Introduction

There are any number of “crises” in the world today that seem to have established themselves as permanent features of the international landscape. Somalia ’s is one such phenomenon. At the time of writing, what is generally accoun ted the fourteenth attempt at “rebuilding the Somali state”, an enterprise in itself suggestive of diplomatic hubris, appears again to risk frustration at the hands of men who wield the power of violent veto. Though theirs may not be the final say in matters, these warlords remind us of the fragility of agreements struck by peacemakers who lack the will or the means forcefully to defend the peace.

The current, intermittent round of negotiations has been convened under the auspices of the Intergovernmental Authority on Development (IGAD). This organisation’s task has been a thankless one, hedged around with all manner of imposed limitations, not least of which has been the unpromising material composing the various Somali faction leaders, for whom the broader interests of Somalia and its people appear to rank low among their priorities.

In this regard we have to bear in mind that the role IGAD has assumed is that of facilitating, not of driving, a negotiating process in which the decisions and actions of the Somalis themselves will be the principal determinant of relative success or failure. That protrac ted and repea ted efforts to achieve a peaceful settlement have met with so little success bears witness to the extreme complexity and fragility of these peace negotiations. Agreements have been made before, only to be viola ted almost immediately. Either that, or elements outside the negotiations have engaged in a spoiling role to assert or improve their bargaining position in the next round.

Another matter to be considered is that the three IGAD states principally involved in this round of negotiations — Kenya , Ethiopia and Djibouti — have internal and regional interests to consider. All have sizeable ethnic Somali populations of their own, either indigenous or refugee, and all have a role in the diplomatic framework of the Horn and the surrounding region. This makes a neutral attitude to the reconstruction of Somalia as a state very difficult to sustain, and makes these states an easy target for those who would discredit them on grounds of alleged bias.

Finally, there is the issue of funding. In this respect IGAD is almost totally dependent on foreign donors, particularly the European Union. Foreign donors may pledge funds but are reluctant to disburse them without some hope of a return on their investment. As we shall see, the round of negotiations that began in Eldoret in 2002 proved more expensive than had been anticipa ted . It remains to be seen how long foreign governments will be willing to foot the bill for an enterprise in which so many of the Somali participants appear to be pursuing agendas of personal aggrandisement at the expense of the common good. Neither should it be imagined that the international community is totally disinteres ted when it comes to Somalia ’s future. There are business interests that are keen to preserve or increase their stakes in whatever profitable enterprises may be identified, not merely in Somalia but in the self-proclaimed state of Somaliland ; Italy and the People’s Republic of China come immediately to mind, though other business interests in the Arabian peninsula have also formed useful alliances in greater Somalia .

Colonial interruption and partial unity

Though it is not the intention to provide anything like an exhaustive account of modern Somali history, no analysis of the current situation would make any sense without touching upon events in the region since the advent of colonial rule.

During the 19th and early 20th centuries, the Somali nation became divided under the rule of four imperial powers: Britain , France , Italy and Ethiopia . This remains an important consideration today, not least because there are substantial Somali populations living outside the boundaries of the, now notional, state of Somalia : in Djibouti , in Kenya and in Ethiopia . At one time or another, this has crea ted considerable regional and domestic problems for all four states, and it remains an important consideration to this day in the approaches and policies adop ted by the latter three states towards their “collapsed” neighbour.

British Somaliland became independent on 26 June 1960 . The Somali Protectorate (previously Italian Somaliland ) followed suit five days later, and the two territories amalgama ted as the Republic of Somalia , in accordance with agreements struck between the two leaderships before independence. This merging of two ex-colonial states was not without its problems. The legal, bureaucratic and educational systems of the British and Italian colonies were made compatible only with difficulty. More importantly the merger impac ted heavily on the political status of clans and lineages that formed the basis of Somali politics. To take one example, which was to be very significant in later years: the Isaaq clan family had domina ted the politics of British Somaliland against the opposition of the Dir and Darod, but union with Somalia dilu ted Isaaq influence and allowed the Dir and Darod to form potent alliances with their fellow clansmen in Somalia and others. It was scarcely surprising, then, that popular enthusiasm for the unity project waned rapidly among the Isaaq. In the south, too, there was unhappiness among the Rahanweyn of the inter-riverine region, who felt that they had lost out in influence to the Hawiye and Darod.

That a sense of Somali nationalism was integral to the decision to join the two states was left in no doubt in the constitution, which called for the union of all Somali territories, by legal and peaceful means. This aspiration was quickly frustra ted by Djibouti, which vo ted to remain within the French fold, and by Ethiopia and Kenya, neither of whom had any intention of giving away territory, albeit that it was occupied by people they persis ted in treating as second-class citizens.

The decision to pursue a greater Somali unification by peaceful and legal means, however, was not Mogadishu ’s alone to take, and by 1964 the Ogadeni resistance to Addis Ababa ’s increasingly oppressive rule had led to a number of border clashes. In Kenya ’s Northern Frontier District the local Somali population waged a desultory guerrilla campaign with clandestine assistance from across the border.

By the mid-1960s politics in the Republic of Somalia was becoming affec ted by the growing fragmentation of clans and clan alliances, which led to a massive proliferation of political parties, and unstable governments. In 1967 a northerner, Muhammad Haji Ibrahim Egal, was appoin ted prime minister by the southern president, Abd ar-Razaq Husseyn. Egal’s pursuit of a more conciliatory policy on unification cost him a great deal of popular support. More damaging, however, was the perception that a narrow political class was exploiting the national political system for personal benefit. Attempts to suppress discontent by adopting a more authoritarian style of rule left the leadership even more dangerously isola ted . On 15 October 1969 the president was assassina ted by one of his guards, and less than a week later the army seized power.

Siyad Barre and the assault on “tribalism”

When the new ruling Supreme Revolutionary Council was announced on 1 November 1969 it was headed by the chief of the army, General Muhammad Siyad Barre. The following year Siyad Barre announced that the country was embracing “Scientific Socialism”, a choice that probably owed as much to Somalia ’s increasing dependence on Soviet aid as to any ideological conviction. This also signalled a massive assault on the dominance of clan and lineage considerations in Somali life, which was officially deplored as tribalism. It is not easy to undo the social fabric by presidential fiat, nevertheless the regime persis ted in its attempts at social engineering, including the sedentarisation of nomadic populations, all of which further complica ted the clan map. As the years passed, Siyad Barre’s centralising tendencies became ever more noticeable and the regime passed gradually into tyranny, with its usual trappings of the cult of the hero leader.

Another method of promoting national unity and rescuing the faded popularity of the regime soon presen ted itself. On 12 September 1974 , the Ethiopian emperor was deposed by his armed forces, a development that encouraged the Ogadeni Somalis and their Oromo cousins in their hopes for autonomy. The ensuing repressive Mengistu government drove the Ogadenis and Oromo to desperation, and in 1976, with Addis distrac ted by a successful insurgency in Eritrea , Siyad Barre began to give considerable material support to the Ogadeni rebels.

The following year, however, the US administration withdrew its support for Mengistu, who now turned to the Soviet Union for assistance. Moscow promptly changed clients in the Horn, deserting Somalia and, with Cuban help, rearming, retraining and supporting Ethiopia ’s forces. Ethiopia and Kenya both denounced Somalia ’s aggression against the former, though it was not until February 1978 that Somalia officially entered the conflict. By then the tide of war had turned and within a few months an Ethiopian invasion of Somalia seemed a distinct possibility.

Resistance

Another consequence of Somalia ’s Ogaden debacle was the arrival from Ethiopia of more than a million refugees, most of whom settled in the north, which increased the sense of alienation among the Isaaq. Disquiet among the Majerteen manifes ted itself in an attemp ted coup by senior officers in April 1978. The survivors of this failure fled to Ethiopia , where they established the Somali Salvation Democratic Front (SSDF). Based in Ethiopia , and with assistance from that country’s government, the SSDF began launching guerrilla raids across the border. This immediately triggered violent reaction against the Majerteen of Mudug from Siyad Barre’s security forces.

In 1981 another resistance group sprang up among the Isaaq: the Somali National Movement (SNM). This, too, opera ted from Ethiopia and again invi ted savage reprisals against the Isaaq of northern Somalia . The Hawiye clan was the next to form an organisation, in Italy in 1987. This was the Uni ted Somali Congress (USC), which split almost immediately along clan lines led by the Abgal and the Habar Gidir. The latter’s General Muhammad Farah “Aideed” received weapons from the SNM, with which it became allied. Finally there was the Somali Patriotic Movement (SPM) led by the Ogadeni and formed in 1989. This organisation, too, coopera ted with the SNM.

It is at this point that IGADD enters the story. The Intergovernmental Authority on Drought and Development (IGADD) had been established in 1986 with a view to promoting a regional approach to the common problems of drought and desertification. Its membership consis ted of seven states: Kenya , Uganda , Ethiopia , Eritrea , Sudan , Djibouti and Somalia .

Despite the organisation’s narrow initial focus, IGADD summits could provide venues for meetings between member states to address other issues of mutual concern. One such meeting was the extraordinary IGADD summit in April 1988 at which the governments of Ethiopia and Somalia undertook to cease their support for each other ‘s Somali dissidents. This agreement was to have several unintended consequences, one of which was ultimately to lead to the effective collapse, firstly, of the Somali state, and secondly of President Siyad Barre’s regime.

Faced with the sudden withdrawal of support by their Ethiopian sponsors, in May 1988 the SNM launched an offensive from Ethiopia against Barre’s government. Barre’s riposte to the 1988 invasion was to launch an indiscriminate offensive using artillery and aircraft against all the principal Isaaq towns and villages in the north, killing thousands of civilians. These savage reprisals had the perverse effect of drawing ever larger numbers of Isaaq to the rebellion. The government, assis ted by the arrival of arms shipments from the US , also equipped many Ogadenis to fight against the Isaaq.

The fall of Siyad Barre

The bloody suppression of the Majerteen and the Isaaq led Siyad Barre’s erstwhile supporters in the donor community to desert him and by 1990 he could barely claim to control the capital, Mogadishu . In January 1991, Aideed’s USC hounded the Somali dictator out of office, by which time Somalia was already reverting to a patchwork of clan and lineage structures, which took over the security role usually monopolised by the state. Mogadishu ’s huge arsenal now fell into the hands of the victors and was augmen ted by military equipment readily available from the remains of Mengistu’s collapsed regime.

The split within the USC now widened, as its leaders, General Aideed (Hawiye/Habar Gidir) and the businessman Ali Mahdi Muhammad (Hawiye/Abgal), could not agree on how power should be shared. The latter seized the opportunity offered by Aideed’s continued pursuit of Siyad Barre’s forces to set up a government. This split the capital in two and led to a bloodbath in the ruined city that claimed some 14,000 lives.

To the south of Mogadishu unpredictable violence was visi ted upon the sedentary Digil Mirifle of the Rahanweyn clan family. Those of Siyad Barre’s Darod/Marehan who remained in Mogadishu also suffered atrociously at the hands of Hawiye militias. The southern agricultural lands were also laid waste by the Marehan who had gathered along the Juba River under the command of the deposed president’s son-in-law, General Hersi “Morgan”. The devastation of this vital agricultural and pastoral region led to a famine in which as many as 300,000 may have died.

Somaliland

The violence of Mogadishu and the riverine regions was not mirrored in the north of the country. On 18 May 1991 the leaders of what had once been British Somaliland repudia ted the 1960 Union and declared their region independent as Somaliland , with Abdarahman “Tur” as interim president for two years. Here traditional clan leaders proved essential to assist in overcoming local clan rivalries and running a series of local peace conferences. Though the self-proclaimed republic remained relatively quiet, it also had its share of violence as a result of a faulty demobilisation plan for the local militias, though hostilities were again ended by the intervention of the influential council of elders. In 1993 Somaliland held a major reconciliation conference at Borama to establish the groundwork for peace and a new form of government. Abdarahman “Tur” was replaced as president by Muhammad Haji Ibrahim Egal, prime minister of Somali before the Siyad Barre coup.

I M Lewis makes the following point on page 266 of his Modern history of the Somali, (James Currey: Oxford , 2002) and one that seems apposite to the present context.

Particularly striking was Somaliland ’s success with low-cost, local clan-based peace initiatives in contrast to the high-profile, internationally sponsored and highly unsuccessful conferences which came to dominate what was optimistically called the “peace process” in southern Somalia . Such high-profile “peace conferences” were destined to become a major local industry in southern Somali politics for over a decade.

As Lewis also remarks, the north was favoured in the sense that it was relatively free of the competing military leaders who exploi ted their clan connections for personal benefit, mobilising foreign aid to sustain their war machines.

There is neither space nor reason, in the present context, to chronicle the Uni ted Nations Operation in Somalia (UNOSOM) and Uni ted Task Force (UNITAF) interventions, save to say that local militias proved extremely adept at manipulating humanitarian intervention to their own advantage. This period also saw the proliferation of armed factions, as external resources became increasingly available. The more confrontational approach eventually adop ted by the US military in Mogadishu was also instrumental in pushing the armed faction leaders into pole position when it came to negotiations, which often yielded signed pieces of paper, but little else. Humilia ted and outmanoeuvred, the UN forces comple ted their withdrawal from Somalia in March 1995, abandoning the ground to entrepreneurs who scavenged the remains of the ruined capital to provide the wherewithal to establish new militias.

It was by this process that Osman Ali “Ato” star ted to make the transition from financier and garage workshop owner to Habar Gidir warlord. The Abgal stone merchant Muse Soodi Yalahow also began to make his move from businessman to entrepreneur of violence.

The progress of the self-proclaimed republic of Somaliland stood in marked contrast to what was happening in the south. Militias were partially demobilised, roadblocks removed, and a basic governance structure established. A police force was formed and courts resumed operations. There was a brief hiatus at the end of 1994 when clans dispu ted the control of Hargeisa airport, a conflict exacerba ted by ex-president, Abdarahman “Tur”, who now opposed independence, and his ally of the moment, Farah “Aideed”. The role of the UN in trying to destabilise Somaliland at this point remains a matter for conjecture.

Ethiopia had adop ted a generally helpful attitude towards Somaliland and hos ted a Peace Committee in 1996, which succeeded in restoring order. This meeting also gave rise to constitutional discussions and arrangements for presidential elections, which were won by Egal in early 1997, a result that precipita ted the withdrawal of the aggressors.

Since 1991 a similar restoration of regional order had been developing in the north-east, which had been libera ted by the Majerteen-led SSDF. Subsequently the SSDF coopera ted with clan elders to establish a form of local administration. As in Somaliland , the lack of formal institutions placed revived authority in the hands of the lineage elders, with all the implications this had for the salience of clanship. In March 1998 a conference of the Majerteen, held in Garowe, led to the establishment of the “Puntland state of Somalia ”. This state regarded itself not as independent, but as autonomous, largely out of concern for the numerous fellow Majerteen clansmen in Kismayu. There were also rela ted Darod clans in Somaliland , which led to an, as yet unresolved, dispute over the status of the regions of Sool and Sanaag. Colonel Abdillahi Yusuf was elec ted as Puntland’s first president, though his leadership remained dispu ted by his erstwhile colleague Muhammad Abshir.

Re-enter IGADD

It is now time to return the focus to IGADD. By 1994 the member states of the organisation had already realised that the developmental problems of the region went much further than drought. Indeed, in that year IGADD began its long involvement in attempts to resolve Sudan ’s protrac ted civil war. Almost as a logical consequence of this diplomatic initiative, an extraordinary summit of IGADD was held in Addis Ababa in April 1995 to discuss ways of revitalising the organisation and expanding its regional involvement into other, rela ted , spheres. On 21 March 1996 a second extraordinary summit was held in Nairobi , at which it was decided to reconstitute the organisation as the Intergovernmental Authority on Development (IGAD). An amended charter was signed, outlining the new tasks and alterations to the organisation’s structures.

The official launching of the new IGAD took place at a full summit held in Djibouti on 25 and 26 November 1996 . Speakers at the summit laid great emphasis on the need for peace and security as an essential prerequisite for development. Three priority areas were identified: Conflict Prevention, Management and Resolution and Humanitarian Affairs; Infrastructure (Transport and Telecommunications) Development; and Food Security and Environment. Particular mention was made of the need to reactivate peace and security initiatives in southern Sudan and Somalia . In his address to the summit, Kenya’s President Moi announced that he had recently hos ted three of the Somali faction leaders in Nairobi, Hussein “Aideed” (who had succeeded his father following the latter’s mortal wounding in combat), Ali Mahdi and Osman “Ato”, and that they had agreed to observe a cease-fire and enter into dialogue. Moi was of the opinion that the agreement reached in Nairobi could serve as the basis for a negotia ted settlement, and called on IGAD and its friends to put pressure on the warring factions to consider seriously the future of the country.

Arta

By 1998, in the wake of the failure of some thirteen international efforts to negotiate some kind of generalised peace in Somalia , academics and diplomats began to toy with new ideas, based to an extent on the more successful experiences of Somaliland and, to a lesser extent, Puntland. What became widely known as “The Building Block Approach” began from the realisation that, far from being reduced to a state of general anarchy as represen ted in the media, there were many areas of Somalia in which rudimentary administration and systems of order had been established.

In September 1999 a new peace initiative was launched by Djibouti ’s President Ismail Omar Guelleh. This followed a visit to Djibouti on 23 August by a group calling itself the Somali Peace Alliance (SPA), which comprised representatives of Puntland, the “Somali Consultative Body”, the Rahanweyn Resistance Army (RRA) and the Somali National Front (SNF). What was fairly novel in President Guelleh’s approach was that it no ted the failure of the warlords to make progress towards peace and emphasised the role that had to be played by civil society. He also urged that the warlords should not enjoy impunity for their crimes and that anyone obstructing the peace process should be the object of international sanctions.

President Guelleh’s initiative received a warm response from Somalis inside and outside the country. Even Somaliland ’s President Muhammad Ibrahim Egal indica ted his support, saying that, if successful, it could establish a political body and leadership in the south of Somalia with which he could negotiate. He subsequently qualified his position to make it clear that there would be no delegates from Somaliland present at such negotiations.

IGAD, too, endorsed Djibouti ’s proposal, first on 30 September, through its Standing Committee on Somalia , and later on 26 November at its summit in Djibouti , when it no ted the initiative’s conformity to the general approach approved by IGAD at its March 1998 summit. An IGAD ministerial meeting in Djibouti formally endorsed the initiative on 27 March 2000 .

In preparation for the talks, a Technical Consultative Symposium was established to advise the Djibouti government. This consis ted of some 60 Somalis invi ted as individuals, from inside and outside Somalia . They were joined by Mohammed Sahnoun as the representative of the UN secretary-general. Among the recommendations put forward by the Symposium was that the process should include those faction leaders genuinely commit ted to peace, but that it should provide for an enhanced role for civil society inside Somalia and in the diaspora. It also sketched the outlines of what might be expec ted to emerge from such a conference: a decentralised political dispensation which would consolidate those areas in which peace had been restored.

The Symposium also urged that a human rights commission be set up to monitor violations of the peace process, and called for the strict enforcement of the UN arms embargo on Somalia . Somewhat less realistic were its recommendations that Somalis occupying the lands and properties of others should withdraw to their areas of origin, and that any transitional national government should be prepared to call for an international force to assist in providing security. Not only were the beneficiaries of years of mayhem unlikely to yield their ill-gotten gains merely upon request, but in the wake of the catastrophic military interventions of the early 1990s, there would be few, if any, foreign nations willing to insert troops into so unpredictable a situation.

The peace conference began on 2 May 2000 in Arta, just outside the Djibouti capital. The first phase consis ted of a meeting of traditional and clan leaders, including elders from across the country. For six weeks the participants worked on clan reconciliation and on drawing up an agenda and lists of the delegates to represent the various clans. Delegates were to include political, business and religious leaders along with representatives of civil society. When President Guelleh opened the second phase of the Arta process on 15 June 2000, there were no fewer than 810 delegates: four delegations of 180 each (including 20 women) representing the major clan families (Darod, Hawiye, Rahanweyn, Dir) plus 90 delegates representing smaller groups (including ten women).

The delegates at Arta spent the next month in discussion before approving a Transitional National Charter (TNC) to provide the basis for governance over the next three years, following which elections would be held. The Charter provided for 18 autonomous regions based on the borders that exis ted at the fall of the Barre regime in 1991. Of particular interest was the reservation of 25 seats for women in the 225-seat Transitional National Assembly (TNA), which represen ted a precedent in Somali history. Twenty-four seats were also reserved for the minority clans.

Early in August 2000 delegates began to select parliamentarians on the basis of clan nominations, which raised the difficulty of deciding the number of seats to be alloca ted to each clan. In order to defuse tensions the peace conference gave Djibouti ’s President Guelleh authority to nominate a further 20 parliamentarians at his own discretion.

The TN Government of Mogadishu

On 13 August the TNA met for the first time and elec ted Abdalla Deerow Issaq as Speaker. Of the 45 nominees for president, 16 entered the poll held on 26 August and Abdikassim Salad Hassan emerged as victor, to be sworn in the following day at a ceremony in Arta attended by the presidents of Djibouti, Eritrea, Sudan and Yemen, the prime minister of Ethiopia and diplomats and officials from a number of African, European and Arab countries. Senior representatives of the OAU, the Arab League and IGAD were also present and the UN secretary-general’s special representative read out a message on his behalf.

Two days later President Abdikassim Salad Hassan asked those with arms to surrender them and promised to rehabilitate those militiamen that could not be absorbed into the new army. Over the next few months he visi ted a number of regional states and was allowed to consolidate his position internationally by taking up Somalia ’s vacant seats in the UN, Arab League, Organisation of the Islamic Conference and IGAD itself.

On 8 October 2000 , the president nomina ted Ali Khalif Galaydh as prime minister. Two weeks later the new prime minister announced a cabinet including representatives of all the major clans.

Somaliland ’s continued refusal to talk to the TNG was soon mirrored by Puntland’s position, which was to withdraw any support for the Arta process on 23 March 2000 , claiming that the Symposium delegates had been hand-picked to suit the organisers’ ends. Puntland’s Colonel Abdillahi Yusuf relen ted slightly in the face of popular pressure, but on 17 June announced the withdrawal of his delegation from the national peace conference, refusing to recognise its outcome.

A number of prominent faction leaders in Mogadishu refused to participate in the Arta process, and on 30 October six of them, including Hussein “Aideed” and Ali “Ato”, issued a statement claiming that President Hassan was pushing Mogadishu back to war. It soon became apparent that it was one thing to reduce the role of the armed factions in negotiations, and quite another to secure their compliance with agreements reached. Doubts also began to be raised about the methods used to identify the representatives of civil society for the Arta conference, and allegations were aired about the close business links that exis ted between the leaders of the TNG and the Djibouti presidency.

As the UN secretary-general’s report of 19 December 2000 was constrained to admit, the TNG faced some daunting tasks. How were they to persuade those who had rejec ted the Arta process, many of them heavily armed, to join it? In other words, how was the TNG to broaden its slender territorial base beyond that part of Mogadishu under its tenuous control, and what sort of relationship could be worked out with the territories of Somaliland and Puntland without endangering the relative peace and stability enjoyed by these two areas?

Lacking the ability to open or control the sea and airports of Mogadishu , the TNG was compelled to turn to outsiders for financial assistance. This was forthcoming from Libya and other members of the Arab League, though most of it was squandered and made little difference to the effective administration of the limi ted areas over which the TNG could claim some ephemeral control. In the event, such improvements as were achieved in restoring order were to be credi ted to the Islamic courts and police financed by local businessmen. Though this made for a modest improvement in the lives of ordinary citizens, it set off alarms in Ethiopia , already involved in sporadic incursion into central and southern Somalia in pursuit of Islamist guerrillas linked to the Oromo separatist movement. Thus, if the TNG enjoyed the support of the Arab states, its enemies could count on moral and material support from Ethiopia , which had also waged a proxy war with Eritrea on Somali soil.

IGAD in Kenya

Prime Minister Galadayh’s government was to last a little over a year before being toppled by a vote of no confidence on 28 October 2001 . This led President Hassan to initiate talks that, with the assistance of Kenya ’s President Moi, led to the panning of a National Reconciliation Conference in Kenya . The IGAD summit of January 2002 endorsed the idea and urged all haste. As Ethiopia and Djibouti were generally ranked behind conflicting factions, the onus of organising matters and setting them in train fell on Kenya, which was regarded as more neutral in internal Somali affairs.

On 15 October 2002 , the Somalia National Reconciliation Conference opened to great fanfare in the Kenyan town of Eldoret , in the presence of the presidents of Kenya , Sudan and Uganda and the prime minister of Ethiopia . IGAD’s executive secretary opened the conference by noting that for the first time all the member states of IGAD were involved and commit ted to solving the problem, and that the IGAD Frontline States ( Kenya , Ethiopia and Djibouti ) were working closely together to achieve peace in Somalia . He emphasised, however, that IGAD and its supporters in the international community could do no more than facilitate a Somali-owned process, in which the delegates would have to decide for themselves the core issues for discussion.

That proceedings would encounter difficulties was hardly in doubt, and European governments took the precaution of warning the warlords that should they fail to attend they would face a travel ban and a freezing of their assets. The head of the TNG, Abdiqassim Salad Hassan, sent in his stead his prime minister, Hassan Abshir Farah, since the conference organisers, insisting that he was merely one among many faction leaders, refused to accord him the status of a head of state.

The overall plan, as laid out before the IGAD Council of Ministers, was for a three-phase process which would begin with the 300 delegates agreeing on the outcome to be sought, identifying the key issues and concluding a cease-fire. The next phase would centre upon reconciliation. Some 75 delegates chosen by the plenary would then form technical working groups, each of which would address an aspect of the peace process: the constitution; disarmament, demobilisation and reintegration; land and property rights; conflict resolution and reconciliation; regional and international relations; and economic recovery, institution building and resource mobilisation. The third and final phase would be plenary sessions to consider and approve the reports of the technical working groups. The delegates would then address the sensitive matter of power sharing and the formation of broad-based government.

The whole process would be managed by an IGAD Technical Committee consisting of the Frontline States, Kenya, Ethiopia and Djibouti , under the former’s chairmanship. It was estima ted that the process would take between six and nine months.

Transitional Charter Mk I

It was hoped that the design of the Eldoret talks would overcome some of the obstacles previously experienced. The large plenary was designed to allow for broad participation, for there would have to be detailed agreement on the new Somalia ’s constitutional structure and thorough debate about issues concerned with reconciliation. Power-sharing would ideally come at the end of the agenda, though as experience was to show, the issue of who would get what was to prove a leitmotiv for many of the discussions and manoeuvrings in the first two phases.

Optimism was high when, after only two weeks, the delegates signed a “Declaration on Cessation of Hostilities and the Structures and Principles of the Somali National Reconciliation Process”. Among other things, this established the principle that the new state would be federal and decentralised in character. It also commit ted the signatories to combating terrorism, which reflec ted a priority concern of the international community in the light of the suspec ted presence of al-Qaeda cells and sympathisers in Somalia .

Only a few days later fighting broke out in a number of centres in Mogadishu and the south, though these were soon quelled by faction leaders at the talks. There was also protrac ted conflict between rival administrations in Puntland, where Colonel Abdillahi Yusuf had refused to stand down after electoral defeat.

A “Leaders’ Committee” of 22 members was formed of the signatories of the Declaration, to establish the rules of procedure and set up a Somali Advisory Group. In the event, the Leaders’ Committee determined to make the most of its own eleva ted status, and assumed the role of the decision-making body at the conference, leaving the numerous other delegates at something of a loss. In essence this gave the faction leaders the ability to veto the whole process should it be seen as threatening their interests. As if to make their intentions clear, the Leaders’ Committee eventually decided that power-sharing and reconciliation should be discussed in parallel.

By November 2002 the second phase of the talks, involving the technical working groups, had run into difficulties, ostensibly over the arrival of much larger groups of delegates than had been planned for, which led to arguments about the allocation of places. In addition, there were Somalis who were not affilia ted to the major factions or movements, demanding that their voices be heard. Ethiopia and Djibouti moun ted furious defences of their own clients’ presence and no agreement could be reached on the representation to be accorded to the different clan families.

Eventually a compromise was reached, to be known as the “4.5 Formula”, whereby each of the principal clan families — Hawiye, Darod, Rahanweyn and Dir — provided 84 delegates and the minority clans 42 between them. Actually reducing the number of attendees to this total of 378 was quite another matter.

The decision to take this route was not without its critics, who argued that it legitimised faction leaders who lacked a significant civilian following and made the Leaders’ Committee over-powerful. The issue of power-sharing was obviously already prominent in everyone’s calculations, as they perceived that the allocation of seats in the second phase might be reflec ted in the final dispensation. On the other hand, the clan-based formula did make it far more difficult for Ethiopia and Djibouti to manipulate their clients and proxies.

It was 15 December 2002 before the second phase could get under way, the Leaders’ Committee having decided to reduce the number of delegates in this phase to 300, also in accordance with the 4.5 Formula. An attempt by the facilitator to add another 100 delegates to represent civil society was rejec ted on the grounds that civil society was already represen ted in the clan allocations. Eventually 16 extra civil society delegates were admit ted . The Committee also insis ted that the same formula be used in establishing the transitional parliament. Each major clan family would thus receive 100 seats in a parliament of 450 members (subsequently reduced to 351). This gave the larger clans a 90% stake in parliament as opposed to the 70% they had held in the TNA.

By now the costs of the Eldoret conference were out of hand, and the decision was taken to reduce these by moving the talks to a government facility at Mbagathi, on the outskirts of Nairobi .

By March the technical working committees had produced their papers, though whether these represen ted a broad consensus seems open to doubt. Special difficulties were apparent in the case of the committee, or committees, working on the constitutional charter. They produced two drafts, one advocating a more centralised form of government (which was especially favoured by the incumbents in Mogadishu ), and one a loose federation based on the principles of clan-based protectionism. Both drafts included Somaliland within the boundaries of the new state, a piece of wishful thinking to all but the “politically correct”.

The committee on land and property faced an unenviable task in trying to unravel the complexities of various types of ownership, which had been further complica ted by unpreceden ted population movements in the past decade or so. They made no findings on substantive issues, but explored possible methods of addressing the problem. They also sugges ted that all militias that had seized areas by force should withdraw prior to arbitration.

The committee on economic recovery had very little reliable information on which to work. Nor, in the absence of substantive recommendations from its counterpart dealing with land and property, was there much basis on which to design revenue systems. The annual budget proposed by the committee seemed, at $1,132,158,605 for the first two years, extremely optimistic, to say the least.

On demobilisation and disarmament the conclusion of the relevant committee was that international assistance would be required to deal with the

100,000 or so men under arms. Given the fluidity of a situation in which so many militias are under no permanent or certain control, and where violent crime had become a livelihood for many young men, this was a massive task.

The plenary for the second phase began on 14 May 2003 , and several prominent faction leaders who had been absent for the previous two months made their reappearance at Mbagathi. By mid-June the delegates had approved five of the technical committees’ reports. Disagreements arose over the transitional charter, finally presen ted on 5 July, however, with the president of the TNA repudiating his representatives’ signature of an agreement that he said would lead to the dismemberment of Somalia .

Other Mogadishu faction leaders signalled similar unhappiness when the negotiations resumed at the end of the month, and demanded that the talks be shif ted to Somalia , so that Somalis themselves could take ownership of the process. They also claimed that a loose federal charter was of benefit to Ethiopia , which allegedly wan ted Somalia to remain weak and disuni ted . Failing to get their way, the malcontents began to tout for the support of Arab states. IGAD’s compromise was to add Uganda , Sudan and Eritrea to the mediation committee. Nevertheless, the portents were not favourable when the conference resumed after Ramadan, as the Mogadishu factions and their allies indica ted their lack of interest in compromise.

Transitional Charter Mk II

Accordingly, by the end of 2003 there were few observers who held out much hope of a successful outcome to the diplomatic process. In desperation the organisers decided to hold a retreat in Mombasa at which a core of delegates would attempt to iron out the essential differences, failing which the international donors made it clear that they would reconsider their financial commitments in the face of the Somali leadership’s lack of seriousness.

Several delays and petty objections followed, but eventually the meeting was convened in Nairobi on 9 January 2004 . To the great surprise of virtually everyone concerned, by 29 January consensus had been reached on the shape of a