Pastoral society and transnational refugees: population movements in Somaliland and eastern Ethiopia 1988 - 2000

from UN High Commissioner for Refugees
Published on 31 Aug 2002

These working papers provide a means for UNHCR staff, consultants, interns and associates to publish the preliminary results of their research on refugee-related issues. The papers do not represent the official views of UNHCR. They are also available online under 'publications' at <>.

ISSN 1020-7473


The classical definition of refugee contained in the 1951 Refugee Convention was ill-suited to the majority of African refugees, who started fleeing in large numbers in the 1960s and 1970s. These refugees were by and large not the victims of state persecution, but of civil wars and the collapse of law and order. Hence the 1969 OAU Refugee Convention expanded the definition of "refugee" to include these reasons for flight.

Furthermore, the refugee-dissidents of the 1950s fled mainly as individuals or in small family groups and underwent individual refugee status determination: in-depth interviews to determine their eligibility to refugee status according to the criteria set out in the Convention. The mass refugee movements that took place in Africa made this approach impractical. As a result, refugee status was granted on a prima facie basis, that is with only a very summary interview or often simply with registration - in its most basic form just the name of the head of family and the family size.1

In the Somali context the implementation of this approach has proved problematic. Somalis are a rather homogeneous ethnic group from a cultural-linguistic point of view, stretching across at least four countries in the Horn of Africa: Somalia, Ethiopia, Kenya and Djibouti. Their main internal social differentiation is on the basis of clans and sub-clans, but even within the clan system, most clans and sub-clans are transnational.

It almost impossible to police the long porous borders in the Horn of Africa and very few people carry any ID document at all. The documents that are produced are often questionable, given the collapse of state structures in Somalia since 1991. It is therefore very difficult to distinguish between bona fide refugees from across the border and locals who are "joining the queue" in order to gain a ration card. In fact, even with lengthy individual interviews, it would be very difficult to distinguish between Ethiopian Somalis and Somali Somalis.

In 1994 there was a revealing event in which a group of Somali(land) refugees from the camps in eastern Ethiopia staged a demonstration in front of the presidential place in Hargeisa requesting the Somaliland government to lobby UNHCR and WFP to increase the food ration in the camps that had been reduced, obviously oblivious to the fact that they were not supposed to return to their country of origin if they wished to enjoy continuous refugee status, let alone ask support from their governmental authorities! Furthermore, many ethnic Somalis (either from Ethiopia or Somalia) have grown accustomed to refugee assistance since the late 1970s and developed "skills" to circumvent various registration techniques.

Yet the need to try to count refugees as accurately as possible and produce reliable statistics is an unavoidable reality in today's humanitarian world2. Much of the same argument could be made for counting IDPs and returnees, with the additional difficulty that no legal definitions of these terms have yet been formulated. This paper attempts to describe and analyse some of these issues in the context of refugee outflows and repatriation movements in the Horn of Africa and particularly between northern Somalia and eastern Ethiopia. But before we do so, we need to take a brief look at the main features of Somali society.

Somali social structure

The social structure that accompanied the prevailing pastoralist way of life, is what social anthropologists have defined as a "patrilineal segmentary opposition". In this system, lacking a hierarchical chain of authority or anything resembling the state or a judiciary, social relationships are defined in terms of kinship based on descent from a common ancestor.In Somali society, as in most pastoral societies, kinship is traced through the male line, that is patrilineal descent. The genealogies, which traditionally both Somali boys and girls have to learn by heart as part of their initiation to adulthood, define an individual's place in society as well as political relations. They are in Lewis's apt definition "what a person's address is in Europe" and - we may add - their only ID card.3

Following Lewis' definition we may identify the following levels: clan-families or federations, clans, sub-clans and lineages. The bottom-line is constituted by the group that has the collective duty to obtain compensation in disputes. If the issue at stake is a murder, the group has the duty to obtain "blood money" (diya) or to seek revenge by killing a member of the other group, often sparking endless feuds. In this system, also described as "pastoral democracy", the legitimate power of chiefs and institutions is rather weak. The clan council, known as shir, and the elders, are some of the most important conflict-resolution mechanisms, but in the end fighting strength is what makes the real difference. This social system was well adapted to a nomadic-pastoralist mode of production but ill suited to the needs of a modern nation state4. Let us now analyse the main clan federations.

The first distinction is between clan federations with a pastoral origin and the "others". The main pastoral federations are the following:

  • Dir: the main clans are the Isaq5 (the hegemonic clan in north-west Somalia/Somaliland), the Issa (hegemonic in Djibouti, but also with a large presence in Ethiopia), the Gadabursi (sandwiched between the first two) and the Bimal, the only Dir clan inhabiting southern Somalia.

  • Darod: it is the largest federation if we take into account its presence in Ethiopia and Kenya. The main clans are the Harti (particularly the Majertein in north-eastern Somalia, but also the Dulbahante and Warsangeli living in present "Somaliland"), the Ogaden (mainly in south-eastern Ethiopia) and the Marrahan (Siyad Barre's paternal clan).

  • Hawiye: living mainly in central and southern Somalia The main clans are the Habar Gidir (Aidid's clan), the Abgal (of his rival Ali Mahdi), the Murusade, the Galjel and the Hawadle.

  • Digil-Mirfle: also known as Rahanweyn they adapted to a more sedentary and agricultural life-style in the fertile regions between the rivers Juba and Wabi Shebeli in southern Somalia.

From a spatial point of view, all clan-families as well as most clans and sub-clans are oriented from the coast towards the interior and are transnational, spanning across the Somali-Ethiopian border and Somali-Kenyan. This seems to lend credence to the oral tradition reporting that clans were founded by Arabian sheikhs. It also makes it very difficult to establish the citizenship of individuals, particularly in a context where lack of proper identity documents is the norm rather than the exception.

Groups of non-nomadic origin often referred to as "minorities" include "mercantile" clans, (e.g. the Reer Hammar and the Reer Brawa) "clergy" clans (like the Shekhal and the Ashraf) or low-caste occupational groups performing culturally "impure" jobs (Gaboye/Mitgan, Tumal and Yiber6) and some clans of Bantu practising fishery and agriculture along the river banks. Most of these groups, without a warrior tradition and without a militia of its own have to rely on the "protection" of a pastoral clan, but can become their victims in case of conflict.

Conflict and progress in North West Somalia

In March 1978 Somalia conceded defeat at the hands of the Ethiopian army, supported by Cuba and the USSR, in the Ogaden War. This defeat ended the "pan-somali" dream, that is the hope to unify all ethnic Somali areas in Ethiopia, Kenya and Djibouti, with Somalia proper, which itself used to be divided between British Somaliland and Italian Somalia (Trusteeships under UN mandate in the 1950s) that became independent and merged in 1960.

Another consequence of the Ogaden War was that Siyad Barre's regime, which had by and large enjoyed popular support up to then, started running into political troubles with a coup attempt by a group of Majertein/Darod officers led by Colonel Abdillahi Yussuf, who founded the Somali Salvation Democratic Front (SSDF). As a result of this insecurity, Barre started relying more and more on clanship to defend his power, particularly on his paternal clan (Marrahan/Darod), maternal clan (Ogaden/Darod) and that of his wife (Dulbahante/Darod), reversing his earlier policy of "detribalisation" and fight against "clanism".

Meanwhile in Ethiopia, the Derg military junta led by Mengistu Hailemariam, started retaliating against the mainly Ogaden/Darod ethnic Somali populations who supported the Western Somali Liberation Front (WSLF) that launched the first guerrilla operations of the Ogaden War before being joined by the regular Somali army. The repression and counter-insurgency tactics of the Derg caused a massive influx of ethnic-Somali Ethiopian refugees towards various areas of Somalia, which aggravated the already difficult economic situation Somalia was facing. The refugees, also thanks to their clan membership, were given the choice of benefiting from international assistance as refugees, or integrating in the society as Somali nationals.

The actual number of refugees soon became a bone of contention between the Somali government, claiming that there were at least one and a half million, and the international community providing aid for 700,000 but privately estimating the correct number at less than 400,000 owing to the spontaneous repatriation of many refugees with a pastoral background7. Siyad Barre attempted to counter these difficulties by allowing some degree of liberalisation of the economy in order to attract western support while at the same time maintaining the single party system and firm control of central power. Towards this aim, he conceded to the US Navy the utilisation of the Berbera air/naval base in 1980, formerly manned by the Soviets. This move helps to explain why the US government and aid agencies financing the UN were turning a blind eye to the fact that the surplus food sent for refugees that was flooding the country was being pocketed by Barre for his entourage and the army8. He also managed to attract large-scale aid from the Italian Co-operation, controlled in Somalia by the Italian Socialist Party in financially questionable "joint ventures" for development projects that seldom worked.

Italy supported Barre almost until the bitter end, thereby attracting the hostility of most opposition groups9. By the mid 1980s, fearing attempts and plots, Barre closed himself even more in his clan fortress which showed the first cracks in 1988 in North West Somalia when an Isaq-led secessionist group, the Somali National Movement (SNM), attacked some urban centres of former British Somaliland.

The Isaq (the hegemonic clan in former British Somaliland) inhabit the central regions of Waqooy Galbeed and Togdheer of NW Somalia. Among them, the Habar Awal, and in particular the Saad Mussa sub-clan, are the most numerous and sedentarised and have partially adopted agriculture in the western district of Gabiley. On the other hand the eastern Isaq clans such as the Habar Yonis and Idagalle/Garhajis and the Habar Ja'lo and their Dulbahante and Warsangeli/Darod neighbours to the east have retained pastoralism as the main mode of production. In the western region of Awdal the Gadabursi were the pioneers of agriculture at the end of the last century while their Issa neighbours, living in the coastal areas neighbouring Djibouti on the other hand have maintained a more pastoralist life-style. Table 1 summarises the main clans in Somaliland:

Table 1: Somaliland's clans by region

Clan Main sub-clan(s) Region(s) Main districts
Isaq Habar Awal Waqooy Galbeed Gabiley, Hargeisa, Berbera
Isaq Garhajis W.Galbeed, Togdheer, Sanaag Hargeisa, Salhaley, Sheikh, Burao, Erigavo
Isaq Arab W. Galbeed Hargeisa, Balli Gubadley
Isaq Habar Ja'lo Togdheer, Sanaag Burao, Erigavo
Isaq Tol Ja'lo W. Galbeed Gabiley
Gadab-ursi All Awadal Borama, Baki, part. Gabiley, Zeila, Lughaya
Issa Mamasan, Khodahgob Awdal Zeila, Lughaya
Harti/Darod Dulbahante Sool, Sanaag Las Anod, Erigavo.
Harti/Darod Warsangeli Sool Erigavo, Las Korey

The relative political and numerical strength of the various clans is reflected by their number of seats in the two legislative institutions of independent "Somaliland" (see further below), the Lower House and the House of Elders (Guurti), each with 82 seats, as illustrated by Table 2.

The SNM was founded in April 1981 by a group of Isaq dissidents from Britain and the Arab states, who met in London. The aim of the movement was the independence of former British Somaliland that had united with the rest of Somalia (under Italian Trusteeship) on 1 July 1960. The reasons behind this stance included the perceived marginalisation and oppression of the northern Isaq vis-a-vis Mogadishu and the Darod in power and the dictatorial and repressive policies of General Siyad Barre who took power in a bloodless coup d'etat in October 1969. The movement was financed by Isaq businessmen in Arab states and established its operating bases in Ethiopia.

However, at the beginning the struggle received little popular support as the Isaq were traditionally more interested in business than politics; at least until 1983 when Barre prohibited the commerce and plantations of "chat" (a mildly stimulant leaf, widely chewed throughout the Horn), to the annoyance of the local population and Isaq businessmen. But the most important factor was an event that took place in northern Ethiopia. In March 1988 the Eritrean Peoples Liberation Front (EPLF) dealt a crushing blow to the Ethiopian army in the battle of Afabet. As a result Mengistu rushed to sign a peace agreement with Siyad Barre in order to secure the southern front and be able to concentrate on his problems with the Eritreans in the north.

Table 2: Distribution of seats in the Lower House and House of Elders by clan in 1999

Total no. of seats
Habar Awal/Isaq
Tol Ja'lo/Isaq
Habar Ja'lo/Isaq

Total Isaq: 55.2%

In April news broke out that the agreement included a termination of hostilities and the end of sanctuaries for armed rebel groups in their respective territories. Deprived of its rear bases, in May 1988 the SNM launched a pre-emptive strike and was able to seize Hargeisa and Burao, the main urban centres in the North West. Given that the army was unable to recapture the cities with the infantry, it employed artillery and air bombardments.

By the beginning of June the two cities and many other urban centres in the regions of Waqooy Galbeed and Togdheer, were almost raised to the ground with an estimated 30,000 victims, according to Africa Watch10. Hundreds of thousands of refugees were given asylum in Ethiopia and the persecution against the Isaqs continued to such an extent, that UNHCR decided to grant prima facie refugee status to all the Isaqs. The rebellion subsided until the final overthrow of Siyad Barre in 1991.

In January 1991, as the southern Hawiye-based United Somali Congress (USC) finally chased Siyad Barre's forces from Mogadishu (but at the same time started a new chapter of a seemingly endless civil war) the SNM liberated Hargeisa. In February its troops entered the Gadabursi dominated region of Awdal, a clan that had given low-key support to Barre's regime and clashed with Gadabursi militia, but then withdrew.

On 18 May 1991 Abdirahman Tur, chairman of the SNM, proclaimed the independence of "Somaliland", the former British Protectorate, while other Somali leaders were meeting in Djibouti. From a legal point of view, Somaliland's argument for independence was based on the fact that they had been independent for four days between 26 and 30 June 1960 (the day of unification) and that they were simply reverting to the old colonial boundaries. The self-proclaimed independence and freedom from "southern" domination did not however bring international recognition, nor instant peace and stability to "Somaliland" and factional fighting, clashes and occasional looting of humanitarian assistance, marked the period from the end of 1991 to the beginning of 1993, although on a smaller scale than in the south.

However the situation started improving in February 1993, when clan elders elected Mohammed Haji Ibrahim Egal as the new president instead of Tur in a reconciliation conference in Borama. Egal, the veteran politician of the British Protectorate and united Somalia, managed to achieve a new system of power sharing based not on nationalism, but on the traditional clan balancing. The SNM relinquished power peacefully (one of the rare cases in Africa of a successful liberation movement to do so), although it remained an important political, semi-tolerated opposition force. At this time that the UNHCR Office in Hargeisa started becoming operational again and other humanitarian agencies that pulled out in 1992 started moving back in.

Unfortunately the fragile peace did not last long and resentment was brewing among the Garhajis sub-clans of the Isaq (Idagalle and Habar Yonis) who felt deprived not only of political power, but also of economic opportunities by the rival Habar Awalsub-clan. As some refugees in the Aware camps conveyed to me in December 1994, the Garhajis thought that they were not adequately represented in the Borama conference and subsequently in Somaliland's political institutions, but there were also some other motives as the Garhajis feared that the Habar Awal were trying to monopolise the economy.

The casus belli became the control of the Hargeisa airport that was in the hands of Idagalle clan militia charging illegal fees and harassing passengers, particularly expatriates. The Idagalle deemed that since the governmental-backed Habar Awal were already controlling tax revenues from Berbera, they had the right to do the same in the airport that is located in their deghan (clan territory). The government - on the other hand - wanted to control the airport not only to secure revenues, but also to show to outsiders that it was in control of the situation.

The show-down came on 15 November when, after governmental troops managed to occupy the airport, Idagalle militia retaliated by shelling the chat market and other targets. The two sides exchanged several rounds of artillery fire across the dry river-bed that divides the city until the end of December, causing hundreds of casualties. This clash sent some 90,000 refugees across the border to the Aware camps, while UNHCR was preparing for voluntary repatriation. UNHCR Hargeisa had to temporarily evacuate to Borama. The conflict spread also to Burao, with he Habar Ja'lo taking sides with the governmental Habar Awal against the Habar Yonis/Garhajis, the common opponent.

At this juncture it appeared that the Mogadishu cancer had reached Hargeisa. In Burao too, the Habar Ja'lo preferred to set up their own shanty town/displaced camp in Yarowe some 10 km. to the east of Burao, than remain with the Habar Yonis in the contested city. Instability and clashes, particularly in Burao, continued until February 1997 when a new reconciliation conference and elections confirmed Egal as President who nominated a new cabinet and redistributed the balance of power.

The House of Elders (Guurti) and the Chamber of Deputies became the main fora of governance and of settlement of disputes. A Supreme Court was also established and a forward-looking constitution (including the possibility of impeaching the president) was adopted. Peace and stability started spreading in the western regions of the country, where self-help reconstruction activities and private investment grew by the day. In the eastern regions security also improved on a whole, although the situation was more fragile and the government had a more tenuous grip on power.

From an economic point of view, livestock exports was the main source of revenues estimated at US $155 million in 1996 and 176 in 1997. The revenues were used by the government to pay civil servants and security personnel thereby keeping them away from militia and banditry. However, in January 1998 Saudi Arabia declared a ban on livestock imports from the Horn because of a few cases of Rift Valley Fever, causing economic hardship. The ban was lifted in May 1999 giving rise to a short economic boom, but was re-imposed in September 2000 for the same reason. After livestock, the second source of economic revenue was remittances from the Somaliland diaspora estimated at US $93 million in 199711.

Third in place were agriculture and trade. Somalilanders, and in particular Isaq, are very skilled and dynamic businessmen also thanks to their connections with Dubai and the virtual absence of taxation. In Hargeisa it is possible to buy PCs at Dubai wholesale prices. Telecommunication companies also flourished. But there are also other manifestations of a vibrant local civil society. Somaliland is a rare example in Africa of a "country" with a relative degree of freedom of press and the main daily Jamhuriya, close to the radical SNM, often runs stories critical of the government. A retired WHO official almost completed a fully equipped maternity hospital through her own fund raising efforts. A group of "returnees" from north America set up a basketball association that could be joined by all players irrespectively of clan origin.

Urbanisation and its related way of life grew and posed a further challenge to pastoralism. It is estimated that in the late 1990s the population of Hargeisa reached 250,000 inhabitants, while in the mid 1950s, according to Lewis, its population ranged between 30,000 and 40,000. Regarding Somaliland's overall population figures, the estimation by region is the following12, to which we have to add between 300,000 and 500,000 Somalilanders living elsewhere in the Horn, in the Middle East, North America and the UK.

Table 3: Estimate of Somaliland's population by region

Waqooy Galbeed

What is the secret for Somaliland's relative success in brining a minimum of peace and stability? In our opinion we can consider three factors. First, the British colonial tradition of "indirect rule", with its "minimalist" approach emphasising self-governance. Second, given its politically and geographically peripheral position vis a vis Mogadishu, it was left in a state of "benign neglect" without too many external interference. Finally, Somaliland could rely on a class of very skilled businessmen and on the political astuteness of president Egal. His recent death (3 May 2002) will put to test the solidity of Somaliland's institutions and the maturity of its people.


1 This paper is written in a personal capacity and does not represent the views of UNHCR. It draws on the author's experience in Djibouti (1992) and Jijiga, eastern Ethiopia (1993-95) as Protection Officer and in Hargeisa, North West Somalia (1997-99) as Repatriation Officer. The term "Somaliland" will be used interchangeably with "North West Somalia" and that of "Puntland" with North East Somalia", without implying a position on the issue of recognition. Information about clan genealogies and boundaries are only indicative.

2 See Jeff Crisp, "Who has counted the refugees? UNHCR and the politics of refugee numbers", New Issues in Refugee Research, UNHCR, Geneva, 1999.

3 I.M. Lewis A Pastoral Democracy, International African Institute, New York, 1961, reprint 1982, p. 2

4 G. Prunier, "Somalia: civil war, intervention and withdrawal" Refugee Survey Quarterly 1996.

5 While I.M. Lewis (A Pastoral Democracy, op. cit. and Blood and Bone; the Call of Kinship in Somali Society, Red Sea Press, 1994) treats the Isaq as a clan-family at the level of Darod or Hawiye, most Somalis, including Isaq I interviewed in Hargeisa, agree that they are genealogically part of Dir and that sheikh Isaq was a brother of Issa and (probably) of Gadabursi. However, in the Arta (Djibouti) 2000 conference, they maintained to be a clan-family directly related to the Prophet's line, claiming the same number of seats as the Darod or the Hawiye, instead of having to share them with other Dir. This would seem to confirm Lewis' contention that while Somali genealogies - being rather univocal among different informers - are less "fictitious" than those in other African societies, they acquire more of a mythical character at the top of the genealogical tree where direct descent from an Arab sheikh confers politico-religious legitimacy and prestige.

6 See further paragraph 4.3 below on these low-caste clans.

7 See M. Maren, The Road to Hell: the Ravaging Effects of Foreign Aid and International Charity, Free Press, New York,1997, and UNHCR, The State of the World's Refugees 2000: Fifty Years of Humanitarian Action, Oxford University Press, 200, pp. 106-110.

8 Maren, The Road to Hell, op. cit.

9 See A. Del Boca , Una Sconfitta dell'Intelligenza: Italia e Somalia, Laterza, Bari, 1993..

10 Africa Watch, Somalia: a Government at War with its own People, Human Rights Watch, New York, 1990

11 Somaliland Republic, Somaliland's Two Years Development Plan, Hargeisa, 1998.

12 The figures provided here are adapted from WHO/UNICEF estimates of 1,100,000 – 1,200,000 in 1998, while UNDP puts the population of Somalia as whole at 6.38 million (2001 Human Development Report - Somalia).

Guido Ambroso
UNHCR Brussels
E-mail :

Evaluation and Policy Analysis Unit
United Nations High Commissioner for Refugees
CP 2500,1211 Geneva 2

Web Site:

(pdf* format)