Review of UNHCR's Kenya-Somalia Cross-border Operation

EVAL/CROS/14

SCOPE OF THE REVIEW

This review analyzes effectiveness and impact of the cross-border strategy followed by UNHCR in response to the 1992 influx of Somali refugees into Kenya. It also seeks to draw planning and management lessons that could be applied to future cross-border operations. Finally, the review makes suggestions regarding the orientation the programme could take in the near future.

In carrying out the evaluation key documents were reviewed and in-depth discussions were held with UNHCR staff in headquarters and the field. Extensive discussions were also held with officials of other United Nations agencies, donors, the governments, and staff of non-governmental organizations both international and Somali. Information covering essentially all activities was collected and analysed during the review, but in the interest of brevity, only the information required to facilitate an analysis of major issues and support conclusions is provided.

BACKGROUND TO THE OPERATION

(1). The Cross-Border Operation from Kenya into Somalia was arguably one of the more innovative programmes recently undertaken by UNHCR. It has translated into practice a new approach to humanitarian assistance founded on prevention and the non-differentiation of mandates. Furthermore, the operation was a dynamic and complex intervention that addressed simultaneously various aspects of refugee protection and assistance in both the country of refuge and the country of origin. It is also noteworthy that all three elements of UNHCR's global strategy, prevention, protection and solutions, have been dealt with in the Cross-Border Operation.

(2). The central government of Somalia was overthrown by rebel factions in January 1991, following years of a civil war that culminated with the fall of Mogadishu. What began as a general uprising against an unpopular dictatorship, then evolved into violent anarchy. Instead of forming a governing coalition, the many Somali liberation movements turned their arms against each other in what became an endless struggle for the control of land and resources.

(3). About three quarters of the population of Somalia have been affected by the conflict and hundreds of thousands have died. Famine and persisting insecurity led to the displacement of millions of people inside Somalia as well as to refugee flight neighbouring countries.

(4). The deterioration of the situation in Somalia and its effects on the region soon became causes of serious concern for the international community. By mid-1991, the United Nations launched its consolidated appeals for the Horn of Africa and Somalia. The funds raised for relief activities were mostly disbursed to the ICRC and international NGOs operating in Somalia. Later in the year, assessment missions were despatched to the region, and their findings consolidated in reports prepared by the Special Emergency Programme for the Horn of Africa (SEPHA) Unit assistance DHA. UNHCR activities for the relief of Somali refugees in Kenya, and later the Cross-Border Operation from Kenya into Somalia, were included in the SEPHA appeals.

(5). The developments in Somalia began affecting UNHCR's programme in Kenya at the end of 1991. Somali refugees were crossing the southern part of the border to the Liboi area or sailed boats to the Kenyan coast.

(6). In early 1992, fighting between major factions spread out into the Gedo region. Members of the deposed President's clan who had sought refuge in Gedo, and scores of local residents fled to other parts of the country or crossed the border into the Northeast Province of Kenya. Most of the refugees who were part of this second influx settled in sites around the Kenyan towns of Mandera and El-Wak, located on or near the border with Somalia. By mid-year, more than 300,000 Somali refugees had entered Kenya and their number increased assistance the rate of 1,000 per day.

(7). UNHCR and the Kenyan authorities adopted a strict assistance policy assistance the border sites to deter nationals from posing as refugees. Refugees crossing to Liboi were screened upon arrival. In the Mandera and El-Wak areas UNHCR had hoped that a minimalist approach would reduce the attraction caused by relief distribution while suitable camp sites could be identified farther from the border. However, refugees were not moved from the border area because the Kenyan government feared that they pose a threat to the region's stability. Furthermore, assistance in the border sites was soon significantly increased after some relief agencies and the media criticized UNHCR's approach.

(8). The UNHCR Cross-Border Operation from Kenya into Somalia was planned in response to the population movements and refugee flows of spring 1992. The operation aimed at stabilizing the Somali communities inside their country to curtail the refugee influx to Kenya. UNHCR planners also hoped to prevent further influxes and to establish in Somalia conditions conducive to the voluntary repatriation of refugees.

(9). Rehabilitation Quick Impact Projects (QIPs) were selected as the main tactical instrument of the operation in Somalia. Beginning with fall 1992 cross-border field teams were deployed assistance the Mandera and El-Wak locations in Kenya. The teams made exploratory incursions into the Gedo region, and soon followed with the fielding of QIPs.

(10). The establishment in Somalia of the first UNHCR outposts, in early 1993, coincided with the deployment of the Unified Task Force (UNITAF) with had landed earlier in Mogadishu. UNITAF, and later the military contingent of the United Nations Operations in Somalia (UNOSOM), deployed units in the major towns of the Gedo region. It also came soon after ICRC ended the cross-border feeding programme that it had operated from El-Wak to stem potential refugee outflows in south Gedo.

(11). Implementing the Cross-Border Operation has required the quick build-up of a much larger staff than that of the Branch Office in Kenya. At the peak of the operation, in late summer 1993, some 65 international and 30 local staff worked in field outposts or in the management unit in Nairobi. The end of the programme in the Gedo region and the closure of field offices in 1994 brought the total number down to less than a dozen.

(12). As of April 1994 the UNHCR Cross-Border Operation had cost a total of USD 35.6 million. From end 1992 to fall 1994 some 360 QIPs worth a total of USD 7.9 million were launched. During the same period UNHCR assisted the voluntary repatriation of about 58,000 refugees assistance a cost of USD 11 million.

(13). Funding objectives have not been met in 1994. Donor response to the USD 14,5 million appeal for the cross-border programme has amounted only to USD 5,1 million. Donor "fatigue" has been caused by the quick deterioration of the general situation in Somalia. Observers predict that a reversal of the progress toward rehabilitation made in the last two years will occur after UNOSOM's scheduled withdrawal. Furthermore, the attention of donor states has switched from the Horn of Africa to the Rwanda emergency. In this unfavourable context, UNHCR is seeking funds to complete the repatriation of some 50,000 Somali refugees who wish to leave south Kenya for their home grounds in the Juba and Kismayo regions.

OVERVIEW

THE CRISIS WAS QUICKLY RESOLVED

(14). The Cross-Border Operation from Kenya into Somalia has shown that UNHCR could by means of proactive programmes seek solutions to refugee crises even before they had fully developed. The operation succeeded in quickly stemming the flow of refugees from the Somali region of Gedo into Kenya. It has also so far prevented further influxes and contributed to create in south-west Somalia conditions that encouraged refugees to return to their country.

(15). The operation brought a solution satisfactory to both the refugees and the assistance community. Less than two years after the beginning of the operation, most of the refugees who had crossed into north-east Kenya had returned to Somalis, and the border sites where they had regrouped were closed. The alternative to prolonged exile UNHCR offered to Somali communities through the rehabilitation of their hometowns had spared refugees the debilitating effects of camp life. The voluntary repatriation of Somali refugees also significantly alleviated the burden placed upon the Government of Kenya and the international community by costly care and maintenance programmes.

(16). UNHCR cross-border rehabilitation activities have had an immediate impact on the Somali communities. The choice of QIPs as rehabilitation tools has allowed the operation's planners to act with speed and flexibility. The first QIPs were launched early in the operation in areas where UNHCR expected refugees to return. While the rehabilitation effects of the programme were not immediately perceptible, QIPs quickly attracted the attention of Somali populations whether refugees, displaced or residents. As a result, spontaneous returns from the Kenya border sites or from other regions in Somalia became noticeable soon after the fielding of the cross-border teams.

THE OPERATION MAY NOT HAVE A LASTING EFFECT

(17). The achievements of the operation may be short-lived, however, for a number of reasons that include the uncertainty prevailing in Somalia and the absence of adequate programme planning.

(18). First, conditions in Somalia have started deteriorating again in the second half of 1994. The peace initiatives that brought Somali warlords together failed to produce government structures. Tensions between various clans have increased, and observers foresee a reversal to an all-out civil war after the withdrawal of the United Nations' military contingent in April 1995.

(19). Second, most of the rehabilitation activities initiated under the Cross-Border Operation are not self-sustainable or will not be sustained by generally unmotivated local communities. Pressed by the scale of the influx into Kenya, UNHCR planners have primarily invoked the short-term attraction power of QIPs on the Somali populations rather than their lasting rehabilitation impact. On one hand Somalis reacted encouragingly to the QIPs because they saw the immediate benefits that were available through project generated employment and contracts. On the other hand, the pressure of impact-driven activities and managerial constraints within UNHCR and its international implementing partners hampered project planning. Many QIPs were quickly designed by the field teams, and have lost their effect after the project implementation phase. Furthermore, activities which need the continued input of local communities after project completion are likely to be abandoned by a population concerned essentially with immediate survival.

(20). Third, the operation's planners did not succeed in coopting international agencies capable of consolidating and developing the activities initiated by UNHCR. The integrated cross-border and cross-mandate approach to the emergency in the Horn of Africa envisaged by the international community failed to materialize. Designated as the lead agency in the Cross-Border Operation from Kenya into Somalia, UNHCR in effect launched the programme alone.

AN ALTERNATIVE STRATEGY MUST NOW BE DEVELOPED

(21). In the future, UNHCR cross-border activities in Somalia should be limited to the preparation of repatriation movements. The QIPs phase of the operation has in fact ended with the closure of the UNHCR outposts in Gedo and with the inability of the teams to operate in the Juba regions. Continuing rehabilitation QIPs in the current Somali imbroglio would neither serve effectively the needs of the communities nor facilitate voluntary repatriation.

(22). An encashment programme for the benefit of the many refugees in the Kenyan coastal camps who have decided to repatriate to Somalia may be the most appropriate form of action. The experience of Gedo has shown that QIPs were more valued by the communities for the resources brought in during implementation than for their actual rehabilitation achievements. QIPs, however, were continued by UNHCR because they were poles of attraction for populations on the move. In the Juba and Kismayo regions the communities are made up of either sedentary rural or semi-urban populations .Without a sustainable rehabilitation impact on these communities QIPs would not be useful.

(23). A minimum UNHCR presence will continue to be needed in Somalia for coordinating voluntary repatriation movements. Experienced staff should be deployed both in Kismayo and in Mogadishu to prepare repatriation movements in consultation with the local forms of authority. As much as conditions permit, the monitoring of returns should be continued.

(24). Securing funds for the operation may be easier information rehabilitation objectives are left out. Donors' "fatigue" with Somalia and the shifting of funding priorities towards Central Africa have been the main cause of a serious funding shortfall in 1994. The lack of a solid rationale for continuing the QIPs programme may, however, be another cause. Major donor states officials view with scepticism the continuation of rehabilitation activities in Somalia, but approve of UNHCR's support to voluntary repatriation.

OPERATIONAL FEATURES ARE REPLICABLE

(25). Many of the initiatives taken under the Cross-Border Operation from Kenya into Somalia can be qualified as daring. The programme has quickly achieved the important objectives set by its planners. It has been based on an innovative policy approach and a pragmatic strategy. Clearly, a comprehensive approach to refugee crises founded on prevention and the non-differentiation of mandates is increasingly required in the type of situations UNHCR is facing today.

(26). Observers generally agree that in Somalia the cross-border approach was put to test under some of the worst imaginable conditions. Many of the operation's organizational components are sound blueprints that require some adjusting to be replicated elsewhere. The creation of a cross-border unit, the organization of mobile field teams and the establishment of temporary outposts in Somalia have allowed for quick and flexible action. In future cross-border programmes the Organization could improve its response by giving special attention to planning and policy input, to coordination with field teams and to the selection experienced staff.

ORIGINS AND IMPACT OF THE CROSS-BORDER OPERATION

(27). The operation had many positive achievements. Most importantly, it made possible the voluntary repatriation of Somali refugees and the closure of the border sites in the Northeast Province of Kenya. This approach spared refugees years of camp life under the relatively poor conditions prevailing in that region. It also eased the burden forced upon the Government of Kenya and the international community by the events in Somalia. Finally, rehabilitation activities under the Cross-Border Operation from Kenya into Somalia contributed to stabilizing populations in the Gedo region, and assistance least in the short term to the prevention of further outflows.

INNOVATIVE APPROACHES WERE PUT TO PRACTICE

(28). The operation adopted a number of innovative approaches to humanitarian assistance. Among those, the most important is the Preventive Zone Concept. This approach aimed at stabilizing, in selected zones covering both sides of a border, populations that included refugees, displaced persons and victims of drought. In addition, it was meant to prevent further refugee flows and to prepare conditions conducive to voluntary repatriation. The innovative character of the approach was further enhanced by the absence in Somalia of government structures, and the prevailing condition of civil strife.

(29). The Cross-border Operation from Kenya into Somalia quickly evolved as a self-contained UNHCR programme. Initially, it had been envisaged as part of a broad cross-mandate and cross-border strategy aimed assistance two preventive zones: one extending from Kenya into Ethiopia, the other one from Kenya into Somalia. The humanitarian assistance framework developed in meetings that brought together representatives of all major United Nations agencies, called for the harmonization of programmes across borders and among agencies, and for the non-discrimination among categories in need. The cross-mandate element, however, failed to materialize in Kenya and Somalia. As a result, UNHCR, which was designated the lead agency

(30). Operating across the border in countries of origin is not entirely new to UNHCR. In the late 1980s UNHCR had launched a number of rehabilitation projects from Pakistan into non-government controlled areas in Afghanistan. These projects had substantially contributed to the establishment of conditions favourable to the return and reintegration of Afghan refugees in these areas.

THE CROSS-BORDER STRATEGY WAS TAKEN A STEP FURTHER

(31). There are nevertheless major differences between UNHCR cross-border activities in Afghanistan and the Kenya to Somalia operation. In the case of the Afghanistan rehabilitation programmes, UNHCR's inability to establish a working presence in the target areas and the existence of highly developed logistic capabilities in Pakistan imposed the cross-border formula. In Kenya and Somalia, however, the cross-border characteristics were determined from the start through the design of preventive zones of uniform humanitarian assistance stretching across the border. Furthermore, UNHCR staff was directly involved on the sites of projects, and their presence was consolidated through the establishment of compounds.

(32). In previous cross-border operations such as those in Afghanistan, or in Operation Lifeline to Sudan, UNHCR's activities had made up for only fractions of much larger international efforts. In the Kenya to Somalia operation, UNHCR was the lead agency and the only one able to deploy quickly on the Somali side of the border.

PREVENTION, REPATRIATION AND PROTECTION

PROTECTION COULD HAVE BEEN BETTER DEFINED

(33). Although the Cross-Border Operation has undeniably contributed to the protection of the refugees who returned to Somalia, there has been little debate about how this mandate should have been translated into practice. Planners have been mainly concerned with social and economic rehabilitation. The immediate survival needs of the communities, the issues of short-term integration and returnee aid and development have all been discussed. An assistance strategy was designed and implemented. Evidently, the preventive approach developed in the Cross-Border Operation has included protection components. However, a clear protection policy that world have addressed goals and defined means had not been spelled out.

(34). While involvement in countries of origin has been the subject of much debate within UNHCR, lessons drawn from past protection roles in programmes such as Afghanistan and Cambodia have not been applied. Clearly, UNHCR becomes increasingly involved in situations where it must engage de facto in community-based protection activities akin to human rights monitoring. The formulation of policy guidelines in that domain would assist operations planners in systematically meeting the protection needs of the populations UNHCR is assisting.

(35). In the context of the cross-border Operation, deploying field teams and encouraging international NGOs to step-in has amounted to "protection through presence". Field officers have monitored the return of refugees and the displaced. They have also interceded with the warlords and other local forms of government. Nonetheless, the attribution of specific protection roles would have enabled UNHCR not only to protect the populations better, but also to measure better its achievements for the benefit of overall planning.

PREVENTIVE GOALS WERE REACHED

(36). The prevention of further refugee flows into Kenya has been the main achievement of the Cross-Border Operation. Stabilizing the populations present in the Gedo region was important in order to break the cycle of displacements which caused refugee outflows. Preventing the movements of populations in Somalia was also an ambitious undertaking because since the outbreak of the civil war most of the Somalis have been on the move. Many had fled armed conflict by moving to their ancestral territories, but then were rejected by the local population, or driven farther by rival clans.

THE OPERATION FOCUSED ON THE COUNTRY OF ORIGIN

(37). The approach taken by UNHCR aimed assistance countering the attraction that resulted from the dispensing of assistance to refugees in Kenya through relief delivery inside Somalia. Moreover, the activities under the Cross-Border Operation sought to assist local populations in restoring basic community structures in their places of origin. As a result, Somali populations were spared the debilitating effects of long-term care and maintenance they would have faced had they moved to refugee camps.

(38). UNHCR planners favoured community-targeted aid in Somalia over individual assistance hoping that community projects would restore socio-economic interdependence and encourage populations to settle. In addition, such projects assisted in the rehabilitation of the infrastructures that had been destroyed during the civil war. Finally, the implementation of community-geared activities required the development of some basic forms of representation, and thus also contributed to reinstate some degree of law and order in a previously anarchic environment.

(39). The deployment of field teams, and the establishment of compounds, in major towns of the Gedo region assistance the beginning of 1993 provided UNHCR with the operational capabilities required for implementing the cross-border approach. Logistical bases that world cater rehabilitation projects were set up furthermore, the permanent presence of staff allowed for close interaction with local populations on the issues of rehabilitation and community development. Relations were developed with representatives of the local communities, and the organization of local implementing partners was encouraged.

CONTINUED INTERCLAN FEUDING HAMPERED THE OPERATION

(40). Finding truly representative interlocutors was not always an easy task. UNOSOM whose mandate included assisting with the restoration of basic government structures sought to organize town councils throughout Somalia. However, it took much time and internal struggling before these councils could establish their legitimacy and become suitable counterparts. The case of the Garba Harre District illustrates how painful that process has been. Staff had to intrench themselves in the UNHCR compound for four days while heavy fighting between various Somali factions raged on the town's streets in support of their bids for council seats.

(41). A leading impediment to the cross-border efforts has been the large gap between the international community's objectives and the expectations of the Somali population. Most international staff involved in the Cross-Border Operation felt that the Somalis valued the income they derived from implementing rehabilitation projects, more than the return to normal living conditions these projects aimed assistance achieving. In effect, the UNHCR operation was widely regarded by the Somalis as a mere vehicle for cash and marketable goods. Clearly, in such a context the preventive impact of the Cross Border Operation appears somewhat ephemeral.

(42). Major donor states' officials, as well as UNHCR and NGO staff, generally agree that work carried out under the Cross-Border Operation has significantly contributed to restoring stability in the Gedo region and prevented refugee outflows. Notwithstanding, they remain sceptical of its long-term impact. One commonly shared opinion is that no assistance will carry sustainable effects until the Somalis themselves overcome interclan feuding and start rebuilding a civil society.

(43). The cross-border activities of UNHCR and its international partners in south-west Somalia exceeded their preventive objectives and contributed to attract refugees back home. More than 90,000 refugees returned voluntarily to Somalia since the beginning of the operation, most of them spontaneously. The prospect of improving security conditions in the Gedo region, poor living conditions and the lack of economic opportunities in the Kenyan border sites have been the main factors of return.

REPATRIATION STARTED UNEXPECTEDLY

(44). Repatriation started as a trickle assistance the end of 1992, and reached its peak during the summer of 1993. By mid-1994 Somali refugees had left the border sites in Kenya to return primarily to the Gedo region, and many of the residents of the coastal camps had also registered for repatriation.

(45). Assistance its inception the Cross-Border Operation was not aimed assistance facilitating the return of Somali refugees from the Kenyan border camps. UNHCR and international NGOs agreed that conditions in Somalia were so precarious that voluntary repatriation could not be encouraged or supported. Repatriation was considered a distant future option. However, assistance the time the cross-border teams stared to operate, field staff noticed that a number of refugees were leaving the camps for Somalia. Subsequently, UNHCR opened registration and developed a repatriation programme.

(46). Initially, the idea that UNHCR would launch a repatriation programme was not well received by various groups within the Organization itself. Some officers of the Division of International Protection believed that repatriation to war-torn Somalia would be premature because the minimum standards of protection generally sought by UNHCR for repatriants could not be ensured. In the field, staff involved in the care and maintenance programmes thought that refugees were not ready to move out of the camps.

ORGANIZING REPATRIATION WAS DIFFICULT

(47). Preparing repatriation movements to the Gedo region has not been an easy task for UNHCR personnel, both in the Kenyan camps and in the areas of return. Weak community structures in the camps made participative planning difficult because of constant changes in refugee representation. In Somalia, securing the agreement of the receiving communities required lengthy negotiations with groups of elders primarily concerned with the material benefits they believed could be derived from cooperating with UNHCR.

(48). Some staff also point out that the lack of cooperation between the UNHCR Sub-Offices and the cross-border field units in the early planning stages of repatriation delayed the process. By the end of 1992, cross-border teams began undertaking missions to the Gedo region that brought them in communities. The teams, however, did not share their plans or findings with Sub-Office staff, and separate missions had to be organized to assess the feasibility of repatriation.

(49). Major logistical problems have plagued repatriation. Instead of assuming a supportive role, UNHCR has been directly involved in arranging the logistics of repatriation movements. Selecting dependable transport from the many offers pressed upon UNHCR by rival clans was a first obstacle. Convoys would be sabotaged information all the parties did not find their expected benefits. The, ensuring the safe passage of convoys through territories controlled by local warlords was an additional challenge. Typically, each warlord on the repatriation route would convincingly offer to protect the convoys in exchange for some retribution.

(50). Escort arrangements made with UNOSOM somewhat alleviated the security threats encountered on the repatriation routes. However, UNOSOM had its own mind on the pace of rehabilitation in Somalia and viewed with scepticism early repatriation movements. Political Affaires staff thought that tangible institution-building should be completed before repatriation could start. As a result, UNOSOM would respond selectively to Anchor's requests for convoy protection, and alternative arrangements had to be negotiated.

(51). Logistical problems were further exacerbated by the unpredictability of the refugees themselves. Repeated no-show of candidates to repatriation assistance the agreed departure times caused UNHCR to pay large standby fees to the transporters. This has been particularly disastrous in the context of the attempted airlift to Somalia, and also made escort arrangements with UNOSOM more difficult to negotiate.

(52). UNHCR field teams often felt that Nairobi and headquarters staff could not fully appreciate the magnitude of the challenges involved in organizing voluntary repatriation to Somalia. Frustration was expressed at critiques of over-expending against the transport budgets. Staff argued that ensuring the safe return of refugees under the anarchic conditions prevailing in Somalia left them with no alternative but buying off protection from local warlords at the going rates.

(53). There are contending views among UNHCR staff of the bearing of individual assistance on the refugees' willingness to repatriate. Staff who had been involved with the care and maintenance programmes in the Kenya border sites tend to argue that the abundance of food in the camps kept refugees from returning home. On the other hand, UNHCR officers who took part in the planning of the Cross-Border Operation minimized the delays caused by the food factor and contend that the number of refugees wishing to repatriate has grown in consistent relation with the pace of normalization in the Geed region.

RETURNEE AID AND REHABILITATION

(54). The rehabilitation activities carried out in Somalia under the Cross-Border Operation have significantly contributed to the revival of the Gedo region. Somalis saw in the programme both the opportunity to reconstruct their towns and a chance to derive immediate benefits from the resources brought in by international organizations. It is arguable that the actual rehabilitation of community structures could have been better served by a well planned multi-sectoral programme than by QIPs. The prevailing sense of urgency given the situation in Somalia, however, compelled planners to somewhat forfeit programme design quality in return for quick deployment.

QIPS WERE EXTENSIVELY USED

(55). A main feature of the Cross-Border Operation has been the exclusive use of QIPs for aid and rehabilitation purposes in south-west Somalia. From the operation's inception until the recent closure of UNHCR outposts in the Gedo and Juba regions, the administration of QIPs has been the principal occupation of all field teams. At present, international NGOs who continue activities initiated by UNHCR still process QIP proposals, and so do UNHCR field staff in the coastal region of Kismayo.

(56). Undoubtedly, UNHCR-funded QIPs have significantly contributed to restoring locally acceptable living standards in the Gedo region. Rehabilitation efforts have targeted all major community infrastructures and services, as well as the traditional occupational sectors. Activities have ranged from the rehabilitation of health centres to chicken farming, and have included such varieties as the building of camel slaughter slabs, the clearing of bush roads or metal piping collection. In some instances QIPs have provided the region's populations with elements of comfort that had been unknown in the pre-war era.

REHABILITATION PRECEDED REPATRIATION

(57). Contrary to previous situations, QIPs were launched in Somalia before repatriation could be envisaged. Since the formulation of the Quick Impact Project approach in Nicaragua, the QIP has been regarded as the leading model of post-return assistance capable of providing returnee-populated communities with timely reintegration assistance. The QIP formula has gained wide recognition and support both within UNHCR and among donor states. In Nicaragua, UNHCR sought to consolidate the completed repatriation process by addressing the immediate needs of the communities where returnees had settled. In that case, returnees had already received an initial relief package, and communities in areas of return had benefited from a board rehabilitation programme.

(58). In the context of the Cross-Border Operation, QIPs aimed initially at stabilizing populations within Somalia and at preventing further refugee outflows to Kenya. UNHCR planners had been primarily concerned with the flow of refugees into the Kenyan border sites, the camps' limited absorption capacity and the reluctance of the Government of Kenya to accommodate a large fleeing population. At that stage repatriation was only a remote possibility.

(59). The QIP approach was selected by the planners of the operation because they felt that the unfolding of events in Somalia required an immediate response. QIP programmes are modular in structure and therefore appear flexible. They also imply limited commitments and thus limited financial risks. As a result, the lead time from programme planning to implementation could be reduced to a minimum.

CONDITIONS PREVENTED COMPREHENSIVE PLANNING

(60). A comprehensive rehabilitation programme might have been more desirable, but conditions in Somalia would not have permitted its implementation. Some critics of the Cross-Border Operation have suggested that an integrated muti-sectoral approach would have been more effective and less costly. At present indeed, with relatively free access to the Gedo region and its communities, aid officials can assess rehabilitation needs with the precision required for the designing of a comprehensive programme. This was not the case at earlier stages.

(61). At the beginning of the operation, a number of factors impeded the formulation of extensive needs assessments. First, high degree of insecurity prevented the cross-border teams from covering more than a few locations in the Gedo region. Only later, several district capitals could be reached by air and from these places quick incursions were made into less accessible rural areas. Second, the few attempts mad eat collecting data were often misunderstood by the villagers who expected money and materials to come with the teams. In further visits to the same areas the data collectors would de greeted with such animosity and threats that assessment missions could not proceed. Third, throughout the implementation period, no local assessment capacity could be relied upon beyond that of the Somali elders committees.

(62). While conditions gradually improved with the constitution of local forms of government and the emergence of local NGOs, the quality of the information feedback received from the field outposts did not meet the standards required for comprehensive planning. This shortcoming was partly due to the significant pressure for immediate delivery the communities put on field teams, and partly to insufficient guidance and training. Almost a year after the beginning of the operation, several sets of guidelines were issued to the field. However, instructions mostly spelled out submission and reporting criteria to be applied to QIPs, bur remained weak on the subject of needs assessment, the evaluation of the communities' wants and of the projects' contributions to overall rehabilitation objectives.

ATTEMPTS AT PLANNING CAME LATE

(63). The establishment within the Cross-Border Operation Unit of a Technical Team aimed at capitalizing on QIPs' experiences with a view to develop a comprehensive rehabilitation plan. The team comprised Sectoral specialists seconded to UNHCR from international NGOs. Sectoral guidelines were issued by the end of 1993, and distributed to all the outposts and concerned NGOs. Notwithstanding, the Technical Team field to formulate a comprehensive plan primarily because it was not able to find, or interface with, local institutions and international agencies committed to integrated multi-sectoral rehabilitation work. The team was eventually dissolved in the summer of 1994.

(64). There are different views on whether QIPs could have been better coordinated. Some cross-border staff have referred to an unavoidable " scattered and ad-hoc QIP s phase " that should have preceded a " reconstruction and development phase ". Other staff members have suggested that there has been an over-zealous implementation of the programme. It is likely that the operation's planners did not feel a pressing need for coordination because they were less concerned with actual rehabilitation achievements than with fixing down populations and encouraging repatriation. The disparities shown between rehabilitation levels in various locations at the closure of the programme would confirm that the QIPs were mere tactical instruments.

(65). In the absence of region-wide plans, the degree to which QIPs effectively contributed to rehabilitation was largely determined by the field teams' common sense, and by the pressures applied by Somali individuals and groups on these teams. Overall, QIPs have allowed for the proper rehabilitation of many essential services and structures. Nonetheless, in some areas, where teams had formulated and implemented well-structured plans, the return of fewer people than had been anticipated resulted in over-assistance and dependency. In other areas, the selection and approval of QIPs was somewhat dictated by security threats and contributed to the local warlords' version of rehabilitation activities.

THE RESPONSE OF SOMALI COMMUNITIES PROVED WEAK

(66). The rehabilitation effect of individual QIPs in Somalia has not been commensurate with the magnitude of the resources committed because the communities were weak or non-existent. In recent years, UNHCR programmes have gradually evolved from individual support to community assistance. Aid planners generally believe that the collective approach boosts beneficiary involvement and reduces the risk of seeing the assisted population fall into a state of perpetual dependence. Planners work on the premise that individuals inevitably organize them selves into social communities.

(67). For a number of reasons, Somali communities have proven much weaker than was anticipated by aid planners. First, the Somalis' semi-nomadic tradition makes for an extremely fragmented society. Second, Islamic tradition generally excludes women from direct participation in community decisions. Third, the civil war has forced many Somalis to seek refuge in ancestral areas of abode they had never seen before. These in country migrations often resulted in conflicts with the resident populations. Finally, the clan structure of Somali society has prevented new communities from forming or functioning.

(68). The impact of the Cross-Border Operation on community rehabilitation could have been enhanced had more QIPs been gender-sensitive. Aid workers have often observed that Somali women do not share the men's confrontational attitudes, and instead are genuinely committed to improving their families' 'ell being. Furthermore, because women carry out all of the domestic chores and most of the subsistence activities, they are natural partners for community rehabilitation work. While administrating QIPs with male partner was generally described by field workers as endless negotiations exercises to determine precisely who would get what, so men were seen as relatively pragmatic and efficient counterparts.

LOCAL NGO'S EMERGED QUICKLY

(69). The International Rescue Committee (IRC) has been anchor's main implementing partner for the cross border QIP programme. Throughout the operation, however, IRC has in turn come to rely increasingly on Somali NGOs. IRC field teams in Somalia identified needs, developed responses and submitted projects proposals to UNHCR. In the process IRC teams contracted local labour and services or sought out local counterparts who would implement the QIPs. Somalis soon understood that they could organize themselves in a manner that would make them eligible for the NGO label, propose rehabilitation activities and derive cash benefits from project implementation.

(70). The emergence of many Somali NGOs in response to the QIP programme has been a positive factor in the operation because it spared international agencies some of the interminable negotiating required for direct implementation. Local NGOs, often made up of only a few individuals, knew better than expatriate staff how to hire and purchase the resources required for the QIPs without disturbing the local balance of power. Furthermore, they were able to enrol local communities more effectively than international aid workers could. Despite many misgivings with regard to accountability, local NGOs have played an important role in bringing fragmented communities to negotiate and even to cooperate.

MANY QIPS HAD UNINTENDED EFFECTS

(71). Though many QIPs have achieved the rehabilitation goals stated in the project descriptions, they sometimes appear to outside observers of no value to the community. Among the most visible examples are the covered markets that were built and remain unused in several locations. Similarly, some schools are squatted by displaced families, and water hand-pumps are removed to give place to the traditional open wells.

(72). Notwithstanding, such seemingly failed projects are always shown as important contributions to the communities by the residents who had been associated with their implementation. In effect, many QIPs are valued by the local populations more for the fringe benefits that could be derived during the implementation phase than for their final outcome. Thus, the construction of schools and markets has provided a number of residents with substantial wages and procurement contracts. As for water pumps, the metal was sold as scrap or used by blacksmiths for the confection of axes and knives.

(73). This unintended impact of QIPs was not negligible. Wages and contracts have fueled local cash-starving economies with the resources needed for a return to basic production and trade standards. The benefits were then distributed among members of the communities in accordance with the traditional social structure. Ultimately, the effects of these alternative circuits would also contribute to the overall objectives of the Cross-Border Operation: namely the stabilization of populations inside Somalia.

REHABILITATION COSTS WERE HIGH

(74). The costs of project delivery in Somalia have run significantly higher than in any other part of the world. Furthermore, the projects' end products have been valued by the beneficiaries at only a fraction of the value ascribed to them in anchor's and the donor states' books. The example of the hand-pumps illustrates it well. Forged hand-pumps are relatively more costly to donor states than were worth only their weight. Viewed from this angle, the QIP approach was not cost-effective.

(75). Opinions differ on whether a QIP-only rehabilitation programme was the correct approach in Somalia. Some staff point out that making cash and materials available to the communities in the Gedo would have supported the return to normality as much as the QIPs did. In addition the communities would have been able to manage the region's rehabilitation in accordance with their own priorities. On the other hand, many field staff contend that local decision-makers would have under-prioritized vital care services and structures. Since most of the QIPs are not self-sustainable, services such as health, sanitation and education are likely to stop as soon as external funding is cut.

CASH AND RESSOURCES COULD HAVE BEEN AN ALTERNATIVE

(76). It is likely that many of the Cross-Border Operation's QIPs could have been efficiently replaced by various cash-dispensing schemes. International aid and development agencies' experts familiar with the prevailing conditions I south-west Somalia concur that restoring some degree of normalcy requires primarily the injection of cash and resources. Because of the scarcity of the country's natural resources, Somalis have always been traders. Before the civil war, economic activity in the Gedo region centered on the commerce of goods bought in Kenya and delivered to the markets in Mogadishu. In addition, local economies derived cash from foreign aid that trickled down the intricate clan and government networks. With war-time destruction and looting, the Somalis' purchasing power has dropped to minimum levels; markets have collapsed and merchants have lost their capital investments.

(77). Major aid agencies have encouraged the creation of networks of credit unions as a means to increase the availability of funds to Somali entrepreneurs. With funds allocated under the USAID food monetization programme for Somalia, the IRC has established credit unions in several major towns of the Gedo region. The credit unions are run by local boards of directors. They function as lending institutions with the difference that they have no shareholders, and thus profits gained from the payment of interests on loans are reinvested in the union's fund. Furthermore, successful loan recipients are bound to pay a small percentage of their operations' profits to the fund as well.

(78). Despite initial scepticism, observers agree that the credit unions experiment has been highly successful. Aid workers who had been involved in assistance to Somalis generally doubted that a system calling on the beneficiaries' sense of accountability would function. This assumption, however, overlooked the extreme stringency of the codes of ownership and behaviour that determine relations between Somalis. In fact, because in the credit unions scheme individuals are accountable to their own community, represented by the local board of directors and other borrowers, the default rate has been low.

(79). UNHCR was not able to consider such unconventional alternatives, because staff involved in the Cross-Border Operation had neither the planning nor the implementing capacity required. Whereas QIPs generally entail fairly simple assessments, some procurement work and some elementary wages and contracts administration, establishing projects like the credit unions requires systems development skills. Evidently, the Operation's planners selected what they saw as a quickly fielded and easily managed programme option.

(80). There is no doubt that the benefits of the QIP programme largely outweigh its costs. The alternative of long-term care and maintenance programmes and continued ordeal for refugees in the Kenyan border sites have been achieved more efficiently through better programme planning and management. QIPs have been particularly costly in Somalia because they were meant to serve a serve a wider range of goals than simple rehabilitation ones. In the process, the planners of the operation have knowingly traded off some of the generally accepted standards of project management for maximum delivery speed and impact.

INTERAGENCY COORDINATION AND THE CONTINUUM

(81). Relations with other agencies involved in the Somalia emergency have brought only marginal benefits to the Cross-Border Operation. At best, UNHCR obtained limited operational support from some, while at worst, the Organization found itself accused of "crimes against humanity". Designated by the Security Council as the lead agency for the operation, UNHCR has carried out the cross-border programme on its own. The weakness of its implementing partners has prevented UNHCR from developing a sustainable rehabilitation plan to be continued by other agencies.

OPERATIONAL CONSTRAINTS HAMPERED INTEGRATION

(82). Coordination with various United Nations agencies was more effective I the area of fund raising than in the planning and implementation of programmes. The Special Emergency Programme for the Horn of Africa (SEPHA) appeals consolidated by DHA presented convincingly the integrated response United Nations agencies operating in the Horn of Africa intended to develop in face of the emergency. A UNHCR summary paper on the Preventive Zone Concept emphasized the need of an integrated approach to the Somali crisis, and envisaged the creation of joint management structures. On the specific issue of cross-border activities in Somalia, the paper foresaw joint field teams composed of UNHCR, WFP, UNICEF, NGO staff and other UN agencies if necessary. Notwithstanding, the integrated approach did not materialize in the Kenya to Somalia Cross-Border Operation.

(83). Generally, the agencies which participated in the SEPHA lacked the operational capacity to commit themselves to the operation. Furthermore, their programme priorities often did not match the operation's strategy. As a result, the lead role conferred to UNHCR by the Security Council was welcomed by these agencies who saw this as an opportunity to step back rather than revise their established programmes.

(84). The creation in Nairobi of a Disaster Management Team (DMT) under the auspices of UNDP, and linked to the United Nations Special Envoy to Somalia, was a genuine effort at strengthening interagency coordination. The DMT was intended to coordinate both relief and development activities in Somalia. All major United Nations agencies are represented on the team which meets regularly. UNHCR has often used this structure to keep the United Nations community abreast of developments in the Cross-Border Operation. In return, the benefits the Organization has derived from the DMT have been only marginal. Furthermore, UNHCR staff contend that the team was set up too late to effectively coordinate already well engaged programmes.

(85). The efforts deployed by the DMT to bring an integrated response of some scale to the relief needs at the Kenyan border sites have not been matched in Somalia. The fact that all agencies aside from UNHCR covered Somalia until recently from branch offices in Mogadishu has prevented their representatives in Nairobi from making any significant contribution to the Cross-Border Operation. On the other hand, agency staff in Mogadishu could neither operate effectively in south-west Somalia, nor manage the integration of their activities with UNHCR's

WFP PROVIDED VALUABLE SUPPORT

(86). Despite a controversy over food for work programmes, cooperation with WFP was generally effective in the Gedo region. WFP had opened offices in several major towns at about the same time UNHCR established its cross-border outposts. Its main activities included providing supplementary food for vulnerable groups as well as food for work. The presence of another United Nations relief agency in the region also reinforced the stabilizing effect of international civilian assistance. Nonetheless, WFP's food for work programme has sometimes been questioned. While it is generally admitted that food for work is a valuable complement to rehabilitation efforts, critics contended that in the Gedo region, excess food derived by the population from generous food derived by the population from generous food for work schemes saturated the local markets and discouraged cash crops farming.

UNOSOM WAS A MIXED BLESSING

(87). Coordination with UNOSOM in the field has been somewhat frustrating to UNHCR teams. Although UNOSOM's mandate evolved from purely military to humanitarian objectives, its structure never gained the flexibility required for relief work. Furthermore, after its failure to disarm the Somali militias, UNOSOM fell back on simple military escort tasks while getting itself increasingly involved in political institutions development and rehabilitation activities. The scale of UNOSOM's contribution to rehabilitation has been negligible, but the reluctance of its Humanitarian Affairs staff to consult with UNHCR and the NGOs has undermined the negotiating of some QIPs with local communities.

(88). UNOSOM's paramount concern with security has also hampered the work of international relief agencies. Clearance for field staff movements would be withheld, or escorts not granted. Eventually, the scale of forces commonly displayed in UNOSOM operations was deemed threatening and was resented by the Somali population. Furthermore, because UNOSOM regarded with apprehension any movements of population. UNHCR-supported repatriation activities often became the object of tedious negotiations with post commanders and of referrals to the high Command.

RELATIONS WITH IRC PROVED AMBIGUOUS

(89). The choice of IRC as anchor's main implementing partner for the QIP programme was not planned. UNHCR first approached the Lutheran World Federation (LWF) . LWF had gained significant logistics experience in Somalia through organizing a food and relief air-lift. An agreement was expected to be reached, but LWF retracted at the last moment explaining that it did not have the capacity required for the operation.

(90). A number of problems have, from the outset, plagued anchor's cooperation with IRC. First, the agreement drawn with IRC did not clearly define the parties' roles and responsibilities. Second, the project management practice of IRC teams differed significantly from UNHCR's. Third the IRC team present at the time in Kenya was inexperienced and would have benefited from greater support and guidance from its headquarters. Fourth, the international staff IRC recruited in Nairobi for the Cross-Border Operation somewhat lacked maturity and experience.

(91). The agreement signed with IRC was flawed because it mislead IRC staff into believing that they had the same decision-making prerogatives as the UNHCR cross-border officers. The planners of the Cross-Border Operation had thought that conditions in Somalia called for fully integrated teams rather than the usual contractual relations between implementing partners and UNHCR. Thus, the first cross-border teams based in Kenya were headed jointly by UNHCR and IRC leaders. Differences in the approaches and methods of work of the two organizations quickly made such arrangements unworkable. The problem was eventually resolved with the creation in Gedo of UNHCR-only outposts.

(92). Resentment of what was perceived by IRC staff as the UNHCR bureaucracy hampered project management and the planning efforts of the Cross-Border Unit in Nairobi. IRC staff generally saw themselves as doers distrusted by their UNHCR counterparts. Partly due to their impatience with the QIPs approval and financial disbursement procedures, IRC staff have for long neglected project reporting. In turn, the lack of feedback has prevented UNHCR planners from properly assessing the effects of the QIP programme.

(93). Finally, IRC staff could have been selected more carefully for this delicate operation. The situation improved with the arrival of more qualified personnel, but the support provided by IRC headquarters has not matched the needs of the operation. Most of the IRC expatriate field workers were recruited in the region. Although generally dedicated to their new assignments, they lacked strong project planning and administration skills.

ROLES COULD HAVE BEEN BETTER DEFINED

(94). Much of the misgivings that characterized relations between the IRC and UNHCR could have been tackled through clearer definitions of objectives, roles and rules. The drafting of project submission and reporting guidelines to be shared with all implementing partners involved in the QIP programme illustrates that UNHCR planners have tried to address these issues. Notwithstanding, the pressure to "get things done now" has on the whole prevailed over careful operations' planning. Furthermore, since some of the difficulties with the IRC cross-border team resulted from the agency's internal organization, UNHCR could have raised its concerns more effectively with IRC headquarters.

(95). There has been a variety of views among planners and observers of the Cross-Border Operation on how the UNHCR rehabilitation efforts stand in the continuum from relief to development. While UNHCR has repeatedly spelled out its commitment to the continuum approach, operations planners find it somewhat difficult to apply in emergency environments. In the context of Somalia, the absence of central authorities and persistent warring between factions preclude any development initiative. Officials from major aid agencies point out that development programmes cannot be forced upon populations unwilling to cooperate even for the relief of their own communities. Aid planners speak of a "suspended continuum" in Somalia.

SUSTAINABILITY WAS NOT A PRIORITY

(96). The pressing developments that prompted the launching of the Cross-Border Operation understandably eclipsed anchor's concern with the continuum. Nonetheless, designing sustainable rehabilitation QIPs could have made them a meaningful contribution to future multi-sectoral development programmes. Along efforts to restore normal living conditions in the Gedo region, UNHCR cross-border teams launched a number of "quasi-developmental" QIPs. These initiatives aimed at self-sustainability and reached farther than the rehabilitation objectives of the operation. To the cross-border planners they were primarily justified by the need to consolidate progress towards the region's stabilization. Without necessarily engaging in development activities, more QIPs could have been geared to self-sustainable activities.

(97). Encouraging the creation of national aid and development institutions is necessary, but in a country fragmented into a multitude of clan-controlled areas such initiatives can be only viable locally. UNHCR has, through the QIP programme, encouraged the formation of a multitude of Somali NGOs. The support of an international NGO such as IRC has enabled Somali organizations to undertake work locally and sometimes in other towns. However, clan pressure seriously restricts the scope of their potential involvement.

(98). The rehabilitation achievements of the Cross-Border Operation may be short-lived because UNHCR was unable to involve at the outset international agencies capable and willing to continue work in Somalia after the Organization's withdrawal. The continuum approach requires the involvement of development institutions alongside UNHCR from the start. The commonly held view that UNHCR can field a programme and then "hand it over" to another agency should be questioned. International aid agencies are seldom willing to continue projects that were conceived and administered on standards which are not their own. As a result UNHCR often faces situations where it should but cannot disengage, or where it withdraws but faces the risk of seeing its achievements crumble.

FUTURE DIRECTIONS

(99). The Cross-Border Operation, as it was carried out until mid-1994, has in effect ended. With the achievement of anchor's main objectives in the north, the centre of gravity of the operation has shifted down to the Lower-Juba and Kismayo areas. Repatriation from the border sites in the Northeast Province of Lenya was completed in the first half of 1994, and the camps closed. By mid-year, returnees to the Gedo region had settled in a somewhat stable environment, and UNHCR cross-border teams were withdrawn. The compounds were handed over to international NGOs willing to continue rehabilitation work for some time, or to local forms of government. The situation has not evolved as favourably in the southern regions to which about 30,000 refugees residing in the coastal area of Kenya wish to return.

(100). Maintaining a permanent presence in Somalia to facilitate voluntary repatriation may be the last cross-border component still needed. Prevailing instability will not allow UNHCR to continue fielding QIPs in the southern regions. Even information conditions appeared more favourable, the experience of Gedo has shown that in Somalia QIPs are not an efficient means to achieve the rehabilitation of local communities. In Gedo, QIPs were used to attract semi-nomadic populations to their homelands or to discourage them from leaving. In many instances the rehabilitation effect of QIPs vanished after the project was completed or international assistance withdrawn.

(101). Continuing the Cross-Border Operation in the south by means of a QIP programme is neither feasible nor desirable. Little progress was made by the cross-border teams in the southern regions of Lower-Juba and Kismayo. Feuding between militias and uncontrolled violence have made the region so unsafe that teams had to be withdrawn or confined to perimeters protected by the UNOSOM military contingent. Monitoring the many agricultural QIPs implemented on the banks of the Juba river has proven impossible. In Kismayo, UNHCR staff are unable to maintain regular contact with the resident communities, or to develop working relationships with their representatives.

(102). QIPs as all other forms of assistance have become additional causes of conflict among the warring factions in the Kismayo region. Despite a local peace conference held in June 1997, warlords have not been able to cooperate. As a result, UNHCR and other agencies represented in Kismayo are constantly under the pressure of opposed parties eager to derive personal benefits from assistance programmes.

(103). The prospect that conditions would improve to allow international agencies to work effectively in the region is dim. A viable agreement between the warlords is not likely to be reached in the near future. Since the withdrawal of UNOSOM has been announced, tensions have increased. Observers predict that the militias are preparing themselves for the looting of the materials and structures that will be left by UNOSOM. In addition, the warlords will fight for the control of the maritime and air ports in which they foresee a substantial source of revenues through the taxing of all aid traffic.

(104). Continuing QIP s for the sake of encouraging voluntary repatriation from the Kenyan coastal camps to the southern regions is not necessary because the refugees are determined to return home. The use of QIPs as an incentive to repatriation would be would not be effective in a context where immediate survival is the prime concern. Furthermore, the populations of the Juba and Kismayo regions are either sedentary rural or semi-urban, and do not need to be stabilized by assistance.

(105). Many of the premises on which the Cross-Border Operation was developed are no longer valid. In summer 1992 planners pressed to stem the massive influx of refugees into Kenya decided on fielding assistance inside Somalia. Voluntary repatriation was not an immediate objective, and in fact UNHCR developed a repatriation programme only after refugees had began to return spontaneously to Somalia. At present, the refugees' desire to repatriate is the main issue UNHCR must address.

Recommendations

A. The operation's objectives should be redefined in light of the current situation and refugee needs. The Branch Office in Kenya should decide whether continuing a cross-border rehabilitation programme is still necessary, or whether a conventional repatriation project would better suit present needs.

B. The Branch Office in Kenya should step up the organization of voluntary repatriation to meet the pressing demands of the refugees. The official closure of the camps by the Kenyan authorities last October leaves refugees with the choice between repatriation and transfer to a closed camp in the desertic Dadaab area of the Northeast Province.

C. The Branch Office in Kenya should consider developing an encashment programme for those willing to return. Making cash available to returnees under the current conditions in Somalia is likely to have more of an impact on their well-being than rehabilitation activities. Encashment costs would be significantly lower than those of fielding projects. Many Somalis live of trade and would favour the cash option.

D. As security conditions in Somalia, a solid UNHCR presence in Kismayo and in Mogadishu should be maintained to ensure that repatriation movements are properly organized and monitored. Staff should be capable of developing and maintaining relations with representatives of the factions controlling the return areas. Negotiating with the local forms of government will be crucial after UNOSOM withdraws in spring 1995.

E. Re-focusing the Cross-Border Operation from rehabilitation to repatriation could require some organizational changes in the Nairobi Branch Office and in the field. It is conceivable that programme tasks pertaining to the closure of the QIPs could be assumed by the Programme Unit, while the organization of repatriation would be continued by the Protection Unit.

F. Funding appeals should clearly introduce a new and distinct repatriation programme. Major donor states officials in Nairobi are sceptical of the continuation of rehabilitation efforts in Somalia. An efficient alternative to QIPs, and a stronger emphasis on voluntary repatriation could make the appeals more convincing.