Kenya-Somalia Border Conflict

Analysis

Conflict Prevention, Mitigation, and Response Program

for East and Southern' Africa (CPMRlESA)
Managing African Conflict IQC
Contract No. 623-1-00-03-00050-00
Task Order 001

This report was prepared for the U.S. Agency for International Development (USAID) with
funding provided through the USAID Regional Economic Development Services Office
(REDSO) in Nairobi, Kenya. Work was contracted to Development Alternatives Inc. under Task
Order 001 of the Managing African Conflict (MAC) Program. The report is a component of the
Peace in East and Central Africa (PEACE) Program which is managed by the USAID/REDSO
Regional Conflict Management and Governance Office.

DISCLAIMER
The Author's views expressed in this publication do not necessarily reflect the views of the United States
Agency for International Development or the United States Government

~
DAI
Development Alt(!rnatives, Inc.
Kalson Towers 8tn Floor; P,O. Box 13403; 00800 Nairobi, Kenya
Tel: (254-20) 3755541/2; Fax: (254-20) 3755543 Email: info@daLcom
 CONTENTS

EXECUTIVE SUMMARY

SECTION I: Background, Context, and History ofthe Kenya-Somalia Border Area

1. Context
1.1. A Troubled Region 1
1.2 Limited Tools of Explanation 2
1.3 Geography 3
1.4 Human Settlement and Productive Activities 4
1.5 Identity and Ethnicity 6
1.6 Key Historical Themes 9

2. Key Changes in the Kenya-Somalia Border Area since 1990

2.1 Changes in Southern Somalia 11
2.2 Changes in Northeast Kenya 14

SECTION II: Contemporary Conflict Analysis ofthe Border Area

3. Contemporary Trend Analysis of the Border Area 17

4. Conflict Mapping - "First-Generation" Post-I 990 Conflicts in Northeast Kenya

4.1 Wajir Conflicts, 1992-93 and 2000-01 20
4.2 Isiolo District Conflicts and Displacement. 24
4.3 Aulihan-Abdwaq clashes, Garissa, 1998-2000 25

5. Conflict Mapping - Contemporary Conflict Zones along the Kenya-Somalia Border

5.1 Mandera 27
5.2 El Wak 29
5.3 Beled Hawa 36
5.4 The Southern Border/Dobley 38

6. Key Sources of Conflict in the Border Area

6.1 Environmental Stress and Poverty 43
6.2 Kenyan State and Local Government Policies 44
6.3 State Collapse and State-Building in Somalia 45
6.4 Hardening of Ethnic Identity 46
6.5 Regional Economy/Cross Border Trade 46
6.6 Outsiders 46
6.7 Contested Urban Space 47
6.8 Regional Spillover. 48
6.9 Crime 48
6.10 Small Arms Proliferation 49
6.11 Borders 49
7. Sources of Peace, Security, and Conflict Management
7.1 Customary Law and Traditional Elders 50
7.2 Civic Groups 52
7.3 Peace Committees 52
7.4 Kenyan State 52
7.5 Business Interests 53
7.6 Islam 53
7.7 "Cosmopolitan" Towns 53

SECTION III: Policy Issues and Considerations

8. Promise and Limits of the "Mediated State"

8.1 The Mediated State as Concept and Explanatory Theory 55
8.2 The Mediated State and Strategies of Peace-building and
Governance-building 58
8.3 Prospects for a Mediated State in Somalia 60
8.4 Addressing Underlying Drivers of Conflict. 62

BIBLIOGRAPHY
Acronyms

ADRA - Adventist Development and Relief Association

AlAI - al-Ittihad al-Islamiyya
ALRMP - Arid Lands Resource Management Project
ASEP - Advancement of Small Enterprise Project
AU - African Union
CMC - Coordinating and Monitoring Committee
CBO - Community-Based Organization
CRD - Center for Research and Development
CT - Counter-terrorism
DC - District Commissioner
DSDO - District Social Development Office
EPAG - Emergency Pastoralist Assistance Group
GHC - Gedo Health Consortium
GPG - Gedo Peace Group
GSU - General Service Unit
IGAD - Inter-Governmental Authority on Development
IDP - Internally Displaced Person
ITDG - Intermediate Technology Development Group
JIST - Joint Initiative Strategic Team
JVA - Jubba Valley Alliance
KHRC -- Kenya Human Rights Commission
MDPC - Mandera District Peace Committee
MEDS - Mandera Educational Development Society
MP - Member of Parliament
NORDA - Northern Region Development Agency
NGO - Non-governmental organization
NSS - National Security Service
OCHA - Office of Coordination for Humanitarian Affairs
OLF - Oromo Liberation Front
PDC - Peace and Development Committee
SACB - Somalia Aid Coordination Body
SNA - Somali National Alliance
SNF - Somali National Front
SRRC - Somali Reconciliation and Reconstruction Council
SUPKEM - Supreme Council of Kenya Muslims
TFG - Transitional Federal Government
TFP - Transitional Federal Parliament
TNG - Transitional National Government
TFI - Transitional Federal Institutions
UNDP - United Nations Development Programme
WASDA - Wajir Southwestern Association
WPDC - Wajir Peace and Development Committee
WCC - Women Care and Concern
WFPD - Women for Peace and Development
 Executive Summary

General

• Chronic instability along the Kenya-Somalia border zone is part of a larger pattern
of state failure, lawlessness, and communal violence afflicting the Kenyan border
areas from Uganda to Somalia, frequently described as "not peace not war." Local
communities suffer levels of displacement and casualties akin to civil war, but in
a context of sporadic, low-intensity communal clashes punctuated by extended
periods of uneasy peace. Spoilers embrace armed conflict not in pursuit of victory
but to create conditions of "durable disorder" from which they profiteer.
Conventional conflict prevention and management approaches have generally
been frustrated in the face of these unconventional conflict dynamics.
• Semi-arid, pastoralist zones in the border areas of Kenya constitute the "frontier"
area, where state capacity to exercise authority is weak to non-existent. The
absence of the state in these areas breeds lawlessness and compels local
communities to rely upon informal systems of protection, usually involving a
combination of tribal or clan militias (for deterrence and retaliation) and
traditional authorities and customs (for conflict management and justice).
• Distinguishing characteristics of the Kenya-Somalia border areas include: the
complete absence of a state counterpart on the Somali side of the border; the
existence of more robust forms of local, informal governance and conflict
management than anywhere else in Kenya's border regions; the rise of vibrant
cross-border trade of commercial goods and cattle; and the dominance of a single
ethnic group, (the Somali), on both sides of the border.
• Since 1995, a number of local factors have contributed to improved security and
informal governance on both sides of the Kenya-Somalia border, especially in
northeastern Kenya. Progress suggests that local peace and conflict prevention
mechanisms have real promise; however, since 2004, serious armed clashes in
Mandera and EI Wak have rendered the region highly insecure and are indicators
that local conflict prevention mechanisms are not a panacea and face limits in
their ability to stem conflicts born of much broader, structural forces at play in the
regIOn.

Key Structural Sources of Border Area Conflict

• The level of poverty, unemployment, and underdevelopment in the Kenya-

Somalia border area is among the highest in the country and is a major contributor
to crime, insecurity, and alienation.
• Environmental degradation of rangelands contributes to increased communal
competition and pastoral conflicts over water and rangeland are endemic. A long
and on-going western expansionism by Somali pastoralists at the expense of other
groups contributes to periodic clashes over land.
• Competition over new and growing urban settlements is a more immediate driver
of conflict in the Kenya-Somalia border area. Towns and villages are important
sites of trade and aid. On the Kenyan side of the border "locations" serve as seats

-i-
 of local government, conferring upon those who control them paid positions as
chiefs and assistant chiefs and control over local patronage.
• Dramatic expansion of cross-border commerce from Somalia into Kenya has had
a variable affect on conflict, at time serving as a force for cross-clan collaboration
and basic security, and at other times producing conflict over control of key trade
routes.
• Competing clans increasingly view control over locations not merely in
administrative terms but as a means for establishing exclusionary zones within
which they can evict or block other clans from access to pasture and business
activities. The result is misuse of locations to engage in localized ethnic cleansing,
which in tum greatly increases the political stakes for control over locations.
• Recent attempts to revive the state-building exercises inside Somalia have
contributed to armed violence on the Somali side of the border, especially in EI
Wak, where the Marehan and Garre clans are jostling fiercely to expand or
maintain their control over land in order to maximize the number of parliamentary
and cabinet seats they hope to claim.
• The trend toward clan or tribally-based locations in Kenya, ethno-states in
Ethiopia, and proportional clan-based representation in Somalia's nascent federal
government has led to a "hardening" of ethnic identities in northern Kenya (where
identity was previously more flexible and nuanced) and some ethnic groups now
face an increasingly exclusionist political environment.
• Spillover from protracted state collapse in Somalia has been a major driver of
conflict in the border areas, producing destabilizing flows of refugees, gun-
smuggling, banditry, warfare, and clan tensions. But Somalia's collapsed state is
not the sole or even most important source of insecurity in northern Kenya.
• Some Kenyan state actors have been a source of conflict rather than a source of
prevention and mediation. Under the previous administration, government officers
reportedly were complicit in commercialized livestock rustling in the region.
Today, some Members of Parliament in the region are accused of fomenting
ethnic tensions in pursuit of parochial political gain.
• On both sides of the border, the arrival of newcomer, or galti, clansmen has been
a major source of destabilization. The outsiders are members of local clans but
hail from other regions, are typically much better armed, and are not stakeholders
in local peace processes. Much of the internal conflicts plaguing the Marehan clan
in Gedo region are animated by tensions between indigenous (guri) and galti
Marehan. The current conflict in EI Wak has drawn heavily on outside Marehan
from Kismayo and Garre militia from Ethiopia.

Key Precipitating Factors and Accelerators of Border Area Conflict

• Acts of crime - principally stolen vehicles, rape, and murder - are typically the
sparks which produce widespread communal violence. Even clans with
historically close ties have had difficulty preventing reprisal attacks which then
provoke larger cycles of violence. Use of the border by criminals to escape
apprehension aggravates the problem.

- ii -
 • Outside elements - political and business leaders in Nairobi, merchants from
outside clans, the diaspora, and the galti interests noted above - have exploited or
fomented inter-clan tensions in the region for a variety of reasons. Though local
communities at times exaggerate this factor to absolve themselves of
responsibility, there have been several instances in which interests in Nairobi or
Mogadishu have accelerated local conflicts with military assistance or political
meddling.
• Local spoilers have exploited local tensions and blocked reconciliation efforts in
pursuit of their political or economic interests. Warlordism is no longer as acute a
problem today as in the early 1990s, but some spoilers remain, particularly those
local actors operating businesses which rely on humanitarian aid contracts.
• The proliferation of small arms in the border area has increased the flammability
of local conflicts and increased the carnage of local raids. Customary clan law and
blood compensation mechanisms, designed to manage small numbers of
casualties, are overwhelmed by raids and attacks in which dozens of people die.

Key Sources of Conflict Prevention and Management

• Though the Kenya-Somalia border area remains chronically insecure and prone to
flare-ups of deadly violence, the region is dramatically more secure than was the
case in the early 1990s. Understanding the sources of this improved peace and
security, offers important clues to managing insecurity in other troubled,
ungoverned border zones in the Hom of Africa.
• Since the mid 1990's, the Kenyan government has been willing but unable to
extend its authority into the border regions. The result has been a "mediated state"
arrangement, in which the Kenyan government partners with local non-state,
civic, and traditional actors to fulfill core functions including conflict mediation,
cross-border diplomacy, and the dispensation ofjustice, normally associated with
the state, For communities along the Kenyan-Somali border the mediated state
approach is a major departure in local experience with the state.
• The local partners which the Kenyan government works through in this mediated
state system are organized in local Peace Committees (PCs), umbrella groups of
local CBOs, including traditional clan elders and a government representative.
The PCs have varied in performance', but overall have been the single most
important factor in the dramatic reversal of anarchy and insecurity in the region.
• The PCs success is due to several key features - a good functional relationship
with the Kenyan government, which has generally given the PCs adequate space
and autonomy to operate; strong local ownership; commitment and knowledge of
local conflicts; open, flexible membership combining traditional and civic
leadership; international financial support; a strong spillover effect, in which one
successful PC is emulated by others in other regions and across the border; and a
nascent institutional learning capacity, in which lessons are shared by PCs from
one region to the next.
• Reliance on customary clan law and traditional elders to enforce it has at times
played an important role in managing conflict and reducing or deterring crime.

- iii -
 • Religious leadership, including Islamic leadership, has played a prominent role in
pressuring local parties to reach accords. For example, in Mandera, the recent
Murille-Garre accord was reached largely due to mediation by the Supreme
Council of Kenyan Muslims (SUPKEM).
• Because most of the economic activity in the region is long-distance commerce,
which requires safe roads, business partnerships and interests generally work
toward peace and security in the region. Most commercial ventures require cross-
clan partnerships to insure access and protection, and those cross-clan
partnerships provide lines of communication and shared interests.

Policy Considerations: Addressing the Mediated State

• For Kenya's northern border areas, the mediated state approach may be the only
alternative to anarchy in the short to medium term. As an approach to state-
building, it has a number of advantages. It is flexible with regard to selection of
local partners; it provides the state with governing partners who possess deep
knowledge of local affairs and who are real stakeholders in promoting peace; and
it allows external aid agencies seeking to assist with both state-building and
conflict prevention the opportunity to work both to improve government capacity
and provide local CBOs with needed support.
• The flexible mediated-state system of governance is evolving in a number of other
"willing but not able" states in Africa, but Kenya appears to be the most advanced
case. The Kenyan experiment thus warrants close monitoring and could develop
as an important model for lessons learned that could be applied elsewhere.
• Assistance to the mediated state can and should take a number of forms, from
direct financial support to the PCs to support for state offices liaising with PCs to
the convening of workshops allowing PCs from different regions to share lessons.
• Care must be taken not to compromise the PCs through too close an affiliation
with external aid agencies. This is especially the case in sensitive border areas like
northern Kenya, where Western counter-terrorism (CT) efforts are prominent and
where any local linkage with Western agencies may be misconstrued and lives put
at risk.
• As a framework for state-building, the mediated state model may be of even
greater utility in Somalia, where most anticipate the emergence of a weak
government which will depend on local intermediaries to help govern its remote
frontier zones for the foreseeable future.
• In both Somalia and Kenya, support to local PCs is a complement to, not a
substitute for, the larger enterprise of building state capacity to govern its
hinterlands.

Policy Considerations: Addressing Structural Drivers of Conflict in the Border Area

• Strengthening the governing capacity of the mediated state helps local

communities manage conflicts, but does little to address the underlying sources of
conflict afflicting the border area. A more comprehensive conflict prevention
policy must also address key conflict drivers themselves.

- iv -
• Several of the conflict drivers noted in this analysis may be ripe for external
assistance, including the need for clarification of political, economic, and pastoral
rights within locations in Kenya. Programs which provide the Kenyan
government and legal aid offices with the capacity to reshape local understanding
of the rights of citizenship and to enforce the laws will go a long way toward
eliminating the emerging threat of ethnic cleansing in Kenya's burgeoning new
locations in the northeast.
• Endemic poverty and low levels of education in the border areas are other major
conflict drivers which urgently need attention. Local residents consistently cite
lack of access to education as a major impediment for regional development, and
international aid agency data back them up.
.• Aid interventions which can build upon existing commercial cross-clan networks
and increase the business community's capacity to serve as a line of
communication and promoter of open roads and peace would serve the region
well.

-v-
 SECTION I: Context and History ofthe Kenya-Somalia Border Area

1. Context

1.1 A Troubled Region

The instability which periodically plagues the Kenya-Somalia border area is part
of a broader, complex pattern of state failure and communal violence afflicting much of
the Horn of Africa. Violence and lawlessness are particularly acute in remote border
areas where states in the region have never projected much authority. When they have,
state authorities have sometimes been the catalysts of insecurity rather than promoters of
peace. On the Somali side of the border, the central government collapsed in January
1991 and has yet to be revived.! In Kenya, the vast, remote, and arid frontier areas
bordering Somalia, Ethiopia, Sudan, and Uganda were never entirely brought under the
control of the state in either colonial or post-colonial eras. Thousands of Kenyans have
died in periodic communal violence in these border areas over the past fifteen years, in
clashes which sometimes produce casualties levels normally associated with civil wars.
Kenyan government administration of its peripheral territory ranges from weak to non-
existent. There, government outposts are essentially garrisons; police and military units
are reluctant to patrol towns after dark, and are badly outgunned by local militias. "Even
the police are never safe here," lamented one Kenyan newspaper headline. 2

The Somali-Kenya border thus is doubly troubled - by the chronic failure of the
Kenyan government to establish a meaningful administration in its border areas, and by
the complete and prolonged collapse of the state in Somalia. But the region's general
crisis of instability is by no means unique. Similar patterns of communal violence and
lawlessness occur throughout Somalia and all along Kenya's "arc of crisis" from the
Karamoja cluster through Marsabit to Wajir. That no central government authority exists
on the Somali side of the border is unquestionably a major part of the problem, but
cannot be blamed as the sole source of the crisis. Were that true, Kenya would only be
plagued with insecurity in North East Province, not along its borders with Uganda,
Sudan, and Ethiopia as well.

The broader nature of the crisis of Kenya's border areas was made painfully clear
during the course of the research for and writing of this study. On July 22, 2005, armed
conflict between the Garre and Marehan clans exploded along the Somalia-Kenyan
border over the disputed town of EI Wak, Somalia, the third major armed clash over the
town in six months. The fighting produced 30 deaths and an estimated 17,000 refugees
fleeing into Kenya. 3 Only two weeks earlier, tensions between the Gabra and Borana in
Marsabit, Kenya (near the Ethiopia border) exploded in what some observers claim is the
single worst incident of communal violence in the history of post-colonial Kenya. The

I A Transitional Federal Government was declared for Somalia in October 2004 but has yet to become
operational.
2 "Even the Police Are Never Safe Here," The East African Standard (July 17 2005). Accessed via Lexis-
Nexus.
3 "Kenya: Conflict over Resources in Border Areas," IRIN(August 1 2005), http://www.irinnews.org.

- 1-
massacre at Marsabit resulted in over 90 Gabra deaths and 9,000 displaced persons. It
also produced an outburst of soul-searching in the Kenya media and parliament over
"Kenya's killing fields" in its border regions and the costs incurred for the failure of the
government to extend its authority into what many Kenyans still refer to as the country's
"frontier.,,4 "That hundreds of armed criminals can terrorize a town for hours without the
intervention of the country's security forces is a clear indication that the government has
little or no authority in the North Eastern region," argued the Kenya Human Rights
Commission (KHRC). 5

1.2 LiInited Tools of Explanation

The Marsabit tragedy also highlights the difficulty of accurately identifying the
causes of these armed conflicts, a point of departure for crafting effective conflict
management and prevention initiatives. Dozen of underlying, contributing, and
precipitating factors in the border areas combine to create a witch's brew of tensions
which can easily ignite into violence. Not surprisingly, virtually all of these potential
conflict drivers were invoked in a flood of commentaries in the Kenyan and international
media immediately after the Marsabit massacre. Blame was placed on resource scarcity;
competition over trade; manipulation of ethnic tensions by political elites; ancient tribal
animosities; the warrior culture of pastoral groups; pastoralism itself; the cultural practice
of livestock raiding; influx of small arms and automatic weapons; commercialization of
livestock trade; foreign criminals and guerrilla movements (in this case the Oromo
Liberation Front, or OLF) exploiting unpoliced borders; spillover of conflict and
lawlessness from troubled neighboring countries; failure of government to provide
effective security; failure of the government to heed warning signs of conflict; slow
government response once the crisis exploded; failure of government to drill enough
boreholes; government practice of drilling too many boreholes as a form of political
patronage, resulting in rangeland degradation; administrative boundaries that lump rival
tribes together; proliferation of "locations" (local government units) that become the
source of competition and ethnic cleansing between ethnic groups; misuse of location
boundaries to block pastoralists from previously shared rangeland; lack of
comprehensive livestock and range management policy; poverty and unemployment; low
educational opportunities and levels; vicious cycles of violence created by a culture of
revenge killings; arbitrary colonial boundaries; discrimination against "low country"
communities in the border areas; decline of traditional authority; legacy of decades of
emergency rule and government neglect; lack of understanding of local politics by non-
native government administrators; corruption; rowdy youth; poor remuneration of police
and security forces; inadequate arms and ammunition to Kenyan security forces; a culture
of collective culpability; and tribalism. 6

Most of these claims have some merit in helping to explain both the Marsabit tragedy

4 "Kenya's Killing Fields," The Nation (July 142005). Accessed via Lexis-Nexus.
5 "Thousands Flee as New Clan Attacks Hit Northern Kenya after Village Massacre," Agence France
Presse (July 142005). Accessed via Lexis-Nexus.
6 Each of these causes of conflict were invoked in at least one international wire report or Kenyan
newspaper article or op-ed in the two weeks following the Marsabit massacre.

-2-
and other Kenyan border crises. But detennining which causes are incidental and which
are at the heart of an anned conflict, which are precipitating and which are underlying
causes of conflict; and under what circumstances these causes of conflict are most likely
to ignite violence is not an easy task. The study of internal conflict and conflict
prevention has long been hamstrung by the difficulty of causation, and the Kenyan-
Somali border conflicts are no exception to the rule.
1.3 Geography

The geography, rainfall, and vegetation of the Somali-Kenya border area varies
considerably from the coastal area to the Ethiopian border and plays an important role in
shaping human activities including, as will be argued below, anned conflict. From
coastal south to northern interior, rainfall steadily decreases, temperatures rise, and
vegetation patterns shift from dense bush to semi-arid conditions. Portions of the coastal
districts and regions are quite inaccessible, featuring dozens of inlets, wadis, seasonal
streams, swamps, and dense bush. A large seasonal lake and swamp, Dhesheeq Waamo,
fonns from the Lower Jubba River toward the Kenyan border and, depending on riverine
flood levels, can inundate a long ribbon of low-lying land into the Lorian swamp in
northeast Kenya. The coast of northern Kenya averages 1200mm of rain annually.7
Overland travel is quite difficult in the southernmost part of the border area.

To the north of the coastal area, rainfall levels on the Somali side of the border (in
the Mareerey-Afmadow-Dobley corridor) are high enough (typically 500-750mm/year) to
sustain some of the best grasslands in all of Somalia, as well as scattered rainfed
agriculture. Rangeland is drier and quite open west of the border in Kenya, and in the
northern Somali region of Gedo and the Kenyan districts of Wajir and Mandera. Rainfall
levels drop off quickly in the northern interior of the border area. Wajir district (Kenya)
averages only 200mm of rain per year; Gedo region (Somalia) averages 200-
300mm/year. Temperatures throughout the border area are hot, especially into the
interior; the mean temperature in Luuq Somalia ranges from 28 C (82 F) in the coldest
months (July-August) to 33 C (92 F) in the hottest months (February-March.).8

The Kenya-Somali border area is partially framed by two perennially flowing

rivers, the Tana River in Kenya (to the west) and the Jubba River in southern Somalia (to
the east). The riverine valleys are narrow further upstream but widen as they approach
the Indian Ocean, creating two fertile ribbons of tropical flora and forest. The lower
portions of the river valleys also harbor tsetse flies and so have historically been avoided
by pastoralists.

The region's main rainy season (gu in Somali) occurs between late March to June.
A second short and less reliable rain (deyr) falls in October -November. The deyr season
rains are especially heavy in the highlands of Ethiopia, the headwater of the Jubba River,
so that flooding in the Jubba valley (and filling of large river valley depressions or
dhesheeq) is most common in October-November. As is true of the entire Horn of Africa,

7Ruto Pkalya et a!., "Conflict in Northern Kenya," (Nairobi: Greenwood Press for ITDG, 2003), p. 20.
8Ken Menkhaus, "Gedo Region," UNDOS Studies in Governance #5 (Nairobi: UNDOS, December 1999),
section 2.1.

-3-
rains are highly variable in the Kenya-Somali border area. About one in five years rains
fail and produce serious drought; likewise, serious riverine flooding occurs about once
every five years. Between recurring drought and flooding, the Tana-Jubba inter-riverine
zone is the site of frequent natural disasters requiring humanitarian response in a "non-
permissive environment" of lawlessness and contested local authority. During the rainy
seasons, overland travel on track roads can be difficult to impossible, slowing both
commerce and war.

Human activity has dramatically altered and damaged much of the region's
ground cover, reducing the carrying capacity of the land. Predictably, this has led to an
increase in communal conflict over access to increasingly scarce pasture. Increased
human population, larger livestock herds, and inappropriate placement of boreholes (for
political reasons, mainly in Kenya) have for decades resulted in severe overgrazing in
some areas. 9 Harvesting of acacia trees for commercial export of charcoal or for firewood
has led to heavy erosion and rangeland degradation in the southern half of the border
area, from the Dadaab refugee camp (near Dobley/Liboi) to the coast.

The isolated and often inaccessible terrain of the border area makes for an
appealing location for terrorist and armed criminal activity, though the extent of terrorist
activity in the border area has at times been overstated. Coastal areas of the border feature
numerous small islands and remote inlets where dhows and fishing vessels can freely
come and go undetected; the dense bush and lack of roads in much of the Lower Jubba
region and coastal zone provides safe haven. The radical Somali Islamist group AI-Ittihad
al-Islamiyya (AlAI) has periodically exploited the Lower Jubba region as a safe haven
and transit point, while al Qaeda operatives are known to have operated out of Kenya's
northern coastal zone and cross into Somalia from there. To the north, the remote
settlements of EI Wak and Luuq (Somalia) have in the past been sites controlled by AlAI.

The border itself is in no way based on any natural geographic or socio-economic

boundary; it is essentially a "line drawn in the sand" by way of a colonial-era treaty (the
Treaty of London in 1924) in which the U.K. ceded territory from the Jubba River to the
current border to Italy, as part of an agreement insuring Italian alliance during World War
I. From 1895 to 1924, the border between British Kenya Colony and Italian Somaliland
was the Jubba river.

1.4 Human Settlement and Productive Activities

The border area is relatively lightly populated. Kenya's North-Eastern Province

has a population density of only 5 persons per square kilometer and hosting a total

9En vironmental degradation of the semi-arid pastoral zones of northeastern Kenya and southern Somalia,
and the reduced carrying capacity of the region, have been extensively documented. For a few examples,
see US Agency for International Development, Jubba Environmental and Socio-Economic Studies
(Burlington VT: Associates in Rural Development, 1989; and Robert Walker and Hassan G. Omar.
"Pastoralists Under Pressure: The Politics of Sedentarisation and Marginalisation in Wajir District,
Northeast Kenya," (Nairobi: Oxfam-GB, July 2002).

-4-
population estimated at 600,000 in 1993. 10 In the Somali "Transjubba" regions of Gedo,
Middle Jubba, and Lower Jubba, the total population is unlikely to exceed 600,000 as
well. 11 The vast majority of the population in both Kenyan and Somali border regions are
rural. In Somalia, about half are subsistence farmers, concentrated along the Jubba river.
Pastoralists and agro-pastoralists make up about 300/0 percent of the population in the
Somalia border regions. The remaining 20% percent of the Transjubba Somalia
population are settled in one of several small towns, nearly all along the Jubba river. The
port city of Kismayo is the largest urban area on the Somali side of the border, with a
variable population, usually in the range of 50,000 to 80,000. Other towns with
populations over 10,000 include Bardhere, Luuq, and Beled Hawa (also referred to as
Bulo Hawa), all in Gedo region. Significant urban growth in these Somali regions has
occurred in Beled Hawa and to a lesser extent Bardhere.

In Kenya's North-East Province, the majority of the population is pastoral.

Northeastern Kenya, not long ago almost entirely rural, has experienced a significant
urbanization trend, with several previously small settlements now housing 40,000 or
more inhabitants. Garissa is the largest city in the border area and now a major
commercial hub. Mandera and Wajir now also exceed 40,000 residents. The single
biggest collection of settled households in the entire border area is not in a town, but
rather at the refugee camps at Dadaab, Kenya, where about 100,000 refugees (mainly
from Somalia) have resided for over a decade.

Pastoral production varies regionally. In the rich grasslands of the southern border
area, particularly on the Somali side of the border, cattle herding predominates. Somali
cattle production is now commercialized, with a vibrant cross-border trade into Kenya,
where the cattle fetch good prices in the Nairobi market. Cattle are walked from Somalia
to Garissa, where they are sold. The cross-border cattle trade, which only developed after
the fall of the Barre government in 1991, has been a real benefit to Somali cattle herders
in the Lower and Middle Jubba. To the north, where semi-arid conditions are not
conducive for cattle, camels, sheep, and goats predominate. In contrast to the cattle trade,
which moves from Somalia to Kenya, camels are brought in for sale at major livestock
markets in Mandera and Moyale from Kenya and Ethiopia, destined for the Somali
market in Mogadishu.

Over the past three decades, a growing percentage of the border area population is
engaging in petty commerce, artisan work, construction, transport, and the service
economy (ranging from hotels and restaurants to internet cafes). Livelihoods earned in
urban commerce have increased considerably in the past 10 years in the border region
due to the emergence of a major transit trade system of consumer goods moved from
abroad through Somalia and into Kenya (see chapter 2). Many of the urban households in

10 Mohamed I. Farah, From Ethnic Response to Clan Identity: A Study ofState Penetration Among the
Somali Nomadic Pastoral Society ofNortheastern Kenya, Studia Sociologica Upsaliensia No. 35 (Uppsala:
Graphic Systems AB, 19930, p. 40.
11 No census has been possible in Somalia for decades. The figure of 600,000 is a ball-bark estimate arrived
at by the author in 1999 based on review of available local/district population estimates produced by
international NGOs; see Menkhaus, "Gedo region" and "Middle Jubba Region." The Transjubba
population has declined since the pre-war period, due to displacement by war and instability.

-5-
the region enjoy access to remittances sent from family members working abroad.
Indeed, in some Somali towns in the Transjubba, remittances arguably constitute the most
important source ofincome. 12 Far more Somalis than Kenyans (including Kenyan
Somalis) live and work abroad, so that the remittance economy is more important to
households holding Somali citizenship (including the many refugees and others who live
in Kenya). Very few rural households receive remittances.

1.5 Identity and Ethnicity

Identity politics are central to Kenya-Somalia border area conflicts, and are also
extremely complex and nuanced. What follows is only a brief survey of the topic.

The Kenya-Somali border area is dominated by the Somali ethnic group. But a
number of other ethnic groups live in the border areas, especially in the northern Kenya
border zone. Many of these groups - such as the Garre, Gabra, and Rendille -- possess
highly ambiguous and fluid ethnic identities, making it difficult to categorize them as
"Somali," "Oromo" or other. The Garre, for instance, are considered a Somali clan but
speak a dialect of Oromiyya. The flexible, fluid nature of ethnic identity among the
Garre, Gabra, and Rendille has historically been a useful tool for negotiating relations
between the dominant groups.

The Somalis themselves are much more hybrid in the Tana-Jubba interriverine
area than in central and northern Somalia. In the process of south-western expansion
across the Jubba river and into present day Kenya, Somali clans freely employed the
practice of clan "adoption" (shegad) either as newcomers seeking protection from a
stronger clan or as a means of absorbing weaker groups. As Cassanelli notes, "during the
periodic migrations of Somali nomads from the drier central plains into the interriverine
area, the incidence of contractual clientship multiplied.,,13

The result is that many members of Somali clans in the border areas are shegad -
some are originally Orma, Wardei, while others are adopted members from another
Somali clan. Occasionally, when political advantage dictates, adopted clans can
"rediscover" their original identity and revoke their old clan identity. Ethnic identity in
the region is not nearly as fixed and immutable as observers often assume, but is rather
used as a tool by communities to pursue what they need - protection and access to
resources. As Laitin and Samatar noted two decades ago, "the essence of great politics in
the Somali context is the clever reconstruction of one's clan identity.,,14

12 Nationally, remittances constitute by far the most important source of hard currency in Somalia. An
estimated $500 million to $1 billion flows into the country annually. In the transJubba area, Kismayo,
Bardhere, and to a lesser extent Bulo Hawa are the most remittance-dependent settlements. In Kenya's
border districts, the refugee camps at Dobley arguably receive the most remittances, though Garissa, with
its growing population of resettled Somalis, may now earn as much or more in remittances. These estimates
are based on the observations of remittance company employees interviewed by the author in Bardhere,
Bulo Hawa, Saakow, and Nairobi in 1998.
13 Lee Cassanelli, The Shaping ofSomali Society (Philadelphia: Univ. of Pennsylvania Press, 1982), p. 77.
14 David Laitin and Said Samatar, Somalia: Nation in Search ofa State (Boulder: Westview, 1986), p. 31.

-6-
 What follows is a simplified explanation of clan and ethnic settlement in the
Tana-Jubba region. The Somali Darood clan-family dominates most of the southern tier
of the Jubba-Tana area. The Harti/Darood historically resided in the Kismayo area and its
hinterland, but since 1991 the town has been under the control of outside clan militias
(Marehan/Darood and Haber Gedir Ayr/Hawiye). The rest of the southern interior, from
Garissa on the Tana River to Marerey on the Jubba river, is inhabited mainly by a number
of Absame/Darood clans, including Mohamed Zubeir, Makabal, Aulihan, Talamoge, and
Jidwak. Along the lower Tana and Jubba river valleys, the main ethnic groups are non-
Somali. Bantu farmers reside along the Jubba river, and in the Tana river valley the
Pokomo (Bantu farmers), Orma and Wardey (Cushitic/mainly pastoral) are the principal
inhabitants. A small group of hunter-gatherers, the Boni, live in the northeastern coastal
comer of Kenya. Along the Somali and Kenyan coast, the Bajuni live as fishermen.

Further north in the border areas, ethnic settlement becomes more complex. In
Gedo region, the Marehan are the single largest group, probably constituting half or more
of the total population. Is Other Somali clans in Gedo region include the Rahanweyn,
Bantu, and Ajurann (along the river and east of the Jubba); a small group ofDir near the
Ethiopian border; and the Garre in El Wak district.

On Kenyan side of border, in Wajir district, Somali and proto-Somali clans

include Ajuraan, Degodia, and Garre. In Mandera district, Murille and Garre are the main
proto-Somali clans. To the east, in Marsabit Province, the Borona, Rendille, Gabra, (all
related to the Oromo) and Oromo predominate.

In addition to clan-based identities, these border communities also possess salient

identities based on citizenship, geography, and caste. One important identity marker is
citizenship in either Kenya, Somalia, or Ethiopia. This distinction is important even
within the same Somali sub-clans, and has two dimensions. One has to do with political
culture, the other with political rights. Both are a source of local tensions, but rarely
armed conflict. Culturally, the "reer Somali" (Somali citizens) have been viewed as much
more politically active, aggressive, and clannish than their "reer Kenya" kinsmen. By
contrast, the Kenyan Somalis had lived under emergency rule in Kenya for nearly thirty
years (until 1990) and could fairly be described as politically quiescent at the time. Sub-
clan identity for the Kenyan Somalis was in the past of little importance; many were not
even aware of their sub-clan lineage. Kenyan Somalis attribute their new-found
assertiveness in Kenyan politics to the catalyzing impact of the reer Somalis, but
nonetheless resent what they perceive to be pushiness and lack of respect for rule of law
on the part of the reer Somali. The split over political rights has to do with the fact that
Somali Kenyans are entitled to access to public school and other rights of citizenship
inside Kenya while the reer Somali are not. Reer Somali in border areas have predictably
sought to acquire Kenyan papers so as to access these benefits and facilitate their travel
inside Kenya and abroad. Kenyan clan elders are quite willing to take a bribe and vouch
to local authorities (who may also be willing to accept a bribe) that a reer Somalis is
actually a nomad from the Kenyan side of the border who needs to be registered as a

15 This calculation that the Marehan are roughly half of the total Gedo region population is based on an
estimate made in Menkhaus, "Gedo Region" (1999), section 2.

-7-
Kenyan. Tensions arise when reer Somalis in the border area exploit facilities such as
primary schools which become overcrowded.

Caste or hierarchical identities within clans matter a great deal as well. In the
transJubba regions of Somalia and northern Kenya, social hierarchies are quite complex.
Arguably the lowest status group are the Somali "Bantu" along the Jubba river and now
in Kenyan refugee camps. Somali Bantu are still casually referred to by ethnic Somalis as
addoon, or slaves, are subjected to blatant discrimination, and have been the victims of
land grabbing, forced labor, and predation at the hands of Somali militias since 1990. 16
Even in Kenyan refugee camps at Dadaab and Kakuma, Somali Bantu are subject to
abuse, ranging from appropriation of food rations by ethnic Somalis to rape of their
women. A 2003 study found that 650/0 of Somali Bantu children in Kakuma camp are
chronically malnourished, a rate five times higher than the camp's general population. 17

Within a clan, sub-clans are divided not only by lineage but by status, with some
lineages considered low caste or boon (among the Marehan clan, the Habr Yacoub are an
example). Intermarriage between boon and the "noble" lineages is rare. In other
instances, sub-clans which are not boon but which occupy a weak position in the clan (the
Urmilig sub-clan of Marehan, which are currently embroiled in the war over El Wak, are
an example). Collectively, these and other forms of discrilnination and social hierarchy
matter a great deal, as they impact communal security, economic rights, legal redress of
crime's committed, and access to humanitarian aid in times of crisis.

Yet another important dimension of identity politics in the border area is the
notion of territorial or local citizenship and rights which can be claimed thereby. Somalis
in the trans-Jubba region make a sharp distinction between guri ("indigenous") and galti
("newcomer") residents. This distinction - which occurs within clans and sub-clans -- has
taken on new importance since 1990, when hundreds of thousands of mainly Darood
clansmen fled into the Jubba regions of southern Somalia. Most of these displaced groups
were galti, and they have struggled with the guri kinsmen, sometimes in lethal conflicts,
for 15 years over control of local political and economic interests. The guri-galti tension
is part of a broader debate within Somali society in both Somalia and Kenya over the
nature of political rights and entitlements. One discourse invokes lineage and the
principle of u dhashay, (" born for a region," or jus sanguinis in Western legal terms),
the notion that one may claim full rights in one's clan's home region, and no other.
Territory thus becomes a vehicle for ethnic exclusionary rights and land access. A
second discourse invokes a birthright claim, ku dhashay, ("born in a region," roughly
equivalent to the legal concept of jus solis), or the notion that one may claim full rights
in one's region of birth, even if one is from an "outside" clan. A third Somali discourse is
ku dhaqrnay, which holds that Somalis may naturalize in any region and enjoy full rights
there; no ethnic or birthright claims may be made to restrict rights and land access to any

16 Ken Menkhaus, "From Feast to Famine: Land and the State in Somalia's Lower Jubba Valley," in The
Struggle for Land in Southern Somalia: The War Behind the War, edited by Catherine Besteman and Lee
Cassanelli (Boulder, CO: Westview Press, 1996), pp. 133-154.
17 International Rescue Committee, "IRC Works to Reduce Malnutrition among Somali Bantus in Kenya,"
press statement (New York, 20 June 2003).

-8-
citizen. This principle is accepted as a standard in most modern legal systems, but in
countries where weak ethnic groups have been pushed off their land by stronger groups
(such as parts of southern Somalia and northern Kenya), the principle is viewed by weak
groups as a legal cover for land grabs. 18 In Kenya, a similar controversy over land and
birthri~hts has been framed within the majimboism ("regionalism") policy debates since
1991. 1

In several towns in the Kenya and Somali border areas, a specific, more
cosmopolitan identity exists in which residents view themselves primarily as citizens of
that town. This strong urban-place identity invariably occurs in multi-clan towns, and
appears to be designed to reinforce local solidarity and peace and minimize the
importance of potentially divisive lineage identities. One of the earliest such instances of
civic identity was in the Kismayo area, where Absame and Harti clans embraced a shared
"reer Waamo" identity. More recently, residents of Luuq (Rahanweyn, Marehan, and
other) speak of being "Reer Luuq," and the multi-clan Kenyan town ofWajir appears to
possess a growing sense of "Reer Wajir" as well.

Finally, on the Kenyan side of the border the many ethnic and clan groups
inhabiting and frequently fighting over that territory share a common identity as "low
country" Kenyans, part of the vast expanse of marginalized, generally pastoral or coastal
groups living on the edge of a country dominated by "up-country" Kenyans. The mutual
disdain between these two categories of Kenyans can be fierce at times.

1.6 Key Historical Themes

Several key themes from the border region's history are essential for making
sense of contemporary conflicts.

Isolation. The border region's role as a vibrant trade route is new. Historically, the
territory west of the Jubba to the Tana river was extremely isolated, possessing little of
value to outsiders. Pre-colonial trade routes ran from Somali seaports at Brava and
Mogadishu up into Ethiopia through Luuq, but not across the current border zone. An
important, pre-colonial Islamic communal settlement (jamaaca) was established in the
early 19 th century at Bardhere, along the Jubba river. 20 Italian and British colonial
penetration of the border area was very light, except for attempts to develop irrigated
plantation production along the lower Jubba river. The Italians built an all-weather road
through Luuq and Doolo in the mid-1930s in order to invade and occupy Ethiopia in
1936, and a seaport was constructed at Kismayo. Government presence in the border
territories was extremely limited; British and Italian authorities relied on clan elders to
maintain basic law and order. Punitive expeditions, typically involving the confiscation

18 See International Crisis Group, "Somalia: Continuation of War by Other Means?" (Nairobi/Brussels:
ICG Africa Report #88,21 December 2004), pp. 19-20.
19 Jacqueline Kloop, "Can Moral Ethnicity Trump Political Tribalism? The Struggle for Land and Nation in
Kenya," African Studies, vol,. 61, no. 2 (2002).
20 Lee Cassanelli, The Shaping ofSomali Society (Philadelphia: Univ. of Pennsylvania Press, 1982), pp.
139-43.

-9-
of a large number of a clan's herd as punishment for a crime or insurrection, was the
principle means of enforcing colonial rule. The region had a reputation then as a territory
of little value and much potential trouble, leading to a colonial policy of containment and
neglect.

Migration and conquest. In the early to mid-19 th century, a major southward migration
of Somali clans from the semi-arid zones of central Somalia and eastern Ethiopia
dramatically reshaped population settlement in the Jubba-Tana River zone. Prior to the
1840s, the territory west of the Jubba river was inhabited by Wardey, Orma, Oromo, and
Boroma?1 Somali clans crossed over the Jubba river in the 1800s and quickly pushed
westward, displacing or absorbing existing pastoral groups in a migration that produced
considerable conflict. By the tum of the century, Somalis reached the Tana River and
would have pushed further had British colonial figures not banned Somalis from crossin~
the Tana river and ending what one 1910 colonial report termed "the Darood invasion." 2

Colonial rangeland and conflict management. To reduce clashes over wells and pasture,
the British demarcated specific zones of grazing by clan. They also exercised the right to
open up access to viable rangeland to outside pastoralists in times of drought as a low-
cost form of rangeland management. The result of the fixed colonial rangeland borders
was that clans today view contemporary political and administrative boundaries
("locations") as an extension of the colonial-era exclusionist zones, and invoke those
boundaries to oust other clans from rangeland.

Irredentism and the shifta wars. The Somali nation was divided among five separate
states by colonialism - Ethiopia, Djibouti, British Somaliland, Italian Somalia, and
Kenya. At independence, a central pillar of the Republic of Somalia was irredentism - a
rejection of the colonial boundaries and an insistence on the political unity of all the
Somali people of the Hom of Africa. In Kenya, some members of the Somali population
in the North-East Province mounted a low-level insurgency against the Kenyan
government in the mid-1960s known as the shifta wars, a pejorative term which today is
used to describe any armed banditry. The insurgency failed to attract much direct support
from the new Somali government and was quelled. Worse, it resulted in the imposition of
draconian emergency rule in North-East Province which was only lifted in 1992. Somali
Kenyans felt doubly betrayed - by the Somali government, which talked the talk of
irredentism but failed to back up the shifta fighters, and by the Kenyan government,
which treated administration of the North-East province as a form of military occupation.

Political repression and expropriation. Political repression was the norm on both the
Somali and Kenyan sides of the border from the 1960s through 1990. In Somalia, the
military government of Siyad Barre came to power in a coup in 1969, ushering in a 21
year period of brutal dictatorship and human rights abuses. In the Lower and Middle

21 One exception may be the "proto-Somali" clans such as the Garre, Ajuraan, and others For the history of
Somali migration throughout the Hom of Africa, see E. Turton, "Bantu, Galla, and Somali Migrations in
the Hom of Africa: A Reassessment of the Juba/Tana Area," Journal ofAfrican History 16,5 (1975), pp.
519-37.
22 "The Darood Invasion" (unpublished report, 1910).

- 10-
Jubba valley, the Barre regime expropriated tens of thousands of hectares of land from
mainly Bantu small-holders for large mechanized state farms, and well-placed civil
servants exploited new land tenure and registration laws in the early 1990s to engage in
massive land-grabs at the expense of villagers. A member of the Marehan clan, President
Barre used the authority of the state to advance the clan's power and interests in the
Jubba regions. Superior fire-power and political muscle allowed the Marehan to engage
in expansion in the region. "Marehanization" policy continues to be the source of
conflicts today. More broadly, the entire Somali experience of the state for 21 years under
Barre was not as a source of rule of law and catalyst of development, but rather as a
source of oppression, terror, and expropriation of land, a weapon used by clans in power
at the expense of rivals.

On the Kenyan side of the border, nearly three decades of emergency rule was
equally disastrous, creating an environment of repression and a collective sense of fear of
and alienation from the state. The Kenyan state did not engage in expropriation of local
resources, but allowed no free speech and dealt harshly with dissent. In the infamous
Bagalla massacre of 1984,400 Somalis of the Degodia clan died in a punitive military
operation by Kenyan forces. Any manifestation of Somali nationalist sentiment was
smashed. When emergency rule was lifted in 1992, the North-East Province was devoid
of community organization, and the authority of traditional clan elders as legitimate
leaders in the community (and not just political cronies in the service of the state) was
badly, though not permanently, compromised.

2. Key Changes in the Kenya-Somalia Border Area since 1990

2.1 Changes in Southern Somalia

By far the most important and dramatic change on the Somali side of the border
occurred in January 1991, with the fall of the Barre regime and the subsequent collapse of
the Somali state. The impact on the Kenya-Somalia border area was immediate and
disastrous, especially in the first two years of civil war and famine.

Refugee flows and humanitarian crises. The collapse of the Barre government triggered
a massive exodus of hundreds of thousands of Somalis into the Jubba regions and
northern Kenya. Most were from the Darood clan-family, fleeing from Mogadishu. Their
arrival into the border area overwhelmed the region. Kismayo town briefly swelled in
size from 80,000 to an estimated 800,000 people. Tens of thousands of Somalis poured
into Kenya by foot, ship, and air, seeking safe haven or passage to third countries.

Armed conflicts, famine, and lawlessness. Upon the fall of the government, southern
Somalia fell into heavily-armed chaos. Swarms of uncontrolled gunmen and residents
looted everything of value in government buildings and in Mogadishu's residential
neighbourhoods. Inter-clan violence led to massacres, ethnic cleansing, and a massive
exodus of displaced persons in all directions. Armed battles pitting factions of the Darood
and Hawiye clan-families swept across the countryside. In several instances, fighting

- 11 -
briefly spilled across the Kenyan border. The area between Mogadishu and the Kenyan
border became a "shatter zone" within which residents were exposed to repeated rounds
of looting until they began to starve. The massive famine which occurred from late 1991
through 1992, and which ultimately claimed an estimated 240,000 Somali lives, was thus
almost entirely due to armed conflict and wartime plundering?3

One of the hallmark features of the crisis of 1991-1992 was the rise of an
economy of plunder, in which a wide range of social groups - from illiterate gunmen who
fought to loot, to merchants of war who made millions of dollars exporting scrap metal
from dismantled factories - came to have a vested economic interest in continued
lawlessness and armed conflict. International relief supplies became part of this economy,
as warlords fought to control key ports of entry and transit of the valuable food shipments
brought into the country. Militias charged exorbitant fees to "guard" the food aid, and
were complicit in diversion of relief supplies. By 1992, the food aid had become the
principal commodity over which warlords fought. Emergency relief became part of the
problem rather than part of the solution. In the Jubba regions, several sites -- Kismayo
port, the Kismayo-Jilib highway, Bardhere, Buale, and Be1ed Hawa - were the principal
food relief distribution hubs and attracted the most militia attention. Kismayo in
particular became a chronically contested town.

Another important aspect of the civil war of 1991-1992 was the almost complete
breakdown of authority at all levels. Militias were under only the loosest control of
militia commanders, and fought mainly in order to loot. Clan elders lost control of young
teen-age gunmen. Both clan customary law (xeer) and Islamic law were rendered largely
irrelevant as constraints on lawless behavior. The result was an epidemic of massacres,
rape, and other previously taboo brutalities.

Arms flows. The Somali civil war produced a major weapons flow in the Kenya-Somalia
border area. Both government troops and liberation fronts looted the enormous Cold War
armories of the army, producing a free flow of weapons and ammunition on the street. At
the same time, the fall of the Mengistu government in Ethiopia and the disbanding of the
Ethiopian army in 1991 flooded the regional market with cheap weaponry. Still more
arms found their way into Somalia via the rapidly growing global arms trafficking in the
immediate post Cold War era. Some of these weapons found their way into Kenya, where
they helped to produce destabilization in the border area and gave criminal elements in
Nairobi greater access to cheap semi-automatic weapons. By 1991, the Kenyan police
and military in northern Kenya were outgunned by clan militias and criminal gangs.

"Galti" factor. The Somali civil war produced massive displacement inside the country,
with Somalis fleeing to their clan's "home areas" for protection. In the case of the Jubba
valley, this produced a destabilizing factor. Specifically, Darood clans originally from
Ethiopia or central Somalia and who lived for years in Mogadishu opted to flee
southward to the Jubba valley, where their lineages had home areas. These new arrivals
were on the one hand members of the local clan and thus made claims on rights to live

23 Refugee Policy Group, Hope Restored? Humanitarian Aid in Somalia 1990-1994 (Washington DC:
RPG, November 1994), p. 5.

- 12 -
and secure resources in these areas, but on the other hand were outsiders or guests (galti)
from a distant region. Worse, many of the galti Marehan, Absame, and Harti Somalis
pouring into the Jubba valley were former members of the Barre government or army,
were well-armed and in some cases very well-funded, and were generally more organized
and active politically than their indigenous (or guri) kin in the Jubba valley. The galti
quickly came to dominate the factions representing the Marehan, Harti, and Absame
clans in the valley, marginalizing the interests of the guri. The interests of the guri and
galti were not synonymous - the guri had long-standing and valued relations with
neighboring clans and were stakeholders in local peace, while the galti 's interests were
focused on recapturing Mogadishu. The galti were often dismissive of local customary
law (xeer) and insouciant about the impact of looting on local clan relations. Guri-galti
tensions became an enduring dynamic in clan politics in Transjubba politics, from
Kismayo to Gedo region, and are a major factor in the current Beled Hawa and EI Wak
conflicts.

Rise ofthe Islamist factor. The Jubba valley became a major site of activity for the small
but important Islamist movement which emerged in post-Barre Somalia. Al Ittihad Al
Islamiyya (AlAI) briefly controlled Kismayo seaport in 1991, controlled the town and
district of Luuq (Gedo region) from 1991 to 1996, and vied with the Marehan faction
Somali National Front (SNF) for control of Beled Hawa. Much of AlAI's support came
from guri Marehan who saw the organization as an effective resistance to the galti-
dominated SNF. In Luuq, AlAI imposed sharia law and may have had links with foreign
terrorists from Sudan, but also succeeded in establishing basic security and role of law at
a time when the rest of southern Somalia was in a state of anarchy. AlAI was driven out
of Luuq by Ethiopian forces in 1996 following a bombing and assassination attempt in
Addis Ababa by a local branch of the AlAI there. Since 1996, AlAI has essentially
disbanded, forming a loose network of "alumni" who are integrated into their Somali
communities. Some small cells of radical Islamists possibly linked to AlAI nonetheless
maintained a presence in the Kenya-Somalia border areas, and intermittently placed
training or staging camps along the isolated coastal area at Ras Kamboni and EI Wak. An
American aid worker, Deena Umbarger, was killed by Islamists along the Kenyan border
in the Lower Jubba region in 1999, and rumors persist that radical Islamist commercial
networks operate along sections of the border area at Dobley and elsewhere. Terror
suspects in the attacks on the US Embassy in Nairobi in 1998 and on the Paradise Beach
Hotel in Mombasa in 1999 crossed the Kenya-Somali border and used southern Somalia
as both a transshipment point and safe haven in those attacks. Fears that the unpoliced
border provides foreign or Somali terrorists with easy entrance into Kenya and an easy
escape route remain strong and have been a major preoccupation of Western counter-
terrorism partnership with the Kenyan government. But it is also the case that the
Ethiopian government and local Somali factions hoping to secure advantage against local
rivals often exaggerate the threat of radical Islamism in the region, making accurate
assessment of the threat of terrorism in the Jubba area more difficult. Even more difficult
has been differentiating between legitimate Islamist movements in the region and those
with links to terrorism. 24

24 The topic of radical Islamist movements in Somalia has received considerable attention since the 9/11
attacks. See for instance International Crisis Group, "Somalia: Combating Terrorism in a Failed State,"

- 13 -
Rise oftransit trade into Kenya. The Kenya-Somalia border region, once an isolated area
with little trade, has since the early 1990s been transfonned economically by the rise of a
vibrant and profitable transit trade into Kenya. Somali entrepreneurs exploit the absence
of customs and taxes in Somalia to move a range of consumer goods - sugar, dry
foodstuffs (rations), cloth, basic household items, fuel, cigarettes, and light electronics -
across the Kenyan border into the lucrative Nairobi market. Goods are either smuggled
over the Kenyan border via unpoliced track roads, or bribes are paid to customs officers
at official crossings. Most of this transit trade arrives at beach ports near Mogadishu, but
the all-weather seaport at Kismayo is used to import sugar destined for Kenya. The trade
has helped to create or expand a number of border towns from Dobley to Beled Hawa,
where goods are offloaded into small warehouses and reloaded .onto Kenyan trucks. This
commerce has also helped to produce a network of cross-clan business partnerships with
a vested interest in safe and open roads.

2.2 Changes in Northeastern Kenya

Repeal ofEmergency Law/Withdrawal ofKenyan state. Repeal of emergency rule in

1992 ushered in a new period of government retrenchment from the border area. While
liberation from the harsh emergency laws was welcomed by local populations, the timing
of the retrenchment was disastrous, coinciding as it did with the collapse of the Somali
state and the spillover of anns, violence, and criminality across the border. In truth, even
had the Kenyan government attempted to maintain the control it exercised via emergency
rule it would have been overwhelmed by the tidal wave of refugees, militia, and guns
from Somalia in 1991 and 1992. By late 1991, the Kenyan government had essentially
lost control of hundreds of kilometers of territory in Northeast Province. Even in major
towns like Mandera, Kenyan police and military could not enter certain parts of the town
after dark.

Lawlessness. The first half of the 1990s was a period when, at times and in some
locations, northern Kenya was widely viewed by both locals and international aid
workers as less safe than southern Somalia. Heavily anned clan-based militias and
gangs, sometimes organized by business and political elites, engaged in looting of
livestock and vehicles, terrorizing both Somali and non-Somali communities beyond the
Tana River. At one point the range of Somali bandits engaged in cattle-rustling reached
as far south as northern Tanzania. In an infamous incident in December 1996, a band of
600 Somali militia launched a raid against the Samburu, killing fifty people, stealing
10,000 head of cattle, and shooting down an aircraft carrying the Samburu MP. Land
travel from Nairobi to Dadaab or Mandera could only be conducted with anned military
escorts in convoys. Kenyan police and anny outposts were themselves not immune from
attack.

(Brussels: ICG, May 2002); International Crisis Group, "Counter-Terrorism in Somalia: Losing Hearts and
Minds?" ICG Africa Report no. 95 (Nairobi/Brussels: ICG, 11 July 2005); Matt Bryden, "No Quick Fixes:
Coming to Terms with Terrorism, Islam, and Statelessness in Somalia," Journal of Conflict Studies vol. 23,
no. 2 (Fall 2003); and Andre Le Sage, "Somalia and the War on Terrorism: Political Islamic Movements
and US Counter-Terrorism Efforts," (Cambridge University: PhD. diss, June 2004).

- 14 -
Ethnic clashes. The northern tier of the Kenyan border area - Wajir and Mandera
Districts, as well as adjacent districts such as Moyale, Marsabit, and Isiolo - became the
scene of serious and uncontrolled clashes between rival ethnic groups in the 1990s. This
was part of a broader pattern of ethnic clashes throughout much of rural Kenya over the
course of the 1990s, instigated by political elites and fueled by competition for land,
political representation, and control of local administration (discussed in chapter 4).

Rangeland clashes. Conflicts over pasture and wells in Northern Kenya have been
endemic since independence, when the government lifted old colonial clan boundaries for
rangeland, introducing an era of unclear tenure on land that is formally government
trustland and hence open to universal use, but in practice informally understood to
"belong" to one clan or another. "The lack of clarity over modem land tenure systems
and the breakdown of old ways has led to large clans trying to expand their land by
attacking and terrorizing their weaker neighbors" notes one analyst,25 This confusion has
contributed to misuse of locations as zones of ethnic exclusion, either by weaker
"indigene" clans seeking to protect their land rights from stronger newcomers, or by
dominant clans seeking to institutionalize their claim to land and seal their victory. In
either case, it can and does produce localized ethnic cleansing. The fact that some
Kenyan Somali clans have greatly increased their firepower and numbers thanks to
refugee flows from Somalia since 1991 has exacerbated conflict over rangeland.

Dadaab camp. The territory around Liboi, across from Dobley Somalia in the southern
portion of the border zone, was dramatically transformed in the early 1990s by the
establishment of an enormous refugee camp called Dadaab. Over 100,000 mainly Somali
refugees were encamped there, making Dadaab the largest settlement in the entire
Northeast Province-Jubba Valley area. Dadaab's impact on the local population and
economy was immediate and profound. Refugees' demand for firewood created
environmental degradation in the area; the militias attracted to the area brought horrific
levels of crime, including widespread rape; the food rations and health and education
services the refugees enjoyed for free stoked resentment in local communities, which had
no such access to health and education; and the food rations and remittances flowing into
the camp produced a new regional economy involving trade in foodstuffs and services
between Dadaab and Garissa. Over time, Dadaab has become increasingly integrated into
and integral to the regional economy on both sides of the border. Somali men keep their
families in the camp to access the food and services, while they return to the Jubba
valley; Garissa merchants benefit from the large new market the camp affords them; and
the entire region exploits the availability ofWFP food rations which are sold by refugees
or diverted from them.

Heightened clannism and political activism. One political spillover from Somalia into
northeastern Kenya was a greatly enhanced and politicized sense of clannism among
Kenyan Somalis. The arrival of the "Reer Somali" in large numbers contributed to the
new assertiveness and political mobilization of Somali Kenyans in the years following

25 Abdi Umar, "Resource Utilization, Conflict, and Insecurity in Pastoral Areas of Kenya," Paper presented
at the USAID Seminar "Conflict Resolution in the Hom of Africa," (Nairobi, 27-29 March 1997), p. 17.

- 15 -
the lifting of emergency rule. Additionally, the explosion of new locations in the 1990s,
each earmarked for a specific sub-clan, further mobilized levels of clannism that in the
past were dormant.

Multi-partyism. The advent of competitive elections in 1992 had a major ilnpact on the
Kenyan border region, principally as a driver of conflict. A number of factors -- the
district-based, first-past-the-post electoral system of electing MPs; the growing capacity
of MPs to inject themselves into local politics to the advantage of their constituents; the
increasing tendency to view electoral districts as ethnic or clan home bases; and the
practice by entire sub-clans to vote as a bloc on orders from clan elders - all combined to
create a high-stakes, winner-take-all mentality to parliamentary elections. Communal
violence was and remains closely linked to the electoral cycle in Kenya.

Tensions between Reer Somali and the Reer Kenya. The arrival of hundreds of thousands
of Somali refugees into northeastern Kenya and Nairobi in the early 1990s created
significant tensions between the Reer Somali and Kenyan Somali, even though many
shared the same clan and sub-identity. For Kenyan Somalis, the Reer Somali brought
unwanted levels of violence, a predisposition to engage in illegal activities, and
aggressive demands on aid agencies and local resources. In the process, the Somali
refugees were blamed for stigmatizing all ethnic Somalis, making life much harder for
Kenyan Somalis, who were increasingly seen by the rest of Kenyans - including the
police - as indistinguishable from the Somali refugees.

Urban drift. The period since 1990 has witnessed a significant trend toward urbanization
in northeast Kenya, an area which until recently was almost exclusively rural. Garissa,
Wajir, and Mandera have all grown rapidly due to a rising commercial and service sector,
a growth in the remittance economy, and an increase in destitution among pastoralists.

- 16 -
 SECTION II: Contemporary Conflict Analysis ofthe Border Area

3. Contemporary Trend Analysis of the Border Area

Conflicts in the border area are shaped by and in some instances triggered by
broader trends in the region. The following are among the most significant political and
economic trends in the contemporary Kenya-Somalia border area.

Conflict trends. One of the most remarkable trends in the border area has been the ebb
and flow of armed conflict and insecurity in the past 15 years. After a five year period
from 1991 through 1995 when the region was buffeted by insecurity and lawlessness,
communities in the border region began to enjoy an incremental increase in security and
an uneven but gradual decline in armed conflict. This trend was most dramatic on the
Kenyan side of the border, where, for reasons explored below, livestock raiding, theft of
vehicles, and communal resource conflicts subsided. In the Somali Transjubba regions,
conflict patterns mirrored a trend throughout south-central Somalia - namely, a decline in
sustained and heavy armed clashes, replaced by a chronic localized armed conflict and
insecurity. Part of the insecurity in the Jubba regions has been a function of criminality
and reprisals; much of it is driven by contested control over towns and the resources
towns attract - aid, trade, and taxation of goods in transit. Security in Gedo region was
remained poor since 1996 due to unresolved intra-Marehan clashes over political control
of towns and districts. Middle Jubba has been the site of periodic eruptions of violence in
the three main towns of Saakow (which was burnt to the ground in intra-Rahanweyn
fighting in 1999); Bualle; and Jilib. In the Lower Jubba region, the prized port city of
Kismayo was the site of several brief armed clashes by militias attempting to take or
regain control of the city.

The ebb of armed conflict and general insecurity in Kenya's Northeast Province
has been reversed since 2003, mainly in the northern tier of the border area. Banditry
remains relatively controlled, but in both Mandera and Wajir districts, serious and deadly
armed clashes have erupted. These clan conflicts - explored in more detail below - have
been much larger and more lethal in scale than in the past, have drawn in outside
elements, have produced localized ethnic cleansing in some cases, and have involved
levels of violence which were previously uncommon or unknown in the region. This
reversal of the dramatic progress which had been achieved in the northern two-thirds of
the Kenyan border area since 1995 is one of the most worrisome conflict trends identified
in this analysis and requires careful scrutiny.

On both sides of the border area, communities continue to deal with an endemic,
low level of insecurity which local populations sometimes characterize as "not peace not
war." By this they mean that protection from acts of criminal violence and property theft
remain precarious even in the best of times, and that conditions remain ripe for periodic
eruptions of communal violence even where a peace has held for a number of years. In
some areas, criminal violence by residual combatants produces levels of insecurity akin
to war. There are no zones of "consolidated peace" in the border area, and only a few
areas could generously be described as "post-conflict." Everywhere in the region, the

- 17 -
political claims and historical narratives of clans and ethnic groups are spiked with often
intense and bitter grievances over lost land, wartime atrocities, unfair allocation of
political or economic opportunities, and outsiders.
This "not peace not war" condition is not unique to the Kenya-Somalia border
area, but characterizes much of Kenya's northern frontier area and most of south-central
Somalia. "The frightening fact is that Somalia is officially not even at war," commented
one aid worker recently. "This level of violence is simply a reflection of the brutality of
everyday life for the people living in this country.,,26
Economic trends. Economically, the border area has remained badly impoverished and
underdeveloped over the past fifteen years. Economic gains made in the booming transit
trade have directly benefited only a small percentage of the regional population, mainly
in urban areas. Most of the profits from cross-border trade are made by businesspeople in
Mogadishu and Nairobi. Still, the transit trade has generated significant levels of
employment and small business opportunities for petty traders and others in the transport,
hospitality, and other service sectors in border region towns. The impressive growth of
Garissa, Kenya is due in large part to the new regional commerce. Importantly,
commerce in general and the livestock trade in particular has generally been accorded
safe passage even in periods of tension and clashes. 27

Trends in the pastoral economy vary by region. In the northern interior, pastoralists
herding camels and goats have faced declining terms of trade for their livestock (against
the value of dry foodstuffs), at the same time competition for scarce pasture and water is
growing. Household food security surveys report that a class of impoverished pastoral
households is growing in the region. The Food Security Assessment Unit (FSAU) has
consistently identified Gedo region in the past five years as having among the highest
levels of malnutrition and food insecurity in all of Somalia. Impoverished pastoralists are
a major source of destitute families appearing in the region's numerous towns.

By contrast, the pastoral economy in the southern portion of the border area,
especially on the Somali side of the border, is enjoying better conditions than at any time
in recent memory. This is because the wet grasslands near Afmadow support cattle
herding, and cattle are fetching consistently strong prices in the nearly Kenyan market.

Agricultural households in the border area - mainly concentrated along the Jubba
river valley - have experienced a very negative trendline since 1990, due mainly to
chronic insecurity and predatory banditry. Much of the rich riverine farmland lies
abandoned, and up to 400/0 of the pre-war riverine population is dead, displaced, or in
refugee camps. This group makes up the largest influx of urban migrants into Kismayo.
Overall agricultural production in the Transjubba is, not surprisingly, far below pre-war

26 Quoted in "Somalia's Violence Catastrophic," BBC News (August 22,2005). Accessed at

http://ncws.bbc.co.uk/2ihi/africa/4173230.stm.
27 Paul Goldsmith has noted a similar dynamic in other areas of northern Kenya, where "commerce is
accorded special respect." Goldsmith, "Cattle, Khat, and Guns: Trade, Conflict, and Security on Northern
Kenya's Highland-Lowland Interface." Isiolo, Kenya: APPEAL-KENYA, Conflict and Conflict
Management in the Horn of Africa Case Study, May 1997.
http://paysol1.tulane.edu/col1f1ict/Cs%20St/GOLDSFTN2.html

- 18 -
levels, and almost exclusively subsistence in nature.

The growth of a remittance economy is one of the most important aggregate

trends in the region, fueling and sustaining urbanization in the area.· The remittance
economy is one of the most resilient and reliable sources of revenue in the short to mid-
term, but directly benefits only a fraction of the overall population.

The nature of the transit trade in the border area renders the regional economy
highly reliant on three financial and market centers - Nairobi as a principal market, site
of banks, and source of some investment funds; Mogadishu, as transit trade warehousing
site and home of most of the top businesspeople investing in the trade; and Dubai, the
financial center and point of wholesale purchasing by traders. In a short period of time,
the once remote subsistence economy of the border area has become "globalized" to an
impressive degree.

The proliferation of border towns is entirely a function of the cross-border trade

and is one of many indicators that local communities view the border more as an
opportunity than a barrier or constraint. The multiple routes that have developed reflects
risk management by traders - by maximizing the number of tracks over the border they
reduce the incentive of local militias to engage in extortion, and expand their options in
the event instability renders anyone route too risky. It is also a function of border clans
seeking to possess their "own" trade route from which they can profit by providing
security and imposing modest taxes.

A more recent trend in commerce is the increased use by Somali merchants of

aircraft to fly higher value consumer goods directly from Dubai into Kenya. Should this
trend continue, cross-border commerce will decline in importance.

Overall, the Transjubba region of southern Somalia has rapidly fallen into the
orbit of the Kenyan economy, and today is less linked to the economy of the rest of
Somalia. Only a sustained effort by the Kenyan authorities to crack down on smuggling
from Somalia would reverse this trend. A combination of factors - the national security
imperative to police the border following multiple terrorist attacks in Kenya, western
pressure and support to do the same following the 9/11 attacks, pressure from some
Kenyan merchants to crack down on smuggling, and public pressure to increase its
presence in border areas following the shocking massacre at Marsabit - may lead to more
sustained Kenyan efforts to police the border. A scenario in which smuggling is actually
curtailed is considered unlikely, but it does serve as a reminder that much of the
economic dynamism the border area is exhibiting is, technically, illegal.

Political trends. On the Kenyan side of the border, several important political trends
have manifested themselves. The introduction of competitive elections for Parliament has
had the positive effect of opening up political space for debate in the region, and of
generating legislative representatives seeking to serve the interests of their home
constituencies. The democratic opening in Kenya has also energized local awareness and
interest in public policies, producing more lobbying and interest group action in the

- 19 -
region than ever before. The negative side of this trend is a marked tendency for elected
MPs to engage in corrupt patronage politics and to foment ethnic or clan tensions to
solidify their base for re-election. As discussed below, the particular patronage tactic of
creating new locations to give sub-clans their "own" local seat of governance is a major
contributor to armed clashes and ethnic cleansing in the region.

On the Somali side of the border, the most obvious and enduring political trend
has been ongoing state collapse. Whether the TFG succeeds or fails, a condition of de
facto state collapse will endure for some time to come in the remote border areas of the
Jubba. A more recent political trend worth monitoring is the ascendance of self-declared
regional states, usually but not always formed as clanustans. These have been
successfully used in parts of Somalia as a political base for political figures with national
ambitions, and the trend is spreading. In the Jubba valley, several possible permutations
of regional states could emerge. In the North, a Gedo regional authority may emerge as
an objective from current intra-Marehan talks, or Gedo region may form part of a broader
regional state comprising Bakool region. To the south, the success of the Jubba Valley
Alliance in claiming control over Kismayo and its riverine hinterland may eventually
produce a Lower Jubba-Middle Jubba transregional authority. The key feature of most
regional states in contemporary Somalia is that few actually govern beyond a few of the
largest towns.

4. Conflict Mapping - "First-Generation" Post-1990 Conflicts in Northeast Kenya

Many of the worst instances of armed conflict in the border region occurred in the
1990s and are now either resolved or at least in a state of suspended animation. They
provide invaluable clues to conflict drivers and management for contemporary crises in
the region, and in some cases have contributed to current problems. A brief inventory and
assessment of the most important of these "first generation" post-1990 conflicts is
provided below.

4.1 Wajir Conflicts, 1992-93 and 2000-01

Wajir district is shared by a number of Somali clans, principally the Ajuraan,

Degodia, and Ogaden. The Ajuraan consider themselves to be the "original" inhabitants
of much of the land (though in fact they displaced the Borana, who once inhabited the
entire zone). In any event, the Ajuraan enjoyed protected access to Wajir-West under the
British colonial system and since independence have faced long-term migratory pressure
and chan~ing demographics from westward-expanding neighbors, especially the
Degodia. 8 The district has historically been almost entirely rural and pastoral, with only
four settlements in the entire district in 1940 (a district comprising 56,601 square
kilometers). Today, there are 71 settlements, of which 26 are new since 1996; a total of
about 380,000 live in the district. 29

Goldsmith, "Cattle, Khat, and Guns," p. 30.

28
Robert Walker and Hassan G. Omar, "Pastoralists Under Pressure: The Politics of Sedentarisation and
29
Marginalisation in Wajir District, Northeast Kenya" (Nairobi: Oxfam-GB, July 2002), pp. 9, 18.

- 20-
 Migratory pressures on the Wajir rangeland have been exacerbated by the
firepower and changed clan demographics arising from the Somali civil war, and have led
to endemic tensions between the three clans over rights to pasture and wells. Land
pressure was worsened in the 1980s when the Degodia were pushed out of Isiolo district
and into Wajir by the Borana. Anxiety over land access is clearly a major underlying
factor in district conflicts. But the clashes which erupted in 1992 and 1993 between the
Degodia, Ajuraan, and Ogaden clans were triggered by the arrival of multi-party politics
and competition over MP constituencies. As noted above, these elections were viewed as
high-stakes, zero-sum contests by clans fearful that victory by rival clans would
institutionalize the rival's hold on resources and eventually disenfranchise the losers. In
1992, general elections led to heightened tensions in a number of electoral districts
("constituencies" in the Kenyan system) where two or more clans shared residency and
where demographics were either shifting or were actively manipulated to produce a
desired outcome for a clan and it's MP candidate.

In Wajir-West constituency, tensions between the Degodia and Ajuraan had

already led to an alarming number of assassinations in the 1980s, rendering the area one
of the most unstable in Kenya. In 1992, the demographically ascendant Degodia clan
sought to increase its numbers by bringing in Degodia from outside the constituency to
vote, in some cases even from Ethiopia. It won the seat, and the Ajuraan loss was seen as
a sign of the declining fortunes of the Ajuraan. Degodia chiefs were subsequently
appointed to an exploding number of new locations, thanks to the influence of the MP.
For the Degodia, this was merely a function of time-honored political patronage by an
MP in service to his base. For the Ajuraan, it appeared to be a large-scale, politically-
sanctioned land grab at their expense. The ethnic clashes which ensued rocked much of
Wajir district, spreading to other clans and overwhelming local government. Violence
even spread among the market women in Wajir town. From 1992 to 1995, a total of 500
businesses in Wajir were looted or destroyed; livestock estimated at a value of $900,000
were lost to rustling; and Wajir town was nearly emptied of professionals and middle-
class residents. During that period, 165 civil servants and teachers either left their posts or
refused to go when assigned to Wajir. 3o Wajir was the epicenter of the descent of much
of Northeast Province into anarchy.

What happened next is one of the more extraordinary tum of events in Kenya's
troubled frontier violence, and is well-documented in print and now film. 3 ! An initially
small women's civic group helped set in motion a peace process which eventually
culminated not only in a relatively durable peace among the three main clans in Wajir,
but also helped produce a new type of civic-government partnership for conflict
management that went on to become a model for peace committees throughout much of
Kenya.

Two women intervened to stop the market violence. The Wajir Women for Peace

30 Delma Ibrahim and Janice Jenner, "Wajir Community-based Conflict Management," paper presented to
the USAID Conference "Conflict Resolution in the Greater Hom of Africa (June 1997).
31 A documentary film entitled "The Wajir Story" was produced by Trojan Horse Productions and

commissioned by Responding to Conflict.

- 2\ -
Group was fonned out of those talks, which expanded to include other women in the
town. 32 This women's group was then joined by a group of professionals who fonned the
multi-clan Wajir Peace Group (WPG), with members from all clans in the district. They
facilitated a meeting of clan elders from all the lineages in the district which culminated
in the Al Fatah declaration, which set out guidelines for the return of peace and future
relations between the clans. Other groups also began to fonn, involving elders and youth,
while a group of businessmen began raising money for peace activities.

In April 1994, a new DC was appointed to Wajir, who sought to partner with local
civic groups and traditional authorities to keep the peace. A rapid response team
composed of both government and civic leaders was fonned on the assumption that early
response could prevent many manageable conflicts from spiraling out of control.
Disputes were handled not according to the letter of Kenyan penal code, but "the Somali
way" -- customary law and blood compensation payment was utilized to manage
murders, and collective punishment in the fonn of confiscation of a clan's cattle until a
culprit was apprehended and stolen animals or goods returned. The result was a steady
decline in banditry and crime. While the deeper, underlying conflict drivers were not
addressed, at least one of the main triggers of communal violence - violent crime - was
greatly reduced.

The Wajir experiment in civic-governmental collaboration - or, in some respects,

government sub-contracting out of key functions to local civic and traditional authorities
-- was fonnalized via a decision to unite the peace groups as a sub-committee of the
District Development Committee (DDC), a forum within the district administration
bringing together government and civil society. The Wajir Peace and Development
Committee (WPDC) was also established in 1995. Chaired by the District Commissioner,
it includes representatives from the District Security Committee, heads of government
departments, NGOs, elders, women, youth, religious leaders, the business community,
and the district's four MPs. The committee thus fonns an umbrella of different peace
activities in the district, enshrining civic-local government collaboration while giving
official government blessing to largely autonomous civic and traditional action on
matters nonnally considered core functions of the state - policing, the judiciary (even
over capital offenses like murder, employing extra-constitutional customary law), and
cross-border diplomacy, to name a few. In the process, social groups not nonnally given
voice in fonnal government - elders, women, and youth - were accorded a central place
in the civic-government collaboration. The WPDC also catalyzed traditional clan elders
in the district to fonn a robust ten-man "Council of Elders," allowing them to routinize
communication and collaboration. Over time, the WPDC received support from
international donors.

The implications of the WPDC experiment are considered below. Here, several
points are worth highlighting:

32 This section draws extensively on Ibrahim and Jenner (1997).

- 22 -
 • the WPDC was unquestionably instrumental in the remarkable turnaround of
Wajir district from one of the most anarchic to one of the more stable border
zones of Kenya;
• the WPDC is a model for similar experiments with peace committees throughout
other troubled rural areas of Kenya, a policy shift that is now in the process of
being enshrined in a national policy on conflict management and peace-building;
• crucial to the WPDC's success was the combined commitment of a top local
government administrator and local civic leaders. The absence of either would
likely have doomed the WPDC to failure;
• the WPDC's success is also due to the fact that it combines both traditional elders
and civic leadership (professionals, businesspeople, local NGO figures, etc).
Though the two are often rival sources of non-state authority, the WPDC
demonstrates that they can work together and that when they do, the partnership is
much more effective;
• the WPDC's chief success has been in reducing incidents of violent crime and
banditry, and facilitating rapid, effective conflict management response where
conflicts have emerged;
• the WPDC has not, however, been in a position to prevent large-scale communal
clashes (as the following case studies will demonstrate) nor address the
underlying causes of armed conflict in the region. It is, in other words, more
effective at conflict management and crime prevention than conflict prevention.

As for the conflict which prompted the creation of the WPDC, relations between the
Ajuraan and Degodia were stabilized and no armed clashes have occurred between the
two since 1994. In an effort to resolve the source of the conflict - the MP seat over which
the two clans fought - the Kenyan government created a new parliamentary constituency,
Wajir-North, intended to give the Ajuraan their "own" seat in parliament. This tactic,
embraced at the urging of many Somali Kenyan politicians, has resulted in Wajir district
sporting four constituencies: Wajir-North (Ajuraan); Wajir-East (Degodia); Wajir-South
(Ogaden); and Wajir-West (split between Ajuraan and Degodia). The Ajuraan were split
over the proposal to assign them a constituency; some worried that the this not only failed
to address the underlying cause of the conflict - grazing land and access to resources -
but actually institutionalized the loss of Ajuraan of land to the Degodia, who, as one
Ajuraan figure put it to the Kenyan media, "are aliens to the area.,,33

The 1992-1993 Wajir conflict left a legacy of secondary conflict issues which
were not resolved and eventually produced armed clashes between the Ajuraan and Garre
in 2000. The fact that the Garre clan had quietly sided with the Degodia in the 1992
election and clashes remained a festering grievance within the Ajuraan; the Ajuraan
responded by using their new power in Wajir-North constituency to push the Garre out of
the area, employing the same misuse of administrative (or in this case electoral) units to
engage in ethnic cleansing that they feared would occur at their expense in Wajir-West.

33 Quoted in Umar, "Resource Utilization" (1997), p. 18.

- 23 -
 In July 2000, clashes erupted between the Garre and Ajuraan in northern Wajir
district. The clashes were sparked by a spiraling cycle ofbandit:l raids and counter-raids,
involving as many as 100 armed men and producing 30 deaths. 3 The violence continued
into early 2001, when Ajuraan residents of north Wajir were the victims of a cross-border
raid, reportedly involving gunman dressed in Ethiopian military uniforms (suspected to
be Ethiopian Garre). Fifteen villagers died, 3,300 were displaced, and 15,000 cattle were
stolen and moved back across the Ethiopian border in the attack. 35 Garre and Ajuraan
tensions over grazing land, control of constituencies and locations, and ethnic cleansing
in Wajir-West were sparked in the latter case by a dispute over a newly declared location
along the border ofWajir and Mandera districts. Garre clan leaders and administrators
sought to settle Garre there, while the Ajuraan expected that the location would be theirs
to govern by dint of their control over Wajir-West constituency. Because the Ires Teno
location is the site of valuable grazing land, the stakes were high for the two clans,
helping to spark renewed violence. Adding to the conflict is the fact that the disputed
territory is adjacent to the Ethiopian border. The two clans (especially the Garre) can call
on Ethiopian kinsmen for aid against their rivals introduces outside elements into the
conflict who are not stakeholders in local peace, who are principally motivated by the
opportunity to loot, and who can return across the Ethiopian border to avoid retaliation or
arrest. This places the conflict beyond the ability of both the WPDC and the Kenyan
government to manage. The international or cross-border dimension to the conflict is
complicated still further by the fact that the Garre accuse the Ajuraan of harboring OLF
militia, which the Ajuraan deny, but which is likely a factor in Ethiopian government
tacit support of or acquiescence to Ethiopian Garre irregulars engaging in the cross-
border attacks. 36

4.2 Isiolo District Conflicts and Displacement

The series of clashes and evictions which have occurred in Kenya's troubled
Isiolo District since the 1980s is technically beyond the Kenya-Somali border area, but is
briefly summarized here because of its spillover impact on Wajir and Mandera districts.
Communal clashes have rendered Isiolo district - a faultline area where a number of
major ethnic groups share uneasy and shifting boundaries, where both urban space and
rangeland is contested, and where competition over seats in parliament and locations is
acute - one of the most unstable areas of northern Kenya. Analysts differ over whether
land disputes or political competition is the main driver of conflict in Isiolo, but the fact
that ethnic claims on land and ethnic control of political representation at the
constituency and location levels are so closely intertwined makes the argument somewhat
artificial. The Borona in particular have felt squeezed by long-term west-ward migration
and settlement by Somali pastoralists. The population of Somalis in Isiolo district has
expanded since 1960 from 10 percent to 35 percent in 1989. 37 Not surprisingly, disputes

34 Tervil Okono, "Kenya Clan Fight Leaves 30 Dead," PANA news service ((July 21 2000).
35 "Life in the Shadow of Bandit Attacks," Kenyan Daily Nation (January 13,2001).
36 Ibid.

37 Umar, "Resource Utilization," p. 16. Most of the Somali population in Isiolo in 1960 were Isaaq from

northern Somalia (at the time, the colony of British Somaliland). They were demobilized soldiers who had
served under the British in WW II and who were given land in Isiolo by the British.

- 24 -
over rangeland and wells are endemic. Clashes in the district have at times degenerated
into what one observer termed "tribal terrorism. ,,38

Two evictions have had a particularly powerful ripple effect on the Kenya-Somali
border area. The first was the eviction of the Somali Murille clan from Isiolo by the
Boran in the mid-1990s. The Murille had for decades migrated westward into Isiolo from
Mandera district, and in the better-watered region of Isiolo had turned to cattle herding.
Forced back to the much more arid conditions of Mandera district, the Murille
experienced severe economic hardship. They also arrived as displaced and distressed
pastoralists at about the same time the Garre were also being displaced from Wajir-North
into Mandera. The subsequent clashes between Garre and Murille (explored below) are
partially a function of this spillover of ethnic cleansing in neighboring territories.

Degodia Somalis have also been displaced from Isiolo, in several separate
incidents since the 1980s. In the aftermath of the Bagalla massacre, an attempt was made
by ICRC and the Kenyan government to reduce the need for Degodia herders to migrate
into Isiolo, where their movements often produced conflict with the Borona. Boreholes
intended for Degodia use were drilled in the western border of Wajir as a conflict
prevention tactic, though they were not welcomed by local pastoralists, who believed,
correctly, that they would alter pastoral migration in ways that would damage
rangelands. 39 Tensions with the Borona persisted. In October 1998, a large group of
Borona launched a large-scale raid on several Degodia settlements in Isiolo, killing over
140 and stealing 17,500 cattle. The Kenyan government accused the attackers of being
members of the OLF, a claim which remains a matter of dispute. The area was rocked by
conflict between the two groups for over a year. Efforts by local peace committees and
external aid agencies to negotiate a peace between the two met with only limited
success. 40 The partial displacement of Degodia into Wajir-West has contributed to local
tensions between Degodia and Ajuraan there.

4.3 Aulihan-Abdwaq clashes, Garissa, 1998-2000

In 1998, conflict over land and access to the Tana River erupted between two
Somali clans, the Aulihan and Abdwaq, in Garissa district. The Aulihan moved their herd
into Sankuri division in search of better pasture during a drought, and met with resistance
from Abdwaq, who claim the area as their traditional grazing zone. Adbwaq resistance to
allow the Aulihan access in a period of drought was unusual, especially since the two
clans are closely related. What contributed to this pastoral inhospitality was rising
political tensions between the two clans. A growing number of Aulihan Somali refugees
had secured Kenyan national identity cards (the Adbwaq are not present in significant
numbers in the Transjubba region and so did not generate a sizable refugee flow on their
own). The rapid growth in the number of these galti Aulihan threatened to upset the

38 Goldsmith, "Cattle, Khat, and Guns," p. 30.

39 Walker and Omar, "Pastoralists Under Pressure," p. 28
40 A brief description of peace efforts in Isiolo is provided in Robert Walker, Dekha Ibrahim, and H.O.A.
Shuria, "Oxfam-GB Funded Peacebuilding Initiatives in the Arid Districts of Kenya: Lessons and
Challenges," (Nairobi: Oxfam-GB, March 2003), pp. 14-15.

- 25 -
balance in upcoming general election in 2002, where Abdwaq MPs in two constituencies,
Fafi and Dujis, might lose out to Aulihan candidates. 41 The conflict which ensued
spread to Garissa town and at its worst was responsible for as many as 30 deaths per day
in the town. Business in Garissa was halted and agricultural land briefly abandoned.

Urban commercial interests became a key force for peace. A group of eminent
Garissa leaders from the Aulihan and Abdwaq clans came together in what became the
PastoralistPeace and Development Initiative (PPDI). As a first step, they brought the
business community into the initiative. Not surprisingly, some of the businesspeople were
eager to stop the fighting while others were actively supporting their clan militia. After
numerous setbacks, clan elders from the two sides were convened in a three day
traditional meeting, along with eminent clan elders from other clans in Wajir and
Mandera, who served as mediators and adjudicators. They concluded that the Abdwaq
must pay a sabeen, or initial installment of blood compensation which serves as an
apology and, as an acknowledgement of wrongdoing, tends to cool tempers and open the
door for full negotiations. After additional setbacks, external peace groups from Nairobi
sponsored a five day meeting at which the two clans reached an agreement to end
hostilities. A joint delegation of elders then traveled to different locations to inform their
kinsmen about the peace, which has since held.

The Garissa conflict is instructive as a conflict similar in some respects to the

current crisis in El Wak (discussed below), in which two clans which have traditionally
been relatively cooperative clash over rising tensions that are mainly over political
representation. Pastoral land and water issues appear superficially to be at issue, but are
in fact largely incidental. In Garissa as in El Wak, commercial interests are strong and
interests in peace or conflict mixed, but in the end Garissa's commercial elite opted to
support the peace effort. Strong civic leadership, enshrined in the PPDI, was almost
certainly essential in expediting a peace accord for a conflict which appeared to be
spiraling out of control. Use of traditional conflict resolution tools - guest clan elders as
mediators, blood compensation negotiated by the clan elders - worked well in this
instance, though slow and prone to setbacks. Finally, the Garissa case is yet another
example of the extent to which contestation over parliamentary constituencies is a
conflict-producing exercise, especially where clans vote in bloc and victory is assumed to
assign the winning clan the right to make exclusivist claims on the territory.

5. Conflict Mapping -- Current Conflict Zones along tbe Kenya-Somalia Border

Political, economic, and conflict dynamics interact across the entire border area,
making the demarcation of separate conflict zones a somewhat artificial exercise.
Nonetheless, it is useful to view the border area as four distinct conflicts.

41The following assessment draws on Walker, Ibrahim, and Shuria, "Oxfam GB Funded Peacebuilding
Initiatives," pp. 13-14.

- 26 -
5.1 Mandera

Background. Despite being at the vortex of regional humanitarian crises, refugee flows,
serious armed conflict in neighboring areas, arms flows, endemic banditry, and Ethiopian
skirmishes with both Islamists and the OLF, Mandera district managed to remain
relatively conflict-free throughout the 1990s. Peace there could never be taken for
granted - the district was chronically tense and not always safe from violent crime or
brief bouts of armed conflict spilling over from adjacent Beled Hawa42 - but for the most
part it avoided the explosions of armed conflict which rocked neighboring Wajir and
Moyale districts in Kenya and Gedo region in Somalia. Instead, Mandera town has
served as an important staging base for international humanitarian relief operations into
southern Somalia, and a vibrant frontier commercial hub. The district thus was plagued
by all of the most dangerous "triggers" of conflict noted throughout this paper -
especially acts of crime -- but underlying conditions were not conducive to outbreaks of
communal violence. In retrospect, however, underlying sources of conflict - the misuse
of locations to create zones of ethnic exclusion and block rival clans from accessing
grazing areas -- were steadily building in the late 1980s and 1990s. That these warning
signs were not recognized or acted upon constitutes a failure of conflict prevention.

Since 2004, the district has experienced a serious setback following clashes
pitting the two largest clans in the district, the Garre and Murille. The two clans have had
a long history of periodic struggles over grazing land, dating back to the 1920s, but also
have had extended periods of peace between them. Prior to the 2004 violence, no serious
armed clashes had occurred between the two clans since 1983.

The 2004 clashes began with pastoral clashes over disputed pasture at Jabibar,
resulting in one death. The death triggered a revenge killing, a clear warning sign that
xeer between the two clans was in danger of breaking down. The spiral of revenge killing
culminated with the killing of a prominent Garre NGO worker by Murille near EI Wak in
December 2004. The Garre, responded with armed attacks that ushered in a period of
wider conflict. From January to March 2005, multiple attacks were responsible for 50
deaths and 30,000 displaced persons in the district. The conflict reached it apex on March
16,2005, when a Murille raid on Garre village ofEI Golicha resulted in a massacre of22
people, of whom 16 were children. The massacre triggered widespread outrage in the
Kenyan media and international press, prompting direct Kenyan government response.
The two clans were convened in peace talks which were eventually mediated by a group
of eminent Muslim leaders from the Supreme Council of Kenya Muslims (SUPKEM).
Under considerable national pressure, the two clans reached an accord which enshrined
open access to pasture throughout the district. But aid agencies on the ground report that
ethnic cleansing at the location level continues, and IDP return to home areas, including
Mandera town, has been slow. The current lack of hostilities may not constitute an
enduring peace. While a return to armed conflict is not viewed as inevitable, it remains a
dangerous possibility.

42 Mandera town has occasionally been showered with stray bullets and even errant mortars from fighting
in Bulo Hawa.

- 27 -
Conflict Drivers. A number of underlying conflict drivers have been at play in Mandera.
Chronic tensions over access to grazing land and wells have unquestionably been a root
cause of the communal violence. Land pressures in the arid region have always been
high. A combination of factors - increased population, the displacement of Garre and
Murille from Wajir-West and Isiolo back into Mandera, an increase in poorly placed
boreholes, and the misuse of locations as zones of exclusive grazing land have all
contributed to growing anxiety among and pressure on pastoralists.

Political competition over constituencies and locations have been a more direct
conflict driver. Following clan clashes in the early 1980s, in 1988 a new constituency,
Mandera-Central, was carved out of Mandera-East to provide a seat in parliament for the
Murille. Prior to that time, the two constituencies in Mandera were routinely held by the
numerically dominant Garre. As in Wajir, it was thought that a separate constituency in a
mainly Murille zone would serve as a conflict prevention device, by eliminating political
competition between the Garre and Murille. Instead, it accelerated it. As elsewhere in the
region, MPs wielded authority to pressure the government to expand the number of
locations in their constituency as a means of rewarding clients and expanding government
services - schools, boreholes, stipends to chiefs - all part of political patronage in Kenya.
But because the "base" of each MP is his clan or sub-clan, rewards of location
chieftainships went exclusively to the clan of the MP. And because control of locations
was used to make exclusionist claims on land within the location borders, the net result
was widespread ethnic cleansing in the three constituencies of Mandera. 43 Conflict was
most acute in locations where valuable, previously shared grazing area was situated.

In the Garre-Murille clashes, competition within the Garre political elite appears
to have played an especially destructive and complex role. In 2002, both the Mandera-
East and Mandera-Central constituencies were won by Garre from the Qoranyo sub-clan,
one of two main Garre sub-clans the other being the Tuuf). The Tuuf had previously
enjoyed prominence in the political realm, and both the sub-clan and its ousted MP, Aden
Nur Mohamed, were unhappy with the outcome. Some local observers contend that
Aden Nur and his Tuuf supporters sought an alliance with Murille and, in an attempt to
demonstrate that the Qoranyo leaders could not rule, provoked security incidents and
tensions between the two clans. Adan Nur's successor, MP Billow Kero, filed a statement
with the CID accusing Aden Nur of inciting violence, and Aden Nur was summoned and
questioned by the police. Nur in turn accuses Kero of using Garre militia to intimidate
rivals. If either or both of the charges are true, it would mirror patterns in a number of
other troubled border regions of Kenya, where MPs and their political rivals are
frequently accused of fomenting ethnic violence. 44

43 IlUG, "30 Killed as Clashes Engulf Mandera," (January 2005). Accessed at:
http://www/itdg.org/id+peace6 mandera
44 "Government Admits its Inability to Find Cause of Feud in NEP," Kenya Times (Jan. 152005). MPs
from Marsabit were brought in for questioning by the Kenyan police under similar suspicion of incitement
following the massacre there in July 2005.

- 28 -
 As with the Garre-Ajuraan clashes in Wajir-North, the fact that Mandera district
clans can call on militia firepower from neighboring Ethiopia, where kinsmen are either
in the Ethiopia military or operate as para-military forces in the border areas, exacerbates
the conflict.

The Garre-Murille conflict was also entangled in and driven by the Garre-
Marehan conflict. The Murille of Mandera district have had a long-running relationship
with the Marehan sub-clans in the border area, especially the Ali Dheere sub-clan. A
series of killings since 2000, and a longer history of rivalry over trade between the Ali
Dheere and the Garre, led to deteriorating relations between the two groups. When anned
clashes between the Marehan and Garre broke out over El Wak, the Garre suspected
Murille complicity with the Marehan, increasing mistrust between the two.

Prospects for Peace. The fact that the recently-brokered peace accord was reached under
considerable external pressure and without adequate follow-up to insure implementation
is worrisome. Most of the underlying factors driving the conflict - political manipulation
of ethnic grievances by politicians, abuse of locations to pursue ethnic cleansing, and
ever-worsening pressures on pastoral households - are still in place. On the other hand, a
number of factors could work in favor of consolidating the fragile peace. The impressive
growth of civic peace groups in Mandera is helping build lines of communication and
their watch-dog role may reduce the space political figures have to manipulate clan
tensions. Business interests in Mandera-town depend on cross-border trade, which has
been interrupted by fighting, and could be convinced to support peace rather than fund
their clan militias. Finally, the fact that the peace accord was mediated by national
Muslim leaders may create a stronger taboo against violating the peace.

5.2 EI Wak

Background. The district of El Wak, Somalia has been the scene of several anned
clashes since December 2004. The fighting has pitted two Somali clans against one
another, the Marehan and Garre clans. As of August 2005, the conflict has produced an
estimated 93 deaths and over 10,000 displaced persons, most of whom have found shelter
with kinsmen in nearby settlements and towns. The first round of fighting in April 2005
produced 20 deaths. In the second clash, the Garre militia retook the town of El Wak
(called Bur Hache by the Garre) on June 12, leaving 43 dead and thousands displaced. A
third round of fighting which erupted on July 22 produced 30 deaths and led to the
Marehan retaking control of the town. Civilians, including women and children, have
been among the victims of the fighting.

El Wak is a remote border district of Gedo region in southwestern Somalia.

Pastoral nomadism is the dominant mode of production for the vast majority of the
population. Both the Somali district ofEl Wak and the adjacent areas in Kenya feature
very high levels of poverty and low levels of development. Chronic instability due to
periodic intra-Marehan clashes in Gedo region since 1996 has exacerbated poverty,
malnutrition, and displacement there.

- 29 -
 El Wak town is the main settlement in El Wak district. Though the current
conflict is driven by a number of factors, control of the town is the main immediate
objective of the two sides. The town's population has grown in recent years due to cross-
border trade and insecurity in other zones of Gedo region; as of 200 1 it was home to an
estimated 2,200 people.

The district of El Wak is an area where rangeland controlled by the Garre and
Marehan overlap. The Garre inhabit territory encompassing parts of El Wak district in
Somalia, Wajir district in Kenya, and the border area of southern Ethiopia. The specific
sub-clan of the Marehan which has historically resided in El Wak, the Urmidig, is a
minority lineage within the broader Marehan clan,

El Wak district has long been co-habited peacefully by the Garre and
Urmidig/Marehan. The two clans have not had notable disputes over grazing areas, and
they have long shared the town ofEl Wak. 45 Indeed, local Urmidig/Marehan and Garre
clans in El Wak have in the past made special efforts to preserve good relations, splitting
positions in the district council and police 50-50, and dropping the blood payment in the
events of a cross-clan killing from 100 to 40 camels. Intermarriage between the two clans
is common as well. These solid relations between the Garre and Urmidig/Marehan
contrast sharply with the conflict-ridden relations the Garre and Marehan have had with
other neighboring communities.

Structural, Environmental, and Political Factors. Beginning in the 1970s, a number of

factors began to place pressure on the Garre-Urmidig peace in El Wak. The first was
(and remains) pressure on the land. Though the Garre and Marehan clans have not fought
over pasture and wells in El Wak, their relations must be viewed against the broader
backdrop of increased pressure on and clashes over land in the region, which has
produced militarized relations between the so-called "comer tribes" of northern Kenya.

A second source of pressure was the rise of the Marehan clan to national political
power in Somalia under the reign of President Siyad Barre (himself a Marehan) from
1969 to 1990. Though Gedo region did not enjoy many direct perks from the Barre regime
--- it remained a poor and underdeveloped region - many Marehan assumed top positions
in Mogadishu and formed part of a powerful political, economic, and military elite. In
Gedo region, this enabled the Marehan to push southward and gradually gain control over
vital towns and rangeland at the expense of neighboring clans such as the Aulihan and
Rahanweyn. The city of Bardhere, an old religious settlement and the largest town in
Gedo region, was in earlier times not considered a "Marehan" town, but by the 1980s the
Marehan had become the dominant clan there. The previously Rahanweyn district of
Luuq was redistricted into Gedo region and increasingly settled by Marehan. Likewise,
the important wells at Fafaduun were lost by the Aulihan to well-armed Marehan
pastoralists. Marehan also gained political control over the valuable port city of Kismayo
in Lower Jubba (Barre appointed only Marehan governors there, and Marehan merchants
were allowed to monopolize livestock export trade at the port) and began populating the
city with newcomer Marehan. Neighboring clans complain that this expansion of

45 As is common in "shared" Somali towns, each clan is clustered on its side of £1 Wak.

- 30-
Marehan territory and power reflects a strategy of "Marehanization" of the entire Gedo
region; some go so far as to argue the Marehan are seeking to dominate all of the
Transjubba region down to Kismayo. To the extent that the "Marehanization" thesis is a
factor, the enduring peace in EI Wak between the Garre and Urmidig/Marehan is
especially impressive, and reflects the fact that until recently EI Wak was considered low
value and that the local Urmidig sub-clan was not party to the expansionist agenda of
other Marehan sub-clans. But the rising power of the Marehan meant that the balance of
power in EI Wak could be easily tipped if and when more powerful Marehan sub-clans
opted to weigh in on local affairs.

A third factor was the collapse of the Somali state and the fall of the Marehan
from power in 1991, which had an enormous impact on Gedo region. As discussed
above, tens of thousands of Marehan fled Mogadishu and arrived in Gedo region. The
displaced Marehan from Mogadishu overwhelmed the local population. They were well
armed, generally wealthier, urban, and better organized politically. Though most of the
top political and economic elite of the Marehan relocated to Nairobi, the newcomers or
galti took over political control of Gedo region under the banner of the Somali National
Front, or SNF. As the Somali crisis dragged on, tensions between the local Marehan (the
guri, or original inhabitants) and the galti Marehan increased. The guri complained that
the galti monopolized political power and economic opportunities, were not stakeholders
in local peace, pursued agendas that served their interests only, and looked down upon
the guri Marehan as weak and incapable.

In northern Gedo region, the growing rivalry between the guri and galti Marehan
- a rivalry which can only partially be explained along sub-clan lines - manifested itself
as a factional struggle between the galti-dominated SNF and the Islamist movement Al- .
Ittihad Al-Islamiyya (AlAI). For years, the community ofEI Wak managed to convince
both of the "factions" of AlAI and SNF not to involve themselves in the town. In an
interview with elders from EI Wak in 1998, they referred to EI Wak as a "faction-free
zone" and were proud of their ability to keep the small multi-clan town out of the
Marehan political fray.46

Sometime in the late 1990s, however, Islamists did move into EI Wak and
establish a presence there. The precise nature of this presence remains the subject of
dispute, but EI Wak has gained a reputation as a safe haven for a small number of
Islamists, some radical in orientation, who were rumored to have used the town to build a
camp and terrorist training base for infiltration into Kenya. Others argued the Islamists
were not so much linked to terrorist training bases but were rather engaged in the
expanding commerce passing through the town. What is clear is that the Islamist presence
in EI Wak at some point included a militia component of some consequence, and that,
unlike in Luuq, where AlAI worked openly with international aid agencies, the Islamists
in EI Wak were hostile to outside visitors. For the community in EI Wak, the presence of
the Islamists was a major problem, earning it an unwanted and probably somewhat
exaggerated reputation as a terrorist lair. This was especially dangerous in the aftermath

46 Ken Menkhaus, "Gedo Region," (Nairobi: UN Development Office for Somalia, December 1999).

- 31 -
of the al Qaeda bombing of the US Embassy in Nairobi in August 1998. The locals claim
that they managed to convince AlAI to depart by 2000.

The rise of cross-border commerce between Somalia and Kenya beginning in

1993 is a fourth factor shaping the conflict in El Wak. This transit trade, which expanded
into a highly profitable and high-volume business involving thousands of Somali and
Kenyan wholesalers, middlemen, small traders, and transporters, initially passed mainly
through Beled Hawa/Mandera. But intra-clan clashes and insecurity among the Marehan
in northern Gedo region rendered that route increasingly unattractive after 1999, pushing
the interstate commerce to other, previously minor trade routes such as the Bardhere-El
Wak route, which is over a track road. El Wak became one of the busiest of the half
dozen or more trade towns which emerged along the Kenyan-Somali border, replete with
small warehouses for storage of non-perishable goods. This increased the value of the
previously uninteresting town and attracted the galti Marehan from northern Gedo region.
By 2005 local residents estimate that about a quarter or more of the Marehan population
of El Wak were newcomers, mainly from sub-clans other than the Urmidig. The galti
Marehan were not stakeholders in the local peace between the Garre and Urmidig and
eventually took actions which helped to trigger the fighting. In addition, the town also
attracted a number of businessmen from Mogadishu, mainly from the Haber Gedir and
Murosade clans. 47 They established partnerships with Marehan and Garre but have also
been accused of stoking rivalries between the Garre and Marehan for their own benefit.
Even without external complications, Garre-Marehan competition for control over the
lucrative trade and over taxes collected on commerce began to erode previously peaceful
relations between the two clans. In sum, cross-border trade increased the value of El
Wak town and in so doing increased possibilities of conflict over its resources.

Fifth, political developments in Somalia, Kenya, and Ethiopia each contributed to

the spiraling of the Garre-Marehan dispute into an unmanageable armed conflict. In
Ethiopia, the Garre have been used by the Ethiopian military to help control their border
with Kenya. The militia leader controlling this Garre unit of some 300 men is Hassan
Koro, and is viewed by neighboring groups as a warlord. Ethiopia, whose principal
concern is thwarting operations by the Oromo Liberation Front in the area, capitalizes on
Garre-Oromo animosities by strengthening and arming Garre militias in its border area.
The result is that the Garre, which previously were not an especially powerful "comer
tribe," now can draw on a powerful militia across the Ethiopian border. Some
neighboring groups complain that - like the Marehan - this has emboldened the Garre to
become more assertive or even expansionist. Marehan insist that the Kenyan government
is complicit with the Garre, pointing to the fact that Koro's Ethiopian Garre militia passes
through and remains in Kenyan territory and no attempt has been made by Kenya to
block them.

In Somalia, the 2003-05 negotiations over power-sharing in the Transitional

Federal Government (TFG) heightened the need for clans to maximize their perceived
power and control of territory. For the Marehan, the loss ofEI Wak to the previously

47 Both Haber Gedir and Murosade are members of the larger Hawiye clan family which today dominates
Mogadishu politically and economically.

- 32 -
weak Garre was an ill-timed humiliation that hurt them nationally and that could not go
unchallenged. For the Garre, loss of control over territory inside Gedo region would
essentially push them out of Somali politics. Aside from a community of Garre in Lower
Shabelle region, El Wak is the main Garre territory inside Somalia. Loss of that territory
to the Marehan would impose enonnous political costs on the Garre of Somalia.
Politically, then, the stakes are extremely high for both Marehan and Garre over El Wak.

Collectively, these underlying factors combined to create a much more

militarized, contentious, and dangerous environment in which to maintain the peace in El
Wak. In many ways what is remarkable about El Wak is how successful the local
community had been in maintaining the peace for 14 years in a context of state collapse,
militarization, resource scarcity, and political manipulation of clannism.

Triggering Events. The Garre-Marehan slide into anned conflict in December 2004 was
in part triggered by spillover from recent outbreaks of anned conflict between the Garre
and Murille clan in Mandera district. It was also triggered by a series of murders which
were not speedily and satisfactorily resolved by clan elders. Growing levels of suspicion
and distrust between the two clans shaped local perceptions of the murders as political
acts, making it much more difficult to resolve the deaths through customary blood
compensation.

Because the Murille are allied with the Ali Dheere sub-clan of the Marehan
(concentrated in the Beled Hawa area), the Garre increasingly suspected the Marehan of
providing support to the Murille as part of a proxy war against them. This suspicion was
increased in October 2003 when a Kenyan Garre employee of the international NGO
ADRA was shot and killed near El Wak by Marehan gunmen from the Ali
DheerelMarehan sub-clan. Though Marehan claim that the killing was simply over a
business dispute, the Garre viewed this as part of the Garre-Murille conflict and believed
the Ali Dheere were complicit with the Murille. Though blood compensation payments
(diya) were paid, more killings ensued, and alanned observers warned that anned conflict
between the Garre and Marehan was imminent. These warnings went largely unheeded.

The stakes were raised further when Garre called on the Garre militia in Ethiopia
for help. In 2004, an Ali Dheere man was killed in El Wak, and the Ali Dheere retaliated
by killing two Garre businessmen in Bulo Hawa. The Garre, emboldened by the addition
of their Ethiopian Garre militia, escalated the crisis by taking El Wak by force in
December 2004. Anned clashes were limited, as the Marehan opted to retreat, but it was
at this point that the local Garre-Unnidig partnership was overwhelmed by broader clan
tensions, heightened by the fact that outside Garre and Marehan interests were now
increasingly driving decisions.

Unresolved tensions between Marehan and Garre erupted in April 2005 when the
Marehan took El Wak in fighting which left 20 dead and over 7,000 displaced. A militia
build-up ensued on both sides, and on June 12 2005 a well-anned Garre militia retook El
Wak in fighting that led to 43 dead and thousands of displaced Marehan.

- 33 -
 Some local Garre leaders at that point sought to initiate peace talks with their
Marehan counterparts, but the Marehan were unwilling to negotiate from a position of
weakness. Instead, the Marehan met at Garbaharrey, Gedo region, ostensibly to help
resolve an intra-Marehan clash over control of a small Gedo settlement, but also with the
aim of closing ranks in order to present a united front against the Garre. The fact that
the Garbaharrey conference simultaneously served as a peace conference and an
opportunity for a war party to forge an alliance serves as a reminder that local
reconciliation in Somalia is often a two-edged sword, a form of alliance building at the
expense of a third party.

The Marehan at that point fell into one of three categories. First were those
voicing a preference for a negotiated solution to EI Wak. This included some - but not all
- of the Urdimig clan on EI Wak. It also included some of the broader Marehan clan
leadership in Gedo region, which wanted to reclaim Marehan co-habitation of EI Wak but
were wary of the political costs of yet another armed conflict in the region involving the
Marehan, especially at a sensitive time in the formation of the Somalia TFG. This group
was not averse to a militia build-up in Gedo region, but saw that as a means of
negotiating from a position of strength. A second group of Marehan were Gedo residents
who insisted that the clan must retake EI Wak by force to save face and only then could
they negotiate a return to co-existence with the Garre. A third group, including Marehan
elites in Nairobi and in the Jubba Valley Alliance, viewed the entire crisis through the
lens of national rather than local interests. For them, the defeat at EI Wak at the hands of
the Garre was an embarrassment and a setback to broader Marehan aspirations for power
at the national level, in the TFG or in a post-TFG government. The loss had to be
reversed decisively to demonstrate the strength of the Marehan generally, and to
consolidate the long-running Marehan goal of rendering Gedo region into a Marehan
regional base, possibly with the longer-term expectation of declaring an autonomous
regional administration there. The latter group's interests won out.

Following a month of inflammatory rhetoric from both sides, in which Garre

accused the Marehan of EI Wak of being AlAI terrorists, while Marehan claimed they
were fighting "non-Somali" Ethiopian militias invading Somali territory, the Marehan
retook EI Wak on July 22. Leading the attack, and possibly launching it without the full
agreement of Marehan clan leaders in Gedo, was a unit of Marehan militiamen from the
JVA in Kismayo. The attack came on two fronts, including one which crossed into
Kenya, a move which apparently took the Garre by surprise and was intended to prevent
their retreat across the border. JVA commander Barre Hirale subsequently expressed a
willingness to initiate peace talks with the Garre, while disputes arose among the
Marehan militia over the distribution of looted camels. A weak Garre counterattack took
place two weeks later and was repulsed with no casualties, a possible indication that the
move was actually intended to test Marehan responses prior to the launching of yet
another attack on the town.

Prospects for Peace. As this paper was in final stages of completion, a local truce was
reported in EI Wak, with the Marehan militia withdrawing from the town as a gesture of
good faith. Kenyan media reported that some of the displaced families were beginning to

- 34-
move back to the town. The Garre are divided over whether to proceed with peace talks
from a position of defeat, but the Marehan gesture is a hopeful sign that progress can be
made toward a comprehensive settlement of the conflict. The fact that this truce was the
result of a meeting facilitated by the Mandera District Peace Committee (MDPC) and
Beled Hawa NGO consortium is a good sign that those nascent umbrella groups are
developing a stronger capacity. One major concern is the fact that the talks involved did
not include Garre militia and political leaders; instead, local Garre elders met with
Marehan elders and militia leaders. 48 As a result, the Garre militia does not consider itself
a party to the accord and could opt to disrupt the truce and the return of residents to the
town.

The fighting over EI Wak has badly damaged the interests of both the Garre and
Marehan. Trade in the EI Wak area has been completely halted, urban residents have
been displaced, and pastoralist grazing patterns have been disrupted and some livestock
looted. The town itself is generally of little value without peace and commerce, so victory
is in a real sense hollow. This should in theory be contributing to a "hurting stalemate"
which could bring both sides to the bargaining table.

The problem is that the "hurting" from the conflict is only felt by local Marehan
and Garre residents, and they are not driving decisions in the crisis. The outside interests
- Ethiopian Garre, galti Marehan, the JVA, and the Garre and Marehan leadership in
Nairobi which are so often accused of using "remote control" to foment these conflicts
are unaffected by the fighting, though it is not clear that anyone is exploiting the conflict
for significant political and economic gain.

Calls for a local solution to a local conflict are not feasible - the conflict is now
entangled in wider interests, and those interests must be addressed and brought into peace
talks. The biggest immediate obstacle is that outsider Garre and Marehan political elites
view the conflict in strictly zero-sum terms, wanting only to negotiate from a position of
strength in order to increase their political capital nationally. One solution may be to
request both sides to withdraw from the disputed town until peace talks are completed.
Local elders from the two clans have stressed that the conflict requires third party
mediation, probably by a trusted set of neutral clan elders from outside the region.
External actors may be able to facilitate peace by providing financial and/or logistical
support to whatever talks emerge.

For their part, local elders are going to have to commit to much speedier and more
effective blood payment for crimes committed, and much greater penalties on kinsmen
who resort to revenge killings. The galti Marehan and Garre whose commitment to local
xeer is weak must be pressed to respect the xeer of their local kinsmen.

The fact that the Garre-Marehan conflict is closely intertwined with the Garre-
Murille conflict in Mandera suggests the possibility that peace talks may need to embrace
a wider scope and set of regional actors than merely the Garre and Marehan. Unless the
broader set of conflict issues which helped to provoke the EI Wak crisis are addressed-

48 Beled Hawa NGO Consortium, "Situation Report, £1 Wak Somalia" (August 2005).

- 35 -
especially the threat of proxy war through third parties - El Wak is likely to remain
unstable for some time to come.

The most hopeful aspect of the El Wak conflict is that it has not been propelled by
some of the intractable structural factors at play in some of the region's other conflicts.
This is not a conflict driven by land and water scarcity, or sharply opposed claims on the
town (both sides agree El Wak is a shared town). It is over a fairly narrow set of issues
related to the prompt management of inter-clan crimes and shared opportunity to profit
from cross-border trade. Both are potentially positive-sum issues. A return to status quo
ante - shared access to and control of rangeland and the town of El Wak, and a return of
control to the local community in El Wak - is the ideal outcome for peace talks. External
militias in both clans will need to be thanked for their support by their kinsmen and
requested to return home.

The principal danger in El Wak is if the conflict is not resolved soon, it will
almost certainly take on a life of its own, with the casualties and losses sustained in the
past six months of fighting becoming the core grievances fueling future fighting. Too
many of the long-running zones of instability in the region have in fact been left to bum
too long and generate too many losses, making it exponentially more difficult for
communities to make peace.

5.3 Beled Hawa

Background. Beled Hawa (also called Bulo Hawa) is a town of about 30,000 people
immediately across the border from Mandera, Kenya. Prior to Somalia's civil war, it was
a small, remote, and relatively unimportant border town, handling a small amount of
imports of Kenyan light manufactured goods traded into Somalia. During the civil war of
1991-92, Beled Hawa swelled with IDPs seeking to cross into Kenya, and then developed
into an important commercial town commanding control of the flow of transit trade from
Mogadishu into Kenya via Mandera. The taxes collected on commerce moving through
the town, the opportunities afforded by the service economy, and proximity to Mandera
(which provided health and education services and access to an airport) greatly increased
the value of Beled Hawa. A power struggle over control of the town developed between
two wings of the Marehan clan. One, the Somali National Front, was composed mainly of
former government officials from Mogadishu; the other, Al-Ittihad Al-Islamiyya (AlAI)
was an admixture of Marehan embracing a political Islamist ideology and local (guri)
Marehan opposing SNF domination of Gedo region. Throughout much of the 1990s, the
rivalry for control over Beled Hawa was tense but contained. For much of the mid-1990s,
the town sported two parallel administrations. An attempt was made by top Marehan clan
elders to broker a peace between the two factions in order to unite the Marehan. This
produced the e1-Ade accords, announced in August 1998, in which it was agreed to
disband AlAI, canton weapons, integrate the two wings into the SNF, and then convene a
delayed congress of the SNF to select or reselect leaders. But hardliners within the SNF,
fearful of a Congress which might result in their removal from power, sabotaged the
accord. Several notable guri leaders in both the SNF and AlAI, including the SNF
sponsor of the peace talks, Ali Nur, were assassinated over the course of the next year.

- 36 -
 Since 2001, armed clashes in Beled Hawa have erupted on several occasions. In
April and May 2002, fighting between the pro-TNG Marehan wing led by Abdirisak Bihi
and the pro-SRRC faction of the Marehan led to Ethiopian intervention and resulted in
the arrest of Bihi and a flood of 10,000 refugees into Mandera Kenya. Poor treatment of
the refugees in Mandera exacerbated tensions and led to numerous deaths there. In
northern Gedo region, commerce was halted, roads mined, local security deteriorated,
and food security remained poor. Gedo region was in 2002 one of the worst
humanitarian emergencies in Somalia due to this outbreak of fighting and the
displacement it caused. Fighting erupted again between May and July 2004, leading to
an estimated 50 deaths and displacing another 3,000 into IDP camps along the Kenya-
Somalia border.

Conflict Drivers. Since 1999, the armed clashes in Beled Hawa have mainly been driven
by factional divisions within the Marehan clan. These factional splits are extremely
complex and are a reflection of the exceptionally fissurable and complex nature of intra-
Marehan politics. Marehan political alliances are informed by multiple faultlines,
including:
• Sub-clan divisions
• Individual leadership rivalries
• Indigenous (guri) Marehan versus those from central Somalia (galti)
• A rough north-south regional division (the Bula Hawa-Dolo-Luuq districts are a
separate trade corridor from the Bardhere-EI Waq districts)
• Ethiopian-backed Marehan versus Marehan aligning with Mogadishu-based factions
• Low-caste (midgan) sub-clans versus dominant sub-clans
• Secular clan elders and faction leaders versus Islamists

These affiliations and cleavages do not always correspond as neatly as they should
and can lead to unusual, even baffling political bedfellows. They also tend to induce very
temporary and unstable alliances and frequent switching of sides. Because the fighting
has been mainly within the Marehan clan, it has generally been constrained, as clan
elders have been able to step in and contain the conflicts. But the disputes have proven
difficult to resolve and easy for outsiders to manipulate.

At stake in these factional disputes is a combination of political and economic

interests. Politically, factional leaders seek to maximize the number of districts and towns
they control in Gedo region to enhance their political standing at the national level, which
is the major preoccupation of Marehan political figures. Economically, control over taxes
collected at the border crossing is an important source of funding and a source of conflict.
Despite the fact that AlAI has in the recent past been one of the actors in the political
divisions in Beled Hawa, ideology is not a particularly important component of the
conflict.

Prospects for Resolution. The conflict over control of Beled Hawa is a classic example
of a broader crisis throughout much of Somalia in which disputes over urban areas of
some political or economic value are primarily driven by Mogadishu exiles seeking to

- 37 -
consolidate a power base and a livelihood in their clan hinterland. In the process their
interests collide with those of their local kinsmen, sometimes producing armed clashes.
Seen from this perspective, the conflict in Beled Hawa will be unraveled the moment
there is a sustained peace in Mogadishu and an opportunity for the Mogadishu exiles to
return to pursue their political and economic ambitions there. Because sustained peace in
Mogadishu is not a likely scenario in the near future, this is not a happy prognosis for
Beled Hawa.

Still, there are a number of factors which might work in favor of a durable peace
in Beled Hawa even in the event Mogadishu remains closed to the Marehan galti. First,
the Beled Hawa feud is intra-clan in nature, and within a clan with strong traditional
elders or ugasses. The paramount chief of the Marehan, Ugaas Omaar, has a strong
political interest in reviving unity among his clan (and in the process enhancing his own
authority over factional leaders); he also has real social and moral authority within the
clan, giving him a certain amount of leverage with factional leaders. Second, the
Marehan face a number of important challenges requiring greater clan unity - in
Kismayo, in EI Wak, and at the national level in the TFG. Third, the prolonged
insecurity in the area has pushed cross-border trade elsewhere - indeed, it has been a
boon to smaller border posts such as Dar es Salaam and Geriley - and in the process has
hurt local economic interests. Fourth, the long-standing presence of several international
NGOs such as Trocaire have helped to build and maintain a network of professional
Somalis working within the NGO community. That network crosses clan and district
lines and can serve as a line of communication and trust-building. Following the killing
of an ADRA NGO staffinember in October 2003 (see the EI Wak case, chapter 5.2),
local and international NGOs operating in Be1ed Hawa formed the Beled Hawa NGO
Consortium, which has formalized the cross-clan cooperation in the NGO community and
serves as an umbrella civil society organization advocating for conflict prevention and
management. Finally, several promising civic peace groups have formed in Gedo region,
organized in a consortium known as the Gedo Peace Group (GPG); they cooperate with
the Mandera District Peace Committee (MDPC), expanding the reach of local civic
groups to promote peace.

5.4 The Southern Border/Dobley

The zone south of EI Wak - referred to here as the Dobley area - constitutes the
longest stretch of the Kenya-Somalia border. It is a particularly interesting case study
because over the past fifteen years it has gone from being the most violent and dangerous
area along the entire border to being the most secure area currently. This border security
cannot be taken for granted - the rapid deterioration in security in EI Wak and Mandera
district in the past few years serves as a cautionary note that peace in the border area is by
no means consolidated. But the impressive reversal of lawlessness in the Dobley area
merits close attention.

Background. Throughout the 1990s, the Dobley area was considered a "badlands" on
both the Somali and Kenyan side of the border. Unlike the northern tier of the border, the
southern border area experienced only a few instances of major armed clashes between

- 38 -
rival clan militias - mostly on the Somali side of the border, where the Aulihan and
Mohamed Zubeir clans fought over control of Dobley.49 But the southern border was
plagued by much more banditry and general lawlessness than northern border areas.
Much of the banditry in Garissa was, according to Paul Goldsmith, "fmanced by well-
connected trade barons who recruit from the pool of retired army personnel and school
leavers--a new class of professional and sophisticated highwaymen."so

The southern border area is distinct from the northern border zone in a number of
important respects. It is inhabited principally by a single Somali clan-family, the Absame
(of which the Ogaden clan is the largest lineage in the area); it is generally cattle
rangeland; it is the site of the enormous Dadaab refugee camp; and its border area
features two rivers, agricultural communities, and the two largest commercial cities in the
entire border area - Kismayo, Somalia, and Garissa, Kenya.

Some of the worst scenes in Somalia's famine and humanitarian crisis of the early
1990s occurred in this area. The route from the Jubba valley to Dobley was one of the
most heavily traveled paths for desperate refugees fleeing by foot from the war and
famine in Somalia. Some of the most notorious Somali warlords - General Mohamed
Said Hersi "Morgan," Siyad Hussein, Col. Omar Jess, Ahmed Hashi - operated in this
region. Their militias only rarely fought with one another; instead, they devoted most of
their energies to preying upon IDPs and refugees. The area around Dobley refugee camp
earned a reputation as one of the most dangerous and violent places in the entire region;
women gathering firewood in the bush were routinely raped by predatory militiamen, aid
convoys were looted, and refugees subjected to extortion and shakedowns. On the
Kenyan side of the border, the roads in and out of Garissa were infested with heavily
armed bandits, and could be traveled only with military escorts. Even the Kenyan police
outpost at Liboi was attacked and looted by Somali militiamen. When the Kenyan police
did venture out, it was often to join the bandits in predatory attacks on the refugees.
Police were frequently accused of rape and extortion, and generally seen as part of the
problem rather than the solution.

Conflict Drivers. Levels of insecurity and banditry were so high in the Dobley area in
part because of a puzzling lack of political cohesion and weak leadership among the
Absame. Most of the top Absame leaders on the Somali side of the border were absent,
living in Nairobi or Addis Ababa. On the Kenyan side of the border, ambitious Absame
leaders focused on national politics in Nairobi. To the extent that they played a role in the
border areas, it was generally destructive, fomenting clan violence with their private
militias for political gain.

In the immediate Dobley/Dadaab area, the source of insecurity can be attributed

in large part to a single warlord, Siyad Hussein. Hussein, an Auhilan militia leader allied
for years with General Morgan, took Dobley by force from the rival Mohamed Zubeir
clan in the early 1990s. Dobley controls an important route for cattle destined for the

49 Further away from the border, the Lower Jubba region has been beset by much more chronic factional
warfare, but that fighting rarely spilled over into the border area.
50 Goldsmith, "Cattle, Khat, and Guns" p. 29.

- 39 -
Kenyan market, as well as overland commercial trade. It also is located close to Dabaab
camp, giving Hussein's militia ample opportunity to prey upon refugees. Hussein was in
1991-92 held personally responsible by some international relief organizations for
orchestrating the widespread diversion of food aid to famine victims in the Jubba valley.
In Dobley, he created a fiefdom of banditry and predation which afflicted the entire zone
for years.

Economically, the southern border region missed out on the expanding transit
trade which was generating so much commercial opportunity along the northern border in
the 1990s. This was partly due to the high level of banditry in the area and the poor roads
connecting Dobley to Mogadishu and Kismayo. But it was mainly the result of the
chronic conflict over control of the port city of Kismayo. In theory, Kismayo is much
better placed to serve as the main entry point for transit trade from Somalia into Kenya. It
possesses the only functioning all-weather seaport in southern Somalia, and is
considerably closer to Kenya than Mogadishu. But since 1991 the city has been the site of
repeated clashes between competing militia, both from the region and from outside areas.
For most of the period since 1991, the clan dominating the hinterland - the Absame-
have been frozen out of Kismayo by some combination of Harti, Marehan, and Haber
Gedir Ayr militias (currently a loose alliance of Marehan and Haber Gedir Ayr militia,
forming the Jubba Valley Alliance or N A, control the city). In response, the Absame
blocked most commerce out of the city toward Kenya throughout the 1990s. Kismayo
became a chronically contested and besieged town, closed to the interior.

Sources ofPeace and Security. Peace and security in the southern border area have
improved considerably since 2000. First, a regime change occurred in Dobley - the
Mohamed Zubeir ousted Hussein's militia, and Hussein himself died. This returned to
power a group of local clan leaders who are guri or indigenous and who have greater
stakes in peace and security, especially to protect the clan's valuable cattle trade across
the border. They have used modest tax revenues from cross-border trade to create a small
police force and administration in Dobley; they have formed a Peace Committee,
modeled on the Kenyan structure, to liaise with Kenyan counterparts to prevent cross-
border crime and maintain the peace; and they have successfully reduced banditry and
crime in the area.

Second, a "pax commerciale" has gradually emerged between the N A in

Kismayo and the Absame in the border areas. This is not by any means a full
reconciliation - the Absame continue to insist that they must be allowed to return to
Kismayo as full partners in governing the city, and that the JVA is an outside occupation
force. But shared commercial interests have allowed the rise of some overland transit
trade from Kismayo through Dobley and Diif, principally of sugar, since the late 1990s. 51

5\The sugar import trade is dominated by a single Mogadishu-based businessman. He opted to shift
imports from the Mogadishu area to Kismayo mainly because off-loading sacks of sugar at beach ports
produced too much water-damaged sugar; the all-weather seaport at Kismayo allows ships and dhows to be
off-loaded directly onto a dry dock. Kismayo has yet to replace other trade items which arrive via
Mogadishu because Mogadishu remains the site of most warehousing facilities and because most of
Somalia businesspeople reside in Mogadishu.

- 40-
As elsewhere in Somalia, hostile clan relations are often trumped by shared business
interests. Dobley and Diif have also benefited from increased overland trade from
Mogadishu via Jilib. Dobley leaders describe their settlement now as a "business center"
with warehouses and internet access.

The strength of this pax commerciale in the region was put to the test in 2003,
when a botched attempt to capture Kismayo by General Morgan resulted in his retreat
through Dobley into Kenya. JVA militia pursued his forces and, in retaliation for what
they perceived (incorrectly) to be Absame complicity with Morgan, proceeded to attack
and bum Dobley. Fourteen Dobley police were killed in the attack, and much of the
settlement destroyed. The losses sustained in this attack remain a worrisome unresolved
issue between the JVA and the Absame clan, but remarkably the Kismayo-Dobley cross-
border trade was only briefly interrupted before resuming.

Third, the profitability of and rapid growth in cross-border cattle from the Lower
Jubba valley into Kenya is increasingly a central pillar of the Absame economy, and one
which requires a modicum of border security and peace. Fourth, there has been a gradual
improvement of relations between local and refugee populations in Dadaab, now notably
better than relations between local residents and refugees at Kakuma camp in northern
Kenya. As a source of remittances and food distribution, Dadaab produces commercial
opportunities for the region and is increasingly integrated into the local economy. The
decline in services provided to the refugees in recent years - a function of donor fatigue -
has had the unintended effect of reducing local resentment over what was earlier
perceived to be favored status for refugees.

Fifth, the fact that the dominant Somali clan in the southern border area - the
Absame - are stakeholders in Kenyan national politics as well as Somali national politics
may be having an ameliorating effect on regional politics. Politically-driven clashes in
the Lower and Middle Jubba involving the Absame are much fewer today than is the case
with the Marehan in Gedo region, in part because the Absame are increasingly focusing
their energies on Kenyan politics, where they hold multiple seats in parliament and
cabinet positions (or shadow cabinet positions). Unlike the Marehan in Gedo region and
most other Somali clans, the Absame in the Transjubba region are currently stakeholders
more in a neighboring country than in Somalia itself. Absame elders have explained the
lack of action regarding their disenfranchisement in Kismayo not as a sign of weakness
but rather as indifference. 'We don't need Kismayo," one noted. 'We have Kenya."

Finally, Garissa town appears to be playing a quiet but substantial role in

promoting peace and security in the southern border zone. Garissa has developed into a
major commercial hub - in many respects it is the commercial capital of the entire border
region. In the process, it has grown quickly and has become a "cosmopolitan" city where,
though the Absame are the main clan there, all Somali clans may live and conduct
business. For urban Absame, especially those displaced from Mogadishu, the fact that
they have access to an increasingly large "Somali" city to reside in reduces the need to
fight for access to Somali cities elsewhere - specifically, Kismayo and Mogadishu.

- 41 -
Outstanding conflict issues. The improved state of peace and security in the southern
border area is impressive but is not a "consolidated" peace and is vulnerable to the kinds
of reversals recently witnessed in Mandera and EI Wak. Preventive measures are thus
especially appropriate for this half of the border area. Most of the threats to the area's
peace and security stem from sources of conflict endemic to the entire border region.
Local leaders stress that acts of crime - thefts, rape, murder - remain the single greatest
threat to local peace. This puts a premium on the ability of local clan elders to act quickly
to negotiate blood payment, in order to prevent retaliatory measures such as revenge
killings. Use of HF radios, though technically illegal in Kenya, has been an important
means of expediting communication and intervention by clan elders in the area.

Resource competition, especially over control of boreholes, is a chronic source of

conflict requiring vigilance on the part of clan elders and civic peace groups. Likewise,
the proliferation of locations on the Kenyan side of the border is a flashpoint for conflict
across the entire border area. And while rival trade routes in the southern border area
have not to date been a source of conflict, the potential remains for groups to be tempted
to destabilize border crossings held by other sub-clans in order to divert trade to their
border town.

A number of conflict issues are partially or wholly specific to the southern border
area. First, control of Dobley remains an unresolved conflict with the possibility of
renewed clashes. Though the area has long been considered part of the Mohamed
Zubeir's territory, the Aulihan have not accepted their ouster from the town and no
reconciliation between the two clans has occurred. The two clans are in fact long-
standing rivals within the Absame clan-family, and Dobley will be the most likely
flashpoint should anned conflict emerge between the two. The unresolved impasse over
control of Kismayo is also a ticking time-bomb for the entire region. As was seen
recently with the sacking of Dobley by the N A, spillover from the Kismayo conflict can
quickly impact the border area. Finally, the harvesting of acacia trees for charcoal
production and export out of Kismayo is expanding toward the Kenyan border and as it
does the charcoal businessmen and their workers will eventually come into conflict with
local pastoralists. The charcoal export industry has had devastating effects on parts of the
Lower Jubba valley and is a long-tenn environmental disaster.

6. Key Sources of Conflict in the Border Area

A comparison of the border area conflicts and conflict dynamics in adjacent areas
helps to identify several key conflict drivers at work in many if not all of these cases.
Importantly, almost all of the factors identified in both this section on conflict drivers and
in the following section on sources of conflict management can, depending on
circumstances, play either a constructive or destructive role in conflict. Commercial
interests can produce pro-peace, cross clan business networks, but can also generate
conflicts over trade and economic incentives to foment lawlessness and ethnic cleansing.
Clan elders and the customary law they apply can be a force for peace or they can
contribute to war-mongering. Ethnic identity itself can be a dangerous, exclusivist force

- 42-
of "political tribalism" or a flexible (even fictional) social instrument of peace and "moral
etlmicity."

Listing conflict drivers in this manner also begs the question of which is in fact
the real source of conflict - the "driver of drivers" -- and which is either a mere
symptom, a secondary cause, or is simply invoked as part of a political discourse to
disguise true motives. Of all the conflict drivers in the border area, no three are as
entangled as pastoral rangeland disputes, environmental stress, and politics. Each
inflames and exacerbates the other and none can be properly explained in isolation from
the others Pressures of migration, increased herds, and environmental degradation
heighten communal disputes over land; the land disputes provide political figures with an
easy tool to mobilize etlmicity; the political proliferation and misuse of locations and
boreholes to create zones of exclusive clan grazing rights both creates and accelerates
communal tensions and worsens environmental stress; and thus a vicious cycle is born.
Understanding how these conflict drivers reinforce one another is as important as
understanding how they contribute to the conflicts themselves.

6.1 Environmental Stress and Poverty

Virtually every analysis of Kenya's troubled pastoral areas emphasizes that

environmental stress and severe poverty and underdevelopment combine to render these
semi-arid zones chronically vulnerable to armed conflict, communal clashes, and violent
crime. The Kenya-Somalia border area partially supports this claim. Most of the
conflicts in the Kenya-Somali border area are driven principally by other, mainly political
factors, but environmental stress and underdevelopment are critical underlying sources
of instability. Particularly in the northern half of the border area, growing evidence
suggest that the poorer tier of pastoralist households are increasingly facing immiseration.
Destitute pastoralists form an important portion of the new urban populations, where
prospects for employment are bleak. Uneducated and unemployed young men are easy
marks for recruitment into militias or criminal gangs. Heightened communal anxiety
over access to scarce resources - pasture, wells, jobs -- are easily exploited by politicians
and others to promote divisions and foment violence. Very poor access to social services,
especially education, and the almost complete absence of a government presence beyond
a few police and military garrisons breeds a profound sense of alienation in much of the
Kenyan border area. On both sides of the border, the rising gap between haves
(merchants, NGO staff, politicians, the professional class, households receiving
remittances) and have-nots (especially destitute pastoralists) is growing and breeds
frustration and resentment as well. Recent research links worrisome environmental
deterioration in the region to the proliferation of boreholes and settlements which are
dispensed for reasons of political patronage and which are disrupting seasonal migration
and degrading what was once prime ungrazed (usub) rainy season pasture. 52

52Robert Walker and Hassan G. Omar, "Pastoralists Under Pressure: The Politics of
Sedentarisation and Marginalisation in Wajir District, Northeast Kenya," Nairobi: Oxfam-GB, (July 2002).

- 43 -
6.2 Kenyan State and Local Government Policies

On a number of levels, political forces have been the single most important driver
of conflict and instability on the Kenyan side of the border area. This fact points to a
paradox in northern Kenya. While retreat of state authority from the border areas is a
factor in the high levels of insecurity there, the presence of state agents has often been a
major driver of conflict as well. It is thus not enough to contend that more robust state
authority is needed in the border zones to improve security. More important is that the
state presence be constructive.

Proliferation and Abuse ofLocations/Ethnic Cleansing. The case of Murille-Garre

conflicts in Mandera district is the clearest examples of how the proliferation of locations
fuels conflicts. Though officially portrayed as "bringing services closer to the people,"
the multiplication of locations has in reality been an easy and effective form of political
patronage to win and maintain support from sub-clans seeking their "own" local
government, settlement, and services. The abuse of locations by local leaders, who view
locations not as administrative units but as zones of exclusive clan prerogatives, has
fueled localized ethnic cleansing and greatly increased the political stakes over control of
locations. This same dynamic has occurred at higher administrative units as well, as the
expulsion of Murille cattle herders from rangeland in Isiolo district in 1995 demonstrates.
All of this is a variation of the national policy debate surrounding the controversial
majimboism ("regionalism") policy promoted by officials in the Moi administration in the
early 1990s. The version of majimboism which was promoted at that time emphasized
the exclusive rights of "indigenes," especially regarding land and property rights at the
local level. This abuse of sub-national governance units, in which local administrations
are transformed into "clanustans," is a distressing example of how political
decentralization can, under the wrong conditions, merely localize ethnic dominance and
repression rather than ameliorate it.

The ethnic exclusivity strategies executed via the locations can be a way to seal a
victory by institutionalizing a clan's claims to territory, but it has at times been embraced
as a defensive strategy by clans fearing that in the current, unclear tenure and rangeland
usage system, militarily stronger clans will push weaker clans off the land.

Parliamentary Constituencies and Rise ofEthno-Representation. Several of the conflicts

examined here were triggered or exacerbated by creation of new parliamentary sub-
districts. Two types of conflict have emerged. One has been inter-clan conflicts over
control over the new seat. Importantly, it is also the case that creation of new
parliamentary seats has occasionally reduced clan tensions by giving two clans in conflict
each their seat. Second, Parliamentary sub-districts have also been misused to deprive
minority clan members the right to access pasture or own businesses in those zones, a
form of ethnic cleansing akin to what has occurred in many locations.

Rangeland policy. There is near universal consensus that the disjoint between formal
state policy on rangeland access and local customs and practices is a major source of
conflict.

- 44-
Politicians as Fomenters ofConflict. A major problem throughout much of Kenya since
the advent of multi-party democracy has been "political violence" - the fomenting of
ethnic clashes and use of private militias by political leaders to attack rivals, both those in
power and their challengers. North-East Province is no exception. Politically-motivated
violence was a major problem in Garissa district in the 1990s; intra-Garre political
rivalries over representation in parliament contributed to tensions in Mandera district in
the past year, and some Wajir MPs have been accused of contributing to militia build-up
in the El Wak conflict. These charges are often difficult to prove but are widespread.

Government Complicity in Communal Violence. This has not been a factor in the post-
Moi era, but at times in the 1990s commercialized cattle rustling in North-East Province
and northern ;Kenya was linked in the Kenyan media to high-ranking government
officials. The previous Kenyan government was also accused of providing arms to local
allies who were given license, sometimes with direct support from the police, to attack
rivals. 53 The fact that the Kenyan government in more recent times has been strongly
committed to bringing peace to its hinterland is a welcome change and serves as a
reminder that, if a central government is either not committed to ending armed conflict
within its borders or is actively complicit in fomenting it, local efforts at peace-building
face an almost insurmountable task.

6.3 State Collapse and State-Building in Somalia

The state is also a factor in conflict on the Somali side of the border, for three
very different reasons. First, as is clear from the case material above, the complete and
prolonged collapse of the Somali state has had disastrous impact on the Transjubba
regions, producing a context of lawlessness and anarchy that is only partially mitigated
by attempts to strengthen local governance. Second, efforts to revive a central
government in Somalia have been conflict-producing. Because representation in Somali
national fora is now explicitly clan-based, state-building negotiations encourage clans to
maximize territory they can claim to control. Loss of a presence in or control of a district
can carry serious consequences for political elites vying for top seats in the transitional
government. This has been a factor in the recent El Wak conflict. Third, when state-
building efforts fail in Somalia, the international community has periodically sought to
work with sub-national, regional or transregional polities. In the late 1990s, this was
termed the "building block approach" to state revival, and indications are strong that a
comparable policy is likely to emerge if and when the TFG fails. That increases local
political competition for control over regional or multi-regional polities, and is likely to
be a conflict issue in both Gedo and Lower Jubba regions.

6.4 Hardening of Ethnic Identity

53Among the many analyses documenting government complicity in Kenya's communal violence, see
Steven Brown, "Quiet Diplomacy and Ethnic Clashes in Kenya," in Donald Rothchild, Chandra Sriram
and Karin Wermester, eds. From Promise to Practice: Strengthening UN Capacitiesfor the Prevention of
Violent Conflict. Boulder: Lynne Rienner, 2002.

- 45 -
 The Kenya-Somali border area is an area where a number of major ethnic groups
overlap. The region offers up ample evidence of how in the past local groups embraced
and utilized flexible ethnic identities to negotiate access to resources and protection. 54
That strategy, which has been so effective for centuries, is today facing a very hostile
political environment. Contemporary political systems of representation, voting,
administration, rights, and land access in Kenya, Ethiopia, and Somalia are increasingly
based explicitly or implicitly on ethnicity. The result is not only an epidemic of localized
and partial but insidious ethnic cleansing, but also a hardening of previously fluid ethnic
identities in the region, and a level of ethnic mobilization not seen previously in the area.
In the hands of the wrong politician, hardened, mobilized ethnic identity in a context of
worsening scarcity and stress is a fonnula for communal violence on a scale that is
unthinkable but which has already occurred in the Greater Hom of Africa. For local
groups which in the past have thrived using ambiguous and flexible ethnic identities - the
Garre, Rendille, Boroma, and others - the current environment of ethno-politics in the
region is an existential crisis with enonnous consequences. Current political forces in the
border area are transfonning ethnic identity into a much more toxic, inflexible, and
dangerous factor than has ever been the case in the past.

6.5 Regional Economy/Cross Border Trade

The transit trade across the Kenya-Somalia border has a complex relationship to
peace and conflict. In some instances - EI Wak, Bulo Hawa, Dobley, and Kismayo -- it
has at times been a source of tension and anned clashes. This is mainly due to the fact
that the new cross-border commerce has introduced a new and lucrative new source of
livelihoods and wealth in a zone of extreme economic scarcity. The specific aspects of
commerce which have tended to serve as an item over which groups fight include control
over border crossings, where tax revenue can be collected; business partnerships with
wealthy traders in Mogadishu, who intentionally play local groups off one another for
their own gain; and competition for contracts, especially with international aid agencies.
A handful of wealthy businessmen profit from and may be complicit in prolonging
humanitarian crises, due to their long-running service to aid agencies transporting food
aid. In several instances - especially in Kenya - businesspeople have opted to exploit the
exclusionist impulse in locations to drive rival businesses out of the area. In the case of
Kismayo, control of the seaport has been a major source of revenue for both the militias
controlling it and the businessmen importing sugar and weapons and exporting charcoal.

6.6 Outsiders

An enduring concern of the border communities is that the chronic conflicts they
suffer from are the result of external agendas, by powerful outsiders or galti who are not
stakeholders in local peace. Outsiders are also held responsible for rendering local feuds
more deadly, for committing crimes leading to cycles of clan violence, for ignoring local

54The single best study on fluid ethnicity in the region is Gunther Schlee, Identities on the Move: Clanship
and Pastoralism in Northern Kenya (Nairobi: Gideon S. Were Press, 1994).

- 46 -
customary law or xeer, for engaging in polemics and fund-raising for war in pursuit of
their parochial political interests, and for exploiting local divisions.

There is strong evidence for all these claims in almost every case of armed
conflict in the border area. From Kismayo to Dobley to EI Wak to Bulo Hawa, "non-
local" interests - displaced Somalis from Mogadishu, Mogadishu-based merchants and
factions, and Nairobi-based interests - have had an often decisive role in provoking
armed violence. The current conflict over EI Wak, which has drawn in Garre militia from
Ethiopia and Marehan militia from Kismayo which was in part triggered by manipulation
of local tensions by Mogadishu-based businessmen, and which has prompted
inflammatory rhetoric and fund-raising for war chests among the diaspora, is the most
compelling example of this problem. Likewise, the long-running conflict in Bulo Hawa is
mainly built around a guri-galti split. In some conflicts along Ethiopia's border with
Kenya and Somalia, Ethiopia itself plays the role of external force which can in some
instances contribute to or become a direct actor in conflicts.

At the same time, it is tempting for local observers to attribute all their woes to
external forces, thereby absolving themselves of responsibility for the conflict. It is also
the case that in some instances, outsiders' interest has been in peace and open access, not
ethnic clashes and instability. As long as the border areas possess economic value as a
trade corridor and political value as a source of seats for national level representation,
outside interests will continue to intrude on local affairs.

6.7 Contested Urban Space

One aspect of the border area conflicts which diverges somewhat from national
trends in Kenya, but which closely mirrors broader patterns in Somalia, is the central role
played by contested urban space. Unlike many of the serious ethnic clashes in Kenya's
rural areas since 1990, land itself is not as central to Kenya-Somalia border conflicts.

Instead, most of these conflicts are struggles over control of a coveted urban space
or settlement. The intra-Marehan fight in Bulo Hawa is entirely over the town of Bulo
Hawa, not surrounding rangeland. Likewise, the clashes at EI Wak have had nothing to
do with dispute pasture and everything to do with control over the town itself. Further
south, Mohamed Zubeir-Aulihan clashes were over the town of Dobley and control of
trade, not pasture; the long-running conflict over Kismayo is also entirely over the city,
not the hinterland.

What this suggests is that, at least in the border area, urban interests and
populations are the main protagonists in armed clashes. The pastoralists serve as a
principal source of militiamen, but in the service of someone else's interests. While this
finding is not universally applicable in the border areas, instances where it is true
contradict some of the conventional wisdom about Kenya's conflict-ridden border area.
Conventional wisdom observes that the conflict-prone border areas are all zones of
nomadic pastoralism, and then, conflating correlation with causality, leaps to the
conclusion that pastoralism must therefore be the cause of the endemic conflict.

- 47 -
Pastoralism as a mode of production is unquestionably a contact sport, and land pressures
in the pastoral sector are severe, but the troubles of the Kenya-Somalia border area are
more closely linked to the interests of urban elites. To the extent that these conflicts over
urban centers involve displaced urbanites from Mogadishu, a permanent peace in
Mogadishu would immediately reduce pressures on towns on the Somali side of the
border.

6.8 Regional Spillover

The border area is highly susceptible to spillover of conflict from adjacent

regions. This is most in evidence in the northern tier, where mujimbo-style ethnic
cleansing as far away as Isiolo triggered a shockwave of displacement and conflict over
resources and political rights in Mandera district years later. On the Somali side of the
border, spillover from the conflict in Kismayo sends ripple effects to the Kenyan border.

6.9 Crime·

Virtually all cases of armed conflict in the border area involve acts of crime as a
precipitating cause. In some cases, a murder - usually of a prominent figure - triggers
reprisal killings which spiral into a cycle of inter-clan violence. In other cases, murders or
other crimes contribute to deteriorating clan relations and rising distrust, serving as an
emotionally powerful part of the build-up to armed clashes.

In normal circumstances, customary law (xeer) is adequate to resolve a crime

between clans. But when circumstances are not normal - when the murdered person was
a very prominent figure in the clan, when inter-clan relations were already strained by
other issues, when the murder appears politically motivated rather than the result of a
personal feud, when clan elders take too long to negotiate the diya payment, or when the
perpetrator's diya-paying group drags its feet or refuses to pay diya -- reprisal killings
are almost inevitable. Once two clans begin to travel that route as opposed to blood
payment, the cycle of killings often spirals into war. The armed conflicts in the border
are which were triggered by acts of crime cannot therefore be blamed on the killings
themselves, which serve as the match lit and set upon very dry kindling. But measures
which accelerate and facilitate the often demanding work of clan elders to negotiate diya
can reduce the likelihood of reprisal acts. In cases were xeer is inadequate to deal with an
inflammatory crime - rape is the most apt case in point - clan elders in the border areas
are adamant that only formal government judicial processes are in a position to handle the
crimes. This recognition of the limits of customary law as a conflict prevention
mechanism is an important step in establishing a division of labor between the weak state
and non-state local actors with limited capacities.

One of the most important observations to glean from the case material of the
border areas is that the distinction between organized, violent crime and low-intensity
communal warfare is increasingly hard to make, especially in the eyes of local
populations whose security is imperiled equally by both.

- 48 -
6.10 Small Arms Proliferation

The widespread ownership and easy availability of small anns in the border area
is widely cited as an intensifier of anned conflict. 55 The flood of small anns in the Hom
of Africa is well-documented, as is the devastating impact of semi-automatic weaponry
on communal conflicts in the region. This report can only repeat the observation that the
anns do not cause the conflicts, but multiply their negative impact. Criminal violence
produces much higher casualty levels, criminal and militia gangs now often outgun police
and military units, and the number of people a militia can massacre in a raid is vastly
higher thanks to semi-automatic guns. The result is that casualty rates in contemporary
violence in the border area are much higher than was the case fifty years ago,
overwhelming customary law designed to handle conflicts from an earlier, less lethal era.

Given the chronic insecurity and porous borders in the region, small anns
proliferation will remain a dangerous reality for border area communities for the
foreseeable future. Some modest efforts in Wajir have succeeded in disanning youth, but
prospects for large-scale disannament in the region are remote for now. In southern
Somalia, the only factor which has reduced the lethal risk posed by near-universal
ownership of semi-automatics is the substantial rise in the cost of ammunition, which has
shortened the duration of anned clashes since the early 1990s. 56

6.11 Borders

The existence of the borders shared by Somalia, Ethiopia, and Kenya are
themselves a periodic contributor to conflict. The principal role they play in conflict is as
a source of safe haven for anned groups - criminal gangs and militia - which commit
acts of violence and then cross to the safety of their home country and disappear. The
shared border area near Mandera is known locally as the "Bennuda triangle" for precisely
this reason. This dynamic points to a curious aspect of the borders - they are relatively
unpatrolled, ungoverned and porous, but not irrelevant. Kenyan military do not cross the
border in hot pursuit of Somali bandits for fear of attack; Somali militias cross into
Kenya in pursuit of rival combatants only reluctantly, for fear of encountering the
Kenyan military or police; and Somali anned bandits take the same risk when mounting a
cross-border raid. In practice, this has meant that the Kenyan Garre have been able to
use the border to launch attacks on EI Wak in Somalia and retreat across the border with
little fear of Marehan counterattack, a tactic which infuriates the Marehan and has led
them to accuse Kenya of favoring the Garre. Further south, Somali bandits have until
recently stolen vehicles and even looted the Kenyan police station at Liboi and then
retreated back to Dobley. On the Kenyan-Ethiopian border, militia and possibly
Ethiopian paramilitary forces have crossed into Kenya to commit devastating livestock
raids.

55 A caveat to this claim is that in a few instances easy availability of small arms has served as an equalizer
between clans, and hence a deterrent to fighting. See Goldsmith, "Cattle, Khat, and Guns."
56 Menkhaus, Somalia: State Collapse, p. 30.

- 49 -
7. Sources of Peace, Security, and Conflict Management

Given the confluence of crises in the Kenya-Somalia border regions - the

complete and prolonged collapse of the Somali state, years of civil war and famine in the
Jubba regions, massive refugee flows and displacement, copious arms trafficking, large-
scale-smuggling, intermittent terrorist activity, inability of the Kenyan state to govern its
frontier areas, and the rise of majimbo-inspired land clashes and ethnic violence across
much of rural Kenya -- the border zone ought to be one of the most anarchic, violent, and
dangerous places on Earth. Yet it is not. Remarkably, on both sides of the border, a
variety of factors have worked to produce local systems and practices which provide
uneven, fragile, but real security, predictability, rule of law, and conflict management.
The emergence of "governance without government" has been observed and documented
in southern Somalia in some detail. There, a mosaic of formal and informal local
authorities has emerged from the protracted collapse of the state. In the Transjubba
regions, clan elders and customary law have re-emerged throughout the area and provide
the principal source of rule of law. Municipal authorities have in several places - Luuq,
Kismayo, Beled Hawa, Dobley - provided more structured formal governance. Civic and
business groups have been weaker as sources of governance in the Transjubba regions
than in much of the rest of southern Somalia, but in recent years have become more
visible and active. Finally, sharia courts, operated by local clerics, have been periodically
active in parts of Gedo region but have not been as widespread as in the Mogadishu and
Shabelle river valley.

On the Kenyan side of the border, an even more impressive change has occurred.
Notwithstanding the recent setbacks in Mandera district and EI Wak, the region has gone
from being one of the most lawless of Kenya's troubled border areas to being one of the
more stable and safe. This improvement is the result of concerted efforts at the local level
to control and minimize the triggers of armed conflict and successfully deter violent
crime. Of the many factors which improve security in the border areas, either by
preventing or managing armed conflict or by reducing the threat of violent crime, the
following have been most important. As with the conflict drivers enumerated above, most
of these are capable of fomenting violence and conflict as well as peace and security.

7.1 Customary Law and Traditional Elders

Xeer, or customary law, developed between clans in the pre-colonial era as a

system for managing conflict, deterring crime, and dispensing justice in an environment
of statelessness. Xeer serves roughly the same purpose as "regimes" which help govern
relations between states in international politics. All parties benefit from routinized and
predictable cooperation on key issues of importance, such as governing use of common
resources.

Customary law was partially undermined in both Somalia and Kenya through
colonial and post-colonial state manipulation of the clan elders who traditionally dispense
and negotiate xeer. Contemporary state administrations tended to view xeer as a vestige
of an old and inferior system rendered obsolete by modem legal codes, police forces, and

- 50-
court systems. Use of paid chiefs or elders by the state to manipulate and control local
populations eroded the credibility and legitimacy of clan elders in their communities.

The collapse or retreat of the Somali and Kenya states in the border areas since
1990 has recreated in some manner the pre-colonial pastoral anarchy which xeer was
designed to address. The revival of the authority of clan elders and xeer in the border
areas has probably been the single most important source of security and conflict
management since 1990. In stateless Somalia, the dominant role of clan elders as
authoritative representatives of their clans in peace negotiations, as quasi-diplomats
managing relations with neighboring clans, and in adjudicating or mediating disputes
within their lineage has been largely unquestioned since the civil war of 1992, and
nowhere in the Transjubba regions is xeer anything less than central as a pillar of law
governing inter- and intra-clan relations.

On the Kenyan side of the border, reliance on xeer to manage conflicts, especially
its use to resolve cases of murder, has been more controversial, in part because it is a
direct challenge to the state legal system, in part because its defining features - the use of
blood payment to handle a crime in lieu of incarceration, and the principle of collective
rather than individual culpability that blood payment groups are based on - violate basic
precepts of modem jurisprudence and human rights. For some Kenyans, including some
Somali Kenyans, allowing xeer to supersede the state legal code is seen as a giant step
backward.

There are many other criticisms of both xeer and reliance on traditional authorities
as sources of representation and social authority. Elders can be venal, corrupt, and
inclined to foment ethnic divisions; their enforcement capacity is variable, and they have
rarely been able to prevent determined politicians and militia leaders from outflanking
them and sabotaging peace-building that does not serve their interests; and elevating the
status of sometimes illiterate elders over the educated and professionals within their clan
consigns communities to mediocre and parochial leadership. For its part, xeer enshrines
rather than overcomes clannism; xeer is limited in its jurisdiction to local clans, and is
overwhelmed by the sheer scale of death and destruction associated with modem
weaponry and war; it is woefully inadequate for protecting women's rights, especially in
dealing with the widespread use of rape as a weapon; and it is least effective in protecting
those whom the law is most obliged to protect - the weak and powerless.

All this is true. Yet the border areas provide persuasive evidence that the revival
and application of customary law by clan elders has, in the absence of an effective state
police and judiciary, been the single most powerful deterrent of crime. 57 Likewise, the
border area is replete with examples of how clan elders have used xeer to manage
conflicts in their early stages or resolve seemingly intractable conflicts. Clan elders have
also been effective outside mediators in peace talks. For all their defects, customary law
and traditional elders have been crucial to keeping what peace exists in the border zones.

57 Lineages have on occasion even executed a repeat offender in their clan in order to stop having to pay
onerous diya payments.

- 51 -
7.2 Civic Groups

Local CBOs have been remarkably active in the region, especially on the Kenya
side, where they fonn part of a vibrant network of the Kenyan non-governmental sector.
As elsewhere, they are a mixed bag. Some are little more than "pocket" or "briefcase"
NGOs composed of one or two people; some are politically compromised and serve as
fronts for clan interests or politicians; others are essentially local businesses seeking a
livelihood as sub-contractors for international NGOs. But local civic groups have clearly
matured and developed since the early 1990s, and today are playing a growing role in
promoting peace. They assist in a number of different ways. They are cross-clan
networks facilitating both routinized and "rapid response" communication, a critical and
often missing function in times of conflict; they help build trust and shared interests in
functional issues such as health care and education which transcend clan lines; they tap
into professional skills and expertise among local populations that otherwise can go
underused; they can provide good offices or mediation; they are flexible and open and
provide opportunity for social groups nonnally excluded from politics - such as women
and youth - to playa more central role; they are perhaps the best forum for engaging
local communities in discussion of underlying sources of conflict and strategies for
addressing them; and, thanks to their linkages outside the region, they are developing a
capacity as repositories of "best practices" for conflict prevention and management.

7.3 Peace Committees

The emergence of peace committees - umbrella groups of civic organizations,

traditional elders, and local governmental officials - has been the single most important
and effective community response to insecurity, crime, and anned conflict in the border
areas. The structure has spread from its initial experiment in Wajir to Garissa and
Mandera, has been emulated by cross-border communities in Somalia, and has spread to
many districts in rural Kenya. Peace committees fonn a central part of the "mediated
state" strategy emerging in Kenya (discussed below in chapter 8). They have a number
of virtues. They encourage and enshrine local ownership of conflict management; they
tap into local knowledge of conflict dynamics; they create a "multiplier effect" by
structuring collaboration among different types of local actors, each of which brings
different strengths; they are flexible and open regarding community participation; and
they encourage collaboration between state actors and civic groups. They also have a
number of weaknesses, discussed below. Still, the evolution of peace committees in the
border area has been central to improvements in public security and peace.

7.4 Kenyan State

The Kenyan government and Kenyan political dynamics have, on a number of

levels, been a major conflict driver in the border area. But it is equally true that the
impressive improvement in regional peace and security has been the result of positive
Kenyan government intervention and policies. This has been due in part to a greater level
of commitment to regional security and conflict prevention on the part of some branches
of the Kenyan government. District Commissioners have been impressively active in

- 52 -
promoting peace and security; the Arid Lands Resource Management Project (ALRMP),
in the Office of the President, has expanded its work beyond rangeland and water issues
to include conflict issues; and the Kenyan military and police, though still struggling with
capacity problems, possess a greater commitment to border security as part of counter-
terrorism policies. At least as important as renewed government willingness to improve
security and rule of law in the border area is the Kenyan government's realistic
assessment of its current capacities, and its openness to partnership with civic groups to
promote peace and security, including accepting "internally regulated group relations to
bypass state legal institutions. ,,58

7.5 Business Interests

Business interests may be the most variable of all the peace and conflict drivers in
the border area. Business interests in the border area have generally been effective as a
source of conflict prevention and management for a number of reasons. First, the two
most important economic activities in the region, livestock trade and cross-border
commerce, require basic security, safe markets, and open roads. To the extent that war
and banditry is bad for business - as is now clearly the case in Dobley and Garissa, and
hopefully true in Mandera and EI Wak - business leaders will back peace efforts. Second,
where businesspeople have invested in fixed assets such as telecommunications offices,
shops, hotels, and houses, warfare is undesirable. Most importantly, commercial activity
in the border area relies upon multi-clan partnerships and networks to move goods safely
across the region. Those partnerships and networks can be a valuable channel for cross-
clan dialogue and cooperation on matters of peace and security.

7.6 Islam

The use of national Islamic leaders in Kenya to mediate the Mandera peace in
2005 helped to highlight the possibilities of Islam as a force for peace in the region. In
some districts of Gedo region, sharia courts have been employed to maintain rule of law.

7.7 "Cosmopolitan" Towns

Contested urban space is listed in this study as a source of conflict. But in some cases,
urban areas have come to play an important role for peace, rule of law, and - most
importantly - ethnic co-existence. Towns which develop a culture of "cosmopolitanism"
- that is where members of any clan or ethnic group are free to settle and do business
there, and where an identity with the town begins to rival clan identity - are critical
sources of peace in the border area. Today, Garissa, Luuq, and arguably Wajir and
Mandera are the most cosmopolitan towns in the border area. They are important as
centers of cross-clan communication and networking, and as antidotes to the exclusionist
ideology which infonns clan attitudes towards smaller settlements and locations. To the
extent that cosmopolitan towns flourish while single-clan settlements stagnate, these
islands of co-existence may eventually help reverse dangerous trends of ethno-politics in
the region.

58 Goldsmith, "Cattle, Khat, and Guns," p, 41.

- 53 -
 SECTION III: Policy Issues and Considerations

8. Promise and Limits of the "Mediated State"

The development of a government-civic partnership to execute core state

functions in the Kenyan border area is very significant, not only in the immediate region
but also as perhaps the most advanced and formalized variation of a broader trend in
governance by weak states in their frontier areas. This trend has been until recently
largely invisible to outside observers and remains poorly understood, both in Kenya and
elsewhere. 59 Observers are gradually becoming more aware of the new phenomenon,
but have struggled to explain it and adequately convey its significance. A recent article
by Letitia Lawson and Donald Rothchild captures both the new awareness of this trend
and the difficult of translating it into words:

"Africans have begun moving away from colonially designed juridical statehood
to fashion empirical formulas that respond to the messiness of their current
realities. Only time will reveal whether these new, flexible structures prove an
effective response to ...state weakness.,,60

In northern Kenya, where communities are further along in this process of

fashioning new formulas to respond to the "messiness of their current realities,"
observers have also struggled to explain exactly what the state-civic partnership is and is
not, producing sometimes clumsy descriptions which tend to obscure and understate the
importance of this experiment. Some have described it as an "ad hoc" arrangement by
local officials overwhelmed by borderland lawlessness. Those coming from a peace-
building and NGO perspective have tended to explain it as local or grassroots movement
to which government has acquiesced on limited matters of conflict management. Those
not following developments in the peace-building sector are thus apt to miss it entirely.

It is the conclusion of this study that the governance model being negotiated,
implemented, and experimented with in northern Kenya is· more than an ad hoc peace-
building strategy, more than a post-colonial version of "indirect rule," and more than a
convenient division of labor between local government, civic groups, and traditional
authorities. It is better understood as a type of "mediated state," a concept initially used
to explain state-frontier governance in early modem Europe. Far from being a relic of
medieval Europe, however, the mediated state model in Kenya today may be at the
forefront of an emerging, largely unrecognized, hybrid form of state-building in weak
states.

59 There are several reasons why this is so. Reliance on non-state sources of governance is viewed by many
observers as a symptom of state failure, not an emerging mode of governance. Second, it is essentially an
indigenous process, beyond the purview of most formal state-building projects sponsored by the World
Bank and UNDP. To the extent that it has been documented, it has tended to be seen through the more
limited lens of grass-roots peace-building. Finally, it transpires in remote, dangerous border areas where
few journalists and analysts spend time. .
60 Letitia Lawson and Donald Rothchild, "Sovereignty Reconsidered," Current History (May 2005), p. 228.

- 54-
8.1 The Mediated State as Concept and Explanatory Theory

The concept of the mediated state is rooted in the study of pre-modern and early-
modern state formation in Europe, where ambitious monarchs with limited power were
forced to manipulate, maneuver, and make deals with local rivals to their authority.
Those rivals, notes Swen Voekel, "often mediated state authority, and did so both as
over-powerful purveyors of royal prerogatives, as 'private' citizens exercising '~ublic'
jurisdiction, or as members of extra-national bodies like the Catholic Church." I This
produced situations in early modern Europe that sound oddly familiar in contemporary
Somalia - France, for instance, is described as "a nation characterized by parcellized and
overlapging jurisdictions, multiple legal codes, and a plethora of internal tariffs and
taxes." As such, the mediated state is considered by historians as a major obstacle to
state-building, a syndrome to be overcome, usually by superior force of arms. Charles
Tilley observes that European state formation "consisted of the states' abridging,
destroying, or absorbing rights previously lodged in other units.,,63 Whether the mediated
states is in an obstacle to state-building or a possible route to state-building is an
intriguing question in contemporary zones of weak and collapsed state authority.

Though extreme caution must be taken in drawing historical parallels between

state formation in early modern Europe and contemporary Africa, the general concept of
a mediated state appears to have some utility in explaining actual politics on the ground
from Kenya to the DRC to Mozambique. 64 In these locations, central governments with
very limited power rely on a diverse range of local authorities to execute core functions
of government and "mediate" relations between local communities and the state. It is
usually an unspoken strategy, not enshrined in a national policy as may soon be the case
in Kenya. 65

Until recently, there were compelling reasons why weak African states did not
attempt some variant on a mediated state. One reason was ideological- the project of the
modern nation-state in independent Africa could not accept less than the full range of
sovereignty and monopoly on the legitimate use of violence within it borders that the
West and East bloc enjoyed. Modernization theories which infused thinking about
political development were predicated on the "passing of traditional society" and the
paramount authority of the state; to the extent that customary law was employed in

61 Swen Voekel, "Upon the Suddaine View: State, Civil Society, and Surveillance in Early Modem
England," Early Modern Literary Studies 4, 2 (September 1999), pp. 1-27. Accessed at
http://purl.oclu.org/emls/04-2/voekupon .htm
62Ibid.
63 Charles Tilly, "Reflection on the History of State Making" in The Formation ofNation States in Western
Europe ed. Charles Tilly (Princeton: Princeton University Press, 1975), p. 35. Importantly, the use of
coercion by emerging states in Europe had the essential secondary effect of consolidating actual
administration (as opposed to mere warlordism) over the citizenry, in order better to tax it to finance the
war efforts. That salutary political effect of war-waging is absent in cases like Somalia, where armed
conflict is financed largely by a combination of international funding and pillaging.
64 To be precise, the mediated state is concentrated in but not limited to Africa - one of the most routinized

forms of the mediated state is in Yemen, where the government must work through tribal leaders for access
to and control of the entire rural northern portion of the country.
65 "Conflict Management Policy Framework Unveiled," Peace Bulletin no 5 (Sept. 2004), pp. 1,3.

- 55 -
remote areas, it was tolerated but not sanctioned by the state, and at any rate viewed as a
dying system to be replaced in due course by the expanding modem state, not as a
building block for state-building. The other reason, as described by Jeffrey Herbst, is that
African states - unlike early modem European states - inherited fixed boundaries finnly
protected by international law from encroachment by more ambitious or effective
neighbors or break-away secessionists. 66 In an earlier period, failure to extend and
maintain authority in frontier areas risked loss of that territory, leading rulers to devote
considerable treasure and manpower to protecting (and expanding) borders. Today,
African state authorities have in some respects pursued a rational strategy by allowing
frontier zones to go ungoverned, especially if the frontier has little economic value, the
cost of establishing rule of law is higher than whatever revenue can be earned from the
area, the state faces a serious shortage of money and capacity, and negative spillover of
anarchy in the frontier does not unduly ilnpact the core areas of the state. The result has
been enonnous tracts of territory in the hinterlands of many African states that are in a
condition of de facto state collapse. The retreat of the state from its frontier areas and the
anned anarchy which ensued in Kenya in the early 1990s was particularly shocking only
because it occurred in a "middle-income" country with greater governmental capacities
than in most of the rest of the continent.

This explanation for the governance vacuum in African border areas emphasizes
the role of state indifference or disinterest in border areas, though lack of state capacity is
also a factor. But when that "frontier governance" calculus changes -- when state
authorities develop an interest in asserting or reasserting security and rule of law in their
hinterland, but lack the capacity - conditions improve for an alternative, "mediated state"
approach.

In the case of Kenya, several factors increased government interest in expanding

law of rule in its border areas. These included the enonnous costs of spillover of crime,
displacement, and anns flows into the "core" of Kenya from the running sores in its
frontier area; a sense of professional duty on the part of a small number of committed
civil servants; pressure to "do something" about the violence and lawlessness in Kenya's
borderlands from international donors, the national media, civic groups, and religious
leaders; and, in the aftennath of multiple terrorist attacks in Kenya in 1998 and 2000 and
the 9/11 attacks, a heightened security concern that terrorists were exploiting Kenya's
porous borders and lawless interior.

State authorities which are willing but unable to govern their remote hinterland
are forced to pursue a mediated state strategy, not out of sudden enlightened appreciation
for the virtues of civil society and traditional authority, but because it is their only
effective option, at least in the short-tenn. This aspect of the mediated state - that it is not
a policy preference but rather a default position for weak states seeking to promote
governance and security in its frontier areas - sets it apart from other contemporary fonns
of state "outsourcing" of governance. The key difference is that states opting to contract
out functions to non-state actors (such as corporations operating seaports or non-profits
delivering social services paid for by the state) usually do so as a matter of public policy

66 Jeffrey Herbst, States and Power in Africa (Princeton: Princeton University Press), 2000.

- 56-
choice, and ostensibly with the objective of providing the service more efficiently. This is
the choice of a state authority which "has acquired the competence to decide the limits of
its own competence. ,,67 By contrast, a mediated state strategy is the recourse of a state
authority which lacks options. It has no choice but to work through local intermediaries if
it is to have even token jurisdiction in an area within its borders.

Sub-contractors, moreover, operate within the legal framework of the state. Sub-
contracting firms and NGOs workfor the state. By contrast, local authorities in a
mediated state arrangement operate beyond the state, its legal code, and its most coveted
possession - its "monopoly on the legitimate use of violence within its territory." The
fact that government-civic partnership in northern Kenya includes implicit government
acceptance of the application of extra-judicial legal codes by non-state actors and the use
of local home-guards to employ lethal force in defense of public security is no mere
"subcontracting" exercise. It is a fundamental revision of basic precepts of sovereign
state authority; a forfeiting of a state's claim to "omni-competence" within its borders; a
new, flexible way to deal with the current "messiness" that conventional inherited
political structures have had great difficulty managing.

State interest is a necessary but not sufficient condition for mediated governance.
Whether a mediated state strategy is actually an option then depends on the presence of
reasonably authoritative local actors which the state can accept (a secessionist polity, for
instance, would not be acceptable; nor would a liberation or radical movement bent on
the overthrow of the government or entire political system). Finally, for a mediated state
strategy to succeed as a peace-building and governance strategy, the sources of local
authority must be relatively legitimate and committed to peace and good governance, not
predatory or corrupted local elites. Otherwise, the strategy produces a patch-quilt of state-
sponsored warlord fiefdoms.

In the case of northern Kenya, the very fortunate combination of revived

customary law and role of traditional clan elders and the ascendance of capable,
responsible, and dedicated civic leadership within the region's small urban professional
population has provided the ideal conditions for a mediated state to work since the mid-
1990s.

A hallmark of the mediated state as it evolved in early modem Europe was

flexibility and pragmatism. State rulers brokered deals with whatever authorities existed
in the periphery of their realm. The approach by definition defies templates and
standardization. The same characteristic obtains in northern Kenya. There, the types and
combinations of local actors in peace committees, and the types of relationships
developed between these local authorities and the state, have varied considerably from
place to place. 68

67 Wolfgang Reinhold, "Introduction: Power Elites, State Servants, Ruling Classes, and the Growth of State
Power," in Reinhold, ed., Power Elites and State Building (Oxford: Oxford University press, 1999), p. 1.
68 The range of peace committee models in Kenya is surveyed and assessed in Walker, Ibrahim, and Shurio
(2003).

- 57 -
 By viewing the current government-civic partnership to conduct core functions of
the state in Kenya's border areas through the lens of the "mediated state" model, we
accord this hybrid governance approach the significance it is due. Though its initial
purpose was to promote peace-building, the mediated state in Kenya has expanded into
core functions of the state - the judiciary, police, cross-border diplomacy. The Kenyan
state is also employing the approach to promote range-land management. 69

Because the strategy involves ceding responsibility to non-state actors, it is

difficult to cast this trend as a contribution to "state-building." But it may well be a form
of "governance-building." Since the objective of state-building projects is not to
strengthen state capacity for its own sake, but rather as a means of promoting good
governance, the possibility that the mediated state can help promote the latter by by-
passing the fonner is an interesting challenge to standard state-building interventions,
which tend to conflate reviving formal state capacity with promotion of governance.

8. 2 The Mediated State and Strategies of Peace-building and Governance-building

The mediated state model is thus a more satisfactory and comprehensive theory to
explain what is already occurring on the ground in Kenya's border areas. But is it also
useful as a state-building strategy? Is mediated governance a trend to support and
encourage, or are criticisms of the approach serious enough to warrant caution?

There is now adequate documented evidence from northern Kenya to sustain two
not entirely incompatible claims - first, that the government-civic partnership there has
unquestionably produced impressive gains in public security and conflict management,
and second, that the approach has serious shortcomings. The achievements of the
mediated state in Kenya's border areas have been documented above and need not be
repeated. As for concerns and criticisms, several stand out as especially noteworthy:

• Allowing customary law to be applied above the laws of the state by non-state
actors, in which collective responsibility is privileged over individual rights and
responsibilities and in which not all are equal before the law is a serious erosion
of civil liberties, human rights, and rule of law. It is also illegal and extra-
constitutional and cannot be sanctioned by the Kenyan government.
• These objections aside, customary law also has a limited reach and is often over
whelmed by the scale of current armed conflicts.
• The peace committees serving as umbrella groups for local actors are very
uneven in capacity, lack an independent, sustained flow of resources, and are
plagued by high turnover.
• Some peace committees are politically compromised and not neutral.
• No local authorities are in a position to cope with conflicts instigated by
powerful outsiders in Nairobi and elsewhere.

69Guyo Haro et ai, "Linkages Between Community, Environmental, and Conflict Management:
Experiences from Northern Kenya," World Development 33,2 (2005), pp.285-299.

- 58 -
 • Some border regions lack the key prerequisite of legitimate, capable local
authorities and are instead either zones of leadership vacuum or are dominated
by predatory politicians and warlords.
• The entire enterprise of constructing mediated governance in northern Kenya
has been designed to manage conflict, but lacks any real capacity to address
underlying causes of armed conflict, which require effective and sustained
central state engagement. 70

The fact that mediated governance is problematic and limited and yet the sole
source of security in weak state frontier areas makes it the "best of bad options," a
position that both its advocates and detractors can share. How then to view the long-term
prospects of mediated governance? If its shortcomings are intrinsic (legal and human
rights objections, for instance), and if better options are on the horizon, then the only
justification for support to mediated governance is as a stop-gap measure, designed to
provide short-term improvements in security and conflict prevention.

If, on the other hand, the prognosis for state expansion and consolidation in the
border areas of weak states is poor, then a different set of policy considerations emerge.
In that case the trend toward mediated governance is longer-tenn in nature, a fact which
citizens and international observers may not prefer but which is not a matter of
preference. Realistically, the focus then becomes improving the mediated state as a
governance and peace-building strategy in frontier zones - taking the rough edges off of a
governance structure that is and will remain flawed and messy. This can be done in a
number of ways. First, support to peace committees (or whatever other pennutation of
local authority emerges by district) is vital if they are to be sustainable. Training and
financial support for appropriate operational expenses are two of the most important
types of support. Recognition is also important - these are mainly voluntary efforts, by
civic figures willing to put themselves in harm's way in very difficult circumstances.
Second, external efforts can help to harmonize, at least partially, the inconsistencies
between customary, sharia, and national legal codes. 7 ! Clarification of the jurisdiction of
customary law is also essential -- for example, is a Kenyan Somali living in Nairobi
subject to xeer, the Kenyan legal code, or both?

Importantly, the mediated state as peace-building and governance-building

strategy offers opportunities to international aid agencies, not merely Faustian choices. In
the past, agencies engaged in "capacity-building" frequently found themselves working at

70 Several studies provide thoughtful critiques of the strengths and weaknesses of peace committees and the
government-civic partnership we are calling a mediated state. See Walker, Ibrahim, and Shuria (2003);
Haro et al (2005); Goldsmith (1997), and Nyunya, Joshua. "Reports by National Experts: Kenya. Conflict
Prevention, Management, and Resolution: Capacity Assessment Study for the IGAD Sub-Region." Leeds:
University of Leeds, Centre for Development Studies, October 2001).
71 A few analysts and organizations have recently produced excellent reports calling for "harmonization" of
traditional/informal and formal governance in the Somali judicial sector. See Andre Le Sage, "Stateless
Justice in Somalia: Formal and Informal Rule of Law Initiatives," (Geneva: Centre for Humanitarian
Dialogue, January 2005), and Puntland Development Research Centre, "Pastoral Justice: A Participatory
Action research Project on Harmonization of Somali Legal traditions: Customary Law, Sharia, and Secular
Law" (Garowe Somalia: PDRC, 2002).

- 59 -
cross-purposes. Projects building capacity of local civic and non-governmental
organizations was viewed by other actors has working against state-building, by diverting
human resources, money, and roles to non-state actors. This debate over whose capacity
is being built has plagued governance projects in Somalia, Kenya, and elsewhere for
years. In a mediated state, however, there is no contradiction between improving the
governance capacity of non-state authorities in border area and simultaneous efforts to
build formal state capacity. One is part of a short to medium-term governance and peace-
building strategy, the other is part of the long-term state-building project which, under the
right circumstances, may eventually displace mediated governance. Whether mediated
governance is in fact an interim strategy or part of long-term alternative form of
governance in parts of Africa and the developing world is a decision which citizens of
these states must make for themselves.

8.3 Prospects for a Mediated State in Somalia

The Kenyan experiment with mediated governance has immediate and interesting
implications for state-building in Somalia.

Two political trends have clearly emerged in Somalia over the past fifteen years
of state collapse. The first is the abject failure of repeated external efforts to revive a
conventional central government in the country via a "top-down" process of power-
sharing among Somalia's quarreling political elites. Though this track record of failed
state-building can partially be attributed to myopic Somali leadership and uninspired
external diplomacy, it is also apparent efforts to revive a central government in Somalia
face important structural obstacles as well. One of the main obstacles is the extremely
weak resource base a Somali state can draw on, a constraint which makes the revival of a
large, conventional state claiming "omni-competence" across a wide range of policy
areas a pipe-dream. Though insistence on such a state structure is understandable from
the perspective of Somali leaders desperate to revive an expansive patronage system and
build a capacity for repression - the only means of securing political control they have
ever known - it is simply untenable for the near future.

The second trend is the rise of local, informal polities which have, in fits and
starts, increasingly provided many Somali communities with variable levels of
governance, public security, and even social services. The problem with this mosaic of
informal polities is that it does not add up to anything resembling a conventional state,
and at this point in time does not appear capable of serving as the building blocks for an
organically-developed state. Local polities in Somalia have remained eminently local.
And even the most impressive, functional examples of sub-national and/or informal
governance in Somalia cannot perform a number of badly needed functions of an
internationally recognized sovereign state, from the issuing of passports to the securing
of loans from international financial institutions.

Up to now, Somalia's informal systems of governance have generally been

accorded little to no role in external efforts to revive a conventional state. The accepted,
unspoken wisdom has been that these local systems of governance are of little

- 60-
significance, mere variations on a broader theme anarchy. They are viewed as short-term
coping mechanisms to be replaced by formal state authority once the elusive state-
building project succeeds.

An alternative approach to state-building which the Kenyan experience provides

is one which combines what is already working locally with what is essential nationally.
In fact, mediated governance in Somalia may be not so much a policy option as the only
viable route to state-building in Somalia under present circumstances. Somalia's best
hope for state revival may lie in the explicit pursuit of a state in which a central
government with very limited power relies on a diverse range of local authorities to
execute core functions of government and "mediate" relations between local communities
and the state. The nascent central state limits itself to a few essential competencies not
already provided by local, private sector, or voluntary sector actors. Central state
authorities resist the temptation to insist on sovereign controls over social and political
realms and entire communities that they cannot realistically exercise. For their part, local
mediators gain recognition from the state by effectively providing core functions of
public security or other services demanded by local communities, and earning legitimacy
as a result.

As in Kenya's border areas, precisely how a formal, "top-down" state structure

can and should co-exist with existing practices and structures of informal governance
would be a matter for Somali authorities to work out, town by town, district by district.
The result would be quite complex and, from a state-building perspective, invariably
"messy," with a wide range of parallel, overlapping, and in some cases contested political
authorities. External actors tasked with supporting state-building in Somalia would
simply not be able to import fixed state-building project templates, could not insist on
standardized judicial and other systems, and would have to learn to work with local
polities in Somalia on their own terms, rather than attempt to transform them into images
in their own likeness. That level of programmatic flexibility and local knowledge has not
been a strong suit of international aid agencies in the past.

A mediated governance approach would enable external aid agencies to support

state-building in Somalia even if the TFG collapses, in that support to local governance
bodies and systems would be understood not as support to a rival to state authority but as
a local partner which can be "plugged in" to an emerging central government if and when
it is revived.

In the Transjubba areas, local communities have paid close attention to the
emerging form of mediated governance in northern Kenya and have actively sought to
emulate it, creating local peace committees in border towns. Some of this is little more
than a fishing expedition for expected external assistance, but much of it appears to
reflect a bona fide hope that the kinds of umbrella-group efforts which have succeeded in
Wajir and Mandera can help communities on the Somali side of the border better manage
both local and cross-border conflict.

- 61 -
 External support to these initiatives must exercise care - in the complete absence
of a state, local power struggles are endemic and foreigners are easily exploited and
drawn into conflicts. Direct provision of financial support or other highly "liquid"
resources are most likely to produce conflicts. But with close knowledge of the
communities in question and due diligence, external actors could help to provide training
and other support to these promising sources of local governance. To the extent that
mediated governance in northern Kenya serves as a model and inspiration for trans-Jubba
Somalia, a new and more constructive form of "spillover" can help counterbalance the
negative spillover of war, crime, and arms flows which have defined the border area for
years.

8.4 Addressing Underlying Drivers of Conflict

Assessments of conflict in northern Kenya all point to an array of underlying

factors which make the region exceptionally prone to armed conflict. Efforts to
strengthen mediated governance have improved local capacity to reduce crime - one of
the main triggers of communal clashes - and have helped manage and mediate conflicts
once they erupt. But a more comprehensive peace building strategy requires policies
which address the most dangerous underlying drivers of conflict. This analysis points to
several possible entry points where external aid may help reshape sources of conflict into
factors promoting peace and security. As was emphasized in chapters 6 and 7, some of
the forces at play in the border area - such as commerce, urban space, and ethnicity -- can
either promote peace or conflict, depending on circumstances. It follows then that aid
interventions should seek to work with the Kenyan government to help shape the context
in ways which tap into the constructive aspects of these factors. Business interests in the
border area are especially amenable to a "shaping" strategy.

The single most powerful conflict driver on the Kenya side of the border has been
political - specifically, proliferation and abuse of locations by MPs as a form of political
patronage and a means of engaging in localized ethnic cleansing. That in tum has
contributed to deterioration of access to grazing areas, environmental degradation, and
deterioration of livelihoods among pastoral households that constitutes a genuine crisis in
the region. It has also created ideal conditions for hardened ethnic identities, communal
clashes, and violent crime. Aid interventions which help produce fundamental changes in
Kenyan policies on locations and rangeland access will go a long way to reversing one of
the most insidious, and preventable, sources of spiraling conflict in the region.

- 62-
 Kenya-Somalia Border Area-Selected Bibliography

Africa Watch. Divide and Rule: State-Sponsored Ethnic Violence in Kenya. New York:
Human Rights Watch, November 1993.

AMANI Forum. "A Conflict Map of Kenya: An Overview of the Conflict Situation in
Kenya: Issues, Extent, and Effect." Centre for Conflict Research, August 2002.

Anderson, David M. and Vigdis Broch-Due, eds. The Poor Are Not Us: Poverty and
Pastoralism in Eastern Africa. Oxford: James Currey Press, 1999.

Besteman, Catherine, and Lee V. Cassanelli. The Struggle for Land in Southern Somalia:
The War Behind the War. Boulder, CO: Westview Press, 1996.

Brown, Steven. "Quiet Diplomacy and Ethnic Clashes in Kenya." In Donald Rothchild,
Chandra Sriram and Karin Wermester, eds. From Promise to Practice:
Strengthening UN Capacities for the Prevention of Violent Conflict. Boulder:
Lynne Rienner, 2002

Cani, Ruggaro. Ii Guibaland. Naples: Tipografia Angelo Treni, 1921.

Dalleo, Peter. "Trade and Pastoralism: Economic Factors in the History of the Somali of
Northeastern Kenya, 1892-1948." Ph.D. diss, University of Syracuse, 1975.

Development Alternatives Inc. "Kenya-Somalia Cross Border Consultation Workshop."

Mandera, Kenya. Unpublished proceedings of the workshop, 24-26 June 2004.

Development Alternatives Inc. "Initiative Summary: Gedo (Somalia) - Mandera (Kenya)

Cross Border Conflict Mitigation." Unpublished paper, 2004.

Farah. M.1. "From Ethnic Response to Clan Identity: A Study of State Penetration of
Somali Nomadic Society of Northern Kenya." Uppsala, Sweden: University of
Uppsala, 1993.

Gebreselassie, Seyoum. "Case Studies on Tensions along the Ethiopia-Kenya-Somalia

Borders." Paper presented at the workshop "Towards an IGAD Conflict Early
Warning and Response Mechanism (CEWARN)," Addis Ababa, Ethiopia, 30-31
Augusts 2001.

Goldsmith.Paul. "The Somali Impact on Kenya, 1990-1993: The View from Outside the
Camps." In Mending Rips in the Sky: Optioonsfor Somali Communities in the 21 st
Century, pp. 461-484. Edited by Hussein Adam and Richard Ford. Lawrenceville
NJ: The Red Sea Press, 1997.

- 63 -
Goldsmith, Paul. "The Land of JUali; Travels through Kenya's Drought Stricken North."
African Studies Quarterly [on-line] 4 (3) URL:
http://web.africa.ufledu/asq/v4/v4i3a3.htm

Goldsmith, Paul. "Cattle, Khat, and Guns: Trade, Conflict, and Security on Northern
Kenya's Highland-Lowland Interface." Isiolo, Kenya: APPEAL-KENYA,
Conflict and Conflict Management in the Hom of Africa Case Study, May 1997.
http://payson.tulane.edu/conflict/Cs%20St/GOLDSFIN2.html

Helander, Bernhard. "The Hubeer in the Land of Plenty: Land, Labor, and Mobility
among a Southern Somali Clan. In The Struggle for Land in Southern Somalia:
The War Behind the War, pp. 47-72, edited by Catherine Besteman and Lee
Cassanelli. Boulder, co: Westview Press, 1996.

Ibrahim, Delma, and 1. Jenner. "Wajir Community-Based Conflict Management." Paper

presented to the USAID Conference on Conflict Resolution in the Greater Hom of
Africa, June 1997.

Janzen, Jorg. "Pastoral Economy and Seasonal Livestock Movements in the Jubba
Valley. Part C of US Agency for International Development, Jubba
Environmental and Socioeconomic Studies. Volume III: Socioeconomic Studies.
Burlington VT: Associates in Rural Development, 1989.

Kamenju, Jan, M. Singo, and F. Wairagu. "Terrorized Citizens: Profiling Small Arms
and Insecurity in the North Rift region of Kenya." Nairobi: Security Research and
Information Centre, 2003.

Khadiagala, Gilbert. "Protection and Poverty: The Experiences ofCommunity Weapons

Collection Initiatives in Northern Kenya. "Nairobi: Oxfam GB, 2003

Kloop, ..Jacqueline. "Can Moral Ethnicity Trump Political Tribalism? The Struggle for
Land and Nation in Kenya." African Studies, vol,. 61, no. 2 (2002).

Kore, Abdul Ebrahim. "Wajir North Conflict: Naturalized Displacement." Unpublished

paper, 200 1.

Kratli, S. and J. Swift. "Understanding and Managing Pastoral Conflict in Kenya: A

Literature Review." Sussex, UK: IDS, University of Sussex, 1999.

Lewis, I.M. Peoples ofthe Horn ofAfrica. London: International African Institute, 1955.

Lewis, I.M. A Modern History ofSomalia: Nation and State in the Horn ofAfrica.
Boulder: Westview, 1988.

Little, Peter. Somalia: Economy without State. Bloomington, IN: Indiana University
Press, 2004.

- 64-
Mahmoud, Hussein A. "Livestock Trading and Trader Networks in Northern Kenya and
Southern Ethiopia." Research Brief 01-06 PARIMA, "Global Livestock
Collaborative Research Support Program," December 2001.

Marchal, Roland. "A Brief Overview of the Gedo Political Economy." Chapter 1 of "The
Need for a Food Monetarization Program in Somalia." Nairobi" EC-Somalia Unit,
1996.

Marchal, Roland. "An Overall View of the Political Situation in Gedo." Unpublished
Report, 1996.

Marialn, Ayele Gebre. "Ethiopia: Report by National Experts." Study Commissioned for
IGAD, "Conflict Prevention, Management, and Resolution: Capacity Assessment
Study for the IGAD Sub-Region." Leeds, UK, October 2001.

Mburugu, E. K. and M. Hussein. "Enhancing Better Natural Resource Use to Prevent

Conflict Among Pastoralist Communities in Kenya." Nairobi: Oxfam GB, 2002.

Menkhaus, Ken. Gedo Region. United Nations Development Office for Somaha,
"Studies on Governance" Series no. 5. Nairobi: UNDOS, December 1999.

Menkhaus, Ken. Middle Jubba Region. United Nations Development Office for Somalia,
"Studies on Governance" Series no. 6. Nairobi: UNDOS, December 1999.

Menkhaus, Ken. "From Feast to Famine: Land and the State in Somalia's Lower Jubba
Valley." In The Struggle for Land in Southern Somalia: The War Behind the War,
pp. 133-154, edited by Catherine Besteman and Lee Cassanelli. Boulder, CO:
Westview Press, 1996.

Merryman, James. "The Economy of Geedo Region and the Rise of Smallholder
Irrigation." In The Struggle for Land in Southern Somalia: The War Behind the
War, pp. 73-89, edited by Catherine Besteman and Lee Cassanelli. Boulder, CO:
Westview Press, 1996.

Morarji, Maitri. "Alchemy Project Field Assessment: Arid Lands Development Focus
(ALDEF), Wajir Kenya." Boston: Tufts University, July 2004.

Muse, Mohamed Barre. "Baseline Assessment, Gedo Region Somalia, 17-29 January
2005." Mogadishu: Center for Research and Dialogue, 2005.

NGO Consortium-Mandera. "Situation Report." Unpublished paper, 4 July 2005.

"Northern Frontier is Kenya's Shame." The Daily Nation, 20 July 2005.

Nyaba, Peter A. and Peter O. Otim. "Cross-Border Pastoral Conflicts: A Case Study of

- 65 -
 the Kenya-Sudan-Uganda Border Sub-Regions." Paper presented at the
Workshop "Towards an IGAD Conflict Early Warning and Response Mechanism
(CEWARN)," Addis Ababa, Ethiopia, 30-31 Augusts 2001.

Nyunya, Joshua. "Reports by National Experts: Kenya. Conflict Prevention,

Management, and Resolution: Capacity Assessment Study for the IGAD Sub-
Region." Leeds: University of Leeds, Centre for Development Studies, October
2001).

Onyango, Carey F. "Law Could Ease Land Rows." Op-ed, East African Standard 9
(November 23, 2002).

Pkalya, Ruto et al. "Conflict in Northern Kenya: A Focus on the Internally Displaced
Conflict Victims in Northern Kenya." Nairobi: ITGD, 2003.
http://payson.tulane.edu/contlict/Cs%20St/GOLDSFIN2.htInl

Pkalya, Ruto et al. "Indigenous Democracy: Traditional Conflict Mechanisms: Pokot,

Turkana, Samburu, and Marakwet." Nairobi: ITDG, January 2004.

Richburg, Keith. "Threat of Rape Haunts Somali Refugee Women in Kenya."

International Herald Tribune, Oct. 2-3, 1993.

Save the Children Fund -UK. "The Food Economy Groups of Bardhere District, June
1996." Nairobi: SCF-UK, 1996.

Schlee, Gunther. Identities on the Move: Clanship and Pastoralism in Northern Kenya.
Nairobi: Gideon S. Were Press, 1994.

Schlee, Gunther. "Interethnic Clan Identities Among Cushitic-Speaking Pastoralists,"

Africa, 55 (1), 1985.

Terra Nuova. "Rapid Assessment of the Situation of the Livestock Sector in Trans-Jubba
Region of Somalia." Nairobi: Unpublished report, 1998.

"Thirty Killed as Clan Clashes Engulf Mandera." Peace Bulletin (ITDG Newsletter) issue
no. 6 (January 2005), pp. 1, 8.

"Thirty Thousand Displaced in Mandera Clan Violence." Peace Bulletin (ITDG

Newsletter) issue no. 7 (April 2005), pp. 1, 12.

Trocaire. "Trocaire Integrated Rehabilitation Programme, Gedo Region, Somalia,

Programme Report January 1997-February 1998. Nairobi; Trocaire, 1998.

Trocaire. "Trocaire Survey, Gedo Region." Nairobi: Trocaire, 1994.

Turton, E.R. "The Pastoral Tribes of Northern Kenya, 1800-1916." Ph.D. diss, University

- 66-
 of London, 1970.

Turton, E.R. "Bantu, Galla, and Somali Migrations in the Hom of Africa: A
Reassessment of the Juba/Tana Area." Journal ofAfrican History 16, 5 (1975),
pp.519-37.

UNDP. Somalia Human Development Report 2001. Nairobi: UNDP, 2001.

UNDP/Somali Rehabilitation Programme. "District Development Profile: EI Waq."

Mandera: SRP, May 21 1997.

UNDP/Somali Rehabilitation Programme. "District Development Profile: Belet Hawa."

Mandera: SRP, May 8 1997.

UNDP/Somali Rehabilitation Programme. "UNHCR-UNDP/SRP Mission to Dadaab

Refugee Camps, 22-26 June 1998." Nairobi: Internal report.

UN Operation in Somalia (UNOSOM). "Regional Report: Gedo." Memo, Department of

Humanitarian Affairs, Aug. 9 1994.

USAID/Regional Economic Development Services Office (REDSO). "RCMG -DAI

Cross-Border Initiatives in the Karamoja and Somali Clusters." Unpublished
memo, 2004.

Walker, Robert and Hassan G. Omar. "Pastoralists Under Pressure: The Politics of
Sedentarisation and Marginalisation in Wajir District, Northeast Kenya."
Nairobi: Oxfam-GB, July 2002.

Walker, Robert., Dekha Ibrahim A. and H. O. A. Shuria. "Oxfam GB Funded

Peacebuilding Initiatives in Arid Districts of Kenya: Lessons and Challenges."
Nairobi: Oxfam GB, 2003.

Waqo, Halakhe D. "Peacebuilding and Small Arms: Experience from Northern Kenya."
Paper presented at the workshop "UN Biennial Conference of States on Small
Arms," New York, 7-11 July 2003.

- 67 -