argue that the different approaches should
not be construed as competing ways of answering the same type of
questions; rather, they often have different aims, and they pose (and answer)
different questions about social phenomena. Next, the comparative, or
diversity-oriented, approach is outlined in some detail, including the
dichotomous analysis of necessary conditions and Boolean algebra as used in
Quantitative Comparative Analysis (QCA). In the final part of the chapter, I
provide a brief introduction to fuzzy sets and how to conduct fuzzy-set
analysis.
There is obviously a lot more to these themes than could possibly fit into
this chapter, and I do not claim to offer a comprehensive introduction to any
of them. Indeed, several concepts that are important to fully grasping fuzzyset analysis are not mentioned in this overview. My ambition is limited to
providing the reader with enough insight into these issues to be able to
understand what I do in chapters 8–10 and why I do so.
My view on set theory and necessary conditions is strongly inspired by
Ragin, (such as Ragin 1987; 2000; 2003; 2006b; 2008a; Rihoux & Ragin
2009) and Goertz (such as Goertz 2006a; Goertz 2006b; Goertz & Levy
2007b; Goertz & Starr 2003). They argue that a large number of the
questions that contemporary social scientists try to answer are, in fact,
questions about set relations and necessary (or sufficient) conditions (see
e.g. Goertz 2003, 76–94). Despite this fact, these questions are rarely
analyzed as questions about set relations or necessary conditions. Instead,
51 Black (1962, 242).
85
PEACE BY REPATRIATION
they are habitually reformulated as questions about correlations and
probabilities. According to Ragin (2008a, 15) this is one of the most common
mistakes in contemporary social science.
The Comparative Approach
A distinction is often made between qualitative and quantitative research.
Most commonly, this distinction is based on the number of cases studied, as
illustrated by the labels small-N and large-N research. Almost as common a
marker is the number of variables, with qualitative studies being described
as embracing many variables in the few cases they contain, and quantitative
studies focusing on a limited number of variables in a large number of cases.
Much social science research can be fit into either of these categories,
although there are significant intra-category variations.
While most research is comparative in one way or other, there is also
what is referred to as the comparative approach, or the comparative
method. As regards numbers of cases and variables, the comparative
approach falls somewhere between the qualitative and the quantitative
traditions, analyzing a medium number of variables in a medium number of
cases. More important, however, than the number of cases and variables, are
the goals of the different approaches. According to Ragin (1994), we use
qualitative methods to study commonalities, comparative methods to study
diversity, and quantitative methods to study covariation.
Discussions of what characterizes the comparative approach often
devote more attention to how it differs from quantitative research than they
do to what sets it apart from qualitative research. This is probably because
the methodological standards for good quantitative research are regularly
presented by their proponents in the quantitative camp as being equally
valid for qualitative and comparative research, an argument most famously
made by King, Keohane and Verba in Designing Social Inquiry (1994). To a
significant extent, however, qualitative and comparative research adhere to a
common set of standards that sets them apart from quantitative analysis—
but there are differences between qualitative and comparative research, too.
Looking first at how the comparative approach differs from quantitative
research, it is important to realize the difference between a set-theoretic
argument and a correlational argument. As an illustration, consider the
relationship between democracy and freedom. Using a correlational
approach we could pursue the argument that the more democratic a country,
the freer it will be. In order to test that argument we would give each country
a value on the variable “democracy” and a value on the variable “freedom,”
and then analyze our data to see whether the variables correlate, that is,
whether cases with a high value on “democracy” also have a high value on
“freedom” and vice versa.
86
CONDITIONS
Conversely, a set-theoretic approach involves a reformulation of the
argument into, for example, a claim that democracy is a necessary condition
for freedom. To test that argument we would classify each country as either
belonging or not belonging to the set of “democratic countries” and as either
belonging or not belonging to the set of “free countries.” We would then see
whether all free countries are in fact also democratic.
A few points need to be stressed. First, a set-theoretic approach does not
mean that a correlational argument is tested by unorthodox means. Rather,
it means that a substantially different argument is made. A correlational
argument should be tested as a correlational argument, and a set-theoretic
argument should be tested as a set-theoretic argument. In order to approach
a scientific problem from different angles, we can obviously reformulate a
correlational argument as a set-theoretic one or vice versa—as long as we are
aware that this entails testing a different argument than the one originally
made.
Second, not all cases are equally relevant to both types of arguments.
Regarding the arguments about democracy and freedom, all countries are
relevant to the test of the correlational argument, because the correlational
argument is an argument about all countries. Countries with a high value on
one variable should have a high value on the other, and countries with a low
value on one variable should have a low value on the other. A country with a
high value on democracy and a low value on freedom is problematic to the
correlational argument that the more democratic a country, the freer it will
be.
However, in the set-theoretic approach, such a country (being a member
of the set of “democratic countries” but not of the set of “free countries”)
does not undermine the argument that democracy is a necessary condition
for freedom. The set-theoretic argument is not an argument about all
countries. It is an argument about the members of the set of “free countries,”
and the claim it makes is that all these countries are also members of the set
of “democratic countries.” It does not make the opposite claim, that all
members of the set of “democratic countries” are also members of the set of
“free countries.”
In other words, necessary causation
is not a claim about the frequency of the cause, the frequency of the outcome, or even
the frequency of the outcome given the cause. It is only a claim that the outcome does
not occur without the cause (Seawright 2002, 183).
The differences between comparative and qualitative research are less
striking; a method like structured, focused comparison illustrates the overlap
between the two approaches. There are differences however. To continue
with the example of democracy and freedom, qualitative research would
87
PEACE BY REPATRIATION
probably be more interested in clarifying the relationship between
democracy and freedom with the help of one or a few particularly relevant
cases. A qualitative study would ask, for example, what is it about democracy
that makes freedom possible or more likely?
This cursory look at qualitative, comparative, and quantitative research
is not intended to imply that one approach is inherently better than the
other; they are good at different things. The purpose of making these points
very explicit is simply to avoid confusion about the nature and aim of the
analysis conducted in the following chapters. For more on the different
approaches see, for example, Ragin (1994, 77–153; 2000, 311–316; 2004b).
7.2 The Methodology of Necessary Conditions
The most elementary test of necessity involves one dichotomous condition
and one dichotomous outcome, each coded as either present (1) or absent
(0). To test for necessity we gather all instances of the outcome—that is, all
cases where the outcome is present—and check whether the condition is also
present in each of these cases. If the condition is present in every instance of
the outcome, the condition can be seen as necessary. In set-theoretic terms, a
condition is necessary for the outcome if the outcome constitutes a subset of
the condition. Note that cases where the outcome is absent are irrelevant to a
basic test of necessity (cf. below).
Several points should be stressed. First of all, a necessary condition can
be either constitutive or causal.52 A constitutive necessary condition is part
of the definition of a concept. Free and fair elections are a constitutive
necessary condition of most definitions of democracy. In principle, we could
use the basic test of necessity to see whether we could find any democracies
that did not have free and fair elections (although this would not make much
theoretical sense). Conversely, a minimum level of economic development
may or may not be a causal necessary condition for democracy, and it would
make theoretical sense to conduct a test of that set relation. The point is that
the test of necessity does not in itself tell us the difference between
constitutive and causal conditions; that is a task for the researcher.53 Also,
even though the term “causal necessary condition” is often used, necessary
conditions are best seen as permitting or enabling the outcome rather than
as actually bringing it about. In this study, I avoid the term causal condition.
Second, and related, the importance of conditions can vary. A necessary
condition (as well as a sufficient condition) can be more or less relevant and
more or less trivial. These concepts refer to the importance of the condition
for explaining the outcome. Explaining this notion, Goertz and Levy (2007a,
52 This corresponds to the distinction between “analytic” and “synthetic” statements, as described by Hollis &
Smith (1991, 57). See also the discussion of concept structure in chapter 8.
53 As seen in chapter 2, some researchers use repatriation as a constitutive condition for peace. This, of
course, must be avoided in the present study (cf. the operationalization of sustainable peace in chapter 8).
88
CONDITIONS
Table 7.1
A basic test of necessity (cf. Ragin 2003, 182).
Condition absent
Condition present
Outcome present
cases contradict necessity
cases support necessity
Outcome absent
cases are irrelevant
cases are irrelevant
40) argue that if 2’s were rarer than 3’s, then 2 would be a more important
cause of 6 in the expression 2*3=6. They go on to illustrate the intuition of
this argument:
A smoker lights up and there is a gas leak in his house; the result is the explosion of his
house. Lighting a match and the presence of gas due to the leak (like 2 and 3 in the
production of 6) are both causes of the explosion. Yet when asked for the cause of the
explosion people will say it was the gas leak. Gas leaks are relatively rare while the
smoker has lit thousands of matches. On an oil rig, where gas is often present, the
cause of an explosion will be the careless worker who lights a cigarette (Goertz & Levy
2007a, 40).
Third, studying cases where the outcome is present and ignoring cases where
the outcome is absent means “selecting on the dependent variable,” a mortal
sin in correlational research. However, in tests of necessity it usually makes
sense (see e.g. Dion 2003). This is so because, as noted above, all cases are
not equally relevant to a set-theoretic argument. Cases where both outcome
and condition are present support an argument about necessity, and cases
where the outcome is present but the condition is absent contradict the
argument. Cases where the outcome is absent cannot tell us anything about
the necessity of the condition (cf. table 7.1).
The fact that not all cases are equally relevant to a set-theoretic
argument is central to fuzzy-set analysis and the concepts of consistency and
coverage (explained further below). It is worth taking a closer look at this
notion.
An illustrative example is provided by Hempel (1945a), namely the settheoretic argument that “twins always resemble each other.” Stated
differently: the set of “twins” is assumed to constitute a subset of the set of
“people who resemble each other.” This means that “resembling each other”
is the necessary condition and “being twins” is the positive outcome—two
people cannot be twins unless they resemble each other. In order to test this
argument, we would gather all pairs of twins and investigate whether the
respective pairs of twins in fact do resemble each other. In this example it is
quite obvious that we would not be interested in comparing married couples,
colleagues, neighbors, or other people who are not twins, because, in
Hempel’s (1945a, 10) words, “any two persons not twins—no matter whether
89
PEACE BY REPATRIATION
Table 7.2
Twins always resemble each other.
Outcome present
Twins
Outcome absent
Non-twins
Condition absent
Two persons do not
resemble each other
cases contradict necessity
Twins who do not resemble
each other
cases are irrelevant
Non-twins who do not
resemble each other
Condition present
Two persons do resemble
each other
cases support necessity
Twins who do resemble
each other
cases are irrelevant
Non-twins who do resemble
each other
they resemble each other or not—would constitute irrelevant evidence.” This
is illustrated in table 7.2.
This means that we have to distinguish not only between positive and
negative cases, but between negative and irrelevant cases, where negative
cases are “nonpositive cases that are relevant” and irrelevant cases are
“nonpositive cases that are irrelevant” (cf. Mahoney & Goertz 2004). The
distinction is important to this study because fuzzy-set analysis does not
exclude negative cases, but it should exclude irrelevant ones.
Making the distinction between negative and irrelevant cases is
complicated by the fact that a necessary condition argument can be
reformulated in a few logically equivalent ways. First, if the presence of X is
necessary for the presence of Y, then the absence of X is sufficient for the
absence of Y. For example, if being a citizen (X) is a necessary condition for
being allowed to vote in parliamentary elections (Y), then being a non-citizen
(non-X) is sufficient for not being allowed to vote (non-Y).
Further, if the presence of X is necessary for the presence of Y, then the
absence of Y is necessary for the absence of X; if being a citizen (X) is
necessary for being allowed to vote (Y), then not being allowed to vote (nonY) is necessary for not being a citizen (non-X). This latter phrasing is rather
cryptic, and it is clearly not the way we would choose to express the
relationship, but it remains valid. Since we cannot find a non-citizen who is
allowed to vote, we must look among persons not allowed to vote if we are to
find a non-citizen. (The fact that some citizens are also not allowed to vote is
not contradicted by any of these statements.)
To illustrate how different ways of expressing the same set relation can
complicate the distinction between negative and irrelevant cases, consider
the so called raven paradox, also introduced by Hempel, in a series of
articles in Mind (Hempel 1945a; 1945b; 1946). The paradox begins with the
claim that all ravens are black, or more formally that “being black is a
necessary condition for being a raven.” In order to confirm this hypothesis,
we need to consider all ravens (all cases where the outcome is present) and
90
CONDITIONS
establish that they are all black (that the necessary condition is present in
each case).
At first glance it does not seem to make any sense to study parrots, fish,
or pieces of clothing if our aim is to verify that all ravens are black—clearly,
only ravens are relevant. However, if being black is a necessary condition for
being a raven, it follows logically that being a non-raven is necessary for
being non-black—that is, in order for any object to be of another color than
black, that object must be something else than a raven. Thus,


being black is a necessary condition for being a raven, and
being a non-raven is a necessary condition for being non-black.
Because these two statements are logically equivalent, evidence of one of
them constitutes evidence of the other. Consequently, because a green parrot
supports the second hypothesis, it also supports the first. The only thing that
can disconfirm either hypothesis remains a non-black raven; any other
object of whatever color supports both hypotheses.54
To distinguish between positive and negative cases, on the one hand, and
irrelevant cases, on the other, Mahoney and Goertz (2004) propose the
possibility principle, which states that “negative cases should be those where
the outcome has a real possibility of occurring.” To exemplify, they argue
that a researcher interested in the causes of industrialization need not bother
to study such cases as the pre-colonial Americas or contemporary Antarctica.
Because industrialization is impossible in these cases they cannot tell us
anything about the causes of industrialization. Similarly, because a parrot
cannot be a black raven, or a raven of whatever color, it cannot tell us
anything about ravens, and the parrot is therefore an irrelevant case.
However, the brown-necked raven is a negative case—it is definitely a raven,
but it can be more brown than black.
In chapter 9 I argue, with reference to this discussion, that not all
conflicts are relevant to an analysis of the peace-by-repatriation thesis, but
only those that have resulted in substantial displacement.
Set Theory and Boolean Algebra
The method of Qualitative Comparative Analysis (QCA) has been developed
to allow the systematic analysis of set-theoretic relations. Fuzzy-set analysis,
which is used in chapter 10, is based on QCA, and a brief overview of QCA
will therefore provide a good basis for the introduction of fuzzy-set analysis
in the next section. I use the conventional abbreviations cs/QCA and fs/QCA
to refer to crisp-set, or dichotomous, QCA (where cases are either in or out of
54 This discussion has inspired a lot of writing, including the article “A Green Parrot Is Just as Much a Red
Herring as a White Shoe” (French 1988).
91
PEACE BY REPATRIATION
sets) and fuzzy-set QCA (which allows for partial membership). QCA refers
to both.
The above discussion of necessary conditions was limited to one
outcome and one condition. QCA makes it possible to study more than one
condition at a time, and to analyze how several conditions can combine in
various ways to produce an outcome.
In order to do this, cs/QCA uses Boolean algebra (and fs/QCA uses fuzzy
algebra, cf. below). Two important Boolean operators are addition and
multiplication. Boolean addition, represented by a + sign, is equivalent to
the logical operator OR. The expression A + B → Y means that the outcome Y
is the result of the presence of either condition A or condition B. The
presence of either A or B is also referred to as the intersection of conditions
A and B. This is different from Boolean multiplication, represented by a *
sign, which is equivalent to the logical operator AND. Thus, A*B → Y means
that the outcome Y is the result of the combined presence of conditions A
and B, which is also referred to as the union of A and B.
Two things should be noted. First, strictly speaking A+B means A and/or
B. The logical OR means that the presence of one condition is enough, but
the presence of both is okay, while the logical AND requires the presence of
both conditions. Second, the * sign is usually implied by conditions being
placed next to each other; thus the expression A*B → Y will most often be
written as AB → Y.
A third Boolean operator is negation, which changes presence to absence
and vice versa. The negation of the crisp set of “democracies” is the crisp set
of “non-democracies”; a case which is a member (1) of the set of
“democracies” is a non-member (0) of the set of “non-democracies.” The
Table 7.3
Hypothetical dataset on ethnic rebellion: dichotomous data.
92
Case
A
B
C
Y
1
1
0
1
1
2
0
0
1
0
3
0
0
0
0
4
1
0
1
1
5
1
1
1
0
6
1
0
1
1
7
1
1
0
0
8
0
0
1
0
9
1
1
0
0
10
1
0
0
1
CONDITIONS
notation of negation varies; in this study I use upper case letters for presence
and lower case letters for absence.
To illustrate the logic of cs/QCA, I have constructed a hypothetical
dataset, which is reproduced in table 7.3. Let us assume that the example
concerns the “onset of severe ethnic rebellion” (outcome Y), and that the
three conditions are “discrimination against an ethnic group” (condition A),
“a militarily strong central power” (B), and “the ethnic group having strong
ties to a particular territorial unit” (C).
The first step of cs/QCA analysis is to create a truth table, as the one in
table 7.4. A truth table is an overview of all logical combinations of the
presence and absence of the conditions. These combinations are referred to
as configurations. Each logical configuration is represented by one row in the
table. The number of rows in a truth table is therefore always 2k, where k is
the number of conditions. In the example there are three conditions, A, B,
and C, which makes 23 = 8 rows. An output value is added to each row,
namely the value on the outcome (Y) for the respective configuration
according to the dataset in table 7.3. Often, a final column is added
indicating the number of cases displaying the particular configuration. This
illustrates the key to configurational thinking: an analysis based on this truth
table is not an analysis of three independent variables, but of eight
configurations conceived as different types of cases (Ragin 2000, 73).55
Table 7.4
Hypothetical truth table based on dichotomous data.
A
B
C
Y
N
1
1
1
0
1
1
1
0
0
2
1
0
1
1
3
1
0
0
1
1
0
1
1
?
0
0
1
0
?
0
0
0
1
0
2
0
0
0
0
1
55 Obviously, the same configuration of conditions is often found in several cases, and it is not uncommon for
these cases to display different values on the outcome. This is referred to as the problem of contradictions and
means that these particular configurations cannot easily be given a value on the outcome in the truth table.
There are ways to address contradictions (see e.g. Ragin 1987, 113–118). However, the problem of
contraditions is only relevant in cs/QCA. In fs/QCA the values in the outcome column in the truth table are
determined in a different way, as illustrated in section 7.3 below. Therefore, I do not go into detail regarding
the problem of contradictions in this overview.
93
PEACE BY REPATRIATION
The truth table gives a better overview than the data matrix of which
configurations lead to the outcome and which do not. In the example, there
are two configurations that lead to the outcome, namely 1-0-1 and 1-0-0, or,
expressed differently, AbC and Abc (with upper case letters indicating
presence, and lower case letters indicating absence). All other configurations
are associated with the absence of the outcome (except the two
configurations aBC and aBc, for which there are no empirical examples; cf.
below).
In the example, the outcome Y is associated with configurations AbC and
Abc; thus AbC → Y and Abc → Y. However, there is a simpler way to express
this, namely Ab → Y. Whenever a case displays the configuration Ab—the
presence of condition A and the absence of condition B—it also displays the
presence of the outcome. In these cases it does not matter whether condition
C is present or absent, and, therefore, condition C can be taken out of the
expression. The process of reducing the solution to as short an expression as
possible is called minimization.56
As noted, the truth table also shows that there are two configurations for
which empirical examples have not been found, 0-1-1 and 0-1-0, or aBC and
aBc. This is a result of limited diversity; the real world simply does not
contain empirical examples of all logically possible configurations. Limited
diversity means that we need to be careful when making generalizations,
because generalizations often cover configurations that lack empirical cases.
In the name of parsimony we make “simplifying assumptions” about the
likely outcome of such cases, were they to exist. Without simplifying
assumptions, we cannot generalize.
Because of limited diversity, statements about causation (in the absence of simplifying
assumptions) are necessarily restricted to the combinations of causally relevant
conditions that actually exist (Ragin 1987, 105).
Limited diversity is a quality of the empirical world that social scientists
study, not of set-theoretic analysis. The problem as such is equally relevant
to case studies or correlational research, though it obviously becomes more
important when ambitions to generalize are high (cf. Ragin 1987, 104–113;
2000, 76–87, 198–202).
In QCA, configurations that lack empirical referents are referred to as
“logical remainders.” Cases displaying these configurations could logically
exist, but in reality they do not—at least not within the population covered by
our investigation. However, they might exist outside that population. For
example, they might exist outside the geographical area covered by the
study, or they might have existed during a time period before the one
56 Minimization is done automatically by the fs/QCA software, but it is useful to know what the concept
entails. For more on minimization, see Ragin (1987, 93–98).
94
CONDITIONS
covered by the study, or they could come into existence at a later stage. A
study conducted in 1995 would not have found any empirical examples of the
configuration “member of the European Union AND former member of the
Warsaw Treaty Organization.” Today, there are several.
There are a few options for the treatment of logical remainders in the
minimization process. The researcher can either (manually) assign these
configurations a value on the outcome, 1 or 0, or let the software assign them
the value which yields the simplest solution. This choice of strategy is made
by the researcher on the basis of theoretical and substantive knowledge. The
researcher must ask: if such cases were to exist, would they most likely be
associated with the presence of the outcome or with the absence of the
outcome, or is it impossible to tell?57 When the researcher decides not to
assign a particular value to logical remainders, but leaves it up to the
software, this is referred to in QCA lingo as “setting logical remainders to
‘don’t care’.” Setting logical remainders to “don’t care” yields the most
“parsimonious solution.” Assigning a particular value to them—on the basis
of theoretical and substantive knowledge—yields a “complex solution.”
Causal Complexity
Few if any social phenomena can be attributed to a single cause. The
diversity-oriented approach therefore emphasizes the possibility of causal
complexity, where there might be more than one causal path to the outcome,
and where several conditions can jointly lead to the outcome. This means
that, in contrast to much correlational research, the diversity-oriented
approach is not looking for the net effect of different independent variables
(cf. Ragin 2006a).
For illustration, let us assume that one of the logical remainders in the
example above, aBC, had also been associated with the presence of the
outcome (even though this does not make theoretical sense). That would give
us two alternative paths to the outcome, Ab and aBC; thus Ab + aBC → Y, an
expression that cannot be minimized further.58 The two different paths to
the outcome, Ab and aBC, is an example of multiple causation. Further, the
expression Ab—the presence of ethnic discrimination and the absence of a
militarily strong central power—is an example of two conditions jointly
leading to the outcome. This is referred to as conjunctural causation (which
obviously applies the expression aBC as well). In conclusion, the expression
Ab + aBC → Y is an example of multiple, conjunctural causation.59
57 In the example with ethnic rebellion, it would make most theoretical sense to assume that ethnic rebellion
would not result from the two configurations aBC and aBc, minimized to aB (the absence of ethnic
discrimination combined with a militarily strong central power).
58 In the fuzzy-set analysis in chapter 10, different paths to the outcome are referred to as ”solution terms.”
59 In the expression “multiple, conjunctural causation,” “multiple” obviously refers to the multiple paths to
the outcome and “conjunctural” refers to the fact that several conditions operate in conjunction, but even
experts sometimes confuse them (cf. Berg-Schlosser & De Meur 2002, 272).
95
PEACE BY REPATRIATION
7.3 Fuzzy Sets
QCA has improved the systematic analysis of set-theoretic relations.
However, many social science concepts do not easily lend themselves to the
“either-or division” required by conventional set theory using crisp sets. For
example, while Sweden and the United States are both clearly members of
the set of “democratic states,” and Burma and North Korea are not, many
states are usually considered less democratic than Sweden but could still be
considered members of the crisp set of “democratic countries”: India,
Argentina, Kenya, Estonia, and Haiti, to mention a few. For 2007, India has
a Polity score of +9, Argentina +8, Kenya +7, Estonia +6, and Haiti has a
score of +5.60 Are they members or non-members of the set of “democratic
states”?
The use of crisp sets requires that each state is coded as being either a
member or a non-member of the set of “democratic states”—there is no inbetween. Once all states are divided into members and non-members of the
set of “democratic states,” no distinction is made between the members of
the set. In other words, crisp sets are insensitive to differences in degree of
democracy among states coded as being members of the set of “democratic
states” (and they are of course equally insensitive to differences among nonmembers). Assuming that the five states mentioned above are all coded as
members of the set of “democratic states,” any claim about democratic states
is expected to apply equally to Sweden, Argentina, and Haiti. If an argument
about democratic states does not hold for Haiti, this is just as damaging to
the argument as if it does not hold for Sweden. Clearly, much information
about cases is lost in dichotomization.
Enter fuzzy sets. Fuzzy sets combine the need to make a qualitative
distinction between members and non-members of sets with a sensitivity to
varying degrees of membership in these sets. They make it possible to argue
that Sweden, India, Argentina, Kenya, Estonia and Haiti all belong to the set
of “democratic states” and still conduct an analysis that is sensitive to the
varying degrees of democracy in these countries. This is because fuzzy sets
allow partial membership in sets. Claims about, in this case, democracy
should then apply with full force to states that are full members of the set of
“democratic states,” but they apply to a lesser extent to states that are weak
members of the set. As in the crisp-set situation, states that are full nonmembers of the set of “democratic states” are irrelevant to an assessment of
a claim made about democratic states (see e.g. Ragin 2003, 181–182).
In fuzzy set analysis this commonsense thinking—that cases with strong
membership in a set constitute the most relevant cases for analyzing
60 The 2008 version of the Polity IV dataset (Marshall 2008) contains worldwide data on political regime
characteristics and transitions from 1800–2008. A state’s “Polity score” is a combination of its autocratic and
democratic characteristics (see Marshall & Jaggers 2009).
96
CONDITIONS
Table 7.5
Crisp versus Fuzzy Sets (from Ragin 2000, 156).
Crisp set
Three-value
fuzzy set
Five-value
fuzzy set
1.00 = fully in
1.00 = fully in
1.00 = fully in
0.75 = more in
than out
0.50 = not fully
out or fully in
0.50 = crossover;
neither in nor out
0.25 = more out
than in
0.00 = fully out
0.00 = fully out
0.00 = fully out
Seven-value
fuzzy set
“Continuous”
fuzzy set
1.00 = fully in
1.00 = fully in
0.83 = mostly but
not fully in
0.17 = mostly but
not fully out
Numerical scores
indicating that degree
of membership is
more “in” than “out”
(0.50<xi<1.00)
0.50 =crossover;
neither in nor out
Numerical scores
indicating that degree
of membership is
more “out” than “in”
(0.00<xi<0.50)
0.00 = fully out
0.00 = fully out
0.67 = more or
less in
0.50 =crossover;
neither in nor out
0.33 = more or
less out
arguments about that set—can be incorporated into the analysis through the
concepts of consistency and coverage, to which I return below (cf. Ragin
2006b, 295).
Fuzzy membership values vary from 0.0 (full non-membership) to 1.0
(full membership). The most basic type of fuzzy sets is three-value fuzzy sets.
They allow full membership (1.0), full non-membership (0.0), and the point
of maximum ambiguity (0.5), also referred to as the crossover point. Next is
a five-value fuzzy set, where one value (0.75) is inserted between full
membership and the crossover point for cases that are more in than out of
the set but not full members, and another value (0.25) is inserted between
full non-membership and the crossover point for cases that are more out
than in the set but not full non-members. In a seven-value fuzzy set, two
values are inserted between full membership and the crossover point, as well
as between full non-membership and the crossover point, et cetera. A fuzzy
set can also be continuous (cf. table 7.5).61
To be clear, the term membership value refers to the available degrees of
membership in a fuzzy set. Determining cases’ degrees of membership in sets
is referred to as assessing fuzzy membership scores. Importantly, even a
continuous fuzzy set should not be confused with a conventional variable.
According to Ragin,
the strength of fuzzy-set analysis derives from the close correspondence between the
content of theoretical concepts, on the one hand, and the assessment of fuzzy
membership scores, on the other. In order to achieve this correspondence, it is
61 It is also possible to construct four-value and six-value fuzzy sets, et cetera, which do not employ the
membership value 0.5. This has become more common in the last few years. I discuss this option further at
the end of this chapter, and argue in favor of the membership value 0.5.
97
PEACE BY REPATRIATION
important for researchers to view the measurement of fuzzy membership as a
fundamentally interpretive act, not as a mechanical exercise (Ragin 2000, 165–166).
As opposed to most correlational analyses, fuzzy-set analysis embraces the
notion of irrelevant variation. As noted above, states that are fully out of the
set of “democratic states” are irrelevant to an assessment of an argument
about democratic states. For example, Saudi Arabia, Belarus, Fiji and
Rwanda would in most accounts be fully out of the set of “democratic states,”
despite significant variations in Polity scores (-10, -7, -4, and -3 respectively).
This variation is irrelevant to their membership scores in the fuzzy set of
“democratic states”; they are all fully out of the set.
Similarly, there can be irrelevant variation among cases with full
membership in a set. For example, despite significant differences in GDP/c,
up to fifty countries could reasonably be considered full members of the
fuzzy set of “rich countries.” Conversely, in variable-oriented research, which
usually considers all variation as being relevant, the country with the highest
GDP/c will have a higher value on the variable “rich” than the country with
the second highest GDP/c, which in turn will have a higher value that the
country with the third highest GDP/c, et cetera.
A related point is that the fuzzy membership value 0.5, the crossover
point, is a conceptual midpoint of each fuzzy set, and should not be confused
with the mean of a conventional variable. In a fuzzy set it does not matter if
all cases have membership scores greater (or lower) than 0.5 (Ragin 2000,
168–169). For example, the vast majority of countries would have
membership scores lower than 0.5 in the fuzzy set of “environmentally
sustainable societies.” Conversely, an a priori requirement that a minimum
share of empirical cases have membership scores higher than 0.5 in that set
would be a serious misrepresentation of the concept of environmental
sustainability.62 In short, “variables are calibrated according to sample
means and standard deviations; fuzzy sets are calibrated according to
theoretical and substantive knowledge” (Ragin 2000, 169).
To illustrate the basic features of fuzzy set analysis (but ignoring for the
moment any theoretical and substantive knowledge), the hypothetical crisp
data from table 7.3 have been translated into fuzzy membership scores in
table 7.6. Again, the outcome (Y) is the onset of severe ethnic rebellion,
condition A is ethnic discrimination, condition B is a militarily strong central
power, and condition C is the ethnic group having strong ties to a particular
territorial unit.
62 Cf. Ragin (2003, 188–193) who assesses membership in “the set with severe IMF protest, thus ensuring
that only the clearest instances of IMF protest receive high membership scores.” As intended, only five of
thirty-five cases (14 percent) get membership scores higher than 0.5 in the fuzzy set of “severe IMF protest.”
One of the six conditions is even more asymmetric; only four of thirty-five cases (11 percent) get membership
scores higher than 0.5 in the fuzzy set of “countries that experienced political liberalization (1978–1988).”
98
CONDITIONS
Table 7.6
Hypothetical dataset on ethnic rebellion: fuzzy data.
Case
A
B
C
Y
1
1.00
0.25
0.75
0.75
2
0.25
0.25
1.00
0.25
3
0.00
0.50
0.25
0.00
4
1.00
0.25
1.00
0.75
5
0.75
0.75
1.00
0.50
6
1.00
0.00
0.75
1.00
7
0.75
0.75
0.50
0.00
8
0.25
0.25
0.75
0.25
9
0.75
1.00
0.25
0.25
10
0.75
0.25
0.75
1.00
As a first example of a fuzzy set analysis, consider a test of whether
“discrimination against a certain ethnic group” (condition A) is a necessary
condition for “severe ethnic rebellion” (outcome Y). The set-theoretic
relationship between the condition and the outcome can be illustrated in an
XY plot, as illustrated in figure 7.1.
Testing necessity with the use of fuzzy sets means testing whether the
outcome constitutes a subset of the condition, or expressed differently,
whether each case has an equal or higher membership score in the condition
than in the outcome. If this is the case, then all cases will end up on or below
the diagonal X=Y in the XY plot. When testing sufficient conditions, we want
Figure 7.1 Fuzzy set analysis of necessity, XY plot.
99
PEACE BY REPATRIATION
cases to end up on or above the diagonal X=Y. With a single exception, the
result of the analysis in the example fits this pattern, which means that,
hypothetically, ethnic discrimination is almost always a necessary condition
of severe ethnic rebellion.
The conclusion that a condition is “almost always necessary” might
sound strange. Intuitively, a necessary condition is always necessary—
otherwise it is not necessary. However, the way the result appears in the XY
plot in figure 7.1 is not uncommon in actual analyses of necessary conditions:
with a few exceptions cases are located on or below the diagonal. This means
that only in rare instances will the outcome appear without the condition
being present. In such cases, referring to the condition as “almost always
necessary” is a better description of the result of the analysis than claiming
that there is no relationship at all between the condition and the outcome.
One thing that is not directly observable from the XY plot is the number
of cases that end up at particular intersections in the plot. In the example,
the ten cases end up at eight different intersections; cases 1 and 4 have the
same membership scores in condition A and outcome Y, as do cases 2 and 8.
Where cs/QCA uses Boolean algebra, fs/QCA uses fuzzy algebra, and
employs (among others) the same three operators described above (cf. table
8.1 below). Fuzzy addition (intersection), the equivalent of the logical OR, is
computed using the maximum membership score. Consider case 1 in table
7.6. It has a membership score of 1.00 in set A (ethnic discrimination) and a
membership score of 0.25 in set B (a militarily strong central power). Its
membership score in the fuzzy set of “ethnic discrimination OR a strong
central power” is therefore 1.00. The weak membership in condition B is
compensated by the strong membership in condition A, and yields a strong
membership in the intersection of conditions A and B.
Conversely, fuzzy multiplication (union) corresponds to the logical AND
and uses the minimum, which means that a weak membership in one set
cannot be compensated by a strong membership in the other. Consequently,
case 1 has a membership score of 0.25 in the union of conditions A and B,
that is, in the fuzzy set of “ethnic discrimination AND a strong central
power.”
Negation of fuzzy membership scores is simply 1 minus the original
membership score. A membership score of 0.75 in the set of “rich countries”
(to enhance illustration of a few points) gives a membership score of 0.25 in
the set of “non-rich countries.” Note that the fuzzy set of “non-rich countries”
is not necessarily the same as the fuzzy set of “poor countries,” for several
reasons.63 First, fuzzy sets are often constructed on the basis of more than
one variable or component, and the components of the fuzzy set of “rich
countries” need not be the same as the components of the fuzzy set of “poor
63 The same applies to crisp sets.
100
CONDITIONS
countries.” Second, fuzzy sets are based more on the theoretical concept than
on case distribution along a continuous variable, such as GDP/c (cf. Ragin
2000, 163–165). Third, while the concept of “rich” has a natural opposite in
the concept of “poor,” many fuzzy sets will not have similar opposites. For
example, negating the fuzzy set of “Protestant countries” gives the fuzzy set
of “non-Protestant countries,” but this is clearly not the same as “Catholic
countries,” “Buddhist countries,” or “atheist countries.”
Consistency and Coverage
The concepts of consistency and coverage are used to measure the extent to
which a condition is necessary or sufficient for an outcome, alone or in
combination with other conditions. Consistency can be said to measure the
existence of subsets, while coverage measures their relevance. The most
straight-forward measure of consistency is the proportion of cases that
confirm the hypothesized subset relation. In a test of necessity, this
corresponds to the proportion of cases where the outcome constitutes a
subset of the condition—and vice versa for a test of sufficiency.64
For illustration, consider the Venn diagrams in figure 7.2. The black
circle represents the hypothesized superset, and the white circle represents
the hypothesized subset. Consistency corresponds to the share of the subset
that actually constitutes a subset of the superset, that is the share of the
white circle that is grey. Coverage, in turn, is the share of the superset that is
covered by the subset, that is the share of the black circle that is grey.
Consistency must be high enough (cf. below) for coverage to be a relevant
measure. In other words, it is “pointless to compute the coverage of a cause
or combination of causes that is not a consistent subset of the outcome”
(Ragin 2006b, 299).
To make the four examples in figure 7.2 more concrete, let us consider
them as being tests of four hypothetical sufficient conditions for getting a
high salary. The outcome, “getting a high salary,” is then the superset, the
black circle, and the various conditions are the white circle subsets. In the
first example the sufficient condition tested is “having a high education.”
Figure 7.2 Consistency and coverage, Venn diagrams.
64 This is sometimes referred to as the ”inclusion algorithm.”
101
PEACE BY REPATRIATION
Most (but not all) people who have a high salary have a high education (most
of the white circle is inside the black circle) This means that consistency is
high; there is a consistent subset relation. Also, most (but not all) people
with high education have high salaries (most of the black circle is covered by
the white circle). This means that coverage is high; the subset relation is
relevant.
In the second example, the sufficient condition is “having a degree from
the top five European universities.” Again, most of the white circle subset is
inside the black circle superset, meaning that consistency is high; most
people with degrees from the top five European universities have high
salaries. However, the white circle subset only covers a small part of the
black circle superset, which means that coverage is low; this particular
condition applies only to a very small minority of all those who have high
salaries. In general, this condition is rather irrelevant as a condition for
having a high salary.
The sufficient condition in the third example is “having a first name
beginning with the letter P.” Here, consistency is low: most people whose
names begin with the letter P do not have high salaries (only a small part of
the white circle is inside the black circle). This means that there is no subset
relation, and that coverage, the part of the black circle covered by the white
circle, does not really matter. Evaluating the relevance of a subset relation is
irrelevant when there is no subset relation. The same is true in the fourth
example, where “being male” is the hypothesized sufficient condition for
having a high salary. Again, most of the white circle is outside the black
circle—a majority of men do not have high salaries: consistency is low,
“being male” is not a consistent subset of “having a high salary”; coverage is,
therefore, uninteresting.
An important weakness of measuring consistency and coverage as the
proportion of cases that confirms the hypothesized subset relation is that it
treats all cases (both consistent and inconsistent cases) as equally important
to the consistency and coverage scores, even though, as argued, their
relevance to the set-theoretic argument can vary substantially (cf. above, and
Ragin 2006b, 295). To incorporate the argument that cases with strong
membership in a set constitute the most relevant cases, the consistency
measure used by the fs/QCA software is sensitive to the extent to which
consistent cases are consistent, and to the margin by which inconsistent
cases are inconsistent; and likewise for the measure of coverage. These
measures are presented in table 7.7.
The formulas in the table illustrate the close similarity between the
computing of consistency and coverage. Consistency is computed by dividing
the sum of minimum scores (fuzzy multiplication of several fuzzy conditions)
by the sum of scores in the subset; coverage is computed by dividing the
same sum of minimum scores by the sum of scores in the superset. In a test
102
CONDITIONS
Table 7.7
Formulas for computing consistency and coverage.
Measure
Subset relation
Computation
Consistency of a necessary condition
Yi ≤ Xi =
∑ (min (Xi,Yi)) / ∑ (Yi)
Coverage of a necessary condition
Yi ≤ Xi =
∑ (min (Xi,Yi)) / ∑ (Xi)
Consistency of a sufficient condition
Xi ≤ Yi =
∑ (min (Xi,Yi)) / ∑ (Xi)
Coverage of a sufficient condition
Xi ≤ Yi =
∑ (min (Xi,Yi)) / ∑ (Yi)
of necessity the outcome (Y) is the subset, and the condition (X) is the
superset. In a test of sufficiency the condition (X) is the subset, and the
outcome (Y) is the superset.
Table 7.7 shows that the consistency score of a necessary condition is
computed in exactly the same way as the coverage score of a sufficient
condition, and that the consistency score of a sufficient condition is the same
as the coverage of a necessary condition. This means, again, that the analysis
itself does not tell us whether a condition is sufficient or necessary. Instead,
the researcher decides on theoretical grounds whether to conduct a test of
necessity or a test of sufficiency.
Figure 7.3 illustrates consistency and coverage in a test of sufficiency,
where we want cases to end up on or above the diagonal. The left part of the
figure measures the consistency of the sufficient condition. Cases on or
above the diagonal are consistent with a claim of sufficiency, and cases below
the diagonal are inconsistent. For purposes of illustration, the two cases that
are inconsistent with sufficiency have the same value on the outcome. The
distance between the x-axis and the case (the unbroken line) shows the
extent to which the cases are consistent, and the distance between the
diagonal and the case (the broken line) corresponds to the extent to which
they are inconsistent. This illustrates the notion that cases with strong
Figure 7.3 Measuring consistency and coverage, XY plots (cf. Ragin 2006b, 302).
103
PEACE BY REPATRIATION
memberships are more important, because the longer the distance between
the x-axis and the diagonal, the more a case contributes to the consistency
score—irrespective of whether the case is consistent or not. The right part of
figure 7.3 measures the coverage of the same sufficient condition. Here, the
consistent cases (above the diagonal) contribute to coverage to the extent
that the horizontal lines are unbroken, and they weaken coverage to the
extent that lines are broken.
In a test of necessity, where cases should end up on or below the
diagonal, consistency would be based on the horizontal position of cases
above the diagonal, and coverage would be based on the vertical position of
cases below it.
Figure 7.3 illustrates that all cases not situated on the actual diagonal
will always undermine either consistency or coverage. The consistency score
of a subset relation decreases as cases end up further from the diagonal on
the wrong side of it, and coverage decreases as cases end up further from the
diagonal on the right side of it. Conversely, the closer to the diagonal cases
end up, the better the consistency and coverage (cf. Goertz 2006a).
As noted, a subset relation exists when large enough a part of the subset
in fact constitutes a subset of the superset. How large is large enough? There
is no definite rule. According to Ragin (2006b, 293), “with observed
consistency scores below 0.75, it becomes increasingly difficult on
substantive grounds to maintain that a subset relation exists, even a very
rough one.” Elsewhere, he recommends using a cutoff point of 0.85, arguing
that scores of 0.00–0.75 indicate substantial inconsistency (Ragin 2004a,
11). In a more recent publication (Ragin 2009, 112–115), he compares the
results of using consistency scores of 0.70 and 0.80, arguing that 0.70 is
“relatively low,” or even “very low.” In this study I use a consistency cutoff of
0.80, the default option of the fs/QCA software.
Conducting Fuzzy-Set Analysis
Finally, I will briefly discuss the use of the fs/QCA software. This will be done
using the example of severe ethnic rebellion introduced above. A fuzzy-set
analysis of all three conditions proceeds via a crisp truth table. The
hypothetical fuzzy data in table 7.6 yield the truth table shown in table 7.8.
The truth table again consists of 2k rows. In the table they are sorted
according to the number of cases (N) that display each particular
configuration. The researcher then selects a “frequency cutoff” and a
“consistency cutoff.” Selecting the frequency cutoff means deciding the
minimum number of cases that should be associated with a particular
configuration in order for that configuration to be considered relevant for the
analysis. A frequency cutoff of 1 means that all configurations that have no
cases are excluded (these are the logical remainders, italicized in the table).
104
CONDITIONS
Table 7.8
Hypothetical truth table based on fuzzy data.
A
B
C
N
Y
raw consistency
1
0
1
4
1
0.937500
0
0
1
2
0
0.400000
1
1
0
1
0
0.500000
1
1
1
1
0
0.727373
0
0
0
0
0.400000
0
1
0
0
0.500000
0
1
1
0
0.714286
1
0
0
0
0.800000
When the analysis contains a large number of cases, a higher frequency
cutoff can be selected.
The researcher also needs to select a consistency cutoff, such as 0.75,
0.80, or 0.85—a minimum consistency score required for a subset relation to
be considered consistent; the default option in the software is 0.80. When
the consistency cutoff is selected, the outcome value for each configuration is
entered into the truth table, “1” for consistent subset relations and “0” for
non-consistent subset relations, as illustrated with bold figures in the (Y)
column of table 7.8. 65
Next, the researcher decides (as in cs/QCA) whether to assign a specific
value, 0 or 1, to the logical remainders, or to set them to “don’t care.” Once
this is done, the final step of the analysis yields the solution, in this case the
parsimonious solution Ab → Y, with a consistency of 0.9412 and a coverage
of 0.8421. Table 7.9 shows what the output looks like in the software. In
Table 7.9
Truth table solution on ethnic rebellion.
Data: conflicts over government, refugees only
Model: Y = (A, B, C)
frequency cutoff: 1
consistency cutoff: 0.937500
raw coverage
A*b
0.842105
solution coverage: 0.842105
solution consistency: 0.941176
unique coverage
0.842105
consistency
0.941176
65 Often, several configurations have consistency scores above the consistency cutoff, and then they are all
given the value “1” in the outcome column. It also happens that no consistency score in the truth table is high
enough to meet the selected consistency cutoff. In such cases, there is no consistent subset relation, and it
does not make sense to go on to the final step of the analysis.
105
PEACE BY REPATRIATION
chapter 10 where a large number of tests are conducted I opt for a simpler
way of presenting the results in the text itself. (Detailed results are given in
Appendix 5.) For example, the information in table 7.9 will be summarized
thus (Formula 7.1):
A*b → Y
consistency 0.9412, coverage 0.8421
In other words, “severe ethnic rebellion” (Y) is a highly consistent and very
relevant subset of “discrimination against a certain ethnic group” (A)
combined with “the absence of a militarily strong central power” (b). This is
the same solution that was found in the crisp set analysis above.
A few things remain to be mentioned about fuzzy-set analysis. Note that
the number of cases represented in the truth table is 4 + 2 + 1 + 1 = 8, but
there were ten cases in the dataset. Also, with three five-value fuzzy sets,
there are obviously more than eight possible configurations, yet the truth
table contains only eight rows. Both discrepancies are consequences of the
procedure for creating a crisp truth table from fuzzy data.
In order to create the truth table, the membership score of each case in
each configuration is computed through fuzzy multiplication, as the
minimum of each case’s membership score in each condition. The
membership scores of case 1 (see table 7.6) was 1.00 in condition A, 0.25 in
condition B, and 0.75 in condition C. Thus membership in ABC is min(A, B,
C) = 0.25, membership in ABc is min(A, B, 1-C) = 0.25, membership in AbC
is min(A, 1-B, C) = 0.75, et cetera. Each empirical case can have partial
membership in several configurations, but it can have a membership > 0.50
in only one configuration. This is illustrated in table 7.10.
When the membership scores of all cases in all configurations have been
computed the truth table is constructed on the basis of the number of cases
with memberships > 0.50 in the various configurations. Thus, the top row in
the truth table (table 7.8) corresponds to the four cases with memberships
> 0.50 in the configuration AbC (table 7.9), the second row of the truth table
corresponds to the two cases with memberships > 0.50 in abC, et cetera.
A final point concerns the two cases that do not have memberships
> 0.50 in any configuration (cases 3 and 7). This is because they have
membership scores of exactly 0.50 in at least one of the conditions. A
membership score of 0.50 remains a membership score of 0.50 even after
negation, and, consequently, a case with a membership score of exactly 0.50
in at least one condition cannot have a minimum score higher than 0.50 in
any configuration. Therefore, such cases are not taken into account when
membership scores in configurations are translated into a truth table—
though in the final step of the analysis the consistent configurations are
tested on all cases.
106
CONDITIONS
Table 7.10
Membership in configurations.
Case
ABC
ABc
AbC
Abc
aBC
aBc
abC
abc
1
0.25
0.25
0.75
0.25
0.00
0.00
0.00
0.00
2
0.25
0.00
0.25
0.00
0.25
0.00
0.75
0.00
3
0.00
0.00
0.00
0.00
0.25
0.50
0.25
0.50
4
0.25
0.00
0.75
0.00
0.00
0.00
0.00
0.00
5
0.75
0.00
0.25
0.00
0.25
0.00
0.25
0.00
6
0.00
0.00
0.75
0.25
0.00
0.00
0.00
0.00
7
0.50
0.50
0.25
0.25
0.25
0.25
0.25
0.25
8
0.25
0.25
0.25
0.25
0.25
0.25
0.75
0.25
9
0.25
0.75
0.00
0.00
0.25
0.25
0.00
0.00
10
0.25
0.25
0.75
0.25
0.25
0.25
0.25
0.25
Because a membership score of exactly 0.50 in a single condition is
enough for such temporary exclusion, increasing the number of conditions
will likely increase the number of cases with membership scores of exactly
0.50 in at least one condition, and consequently decrease the number of
cases actually considered in the creation of the truth table. There is some
concern about this issue in the QCA community. Skeptics ask what
conclusions we can draw on the basis of cases that we know so little about
that we cannot say whether they are more in or more out of a particular fuzzy
set. However, explicit discussions of the issue are rare. While the use of the
fuzzy membership value 0.50 appears to be decreasing in practical use, I am
aware of only one example of an explicit proposal to actually stop using it
(Schneider & Wagemann 2007, 183). This could be accomplished, for
example, by using six- instead of five-value fuzzy sets, thus employing the
membership values 0.00, 0.20, 0.40, 0.60, 0.80, and 1.00. Alternatively, the
most ambiguous cases could be assigned values close to 0.50, such as 0.55 if
slightly more in and 0.45 if slightly more out.
However, excluding the 0.50 value is not without consequences. Both
alternatives described above involve force-fitting the most ambiguous cases
as either more in or more out of the set. The truth table would then be
constructed on the basis of a larger number of cases, but at the price of
altering the operationalizations of central concepts in the analysis. According
to Ragin (2000, 259),
rescalings have important substantive and theoretical consequences in fuzzy-set
analysis, in effect changing the theoretical and substantive status of cases. For
example, a case that was only “partially in” a set might become “fully in” a set, after
rescaling membership scores. As noted previously, such rescalings often call for
reconceptualizing the set that is rescaled.
107
PEACE BY REPATRIATION
Further, as observed by Cioffi-Revilla (cf. epigraph to chapter 9), “the
problem of intrinsic fuzziness in IR cannot be solved by any arbitrarily large
increase in the accuracy of our measurements.” More to the point, if we do
not actually have enough knowledge to decide whether a case is more in or
out of a set, it seems prudent to leave it out of the truth table. That way, only
cases about which we have such knowledge are used to identify
configurations that are consistent subsets of the outcome. A principal
argument against excluding the 0.50 value is that force-fitting the most
ambiguous cases into being more in or more out of a particular fuzzy set does
not improve our knowledge about these cases, it merely reduces our ability
to be explicit about their ambiguity in the analysis. Surely, this is not to be
taken lightly.
In this study I retain the fuzzy membership value 0.50.
108
CONDITIONS
Theoretical Concepts to Fuzzy
8 |From
Sets
One cannot neatly separate the ontology of a concept from the role it plays in causal
theories and explanations.
Gary Goertz66
8.1 Concept Structures
Having described the methodology of necessary conditions and fuzzy sets in
chapter 7, in this chapter I construct the fuzzy sets that will be used in the
analysis of the peace-by-repatriation thesis in chapter 10. More precisely, I
operationalize the outcome and conditions developed in chapters 2–6 as
fuzzy sets. Before proceeding to that particular exercise, however, I devote a
few pages to a discussion of what such an operationalization entails.
Basically, operationalizing a social science concept means making it
observable, often also measurable. According to Goertz (2006b) social
science concepts can be fruitfully analyzed as consisting of three levels. The
first level, the basic level in Goertz’s terminology, is the concept itself—such
as peace, war, democracy, or power. This level is used in theoretical
propositions such as “democratic states do not fight wars against each
other.” At the secondary level are the components of the concept. For
example, explains Goertz, democracy may be seen as consisting of civil
rights, competitive elections and so on. These “secondary-level dimensions
remain part of the theoretical edifice, but they are concrete enough to be
operationalized by the indicator/data level” (Goertz 2006b, 6–7). The third
level, then, is the indicator/data level, or operationalization level. This is
where we find the actual data, specific enough to observe and measure the
original concept.
As an illustration, consider the concept of democracy as used in the
Polity IV Project—which I use in the construction of the fuzzy set of
“sustainable peace.” The Polity IV Project assigns each country a democracy
score ranging from 0 to 10 and an autocracy score ranging from 0 to 10. A
combined “Polity score” is then computed by subtracting the autocracy score
from the democracy score. The Polity score can thus vary from -10 (strongly
autocratic) to +10 (strongly democratic).
The secondary level of the concept of democracy/autocracy has five
components: competitiveness of executive recruitment, openness of
executive recruitment, constraints on chief executive, regulation of
participation, and competitiveness of political participation (Marshall &
Jaggers 2009, 14–15). These components show what dimensions of
66 Goertz (2006b, 28).
109
PEACE BY REPATRIATION
democracy are important in the Polity IV Project. Other aspects of plural
democracy, such as freedom of the press or the rule of law, are not part of the
Polity IV conceptualization of democracy.
On the indicator/data level the competitiveness of executive recruitment
is operationalized into a variable consisting of eight forms, or degrees, that
such competitiveness can take, ranging from “succession by birthright” via
“informal competition within an elite” to “formal competition among
publicly supported candidates” (Marshall & Jaggers 2009, 22).
Corresponding operationalizations are made for the other secondary-level
components, allowing researchers to pin-point the level of democracy/
autocracy in the states of the world.
The relationship between the secondary-level components is an
important, but occasionally overlooked aspect of concept formation and
operationalization. Goertz (2006b) describes two principles for constructing
this kind of multidimensional concept, one based on necessary and
sufficient conditions and one based on family resemblance.
In the necessary and sufficient conditions structure, secondary-level
components are individually necessary and jointly sufficient for the presence
of the concept. An example is the definition of “armed conflict” used by the
UCDP (2010):
An armed conflict is a contested incompatibility that concerns government and/or
territory where the use of armed force between two parties, of which at least one is the
government of a state, results in at least 25 battle-related deaths in one calendar year.
The definition contains several components, each of which must be present
for a conflict to fit the definition. For example, if an armed conflict does not
result in at least 25 battle-related deaths in one calendar year, or if none of
the parties to the conflict is the government of a state, then the conflict does
not meet the criteria for inclusion in the UCDP dataset. (This definition is
used in the selection of cases in chapter 9).
Another example is Walter’s (2002, 51) conceptualization of a formal
attempt to negotiate:
A war was coded as having experienced negotiations if three criteria were met: (1) the
leaders or representatives of each of the main fighting factions met in face-to-face
talks, (2) these individuals were willing to discuss both a cease-fire and a political
solution to the war, and (3) their respective factions had the capabilities to continue
the war if talks broke down [emphasis in original].
A situation that fails to meet even one of these criteria is coded as having
experienced “no negotiation,” and cases where all three were met constitute
cases of “active negotiations.” Each criterion is individually necessary, and
they are jointly sufficient.
110
CONDITIONS
As an example of the family resemblance structure Goertz refers to Hicks
(1999), who
defines a welfare state circa 1930 as one that provides at least three of the four
following services: (1) unemployment compensation, (2) old age pensions, (3) health
insurance, or (4) workman’s compensation (Goertz 2006b, 38; cf. Hicks 1999, 67).
Here, no secondary level component of the concept is individually necessary,
but the combination of a minimum number of components is sufficient. In
the family resemblance structure, failing to meet one criterion may be
compensated by meeting a sufficient number of other criteria; this is not
possible in the necessary and sufficient condition structure.67
The Polity IV concept of democracy/autocracy is another example of the
family resemblance structure. In the Polity IV Project, “there is no ‘necessary
condition’ for characterising a political system as democratic, rather
democracy is treated as a variable” (Marshall & Jaggers 2009, 14). This
means that a low value on, for example, competitiveness of executive
recruitment can be compensated by a high value on constraints on chief
executive.
The different logics of the two concept structures can be illustrated more
formally. The necessary and sufficient condition structure operates
according to the logical AND in classic Aristotelian logic. The family
resemblance structure works according to the logical OR. The difference can
be expressed in terms of fuzzy logic and set theory as well, as in table 8.1.
The crux of the matter is the need for consistency. If the secondary-level
components are described and organized according to the family
resemblance structure, then the concept should be operationalized and
measured according to the family resemblance structure; if the concept is
Table 8.1
The logic of concept structures (from Goertz 2006, 44).
Mathematics
Classic logic
Fuzzy logic
Set theory
Necessary and Sufficient Condition
AND
minimum
intersection
Family Resemblance
OR
maximum
union
67 The necessary and sufficient condition structure of concept construction should not be confused with the
analysis of necessary and sufficient conditions. While the analysis of necessary and sufficient conditions asks
whether certain conditions are necessary or sufficient for a certain outcome, the necessary and sufficient
condition structure of concept construction rather defines certain conditions as necessary or sufficient
components of the presence of a phenomenon. It is important to note that when a concept is constructed
according to the necessary and sufficient condition structure, the particular conditions that are part of that
concept cannot be made part of an analysis of necessary and sufficient conditions for the presence of that
same concept; as argued below, in order to be able to analyze whether repatriation is a necessary condition for
sustainable peace, refugee repatriation cannot be a necessary component of the concept of sustainable peace.
111
PEACE BY REPATRIATION
constructed according to the necessary and sufficient conditions structure,
then it should be operationalized and measured in terms of necessary and
sufficient conditions. We should aim for “concept-measure consistency.”
However,
in general most concepts of democracy use the necessary and sufficient condition
theoretical structure, while most measures use the family resemblance strategy, which
as we have seen is typified by addition or the logical OR. Thus in terms of conceptmeasure consistency they do poorly (Goertz 2006b, 98).
Developing his argument further in collaboration with Mahoney, Goertz
demonstrates the usefulness of fuzzy set analysis in the testing of two-level
theories using different concept structures for different concepts. This is
done by conducting a fuzzy set analysis of Skocpol’s theory of social
revolutions, parts of which are briefly recounted here. For the full exercise
see Goertz and Mahoney (2006, 257–266).
Skocpol’s theory is succinctly summarized as: “State breakdown and
peasant revolt are individually necessary and jointly sufficient for social
revolution.” The theory contains three basic-level concepts: the outcome
social revolution and the causes state breakdown and peasant revolt. These,
in turn, consist of several secondary-level components: social revolution
consists of (1) class-based revolts from below, (2) rapid and basic
transformation of state structures, and (3) rapid and basic transformation of
class structures; state breakdown consists of (1) international pressure, (2)
dominant class leverage within the state, and (3) agrarian backwardness;
and peasant revolt consists of (1) peasant autonomy and solidarity, and (2)
landlord vulnerability (Goertz & Mahoney 2006, 249).
Goertz and Mahoney operationalize each secondary-level component as
a five-value fuzzy set and assign membership scores to the various cases in
the analysis. They then demonstrate how the fuzzy membership scores in the
secondary-level components can be aggregated into fuzzy membership
scores in the basic-level components in two different ways, depending on
concept structure.
One option is to argue that each secondary-level component is
individually necessary for the basic-level concept. If so, all three components
must be present in order for the concept to be present. This requires the use
of the logical AND, which in fuzzy logic is calculated as the minimum (cf.
table 8.1). Alternatively, each secondary-level component can be considered
individually sufficient for the basic-level concept, in which case the concept
can be present even when one or more components are absent–as long as at
least one sufficient component is present. This procedure works according to
the logical OR, calculated as the maximum in fuzzy logic.
Usually, the family resemblance structure does not mean that each
secondary-level component is individually sufficient. More often m of n
112
CONDITIONS
components must be present, as in Hicks’s definition of a welfare state cited
above. To handle such concepts, Goertz & Mahoney (2006, 243–244)
present a third aggregation procedure using the min(sum Xi, 1), which
calibrates the fuzzy membership scores in the secondary-level components to
reflect the number of components that must be present for a case to be a
member of the basic-level concept.
The details of this procedure need not concern us here. The point is that
strict applications of the minimum and maximum are not the only options;
there are “in-betweens.” And as Goertz and Dixon explain, the same applies
to concept structures. The necessary and sufficient conditions structure and
the family resemblance structure are not the only choices available. Instead,
they are situated “at opposite poles of the concept structure continuum. This
continuum consists of the degree of substitutability between secondary-level
factors” (Goertz & Dixon 2006, 154).
Goertz first constructs a fuzzy set of each secondary-level component,
then assesses the membership score of each case in these components.
Finally, he uses these membership scores to calculate the membership in the
basic-level concept, using minimum, maximum, or some alternative
procedure. In this study, the secondary-level components largely consist of
indicator/data-level measures, which means that fuzzy sets are constructed
of the basic-level concepts rather than of the secondary-level components.
The fuzzy sets are constructed not by aggregating memberships in
secondary-level components, but by formulating explicit minimum criteria
for each fuzzy membership value in each fuzzy set. In some instances, the
concept structure changes between different membership values of a single
fuzzy set.
The rationale behind this procedure is that all secondary-level
components are not equally important to the basic-level concept. In some
instances (such as the fuzzy set of “sustainable peace”), a minimum level of
one secondary-level component is necessary for any non-zero membership in
a fuzzy set, whereas higher membership scores are assessed on the basis of
additional components that are not relevant to low membership values. In
other instances (such as the fuzzy set of “repatriation”), membership values
are based almost entirely on a single component, but in its absence another
component can warrant a low, but not high, membership. In one instance
(the fuzzy set of “partition”) the necessary condition for full membership and
the respective necessary conditions for each lower membership value are
mutually exclusive. While this may seem confusing at first, things will
become clearer as the fuzzy sets are constructed in the rest of this chapter.
Each section ends with a table summarizing the respective fuzzy set.
To the best of my knowledge, these concepts have not been
operationalized as fuzzy sets before, so in this regard I am breaking new
ground. This is a challenge that needs to be taken seriously. Further, fuzzy
113
PEACE BY REPATRIATION
sets should be the result of dialogue between theory and data, and many
judgment calls are required when the fuzzy sets are constructed. For these
reasons I find it particularly important to explicate the reasoning that
underlies each fuzzy set and I do this in significantly greater detail than what
is common in studies using fuzzy sets. Alternative operationalizations are
conceivable, but to seriously question the ones I use in this study they need
to be accounted for in corresponding detail.
Six sets are constructed: the fuzzy-set outcome “sustainable peace,” and
the five fuzzy-set conditions “repatriation,” “ended displacement situation,”
“partition,” “fractionalization,” and “group concentration.” All six are
constructed as five-value fuzzy sets. There are a few studies that use sets with
different numbers of membership values. While this might be motivated in
special circumstances, I believe that the consequences it has for the analysis
outweigh the benefits, and that the same number of membership values
across all fuzzy sets should be the default option. Anyway, I intend to stick to
that option in the present study. Also, I agree with Ragin that five-value
fuzzy sets are
especially useful in situations where researchers have a substantial amount of
information about cases, but the evidence is not systematic or strictly comparable
from case to case (Ragin 2000, 157).
Finally, and related to Goertz’s statement in the epigraph to this chapter, the
operationalizations here should not deviate too much from previous
operationalizations. In one respect they will, namely that they are fuzzy sets.
However, “content-wise” they should be comparatively close to current
research in order to contribute to the accumulation of knowledge. The
concept of sustainable peace is probably the best example. It would be quite
possible to elaborate on various complex definitions of the concept,
including the meaning of democracy in different cultural settings. However,
an operationalization of that kind would not measure the same phenomenon
being measured in other research in the field (see also George & Bennett
2004, 70–71; Yin 2009, 32–33). It would therefore contribute less to our
accumulated understanding of sustainable peace.
8.2 Sustainable Peace
The most elementary way to define peace is probably as the absence of war.
Peace as the absence of war is often referred to as negative peace. It is also
possible to talk about positive peace. Where negative peace is usually defined
as the absence of direct violence, positive peace may be defined as the
absence of structural violence (see e.g. Galtung 1969; Wallensteen 2007;
Wiberg 1990). Both are examples of peace defined negatively, that is, as the
114
CONDITIONS
Table 8.2
Definitions of peace.
Negative peace
Positive peace
Negatively defined
absence of
direct violence
absence of
structural violence
Positively defined
presence of
military stability
presence of
social justice
absence of something. Peace can also be defined positively as the presence of
something, although this is less common.
While there is been peace between Sweden and Norway for more than a
century, this is not the kind of peace that is of interest in this study. Here,
focus is on peace after armed conflict, which in the field of peace and conflict
research implicitly means peace after recent armed conflict.
The conventional way of conceptualizing peace after armed conflict is as
the absence of war, that is, negatively defined negative peace. From World
War II until the end of the Cold War, the ambition of mediators and
peacemakers was to put an end to the use of armed force, and if that
succeeded then peace was achieved. The common way to measure this
success was to evaluate whether the absence of armed force became durable,
that is, if it lasted for at least a minimum amount of time. According to
Walter (2002, 53) a “five-year time frame is the standard measure of success
in the literature” (cf. Hartzell, Hoddie & Rothchild 2001; Licklider 1995;
Walter 1997).
While a focus on the initial five years after a peace settlement (whether
in the form of agreement, victory or otherwise) has been a common practice
in the literature, both shorter and longer periods have also been used.
Nordquist (1992) employs three measures of durability, one of which is
whether agreements remain operational for at least 15 years. Ohlson (1998)
requires the absence of war for at least 30 months in order to call an
agreement successful. Also, some researchers use several cutoff points. For
example, Doyle and Sambanis (2000; cf. Sambanis 2000) measure
peacebuilding success two, five, and ten years after a war. In recent years, it
has become more common not to use specific cut-off points, but simply to
measure how long a peace agreement holds (Fortna 2003; Gurses, Rost &
McLeod 2008; Hartzell & Hoddie 2003; Jarstad & Nilsson 2008; Werner
1999), though some such studies, such as Druckman & Albin (2010), limit
the continuous measure to the first five years.
The recurrence of war is the most common indicator of failure. A peace
settlement fails if war restarts within a period of five years when such a fixed
measure is used, or when war restarts if a continuous measure is used. Some
studies do not require the use of force in order to code a peace settlement as
failed. For Nordquist (1992) it is enough for the agreement to be formally
115
PEACE BY REPATRIATION
invalidated by either party, for Werner (1999) the mere threat of force
suffices. However, such measures are unusual. Also, when the measure used
is the recurrence of armed force it can be adjusted to a lower level than the
conventional definition of war. Walter (2002, 53, note 15) uses the
conventional measure and finds that violence has ended “if fewer than one
thousand war deaths occurred each year for a period of five consecutive
years.” This can be compared to the UCDP definition of armed conflict cited
above: at least 25 battle-related deaths in one calendar year. The UCDP
distinguishes between “wars” that cause at least 1,000 battle-related deaths
in one calendar year—a definition that corresponds to the convention in the
literature—and “minor armed conflicts” conflicts that cause 25–999 battlerelated deaths.
It would be possible to construct a fuzzy set of “sustainable peace” on the
basis of settlement duration. However, while it would be fairly simple to
assign fuzzy membership scores in such a set, conceptually it would be off
target in important respects. This is because, in addition to different
thresholds being used for coding if and when peace settlements fail, a broad
range of criteria can be added to the “absence of violence” standard in the
measure of success.
Fortna acknowledges that her definition of peace as merely the absence
of war has its limitations, because “not all varieties of peace are equally
desirable, nor does stability necessarily coincide with social justice” (Fortna
2003, 339). Ohlson voices similar concerns when he argues that the idea of
negative peace “suggests that highly unacceptable social orders and peace
could be compatible. This is problematic, even unsatisfactory” (Ohlson 1998,
31).
Similarly, while negative peace was often the reasonable aim of
peacekeeping during the Cold War, over the past twenty years the
international community has developed higher ambitions in areas such as
post-conflict reconstruction, transitional justice, democratization, et cetera.
There is little consensus regarding the precise meaning of and differences
between these various labels, but if any single one of them can be used as an
overarching concept it is peacebuilding.
To some researchers peacebuilding involves a whole new range of means
at the disposal of the international community, but the aim is more or less
the same as before. Cousens’s emphasis on the original purpose of
peacebuilding was referred to in chapter 2. In a similar manner, Jarstad and
Sisk (2008) question the common perception that peacebuilding and
democratization are mutually reinforcing, and argue that when these two
aims collide the fortification of peace should take priority over
democratization.
While most researchers would agree that the durable absence of war is
the most important objective of peacebuilding (if only because a return to
116
CONDITIONS
warfare will undo achievements in other areas), democracy, respect for
human rights, the rule of law, gender equality, et cetera are often placed on
the output side of the equation. Rather than being means to an end
(durability), they are goals in their own right in peacebuilding processes.
In this way the ambitions of peacebuilding have expanded beyond the
mere cessation of armed violence. This has effects on how the success and
failure of peacebuilding operations are conceptualized. A recurring theme is
that peace must be made sustainable, a term which, within this field of
research, has different connotations than durable. Paris has expressed this
in the following way:
If the test of ‘successful’ peacebuilding is simply whether large-scale conflict resumed
in the aftermath of a peacebuilding mission, then most of the operations conducted in
the 1990s were successful, because in all but three cases (Angola, Rwanda, and
Liberia), large-scale hostilities have not resumed. But if we use instead the standard of
success articulated by Kofi Annan and Boutros Boutros-Ghali—namely, the
establishment of a ‘sustainable’ peace, or a peace that will endure long after the
peacebuilders depart from the country—then the picture becomes less favorable (Paris
2004, 6).
But what is really the difference between the traditional measure of success
and failure, “whether large-scale conflict resumed in the aftermath of a
peacebuilding mission,” and the more ambitious aim of establishing “a
‘sustainable’ peace, or a peace that will endure long after the peacebuilders
depart from the country”? Instead of asking whether armed conflict resumed
within five years of the end of active hostilities, should we simply ask
whether armed conflict resumed within five years of the departure of the
peacebuilders?
Yes and no. Yes, in much of the peacebuilding literature the fundamental
criterion of success remains the durable absence of war. What is new is that
peacebuilding theory considers a broader range of activities in and aspects of
the post-conflict civil society as instrumental for that durability, although
few such activities and aspects make it through the operationalization phase
into indicators and measurements. But no, some studies have employed
measures of sustainable peace that require more than the absence of armed
conflict for a peace settlement to be deemed fully successful—as opposed to
partially successful processes, where armed conflict does not resume, but the
“quality” of the peace is nevertheless not entirely satisfactory.
Grading Peace
One implication of the discussion thus far is that the success and failure of
peacebuilding is not necessarily a question of either/or; it can also be a
matter of more or less, of varying degrees of success. Hampson, for one,
argues that “the ending of civil violence and armed conflict” should be seen
as only partial success.
117
PEACE BY REPATRIATION
The ultimate success of the peace-building process in situations of civil conflict is thus
directly related to a society’s ability to make an effective transition from a state of war
to a state of peace marked by the restoration of civil order, the reemergence of civil
society, and the establishment of participatory political institutions (Hampson 1996,
10).
Such an understanding of peacebuilding success is key in the construction of
the fuzzy set of “sustainable peace,” and it is not entirely new.
Downs and Stedman use three degrees of success in their study of the
implementation of peace settlements (Downs & Stedman 2002, 50–51). They
operationalize “mission success” on the basis of two dichotomous variables,
or rather, conditions,
(1) whether large-scale violence is brought to an end while the implementers are
present; and (2) whether the war is terminated on a self-enforcing basis so that the
implementers can go home without fear of the war rekindling.
If neither condition is fulfilled a case is coded as a failure; if condition one is
fulfilled, but not condition two, a case is coded as a partial success; and if
both conditions are fulfilled, and the war did not restart within two years, a
case is coded as a success.
In important respects, the fuzzy set of “sustainable peace” used in this
study builds on work by Doyle and Sambanis, who argue that “‘sustainable
peace’ is the measure of successful peacebuilding” (Doyle & Sambanis 2006,
4). They introduce two concepts of peace: sovereign peace and participatory
peace (2006, 73). First, sovereign peace (also referred to as lenient peace, cf.
Doyle & Sambanis 2000) is “a negative standard of peace, and refers to the
absence of large-scale violence.” Second, participatory peace (or strict peace)
“requires sovereign peace plus a minimum level of political openness,” coded
on the basis of the Polity IV index.
The concept of participatory peace is constructed according to the
necessary and sufficient condition structure. Participatory peace requires
sovereign peace plus a minimum level of political openness. A high level of
political openness cannot compensate for a failure to fulfill the “negative
standard of peace” for at least two years, the cutoff point used by Doyle and
Sambanis. This means that the same case can be coded as a success
according to the sovereign peace standard and a failure according to the
participatory peace standard, but not vice versa (cf. Doyle & Sambanis 2006,
table 3.1, pp 76–81).
An important conclusion from their earlier study is that
higher order, or democratic, peacebuilding is more successful after nonidentity wars,
after long and not very costly wars, in countries with relatively high development
levels, and when UN peace operations and substantial financial assistance are
available. Lower order peacebuilding – an end to the violence – is more dependent on
118
CONDITIONS
muscular third-party intervention and on low hostility levels rather than on the
breadth of local capacities (Doyle & Sambanis 2000, 795).
This shows that the lenient and strict types of peacebuilding are dependent
on different factors. In other words, to achieve democracy and justice after a
civil war, what is required is not simply more of whatever is associated with
an end to violence, but something qualitatively different.
Clearly, many options are open for the operationalization of sustainable
peace as a fuzzy set. Two concerns have influenced the process in this study.
First, the post-conflict context of the study should be reflected in the
construction of the fuzzy set; all degrees of membership, including full
membership, refer to situations in countries that relatively recently
experienced armed conflict. Second, the assumption under scrutiny in this
study comes from a field of research where peace is measured according to
variations on a common standard, as described above. The fuzzy set of
“sustainable peace,” while original in some ways, should not deviate too
much from these conventional measures.
As noted above, Doyle and Sambanis use a cutoff point of two years.
While it seems far-fetched to talk about sustainable peace in a situation
where large-scale conflict resumes after three years, the absence of war for at
least two years is a measure that has been used in the literature; here, it will
be referred to as “sovereign peace for at least two years” and yield a
membership score of 0.25 in the fuzzy set of “sustainable peace”—such
situations have some features of sustainable peace, but they are more out of
the set than in. Conflicts that restart within two years are considered fully
out of the set. Next, the membership value 0.50, or the cross-over point, will
be what has been described as “the standard measure of success in the
literature,” namely the absence of armed force for at least five years—here
“sovereign peace for at least five years.”
For the remaining membership values, 0.75 and 1.00, certain qualities
will be added to the five year measure, but a longer time-period will not be
employed, for two reasons. First, as Downs and Stedman (2002, 48) aptly
point out, “the passage of time is the enemy of inference.” The longer the
time period studied, the more difficult to relate the outcome to the
conditions. Secondly, Hartzell, Hoddie and Rothchild (2001, 187–188) argue
that instances of civil wars resuming after more than five years are rare. And
in fact, of the 138 conflict terminations in the UCDP termination dataset,
seventy-nine were durable according to the five year measure, and only six of
these restarted at a later date.68 This means that, although there are
exceptions, if a terminated conflict has not restarted within five years it will
usually remain terminated. Adding, say, another five years for higher
68 Ethiopia (Oromiya), 1991; Haiti, 1991; Indonesia (Aceh), 1991; U.K (Northern Ireland), 1991; Georgia
(South Ossetia), 1992; and Iraq, 1996.
119
PEACE BY REPATRIATION
Table 8.3
The fuzzy set of “sustainable peace.”
Value
1.00
0.75
0.50
0.25
0.00
Operationalization
participatory peace for at least five years, and respect for
human rights, and no peace operation
participatory peace for at least five years
sovereign peace for at least five years
sovereign peace for at least two years
conflict recurrence within two years
membership values would therefore change little in terms of results (but,
since all terminations should theoretically be given the chance to hold for ten
years, it would reduce the number of cases that could be included in the
study).
Instead of using a longer period than five years for the higher degrees of
membership, I will employ Doyle and Sambanis’s concept of participatory
peace, that is sovereign peace plus a certain level of democracy. Doyle and
Sambanis use two alternative levels of democracy: a combined Polity score of
at least +3, and a combined Polity score of at least +6. For the membership
value 0.75 in the fuzzy set of “sustainable peace,” I will require five years of
sovereign peace plus a combined Polity score of at least +6.69
For full membership in the fuzzy set of “sustainable peace,” a case should
experience participatory peace—that is, sovereign peace for at least five years
and a Polity score of at least +6 by the end of that period—plus there should
be no ongoing peace operation at the end of the five year period (also based
on Doyle & Sambanis 2006, 74) and a minimum level of respect for human
rights. To measure respect for human rights, I use the Cingranelli-Richards
(CIRI) index (Cingranelli & Richards 2004; 2007) and require at least 5 on a
combined measure standardized to vary between 0 and 10 (see Appendix 3
for details on the fuzzy set of “sustainable peace”).
Like the examples of graded peace above, this fuzzy set is based on the
necessary and sufficient condition structure. For example, a high Polity score
cannot compensate for short durability of the absence of conflict.
8.3 Repatriation
At first glance the operationalization of repatriation seems obvious:
repatriation takes place when people repatriate. People who repatriate can
be observed, even counted if necessary (though precise figures are often
69 Using the lower threshold, a Polity score of +3, changes the membership scores of no more than three
cases compared to the threshold +6. It should also be noted that the threshold +6 is not arbitrary; the Polity
IV web page (http://www.systemicpeace.org/polity/polity4.htm) proposes a three-part categorization of
states as autocracies (-10 to -6), anocracies (-5 to +5 and the three special values: -66, -77, and -88), and
democracies (+6 to +10), that is, a threshold of +6 for democracies. Further, with a threshold of +6 it does not
matter if I use the old Polity score or the new Polity2 score, so I will simply refer to “Polity scores.”
120
CONDITIONS
difficult to estimate). However, the challenge in this section is that
repatriation needs to be operationalized as a condition. It must be possible to
code cases as being cases of repatriation or cases of non-repatriation—or
more specifically, since fuzzy sets are used, we must be able to determine the
extent to which cases are cases of repatriation. The original assumption
investigated in this study is formulated in terms of conditionality and not as
a correlation, and in order to analyze whether repatriation is a necessary
condition for peace, I must be able to determine whether that condition is
present or absent in each case analyzed.
When it is assumed that repatriation is a necessary condition for peace,
what is meant by “repatriation”? As soon as the first refugee has repatriated,
repatriation has taken place to some extent, but this is clearly not what is
intended. Likewise, the assumption does not require the return of every
single displaced person to their original homes. As is so often the case, the
answer lies somewhere in between, but the literature is not very enlightening
in this regard.
While the construction of the fuzzy set of “sustainable peace” could draw
on earlier operationalizations of peace, a somewhat different approach is
needed in the construction of the fuzzy set of “repatriation.” A useful starting
point is to consider a few cases where successful repatriation processes are
often cited as having been important for peace. Cambodia, Guatemala, and
Mozambique are prominent examples. While case studies of repatriation
during peace processes do not often use the language of necessary
conditions, given the way these three cases are considered in the
peacebuilding literature they would clearly qualify as cases of repatriation.
So, what did repatriation look like in these cases?
In the case of Cambodia “the repatriation in safety and dignity of
Cambodian refugees and displaced persons [was] an integral part of the
comprehensive political settlement” of the conflict (Article 20 of the Paris
Agreement). Repatriation was explicitly related to the ability of refugees and
displaced persons to participate in elections, as provided for in the peace
agreement. “The effectiveness of the [Cambodian] repatriation operation has
been widely recognized and it has been hailed as a model of its kind”
(Eastmond & Öjendal 1999, 38–39). In 1991, the Cambodian refugee
population consisted of more than three hundred and ninety thousand
refugees. During 1992 and 1993 the UNHCR assisted more than three
hundred and sixty thousand of them to repatriate, 95 percent in time to vote
in the May 1993 elections. Several thousand more repatriated spontaneously
(that is, without UNHCR assistance), bringing the total number of returnees
to three hundred and sixty-seven thousand by the end of 1993 (USCRI 1994,
78). While generally considered very successful, the repatriation process was
tainted by Thailand’s forcible return of the few hundred Cambodians who
did not repatriate voluntarily, and it has been criticized for failing to
121
PEACE BY REPATRIATION
reintegrate the returning refugees properly (see e.g. Eastmond & Öjendal
1999).
In the case of Guatemala the active participation of refugees in planning
and implementing repatriation has contributed to the assessment of it as a
success.70 After a 1992 agreement between the Guatemalan government and
representatives of the Guatemalan refugee community in Mexico, organized
repatriation began, despite the fact that the conflict was not brought to an
end until 1996. In 1989, the number of refugees was fifty-nine thousand. By
1999, the U.S. Committee for Refugees and Immigrants71 argued that those
Guatemalans still remaining in Mexico in 1998 should no longer be
considered refugees, despite the fact that both the Mexican government and
the UNHCR still considered them as such. The USCRI argued that nothing
prevented repatriation, and that those who had not repatriated by 1998
should be considered to have voluntarily decided to resettle elsewhere. Thus,
they should no longer be considered displaced. The same was said about
IDPs (USCRI 1999, 266–267). By this time, more than thirty-one thousand
refugees had repatriated.
The Mozambican repatriation process is one of the most massive to date.
Out of a total refugee population of 1.7 million in 1992, some 1.4 million
returned between the signing of the peace agreement in 1992 and the end of
1995, in a process regarded as very successful. 3 million IDPs returned home
during the same period. In 1996 the UNHCR invoked the cessation clause for
refugees from Mozambique. The UNHCR cited the “uninterrupted peace and
stability” in Mozambique since the signing of the peace agreement, and the
successful return and reintegration of refugees, and concluded that
these developments, as well as their broad international recognition and appreciation,
are indicative of the fundamental nature and durability of the changes which have
taken place in Mozambique, and call therefore for the application of the cessation
clause to Mozambican refugees (UNHCR 1996).
The USCRI applies the same argument to remaining IDPs (USCRI 1997, 82).
Two secondary-level components of the concept “repatriation” can be
extracted from this very brief overview. First, the extent of repatriation that
has taken place is obviously important, and second, the successful
completion of an organized repatriation program should also be taken into
account.72
70 In Guatemala the terms “return” and “repatriation” took on different connotations with return implying an
active subject deciding to return home as opposed to the passive object being repatriated. This study adheres
to the conventional use of the terms as synonymous.
71 The World Refugee Survey, a yearly publication by the USCRI, has been the primary source of
displacement data in this study (cf. chapter 9).
72 Additionally, a decision by the UNHCR to invoke the cessation clause is seemingly critical in that it reveals
a belief that the problem is solved. However, that is exactly the crux of the cessation clause in the present
context: problem solved. It does not presuppose that the problem was solved through repatriation; any
122
CONDITIONS
The absolute number of refugees who have repatriated is not irrelevant,
but the share that has done so is more pertinent. If thirty thousand out of a
displaced population of forty thousand have repatriated, this is more to the
point than if fifty thousand out of two hundred thousand have repatriated.
Still, although a 75 percent repatriation rate is a better indicator of the
condition “repatriation” than a 25 percent repatriation rate, it is not obvious
what rate to require for full or partial membership in the fuzzy set of
“repatriation.” Previous research offers no guidance since repatriation has
not been operationalized as a condition in this way before. However, the
three repatriation successes can serve as a starting point.
In the Cambodian case the repatriation rate was almost 95 percent, in
the Guatemalan case the rate was 54 percent; and Mozambique achieved a
repatriation rate of 81 percent. There is a considerable difference between
slightly over 50 percent of refugees returning home to Guatemala and more
than 80 percent returning home to Mozambique and Cambodia.
Nevertheless, all three cases are usually considered successful, and should
definitely be (at least) more in than out of the fuzzy set of “repatriation.”
Therefore, in this study, full membership in the fuzzy set of
“repatriation” will require a repatriation rate of at least 80 percent. This
means that the membership scores of Cambodia and Mozambique are 1.00.
Guatemala receives a membership score of 0.75 ending up between full
membership and the crossover point, at a level which will require a
repatriation rate of at least 50 percent. With repatriation rates of less than
50 percent it becomes increasingly difficult to categorize cases as “cases of
repatriation,” and in cases where less than one-third of displaced persons
return repatriation is clearly not the most common solution. Therefore, cases
with repatriation rates between 30 and 50 percent will receive a membership
score of 0.50—the crossover point.
The other secondary-level component of the concept of repatriation, the
successful completion of an organized repatriation program, is arguably less
important than the displacement rate, most obviously because the ambitions
of organized repatriation processes may vary substantially. As an illustration,
consider the case Bosnia and Herzegovina (Croat), 1994 (part of the fuzzyset analysis in chapter 10). In this case a few thousand Bosnian Croats
repatriated from Croatia to Republika Srpska (RS) under a June 1998
agreement between Croatia and RS (USCRI 2000, 225; 2001, 211). This
program was not aiming at returning several hundred thousand Bosnian
Croats from Croatia to Bosnia, and in terms of finding durable solutions for
the entire Bosnian Croat refugee population in Croatia the impact of the
program was practically negligible. The overall repatriation rate from Croatia
durable solution will do. Consequently, the cessation clause, critical as it is, will be part of the fuzzy set of
“ended displacement situation” operationalized in the next section, but not of the fuzzy set of “repatriation.”
123
PEACE BY REPATRIATION
to Bosnia was 10–15 percent, and this particular program probably
contributed less than one-tenth of that. Yet, in the context of the situation in
former Yugoslavia (to which assessments of fuzzy membership scores should
be sensitive), an exchange program of this kind should count for something.
On the basis of such considerations the successful completion of an
organized repatriation program is used as a basis for membership scores of
0.25 in the fuzzy set of “repatriation” for cases where the repatriation rate is
lower than 30 percent. What, more precisely, constitutes a successful
completion of an organized repatriation program will have to be determined
on a case by case basis. Similarly, the three thresholds based on the
repatriation rate are not set in stone, but can be employed flexibly depending
on such factors as the size of the total displaced population, the pace of
repatriation, or whether repatriating refugees are unable to return to their
areas of origin and instead become internally displaced—which raises a final
concern before the fuzzy set of “repatriation” can be presented:
What about the distinction between refugees and internally displaced
persons (IDPs)? Should only refugees count, or both refugees and IDPs? The
assessment of cases can differ substantially depending on whether one or
both categories of displaced persons are included in the analysis. The
Democratic Republic of Congo (DRC), 2001 (also part of the fuzzy-set
analysis in chapter 10) is a case in point. The vast majority of displaced
persons in/from DRC were IDPs (more than 3 million, or 85 percent of the
total displaced population in 2003), and while more than 2 million IDPs
returned home, only a handful (relatively speaking) refugees returned. Also,
there are important differences (as well as similarities) between refugees and
IDPs regarding return and protection in relation to peace processes (Fagen
2009).
On the one hand, general statements about repatriation as a necessary
condition for peace typically allude to refugee repatriation, and do not
mention IDPs. This suggests that the analysis be limited to refugees. On the
other hand, in the case where the argument about peace being dependent on
repatriation has been pronounced more forcefully than anywhere else,
namely Bosnia-Herzegovina, a strong emphasis has been placed on minority
Table 8.4
The fuzzy set of “repatriation.”
Value
1.00
0.75
0.50
0.25
0.00
124
Operationalization
repatriation rate at least 80 percent
repatriation rate at least 50 percent
repatriation rate at least 30 percent
repatriation rate less than 30 percent, and successful repatriation program
repatriation rate less than 30 percent, and no successful repatriation program
CONDITIONS
return and the remixing of ethnic groups; a mere returning of refugees to
their country of origin has clearly not been considered good enough. This
calls for including both refugees and IDPs.
In keeping with Goertz’s proposition in the epigraph to this chapter, I
believe that the decision should depend on the role the concept repatriation
plays in causal theories and explanations. Unfortunately, the causal theories
and explanations at hand do not provide a decisive answer either way. The
argument about refugee warriors is more relevant to refugee populations
than internally displaced populations, not in principle, but because displaced
persons are more likely to be able to arm themselves when outside their
country of origin. This is in line with Lischer’s argument that governmentsin-exile are the most likely to use force—governments-in-exile are also more
likely to be abroad, whereas people fleeing generalized violence may or may
not cross an international border (cf. chapter 2). Conversely, the argument
about expressing confidence in the future of one’s country applies at least as
much to IDPs as to refugees. This is particularly evident in cases where
refugees return to their country of origin, but end up in internal
displacement because they are unable or unwilling to return to their actual
homes. Consequently, I will employ two alternative codings for the fuzzy set
of “repatriation,” one based on all displaced persons, and one based on
refugees and asylum seekers only, excluding IDPs.
In this set, a very low rate of return can be compensated by a successful
repatriation program. However, such a program cannot by itself yield a high
membership score in the set. This means that while successful repatriation
programs are not entirely irrelevant to membership scores in the fuzzy set of
“repatriation” those membership scores are nevertheless primarily assessed
on the basis of repatriation rates.
8.4 Ended Displacement Situation
The fuzzy set of “repatriation” is obviously based particularly on the durable
solution repatriation. Conversely, the fuzzy set of “ended displacement
situation” is equally compatible with the two other durable solutions, local
integration and resettlement. What matters is that the displacement
situation has ended, not how it ended.
The share of displaced persons who find a solution—any solution—is a
significant secondary-level component of the concept ended displacement
situation. However, the “any-solution-rate” is less central to the fuzzy set of
“ended displacement situation” than the repatriation rate is to the fuzzy set
of “repatriation.” This is so because theoretically, the focus of repatriation as
a condition for peace is primarily on how those who do return can contribute
to sustainable peace, but the focus of ended displacement situation as a
condition for peace is rather on how those who do not find durable solutions
125
PEACE BY REPATRIATION
(or how the fact that not all displaced persons have found durable solutions)
can hamper peace efforts.
A more important secondary-level component is therefore the absolute
number of people who have not found durable solutions by the time the
outcome is assessed. As in the case of repatriation, there is little guidance in
previous research regarding what numbers would be relevant to different
fuzzy membership values in the set of “ended displacement situation” (and,
again, a fair amount of flexibility is called for in the coding process, as the
assessment of fuzzy membership scores should reflect case knowledge).
In 2004 the UNHCR introduced the measure of “25,000 persons or
more who have been in exile for five or more years in developing countries”
as a way to identify what was termed “major protracted refugee situations”
(Loescher & Milner 2005; UNHCR Executive Committee 2004). The number
twenty-five thousand is the basis for the designation “major,” and the five
year duration makes a situation “protracted.” The rationale behind the figure
twenty-five thousand is not explicated,73 but the UNHCR’s use of it makes it
less arbitrary than most other figures as a starting point for an
operationalization of the concept ended displacement situation into a fuzzy
set. Crude as the measure of twenty-five thousand may be, a major
displacement situation is hardly an ended displacement situation.
In view of the many refugee populations that number several hundred
thousand, even millions, the measure of twenty-five thousand appears
relatively low and sets a high standard for membership in the fuzzy set of
“ended displacement situation.” If millions of displaced persons succeed in
finding durable solutions it might seem less important whether ten
thousand, twenty-five thousand or fifty thousand fail to do so. However,
once again the role the concept plays in causal theories and explanations
needs to be kept in mind, and then the absolute number of people who
remain displaced is more important than the relative number who find
durable solutions. In addition it is quite likely that the last remaining
displaced persons are the most politically contentious, and in such
circumstances their number does not have to be high for the issue to be very
salient.
Low membership values in the fuzzy set of “ended displacement
situation” are based on a combination of these two secondary-level
components. As argued above, a major displacement situation is hardly an
ended displacement situation. Therefore, cases where the remaining
displaced population is substantially larger than twenty-five thousand
persons will be considered full non-members of the set—unless at least 50
percent of the total displaced population have managed to find durable
73 An unconfirmed rumor has it that the figure twenty-five thousand was chosen because at the time it
resulted in a list of protracted refugee situations that neatly fit a standard page U.S. letter. The UNHCR
admits, in the original document, that the measure is “crude.”
126
CONDITIONS
solutions, in which case membership will be 0.25. If the remaining displaced
population is around twenty-five thousand, and at least 50 percent have
found solutions, it is difficult to determine whether a case should be
considered more in or out of the set, and, consequently, membership will be
0.50. For higher membership values, the size of the remaining displaced
population is key, and the any-solution-rate is disregarded. For full
membership in the set, the remaining displaced population should not
exceed ten thousand persons. Again, the figure ten thousand is not based on
any established convention in the literature, because there is none. However,
reporting on displacement becomes increasingly scant as numbers decrease,
and many summaries of displacement situations do not contain displaced
populations smaller than ten thousand.
In addition to the size of the remaining displaced population and the
any-solution-rate, the case of Mozambique in the previous section suggested
a third secondary-level component of the concept ended displacement
situation, namely “the cessation clause.” Article 1. C of the Refugee
Convention lists the circumstances under which the Convention shall cease
to apply to a person. This shall be the case, inter alia, if he has voluntarily reavailed himself of the protection of the country of his nationality, or if he has
acquired a new nationality and is protected by the country of that new
nationality. Most important in the present context, Article 1.C (5) states that
the Convention shall cease to apply to a person if
he can no longer, because the circumstances in connexion with which he has been
recognised as a refugee have ceased to exist, continue to refuse to avail himself of the
protection of the country of his nationality.
Until 2001, the cessation clause had twice been invoked by the UNHCR with
reference to the settlement of civil conflict: in 1973 concerning Sudan and in
1996 concerning Mozambique and Malawi. It had also been invoked seven
times with reference to “independence” and twelve times with reference to
“regime change/democratization” (Bonoan 2001, 22).
Article 1.C (5) is obviously more closely related to changed circumstances
in the country of origin than to a solution to the displacement situation. The
basis for invoking the cessation clause in the cases of Sudan and
Mozambique/Malawi was not the large number of displaced persons that
had repatriated, but the “fundamental nature and durability of the changes”
that had taken place (UNHCR 1996). Before invoking the cessation clause,
however, the UNHCR tries to find durable solutions for the displaced; the
cessation clause is not invoked at the initiation of, say, a repatriation
program, but in relation to the completion of it. By invoking the cessation
clause, the UNHCR sends the message that any circumstances related to the
creation of the displacement situation have ceased to exist, and that anyone
127
PEACE BY REPATRIATION
Table 8.5
The fuzzy set of “ended displacement situation.”
Value
1.00
0.75
0.50
0.25
0.00
Operationalization
< 10,000 remaining displaced persons, or cessation clause invoked
10,000–25,000 remaining displaced persons
approximately 25,000 remaining displaced persons, and any-solution-rate ≥50 percent
> 25,000 remaining displaced persons, and any-solution-rate ≥50 percent
> 25,000 remaining displaced persons, and any-solution-rate < 50 percent
who still resides outside his or her country of nationality does so by choice,
not because of well-founded fear. In essence, it means that the UNHCR
considers the problem solved.
More specifically, it means that the displacement situation is solved, not
necessarily that the refugees’ problems are solved. In the present context, the
UNHCR invoking the cessation clause is important enough an act to warrant
full membership in the fuzzy set of “ended displacement situation,”
irrespective of the size of the remaining displaced population and the anysolution-rate. Similarly, although the USCRI cannot invoke the cessation
clause, they occasionally make explicit assertions that certain groups of
people should no longer be considered displaced, as in the case of refugees
from Guatemala and, as discussed in the previous section, IDPs in both
Guatemala and Mozambique. Such assertions are made on the same basis
used by the UNHCR to invoke the cessation clause, and will equally merit
full membership.
The fuzzy set of “ended displacement situation” is summarized in table
8.5. In the same way as for “repatriation,” the assessment of membership in
the fuzzy set “ended displacement situation” is based on developments up to
five years after the conflict termination, corresponding to the five-year
measure of the outcome, sustainable peace (with the obvious exception of
cases where the termination fails within five years). Also, as in the case of
repatriation, two alternative codings are used, one including all displaced
persons, and one based on refugees and asylum seekers only, excluding
IDPs.
8.5 Partition
When it comes to partition there is more previous research to draw on than
was the case with repatriation and ended displacement situation. As a
starting point for the operationalization of the fuzzy set of “partition,” I use
Kaufmann’s list of terminated conflicts 1944–1994, in table 8.6.74
74 In addition to the thirteen cases presented in here, Kaufmann counts twelve cases that ended by complete
military victory by one side, and two that were suppressed by military occupation by a third party. However,
those latter types of conflict terminations are not relevant to a discussion of different degrees of partition.
128
CONDITIONS
Table 8.6
Outcomes of ethnic civil wars (adapted from Kaufmann 1996b, 160).
De jure partition
Ukrainians vs. USSR
Lithuanians vs. USSR
Eritreans vs. Ethiopia
independent 1991
independent 1991
independent 1993
De facto partition
Armenians vs. Azerbaijan
Somali clans
de facto partition
de facto partition in the north
Autonomy
Nagas vs. India
Basques vs. Spain
Tripuras vs. India
Palestinians vs. Israel
Miskitos vs. Nicaragua
Abkhazians vs. Georgia
autonomy 1972
autonomy 1980
autonomy 1972
autonomy 1993, partly implemented
autonomy 1990
autonomy 1993
Limited autonomy
Moros vs. Philippines
Chittagong Hill peoples vs. Bangladesh
limited autonomy 1990
limited autonomy 1989
Clearly, the creation of a new independent state constitutes full
membership in the fuzzy set of “partition,” and various forms of autonomy
can be used to define partial membership in the set.75 Because the role of
partition as a condition for peace is related to the security dilemma, two
aspects of autonomy should be considered in the operationalization process:
territoriality and power.
As discussed in chapter 2 political/territorial partition and ethnic
separation are different phenomena. I address partition in this section before
turning to fractionalization and group concentration in the following two
sections. The reason for this order is that partition, like repatriation and
ended displacement situation, can be directly related to the termination of
the conflict in question. However, as discussed further below, available data
on fractionalization and group concentration cannot be related to the
individual conflict terminations, but rather refer to certain aspects of the
conflict. This difference is important, and I therefore treat repatriation,
ended displacement situation, and partition as one set of conditions, and
fractionalization and group concentration as another set.
75 Autonomy as part of the operationalization of the fuzzy set of “partition” should not be confused with the
autonomy of municipal government, or states in a federation. While the latter refers to how centralized versus
decentralized a country is, what is relevant to the concept of partition is whether autonomy arrangements are
related to a specific territorial unit within the state, which sets that unit apart from other units in the state (or
to a specific group, setting that group apart).
129
PEACE BY REPATRIATION
Table 8.7
The fuzzy set of “partition.”
Value
1.00
0.75
0.50
0.25
0.00
Operationalization
de jure partition (new state created and recognized)
de facto partition, or autonomous region responsible for external security
autonomy arrangements related to territory
autonomy arrangements related to ethnic group
no autonomy arrangements
First, while most autonomy arrangements are territorially bound there
are alternatives, such as cultural autonomy which is awarded a certain group
in a society without geographic specifications (see e.g. Cornell 2001;
Hannum 1990; Lapidoth 1997). Though not irrelevant in this context,
cultural autonomy clearly warrants a lower membership in the fuzzy set of
“partition” than territorial autonomy. Second, the powers delegated to or
otherwise assumed by the autonomous entity can include anything from the
responsibility for basic social services, language and education, taxation, to
internal (police) and external (military) security and the right or ability to
enter into agreements with other states or international organizations. As a
basis for assessing membership scores in the fuzzy set of “partition,”
responsibility for external security, including control of border crossings, is
more pertinent than responsibility for education. Obviously, full
membership in the set is also based on territoriality and power.
8.6 Fractionalization
As mentioned in chapter 6, one dimension of ethnic structure concerns the
heterogeneity or fractionalization of a country and another concerns the
geographic distribution of ethnic groups. Fractionalization, operationalized
in the present section, is based only on the number and relative size of ethnic
groups in a country. In the next section, the concept of group concentration,
developed within the framework of the Minorities at Risk (MAR) Project
precisely to enable consideration of ethnic separation or intermixing, is
described and operationalized.
The traditional measure of ethnic fractionalization is calculated as 1
minus the standardized Herfindahl-Hirschman index76 of ethnolinguistic
group shares, and reflects the probability that two random individuals from
the population of a country belong to two different ethnic groups; the higher
the value, the more heterogeneous a country (see Alesina et al. 2003, 156–
158). This formula for calculating ethnic fractionalization is often referred to
as the ELF index (for ethnolinguistic fractionalization). The ELF index has
usually been based on data from the Atlas Narodov Mira, published by the
76
2
ELF  1   in1 s i where s i is the share of group ( i  1,..., n ) .
130
CONDITIONS
State Geological Committee of the Soviet Union in 1964, and reproduced in
Taylor and Hudson (1972; cf. Bridgman 2008; Posner 2004).
However, the traditional measure, most importantly the data from the
Atlas, have been criticized, and alternative measures are continuously being
developed.77 They are different in at least three important and partly related
respects. First, the types of cleavages used to identify various groups differ.
In addition to ethnicity, language is often taken into account and
occasionally also religion—which is not uncomplicated, since language and
religion are often part of the concept of ethnicity. The ethnolinguistic index
in the Atlas, for example, is based on “the common Eastern European
assumption that native language marks ethnicity” (Fearon 2003, 210).
Second, different data sources are used, often global ones such as the Atlas,
the CIA World Factbook, Encyclopedia Britannica, Barrett (1982), Levinson
(1998), or Vanhanen (1999), but also country studies and census data. Third,
various additional considerations are made when the actual indices are
constructed. For example, Rae and Taylor (1970, 26–33) propose different
formulae for large and small populations, Fearon (2003; cf. 2000) uses the
structural distance between languages to sensitize his measure to the
cultural distances between groups in a country, and Posner (2004)
constructs an index of “politically relevant groups,” which by implication
classifies some culturally or linguistically distinct groups as irrelevant
(depending on the issue at hand).
In this study I use data from Alesina et al. (2003), who present three
separate indices for ethnic, linguistic and religious fractionalization. These
indices are computed in the same way as the ELF index, described above, but
are not based on Atlas data. Instead, the Encyclopedia Britannica
constitutes the primary source, supplemented by various other sources as
needed (see Alesina et al. 2003, 158–160). I consider both the average score
of all three indices combined, and the ethnicity score on its own. The fuzzy
membership values are based on the distribution of the cases across the
ethnic fractionalization index. 78
77 See, for example, Posner (2004, 856) who presents the fractionalization scores for 42 African countries
according to seven different indices. For many countries the scores vary substantially from one index to the
other.
78 The thresholds are based on the distribution of the thirty-seven countries included in the present study (cf.
chapter 9), not on all two hundred fifteen states in Alesina’s dataset. This is because the distribution of all
states across the ethnic fractionalization index basically covers the entire range from zero to one, without
interruption, but when only the thirty-seven cases are analyzed it is possible to discern a few thresholds that
can be used to define the different fuzzy membership values in the set of “fractionalization.” It should be
noted, that the distribution of the cases is not very different from the distribution of all states, although the
average ethnic fractionalization score is higher, 0.54 compared to 0.44 for all states. Standard deviations are
0.24 and 0.25 respectively. The same is true for the average combined ethnic-linguistic-religious
fractionalization index, where cases score 0.50 compared to 0.42 for all states, and standard deviations are
0.21 and 0.19.
131
PEACE BY REPATRIATION
Table 8.8
The fuzzy set of “fractionalization.”
Value
1.00
0.75
0.50
0.25
0.00
Operationalization
>0.85 on ethnic fractionalization or > 0.85 on combined fractionalization
>0.60 on ethnic fractionalization or > 0.60 on combined fractionalization
>0.35 on ethnic fractionalization or > 0.35 on combined fractionalization
>0.20 on ethnic fractionalization or > 0.20 on combined fractionalization
<0.20 on ethnic fractionalization and < 0.20 on combined fractionalization
Importantly, available datasets on fractionalization are not based on
yearly fluctuations—which, for the purposes of this study, would have been
preferable. On the contrary, fractionalization is measured at intervals of a
decade or more. (Alesina uses the most recent data for two hundred fifteen
states, and time of data collection varies from 1979 to 2001.) As illustrated by
the case study on Bosnia-Herzegovina and the plans for a census in 2011, it is
particularly difficult to trace changes in fractionalization in the aftermath of
conflict. Hence, it was not practical within the framework of this study to
assemble yearly data on fractionalization for thirty-seven countries over a
period of several years (as was done for repatriation and ended displacement
situation). Therefore, rather than tracing how fractionalization changed
during the first several years after conflict termination, the condition
fractionalization in this study serves to describe the various countries where
the conflicts took place. I return to this caveat in chapter 10.
8.7 Group Concentration
As seen above, fractionalization is insensitive to the geographic distribution
of ethnic groups within a country. However, as Kaufmann argues: “The
severity of the ethnic security dilemma is greatest when demography is most
intermixed, weakest when community settlements are most separate.”
(Kaufmann 1996b, 148) Therefore, the concept of group concentration comes
closer to the essence of the security dilemma. As in the case of
fractionalization, the measure of group concentration refers to a specific
point in time and does not permit the analysis of change from one year to the
next. Thus, the condition group concentration will be used in the same way
as fractionalization to describe the context of the respective conflicts.
For data on group concentration I turn to the MAR project, which is the
only reasonable option. In a recent study, Melander measures regional
ethnic diversity and argues that “by relying on the MAR codings as the basis
for the operationalization of the key independent variable [of regional ethnic
diversity], the present study follows the state of the art in the field”
(Melander 2009, 103). The latest version of the MAR dataset has seen the
introduction of new group concentration indicators, the foundation of which
132
CONDITIONS
is the “regional base,” originally devised by Laitin. According to Laitin (2006,
213), the existence of a regional base requires that
(1) at least 25% of the minority population should live in a certain spacially contiguous
region, and (2) the percentage of the country’s minority population living in that
region should be significantly higher than the percentage of the country’s dominant
group living in that region.
The fuzzy set of “group concentration” used in this study is based on the
regional base and other group concentration indicators from the MAR
dataset. The minority populations referred to in the definition of a regional
base are ethnopolitical groups. The UCDP (used for case selection in chapter
9) distinguishes between conflicts over government and conflicts over
territory. In conflicts over government there are no ethnopolitical groups
and such conflicts are therefore, by definition, fully out of the fuzzy set of
“group concentration.” In (intrastate) conflicts over territory there are
ethnopolitical groups, and such conflicts will therefore have at least partial
membership in the fuzzy set of “group concentration.”
The most difficult type of case to assess is that in which there is a
regional base, but the ethnopolitical group is only weakly concentrated in it.
Such cases will therefore receive membership scores of 0.50 in the fuzzy set
of “group concentration.” Two MAR indicators are use to measure the
concentration of groups in the regional base. One concerns the ethnopolitical
group’s share of the total population in the region: <50 percent, 50–75
percent, or 75–100 percent (MAR indicator gc6b, cf. Davenport 2003, 18–
21). The other indicator concerns the proportion of the group that lives
outside the regional base: <25 percent, 25–50 percent, or 50–75 percent
(MAR indicator gc7). An ethnopolitical group is considered to be weakly
concentrated to the regional base if the group constitutes less than 50
percent of the population in the region and 50–75 percent of the group lives
outside the region (that is, the weakest concentration scores, respectively).
The membership scores of such cases will be 0.50.
Full membership, that is 1.00, will require the strongest group
concentration on both indicators—the group should constitute 75–100
Table 8.9
The fuzzy set of “group concentration.”
Value
1.00
0.75
0.50
0.25
0.00
Operationalization
regional base and strong concentration
regional base and medium concentration
regional base and weak concentration
no regional base
no group (conflicts over government)
133
PEACE BY REPATRIATION
percent of the region’s population and less than 25 percent of the group
should live outside the region. Other combinations of scores than the two
weakest or the two strongest on the two MAR indicators will result in
membership scores of 0.75 in the fuzzy set of “group concentration.” Finally,
conflicts over territory where the there is no regional base—that is, where
Laitin’s two criteria are not fulfilled—will receive membership scores of 0.25.
This concludes the operationalization of concepts into fuzzy sets. The
next chapter goes on to select the cases for the fuzzy-set analysis and to
assess their membership scores in each of the fuzzy sets constructed in this
chapter.
134
CONDITIONS
9 |Case Selection and Coding
The problem of intrinsic fuzziness in IR cannot be solved by any arbitrarily large increase
in the accuracy of our measurements. This fuzziness is embedded in the very nature of
foreign policy and international politics; from this perspective, it is “empirical” in a
very real sense.
Claudio A. Cioffi-Revilla79
9.1 Selection of Cases
As explained in Chapter 7, not all cases are equally relevant to a set-theoretic
argument. In an analysis of the peace-by-repatriation thesis relevant cases
should be armed conflicts that have both caused substantial displacement
and ended at one time or other, and which may or may not have restarted
(cf. Gerring 2007, 19). In this way, it will be possible to analyze both to what
extent each case is a case of repatriation and to what extent it is a case of
sustainable peace. Conflicts that either have not ended or did not cause
substantial displacement would be irrelevant (cf. the possibility principle,
chapter 7).
In order to identify cases for the fuzzy set analysis, I turned to the UCDP,
a collection of datasets about armed conflicts since the World War II. The
UCDP website,80 as well as each issue of States in Armed Conflict, an annual
publication by the Department of Peace and Conflict Research at Uppsala
University, lists a number of definitions used by the UCDP. Since the case
selection procedure is partly based on these definitions, two are presented
here:
An armed conflict is a contested incompatibility that concerns government and/or
territory where the use of armed force between two parties, of which at least one is the
government of a state, results in at least 25 battle-related deaths in one calendar year.
The intensity variable is coded in two categories: Minor: At least 25 but less than
1,000 battle-related deaths in a year. War: at least 1,000 battle-related deaths in a
year.
Until 2006 the UCDP used three levels of intensity: minor, intermediate, and
war. The intermediate level introduced a temporal dimension to the coding
of conflict intensity, defined as “more than 1,000 battle-related deaths
recorded during the course of the conflict, but fewer than 1,000 in any given
year (Sollenberg 2001, 9–11). However, because the intensity variable was
often misconceived as ordinal (as three levels of increasing intensity), the
79 Cioffi-Revilla (1981, 131); IR means International Relations.
80 http://www.pcr.uu.se/research/UCDP/index.htm.
135
PEACE BY REPATRIATION
intermediate level was dropped, and is now instead represented by a
separate dummy denoting cumulative intensity (UCDP 2009c, 7, footnote 8).
This dummy variable has been employed in the selection of cases for the
present analysis.
I used the Armed Conflict Termination Dataset, version 1.0, published in
2005, and employed several criteria to select cases. From the comprehensive
“conflict year” data, I first excluded all non-terminated conflicts (coded 0 on
“Term”). This resulted in a list of all conflict terminations 1946–2004, as
identified by the UCDP.
Next, though the claim about repatriation and peace is not made with
explicit reference to any specific period in time, the analysis needs to focus
on a limited number of years in order to be feasible. I decided to focus on the
period 1989–2001 (excluding cases coded <1989 or >2001 on “Year”). This
period was selected for two reasons. First, comprehensive peacebuilding
efforts aimed at “sustainable peace” rather than merely ending the use of
force are largely a post-Cold War phenomenon (see e.g. Doyle & Sambanis
2006, 22), and as argued in chapter 1, repatriation became the default
solution to refugee problems in the late 1980s. Hence, limiting the study to
1989 and later means that the analysis will focus on the time period during
which the argument about repatriation and peace has been most
emphatically pronounced and employed. The latter cutoff point, 2001, is
motivated by the operationalization of the fuzzy set of “sustainable peace”; it
must be possible to determine whether the conflict termination lasted for at
least five years.
The second reason for limiting the analysis to the period 1989–2001 has
to do with the main source used for information on displacement and return
(or other durable solutions) in this study, namely the World Refugee Survey
(WRS). This is an annual publication by the U.S. Committee for Refugees
and Immigrants. The WRS has been particularly useful during the period of
interest in this study. Each issue consists of three parts: (1) a statistical
overview of world refugee data covering both refugees and asylum seekers
worldwide by country of origin and country of asylum, main sources of
refugees and IDPs, major return movements, et cetera, (2) several thematic
articles about various aspects of displacement, and (3) detailed country
reports with information about displacement within and from a large
number of countries worldwide, based on critical assessments of a number of
different sources as well as field visits by USCRI staff. The WRS is
particularly suitable for the purposes of this study because it focuses on
conflict-induced displacement and attempts to relate displacement to
particular conflicts, even in countries with several simultaneous conflicts.
Unfortunately, its usefulness (for these purposes) is largely limited to the
period 1989–2001. Having consisted of around one hundred pages of less
detailed information in the late 1980s, the WRS became ever more extensive
136
CONDITIONS
and detailed during the 1990s, not least in terms of describing the
development of conflicts with consequences for displacement, following
various groups of displaced persons from year to year, and back-dating
displacement figures as estimates became more reliable over time. The 2000
issue is more than three hundred pages long. A few years after the turn of the
century, however, the publication became less extensive, and its focus shifted
to the asylum policies of host countries.
Finally, the resolution of intrastate and interstate conflicts works
differently (Wallensteen 2007, 81–189), and comprehensive peacebuilding
efforts are usually made after intrastate, or civil, conflicts rather than after
interstate ones. Therefore the latter type (coded 2 on “Type”) was also
excluded. Remaining conflicts are either internal conflicts or
internationalized internal conflicts.81 In order to further limit the number of
cases I decided to also exclude cases that had not resulted in at least one
thousand battle-related deaths during the entire course of the conflict (coded
0 on “CumInt”), based on the dummy variable on cumulative intensity.82
This process left me with a list of sixty-five conflict terminations in fiftytwo different conflicts, all of which had resulted in at least one thousand
battle-related deaths (see table 9.1). While several theoretical arguments
about the importance of return (or of avoiding it) are primarily related to
ethnic (territorial) conflicts, some are equally valid for ideological
(governmental) conflicts. In chapter 10 tests are conducted on the two types
of conflict both separately and together.
Further Reduction of Cases
The list of cases in table 9.1 was the result of limitations made on the basis of
the conflict termination dataset. However, in order for a conflict termination
to be relevant to a test of the peace-by-repatriation thesis the conflict should
have caused substantial displacement. In this study I use the threshold
twenty-five thousand displaced persons to identify substantial displacement.
As discussed in relation to the operationalization of the fuzzy set of “ended
displacement situation,” a “major, protracted refugee situation” consists of at
least twenty-five thousand displaced persons. Again, the figure twenty-five
thousand is arbitrary, but less so than any other conceivable figure.
On the basis of the substantial displacement criterion, fifteen conflict
terminations in nine conflicts were excluded from the list above, namely
Ethiopia (Ogaden), 1996; Senegal (Casamance), 2001; UK (Northern
81 Extraterritorial (or anti-colonial) conflicts (coded 1 on “Type”) were excluded indirectly by the exclusion of
pre-1989 conflict terminations.
82 The UCDP threshold of twenty-five battle-related deaths per year is low in comparison with other datasets.
Otherwise one thousand deaths per year has often been used. By excluding most minor armed conflicts, but
including those with a history of high intensity, the present study will analyze wars as well as less intensive
but protracted conflicts.
137
PEACE BY REPATRIATION
Table 9.1
A preliminary list of cases.
Afghanistan, 2001
Laos, 1990*
Angola, 1995
Lebanon, 1990
Azerbaijan (Nagorno-Karabakh), 1994
Liberia, 1995
Bangladesh (Chittagong Hill Tracts), 1992
Morocco (Western Sahara), 1989
Bosnia-Herzegovina (Croat), 1994
Mozambique, 1992
Bosnia-Herzegovina (Serb), 1995
Myanmar (Kachin), 1992
Cambodia, 1998
Myanmar (Karen), 1992, 1995
Chad, 1994
Myanmar, 1992*, 1994*
Congo, 1999
Nicaragua, 1989
Democratic Republic of Congo, 2001
Paraguay, 1989
El Salvador, 1991
Peru, 1999
Ethiopia (Eritrea), 1991
Philippines (Mindanao), 1990
Ethiopia, 1991
Philippines, 1995*, 1997*
Ethiopia (Ogaden), 1996*
Russia (Chechnya), 1996
Georgia (Abkhazia), 1993
Rwanda, 1994
Guatemala, 1995
Senegal (Casamance), 2001*
Guinea-Bissau, 1993
Sierra Leone, 2000
India, 1994*
Somalia, 1996*
India (Nagaland), 1997*, 2000*
Sri Lanka (Eelam), 2001
India (Punjab/Khalistan), 1993*
Sri Lanka, 1990*
Indonesia (Aceh), 1991*
Tajikistan, 1996, 1998
Indonesia (East Timor), 1989*, 1992*, 1998*
Uganda, 1991
Iran (Kurdistan), 1990, 1993, 1996
UK (Northern Ireland), 1991*, 1998*
Iran, 1993*, 1997*, 2001*
Yemen (South Yemen), 1994
Iraq (Kurdistan), 1993, 1996
Yugoslavia (Croatia), 1991**
Iraq, 1996
Yugoslavia (Kosovo), 1999
* Cases were excluded as the coding process continued; each is briefly explained below.
** Yugoslavia (Croatia) is a special case discussed further below.
Throughout, I use the original UCDP “name” of each conflict. This means that conflicts over
territory are those where a territory (or, in a few cases, an ethnic group) is added in parenthesis,
and conflicts over government are referred to simply by the name of the country, and that the
various conflicts in Burma/Myanmar are referred to by the name Myanmar.
Ireland), 1991, 1998; India (Punjab/Khalistan), 1993; Indonesia (East
Timor), 1989, 1992; Iran, 1993, 1997, 2001; Myanmar, 1992, 1994;
Philippines, 1995, 1997; and Sri Lanka, 1990.
In several of these cases it was difficult to find information about
displacement. It was particularly difficult to find information that allowed
displacement to be attributed to the conflict in question. The conflicts over
government in Iran and Myanmar can be used to illustrate this problem. In
both countries there were territorial conflicts going on at the same time.
These were of a greater magnitude and caused large-scale displacement. The
armed opposition in the conflict over government in Iran, the Mujahedin eKhalq (MEK), was not so much involved in warfare as in targeted attacks, the
two most important of which were carried out in 1981 (Keesing’s Record of
138
CONDITIONS
World Events May 1982). There are also reports that the MEK carried out a
few military operations in Iran during the final year of the Iran-Iraq war
(Keesing’s Record of World Events September 1988), but nowhere have I
been able to find information about substantial displacement related to the
conflict over government in Iran. In Myanmar, in addition to the conflict
between the government and the All Burma Students Democratic Front, the
UCDP identifies seven conflicts over territory (Arakan, Kachin, Karen,
Karenni, Mon, Shan, and Wa), two of which meet the criteria for inclusion in
the present analysis, namely Myanmar (Kachin) and Myanmar (Karen).
Regarding the conflict over government there are occasional reports of
displaced dissidents, students, and democracy activists, but any figures
provided are usually well below ten thousand. These figures are obviously
highly uncertain, but since I have no basis for making higher estimates the
case is excluded from the analysis.
Further, the case of Indonesia (East Timor), 1998 had produced some
displacement by 1998, although perhaps not as much as twenty-five
thousand displaced persons. However, the main reason for excluding the
case was the drastic change in the situation in 1999 in the aftermath of the
referendum on independence; out of a total population of eight hundred
thousand, up to seven hundred ninety thousand people were displaced by
anti-independence militia riots. Under those circumstances it would not
make sense to try to track the fate of those who were displaced prior to 1999
by the conflict between the Government of Indonesia and Fretilin.
Another four cases were excluded because they reached a high level of
intensity only after the conflicts had been terminated (and, obviously,
restarted), namely India, 1994, India (Nagaland), 1997 and 2000, and
Indonesia (Aceh)1991. In other words: at the time of the coded conflict
terminations these cases did not meet the substantial displacement criterion.
When remaining cases were cross-checked with the main conflict dataset
from the UCDP, Paraguay, 1989 and Laos, 1990 turned out not to meet the
cumulative intensity criterion. For example, in the case of Laos there was a
history of high intensity conflict (war between the government and the
Pathet Lao in the 1960s and early 1970s). However, the conflict that was
terminated in 1990 began only in 1976 and total battle-related deaths for
that conflict are given as 25–999.
The case of Somalia, 1996 was excluded because the conflict was not
actually terminated; it merely ceased to meet one of the criteria for inclusion
in the UCDP dataset, namely that one of the parties to the conflict is the
government of a state. The general anarchy in Somalia in the late 1990s
made it impossible to identify any party as a government. However,
according to Wallensteen and Sollenberg (2001, 642), “if the government
criterion is disregarded, Somalia would be included for the years 1997–2000
as an intermediate armed conflict, since it meets the other criteria.”
139
PEACE BY REPATRIATION
In addition to these exclusions, one of the conflicts in table 9.1 was
replaced by a previously excluded case. The conflict Yugoslavia (Croatia),
between the Yugoslav National Army and Croat irregulars over the secession
of Croatia from Yugoslavia was active at war level during 1991 and coded as
terminated by the end of 1991. In 1992 Croatia gained international
recognition, with Serb irregulars in the Krajina region breaking away from
the new state of Croatia. This latter conflict, labeled Croatia (Serb) was
active in the years 1992, 1993, and 1995—but at minor level only, and hence
excluded from the present analysis according to the cumulative intensity
criterion. The question is whether, for the present purposes, it is reasonable
to treat these two cases as different conflicts. Is the conflict Yugoslavia
(Croatia) best seen as terminated in 1991—and a durable termination at that,
giving the case a membership score of 0.5 in the fuzzy set of “sustainable
peace”? And is the conflict Croatia (Serb) best seen as having started in 1992
and as having resulted in a cumulative total of less than one thousand battlerelated deaths? On termination type the 1991 termination is coded as
“other,” indicating that the termination is a consequence of changes in
criteria rather than a cessation of violence.
My overall aim in the coding process is to avoid second-guessing
individual cases from the UCDP dataset, because if I make my own
evaluation of one case I ought to do the same for all cases. Nevertheless, I
believe the least arbitrary way of handling this issue is to treat the two
conflicts as one, even though the formal status of the parties and the
particular issue of contention changed in 1992. Further, I believe it makes
most sense to argue that the 1991 termination of the Yugoslavia (Croatia)
conflict not be counted as a conflict termination in the present context, but
instead that the two terminations of the conflict Croatia (Serb), in 1993 and
1995, are included in the analysis; they are treated as terminations of a
conflict which in 1991 was active at war level and should therefore be
included in the analysis. The most difficult consequence of merging these
two cases concerns the fuzzy set of “partition” (see below).
Finally, note that conflict termination for Cambodia is 1998, several
years after the major repatriation subsequent to the Paris agreement in the
early 1990s. The case still gets a high membership score in the fuzzy set of
“repatriation,” but the assessment is based on figures from 1997–2000.
9.2 The Coding Procedure
After the further reduction of cases just described, forty-three conflict
terminations in thirty-seven different conflicts remain and need to be given
membership scores in the fuzzy sets operationalized in chapter 8. The result
140
CONDITIONS
of the coding procedure is summarized in table 9.3, with “repatriation” and
“ended displacement situation” referring to both refugees and IDPs.83
Even with explicit criteria for what constitutes different degrees of
membership in the various fuzzy sets, the coding process was often anything
but straightforward. The most common problem is that data on
displacement are inherently uncertain and imprecise. Therefore, assessing
membership scores in the fuzzy sets of “repatriation” and “ended
displacement situation” was the most complicated, but in a few cases
difficulties were also related to the other fuzzy sets, including “sustainable
peace.”
It is very difficult to determine yearly figures of displacement and
repatriation during and after a conflict, especially when that conflict went on
for several years and perhaps overlapped with several other conflicts.
Displacement is not static. People move back and forth; they may or may not
choose to register with the UNHCR or other agencies when they flee and
when they return, and various factors related to personal security, political
conviction, economic situation et cetera, simultaneously affect their
decisions. The figures used for coding the cases in this study vary
considerably in precision and reliability. They are sometimes authoritative,
but often merely indicative.
As far as possible I have tried to rely on the information provided by the
USCRI in the World Refugee Survey mentioned above. This is because I
believe the coding will be more reliable if based on similar data, compiled
and evaluated according to the same method. Nevertheless, in many cases I
have had to look elsewhere for additional information. In fact, this is
particularly true for the cases that were excluded on the basis of not having
caused substantial enough displacement, such as the conflicts over
government in Iran and Myanmar, but it was necessary in several other
cases as well.
While it would take up too much space to discuss each case in detail and
explain how I chose to address the various difficulties I encountered, this
section describes the coding of six cases—three of the most complicated and
three of the easiest ones. This illustrates the kinds of problems I was
confronted with during the coding process and how I handled them.
Complicated Cases
The case of Iraq (Kurdistan) was first coded as terminated in 1993. It then
restarted briefly in 1996 and was terminated again during that same year; it
has not restarted since. During the period in question, Iraq consistently had
a Polity score of -9. The two terminations therefore receive membership
83 Appendix 2 presents all cases’ membership scores in all fuzzy sets, including those based on alternative
operationalizations discussed in chapter 10.
141
PEACE BY REPATRIATION
scores in of 0.25 and 0.50, respectively, in the fuzzy set of “sustainable
peace.”
Concerning the two fuzzy sets related to displacement, the case was more
complicated. The two Iraqi conflicts included in this analysis—the one
between the government and the Shiite Supreme Assembly/Council for the
Islamic Revolution in Iraq (ASIRI/SCIRI) over government, and the one
between the government and the Kurdistan Democratic Party (KDP), and the
Patriotic Union of Kurdistan (PUK)—are both characterized by a peak in
1991.
After the allies of Kuwait had used all means necessary to ensure Iraqi
withdrawal from Kuwait in 1991, both the Shiites and the Kurds seized the
opportunity to revolt against the government of Saddam Hussein, expecting
the support of the international community (cf. Malanczuk 1991; Strauss
2002). The international community, however, was not forthcoming, and the
government brutally crushed both uprisings, causing displacement of
immense proportions. However, most of those displaced during late spring
of 1991 had returned by the end of the year (USCRI 1992, 96–100).
The displacement related to the 1991 Kurdish uprising was exceptionally
large. At the peak of displacement in mid May 1991, more than 1.3 million
Iraqis had fled to Iran, and some four hundred fifty thousand to Turkey. The
situation was further complicated by the fact that this took place in a context
already characterized by large-scale conflict-induced displacement, some of
it related to the conflict Iraq (Kurdistan), and relevant to the present
analysis, and some not so related. Most importantly, some three hundred
fifty thousand Iraqis, mostly Kurds, were expelled to Iran during the 1980–
1988 Iran-Iraq war on suspicion of being of Iranian origin. For most of the
1990s these “expellees” continued to constitute the bulk of Iraqi refugees or
Iraqis living in refugee-like situations in Iran (USCRI 1998, 142), but they
are not counted in this study.
Detailed figures show that three Iranian provinces—West Azerbaijan,
Kordestan, and Bakhtaran84—together hosted 1,342,000 Iraqis at the peak of
displacement in 1991 (USCRI 1992, 97). These three provinces more or less
coincide geographically with Iranian Kurdistan and are just across the
border from Iraqi Kurdistan. It therefore seems reasonable to estimate that
most of the 1,300,000 refugees were Iraqi Kurds, even though this is not
expressly stated in the displacement data. By the end of the year there were
no more than 69,929 Iraqis left in the three provinces, a majority of whom
had arrived prior to 1991 (USCRI 1992, 97). By the beginning of 1992, some
44,000 Iraqis from the 1991 influx remained in Iran—30,000 Shiites and
14,000 Kurds (USCRI 1993, 98). While figures are uncertain, this means that
84 From 1979 to the early 1990s, Bakhtaran was the name used for the province otherwise known as
Kermanshah.
142
CONDITIONS
nearly all of the Iraqi Kurds who fled to Iran during 1991 repatriated within
the year—an extraordinary record, indeed. “The speed of the repatriation
was a source of wonder,” according to the USCRI (1992, 97). In addition to
those fleeing to Iran some 450,000 Iraqis, mainly Kurds, sought refuge in
Turkey in 1991. Here, too, almost all repatriated very soon; by the end of the
year only 4,079 of them remained in Turkey (USCRI 1992, 82–83).
However, Iraqi Kurds had been displaced by the conflict Iraq
(Kurdistan) also before the 1991 uprisings, primarily to Iran, making the
total number of displaced Iraqi Kurds close to 2 million in late spring 1991.
However, even when they are counted, the repatriation rate remains over 90
percent.
It could be argued that the repatriation that took place in 1991 and 1992
is of limited relevance to the establishment of peace several years later, and
particularly after the 1996 termination. However, I have argued that without
the massive repatriation that took place in 1991 and 1992 the situations in
1993 and 1996 would have been quite different, and that events in 1991 and
1992 should be considered when the membership scores of these
terminations are assessed. Therefore, despite continued new displacement
and repatriation during the rest of the 1990s, both terminations are given a
membership scores of 1.00 in the fuzzy set of “repatriation,” when only
refugees are considered.
There was also large-scale internal displacement related to the conflict
Iraq (Kurdistan). Internally displaced Iraqi Kurds can clearly be counted in
the hundreds of thousands, but any figures are rough approximations at
best. By the end of 1992, there were an estimated four hundred thousand
IDPs in Iraq (USCRI 1993, 52), at least half of which were Shiites in the
south (USCRI 1994, 105; cf. 1995, 116). The reported number of IDPs in
northern Iraq increased during the 1990s. By the turn of the century the UN
Center for Human Settlements estimated that more than 1 million had been
displaced at one time or other (see USCRI 2000, 187; cf. King 2007, 408).
The number of displaced at any one time may have been as high as eight
hundred thousand, many of which were uprooted not as a result of the
conflict between the Kurds and the Iraqi government, but by Kurdish infighting or by incursions and shelling of Kurdish-controlled areas from Iran
and Turkey (USCRI 2000, 187; cf. King 2007, 408).
This means that when IDPs are included, the repatriation rate decreases
and becomes more difficult to determine, but in light of the large numbers of
refugees that were displaced and repatriated during 1991, I have not found
enough evidence to suggest that adding IDPs to the equation should reduce
the membership score in the fuzzy set of “repatriation.” Again, this goes for
both terminations.
Then there is the fuzzy set of “ended displacement situation.” As noted,
Iraqi Kurds had been displaced before the 1991 wave of displacement and
143
PEACE BY REPATRIATION
repatriation, and they continued to be displaced in the following years. By
the end of 1995, there were still more than one hundred fifty thousand Iraqi
Kurds in Iran, not counting the “expellees” (USCRI 1996, 111; 1999, 149).
Substantial repatriation took place over the following years, but the number
remained above twenty-five thousand.85 Adding IDPs increases that number.
With more than twenty-five thousand people still displaced, but with more
than 50 percent having found solutions (though the durability of those
solutions may be questioned) the case gets a membership score of 0.25 in the
fuzzy set of “ended displacement situation,” for all displaced persons as well
as for refugees only and for both terminations.
Turning to partition, assessing the fuzzy membership scores is easier.
After the failed uprisings by Iraqi Kurds and Shiites and the massive
displacement that resulted, no-fly zones were created in the north and south
of Iraq. The northern one soon developed into a de facto Kurdish state
(Carpenter 2009, 23; Gunter 1993; 2003, 156) controlling, among other
things, strategic border crossings to Iran and Turkey (Dahlman 2002, 285–
292). Both terminations are given membership scores of 0.75 in the fuzzy set
of “partition.”
Iraq is a 0.50 member in the fuzzy set of “fractionalization,” and, in the
case of the Kurds, it is a full member of the fuzzy set of “group
concentration.”
Lebanon is one of several cases where data on democracy/autocracy and
respect for human rights are missing. Such missing data are unproblematic
in cases where the conflict restarts within five years, but when the conflict
termination is durable such data are needed in order to qualify the case’s
membership in the fuzzy set of “sustainable peace.” In the case of Lebanon,
the conflict was terminated in 1990, and Polity scores are missing for the
entire period 1975–2004. The first fifteen years, 1975–1989, are coded -77,
“interregnum, or anarchy,” and the next fifteen years, 1990–2004, are coded
-66, “foreign interruption.”
A case coded (-66) or (-77) is not necessarily highly authoritarian, but it
is definitely not strongly democratic (cf. Marshall & Jaggers 2009, 17–18). In
the case of Lebanon this fits Zahar’s (2002, 567–568) description of the
Syrian presence in Lebanon in the aftermath of the civil war, and Syria’s
stake in the stability of Lebanon: “In the short term, this stability has come
at the expense of human rights, democracy, and national reconciliation.”
85 Some seventy-five thousand Iraqi Kurds fled to Iran in 1996 as both the Iraqi government and Kurdish
troops entered the safe zone established through Operation Provide Comfort in 1991 (USCRI 1997, 151–152).
This displacement—and the internal displacement of perhaps one hundred fifty thousand people—was largely
the result of Kurdish in-fighting, and not related to the conflict Iraq (Kurdistan). However, it mentioned
because repatriation figures for 1996 and the following years relate both to the more than one hundred fifty
thousand Iraqi Kurds in Iran before 1996 and to the seventy-five thousand newly displaced.
144
CONDITIONS
Lebanon is given a membership score of 0.50 in the fuzzy set of “sustainable
peace.”
Regarding displacement caused by the conflict in Lebanon, figures are
highly uncertain. More than half the country’s population is estimated to
have been displaced at some point during the war (USCRI 1990, 84–85), and
the total population reportedly decreased by one-third, from 4 million to 2.7
million, during the 1980s (see USCRI 1993, 107).
Most of those who left Lebanon simply emigrated, that is, they did not
seek asylum and therefore did not become part of refugee statistics.
However, in 1994, the Social Affairs Ministry reported that around half of the
originally nine hundred thousand persons then estimated to have fled
abroad during the war had been able to return. It was not clear whether
these four hundred fifty thousand returnees had been able to return to their
pre-war homes (USCRI 1995, 120). There are no reports of further
significant return to Lebanon over the next few years.
Internal displacement was also massive, but figures are imprecise. By the
end of 1989, some five hundred thousand to 1 million Lebanese are believed
to have been internally displaced (USCRI 1990, 32), and the average number
of IDPs at any one time during most of the war is estimated at eight hundred
thousand (USCRI 1993, 106). According to the United Nations Development
Program (UNDP), by the end of 1996 some four hundred fifty thousand
remained internally displaced since the civil war (see USCRI 1997, 161).
Again, it is unclear whether those no longer considered displaced had
returned home or found other solutions.
On the basis of these uncertain figures, I estimate that more than half of
all displaced persons had found durable solutions five years after conflict
termination. However, with several hundred thousand remaining displaced,
the case’s membership score in the fuzzy set of “ended displacement
situation” is no higher than 0.25. Further, not all of those who found durable
solutions did so through repatriation, though many probably did. I believe it
is reasonable, therefore, to assume a repatriation rate at somewhere between
30 and 50 percent, for both cross-border and internal displacement. This
translates to a membership score in the fuzzy set of “repatriation” of 0.50.
Regarding partition, the 1989 Ta’if Agreement, which ended the conflict,
did not result in the partition of Lebanon. Rather, an intricate system of
power-sharing was introduced. The various religious communities in
Lebanon each has its own jurisdiction regarding several civil matters, and
the country has been referred to as “an example of a consociational system
with cultural autonomy based on the personality principle” (Hanf 1991, 46).
However, this cultural autonomy was a general facet of the Lebanese political
order, and not related to one particular ethnopolitical group. Therefore, the
case is given a membership score of 0.00 in the fuzzy set of “partition.”
145
PEACE BY REPATRIATION
Lebanon’s membership score in the fuzzy set of “fractionalization” is
0.50, and as in other conflicts over government, the membership score in the
fuzzy set of “group concentration” is 0.00.
The conflict in Tajikistan is coded in the UCDP as terminated twice, in
1996 and in 1998. Assessing membership scores in the fuzzy set of
“sustainable peace” is simple. The 1996 termination lasted for one year and
gets a membership score of 0.00; the 1998 termination was durable, but the
Polity score for Tajikistan was, and remains, negative, and the case gets a
membership score of 0.50 in the fuzzy set of “sustainable peace.”
The case of Tajikistan presents a striking example of discrepancies
between different sources of displacement data. For 1994, the USCRI reports
145,000 Tajik refugees and asylum seekers in Russia (1995, 42), while the
corresponding figure from UNHCR is zero (1995a, 54; 2005, 337). I have
handled this and similar instances by relating figures for several consecutive
years, trying to discern trends in total numbers of displaced, new
displacement, repatriation, and other durable solutions. In several cases,
more or less explicit corrections have been published a few years after a
complex crisis, and this also helps to reduce the margin of error. The 1994
discrepancy between the USCRI and the UNHCR can then likely be ascribed
to disagreement over who should count as refugees and asylum seekers,
rather than disagreement over how many people had left Tajikistan for
Russia because of the conflict in Tajikistan.
The number of refugees and asylum seekers from Tajikistan appears to
have peaked at nearly, or possibly above, 250,000 in 1995/1996 (USCRI
1997, 4–6, 140–141). By 2001, five years after the 1996 termination, that
figure was roughly halved; and during the next two years it was halved again.
This development is not immediately visible from the statistics sections of
the WRS, but it is a reasonable conclusion based on comparing country
reports of several issues of the publication. In the statistics section of the
WRS 2000 (covering the year 1999), Tajikistan is presented as the source of
63,000 refugees and asylum seekers, and Russia is said to host 104,300
refugees and asylum seekers, 89,400 from the former USSR (with no further
specifications) and 14,900 from other places (USCRI 2000, 2–5). However,
the country report for Russia sets the number of Tajik refugees and asylum
seekers in Russia at 107,355. Of these, 12,299 were recognized refugees and
95,056 were so called forced migrants86 (USCRI 2000, 265).
According to the statistics section of the WRS 2001 (covering 2000),
there were 60,000 refugees and asylum seekers from Tajikistan, 2,100 of
86 Despite the distinction in Russian asylum legislation between forced migrants or involuntarily relocated
persons (which applies to persons from within the former Soviet Union, or the “near abroad”) and refugees
(which applies to persons from countries outside the former Soviet Union, or the “far abroad”), displaced
persons from Tajikistan are found in both categories (see e.g., USCRI 1998, 199).
146
CONDITIONS
which were in Russia (USCRI 2001, 2–5). However, the country report for
Russia refers to 2,061 refugees and 83,039 forced migrants from Tajikistan,
a total of 85,100 (USCRI 2001, 247).
The statistics section of WRS 2002 (covering 2001) identifies Tajikistan
as the origin of 55,000 refugees and asylum seekers, but does not specify
how many of them are in Russia; of a total of 28,200 refugees and asylum
seekers in Russia, 14,800 are from Georgia, and 13,400 are “other” (USCRI
2002, 2–5). However, the country report for Russia sets the number of
forced migrants from Tajikistan at 59,190 (USCRI 2002, 236). The number
of refugees is not specified, but only 236 persons in all were granted refugee
status in Russia during 2001 (USCRI 2002, 235).
While the largest number of Tajik refugees and asylum seekers found
their way to Russia, significant numbers also went to the neighboring Central
Asian states of Kazakhstan, Kyrgyzstan, Turkmenistan and particularly to
Uzbekistan. The WRS 1995 sets the number of Tajiks who fled to Uzbekistan
at 50,000 (USCRI 1995, 109). The WRS 1996 makes no mention of Tajiks in
Uzbekistan. The following two years, there are reports of 30,000 Tajiks
living in “refugee-like situations” in Uzbekistan (USCRI 1997, 11; 1998, 11).
The WRS 1999 again does not mention them, but from the 2000 issue, they
are back and considered refugees, without any explanation of why they are
now refugees rather than people in refugee-like situations (e.g., USCRI
2000, 3, 173; 2001, 3, 166).
So, to sum up, to assess the fuzzy membership scores for Tajikistan, I
have consulted the statistics sections of the WRS, I have considered the
numbers of “refugees” and “forced migrants” from the country reports on
Russia, and I have counted 30,000 Tajiks in Uzbekistan each year,
irrespective of whether they are listed in the WRS as refugees, people in
refugee-like situations, or not listed at all.
Table 9.2 shows the numbers of refugees and asylum seekers from
Tajikistan from 1994 through 2002 according to the statistics section (left
column) and on the basis of the considerations just mentioned (right
column). While the figures in both columns are rough approximations, I
believe the right column makes more sense as a realistic development of
displacement trends than the left column.
With more than twenty-five thousand remaining displaced persons, and
an any-solution-rate of at least 50 percent, the membership score in the
fuzzy set of “ended displacement situation” is 0.25 for both terminations,
1996 and 1998.
In 1997 a Protocol on Refugees was signed as part of the ongoing peace
process, aiming at the voluntary return of all displaced persons to their
homes (Bell 2000, 372–373). Nevertheless, repatriation to Tajikistan was
very limited. The majority of refugees and asylum seekers from Tajikistan
who found durable solutions did so through local integration, largely in
147
PEACE BY REPATRIATION
Table 9.2
Refugees and asylum seekers from Tajikistan, 1994–2002.
Year
1994
1995
1996
1997
1998
1999
2000
2001
2002
Statistics section
165,000
170,400
245,600
68,400
35,000
60,600
57,650
42,500
97,342
Other considerations
248,000
243,100
245,600
218,400
172,212
167,955
142,750
102,500
97,342
countries where they belonged to the ethnic majority; for example, the fifty
thousand Tajiks who fled to Uzbekistan were of Uzbek origin, and the twenty
thousand Tajiks who went to Turkmenistan were of Turkmen descent, and
found solutions in those countries (USCRI 1995, 109). During several years
after 1996 (when the conflict was terminated a first time), no more than an
estimated thirty thousand displaced persons repatriated to Tajikistan (see
e.g., USCRI 1997, 140; 1998, 132; 1999, 139; 2000, 173). This is clearly less
than 30 percent of two hundred fifty thousand, and both terminations are
given membership scores of 0.00 in the fuzzy set of “repatriation.”87
The discussion so far has concerned refugees and asylum seekers.
However, the conflict also resulted in large-scale internal displacement,
some five hundred thousand Tajiks—10 percent of the country’s total
population—became internally displaced. For most of the IDPs the
displacement was of relatively short duration. It has been estimated that
more than 90 percent of IDPs had returned home by the end of 1994, the
vast majority already during 1993 (USCRI 1995, 109). This means that, when
both refugees and IDPs are considered, more than 50 percent of displaced
persons found durable solutions through return, giving both terminations
membership scores of 0.75 in the fuzzy set of “repatriation.” Tajikistan’s
membership score in the fuzzy set of “ended displacement situation” is the
same for all displaced persons as for refugees only: 0.25.
Tajikistan is fully out of the two fuzzy sets of “partition” and “group
concentration,” but its membership score is 0.50 in the fuzzy set of
“fractionalization.”
87 Another estimated thirty thousand refugees and asylum seekers repatriated during 1993–1994, that is
before displacement peaked in 1995–1996. This means that a total of up to sixty thousand Tajik refugees and
asylum seekers repatriated – still below 30 percent of two hundred fifty thousand.
148
CONDITIONS
Easy Cases
The 1995 termination of the conflict in Liberia lasted for four years, which
makes the case’s membership score in the fuzzy set of “sustainable peace”
0.25.
Displacement peaked in 1994 with 784,000 refugees and asylum seekers
and 1,100,000 IDPs (USCRI 1995, 44). By the end of 1999, “approximately
two-thirds of all Liberian refugees and nearly 90 percent of all internally
displaced Liberians had returned to their home areas,” leaving 250,000
refugees and 50,000 IDPs (USCRI 2000, 5–6, 100). Consequently, Liberia’s
membership score in the fuzzy set of “repatriation” is 1.00 for all displaced
persons and 0.75 for refugees only. The membership score in the fuzzy set of
“ended displacement situation” is 0.25, both for all displaced persons and for
refugees only.
There was no partition or autonomy in the case of Liberia, so the
membership score in the fuzzy set of “partition” is 0.00. The case is fully in
the fuzzy set of “fractionalization,” but being a political conflict, it is fully out
of the fuzzy set of “group concentration.”
The conflict in Mozambique was terminated in 1992 and has not restarted.
Further, since 1994 the country has consistently had a Polity score of +6, and
the standardized CIRI value for Mozambique is 7.0. Finally, the United
Nations Operation in Mozambique (ONUMOZ) left by the end of 1994. This
means that the case of Mozambique is a full member of the fuzzy set of
“sustainable peace.”
As already mentioned in chapter 8 Mozambique is considered a very
successful case of repatriation. Over a few years beginning in 1992 some 80
percent of the 1.7 million refugees and the 3.5 million IDPs returned home.
In 1996 the UNHCR invoked the cessation clause for Mozambican refugees
worldwide. By then, there were no IDPs in Mozambique, and the UNHCR
concluded that the one hundred thousand odd refugees who remained in
South Africa, Zambia, and Tanzania were there “for economic rather than
security reasons.” This makes the case a full member of the fuzzy sets of
“repatriation” and “ended displacement situation,” both when all displaced
persons are considered and when only refugees are taken into account.
There was no partition or autonomy, and the case is also fully out of the
fuzzy set of “group concentration.” It gets a membership score of 0.75 in the
fuzzy set of “fractionalization.”
Also in 1992, a truce was reached in the conflict Bangladesh (Chittagong
Hill Tracts). Over the following years negotiations continued, resulting in a
peace agreement in 1997. No peacekeeping operation was ever deployed to
Bangladesh and the Polity score and the standardized CIRI value are both
149
PEACE BY REPATRIATION
above the thresholds set for full membership in the fuzzy set of “sustainable
peace.”
The conflict resulted in the displacement of tens of thousands of the
tribal population, primarily Buddhist Chakma. For several years before the
conflict ended, some fifty thousand Chakma refugees were living in India.
That number increased temporarily to seventy-five thousand during 1991,
but by the end of 1992 it was back to fifty thousand (USCRI 1991, 91–92;
1992, 95–96; 1993, 89, 93). The 1997 peace agreement gave new momentum
to the repatriation process. Ten thousand Chakma repatriated during 1997,
and the remaining forty thousand returned by late February 1998, effectively
bringing the displacement situation to an end (USCRI 1998, 122–123; 1999,
125).88
The peace accord contained provisions for regional autonomy (see e.g.,
Bell 2000, 26; Rashiduzzaman 1998), and the case gets a membership score
of 0.50 in the fuzzy set of “partition.” Bangladesh is fully out of the fuzzy set
of “fractionalization” and the case has a membership score of 0.75 in the
fuzzy set of “group concentration.”
A Few Final Comments
Among the terminated conflicts in the UCDP dataset, several cases are coded
as having restarted with new parties. Examples are Chad, Congo, Liberia,
and Rwanda. It could be argued in these cases that the conflict terminations
remained durable and that they should be given membership scores in the
fuzzy set of “sustainable peace” on that basis. However, these renewed
conflicts were often fought between a government consisting of the former
rebels fighting rebels consisting of the former government and/or its allies—
”basically a continuance of the first phase, albeit under new banners” (UCDP
2009a). In this study, I have considered such cases the same conflict.
The “combined” case of Yugoslavia (Croatia) and Croatia (Serb) was the
most complicated case to code on partition. The former conflict would be a
full member of the fuzzy set partition, and the latter conflict a full nonmember. I argued that by 1995 and even 1993—the years of the conflict
terminations included in the analysis—Croatian independence was no longer
an issue of contention, at least not one that the Serb side realistically fought
against. Rather, the conflict that was terminated in 1993 and 1995 concerned
the regions of Eastern Slavonia, Baranja, and Western Sirmium. Therefore,
the combined case is given a membership score of 0.00 in the fuzzy set
“partition.”
88 In January 1996, sixty-five thousand Chakma living in Arunachal state in India became the focus of
attention (USCRI 1997, 131). They were generally considered immigrants rather than refugees, but in 1996 a
local group in Arunachal Pradesh tried to forcibly repatriate them to Bangladesh. India’s National Human
Rights Commission brought the case before the Supreme Court, which ordered a halt to the forced
repatriation. These sixty-five thousand are not included in the present analysis.
150
CONDITIONS
Ended displacement
situation
Partition
Fractionalization
Group concentration
Afghanistan, 2001
Angola, 1995
Azerbaijan (N.-Karabakh), 1994
Bangladesh (CHT), 1992
Bosnia-Herzegovina (Croat), 1994
Bosnia-Herzegovina (Serb), 1995
Cambodia, 1998
Chad, 1994
Congo, 1999
Croatia (Serb), 1993
Croatia (Serb), 1995
Democratic Republic of Congo, 2001
El Salvador, 1991
Ethiopia, 1991
Ethiopia (Eritrea), 1991
Georgia (Abkhazia), 1993
Guatemala, 1995
Guinea-Bissau, 1999
Iran (Kurdistan), 1990
Iran (Kurdistan), 1993
Iran (Kurdistan), 1996
Iraq, 1996
Iraq (Kurdistan), 1993
Iraq (Kurdistan), 1996
Lebanon, 1990
Liberia, 1995
Morocco (Western Sahara), 1989
Mozambique, 1992
Myanmar (Kachin), 1992
Myanmar (Karen), 1992
Myanmar (Karen), 1995
Nicaragua, 1989
Peru, 1999
Philippines (Mindanao), 1990
Russia (Chechnya), 1996
Rwanda, 1994
Sierra Leone, 2000
Sri Lanka (Eelam), 2001
Tajikistan, 1996
Tajikistan, 1998
Uganda, 1991
Yemen (South Yemen), 1994
Yugoslavia (Kosovo), 1999
Repatriation
Case
Peace
Table 9.3
Coding of selected cases.
0.00
0.25
0.50
1.00
0.50
0.50
0.50
0.25
0.25
0.00
1.00
0.25
1.00
0.50
0.50
0.50
1.00
0.50
0.25
0.25
0.50
0.50
0.25
0.50
0.50
0.25
0.50
1.00
0.50
0.25
0.00
1.00
1.00
0.25
0.25
0.25
0.50
0.00
0.00
0.50
0.25
0.50
0.75
0.75
0.00
0.00
1.00
0.25
0.25
1.00
0.25
1.00
0.00
0.25
0.75
0.25
1.00
0.75
0.00
0.25
1.00
0.00
0.00
0.00
0.00
1.00
1.00
0.50
1.00
0.00
1.00
0.00
0.00
0.00
0.75
0.00
0.00
0.75
1.00
0.75
0.75
0.75
0.75
1.00
1.00
0.75
0.00
0.00
0.00
1.00
0.25
0.25
0.75
0.50
0.25
0.00
0.00
0.25
1.00
0.75
0.25
0.00
1.00
1.00
0.25
0.50
0.50
0.00
0.25
0.25
0.25
0.25
0.00
1.00
0.00
0.00
0.00
0.50
1.00
0.00
0.25
0.25
0.50
0.25
0.25
0.25
0.50
1.00
0.25
0.00
0.00
0.75
0.50
0.00
0.50
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
1.00
0.50
0.00
0.00
0.00
0.00
0.00
0.00
0.75
0.75
0.25
0.00
0.75
0.00
0.75
0.75
0.75
0.00
0.00
0.50
0.75
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.50
0.75
0.75
0.25
0.00
0.75
0.75
0.25
1.00
1.00
0.50
0.50
1.00
0.25
0.75
0.75
0.50
0.50
1.00
0.75
0.75
0.75
0.50
0.50
0.50
0.50
1.00
0.50
0.75
0.50
0.50
0.50
0.50
0.75
0.50
0.25
0.50
1.00
0.50
0.50
0.50
1.00
0.00
0.50
0.00
0.00
1.00
0.75
0.75
0.75
0.00
0.00
0.00
0.75
0.75
0.00
0.00
0.00
1.00
0.75
0.00
0.00
1.00
1.00
1.00
0.00
1.00
1.00
0.00
0.00
1.00
0.00
0.75
0.25
0.25
0.00
0.00
1.00
1.00
0.00
0.00
0.75
0.00
0.00
0.00
0.00
1.00
151
PEACE BY REPATRIATION
Regarding fractionalization, it should be noted that for the case
Yugoslavia (Kosovo) I use the figure given at Serbia/Montenegro in
Alesina’s data set. Both “Yugoslavia” and “Serbia/Montenegro” are cases in
the data set; data for Yugoslavia is from 1995, and data for Serbia/
Montenegro is for 1991 (which would suggest that I use the more recent
figure, related to “Yugoslavia”). However, more importantly, Yugoslavia
corresponds to pre-1991 Yugoslavia, that is, including Slovenia, Croatia, et
cetera. Serbia/Montenegro corresponds to the country in which the Kosovo
war was fought. Therefore, it makes more sense to use the fractionalization
measure for Serbia/Montenegro.
The fuzzy membership scores of all cases are presented in table 9.3.
152
CONDITIONS
10 |Systematic Comparative Analysis
While it may be a lamentable comment on the human condition, the fact is that population
flows that increase ethnic homogeneity are much less likely to be reversed than those that
create greater heterogeneity.
USCRI89
10.1 Testing the Peace-by-Repatriation Thesis
The previous three chapters have laid the groundwork for the fuzzy-set
analysis conducted in the present chapter. In chapter 7 the methodology of
necessary conditions and set theory was introduced, the method of
Qualitative Comparative Analysis (QCA) was presented, and fuzzy sets and
fuzzy-set analysis were described. The concepts of the peace-by-repatriation
framework, which I developed in chapters 2–6, were operationalized as fivevalue fuzzy sets in chapter 8. Finally, in chapter 9 I selected and coded cases
for the fuzzy-set analysis.
This chapter begins with a few introductory comments on how the
analysis was conducted. The peace-by-repatriation thesis is then tested on
the entire set of cases, followed by separate tests for conflicts over territory
and conflicts over government. After discussing the findings of the analysis,
a few alternative tests are conducted to see whether the results are affected
by alternative operationalizations of key concepts.
Preliminaries
Of the forty-three cases—conflict terminations—analyzed in this chapter,
twenty-two are terminations of conflicts over territory, and twenty-one are
terminations of conflicts over government (cf. chapter 8). Because the role of
ethnicity differs in these types of conflict, after I test the peace-byrepatriation thesis on all conflicts, I test it separately on conflicts over
territory and conflicts over government. I also conduct tests involving all
displaced persons as well as tests limited to refugees. On the one hand, in
light of the explanations of how repatriation contributes to peace, it is
reasonable to presume that the peace-by-repatriation thesis applies to
refugees and IDPs alike. On the other hand, because explicit reference is
often made only to refugees there is reason to examine the thesis also
without including IDPs. Note that five conflicts—Georgia (Abkhazia),
Guinea-Bissau, Myanmar (Kachin), Peru, and Russia (Chechnya)—
produced almost no refugees, and therefore do not fulfill the “minimum
displacement criterion” for inclusion in the analysis when it is restricted to
89 USCRI (1994, 29).
153
PEACE BY REPATRIATION
Table 10.1
Number of cases in different tests.
Conflicts over territory
Conflicts over government
All cases
All displaced persons
22
21
43
Refugees only
19
19
38
refugees. The number of cases covered by different combinations of cases
and types of displaced persons are most easily presented in tabular form.
This is done in table 10.1.
An important distinction between some of the conditions should also be
reiterated. Three conditions—repatriation, ended displacement situation,
and partition—are clearly connected to their respective cases, that is, to the
conflict terminations. The two remaining conditions—fractionalization and
group concentration—are not so connected, and serve rather to describe part
of the context of the cases. The reason for this, as described in chapter 8, is
that there are no readily available data that allow changes in fractionalization
and group concentration to be traced from one year to the next. Clearly, this
is not ideal, but I believe that using the available, albeit less-than-ideal, data
on fractionalization and group concentration is the best realistic option
under the circumstances. To illustrate the difference between the conditions
consider the hypothetical solution REPALL*FRAC → PEACE (cf. table 10.2).
This should be interpreted as meaning that post-conflict repatriation is
related to peace in cases where fractionalization is high; the measure of
fractionalization, however, does not permit the conclusion that post-conflict
repatriation should be pursued with the aim of increasing fractionalization.
The software used to conduct the fuzzy-set analysis in this study is
fs/QCA, version 2.5.90 All tests are run using the software’s default frequency
and consistency cutoffs (frequency cutoff 1; consistency cutoff 0.80). I use
the (recommended) “truth table algorithm.”91 However, having selected the
frequency and consistency cutoffs (as described in the final section of
chapter 7), I use the “specify analysis” option rather than the
(recommended) “standard analyses” option. I do this because the latter
procedure requires estimations of the plausibility of counterfactuals (Ragin
2008b, 49–53, 81). What this means is that, since there will most often be
configurations that are not represented in the empirical data (so called
logical remainders), the researcher needs to estimate how likely it is that
these configurations would be associated with the positive outcome were
they to exist. When the researcher possesses enough theoretical and
90 The fs/QCA software can be downloaded from http://www.u.arizona.edu/~cragin/fsQCA/software.shtml.
91 The alternative ”inclusion algorithm” simply calculates the proportions of cases consistent with the claim of
necessity or sufficiency, cf. chapter 7.
154
CONDITIONS
Table 10.2
Abbreviations of concepts used in the analysis.
PEACE
PEACE 2
REPALL
REPREF
ENDALL
ENDREF
ENDALL 2
ENDREF 2
PART
FRAC
FRAC 2
CON
Sustainable peace
Alternative coding of PEACE
Repatriation (all displaced persons)
Repatriation (refugees only)
Ended displacement situation (all displaced persons)
Ended displacement situation (refugees only)
Alternative coding of ENDALL
Alternative coding of ENDREF
Partition
Fractionalization
Alternative coding of FRAC
Group concentration
substantive knowledge to be able to make such estimations then the analysis
is strengthened by the ability to take that knowledge into consideration.
However, I do not believe that available knowledge about the peace-byrepatriation thesis is sufficient to permit such estimations. For the same
reason, I choose the parsimonious solution throughout (cf. chapter 7 and
Appendix 4).
I use abbreviations for the various conditions when I present formulas,
with upper case letters signifying the presence of conditions and lower case
letters representing their absence. The abbreviations are presented in table
10.2.
Because I test numerous different combinations of conditions on several
different groups of cases, the chapter is interspersed with formulas and
consistency and coverage scores. The results covering a specific group of
cases are summarized at the end of each section. The first tests are described
in some detail (cf. also Appendix 5) to enhance understanding of how to
interpret the solutions.
10.2 All Conflicts
As portrayed above, the aim of diversity-oriented research is not to measure
the isolated effect of independent variables. However, “good practices” of
fs/QCA prescribe that any necessary conditions be eliminated from the truth
table analysis and addressed separately (see Ragin 2009, 118). This requires
testing for necessary conditions before conducting the truth table analysis. In
the test of necessity a high consistency threshold should be employed.
Testing each of the five conditions for necessity produces consistency
scores varying from 0.2949 (partition) to 0.7436 (fractionalization), but
none of them is close to the default 0.80 threshold. In addition, 0.90, or even
0.95, is probably more suitable in this test, where a high consistency
threshold is explicitly called for. In other words, none of the conditions is by
155
PEACE BY REPATRIATION
itself a necessary condition for sustainable peace when tested on all cases,
and all five are therefore retained for further analysis.
From a deterministic perspective these results are all we need to refute
the peace-by-repatriation thesis in its unidirectional Hard I form, that
repatriation is a necessary condition for peace. The analysis demonstrates
that there are cases of sustainable peace which are not at the same time cases
of repatriation. Peace is therefore possible in the absence of repatriation,
and, hence, repatriation is not necessary for peace. This is in line with
Adelman’s conclusion that no one relationship between repatriation and
peace is valid across all cases. However, the rationale for this study is to go
beyond that basic conclusion and perform a more detailed study of the
peace-by-repatriation thesis. Before proceeding to the fuzzy-set analysis,
however, the “bivariate” relationship between repatriation and peace will be
discussed a bit further.
This relationship can be illustrated graphically in an XY plot as in figure
10.1 (similar to the one used to illustrate fuzzy-set analysis of necessity in
figure 7.1). In order to support the peace-by-repatriation thesis the analysis
would have had to result in all cases ending up on or below the diagonal
X=Y. The figure clearly shows that they do not.
However, there are at least two things that the figure does not reveal.
First, it does not tell us how many cases end up at each intersection of the
XY plot. From the sixteen intersections marked in the figure, we know that
ten of the forty-three cases analyzed end up on or below the diagonal, and
that six end up above it, but we do not know where (that is, in which of the
sixteen marked intersections) the remaining twenty-seven cases end up.
Whether they end up primarily on-or-below or above the diagonal
Figure 10.1 Fuzzy-set analysis of peace by repatriation, XY plot.
156
Croatia (Serb) ‘93
Myanmar (Karen) ‘95
0.25
0.00
Chad
0.25
BiH (Croat)
BiH (Serb)
Croatia (Serb) ‘95
El Salvador
Guatemala
Figure 10.2 XY plot of membership in PEACE / REPALL.
PEACE
0.00
Iran (Kurd) ‘90,
Iran (Kurd) ‘93
Myanmar (Karen) ‘92
Philippines
(Mindanao)
Angola
0.50
Peru
Azerbaijan (N-K)
Georgia (Abkhazia)
Iran (Kurd) ‘96
Morocco (W Sahara)
Myanmar (Kachin)
Iraq
0.75
1.00
REPALL
0.50
Lebanon
0.75
Sri Lanka (Eelam)
Afghanistan
Tajikistan ‘96
Russia (Chechnya)
DRC
Ethiopia (Eritrea)
Sierra Leone
Tajikistan ‘98
Yugoslavia (Kosovo)
Nicaragua
1.00
Iraq (Kurd) ‘93
Congo
Liberia
Rwanda
Uganda
Iraq (Kurd) ‘96
Yemen (S Yemen)
Cambodia
Ethiopia
Guinea-Bissau
Bangladesh (CHT)
Mozambique
CONDITIONS
determines, together with their distance from the diagonal, the character of
the subset relation. That character, however, is captured by the concepts of
consistency and coverage, as explained in chapter 7. The software calculates
the consistency and coverage scores, making a more detailed graphic display
of the exact distribution of cases across the XY plot redundant. In this
particular instance, fifteen of the remaining twenty-seven cases end up on or
below the diagonal, and twelve end up above, yielding a consistency of
157
PEACE BY REPATRIATION
0.5769 and a coverage of 0.5233, which is, as has been noted, far from a
consistent subset relation.
The other thing the figure does not tell us is which cases end up where in
the XY plot. While this does not affect consistency and coverage, potentially
interesting patterns can be revealed by illustrating the position of different
cases. This is done in figure 10.2 (which also shows how many cases end up
at each intersection).
The figure shows, for example, that half of the conflicts over territory
(bold font) are fully out of the fuzzy set of “repatriation,” whereas two-thirds
of conflicts over government have a membership score of at least 0.75 in that
same set. Further, all Latin American cases are full members of the fuzzy set
of “sustainable peace,” but have on average low membership scores in
“repatriation.” Of the thirteen African cases, more than half are full members
in “repatriation,” but only one has a membership of >0.50 in “sustainable
peace.”
Clearly, to repeat, the peace-by-repatriation thesis in its unidirectional
Hard I version does not hold across all cases. The more detailed study of the
thesis now proceeds with fuzzy-set analysis. First, analyzing all five
conditions together yields no consistent solutions. However, testing various
combinations of two, three, or four conditions does. For example, a test of
REPALL and ENDALL yields a consistent solution (Formula 10.1):
ENDALL → PEACE
consistency 0.8226, coverage 0.6538
Five cases have full membership in this solution (1.00 membership in
ENDALL and 1.00 membership in PEACE), namely Bangladesh (Chittagong
Hill Tracts), El Salvador, Guatemala, Mozambique, and Peru, and they
make a significant contribution to the consistency and coverage scores of the
solution.92 Four other cases have >0.50 membership in this solution, namely
Guinea-Bissau (1.00 in ENDALL/0.50 in PEACE), Yemen (South Yemen)
(1.00/0.50), Cambodia (0.75/0.50), and Ethiopia (0.75/0.50).
The solution ENDALL → PEACE means that ENDALL is a consistent
and relevant subset of PEACE, which suggests that ending displacement goes
a long way toward bringing sustainable peace. Sustainable peace is obviously
a function of many conditions not included in the present analysis. I would
not, therefore, conclude that ending displacement is “almost always
sufficient” for peace. However, the result suggests a significant empirical
relationship between ending displacement and peace.
92 The subset relation ENDALL ≤ PEACE holds for a majority of cases in the analysis, including Nicaragua
(0.50 ≤ 1.00), Yugoslavia (Kosovo) (0.25 ≤ 0.75), and Afghanistan (0.00 ≤ 0.00), but cases with stronger
membership are more relevant to the argument.
158
CONDITIONS
Combining REPALL with any one of the other three conditions does not
produce any consistent solutions. However, ENDALL yields consistent
solutions in combination with each of them. ENDALL and PART produces
the same solution as presented in formula 10.1, as does ENDALL and FRAC.
ENDALL and CON yields a more complex solution—though still the
parsimonious option in the software (Formula 10.2):
ENDALL*CON → PEACE
consistency 0.8750, coverage 0.1795
Bangladesh (CHT) has strong, though not full membership in this solution
(ENDALL 0.75/PEACE 1.00). No other case has >0.50 membership in the
solution term ENDALL*CON. The consistency of this solution is somewhat
higher than that presented in formula 10.1, but coverage is markedly lower;
that is, the subset relation is consistent, but not very relevant. This means
that cases where displacement has ended after a conflict involving a
regionally concentrated group are most often also cases of sustainable peace.
However, this particular combination of conditions is found in only a small
percent of all cases of sustainable peace. (Instead, sustainable peace is
usually related to other conditions).
As noted above, Bangladesh (CHT) is a full member of both PEACE and
ENDALL. However, its membership score in CON is only 0.75, and since
membership in a fuzzy union is based on the minimum, Bangladesh (CHT)’s
membership score in ENDALL*CON is MIN(1.00, 0.75) = 0.75. Again, the
subset relation ENDALL*CON ≤ PEACE holds for a majority of cases in the
analysis,93 but membership in ENDALL*CON is generally very low, which
also makes coverage low. One way of illustrating the low coverage is to note
that Bangladesh (CHT) is the only case with a membership score of 0.75 in
ENDALL*CON, two cases have a membership score of 0.50, nine have a
membership score of 0.25, and the remaining thirty-one cases are fully out of
that configuration.
Several combinations of three conditions also yield consistent solutions.
REPALL, ENDALL, and PART produces the following solution (Formula
10.3):
repall*ENDALL → PEACE
consistency 0.8621, coverage 0.3205
Peru is a full member (repall*ENDALL 1.00/ PEACE 1.00) in this solution,
and two other cases have >0.50 membership in the solution term
repall*ENDALL, namely El Salvador (0.75/1.00) and Guatemala
93 In fact, the Iran (Kurdistan) 1993 termination is the only exception.
159
PEACE BY REPATRIATION
(0.75/1.00). This particular solution, which is not very relevant, suggests that
it is important to end displacement, but that this should not be done through
repatriation. As described above, membership in the absence of a fuzzy
condition equals 1 minus membership in the (presence of the) condition. El
Salvador has a membership score of 0.25 in REPALL and a membership
score of 1.00 in ENDALL. Hence El Salvador’s membership score in the
configuration repall*ENDALL is MIN(1 – 0.25; 1.00) = 0.75.
Testing REPALL, ENDALL, and FRAC again yields the same solution as
in formula 10.1 (and, as long as the same set of cases is tested, consistency
and coverage are also the same). Combining REPALL, ENDALL, and CON
results in a solution consisting of two solution terms (Formula 10.4):
repall*ENDALL + ENDALL*CON → PEACE
consistency 0.8824, coverage 0.3846
When an analysis results in a solution containing two or more solution
terms, this means that the outcome is associated with either of the solution
terms. In this case, PEACE is often the result of either repall*ENDALL or of
ENDALL*CON. This means that cases characterized either by displacement
having been solved, but through other means than repatriation, or by
displacement having been solved after a conflict involving a regionally
concentrated minority are often also cases of peace. The first solution term,
repall*ENDALL, is the same as in formula 10.3, and the second solution
term, ENDALL*CON, is the same as in formula 10.2, again with the same
consistency and coverage for the individual solution terms, and obviously
with the same cases displaying full/strong membership.
A few other combinations of three conditions again reproduce the
solutions presented in formulas 10.1 and 10.2, but no new consistent
solutions are found. The same is true for all combinations of four conditions.
Overall, ending displacement stands out as the most important condition
when all cases and all displaced persons are analyzed.
Shifting our attention to a strict focus on refugees, that is excluding
IDPs, it is again important to test for necessary conditions before conducting
the truth table analysis. The analysis of all displaced persons above found no
necessary conditions. The corresponding tests of only refugees yield lower
consistency scores for all conditions (except repatriation, which increases to
from 0.5769 to 0.6418); there are no necessary conditions here either.
The fuzzy-set analyses of various combinations of conditions produce a
few consistent solutions. Of all combinations of any two conditions, two
produce consistent solutions. First, the test of ENDREF and FRAC yields the
solution (Formula 10.5):
160
CONDITIONS
ENDREF*frac → PEACE
consistency 0.8378, coverage 0.4627
Bangladesh (CHT) has full membership in this solution (1.00/1.00), and
three cases have >0.50 membership in the solution term ENDREF*frac,
namely Yemen (South Yemen) (1.00/0.50), El Salvador (0.75/1.00) and
Cambodia (0.75/0.50). The solution suggests that cases where displacement
has ended after a conflict in a country with a low degree of ethnic
fractionalization are usually cases of sustainable peace. The coverage of this
particular solution is not very high, which means that the solution is not
relevant to most cases.
Second, testing ENDREF and CON yields the solution (Formula 10.6):
ENDREF*CON → PEACE
consistency 0.8750, coverage 0.2090
The high consistency means that cases where displacement has ended after
conflict involving a regionally concentrated minority tend to be associated
with peace, but the low coverage shows that most cases of peace after armed
conflict do not fit the description of displacement having ended after a
conflict involving a regionally concentrated minority. Indeed, Bangladesh
(CHT) is the only case with >0.50 membership in this solution (0.75/1.00).
The corresponding solution was found in the analysis involving all displaced
persons: cases that combine the ending of conflict-induced displacement
with the conflict having involved a regionally concentrated group are most
often cases of sustainable peace, but this particular combination of
conditions is found in only a small percent of all cases of sustainable peace.
The tests of various combinations of three conditions produce only one
consistent solution. ENDREF, FRAC, and CON yields the same solutions as
the test of only ENDREF and CON (cf. formula 10.6). Finally, combining the
four conditions REPREF, ENDREF, FRAC, and CON yields two alternative
consistent solutions,94 namely (Formula 10.7):
94 Alternative solutions to the same truth table analysis are possible when the minimization process (cf. p 92)
leads to more reduced expressions—or ”prime implicants”—than are needed to cover all the original
”primitive expressions.” Ragin (1987, 95–98) illustrates this with the following example. Four of the eight
possible configurations of the three conditions A, B, and C are related with the positive outcome Y, like this:
AbC + aBc + Abc + ABC → Y. These four configurations are the ”primitive expressions” that we want to
minimize in order to achieve as simple a solution as possible. Through minimization, ABC combines with AbC
to produce AC; ABC combines with ABc to produce AB; and ABc combines with aBc to produce Bc. Hence
AC + AB + Bc → Y. AC, AB, and Bc are the ”prime implicants.” However, two prime implicants, AC and Bc,
together cover all four primitive expressions, making AB redundant. When such a situation occurs—as it does
in a test of REPREF, ENDREF, FRAC, and CON—the software allows the researcher to select which prime
implicants to use. As always, this should be done on the basis of theoretical and substantive knowledge, and
because I do not feel confident excluding any prime implicants on the basis of such knowledge, I present both
alternative consistent solutions.
161
PEACE BY REPATRIATION
repref*ENDREF + ENDREF*CON → PEACE
consistency 0.8387, coverage 0.3881
and (Formula 10.8):
repref*ENDREF + REPREF*frac*CON → PEACE,
consistency 0.8000, coverage 0.4179
El Salvador is the only case with strong membership in the first term of both
solutions, repref*ENDREF (0.75/1.00). Bangladesh (CHT), in turn, has
strong membership in the second terms of both solutions (0.75/1.00). The
analysis shows that if a case is either a case of repref*ENDREF or a case of
ENDREF*CON (formula 10.7), it is usually also a case of PEACE (and the
analogous for formula 10.8).
In sum, limiting the analysis to refugees produces the same main
findings as the analysis including all displaced persons. Of the conditions
analyzed, ending displacement is the most important condition for peace.
These findings are summarized in table 10.3. What is interesting so far is
that repatriation appears to play a rather marginal role in sustaining peace,
and the importance of partition is even smaller. Ending displacement is part
of almost every consistent solution; repatriation is part of a few (in both
present and absent forms), and partition is not part of a single consistent
solution.
Table 10.3
Consistent solutions—all conflicts.
Solution
ENDALL
ENDALL*CON
repall*ENDALL
repall*ENDALL +
ENDALL*CON
Solution
ENDREF*CON
ENDREF*frac
repref*ENDREF +
ENDREF*CON
repref*ENDREF +
REPREF*frac*CON
162
All conflicts, all displaced persons
Consistency
→ PEACE
0.8226
→ PEACE
0.8750
→ PEACE
0.8621
→ PEACE
0.8824
All conflicts, refugees only
Consistency
→ PEACE
0.8750
→ PEACE
0.8378
Coverage
0.6538
0.1795
0.3205
0.3846
Coverage
0.2090
0.4627
→ PEACE
0.8387
0.3881
→ PEACE
0.8000
0.4179
CONDITIONS
Thus far, the tests have been conducted on all cases (forty-three cases
when all displaced persons are analyzed, and thirty-eight when only refugees
are considered). However, a key rationale for the peace-by-repatriation
thesis is the need to remix ethnic groups after armed conflict. This makes the
thesis particularly pertinent to conflicts over territory. Similarly, partition
theory has an explicit focus on conflicts over territory. In the next section,
therefore, I repeat the analyses conducted above, but only include the
twenty-two cases of conflict over territory.
10.3 Conflicts over Territory
Excluding now conflicts over government and concentrating on conflicts
over territory, I begin by testing all conditions for necessity. This yields a
consistency score of 0.8919 for group concentration (the other conditions get
lower scores). Group concentration is by definition higher in conflicts over
territory than in conflicts over government (and for necessary conditions, the
outcome should be a subset of the condition). Therefore, it is no surprise that
group concentration gets a higher consistency score as a necessary condition
when the analysis is limited to conflicts over territory. The consistency of
group concentration as a necessary condition for peace is high, but it is not
0.90 or 0.95. I will therefore retain it in the analysis of conflicts over
territory.
When all displaced persons are considered, the analysis of conflicts over
territory results in two consistent solutions, both of which were also found in
the analysis of all cases. The first contains one condition (Formula 10.9):
ENDALL → PEACE,
consistency 0.8095, coverage 0.4595
and the other contains two conditions (Formula 10.10):
ENDALL*CON → PEACE,
consistency 0.8750, coverage 0.3784
Bangladesh (CHT) is the only case (among conflicts over territory) with
strong membership in either solution (1.00/1.00 in ENDALL → PEACE and
0.75/1.00 in ENDALL*CON → PEACE).
Combining the analysis of conflicts over territory with an exclusive focus
on refugees yields four consistent solutions. One singles out the need to end
displacement (Formula 10.11):
ENDREF → PEACE,
consistency 0.8095, coverage 0.5313
163
PEACE BY REPATRIATION
Table 10.4
Consistent solutions—conflicts over territory.
Conflicts over territory, all displaced persons
Solution
Consistency
ENDALL → PEACE
0.8095
ENDALL*CON → PEACE
0.8750
Solution
ENDREF
ENDREF*CON
REFREF*frac*CON
ENDREF*CON +
REPREF*frac*CON
Conflicts over territory, refugees only
Consistency
→ PEACE
0.8095
→ PEACE
0.8750
→ PEACE
0.8000
→ PEACE
0.8000
Coverage
0.4595
0.3784
Coverage
0.5313
0.4375
0.3750
0.5000
Bangladesh (CHT) has full membership in this solution (1.00/1.00) and
Yemen (South Yemen) (1.00/0.50) is the only other case with >0.50
membership in the solution term ENDREF. Another solution was found also
in the analysis of all cases (limited to refugees), namely (Formula 10.12):
ENDREF*CON → PEACE,
consistency 0.8750, coverage 0.4375
A third solution constituted one of the solution terms in a solution with two
terms (cf. formula 10.8), namely (Formula 10.13):
REPREF*frac*CON → PEACE,
consistency 0.8000, coverage 0.3750
The two solutions in formulas 10.12 and 10.13 also combined to form one
consistent solution with two solution terms (Formula 10.14):
ENDREF*CON + REPREF*frac*CON → PEACE
consistency 0.8000, coverage 0.5000
Bangladesh (CHT) is the only case with strong membership in either of these
solutions. Table 10.4 summarizes the analyses of conflicts over territory.
10.4 Conflicts over Government
Finally, to make the analysis complete, the same tests are run exclusively on
the twenty-one cases of conflicts over government. No condition is necessary
for peace in an analysis restricted to conflicts over government. ENDALL
gets the highest consistency, scoring 0.8293.
164
CONDITIONS
However, two other conditions are excluded from the analysis, namely
partition and group concentration. As these concepts are operationalized in
this study (cf. chapter 8), they are irrelevant to an analysis of conflicts over
government, and all of them are fully out of the set of “group concentration.”
The analysis of conflicts over government produces three consistent
solutions. The first was also found in the analysis of all conflicts as well as in
the analysis of conflicts over territory (Formula 10.15):
ENDALL → PEACE,
consistency 0.8293, coverage 0.8293
El Salvador, Guatemala, Mozambique, and Peru are full members
(1.00/1.00) of this solution. Another three have >0.50 membership in
ENDALL (cf. formula 10.1), namely Guinea-Bissau (1.00/0.50), Cambodia
(0.75/0.50), and Ethiopia (0.75/0.50). The second solution was also found
in the analysis of all cases, but not in the analysis of conflicts over territory
(Formula 10.16):
repall*ENDALL → PEACE,
consistency 0.8889, coverage 0.3902
Peru is a full member (1.00/1.00) of this solution, and El Salvador and
Guatemala are strong members (both 0.75/1.00). A third solution, not found
in the previous analyses was (Formula 10.17):
repall*frac → PEACE,
consistency 0.8667, coverage 0.3171
El Salvador is the only case with >0.50 membership in this solution
(0.75/1.00), and, unsurprisingly, coverage is low. The solution suggests that
after conflicts in countries characterized by low fractionalization, displaced
persons should not return home.
Finally, limiting the analysis of conflicts over government to refugees
yields one consistent solution (Formula 10.18):
REPREF*frac → PEACE,
consistency 0.8235, coverage 0.4000
One case has >0.50 membership in this solution, namely Cambodia
(0.75/0.50). It suggests that refugees should return home after conflicts in
countries with a low degree of fractionalization. This solution largely
contradicts the one found in formula 10.17, which suggested that displaced
persons—refugees and IDPs considered together—should not return home
165
PEACE BY REPATRIATION
Table 10.5
Consistent solutions—conflicts over government.
Conflicts over government, all displaced persons
Solution
Consistency
ENDALL → PEACE
0.8293
repall*ENDALL → PEACE
0.8889
repall*frac → PEACE
0.8667
Coverage
0.8293
0.3902
0.3171
Conflicts over government, refugees only
Solution
Consistency
REPREF*frac → PEACE
0.8235
Coverage
0.4000
after the end of such conflicts. The results of the analysis of conflicts over
government are summarized in table 10.5.
10.5 Findings
The most consistent result of the analyses presented in this chapter is the
importance for sustainable peace of ending displacement. This importance
can be illustrated in several ways.
First, ENDALL/ENDREF is the only condition that comes out on its own
as a consistent subset of PEACE. REPALL/REPREF, FRAC, and CON appear
in solution terms only in combination with another condition. Second, either
ENDALL or ENDREF are part of eleven of the thirteen different solutions
produced by the various analyses, more than any other condition. Third,
these eleven solutions all contain ENDALL or ENDREF as present (upper
case letters) conditions; the absence of either ENDALL or ENDREF is not
part of any solution. REPALL/REPREF is part of eight solutions—REPREF
appears variously as present or absent, while REPALL is always absent. The
absence of FRAC is part of six solutions, and the presence of CON is part of
seven solutions. It is worth noting that PART does not appear in a single
consistent solution, including tests restricted to conflicts over territory.
Figure 10.3 illustrates the subset relation ENDALL ≤ PEACE, which
(unlike the one in figure 10.2) is a consistent subset relation and an actual
solution of several truth table analyses—most cases are clearly on or above
the diagonal.
As already mentioned, various combinations of the same nine cases—
Bangladesh (CHT), Cambodia, El Salvador, Ethiopia, Guatemala, GuineaBissau, Mozambique, Peru, and Yemen (South Yemen)—have >0.50
membership in the solutions produced in the tests of the peace-by-
166
0.00
0.25
Sri Lanka (Eelam)
Tajikistan ‘96
Croatia (Serb) ‘93
Myanmar (Karen) ‘95
Afghanistan
0.25
Figure 10.3 XY plot of membership in PEACE / ENDALL.
PEACE
0.00
Iran (Kurd) ‘90,
Iraq (Kurd) ‘93
Russia (Chechnya)
Congo
DRC
Liberia
Rwanda
Myanmar (Karen) ‘92
Philippines
(Mindanao)
Angola
0.50
Yugoslavia (Kosovo)
BiH (Croat)
BiH (Serb)
Ethiopia (Eritrea)
Iraq (Kurd) ‘96
Lebanon
Tajikistan ‘98
Croatia (Serb) ‘95
Azerbaijan (N-K)
Georgia (Abkhazia)
Morocco (W Sahara)
Myanmar (Kachin)
Iraq
0.75
1.00
ENDALL
0.50
Iran (Kurd) ‘93
Chad
Uganda
Iran (Kurd) ‘96
Sierra Leone
Nicaragua
0.75
Cambodia
Ethiopia
1.00
Yemen (S Yemen)
Guinea-Bissau
Bangladesh (CHT)
El Salvador
Guatemala
Mozambique
Peru
CONDITIONS
repatriation thesis conducted above. No other case has >0.50 membership in
any consistent solution.95
The five cases in the upper-right corner of figure 10.3 are full members
of both ENDALL and PEACE. Four of them are conflicts over government
95 Incidentally, four of these cases (Cambodia, Ethiopia, Guinea-Bissau and Yemen) belong to the roughly
twenty percent of cases for which the subset relation ENDALL ≤ PEACE does not hold. Still, the software
highlights them because they have >0.50 membership in some of the consistent solutions.
167
PEACE BY REPATRIATION
and one is a conflict over territory. This is an important reason for the
different coverage scores of the solution ENDALL → PEACE for conflicts
over territory and conflicts over government, respectively (tables 10.4 and
10.5). Of all consistent solutions found, ENDALL → PEACE in conflicts over
government has by far the greatest coverage. Conversely, the case Croatia
(Serb) 1995 is fully in PEACE and fully out of ENDALL, making a significant
contribution to the lower coverage of the solution ENDALL → PEACE in
conflicts over territory.
The analysis also highlights the difference between fractionalization and
group concentration. At the operationalization stage it was argued that they
represent different dimensions of ethnic heterogeneity. However, FRAC only
appears as absent and CON only as present when they are part of consistent
solutions, indicating that fractionalization and group concentration are less
related than might have been expected. As noted above, group concentration
is not relevant to conflicts over government, at least not as the concept is
operationalized in this study— however, fractionalization is. Interestingly,
conflicts over government have higher average membership scores in the
fuzzy set of “fractionalization” than do conflicts over territory: 0.71 and 0.50,
respectively. Indeed, seven conflicts over government are full members of
that set, but not a single conflict over territory is. In the end, however, this
does not affect the main finding that ending displacement is the most
important condition.
Finally, it should be emphasized that the solutions presented in tables
10.3–5 do not reflect to the only combinations of conditions that have been
tested. Instead, they are all the consistent solutions. See Appendix 5 for a
more extensive summary of these tests.
10.6 Modifying the Analysis
As discussed in chapter 8 the concepts used as conditions and outcome in
this study have not been operationalized as fuzzy sets before. For example, I
could have focused on other aspects of the first-level concepts, or selected
different thresholds to define membership values in the various fuzzy sets. I
am the first to admit that alternative operationalizations are conceivable.
However, having devoted quite some time to operationalizing the concepts
and coding the cases I believe the operationalizations used in this study are
reasonable in the context that they are used. Nevertheless, in light of the
exploratory character of this study it is worth conducting a few alternative
tests. These involve alternative operationalizations of sustainable peace,
ended displacement situation, and fractionalization.
168
CONDITIONS
Alternative Operationalization of Sustainable Peace
The fuzzy membership value 0.50 is considered problematic by some
researchers. At the end of chapter 7 I explained that I consider it equally
problematic to exclude the 0.50 value even for the most ambiguous cases.
Nevertheless, a look at table 9.3 shows that seventeen of forty-three
cases in this study have a membership score of 0.50 in the outcome
“sustainable peace.” This is not surprising, considering the elusive nature of
the concept of peace. (It is perhaps more surprising that I have dared to code
some cases as fully in that set.) Still, almost forty percent of cases undecided
on outcome is a lot. This section asks whether reducing that share affects the
result of the analysis.96
I retain the fuzzy membership value of 0.50, but reduce the number of
cases that are assigned that value. In the original operationalization, the 0.50
value comprised cases where sovereign peace (defined as the absence of
armed conflict) had lasted for at least five years, but which had a Polity score
of less than +6. Cases with a shorter period of sovereign peace received a
lower membership score in the set, and cases with a Polity score of +6 or
higher received a higher membership score.
The alternative operationalization separates these cases only on the basis
of their Polity score, and employs a lower threshold. As before, cases where
conflict recurred within two years are fully out of the set of “sustainable
peace”; only cases where sovereign peace lasted for at least two years have
nonzero membership. Among these cases, those with a Polity score of 0 or
lower are assigned a membership score of 0.25 in the set, and those with a
Polity score of +1 or higher are assigned a membership score of 0.75 in the
set. The requirements for full membership are the same as with the original
operationalization.97
Coding the cases according to this alternative operationalization changes
the membership scores of fourteen cases; seven change from 0.50 to 0.25
and seven change from 0.50 to 0.75. This leaves three cases in the 0.50
category, —Bosnia-Herzegovina (Croat), Bosnia-Herzegovina (Serb), and
Lebanon—cases where Polity scores are missing for the relevant years. Fuzzy
membership scores according to the alternative coding of sustainable peace
are presented in Appendix 2, column “PEACE 2.”
I will not go into detail about these alternative tests. They are simply
conducted to see whether alternative operationalizations fundamentally
96 With reference to the discussion at the end of chapter 7, it should be noted that only cases with a
membership of 0.50 in one or more conditions are excluded from the creation of the truth table. Cases are not
excluded from that procedure on the basis of their membership in the outcome. The exercise in the present
section, therefore, is not a response to that particular critique of the 0.50 value.
97 The difference between the Polity score and the Polity2 score, again, does not affect the coding of cases
according to the alternative operationalization of sustainable peace.
169
PEACE BY REPATRIATION
Table 10.6
Consistent solutions—alternative operationalization of “sustainable peace.”
Solution
ENDALL
ENDALL*FRAC
ENDALL*FRAC +
ENDALL*CON +
repall*ENDALL
ENDALL*FRAC +
repall*ENDALL +
REPALL*FRAC*CON
ENDALL*FRAC +
ENDALL*CON +
repall*frac*con
All conflicts, all displaced persons
Consistency
→ PEACE 2
0.8548
→ PEACE 2
0.8667
Coverage
0.6795
0.5000
→ PEACE 2
0.8846
0.5897
→ PEACE 2
0.8393
0.6026
→ PEACE 2
0.8095
0.6538
change the results of the original analysis. In the case of PEACE 2, they do
not.
The analysis, summarized in table 10.6, yields a few individual solution
terms that do not contain the presence of ENDALL, but these solution terms
have rather low coverage—particularly unique coverage (cf. Appendix 5).
Also, they are part of consistent solutions where ENDALL appears as the
most important solution term. Of the twenty-two combinations of conditions
tested, ten analyses yield the parsimonious solution ENDALL → PEACE.
Alternative Operationalization of Ended Displacement Situation
Of the conditions included in this study, “ended displacement situation”
stands out as the most important condition for sustainable peace. The subset
relation ENDALL ≤ PEACE holds for around eighty percent of all cases. For
a condition to constitute a subset of the outcome, membership in the
condition should be equal to or lower than membership in the outcome.
Does this mean that the lower the average value of memberships in a
condition, the more likely it is to constitute a subset of the outcome? A look
Table 10.7
Average membership scores.
PEACE
REPALL
ENDALL
PART
FRAC
CON
170
Average membership
0.45
0.50
0.36
0.22
0.60
0.41
Standard deviation
0.30
0.42
0.34
0.33
0.25
0.44
CONDITIONS
Table 10.8
Alternative operationalization of the fuzzy set of “ended displacement situation.”
Value
1.00
0.75
0.50
0.25
0.00
Operationalization
< 25,000 remaining displaced persons, or cessation clause invoked
approximately 25,000 remaining displaced persons
25,000–100,000 remaining displaced persons
> 100,000 remaining displaced persons, and any-solution-rate ≥50 percent
> 100,000 remaining displaced persons, and any-solution-rate < 50 percent
at the average membership scores in the conditions and in the outcome
analyzed in this study (table 10.7) reveals that membership in ended
displacement situation is comparatively low. However, it is worth noting that
the average membership in ENDALL is not as low as that of PART, which
does not show up in any consistent solutions. This shows that it is also
important in which cases membership scores are high or low.
However, the fact that a very low average membership in PART does not
translate into partition being an important condition in the analysis does not
preclude the possibility that the comparatively low average membership in
ENDALL unduly influences the results. Therefore, the analysis will be
repeated with an alternative operationalization of ended displacement
situation, which increases average membership.
The original operationalization of ended displacement situation took the
concept of “major, protracted refugee situation” as a starting point, arguing
that a major refugee situation is clearly not ended. This meant that cases
where more than twenty-five thousand people remained displaced after five
years were automatically considered more out than in the fuzzy set of ended
displacement situation (cf. table 8.5). However, considering the magnitude
of many displacement situations this might have been asking too much. A
more generous operationalization (which I nevertheless find theoretically
less satisfying) would be that a major refugee situation is clearly not fully
ended. With the original operationalization, a case with more than twentyfive thousand remaining displaced persons would receive a membership
score of no more than 0.25 in the fuzzy set of “ended displacement
situation.” With the alternative operationalization a membership of 0.25 is
possible for cases with a remaining displaced population four times larger.
The alternative operationalization of “ended displacement situation” is
presented in table 10.8.
Coding the cases according to this alternative operationalization
increases the membership scores of twenty cases, nearly half of all cases in
the analysis—eighteen cases are assigned higher membership scores in
ENDALL 2 than they were in ENDALL, and fifteen are stronger members in
ENDREF 2 than in ENDREF. Average membership scores in the fuzzy set of
“ended displacement situation” increases to 0.49, similar to that of
171
PEACE BY REPATRIATION
Table 10.9
Consistent solutions—alternative operationalization of “ended displacement situation.”
Solution
ENDALL 2*frac
REPALL*ENDALL 2*CON
ENDALL 2*frac*CON
repall*ENDALL 2*frac
All conflicts, all displaced persons
Consistency
→ PEACE
0.8085
→ PEACE
0.9167
→ PEACE
0.8947
→ PEACE
0.8519
Coverage
0.4872
0.1410
0.2179
0.2949
repatriation (std remains 0.35). Fuzzy membership scores are presented in
Appendix 2, columns “ENDALL 2/ ENDREF 2.”
The presence of ENDALL 2 appears in every consistent solution. Some
consistency scores are unusually high (for this study). The consistency of the
subset relation REPALL* ENDALL 2*CON ≤ PEACE is over 0.90. However,
the coverage of the various solutions are low or even very low. One case,
Bangladesh (CHT), has a membership score of 0.75 in the configuration
REPALL*ENDALL 2*CON, nine cases have membership scores of 0.25, and
the remaining thirty-three cases are fully out of that particular configuration.
Such a solution is far from relevant. In sum, the results show that the
alternative operationalization of ended displacement situation does not
undermine the results of the original analysis.
Alternative Operationalization of Fractionalization
As noted in chapter 8, there are different ways of describing and measuring
the ethnic structure of countries. In the original analysis, I used data from
Alesina et al (2003), which, while based on more recent sources than the
Atlas Narodov Mira, are calculated as the traditional ELF index, which is
widely used in the literature.
An interesting alternative to the ELF index is introduced by Fearon
(2003). Fearon notes that the traditional measure does not take into account
how different various ethnic group are. He illustrates this by comparing
Belarus with Cyprus. With seventy-eight percent Belarusians, thirteen
percent Russians, four percent Poles, and three percent Ukrainians, Belarus
receives a fractionalization score of 0.37.98 Cyprus’s fractionalization score is
Table 10.10
Consistent solutions—alternative operationalization of “fractionalization.”
All conflicts, all displaced persons
Solution
Consistency
ENDALL → PEACE
0.8226
98 1 - (0.782 + 0.132 + 0.042 + 0.032) = 0.3722
172
Coverage
0.6538
CONDITIONS
practically the same, 0.36, based on seventy-eight percent Greeks and
eighteen percent Turks.99 However,
[Belarusians], Ukrainians, and Russians are quite similar in terms of religion,
language, and customs, and Poles speak a Slavic language and share many of the same
customs. By contrast, Greeks and Turks speak languages that come from completely
different families (Indo-European and Altaic), subscribe to two different world
religions (Orthodox Christianity and Islam), and have very different customs (Fearon
2003, 211).
Because the traditional measure of ethnic fractionalization fails to capture
differences such as those between Belarus and Cyprus, Fearon constructs a
measure of “cultural fractionalization,” using the structural divergence
between languages as a proxy for the cultural distance between groups in a
country. The cultural distance between groups is potentially relevant in a test
of the peace-by-repatriation thesis, and it is therefore used as an alternative
operationalization of the condition “fractionalization.” Appendix 2, column
“FRAC 2” presents the fuzzy membership scores. Six cases received higher
membership scores in “FRAC 2” than they did in the original “FRAC,” and
thirteen cases received a lower membership score in “FRAC 2.”100 The
analysis results in one consistent solution, which supports the results of the
original analysis.
Summary
The main finding of the analysis conducted in this chapter is that the subset
relation ENDALL≤PEACE is consistent and fairly relevant. This means that
ending displacement is more important than doing so particularly through
repatriation. In a way, this goes back to Adelman’s Soft I position (see p 14),
that any solution to the displacement situation is a sign that peace is in place,
though the results of this study indicate that “any solution to displacement”
is not merely a sign of peace, but more likely an important contributing
factor in the establishment of peace—more important than repatriation,
anyway.
A total of six cases are full members of the fuzzy set of “sustainable
peace,” namely Bangladesh (CTH), El Salvador, Guatemala, Mozambique,
Nicaragua, and Peru. Five of these are also full members of the fuzzy set of
“ended displacement situation,” but only two of them are full members of the
fuzzy set of “repatriation.” Similarly, as just seen from another angle, seven
cases are full members of the fuzzy set of “ended displacement situation” and
99 1 - (0.782 + 0.182) = 0.3592
100 Regarding the case Yugoslavia (Kosovo), discussed at the end of chapter 9, for FRAC 2 I use data for
”Yugoslavia” on line 7 under Eastern Europe and the Former Soviet Union in Fearon’s appendix, which
corresponds to “Serbia/Montenegro” in Alesina’s dataset; ”Yugoslavia” on line 1 in Fearon corresponds to
Alesina’s ”Yugoslavia,” which I do not use.
173
PEACE BY REPATRIATION
in five of these cases the condition constitutes a subset of the outcome.
Twelve cases, nearly twice as many, are full members of the fuzzy set of
“repatriation,” but that condition constitutes a subset of the outcome in only
two of these cases.
In light of the fact that cases with strong membership in outcome and
condition are the most relevant to the theoretical argument in set-theoretic
analysis, this goes a long way toward understanding the importance of
ending displacement as a result of the present analysis—because clearly,
ending displacement has been found to be considerably more important to
achieving sustainable peace than is repatriation.
Further, as noted, four of the five cases with full membership in the most
general solution, ENDALL ≤ PEACE, are conflicts over government. In light
of this finding, it is not surprising that ending displacement comes out as
particularly relevant after conflicts over government (the highest coverage
score of any solution).
Like ENDALL, PART is a subset of sustainable peace in around seventyfive percent of the cases. There are only a few cases with >0.50 membership
in the fuzzy set of “partition” (cf. the low average membership in partition,
table 10.8). Even among conflicts over territory, average membership in the
fuzzy set of “partition” is no higher than 0.43; only forty percent of conflicts
over territory have a membership of >0.50 in partition. However, unlike
ended displacement situation, partition is not a relevant subset.
Despite the findings of this analysis, ending displacement cannot be said
to constitute a sufficient condition for achieving sustainable peace after
armed conflict. Whether peace will be sustainable or not depends on a large
number of factors not considered in this study—power sharing
arrangements, security guarantees, justice and reconciliation processes,
foreign investments and economic recovery to mention a few. The present
analysis says nothing about the importance of addressing displacement in
different ways relative to the importance of such other variables.
However, in light of the common assumption that repatriation is
necessary for peace—the starting point of this study—the crucial finding of
the analysis is that ending displacement is more important for sustainable
peace than is repatriation. The need to end displacement is the most
coherent finding across the various tests conducted in this chapter—whether
we analyze conflicts over government or conflicts over territory, whether we
include all displaced persons or only refugees, and whether I use my original
or alternative operationalizations of peace, fractionalization, and ended
displacement situation. As discussed above there are exceptions, but they are
few and not very relevant. Additionally, several tests yield solutions
consisting of only ended displacement situation. Conversely, the importance
of repatriation is not at all constant across the various tests.
174
CONDITIONS
This means that at the end of armed conflicts that have caused largescale displacement, the sustainability of peace will generally depend less on
whether displaced persons return home than on whether displaced persons
are able to find durable solutions, be it repatriation, local integration, or
resettlement.
175
PEACE BY REPATRIATION
11 |Summary and Conclusions
Newcomers to the field of international relations may be surprised to learn that efforts to
advance human rights are often at odds with attempts to halt violent conflicts. In the
experience of practitioners, however, it is unfortunately a commonplace that the
promotion of human rights and the practice of conflict resolution, while both admirable
endeavors in themselves, are by no means necessarily complementary objectives.
Richard H. Solomon101
11.1 Summary
The point of departure for this study was the assumption that the return of
displaced persons is necessary for the establishment of sustainable peace
after armed conflict. I labeled this assumption the “peace-by-repatriation
thesis” and argued that, unfortunately, it is most often presented as a
statement of fact rather than as a hypothesis to be tested. I then set out to
conduct such a test.
In order to test the peace-by-repatriation thesis, I first needed to make it
more precise. I began by tracing the theoretical foundations of the thesis in
previous research and discerned several concepts of potential relevance to an
analysis of it. I argued that the theoretical foundations of the peace-byrepatriation thesis are weak. However, on the basis of previous research I
was able to construct a preliminary framework, or pre-theory, of peace by
repatriation. To refine the preliminary framework I applied it to two case
studies of conflicts that caused large-scale displacement, namely BosniaHerzegovina and Nagorno-Karabakh. While similar in many respects, the
cases represent two very different perceptions of the relationship between
repatriation and peace, they are, to use Seawright and Gerring’s term,
diverse cases. Adding insights from the case studies to the conclusions from
previous research, I argued that a theoretical framework suitable for a test of
the peace-by-repatriation thesis would entail—in addition to repatriation
and sustainable peace—four other potentially relevant conditions, namely
ended displacement situation, partition, fractionalization, and group
concentration.
Proceeding to a test of the peace-by-repatriation thesis, I began by
arguing that the claim made by the thesis was one of necessary conditions,
best analyzed in set-theoretic terms. I argued that fuzzy-set analysis would
be particularly suitable, and I operationalized the concepts of the theoretical
framework as fuzzy sets. Forty-three conflict terminations were selected and
coded for the fuzzy-set analysis, which clearly showed that repatriation is not
101 Foreword, Mertus & Helsing (2006, ix).
176
CONCLUSIONS
necessary for peace. Further, repatriation did not seem to be of general
importance in combination with other conditions either. Nor was there any
support for the alternative claim that ethnic groups should be kept separated
through partition. Instead, the one condition that turned out to be
consistently relevant to the establishment of sustainable peace was to end
displacement—irrespective of how this was done. Three of the consistent
solutions resulting from the fuzzy-set analysis in chapter 10 have consistency
scores > 0.50, namely ENDALL → PEACE in the analysis of conflicts over
government (0.8292), ENDALL → PEACE in the analysis of all conflicts
(0.6538), and ENDREF → PEACE in the analysis of conflicts over territory
(0.5313). The various conditions were combined in different ways, separate
tests were conducted on conflicts over territory and conflicts over
government, respectively, and the analysis was repeated with alternative
operationalizations of some of the concepts. None of this changed the main
findings, that repatriation is not necessary, but that ending displacement
seems to be an important condition for peace.
11.2 Conclusions
This study raises serious doubts about the generality of the peace-byrepatriation thesis. The systematic analysis in chapter 10 found no support
for the necessity of repatriation. This supports Adelman’s conclusion that no
one relationship between repatriation and peace is valid across all cases. This
is not to deny that repatriation can indeed be important for peace in some
cases, depending on the circumstances. The case studies suggest that prewar interethnic relations are critical in this regard.
Expressed in terms of Adelman’s positions, the fuzzy-set analysis does
not support the unidirectional Hard I position, which holds that reptriation
is a necessary condition for peace. Instead, the hard version of the Soft I
position, that ending displacement is important for peace, comes out strong
in the analysis. This means that the grand rhetoric of the repatriation culture
does not reflect the empirical relationship between displacement and peace.
This conclusion can be developed a bit further with a look back at the
case studies. Bosnia is the clearest example of the peace-by-repatriation
rhetoric, and the case study suggested that repatriation indeed has an
important role to play in the consolidation of peace there. Conversely,
repatriation is not likely to be very important for peace in the case of
Nagorno-Karabakh, where the the predominant perception of the
relationship between repatriation and peace is Adelman’s Soft II position,
that repatriation will be a sign that peace is in place.
However, a few observations in the latter case also hint at the main
finding from the fuzzy-set analysis, that any solution to displacement is
important for peace. Some of the interviewees noted that the issue of
Nagorno-Karabakh is not as high on the political agenda in Armenia as it is
177
PEACE BY REPATRIATION
in Azerbaijan. Part of the explanation for this might be the fact that the
Armenian IDPs have largely found durable solutions and that the refugees—
that is, Armenian refugees from Azerbaijan—are in the process of doing so.
In Azerbaijan, the issue of internal displacement is much more salient. At the
same time, Armenian control over most of the disputed territory also helps
explaining why the issue is more acute to Azerbaijanis.
Regarding UNHCR’s explanations of how repatriation contributes to
peace, I argued initially that two are clearly related to the particular solution
of repatriation, namely that returning refugees vote with their feet to validate
the post-war political order and that they contribute to rebuilding their
country. Refugees returning in time to vote in the first post-war elections in
Cambodia is an example of the former, and the Return of Qualified Nationals
in Bosnia is an example of the latter. The other two explanations—the need
to restore protection and the issue of refugee warriors—are not specifically
related to repatriation.
Reversing Ethnic Cleansing
In addition to these explanations, the reversal of ethnic cleansing and the
remixing of ethnic groups were described as critical for the establishment of
sustainable peace, most notably in the case of Bosnia. In this way,
repatriation is both an important objective in its own right and a condition
for achieving other ends.
The case of Bosnia illustrates how return can represent a right of the
individual displaced person. As Lord Ashdown exclaimed: “We’ve invented a
new human right here, the right to return after a war.” Similarly, the
approach adopted to bring about return to and within Bosnia—as well as to
promote property repossession—was to treat the issue as a question of the
rule of law. In chapter 4 I quoted Prettitore, who argues that serious progress
on return and repossession came about only when these processes were
grounded in the rule of law and no longer subject to political agreements. In
analyses of the international involvement in Bosnia this is one of the
important lessons. Hopefully it has been learned.
There is an element of this line of reasoning in the case of NagornoKarabakh, too. The international community’s support in principle for the
return of displaced persons is based on a perception of such return as a right,
once it becomes possible. Further, it is clear from the case study in chapter 5
that in the case of Nagorno-Karabakh, this right to return is unrelated to the
peace-by-repatriation thesis.
The success of the rule-of-law approach to the Bosnian return process
should not hide the fact that the return of displaced persons—and hence,
unavoidably, displaced persons themselves—have also been used as means to
an end. Nevertheless, the case bears witness to a promising ambition to treat
return as a right. The rule-of-law approach to repatriation means that any
178
CONCLUSIONS
person who has been forced from his or her home during armed conflict has
an inherent right to reclaim that property and to return voluntarily and in
dignity to live there—irrespective of whether this contributes to a particular
political goal, including the establishment of sustainable peace.
The case study of Bosnia also highlighted how repatriation can be a
condition for successful reconciliation, an aspect that was not equally
relevant in the case of Nagorno-Karabakh. The argument was that
reconciliation and the rebuilding of trust between the ethnic groups require
renewed intermixing. From this perspective, reversing ethnic cleansing is not
only a goal in its own right, but also a condition for reconciliation. The high
level of social interaction among the ethnic groups in pre-war Bosnia was
presented as essential to this line of reasoning. Conversely, if ethnic groups
led separate lives before a conflict, even when they happened to inhabit the
same parts of a country, there are fewer former social bonds to restore and
build peace upon in a reconciliation process. This was the situation in
Nagorno-Karabakh. In such cases, the return of displaced persons will play a
smaller role for the establishment of sustainable peace. These observations
led me to suggest that pre-war interethnic relations are one of the
circumstances that determine whether repatriation will contribute to peace
or not. Interestingly, this argument has not been an important part of the
rhetoric surrounding the need for repatriation to and within Bosnia, nor the
discussions about possible return to Nagorno-Karabakh.
In addition, repatriation leading to ethnic remixing can contribute to
peace by undermining the power of ethnonationalist elites. Fifteen years
after the end of the war Bosnia is still largely characterized by politicians
appealing to nationalist sentiments, reinforcing ethnic stereotypes, and
keeping mutual fear and suspicion alive. Jenne (2009, 285) studies post-war
Bosnia and Kosovo and finds that
de facto partition has ensured the electoral success of nationalist parties and policies;
impeded property restitution and refugee return; permitted rent-seeking and
corruption by nationalist elites; and segregated security and police forces along ethnic
lines—creating a climate of extreme insecurity for ethnic minorities residing in the
“wrong” territory [emphasis in original].
Interviewees argued that the most reasonable way to overcome this is
through remixing the ethnic groups and restoring pre-war relations. The
situation in the case of Nagorno-Karabakh is similar, with deep distrust
between Armenians and Azerbaijanis, and attempts to overcome these
divisions are considered suspect or even treacherous.
In the case study of Bosnia it was argued that temporary protection for
displaced persons makes sense where there are positive pre-war interethnic
relations to rebuild. If alternative durable solutions had been available to
refugees from Bosnia early on, many of them would probably have opted for
179
PEACE BY REPATRIATION
them. As a consequence of this, the argument went on, far fewer would have
been interested in returning once return became a real possibility. Despite
this argument, displaced persons enjoying temporary protection cannot be
held in limbo indefinitely. If repatriation does not become a viable option
within a reasonable period of time the rationality of temporary protection
needs to be reassessed.
Future Research
The results of this study show that repatriation as the default solution to
displacement cannot be justified with reference to the role of repatriation for
building sustainable peace, certainly not to promote the return of displaced
persons to dangerous environments in the name of peace. Whether
repatriation contributes to peace or not will vary from case to case depending
on the circumstances. The case studies indicate that pre-war interethnic
relations are important, but this conclusion cannot easily be generalized
beyond those two.
Future research should deepen the analysis of the peace-by-repatriation
thesis in two ways. First, by developing the theoretical argument about the
relationship between repatriation and reconciliation and, second, by adding
to the systematic knowledge about interethnic relations in specific cases. A
possible starting point for developing the theoretical argument is Lederach’s
notions of grass roots peacebuilding and reconciliation through encounters.
Lederach (1997, 24) calls for a paradigmatic shift “away from a concern with
the resolution of issues and toward a frame of reference that focuses on the
restoration and rebuilding of relationships.” Similar concerns have been
voiced by Prendergast and Plumb (2002, 327), who argue that
in order for peace agreements between warring parties to lead to durable peace, there
needs to be, alongside the top-down implementation of the peace agreement,
concurrent bottom-up processes aimed at constructing a new social contract and
healing societal divisions.
The implication for this for the peace-by-repatriation thesis is that in the
absence of repatriation, there will be no encounters, and hence relationships
cannot be renewed.
An alternative theoretical perspective is feminist peace research. Most
important in this context, feminist peace research calls for a focus on other
issues than the ones at the center of armed conflict and the involvement of
other actors than the primary warring parties. In important respects, the
fields of grass roots peacebuilding and feminist peace research make similar
recommendations—especially regarding the issues relevant to the findings of
this study. Where feminist peace research calls for the participation of
women, grass roots peacebuilding emphasize the participation of civil society
(Belloni 2001; Orjuela 2003; Richmond 2005).
180
CONCLUSIONS
Jarstad and Sisk (2008, 1) argue that the theory and practice of postconflict peacebuilding emphasizes, inter alia, “the elite and public
negotiation of comprehensive peace agreements.” According to Richmond
(2005, 129), non-governmental actors and agencies play an important role in
peacebuilding because they often have “unparalleled access to conflict zones,
far beyond those actors which form part of the official political, economic
and development discourse.” Similar roles for displaced persons are
suggested by Fagen (2009) and Newman (2003).
In addition, more empirical research is important, beginning with case
studies, structured in a way that allows systematic comparison. The case
studies of Bosnia and Nagorno-Karabakh suggest that the relationship
between repatriation and peace is influenced by pre-war interethnic
relations, but it is difficult to test this in a systematic way. What is needed is
information on how ethnic groups interacted before a conflict, and how
people intermix in the aftermath of conflict. Global datasets based on census
data cannot provide such information, if only because censuses are not
conducted even on a decennial basis. Further, existing datasets on ethnic
intermixing—whether various forms of data on fractionalization or the MAR
data on group concentration—tell us little, if anything, about the quality of
relations between ethnic groups. Quantitative studies, or systematic studies
of a medium number of cases, will be important once there is enough
systematically collected data on relevant variables; presently there is not. I
believe, therefore, that the next step in the furtherance of research on the
peace-by-repatriation thesis needs to be case studies based on a common
format. Such studies will be a first indication of whether and how the
findings from the two cases studies of Bosnia and Nagorno-Karabakh apply
to a broader empirical domain, thereby further refining the theoretical
argument. In addition, such studies will help pinpointing what kinds of data
are needed to compile relevant datasets that would allow more quanitative
analyses of the peace-by-repatriation thesis.
In this regard, I consider Fearon’s measure of cultural fractionalization
to be more to the point than the traditional ELF measure. This can be seen in
the cases of Bosnia-Herzegovina and Nagorno-Karabakh. According to the
ELF measure Bosnia is more fractionalized than Azerbaijan, with
fractionalization scores of, respectively, 0.681 and 0.188. Conversely, the
cultural fractionalization scores are 0.146 for Bosnia and 0.187 for
Azerbaijan. However, cultural fractionalization is still related to the state,
just like the ELF measure, and group size remains important to the
fractionalization score. Furthermore, it does not reflect pre-war interethnic
relations.
Finally, I want to stress again that while this study seriously questions
repatriation as a blanket solution to displacement, it should not be taken as
making a case against the return of displaced persons in the aftermath of
181
PEACE BY REPATRIATION
armed conflict. The potential contribution of return to the establishment of
sustainable peace might vary considerably from one case to the other, but
that contribution is only one of many conceivable reasons to promote return.
The right of displaced persons to return is possibly an example of what
Wallensteen (2007, 11) refers to when he argues that certain standards are
being established for internationally accepted peace agreements. Indeed, as
has been noted, provisions for return are included in most major peace
agreements of late. The rule-of-law approach adopted to promote return and
property repossession in Bosnia is an important model, even though it was
motivated by the peace-by-repatriation thesis. It is essential, however, that
such a right of displaced persons to return is not misrepresented as an
obligation to return.
182
REFERENCES
References
Written Material
Abiri, Elisabeth. 2000. The Securitisation of Migration: Towards an
Understanding of Migration Policy Changes in the 1990s. The Case of
Sweden, Göteborg: Department of Peace and Development Research.
Adelman, Howard. 1992. “Ethnicity and Refugees.” In World Refugee
Survey 1992, Washington, D.C.: U.S. Committee for Refugees.
Adelman, Howard. 2002. “Refugee Repatriation.” In Stephen John Stedman,
Donald Rothchild & Elizabeth M. Cousens (eds.), Ending Civil Wars: The
Implementation of Peace Agreements, Boulder: Lynne Rienner, 273–302.
Alesina, Alberto, et al. 2003. “Fractionalization.” Journal of Economic
Growth, 8(2), 155–194.
Alexseev, Mikhail A. 2006. Immigration Phobia and the Security Dilemma:
Russia, Europe, and the United States, Cambridge: Cambridge University
Press.
Ali, Taisier M. & Robert O. Matthews. 2004. “Conclusion: The Long and
Difficult Road to Peace.” In Taisier M. Ali & Robert O. Matthews (eds.),
Durable Peace: Challenges for Peacebuilding in Africa, Toronto: University
of Toronto Press, 393–425.
Ambroso, Guido. 2006. “The Balkans at a Crossroads: Progress and
Challenges in Finding Durable Solutions for Refugees and Displaced Persons
from the Wars in the Former Yugoslavia.” New Issues in Refugee Research,
No. 133, <http://www.unhcr.org>.
Arbitral Tribunal for Dispute Over Inter-Entity Boundary in Brčko Area.
1999. “Final Award.” <http://www.ohr.int>.
Baker, Pauline H. 2001. “Conflict Resolution versus Democratic Governance:
Divergent Paths to Peace?” In Chester A. Crocker, Fen Osler Hampson &
Pamela Aall (eds.), Turbulent Peace: The Challenges of Managing
International Conflict, Washington, D.C.: United States Institute of Peace
Press, 753–764.
Barnett, Michael & Martha Finnemore. 2004. Rules for the World.
International Organizations in Global Politics, Ithaca & London: Cornell
University Press.
183
PEACE BY REPATRIATION
Barrett, David B (ed.). 1982. World Christian Encyclopedia, New York:
Oxford University Press.
Bayefsky, Anne & Michael W. Doyle. 1999. Emergency Return: Principles
and Guidelines, Princeton: Center of International Studies, Princeton
University.
Bell, Christine. 2000. Peace Agreements and Human Rights, Oxford: Oxford
University Press.
Belloni, Roberto. 2001. “Civil Society and Peacebuilding in Bosnia and
Herzegovina.” Journal of Peace Research, 38(2), 163–180.
Berdal, Mats. 2004. “Bosnia.” In David M. Malone (ed.). The UN Security
Council: From the Cold War to the 21st Century, Boulder & London: Lynne
Rienner, 451–466.
Berg-Schlosser, Dirk & Gisèle De Meur. 2002. “Reduction of Complexity.” In
Dirk Berg-Schlosser & Jeremy Mitchell (eds.), Authoritarianism and
Democracy in Europe, 1919–1939: Comparative Analyses, Basingstoke:
Palgrave Macmillan, 270–284.
Bhatia, Michael. 2003. “Repatriation under a Peace Process: Mandated
Return in the Western Sahara.” International Journal of Refugee Law,
15(4), 786–822.
Biermann, Frank & Ingrid Boas. 2010. “Preparing for a Warmer World:
Towards a Global Governance System to Protect Climate Refugees.” Global
Environmental Politics, 10(1), 60–87.
Black, Max. 1962. Models and Metaphors: Studies in Language and
Philosophy, Ithaca, NY: Cornell University Press.
Black, Richard. 2001. “Return and Reconstruction in Bosnia-Herzegovina:
Missing Link, or Mistaken Priority?” SAIS Review, 21(2), 177–199.
Black, Richard. 2006. “Return of Refugees: Retrospects and Prospects.” In
Michael Dumper (ed.). Palestinian Refugee Repatriation: Global
Perspectives, London and New York: Routledge, 23–40.
Black, Richard & Khalid Koser (eds.). 1999. The End of the Refugee Cycle?
Refugee Repatriation and Reconstruction, New York and Oxford: Berghahn
Books.
Bonoan, Rafael. 2001. “When Is International Protection No Longer
Needed? The ‘Ceased Circumstances’ Provisions of the Cessation Clauses:
184
REFERENCES
Principles and UNHCR Practice, 1973–1999.” The Rosemary Rogers
Working Paper Series, No. 8, <http://web.mit.edu/CIS/www/migration/
publications.html>.
Brecher, Michael. 1999. “International Studies in the Twentieth Century and
Beyond: Flawed Dichotomies, Synthesis, Cumulation.” International Studies
Quarterly, 43(2), 213–264.
Bridgman, Benjamin. 2008. “What Does the Atlas Narodov Mira Measure?”
Economics Bulletin, 10(6), 1–8.
Brounéus, Karen. 2008. Rethinking Reconciliation: Concepts, Methods, and
an Empirical Study of Truth Telling and Psychological Health in Rwanda,
Uppsala: Department of Peace and Conflict Research.
Byrne, Rosemary, Gregor Noll & Jens Vedsted-Hansen. 2004.
“Understanding Refugee Law in an Enlarged European Union.” European
Journal of International Law, 15(2), 355–379.
Carpenter, Ted Galen. 2009. “Middle East Vortex: An Unstable Iraq and Its
Implications for the Region.” Mediterranean Quarterly, 20(4), 22–31.
Chimni, B.S. (ed.). 2000. International Refugee Law: A Reader, New Delhi,
Thousand Oaks & London: Sage Publications.
Chimni, B.S. 2003. “Post-Conflict Peace-Building and the Return of
Refugees: Concepts, Practices, and Institutions.” In Edward Newman &
Joanne van Selm (eds.), Refugees and Forced Displacement: International
Security, Human Vulnerability, and the State, Tokyo: United Nations
University Press, 195–220.
Chimni, B.S. 2004. “From Resettlement to Involuntary Repatriation:
Towards a Critical History of Durable Solutions to Refugee Problems.”
Refugee Survey Quarterly, 23(3), 55–73.
CIA. 2010. “The World Factbook.” <http://www.cia.gov>. Accessed 21 July,
2010.
Cingranelli, David L. & David L. Richards. 1999. “Measuring the Level,
Pattern, and Sequence of Government Respect for Physical Integrity Rights.”
International Studies Quarterly, 43(2), 407–417.
Cingranelli, David L. & David L. Richards. 2004. The Cingranelli-Richards
(CIRI) Human Rights Database Coder Manual, Version 8.01.04.
Binghamton, NY: Department of Political Science, Binghamton University,
<http://ciri.binghamton.edu>.
185
PEACE BY REPATRIATION
Cingranelli, David L. & David L. Richards. 2007. CIRI Project: Human
Rights Data, Version 2007.04.12 [computer file]. Binghamton, NY:
Department of Political Science, Binghamton University, <http://ciri.
binghamton.edu>.
Cioffi-Revilla, Claudio A. 1981. “Fuzzy Sets and Models of International
Relations.” American Journal of Political Science, 25(1), 129–159.
Clark, Tom. 2004. “Rights Based Refuge, the Potential of the 1951
Convention and the Need for Authoritative Interpretation.” International
Journal of Refugee Law, 16(4), 584–608.
Cockell, John G. 1998. “Peacebuilding and Human Security: Frameworks for
International Responses to Internal Conflict.” In Peter Wallensteen (ed.),
Preventing Violent Conflicts: Past Record and Future Challenges, Uppsala:
Department of Peace and Conflict Research, 203–230.
Cohen, Roberta & Francis M. Deng. 1998a. Masses in Flight. The Global
Crisis of Internal Displacement, Washington D.C.: Brookings Institution
Press.
Cohen, Roberta & Francis M. Deng (eds.). 1998b. The Forsaken People. Case
Studies of the Internally Displaced, Washington D.C.: Brookings Institution
Press.
Cornell, Svante E. 1999. The Nagorno-Karabakh Conflict, Uppsala:
Department of East European Studies.
Cornell, Svante E. 2001. Autonomy and Conflict: Ethnoterritoriality and
Separatism in the South Caucasus—Cases in Georgia, Uppsala: Department
of Peace and Conflict Research.
Cornell, Svante E. 2002. “Autonomy as a Source of Conflict: Caucasian
Conflicts in Theoretical Perspective.” World Politics, 54(2), 245–276.
Costello, Cathryn. 2005. “The Asylum Procedures Directive and the
Proliferation of Safe Country Practices: Deterrence, Deflection and the
Dismantling of International Protection?” European Journal of Migration
and Law, 7(1), 35–69.
Council of the European Union. 2005. “Council Directive 2005/85/EC of 1
December 2005 on minimum standards on procedures in Member States for
granting and withdrawing refugee status.” Official Journal of the European
Union, 48(L 326), 13–34.
186
REFERENCES
Cousens, Elizabeth M. 2001. “Introduction.” In Elizabeth M. Cousens &
Chetan Kumar (eds.), Peacebuilding as Politics: Cultivating Peace in Fragile
Societies, Boulder: Lynne Rienner, 1–20.
Crawley, Heaven. 2005. “Europe Looks to Africa to Solve the ‘Asylum
Problem’.” Forced Migration Review, 23, 23–26.
Dahlman, Carl. 2002. “The Political Geography of Kurdistan.” Eurasian
Geography and Economics, 43(4), 271–299.
Darby, John & Roger Mac Ginty (eds.). 2003. Contemporary Peacemaking:
Conflict, Violence and Peace Processes, New York: Palgrave Macmillan.
Davenport, Christian. 2003. Minorities At Risk Dataset Users Manual
030703, College Park, MD: Center for International Development and
Conflict Management (CIDCM), <http://www.cidcm.umd.edu/inscr/mar/>.
de Waal, Thomas. 2003. Black Garden: Armenia and Azerbaijan through
Peace and War, New York and London: New York University Press.
de Waal, Thomas. 2009. “The Karabakh Trap: Dangers and Dilemmas of the
Nagorny Karabakh Conflict.” Occasional Papers, Conciliation Resources
<http://www.c-r.org>.
Dayton Agreement. 1995. “General Framework Agreement for Peace in
Bosnia and Herzegovina.” <http://www.usip.org/publications/peaceagreements-bosnia-herzegovina>.
Dehdashti, Rexane. 1997. “Nagorno-Karabakh: A Case-Study of OSCE
Conflict Settlement.” In Michael Bothe, Natalino Ronzitti & Allan Rosas
(eds.), The OSCE in the Maintenance of Peace and Security: Conflict
Prevention, Crisis Management and Peaceful Settlement of Disputes, The
Hague, London & Boston: Kluwer Law International, 459–478.
Dehdashti, Rexane. 2000. Internationale Organisationen als Vermittler in
innerstaatlichen Konflikten: Die OSZE und der Berg Karabach-Konflikt,
Frankfurt and New York: Campus Verlag.
del Castillo, Graciana. 2008. Rebuilding War-Torn States: The Challenge of
Post-Conflict Economic Reconstruction, Oxford: Oxford University Press.
Dion, Douglas. 2003. “Evidence and Inference in the Comparative Case
Study.” In Gary Goertz & Harvey Starr (eds.), Necessary Conditions: Theory,
Methodology, and Applications, Lanham: Rowman & Littlefield, 95–112.
187
PEACE BY REPATRIATION
Downs, George & Stephen John Stedman. 2002. “Evaluation Issues in Peace
Implementation.” In Stephen John Stedman, Donald Rothchild & Elizabeth
M. Cousens (eds.), Ending Civil Wars: The Implementation of Peace
Agreements, Boulder & London: Lynne Rienner, 43–69.
Doyle, Michael W. & Nicholas Sambanis. 2000. “International
Peacebuilding: A Theoretical and Quantitative Analysis.” American Political
Science Review, 94(4), 779–801.
Doyle, Michael W. & Nicholas Sambanis. 2006. Making War and Building
Peace: United Nations Peace Operations, Princeton and Oxford: Princeton
University Press.
Druckman, Daniel & Cecilia Albin. 2010. "Distributive Justice and the
Durability of Peace Agreements." Review of International Studies, Doi:
10.1017/S0260210510000549, published online by Cambridge University
Press (print publication to follow).
Dzumhur, Jasminka. 2010. “The Protection of Social and Economic Rights in
Post-Conflict Bosnia and Herzegovina,” presented at the conference
Combating Sexual Violence: How Can We Make a Lasting Impact in the
DRC?, Stockholm, 25 May 2010.
Eastmond, Marita & Joakim Öjendal. 1999. “Revisiting a ‘Repatriation
Success’: The Case of Cambodia.” In Richard Black & Khalid Koser (eds.),
The End of the Refugee Cycle? Refugee Repatriation & Reconstruction, New
York and Oxford: Berghahn Books, 38–55.
Fagen, Patricia Weiss. 2009. “Peace Processes and IDP Solutions.” Refugee
Survey Quarterly, 28(1), 31–58.
Fearon, James D. 2000. “What Is a Language Community?” American
Journal of Political Science, 44(1), 142–155.
Fearon, James D. 2003. “Ethnic anc Cultural Diversity by Country.” Journal
of Economic Growth, 8(2), 195–222.
Fekete, Liz. 2005. The Deportation Machine: Europe, Asylum and Human
Rights, London: Institute of Race Relations.
Feller, Erika. 2009. “Giving Peace a Chance: Displacement and Rule of Law
during Peacebuilding.” Refugee Survey Quarterly, 28(1), 78–94.
Fletcher, Laurel E. & Hervey M. Weinstein. 2002. “Violence and Social
Repair: Rethinking the Contribution of Justice to Reconciliation.” Human
Rights Quarterly, 24(3), 573–639.
188
REFERENCES
Foreign Ministry. 2005. Letter to UNHCR and UNMIK, UD2005/28923/
MAP, Stockholm.
Fortna, Virginia Page. 2003. “Scraps of Paper? Agreements and the
Durability of Peace.” International Organization, 57(2), 337–372.
Frelick, Bill. 1997. “Assistance Without Protection: Feed the Hungry, Clothe
the Naked, and Watch Them Die.” In World Refugee Survey 1997,
Washington, D.C.: USCRI, 24–33.
French, Steven. 1988. “A Green Parrot Is Just as Much a Red Herring as a
White Shoe: A Note on Confirmation, Background Knowledge and the
Logico-Probabilistic Approach.” The British Journal for the Philosophy of
Science, 39(4), 531–535.
Galtung, Johan. 1969. “Violence, Peace, and Peace Research.” Journal of
Peace Research, 6(3), 167–191.
George, Alexander L. & Andrew Bennett. 2004. Case Studies and Theory
Development in the Social Sciences, Cambridge, MA & London: MIT Press.
Gerring, John. 2007. Case Study Research: Principles and Practices,
Cambridge: Cambridge University Press.
Global IDP Project & Norwegian Refugee Council. 2002. Internally
Displaced People: A Global Survey, 2nd ed. London: Earthscan.
Goertz, Gary. 2003. “The Substantive Importance of Necessary Condition
Hypotheses.” In Gary Goertz & Harvey Starr (eds.), Necessary Conditions:
Theory, Methodology, and Applications, Lanham: Rowman & Littlefield,
65–94.
Goertz, Gary. 2006a. “Assessing the Trivialness, Relevance, and Relative
Importance of Necessary and Sufficient Conditions in Social Science.”
Studies in Comparative International Development, 41(2), 88–109.
Goertz, Gary. 2006b. Social Science Concepts: A User’s Guide, Princeton:
Princeton University Press.
Goertz, Gary & William F. Dixon. 2006. “Substitutability, Aggregation, and
Weakest-Link Measures.” In Gary Goertz, Social Science Concepts: A User’s
Guide, Princeton: Princeton University Press, 129–155.
Goertz, Gary & Jack S. Levy. 2007a. “Causal Explanation, Necessary
Conditions, and Case Studies.” In Gary Goertz & Jack S. Levy (eds.),
189
PEACE BY REPATRIATION
Explaining War and Peace: Case Studies and Necessary Condition
Counterfactuals, London & New York: Routledge, 9–45.
Goertz, Gary & Jack S. Levy (eds.). 2007b. Explaining War and Peace: Case
Studies and Necessary Condition Counterfactuals, London & New York:
Routledge.
Goertz, Gary & James Mahoney. 2006. “Concepts in Theories: Two-Level
Theories.” In Gary Goertz, Social Science Concepts: A User’s Guide,
Princeton: Princeton University Press, 237–268.
Goertz, Gary & Harvey Starr (eds.). 2003. Necessary Conditions: Theory,
Methodology, and Applications, Lanham: Rowman & Littlefield.
Goodwin-Gill, Guy S. & Jane McAdam. 2007. The Refugee in International
Law, 2nd ed. Oxford: Oxford University Press.
Gunter, Michael M. 1993. “A de facto Kurdish State in Northern Iraq.” Third
World Quarterly, 14(2), 295–319.
Gunter, Michael M. 2003. “Middle East: The Kurds Struggle for ‘Kurdistan’.”
In Joseph R. Rudolph (ed.), Encyclopedia of Modern Ethnic Conflicts,
Westport, CT & London: Greenwood Press, 151–159.
Gurr, Ted Robert. 2000. Peoples versus States: Minorities at Risk in the
New Century, Washington, D.C.: United States Institute of Peace Press.
Gurses, Mehmet, Nicholas Rost & Patrick McLeod. 2008. “Mediating Civil
War Settlements and the Duration of Peace.” International Interactions,
34(2), 129–155.
Haddad, Emma. 2008. The Refugee in International Society: Between
Sovereigns, Cambridge: Cambridge University Press.
Hampson, Fen Osler. 1996. Nurturing Peace: Why Peace Settlements
Succeed or Fail, Washington, D.C.: United States Institute of Peace Press.
Hanf, Theodor. 1991. “Reducing Conflict through Cultural Autonomy: Karl
Renner’s Contribution.” In Uri Ra’anan et al (eds.), State and Nation in
Multi-Ethnic Societies: The Breakup of Multinational States, Manchester:
Manchester University Press, 33–52.
Hannum, Horst. 1990. Autonomy, Sovereignty, and Self-determination:
The Accommodation of Conflicting Rights, Philadelphia: University of
Pennsylvania Press.
190
REFERENCES
Hannum, Horst. 2006. “Human Rights in Conflict Resolution: The Role of
the Office of the High Commissioner for Human Rights in UN Peacemaking
and Peacebuilding.” Human Rights Quarterly, 28(1), 1–85.
Harbom, Lotta, Stina Högbladh & Peter Wallensteen. 2006. “Armed Conflict
and Peace Agreements.” Journal of Peace Research, 43(5), 617–631.
Harrell-Bond, Barbara E. 1989. “Repatriation: Under What Conditions Is It
the Most Desirable Solution for Refugees?” African Studies Review, 32(1),
41–69.
Hartzell, Caroline & Matthew Hoddie. 2003. “Institutionalizing Peace:
Power Sharing and Post-Civil War Conflict Management.” American Journal
of Political Science, 47(2), 318–332.
Hartzell, Caroline, Matthew Hoddie & Donald Rothchild. 2001. “Stabilizing
the Peace After Civil War: An Investigation of Some Key Variables.”
International Organization, 55(1), 183–208.
Hathaway, James C. 2007. “Forced Migration Studies: Could We Agree Just
to ‘Date’?” Journal of Refugee Studies, 20(3), 349–369.
Hempel, Carl G. 1945a. “Studies in the Logic of Confirmation (I.).” Mind,
54(213), 1–26.
Hempel, Carl G. 1945b. “Studies in the Logic of Confirmation (II.).” Mind,
54(214), 97–121.
Hempel, Carl G. 1946. “A Note on the Paradoxes of Confirmation.” Mind,
55(217), 79–82.
Herz, John H. 1950. “Idealist Internationalism and the Security Dilemma.”
World Politics, 2(2), 157–180.
Hettne, Björn. 1993. “Dynamics of Ethnic Conflict.” In Helena Lindholm
(ed.), Ethnicity and Nationalism: Formation of Identity and Dynamics of
Conflict in the 1990s, Gothenburg: Nordnes, 68–89.
Hicks, Alexander. 1999. Social Democracy and Welfare Capitalism: A
Century of Income Security Politics, Ithaca, N.Y.: Cornell University Press.
Holbrooke, Richard. 1998. To End a War, New York: Random House.
Hollis, Martin & Steve Smith. 1991. Explaining and Understanding
International Relations, Oxford: Clarendon Press.
191
PEACE BY REPATRIATION
Horowitz, Donald L. 1985. Ethnic Groups in Conflict, Berkeley, Los Angeles
& London: University of California Press.
Howard, Lise Morjé. 2008. UN Peacekeeping in Civil Wars, Cambridge:
Cambridge University Press.
HRW. 2009. Pushed Back, Pushed Around: Italy’s Forced Return of Boat
Migrants and Asylum Seekers, Libya’s Mistreatment of Migrants and
Asylum Seekers, New York: Human Rights Watch, <http://www.hrw.org>.
Huntford, Roland. 1997. Nansen: The Explorer as Hero, London:
Duckworth.
Huysmans, Jef. 2002. “The European Union and the Securitization of
Migration.” Journal of Common Market Studies, 38(5), 751–777.
Hyndman, Jennifer. 2000. Managing Displacement: Refugees and the
Politics of Humanitarianism, Minneapolis & London: University of
Minnesota Press.
Hyndman, Jennifer. 2010. “Introduction: The Feminist Politics of Refugee
Migration.” Gender, Place and Culture, 17(4), 453–459.
ICG. 1998a. “A Hollow Promise? The Return of Bosnian Serb Displaced
Persons to Drvar, Bosansko Grahovo and Glamoc.” Bosnia Report 29,
International Crisis Group <http://www.crisisgrop.org>.
ICG. 1998b. “Minority Return or Mass Relocation.” Balkans Report 33,
International Crisis Group <http://www.crisisgrop.org>.
ICG. 1998c. “A Tale of Two Cities: Return of Displaced Persons to Jajce and
Travnik.” Balkans Report 34, International Crisis Group <http://www.
crisisgrop.org>.
ICG. 1999. “Preventing Minority Return in Bosnia and Herzegovina: The
Anatomy of Hate and Fear.” Balkans Report 73, International Crisis Group
<http://www.crisisgrop.org>.
ICG. 2000. “Bosnia’s Refugee Logjam Breaks: Is the International
Community Ready?” Balkans Report 95, International Crisis Group
<http://www.crisisgrop.org>.
ICG. 2002. “The Continuing Challenge of Refugee Return in Bosnia &
Herzegovina.” Balkans Report 137, International Crisis Group
<http://www.crisisgrop.org>.
192
REFERENCES
ICG. 2003. “Bosnia’s Brčko: Getting In, Getting On and Getting Out.”
Balkans Report 144, International Crisis Group <http://www.
crisisgrop.org>.
ICG. 2004. “Macedonia: Make or Break.” Europe Briefing 33, International
Crisis Group <http://www.crisisgrop.org>.
ICG. 2005a. “Nagorno-Karabakh: Viewing the Conflict from the Ground.”
Europe Report 166, International Crisis Group <http://www.
crisisgrop.org>.
ICG. 2005b. “Nagorno-Karabakh: A Plan for Peace.” Europe Report 167,
International Crisis Group <http://www.crisisgrop.org>.
ICG. 2007. “Nagorno-Karabakh: Risking War.” Europe Report 187,
International Crisis Group <http://www.crisisgrop.org>.
IDMC. 2008. “Bosnia and Herzegovina: Quick Facts.” Internal Displacement
Monitoring Centre. http://www.internal-displacement.org (“At a Glance”).
Accessed 20 January, 2010.
IDMC. 2010. Armenia: Need to Monitor Progress towards Durable
Solutions, Geneva: Internal Displacement Monitoring Centre.
IDMC. Undated. “Azerbaijan: Country Page.” <http://www.internaldisplacement.org>. Accessed 3 February, 2010.
IOM. Undated. “IOM Regional: Missions in the Region: IOM in Bosnia and
Herzegovina.” <http://www.iom.hu>. Accessed 2 February, 2010.
Jarstad, Anna K. 2008. “Dilemmas of War-to-Democracy Transitions:
Theories and Concepts.” In Anna K. Jarstad & Timothy D. Sisk (eds.), From
War to Democracy: Dilemmas of Peacebuilding, Cambridge: Cambridge
University Press, 17–36.
Jarstad, Anna K. 2009. “Post-Accord Elections and Armed Conflict.” In
Ashok Swain, Ramses Amer & Joakim Öjendal (eds.), The Democratization
Project: Opportunities and Challenges, London, New York and Delhi:
Anthem Press, 153–163.
Jarstad, Anna K. & Desirée Nilsson. 2008. “From Words to Deeds: The
Implementation of Power-Sharing Pacts in Peace Accords.” Conflict
Management and Peace Science, 25(3), 206–223.
Jarstad, Anna K. & Timothy D. Sisk (eds.). 2008. From War to Democracy:
Dilemmas of Peacebuilding, Cambridge: Cambridge University Press.
193
PEACE BY REPATRIATION
Jenne, Erin K. 2009. “The Paradox of Ethnic Partition: Lessons from de
facto Partition in Bosnia and Kosovo.” Regional & Federal Studies, 19(2),
273–289.
Johansson, Patrik. 2007. “Refugee Repatriation as a Necessary Condition for
Peace.” In Ashok Swain, Ramses Amer & Joakim Öjendal (eds.),
Globalization and Challenges to Building Peace, London, New York and
Delhi: Anthem Press, 91–106.
Johansson, Patrik. 2009a. “The Humdrum Use of Ultimate Authority:
Defining and Analysing Chapter VII Resolutions.” Nordic Journal of
International Law, 78(3), 309–342.
Johansson, Patrik. 2009b. “Putting Peace to the Vote: Displaced Persons and
a Future Referendum on Nagorno-Karabakh.” Refugee Survey Quarterly,
28(1), 122–139.
Kabachnik, Peter, Joanna Regulska & Beth Mitchneck. 2010. “When and
Where Is Home? The Double Displacement of Georgian IDPs from
Abkhazia.” Journal of Refugee Studies, 23(3), 315–336.
Kaldor, Mary. 2007. “Oil and Conflict: The Case of Nagorno Karabakh.” In
Mary Kaldor, Terry Lynn Karl & Yahia Said (eds.), Oil Wars, London and
Ann Arbor: Pluto Press, 157–182.
Kaufman, Stuart J. 2001. Modern Hatreds: The Symbolic Politics of Ethnic
War, Ithaca & London: Cornell University Press.
Kaufman, Stuart J. 2006. “Escaping the Symbolic Politics Trap:
Reconciliation Initiatives and Conflict Resolution in Ethnic Wars.” Journal
of Peace Research, 43(2), 201–218.
Kaufmann, Chaim. 1996a. “Intervention in Ethnic and Ideological Wars:
Why One Can be Done and the Other Can’t.” Security Studies, 6(1), 62–100.
Kaufmann, Chaim. 1996b. “Possible and Impossible Solutions to Ethnic Civil
Wars.” International Security, 20(4), 136–175.
Kaufmann, Chaim. 1998. “When all else Fails. Ethnic Population Transfers
and Partitions in the Twentieth Century.” International Security, 23(2),
120–156.
Keesing’s Record of World Events. Cambridge: Keesing’s Worldwide LLC.
194
REFERENCES
King, Diane E. 2007. “Internal Displacement: Kurds.” In Suad Joseph (ed.),
Encyclopedia of Women and Islamic Cultures, vol. IV: Economics,
Education, Mobility and Space, Leiden: Brill, 407–409.
King, Gary, Robert O. Keohane & Sidney Verba. 1994. Designing Social
Inquiry: Scientific Inference in Qualitative Research, Princeton: Princeton
University Press.
Knight, Mark & Alpaslan Özerdem. 2004. “Guns, Camps and Cash:
Disarmament, Demobilization and Reinsertion of Former Combatants in
Transitions from War to Peace.” Journal of Peace Research, 41(4), 499–516.
Koser, Khalid. 2003. “Reconciling Control and Compassion? Human
Smuggling and the Right to Asylum.” In Edward Newman & Joanne van
Selm (eds.), Refugees and Forced Displacement: International Security,
Human Vulnerability, and the State, Tokyo, New York & Paris: United
Nations University Press, 181–194.
Koser, Khalid, Martha Walsh & Richard Black. 1998. “Temporary Protection
and the Assisted Return of Refugees from the European Union.”
International Journal of Refugee Law, 10(3), 444–461.
Kostić, Roland. 2007. Ambivalent Peace: External Peacebuilding,
Threatened Identity and Reconciliation in Bosnia and Herzegovina,
Uppsala: Department of Peace and Conflict Research, Uppsala University.
Kritz, Neil J. 2001. “The Rule of Law in the Postconflict Phase: Building a
Stable Peace.” In Chester A. Crocker, Fen Osler Hampson & Pamela Aall
(eds.), Turbulent Peace: The Challenges of Managing International
Conflict, Washington, D.C.: United States Institute of Peace Press, 801–820.
Kumar, Krishna (ed.). 1997. Rebuilding Societies After Civil War: Critical
Roles for International Assistance, Boulder: Lynne Rienner.
Kuperman, Alan J. 2004. “Is Partition Really the Only Hope? Reconciling
Contradictory Findings About Ethnic Civil Wars.” Security Studies, 13(4),
314–349.
Kurki, Milja. 2008. Causation in International Relations: Reclaiming
Causal Analysis, Cambridge: Cambridge University Press.
Kvale, Steinar. 2007. Doing Interviews, Thousand Oaks, CA: Sage
Publications.
Kälin, Walter. 2008. “Durable Solutions for Internally Displaced Persons: An
Essential Dimension of Peacebuilding.” Briefing Paper Submitted to the
195
PEACE BY REPATRIATION
Peacebuilding Commission Working Group on Lessons Learned Meeting of
13 March 2008, Brookings <http://www.brookings.edu/papers/2008/
0313_peacebuilding_kalin.aspx>.
Laitin, David D. 2006. “Ethnic Unmixing and Civil War.” In Roy Licklider &
Mia Bloom (eds.), Living Together After Ethnic War: Exploring the Chaim
Kaufmann Argument, London & New York: Routledge, 210–226.
Lapidoth, Ruth. 1997. Autonomy: Flexible Solutions to Ethnic Conflicts,
Washington, D.C.: United States Institute of Peace Press.
Leckie, Scott (ed.). 2007. Housing, Land, and Property Restitution Rights of
Refugees and Displaced Persons: Laws, Cases, and Materials, Cambridge:
Cambridge University Press.
Lederach, John Paul. 1997. Building Peace: Sustainable Reconciliation in
Divided Societies, Washington, D.C.: United States Institute of Peace Press.
Levinson, David. 1998. Ethnic Groups Worldwide: A Ready Reference
Handbook, Phoenix, AZ.: Oryx Press.
Library of Congress. “Country Studies.” Federal Research Division, Library
of Congress, <http://lcweb2.loc.gov/frd/cs/cshome.html>. Accessed 23
April, 2010.
Licklider, Roy. 1995. “The Consequences of Negotiated Settlements in Civil
Wars, 1945–1993.” American Political Science Review, 89(3), 681–690.
Licklider, Roy & Mia Bloom (eds.). 2006. Living Together After Ethnic
Killing: Exploring the Chaim Kaufmann Argument, London & New York:
Routledge.
Lindholm, Helena. 1993. “Introduction: A Conceptual Discussion.” In
Helena Lindholm (ed.), Ethnicity and Nationalism: Formation of Identity
and Dynamics of Conflict in the 1990s, Gothenburg: Nordnes, 1–39.
Lischer, Sarah Kenyon. 2005. Dangerous Sanctuaries: Refugee Camps, Civil
War, and the Dilemmas of Humanitarian Aid, Ithaca: Cornell University
Press.
Loescher, Gil, Alexander Betts & James Milner. 2008. The United Nations
High Commissioner for Refugees (UNHCR): The Politics and Practice of
Refugee Protection into the Twenty-First Century, London & New York:
Routledge.
196
REFERENCES
Loescher, Gil & James Milner. 2005. “Protracted Refugee Situations:
Domestic and International Security Implications.” Adelphi Paper 375.
Lynch, Dov. 2004. Engaging Eurasia’s Separatist States: Unresolved
Conflicts and De Facto States, Washington, D.C.: United States Institute of
Peace Press.
Lyons, Terrence. 2002. “The Role of Postsettlement Elections.” In Stephen
John Stedman, Donald Rothchild & Elizabeth M. Cousens (eds.), Ending
Civil Wars: The Implementation of Peace Agreements, Boulder: Lynne
Rienner, 215–235.
Mahoney, James & Gary Goertz. 2004. “The Possibility Principle: Choosing
Negative Cases in Comparative Research.” American Political Science
Review, 98(4), 653–669.
Malanczuk, Peter. 1991. “The Kurdish Crisis and Allied Intervention in the
Aftermath of the Second Gulf War.” European Journal of International
Law, 2(1), 114–132.
Malcolm, Noel. 2002. Bosnia: A Short History, London: Pan Books.
Marko, Joseph. 2000. “Bosnia and Herzegovina: Multi-Ethnic or
Multinational?” In Council of Europe (ed.), Societies in Conflict: The
Contribution of Law and Democracy to Conflict Resolutions, Strasbourg:
Council of Europe Publishing, 92–118.
Marshall, Monty G. 2008. Polity IV Project: Political Regime
Characteristics and Transitions, 1800–2008, Version p4v2008 [computer
file]. Severn, MD: Center for Systemic Peace, <http://www.systemicpeace.
org/inscr/inscr.htm>.
Marshall, Monty G. & Keith Jaggers. 2009. Polity IV Project: Dataset Users’
Manual, Severn, MD: Center for Systemic Peace, <http://www.
systemicpeace.org/polity/polity4.htm>.
Martin, Susan & Andrew I. Schoenholtz. 2006. “Promoting the Human
Rights of Forced Migrants.” In Julie A. Mertus & Jeffrey W. Helsing (eds.),
Human Rights and Conflict: Exploring the Links between Rights, Law and
Peacebuilding, Washington, D.C.: United States Institute of Peace Press,
405–429.
McNamara, Karen Elizabeth. 2007. “Conceptualizing Discourses on
Environmental Refugees at the United Nations.” Population and
Environment, 29(1), 12–24.
197
PEACE BY REPATRIATION
Meernik, James. 2005. “Justice and Peace? How the International Criminal
Tribunal Affects Societal Peace in Bosnia.” Journal of Peace Research, 42(3),
271–289.
Melander, Erik. 1999. Anarchy Within. The Security Dilemma Between
Ethnic Groups in Emerging Anarchy, Uppsala: Department of Peace and
Conflict Research.
Melander, Erik. 2009. “The Geography of Fear: Regional Ethnic Diversity,
the Security Dilemma and Ethnic War.” European Journal of International
Relations, 15(1), 95–124.
Mertus, Julie. 2003. “Sovereignty, Gender, and Displacement.” In Edward
Newman & Joanne van Selm (eds.), Refugees and Forced Displacement:
International Security, Human Vulnerability, and the State, Tokyo, New
York & Paris: United Nations University Press, 250–273.
Mertus, Julie A. & Jeffrey W. Helsing (eds.). 2006. Human Rights and
Conflict: Exploring the Links between Rights, Law and Peacebuilding,
Washington, D.C.: United States Institute of Peace Press.
Milner, James. 2009. Refugees, the State and the Politics of Asylum in
Africa, Basingstoke: Palgrave Macmillan.
Mooney, Erin. 2003. “Towards a Protection Regime for Internally Displaced
Persons.” In Edward Newman & Joanne van Selm (eds.), Refugees and
Forced Displacement: International Security, Human Vulnerability, and
the State, Tokyo, New York & Paris: United Nations University Press, 159–
180.
Moses, Jonathon W. & Torbjørn L. Knutsen. 2007. Ways of Knowing:
Competing Methodologies in Social and Political Research, Basingstoke:
Palgrave Macmillan.
Murakami, Haruki. 2006. “A Folklore for my Generation: A Prehistory of
Late-Stage Capitalism.” In Blind Willow, Sleeping Woman, London: Vintage
Books, 75–101.
Naqvi, Yasmin. 2004. “Between a Rock and a Hard Place? A Legal Analysis
of the Voluntary Repatriation of Guatemalan Refugees.” Refugee Survey
Quarterly, 23(2), 74–99.
Naudé, Wim. 2009. “Entrepreneurship, Post-Conflict.” In Tony Addison &
Tilman Brück (eds.), Making Peace Work: The Challenges of Social and
Economic Reconstruction, Basingstoke: Palgrave Macmillan, 251–263.
198
REFERENCES
Newman, Edward. 2003. “Refugees, International Security, and Human
Vulnerability: Introduction and Survey.” In Edward Newman & Joanne van
Selm (eds.), Refugees and Forced Displacement: International Security,
Human Vulnerability, and the State, Tokyo, New York & Paris: United
Nations University Press, 3–30.
Noll, Gregor. 1999. “Rejected Asylum Seekers: The Problem of Return.”
International Migration, 37(1), 267–288.
Noll, Gregor. 2003. “Securitizing Sovereignty: States, Refugees, and the
Regionalization of International Law.” In Edward Newman & Joanne van
Selm (eds.), Refugees and Forced Displacement: International Security,
Human Vulnerability, and the State, Tokyo, New York & Paris: United
Nations University Press, 277–305.
Nordquist, Kjell-Åke. 1992. Peace after War. On Conditions for Durable
Inter-State Boundary Agreements, Uppsala: Department of Peace and
Conflict Research.
NRC. 2008. While Waiting for Return: Achievements and Further
Challenges in the National Reponse to IDP Needs in Azerbaijan, Baku:
Norwegian Refugee Council.
Ogata, Sadako. 1997. “Introduction: Refugee Repatriation and PeaceBuilding.” Refugee Survey Quarterly, 16(2), vi–x.
Ohlson, Thomas. 1998. Power Politics and Peace Policies: Intra-State
Conflict Resolution in Southern Africa, Uppsala: Department of Peace and
Conflict Research.
OHR. 2000. “Property Law Implementation Plan (PLIP).” OHR, UNHCR,
OSCE, UNMIBH, and CRPC, <http://www.ohr.int/plip>.
Orchard, Philip. 2008. A Right to Leave: Refugees, States, and
International Society, Vancouver: University of British Columbia,
<https://circle.ubc.ca/handle/2429/1261>.
Orjuela, Camilla. 2003. “Building Peace in Sri Lanka: A Role for Civil
Society?” Journal of Peace Research, 40(2), 195–212.
OSCE. 2006. “Statement by the Minsk Group Co-Chairs, 3 July 2006.”
<http://www.osce.org>. Accessed 4 February, 2010.
OSCE. Undated. “Property Law Implementation Completion: Essential to
Securing the Right of Return.” OSCE Mission to Bosnia and Herzegovina,
199
PEACE BY REPATRIATION
<http://www.oscebih.org/human_rights/propertyrepossession.asp?d=1>.
Accessed 20 January, 2010.
Pape, Robert A. 1997. “Partition: An Exit Strategy for Bosnia.” Survival,
39(4), 25–28.
Paris Agreement. 1991. “Agreement on a Comprehensive Political Settlement
of the Cambodia Conflict.” <http://www.usip.org/library/pa/cambodia/
pa_cambodia.html>. Accessed 14 December, 2005.
Paris, Roland. 2004. At War’s End: Building Peace After Civil Conflict, New
York: Cambridge University Press.
Peterson, Jenny H. 2010. “‘Rule of Law’ Initiatives and the Liberal Peace:
The Impact of Politicised Reform in Post-Conflict States.” Disasters, 34(1),
15–39.
Petrin, Sarah. 2002. “Refugee Return and State Reconstruction: A
Comparative Analysis.” New Issues in Refugee Research, No. 66,
<http://www.unhcr.org>.
Phuong, Catherine. 2000. “‘Freely to Return’: Reversing Ethnic Cleansing in
Bosnia-Herzegovina.” Journal of Refugee Studies, 13(2), 165–183.
Posen, Barry R. 1993. “The Security Dilemma and Ethnic Conflict.” In
Michael E. Brown (ed.), Ethnic Conflict and International Security,
Princeton, NJ: Princeton University Press, 103–124.
Posner, Daniel N. 2004. “Measuring Ethnic Fractionalization in Africa.”
American Journal of Political Science, 48(4), 849–863.
Pouligny, Béatrice. 2005. “Civil Society and Post-Conflict Peacebuilding:
Ambiguities of International Programmes Aimed at Building ‘New’
Societies.” Security Dialogue, 36(4), 495–510.
Prendergast, John & Emily Plumb. 2002. “Building Local Capacity: From
Implementation to Peacebuilding.” In Stephen John Stedman, Donald
Rothchild & Elizabeth M. Cousens (eds.), Ending Civil Wars: The
Implementation of Peace Agreements, Boulder: Lynne Rienner, 327–349.
Prettitore, Paul. 2006. “Refugee Return in Bosnia and Herzegovina.” In
Michael Dumper (ed.), Palestinian Refugee Repatriation: Global
Perspectives, London & New York: Routledge, 180–208.
Putnam, Tonya L. 2002. “Human Rights and Sustainable Peace.” In Stephen
John Stedman, Donald Rothchild & Elizabeth M. Cousens (eds.), Ending
200
REFERENCES
Civil Wars: The Implementation of Peace Agreements, Boulder: Lynne
Rienner, 237–271.
Rae, Douglas W. & Michael Taylor. 1970. The Analysis of Political
Cleavages, New Haven and London: Yale University Press.
Ragin, Charles C. 1987. The Comparative Method: Moving Beyond
Qualitative and Quantitative Strategies, Berkeley: University of California
Press.
Ragin, Charles C. 1994. Constructing Social Research: The Unity and
Diversity of Method, Thousand Oaks, London & New Delhi: Pine Forge
Press.
Ragin, Charles C. 2000. Fuzzy-Set Social Science, Chicago: University of
Chicago Press.
Ragin, Charles C. 2003. “Fuzzy-Set Analysis of Necessary Conditions.” In
Gary Goertz & Harvey Starr (eds.), Necessary Conditions: Theory,
Methodology, and Applications, Lanham: Rowman & Littlefield, 179–196.
Ragin, Charles C. 2004a. “From Fuzzy Sets to Crisp Truth Tables.”
<http://www.compasss.org>.
Ragin, Charles C. 2004b. “Turning the Tables: How Case-Oriented Research
Challenges Variable-Oriented Research.” In Henry E. Brady & David Collier
(eds.), Rethinking Social Inquiry: Diverse Tools, Shared Standards,
Lanham, MD: Rowman & Littlefield, 123–138.
Ragin, Charles C. 2006a. “The Limitations of Net-Effects Thinking.” In
Benoît Rihoux & Heike Grimm (eds.), Innovative Comparative Methods for
Policy Analysis: Beyond the Quantitative-Qualitative Divide, New York:
Springer, 13–41.
Ragin, Charles C. 2006b. “Set Relations in Social Research: Evaluating Their
Consistency and Coverage.” Political Analysis, 14(3), 291–310.
Ragin, Charles C. 2008a. Redesigning Social Inquiry: Fuzzy Sets and
Beyond, Chicago & London: University of Chicago Press.
Ragin, Charles C. 2008b. User’s Guide to Fuzzy-Set/Qualitative
Comparative Analysis, Version September 2008, Tucson, AZ: Department
of Sociology, University of Arizona, <http://www.u.arizona.edu/~cragin/
fsQCA/download/fsQCAManual.pdf>.
201
PEACE BY REPATRIATION
Ragin, Charles C. 2009. “Qualitative Comparative Analysis Using Fuzzy Sets
(fsQCA).” In Benoît Rihoux & Charles C. Ragin (eds.), Configurational
Comparative Methods: Qualitative Comparative Analysis (QCA) and
Related Techniques, Los Angeles, London, New Delhi & Singapore: Sage
Publications, 87–121.
Ramcharan, B.G. (ed.). 1997. The International Conference on the Former
Yugoslavia: Official Papers, Volume 1, The Hague, London & Boston:
Kluwer Law International.
Rashiduzzaman, Mohammad. 1998. “Bangladesh’s Chittagong Hill Tracts
Peace Accord: Institutional Features and Strategic Concerns.” Asian Survey,
38(7), 653–670.
Richards, David L., Ronald D. Gelleny & David H. Sacko. 2001. “Money with
a Mean Streak? Foreign Economic Penetration and Government Respect for
Human Rights in Developing Countries.” International Studies Quarterly,
45(2), 219–239.
Richmond, Oliver P. 2005. The Transformation of Peace, Basingstoke:
Palgrave Macmillan.
Riggs, Fred W. 1991a. “Ethnicity, Nationalism, Race, Minority: A
Semantic/Onomantic Exercise (Part One).” International Sociology, 6(3),
281–305.
Riggs, Fred W. 1991b. “Ethnicity, Nationalism, Race, Minority: A
Semantic/Onomantic Exercise (Part Two).” International Sociology, 6(4),
443–463.
Rihoux, Benoît & Charles C. Ragin (eds.). 2009. Configurational
Comparative Methods: Qualitative Comparative Analysis (QCA) and
Related Techniques, Los Angeles, London, New Delhi & Singapore: Sage
Publications.
Roeder, Philip G. & Donald Rothchild (eds.). 2005. Sustainable Peace:
Power and Democracy after Civil War, Ithaca & London: Cornell University
Press.
Rogel, Carole. 2003. “Kosovo: Where It All Began.” International Journal of
Politics, Culture and Society, 17(1), 167–182.
Roht-Arriaza, Naomi & Javier Mariezcurrena (eds.). 2006. Transitional
Justice in the Twenty-First Century: Beyond Truth versus Justice,
Cambridge: Cambridge University Press.
202
REFERENCES
Rosenau, James N. 1966. “Pre-theories and Theories of Foreign Policy.” In
R. Barry Farrell (ed.), Approaches to Comparative and International
Politics, Evanston, IL: Northwestern University Press, 27–92.
Rowland, Richard H. 2008. “Population Trends in a Contested PseudoState: The Case of Nagorno-Karabakh.” Eurasian Geography and
Economics, 49(1), 99–111.
Sahovic, Dzenan. 2007. Socio-Cultural Viability of International
Intervention in War-Torn Societies: A Case Study of Bosnia Herzegovina,
Umeå: Department of Political Science.
Sambanis, Nicholas. 2000. “Partition as a Solution to Ethnic War: An
Empirical Critique of the Theoretical Literature.” World Politics, 52(4), 437–
483.
Sambanis, Nicholas. 2006. “Partition as a Solution to Ethnic War: An
Empirical Critique of the Theoretical Literature.” In Roy Licklider & Mia
Bloom (eds.), Living Together After Ethnic Killing: Exploring the Chaim
Kaufmann Argument, London & New York: Routledge, 41–88.
Sannerholm, Richard. 2007. “Beyond Criminal Justice: Promoting the Rule
of Law in Post-Conflict Societies.” In Ashok Swain, Ramses Amer & Joakim
Öjendal (eds.), Globalization and Challenges to Building Peace, London,
New York & Delhi: Anthem Press, 55–72.
Schnabel, Albrecht & Hans-Georg Ehrhart (eds.). 2005. Security Sector
Reform and Post-Conflict Peacebuilding, New York: United Nations
University Press.
Schneider, Carsten Q. & Claudius Wagemann. 2007. Qualitative
Comparative Analysis (QCA) und Fuzzy Sets: Ein Lehrbuch für Anwender
und jene, die es werden wollen, Opladen & Farmington Hills: Verlag Barbara
Budrich.
Schreuer, Christoph. 1998a. “The Brčko Award of 14 February 1997.” Leiden
Journal of International Law, 11(1), 71–80.
Schreuer, Christoph. 1998b. “The Brčko Supplemental Award of 15 March
1998.” Leiden Journal of International Law, 11(3), 493–495.
Schreuer, Christoph. 1999. “The Brčko Final Award of 15 March 1999.”
Leiden Journal of International Law, 12(3), 575–581.
Seawright, Jason. 2002. “Testing for Necessary and/or Sufficient Causation:
Which Cases Are Relevant?” Political Analysis, 10(2), 178–193.
203
PEACE BY REPATRIATION
Seawright, Jason & John Gerring. 2008. “Case Selection Techniques in Case
Study Research.” Political Research Quarterly, 61(2), 294–308.
Silber, Laura & Allan Little. 1996. The Death of Yugoslavia, 2nd ed. London:
BBC/Penguin.
Smith, Merrill. 2004. “Warehousing Refugees: A Denial of Rights, a Waste of
Humanity.” In World Refugee Survey 2004, Washington, D.C.: U.S.
Committee for Refugees and Immigrants, 38–56.
Sollenberg, Margareta (ed.). 2001. States in Armed Conflict 2000, Uppsala:
Department of Peace and Conflict Research.
Spear, Joanna. 2002. “Disarmament and Demobilization.” In Stephen John
Stedman, Donald Rothchild & Elizabeth M. Cousens (eds.), Ending Civil
Wars: The Implementation of Peace Agreements, Boulder: Lynne Rienner,
141–182.
Stake, Robert E. 1995. The Art of Case Study Research, Thousand Oaks,
London, and New Delhi: Sage Publications.
Stedman, Stephen John. 2002. “Introduction.” In Stephen John Stedman,
Donald Rothchild & Elizabeth M. Cousens (eds.), Ending Civil Wars: The
Implementation of Peace Agreements, Boulder: Lynne Rienner, 1–40.
Stedman, Stephen John, Donald Rothchild & Elizabeth M. Cousens (eds.).
2002. Ending Civil Wars: The Implementation of Peace Agreements,
Boulder: Lynne Rienner.
Stein, Barry N. 1997. “Reintegrating Returning Refugees in Central
America.” In Krishna Kumar (ed.), Rebuilding Societies After Civil War:
Critical Roles for International Assistance, Boulder: Lynne Rienner, 155–
180.
Strauss, Mark. 2002. “Attacking Iraq.” Foreign Policy, (129), 14–19.
Swain, Ashok (ed.). 2007. Diasporas, Armed Conflict and Peacebuilding in
Their Homeland, Uppsala: Department of Peace and Conflict Research,
Uppsala University.
Taylor, Charles Lewis & Michael C. Hudson. 1972. World Handbook of
Political and Social Indicators, 2nd ed. New Haven and London: Yale
University Press.
204
REFERENCES
Thielemann, Eiko R. 2004. “Why Asylum Policy Harmonisation Undermines
Refugee Burden-Sharing.” European Journal of Migration and Law, 6(1),
47–65.
UCDP. 2009a. “Conflict Database.” Department of Peace and Conflict
Research, <http://www.pcr.uu.se>. Accessed 9 February, 2010.
UCDP. 2009b. “UCDP Peace Agreement Dataset, 1989–2005.” Uppsala
Conflict Data Program <http://www.pcr.uu.se/research/UCDP/index.htm>.
UCDP. 2009c. “UCDP/PRIO Armed Conflict Dataset Codebook, version 42009.” Uppsala Conflict Data Program <http://www.pcr.uu.se/research/
UCDP/index.htm>.
UCDP. 2010. “Definitions.” Department of Peace and Conflict Research,
<http://www.pcr.uu.se/research/ucdp/definitions/>. Accessed 12 October,
2010.
UN. 1992. An Agenda for Peace: Preventive Diplomacy, Peacemaking and
Peace-Keeping, A/47/277 - S/24111, New York: United Nations.
UN. 2000. Report of the Panel on United Nations Peace Operations, “The
Brahimi Report,” U.N. Doc. A/55/305-S/2000/809.
UN. 2004. Letter dated 16 March 2004 from the Permanent Representative
of Azerbaijan to the United Nations addressed to the Secretary-General and
the President of the Security Council, United Nations Doc. A/59/66S/2004/219.
UN. 2005. United Nations Secretary-General’s address to UNHCR
Executive Committee, 6 October 2005, SG/SM/10145, REF/1185, Geneva:
United Nations.
UN Security Council. 2006. Briefing by the United Nations High
Commissioner for Refugees, 24 January 2006, S/PV.5353, New York:
United Nations.
UN Security Council. 2009. Resolution 1895, 18 November 2009.
UNHCR. 1955. Nobel Lecture by Dr. van Heuven Goedhart, 12 December
1955, <http://nobelprize.org>.
UNHCR. 1992. Statement of Mrs. Sadako Ogata, United Nations High
Commissioner for Refugees, to the International Meeting on Humanitarian
Aid for Victims of the Conflict in the former Yugoslavia, Geneva, 29 July
1992, <http://www.unhcr.org/admin/ADMIN/3ae68fac1a.html>.
205
PEACE BY REPATRIATION
UNHCR. 1995a. “Populations of Concern to UNHCR: A Statistical Overview
(1994).” <http://www.unhcr.org>. Accessed 6 January, 2010.
UNHCR. 1995b. The State of the World’s Refugees: In Search of Solutions,
Oxford: Oxford University Press.
UNHCR. 1996. Applicability of the Cessation Clauses to Refugees from the
Republics of Malawi and Mozambique, UNHCR/IOM/88/96, New York:
United Nations.
UNHCR. 1997. The State of the World’s Refugees: A Humanitarian Agenda,
Oxford: Oxford University Press.
UNHCR. 2000. The State of the World’s Refugees: Fifty Years of
Humanitarian Action, Oxford: Oxford University Press.
UNHCR. 2004. UNHCR Position on the Continued International Protection
Needs of Individuals from Kosovo, August 2004, Geneva: UNHCR.
UNHCR. 2005. Statistical Yearbook 2003: Trends in Displacement,
Protection and Solutions, Geneva: United Nations High Commissioner for
Refugees.
UNHCR. 2006. Country Operations Plan: Armenia, <http://www.unhcr.
org/434138e32.html>.
UNHCR. 2009. “Internally Displaced Persons and Returnees in Bosnia and
Herzegovina.” UNHCR Representation in Bosnia and Herzegovina,
<http://www.unhcr.ba/return/IDPs.pdf>. Accessed 20 January, 2010.
UNHCR. 2010. “UNHCR Global Appeal 2010–11.” <http://www.unhcr.org/
ga10/index.html>.
UNHCR Executive Committee. 1985. Voluntary Repatriation, Conclusion
No. 40 (XXXVI), New York: United Nations.
UNHCR Executive Committee. 2004. Protracted Refugee Situations,
EC/54/SC/CRP.14, UNHCR.
USCRI. 1988–2006. World Refugee Survey, various editions. Washington,
D.C.: U.S. Committee for Refugees and Immigrants; <http://refugees.org/>.
Wallensteen, Peter. 2007. Understanding Conflict Resolution, 2nd ed.
London, Thousand Oaks & New Delhi: Sage Publications.
206
REFERENCES
Wallensteen, Peter & Patrik Johansson. 2004. “Security Council Decisions in
Perspective.” In David M. Malone (ed.), The UN Security Council: From the
Cold War to the 21st Century, Boulder: Lynne Rienner, 17–33.
Wallensteen, Peter & Margareta Sollenberg. 2001. “Armed Conflict, 1989–
2000.” Journal of Peace Research, 38(5), 629–644.
Walter, Barbara F. 1997. “The Critical Barrier to Civil War Settlement.”
International Organization, 51(3), 335–364.
Walter, Barbara F. 2002. Committing to Peace: The Successful Settlement of
Civil Wars, Princeton & Oxford: Princeton University Press.
van Selm, Joanne. 2003. “Refugee Protection Policies and Security issues.”
In Edward Newman & Joanne van Selm (eds.), Refugees and Forced
Displacement: International Security, Human Vulnerability, and the State,
Tokyo, New York & Paris: United Nations University Press, 66–92.
Vanhanen, Tatu. 1999. “Domestic Ethnic Conflict and Ethnic Nepotism: A
Comparative Analysis.” Journal of Peace Research, 36(1), 55–73.
Weiss, Thomas G. & David A. Korn. 2006. Internal Displacement:
Conceptualization and Its Consequences, London & New York: Routledge.
Weiss, Thomas G. & Amir Pasic. 1998. “Dealing with the Displacement and
Suffering Caused by Yugoslavia’s Wars.” In Roberta Cohen & Francis M.
Deng (eds.), The Forsaken People. Case Studies of the Internally Displaced,
Washington D.C.: Brookings Institution Press, 175–231.
Werner, Suzanne. 1999. “The Precarious Nature of Peace: Resolving the
Issues, Enforcing the Settlement, and Renegotiating the Terms.” American
Journal of Political Science, 43(3), 912–934.
Wiberg, Håkan. 1990. Konfliktteori och fredsforskning, 2nd ed. Solna:
Almqvist & Wiksell.
Woodward, Susan L. 2002. “Economic Priorities for Successful Peace
Implementation.” In Stephen John Stedman, Donald Rothchild & Elizabeth
M. Cousens (eds.), Ending Civil Wars: The Implementation of Peace
Agreements, Boulder: Lynne Rienner, 183–214.
Yin, Robert K. 2009. Case Study Research: Design and Methods, 4th ed. Los
Angeles, London, New Delhi, Singapore, Washington D.C.: Sage
Publications.
207
PEACE BY REPATRIATION
Zahar, Marie-Joëlle. 2002. “Peace by Unconventional Means: Lebanon’s
Ta’if Agreement.” In Stephen John Stedman, Donald Rothchild & Elizabeth
M. Cousens (eds.), Ending Civil Wars: The Implementation of Peace
Agreements, Boulder & London: Lynne Rienner, 567–597.
Zolberg, Aristide R., Astri Suhrke & Sergio Aguayo. 1989. Escape from
Violence: Conflict and the Refugee Crisis in the Developing World, New
York & Oxford: Oxford University Press.
Interviews
Interview 1:
Country Representative of IC organization in Azerbaijan. 18
August 2008, Baku.
Interview 2:
Country Director of international NGO in Azerbaijan. 20
August 2008, Baku.
Interview 3:
Project Manager plus two Coordinators at IC organization in
Azerbaijan. 20 August 2008, Baku.
Interview 4:
Country Director of international NGO in Azerbaijan. 21
August 2008, Baku.
Interview 5:
High level Representative of Azerbaijani Government
Ministry. 21 August 2008, Baku.
Interview 6:
Chairperson of Azerbaijani NGO. 23 August 2008, Baku.
Interview 7:
Head of Delegation international NGO in Azerbaijan. 25
August 2008, Baku.
Interview 8:
Program Officer at IC organization in Azerbaijan. 25 August
2008, Baku.
Interview 9:
Chairperson of Azerbaijani NGO. 26 August 2008, Baku.
Interview 10:
Head of Azerbaijani NGO. 26 August 2008, Baku.
Interview 11:
Chairperson of Azerbaijani NGO. 27 August 2008, Baku.
Interview 12:
Director of Azerbaijani NGO. 28 August 2008, Baku.
Interview 13:
Desk Officer at Azerbaijani Government Ministry. 28 August
2008, Baku.
Interview 14:
Director of Azerbaijani NGO. 29 August 2008, Baku.
Interview 15:
Chairperson Armenian NGO. 3 September 2008, Yerevan.
Interview 16:
Deputy Head of Armenian Government Ministry. 3
September 2008, Yerevan.
208
REFERENCES
Interview 17:
Director of Armenian NGO. 4 September 2008, Yerevan.
Interview 18:
Representative of IC organization in Armenia. 4 September
2008, Yerevan.
Interview 19:
Protection Officer at IC organization in Armenia. 4
September 2008, Yerevan.
Interview 20:
Director of Armenian NGO. 5 September 2008, Yerevan.
Interview 21:
Head of de facto NKR Government Ministry. 8 September
2008, Stepanakert.
Interview 22:
Head of the de facto Stepanakert Municipality Department.
8 September 2008, Stepanakert.
Interview 23:
Head of Nagorno-Karabakh NGO. 8 September 2008,
Stepanakert.
Interview 24:
Head of Nagorno-Karabakh NGO. 9 September 2008,
Shoushi.
Interview 25:
Expert Adviser at Bosnian (State level) Government
Ministry. 15 October 2008, Sarajevo.
Interview 26:
Social Worker at Bosnian NGO. 15 October 2008, Sarajevo.
Interview 27:
Head of Bosnian NGO. 20 October 2008, Sarajevo.
Interview 28:
Expert Adviser at IC organization in BiH. 21 October 2008,
Sarajevo.
Interview 29:
Representative of IC organization in BiH. 23 October 2008,
Sarajevo.
Interview 30:
Protection Officer plus Program Officer at IC organization in
BiH. 23 October 2008, Sarajevo.
209
PEACE BY REPATRIATION
210
APPENDICES
Appendix 1: Interview Guide
General
What is the relation between the solution of displacement and the resolution
of the conflict?
In what ways is the solution of displacement dependent on the resolution of
the conflict—and in what ways is the resolution of the conflict dependent on
the solution of displacement?
Can the conflict be resolved without the return of DPs? How? Why not?
Would resolution of conflict have been different/easier if there were no DPs?
Why?
Is the IDP/refugee/returnee community a political actor?
Do IDPs/refugee/returnees have an agenda that they pursue?
Are they a force for peace or for conflict; compromise, principles, rights…?
In what ways could the resolution of this conflict become a model for other
cases?
Nagorno-Karabakh
What do you think will happen to the displaced persons (various groups)?
When and how will displacement end?
In what ways can the return of displaced persons contribute to peace?
In what ways can displaced persons and returnees contribute to peace?
How do you envision future relations between Armenians and Azerbaijanis?
The present plan for peace envisages a future referendum on the status of
Nagorno-Karabakh. Why is this /not a good idea?
Bosnia-Herzegovina
In what ways has the return of displaced persons contributed to peace?
In what ways have displaced persons and returnees contributed to peace?
What effects have the strong emphasis on return had on peace
implementation?
211
PEACE BY REPATRIATION
ENDREF
ENDALL
REPREF
REPALL
PEACE 2
CASE
PEACE
Appendix 2: Fuzzy Membership Scores
Afghanistan, 2001
0.00
0.00
0.75
0.75
0.00
Angola, 1995
0.25
0.25
0.00
0.00
0.00
0.00
0.00
Azerbaijan (N.-Karabakh), 1994
0.50
0.25
0.00
0.00
0.00
0.00
Bangladesh (CHT), 1992
1.00
1.00
1.00
1.00
1.00
1.00
Bosnia-Herzegovina (Croat), 1994
0.50
0.50
0.25
0.25
0.25
0.25
Bosnia-Herzegovina (Serb), 1995
0.50
0.50
0.25
0.25
0.25
0.25
Cambodia, 1998
0.50
0.75
1.00
1.00
0.75
0.75
Chad, 1994
0.25
0.25
0.25
0.25
0.50
0.50
Congo, 1999
0.25
0.25
1.00
1.00
0.25
0.75
Croatia (Serb), 1993
0.00
0.00
0.00
0.00
0.00
0.00
Croatia (Serb), 1995
1.00
1.00
0.25
0.25
0.00
0.00
Democratic Republic of Congo, 2001
0.25
0.25
0.75
0.00
0.25
0.00
El Salvador, 1991
1.00
1.00
0.25
0.25
1.00
1.00
Ethiopia, 1991
0.50
0.75
1.00
1.00
0.75
0.75
Ethiopia (Eritrea), 1991
0.50
0.75
0.75
0.75
0.25
0.25
Georgia (Abkhazia), 1993
0.50
0.75
0.00
—
0.00
—
Guatemala, 1995
1.00
1.00
0.25
0.75
1.00
1.00
Guinea-Bissau, 1999
0.50
0.75
1.00
—
1.00
—
Iran (Kurdistan), 1990
0.25
0.25
0.00
0.00
0.25
0.25
Iran (Kurdistan), 1993
0.25
0.25
0.00
0.00
0.50
0.50
Iran (Kurdistan), 1996
0.50
0.75
0.00
0.00
0.50
0.50
Iraq, 1996
0.50
0.25
0.00
0.50
0.00
0.25
Iraq (Kurdistan), 1993
0.25
0.25
1.00
1.00
0.25
0.25
Iraq (Kurdistan), 1996
0.50
0.25
1.00
1.00
0.25
0.25
Lebanon, 1990
0.50
0.50
0.50
0.50
0.25
0.25
Liberia, 1995
0.25
0.25
1.00
0.75
0.25
0.25
Morocco (Western Sahara), 1989
0.50
0.25
0.00
0.00
0.00
0.00
Mozambique, 1992
1.00
1.00
1.00
1.00
1.00
1.00
Myanmar (Kachin), 1992
0.50
0.25
0.00
—
0.00
—
Myanmar (Karen), 1992
0.25
0.25
0.00
0.00
0.00
0.00
Myanmar (Karen), 1995
0.00
0.00
0.00
0.00
0.00
0.00
Nicaragua, 1989
1.00
1.00
0.75
0.75
0.50
0.50
Peru, 1999
1.00
1.00
0.00
—
1.00
—
Philippines (Mindanao), 1990
0.25
0.25
0.00
0.00
0.00
0.00
Russia (Chechnya), 1996
0.25
0.25
0.75
—
0.25
—
Rwanda, 1994
0.25
0.25
1.00
1.00
0.25
1.00
Sierra Leone, 2000
0.50
0.75
0.75
0.75
0.50
0.50
Sri Lanka (Eelam), 2001
0.00
0.00
0.75
0.50
0.25
0.25
Tajikistan, 1996
0.00
0.00
0.75
0.00
0.25
0.25
Tajikistan, 1998
0.50
0.25
0.75
0.00
0.25
0.25
Uganda, 1991
0.25
0.25
1.00
0.50
0.50
0.50
Yemen (South Yemen), 1994
0.50
0.25
1.00
1.00
1.00
1.00
Yugoslavia (Kosovo), 1999
0.75
0.75
0.75
1.00
0.25
0.50
212
PART
FRAC
FRAC 2
CON
ENDREF 2
ENDALL 2
APPENDICES
Afghanistan, 2001
0,25
0,25
0.00
0.75
1.00
0.00
Angola, 1995
0,00
0,00
0.00
0.75
0.25
0.00
Azerbaijan (N.-Karabakh), 1994
0,00
0,00
0.75
0.25
0.25
1.00
CASE
Bangladesh (CHT), 1992
1,00
1,00
0.50
0.00
0.25
0.75
Bosnia-Herzegovina (Croat), 1994
0,25
0,25
0.00
0.75
0.25
0.75
Bosnia-Herzegovina (Serb), 1995
0,25
0,25
0.50
0.75
0.25
0.75
Cambodia, 1998
1,00
1,00
0.00
0.25
0.25
0.00
Chad, 1994
0,75
0,75
0.00
1.00
1.00
0.00
Congo, 1999
0,50
0,75
0.00
1.00
0.75
0.00
Croatia (Serb), 1993
0,25
0,00
0.00
0.50
0.25
0.75
Croatia (Serb), 1995
0,25
0,00
0.00
0.50
0.25
0.75
Democratic Republic of Congo, 2001
0,25
0,00
0.00
1.00
1.00
0.00
El Salvador, 1991
1,00
1,00
0.00
0.25
0.25
0.00
Ethiopia, 1991
0,75
0,75
0.00
0.75
0.75
0.00
Ethiopia (Eritrea), 1991
0,25
0,25
1.00
0.75
0.75
1.00
Georgia (Abkhazia), 1993
0,00
—
0.50
0.50
0.50
0.75
Guatemala, 1995
1,00
1,00
0.00
0.50
0.75
0.00
Guinea-Bissau, 1999
1,00
—
0.00
1.00
0.75
0.00
Iran (Kurdistan), 1990
0,50
0,50
0.00
0.75
0.75
1.00
Iran (Kurdistan), 1993
0,75
0,75
0.00
0.75
0.75
1.00
Iran (Kurdistan), 1996
0,75
0,75
0.00
0.75
0.75
1.00
Iraq, 1996
0,50
0,50
0.00
0.50
0.50
0.00
Iraq (Kurdistan), 1993
0,25
0,25
0.75
0.50
0.50
1.00
Iraq (Kurdistan), 1996
0,25
0,25
0.75
0.50
0.50
1.00
Lebanon, 1990
0,25
0,25
0.25
0.50
0.25
0.00
Liberia, 1995
0,25
0,25
0.00
1.00
1.00
0.00
Morocco (Western Sahara), 1989
0,00
0,00
0.75
0.50
0.50
1.00
Mozambique, 1992
1,00
1,00
0.00
0.75
0.25
0.00
Myanmar (Kachin), 1992
0,50
—
0.75
0.50
0.50
0.75
Myanmar (Karen), 1992
0,00
0,00
0.75
0.50
0.50
0.25
Myanmar (Karen), 1995
0,00
0,00
0.75
0.50
0.50
0.25
Nicaragua, 1989
1,00
1,00
0.00
0.50
0.25
0.00
Peru, 1999
1,00
—
0.00
0.75
0.75
0.00
Philippines (Mindanao), 1990
0,25
0,25
0.50
0.50
0.25
1.00
Russia (Chechnya), 1996
0,25
—
0.75
0.25
0.25
1.00
Rwanda, 1994
0,50
1,00
0.00
0.50
0.00
0.00
Sierra Leone, 2000
0,75
0,75
0.00
1.00
0.75
0.00
Sri Lanka (Eelam), 2001
0,25
0,50
0.00
0.50
0.50
0.75
Tajikistan, 1996
0,50
0,50
0.00
0.50
0.75
0.00
Tajikistan, 1998
0,50
0,50
0.00
0.50
0.75
0.00
Uganda, 1991
1,00
1,00
0.00
1.00
1.00
0.00
Yemen (South Yemen), 1994
1,00
1,00
0.00
0.00
0.25
0.00
Yugoslavia (Kosovo), 1999
0,25
0,75
0.50
0.50
0.50
1.00
213
PEACE BY REPATRIATION
SOV. PEACE
5 YEARS
POLITY
SCORE
CIRI VALUE
INTERNAT.
PRESENCE
SUSTAINABLE
PEACE
Afghanistan, 2001
2003
no
no
-66
3.50
yes
0.00
Angola, 1995
1998
yes
no
-3
2.13
no
0.25
Azerbaijan (N.-Karabakh), 1994
2005
yes
yes
-7
4.50
no
0.50
Bangladesh (CHT), 1992
-
yes
yes
6
5.38
no
1.00
Bosnia-Herzegovina (Croat), 1994
-
yes
yes
-66
6.75
yes
0.50
Bosnia-Herzegovina (Serb), 1995
-
yes
yes
-66
6.75
yes
0.50
Cambodia, 1998
-
yes
yes
2
4.50
no
0.50
Chad, 1994
1997
yes
no
-2
4.25
no
0.25
Congo, 1999
2002
yes
no
-4
5.63
no
0.25
Croatia (Serb), 1993
1995
no
no
-5
7.25
yes
0.00
Croatia (Serb), 1995
-
yes
yes
7
7.75
no
1.00
2006
yes
no
5
1.63
yes
0.25
El Salvador, 1991
-
yes
yes
7
8.88
no
1.00
Ethiopia, 1991
-
yes
yes
1
4.75
no
0.50
1998
yes
yes
1
4.75
no
0.50
Georgia (Abkhazia), 1993
-
yes
yes
5
3.63
yes
0.50
Guatemala, 1995
-
yes
yes
8
6.38
no
1.00
Guinea-Bissau, 1999
-
yes
yes
1
6.63
yes
0.50
Iran (Kurdistan), 1990
1993
yes
no
-6
1.25
no
0.25
Iran (Kurdistan), 1993
1996
yes
no
3
0.63
no
0.25
Iran (Kurdistan), 1996
-
yes
yes
3
0.63
no
0.50
Iraq, 1996
2004
yes
yes
-9
0.00
yes
0.50
Iraq (Kurdistan), 1993
1996
yes
no
-9
0.00
yes
0.25
Iraq (Kurdistan), 1996
-
yes
yes
-9
0.00
yes
0.50
Lebanon, 1990
-
yes
yes
-66
-
yes
0.50
2000
yes
no
0
3.13
no
0.25
Morocco (Western Sahara), 1989
-
yes
yes
-7
2.50
yes
0.50
Mozambique, 1992
-
yes
yes
6
7.00
no
1.00
Myanmar (Kachin), 1992
-
yes
yes
-7
1.13
no
0.50
Myanmar (Karen), 1992
1995
yes
no
-7
1.13
no
0.25
Myanmar (Karen), 1995
1997
no
no
-7
0.63
no
0.00
Nicaragua, 1989
-
yes
yes
7
7.13
no
1.00
Peru, 1999
-
yes
yes
9
5.50
no
1.00
Philippines (Mindanao), 1990
1993
yes
no
8
5.25
no
0.25
Russia (Chechnya), 1996
1999
yes
no
7
3.75
no
0.25
Rwanda, 1994
1997
yes
no
-5
3.75
no
0.25
-
yes
yes
5
6.63
yes
0.50
Sri Lanka (Eelam), 2001
2003
no
no
6
3.38
no
0.00
Tajikistan, 1996
1998
no
no
-1
3.38
no
0.00
Tajikistan, 1998
-
yes
yes
-2
4.50
no
0.50
1994
yes
no
-4
3.88
no
0.25
Yemen (South Yemen), 1994
-
yes
yes
-2
2.25
no
0.50
Yugoslavia (Kosovo), 1999
-
yes
yes
6
6.50
yes
0.75
CASE
Democratic Republic of Congo, 2001
Ethiopia (Eritrea), 1991
Liberia, 1995
Sierra Leone, 2000
Uganda, 1991
214
RESTARTED
SOV. PEACE
2 YEARS
Appendix 3: Coding Sustainable Peace
APPENDICES
Explanatory note for Appendix 3.






RESTARTED: Year when conflict restarted (if at all) according to the
UCDP.
SOVEREIGN PEACE 2 YEARS: Did the conflict termination last for
at least two years?
SOVEREIGN PEACE 5 YEARS: Did the conflict termination last for
at least five years?
POLITY SCORE: Average Polity score at termination plus 4, 5, and 6
years.
CIRI VALUE: Average CIRI value at termination plus 4, 5, and 6
years; a weighted value max 10, with PHYSINT and EMPINX given
equal weight: (PHYSINT * 5 + EMPINX * 4) / 8.102
INTERNATIONAL PRESENCE: was the international community
present at termination plus 5 years?
The concepts are operationalized according to the necessary and sufficient
condition structure. If the original termination did not last for five years,
then POLITY, CIRI and INTERNATIONAL PRESENCE are irrelevant. If the
original termination lasted for five years but POLITY is <6, then CIRI and
INTERNATIONAL PRESENCE are irrelevant.
102 The Physical Integrity Index (PHYSINT) is an additive index based on the Torture, Extrajudicial Killing,
Political Imprisonment, and Disappearance indicators, and it ranges from 0 (no government respect for these
rights) to 8 (full government respect for these rights). For details see Cingranelli and Richards (1999). The
Empowerment Rights Index (EMPINX) is an additive index based on the Freedom of Movement, Freedom of
Speech, Workers’ Rights, Political Participation, and Freedom of Religion indicators, ranging from 0 (no
respect) to 10 (full respect). For details see Richards, Gelleny and Sacko (2001).
215
PEACE BY REPATRIATION
Appendix 4: Complex versus Parsimonious Solutions
As explained in chapter 7, a fuzzy set analysis can be designed to produce
either a complex or a parsimonious solution, the difference being whether or
not the researcher manually assigns a particular value on the outcome for
logical remainders, that is, theoretically possible configurations not
represented among the empirical cases. To illustrate, consider an analysis of
the three conditions repatriation, ended displacement situation and
partition. The truth table reproduced in table A4.1 shows that there are two
logical remainders, namely 0-1-1 (repall*ENDALL*PART) and 1-1-1
(REPALL* ENDALL*PART).
Table A4.1
Truth table (all cases, all displaced persons, three conditions).
Repall
1
0
0
1
1
0
0
1
Endall
0
0
0
1
0
1
1
1
Part
0
0
1
0
1
0
1
1
N
8
6
5
5
4
3
0
0
Peace
Consistency
0.558140
0.612245
0.640000
0.790698
0.625000
0.857143
1.000000
1.000000
In order to achieve a complex solution I need to decide whether, if such cases
were to exist, they would most likely be associated with sustainable peace or
not. For example, assigning them a positive value on the outcome yields the
following complex solution (Formula A4.1):
repall*ENDALL → PEACE
consistency 0.8621, coverage 0.3205
Alternatively, giving the logical remainders a negative value on the outcome
results in the following complex solution (Formula A4.2):
repall*ENDALL*part → PEACE
consistency 0.8571, coverage 0.3077
Finally, not assigning any particular value to these configurations, but
instead setting logical remainders to “don’t care,” gives the following
parsimonious solution (Formula A4.3):
216
APPENDICES
repall*ENDALL → PEACE
consistency 0.8621, coverage 0.3205
The solutions in formulas A4.1 and A4.3 are identical. The only difference is
that in the complex solution (formula A4.1), I, the researcher, decided that
the logical remainders should have a positive value on the outcome, whereas
in the parsimonious solution (formula A4.3) this decision was taken by the
software. The software makes that decision on the basis of which solution is
the most parsimonious. As formula A4.2 shows, giving the logical
remainders a negative value on the outcome results in a more complex
solution, and therefore, the software prefers to give them a positive value.
For the researcher to make this decision he or she needs to possess
enough theoretical and substantive knowledge to be able to assess whether
the logical remainders are, in fact, most likely to be associated with the
presence of the outcome or with the absence of the outcome. Because I do
not possess that kind of knowledge, I set logical remainders to “don’t care”
throughout in this study.
217
PEACE BY REPATRIATION
Appendix 5: Truth Table Solutions
This appendix presents the truth table solutions that constitute the basis for
formulas 10.1–10.18 in chapter 10. In several cases the same solution has
resulted from separate tests of different combinations of conditions. Below, I
present one table per solution (that is, one table per formula in chapter 10)
and specify for each solution which tests—or “models”—yield that particular
solution.
The parsimonious solution option is used throughout, as are the default
frequency cutoff (1) and consistency cutoff (0.80). Two things should be
noted. First, a consistency cutoff of 1 means that a configuration should be
represented by at least one empirical case in order to be included in the
analysis, while a consistency cutoff of 0.80 means that an individual
configuration should have a raw consistency of more than 0.80 to be
counted as consistent in the truth table analysis. These cutoffs are used
throughout. Once solutions are found, however, I have required a solution
consistency of at least 0.80 to count a solution as consistent. (Consistencies
of >0.80 in the truth table are no guarantee that the resulting solution will
be consistent. Second, the software reports de facto cutoffs, rather than the
default ones. For example, if four configurations are represented by,
respectively, 0, 0, 3, and 7 cases, then the reported frequency cutoff will be 3.
This is what is reproduced in the tables below.
Table A5.1
Truth table solution, basis for formula 10.1.
Data: all conflicts, all displaced persons
Model: PEACE = (REPALL, ENDALL)
frequency cutoff: 3
consistency cutoff: 0.804348
Model: PEACE = (ENDALL, PART)
frequency cutoff: 8
consistency cutoff: 0.813559
Model: PEACE = (ENDALL, FRAC)
frequency cutoff: 2
consistency cutoff: 0.822222
Model: PEACE = (REPALL, ENDALL, FRAC)
frequency cutoff: 1
consistency cutoff: 0.827586
Model: PEACE = (ENDALL, PART, FRAC)
frequency cutoff: 1
consistency cutoff: 0.818182
raw coverage*
ENDALL
0.653846
solution coverage: 0.653846
solution consistency: 0.822581
unique coverage*
0.653846
consistency
0.822581
* Raw coverage corresponds to the coverage of an individual solution term irrespective of the coverage of
other solution terms. Unique coverage refers to the coverage related exclusively to that particular solution
term (cf. Ragin 2006b, 304–307).
218
APPENDICES
Table A5.2
Truth table solution, basis for formula 10.2.
Data: all conflicts, all displaced persons
Model: PEACE = (ENDALL, CON)
frequency cutoff: 1
consistency cutoff: 0.875000
Model: PEACE = (ENDALL, FRAC, CON)
frequency cutoff: 1
consistency cutoff: 0.928571
raw coverage
ENDALL*CON
0.179487
solution coverage: 0.179487
solution consistency: 0.875000
unique coverage
0.179487
consistency
0.875000
Table A5.3
Truth table solution, basis for formula 10.3.
Data: all conflicts, all displaced persons
Model: PEACE = (REPALL, ENDALL, PART)
frequency cutoff: 3
consistency cutoff: 0.857143
Model: PEACE = (REPALL, ENDALL, PART, CON)
frequency cutoff: 1
consistency cutoff: 0.857143
raw coverage
unique coverage
repall*ENDALL
0.320513
0.320513
solution coverage: 0.320513
solution consistency: 0.862069
consistency
0.862069
Table A5.4
Truth table solution, basis for formula 10.4.
Data: all conflicts, all displaced persons
Model: PEACE = (REPALL, ENDALL, CON)
frequency cutoff: 1
consistency cutoff: 0.857143
Model: PEACE = (REPALL, ENDALL, FRAC, CON)
frequency cutoff: 1
consistency cutoff: 0.823529
raw coverage
unique coverage
repall*ENDALL
0.320513
0.205128
ENDALL*CON
0.179487
0.064103
solution coverage: 0.384615
solution consistency: 0.882353
consistency
0.862069
0.875000
219
PEACE BY REPATRIATION
Table A5.5
Truth table solution, basis for formula 10.5.
Data: all conflicts, refugees only
Model: PEACE = (ENDREF, FRAC)
frequency cutoff: 1
consistency cutoff: 0.837838
raw coverage
ENDREF*frac
0.462687
solution coverage: 0.462687
solution consistency: 0.837838
unique coverage
0.462687
consistency
0.837838
unique coverage
0.208955
consistency
0.875000
Table A5.6
Truth table solution, basis for formula 10.6.
Data: all conflicts, refugees only
Model: PEACE = (ENDREF, CON)
frequency cutoff: 1
consistency cutoff: 0.875000
Model: PEACE = (ENDREF, FRAC, CON)
frequency cutoff: 1
consistency cutoff: 0.928571
raw coverage
ENDREF*CON
0.208955
solution coverage: 0.208955
solution consistency: 0.875000
Table A5.7
Truth table solution, basis for formula 10.7.
Data: all conflicts, refugees only
Model: PEACE = (REPREF, ENDREF, FRAC, CON)
frequency cutoff: 1
consistency cutoff: 0.833333
raw coverage
unique coverage
repref*ENDREF
0.283582
0.179104
ENDREF*CON
0.208955
0.104478
solution coverage: 0.388060
solution consistency: 0.838710
consistency
0.791667
0.875000
Table A5.8
Truth table solution, basis for formula 10.8.
Data: all conflicts, refugees only
Model: PEACE = (REPREF, ENDREF, FRAC, CON)
frequency cutoff: 1
consistency cutoff: 0.833333
raw coverage
unique coverage
repref*ENDREF
0.283582
0.238806
REPREF*frac*CON
0.179104
0.134328
solution coverage: 0.417910
solution consistency: 0.800000
220
consistency
0.791667
0.800000
APPENDICES
Table A5.9
Truth table solution, basis for formula 10.9.
Data: conflicts over territory, all displaced persons
Model: PEACE = (REPALL, ENDALL)
frequency cutoff: 2
consistency cutoff: 0.812500
Model: PEACE = (ENDALL, FRAC)
frequency cutoff: 2
consistency cutoff: 0.842105
Model: PEACE = (REPALL, ENDALL, FRAC)
frequency cutoff: 1
consistency cutoff: 0.812500
Model: PEACE = (ENDALL, PART, FRAC)
frequency cutoff: 1
consistency cutoff: 0.812500
raw coverage
ENDALL
0.459459
solution coverage: 0.459459
solution consistency: 0.809524
unique coverage
0.459459
consistency
0.809524
Table A5.10
Truth table solution, basis for formula 10.10.
Data: conflicts over territory, all displaced persons
Model: PEACE = (ENDALL, CON)
frequency cutoff: 1
consistency cutoff: 0.875000
Model: PEACE = (REPALL, ENDALL, CON)
frequency cutoff: 1
consistency cutoff: 0.909091
Model: PEACE = (REPALL, ENDALL, FRAC, CON)
frequency cutoff: 1
consistency cutoff: 0.909091
raw coverage
unique coverage
ENDALL*CON
0.378378
0.378378
solution coverage: 0.378378
solution consistency: 0.875000
consistency
0.875000
Table A5.11
Truth table solution, basis for formula 10.11.
Data: conflicts over territory, refugees only
Model: PEACE = (REPREF, ENDREF)
frequency cutoff: 2
consistency cutoff: 0.812500
Model: PEACE = (ENDREF, FRAC)
frequency cutoff: 1
consistency cutoff: 0.842105
Model: PEACE = (REPREF, ENDREF, FRAC)
frequency cutoff: 1
consistency cutoff: 0.812500
raw coverage
ENDREF
0.531250
solution coverage: 0.531250
solution consistency: 0.809524
unique coverage
0.531250
consistency
0.809524
221
PEACE BY REPATRIATION
Table A5.12
Truth table solution, basis for formula 10.12.
Data: conflicts over territory, refugees only
Model: PEACE = (ENDREF, CON)
frequency cutoff: 1
consistency cutoff: 0.875000
Model: PEACE = (REPREF, ENDREF, CON)
frequency cutoff: 1
consistency cutoff: 0.909091
Model: PEACE = (ENDREF, FRAC, CON)
frequency cutoff: 1
consistency cutoff: 928571
Model: PEACE = (REPREF, ENDREF, FRAC, CON)
frequency cutoff: 1
consistency cutoff: 0.909091
raw coverage
unique coverage
ENDREF*CON
0.437500
0.437500
solution coverage: 0.437500
solution consistency: 0.875000
consistency
0.875000
Table A5.13
Truth table solution, basis for formula 10.13.
Data: conflicts over territory, refugees only
Model: PEACE = (REPREF, ENDREF, FRAC, CON)
frequency cutoff: 1
consistency cutoff: 0.909091
raw coverage
unique coverage
REPREF*frac*CON
0.375000
0.375000
solution coverage: 0.375000
solution consistency: 0.800000
consistency
0.800000
Table A5.14
Truth table solution, basis for formula 10.14.
Data: conflicts over territory, refugees only
Model: PEACE = (REPREF, ENDREF, FRAC, CON)
frequency cutoff: 1
consistency cutoff: 0.909091
raw coverage
unique coverage
ENDREF*CON
0.437500
0.125000
REPREF*frac*CON
0.375000
0.062500
solution coverage: 0.500000
solution consistency: 0.800000
222
consistency
0.875000
0.800000
APPENDICES
Table A5.15
Truth table solution, basis for formula 10.15.
Data: conflicts over government, all displaced persons
Model: PEACE = (ENDALL, FRAC)
frequency cutoff: 2
consistency cutoff: 0.812500
Model: PEACE = (REPALL, ENDALL, FRAC)
frequency cutoff: 1
consistency cutoff: 0.814815
Model: PEACE = (REPALL, ENDALL, PART, FRAC, CON)
frequency cutoff: 1
consistency cutoff: 0.814815
raw coverage
unique coverage
ENDALL
0.829268
0.829268
solution coverage: 0.829268
solution consistency: 0.829268
consistency
0.829268
Table A5.16
Truth table solution, basis for formula 10.16.
Data: conflicts over government, all displaced persons
Model: PEACE = (REPALL, ENDALL)
frequency cutoff: 2
consistency cutoff: 0.888889
raw coverage
unique coverage
repall*ENDALL
0.390244
0.390244
solution coverage: 0.390244
solution consistency: 0.888889
consistency
0.888889
Table A5.17
Truth table solution, basis for formula 10.17.
Data: conflicts over government, all displaced persons
Model: PEACE = (REPALL, FRAC)
frequency cutoff: 1
consistency cutoff: 0.866667
raw coverage
unique coverage
repall*frac
0.317073
0.317073
solution coverage: 0.317073
solution consistency: 0.866667
consistency
0.866667
Table A5.18
Truth table solution, basis for formula 10.18.
Data: conflicts over government, refugees only
Model: PEACE = (REPREF, FRAC)
frequency cutoff: 1
consistency cutoff: 0.823529
raw coverage
REFREF*frac
0.400000
solution coverage: 0.400000
solution consistency: 0.823529
unique coverage
0.400000
consistency
0.823529
223
PEACE BY REPATRIATION
224
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Related manuals

Download PDF

advertisement