31
Issue 30, January 2020
Abstract: The article aims to present four con licts in East and South Africa that spilled over
into neighboring states. The article covers a limited period of time from the end of the Cold War
to the present and focuses on the area of East and South Africa. All these con licts are brie ly
introduced and the author concentrates mainly on the description of the spillover effect (it includes,
for example, the phase of the con lict in which the con lict spilled over to a neighboring
state, timing, interconnection or cooperation of armed groups). The most important goal of the
article is to analyze conditions based on con lict theories and spillover effect theories that led
to the spillover into these states and compare them with other neighboring states, where the
con licts did not occur.
Keywords: Con lict; Cooperation; Neighboring state; Security; Spillover effect; War.
Introduction
After the end of the Cold War, there were several positive but also some negative trends
in con lict studies. Positive is that there is a reduction in the number of interstate wars
and the number of deadly wounded soldiers on the battle ield has been decreasing
as well. On the other hand, the number
of national wars is higher and, moreover,
these con licts negatively affect the civilian
population that has to be resettled. At
the same time, civilian losses are increasing,
which are several times higher than in
the wars before the year 1990. Africa has
the highest con lict potential, the end of
the Cold War has not brought a peaceful
period there, while many con licts, which
were suppressed due to bipolar rivalry and
the interest of the superpowers, broke out
with even greater intensity. Not only the end
EAST and SOUTH AFRICA:
Spillover Effect of Con licts
Lucie KONEČNÁ
Lucie KONEČNÁ M.A.
Ph.D. candidate
Faculty of Social Science
Masaryk University
E-mail: 397816@mail.muni.cz
Con lict Studies Quarterly
Issue 30, January 2020, pp. 31-54
DOI:10.24193/csq.30.2
Published First Online: 05/01/2020
32
Con lict Studies Quarterly
of the Cold War but also other factors such as natural conditions and inef icient economic
development, ethnic hatred and tribalism (largely in luenced by colonialism and
decolonization), high level of poverty or instability of state institutions are important
factors affecting Africa’s con lict environment. This is con irmed by statistics that show
that over 40% of all African countries have been affected by at least one civil war over
the past 40 years (Elbadawi & Sambanis, 2009, p. 3).
One of the phenomena that occur in con licts in Africa is its spillover effect. In a few
cases, we may have seen that a con lict that began in a domestic state has spilled over
into neighboring states. I do not mean the internationalization of a con lict that involves
governments of two states ighting against each other. In these cases, the national war
caused destabilization and chaos in a neighboring state, and these negative effects
caused the outbreak of a new civil war in that state. In these new con licts, in the vast
majority of cases, militias from neighboring states (not state ones), abuse the unstable
situations in the neighboring state to expand their activity. In some cases they are
directly involved, in other cases, they only support rebel groups in neighboring countries.
The key to this phenomenon of con lict spillover is the continuity of time and the
involvement of actors from the neighboring state. However, why is in some cases, the
neighboring state destabilized and con lict spilled over, and in others not? I will describe
this on the example of con licts from East and South Africa.
East Africa is a region rich in human capital, but other characteristics include harsh
climatic conditions (frequent droughts), increasing environmental degradation, poverty,
political instability or high con lict potential. Somalia is part of this area. It is a very
unstable state and a civil war has been going on there since 1991. The war spilled over
into neighboring Ethiopia. Another con lict that spilled over to a neighboring state is
the civil war in Ethiopia that spilled over to Djibouti. In contrast, South Africa is a more
stable region than the Eastern one. However, even there, the security situation is not
good, there is a very high crime rate, a high in lux of refugees and uneven population
geography. All these factors increase the con lict potential, and therefore we can ind
con licts that have spilled over also in this region. Speci ically, it is a civil war in Angola
that spilled over to Namibia and the con lict in Cabinda (Angola), which spilled over to
the Republic of Congo. In this article, I will describe these four case studies and focus
on spillover analysis.
Theoretical Framework
A number of world authors have been concerned with the de inition and classi ication
of con licts over the years. In the most general sense, con lict can be expressed as a state
of competition between two or more actors who have incompatible interests. However,
not every dispute or tense situation can be considered as a con lict. The con lict has
certain characteristics. According to Holsti (1991), the con lict must have these speci ic
33
Issue 30, January 2020
characteristics: actors (at least one of them is state actor), a de inable area of con lict
interests, present tensions, and con licting actions. Very often used criteria are those
from Smolík and Šmíd (2010). They divide them into the qualitative and quantitative
categories. On the basis of these criteria, national con licts can be considered as a war
when at least one of the armed forces is subordinate to governmental power, at the
same time showing a managed organization, a determined strategy, and continuity
of operations. The quantitative criterion of 1,000 dead per year determines the most
intense kind of con lict – war (Smolík & Šmíd, 2010, pp. 18-25). The civil war is a kind
of con lict that, like con lict, has been described by many authors. In my opinion, the
aptest criteria of the civil war were determined by Fortna (2008). The civil war has
killed over 1,000 people over its lifetime, taking place within the internationally recognized
borders of a sovereign state, threatening the sovereignty of that state, one of the
struggling parties being state troops, rebels are organized and capable of causing losses
to the state (Fortna, 2008, pp. 2-5). These criteria clearly determine what civil war is.
The Spillover Effect
The Spillover Effect is a relatively new concept in security studies that have been taken
from economics. This concept is very often mentioned but without some theoretical
anchoring. Ko i Annan’s statement is well known in this context. He claimed that wars
in neighboring states will not leave the surrounding states intact (Whitaker, 2003, p.
1). One of the most famous con lict scientists, Paul Collier, works with the spillover
effect too. According to Collier (2007), civil war lows from the countries concerned
not only to neighboring countries but also to more distant states through terrorism,
drug traf icking and the spread of epidemics (Collier, 2007, p. 147). Thus, there is a
spillover of threats from states in con lict. Collier, however, did not pay close attention
to this phenomenon and focused primarily on the trap of con lict and the causes of
civil wars. Another well-known author who mentions this phenomenon in his work is
Stewart Patrick. Patrick (2006) states that weak and failing states are very often the
place where threats such as terrorism, proliferation, and crime are spilling from. Thus,
the author does not associate the concept of spillover with the spillover of con licts and
wars, but rather with the spillover of threats that may cause instability in neighboring
countries and undermine security. Patrick argues that the relationship between state
weakness and spillover is not linear. It always depends on the type of threats, and on
religious, geographical, and cultural factors (Patrick, 2006, pp. 7-17). A similar study
was published by the African Development Bank, which describes the impact on the
neighboring state and the spillover of threats from con lict-affected states. They divide
these threats into social and economic. Economic threats and impacts include e.g. disruption
of infrastructure or decline in economic performance. Social threats include,
for example, the in lux of refugees or the spread of diseases and HIV / AIDS (African
Development Bank, 2008, pp. 11-19).
34
Con lict Studies Quarterly
There are several authors who dealt with the spillover effect in more detail. Among
the most important work is the work of Bosker and de Ree, who conducted research
on spillovers in 2009. Based on historical events and mathematical calculations, they
tried to determine whether con licts spilled over and how probable it is. As a result
of their study, con lict spillover is a phenomenon that can only be con irmed in Africa.
It can also be argued that if a given African country is adjacent to at least one con lict
country, the risk of con lict in that country is about 4% higher (Bosker & de Ree, 2009).
Silve and Verdier (2018) studied the theory of regional security complexes. At the same
time, however, they also focused on the spillover threats that spread within the complex
when a civil war is taking place in one of the states. They determined the so-called main
elements of instability, which undermine security within the complex and can lead to
the outbreak of war in another state. These elements of instability include the proliferation
of small arms, the high in lux of refugees and ethnic cross-border ties. They also
described the negative impact of the size and porosity of the border and the instability
of state institutions. Terrill (2008) described con lict spillovers over Iraq. According
to him, there are four spillover factors, which can then lead to con lict. The irst is the
in lux of refugees and displaced persons, the second is cross-border terrorism, the third
is an international crime and organized crime, and the fourth is the intensi ication of
separatism (Terrill, 2008).
According to available research, threats such as enormous migration, terrorism, and
arms traf icking spillover when a neighboring African state is hit by the con lict. At the
same time, these threats weaken the economic performance of neighboring states. All
these factors weaken neighboring states, and in some cases, the con lict itself actually
spillover, especially in Africa. In order to discuss con lict spillovers and not just the
outbreak of a new con lict in a neighboring state, certain conditions must be met. The
vast majority of spillover authors de ine a certain continuity of time. This is de ined
by the period during which the con lict in the neighboring state takes place, and most
of them add to this period ive another years after the con lict itself ends. This period
of time, therefore, clearly determines the continuity of the spillover effect. I also work
with this de inition of time continuity in my research. At the same time, I add the second
and third conditions, which, in my opinion, are crucial. Otherwise, it cannot be clearly
demonstrated that this is a spillover of the con lict and not a mere outbreak of a new
con lict in a state which has been weakened, for example, by unfavorable regional developments.
The second condition is the involvement of some of the same actors. This
is mostly a certain anti-state militia that can physically engage in the ighting itself in a
neighboring state, but indirect involvement is also suf icient, it usually involves inancial
technical or logistical support. The second key factor is, therefore, the involvement of
some actors from the con lict that spilled over to a neighboring state. The third condition
for a clear de inition of the spillover effect is cooperation, a certain degree of cooperation
between the actor from the original con lict and one of the actors involved in the
35
Issue 30, January 2020
con lict in the neighboring state where the con lict spilled over. These three conditions
for de ining the spillover effect have been determined on the basis of the analysis of the
above-mentioned literature on spillover effect, as well as on the basis of the analysis of
several case studies of a spillover effect from East and South Africa.
Methodology
The irst goal of this work is to describe the spillover effect of con licts on cases from
East and South Africa. This describes how the spillover effect works and focuses on the
time continuity of con lict’s spillover and the involvement of cross-border actors. As a
result, the spillover effect of con licts in Africa can be described. Descriptive-analytical
techniques and comparisons are used for this purpose. The cases were selected on the
basis of similarity, namely Mill’s method of matching, the two examined phenomena
have one factor in common (Kouba, 2008, p. 116). In this case, the dependent variable
(spillover of con lict) is common for all cases. All cases are geographically close and
meet the necessary conditions that differentiate the spillover effect from the outbreak
of a new con lict. In includes time continuity, the involvement of cross-border actors
and cooperation.
The second goal of my thesis is to determine which factors characterizing neighboring
states were present at the moment when the con lict spilled over, and how these factors
differed from the states without spillover effect. Based on theories about spillover effect,
causes of con licts and regional security complexes, I have identi ied factors characterizing
a neighboring state that may affect the spillover and unleashing of a con lict in
a neighboring state. I have divided these factors into four categories: social, politicalsecurity,
economic and geographical. I have worked with several dozen authors to cover
the widest possible range of factors. Among the theorists mentioned in the theoretical
part about spillover, it includes Patrick (2006), Silve and Verdier (2018), Terrill (2008),
Millett (2002), African Development Bank, Carmignani and Kler (2013) and Whikater
(2003). At the same time, I have also worked with publications by other well-known
scientists, like Galula (1964), which describes the conditions in states where it is easier
for rebels to prepare a successful uprising. Galula de ines a number of geographical factors.
From these mentioned factors I have chosen ive, which are easily mathematically
expressed and therefore can be incorporated into my analysis. Galula also describes the
economy and level of development. It is the reason why I work with the value of GDP
(Galula, 1964). In addition to GDP, I have added two factors related to the occurrence
of natural resources, which Collier describes in his work, saying that if these factors
are met, the chances of a con lict eruption are much higher (Collier and Hoef ler 1999).
Other authors, included in my work, are Ncube and Jones, who characterize the causes
of con licts in Africa and de ine the factors that in luence the outbreak of con licts. Based
on the theories of these authors, I, therefore, characterized other factors that increase
the chances of con lict’s spillover and its outbreak in a neighboring state. Among the
36
Con lict Studies Quarterly
above-mentioned factors, I added a high incidence of terrorist attacks and a human
development index. Other factors, with the exception of weapons proliferation, have
already been included. Proliferation cannot be included because there is no accurate
and relevant data on Africa for this phenomenon. At the same time, other factors not
mentioned by these authors have been included. Many other authors and institutions
consider these factors as a weakening element of the state, increasing the likelihood of
con lict. This is, for example, a high in lux of refugees, described by Enghoff, Hansen,
Umar, Gildestad, Owen and Obara (2010, pp. 67-76), Gomez, Christensen, Araya, and
Harild (2010, pp. 7-20) or FFP, which measures the Fragile State Index. In addition,
Carmignani and Kler (2013) or Whitaker (2003), which I have already mentioned,
mention the in lux of refugees as a risk factor. Some factors are also included on the
basis of the Fragile State Index (FFP) theories, namely humanitarian disaster, high crime
rates and the presence of armed active militias. The overall distribution of factors can
be seen here:
SOCIAL FACTORS:
a) High occurrence of the same ethnicity in neighboring states
b) High occurrence of the same religion in neighboring states
c) Humanitarian disaster (e.g. famine)
d) Common war history
e) Low human development index (life expectancy, literacy + income)
f) A high in lux of refugees
POLITICAL-SECURITY FACTORS:
a) The different political ideology of the neighboring state
b) High level of terrorism
c) The political instability of a neighboring country
d) The high crime rate of a neighboring country
e) Existence of armed active militias
f) Small army
ECONOMIC FACTORS:
a) Presence of important natural resources in the border area
b) Dependence on exports of one natural resource
c) Low GDP
GEOGRAPHIC FACTORS:
a) Porous and long borders
b) Size of neighboring state (large state)
c) The population of the neighboring state (populated)
d) Uneven population demography
e) Uneven ethnic demography
37
Issue 30, January 2020
For the purposes of this research, I use qualitative comparative analysis with sharp
sets, which uses Boolean logic to determine the necessary and suf icient conditions for
the occurrence of a given phenomenon.1
The characteristics of the neighboring states
are divided into four categories based on social, political, economic and geographical
factors. Subsequently, truth tables are created for individual cases. Columns consist
of individual variables (factors and result), rows represent all hypothetically possible
combinations of factors (Beneš & Drulák, 2016, p. 86). These tables are used for all
factors involved in all neighboring states. Data on social, political-security, economic
and geographical factors are converted, calibrated to 0 and 1 (1 indicating the occurrence
and 0 the opposite). Thus, dichotomous data encoding occurs (Ragin, 2008, p.
131). Then I compare what factors (within the QCA condition) were present in the
con lict that spilled over and whether these factors were present in situations where
the spillover effect did not occur. The cases of spillover effects will be compared with
each other – whether there are factors that are the same for each case. For some of
these factors, values can simply be converted to 0 or 1, depending on whether or not
the phenomenon occurs in the neighboring state. For others, a threshold is created
based on the African or regional average.
Spillover Effect of Con licts in East Africa
After the end of the Cold War, two con licts spilled over into neighboring states in East
Africa. The irst is the civil war in Somalia, which has been going on from 1991 to the
present. After the overthrow of the Barre regime in 1991, a power vacuum emerged in
a country that various armed factions and militias tried to ill. The most famous militias
were those of Ali Mahdi Muhammad and Muhammad Farah Aidid (Pham, 2013, pp.
135-137). Individual clans and militias divided the territory of the country and fought
each other in almost all parts of the state. There were total anarchy, chaos and mass
movements in the country. After the end of the foreign intervention in 1995, clashes
between individual militias continued, but not to the same intensity as in previous
years. Until 2000, ighting took place only in certain areas. In 2001, the Union of Islamic
Courts (UIC) was created (Williams, 2009, p. 515). UIC also formed armed factions,
which during 2006 conquered the most important strategic sites, including the capital
Mogadishu. The UIC subsequently disintegrated into several fractions. The most radical
faction that has waged war against the government since 2007 until now is the Islamic
terrorist group, Al-Shabaab. Another, previously active, terrorist organization, which
has already disappeared, was Hizbul Islam. This con lict, spilled over into one of the
federal states of Ethiopia, into Ogaden.
1 Boolean logic deals with logical operations on a set of values {0, 1}. I also use this set in my research
(Caramani, 2009, p. 25).
38
Con lict Studies Quarterly
The second con lict that spilled over was the civil war in Ethiopia, which took place in
1974-1991. Spillover to neighboring Djibouti occurred after the end of the Cold War,
so I work with this case as well. In 1974, part of the army formed the Coordinated
Council of the Armed Forces of the Police and Territorial Army, which they called the
Derg. Derg carried out a coup d’état and overthrew the emperor. There were several
opposition groups in the country trying to weaken the power of the Derg and ight
against it. One of the most famous groups was the WSLA (Western Somali Liberation
Front), an irredentist movement that operated in the Ogaden region. In the 1980s, a
series of uprisings took place mainly in the Eritrean area (which at that time belonged
to Ethiopia) and in the Tigrai area. At that time, the EPLF (Eritrean People’s Liberation
Front), whose goal was the secession of Eritrean territory, was particularly active. TPLF
(Tigray’s People Liberation Front), which had similar secessionist goals, cooperated
with EPLF. The government has prepared several attacks against the EPLF and TPLF.
The aim of these attacks was the total destruction of these organizations. Operations
such as Shiraro, Lash and Adwa were launched, but these attacks were not successful.
TPLF together with several other groups formed at the turn of 1988/1989 organization
Ethiopian People’s Revolutionary Democratic Front (EPRDF), which, with the support
of local residents and the United States of America, managed to depose Derg in 1991.
During 1990, Mengistu’s army was pushed into the capital, then Mengistu led the
country, and the army surrendered to the EPRDF (Vestal, 1999, p. 75). The following
sections describe the spillover of these two con licts to neighboring states.
Spillover of Con lict from Somalia to Ethiopia
The con lict in Ogaden began in 1994, three years after the outbreak of the civil war in
Somalia. One of the ighting parties of the con lict is ONLF (Ogaden National Liberation
Front), the other is the military-led party EPRDF. ONLF was of icially established in
1984, but until 1994 it was a non-violent organization that tried to promote its interests
mainly through political channels. ONLF took part in the 1992 elections, the
new elections following the overthrow of the previous communist regime, and ONLF
became the main party in the regional parliament after these elections. The main aim
of ONLF is the right to self-determination of Somalis in Somali federal state Ogaden,
the secession of the area, or at least gaining autonomous status. This, of course, is not
compatible with the goals of the central government (Vaughan, 2011). The situation
became more pronounced during and immediately after the irst elections in 1992.
Several members of ONLF disappeared, others were found dead under unclear circumstances.
The situation in the region began to be alarming, the population rebelled
and therefore the central government led by the EPRDF decided in 1994 to abolish
the regional parliament in Somalia. From this year ONLF declared a violent campaign
against the central government (Dagne, 2010). Because the organization did not have
enough means and capacities, it resorted to the strategy of hit and run during the 90s.
39
Issue 30, January 2020
During the 1990s, the organization carried out a series of attacks on government and
military targets. Most of them were bombings. Between 2003 and 2004, the government
took an offensive to destroy ONLF troops. However, this offensive was neither
extensive nor systematic, so the declared objectives were not achieved (Home Of ice
UK, 2004). In 2006, there were mainly attacks on military bases, but also an attack on
the Chinese oil company Sinopec Limited in Obole, which killed more than 60 workers
(Opalo, 2010). In response to these events, the Ethiopian army launched a very brutal
offensive in early 2007, which was directed not only against ONLF but against all residents
of the Somali Federated States. A trade blockade was launched, which, together
with the drought that prevailed in Ogaden at the time, caused famine. In addition,
there were mass arrests of people who were imprisoned for several years without trial.
Another collective punishment was the burning of villages, where residents allegedly
sympathized with ONLF or villages that allegedly supported or cooperated with ONLF.
In 2010, government representatives met with representatives of ONLF in Frankfurt,
Germany, where they signed a peace agreement to end the ighting. However, part of
ONLF led by Muhammad Omar Osman disagreed with this agreement and created its
own offshoot of ONLF called the Ogaden National Liberation Army (ONLA), which is
still active in the area (Home Of ice UK, 2017). In 2013, ONLA relocated to the Somali
capital of Mogadishu, where it plans and conducts all operations against the Ethiopian
government. Clashes continue at low intensity, e.g. ONLA killed 23 Ethiopian soldiers
in 2016 and more than 30 in 2017 (Harar24, 2017).
In order to discuss the spillover effect, there must be a temporal link between the conlict
in Somalia and Ethiopia, as well as the involvement of some of the same actors.
The con lict in Somalia began in 1991, after the collapse of the Barre regime and with
varying intensity, this con lict has continued. The con lict was the most intensive in
the irst half of the 1990s and during the UIC’s rise to power. During this irst intensive
period, spillover to neighboring Ethiopia began, speci ically at the turn of 1993-1994.
At the same time, it has been noted that the most intense ighting took place in Somalia
in its central and southern parts, especially along the borders with Ethiopia and Kenya,
which negatively affected the security situation in both countries (Menkhaus, 2007).
Security threats have been spilled in border regions of Kenya and Ethiopia. This included
arms and counterfeit trade, drug traf icking, increased terrorist activity, increased crime
and an overall deterioration in the security situation compared to the national average
(Gastrow, 2011). In the case of Ethiopia, the con lict itself was spilled over, which
was also in luenced by internal factors. It is obvious that there is a continuity of time
in con lict’s spillover.
The second factor discussed is the involvement of the same actors. According to the
latest statistics, the Somali federal state occupies more than 95% of ethnic Somalis,
in the 1990s this igure was a few percent lower, slightly above 90%, but even in this
40
Con lict Studies Quarterly
period, the Somalis represented a majority group (Ethiopian Government, 2018). The
vast majority of Somalia’s people are ethnic Somalis, for whom ethnicity is not a key
element, but clan membership is crucial. Major clans in Somalia include Darod, Dir,
Hawiye, Isaaq, and Rahanweyn (Pham, 2013). The clans are then divided into several
dozen sub-clans. The area along the border with Ethiopia is inhabited primarily by
the Darod clan (roughly 2/3), followed by the Rahanweyn clan. A very similar ethnic
demography is also along the border with Somalia in Ethiopian territory. This territory
consists only of the Somali federal state, where the vast majority of the population are
Somalis from the Darod tribe, namely the sub-clans of Ogaden, Jidwaq, and Marehan
(Lewis, 1999). During the beginning of the Somali Civil War, the territory of Somalia
was divided among the factions, which were created on a clan basis. Members of the
individual clans worked together across Somalia, Ethiopia, and Kenya. Ethiopia’s people
were involved in the Somali con lict in Somali territory. At the same time, the Somalis
very often crossed the international border and used the territory of the Somali Federal
State to recruit new members, gain inancial support and an environment for illegal
trade, which was one of the main sources of income (IRB, 1999). Therefore, cooperation
across the states was clear.
In addition to this clan involvement, cooperation and interdependence of some organizations
are also evident. Speci ically, it is a cooperation between ONLF and AIAI
(Al-Ittihad Al-Islam - an armed Islamic organization that was established in 1990 in
Somalia). Since 1990, the AIAI has operated not only in Somalia but also in the Somali
Federal State of Ethiopia. The AIAI supported ONLF activity there until 1997 and the
organizations discussed their activity together. There was also a common trade and
training. It is therefore evident that one of the most active Somali organizations of the
beginning of the civil war in Somalia has also in luenced the Ethiopian organization
ONLF (Terdman, 2008).
Spillover of Con lict from Ethiopia to Djibouti
Djibouti gained independence in 1977 until then it was a French colony. From the very
beginning, there have been disputes between the two main ethnic groups. The Issas
seized power, formed a single-party government in the country - the People’s Union
for Development (RPP – Rassemblement Popuilare pour le Progres) and oppressed the
second largest ethnic group, Afars.2
Afars’ disagreement with the central government
resulted in a civil war between 1991 and 1994. It was a low-intensity con lict, as the
number of victims does not exceed 260 (according to the UCDP), some estimates are
higher but do not exceed 1000 (UCDP, 2017).
2 The Issas are a sub-clan of the Somali Clan Dir, so if we speak of an ethnic group, the Issas should
be labeled as Somalis, but the Issas label is used more often (Lewis, 1999, p. 191).
41
Issue 30, January 2020
The situation within Djibouti was very tense in the 1980s. The Afars were increasingly
marginalized. They were not represented in the government and parliament and were
disadvantaged also in other areas. Tensions increased and also neighboring states had
a negative impact on the situation as well. In Somalia, Barre was deposed in 1991 and
civil war began in the country. In Ethiopia, the Mengistu was deposed and the regime
changed. The unfavorable conditions in the immediate vicinity of the state, together with
the worsening economic situation and economic crisis, had an impact on the already
tense situation in Djibouti. The catalyst for the outbreak of the con lict itself was the
wave of arrests of the Afars during January-August 1991. The arrests were based on
alleged anti-state actions and disruptive tendencies of arrested people (Kadamy, 1996).
At that time, three Afar organizations formed the Front for the Restoration of Unity and
Democracy (FRUD), which began a violent guerrilla struggle against the government in
the Obock region in November 1991 (ARDHD, 2008). In the initial phase of the con lict,
there were series of bloody ighting between the Djibouti army and FRUD, especially in
the north of the country near Obock and Tadjour. FRUD’s units were very successful and
largely managed to occupy this area. Therefore, the central government continued to
arrest all FRUD supporters in all parts of the country. During the irst six months of the
con lict, Djibouti soldiers were very brutal towards the civilian population, with several
incidents in the north of the country, they were iring into an unarmed crowd of civilians.
Since 1992, France has evidently become involved in the con lict. The French representatives
were trying to facilitate peace talks between the rebels and the government
and also deployed French troops to help government troops to stop FRUD (ApNews,
1992). The last major government offensive was launched in July 1993 to clean up the
last cities in the north of the country, which was still under FRUD’s control. Government
troops actually managed to push most of the FRUD’s troops out of the territory, but
the destruction of FRUD’s units did not occur. FRUD retreated to mountain areas while
crossing the boundaries of Eritrea and Somalia, where it began to manage its activity
(Leta, Zeray and Haile, 2015, p. 5). The peace agreement was signed in December 1994,
among other things, the ruling party of the RPP committed to liberalize and democratize
the country. However, this transformation cannot be considered as successful, and
despite tolerated opposition, elections in the country are not free and democratic, and
the RPP continues to control its position of power (Freedom House, 2018).
In this section, two factors are discussed, the timing of the con lict in Djibouti with the
con lict in Ethiopia, as well as the involvement of consensus actors. The Ethiopian Civil
War took place between 1974-1991 and the most intense ighting took place in the late
1980s in three Ethiopian regions in the north - Amharia, Afar, and Tigray. From these
areas that the con lict subsequently spilled over to Djibouti. There was a civil war in
Djibouti between 1991-1994. The time continuity of the con lict is therefore evident, it
is interesting that the con lict spilled at the end of the Ethiopian War, which is justi ied
by the fact that in the 1970s most of the clashes took place in the central part of the
42
Con lict Studies Quarterly
country and eastern Ethiopia, while in the 1980s ighting has moved to the aforementioned
north of the country. The key was the porosity of the border and the occurrence
of the same ethnic groups in a neighboring state. Crossing the Djibouti border occurred
early in the 1980s, particularly the ALF (ALF - Afar Liberation Front) used the border
area and territory of Djibouti as a base for the preparation of attacks and they also recruited
new members from that territory (Yohannes, Hadgu and Ambaye, 2005, p. 35).3
Disputes between the Somalis of the Issas and Afars tribes not only took place in the
territory of Djibouti but also negatively affected the Afar region in Ethiopia. These
disputes date back to the 19th century. Both groups form a nomadic community, while
the vast majority of them profess Islam. Very often there have been disputes over water
and pastures to which the Afars have historically claimed. The situation in Ethiopia and
Djibouti was different. In Ethiopia, Afars have a majority in a regional state, so Issas
are oppressed and discriminated against. They migrated to Djibouti during the second
half of the 20th century to ind a safer place. By contrast, Afars is a minority in Djibouti,
where they are disadvantaged compared to Issas (Ahmed, 2009). During the civil war
in Ethiopia, there was an exchange of inhabitants between the two areas and ALF used
this border area as its base. At the end of the civil war in Ethiopia, some soldiers from
the former Derg regime withdrew from Ethiopia to Djibouti. They left there a large
number of weapons and military equipment after the war. At the same time, some of
the soldiers joined the newly formed FRUD and, in the early stages of the con lict, represented
the main combat component of the FRUD alongside the Eritreans (Yasin, 2010).
Some ALF members also participated in the con lict and supported FRUD inancially,
technically and personally (Vaughan, 2011). This was also because they were dissatisied
with the EPRDF’s progress and developments in Ethiopia after 1991. The Central
Ethiopian government was concerned about the success of the FRUD and weakening
its own position in the country since the Afars are not only a minority in Djibouti but
in Ethiopia and Eritrea. Therefore, they asked the French to intervene against FRUD.
In addition, they arrested several members and supporters of FRUD in their territory
(Mohammed, 1996). Finally, the EPRDF ended cooperation with the ALF and created
a new puppet party representing the Afars’s interests in the country, the Afar People’s
Democratic Organization (APDO). From these data, it is evident that some ALF and
Derg militants who participated in the war in Ethiopia moved after the end of the war
to Djibouti, where they joined the local militias.
3 The Afar Liberation Front (ALF) was established in 1975 in Ethiopia as an armed militia ighting
the Derg in the Afar region. Their main goal was to enforce more rights for the Afar ethnic group,
which was particularly disadvantaged under the regime of Haile Selassie.
43
Issue 30, January 2020
Spillover Effect of Con licts in South Africa
After the end of the Cold War, two con licts spilled over into neighboring states in South
Africa. The irst one is a civil war in Angola, which took place between 1975-2002. In
the initial phase, the main players were three parties, the MPLA (People’s Movement
for the Liberation of Angola), the FNLA (National Liberation Front of Angola) and the
UNITA (National Union for the Total Independence of Angola) (Fiala, 2005). These three
groups concluded an agreement in 1975, which failed after only a few months, and
since the beginning of Angola’s independence (11 November 1975), there has been an
armed struggle in the country. MPLA took advantage of controlling the capital of Luanda
and proclaiming all other parties to be rebels (Dyčka, 2010). In early 1976, the MPLA
launched an offensive with the help of Cuban troops to push UNITA and FNLA out of the
country. In the case of the FNLA, this offensive was truly successful and FNLA leader
Holden Roberto had to lee the country. In the case of UNITA, the organization was
pushed to the south but not pushed out of the country (Ferreira, 2006). In the following
years, neither of these two parties gained a signi icant advantage, UNITA controlled the
south of the country, MPLA the north. In 1988, South Africa agreed to grant Namibia
independence and to stop support UNITA in exchange for withdrawing Cuban troops
from Angola. This stopped the support of the two largest international actors in this
con lict, Cuba and South Africa. The con lict itself ended in April 2002 with an attack
on Savimbi, who died during the attack. An overwhelming majority of UNITA members
followed the Lusaka agreement signed in 1994 (Saunders, 2009). This war spread into
Namibia, where the con lict took place during 1999.
The second case is con lict in Cabinda, which spilled over to the Republic of Congo.4
In
Cabinda, Front of the Liberation Enclave Cabinda (FLEC) has been ighting MPLA since
the 1970s. FLEC attacked MPLA’s units, but the intensity of the con lict was very low
compared to the intensity of the ighting between UNITA and MPLA. FLEC controlled
territories in the densely forested inland areas of the province and sporadically attacked
the army and state targets. In 1983 there were personnel disputes within the
FLEC leadership and the organization was divided into two groups, FLEC-Renovada
(FLEC-R) and FLEC-FAC (Cabinda Armed Forces) (HRW, 2004). At the end of the 1990s,
the intensity of the ighting increased, some UNITA members withdrew to the area
and mass population movements took place. In October 2002, following the end of the
civil war in Angola and the ultimate defeat of UNITA, the Central Angolan government
regrouped its attention to Cabinda and sent 30,000 troops to the area to inally defeat
FLEC. In 2004, the two wings of FLEC were merged, and at the same time, a series of
4 Cabinda is an enclave in the north, which has of icially been part of Angola since 1956, although
it has a different history and is not geographically linked to Angola as it is separated by the DRC
zone (Porto, 2003, p. 11).
44
Con lict Studies Quarterly
peace negotiations between FLEC and MPLA began. The United FLEC, together with
representatives of the Catholic Church of Cabinda and members of the Cabinda Civic
Movement Mpalabanda, subsequently created the Cabindian Dialogue Forum (FCD).
It should represent the interests of Canadians in peace negotiations (OMCT, 2015).
Subsequently, in 2006, a peace treaty was signed between MPLA and FCD representative
António Bento Bembe, who reportedly signed the agreement under duress without the
consent of the majority of the FCD member (Southern Africa Monitor, 2006). Since this
year, the government of icially declares the end of the Cabinda’s war, but members of
the FLEC-FAC formation have not agreed to sign the treaty and continue to run attacks
against government and military targets. Over the past few years, they have attacked
government units and kidnapped foreign workers. They also attacked the national football
team of Togo in 2010 (Guardian, 2010).
Spillover of Con lict from Angola to Namibia
The con lict in Angola spilled into neighboring Namibia into the Caprivi region. The
people in this area have different ethnicities than the rest of Namibia, and since the time
of German colonization, they have longed for independence and the right to self-determination.
After Namibia’s independence in 1990, SWAPO (South-West African People´s
Organization) seized power. Part of the Caprivi local political party DTA (Democratic
Turnhalle Alliance) was dissatis ied with the full domination of SWAPO, and in 1994
they formed an organization called the Caprivi Liberation Army (CLA), which from the
very beginning established a violent solution of the problem. On the 2nd
of August 1999,
the CLA launched an offensive on the city of Katima Mulilo, the capital of the Caprivi
region. CLA’s units attacked several buildings, an army base, a police station, and a state
radio station. During the ighting, 14 people died and more than 3,000 people were
moved to neighboring Botswana (IRIN, 1999). A state of emergency was declared in
the province and the government began to arrest all alleged members of the CLA. In
August and September 1999, there were several other armed clashes between the CLA
and the NDF (National Defense Forces), the last one took place in September. Major CLA
members, including Mishake Muyongo, led to neighboring states, especially Botswana.
Some of them got granted asylum in European countries. Since these events, no other
armed clashes have occurred.
The civil war in Angola took place between 1975-2002. As it was written in previous
sections of the work, MPLA, which was for most of the con lict the ruling party controlled
the north of the country, UNITA controlled the south. The con lict itself was characterized
by a very different intensity in the course of time, the most intense ighting took
place in three periods of time. The irst of these periods was at the beginning of the
con lict between 1976-1977, the second involved the offensive at Cuito Cuanavale, which
represents the years 1985-1988. The third period was the most intense (in terms of
the number of casualties and armed clashes) and this took place between 1998-2000.
45
Issue 30, January 2020
The spillover to Namibia occurred in 1998 when the irst armed clashes between the
rebels and the government began in Namibia. Thus, as in most previous cases, the
con lict spilled over at the stage of its greatest intensity. At the same time, there is also
an apparent time sequence. The interesting thing about this case is that the civil war
in Angola was very long and the con lict spilled over almost at its end. The explanation
is logical, as over the years UNITA had international support and controlled a large
part of the territory. The organization used the territory and got from the territory
inancial resources and physical support. Moreover, South Namibia was dominated by
South African SADF (until the early 1990s). UNITA had no reason to expand beyond
Angola. Everything just changed in the 1990s, when Namibia gained independence,
there were dozens of deserted military camps in its territory, used in the past by SADF
units. These old camps were a great opportunity for UNITA when the organization was
pushed out of most of Angola’s territory ( 1998-2000). For these reasons, the spillover
occurred in the late 1990s.
The Namibian population has been involved in the war in Angola since its inception.
The SADF’s government units consisted mainly of the people of South Africa, but also
of the Namibians. However, a large part of the Namibians was involved in the ighting
on the MPLA’s side, supported by the revolutionary Namibian SWAPO. Even today it
is relatively common in Namibia to meet someone who participated in these battles
and remembers these events well. This is con irmed by the fact that it was indeed a
very high involvement of several thousand. However, after independence, SWAPO and
SADF withdrew from Angolan’s territory in the vast majority of cases. SWAPO began
to concentrate on events in the new state, irst elections, constitutional creation and
consolidation of power. However, there were several personal con licts in the party
and some of its major leaders left the party in the late 1980s and early 1990s. The best
known is Mishake Muyongo, who was until 1980 a vice president of SWAPO. All these
people withdrew to the north of the country, where they formed their own political
organization DTA. Part of the DTA subsequently formed the CLA. In the 1990s, UNITA
lost control of much of its territory, diamond mines, and parts of its physical base, and
it was forced to withdraw to Namibia and Botswana. In Caprivi, UNITA used the former
camps of SADF, which also served to train CLA units. In the mid-1990s, UNITA provided
CLA with ammunition and equipment (Prendergast, 1999). The reason was UNITA’s
hatred of SWAPO for its involvement in MPLA ighting. In 1998, the Namibian NDF army
discovered a training camp in the Caprivi area called Mudumu Game Park, where several
CLA and UNITA members were involved (IRIN, 1999). The link between CLA and UNITA
is, therefore, more than obvious. Moreover, experts from the University of Pretoria
and Cape Town agree on the cooperation between UNITA and CLA (Irishtimes, 1999).
In addition, four CLA members con irmed this link during an interview in Botswana
(IrinNews, 1999). In this case, the rule is that the common enemy, which was (for UNITA
and CLA) the SWAPO party, makes friends from two opposing parties.
46
Con lict Studies Quarterly
Spillover of Con lict from Angola (Cabinda) to the Republic of Congo
The war in Angola and the war in the province of Cabinda spilled over into the Republic
of Congo. There were two civil wars in the 90s. The irst one took place between 1993-
1994, and in terms of scale and intensity, the war was much milder and not so intense
compared to the second civil war that took place between 1997-1999. The irst civil
war took place because of the disputes that followed the 1992 elections. The country
was divided into three areas based on the support of individual candidates, the key
was the ethnic division. In addition, militias were formed in these areas, the so-called
Zulu militias supported Pascal Lissouba, Ninja militias supported Bernard Kolelas and
Cobra militias supported President Sassou Nguessa (Englebert & Ron, 2004). Lissouba’s
supporters have created other militias, such as the Cocoyes militia or Mambas militias.
Fighting began between the Zulu and Ninja militias in the capital Brazzaville in
the second half of 1993. During these ights, 2,000 people died. Ninja militias did not
participate too much in the ighting. The violence ended in 1994.
Over the next two years, there was a restless peace in the country, which was occasionally
disrupted by brief episodes of violence. New elections were to be held in July 1997,
but the Lissouba’s and Sassou’s supporters had been preparing for armed confrontation
since March. In May 1997, Sassou decided to candidate in the northern city of
Owando, where the former President and former Prime Minister of Lissouba, Jacques
Yhombi-Opango, came from. His speech led to the outbreak of violence between local
supporters of Sassou (mostly Mbochi) and supporters of Yhombi (mostly Kouyou). In
addition, Lissouba discovered that Sassou was arming his militias, so he decided to send
an army to Sassou’s residence to arrest some of his collaborators who were reportedly
involved in the violence in Owando. Everything ended in a bloody shootout that
unleashed violence throughout the capital. Thus the second civil war in Congo began
(Trapnell & P lepsen, 2005). At the beginning of the war, the militias of Kolelas and
Sassou cooperated against the militias of Lissouba, and Sassou was also supported by
France. Lissouba cooperated with UNITA and FLEC troops, who he allowed since 1992
to legally reside in Congo. Lissouba and UNITA rebels have been working together since
the 1990s. In the second largest city of Pointe-Noire, they smuggled weapons to assist
local ighters. Angola joined the war in October 1997, sending about 3,000 troops to help
Sassou. This induced Kolelas’s cooperation with Lissouba. On the one side, the Ninja
militia and the Cocoyes fought, on the other, the Angolan army and the Cobra militia.
Sassou, thanks to the help of Angola, quickly gained the upper hand and Lissouba and
Kolelas led into exile in 1998. In the second half of 1998, the ighting shifted to the
south of the country, but since January 1999 the intensity of the ighting decreased,
until the end of 1999 rebellion ended in an agreement that provided amnesty to all
commanders of Ninja and Cocoyes militias (Atzili, 2007). Sassou Nguesso came back
to power and he reigns in the country until nowadays.
47
Issue 30, January 2020
The con lict in Angola itself took place between 1975-2002, and in Cabinda, the conlict
lasted four years later, until 2006 (it is the of icial end on the basis of signing the
peace treaty and the government declaration). Two civil wars took place in the Republic
of Congo, in 1993-1994 and in 1997-1999. In the Second Civil War, UNITA and FLEC
troops were involved, but also Angolan soldiers from the MPLA-controlled state army.
Their presence decided the outcome of the con lict. The timing of the spillover of the
con lict is evident in this case. The con lict spilled over at a time when there was still
war in Angola. This is a very special period in terms of intensity since there was still
a cease ire in Angola in 1997, but since the beginning of 1998, bloody ights started
again and the intensity of the con lict was very high. The key thing is that during this
period, UNITA was pushed out of 2/3 of the country’s territory, and that is why it
largely occupied the territory of the Republic of Congo and the DRC. At this time, FLEC
controlled very limited, mostly rural areas of Cabinda and, like UNITA, made extensive
use of neighboring states. So it can be said that this is a period when the presence of
UNITA and FLEC troops in Congolese territory was the highest. In terms of intensity,
however, the con lict was not spilled in its most intense phase.
Since the beginning of the Second Civil War in Congo, the cooperation of various militias
with UNITA and FLEC troops has been evident. There was very extensive cooperation
between the Lissouba’s militias, especially the Zulu and Cocoyes militias, with the
Angolan groups. UNITA and FLEC troops have been ighting on Lissouba’s side since
early 1997 and were included in his militia. As a result, UNITA received air shipments
of weapons from the town of Pointe Noire, which was completely under the control of
Lissouba (Turner, 2002). FLEC troops, on the other hand, received primarily inancial
support, but between them and the Lissouba’s militias was cooperation since 1992,
which included mainly arms trade, but in some cases, joint training of units. Angolan
President Dos Santos was very worried about how the Congo war could affect the outcome
of the civil war in Angola. Angolan military units (FAA - Angolan Armed Forces)
involved in the ighting on the side of the Cobra militia. The intervention itself took
place on October 12, 1997, when 3000 FAA soldiers entered the country. As a result,
Sassou gained a very quick advantage and subsequently won.
Analysis of Spillover Effect
In this part of the work, a qualitative comparative analysis is used to compare conditions
in all neighboring states of countries where con licts took place and subsequently spilled
over and in the other neighboring countries where it did not happen. For a detailed
comparison of the factors described in the methodological section, the truth table has
been created. If a given factor occurred in a neighboring state, it is assigned a value of
1, if not a value of 0. At the same time, it should be noted that I work with data from the
time, when the con lict spilled over, not the latest data. The form of truth table can be
seen in Annex 1. A computer program (Tosmana) was used to evaluate the results. In
48
Con lict Studies Quarterly
the beginning, it is necessary to explain the crosses in the table. In the case of the civil
war in Somalia, a different political ideology was not included in the analysis. This is
because there was no central government in Somalia in the 1990s to control most of
the country and maintain a certain ideology. The country was divided among various
factions that had different ideas and motives. Therefore, the state ideology cannot be
compared with the ideologies of neighboring states. It is evident, at irst sight from
the table that economic factors did not play any role in these cases and therefore we
can say there is no link between economic factors and the occurrence of the spillover
effect. Therefore, the natural resources dimension does not play a signi icant role in
these cases as it was only present in the case of the Republic of Congo. According to
Collier, dependence on the export of one commodity increases the chance of con lict,
but in practice of spillover effect, it does not apply. At the same time, it is obvious that
there is no single explanation to clarify all the cases. Therefore, it is not possible to talk
about causality and generalize the conclusions. There is only a coherence between the
combination of factors and the occurrence of a spillover effect. To be able to talk about
causality we would have to work with a greater number of cases. However, it is also
clear from this analysis that some factors are more important than others.
Annex 1: Truth Table of Spillover in East and South Africa
Case ID
Soca
Socb
Socc
Socd
Soce
Socf
Pola
Polb
Polc
Pold
Pole
Polf
Ecoa
Ecob
Ecoc
Geoa
Geob
Geoc
Geod
Geoe
Spillover
Ethiopia 1 1 0 1 1 1 X 1 0 1 1 0 0 0 0 1 1 1 1 1 1
Kenya 0 0 0 0 0 1 X 1 0 1 0 0 0 0 0 0 0 1 1 1 0
Djibouti 1 1 0 0 1 0 X 0 1 1 1 1 0 0 1 0 0 0 0 0 0
Djibouti 2 1 1 0 1 1 1 1 0 1 1 1 1 0 0 1 0 0 0 0 0 1
Kenya 0 1 0 0 0 0 1 1 0 1 0 0 0 0 0 0 0 1 1 1 0
Somalia 1 1 0 1 1 0 0 1 1 1 1 0 0 0 1 1 1 0 0 0 0
Sudan 0 1 0 0 0 0 1 1 1 1 1 0 1 1 0 0 1 1 1 1 0
Congo 0 1 0 0 0 0 0 0 1 1 1 1 1 1 0 0 0 0 0 0 1
DRC 0 1 0 0 1 1 0 1 1 1 1 0 1 0 1 1 1 1 1 1 0
Namibia 1 1 0 0 0 0 0 0 1 1 1 1 0 0 0 1 0 0 1 1 1
Zambia 0 1 0 0 0 0 0 0 1 0 0 1 0 0 1 1 0 0 1 1 0
Source: Author
Overall, there are twelve combinations of factors whose presence in a neighboring state
increases the likelihood of con lict’s spillover. The political-security factors that have
the most frequent representation seem to be crucial. Among these factors, the most
represented factors are different political ideology, political instability, the existence of
armed active militias and having a small army. Besides political-security factors, social
and geographical are important as well. The most frequented social factor is a high
49
Issue 30, January 2020
occurrence of the same ethnicity in a neighboring state and the most frequented geographical
factors are uneven ethnic demography and having a porous and long border.
These are a combination of factors that increase the likelihood of spillover con lict and
can even lead to spillover itself:
1. Political instability + existence of armed active militias + small army
2. Political instability + different ideology + small army
3. High occurrence of the same ethnicity + high crime rate + small army + porous
and long border
4. Existence of armed active militias + small army + uneven ethnic demography
5. High crime rate + small army + uneven ethnic demography
6. Existence of armed active militias + small army + porous and long border
7. High occurrence of the same ethnicity + different ideology + small army
8. High occurrence of the same ethnicity + small army + porous and long border
9. Porous and long border + uneven population demography + uneven ethnic de-
mography
10. Existence of armed active militias + porous and long border + uneven ethnic de-
mography
11. High occurrence of the same ethnicity + porous and long border + uneven ethnic
demography
12. High occurrence of the same ethnicity + high occurrence of the same religion +
high in lux of refugees
As you can see it at the combinations of factors, there are scenarios created just from
political-security factors, but also scenarios that contain just social or geographical
factors. It is therefore not possible to say which category is the most important and it is
clear that all three categories (political-security, social and geographical) are important
during con lict spillovers. On the other hand, economic factors are marginal compared to
other factors. If we apply this analysis to theories about the description of the spillover
effect and the causes of con licts, we can say that the most accurate theories are the
theories that Silver and Verdier created. However, even these theories do not fully cover
our conclusions, and it is clear that spillover is a much more complex phenomenon,
which must be explained by a combination of several factors, moreover, as can be seen
from the analysis, there are several typologies of spillover.
Conclusion
The spillover effect is a phenomenon that accompanies some con licts. It is most often
seen in Africa and that is why this article deals with the analysis of the spillover effect
on cases from South and East Africa. In the past, very few surveys on spillover effect
analysis with the use of real examples have been conducted. Therefore, understanding
and analyzing this phenomenon is very important.
50
Con lict Studies Quarterly
In the theoretical part of this work, the conditions that distinguish the spillover effect
from the outbreak of a new con lict were clearly de ined. This includes time continuity,
the involvement of some of the same actors and the cooperation of local and neighboring
actors. At the same time, I also focused on describing and determining the moment
when the con lict spilled over (in terms of the intensity) to the neighboring state.
Whether the con lict spilled over when the con lict was most intense, at the time of the
cease ire or after the con lict ended. The analysis shows that in all cases the spillover
time continuity has been met. In three cases the spillover effect occurred at a time when
the con lict had the greatest intensity. This is con irmed by a spillover of the con lict
from Somalia, from Ethiopia and also by spillover from Angola to Namibia. The case
of the spillover of the con lict from Angola to the Republic of Congo is speci ic. There
was still a cease ire in Angola in 1997, but since the beginning of 1998, bloody ights
have started again and the intensity of the con lict has been very high. The key thing
is that during this period, UNITA was pushed out of 2/3 of the country’s territory, and
that is why it largely occupied the territory of the Republic of Congo. In addition, other
organizations, such as MPLA units, were not involved in the con lict until 1998. In all
cases, there was a spillover, when the ighting took place near the border or when the
troops were pushed beyond the border. In all cases, the actors from the neighboring
state were involved in con licts and worked with local militias. They physically engaged
in the ighting and fought on the side of their allies in all four cases. In some cases,
inancial and logistical support or bartering has also taken place, but physical involvement
is crucial. From these cases, it is concluded that if there is intense ighting along
the border, the chance of spillover is higher.
The second part of the work deals with the analysis of factors in neighboring states that
increase the probability of spillover effect. In total, there are twelve combinations of
factors that are key to spillover. The analysis shows that economic factors are irrelevant.
On the other hand, we cannot say which of the remaining category of factors is the
most important as there is always a scenario that increases the likelihood of spillovers
and involves combinations of factors from only one category (either political-security
or only social or geographic). However, the most prevalent factors can be identi ied.
Among political- security factors it is: different political ideology, political instability,
the existence of armed active militias and having a small army. The most frequented
social factor is a high occurrence of the same ethnicity in a neighboring state and the
most frequented geographical factors are uneven ethnic demography and having a porous
and long border. All these factors are logical, because the long and porous borders
make it easier to spillover the con lict, especially when ighting takes place along the
border. If the era on a second side of the border is inhabited by the same ethnic group
this area can serve as a good base for the militia. The same ethnicity increases the sense
of belonging and this often leads to support for the militia. In cases, when the army is
small it is not able to control its entire territory. Political instability leads to dissatisfac-
51
Issue 30, January 2020
tion among the local population and often creates space for new (sometimes violent)
changes. In cases where there are already present armed active militias in the country,
it is clear that something is not working in that state. This, in combination with other
factors, can lead to the spillover of the con lict itself. It is therefore clear that these factors
are not coincidental, and states in the immediate vicinity of the con lict should try
to work on these issues to avoid the spillover of the con lict.
References
1. African Development Bank. (2008). The consequences of con lict. Abidjan: African Development
Bank.
2. Ahmed, R. (2009). Undemocratic Threats are Hanging Over the Issa Tribe in Ethiopia.
Paris: ARDHD.
3. ApNews. (1992, February 25). 250 French Troops Sent to Djibouti War Zone. Retrieved
from https://apnews.com/d7b4b1b6a29a468f05130adad5ad7975.
4. ARDHD - Association pour le Respect des Droits de l’Homme à Djibouti. (2008, March
12). Pour la 12ème commémoration du massacre d’Arhiba.. Retrieved from http://www.
ardhd.org/arhiba.asp.
5. Atzili, B. (2007). When Good Fences Make Bad Neighbors Fixed Borders, State Weakness,
and International Con lict. International Security, 31, 139-173.
6. Bosker, M., & de Ree, J. (2009). Localizing con lict spillovers: introducing regional heterogeneity
in con lict studies. Groningen: University of Groningen.
7. Beneš, Vít, & Drulák, Petr. (2016). Metodologie výzkumu politiky. Brno: Masaryk Uni-
versity.
8. Caramani, D. (2009). Introduction to the Comparative Method with Boolean Algebra. Los
Angeles: Sage Publications.
9. Carmignani, F., & Kler, P. (2013). Surrounded by wars: quantifying the role of spatial
con lict spillovers. South East Queensland: Grif ith University.
10. Collier, P., & Hoef ler, A. (1999). Greed and Grievance in Civil War. NY: WorldBank.
11. Collier, P. (2007). Miliarda Nejchudších: Proč se některýmzemímnedaří a co s tím. Prague:
Vyšehrad.
12. Dagne, T. (2010). Horn of Africa: Current Conditions and U.S Policy Committee on Africa
and Global Health. Washington: FAC.
13. Dyčka, L. (2010). Aktéři surovinového kon liktu v Angole. Brno: Masaryk University.
14. Elbadawi, I., & Sambanis, N. (2009). Why Are There So Many Civil Wars in Africa? Understanding
and Preventing Violent Con lict. Washington: World Bank.
15. Enghoff, M., Hansen, B. et. al. (2010). Socio-economic and Environmental Impacts of
Dadaab Refugee Camps on Host Communities. Kenya: Kenya Red Cross Society.
16. Englebert, P., & Ron, J. (2004). Primary Commodities and War: Congo-Brazzaville’s Ambivalent
Resource Curse. Comparative Politics, 37, 61-81.
17. Ethiopian Government. (2018, January 25). Somali Regional State. Retrieved from
http://www.ethiopia.gov.et/somali-regional-state.
52
Con lict Studies Quarterly
18. Ethiopian Parliament. (2017, Match 7). Ethiopian Parliament State Information. Retrieved
from http://www.ethiopar.net/.
19. Ferreira, M. E. (2006). Angola Con lict and Development. The Economics of Peace and
Security Journal, 1, 3-12.
20. Fiala, V. (2005). Vznik politických stran v zemích lusofonní Afriky. Politologická revue,
11, 91-138.
21. Fortna, V. P. (2008). Peacekeeping and the Peackept: Data on Peacekeeping in Civil Wars.
New York: Columbia University Press.
22. Freedom House. (2018, January 25). Djibouti. Retrieved from https://freedomhouse.
org/report/freedomworld/2011/djibouti?page=22&year=2011&country=8026.
23. Galula, D. (1964). Counterinsurgency Warfare. New York: Frederick A. Praeger.
24. Gastrow, P. (2011). Termites at Work: Transnational Organized Crime and State Erosion
in Kenya. New York: International Peace Institute.
25. Gomez, M. P., Christensen, A., Araya, Y., & Harild, N. (2010). The Impacts of Refugees
on Neighboring Countries: A Development Challenge. Copenhagen: Social Development
Department.
26. Guardian. Togo footballers were attacked by mistake, Angolan rebels say. (2010, June 5).
Retrieved from https://www.theguardian.com/world/2010/jan/11/two-arrested-togo-
football-attack.
27. Harar24. (2017, September 11). Ethiopian Soldiers killed in a heavy ighting between
ONLF and Ethiopian Soldiers. Retrieved from http://harar24.com/?p=16329.
28. Home Of ice UK. (2004). Ethiopia Country Report. London: Home Of ice.
29. Home Of ice UK. (2017). Country Policy and Information Note: Ethiopia Opposition to
the Government. London: Home Of ice.
30. Holsti, K.J. (1991). Armed Con licts and International Order. Cambridge: Cambridge Uni-
versity.
31. Irin. (1999). Weekly Round-up 32 covering the period 7-13 August 1999. Johannesburg:
IRIN.
32. IrinNews. (1999, September). UNITA training Caprivi rebels. Retrieved from http://
www.irinnews.org/report/7387/angola-namibia-unita-training-caprivi-rebels.
33. Irish Times. (1999, May 7). Namibia blames Savimbi for Caprivi secessionist rebellion.
Retrieved from https://www.irishtimes.com/culture/namibia-blames-savimbi-for-
caprivi-secessionist-rebellion-1.213335.
34. HRW. (2004). Angola Between War and Peace in Cabinda. Luanda: HRW.
35. IRB - Immigration and Refugee Board. (1999). Ethiopia: The Shekhash, a clan of Somali
origin and their treatment in Ethiopia. Ottawa: IRB.
36. Kadamy, M. (1996). Djibouti Between War and Peace. Review of African Political Economy,
23, 511-521.
37. Kouba, K. (2008). Využití Millových metod ve srovnávací politologii: metodologické
předpoklady a problémy. Olomouc: University of Palacky.
38. Leta, A., Zeray, Y., & Haile, Y. (2015). The Politics of Dyadic Community: The Case of Djibouti.
Ethiopia: Mekelle University.
53
Issue 30, January 2020
39. Lewis, I.M. (1999). A Pastoral Democracy: A Study of Pastoralism and Politics Among the
Northern Somali of the Horn of Africa. Munster: International African Institute.
40. Menkhaus, K. (2007). The crisis in Somalia: Tragedy in ive acts. Oxford: Oxford University
Press.
41. Millett. R. L. (2002). Colombia’s con lict – The spillover effects of wider war. Carlisle: US
Army War College.
42. Mohammed, R. (1996). Djibouti: Between War and Peace. Connecticut: Praeger.
43. Ncube, M., & Jones, B. (2013). Drivers and Dynamics of Fragility in Africa. Pretoria: AfDB.
44. OMCT. (2015). Angola – They Want to Keep Us Vulnerable. Luanda: OMCT.
45. Opalo, K. (2010). The Ogaden Region: A Fragile Path to Peace. Pretoria: Africa Back-
grounder.
46. Patrick, S. (2006). Weak States and Global Threats: Assessing Evidence of „Spillovers“.
London: Center for Global Development.
47. Pham, J. P. (2013). State Collapse, Insurgency and Counterinsurgency Lessons from Somalia.
Carlisle PA: SSI.
48. Prendergast, J. (1999). Angola’s Deadly War. Washington: USIP.
49. Porto, J. G. (2003). Cabinda, Notes on a soon-to-be-forgotten war. Pretoria: Institute for
Security Studies.
50. Ragin, C. (2008). Redesigning Social Inquiry: Fuzzy Sets and Beyond. Chicago: University
of Chicago Press.
51. Saunders, S. (2009). Angola from war to peace. Cape Town: University of Cape Town.
52. Silve, A., & Verdier, T. (2018). Theory of regional con lict complexes. Québec: CRREP.
53. Smolík, J., & Šmíd, T. (2010). Vybrané bezpečnostní hrozby a rizika 21. století. Brno: Mezinárodní
politologický ústav.
54. Southern Africa Monitor. (2006). Peace Deal Lacks Credibility, 11, 1-3.
55. Terdman, M. (2008). Somalia at War – Between Radical Islam and Tribal Politics. Tel
Aviv: Daniel Abraham Center for International and Regional Studies.
56. Terrill, A. (2008). Regional spillover effect of the Iraq war. Washington: U.S. Government.
57. Trapnell, T., & P lepsen, L. (2005). The Republic of Congo. New Jersey: Princeton Uni-
versity.
58. Turner, T. (2002). Angola’s Role in the Congo War. In J.F. Clark (Ed.), The African Stakes
of the Congo War. New York: Palgrave Macmillan.
59. UCDP -Uppsala Con lict Data Program (2017, January). Battle Related Deaths Dataset
1989-2016. Retrieved from http://ucdp.uu.se/downloads/.
60. Vaughan, S. (2011). Ethnicity and Power in Ethiopia. Edinburgh: University of Edin-
burgh.
61. Vestal, T. (1999). Ethiopia: A Post Cold War African State. Santa Barbara: Praeger Pub-
lishers.
62. Williams, Paul D. (2009). Into the Mogadishu Maelstrom: The African Union Mission in
Somalia. International Peacekeeping, 16, 514-530.
54
Con lict Studies Quarterly
63. Whitaker, E.B. (2003). Refugees and the Spread of Con lict: Contrasting in Central Africa.
North Carolina: University of North Carolina.
64. Yasin, M. (2010). Regional Dynamics of Inter-ethnic Con licts in the Horn of Africa: An
Analysis of the Afar-Somali Con lict in Ethiopia and Djibouti. Hamburg: The University
of Hamburg.
65. Yohannes, G. M., Hadgu, K., & Ambaye, Z. (2005). Addressing Pastoralist Con licts in Ethiopia.
London: Safeworld.